586
Langmuir 2005, 21, 586-594
Solvent/Gelator Interactions and Supramolecular Structure of Gel Fibers in Cyclic Bis-Urea/Primary Alcohol Organogels Yeonhwan Jeong,† Kenji Hanabusa,‡ Hiroyasu Masunaga,† Isamu Akiba,† Kentaro Miyoshi,† Shinichi Sakurai,§ and Kazuo Sakurai*,† Department of Chemical Processes & Environments, The University of Kitakyushu, 1-1, Hibikino, Wakamatsu-ku, Kitakyushu, Fukuoka, 808-0135 Japan, Department of Functional Polymer Science, Faculty of Textile Science and Technology and Graduate School of Science and Technology, Shinshu University, Ueda, Nagano, 386-8567 Japan, and Department of Polymer Science and Engineering, Kyoto Institute of Technology, Matsugasaki, Kyoto, 606-8585, Japan Received October 5, 2004 An organogel system consisting of trans-(1S,2S)-bis(ureidododecyl)cyclohexane (SS-BUC) and a series of primary alcohols was explored with optical polarizing microscopy (OPM), electron microscopy, circular dichroism (CD), wide-angle X-ray scattering (WAXS), and synchrotron small-angle X-ray scattering (SAXS). OPM, SAXS, and especially WAXS showed that the gel fiber of SS-BUC/methanol gels essentially consists of SS-BUC crystal itself. SAXS showed that the SS-BUC crystal in the gel takes a lamella with a domain spacing of 5.2 nm. When we left the gel at room temperature, the spacing decreased to 3.1 nm after several months. This distance change may correspond to the structural transition from a double-layer structure to an intercalated-layer structure, which was proposed by Feringa et al. (Chem.sEur. J. 1999, 5, 937-950) as a possible arrangement of the molecular packing. When the gels in ethanol, propanol, butanol, or octanol were examined, they never showed crystalline peaks in WAXS and SAXS, indicating the amorphous nature of the gels. With increasing the alkyl chain length from ethanol to octanol, dramatic changes were observed in the CD spectrum in the 200-500-nm range. Because these CD changes are correlated to the absorbance of urea, those can be considered as the evidence that the solvents strongly relate to the spatial arrangement between the adjacent urea groups. For the amorphous gels, the cross-sectional correlation function [γC(u)] was directly obtained by the inverse Hankel transform of the SAXS data. The value of γC(u) for the gels is decreased with increasing u (distance between the two scattering bodies, see eq 5). Furthermore, it more rapidly decreases than that of the rigid cylinder model. This feature can be explained by the speculation that many solvent molecules permeate into the SS-BUC fiber. There was a clear difference between ethanol and the other gels, indicating that the solvents with a longer alkyl chain give the more permeated and diffused fiber. This permeated fiber (i.e., wet fiber) can rationalize the dramatic CD change, by presuming that the permeated solvent molecules alter the molecular stacking form.
Introduction Organogels are the materials in which three-dimensional networks are formed through self-assembling of low-molecular-weight compounds (hereafter denoted by organogelators), and the network can absorb a large amount of solvent. Hence, their macroscopic property becomes viscoelastic.2 When temperature is increased, the viscoelastic material suddenly changes to liquid. This thermal transition can be explained by lowering the network density and finally becoming unable to support the gel phase. This transition temperature is called Tgel, and the relation between Tgel and the organogelator concentration is apparently described by the FerryEldridge equation.3,4 Well-known organogelators include, for instance, certain cholesterol and anthracene derivatives,5,6 surfactants,7 * Corresponding author. E-mail:
[email protected]. † The University of Kitakyushu. ‡ Shinshu University. § Kyoto Institute of Technology. (1) van Esch, J.; Schoonbeek, F.; de Loos, M.; Kooijman, H.; Spek, A. L.; Kellogg, R. M.; Feringa, B. L. Chem.sEur. J. 1999, 5, 937-950. (2) Terech, P.; Weiss, R. G. Chem. Rev. 1997, 97, 3133-3159. (3) Eldridge, J. E.; Ferry, J. D. J. Phys. Chem. 1958, 58, 992-995. (4) Kobayashi, T.; Takenaka, M.; Hashimoto, T. Kobunshi Ronbunshu 1998, 55, 613-627. (5) Lin, Y.-C.; Weiss, R. G. Macromolecules 1987, 20, 414-417. (6) Murata, K.; Aoki, M.; Nishi, T.; Ikeda, A.; Shinkai, S. Chem. Commun. 1991, 1715-1718.
