This article was downloaded by: [Mohammad Shishesaz]

On: 15 September 2013, At: 01:40

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Adhesion Science and

Technology
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/tast20

The effect of viscoelasticity of

adhesives on shear stress distribution
in a double-lap joint using analytical
method
a a
Mohammad Shishesaz & Arash Reza
a
Department of Mechanical Engineering , Shahid Chamran
University , Ahvaz , Iran
Published online: 25 Feb 2013.

To cite this article: Mohammad Shishesaz & Arash Reza (2013) The effect of viscoelasticity of
adhesives on shear stress distribution in a double-lap joint using analytical method, Journal of
Adhesion Science and Technology, 27:20, 2233-2250, DOI: 10.1080/01694243.2013.769085

To link to this article: http://dx.doi.org/10.1080/01694243.2013.769085

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the
Content) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
 Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013
 Journal of Adhesion Science and Technology, 2013
Vol. 27, No. 20, 22332250, http://dx.doi.org/10.1080/01694243.2013.769085

The effect of viscoelasticity of adhesives on shear stress distribution in a

double-lap joint using analytical method
Mohammad Shishesaz* and Arash Reza

Department of Mechanical Engineering, Shahid Chamran University, Ahvaz, Iran

(Received 5 November 2012; nal version received 18 January 2013; accepted 22 January 2013)
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

In this paper, the effect of viscoelasticity of polymeric adhesives on shear stress distribution
in the adhesive layer of a double lap joint under shear is studied. The postulated model is
composed of three single isotropic adherend layers joined by polymeric adhesives. Each
adherend layer is assumed to behave as linear elastic. The results are deduced for epoxy
adhesive, and hence, 3-parameters solid model is used to model the viscoelastic behavior
of the adhesive layers. The derived equilibrium equations are transformed into Laplace
domain and are solved using residue theorem. Results show that for a step load of
P0 = 100 N, the peak shear stress in the epoxy adhesive layer reduces to its steady value
over a time span of almost 1000 s. Moreover, if the applied load is raised gradually from
zero to its nal value of 100 N in sp seconds, its rate of increase has a direct effect on the
peak shear stress developed in the adhesive layers. The inuence of the adhesive thickness,
as well as the effect of adhesive viscosity and its shear modulus (in terms of their ratio),
on shear stress distribution in the adhesive joint are also investigated.
Keywords: adhesive joint; double-lap joints; polymeric resins; viscoelasticity; residue
theorem

1. Introduction
One of the most efcient methods of joining structures is the use adhesives. Compared to
mechanical joints such as bolts or rivets, adhesive joints are marked by their ease of assem-
bling and spreading the stress over a larger region. Consequently, the use of adhesives for
bonding structures is steadily growing in more recent applications. Adhesive bonded joints
are already playing a signicant role in development and production of various structures.
When used to bond polymer-matrix composites, the adhesives can be selected from the same
family of materials to assure a good compatibility. However, adhesives are aw dominated
and due to their viscoelastic behavior, time-dependent failures may occur. Thus, in the design
of adhesively bonded joints, to ensure their safety, it is necessary to examine stress distribu-
tion developed in the joint. Many researchers have performed different theoretical analyses
for single-lap and double-lap adhesive joints. To achieve tractable results and quantitative
agreement with experimental results, their works have been based on certain assumptions
and/or simplications.
One of the earliest analytical methods present in the literature for stress analysis of
bonded joints is the work of Volkersen.[1] He proposed a shear lag model to account for the

*Corresponding author. Email: [email protected]

2013 Taylor & Francis

 2234 M. Shishesaz and A. Reza

non-uniform shear stress distribution along the adhesive bondline. The major assumptions in
his model were that the adhesive deforms only in shear, while due to eccentric load paths
applied to each adherend layer, bending effect on the joint was neglected. Goland and
Reissner [2] took this effect into account and predicted a more non-uniform shear stress
prole in the adhesive bondline than that postulated by Volkersen. One of the major aspects
of Goland and Reissner analysis is that it predicts the excessive adhesive shear strains at the
edges of the joint, which is due to the elastic bending of the adherends. Allen and Shanahan
[3,4] studied the creep behavior of a phenolic resin/polyvinyl formal composite and a modi-
ed epoxy novolac adhesive in their glassy state when used to bond high tensile steel single
lap joints. Both characteristic delay times and creep behavior have been explained by zero
order rate theory using an asymmetric potential energy barrier. It has been shown that adher-
end separation is associated with the maximum adhesive shear stresses found at the ends of
the overlap.
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

