Thiourea Aqueous Solution

Feb 18, 2004 - Given enough time, G' of all solutions can exceed G' ' at a certain temperature slightly lower than the gelation temperature, indicatin...
0 downloads 0 Views 364KB Size
2086

Langmuir 2004, 20, 2086-2093

Thermal Gelation of Cellulose in a NaOH/Thiourea Aqueous Solution Lihui Weng,†,‡ Lina Zhang,*,† Dong Ruan,† Lianghe Shi,‡ and Jian Xu‡ Department of Chemistry, Wuhan University, Wuhan 430072, People’s Republic of China, and State Key Laboratory of Polymer Physics and Chemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100080, People’s Republic of China Received October 25, 2003. In Final Form: January 2, 2004 Utilizing a novel solvent of cellulose, 6 wt % NaOH/5 wt % thiourea aqueous solution, for the first time, we prepared the thermally induced cellulose gel. We investigated the thermal gelation of cellulose solutions with rheometry and the structure of the gel with 13C NMR, wide-angle X-ray diffraction, environmental scanning electron microscopy, and atomic force microscopy. The cellulose solutions revealed an increase in both the storage modulus (G′) and the loss modulus (G′′) with an increase in the temperature during gelation. The temperature at the turning point, where G′ overrides G′′ because of the onset of gelation, decreased from 38.6 to 20.1 °C with an increase of cellulose concentration from 4 to 6 wt %. Given enough time, G′ of all solutions can exceed G′′ at a certain temperature slightly lower than the gelation temperature, indicating that the occurrence of the gelation is also a function of time. Each of the assigned peaks of NMR of the cellulose gel is similar to that of the cellulose solution, suggesting that the gelation resulted from a physical cross-linking. The gels were composed of relatively stable network units with an average diameter of about 47 nm. At either a higher temperature (at 60 °C for 30 s) or a longer gelation time (at 30 °C for 157 s), the gel in the 5 wt % cellulose solution could form. A schematic gelation process was proposed to illustrate the sol-gel transition: the random self-association of the cellulose chains having the exposed hydroxyl in the aqueous solution promotes the physical cross-linking networks.

Introduction Renewable resources have attracted much attention because of the great importance for a sustainable development and environmental conservation. Cellulose is the richest natural polymer on earth, whereas some of its characteristic features retard its applications such as the rigidity of the chain and the insolubility in many solvents. Recently, we have developed a new solvent of cellulose (6 wt % NaOH/5 wt % thiourea aqueous solution) for preparation of the regenerated cellulose film.1 The 6 wt % NaOH/5 wt % thiourea aqueous solution can completely dissolve cellulose I with viscosity-average molecular weight (Mη) less than 8.0 × 104 at -8 °C because NaOH breaks the intermolecular hydrogen bonding of cellulose while thiourea prevents the self-association of cellulose molecules. Interestingly, the cellulose solution is transparent at room temperature but turns into an opaque gel that retains a definite shape at an elevated temperature. Furthermore, the opaque gel turns again into a transparent solution when frozen at about -8 °C for 12 h. A basic understanding of the process and mechanism of the solgel transition is essential for a successful research and development of a cellulose solution and gel. In the formation of physical gels, there are no covalent bonds forming or breaking and the polymers form a crosslinked network through ionic bonding, hydrogen bonding, or an associative interaction of polymer-polymer. In contrast to the covalent bond in a chemical hydrogel, the binding energy of cross-linking in physical gels is on the level of thermal energy so that the network junctions can be created and destroyed by the thermal motion of the * Correspondence to L. Zhang. E-mail: [email protected]. Tel.: +86-27-87219274. Fax: +86-27-87882661. † Wuhan University. ‡ Chinese Academy of Sciences. (1) Zhang, L.; Ruan, D.; Gao, S. J. Polym. Sci., Part B: Polym. Phys. 2002, 40, 1521.

