Water Dynamics in Bolaamphiphile Hydrogels Investigated by 1H

Dec 13, 2010 - Institute of Chemistry, Martin-Luther-UniVersity Halle-Wittenberg, Von-Danckelmann-Platz 4,. 06120 Halle/Saale, Germany, and Institute ...
0 downloads 0 Views 1MB Size
Download PDF
14

J. Phys. Chem. B 2011, 115, 14–22

Water Dynamics in Bolaamphiphile Hydrogels Investigated by 1H NMR Relaxometry and Diffusometry Martin Bastrop,†,‡ Annette Meister,† Hendrik Metz,‡ Simon Drescher,†,‡ Bodo Dobner,‡ Karsten Ma¨der,*,‡ and Alfred Blume*,† Institute of Chemistry, Martin-Luther-UniVersity Halle-Wittenberg, Von-Danckelmann-Platz 4, 06120 Halle/Saale, Germany, and Institute of Pharmacy, Martin-Luther-UniVersity Halle-Wittenberg, Wolfgang-Langenbeck-Str. 4, 06120 Halle/Saale, Germany ReceiVed: August 16, 2010; ReVised Manuscript ReceiVed: NoVember 5, 2010

The dynamical properties of water in hydrogels formed by the bolaamphiphiles dotriacontane-1,32-diyl-bis[2(trimethylammonio)ethylphosphate] (PC-C32-PC) and its head group derivative dotriacontane-1,32-diyl-bis[2(dimethylammonio)ethylphosphate] (Me2PE-C32-Me2PE) were investigated by determining the transverse relaxation times and the mean diffusion coefficients of the water protons. T2 distributions obtained for hydrogels containing only 1 mg/mL of PC-C32-PC or Me2PE-C32-Me2PE were found to be almost unchanged compared to that of pure deionized water. In a hydrogel containing 30 mg/mL PC-C32-PC, a small dynamical perturbation of the water molecules could be observed, since the obtained major peak in the T2 distribution was found to be slightly broadened and shifted to lower T2 values with respect to pure water. Moreover, a slightly decreased mean diffusion coefficient of the water molecules was determined for the hydrogel with 30 mg/mL PC-C32PC. For 30 mg/mL Me2PE-C32-Me2PE at pH 5, the dynamical properties of water are significantly influenced by the presence of the bolaamphiphiles. Whereas the mean diffusion coefficient of water was again only slightly decreased, the relaxation behavior was found to be considerably changed. At room temperature, the major T2 peak obtained for 30 mg/mL Me2PE-C32-Me2PE at pH 5 was significantly broadened and shifted to lower T2 values compared to pure buffer. This broad monomodal peak becomes bimodal when the temperature is decreased to 5 °C. Increasing the temperature revealed that the structural changes of the bolaamphiphilic self-assemblies at 45 °C are reflected in the determined mean relaxation times (T2m). These results indicate the presence of a significant degree of structural heterogeneity in a hydrogel formed by 30 mg/mL Me2PE-C32-Me2PE at pH 5. By comparing the mean diffusion coefficients with the mean transverse relaxation times, the distance scale of these heterogeneities could be estimated to be in the range of 50 to 120-140 µm. They can most probably be ascribed to the existence of a considerable number of lamellar domains being formed besides the nanofibers in hydrogels with 30 mg/mL Me2PE-C32-Me2PE at pH 5. Upon storage, the phenomenon of syneresis was observed for this hydrogel, which corresponds to a growth of the lamellar domains to sizes between 140 and 190 µm. Introduction Gels combine the properties of fluids and solids. Despite the fact that the major portion by weight and volume is liquid, no flow is observed, since a gel can maintain its form under the stress of its own weight. Gels are used in food products, in cosmetics, and also in agriculture and oil industries. Gels containing water as the dispersive phase are called hydrogels. Due to their high water content and their soft consistency, hydrogels usually possess good biocompatibility and resemble natural living tissue more than any other class of synthetic biomaterials.1 Therefore, they have found numerous applications in the biomedical field, for example as contact lenses,2 as drug delivery systems,3 and in regenerative applications.4 There are many different hydrogel systems available today, which are all composed of an aqueous phase and one or more hydrophilic or amphiphilic components that are cross-linked into a threedimensional network by either covalent bonds or noncovalent interactions.5 Most of these hydrogelators are natural or synthetic * To whom correspondence should be addressed. E-mail:alfred.blume@ chemie.uni-halle.de (A.B.), [email protected] (K.M.). † Institute of Chemistry, Martin-Luther-University Halle-Wittenberg. ‡ Institute of Pharmacy, Martin-Luther-University Halle-Wittenberg.

polymers, but the ability to gel water was also reported for several low molecular weight (LMW) molecules as well.6,7 The common feature of these LMW hydrogelators is their amphiphilic character that determines their specific self-assembly into a three-dimensional network of fibers, which is required for hydrogelation. Recently, a class of symmetrical bipolar amphiphiles, socalled bolaamphiphiles, was synthesized, which turned out to be highly efficient LMW hydrogelators.8-10 These molecules consist of a long alkyl chain with phosphocholine head groups attached on either side. The capability of network formation and its dependence on various factors, such as chain length or head group variations, concentration, pH, and temperature, were intensively studied by various methods.10-13 Most of these investigations were carried out on the bolalipid dotriacontane-1,32diyl-bis[2-(trimethylammonio)ethylphosphate] (PC-C32-PC, Figure 1) and its head group derivative dotriacontane-1,32-diylbis[2-(dimethylammonio)ethylphosphate] (Me2PE-C32-Me2PE). At room temperature, both bolaamphiphiles self-assemble into well-defined nanofibers that are several micrometers long. By increasing the temperature, reversible transitions from the long

10.1021/jp107755k  2011 American Chemical Society Published on Web 12/13/2010

Water Dynamics in Bolaamphiphile Hydrogels

J. Phys. Chem. B, Vol. 115, No. 1, 2011 15

Figure 1. Chemical structures of the symmetrical single-chain bolaamphiphiles PC-C32-PC (R ) CH3) and Me2PE-C32-Me2PE (R ) H).

