TMCnet News
Stratified Dipole-Arrays Model Accounting for Bulk Properties Specific to Perfluoroalkyl Compounds [ChemPlusChem](ChemPlusChem Via Acquire Media NewsEdge) Perfluoroalkyl compounds are known to exhibit a hydrophobic character on the surface of the material, although the C-F bond has a large dipole, which should make the molecular surface polar and hydrophilic. This inconsistency has long been a chemical matter to be solved. Herein, a stratified dipole-arrays model is proposed: the molecular polar surface can be fully hidden by forming a two-dimensional aggregate of perfluoroalkyl (Rf) groups ; this aggregate is spontaneously induced by dipole-dipole interaction arrays owing to the helical structure of the Rf group. In this model, a 'short' Rf group should play the role of a single Rf group with a hydrophilic character, whereas a 'long' Rf group should spontaneously form a hexagonal aggregate. To examine this model, Rf -containing myristic acids with various Rf lengths have been synthesized and their aggregation properties are analyzed by using the Langmuir monolayer technique aided by precise IR spectroscopic analysis. Keywords : bulk properties · fluorine · IR spectroscopy · monolayers · surface chemistry Perfluoroalkyl compounds are known to exhibit unique proper- ties, such as high melting temperature, high hydrophobicity, and low electric permittivity.[1, 2] Researchers have long tried to describe these material properties through the elemental char- acteristics of fluorine.[1] The large electronegativity of fluorine is sometimes considered to induce a partial negative charge on the molecular surface to make the London dispersive force weak, which explains the uniquely weak adhesive property of poly(tetrafluoroethylene) (PTFE ; Teflon). This mechanism, how- ever, is inconsistent with the high melting temperature of Teflon (ca. 327 8C). There has been no uniform theory to under- stand the bulk properties of fluorinated compounds. In the present study, a stratified dipole-arrays model is proposed, so that the inconsistency would readily be removed. To discuss the hydrophobicity, the opposite concept of hy- drophilicity is considered first. One of the most important fac- tors to generate hydrophilicity is electrical dipoles in a mole- cule. A water molecule, consisting of two O^H bonds, has a bent structure, which generates an apparent dipole over the molecule. Hydration occurs as a result of molecular interactions with water through dipole-dipole interactions, and thus, a com- pound with a large dipole has a hydrophilic character. When a compound has many dipoles, the collection of dipoles indu- ces polarity, which contributes to electric permittivity. There- fore, in general, a compound with a large electric permittivity exhibits a hydrophilic property. When a molecule has a highly symmetric structure, on the other hand, the spatial average of the dipoles makes the polar- ity weak, which yields a low permittivity and a hydrophobic property. Carbon tetrachloride is a representative compound : the four C^Cl bonds are arranged in the tetrahedral (Td) molec- ular symmetry, which yields a low electric permittivity (e = 2.24).[3] In a similar manner, normal alkyl chains with the D2h symmetry also exhibit a hydrophobic character, yielding a low permittivity (e = 2.27).[4] Regardless, a fully fluorinated alkyl (perfluoroalkyl ; Rf) group has a different character. Owing to the strongest electronegativity, an apparent elec- tric dipole is generated along the C^F bond. In addition, the van der Waals radius is about 1.2 times larger than that of hy- drogen, which results in a 1.7 times greater volume.[1, 5] As a result of both electrorepulsive and steric effects, the back- bone of an Rf group favors a twisted helical structure[1, 6-9] in- stead of a planer structure when the length is ^C6F13 or longer.