3322
J. Phys. Chem. B 2008, 112, 3322-3327
DNMR Study of Hydrophilic and Hydrophobic Silica Dispersions in EBBA Liquid Crystals Jonathan Milette,† C. T. Yim,‡ and Linda Reven*,† Centre for Self-Assembled Chemical Structures (CSACS), Department of Chemistry, McGill UniVersity, 801 Sherbrooke Street West, Montreal, Que´ bec, Canada H3A 2K6, and Department of Chemistry, Dawson College, 3040 Sherbrooke Street West, Westmount, Que´ bec, Canada H3Z 1A4 ReceiVed: September 24, 2007; In Final Form: December 12, 2007
Dispersions of hydrophilic (A300) and hydrophobic (R812) silica aerosils in a Schiff-base-type liquid crystal (LC), p-ethoxy(benzylidene)-p-n-butylaniline (2O.4), EBBA, were characterized by deuterium nuclear magnetic resonance (DNMR). The formation and stability of random (RAN) versus anisotropic (AAN) aerosil networks under zero- versus in-field cooling was studied as a function of aerosil density and compared to previous studies of n-alkylcyanobiphenyl (nCB) dispersions. Whereas the LC directors of the hydrophobic R812 dispersions are almost completely annealed after in-field cooling, the hydrophilic A300 silica in EBBA gives rise to a mixture of RAN and AAN. The more complete R812 AAN partially breaks under in-field sample rotation, but the partial AAN formed by the A300 silica is stable. Weakening the aerosil network to compensate for weaker LC surface anchoring results in a complete network, but a strong LC/silica surface interaction must be combined with hydrophilic aerosils to produce AANs which are both complete and stable.
1. Introduction Liquid crystals (LCs) containing silica (aerosil) suspensions differ from LCs confined in polymer networks, aerogels or other porous solids in that the solid network can rearrange. Upon application of an external field, these materials become transparent due to LC reorientation and the transformation of defects. The persistence of this state after the field is turned off is the so-called memory effect, first observed for nematic systems1 and later proposed for applications such as laser addressing.2 Silica aerosils consist of irregularly branched strings of spherical, nanoparticles of amorphous silica linked together by -Si-OSi- bonds.3 At a high enough density, these aggregates can further agglomerate via surface H-bonds through a diffusionlimited process to create an aerogel-like network of void spaces, known as a random aerosil network (RAN). One current interpretation of the behavior of aerosil-filled LCs is that the specific interactions between the embedded LC and the silica network are believed to lead to the memory effect. The surface hydroxyl groups of the aerosil particles and the polar headgroups of LC molecules such as n-alkylcyanobiphenyl (nCB) can interact to form a homeotropic alignment of the LC at the silica surface and an elastic strain on the LC directors.4 Upon application of a field and within a specific range of aerosil densities, the LC alignment results in the reorientation of the aerosil network due to the interaction between the matrix and the agglomerates. The relatively weak H-bonding between the silica particles allows the network to rearrange and respond elastically to the LC reorientation.5 The random aerosil network, characterized by a mean void size, l0,6 is proposed to transform into an anisotropic aerosil network (AAN), stabilized by the LC elastic forces. The aligned silica surfaces of the AAN in turn stabilize the LC alignment, locking it in the voids and leading to the memory effect. At higher silica densities, the void * To whom correspondence should be
[email protected]. † Department of Chemistry, McGill University. ‡ Department of Chemistry, Dawson College.
addressed.
