UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Photodegradation of adhesives used in wood composite materials Miesner, Martin 2008

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


ubc_2008_fall_miesner_martin.pdf [ 10.31MB ]
JSON: 1.0066630.json
JSON-LD: 1.0066630+ld.json
RDF/XML (Pretty): 1.0066630.xml
RDF/JSON: 1.0066630+rdf.json
Turtle: 1.0066630+rdf-turtle.txt
N-Triples: 1.0066630+rdf-ntriples.txt
Original Record: 1.0066630 +original-record.json
Full Text

Full Text

PHOTODEGRADATION OF ADHESIVES USEDIN WOOD COMPOSITE MATERIALSbyMartin MiesnerDipl. Ing. (FH), University of Applied Science Rosenheim, 2005A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THEREQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(FORESTRY)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)May, 2008© Martin Miesner, 2008AbstractThe weathering of wood composites is caused by a complex combination of chemicaland mechanical effects. Wood composites such as glulam beams are increasinglybeing used outdoors where their service life depends to some extent on the durability ofthe adhesive used in the composite. Increases in the durability of adhesives used insuch composite materials would prolong their service life and enable them to competemore effectively with other structural materials such as concrete and steel. This studyattempted to improve our understanding of the photodegradation of adhesives and therelationship between wood and adhesive photodegradation. The effectiveness of a UVlight absorber and hindered amine light stabilizer (UVA and HALS) at protectingadhesives from photodegradation was also investigated. First, the effect of adhesivetype (melamine formaldehyde, epoxide, and emulsion polymer isocyanate), stabilizerand adhesive stabilizer interaction on tensile strength, weight loss and discoloration ofadhesive dog-bone samples exposed in two different weatherometer devices (QUV andXenon-arc) was examined. Structural and chemical changes of the adhesive specimenswere examined using Fourier Transform Infrared (FTIR) spectroscopy and ScanningElectron Microscopy (SEM). Secondly, the effects of adhesive type (melamineformaldehyde, epoxide), stabilizer and adhesive stabilizer interaction on surfaceroughness and discoloration of wood-adhesive-dowel samples exposed to solarradiation was examined. Profileometry and SEM was used to examine the surface ofdowels in the region where they were exposed to both wood and sunlight. An epoxideadhesive (butyl glycidyl ether of bisphenol-A with polyamide) used in the aircraftindustry showed outstanding resistance to weathering. The other adhesives were not asresistance to weathering, but the addition of a UVA/HALS photostabilizer to theadhesives generally increased their photostability (particularly color changes of theepoxy adhesives and weight loss of the MF adhesive). Greater degradation of adhesivesamples occurred when they were exposed in a QUV weatherometer than in a Xenon-arc weatherometer. The synergistic effect of moisture and UV radiation on thedegradation of adhesives may account for this observation. Adhesive dowels embeddedin wood did not show greater degradation (erosion) in the region where they wereexposed to both wood and sunlight. Therefore the hypotheses that woodphotosensitizes adhesives could not be supported by experimental findings. Furtherrefinement of the experimental methodology developed in this thesis would be desirableto retest this hypothesis. All of the four adhesives that were tested possessed someinteresting characteristics that might make them suitable for use in glulam exposedoutdoors, but out of the four the two epoxy adhesives appeared to have the greatestpotential.iiTable of ContentsAbstract ^ iiTable of Contents ^  iiiList of Figures viiAcknowledgements^ x1. General Introduction  12. Literature Review ^ 52.1. Introduction  52.2. Adhesives/Polymers - Adhesion ^  52.3. Types of Adhesives ^  82.3.1. Epoxy Adhesive  92.3.2. Emulsion Polymer Isocyanate (EPI) Adhesive ^  132.3.3. Melamine Adhesive ^  132.4. Degradation of Polymers/Adhesives ^  162.5. Weathering of Adhesives ^  172.5.1. Solar Radiation/Heat Photodegradation of Adhesives ^ Factors Responsible for Photodegradation of Adhesives ^ Free Radical Generation ^  212.5.2. Water/Humidity ^  232.5.3. Oxygen ^ 242.5.4. Pollutants 252.6. Weathering of Wood ^ 262.6.1. Photodegradation of Wood ^ 272.7. Photo-stabilization ^ 292.7.1. UV Absorbers (UVA) 302.7.1.1. Hydroxyphenyl-benzotriazole ^  312.7.1.2. Hydroxyphenyl-s-triazine  322.7.2. Hindered Amine Light Stabilizers (HALS) ^ 322.7.2.1. 2,2,6,6-Tetra-methyl-piperidine  332.7.3 Antioxidants ^  332.8. Weathering of Wood Polymer Composites ^  34iii2.8.1. Wood Plastic Composites (WPC) ^ 342.8.2. Glulam Beams ^ 352.8.3. Photocatalytic Effect of Wood on Polymer Degradation ^ 362.9. Conclusion ^  373. Weathering and Photostability of Adhesives ^ 383.1. Introduction ^  383.2. Materials and Methods ^ 393.2.1. Experimental Design and Statistical Analyses ^ 393.2.2. Sample Preparation ^  403.2.3. Colour Measurement 433.2.4. Artificial Weathering ^  433.2.5. Tensile Strength Testing  463.3 Results ^  473.3.1 Tensile Strength ^ 483.3.2. Weight Loss 493.3.3 Color Change ^ 523.3.4 Structural Changes (SEM) ^ 623.3.5 Chemical Changes (FTIR)  643.4 Discussion ^ 693.5 Conclusions  754. Relationship between Wood and Adhesive Photodegradation ^ 764.1. Introduction ^  764.2. Materials and Methods ^ 774.2.1. Experimental Design and Statistical Analyses ^ 774.2.2. Sample Preparation ^  774.2.3. Natural Weathering  804.2.4. Roughness Measurement^ 814.2.5. Colour Measurement 814.3. Results ^  824.3.1 Roughness Changes ^ 824.3.2 Color Changes 834.3.3 Morphological Changes ^ 884.4 Discussion ^  90iv4.5 Conclusions^ 925. General Discussion and Conclusions ^  935.1. Discussion and Suggestions for Further Research ^ 935.2. Conclusions ^ 966. References 97Appendices ^ 106vList of TablesTable 3.1: Adhesives assessed for their resistance to weathering ^ 40Table 3.2: Significant effects of and interactions between adhesive type and stabilizeron resistance of adhesive films to accelerated weathering in a Xenon-arcdevice 47Table 3.3: Significant effects of and interactions between adhesive type and stabilizeron resistance of adhesive films to accelerated weathering in a QUV device .. 47Table 3.4: Percentage weight and tensile strength losses of glue film specimensexposed for 200h in a Xenon-arc or QUV weatherometer ^ 51Table 4.1: Dates when adhesive dowels were exposed to sunlight, and total sunshinehours and mean temperature during the exposure trial ^ 80Table 4.2: Significant effects of and interactions between adhesive type and stabilizeron resistance of adhesive dowels to solar radiation ^  82Table 4.3: Lightness/color values of fully exposed unstabilized and stabilized MF andEpU samples ^ 85viList of FiguresFigure 1.1: The Expo Dach built for the World exhibition 2000 in Hannover/Germany ....(Janberg, n.d.) ^  3Figure 2.1: Mechanical interlocking between adhesive and substrate (SpecialChem,n.d.) ^  7Figure 2.2: Chemical structure of epoxide resin derived from DPP and ECH ^ 9Figure 2.3: External cross-linking of epoxide resin with amine hardener (primary aminegroup) ^  10Figure 2.4: FTIR spectra of epoxy polymer before and after a "standard" curingschedule; a, freshly mixed adhesive; b, after being cured at 160 °C for 2.5h(Hon 1994) ^  11Figure 2.5: Reaction scheme of melamine and formaldehyde ^  14Figure 2.6: Reaction of methylol melamine by condensation to form intermediates ^ 14Figure 2.7: Final curing process through reaction of amino and methyl groups ^ 15Figure 2.8: Autooxidation scheme for polymer degradation showing the cyclic nature ofthe process (Horsey 1994) ^ 22Figure 2.9: Pathways of lignin photo-oxidation ^ 28Figure 2.10: Chemical structure of hydroxyphenyl-benzotriazole ^ 31Figure 2.11: Internal hydrogen bond formation and tautomerization in a hydroxyphenyl-benzotriazole^  32Figure 2.12: Chemical structure of hydroxyphenyl-s-triazine ^  32Figure 2.13: Formation of hindered piperidinoxy radicals from hindered piperidine ^ 33Figure 3.1: Chemical structures of (1) HALS (Tinuvin 292) and (2) UVA (Tinuvin 384) 41Figure 3.2: Preparation dog-bone adhesive samples; 1, mixing adhesive formulation; 2,spreading out adhesive with a paint gauge; 3, removing adhesive film aftercure with a razor blade; 4, samples punched out with a dog-bone shapedcutter; 5, adhesive dog-bone specimen ^ 42Figure 3.3: QUV weatherometer and dog-bone specimens in sample holder ^ 45Figure 3.4: Xenon-arc weatherometer and dog-bone specimens in sample holder ^ 45Figure 3.5: QTS 3 Quick test tensile strength tester from Pruefpartner GmbH ^ 46viiFigure 3.6: Tensile strength of glue-film specimens (X) after exposure in Xenon-arcweatherometer for 200h (results averaged across stabilized specimens andunstabilized controls). Unexposed controls (c) are shown for comparison . 48Figure 3.7: Tensile strength of glue-film specimen (Q) after   exposure in QUVweatherometer for 200h (results averaged across stabilized specimens andunstabilized controls). Unexposed controls (c) are shown for comparison . 49Figure 3.8: Weight losses of glue film specimens after exposure in a Xenon-arcweatherometer for 200h (results averaged across stabilized and unstabilizedcontrols)  50Figure 3.9: Weight losses of glue film specimens after exposure in a QUVweatherometer for 200h (results averaged across stabilized and unstabilizedcontrols)  50Figure 3.10: Interaction of adhesive type and stabilizer on weight losses of glue filmspecimens exposed in a QUV device for 200h ^ 51Figure 3.11: Effect of adhesive type on L* ratio, a* ratio, b* ratio and deltaE afterexposure to QUV ^ 53Figure 3.12: Effect of stabilizer on b* ratio and deltaE of glue films after exposure toQUV (averaged across glue types) ^ 54Figure 3.13: Effect of Adhesive/stabilizer on L* ratio, a* ratio, b* ratio and deltaE afterexposure to QUV ^ 56Figure 3.14: Effect of adhesive type on L* ratio, b* ratio and deltaE of adhesive filmsafter exposure in a Xenon-arc weatherometer ^ 58Figure 3.15: Effect of stabilizer on L* ratio, b* ratio and deltaE of adhesive films afterexposure in a Xenon-arc weatherometer ^ 59Figure 3.16: Interaction of adhesive type and stabilizer on L* ratio, a* ratio, b* ratio anddeltaE of adhesive films after exposure in a Xenon-arc weatherometer ^ 61Figure 3.17: Unstabilized MF before (left) and after QUV exposure (right) ^ 62Figure 3.18: Unstabilized EpE before (left) and after QUV exposure (right) ^ 62Figure 3.19: Unstabilized EPI before (left) and after QUV exposure (right) ^ 63Figure 3.20: Unstabilized EpU before (left) and after QUV exposure (right) ^ 63Figure 3.21: Unstabilized MF after QUV (left) and Xenon-arc exposure (right) ^ 63Figure 3.22: Unstabilized (left) and stabilized (right) EpU after QUV exposure ^64viiiFigure 3.23: Unstabilized MF before (bottom) and after QUV exposure (top) ^ 65Figure 3.24: Unstabilized EpE before (bottom) and after QUV exposure (top) ^ 66Figure 3.25: Unstabilized EPI before (bottom) and after QUV exposure (top) ^ 66Figure 3.26: Unstabilized EpU before (bottom) and after QUV exposure (top) ^ 67Figure 3.27: Unstabilized (bottom) and stabilized EpU after exposure to QUV (top) ^ 68Figure 3.28: Unstabilized MF after exposure to QUV (top) and Xenon-arc (bottom) ^ 69Figure 3.29: Evolution of acetic acid from vinyl acetate (Copuroglu and Sen 2004) ^ 72Figure 4.1: Sample board showing a dowel and interfaces between the dowel, woodand the atmosphere ^ 79Figure 4.2: Effect of stabilizer on surface roughness ratio of adhesive dowels afterexposure to sunlight ^  83Figure 4.3: Effect of adhesive type on b* ratio and a* ratio after full exposure of dowelsto solar radiation ^  84Figure 4.4: Effect of adhesive type on a* ratio after exposure to solar radiation (region3)^ 85Figure 4.5: Effect of adhesive/stabilizer on L* ratio, a* ratio and deltaE after exposure tosolar radiation ^  86Figure 4.6: Unstabilized EpU before (bottom) and after (top) exposure to solar radiation^  87Figure 4.7: Unstabilized MF before (bottom) and after (top) exposure to solar radiation..^  87Figure 4.8: Unstabilized MF before (top) and after exposure (bottom) ^ 88Figure 4.9: Stabilized EpU before (top) and after exposure (bottom) 89Figure 4.10: Unstabilized MF after exposure to sunlight. Complete dowel (left), interfaceregion (right) ^  89Figure 4.11: Unstabilized EpU after exposure to sunlight. Complete dowel (left),interface region (right) ^  90ixAcknowledgementsThe research reported in this thesis was made possible through the funding of theWood Based Composite Centre (WBC).I would like to thank the following individuals and organizations: Dr. Philip Evans for the guidance and support throughout my studies Dr. Mohammed Jahangir Chowdhury and Dr. Hans Krause for their support andguidance throughout my studies The graduate students and research assistants in Phil's research group United Resin Corporation, Ellsworth Corporation, Henkel and Dynea for providingadhesive samples My family and friendsx1. General IntroductionAccording to Amstock (2001) the term glue refers to a sticky material, whereasadhesives don't have to be sticky. Adhesive is more inclusive, but in the past these twoterms have been considered synonymous (Amstock 2001). For example, organicadhesives, coatings or cellulose are based on polymers and their organic chemistryshares some similarities (Stevens 1999). Therefore, they are sometimes treated asequivalent in this work.As Keimel (1994) mentioned, the history of adhesives is closely related to thehistory of humankind. The first commercial glue plant was founded in 1690 in Holland.At that time the majority of glue applications were for the manufacturer of furniture(Amstock 2001). Today the fields of applications for adhesives are manifold and theysurround us in nature and in our daily lives. Applications abound from office "post-itnotes" to automotive safety glass to footwear to aerospace structures to "no-lick"postage stamps to bonding wood. Many products that we take for granted could neverexist if it were not for adhesive bonding (Petrie 2007). According to market researchconducted by Freedonia, the World market value for adhesives was estimatedapproximately to 25 billion U.S. dollars in 2004 (Phanomen Farbe 2001). The largestregional markets are North America, Western Europe, and the Far East. These regionsaccount for 85 % of the global demand by volume and 78 % of the value of adhesives(Petrie 2007). Furthermore, the world market for adhesives (and sealants) is growing by3 % annually (Petrie 2007).Major advancements in adhesive technology were made in the twentieth century(Amstock 2001). The development of modern and structural adhesives was closely tiedto the development of aircraft and aerospace industries (Keimel 1994, Dowling 1945).High performance adhesives are particularly useful for this industry, because they saveweight and are better in distributing stresses than mechanical attachment techniquessuch as bolting and riveting (Schultz and Nardin 1994). Another large-scale applicationof adhesives is in the construction industry for materials such as engineered woodenpanels or glue laminated beams (glulam) (Amstock 2001, Phdnomen Farbe 1999).Glued laminated wood construction had its beginning at the turn to the twentiethcentury as a result of the activities of Otto Karl Hetzer in 1909, who obtained his firstpatent for glulam construction (Rhude 1996). Today large dimension engineered wood1composites, such as glulam beams, act as substitutes for steel and concrete for theconstruction of industrial and agricultural buildings and bridges (Mueller 2000). Petersenand Solberg (2002) reported that the manufacture and use of these materials consumesfar less energy and thus generates lower levels of carbon dioxide than comparable steelproducts. Glulam also looks nicer than steel beams and it can be used to createdistinctive buildings that people enjoy working in (Mueller 2000). Steel and concrete,however, are less susceptible to decay, fire and weathering than engineered woodcomposites.A topic that is comprehensively discussed in the literature, which is of relevanceto the performance of glulam exposed outdoors, is the complex effect of environmentalfactors (water, oxygen, heat and radiation) on adhesives and polymers. Such a topic isimportant because of the negative impact of environmental factors on the long termdurability of adhesive bonds (Cotter 1977, Comyn 1998, Kerr, MacDonald, and Orman1970, Althof 1981, Jellinek 1983, Kockott 1988, Dorn and Breuel 1992, White andTurnbull 1992, Stevens 1999, Akmal and Usmani 2000, Searle 2000, Keene et al.2001). Also the price dependency of synthetic adhesives on their raw-materialpetroleum (Phanomen Farbe 2001); calls for a better understanding of the degradationmechanism of adhesives, leading to improvements in the protection of adhesives ofadhesives and longer service lives for composites.UV light is the driving force behind the degradation that occurs when polymersare exposed to the elements (Keene et al. 2001 and Searle 2000). According to Torikai(2000), polymeric materials contain functional groups (chromophores), which arecapable of absorbing UV light and leading to chemical changes in the parent polymer.The two critical modes of photodamage applicable to most natural and syntheticmaterials exposed to solar radiation are yellowing discoloration and loss in mechanicalintegrity (Andrady et al., n.d.). More solar radiation and solar radiation of shorter-wavelengths now reaches the earth surface, as a result of partial ozone depletion in theatmosphere, which further decreases the lifetimes of polymeric materials such asadhesives (Torikai 2000). Different approaches have been taken to protect adhesivesagainst UV radiation. Light stabilizers may be incorporated into the adhesive matrix inorder to protect it against photo-induced degradation (Horsey 1994). Other additivessuch as antioxidants are also used to protect adhesives from oxidation (Horsey 1994,Comyn 1998). The patent literature on additives such as stabilizers and antioxidants is2voluminous (Rabek 1987). Glulam beams exposed outdoors may be coated withpigmented finishes, which provides some protection from photooxidative degradation,but such a coating also destroys the visual appeal of the glulam. The "Expo Dach" builtfor the World exhibition 2000 in Hannover/Germany (Figure 1.1), is constructed mostlyfrom glulam (Bogusch and Seidel 2000).Figure 1.1 has been removed due to copyright restrictionsFigure 1.1: The Expo Dach built for the World exhibition 2000 in Hannover/Germany(Janberg, n.d.)The designers of this structure recognized the need to protect the glulam in theroof from weathering and to do this they stretched a thin translucent plastic over it(Burger 2000). This covering, which can be seen in the small picture in Figure 1.1, hashelped to improve the service life of the glulam. In many situations, however, it is notpossible to protect glulam by design and construction, and there may be a need todirectly enhance the composite by increasing the photostability of the wood andadhesive.However, there has not been much research on the photostability of adhesivesused in glulam. Furthermore, as pointed out by Jordan (1989) it is important to examinenot only the adhesive, but also the composite as a whole including adhesive andsubstrate (wood). In addition, it is mentioned in the literature, that individualobservations of each material combination in a composite exposed to specificconditions is required, in order to evaluate its performance under these conditions (Dorn3and Breuel 1992). Therefore, this study was designed to examine the performance ofselected adhesives under different artificial weather conditions and also to evaluate theinteraction between wood and adhesive photodegradation, as it occurs e.g. in glulambeams exposed outdoors.This study aimed to investigate the performance of adhesives exposed toartificial weather conditions including ultraviolet light. The ability of selective additives tostabilize adhesives and the influence of the wood matrix on adhesive photodegradationwere also examined. Following this introduction, Chapter 2 reviews relevant literatureon the effect of weathering factors (water, oxygen, radiation, pollution) on polymericmaterials such as wood, adhesive and wood-adhesive composites and ways ofstabilizing them. Experimentation in Chapter 3 examined the photostability of fourdifferent adhesives that could be used in glulam exposed outdoors. This research alsoexamined whether a photo-stabilizer could protect the adhesive from photodegradation.The aim was to determine which adhesive type was most resistant to photodegradationand how effective the selected stabilizer mixture was at restricting photodegradation.It is well known that metal compounds can accelerate the photooxidativedegradation of polymers such as polyethylene and polypropylene (Shlyapnikov,Kiryushkin, and Martin 1996). Previous studies have found that the photodegradation ofsuch polymers when used in wood plastic composites is accelerated by the presence ofwood (Matuana, Kamdem, and Zhang 2001, Kiguchi et al. 2006). Therefore, it isreasonable to hypothese that the photodegradation of adhesives in glulam could becatalyzed by wood. Therefore, Chapter 4 examined the effect of solar radiation on thephotostability of adhesive in contact with wood. A better understanding of thephotostability of adhesives used in glulam and the factors that influence theirphotodegradation could help to prolong the service life of glulam exposed outdoors.More importantly, an extended service life for such products will help wood to becomemore cost effective and compete better with structural materials such as concrete andsteel that cannot be produced sustainably.42. Literature Review2.1. IntroductionAdhesives are widely used today for a variety of different purposes. Woodcomposites such as glulam beams or wood plastic composites for example combinewood and adhesives (polymer) in one composite material. Organic materials, includingmost adhesive types and wood, however, are prone to degradation processesespecially under outdoor conditions, and thus the lifetime of such materials is limited.Degradation caused by the environment comprises different factors; so the collectiveterm weathering is used to describe it. Photodegradation refers to the degradation of amaterial caused by photons from highly energetic UV light. Better knowledge of thedegradation of polymers, especially in composite materials, is not only of academicinterest, but it is also of great practical importance. Rabek (1987) suggested that suchknowledge could be used to improve stabilization through the development of morestable polymer structures, or by the introduction of more effective stabilizers.Alternatively, it could be used to help develop more "environmentally friendly" polymerswhich are designed to rapidly disintegrate outdoors when exposed to daylight, but areable to maintain their desirable mechanical properties during use under indoorillumination (Akmal and Usmani 2000).2.2. Adhesives/Polymers - AdhesionAccording to Schultz and Nardin (1994) the construction industry is the main fieldof application of adhesive bonding. The role of an adhesive is to bond two pieces in afixed position. To achieve this goal the molecular attraction between the substrate andthe adhesive (adhesion forces) and the molecular attraction by the adhesive throughoutits mass (cohesion forces) have to be balanced (Sellers 