Journal of Solid Mechanics and Materials Engineering

Journal of Solid Mechanics and Materials Engineering Vol. 4, No. 7, 2010 Fracture Toughness Test of Epoxy Adhesive Dissimilar Joint with Various Adh...
Author: Chester McBride
3 downloads 3 Views 407KB Size
Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

Fracture Toughness Test of Epoxy Adhesive Dissimilar Joint with Various Adhesive Thicknesses* Mohd AFENDI**, *** and Tokuo TERAMOTO** **Department of Engineering Mechanics and Energy Graduate School of Systems and Information Engineering University of Tsukuba, Tennodai 1-1-1, Tsukuba, Ibaraki, 305-8573, Japan E-mail: [email protected] ***School of Mechatronic Engineering Universiti Malaysia Perlis, 02600, Arau, Malaysia

Abstract In this study, effect of bond thickness upon the strength and fracture toughness of epoxy adhesive dissimilar joint is investigated. Tensile and three-point-bending (abbreviated as 3PB hereafter) fracture tests are conducted. Finite element method (abbreviated as FEM hereafter) analysis is also executed to analyze the stress distribution at an interface corner of dissimilar joint. From FEM analysis results, it is found that the stress singularity in the dissimilar joint exists pronouncedly at the SUS304/adhesive interface corner and the order of stress singularity in the tensile model is higher than that in the 3PB model. Moreover, the order of stress singularity in the dissimilar joint having bond thickness of 1.0 mm is quite identical to the value obtained from analytical solution under the plane stress condition. From 3PB test and tensile test, it has been confirmed that the failure stress of dissimilar joint slightly increases with the decreasing bond thickness and can be well predicted by using the interface corner toughness, Hc parameter. The failure of dissimilar joints always originates from the SUS304/adhesive interface corner and the failure stress for dissimilar joint of 3PB test is higher than that of tensile test. For the specimens failed at the ALU/adhesive interface corner, the poor wettability of ALU adherend’s surface plays an important role. For the dissimilar joint with an interfacial crack, the fracture toughness, Jc is calculated by J integral method in FEM analysis. Fracture toughness, Jc for cohesively fractured specimens is more or less constant but shows some dependency on bond thickness for interfacially fractured specimens. Locus of fracture can be best interpreted in terms of stress singularity order at the interfacial crack tip. Key words: Adhesive Joint, Tensile, Three-Point-Bending, Stress Singularity, Joint Strength, Fracture Toughness, Interface Fracture, Cohesive Fracture

1. Introduction

*Received 16 Nov., 2009 (No. e50) [DOI: 10.1299/jmmp.4.999]

Copyright © 2010 by JSME

Adhesive joint is definitely the ideal substitute for any conventional bonding methods (e.g. rivet, welding, diffusion bonding, etc.) in structural engineering applications, particularly in dissimilar materials joining[1]. Nevertheless, adhesive joint inevitably contains flaws or discontinuities at the interfaces. Moreover, stress singularity which develops at the interface corner due to elastic mismatches may initiate failure. As such, adhesive joints often fail unexpectedly and severely under a relatively low mechanical or thermal load in service. In the literature, numerous works have been directed on determining strength and fracture behavior of similar material sandwiched joint. These

999

999

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

include investigations upon the effect of bond thickness[2-4], crack path propagation[5-7] and assessment of fracture initiation criteria[8-12]. It has been reported that the strength and fracture behavior of the rubber modified adhesive joint is greatly dependent on the adhesive bond thickness and existence of cracks or flaws[2,3]. However, the mechanisms of the dependency for brittle adhesive joint are not yet clarified and study on dissimilar materials sandwiched joint is hardly available thus motivated this work. In order to ensure high reliability and significant strength performance of adhesive joints, the strength and fracture toughness of adhesive joints should first be properly determined. Hence, in this study, the strength of epoxy adhesive-bonded butt joints of dissimilar adherends under the tensile load and 3PB was examined on several adhesive bond thicknesses. Since the singularity due to the mismatch of materials at the interface corner of adhesive joint also contributes to interfacial crack growth, fracture toughness of adhesive joints with an interfacial crack was also evaluated. From the experimental and FEM analysis results, the fracture mechanisms of dissimilar adherends bonded joints will be qualitatively discussed.