polymers, leading to the unique properties of physical gels.2 Most of the gelation in the cellulose solution usually takes place by cooling the solution below a certain temperature. For example, the cellulose solutions in ammonia/ammonium thiocyanate solvent form thermoreversible gels at temperatures below 30 °C;3 the physical gelation of the cellulose derivative of 6-hydroxyl groups takes place in a tetrahydrofuran solution below 270 K.4 However, aqueous solutions of methylcellulose (MC) are well-known to undergo thermoreversible gelation in an aqueous solution upon heating. The gelation of the MC solution is primarily caused by the hydrophobic interaction between the molecules with methoxyl substitution.5-8 Moreover, MC solutions with different concentrations gel above 60 °C, and the gelation temperature of MC solutions decreases with an increase of the MC concentration.7 Other cellulosic derivatives that form gels at an elevated temperature have been reported, for example, rheological behavior during thermoreversible gelation of aqueous systems of ethyl(hydroxyethyl)cellulose (EHEC) in the presence of sodium dodecyl sulfate (SDS) at 10-45 °C,9 a temperature-induced gelation of the system EHEC + SDS + water,10 and the gel made of EHEC and cationic surfactants.11 In addition, thermal gelation also occurs in cellulose/cuprammonium (2) Guenet, J. Thermoreversible Gelation of Polymers and Biopolymers; Academics Press: London, 1992. (3) Frey, M. W.; Cuculo, J. A.; Khan, S. A. J. Polym. Sci., Part B: Polym. Phys. 1996, 34, 2375. (4) Itagaki, H.; Tokai, M.; Kondo, T. Polymer 1997, 38, 4201. (5) Li, L. Macromolecules 2002, 35, 5990. (6) Li, L.; Thangamathesvaran, P. M.; Yue, C. Y.; Tam, K. C.; Hu, X.; Lam, Y. C. Langmuir 2001, 17, 8062. (7) Takahashi, M.; Shimazaki, M.; Yamamoto, J. J. Polym. Sci., Part B: Polym. Phys. 2001, 39, 91. (8) Li, L.; Shan, H.; Yue, C. Y.; Lam, Y. C.; Tam, K. C.; Hu, X. Langmuir 2002, 18, 7291. (9) Nystrom, B.; Walderhaug, H.; Hansen, F. K. Langmuir 1995, 11, 750. (10) Lindell, K.; Cabane, B. Langmuir 1998, 14, 6361. (11) Cabane, B.; Lindell, K.; Engstom, S.; Lindman, B. Macromolecules 1996, 29, 3188.

10.1021/la035995o CCC: $27.50 © 2004 American Chemical Society Published on Web 02/18/2004

Thermal Gelation of Cellulose/NaOH/Thiourea Solution

Langmuir, Vol. 20, No. 6, 2004 2087

solution 12 due to the partial destruction of the complex that leads to the formation of intermolecular hydrogen bonds of cellulose at above 40 °C. Fundamental research and industrial application drive the interest of the physical gels in aqueous systems of cellulose. The transient nature of the physical junctions makes it difficult to study physical gels near their point.9 Concerning the gelation in cellulose solution, it is crucial and complicated to solve many problems, especially what leads cellulose in NaOH/ thiourea aqueous solution to remain a transparent solution at a low temperature, whereas it forms a gel upon heating, and how do the polymer concentration and temperature influence the thermally induced gelation. In this paper, we studied the formation and mechanism of the gelation from the cellulose solution in the NaOH/thiourea aqueous solution system with various tools [rheometry, 13C NMR, environmental scanning electron microscopy (ESEM), wide-angle X-ray diffractometry (WAXD), and atomic force microscopy (AFM)] and hope our work may contribute some meaningful information to deeply understanding physical gelation.

respectively. The gel sample for ESEM was prepared by heating each solution at 60 °C for 5 min, and the size was about 0.5 × 0.5 cm2. AFM measurements were performed with a Dimension 3000 with a NanoScope IIIa controller by Digital Instruments (U.S.A.) in contact mode under ambient conditions. The spring constant of the V-shaped nanoprobe was 0.16 N/m. The gel sample from a 5 wt % cellulose solution was dispersed in distilled water and in the diluted solvent (NaOH/thiourea/water, 0.5:1:98.5 mL) to form a suspension solution with stable particles, and then a drop of the resulting suspension was spread on a mica substrate for AFM measurement. The samples were placed on top of the piezoelectric scanner, brought into contact with the nanoprobe cantilever tip, and scanned at the present cantilever deflection. The statistics were done by measuring the diameters of 20 particles in one top-view AFM image and then averaging.

