Performing lap shear tests on adhesives

For the estimation of the mechanical behavior of adhesively bonded joints, simple polymer samples (e.g., dog bone test) are of limited use, as chemical composition and surface pretreatment of the adherends cannot be considered. This is of special importance if substrate-sensitive adhesives, such as anaerobics, are applied. In technical practice, joints should, wherever possible, be loaded in shear; therefore such destructive lap tests have established themselves as the standard in-situ test methods.

Thick- and Thin-Adherend Tensile-Shear Tests

Shear tests, like the napkin ring torsion technique with butt-bonded hollow cylinders according to ISO 11003-1 (ASTM E 229-92) or the Iosipescu method with notched substrates1, are seldom used in adhesive technology. This is also true for double-lap joints, which are difficult to prepare and cannot avoid principal problems such as differential sample straining. Industry demands simple experimental procedures; thus, static single-lap tests with plate adherends represent the preferential methods.

For the design of a joint, the elastic parameters (two out of Young's modulus E, shear modulus G, and Poisson's ratio v) and - in some cases - the failure stress and strain are usually required. For the determination of these characteristics, the thick-adherend tensile-shear test according to ISO 11003-2 (ASTM D 3983-81) is often applied. Substrate deformation can be taken into account by appropriately attached extensometers2 and numerical correction.3

However, in many cases, such as quality control by measuring bond strengths, the thin-adherend tensile-shear test in accordance with ISO 4587 (ASTM D 1002-72) or ASTM D 3165-73 (laminated assembles to prevent misalignment in the tension tester) is still widely used. Thus, to realistically assess data derived from these two destructive test methods of adhesive technology, one must compare the stresses acting along the glue line and the resulting failure stress. This article examines a linear-elastic finite element (FE) calculation of the stress distributions and an experimental determination of bond strengths at room temperature (296 K). As a concrete example, steel substrates (structural steel S235) and a commercially available anaerobic adhesive are chosen. The mechanical characteristics are presented in Table 1.

Figure 1. FE Model of the Thick-Adherend Tensile-Shear Test with Stepped Substrates

FE Modeling

The commercial FE code MSC.Marc is used. Figures 1-2 show the sample models with the definition of coordinates, boundary conditions and applied loads, as well as mesh details for both single-lap joints with thick and thin adherends, respectively. With a glue length of 12 mm and specimen width of 25 mm, the glue surface area amounts to 300 mm2. The thickness of the adhesive layer (bond gap) equals 30 µm, which is typical of anaerobic products. The essential difference between both types of test joints is the substrate thickness: 12 mm (thick adherends) and 1.6 mm (thin adherends).

Figure 2. FE Model of the Thin-Adherend Tensile-Shear Test
The mesh density is shown in Table 2. The utilized element type No. 114 is a four-node isoparametric arbitrary quadrilateral written for plane stress applications using reduced integration. This element employs an assumed strain formulation developed in natural coordinates, which ensures good representation of the shear strains within the element. Note that the applied plane stress state corresponds to the situation prevailing in the region near the surface, where the joint actually fails.

By supporting the specimen at all nodes with y=0 in the gripping zone, the boundary conditions are chosen in a way that no rotation around the z axis (perpendicular to the x and y direction in Figures 1-2) can occur in this region. A point load of 100 N in x direction is applied to the nodes, thus resulting in an average shear stress for each calculation in the adjacent equation.

Figure 3. Comparison of the Shear Stress-Distance Curves of Thick and Thin Adherends

FE Results

The computed stress curves are normalized with respect to the mean shear stress, tave. In Figures 3-5, ladenotes the glue length (12 mm). For all calculated stress components, the edge raise in the case of the joined thin substrates is much higher if compared with thick adherends.

Figure 4. Comparison of the Tensile Stress sx of Thick and Thin Adherends
Figure 3 reveals that the shear stress is nearly constant across the bond line: even for the thin steel plates, the rising buildup of stress in the rim zone is not particularly pronounced. On the other hand, Figures 4-5 show that the distance distributions of the tensile stresses, especially ox, reveal a large increase in the edge region of the joint, x Æ ± 1⁄2 la.

Figure 5. Comparison of the Tensile Stress sy of Thick and Thin Adherends
However, for the thick adherends, due to the lowered influence of differential straining in these substrates, a rather uniform stress state is formed as an indispensable prerequisite for the derivation of realistic mechanical characteristics. Here, the normalized tensile stresses in the middle of the adhesive layer (x»0) approximately reach zero, whereas marked compression occurs for the thin plate specimens there.

Figure 6. Comparison of the Test Results According to ISO 4587 (Thin Adherends) and ISO 11003-2 (Thick Adherends)

Experimental Results

Single-lap tensile-shear tests with thick and thin S235 adherends are performed at room temperature (296 K) according to ISO 11003-2 with stepped substrates4and ISO 4587, respectively. The anaerobic product defined in Table 1 is applied with a 30 µm gap. Results are shown in Figure 6. In agreement with the findings of the FE simulation presented previously, the measured fracture strength of the adhesively bonded thick substrates (ISO 11003-2) is more than two times higher than the one of the glued thin adherends (ISO 4587).

Conclusion

The present comparison of the static single-lap tensile-shear tests according to the standards ISO 11003-2 and ISO 4587 indicate that adherend thickness considerably influences the effect of differential straining in the substrate and thus the uniformity of the stress state. As a concrete example, structural steel specimens (S235) joined together with an anaerobic adhesive (30 µm gap) are chosen. The performed finite element simulation confirms that for the thick adherends (ISO 11003-2) a much more homogeneous stress distribution across the bond line exists, whereas the thin plane sheet assembly reveals a high stress increase (particularly tensile component sx) at the edge of the joint and compression in the middle region. As an important consequence, under the same test conditions, the experimentally determined fracture strength of glued thick stepped substrates exceeds the one of thin plate samples by a factor greater than two.

Simplified performance is the main advantage of the ISO 4587 thin-adherend tensile-shear test. It is widely used, so a huge variety of measurements is available in literature. However, the obtained data has very limited connection to the desired intrinsic adhesive parameters. As this article emphasizes, realistic shear characteristics, necessary for engineering joint design, can only be determined by applying the ISO 11003-2 thick-adherend test.

For more information on lap shear tests, contact Dr. Jürgen Gegner, SKF GmbH, Ernst-Sachs-Str. 5, 97424 Schweinfurt, Germany; or Dr. Andreas