Fluorinated Polar Heads Can Strikingly Increase or Invert the Dipole

Jan 17, 2007 - The N-trifluroethyl docosanamide (fea) and N-ethyl docosanamide (ea) were synthesized by Dr. R. Wagner and Mrs. Y. Wu at the Max-Planck...
0 downloads 0 Views 808KB Size
Download PDF
Langmuir 2007, 23, 2623-2630

2623

Fluorinated Polar Heads Can Strikingly Increase or Invert the Dipole Moments at the Langmuir Monolayer-Water Boundary: Possible Effects from Headgroup Conformations J. G. Petrov,* E. E. Polymeropoulos,† and H. Mo¨hwald Max-Planck Institute of Colloids and Interfaces, D-14476 Golm/Potsdam, Germany ReceiVed October 26, 2006. In Final Form: NoVember 27, 2006 The dipole potential of lipid monolayers and bilayers is positiVe toward their nonpolar moiety. In previous papers, we have shown that designed molecules with fluorinated polar heads can invert the polarity of un-ionized Langmuir films. Monolayers of long-chain trifluoroethyl ester RCOOCH2CF3 and trifluoroethyl ether ROCH2CF3 exhibit large negatiVe ∆V values, shifted by 150-200% from the positive dipole potentials of their non-fluorinated analogs (Petrov and Mo¨hwald J. Phys. Chem. 1996, 100, 18458; Petrov et al. J. Phys. Chem. B 2005, 109, 14102). Here we report large positiVe surface (dipole) potentials of monolayers of N-trifluoroethyl docosanamide RCONHCH2CF3 and a 300% ∆V shift with respect to the non-fluorinated N-ethyl docosanamide films. Comparing the dipole potentials and normal dipole moments of the RCONHCH2CF3 and RCOOCH2CF3 monolayers and the maps of the local electrostatic potential (MEP) and lipophilicity (MLP) of their molecules in vacuum, we conclude that the opposite ∆V shifts and the difference of 1480 mV between the films of these structurally similar amphiphiles seem to be due to strongly different conformations of their heads. The large positive ∆V values of the N-trifluoroethyl amide monolayer was related to the network of sNH...OdCs bonds fixing the orientation of the hydrophobic δ+C-F3δ- dipoles toward water. The trifluoroethyl ester heads do not form H-bonds and can adjust their energetically optimal conformation orienting the hydrophobic δ+C-F3δ- dipoles toward air. The opposite signs of the dipole potential and the apparent normal dipole moments of the trifluoroethyl ester and ethyl ester monolayers were explained via energy minimization of 36 upright closely packed molecules with “hook-like” heads. The equilibrium architecture of this ensemble shows statistical distribution of the headgroup conformations and a nano-rough monolayer-water boundary as known from X-ray reflectivity experiments and molecular dynamic simulations of phospholipid monolayers and bilayers. The average of the vertical molecular dipole moments at equilibrium agree fairly well with the measured values of µ⊥, and the mean molecular area in the ensemble 19.3 Å2 matches the value of 18.9 ( 0.2 Å2 determined via X-ray diffraction at gracing incidence surprisingly well. These results reflect the balance of the attractive and repulsive forces between the closely packed “dry” amphiphilic molecules, but a more sophisticated molecular modeling explicitly including water would better serve to reveal the mechanism of the observed effects.

Introduction Amphiphilic substances having a hydrophobic moiety and a polar or charged hydrophilic “head” form insoluble Langmuir monolayers and Gibbs adsorption layers at the air-water and oil-water interfaces, micelles, uni- or multi-lamellar liposomes, and liquid crystals in solution. These molecular assemblies have a wide spectrum of applications as biomembrane models, drug carriers, detergents, lubricants, optoelectronic devices, etc. The electrostatic potential arising from the dipoles at the headswater boundary is an important characteristic of these systems.1-4 It determines the binding energy and transport kinetics of hydrophobic ions and ion carriers at/through membranes,5 the * Corresponding author. E-mail: [email protected]. Present address: Institute of Biophysics, Bulgarian Academy of Sciences, 1 Acad. G. Bonchev Str., Block 21, 1113 Sofia, Bulgaria. † Present address: Zentaris GmbH, Weismu ¨ llerstr. 50, 60314 Frankfurt, Germany. (1) (a) Brockman, H. Chem. Phys. Lipids 1994, 73, 57. (b) Brockman, H. Curr. Opin. Struct. Biol. 1999, 9, 438-443. (c) Brockman, H. L.; Applegate, K. R.; Momsen, M. M.; King, W. C.; Gromset, G. A. Biophys. J. 2003, 85, 2384-2396. (2) (a) Clarke, R. J. AdV. Colloid Interface Sci. 2001, 89-90, 263. (b) Clarke, R. L. Biochim. Biophys. Acta 1997, 1327, 269. (c) Clarke, R. J.; Luepfert Ch. Biophys. J. 1999, 76, 2614. (d) Schambereger, J.; Clarke R. J. Biophys. J. 2002, 82, 3081. (e) Starke-Peterkovic, T.; Turner, N.; Vitha, M. F.; Waller, M. P.; Hibbs, D. E.; Clarke, R. J. Biophys. J. 2006, 90, 4060-4070. (3) (a) Taylor, D. M. AdV. Colloid Interface Sci. 2000, 87, 183. (b) Taylor, D. M. Thin Solid Films 1998, 331, 1. (4) Dynarowicz-Latka, P.; Dhanabalan, A.; Oliveira, O. N., Jr. AdV. Colloid Interface Sci. 2001, 91, 221. (5) Flewelling, R. F.; Hubbell, W. L. Biophys. J. 1986, 49, 541.

