Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804 www.springerlink.com/content/1738-494x(Print)/1976-3824(Online) DOI 10.1007/s12206-017-0927-1 Adhesion strength characterization for different frame materials of handheld products† Ngoc San Ha1, Thanh Duc Dao1, Nam Seo Goo1,*, Jae Kwak2 and Soonwan Chung2 1 Smart Microsystem Research Laboratory, Department of Advanced Technology Fusion, Division of Interdisciplinary Studies, Konkuk University, 120 Neungdong-ro, Gwangjin-gu, Seoul 143-701, Korea 2 Global Technology Center, Samsung Electronics Co. LTD, Korea (Manuscript Received January 22, 2017; Revised May 13, 2017; Accepted May 31, 2017) ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Abstract In this study, the adhesion strength for different frame materials of handheld products was investigated. To characterize the adhesive strength, a simple model of a cellphone that consists of a glass window attached to body frame using double-sided tape was designed and fabricated. The adhesion strength with different pull out speeds and aging times was measured using a conventional pull out testing method. To investigate the detailed delamination process of the adhesive layer, the digital image correlation technique was applied. Moreover, in recent years, the trend of using metal for a body frame of a cellphone has gradually increased due to the fashion of metal frames. Therefore, two materials for body frames were considered for testing in this study: Aluminum was the representative metal and polycarbonate was the representative conventional material. The results showed that the strain at the interface between the adhesive layer and body frame is higher than that at the interface between adhesive layer and glass window for both cases of aluminum and polycarbonate frames. Moreover, the fracture energy in the aluminum body frame is higher than that in the polycarbonate body frame. In order to validate the experimental results, the cohesive elements in ABAQUSTM were used for the modeling bonding layer. The results showed a good agreement between simulation and experiment. Keywords: Adhesion strength; Cohesive element; Double-sided tape; Digital image correlation (DIC) ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 1. Introduction Recently, portable electronic products such as cellular phones, notebook computer and Personal digital assistants (PDAs) have progressively become indispensable devices for daily use. Because most of these devices are costly, the demand for quality in these products is comparatively high. Consequently, manufactures and researchers worked diligently to improve the reliability and durability of these products. Cellular phones are the most commonly used of these devices, and they often accidentally dropped onto hard surfaces. The shock impact can cause the physical damage to the internal Printed circuit boards (PCBs) as well as the external body. To improve the mechanical reliability of the cellphones under impact loading, numerous experiments and simulations have been conducted [1-4]. However, most such studies have focused on the component interconnection and the response of PCBs. Studies on the connection between the window and body are rare. Commonly, the window is attached to the body * Corresponding author. Tel.: +82 2 450 4133, Fax.: +82 2 444 7091 E-mail address: [email protected] † Recommended by Associate Editor Jin Weon Kim © KSME & Springer 2017 frame using tape or bonding material in cellphones. Therefore, adhesive tape and bonding material strength play an important role in the reliability improvement of cellphones. In this study, the double-sided tapes was focused to study because they have a quick and easy installation procedure. In fact, the pressure sensitive adhesives do not need curing that affects the production efficiency. Due to their advantage, the pressure sensitive adhesives is increasingly being used as an assembling technique for various applications [5-8], especially, adhesively bonded joints of dissimilar materials [6, 9, 10]. The dissimilarmaterial joint play an important role in cellphone while the window and the body frame are often made from different materials. For the dissimilar-material joint, a variety of studies have showed that failure often occurs along the interface between two different materials with high property mismatch and the bonding strength of dissimilar material joint is far less than the bond strength established by the adhesive manufacturer for joints with similar material. In order to investigate the behavior of adhesive joint, several methods were proposed such as video camera, photoelasticity, and Digital image correlation (DIC). Karachalios et al. [11, 12] used a video camera to investigate the failure mechanisms of different configurations of joints. Creachcadec et al. 4796 N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804  characterized the behavior of a thick flexible adhesive under different monotonic tensile/compression-shear loadings using an optical method. Krishnan and Xu  employed phoelasticity technique and coherent gradient sensing to investigate the fracture properties of adhesive joint between metal and polymer materials. Kashfuddoja and Ramji  investigated the strain field in