9 Wood Adhesion and Adhesives Charles R. Frihart USDA, Forest Service, Forest Products Laboratory, Madison, WI CONTENTS 9.1 General..................................................................................................................................216 9.2 Wood Adhesive Uses............................................................................................................217 9.3 Terminology..........................................................................................................................219 9.4 Application of the Adhesive.................................................................................................220 9.4.1 Adhesive Application to Wood ................................................................................220 9.4.2 Theories of Adhesion ...............................................................................................221 9.4.3 Wood Adhesion ........................................................................................................225 9.4.4 Wood Surface Preparation........................................................................................225 9.4.5 Wood Bonding Surface ............................................................................................226 9.4.6 Spatial Scales of Wood for Adhesive Interaction ....................................................229 9.4.7 Wetting and Penetration in General .........................................................................230 9.4.8 Wetting, Flow, and Penetration of Wood .................................................................231 9.5 Setting of Adhesive ..............................................................................................................234 9.5.1 Loss of Solvents .......................................................................................................235 9.5.2 Polymerization ..........................................................................................................236 9.5.3 Solidification by Cooling .........................................................................................237 9.6 Performance of Bonded Products ........................................................................................238 9.6.1 Behavior under Force ...............................................................................................238 9.6.2 Effect of Variables on the Stress-Strain Behavior of Bonded Assemblies..............................................................................................241 9.6.3 Bond Strength...........................................................................................................242 9.6.4 Durability Testing .....................................................................................................245 9.7 Adhesives..............................................................................................................................246 9.7.1 Polymer Formation ...................................................................................................246 9.7.2 Self-Adhesion ...........................................................................................................248 9.7.3 Formaldehyde Adhesives..........................................................................................249 9.7.3.1 Phenol Formaldehyde Adhesives ..............................................................250 9.7.3.2 Resorcinol and Phenol-Resorcinol Formaldehyde Adhesives..................252 9.7.3.3 Urea Formaldehyde and Mixed Urea Formaldehyde Adhesives .............254 9.7.3.4 Melamine Formaldehyde Adhesives .........................................................255 9.7.4 Isocyanates in Wood Adhesives ...............................................................................257 9.7.4.1 Polymeric Diphenylmethane Diisocyanate...............................................258 9.7.4.2 Emulsion Polymer Isocyanates .................................................................260 9.7.4.3 Polyurethane Adhesives ............................................................................260 9.7.5 Epoxy Adhesives ......................................................................................................261 9.7.6 Polyvinyl and Ethylene-Vinyl Acetate Dispersion Adhesives.................................263 9.7.7 Biobased Adhesives ..................................................................................................265 0-8493-1588-3/05/$0.00+$1.50 © 2005 by CRC Press LLC 215 216 Handbook of Wood Chemistry and Wood Composites 9.7.7.1 Protein Glues.............................................................................................265 9.7.7.2 Tannin Adhesives ......................................................................................266 9.7.7.3 Lignin Adhesives.......................................................................................267 9.7.8 Miscellaneous Composite Adhesion ........................................................................267 9.7.9 Construction Adhesives ............................................................................................268 9.7.10 Hot Melts ................................................................................................................ 