Aggregation of Perfluoroctanoate Salts Studied by 19F NMR and DFT

pubs.acs.org/Langmuir © 2010 American Chemical Society

Aggregation of Perfluoroctanoate Salts Studied by 19F NMR and DFT Calculations: Counterion Complexation, Poly(ethylene glycol) Addition, and Conformational Effects Gerardo Abbandonato,† Donata Catalano,* and Alberto Marini Dipartimento di Chimica e Chimica Industriale, Universit
a di Pisa, via Risorgimento 35, 56126 Pisa, Italy. † Present address: Scuola Normale Superiore, Piazza dei Cavalieri, 7 56126 Pisa, Italy. Received June 25, 2010. Revised Manuscript Received September 10, 2010 The aggregation of perfluoroctanoate salts in H2O is studied by 19F NMR on solutions of LiPFO, NaPFO, and CsPFO, without and with the addition of two poly(ethylene glycol) (PEG) oligomers of molecular weight 1500 and 3400 Da, respectively, and with the addition of suitable crown ethers. The 19F chemical shift (cs) trends are monitored, at 25 °C, in a concentration range including the critical micellar concentration (cmc) or, in the presence of PEG, the critical aggregation concentration (cac). The cac values in the samples with PEG are lower than the cmc values of the corresponding samples without PEG; moreover, the 19F cs trends above the cac and above the polymer saturation concentration reveal and help to explain some peculiarities of the aggregation process of PEG on PFO micelles, which, in the first step, seems to occur while the surfactant concentration in water is still increasing. Also in LiPFO/H2O or NaPFO/H2O solutions containing 12-crown-4 or 15-crown-5 ethers, suitable to complex Liþ or Naþ ions, respectively, the cmc decreases. On the other hand, the micellization process in the presence of crown ethers does not show other peculiarities. The prevailing conformations of the PFO chain are discussed on the basis of quantum-mechanical calculations. The theoretical chemical shifts were computed at the DFT level of theory, taking into account the effects of the environment by means of the IEF-PCM method. The helical structure is the most stable one, but anti conformations are easily accessible, in both the aqueous and fluorinated environment. The comparison between computed and experimental chemical shifts indicates that anti conformations are more important in the micelles than in water and in CsPFO micelles than in LiPFO or NaPFO ones.

Introduction Polymer-surfactant systems in aqueous solutions constitute an attractive research subject owing to their various applications.1,2 Among surfactants, fluorinated and perfluorinated compounds deserve special attention3 for many peculiar physical and chemical properties, which determine their massive industrial use, with consequent environmental and medical implications.4,5 Such properties are essentially related to the high conformational stiffness of fluorinated chains,6 to their highly hydrophobic nature, and, more generally, to the repellent behavior toward non-fluorinated compounds.3 In this paper we investigate the aggregation process of the perfluoroctanoate salts LiPFO, NaPFO, and CsPFO in water. A short summary of the features exhibited by the micelles of the three salt/ water binary mixtures can be useful for introducing our work. At a fixed temperature, the critical micellar concentration (cmc) depends on the counterion, being cmc(Liþ) ≈ cmc(Naþ) > cmc(Csþ).7,8 *Corresponding author: Tel þ39-050-2219266; Fax þ39-050-2219260, e-mail [email protected]. (1) Goddard, E. D. Colloids Surf. 1986, 19, 255. (2) Kwak, J. C. T., Ed.; Polymer-Surfactant Systems; Surfactant Science Series; Marcel Dekker, Inc.: New York, 1998; Vol. 77. (3) Kissa, E. Fluorinated Surfactants and Repellents; Surfactant Science Series; Marcel Dekker, Inc.: New York, 2001; Vol. 97. (4) Goecke, C. M.; Jarnot, B. M.; Reo, N. V. Chem. Res. Toxicol. 1992, 5, 512. (5) Kreckmann, K. H.; Sakr, C. J.; Leonard, R. C.; Dawson, B. J. J. Occup. Environ. Hyg. 2009, 6, 511. (6) Schwarz, R.; Seeling, J.; K€unnecke, B. Magn. Reson. Chem. 2004, 42, 512. (7) Gianni, P.; Barghini, A.; Bernazzani, L.; Mollica, V.; Pizzolla, P. J. Phys. Chem. B 2006, 110, 9121. (8) Gianni, P. A.; Bernazzani, L.; Guido, C. A.; Mollica, V. Thermochim. Acta 2006, 451, 73. (9) Hoffmann, H.; Kalus, J.; Thurn, H. Colloid Polym. Sci. 1983, 261, 1043. (10) Berr, S. S.; Jones, R. R. M. J. Phys. Chem. 1989, 93, 2555. (11) Everiss, E.; Tiddy, G. J. T.; Wheeler, B. A. J. Chem. Soc., Faraday Trans. 1 1976, 72, 1747.

16762 DOI: 10.1021/la102578k

LiPFO and NaPFO are known to form spherical micelles9,10 in diluted water solutions, in a wide range of concentrations11 above the cmc. On the other hand, the CsPFO micelles are oblate ellipsoids,12 which, at high surfactant concentration, form nematic and lamellar crystalline phases.13 However, there is also evidence that the first CsPFO aggregates just above the cmc are probably spherical.14,15 For the three studied surfactants, the aggregation number grows with increasing surfactant concentration, but it is higher and increases much more steeply in the ellipsoidal CsPFO micelles than in spherical NaPFO ones; the degree of counterion binding similarly grows for these two types of micelles, but it is higher for the CsPFO ones.10,12 The nature of the counterion only slightly affects both the micellization enthalpies and entropies,7,8,16 but the CsPFO micelles present slightly less positive aggregation enthalpy.7,8 This and many other peculiarities of the CsPFO/ water system can be related to the smaller radius of hydrated Csþ ion with respect to those of hydrated Naþ or Liþ, an argument often brought up when the counterion effect on surfactants behavior is discussed.8,14,17-20 As far as the aggregation of nonionic flexible polymers with anionic surfactants in water is concerned, it is generally found that (12) Iijma, H.; Kato, T.; Yoshida, H.; Imai, M. J. Phys. Chem. B 1998, 102, 990. (13) Boden, N.; Corne, S. A.; Jolley, K. W. J. Phys. Chem. 1987, 91, 4092. (14) Iijma, H.; Koyama, S.; Fujio, K.; Uzu, Y. Bull. Chem. Soc. Jpn. 1999, 72, 171. (15) Iijma, H.; Kato, T.; S€oDeman, O. Langmuir 2000, 16, 318. (16) Blanco, E.; Gonzales-Perez, A.; Ruso, J. M.; Pedrido, R.; Prieto, G.; Sarmiento, F. J. Colloid Interface Sci. 2005, 288, 247. (17) Missel, P. J.; Mazer, N. A.; Carey, M. C.; Benedek, G. B. J. Phys. Chem. 1989, 93, 8354. (18) Ropers, M. H.; Czichocki, G.; Brezesinski, G. J. Phys. Chem. B 2003, 107, 5288. (19) Lu, J. R.; Marrocco, A.; Su, T. J.; Thomas, R. K.; Penfold, J. J. Colloid Interface Sci. 1993, 158, 303. (20) Pandey, S.; Bagwe, R. P.; Shah, D. O. J. Colloid Interface Sci. 2003, 276, 160.