a thin adhesive layer in the single-sided patch that repaired carbon fiber reinforced polymer panel under tensile load using the Digital image correlation (DIC) technique. Fedele et al.  presented an experimental– numerical methodology to identify the parameters of a cohesive law of an adhesive layer within a joined assembly on the basis of kinematic data provided by DIC. Comer et al.  successfully used 3D-DIC to measure full-field in-plane surface strains and out-of-plane surface deformations for composite single lap bonded joints. Blyberg et al.  studied the adhesive joints by using 3D-DIC in timber/glass applications with three different adhesives: Silicone, acrylate, and polyurethane. They concluded that the behavior of silicone adhesive is highly influenced by its nearly incompressible behavior. Sadoski et al.  conducted experiments with the application of the DIC-based ARAMIS® systems to monitor the deformation process of steel adhesive double lap joints reinforced by rivets. They realized that the application of an additional rivet leads to a substantial increase in energy absorption by about 35 % in comparison to simple adhesive double lap joints. Budzik et al.  applied DIC technique to investigate micro/macroscopic behavior at the interfaces between rigid and soft materials. It is noteworthy that the investigation on the adhesively bonding joints focused on the different kinds of adhesive and bonding methods. Yang et al.  performed the experiments with four types of adherends with steel substrates. They found that interfacial fractures were prone to occur on the substrate with relatively weak yield strength. To the best our knowledge, no study has been found that investigates adhesion strength for the window and body frame assembly of cellphones. Moreover, in the reality, the most common causes of failure for cell phones is under dynamic condition. However, in the primary study of adhesive, the quasi-static condition was considered in the experiment to investigate the fracture properties of adhesive. To simulate the fracture of adhesive, Cohesive zone models (CZMs) have been used frequently in recent years. CZMs are based on the assumption that the stress transfer capacity between the two separating faces of a delamination is not lost completely at damage initiation, but rather is a progressive event governed by progressive stiffness reduction of the interface between the two separating faces. Liljedahl et al.  modeled the moisture degradation of adhesively-bonded aluminum and composite joints using a CZM approach. Hu et al.  used a bi-linear traction-separation response of a CZM integrated with a response surface methodology to simulate the joints progressive damage and environmental degradation process. Campilho et al.  simulated the behavior of adhesively-bonded single- and double-lap joints between alumi- num adherend using CZMs and extended finite element method (XFEM). Xu and Wei  implemented FEM based on CZMs to systematically study the overall strength and interface failure mechanism of single-lap joints with various system parameters including the fracture energy of the adhesive layer, overlap length, and adhesive layer thickness on the load-bearing capability of the joints. Cohesive zone models have been implemented in commercial software including ABAQUSTM  and Genoa® . Recently, there have been numerous studies using ABAQUSTM software to investigate adhesive behavior [27-31]. Carrere et al.  and Liao et al.  used a mixed-mode CZM with a bilinear shape coupled with a finite element subroutine in ABAQUSTM to investigate the load-bearing capacity and damage level of a double scarf joint with various scarf angles and adhesives. Hasegawa et al.  used a cohesive element implemented in ABAQUSTM software to investigate the mechanical behavior of rubbery adhesive and bonded joints under peeling and shearing tests. Bang et el.  studied aluminum foam with initial crack, which has a closed cell form bonded adhesively, to analyze the crack propagation behavior. Xu and Wei  used ABAQUSTM software to explore the influence of adhesive thickness on the overall strength of the adhesive joints. The details and history are referred to in the review written by Elices et al. . In this study, the adhesion strength was characterized using a universal testing machine and DIC technique. To simplify the measurement, a simple cellphone model was designed and fabricated. The model consisted of a glass window and a body frame. The glass window was glued on the body frame using adhesive tape. Nowadays, the trend of using metal for the body frame of cellphone has gradually increased due to the popularity of metal frames. Therefore, two materials for the body frame were considered in this study: Aluminum was the representative metal and polycarbonate was the representative conventional material. The adhesion strength was investigated through pull out tests. The body frame of the sample was fixed using a special clamp, and the glass window was pulled out using a rod and hook system. The strain, failure mechanisms, and damage initiation at the adhesive layer were determined using the ARAMIS® (GOM, Germany)DIC based system. Finally, in order to validate the experimental results, the cohesive element implemented in ABAQUSTM software was used to model the adhesive layer based on the experimental data. Simultaneously, a parametric study was analyzed to investigate the effect of each parameter (stiffness, strength, and fracture toughness) on the delamination of the adhesive layer. 