By the steady shear test, the zero shear viscosities of the solutions were characterized and shown in Fig. 1(a). The inserted images show the polarized optical micrographs of the solutions. The concentration range of the solutions can be divided into four regions; non-contact region (3 ~ 9 wt%), entanglement region (9 ~ 14 wt%), anisotropic region (14 ~ 16 wt%), and gel region (16 ~ 18 wt%). The zero-shear viscosities changed significantly with the region.
In the non-contact region, the zero-shear viscosities slowly increased as the concentration increased. The cellulose chains with low concentration have little chance to contact with each other due to the small number of molecules in unit volume. In the entanglement region, the zero-shear viscosities rapidly increased with the concentration increase. The cellulose chains started to contact and entangle each other, resulting in large viscosity increase. As the concentration increased from 14 to 16 wt% (anisotropic region), the zero-shear viscosities decreased mainly because of the phase transition to LC. It is well-known that cellulose forms LC phase over the critical concentration. The cellulose chains become rigid by the intramolecular hydrogen bonds. The rigid chains were easily aligned and showed the cholesteric LC transition (Luo et al. 2014; Song et al. 2011; Song et al. 2010). In turns, the interchain contacts decreased in aligned chain conformation. It resulted in the viscosity drop at 16 wt%. With the further increase of the concentration, the zero-shear viscosity rapidly increased again in the gel region. In the highly concentrated solution, cellulose chains were physically networked by the intermolecular hydrogen bonds and gave in the organic gel. The 3-D physical network structure needed more intensive shear to deform. As a result, the viscosity rapidly increased in the gel region.
The polarized optical micrographs confirm the phase transition of the solution along the concentration. (See Fig. 1(a) and Figure S1). In the non-contact and the entanglement regions (≤14 wt%), the solution showed the fully isotropic phase regardless of the concentration. In those regions, the cellulose concentrations were too low to form an anisotropic phase. As the concentration increased, the anisotropic phase was observed in the concentration range from 14 to 16 wt%. As described above, cellulose chains were aligned in the anisotropic region. The aligned chains formed the LC phase and showed anisotropy in the polarized optical micrographs. With the further increase of concentration (the gel region), the solution showed a fully isotropic phase again. The cellulose chains formed a strong 3-D physical network by the gelation. Due to the 3-D chain network, the chain movements were seriously restricted and lost a chance to be aligned. In other words, the chains were randomly oriented by the strong intermolecular hydrogen bonds.
Figure 1(b) shows the dynamic viscoelastic behavior of the solutions in the frequency range of 10− 1 to 102 rad/s at 30 ℃. As well known, the crossover (ωc) of the storage (G') and loss (G") moduli have strong dependency on the physical states such as the LC and gel (FitzGerald et al 2013; Huang et al. 2018). In the non-contact region (3 ~ 9 wt%), ωc was not observed in the measured range because of the low chain-to-chain interaction as explained in Fig. 1(a). Above the entanglement region (9 wt%≤), ωc began to appear. The appearance of ωc in the measured range means that the interaction between the chains was drastically increased (Ahn et al. 2016; Kim et al. 2018). As the concentration increased up to 14 wt%, the cellulose chains can have more chance to entangle by the close intermolecular distances. In the anisotropic region (LC phase) from 14 to 16 wt%, ωc was shifted to higher frequency (See Fig. 1(c)). The increase of ωc at 16 wt% was originated from the LC formation. In the LC phase, the aligned cellulose chains had less chance to contact each other, resulting in fast relaxation under the dynamic test. Although the zero-shear viscosity of 16 wt% solution was similar to the isotropic solution, G' and G" were crossed in the measured range. It indicates that the chains were physically networked and gelled even in the LC phase. In other words, it can be said that the intermediate physical state of gel and LC existed in 16 wt% solution. In the gel region (16 ~ 18 wt%), ωc rapidly decreased again. The solution was fully gelled and then the relaxation was significantly hindered.
