3154
Langmuir 2006, 22, 3154-3160
Surface Forces Measurements of Spin-Coated Cellulose Thin Films with Different Crystallinity Shannon M. Notley,*,† Malin Eriksson,‡ Lars Wågberg,‡ Stephanie Beck,§ and Derek G. Gray§ Department of Applied Mathematics, Research School of Physical Sciences and Engineering, Australian National UniVersity, Canberra 0200 ACT, Australia, Department of Fiber and Polymer Technology, Royal Institute of Technology (KTH), SE-10044 Stockholm, Sweden, and Department of Chemistry, Pulp and Paper Research Centre, McGill UniVersity, 3420 UniVersity Street, Montreal, QC, Canada ReceiVed October 26, 2005. In Final Form: January 15, 2006 A systematic study of the surface forces between a cellulose sphere and cellulose thin films of varying crystallinity has been conducted as a function of ionic strength and pH. Semicrystalline cellulose II surfaces and amorphous cellulose films were prepared by spin-coating of the precursor cellulose solutions onto oxidized silicon wafers before regeneration in water. Crystalline cellulose I surfaces were prepared by spin-coating wafers with aqueous suspensions of sulfate-stabilized cellulose I nanocrystals. These preparation methods produced thin, smooth films suitable for surface forces measurements. The interaction with the cellulose I was monotonically repulsive at pH 3.5, 5.8, and 8.5 and at 0.1, 1, and 10 mM ionic strengths. This was attributed to the presence of strongly ionizable sulfur-containing groups on the cellulose nanocrystal surfaces. The amorphous film typically showed a steric interaction up to 100 nm away from the interface that was independent of the solution conditions. A range of surface forces were successfully measured on the semicrystalline cellulose II films; attractive and repulsive regimes were observed, depending on pH and ionic strength, and were interpreted in terms of van der Waals and electrostatic interactions. Clearly, the forces acting near cellulose surfaces are very dependent on the way the cellulose surface has been prepared.
Introduction The detailed surface interactions between cellulose surfaces are of great importance in papermaking and also in the textile and composites industries.1 These surface forces interactions are poorly understood, yet form the basis for a range of surface modifications that are often empirically used to alter or improve the performance of many cellulose-based products. The major reason for this lack of understanding is the inherent difficulty in conducting experiments on real fiber systems due to chemical and structural heterogeneities of the fibers. This has led to an increasing interest in developing well-defined model systems specifically for surface forces measurements. Along with the attractive interactions that are ubiquitous in nature,2 repulsive interactions may also be expected to act between fiber surfaces in aqueous solution. The presence of charged groups on the fiber surfaces and within the fiber wall (originating from other species naturally occurring in the wood fiber such as hemicellulose and lignin or those formed due to pulping procedures3) can result in a repulsive force when the respective electrical double layers overlap. Factors affecting the magnitude of this electrostatic repulsion include the surface charge or potential and ionic strength of the aqueous solution.4,5 These attractive and repulsive interactions can be characterized by * Corresponding author. Phone: +61 261257583. Fax: +61 261250732. E-mail:
[email protected]. † Australian National University. ‡ Royal Institute of Technology. § McGill University. (1) Wood and Cellulosic Chemistry; Hon, D. N., Shiraishi, N., Eds.; Marcel Dekker: 1991. (2) Israelachvili, J. Intermolecular and Surface Forces, 2nd ed.; Academic Press: 1991. (3) Chemical Modification of Lignocellulosic Materials; Hon, D. N.-S., Ed.; Marcel Dekker: 1991. (4) Deryagin, B.; Landau, L. Acta Physicochim. URSS 1941, 14, 633-662. (5) Verwey, E. G. W.; Overbeek, J. T. G. Theory of the Stability of Lyophobic Colloids; Elsevier: New York, 1948.
