pubs.acs.org/Langmuir © 2009 American Chemical Society
Nanoscale Cellulose Films with Different Crystallinities and Mesostructures;Their Surface Properties and Interaction with Water § † Christian Aulin,†,‡ Susanna Ahola, Takashi Nishino,# Yasuo Hirose,# :: Peter Josefsson, § Monika Osterberg, and Lars Wagberg*,† †
Fibre Technology, Royal Institute of Technology, Teknikringen 56, 100 44 Stockholm, Sweden, ‡BIM Kemi AB, Box 3102, SE-443 03 Stenkullen, Sweden, §Department of Forest Products Technology, Faculty of Chemistry and Materials Sciences, Helsinki University of Technology, P.O. Box 3320, FIN-02015 TKK, Espoo, Finland, and #Department of Chemical Science and Engineering, Graduate School of Engineering, Kobe University, Rokko, Nada, Kobe 657-8501, Japan Received January 26, 2009. Revised Manuscript Received March 9, 2009
A systematic study of the degree of molecular ordering and swelling of different nanocellulose model films has been conducted. Crystalline cellulose II surfaces were prepared by spin-coating of the precursor cellulose solutions onto oxidized silicon wafers before regeneration in water or by using the Langmuir-Schaefer (LS) technique. Amorphous cellulose films were also prepared by spin-coating of a precursor cellulose solution onto oxidized silicon wafers. Crystalline cellulose I surfaces were prepared by spin-coating wafers with aqueous suspensions of sulfate-stabilized cellulose I nanocrystals and low-charged microfibrillated cellulose (LC-MFC). In addition, a dispersion of high-charged MFC was used for the buildup of polyelectrolyte multilayers with polyetheyleneimine on silica with the aid of the layerby-layer (LbL) technique. These preparation methods produced smooth thin films on the nanometer scale suitable for X-ray diffraction and swelling measurements. The surface morphology and thickness of the cellulose films were characterized in detail by atomic force microscopy (AFM) and ellipsometry measurements, respectively. To determine the surface energy of the cellulose surfaces, that is, their ability to engage in different interactions with different materials, they were characterized through contact angle measurements against water, glycerol, and methylene iodide. Small incidence angle X-ray diffraction revealed that the nanocrystal and MFC films exhibited a cellulose I crystal structure and that the films prepared from N-methylmorpholine-N-oxide (NMMO), LiCl/DMAc solutions, using the LS technique, possessed a cellulose II structure. The degree of crystalline ordering was highest in the nanocrystal films (∼87%), whereas the MFC, NMMO, and LS films exhibited a degree of crystallinity of about 60%. The N,N-dimethylacetamide (DMAc)/LiCl film possessed very low crystalline ordering (99.9%, Sigma Aldrich). The cellulose dissolution procedure was adapted without the derivatizing agent of Berthold et al.,31 and the films were then prepared according to the method of Eriksson et al.14 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 in 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. The pulp was dissolved in DMAc/LiCl by adding 0.5 g of pulp to 18 g of DMAc, heating (31) Berthold, F.; Gustafsson, K.; Berggren, R.; Sjoeholm, E.; Lindstroem, M. J. Appl. Polym. Sci. 2004, 94, 424–431.
Langmuir 2009, 25(13), 7675–7685
Aulin et al. the mixture to 160 °C, and then leaving the solution to cool to 100 °C, after which 1.5 g of LiCl was added. The mixture was then left to cool to 25 °C under stirring overnight, and this resulted in a clear, highly viscous, cellulose solution. 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 PEI anchoring polymer layer. These cellulose surfaces were then placed in Milli-Q water to remove any excess LiCl and solvent before being blown dry with nitrogen. Langmuir-Schaefer (LS) Cellulose Films. Substrates used for LS film preparation were either QCM-D crystals, which were spin-coated with polystyrene by the supplier (Q-Sense AB), or smooth silica wafers that were hydrophobized by immersion in 0.04% dichloromethylsilane in xylane for 2 h, rinsed with xylane, and dried with nitrogen gas. Trimethylsilyl cellulose (TMSC) was deposited on the polystyrene-coated crystals using the horizontal Langmuir-Schaefer deposition technique. For the QCM-D and the crystalline ordering measurements, 30 and 98 layers of TMSC, respectively, were deposited. The preparation of films is described by Tammelin et al.32 Prior to use in QCM-D, TMSC deposited on the polystyrene crystal was converted to cellulose by desilylation using hydrochloric acid vapor.27 Low-Charged (LC) Microfibrillated Cellulose Films. Cellulose microfibrils were disintegrated from delignified sulfite pulp using a high-pressure fluidizer (Microfluidizer M-110EH, Microfluidics Corp.) at STFI-Packforsk, Stockholm, Sweden.33 The homogenized 2% microfibril gel was diluted with water from a Milli-Q Gradient system (resistivity > 18 MΩ) to 1.67 g/L and centrifuged for 45 min at 10400 rpm to remove remaining fibril aggregates using an Optima L-90K ultracentrifuge from Beckman Coulter. The clear supernatant was used for spin-coating. Substrates used for film preparation were either silica QCM-D crystals (Q-Sense AB) or smooth silica wafers (Okmetic Oy, Helsinki, Finland). 3-Aminopropyltrimethoxysilane (APTS, 97%, Sigma Aldrich) was used as an anchoring substance to improve the coverage of the silicon surface with fibrils. Washed silica substrates were immersed in 1% v/v APTS/ toluene solution for 40 min, rinsed with toluene, and dried in an oven at 60 °C for 30 min. The model films were prepared by spincoating (Chemat Technology, KW-4A) the fibril dispersion onto the substrates at 3000 rpm for 45 s. The spin-coated surfaces were rinsed with water, dried gently with nitrogen gas, and heattreated in an oven at 80 °C for 10 min. The preparation of the nanofibril films is presented in more detail by Ahola et al.5
High-Charged (HC) Microfibrillated Cellulose Multilayer Films. Model films of carboxymethylated microfibrillated cellulose were prepared according to a method described by Wagberg et al.29 Briefly, the alternating deposition of polyethyleneimine (1 g/L, Mw = 60000 g/mol, 50% aqueous solution, Acros Organics) and dispersed MFC (1 g/L) on silica surfaces leads to the formation of polyelectrolyte multilayers (PEM).28 A silica wafer was treated with PEI and MFC up to one bilayer in Milli-Q water and another wafer with PEI and MFC up to seven bilayers to ensure sufficient surface coverage of the microfibrils. Nanocrystal Cellulose Films. A colloidal suspension of cellulose nanocrystals was prepared by acid hydrolysis of a dissolving grade pulp according to a previously described method.34 The colloidal sol of cellulose nanocrystals, stabilized by the presence of surface sulfate groups with sodium counterions and washed with deionized water, was used to prepare cellulose I surfaces using a modified procedure based on the method of Edgar and Gray.24 One percent (w/w) of this colloidal (32) Tammelin, T.; Saarinen, T.; Oesterberg, M.; Laine, J. Cellulose 2006, 13, 519–535. (33) Paeaekkoe, M.; Ankerfors, M.; Kosonen, H.; Nykaenen, A.; Ahola, S.; Oesterberg, M.; Ruokolainen, J.; Laine, J.; Larsson, P. T.; Ikkala, O.; Lindstroem, T. Biomacromolecules 2007, 8, 1934–1941. (34) Dong, X. M.; Kimura, T.; Revol, J.-F.; Gray, D. G. Langmuir 1996, 12, 2076–2082.