adsorption of amphiphilic peptides,1b,6-8 and the enzymatic activity at liquid interfaces.9 Numerous studies show that the polarization of the hydration water of the headgroups controls the short-range repulsion between hydrophilic surfaces.10,11 The dipole potential of lipid bilayers and Langmuir monolayers12 is positiVe toward their nonpolar moiety.1-4,11-18 Fluorination of the terminal methyl group or the upper part of the hydrocarbon chain reverses the dipoles at the monolayer-air boundary from positive to negative.19-22 Theoretical and (6) Cafiso, D. S. Influence of charges and dipoles on macromolecular adsorption and permeability. In Permeability and Stability of Lipid Bilayers; DiSalvo, E. A., Simon, S. A., Eds.; CRC Press: Boca Raton, FL, 1995. (7) Cladera, J.; O’Shea, P. Biophys. J. 1998, 74, 2434. (8) Allende, D.; Vidal, A.; Simon, S. A.; McIntosh, T. J. Chem. Phys. Lipids 2003, 122, 65. (9) Maggio, B. J. Lipid Res. 1999, 40, 930. (10) Belaya, M. L.; Feigel’man, M. V.; Levadnyii, V. G. Langmuir 1987, 3, 648. (11) Simon, S. A.; McIntosh, T. J. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 9263. (12) Usually ∆V designates the surface potential of Langmuir and Gibbs monolayers, but here we use the term “dipole potential” adopted for model and native membranes1,2,18 to stress that it originates from the dipoles introduced by the amphiphiles at the interface. (13) Papahadjiopoulos, D. Biochim. Biophys. Acta 1968, 163, 240. (14) Smaby, J. M.; Brockman, H. L. Biophys. J. 1990, 58, 195. (15) Adam, N. K.; Askew, F. A.; Danielli, J. F. Biochem. J. 1935, 29, 1786. (16) Stenhagen, E. Determination of Organic Structures by Physical Methods; Braude, E. A., Nachod, F. C., Eds.; Acadmic: New York, 1955. (17) Demchak, R. J.; Fort, T., Jr. J. Colloid Interface Sci. 1974, 46, 191. (18) Haydon, D. A.; Hladky, S. B. Q. ReV. Biophys. 1972, 5, 187-282. (19) Fox, H. W. J. Phys. Chem. 1957, 61, 1058. (20) Bernett, M. K.; Zisman, W. A. J. Phys. Chem. 1963, 67, 1534.

10.1021/la063135c CCC: $37.00 © 2007 American Chemical Society Published on Web 01/17/2007