porphyrins and phthalocyanines,8,9 carbohydrate,10 peptide derivatives,11 bis-urea compounds,1,12,13 phenylenevinylene derivatives,14,15 and oligoamides.16,17 A common feature for these molecules is that they can self-assemble through highly specific noncovalent interactions into long fibrous structures, which in turn form an entangled network in the liquid. The presence of strong selfcomplementary and unidirectional intermolecular inter(7) Oda, R.; Huc, I.; Candau, S. J. Angew. Chem., Int. Ed. 1998, 37, 2689-2691. (8) Terech, P.; Gebel, G.; Ramasseul, R. Langmuir 1996, 12, 43214423. (9) Engelkamp, H.; Middelbeek, S.; Nolte, R. J. M. Science 1999, 284, 785-787. (10) Yoza, K.; Amanokura, N.; Ono, Y.; Akao, T.; Shinmori, H.; Shinkai, S.; Reinhoudt, D. N. Chem.sEur. J. 1999, 5, 2722-2729. (11) Aggeli, A.; Bell, M.; Boden, N.; Keen, J. N.; Knowles, P. F.; McLeish, T. C. B.; Pitkeathly, M.; Radford, M. Nature 1997, 386, 359363. (12) Hanabusa, K.; Shimura, K.; Hirose, K.; Kimura, M.; Shirai, H. Chem. Lett. 1996, 885-886. (13) de Loos, M.; van Esch, J.; Stokroos, I.; Kellogg, R. M.; Feringa, B. L. J. Am. Chem. Soc. 1997, 119, 12675-12676. (14) Ajayaghosh, A.; George, S. J. J. Am. Chem. Soc. 2001, 123, 51485149. (15) Ajayaghosh, A.; George, S. J.; Praveen, V. K. Angew. Chem., Int. Ed. 2003, 42, 332-335. (16) Schmidt, R.; Schmutz, M.; Michel, M.; Decher, G.; Me´sini, P. J. Langmuir 2002, 18, 5668-5672. (17) Schmidt, R.; Schmutz, M.; Mathis, A.; Decher, G.; Rawiso, M.; Me´sini, P. J. Langmuir 2002, 18, 7167-7173.
10.1021/la047538t CCC: $30.25 © 2005 American Chemical Society Published on Web 12/13/2004
Gel Fibers in Bis-Urea/Primary Alcohol Organogels
Figure 1. Chemical structure of the organogelator, SS-BUC, used in this work.
Langmuir, Vol. 21, No. 2, 2005 587
gel fiber surface determines how the fibers self-assemble. Wang et al.23 observed the sol-to-gel conversion of a cholesterol derivative with atomic force microscopy (AFM) and showed that solvent molecules are permeated inside the bundles, either within the elemental fibers or in “channels” between them. Their work also indicates the manifest responsibility of the solvents to determine the supramolecular structure in the organogels. In this paper, we examine the physical properties of SS-BUC gels in five (six for SAXS) different alcohols by using electron microscopy, Fourier transform infrared (FTIR) spectroscopy, wide-angle X-ray scattering (WAXS), SAXS, and circular dichroism (CD) and discuss the solvent effect on the gels.
actions has, thus, been identified as a molecular requirement for gelation ability.2,18 The molecular interactions can be classified into three types: dipole-dipole, van der Waals, and hydrogen-bonding interactions. The hydrogenbonding interaction generally takes place when a hydrogen atom locates between two oxygen atoms and these three atoms align linearly. This strict unidirectional orientation leads the molecules to align in a specific manner, which is the major cause of highly organized supramolecular structures.12,19 Among others, cyclic bis-urea or -amino compounds have been studied extensively as a model compound for the hydrogen-bonding-type organogelators.1,12,20,21 Hanabusa et al.12,20 are the first to synthesize a series of bis-urea cyclohexane organogelators including trans(1S,2S)-bis(ureidododecyl)cyclohexane (the chemical structure is presented in Figure 1 and is denoted by SS-BUC). In succeeding years, Feringa’s group1,13,21 extensively studied the thermotropic and rheological properties of trans-(1R,2R)-bis(ureidododecyl)cyclohexane (RR-BUC), which is the enantiomer of SS-BUC and has the same physical properties as SS-BUC, except for the helicity of the gel fiber. They showed that the thermal stability of the RR-BUC gels increases in the order of 1-octanol < 1-hexanol < 1-butanol < 1-propanol, corresponding to increasing polarity of the solvent. This result seems to contradict the molecular model that hydrogen-bond formation between urea groups is the primary driving force for self-assembling, and, hence, the thermal stability should decrease with increasing the solvent polarity. They interpreted this solvent dependence according to the hypothesis that the increased network stability is ascribed to increased strength of the junction zones in more polar solvents and, hence, these junction zones are stabilized by solvophobic forces. Solvent-mediated intermolecular interactions seem to play a critical role in the formation of organogels. Sakurai et al.22 measured small-angle X-ray scattering (SAXS) from an organogel made of methyl 4,6-O-benzylidene-R-Dmannopyranoside in p-xylene and found that the elemental gel fibers are hexagonally packed with a spacing of 6.0 nm. This length is about 6 times larger than that of the gelator molecule and almost independent of both temperature and the gelator concentration. These features suggested that the solvophobic/solvophilic balance on the