1994). Adhesives havereplaced in part the more classical mechanical bonding techniques of bolting or riveting.Adhesives are competitive primarily because they allow users to save weight, providebetter stress distribution, and they offer better aesthetics since the glue-line can bepractically invisible. Schultz and Nardin (1994) mentioned that adhesion became aserious area in scientific circles only about 50 years ago.5To perform satisfactorily, adhesives must be capable of both, wetting andpenetrating the wood substrate and being converted into a rigid solid through eitherevaporation of the solvent, physical cooling in the case of a thermoplastic-typeadhesive, or a chemical reaction that may require a catalyst and/or heat. Furthermore,the adhesives also need to polymerize to produce a stiff solid with adequate mechanicalproperties. As adhesives increase to a molecular weight of 500 g/mol or above usefulcohesive strength is developed (Sellers 1994).Schultz and Nardin (1994) pointed out that one of the main difficulties in thestudy of adhesion mechanisms lies in the fact that the subject is at the boundary ofseveral scientific fields. This diversity is apparent in the number of theoretical models ofadhesion that have been proposed (Schultz and Nardin 1994). The mechanism bywhich materials adhere together was classified by Filbey and Wightman (1988) into fourcategories: 1, electrostatic; 2, diffusion; 3, mechanical interlocking; and 4, adsorption.There are some discrepancies in the literature, however, regarding the classification ofadhesion mechanisms. Petrie (2007) suggested that there are several theoriesattempting to describe adhesion. To the most common theories (mentioned above) headded the weak-boundary-layer theory, which is not described here.Electrostatic adhesion results from the redistribution of charges that occurs whendissimilar materials are brought close together (Filbey and Wightman 1988). Theelectrostatic theory was proposed by Deryaguin and co-workers in 1948 and theysuggested that electrostatic forces can have a significant contribution to adhesivestrength (Schultz and Nardin 1994).The diffusion theory states that polymer/polymer adhesion occurs due tointerdiffusion of the macromolecular chains at the interface (Filbey and Wightman1988). This theory, which was proposed by Voyutskii (1963), implies that themacromolecular chain or chain segments are sufficiently mobile and mutually soluble.The joint strengths mainly depend on factors such as contact time, temperature, natureand molecular weight of polymers, and pressure (Schultz and Nardin 1994).The mechanical interlock theory implies that the physical structure of theadhesive interface influences adhesion. The mechanical interlock theory is shown inFigure 2.1.6Figure 2.1 has been removed due to copyright restrictionsFigure 2.1: Mechanical interlocking between adhesive and substrate (SpecialChem,n.d.)The adherent is structured so that adhesives can flow into crevices or pores, but cannot easily be pulled out (Filbey and Wightman 1988). Some studies have suggestedthat the effective submicroscopic penetration of adhesives in cellulose in wood occursat a molecular weight of 1000 g/mol or less (Sellers 1994). In most cases adhesion canbe enhanced by an increase in interfacial area resulting from increased surfaceroughness, (because wetting conditions permit penetration of the adhesive into poresand cavities). An increase in the effective surface area of the adherent also increasesthe number of primary and secondary bonds available to the adhesive. According toDavis and Shaffer (1994), surfaces with features on the order of tens of nanometersexhibit superior performance to those with features on the order of micrometers.Several factors contribute to this difference in performance. The larger-scale featuresare fewer in number and are generally smoother (even on a relative scale), so thatinterlocking is less effective. Larger-scale features also allow trapped air and surfacecontaminants to remain at the bottom of the troughs and pores. These unbondedregions limit joint performance by reducing both chemical and physical bonds andserving as stress concentrators (Davis and Shaffer 1994). The interlock theory was firstproposed by MacBain and Hopkins (1925). Several years later experiments conductedby Borroff and Wake (1949) clearly showed that penetration of the adhesive into theadherent was the most important parameter for tested adhesive joints (Schultz andNardin 1994). Borroff and Wake (1949), however, concluded that the mechanicalinterlock theory cannot be considered as universal, because good adhesion can occurbetween smooth surfaces.The adsorption theory of adhesion, generally attributed to Sharpe andSchonhorn, is, according to Schultz and Nardin (1994), the most widely used approachin adhesion science at present. This theory postulates that primary and secondary7bonds formed at interfaces result in good adhesion. The secondary bonds are Van derWaals forces which include London dispersion forces, dipole-dipole, acid-base, andhydrogen bonding. The primary bonds are covalent or ionic bonds which form betweenthe adherent and the primer and/or adhesive (Filbey and Wightman 1988). The typicalstrength of a covalent bond, for example, is of the order of 100 to 1000 kJ/mol.,whereas those resulting from Van der Waals interactions and hydrogen bonds do notexceed 50kJ/mol. The formation of chemical bonds depends on the reactivity of bothadhesive and substrate. Generally, the formation of an assembly goes through a liquid-solid contact step, and therefore the criteria for good adhesion become essentiallythose for criteria of good wetting, although this is a necessary, but not sufficientcondition (Schultz and Nardin 1994).Schultz and Nardin (1994) concluded that adhesion is a very complex fieldbeyond the reach of any single model or theory. In practice, several adhesionmechanisms can be involved simultaneously. However, it is generally assumed that theadsorption theory defines the main mechanism exhibiting the widest applicability. Itdescribes the achievement of intimate contact and the development of physical forcesat the interface. This is a necessary step for interlocking, inter-diffusion, and chemicalbonding mechanisms to occur subsequently, further increasing adhesive strength(Schultz and Nardin 1994).2.3. Types of AdhesivesThe process of bonding wood materials with glue has been practiced since theearly periods of civilization. Examples of glued wood materials exist which arethousands of years old, for example scarf-jointed wooden chariot wheels anddecorative, thin-veneer-covered artifacts from the Pharaonic periods of Egypt (Sellers1994). The adhesives examined in this study can be used for bonding wood, but theycan be used for other purposes as well. The development of adhesives has shiftedaway from stiff, brittle adhesives towards ductile products. According to Jordan (1989)this shift has occurred mainly due to the susceptibility of the former adhesives to fatiguecaused by dynamic load. The development of ductile adhesives is particularly beneficialfor adhesives used in wood composites such as glulam, because it allows the wood tomove in its dimension without putting too much stress on the gluline.8OH Condensation 1CH2 — CH — CH2 — CI + 1-10/\OHO- CH2 - CH - CH2/0\O—CH2— CH—CH2ACH2—CH—CH2—O 01-1 +^HCI Polymerization •n/0\CH2 — CH— CH2 02.3.1. Epoxy AdhesiveIn the nineteen thirties Schlack of I.G. Farben (today BASF) patented polyglycidylethers of polyphenols, including diphenylpropane (DPP). As mentioned by Potter(1970), 95% of epoxide resins were once diglycidyl ethers of DPP. Today most of theepoxide resins are made by the interaction of epichlorohydrin (ECH) anddiphenylpropane (DPP) also called bisphenol-A (Kollek 1988, Goulding 1994) (Figure2.2).Figure 2.2: Chemical structure of epoxide resin derived from DPP and ECH9Goulding (1994) mentioned that epoxy resin derived from DPP and ECH iscapable of reacting with various products, or itself, to form a solid, infusible product ofconsiderable strength. The reaction with itself, as a result of homopolymerization, istypical of catalyzed cross-linking. Under the external cross-linking process, with thepresence of hardeners, the oxirane group reacts with active hydrogen available e.g. inamines, and amides (Goulding 1994). The external cross-linking is shown below inFigure 2.3.lo\RN H2 + H2C—CH—C."ArtH2 H/RNC—CH—Cavv%H2^I^H2OHThis product can react with an additional epoxide group to continue the cross-linkingprocessOHI/ H^/ /CH2—CH—CH2'fvy0RNN^/ \^RN\CH2—CH—CH2uvv + H2C—CH—CH2uvv^CH2—CH—CH2'wI^ IOH OHFigure 2.3: External cross-linking of epoxide resin with amine hardener (primary aminegroup)The wide and increasing use of epoxy resins in industry, despite the relativelyhigh cost of manufacturing them, is according to Paul and his colleges (1979) due totheir unique combination of properties. Furthermore, Paul et al. (1979) mentioned thattheir properties such as good molding characteristics are due to the absence of volatileby-products and low shrinkage during cure, and the consequent minimization of internalstresses. These attributes give epoxy resins high mechanical properties. Paul et al.(1979) pointed out that the cured resins have high adhesive strengths due mainly to thegeneration of polar hydroxyl groups during cure, and they possess high electricalinsulation and good chemical resistance. These properties of epoxy adhesives were10also reported by Petrie (2007). As suggested by Paul et al. (1979) many of theseproperties can be modified by blending resin types, selecting curing agents, or usingfillers that confer a versatility on epoxy resins that is not found in other thermosettingsystems. According to Davis and Shaffer (1994) epoxy adhesives are also used due totheir greater toughness, and lower temperatures and pressures required during cure.Water thinnable epoxy systems are potentially capable of being used as woodadhesives and they form strong durable joints. The very slow attainment of maximumjoint strength, however, may be a limiting factor for certain applications (Mynott and Vander Straeten 1984).The extent of cross-linking of an adhesive describes the degree of cure. Thenature of the resin molecules between cross-links and the nature of the curing agentdetermine the properties of the cured material (Potter 1970, Kollek 1988). To obtain themost favorable properties in any cured resin it is important to achieve maximum cross-linking. Infra-red spectra taken at various times during the curing process of epoxyadhesives have shown a loss of epoxy and an increase in hydroxyl concentration(Figure 2.4).Figure 2.4 has been removed due to copyright restrictionsFigure 2.4: FTIR spectra of epoxy polymer before and after a "standard" curingschedule; a, freshly mixed adhesive; b, after being cured at 160 C for 2.5h (Hon 1994)11The extent of hydrogen bonding in cured epoxide networks has beeninvestigated by Harrord (1963). He concluded that in the temperature range of 30-200 Cthe hydroxyl group is extensively involved in hydrogen bond formation. Paul et al.(1979) stated that 'the cured resin has high adhesive strengths due mainly to thegeneration of polar hydroxyl groups during cure'.Amstock (2001) pointed out that epoxies are thermosetting and can be definedas any molecule containing more than one a-epoxy group capable of being converted toa thermoset form. The epoxide group is a three-membered oxide ring. The resincompounds contain on average more than one epoxide group per molecule, and theyare polymerized through these epoxide groups, using a cross-linking agent (also calleda curing agent or hardener). Epoxy resin is mostly used in the cured state when allepoxide have reacted; in the uncured non-cross-linked state they are of limited utility.With increasing degree of polymerization the resins become solids of increasing meltingpoints (Potter 1970, Amstock 2001). The characteristic physical, chemical, andelectrical properties of the cured resin all stem from the basic molecular structure of thepolymerized resin. Important factors at the molecular level which determine theseproperties are: 1, the extent of cross-linking, i.e., the degree of cure; 2, the cross-linkdensity; 3, the nature of the resin molecule between cross-links; and 4, the nature of thecuring agent molecule (Potter 1970). Comyn et al. (1979) suggested that during curingthe epoxide ring may get opened to produce either alcohol or ether units. These groupsprobably contribute to both the adhesion and water absorption associated withepoxides. Both of these chemical groups are hydrophilic. According to Althof (1981)epoxy nylon adhesive bonded joints suffer a decrease of strength when they areexposed to a wet, hot, climate (Althof 1981).According to Paul et al. (1979) the versatility of epoxy adhesives can beimproved by blending them with different resin types, selecting curing agents, or usingfillers. Such versatility is found in other thermosetting systems.122.3.2. Emulsion Polymer Isocyanate (EPI) AdhesivePoly vinyl acetate (PVAC) was used as a solvent-based adhesive in the 1930s,and later as a hot melt, but it did not become commercially important until its use in the1940s as an emulsion adhesive for paper and wood. Today, in emulsion form as whiteglue, it is the world's most widely used thermoplastic adhesive (Keimel 1994).According to Sellers (1994) PVAC adhesive impart relatively high bond strength, curesquickly and forms colorless bond-lines. They are free from the environmentallyquestionable compound formaldehyde. PVAC is compatible with plasticizers, wettingagents and other additives, which allow different formulations to be created.Increasingly, copolymer forms are offered commercially to decrease stress-relatedcreep and to enhance exterior durability (Sellers 1994).At the end of the last century EPI adhesives were developed based on PVACdispersions containing isocyanate hardener. These were able to form some water-resistant bonds, unlike PVAC (Krystofiak, Proszyk, and Mariusz 2003, Dynea, n.d.).Sellers (1994) mentioned that polymeric diisocyanate (PMDI) can continue to cross-linkby reaction with itself in the presence of absorbed moisture and that PMDI is used as acatalyst in vinyl emulsions. Krystofiak, Proszyk, and Mariusz (2003) found that the gluelines formed from EPI adhesives have high thermal and water resistance, which is, tosome extent, influenced by type and amount of isocyanate hardener. EPI can beclassified as a thermoplastic adhesive with properties very close to other thermosettingadhesives. EPI adhesives can be formulated to alter their water resistance, curingspeed, strength and viscosity to suit different substrates. EPI adhesives are cold-settingwith very good adhesion to difficult-to-bond hardwoods and they are also fast curingeven at high wood moisture content (Dynea, n.d.). EPI adhesives have not been usedextensively for glulam beams.2.3.3. Melamine AdhesiveMelamine formaldehyde resins were commercially used for the first time in the1930's (Chugg 1964). Melamine reacts with formaldehyde to form methylol melamine(hexamethylolmelamine), as shown in Figure 2.5. Two to six methylol groups can beheld by a melamine molecule. Methylol melamine reacts by condensation to formhydrophobic intermediates, with splitting of water and formaldehyde (Pizzi 1994, Chugg131964), see Figure 2.6. Garnier et al. (2001) demonstrated that methylene ether bridges(-CH2OCH2-) occur, which rearrange themselves with relative ease to form methylene (-CH 2-) bridges with liberation of formaldehyde. The final curing process transforms theseintermediates into an insoluble cross-linked MF resin through the reactions of aminoand methylol groups (Pizzi 1994, Chugg 1964, Figure 2.7). According to Pizzi (1994)fully cured melamine resins possess high hardness and abrasion resistance. They arenot flammable and are resistant to chemicals./CH2OHNH2^ HNN^........' N 0Addit ion N^NHN/N,\ NHCH2OH^CH2OHH2N/ N\ NH2+ 3Figure 2.5: Reaction scheme of melamine and formaldehyde+CondensationHeat- H20-CH2O/CH2OHHNNN/,\H NI^NHICH2OH^CH2OH/CH2OHHN\ N,N tijHHNI CH2OHCH2OH/CH2OH/CH2OHHN^ HNN^ NI^NHt /"NH..,,... ,....., HN/"ilCH2OH^H2NH1CH2OHFigure 2.6: Reaction of methylol melamine by condensation to form intermediates14/CH2OHHN/CH2OHHNN-------''''..'..*'. '''NN/N .\HNCH2HOH2CN------.1-*'''` NIHN/'N\NH CH2N^NHN/,\NHICH2OHnPolymerizationHeat- H2O/CH2OHHNHNNOSS\N "NH^N'-' A'.':NZ\N/N NHN".--.......k........-`'N^H^HN)I\N/,\NH N '`.. NHN/^N/I'll'H^HFigure 2.7: Final curing process through reaction of amino and methyl groupsPizzi (1994) mentioned that melamine-formaldehyde and melamine ureaformaldehyde resins are among the most widely used adhesives for wood panels, bothin exposed and semi-exposed conditions because they are more resistant to hydrolysisthan urea formaldehyde adhesives.N^N/\ /'"HN^HN^NHK NH152.4. Degradation of Polymers/AdhesivesPolymers can be degraded chemically by photo-oxidation, thermaldecomposition and oxidation, hydrolysis, or attack by pollutants (Davis and Shaffer1994, Stevens 1999). Jordan mentioned that chemical degradation processes can havephysical origins such as those resulting from heat and light (Jordan 1989).Degradation is an irreversible process, and in general it involves chemicalalteration of the polymer caused by the service environment. Frequently, but notalways, these changes reduce the performance of polymeric materials (Akmal andUsmani 2000). Akmal and Usmani (2000) stated that polymers, whether man-made ornatural, often lose their properties such as appearance, mechanical strength, andoverall integrity when they are exposed to UV light.Some polymers degrade more rapidly than others (Akmal and Usmani 2000).The rate of degradation depends on a number of variables that can be grouped intothree categories: environment, material, and stress (Davis and Shaffer 1994). Stevens(1999) stated that the lower the molecular weight, the more rapidly the polymerdegrades. Nguyen and Rogers (1988) proposed that mechanical stress on melamine-acrylic coatings on an aluminum substrate accelerated degradation. Hence, withincreasing stress the loss of melamine triazine rings and the rate of formation ofhydroxyl and amine groups increased. The acceleration of degradation was consideredto be due to enhanced oxygen accessibility and, possibly, the more direct effects ofstress on the reaction mechanism and its micro reversibility (Nguyen and Rogers 1988).Extensive work on the degradation of many polymers has been undertaken andkey degradation reactions have been identified. According to Akmal and Usmani (2000),it is possible from these sets of reactions, to accurately anticipate the major causes ofdegradation of any new or modified polymer. To investigate the degradation of anadhesive used to bond two materials together, however, it is important to focus on thesystem adhesive and substrate and not only on the adhesive itself (Jordan 1989).162.5. Weathering of AdhesivesMany polymers are prone to photo-chemical degradation, caused by ultra-violetsolar radiation and atmospheric oxygen, which lead to chain scission of the polymerbackbone (White and Turnbull 1992). Such chemical reactions may be accelerated byelevated temperatures and by stress, as mentioned above. Stress may be appliedexternally or can result from a temperature gradient or difference in thermal expansioncoefficients at different locations in the polymer (White and Turnbull 1992).The weathering of polymeric materials is generally due to a complex interactionof the effects of all environmental factors, including solar radiation, moisture, heat,atmospheric pollutants, and others. Some of them only contribute to degradationthrough their effects on secondary reactions that follow bond breakage (Searle 2000).2.5.1. Solar Radiation/HeatSolar radiation, particularly ultraviolet (UV) radiation, is mainly responsible forlimiting the lifetime of materials exposed to the environment (Osawa 1983, Searle2000). UV radiation has enough energy to initiate degradation processes by directlybreaking organic chemical bonds. However, light that is not absorbed can not causeany damage Searle (2000).Approximately 8% of the solar radiation on earth is in the ultraviolet range (280 -400 nm). Its photochemical effect is decisive in the degradation of polymers because ofits higher quantum energy. The photon energy of solar radiation is inverselyproportional to its wavelength (I):(I): E=2.861 04/lambdawhere, E = energy of a photon (kcal/mol) and A wavelength (nm)The cut-off for short-wave UV radiation at the earths surface is determined by theUV absorption of the ozone layer in the atmosphere (Searle 2000). Even long-wave UVradiation, however, may causes photochemical reactions that can modify the characterof polymers. The partial depletion of ozone may lead to the enhancement of shorterwavelength solar radiation reaching the earth. This may lead to acceleration ofphotodegradation of polymeric materials, resulting in a decrease of useful lifetimes for17polymers exposed outdoors (Torikai 2000). Photosensitization of polymers to long-waveradiation caused by dyes, pigments, or processing additives may also reduce thelifetimes of polymers exposed to solar radiation Kockott (1988). The spectral absorptionproperties of polymers are determined by their chemical structure. Pure types ofaliphatic-type polymers (e.g. polyethylene, polypropylene, PVC), and others, are notcapable of absorbing terrestrial solar radiation. The effect of solar radiation on thesepolymers is due to the presence of UV-absorbing impurities and thermal oxidationproducts introduced during polymerization and processing. Furthermore, hydroperoxides and carbonyl groups in polymer chains increase the susceptibility of polymersto photodegradation.Visible radiation absorbed by colored materials can cause deterioration ofpolymers (including color changes). Visible radiation, between 400 and 800 nm,generally makes up 52% of total solar radiation. It also contributes to heating (Kockott1988, Searle 2000). Over 40% of solar radiation has wavelengths longer than 800nm,which is not visible to human eyes, e.g. infrared (IR) radiation. It has a higher irradiancethan UV and visible radiation (Williams 2005). As a result, the temperature of materialsexposed to IR radiation rises automatically (Kockott 1988). Kockott (1988) found that atemperature rise at the surface of an organic material will increase the rate of chemicalreactions, initiated by UV radiation. Davis and Shaffer (1994) concluded that elevatedtemperature conditions tend to degrade adhesive joint strength at a faster rate. Anapproximate rule drawn from reaction kinetics states that a temperature increase of 10°C will approximately double reaction rates. Therefore heat plays a major role in theageing of polymers. In amorphous and partially crystalline polymers, the glass transitiontemperature plays a role, since the mobility of the chains and radicals increases abovethe glass transition temperature and, simultaneously, oxygen diffusion occurs morerapidly (Kockott 1988).UV energy incident upon the surface of a polymer will penetrate into the surfaceto an extent dependant upon its incident intensity. As the UV energy penetrates furtherinto the sample, its intensity becomes increasingly attenuated due to polymerabsorption (Keene et al 2001). The intensity of radiation absorbed by a material obeysLambert-Beer's law; which governs the relationship between absorbance andpenetration depth, absorbtivity, and concentration of material (Osawa 1983). The18absorption of light equals a constant multiplied by the UV absorber concentrationmultiplied by the path lengths (Horsey 1994). As a result the magnitude of chemicalchange to polymer exposed to solar radiation decreases as a function of depth (Keeneet al. 2001). Photodegradation of AdhesivesPhotodegradation is defined as a change in surface chemistry caused by UVradiation, which may result in discoloration and, following prolonged UV exposure, tolosses in mechanical integrity (Muasher and Sain 2006). Stark and Matuana (2006)pointed out that the physical result of photodegradation of polymers is surface crackingand loss in strength and stiffness. Ranby and Rabek (1975) mentioned that in mostmechanistic studies photochemical reactions are carried out in gas or liquid phase, andrarely in the solid phase, whereas in most empirical studies of the susceptibility ofpolymers to degradation expose materials as solids.The initial step involved in the photodegradation of polymers is photo-absorptionat a specified wavelength, depending on the polymer structure and the presence ofimpurities (Torikai 2000). According to Ranby and Rabek (1975) most polymers onlycontain molecules with C-C, C-H, C-0, C-N and C-CI bonds and hence are notexpected to absorb light of wavelengths longer than 190 nm. Furthermore, Searle(2000) mentioned that only short wavelength photons absorbed by the main structuralcomponents are capable of causing severe degradation, such as bond scission. Freeradicals, however, are formed after polymers are irradiation with UV light ofwavelengths > 300nm (Ranby and Rabek 1975). Ranby and Rabek (1975) and Horsey(1994) suggested that chromophores and small quantities of impurities are responsiblefor the absorption of light quanta in polymers. Adhesives can degrade via the photo-oxidative pathway and/or photolytically. In the latter, polymers which are chromophoresabsorb UV light directly without first undergoing oxidation reactions. Examples of thesepolymers are aromatic polyurethanes or polyesters (Horsey 1994), and lignin in wood.Only a small portion of the radiation absorbed by a polymeric material causesphotochemical changes. Other absorbed radiation is eliminated in harmless ways.According to Searle (2000) the activation spectrum of a material, is the point at whichphotochemical reactions occur, depending on the absorption properties of the material,19the spectral emission properties of the light source and the quantum efficiencies of theabsorbed wavelength. The Grotthus and Draper's law states that photochemicalreactions only take place, when molecules are excited to higher states via absorption oflight quantum of sufficient energy. Therefore, the absorption of light is the indispensablefirst step in the initiation of photochemical reactions (Jellinek 1983, Ranby and Rabek1975). Torikai (2000) pointed out the importance of finding out the wavelengthsensitivity of a polymer; which means to clarify what reactions occur and whichwavelength is responsible in order to develop an effective and appropriate protectionsystem. Factors Responsible for Photodegradation of AdhesivesGenerally photodegradation processes in polymers exposed to light are verycomplex because different types of impurities, additives, or abnormal bonds in polymersabsorb UV radiation (Torikai 2000). Industrially produced polymers contain a number oflight-absorbing impurities produced in side-reactions during polymerization, processing,and storage. These impurities can be divided into two groups: Internal impuritiescontaining chromophoric groups, such as hydroperoxides, carbonyls, unsaturatedbonds, and catalyst residues; and external impurities, which may contain chromophoricgroups. These may be traces of solvents, catalyst, and metals or metal oxides fromprocessing containers and equipment. Compounds from polluted urban atmospheresand smog, and various additives may also act as photosensitizers.Torikai (2000) mentioned that a range of other factors also influence thephotodegradation of polymers. These include the molecular weight of the polymer andits distribution, the processing technique, the temperature maintained duringprocessing, the mechanical tension during the preparation of the film, its density, theextent and distribution of crystallinity, the size of the crystallites, the structure at thesurface of the film, the boundary region between crystalline lamella, any defects withinthe crystal cores and the orientation or mobility of the chains. Some of these factors willbe discussed later on in this thesis. Free Radical GenerationAkmal and Usmani (1994) stated that the origin of all the degradative processesin polymers following absorption of light is the initial bond-breaking reaction. Bondcleavage, to produce a free radical pair, can be induced by radiation energy (ultraviolet,gamma, x-ray, and electron beam), mechanical action (e.g. any type of shearingprocess extending the cohesive forces of the molecules such as stretching, grinding,milling, or sheeting), or heat (Horsey 1994, Akmal and Usmani 1994). Photooxidativedegradation of polymers includes such processes as chain scission, crosslinking, andsecondary oxidative reactions (Rabek 1987).Under these circumstances, internal and or external chromophoric groups canproduce low molecular radicals (R*) and or polymeric radicals (1 3*). Photon energyabsorbed by a given chromophoric group can cause polymer dissociation into freeradicals by energy transfer to another group (energy transfer process), or the energycan be accumulated at a given bond (by an energy migration process) (Rabek 1987).In the presence of oxygen and in an extremely fast reaction molecular radicalsand or polymeric radicals form peroxyradicals (ROO*) (Horsey 1994). Abstraction ofhydrogen from a polymer molecule transforms the peroxyradicals into a hydroperoxide.This can further decompose to generate two new free radicals (RO*, HO*), which caninitiate further reactions (Rabek 1987, Horsey 1994). This autocatalytic mechanismresults in autooxidation and an increase in the number of free radicals available toinitiate new degradation reactions. The general autooxidation scheme for hydrocarbonsshown in Figure 2.8, represents the core reaction for hydrocarbon degradation (Horsey1994) and is applicable to both wood and adhesive degradation.21Figure 2.8 has been removed due to copyright restrictionsFigure 2.8: Autooxidation scheme for polymer degradation showing the cyclic nature ofthe process (Horsey 1994)Hydroperoxide can dissociate by thermolysis at temperatures above 100 °C andvia a photolytic pathway under ambient temperatures. According to Klemchuck andGande (1988) the latter is the major cause of photo-induced polymer degradation. Otherphoto-induced reactions can occur due to light absorption by trace levels of carbonylimpurities resulting from thermal degradation of the polymer (Horsey 1994).Subsequent reactions of free radicals amongst each other can result incrosslinking; which is understood as the termination reaction (Rabek 1987). In general,termination mechanisms for free radical reactions are chain scission and cross-linking.The ramification of degradation of polymers via cross-linking includes hardening,skinning, gel formation, a decrease in tack, and an increase in viscosity. Degradationvia chain scission results in softening, a viscosity decrease, an increase in tack, and aloss of cohesive strength. In addition to these physical changes, discoloration is alsopossible (Kockott 1988). According to Stark and Matuana (2006) chain scission canresult in more chain mobility and secondary recrystallization. While cross-linking andchain scission are competitive mechanisms during UV degradation, cross-linking hasbeen shown to be the preferred mechanism during accelerated weathering (Stark andMatuana 2006).222.5.2. Water/HumidityThe effect of humid air on certain adhesives joints was first reported in 1964(Kerr, MacDonald, and Orman 1970). Water is omnipresent in different states in theenvironment and it has a direct impact on polymers which are exposed to it (Kockott1988). As Davis and Shaffer (1994) suggested, moisture is the cause of mostenvironmentally induced bond failures. It can weaken or disrupt secondary (dispersionforces) bonds across the adhesive-adherent interface.Kockott (1988) recommended differentiating between two models of action ofwater on organic materials. First there is the mechanical stressing of the materialcaused by swelling due to water uptake. Secondly, the chemical reaction of water withthe polymer, as seen e.g. in hydrolysis; where water, oxygen and short wave radiationare responsible for free radical generation which can lead for example to chalking ofsome polymers. The latter was observed in TiO 2 (titanium oxide) pigmented coatingsand plastics (Kockott 1988). Hydrolytic attack can also cause chain scission in somepolymers, leading inevitably to deterioration in properties; which can occur unacceptablyfast at elevated temperatures (White and Turnbull 1992). Water is especially aggressiveat temperatures of 70 C and higher. Even rather hydrophobic materials can be affectedby water, as suggested by Jordan (1989).Organic materials have a relatively low diffusion coefficient of 10 -8 cm2/s. In factrapidly changing humidity conditions at the surface of polymers, results in lesspenetration of water into the material, and lower mechanical stresses. According toKockott (1988) in a polymer with a diffusion coefficient of 10 .8 cm2/s, seasonal humidityfluctuations penetrate approximately 1.7 mm into the material. Kockott (1988) pointedout that water absorption capacity and diffusion coefficients are important factors thatinfluence polymer ageing. Moisture can diffuse through the adhesive or the adherent(Moloney et al. 1981), or it can wick or travel along the interface where it can migratevia capillary action through cracks and crazes in the adhesive (Davis and Shaffer 1994).Once moisture is present, it can attack adhesive bonds by:1. Plasticizing the adhesive2. Swelling the adhesive and inducing concomitant stresses233. Disrupting secondary bonds across adherent-adhesive interfaces4. Irreversible altering the adhesive (e.g. hydrolysis, cracking, or crazing)5. Hydrating or corroding the adherent surfaceAccording to Davis and Shaffer (1994) plasticization of the adhesive, whichdepresses Tg (glass transition temperature) and lowers the modulus and strength of theelastomer, is of primary importance in moist environments. Plasticization of theadhesive may also allow disengagement from a micro-rough adherent surface to reducephysical bonding and thus reduce joint strength and durability. On the other hand it mayalso cause stress relaxation or crack blunting and improve durability (Davis and Shaffer1994).2.5.3. OxygenAtmospheric oxygen causes oxidation of polymer matrices and this effect isstrongly accelerated by UV irradiation especially at elevated temperatures and by thepresence of certain impurities that catalyze the oxidation process (Horsey 1994, Jordan1989, Stevens 1999). Hydrocarbon-based adhesive formulations that contain doublebonds are particularly susceptible to oxidation (Jordan 1989); whereas Stevens (1999)pointed out that tertiary carbon atoms are most susceptible to attack. Saturatedpolymers are degraded very slowly by oxygen, and the reaction is autocatalytic.Unsaturated polymers undergo oxidative degradation much more rapidly as a result ofcomplex free radical processes involving peroxide and hydroperoxide intermediates(Stevens 1999). Stevens (1999) stated that singlet oxygen, whose concentration in theatmosphere is negligible (but there are several sources to produce it e.g. decompositionof ozonides), reacts with unsaturated carbon-carbon bonds to produce hydroperoxides.It is further known that polymers containing hydroperoxides have reduced photostability.This was confirmed by Williams (2005) who further mentioned that singlet oxygenquenchers could preclude the formation of hydroperoxides, thereby stabilizing woodagainst photodegradation.Oxidative degradation can occur during the isolation, storage, compounding, andend use of polymers and adhesives. Also for most hydrocarbon polymers, as noted by24Carlsson (1990), photodegradation is not a function of their tendency to absorb light;rather it is a function of their susceptibility to oxidation. To prevent thermal oxidationreactions, antioxidants are used in adhesive formulations (Earhart, Patel, and Knobloch1994); but Jordan (1989) suggested that antioxidants are not capable of stoppingoxidation reactions completely. Turton and White (2001) mentioned that the reactionrate near the surface of a polymer exposed to UV is very high and most of the oxygen isconsumed before it can penetrate far into the material. The most aggressive form ofoxygen is ozone, for which special antioxidants have been developed (Jordan 1989,Stevens 1999). Nitrogen (N) has no significant affect on adhesives. The impact of otheratmospheric air constituents such as sulfur dioxide and nitrogen oxide on thedegradation of polymers has not been fully investigated (Jordan 1989).Stark and Matuana (2006) found out that carbonyl growth shows a linearrelationship with growth in surface oxidation. According to Dorn and Breuel (1992) thepresence of carbonyl groups in a degraded polymer indicates that oxidation has takenplace and, furthermore, that the material is vulnerable to further deterioration becausesuch groups are photo-labile.2.5.4. PollutantsIndustrial pollutants have a direct impact on the weathering of polymers withoutany interaction with other weathering factors. Laboratory experiments have shown thatpollutants cause degradation of polymers (such as Nylon 6), even in the absence of UVradiation (White and Turnbull 1992). Nichols (2005) pointed out that atmosphericpollution can lower the pH of rainfall, leading to acid etching of coatings. However,significant loss of properties occurs only as a result of the combination of differentweathering factors. For example sulfuric acid in combination with water and solar UVradiation is responsible for rapid fading (Kockott 1988). Some of the pollutantsthemselves are photolytic, leading to the generation of additional products that maycause degradation (White and Turnbull 1992). The common atmospheric pollutants inindustrialized regions are sulphur dioxide, oxides of nitrogen and carbon, and ozone.There is significant synergy between their activity and photodegradation.252.6. Weathering of WoodWood undergoes rapid surface degradation during exterior exposure dueprincipally to the effects of light and water (Feist and Hon 1984). The term weatheringdescribes the surface degradation that occurs when wood is used outdoors and aboveground. As Feist and Horn (1984) pointed out weathering should not be confused withdecay caused by fungi, which can extend deeply into wood and significantly reduce thestrength of structural timber. Evans (2005) stated that microorganisms play a minor rolecompared to environmental factors in the degradative processes involved in weathering.According to Yusuf and co-workers (1995), the main factors that are responsible for thenatural weathering of wood are solar radiation, especially the UV component, moisturein its different states, heat and atmospheric gases. They emphasized that the UVportion of the solar spectrum is more energetic than visible light and most effective atcausing degradation of wood (Yusuf et al. 1995). Visible light is also involved inweathering, according to Derbyshire and Miller (1981). Nearly 30% of softwood iscomposed of lignin, which contains numerous chromophores that efficiently absorb UVradiation (Heitner 1993). Water washes degraded products such as lignin andhemicelluloses from weathered wood, but cellulose remains on the surface becausesome of its degradation products are less water soluble (Muasher and Sain 2006). Feist(1990) mentioned that the most obvious features of weathered wood are its greycoloration, rough surface texture and presence of checks. Extractives are leached fromweathered wood surfaces and this, in combination with loss of lignin makes the surfacemore hydrophilic (Williams 2005). Wood is naturally a hydrophilic material and wood cellwalls swell when they absorb water. This facilitates deeper light penetration andprovides sites for further degradation (Stark and Matuana. 2006). The effects ofweathering are superficial in nature and are confined to the upper 2-3 mm of wood,except for checks which can extend more deeply into wood. Kataoka, Kiguchi, andEvans (2004) found that the first 75 pm thick layer of wood absorbed 90% of UV light(350 nm), and the first 220 pm thick layer absorbed 90% of visible light (420 nm). Onepercent of UV and visible light, however, was capable of penetrating wood to depths of150 and 440 pm, respectively. It is generally understood that low density earlywooddegrades more rapidly than denser latewood, but it is not clear to what depths light isstill photo-chemically active in wood (Kataoka, Kiguchi, and Evans 2004). Hon and Feist26(1981) stated that crystalline regions of cellulose are impervious for UV light and theysuggested that this might explain why denser late wood degrades less rapidly than earlywood.2.6.1. Photodegradation of WoodPhoto-oxidation of lignin refers to a process where lignin undergoes chemicalmodifications such as bond cleavage, and hydrogen abstraction resulting in radicalformation, creation of peroxides with oxygen and finally its decomposition and theproduction of colored and hydrophilic by-products, so called chromophores. Thesemodifications induce surface property changes such as discoloration, increased watersensitivity followed by hydrolysis, leaching and cracking (Hon 1991). Lignin is reportedto absorb UV light with a maximum at 280 nm and decreasing absorption extendingbeyond 380 nm into the visible region of the solar spectrum (Kalnins 1966). Generally,photodegradation is triggered by the absorption of UV light by lignin and the formationof free radicals as suggested by Yusuf and his co-workers (1995). Lignin has variousreactive groups that react with light to form free radicals (Stark and Matuana 2006, Hon1981).The mechanisms of photo-oxidation of lignin (Figure 2.9) are only partiallyunderstood and documented (Leary 1994). It is generally accepted that at least fourreaction pathways are involved in the degradation of lignin: (1) direct absorption of UVlight by conjugated phenolic groups to form phenoxyl free-radicals; (2) abstraction ofphenolic hydrogen as a result of aromatic carbonyl triplet excitation to produce a ketyland phenoxyl free-radical (Kringstad and Lin 1970); (3) cleavage of non-phenolicphenacyl-a-O-arylethers to produce phenacyl phenoxyl free-radical pairs (Gierer andLin 1972), and; (4) abstraction of benzylic hydrogen of the a-guaiacylglycerol-p-arylether group to form ketyl free-radicals which then undergoes cleavage of the r3-0-4ary-lether bond to produce an enol and phenoxyl free-radical (Scaiano, Netto-Ferreira,and Wintgens 1991). The enol then tautomerizes to a ketone. Alkoxyl and peroxyl free-radicals produced from the reaction of oxygen and lignin free-radicals react with thephenoxyl free-radical formed to produce colored chromophores, e.g. quinoides,aromatic ketones, aldehydes, and acids as photo-degradation products (personalcommunication 2007, Schaller).27a-Guaiacylaceto-veratiora (a-GAV)-50% in lignina-Guaiacylaceto-veritrore (a-GAV)-7% in lignin CtoeR- H+-R* VisBtle lightlit"Pheno xy"(1990)*CI"Ketyl"(1992)"9^0B-0-4cleavage HO''1'Phew}lyradical^i.igninOxidationbyRO*, ROO*Ketoneradical CLLigninRtoa2Figure 2.9: Pathways of lignin photo-oxidationAs described by Derbyshire and Miller (1981) there are two mechanisms for thephotodegradation of wood, direct photolysis of lignin (as described above) and photo-sensitized degradation of cellulose. Photolysis involves the rupture of covalent bonds inlignin and extractives (Derbyshire and Miller 1981). The rate at which wood cellulosecleaved is reported to be dependent on the wavelength of the incident light.Wavelengths less than 280 nm increase the rate of photolysis, but the rate is slower onexposure to wavelengths longer than 340 nm (Evans et al. 2005). Photo-sensitizeddegradation is the most likely explanation for the breakdown of cellulose (Derbyshireand Miller 1981). Also Hon (1981) suggested that UV light absorbed by lignin helped todegrade cellulose by energy transfer.282.7. Photo-stabilizationPotter (1970) suggested that an understanding of the mechanism by which apolymer deteriorates may lead to new approaches to stabilization. Light stabilizers areadditives to polymeric materials which prevent photodegradation or crosslinking causedby UV light (Ranby and Rabek 1975, Rabek 1987). A number of books and reviewsdiscuss photostabilization mechanisms and photostabilizer properties. As pointed out byRabek (1987), 'the practical problem in the study of photostabliization of polymers is avoluminous patent literature on these products, whereas a highly fragmented and oftencontradictory scientific literature has attempted to answer mechanistic questions'.However, only a few classes of photostabilizers are of commercial significance (Ranbyand Rabek 1975).Any degradation by UV radiation is close to the surface of a material because theattenuation of light with increasing substrate depth. Therefore, the concentration of UVabsorbers should be highest at the surface (Rabek 1987). In general, the concentrationof stabilizers added to the materials is a function of the thickness of the substrate. Thethinner the polymer, the higher the concentration of UV absorber needed to provideadequate protection. In a thin film the surface layer is a much larger fraction of the totalfilm volume than it is for a thick film. Therefore, the surface layer plays a more importantrole in affecting the total properties of the film (Rabek 1987). Or as Decker (2001) stated'the radiation filter effect of UVA is much greater when the entire light stabilizer isconcentrated in a thin layer on the materials face exposed to sunlight'. The radiationfiltering effect of UVA increases exponentially with its concentration according to theLambert-Beer law (Decker, Zahouily, and Valet 2001). Also any type of additive(photostabilizer, antioxidant, thermal stabilizer, etc.) must be evenly distributed withinthe polymer, which requires it to be compatible with the polymer. Stabilizers are usuallymore compatible with the amorphous part of a polymer, and may be excluded fromsemi-crystalline polymers. Furthermore, oxygen diffusion is higher in the amorphousregion. It is clear that much of the photooxidative reaction and stabilization occurs in theamorphous polymer region, and high stabilizer concentrations are required there(Rabek 1987).Nowadays, two general classes of light stabilizers are mainly used to protectpolymers from photodegradation, UV absorbers (UVA) and hindered amine light29stabilizers (HALS) (Horsey 1994, Rabek 1987). It is known that the use of HALS with aUV absorber provides excellent stabilization in many polymer compositions according toBerner and Rembold (1981). Muasher and Sain (2006) have shown a synergetic effectof UV absorber in combination with HALS in the UV protection of low densitypolyethylene. Also Decker, Zahouily, and Valet (2001) stated that the combination ofUVA and HALS has a pronounced stabilizing effect on the photodegradation ofmethoxylated melamine, trifunctional isocyanate and polyurethane coatings. It isunderstood that UV stabilizer operate most