[1, 10] The twisting rate depends on the crystal phase as a function of temperature and pressure,[11] and the direction of aCF2 group is rotated by 1808 over 13 CF2 groups (13/6 ; pha- se II) below a temperature of 19 8C under atmospheric pressure (Table 1).[6, 11] With the helical structure, the local dipoles cannot be cancelled out, which means that the 'molecular surface' re- mains polar (Figure 1 a). Figure 1 b shows a nearly top view of Rf-containing myristic acid, in which the bottom and top CF2 groups are drawn in dark orange with arrows representing the dipoles (red and blue, respectively). For better visibility, a simple schematic top view is available in Figure 1 c. In the crystal phase II, the helical structure of an Rf group has the 13/6 structure :[1, 6, 11] the dipole direction is rotated by 1808 for 12 CF2^CF2 'intervals' (Table 1). When an Rf group of ^(CF2)9CF3, which has 8 intervals, is con- sidered, as illustrated in Figure 1 a, the bottom and top dipoles are thus rotated by 1208. This situation leads to Rf groups with hexagonal packing owing to the 2D dipole-dipole interaction 'arrays' (pink and blue stripes in Figure 2 c). Rf groups are known to favor hexag- onal packing.[1] When the length, n,of^(CF2)nCF3 is nine or longer, the Rf chains spontaneously aggregate to fully hide the molecular polar surface, whereas a short Rf chain exhibits single-molecule character owing to weak aggregation proper- ties for maintaining the intrinsic polar character. In particular, for n ^ 3 (Figure 2 a), the polar character should explicitly remain because the molecule has a nearly planer zigzag struc- ture, which has much lower aggregation properties than the helical one.[1] We note the intermediate state has a twisting angle of 908 (Figure 2 b), which is realized by the Rf group of n= 7 (Table 1).[6] This Rf group is also expected to aggregate two-di- mensionally because the dipoles span the orthogonal two de- grees of freedom in space (plane). This molecular packing, however, requires tetragonal packing with a cross-sectional area larger than the hexagonal one by a factor of 2= 3. Be- cause this is not the closest packing, it should be less stable than the hexagonal one when the aggregate is two-dimen- sionally compressed. This is discussed below in relation to ex- perimental results. Another note is that the even-numbered CF2 groups direct the odd-numbered ones to the opposite side. Therefore, the twisting angle for the even-numbered CF2 group should be measured from the direction of the 'second' CF2 group. For ex- ample, the n = 10 compound has the 'tenth' CF2 group twisted by 1208 from the 'second' CF2 group, because there are 8 in- tervals, which is the same logic as that for an odd Rf length. Therefore, no even-odd effect is expected, which is confirmed by the dependence of the melting temperature on the Rf length.[12] As well as the two directions of the top and bottom CF2 groups (Figure 2 b or c), the rest of the CF2 groups individually spread into similar linear arrays with different directions. As a result, the polarity of the aggregate is weakened by the 2D direction average, which makes the electrical permittivity low. The high melting temperature property is also elucidated by systematic 2D interactions. In this manner, the Rf-compound- specific properties can readily be elucidated by 2D spontane- ous aggregation and not by a single molecule. Once the 2D aggregate is generated, only the ^CF3 group appears on the surface of the aggregate. Because the ^CF3 group is chemically connected to the ^(CF2)n^ group, the situa- tion can roughly be approximat- ed to be CF4, which is insoluble in water,[13] through similar logic to that used for CCl4. As a result, the surface of the 2D aggregate of the Rf groups should exhibit a hydrophobic or very weak hy- drophilic character. To examine the stratified dipole-arrays model, four com- pounds (n = 3, 5, 7, and 9) were prepared (for details see the Supporting Information). The ter- minus of myristic acid was par- tially fluorinated, such as, CF3(CF2)n^(CH2)m^COOH, in which n+ m =12 (Figure 1 a). Notably, normal myristic acid (nonfluorinated) generates a Langmuir monolayer[14] on water with a nearly perpendicular molecular stance when the monolayer area is decreased.