Synthesis and Characterization of SS-BUC. SS-BUC was synthesized according to the established method.12 The 1H NMR spectra and specific rotation were measured on a JNM-ECP500 MHz NMR spectrometer (JEOL Co., Japan) and on a SEPA-300 polarimeter (HORIBA Co., Japan), respectively. 1H NMR (500 MHz, CDCl3, 60 °C) showed 4.93 (d, 2H), 4.38 (t, 2H), 3.44 (br, 4H), 3.09 (m, 4H), 2.04 (d, 2H), 1.71 (br, 2H), 1.45 (br, 4H), 1.25 (br, 40H), 0.88 (t, 6H), and [R]D was evaluated to be -0.700 (21.5 °C, 10 mg/mL). These values agreed with the reported ones.1 We used methanol (MeOH), ethanol (EtOH), propanol (PrOH), butanol (BuOH), heptanol (HeOH), and octanol (OcOH) as solvents, and all were of spectroscopic grade. Tgel values for 3 wt % MeOH, EtOH, PrOH, BuOH, and OcOH gels were determined with the test tube tilting method24 and coincide with the reported values for RR-BUC.21 Microscopic Observations, CD, and FT-IR. A Nikon ECLIPSE E6003 was used to carry out optical polarizing microscopy (OPM) for the 3 wt % MeOH and EtOH gels at room temperature. After a specimen was coated with Pt for 30 s, scanning electron microscopy (SEM) was carried out on a Hitachi FE-SEM S-5200 at 15 kV. CD measurements were carried out with a J-820 CD spectrometer. The 3 wt % gels were put into a glass cell with the thickness of 0.05 mm, and the spectrum was obtained at room temperature in the range of 200-500 nm at a scanning speed of 20 nm/min. We confirmed that linear dichroism was negligible in the CD of the present samples. Reflected FT-IR measurements with a Perkin-Elmer FT-IR spectrometer were executed in the range of 1560-1660 cm-1 at room temperature. WAXS and SAXS Measurements. We prepared 1.0, 2.0, and 3.0 wt % SS-BUC gels using MeOH, EtOH, PrOH, BuOH, HeOH, or OcOH. The samples were loaded into a quartz cell (MarkRo¨hrchen) with a 2-mm diameter, and then the cell was sealed with an epoxy adhesive. SAXS experiments were carried out at the BL40B2 and BL45XU stations at SPring-8 in Japan.25,26 At the BL40B2 station, the camera length and the X-ray wavelength were adjusted to 100 cm and 1.0 Å, respectively. The scattering intensity was accumulated for 300 s in the range of q ) 0.12-6.7 nm-1 with a Rigaku R-AXIS IV++ system (a 30 cm × 30 cm imaging plate), where q is the magnitude of the scattering vector defined by eq 2. The BL45XU station was equipped with an X-ray image intensifier with a cooled charge-coupled device camera, and the camera length and the X-ray wavelength were adjusted to 150 cm and 1.0 Å, respectively. WAXS experiments were performed on a Rigaku XRD-DSC-II. The data were accumulated with a slit of 0.15-mm width at scanning speed of 0.02°/min in the scattering angle range of 2θ ) 2-30° at room temperature. Owing to a low concentration of the organogelator, it was
(18) Luboradzki, R.; Gronwald, O.; Ikeda, A.; Shinkai, S.; Reinhoudt, D. N. Tetrahedron 2000, 56, 9595-9599. (19) Gronwald, O.; Shinkai, S. Chem.sEur. J. 2001, 7, 4328-4334. (20) Hanabusa, K.; Yamada, M.; Kimura, M.; Shirai, H. Angew. Chem., Int. Ed. Engl. 1996, 35, 1949-1951. (21) Brinksma, J.; Feringa, B. L.; Kellogg, R. M.; Vreeker, R.; van Esch, J. Langmuir 2000, 16, 9249-9255. (22) Sakurai, K.; Jeong, Y.; Koumoto, K.; Friggeri, A.; Gronwald, O.; Sakurai, S.; Okamoto, S.; Inoue, K.; Shinkai, S. Langmuir 2003, 19, 8211-8217.
(23) Wang, R.; Geiger, C.; Chen, L.; Swanson, B.; Whitten, D. G. J. Am. Chem. Soc. 2000, 122, 2399-2400. (24) Murata, K.; Aoki, M.; Suzuki, T.; Harada, T.; Kawabata, H.; Komori, T.; Oseto, F.; Ueda, K.; Shinkai, S. J. Am. Chem. Soc. 1994, 116, 6664-6676. (25) Fujisawa, T.; Inoko, Y.; Yoji, N. J. Synchrotron Radiat. 1999, 6, 1106-1114. (26) Fujisawa, T.; Inoue, K.; Oka, T.; Iwamoto, H.; Uruga, T.; Kumasaka, T.; Inoko, Y.; Yagi, N.; Yamamoto, M.; Ueki, T. J. Appl. Crystallogr. 2000, 33, 797-800.
Experimental Section
588
Langmuir, Vol. 21, No. 2, 2005
Jeong et al.
sometimes difficult to observe the crystalline peaks in organogels. To obtain the trace of the crystal diffraction, we had to carefully measure the samples with the narrow slit and slow scanning rate.
∫0∞γC(u) J0(qu)u du
(4)
∫-∞∞∆FC(xb) × ∆FC(xb + u) dxb
(5)
I(q) ∝ γC(u) )
1 q
Data Analysis of SAXS SAXS at low angles from cylindrical particles with infinite length is given by the following equations:2,8,27
I(q) ∝ q-1 exp(-q2rc2/2) q)
θ 4π sin λ 2
()
(1) (2)
I(q) ∝
where rc is the cross-sectional radius of gyration for the cylindrical scatterer and θ is the scattering angle. Equation 1 is an expansion for the low-angle limit of the general scattering function from rodlike scatterers given by eq 4, when rc is defined by the following equation, using the radial scattering length (or electron) density difference between the scatterer and the solvent, ∆FC(x):
2
rc
∫C∆FC(x) dC2(x) dx ) ∫C∆FC(x) dx
where γC(u) and J0(qu) are the cross-sectional correlation function and the zeroth-order Bessel function. The function of γC(u) is directly obtained by the inverse Hankel transform of the measured I(q). For a solid cylinder with the radius of R, eq 4 is analytically expressed by the firstorder Bessel function J1(Rq) as follows:
(3)