269 9.7.11 Pressure Sensitive Adhesives ...................................................................................269 9.7.12 Other Adhesives........................................................................................................270 9.7.13 Formulation of Adhesives ...................................................................................... 270 9.8 Environmental Aspects .........................................................................................................272 9.9 Summary...............................................................................................................................272 References ......................................................................................................................................273 9.1 GENERAL The recorded history of bonded wood dates back at least 3,000 years to the Egyptians (Skeist and Miron 1990, River 1994a), and adhesive bonding goes back to early mankind (Keimel 2003). Although wood and paper bonding are the largest applications for adhesives, some of the fundamental aspects are not fully understood. Better understanding of the critical aspects in wood adhesion should lead to improved composites. The chemistry of adhesives has been covered in detail; however, how the adhesives hold wood together when under external and internal stresses need to be better understood from the basic scientific principles. This chapter is aimed at more in-depth coverage of those items that are not covered elsewhere. It will touch briefly on topics covered by other writers and the reader should examine the recommended books and articles for more details. Many of the books on adhesives and adhesion are long and complicated, but at least one is briefer, while still being quite thorough (Pocius 2002). Adhesives are designed for specific applications, leading to thousands of products (Rice 1990). Petrie has broken adhesives into 20 groups of synthetic structural, 11 groups of elastomeric, 12 groups of thermoplastic, and six groups of natural adhesives (Petrie 2000). Brief has summarized the vast number of markets for adhesives (Brief 1990). Understanding how an adhesive works is difficult since adhesive performance is not one science of its own, but the combination of many sciences. Adhesive strength is defined mechanically as the force necessary to pull apart the substrates that are bonded together. Mechanical strength is dependent upon primary and secondary chemical bonds of the polymer chains in the adhesive, wood and adhesive-wood interphase. Thus, one needs to consider both the chemical and mechanical aspects of bond strength, and the interrelation of the two factors. Because adhesive strength is a measurement of failure, the process determines where the localized stress exceeds the bond strength under specific test conditions. One concept is the idea of the bonded assembly being a series of links representing each phase with the failure occurring in the weakest link (Marra 1980). Although the bend is actually more a continuum than discrete links. The localized stress is usually very different from applied stress due to stress distribution and concentration (Dillard 2002). It is generally preferred that the adhesive bond be stronger than the substrate so that the failure mech- anism is one of substrate fracture. There are generally three steps in the process of adhesive bonding. The first is usually the preparation of the surface to provide the best interaction of the adhesive with the substrate. Even though a separate treatment step may not be used in some cases, the knowledge of material science (surface chemistry and morphology) is important for understanding this interaction. Preparation of the surface can involve either mechanical or chemical treatment or a combination of the two. In some cases, the adhesive is modified to deal with problems in wetting of the surface or contamination on the surface. Surface analysis techniques are often more difficult on wood than other materials due to the complex chemistry and morphology of the wood. Wood Adhesion and Adhesives 217 The second step is that the adhesive needs to form a molecular-level contact with the surface; thus, it should be a liquid so that it can develop a close contact with the substrates. This process involves both the sciences of rheology and surface energies. Rheology is the science of the defor- mation and flow of matter. Surface energies are determined by the polar and non-polar components of both the adhesive and the wood. Improving the compatibility by changing one or both of the components can lead to stronger and more durable bonds. The third step is the setting, which involves the solidification and/or curing of the adhesive. Most adhesives change physical state in the bonding process, with the main exception being pressure sensitive adhesives that are used on tapes and labels. The solidification process depends on the type of adhesive. For hot melt adhesives, the process involves the cooling of the molten adhesive to form a solid, whether this is an organic polymer as in craft glues, or an inorganic material as in the case