Fracture mechanics applied to the fracture of adhesively bonded joints 2009

Introduction

This study is about the effect of the adhesive thickness and adhesive ductility on the mixed
mode loading of Double Cantilever Beam joints. The project is divided into three main tasks.
The first task is about the development of a finite element model using a cohesive zone model
to design an experimental device based on the standard ASTM D6671D. The second task is
concerned with the generation of experimental results in mixed mode varying the type of
adhesive and the adhesive thickness. In the third task, the experimental results will be used to
develop an adhesive toughness model as a function of the mode mixity.

Objectives

What I have already done

a. Numerical analysis with ABAQUS

The  adhesive  is  modeled  by  interface  elements  distributed  along  the  bondline,  as  shown  in 
Figure 1, with the crosses at the interface, leaving a pre‐crack length a0 of 118 mm.  

 
Figure 1. Mixed Mode Bending model.

The  virtual  beam  supposed  to  simulate  the  experimental  device  proposed  in  ASTM 
D6671D (Figure 2) is defined by three cylinders (A, B and C) physically attached. This beam acts 
like a lever which is loaded at node A, transmitting a load at node B, and a pull force at node C. 

 
Figure 2. Experimental set-up
 

P.1 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

Figure 3 shows von Mises stress distribution obtained in ABAQUS with this model. The zoomed 
area shows clearly the damaged zone and the propagation of the crack tip.  

Figure 3. ABAQUS von Mises stress plot (bottom is a zoom of top).

 

b. Virginia Tech Tests

Figure 4. Dual actuator with specimen (a) Different mixity test with opposite loading (b) Different mixity test (c)

P.2 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

What to do now?

1. Testing specimens in mode II

2. Perfect the data acquisition process with the high speed camera image record and post-
processing.
3. Design a new apparatus to perform mix mode tests with steel adherents in a MTS
instrom machine.

1.  Testing Specimens in Mode II  

Mode II adhesive joint failure testing has no standard to define it. Mode II interlayer failure of
composite materials testing as end notched flexure (ENF), end load split (ELS) and four-point
end notched flexure showed in 0 are used.

a) End loaded split 

 
 
 

b) End notched flexure    c) Four‐point end notched flexure 

Figure 5. Mode II interlayer failure of composite materials tests.

 

The end loaded split test is sensitive to the griping conditions and presents some troubles to
determine GIIc for large displacements. The Four-point end notched flexure requires a special
apparatus and presents some difficulties with friction at the pre-crack region [i]. Despite the
difficulty to measure de crack length, the end notched flexure test (ENF) is widely used, mostly
by its simplicity (see figure 5).

 
Figure 6. ENF test diagram.

P.3 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

1.1 Closed Form Solution for GIIC  

Applying the beam theory analysis to each specimen adherend which supports half of the
applied load (P), the strain energy is calculated according to equation 1.

    1
2 2
 

Analyzing the beam in three segments (0  ;    ;   2 ), equation 1

becomes equation 2 for the flexure component and equation 3 for the shear component of the
energy.

⁄4 ⁄2 ⁄2
2   2
2 ⁄8 2 2
 

/
/
2         3
/ 2 2
 

Using equation 3 for τ and applying Castigliano’s theorem, the displacement of the load
application point (shown in figure 6) is given by equation 4 that differentiates into equation 5.

3 2 3 
  4
12 10 
 

9 
  5
16 

Besides the difficulty to measure the crack length, there is another drawback related with the
existence of a region damaged by non-elastic processes, named fracture process zone (FPZ).
Silva et al. [ii] found that the energy released at this fracture process zone must be accounted
and an equivalent crack length (ae) that contemplates this effect must be defined.

 
Figure 7. Schematic of the fracture process zone.
 

P.4 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

1.2  Test method  
 

ENF is the easier test and therefore it is widely used. The specimen is placed over two rods that
will support its weight and the loading, allowing the rotation. The rods are placed symmetrical in
relation to the loading point that should be half length of the specimen. During load history the
values of force, applied displacement and crack length (P, d and a) shall be registered in order
to calculate the critical strain energy release rate. Most of times it is difficult to accurately
measure the crack length.

1.2  Manufacture Program  

We will try to obtain the results for 1 mm and 0.5 mm thickness in order to adjust and determine
the compliance for these two thicknesses. Raul Campilho has already determined the
compliance for the 0.2 mm thickness of Araldite 2015

Tests must be done with proper specimens that shall be manufactured in accordance with the
following table.

Table1. Specimens for mode II testing

# Specimen Adhesive Thickness [mm] Due date

3 SIKAFLEX 255 1 26-03-2009
3 AV138+HV998 1
3 Araldite 2015 1
3 SIKAFLEX 255 0.5
3 AV138+HV998 0.5
3 Araldite 2015 0.5

18 TOTAL

1.3  Specimen Manufacture  
 

A good specimen is fundamental to achieve cohesive failure and obtain the best test data.
Therefore manufacture of specimens must be very well attended. The specimens manufacture
should be done according with the next steps:

1. Mould Preparation;
2. Beams preparation;
3. Spacers and razor blade preparation;
4. Adhesive mixture (when needed) and deposition;
5. Press placement and pressing stage;
6. Curing of the adhesive;
7. Removal and cleaning;
8. Preparation for testing.

