The Build-Up of Polyelectrolyte Multilayers of Microfibrillated Cellulose

Jan 11, 2008 - Relationship between Young's Modulus and Film Architecture in Cellulose ..... Tom Lindström. ACS Applied Materials & Interfaces 0 (pro...
0 downloads 0 Views 469KB Size
Download PDF
784

Langmuir 2008, 24, 784-795

The Build-Up of Polyelectrolyte Multilayers of Microfibrillated Cellulose and Cationic Polyelectrolytes Lars Wågberg,*,† Gero Decher,‡ Magnus Norgren,†,§ Tom Lindstro¨m,| Mikael Ankerfors,| and Karl Axna¨s† Fibre and Polymer Technology, KTH, Teknikringen 56, 10044 Stockholm, Sweden, Institut Charles Sadron (CNRS-ULP), 6 rue Boussingault, 67083, Strassbourg Cedex, France, Department of Natural Sciences, Fibre Science and Communication Network, Mid Sweden UniVersity, 85170 SundsVall, Sweden, and STFI Packforsk AB, Box 5604, 114 86 Stockholm, Sweden ReceiVed August 13, 2007. In Final Form: October 16, 2007 A new type of nanocellulosic material has been prepared by high-pressure homogenization of carboxymethylated cellulose fibers followed by ultrasonication and centrifugation. This material had a cylindrical cross-section as shown by transmission electron microscopy with a diameter of 5-15 nm and a length of up to 1 µm. Calculations, using the Poisson-Boltzmann equation, showed that the surface potential was between 200 and 250 mV, depending on the pH, the salt concentration, and the size of the fibrils. They also showed that the carboxyl groups on the surface of the nanofibrils are not fully dissociated until the pH has reached pH ) ∼10 in deionized water. Calculations of the interaction between the fibrils using the Derjaguin-Landau-Verwey-Overbeek theory and assuming a cylindrical geometry indicated that there is a large electrostatic repulsion between these fibrils, provided the carboxyl groups are dissociated. If the pH is too low and/or the salt concentration is too high, there will be a large attraction between the fibrils, leading to a rapid aggregation of the fibrils. It is also possible to form polyelectrolyte multilayers (PEMs) by combining different types of polyelectrolytes and microfibrillated cellulose (MFC). In this study, silicon oxide surfaces were first treated with cationic polyelectrolytes before the surfaces were exposed to MFC. The build-up of the layers was monitored with ellipsometry, and they show that it is possible to form very well-defined layers by combinations of MFC and different types of polyelectrolytes and different ionic strengths of the solutions during the adsorption of the polyelectrolyte. A polyelectrolyte with a three-dimensional structure leads to the build-up of thick layers of MFC, whereas the use of a highly charged linear polyelectrolyte leads to the formation of thinner layers of MFC. An increase in the salt concentration during the adsorption of the polyelectrolyte results in the formation of thicker layers of MFC, indicating that the structure of the adsorbed polyelectrolyte has a large influence on the formation of the MFC layer. The films of polyelectrolytes and MFC were so smooth and well-defined that they showed clearly different interference colors, depending on the film thickness. A comparison between the thickness of the films, as measured with ellipsometry, and the thickness estimated from their colors showed good agreement, assuming that the films consisted mainly of solid cellulose with a refractive index of 1.53. Carboxymethylated MFC is thus a new type of nanomaterial that can be combined with oppositely charged polyelectrolytes to form well-defined layers that may be used to form, for example, new types of sensor materials.

Introduction The formation of polyelectrolyte multilayers (PEMs) by the consecutive treatment of solid surfaces with oppositely charged polyelectrolytes/particles has been known for some time,1,2 and it has received remarkable attention because of its simplicity and ingenious applicability to surface modification and surface engineering.3-7 By combining different polyelectrolytes or polyelectrolytes and different nanoparticles it has been possible to prepare electrochromic devices,6 highly efficient membranes,4,8 * Corresponding author. † KTH. ‡ Institut Charles Sadron. § Mid Sweden University. | STFI Packforsk AB. (1) Decher, G.; Hong, J. D.; Schmitt, J. Thin Solid Films 1992, 210/211, 831. (2) Iler, R. K. J. Colloid Interface Sci. 1966, 21, 569. (3) Decher, G. Science 1997, 277, 1232. (4) Decher, G.; Schlenoff, J. B. Multilayer Thin Films: Sequential Assembly of Nanocomposite Materials; Wiley-VCH: Weinheim, Germany, 2003. (5) Bertrand, P.; Jonas, A.; Laschewsky, A.; Legras, R. Macromol. Rapid Commun. 2000, 21, 319. (6) Hammond, P. T. AdV. Mater. 2004, 16, 1271. (7) von Klitzing, R. Phys. Chem. Chem. Phys. 2006, 8, 5012. (8) Stranton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. Langmuir 2003, 19, 7038.

hollow capsules for the controlled release of different active chemicals,4,9,10 and solid-state dye-sensitized solar cells.11 Furthermore, by controlling the conditions during their preparation, it has been possible to control the structure and thickness of the multilayers.7 A higher charge density of the polyelectrolytes leads to a thinner layer as shown in experiments where polystyrene sulfonate (PSS) was combined with copolymers of poly(diallyldimethylammonium chloride) (PDADMAC) and N-methyl-N-vinylformamide with different charge densities.12 Initially, there is a large decrease in the thickness/layer when the charge density is increased, but a constant thickness is reached at a PDADMAC content of 60 mol %. It was also found that a minimum charge density of the cation of 0.036 elementary charges per angstrom of contour length of the polyelectrolyte was needed for the formation of PEMs.12 The type of polymer also affects the change in thickness with the number of polyelectrolyte bilayers. As an example of this, a combination of PSS and PDADMAC in 1 M NaCl on silicon (9) Donath, E.; Sukhorukov, G. B.; Caruso, F.; Davis, S. A.; Mo¨hwald, H. Angew. Chem., Int. Ed. Engl. 1998, 37, 2202. (10) Caruso, F.; Caruso, R. A.; Mo¨hwald, H. Chem. Mater. 1999, 11, 3309. (11) Lowman, G. M.; Hammond, P. T. Small 2005, 1, 1070. (12) Glinel, K.; Moussa, A.; Jonas, A. M.; Laschewsky, A. Langmuir 2002, 18, 1408.