where dC(x) is the distance from the cross-sectional mass center to the scattering point x and the integral is carried out over the plane perpendicular to the cylindrical axis. According to eq 1, rc can be evaluated from the initial slope of the ln q × I(q) versus q2 plot (cross-sectional Guinier plot). The q region where eq 1 is valid depends on the magnitude of rc, and this region is called the crosssectional Guinier region. Generally, the Guinier regions are approximately equal to that of (rc × q)2 < 2 (see Figure S1 in Supporting Information); thus, for the case of rc ) 10 nm, q < 0.15 nm-1. A straight (nonflexible) and rigidbody cylinder with a radius of R is the simplest model for cylindrical scatterers and has been widely used to describe SAXS from organogels by Terech et al.2,28-32 In this model, ∆FC(x) is a constant within the cylinder and suddenly (stepwise) drops to 0 outside of the cylinder, and the relation between R and rc is given by rc ) R/x2. This model predicts that I(q) should decrease as ∼q-4 in the high-q region (Porod law). The Porod law holds not only for straight rigid-body cylinders but also for any scatterers with sharp interfaces that can be well-described by step functions (ideal two-phase model). However, in real materials, I(q) fulfills the prediction in a very few cases. The deviation from the Porod law is generally ascribed to the diffused interface between the scatterer and the solvents. The general expression of the scattering from a cylindrical particle with ∆FC(x) is given by eq 4 if the cylinders have a considerably larger dimension along z axis than along other two axes: (27) Roe, R.-J. Methods of X-ray and Neutron Scattering in Polymer Science; Oxford University Press: New York, 2000. (28) Terech, P.; Furman, I.; Weiss, R. G. J. Phys. Chem. 1995, 99, 9558-9566. (29) Terech, P.; Ostuni, E.; Weiss, R. G. J. Phys. Chem. 1996, 100, 3759-3766. (30) Terech, P.; Coutin, A.; Giroud-Godquin, A. M. J. Phys. Chem. B 1997, 101, 6810-6818. (31) Terech, P.; Allegraud, J. J.; Garner, C. M. Langmuir 1998, 14, 3991-3998. (32) Terech, P.; Coutin, A. Langmuir 1999, 15, 5513-5525.
J12(Rq) q3
(6)
Results Visual and Microscopic Observations. Figure 2 presents photographs for the five SS-BUC gels made of the five primary alcohols: MeOH, EtOH, PrOH, BuOH, and OcOH. The MeOH gel is white and opaque, indicating the presence of large aggregates with a size comparable with the wavelengths of visible light. With increasing the alkyl chain length, the gel becomes transparent. This means that the aggregate size presumably becomes smaller or solubility of the gelator becomes larger with increasing the alkyl chain length (i.e., the larger solubility means less gelator in the aggregate; thus, the gel becomes more transparent). When we observed the gels in a thinner vial than those in Figure 2, the gels were transparent, except for the MeOH gel. These results suggest that the MeOH gel is different from the others. Figure 3a,b compares the OPM images between the MeOH and the EtOH gels. Although the data are not presented, the microscopic images of the other alcohol (PrOH, BuOH, and OcOH) gels were similar to that of the EtOH gel. The MeOH gel contains many birefringent objects, and some of them exhibit a typical spherulite pattern. This feature indicates that the gel consists of crystals, and they grow from a crystal nucleus to the radial direction in an almost symmetrical manner. These features are in marked contrast to the image of the EtOH gel. The EtOH gel exhibits birefringence. However, there are no bright and spherulite patterns. When we observed the SEM image, there were clear differences between MeOH and the other gels. The typical examples are presented in panels c-e of Figure 3: MeOH, EtOH, and OcOH gels, respectively. All gels show a typical network usually observed for organogels. However, the EtOH and OcOH gels show a helical structure while MeOH gel does not show such a pattern at all. Helicity of the fiber in the SEM images of EtOH and OcOH gels is always right-handed, and the pitch in the OcOH gel is slightly larger than that of the EtOH gel (see Table 1). As shown in all the SEM images, the thicker fibers consist of thinner ones, and the smallest fiber is about 40-50 nm. Because the thickness of the platinum coating is about 20 nm, the net diameter of the thinnest fiber is 20-30 nm. This size is one order of magnitude larger than that of the SS-BUC molecule. Feringa et al.21 studied a similar system with transmission electron microscopy (TEM) after Pt shadowed and observed helical patterns for the 1-propanol, 1-butanol, 1-hexanol, and 1-octanol gels. They did not show TEM images for the mehanol gel. However, Hanabusa et al.12 carried out TEM for the MeOH gels after they were stained by osmic acid. They did not observe a helical pattern, although they observed bundled and entangled fibers.
Gel Fibers in Bis-Urea/Primary Alcohol Organogels
Langmuir, Vol. 21, No. 2, 2005 589
Figure 2. Comparison of the photographs of 3 wt % SS-BUC gels among five different alcohols. Each gel was prepared by heating the mixture of SS-BUC and alcohols, followed by standing at room temperature.
Figure 4. Comparison of WAXS profiles of the SS-BUC solid, the MeOH and EtOH gels, and the solvents.