Experimental Section Materials and Sample Preparation. Cellulose (cotton linters) was purchased from Hubei Chemical Fiber Co., Ltd, and the R-cellulose content was more than 95%. The cellulose was used without being further purified. Mη was measured to be 8.0 × 104 with cadoxen as the solvent at 25 °C by using viscometery according to [η] ) 3.85 × 10-2Mw0.76 (mL g-1).13 The aqueous solution of NaOH (AR grade)/thiourea (AR grade)/H2O (6:5:89 by weight) was used as the solvent of cellulose. The mixture of cellulose and NaOH/thiourea aqueous solution was stored at -8 °C for 12 h and then stirred at 20 °C to obtain a transparent solution.1 Cellulose gel for measurements was prepared by heating the concentrated solution of cellulose at 60 °C or the desired temperature for 5 min until the solution gelled. Characterization. Dynamic rheology experiments were carried out on an ARES rheometer (Rheometric Scientific, Inc., NJ) at temperatures from 0 to 70 °C. A couette (two concentric cylinders with a gap of 2 mm) was used for monitoring the moduli evolution during the gelation process. The temperature was controlled within (0.5 °C. The shear storage modulus (G′) and loss modulus (G′′) were measured as a function of the frequency and temperature, as well as time. Degassed cellulose solution was heated to the desired temperature directly in the rheometer (without shearing or oscillating). For the frequency and time sweep measurements, it is defined as time t ) 0 s when the temperature reaches the desired temperature. The dynamic temperature and time sweep measurements were both performed at a shear strain amplitude of 10% with an angular frequency of 1 rad/s within the linear viscoelastic region. To prevent dehydration from the solution, a thin layer of low-viscosity mineral oil was spread on the exposed surface of the solution. The cellulose gel for 13C NMR measurements was directly prepared in the specific tube by heating the cellulose solution with 6 wt % NaOH/5 wt % deuterium oxide (D2O) as the solvent to 60 °C for 5 min. The measurements of the gel and solution were carried out with a Bruker DMX-300 NMR spectrometer, which was stabilized by Bruker temperature control units. The resonance frequency was 75.00 MHz. To clarify the structure of the cellulose gel, the WAXD measurement of the cellulose gel film about 0.5 mm in thickness was performed with an X-ray diffractrometer (Rigaku D/max-II, Japan) using a Cu KR target at 40 V and 50 mA. The diffraction angle ranged from 5 to 45°. The cross-section morphology of the cellulose gel was observed by an environmental scanning electron microscope (ESEMKYKY1600, China) at 25 °C. The pressure in the specimen chamber and the accelerating voltage were 1000 Pa and 25 kV, (12) Miyamoto, Y.; Matsui, T.; Saito, M.; Okajima, K. Seni Kikai Gakkai Shi 1996, 49, 45. (13) Brown, W.; Wiskston, R. Eur. Polym. J. 1965, 1, 1.

Results and Discussion Thermal Gelation of the Cellulose Solution. The sol-gel transition was crucial for studying the structure and predicting the properties of gels. The rheological method is the most direct and reliable way for determination of the sol-gel transition and characterization of rheological properties of gels.5 Figure 1 shows the angular frequency (ω) dependence of the shear storage modulus (G′) and loss modulus (G′′) for a 5 wt % cellulose solution at 30 °C at four different time periods during gelation. The cellulose solution initially (0 s) behaved as a kind of viscous liquid, and the G′ was smaller than the G′′ in the whole range of frequencies measured. With the gelation proceeding, the storage modulus increased more rapidly than the loss modulus, and G′ exceeded G′′ at low frequencies (720 s). Moreover, the curve of the storage modulus was also flatter than that of the loss modulus, implying that the gelation has started. With an increase of gelation time (3480 s), the range where G′ exceeded G′′ expanded to a higher frequency. Finally, G′ became flatter and less dependent on the frequency at a low frequency (22 740 s) and is considerably larger than G′′ in most of the range of frequencies, a characteristic that indicates the gel state.4 Detection of the gel point has often relied on very simple criteria. For example, it is an indication of the gel point when the sample gelling signal becomes just greater than the background noise or when G′ becomes greater than a pre-assigned threshold value.14 Winter et al.15-18 defined the sol-gel transition as the point at which both G′ and G′′ scale with ωn so that the ratio of G′′ to G′ (i.e., tan δ) is independent of the frequency. However, for the cellulose/NaOH/thiourea aqueous solution, there is not such a condition that can be found to satisfy the requirement in applying the Winter’s scaling law at the sol-gel transition. Therefore, the traditional definition, that is, the crossover of the G′ and G′′ curves,19 was chosen for the determination of the gel point. It should be pointed out that the gel point determined by this method is frequency-dependent, as shown in the Figure 1. For consistency of comparison, we choose the frequency 1 rad/s for all the samples in determination of the gel point and gelation time. An explanation for thermally induced gelation is that gel formation involves the association of chain segments, which results in a three-dimensional network that contains solvent in the interstices. The associated regions are known as junction zones and formed with of two or (14) Kavanagh, G. M.; Ross-Murphy, S. B. Prog. Polym. Sci. 1998, 23, 533. (15) Chambon, F.; Winter, H. H. Polym. Bull. (Berlin) 1985, 13, 499. (16) Winter, H. H.; Chambon, F. J. Rheol. 1986, 30, 367. (17) Scanlan, J. C.; Winter, H. H. Macromolecules 1991, 24, 47. (18) Mours, M.; Winter, H. H. Macromolecules 1996, 29, 7221. (19) Ferry, J. D. Viscoelastic Properties Of Polymers, 3rd ed.; John Wiley & Sons: New York, 1980.