An extensive literature review on existing analytical models for both single and double-
lap joints has been made by Baldan [5] and da Silva et al. [6,7] to assist any designer to
choose the proper model for a particular application. A simplied one-dimensional approach
was developed by Her [8] to model the adhesively bonded single and double-lap joints. He
studied the effect of various parameters such as thickness and moduli of the adhesive and
adherend layers, on stress distribution in the joint. Good agreement was observed between
analytical solution and nite element (FE) results. Tsai and Morton [9] investigated stress dis-
tributions in laminated composite double-lap joints with unidirectional and quasi-isotropic
adherends. They performed theoretical and experimental approaches. According to their
experimental and FE results, each adherend layer experiences a signicant shear deformation
which has been neglected in their theoretical model.
Most of the adhesives used in structural applications have a viscoelastic (viscoplastic)
behavior in nature, especially at high temperatures and high stress levels.[10] Redistribution
of stresses and strains that occur in such joints, inuence the joint strength considerably. Most
of the engineering viscoelastic materials, such as polymers, follow the Boltzmanns superposi-
tion principle [1114] and so their constitutive stress-strain relationships can be expressed in
terms of either linear differential or linear integral operators. Yadagiri et al. [15] used a direct
formulation for viscoelastic analysis of adhesively bonded joints using FE method. They
developed a six-noded quadratic isoparametric element and used hereditary integrals to repre-
sent the stress-strain relations. Groth [16] studied stress distribution in single lap joints (with
and without a crack) by 3D FE method. He used viscoelastic and viscoplastic behaviors for
the adhesive layer. Pandy and Narasimhan [17] applied 3D viscoplastic analysis of an adhe-
sively bonded single lap joint, using FE method. Their model included the material and geo-
metric nonlinearity. They treated the adhesive layer as elasto-viscoplastic and the adherend as
linearly elastic. The geometric nonlinearity in the single lap joint was due to nite rotation of
the joint. The stress-strain relation in their adhesive material was selected according to
RambergOsgood equation. Compared to elastic solution, their viscoplastic analysis resulted
in smaller stresses at the end of the overlap region. Sato [18] investigated a mathematical
method to treat time dependent models with residual stresses induced in an adhesive layer
conned between two adherends with different curvatures.
In this work, using residue theorem, the effect of viscoelasticity of polymeric adhesives
on shear stress distribution in the adhesive layer of a double lap joint is studied. The pro-
posed model is consisted of three similar adherend layers joined by two adhesive layers, as
shown in Figure 1. Each adherend layer is modeled as linear elastic, while the adhesive layers
are assumed to behave as linearly viscoelastic. The results are deduced for epoxy adhesive,
and hence, three parameters viscoelastic solid model is used to derive the governing
 Journal of Adhesion Science and Technology 2235

differential equations. These equations are then transformed into Laplace domain and residue
theorem is used to get a distribution for the shear stress in each adhesive layer (in the time
domain). The inuence of two loading conditions, adhesive thickness, loading rate and the
ratio of polymer viscosity and its shear modulus are studied on shear stress (strain)
distribution in each adhesive layer.

2. Model description
The proposed model is consisted of three adherend layers joined by two polymeric adhesive
layers, as shown in Figure 1. The subscripts t and b are used to correspond to the top and
bottom adhesive layers. Also, the adherend layers are numbered as 1, 2 and 3, from top to
bottom respectively. The following assumptions are used to model the problem:
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

Each adherend layer behaves as a linear elastic material.

Each adhesive layer behaves as a linear viscoelastic material.
The tensile load P0 acts along the x-axis and is applied to mid-adherend layer.
The 3-parameter viscoelastic solid model is used to show the viscoelastic behavior of
each adhesive layer.[15,16]
The adhesive layers do not sustain any axial load, and hence, undergo only pure shear.
Due to the state of plane stress condition, out-of-plane normal stresses are ignored in
both the adhesive and adherend layers.[10]

3. Governing equations in the adhesive layer

To derive the governing equations, it is also assumed that the shear stress in the adhesive
layer is constant along the width of the joint. Moreover, it is assumed that ov=ox is negligible
compared to ou=oy (v is the vertical deformation in the adhesive along y-axis). Using this
assumption along with a linear distribution for axial deformation u along the adhesive
thickness, lead into the fact that shear stress at any section along the thickness of the adhesive
layer remains constant (sxy ou=oy).[5,10] According to these assumptions, as shown in
Figure 2, a force balance on each adherend layer of the joint, with a length x, gives;

Figure 1. A double lap joint under shear.