Figure 3. Photographs of 3 wt % SS-BUC gels in MeOH and EtOH taken by OPM and SEM. For comparison, the SEM image for the OcOH gel is presented in panel e. Table 1. Comparison of Pitches for SS-BUC Gels in EtOH and OcOH
EtOH OcOH
no. of measurement
average of pitches (nm)
standard deviation
10 10
140 230
0.28 0.17
These previous studies indicate that the MeOH gel does not show a helical pattern and the others show the patterns, being consistent with our SEM results. WAXS: Crystalline versus Amorphous Gels. Figure 4 presents WAXS profiles from the SS-BUC solid, MeOH gel, and EtOH gel, compared with those of their solvents. The EtOH gel shows a profile almost identical with that of the solvent. The main peaks around 2θ ) 22° seem identical with each other; while the lower-angle peak around 10° in the gel is slightly larger than that in the solvent. Contrary to the EtOH gel, the MeOH gel has two
sharp peaks at 2θ ) 20.9 and 22.8° and one broad peak at 8.6° overlapping the broad solvent peak. The 20.9° peak appears at the same positions as that of the solid, while the solid has no peak corresponding to 8.6 and 22.8°. The solid has a sharp and large peak at 22°, which does not exist in the gels. The difference between the MeOH and the EtOH gels indicates that the MeOH gel contains a crystal and the EtOH gel does not. The other gels such as PrOH, BuOH, and OcOH showed the same results as that of EtOH. The present WAXS result indicates, by combining with the microscopic images, that the gel fibers in the MeOH gel consist of SS-BUC crystal and the other gel fibers can be considered amorphous. Although the MeOH gel shows a crystalline nature, the peak positions are not identical between the gel and the solid. This discrepancy suggests that the crystalline structure in the MeOH gel is different from that of the solid and can be ascribed to polymorphism of SS-BUC crystals. We carried out SAXS and WAXS from the SS-BUC gels in toluene, cyclohexane, dimethylsulfoxide (DMSO), and dimethylformamide (DMF), although the data were not shown in the text. Their results indicated that the SSBUC/toluene, DMSO, and DMF gels are crystalline but the cyclohexane gel is amorphous. The amorphous nature of the cyclohexane gel may be explained by the speculation that the molecular stacking of the gel fibers in cyclohexane is more diffused than that of the other solvents because cyclohexane can interact with the alkyl chains of gelators. Alkyl Chain Length Dependence of CD. Figure 5 compares the CD spectra among the EtOH, PrOH, BuOH, and OcOH gels. Because we could not obtain a reliable spectrum for the MeOH gels at any concentration owing to turbidity and birefringence, the data for the MeOH gel is not included. The origin of the CD spectrum for SSBUC should be exciton coupling between the adjacent urea bonds, and the CD spectrum can be related to the angle between the two induced dipole moments and the distance between them. Generally, the angle determines the shape of the spectra and the distance determines the intensity. The positive Cotton bands for the EtOH, PrOH, and BuOH gels gradually decrease with increasing the alkyl
590
Langmuir, Vol. 21, No. 2, 2005
Jeong et al.
Figure 5. CD spectra of 3 wt % SS-BUC gels in four primary alcohols, measured at room temperature with a 0.05-mm cell. The MeOH gel was too opaque to give a liable spectrum.
chain length. The negative band around 210 nm increases in magnitude going from EtOH to BuOH. These trends seem to systematically line up in the order of the chain length. On the other hand, the OcOH gel apparently shows a different feature from the others. The OcOH gel has no positive band, and the 210-nm negative band has a weaker intensity than that of the BuOH gel. When one dipole moment is involved in the CD spectrum, the resultant spectrum should be symmetrical about the point that the spectrum crosses the horizontal axis. However, the spectra in the figures are not symmetric at all, indicating that multiple dipole moments contribute to the spectrum. Therefore, the positive band around 300250 nm and the negative band at 210 nm are considered to have different origins or are due to the overlaying of multiple bands. Taking into account these features, the decreasing of the positive band from the EtOH to the BuOH gels and the disappearance in the OcOH gel can be considered as a continuous series of changes owing to the increment of the alkyl chain length. The negative Cotton band increases going from EtOH to BuOH, but it decreases for OcOH. The reason is not clear. However, it could be due to the solubility difference between the OcOH and the others. As shown in Figure 2, the OcOH gel is less opaque than the others. This feature is consistent with the lower Tgel of this system than those of the others. If less of the gelators are self-assembled, the CD intensity becomes lower. This may be an explanation of the lower CD for the OcOH gel. Important conclusions derived from the CD measurements are that there is some loss or change in chiral organization on increasing the alkyl chain length of the solvent. This implies that SS-BUC stacking is more loosened by the solvents with the longer alkyl chain, which is consistent with other measurements. IR Spectrum and Hydrogen-Bond Formation. Figure 6 compares IR spectra in the range of 1660-1560 cm-1, where amide I and II bands show characteristic peak shifts upon the hydrogen-bond formation.33-35 Generally, the amide I and amide II peaks are located in the (33) Kalnin, N. N.; Baikalov, A.; Venyaminov, S. Y. Biopolymers 1990, 30, 1273-1280.
Figure 6. Comparison of the FT-IR spectra of the SS-BUC solid and 3 wt % SS-BUC gels in various primary alcohols. The amide I band of the SS-BUC gel in MeOH is blue-shifted by about 2 cm-1 compared to gels in other alcohols.
1630-1650 and 1560-1590 cm-1 regions, respectively; the former shifts to the lower wavenumber (red-shift), and the latter shifts to the higher wavenumber (blueshift) upon transition from the sol (isolate) to gel (hydrogenbonding) states.1 On the other hand, in the solid state (crystalline state highly packed with hydrogen bonds), the amide I and II peaks appear at 1620-1630 and 15901600 cm-1, respectively. In the MeOH gel, the amide I and II bands are blueshifted by 7 and 5 cm-1, respectively, compared with those of SS-BUC solid. The blue-shifted amide I and II bands show that SS-BUC is less packed than the solid, although both are crystalline. This suggests that the SS-BUC crystal in the MeOH gel is somehow different from the solid crystal, which will be more clearly demonstrated in the time course of the SAXS diffraction in the MeOH gel (Figure 10). When the solvent is changed from EtOH to OcOH, the amide I band shifts toward the higher wavenumber, indicating that the hydrogen bond is loosening. The amide II band moved from 1592 to 1597 cm-1 when the MeOH gel was formed. However, it came back to 1592 cm-1 for the other solvents. The amide I band is mainly associated with the CdO stretching vibration (70-85%) and is directly related to the backbone conformation. Therefore, it can be directly related to the strength of the hydrogen bond. However, the amide II band results from the NsH bending vibration (40-60%) and from the CsN stretching vibration (18-40%). This band is conformation-sensitive. Hence, it may be difficult to interpret the changes on the basis of the gelation. SAXS. Figure 7 compares the SAXS profiles among the six gels. The MeOH gel has a sharp peak at q ) 1.34 nm-1, (34) Venyaminov, S. Y.; Kalnin, N. N. Biopolymers 1990, 30, 12431257. (35) Venyaminov, S. Y.; Kalnin, N. N. Biopolymers 1990, 30, 12591271.