Published on Web 10/01/2010

Langmuir 2010, 26(22), 16762–16770

Abbandonato et al.

the beginning of the process occurs at a concentration called critical aggregation concentration (cac), lower than the corresponding cmc. Beyond the so-called polymer saturation concentration (psc), the surfactant in excess of the psc itself forms regular micelles in water.2 The aggregation processes of poly(ethylene glycol) (PEG) of various molecular weights on LiPFO,8 CsPFO,7 NaPFO,21 and LiPFN (litium perfluorononanoate)22 in water have been extensively studied by isothermal calorimetric tritations. The polymer-surfactant aggregation is endothermic at concentrations just above the cac but becomes progressively less endothermic at higher concentrations below the psc, becoming even exothermic in the case of CsPFO. A similar behavior had been already found for the aggregation of PEG oligomers with sodium dodecyl sulfate (SDS) in water23 (the most extensively studied systems1,23-27). Such aggregates were also investigated by 13 C NMR spectroscopy with increasing surfactant concentration23 and, in spite of the peculiar calorimetric effect, were found to be regular SDS micelles, wrapped by the polymer, as previously reported.1,28,29 The first aim of the present work was to verify whether this simple model of aggregation is adequate also for PEG-PFO systems. This was not granted, since the conformational stiffness and the hydrophobic hydrocarbon-repellent nature of the fluorinated chains might promote the formation of peculiar aggregates. To this regard, surprising differences are reported between some properties of NaPFN and of the analogous compound with CF2H terminal group30 as well as between the aggregation process of PEG-SDS23 and that of PEG-F3SDS.31 Therefore, we have studied the aggregation of PEG with LiPFO, NaPFO, and CsPFO by NMR spectroscopy, by following the 19F chemical shift (cs) trends with increasing surfactant concentration and comparing the trends obtained with and without the addition of PEG. This technique is simple, sound, and widely applied to 19F14,15,31-34 and to other nuclei as well.14,15,23,34,35 In fact, the cs trends can efficiently reveal variations and discontinuities in the average chemical environment around the investigated nuclei: in the present study, the 19F nuclei migrate, with increasing surfactant concentration, from the solution aqueous bulk phase to the PEG-PFO aggregate or to the micelle fluorinated core. The second problem tackled in this work concerns the role of the counterion in the interaction between PEG and PFO. This issue is discussed in the literature for PEG-SDS and similar systems.25,36,37 Remembering the well-known ability of PEG to complex alkaline cations, even if Li-PEG complexes are hardly detected in protic solvents,38-42 it is worth studying the LiPFO (21) De Lisi, R.; De Simone, D.; Milioto, S. J. Phys. Chem. B 2000, 104, 12130. (22) Gianni, P.; Barghini, A.; Bernazzani, L.; Mollica, V. Langmuir 2006, 22, 8001. (23) Bernazzani, L.; Borsacchi, S.; Catalano, D.; Gianni, P.; Mollica, V.; Vitelli, M.; Asaro, F.; Feruglio, M. J. Phys. Chem. B 2004, 108, 8960. (24) Minatti, A.; Zanette, D. Colloids Surf., A 1998, 113, 237. (25) Froehner, F. J.; Belarmino, A.; Zanette, D. Colloids Surf., A 1998, 137, 131. (26) Meszaros, R.; Varga, I.; Gilanyi, T. J. Phys. Chem. B 2005, 109, 13538. (27) Peron, N.; Meszaros, R.; Varga, I.; Gilanyi, T. J. Colloid Interface Sci. 2007, 313, 389. (28) Cabane, B.; Dulessix, R. J. Phys. (Paris) 1982, 43, 1529. (29) Chari, K. J. Colloid Interface Sci. 1992, 151, 294. (30) Downer, A.; Eastoe, J.; Pitt, A. R.; Penfold, J.; Heenan, R. K. Colloids Surf., A 1999, 156, 33. (31) Smith, M. L.; Muller, N. J. Colloid Interface Sci. 1975, 52, 507. (32) Guo, W.; Brown, T. A.; Fung, B. M. J. Phys. Chem. 1991, 95, 1829. (33) Segre, A. L.; Proietti, N.; Sesta, B.; D’Aprano, A.; Amato, M. E. J. Phys. Chem. B 1998, 102, 10254. (34) Xing, H.; Lin, S.-S.; Lu, R.-C.; Xiao, J.-X. Colloids Surf., A 2008, 318, 199. (35) Cabane, B. J. Phys. Chem. 1977, 81, 1639. (36) Dubin, P. L.; Gruber, J. H.; Xia, J.; Zhang, H. J. Colloid Interface Sci. 1992, 148, 35. (37) (a) Benrraou, M.; Bales, B. L.; Zana, R. J. Colloid Interface Sci. 2003, 267, 519. (b) J. Phys. Chem. B 2003, 107, 13432. (38) Cross, J. Nonionic Surfactants; Marcel Dekker, Inc.: New York, 1987.