10.1021/la702481v CCC: $40.75 © 2008 American Chemical Society Published on Web 01/11/2008

PEMs of MFC and Cationic Polyelectrolytes

oxide gave a material with a thickness of 270 Å,13 whereas a combination of PSS and poly(allylamine hydrochloride) (PAH) under similar conditions gave a thickness per layer of around 40 Å.14 A third factor that has a considerable influence on the incremental build-up of the layers in a PEM is the salt concentration in the solution during PEM formation. It has been demonstrated that there is a linear increase in thickness per layer when the NaCl concentration is increased for both PSS/ PDADMAC12 and the PSS/PAH system,14 whereas earlier measurements showed a square relationship between salt concentration and thickness for the PSS/PAH layer.15 The reason for the large increase in layer thickness with increasing salt concentration is not clear, but it has been suggested13 that the polymers were not adsorbed with a larger amount of loops and tails at higher salt concentrations in a single layer, as might be expected from polyelectrolyte adsorption theory,16 but more as nonequilibrium structures within several polymeric layers in the PEM. This hypothesis has not been critically tested to our knowledge. It has also been established that the exposure of PEMs of PSS and PDADMAC, formed at moderate salt concentrations, to solutions of higher concentrations of NaCl resulted in a smoothening, i.e., annealing, of the PEM layer.17 The molecular mass of the polyelectrolyte also affects the build-up of the PEM.18 By testing combinations of PSS in a range of molecular masses from 7200 to 801 100 g mol-1 with methylated poly(vinylpyridine) of 5060 or 46 700 g mol-1, it was found18 that the high molecular polyelectrolytes (L+/L-) gave thicker layers than combinations of high molecular mass cations (L+) and low molecular mass anions (S-) and vice versa (L-/S+). A combination of low molecular mass polyelectrolytes (S+/S-) gave no steady build-up of PEM at all at 0.5 M NaCl. This was interpreted as being due to the establishment of a kinetic balance between the formation of PEM and the formation of soluble polyelectrolyte complexes (PECs). If the interpenetration of polyelectrolytes into the PEM was nonexisting or limited, i.e., in S+/S- combinations, the formation of PECs was favored, whereas the opposite was found for high molecular mass combinations (L+/L-). This means that both the kinetics and the salt concentration have a large influence on the balance between PEM and PEC formation. It has also been established that the type of counterion has a large influence on the thickness of the PEM formed from PSS and PDADMAC.19 Using a set of negative counterions from the Hofmeister series, it was shown that the more hydrated counterions, i.e., F-, resulted in much thinner layers than the less hydrated Br-.19 It was suggested that this was mainly due to the influence of the counterions on the solution structure of the polyelectrolyte rather than on the PEM formation per se.7,19 It has also been established that the type of polyelectrolyte in the anchoring layer has a large effect on the build-up of the PEM of PAH and PSS.20,21 When a cross-linked polyelectrolyte, polyethyleneimine (PEI), was used as an anchoring layer, the PEMs of PSS/PAH showed a more direct formation of multilayers (13) Dubas, S. T.; Schlenoff, J. B. Macromolecules 1999, 32, 8153. (14) Lo¨sche, M.; Schmitt, J.; Decher, G.; Bouwman, W. G.; Kjaer, K. Macromolecules 1998, 31, 8893. (15) Lvov, Y. M.; Decher, G. Crystallogr. Rep. 1994, 39, 628. (16) Fleer, G. J.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces; Chapman & Hall: London, 1993. (17) Dubas, S. T.; Schlenoff, J. B. Langmuir 2001, 17, 7725. (18) Sue, Z.; Salloum, D.; Schlenoff, J. B. Langmuir 2003, 19, 2491. (19) Saloma¨ki, M.; Tervasma¨ki, P.; Areva, S.; Kankare, J. Langmuir 2004, 20, 3679. (20) Bosio, V.; Dubreuil, F.; Bogdanovic, G.; Fery, A. Colloids Surf., A 2003, 243, 147. (21) Kolasinska, M.; Krastev, R.; Warszynski, P. J. Colloid Interface. Sci. 2007, 305, 46.

Langmuir, Vol. 24, No. 3, 2008 785

at lower layer numbers, and the formed layers also had a lower surface roughness.20,21 The thickness of 13 layers was 126 Å without PEI and 153 Å with PEI as an anchoring layer.21 The initial layer can clearly also have a large influence on layers further out in the PEM. There are thus many factors affecting the build-up of PEMs from oppositely charged polyelectrolytes and/or nanoparticles and, despite several theoretical approaches22-24 and molecular dynamics simulations,25-28 there is still no general theory available to describe the formation of PEMs on solid surfaces. Recently, the layer-by-layer technique has received a lot of interest in biomedical applications,29,30 and it has even been used for tailoring surfaces for interactions with human chondrosarcoma cells.31 In these applications, it is essential that the polyelectrolytes have an inert character and that the material used shows no interaction with human tissue, apart from the substance with the bioactive property. The building blocks of the PEM should preferably also be renewable. In this respect, cellulose, the most abundant polymer on earth, should be an appealing polymer, provided that it can be given a sufficient charge to make it dispersible/soluble in water and that it has a physical form suitable for the formation of PEMs. Recently it has been shown32,33 that nanocrystals produced from cotton via hydrolysis in sulfuric acid and subsequently washed and ultrasonicated can be incorporated into PEMs together with PDADMAC32 or PAH.33 The nanocrystals, with average dimensions of 129 × 10 nm,2,32 formed typical PEMs with the different polyelectrolytes, and it was found by ellipsometry that the bilayer thickness of PDADMAC/nanocrystals was 11 nm, whereas that of PAH/ nanocrystals was 2 nm. Another interesting cellulose material is microfibrillated cellulose (MFC), composed of liberated semicrystalline microfibrils, normally produced using high-pressure homogenization of wood fibers in water. MFC was first introduced by Turbak34 who reported that the width of the MFC fibrils was between 25 and 100 nm, while the length was much longer. A new method for the manufacture of a smaller and more homogeneous MFC has recently been developed,35 in which a combination of mechanical and enzymatic pretreatment is used followed by highpressure homogenization.35 To our knowledge, the first carboxylated MFC was manufactured by Wågberg et al.36 but this material was not further characterized, and it was used only for studies of polyelectrolyte accessibility to cellulose microfibrils. It was found that high (22) Joanny, J. F. In Multilayer Thin Films: Sequential Assembly of Nanocomposite Materials; Decher, G.; Schlenoff, J. B., Eds.; Wiley-VCH: Weinheim, Germany, 2003. (23) Netz, R. R.; Joanny, J.-F. Macromolecules 1999, 32, 9013. (24) Castelnovo, M.; Joanny, J.-F. Langmuir 2000, 16, 7524. (25) Dobrynin, A. V.; Rubinstein, M. Prog. Polym. Sci. 2005, 30, 1049. (26) Patel, P. A.; Jeon, J.; Mather, P. T.; Dobrynin, A. V. Langmuir 2006, 22, 9994. (27) Patel, P. A.; Jeon, J.; Mather, P. T.; Dobrynin, A. V. Langmuir 2005, 21, 6113. (28) Abu-Sharkh, B. Langmuir 2006, 22, 3028. (29) Tang, Z.; Wang, Y.; Podsialo, P.; Kotov, N. A. AdV. Mater. 2006, 18, 3203. (30) Schneider, A.; Vodouhe, C.; Richert, L.; Francius, G.; Le Guen, E.; Schaaf, P.; Voegel, J.-C.; Frisch, B.; Picart, C. Biomacromolecules 2007, 8, 139. (31) Richert, L.; Lavalle, P.; Vautier, D.; Senger, B.; Stoltz, J.-F.; Schaaf, P.; Voegel, J.-C.; Picart, C. Biomacromolecules 2002, 3, 1170. (32) Podsialo, P.; Choi, S.-O.; Shim, B.; Lee, J.; Cuddihy, M.; Kotov, N. A. Biomacromolecules 2005, 6, 2914. (33) Cranston, E. D.; Gray, D. G. Biomacromolecules 2006, 7, 2522. (34) Turbak, A. F.; Snyder, F. W.; Sandberg, K. R. J. Appl. Polym. Sci.: Appl. Polym. Symp. 1983, 37, 815. (35) Pa¨a¨kko¨, M.; Ankerfors, M.; Kosonen, H.; Nyka¨nen, A.; Ahola, S.; O ¨ sterberg, M.; Ruokolainen, J.; Laine, J.; Larsson, P. T.; Ikkala, O.; Lindstro¨m, T. Biomacromolecules 2007, 8, 1934. (36) Wågberg, L.; Winter, L.; Lindstro¨m, T. Colloids Surf. 1987, 27, 163.

