mparative Tests of Shear Strength of Adhesive Lap Joints After Thermal Shocks

This work presents the results of comparative tests for the determination of Young’s modulus and the static shear strength of adhesive lap joints, based on grade 316L steel. The tests also concerned the determination of the glass transition temperature of a certain adhesive composition: Epidian 57 epoxy resin with a 10% Z1 hardener content. The paper shows the test results for the surface free energy and selected surface roughness parameters, including photographs of the test specimens after destructive testing. The tests were comparative and performed on adhesive joints, with and without exposure to thermal cycling. The scope of the testing included a relatively short thermal cycling run of 500 cycles with a temperature variation of -40°C to +60°C. An analysis was carried out of the results from testing the static shear resistance of specimens manufactured using various methods of adhesive joint seasoning. The experimental test results were statistically processed in compliance with good research practice.


INTRODUCTION
Adhesive compositions based on epoxy resins are one of the most popular materials used in adhesive processes. Adhesive joints are frequently used in many modern sectors of industry, especially in aerospace, automotive, and machine engineering. Adhesives and adhesive processes are utilised not only for bonding components of machines, but also to seal and encapsulate mechanical constructions. This is related to the many advantages provided by adhesives and adhesive joints. How good the adhesive bond is depends largely on the process of pretreating the faces of the materials to be joined [1][2][3]. As a rule of thumb, the adhesive processes are intended to "develop" the interface geometrically and provide an optimum energy state of the faces of the joined components [1,4].
The design and process engineers decide when and how adhesive joints should be applied for operation under thermomechanical loads (or thermal shocks). The process of adhesive joint fatigue is an eff ect caused by stresses which change cyclically in time. This process results in a reduction of adhesive joint strength and durability, a consequence of which is structural failure at a stress lower than the static strength of the joint [1,4]. The issue of thermal fatigue in adhesive joints and its eff ects on long-term, reliable operation still remains an object of research [5][6][7][8].
Temperature signifi cantly aff ects the strength of single adhesive lap joints. Changes in temperature by several to several dozen degrees Celsius which do not largely aff ect the properties of metals can change the properties of the long-chained plastics used as industrial adhesives. Varying thermal loads aff ect adhesive joints: fi rst they introduce thermal stresses into the joint system and then they change the mechanical properties [9].
With the state of the art today it is possible to make hybrid joints by combining adhesive processes and traditional joining methods, like pressure welding or riveting [10][11][12].
The adhesive joint strength does not exceed the nominal tensile strength of the stainless steel given by the literature [13] and is far bellow the values of ultimate tensile strength reported for structural steels [14].