efficiently by energy transfer from excitedchromophores in the polymer, rather than solely by a screening mechanism (Ranby andRabek 1975).Another additive, which is used to protect adhesives from oxidation when theyare subjected to high temperatures during processing and compounding, is an anti-oxidant. These only provide limited protection against UV-initiated oxidation (Horsey1994).2.7.1. UV Absorbers (UVA)UV absorbers must fulfill certain requirements. They have to absorb strongly inthe UV region (290 to 400 nm) and they need a sharp cut-off in the visible region(>400nm), so as not to color the polymer. They also must be quite photo-stable anddissipate the photoexcitation energy in a harmless way (Horsey 1994, White andTurnbull 1992, Renz 2001). Furthermore, Horsey (1994) explains that UV absorbers aremore effective in protecting thick polymers than thin films, for reasons that arediscussed above. This explains why UV absorbers are widely added to sealants buthave limited efficacy in adhesive films, where typical thickness may be about 0.025mm.Decker, Zahouily, and Valet (2001) pointed out that UV absorbers also need to have thelongest possible lifetime besides their good photostability. Furthermore he mentionedthat the light stability of a UVA is not only dependant on its chemical structure, but alsoon the type of polymer it is applied to. The lifetime of an UVA is increased by theaddition of HALS radical scavenger, because it helps the UVA to remain in theprotective film, thus ensuring a long-lasting filtering effect (Decker, Zahouily, and Valet2001). Information on the spectral response of a polymeric material to the source ofradiation to which it will be exposed under use conditions is important in the30development of stabilizer systems to prolong its lifetime. The activation spectrumdetermines the wavelengths in the exposures source that are harmful to a specificmaterial and thus the type of UV absorber needed for optimum screening protection(Searle 2000). The effectiveness of a UV absorber at protecting a material depends onits ability to compete with the polymer in absorbing actinic wavelengths. Differentspectral regions may be responsible for the various types of degradation such asyellowing, loss in impact strength, and bleaching, as shown with ABS, polyethylene, andpolysulfone exposed to borosilicate-glass filtered xenon arc radiation (Searle 2000).Classes of UV absorber include salicylate, cyanoacrylate, malonate, oxanilide,benzophenone, s-triazine and benzotriazole (Renz 2001). From a technical point ofview the hydroxyphenyl-benzotriazoles and hydroxyphenyl-s-triazines are the mostimportant classes of UVA (personal communication 2006, Krause). Hydroxyphenyl-benzotriazoleBenzotriazole UV absorbers are particularly useful for protecting polymersbecause of their increased photostability in comparison to other UV absorbers (Renz etal. 2001). However, Renz and colleges (2001) reported that there is a need for abenzotriazole that is highly soluble and provides added protection in the 350 to 400 nmregion of the ultraviolet spectrum. The chemical structure of a hydroxyphenyl-benzotriazole is presented in Figure 2.10.Figure 2.10: Chemical structure of hydroxyphenyl-benzotriazoleRanby and Rabek (1975) mentioned that hydroxyphenyl-benzotriazole may forminternal hydrogen bonds (Figure 2.11), as in the case of hydroxybenzophenones. Theirphotostabilization mechanism is also considered to be rapid tautomerism in their excitedstates.31+ hv-AFigure 2.11: Internal hydrogen bond formation and tautomerization in a hydroxyphenyl-benzotriazole2.7.1.2. Hydroxyphenyl-s-triazineThe mechanisms by which hydroxyphenyl-s-triazine UV absorbers stabilizepolymers are similar to those of hydroxyphenyl-benzotriazoles. The chemical structureof hydroxyphenyl-s-triazine is shown in Figure 2.12.Figure 2.12: Chemical structure of hydroxyphenyl-s-triazine2.7.2. Hindered Amine Light Stabilizers (HALS)The second most important type of light stabilizer, HALS, provides stability topolymers and can function in thin films or on sample surfaces. Unlike UVA, HALS don'tabsorb UV radiation and therefore they have no effect on the incident light. Theprotective mechanisms of HALS involve the destruction of unstable and radical initiatinghydroperoxides and scavenging free radicals before they can be involved inpropagation reactions (Horsey 1994, White and Turnbull 1992). HALS were developedon an industrial scale during the 1980's. According to Rabek (1987) the simplest modelcompound for a HALS is 2,2,6,6-tetra-methyl-piperidine.322.7.2.1. 2,2,6,6-Tetra-methyl-piperidineIt is generally accepted that hindered piperidine produces hindered piperidinoxyradicals during uv irradiation in the presence of oxygen (02) and free radicals (R*) asshown in Figure 2.13.CH3 + hv +02 + R • Figure 2.13: Formation of hindered piperidinoxy radicals from hindered piperidinePiperidinoxy radicals belong to the class of stable organic free radicals and theirformation mechanism is not fully understood. Their photostabilization mechanism andits kinetics are complicated because different piperidine derivates are participating instabilization reactions and they are still being investigated (Rabek 1987, Horsey 1994).Nonetheless, Horsey (1994) pointed out that the primary protection mechanism ofHALS involves free radical scavenging and hydroperoxide decomposing.2.7.3 AntioxidantsEarhart, Patel, and Knobloch (1994) mentioned that antioxidants are used toinhibit degradation as a result of thermal oxidation. Ranby and Rabek (1975) classifiedantioxidants into two groups according to their protection mechanisms: 1, kinetic chainbreaking antioxidants or primary antioxidants; 2, peroxide decomposers or secondaryantioxidants. The primary antioxidants inhibit oxidation via a rapid chain-terminatingreaction. Stabilization is achieved through proton donation from the —OH group of theprimary antioxidant to a peroxy or alkoxy radical (Earhart, Patel, and Knobloch 1994).The performance of primary antioxidants can be improved by the use of secondaryantioxidants as suggested by Earhart, Patel, and Knobloch (1994). The latter do not actas radical scavengers but they undergo redox reactions with hydroperoxides to formnon-radical, stable products. Secondary antioxidants are generally used exclusively incombination with primary antioxidants (Earhart, Patel, and Knobloch 1994).332.8. Weathering of Wood Polymer CompositesThe photodegradation of wood and plastic separately are well documented inliterature (Muasher and Sain 2006), but the photodegradation of blends of wood andpolymers has received little attention.2.8.1. Wood Plastic Composites (WPC)It has been estimated that between the years 2004 and 2009, wood plasticdecking will grow by 23% annually; this is mainly due to the fact that compared to otherdecking products wood plastic decking promises lower maintenance and greaterdurability (Stark and Matuana 2006).Weathering resistance of WPC is generally poor despite claims to the contrary.Kiguchi et al. (2000) point out that discoloration, chalking and dimensional changecaused by exterior exposure are major problems for the use of wood plastic compositesoutdoors. The color changes and losses in mechanical properties of wood plasticcomposites exposed to accelerated weathering conditions were examined by Muasherand Sain (2006), and Stark and Matuana (2006). Muasher and Sain (2006) mentionedthat the photodegradation of wood plastic composites is complicated because eachcomponent, namely wood and plastic, may degrade via a different mechanism.Experiments have shown that both wood and plastic caused discoloration (whitening) ofthe composite. Dark pigments improved color stability, however, chalking on thesurfaces of wood plastic composites still occurred (Kiguchi et al. 2000). Also Stark andMatuana (2006) pointed out that exposure of wood plastic composites to UV radiationcan result in changes in both the polymer matrix and the wood component. Further theymentioned that wood plastic composites experienced much less color changes whenexposed to UV (xenon arc type) alone. The majority of losses in mechanical propertiesduring weathering were caused by moisture effects; in addition to the acceleration ofoxidation reactions caused by water (Stark and Matuana 2006). The water also washedaway the degraded polymer layer and removed wood extractives. The swelling of thewood fiber created micro-cracks in the composite matrix which facilitated deeperpenetration of light (Stark and Matuana 2006).342.8.2. Glulam BeamsThe movement away from solid wood to engineered wood products forconstruction has increased the consumption of adhesives (Frihart 2005). Given theweight of adhesive (2 — 8%) compared to the product weight, cost is an issue accordingto Frihart (2005); and in the context of wood composites, adhesive development isdriven by adhesive cost-reduction (Berglund and Rowell 2005). Glued laminated woodconstruction (glulam) had its beginning at the turn of the 20th century when Otto KarlHetzer (1846 — 1911) of Weimar, Germany, obtained his first patent for this method ofconstruction. Max Hanisch, Sr. was the first person who brought this new technology toAmerica (Rhude 1996).The effect of UV light on glulam that is exposed to the sun or other sources of UVlight are of great concern to the manufactures of such articles (Renz et al. 2001).Hence, the degradation of glue-laminated beams during weathering has been studiedby several scientists. The delamination of glue lines is of particular concern for glulamexposed outdoors. Frangi, Fontana, and Mischler (2004) found out that the shearresistance of gluelam decreased with increasing temperature, when 1-K-polyurethanewas used to bond laminae together. Selbo (1965), however, mentioned that melamineformaldehyde glue joints in laminated Douglas-fir beams exposed to the weathershowed practically no delamination after about 20 years exposure. Changes of woodmoisture content may result in bondline stresses which influence bond-line durability.Phenol resorcinol formaldehyde resin with resorcinol content greater than 16%enhanced the durability and lowered the delamination of CCA treated radiata pinegluelams (Lisperguer and Becker 2005). According to Hofferber et al. (2006) acetylatedwood bonded with epoxy adhesives may be expected to show a reduction of theswelling strain and thus the bondline stress, compared to untreated wood. Meierhoefer(1986) demonstrated that waterborne and oil preservatives have a positive effect onmoisture change, linear expansion as well as development of checks and delaminationof glulam. It is also possible to protect gluelam to some degree against surfacechecking by using an appropriate coating formulation. Nevertheless, due to the slowdiffusion of moisture through the coating, swelling of the wood and thus cracks in thecoated surface may still occur. To prevent this phenomena from occurring the surfacecoating needs to be frequently renewed. It has been found that the diffusion resistanceincreases linearly with increasing coating thickness. Thus the protection mechanisms of35the coating are assured for a longer period of time when thicker coatings are used (Sell1983).Tests have been developed to predict the lifetime of gluelam beams. Accordingto Deppe and Schmidt (1987) the results from short-term tests with glulam may notpredict their long term behavior. Hence, they developed an accelerated weathering testfor glulam with the aim of predicting the long term behavior of glulam exposed outdoors.The accelerated weathering test involved preliminary exposure of samples in a xenotestweatherometer followed by moisture cycling. Such a test was more effective thanmoisture cycling alone at reducing glue-bond strength of small glulam specimens(Deppe and Schmitt 1987).2.8.3. Photocatalytic Effect of Wood on Polymer DegradationThere has been very little research on the photocatalytic effect of wood onpolymer degradation. Kiguchi and coworkers (2006) suggested that lignin generatedfree radicals during photo-irradiation that attacked polypropylene (PP) in wood plasticcomposites. Polypropylene containing photostabilizers that was artificially weathered for500 hours didn't show any color changes, which was quite different from that observedfor wood plastic composites containing polypropylene. According to Kiguchi et al. (2006)the bleaching of WPC was not only caused by photodegradation of woodfibre, but alsoby degradation of polypropylene.Matuana, Kamdem, and Zhang (2001) have done experiments with PVC/woodfiber composites and they concluded that wood fibers are effective sensitizers, andincorporation of wood into a PVC matrix accelerates the photodegradation of the PVCmatrix. Their explanation is that the photoinduced breakdown of lignin in wood fibersleads to the formation of chromophoric functional groups such as carbonyls, carboxylicacids, quinones, and hydroperoxy radicals. They suggested that these chromophoricgroups accelerate the degradation of polymers in the composite material (Matuana,Kamdem, and Zhang 2001).362.9. ConclusionAccording to Dorn et al. (1992) high humidity and heat have the greatest impacton the strength of adhesives. Osawa (1983) and Searle (2000) both mentioned that UVlight is in particular responsible for limiting the lifetime of polymers. Weathering involvescomplex processes that are still not completely understood for most materials (Searle2000). Different climatic zones show different magnitudes of each weathering factor,which makes it difficult to develop stable adhesive formulations for all conditions(Jordan 1989). Furthermore, different weathering parameters impact the adhesiveusually in combination with each other. Results for individual adhesive types testedunder certain weather conditions are not necessarily applicable to other types ofadhesives. Therefore, it is recommended to conduct aging tests individual adhesivesunder different end uses (Dorn and Breuel 1992).At present the durability of glue lines in wood composites such as glulamexposed outdoors is not of great concern for industry. However with an increasing useof wood-adhesive composites outdoors it will be necessary to increase ourunderstanding of the photodegradation of adhesives used in glulam and the interactionof wood and adhesive degradation.373. Weathering and Photostability of Adhesives3.1. IntroductionThe history of adhesives is closely related to the history of humans (Keimel1994). Adhesives surround us both in nature and in our daily lives. The applications ofadhesives are numerous as mentioned in the introduction to this thesis. Petrie (2007)concluded that the study of adhesives and the sciences surrounding their applicationhas never been more important.Adhesives used outdoors are exposed to solar radiation, rain, heat, atmosphericpollutants etc. The combination of these parameters is deleterious to adhesives, andcan result in a significant reduction in the performance of bonded products (Dorn andBreuel 1992). Hence, the weathering of adhesives is a major concern (Kockott 1988,Dorn and Breuel 1992, White and Turnbull 1992). Different approaches have beentaken to improve the stability of adhesives used outdoors (Horsey 1994, Comyn 1998,Burger 2000). Accelerated aging devices are used routinely to provide predictive andreproducible data on the performance of adhesives in a reasonable amount of time(Horsey 1994). The performance of individual adhesive types tested under specificweather conditions, however, is not necessarily applicable to other adhesive types.Therefore, it is prudent to conduct aging tests for different adhesive types (Dorn andBreuel 1992). In order to understand the resistance of adhesives in wood composites toweathering, which is one of the aims of this study, it is important to evaluate theirphotostability. This chapter examines the photostability of four clear, light colored,adhesives that may have potential for use in glulam exposed outdoors. This chapteralso examines the ability of a selected UVA/HALS stabilizer to protect adhesivesexposed to artificial weather conditions. A secondary aim of this chapter was toevaluate the effect of different weathering regimes in two different weatherometers onthe degradation of the selected adhesives.383.2. Materials and Methods3.2.1. Experimental Design and Statistical AnalysesAn experiment was designed to examine the effects of three factors: 1, adhesivetype (melamine formaldehyde, epoxide (two types) and, emulsion polymer isocyanate);2, stabilizer and 3, the interaction of adhesive and photostabilizer on weatheringresistance of adhesive films. Loss in weight and tensile strength of adhesive films andchanges in color following exposure in two different weathering devices (QUV andXenon-arc) were used to assess weathering resistance of the different adhesives. Twosets of adhesive samples were prepared and exposed in the different weatherometers.SEM (Scanning Electron Microscopy) and FTIR (Fourier Transform Infrared)spectroscopy were also used to investigate the effects of weathering on the differentadhesives. Results from SEM and FTIR, however, were not subject to statisticalanalysis.Analysis of variance was used to examine the effects of adhesive type, stabilizer,and stabilizer/adhesive interaction on factors of interest (tensile strength, weight loss,and changes in color). The results from specimens exposed in the differentweatherometers were analysed separately. Computation was performed using Genstat5 (Genstat 2000). Before the final analysis diagnostic checks were performed todetermine whether results conformed to the assumptions of analysis variance, i.e.normality with constant variance. As a result of such checks, tensile strength and weightloss data was converted to natural logarithms before analysis of variance. Significantresults are plotted graphically and bars representing the least significant differences areincluded on graphs to facilitate comparison of means. Appendix 1 contains all data andstatistical output.393.2.2. Sample PreparationFour different adhesives were assessed (Table 3.1). All of them were nearlycolorless because of the market preference for clear glue lines in laminated woodproducts. Three of them were thermosetting adhesives and the other one, was athermoplastic adhesive with properties close to thermosetting adhesives (Dynea, n.d.).Table 3.1: Adhesives assessed for their resistance to weatheringAbbreviation Full name Chemical name Current applicationEpE Epoxy resin andhardenerButyl glycidyl ether ofbispenol-A withpolyamide blendMetal bonding, aircraftindustryEpU Epoxy resin andhardenerEpoxy resin withglycidyl ether andpolyoxy propyleneamine typeMetal bonding aircraftindustryMF MelamineformaldehydeMethylated melamineformaldehyde polymerand catalystImpregnating papersheets, particleboard,plywoodEPI Emulsion PolymerIsocyanatePolyvinyl acetate anddipenylmethane-diisocyanateWood glue forenvironmentallydemanding situations,plastic-to-woodThe photostabilizer that was added to adhesive films contained two compounds:(1) Hindered amine light stabilizer (Bis(1,2,2,6,6-pentamethy1-4-piperidyl) sebacate)(trade name: TINUVIN 292); and (2) benzotriazole UV absorber (Iso-octy1-3-(3-(2H-benzotriazol-2-y1)-5-tert. Butyl-4-hydroxyphenyl propionate) (trade name: TINUVIN 384),in a ratio of 1:1 (2% by weight of the adhesive film). Their chemical structures areshown in Figure 3.1.40CH3 H3C CH30^CH2 rj 0 ^ N-CH3H3C-NH3CH3CCH2CH2C00C8H1 7(1) HALS^ (2) UVAFigure 3.1: Chemical structures of (1) HALS (Tinuvin 292) and (2) UVA (Tinuvin 384)The different adhesives and hardeners were mixed together in separatepolystyrene cups with a total tare weight of -12 grams for each adhesive. Theadhesives were spread out with a paint gauge on a glass surface to form filmsapproximately 450 pm in thickness (Figure 3.2). The two epoxy adhesives were curedat 60 t in an oven for 150 minutes (EpE) or 180 minutes (EpU). The MF and EPIadhesives were cured under ambient temperatures for 180 minutes. After the adhesivefilms were cured they were carefully removed from the glass surface with a razor blade(Figure 3.2). Afterwards they were placed on a piece of polyvinylchloride (flooring gradePVC) and 320 samples were punched out from the film strips using a dog-bone shapedcutter (Figure 3.2). One sample per film strip was punched out and allocated toscanning electron microscopy. The samples were labeled on their roughest side(opposite to the one that was adjacent to the glass) using a permanent marker. Allsamples were placed between plotting paper under weight for three to four weeks in aconditioning room 20 ± 1 C and at 65 ± 5% relative humidity. The procedure forpreparing the dog-bone adhesive samples is shown in Figure 3.2. Before and after theadhesive dog-bone samples were exposed to artificial weathering they were weighedon an AND GR-200 balance, and their color was measured.41Figure 3.2: Preparation dog-bone adhesive samples; 1, mixing adhesive formulation; 2,spreading out adhesive with a paint gauge; 3, removing adhesive film after cure with arazor blade; 4, samples punched out with a dog-bone shaped cutter; 5, adhesive dog-bone specimen423.2.3. Colour MeasurementColor measurements were performed using a spectrophotometer (MINOLTA CM2600d). A total of 505 color measurements were made. Color is expressed using theCIE (Commission Internationale de l'Eclairage) LAB space system, which consists ofthree parameters; L* is Lightness (0=black; 100=white), a* is greenness/redness (-60=green; 60=red), and b* is blueness/yellowness (-60=blue; 60=yellow). Thespectrophotometer was set to SCI (specular component included) and measurementswere performed on the side of samples which was exposed to artificial weathering(smoother glass side). Two measurements on the same area of each sample werecarried out before and after weathering. The samples were always placed on a yellowplotting paper when their color was measured. Areas of dog-bone specimens thatcontained voids were avoided when color measurements were made. The total colorchange deltaE occurring as a result of weathering was calculated according to ASTMD2244 (ASTM 1993). The equation for the calculation of deltaE is (II):(II): DeltaE = ((l—before — Lafter) 2 + (abefore — aafter) 2 + (bbefore — befter) 2r 53.2.4. Artificial WeatheringAccelerated aging devices are used routinely to provide predictive andreproducible data on the weathering resistance of materials in a reasonable amount oftime (Horsey 1994). Laboratory test data from such devices is only applicable toprediction of in-service performance if the mechanism and type of degradation are thesame under both types of exposure conditions. The opposing effects of differentspectral regions, on the same material shown by the activation spectrum, demonstratesthe importance when using accelerated weathering devices of closely simulating naturalconditions over the full range of actinic wavelengths. Because the net effect of theradiation source depends on the relative intensities of the short versus longwavelengths, simulation of long wavelength UV and visible radiation can be just asimportant as simulation of short wavelength UV radiation (Searle 2000). Reliablemethods for lifetime prediction of many materials have not yet been developed (Whiteand Turnbull 1992, Kaempf 1984).43In this research, samples were exposed in two different types of weatherometersto increase confidence on conclusions drawn about the weathering resistance ofdifferent adhesive/stabilizer combinations. The QUV device is cheaper to buy andhence is commonly used by industry. Xenon-arc