E-mail:
sizes become too small to allow the reorientation of the LCs by the field, silica disorder dominates and the memory effect is no longer observed.7 An alternative source of the memory effect does not require the reorganization of the aerosil network. Instead, Bellini and others have found both experimentally and theoretically that the introduction of a dilute “quenched” random disorder where the hosting structure cannot reorganize is sufficient to give rise to memory effects.8,9 The “memory parameter” was first calculated for an aerosil dispersion in 5CB using light transmission spectroscopy.10,11 The maximum parameter is found at silica concentrations of ∼2.5 wt % for hydrophilic silica and much higher concentrations of ∼10 wt % for hydrophobic silica. A minimum parameter is measured for nonpolar LCs unable to form hydrogen bonds with the silica surface. Intensity shifts of the aerosil IR bands (OH and Si-O-Si groups) at the maximum memory parameter have been used to support the role of H-bonding and the proposed rearrangement of the network.11,12 Evidence for the anchoring of nCB to the aerosil surface via H-bonds between the cyano head groups and the silica surface OH groups comes from IR band shape analysis. The stability of the memory effect of nCB/ silica dispersions was found to correlate to the surface silanol density7 and higher silica densities, Fs, were required to form a network in the case of hydrophobic silica dispersions in nCB. Despite the important role the molecular surface anchoring strength may play, relatively little work has been done to tune these interactions through variation of the LC and/or filler surface chemistry. Characterization at the molecular level, which is required to understand these interactions, has also been lacking. Wide-line 2H NMR is a powerful tool to directly probe LC alignment and to detect rearrangement of the solid network in the presence of an applied magnetic field. Although 2H NMR has been extensively used to study other confined LCs, only studies of aerosil dispersions using the typically studied nCB and hydrophilic silica have been reported.4 The 2H NMR line shapes reflect a competition between a surface-induced order and random disorder. At low concentrations, the silica acts as a random impurity and the 2H NMR
10.1021/jp077682y CCC: $40.75 © 2008 American Chemical Society Published on Web 02/26/2008
Silica Dispersions in EBBA Liquid Crystals spectrum is essentially the same as the pure LC, a sharp, full splitting doublet.13 At silica densities above the percolation threshold in the soft gel regime, a RAN forms. If the average void size is less than the magnetic coherence length, the surface/ LC interactions dominate and the 2H line shape reflects a distribution of director orientations, leading to a broad Pake pattern. When the sample is heated to the isotropic state and then cooled in field, an AAN can form and the Pake pattern is replaced by the full spitting doublet, showing that the LC directors are again aligned with the field.4 In this study, we have used wide-line 2H NMR to characterize a Schiff-base-type LC, p-ethoxy(benzylidene)-p-n-butylaniline (2O.4), EBBA, which is expected to interact differently with the surface, since it has a polar group in the center and a moderate dipole moment (2.1 D), as compared to nCB LCs (∼6 D) which have the polar group at an exposed headgroup position. The effect of silica dispersions on phase transitions of related alkoxybenzylidene alkylanilines have been reported, but no studies of the memory effect have been done.14,15 Calorimetric and dielectric measurements of weakly polar LCs which also have a polar group in the center displayed different behavior from nCB which was attributed to the different LC/surface.16 Our 2H NMR study of silica-filled EBBA reveals the formation of an anisotropic network but with only partial alignments in both hydrophilic and hydrophobic fumed silica dispersions. 2. Experimental Section Materials. Hydrophilic silica A300 and hydrophobic silica R812 were obtained from Degussa.17 Silica A300 has a specific surface area of a ) 300 ( 30 m2/g and a surface hydroxyl (OH) density of 5 groups/nm2. Silica R812 (a ) 260 ( 30 m2/g) has a fraction of its surface OH groups methylated, to give a surface OH density of 0.44 groups/nm2. N-(p-ethoxybenzylidene)-p-nbutylaniline, EBBA, was obtained from Sigma-Aldrich, recrystallized in ethanol three times and dried under high vacuum (∼0.01 mmHg) for 24 h. The EBBA-d2 was synthesized by first refluxing p-n-butylaniline hydrochloride in D2O for several hours.18 The resulting p-n-butylaniline-d4 was extracted with ether, purified and then used to prepare EBBA-d2 by the reaction with 4-ethoxybenzaldehyde. The product was further purified by recrystallization several times in ethanol. Sample Preparation. The EBBA/aerosil