The fracture behaviour of nanostructure added adhesives under ambient temperature and thermal cyclic conditions

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

Contents lists available at ScienceDirect

Theoretical and Applied Fracture Mechanics journal homepage: www.elsevier.com/locate/tafmec

The fracture behaviour of nanostructure added adhesives under ambient temperature and thermal cyclic conditions

T

⁎

Belin Kanara, Salih Akpinara, , Iclal Avinc Akpinarb, Hamit Akbulutc, Adnan Ozeld a

Dept. of Mechanical Eng., Erzurum Technical University, 25050 Erzurum, Turkey Independent Researcher, 25240 Erzurum, Turkey c Dept. of Mechanical Eng., Ataturk University, 25240 Erzurum, Turkey d Dept. of Mechanical Eng., Erzincan University, 24100 Erzincan, Turkey b

A R T I C LE I N FO

A B S T R A C T

Keywords: Nanocomposites Adhesive joints Fracture mechanics Numerical modelling Double cantilever beam test

In this study, the fracture behavior of nanocomposite adhesives produced by adding nanostructure in to the adhesives were investigated using Double Cantilever Beam (DCB) test under ambient temperature and thermal cycle conditions. Adhesively bonded DCB joints were produced using DP460 toughened adhesive type and DP125 flexible adhesive type as the adhesives; AA2024-T3 aluminum alloy was used as the adherend, and 1 wt. % Graphene-COOH, Carbon Nanotube-COOH and Fullerene C60 were used as the added nanostructures. As a result, when the experimental fracture energy was examined, the nanocomposite adhesives obtained by adding nanostructure were found to have increased the fracture energy of the joint. In the joints bonded with the flexible DP125 adhesive that were subjected to thermal cycles, the addition of Graphene-COOH into the adhesive increases the fracture energy of the joint by 55%, the addition of Fullerene increases the fracture energy of the joint by 135%. Also, it was observed that there is a significant difference between the displacements that were obtained directly from the test machines stroke and measured via video extensometer giving the crack opening between the top and bottom adherends during the DCB test. This situation significantly effects the correct calculation of the fracture energy of adhesive.

1. Introduction In recent years, adhesively bonded joints are frequently used the aviation and automotive industry. Therefore, published literature shows numerous studies that have been conducted to increase the loadcarrying capacities of adhesively bonded joints. The majority of these works have been aimed at increasing the load-carrying capacity of the joint by changing the joint geometry. However, studies carried out in recent years have been based on increasing the load-carrying capacity of the joint by adding nanostructures into an adhesive [1–13]. It is important to do experimental and numerical analyses of adhesively bonded joints, and look at the compatibility of the experimental and numerical data. In published literature, there are several methods used in the numerical analysis of adhesively bonded joints. However, the Cohesive Zone Model (CZM)—which is considered the most suitable—has been widely used for numerical analysis of adhesively bonded joints. In published literature, there are many studies conducting numerical analyses using CZM [14–19]. The CZM parameters of an adhesive must be determined to perform numerical

⁎

analyses of the CZM. There are several different methods for determining CZM parameters (Mode I-Double Cantilever Beam, Mode-II End Notch Flexure and Mixed Mode-Four Point Bend). One of the methods to determine CZM parameters is the Mode I-Double Cantilever Beam (DCB) test. In published literature, there are few studies that measure the fracture energy, as well as the displacement (opening between the lower and upper adherend) values corresponding to the maximum stress and failure of the adhesive, by conducting DCB tests of the adhesively bonded joint [20–27]. Some of them are summarized below. In a study by Khoramishad et al. [28], the influence of graphene oxide nano-platelets on fracture energy was investigated experimentally and numerically in adhesively bonded joints. According to the result of this study, adding 3 wt.% graphene oxide nano-platelets into the adhesive increases the fracture energy of the link by about 69%. In a study done by Kim et al. [29], an adhesive was reinforced with fiber, and its DCB test was conducted. A substantial increase was observed in the fracture energy of the fiber reinforced adhesive, and the amount of the increase varied with respect to the amount of fiber used

Corresponding author. E-mail address: [email protected] (S. Akpinar).

https://doi.org/10.1016/j.tafmec.2018.08.006 Received 17 May 2018; Received in revised form 7 August 2018; Accepted 8 August 2018 Available online 10 August 2018 0167-8442/ © 2018 Elsevier Ltd. All rights reserved.

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

Nomenclature DCB GIC CBT STM P δ B ɑ Δ

Double Cantilever Beam the tensile critical strain energy release rate Corrected Beam Theory Standard Test Method (for Fracture Strength in Cleavage of Adhesives in Bonded Metal Joints) applied load, N displacement at the point the load was applied, in mm specimen width, in mm crack length (from the point where the load was applied), in mm crack length correction for crack tip rotation and deflection (Δ is a function of C and ɑ, and can be found experimentally)

C Pmax E B h Kn

compliance of the specimen, δ/P load to start crack, N tensile modulus of adherend, MPa specimen width, mm thickness of adherend, mm max normal cohesive stiffness Tn ∗

Tnmax δn∗

maximum normal cohesive traction σmax normal displacement jump at maximum normal cohesive tractioncohesive traction normal displacement jump at the completion of debonding maximum normal displacement jump attained in deformation history damage parameter associated with Mode I dominated bilinear cohesive law.