Gel Fibers in Bis-Urea/Primary Alcohol Organogels
Figure 7. SAXS profiles of 3 wt % SS-BUC gels in six primary alcohols.
Figure 8. Cross-sectional radii of gyration rc of SS-BUC gels in EtOH, PrOH, BuOH, and OcOH. As alkyl chain lengths in alcohols are longer, cross-sectional radii of gyration are smaller.
corresponding to d ) 4.7 nm. Here, all samples were measured approximately a few hours after being cooled from the sol state. As shown in Figure 10, where SAXS was carried out in a wider q range than that of the present figure, the 1.34-nm-1 peak can be assigned to the first order of a lattice scattering from a lamella structure. On the other hand, the other gels had no lattice peak, and these amorphous gels show upward curvatures in the low-q range and broad and small second peaks around 0.8 nm-1, except for the EtOH gel. When we examined the gelator concentration dependence of the scattering profiles (see Figures S3 and S4 in Supporting Information) for the amorphous gels, there was no obvious concentration dependence in the scattering profiles, except for the OcOH
Langmuir, Vol. 21, No. 2, 2005 591
Figure 9. Comparison of experimental SAXS profile of the 3 wt % OcOH gel with the theoretical SAXS profile of R ) 10 nm (a) and cross-sectional correlation function γC(u) for rigid cylinder of gels in EtOH, BuOH, and OcOH (b). R is the radius of the solid cylinder. The cross-sectional correlation function γC(u) is the correlation between two ∆FC(x) values when the distance between these two is u. It is directly obtained by the inverse Hankel transform from the SAXS data.
gel. Therefore, the scattering profiles for different concentrations could be superimposed with each other, although the lower-concentration gels gave the lower S/N ratio. There is a possibility that we may observe an interfiber lattice scattering with increasing the gelator concentration. However, the concentration independence of the SAXS implies that this is not the case for the present system. Because SEM shows that the gels consist of long fibers with a radius of 10-20 nm, the upward curvatures and second peak in the amorphous gels can be ascribed to a particle scattering from the cylindrical objects. In the high-q region (q > 0.6 nm-1), where the Porod region appears, the intensities decay with the lower exponent than expected from the Porod law. This feature can be explained by a diffused interface between the gel fiber (scatterer) and the solvent. As depicted in the figure, the exponent is decreased with increasing the alkyl chain length of the alcohols, indicating that the longer alkyl chain provides the more diffused interface. Figure 8 shows the cross-sectional Guinier plot for the amorphous gels. From the initial slopes, the radius of gyration for the cylindrical fiber (rc) is evaluated for each system, and the resultant values are indicated in the figure. The value of rc is decreased with increasing the alkyl chain length. For all rc evaluations, the data points are deviated upward from the initial slopes in the range of q2 > 0.04 nm-2. This is because the cross-sectional Guinier expression, eq 1, does not hold in the higher-q region and normally dispersities in the radius give the upward curvatures. When the distribution in rc is small enough (generally the standard deviation is less than 0.15), we can observe
592
Langmuir, Vol. 21, No. 2, 2005
Jeong et al.
Figure 10. Time course of the long spacing from SAXS measurement for the 3 wt % MeOH gels (left side) and proposed model from SAXS data (right side). After several months, the lamella spacing varied from 5.2 to 3.1 nm. This dwindling spacing may correspond to the structural transition from a double layer to an intercalated layer.