P.5 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

1.3.1   Mould Preparation 
 

The mould (Figure 8), used to produce the bonded joints was design by Dr. Lucas. With this
mould we can make up to 6 specimens.

Figure 8 – Mould used to prepare de bonded joints

 

Before any use of the mould we have to guarantee e perfect clean surface without any burs and
remains portions of adhesive. If that happens we must remove it with an aluminium spatula
(Figure 9) and then pass acetone using a clean cloth (Figure 10).

Figure 9 – Remove remains portion of adhesive

 

P.6 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

 
Figure 10 – Cleaning with acetone and paper sheet.
 

After cleaning the mould, we covered the surface with a release agent. In this case the
FREKOTE 770NC.

The parts were heated to 60ºC in a hot press to speed up the cure reaction of the release agent.

After the parts reached 45 ºC, three coats are applied with 5 minutes interval between. The time
was the necessary to the complete solvent evaporation. All parts must have coated carefully. All
security procedures are used to avoid health damage.

The razor blade and the spacers had been surface-treated with release agent before the
specimen parts assembly.  

Figure 11 – Mould release agent application

 

P.7 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

1.3.2 Beams preparation ­ Sand blasting 
 

The beams must be prepared before applying the adhesive. Sand blasting is a common 
treatment, allowing to obtain a good surface for adhesion. 

Beams must be placed inside the sandblast machine and the surface must be  sandblasted. 

Specimens Sandblast air gun

Air gun 

Protective gloves 

Figure 12 - Specimens placement inside the sandblasting machine in INEGI.

 

 
Figure 13 - Working with the sandblaster 

P.8 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

Figure  13  shows  the  operator  with  his  hands  in  the  sandblaster  protective  gloves,  looking 
through the screen and pressing the foot actuator for the jet airflow. 

The  surface  should  be  sandblasted  in  a  manner  that  the  final  finishing  looks  like  sand,  with 
high rugosity. 

Sandblasted surface   
Figure 14 - Sandblasted surface in the beams.

1.3.3 Spacers and razor blade preparation 
 

The glue line thickness must be calibrated with a spacer in the top end and a razor blade with 
spacer in the initial crack length tip. This is an important step of the specimens manufacture, 
because it is crucial to obtain a correct thickness for the adhesive layer. 

There are several feeler gauges thicknesses (Figure 15) in order to measure different gapes
and by this allowing to calibrate the gape in between the two beams of the specimen.

Figure 15 - Feeler gauges stock.

P.9 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

It is known that a razorblade has 0.10 mm in thickness.

Figure 16 - Razor blades used to promote the initiation of the crack (placed at the crack tip).

The next table explains how to obtain the correct combination for spacers and razorblades for
the different layer thicknesses.

Table 2. Spacer combinations to obtain the wanted adhesive thickness.

Adhesive
Razor
layer
Top spacer blade Razor blade bottom spacer Razor blade upper spacer
thickness
[mm]
[mm]

2 2x 1 mm 0.1 2x 0.4 mm + 1x 0.1 mm 1x 0.05mm 2x 0.4 mm + 1x0.1 mm + 1x 0.05mm

1 1x 1 mm 0.1 1x 0.4 mm + 1x 0.05 mm 1x 0.4 mm + 1x 0.05 mm
0.5 1x 0.5 mm 0.1 1x 0.2 mm 1x 0.2 mm
0.2 1x 0.2 mm 0.1 1x 0.05 mm 1x 0.05 mm

Razorblades are cut in two symmetric blades. This allows to use one razorblade for two crack
initiation spacers.

Figure 17 - Cutting the razorblade in two symmetric blades.

P.10 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

Figure 18 - Cutting a 0.20 mm spacer.

Figure 19 - First and second operation of razor blade spacer gluing.

First operation is to glue the razorblade to the bottom spacer. In the second operation the
razorblade glued to the bottom spacer is glued to the top spacer resulting in the assembly
shown in figure 20 scheme.

Figure 20 – Razorblade, top and bottom gauge feeler assembly scheme.

P.11 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

The razorblade should be glued in such a manner that leaves the blade exposed as seen in the
next figure.

Figure 21 – Razorblade, top and bottom gauge feeler assembly scheme.

1.3.4 Adhesive preparation and deposition 

After preparing the surface by sandblasting, having the spacers and the razorblade assembly
we are ready to setup the bottom beams in the mould to apply the adhesive and close the upper
and bottom beams together.

To avoid problems with the adhesive pot life, the bottom beam must be already placed in the
mould and prepared with the spacers and razorblades before we can apply the adhesive.

