ARTICLE IN PRESS

International Journal of Mechanical Sciences 50 (2008) 641–648

www.elsevier.com/locate/ijmecsci

Stress concentration analyses of bi-material bonded joints

without in-plane stress singularities
Zhixue Wu
Mechanical Engineering College, Yangzhou University, Yangzhou 225009, PR China
Received 4 February 2007; received in revised form 20 December 2007; accepted 11 January 2008
Available online 19 January 2008

Abstract

The problem of stress concentration in bi-material bonded joint is investigated under the condition of without stress singularities.
Disappearance conditions of stress singularity near interface corners and edges are determined based on analyses of eigenvalue equations.
Straight-side and curved interface of materials are designed for the bi-material models to avoid singular stress fields around the interface
corner and edge. Assuming that one stress component or combined stresses are responsible for failure at or near the interface, the stress
concentration becomes critical for the design of bi-material joints with higher interfacial strength. Numerical results show that the stress
state near the interface depends strongly on both the interface geometry and the combination of materials, and stress concentration may
always occurs at or near the interface. Emphasis is placed on the necessity for geometric optimization of an interface in order to design
singularity-free junction with higher interfacial strength.
r 2008 Elsevier Ltd. All rights reserved.

Keywords: Stress singularity; Interfacial strength; Bonded joint; Stress concentration; Finite element analysis

1. Introduction elimination of stress singularities may be an effective

approach. Based on this idea, some novel specimen
Bonded components are widely used in the automotive, designs have been proposed to evaluate the interfacial
aerospace, and electronic industries. Compared to mechan- strength. A cruciform specimen was suggested by Tandon
ical fastening, e.g. riveting or bolting, adhesive bonding et al. [2,3] to measure the normal strength of fiber/matrix
may provide more uniform and efficient load transfer into interface in unidirectional composites. A convex interface
the patch and can reduce the risk of high stress concentra- joint was designed to measure interfacial tensile strength
tions. This leads to wide use of bonded repairs instead of [4–6] and interfacial shear strength [7]. Steven and
using riveted repairs in aircraft structures [1]. The strength Stefan [8] investigated the interfacial strength of butt
and reliability of such components are of particular joints with different interface bias. Schüller and Lauke
concern since failure is often associated with material and presented a novel method to determine the adhesion
geometric discontinuities. It is found that the failure of strength of a polymer joint [9], where the stress state at
these material systems often initiates near the interface the interface is purely tensile and does not contain shear
corners or free edges due to stress singularities (the very components. The above results showed that the static load
high stresses) developed near these regions under external transfer capacity of dissimilar material joints could be
loading. Hence, reducing local stress levels may result in improved after removing or reducing the influence of the
higher joint strengths of dissimilar materials. stress singularity.
In order to evaluate intrinsic interfacial strengths or to In fact, the stress field near a general 3-D interface
improve load capacity of dissimilar materials, reduction or corner is considerable complex. Hideo [17] has reported
that the order of stress singularity at a corner of 3-D
Tel.: +86 514 7851796; fax: +86 514 7887937. joints is stronger than that at a corner of 2-D ones.
E-mail address: [email protected] Recently, Kotousov [18] reported that there may be a new

0020-7403/$ - see front matter r 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijmecsci.2008.01.004
 ARTICLE IN PRESS
642 Z. Wu / International Journal of Mechanical Sciences 50 (2008) 641–648