nanofibers to short rods or micellar aggregates can be induced and detected as endothermic peaks in the DSC thermograms.9,10 For all potential biomedical applications, the physicochemical properties of water inside these hydrogel systems are of crucial importance, since they mainly determine the transport processes through the gel, such as the release of drugs or the nutrient supply for cells and tissues. The properties of water in a hydrogel are affected by its interaction with the gel network, by the pore sizes, and by the pore interconnections.14 One can generally distinguish two classes of water inside a hydrogel, which are in exchange with each other: one class of water molecules that is remote from the network and is not influenced by its presence (unperturbed water) and one that is interacting with the network and therefore changed in its dynamical properties compared to pure water (perturbed water).15 For hydrogels with very low water content, only the second class might be present. Hydrogels formed by PC-C32-PC and Me2PE-C32-Me2PE have a high water content of about 95-99.9%. Hence, one can expect a considerable amount of unperturbed water. The physicochemical properties of water in hydrogel systems were studied by a number of techniques, such as calorimetry, dilatometry, differential thermal analysis, specific conductivity, neutron scattering, and nuclear magnetic resonance spectroscopy. We have chosen the proton nuclear magnetic resonance (1H NMR) technique to gain information about the dynamics of the water molecules in hydrogels of PC-C32-PC and Me2PE-C32Me2PE. 1H NMR relaxometry allows the measurement of the longitudinal (T1) and the transverse relaxation times (T2) of the protons. Nuclear magnetic relaxation requires fluctuating magnetic fields of proper frequencies. These fluctuating magnetic fields result from rotations and Brownian motion of the water molecules. Therefore, NMR relaxation is highly sensitive to the dynamics of the water molecules. Relaxation processes in 1H NMR are dominated by dipole-dipole interactions between neighboring protons. In the present study, the transverse relaxation behavior of the water protons was studied for hydrogels formed by PCC32-PC and Me2PE-C32-Me2PE at concentrations of 1 or 30 mg/mL in the temperature range between 5 and 80 °C. The transverse magnetization decays were obtained with the spin echo technique by applying the Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence.16,17 These magnetization decays were fitted with the program WinDXP. The obtained T2 distributions were found to be dominated by the water protons and could therefore be used to determine the mean T2 values (T2m) of the water protons. Furthermore, the shapes of the T2 peaks provide additional information about the proton relaxation behavior. Besides the transverse relaxation times, the mean diffusion coefficients of the water molecules were measured for the hydrogels at different temperatures by means of the pulsedgradient stimulated-echo (PGSTE) method developed by Stejskal and Tanner.18,19 The observed differences regarding the relaxation behavior of water in hydrogels of PC-C32-PC and Me2PE-C32-Me2PE were interpreted in terms of the presence of a certain degree of heterogeneity inside the samples with 30 mg/mL Me2PE-C32Me2PE at pH 5, but not inside that of PC-C32-PC at the same concentration. As will be discussed below, these heterogeneities

can be ascribed to extended domains of lamellar aggregates that are formed by Me2PE-C32-Me2PE besides the nanofibers. Materials and Methods Chemicals. Dotriacontane-1,32-diyl-bis[2-(trimethylammonio)ethylphosphate] (PC-C32-PC) and dotriacontane-1,32-diylbis[2-(dimethylammonio)ethylphosphate] (Me2PE-C32-Me2PE) were synthesized as described previously.8,13 Sodium acetate and acetic acid (100%) were purchased from Merck (Darmstadt, Germany) and Riedel-de Hae¨n (Seelze, Germany), respectively. Instrumental Settings. All 1H NMR measurements were carried out with a low-field benchtop Maran Ultra 1H NMR spectrometer from Oxford Instruments, UK. The spectrometer is equipped with an air flow temperature regulation and a 3D imaging unit. Transverse magnetization decays were obtained between 5 and 80 °C by applying the CPMG pulse sequence, which involves a single 90° radio frequency (rf) pulse and a series of 180° rf pulses. Each pulse sequence was detecting 36 864 echoes with the time 2τ ) 0.27 ms between two echoes and a relaxation delay time of 20 s. The transverse magnetization decays were fitted with the program Windows Distributed EXPonential analysis software (WinDXP, Oxford Instruments, Abingdon, UK) and T2 distributions with 256 points are calculated in the relaxation time range from 10 µs to 20 s. Mean diffusion coefficients of water were determined by means of the pulsed-gradient stimulated-echo (PGSTE) method. These measurements were performed in the temperature range from 30 to 52 °C with the following parameters: The time D1 (also denoted as τ1 in Figure S1 in the Supporting Information) between the 90° rf pulse and the first field gradient pulse was always 0.1 ms The time D2 (τ1 + δ + τ2,1) was 7 ms. D3 (δ), the duration of the field gradient pulse, was varied in the range from 0.28 to 1.5 ms in order to obtain decreasing echo intensities. The theoretical amplitude g of the gradient field was 0.922 T/m. Calibration of the measurement with pure deionized water leads to a correction factor of 0.784. Therefore, the value of g was calculated to be 0.723 T/m. The time D4 (∆), which is the time between the two gradient pulses, was adjusted to 20 ms. Sample Preparation. One milligram or 30 mg PC-C32-PC was weighed into a small glass vessel and 1 mL of deionized water was added to obtain a PC-C32-PC concentration of 1 or 30 mg/mL, respectively. The carefully closed glass vessel was alternately heated in a water bath to a temperature above 80 °C and vortexed until all PC-C32-PC was homogeneously dispersed. No efforts were made to exclude oxygen from the samples. Prior to the NMR measurements, the samples were stored between 2 and 8 °C for 2 h. For Me2PE-C32-Me2PE, the preparation was the same, but 100 mM acetate buffer at pH 5 was used instead of deionized water. Each sample was prepared and measured twice. Results and Discussion PC-C32-PC. Initially, the transverse relaxation behavior of pure deionized water was determined at temperatures ranging

16

J. Phys. Chem. B, Vol. 115, No. 1, 2011

Figure 2. Normalized T2 distributions obtained at 25 °C for hydrogels containing 1 or 30 mg/mL PC-C32-PC compared to that of pure deionized water.