Langmuir 2009, 25(13), 7675–7685
Article suspension of cellulose I nanocrystals was spin-coated onto a silica wafer pretreated with PEI at 4000 rpm for 1 min. The films were subsequently heat-treated at 90 °C for 4 h to ensure that the films did not delaminate upon exposure to an aqueous electrolyte solution. Atomic Force Microscopy (AFM). The cellulose model films were imaged with tapping-mode AFM (Multimode IIIa, Veeco, Santa Barbara, CA) to determine the morphology and surface roughness. Topographic (height) and phase images were recorded under ambient air conditions (23 °C and 50% relative humidity). RTESP silica cantilevers (Veeco), each with tip radius of 8 nm and spring constant of 40 N/m (values provided by the manufacturer) were oscillated at their fundamental resonance frequencies, which ranged between 200 and 400 kHz. The roughness values were determined from the height image over 1 μm2 images and are presented as a root-mean-square (rms) value. No image processing except flattening was made. Ellipsometry. A manual nulling photoelectric ellipsometer with a mercury lamp (546.1 nm) (Type 43702-200E, Rudolph Research) was used to determine the thickness of the cellulose films. The ellipsometrical delta (Δ) and psi (ψ) values were collected, and the thicknesses of the films was determined by fitting the measured data to an optical three-layer model: silicon/ cellulose/air (Figure 1). The silicon and cellulose layers were expressed mathematically by using the complex refractive index (N = n + ik). If the ambient media (air) and the bulk surface (silicon) are regarded as infinite, and assuming that the layers in the optical model are homogeneous and optically isotropic, the ellipsometrical angles Δ and ψ can be related to the silicon and cellulose film conditions according to eq 1,35,36 where Δ is related to the phase shift and ψ to the change in amplitude upon reflection: tanðψÞ eiΔ ¼
Rp ¼ FðN0 ;Nf ;N1 ;df ;d1 ;φ;λÞ Rs
ð1Þ
Rp = complex reflection coefficient for the parallel component, Rs = complex reflection coefficient for the perpendicular component, F = ratio of Rp and Rs, N0 = complex refractive index of ambient (air), Nf = complex refractive index of the film (cellulose), N1 = complex refractive index of the bulk surface (silicon), df = cellulose film thickness, d1 = silicon thickness, φ = angle of incidence, and λ = wavelength of light. On the basis of an earlier work, the cellulose and silicon were assumed to have refractive indices of 1.53 and 3.85, respectively.13 Before the cellulose films were considered, the thickness of the native silicon oxide layer and the anchoring polymers, PEI, APTS, or polystyrene, was measured to be constantly below 4 nm and was thus assumed to be negligible in the calculation of the cellulose layer thickness. The calculations were made by assuming start values for the bulk surface (n1 and k1), and an iteration was performed until a solution with a positive thickness (d1) with a minimal imaginary part was found. Once the properties of the silicon substrate were determined, the average thickness and mean refractive index of the cellulose film (df, nf) were calculated by numerical iteration from the changes in Δ and ψ compared to the clean substrate. For each substrate studied, at least three points were measured to obtain an average film thickness and to determine the film homogeneity. Contact Angle Measurements. Contact angles were measured to estimate the surface energies of the tested substrates. A CAM 200 (KSV Instruments Ltd., Helsinki, Finland) contact angle goniometer was used for the advancing contact angle measurements. The software delivered by the instrument manufacturer calculates the contact angles on the basis of a (35) McCrackin, F. L.; Passaglia, E.; Stromberg, R. R.; Steinberg, H. L. J. Res. Natl. Bur. Stand., Sect. A: Phys. Chem. 1963, A67, 363–377. (36) Azzam, R. M. A.; Bashara, N. M. Ellipsometry and Polarized Light; North Holland: Amsterdam, The Netherlands, 1977.
DOI: 10.1021/la900323n
7677
Article
Aulin et al.