2. Methods The epoxy adhesive resin used in this study was Hi-Super 30 produced by Cemedine Co., Japan. This is a commercial brittle epoxy adhesive which can be cured at room temperature approximately in 30 minutes. The adhesive was prepared prior to bonding by mixing the epoxy resin and hardener with the conditioning mixer for 1 min: 3min schedule of diffusion and de-foaming, respectively. The mechanical properties of the bulk epoxy adhesive have been reported in the previous study[13]. The pertinent mechanical properties of epoxy adhesive and adherends are given in Table 1. Tensile and 3PB test specimens were prepared to obtain the strength and fracture toughness of adhesive joints. The dimensions of 3PB test specimens and tensile are shown in Fig. 1(a) and (b), respectively. The adherends were consisted of SUS304 stainless steel and YH75 aluminium alloy. YH75 is the trade name of aluminium alloy which is identical to A7075P-T651 aluminium alloy in terms of mechanical properties according to JIS H 4000:2006 specification[14]. Adhesive bond thickness, t inside an adhesive joint was controlled by using a developed fixture and varied from 0.1 mm to 1.5 mm. For specimen with interfacial crack, the interfacial crack was introduced by pasting a strip with 0.05 mm thick Teflon tape on the adherend surface prior to bonding. Three different interfacial crack length, a were chosen to study their effects on each test. Thus, a/W was given as 1/8, 1/4 and 3/8, where W was the width of the specimen. Tensile and 3PB fracture tests of adhesive joints were carried out with a universal tensile test machine (INSTRON 4206) and a compact 3PB test machine (Little Senstar LSC-1/30), respectively. Both fracture tests were conducted at room temperature with the crosshead speed of 1.0 mm/min. 2D non-linear elastic FEM analysis was performed using ANSYS 11 code. The true stress-strain curve as shown in Fig. 2 was extrapolated from the actual uniaxial tensile test data to constitute the adhesive layer in FEM model[13]. The adherends were assumed to remain elastic materials and the data of mechanical properties were taken from Table 1. We Table 1 Mechanical properties of adhesive and adherends Material

Elastic modulus / GPa

Yield strength / MPa

Poisson’s ratio

Epoxy[13]

3.4

36.5

0.396

SUS304

206

295

0.3

ALU (YH75)

71

505

0.33

1000 1000

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

x (a) 3PB

SUS304

16

YH75

70

(b) Tensile

y 8

adhesive layer

20

SUS304

40 18

x

YH75

Φ 16

y t

5

Fig. 1 Geometry of specimens. The span of 3PB is 64 mm 60 50 15

σ true (MPa)

40 40

30 30

11

20 10

0

2

0.01

0.02

ε true

0.03

0.04

Fig. 2 True stress-strain curve of bulk epoxy adhesive employed internal multipoint constraint (abbreviated as MPC hereafter) approach to define the contact assembly in FEM model of adhesive joint. These MPC elements ignore any friction and the interaction between adhesive and adherend is always bonded (i.e. no separation at the interface). With this feature, the stress of each interface nodes can be obtained from its closest integration point. 2D-FEM simulations were carried out to investigate the stress-y distribution along the joint interfaces and near the interface corner region, and also to evaluate the fracture toughness, Jc of FEM model with an interfacial crack.

3. Results and discussion 3.1 Strength of adhesive dissimilar joint without crack First, the stress-y distribution in the adhesive joint of tensile and 3PB was analyzed by FEM analysis. Figure 3 shows the stress-y distribution near the interface corner in (a) tensile, and (b) 3PB adhesive joint of 1.0 mm bond thickness with similar and dissimilar adherends. The applied surface tensile stress, σ0 was 1 MPa and only plane stress condition was considered. The finest mesh size of 0.01 x 0.01 mm was used. SES and AEA refer to the adhesive similar joint of SUS304 stainless steel and YH75 aluminium alloy adherends, respectively. SEA refers to the adhesive dissimilar joint bonded with SUS304 and YH75 aluminium alloy. In the case of the tensile model, at the vicinity of the interface corner (i.e. r < 0.01 mm), we noticed that the stress-y elevates proportionally up to three times of the

