Timber has been used as a construction material in civil infrastructure throughout the world for several millennia and it is still a popular construction material to this day. Degradation of timber due to mechanical and environmental actions, as well as possible higher loads, can necessitate the need for strengthening or repair. The external bonding of fibre-reinforced polymer (FRP) composites offers a viable solution. A lack of understanding of the strength and behaviour of FRP-to-timber bonded interfaces is, however, hindering the safe and rational design of FRP strengthening measures for timber structures. The aim of this research project is to enhance understanding of the strength and behaviour of the bonded interface between timber and FRP. Important tangible outcomes of the project include the development of effective bonding systems as well as bond stress-slip models and bond strength model which quantify the bonded interface. In order to achieve these outcomes, an extensive experimental and analytical investigation is conducted. Tests are performed and reported on FRP-to-timber joints as well as FRP-strengthened beams. For the former, softwood (Pine), hardwood (Camphorwood) and glulam timber products have been tested. Variables include (i) externally bonded (EB) plates and near-surface mounted (NSM) plates, (ii) FRP plates formed in a wet lay-up procedure and pultrusion, (iii) bonded length of FRP, (iv) adhesive type, (v) adhesive thickness, (vi) timber species, and (vii) natural growth characteristics of the timber such as annual growth rings and knots. The concept of an effective bond length has been verified from the tests as well as effective procedures and materials for bonding FRP to softwoods and hardwoods. Models are also proposed and validated to quantify the bond strength and bond stress-slip relationships of the joint tests. The strengthening methods are then applied to glulam beams in order to observe the behaviour and strength of the bonding systems on a larger scale as well as on a system that bends. The bond strength model proposed from the joint tests is then assessed against the beam tests. Finally, conclusions are made on the entire program of study. Then, recommendations for future research are proposed. / published_or_final_version / Civil Engineering / Doctoral / Doctor of Philosophy

Polymeric diphenylmethane diisocyanate, pMDI, is a wood adhesive that provides excellent composite board properties. Much is unknown about the specific mechanism of pMDI/wood adhesion under conditions that are typical of wood gluing operations. The present research describes the use of ¹⁵N cross-polarization, magic-angle spinning (CP/MAS) NMR as a technique for probing the cure chemistry and bondline morphology of pMDI-bonded wood composites. A 99% ¹⁵N-enriched pMDI resin with desirable adhesive properties was synthesized. A series of model cellulose/¹⁵N-pMDI composites, cured as a function of cellulose precure moisture content, were tested prior to solid wood composites in order to test the feasibility of this technique. Solid wood/¹⁵N-pMDI composites were then cured as a function of wood precure moisture content, cure temperature, and cure time. The ¹⁵N CP/MAS NMR spectra clearly show the dominance of the isocyanate/water reaction on the cure chemistry of all composites tested, both cellulose and solid wood. Four prominent resonances are observed in each spectrum: residual isocyanate, polyurea, and the amide and imide nitrogens of biuret type structures. Different trends in the relative intensities of these resonances are observed as a function of the press variables. Significant amounts of urethane formation are not detected; however, low amounts could be obscured by signal overlap. Relaxation studies using variable contact time experiments were complicated by excessively long cross-polarization rates for nonprotonated nitrogens. However, experiments using variable spin lock times prior to fixed contact periods indicate that the cured resin in these composites is probably a homogeneous continuum. The utility of ¹⁵N CP/MAS NMR for elucidating fine structural and morphological information from complex isocyanate-cured wood composites is clearly demonstrated. / Master of Science

Hydroxyakyl lignin derivatives were reacted with polymeric methylene diphenyl diisocyanate (PMDI) and hexamethoxy-methyl-melamine (HMMM) to form polyurethane and polyether wood adhesives respectively. Adhesive performance in shear block tests indicated: (a) that the combination of lignin and PHDI reduced the adhesive strength shown by neat PMDI. The HMMM failed to produce an acceptable wood adhesive in the absence of lignin, requiring 50-60% lignin derivative co-substrate for peak performance; (b) adhesive performance was related to molecular weight, if an organic solvent was the carrier, or solubility if the formulation was emulsified; (c) adhesive performance for the lignin based adhesives was better than a urea formaldehyde reference. Structure property relationships were determined by correlating data obtained by the analysis of (in vivo) cured adhesive films and (in vitro) adhesive strength data resulting from shear block testing. The results indicated that: (a) glass transition temperatures of the in vivo cured adhesives were inversely related to the strength of the adhesives cured in vitro; (b) variations in infrared analysis of the in vivo cured adhesives were used to determine the levels of products from the cross linking reaction. In vitro adhesive strength was directly related to the level of reaction products determined to be present in the in vivo wood adhesives; (c) the relationships between the analysis of in vivo and in vitro cured adhesives indicated that the lignin component may act as a soft segment blocks or domains in a more rigid polymer matrix. Particle board was produced with the lignin adhesives with: (a) properties equal to those produced with commercial OF resins; (b) spray application greatly reducing the effects of carrier compatibility; (c) none of the lignin based adhesives were water resistant. / Master of Science