Influence of wood-adhesive system, pre-crack configuration, and data-reduction method on the Mode II fracture behavior of bonded North American hardwood joints for structural applications

Fracture parameters of bonded hardwood joints are essential for the numerical modeling of hardwood-based structural wood composites, yet limited data are available for North American hardwood species under Mode II loading. Therefore, this study evaluated the effects of wood species, adhesive type, initial pre-crack configuration, data-reduction method, and cohesive-element implementation on the Mode II fracture response of bonded hardwood joints. End-notched flexure tests were performed on red maple, red oak, and yellow poplar specimens. Red maple and yellow poplar were bonded with one-component polyurethane, while red oak was bonded with phenol-resorcinol-formaldehyde. The fracture toughness ( G IIc ) was calculated using an adapted ASTM D7905 procedure, the compliance-based beam method (CBBM), and digital image correlation (DIC). The initial pre-crack configuration did not show statistically significant differences in the average fracture toughness trends within the investigated a ₀/ L range. In contrast, the data-reduction method clearly affected the calculated G IIc values. CBBM yielded the highest average values, DIC the lowest, and the adapted ASTM approach predicted intermediate results. Red maple and red oak joints showed comparable fracture toughness, whereas yellow poplar joints were significantly lower. DIC-derived cohesive laws were implemented through cohesive elements in finite element simulations. The simplified bilinear cohesive formulation captured crack initiation and maximum load with reasonable accuracy, showing that G IIc and peak shear stress can serve as practical input parameters for modeling bondline fracture in bonded hardwood joints.