1001 1001

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

applied stress. However, in the 3PB model, instead of linear increase, when r approaches the interface corner edge (i.e. r → 0), stress concentration is seen to be relatively low compared with the results of tensile model. The interface corner that has higher stress concentration in order is SUS304/adhesive of dissimilar joint, SUS304/adhesive of similar joint, ALU/adhesive of similar joint and ALU/adhesive of dissimilar joint. This is in close agreement with the results obtained by Kyogoku et al.[15]. In their study, they have examined both experimentally and analytically the strength of scarf joints of dissimilar adherends bonded with epoxy adhesive. They reported that the deformation state and fracture process of these joints were remarkably different from those joints of identical adherends. The mesh size constructed in FEM analysis is crucial and may affect the stress distribution near the region of interface corner. Therefore, we re-analyzed the above FE model but with a coarser mesh size of 0.1 x 0.1 mm to study its effect. We noticed a moderate change in stress-y value and found that it was difficult to evaluate the interface corner stress singularity, λ if coarse mesh size was used. Therefore, the dissimilar joint FE model with the highly refined mesh size of 0.01 x 0.01 mm will be used throughout this study. The comparison between the stress-y distribution at the SUS304/adhesive interface corner in the dissimilar joint FE model of tensile and 3PB specimen using mesh size of 0.01 x 0.01 mm is shown in Figure 4 (a). In the 3PB model, only half of the cross section is subjected to tensile stress while the other half is in compressive stress states. Thus, 3PB dissimilar joint is estimated to gain higher failure stress than that of tensile dissimilar joint. The slope of plots within the range of 0.001 mm ≤ r ≤ 0.01 mm in log-log scale is nearly equal to the interface corner stress singularity, λ as shown in Figure 4 (b). From this figure, obviously the order of stress singularity in the tensile dissimilar joint is higher than that in the 3PB. The stress singularity obtained from FEM analysis was verified with that obtained analytically from Bogy’s characteristic equation[16], as listed in Table 2. From asymptotic solution of Bogy, assuming the plane stress condition, we measured the order of stress singularity, λ at the interface corner of SUS304/adhesive and ALU/adhesive as 0.2613 and 0.2321, respectively. Therefore, the order of stress singularity in the tensile dissimilar joint having bond thickness of 1.0 mm is quite close to the value obtained from analytical solution under the plane stress condition. The order of stress singularity for dissimilar joint having 0.5 mm bond thickness was also listed. If the value of λ between 0.5 mm and 1.0 mm thick dissimilar joint is compared, one may notice that there is a slight different in the value of λ. This indicates that the dependency of λ upon bond thickness cannot be totally neglected. When the bond thickness is reduced (i.e. t → 0), λ will also decrease to the extent approaches 0.1309 which is the λ value of SUS304/ALU interface corner. Thus, the interface corner of dissimilar joint with thinner bond thickness will experience less stress concentration and consequently high failure stress can be anticipated. As already discussed in the previous section, it is essential to determine the failure stress of adhesive joints. The relation between failure stress and bond thickness for adhesive joints without crack is shown in Fig. 5. The solid triangle plots are referred to the results obtained from 3PB test. The open square and solid square plots are referred to the results of tensile test with failure of type A and type B, respectively. Type A corresponds with interface failure initiated at the ALU/adhesive interface corner while type B corresponds with cohesive failure initiated at the SUS304/adhesive interface corner. As for the specimens failed at ALU/adhesive interface corner, this feature can be explained in terms of surface wettability[17]. Following a simple drop test of adhesive onto the surface of SUS304 stainless steel adherend and YH75 aluminium alloy adherend, we found that the surface of YH75 aluminium alloy adherend has a poor wettability against adhesive employed in this study.

1002 1002

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

Table 2 Interface corner stress singularity parameter of tensile dissimilar joint Method Theoretical FEM Condition Plane Strain Plane Stress Plane Stress Plane Stress Side t = 0.5 mm t = 1.0 mm SUS/ADH 0.3289 0.2613 0.2557 0.2652 ALU/ADH 0.2963 0.2321 0.1851 0.2250 (a)

(b) 3.5

t =1.0mm

3

SES-SUS AEA-ALU SEA-SUS SEA-ALU

t =1.0mm

3 2.5 σy (MPa)

2.5 σy (MPa)

3.5

SES-SUS AEA-ALU SEA-SUS SEA-ALU

2

2

1.5

1.5

1

1 0.5

0.5 0

0.02

0.04

0.06

0.08

0

0.1

0.02

0.04

0.06

0.08

0.1

r (mm)

r (mm)

Fig. 3 Stress-y distribution near the interface corner in adhesive joint. (a) Tensile, (b) 3PB. r is the distance from interface corner edge (a)