2088

Langmuir, Vol. 20, No. 6, 2004

Weng et al.

Figure 1. Storage modulus (G′, b) and loss modulus (G′′, O) of the 5 wt % cellulose solution as a function of the angular frequency (ω) at 30 °C for (a) 0, (b) 720, (c) 3480, and (d) 22 740 s.

more chains.20 Different types of association responsible for the junction zones have been discussed.21,22 Cabane et al.11 have indicated that raising the temperature causes some macromolecules or some sections of macromolecules to dump into lumps of a polymer-rich phase, which leads to a lower critical solution temperature behavior of the system where there are always interactions of hydrophobic groups. However, because there is no hydrophobic group in our system, the self-association tendency of cellulose plays a key role in the formation of gel. Figure 2 shows the temperature dependence of the storage and loss moduli for the 4, 5, and 6 wt % cellulose solutions with a heating rate of 2 °C/min. Both the G′ and the G′′ values increased with an increase in the temperature and exhibited three regions. The first region was characterized by a relatively slow increase of G′ and G′′, where G′ is lower than G′′, showing the common viscoelastic behavior of a liquid. Subsequently, with the partial formation of aggregates or clusters through self-association of cellulose, both G′ and G′′ increased gradually. However, G′ exhibited a much higher increasing rate, suggesting that the evolution of the self-association mainly contributes to the increase in the elasticity of the system. The intersection temperature of the G′ and G′′ curves, that is, the gel point, decreased from 38.6 to 20.1 °C with an increase of the cellulose concentration from 4 to 6 wt %. G′ and G′′ increased dramatically with the increase in the temperature until about 50 °C. In the second region, G′ is higher than G′′, indicating that an elastic gel network formed. When the (20) Rees, D. A. Adv. Carbohydr. Chem. Biochem. 1969, 24, 267. (21) Flory, P. J. Faraday Discuss. Chem. Soc. 1974, 57, 1. (22) Rees, D. A. Chem. Ind. (London) 1972, 630.

Figure 2. Dependence of the storage and loss moduli of the cellulose solutions in the NaOH/thiourea aqueous solution system on the elevated temperature with a heating rate of 2 °C/min.

temperature was higher than 50 °C, the increasing speed of G′ decreased but never equaled 0, namely, no plateau was observed, suggesting that the gelation process is much different from a chemically cross-linked gelation process, in which there is always a G′ plateau following the second rapid increase of the G′ region. This probably resulted from the gradual formation of the condensed gel network accompanied with microphase separation.23 The G′ value of each solution is temperature-dependent, which is attributed to the fact that the thermal disposal created (23) Roy, C.; Butdtova, T.; Navard, P. Biomacromolecules 2003, 4, 259.