mode of singularities, the out-of-plane mode, which constants,

depends on Poisson’s ratio for plates of finite thickness. 8
These factors may have important effect on the design of >
> E1  E2
>
<a ¼ E þ E ;
dissimilar material joints without stress singularities. 1 2
(1)
Apparently, the first important step for measuring or >
> m1 ðk2  1Þ  m2 ðk1  1Þ
>
: b ¼ m ðk þ 1Þ þ m ðk þ 1Þ:
improving interfacial strength is the elimination or reduc- 1 2 2 1
tion of stress singularities. The strength concept also
requires the stress distribution as uniformly as possible Here E j ¼ E j for plane stress and E j ¼ E j =ð1  v2j Þ for
along the interface of bi-material bonded joint. Assuming plane strain; Ej and vj (j ¼ 1, 2) are the elastic moduli and
that one stress component or combined stresses are Poisson’s ratios of materials 1 and 2, mj ¼ Ej/(2(1+vj)) are
responsible for failure at or near the interface after the material shear moduli; kj ¼ (3vj)/(1+vj) for plane
removing the stress singularity, the stress concentra- stress and kj ¼ 34vj for plane strain. As reported by Bogy
tion becomes critical for the design of bi-material [11], the stress singularity around a wedge (the intersection
joints with higher interfacial strength. In this work, the of interface with free side, O or O0 ) is determined by the
problems of stress concentration in bi-material bonded material mismatch and two joint angles of the bi-material
joint after eliminating the stress singularity are investi- corner. For the bi-material geometry with straight-side, the
gated using axisymmetric finite element models with order of the stress singularity can be determined from the
straight-side and curved interface of materials. The analysis characteristic equation [11]:
is based on linear elastic theory and the interface is Dðg; a; b; lÞ ¼ f4l2 ðl2  1Þ sin4 g þ sin2 lðp  2gÞ
considered as perfectly bonded. The aim is to investi-
gate the influences of the interface geometry and the  l2 sin2 2gga2 þ 4l2 sin2 gfsin2 lg
combination of materials on the stress concentration, þ sin2 lðp  gÞ  2l2 sin2 gÞgab
with the goals of providing useful information for
þ 4fsin2 lðp  gÞ  l2 sin2 gg
engineers to design bi-material joints with higher interfacial
strength.  ðsin2 lg  l2 sin2 gÞb2
þ 2fsin2 lg  sin2 lðp  gÞg
2. Theoretical background and analytical models  ð1  2l2 sin2 gÞa  4l2 sin2 g
 fsin2 lðp  gÞ  sin2 lggb þ sin2 lp ¼ 0,
Consider the bi-material geometry with straight-side and
interface shown in Fig. 1(a), which consists of two elastic, (2)
isotropic and homogeneous materials. As first noted by where g and pg are the two wedge angles, l is a complex
Dundurs [10], the material properties can be completely parameter representing the stress singularity intensity.
described in terms of two dimensionless material mismatch In general, l may be real or complex. The contribution
parameters which are combinations of the four elastic to the singular stress field from the smallest Re (l) is
significantly greater than from all the other values of l and
the minimum |Re(l1)| is generally used to represent the
order of the stress singularity. According to Eq. (2), the
stress singularity for a given combination of materials will
disappear if Re(l)X1. In fact, a preliminary analysis
showed that a range of the two wedge angles, with which
Material 1 Material 1 the stress singularities will disappear, might always exist
even though under the constraint of straight-side geometry.
π- For a given combination of material properties, the critical
1

O′   wedge angle g can be obtained using the searching program
developed by the author [12].
Interface O″ It is found that g seems always less than 901 if material 1
O′
O  O 2 is a typical soft material (with smaller elastic modulus) and
r material 2 is a hard material (with larger elastic modulus),
Interface
though there is no mathematical or physical theory behind
this statement. This means that it is not practical to design
Material 2 Material 2
a bi-material joint with straight-side and interface, as
shown in Fig. 1(a), in which stress singularities around
both O and O0 are removed. That is to say, the interface has
to be profiled, as shown in Fig. 1(b), if we want to design a
Fig. 1. General model of bi-material joint: (a) the definition of interface bonded joint with constant section (instead of convex
wedge and (b) the definition of interface corner. interface joint) and without stress singularities. In this case,
 ARTICLE IN PRESS
Z. Wu / International Journal of Mechanical Sciences 50 (2008) 641–648 643