 2236 M. Shishesaz and A. Reza
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

Figure 2. Force balance on an element of the joint.

oT1 x; t oT2 x; t
w sx; t 0 and  2w sx; t 0 1
ox ox

In Equation (1), w refers to the width of the joint (see Figure 1). Due to symmetry in
arrangement of the layers in the joint, T^1 x; s T^3 x; s. Indices 1, 2 and 3 refer to the top,
middle and bottom adherend layers respectively while T^ x; s is the tensile load in each
adherend layer in Laplace domain. Assuming linear elasticity, the strain in each adherend
layer may be written as:

ou1 x; t T1 x; t
e1 x; t where e1 x; t 2a
ox E 1 d1 w

ou2 x; t T2 x; t
e2 x; t where e2 x; t 2b
ox E 2 d2 w

According to Equation (2), one may write;

ou1 x; t T1 x; t ou2 x; t T2 x; t
; 3
ox E1 d1 w ox E 2 d2 w

As a rst approximation, the adhesive shear strain is taken as

u2 x; t  u1 x; t
cx; t 4
h

where, h is the adhesive thickness. For epoxy adhesive with viscoelastic properties, the rela-
tionship between shear stress and strain is given as [11]:
 Journal of Adhesion Science and Technology 2237
Z t
ocx; 1
sx; t Gt  1 d1 5
0 o1

In Equation (5), sx; t is the shear stress, G and c are the adhesive relaxation shear modu-
lus and shear strains respectively, and 1 is the variable of integration. Using the Laplace trans-
form, Equation (5) becomes;

^  s^
s^x; s G ^ cx; s
cx; s sGs^ 6

^  s in the
Equation (6) is called Alfreys correspondence principle.[19] The quantity G
transformed domain is analogous to the Youngs shear modulus for a linear elastic material.
Moreover, in Laplace domain, Equations (1)(4) may be written as;
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

oT^ 1 x; s oT^ 2 x; s
w s^x; s 0 and  w^
sx; s 0 7
ox ox

u1 x; s T^ 1 x; s
o^ o^u2 x; s T^ 2 x; s
and 8
ox E 1 d1 w ox E2 d2 w

u2 x; s  ^u1 x; s
^
c^x; s 9
h

Upon differentiating Equations (6) and (7), one may write;

o^ ^ o^
sx; s sGs u2 x; s o^u1 x; s

 10
ox h ox ox

o2 T^ 1 x; s o^
sx; s
w 11
ox2 ox

Substituting Equation (10) back in (11) we have;

o2 T^ 1 x; s ^ T^ 2 x; s T^ 1 x; s
sGs
  12
ox2 h E 2 d2 E1 d1

Since;

T^ 2 x; s Ps
^  2T^ 1 x; s 13

Then, substituting for T^ 2 x; s from Equation (13) back in Equation (12) gives;

^  2
o2 T^ 1 x; s sGs 1 ^
 ^ Ps
sGs ^
 T 1 x; s  14
ox2 h E2 d2 E1 d1 E2 d2 h

The following boundary conditions are used to solve Equation (14).

 2238 M. Shishesaz and A. Reza

^
Ps
T^ 1 l; s 0; T^ 1 l; s 15
2

Upon proper application of boundary conditions, one may write;

   
^
Ps 1 sinhksx 1 E2 d2  2E1 d1 coshksx 2E1 d1
T^ 1 x; s  16a
2 2 2sinhksl 2 E2 d2 2E1 d1 coshksl E2 d2 2E1 d1

   
^ ^ 1 sinhksx 1 E2 d2  2E1 d1 coshksx 2E1 d1
T 2 x; s Ps 1   16b
2 2sinhksl 2 E2 d2 2E1 d1 coshksl E2 d2 2E1 d1

Where
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

s
 
q
^ 1 2 1
ks c sGs & c 17
h E2 d2 E1 d1

With the help of Equation (1), shear stress in the adhesive layer may be expressed as:
   
^
Psks cos hksx E2 d2  2E1 d1 sinhksx
s^x; s  ; l\x\1 18
4 sinhksl E2 d2 2E1 d1 coshksl

where, at x l; l; s^x; s 0. In this study the effect of two different loading conditions
are examined on the shear stress distribution in each adhesive layer. In the rst case, the
applied load (see Figure 3) is expressed in terms of a step function Pt P0 , where, in
Laplace domain it may be represented by Equation (19).