786 Langmuir, Vol. 24, No. 3, 2008

molecular weight polyelectrolytes were accessible to all carboxyl groups, indicating total delamination of the cell wall. This method was also based on carboxymethylated pulps and is followed up on in the present paper. Other ways of introducing carboxyl groups such as TEMPO oxidation have also been explored.37 The purpose of the present work was to determine the properties of carboxylated MFC in order to establish the colloidal stability of this material and to utilize these fully liberated microfibrils to form multilayers with cationic polyelectrolytes. It was also intended to clarify how this build-up was affected by the selection of polyelectrolytes and the solution concentration of NaCl used during cationic polyelectrolyte adsorption. In addition, the MFC was carefully characterized with different techniques. Experimental Section Materials. Fibers. In the manufacture of MFC, a commercial sulfite softwood-dissolving pulp (Domsjo¨ Dissolving Plus; Domsjo¨ Fabriker AB, Domsjo¨, Sweden), from 60% Norwegian spruce (Picea abies) and 40% Scottish Pine (Pinus sylVestris), with a hemicellulose content of 4.5% (measured as solubility in 18% NaOH) and a lignin content of 0.6% was used. The pulp was thoroughly washed with deionized water and used in its never-dried form. Polyelectrolytes. Three different cationic polyelectrolytes were used: PEI, Lupasol WF (Mw ) 25 000 g mol-1 according to the supplier) from BASF, Ludwigshafen, Germany; PAH from Aldrich (Mw 70 000 g mol-1 according to the supplier); and PDADMAC from Aldrich (Mw ) 100 000-200 000 g mol-1 according to the supplier). All solutions were prepared using ultrapure water (Milli-Q plus system, Millipore), and the polyelectrolytes were used without further purification. The polyelectrolyte concentrations used were PEI, 2.5 mg mL-1; PDADMAC, 1.25 mg mL-1; and PAH, 1.01 mg mL-1. These concentrations for the PAH and PDADMAC gave solutions with concentrations of 0.01 monomol L-1 (monomol ) moles of the repeating monomer). In a set of experiments where the influence of ionic strength on the multilayer build-up was investigated, solutions of PAH and PDADMAC containing 0.1 and 0.5 M NaCl were used. Silicon Wafers. Silicon wafers for use as substrates in the formation of multilayers were purchased from WaferNet, Inc. (San Jose´, CA). The silicon wafers used were initially degreased with acetone, lowered into a bath of MeOH/HCl (1:1) for 10 min, followed by another 10 min in H2SO4, and finally rinsed extensively with Milli-Q water and blown dry with nitrogen. Methods. Preparation of MFC. The MFC was prepared at STFI Packforsk, Stockholm, Sweden, with the aid of a high-pressure homogenization technique similar to an previously described procedure35 but using a carboxymethylation38 pretreatment of the fibers. A total of 110 grams of fibers were pretreated. The neverdried fibers were first dispersed in deionized water at 10 000 revolutions in an ordinary laboratory reslusher. This was conducted with batches of 30 g of fibers in 2 L of deionized water. The fibers were than solvent-changed to ethanol by washing 110 g of fibers in 1 L of ethanol four times with an intermediate filtration step. The fibers were then impregnated for 30 min with a solution of 10 grams of monochloroacetic acid in 500 mL of isopropanol. These fibers were then added in portions to a solution of 16.2 g of NaOH in 500 mL of methanol mixed with 2 L of isopropanol that had been heated to just below its boiling temperature in a 5 L reaction vessel fitted with a condenser. This carboxymethylation reaction was allowed to continue for 1 h. Following this carboxymethylation step, the fibers were filtered and washed in three steps: first with 20 L of deionized water, then with 2 L of acetic acid (0.1 M), and finally with 10 L of deionized water. The fibers were then impregnated with a 2 L NaHCO3 solution (4 wt % solution) for 60 min in order to convert the carboxyl groups (37) Saito, T.; Nishiyama, Y.; Putaux, J.-L.; Vignon, M.; Isogai, A. Biomacromolecules 2006, 7, 1687. (38) Walecka, J. A. Tappi 1956, 39, 458.

Wågberg et al. to their sodium form. Finally, the fibers were washed with 15 L of deionized water and drained on a Bu¨chner funnel. After this pretreatment, the fibers were homogenized using a high-pressure fluidizer (Microfluidizer M-110EH, Microfluidics Corp.). The fluidizer was equipped with two chambers of different sizes connected in series (200 and 100 µm). Full homogenization could be achieved with a single pass at a fiber consistency of 2 wt % and an operating pressure of 1650 bar. Dispersion of the MFC. The fibrils from the homogenization were dispersed by sonication. One gram of MFC stock sample (approximately 2 wt %) and 12 mL of Milli-Q water were mixed by shaking in a 15 mL glass jar. The apparatus used was a Bioblock Scientific, Vibracell 742 412 with a microtip probe (3 mm diameter). The mix was sonicated for 10 min at 25% of the amplitude setting. After this, the fibrils were quite well dispersed, as observed by ocular inspection, but a very small fraction of undispersed material and some contamination from the Titanium microtip probe remained. The contamination is due to a slow disintegration/destruction/erosion that occurs at high amplitudes. Thereafter, the dispersion was transferred to 30 mL flasks with a Teflon lining and submitted to centrifugation for 2 h at 8030g. The clear dispersion was then removed by pipet. The influence of sonication time on the dispersion efficiency was investigated at 3, 10, and 30 min for three samples. Charge Determination of the MFC. The total charge of the MFC was measured by conductometric titration39 of the fibers, prior to homogenization, at STFI-Packforsk, Stockholm, Sweden. Before titration, the fibers were treated both at low pH, to remove unwanted adsorbed metal ions, and at high pH, to remove possible contaminants, i.e., hemicellulose, from the fibers. In this procedure, a sample containing 2 g of dry pulp was dispersed in 1000 mL of deionized water with a subsequent addition of 0.01 M HCl to set the pH at pH ) 2. The excess HCl was removed after 30 min by repeated filtration and dispersion in deionized water until the conductivity was below 5 µS cm-1. In order to transfer the carboxyl groups to their sodium form, the fibers were dispersed with deionized water, after which 0.001 M NaHCO3 was added and the pH was adjusted to pH ) 9 by the addition of NaOH. After 30 min, the excess NaOH and the NaHCO3 were removed by repeated filtration and dispersion in deionized water until the conductivity was below 5 µS cm-1. After this, the sample was once more set to its hydrogen counterion form and washed to a conductivity below 5 µS cm-1. Transmission Electron Microscopy (TEM). In the preparation for TEM, 5 µL of the dispersed MFC sample was deposited onto a freshly glow-discharged carbon-coated Cu grid (400 mesh). After adsorption, the grid was stained with a droplet (5 µL) of 1% uranyl acetate and dried with a filter paper (Whatman 4 or 5). Finally, the grid was observed under standard conditions in a CM 12 Philips TEM operating at 120 kV. The images were recorded on SO163 films (Kodak) or with a Megaview III CCD camera (SIS). Atomic Force Microscopy (AFM). The bilayers consisting of PEI and MFC formed by the same procedure as was used for multilayering of the materials were imaged by tapping-mode AFM using a Nanoscope IIIa from Digital Instruments (Santa Barbara, CA). The images were recorded under ambient air conditions (temperature and relative humidity) with an RTESP silica cantilever (Vecco, Santa Barbara, CA). Multilayering of MFC and Different Cationic Polyelectrolytes. The multilayered films were formed either by simple dipping or by spraying. All the films were deposited on a precursor layer of PEI. For the films formed by dipping, the contact time with the MFC was 20 min, and the dispersed MFC was used without further dilution. The MFC dipping was followed by two consecutive 5-min dips in two separate vessels with Milli-Q water in order to rinse off the superfluous MFC. In all dipping experiments, only about half the wafer was covered by the liquid. The wafers were simply left standing up, leaning on the inside of the vessel. This was done in order not to contaminate the surface or the liquid with the tweezers. When applying a layer of polyelectrolyte (PEI, PAH, or PDADMAC), the (39) Katz, K.; Beatson, R. P.; Scallan, A. M. SVen. Papperstidn. 1984, 87, R48.