there is an interface corner (point O00 in Fig. 1(b)) and stress y1 ¼ y2 ¼ p or a ¼ b (a logarithmic singularity may be
singularities may arise around this point. The order of the present in this case [13,14]). One example is shown in Fig. 2
stress singularity around an interface corner can be to demonstrate the variation of l with y1. The first
obtained from finding the Eigenvalue l from the char- combination of materials is ceramic (material 1, E1 ¼
acteristic equation provided by Bogy and Wang [13]: 359 GPa, v1 ¼ 0.20) and quartz (material 2, E2 ¼ 73.1 GPa,
v2 ¼ 0.17) and the second one is glass (material 1,
Dðy1 ; y2 ; a; b; lÞ ¼ ½l2 ða  bÞ2 sin2 y1  ð1 þ bÞ2 sin2 ly1  E1 ¼ 64 GPa, v1 ¼ 0.20) and polycarbonate (material 2,
 ½l2 ða  bÞ2 sin2 y2  ð1  bÞ2 sin2 ly2  E2 ¼ 2.4 GPa, v2 ¼ 0.34). This figure shows that l ¼ 1
(no stress singularity) if y1 ¼ p, though the variation of l
þ ð1  a2 Þ sin2 lðp  y2 Þ½2l2 ða  bÞ2 sin2 y2
with y1 is complicated. This suggests that the stress
þ 2ð1  b2 Þ sin ly1 sin ly2 þ ð1  a2 Þ singularity can be eliminated by means of smoothing the
 sin2 lðp  y2 Þ ¼ 0, (3) sharp interface corner, like the method for eliminating the
effect of 3-D bi-material interface corner geometries on
where the plane strain condition is assumed. Liu et al. [14] stress singularity [15].
have also obtained the identical characteristic equation for The author has performed stress analyses of singular-free
an axisymmetric interface corner. bi-material joints with flat interface [12]. It was shown that
Examining Eq. (3), it is found that the stress singularity the effect of the mismatch between material properties on
seems always exist around an interface corner except the magnitude of stress components can be much stronger
associated with the influence of geometric factors. Con-
1 sidering that the stress distribution in a convex joint cannot
be uniform even for mono-material specimen and that this
0.9 shape may not be widely adopted in engineering applica-
tion, models with straight-side and curved interface, as
shown in Fig. 3, are used to analyze the stress field near the
0.8
interface. Comparing with convex joint, straight-side joint
may be more useful. It should be noted that Bogy equation


0.7 is also applicable to curved interface, though this formula

is derived for straight-side geometries. Three geometrical
0.6 Ceramic/Quartz configurations of Fig. 3 are used to study the possible
Glass/Polycarbonate influence of interface geometry on stress concentration.
The determination of the best interface geometry is a very
0.5
0 60 120 180 240 300 360 complex problem. In this work, we focus only on the
1 (degree) curved interface that consists of a circular arc and a
straight line. The variation of shape then decreases to a
Fig. 2. The variation of l with interface corner y1. variation of the arc radius r (0orpR) and each shape is

0 = 1MPa 0 = 1MPa

Material 1 Material 1 Material 1

H1

y y




Interface
r
r O O R Interface
x x
Interface
H2

Material 2 Material 2 Material 2

Fig. 3. Schematic of analytical geometries: (a) the curved section of the interface near the surface, (b) the curved section of the interface at the center, and
(c) the case of r/R ¼ 1.
 ARTICLE IN PRESS
644 Z. Wu / International Journal of Mechanical Sciences 50 (2008) 641–648

employed to predict initiation of failure as a first

approximation [2,3,8,9]. It predicts failure when one stress
component or combined stresses reach its corresponding
intrinsic strength property. The stress analysis provides the
connection between the applied load and the created
interfacial stress. Several failure mechanisms may occur
in the joint depending on the adhesion quality at the
Interface
interface and the mechanical properties of the constituents:
debonding at the interface or the local stress within the
constituents exceeding the allowable value. Which kind of
failure really occurs depends on the local strengths of the
constituents respectively of the interface.
Here, stress concentrations are defined by
sm sm
eqv sNor
Km
1 ¼
1
; Km
eqv ¼ ; K Nor ¼ , (4)
s0 s0 s0
where m=1, 2 is index for the constituents; sm m
1 , seqv and
sNor are the first principal stress, the von Mises stress and
the interfacial normal stress, respectively. All the stresses
Fig. 4. A typical finite element mesh. are obtained using the solution for the stress field within
the corresponding constituent. For the interfacial normal
stress, the larger value is adopted.
uniquely characterized by the shape parameter r/R, where
R ¼ W/(2 cos g) (go901) is the maximum arc radius and g 3.1. The influence of material mismatch
(go901) is the maximum wedge angle of the hard material.
Fig. 3(a) and (b) are the designs that the curved sections are The influence of the material mismatch on stress
near the surface and the center, respectively. Fig. 3(c) is the concentration is studied using the model of Fig. 3(c) by
case of r/R ¼ 1. means of analyzing of the stress field near the interface. The
Many interface corner problems are of interest in applied Poisson’s ratios of the two materials are fixed with
mechanics and engineering application. The complexity of v1 ¼ v2 ¼ 0.3, and E1 and E2 are varied in order to study
stress singularity at a vertex of general 3-D joint can be their influence. The critical wedge angle g and the
avoided for an axisymmetric joint [15]. Finite element corresponding Dundurs’ parameters for various material
analyses of the models in Fig. 3 have been performed combinations are shown in Table 1. It is interesting to note
employing the software package ANSYS 8.0 using that the critical wedge angles are almost the same for the
axisymmetric models. For the axisymmetric interface selected material combinations. For simplicity, critical
corner of perfectly bonded dissimilar materials, it was wedge angles for various material combinations are just
shown that the order of the stress singularity at interface set g ¼ 56.61 in all the models to avoid the possible effect of
corner coincides with that of the corresponding plane different interface geometry.
strain problem [14]. Thus, the critical wedge angle g can be As an example, Fig. 5(a) and (b) shows, respectively, the
obtained with Eq. (2) by setting l1 ¼ 104 [12]. Only a contours of the first principal stress and the von Mises
half of the joint in Fig. 3 was analyzed because of the stress in the joint of E1/E2 ¼ 0.1. It is found that the
symmetry. Six-node triangular elements are used in all maximum first principal stress and the maximum von
models, as shown in Fig. 4. Preliminary analysis showed Mises stress appear away from the interface in the soft
that singular element or extremely fine mesh division is not material but they occur at the interface in the hard material
necessary due to the disappearance of the stress singularity. under axisymmetric condition. Fig. 6 gives the variations of
However, suitable refinement is needed along the interface the maximum stress concentration with elastic modulus
because strong stress concentration may occur at or near ratio of the bounded materials E1/E2. It can be observed
the interface. Values of (H1, H2)X1.5W are used so that that all the maximum stress concentrations increase
the influence of the top and bottom edge boundary approximate linearly as E1/E2 decreases. This means that
conditions on the stress near the interface is at a negligible material mismatch has effect on the stress concentration,
level. The load s0 is assumed to be uniform tension of even thought specimen geometries are the same and stress
1 MPa. singularity is eliminated.
Strictly speaking, the measurement of a material strength
3. Results and discussion requires a uniform and uniaxial stress state. However, it
seems to be impossible to achieve this kind of stress state
There are several criterions for the measurement of along an interface due to the material mismatch. Fig. 7
interfacial strength. The stress-based criterion is frequently shows the distribution of the first principal stress, the von
 ARTICLE IN PRESS
Z. Wu / International Journal of Mechanical Sciences 50 (2008) 641–648 645