employing the DLVO theory of colloidal stability provided that the interaction geometry is accurately known.6 A number of recent studies have involved the measurement of surface forces in cellulosic systems. Both the surface forces apparatus (SFA)7-9 and colloidal probe technique10-16 have been used to study these interactions. Furthermore, many and varied techniques have been employed to prepare model cellulose thin films and spheres that were suitable for these measurements.7,17-21 For the most part, these cellulose surfaces have been regenerated from precursor cellulose derivatives such as trimethylsilylcellulose or from solutions of cellulose in solvents such as lithium chloride/ dimethylacetamide (LiCl/DMAc) or amine oxides such as N-methylmorpholine oxide (NMMO). Regeneration techniques lead to changes in the crystal structure of the native cellulose (6) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature 1991, 353, 239. (7) Neumann, R. D.; Berg, J. M.; Claesson, P. M. Nord. Pulp Pap. Res. J. 1993, 8, 96-104. (8) Holmberg, M.; Berg, J.; Stemme, S.; Odberg, L.; Rasmusson, J.; Claesson, P. J. Colloid Interface Sci. 1997, 186, 369-381. (9) Osterberg, M. J. Colloid Interface Sci. 2000, 229, 620-627. (10) Rutland, M. W.; Carambassis, A.; Willing, G. A.; Neumann, R. D. Colloids Surf., A 1997, 123-124, 369-374. (11) Carambassis, A.; Rutland, M. Langmuir 1999, 15, 5584-5590. (12) Zauscher, S.; Klingenberg, D. J. J. Colloid Interface Sci. 2000, 229, 497510. (13) Notley, S. M.; Petterson, B.; Wågberg, L. J. Am. Chem. Soc. 2004, 126, 13930-13931. (14) Notley, S. M.; Wågberg, L. Biomacromolecules 2005, 6, 1586-1591. (15) Notley, S. M.; Wågberg, L. In Transactions of 13th Fundamental Research Symposium of the Pulp and Paper Fundamental Research Society; Cambridge, U.K., 2005; pp 1337-1350. (16) Nigmatullin, R.; Lovitt, R.; Wright, C.; Linder, M.; Nakari-Setala, T.; Gama, M. Colloids Surf., B 2004, 35, 125-135. (17) Gunnars, S.; Wågberg, L.; Cohen-Stuart, M. A. Cellulose 2002, 9, 239249. (18) Edgar, C. D.; Gray, D. G. Cellulose 2003, 10, 299-306. (19) Fa¨lt, S.; Wågberg, L.; Vesterlind, E. L.; Larsson, P. T. Cellulose 2004, 11, 151-162. (20) Eriksson, J.; Malmsten, M.; Tiberg, F.; Honger Callisen, T.; Damhus, T.; Johansen, K. S. J. Colloid Interface Sci. 2005, 284, 99-106. (21) Freudenberg, U.; Zschoche, S.; Simon, F.; Janke, A.; Schmidt, K.; Behrens, S. H.; Auweter, H.; Werner, C. Biomacromolecules 2005, 6, 1628-1634.
10.1021/la052886w CCC: $33.50 © 2006 American Chemical Society Published on Web 02/22/2006
Cellulose Thin Films Surface Forces Measurements
and, in some cases, complete removal of crystallinity. Another technique for preparing cellulose thin films has been by spincoating a colloidal sol of cellulose nanocrystals prepared through acid hydrolysis of wood pulp.18 With the use of this preparation procedure, the native cellulose crystal structure (cellulose I), was retained. The measured forces when using amorphous cellulose surfaces such as those prepared from the Langmuir-Blodgett technique or using cellulose spheres prepared by regeneration from LiCl/ DMAc have typically been dominated by a steric interaction extending up to 100 nm away from the interface.7,11 This has been interpreted using a “dangling tail” model where a few chains from the highly swollen film extend significantly away from the surface. Since cellulose is insoluble in water it is still not clear which structural element of the model surface is causing this steric interaction. Outside of this steric region, Rutland and coworkers10,11 have shown that the measured interaction profile can be reasonably fit by using DLVO theory and, furthermore, that the steric interaction was dependent on the ionic strength of the solution. The omnipresent nature of the steric force at short surface separations, however, complicates the analysis if other surface forces such as dispersion forces or those due to the addition of polymer are of greater interest. Furthermore, the high swelling ability of these amorphous films, due to Donnan equilibrium effects, leads to poor mechanical properties when exposed to aqueous electrolyte solutions. This all but eliminates the possibility of measuring contact forces such as those due to adhesion and friction. Cellulose I thin films were recently used in surface forces measurements.22 In this study, an unmodified AFM cantilever was used which led to an asymmetrical interaction (cellulose vs silicon nitride). Furthermore, the very small magnitude of the observed forces, due to the small radius of the cantilever tip, limited the ability to draw accurate conclusions on the apparent potential energy of interaction. However, it was possible to qualitatively observe differences in the force-distance curves due to the addition of adsorbing polymer, specifically in terms of the increase in adhesion upon tip retraction due to bridging between the cellulose and silicon nitride surfaces.23 Model spin-coated cellulose II (the thermodynamic crystalline cellulose allomorph) films have been prepared and used to measure surface forces.13,14 These films can be characterized as both crystalline and amorphous in nature. The solution of cellulose in NMMO is not a true solution as such, but rather it is a fringe micellar type arrangement with some aggregates present. Furthermore, dissolution and subsequent regeneration in water results in a change in the crystal structure to the more thermodynamically favorable cellulose II allomorph.24 With the use of these cellulose II surfaces, dispersion forces under conditions where electrostatic interactions are unfavorable were measured.13 From the force-distance data, a Hamaker constant for cellulose interacting in water was determined, 3 × 10-21 J, that agreed well with previous estimations from scanning ellipsometry measurements.25 In another study, carboxyl groups were introduced into the pulp prior to solution formations and spin-coating in order to increase the charge over native cellulose hence mimicking the surface chemistry of real fibers.14 By choosing appropriate solution conditions in the force-distance (22) Lefebvre, J.; Gray, D. G. Cellulose 2005, 12, 127-134. (23) Claesson, P. M.; Poptoshev, E.; Blomberg, E.; Dedinaite, A. AdV. Colloid Interface Sci. 2005, 114-116, 173-187. (24) Fink, H.-P.; Weigel, P.; Purz, H. J.; Ganster, J. Prog. Polym. Sci. 2001, 26, 1473-1524. (25) Bergstro¨m, L.; Stemme, S.; Dahlfors, T.; Arwin, H.; O ¨ dberg, L. Cellulose 1999, 6, 1-13.
