Introduction

Cellulose and its family are the most abundant natural resource in the world. The generation of new green materials from such a non-petroleum source has become an issue that has attracted substantial interest. Hydroxypropyl cellulose (HPC), one of the most important cellulose derivatives, has been studied extensively and many applications have been found, including thickening agents, fillers, dietary fiber, anti-clumping agents and emulsifiers*. Numerous investigations have examined the bicomponent phase diagram of aqueous HPC solution, i.e. HPC in H2O, to determine the cloud temperature, TC, and the liquid–crystal critical concentration, C*(Fortin and Charlet 1989).

Many polysaccharides, including HPC, form a liquid crystalline (LC) phase in various solvents (Aharoni 1980, 1981, 1982; Werbowyj and Gray 1976; Chang and Gray 1978; Werbowyj and Gray 1980; Conio et al. 1983; Marsano and Fossati 2000; Rwei et al. 2009). This characteristic has made the cellulose family attractive for some specific optical applications. Many researchers are interested in its mesophase structure under certain thermal or solvent conditions because it provides useful information about the chain conformation of polymeric materials. HPC is known to be one of the few cellulose derivatives that can form LC phases with both thermotropic and lyotropic properties (Flory 1956, 1978; Budgell 1989). In our earlier study, both 2-D and 3-D phase diagrams of the HPC/H2O/H3PO4 system at various temperatures were successfully obtained using a light transmission device and a polarized light microscope (Rwei and Lyu 2012). The addition of H3PO4 was found to inhibit the formation of the LC phase. However, as the temperature was raised from 50 to 70 °C, the LC phase could only be maintained at high H3PO4 concentrations. Furthermore, adding H3PO4 suppressed the formation of the cloudy suspension (CS) phase and increased the lower critical solution temperature (LCST) value.

The rheological technique is another important method for characterizing the phase transition of cellulosic solutions (Guido and Grizzuti 1995). Clasen and Kulicke (2001) extensively examined the rheological investigations of cellulosic solutions. Thermoreversible gelation has been widely studied with reference to methylcellulose (MC) and hydroxypropylmethylcellulose (HPMC) (Heymann 1935; Sarkar 1979, 1995; Haque and Morris 1993; Haque et al. 1993; Hirrien et al. 1998; Desbrières et al. 2000; Hussain et al. 2002). Fujii reported on the LC behavior of hydroxypropylcellulose (HPC) in a mixed solvent of glycerol and water. He studied HPC orientation was studied in both an elongational flow field and a shear flow field and determines that the elongation flow promoted the formation of the LC phase. Although the phase behavior of aqueous polysaccharide materials, especially HPC solution, has attracted substantial interest, no study has focused on the rheology of HPC/H2O/H3PO4 tertiary system.

The goal of this paper is to examine the rheological behavior of the HPC/H2O/H3PO4 tertiary system based on the previously obtained 3-D phase diagram. The effects of temperature, HPC concentration, liquid composition (H2O/H3PO4 ratio), frequency, and phase transition on the rheological parameters are studied comprehensively. The ultimate purpose is to elucidate, by making rheological measurements, the evolution or molecular interaction throughout the phase change when an external force is applied.

Experiments

The HPC materials that were adopted in this study were purchased from TCI and had molecular weights (Mw) of 110,000–150,000 (g/mol). The solvents were deionized water and phosphoric acid (Acros). The ranges of concentrations of the HPC polymer, phosphoric acid and deionized water were 0.5 %–70 wt%, 0 %– 55.2 wt%, and 8.3 %–99.8 wt%, respectively. The phase diagram was taken from our previous work. Notably, the 2-D phase diagrams that were obtained herein were based on rising temperature: the phase separation was therefore a binodal separation rather than a spinodal separation. Hence, the effect of temperature on the rheological measurements was examined.

Rheological measurements were made using a Physica MCR301 rheometer that was fitted with a C-CC 27/T200 cup and B-CC 27/Q1 bob attachment. (Anton Paar, Graz, Austria). The Couette device that was used in a dilute solution (<5 wt%) was composed of a single gap cylinder with a concentric cylindrical cup and a hollow bob with internal and external diameters of 2.66 and 2.89 cm, respectively. A parallel plate with a diameter of 2.5 cm was used for a concentrated solution (at least 6 wt%). The investigated frequency range was 0.1–100 Hz.