Table 1
The critical wedge angle g for various material combinations

E1/E2 0.001 0.01 0.02 0.05 0.10 0.20 0.50 0.75 0.90
g(1) 56.7823 56.7822 56.7820 56.7816 56.7807 56.7787 56.7676 56.7378 56.6483
a 0.9980 0.9802 0.9608 0.9048 0.8182 0.6667 0.3333 0.1429 0.0526
b 0.2851 0.2801 0.2745 0.2585 0.2338 0.1905 0.0952 0.0408 0.0150

1.3
2
1 eqv
1.2 1

1.1

Stress concentration
1.0
Nor 12
0.9

0.8

1
0.7 eqv

0.6

0.5
0.0 0.2 0.4 0.6 0.8 1.0
2x/W

Fig. 7. Stress concentrations along the interface for the model of Fig. 3(c),
E1/E2 ¼ 0.1.

the design of the convex joints with flat interface cannot

generate a uniform stress distribution at a bi-material
Fig. 5. Contours of the stress components in the joint of Fig. 3(c) with E1/
E2 ¼ 0.1: (a) the first principal stress and (b) the von Mises stress.
interface either [12].

1.4 3.2. The influence of interface geometry

1
1(Max)
1.35
2
1 (Max)
Various possible shapes of interfacial geometries, from a
The maximum stress concentration

1.3 1
technological point of view, may be adopted to design a bi-
eqv (Max) material bonded joint without stress singularity. The
2
1.25 eqv (Max) influence of interface geometry on stress concentration is
Nor (Max) studied using the models of Fig. 3(a) and (b) with various
1.2 shape parameters r/R. Considering that the critical wedge
angle varies little for many material combinations,
1.15
g ¼ 56.61 is used in the models if not mentioned. It is
1.1
noted that the position of the maximum first principal
stress may be changed due to the influence of interface
1.05 geometry. For example, the maximum first principal stress
in the hard material appears near the interface if r/Ro0.7
1 (see Fig. 8) in the case of E1/E2 ¼ 0.1 for the model of
0 0.2 0.4 0.6 0.8 1
Fig. 3(a). Fig. 9 shows the variations of the maximum
E1/E2
stress concentrations with r/R in the case of E1/E2 ¼ 0.1. It
Fig. 6. Variations of the maximum stress concentration with elastic can be seen that the effect of the relative curvature radius
modulus ratio of the bounded materials. r/R on the maximum stress concentration is different for
different stress component. It is generally supposed that the
maximum normal stress at the interface is responsible for
Mises stress and the interfacial normal stress along the debonding [2,3,9]. The results in Fig. 9 reveal that the
interface in the joint of E1/E2 ¼ 0.1. It can be seen that debonding strength can be increased via the optimization
none of the stresses is uniform along the interface. In fact, of the interface geometry. This also shows that the problem
 ARTICLE IN PRESS
646 Z. Wu / International Journal of Mechanical Sciences 50 (2008) 641–648