(b)

3

10

SEA-3PB SEA-Tensile

t =1.0mm

2.5

t =1.0mm

SEA-3PB SEA-Tensile

1

σy (MPa)

σy (MPa)

2 1.5

0.1

1 0.5 0 0

2

4 r (mm)

6

8

0.01 0.001

0.01

0.1 r (mm)

1

10

Fig. 4 Stress-y distribution near SUS304/adhesive interface corner in dissimilar joint. (a) comparison between tensile and 3PB specimens (b) slope of the curves from log(stress-y)-log(r) of the respective plots. r is distance from the interface corner edge In the 3PB test, only failure of type B was observed. This kind of failure can be explained as the existence of stress singularity at the SUS304/adhesive interface corner. Since the stress singularity at the SUS304/adhesive corner is larger than the ALU/adhesive interface corner, the failure will always propagate from this apex. At a distance ahead of the interface corner line, the stress singularity is vanished, thus the crack deviates into the adhesive layer where the stress concentration is relatively high. Furthermore, it can be noted in all cases, the failure stress is gradually increased with the decreasing bond thickness. This indicates a typical influence of bond thickness upon the strength of adhesive joints and has been well accepted. Moreover, the failure stress from 3PB test is higher than the results of tensile test. This is due to that the adhesive layer inside the tensile test specimens has experienced far greater stress concentration in comparison to adhesive layer inside the 3PB test specimens. In addition, in 3PB test specimens, only half of the cross section is subjected to tensile stress (see Fig. 4 (a)) while the other half is in the compressive stress state.

1003 1003

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

The interface corner toughness has been widely employed to predict the strength of adhesive joint and its relation to the bond thickness[8-12]. According to Akisanya and Meng[10], the interface corner toughness, Hc is defined by

H C = σ C t λ Q(α , β )

(1)

where Q is a non-dimensional constant function of the material elastic parameters (i.e. Dunder’s parameter). Since adherends are much more rigid than epoxy adhesive, α=1 and β=α/4 is considered. For these material combinations, the value of Q is taken from ref. [10] as 0.5. As for the λ, we used the value obtained from analytical solutions under the plane stress assumption as listed in Table 2. From the calculation, the average values of Hc for 3PB test and tensile test with type B failure are 7.51 MPa.mm0.26131 and 4.66 MPa.mm0.26131, respectively. Using the value of Hc in conjunction with Eq. (1), inversely, joint strength of each test can be predicted. Prediction lines based on Hc parameter for 3PB test and tensile test are represented by black solid line and red solid line, respectively, as shown in Fig. 5. Obviously, the prediction is in good agreement with the measured data with the 3PB prediction line is higher than Tensile (B). Besides, the average value of Hc for tensile test with type A failure is 3.03 MPa.mm0.23209 and cannot be directly compared with other results due to the different in stress singularity order.

40

Failure stress (MPa)

SUS304 Adhesive

30

B

A

3PB (B) Tensile (B) Tensile (A)

ALU

20

10

0 0

0.2

0.4 0.6 0.8 1 1.2 Bond thickness, t (mm)

1.4

1.6

Fig. 5 Failure stress of adhesive dissimilar joint against bond thickness The influence of residual stress in the strength of adhesive joint is now discussed. Residual stress arises in an adhesive joint due to significant temperature change during cooling from their processing (i.e. curing) temperature[18]. Residual stress within the adhesive layer bonded between similar adherends may be estimated by[19]

σ res =

Eadh (α adh − α )(Tsft − Tt ) 1 −ν adh

(2)

Here, α denotes the coefficient of thermal expansion (abbreviated as CTE hereafter). When dissimilar adherends are bonded together, α is expressed by

1

α

=

1

α1

+

1

α2

(3)

(i.e. subscripts 1 and 2 refer to SUS304 adherend and aluminium adherend, respectively). Tsft is the stress free temperature which in this study is the curing temperature. Tt is the