of solder. Other types of adhesives have polymers dissolved in a liquid, which may be water (e.g., white glues) or an organic (e.g., rubber cement). The loss of the solvent converts these liquids to solids. The third type of adhesive is made up of small molecules that polymerize to form the adhesive, for example, super glues or two-part epoxies. Most wood adhesives involve both the polymerization and solvent loss methods. Understanding the conversion of small molecules into large molecules requires knowledge of organic chemistry and polymer science. Once the bond is prepared, the critical test is the strength of the bonded assembly under forces existing during the lifetime use of the assembly. This involves internal forces from shrinkage during the curing of the adhesive and differential expansion/contraction of the adhesive and substrate during environmental changes, or externally applied forces. Understanding the performance of a bonded assembly requires knowledge of both chemistry and mechanics. Often the strength of a bonded assembly is discussed in terms of adhesion. Adhesion is the strength of the molecular layer of adhesive that is in contact with the surface layer of the substrate, such as wood. The internal and applied energies may be dissipated at other places in the bonded assemblies than the layer of molecular contact between the adhesive and the substrate. However, failure at the interface between the two is usually considered unacceptable. Understanding the forces and their distribution on a bond requires knowledge of mechanics. An appreciation of rheology, material science, organic chemistry, polymer science, and mechan- ics leads to better understanding of the factors controlling the performance of the bonded assemblies; see Table 9.1. Given the complexity of wood as a substrate, it is hard to understand why some wood adhesives work better than other wood adhesives, especially when under the more severe durability tests. In general, wood is easy to bond to compared to most substrates, but it is harder to make a truly durable wood bond. A main trend in the wood industry is increased bonding of wood products as a result of the use of smaller diameter trees and more engineered wood products. 9.2 WOOD ADHESIVE USES Because adhesives are used in many different applications with wood, a wide variety of types are used (Vick 1999). Given the focus of this book on composites, the emphasis will be more on adhesives used in composite manufacturing than on those used in product assembly. Factors that influence the selection of the adhesive include cost, assembly process, strength of bonded assembly, and durability. The largest wood market is the manufacturing of panel products, including plywood, oriented strandboard (OSB), fiberboard, and particleboard. Except for plywood, the adhesive in these applications bonds small pieces of wood together to form a wood-adhesive matrix. The strength of the product depends on efficient distribution of applied forces between the adhesive and wood phases. The composites (strandboard, fiberboard, and particleboard) have adhesive applied to the wood (strands, fibers, or particles); then they are formed into mats and pressed under heat into the final product. This type of process requires an adhesive that doesn’t react immediately at room temperature (pre-mature cure), but is heat-activated during the pressing operation. Given the weight 218 Handbook of Wood Chemistry and Wood Composites TABLE 9.1 Wood Bonding Variables Resin Wood Process Service Type Species Adhesive amount Strength Viscosity Density Adhesive distribution Shear modulus Molecular weight Moisture content Relative humidity Swell–shrink resistance distribution Mole ratio of Plane of cut: radial, Temperature Creep reactants tangential, transverse, mix Cure rate Heartwood vs. sapwood Open assembly time Percentage of wood failure Total solids Juvenile vs. mature wood Closed assembly time Failure type Catalyst Earlywood vs. latewood Pressure Dry vs. wet Mixing Reaction wood Adhesive penetration Modulus of elasticity Tack Grain angle Gas-through Temperature Filler Porosity Press time Hydrolysis resistance Solvent system Surface roughness Pretreatments Heat resistance Age Drying damage Posttreatments Biological resistance: fungi, bacteria, insects, marine organisms pH Machining damage Adherend temperature Finishing Buffering Dirt, contaminants Ultraviolet resistance Extractives pH Buffering capacity Chemical surface Note: Norm Kutscha contributed most of the information for this table. of adhesive (2–8%) compared to the product weight, cost is an issue. In addition, since the wood surfaces are brought close together, gap filling is not an important issue, but over penetration is. On the other hand, for plywood, the surfaces are not uniformly brought in such close contact, requiring the adhesive to remain more above the surface. Light colored adhesives are important for some applications, but many of these products have their surfaces covered by other materials. Most of the adhesives used in wood bonding have formaldehyde as a co-monomer, generating concern about formaldehyde emissions. Dunky and Pizzi have discussed many of the commercial issues relating to the use of adhesives