2624 Langmuir, Vol. 23, No. 5, 2007

experimental studies23-25 show that remote dipoles located at 15-20 Å from the water substrate do not affect the polarity and dielectric permittivity of the hydrated headgroups. To achieve a maximum modification of this biorelevant region we synthesized and studied two long-chain amphiphiles with Cδ+-F3δ- dipoles in the hydrophilic heads.26-30 Langmuir monolayers of the trifluoroethyl ester of docosanoic acid RCOOCH2CF3 and docosyl trifluoroethyl ether ROCH2CF3 showed large negatiVe dipole potentials ∆V at the air-water interface and a negatiVe ∆V shift of 150-200% from the positive dipole potentials found for the non-fluorinated analogs RCOOCH2CH3 and ROCH2CH3.26,30 In this paper, we report large positiVe dipole potentials of Langmuir films of a new amphiphile with fluorinated head, the Ntrifluoroethyl amide of docosanoic acid, RCONHCH2CF3, whose monolayers show a dramatic positiVe ∆V shift of 300% with respect to the non-fluorinated RCONHCH2CH3 films. The surface pressure-molecular area π/A and dipole potential-molecular area ∆V/A isotherms of these monolayers were compared, and their differences to the corresponding dependencies of the trifluoroethyl ester and ethyl ester films were analyzed. The four amphiphiles were compared also on the molecular level via maps of the molecular electrostatic potential (MEP) and molecular lipophilicity (MLP), displaying the local electrostatic potential and the local lipophilicity pseudo-potential on the solvent accessible molecular surface, respectively.31-38 The aim was to understand why the same CF3 for CH3 substitution in the similar N-trifluoroethyl amide and trifluoroethyl ester heads caused opposite shifts of the dipole potential of their monolayers. The experimental data and the molecular models suggest that the dramatic difference of 1480 mV between the dipole potentials of the N-trifluoroethyl amide and trifluoroethyl ester monolayers seems to be due to strongly different conformations of their hydrophilic heads. The positive ∆V value of the N-trifluoroethyl amide film was related to the formation of a network of sNH...Od Cs bonds39-41 in the monolayers fixing the Cδ+-F3δ- terminals of the heads to point with the negative ends toward water. The trifluoroethyl and ethyl ester heads do not form H-bonds41-43 (21) Vogel, V., Moebius, D. J. Colloid Interface Sci. 1988, 126, 408. (22) McIntosh, T. J.; Simon, S. A.; Vierling, P.; Santaella, C.; Ravily, V. Biophys. J. 1996, 71, 1853. (23) Taylor, D. M.; Bayes, G. F. Phys. ReV. E 1994, 49, 1439. (24) Taylor, D. M.; Bayes, G. F. Mater. Sci. Eng. C 1999, 8-9, 65. (25) Petrov, J. G.; Polymeropoulos, E. E.; Mo¨hwald, H. J. Phys. Chem. 1996, 100, 9860. (26) Petrov, J. G.; Mo¨hwald, H. J. Phys. Chem. 1996, 100, 18458. (27) Petrov, J. G.; Polymeropoulos, E. E.; Mo¨hwald, H. Langmuir 2000, 16, 7411. (28) Petrov, J. G.; Brezesinski, G.; Krasteva, N.; Mo¨hwald, H. Langmuir 2001, 17, 4581. (29) Petrov, J. G.; Brezesinski, G.; Andreeva, T. D.; Mo¨hwald, H. J. Phys. Chem. 2004, 108, 16154-16162. (30) Petrov, J. G.; Andreeva, T. D.; Kurt, D.; Mo¨hwald, H. J. Phys. Chem. B 2005, 109, 14102. (31) (a) MNDO. MOPAC (version 5.0) QCPE, No. 455, 1989. (b) Dewar, M. J. S.; Thoel, W. W. J. Am. Chem Soc. 1985, 107, 4899. (32) Connoly, M. L. Science 1983, 221, 709. (33) Ghose, A.; Crippen, G. J. Comput. Chem. 1986, 7, 565. (34) Viswandhan, V. N.; Ghose, A. K.; Revankar, G. R.; Robins, R. K. J. Chem. Inf. Comput. Sci. 1989, 29, 163. (35) Heiden, W.; Moeckel, G.; Brickmann, J. J. Comput.-Aided Mol. Des. 1993, 7, 503. (36) Heiden, W.; Brickmann, J. J. Mol. Graphics 1994, 12, 106. (37) Waldherr-Teschner, M.; Goetze, T.; Heiden, W.; Knoblauch, M. Volhardt, H.; Brockmann, J. In AdaVances in Scientific Visualisation; Post, F. H., Hin, A. J., Eds.; Springer: Heidelberg, 1992, pp 58-67. (38) Testa, B.; Carrupt, P.-A.; Gaillard, P.; Tsai, R.-S. In Lipophilicity in Drug Action and Toxicology; Pliska, V., Testa, B., Van Waterbeemd, H., Eds.; VCH: Weinheim, 1996; Chapter 4, pp 49-71. (39) Alexander, A. E.; Rideal, E. K. Nature 1941, 147, 541. (40) Alexander, A. E. Proc. R. Soc. London 1942, A179, 470. (41) Gehlert, U.; Weidemann, G.; Vollhardt, D.; Brezesinski, G.; Wagner, R.; Mo¨hwald, H. Langmuir 1998, 14, 2112. (42) Gericke, A.; Huenerfuss, H. Ber. Bunsen-Ges. Phys. Chem. 1995, 99, 641.

PetroV et al.

and can adjust their conformation via rotation around the covalent bonds. The opposite sign of the dipole potential of the trifluoroethyl ester and ethyl ester monolayers could be explained on the basis of a simple model of an energy minimized ensemble of 36 closely packed upright molecules with “hook-like” heads. The simulation shows a statistical distribution of the equilibrium conformations of the headgroups and a nano-rough monolayerwater boundary yielding reasonable mean values of the apparent normal dipole moment of the monolayers. Experimental Section Materials and Methods. The N-trifluroethyl docosanamide (fea) and N-ethyl docosanamide (ea) were synthesized by Dr. R. Wagner and Mrs. Y. Wu at the Max-Planck Institute of Colloids and Interfaces, Golm, Germany. The melting temperatures are 96-97 °C for fea and 89-90 °C for ea. Elemental analysis. fea: 68.22% C, 10.79% H, 3.54% N. Calcd 68.37% C, 11.00% H, 3.32% N. ea: 77.81% C, 13.04% H, 4.09% N. Calcd 78.41% C, 13.43% H, 3.81% N. The IR spectra in KBr and the NMR spectra confirmed the products. The synthesis and characterization of the trifluoroethyl docosanoate (fee) was described in a previous publication,26 and the ethyl docosanoate (ee) was purchased from Sigma and used as received. The amphiphiles were spread on Milli-Q Millipore water as 1 mM chloroform solution at initial area of 60-70 Å2/ molecule. After 5 min of evaporation of the solvent, the monolayers were compressed at a velocity of 2.8 Å2/molecule‚min-1. The π/A and ∆V/A isotherms were recorded at 23 °C with a Langmuir film balance with Wilhelmy dynamometric system and a Kelvin type vibrating electrode. The maps of molecular electrostatic potential (MEP) of single molecules in vacuum were generated by means of the MOLCAD graphic program31 from the partial atomic charges calculated via MNDO. The maps of molecular lipophilicity (MLP) were determined also via MOLCAD assuming additive atomic contributions.31 Details on the MEP and MLP preparation can be found in a previous publication.25

