Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
775530 | Engineering Fracture Mechanics | 2009 | 22 Pages |
Structural adhesive joints were subjected to high loading rates in mode I and their resulting fracture behaviour was studied in detail. Joints were formed between unidirectional carbon-fibre epoxy composites and between aluminium alloy substrates bonded with a tough, single-part automotive adhesive (XD4600) from Dow Automotive. Double cantilever beam (DCB) and tapered double cantilever beam (TDCB) tests were performed, from quasi-static loading rates up to 15 m/s, and a test rig was developed incorporating high-speed video acquisition for the high-speed tests. A detailed data reduction strategy was developed to account for (i) the types of different fracture behaviour regimes encountered, (ii) the dynamic effects in the test data, and (iii) the contribution of kinetic energy in the specimen arms to the energy balance. Using the above data reduction strategy, increasing the test rate over six decades (from 10−5 to 101 m/s) was found to lead to a reduction in the value of the adhesive fracture energy, GIc, by about 40% of its quasi-static value, i.e. from 3.5 to about 2.2 kJ/m2. Further, at quasi-static loading rates, the measured adhesive fracture energies were independent of substrate material and test geometry (i.e. DCB or TDCB). However, at faster loading rates, the TDCB tests induced higher crack velocities for a given loading rate compared with the DCB test geometry, and neither the test rate nor the crack velocity were found to be the parameter controlling the variation in GIc with increased test rate. Thus, an isothermal–adiabatic model was developed and it was demonstrated that such a model could unify the DCB and TDCB test results. Indeed, when the GIc values were plotted as a function of 1/√time, where the time was defined to be from the onset of loading the material to that required for the initiation of crack growth, the results collapsed onto a single master curve, in agreement with the isothermal–adiabatic model.