See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/279208954

Joining performance evaluation of different types of GEP224 glass/epoxy-to-

AZ31B magnesium alloy singlelap joints

Article in International Journal of Precision Engineering and Manufacturing · June 2015

DOI: 10.1007/s12541-015-0147-5

CITATIONS READS

3 167

3 authors, including:

Jungseok Kim Jae Yong Lim

Korea Railroad Research Institute Seoul National University of Science and Technology
113 PUBLICATIONS 1,538 CITATIONS 48 PUBLICATIONS 315 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Jungseok Kim on 31 July 2017.

The user has requested enhancement of the downloaded file.

 INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 16, No. 6, pp. 1135-1140 JUNE 2015 / 1135
DOI: 10.1007/s12541-015-0147-5 ISSN 2234-7593 (Print) / ISSN 2005-4602 (Online)

Joining Performance Evaluation of Different Types of

GEP224 Glass/Epoxy-to-AZ31B Magnesium Alloy Single-
lap Joints

Jung-Seok Kim1,#, Jae-Young Lim1, and Woo-Geun Lee2

1 New Transportation Systems Research Center, Korea Railroad Research Institute, 176, Cheoldobangmulgwan-ro, Uiwang-si, Gyeonggi-do, 437-757, South Korea
2 Railway System Engineering, University of Science & Technology, 217, Gajeong-ro, Yuseong-gu, Daejeon, 305-350, South Korea
# Corresponding Author / E-mail: [email protected], TEL: +82-31-460-5663, FAX: +82-31-460-5289

KEYWORDS: Joint, Magnesium, GFRP, Hybrid, Strength

The joining performance was evaluated for different types of GEP224 glass/epoxy-to-AZ31B magnesium alloy single-lap joints. To
achieve this goal, three types of single-lap joints with adhesive bonded, mechanically fastened and hybrid (bond/bolt) joints were
fabricated. The strength and failure mechanism of the joints was investigated and compared with each other. Additionally, the width
and joining torque effect was evaluated. From the test results, the load-displacement curves of the bonded and the bolted single-lap
joints showed a nearly linear and logarithmic curve shape until final failure, respectively. In the hybrid joints, they initially behaved
like bonded joints, while they followed a logarithmic curve shape, like the bolted joints, after a sudden load drop. The failure strength
of the bonded joint showed the biggest increase as the width was getting wider. In the bolted joint, the failure load showed a decreased
trend as the torque increased due to the early net tension failure of the GFRP adherends. In contrast, the failure load of the hybrid
joint increased as the torque increased.

Manuscript received: July 29, 2014 / Revised: December 24, 2014 / Accepted: March 11, 2015

1. Introduction of a structure is less than the mechanical joints. The adhesive joint has
a sealing effect, no stress concentration due to bolt holes, and no
Magnesium is the third most commonly used structural metal, damage in the FRP from the bonding process. These advantages allow
following iron and aluminum.1 It is also the lightest metal.2 Automotive the adhesive joints to have high fatigue strength. In the adhesive joints,
makers have focused on the application of magnesium alloys to however, each adherend has its own suitable adhesive. The selection of
automotive components since the early 1990s, especially in terms of adhesive is difficult for joints of different materials. In addition,
fuel efficiency and weight reduction. Though many magnesium parts degreasing and etching processes are necessary to obtain high joining
have been mass-produced for automobiles, few parts are widely used. strength during the general bonding process of metal adherends. To
Namely, such parts can be indicated as the steering wheel core and overcome the potential weakness of adhesive bonding, the bonded/
instrument panel (or cowl cross beam) in the interior part, and the engine bolted hybrid joints were proposed.5 In the hybrid joint, the mechanical
head cover in the powertrain parts.1 There are two major methods for fastening is added to the bonded joints to improve the joining strength.
joining different materials: mechanical fastening and adhesive bonding. However, fabrication cost would be increase in case of the hybrid joints
Mechanical fastening using bolts or rivets is simple, and it is possible The previous researches on the composite to metal joints were
to obtain high joining strength with a small scatter. Therefore, it is focused on the composite to steel or aluminum combination since they
widely used in metal structures. The disadvantages of mechanical joints are widely used metal. The composite to magnesium configuration in
are an increase in the weight of the entire structure and low sealing the bi-material joints was relatively rare so more study is needed.
performance. In addition, the presence of the bolt holes in mechanical Amanico et al.12 performed a feasibility study of the friction spot joining
joints decreases the cross-sectional area of structures, and increases technique on magnesium AZ31–O/glass fiber and carbon fiber reinforced
stress concentration.4-11 Using the adhesive bonding method, the weight poly(phenylene sulfide) joints. They evaluated the thermo-mechanical

© KSPE and Springer 2015

 1136 / JUNE 2015 INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 16, No. 6

Table 1 Material properties of GFRP and AZ31B

Material properties GEP224 (glass/epoxy) AZ31B*
E11 (GPa) 34.4 45
E22 (GPa) 13.2 45
G12 (GPa) 7.05 16.7
ν12 0.27 0.35
*Refer to the manufacturer’s sheets.

phenomena associated with the friction spot joining process promoted

metallurgical and polymer physical-chemical transformations.
In this study, the joining performance was experimentally investigated
and compared for different types of joining configurations. To achieve
these goals, three types of single-lap joints with GFRP (Glass fiber
reinforced plastics) and magnesium alloy adherends were fabricated.
The joints studied in this study are adhesively bonded joints, mechanical
fastened joints and hybrid joints. Through the tensile tests, the static
strength and failure mechanism was then investigated and compared
Fig. 1 Preparation process of the GFRP-magnesium bonded joints
with each other. Failure modes were observed using a microscope.