Langmuir 2010, 26(22), 16762–16770

Article

and NaPFO aggregation processes also in the presence of cyclic ligands, namely in the presence of 12-crown-4 or 15-crown-5 ether, capable of catching Liþ or Naþ ions, respectively.43 Such crown ethers and PEGs are built up by the same monomeric ethylenoxy (EO) unit, and their use as complexing agents may have similar outcomes (PEG was even called “the poor chemist’s crown”),44 but their effects can obviously diverge whenever the chain structure, open and highly flexible in the case of PEG, cyclic and constrained for crown ethers, becomes important.38,41 (This point will be presented as first in the Results and Discussion section.) The last problem faced concerns the PFO chain conformation in water and inside aggregates and micelles. Perfluorinated linear alkyl chains are known to be stable in a helical conformation, obtainable starting from the chain skeleton in its planar, all-anti (zigzag) conformation and successively twisting all the CC-CC dihedral angles of 15°-20° in the same direction.6,45,46 However, the prevalence of the helical structure has been sometimes discussed,46,47 and the all-anti conformation has been found to dominate for perfluorinated chains shorter than nine carbon units.46 We have asked ourselves if the different 19F cs values recorded for surfactants outside and inside the various micelles or aggregates, besides reflecting the aqueous or fluorinated environment surrounding the molecule, could be sensitive to the prevailing conformational arrangement of the PFO chains. We have tried to give an answer to this question by comparing experimental and computed chemical shifts. In effect, the current quantummechanical (QM) methods and the robustness and reliability of the available computational packages more and more frequently enable us to explain details of experimental findings or to discriminate among different explicative hypotheses by comparing experimental and theoretical results. Today’s advances of theory and methodologies make it possible to investigate not only isolated but also embedded systems and composite environments.48-50 Moreover, the degree of correlation between sets of computed chemical shielding tensors and the corresponding experimental observables has been recently used to solve conformational problems in complex systems.51-55 In the present work, the chemical shifts were calculated at the DFT level of theory, taking into account the effect of the environment by means of the IEF-PCM method.56 (39) Chan, K. W. S.; Cook, K. D. Macromolecules 1983, 16, 1736. (40) Sartori, R.; Sepulveda, L.; Quina, F.; Lissi, E.; Abuin, E. Macromolecules 1990, 23, 3878. (41) Okada, T. Macromolecules 1990, 23, 4216. (42) Wyttenbach, T.; von Helden, G.; Bowers, M. T. Int. J. Mass Spectrom. Ion Processes 1997, 165/166, 377. (43) Steed, J. W. Coord. Chem. Rev. 2001, 215, 171. (44) Balasubramanian, D.; Chandani, D. J. Chem. Educ. 1983, 60, 77. (45) Jang, S. S.; Blanco, M.; Goddard, W. A., III; Cadwell, G.; Ross, R. B. Macromolecules 2003, 36, 5331. (46) Ellis, D. A.; Denkenberger, K. A.; Burrow, T. E.; Mabury, S. J. Phys. Chem. A 2004, 108, 10099. (47) Knochenhauer, G.; Reiche, J.; Brehmer, L.; Barberka, T.; Wooley, M.; Tredgold, R.; Hodge, P. J. Chem. Soc., Chem. Commun. 1995, 1619. (48) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105, 2999. (49) Cavalli, A.; Carloni, P.; Recanatini, M. Chem. 2006, 106, 3497. (50) Kamerlin, S. C. L.; Haranczyk, M.; Warshel, A. J. Phys. Chem. B 2009, 113, 1253. (51) Kaupp, M., Buhl, M., Malkin, V. G., Eds.; Calculation of NMR and EPR Parameters; Wiley VHC Verlag GmbH & Co. KGaA: Weinheim, 2004. (52) Marini, A.; Prasad, V.; Dong, R. Y. A Combined DFT and Carbon-13 NMR Study of a Biaxial Bent-core Liquid Crystal. In Nuclear Magnetic Resonance Spectroscopy of Liquid Crystals; Dong, R. Y., Ed.; Publisher World Scientific Publishing Co.: Singapore, Nov 2009. (53) Marini, A. Understanding Complex Liquid Crystalline Materials: A multinuclear NMR spectroscopy and ab initio calculations approach; VDM Verlag Dr. Muller Aktiengesellschaft & Co. KG: Saarbr€ucken, Germany, 2010 (ISBN-13: 9783639254228). (54) Dong, R. Y.; Marini, A. J. Phys. Chem. B 2009, 113, 14062. (55) Dahlberg, M.; Marini, A.; Mennucci, B.; Maliniak, A. J. Phys. Chem. A 2010, 114, 4375. (56) Canc
es, E.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107, 3032.

DOI: 10.1021/la102578k

16763

Article

Abbandonato et al.

Materials and Methods Materials and Samples Preparation. Poly(ethylene glycol) (PEG) samples with nominal molecular weight of 1500 and 3400 Da, as well as 12-crown-4 and 15-crown-5 ethers, were commercial products from Aldrich. To obtain LiPFO, NaPFO, and CsPFO, pentadecafluorooctanoic acid (99%, Fluorochem Ltd.) was neutralized with lithium hydroxide monohydrate (99.95%, Aldrich), sodium hydroxide (g98%, Aldrich), and cesium hydroxide (99.7%, Aldrich), respectively, following the reported procedure.7 From aqueous stock 0.41-0.46 m solutions of the three salts, three series of solutions with concentrations between 0.005 and 0.100 m were prepared by dilution. Other series of solutions with the same surfactants concentration were prepared, also containing respectively PEG1500 0.1 wt %, PEG3400 0.1 wt %, PEG3400 0.2 wt %, or the suitable crown ether 0.4 wt %. All the solutions were allowed to homogenize at least overnight before recording the 19F NMR spectrum. Moreover, some samples were prepared twice in order to check the reproducibility of the procedure. NMR Measurements. 19F NMR spectra of all the described solutions were recorded using a Varian spectrometer operating at 282.17 MHz. A single pulse sequence was applied, with a number of transients ranging from 4 to 64, depending on the concentration. All spectra were recorded at 25 °C ( 0.1 °C without frequencyfield locking, adjusting the homogeneity of the magnetic field on the 1H water signal. The signal of CF3COONa in water was used as external reference for the 19F cs values. Most measurements were repeated on the same samples days or weeks later in order to test the stability of the solutions and, more generally, the reproducibility of the results. 19F NOESY spectra were also recorded on samples of LiPFO and CsPFO without and with PEG3400 at high, intermediate, and low surfactant concentration. Mixing times of 1.1 or 0.7 s were used; these values were chosen after having estimated typical 19F T1 values (CF2 groups: T1 about 0.9 and 0.5 s below and far above the cmc (or cac), respectively; CF3 group: T1 about 1.5 and 1 s). Computational Details. All the quantum-mechanical calculations were performed by the Gaussian 09 package (A.01),57 with molecular models of PFO anion and LiPFO built up by the GaussView 5.0 program. The geometry optimizations and vibrational frequencies calculations were performed at the DFT level of theory, employing the combination of B3LYP functional58,59 with the 6-311þG(d) basis set. The calculations were done both in vacuo and in solution by taking into account the effect of the solvent with the IEF-PCM model,56 using the “G03Defaults” setting. The investigated solvent environments were constituted by perfluoroctane (PF, dielectric constant ε = 1.8260,61) and by water (ε=78.54). The PCM cavities were built up applying the United Atom Topological Model to the atomic radii of the UFF force field,62 as implemented in the Gaussian 09 code. GIAO-DFT calculations of 19F chemical shielding tensors (CSTs) were performed on the PFO anion in various conformations, chosen as discussed in the Results and Discussion section, by exploiting a modified Perdew-Wang exchange-correlation functional,63 called MPW1PW91,64 which has been shown to be particularly accurate in predicting chemical shielding tensors. All the CST calculations were obtained using the 6-311þG(d,p) basis set, which seems to accomplish the best compromise between accuracy and CPU time. The isotropic part (σF) of each computed CST was evaluated as one-third of the tensor trace. In order to relate the σF values to the (57) Frisch, M. J.; et al. Gaussian 09 Revision A.01; Gaussian, Inc.: Wallingford, CT, 2009. (58) Becke, A. D. Chem. Phys. 1993, 98, 1372. (59) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (60) Camerons, D. G.; Umemura, J.; Mantsch, H. H. Can. J. Chem. 1981, 59, 1357. (61) Brady, J. E.; Carr, P. W. Anal. Chem. 1982, 54, 1751. (62) Casewit, C. J.; Colwell, K. S.; Rappe, A. K. J. Am. Chem. Soc. 1992, 114, 10046. (63) Perdew, J. P.; Burke, K.; Wang, Y. Phys. Rev. B 1996, 54, 16533. (64) Adamo, C.; Barone, V. J. Chem. Phys. 1998, 108, 664.