in manufacture and the use of wood composites (Dunky and Pizzi 2002). For laminating lumber and bonding finger joints, the adhesive can either be heat or room- temperature cured. The cost of the adhesive has become more critical as the thickness of the wood has decreased from glulam to laminated veneer lumber and parallel strand lumber (Moody et al. 1999). Generally, color is not critical unless it is in a trim application, but moisture and creep resistance are more important because these products are often used structurally. Adhesives used in construction and furniture assembly usually have long set times and are room-temperature cured. Furniture adhesives are light-colored, low-viscosity, and generally do not need much moisture resistance. On the other hand, construction adhesives generally have a high viscosity and need flexibility, but can be dark-colored. The movement away from solid wood for construction to engineered wood products has increased the consumption of adhesives. A wooden I-joist can have up to five different adhesives in its construction; see Figure 9.1. The wood laminates that form the top and bottom members may be finger joined with a melamine-formaldehyde adhesive and glued together with a phenol-resorcinol- formaldehyde adhesive. The OSB that forms the middle part is often produced using both phenol- formaldehyde and polymeric diphenylmethane diisocyanate adhesives. This middle section is then Wood Adhesion and Adhesives 219 FIGURE 9.1 The importance of adhesives is illustrated by the need for different adhesives to make the flange by the bonding of laminate pieces and the oriented strandboard from the flakes and the final I-joist by attachment of the strandboard to the flange. joined to the top and bottom members with emulsion-polymer isocyanate. Each of these adhesives has different chemistries, and some are bonded under different conditions of time, temperature, and pressure to a variety of wood surfaces, and is subjected to different forces during use. Thus, it is not surprising that a simple model for satisfactory wood adhesion has been difficult to derive. 9.3 TERMINOLOGY Confusion can be caused if there is not a clear understanding of the terminology; this chapter generally follows that given in the ASTM Standard D 907-00 (ASTM International 2000a). Adhesive joint failure is “the locus of fracture occurring in an adhesively-bonded joint resulting in the loss of load-carrying capability” and is divided into interphase, cohesive, or substrate failures. Cohesive failure is within the bulk of the adhesive, while substrate failure is within the substrate or adherend (wood). The least clear failure zone is that occurring within the interphase, which is “a region of finite dimension extending from a point in the adherend where the local properties (chemical, physical, mechanical, and morphological) begin to change from the bulk properties of the adherend to a point in the adhesive where the local properties are equal to the bulk properties of the adhesive.” Figure 9.2 shows the various regions of a bonded assembly. The bulk properties are the properties of one phase unaltered by the other phase. The assembly time is “the time interval between applying adhesive on the substrate and the application of pressure, or heat, or both, to the assembly.” This time can be closed with substrates brought into contact or open with the adhesive exposed to the air; these times are important to penetration of the adhesive and evaporation of solvent. Set is “to convert an adhesive into a fixed or hardened state by chemical or physical action, such as condensation, polymerization, oxidation, vulcanization, gelation, hydration, or evaporation of volatile constituents.” Cure is “to change the physical properties of an adhesive by chemical reaction....” Note that cure is only one way in the adhesive setting step. However, because cure is a function of how it is measured, there is no universal value for an adhesive. Separating partial cure from total cure is important because they usually have very different properties, and in most bonded products, total cure is not usually obtained. Tack 220 Handbook of Wood Chemistry and Wood Composites FIGURE 9.2 A transverse scanning electron microscope image of a resorcinol bond of yellow-poplar, showing the zones of bulk wood, interphase region, and bulk adhesive. is “the property of an adhesive that enables it to form a bond of measurable strength immediately after the adhesive and adherend are brought into contact under low pressure.” Tack is important for holding composites together during lay up and pre-pressing. A structural adhesive is “a bonding agent used for transferring required loads between adherends exposed to service environments typical for the structure involved” (ASTM International 2000a). For wood products, structural implies that failure can cause serious damage to the structure, and even loss of life (Vick 1999), while semi-structural adhesives need to carry the structural load, but failure is not as disastrous, and nonstructural adhesives typically support merely the weight of the bonded product. Other terms are used in different ways that can also cause confusion. The term adhesive can refer to either the adhesive as applied or the cured product. On the other hand, a resin is often used to refer to the uncured adhesive, although the ASTM defines a resin as “a solid, semisolid or pseudosolid organic material that has an indefinite and often high molecular weight, exhibits a tendency to flow when subject to stress, usually has a softening or melting range, and usually fractures conchoidally” (ASTM International 2000a). Thus, a crosslinked adhesive is not a resin, but the adhesive in the uncrosslinked state may be. Glue was “originally, a hard gelatin obtained from hides, tendons, cartilage, bones, etc. of animals,” but is now generally synonymous with the term adhesive. 