Thermal Gelation of Cellulose/NaOH/Thiourea Solution

Figure 3. Gelation kinetics monitored in terms of the storage and loss moduli during the gelation of 4 wt % (G′, 1; G′′, 2) and 5 wt % (G′, 3; G′′, 4) cellulose solutions at 30 °C.

the physical interaction junctions in the concentrated solutions. The sol-gel transition of the cellulose solution occurred in the range from 20 to 40 °C on the whole. Above 50 °C, all the G′ curves showed a much lower increasing rate, indicating the existence of the gel network with junctions of cellulose chains. It is worth noting that the sol-gel transition temperature significantly shifts to lower temperatures with an increase in the polymer concentration. This implies that the association is caused not only by heating but also by elevating the concentration because elevating the concentration increases the chance of collision of the cellulose chains. The kinetics of the gelation process for 4 and 5 wt % cellulose solutions with G′ and G′′ as a function of the gelation time with a shear strain amplitude of 10% and an angular frequency of 1 rad/s at 30 °C are illustrated in Figure 3. Given enough time, at 30 °C, a temperature below but close to the gel point determined in Figure 2, the gelation could take place both in the 4 and in the 5 wt % cellulose solutions, implying that the solution is not an exactly equilibrium system. Meanwhile, the gelation time decreased from 4215 to 157 s with an increase in the cellulose concentration from 4 to 5 wt %. As expected, the higher the concentration, the faster the gelation, implying that the gelation process relates closely to an increasing chance for collision of macromolecular chains. The G′ curves of the concentrations began with an increasing region directly without showing any lag time, which is not typical for a “normal” gelling system. Similar to the results observed in Figure 2, even staying after 3000 s for 5 wt % cellulose solution, the G′ is still increasing gradually without showing a plateau, indicating that the system hardly reaches its equilibrium state, which is one of the main reasons why it is difficult to characterize the behavior of a physical gel.10 There is a small decrease in G′ of 5 wt % cellulose solution at about 3500 s, which can be attributed to instantaneous partial destruction of the network formed under shear strength. Compared with the results from the temperature sweep in Figure 2, the ultimate values of G′ of both 4 and 5 wt % at 30 °C are much lower, indicating that the gels formed in this case are much weaker in strength than that formed at a high temperature. Furthermore, the comparison of Figures 2 and 3 reveals that the gelation of the cellulose solution is a function of not only the temperature and concentration but also time. Therefore, the time control is important during the measurements; namely, the measurements should start as soon as the temperature reaches the desired value. It is well-known that low-frequency sweeps are a time-consuming test, which may affect the measured

Langmuir, Vol. 20, No. 6, 2004 2089

Figure 4. Frequency spectra of the storage (G′) and loss (G′′) moduli for the 5 wt % cellulose solution at 30 °C after 0 (b, O) and 7083 (2, 4) s.

results. Low-frequency sweeps (0.1-10 rad/s) were, thus, performed on the same system used in Figure 1 at the same temperature of 30 °C for times 0 and 7083 s. The frequency dependences of G′ and G′′ are shown in Figure 4. At a time of 0 s, the gelation of the solution has already started with G′ higher than G′′ in the part of the frequency range examined. For each frequency sweep, several minutes pass before the measurements begin, whereas the gelation time of the 5 wt % cellulose solution at 30 °C with an angular frequency 1 rad/s is 157 s as shown in Figure 3, implying that the gelation has already started before the measurements. The data at 7083 s exhibited characteristics of the gel state, similar to the results in Figure 1. Morphology and Structure of the Cellulose Gel. Figure 5 shows the 13C NMR spectra of 6 wt % cellulose in the solution and gel states. The peaks at 103.7, 79.0, 75.3, 73.8, and 60.7 ppm are assigned to the C1, C4, C5,3, C2, and C6 of cellulose, respectively. Compared with the solid-state 13C NMR spectrum of cellulose (cotton linters)1 shown in Figure 5, the peaks for C4 and C6 of cellulose in the solution and gel states obviously shifted to a higher magnetic field than those of the original cellulose, similar to cellulose in LiCl-DMAc,24 which is a true solution of cellulose. Moreover, there is no obvious difference in the chemical shifts or peak shapes between the 13C NMR spectra of the solution and those of the gel states,1,25 indicating that the chemical environments of carbon atoms on the macromolecular chain practically remain the same during the transition. In other words, the networks at the gel state were caused by a physical interaction junction rather than by a covalent bond.1,25 The presence of noncovalent cross-links complicates any physical description of the network properties enormously because their number and position fluctuate with time and temperature, which has been reflected in the dependence of the storage modulus on the temperature, time, and concentration in Figures 2 and 3. The consistent evidence of the physical junction structure can be seen on the WAXD pattern of 5 wt % cellulose/ NaOH/thiourea at the gel state, as shown in Figure 6. Compared with the original cellulose, the gel did not exhibit any crystallinity, namely, there are no crystals, indicating a random junction of the intra- and intermol(24) McCormick, C. L.; Callais, P. A.; Hutchinson, B. H. Macromolecules 1985, 18, 2394. (25) Cuculo, J. A.; Smith, C. B.; Sangwatanaroj, U.; Stejskal, E. O.; Sankar, S. S. J. Polym. Sci., Part A: Polym. Chem. 1994, 32, 229.