Langmuir, Vol. 22, No. 7, 2006 3155
experiment, the interaction potential could be tuned to give either attractive or repulsive surface forces. No systematic comparison of the observed surface forces as a function of cellulose film preparation technique has previously been undertaken. In this study, three types of spin-coated cellulose films, cellulose I films prepared from nanocrystals, cellulose II from the regeneration of a NMMO cellulose solution, and amorphous cellulose films prepared by regeneration from LiCl/ DMAc, have been used to measure the forces of interaction as a function of pH and ionic strength. The colloidal probe technique was used in this study with an amorphous cellulose sphere. These spheres have been extensively characterized and treated to minimize swelling which could result in the presence of an unwanted short-range steric interaction. Through comparison of the measured interaction forces between cellulose surfaces using similar raw materials for the film preparation any differences due to film crystallinity will be observed. Materials and Methods Materials. A dissolving grade pulp from Domsjo¨ Fabriker, Sweden, was used as the raw material for the preparation of the amorphous cellulose thin films and the cellulose II thin films. A similar northern softwood dissolving grade pulp, Temalfa 93 (Tembec Inc., Temiscaming, Canada) was used to make the cellulose I nanocrystal suspension. Glyoxolated polyacrylamide (G-PAM, Parez 631 NC) or polyvinylamine (PVAm) was used as the cationic anchoring polymer and was supplied by Cytec and BASF Germany, respectively. NMMO (N-methylmorpholine N-oxide) was supplied as a 50% w/w aqueous solution by Aldrich, Sweden, while dimethyl sulfoxide (DMSO) was supplied by KeBo, Sweden. NaCl, HCl, and NaOH were all of analytical grade and were used as supplied by Aldrich, Sweden. Lithium chloride/dimethylacetamide solution was used as the solvent in the preparation of amorphous cellulose surfaces, and these chemicals were supplied by Sigma-Aldrich. Milli-Q water was used in all experiments. Amorphous cellulose spheres, prepared from lithium chloride/dimethylacetamide solution, were provided by MonoGel AB, Helsingborg, Sweden. The spheres were solvent exchanged to ethanol before a light heat treatment at 40 °C for 2 h to remove the solvent prior to attachment to the cantilever. Reverse imaging of the cellulose spheres using atomic force microscopy was performed to determine the surface roughness.26 Cellulose Surface Preparation. A colloidal suspension of cellulose nanocrystals was prepared by acid hydrolysis of a dissolving grade pulp by a previously described method.27 The colloidal sol of cellulose nanocrystals, stabilized by the presence of surface sulfate groups with sodium counterions and exhaustively deionized, was used to prepare cellulose I surfaces using a modified procedure based on the method of Edgar and Gray.18 A 3% w/w of this colloidal suspension of cellulose I nanocrystals was spin-coated onto a silica wafer pretreated with a cationic polyelectrolyte at 4000 rpm for 1 min. The films were subsequently heat treated at 90 °C for 4 h to ensure the films did not delaminate upon exposure to aqueous electrolyte solution. This procedure produced cellulose I films of approximately 120 nm thickness according to ellipsometry measurements (Beaglehole Instruments, New Zealand). Paracrystalline cellulose II thin films were prepared according to a previously developed method.17 A cellulose II solution was first prepared by dissolving 0.5 g of pulp in 25 g of NMMO at 115 °C for approximately 2 h. An amount of 75 g of DMSO was then added to the solution in order to decrease the viscosity before spin-coating to give a final cellulose solution concentration of 0.5% w/w. Films were spin-coated at up to 3500 rpm for 30 s onto oxidized silicon wafers pretreated with an anchoring layer G-PAM. This produced films with an approximate thickness of 30 nm, as measured using ellipsometry, while minimizing the surface roughness. (26) Neto, C.; Craig, V. S. J. Langmuir 2001, 17, 2097-2099. (27) Dong, X. M.; Kimura, T.; Revol, J.-F.; Gray, D. G. Langmuir 1996, 12, 2076-2082.