1004 1004

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

testing temperature. Since all specimens in this study were cured and tested at room temperature, Tsft ≈ Tt. Therefore, from Eq. (2), it can be considered that residual stress is almost negligible in this experiment. With a slight change in curing and testing temperature will inevitably generate the residual stress. However, the resulting residual stress for specimens cured and tested at room temperature has only little effect on adhesive joint strength so much so it is less important to be considered[20,21]. Moreover, since the bond thickness employed was small, the residual stress resulted from adhesive shrinkage associated with curing can also be neglected. It is noteworthy that the pre-exist residual stress may affect the stability of the crack propagation path[22]. 3.2 Fracture toughness of adhesive dissimilar joint with interfacial crack Next, we will discuss the effect of interfacial crack length on the fracture toughness of tensile dissimilar joints with an interfacial crack. Figure 6 shows the results of fracture toughness, Jc versus percentage of interface fracture (abbreviated as IF hereafter) for the tensile dissimilar joint having interfacial crack length of 5, 10 and 15 mm. Note that the IF is the area of interface fracture over the entire bonding area. Here, SEA and AES represent the dissimilar joint with an interfacial crack at the SUS304/adhesive interface and ALU/adhesive interface, respectively. In the case of SEA, as shown in Fig. 6, Jc values fall

60

Area of IF

Jc (N/m)

50

adhesive

40

crack

5mm 10mm 15mm

30 20 10 0 0

20

40 60 IF (%)

80

100

Fig. 6 Fracture toughness, Jc against fraction of interface fracture (Tensile-SEA)

60

Jc (N/m)

50

Area of IF

40

adhesive

5mm 10mm 15mm

crack

30 20 10 0 0

20

40 60 IF (%)

80

100

Fig. 7 Fracture toughness, Jc against fraction of interface fracture (Tensile-AES)

1005 1005

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

within 10-30 (N/m) when IF is less than 60%. Meanwhile, for AES case, the Jc value is apparently constant for IF ranges between 40-90%, as can be seen in Fig. 7. In contrast to the SEA case, most joints of AES fractured entirely at the joint interface. Furthermore, it was also observed that from all specimens, there was no established dependency of Jc upon interfacial crack length. Thus, the following discussion will be restricted only to the effect of bond thickness on the Jc of dissimilar joints with an interfacial crack. We divided this discussion with regard of the locus of fracture into cohesive fracture and interface fracture. Here, when the specimen fractured within the adhesive, it was distinguished as cohesive fracture (abbreviated as CF hereafter). The relationship between fracture toughness, Jc and bond thickness, t for cohesively fractured specimens is shown in Fig. 8. For Tensile-SEA, despite some variance in data, the value of Jc is more or less constant, which is about 10~30 (N/m). Meanwhile, for Tensile-AES, the Jc values are constant even though the number of data is small. In addition, the number of 3PB data is also remarkably very small. For the interfacially fractured specimens, the relationship

60

3PB-SEA 3PB-AES Tensile-SEA Tensile-AES

50

Jc (N/m)

40

adherend adhesive adherend

30 20 10 0 0

0.2

0.4 0.6 0.8 Bond thickness, t (mm)

1

1.2

Fig. 8 Fracture toughness, Jc against bond thickness (Cohesive fracture)

60 adherend

50

adhesive adherend

Jc (N/m)

40

3PB-SEA 3PB-AES Tensile-SEA Tensile-AES

30 20 10 0 0

0.2

0.4 0.6 0.8 Bond thickness, t (mm)

1

1.2

Fig. 9 Fracture toughness, Jc against bond thickness (Interface fracture)

1006 1006

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

between fracture toughness, Jc and bond thickness, t is shown in Fig. 9. For the tensile data, only the joints of AES were interfacially fractured. Also, we can observe that there is a trend where Jc increases with the decreasing bond thickness. In contrast, for 3PB data, there is no significant relationship between Jc and t, and the data are quite scattered in comparison to the tensile data. Overall, the Jc values for cohesively fractured joint are higher than the Jc values of interfacially fractured joint. To elucidate the mechanisms of fracture for dissimilar joints with an interfacial crack as discussed above, we have analyzed the stress-y distribution at the interfacial crack tip and subsequently determine their stress singularity λ by FEM analysis. Figure 10 (a) shows the stress-y distribution in front of an interfacial crack originated at the SUS/adhesive interface corner in the dissimilar joint of 3PB and tensile. The model with smallest mesh size of 0.01 x 0.01 mm was employed. In Fig. 10 (b), within 0.001 < r/W < 0.01, we can obtain two slopes which equal to the λ value for tensile and 3PB. Since the mesh size constructed in FEM analysis is crucial and may affect the stress distribution in front of the interfacial crack tip, we re-analyzed the same model but now with the finest mesh size of 0.1 x 0.1 mm to study its effect and the results obtained are plotted in Fig. 10 (c). From Fig.10 (c), we noticed a moderate change in stress-y value and found that it was difficult to evaluate λ at this time. Therefore, the dissimilar joint model with the highly refined mesh size of 0.01 x 0.01 mm will be used throughout this study. Results of λ at the interfacial crack tip of 1.0 mm thick dissimilar joint are summarized in Table 3 below. (a) 5 SEA-3PB SEA-Tensile

