Article pubs.acs.org/Biomac
Direct Adhesive Measurements between Wood Biopolymer Model Surfaces Emil Gustafsson,*,†,‡ Erik Johansson,‡ Lars Wågberg,†,‡ and Torbjörn Pettersson†,‡ †
Wallenberg Wood Science Center and ‡Department of Fibre and Polymer Technology, School of Chemical Science and Engineering, KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden S Supporting Information *
ABSTRACT: For the first time the dry adhesion was measured for an all-wood biopolymer system using Johnson−Kendall−Roberts (JKR) contact mechanics. The polydimethylsiloxane hemisphere was successfully surface-modified with a Cellulose I model surface using layer-by-layer assembly of nanofibrillated cellulose and polyethyleneimine. Flat surfaces of cellulose were equally prepared on silicon dioxide substrates, and model surfaces of glucomannan and lignin were prepared on silicon dioxide using spin-coating. The measured work of adhesion on loading and the adhesion hysteresis was found to be very similar between cellulose and all three wood polymers, suggesting that the interaction between these biopolymers do not differ greatly. Surface energy calculations from contact angle measurements indicated similar dispersive surface energy components for the model surfaces. The dispersive component was dominating the surface energy for all surfaces. The JKR work of adhesion was lower than that calculated from contact angle measurements, which partially can be ascribed to surface roughness of the model surfaces and overestimation of the surface energies from contact angle determinations.
■
INTRODUCTION The fibers found today in trees and annual plants are the result of an evolutionary process that has created an intricate chemistry and advanced nano- and mesostructures to optimize the fiber properties, and both fiber structure and chemistry are crucial for the final properties. The interactions between the wood biopolymers will have a profound effect on the mechanical properties of the fibers. However, we possess limited fundamental understanding of the importance of the adhesive interactions between the various chemical constituents of the fiber wall. In plant fiber walls, highly crystalline cellulose nanofibrils, with high stiffness and a Young’s modulus close to 130 GPa,1,2 are embedded in a matrix of lignin and hemicelluloses in a lamellar structure3 that lets the plant utilize the high Young’s modulus of the cellulose nanofibrils. The adhesive interactions are also crucial for the properties of fiber/ fiber joints in paper and for the interactions in new biodegradable and renewable materials containing lignocellulosic fibers. Because of the importance of this adhesive interaction and to fill this gap in knowledge, the focus of the present work has been to measure the adhesion between flat wood biopolymer model surfaces and hemispherical polydimethylsiloxane (PDMS) caps coated with nanofibrillated cellulose (NFC), analyzing the results according to Johnson−Kendall− Roberts (JKR) theory.4 Wood fibers are chemically heterogeneous with high surface roughness and are therefore not suited for fundamental adhesion studies, which require smooth, chemically homoge© XXXX American Chemical Society
neous surfaces. Adequate model surfaces prepared from different wood biopolymers are therefore necessary for such measurements. Various procedures for preparing model surfaces of cellulose, lignin, and hemicelluloses are described in the literature.5−16 Previous work on cellulose model surfaces has mostly been based on spin-coated surfaces produced from regenerated cellulose5−8 and from suspensions of cellulose nanocrystals (CNCs),9 and, using these starting materials, one can prepare cellulose model surfaces with different degrees of crystallinity.10 The interaction between cellulose surfaces in water has been studied using surface force apparatus (SFA)5 and colloidal probe atomic force microscopy (AFM),17−20 and it was demonstrated that the surface charge and roughness was crucial for the cellulose interactions.19,20 The colloidal probe technique has also been used to measure the adhesion in air between a silica sphere and chemically treated hemp fibers.21 It has been demonstrated that cellulose model surfaces can be made from CNC11 or NFC12,13 in combination with cationic polyelectrolytes using the layer-by-layer (LbL) technique.22 NFC-containing LbL films are relevant model surfaces because they retain both the native cellulose I crystalline ordering10 and the natural fibrillar structure. The interactions between surfaces coated with LbL films containing NFC23 and CNC24 in water have been investigated. Received: May 15, 2012 Revised: August 20, 2012
A
dx.doi.org/10.1021/bm300762e | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