^ P0
Ps 19
s

Figure 3. Types of applied loadings to the joint.

 Journal of Adhesion Science and Technology 2239

In the second case, the ramp portion of the applied load varies from zero to P0 at
0 6 t 6 sp , and remains constant afterwards (see Figure 3). Here, the applied load in Laplace
domain may be expressed as in Equation (20).
 
^ P0 1  esp s
Ps 20
sp s2

Now, substituting Equations (17) and (19) in Equation (18), the expression for shear stress
(in adhesive layer) in Laplace domain is equal to;
2  q   q 3
s ^ x ^ x
^ 6 cosh c s Gs sinh c sGs
Gs 7
s^x; s a 6  q   b  q 7
s 4 5; l\x\1 21
^ ^
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

sinh c sGsl cosh c sGsl

In Equation (21), a and b are constants dened as in Equation (22).

 
cP0 E2 d2  2E1 d1
a ;b 22
4w E2 d2 2E1 d1

Taking the inverse Laplace transform of Equation (21), an expression for shear stress as a
function of time in adhesive layer is obtained. The three parameters linear viscoelastic solid
model is used to model the epoxy resin.[20,21] This model is shown in Figure 4. The agree-
ment between this model and actual time-dependent material behavior is acceptable for a
wide range of materials with low as well as high viscosity. The governing differential equa-
tion for this model is given in Equation (23).

r p1 r_ q0 e q1 e_ 23

Where the coefcients q0 , q1 and p1 are constants related to the viscoelastic properties of
the epoxy adhesive. These coefcients are given in Equation (24).

g2 G1 G2 g2 G 1
p1 ; q0 ; q1 24
G1 G2 G1 G2 G1 G2

The relaxation module of this model is given by;

^ q0 q1 s
Gs 25
s1 p1 s

Figure 4. Three parameter linear solid model.

 2240 M. Shishesaz and A. Reza

Now transforming Equation (21) into the time domain, the adhesive shear stress may be
expressed as a function of time. To do so, the inverse Laplace integration is obtained using
the residue theorem.[2225] On proper application of this theorem (see appendix A), shear
stress distribution in each adhesive layer, in time domain, may be expressed as in Equation
(26).
p p p p p p
a q0 e2cx q0 1 ec q0 lx ab q0 e2cx q0  1ec q0 lx
sx; t p  p
e2cl q0  1 e2cl q0 1
s s
X 1
1n q 0 q 1 s n sn t npx 1n q0 q1 s0n s0n t
2a e  cos b e
n1
usn s2n 1 p1 sn l us0n s0n2 1 p1 s0n
 !
2n  1px
 sin 26
2 l
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

Now, Equation (6) is used to calculate the creep response of the adhesive layer. Hence,
the shear strain in each adhesive layer may be written as;

s^x; s
c^x; s 27
^
sGs

Substituting Equation (21), in (27), shear strain in the adhesive layer, in Laplace domain
is equal to;
2 q q 3
coshc s ^
Gs x sinhc ^ x
sGs
1 6 7
c^x; s aq4 q  b q 5 28
^
s3 Gs ^ l
sinhc sGs ^ l
coshc sGs

Using the same procedure as before, the creep response of the adhesive layer in the time
domain is;
p p p p
ae2cx q0 1ec q0 lx abe2cx q0  1ec q0 lx aq1  p1 q0 qq0 t
cx; t p 2clp  p 2clp e 1
q0 e q0  1 q0 e q0 1 q0 q1
s
X 1
1n q 0 q 1 s n sn t npx
2a e  cos
n1
usn s2n 1 p1 sn l
n
s  !
1 q0 q1 s0n s0n t 2n  1px
b e  sin 29
us0n s0n2 1 p1 s0n 2 l

4. Finite element analysis

The 2-D FE model of the joint was prepared and solved for stress distribution in the adhesive
layer using the commercial FE code, ANSYS v.11. The applied load and boundary conditions
are as shown in Figure 5(a) and (b). The imposed boundary conditions were selected to simu-
late the state of pure shear in the joint (this assumption was used to derive the governing dif-
ferential equations). This is shown in terms of zero lateral displacements on all nodes, to
prevent bending of the adherend and adhesive layers. Each adherend and adhesive layer was
 Journal of Adhesion Science and Technology 2241
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

Figure 5. Finite element model of the joint and its meshed state (a): nite element model of the lap
joint (b): portion of the meshed model.

meshed using PLANE183 element. This is an 8-node element and has quadratic displacement
behavior and may be used as a plane element in plane stress, plane strain and generalized
plane strain conditions. Prony series, as expressed in Equation (30), is used to present the vis-
coelastic behavior of the adhesive layers.
" #
X
N
t=sj
Gt G0 n1 nj e 30
j1

The initial shear elastic modulus G0 and nj are dened as

X
N
Gj
G0 G1 Gj ; nj 31
j1
G1

nj and sj are the relative shear modulus and relaxation time respectively. Properties of
similar layers were considered to be the same. The FE ndings are compared with those of
analytical solution, as will be discussed in the subsequent section.

