This work describes the development of an improved mode I fracture testing procedure for adhesively-bonded wood, and demonstrates the sensitivity of this approach. The two significant improvements were: 1) the use of the flat double cantilever beam (DCB) geometry, which has been uncommon for wood and 2) the application of an established and powerful data analysis using a corrected compliance method from beam theory. Three studies were conducted using various wood adhesives and DCB specimens were fabricated from yellowpoplar (Liriodendron tulipifera) sapwood. The sensitivity of this methodology showed significant differences in fracture performance as the degree of cure increased for a phenol formaldehyde adhesive, and yielded maximum strain energy release rate (SERR) values of 370 - 560 J/m2. A second study showed performance differences between two polymeric diphenylmethane diisocyanate (pMDI) adhesives and one polyurethane adhesive. Typical maximum SERR values were 160 and 130 J/m2 for the pMDI adhesives and 160 J/m2 for the polyurethane adhesive. A third study investigated the effect of loading rates on a cross-linked polyvinyl acetate adhesive and maximum SERR values of 370 - 560 J/m2 were achieved. Adhesive penetration and cure were determined by image analysis with fluorescence microscopy, and by micro-dielectric analysis, respectively. Since the geometry of the fracture procedure dictates the absence of wood failure, the resulting fractured surfaces were readily analyzable. The surface analysis techniques of laser ionization mass analysis, solid-state nuclear magnetic resonance and field emission scanning electron microscopy were used to investigate the locus of failure for the smooth fractured surfaces. / Master of Science

The primary direction of this thesis was towards exploring qualitative and quantitative characteristics necessary for refining and understanding the flat wood double cantilever beam (DCB) as a valid means for testing Mode I fracture energy in wood adhesive bonds. Southern yellow pine (SYP) adherends were used with epoxy and phenol formaldehyde (PF) impregnated films, providing two systems with different characteristics for investigation. An adhesive penetration analysis was performed for both the epoxy and PF bonds. The PF penetration into the SYP was shown to be relatively shallow. The epoxy penetration was shown to be deeper. Epoxy-SYP DCBs were quasi-statically tested with varying widths (10 mm, 15 mm, and 20 mm), showing decreases in scatter of critical and arrest strain energy release rates, GIc and GIa, with increases in specimen width. Quasi-static fracture testing was also performed on PF SYP-DCBs, showing much higher critical and arrest fracture energy values than the epoxy-SYP DCBs, indicating that deep adhesive penetration is not necessarily a requisite for higher Mode I fracture energy values. Grain distribution influences were computationally investigated because of the stiffness difference between latewood and earlywood growth and the grain angle along the length of the beams. The grain angle and the stiffness difference between latewood and earlywood growth caused the effective stiffness, (ExxI)eff, to vary along the length of the beam. The effective stiffness variation caused variations in the beam's ability to receive and store strain energy, complicating and confounding determination of experimental results. Cyclic loading tests were performed on PF-SYP DCB's. The cycle frequency was 3Hz, with a valley to peak load ratio of R = 0.5. Specimen softening was observed with cycling, with re-stiffening occurring with crack growth. Contrary to expectations, specimen compliance occasionally decreased with small crack extensions. A toughening mechanism was frequently observed, whereby subsequent crack lengths required more cycles to failure than the previous crack length. Monotonically extending the crack length far from the fatigued region created a fresh crack that did not show the toughened behavior. But toughening did resume with subsequent crack lengths. / Master of Science