from 25 to 80 °C. The measured decays of the transverse magnetization were fitted with the program WinDXP. Analogous to the well-known program CONTIN,20,21 WinDXP uses Laplace inversion to calculate continuous T2 distributions. WinDXP was used before to investigate the relaxation behavior of water in a variety of samples, such as starch,22,23 silica gels,24 soil,25 and chicken breast.26 Figure 2 displays a typical full-range T2 distribution obtained for deionized water at 25 °C. It shows one major peak at approximately 2.8 s, which is representing the relaxation behavior of the majority of the water protons. In addition, two small peaks are visible. Experiments with increasing glass surface-to-volume (A/V) ratios indicate that these peaks are related to the relaxation of water protons at the glass surface, which acts as a relaxation sink.21 At A/V ratios of about 0.6 to 4.1 mm-1, the peak areas of the small peaks take up only 5-12% of the total peak area, whereas at a higher ratio of about 32.1 mm-1, the peak area increases to about 55% (Figure S2 in the Supporting Information). Only for the latter sample, the shape and the position of the major T2 peak were found to be significantly influenced by the process of surface relaxation. Therefore, the major T2 peak obtained for the investigated samples can be fully attributed to the dynamical properties of the water molecules in the particular sample and not to surface phenomena. The center of this major peak of the T2 distributions was used to determine the mean T2 values, T2m, of the particular sample. For pure deionized water, the major T2 peak was found to be gradually broadened and shifted to higher T2 values when the temperature is increased from 25 to 80 °C (Figure 3). The corresponding T2m values are linearly increasing by approximately 70 ms/K from 2.84 s at 25 °C to 6.74 s at 80 °C (Figure 4). This is due to the fact that the dipole-dipole relaxation process requires fluctuating local magnetic fields of low frequency and short distances between the interacting species. Therefore, transverse relaxation becomes less efficient with increasing thermal motions of the water molecules and as a consequence, the relaxation rate R2 ) 1/T2, decreases. The observed broadening of the major T2 peak with increasing temperature is due to a wider distribution of the rotational correlation times. PC-C32-PC is a very efficient LMW hydrogelator, since it gels water at concentrations as low as 1 mg/mL.9 After dispersion at temperatures above 80 °C, where micelle-like aggregates are formed, the PC-C32-PC molecules self-assemble into helically structured nanofibers when the sample is cooled

Bastrop et al.

Figure 3. Deionized water at temperatures from 25 to 80 °C: Major peaks of the calculated T2 distributions normalized to the same peak area.

Figure 4. Deionized water and 30 mg/mL PC-C32-PC in deionized water at temperatures from 25 to 80 °C: the determined T2m values (accuracy ) (0.02 s; g60 °C, (0.035 s) are plotted against temperature. For comparison, the T2m values for pure buffer and the resulting REFs are also shown.

down to room temperature. Unlike other LMW hydrogelators, PC-C32-PC cannot form hydrogen bonds between the head groups. Therefore, its self-assembly into fibers and the crosslinking between these fibers is solely driven by hydrophobic interactions between the long alkyl chains. The role of water in these processes is not fully understood up to now. Therefore, the question whether the presence of the three-dimensional network of bolaamphiphilic nanofibers influences the relaxation behavior of the water molecules is of importance. By increasing the temperature, the reversible fiber-micelle transition can be induced and the gel character is lost. Hence, the question whether this macroscopic change from a hydrogel to a micellar fluid causes a detectable change in the relaxation behavior of the water molecules in this system was also addressed. As Figure 2 shows, there is no significant difference observable between the T2 distributions of pure deionized water and the hydrogel containing only 1 mg/mL PC-C32-PC. Increasing the PC-C32-PC concentration to 30 mg/mL results in some small differences in the T2 distribution. The major peak becomes slightly broadened and is shifted to lower T2 values. This peak again represents the majority of the protons of the water molecules and the determined T2m value is decreased by about 0.3 s with respect to that of pure water. Besides the peaks that correspond to proton relaxation at the glass surface, an additional T2 peak is detected at approximately 1 ms. This peak represents

Water Dynamics in Bolaamphiphile Hydrogels

J. Phys. Chem. B, Vol. 115, No. 1, 2011 17

Figure 5. PC-C32-PC, 30 mg/mL, in deionized water at temperatures from 25 to 80 °C: major peaks of the calculated T2 distributions normalized to the same peak area.

the relaxation of the protons of the bolaamphiphile molecules as could be confirmed by experiments with D2O (data not shown). The temperature dependence of the relaxation behavior was only investigated for the hydrogel containing the higher concentration, namely 30 mg/mL PC-C32-PC. In analogy to pure H2O, the major T2 peaks become significantly broadened and gradually shifted to higher T2 values with increasing temperature from 25 to 80 °C (Figure 5). The T2m values also increased linearly by about 60 ms/K with increasing temperature (Figure 4). Above 50 °C, the fibers are replaced by micellelike aggregates and the hydrogel character is lost. However, this change in the aggregate morphology is not reflected in the T2 distributions or in the determined T2m values, which show no discontinuity at this temperature. Therefore, there is obviously no relationship observable between the formed aggregates and the relaxation behavior of the water protons for hydrogels formed by PC-C32-PC. The conclusion is that the relaxation behavior of the water protons seems to be almost unaffected by the presence of the dense fiber network even when the PC-C32-PC concentration was increased to 30 mg/mL. There is only a slight decrease of the T2m values observable for the hydrogel with 30 mg/mL PCC32-PC compared to pure water. As shown by Roorda et al. for poly(hydroxyethyl methacrylate) (pHEMA) hydrogels,27 the observed small effect on the relaxation of the water protons can be fully explained as a concentration effect. These authors determined the relaxation rate enhancement factors (REF)

R1,2 (hydrogel) T1,2 (water) REF ) ) R1,2 (water) T1,2 (hydrogel)

(1)

for a large number of hydrogel samples with varying polymer contents, cross-linking densities and number of hydrophilic binding sites. Roorda and co-workers found the REF to be dependent only on the polymer content, but independent of the number of binding sites or on the degree of cross-linking. The authors therefore concluded the decrease in the average mobility of water in the pHEMA gel to be mainly determined by the concentration of the polymer and not by the cross-linker content. Another important conclusion was drawn by Roorda et al. from their 17O relaxation study: All the water in pHEMA hydrogels at 25 °C was found to be in exchange on a time scale of a millisecond or less. This means that, if different states of water are present at this temperature, the exchange between these states is so fast that it results in the detection of only one averaged