3156 Langmuir, Vol. 22, No. 7, 2006 Amorphous cellulose films were prepared from cellulose dissolved in a solution of lithium chloride in dimethylacetamide (DMAc).The cellulose dissolution procedure was adapted without the derivatizing agent after Berthold et al.28 and the film preparation after Eriksson et al.20 An amount of 0.5 g of acetone-extracted dissolving grade pulp was initially immersed in Milli-Q water to allow the pulp to swell. After 24 h, the pulp suspension was filtered to remove most of the water. The pulp was then placed into methanol for 30 min with stirring before filtering. This was repeated three times. The pulp was then solvent exchanged to DMAc by immersion for 30 min. This was repeated three times with filtration between each step. An amount of 1.5 g of LiCl was added to 18 mL of DMAc which was heated to 110 °C. When a clear solution was obtained, the solution was removed from heat and allowed to cool. The solventexchanged pulp was then added to the solution of LiCl in DMAc in small portions. The pulp was allowed to dissolve at room temperature for 24 h after which a clear, colorless, but highly viscous cellulose solution was formed. This solution was subsequently diluted with 80 mL of DMAc (a final cellulose concentration of 0.5% w/w) and heated to a temperature of 110 °C before spin-coating onto an oxidized silicon wafer pretreated with a cationic anchoring polymer layer. These cellulose surfaces were then placed into Milli-Q water to remove any excess LiCl and solvent before being blown dry with nitrogen. The dry thickness of the amorphous cellulose thin film was measured to be 44 nm using ellipsometry. Silica Surface Preparation. Silica has been extensively studied previously in surface forces applications.29-32 An oxidized silicon wafer was used as a flat substrate to determine the suitability of the cellulose spheres in these surface forces measurements. The wafers were supplied by Peregrine Semiconductors Pty Ltd. (Sydney, Australia) and had a native oxide layer of 4 nm as measured by ellipsometry. Prior to use, the silica surface was rinsed in 10% w/w aqueous NaOH for 30 s followed by rinsing with copious amounts of Milli-Q water before being blown dry with N2. As a final step, the surface was subjected to a mild water plasma treatment to ensure the silica substrate was clean and wetting. These cleaning treatments did not appreciably increase the surface roughness of the silica substrate. Surface Forces Measurements. The interaction between a cellulose sphere and the three different cellulose thin films was measured using colloidal probe microscopy. A Multimode Scanning Probe microscope (Veeco Ltd., Santa Barbara, CA) was used for all force experiments conducted in this study. A cellulose sphere (of size between 5 and 20 µm) was attached to the end of an AFM cantilever with a small amount of epoxy adhesive according to the method of Ducker et al.6,29 Standard, tipless, contact SiO2 cantilevers (Mikromasch Tallinn, Estonia) were used for force measurements in this study. The data were typically collected using the 250 µm long, rectangular-shaped cantilever with a spring constant of 0.6 N/m as measured by the thermal noise method.33 Typically, the flat surface is ramped toward the sphere-modified cantilever by the expansion of a piezoelectric scanner. Surface forces cause a measurable deflection of the cantilever. The deflection of the cantilever is converted to a force by Hookes’ law with the aid of the determined spring constant. The point of zero separation is defined as the onset of the constant compliance region, where a linear deflection of the cantilever is measured due to the linear motion of the flat substrate. This constant compliance region is also used to determine the optical sensitivity of the system. Once a predetermined load is reached, the motion of the flat substrate is reversed and the adhesion between the surfaces can be measured. Zero force is defined as when the motion of the flat surface does not cause a (28) Berthold, F.; Gustafsson, K.; Berggren, R.; Sjoholm, E.; Lindstrom, M. J. Appl. Polym. Sci. 2004, 94, 424-431. (29) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmuir 1992, 8, 18311836. (30) Hartley, P. G.; Larson, I.; Scales, P. J. Langmuir 1997, 13, 2207-2214. (31) Notley, S. M.; Biggs, S.; Craig, V. S. J.; Wågberg, L. Phys. Chem. Chem. Phys. 2004, 6, 2379-2386. (32) Toikka, G.; Hayes, R. A. J. Colloid Interface Sci. 1997, 191, 102-109. (33) Hutter, J. L.; Bechhoefer, J. ReV. Sci. Instrum. 1993, 64, 1868-1873.
