Introduction

The molecular structure of polysaccharides has led to the belief that these polymers are stiff. The characteristic ratio (C∞) for polysaccharides is rather high, although reported C∞ values are limited to typical polysaccharides such as cellulose1 and gellan.2 If we can prepare concentrated solutions of polysaccharides, we can determine the molecular weights between entanglements in the molten state (Me,melt) for these polysaccharides,3, 4, 5 which reflects the chain properties, that is, the flexibility of the chain. Since the first report that cellulose has good solubility in ionic liquids,6 many studies have been conducted to determine the solution properties of polysaccharides at high concentrations. We prepared concentrated solutions of several kinds of polysaccharides in an ionic liquid 1-butyl-3-methylimidazolium chloride (BmimCl) and carried out rheological measurements for the solutions to determine Me,melt for the polysaccharides.7, 8, 9 The experiments yielded rather small values of Me,melt for the polysaccharides; for example,8 Me,melt was 2.3 × 103 for agarose, which is composed of (1,3)-β-D-galactopyranose and (1,4)-3,6-anhydro-α-L-galactopyranose.10 This value appears to be much smaller than expected for a stiff polymer and is actually comparable to the Me,melt for the simplest flexible polymer polyethylene (Me,melt=1.3 × 103),3 although no reported value of C∞ for agarose is available. Why the Me,melt for polysaccharides is so small is not clear at present. The Me,melt for amylose was also determined and was much larger than that for cellulose.7 To explain the Me,melt values, it is necessary to examine whether the use of an ionic liquid as a solvent affects the estimation of Me,melt. It has been proposed that ionic liquids form three-dimensional structures even in the liquid state because of cation–anion hydrogen bonding and cation–cation ring stacking.11, 12, 13 The three-dimensional structure may contribute to the modulus of the solutions; if the elasticity due to the network of solvent molecules contributes to the plateau modulus (GN0), then Me,melt is reduced. The aim of this study is to clarify whether or not the network formed by the ionic liquid molecules contributes to GN0. Concentrated solutions of agarose were prepared by using formamide (FA), N-methylformamide (MFA) and BmimCl as solvents. Angular frequency (ω) dependence curves of the storage and loss moduli (G′ and G″, respectively) were compared among these solutions.

Experimental procedure

Agarose (Research Organics, Cleveland, OH, USA) was used without further purification. FA (Wako, Osaka, Japan) and MFA (Aldrich, St Louis, MO, USA) were used as received. The ionic liquid BmimCl (Aldrich) was the same as that described in the previous studies.7, 8 Agarose solutions were prepared as follows: agarose was added to each solvent (for BmimCl, the solvent was preheated to convert it into the liquid state) in a dry glass vessel, and the mixture was quickly stirred with a stainless spatula on a hot plate at ∼80 °C. Then, the vessel was sealed and left on the hot plate for complete melting. The concentration of agarose (c) was 1.0 × 102 or 2.0 × 102 kg m−3, that is, ca 10 or 20 wt %. In the calculation of c, 1.0 × 103 kg m−3 was assumed for the density of agarose, and 1.14 × 103, 1.00 × 103 and 1.08 × 103 kg m−3 were used for the densities of FA,14 MFA14 and BmimCl,7, 8 respectively. To minimize the effect of moisture absorption, a fresh bottle of solvent was used every time agarose samples were prepared, and the rheological measurements were made just after finishing the sample preparation.

Rheological measurements (dynamic viscoelasticity measurements in this study) were carried out with an ARES rheometer (now TA Instruments, New Castle, DE, USA) using the dynamic operation-testing mode under a nitrogen atmosphere. A cone–plate geometry with a diameter of 25 mm and a cone angle of 0.1 rad was used. The ω-dependence of G′ and G″ for the solutions was measured with a strain amplitude (γ) of 0.1. Because the melting temperatures (Tm) of FA14 and MFA14 are reported to be 2.6 and −5.4 °C, respectively, the temperature (T) ranged from 5 to 40 °C for the FA solution and from 0 to 30 °C for the MFA solution. The T-range for the BmimCl solution was 20–120 °C. For BmimCl, Tm is ca 70 °C, but the supercooled state of the BmimCl solutions was rather stable. Thus, viscoelasticity measurements were successfully taken even at 20 °C. The master curves in each solvent were prepared using only a horizontal shift with the shift factor aT.

Results and Discussion

In Figure 1a the master curves of the ω dependence of the dynamic viscoelastic functions (G′ and G″) for the BmimCl and FA solutions of agarose at c=100 kg m−3. The time–temperature superposition principle held very well for the G′ and G″ curves for these solutions and for the other solutions examined in this study, as will be discussed below. The reference temperature (Tr) in the figure was set to 80 °C for both solutions. For the FA solution, the dynamic viscoelasticity measurements were taken in the T-range of 5–20 °C. The master curves were prepared initially at 20 °C, and then the master curves were shifted so that Tr=80 °C using the extrapolated values of aT. In the master curves for BmimCl solution (BmimCl curves), the rubbery plateau and the terminal (flow) zones are clearly observed; the rubbery plateau is designated as the region of G′ at log ωaT of 2.2–4.7. Here, the plateau modulus GN0 can be defined as the G′ value at ωaT where the loss tangent (tanδ=G″/G′) curve attains a minimum7, 8, 9, 15 and is related to the molecular weight between entanglements (Me) by