δnc δnmax Dn

δn

g), 1 wt.% Carbon Nanotube-COOH (diameter: 10–20 nm; length: 10–30 mm; purity: 95%; surface area: 200 m2/g; 2 wt.% COOH content) and 1 wt.% Fullerene C60 (purity: 99%) were used. The material properties of the adhesives used in the experimental studies are shown in Table 1 [7]. Structural adhesives are subject to curing depending on the temperature and time. Curing conditions and composition rates of the adhesives are given in Table 2.

in the adhesive. There are several theories used to calculate the fracture energy of an adhesive by using the DCB test. Mohammadreza et al. [30] used the Modified Beam Theory (MBT) and Compliance Calibration Method (CCM) to calculate the fracture energy of the adhesive, comparing the results with 2D and 3D numerical analyses. According to the experimental and numerical analysis results, it was found that 3D analysis was more consistent. Lopes et al. [31] conducted DCB tests for three different adhesives, and the fracture energies of each adhesive were obtained by using three different energy methods: the Compliance Calibration Method, the Corrected Beam Theory and the Compliance-Based Beam Method. Based on the results of the study, the DCB test was deemed appropriate for soft and hard adhesives, while the Tapered Double-Cantilever Beam (TDCB) test was more appropriate for very hard adhesives. Furthermore, Blackman et al. [32] investigated the fracture behavior of the structure of adhesive joints using the DCB test specimens at test rates of 1 m/s up to 15 m/s. DCB test rates changed the fracture energy of the adhesive. Adhesively bonded joints are exposed to the thermal cycle like environment, warm and cold, it is crucial to investigate the thermal cycle performances. In this study, the fracture behavior of nanocomposite adhesives produced by adding nanostructure in to the adhesives were investigated using Double Cantilever Beam (DCB) test under ambient temperature and thermal cycle conditions. Adhesively bonded DCB joints were produced using DP460 toughened adhesive type (strain rate about 4.7%) and DP125 flexible adhesive type (strain rate about 78.5%) as the adhesives; AA2024-T3 aluminum alloy was used as the adherend, and 1 wt.% Graphene-COOH, Carbon Nanotube-COOH and Fullerene C60 were used as the added nanostructures. In the experiments while the crack growth is measured by using a high-speed video (HSV) camera, the displacements were measured by extensometer. The total thermal cycling operation comprises five cycles; for one cycle, the sample is held at 21 °C for 10 min/40 °C for 30 min/21 °C for 10 min and −50 °C for 30 min. The ambient temperature is set at 21 °C.

2.2. Fabrication of the Double Cantilever Beam (DCB) joints The adherend material used in the Double Cantilever Beam (DCB) was AA2024-T3 aluminum alloy, and the DCB joint samples’ geometry and dimensions are shown in Fig. 1. DCB joint samples’ geometry and dimensions are currently standardized [33,34]. In this study, the crack length is 55 mm (the part from where the load is applied to the beginning of the adhesive) and the pre-crack is not found. The most critical part in preparing adhesives reinforced with nanostructures is to homogeneously distribute the carbon nanostructures in the adhesive to prevent flocculation between nanostructures. In a study performed by Gültekin et al. [4], the standard deviation was decreased to 1–2% in the joints produced by using a new method. This new method was developed together with colleagues at the department of chemistry education. Considering this new method, both non-reinforced and nanostructure reinforced adhesives was prepared. A full discussion can be found elsewhere [7]. For the adhesively bonded joints to display their high performance, surface machining methods were applied to the adhesive (AA2024-T3 aluminum alloy) before bonding. To clean burrs and remove oil, grease and dirt resulting from cutting the specimens into the desired dimensions, specimens were first ground with 600 SiC sandpaper followed by 1000 SiC sandpaper to obtain a smooth surface. After grinding, specimens were washed under flowing water and kept in acetone for twenty minutes. The mold shown in Fig. 2 was used to produce the DCB joint sample with precision. The adhesive thickness in all joint types is 0.16 mm. In order to obtain an adhesive layer thickness of 0.16 mm after curing,

2. Experimental work 2.1. Materials

Table 1 Material properties of the adhesives [7].