Figure 1. Optical three-layer model used for the determination of cellulose film thickness. numerical solution of the full Young-Laplace equation. Measurements were performed at room temperature with three probe liquids: Milli-Q water, glycerol (99.5%, Kebo Laboratory), and diiodomethane (99%, Sigma Aldrich). The contact angle was measured at three different positions on each sample. The surface energies were calculated from the contact angle data at equilibrium by the Van Oss method,37,38 which divides the total surface free energy into two components, the dispersive and the polar components, where the polar interactions originate from the Lewis acid-base interactions γtotal ¼ γps þ γds s
ð2Þ
where pffiffiffiffiffiffiffiffiffiffiffiffi γps ¼ 2 γþ s γs
ð3Þ
The subscripts indicate the solid, s, or liquid, l, phase, and γ and γp refer to the dispersive and polar components of the total surface energy, respectively. γ+ is the acceptor and γ- the donor part of the Lewis acid-base interactions. When combined with Young’s equation, the equations developed by Chaudhury, Good, and Van Oss yield the equation d
qffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffi! þ γl ð1 þ cos θÞ ¼ 2 γdl γds þ γsþ γγs γl l þ
ð4Þ
where θ is the contact angle, γl is the liquid surface tension, and + γ+ s , γl and γs , γl are the acid and base contributions to the surface energies of the solid and liquid, respectively. The values of the surface tension and its components for the probe liquids that were used in calculations were determined by Della Volpe and Siboni.39,40 Water, glycerol, and diiodomethane, with known γdl , γ+ l , and γl values, were used to determine the corresponding parameters of the solid, because the ratios of the basic and acidic parts of the interfacial free energy of test liquids should be large.41 The values for contact angle measurements against methylene iodide (γdl ≈ γl = 50.8 mJ/m2) were used to calculate the dispersive part of the surface energy (γds ) according to eq 5, which is valid when the interactions between the liquid and the surface are dominated by dispersive interactions: qffiffiffiffiffiffiffiffiffiffi γð1 þ cos θÞ ¼ 2 γdl γds
ð5Þ
Contact angles against water and glycerol were used to calculate the set of γ+ s and γs parameters. Small Incidence Angle X-ray Diffraction. The monochromatic synchrotron radiation (wavelength = 0.12398 nm), generated on the beamline BL-46 XU at the Super Photon ring 8 GeV (SPring-8, Nishi Harima, Hyogo, Japan), was irradiated onto the cellulose model films using a small incidence angle (37) Van Oss, C. J. Interfacial Forces in Aqueous Media; Dekker: New York, 1994. (38) Van Oss, C. J.; Ju, L.; Chaudhury, M. K.; Good, R. J. J. Colloid Interface Sci. 1989, 128, 313–19. (39) Della Volpe, C.; Siboni, S. J. Colloid Interface Sci. 1997, 195, 121–136. (40) Della Volpe, C.; Maniglio, D.; Brugnara, M.; Siboni, S.; Morra, M. J. Colloid Interface Sci. 2004, 271, 434–453. (41) Hollander, A. J. Colloid Interface Sci. 1995, 169, 493–496.
7678 DOI: 10.1021/la900323n
geometry. The beamline provides a Huber 8-axis goniometer equipped with a scintillation counter. The incidence angle was 0.15°, which is greater than the critical angle (0.14°) of cellulosic material. All of the out-of-plane diffraction profiles were obtained at a scanning rate of 1 °/min. After subtraction of the air scattering, the diffraction profile was curve-resolved into noncrystalline scattering and crystalline reflections using Rigaku multipeaks separation software. The apparent crystallinity, Xc, was evaluated from their area ratio. The crystallite size was determined from the integral width using Scherrer’s equation.42 To assess the reproducibility of the experimental results, similar measurements were performed for several samples using a laboratory-scale rotating anode type X-ray diffraction apparatus (Mac Science, Ltd., Sra-Ml8X) combined with a Ge (111) incident monochromator and a goniometer with a thin film attachment. The films were irradiated with X-ray beams (wavelength = 0.15418 nm) from a Cu target, operated at 40 kV and 200 mA, with an incidence angle of 0.23°. QCM-D. Swelling, that is, the interaction between cellulose films and water, was studied using a quartz crystal microbalance :: :: with dissipation (QCM-D) from Q-Sense AB, Vastra Frolunda, Sweden. The experiments were performed using the Q-Sense E4instrument, which is designed for controlled flow measurements. The QCM-D measures simultaneously changes in frequency and dissipation (frictional losses due to viscoelastic properties of the adsorbed layer) at the fundamental resonance frequency, 5 MHz, and its overtones 15, 25, 35, 45, 55, and 75 MHz. Without adsorbate, the crystal oscillates at a resonant frequency f0, and after adsorption, the resonance frequency decreases to f. For uniform, rigidly adsorbed films the change in frequency, Δf, is proportional to the adsorbed mass per unit surface, Δm, according to the Sauerbrey equation43,44 Δm ¼ -
CΔf n
ð6Þ
where n is the overtone number (1, 3, 5, 7, 9, 11, or 13) and C is a device-sensitive constant. The relationship is valid when the adsorbed mass is small compared to the mass of the crystal. The Δm value in the present work is affected both by the change in viscosity from air to water and by the adsorbed water in the film. Due to the good temperature control of the E4 instrument, it is possible to separate these different processes with good accuracy, and the frequency changes used to estimate the water uptake by the films have been corrected for the viscosity changes. The cellulose film samples were stored in a desiccator and taken out 15 min before measurement was begun. Each measurement was started in air, after which 10-5 M NaHCO3 solution was injected. The films were allowed to swell and reach equilibrium in the buffer solution, which was continuously fed to the QCM chamber at a flow rate of 0.1 mL/min. Each measurement was repeated at least twice.
Results Morphology of the Films. AFM height imaging was recorded to determine the morphology and surface roughness of the cellulose films (Figure 2). Phase images were generated for additional structural information, as a consequence of variations in the cellulose material properties such as adhesion, friction, and viscoelasticity.45,46 All images were recorded with (42) Cullity, B. D. Elements of X-Ray Diffraction, 2nd ed.; Prentice Hall: Upper Saddle River, NJ, 1978. :: :: (43) Hook, F.; Rodahl, M.; Brzezinski, P.; Kasemo, B. Langmuir 1998, 14, 729– 734. (44) Sauerbrey, G. Zeitschrift fuer Physik 1959, 155, 206–22. (45) Raghavan, D.; Gu, X.; Nguyen, T.; VanLandingham, M.; Karim, A. Macromolecules 2000, 33, 2573–2583. (46) Bar, G.; Thomann, Y.; Brandsch, R.; Cantow, H. J.; Whangbo, M. H. Langmuir 1997, 13, 3807–3812.
Langmuir 2009, 25(13), 7675–7685
Aulin et al.
Article
Figure 2. AFM tapping mode height (left) and phase (right) images of the cellulose films on silica. The scanned surface areas were 1 μm2, and the z ranges are 25 nm for the LC-MFC, HC-MFC, and nanocrystal surfaces, 40 nm for the NMMO and LiCl/DMAc surfaces, and 3 nm for the LS surfaces. Typical height profiles are shown in the middle, all with a z range of 25 nm.
an area of 1 μm2, and the images and height profiles presented in Figure 2 are representative of the films. AFM images of low-charged (LC) MFC show a fibrillar network structure of the film with an rms roughness of 3.2 nm. Compared with the LC model film, a bilayer of PEI/HC MFC formed a denser and smoother surface, with a roughness of 1.6 nm. The carboxymethylation pretreatment makes the fibrils highly charged and easier to liberate, and this results in smaller and more uniform fibril dimensions (10-15 nm) compared to the LC MFC, where the fibril width was 10-30 nm, as measured directly from the AFM height image. A possible broadening effect due to the geometry of the tip has not been considered. Compared to the single-bilayer PEI/MFC film, the seven-bilayer PEI/HC MFC film had a slightly rougher structure with an rms roughness value of 3.0 nm. The more compact fibrillar structure is shown in the phase image, where the surface Langmuir 2009, 25(13), 7675–7685
is fully covered with the fibrils, with no empty intermediate spaces. Figure 2 also shows images of the film prepared from the dimethylacetamide/lithium chloride solvent. As seen, the substrate displayed a structured feature in the nanometer/mesoscale but with no preferential orientation. The film exhibited a nonfibrillar spherically shaped structure totally covering the silica substrate. Compared with the other surfaces, the LiCl/DMAc film was rather rough with an rms roughness of about 5.9 nm, which is similar to that previously reported in the literature.14,47 The height and phase mode AFM images of 1 μm2 regenerated cellulose surfaces from NMMO solutions are also shown in Figure 2. The average roughness, rms, of the NMMO cellulose (47) Eriksson, M.; Notley, S. M.; Wgberg, L. Biomacromolecules 2007, 8, 912– 919.