2090

Langmuir, Vol. 20, No. 6, 2004

Weng et al.

Figure 6. WAXD patterns of the hydrogel from the 5 wt % cellulose solution and the original cellulose.3

Figure 5. 13C NMR spectra of the 6 wt % cellulose in the NaOH/ thiourea/D2O solution and gel-state and solid-state CP/MAS 13 C NMR spectra of cellulose.3

ecules rather than an ordered packed chain structure in the gel state. In addition, the polarization microscopy examination of the cellulose solution at the pseudogel state was also conducted, and no signal of birefringence was observed, indicating there is no liquid crystal in the gel. In view of the results mentioned previously, physical crosslinks exist in the cellulose gel. ESEM is a powerful tool for the investigation of polymers in a number of novel and previously inaccessible situations.26 The ESEM photographs of the gels from three cellulose solutions are shown in Figure 7. The gel from the 4 wt % cellulose solution as shown in Figure 7a exhibited a certain extent of phase separation, which appeared as irregularly shaped particles of about 4 µm with some even connected together. The particles packed loosely with some void. With an increase of the cellulose concentration, the gels aggregated together to form a gel lump for the 5 wt % solution in Figure 7b and 6 wt % solution in Figure 7c. This is also attributed to the fact that the physical junction from cellulose self-association in the solution strengthened with an increase in the concentration, similar to hydroxypropylcellulose (HPC) solution, in which HPC particles would tend to aggregate together as the solution concentration increases.27,28 (26) Karlsson, J. O.; Andersson, M.; Berntsson, P.; Chihani, T.; Gatenholm, P. Polymer 1998, 39, 3589.

For further investigating the gel, a precise characterization technique, AFM, was employed to analyze the gel structure. Figure 8a shows the top view image of the height scan of the cellulose gel spread on the substrate after being dispersed in pure water. The visualized gels exhibited a structure of nanoparticles in size, and the average dimension of the cross section for those nanoparticles is 47.7 ( 5.9 nm. The particles packed together to form a peak-valley structure, which has also been observed by Matsumura et al.29 in studying the partially esterified pulp fibers. Although the probable orientation flow or extension occurring during sample preparation may have some effects on the dimensional shapes of the gel,30 the observed gels were generally composed of nanoparticles that connected to form a peak-valley structure, confirming that the formation of the cellulose gel involves the aggregation of small network units. Furthermore, the results of ESEM also revealed connecting particle structures as shown in Figure 7a, while there is no shearing motion during sample preparation. The height profile of the surface along the cross section of the top view image was also shown in Figure 8a. The peak-valley height, as indicated, is 2.36 nm, while the average statistic crosssection size of the nanoparticles is 47.7 ( 5.9 nm, indicating that those nanoparticles are relatively flat. The peaks and valleys became more apparent in the three-dimensional image shown in Figure 8b. Moreover, AFM is also performed on the cellulose gels dispersed in the NaOH/ thiourea aqueous solution diluted by copious water. As shown in Figure 9, the structure and size of the gel particles are almost the same as those in the pure water, and the average dimension of the cross section for those nanoparticles is 45.9 ( 8.5 nm, similar to that in pure water. Moreover, it is clear that those particles packed loosely together to form gel aggregates. This revealed that one nanoparticle represents a stable network unit, which further packed to form a gel lump. Interestingly, the mean (27) Gao, J.; Haidar, G.; Lu, X. H.; Hu, Z. B. Macromolecules 2001, 34, 2242. (28) Hirsh, S. G.; Spontak, R. J. Polymer 2002, 43, 123. (29) Matsumura, H.; Glasser, W. G. J. Appl. Polym. Sci. 2000, 78, 2254. (30) Ikeda, S. Food Hydrocolloids 2003, 17, 399.

Thermal Gelation of Cellulose/NaOH/Thiourea Solution

Langmuir, Vol. 20, No. 6, 2004 2091

Figure 7. ESEM images of the concentrated solutions of cellulose after heating over their opaque temperature point: (a) 4, (b) 5, and (c) 6 wt %.