PEMs of MFC and Cationic Polyelectrolytes

Figure 1. A dispersion of MFC, 1.9 g L-1, in deionized water after homogenization, ultrasonication, and centrifugation. dipping time was 10 min, and the rinsing procedure was the same. The sequence used in these experiments was as follows: formation of an initial layer of PEI and then an MFC layer, followed by alternating polyelectrolyte and MFC dipping. In all the treatment steps, the pH value was between 7 and 8. The thickness was measured by ellipsometry after each polyelectrolyte or MFC layer had been applied. No true statistical evaluation of the formation of PEMs of polyelectrolytes and MFC was performed, i.e., more than five experimental series per measuring point. However, the formation of PEM with PEI and MFC was repeated several times during the course of the work, and only minor differences were found between the series. By plotting the thickness as a function of layer number, it was actually difficult to distinguish the different measuring series from each other. In this respect, the dipping experiments were very repeatable. The spraying procedure was similar in principle but much quicker. Analogous with the dipping, PEI was used as a precursor layer. The wafer was held up by a clip from underneath. All the spraying was performed horizontally and perpendicular to the receiving surface. The distance from the spraying nozzle to the surface was about 15 cm. The spraying time for both polyelectrolyte and MFC was 3 s. This was followed by 30 s drainage. The surface was then sprayrinsed with Milli-Q water for 20 s with an additional contact/drainage time of 10 s.40 The wafer was finally removed from the clip and dried under a flow of nitrogen. Ellipsometry. The thickness of the multilayered films was measured with a PLASMOS SD 2100 ellipsometer operated at a single wavelength of 632.8 nm and a constant angle of 45°. The refractive index was assumed to be constant at n ) 1.465. Although this procedure leads to slightly incorrect values with respect to the absolute film thickness, it gives a quick and precise determination of the relative film thicknesses. Thickness values obtained on the assumption that the refractive index is the same for all films are of better precision than is required for the comparison of film growth data as in this report. For each substrate studied, 10 points were measured to obtain an average film thickness and to determine the film homogeneity.

Langmuir, Vol. 24, No. 3, 2008 787

Figure 2. A TEM image of the liberated MFC. The scale bar corresponds to 0.5 µm.

Characterization of the Anionic MFC Material. The purpose of the carboxymethylation procedure is twofold. First, a charged MFC material is desirable for building PEMs; second, charging of the pulp makes the microfibrils much easier to liberate and a much more uniform microfibrillar material can thus be obtained. Carboxymethylation of the starting wood pulp material, followed by homogenization in the microfluidizer, ultrasonication, and subsequent centrifugation of the MFC resulted in a clear dispersion of MFC, as shown in Figure 1, with a concentration

of 1.9 g L-1. The clarity of this dispersion means that the fibrils of the MFC are too small to scatter light. The TEM investigation also showed that the fibrils are indeed liberated from each other after the homogenization and ultrasonication procedure. This is shown in Figure 2. A simple measurement of the size of the fibrils shows that they have a cross-section between 5 and 15 nm and that the length can be more than 1 µm. This means that this is a new, rather unique nanomaterial from a renewable resource. Since cellulose crystals are known to have a high Young’s modulus of 138 GPa,41 it also means that the material is very interesting for fibril-reinforced composites. This use of the material is, however, beyond the scope of the present work. The results of the conductometric titrations showed that the charge of the MFC was 515 µequiv g-1. Recalculated to the degree of substitution of the cellulose, this corresponds to 0.087, which is significantly higher than what is usually found in native fibers. It should also be stressed that the charge of the cellulose in native fibers is close to 0, since the charge of the fibers stems mainly from the presence of charged hemicelluloses. The presence of these carboxyl groups naturally influences the dispersion properties of the MFC and the interaction between the MFC and oppositely charged polyelectrolytes. Since the dissociation of the carboxyl groups is pH-dependent, the charge of the MFC will also be pH-dependent and, since the fibrils contain all their charge on the surface of the fibrils, the dissociation of these charges will be highly dependent on the presence of other charges in a manner similar to that found in polyelectrolytes. This also means that the degree of dissociation of the carboxyl groups will significantly depend on the ionic strength in solution. This will be treated further under the Discussion section. Formation of PEMs of Polyelectrolytes and MFC. In order to determine whether the MFC could interact with oppositely charged polyelectrolytes to form PEMs, as was previously shown to be possible with cellulose nanocrystals,32,33 a series of experiments were conducted where silicon wafers were consecutively treated with cationic polyelectrolytes and MFC with thorough rinsing between each step. The thickness of the layers formed on the wafers was determined with ellipsometry, and the results are summarized in Figure 3, where the thickness of PEMs from MFC and PDADMAC, PAH, and PEI are shown. In these experiments, no extra salt, apart from the counterions to the charges on the MFC and on the polyelectrolytes, was added. As can be seen in the figure, there is a large difference in multilayer thickness between the different polyelectrolytes,

(40) Izquierdo, A.; Ono, S.; Voegel, J.-C.; Schaaf, P.; Decher, G. Langmuir 2005, 21, 7558.

(41) Nishino, T.; Takano, K.; Nakamae, K. J. Polymer Sci. Part B: Polymer Phys. 1995, 33, 1647.

Results

788 Langmuir, Vol. 24, No. 3, 2008

Wågberg et al.

Figure 3. Thickness of a PEM from MFC and, respectively, PEI, PAH, and PDADMAC, measured by ellipsometry. No extra salt was added, and the pH was between 7 and 8 in all the treatment steps.

Figure 4. (a) Thickness of a PEM of MFC and PDADMAC at different NaCl concentrations measured with ellipsometry. The pH was between 7 and 8 in all the treatment steps. (b) Thickness of a PEM from MFC and PAH at different NaCl concentrations measured with ellipsometry. The pH was between 7 and 8 in all the treatment steps.

particularly between PEI and the other two polyelectrolytes. There is also a difference not only in total thickness but also in the development of thickness with layer for the different polyelectrolytes. The increase in thickness is small for the first layers and then reaches a steady build-up, with a larger increase in thickness for the MFC layer than for the polyelectrolyte layer. The thickness of each MFC layer is much higher with PEI than with PDADMAC. With PAH, there is a steady increase in thickness with both the MFC and the polyelectrolyte layer from the start of the build-up of the PEM; i.e., the thickness of each MFC layer is similar to that of the polyelectrolyte layer.