relaxation rate. The authors could therefore confirm earlier results obtained by thermal analysis of pHEMA hydrogels28,29 that rule out the actual presence of thermodynamically different classes of water, which were proposed before.30 The relaxation rate enhancement factors were also determined for PC-C32-PC. The results are shown in Figure 4 and confirm that the relaxation behavior of the water protons is independent of the type of the PC-C32-PC aggregates, since no discontinuity could be observed. The uniformly low REFs as well as the relatively narrow major T2 peak obtained for the sample with 30 mg/mL PCC32-PC can be explained as follows: because of the small volume fraction of PC-C32-PC, the particular bolaamphiphilic aggregates can be expected to be uniformly distributed throughout the water volume. It can be assumed that only the first layers of water around the aggregates are dynamically perturbed by the aggregate surface.31-33 These perturbed water molecules are in a diffusional exchange with the surrounding unperturbed water on a submillisecond level. Extremely rapid exchange can be assumed, since the diffusion distance from the surface layers to the surrounding water volume is very short. As a consequence, the water molecules will experience only a small perturbation by the aggregate surface. The resulting influence on the proton relaxation behavior is therefore very small, which is reflected in a major T2 peak that is only slightly broadened and shifted with respect to that of pure water. As described above, diffusion of the water molecules has a large influence on the obtained T2 distributions. Therefore, the mean diffusion coefficient of water was determined for the sample with 30 mg/mL PC-C32-PC by applying the pulsedgradient stimulated-echo (PGSTE) method.18,19 Stimulated echoes were recorded for different values of the duration (D3) of the gradient pulses. With increasing D3, the echo amplitudes (A2τ) decrease substantially compared to the echo amplitude obtained without the application of the gradient pulses (A0). The decreasing echo amplitudes are related to the mean diffusion coefficient of water (Dm) by18

( )

ln

A2τ 1 ) -γ2g2Dmδ2 ∆ - δ ) A0 3

(

)

1 - γ2g2DmD32 D4 - D3 3

(

)

(2)

The parameters D4 (∆), D3 (δ), and g were specified above (Materials and Methods) and γ is the magnetogyric ratio. The obtained values of ln(A2τ/A0) were plotted against the term D32(D4 - 1/3D3). By linear fitting, the values for the slope were obtained. According to eq 2, the slope is corresponding to -γ2g2Dm. With γ ) 267 538 030 s-1 T-1 and g ) 0.723 T/m, the values for the diffusion coefficient Dm of the water molecules in the hydrogels could be calculated. Figure 6 shows that the temperature-dependent mean diffusion coefficient determined for water in the PC-C32-PC hydrogel is slightly decreased with respect to the values of pure water. This small decrease in Dm of about 3-7% is in agreement with the results obtained for other hydrogels at comparable gelator concentrations.34 In these investigations, the diffusion coefficient was found to depend only on the solid volume fraction, but not on the macroscopic state35 or on the cross-linker content.36 Me2PE-C32-Me2PE. The transverse relaxation times were also determined for hydrogels formed by Me2PE-C32-Me2PE

18

J. Phys. Chem. B, Vol. 115, No. 1, 2011

Figure 6. Mean diffusion coefficients for deionized water, 30 mg/mL PC-C32-PC in deionized water, and 30 mg/mL Me2PE-C32-Me2PE in acetate buffer pH 5 plotted against temperature.

Figure 7. Normalized major T2 peaks obtained at 25 °C for hydrogels containing 1 or 30 mg/mL Me2PE-C32-Me2PE compared to PC-C32PC.

at pH 5. At first, the influence of the changed aqueous medium was tested. The temperature-dependent relaxation behavior of the water protons was found to be essentially the same for acetate buffer at pH 5 and pure deionized water (Figure S3 in the Supporting Information). Me2PE-C32-Me2PE forms a hydrogel in acetate buffer at pH 5 when the head groups are in their zwitterionic state. In analogy to PC-C32-PC, hydrogelation of Me2PE-C32-Me2PE at pH 5 is due to the formation of a three-dimensional network of helical nanofibers as could be shown by electron microscopy10 and SANS.12 Its self-assembly into fibers and the cross-linking between these fibers is again mainly driven by hydrophobic interactions. Additionally, a stabilizing effect by the formation of hydrogen bonds between NH(CH3)2+ and PO2- groups was observed as an increase of the transition temperatures with respect to PC-C32-PC, which causes the hydrogel state to persist up to higher temperatures.10 This is also reflected in a decreased critical gel concentration of approximately 0.5 mg/mL Me2PEC32-Me2PE10 compared to 1 mg/mL for PC-C32-PC. Because of the similarities in the aggregation behavior of Me2PE-C32-Me2PE at pH 5 compared to PC-C32-PC, one could also expect similarities in the relaxation behavior of the water protons inside their hydrogels. Figure 7 shows that this is the case only for the two samples containing 1 mg/mL of the particular bolaamphiphile. At a higher concentration of 30 mg/ mL Me2PE-C32-Me2PE, however, the relaxation behavior of the water protons was found to differ distinctly from that of a

Bastrop et al. sample with 30 mg/mL PC-C32-PC. The major T2 peak is considerably broadened and shifted to lower T2 values. Therefore, at a bolaamphiphile concentration of 30 mg/mL one can assume significant differences to be present between the two bolaamphiphiles regarding their interaction with the water molecules. In contrast to PC-C32-PC, Me2PE-C32-Me2PE molecules are capable of forming hydrogen bonds, which may lead to a stronger dynamical perturbation of the water molecules. Furthermore, they contain dissociable protons, which can influence the relaxation of the water protons via a proton exchange. For the sample with 30 mg/mL Me2PE-C32-Me2PE, these effects can be expected to cause a noticeable shift of T2m to lower values compared to PC-C32-PC. This could be clearly demonstrated by determining the relaxation behavior of water in a transparent hydrogel with 30 mg/mL Me2PE-C28-Me2PE at pH 5, which can be expected to be free of lamellar aggregates. For this sample, a narrow major T2 peak is observed (Figure S4 in the Supporting Information). The width of this peak roughly corresponds to that obtained for PC-C32-PC, but it is centered at approximately 1.4 s and therefore down-shifted by more than 1 s with respect to PC-C32-PC due to the effect of hydrogen bonding and proton exchange. For Me2PE-C32Me2PE, however, the decrease of the T2m value is even more pronounced than that of Me2PE-C28-Me2PE and, additionally, the extreme broadening of the major T2 peak is also observed. Therefore, more fundamental differences than the formation of hydrogen bounds or the proton exchange seem to be present between PC-C32-PC and Me2PE-C32-Me2PE. The changed relaxation behavior of the water protons in hydrogels containing 30 mg/mL Me2PE-C32-Me2PE at pH 5 is also reflected in the temperature dependence of the major T2 peak (Figure 8a). In accordance to the results discussed above, this major T2 peak was found to be shifted to higher values with increasing temperatures from 25 to 80 °C. This indicates an increasing mobility of the water molecules in the hydrogel sample. In contrast to the results obtained for the hydrogel of PC-C32-PC, the major T2 peak obtained for 30 mg/mL Me2PEC32-Me2PE was found not to be gradually broadened with increasing temperature, but to show a two-phase behavior. Between 25 and 50 °C, the peak becomes substantially narrower and its height is increased. In the second phase between 50 and 80 °C, the width of the distribution peak increases again accompanied by a decrease of the peak height. At 80 °C, the major T2 peak corresponds to that of 30 mg/mL PC-C32-PC at the same temperature (Figure S5 in the Supporting Information). The two-step behavior of the relaxation also becomes evident when the T2m values and the REFs are plotted against temperature (Figure 8b). For 30 mg/mL PC-C32-PC, the T2m values increase almost linearly by about 60 ms/K with increasing temperature. For Me2PE-C32-Me2PE, the T2m values also initially increase linearly (72 ms/K), but between 42 and 48 °C a clearly visible discontinuity indicates distinct changes in the relaxation behavior. In this region, T2m increases by approximately 150 ms/K and, simultaneously, a significant decrease in the REF values is detected between 40 and 50 °C. This points to drastic changes in the organization of the bolaamphiphile molecules. In fact, DSC thermograms show an endothermic peak at 45.5 °C, which is macroscopically related to a decrease in the number of cross-links inside the threedimensional network of nanofibers.10 Microscopically, an increase in the motional freedom of the alkyl chains and of the head groups was observed.10,12,37 However, the hydrogel state still persists above 45.5 °C. The transition temperature of 45 °C, determined from the T2m values, agrees well with that