dispersions were prepared according to the solvent method.6 The aerosil was dried at 150 °C for 24 h, and stock solutions using ethanol as the solvent were sonicated with an ultrasonic probe (Cole-Parmer 25 kHz Ultrasonic Processor, model 4180) for 20 min at room temperature. The proper volume of the aerosil stock solution was mixed with EBBA-d2 in crystalline (Cr) phase and put under low vacuum for 36 h at 70 °C with gradually decreasing pressure down to 154 mmHg while stirring. To ensure complete evaporation of the solvent, the sample was transferred into the NMR tube and to a high vacuum system of 0.1 mmHg for 6 h at 100 °C. The final step in the sample preparation involved the zero-field cooling process; i.e., the sample is cooled from the isotropic phase to the nematic phase at ∼45 °C outside the magnetic field. Between experiments, samples were kept at ∼45 °C in an incubator. DNMR. Deuterium nuclear magnetic resonance (DNMR) spectra were acquired with a Chemagnetics CMX-300 spectrometer operating at 45.99 MHz. The solid echo pulse sequence with 4 µs (∼π/2) pulse, an acquisition recycle time of 0.2 s and a spectral width of 50 kHz was used with broadband proton decoupling. The in-field cooling was done by preheating the sample in the magnetic field from 38 °C, the nematic (N) phase,
J. Phys. Chem. B, Vol. 112, No. 11, 2008 3323 TABLE 1: Void Size of A300 and R812 Dispersions in LCs mean void size, l0 silica density, Fs (g/cm3)
A300 (nm)
R812 (nm)
0.010 0.025 0.050 0.075 0.100
667 267 133 89 67
769 308 154 103 77
to 78 °C, the isotropic (I) phase, and then cooling it down at an average rate of 0.2 °C/min across the TN-I phase transition and 1.3 °C/min down to 38 °C. 3. Results and Discussion EBBA is a thermotropic LC with a nematic mesophase. This liquid crystal was chosen in view of its likely weaker interaction with the silica surface as compared to the cyanobiphenyl-type liquid crystals studied by other groups. The reported phase transition temperatures of EBBA are TCr-N ) 36 °C and TN-I ) 77 °C.19 The sample preparation procedure is crucial for creating homogeneous samples, which is not always achievable for certain LCs and silica densities. The methods that have been used for preparation of aerosil dispersions include (a) simple mechanical mixing, (b) homogenizing with an ultrasonic probe during evaporation and (c) the solvent method. The solvent method was adopted because it has been shown to generate more uniform dispersions, particularly for samples of high silica density. The speed at which the solvent is evaporated under low vacuum and constant stirring influences the homogeneity of the dispersion. As reported by other groups,4,6 heterogeneous samples visually displayed a phase separation and/or change in color from light to dark beige. Such samples also showed an additional splitting in the DNMR spectra. All of the samples were prepared several times to check for reproducibility. Dispersions ranging from 0 to 0.900 g of SiO2/cm3 of LCs (Fs) were prepared. At low silica densities, Fs e 0.01, the “no gel” regime, the nanoparticles behave as impurities in the LC matrix and no silica network is expected.4 At higher silica concentrations where Fs g 0.01 in the “soft gel” regime, the silica agglomerates interact via surface hydrogen bonds to form a random network such that the host LC molecules are separated into many random domains (voids). With increasing silica density, the mean void size becomes smaller and the network more rigid until it reaches the “stiff gel” regime, Fs g 0.1, where the disorder induced by the silica dominates and a complete network is formed, locking in an isotropic distribution of the LC directors. In the soft gel regime, the average void size of the network can be approximated as l0 ) 2/aFs.6 The estimated void sizes of the soft gel samples produced for this study are listed in Table 1. In confined LC systems, the director orientations of the LCs in the aerosil cavities are influenced by the external magnetic field only when the magnetic coherence length, ξM ) (K/B02∆χ)1/2, is smaller than the mean void size, l0.13 This length is defined by the average Frank elastic constant, K, the magnetic susceptibility anisotropy, ∆χ, and the magnetic field, B0. In the case of the DNMR studies of aerosil-filled EBBA reported here, ξM ∼ 1 µm. 3.1. “No Gel” Regime: Zero-Field Cooling (ZFC). For aerosil densities below Fs ) 0.01 in the “no gel” regime, no network is formed and the induced disorder is small. The DNMR spectrum at 38 °C (N phase) after zero-field cooling (ZFC) is a full splitting doublet, similar to the pure LC (Figure 1a). The LC directors are fully oriented by the applied field,
3324 J. Phys. Chem. B, Vol. 112, No. 11, 2008
Milette et al.
Figure 1. (a) DNMR spectrum of EBBA and EBBA/aerosil (b) R812 and (c) A300 dispersions, Fs ) 0.005, at 38 °C.