[15] When a 'short' Rf group is introduced, however, the Rf part is expected to have the hydrophilic character of a single mole- cule (Figure 2 a), which makes both ends of the compound hy- drophilic. When the compound with n = 3 is spread on water, the molecules should lie on the water surface.[16] On the other hand, when a 'long' Rf group with n= 9 is introduced, the com- pound should spontaneously aggregate (Figure 2 c); this would result in the perpendicular molecular stance in the monolayer even at a low surface pressure. Figure 3 shows the surface pressure (p)-surface area (A) iso- therms of the monolayer prepared by using the four com- pounds. The molecular packing should mainly be governed by the Rf group because the Rf group has a 1.4 times larger cross- sectional area than that of the alkyl part.[17] When the monolay- er is compressed, the isotherms of n = 3 and 5 exhibit an in- crease in surface pressure, even at a large surface area (the gas phase), as shown in the inset in Figure 3, which indicates that the molecules are lying on the water surface.[16] Because non- fluorinated myristic acid did not exhibit a gas phase,[14] the lying molecular stance proves the hydrophilic properties of both ends of the compound. The liquid expansion (LE) region[14] is fairly large in the two isotherms, especially for n = 3. In addition, the limiting molecular area, that is, the cross-sec- tional area of a straight molecule is 0.482 (n = 3) and 0.402 nm2 (n= 5) per molecule, which is much larger than the theoretical cross section of 0.296 nm2 per molecule.[1, 18-20] In this manner, a short Rf group has poor aggregation and apparent hydrophil- ic properties ; these are supported by IR spectroscopy data dis- cussed below. When the n = 7 compound is employed, both the gas and LE phases disappear, and the limiting molecular area decreases to 0.364 nm2 per molecule. The n = 9 compound exhibits a fur- ther aggregated character : the slope of the linearly increasing part is steep, and the limiting molecular area is 0.296 nm2 per molecule. Here, the limiting molecular area of n = 7 may be physically ambiguous because tetragonal packing would be changed into hexagonal packing by further compression. The slope of the linear part of n=7 is smaller than that of n = 9, which means that the packing is not stable. In this case, in- stead of discussing the limiting molecular area, the surface area at an appropriate surface pressure should be discussed. The appropriate point should be the low-end linear part of the isotherm, which is marked by a red circle in Figure 3. At this surface pressure, the areas of n = 7 and 9 compounds are 0.286 and 0.330 nm2 per molecule, respectively. The ratio of the areas (1.15) perfectly agrees with the theoretical expecta- tion of 2= 3 = 1.15, which is another strong support for the 2D aggregation model. The molecular orientations suggested by the p-A isotherms were further investigated by using IR spectroscopy. The mono- layer was transferred onto a hydrophilic gold surface[21] by the Langmuir-Blodgett (LB) technique,[14, 22] and subjected to IR re- flection-absorption (RA) spectrometry.[23, 24] IR RA spectrometry has adequately high sensitivity for measuring a monolayer (Figure 4). Before discussing the IR RA spectra of the monolayers, the IR spectra of bulk solids should be discussed because a bulk solid yields a molecular-orientation-free spectrum of highly packed molecules. The bulk sample spectra were measured by using the attenuated total-reflection (ATR) technique.