Fig. 8. Contours of the first principal stress in the half joint of Fig. 3(a) with E1/E2 ¼ 0.1: (a) r/R ¼ 0.1 and (b) r/R ¼ 0.6.

2 that in model of Fig. 3(a) if r/R is the same. Model of

1
1 (Max) Fig. 3(a) seems to have some advantages over model of
1.9
2
1 (Max) Fig. 3(b), if the maximum normal stress is critical.
The maximum stress concentration

1.8 In fact, there is a wide choice of the wedge angle free of

1eqv (Max)
1.7 singularity since the stress singularity will vanish if the
2
eqv (Max) wedge angle is less than the critical value. Fig. 12 illustrates
1.6
Nor (Max) the effect of different wedge angle (and different R, since
1.5 R ¼ W/(2 cos g)) on the maximum stress concentrations
1.4
using the model of Fig. 3(c) with E1/E2 ¼ 0.1 and
v1 ¼ v2 ¼ 0.3. As shown in Fig. 12, it seems to be the best
1.3 choice to design a bi-material joint free of singularity at the
1.2 critical wedge angle.
The above results demonstrate that strong stress
1.1
concentration will appear at or near the interface for the
1 studied bi-material models if r/R or E1/E2 is too small. This
0 0.2 0.4 0.6 0.8 1 allows the optimization of the interfacial stress state by
r/R changing the interface geometry of a bi-material joint when
Fig. 9. Variations of the maximum stress concentration with r/R in the the combination of materials is chosen. Optimization
model of Fig. 3(a) for E1/E2 ¼ 0.1. means, here, generation of the minimum stress concentra-
tion at or near the interface according to the employed
failure criterion. Also, the determination of a suitable
of stress concentration around an interface is more failure criterion for a bi-material joint is very important in
complex than that at a notch in mono-material structure. the situation without stress singularity since the influence
For the model of Fig. 3(b), the maximum first principal of material mismatch and interface geometry on different
stress and the maximum von Mises stress appear away stress component is not the same. Thus, this work will
from the interface in the soft material but they occur at the provide useful information for engineers to design bi-
interface in the hard material, as shown in Fig. 10. The material joints with higher interfacial strength.
variations of the maximum stress concentrations with r/R However, the practical applications of the above results
in the case of E1/E2 ¼ 0.1 are shown in Fig. 11. It can be may be rather limited as in reality bonded joints do not
seen that all the maximum stress concentrations increase as contain this type of geometry for several reasons. The first
r/R reduces. Comparing with Fig. 9, the maximum normal is that at the micron scale substrate edges are irregular and
stress at the interface in model of Fig. 3(b) is larger than it would be impossible to control this angle of the substrate
 ARTICLE IN PRESS
Z. Wu / International Journal of Mechanical Sciences 50 (2008) 641–648 647

Fig. 10. Contours of the stress components in the half joint of Fig. 3(b) with E1/E2 ¼ 0.1 and r/R ¼ 0.4: (a) the first principal stress and (b) the von Mises
stress.

2.5 1.4
1
1 (Max)
2
1 (Max)
The maximum stress concentration
The maximum stress concentration

1
eqv (Max) 1.3
2 2
eqv (Max)

Nor (Max)
1.2 1
1 (Max)
2
1 (Max)
1.5
1
1.1 eqv (Max)
2
eqv (Max)

Nor (Max)

1 1
0 0.2 0.4 0.6 0.8 1 10 20 30 40 50 60
r/R  (°)

Fig. 11. Variations of the maximum stress concentration with r/R in the Fig. 12. The effect of wedge angle on the maximum stress concentration in
model of Fig. 3(b) for E1/E2 ¼ 0.1. the model of Fig. 3(c) with E1/E2 ¼ 0.1.