2. Material and methods 2.1 Sample preparation To characterize the adhesion strength for drop impact reliability, a simple model of a mobile phone was designed and manufactured, as shown in Fig. 1(a). The model consisted of N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804 (a) 4797 (b) Fig. 1. (a) Specimen configuration; (b) fabricated phone model: The left model is the aluminum body frame and the right model is the polycarbonate body frame. two main parts: body frame and glass window. The glass window was glued on the body frame using double-sided tape. The glue interface to be attached to the glass window of the frame was around 2 mm on each side. The assembly procedure was as follows: Firstly, the double-sided tape was cut into small strips 2.5 mm in width. One side of the tape was attached to a glass window and the other side was attached to a frame. After that, 5 kg of dummy weight was put on the top of sample for 30 seconds to bond the window to the frame firmly. The size of the fully assembled model was 50´52.2´8 mm3, as shown in Fig. 1(b). The model was then kept at room temperature for three days before testing to achieve maximum bond strength. In this study, two kinds of material cases were considered: Aluminum and polycarbonate. 2.2 Experimental setup The adhesion strength was measured using a universal testing machine (MTS R&D, Korea). The experimental setup was shown in Fig. 2. A special fixture was designed to clamp the sample to the jig of machine. The fixture was fabricated from a 5 mm thick aluminum plate. The fixture was designed so that it can hold the body frame, as shown in Fig. 10. The glass window of sample was then pulled out using a rod and hook system. Firstly, the rod was placed on the center line of the glass window that was initially marked. After that, the hook was lifted up until the load cell changed to 0.1 N. At this time, the cable was tensioned and kept the rod in the right position. Finally, the load and displacement data was set to zero to start the experiment. To investigate the effect of pull out speed on the adhesive strength, three pull out speeds were considered: 0.5, 1 and 5 mm/min. 2.3 Digital image correlation The strain and failure mechanisms were investigated using Fig. 2. Experimental setup for strength measurement. DIC technique and a computer-controlled MTS test machine. The 2D DIC system consists of a controller box and a CCD camera. The CCD camera was fixed on a stand with an inbuilt spirit level to ensure a horizontal level. The CCD camera has a spatial resolution of 2024×2024 pixels. To capture a small area of 10´10 mm2, the CCD camera was equipped with a 50 mm lens and an extension tube. Two halogen lambs were placed in front of sample so that the camera could capture a high contrast image. The deformation of adhesive layer during the pull out test was then analyzed in ARAMIS® software (version 6.0.2). Before measurement, the sample surface was covered with a random black and white speckle pattern. The procedure for making the speckle pattern is discussed in Refs. [33, 34]. 2.4 Finite element method The cohesive behavior of the tape was defined using a traction-separation law that can be used to model the delamination at the interface between the glass window and body frame. Traction-separation law assumes initially linear elastic behavior followed by the initiation and evolution of damage. The elastic behavior can be defined as  4798 N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804 traction t n0 , t s0 d n0 , d s0 d nf , d sf separation Fig. 3. Traction law with linear softening law available in ABAQUSTM. K ns ù ìe n ü ú í ý = Kε K ss û îe s þ ìt ü é K t = í n ý = ê nn îts þ ë K ns (1) (a) where t is the traction stress vector and where tn, ts are two components of traction stress vector and represent the normal and shear tractions, respectively, while en and es are the normal and shear strain, respectively. The matrix K consists of the stiffness parameters that are related to the elastic moduli of the adhesive layer. Damage initiation can be characterized by different criteria. In this study, the maximum nominal stress criterion was considered and defined as  ìï t t üï MAX í n0 , s0 ý = 1 . ïî tn ts ïþ (2) The damage is assumed to initiate when the maximum normal stress ratio reaches a value of one. The symbol áñ represents the Macaulay bracket and indicates that a pure compressive deformation or stress state does not initiate damage. This means that the damage initiation is coupled between tension and shear. After the initiation criterion is reached, the material stiffness degrades under different laws. We assume that the tape behavior is followed a linear softening law, as shown in Fig. 3. Complete separation is predicted by an effective displacement-based evolution or energy-based evolution. In this study, the energy-based evolution is used to simulate the damage separation of the tape  a a ïì Gn ïü ïì Gs ïü í C ý + í C ý =1 ïî Gn ïþ ïî Gs ïþ (3) where Gn and Gs, denote the work done by the traction and their conjugate relative displacements in normal first and normal second shear directions, respectively. Gnc and Gsc are the critical fracture energies required to cause failure in the normal and shear directions, respectively. In Fig. 3, the critical fracture energy is defined as the total area under the bilinear traction-separation model. The critical fracture energy is the most important parameter that influences the behavior of the cohesive element . The relationship between critical fracture energy, traction stress, and failure separation displace- (b) Fig. 4. Force versus displacement: (a) Aluminum body frame; (b) polycarbonate body frame. ment δf is defined as Gc = t 0d f . 