Figure 2. DNMR spectra of EBBA/aerosil R812 at 38 °C as a function of silica density (a) before and (b) after in-field cooling.
and the dispersion has no effect on the spectral pattern (Figure 1b and c). For dispersions with concentrations in the higher range of the no gel regime, a partial network with a large average void size may form but results in no change in spectral pattern relative to the pure LC. The silica links may be too weak to create elastic strains on the LCs and no thixotropic gel forms. Only a small disorder is introduced, and the system retains bulk LC-like behavior. Furthermore, the magnetic field of the NMR can overcome the elastic strains induced by the weak network to produce an alignment similar to the pure LC. These DNMR results are similar to those observed for no gel silica dispersions of nCB in the nematic phase.4 3.2. “No Gel” Regime: In-Field Cooling. An in-field temperature cycle study was performed in the no gel regime by heating and cooling the sample between 38 and 78 °C under the influence of the NMR magnetic field. As the temperature increases, the LC molecules lose their alignment, resulting in a smaller doublet energy splitting. Once the temperature reaches the TN-I transition, the LC has lost all orientational order and a singlet is observed. As the temperature is subsequently lowered back to the nematic phase, the doublet reappears and the size of the splitting gradually increases and reaches its original value once the cycle is finished. No hysteresis is observed. At the N-I transition, there is a plateau where the doublet and singlet coexist. Jin et al. interpret this plateau as the signature of the magnetic field annealing effect on the elastic strains.4 Once the cycle was completed, an angle dependence study was performed by rotating the sample in the magnetic field. As expected, the spectrum did not display any angle dependence, since the LC directors can freely realign themselves. The nematic phase of 8CB in the “no gel” regime also showed no angle dependence after the in-field cooling; however, a strong angle dependence in the higher silica density range of the same regime was observed for the smectic phase of 8CB. This
difference was attributed to the higher viscosity of the smectic phase as compared to the nematic phase.4 3.3. “Soft Gel” Regime: Zero-Field Cooling. In the soft gel range, 0.01 < Fs < 0.1, a partial to complete silica network is established. The system is more complex, since (1) the LC orientational order is altered by a surface-induced order that decays with distance and (2) elastic strains are created by the competition between these surface interactions and distortion of the local directors.4 After zero-field cooling to 38 °C in the nematic phase, the DNMR spectra shown in Figures 2a and 3a for the hydrophobic (R812) and hydrophilic (A300) dispersions consist of a superposition of a Pake pattern and a full splitting doublet. We propose that these spectra, which are reproducible, arise from a distribution of void sizes or the coexistence of a RAN and an AAN rather than sample heterogeneity. Samples which were visually heterogeneous displayed additional splittings of the doublet and Pake pattern. In the case of a distribution of void sizes, the Pake pattern arises from the isotropic distribution of the local directors of the EBBA that is embedded in cavities smaller than the magnetic coherence length, whereas the full splitting doublet arises from LC molecules in cavities larger than the magnetic coherence length that can be aligned by the applied field. With increasing silica density, the average void size decreases, resulting in a dominant Pake pattern contribution. An alternative interpretation, given by Jin et al., who observe very similar superpositions of a doublet and Pake pattern in the nematic phase of ZFC 8CB/hydrophilic silica dispersions, is that the silica configuration consists of a mixture of a RAN (Pake pattern) and an AAN (doublet).4 In their case, the room temperature ZFC spectra (smectic phase) consist of a Pake pattern which transforms to the superposition of the Pake pattern and doublet as the sample is heated to the less viscous nematic phase. They attributed the appearance of the doublet to the breaking of silica links and rearrangement of the silica agglomerates into an AAN through elastic coupling with the magnetic field. With increasing silica density, smaller void sizes
Silica Dispersions in EBBA Liquid Crystals
J. Phys. Chem. B, Vol. 112, No. 11, 2008 3325
Figure 3. DNMR spectra of EBBA/aerosil A300 at 38 °C as a function of silica density (a) before and (b) after in-field cooling.