Water Dynamics in Bolaamphiphile Hydrogels

J. Phys. Chem. B, Vol. 115, No. 1, 2011 19

Figure 9. Observation of a bimodal major T2 peak by decreasing the temperature from 25 to 5 °C for 30 mg/mL Me2PE-C32-Me2PE at pH 5. For 30 mg/mL PC-C32-PC in deionized water, a bimodal major T2 peak is not obtained at 5 °C.

Figure 8. Me2PE-C32-Me2PE, 30 mg/mL, in acetate buffer pH 5 at temperatures from 25 to 80 °C. (a) Major peaks of the calculated T2 distributions normalized to the same peak area. (b) The determined T2m values (accuracy ) (0.02 s; g60 °C, (0.035 s) are plotted against temperature. For comparison, the T2m values for pure buffer and the resulting REFs are also shown.

obtained by DSC. This allows to conclude that the determination of the relaxation behavior of the water protons by 1H NMR is capable of detecting the temperature-induced transition in the investigated Me2PE-C32-Me2PE hydrogel. Similar results were reported, for example, for hydrogels formed by carrageenan38 or acrylamide.39,40 The second transition of Me2PE-C32-Me2PE, detected by DSC at 69.5 °C, is known to be related to a complete breakdown of the three-dimensional network. This transition is not detected by the plotted T2m and REF values. One can therefore conclude that there is no correlation between the hydrogel state and the observed relaxation behavior analogous to the results obtained for the PC-C32-PC sample. On the other hand, there is obviously a correlation between the relaxation behavior of the water protons and the self-assembled Me2PEC32-Me2PE at temperatures up to the first transition temperature, which was not observed for PC-C32-PC. By decreasing the temperature from 25 to 5 °C, the broad monomodal major T2 peak becomes bimodal for the hydrogel with 30 mg/mL Me2PE-C32-Me2PE (Figure 9). The two separated peaks are centered at approximately 0.43 and 1.75 s. For 30 mg/mL PC-C32-PC, however, only a narrow monomodal major T2 peak was obtained at 5 °C. For the sample with 30 mg/mL Me2PE-C32-Me2PE at pH 5, the mean diffusion coefficients of the water molecules were also determined. As Figure 6 shows, the translational mobility of the water molecules is essentially the same for both hydrogels, although there are distinct differences observable in the relaxation behavior. Hence, one can expect the diffusion of the water molecules in the Me2PE-C32-Me2PE hydrogel to be almost unrestricted compared to the pure aqueous medium.

Figure 10. Syneresis upon storage observed for the hydrogel containing 30 mg/mL Me2PE-C32-Me2PE at pH 5. (a) Freshly prepared NMR sample. (b) NMR sample after storage at room temperature for 48 h. (c) After syneresis, two separated populations of water protons become visible in the T2 distribution at 25 °C, but at 40 °C a broad monomodal major T2 peak is again obtained.

The hydrogel sample with 30 mg/mL Me2PE-C32-Me2PE at pH 5 also revealed interesting behavior upon storage. Samples for NMR measurements were completely transparent for PCC32-PC and did not show any changes in their properties upon storage for several weeks. The NMR sample of Me2PE-C32Me2PE was only transparent directly after the preparation, but became highly turbid during the next 2 h of storage time prior to the NMR measurements (Figure 10a). Upon storage at room temperature for only 2 days, this sample was found to show shrinkage of the hydrogel volume, which was accompanied by an expulsion of part of the water. This well-known phenomenon of syneresis was not observed before for Me2PE-C32-Me2PE at pH 5 due to the low concentrations (e10 mg/mL) or small sample volumes (e100 µL) that were used for other investigations. Instead, the gel structure was found to be stable for at least several months for hydrogels formed by Me2PE-C32-