2 (4) The ratio between traction stress and damage initiation d0 is presented as initial stiffness K eff = t0 d0 . (5) The above formulas are specific to the normal and shear directions. 3. Results and discussion 3.1 Adhesive strength The strength of the adhesive layer was measured after 4, 5, 6 and 11 days of aging. The effect of aging time on the failure load will be discussed in the next section. Fig. 4 shows the force versus displacement curves when the aluminum and polycarbonate body frames were used. The tendency of these 4799 N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804 The results of adhesive properties are then listed in Table 1. The measured initial stiffness of the adhesive layers for the two cases aluminum and polycarbonate are similar. This means that the experiment was conducted well. Moreover, the maximum traction in the aluminum case is higher than that in the polycarbonate case. Meanwhile, the fracture energy for the aluminum case is smaller than that for the polycarbonate case. This is due to the second peak in the force-displacement curve from polycarbonate case is higher than that from aluminum case. The second peak in the force-displacement curve from polycarbonate case is higher than that from aluminum case because the elongation of fibrils in case of polycarbonate case Table 1. Adhesive layer properties. Polycarbonate Aluminum Maximum stress (MPa) 0.385 0.428 Initial stiffness (MPa/mm) 0.891 0.895 Fracture energy (J/m2) 400 355 (b) (a) Pull out direction curves was consistent, and there are two peaks on the curve. In the initial stage of loading, the microscopic voids are generated. Afterward, the voids grow gradually and reach the limit at the first peak of the force–displacement curve. The first peak determines the maximum strength of the adhesive. When the force increases higher than this peak, the delamination behavior between glass window and aluminum case begins. It is clear to see that the force was increased again due to the elongation of the fibrils (Fig. 5(b)). The second peak shows that the glass was fully detached from the body frames. For the polycarbonate case, the second peak is higher than the first peak because the elongation of fibrils is large before the glass was fully detached from the body frames. In order to determine the adhesive properties, the toe region that was identified as a small section of very low slope before the initial linear portion of the curves was ignored . The toe region on the force-displacement curves was caused from the cable-hook system. The adhesive properties were characterized by fracture stress, initial stiffness, and fracture energy . The parameters of cohesive zone model were identified as Hayashida’s work . Although the stress–displacement curves of the butt joints shown in these figures are very complicated, the curves were modeled with the simplest traction-separation rule, known as the bi-linear model because we only intended to verify the order of the experimentally obtained fracture energy. In this case, sophisticated, but complicated models seemed unnecessary. The parameters for the modeling were determined as follows: ·The fracture stress is defined as the average value of fracture load obtained experimentally. The maximum tractions used for the calculation are 0.385 MPa and 0.428 MPa for polycarbonate and aluminum frames, respectively. ·The initial stiffness of the bi-linear traction-separation law was calculated to be 0.891 and 0.895 MPa/mm for polycarbonate and aluminum frames, respectively, based on Eq. (5). ·The area surrounded by the lines is defined as the critical fracture energy obtained experimentally. The critical fracture energy values were calculated to be was 400 J/m2 and 355 J/m2 for polycarbonate and aluminum frames, respectively. O y 1 mm x Fig. 5. Strain distribution in the y direction (a) aluminum body frame; (b) polycarbonate body frame. is higher than that from aluminum case. The delamination behavior of adhesive layer was investigated using the DIC technique. Fig. 5 shows the Y-direction (force direction) strain of the sample before failure (at 60 % failure load). It is clear that the strain at the interface between the adhesive layer and body frame is higher than that at the interface between adhesive layer and glass window for both cases of aluminum and polycarbonate frames. However, delamination occurred only at the body frame of the aluminum case. Meanwhile delamination occurred at the body frame and glass window