Figure 4. DNMR of Fs ) 0.075 EBBA/aerosil (a) R812 and (b) A300 ZFC dispersions at 38 °C.
and a stronger network, there is less reorientation of the local directors or breaking of the silica strands by an applied magnetic field. The directors and network remain randomly oriented, leading to a dominant Pake pattern contribution. Additional insight is gained by comparing the ZFC spectra of hydrophilic (A300) and hydrophobic (R812) silica dispersions (Figure 4). At Fs ) 0.075, the hydrophilic dispersion induces more disorder than the hydrophobic one, as evidenced by a larger Pake pattern contribution. The higher surface density of OH groups provides more H-bonding linkages and a stronger and more complete network. The isotropic distribution of local directors and hence the Pake pattern contribution is larger in the case of the hydrophilic dispersions which is in agreement with previous studies.20 3.4. “Soft Gel” Regime: In-Field Cooling (IFC). The thermal behavior of the soft gels is similar to the no gel samples in that there is no hysteresis of the size of the energy splitting for the doublet or Pake pattern contributions during the in-field heating and cooling cycle (Figure 5). On the other hand, the relative contributions of the two components change. Comparison of the ZFC and IFC spectra in Figures 2 and 3 reveals that the Pake pattern contribution is smaller after the in-field cooling. As supported by the angle dependence measurements described below, the RAN that was not annealed under the influence of an applied field alone becomes partially transformed to an AAN.
Figure 5. Temperature cycle of a soft gel dispersion: Fs ) 0.050 EBBA/aerosil A300.
If the annealed sample is removed and then returned to the magnetic field of the NMR spectrometer at a much later time, the spectrum remains identical. The persistence of this alignment of the network is a manifestation of the memory effect as detected by NMR spectroscopy which differs from the optical measurements in that the sample is always subjected to an external field. We also note that there have been studies of the influence of the heating rate on the phase transitions of aerosil/ LC dispersions; however, this aspect was not examined in this paper.21 Whereas the LC directors in the hydrophobic silica dispersions up to Fs ) 0.1 are almost completely annealed after the temperature cycle, a strong Pake contribution remains for the hydrophilic dispersions. Soft gel dispersions of 8CB/A300 aerosils with silica densities up to Fs ) 0.1 were completely annealed by IFC, unlike the partial annealing for EBBA/A300
3326 J. Phys. Chem. B, Vol. 112, No. 11, 2008
Milette et al.
Figure 6. (a) Stiff gel EBBA/aerosil R812 Fs ) 0.150 and (b) soft gel EBBA/aerosil A300 Fs ) 0.100 dispersions at 38 °C. The ZFC spectrum, IFC 0° angle spectrum and the IFC spectrum after 90° rotation.
soft gels shown in Figure 3. The stronger network formed by the hydrophilic silica combined with the weaker surface interaction of EBBA, as compared to nCB, makes it more difficult to break and reform the network in order to create an AAN. 3.5. “Soft Gel” Regime: Angle Dependence. To test for the presence of a stable AAN after IFC of the soft gels, an angle study was carried out. If the RAN is transformed into an AAN and the average void size is smaller than the magnetic coherence length, then the local LC directors are locked in the AAN voids. The energy splitting of the doublet arising from these domains should display a P2(cos θ) dependence, where θ is the angle between the nematic director and the applied field direction. If the LC directors are locked into a stable AAN, then rotating the sample by 90° should reduce the doublet splitting by half. Such an angle dependence is shown in Figure 6b for a hydrophilic silica gel dispersion. Recalling that the IFC spectrum of the EBBA/aerosil A300, Fs ) 0.100, consists of a superposition of a Pake pattern and a full splitting doublet (where θ ) 0°), the rotation of the sample by 90° reduced the doublet splitting by half such that it overlaps with the Pake pattern. This angle dependence supports the partial formation of a stable AAN in the case of the hydrophilic silica soft gel. The behavior of the hydrophobic dispersions which display a full splitting doublet after IFC is more complex. In this case, only a portion of the