when properly filtered closelysimulates the full spectrum of the solar radiation. The fluorescent UVA 340 lamps usedin QUV devices only match well with daylight in the 300 — 350 nm wavelength region(Searle 2000).One set of specimens was exposed for 200 hours in a QUV/spray acceleratedweathering tester (Q-Panel Lab Products, Cleveland, USA) equipped with fluorescentUVA 340 lamps. The exposure cycle started with a 6 hour conditioning phase at 50 C tocreate moisture stress in subsequent cycles. The second step consisted of 8 hours UVlight at 70 C with 0.68 W/m 2/nm, followed by 4 hours conditioning at 50 C. This cyclewas repeated 16 times (without the use of water spray).The other set of samples was exposed in a Xenon-arc weatherometer (Model:65-W) for 200h. This weatherometer contained a 6500 watt xenon arc lamp withborosilicate inner and outer filters (Atlas Electric Devices, Chicago, USA). The sampleswere exposed to a slightly different cycle compared to those exposed in the QUVdevice. This was mainly due to the different technical configurations of the twoweatherometers. The main differences between these two weatherometers, besidestheir different light sources, are that the Xenon-arc device doesn't have a humidifier andits maximum inner temperature is close to 50 C. In the first step of the Xenon-arc cycle,the samples were exposed to 6 hours of darkness under ambient temperatures. Duringthis stage water spray was used to increase the humidity of the chamber. The secondstep consisted of 8 hours UV light where the chamber temperature increased to amaximum of 50 C, followed by 4 hours of darkness with water spray. The water spraywas directed onto a shield, which increased the humidity in the chamber and avoidedwashing the specimens out from their holders. This cycle was repeated 16 times. Figure3.3 and Figure 3.4 show the QUV and Xenon-arc weatherometers, respectively. Inorder to remove specimens from their specimen holders after artificial weathering, theywere immersed in warm water at about 40C for a maximum of 5 minutes.44Figure 3.3: QUV weatherometer and dog-bone specimens in sample holderFigure 3.4: Xenon-arc weatherometer and dog-bone specimens in sample holder453.2.5. Tensile Strength TestingTensile strength measures the stress required to pull the adhesive dog-bonesamples to a point where they break. Tensile strength forces act perpendicular to theplane of the tested material and, ideally, are distributed uniformly over the cross-sectional area. The fracture will develop where the local stress exceeds the localstrength (Petrie 2007). Several studies have shown a relationship, which is not alwayslinear, between loss in tensile strength and increasing exposure time to UV light (Turtonand White 2001, Andrady, n.d.).A QTS 3 Quick Test device from Pruefpartner GmbH (see Figure 3.5) was usedto measure the tensile strength of specimens after weathering. Strength tests were alsoperformed on matched, unexposed, controls. Samples were mounted vertically in thejaws of the device and a tensile force was applied to specimens. The testing speed was20 mm/min, and the maximum testing span of the jaws was 10mm.Figure 3.5: QTS 3 Quick test tensile strength tester from Pruefpartner GmbH463.3 ResultsThe effects of adhesive type, stabilizer and the interaction of adhesive/stabilizeron tensile strength, weight loss and L* (lightness), b* (yellowness), a* (redness) anddelta E (total color change) of specimens after accelerated weathering in the Xenon-arcor QUV weatherometer are summarized in Table 3.2 and Table 3.3.Table 3.2: Significant effects of and interactions between adhesive type and stabilizeron resistance of adhesive films to accelerated weathering in a Xenon-arc deviceResponse variable Experimental factorsAdhesive type^Stabilizer^Adhesive/StabilizerTensile strength .„„ NS (p = 0.129) NS (p = 0.537)Weight change *** NS (p = 0.731) NS (p = 0.683)L* ratio (Lightness) *** ...., „,,B* ratio(blue — yellow) „, ,,, ,,,A* ratio (green — red) NS (p=0.143) NS (p=0.267) ,,DeltaE ,,, .., ,,,p < 0.01; — = p < 0.001; NS = not significant (p > 0.05)Table 3.3: Significant effects of and interactions between adhesive type and stabilizeron resistance of adhesive films to accelerated weathering in a QUV deviceResponse variable Experimental factorsAdhesive type^Stabilizer^Adhesive/StabilizerTensile strength ,,, NS (p = 0.300) NS (p = 0.825)Weight change ,,, NS (p = 0.499) *L* ratio (Lightness) „, NS (p = 0.569) ,,,B* ratio(blue — yellow) *** *** ,„A* ratio (green — red) *** NS (p=0.589) ***DeltaE 1r** *** ***= p< 0.05; *— = p < 0.001; NS = not significant (p > 0.05)47148.0054.6020.107.402.700.003.3.1 Tensile StrengthThe adhesive type had a significant (p<0.001) effect on the tensile strength of theadhesive films exposed in both types of weatherometers (Table 3.2 and Table 3.3).Figures 3.6 and 3.7 show the tensile strength of glue-film specimens before and afterexposure in the Xenon-arc and QUV devices, respectively. The adhesive EpE had thehighest tensile strength followed by the melamine formaldehyde specimens. The tensilestrength of the EpU epoxy adhesive and the EPI emulsion polymer adhesive specimenswere significantly lower. The strength of the different adhesive types was similar in thetwo different weatherometers (compare Figure 3.6 and Figure 3.7). There were smalllosses in strength of EPI, EpE and MF specimens exposed in the Xenon-arcweatherometer. Strength losses of these adhesives were more pronounced afterexposure in the QUV weatherometer, particularly in the case of the EPI adhesive. Incontrast the tensile strength of EPU specimens increased after exposure in the Xenon-arc and QUV devices.5.00 ^LSD=0.3515 X CX CEPI^EpE^EpU^MFAdhesive typeFigure 3.6: Tensile strength of glue-film specimens (X) after exposure in Xenon-arcweatherometer for 200h (results averaged across stabilized specimens and unstabilizedcontrols). Unexposed controls (c) are shown for comparison. Y1 axis shows result on alog scale (natural logarithms, In) because results were analysed as logs. Y2 axis showsresults on a normal scale. The LSD bar is an error bar that can be used to comparemeans and is used in preference to other types of error bars2 4.00g 3.002.00I 1.000.00485.002. 2.700.00 EPI MFa C Ca Ca148.0054.6020.10111;7.40 0us60.00EpE^EpUAdhesive typeCLSD=0.2984Figure 3.7: Tensile strength of glue-film specimen (0) after exposure in QUVweatherometer for 200h (results averaged across stabilized specimens and unstabilizedcontrols). Unexposed controls (c) are shown for comparison. Y1 axis shows result on alog scale (natural logarithms, In) because results were analysed as logs. Y2 axis showsresults on a normal scale. The LSD bar is an error bar that can be used to comparemeans and is used in preference to other types of error bars3.3.2. Weight LossAdhesive type also had a significant (p<0.001) effect on loss of weight of glue-film specimens during accelerated weathering. The weight losses are expressed asratio of weight of specimens after weathering divided by weight of the same specimenbefore weathering (Weightratio = (-weightbef ore + weightafter)/weightbefore). Figure 3.8 andFigure 3.9 show the weight ratios of glue-film specimens after exposure in the Xenon-arc and QUV devices, respectively.Weight losses of EpE adhesive films exposed in the Xenon-arc weatherometerwere significantly (p<0.05) lower compared to the other three adhesive types. Therewas no significant difference (p>0.05), between the weight losses of the MF and EpUspecimens, but both lost significantly (p<0.05) more weight than the EPI glue filmspecimens (Figure 3.8).49^-0.08^ 0.92-0.10^ 0.900.00-0.02""eo -0.040' -0.06  ^   LS D=0.01561.000.980.960.9400)EPI^EpE^EpU^MFAdhesive typeFigure 3.8: Weight losses of glue film specimens after exposure in a Xenon-arcweatherometer for 200h (results averaged across stabilized and unstabilized controls),and expressed as natural logarithmsWeight losses of EpE and EPI adhesive films exposed in the QUVweatherometer were significantly (p<0.05) lower than the EpU and MF specimens.There was no significant difference in the respective weight losses of the EpE and EPIfilms and the EpU and MF films (Figure 3.9). -0.04  O- 0.96^CC.0, -0.08 - 0.92^C)-0.12 LSD=0.0419  - 0.89 -0.16 EPI EpEEpU MF 0.85Adhesive typeFigure 3.9: Weight losses of glue film specimens after exposure in a QUVweatherometer for 200h (results averaged across stabilized and unstabilized controls) ,and expressed as natural logarithmsIn general there were similarities in weight losses of samples exposed in the twodifferent types of weatherometers, but MF and EpU samples exposed in the QUVweatherometer lost more weight than those exposed in the Xenon-arc device (compareFigure 3.8 and Figure 3.9).500.04   LS D=0.0428 LSD=0.0179 ■s 0.00-0.04 -Zco -0.08-0.12 - -0.16 Control■ UVA/HALS11.041.00O0.96 V0.92 0)0.890.85The percentage losses in weight and tensile strength of glue film specimensexposed in both types of weatherometer are summarized in Table 3.4. Positive valuesrepresent weight or strength gains.Table 3.4: Percentage weight and tensile strength losses of glue film specimensexposed for 200h in a Xenon-arc or QUV weatherometerWeatherometer Factor Adhesive typeEPI EpE EpU MFXenon-arc Weight loss (%) 3 0.1 5.3 5.8Strength loss (%) 13.2 2.4 +9.5 10.6QUV Weight loss (%) 3.2 +0.4 11.8 12.8Strength loss (%) 44.6 8.6 +77.8 18.4There was a significant (p<0.05) interaction of adhesive type and stabilizer on theweight losses of specimens exposed in the QUV device (Figure 3.10).EPI^EpE^EpU^MFAdhesive typeFigure 3.10: Interaction of adhesive type and stabilizer on weight losses of glue filmspecimens exposed in a QUV device for 200hThe stabilizer had a significant effect at reducing weight loss of MF adhesivedog-bone samples, but the opposite was the case for the EpU adhesive. For EPI andEpE adhesives there was no significant difference in weight losses of the controls andthe stabilized glue film specimens. Both the stabilized and unstabilized MF and EpUadhesive samples lost significantly more weight than the EPI and EpE specimens.There were no significant (p>0.05) interactions of adhesive type and stabilizer on tensilestrength losses of specimens exposed in the QUV device or weight and tensile strengthlosses of specimens exposed in the Xenon-arc weatherometer.513.3.3 Color ChangeThe colors of the tested adhesives which were visible to the naked eye were: MF= yellow, EPI = light yellow, EpE = grey and EpU = transparent. These colors changedafter weathering. Adhesive type had a significant (p<0.001) effect on L* ratio, a* ratio, b*ratio and deltaE of glue film specimens exposed in the QUV. The ratios were calculatedby taking the value after weathering and dividing it by the value before exposure. Thelightness of the EPI, EpU and MF films decreased significantly when they were exposedin the QUV weatherometer but there was little change in the lightness of the EpEadhesive films. The lightness of all four adhesive types was significantly different fromeach other. There were significant changes in the redness of EpE, EPI and MF films butnot the EpU film. The redness of the EpE film was significantly different from that of theother three adhesive films. There was no significant difference in the redness of theEpU and EPI and the EPI and MF films. The yellowness or b-value of all four adhesivetypes changed significantly after exposure. There was no significant difference in the b-values of the EPI and EpU films after exposure. The EPI films experienced the greatestcolor-change, followed by MF, EpU and EpE films. The total color change of all fouradhesive types was significantly different from each other. The effect of adhesive typeon L* ratio, a* ratio, b* ratio and deltaE of adhesive films after exposure to QUV isshown in Figure 3.11.521. 0.80io0.700.600.500.40 EPIEpEAdhesive typeEpU^MF  LSD=0.8729   EPI^EpE^EpU^MFAdhesive type3.002.502.0001.501.000.500.001.80 1.60 - LSD=0.17242 EpEAdhesive typeEPI 0.80 -0.60 EpU^MFLSD=5.539000EPI^EpE^EpU^MFAdhesive type35.0030.0025.00w 20.00.13-o 15.0010.005.000.00Figure 3.11: Effect of adhesive type on L* ratio, a* ratio, b* ratio and deltaE afterexposure to QUV531.501.401.301.000.90The stabilizer significantly (p<0.001) reduced the yellowing of the glue filmsexposed in the QUV device and this is reflected in the lower b* ratio for stabilized filmscompared to unstabilized ones (Figure 3.12).ControlTreatmentLSD=0.0687^  UVA/HALS0.8022.0020.0018.00a 16.00-014.0012.0010.00•LSD=0.905UVA/HALS^ControlTreatmentFigure 3.12: Effect of stabilizer on b* ratio and deltaE of glue films after exposure toQUV (averaged across glue types)There was a significant (p<0.001) interaction of adhesive and stabilizer on L*ratio, a* ratio, b* ratio and deltaE of glue film specimens exposed in the QUV. Theseinteractions occurred because the photostabilizer restricted changes in color of the twoepoxy adhesives, but it had little effect in restricting color changes of MF and EPIadhesives (Figure 3.13). The stabilized and unstabilized EPI films showed the largestchanges in lightness during exposure followed by MF, EpU and EpE films. Thestabilized EpE and EpU showed no significant differences in their L-value duringexposure, whereas the other stabilized and unstabilized adhesive films weresignificantly different. A significant difference in the lightness of stabilized andunstabilized films was only observed for the two epoxide adhesives (EpE, EpU). Thestabilized and unstabilized MF and EPI films showed the biggest changes in a* duringexposure in the QUV device; followed by unstabilized EpU and EpE films. The redness54of the stabilized EpU and stabilized and unstabilized EpE films did not changesignificantly during exposure. The unstabilized EpE films showed a significant differencein redness compared to the other three unstabilized adhesive films, but EPI, EpU andMF control films were not significantly different from each other. The stabilizer wasparticularly effective at preventing the reddening of the EpE and EpU (LSD=0.4025, forcomparing means with the same level of adhesive) adhesive films, as mentioned above.The stabilizer was also very effective at preventing the yellowing of the epoxyadhesives. In contrast, stabilized EPI films showed slightly greater yellowing thanunstabilized specimens and there was little effect of the stabilizer on the yellowing of theMF adhesive films. EPI films showed the greatest total color change during exposure inthe QUV device followed by MF, unstabilized EpU, EpE, stabilized EpU and EpE films.The color change of the stabilized EpE and EpU films were significantly different fromeach other and they exhibited the smallest total color change. The stabilizer was onlysignificantly effective at reducing the total color change in the case of EpE and EpUfilms.551.201.00   ■0.80 I01 0.60 • UVA/HALS0.40  - ■ ControlI LSD=0.05100.20' LSD=0.03420.00 EPI EpEEpU MFAdhesive type3.002.50   ■■ ■2.001.501.00  LSD=0.89590.50  0.00 LSD=0.4025-0.50 EPI EpE^EpU^MFAdhesive type2.502.00 ■^LSD=0.1883 I LSD=0.13731.50 -  ■  la^1.00  0.50 -0.00 EPI EpE^EpU^MFAdhesive type35.00  30.0025.0020.00■■15.0010.00 ■5.00   0.00 EPI EpE EpUAdhesive typeI LSD=5.607LSD=1.811MFFigure 3.13: Effect of Adhesive/stabilizer on L* ratio, a* ratio, b* ratio and deltaE afterexposure to QUV. The smaller numbered LSD is for comparing means of the sameadhesive56Adhesive type had a significant (p<0.001) effect on L* ratio, b* ratio and deltaE ofglue film specimens exposed in the Xenon-arc weatherometer. The EPI films becamedarker during exposure in the Xenon-arc device and hence their L ratio was significantlylower than those of the other adhesives. Weathering had no significant effect on thelightness of the epoxy (EpE and EpU) and MF adhesive films and the lightness of theseadhesives were not significantly different from each other. All four adhesives becameyellow during exposure in the Xenon-arc device and hence there were significantchanges in their b-ratios. The EPI showed the greatest total color change, followed byMF, EpU and EpE adhesive films (Figure 3.14). The total color change of the EpE, EpUand MF adhesive films were not significantly different from each other.571.000.900.80   00.70  LSD=0.04430.600.500.40 EPI EpE EpU MFAdhesive type1.601.501.401.301.201.1012 1.000.900.800.700.60 EPI^EpE^EpU^MFAdhesive type30.00 ^25.0020.0015.0010.005.000.00LSD=4.471EPI EpE^EpUAdhesive typeMFFigure 3.14: Effect of adhesive type on L* ratio, b* ratio and deltaE of adhesive filmsafter exposure in a Xenon-arc weatherometerStabilizer had a significant (p<0.001) effect on L* ratio, b* ratio and deltaE of gluefilm specimens exposed in the Xenon-arc weatherometer. The stabilizer was effective atreducing the darkening (L*) and yellowing (b*) of the glue films during exposure.Accordingly, the total color change (deltaE) of stabilized samples during exposure wassignificantly smaller than that of unstabilized films. The effect of stabilizer on L* ratio, b*ratio and deltaE after exposure of films in the Xenon-arc weatherometer is shown inFigure 3.15.581.451.401.351.30 -1.25 LSD=0.02171.201.151.10  -1.051.00 —  UVA/HALS^ControlTreatment0.910.900.9000.89 -J0.89 -LSD=0.00490.880.88 UVA/HALS ControlTreatment15.00 -14.00 -13.00 -w 12.0011.0010.009.008.00 UVA/HALS^ControlTreatmentFigure 3.15: Effect of stabilizer on L* ratio, b* ratio and deltaE of adhesive films afterexposure in a Xenon-arc weatherometerThere was a significant adhesive/stabilizer (p<0.001) interaction on L* ratio, b*ratio, deltaE and a* ratio (p<0.01) of glue film specimens exposed in the Xenon-arcweatherometer. In terms of lightness the interaction occurred because the stabilizer waseffective at reducing the darkening of the EPI, EpE and EpU films, but it was noteffective with the MF films. In the case of a-value, the interaction occurred because thestabilizer was effective at reducing the reddening of the EpU films, but it was ineffectiveat reducing the reddening of the other adhesive types. The interaction of adhesive andstabilizer on the b-value occurred because the stabilizer reduced the yellowing of the59EpE and EpU films but it had no significant effect on the yellowing of the EPI and MFfilms. The interaction between an adhesive and stabilizer on total color change occurredbecause the stabilizer was effective at reducing the total color change of EpE and EpUfilms but it was ineffective at reducing the discoloration of the EPI and MF films. Theinteractive effect of adhesive type and stabilizer on L* ratio, a* ratio, b* ratio and deltaEafter exposure to Xenon-arc is shown in Figure 3.16.60•LSD=0.7996   LSD=0.6426EPI2.502.302.101.90O 1.70E 1.50la 1.301.100.900.700.50  EpE^EpU^MFAdhesive type2.001.801.601.000.800.60LSD=0.1481• I LSD=0.0433 •^LSD=0.0445^ ^LSD=0.0097  • UVA/HALS• Control  EPI^EpE^EpU^MFAdhesive type^EpE^EpU^MFAdhesive type• LSD=4.509  I LSD=1.224EPI^EpE^EpU^MFAdhesive typeFigure 3.16: Interaction of adhesive type and stabilizer on L* ratio, a* ratio, b* ratio anddeltaE of adhesive films after exposure in a Xenon-arc weatherometer. The smallernumbered LSD is for comparing means of the same adhesive30.0025.0020.0015.0010.005.000.00  61004481 ND17.5mm 10.08V x2.^80OmSE 28-806-0020-8.,48-08 004480 WD17.8mm 10.0W x1.5k 208mSE SE^28-Feb-08^004178 8018.3mm 10.0kV xl.5k 208m3.3.4 Structural Changes (SEM)A HITACHI S-2600N Scanning Electron Microscope was used to investigatestructural characteristics of different adhesive films before and after weathering. Thecross-sections where the dog-bone samples fractured were examined. The preparedsamples were coated with a 5 pm thick gold layer after they were glued on a SEM stud.Figure 3.17 — Figure 3.20 show SEM images of unstabilized MF, EpE, EPI and EpUspecimens before and after exposure in a QUV weathering device, respectively. Figure3.21 shows unstabilized MF samples after exposure in QUV and Xenon-arc weatheringdevices. Figure 3.22 shows unstabilized and stabilized EPI after exposure in a QUVweathering device.Figure 3.17: Unstabilized MF before (left) and after QUV exposure (right)Figure 3.18: Unstabilized EpE before (left) and after QUV exposure (right)62004483 01117.7an 10:011, xj...5k .2011T:SE 28-Feb-08 004484 11015.5am 10.0kV x1.