Results and discussion

Effect of phase change on rheological parameters

The 3-D phase diagram of the HPC/H2O/H3PO4 tertiary system against temperature was obtained in our previous study. Four distinct phases—the completely separated (S) phase, the cloudy suspension (CS) phase, the liquid crystalline miscible (LC) phase, and the isotropically miscible (I) phase–were identified, as presented in Fig 1. Five points, a–e, with various HPC/liquid compositions, were selected to observe the effect of increasing temperature on targeted rheological parameters. Among all of the rheological parameters—storage modulus G′, loss modulus G″, tanδ, and complex viscosity η*—the latter two were found to vary distinctly as a function of frequency (ω) during phase changes, and were plotted in Figs. 2 and 3, respectively. Figures 2a and 3a plot typical tanδ and η* versus ω for the pure LC phase. The complex viscosity revealed typical shear thinning behavior with no special characteristic. The parameter tanδ, interestingly, was independent of ω and insensitive to temperature. The liquid crystalline state of the biopolymer is normally regarded as a condensed state with highly anisotropic order. The anisotropic phase with a regularly aligned domain is not easily affected by frequency from 0.1 to 100 Hz, or by the sweep of temperature from 25 to 70 °C. The value of tan δ therefore remains constant.

Fig. 1
figure 1

Five points, labeled from a to e, were selected based on different phase regions of the 3-D HPC/H2O/H3PO4 tertiary diagram (Each phase is represented in a specific color. dark grey the completely separated phase (S); yellow the liquid crystalline miscible phase (LC); light grey the cloudy suspension phase (CS); white the isotropically miscible phase (I); blue the overlapping area of dark grey and yellow (S + LC); purple the overlapping area of light grey and yellow (CS + LC)). They experienced different phase changes when the temperature increased. (Color figure online)

Full size image
Fig. 2
figure 2

The tanδ versus frequency (ω) of points (a–e) shown in Fig. 1 at various temperatures. a The pure LC phase (point-a: 50 % HPC/25 % H2O/25 % H3PO4), b the pure I phase (point-b: 30 % HPC/40 % H2O/30 % H3PO4), c from LC phase to I phase (point-c: 40 % HPC/35 % H2O/25 % H3PO4), d from LC phase to CS phase (point-d: 65 % HPC/31.5 % H2O/3.5 % H3PO4), e from I phase to CS phase (point-e: 20 % HPC/70 % H2O/10 % H3PO4)

Full size image
Fig. 3
figure 3

The complex viscosity η* versus frequency (ω) of points (a–e) shown in Fig. 1 at various temperatures. a The pure LC phase (point-a: 50 % HPC/25 % H2O/25 % H3PO4), b the pure I phase (point-b: 30 % HPC/40 % H2O/30 % H3PO4), c from LC phase to I phase (point-c: 40 % HPC/35 % H2O/25 % H3PO4), d from LC phase to CS phase (point-d: 65 % HPC/31.5 % H2O/3.5 % H3PO4), e from I phase to CS phase (point-e: 20 % HPC/70 % H2O/10 % H3PO4)

Full size image

Figure 2b demonstrates that the tanδ of the single I phase decreased slowly as the frequency increased but only up to around 100 Hz where it increased abruptly. An increase of dynamic frequency reveals a decrease in the processing time, indicating the time to observe the response from the deformed HPC polysaccharide would be reduced. Subsequently, the loss modulus G″, representing the soft or fluid character of the HPC polymer, would gradually decrease. The parameter tanδ, defined as G″/G′, therefore declines. The sharp increase in tanδ at high frequency in Fig. 2b, however, may be associated with the fact that increasing the frequency over a threshold eliminates the interaction among the HPC polymers, significantly increasing the loss modulus G″.