The type of polyelectrolyte thus has a large influence on the development of the thickness of the PEM and it was therefore considered important to investigate how the salt concentration in the polyelectrolyte solution affected the thickness of the PEM. Note that no salt was added to the MFC solutions and that the adsorption of MFC was always in Milli-Q water. The results of these studies for PDADMAC and PAH are shown in Figure 4a,b, respectively. These figures show that there is indeed a large influence of salt concentration on the build-up of the PEM. With PDADMAC, the thickness of the MFC layers increases when changing from

PEMs of MFC and Cationic Polyelectrolytes

Figure 5. AFM image of one bilayer of PEI and MFC on an oxidized silicon wafer. In this case, no additional electrolyte has been added.

Figure 6. Interference colors of films of MFC and PEI as a function of the number of layers. For example, 12 in this figure hence means a combination of six layers of PEI and six layers of MFC. No additional electrolyte was added.

no salt to 0.1 M NaCl, but decreases when the NaCl concentration is further increased to 0.5 M NaCl. With PAH, the layer thickness increases when the electrolyte concentration is increased to 0.1 M NaCl and when it is further increased to 0.5 M NaCl. These results show that the adsorption and adsorption conformation of the polyelectrolyte have a profound effect on the build-up of the multilayers of MFC and polyelectrolyte. Properties of the MFC/Polyelectrolyte Layers. In order to further characterize the polyelectrolyte/MFC layers, they were subjected to AFM analysis and simple light reflectance characterization. Figure 5 shows an AFM image of a bilayer of PEI/ MFC deposited on a silicon wafer. The fibrils can be clearly distinguished, and there seems to be a rather open fibrillar network on the surface for this bilayer. A direct measurement of the width of the fibrils in this image shows values between 4 and 11 nm with an average of 6 nm, taking into consideration the broadening due to the geometry of the tip. This is in good agreement with the data from the TEM micrographs. Films consisting of a combination of MFC and PEI also showed the development of clear interference colors, as shown in Figure 6. This indicates that the films formed with an increasing number of bilayers of MFC and PEI are smooth and give a well-defined increase in thickness for each bilayer so that, once an initial thickness has been achieved, a color change occurs with each additional bilayer.

Discussion Influence of the Presence of Charged Groups on the Surface Potential of the MFC and the Interaction between MFC Fibrils. A deeper understanding of the dispersion characteristics of MFC is important not only in an academic context but also when optimizing the use of this material in different applications.

Langmuir, Vol. 24, No. 3, 2008 789

The results in Figures 3-6 also show that it is possible to form well-organized layers of MFC and cationic polyelectrolytes, despite the strongly anisotropic nature of the MFC colloids. This implies that the fibrils do not aggregate as they approach the surface, despite their large specific surface area. In turn, this indicates that the adsorption of the MFC to the presaturated surface is a type of self-organization process where the electrostatic repulsion between the fibrils prevents them from aggregating. In Figure 3 it can also be seen that the fibrils are adsorbed as single entities and not as bundles or fibril flocs, and the development of clear interference colors, as shown in Figure 6, also shows that well-defined layers are formed. All this implies that the electrostatic repulsion between the fibrils is very strong and of large importance for the layer formation, and, in order to quantify this repulsion, it was decided to carefully characterize the surface potential of the fibrils and to estimate the interaction energy between fibrils. The relatively well-defined size and shape and uniform physicochemical characteristics of MFC facilitates the use of theoretical models to describe the dispersion properties and physical interactions. In a colloidal dispersion, the charge of a colloidal particle bearing weakly dissociating groups is dependent on a whole range of parameters that influence the dissociation. These parameters include the properties of the colloids themselves, the characteristics of the dispersing agent and the interplay between colloids and dispersing agent. In general, when performing theoretical calculations in order to predict interactions in a colloidal system consisting of weakly dissociating anionic colloids in an aqueous dispersion, it is convenient to start by describing the behavior of their weakly dissociating groups. If only one type of dissociating group exists and if the total amount is known, together with the surface area of the colloidal particles, the dissociation constant, Ka, can be expressed as -

Ka(T) )

ΓA [H+ surf] ΓHA

(1)

where ΓHA and ΓA- are the number of moles per square meter of the surface of species HA and A-, respectively, T is the absolute + ] is the concentration of protons at the temperature, and [H surf interface. Moreover, the degree of dissociation, R, is defined as

ΓA

R) Γ

A-

-

+Γ

) HA

Ka(T) Ka(T) + [H+ surf]

(2)

The MFC fibrils consist mainly of pure crystalline cellulose where all the carboxyl groups have been introduced at the outermost layer of the surface during the preparation procedure. The pKa value of the carboxyl groups is assumed to be 4.8. For such colloidal particles, the surface charge density, σ, can be calculated according to

σ ) eNAΓA

-

(3)

where e and NA are the charge of an elementary particle and Avogadro’s number, respectively. The above relations describe the colloidal particle in terms of its own characteristics, independent of the surrounding system. To model the properties of these colloids dispersed in water containing low molecular electrolytes (salt and protons/hydroxide ions) at arbitrary concentrations, it is necessary to consider the system as a whole, for instance, by solving the Poisson-

790 Langmuir, Vol. 24, No. 3, 2008

Wågberg et al.

Boltzmann (PB) equation:42-44

∇2ψ(r) ) -

eNA 0r(T)

∑j

(

)

- ezjψ(r)

zjcj exp

kT

(4)

where ψ is the electrostatic potential at an arbitrary distance, r, from the surface of a colloidal particle, 0 is the permittivity in vacuum, r is the relative permittivity of the dispersant, zj is the valence of ion j of the electrolyte, and k is the Boltzmann constant. The PB equation describes the electrostatic mean potential and all the charge interactions in a dispersion of colloidal particles. The electrostatic potential of a flat surface, ψsurf, is derived from Gauss law as

dψsurf σ )dr 0r(T)

(5)

In a low charged system, i.e., when |ezjψsurf/kT| < 1, it is possible to linearize the PB equation to give the Debye-Hu¨ckel relation, where the surface potential is simply obtained as ψsurf ) σκ/ 0r(T), where κ is the inverse Debye length.44 The limiting potential is then 25.6 mV for a 1:1 electrolyte at 298 K. This criterion is, however, often not fulfilled, and eq 4 must be solved numerically. This is the case for MFC over large parts of the dissociation range. The variation in the electrostatic potential with distance from the particle surface at a given ionic strength and degree of dissociation was calculated by dividing the MFC dispersion into cylindrical cells and solving eq 4 in cylindrical geometry43 at a volume fraction of 10-4 (0.15 g MFC/dm3 aqueous electrolyte solution). Since the diameters of the cylindrical MFC fibrils were found to be in the range of 5-15 nm, two maximum surface charge densities at complete dissociation (93.2 and 280 mC m-2) were calculated for the fibrils based on the total charge of the MFC material divided by the total surface area of fibrils of one specific diameter. As a result, the differences that are extracted in further calculations are expected to describe the extremity of the MFC material. From this, the pH value of the bulk dispersion could be calculated from the negative logarithm of the Boltzmann distribution of protons in the system according to

pH ) -log[H+ surf] - log exp

( ) eψsurf kT

Figure 7. Calculated degree of dissociation as a function of pH in the bulk dispersion of MFC rods of two different diameters at different ionic strengths and a temperature of 298 K: (a) diameter 5 nm and (b) diameter 15 nm.