^217;m:SE 08 -Feb-08SE^28-Feb-08^004479 )0117.7ma 10.01N SE^28-Fab-08^004477 $1017.Sam 10.0W x1.5k 201asFigure 3.19: Unstabilized EPI before (left) and after QUV exposure (right)•^•^•SE^19-1,46-08^0049,30 2. lam 10-0kV 2500 100um 28- -OE^tames 14818.01m4 1070W x510 100uaFigure 3.20: Unstabilized EpU before (left) and after QUV exposure (right)Figure 3.21: Unstabilized MF after QUV (left) and Xenon-arc exposure (right)63SE^20-Feb-08^004422 MD18,8ma 13.0kV x1.5k 20un SR^20 -Fob -OS^004423 MD17-0mm 13.0k^ xl.Sk 2011mFigure 3.22: Unstabilized (left) and stabilized (right) EpU after QUV exposureSEM images of adhesive films (Figure 3.17 — Figure 3.22) didn't show any majorchanges as a result of exposure in the QUV device (Figure 3.17 — Figure 3.20), orbetween QUV and Xenon-arc exposure (Figure 3.21). Furthermore, there was nodifference in the appearance of unstabilized and stabilized films (Figure 3.22).Nonetheless, each adhesive type had a characteristic appearance. The MF appeared tobe dense, flaky and rather brittle with some crystalline domains; EpE possessed acompact appearance with larger spots of higher crystallinity. The EPI showed a porousand spongy cross-sectional face and the EpU appeared to be homogeneous over itscross-section. The cloudy patches seen in the EpU image may be an artifact of the waythat samples were cut with a razor-blade.3.3.5 Chemical Changes (FTIR)The FTIR spectra of adhesive dog-bone samples were taken with a Perkin ElmerSpectrum One FTIR Spectrometer. The resolution was 16 cm -1 and 12 scans wereperformed for each sample and then averaged to produce a spectrum. The followingFTIR spectra are given in the range of 650 — 4000 cm -1 . Some of the main changes inthe functional group of the adhesives were observed in this region.Figure 3.23 — Figure 3.26 show spectra of unstabilized MF, EpE, EPI and EpU beforeand after exposure in a QUV device, respectively.640a0to.02926\fr173410343346\fr 813/1734^V \/ h\,\ A\LA10343346 2926\fr^ \fr r^8\fj^13\l/-\41VI4000 3600 3200 2800 2400 2000 1800 1600 1400 1200 1000 800 650Wavenumber cm'Figure 3.23: Unstabilized MF before (bottom) and after QUV exposure (top)A slight decrease of the peak at a wavenumber of 1034 cm -1 was observed in theMF adhesive after weathering in a QUV device, which may be caused by a loss of etherC-0 stretching. Furthermore, a slight increase in the 1737 cm -1 band occurred, whichmay be due to an increase of carbonyl groups. The peak at 2926 cm -1 (CH stretching)increased. The peak at 813 cm -1 (melamine triazine ring) decreased after acceleratedweathering. The band around 3300 cm -1 (hydroxyl group) stayed almost constant duringexposure (Figure 3.23).651653\1/ 16101041335133511734\l/16101653 ,1734 \I4945 83010418301945\173341 29262845\l/^227812391737^1373 \i/16421509 \l/1737 3341 2926\i/227813731642 15° ` 4, 12394000 3600 3200 2800 2400 2000 1800 1600 1400 1200 1000 800 650Wavenumber cm -1Figure 3.24: Unstabilized EpE before (bottom) and after QUV exposure (top)The spectra for unstabilized EpE showed an increase in the band at awavenumber of 3300 cm -1 (hydroxyl group) after exposure. The peak at 1734 cm -1(carbonyl group) increased, whereas, a peak at 1653 cm -1 (carbonyl group) appearedafter exposure. The peaks at 1610 (aromatic C=C), 1041 (C-0 group), and 830 cm-1(aromatic CH) also decreased significantly after exposure (Figure 3.24).4000^3600 3200 2800 2400 2000 1800 1600 1400 1200 1000 ^800 650Wavenumber cm -1Figure 3.25: Unstabilized EPI before (bottom) and after QUV exposure (top)•«--•0a_o0aco664,3331292629261'333116081653^\l/1726928AI.16081653 h I1726 \k/ ^I,i Ij^f1VIn the case of unstabilized EPI the absorption peak at 3300 cm -1 (hydroxyl group)stayed constant during exposure in the QUV device. Peaks at 2926 (trans isomer C=C)and 2845 cm -1 (trans isomer C=C) decreased slightly after exposure. The peaks at 1737(carbonyl group), 1373, and 1239 cm -1 (amide) decreased significantly and the peaks at2278 and 1509 cm -1 (amide) completely disappeared after exposure. A peak with awavenumber at 1642 cm -1 (carbonyl group) appeared after exposure (Figure 3.25).COO_o4000 3600 3200 2800 2400 2000 1800 1600 1400 1200 1000 800 650Wavenumber cm -1Figure 3.26: Unstabilized EpU before (bottom) and after QUV exposure (top)The spectra for unstabilized EpU showed a slight increase in the peak at a wavenumber3300 cm-1 (hydroxyl group) after exposure. The peak at 928 cm-1 (oxirane ring)decreased slightly. The peak at 1726 cm -1 (carbonyl group) slightly increased, and apeak at 1653 cm -1 appeared after exposure. The peaks at 1608 (aromatic C=C) and2926 cm-1 (CH2) decreased significantly after the glue films were exposed in the QUVdevice (Figure 3.26). Both epoxide adhesive types showed different characteristics. TheEpE was a grey color and rather brittle and the EpU was clear and very flexible.Nevertheless, the EpU and EpE adhesives have almost the same spectra (compareFigure 3.24 with Figure 3.26), probably because commercial formulations of epoxyadhesives are generally composed of a glycidyl ether epoxy type resin with anamide/amine blend hardener.672926^ 16531608 i\l/ ,3331^ Ai 928\l/3331928Figure 3.27 shows spectra of unstabilized and stabilized EpU films after exposure in theQUV device.4000 3600 3200 2800 2400 2000 1800 1600 1400 1200 1000 800 650WavenumberFigure 3.27: Unstabilized (bottom) and stabilized EpU after exposure to QUV (top)The spectra of the unstabilized EpU after exposure in the QUV device exhibiteda higher peak around 3300 cm -1 . Peaks at 1653, 1608 and 2926 cm -1 are notably highercompared to those in the spectrum of stabilized EPI (Figure 3.27). A peak at 928 cm -1stayed constant.0068Figure 3.28 shows spectra of unstabilized MF after exposure to QUV and Xenon-arc4000 3600 3200 2800 2400 2000 1800 1600 1400 1200 1000 800 650Wavenumber cm -1Figure 3.28: Unstabilized MF after exposure to QUV (top) and Xenon-arc (bottom)The peak at 1732 cm -1 has almost the same height after exposure of adhesive filmsin both types of weathering devices. The peaks at 1079 and 814 cm -1 are markedlysmaller after exposure in QUV compared to those of samples exposed in the Xenon-arcweatherometer (Figure 3.28).3.4 DiscussionIf the quantum energy present in UV radiation is similar to the dissociationenergy of atomic bonds, then bonds can break and a free-radical degradation processmay commence (Zeus Industrial Products 2006). Also light absorbing impurities(chromophores) are capable of generating free radical chain scission processes asmentioned in Chapter 2 (Turton and White 2001). Generally, photodegradation cancause reductions in the mechanical properties of polymers (Zeus Industrial Products2006, Turton and White 2001, Andrady, n.d.). Weathering in the Xenon-arc device,however, did not significantly reduce the tensile strength of the different adhesives.Furthermore, only EPI films lost tensile strength after exposure in the QUV device(Figure 3.5 and Figure 3.6). Some polymers are inherently resistant to UV induced69photodegradation, because of their high bond strength and the absence of UVabsorbing structures or impurities (Zeus Industrial Products 2006). The two epoxideadhesives and the MF seemed to have these characteristics. For both epoxy adhesivethere were increases in absorption bands at 3300 cm -1 (hydroxyl) following weatheringand, increases were observed for the EpE adhesive in the peaks at 1724 cm -1 and 1734cm-1 (carbonyl). These changes are indicative of photochemical aging of the polymer.As proposed by 011ier-Dureault and Gosse (1998) polymers which photo-chemically agein the presence of oxygen give rise to oxidation products such as alcohols,hydroperoxides, ketones, aldehydes, carboxylic acids, esters, peracids, peroxyesters,and anhydrides. The epoxy adhesive also showed a slight decrease in the etherabsorption in the epoxide ring (oxirane ring) around 945 cm -1 following weathering. Thismay indicate that the ring opened up. The major decreases of the characteristicaromatic bands (C=C at 1610 cm -1 and CH bands at around 3000 (only EpU) and 830cm-1 (only EpE)) also indicate loss of aromatic groups from the epoxy backbonestructure. The ether peak at 1042 cm-1 also decreased in the case of the EpE adhesiveafter weathering. It is assumed that this change resulted from chain cleavages at theoxygen atom. The appearance of ketones and amide stretching at 1660 cm 1 for bothepoxides are in good agreement with the photo-oxidative mechanism for epoxy curedwith aliphatic amines, proposed by Bellenger and Verdu (1983). Ketone formationresults from the secondary hydroxyl groups, and amide is generated from theabstraction of hydrogen from methylene groups (CH2) adjacent to the crosslink.Radicals and oxygen are required for both reactions. The appearance of a band near3300 cm-1 , assigned to NH stretching and in the same region as the hydroxyl groups,supports the conclusion of amide formation. As suggested by Rezig et al. (2006) adecrease in the peak at 2925 cm -1 (due to CH2 stretching), may be caused by mass lossof the cured material. On the other hand a decrease of both the 1510 cm -1 (benzenering stretching) and 1245 cm -1 (aryl-ether stretching) peaks signify chain scissionbecause these bands represent the chemical groups present in the main chains ofepoxy resins. This was not observed for both epoxy adhesives. During curing theepoxide ring opens up to produce either alcohol or ether units. Both of these chemicalgroups are hydrophilic (Comyn et al. 1979) and probably contribute to both theadhesion and water absorption associated with epoxides. Also the slight increase of the1740 cm-1 , 1434 cm -1 (EpE) and 1726 cm -1 (EpU) stretching bands might be due to the70presence of anhydride groups resulting from oxidation according to Le Huy and coworkers (1992).Two of the three thermosetting adhesives (EpE and MF) showed higher tensilestrength than the 'thermoplastic' EPI, as expected (Figure 3.6 and Figure 3.5).Thermoset adhesives possess a higher degree of chemical cross-linking, which resultsin a higher cohesion force between molecular chains. Pure thermoplastics develop nochemical bonds and only secondary bonds that hold the molecular chains together. Theformer exhibit higher bond energies compared to the latter (Schwarzer 2002). Theepoxide adhesive EpU did not break during tensile testing. After force relaxation thesample went back to its original size. This EpU adhesive was initially expected to be athermoset, but it behaved more like an elastomer, because it exhibited smooth, rubbery-elastic properties. The EpU dog-bone samples were transparent which suggested it hadan amorphous structure (Schwarzer 2002). SEM images of EpU cross-sections alsoindicated its homogenous/amorphous nature (Figure 3.20).It seems reasonable to assume, that high weight losses of adhesive films duringweathering would be associated with losses in tensile strength. In case of EpE thisassumption seemed to be correct. The MF adhesive films lost more weight than the EPIfilms; however, they lost less tensile strength than the EPI films (Table 3.4). It ispossible that the MF films weren't completely cured when their weights were taken andloss of water and formaldehyde could account for the relatively high weight losses of theMF films. The higher cross-linking of the MF molecules compared to EPI, may havemaintained a stronger structure resulting in less tensile strength losses of MF filmsduring exposure. The emulsion polymer isocyanate consisted of vinyl acetate and anisocyanate type hardener. There is not much information on the FTIR spectra of EPIadhesives, possibly because this adhesive type is relatively new. The band around3300 cm-1 is due to hydroxyl groups and stayed relatively constant or decreased slightlyduring exposure. The CH stretching bands at 2926 cm -1 and 2845 cm -1 represent thetrans isomer bands of vinyl and these decreased in height after weathering. A smallpeak at 2278 cm-1 was observed before exposure and this completely disappeared afterweathering. This band might be due to aliphatic nitrites (RCN) between the vinyl acetateand the isocyanate or due to an aliphatic isonitrile (RNC-), which is at the end of a notcompletely cross-linked isocyanate where the oxygen atoms dissociate (Nagle et al.2007). A decrease of the 1740 cm -1 band occurred after weathering. According to71Copuroglu and Sen (2004) this decrease is due to ester elimination and the formation ofacetic acid. Acetic acid is the main UV degradation product of vinyl acetate. Figure 3.29shows the evolution of acetic acid from vinyl acetate.Figure 3.29 has been removed due to copyright restrictionsFigure 3.29: Evolution of acetic acid from vinyl acetate (Copuroglu and Sen 2004)The formation of aldehydes (1737 cm -1 ) are thought to be associated with thedecrease of the ester peak in vinyl acetate (Copuroglu and Sen 2004). The appearanceof the peak at a wavenumber of 1642 cm -1 might be due to carboxyl groups (carboxylateion (RC00-)). Also the strong increase of the peak at a wavenumber of 1371 cm -1 mightbe due to an increase of carboxylic acid in the polymeric material. In this case the band1373 cm -1 showed a decrease. The amide 2 absorption band (1517 cm -1 ) and the amide3 absorption band (1236 cm -1 ) decreased following exposure of the EPI films in theQUV device. Both stem from isocyanate molecules in the polymeric matrix.Both natural biopolymer (e.g. wood fiber) and synthetic polymers undergo UVinduced discoloration, which is evidence for chemical changes in the polymer matrix(Andrady, n.d., Zeus Industrial Products 2006, Althof 1981). According to Pospisil et al.(2002) there two main contributors to the discoloration of polymers: (a) structuralinhomogeneities and impurities absorbing UV-VIS light and (b) products consumingstabilizers which arise from a reaction of stabilizers with alkylperoxy radicals. Asmentioned by Clough et al. (1996) the formation of conjugated structures is important inradiation-induced color formation. Generally it was observed, that adhesives with highertensile strength exhibited less discoloration, as seen in Figure 3.7 and Figure 3.11. EPIwas the adhesive which exhibited a significant reduction in tensile strength after QUVexposure, and it also showed the greatest color change (deltaE), compare Figure 3.7with 3.11. The other adhesives tested here showed far less discoloration than EPI, andlittle if any changes in tensile strength after exposure. On the other hand, Clough et al.(1996) investigated several polymers including epoxides, and they found thatsusceptibility to radiation-induced discoloration of polymers had little or no correlationwith radiation-induced degradation of their mechanical properties. In support of this72observation Althof (1981) stated that sample discoloration, in itself, may not be relatedto physical changes in the properties of an adhesive, rather it is more of an indication ofmodification of the chemical structure of the material. The observations of the behaviorof the EpE, EpU and MF film specimens during exposure accord with this statement(compare Figure 3.6 / 3.7 with Figure 3.11 / 3.14). Stark and Matuana (2006) mentionedthat cross-linking and chain scission are competitive mechanisms involved in the UVdegradation of polymers. Cross-linking has also been shown to be the preferredmechanism during accelerated weathering. Cross-linking and the formation ofconjugated structures during accelerated weathering can create coloredphotodegradation products whereas, mechanical changes would occur due to chainscission. It is possible that cross-linking rather than chain scission reactionspredominated in the EpE, EpU and MF specimens and this may account for theobservation that discoloration of the adhesives occurred with little loss of mechanical(tensile) properties. Discoloration of EpU due to weathering has been observedpreviously by Clough et al. (1996). They tested clear amine cured epoxy adhesive andfound that they became dark yellow during weathering.The QUV weatherometer was more destructive towards the adhesive samplesthan the Xenon-arc weatherometer, as indicated by the weight loss results for the MFand EpU samples (Figure 3.8 and Figure 3.9), and also by FTIR spectroscopy of MFsamples before and after weathering (Figure 3.28). The peak at 1732 cm -1 (carbonylgroups) for MF samples stayed nearly unchanged after exposure in both types ofweatherometers, but peaks at 1079 cm-1 (ether C-0 stretching) and 814 cm -1 (melaminetriazine ring) were markedly smaller after exposure in the QUV device, which isconsistent with the photo-oxidation mechanism for MF adhesives proposed by Lemaireand Siampiringue (1999). Furthermore, total color change (deltaE) of samples exposedin the QUV was higher compared to samples exposed in the Xenon-arc weatherometer,compare Figure 3.11 and 3.14. The major differences between the weathering cyclesused in the two weathering devices are described in detail (above). In brief, however,the QUV exposed samples to light from fluorescent tubes (340 nm) whereas light fromthe Xenon-arc lamp more closely simulated the characteristics of sunlight. Samplestested in the QUV weatherometer may also have experienced higher humidity andtemperatures than those exposed in the Xenon-arc device. The finding that adhesivesamples were more degraded in the QUV device than in the Xenon-arc weatherometer73may indicate that short wavelength light and the synergetic effects of heat and moistureare required to accelerate the degradation of adhesives. This observation is inagreement with a statement made by (Searle 2000, Kockott 1988), who mentioned thatthe synergistic effect of all weather factors is required for decisive degradation ofpolymeric materials.Scanning electron microscopy of samples before and after weathering wasineffective at discerning differences between weathered and unweathered samples.The UVA/HALS stabilizers was only effective at reducing discoloration of the EpEand EpU adhesives (Figure 3.13 and Figure 3.16). The EPI and MF adhesives were theones which showed the most discoloration, regardless of the addition of a stabilizer.FTIR spectroscopy showed a positive effect of adding stabilizer to EpU (Figure 3.27).Unstabilized EpU adhesive exhibited notably higher peaks at 3300 cm -1 (hydroxyl) and1653 cm -1 (carbonyl). A higher peak at 2926 cm -1 as shown in the unstabilized spectrais evidence for more complete degradation (above). The effects of the same stabilizeron different adhesive formulations varied and this indicates that different adhesives mayrequire specific and different stabilizer systems to protect them from photodegradation.Generally changes in the hydroxyl (-OH) (around 3300 cm -1 ) and carbonyl (C=0)regions (1600 — 1800 cm -1 ) are evidence for photo-oxidative degradation (Kaczmarek,Podgorski, and Bajer 2005, Nagle et al. 2007). Furthermore, it is mentioned byKaczmarek, Podgorski, and Bajer (2005) that the presence of carbonyl groups isevidence for impurities in the polymeric material. Samples which were subjected to UVlight exhibited yellowing, which is a consequence of the formation of unsaturatedgroups. These unsaturated groups can be, for example, vinyl, vinylidene, and trans-vinylene groups (Copuroglu and Sen 2004). All four of the adhesives that were testedhere are used each for specific purposes, as mentioned above. Based on the resultshere the EpU adhesive exhibited some promising characteristics, which make it apotential candidate for use in wood composites used outdoors, such as glulam beams.EpU is a clear type of adhesive which showed relatively little discoloration even after200 hours exposure in the QUV device. This lack of discoloration may be important incomposites, where a clear gluline is required. The EpU adhesive also maintained itshigh flexibility after weathering, whereas often polymers become brittle after exposure(Zeus Industrial Products 2005, Shangguan et al. 2006). A rather flexible adhesive ismore capable of accommodating wood swelling and shrinking. The capacity of the EpU74to bond wood exposed outdoors and the economics of using this adhesive need to beinvestigated further.3.5 ConclusionsThe aim of this study was to investigate the performance of four clear adhesivesexposed to two different artificial weathering devices, and also examine the potential ofa UVA/HALS additive to photostabilize them. The EpE showed the best tensile strengthresults followed by MF, EPI and EpU adhesive films. The EpE adhesive also exhibitedthe least discoloration, followed by EpU, MF and EPI adhesive films. The EpE alsoshowed the lowest weight loss. EPI showed the second lowest weight loss followed byMF and EpU. A stabilizer consisting of a UV absorber and hindered amine lightstabilizer was effective at reducing weight loss of some of the adhesives and it waseffective at reducing discoloration of the adhesives particularly the epoxy adhesives.QUV exposure degraded the MF and EpU adhesives more than Xenon-arc exposure.There was no direct correlation between tensile strength and weight loss. The EpEadhesive showed the best performance overall. Chemical changes still occurred withthe EpE adhesive, but they had little impact on its tensile strength and weight loss.754. Relationship between Wood and Adhesive Photo-degradation4.1. IntroductionChapter 3 examined the performance of four clear adhesives under artificialweather conditions. There was evidence of chemical changes and degradation of thefour adhesive types. One of the adhesives (EpU) was very resistant tophotodegradation. The performance of the MF adhesive was also promising, particularlywhen a stabilizer was added to it. The experimentation in Chapter 3 focused on thephotodegradation of the adhesive alone. The emphasis in this chapter is on thephotodegradation of wood and adhesive, as it might occur in a wood composite such asglulam exposed outdoors.Previous research has not extensively examined the relationship between photo-degradation of wood and adhesives in composites. As mentioned in Chapter 2, a studyconducted by Kiguchi and co workers (2006) suggested a relationship between thephotodegradation of wood fibre and polymer in wood plastic composites (WPC) (Kiguchiet al. 2006). Matuana, Kamdem, and Zhang (2001) concluded that wood fibers areeffective photo-sensitizers and their incorporation into a polymer like PVC acceleratesthe photodegradation of the polymer when the composite is exposed to UV irradiation.One method of testing whether wood photosensitizes polymers would be to examinechanges in the physical (roughness) and chemical properties (discoloration) at thewood-adhesive interface. Materials become rougher when they are exposed to light dueto uneven decomposition and loss of molecules/atoms near the exposed surface.Geretovszky et al. (2002) and Wu et al. (2001) both found that mean surface roughnessof polymeric films increased almost linearly with irradiation time. Therefore, it seemsreasonable to hypothesize that the surface roughness of polymers may be higher inareas where the adhesive photodegradation is accelerated by the photodegradation ofwood. Chemical changes which are reflected in discoloration may also be morepronounced at the wood/adhesive interface, for similar reasons.The aim of this research was to test this hypothesis by examining changes inroughness and color of wood-adhesive samples exposed to natural solar radiation. Theresults of this experiment are expected to improve the understanding of the relationship76between photodegradation of adhesives in contact with wood, which may lead to thedevelopment of wood composites (glulam) with improved durability.4.2. Materials and Methods4.2.1. Experimental Design and Statistical AnalysesAn experiment was designed to examine the effects of three factors: 1, adhesivetype (MF and EpU); 2, stabilizer and 3, the interaction of adhesive and photo-stabilizeron the photodegradation of adhesive dowels in intimate contact with wood. Changes incolor and surface roughness of dowels were examined to evaluate the extent ofphotodegradation of the adhesive.Analysis of variance was used to analyze the effects of adhesive type, stabilizer,and the interaction of stabilizer/adhesive on factors of interest (changes in color andsurface roughness). Computation was performed using Genstat 5 (Genstat 2000).Before the final analysis diagnostic checks were performed to determine whetherresults conformed to the assumptions of analysis variance, i.e. normality with constantvariance. Significant results are plotted graphically and bars representing the leastsignificant difference are included on graphs to facilitate comparison of means.Appendix 2 contains all data and statistical output.4.2.2. Sample PreparationBased on results in Chapter 3, two out of the four adhesive types were selectedfor this experiment. The MF and EpU were chosen because they exhibited reasonableresistance to photodegradation. Furthermore, it was possible to easily cast adhesivedowels from the MF and EpU adhesives whereas the EpE and EPI adhesive were moredifficult to cast into dowels. In addition the EpE adhesive was very resistant tophotodegradation and it might be difficult to observe the hypothesed phenomenon withsuch a "photostable" adhesive.The preparation of adhesives was similar to that described in Chapter 3. Theadhesive and adhesive-stabilizer mixtures were placed separately in 20 glass tubes (55mm in length and with 6 mm inner diameter) with one end of the glass tube closed up77with a paper-stopper. A glass tube was chosen, in order to reduce chemical reactionsbetween the adhesive mixture and the cast. The filled glass tubes were stored verticallyfor 12 hours at ambient temperature in a rack. Afterwards the glass was removed fromthe cured adhesive dowel by destroying the glass-tube carefully with light hammerstrokes. Then the 20 dowels were cut into half using an ordinary band saw. One half ofthe original dowel was allocated to UV exposure and the other acted as a control. Afterlabeling the 40 dowels with a permanent marker at their ends, they were stored in aconditioning room at 20 ± 1 t and 65 ± 5% relative humidity for a period of 7 days.Wooden samples, 70 (width) x 150 (length) x 15 mm (thickness) were cut fromAustralian radiata pine boards (Pinus radiata D. Don) using an Altendorf table saw.Radiata pine was chosen because it had very wide growth rings, allowing dowels to beinserted exclusively into earlywood. Both Browne and Simonson (1957) and Kataoka,Kiguchi, and Evans (2004) have shown that the penetration depth of light intoearlywood is deeper compared to latewood, because of the lower density of earlywood.Deeper penetration of light into wood will presumably enlarge the interface betweenwood and the adhesive dowel, which should make it easier to distinguish interfacialdegradation from that occurring in other parts of the dowel. Before holes were drilledinto the wooden samples, the exact diameter of the adhesive dowels was measuredusing an electronic caliper. According to these measurements an appropriately sizeddrill bit was selected (5 mm drill bit for MF dowels and 6 mm drill bit for EpU dowels).Four holes of 5 mm depth were drilled into the earlywood of each wooden sample. Twoof them with diameters of 5 mm were allocated at random to the MF dowels and theother two, which had diameters of 6 mm were used for the EpU dowels. The holes weredrilled using an ordinary manual drill press. Prior to exposure, the dowels were insertedtightly into the holes so that they protruded at least 5 mm above the surface of theboards. Figure 4.1 shows a sample board containing a dowel. Three regions can beidentified within each dowel; (1) Region 1 is fully exposed to sunlight; (2) Region 2 is theinterface between wood and adhesive where sunlight acts on both wood and adhesive;(3) Region 3 is a zone where sunlight does not act on the wood and adhesive. Fivewood-adhesive-dowel samples were exposed to solar radiation, and the remaining 5samples were stored in a conditioning room.781. UV + 022. UV + 02 + Wood3. 