Notley et al.
Figure 1. AFM tapping mode image of the interacting area of a cellulose sphere used in the measurement of surface forces. The left image is 1 µm × 1 µm, and the right image is 10 µm × 10 µm, with a z contrast of 50 and 100 nm, respectively. The root-mean-square (rms) surface roughness of the sphere was 5.9 nm over a 1 µm2 image after plane-fitting to remove the bow due to the curvature of the sphere. measurable deflection of the cantilever. Both the forces on approach and retraction of the surfaces were measured; however, only the approach data will be discussed in this paper. At least 20 forcedistance curves were measured, and the data presented in this work are representative of these curves. Measured forces are normalized by the probe radius for comparison to theoretical fits of the surface force by the Derjaguin approximation.34 The Derjaguin approximation relates the energy per unit area for two planar surfaces, EPP, to the energy between a sphere and a flat plate by eq 1, where FSP is the surface force and r is the radius of the sphere. EPP )
1 FSP 2π r
(1)
The normalized force-distance data was fit to DLVO theory within the boundary limits of constant charge and constant potential for the solution of the nonlinear Poisson-Boltzmann equation.35 The procedure allowed fitting of the asymmetric surface potentials for the two different surfaces as a function of ionic strength.
Results Previous surface forces measurements involving cellulose have often been limited by the roughness or swelling of either the cellulose films or spheres or both. Thus, it is important to have a thorough understanding of these properties of the films and spheres in order to correctly interpret the observed surface forces measurements. The roughness (root-mean-square) of the three differently prepared cellulose surfaces was measured here and was typically less than 2 nm over 1 µm2 as determined by atomic force microscopy (AFM) imaging. Furthermore, AFM reverse imaging in air was used to determine the roughness of the interacting area of the cellulose sphere used in these surface forces measurements. Figure 1 shows an example image. It is interesting to note that the roughness of the sphere is significantly greater than that of the flat cellulose thin films. However, these cellulose spheres have previously been successfully used to determine van der Waals type interactions.13 Interaction Forces between a Cellulose Sphere and Silica Flat Substrate. The interaction between a cellulose sphere and a silica substrate was measured as a function of pH and ionic strength. Figure 2 shows the measured force-distance data for the interaction between a cellulose sphere and an oxidized silicon wafer at pH 3.5 and pH 8.5 in an aqueous background electrolyte solution of NaCl with a concentration of 1 mM. The normalized force-distance curve for the data measured at pH 3.5 shows an (34) Derjaguin, B. V. Kolloid-Z. 1934, 69, 155. (35) Chan, D. Y. C.; Horn, R. G. J. Chem. Phys. 1985, 83, 5311.
Cellulose Thin Films Surface Forces Measurements
Figure 2. Normalized force-distance curves for the interaction of a cellulose sphere and silica flat substrate as a function of pH in a background aqueous electrolyte solution with a concentration of 1 mM. Open squares are for pH 3.5, and filled diamonds are for pH 8.5. The solid line is a fit to the nonretarded Hamaker equation with a Hamaker constant of 5 × 10-21 J.
Figure 3. Normalized force-distance curves for the interaction of a cellulose sphere and silica flat substrate at pH 8.5 as a function of ionic strength. Open squares are for 10 mM, and filled diamonds are for 1 mM. The data have been fit to DLVO theory in the limits of constant charge and potential (solid lines). The fitting parameters were κ-1 of 2 and 8 nm for the 10 and 1 mM ionic strengths, respectively.
attractive interaction that can be well described using the nonretarded van der Waals equation with a fitted Hamaker constant of 5 × 10-21 J and a nominal electrostatic contribution using a surface potential of -1 mV. By increasing the solution pH, the measured interaction becomes monotonically repulsive with a small jump into contact at apparent surface separations of less than 2 nm. This interaction curve was fit using DLVO theory between the constant charge and potential limits for the solution to the nonlinear Poisson-Boltzmann equation using surface potentials for silica and cellulose of -42 and -3 mV, respectively. The surface potential of silica has been measured previously using the colloidal probe technique, and the magnitude, -42 mV, was taken as constant under these solution conditions with the magnitude of the surface potential of the cellulose sphere varied to give the best fits. The Debye length, κ-1, was also used as a fitting parameter and was determined to be 8 nm which results from the increase in electrolyte concentration due to the adjustment of the solution pH. The low potential of the cellulose sphere is not surprising given that the raw material for preparing the spheres is a highly extracted pulp with little hemicellulose or other impurities meaning that under the solution conditions in this study, virtually no ionizable groups on the native cellulose are dissociated. Figure 3 shows the measured interaction forces between cellulose and silica as a function of ionic strength at pH 8.5. The measured force-distance curves can be satisfactorily fit using
Langmuir, Vol. 22, No. 7, 2006 3157
Figure 4. Normalized force-distance curves between a cellulose sphere and cellulose I flat surface at a background ionic strength of 0.1 mM as a function of pH. The data can be fit to DLVO theory with a measured Debye length of 32 nm; surface potentials are given in Table 1.