(6)

Figures 7 and 8 respectively show the degree of dissociation and the surface potential versus the pH value of the dispersion for MFC rods with diameters of 5 and 15 nm at different ionic strengths. Figure 7a shows that the degree of dissociation increases with increasing ionic strength of the dispersion, giving an apparent pKa shift for the 5 nm MFC rods from approximately 7.4 at 1 mM to around 5.3 at a monovalent electrolyte concentration of 300 mM. Once more it has to be pointed out that the calculations are based on a typical dissociation behavior of carboxyl groups with an assumed pKa of 4.8, and that the relatively large differences between the pKa value of the individual carboxyl groups and the calculated apparent pKa values obtained for the MFC material are due to the electrostatic effects arising from the surface coverage of carboxyl groups on the fibrils. By increasing the ionic strength, the electrostatic surface potential decreases, and the apparent (42) Gunnarsson, G.; Jo¨nsson, B.; Wennerstro¨m, H. J. Phys. Chem. 1980, 84, 3114. (43) Jo¨nsson, B.; Wennerstro¨m, H. J. Colloid Interface Sci. 1981, 80, 482. (44) Evans, D. F.; Wennerstro¨m, H. The Colloidal Domain; VCH Publishers, Inc.: New York, 1994.

Figure 8. Calculated surface potential as a function of pH in the bulk dispersion for MFC rods of two different diameters at different ionic strengths and a temperature of 298 K: (a) diameter 5 nm and (b) diameter 15 nm.

pKa value of the MFC moves closer to the pKa value of the individual carboxyl groups. As shown in Figure 7b, the effect of the electrostatics on the dissociation behavior is even more pronounced for the fibrils with a diameter of 15 nm, where the apparent pKa is shifted to higher values approximately one unit more than in the former case.

PEMs of MFC and Cationic Polyelectrolytes

Langmuir, Vol. 24, No. 3, 2008 791

Figure 9. The calculated net interaction energy at different ionic strengths versus separation distance between two MFC rods with the same diameters (5 or 15 nm) at 298 K and at two different pH values (pH 3.5 or 7.0), as predicted by the DLVO theory: (a) I ) 1.0 mM, (b) I ) 10 mM, (c) I ) 100 mM, (d) I ) 300 mM.

On the other hand, the change of the surface potential over the pH scale shown in Figure 8a,b indicates that, despite a lower degree of dissociation, the surface potential is higher for a MFC rod with a diameter of 15 nm than for a 5 nm MFC-rod at a given pH value. These differences in dissociation behavior and in surface potential are expected and due mostly to a higher calculated surface charge density but, to a minor extent, are also due to geometrical effects, depending on the difference in curvature of rods of different diameters. In order to estimate the colloidal stability in the MFC system, the Derjaguin-Landau-Verwey-Overbeek (DLVO) theory44 was applied, and the attractive and repulsive interaction potentials were calculated using the geometry of crossed cylinders to model the attractive van der Waals interaction potential according to the equation

{

V(r) ) πR -

}

H121 1 64kTcbγsurf2 exp(-κr) + 6π r κ2

(7)

where R denotes the radius of the MFC rods, H121 is the effective Hamaker constant for two cellulose surfaces separated by water, cb denotes the electrolyte bulk concentration in moles per cubic meter, and γsurf is defined as being equal to tanh(zeψsurf/4kT). Figure 9 shows the calculated interaction energy as a function of the separation distance of two MFC rods dispersed in water at four different concentrations of a 1:1 electrolyte and at two different pH values. The interaction curves show no secondary energy minimum, and the primary minimum is situated close to the particle surfaces. In the calculations, an effective Hamaker constant of H121 ) 3.5 × 10-21 J was used,45 and a temperature of 298 K was chosen.

As shown in Figure 9a, the theory predicts that, already at an ionic strength of 1.0 mM, the repulsive barrier between MFC rods 5 nm in diameter is probably too low to avoid irreversible flocculation at pH 3.5. In Figure 9b, where I ) 10 mM, the 5 nm fibrils and possibly the 15 nm fibrils are expected also to coagulate at pH 3.5. Figure 9c,d shows that, when the ionic strength is further increased, irreversible flocculation is also predicted by the calculations for the 15 nm MFC rods at the lowest pH value, and the repulsive threshold for the 5 nm fibrils at pH 7 is reduced drastically, but stability is still expected. These calculations, based on the charge and size characterization of the MFC, show that the MFC used in the present work will have a very high surface potential when all the charges have been dissociated. It is also obvious that the pH must be increased to above 7-10, depending on the salt concentration, before all the carboxyl groups are fully dissociated. In the present experiments, the pH was kept at ca. pH ) 7, which means that the degree of dissociation is sometimes far from 1. These calculations also clearly show that increasing the salt concentration from 1 mM NaCl to 300 mM NaCl and keeping the pH constant at 7 would lead to a moderate decrease in surface potential from -170 mV to about -100 mV (Figure 7b). This is a combined effect on the surface potential of an increased degree of dissociation of the carboxyl groups at higher salt concentrations and the influence of an increased ionic strength of the solution. The results also show that the stability, i.e., the electrostatic repulsion between the fibrils, is high once the charged groups are dissociated. This can explain the ability of these fibrils to form smooth layers, since the fibrils will strongly repel each other when they approach the oppositely charged polyelectrolyte(45) Notley, S. M.; Pettersson, B.; Wågberg, L. J. Am. Chem. Soc. 2004, 126, 13930.

792 Langmuir, Vol. 24, No. 3, 2008

Wågberg et al.

Figure 10. Differential thickness measured with ellipsometry for the PEI and MFC layers as a function of layer number. No extra salt was added, and the pH value was kept between 7 and 8 in all the steps.

covered silicon wafer. As long as the pH is sufficiently high and the electrolyte concentration is at a moderate level, the fibrils will avoid each other and adapt to the positively charged interface in a self-avoiding way, leading to the formation of a well-defined MFC layer. Influence of the Type of Polyelectrolyte and Ionic Strength on the Structure of the Formed Multilayers. As was shown in Figures 3 and 4, the type of polyelectrolyte and the salt concentration during the adsorption of the polyelectrolyte have a large influence on the structure of the multilayers formed. The differential increases in thickness when adsorbing PEI and MFC, respectively, calculated from the data in Figure 3, are shown in Figure 10. As can be seen in this figure, the differential thickness for the MFC reaches an equilibrium level of about 20 nm after 10 layers, and the equilibrium thickness for the PEI is about 3 nm after 10 layers. These levels are very interesting, considering previously published data regarding the thickness of PEI layers and the thickness data for the MFC discussed earlier in this report. According to Horn et al.,46 the dry thickness of a PEI molecule with a molecular mass of 37 000, compared with 20 000 in the present work, was 0.6 nm, but the patch size of the same polymer molecule in the dry state on a mica surface was 13 nm. The hydrodynamic diameter of the polymer, determined with dynamic light scattering, was 6 nm, and the change in the dimensions upon adsorption was ascribed to a flattening of the polyelectrolyte due to the interaction with the surface.46 However, considering the differences in substrates and measuring methods, these dimensions are in the same size range as the 3 nm found for the PEI layer in the present work. The data for the MFC shown in Figure 10 are also very close to the data for the MFC determined from the TEM micrographs that showed diameters of MFC between 5 and 15 nm. These data thus indicate that it takes about five bilayers to fully saturate the surface with PEI and to reach an equilibrium situation where each subsequent layer has the same change in thickness for both the PEI and the MFC. This is also to some extent supported by the AFM image shown in Figure 5, which indicates that, with a single bilayer, there is still a fairly large number of noncovered areas on the surface. Considering the anisotropic nature of the MFC particles, this development of very well-defined layers of MFC is a somewhat (46) Pfau, A.; Schrepp, W.; Horn, D. Langmuir 1999, 15, 3219.