5. Results and discussion

The main objective of this work is to determine the creep response of polymeric adhesives
under shear in a double lap joint under various conditions. Equilibrium equations indicate that
 2242 M. Shishesaz and A. Reza

in a double lap joint, shear stress distribution in the adhesive layer is a function of physical
and mechanical properties of all layers. The required values for epoxy adhesive used in this
work were deduced from graphs in Ref. [14] and after some manipulation, along with value
of other parameters, are given in Table 1. The overlap length of the joint was assumed to be
20 mm, with 1 mm thick aluminum adherend layers adhered to epoxy adhesive. The effect of
adhesive thickness, as well as other parameters, is then studied on the creep response of each
adhesive layer. In this process, two type of loading conditions were examined. In the rst
type, a step load of P0 = 100 N was applied to the joint (load type 1), while in the second case
(load type 2), a ramp load with a nal value of P0 = 100 N was used. Here it was assumed
that the applied load rises from zero to its nal value of P0 over a time period of 0 6 t 6 sp ,
and remains constant afterwards (see Figure 3).
Figure 6 compares shear stress distribution in the adhesive layer along the overlap region.
The results were obtained based on analytical method and those of FE solution at different
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

time instants. The magnitude of the step load was taken to be P0 = 100 N. As it appears, in all
cases, there is an excellent match between the results of the two methods. In addition, the
peak values of shear stress (at different time instants) appear to be at the right end of the
overlap region. As realized, the elastic analysis of the joint (results at t = 0), overestimate
the peak shear stress developed in the adhesive layer. This value substantially decreases with
an elapse in time. For further comparison, the results based on elastic behavior of the adhe-
sive layer [8] are superimposed. According to this gure, the results of linear elasticity solu-
tion overlap those of viscoelastic analysis at time equal to zero. As time elapses to almost
1000 s, the magnitude of the peak shear stress is reduced substantially to about 35% of its

Table 1. Mechanical and physical properties of adherend and adhesive layers.[14]

Material E (GPa) b (mm) G1 (GPa) G2 (GPa) g2 (GPa s)

Adherends: aluminum 69 40 0.33
Adhesive: epoxy 3.2 40 0.4 1.1429 0.1714 58.82

Figure 6. Comparison of FE and analytical results on shear stress in adhesive layer.

 Journal of Adhesion Science and Technology 2243

initial value. Beyond this point, any further increase in time does not seem to affect shear
stress distribution along the joint. Figure 7 shows the peak values of these stresses (based on
analytical method and FE solution) with an increase in time, for the same step load given
above. An excellent match is observed between the results of the two methods. According to
this gure, based on the selected data, the relaxation time appears to be about 1000 s for
epoxy adhesive.
According to Figure 8, for load type 2, the rate of increase in the applied load P0 has a
direct effect on the peak shear stress developed in the adhesive layer. For example, increasing
the loading rate from 3.12 to 100 N/s will increase the peak shear stress in the joint by almost
18%. This means that holding P0 constant, increasing sp will lower the induced peak shear
stress in the adhesive layer. As the loading rate increases, the peak value of shear stress in
the joint, based on load type 2, approaches that of load type 1 (step load). Figure 9 shows
the effect of adhesive thickness on shear stress distribution in the adhesive layer. The applied
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

