PSI - Issue 25

F.J.C.F.B. Loureiro et al. / Procedia Structural Integrity 25 (2020) 63–70 Loureiro et al. / Structural Integrity Procedia 00 (2019) 000 – 000

64

2

1. Introduction

Several benefits over conventional processes such as welding, fastening or riveting promote the use of the adhesive technology. In fact, the capability to obtain lighter and stronger structures, the possibility to join dissimilar materials or thin plates, the skill of enduring higher loads due to a more uniform stress fields along the bonded area, a good corrosion resistance and better aesthetics, are characteristics that promote the application of this method. The drawbacks include the careful joint design oriented to suppress peel stresses, low resistance to temperature and humidity, requirement of a surface treatment and disassembly issues (Petrie 1999). To allow a higher reliability, accurate predictive tools should be provided to engineers. The CZM technique, proposed by Barenblatt (Barenblatt 1959, Barenblatt 1962) and Dugdale (1960), is capable of simulating damage initiation and growth. Through CZM, a fracture can be artificially introduced in structures, in which damage growth is allowed by the introduction of a possible discontinuity in the displacement field. CZM reproduce the damage along a given path, establishing a traction-relative displacement law, by specification of several parameters ruling the crack growth process such as fracture toughness in mode I ( J IC ) and mode II ( J IIC ). Therefore, to determine the cohesive law parameters, fracture characterization tests must be performed. Within this scope, the R-curve of a specific fracture test plots J I or J II against the crack length ( a ). The R -curve ideally describes a perfect horizontal J - a curve during damage growth, whose steady state value gives the measurement of J C . Typically, failure in adhesive bond takes place under mixed-mode due to the joint geometry, the different properties of the joint constituents and applied load. Therefore, to accurately predict the joint strength, it is necessary to take into account both J IC and J IIC , and also to use mixed-mode criteria (da Silva and Campilho 2012). Mixed-mode test methods include the asymmetric double-cantilever beam (ADCB), the mixed-mode bending (MMB) and the SLB tests. The SLB test proposed by Yoon and Hong (1990) is based on the End-Notched Flexure (ENF) test. However, the specimen’s lower beam has a smaller length, resulting in an unsupported edge which leads to an opening mode and a shear mode event at the same time at the crack tip (Szekrényes and Uj 2004). To estimate J C , several reduction methods are available, such as the J -integral, proposed by Rice (1968). The J -integral has been used to determine the strain energy release rate using a path-independent integral contour around the crack tip. Ji et al. (2012) used the SLB test to investigate the t A effect on the interfacial J , joint stresses and shapes of the interfacial traction-separation laws under mixed-mode loading, using CZM-based nonlinear numerical models. J C was estimated by a J -integral formulation. Specimens with five different t A (0.1, 0.2, 0.4, 0.6 and 0.8 mm) of the adhesive Loctite ® Hysol 9460 (Loctite ® from Düsseldorf, Germany) were built with 1018 low carbon steel adherends. The t A dependent CZM laws were determined by the direct method, i.e. by the differentiation of the experimental curves. The results showed that J C increases with t A . In addition, the tensile cohesive strength ( t n 0 ) decreases, while the maximum shear stress ( t s 0 ) increases as t A increases. This work presents an experimental study using the SLB test on specimens bonded with the Araldite® 2015, to study their mixed-mode fracture properties. For this purpose, the J -integral method was applied to estimate J I and J II . Framing the obtained values in the fracture envelope enabled to select which power-law failure criterion is more appropriate for this adhesive. Moreover, the tensile and shear CZM laws were obtained by the direct method.

2. Experimental Work

2.1. Joint materials

The adherends were manufactured in composite material, created from a unidirectional carbon-epoxy pre-preg (SEAL ® Texipreg HS 160 RM; Legnano, Italy). The pre-preg was supplied in roll form, with a ply thickness of 0.15 mm. To obtain the adherends’ thickness ( h ) of 3 mm, 20 layers of carbon-epoxy pre-preg we re cut with the adherends’ dimensions, hand-lay-up and cured in a hot-plates press for 1 hour at 130 °C and pressure of 2 bar. The elastic orthotropic properties of a unidirectional lamina for similar curing conditions were retrieved from the work of Santos and Campilho (2017).The adherends were bonded with the intermediate ductile epoxy Araldite ® 2015. The mechanical and toughness properties were characterized in previous researches by Campilho et al. (Campilho et al. 2011, Campilho et al. 2013). Characterization of the adhesive was undertaken by performing bulk tensile tests, Thick Adherend Shear Tests (TAST) and dedicated fracture tests (the Double-Cantilever Beam (DCB) and ENF). The bulk