for the polycarbonate case. However, failure started at the frame first and then occurred at the glass window later. For clear observation, the strain data of three section across the adhesive layer was investigated and is shown in Fig. 6. In the case of the aluminum body frame, the strain at the interface between the body frame and adhesive layer is around 25 %, which is higher than the strain (2.5 %) at the interface between the glass window and adhesive layer. Moreover, strain at the interface between the body frame and adhesive layer in Sec. 0 is higher than that in Sec. 2. This means that the failure will occur at corner first and propagate along the interface. Similar to the aluminum body frame case, the strain at the interface between the body frame and adhesive layer is higher than the strain at the interface between the glass window and adhesive layer for the polycarbonate body frame case. Therefore, the failure will also start at the bottom of the adhesive layer. However, the strain at the bottom interface is similar to that at the top interface in Sec. 2 (Fig. 6). This indicates that the failure also occurs at the top interface. 3.2 Effect of aging time As we presented above, to investigate the effect of aging 4800 N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804 100 Force (N) 80 0.5 mm/min 1 mm/min 5 mm/min 60 40 20 0 0.0 (a) 0.5 1.0 1.5 Displacement (mm) 2.0 Fig. 8. Force-displacement curve with different pull out speeds. maximum failure force was almost constant. It is thought that the failure force increases with aging time because the increment in aging time allows the adhesive to flow into the peaks and valleys of the frame by capillarity. 3.3 Effect of strain rate (b) Fig. 6. Strain at adhesive layer: (a) Aluminum body frame; (b) polycarbonate body frame. 80 Force (N) 70 60 50 40 30 20 2 4 6 8 Aging time (days) 10 12 In order to investigate the effect of strain rate on the adhesive strength, one batch of eight sample was made with the same initial conditions. As described above, the samples were tested after 6 days for achieving the maximum fracture load. Fig. 8 shows the relationship between force and displacement with different pull out speeds. Because a pull out speed of 1 mm/min was investigated before, one sample was tested again to confirm the previous results. The maximum failure load was around 63.7 N, which was consistent with previously measured data (Fig. 4). It is clear that the tendencies of these curves are similar. When the displacement increases, the force increases dramatically, especially after 0.5 mm, and the force reaches a maximum value at a certain displacement for three pull out speeds. At this time the glass window is separated from the body frame. If the displacement is increased continuously, the window is fully delaminated from the body frame. The behavior of the adhesive in this stage for three pull out speed is slightly different. For a small pull out speed of 0.5 mm/min, the force gradually fell down while the force sharply decreased for a higher pull out speed 1 mm/min and 5 mm/min. Fig. 9 illustrated the maximum force with different pull out speeds. As can be seen from the graph, the force is logarithmically related to the applied pull out speeds. Fig. 7. Effect of aging time on the failure load. 3.4 Finite element method time on adhesive strength, the samples were tested at 3 days, 4 days, 6 days and 11 days after the initial assembly force was applied. The results of failure forces and aging times are shown in Fig. 7. The non-linear curve fitting implemented in Origin Pro 8.5 was used to fit the result (the red line). When the aging time increased from 4 days to 6 days, the maximum failure force increased by around 19 %. After 6 days, the To simulate the adhesive behavior, a sample under pull out test was modeled in ABAQUSTM v.6.12 software under 2D plane strain condition. There are several ways to simulate the adhesive layer such as using elastic elements [39-41], cohesive elements  and cohesive surfaces . However, the using elastic elements only show the stress information while the cohesive element and cohesive surface can simulate the 4801 N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804 100 (a) (a) Force (N) 80 (b) (b) 60 40 20 0 0 1 2 3 4 5 Tensile speed (mm/min) 6 Fig. 11. Deformation of a adhesive layer in the aluminum body frame just before failure (a) FEM; (b) experiment. Fig. 9. Effect of pull out speed on the fracture force. Adhesive layer Window Rod Sym. Clamp Body frame Fig. 10. Mesh of half model. delamination behavior. Therefore, in this study, the cohesive element COH2D4 (4-node two-dimensional cohesive element) was used for the adhesive layer. Meanwhile the glass window, frame, rod, and clamp were modeled as CPE4R element (4-node bilinear plane strain quadrilateral, reduced integration, hourglass control). The mesh of sample is shown in Fig. 10. The total number of elements is 1748. For the material properties of the adhesive layer, the mode-independent assumption in ABAQUSTM is used for mixed-mode behavior in damage evolution. Therefore, the fracture energies of 400 and 355 J/m2 were used for polycarbonate and aluminum cases. In the experiment, the case of the sample was fixed at the bottom using