20

J. Phys. Chem. B, Vol. 115, No. 1, 2011

Me2PE at lower concenrations.10 A picture of the hydrogel after syneresis is presented in Figure 10b. The cylindrical shape is preserved upon shrinkage and about one-third of the total buffer content is expelled. The resulting T2 distribution of this sample at 25 °C shows a splitting of the major peak into two separated peaks (Figure 10c). These separated peaks at 0.95 and 2.75 s obviously correspond to one fraction of water protons that are highly perturbed by the presence of the bolaamphiphiles and a second fraction that is almost unperturbed, respectively. At 40 °C, however, only a monomodal major T2 peak is obtained for the same sample (Figure 10c), although these two fractions can be expected to be still present at this temperature. The explanation for the observed broad peak at room temperature and its splitting at 5 °C as well as for the phenomenon of syneresis is the presence of a significant degree of heterogeneity inside the hydrogel formed by 30 mg/mL Me2PE-C32-Me2PE at pH 5. These heterogeneities are most probably due to a considerable number of lamellar aggregates, which coexist with the nanofibers. Only a small number of these lamellar aggregates was observed in the cryo-TEM images of samples containing 0.3 mg/mL Me2PE-C32-Me2PE as darker almost rectangular areas, but their existence could not be explained at that time.10 Meanwhile, lamellar aggregates could be identified for other symmetrical single-chain bolaamphiphiles. MePE-C32-MePE, the monomethyl derivative of PC-C32-PC, is not self-assembling into nanofibers, but only into lamellar aggregates.13 Higher numbers of rectangular lamellar aggregates were also found for the longer chain bolaamphiphiles Me2PE-C34-Me2PE and Me2PE-C36-Me2PE at pH 5.41 For these bolaamphiphiles, the phenomenon of syneresis was also observed and could be attributed to the transfomation of nanofibers, which seem to be the kinetically favored aggregates observed after cooling, into the thermodynamically stable lamellar aggregates observed after storage. The shrinkage of the gel volume is directly linked to the formation of the lamellae, which may be present in large clusters of stacked single lamellae. A model for the arrangement of the bolaamphiphile molecules in these lamellar aggregates was also proposed. According to this model, the lamellae are formed by tilted bolalipids interdigitated in a crossed fashion.41 The formation of such lamellae leads to a significant reduction of the surface that is exposed to the aqueous medium compared to the nanofibers. The first transition, detected for Me2PE-C34Me2PE and Me2PE-C36-Me2PE at approximately 50 and 60 °C, respectively, is obviously related to a disintegration of the lamellae, since only fibers were found above this transition temperature.41 Using the negative staining technique, we tried to obtain transmission electron images for the hydrogel with 30 mg/mL Me2PE-C32-Me2PE at pH 5. This approach was not successful due to the high bolaamphiphile concentration that prevented the observation of well-defined aggregates. Therefore, there is still no direct experimental evidence for the existence of lamellar domains in this hydrogel. However, the results obtained for the longer chain analogues Me2PE-C34-Me2PE and Me2PE-C36Me2PE clearly support the assumption that also for 30 mg/mL Me2PE-C32-Me2PE at pH 5, a large number of lamellar aggregates is formed. The highly turbid appearance and the occurrence of syneresis for this sample are in agreement with the observations reported for Me2PE-C34-Me2PE and Me2PE-C36-Me2PE. The major T2 peaks obtained for water in samples with 30 mg/mL Me2PEC34-Me2PE and Me2PE-C36-Me2PE (Figure S4 in the Supporting Information) clearly confirm the interpretation of the

Bastrop et al. NMR results of Me2PE-C32-Me2PE in terms of the presence of the lamellae. For Me2PE-C34-Me2PE, a further shift and broadening is observed compared to Me2PE-C32-Me2PE, whereas the major T2 peak obtained for Me2PE-C36-Me2PE is even bimodal due to syneresis during storage of this sample. Therefore, regarding the nature of the formed aggregates that result from the self-assembly of the bolaamphiphiles, Me2PEC32-Me2PE appears to represent an intermediate state in the Me2PE-Cn-Me2PE family. Whereas for the aqueous samples with 0.3 mg/mL of the polymethylene-1,ω-(phosphodimethylethanolamines) with shorter chain lengths (n ) 22-30) only fibers and micelles were observed in the cryo-TEM images,13 a considerable number of lamellar aggregates could additionally be found for Me2PE-C34-Me2PE and Me2PE-C36-Me2PE when the samples were quenched from 25 °C.41 For 0.3 mg/mL Me2PE-C32-Me2PE at pH 5, such lamellar aggregates were also observed, but only at a relatively low number.10 The NMR sample containing only 1 mg/mL Me2PE-C32-Me2PE can therefore be expected to also contain these lamellar aggregates. However, their number seems to remain rather low even upon storage for a month, since no turbidity or syneresis was observed for this sample. This explains the relaxation behavior of the water protons in this sample, which was found to be almost unchanged compared to the pure buffer and to the sample with 1 mg/mL PC-C32-PC (Figure 7). As proposed here, the number of the lamellar aggregates is dramatically increasing when the concentration of Me2PE-C32Me2PE is increased from 1 to 30 mg/mL, which leads to the turbid appearance of the sample and to the observed broad major T2 peak. The presented NMR relaxation data indicate that the lamellar aggregates are not uniformly distributed throughout the hydrogel volume, but are, at least in part, organized in lamellar domains, which may consist of stacked single lamellae. The broad major T2 peak at 25 °C and its splitting into two separated peaks at 5 °C gives evidence that these lamellar domains are in the micrometer range. Lillford et al.36 suggested that multiexponential relaxation will be observed, if there are heterogeneities within the sample over spatial regions larger than the region sampled by one water molecule during its intrinsic relaxation time. This was confirmed by Belton et al.,42 who provided the theoretical framework to explain the change from monoexponential to biexponential relaxation for agarose gels that were freeze-thawed. They were able to show that heterogeneities on a distance scale of about 110-140 µm are necessary to obtain multiexponential relaxation behavior at room temperature. The authors showed that the mean diffusion distance (dm) of 110 µm, as the limit for diffusional averaging of the relaxation, can be easily derived from the mean diffusion coefficient (Dm). It is calculated as the average distance that a freely diffusing water molecule travels during the time t equal to T2m:

dm2 ) 6Dmt ) 6DmT2m

(3)

The broad major T2 peak at room temperature and its splitting into two peaks at 5 °C may therefore be explained as follows: the relaxation of water protons in the hydrogel formed by 30 mg/mL Me2PE-C32-Me2PE is strongly influenced by the presence of the extended lamellar domains. Besides a fraction of more or less unperturbed water, a considerable number of water molecules will be present inside these domains, most probably in the thin layers of water between two adjacent lamellae. The dwelling time of a particular water molecule inside such a cluster of lamellae will be increased compared to the dwelling time at