DOI: 10.1021/la900323n
7679
Article
surface was 6.2 nm (Table 1). The phase image indicates fibrillarlike cellulose structures with a width in the range of 10-20 nm overlapping each other, and it appears that the fibrils prefer to adopt a worm-like structure. Cellulose nanocrystals seem to a have a dimension of 15 nm by 100-200 nm. The spin-coated cellulose film has a large amount of stiff, randomly oriented nanocrystals adsorbed, and no bare silica substrate can be seen. The rms value was calculated to be 4.0 nm. Finally, the LS-deposited cellulose film was also imaged by AFM. The measured rms roughness was 0.5 nm, indicating that the surface was very smooth, and no mesoscale cellulose structure, comparable to the structure of the other surfaces, could actually be seen in the images. The thickness is controlled by the deposition of one bilayer at a time horizontal to the silica surface, leading to the formation of very smooth and dense films, giving low rms values with AFM. The thicknesses of the cellulose films determined by ellipsometry are presented in Table 1. The thickness of the LiCl/DMAc film and of the NMMO film was simply controlled by adjusting the concentration of the cellulose in the LiCl/DMAc and NMMO, respectively, during the dissolving processes. The results are similar to earlier published results.20 Notley et al. reported film thicknesses of NMMO cellulose, LiCl/DMAc cellulose, and cellulose nanocrystals films to be 30, 44, and 120 nm, respectively.20 The greater thickness of the nanocrystal cellulose film presented by Notley et al. was due to a higher concentration of the nanocrystal suspension used in the spincoating process, which resulted in a thicker film than in the present study. The ellipsometry measurements also show that the thickness of the PEI/MFC films increases with increasing number of layers, as expected. Earlier ellipsometry studies showed similar increases in film thickness, although a different refractive index was assumed for the PEI/MFC film and different software was used to calculate the film thickness from the ellipsometric data. 29 To evaluate the ability of the cellulose surfaces to interact with liquids and other solid surfaces, the surface energies and their components were determined through contact angle measurements with water, glycerol, and methylene iodide. The contact angle data and the surface energies obtained are presented in Table 2. These values show that the dispersive part of the cellulose surface energy is significantly larger than the polar contribution. The cellulose films clearly exhibit basic surface characteristics and electron donor properties. Crystalline Ordering. Figure 3 shows the small incidence angle synchrotron X-ray diffraction profiles of the cellulose model films. Using SPring-8, it is obvious that sufficient diffraction intensity could be obtained from the films with a thickness of ca. 10 nm. The incidence angle employed in this study was higher than the critical angle, so the incidence X-ray beam penetrated into the film, and diffraction arose from the whole of the film without interference from the surface roughness. Among the profiles, the profiles of PEI/MFC multilayer films, LC-MFC films, and the nanocrystal film correspond to that of cellulose crystal I organization. The diffraction peaks are particularly sharp for the nanocrystal film, which implies a large crystallite size. (The sharp spike at the 2θ of ca. 22.7° was artificial due to electric noise, which was established using the laboratory scattering equipment.) In contrast, NMMO and LS films could be regarded as having a cellulose crystal II organization. This is reasonable considering the sample preparation, by which cellulose was dissolved and/or regenerated from native cellulose. 7680 DOI: 10.1021/la900323n
Aulin et al. Table 1. Cellulose Film Thickness (with an Error of (3 nm) and the Average Surface Roughness (rms), Determined by Ellipsometry and Atomic Force Microscopy, Respectively cellulose film
thickness (nm)
surface roughness, rms (nm)
LC-MFC (1 layer) HC-MFC, 1 bilayer PEI/MFC HC-MFC, 7 bilayer PEI/MFC LiCl/DMAc cellulose nanocrystal cellulose NMMO cellulose LS cellulose
11.9 12.5
3.2 1.6
75.0
3.0
53.8 23.8 32.5 17.8
5.9 4.0 6.2 0.5
Table 3 summarizes the apparent crystallinity Xc and the crystallite sizes of the cellulose model films obtained using the Scherrer equation. The Xc values from SPring-8 agree with those obtained using a laboratory apparatus within experimental error. This shows the high reliability and reproducibility of these values. The high crystalline ordering and large crystallite size in the nanocrystal film compared with those of other films could also be quantified. The diffraction profile of the film spin-coated from the DMAc/LiCl solution was very diffuse. The miocrostructure of this type of the film is under discussion, that is, whether it is completely amorphous or not. At least, the Xc value was very low, and the noncrystalline part was highly predominant. Interaction with Water. QCM-D measurements were performed to study the effect on swelling of the cellulose model surface crystallinity, chemistry, and morphology. Figure 4 shows QCM-D swelling curves for all of the cellulose films studied. The measurements were started in air, after which 10-5 M NaHCO3 buffer solution was injected (t = 0 min). The injection of this solution caused a sudden drop in frequency and an increase in energy dissipation due to the viscosity difference between air and aqueous buffer solution and also because of initial water uptake in the films. The change in frequency caused by the bulk effect (i.e., viscosity) was determined to be -386 Hz (third overtone) for both silica and gold crystals, and the rest of the frequency change was due to swelling of the film. A two-phase swelling behavior was observed: after the initial rapid water uptake, the cellulose surfaces started to slowly swell in the buffer solution. Most of the water uptake occurred in the initial phase, whereas the swelling in the second phase was minor for all of the films studied (see Figure 4). The time required to reach equilibrium varied from a few seconds to several hours, depending on the type of surface used. In the Langmuir-Schaefer (Figure 4a) and nanocrystal films (Figure 4c), no swelling was observed after the initial water uptake; equilibrium was reached immediately, which was observed as a sharp 90° angle in the frequency curve. LiCl/DMAc (Figure 4a) and PEI/MFC seven layers (Figure 4c) showed the highest swelling, and the swelling continued for several hours. In the NMMO cellulose and PEI/MFC films, a second swelling phase occurred, but it was minor. For the LC-MFC film a “bump” in the frequency curve was observed at first, after which the frequency slowly stabilized. Similar behavior has been observed and discussed by Ahola et al.5 The energy dissipation of the films (Figure 4b,d) increased due to changes in the viscoelastic properties of the films as a result of swelling in the aqueous buffer solution. The initial water uptake at t = 0 min caused a rapid increase in energy dissipation, corresponding to the change in frequency in Figure 4a,c. The energy dissipation increase due to the viscosity difference Langmuir 2009, 25(13), 7675–7685
Aulin et al.