Figure 2c plots the tanδ of a given HPC/fluid composition experienced two homogeneous phases- the LC and I phases. The typical behaviors of HPC in the LC and I phases that were discussed above are elucidated again herein. The existence of I phase at high temperature is easily identified by the abrupt increase in tanδ. Notably, the change in viscosity from Fig. 3c reflects difficulty in distinguishing the I phase from the LC phase, since both phases display typical shear thinning behaviors. Briefly, tanδ is the only rheological parameter that can be used to distinguish between I and LC phases.

Figures 2d and e reveal the significant irregular damping in tanδ, defined as G″/G′, at high temperatures, which can be used to identify the CS phase. Carotenuto and Grizzuti (2006) made similar observations, and identified an unstable perturbation of G′ and G″ against temperature when the CS phase had been formed. Notably, due to the damping situation at 70 °C is too serious to be accurately detected, the highest temperature shown in Fig. 2 d and e are 65 and 60 °C, respectively.

Figure 3d and e display the occurrence of serious dropping for η*at 65 and 60 °C, respectively. All results indicate that the agglomeration of the CS phase at high temperatures is too serious to generate a smooth flow. Figure 3d and e show that η* increased with temperature when the HPC fluid was heated over the CS phase boundary, which a behavior is not typical of polymer solutions. When the temperature was increased over the LCST boundary, the interaction, usually hydrogen bonding between water molecules and HPC, decreased and the water molecules were expelled by the HPC network. Poor solvent behavior and the CS phase were observed. Accordingly, the dynamic viscosity increased with temperature when solution was heated above the LCST point.

Figure 4a and b plot the effect of increasing the HPC concentration, at a given liquid composition, on the tanδ versus ω curve at 25, and 70 °C, respectively. The data reveal that the tanδ of the LC phase is insensitive to the HPC concentration, independently of the liquid composition. The tanδ of the I phase, however, decreased as the HPC concentration increased and again increased abruptly at high frequencies, as discussed earlier. In short, the tanδ was screened out to become the only useful rheological parameter for distinguishing the I phase from the LC phase.

Fig. 4
figure 4figure 4

The effect HPC concentration, at a given liquid composition, on the plot of tanδ versus ω at a 25, 1 phase diagram at 25 °C, 2 line-a (H2O:H3PO4 = 0.43:1), 3 line-b (H2O:H3PO4 = 1:1), 4 line-c (H2O:H3PO4 = 9:1) and b 70 °C, 1 phase diagram at 70 °C, 2 line-d (H2O:H3PO4 = 0.43:1) (70 °C), 3 line-e (H2O:H3PO4 = 1:1) (70 °C), respectively

Full size image

Figure 4b-2 plots tanδ at the four marked points, d1 to d4, spread along line-d with a fixed liquid composition of H2O/H3PO4 = 0.43. The HPC concentration increased from d1 to d4, enabling tanδ to experience the phase changes in the sequence of I–LC–I. Figure 4b-2 indeed reveals that the only point in the middle, d2, exhibited only LC character, and the rest of the points, d1, d3, and d4, exhibited the characteristics of the I phase. This result reconfirms the accuracy of the previously obtained 3-D phase diagram.

Arrhenius model of temperature-dependence

The temperature-dependence of the fluidity of liquid concerns the decrease in liquid viscosity with increasing temperature. This relationship is normally expressed using the Arrhenius model, in which the viscosity depends exponentially on the reciprocal of temperature. The model is based on the assumption that the fluid flow obeys the Arrhenius equation for molecular kinetics (Giner et al. 1996; Saravacos 1970; Kaya and Belibaǧlı 2002):

$$ \eta = \eta_{\infty } \,.\,{\text{e}}^{{\frac{\text{E}}{\text{RT}}}} $$
(1)
$$ \ln \eta = \ln \eta_{\infty } + \frac{\text{E}}{\text{R}}\,.\,\left( {\frac{1}{\text{T}}} \right) $$
(2)

where T is temperature; η∞ is a coefficient; E is the activation energy, and R is the universal gas constant. Figures 5a–c plot lnη against 1/T for the five HPC/liquid compositions in Fig. 1. Figures 5a and b exhibit a linear relationship, indicating that flows in a pure homogeneous phase, whether the I phase, the LC phase, or a combination of both, obey the Arrhenius model. However, Fig. 5c clearly demonstrates that when the temperature was increased over the CS boundary, the Arrhenius model no longer applied because the agglomeration of suspended material in the CS caused viscosity damping, as mentioned above. Since the CS phase does not follow the Arrhenius model, the remaining studies of E value herein focus on the I and LC phases only.