Water Dynamics in Bolaamphiphile Hydrogels the fiber surface, since the diffusion distance to the fraction of unperturbed water is now longer. As a result, the motional perturbation exerted by the surface of the lamellae on the water molecules is considerably increased, leading to a relaxation rate enhancement. T2m as well as Dm are temperature dependent, which results in significant differences in the T2 distributions obtained at different temperatures. For 30 mg/mL Me2PE-C32Me2PE at 25 and 30 °C, for example, the mean diffusion coefficient of the water molecules is large enough to enable a sufficiently fast exchange between the populations of perturbed and unperturbed water, which results in a broad, but monomodal major T2 peak, which is the weighted sum of all relaxation times present in the sample. Due to the dynamical perturbation exerted by the presence of the lamellar aggregates, the obtained T2m value for this sample with 30 mg/mL Me2PE-C32-Me2PE at pH 5 was found to be considerably decreased with respect to that determined for the transparent hydrogel formed by 30 mg/ mL Me2PE-C28-Me2PE at pH 5, which contains only nanofibers. At 30 °C, the determined values of the mean diffusion coefficient and the mean transverse relaxation time are about 2.5 × 10-9 m2/s and 1.3 s, respectively. Using eq 3, the mean diffusion distance dm for a freely diffusing water molecule is calculated to be about 1.4 × 10-4 m or 140 µm, which is in agreement with the results of Belton and co-workers.42 For 25 °C, no mean diffusion coefficient was determined, but it can be estimated to be approximately 2.1 × 10-9 m2/s. With the T2m value of 1.11 s, dm is calculated to be 120 µm. Therefore, the distance scale of heterogeneities that are present in a freshly prepared hydrogel of 30 mg/mL Me2PE-C32-Me2PE at pH 5 is below 120-140 µm. As shown in Figure 10c, a bimodal major T2 peak is obtained at 5 °C for the same sample due to a substantially decreased mean diffusion coefficient. The value of Dm at 5 °C can be estimated to be approximately 10-9 m2/s. If one uses the lower T2m value of 0.42 s from the bimodal major T2 peak, the mean diffusion distance at 5 °C is only 50 µm. One can therefore conclude that the distance scale of the heterogeneities present in the freshly prepared hydrogel of 30 mg/mL Me2PE-C32-Me2PE at pH 5 lies in the range between 50 and 120-140 µm. Upon storage, however, syneresis occurs, which results in a bimodal T2 peak at 25 and 30 °C. The lamellar stacks have obviously become larger in their dimensions than 140 µm. At 40 °C, however, a monomodal T2 peak was observed for the sample after syneresis. This is due to the increased diffusion coefficient of water at this temperature (Dm ) 3.1 × 10-9 m2/ s), which enables the water molecules to exchange between their sites in the heterogeneous system. Comparison of the diffusion coefficient with the T2m value of 1.95s at 40 °C indicates that the distance scale of the heterogeneities does not exceed 190 µm. Therefore, the observation of syneresis for the hydrogel of 30 mg/mL Me2PE-C32-Me2PE at pH 5 can be explained by a significant increase in the dimension of the domains, which are formed by stacks of lamellae. These lamellar domains seem to reach average sizes in the range between 140 and 190 µm after storage for 2 days. Summary The dynamics of the water molecules in hydrogels formed by the bolaamphiphiles PC-C32-PC and Me2PE-C32-Me2PE at concentrations of 1 and 30 mg/mL were investigated by determining the relaxation behavior and the mean diffusion coefficients of the water protons. The dynamical properties of the water molecules in PC-C32-PC hydrogels were found to be almost unchanged compared to pure H2O. Only for the sample

J. Phys. Chem. B, Vol. 115, No. 1, 2011 21 containing 30 mg/mL PC-C32-PC, a small dynamical perturbation could be detected by slightly decreased values for the mean transverse relaxation time constant T2m and the mean diffusion coefficient Dm. This perturbation was shown to be independent of the type of aggregates and did therefore not respond to the temperature induced change from a hydrogel formed by nanofibers to a solution of micellar aggregates at higher temperatures. For Me2PE-C32-Me2PE at pH 5 and a concentration of 1 mg/ mL, again, no significant changes in the T2 distribution at 25 °C are observed compared to pure buffer due to the low bolaamphiphile content. For a sample with 30 mg/mL Me2PEC32-Me2PE, the mean diffusion coefficient of the water molecules was similar to that obtained for 30 mg/mL PC-C32PC, but the relaxation behavior was found to be considerably changed. This change in the relaxation behavior can be attributed to a considerable degree of heterogeneity of structures in the Me2PE-C32-Me2PE hydrogel, which is most probably caused by the formation of domains of stacked lamellar aggregates. In previous studies, only lower concentrations of Me2PE-C32Me2PE or considerably smaller sample volumes were used. Therefore, it is the first time that the transformation of a considerable number of nanofibers into lamellar aggregates and the phenomenon of syneresis is observed for Me2PE-C32-Me2PE at pH 5. At 25 °C, a broad monomodal major T2 peak was obtained, which became split when the temperature was decreased to 5 °C. By comparing the values determined for T2m and Dm at the particular temperature, these findings allowed to estimate the size range of the lamellar domains. These domains can be expected to reach dimensions larger than 50 µm and smaller than 120-140 µm. When the sample is heated, these lamellar domains become disintegrated during the first transition that was observed before by DSC. As a result, the relaxation behavior of the water protons was found to be considerably changed above this transition temperature and is reflected in the obtained T2m values and relaxation rate enhancement factors. Moreover, the hydrogel containing 30 mg/mL Me2PE-C32-Me2PE at pH 5 was found not to be stable upon storage, but to show the phenomenon of syneresis due to a significant growth of the lamellar domains. By analysis of the NMR relaxation data, the lamellar domains could be shown to reach sizes between 140 and 190 µm for the sample after storage for two days. It could therefore be demonstrated that 1H NMR relaxometry and diffusometry can provide useful information about the dynamical properties of the water molecules and can be used to reveal the presence of heterogeneities in the hydrogel samples. For further investigations of the water dynamics of the bolaamphiphilic hydrogels, sophisticated methods have to be applied. Whereas NMR relaxation probes a time window of a few milliseconds to seconds, the relation of water dynamics and structural features of hydrogel systems can be studied on the time scale of hundreds of picoseconds in a spatial window of a few ångstroms by means of incoherent quasielastic neutron scattering (QENS) as was shown by Paradossi and co-workers for other heterogeneous systems like Sephadex hydrogels.43,44 This method can also be used in combination with molecular dynamics (MD) simulations in order to yield a view of the investigated system on a molecular level.45 Acknowledgment. This work was supported by grants from the Cluster of Excellence “Nanostructured Materials” of the state of Saxony-Anhalt (M.B., B.D., K.M., and A.B.) and from the Deutsche Forschungsgemeinschaft (A.M., S.D., B.D., and A.B.).