Article
Table 2. Total Surface Energies of the Cellulose Films and Their Respective Polar and Dispersive Components, Calculated from Contact Angle Measurements against Water, Glycerol, and Methylene Iodide with an Error of (2°a cellulose film
water glycerol (deg) (deg)
methylene iodide (deg)
dispersive component (mNm/m2)
acceptor, γ+ donor, polar component total surface energy s 2 (mNm/m2) γ(mNm/m2) (mNm/m2) s (mNm/m )
LC-MFC (1 layer) 45.8 59.5 45.7 36.6 ( 1.1 4.8 ( 0.4 25.2 ( 0.2 22.1 ( 2.5 58.8 ( 3.6 HC-MFC, 1 bilayer 31.5 23.1 38.6 40.3 ( 1.0 6.2 ( 0.4 13.8 ( 0.5 18.5 ( 0.6 58.8 ( 1.6 PEI/MFC HC-MFC, 7 bilayer 31.8 25.2 30.8 43.9 ( 1.3 3.2 ( 0.2 14.6 ( 0.4 13.6 ( 1.4 57.5 ( 2.7 PEI/MFC nanocrystal cellulose 23.7 20.1 27.8 45.1 ( 0.8 2.4 ( 0.1 17.2 ( 0.4 13.0 ( 0.6 58.0 ( 1.4 LiCl/DMAc 39.7 31.7 37.7 40.7 ( 1.0 4.4 ( 0.2 12.0 ( 0.6 14.5 ( 0.9 55.2 ( 1.9 cellulose NMMO cellulose 23.8 28.7 37.4 40.9 ( 0.9 1.3 ( 0.1 20.2 ( 0.2 10.3 ( 1.3 51.2 ( 1.4 LS cellulose 49.7 41.2 34.1 42.4 ( 0.9 2.0 ( 0.1 9.6 ( 0.5 8.8 ( 1.0 51.2 ( 1.9 a Apart from the dispersive and polar component of the total surface energies, the acid/base (electron acceptor and donor) properties are summarized.
Figure 3. Small incidence angle synchrotron X-ray diffraction profiles of the cellulose films. Table 3. Apparent Crystallinity and Crystallite Size of the Model Cellulose Films crystalline ordering (%) SPring-8 cellulose I crystal structure LC-MFC 1 layer LC-MFC 6 layers PEI/HC-MFC 1 bilayer PEI/HC-MFC 7 bilayers nanocrystal cellulose II crystal structure NMMO cellulose LS cellulose LiCl/DMAc cellulose a Based on cellulose Iβ.
crystallite size (nm)
lab
SPring-8 11h0 2.7
110a 5.2
200a 2.7
2.1 3.2
4.1 6.9
3.0 3.5
2.8 2.9
2.0 2.1
2.3 3.0
a
70.0 62.0 87.0 58.0 63.0
66.5 60.5 64.5 54.4 85.1 60.0 53.5 14.8
between air, and the buffer solution was determined to be 159 10-6 and 158 10-6 for silica and gold, respectively. In LiCl/ DMAc (Figure 4b) and PEI/MFC multilayer films (Figure 4d) the increase in dissipation was larger than in the films that showed a lower degree of swelling, that is, Langmuir-Schaefer, NMMO, nanocrystals, and one bilayer PEI/MFC. The LC-MFC film showed the highest energy dissipation values (Figure 4d), indicating an extremely loose and viscous swollen film compared to the quantity of the water taken up (Figure 4c). Equation 6 was used to calculate the mass of water adsorbed in the films. To exclude the effect of the film thickness and the amount of cellulose in the films, the Sauerbrey mass was Langmuir 2009, 25(13), 7675–7685
normalized by dividing the mass by the thickness of the films. The percentage swelling (mass of water/dry mass of the film) was also calculated. The dry mass of the films was calculated using the thickness and area of the films and assuming that the densities of celluloses I and II are 1.592 and 1.582 g/cm3,48,49 respectively, in all of the films studied. The results in Table 4 show that the amount of water in the LiCl/DMAc film was high compared to that in the other films. NMMO, MFC, one-bilayer PEI/MFC, and nanocrystals show alomost the same swelling, whereas the Langmuir-Schaefer film showed the lowest percentage swelling. Surprisingly, the percentage swelling in the seven-bilayer MFC/PEI multilayer film was relatively low. The stiffening could be partly due to water removal from the nanofibril film caused by charge neutralization on the surface as highly cationic PEI is adsorbed on anionic MFC. The charge neutralization leads to a reduction in the swelling force.50 It is noteworthy that the Sauerbrey equation is valid only when the adsorbed amount is small compared to the mass of the crystal and the layer is rigid. If the layer is loose and the adsorbed amount is large, the equation can underestimate the mass of the layer, which can be the case with LiCl/DMAc and MFC films. Johannsmann’s model was used to calculate the “true” adsorbed mass of water.51-54 The calculations showed that the adsorbed mass was slightly higher when calculated with Johannsmann’s model, but the difference from the values obtained with the Sauerbrey equation was very small, and Johannsmann’s mass values are hence not shown.