Fig. 5
figure 5

The plot of lnη against 1/T for points (a–e) shown in Fig. 1. a Under the same phase: LC–LC (point-a) and I–I (point-b). b Homogeneous phase transition: LC-I (point-c). c heterogeneous phase transition: LC-CS (point-d) and I-CS (point-e)

Full size image

Figure 6 reveals that at a fixed liquid composition, meaning a constant H2O/H3PO4 ratio, the increase in HPC concentration in the I phase reduces the activation energy, E. However, once the HPC concentration is in the LC regime, the E value of the formation of the LC phase becomes highly insensitive to the HPC concentration. The activation energy of flow represents the sensitivity of the viscosity to the temperature. An increase of HPC concentration in the I phase implies a decrease in the total amount of liquid and a consequent decrease in sensitivity to the temperature. In the LC region, the anisotropic phase with a regularly aligned domain makes the conformation of HPC molecules insensitivity to the temperature, the activation energy E of viscosity is therefore independent of the liquid composition. Notably, the crossover point at 36 wt% of HPC is exactly the LC-I boundary of line-f shown in the phase diagram of Fig. 6.

Fig. 6
figure 6

The plot of activation energy E against HPC concentration at the H2O/H3PO4 ratio of 0.43/1. The boundary between I phase and LC phase can be determined

Full size image

Figure 7 plots the effect of liquid composition on the E value at a given HPC concentration. It reveals that E decreases as the H2O/H3PO4 ratio increases, indicating that the addition of H3PO4 makes aqueous HPC solution more sensitive to temperature. In our previous study (Rwei and Lyu 2012), adding H3PO4 was found to suppress the formation of LCST, indicating that H3PO4 sticks to HPC more stable than a water molecules does. In general, given a particular increase in temperature, a good solvent would cause the polymer solute to have a larger entropy (ΔS) than would a theta solvent. Consequently, the free energy term, TΔS, would be larger. Haque et al. provided the same hypothesis to explain the occurrence of the LCST of HPC (Haque et al. 1993; Haque and Morris 1993). H3PO4 is a better solvent than H2O at higher temperature. Its inclusion certainly makes the HPC solution more sensitive to a change in temperature. Moreover, Fig. 7 reconfirms that the E value of the I phase exceeds that of the LC phase for any given liquid composition, as mentioned previously.

Fig. 7
figure 7

The plot of activation energy E against H2O/H3PO4 ratio for HPC 30 wt% (line g, I phase), and 40 wt% (Line h, LC phase), respectively

Full size image

Sol/Gel transition

A simple and commonly used means of determining the ‘‘sol/gel transition’’ in the dynamic modulus spectrum is based on the point at which G′ and G″ are equal over a frequency range of more than 2 orders. Figure 8b–f plot the evolution of the dynamic moduli, G′ and G″, near the sol/gel transition (SGT) at room temperature at frequencies of between 0.1 and 10 Hz (Shukla et al. 2002, 2003; Rwei et al. 2005; Carotenuto and Grizzuti 2006). Five liquid compositions, labeled in “line-i” to “line-m” in the phase diagram of Fig. 8a, were selected, and the SGT points were determined for each line thereafter. The experimental results show that the SGT concentration decreases with the H2O/H3PO4 ratio. Notably, “line-i” in Fig. 8b indicates that an H3PO4-rich state has no SGT point. These results reconfirm that H3PO4 is an effective solvent that increases the SGT concentration in the HPC solution.