Results and Discussion Comparison of the π/A and ∆V/A Isotherms of the Amide and Ester Monolayers. Figure 1 presents the π/A and ∆V/A isotherms of the fluoroethyl amide (fea) and ethyl amide (ea) monolayers. Neither π/A curve shows the typical surface pressure maximum at collapse. The points at which the slope dπ/dA starts decreasing (see the horizontal arrows) indicate the reversal of the sign of d2π/dA2 and the loss of mechanical stability of the monolayers. They correspond to the maximum surface pressure πm and the minimum molecular area Am, before the 2D-3D transition. The values of πm ) 21 mN/m for the fea monolayers and of πm ) 24 mN/m for the ea monolayers suggest slightly smaller affinity of the fluorinated amide heads toward water. The larger molecular area of fea, Am(fea) ) 20.0 Å2 versus Am(ea) ) 18.7 Å2, could mean tilted hydrocarbon chains or existence of voids in the fea film. The values of the dipole potential at Am, corresponding to maximum monolayer density, are ∆Vm ) +1130 mV for the fea and ∆Vm ) +280 mV for the ea film (see the vertical arrows).44 They show that the CF3 for CH3 substitution in the N-ethyl amide head dramatically increases the positive dipole potential by +850 mV or 300%. Figure 2 presents the π/A and ∆V/A isotherms of the monolayers of the docosanoyl trifluoroethyl ester (fee) and ethyl ester (ee). The characteristics of the fee monolayer at maximum 2D density are πm ) 38 mN/m, Am ) 19.1 Å2, and ∆Vm ) -355 mV (see (43) Mendelsohn, R.; Brauner, J. W.; Gericke, A. Annu. ReV. Phys. Chem. 1995, 46, 305. (44) The presented value of ∆Vm for the N-ethyl amide monolayer agrees well with ∆V ) +290 mV reported years ago by Mitchell for monolayers of the methyl amide of stearic acid at pH 6.0. (Mitchell, J. S. Tabulae Biolojicae 1939, 19, 276.)

Fluorinated Polar Heads

Langmuir, Vol. 23, No. 5, 2007 2625

Figure 1. Surface pressure π and dipole potential ∆V vs average molecular area A of Langmuir monolayers of RCONHCH2CF3 (a) and RCONHCH2CH3 (b). The arrows indicate the values corresponding to maximum density of the monolayers.

the arrows). Their comparison to the values of the ee film (πm ) 46 mN/m, Am ) 19.0 Å2, ∆Vm ) +335 mV) delineates the effect of the headgroup fluorination in this system. The value of πm of the fee monolayer is significantly lower, indicating considerable decrease of the affinity of the fluorinated head toward water. However, here the same CF3 for CH3 substitution reverses the sign of the dipole potential and changes its magnitude by -700 mV or 210%. Repeated measurements give the scatter of the above characteristics of the monolayers for at least five isotherms (Table 1). The minimum molecular areas of the ester films reproduce within 0.2 Å2; those of the amide monolayers are slightly more scattered. All mean values of A h m coincide within the scatter limits with the real molecular areas AXY determined via GIXD for the compact monolayers,28,45 showing negligible effect of h m for the voids on A h m. Therefore, the larger value of A N-trifluoroethyl amide monolayer indicates tilted closely packed hydrocarbon chains. Since the cross-section of an all-trans chain is 18.5 Å2, the tilt angle from the normal to the film is τ ) arccos(18.5/20.0) ) 22.3°. Under the same conditions the N-ethyl amide and both trifluoroethyl and ethyl ester monolayers consist of upright closely packed molecules. Comparison of the Apparent Dipole Moments of the Amide and Ester Monolayers. From the values of ∆Vm and Am one can determine the vertical components of the apparent molecular dipole moment at maximum monolayer density:

µ⊥,m ) 0Am∆Vm

(1)

Here 0 is the absolute permittivity of vacuum and  is the relative permittivity of the monolayer averaging the contributions of the hydrocarbon tails, polar heads, and hydration water and accounting (45) Preliminary GIXD data for fea and ea monolayers.

Figure 2. Isotherms of surface pressure π and dipole potential ∆V vs average molecular area A of Langmuir monolayers of RCOOCH2CF3 (a) and RCOOCH2CF3 (b). The arrows indicate the values corresponding to the maximum density of the monolayers. Panel a is reproduced with permission from J. Phys. Chem. 1996, 100, 18458-18463. Copyright 1996 by the American Chemical Society. Table 1. Comparison of the Structural and Electrostatic Characteristics of the Amide and Ester Monolayers at Maximum Densitya substance

A h m (Å2)

RCONHCH2CF3 RCONHCH2CH3 RCOOCH2CF3 RCOOCH2CH3

20.5 ( 0.5 18.8 ( 0.6 19.1 ( 0.2 19.2 ( 0.2

a

AXY (Å2) ∆Vm (mV) 20.0 18.6 18.9 18.8

µ j ⊥,m ( ) 7) (D)

1100 ( 40 4.06 ( 0.10 280 ( 10 0.98 ( 0.03 -355 ( 10 -1.26 ( 0.03 345 ( 10 1.20 ( 0.03

Designations of the columns are explained in the text.

for the dipole-dipole interactions in the films. One often sets  ) 1 following Helmholtz, but the obtained values of µ⊥,m are much lower than the molecular dipole moments determined independently.46 Summarizing the results of numerous experiments Adam et al.47 concluded that a value of  ≈ 5 ÷ 10 is more appropriate for uncharged condensed monolayers. The last column h m was of Table 1 shows µ⊥,m data calculated with eq 1 at  ) 7. A replaced by AXY to completely exclude the effect of the voids. They do not contribute to ∆Vm but could make the average molecular area Am measured in the Langmuir trough larger than the real molecular area AXY determined by GIXD. Such substitution of AXY for Am yields apparent dipole moments µ⊥,m, which are strictly representative for the compact parts of the film even in presence of voids. The last column of Table 1 shows that the fluorination of the N-ethyl amide and ethyl ester heads dramatically changes the (46) Gaines, G. L., Jr. Insoluble Monolayers at Liquid-Gas Interfaces; Interscience Publishers: New York, 1966; p 190. (47) Adam, N. K.; Danielli, J. F. Harding, J. B. Proc. R. Soc. London A 1934, 147, 491.