step load is equal to P0 = 100 N. According to this gure, although the peak shear stress is
higher at the right end of the joint, increasing the thickness of the adhesive layer from 0.01
to 0.1 mm, lowers the peak shear stress to almost 1/3 of its initial value at the both ends.
Higher values of adhesive thickness seem to smooth out the rise in shear stress which occurs
near the overlap ends. According to Figure 10, for the step load of P0 = 100 N, with an
increase in adhesive thickness, not only the peak values of shear stresses which occur at the
right end of the joint are reduced, but also their steady state values seem to get close to the
value of 0.22 MPa, after relaxation time is reached (as stated earlier, based on the given data,
this time appears to be 1000 s). This gure also indicates that the effect of any change in
adhesive thickness on the its induced shear stress is more pronounced at t = 0. According to
Figure 11, for load type 2, for a loading rate of 8 N/s, the peak shear stress in the adhesive
layer occurs after almost ten seconds (t = 10 s). The occurrence of this stress seems to be inde-
pendent of the adhesive thickness. The magnitude of these peak stresses seem to be in the
same order as those for load type 1 (see also Figure 10).
The effects of adhesive viscosity and its extensional stiffness (in terms of their ratio), on
the peak shear stress in the adhesive layer are shown in Figures 12 and 13, for load types 1

Figure 7. Comparison of peak shear stress vs. time.

 2244 M. Shishesaz and A. Reza
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

Figure 8. The effect of loading rate on the peak shear stress developed in the adhesive layer.

Figure 9. The effect of adhesive thickness on its induced shear stress at time zero (a): left side of
adhesive layer (b): right side of adhesive layer.

and 2 respectively. Assuming q g2 =G2 , then according to Figures 12 and 13, doubling the
value of q (from 173 to 346), will lower the peak value of shear stress by about 27% for load
type 1, and 23% for load type 2. The value of q 346 corresponds to epoxy adhesive. Other
 Journal of Adhesion Science and Technology 2245
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

Figure 10. The effect of thickness of adhesive on shear stress in adhesive layer with respect to time
(load type 1).

Figure 11. The effect of thickness of adhesive on shear stress in adhesive layer with respect to over
time (load type 2 with sp 8 s).

values of q in these two gures do not necessarily represent a certain material. They are only
selected to demonstrate the effect of viscous to extensional properties of polymers, in terms
of their ratio, on the peak shear stress developed in the adhesive layer.
 2246 M. Shishesaz and A. Reza
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

Figure 12. The effect of q on the peak shear stress developed in the adhesive layer, load type 1.

Figure 13. The effect of q on the peak shear stress in the adhesive layer, load type 2.

6. Conclusions
This paper presents the analytical and FE study of shear stress distribution in a viscoelastic
adhesive layer, namely epoxy resin, in a double lap joint under shear. The shear-lag model
was used to obtain the governing differential equilibrium equations. These equations were
solved using residue theorem. The results for shear stress distribution based on analytical
method match exactly those obtained by FE solution. Also, for epoxy adhesive, based on a
step load of P0 = 100 N, over a time span of 1000 s (relaxation time), the peak value of shear
stress in the adhesive layer is reduced to 1/3 of its initial value and remains constant
 Journal of Adhesion Science and Technology 2247

afterwards. Moreover, if the applied load is gradually increased to its nal value of P0, its rate
of increase has a direct effect on the peak shear stress developed in the epoxy adhesive layer.
For an increase in loading rate from 3.12 to 100 N/s, this increase is about 18%. The results
indicate that in a double lap joint, although the peak shear stress is higher at the right end of
the joint, for both load types, increasing the thickness of the adhesive layer from 0.01 to
0.1 mm, lowers the peak value of shear stress to almost 1/3 of its initial value (at both ends
of the overlap region). Also, the ratio of viscous effect to the shear modulus of the adhesive
layer, namely q, has an adverse effect on the peak shear stress developed in the joint. Dou-
bling this value (from 173 to 346), lowers the peak shear stress by 27% for load type 1, and
23% for load type 2. According to the results, for the step load of P0 = 100 N, it appears that
the shear stress in epoxy adhesive reaches its steady value after 1000 s, regardless of the
adhesive thickness. This behavior was also observed for load type 2.
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

Nomenclature

c parameter dened in Equation (17)

d adherend thickness
E elastic modulus of the adherend
G0 adhesive initial shear modulus
G1 shear modulus of spring 1 (in 3 parameter linear solid model, see Figure 4)
G2 shear modulus of spring 2 (in 3 parameter linear solid model, see Figure 4)
G1 equilibrium shear modulus of the adhesive
Gt shear relaxation modulus function
^
Gs shear relaxation modulus function in Laplace domain
h adhesive thickness
l half length of the adhesive layer
p1 coefcient dened in Equation (24)
P0 tensile applied load
^
Ps tensile applied load in Laplace domain
q0 coefcient dened in Equation (24)
q1 coefcient dened in Equation (24)
t time parameter
T tensile load in each adherend layer
u longitudinal displacement of each adherend layer
v lateral displacement
w adherend width
x longitudinal global coordinate
y transverse global coordinate
a parameter dened in Equation (22)
b parameter dened in Equation (22)
c adhesive shear strain
e axial strain in each adherend layer
g2 shear damping of the dashpot (in 3 parameter solid model)
k parameter dened in Equation (17)
nj relative shear modulus in Prony series
q viscosity coefcient of polymeric adhesive
sj relaxation time in Prony series
sp time at which the ramp load becomes constant
sxy shear stress in xy plane
 2248 M. Shishesaz and A. Reza