The entanglement molecular weight (Me) was calculated using Eq. (1) and shown in Fig. 1(c). The measured values are summarized in Table 1. For the calculation, the plateau modulus (G0N) was obtained as the storage modulus having the lowest tangent delta in the rubbery plateau region (See Figure S3) (Liu et al 2006). As similar to ωc, the entanglement molecular weight decreased as the concentration increased up to 14 wt%. In general, polymers in concentrated solution have many entanglements due to the increase of chain-to-chain contacts. It results in a lower entanglement molecular weight. In other words, the degree of entanglement is inversely proportional to the entanglement molecular weight (Huang et al. 2013; Wool 1993). As the concentration of cellulose increased up to 14 wt%, the entanglements increased which led to an increase of the contacts between neighboring chains. Whereas Me was decreased in the anisotropic region (14 ~ 16 wt%). Cellulose chains in the anisotropic region maintained the aligned state during the dynamic test. It led to the increase of Me. Me decreased again by the gelation with the concentration increase (18 wt%). As the cellulose chains were gelled, the contact distance between the chains drastically decreased by the formation of physical network.
(1)
Table 1
The plateau modulus and entanglement molecular weight of solutions for the different concentrations
Concentration
(wt%)
|
G0N
(103ᆞPa)
|
Me
(103 gᆞmol− 1)
|
3
|
-
|
-
|
5
|
-
|
-
|
7
|
-
|
-
|
9
|
15.27
|
115.14
|
12
|
21.34
|
70.86
|
14
|
39.32
|
38.46
|
15
|
38.27
|
39.51
|
16
|
33.22
|
45.52
|
17
|
34.53
|
43.79
|
18
|
45.67
|
33.11
|
The fluidity and gelation of the solution were confirmed by the classical “vial upside-down” method (See Fig. 2 (a)). The solution with low concentration (9 wt%) flowed down immediately after the upside-down of the vial, while the solutions with 12 and 14 wt% flowed down very slowly. After 12 hours, most parts of the solutions moved to the bottom. However, the solutions with high concentrations (16 and 18 wt%) did not flow and remained on the top side of the bottle even after 12 hours. It indicates that the solutions with high concentration were gelled by the strong physical network between the chains. In the high concentrations, the intermolecular hydrogen bonds and the chain entanglements were dramatically increased, resulting in the gelation of the solutions. It can be a solid evidence that the intermediate state of gel and LC was formed in 16 wt% solution.
Figure 2(b) shows the time-dependent stress sweep results. The test results of each solution were shown in Figure S4(a). The time-dependent stress sweep was performed to study the change of the physical states and the interaction of the chains along time. During the test, the stress was recorded for 300 s at the fixed shear rate of 0.1 s-1. After that, the shear deformation was stopped, and the solution was kept for the designated rest time (0, 60, and 360 s). The ‘60 s’ means that the shear was re-applied after a temporal pause for 60 s. After the rest time, the shear stress was recorded again under the same condition of the initial run. The shear stress was normalized by the relaxed stress (τ∞). τ∞ means the shear stress when the solution was fully relaxed (See Figure S4(b)). In the initial run, the solutions with the concentration under 14 wt% reached steady state right after a small stress overshoot around 60 s. The stress overshoot was originated from the chain disentanglement by the applied shear. When the concentration was higher than 16 wt%, the overshoot dramatically increased. It means that large stress was required to break the strong gel network. Comparing with another anisotropic solution of 14 wt%, the solution of 16 wt% showed more intensive overshoot. It is again resulted from the anisotropic gel. The concentration of 14 wt% was not high enough for physical networking. The second sweep run started right after the first pause without resting. As shown in Figure 2(b), the overshoot was not observed for the solutions with the concentrations from 5 to 16 wt%. It is because the chain had no time to recover the intermolecular interaction destructed during the initial run. However, the stress overshoot was still observed at 18 wt% solution in the second run. This overshoot was originated from the unbroken chain network rather than the spontaneous recovery of the network. It showed that the chain network in the gel state was not completely destroyed by the applied shear due to the strong hydrogen bonds and the entanglement. The third run was carried out after 6 minutes of resting following the second run. The overshoot was also strongly dependent on the concentration. When the solution concentration was lower than 12 wt%, the size of each overshoot was similar to that of the initial run. It indicates that the intermolecular entanglement was almost recovered by the chain relaxation. In 14 and 16 