2626 Langmuir, Vol. 23, No. 5, 2007

PetroV et al.

magnitude of the apparent dipole moment at the monolayerwater boundary, but this change has opposite signs. For the N-amide films the shift ∆µ⊥,m ) +3.08 ( 0.13 D is positiVe, whereas for the ester monolayers ∆µ⊥,m ) -2.46 ( 0.03 is negatiVe. The effect of the fluorination of the headgroups can be also analyzed on the basis of the three-capacitor model of condensed uncharged monolayers,17 which separates µ⊥ and  in contributions of hydration water (w), headgroups (h), and hydrocarbon tails (t):

µ⊥ µw µh µt + + ) 0A∆V )  w h t

(2)

The original paper of Demchak and Fort17 gives (µw/w) ) 0.040 D, h ) 7.6, and t ) 5.3 and suggests that these values are universal for condensed uncharged monolayers with different headgroups. Oliveira et al.48 proposed other values, (µw/w) ) -0.065 D, h ) 6.4, and t ) 2.8, but retained the universal character of these parameters. Such a postulate implies that the fea, ea, fee, and ee monolayers having the same hydrocarbon chains but different uncharged headgroups should have the same values of (µw/w) and (µt/t). The mean of the above values of h ) 7.0 yields the same opposite shifts for the amide and ester films, ∆µ⊥,h(amides) ) +3.08 D and ∆µ⊥,h(esters) ) -2.46 D but related to the dipole moments of the headgroups. The dramatic difference, µ⊥,h(fea) - µ⊥,h(fee) ) +5.4 D, between the vertical dipole moment components of the N-trifluoroethyl amid and trifluoroethyl ester films suggests that the two fluorinated heads have strongly different conformations. Different polar heads should have different structure of the hydration shells, so that the assumption of universal µw/w values is not sound enough. This problem is avoided in the papers of Burnett and Zisman20 and Vogel and Moebius,21 who do not distinguish between the electrostatic contributions of the headgroups and the aqueous subphase and present µ⊥ as a sum of the dipole moments of hydrated heads and the hydrocarbon tails. However, they do not specify the dielectric permittivity of these regions and use the Helmholtz value  ) 1 that yields effective normal dipole moments. Maps of the Electrostatic Potential and Lipophilicity of the Amide and Ester Molecules in Vacuum. Figure 3 presents the maps of molecular electrostatic potential (MEP) and molecular lipophilicity potential (MLP) of the non-fluorinated and fluorinated amide and ester molecules whose heads have the same conformation. The conclusions presented below hold also for other conformations if they are the same for all ea, ee, fea, and fee molecules. The color scales on the right-hand side should facilitate their comparison. The MEPs in panel a show that the non-fluorinated ethyl amide (ea) and ethyl ester (ee) have rather similar potential (respectively charge) distribution, implying the same sign and close magnitudes of the molecular dipole moments components along the chains µ|. This similarity concurs the experimental values µ⊥,m( ) 7) ) 0.98 D (ea) and 1.20 D (ee). However, close similarity of the MEPs can also be seen for both fluorinated amphiphiles although the experimental dipole moments, µ⊥,m( ) 7) ) +4.06 D (fea) and -1.26 D (fee), strongly differ in magnitude and have opposite signs. Comparison of the MEPs in the headgroups region of the N-trifluoroethyl amide (fea) and trifluoroethyl ester (fee) suggests that the CF3 group does not yield opposite signs of the dipole (48) Oliveira, O. N., Jr.; Taylor, D. M.; Lewis, T. J.; Salvagno, S.; Stirling, C. J. M. J. Chem. Soc. 1989, 85, 1009.