at this level. The second and more important reason is that become quite different from the present result if taking into
bonded joint ends tend to contain embedded substrate account the effect of a thin plastic adhesive layer on the
corners rather than the idealized planar configuration stress singularities [16].
shown in Fig. 3. The third is that a linear elastic behavior
of the perfectly bonded bi-material joint is supposed during 4. Conclusions
all considerations. For interface adhensions with adhensive
layer, nonlinear deformations may develop in the consti- The problem of stress concentration in perfectly bonded
tuents, which will cause a change of the edge angles during bi-material joint is investigated using finite element method
loading. For example, the critical wedge angle g will under axisymmetric state with straight-side and curved
 ARTICLE IN PRESS
648 Z. Wu / International Journal of Mechanical Sciences 50 (2008) 641–648

interface under the conditions of linear elastic and no stress [4] Xu LR, Kuai H, Sengupta S. Dissimilar material joints with and
singularity. The following conclusions can be made based without free-edge stress singularities: Part I. A biologically inspired
design. Experimental Mechanics 2004;44(6):608–15.
on the obtained results.
[5] Xu LR, Sengupta S. Dissimilar material joints with and without free-
edge stress singularities: Part II. An integrated numerical analysis.
 Stress concentration may always exist because of the Experimental Mechanics 2004;44(6):616–21.
dependence of the interfacial stress state on the [6] Wang P, Xu LR. Convex interfacial joints with least stress
combination of materials, even thought the stress singularities in dissimilar materials. Mechanics of Materials 2006;38:
singularity is eliminated. 1001–11.
[7] Xu LR, Kuai H, Sengupta S. Free-edge stress singularities and edge
 The influence of interface geometry in conjunction with modifications for fiber pushout experiments. Journal of Composite
material mismatch on the interfacial stress state is very Materials 2005;39(12):1103–25.
strong. Thus, the interface geometry of a bi-material [8] Steven R-A, Stefan H. Strength prediction of beams with bi-material
joint allows further optimization when the combination butt-joints. Engineering Fracture Mechanics 2003;70:1491–507.
of materials is fixed. [9] Schüller T, Lauke B. Measuring the interfacial transverse strength: a
novel test for evaluating polymer joints. International Journal of
 The determination of a suitable failure criterion for a bi- Adhesion & Adhesives 2002;22:169–74.
material joint is important in the situation without stress [10] Dundurs J. Discussion of edge-bonded dissimilar orthogonal elastic
singularity since the influence of material mismatch and wedges under normal and shear loading. Journal of applied
interface geometry on different stress component is not Mechanics 1969;36:650–2.
[11] Bogy DB. Two edge-bonded elastic wedges of different materials and
the same.
angles under surface tractions. Journal of applied Mechanics 1971;
38:377–86.
[12] Zhixue W. Design free of stress singularities for bi-material
Acknowledgments components. Composite Structures 2004;65:339–45.
[13] Bogy DB, Wang KC. Stress singularities at interface corners in
This research was jointly sponsored by State Leading bonded dissimilar isotropic elastic materials. International Journal of
Solids and Structures 1971;7:993–1005.
Academic Discipline Fund and Shanghai Leading Aca- [14] Liu Y-H, Xu J-Q, Ding H-J. Order of singularity and singular stress
demic Discipline (Project nos. Y0102 and BB67), which are field about an axisymmetric interface corner in three-dimensional
greatly appreciated by the author. isotropic elasticity. International Journal of Solids Structures 1999;
36:4425–45.
[15] Zhixue W. A method for eliminating the effect of 3-D bimaterial
References
interface corner geometries on stress singularity. Engineering
Fracture Mechanics 2006;73:953–62.
[1] Vlot A, Verhoeven S, Ipenburg G, et al. Stress concentrations around [16] Kotousov A. Effect of a thin plastic adhesive layer on the stress
bonded repairs. Fatigue & Fracture of Engineering Materials & singularities in a bi-material wedge. International Journal of
Structures 2000;23:263–76. Adhesion & Adhesives 2007;27:647–52.
[2] Tandon GP, Kim RY, Bechel VT. Evaluation of interfacial normal
[17] Hideo K. Stress singularity analysis in three-dimensional bonded
strength in a SCS-0/epoxy composite with cruciform specimens. structure. International Journal of Solids and Structures 1997;34(4):
Composites Science and Technology 2000;60:2281–95. 461–80.
[3] Tandon GP, Kim RY, Warrier SG, Majumdar BS. Influence of free
[18] Kotousov A. Fracture in plates of finite thickness. International
edge and corner singularities on interfacial normal strength:
Journal of Solids and Structures 2007;44(25–26):8259–73.
application in model unidirectional composites. Composites, Part
B: Engineering 1999;30:115–34.