full splitting doublet component present after IFC displays the P2(cos θ) dependence upon rotation. The component displaying P2(cos θ) dependence increases with the R812 aerosil density. As shown in Figure 6, an R812 dispersion in the lower range of the stiff gel regime displays similar angle dependence behavior to the A300 soft gel. A similar behavior was reported for the IFC 8CB/aerosil A300 soft gels at low silica densities in the nematic phase.4 This behavior contrasts with the more viscous smectic phase where the entire full splitting doublet of the 8CB soft gels display the P2(cos θ) dependence. The full splitting doublet component of the 8CB dispersions in the nematic phase with no angle dependence was attributed to LC molecules which are contained in a weakly
linked AAN. As the sample is rotated, the weakly linked silica strands break and the LC molecules are reoriented to lie parallel to the applied field. This mechanism would be more strongly at play in the case of the EBBA/hydrophobic silica dispersions where the network is weakly linked even at high filler densities due to the small quantity of surface hydroxyl groups. 3.6. “Stiff Gel” Regime. The hydrophilic aerosil dispersions with Fs > 0.1 are classified as stiff gels. In this regime, the silica concentration is high enough to create a complete network, the induced random disorder dominates and the orientation order of the LC directors is low. Only weak in-field annealing of the LC directors can occur, since the magnetic coherence length exceeds the mean void size. In the lower aerosil density range of stiff gel dispersions, Fs > 0.1 for A300 and Fs > 0.15 for R812, the ZFC spectra consist of a superposition of a strong Pake pattern and a weak full splitting doublet (see Figures 2 and 3). As discussed earlier, these spectra arise from a void size distribution or the coexistence of random and anisotropic aerosil networks. The ZFC spectra for both the hydrophilic and hydrophobic dispersions for Fs ) 0.250 are perfect powder patterns with flat shoulders, and no angular dependence is observed after the IFC. All nematic domains are random, the average size of the voids is smaller than the magnetic coherence length and there are no reorientations or breaking of silica strands once the sample is placed in the magnetic field. The behavior of the system during the temperature cycle is again similar to the no gel regime in that no hysteresis is observed. An important difference with the soft gel regime occurs when the temperature approaches the TN-I transition. As the splitting decreases with an increase temperature, all of the peaks collapse into one broad peak. The full width at halfmaximum (fwhm) of the broad singlet decreases as the phase transition is approached. With increasing aerosil concentration, the temperature at which the collapse occurs decreases and the temperature range over which isotropic- and nematic-like peaks coexist increases. The phase transition than becomes less sharp, and a continuous evolution of the orientational order is observed during the temperature cycles. A similar behavior is observed
Silica Dispersions in EBBA Liquid Crystals
J. Phys. Chem. B, Vol. 112, No. 11, 2008 3327
Figure 7. DNMR spectra of the stiff gel dispersions of the aerosils (a) R812 and (b) A300 in EBBA at 38 °C.
with the stiff gel regime of 8CB/A300 dispersions. The interpretation given by Jin et al., who observed collapsing of the nematic doublet into a singlet at this gel regime for the 8CB/ A300 system, is that the LC domains are size limited such that the spectra are influenced by translational diffusion.4 In the stiff gel, the nematic order is unable to grow beyond the domain size. The LCs centered in the voids therefore have a distinct orientation compared to the molecules at the silica surface. In this case, the scale at which the nematic director varies is proportional to the diffusion length of the molecules (8CB diffusion length, χ0 ≈ 10 nm). This leads to a motional narrowing of the energy splitting and the broadening of the phase transition and accounts for the transformation of the Pake pattern into a broad singlet at very high aerosil densities in the stiff gel regime, as seen in Figure 7. The narrowing effect is observed at lower aerosil densities for the hydrophilic silica when compared to the hydrophobic silica due to the smaller void size and higher disorder induced. 