Figure 22 - Cleaning with a paper tissue and placing the bottom beam on the mould.

P.12 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

Figure 23 – Placing the top spacers (gauge feeler in the left and razorblade assembly in the right)

Figure 24 – Three bottom beams prepared with spacers and razorblades assemblies.

Next step is to apply the adhesive. The majority of the structural adhesives is obtained by
mixing a resin with an hardener that works like a catalizer. Some adhesives are supplied in
dual container cartridges that are attached to a deposition mixing nozzle (Figure 25) which
promotes the mixture of the two components, avoiding the operator to mix them manually. Both
Araldite 2015 and SIKAFLEX 255 have mixing nozzles.

P.13 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

Figure 25 - Mixing nozzle.

Araldite AV138M is supplied in a separate container from the hardener HV 998,as seen in figure
26 a.

Figure 26 – a) Adhesive and hardener separate containers and b) mixture tools to mix the two components.

These two components need to be mixed manually in a proportion of 100 parts of AV138M to
40 parts of Hardener HV998, as seen in figure 26 b.

Figure 27 – Araldite 2015 application.

P.14 www.adfeup.co.cc
Fra
acture mech
hanics applie
ed to the fracture of adhesively
a bonded
b joints 2
2009
 

The Araldite
A 2015
5 is applied in strings side by side fro
om bottom to top and then back from top to
bottom (Figure 277). After the application of
o the adhes sive the uppe
er beam of th
he specimenn must
aced over the adhesive and
be pla a pressed to close it to ogether (Figu
ure 28).

F
Figure 28 – Pla
acing and pres
ssing the top beam.
b

For SIKAFLEX
S 25
55 FX there is the need tot apply prev 6 G+P primer. The primer (206
viously a 206
G+P)) must be applied just as painting the surface with
h a brush, shown in figure
e 29.

9 - Applying the
Figure 29 t primer fo
or SIKFLEX 255
2 - FX

1
1.3.5 Presss placemen
nt and pressing stage  

Whenn all the top beams are placed,

p the mould
m top pla
ate must also
o be placed u
using the thre
eaded
handles, that should be removved (unscrew
wing).

P.15 www.adffeup.co.cc
Fra
acture mech
hanics applie
ed to the fracture of adhesively
a bonded
b joints 2
2009
 

Threeaded 
handle 

Threaded 
handle 

Figure
e 30 – Mould closed.
c

The assembly
a of the mould closed
c and th
he specimenns must be ta
aken into the
e hydraulic heated
h
presss from intoco
o and a presssure shall be
b applied by
y closing the
e pressure va
alve until the
e load
cell re
eading gets to
t 7,000 psi.

Movement
Temperatu ure buttons and
controllers load cell

Pressure valve
ON/OFF
button

Figure 31 – Intoco hea

ated plates hy
ydraulic press machine.

P.16 www.adffeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

There should be made some verification about the alignment of the specimens in the mould,
before applying the full pressure.

Figure 32 - Verifying the lateral (left) and top (right) alignments.

1.3.6 Curing of the adhesive  

The temperature must be set in accordance with the adhesive curing time specifications defined
by the manufacturer. The next table summarizes the temperature and curing time for each
adhesive.

Table 3. temperatures and curing times.

Curing Time
Adhesive Temperature [ºC]
in the press outside the press

SIKAFLEX 255 FX Ambient (20) 24 Hours one week

AV138 + HV998 80 30 minutes two days

Araldite 2015 60 45 minutes two days

1.3.7 Removal and cleaning 
 

After the curing time in the press, the mould should be removed and the specimens can be 
taken out the mold. There may be the need for removing the pins of the mold. There is a 
proper tool from BETA and a hammer for that operation. 

Cleaning and removing the excess of adhesive that remains in the specimen sides, must be 
done only after the curing time outside the press is over. A knife or a piece of aluminum can be 
used to remove the excess of adhesive.  

P.17 www.adfeup.co.cc
Fracture mechanics applied to the fracture of adhesively bonded joints 2009
 

1.3.8 Preparation for testing 
 

Once the specimen is free of the adhesive in excess, the glue line must be clearly visible and 
maintain a constant thickness. A coat of white tint will be painted over the glue line. A 
corrector painter is usually the best option for that.  

After being painted in white, side rulers or scales can be attached or glued, helping the 
measurement of the crack length while it is growing during the test. 

                                                            
[i] Schuecker C., Devidson, B. D., (2000), in: Composite Materials: Theory and Practice (Grant P,
Rousseau C.Q., eds), ASTM STP 1383, p 334, ASTM, Philadelphia, PA.

[ii ] Silva, M. A. L., de Moura, M.F.S.F., Morais, J.J.L., (2006), Numerical analysis of the ENF test for
mode II wood fracture, Composites Part A, Vol In Press.

P.18 www.adfeup.co.cc