σy (MPa)

4 3

0.01 x 0.01mm t = 1.0 mm

2 1 0 0

0.1

0.2

0.3

0.4

0.5

r /W

(b)

(c)

10

10 SEA-3PB SEA-Tensile

σy (MPa)

σy (MPa)

SEA-3PB SEA-Tensile

1

1

0.1 x 0.1mm t = 1.0 mm

0.01 x 0.01mm t = 1.0 mm 0.1 0.0001

0.001

0.01 r /W

0.1

1

0.1 0.0001

0.001

0.01 r /W

0.1

1

Fig. 10 Stress-y distribution in front of the interfacial crack originated at SUS/adhesive interface corner in dissimilar joint. (a) comparison between tensile and 3PB specimens, (b) log(stress-y)-log(r/W) of the respective plots, (c) plots for model with the coarse mesh of 0.1 x 0.1 mm. Abscissa is the normalized distance from an interface corner

1007 1007

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

Table 3 Stress singularity parameter at interfacial crack tip of 1.0 mm thick dissimilar joint Method Theoretical FEM (3PB) FEM (Tensile) Condition Plane Plane P. Stress P. Stress P. Stress P. Stress Strain Stress 0.1x0.1mm 0.01x0.01 0.1x0.1mm 0.01x0.01 Side mm mm SUS/ADH 0.3534 0.2876 0.3161 0.3196 0.2448 0.2542 ALU/ADH 0.3623 0.3033 0.2797 0.3187 0.2472 0.2703 It is well known that, the fracture will initiate at a point where the stress concentration is critical. Since an interfacial crack introduced in a dissimilar joint using Teflon tape in our study is not a true natural crack, there are actually two potential locations where the fracture is most likely initiated, i.e. either at the upper crack tip (T1) or at the lower crack tip (T2). Figure 11 (a) shows the contour plots of stress-y in 3PB-SEA. Obviously, the high stress concentration can be seen at the region of T1 and T2. In the case of T1, the fracture initiates and propagates along the interface which in the final form reveals as an interface fracture. On the other hand, if fracture initiates at T2, the fracture grows within the adhesive layer and from the appearance of fracture surface, this can be characterized as a cohesive failure. Table 4 Stress singularity parameter at interfacial crack tip of 1.0 mm thick dissimilar joint Model FEM (3PB) FEM (Tensile) Condition Side SEA AES

crack

T1 0.3196 0.3187

crack

T2 0.2875 0.2995

crack

T1 0.2542 0.2703

crack

T2 0.2546 0.2552

To illustrate further this point, we evaluate the λ at T1 and T2 for both 3PB and tensile model, and the results are given in Table 4. Accordingly, the value of λ at T1 is larger than T2 in all configurations under our consideration except in the case of Tensile-SEA. As for the case where T1 > T2, interface fracture is dominant and this is in agreement with the fracture surface observation of Tensile-AES, 3PB-SEA and 3PB-AES specimens where these joints failed interfacially (see Fig. 8). For the Tensile-SEA, even though T2 > T1, but the different is not very distinct, so we still need another supporting evidence to interpret this fracture behavior. With this regard, we investigated the contour plots of stress-y in dissimilar joint and the results are shown in Fig. 11 (b). In Fig. 11 (b), it is interesting to note that, clearly the stress-y is locally concentrates at the T2, which means the cohesive fracture will take place. This feature is very consistent with the fractography observation of Tensile-SEA specimen surface (see Fig. 8) where most of the joints failed cohesively. Finally, it should be emphasized that the quality of the bonding surface affects strongly the strength and toughness of adhesive joint. In other words, any undetectable defects on the bonding surface or insufficient bonding achieved may also contribute to not only a change in the locus of fracture but also reduce the Jc values dramatically. Further study is greatly needed to fully understand the strength and fracture characteristics of adhesive joint.