Discussion Molecular and Meso Structural Levels of Cellulose Model Films and Their Interaction with Polar and Nonpolar Liquids. The crystalline ordering of cellulose in nature and in cellulose materials following different treatments is a complex issue, but it has been shown25 that the ordering has a large influence on the macroscopic properties of cellulose materials. As was mentioned earlier, four different polymorphs or crystal structures are known: celluloses I, II, III, and IV. The complex molecular features of cellulose also provide obstacles for model film preparation because it will influence the solubility of the (48) Sugiyama, J.; Vuong, R.; Chanzy, H. Macromolecules 1991, 24, 4168–4175. (49) Mwaikambo, L. ::Y.; Ansell, M. P. J. Mater. Sci. Lett. 2001, 20, 2095–2096. (50) Ahola, S. M. P.; Osterberg, M.; Teerinen, T.; Laine, J. Bioresources 2008, 3, 1315–1328. :: (51) Aulin, C.; Varga, I.; Claesson, P. M.; Wagberg, L.; Lindstrom, T. Langmuir 2008, 24, 2509–2518. (52) Naderi, A.; Claesson, P. M. Langmuir 2006, 22, 7639–7645. (53) Kaufman, E. D.; Belyea, J.; Johnson, M. C.; Nicholson, Z. M.; Ricks, J. L.; Shah, P. K.; Bayless, M.; Pettersson, T.; Feldotoe, Z.; Blomberg, E.; Claesson, P.; Franzen, S. Langmuir 2007, 23, 6053–6062. (54) Johannsmann, D.; Embs, F.; Willson, C. G.; Wegner, G.; Knoll, W. Makromol. Chem., Macromol. Symp. 1991, 46, 247–51.
DOI: 10.1021/la900323n
7681
Article
Aulin et al.
Figure 4. Swelling of dry cellulose model films upon exposure to 10-5 M NaHCO3 buffer solution (at t = 0 min) as determined by the QCMD technique: (a, c) change in frequency, Δf, as a function of water uptake in the cellulose films (third overtone); (b, d) change in energy dissipation as the cellulose films swell in aqueous solution. In dissipation curves b and d the scale of the y-axis starts from 150 (instead of 0) to show the differences more clearly. Table 4. Thickness and Dry Mass of Cellulose Films, Sauerbrey Mass of Adsorbed Water, Thickness-Normalized Mass of Adsorbed Water, and Percentage Swelling of Cellulose Model Surfaces
surface LC-MFC PEI/MFC 1 bilayer PEI/MFC 7 bilayers LiCl/DMAc NMMO cellulose nanocrystals LS cellulose
normalized dry mass adsorbed thickness of cellulose adsorbed mass 10-8 swelling 2 2 (nm) (%) (mg/m ) mass (mg/m ) (mg/m3) 11.9 12.5
18.9 19.9
4.6 4.2
3.9 3.4
24 21
75
119.4
12.9
1.7
11
53.8 32.5
85.2 51.4
40.7 11.2
7.6 3.5
48 22
23.6 17.8
37.6 28.2
9.9 2.1
4.2 1.2
26 7
cellulose and the reprecipitation procedure will effect the crystalline ordering of the solid film. The methodology of the preparation is not a straightforward issue; several different preparation methods were therefore compared, and the properties of the cellulose films obtained were carefully characterized in the present study. The molecular structures of the cellulose films were resolved by small incidence angle X-ray diffraction, and the results clearly show that the cellulose molecular structure and degree of crystalline ordering of the films depend on the choice of dissolving process. From these XRD studies it has been found that the cellulose film from LiCl/DMAc almost lacks crystalline ordering (14.7%), which indicates that the dissolving chemicals in this case penetrate or destroy the domains in the native cellulose. After the dissolving and spin-coating process, this further results in a dense film consisting of spherically shaped cellulose domains on the mesoscale. The NMMO cellulose film 7682 DOI: 10.1021/la900323n
clearly exhibited a cellulose II structure and a significantly higher degree of crystalline ordering than the LiCl/DMAc cellulose II films. The NMMO film thus recrystallizes during or after the dissolving process in NMMO. In addition to the change in crystalline alignment of the cellulose chains, there is a structural change at the mesolevel in the film. As seen in the AFM images, the sizes of the cellulose domains suggest that they are agglomerates of a few tens of individual cellulose chains and clearly not individual molecules. As the cellulose domains are conspicuous yet very small, they provide a change from the molecular level to a cellulose network or secondary structure in the mesolevel. Arndt et al.55 used static light scattering to characterize the size and shape of cellulose dissolved in NMMO and proposed a hyperbranched fringe-micellar model for the structure. This type of structure is not in contradiction to our AFM image of the paracrystalline film in Figure 2. On the contrary, it supports the results from the present work and suggests that the cellulose molecules are never totally dissolved in the NMMO solution but are rather incorporated in a fringemicellar type of structure already in the “dissolved” state and that this ordering is increased further as the film is regenerated. Figure 2 shows an AFM image of the LS-deposited cellulose film, with a very smooth surface with a very small scaled structure. This is, as previously mentioned, a consequence of the carefully controlled deposition of one monolayer at a time, whereas the roughness of the films remains constant, most probably creating a highly dense, well-ordered sheet structure.27 It is also interesting to note that, despite the fact that the molecules are totally dissolved in the LiClDMAc solution, they (55) Arndt, K.-F.; Morgenstern, B.; Roder, T. Macromol. Symp. 2000, 162, 109– 119.
Langmuir 2009, 25(13), 7675–7685
Aulin et al.