wt% solutions, the overshoot during the third run did not appear due to the stable chain alignment of the LC phase. The cholesteric phases of the anisotropic solutions (14 and 16 wt%) were changed to the nematic structure by the shear deformation in the first and second runs. In the nematic structure, 3-D chain network formation is limited by the lack of physical interaction between the chains. As a result, the complete recovery of the stress overshoot was not observed at the third run. It means that the nematic structure derived by shear deformation can remain stable in the resting time. This behavior is a key point to produce regenerated films with high chain orientation. Especially, 16 wt% solution showed a strong stress overshoot at the initial run and the slow recovery which were originated from the intermediate state of gel and LC. In 18 wt% solution, the overshoot was recovered completely. The randomly oriented chains in 18 wt% solution can form the 3-D network again. The network structure spontaneously recovered by the chain relaxation resulting in complete recovery of the stress overshoot in the third run.
Figure 3(a) shows the strain amplitude sweep results of the cellulose solution under large dynamic strain conditions. The test results of all solutions are shown in Figure S5. Hyun et al grouped the LAOS (large amplitude oscillatory shear) behavior of polymer solution or melts into four types; type I (strain thinning), type II (strain hardening), type III (weak strain overshoot), and type IV (strong strain overshoot) (Hyun et al. 2002). The prepared solutions in this study belong to the type I LAOS behavior (Hyun et al. 2011). Type I behavior refers to the strain thinning behavior of typical linear polymer solution or melt. As clearly seen in the Fig. 3(a), all the curves can be divided into two regions; one is steady state region in low amplitude, and the other is strain thinning region in high amplitude. In the steady state region, both G' and G" were independent of the amplitude, indicating that the physical state of cellulose in the solution was not changed by the deformation. In the strain thinning region, the solution became more fluid-like with the amplitude increase due to the disentanglement and the alignment of cellulose chains. The entanglements and gel network of cellulose chains were destroyed due to the large deformation. The moduli were also relied on the concentration. As similar to the zero-shear viscosity, the solution of 16 wt% showed the smallest values in both moduli (G' and G") among the concentrated solutions (14 ~ 18 wt%), mainly because of the phase transition from isotropic to anisotropic phase. The increases in G' and G" were originated from the interchain contacts and entanglements. As the concentration increased, the physical interaction between the chains increased which led to the increase in the moduli. In the anisotropic solutions (14 and 16 wt%), the aligned chains could easily slip between them. The slippage between the chains resulted in the decrease in the moduli of the solutions.
Figure 3(b) shows the tangent delta (tan δ) of the solution under the LAOS condition (presented as a black dotted line). The tan δ decreased as the concentration increased up to 14 wt%. It is because the solution became more elastic with the concentration increase. As the concentration increased, the physical interaction between the chains made the solution more solid-likely. The further increase to 16 wt% resulted in the decrease of tan δ. It also indicates the alignment of the cellulose chain. The gelation at 18 wt% decreased the tan δ again.
The yield stress (τy) was calculated and displayed in Fig. 3(b) as a red dotted line. The yield stress was calculated using Eq. (2). G'0 means the storage modulus at zero strain and γc (critical strain) means the strain to break the chain network of the solution (See Figure S6).
(2)
The yield stress increased monotonously with the concentration. The initial increase was resulted from the entanglement and the later increase was resulted from the gel networking. In the low concentration, the chain entanglement increased the elastic properties of the solution. As the solutions were gelled, the yield stress of the solutions was increased. The yield stress of a polymer solution refers to the strength of the chain network (Chejara et al 2013; Lu and Hsieh 2010). As illustrated in Fig. 3(a), the moduli (especially G'0) at 16 wt% decreased as the liquid crystal was formed. Likewise, γc was expected to decrease by LC transition at 16 wt%. However, the chains were strongly networked in the gel at 16 wt%, which led the increase in γc. The chain network was not destructed by the large shear deformation. As a result, the yield stress increased as the solution was gelled at 16 wt%. With the rheological observation under the LAOS condition, it was clearly revealed that the presence of the intermediate state of the gel and LC.