02 + Wood^ UVFigure 4.1: Sample board showing a dowel and interfaces between the dowel, woodand the atmosphereThe experimental sample was designed to create a good interface between thewood and adhesive dowel. An interface is described [according to ASTM Standard D907-00, ASTM International 2001] as "a region of finite dimension extending from apoint in the adherend where the local properties (chemical, physical, mechanical andmorphological) begin to change from the bulk properties of the adherend to a point inthe adhesive where the local properties are equal to the bulk properties of theadhesive". The bulk properties are the properties of one phase unaltered by the otherphase. The interfacial region has different chemical and physical attributes compared tothe bulk material (adhesive or adherend). Therefore, the interface properties criticallyinfluence the properties of the adhesive bond (Petrie 2007). The least clear failure zoneaccording to (Frihart 2005) is that occurring within the interface. An importantrequirement in order to observe the hypothesed phenomenon (above) is a tight fitbetween the adhesive dowels and the wooden samples. As mentioned by Frihart(2005), in order for an adhesive to chemically bond to a substrate, intimate contactbetween both components is required. Region 2, which is the interfacial zone, was onlyexamined using profileometry. This zone was too narrow to measure its color using aspectrophotometer.794.2.3. Natural WeatheringThe wood-adhesive-dowel samples were exposed to 195 sunshine hours at theCenter of Advanced Wood Processing in Vancouver/Canada from the first of July 2007to the 15th of September 2007 (Table 4.1). The samples were not exposed to any rain.Table 4.1: Dates when adhesive dowels were exposed to sunlight, and total sunshinehours and mean temperature during the exposure trialDateSunshinehoursMean temperature[C]2007-07-31 6.5 17.42007-08-01 6.5 18.92007-08-02 6.5 20.42007-08-03 6.5 18.62007-08-04 6.5 17.42007-08-05 6.5 18.62007-08-06 6.5 18.12007-08-07 6.5 16.72007-08-13 6.5 16.82007-08-14 6.5 17.82007-08-15 6.5 19.12007-08-16 6.5 18.42007-08-24 6.5 19.42007-08-25 6.5 17.12007-08-29 6.5 19.42007-08-30 6.5 20.42007-08-31 6.5 172007-09-01 6.5 15.32007-09-02 6.5 15.82007-09-05 6.5 16.92007-09-06 6.5 15.82007-09-07 6.5 14.92007-09-08 6.5 14.42007-09-09 6.5 16.82007-09-10 6.5 16.92007-09-11 6.5 17.42007-09-12 6.5 16.22007-09-13 6.5 15.42007-09-14 6.5 16.62007-09-15 6.5 15Total 195 Average 17.3804.2.4. Roughness MeasurementAn AltiSurf 500 profilometer was used to scan the adhesive-dowel surface inorder to obtain mean roughness values (Sa), which were analysed statistically, and alsoto visualize the surface profile at the interface between wood and adhesive that wasexposed to sunlight. A 9 mm 2 surface was scanned across the three regions of thedowels (1 through to 3). A 300 pm probe was used to scan samples and measure theirsurface roughness. Measurement of surface roughness was performed before and afterexposure of dowels to sunlight. The samples were placed on a sample-holder in a waythat ensured that the same surface was scanned before and after exposure. Theaverage roughness (Sa) was obtained for each region within the dowel, and thesoftware (Paper Map version 3.2.0) was used to create topographical maps of the areasthat were scanned.4.2.5. Colour MeasurementA Minolta spectrophotometer was used to measure the color of samples bothbefore and after exposure. Each sample was measured twice. Once in region 3 whichwas surrounded by wood and once at the top of the dowel that was directly exposed tonatural sunlight (region 1). The total color change deltaE occurring as a result ofexposure was calculated as described in Chapter 3. The ratio of color before and afterexposure was calculated and analysed statistically.814.3. ResultsThere were significant effects of adhesive type, stabilizer and interactions ofadhesive/stabilizer on L* ratio (lightness), b* ratio (yellowness), a* ratio (redness),deltaE and Sa ratio (roughness) of adhesive-wood samples after exposure to solarradiation which are summarized in Table 4.2.Table 4.2: Significant effects of and interactions between adhesive type and stabilizeron resistance of adhesive dowels to solar radiationResponse variable Experimental factorsAdhesive type^Stabilizer^Adhesive/StabilizerSa ratio (roughness) NS (p = 0.348) NS (p = 0.889)Exposed part of dowel (region 1)L* ratio (Lightness) NS (p = 0.120) NS (p = 0.777)b* ratio(blue — yellow) NS (p=0.795) NS (p=0.305)a* ratio (green — red) NS (p=0.071) ...DeltaE NS (p=0.334) NS (p=0.101) *Unexposed part of dowel (region 3)L* ratio (Lightness) NS (p=0.178) NS (p = 0.773) NS (p=0.829)B* ratio(blue — yellow) NS (p=0.286) NS (p=0.599) NS (p=0.277)A* ratio (green — red) NS (p=0.094 NS (p=0.061)DeltaE NS (p=0.249) NS (p=0.515) NS (p=0.502)= p < 0.05; — = p < 0.01; —* = p < 0.001; NS = not significant (p > 0.05)4.3.1 Roughness ChangesAdhesive type had no significant (p>0.05) effect on the roughness of adhesivedowels that were fully exposed to solar radiation, and there was no significant (p>0.05)interaction of adhesive and stabilizer on surface roughness. The stabilizer, however,had a significant (p<0.05) effect on the surface roughness of adhesive dowels exposedto solar radiation (Table 4.2). Figure 4.2 shows the roughness ratio of stabilized andunstabilized dowel samples after exposure. The unstabilized dowels were rougher thanthe stabilized ones after exposure, as expected.823.002.251.500.750.00-0.75 20.099.494.48 2ilM2.12^cn1.000.47Control^UVA/HALSFigure 4.2: Effect of stabilizer on surface roughness ratio of adhesive dowels afterexposure to sunlight, and expressed as natural logarithms4.3.2 Color ChangesAdhesive type had significant effects on the b* and a* ratio of the part of thedowels that were fully exposed to solar radiation, and the a* ratio of the part of thedowel that was exposed to light and wood. There were also significant interactionsbetween adhesive and stabilizer on L*, a* and deltaE of dowel sections that were fullyexposed to solar radiation. There were no significant (p>0.05) effects of stabilizer onany of the color parameters. The effect of adhesive type on b* ratio and a* ratio ofdowel sections that were fully exposed to solar radiation is shown inFigure 4.3. The exposed part of the EpU dowel showed significantly less yellowing thanthe exposed part of MF dowels, but the opposite was the case for a*, in which case theMF was significantly less prone to redening than EpU adhesive.8321.301. LSD=0.09471.051.000.950.90EpU MFAdhesive type1.501.40 -o 1.30- LSD=0.2093is 1.20 -1.101.00 EpU^ MFAdhesive typeFigure 4.3: Effect of adhesive type on b* ratio and a* ratio after full exposure of dowelsto solar radiationThe effect of adhesive type on a* ratio of dowels after exposure to solar radiationand wood (region 1) is shown in Figure 4.4. Adhesive type also had a significant(p<0.01) effect on the a* ratio of that part of the glue dowel samples which wassurrounded by wood (region 3). The unexposed part of the MF dowel showedsignificantly less redening than the unexposed part of EpU dowels.841.401.301.2001.10la1.000.900.80LSD=0.192EpU^ MFAdhesive typeFigure 4.4: Effect of adhesive type on a* ratio after exposure to solar radiation (region3)There was a significant interaction between adhesive and stabilizer on lightness.Unstabilized MF which was fully exposed to solar radiation showed a smaller change inlightness than unstabilized EpU. Stabilized and unstabilized EpU samples, which werefully exposed to solar radiation were not significantly (LSD=0.0498, for comparingmeans with the same level of adhesive) different from each other, in terms of theirlightness (Figure 4.5). The same trend was observed for fully exposed MF samples.Table 4.3 shows the lightness values of fully exposed unstabilized and stabilized EpUand MF samples.Table 4.3: Lightness/color values of fully exposed unstabilized and stabilized MF andEpU samplesControl UVA/HALSEpU 0.9617 0.9956MF 1.0297 0.9868There was also a significant interaction between adhesive and stabilizer on thea* ratio of samples because the stabilizer restricted the redening of EpU dowels thatwere fully exposed to solar radiation, but the opposite was the case for the MFspecimens. Also unstabilized EpU exhibited significantly (p<0.05) more redening thanunstabilized, fully exposed MF dowels. As for deltaE, there was a significant (p<0.05)interaction between adhesive and stabilizer because the stabilizer restricteddiscoloration of EpU dowels that were fully exposed to solar radiation, but the oppositewas observed for MF specimens. The interactions between adhesive and stabilizer onL* ratio, a* ratio and deltaE after full exposure of dowels to solar radiation are shown inFigure 4.5.851.041.02 o^1.00LSD=0.0498 ■ LS D=0.0473  Li^0.980.96   Control■ UVA/HALS0.94 EpU MFAdhesive type2.00  1.60 - LSD=0.32970^1.20 -  ■  la^0.800.40  ControlLSD=0.4259 ■ UVA/HALS0.00 EpU MFAdhesive type10.00 ■7.50 ■5.00 LSD=6.622 LSD=3.775 2.50 UVA/HALS■ Control0.00 ^EpU^ MFAdhesive typeFigure 4.5: Effect of adhesive/stabilizer on L* ratio, a* ratio and deltaE after exposure tosolar radiation86The discoloration of dowels in region 1 and 3 before and after exposure to solarradiation was visible to the naked eye. Figure 4.6 and Figure 4.7 show photographs ofunstabilized EpU and MF before and after exposure to solar radiation, respectively.Figure 4.6: Unstabilized EpU before (top) and after (bottom) exposure to solar radiationFigure 4.7: Unstabilized MF before (top) and after (bottom) exposure to solar radiation874.3.3 Morphological ChangesA confocal profilometer was used to visualize erosion along the length of eachdowel to determine whether degradation was greater at the interface where theadhesive was exposed to both wood and solar radiation (as hypothesed in theintroduction). Figure 4.8 and Figure 4.9 show profileometry plots for MF and EpUsurfaces (9 mm2) across the interface (region 2), respectively. Greater deterioration atthe interface would appear in the centre of the images. For both adhesive types thispattern of degradation was not observed.Figure 4.8: Unstabilized MF before (top) and after exposure (bottom)880:7,3774 Wn14.9mmFigure 4.9: Stabilized EpU before (top) and after exposure (bottom)A scanning electron microscope (HITACHI S-2600N) was used to investigatestructural characteristics at the surface of the unstabilized adhesive dowels afterexposure. Figure 4.10 and Figure 4.11 show SEM images of unstabilized MF and EpUdowels after exposure to solar radiation, respectively. Besides a few scratches on thesurface of the adhesive dowels, probably resulting from removing the glass tubes, noevidence for distinct erosion in region 2 was found.Figure 4.10: Unstabilized MF after exposure to sunlight. Complete dowel (left), interfaceregion (right)89SS 09 -Nom -07 003785 WO 8.7mm 8.00WV m150 200umFigure 4.11: Unstabilized EpU after exposure to sunlight. Complete dowel (left),interface region (right)4.4 DiscussionThere were no significant changes in the color (deltaE) of the part of theadhesive dowel that was fully inserted into wood (region 3). In contrast, the part thatwas exposed to light (region 1) showed significant color changes, as expected. Nearlyall exposed adhesive dowels also showed visible color changes a few millimeters belowthe surface of the wood. This is probably due to penetration of light through earlywoodcausing chemical changes and discoloration. Nevertheless, results didn't appear tosupport the hypothesis that wood catalyzed the photodegradation of the adhesives.Severe degradation resulting in increased interfacial surface roughness was notobserved with any adhesive specimens exposed to 200 hours of sunlight (Figure 4.2). Itis possible that the duration of exposure (200 hours) of the samples to solar radiationwas not long enough to observe the hypothesed phenomena. However, the exposureperiod couldn't have been easily extended, due to seasonal changes in the weathertowards the end of the trial resulting in longer and more frequent rain periods. Otherresearchers have exposed polymer samples in their weathering experiments for morethan 1000 hours (Stark and Matuana 2006), but Decker, Zahouily, and Valet (2001)reported that distinct changes in IR spectra can be observed after only 200 h of QUVexposure. FTIR analyses of adhesive dowels using a Perkin Elmer Spectrum One FTIRSpectrometer, didn't give any meaningful results, which was probably due to thedifficulty of making a FTIR measurements on a cylindrical shaped adhesive dowel.90It is known that the use of a benzotriazole UV absorber in combination with ahindered amine light stabilizer is capable of protecting many polymer compositions(Renz et al. 2001). The stabilizer in this experiment was effective at restricting thediscoloration of EpU samples from photodegradation, but not the MF adhesive dowels.This accords with the results in Chapter 3, and suggests that other types of stabilizersmight need to be added to MF adhesives to increase their photo-resistance. Clough etal. (1996) mentioned that stabilizer formulations which protect polymeric materialsagainst discoloration need to be individually optimized in order to obtain the lowestpossible extent of discoloration for each polymer. Therefore, findings here are notsurprising. Other additives beside the UVA/HALS stabilizer weren't added to any of theadhesives tested here. Impurities from processing however can be responsible forabsorbing UV light and thus causing severe discoloration of polymers (Searle 2000) It ispossible that the light stabilizer used here was not able to absorb/scavenge all of thelight and free radicals that caused discoloration of the MF adhesiveKaci et al. (1999) mentioned that the HALS, Tinuvin 783, was effective atrestricting the rates of formation of ketones in low density polyethylene (LDPE), whichare involved in photochemical reactions due to photon absorption by its carbonyl group(Kaci et al. 1999). Stabilized MF adhesives showed more chemical modifications uponQUV aging than urethane adhesive (Decker, Zahouily, and Valet 2001). According tothese researchers (Decker, Zahouily, and Valet 2001) ether cross-links in MF adhesivesare sensitive to photodegradation, which is seen as a decreased ether peak around1070 cm -1 (see Chapter 3). Another light stabilizer product needs to be found whichimproves the photostability of MF adhesive.As described earlier in this chapter, EpU and MF dowels were labeled from thebottom with a permanent marker. The label on EpU samples was bleached afterexposure. That was not observed with MF labels. Such bleaching may be attributed tosolar radiation because unexposed EpU dowels inserted in wood and stored in aconditioning room didn't exhibit bleaching of labels. The transparency of EpU seems tobe responsible for solar radiation reaching all the way through to the bottom of thedowel. This bleaching effect was observed in the previous experiment in Chapter 3where samples were exposed to UV light. If UV light is capable of penetrating throughthe dowel, then it is apparent that the UV absorber that was added to the dowels wasnot able to fully protect the clear EpU adhesive. Hence it might be desirable to pigment91glues used in wood composites exposed outdoors to match the color of wood ratherthan use clear glue lines.4.5 ConclusionsThe hypothesis that adhesive photo-degradation is catalyzed by woodphotodegradation, could not be supported by the experiment performed in this Chapter.Profileometry didn't reveal erosion of the adhesive at the interface where dowelsamples were exposed to both sunlight and wood. This finding might be due to lack ofintimate contact between wood and adhesive at a molecular level. Alternatively, theexposure time may have been too short for the hypothesed effect to have occurred. Thestabilizer was effective at restricting discoloration of EpU dowel samples, but it was lesseffective with the MF dowel samples.925. General Discussion and Conclusions5.1. Discussion and Suggestions for Further ResearchIn this thesis it was hypothesed that photodegradation of adhesives is catalyzedby photodegradation of wood. Initial experimentation in Chapter 3 was performed toevaluate the performance of adhesives under artificial weather conditions; these resultswere also used to select adhesives for the second experiment (Chapter 4), whichattempted to answer the hypothesis. The stabilized and unstabilized adhesive dowelsdidn't show distinct surface erosion in the region where they were in contact with woodand exposed to solar radiation for 200 hours. Hence, the results in Chapter 4 don'tsupport the hypothesis that adhesive photodegradation is catalyzed byphotodegradation of wood. Nonetheless, arguments were made in Chapter 4 whichindicate that the experimental results may have been due to lack of intimate contact (atthe molecular level) between adhesive and wood or due to the limited exposure time.Therefore, further refinement of the experimental methodology would be desirable totest the hypothesis advanced in this thesis.The experiments performed in Chapter 3 to test the photostability of differentadhesive types and the photostabilization of adhesives with UVA/HALS additive yieldedinteresting results. The EpE adhesive was outstanding in terms of weight loss, tensilestrength performance and lack of discoloration. The stabilizer was effective at restrictingdiscoloration of EpE specimens, but it did not affect weight loss of samples exposed ina QUV weatherometer. Therefore it would appear that chemical changes in EpEsamples, resulting in discoloration, did not cause significant physical degradation of theadhesive. MF adhesives are less durable outdoors than resorcinol formaldehydeadhesives (Petrie 2007), but the MF adhesive showed the 2 nd best strength results.However, the MF adhesive had the highest weight loss, and second largestdiscoloration. The stabilizer was ineffective at restricting discoloration of the MFadhesive, but effective at reducing weight losses. The thermoplastic/thermoset EPIadhesive showed relatively low weight loss (2 nd best overall) and the largestdiscoloration, and strength losses. A water-based polymer isocyanate adhesive used forplywood has shown bond durability that is comparable to phenolic adhesives (Yoshida1986). However, after 36 months of natural weathering the water-based polymer93isocyanate lost most of its initial bond quality (Yoshida 1986). The EPI appears to beless suitable for glulam exposed outdoors than the other adhesives tested here. Forglulam applications structural adhesives are usually required with high strengthproperties and performance levels that do not change significantly with moderate aging(Petrie 2007). As mentioned by Petrie (2007) structural adhesives are generallythermosetting types with a higher degree of chemical cross-linking. Mynott and Van derStraeten (1984) pointed out that resorcinol formaldehyde and resorcinol phenolformaldehyde adhesives are currently dominating the field of structural wood gluing.They have been shown to be at least as durable as the wood itself under exteriorexposure conditions (Petrie 2007). Epoxides have been used for some specializedapplications such as metal-to-wood bonds (Petrie 2007). According to Frihart (2005)they are less commonly used in wood bonding because they cost more than most woodadhesives. The clear EpU adhesive tested here exhibited characteristics, which make ita potential candidate for use in wood composites used outdoors, such as glulambeams. The EpU showed relatively little discoloration after 200 hours exposure in theQUV device. This lack of discoloration may be important in composites, where a cleargluline is required. The EpU adhesive was unusually elastic which doesn't allow itsstrength characteristics to be compared with those of the other three adhesives testedhere. A rather flexible adhesive is more capable of accommodating wood's dimensionalchanges, whereas other adhesives often become brittle after exposure (Zeus IndustrialProducts 2005, Shangguan et al. 2006). Mynott and Van der Straeten (1984) found thatthe shear strength of epoxide bonded joints in glulam beams didn't decreasesignificantly after 10 years of weathering, but nevertheless the capability of the EpUadhesive to bond wood exposed outdoors will have to be investigated separately. Alsothe economics of using EpU adhesive for glulam beams will have to be studied. Basedon the results described in this thesis all adhesives possessed some interestingproperties for gluelam beams exposed outdoors, but the EpE and EpU appeared tohave the greatest potential.Results in both Chapter 3 and Chapter 4 revealed that stabilized EpU showed farless discoloration than stabilized MF samples. According to Dorn and Breuel (1992) theperformance of specific adhesive types tested under certain weather conditions are notnecessarily applicable to other polymers. For example, Turton and White (2001)investigated the photodegradation of two stabilized polypropylenes, and found94surprising differences in their tensile strength performance after weathering. Decker,Zahouily, and Valet (2001) found that stabilized urethane adhesive showed lessdiscoloration than stabilized melamine adhesive exposed under the same conditions.The discoloration of EpU was greatly restricted by adding a UV stabilizer (UVA andHALS) to the adhesive; whereas the same stabilizer formulation was not effective atreducing the discoloration of the MF adhesive. Some stabilizers are also known to affectthe color of polymers (Pospisil et al. 2002). Phenolic stabilizers are colorless in theiroriginal form and their architecture combines intrinsic activity with physical persistence(Pospisil 2002). Pospisil (2002) mentioned that polymer discoloration can be due to theformation of quinone methides, which can be formed due to trapping of alkylperoxyradicals by phenolic stabilizers. The type and number of radicals created in a polymerduring exposure are quite specific and may vary due to different chromophore typesand amounts in the polymer and also different chemical compositions of the adhesivebackbone. HALS protect polymers by free radical scavenging, and their stabilizationreactions are not completely understood (Rabek 1987, Horsey 1994). It is possible thatdifferences in the discoloration of stabilized EpU and MF adhesives during exposure tolight might be related to the effectiveness of the HALS at scavenging radicals in the twoadhesives. Hence, further research to examine this would be desirable. Furthermore itwould be beneficial to determine the best stabilizer formulation for the differentadhesives. As glulam beams gain greater importance for outdoor use, this questionsmay become increasingly important.Stabilized MF samples experienced less discoloration when exposed only tosolar radiation (Chapter 4) compared to samples exposed to light and moisture in aQUV weatherometer. The synergistic effect of moisture and UV light in degradingpolymers is well documented in the literature (Searle 2000, Kockott 1989). Moisturechanges in polymers can cause swelling and shrinking. Solar radiation can causeembrittlement of the material surface, which means that the tendency of the polymer tocrack under swelling/shrinking stresses is further promoted. Subsequently, crackspromote deeper penetration of light into the polymeric material, resulting in severepolymer degradation (Kockott 1989). Furthermore, Kockott (1989) pointed out asynergistic effect of short-wave radiation, water, and air-oxygen on the discoloration ofpolymers such as titanium-dioxide pigmented coatings. Therefore findings here accordwith results in the literature.955.2. ConclusionsThe EpE adhesive showed the best overall performance after exposure to 200 hours toartificial weathering. A stabilizer (UVA/HALS) was effective at reducing the weight lossof the MF adhesive and was effective at reducing discoloration of the epoxy adhesives.The stabilizer was ineffective at restricting weight losses of EpE, EpU and EPIadhesives and not effective at reducing discoloration in the cases of MF and EPIadhesives. Therefore, it can be concluded that the stabilizer formulation needs to betailored to suit different adhesives. QUV exposure degraded the MF and EpU adhesivesmore than Xenon-arc exposure. The synergistic effect of moisture and UV radiation onthe degradation of adhesives may account for this observation. Profileometry didn'treveal more distinct erosion of the adhesive at the interface where dowel samples wereexposed to both sunlight and wood, as hypothesed. A lack of intimate contact betweenwood and adhesive at a molecular level or the limited exposure time may beresponsible for this observation. Based on these findings it can be concluded that EpEand EpU might be good candidates for glulam beams exposed outdoors, and MFadhesive only if it is stabilized.966. ReferencesAkmal, N., A. M. Usmani. 