9.4 APPLICATION OF THE ADHESIVE 9.4.1 ADHESIVE APPLICATION TO WOOD The first step in bond formation involves spreading the adhesive over the wood surface. The physical application of the adhesive can involve any one of a number of methods, including using spray, roller coating, doctor blade, curtain coater, and bead application technologies. After the adhesive application, a combination of some open and closed assembly times is used depending on the Wood Adhesion and Adhesives 221 specific bonding process. Both give the adhesive time to penetrate into the wood prior to bond formation, but the open assembly time will cause loss of solvent or water from the formulation. Long open times can cause the adhesive to dry out on the surface causing poor bonding because flow is needed for bonding to the substrate. In the bonding process, pressure is used to bring the surfaces closer together. In some cases, heat and moisture are used during the bonding process, both of which will make the adhesive more fluid and the wood more deformable (Green et al. 1999). For any type of bond to form, molecular-level contact is required. Thus, the adhesive has to flow over the bulk surface into the voids caused by the roughness that is present with almost all surfaces. Many factors control the wetting of the surface, including the relative surface energies of the adhesive and the substrate, viscosity of the adhesive, temperature of bonding, pressure on the bondline, etc. Wood is a more complex bonding surface than what is generally encountered in most adhesive applications. Wood is very anisotropic because the cells are greatly elongated in the longitudinal direction, and the growth out from the center of the tree makes the radial properties different from the tangential properties. Wood is further complicated by differences between heart- wood and sapwood, and between earlywood and latewood. Adding in tension wood, compression wood, and slope of grain increases the complexity of the wood adhesive interaction. The manner in which the surface is prepared also influences the wetting process. These factors are discussed in later sections of this chapter and in the literature (River et al. 1991), but for now we will assume that the adhesive is formulated and applied in such a manner that it properly wets the surface. 9.4.2 THEORIES OF ADHESION Adhesion refers to the interaction of the adhesive surface with the substrate surface. It must not be confused with bond strength. Certainly if there is little interaction of the adhesive with the adherent, these surfaces will detach when force is applied. However, bond strength is more complicated because factors such as stress concentration, energy dissipation, and weakness in surface layers often play a more important role than adhesion. Consequentially, the aspects of adhesion are a dominating factor in the bond formation process, but may not be the weak link in the bond breaking process. It is important to realize that, although some theories of adhesion emphasize mechanical aspects and others put more emphasis on chemical aspects, chemical structure and interactions determine the mechanical properties and the mechanical properties determine the force that is concentrated on individual chemical bonds. Thus, the chemical and mechanical aspects are linked and cannot be treated as completely distinct entities. In addition, some of the theories emphasize macroscopic effects while others are on the molecular level. The discussion of adhesion theories here is brief because they are well covered in the literature (Schultz and Nardin 2003, Pocius 2002), and in reality, most strong bonds are probably due to a combination of the ideas listed in each theory. In a mechanical interlock, the adhesive provides strength through reaching into the pores of the substrate (Packham 2003). An example of mechanical interlock is Velcro; the intertwining of the hooked spurs into the open fabric holds the pieces together. This type of attachment provides great resistance to the pieces sliding past one another, although the resistance to peel forces is only marginal. In its truest sense, a mechanical interlock does not involve the chemical interaction of the adhesive and the substrate. In reality, there are friction forces preventing detachment, indicating interaction of the surfaces. For adhesives to form interlocks, they have to wet the substrate well enough so that there are some chemical as well as mechanical forces in debonding. For a mechanical interlock to work, the tentacles of adhesive must be strong enough to be load bearing. The size of the mechanical interlock is not defined, although the ability to penetrate pores becomes more difficult and the strength becomes less when the pores are narrower. It should be noted that generally mechanical interlocks provide more resistance to shear forces than to normal forces. Also, many substrates do not have enough roughness to provide sufficient addition to bond strength from the mechanical interlock. Roughing of the substrate surface by abrasion, such as grit blasting or abrasion, normally overcomes this limitation. 