the second oscillation due to the particle scattering. The upper panel in Figure 9 compares the experimental data of the 3 wt % OcOH gel with a theoretical calculation of R ) 10.0 nm using eq 6. The theoretical curve can describe the downward curvature of the experimental data in the q < 0.25 nm-1 range. However, it fails to describe the second peak position, its peak width, and furthermore, the q dependence of I in the high-q region. The reason of the failure is clear; the rigid cylinder model is not an appropriate model for the gel fiber. It should be mentioned that a correction for the thermal density fluctuation was made empirically according to an established method,27 and any ambiguity for the conclusion mentioned here is not caused from this correction (see Figure S6 in Supporting Information). We calculated the cross-sectional correlation function: γC(u) using a numerical inverse Hankel transform, and the results are plotted in the lower panel (Figure 9) comparing with that of a rigid cylinder. Here, the horizontal axis is normalized with the radius (or radius of gyration) of the scatterer for convenience to compare the correlation functions with different radii. According to the definition in eq 5, γC(u) is the correlation between
two ∆FC(x) values when the distance between these two is u. Therefore, as the inset illustration shows, γC(u) is the summation of the product of the electron density within the overlapping region. With increasing u, γC(u) is decreasing because the overlapping area is decreasing. γC(u) for rigid cylinders is analytically given, and it decays linearly in the u/R < 0.5 region. When we compare the experimental data with those of the rigid cylinders, the data points more radically decay than those of the theory, indicating that the real gel fibers have a considerably diffused interface or that many solvent molecules permeate into the SS-BUC fiber. Furthermore, there was a clear difference between the EtOH and the other gels. The difference indicates that the longer alkyl chain gives the more permeated and diffused fiber. Time Course of the Long Spacing in the MeOH Gels. Owing to polymorphism, the solid-state SS-BUC did not give a unique lamellar spacing. In our experiment, SAXS from the solid had multiple peaks in the 1-2-nm-1 range and the peak distribution depended on the solvents used in the purification process and the other conditions. In this sense, the sharp single peak in SAXS from the gel state surprised us. We measured the time course of this
Gel Fibers in Bis-Urea/Primary Alcohol Organogels
spacing with a securely sealed cell with epoxy. The result is presented in Figure 10. The spacing decreased from 5.2 to 3.1 nm after several months, keeping the lamellar structure. This dwindling spacing may correspond to the structural transition from a double layer to an intercalated layer, as illustrated in the figure (for the definition of the double and intercalated layers, see Figure S7 in Supporting Information). Both structures are proposed by Feringa et al.1 as a possible arrangement of the molecular packing. Discussion Transition from Dry to Wet Gels. Terech et al.2,8,28-32 extensively studied SAXS from the organogels and showed that most of the scattering can be represented by particle scattering from randomly oriented cylinders with the infinite length. They showed that some of their organogelators have crystalline peaks in WAXS. George and Weiss36 showed that clear lattice scattering can exist in the gel state. For instance, their ammonium carbamate compounds can gelatinize silicon oil and provide a clear lamellar diffraction pattern with the domain spacing evaluated to be 45 Å. Their WAXS showed that the crystalline peaks are also present in the gel state. All those organogels are composed of the crystalline fibers; therefore, the solvent molecules are excluded from the gel fiberssin other words, they are classified as a “dry gel”. Kobayashi et al.4 studied an organogel system made from 1,3:2,4-cis-O-inside-bis-O-(p-methylbenzylidene)-D-sorbitol and showed that the gel fiber is essentially made of a needle crystal only containing the gelator. They explained the formation of long crystalline fibers by one-directional growth of the crystal. In fact, Shinkai et al.18 pointed out strong coloration between gelation ability and the molecular nature that allows forming a one-dimensional hydrogen-bond array in the crystal structure. Whitten et al.37 synthesized a gelator containing a cholesterol tethered to a trans-stilbene and observed morphological changes upon the sol-gel transition (induced by temperature drop) for this gelator with AFM. They observed a cross section of the fibril bundle and found that the fiber bundles consist of tens of unit fibers and the cross section suggests a fluidized surface. Their observation indicated that solvent molecules are involved inside the bundles, and the solvent volume fraction was evaluated to be about 30%. Their results evidence that the solvent molecules are incorporated into the gel fiber. As far as we know, this is the first evidence for the “wet-gel”. More recently, Sakurai et al.22 studied a gel made from methyl 4,6-O-benzylidene-R-D-mannopyranoside and p-xylene with SAXS, and the gel can be classified as a “wet gel”. When a gel fiber is dry and crystalline, it is essentially classified into a needle crystal and the gelation should be ascribed to that the fine crystal needles are entangled to induce a capillary effect to absorb solvents. From the standpoint of supramolecular chemistry, the crystal and dry gel fibers may not be interesting. It seems that many previous works did not clarify whether their gel was classified into crystal or amorphous gels. The present study shows that MeOH induces a crystal and dry gel and the other solvents produce wet and amorphous gels. It is interesting that the transition from the dry to wet gels is induced by such a small difference in the solvent. Although the data is not presented in the text, we tried to observe if there was a threshold composition in the MeOH/EtOH (36) George, M.; Weiss, R. G. J. Am. Chem. Soc. 2001, 123, 1039310394. (37) Geiger, C.; Stanescu, M.; Chen, L.; Whitten, D. G. Langmuir 1999, 15, 2241-2245.
Langmuir, Vol. 21, No. 2, 2005 593
mixture for this wet/dry gel transition. The crystalline lamellar peak decreased gradually with increasing the EtOH ratio, and there seemed to be a structure transition between 30 and 25 vol % MeOH. However, the positions of the lamellar peaks were not changed (see Figure S5 in Supporting Information). Helical Fiber and Amorphous Nature. One of the peculiar features of the organogels is that the gel fiber forms periodical helixes, presumably, reflecting on the molecular chirality. Figure 3 shows that the helix pattern is observed in EtOH and OcOH. This feature suggests that the helical pattern is characteristic for the amorphous gels. This is an interesting contrast to that a helical structure (row structure)38 is produced by twisting a crystal lamella along the stacking direction in crystalline polymers. The radius of the fiber is estimated to be 10 nm from the SAXS measurements, and this value almost agrees with the finest fiber that can be observed with SEM. Therefore, we can consider that the fiber with rc ) 10 nm is the basic constituent for the network. It is reasonably considered that the double or intercalated layer (see Figure 10) is slightly twisted along the axial direction (CD supports this speculation), and this is the origin of the helix. If this is the case, the radius of the helix may be around 2-3 nm, which does not agree with the experimental results. One may think the observed helix may be a supercoiling architecture of the twisted molecules; hence, the resultant radius can be larger than that of the twisted molecular layers. Generally, the helicity of the supercoil is the reverse of the original coil helicity and there is no upper limit of the buildup of the supercoil. In this case, microscopy should show a mixture of the right- and lefthanded helixes, because when a right-handed helix makes a supercoil, the resultant coil should be a left-handed one. Obviously this is not the case in this system. As Hanabusa et al.20 and Feringa et al.1,13 pointed out, there is a missing ring to connect the molecular chirality detected with CD and the helix observed with microscopy. Molecular Model to Explain the CD Change and SAXS. The u dependence of γC(u) indicates that many solvent molecules permeate into the SS-BUC fiber. Because the urea groups form hydrogen bonds to tightly bind the molecules, when the solvents enter between the dodecyl chains of SS-BUC, the chains have to change the conformation to accommodate the solvents. It is reasonable to suppose that, when the more solvents invade the gel fiber, the SS-BUC conformation is more altered. This conformational change should affect the spatial arrangement of the urea bonds, causing the CD changes. The distance change between the adjacent urea groups is too small to give an appreciable change in IR. However, CD is considerably sensitive to the dipole-dipole distance as well as the angle between them. This model is illustrated in Figure 11. The driving force of the solvent permeation into the fiber can be affinity between alkyl chains. Because SSBUC has dodecyl chains and these long alkyl chains are supposed to face outside, the longer alkyl alcohols should be more favorable to interact with the dodecyl chains. This is the reason that the longer alcohol gives the more diffused interface; hence, the thermal stability is more decreased. Solvent Effect in the MeOH Dry Gel (Dwelling of the Lamellar Spacing of the MeOH Gel). According to the computational chemistry of SS-BUC,1 the SS-BUC (38) Schultz, J. M. Polymer Crystallization; American Chemical Society: Washington, DC, 2001.