became ordered to a large degree with the LS deposition technique used for these surfaces. Previously it has been found that when these films are immersed in water, they obtain a similar, although maybe slightly thinner, fibrillar-type of structure as the films prepared by spin-coating from NMMO.56 So far, all of the cellulose surfaces discussed have been true dense films, that is, cellulose that uniformly and completely covers the substrate. More open fibrillar network films have also been created by treating the substrates with dispersions of low/ high-charged MFC or nanocrystals, using a spin-coating or dipping technique. The XRD analysis shows that these films were composed solely of cellulose I and indicate that the crystalline ordering of the LC/HC-MFC and nanocrystal remains intact after preparation of these films. The degree of crystalline ordering is, as expected, higher for the nanocrystal films than for the MFC films, because the hydrolysis step during the preparation degrades only the amorphous zones and not the crystalline zones of the native microfibrills. It should also be mentioned that, despite the rather smooth surface of these materials and their high refractive index,51 indicating a high density of the films, the fibrils in these films retain their fibrillar entity; when these materials are exposed to other solids and liquids, this will affect their macroscopic behavior. The crystalline ordering and the mesoscale structure of these materials naturally affect their interaction with other materials. In the ideal case, this interaction is determined by the chemical composition of the surfaces, which induces different types of interactions, such as nonspecific (van der Waals forces) and specific (e.g., polar, including hydrogen bonding) interactions. Contact angle measurements using water, glycerol, and methylene iodide were performed to provide information regarding the respective contributions of dispersive and polar interactions to the total surface energy of the surfaces. The calculated values in Table 2 indicate that the dispersive part of the cellulose surface energy is larger than the polar contribution, because the surface energy of cellulose determined by contact angle measurements is known to be ca. 55 mJ/m2.37 Kontturi et al. also reported a surface energy of nanocrystal cellulose films as high as 65 mJ/m2.3 Reported values for the dispersive part of the surface energy of NMMO cellulose films are 40 and 44 mJ/m2, determined by contact angle measurements.37,57 These values are similar to the values obtained in the present investigation. It is also worth noting that the relative contribution from dispersive interactions is highest for the nanocrystal surface, which also had the highest degree of crystalline ordering, indicating that the cellulose molecules are less prone to specific interactions in these surfaces. Table 2 also shows that the cellulose films exhibit basic surface characteristics, which is rather unexpected because the cellulose surfaces, at least HC-MFC and nanocrystals, should have numerous acidic groups. On the other hand, a basic nature of fibrils/fibers is not surprising, because the Van Oss method38 frequently used to determine these properties often results in surfaces with basic properties.58-60 The difference in total surface energy of the cellulose films, as obtained from contact angle measurements, is small. It is therefore difficult to draw conclusions on the influence of crystallinty and mesostructure :: (56) Osterberg, M. Ph.D. Thesis, Royal Institute of Technology, KTH, 2000. (57) Forsstroem, J.; Eriksson, M.; Wagberg, L. J. Adhes. Sci. Technol. 2005, 19, 783–798. (58) Berg, J. C. Nordic Pulp Pap. Res. J. 1993, 8, 75–85. (59) Berg, J. C. Surfactant Sci. Ser. 1993, 49, 75–148. (60) Luner, P. E.; Oh, E. Colloids Surf., A: Physicochem. Eng. Aspects 2001, 181, 31–48.
Langmuir 2009, 25(13), 7675–7685
Article
on the surface energy of the cellulose films. Further studies are needed where the surface properties are further characterized with, for example, contact angle measurements involving more probe liquids. Inverse gas chromatography may also provide additional information about the acid/base properties of the cellulose films.61-63 Fundamental aspects of adsorption,10,11,32,64 swelling,5,16,65 and surface interactions7,66-68 are the most prominent areas of application of cellulose model films. When the interaction of cellulose and other materials in both the dry and wet states is studied, it is important to consider and include both the molecular and the mesostructural levels of the surfaces. Smallscale roughness and surface heterogeneities may, for example, introduce large variations in the measured adhesion between cellulose surfaces, because the true contact area and adhesion :: mechanisms may vary between measurements.69,70 Rundlof 15 et al. measured the adhesion between a LB cellulose film and a polydimethylsiloxane (PDMS) cap and concluded that the hysteresis between loading and unloading cycles depended on interpenetration and surface roughness, that is, interdigitation, a parameter partly controlled by the secondary structure of the film. Only a few publications exist on adsorption studies on cellulose surfaces demonstrating the difference in adsorption behavior between cellulose films with amorphous71 and crystalline11,23,32,64,72 characteristics. The mesostructural properties of the cellulose films affect the adsorption of, for example, polyelectrolytes, enzymes, and surfactants onto their surfaces, but there is a lack of literature discussing the relationship between the interaction of cellulose and other materials and the mesostructure of the films. In the work by Stiernstedt et al.,73 the effect of roughness and crystallinity of cellulose films on normal and friction forces was discussed, and Ahola et al.18 compared the adsorption of enzymes and enzymatic hydrolysis on different cellulose model films. However, clearly more work is needed to study the adsorption behavior of, for example, polyelectolytes, surfactants, and enzymes onto cellulose surfaces considering the different affects from the mesostructures. Interaction of Cellulose Model Surfaces with Water. Cellulose does not naturally dissolve in water, but water can penetrate inside the amorphous part of a cellulose matrix, causing swelling of the entire material. Crystalline cellulose does not, however, swell because water cannot penetrate into the crystalline structure.74 With this in mind, the amorphous cellulose films should show a higher degree of swelling when exposed (61) Tshabalala, M. A. J. Appl. Polym. Sci. 1997, 65, 1013–1020. (62) Lundqvist, A.; Oedberg, L.; Berg, J. C. Tappi J. 1995, 78, 39–42. (63) Jacob, P. N.; Berg, J. C. Langmuir 1994, 10, 3086–3093. (64) Rojas, O. J.; Ernstsson, M.; Neuman, R. D.; Claesson, P. M. J. Phys. Chem. B 2000, 104, 10032–10042. (65) Freudenberg, U.; Zimmermann, R.; Schmidt, K.; Behrens, S. H.; Werner, C. J. Colloid Interface Sci. 2007, 309, 360–365. (66) Neuman, R. D.; Berg, J. M.; Claesson, P. M. Nordic Pulp Pap. Res. J. 1993, 8, 96–104. (67) Wagberg, L. Nordic Pulp & Paper Research Journal 2000, 15, 586–597. (68) Poptoshev, E.; Rutland, M. W.; Claesson, P. M. Langmuir 2000, 16, 1987– 1992. (69) Gardner, D. J.; Oporto, G. S.; Mills, R.; Samir, M. A. S. A. J. Adhes. Sci. Technol. 2008, 22, 545–567. (70) Kendall, K. Science 1994, 263, 1720–5. (71) Eriksson, J.; Malmsten, M.; Tiberg, F.; Callisen Thomas, H.; Damhus, T.; Johansen Katja, S. Journal Colloid Interface Sci. 2005, 284, 99–106. (72) Geffroy, C.; Labeau, M. P.; Wong, K.; Cabane, B.; Cohen Stuart, M. A. Colloids Surf., A: Physicochem. Eng. Aspects 2000, 172, 47–56. (73) Stiernstedt, J.; Nordgren, N.; Wagberg, L.; Brumer, H.; Gray, D. G.; Rutland, M. W. J. Colloid Interface Sci. 2006, 303, 117–123. (74) Mueller, M.; Czihak, C.; Schober, H.; Nishiyama, Y.; Vogl, G. Macromolecules 2000, 33, 1834–1840.