16764 DOI: 10.1021/la102578k

experimental NMR chemical shifts (δFExp),65 which are referred to CF3COONa in water, the 19F isotropic shielding of trifluoroacetate cation in water (σREF) was also calculated at the same level of theory exploited for PFO. The value obtained for σREF is 252.37 ppm. The computed chemical shifts were finally evaluated as δFCalc = -(σF - σREF).

Results and Discussion 19

F NMR Spectra and Data Analysis. The numbering of the groups in the PFO chain is shown in Figure 1, together with a typical 19F NMR spectrum. In the monodimensional spectra, seven signals with different chemical shifts are present, corresponding to the six CF2 groups and to CF3. The assignment of the signals was carried out according to the literature.66 Only one superimposition, involving the F3 and F6 signals, occurs in the middle of the concentration range examined for each series of spectra. All spectra recorded are in the fast exchange limit:67 therefore, each observed chemical shift δiobs is the weighted average of the values pertinent to the different exchanging sites in the sample.68 When only two exchanging sites are considered, for instance water and micelles, we have δi obs ¼ ðcw δi w þ cmic δi mic Þ=c

ð1aÞ

which can be rearranged to give δi obs - δi w ¼ - cw ðδi mic - δi w Þ=c þ ðδi mic - δi w Þ

ð1bÞ

In eqs 1a and 1b, the index i refers to the ith group of equivalent 19 F nuclei, c is the total PFO concentration in solution, cw and cmic are the PFO concentrations in the aqueous bulk phase and in micelles, respectively, with c = cwþ cmic , and δiw and δimic are the chemical shifts in the two environments. Following the widely applied phase-separation model of micellization,3 micelles form when cw has reached its maximum value, cmc, which remains constant when c is further increased. Moreover, (δimic - δiw) can usually be considered independent of concentration. Then δi obs - δi w ¼ - ½cmcðδi mic - δi w Þð1=cÞ þ ðδi mic - δi w Þ

ð2Þ

and the trend of (δiobs - δiw) vs 1/c is linear; (δimic - δiw) and cmc can be easily determined from the intercept and slope values of the fitting straight line. For a three-site fast equilibrium, as may occur in the systems studied in this work, we have δi obs ¼ ðcw δi w þ cagg δi agg þ cmic δi mic Þ=c

ð3Þ

where cagg is the polymer-PFO aggregates’ concentration (c = cwþ caggþ cmic) and δiagg is the ith 19F chemical shift in the aggregates. If micelles form when cw and cagg have reached their maximum values (cac and (cagg)max, respectively) and if (δimic δiw) and (δiagg - δiw) are independent of concentration, eq 3 becomes δi obs - δi w ¼ - ½cacðδi mic - δi w Þ þ ðcagg Þmax ðδi mic - δi agg Þð1=cÞ þ ðδi mic - δi w Þ

ð4Þ

(65) Facelli, J. C. Concepts Magn. Reson. 2004, 20A, 42. (66) Buchanan, G. W.; Munteanu, E.; Dawson, B. A.; Hodgson, D. Magn. Reson. Chem. 2005, 43, 528. (67) To our knowledge, the occurrence of the slow exchange limit in systems similar to the present ones is reported only in ref 33. (68) Sanders, J. K. M.; Hunter, B. K. Modern NMR Spectroscopy; Oxford University Press: Oxford, 1993.

Langmuir 2010, 26(22), 16762–16770

Abbandonato et al.

Article

Figure 1. Group numbering in the PFO unit and 19F NMR spectrum of aqueous CsPFO (0.032 m). The chemical shift scale is referred to aqueous CF3COONa. The insets show the excellent resolution achieved in this spectrum. Table 1. Relevant Surfactant Concentrations and 19F Chemical Shifts of the F8 Group in Micelles with Respect to the Corresponding Values in Watera b

system

cmc (mol/kg)

|δmic - δw| (ppm) [F8]

b

system

cac (mol/kg)

C2 (mol/kg)

Cw d (mol/kg)

|δmic - δw| ≈ |δagg - δw| (ppm) [F8]

LiPFO/w LiPFO/w/c4 NaPFO/w NaPFO/w/c5 CsPFO/w

0.030 2.56 LiPFO/w/PEG1500 0.018 0.038 0.029 2.40 0.024 2.47 LiPFO/w/PEG3400 0.014 0.038 0.030 2.47 0.031 2.52 NaPFO/w/PEG3400 0.1 wt % 0.014 0.035 0.028 2.47 0.014 2.32 NaPFO/w/PEG3400 0.2 wt % 0.013 0.060 0.028 2.51 0.025 CsPFO/w/PEG1500 0.017 0.040 0.027 2.62 CsPFO/w/PEG3400 0.013 0.060 0.028 2.57 0.060c 2.84 0.028e a The estimated uncertainties on the reported values are of some units of the last digit quoted. The couples of values in the second and third columns and in the last two columns were obtained by fitting selected series of data from Figures 2, 3 and 4 to eq 2 (or eq 4). Each concentration not corresponding to a chemical shift difference value in the table was directly determined from a break point in the trends of Figures 2, 3, and 4. b w, c4, and c5 indicate water, 12-crown-4, and 15-crown-5 ethers, respectively; PEG concentration is 0.1 wt % where not differently specified. c Concentration called cmc II in the text. d Surfactant concentration in the aqueous bulk phase for c g C2. e Concentration identified as Cw in the text.