Since the early 1970s,14 lignin model surfaces have been prepared in many ways using various lignin sources and isolation techniques, as summarized by Norgren et al.,15 who also prepared smooth, stable lignin model surfaces by spincoating softwood kraft lignin dissolved in aqueous alkaline solutions. Later, Notley and Norgren studied the interaction between these surfaces and cellulose in water at various salt concentrations25 and used these surfaces to determine the surface energy of lignin from contact angle measurements.26 These lignin films were demonstrated to be only partially wetted by water due to the lower surface energy and lower polarity of the lignin surface.26 Spin-coated lignin model surfaces have also recently been prepared from milled wood lignin.26,27 Hemicelluloses are found in different amounts in wood. In softwoods, hemicelluloses from the galactoglucomannan (GGM) family are predominant.3 Two main GGMs can be found, one galactose-rich type that is water-soluble and one galactose-poor type, called glucomannan (GM), that is water insoluble.3,28 Very few studies have examined interactions involving hemicellulose model surfaces, but one example is a recent AFM study of model surfaces of xyloglucan grafted to gold.16 Studies have examined the adsorption of various wood hemicelluloses to cellulose29−32 and lignin30 model surfaces, and other studies have examined the interaction between substrates and adsorbed hemicelluloses.33,34 In the early 1990s, Chaudhury and Whitesides35 developed an experimental protocol based on JKR contact mechanics theory,4 which made it possible to measure adhesion using soft hemispherical PDMS caps. The JKR approach has been used to measure the adhesion between PDMS modified with selfassembled monolayers,35 solvent-casted poly(methyl methacrylate) (PMMA) and polystyrene (PS) polymer films,36 and LbL films of PAH/PAA.37 Significant adhesion hysteresis in both the PS−PS and the PMMA−PMMA contacts was reported, and the surface energies of PS and PMMA determined from the JKR unloading data agreed well with the surface energy values obtained from SFA measurements of the same systems.36 In the LbL study, the adhesion decreased when the caps were coated, which was suggested to be due to increased surface roughness in combination with the high Young’s modulus of the LbL film.37 JKR theory has previously been used in measuring the adhesion between unmodified PDMS and cellulose model surfaces with different degrees of crystallinity;38−40 while the different surfaces exhibited similar work of adhesion on loading, the adhesion hysteresis was somewhat higher for the amorphous cellulose surfaces than for the crystalline surfaces, although it was fairly high for all surfaces.40 The mechanisms of adhesion hysteresis between cross-linked PDMS hemispheres41 and between PDMS and silicon oxide substrates42,43 have been studied using JKR theory. In summary, the present work aimed to use the JKR approach to examine the work of adhesion between smooth, well-defined model surfaces of wood biopolymers under ambient conditions to gain fundamental understanding of interactions both within the fiber wall and in fiber/fiber joints in paper. Flat model surfaces of cellulose were prepared using LbL deposition of NFC and polyethylene imine (PEI), and spin-coating was used to prepare flat lignin and GM model surfaces. LbL films of NFC and PEI, previously demonstrated to appear as both uniform and smooth films on PDMS,44 were used to coat the hemispherical PDMS caps used in the JKR experiments. To the knowledge of the authors, this is the first
time the JKR approach has been used to examine the adhesive interaction between two wood components. This new approach of coating PDMS with wood biopolymers using LbL deposition permits further examination of the adhesion between wood biopolymer surfaces and the influence of, for example, physical and chemical modifications, humidity, and drying. The model surfaces were characterized using AFM, X-ray photoelectron spectroscopy (XPS), and contact angle measurements, and the surface energies were calculated using Van Oss−Good theory.45,46 The work of adhesion as calculated from the surface energies was compared with the JKR results.
■
EXPERIMENTAL SECTION
Materials. Poly(allylamine hydrochloride) (PAH; Mw = 70 000 g mol−1) was purchased from Sigma-Aldrich (Munich, Germany) and poly(ethylene imine) (PEI; Mw = 60000 g mol−1, 53% aqueous solution) was obtained from Acros Organics (Geel, Belgium); both polycations were used without further purification. NFC was prepared at Innventia AB (Stockholm, Sweden) according to a method previously described.12,47 In brief, the NFC was prepared from commercial sulfite softwood dissolving pulp (Domsjö Dissolving Plus, Domsjö Fabriker, Ö rnskö ldsvik, Sweden) using high-pressure homogenization similar to a previously described method48 but with carboxymethylation49 pretreatment of the fibers. Commercial kraft lignin, Indulin AT, was obtained from MeadWestvaco (Richmond, VA) and was purified according to a procedure previously described by Norgren et al.25 The lignin was dissolved in a dioxane:water mixture (9:1) and stirred for 2 h at room temperature to remove carbohydrates. The solution was centrifuged, and the residue, containing undissolved carbohydrates, was removed. The dioxane was thereafter evaporated in a rotavapor and the lignin was freezedried. To remove extractives, the freeze-dried lignin was subjected to pentane extraction for 8 h. Glucomannan from Norway spruce (Picea abies) wood holocellulose had been extracted with NaOH/H3BO4 and precipitated with Fehling reagent.28 The precipitates were washed with deionized water, dissolved in 1 M HCl, and further precipitated and washed with ethanol. The purified GM was composed of galactose, glucose, and mannose in a