AA2024-T3 aluminum alloy—commonly used in aerospace, transportation and general engineering industries due to its superior mechanical and physical properties—was used in this study as an adherend. For adhesion, two-part epoxy adhesives (produced by 3M Company, St. Paul, Minnesota, U.S.A.) DP460 toughened adhesive and DP125 flexible adhesive were used. For the nanostructure, 1 wt.% GrapheneCOOH (thickness: 5–7 nm; diameter: 5 µm; surface area: 120–150 m2/

E (MPa) ν σt (MPa) εt (%)

DP 460

DP 125

1984 ± 43 0.37 38.4 ± 1.1 4.7

25.1 ± 2 0.35 12.7 ± 0.4 78.5

E: Young’s modulus; ν: Poisson’s ratio; σt: Ultimate tensile strength; εt: Ultimate tensile strain. 121

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

2.4. Experimental parameters of the DCB joints

Table 2 Curing conditions and composition rates of the structural adhesives. Adhesive

Compound Rate (Epoxy:A/ Hardener:B)

Curing temperature/time

3M™ DP-460 3M™ DP-125

A:B = 2:1 A:B = 2:1

60 °C/120 min 70 °C/120 min

In the study, mechanical behaviors of DCB joint types exposed to tensile load were examined experimentally and numerically. For that, samples were designed in four main groups (DP460-ambient temperature, DP460-thermal, DP125-ambient temperature and DP125-thermal) and the experiment parameters of these samples are given in Table 3. Forty-eight DCB joint samples were prepared by preparing three samples for each parameter given in Table 3.

shims (see Fig. 2) were placed into the mold. Moreover, a hot press was used to provide the curing conditions for the structural adhesives shown in Table 2. Excess adhesive formed during the curing of joint specimens was cleaned from the joint surface (Fig. 3a). Later, a thin film layer was formed on the surface of the joint by spraying paint to help determine the crack growth (Fig. 3b). A measuring tape was bonded onto the surface of the specimen to determine the crack length (Fig. 3c).

2.5. Experimental tests for the DCB joints A Mode-I crack was applied to the DCB joints produced in the study, and the boundary conditions are shown in Fig. 5. All experiments were carried out under a 5 mm/min loading rate by using Shimadzu AG-IS 100 (Tokyo, Japan) test equipment with a 5 kN load cell in a laboratory with an ambient temperature of 21 °C and relative humidty of 30%. The boundary conditions and applied load were the same for all specimens (Type-I, Type-II, Type-III and Type-IV) (Fig. 6). Two cameras were used during the tests of DCB joint specimens (Fig. 6). Camera-1 (video extensometer) measures the displacement between the points A and B along the z-axis shown in Fig. 5, while camera-2 (model: Basler acA4600-10uc, made in Germany, resolution 4608 px × 3288 px, frame rate 10 fps, lens: 12 mm, size of the observed area: width 233 mm × height 167 mm, working distance: 445 mm, software: labview, two-dimensional analysis-2D) measures the crack propagation along the y-axis. Moreover, the displacement at the point the load was applied was obtained from the stroke of the test equipment. GIC (the tensile critical strain energy release rate) values of DCB joints were obtained according to two different standards. The first one is the Corrected Beam Theory (CBT) in the BS 7991 standard [34],

2.3. Ambient temperature and thermal cycling conditions for the DCB joints Adhesively bonded joints used in aerospace applications are subject to thermal loading. For example, it is known that when an airplane is airborne (10,000 feet) the temperature is approximately −60 °C, whereas while it is on the ground the temperature is approximately 35 °C (in summer). Therefore, it is critical to know the thermal behavior of adhesively bonded joints. According to published literature, a thermal cycle is usually applied mechanically by using a drying oven and a coolant. However, due to the automatic operation of the device used in this study, it is easier to get accurate results in terms of temperature and time than by hand. Some technical features of the thermal cycling device, as shown in Fig. 4a, are as follows. The device has three independent chambers (Fig. 4b–d). The device has a touch-screen PLC (programmable logic controller) and programmable features. Maximum number of programmable cycles for the device is 99 (Fig. 4f). Heating and cooling units, as well as the cycle and the timing are adjusted by a control unit. Maximum temperatures of the device are −60 ± 1 °C (Fig. 4b), ambient (∼25 °C) (Fig. 4c) and 250 ± 1 °C (Fig. 4d) for chambers 1, 2 and 3, respectively. The total thermal cycling operation comprises five cycles; for one cycle, the sample is held at 21 °C for 10 min/40 °C for 30 min/21 °C for 10 min and −50 °C for 30 min. The ambient temperature is set at 21 °C.

GIC =

3Pδ 2B (a + Δ)

(1)

The second one is ASTM 3433-99 - Standard Test Method for Fracture Strength in Cleavage of Adhesives in Bonded Metal Joints (STM) - [33],

GIC =

(4P 2max)(3a2 + h2) (EB2h3)

Fig. 1. DCB geometry used in the experimental investigation. 122

(2)

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

Fig. 2. The manufacturing mold of the joint samples.

Fig. 3. a. DCB joint specimen b. spraying paint on the surface and c. bonding a measuring tape on the surface.

3. Numerical work

analyses were the same as those used in the experimental studies. In three-dimensional (3D) analysis, the adherend section of the joint was modeled by using 20-node members through three degrees of freedom and Solid 186. Smaller elements were used in critical regions in terms of stress distribution (Fig. 7). When designing this model, mesh density was increased gradually. After a critical point (ideal mesh density), using a high number of meshes was seen to remain constant for stress analysis. Consequently, in order to decrease the required time for the numeric solutions, the mesh density at this critical point was chosen. The Cohesive Zone Model (CZM) was used for the adhesive region.