Equation 4, as eq 2, describes a linear trend of (δiobs - δiw) vs 1/c. Therefore, a linear trend of the experimental data with respect to 1/c can support a two- or three-site equilibrium model, with cw, cagg, (δimic - δiw) and (δiagg - δiw) constant. Of course, if δimic ≈ δiagg, eqs 2 and 4 become formally equal. Finally, we observe that all the above equations reduce to δiobs = δiw when c < cmc (or c < cac) and to δiobs ≈ δimic when c . cmc (or c . cac þ (cagg)max). The concentration values and chemical shift differences relevant for the present work are collected in Table 1. LiPFO and NaPFO Systems. LiPFO and NaPFO give quite similar results and will be discussed together. In Figure 2 the trends of |δiobs| for F8, F4, and F3 versus c are reported for the binary surfactant/water systems and for the ternary systems also containing the PEG1500 or PEG3400 0.1 wt % or the suitable crown ether. The graphs relative to all the fluorine groups are supplied as Supporting Information, Figure S.1. For the groups from F8 to F4 all trends are very similar and show the same, welldefined discontinuities. Binary Systems. From the series of graphs from F8 to F4 it is evident that the difference between the high and low concentration chemical shifts progressively decreases (from 1.8 to 0.9 ppm), reflecting a reduction of |δimic - δiw|. The trends of the F2 and F3 chemical shifts, very similar to each other, are quite different from those of F8-F4 and show an abrupt reduction of the spanned range of values to only 0.3-0.4 ppm. The variation of (δimic - δiw) Langmuir 2010, 26(22), 16762–16770

along the fluorinated chain is put in evidence in Figure 6 (discussed later) by the changing of the distance between light green and dark blue dashes. All this clearly suggests that the water penetration inside micelles or aggregates of any kind and, more generally, the interaction between the 19F nuclei and the hydration and counterion shell are limited to the second CF2 group from the polar head. A similar remark, concerning PVP/LiPFN/ water systems, is reported in ref 33. For both the binary systems LiPFO/water and NaPFO/water, the cmc is pointed out by the onset of a steep cs variation at 0.030 and 0.031 m for LiPFO/water and NaPFO/water, respectively, in agreement with the literature.7,8 At concentrations higher than the cmc, all the F8-F2 chemical shifts change linearly with respect to 1/c, in agreement with eq 2 and the underlying model. These trends are shown for the F8 group in Figure S.3a,b in the Supporting Information, while the relevant parameters obtained from their analysis are collected in Table 1. A quite moderate dependence of δiw on c is detected below the cmc or cac only in the case of LiPFO solutions. Effects of Crown Ether Addition. In the presence of an amount (0.4 wt %) of the suitable crown ethers, that is 12crown-4 for the LiPFO solution and 15-crown-5 for the NaPFO one, the cmc values decrease, as clearly shown in Table 1, Figure 2, and Figure S.1. The F8-F4 cs trends vs c well resemble those found in the binary systems, while the trends vs 1/c are linear for both surfactants, as predicted by eq 2 (see Figure S.3a,b). DOI: 10.1021/la102578k

16765

Article

Abbandonato et al.

Figure 2. |δobs| values (the experimental values are negative) for F8, F4, and F3 nuclei vs the surfactant concentration c. Notice the quite different vertical scales used for the different groups of nuclei. Colors and symbols: red triangles, LiPFO/w and NaPFO/w system; yellow stars, systems with crown ethers; blue squares, with PEG1500 0.1 wt %; green rhombs, with PEG3400 0.1 wt %.

The cs differences (δimic - δiw) are slightly decreased by the crown ether addition, more significantly in the case of NaPFO. As far as the F3 and F2 groups are concerned, the rather flat cs trends in the presence of crown ether neither point out the cmc decrease nor match the cs trends of the binary systems. Altogether, the complexation of Liþ or Naþ ions does not substantially affect the micellization process beyond the cmc; however, complexed ions stabilize the early formed PFO micelles better than the free ions do. To this purpose, we recall that, for micelles of both DS19 and PFO7,8,14 with different counterions, the cmc values decrease in the order Liþ g Naþ > Csþ, while in the same order, the hydrated counterion radius decreases and the counterion binding degree to the micelle increases.10,12,19,14,17 Moreover, small-angle neutron scattering (SANS) studies on LiDS and SDS micellar solutions have shown that the counterion binding degree is much higher for Liþ and Naþ ions complexed by [2.2.2]-cryptand than for simply hydrated ions.69,70 This causes the decrease of the effective micellar charge, the reduction of the charge separation in solution, and the formation of stable micelles with lower aggregation number. All this can reasonably occur also when crown ethers complex Liþ or Naþ ions in the presence of PFO, with the consequent decrease of cmc. A SANS study of these systems might confirm this explanation. Effects of PEG Addition. In the presence of PEG 0.1 wt %, the cac values are lower than the corresponding cmc values of the binary systems (see Table 1, Figure 2, and Figure S.1). Adapting the above considerations on the effect of crown ethers addition to the case of the polymer (see also ref 36), we suggest that each PEG chain, binding several Liþ or Naþ ions38,39,41 and bringing them close to one another, might play the role of aggregation nucleus, stabilizing one or more small PFO micelles. However, the early aggregation of PFO on PEGs at 25 °C is known to be endothermic, as the “regular” PFO micellization,7,8 and the process is (69) Baglioni, P.; Gambi, C. M. C.; Giordano, R.; Teixeira, J. Colloids Surf., A 1997, 121, 47. (70) Scaffei, L.; Lanzi, L.; Gambi, C. M. C.; Giordano, R.; Baglioni, P.; Teixeira, J. J. Phys. Chem. B 2002, 106, 10771.