594
Langmuir, Vol. 21, No. 2, 2005
Jeong et al.
BUC molecules stack with each other in the screwed form. However, the gel fiber does not take the intercalated form yet. This is probably because the presence of MeOH between the alkyl chains prevents the alkyl chain from taking the intercalated form.
Figure 11. Proposed models for the structure of the gel fibers in alcohols with a shorter alkyl chain and longer alkyl chain, respectively. The solvent with the longer alkyl chain interacts with the alkyl chain of the gelator and disturbs the molecular packing of the gel fiber.
gel can adopt two possible arrangements for the gel fiber, translational and screw axis aggregate (see Figure S7 in Supporting Information). In the screw axis aggregates, SS-BUC molecules stack alternatively along the hydrogen bond between the urea groups. This screwed form enables the gel fiber to take the intercalated form because there is enough room to accommodate other alkyl chains between the adjacent molecules. SAXS showed that the spacing of 3.1 nm was attained after several months, and this spacing can be reasonably explained with the intercalated/screwed model. The initial value of 5.2 nm in Figure 10 happens to be coincident with that of the double-layer model. To change from the intercalated to the double-layer model, the urea hydrogen-bond form has to alter from the translational to the screwed form. We are not sure that such transitions can occur without destroying the gel fiber. Another possibility is as follows. In the initial state, SS-
Concluding Remarks SS-BUC gels in various primary alcohols were investigated with OPM, electron microscopy, CD, FT-IR, and X-ray scattering methods. These gels showed different morphologies according to the polarities of the alcohols used. The SS-BUC gel in MeOH had more birefringence than that in EtOH from the OPM observation. The SSBUC gels in EtOH and OcOH consisted of right-handed aggregates, but the gel in MeOH did not. WAXS showed that the SS-BUC gel in MeOH has crystalline property but the gel in EtOH has an amorphous nature. CD spectroscopy confirmed that gels in primary alcohols comprised of different molecular aggregation types between urea groups. FT-IR spectroscopy showed that amide I band of the gels in EtOH to OcOH was blue-shifted to 2 cm-1 from that of the gel in MeOH, indicating the interference of hydrogen bonding between urea groups by solvents. From the SAXS measurement of SS-BUC gels in primary alcohols, the gel in MeOH formed a lamella structure and the other gels formed an amorphous structure in the range of q ) 0.12-6.7 nm-1. The deviation from the Porod law in a large q range showed that gels in solvents with a longer alkyl chain give a more diffused interface than those in solvents with a shorter alkyl chain. The cross-sectional radii of gyration rc of SS-BUC gels in alcohols were estimated to be 10.2-7.4 nm from EtOH to OcOH. The cross-sectional correlation function [γC(u)] of the amorphous gels indicated that solvent molecules permeate into the SS-BUC gelator. After several months, the lamella spacing for the SS-BUC/MeOH gel changed from 5.2 to 3.1 nm, indicating that the lamella structure of the gel may have two possible arrangements, that is to say, double layer or intercalated layer. Acknowledgment. This work is performed under the approval of SPring-8 Advisory Committee (2003A0314NL2-np, 2003B0792-NL2b-np). SEM and FT-IR were performed at the instrumentation center of The University of Kitakyushu. Supporting Information Available: The Guinier region (Figure S1), temperature dependence of SAXS profiles for the 3 wt % SS-BUC gel in MeOH (Figure S2). Concentration dependence of the SS-BUC gels’ SAXS profiles in PrOH (Figure S3), concentration dependence of the SS-BUC gels’ SAXS profiles in PrOH, BuOH, and OcOH at 3 and 1 wt % (Figure S4), solvent composition dependence of the SAXS profiles for SS-BUC gels in MeOH and EtOH (Figure S5), comparison of the experimental SAXS profile and thermal-density-fluctuation-corrected SAXS profile for 3 wt % SS-BUC OcOH gel (Figure S6), and schematic presentations for the double layer and translational axis (a) and intercalated layer structure and screw axis (b) for the SS-BUC MeOH gel fiber (Figure S7). This material is available free of charge via the Internet at http://pubs.acs.org. LA047538T