2000. Medical polymers and diagnostic reagents.Chap. 13 in Handbook of Polymer Degradation. 2nd ed., revised and expanded,edited by S. H. Hamid. New York: Marcel Dekker.Althof, W. 1981. The influence of moisture on adhesive bonded joints. In Adhesion5, edited by K. W. Allen. London: Applied Science Publishers Ltd.Amstock, J. S. 2001. Introduction to sealants and adhesives. In Handbook ofAdhesives and Sealants in Construction. New York: McGraw - Hill.Andrady, A. L., S. H. Hamid, X. Hu, A. Torikai. n.d. Effects of increased solar ultravioletradiation on materials. Accessed 08th February 2008. Available from:http://www.gcrio.org/UNEP1998/UNEP98p62.html . Internet.ASTM. 1993. Standard test for calculation of color differences from instrumentallymeasured color coordinates. In ASTM D2244. Philadelphia: American Society ofTesting and Materials.ASTM. 2001. Standard terminology of adhesives. In ASTM D907 - 00. Philadelphia:American Society of Testing and Materials.Bellenger, V., J. Verdu. 1983. Photooxidation of amine crosslinked epoxies. II: Influenceof structure. Journal of Applied Polymer Science 28(9):2677 - 2688.Berglund, L., R. M. Rowell. 2005. Wood composites. Chap. 10 in Handbook ofWood Chemistry and Wood Composites, edited by R. M. Rowell. New York.Berner, G., M. Rembold. 1981. Aspects of new developments in light stabilizers forindustrial paints. Farbe and Lack 87(11):930 — 935.Bogusch, W., A. Seidel. 2000. Das Expo Dach in Hannover. Informationsdienst Holz,August, 1 — 7.Borroff, E. M., W. C. Wake. 1949. Journal of Trans Institute of the Rubber Industry25:190 - 210.Browne, F. L., H. C. Simonson. 1957. The penetration of light into wood. Forest ProductJournal7: 308 - 314.Burger, N. 2000. EXPO 2000 - Holzdach des deutschen Pavillions. Accessed 22ndJanuary 2008. Available from: http://www.burger-landshut.de/projekt_expo2000.html .Internet.97Carlsson, D. J., S. Chmela. 1990. Polymers and high-energy irradiation: Degradationand stabilization. Chap 4 in Mechanisms of Polymer Degradation and Stabilization,edited by G. Scott. London and New York: Elsevier.Chugg, W. A. 1964. Adhesion - General information. Chap. 9 in Glulam; the Theory andPractice of the Manufacture of Glued laminated Timber Structures. London: Benn.Clough, R. L., K. T. Gillen, G. M. Malone, J. S. Wallace. 1996. Color formation inirradiated polymers. Radiation Physics and Chemistry 48(5):583 - 594.Comyn, J. 1998. Degradation of adhesives. Plastics, Rubber and CompositesProcessing and Applications 27(3):110 - 115.Comyn, J., D. M. Brewis, R. J. A. Shalash, J. L. Tegg. 1979. The effect of water onepoxide resins and joints. In Adhesion 3, edited by K. W. Allen. London: AppliedScience Publishers Ltd.Copuroglu, M., M. Sen. 2005. A comparative study of UV aging characteristics ofpoly(ethylene-co-vinyl acetate) and poly(ethylene-co-vinyl acetate)/carbon blackmixture. Polymers for Advanced Technologies 16(1):61 - 66.Copuroglu, M., M. Sen. 2004. A comparative study of thermal ageing characteristics ofpoly(ethylene-co-vinyl acetate) and poly(ethylene-co-vinyl acetate)/carbon blackmixture. Polymers for Advanced Technologies 15(7):393 — 399.Cotter, J. L. 1977. The durability of structural adhesives. In Developments inAdhesives, edited by W. C. Wake. London: Applied Science Publishers.Davis, G. D., D. K. Shaffer. 1994. Durability of adhesive jionts. In Handbook ofAdhesive Technology, edited by A. Pizzi, K. L. Mittal. New York: M. Dekker.Decker, C. 2001. Photostabilization of polyvinylchloride by protective coatings. Journalof Vinyl & Additive Technology 7(4):235 - 243.Decker, C., K. Zahouily, A. Valet. 2001. Curing and photostabilization of thermoset andphotoset acrylate polymers. Macromolecular Materials and Engineering 286(1):5 —16.Deppe, H. - J., K. Schmidt. 1987. Zum Sicherheitsaspekt bei Brettschichthoelzern. Holzals Roh-und Werkstoff 45(6):255 - 256.Derbyshire, H., E. R. Miller. 1981. Photodegradation of wood during solar radiation.Holz als Roh-und Werkstoff 39(8):341 - 350.Dorn, L., G. Breuel. 1992. Alterungsverhalten von Klebverbindungen - Einfluss vonKlebstoff, Klima und Belastung. Blech Rohre Profile 39(9):641 - 644.Dowling, A. P. 1945. Wood aircraft assembly glues. Modern Plastics:156 - 158 and 192- 193.98Dynea. n.d. Emulsion Polymer Isocyanate (EPI) adhesive. Accessed 13rd June 2007.Available from: http://www.dynea.com/customers.asp?item_id=15 . Internet.Earhart, N. J., A. Patel, G. Knobloch. 1994. Thermal stabilization of adhesives. InHandbook of Adhesive Technology, edited by A. Pizzi, K. L. Mittal. New York: M.Dekker.Evans. P., M. J. Chowdhury, B. Mathews, K. Schmalzl, S. Ayer, M. Kiguchi, Y. Kataoka.2005. Weathering and surface protection of wood. Chap. 14 in Handbook ofEnvironmental Degradation of Materials, edited by M. Kutz. New York: WilliamAndrew Publishing.Feist, W. C. 1990. Outdoor wood weathering and protection. Chap. 11 in ArchaeologicalWood: Properties, Chemistry, and Preservation, edited by R. M. Rowell, R. J.Barbour. Washington DC: AmericanChemical Society.Feist, W. C., D. N. - S. Hon. 1984. Chemistry of weathering and protection. Chap. 11 inThe Chemistry of Solid Wood (Advances in Chemistry Series), edited by R. M.Rowell. American Chemical Society.Feldman, D., M. Khoury. 1988. Epoxy-lignin polyblends. Part II. Adhesive behavior.Journal of Adhesion Science and Technology 2(2):107 - 116.Filbey, J. A., J. P. Wightman. 1988. Factors affecting the durability of titanium/epoxybonds. In Adhesion 12, edited by K. W. Allen. London: Applied Science PublishersLtd.Frangi, A., M. Fontana, A. Mischler. 2004. Shear behaviour of bond lines in gluedlaminated timber beams at high temperatures. Journal of Wood Science andTechnology 38(2):119 - 126.Frihart, C. R. 2005. Wood adhesion and adhesives. Chap. 9 in Handbook of WoodChemistry and Wood Composites, edited by R. M. Rowell. New York.Garnier, S., A. Pizzi, 0. C. Vorster, L. Halasz. 2001. Rheology of polyflavonoid tannin-formaldehyde reactions before and after gelling. I. Methods. Journal of AppliedPolymer Science 86(4):852 - 863.Geretovszky, Zs., B. Hopp, I. Bertoti, I. W. Boyd. 2002. Photodegradation ofpolycarbonate under narrow band irradiation at 172 nm. Applied Surface Science186(1 - 4):85 - 90.Gierer, J., S. Y. Lin. 1972. Photodegradation of lignin. A contribution to the mechanismof chromophore formation. Papperstidning-Nordisk Cellulosa 75(7):233.Goulding, T. M. 1994. Epoxy resin adhesives. In Handbook of Adhesive Technology,edited by A. Pizzi, K. L. Mittal. New York: M. Dekker.99Grassie, N. 1975. A review of present trends in polymer degradation. Chap. 1 inDegradation and Stabilization of Polymers, edited by G. Geuskens. London: AppliedScience Publishers.Harrod, J. F. 1963. Hydrogen bonding in amine-epoxide adducts. Journal of PolymerScience Part A, General Papers 1(1):385 - 391.Heitner, C., J. C. Scaiano. 1993. Light induced yellowing of wood containing papers. InPhotochemistry of Lignocellulosic Materials, ACS Symposium Series, 531. AmericanChemical Society.Hofferber, B. M., E. Kolodka, B. Rishawn, R. J. Moon, C. R. Frihart. 2006. Effects ofswelling forces on the durability of wood adhesive bonds. Paper presented inProceedings of the 29th annual meeting of The Adhesion Society, Inc., Jacksonville,Florida, 19 - 22 February.Hon, D. N. - S. 1994. Analysis of adhesives. In Handbook of Adhesive Technology,edited by A. Pizzi, K. L. Mittal. New York: M. Dekker.Hon, D. N. - S.. 1991. Photochemistry of wood. In Wood and Cellulosic Chemistry,edited by D. N. - S. Hon, N. Shiraishi. New York: Marcel Dekker.Hon, D. N. - S.. 1984. Photochemical degradation of lignocellulosic materials. Chap. 8in Developments in Polymer Degradation, edited by N. Grassie. London and NewYork: Applied Science Publishers.Hon, D. N. - S., W. C. Feist. 1981. Free radical formation in wood: The role of water.Wood Science 14(1):41 - 47.Horsey, D. 1994. Ultraviolet stabilization of adhesives. In Handbook of AdhesiveTechnology, edited by A. Pizzi, K. L. Mittal. New York: M. Dekker.Janberg, N. n.d. Expo Roof. Accessed 26th March 2007. Available from:http://en.structurae.de/structures/data/index.cfm?ID=s0000918 . Internet.Jordan, R. 1989. Alterungsschutz bei Klebstoffen. Seifen, Oele, Fette, Wachse115(19):714 - 717.Kaci, M., T. Sadoun, S. Cimmino. 2000. HALS stabilization of LDPE films used inagricultural applications. Journal of Macromolecular Materials and Engineering278(1):36 - 42.Kaczmarek, H., A. Podgorski, K. Bajer. 2005. Photochemical reactions in poly(vinylchloride)/poly(vinyl alcohol) blends. Journal of Photochemistry and Photobiology17(2):187 - 195.100Kaempf, G., H. - G. Fitzky, R. Holm, H. Schmid, S. Storp, J. Wagner, T. Werner.1984. Untersuchung der Abbauvorgaenge bei der Bewitterung von Polymeren mitphysikalischen Methoden. Ressort Technische Entwicklung / Angewandte Physikdes Zentralbereiches Zentrale Forschung der Bayer AG, Leverkusen.Kalnins, M. A. 1966. Surface characteristics of wood as they affect the durability offinishes. Part 2: Photochemical degradation of wood. U.S. Forest ProductsLaboratory Report 57:23 — 60.Kataoka, Y., M. Kiguchi, P. D. Evans. 2004. Photodegradation depth profile andpenetration of light in Japanese cedar earlywood (Cryptomeria japonica D. Don)exposed to artificial solar radiation. Surface Coatings International Part B: CoatingsTransactions 87(3):187 - 193.Keene, L., C. R. Clayton, G. Halada. 2001. Novel techniques for the investigation oflong-term photodegradation of multi-layer polymer coatings. Electrochemical Society- Proceedings:197 - 206.Keimel, F. A. 1994. Historical developments of adhesives and adhesive bonding. InHandbook of Adhesive Technology, edited by A. Pizzi, K. L. Mittal. New York: M.Dekker.Kerr, C., N. C. MacDonald, S. Orman. 1970. Effect of hostile environments on adhesivejoints. British Polymer Journal 2(1 - 2):67 - 70.Kiguchi, M., Y. Kataoka, H. Matsunaga, T. Nishimura, H. Matsui, K. Yamamoto, P. D.Evans. 2006. Weathering performance of wood-fibre plastic composites made fromrecycled materials. Paper presented at 2nd International Conference onEnvironmentally - Compatible Forest Products, Porto, Portugal, September.Kiguchi, M., Y. Kataoka, H. Kaneiwa, K. Akita, P. D. Evans. 2000. Photostabilisation ofwoodfibre-plastic composites by chemical modification of woodfibre. Paperpresented at 5th Pacific Rim Bio - Based Composites Symposium, Canberra,Australia, 10 - 13 December.Klemchuk, P. P., M. E. Gande. 1988. Stabilization mechanisms of hindered amines.Polymer Degradation and Stability 22(3):241 - 274.Kockott, D. 1988. Natural and artificial weathering of polymers. Polymer Degradationand Stability 25(2 - 4):181 - 208.Kollek, H. 1988. Chemistry of phenolic and epoxy adhesives. In Adhesion 12, edited byK. W. Allen. London: Applied Science Publishers Ltd.Kringstad, K. P., S. Y. Lin. 1970. Mechanisms in the yellowing of high yield pulps bylight: structure and reactivity of free radical intermediates in the photodegradation oflignin. Tappi Journal 53:2296.101Krystofiak, T., S. Proszyk, J. Mariusz. 2003. Studies of some properties of EPIadhesives. Annals of Warsaw Agricultural University - SGGW, Forestry and WoodTechnology 53:214 - 217.Leary, G. J. 1994. Recent progress in understanding and inhibiting the light-inducedyellowing of mechanical pulps. Journal of Pulp and Paper Science 20(6):154 — 160Le Huy, H. M., V. Bellenger, M. Paris, J. Verdu. 1992. Thermal oxidation of anhydridecured epoxies II - Depth distribution of oxidation products. Polymer Degradation andStability 35(2):171 - 179.Lemaire, J., N. Siampiringue. 1999. Prediction of coating lifetime based on FTIRmicrospectrophotometric analysis of chemical evolutions. Chap. 16 in Service LifePrediction of Organic Coatings: A Systems Approach, edited by D. R. Bauer, J. W.Martin. Washington DC: American Chemical Society.Lisperguer, J. H., P. H. Becker. 2005. Strength and durability of phenol-resorcinol-formaldehyde bonds to CCA-treated radiata pine wood. Forest Products Journal55(12):113 - 116.Matuana, L. M., D. P. Kamdem, J. Zhang. 2001. Photoaging and stabilization of rigidPVC/Wood-fiber composites. Journal of Applied Polymer Science 80(11):943 - 1950.McBain, J. W., D. G. Hopkins. 1925. On adhesives and adhesive action. Journal ofPhysical Chemistry 29(2):188 - 204.Meierhofer, U. A. 1988. Witterungsverhalten von impraegniertem Brettschichtholz nachfuenfjaehriger Freibewitterung. Holz als Roh-und Werkstoff 46(2):53 - 58.Meierhofer, U. A. 1986. Witterungsverhalten von salz-und oel-impraegniertemBrettschichtholz. Holz als Roh-und Werkstoff 44(5):173 - 177.Moloney, A. C., D. M. Brewis, J. Comyn, B. C. Cope. 1981. The effect of carriers on theenvironmental stability of adhesive joints. In Adhesion 5, edited by K. W. Allen.London: Applied Science Publishers Ltd.Muasher, M., M. Sain. 2006. The efficiency of photostabilizers on the color change ofwood filled plastic composites. Polymer Degradation and Stability 91(5):1156 - 1165.Mueller, C. 2000. Otto Hetzer and the birth of laminated timber engineering. InHolzleimbau / laminated Timber Construction. Basel, Berlin, Boston: Birkhaeuser.Mynott, T. I., E. Van der Straeten. 1984. The weathering performance of experimentalglulam beams bonded with an epoxy resin-based adhesive. Journal of AppliedPolymer Science: Applied Polymer Symposium 40, 19 - 25.Nagai, N., T. Matsunobe, T. lmai. 2005. Infrared analysis of depth profiles in UV-photochemical degradation of polymers. Polymer Degradation and Stability 88(2):224 - 233.102Nagle, D. J., M. Celina, L. Rintoul, P. M. Fredericks. 2007. Infrared microspectroscopicstudy of the thermo-oxidative degradation of hydroxy-terminatedpolybutadiene/isophorone diisocyanate polyurethane rubber. Polymer Degradationand Stability 92(8):1446 - 1454.Nguyen, T. L. H., C. E. Rogers. 1988. Effects of mechanical deformation on thephotodegradation of acrylic-melamine coatings. Journal of Polymer Science andTechnology 37:431 - 435.Nichols, M. E. 2005. Paint weathering tests. Chap. 19 in Handbook of EnvironmentalDegradation of Materials, edited by M. Kutz. New York: William Andrew Publishing.011ier-Dureault, V., B. Gosse. 1998. Photooxidation of anhydride-cured epoxies: FTIRstudy of the modifications of the chemical structure. Journal of Applied PolymerScience 70(6):1221 - 1237.Osawa, Z. 1983. Photodegradation and photooxidation of polymers. In Degradation andStabilization of Polymers, edited by H. H. G. Jellinek. Amsterdan and New York:Elsevier.Paul, N. C., P. J. Pearce, D. H. Richards, D. Thompson. 1979. Synthesis of flexibilisedepoxy adhesives. In Adhesion 3, edited by K. W. Allen. London: Applied SciencePublishers Ltd.Petersen, A. K., B. Solberg. 2002. Greenhouse gas emissions, life-cycle inventory andcost-efficiency of using laminated wood instead of steel construction. EnvironmentalScience & Policy 5(2):169 - 182.Petrie, E. M. 2007. An overview of adhesives and sealants. Chap. 1 in Handbook ofAdhesives and Sealants. 2nd ed. New York: McGraw Hill.Petrie, E. M. 2007. Stress and joint design. Chap. 5 in Handbook of Adhesives andSealants. 2nd ed. New York: McGraw Hill.Petrie, E. M. 2007. Epoxy, polyurethane, acrylic, and cyanoacrylate adhesives. Chap.13 in Handbook of Adhesives and Sealants. 2nd ed. New York: McGraw Hill.Petrie, E. M. 2007. Selection of adhesives Chap. 17 in Handbook of Adhesives andSea/ants. 2nd ed. New York: McGraw Hill.Phanomen Farbe. 2001. Kleben Dichten. Accessed 26th March 2008. Available from:http://www.colour-europe.de/pf_113_2001_12.htm . Internet.Phanomen Farbe. 1999. Handel Handwerk. Accessed 26th March 2008. Available from:http://www.colour-europe.de/pf_411_2000_11.htm . Internet.Pizzi, A. 1994. Melamine-formaldehyde adhesives. In Handbook of AdhesiveTechnology, edited by A. Pizzi, K. L. Mittal. New York: M. Dekker.103Pospisil, J., W. - D. Habicher, J. Pilar, S. Nespurek, J. Kuthan, G. - 0. Piringer, H.Zweifel. 2002. Discoloration of polymers by phenolic antioxidants. PolymerDegradation and Stability 77(3):531 - 538.Potter, W. G. 1970. Introduction to epoxide resins. In Epoxide Resins. London: Iliffe.Rabek, J. F. 1987. Photodegradation of Polymers. Chap. 14, Photostabilization ofPolymers. Chap. 15 in Mechanisms of Photophysical Processes and PhotochemicalReactions in Polymers. Chichester and New York: Wiley.Ranby, B. G., J. F. Rabek. 1975. Fundamental principles of photochemistry. InPhotodegradation, Photo-oxidation and Photostabilization (Principles andApplications). London and New York: Wiley.Reizig, A., T. Nguyen, D. Martin, L. Sung, X. Gu, J. Jasmin, J. W. Martin. 2006.Relationship between chemical degradation and thickness loss of an amine-curedepoxy coating exposed to different UV environments. JCT Research 3(3):173 - 184.Renz, W., M. Wood, J. Suhadonlik, R. Ravichandran, R. lyengar, P. Hayoz. l eFebruary 2001. Stabilized adhesive compositions containing highly soluble, red-shifted, photostable benzotriazole UV absorbers and laminated articles derivedtherefrom. US 6,187,845 Bl.Rhude, A. J. 1996. Structural glued laminated timber: History of its origins and earlydevelopment. Forest Products Journal 46(1):15 - 22.Scaiano, J. C., J. C. Netto-Ferreira, V. Wintgens. 1991. Fragmentation of ketyl radicalsderived from a-phenoxyacetophenone: an important mode of decay for lignin-relatedradicals. Journal of Photochemistry and Photobiology A: Chem. 59(2):265 - 268.Schultz, J., M. Nardin. 1994. Theories and mechanisms of adhesion. In Handbookof Adhesive Technology, edited by A. Pizzi, K. L. Mittal. New York: M. Dekker.Schwarzer, J. 2002. Kunststoffe. Fachhochschule Rosenheim FachbereichAllgemeinwissenschaften, Rosenheim.Searle, Norma D. 2000. Activation spectra of polymers and their application tostabilization and stability testing. Chap. 16 in Handbook of Polymer Degradation. 2nded., revised and expanded, edited by S. H. Hamid. New York: Marcel Dekker.Selbo, M. L. 1965. Performance of melamine resin adhesives in various exposures.Forest Products Journal 15(12):475 — 483.Sell, J. 1983. Rissbildung bei wetterbeanspruchten Brettschichttraegern. Holz-Zentralblatt, 20 April, 109(47).Sellers, T. Jr. 1994. Adhesives in the wood industry. In Handbook of AdhesiveTechnology, edited by A. Pizzi, K. L. Mittal. New York: M. Dekker.104Shlyapnikov, Y. A., S. G. Kiryushkin, A. P. Martin. 1996. in Antioxidative Stabilization ofPolymers. London: Taylor and Francis Publishers Ltd.SpecialChem Innovation and Solution in Adhesives and Sealants. n.d. Adhesion theory— Mechanical Interlocking. Accessed 24th March 2008. Available from:http://www.specialchem4adhesives.com/documents/indexables/contents/4/images/interlock.gif. Internet.Stark, N. M., L. M. Matuana. 2006. Influence of photostabilizers on wood flour - HDPEcomposites exposed to xenon-arc radiation with and without water spray. PolymerDegradation and Stability 91(12):3048 - 3056.Stevens, M. P. 1999. Polymer Chemistry an Introduction. 2nd ed. New York: OxfordUniversity Press.Torikai, A. 2000. Wavelength sensivity of photodegradation of polymers. Chap. 15in Handbook of Polymer Degradation. 2nd ed., revised and expanded, edited by S.H. Hamid. New York: Marcel Dekker.Turton, T. J., J. R. White. 2001. Observation of different photo-degradation behavior intwo similar polypropylenes. Journal of Materials Science 36(19):4617 - 4624.Voyutskii, S. S., V. L. Vakula. 1963. The role of diffusion phenomena in polymer-to-polymer adhesion. Journal of Applied Polymer Science 7(2):475 — 491.Wu, Y. C., C. — M. Huang, Y. Li, R. Zhang, H. Chen, P. E. Mallon, J. Zhang, T. C.Sandreczki, D.-M. Zhu, Y. C. Jean, R. Suzuki, T. Ohdaira. 2001. Deterioration ofa polyurethane coating studied by positron annihilation spectroscopy: Correlationwith surface properties. Journal of Polymer Science 39(19):2290 - 2301.Yoshida, H. 1986. Bond durability of water-based polymer-isocyante adhesives(API) for wood. Mokuzai Gakkaishi 32(6):432 — 438Yusuf, S., Y. Imamura, M. Takahashi, K. Minato. 1995. Weathering properties ofchemically modified wood with some cross-linking agents and its decay resistanceafter weathering. Mokuzai Gakkaishi 41(8):785 - 793.Zeus Industrial Products. 2006. Weathering of plastics. Accessed 08th February 2008.Available from: http://www.zeusinc.com/pdf/Zeus_Weathering_of_Plastics.pdf .Internet.Zeus Industrial Products. 2005. UV properties of plastics: Transmission and resistance.Accessed 28th March 2008. Available from:http://www.zeusinc.com/pdf/Zeus_UV_Properties.pdf . Internet.105Appendix 1Statistical Analysis — Chapter 3Effects of adhesive type, stabilizer, and adhesive stabilizer interaction on color (deltaE,L*, a*, and b*), tensile strength and weight loss (see CD)106Appendix 2Statistical Analysis — Chapter 4Effects of adhesive type, stabilizer, and adhesive stabilizer interaction on color (deltaE,L*, a*, and b*), and surface roughness (see CD)107


Citation Scheme:


Usage Statistics

Country Views Downloads
United States 6 1
Netherlands 1 0
Germany 1 29
China 1 0
Ukraine 1 0
Belgium 1 0
France 1 0
Hong Kong 1 1
Japan 1 0
City Views Downloads
Unknown 4 31
Ashburn 2 0
Plano 1 0
Raleigh 1 1
Beijing 1 0
Brussels 1 0
Arlington Heights 1 0
Central District 1 1
Mountain View 1 0
Tokyo 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items