16766 DOI: 10.1021/la102578k

anyway entropically driven. The investigation of the energetic details of the proposed polymer-micelle interaction mechanism, mediated by the counterion, could be particularly interesting in the case of LiPFO, since Li-PEG complexes, found in the vapor phase,42 are hardly detected in protic solvents.39,41 We will now discuss the relevant features of the cs trends in the presence of PEG. Except for a slight difference in the cac values, the cs trends relative to LiPFO solutions with the same amount by weight of PEG1500 and PEG3400 are superimposed, indicating that the aggregation process is basically independ of the polymer molecular weight, at least within the narrow range here investigated. For all the LiPFO and NaPFO systems with PEG, the cs trends of F8-F4 (Figure 2 and Figure S.1) show some salient peculiarities not found either in the presence of crown ethers or in PEG-SDS systems.23 (The F3 and F2 cs trends do not present any comparable feature.) In fact, the chemical shifts begin to vary at the cac, indicating that the average environment around the 19F nuclei is becoming less hydrated, but after a small increment of the surfactant concentration, a plateau is approached or reached. At c = 0.038 and 0.035 m for LiPFO and NaPFO solutions, respectively, the cs trends with PEG have a second break point, beyond which they closely follow those without PEG. We will call this particular concentration C2. Similar trends are reported for PVP/ LiPFN/water systems,33 while the graphs drawn against 1/c in Figure S.3a,b recall those relative to PEG/F3SDS/water systems.31 In order to verify whether C2 coincides with the polymer saturation concentration (psc), a series of samples of NaPFO/ water with a double amount (0.2 wt %) of PEG3400 was were also studied. As shown in Table 1 and in Figure 3a for F8 (the data for all fluorine groups are reported in Figure S.2), the cac does not significantly decrease with increasing the polymer concentration, while C2 shifts from 0.033 to 0.060 m, thus hardly supporting the identification of C2 with the psc. In fact, the concentration of aggregated surfactant, estimated by the difference (C2 - cac), is more than doubled when the polymer concentration is doubled. Moreover, a comprehensive comparison of the systems with different polymer amounts rules out the identification of C2 Langmuir 2010, 26(22), 16762–16770

Abbandonato et al.

Article

Figure 3. (a) Experimental |δobs| trends vs c for F8 nuclei in NaPFO solutions: triangles, NaPFO/w system; rhombs, with PEG3400 0.1 wt %; circles, with PEG3400 0.2 wt %. (b) Computed trends vs c. The dashed curve, simulating a two-site fast exchange behavior (surfactant in water and in micelles), is computed by eq 2, with cmc = 0.030 m and (δmic - δw) = -2.50 ppm. The solid curve between 0.015 and 0.035 m, also simulating a two-site fast exchange behavior (surfactant in water and in PEG-PFO aggregates), is computed with cac = 0.015 m and (δagg δw) = -1.00 ppm; beyond c = 0.035 m, it is computed by eq 4, with cw = cac, cagg = (0.035 m - cac), (δagg - δw) = -1.00 ppm and (δmic δw) = -2.50 ppm. The dotted curve is computed as the solid one, but the width of the two-site exchange interval beyond the cac is doubled.

as psc. Figure 3b and Figure S.3c show the cs trends vs c and 1/c, respectively, expected if a two-site fast exchange described by eq 2 were active in the cac-C2 range, followed, above C2, by a threesite exchange described by eq 4. The numerical assumptions specified in the captions were chosen in order to mimic the cs behavior of F8 with 0.1 wt % of PEG3400. The comparison between parts a and b of Figure 3 clearly shows that such model is not adequate to contemporaneously explain the cs trends of samples with PEG 0.1 and 0.2 wt %. In particular, the plateau values reached in the sample series with less polymer cannot simply correspond to the chemical shifts of the various groups in the PEG-PFO aggregates, since, in the series with double amount of polymer, such plateau values are widely exceeded before C2. Moreover, the trends of F8-F4 chemical shifts vs 1/c are nonlinear in the region between the cac and C2, indicating that the model underlying eq 2 fails (compare Figures S.3b and S.3c). Instead, such nonlinear behavior is compatible with eq 1b if cw and/or (δiagg - δiw) vary with increasing c; furthermore, the approaching or occurrence of a plateau in the cs trends reasonably indicates that both (δiagg - δiw) and cw/c are about constant. Above the C2 concentration, the cs trends vs 1/c are again linear, regular micelles presumably form besides the early formed PEGPFO aggregates and a three-site fast exchange is established, described by eq 4. However, the superimposition of the F8-F4 cs trends with and without PEG, and the results of the linear data fitting (see Table 1) suggest that δiagg ≈ δimic and that cw ≈ cmc when c > C2. (The constant cw value when c > C2 is called Cw in Table 1.) Altogether, our data indicate that, between the cac and C2, the surfactant concentration in water is not constant: in contrast with what usually found or accepted, PFO added to the system partially dissolves in the aqueous bulk phase and only partially contributes to the PEG-PFO aggregates formation and evolution. At C2, the surfactant concentration in water, cw, reaches its maximum value, which is about the same as in binary systems, closely approaching the cmc value. Beyond C2, the aggregation process regularly proceeds. Moreover, since δiagg ≈ δimic, even the early formed PEG-PFO aggregates are probably “regular” micelles, maybe smaller than the free ones, wrapped by the polymer, in spite of the quite peculiar cs trends in Figure 2 and Figure S.1 (see also ref 33). This view is well compatible with isothermal calorimetric tritation (ICT) results:8 the aggregation process of LiPFO on PEG 0.1 wt % is endothermic at its onset, but at higher PFO Langmuir 2010, 26(22), 16762–16770

Figure 4. |δobs| values (the experimental values are negative) vs c for F8 and F2 fluorine groups in CsPFO solutions. Notice that the vertical scales are quite different. Symbols: triangles, binary CsPFO/w system; squares, with PEG1500 0.1 wt %; rhombs, with PEG3400 0.1 wt %.

concentrations where the aggregation process is accompanied by the increase of cw, the endothermic effect decreases and almost vanishes. Beyond C2, the “regular” endothermic micellization process continues. ICT measurements on systems with different PEG concentrations might confirm our analysis. CsPFO Systems. In Figure 4 and Figure S.1, the trends of δiobs vs c are reported for the binary CsPFO/water system and for those also containing PEG1500 or PEG3400. Most remarks relative to the LiPFO and NaPFO binary systems are also valid for the CsPFO one and will not be repeated, while its peculiarities, highlighted also by the data of Table 1, will be discussed. A moderate, but significant, linear dependence of δiw on c is detected below the cmc or cac and taken into account when necessary. DOI: 10.1021/la102578k

16767

Article

Abbandonato et al.

Figure 5. Conformations of PFO (“side” and “front” views): (a) all-gauche (notice the spiral structure of the carbon backbone), (b) all-anti, (c) helical.