a special clamp. The clamp was modeled as a rectangle that was fixed all the degree of freedom. The rectangle interacted with the frame using contact conditions in ABAQUSTM. The rod was modeled as a circle in which the same velocity was applied as in the experiment. The layer of cohesive elements was tied at the top body frame and bottom of glass window using tie constraints . In the experiment, it is clear that the glass window was separated from the body frame from the bottom of adhesive layer, even though there was a small amount of damage on the top of the adhesive layer of the polycarbonate body frame. Therefore, an adhesive layer was modeled to simulate the delamination behavior between the glass window and body frame (Fig. 10). The adhesive stiffness, maximum stress, and fracture energy from experiment was used for cohesive elements (Table 1). The deformation of a adhesive layer in the aluminum body frame just Fig. 12. Mises stress of phone model for the aluminum body frame. before failure is shown in Fig. 11. The cohesive element was elongated until the stiffness degradation is equal to 1. If the stiffness degradation is euqal to 1, the cohesive element is deleted. It is obvious that the delamination behavior will occur from the corner (right to left). The simulation results show good agreement with the experimental results. The Fig. 12 shows the Misses stress of the model. The stress is concentrated on the body frame, while the maximum stress occurs at the contact point between the rod and glass window, as shown in Fig. 12. In order to compare the FEM results to the experimental results, the relationship between force and displacement was determined. The force-displacement curve from the FEM is illustrated in Fig. 13. It is clear that the FEM results were close to the average data from the experiment in terms of fracture stress. To investigate the effect of the properties of the adhesive layer, a parametric study was conducted in ABAQUSTM. To simplify the analysis, only the aluminum body frame was investigated, as shown in Fig. 14. When the stiffness of adhesive was increased two and three times (100 % and 200 % incensement), the initial damage displacement was decreased by around 44.4 % and 55.5 %, respectively. Meanwhile, the fracture load was increased by around 11.5 % and 9.4 %, respectively. As the fracture energy was increased 1.1 and 1.2 times (10 % and 20 % incensement), the initial damage displacement was increased 11.1 % for both cases. Meanwhile, the fracture load was increased by around 13 % and the damage displacement was decreased. Another parameter tested is changing the fracture energy of the adhesive layer. As the fracture stress of adhesive layer was increased 1.1 and 1.2 times (10 % and 20 % incensement), the fracture load was increased 21.1 % and 31.2 %, respectively. 4802 N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804 glass window was detached from the bottom of adhesive layer. Moreover, the fracture energy in the aluminum body frame is higher than that in the polycarbonate body frame. The pull out speed of testing and the aging time of adhesive layer influenced the fracture load. The fracture load is increased as the pull out speed is increased. Meanwhile, the aging time affected the fracture load in the first six days. In order to validate the experimental results, the cohesive elements in ABAQUSTM were used for the modeling bonding layer. The results showed good agreement between simulation and experiment. (a) Acknowledgment This work was also supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (NRF-2016R1A2B4007443). The authors are grateful for the financial support. Nomenclature-----------------------------------------------------------------------t tn ts en es (b) Fig. 13. Force versus displacement between experiment and FEM: (a) Aluminum body frame; (b) polycarbonate body frame. df 80 Original Stiffness x2 Stiffness x3 Fracture energy x1.1 Fracture energy x1.2 Stress x1.1 Stress x1.2 Force (N) 60 40 20 0 0.0 K Gnc Gsc 0.5 1.0 1.5 Displacement (mm) 2.0 2.5 Fig. 14. Effect of fracture stress, initial stiffness, and fracture energy on adhesive properties. 4. Conclusions In this study, adhesion strength and the detailed delamination process of the adhesive layer were investigated using a conventional pull out testing method and digital image correlation technique. The results showed that the strain at the interface between the adhesive layer and body frame is higher than that at the interface between adhesive layer and glass window for both the cases of aluminum and polycarbonate frames. The δ0 : The traction stress vector : Normal of traction stress vector : Shear of traction stress vector : Normal strain : Shear strain : Stiffness matrix : Critical fracture energy in normal direction : Critical fracture energy in shear direction : Failure separation displacement : Damage initiation References  J. Karppinen, J. Li, J. Pakarinen, T. T. Mattila and M. Paulasto-Kröckel, Shock impact reliability characterization of a handheld product in accelerated tests and use environment, Microelectronics Reliability, 52 (1) (2012) 190-198.  T. Mattila, L. Vajavaara, J. Hokka, E. Hussa, M. Mäkelä and V. Halkola, Evaluation of the drop response of handheld electronic products, Microelectronics Reliability, 54 (3) (2014) 601-609.  