Four different types of DCB joints, Type-I Type-II, Type-III and TypeIV, were modeled in 3D using the ANSYS 16 software package [35], see Fig. 7. The numerical analyses of the DCB joints were performed using the non-linear finite element method (elastic-plastic strain analysis) by considering the non-linear (elastic-plastic) material behavior of adherend (AA2024-T3). The used material model is Multilinear Isotropic Hardening (MISO). The dimensions of specimens (Fig. 1), as well as their loading and boundary conditions (Fig. 7) used in finite element

Fig. 4. (a) Thermal cycling device, (b) Cold chamber, (c) Ambient chamber, (d) Hot chamber. 123

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

progression data for the DCB joint samples were measured. The fracture energy release rate (GIC) was obtained by using Eqs. (1) and (2). In all joints types, both the fracture energy–crack length curves and the mechanical properties were obtained by averaging the result of the three samples, and the standard deviation was found to be approximately from 1 to 1.5%, which indicates the homogenous distribution of nanostructures in the adhesive. In terms of thermal cycling, the minimum standard deviation values obtained from the three tests of the same sample revealed that it was more accurate to conduct thermal cycling in the fully automated thermal device. The result obtained also provides an opportunity to compare the fracture energy of the nonreinforced and nanoparticle-reinforced joints more accurately and make joints reliable and reproducible. The fracture energy-crack length curves of the DP460 toughened adhesive obtained from unreinforced and nanostructure-reinforced DCB joints (Type-I and Type-II) under ambient temperature and thermal cycle conditions are shown in Fig. 9. These curves were obtained using the Corrected Beam Theory (CBT). In DCB joints bonded with the DP460 adhesive, the crack length was recorded after the formation of the first crack. In other words, the 90 mm-propagation of the crack formed after the first 10 mm-adhesive length was investigated, and the fracture energy–crack length curves were obtained based on the data obtained. When curves obtained by the CBT are examined, nearly constant values of the energy during the 90 mm-propagation of the crack is related with finding the Δ value used in the Eq. (1) correctly. This phenomenon is proof of the accuracy of the experimental data. At ambient temperature, the addition of Graphene-COOH to the DP460 adhesive decreases the fracture energy of the joint, while the addition of Carbon Nanotube-COOH increases the fracture energy of the joint. The addition of Fullerene to the adhesive has almost no effect on the fracture energy of the joint (Fig. 9a). However, when joints are subjected to thermal cycles, this substantially changes. When the nonreinforced joint (Type-I) is subjected to thermal cycles, the fracture energy of the Type-II joint significantly decreases, while the addition of Graphene-COOH and Fullerene into the adhesive considerably increases the fracture energy of the joint (Type-IIG and Type-IIF) (Fig. 9b). It can be said that this phenomenon is related to the thermal conductivity coefficient of the nanostructure added into the adhesive. Pmax (N), δcn (mm), GIC-N/mm (CBT) and GIC-N/mm (Standard Test Method-STM) values obtained from DCB joints bonded with non-reinforced and nanostructure-reinforced DP460 adhesive are shown in Table 4. When these values are examined at ambient temperature, the addition of Graphene-COOH, Carbon Nanotube-COOH and Fullerene reinforcements into the DP460 adhesive increased the maximum load (Pmax) of the joint by 9%, 16% and 38%, respectively. When DCB joints bonded with non-reinforced DP460 adhesive were subjected to thermal cycles, the maximum load of the joint decreased by 34%. However, the addition of Fullerene into the adhesive increased the maximum load of the joint by 7% (Table 4). Moreover, when fracture energy data obtained from CBT was compared with that obtained from the STM, the fracture energy obtained from the STM was higher by between 8 and 26%. The reason for that is that the STM does not consider the minimum elastic recovery caused by the thickness of the adherend.

Table 3 Experimental parameters used in the experimental investigation. TYPE

Adhesive/Reinforcement/Conditions

Type-I Type-IG Type-IC Type-IF

DP460/Non-reinforced/Ambient temperature DP460/Graphene-COOH/Ambient temperature DP460/Carbon Nanotube-COOH/Ambient temperature DP460/Fullerene/Ambient temperature

Type-II Type-IIG Type-IIC Type-IIF

DP460/Non-reinforced/Thermal DP460/Graphene-COOH/Thermal DP460/Carbon Nanotube-COOH/Thermal DP460/Fullerene/Thermal

Type-III Type-IIIG Type-IIIC Type-IIIF

DP125/Non-reinforced/Ambient temperature DP125/Graphene-COOH/Ambient temperature DP125/Carbon Nanotube-COOH/Ambient temperature DP125/Fullerene/Ambient temperature

Type-IV Type-IVG Type-IVC Type-IVF

DP125/Non-reinforced/Thermal DP125/Graphene-COOH/Thermal DP125/Carbon Nanotube-COOH/Thermal DP125/Fullerene/Thermal

Because the adhesive was only subject to peel forces in the DCB joint, the Mode I Dominated Bilinear CZM Model was selected from the ANSYS 16 program package. Two parameters (Tnmax = maximum normal cohesive traction σmax and δn∗ = normal displacement jump at maximum normal cohesive traction) must be found for CZM. These parameters are determined by Pmax and δcn obtained from the experiments of DCB joints. The Mode I dominated bilinear CZM model assumes that the separation of the material interfaces is dominated by the displacement jump that is normal to the interface, as shown in the following Fig. 8: The relation between tangential cohesive traction Tn and tangential displacement jump δn can be expressed as Eq. (1); (3)

Tn = Kn δn (1−Dn )

Dn-damage parameter- in the Eq. (1) was obtained from the Eq. (2).