Binary System. The maximum differences |δimic - δiw| for all fluorine groups are higher than those found in the LiPFO and NaPFO solutions. This could be related to the shape of the micelles, oblate in the case of CsPFO, spherical in the other cases: the average conformational distribution along the PFO chains is presumably sightly different in the two kinds of aggregates, and the average chemical environment around each fluorine group can reflect this difference. This point will be further discussed in the final subsection. In the cs trends of Figure 4 and Figure S.1, besides the cmc at 0.025 m, which marks the beginning of the micellization process, a second break point is found at c = 0.060 m. This concentration was called cmc II in the literature14,15 and was supposed to mark the transformation of the early formed micelles from spherical to oblate. The cs trend vs 1/c is linear above the cmc II (see Figure S.3d), and the parameters obtained by fitting the data of F8 to eq 2 are shown in Table 1 (last line of the second and the third columns). The concentration value 0.028 m, slightly higher than the cmc, might be interpreted as Cw, the surfactant concentration in the aqueous bulk phase coexisting with CsPFO micelles in their final form. Finally, it must be noted that the F2 cs trend of Figure 4 differs from that reported in refs 14 and 15, which has a minimum instead of a discontinuity at the cmc II. Effects of PEGs Additions. As expected, the cac values recorded in the presence of the two polymers are lower than the cmc (see Table 1). To a certain extent, the F8-F4 cs trends relative to the solution with PEG1500 are analogous to those found for LiPFO and NaPFO: C21500 is located at 0.040 m, and above this point, the trends with the polymer closely resemble those without it (see Figure 4 and Figure S.1). The F8-F4 cs trends relative to the solution with PEG3400 match those with PEG1500, except in the interval between C21500 and C23400 (the last value coincides, maybe accidentally, with the cmc II). Therefore, at least one step of the PEG-CsPFO aggregation process depends on the polymer molecular weight, differently from what found for LiPFO and NaPFO and from what reported for PEG-SDS aggregation.23 As far as the F2 and F3 cs trends are concerned, the ones with the polymers are significantly flatter 16768 DOI: 10.1021/la102578k

than those without them. Altogether, the case of PEG-CsPFO aggregation is more complex than the others here studied, and some unexplained peculiarities should be investigated by other techniques, as indicated for instance in ref 37. Anyway, as found for PEG aggregation on LiPFO and NaPFO, on the basis of the data in Table 1 the concentration of the surfactant in the water bulk might increase while PEG-CsPFO aggregates form, until the value Cw ≈ cmc is reached. Beyond this point, a three-site equilibrium settles, described by eq 4. Remarks on Data of Table 1. The Cw values determined for all the systems containing PEG and for the CsPFO/water binary system, ranging from 0.027 to 0.030 m, are only slightly lower than the cmc values of the NaPFO/and LiPFO/water systems. In the binary systems, the |δimic - δiw| values are slightly higher than in the corresponding ternary ones. However, only for CsPFO/water without and with PEGs and for NaPFO/water without and with 15-crown-5 ether the divergence is numerically significant and clearly detectable. Perfluorinated Chain Conformation and 19F Chemical Shifts. The influence of the solvent environment and of the chain conformation on the (δimic - δiw) values in the binary systems will be investigated in the following. Quantum-Mechanical Study: Populated Conformations. Geometry optimizations were performed starting from two conformations: all-gauche, with CC-CC dihedral angles of 60°, and all-anti, with CC-CC dihedral angles of 180°. During the optimizations, performed without any geometrical constraint, at first only the bond distances and angles varied with a substantial internal energy decrease, producing “relaxed” all-gauche and allanti conformations (Figure 5a,b). In the former conformation, the starting dihedral angles were altered of 3.4° as a maximum, while in the latter one they remained practically unchanged. Then, in the following of the optimization process, for both the all-gauche and all-anti conformations, the CC-CC dihedral angles changed in several steps until the “fully relaxed” helical structure of Figure 5c was reached. This is the most stable conformation in vacuo and in the solvents; as far as the “relaxed” geometrical parameters are concerned, the differences due to the environment are negligible. Langmuir 2010, 26(22), 16762–16770

Abbandonato et al.

Article

Table 2. Relative Energies (kcal/mol) and Populations of PFO at 298 K, Computed in Various Environments for Three Relaxed Conformations, without and with Zero-Point Vibrational Energy (ZPE) Contribution environment in vacuo (ε = 1)

PF (ε = 1.83)

water (ε = 78.54)

reference conformation, helical

all-anti

all-gauche

all-anti

all-gauche

all-anti

all-gauche

0.0 100 0.0 100

2.7 1.0 1.3 10.6

5.2 0.0 5.2 0.0

2.5 1.6 1.1 15.8

5.1 0.0 5.0 0.0

2.4 1.8 0.9 22.3

5.0 0.0 5.0 0.0

energy (kcal/mol) relative population energy with ZPE (kcal/mol) relative population

In vacuo computations repeated without and with counterions (Liþ, Naþ, Kþ, Csþ) demonstrated that the stable conformation is unaffected by the counterion, in contrast with what reported for a similar perfluorinated surfactant.71 The stable helical structure is obtainable from the chain in the relaxed all-anti conformation by twisting the five CC-CC dihedral angles (ordered starting from the polar head) of 16.8°, 16.4°, 16.6°, 17.8°, and 16.7°. This result is in excellent agreement with what reported in ref 45. It must be noted that, the gauche arrangement of a CC-CC dihedral angle corresponds to a relative minimum of the internal energy between two maxima (eclipsed situations). Instead, between the anti arrangement and the helical one, there is no energetic barrier, but just an energy decrement: an anti conformation is a torsional deformation of the helical one. The relative energies of the “relaxed” conformations, computed without and with the contribution of the zero-point vibrational energy (ZPE), are reported in Table 2. Such a contribution amounts to 75.5 kcal/ mol for the helical and all-gauche conformations and to 74.1 kcal/ mol for the all-anti one, which lacks the vibrational coordinates involving the torsions of the CC-CC dihedral angles. As estimated by the Boltzmann distribution, the relaxed all-anti conformation becomes easily accessible when the ZPE is taken into account, while the all-gauche population is negligible at room temperature. In order to estimate the energy increment due to the anti arrangement of a single dihedral angle in a helical chain, other sets of conformers were generated in the two solvents by rotating the dihedral angle C5C6-C7C8 within the range of (15° around the minimum-energy conformation and exploiting the relaxed scan technique.72 The anti arrangement corresponds to an energetic increment in vacuo of about 0.55 kcal/mol without ZPE contribution (as shown in Figure S.4). Five similar additive increments, one for each CC-CC dihedral angle, produce the all-anti relative energy of 2.7 kcal/mol reported in Table 2. Therefore, from the data of Table 2, taking into account the ZPE contribution, single anti or gauche arrangements increase the internal energy of the helical chain of about 0.2-0.3 or 1 kcal/mol, respectively. Altogether, for the minimum-energy helical conformation, a moderate flexibility is predicted: some gauche arrangements in the chain are possible, while more anti arrangements are easily contemporarily accessible. The distribution of the accessible structures should be slightly broader in water than in the perfluorinated environment; however, we must avoid inferring too detailed conclusions from quite small energy differences, since the theoretical PF environment only partially reproduces the conditions inside the micelle where the PFO chains are closely packed and specifically constrained. The relative energy of the various conformers could be consequently altered. Comparison between Averaged Computed Distances and 19 F NOESY Results. All the NOESY 19F NMR spectra show (71) Erkoc-, S-.; Erkoc-, F. J. Mol. Struct. 2001, 549, 289. (72) Cramer, C. J. Essentials of Computational Chemistry: Theories and Models; Wiley: Chichester, England, 2004.