[24-26] No- tably, the ATR spectra are linear combinations of both trans- verse optic (TO ; Im(e)) and longitudinal optic (LO ; Im(^1/e)) energy-loss functions (e : complex electric permittivity).[24] Be- cause the RA spectra of a thin film are governed by the LO function only,[24, 27, 28] the TO contribution to the raw ATR spec- tra (Figure S1 in the Supporting Information) was removed to leave the LO energy-loss function spectra (Figure 5) by calcu- lating e from the Kramers-Kronig relationship[24,27,29] with n = 1.33.[2] The discussion of IR spectra of Rf compounds is different from that of normal alkyl compounds. As found in Figure 5, for example, the spectral pattern greatly changes with increasing the chain length. According to Koenig et al. ,[30] an Rf group generates 11 IR-active bands (GIvRib ¼ 3A2 þ 8E1) for both phase- s II and IV (15/7; > 19 8C), which correspond to the point groups D(12p/13) and D(14p/15), respectively. Of these many bands, the bands at circa ñ= 1340 and 1150 cm^1 are useful, as explained below. The band at circa ñ= 1340 cm^1 is influenced by the C^C^C symmetric stretching vibration (A1) and CF3 symmetric stretch- ing vibration (A2).[31] Fortunately, the A1 mode is IR inactive,[30] and therefore, this band can be recognized to be the CF3 sym- metric stretching vibration (ns(CF3)) band. Because the direction of the transition moment of the ns(CF3) mode is nearly parallel to the Rf group axis, this band can be used for orientation analysis. In the IR RA spectra of the monolayers (Figure 4), this band appears most strongly for n = 9. Because RA spectrome- try has a surface selection rule[23] that only the surface-normal component of a transition moment appears in the spectra, the strong appearance of this mode indicates that the Rf groups have a perpendicular stance to the monolayer surface. On the other hand, this band is largely suppressed for n = 3, which in- dicates that the Rf groups are lying parallel to the surface. The band at circa ñ= 1150 cm^1 is another useful band, which is assigned to the CF2 symmetric stretching vibration (ns(CF2); A2).[30, 32] Because an Rf group has a helical structure, the CF2-related vibrational modes are insensitive to the molec- ular orientation. Instead, the 'position' of this band can be used to discuss molecular packing in the LB monolayer by re- ferring to the spectrum (Figure 5). The monolayer band (Figure 4) for n =3 and 5 appears at an apparently higher posi- tion, by 10 and 9 cm^1, respectively, than that of the bulk band. In other words, short Rf groups have a weak aggregation property. On the other hand, as for the n = 7 compound, the positions of the bulk and monolayer bands become closer and they perfectly meet at ñ= 1153 cm^1 for n = 9. In this manner, the n =9 compound has an outstandingly strong spontaneous molecular aggregation property, as expected. In this manner, the newly proposed stratified dipole-arrays model for 2D molecular aggregation is the key to understand the specific bulk properties of the Rf compound. Rf groups with the magic number of n = 9 or longer will be put together spontaneously to generate a 2D aggregate that exhibits differ- ent characteristics from a single molecule. Acknowledgements This study was financially supported by a Grant-in-Aid for Scien- tific Research (B) (no. 23350031 (T.H.)) from the Japan Society for the Promotion of Science, and Priority Areas (23106710 (T.H.)) from the Ministry of Education, Science, Sports, Culture, and Tech- nology, Japan. We also thank the Sumitomo Foundation for fi- nancial support for our study (no. 120384 (T.H.)). This study was supported by the Collaborative Research Program of Institute for Chemical Research, Kyoto University (grant no. 2013-35 (M.S.)). [1] M. P. Krafft, J. G. Riess, Chem. Rev. 2009, 109, 1714 - 1792. [2] W. Liu, Y. Koike, Y. Okamoto, Macromolecules 2005, 38, 9466 - 9473. [3] P. Perez, T. E. Block, C. M. Knobler, J. Chem. Eng. Data 1971, 16, 333 - 335. [4] D. D. Popenoe, S. M. Stole, M. D. Porter, Appl. Spectrosc. 