moments of the heads µh if their conformation is the same. The fea and fee chains are polarized to the same extent and should yield the same dipole moments of the tails µt. Therefore, for the same conformation of the fea and fee heads the sum µh/h + µt/t in eq 2 cannot have different signs because h and t are positive. Such polarization of the hydrocarbon chains effectively yields negative dipole moment contributions of the tails µt/t because the CH2 groups closer to the heads are more positive than the remote ones (see the color scale). Since the polarization, and hence the negative contribution of the fee tail exceeds the negative contribution of the ee tail, their difference could explain the negative sign of µ⊥,m obtained for the trifluoroethyl ester monolayer. However, practically the same difference between the polarization of the tails µt/t, can be seen respectively for the fea and fee chains, although the values of µ⊥,m for these monolayers are positive. Therefore, the polarized chains do not determine the molecular dipole moment components along the chains µ|. Molecular dynamic simulations of phosphatidylcholine bilayer-water boundary showed that the positive sign of the dipole potential originates from the hydration water dipoles, which overcompensate the headgroups dipoles.49,50 This result supports the conclusion of the NMR and X-ray experimental study of Gawrish et al.51 that the hydration water yields the major contribution to the dipole potential of the phosphocholine bilayer. The above scenario cannot be automatically transferred to our monolayers because their polar heads are much less hydrophilic and their close packed crystalline structure does not allow penetration of water molecules. For this reason we compare the MLPs of the fea, ea, fee, and ee molecules presented in Figure 3b, which characterize the affinity of the headgroups toward water at the same conformation of the heads.38 It can be seen that the headgroup regions of the MLPs significantly differ from each other demonstrating different hydrophilicity. However, the MLP of the most hydrophilic ethyl amide head (ea) by far differs from the MLP of the ethyl ester head (ee), in spite of the same sign and close experimental values of the apparent dipole moments, µ⊥,m(ea) ) 0.98 D and µ⊥,m(ee) ) 1.20 D. In contrast, the local lipophilicity distributions of the terminals of the ethyl ester (ee) and trifluoroethyl amide (fea) heads do not differ substantially, whereas the corresponding experimental values of µ⊥,m are 1.20 and 4.06 D, respectively. The largest lipophilicity difference between the MLP of the ethyl amid (ea) and trifluoroethyl ester (fee) heads does not correlate with the largest experimental difference µ⊥,m(fea) - µ⊥,m(fee) ) +5.4 D between the trifluoroethyl amide and trifluoroethyl ester heads. Therefore, different hydrophilicity of the heads, can merely explain the dramatic effect of their fluorination if the conformation is the same. Headgroup Conformations and a Possible Mechanism of the Fluorination Effect. The early investigations52-54 compared the experimental value of µ⊥,exp( ) 5 ÷10) with µ⊥,clc values calculated via vector summation of the dipole moments of the chemical bonds or atomic groups at different conformations of the heads. It was assumed that the coincidence between of µ⊥,exp and µ⊥,clc indicates the correct conformation of the headgroups. On this basis Alexander and Schulman52 concluded that at (49) Essman, U.; Perera, L.; Berkowitz, M. L. Langmuir 1995, 11, 45194531. (50) Berkowitz, M. L.; Bostick D. L.; Pandit, S. Chem. ReV. 2006, 106, 15271539 (and the references cited therein). (51) Gawrisch, K.; Ruston, D.; Zimmerberg, J.; Parsegian, V. A.; Rand, R. P.; Fuller, N. Biophys. J. 1992, 61, 1213. (52) Alexander, A. E.; Schulman, J. H. Proc. R. Soc. London 1937, A161, 115. (53) Fort, T.; Alexander, A. E. J. Colloid Sci. 1959, 14, 190. (54) Davis, J. T.; Rideal, E. K. Interfacial Phenomena; Academic Press: New York, 1961; p 72.

Fluorinated Polar Heads

Langmuir, Vol. 23, No. 5, 2007 2627

Figure 3. (a) Maps of the local electrostatic potential (MEP) of ethyl amide (ea), ethyl ester (ee), trifluoroethyl amide (fea), and trifluoroethyl ester (fee) molecules in vacuum. (b) Maps of the local molecular lipophilicity potential (MLP) of the same molecules in vacuum. All molecular models have the same conformation of the polar heads.

70 Å2/molecule the ethyl radicals in the heads of ethyl stearate monolayers point toward air (“E” conformation in Figure 4), but at 19-20 Å2/molecule they accommodate themselves beneath the long chains in order to ensure their closest packing (“Z” conformation in Figure 4). The above steric restriction substantiated the choice of the energetically unfavorable Z conformation, which could qualitatively explain the reduced rate of hydrolysis in solid condensed ethyl ester films via screening of the carbonyl bond by the ethyl radicals.52,55 The orientation of the strong Cδ+-F3δ- and H2Cδ+-CF3δdipoles toward water could explain the large positive shift of the dipole potential and apparent dipole moment of the N-trifluoroethyl amide monolayers with respect to the N-ethyl amide films illustrated by Figure 2 and Table 1. However, it contradicts the negative shift between the trifluoroethyl ester and ethyl ester (55) Alexander, A. E.; Rideal, E. K. Proc. R. Soc. London 1937, A163, 70.

monolayers (cf. Figure 2 and Table 1). We rationalize this contradiction by assuming that the conformations of the N-trifluoroethyl and N-ethyl amide heads are the same and close to the Z conformation, but the conformation of the trifluoroethyl ester heads is significantly different. This difference could be due to intermolecular sNH...OdCs bonds creating a lateral network in the monolayers39-41 and restricting the rotation around the C-C bonds in the amide heads. The trifluoroethyl and ethyl ester monolayers do not form H-bonds and networks between their heads and such rotation around the C-C bonds is possible.41-43 The MLPs in Figure 3 show that the CF3-terminal of the trifluoroethyl ester head is strongly hydrophobic56,57 and “prefers” (56) This result is not as obvious as usually assumed, because the CF3 group in the trifluoromethylalkanes decreases their lipophilicity due to the strong Cδ+CF3δ- dipoles.57 On the other side, the CF3CH2OH is considerably more lipophilic than CH3CH2OH (cf. Table 6 in ref 57).