4. Summary The disorder induced by hydrophilic (A300) and hydrophobic (R812) aerosils dispersed in a Schiff-base-type liquid crystal with a small dipole moment, p-ethoxy(benzylidene)-p-n-butylaniline (2O.4), EBBA, was assessed using DNMR as a structural probe. The filler density was varied and the resulting dispersions studied under zero- and in-field cooling. As expected, no network is observed in the no gel regime, Fs < 0.01, whereas, in the soft gel regime, Fs > 0.01, a partial to complete network is formed depending on the type of silica. In-field cooling of the hydrophilic silica dispersions in EBBA leads to the coexistence of random (RAN) and anisotropic (AAN) networks, whereas the hydrophobic aerosil forms a complete AAN under the same conditions. Angle dependence studies demonstrate that the partial AAN of the hydrophilic silica is stable but the hydrophobic silica network readily rearranges under in-field sample rotation. These results indicate that a sufficiently strong LC surface anchoring strength must be combined with hydrophilic silica to form an AAN which is both complete and mechanically robust. Further tuning of the LC surface interactions by trying out other types of fillers, such as alumina aerosils, as well as using LCs whose headgroups contain strongly hydrogen bonding functional groups, are underway. Acknowledgment. Funding for this work was provided by the Natural Sciences and Engineering Council of Canada
(NSERC) and Fonds Que´be´cois de la Recherche sur la Nature et les Technologies (FQRNT). We thank Dr. Frederick Morin of the McGill NMR facility for his assistance. References and Notes (1) Eidenschink, R.; De Jeu, W. H. Electron. Lett. 1991, 27 (13), 1195. (2) Keruzer, M.; Tschudi, T. Mol. Cryst. Liq. Cryst. 1992, 223, 219. (3) Crawford, G. P.; Zˇ umer, S. Liquid crystals in complex geometries: formed by polymer and porous networks; Taylor & Francis: London, 1996; Chapter 15, pp 307-324. (4) Jin, T.; Finotello, D. Phys. ReV. E 2004, 69, 041704. (5) Iannacchione, G. S. Fluid Phase Equilib. 2004, 223, 177. (6) Iannacchione, G. S; Garland, C. W.; Mang, J. T.; Rieker, T. P. Phys. ReV. E 1998, 58, 5966. (7) Glushchenko, A.; Kresse, H.; Puchkovska, O.; Reshetnyak, V.; Reznikov, Yu.; Yaroshchuk, O. Mol. Cryst. Liq. Cryst. 1998, 321, 15. (8) (a) Bellini, T.; Buscaglia, M.; Chioccoli, C.; Mantegazza, F.; Pasini, P.; Zannoni, C. Phys. ReV. Lett. 2000, 85, 1008. (b) Bellini, T.; Buscaglia, M.; Chioccoli, C.; Mantegazza, F.; Pasini, P.; Zannoni, C. Phys. ReV. Lett. 2002, 88 (24), 245506. (c) Rotunno, M.; Buscaglia, C.; Chioccoli, C.; Mantegazza, F.; Pasini, P.; Bellini, T.; Zannoni, C. Phys. ReV. Lett. 2005, 94, 097802. (d) Buscaglia, M.; Bellini, T.; Chioccoli, C.; Mantegazza, F.; Pasini, P.; Rotunno, M.; Zannoni, C. Phys. ReV. E 2006, 74, 011706. (9) (a) Arcioni, A.; Bacchiocchi, C.; Grossi, L.; Nicolini, A.; Zannoni, C. J. Phys. Chem. B 2002, 106, 9245. (b) Arcioni, A.; Bacchiocchi, C.; Vecchi, Venditti, G.; Zannoni, C. Chem. Phys. Lett. 2004, 396, 433. (10) Glushchenko, A.; Kresse, H.; Reshetnyak, V.; Reznikov, Yu.; Yaroshchuk, O. Liq. Cryst. 1997, 23 (2), 241. (11) Puchkovskaya, G.; Reznikov, Yu.; Yakubov, A.; Yaroshchuk, O.; Glushchenko, A. J. Mol. Struct. 1997, 404, 121. (12) Frunza, L.; Kosslick, H.; Bentrup, U.; Pitsch, I.; Fricke, R.; Frunza, S.; Schonhals, A. J. Mol. Struct. 2003, 651, 341. (13) Crawford, G. P.; Zˇ umer, S. Liquid crystals in complex geometries: formed by polymer and porous networks; Taylor & Francis: London, 1996; Chapter 7, pp 159-186. (14) Haga, H.; Garland, C. W. Phys. ReV. E 1997, 56 (3), 3044. (15) Haga, H.; Garland, C. W. Liq. Cryst. 1997, 23 (5), 645. (16) Lobo, C. V.; Prasad, S. K.; Yelamaggad, C. V. J. Phys.: Condens. Matter 2006, 18, 767. (17) Technical Bulletin Fine Particles, Number 11, Basic Characteristics of AEROSIL® Fumed Silica, 4th ed., Degussa Corporation. (18) (a) van der Est, A. J. Ph.D. Thesis, The University of British Columbia, Vancouver, BC, 1987. (b) Kelke, H.; Schuelle, B. Angew. Chem., Int. Ed. Engl. 1969, 8, 884. (19) Dolganov, V. K.; Gal. M.; Kroo, N.; Rosta, L.; Szabon, J. Liq. Cryst. 1987, 2 (1), 73. (20) Jakli, A.; Kali, Gy.; Rosta, L. Physica B 1997, 234, 297. (21) Sharma, D.; MacDonald, J. C.; Iannacchione, G. S. J. Phys. Chem. B 2006, 110, 26160.