DOI: 10.1021/la900323n
7683
Article
Aulin et al.
to water, and the crystalline films should show a lower degree of swelling when in contact with water. The QCM-D results did indeed show that the amorphous LiCl/DMAc film had the highest degree of swelling (see Table 2). However, the second highest degree of swelling was detected for the cellulose nanocrystals. This observed water uptake by the nanocrystals cannot be due to the swelling of the individual crystals; it is most probably due to water penetration between the crystals, causing a separation of the crystals and allowing for a further water imbibition into the film. The nanocrystals contain sulfonic groups from the dissolution procedure, and these groups will aid in creating an osmotic pressure separating the crystals in the film and exposure to liquid water. With regard to the other films, it can be concluded that the Langmuir-Schaefer film has the lowest degree of swelling despite having about the same degree of crystalline ordering as, for example, the NMMO film, which swells considerably more than the LS film. There may be several explanations of this difference, but obviously the large ordering of the cellulose molecules during the formation of the LS films, shown by X-ray diffraction, results in a dense film with a low degree of swelling. The NMMO films are prepared from a fringe-micellar type of solution and, although they have a rather high degree of crystalline ordering, it seems logical that this slightly fibrillated material (see AFM images in Figure 1) will have a higher swelling in water than the LS film; that is, more water is absorbed between the fibrils of the NMMO film than within the fibrils themselves. From these observations, it can be suggested that the molecular structure of cellulose is not the only factor determining the interactions between cellulose films and water. It is rather a balance between the molecular structure and the mesostructure of the films that determines the swelling. According to this line of discussion, the films made from the microfibrillated cellulose should show a degree of swelling similar to that of the NMMO films. For the LC-MFC and the PEI/MFC one layer, this is indeed the situation, whereas the degree of swelling of the PEI/MFC seven bilayers is significantly lower. This latter finding can be explained by deswelling due to charge neutralization of the HC-MFC by the PEI, similar to that found for PDADMAC.50 An alternative explanation is the fact that PEI can act as a wet strength agent for cellulose fibers/ fibrils75 and, for the thicker films with a better packing of the MFC and a higher amount of PEI in g/m2, it can be suggested that the wet strength properties of PEI prevent the film from expanding and absorbing more water when in contact with liquid water. In summary, it may be suggested that the swelling of the films is a combination effect where the water can be incorporated either directly in amorphous parts of the cellulose or between cellulose crystals or cellulose fibrils or a combination of both. Generally, the swelling of the films at equilibrium can be described by the simple equation ΔGnetwork ¼ ΔGmix þ ΔGel
ð7Þ
where ΔGel is the integral Gibbs free energy from the osmotic pressure created by charges in the cellulose, ΔGnetwork is the integral Gibbs free energy describing the restraining action of the molecular/fibrillar structure of the film, and ΔGmix is the swelling due to the integral Gibbs free energy of mixing of cellulose and water There will hence be a balance between the (75) Dunlop-Jones, N. Wet-Strength Chemistry; Chapman and Hall: New York, 1991; pp 76-96.
7684 DOI: 10.1021/la900323n
restraining network energy and the free energy of swelling created by the charged groups and by the mixing of cellulose and water; that is, a higher charge of the cellulose (by, for example, oxidation) will lead to a higher swelling at the same ΔG network, and a higher ΔG network at a similar charge will lead to a lower degree of swelling. These different contributions are naturally average properties of the films, and simple swelling measurements do not reveal at what structural level the films are absorbing water, and as has been demonstrated in the present work, the swelling measurements have to be combined with a careful structural characterization of the films. Despite the high-resolution methods used, in the present work, it is not exactly clear how the swelling occurs at a molecular level and mesolevel of the films. It is clear, however, that the existence of a fibrillar/nanocrystal structure of the film will permit water absorption even though the material in the film has a high degree of crystalline ordering. This may at first seem contradictory but, given the mesostructuring of the film, this can be understood as demonstrated in the present work. Due to the network structure with voids between the fibrils/crystals, nanocapillarity effects will also contribute to the swelling behavior in the case of the fibrillar/nanocrystal films. The swelling of superabsorbent polymers and cellulose fluff pulp, as a result of the capillary pressure present in polymeric networks and interparticle pores, has previously been studied by Buchholz et al.76 To conclude, the swelling of cellulose model films is a complex phenomenon that is affected not only by the molecular structure of cellulose but also by the physical structure of the films. It is evident that the packing of the cellulose material is an important factor, and more information about cellulose model film densities is needed to enable the swelling phenomena to be fully understood. More work is also needed to study the swelling kinetics of the films and cellulosic fibers, but the model systems are definitely useful.
Conclusions Thin and smooth cellulose model films with different degrees of crystalline ordering were prepared. Small incidence angle X-ray diffraction studies clearly revealed that cellulose films consisting of LC-MFC, HC-MFC, and nanocrystals exhibit a cellulose I structure. The cellulose surfaces prepared from dissolving pulp dissolved in NMMO together with the films prepared by the Langmuir-Schaefer technique exhibit cellulose II ordering, whereas the LiCl/DMAc surfaces have amorphous characteristics with a very low degree of crystalline ordering. The surface morphology and the film thickness were characterized in detail using AFM and ellipsometry measurements. The total surface energy of the cellulose films was obtained using contact angle measurements with water, glycerol, and diiodomethane. The dispersive contribution to the surface energy was found to dominate the total surface energy of the films, and this was most pronounced for the nanocrystal films with a high degree of crystalline cellulose I ordering. The swelling of the films was studied by quartz crystal microbalance with dissipation, and some careful interpretations were made to link the swelling with the crystallinity properties of the films, that is, the degree of crystalline ordering. Clearly, the difference in both crystalline ordering and the mesostructure of the films affected the swelling of the cellulose films. Consequently, both of these properties
(76) Buchholz, F. L.; Pesce, S. R.; Powell, C. L. J. Appl. Polym. Sci. 2005, 98, 2493–2507.
Langmuir 2009, 25(13), 7675–7685
Aulin et al.
will also affect the adsorption and interaction behavior of these substrates. Acknowledgment. We thank BIM Kemi Sweden AB and the Knowledge Foundation through its graduate school YPK for financial support. L.W. also acknowledges financial support
Langmuir 2009, 25(13), 7675–7685
Article
:: from the Biomime project at KTH. S.A. and M.O. thank the Finnish Funding Agency for Technology and Innovation (Tekes) for funding through the Designcell project. We :: gratefully acknowledge Professor Lars Odberg for valuable discussions as well as Aila Rahkola from TKK for experimental assistance.
DOI: 10.1021/la900323n
7685