2628 Langmuir, Vol. 23, No. 5, 2007

Figure 4. (a) Schematic presentation of the conformations of the hydrophilic head of ethyl stearate monolayers adopted by Alexander and Schulman52 at 70 Å2/molecule (E) and 20 Å2/molecule (Z). R ) C17H35, Eth ) C2H5. (b) “Hook-like” conformation of the longchain trifluoroethyl ester molecule adopted in the molecular simulation.

to be located in the region of the hydrocarbon chains. Such orientation was found by Adam58 for fatty acid esters with long alkyl chains, but it doubled the area per molecule. Keeping in mind this fact, we performed a complete conformational analysis and examined several energetically stable conformations of the trifluoroethyl ester molecule with CH2CF3 radicals beneath the acyl chains. They all gave positive dipole moments along the long chain µ|.59 For this reason we a priori choose the low energy conformation presented in Figure 4b, which accounts for the hydrophobicity of the CF3 terminal and yields a negative dipole moment along the acyl chain µ| ) - 0.80 D. Thirty-six such molecules were arranged in closely packed upright pattern and the energy of the ensemble was minimized using the TRIPOS force field.60 Figure 5a illustrates the side and top views of the initial configuration; Figure 5b presents the side and top views of the final configuration. One can see that some of the equilibrium conformations differ from the initial one, but the close packing of the long hydrocarbon chains is preserved. The equilibrium values of µ| were calculated for each of the 16 molecules surrounded by neighbors on all sides; the molecules in the periphery of the ensemble were excluded because the absence of neighbors on at least one side changed their conformation. The individual µ| values were scattered, some of them being even positive, but their mean value is negative, µ j | ) -0.66 D. Moreover, the same hook-like conformation of an ethyl ester molecule has a positive dipole moment along the chain µ| ) + 0.60 D, and the energy minimized ensemble of 36 such molecules gives a positive mean dipole moment µ j | ) + 0.36 D for the of 16 inner equilibrium conformations. The calculated mean dipole moment along the chains µ j| ) -0.66 D has the same sign, and its magnitude reasonably agrees with the measured apparent dipole moment of the trifluoroethyl (57) Smart, B. E. J. Fluorine Chem. 2001, 109, 3-13. (58) Adam, N. K. Proc. R. Soc. London 1929, A126, 366. (59) The MNDO calculation of µ| is based on the atomic point charges of energetically optimized molecular geometries but also accounts for hybridization effects which sometimes yield up to 50% of the dipole moment. Details of this procedure are given in ref 25. (60) Tripos, Inc. SYBYL Molecular Modeling Software; Tripos: St Louis, MO.

PetroV et al.

ester monolayer, µ⊥,m( ) 7) ) - 1.26 D, whose molecules are upright and closely packed as in the above ensemble.28 The calculated mean value µ j | ) + 0.36 D for the ethyl ester monolayer, which has the same molecular structure at high surface pressure,28 agrees in sign with µ⊥,m( ) 7) )+ 1.20 D, but the difference of the magnitudes is larger. Recalling the empirical suggestion of Adam et al.47 that for uncharged condensed monolayers  ≈ 5 ÷10, one could find quantitative agreement setting (fee) ) 3.7 and (ee) ) 2.1 in eq 1. Having in mind the simplicity of our model, which implicitly accounts for the headwater interactions but neglects the contribution of the water dipoles, we believe that the correspondence of the experimental and the calculated dipole moments is quite satisfactory. The above simulation yields an interesting result about the average area occupied by a hook-like molecule in the ensemble at equilibrium. Figure 6 presents Figure 5b (bottom) in CPK format. The inner 16 molecules form imperfect hexagonal unit cells distorted to next-neighbors in surprising coincidence with the results of our GIXD investigation.28 The four inner cells give molecular areas of 19.0, 18.3, 18.9, and 20.9 Å2, whose average A h ) 19.3 Å2 quantitatively matches the molecular area of 18.9 ( 0.2 Å2 determined via GIXD at high-surface pressure for the fluorinated and non-fluorinated ethyl ester monolayers.28 Our molecular simulation experiment reveals the statistical character of the average normal dipole moment and dipole potential of condensed Langmuir monolayers. It shows that the equilibrium individual values of µ|| are different and that some of them have opposite signs, but their mean values semiquantitatively agree with the experiment. The statistical protrusion of the trifluoroethyl and ethyl radicals (Figure 5b, left) creates a nano-dimensional roughness of the monolayer-water boundary. This result correlates with the conclusions of the X-ray reflectivity studies of Langmuir films61 and molecular dynamic simulations of phospholipid monolayers and bilayers.49,50 The nano-roughness is a statistical alternative to the fixed Z conformation of the ethyl ester headgroups proposed by Alexander and Schulman.52 Such architecture does not reject the classical explanation52,55 of the slow hydrolysis of closely packed monolayers because the nanorough monolayer-water boundary also effectively screens the carbonyl bonds from the H+ ions attack. The simulation described above justifies the opposite signs of the apparent normal dipole moment µ⊥,m and dipole potential ∆Vm of the fluorinated and non-fluorinated ethyl ester monolayers, but a similar scenario should work for the trifluoroethyl and ethyl ether monolayers30 because their heads also allow free H2C-CF3 rotation. The results of the modeling reflect the balance of the attractive and repulsive forces between the closely packed amphiphilic molecules. In spite of this one should consider these results with caution until more sophisticated simulation calculations, explicitly including water, support or reject its significance.

Conclusions Substitution of the CH3 terminal of the polar head of the N-ethyl docosanamide by a CF3 group dramatically increases the positive values of the dipole potential and the apparent normal dipole moment µ⊥ of the Langmuir monolayers. The large positive ∆V shift is opposite of the negative shift of ∆V and µ⊥ caused by the same CF3 for CH3 substitution in the ethyl docosanoate head. The huge difference of +5.3 D between the apparent normal dipole moments µ⊥,m of the trifluoroethyl amide and trifluoroethyl (61) Als-Nielsen, J.; Moehwald, H. Handbook on Synchrotron Radiation; Ebashi, S., Koch, M., Rubenstein, E., Eds.; Elsevier Science Publishers: Amsterdam, 1991; Vol. 4, Chapter 1, pp 26,42.