Langmuir 2010, 26(22), 16762–16770

connections between 19F nuclei of adjacent groups and of groups separated by one CF2 unit, as already described in the literature.66 The cross-peaks of the second kind are more intense than or comparable to those of the first kind, with two exceptions at the chain ends: the F7-F8 cross-peak is more intense than the F6-F8 one, and the F2-F3 cross-peak is more intense than the F2-F4 one. No significant difference is observed among the NOESY spectra of samples below or above the cmc, with or without PEG3400. This indicates that the micellization and aggregation processes do not imply changes detectable by this technique in the conformational distribution of the PFO chain. The relative intensities of the crosspeaks can be qualitatively explained by the average values of the computed distances between fluorine nuclei of different groups along the chain, for both the helical and the all-anti conformation (see Table S.1 in the Supporting Information). The only exception is the intensiy inversion involving the F2-F3 and F2-F4 cross-peaks, already observed in ref 66. In that work the NOESY spectrum is anyway related only to the all-anti conformation. Comparison between Computed and Experimental Chemical Shifts. The 19F isotropic chemical shifts, computed for the allanti, all-gauche, and helical conformations in vacuo, in PF and in water, are shown in graphic form in Figure 6a,b. Tables of values are supplied as Supporting Information (Table S.2). The experimental values reported in Figure 6a,b are chosen as follows: the 19 F chemical shifts in water (dark blue dashes) are those recorded below the cmc (or cac), common to all solutions, while the values in the micellar fluorinated environment (light green dashes) are those reached in the LiPFO/water system at high concentration. Some interesting observations are possible: (i) for each 19F group, except for those at the chain ends, the interval of values spanned by the computed chemical shifts includes the experimental values; (ii) the values computed for the unpopulated all-gauche conformation (stars in Figure 6a) differ from those of the all-anti and helical ones essentially for their alternating trend along the chain and because, from F2 to F6, they are hardly affected by the environment; (iii) the trends of the chemical shifts computed for the populated helical and all-anti conformations (triangles and squares, respectively, in Figure 6a,b) follow the experimental ones, and their values are more sensitive to the environment; (iv) for each 19 F group, the experimental values δiw and δimic are mostly located close to the values computed for the helix in water and for the allanti conformation in PF, respectively. In other words, Figure 6b clearly shows general agreement between blue dashes and blue triangles as well as between green dashes and green squares. Therefore, the experimental 19F chemical shifts values, affected by the aqueous or perfluorinated environment, also reflect the conformational distribution of the PFO chain: the comparison with the computed values confirms the prevalence of the helical arrangement and highlights the contribution of the anti distortions, much more relevant in micelles than in water. Moreover, since the differences between the cs values in water and in CsPFO/ water micelles are higher than those in the other systems, as noticed in the CsPFO Systems section, anti distortions are DOI: 10.1021/la102578k

16769

Article

Abbandonato et al.

The accurate analysis of the 19F cs trends with increasing surfactant concentration suggests that the PEG-PFO aggregates, even the early formed ones, are “regular” micelles wrapped by the polymer, in spite of some relevant peculiarities. These are explained supposing that, above the cac, the surfactant concentration in water, cw, is not constant, in contrast with what usually found or accepted, but it increases until it reaches its maximum value, Cw ≈ cmc. This is about the same in all the binary systems and in those with PEG. The complexation of Liþ or Naþ ions by suitable crown ethers does not substantially affect the micellization process beyond the cmc, but in the presence of complexed ions, the cmc decreases while the counterion binding degree to the micelle presumably increases. Analogously, each PEG chain, binding several Liþ or Naþ ions, might play the role of aggregation nucleus, causing the decrease in the cmc. However, this point requires further investigations to be definitively proved. The “fully relaxed” helical structure of Figure 5c is the most stable conformation in vacuo and in the solvents. The experimental 19F cs values, influenced by the aqueous or perfluorinated environment, also reflect the conformational distribution of the PFO chain: the anti arrangements are more probable in micelles than in water and in the discotic CsPFO micelles than in the spherical LiPFO and NaPFO ones. We like to stress this last finding since, in the study of supramolecular aggregates, conformational effects are often guessed but not easily proved. The quantum-mechanical approach to the analysis of the 19F chemical shifts emerges as a reliable and promising strategy to investigate this kind of complex system.

Figure 6.

19

F isotropic chemical shifts of the different fluorine groups in PFO, computed for various conformations and environments. The experimental values recorded in micelles and water are also reported. The graph at the bottom (b) is a simplified version of the upper one (a).

suggested to be more important in the discotic CsPFO micelles than in the spherical LiPFO and NaPFO ones.

Conclusions The findings reached by this work are presented in detail throughout the Results and Discussion section. Here we just summarize our answers to the three problems presented in the Introduction.

16770 DOI: 10.1021/la102578k

Acknowledgment. The authors are grateful to Prof. Paolo Gianni for providing most compounds, to Dr. Silvia Borsacchi, who prepared part of the samples, to Dr. A. Mandoli for hospitality in his laboratory, and to Dr. Claudia Forte for useful suggestions. The computational work was partially supported by the Swedish National Infrastructure for Computing (SNIC 00109-49) via PDC. Supporting Information Available: Computed average distances between fluorine nuclei; computed 19F chemical shifts. Graphics of (i) the experimental 19F chemical shifts vs c for the different fluorine and of (ii) the 19F chemical shifts vs 1/c for F8, for all the investigated systems. Torsional energy profile of PFO vs the C5C6-C7C8 dihedral angle. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2010, 26(22), 16762–16770