unexpected result, since it might have been anticipated that these materials would yield a more randomized, network adsorption structure. However, considering the very high length-to-width ratio of these materials, which will result in a high flexibility, and the high surface potential of the material, it can be suggested that these particles will adsorb to the oppositely charged surface in a self-avoiding way that will in turn lead to a more dense packing on the surface. As was also shown in Figure 3, the type of cationic polyelectrolyte chosen has a profound effect on the build-up of the PEM with MFC. The linear polyelectrolytes, PDADMAC and PAH, resulted in a much lower increase in thickness per layer than the PEI. The results in Figure 4a,b also show that a higher concentration of NaCl during adsorption of the polyelectrolyte leads to a larger increase in thickness per layer for the linear polyelectrolytes. In the case of PDADMAC, a decrease was detected at the highest salt concentration, whereas the PAH also showed an increase in thickness per layer with 0.5 M NaCl. In order to show how the thickness per layer changed for PDADMAC and PAH at different salt concentrations, the differential thickness was calculated for the MFC layers, and the results are shown in Figure 11 (PDADMAC in 11a and PAH in 11b). Note that, in this case, the MFC layers were all formed in Milli-Q water. Figures 10 and 11 show that, under salt-free conditions, PEI gives a layer thickness of about 20 nm after five bilayers, whereas PAH gives a layer thickness of around 2 nm, and, for the PDADMAC, no leveling off can be detected, i.e., the layer thickness increases with each layer. Obviously, for polyelectrolytes with a very flat adsorption conformation, i.e., PDADMAC and PAH, the increase in thickness with the MFC will be much lower. It is also clear in Figure 11a,b that a higher salt concentration leads to the formation of much thicker MFC layers. For PDADMAC in 0.1 M NaCl, there is a continuous increase in the layer thickness up to about nine MFC layers, after which a thickness of around 35 nm is reached, whereas, at 0.5 M NaCl, no leveling off is observed, and after 10 layers, there is a rapid decrease in thickness. For PAH, the situation is different, and at 0.1 M NaCl, there is a leveling off at 15 nm after five layers, and at 0.5 M NaCl the leveling off occurs at about 30 nm after eight layers. There may be several explanations of this behavior, but the increase in electrolyte concentration will have different

PEMs of MFC and Cationic Polyelectrolytes

Langmuir, Vol. 24, No. 3, 2008 793

Figure 11. Differential thickness for the formation of each MFC layer (in deionized water) as a function of layer number at different NaCl concentrations during adsorption of the polyelectrolyte layer for (a) PDADMAC and (b) PAH. The pH was between 7 and 8 in all cases.

effects. It is well-known that an increase in salt concentration will cause a coiling of the polyelectrolyte chain, at least in dilute polyelectrolyte solutions.47 This will lead to an adsorption conformation more similar to that of PEI for both PDADMAC and PAH, i.e., a polyelectrolyte with a more pronounced threedimensional (3D) structure at the interface will result in thicker MFC layers. This influence of the chemical structure of the polyelectrolyte on the build-up of the layers has been reported earlier,20,21 where it was shown that a 3D structure of the polyelectrolyte leads to a more direct formation of PEM at low layer numbers,20 and that the use of PEI as an anchoring layer affects the thickness of the entire PEM, even at higher layer numbers.21 It is hence clear that the structure of the adsorbed polyelectrolyte has a large influence on thickness development when combining the polyelectrolyte with MFC in Milli-Q water. The exact reason for this is currently investigated in our laboratory. It is nevertheless evident that this opens the way to tailor multilayers of polyelectrolytes and MFC that may be useful in specific applications. In the case of SiO2 surfaces, an increase in electrolyte concentration leads to a higher charge of the surface,48 and this will no doubt lead to a higher adsorption of the cationic polyelectrolyte that will in turn lead to a higher adsorption of MFC in the next layer. This is valid for the development of the first layers of the PEM, but after three to five layers, the influence of the substrate properties diminishes, and the interaction between the MFC and the polyelectrolyte will determine the build-up of (47) Smits, R. G.; Kuil, M. E.; Mandel, M. Macromolecules 1993, 26, 6808. (48) Bolt, G. H. J. Phys. Chem. 1957, 61, 1166.

the PEM.4 Considering the results in Figures 10 and 11, it is improbable that the change in surface charge can explain the difference in thickness of the MFC layers between 5 and 10 layers of the PEM. This also means that the difference between the different investigated systems is due to the interaction between the polyelectrolytes and the MFC, and not due to the interaction between the polyelectrolytes and the bare surface. In the case of pure electrosorption,16 it is well-known that the adsorption of polyelectrolytes decreases with increasing salt concentration, and the decrease in thickness for the PDADMAC/ MFC combination at 0.5 M NaCl is an indication that the adsorption of this polyelectrolyte is driven mainly by the release of counterions from the polyelectrolyte and from the MFC. The fact that the decrease occurs upon changing from 0.1 M NaCl to 0.5 M NaCl might be traced back to the results shown in Figure 7 that show that the degree of dissociation of the charges on the cellulose is not significantly increased when the electrolyte concentration is increased above 0.1 M NaCl. As the electrolyte concentration is increased to 0.5 M NaCl, the only effect of the electrolyte will hence be to decrease the driving force for polyelectrolyte adsorption.16 Since the MFC was adsorbed at low salt concentrations and the polyelectrolytes were adsorbed at different salt concentrations, it is suggested that the changes in the degree of dissociation of the charges of the MFC occurs in the adsorbed state. The results might also suggest that, at higher salt concentrations, the interaction between the MFC substrate and the PDADMAC is so low so that, at higher layer numbers, there might be a desorption of the already adsorbed

794 Langmuir, Vol. 24, No. 3, 2008

Wågberg et al.