⎧0 ⎪ δnmax ≤ δn∗ Dn = max ⎨ δn ⎪1 ⎩

(

δnmax ≤ δn∗

)(

δnc

δnc − δn∗

)

δn∗ < δnmax ≤ δnc δnmax > δnc

(4)

In the numerical analyses of the DCB joints, the layer between the adherend and the adhesive was perfectly bonded to each of them and the interlayer of the adhesive was modeled as a cohesive member. The fracture energy (GIC) was obtained by using the load–displacement and crack length data obtained from the experiments and the equations in Eqs. (1) and (2). 4. Results and discussions 4.1. Experimental results Three samples of Type-I G, C and F, Type-II G, C and F, Type-III G, C and F, and Type-IV G, C and F were each subjected to Mode-I loading until failure. Afterwards, the loads (P)–displacements (δ) and crack

Fig. 5. Mode-I test boundary conditions. 124

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

Fig. 6. Video extensometer, high-speed video camera and loading used in the experiments.

However, during testing, elastic recovery occurs in the adherend, and the amount of elastic recovery affects the displacement (δ) between the two materials. CBT considers this elastic recovery (Δ = crack length correction for crack tip rotation and deflection). Therefore, it can be concluded that fracture energy results obtained from the CBT are more accurate. Examining the fracture energy data obtained from the CBT at ambient temperature, the addition of Graphene-COOH into the DP460 adhesive decreased the fracture energy of the joint by 35%, while the addition of Carbon Nanotube-COOH and Fullerene increased the fracture energy of the joint by 29% and 20%, respectively. When DCB joints bonded with the non-reinforced DP460 adhesive were subjected to thermal cycles, the fracture energy of the joint decreased by 52%, while the addition of Fullerene increased the fracture energy of the joint by 44% (Table 4). The fracture energy-crack length curves of the DP125 flexible adhesive obtained from unreinforced and nanostructure-reinforced DCB joints (Type-III and Type-IV) under ambient temperature and thermal cycle conditions are shown in Fig. 10. These curves were obtained according to the CBT and the mechanical properties of the joints produced by the DP125 adhesive are shown in Table 5. The addition of Carbon Nanotube-COOH into the DP125 adhesive at ambient temperature decreased the fracture energy of the joint by 8%, while the addition of Fullerene increased the fracture energy of the joint by 91% (Fig. 10a and Table 5). In the joints subjected to thermal cycles (Fig. 10b), the addition of Graphene-COOH into the DP125 adhesive increased the fracture energy of the joint by 50%, while the addition of Fullerene increased the fracture energy of the joint by 127% (Table 5). However, the addition of Carbon Nanotube-COOH into the adhesive decreased the fracture energy of the joint by 24%. According to the obtained experimental data, both the joints obtained with DP460 toughened adhesive and the joints obtained with DP125 flexible adhesive are subjected to thermal cycling, the fracture

Fig. 8. Mode I Dominated Bilinear CZM Law [35].

energy of the joints is reduced. However, when 1 wt.% Fullerene is added to these two adhesives, the fracture energy of the joint increases significantly. Damage modes in joints Type-I, Type-II, Type-III and Type-IV were identified and given in the Fig. 11. By examining defects according to the deformation types specified in ISO 10365 standards [36], in Type-I, Type-IC, Type-II and Type-IIC specimens, adhesion failure (AF) and special cohesion failure (SCF) is observed when the thickness of the adhesive layer is not equal to the adherends (Fig. 11a). Therefore, it can be said that cohesion failure (CF) was observed in Type-IG, Type-IF, Type-IIG and Type-IIF specimens when the adhesive layer was present on the top and bottom members (Fig. 11a). Moreover, adhesion failure and special cohesion failure occurred in all joints bonded with the DP125 adhesive (Fig. 11b).

Fig. 7. Finite element analyses models of DCB joints. 125

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

5

5

4.5

4.5

4

4

3.5

3.5

3

GÕF1PP

*ÕF1PP

B. Kanar et al.

2.5 2 1.5

3 2.5 2 1.5

Type-I Type-IG Type-IC Type-IF

1 0.5 0 55

65

75

85

Type-II Type-IIG Type-IIC Type-IIF

1 0.5 0

95 105 115 125 135 145 155

55

DFUDFNlengthPP

65

75

85

95

105 115 125 135 145 155

D-FUDFNlengthPP

D 

E 

Fig. 9. Fracture energy–crack length curves of DCB joints bonded with the DP460 adhesive based on the CBT: (a) ambient temperature (b) after thermal cycles. Table 4 Mechanical properties of DCB joints produced by the DP460.