DLVO theory within the constant charge and potential limits with Debye lengths of 8 and 2 nm for the interactions in 1 and 10 mM NaCl, respectively. The data in Figures 2 and 3 show no evidence of a short- or long-range steric interaction which has often been observed using cellulose spheres from other sources, implying that the spheres used in these surface forces measurements are relatively nonswelling and nondeformable. Thus, any observation of nonDLVO type forces in the measured interaction between the cellulose sphere and cellulose flat surfaces should be due to the swelling of these flat cellulose surfaces and not of the sphere. Interaction Forces between a Cellulose Sphere and Cellulose Surfaces with Varying Crystallinity. The cellulose thin films have previously been characterized in terms of the surface chemistry and surface morphology. These extensive studies have included measurements of the surface roughness of the films which is typically below 2 nm root-mean-square over a 1 µm2 area as measured by atomic force microscopy tapping mode imaging. High surface roughness may result in difficulty measuring the point of zero contact in force-distance measurements hence making determinations of surface forces parameters such as surface potentials and Hamaker constants unreliable. Prior to interaction measurements, the surface roughness of the cellulose films was measured to ensure that the films were smooth and continuous. Cellulose I. Cellulose I surfaces were prepared by spin-coating a dilute sol of cellulose nanocrystals onto a silica wafer. A short heat treatment was used to stabilize the film. Subsequently, surface forces between the cellulose I film and cellulose sphere were measured as a function of pH and ionic strength. Figure 4 shows the normalized force-distance interaction at an ionic strength of 0.1 mM as a function of pH. All of the interaction curves are monotonically repulsive indicating that the component of the interaction potential due to the overlap of the electrical double layers of these surfaces dominates over the dispersion component. The data can be well fit to DLVO theory with a typical Debye length of 32 nm. The fitted surface potentials for all solution conditions probed in this study are summarized in Table 1. It is interesting to note that there is a general trend for the measured surface potential to increase as a function of increasing ionic strength. This is somewhat surprising as the Grahame equation predicts that for a constant charge surface, the surface potential should decrease as a function of increasing ionic strength.2 This suggests that the cellulose I surface has some swelling affinity as an increasing
3158 Langmuir, Vol. 22, No. 7, 2006
Notley et al.
Table 1. Fitted Cellulose I Surface Potentials from Surface Forces Measurements as a Function of pH and Ionic Strength ionic strength
pH
surface potential/mV
10-4 M
3.5 5.5 8.5
-25 -40 -52
10-3 M
3.5 5.5 8.5
-25 -30 -39
10-2 M
3.5 5.5 8.5
-49 -44 -50
surface potential could indicate that Donnan equilibrium effects must be taken into account, especially for any residual carboxylic groups in the bulk of the film.36 Figure 5 shows the interaction between the cellulose I film and the cellulose sphere at pH 8.5 as a function of ionic strength. Again, the data is well described using DLVO with the predicted Debye lengths.
Figure 6. Normalized force-distance curves for a cellulose sphere interacting with a cellulose II thin film in a background electrolyte concentration of 0.1 mM as a function of pH.
Figure 7. Normalized force-distance curves for a cellulose sphere and a cellulose II surface at pH 5.5 as a function of ionic strength. Open squares are for 0.1 mM, and filled diamonds are for 1 mM. Figure 5. Normalized force-distance curves for the interaction between a cellulose sphere and cellulose I flat surface at pH 8.5 as a function of ionic strength. The fitted Debye lengths were 32, 10.5, and 3.5 nm for 0.1, 1, and 10 mM, respectively.