where R is the gas constant. It should be noted that Me in the above equation should be in kg mol−1. For the BmimCl solution at c=100 kg m−3, we have a GN0 of 1.2 × 104 Pa for the BmimCl solution, which gives an Me of 2.4 × 104. This value is consistent with the Me values obtained in a previous study.8 In the master curves for the FA solution (FA curves), the flow zone is clearly observed. Concerning the rubbery plateau, however, only an onset of the plateau, that is, the intersection of the G′ and G″ curves, emerges at the high ωaT limit, and the height of the intersection is identical to that on the BmimCl curves. Figure 1b shows the same data as Figure 1a, but here the FA curves are horizontally shifted to have the best overlap with the BmimCl curves. The extent to which the curves were shifted is expressed by the parameter A, and log A=−1.14 for the FA curves; for the BmimCl curves, log A is taken to be 0. As can be seen from the figure, the horizontal shift was applied successfully. Our results suggest that the FA solution at c=100 kg m−3 has the same number of entanglements as the BmimCl solution, although the high melting temperature of FA prevents measurements at low temperatures, and thus we were not able to obtain direct evidence of the dynamic viscoelasticity. The number (or more precisely, the number density) of entanglements appears to be independent of the solvent species as long as the solutions are homogeneous, as is well established for the synthetic polymers. Focusing on the properties of the ionic liquid BmimCl, it is well known that BmimCl forms a three-dimensional structure, although the structure is not very strong. The lack of enhancement of GN0 in BmimCl compared with that in FA indicates that the lifetime of the BmimCl structure is sufficiently short, and therefore, the formation of the structure does not contribute to the modulus in the rubbery and terminal zones. In other words, Me in the BmimCl solution depends purely on the chain–chain entanglements.

Figure 1
figure 1

(a) Master curves of the ω-dependence of G′ and G″ for the 1-butyl-3-methylimidazolium chloride (BmimCl) and formamide (FA) solutions of agarose at c=100 kg m−3. The reference temperature, Tr, is 80 °C. (b) The FA curves are horizontally shifted by the parameter A.

Full size image

The parameter A defined previously is related to the ratio of the zero-shear viscosity (η0) for the FA solution to that for the BmimCl solution, because A is a type of horizontal shift factor and therefore has the same meaning as aT. Although we do not have detailed data on η0, pure BmimCl and pure FA are expected to have very different η0 values. However, at c=100 kg m−3, the difference becomes rather small, only approximately one order of magnitude, because log A=−1.14.

The plots of aT versus T−1 for the BmimCl and FA solutions of agarose at c=100 kg m−3 are shown in Figure 2. It should be noted that aT here is the shift factor from Tr (80 °C) even for the FA solutions; namely, the original shift factor was modified so that at 80 °C log aT=0. The data points for the FA solution fall on a single line, indicating that the T-dependence of aT can be well described by an Arrhenius-type equation. The plots of log aT versus T−1 for the BmimCl solution can also be approximated by a single line.

Figure 2
figure 2

Shift factor aT plotted against T−1 for the 1-butyl-3-methylimidazolium chloride (BmimCl) and formamide (FA) solutions of agarose at c=100 kg m−3. The reference temperature, Tr, is 80 °C.

Full size image

Figure 3 shows the master curves of G′ and G″ of agarose in BmimCl, FA and MFA at c=200 kg m−3. The master curves for BmimCl cover a wide range of ωaT compared with the others because of the wider T-range of the measurements. For the plateau zone, we have a GN0 of 5.2 × 104 Pa, leading to an Me of 1.1 × 104 at c=200 kg m−3. This Me value is roughly half of that (2.4 × 104) for the solution at c=100 kg m−3. The master curves for the FA and MFA solutions (FA and MFA curves, respectively) overlapped well with each other, and the flow zone and the onset of the plateau zone are observed for these curves, as in the case of the FA curves at c=100 kg m−3. In addition, the intersections of G′ and G″ for the FA and MFA curves are identical in height to that of the BmimCl curve, suggesting that GN0 is the same for these solutions, and therefore, it is expected that all three solutions will have the same value of Me. Again, Me is independent of the solvent species.

Figure 3
figure 3

Master curves of the ω-dependence of G′ and G″ for the 1-butyl-3-methylimidazolium chloride (BmimCl), formamide (FA) and N-methylformamide (MFA) solutions of agarose at c=200 kg m−3. The reference temperature, Tr, is 80 °C.

Full size image

The plots of aT versus T−1 are shown in Figure 4. Here, aT is the shift factor from Tr (80 °C) for all solutions; namely, the original shift factor was modified so that at 80 °C, log aT=0 for the FA and MFA solutions. It is interesting that the data points for the FA and MFA solutions fall on a single line, indicating that the T-dependence of aT can be well described by the same Arrhenius-type equation.

Figure 4
figure 4

Shift factor aT plotted against T−1 for the 1-butyl-3-methylimidazolium chloride (BmimCl), formamide (FA) and N-methylformamide (MFA) solutions of agarose at c=200 kg m−3. The reference temperature, Tr, is 80 °C.

Full size image

Conclusion

The rheological properties of concentrated solutions of agarose having two different concentrations were compared. Two organic solvents FA and MFA, and an ionic liquid BmimCl were used as the solvents. The ω-dependence curves of G′ and G″ for these solvents overlapped well only after a horizontal shift was introduced. The good agreement in the curves indicates that the entanglement properties of the agarose chains are similar regardless of the solvent. The small Me,melt for polysaccharides can be attributed to the network of polysaccharide chains, not to the network formed by the ionic liquid molecules.