Table 5 Mechanical properties of DCB joints produced by DP125.

δnc (mm)

GIC-N/mm (CBT)

GIC-N/mm (STM)

SampleType

P

2006 2192 2334 2767

0.142 0.128 0.181 0.195

2.56 1.65 3.30 3.07

2.78 2.53 3.47 3.89

Type Type Type Type

III III G III C III F

1494 1997 2380 2965

0.133 0.141 0.167 0.229

1.21 2.18 2.65 4.44

1.34 2.32 2.98 4.79

Type Type Type Type

IV IV G IV C IV F

SampleType

P

Type Type Type Type

I IG IC IF

Type Type Type Type

II II G II C II F

max

(N)

2.4

2.4

2.1

2.1

1.8

1.8

1.5

1.5

GÕF1PP

GÕF1PP

The numerical analysis results were compared with the experimental results for all joint types. However, in this study, comparison results were presented for only Type-I and Type-III joint types. In the numerical analyses, the point C was fixed, a displacement was introduced from the point D across the +z direction (Fig. 12). While comparing the experimental and numerical analyses shown in Figs. 13 and 14, in the experiments of DCB joints, displacements at the point—where the load was applied—were found by using the displacement values obtained from A and B points with extensometer (Fig. 5) and the obtained values were used. The reason for this is that the sliding between the material and the clamp holding it and the gap at

1.2 0.9

0 55

65

75

85

GIC-N/mm (STM)

1014 1240 1043 1510

0.242 0.301 0.189 0.364

0.92 0.97 0.85 1.76

0.79 1.58 0.70 1.90

910 1396 921 1950

0.185 0.341 0.140 0.383

0.65 1.38 0.70 2.09

0.93 1.54 0.63 3.18

1.2 0.9 0.6

Type-III Type-IIIG Type-IIIC Type-IIIF

0.3

GIC-N/mm (CBT)

(N)

the connection points of the equipment change the displacements obtained from the stroke of the equipment. Therefore, a 20% difference is found between the displacement obtained from the stroke of the equipment and that obtained from the numerical analysis (the amount of difference can change with respect to the loading capacity of the joint). It can be seen from Fig. 13a and 14a that the load–displacement curves obtained from non-reinforced joints and those obtained from numerical analyses are significantly consistent with each other (approximately 98–99%). The consistency, on the other hand, varies from 95 to 98.5% in nanostructure-reinforced joints (Figs. 13b–d and 14b–d). As a result, data obtained from numerical analyses strongly confirms the accuracy of experiments and the use of the Mode I Dominated

4.2. Numerical analysis results and comparison with the experimental ones

0.6

δnc (mm)

max

Type-IV Type-IVG Type-IVC Type-IVF

0.3 0

95 105 115 125 135 145 155

D-FUDFNlengthPP

55

65

75

85

95 105 115 125 135 145 155

D-FUDFNlengthPP

E 

D 

Fig. 10. CBT-based fracture energy–crack length curves of DCB joints bonded with the DP125 adhesive: (a) ambient temperature (b) after thermal cycles. 126

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

Fig. 11. Surface damages of the samples examined in this study, (a) DP460, (b) DP125.

127

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

Fig. 12. Numerical analysis of DCB joints.

about 30%.

Bilinear Cohesive Zone Model in numerical analyses is the right method.

• During experiments, elastic recovery occurs in the adherend elastic;

5. Conclusions In this study, the mechanical behaviors of four different Double Cantilever Beam (DCB) joint types subjected to tensile loading were numerically and experimentally investigated. The results obtained from the study were as follows:

•

• In all joints types, both the fracture energy–crack length curves and •

•

the mechanical properties were obtained by averaging the result of the three samples, and the standard deviation was found to be approximately from 1 to 1.5%, which indicates the homogenous distribution of nanostructures in the adhesive. When non-reinforced joints (Type-I) are subjected to thermal cycles (Type-II), the fracture energy of the joint reduced by about 53%. Furthermore, when non-reinforced joints (Type-III) are subjected to thermal cycles (Type-IV), the fracture energy of the joint reduced by

•

and the amount of elastic recovery affects the displacement value (δ) between the two materials. Thus, according to the STM, it can be concluded that the fracture energy data obtained from the CBT are more accurate. The addition of Graphene-COOH, Carbon-Nanotube-COOH and Fullerene into the DP460 adhesive at ambient temperature increased the maximum load of the joint by 9%, 16% and 38%, respectively. When non-reinforced joints (Type-I) are subjected to thermal cycles, the fracture energy of the joint (Type-II) significantly decreases, while the addition of Graphene-COOH and Fullerene into the adhesive substantially increases the fracture energy of the joint (TypeIIG and Type-IIF). It can be said that this phenomenon is related to the thermal conductivity coefficient of the nanostructure added into the adhesive. At ambient temperature, the addition of Carbon Nanotube-COOH into the flexible DP125 adhesive decreases the fracture energy of the

Fig. 13. Numerical analysis results and comparison with the experimental results: (a) Type-I, (b) Type-IG, (c) Type-IC, (d) Type-IF. 128

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

Fig. 14. Numerical analysis results and comparison with the experimental results: (a) Type-III, (b) Type-IIIG, (c) Type-IIIC, (d) Type-IIIF.