Cellulose II. Cellulose II surfaces were prepared from the regeneration of a solution of dissolving grade pulp in NMMO from water. The solution of cellulose in NMMO is understood to be something like a fringe micellar type structure which retains some crystalline structure.24,37 This type of film structure contains both crystalline and amorphous regions providing a model surface resembling the crystalline/amorphous cellulose structures often found in nature. Figure 6 shows the interaction between a cellulose II flat surface and cellulose sphere at an ionic strength of 0.1 mM at varying pH. At low pH, the interaction force is attractive indicating that the cellulose film is essentially uncharged leading to the domination of the van der Waals component over that due to forces arising from electrostatic considerations. However, as the pH of the solution increases, the film’s charge increases resulting in a reversal of the interaction energy from attractive to repulsive. Interaction forces between the cellulose II surface and the amorphous cellulose sphere were also measured at different ionic strengths. Figure 7 shows the interaction at pH 5.5 at an ionic strength of 0.1 and 1 mM. At low ionic strength, the interaction is repulsive and is well fit using DLVO theory with a Debye length of 25 nm, and surface potential of the cellulose II surface was -13 mV. However, on increasing the ionic strength, the (36) Grignon, J.; Scallan, A. M. J. Appl. Polym. Sci. 1980, 25, 2829-2843. (37) Arndt, K.-F.; Morgenstern, B.; Ro¨der, T. Macromol. Symp. 2000, 162, 109-119.
Figure 8. Normalized force-distance curve for the interaction of a cellulose sphere and an amorphous cellulose flat surface at pH 3.5 and an ionic strength of 1 mM. No attempt was made to fit the data to DLVO theory.
interaction becomes attractive due to the screening of the electrical double layers resulting in the van der Waals component dominating the measured force. Amorphous Cellulose. Interaction forces were also measured between a cellulose sphere and amorphous cellulose flat surface prepared from spin-coating of a solution of cellulose in lithium chloride and dimethylacetamide. Figure 8 shows the measured force-distance curve at pH 3.5 and an ionic strength of 0.1 mM. The interaction is monotonically repulsive and is of a steric nature. Indeed, under all solution conditions investigated in this study, the interaction forces were characterized by a steric repulsion for the amorphous cellulose surface. This steric repulsion was somewhat variable, extending up to 100 nm away from the interface in some cases, and its origin is most likely due to the
Cellulose Thin Films Surface Forces Measurements
Langmuir, Vol. 22, No. 7, 2006 3159
significant swelling of the amorphous film. In a previous study, Eriksson et al. showed that an amorphous cellulose film prepared in a similar spin-coating manner swelled by more than 100% by thickness when exposed to aqueous electrolyte solutions.20
Discussion Previous measurements of the apparent surface forces between model cellulose surfaces have proven inconsistent. This is primarily due to the various methods used to prepare the model cellulose thin films. In this study, three differently prepared precursor fluids were used to probe the effect of the film crystallinity on the measured force-distance interaction for the spin-coated surfaces. It is interesting to note that under some solution conditions the surfaces behave very similarly but under others the measured interaction is quite different. The cellulose I films showed somewhat surprising results over the range of solution conditions. All force interaction curves were monotonically repulsive but were well described using DLVO theory. At low pH (3.5), the measured surface potential was quite small yet significant enough for the overlap of the respective electrical double layers of the sphere and flat surface to dominate over the dispersion contribution to the overall interaction potential. This is in accord the fact that the surface of the nanocrystals prepared by sulfuric acid hydrolysis contains some sulfate ester groups and that these groups are preserved during the formation of the films.18,38 The cellulose I allomorph is the most common crystalline structure in wood fibers, but the preparation procedure introduces sulfate ester groups, so caution must be exercised when interpreting results of surface forces measurements. Typically, the pKa of these groups is of the order of 1.9, meaning that, at the solution conditions in this study, they would be almost completely dissociated. It is also likely that the raw material in this study contains some carboxyl groups, in the form of uronic acids, that are retained in the formed film.39 This results in the increasing trend of the measured surface potential as a function of increasing pH, as the pKa of the carboxyl groups is 3.5. The interaction measurements using the cellulose sphere and cellulose II surfaces shown in Figures 6 and 7 perhaps have the greatest potential for clarifying interactions between cellulose surfaces. However, the cellulose II allomorph, while being the thermodynamically favored crystalline form, is less commonly found naturally. Under conditions where relatively longer range surface forces dominate (such as those due to overlap of electrical double layer or indeed dispersion forces when the interacting area is large), the difference in crystal structure between cellulose I and cellulose II is of less importance than overall variations in density, morphology, and swellability of the cellulose samples. These arise from the presence of amorphous regions in the film. The method of cellulose regeneration can greatly influence the degree of crystallinity of the film with even small changes in the temperature of the water bath resulting in significant lowering of the crystallinity. As it can be reasonably expected that any charged groups such as carboxyl groups will be located uniformly throughout the film, swelling pressures may vary considerably. Thus, Donnan equilibrium effects must be accounted for when altering solution conditions.36 However, crystalline regions should act to restrict this swelling. Even though this might complicate the use of the model surfaces, it is exactly mimicking the situation encountered in most cellulose-containing fibers as previously demonstrated.40 (38) Dong, X. M.; Revol, J.-F.; Gray, D. G. Cellulose 1998, 5, 19-32. (39) Fras, L.; Laine, J.; Stenius, P.; Stana-Kleinschek, K.; Ribitsch, V.; Dolecek, V. J. Appl. Polym. Sci. 2004, 92, 3186-3195. (40) Fa¨lt, S.; Wågberg, L.; Vesterlind, E. L. Langmuir 2003, 19, 7895-7903.