Figure 8. AFM image of the gel from the 5 wt % cellulose solution spread on the substrate after being dispersed in pure water: (a) top-view image of the height scan and height profile analysis; (b) three-dimensional topography image.

size of smallest network unit (≈47 nm) is much smaller than the true contour length (L) of cellulose molecules with a molecular weight (Mη) of 8.0 × 104 (about 230 nm for L), which was roughly estimated from the molar mass per unit contour length (ML ) 350 nm-1) of cellulose in 6 wt % NaOH/4 wt % urea aqueous solution,31 according to the definition (L ) M/ML) for a wormlike chain. Note that the Mη value of the cellulose regenerated from the 6 wt % NaOH/5 wt % thiourea aqueous solution was determined by viscosmetry to be 8.0 × 104. The difference of the chain conformation in the solution proved that the intramolecular hydrogen bonds of the cellulose, which sustain the chain stiffness, were destructed, leading to a change from semistiff to very flexible chains (or random coil) in this system at a high temperature. At the same time, new intra- and intermolecular association was created by the self-association force of cellulose to form the relatively stable network units. Process and Mechanism of Thermally Induced Gelation. There are strong intra- and intermolecular hydrogen bonds in cellulose, and the self-association (31) Zhou, J.; Zhang, L.; Cai, J. J. Polym. Sci., Part. B: Polym. Phys. 2004, 42, 347.

tendency of hydroxyl on the cellulose chain is much larger than association between cellulose and water molecules, so cellulose is insoluble in water. However, in the cellulose/ NaOH/thiourea aqueous solution system, the presence of NaOH can break the intermolecular hydrogen bonds of cellulose to create a significant ion-pair interaction, which gives a chance for the water or thiourea molecules to form hydrogen bonds with the released -OH groups on the cellulose chains. The thiourea with highly polar CdS and NH2 groups serves as a hydrogen bonding donor and receptor between thiourea and cellulose to form new intermolecular interactions, which lead to the dissolution of cellulose in the aqueous solution.1 Therefore, sodium ions, water, and thiourea molecules could form an overcoat surrounding the cellulose chain to prevent its selfassociation, resulting in the cellulose to be water soluble. At a low temperature, the solution was relatively stable because the self-association was prevented by the formation of an “overcoat” structure. At an elevated temperature, the overcoat layer was broken to expose the hydroxyl groups of the cellulose to each other as a result of the rapid thermal escape motion of small molecules, leading to the physical cross-linking networks caused by the self-

2092

Langmuir, Vol. 20, No. 6, 2004

Weng et al.

Figure 9. AFM image of the gel from the 5 wt % cellulose solution spread on the substrate after being dispersed in the NaOH/ thiourea aqueous solution: (a) top-view image of the height scan; (b) three-dimensional topography image.

association force of the cellulose macromolecules, in which the exposed hydroxyl serves as junctions. The stable network units were, thus, formed by the random association of cellulose chains and further aggregated to form gels. With an increase in the temperature, more junctions of cellulose chains would be formed, resulting in a loss of the light quantity due to the increase of the particle size in the solution (turbid) and the formation of gel networks, accompanied by the sharp increase in the modulus and cross-linking density. In addition, it can also be explained from the thermodynamic analysis that the self-association of cellulose molecules in an aqueous solution happened only when the free energy change (∆G) is negative at a given constant temperature T by ∆G ) ∆H - T∆S < 0, where ∆H and ∆S are the changes of enthalpy and entropy, respectively. When the cellulose self-associated in the aqueous solution, the change in the entropy (∆S) should be negative as a result of the restriction of molecular chains in aggregates. If the gelation process can take place, the enthalpy should be more negative to meet the requirement of ∆G < 0. However, the gelation occurred at a high temperature with the endothermic heat ∆H > 0, as confirmed by the preliminary results of microdifferential scanning calorimetric analysis (not shown) measured with Setaram Micro-DSC III (Scientex Pty., Ltd.), which revealed a relatively broad endothermic peak. Furthermore, the association of the hydroxyl groups is an exothermic process, that is, ∆H < 0. Such an interesting question can only be resolved by consideration of the small molecules including water, thiourea, and some ions. When the cellulose dissolved in the aqueous solution, the small molecules were fixed on the cellulose chain as an overcoat to give some degree of order in the solution. Heating will supply the activation energy to destroy the overcoat between the small molecules and cellulose, resulting in the self-association of cellulose. The destruction of the overcoat is an endothermic process. Compared with the energy needed in the destruction of the overcoat structure, the energy for the formation of self-association (its intrinsic tendency) is much lower, so the gelation process is an endothermic process (∆H > 0) on the whole. Meanwhile, the release of small molecules from the chain gives an increase in the entropy of the system. Such an entropydriven process resulted in the self-association of cellulose in the aqueous solution upon heating. In addition, as shown in Figure 3, the systems are not in actual equilibrium