• •

joint by 8%, while the addition of Fullerene increases the fracture energy of the joint by 91%. In the joints bonded with the flexible DP125 adhesive that were subjected to thermal cycles, the addition of Graphene-COOH into the adhesive increases the fracture energy of the joint by 50%, the addition of Fullerene increases the fracture energy of the joint by 127%. It is significant that the use of displacement values obtained from DCB joint experiments using an extensometer are consistent when a comparison is made between the experimental and numerical analyses. Moreover, the use of Mode I Dominated Bilinear Cohesive Zone Model in numerical analyses is the right method.

[3]

[4]

[5]

[6]

Acknowledgement

[7]

This study was financially supported by the Scientific and Technological Research Council of Turkey (TUBITAK) with a project number of 114M408.

[8] [9]

Appendix A. Supplementary material

[10]

Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.tafmec.2018.08.006.

[11]

References

[12]

[1] V.K. Srivastava, Effect of carbon nanotubes on the strength of adhesive lap joints of C/C and C/C-SiC ceramic fibre composites, Int. J. Adhes. Adhes. 31 (2011) 486–489. [2] H. Khoramishad, R.S. Ashofteh, M. Mobasheri, F. Berto, Temperature dependence of

[13]

129

the shear strength in adhesively bonded joints reinforced with multi-walled carbon nanotubes, Eng. Fract. Mech. (2018), https://doi.org/10.1016/j.engfracmech.2018. 05.032. S.M.J. Razavia, M.R. Ayatollahi, A. Nemati Giv, H. Khoramishad, Single lap joints bonded with structural adhesives reinforced with a mixture of silica nanoparticles and multi walled carbon nanotubes, Int. J. Adhes. Adhes. 80 (2018) 76–86. K. Gültekin, S. Akpinar, A. Gürses, Z. Eroglu, S. Cam, H. Akbulut, Z. Keskin, A. Ozel, The effects of graphene nanostructure reinforcement on the adhesive method and the graphene reinforcement ratio on the failure load in adhesively bonded joints, Compos. Part B 98 (2016) 362–369. M.R. Ayatollahi, A. Nemati Giv, S.M.J. Razavi, H. Khoramishad, Mechanical properties of adhesively single lap bonded joints reinforced with multi-walled carbon nanotubes and silica nanoparticles, J. Adhes. 93 (2017) 896–913. H. Khoramishad, R.S. Ashofteh, H. Pourang, F. Berto, Experimental investigation of the influence of temperature on the reinforcing effect of graphene oxide nanoplatelet on nanocomposite adhesively bonded joints, Theor. Appl. Fract. Mech. 94 (2018) 95–100. I.A. Akpinar, K. Gültekin, S. Akpinar, H. Akbulut, A. Ozel, Experimental analysis on the single-lap joints bonded by a nanocomposite adhesives which obtained by adding nanostructures, Compos. Part B 110 (2017) 420–428. M.H. Kang, J.H. Choi, J.H. Kweon, Fatigue life evaluation and crack detection of the adhesive joint with carbon nanotubes, Compos. Struct. 108 (2014) 417–422. H. Khoramishad, R.S. Ashofteh, Influence of multi-walled carbon nanotubes on creep behavior of adhesively bonded joints subjected to elevated temperatures, J. Adhes. (2018), https://doi.org/10.1080/00218464.2018.1451333. H. Khoramishad, O. Alizadeh, L.F.M. da Silva, Effect of multi-walled carbon nanotubes and silicon carbide nanoparticles on the deleterious influence of water absorption in adhesively bonded joints, J. Adhes. Sci. Technol. 32 (2018) 1795–1808. K.T. Hsiao, J. Alms, S.G. Advani, Use of epoxy/multiwalled carbon nanotubes as adhesives to join graphite fibre reinforced polymer composites, Nanotechnology 14 (2003) 791–793. H. Khoramishad, O. Alizadeh, Effects of silicon carbide nanoparticles and multiwalled carbon nanotubes on water uptake and resultant mechanical properties degradation of polymer nanocomposites immersed in hot water, Polym. Compos. (2017), https://doi.org/10.1002/pc.24303. L. Zhai, G. Ling, J. Li, Y. Wang, The effect of nanoparticles on the adhesion of epoxy adhesive, Mater. Lett. 60 (2006) 3031–3033.