PEM. Considering the rather high stability of layer build-up, this latter explanation is unlikely. With PAH, the situation is quite different. For this polyelectrolyte, there is still an increase in thickness also at 0.5 M NaCl. This can have several explanations. First of all, the charge of this polymer is highly salt- and pH-dependent, and, as the salt concentration is increased, the charge of the polymer will most probably also increase at pH 7. Since the theoretical charge of this polymer is 23 mequiv g-1 when it is fully protonated, it can be assumed that its charge at high salt concentration at pH 7 could probably be around 10 mequiv g-1. This means that this polymer could tolerate higher salt concentrations before its adsorption starts to decrease at increased salt concentrations,16 despite the fact that the degree of dissociation of the cellulose charges has reached a maximum around a salt concentration of 0.3 M NaCl at pH 7. An alternative explanation could be that there is a nonionic interaction between the PAH and the cellulose, but there are no data available supporting this explanation. However, more experiments are needed to test this hypothesis. All these results thus indicate that it is the conformation of the adsorbed polyelectrolyte in combination with the adsorbed amount of polymer that affects the thickness of the adsorbed MFC layers. They also suggest that the build-up of the layers is controlled by electrostatic interactions between the polyelectrolytes and the MFC, and that the difference between different systems can be traced back to the degree of dissociation of the charges on the polyelectrolytes and the MFC and their dependence on salt concentration. The exact mechanistic reason for this and the relative influence of the structure and the adsorbed amount are currently under investigation in our laboratory. As was shown in Figure 6, the films show the development of different interference colors as the thickness of the PEM is increased. This indicates that the layers are smooth enough on a level less than the wavelength of light to separate the different colors in the white light spectrum from each other. This has previously been reported for multilayers of colloidal particles,2 as well as for thin films on a glass support.49 For monochromatic light, the maximum in reflection can be related to the thickness of the film and the wavelength of the light by the simple relationship49

t)

λm 2n

(8)

where t is the thickness of the film, λ is the wavelength of the incident light, n is the refractive index of the film, and m is an integer related to the number of reflection maxima.

Table 1. The Relation between Interference Colors and Film Thickness Calculated According to eq 8 According to Relations from Vaciceka and a Comparison of These with Values from Ellipsometric Measurementsb color violet dark blue light blue

wavelength thickness (nm) thickness (nm) thickness (nm) (nm) from eq 8 ellipsometry (Vacicek)a 400 450 500

162 (130) 182 (147) 202 (163)

91 113 139

113 (93) 144 (119) 167 (138)

a Reference 46. b The refractive index has been assumed to be a geometrical average of those of air and cellulose or to be solid cellulose (values in brackets in columns 3 and 5).

This is, however, only valid for monochromatic light. For white light, the entire spectrum has to be taken into consideration when calculating the occurrence of the maxima in reflectance.49 In the present work, it has not been possible to determine whether the films are porous, and it is therefore difficult to determine or estimate a refractive index of the film, which is necessary in order to calculate the relationship between the thickness and the interference colors. As a first approximation, it can be assumed that the films are porous and that the refractive index is the geometrical mean of the indices of air and cellulose, i.e., (nair × ncell)0.5. Using this value, the relationship between film thickness and the wavelength of the interference color can be calculated using either eq 8 or the relationships where the entire white light spectrum has been taken into account.49 These values can then be compared with the data from the ellipsometry measurements. A value of ncell ) 1.5350,51 was assumed, and the results of the calculations are summarized in Table 1. This table includes results (in brackets) where it has been assumed that the films consist of solid cellulose, i.e., n ) 1.53. The values for the wavelengths of the colors have been selected from the work of Vacicek.49 The data in column 3 calculated by using eq 8 are, as expected, very different from the values determined by ellipsometry, even if a solid cellulose film is assumed. The values calculated according to Vacicek49 are in very good agreement with the ellipsometry values assuming that the MFC film is solid cellulose. This indicates that the dry films of the MFC are basically solid cellulose, once they are dried. It must also be pointed out that the influence of the polyelectrolyte on the optical properties of the films has been neglected since it constitutes a very small amount by weight and by volume compared with the cellulose. Formation of PEM by Spraying. In practical applications, it may be difficult to use a layer-by-layer deposition where the surfaces are consecutively dipped and rinsed between each deposited layer, and alternative application techniques are

Figure 12. Formation of a PEM from PEI and MFC by dipping and by spraying in the absence of added salt. The pH was between 7 and 8.

PEMs of MFC and Cationic Polyelectrolytes

therefore of interest. Spraying of the multilayers has previously been suggested as an alternative method for PEM formation,40 and it was therefore decided to investigate whether this technique could be also applicable for the formation of PEM from PEI and MFC under salt-free conditions. The results are shown in Figure 12. As can be seen, it is possible to form layers with the spraying technique that are almost as regular as those formed with the dipping technique. The spray-formed PEM also showed clear interference colors, but there was a somewhat larger color variation within each layer.

Conclusions The present work has shown that it is possible to prepare a stable dispersion of nanosized fibrils of cellulose by subjecting carboxymethylated cellulose fibers to a high-pressure homogenization followed by ultrasonication and centrifugation. Clear dispersions with a concentration of 1-2 g L-1 can be prepared directly by this procedure. TEM and AFM characterization of the fibrils have shown that they have a diameter between 5 and 15 nm and a length of up to 1 µm and charge measurements show that the material has a charge of 515 µequiv g-1. Using these data and the PB equation, it was possible to calculate the surface potential of the fibrils and how it changed with changes in the pH and salt concentration. Assuming that all the charges were located on the surface of the fibrils, it was found that the surface potential, at full dissociation of the carboxyl groups, was 200 mV for fibrils with a diameter of 5 nm and ca. 250 mV for the 15 nm fibrils at 1 mM NaCl. It was also shown that a pH higher than 10 was required to dissociate all the charges of the MFC. These values for the surface potential together with the Hamaker constant for cellulose were then used to calculate the interaction (49) Vacicek, A. Optics of Thin Films; North-Holland Publishing Company: Amsterdam, 1960. (50) Bergstro¨m, L.; Stemme, S.; Dahlfors, T.; Arwin, H. Cellulose 1999, 6, 1. (51) Holmberg, M.; Berg, J.; Stemme, S.; O ¨ dberg, L.; Rasmusson, J.; Claesson, P. J. Colloid Interface Sci. 1997, 186, 369.

Langmuir, Vol. 24, No. 3, 2008 795

energy between rods of MFC in water with different salt concentrations, using the DLVO theory. The results of these calculations show that the stability, i.e., the electrostatic repulsion between the fibrils, is high once the charged groups are dissociated. They also show that a moderate increase in NaCl concentration leads only to a minor decrease in colloidal stability of the material, since the degree of dissociation of the charged groups is significantly increased as the salt concentration is increased, and this counteracts the decrease in surface potential caused by the addition of salt. It has also been shown that it is possible to prepare PEMs with the MFC and different cationic polyelectrolytes. The 3D structure of the polyelectrolytes on the solid surface used for PEM preparation seems to have a large influence on the development of layer thickness. The combination of PEI and MFC in deionized water results in the formation of regular layers of MFC and PEI with layer thicknesses of 20 and 3 nm, respectively, after deposition of about 10 layers. By changing the salt concentration during adsorption of PDADMAC and PAH, it was possible to control the thickness of the PEM. The PEMs had different colors depending on the thickness of the multilayers, and simple estimations of the thickness of the PEM from the colors, assuming dense cellulose layers, showed surprisingly good agreement with data from ellipsometry measurements. This indicates that the PEMs are basically compact films of cellulose with some cationic polyelectrolyte mixed/intercalated between the fibrils. Acknowledgment. TEM images were kindly prepared by Marc Schmutz at Institute Charles Sadron, Strasbourg, France, and Dr. J. A. Bristow is gratefully acknowledged for a linguistic revision of the manuscript. Supporting Information Available: Interface colors of PEMs of MFC and PEI. This material is available free of charge via the Internet at http://pubs.acs.org. LA702481V