PSI - Issue 13

Christopher Schmandt et al. / Procedia Structural Integrity 13 (2018) 799–805 C. Schmandt, S. Marzi / Structural Integrity Procedia 230 (2018) ECF22

804

6

for all tests (picture 4). Elastic stored energy is then released by building of new fracture surfaces without additional external energetic input. Overcritical crack growing arrests after the elastic stored energy has fallen below a certain amount and changes to subcritical crack growth anew (picture 5), until further monotonic loading has supplied enough stored energy for a next overcritical dynamic crack jump. This behavior can occur many times giving the curve a saw tooth shape. The amplitude of those self-sustained oscillations can be described by difference of J cs , where subcritical crack growth is initiated and J co , where crack growth gets overcritical. After reaching the first time J co , which is nothing else but the previously prescribed fracture energy J c in that initial case, both J co and J cs seem to uniformly decrease with constant amplitude under rising crack length. J cs is increasing with loading rate as well as J co , which was previously also shown for J c . The amplitude of oscillating crack growth seems to have a maximum in the range of cross head velocities in between 0.01 to 1 mm/s. Even higher or even lower cross head velocities yield lower amplitudes, which implies that the observed jerky crack propagation seems to slowly disappear outside the investigated range of loading rates. Especially for crash loading conditions, it is assumed that the frequency of oscillating crack propagation rate gets that high that the crack just seems to propagate in a stable manner, which is in fact nothing else but a rapidly alternating crack propagation rate. In general, stick-slip instabilities regarding crack propagation rate for adhesives under peel loading are well known from literature. Biel et al. (2012) noticed unstable crack propagation in DCB test bonded with Sikaflex® UHM as well. Maugis and Barquins (1988) found that stable crack propagation abruptly turns over to jerky propagation when increasing the crosshead velocity during peel tests with adhesive tapes. At a certain velocity, amplitude decreased until stable propagation was observed anew under higher crosshead velocities. Kinloch and Yuen (1989) observed micro-stick-slip and macro-stick-slip crack growth while peeling flexible polyimide film laminates, which were bonded to copper foil by a polymeric adhesive. Webb and Alfantis (1995) presented a stick-slip fracture model based on a non-monotonic fracture toughness-velocity curve and an inertia dependent modified equation of motion for the crack tip. They compared their model with experimental observations of Yamini and Young (1977), who observed discontinuous stick-slip propagation of cracks in epoxy resins. Almost all authors additionally found a dependency of discontinuous fracture processes on temperature. A final view on fracture patterns in Fig. 5 confirms, that stick-slip fracture can only be distinctly noticed for moderate cross head velocities within the tested range, but neither for highest nor for lowest loading rates. Just at cross head velocities of about 0.01 to 1 mm/s, alternating fracture patterns can be observed, which indicates a distinctive jerky fracture behavior with subcritical and overcritical crack growth. Crack propagation seems to get stable outside this range of loading rates, though the crack tends to propagate near the interface with adhesively failed regions for very low loading rates, which is commonly observed regarding quasi-static peeling of adhesive joints.

Fig. 5. Fracture patterns for different cross head velocities with subcritical and overcritical crack growth regions.

4. Conclusions The effect of loading rate on the mode I fracture behavior of a semi-structural hyperelastic one-component polyurethane adhesive, which is very tough and possesses high modulus, is investigated by performing DCB tests. In a first step, tests are performed by controlling the testing machine on various constant crack opening velocities to investigate rate effects in the region of rising J . Regarding the cohesive law for mode I, the presented results show