Theoretical and Applied Fracture Mechanics 97 (2018) 120–130

B. Kanar et al.

[14] R.D.S.G. Campilho, M.D. Banea, J.A.B.P. Neto, L.F.M. da Silva, Modelling adhesive joints with cohesive zone models: effect of the cohesive law shape of the adhesive layer, Int. J. Adhes. Adhes. 44 (2013) 48–56. [15] I. Georgiou, H. Hadavinia, A. Ivankovic, A.J. Kinloch, V. Tropsa, J.G. Williams, Cohesive zone models and the plastically deforming peel test, J. Adhes. 79 (2003) 239–265. [16] L. Weidong, Z. Shenguang, S. Zhiwei, W. Xiaoxiao, H. Ping, Experimental and numerical analysis on fatigue durability of single-lap joints under vibration loads, J. Adhes. 93 (2017) 187–203. [17] X. Qiang, N. Guodong, S. Yejie, Q. Shaoxing, P. Hua-Xin, Numerical investigation on the loading-delamination-unloading behavior of adhesive joints, Compos. Part A 90 (2016) 45–50. [18] L. Haibo, Y. Ying, Z. Taotao, L. Zudian, Progressive failure and experimental study of adhesively bonded composite single-lap joints subjected to axial tensile loads, J. Adhes. Sci. Technol. 30 (2016) 894–914. [19] T.E.A. Ribeiro, R.D.S.G. Campilho, L.F.M. da Silva, L. Goglio, Damage analysis of composite–aluminium adhesively-bonded single-lap joints, Compos. Struct. 136 (2016) 25–33. [20] J. Zhengwen, W. Shui, L. Minghong, M. Lei, Analytical solutions for non-uniformity of energy release rate of orthotropic double cantilever beam specimens with an adhesive layer, Eng. Fract. Mech. 164 (2016) 49–59. [21] S. Sylwester, Numerical analysis of the DCB test configuration applicability to mechanically coupled Fiber Reinforced Laminated Composite beams, Compos. Struct. 152 (2016) 477–487. [22] H. Khoramishad, M. Khakzad, Toughening epoxy adhesives with multi-walled carbon nanotubes, J. Adhes. 94 (2018) 15–29. [23] S. Yu, K. Masato, S. Chiaki, Experimental study of the Mode I adhesive fracture energy in DCB specimens bonded with a polyurethane adhesive, J. Adhes. 93 (2017) 235–255. [24] S. Carlos, A.L. Frank, T. Albert, Finite-thickness cohesive elements for modeling thick adhesives, Eng. Fract. Mech. 168 (2016) 105–113. [25] D. Álvarez, B.R.K. Blackman, F.J. Guild, A.J. Kinloch, Mode I fracture in adhesively-

[26]

[27] [28]

[29]

[30]

[31]

[32]

[33]

[34]

[35] [36]

130

bonded joints: a mesh-size independent modelling approach using cohesive elements, Eng. Fract. Mech. 115 (2014) 73–95. M.D. Banea, L.F.M. da Silva, R.D.S.G. Campilho, The effect of adhesive thickness on the mechanical behavior of a structural polyurethane adhesive, J. Adhes. 91 (2015) 331–346. G.F. Dias, M.F.S.F. de Moura, J.A.G. Chousal, J. Xavier, Cohesive laws of composite bonded joints under mode I loading, Compos. Struct. 106 (2013) 646–652. H. Khoramishad, M. Ebrahimijamal, M. Fasihi, The effect of graphene oxide nanoplatelets on fracture behavior of adhesively bonded joints, Fatigue Fract. Eng. Mater. Struct. 40 (2017) 1905–1916. J.G. Kim, Y.J. Hwang, S.H. Yoon, D.G. Lee, Improvement of the fracture toughness of adhesively bonded stainless steel joints with aramid fibers at cryogenic temperatures, Compos. Struct. 94 (2012) 2982–2989. K. Mohammadreza, A.M. Farhad, Fracture analysis in adhesive composite material/ aluminum joints under mode-I loading; experimental and numerical approaches, Int. J. Adhes. Adhes 39 (2012) 8–14. R.M. Lopes, R.D.S.G. Campilho, F.J.G. da Silva, T.M.S. Faneco, Comparative evaluation of the Double-Cantilever Beam and Tapered Double-Cantilever Beam tests for estimation of the tensile fracture toughness of adhesive joints, Int. J. Adhes. Adhes. 67 (2016) 103–111. B.R.K. Blackman, A.J. Kinloch, F.S. Rodriguez Sanchez, W.S. Teo, J.G. Williams, The fracture behaviour of structural adhesives under high rates of testing, Eng. Fract. Mech. 76 (2009) 2868–2889. ASTM 3433-99, Standard Test Method for Fracture Strength in Cleavage of Adhesives in Bonded Metal Joints, ASTM International, West Conshohocken, PA, 2012. BS 7991:2001 standard, Determination of the mode I adhesive fracture energy, GIC, of structural adhesives using the double cantilever beam (DCB) and the tapered double cantilever beam (TDCB) specimens, 2001. ANSYS, The General Purpose Finite Element Software (Version 14.0), Swanson Analysis Systems Inc., Houston, Texas, 2012. ISO 10365, 1992(E) Adhesives – designation of main failure patterns.