EXPERIMENTAL PROCEDURE
The tests were performed on an adhesive composition made of Epidian 57 epoxy resin with a 10% Z1 hardener content. This material was used to manufacture the test specimens for the determination of the material properties. The general formula of the epoxy resin is shown in Figure 1.
The number n depends on the type of epoxy resin and is characterised by the epoxy number. In the initial stage of curing between the amine group and the epoxy group, a hydroxyl group forms. A simplified diagram of this is shown with equation (1): The product of the addition includes secondary amine groups which can react with the epoxy group of the next resin molecule in succession (2): Some of the specimens were only seasoned at room temperature for seven days while the remainder were exposed to thermal shocks. These specimens were exposed to 500 thermal cycles (thermal shocks) in a thermal shock chamber. The set minimum temperature limit was -40°C, and the set maximum temperature limit was +60°C, providing a 100°C temperature variation range. The specimens were conditioned in each chamber for 15 minutes, not including the time to achieve a stable chamber temperature. Moreover, some of the test specimens were heat-treated by reheating (at +80°C for 2 h) to eliminate the effects of high temperature in the thermal shock chamber. Figure 2 shows an overview of the test specimens formed into dumbbells for the determination of the material characteristics before and after the thermal shocks. Table 1 summarises the pretreatment processes used for the test specimens with the adhesive composition (Epidian 57 with a 10% Z1 hardener content) used to determine Young's modulus.
The next stage of the tests was to determine the strength of the adhesive, which was done in compliance with the prevailing reference standard. The test specimens were made from grade 316L steel, each measuring 25x100x1.5 mm. To develop the geometric surface and remove the physical adsorption layer, the test specimens were mechanically processed with a P320 grit tool for 30 seconds. This mechanical cleaning was done in two steps: first each test specimen was washed with a degreaser (Loctite 7063) followed by wiping with a paper towel (both operations were done twice), and second the test specimens were washed with the degreaser (Loctite 7063) and left for it to evaporate. A PGX goniometer, including software, was used to measure the wetting angle of the grade 316L steel surface and determine the SFE (free surface energy). The wetting angle was measured with distilled water and diiodomethane at least ten times on each of the test specimens of interest. The measurements were made on a test panel and followed a level check with an optical level gauge, at an ambient temperature of (19-21)°C and (45-50)% RH. The test liquids used for measuring the wetting angle were applied on the test specimen surface automatically, in drips .5 mJ/m 2 . DSC analysis was performed to determine the glass transition temperature. The cross-linked test specimen was overheated from 0 to 250°C at 10°C/min, followed by cooling and repeating the overheating. During the tests, the glass transition temperature during the first and second heating was determined.
The test specimens measured 100 mm x 25 mm in one plane, each 1.5 mm thick, of grade 316L steel, and were adhesively bonded to form adhesive joints. Figure 3 shows the schematic layout of the single adhesive lap joint made with the adhesive composition of Epidian 57 and a 10% Z1 hardener content. The adhesive layer thickness was g k = 0.1 mm, whereas the lap length was 12.5 mm.
The adhesive composition tested was cured at an ambient temperature of (20-22°C) and (45-55)% RH. The unit pressure applied to the test specimen surface during the adhesive bonding process was 0.2 MPa and the curing time of the adhesive composition was 120 hours. Some of the single adhesive lap joint test specimens were exposed to thermal load cycles. Table 2 summarises the test specimen preparation process to produce single adhesive lap joints for grade 316L steel bonded with the adhesive composition.
For these tests, a Keyence VHX-5000 microscope was used for image work with the grade 316L steel specimens before and after mechanical preparation with the P320 grit tools.
The surface roughness was measured with a Hommel-Etamic 3D T8000 RC-120-400 tool and a dia. 2 µm feeler, while the elementary section length was chosen according to the reference literature [15].  Table 3 is a summary of the 3D isometric maps of the test specimens before and after mechanical preparation with the P320 grit tool and includes 2D profiles. The marks shown by the surface topography maps are typical of this type of surface preparation. Table 4 lists the selected 3D surface roughness parameters for all the studied options of the test specimens. The following 3D parameters were considered: Sq -root mean square value of the 3D profile datums; Sp -maximum 3D profile peak height; Sv -minimum 3D profile valley depth; Sz -maximum 3D profile height; Sa -arithmetic mean of the 3D profile datums. Before the measurements, the surfaces of the test specimens were cleaned and degreased with Loctite 7063. Note that the surface roughness profile indicates that the grit tool mechanical preparation effectively developed the required surface finish and removed contaminants from the superficial layer.

Analysis of selected surface roughness parameters
The tests indicate an unambiguous increase in all surface roughness parameters for all grade 316L steel specimens following the surface preparation with the P320 grit tool in comparison to the condition of test specimens prior to the mechanical preparation. Figure 4 shows the photographic imaging of the grade 316L steel specimens before and after surface preparation with the P320 grit tool.

Test specimen surface imaging
The photographs were taken with a 500x magnification. Figure 4(b) shows the characteristics scratches due to the kinematics of the mechanical preparation process and the nature of the grit tool. The surface preparation with the grit tool was intended to remove the layer of contaminants, or the physical adsorption layer.   As-seasoned, without grit surface preparation, without thermal shock cycling P2 As-seasoned, without grit surface preparation, after 500 thermal shock cycles P3 As-seasoned, after grit surface preparation, without thermal shock cycling P4 As-seasoned, after grit surface preparation, after 500 thermal shock cycles Table 3. 3D isometric maps of the test specimens before and after the surface preparation with the P320 grit tool, with 2D profi les   The tests were performed on the SFE to verify the correct pretreatment of the grade 316L steel superficial layer for application of the adhesive processes, including the adhesive bonding. The tests revealed a small scatter in the results both for the test specimens before and after the surface preparation with the P320 grit tool, which could mean that the energy state was homogeneous over the surfaces of the adhesively bonded specimens. This energy state favours the production of adhesive joints. This analysis revealed an unambiguous increase in SFE for all grade 316L steel specimens following the surface preparation with the P320 grit tool in comparison to the condition of test specimens prior to the surface preparation. The SFE increased by 18.5% and is advantageous in adhesive technologies. Energetic properties of constructional materials surface layer are very important features in adhesive bonds constructions. These connections are for example: contact bonding, sealing, varnishing, etc. The parameter that determines the effectiveness of joining the surfaces of construction materials is the adhesive energy, which is the work necessary to separate the two layers of the adhesive joints material without penetrating the nature of the interactions causing the formation of a permanent joint.