Fluorinated Polar Heads

Langmuir, Vol. 23, No. 5, 2007 2629

Figure 5. (a) Initial arrangement of an ensemble of 36 fee or ee molecules; left, side view; right, top view. (b) Side view (left) and top view (right) of the equilibrium conformations after energy minimization of the ensemble.

Figure 6. Top view of the equilibrium packing of an ensemble of 36 “hook-like” trifluoroethyl docosanoate molecules presented in CPK format. The molecules surrounded by neighbors on all sides form a distorted hexagonal lattice with a mean area per molecule quantitatively matching the value obtained in the GIXD study28 of this monolayer.

ester monolayers at maximum density suggests different conformations of their hydrated headgroups.

The maps of the molecular electrostatic potential (MEP) of single N-trifluoroethyl amide (fea) and trifluoroethyl ester (fee)

2630 Langmuir, Vol. 23, No. 5, 2007

molecules having the same conformation of the heads imply the same sign and close magnitudes of µ⊥,m in contrast to the opposite signs and significantly different values found experimentally, µ⊥,m ) + 4.06 D (fea) and µ⊥,m ) - 1.26 D (fee). This contradiction supports the hypothesis that the trifluoroethyl amide and trifluoroethyl ester heads have different conformations. The maps of MLP of the N-trifluoroethyl amide (fea), N-ethyl amide (ea), trifluoroethyl ester (fee), and ethyl ester (ee) molecules show that the four headgroups having the same conformation significantly differ in affinity toward water. However, the hydrophilicity of heads does not correlate with the values of µ⊥,m and ∆Vm. The MLP of the ethyl amide head strongly differs from the MLP of the ethyl ester head, whereas the experimental dipole moments µ⊥,m(ea) ) 0.98 D and µ⊥,m(ee) ) 1.20 D have the same sign and close magnitudes. The largest experimental difference µ⊥,m(fea) - µ⊥,m(fee) ) +5.4 D, does not correlate with the largest difference between MLP of the N-ethyl amid and trifluoroethyl ester heads. Therefore, even a strong difference in the hydrophilicity of the heads cannot explain the dramatic opposite effects of the fluorination of N-ethyl amide and ethyl ester. The difference between the N-trifluoroethyl amide and trifluoroethyl ester monolayers seems to be due to sNH...Od Cs hydrogen bonds cross-linking the amide heads. They could fix the orientation of the Cδ+-(Fδ-)3 dipoles with the negative ends toward water, strongly increasing the positive values of µ⊥,m and ∆Vm. Since H-bonds do not form between the trifluoroethyl ester heads they can adjust optimal conformation via rotation of the C-C bonds. Comparing different stable conformations of the trifluoroethyl ester molecule in vacuum we found that a hook-like geometry satisfies the energetically preferred orientation of the hydrophobic CF3 terminals toward air and gives a negative dipole moment component along the hydrocarbon chain µ||. Energy minimization of an ensemble of 36 such molecules gives statistical distribution of the equilibrium headgroup conformations. The average of the individual equilibrium µ|| values of the 16 molecules surrounded by neighbors on all sides, µ j || ) -0.66 D, is negative and agrees relatively well with the experimental value, µ⊥,m( ) 7) ) -1.26 D. The agreement becomes quantitative for an effective dielectric permittivity of the monolayer  ) 3.7.

PetroV et al.

The same hook-like conformation of the ethyl ester molecule has a positive dipole moment component along the chain. Energy minimization of the ensemble of 36 molecules yields positive mean value of µ j || ) +0.36 D, which agrees tolerably with the experimental value of µ⊥,m ) + 1.20 D; the values match for  ) 2.1. Considering the simplicity of the model, which implicitly accounts for the headgroup-water interaction but neglects the contribution of the water dipoles, we believe that both calculated results are in good qualitative and fair quantitative agreement with the experimental ones. The CPK presentation of the equilibrium ensemble shows that the first neighbors of the inner 16 molecules form imperfect hexagonal unit cells distorted to next-neighbors in excellent agreement with the molecular structure obtained in our GIXD study of the fluorinated ethyl ester monolayer. The four inner cells give a mean molecular area of 19.3 Å2 that quantitatively matches the GIXD molecular area of 18.9 ( 0.2 Å2. This result rejects arguments that the close packing of the chains of the energetically favorable hook-like molecules is sterically restricted. Our molecular simulation reveals the statistical character of the average normal dipole moment and dipole potential measured for condensed Langmuir monolayers. The individual equilibrium molecular dipole moments along the chain µ|| have different magnitudes and even different signs but their average fairly well agrees with the experimental value. The statistical protrusion of the trifluoroethyl and ethyl terminals shows a nanodimensional roughness of the headgroups region, as suggested by X-ray reflectivity studies of Langmuir monolayers and molecular dynamic simulations of phospholipid monolayers and bilayers. Such roughness presents a statistical alternative to the well-liked Z conformation of headgroups of long-chain ethyl ester monolayers preserving the classical explanation of their slow hydrolysis via screening the carbonyl bonds. Acknowledgment. J.G.P. acknowledges the research stipend from the Alexander von Humboldt Foundation in 2004, which enabled finalization of this study. LA063135C