References
[1] Volkersen O. Die Nietkraftverteiligung in zugbeanspruchten nietverbindungen mit konstanten
Laschenquerschnitten. [Stress distribution of bonded joint under tensile stress with constant cross
section of straps] Luftfahrtforschung. [Aerosp. res.] 1938;15:41.
[2] Goland M, Reissner E. The stresses in cemented joints. J. Appl. Mech. 1944;11:A17A27.
[3] Allen KW, Shanahan MER. The creep behaviour of structural adhesive joints I. J. Adhes.
1975;7:161174.
[4] Allen KW, Shanahan MER. The creep behaviour of structural adhesive joints II. J. Adhes.
1976;8:4356.
[5] Baldan A. Adhesively-bonded joints in metallic alloys, polymers and composite materials: mechan-
ical and environmental durability performance. J. Mater. Sci. 2004;39:47294797.
[6] da Silva LFM, Dasneves PJC, Adams RD, Spelt JK. Analytical models of adhesively bonded joints
Part I: literature survey. Int. J. Adhes. Adhes. 2009;29:319330.
[7] da Silva LFM, Dasneves PJC, Adams RD, Spelt JK. Analytical models of adhesively bonded joints
Part II: comparative study. Int. J. Adhes. Adhes. 2009;29:331341.
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

[8] Her S. Stress analysis of adhesively-bonded lap joints. J. Comput. Struct. 1999;47:673678.
[9] Tsai MY, Morton J. An investigation into the stresses in double-lap adhesive joints with laminated
composite adherends. Int. J. Solids Struct. 2010;47:33173325.
[10] Dillard DA, Pocius AV. Adhesion science and engineering I: the mechanics of adhesion. 1st ed.
Amsterdam: Elsevier Science BV; 2002. Chapter 12.
[11] Drezdov AD. Mechanics of viscoelastic solids. New York (NY): John Wiley; 1998, Chapter 1.
[12] Riande E, Diaz-Calleia R, Prolonco MG, Masegosa RM, Salom C. Polymer viscoelasticity stress
and strain in practice. New York (NY): Marcel Dekker Inc; 2002. Chapter 5.
[13] Shaw MT, Macknight WJ. Introduction to polymer viscoelasticity. 3rd ed. New Jersey (NJ): John
Wiley; 2005. Chapter 2.
[14] Ferry JD. Viscoelastic properties of polymers. 3rd ed. New York (NY): John Wiley; 1980. Chapter
1.
[15] Yadagiri S, Reddy CP, Reddy TS. Viscoelastic analysis of adhesively bonded joints. J. Comput.
Struct. 1987;27:445454.
[16] Groth HL. Viscoelastic and viscoplastic stress analysis of adhesive joints. Int. J. Adhes. Adhes.
1990;10:207213.
[17] Pandey PC, Narasimhan S. Three-dimensional nonlinear analysis of adhesively bonded lap joints
considering viscoplasticity in adhesives. J. Comput. Struct. 2001;79:769783.
[18] Sato C. Stress estimation of joints having adherends with different curvatures bonded with
viscoelastic adhesives. Int. J. Adhes. Adhes. 2011;31:315321.
[19] Brinson HF, Brinson LC. An introduction to polymer engineering science & viscoelasticity. New
York (NY): Springer; 2008.
[20] Huang M, Niu X, Soboyejo WO. Creep induced rate effects on radial cracks in multilayered
structures. J. Mater. Sci. 2007;18:6569.
[21] Andrianov IV, Topol H, Weichert D. Load transfer in bre-reinforced composites with viscoelastic
matrix: an analytical study. J. Appl. Mech. 2009;79:9991007.
[22] Cohen AM. Numerical methods for laplace transform inversion. New York, NY: Springer; 2007,
vol. 5, Chapter 2.
[23] Kythe PK, Schaferkotter MR. Handbook of computational methods for integration. New York, NY:
Chapman & Hall/CRC Press; 2005, Part 1, Sec. 32.
[24] Polyanin AD, Manzhirov AV. Handbook of mathematics for engineers & scientists. London: Chap-
man & Hall/CRC, Taylor & Francis Group; 2007, Chapter 11.
[25] Bellman R, Kalaba RE, Lockett JA. Numerical inversion of the laplace transform: application to
biology, economics, engineering and physics. New York, NY: Elsevier Publishing Company; 1966,
Chapter 4.
 Journal of Adhesion Science and Technology 2249