Glass transition temperature
In this test carried out on the test specimens produced with the T1 process, the determination of the glass transition temperature proved to be difficult as the test specimen underwent curing, which should begin at 60°C in the test setup. The curing enthalpy had the exothermic peak shown in red in Figure 5.
The test specimen (T1) was heated at a rate of 10°C/min, and the cross-linking process enthalpy shown in Figure 6 was 368 J/g. Naturally, the process was exothermic, which meant it produced heat (see the peak pointing upward). The cross-linked test specimen was overheated from 0 to 250°C at 10°C/min, followed by cooling and repeating the overheating (Figure 7). The glass transition temperature of the cured adhesive composition was 79°C and reached 85°C in the second heating run, exceeding the first run by 6 degrees. This meant that the crosslinked test specimen was fully cross-linked in the first 0-250°C heating run.
For the test specimens prepared in the T3 process, which included reheating at 80°C for 2 hours, three tests were performed to determine the glass transition temperature. An example of the glass transition temperature determination on a T3 process test specimen is the trend shown in Figure 8, while a summary of all the test results is given in Figure 9.
The tests led to a conclusion that the T1 process test specimen was not cured fully, whereas all the three test specimens produced with the T3 process were fully cured. Naturally, the glass transition temperature in the second heating run was higher than in the first heating run, because the test specimens were always fully cured in the first overheating run within 0-200°C. For the T3 process test specimens, the glass transition temperatures were, respectively: test specimen #1 (first heating run, glass transition   temperature -91°C; second heating run, glass transition temperature -103°C); test specimen #2 (first heating run, glass transition temperature -93°C; second heating run, glass transition temperature -104°C); test specimen #3 (first heating run, glass transition temperature -93°C; second heating run, glass transition temperature -100°C).

Young's modulus of the adhesive composition (E57+Z1-10%)
Figures 10 to 13 show the trends of stress based on which the Young's modulus values were determined for the adhesive composition of Epidian 57 with a 10% Z1 hardener content. The test specimens for the Young's modulus   Table 1. For each process option, five test specimens of the material were produced, and the mean values are shown in Figure 14 with the measure of scatter.
The tests performed on the adhesive composition (Epidian 57 + 10% of Z1 hardener) revealed an increase in Young's modulus following the reheating run at 80°C for 2 h in comparison to the test specimens before the reheating run. The highest increase (36%) was found for the T4 process test specimens compared to the T1 process test specimens. For the test specimens prior to thermal shock cycling, i.e. the T1 and T3 processes, Young's modulus increased by 29% for the T3 process test specimens compared to the T1 process test specimens. A significant increase in scatter around the mean value was found for the test specimens following the Fig. 13. Young's modulus determination for T4 process test specimens reheating run (T3 and T4 processes) compared to the test specimens before the reheating run (T1 and T2 processes). The measure of scatter was the standard deviation which increased five-fold in this case. Figures 15 to 18 show the failure stress trends in the single adhesive joint specimens prepared in the process shown in Table 2. The processing of the test specimens for this test included surface preparation of grade 316L steel with the P320 grit tool and loading with variable thermal stress. For each process option, 10 test specimens were produced.

Adhesive joint strength
An analysis of the failure stress trends in the single adhesive lap joints revealed larger scatters in the trends for the post-thermal-shock cycled specimens (P2 and P4) than in the test specimens before the thermal shock cycling (P1 and P3). Figure 19 shows the mean values of strength for the single adhesive lap joints from the function of the test specimen process option, plus the standard deviation values.
The tests revealed a reduction of the failure stress values in the single adhesive lap joints after the thermal shock cycling in comparison to the test specimens before the thermal shock cycling. This reduction occurred for the test specimens with and without surface preparation with the P320 grit tool. For the test specimens before the surface preparation, the failure stress was reduced by 20% for the post-thermal-shock cycled specimens compared to the specimens before the thermal shock cycling. For the test specimens with surface preparation with the P320 grit tool, the failure stress was reduced more, by 45%     for the post-thermal-shock cycled specimens compared to the specimens before the thermal shock cycling. Table 5 shows the photographic imaging of the test specimens after the strength tests. The photographs show the test specimens before and after the surface preparation with the P320 grit tool and before and after the 500 cycles of thermal shock treatment. The photographs were taken with a 500x magnification.

Imaging of the test specimens after the strength tests
The photographs revealed the adhesivecohesive nature of the specimen fractures, especially in those with surface preparation by the P320 grit tool and after the thermal shock cycling. This could be evidence of thermal fatigue of the tested single adhesive lap joints. Table 5. Imaging of the test specimens after the strength tests