222 Handbook of Wood Chemistry and Wood Composites If the concept of tentacles of adhesive penetrating into the substrate is transferred from the macro scale to the molecular level, the concept is referred to as the diffusion theory (Wool 2002). If there are also tentacles of substrate penetrating into the adhesive, the concept can be referred to as interdiffusion. This involves the intertwining of substrate and adhesive chains. The interface is strong since the forces are distributed over this intertwined polymer network (Berg 2002). However, the concept can also work if only the adhesive forms tentacles into the substrate. For this to occur, there has to be good compatibility of the adhesive and substrate. This degree of compatibility is not that common for most polymers. When it does occur a strong network is formed from a combination of chemical and mechanical forces. The other theories are mainly dependent upon chemical interactions rather than truly mechanical aspects. Thus, they take place at the molecular level, and require an intimate contact of the adhesive with the substrate. These chemical interactions will be discussed in order of increasing strength of the interaction (Kinlock 1967). The strengths of various types of bonds are given in Table 9.2, along with examples of some of the bond types in Figure 9.3. It is important to remember that the strength of interaction is for just a single interaction. To make a strong bond these interactions need to be large in number and evenly distributed across the interface. The weakest interaction is the London dispersion force (Wu 1982a). This force is the disper- sive force that exists between any set of molecules and compounds when they are close to each other. The dispersion force is the main means of association of non-polar molecules, such as polyethylene (Figure 9.3). Although this force is weak, where the adhesive and the adherend are in molecular contact, the force exists between all the atoms and can result in appreciable total strength. The ability of the gecko to walk on walls and ceilings has been attributed to this force (Autumn et al. 2002). The other types of forces are generally related to polar groups (Pocius 2002). The weakest are the dipole-dipole interactions. For polar bonds, there is a separation of charge between the atoms; this process creates a natural, permanent dipole. Two dipoles can interact if positive and negative ends of the dipole match up with the opposite ends of another dipole. The strength of this interaction TABLE 9.2 Table of Bond Strengths from Literature Bond Types and Typical Bond Energies Bond Energy Type (kJ·mol–1) Primary bonds Ionic 600–1100 Covalent 60–700 Metallic, coordination 110–350 Donor-acceptor bonds Brønsted acid-base interactions Up to 1000 (i.e., up to a primary ionic bond) Lewis acid-base interactions Up to 80 Secondary bonds Hydrogen bonds (excluding fluorines) 1–25 Van der Waals bonds Permanent dipole-dipole interactions 4–20 Dipole-induced dipole interactions Less than 2 Dispersion (London) forces 0.08–40 Source: Data from Fowkes 1983, Good 1966, Kinloch 1987, Pauling 1960, Wood Adhesion and Adhesives 223 H H H H δ_ O C C R R C C δ+ C H H H H R3 R4 a. b. δ_ H H H H O C C R δ+ C R C C R1 R2 H H H H H OH R1 H O O H O H O O O R2 H H H c. H d. H H H O O O N H2 H R R3 C R4 FIGURE 9.3 Examples of various types of bonds, including (a) dispersive bonds between two hydrocarbon chain, such as exist in polyethylene, (b) a dipole bond between two carbonyl group, such as in a polyester, (c) hydrogen bonds between a cellulosic segment and a phenol-formaldehyde polymer, (d) an ionic bond between an ammonium group and a carboxylate group. depends on proper alignment of the dipoles, which is not difficult for small molecules in solution, but can be very difficult between two chains because they have constrained translation and rotation (Wu 1982a). A variation of this concept is the dipole-induced dipole, but this interaction is usually weaker than the permanent dipole interaction and also suffers from the same alignment problem in polymers. Strongest of the secondary interactions is the hydrogen bond formation. This type of bond is common with polar compounds, including nitrogen, oxygen, and sulfur groups with attached hydrogens, and carbonyl groups. This type of bond involves sharing a hydrogen atom between two polar groups, and is extremely likely with wood and wood adhesives because both have an abundance of the proper polar groups. Almost all wood components are rich in hydroxyl groups and some contain carboxylic acid and ester groups. Both of these groups form very strong internal hydrogen bonds that give wood its strength, but are