E. Tempelman, M. Dwaikat and C. Spitás, Experimental and analytical study of free-fall drop impact testing of portable products, Experimental Mechanics, 52 (9) (2012) 1385-1395.  C.-L. Hwan, M.-J. Lin, C.-C. Lo and W.-L. Chen, Drop tests and impact simulation for cell phones, Journal of the Chinese Institute of Engineers, 34 (3) (2011) 337-346.  A. L. Loureiro, L. F. M. da Silva, C. Sato and M. A. V. Figueiredo, Comparison of the mechanical behaviour between stiff and flexible adhesive joints for the automotive industry, The Journal of Adhesion, 86 (7) (2010) 765-787.  A. Krishnan and L. R. Xu, Systematic evaluation of bonding strengths and fracture toughnesses of adhesive joints, The Journal of Adhesion, 87 (1) (2011) 53-71. N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804  C. K. Kok, C. C. Ooi, J. J. Ng and K. W. Liew, Experimental study and modelling of an adhesive-bonded polycarbonate assembly for application in consumer electronics, The Journal of Adhesion, 89 (4) (2013) 247-263.  K. Zhao and L. R. Xu, Size effect of the adhesive bonding strengths of metal/polymer similar and dissimilar material joints, The Journal of Adhesion, 91 (12) (2015) 978-991.  L. F. M. da Silva, P. J. C. das Neves, R. D. Adams and J. K. Spelt, Analytical models of adhesively bonded joints—Part I: Literature survey, International Journal of Adhesion and Adhesives, 29 (3) (2009) 319-330.  R. D. Adams and V. Mallick, A method for the stress analysis of lap joints, The Journal of Adhesion, 38 (3-4) (1992) 199-217.  E. F. Karachalios, R. D. Adams and L. F. M. da Silva, Single lap joints loaded in tension with ductile steel adherends, International Journal of Adhesion and Adhesives, 43 (2013) 96-108.  E. F. Karachalios, R. D. Adams and L. F. M. da Silva, Single lap joints loaded in tension with high strength steel adherends, International Journal of Adhesion and Adhesives, 43 (2013) 81-95.  R. Créac'Hcadec, G. Jamin, J.-Y. Cognard and P. Jousset, Experimental analysis of the mechanical behaviour of a thick flexible adhesive under tensile/compression-shear loads, International Journal of Adhesion and Adhesives, 48 (2014) 258-267.  M. Kashfuddoja and M. Ramji, Assessment of local strain field in adhesive layer of an unsymmetrically repaired CFRP panel using digital image correlation, International Journal of Adhesion and Adhesives, 57 (2015) 57-69.  R. Fedele, B. Raka, F. Hild and S. Roux, Identification of adhesive properties in GLARE assemblies using digital image correlation, Journal of the Mechanics and Physics of Solids, 57 (7) (2009) 1003-1016.  A. Comer, K. Katnam, W. Stanley and T. Young, Characterising the behaviour of composite single lap bonded joints using digital image correlation, International Journal of Adhesion and Adhesives, 40 (2013) 215-223.  L. Blyberg, E. Serrano, B. Enquist and M. Sterley, Adhesive joints for structural timber/glass applications: Experimental testing and evaluation methods, International Journal of Adhesion and Adhesives, 35 (2012) 76-87.  T. Sadowski, M. Kneć and P. Golewski, Experimental investigations and numerical modelling of steel adhesive joints reinforced by rivets, International Journal of Adhesion and Adhesives, 30 (5) (2010) 338-346.  M. K. Budzik, J. Jumel and M. E. R. Shanahan, 4-Point beam tensile test on a soft adhesive, Materials & Design, 46 (2013) 134-141.  X. Yang, L. Yao, Y. Xia and Q. Zhou, Effect of base steels on mechanical behavior of adhesive joints with dissimilar steel substrates, International Journal of Adhesion and Adhesives, 51 (2014) 42-53.  C. D. M. Liljedahl, A. D. Crocombe, M. A. Wahab and I. A. 4803 Ashcroft, Modelling the environmental degradation of adhesively bonded aluminium and composite joints using a CZM approach, International Journal of Adhesion and Adhesives, 27 (6) (2007) 505-518.  P. Hu, X. Han, L. F. M. da Silva and W. D. Li, Strength prediction of adhesively bonded joints under cyclic thermal loading using a cohesive zone model, International Journal of Adhesion and Adhesives, 41 (2013) 6-15.  R. D. S. G. Campilho, M. D. Banea, A. M. G. Pinto, L. F. M. da Silva and A. M. P. de Jesus, Strength prediction of single- and double-lap joints by standard and extended finite element modelling, International Journal of Adhesion and Adhesives, 31 (5) (2011) 363-372.  W. Xu and Y. Wei, Strength and interface failure mechanism of adhesive joints, International Journal of Adhesion and Adhesives, 34 (2012) 80-92.  D. Systèmes, Abaqus analysis user’s manual, Simulia Corp. Providence, RI, USA (2007).  K. Song, C. G. Dávila and C. A. Rose, Guidelines and parameter selection for the simulation of progressive delamination, ABAQUS User’s Conference (2008) 43-44.  N. Carrere, C. Badulescu, J.-Y. Cognard and D. Leguillon, 3D models of specimens with a scarf joint to test the adhesive and cohesive multi-axial behavior of adhesives, International Journal of Adhesion and Adhesives, 62 (2015) 154164.  L. Liao, T. Sawa and C. Huang, Numerical analysis on load-bearing capacity and damage of double scarf adhesive joints subjected to combined loadings of tension and bending, International Journal of Adhesion and Adhesives, 53 (2014) 65-71.  