22

J. Phys. Chem. B, Vol. 115, No. 1, 2011

Supporting Information Available: Parameters for the field-gradient stimulated-echo (PGSTE) method, T2 distributions, and T2 peaks. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Peppas, N. A.; Bures, P.; Leobandung, W.; Ichikawa, H. Eur. J. Pharm. Biopharm. 2000, 50, 27–46. (2) McConville, P.; Whittaker, M. K.; Pope, J. M. Macromolecules 2002, 35, 6961–6969. (3) Jeong, B.; Kim, S. W.; Bae, Y. H. AdV. Drug DeliVery ReV. 2002, 54, 37–51. (4) Baroli, B. J. Pharm. Sci. 2007, 96, 2197–2223. (5) Nayak, S.; Lyon, L. A. Angew. Chem. Int. Ed. Engl. 2005, 44, 7686– 7708. (6) de Loos, M.; Feringa, B. L.; van Esch, J. H. Eur. J. Org. Chem. 2005, 3615–3631. (7) Estroff, L. A.; Hamilton, A. D. Chem. ReV. 2004, 104, 1201–1217. (8) Koehler, K.; Foerster, G.; Hauser, A.; Dobner, B.; Heiser, U. F.; Ziethe, F.; Richter, W.; Steiniger, F.; Drechsler, M.; Stettin, H.; Blume, A. Angew. Chem. Int. Ed. Engl. 2004, 43, 245–247. (9) Koehler, K.; Foerster, G.; Hauser, A.; Dobner, B.; Heiser, U. F.; Ziethe, F.; Richter, W.; Steiniger, F.; Drechsler, M.; Stettin, H.; Blume, A. J. Am. Chem. Soc. 2004, 126, 16804–16813. (10) Koehler, K.; Meister, A.; Foerster, G.; Dobner, B.; Drescher, S.; Ziethe, F.; Richter, W.; Steiniger, F.; Drechsler, M.; Hause, G.; Blume, A. Soft Matter 2006, 2, 77–86. (11) Drescher, S.; Meister, A.; Blume, A.; Karlsson, G.; Almgren, M.; Dobner, B. Chem.sEur.J. 2007, 13, 5300–5307. (12) Meister, A.; Bastrop, M.; Koschoreck, S.; Garamus, V. M.; Sinemus, T.; Hempel, G.; Drescher, S.; Dobner, B.; Richtering, W.; Huber, K.; Blume, A. Langmuir 2007, 23, 7715–7723. (13) Meister, A.; Drescher, S.; Garamus, V. M.; Karlsson, G.; Graf, G.; Dobner, B.; Blume, A. Langmuir 2008, 24, 6238–6246. (14) Hoffman, A. S. AdV. Drug DeliVery ReV. 2002, 43, 3–12. (15) Hills, B. P.; Manning, C. E. J. Mol. Liq. 1998, 75, 61–76. (16) Carr, H. Y.; Purcell, E. M. Phys. ReV. 1954, 94, 630–638. (17) Meiboom, S.; Gill, D. ReV. Sci. Instrum. 1958, 29, 688–691. (18) Stejskal, E. O.; Tanner, J. E. J. Chem. Phys. 1965, 42, 288–292. (19) Tanner, J. E. J. Chem. Phys. 1970, 62, 2523. (20) Provencher, S. W. Comput. Phys. Commun. 1982, 27. (21) Hills, B. P.; Snaar, J. E. M. Mol. Phys. 1995, 84, 141–157. (22) Chiotelli, E.; Pilosio, G.; Le Meste, M. Biopolymers 2002, 63, 41– 58.

Bastrop et al. (23) Tang, H. R.; Brun, A.; Hills, B. Carbohydr. Polym. 2001, 46, 7– 18. (24) Mitchell, J.; Stark, S. C.; Strange, J. H. J. Phys. D: Appl.Phys. 2005, 38, 1950–1958. (25) Schaumann, G. E.; Hobley, E.; Hurrass, J.; Rotard, W. Plant Soil 2005, 275, 1–20. (26) Li, R.; Kerr, W. L.; Toledo, R. T.; Carpentier, J. A. J. Food Sci. 2000, 65, 575–580. (27) Roorda, W. E.; de Bleyser, J.; Junginger, H. E.; Leyte, J. C. Biomaterials 1990, 11, 17–23. (28) Roorda, W. E.; Bouwstra, J. A.; de Vries, M. A.; Junginger, H. E. Biomaterials 1988, 9, 494–499. (29) Roorda, W. E.; Bouwstra, J. A.; De Vries, M. A.; Junginger, H. E. Pharm. Res. 1988, 5, 722–725. (30) Lee, H. B.; Jhon, M. S.; Andrade, J. D. J. Colloid Interface Sci. 1975, 51, 225–231. (31) Denisov, V. P.; Halle, B. J. Mol. Biol. 1995, 245, 682–697. (32) Møller, K. B.; Rey, R.; Masia, M.; Hynes, J. T. J. Chem. Phys. 2005, 122 (11), 1863172. (33) Loeffler, H. H.; Inada, Y.; Funahashi, S. J. Phys. Chem. B 2006, 110, 5690–6. (34) Duval, F. P.; Porion, P.; Fauge`re, A.-M.; Van Damme, H. J. Colloid Interface Sci. 2001, 242, 819–826. (35) Penke, B.; Kinsey, S.; Gibbs, S. J.; Moerland, T. S.; Locke, B. R. J. Magn. Reson. 1998, 132, 240–254. (36) Lillford, P. J.; Clark, A. H.; Jones, D. V. ACS Symp. Ser. 1980, 127, 177–195. (37) Bastrop, M.; Meister, A.; Metz, H.; Drescher, S.; Dobner, B.; Maeder, K.; Blume, A. J. Phys. Chem. B 2009, 113, 574–582. (38) Lewis, G. P.; Derbyshire, W.; Ablett, S.; Lillford, P. J.; Norton, I. T. Carbohydr. Res. 1987, 160, 397–410. (39) Tanaka, H.; Fukumori, K.; Nishi, T. J. Chem. Phys. 1988, 89, 3363– 3372. (40) Tanaka, N.; Matsukawa, S.; Kurosu, H.; Ando, I. Polymer 1998, 39, 4703–4706. (41) Meister, A.; Drescher, S.; Karlsson, G.; Hause, G.; Baumeister, U.; Hempel, G.; Garamus, V. M.; Dobner, B.; Blume, A. Soft Matter 2010, 6, 1317–1324. (42) Belton, P. S.; Hills, B. P.; Raimbaud, E. R. Mol. Phys. 1988, 63, 825–842. (43) Paradossi, G.; Cavalieri, F.; Chiessi, E. Carbohydr. Res. 2005, 340, 921–927. (44) Paradossi, G.; Cavalieri, F.; Chiessi, E.; Mondelli, C.; Telling, M. T. F. Chem. Phys. 2004, 302, 143–148. (45) Chiessi, E.; Cavalieri, F.; Paradossi, G. J. Phys. Chem. B 2007, 111, 2820–2827.

JP107755K