also available for external hydrogen bonds. All major wood adhesives have polar groups that can form internal and external hydrogen bonds. The bio-based adhesives depend heavily on hydrogen bonds for their adhesive and cohesive strength. Many synthetic adhesives are less dependent upon the hydrogen bond for their cohesive strength because they have internal crosslinks, but most certainly form hydrogen bonds to wood. One limitation of the hydrogen bond is its ability to be disrupted in the presence of water. Water and other hydrogen bonding groups can insert themselves between the two groups that are present in 224 Handbook of Wood Chemistry and Wood Composites the hydrogen bond. This process softens the inter-chain bonds so that they are less able to resist applied loads. Thus, a material that adsorbs and absorbs water, like wood, will lose some of its strength when it is wet. The same is true of the adhesion between the wood and the adhesive—it is certainly possible that hydrogen bonds weaken enough to serve as a failure zone. An interesting aspect of secondary bonds (dispersive, dipolar, and hydrogen bonds) is that after disruption, they can reform while fractured covalent bonds usually do not reform. The reformability of hydrogen bonds has been known about for a long time, but recent work has indicated that it can be an important part of wood’s ability to maintain strength even after there is some slippage of the bonds (Keckes et al. 2003), and this process has been referred to as “Velcro” mechanics (Kretschmann 2003). The role of this process in allowing the adhesives to adjust and maintain strength as the wood changes dimensions is not well understood, but could play a significant factor. Strong bonds can be formed from donor-acceptor interactions. The most common of these interactions with wood-adhesive bonds are the Brønsted acid-base interactions. Some acid-base interactions of cations with anions are possible in adhesion to substrates. Wood contains some carboxylic acids that can form salts with adhesives that contain basic groups, such as the amine groups in melamine-formaldehyde, protein, and amine-cured epoxy adhesives. Generally, with most materials, the strongest interaction is when a covalent bond forms between the adhesive and the substrate. However, for wood adhesion, this has been an area of great debate, because of the difficulty in determining the presence of this bond type given the complexity of both the adhesive and the wood and the difficulty of generating a good model system. Because wood has hydroxyl groups in its three main components—cellulose, hemicellulose, and lignin— and many of the adhesives can react with hydroxyl groups, it is logical to assume that these reactions might take place. However, others contend that the presence of large amounts of free water would disrupt this reaction (Pizzi 1994a). More sophisticated analytical methods will be needed to answer this issue (Frazier 2003). It is commonly assumed that the strongest interaction will control the adhesion to the substrate. This overlooks the fact that the adhesion is the sum of the strength of each interaction times the frequency of its occurrence. Thus, covalent bonds that occur only rarely may not be as important to bond strength as the more common hydrogen bonds or dipole-dipole interactions. Hydrogen bonds may be less significant under wet conditions than other bonds if the water disrupts these bonds. It is more important to think about forming stronger adhesion, not by a single type of bond, but by a large number of bonds of different types. Another point to consider is that the adhesive can adhere strongly to a surface and still not form a strong bond overall, due to failure within either the adhesive or the adherend interphases. One model of adhesion that is generally not related to the bond formation step, but is observed during bond breakage, is the electrostatic model. This model assumes that adhesion is due to the adhesive or the adherend being positive while the other is the opposite charge. It is unlikely that such charges generally exist prior to bond formation, and therefore cannot aid in adhesion; however, they can occur during the debonding process. Another model that has limited applicability to most cases of adhesion is deep diffusion, which involves polymers from the adhesive and adherend mixing to form a single, commingled phase. Although it is unlikely that the wood will dissolve in the adhesive, it is quite likely that many of the adhesive molecules will be absorbed into the wood cell walls. This diffusion can form one of several types of structures that more strongly lock the adhesive into the wood. This is one type of penetration, and it will be covered in section 9.4.8. In many cases, the strength of this penetration could be as strong as covalent linkages. Most of these adhesion models play not only a role in bond formation, but also aid the bonded assembly in resisting the debonding forces. The important part to remember is that, depending on the origin of the forces, the stresses can be either concentrated at the interface or dispersed throughout the bonded assembly. If the forces are dispersed, then the force felt at the interface may be quite small.