4. Conclusions The effect of bond thickness upon the strength and fracture toughness of epoxy adhesive dissimilar joint is investigated with tensile and 3PB specimens. The stress distribution at the interface corner of dissimilar joint was analyzed by FEM analysis. From FEM analysis results, it is found that the stress singularity in the dissimilar joint exists pronouncedly at the SUS/adhesive interface corner and the order of stress singularity in the tensile model is higher than that in the 3PB model. Moreover, the order of stress singularity

1008 1008

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

(a) T1

T2

Teflon tape

(b) T1

T2

Teflon tape

Figure 11 Contour plots of stress-y of (a) 3PB-SEA and (b) Tensile-SEA. t is 1.0 mm and the finest mesh size is 0.01 x 0.01 mm. Applied stress is 1 MPa in the dissimilar joint having bond thickness of 1.0 mm is quite close to the value obtained from analytical solution under the plane stress condition. From 3PB test and tensile test, it is confirmed that the failure stress increases slightly with the decreasing bond thickness and can well be predicted by using Hc parameter. The failure stress for dissimilar joint of 3PB test is higher than that of tensile test. Fracture toughness, Jc for cohesively fractured specimens is more or less constant but shows some dependency on bond thickness for interfacially fractured specimens. Locus of fracture can be best interpreted by the assessment of the order of stress singularity at the interfacial crack tip.

1009 1009

Journal of Solid Mechanics and Materials Engineering

Vol. 4, No. 7, 2010

References [1] Shields, J., Adhesives handbook, Butterworths, 1984. [2] Ikeda, T., Lee, D., Miyazaki, N., J. Adhes. Soc. Jpn., vol. 3, 97-105, 2006. [3] Lee, D. B., Ikeda, T., Miyazaki, N., and Choi, N. S., ASME J. Eng. Mater. Technol., vol. 126, 14–18. 2002. [4] Duer, R., Katevatis, D., Kinloch, A.J., Int. J. Frac., vol. 75, 157-162, 1996. [5] Daghyani, H.R., Ye, L., Mai, Y.-W., J. Mat. Sci., vol. 31, 2523-2529, 1996. [6] Fleck, N.A., Hutchinson, J.W., Suo, Z., Int. J. Solids Struct., vol. 27, 1683-1703, 1991. [7] Akisanya, A.R., Fleck, N.A., Int. J. Frac., vol. 58, 93-114, 1992. [8] Hattori, T., Sakata, S., Hatsuda, T., Murakami, G., JSME Int. J., vol. 31, no. 4, 718-723, 1988. [9] Qian, Z. and Akisanya, A.R., Acta mater., vol. 14, 4895-4904, 1998. [10] Akisanya, A.R., Meng, C.S., J. Mech. Phys. Solids, vol. 51, 27-46, 2003. [11] Reedy, E.D. and Guess, T.R., Int. J. Solids Struct., vol. 30, 2929-2936, 1993. [12] Reedy, E.D. and Guess, T.R., Int. J. Frac., vol. 88, 305-314, 1997. [13] Afendi, M., Teramoto, T., in: Proc. of Yamanashi District Conf. (JSME), 243-244, 2008. [14] Information on http://www.jisc.go.jp/ [15] Kyogoku, H., Sugibayashi, T., Ikegami, K., Trans. Jpn. Soc. Mech. Eng., vol. 53, 499-505, 1986. [16] Bogy, D.B., Trans. ASME, J. Appl. Mech., vol. 38, 377-386, 1971. [17] Afendi, M. and Teramoto, T., J. Adhes. Soc. Jpn., vol. 45 (12), 471-477, 2009. [18] Dillard, D.A., in: Adhesion Science and Engineering: Vol. 1- The Mechanics of Adhesion, edited by Pocius, A.V. and Dillard, D.A., Chapter 1, Elsevier Science, 2002. [19] Chen, B., Dillard, D.A., Int. J. Adhes. Adhes., vol. 21, 357-368, 2001. [20] Reedy, E.D., in: Adhesion Science and Engineering: Vol. 1- The Mechanics of Adhesion, edited by Pocius, A.V. and Dillard, D.A., Chapter 5, Elsevier Science, 2002. [21] Kinloch, A.J., Thrusabanjong, E, Williams, J.G., J. Mat. Sci., vol. 26, 6260-6270, 1991. [22] Hutchinson, J.W., Suo, Z., Mixed Mode Cracking in Layered Materials, Academic Press, Inc., 1992.

1010 1010