because of the strong self-association tendency of the cellulose chain and instability of the overcoat structure. Onthebasisoftheinformationobtainedfromrheometry,13C NMR, ESEM, AFM, and WAXD, a schematic process describing the formation of the gel in the cellulose solution is presented in Figure 10. The cellulose chains surrounded with small molecules such as thiourea and NaOH as an overcoat were dispersed in the aqueous solution at a molecular level below 20 °C, as shown in Figure 10A. When part of the overcoat of the small molecules on the cellulose chains was perturbed by raising the temperature, the random junction between the cellulose molecules having exposed hydroxyls occurred on the same chains as a result of the self-association force, resulting in the formation of a network, leading to turbidity of the solution (Figure 10B). At temperatures higher than 50 °C, a number of junction points between the cellulose chains occurred to form randomly cross-linked networks. This results in a transition from the solution to the gel state as shown in Figure 10C. As confirmed by 13C NMR, there is no chemical change during the gelation process, indicating that the gelation results from a physical interaction. In this case, the exposed -OH groups of cellulose play an important role in the formation of the cross-link network structure by random self-association. The self-association is weak at a relatively low temperature as a result of the shield effect of the overcoat, but it is strengthened at the elevated temperatures, leading to the formation of gels. Conclusions The thermally induced gels from the random association of concentrated cellulose solutions in 6 wt % NaOH/5 wt % thiourea were successfully prepared. It is revealed that the sol-gel transition of the cellulose solution with concentrations from 6 to 4 wt % occurred in the temperature range from 20 to 40 °C. The results of rheometry showed that the temperature at which the concentrated solution turned into a gel decreased with an increase of the cellulose concentration. The gelation process contains three typical regions, in the second one the storage modulus G′ increased exponentially with the temperature, and finally the cellulose solution changed into gels over 50 °C. At any temperature slightly lower than the gelation temperature, gelation could take place as long as enough time was given. ESEM indicated that the gel from

Thermal Gelation of Cellulose/NaOH/Thiourea Solution

Langmuir, Vol. 20, No. 6, 2004 2093

Figure 10. Schematic gelation process of the cellulose solution below 20 °C (A), at 25-40 °C (B), and above 50 °C (C).

4 wt % cellulose solution exhibited a structure of packed particles, while the gels from 5 and 6 wt % cellulose solutions aggregated to a lump structure, respectively. Moreover, either at a higher temperature or with a longer gelation time, the cellulose gel could form. The AFM images revealed that the cellulose gel was composed of nanoparticles with a diameter of about 47 nm as a relatively stable network unit. The gelation was caused by physical cross-linking of the cellulose chains in the solution at elevated temperature, that is, self-association of macromolecules having exposed hydroxyls. The heating induced the further destruction of the intra- and intermolecular hydrogen bonds of cellulose in the solution and creation of new random association on the same chains or different chain to form cross-linking structure at the same time.

Acknowledgment. This work was supported by the National Natural Science Foundation of China (Grants 59933070, 20204011, and 20174048), the National High Technology Research and Development Program of China (863 Program; 2003AA333040) and the Foundation of State Open Laboratory of Polymer Physics (Grant 02-B16). We thank Prof. Chi Wu (The Chinese University of Hong Kong) for valuable comments. We also thank Mr. Zhuchuan Li and Mrs. Manjun Shao (Multiphase Reaction Laboratory, Institute of Chemical Metallurgy, Chinese Academy of Sciences, Beijing, China) for assistance in ESEM measurements and Mrs. Chuanfeng Zhu and Prof. Lijun Wan (CAS key laboratory of Molecule Nanostructure and Nanotechnology, Beijing, China) for assistance in AFM measurements. LA035995O