Appendix A

Analytical solution using residue theorem

The rst term in Equation (21) may be written in time domain as in Equation (A.1). This is known
as the Bromwich integral (also as Mellin integral).
Z
1
s1 x; t est s^1 x; sds A:1
2pi C

For simplicity, allowing Xx; s est s^1 x; s, then one can write;
 q 
^ x
r est cosh c sGs
q0 q1 s
Xx; s a 2   q  A:2
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

s 1 p1 s ^ l
sinh c sGs

Hence, Equation (A.1) may be re-written as;

Z
1
s1 x; t Xx; sds A:3
2pi C

The problem of inverting the Laplace transform can often be solved analytically by applying a
partial fraction expansion or integration along some contour in the complex s-plane. Usually, to estimate
the complicated integrals, residue theorem is used. According to this theorem, by summing all the
residues of Xx; s at its poles and multiplying them by 2i, the integral term in Equation (A.3) can be
expressed as;
Z X
n
Xx; sds 2pi Rk A:4
C k1

Where, Rk are the residues of Xx; s at s = sk. Hence, by substituting Equation (A.4) in Equation
(A.3), the rst term in Equation (21) can be obtained in the time domain as;

X
n
s1 x; t Rk A:5
k1

It is evident that Xx; s has a simple pole at s 0, while a second order pole exists at s 1=p1 .
One may show that residue of the pole at s 0 is equal to:
 q 
^ x
r e cosh c sGs
st
q0 q1 s
Rjs0 lims  0  a 2   q 
s!0 s 1 p1 s ^ l
sinh c sGs
p p p
a q0 e2cx q0 1ec q0 lx
p A:6
e2cl q0  1

While at s
 q  1=p1 , the residue of the pole is equal to zero. For other simple poles, it follows that
^
sinh c sGsl 0. Consequently;
 2250 M. Shishesaz and A. Reza
q  2
^ n npi ) sn Gs
sn Gs ^ n  np m2 A:7
cl cl

Hence, with the help of Equation (25), one may write;

q0 m 2
sn  A:8
q1 p1 m 2

It may be shown that residue of the pole at s sn takes the form of;
 q 
r st
e cosh c sGs ^ x
q0 q1 s
Rn Rjssn lims!sn a 2  q
 q  A:9
s 1 p1 s d ^ l cosh c sGs ^ l
c s Gs
Downloaded by [Mohammad Shishesaz] at 01:40 15 September 2013

ds

For simplicity, we now consider a portion of the denominator of Equation (A.9) as;
q

us
d ^ l clq1  qq
c sGs
0 p1
A:10
ds 21 p s q0 q1 s
2
1 1p1 s

Then, Equation (A.9) may be written as;

s s
q0 q1 s n esn t coshnpx i 1
n
q0 q1 s n npx
Rn a 2  l
a  esn t cos A:11
sn 1 p1 sn usn coshnpi usn sn 1 p1 sn
2 l

With the help of Equations (A.11) and (A.1), the rst term in Equation (21) for the shear stress, in
time domain, may be written as in Equation (A.12).

p p p s
a q0 e2cx q0 1ec q0 lx X1
1
n
q0 q1 s n npx
s1 x; t p  A:12
sn t
2a e cos
e2cl q0  1 n1
usn s2n 1 p1 sn l

Using similar procedure, the second term in Equation (21) may be written in time domain. Choosing
a step function for the load, shear stress in the adhesive layer may be written as;
p p p p p p
a q0 e2cx q0 1ec q0 lx ab q0 e2cx q0  1ec q0 lx
sx; t p  p
e2cl q0  1 e2cl q0 1
s s
X 1
1
n
q 0 q 1 s n sn t npx 1
n
q0 q1 s0n s0n t
2a e  cos b e
n1
usn s2n 1 p1 sn l us0n s0n2 1 p1 s0n
 
2n  1px
 sin A:13
2 l

where,

2n  12 p2 0 q0 r 2
r2 and s  A:14
4c2 l 2 n
q1 p1 r 2