Figure 4(a) shows the digital image of the film regenerated from the 16 wt% cellulose/AmimCl solution. The transparent films were successfully prepared using the solutions with the concentration range of 9 to 18 wt% (See Figure S2) However, the film was not fabricated from the dilute solution (≤ 7 wt%) due to the low chain entanglement. In the film process, the prepared cellulose solution was unidirectionally spread on a glass substrate. The chain orientation in the film was estimated using 2D-WAXS. The in-plane orientation was strongly developed from the anisotropic solution (See Figs. 4 (b) and (c), and Figure S7.). Figures 4(d), (e), and (f) show the polarized microscope images of the film along the polarized angle. The magenta-color was obtained by tint plate insertion. The images of 0° and 90° tilted against the polarizing direction show a similar magenta color, indicating that the polarized lights did not pass through the film. Whereas the 45° image shows the fully strong bright green color. It reveals that the film was uniaxial chain orientation in the film originated from the LC phase in the solution.
Figure 4 (g) shows the stress-strain curves of the films, of which important parameters were summarized in Table 2. The initial modulus and the tensile strength of the film were increased with the increase of the concentration up to 16 wt% and then decreased with the further increase in concentration. The tensile strength depends on both chain orientation and the interactions, while the initial modulus mainly on the orientation. In the low concentration (≤ 12 wt%), both modulus and strength were increased with the concentration by the chain-to-chain entanglement. The phase transition to LC increased the initial modulus dramatically. The cholesteric phase in the solution was changed into the nematic phase during the film casting, resulting in uniaxial chain orientation. The uniaxially aligned structure was not easily deformed by the applied stress comparing the randomly oriented structure. The tensile strength was drastically improved when the concentration increased from 14 to 16 wt%. The solution of 16 wt% had a robustly networked structure in aligned chains as described above. This intermediate phase gave a better chance to redistribute the applied stress effectively. The aligned network can maintain the structure without significant destruction even under the large deformation. At 18 wt%, both the tensile strength and initial modulus decreased rapidly. Meanwhile, the toughness was continuously increased in the concentration range in this study. The increase of toughness up to 16 wt% was originated from the entanglement, the chain orientation, and the robust network, respectively. Although the chain in 18 wt% solution did not have preferred alignment, it showed the highest toughness among the prepared films. It was mainly resulted from the large increase of the strain at breakage. During the deformation, the networked structure was partially destroyed and gave partial freedom for the chain. By the applied strain, the randomly oriented chain can be reoriented and extended along the strain direction. It can be confirmed from several yield points in the stress-strain curve (denoted as green arrows line in Fig. 4(g)). At the yield points, the stress was used for reorientation and extension of the chain rather than elongation. By the process, the film can have the aligned and robustly networked structure as similar to the film from 16 wt% solution.
Table 2
The mechanical properties of the regenerated cellulose films
Cellulose content
(wt%)
|
9
|
12
|
14
|
16
|
18
|
Tensile strengtha (± SD)
(MPa)
|
133.0 (± 19.4)
|
170.5 (± 15.2)
|
259.2 (± 22.1)
|
420.2 (± 28.7)
|
228.4 (± 20.1)
|
Initial modulusa (± SD)
(GPa)
|
4.21 (± 1.12)
|
3.69 (± 1.71)
|
9.03 (± 2.09)
|
9.44 (± 1.55)
|
3.25 (± 0.91)
|
Toughness
(MJ/m3)
|
3.32
|
8.81
|
9.05
|
18.6
|
20.3
|
a Data presented are the mean of triplicate measurements (± SD)
|