K. Hasegawa, A. D. Crocombe, F. Coppuck, D. Jewel and S. Maher, Characterising bonded joints with a thick and flexible adhesive layer. Part 2: Modelling and prediction of structural joint responses, International Journal of Adhesion and Adhesives, 63 (2015) 158-165.  H. Bang, S.-K. Lee, C. Cho and J. U. Cho, Study on crack propagation of adhesively bonded DCB for aluminum foam using energy release rate, Journal of Mechanical Science and Technology, 29 (1) (2015) 45-50.  W. Xu and Y. Wei, Influence of adhesive thickness on local interface fracture and overall strength of metallic adhesive bonding structures, International Journal of Adhesion and Adhesives, 40 (2013) 158-167.  M. Elices, G. V. Guinea, J. Gómez and J. Planas, The cohesive zone model: Advantages, limitations and challenges, Engineering Fracture Mechanics, 69 (2) (2002) 137-163.  N. S. Ha, T. Jin and N. S. Goo, Modal analysis of an artificial wing mimicking an Allomyrina dichotoma beetle's hind wing for flapping-wing micro air vehicles by noncontact measurement techniques, Optics and Lasers in Engineering, 51 (5) (2013) 560-570.  V. T. Le, N. S. Ha, T. Jin, N. S. Goo and J. Y. Kim, Thermal interaction of a circular plate-ring structure using digital image correlation technique and infrared heating system, 4804 N. S. Ha et al. / Journal of Mechanical Science and Technology 31 (10) (2017) 4795~4804 Journal of Mechanical Science and Technology, 30 (9) (2016) 4363-4372.  T. Diehl, On using a penalty-based cohesive-zone finite element approach, Part I: Elastic solution benchmarks, International Journal of Adhesion and Adhesives, 28 (4) (2008) 237-255.  N. Ha, T. Jin, N. Goo and H. Park, Anisotropy and nonhomogeneity of an Allomyrina Dichotoma beetle hind wing membrane, Bioinspiration & Biomimetics, 6 (4) (2011) 046003.  Hibbett, Karlsson, Sorensen, ABAQUS/standard: User's Manual, Hibbitt, Karlsson & Sorensen (1998).  S. Hayashida, T. Sugaya, S. Kuramoto, C. Sato, A. Mihara and T. Onuma, Impact strength of joints bonded with highstrength pressure-sensitive adhesive, International Journal of Adhesion and Adhesives, 56 (2015) 61-72.  M. Mokhtari, K. Madani, M. Belhouari, S. Touzain, X. Feaugas and M. Ratwani, Effects of composite adherend properties on stresses in double lap bonded joints, Materials & Design, 44 (2013) 633-639.  M. Zehsaz, F. Vakili-Tahami and M.-A. Saeimi-Sadigh, Parametric study of the creep failure of double lap adhesively bonded joints, Materials & Design, 64 (2014) 520526.  M. Fari Bouanani, F. Benyahia, A. Albedah, A. Aid, B. Bachir Bouiadjra, M. Belhouari and T. Achour, Analysis of the adhesive failure in bonded composite repair of aircraft structures using modified damage zone theory, Materials & Design, 50 (2013) 433-439.  D. Van Truong, T. Kitamura and V. Van Thanh, Crack initiation strength of an interface between a submicron-thick film and a substrate, Materials & Design, 31 (3) (2010) 1450-1456.  M. K. Kim, D. J. Elder, C. H. Wang and S. Feih, Interaction of laminate damage and adhesive disbonding in composite scarf joints subjected to combined in-plane loading and impact, Composite Structures, 94 (3) (2012) 945-953. Ngoc San Ha graduated from department of Aeronautical Engineering of Ho Chi Minh City University of Technology, Viet Nam, 2008, and got Ph.D. degrees from Department of Advanced Technology Fusion, Konkuk University, Korea, 2014. Currently, he is a Research Professor at the Department of Advanced Technology Fusion, Konkuk University, Korea. His topics of interest primarily focus on bioinspired composite material, structural dynamics of small systems, material characterization, and digital image correlation application. Thanh Duc Dao graduated from department of Mechanical Engineering of Ho Chi Minh City University of Technology, Viet Nam, 2014, and got Master degree from Department of Advanced Technology Fusion, Konkuk University, Korea, 2016. He was a Research Professor at the Department of Advanced Technology Fusion, Konkuk University, Korea from September 2016 to April 2017. His topics of interest primarily focus on shape memory composite material, space deployable structure, material characterization, and digital image correlation application. Nam Seo Goo graduated from department of Aeronautics Engineering of Seoul National University with honors in 1990, and earned M.S. and Ph.D. degrees in Aerospace Engineering at the same university in 1992 and 1996, respectively. His Ph.D. was in the structural dynamics of aerospace systems. He is a professor in the Department of Advanced Technology Fusion at Konkuk University, Seoul, Korea. His current research interests are structural dynamics of small systems, smart structures and materials, and opto-mechanics Soonwan Chung graduated graduated from department of Aeronautics and Aerospace Engineering of Seoul National University in 1995, and earned M.S. and Ph.D. degrees in Aerospace Engineering at the same university in 1997 and 2002, respectively. His Ph.D. was related to the nonlinear continuum damage mechanics combined with parallel computing. He is a principal engineer in Global Technology Center of Samsung Electronics Co., Suwon, Korea. His current research interests are waterproof design, drop reliability and material/process optimization for adhesion enhancement.