Figure 9. Comparison of the interaction force-distance curves between a cellulose sphere and the cellulose I, II, and amorphous cellulose surfaces at pH 3.5 and 0.1 mM aqueous NaCl solution. The interaction with the cellulose I sample is characterized by an electrostatic repulsion, while van der Waals forces and steric interactions predominate with the cellulose II and amorphous cellulose samples, respectively.
A range of different surface forces were measurable by choosing appropriate pH and ionic strengths that minimize these swelling pressures. At low pH and consequently low dissociation of carboxyl groups, the measured interaction was satisfactorily fit with the nonretarded Hamaker equation. The fitted Hamaker constant agreed well with previous determinations from surface forces measurements and spectroscopic ellipsometry which were in the range of (3-8) × 10-21 J.13,25 Furthermore, on increasing the pH of the solution, the interaction force could be tuned to give either an attractive or repulsive interaction depending on the ionic strength, in line with predictions from DLVO theory. The low overall charge together with some degree of crystallinity results in a stable cellulose film with sound mechanical properties that may be used for the measurement of these noncontact forces. Another situation arose with the cellulose film regenerated by aqueous coagulation of cellulose solutions in lithium chloride/ dimethylacetamide. The interaction potential between the relatively dense amorphous cellulose spheres and this cellulose film was quite distinct, indicating a steric repulsion over relatively large distances, independent of the solution conditions. It is likely with the absence of any crystallinity and despite a very low overall charge, that these films swell to a significant extent in aqueous electrolyte solutions as previously observed using films prepared by a similar method.20 Figure 9 shows a comparison of the measured interaction potentials for the three different cellulose surfaces used in this study at pH 3.5 and 1 mM electrolyte. Under these solution conditions, the force-distance curves are completely different. The cellulose I interactions are characteristic of an electrostatic repulsion, the cellulose II sample displays a combination of electrostatic and dispersion force interactions, while steric interactions predominate for the amorphous sample. Thus, the preparation procedure of the cellulose thin film clearly has a great influence on the measurable surface forces even though similar raw materials were used for each sample. In many practical applications such as in papermaking and in fiber-reinforced composites, it is likely that a range of surface forces are operating. The relative magnitude of these forces will hence influence the way in which wood-based fibers will come together to form joints in a paper sheet and the interaction between fibers and polymer matrixes in fiber-reinforced composites. While forces which promote adhesion such as those due to van der Waals interactions and hydrogen bonding are of high importance, other noncontact forces can greatly contribute to the strength of
3160 Langmuir, Vol. 22, No. 7, 2006
the paper sheet. For example, the ability of the wood fiber to swell is due in large part to the amorphous cellulose regions. Without these fiber-water interactions, the molecular contact area in a joint would be greatly reduced. Furthermore, a swollen gellike amorphous region is more mobile than the crystalline regions leading to the possibility of creating the favorable orientation necessary for the development of hydrogen bonds. Thus a complete understanding of the surface forces between both crystalline and amorphous cellulose is required to correctly predict the differences in strength due to changes in papermaking conditions.
Conclusions The colloidal probe technique was used to measure the interaction potential between an amorphous cellulose sphere and cellulose thin films prepared in three distinct ways. The characteristics of the force-distance curves depended strongly on the method used to prepare the thin films, with curves characteristic of van der Waals, electrostatic, and steric forces being observed for the different films. Amorphous films resulted
Notley et al.
in a predominantly steric interaction, while the crystalline films adequately agreed with predictions from DLVO theory. The use of cellulose I surfaces resulted in a monotonically repulsive interaction due to the charge introduced through the acid hydrolysis preparation procedure. The surface potential could be determined through fitting between the limits of constant charge and constant potential for the cellulose I film. The measured interaction when using semicrystalline cellulose II films could be tuned to be either attractive or repulsive in line with predictions from DLVO theory. At low pH the measured interaction potential was used to determine a Hamaker constant which agreed well with previous determinations. At high pH, the interaction was well described between the constant charge and constant potential limits. Acknowledgment. S.M.N. would like to acknowledge financial support from the Cooperative Research Centre for Functional Communication Surfaces (CRC SmartPrint). Fruitful discussions with Vincent Craig are also acknowledged. LA052886W