1992, 46,79- 87. [5] A. Bondi, Phys. Chem. 1964, 68, 441 - 451. [6] L. Piseri, B. M. Powell, G. Dolling, J. Chem. Phys. 1973, 58, 158 - 171. [7] C. W. Bunn, E. R. Howells, Nature 1954, 174, 549 - 551. [8] K. Ute, R. Kinoshita, K. Matsui, N. Miyatake, K. Hatada, Chem. Lett. 1992, 21, 1337 - 1340. [9] K. Monde, N. Miura, M. Hashimoto, T. Taniguchi, T. Inabe, J. Am. Chem. Soc. 2006, 128, 6000 - 6001. [10] P. Piaggio, P. G. Francese, G. Masetti, G. Dellepiane, J. Mol. Struct. 1975, 26, 421- 428. [11] a) E. S. Clark, Polymer 1999, 40, 4659 - 4665 ; b) A. A. P^dua, J. Phys. Chem. A 2002, 106, 10116 -10123. [12] N. O. Brace, J. Org. Chem. 1962, 27, 4491- 4498. [13] P. Scharlin, R. Battino, E. Silla, I. TuÇ^n, J. L. Pascual-Ahuir, Pure Appl. Chem. 1998, 70, 1895 - 1904. [14] W. D. Harkins, E. Boyd, J. Phys. Chem. 1941, 45, 20 - 43. [15] G. Brezesinski, B. Dobner, C. Stefaniu, D. Vollhardt, Langmuir 2011, 27, 5386-5392. [16] T. Takiue, T. Fukuda, D. Murakami, H. Inomata, H. Sakamoto, H. Matsu- bara, M. Aratono, J. Phys. Chem. C 2008, 112, 5078 - 5084. [17] Y. Oishi, T. Kato, T. Narita, K. Ariga, T. Kunitake, Langmuir 2008, 24, 1682-1685. [18] T. Takiue, T. Tottori, K. Tatsuta, H. Matsubara, H. Tanida, K. Nitta, T. Uruga, M. Aratono, J. Phys. Chem. B 2012, 116, 13739 - 13748. [19] T. Takiue, D. Vollhardt, Colloids Surf. A 2002, 198-200, 797- 804. [20] S. Barton, J. Chem. Phys. 1992, 96, 1343 -1351. [21] T. Shimoaka, Y. Itoh, T. Hasegawa, J. Phys. Chem. C 2012, 116, 17142 - 17148. [22] K. B. Blodgett, J. Am. Chem. Soc. 1934 , 56, 495. [23] J. Umemura, Handbook of Vibrational Spectroscopy, Vol. 2, Wiley, Chi- chester, 2002, pp. 982 - 998. [24] V. P. Tolstoy, I. V. Chernyshova, V. A. Skryshevsky, Handbook of Infrared Spectroscopy of Ultrathin Films, Wiley, Chichester, 2003. [25] N. J. Harrick, J. Opt. Soc. Am. 1965, 55, 851 - 857. [26] N. J. Harrick, Internal Reflection Spectroscopy, Harrick Scientific, New York, 1987. [27] P. Yeh, Optical Waves in Layered Media, Wiley-Interscience, Hoboken, 1988. [28] Y. Itoh, A. Kasuya, T. Hasegawa, J. Phys. Chem. A 2009, 113, 7810 - 7817. [29] J. S. Plaskett, P. N. Schatz, J. Chem. Phys. 1963, 38, 612 - 617. [30] M. J. Hannon, F. J. Boerio, J. L. Koenig, J. Chem. Phys. 1969, 50, 2829 - 2836. [31] C. S. Kim, R. C. Mowrey, J. E. Butler, J. N. Russell, Jr. , J. Phys. Chem. B 1998, 102, 9290 - 9296. [32] Y. Ishikawa, H. Kuwabara, T. Kunitake, J. Am. Chem. Soc. 1994, 116, 5579-5591. Received : May 28, 2014 Revised : June 20, 2014 Published online on July 30, 2014 Takeshi Hasegawa,*[a] Takafumi Shimoaka,[a] Nobutaka Shioya,[a] Kohei Morita,[b] Masashi Sonoyama,[b] Toshiyuki Takagi,[c] and Toshiyuki Kanamori[c] [a] Prof. T. Hasegawa, Dr. T. Shimoaka, N. Shioya Laboratory of Solution and Interface Chemistry Division of Environmental Chemistry, Institute for Chemical Research Kyoto University, Gokasho, Uji, Kyoto 611-0011 (Japan) Fax: (+ 81) 774-38-3074 E-mail : [email protected] [b] K. Morita, Prof. M. Sonoyama Division of Molecular Science, Faculty of Science and Technology Gunma University, Kiryu, Gunma 376-8515 (Japan) [c] Dr. T. Takagi, Dr. T. Kanamori National Institute of Advanced Industrial, Science and Technology (AIST) AIST Tsukuba Central 5, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565 (Japan) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cplu.201402156. © 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. (c) 2014 Blackwell Publishing Ltd. |