Fig. 8
figure 8

a–f The sol/gel transition (SGT) tests at 25 °C under various solvent compositions. The G′ and G″ were plotted against frequency on b–f, representing line i–line m, respectively. a Sol/gel transition (SGT) test under various solvent compositions at 25 °C, b line-i (H2O:H3PO4 = 0.43:1) (25 °C, No SGT), c line-j (H2O:H3PO4 = 1:1) (25 °C, No SGT), d line-k (H2O:H3PO4 = 2.3:1) (25 °C, SGT at 65 %), e line-l (H2O:H3PO4 = 9:1) (25 °C, SGT at 50 %), f line-m (H2O:H3PO4 = 1:0) (25 °C, SGT at 30 %)

Full size image

An investigation similar to the above was performed to determine SGT at 45 and 75 °C. Figure 9a plots the results. The SGT concentration decreased as the temperature increased. This finding is consistent with those concerning other polysaccharide systems, which reportedly reach a gel state easily at high temperatures. Desbrières has expressed that gelation on heating is thermoreversible and attributable to hydrophobic interactions. Its mechanism is associated with the presence of hydrophobic zones as the temperature increases and their initiation of physical crosslinking (Desbrières et al. 2000).

Fig. 9
figure 9

a Sol–gel transition (SGT) test under various liquid compositions and temperatures. SGT plot on the HPC/H2O/H3PO4 diagram for b 25 °C, c 45 °C and d 70 °C, respectively

Full size image

Figures 9b–d plot the SGT curves that correspond to Fig. 9a on 2-D diagrams for 25, 45 and 70 °C, respectively. The explanations of SGT formation are similar to those of CS phase formation or LCST occurrence. All of the findings were associated with the easy expulsion of the HPC intermolecular network at high temperature and the onset of the CS phase. Some reports have expressed the coupling of SGT and the LC point, or SGT and the CS point. Carotenuto and Grizzuti (2006) pointed out a two-step phase transition mechanism—formation of a CS phase followed by gelation. The onsets of both CS phase and gelation occur at well-defined characteristic temperatures, which do not depend on the overall polymer concentration. Some reports, however, have predicted a sequence of appearances of CS phase and gelation as the HPC concentration increases (Carotenuto and Grizzuti 2006; Conio et al. 1983; Rwei et al. 2009). Figure 9b–d reveal that both SGT and LC points are at well defined but different HPC concentrations. However, no absolute order of the appearance of SGT and LC points or SGT and CS points exists for the HPC/H2O/H3PO4 tertiary system. The only certainty is that at 45 or 70 °C, and a high H3PO4 content to suppress the formation of the CS phase, the LC point is always at a lower HPC concentration than the SGT point. The order of appearance of these points can be best understood by tracing the SGT curve in Fig. 9.

Conclusions

This study characterized the rheological behavior of the HPC/H2O/H3PO4 system. The tanδ parameter was concluded to be the only useful rheological parameter for distinguishing the isotropic (I) phase from the liquid crystalline (LC) phase. The tanδ parameter in the LC phase was insensitive to the process factors, including the frequency, HPC concentration, and the temperature, whereas in the I phase, tanδ decreased as the HPC concentration increased, and slightly increased with temperature; most notably, it decreased gradually as frequency increased but abruptly increase as the frequency was increased into the high-frequency regime (at approximately 100 Hz). The anisotropic character is responsible for the regular alignment of domain in the LC phase, which results in a stably condensed condition that is not strongly influenced by any process factor. Agglomeration of the cloudy suspension (CS) phase at high temperatures was too serious to allow a smooth flow, so tanδ and η* were significantly damped, which result can be regarded as a good indicator of the formation of the CS phase. With an increase of temperature, the viscosity of flow, which was in a single homogeneous phase—either the I phase or the LC phase—or experienced a combination of these two homogeneous phases in sequence, obeyed the Arrhenius model. In contrast, once the temperature rose above a value at which the heterogeneous CS phase was formed, the Arrhenius model no longer applied. The activation energy E for I phase at a fixed liquid composition, decreased as the HPC concentration increased. However, once the HPC concentration rose into the LC regime, the activation energy E for LC phase became very insensitive to the HPC concentration. Finally, the sol/gel transition concentration decreased as temperature increased but increased with the H3PO4 content. Additionally, this tertiary system exhibited no clear order of the onsets of the formations of SGT phase, the LC phase, and the CS phase as HPC concentration was varied.