2. Specimen Preparation and Test

2.1 Configuration
To investigate and compare the static strength and failure mechanism
of the three different types of joints, the joint specimens were made.
The GFRP adherends were machined from glass fibre epoxy laminates
manufactured by the autoclave curing process. The laminates were cured
from four-harness satin fabric glass/epoxy prepregs (GEP224, SK
Chem., Korea). The laminate thickness and stacking sequences were 5
mm and [45/-45/45/0/-45/90/45/0/-45/90]s, respectively. The magnesium
adherends were cut from AZ31B magnesium plates (POSCO, Korea,
t=5 mm). Table 1 lists the material properties of the materials used. Fig. 2 Curing cycle of the bonded joints
For the bonded joint, both GFRP laminate and magnesium plates of
size 254 mm × 254 mm were bonded together in a single-lap joint
configuration. The adherend surfaces were abraded using 400 grit sand
paper and rinsed with acetone prior to bonding. FM73 adhesive film
(Cytec, USA) was used for the bonding of two adherends. Fig. 1 shows
the preparation process of the GFRP-magnesium bonded joint. The
vacuum-packed specimen was cured in an oven based on the curing
cycle as shown in Fig. 2.
In the case of bolted joints, both GFRP and magnesium adherends
Fig. 3 Specimen configuration
were cut from the magnesium plates and GFRP laminates and were
bolted together in a single-lap joint configuration using M6 bolts. For
the fabrication of hybrid joints, the bonded joint specimens were made
using the same process as the bonded joint. Then, they were drilled for
the M6 bolt using a drilling machine. All joint specimens were cut
using a water jet to obtain a good finishing surface.
The length and thickness of all specimens was 135 mm and 10 mm,
respectively. They had three different widths of 36 mm, 40 mm, and 44
mm. The joining length was 35 mm. The dimension and configuration Fig. 4 Joint configurations: (a) bonded, (b) bolted, (c) hybrid
of the specimens is shown in Fig. 3. Fig. 4 shows the joining region of
each single lap joint.
test machine with a capacity of 100kN (RB301 Unitech M, Korea) as
2.2 Tensile test shown in Fig. 5. The universal test machine was equipped with a high
The tensile tests were performed using an instrumented universal speed data acquisition system of load-displacement. All specimens
 INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 16, No. 6 JUNE 2015 / 1137

Fig. 5 Experimental setup for the tensile test

were elongated until fully debonded or adherend failure at a loading

rate of 13 mm/min. The automatic data acquisition system was used to
obtain the load-displacement curve. During the tests, the load-
displacement data were recorded as a function of time at intervals of
0.1 s. Three replicated tests were carried out. A microscope (SEM Tech
TC83) was installed to observe the failure behavior at the joining
regions.

3. Results and Discussions

3.1 Load-displacement curves

Fig. 6 presents load-displacement curves of each single-lap joint
with a width of 36 mm. The abscissa is crosshead displacement while
the ordinate is applied load.
The bonded joint of Fig. 6(a) behaved linearly until it reached the
maximum load. In this test, three replica tests were carried out. The
average maximum load and displacement was 8.65 kN and 1.74 mm.
After the maximum load, there was a dramatic load drop due to the
debonding of the adhesive layer or a combination of the debonding and
delamination of the GFRP adherend.
Fig. 6(b) presents load-displacement curves of the bolted and hybrid
joints under the finger tight torque condition (torque=0). In the case of
the bolted joint, the applied load increased into a logarithmic curve
shape as the displacement increased until catastrophic joint failure. The
average displacement was 6.33 mm at the average maximum load of
12.7 kN. The hybrid joints had two load peaks and initially behaved
like the bonded joints until the first maximum load of 11.0 kN at an
average displacement of 2.08 mm due to the debonding between the
GFRP and magnesium adherends. It followed the logarithmic curve
shape, like the bolted joints, until the second maximum load of 12.2 kN Fig. 6 Load-displacement curves of single-lap joints; (a) bonded, (b)
at an average displacement of 5.28 mm. The bolted joints showed the bolted and hybrid (torque=0 Nm), (c) bolted and hybrid (torque=10
largest displacement before the final failure while the bonded joint was Nm), (d) bolted and hybrid (torque=20 Nm)
the smallest. The first peak value of the hybrid joint was 58.7% higher
than the peak value of the bonded one since the load sustained the bolt
as well as the bond layer. From the load-displacement curves, the bolted Figs. 6(c)-(d) show the load-displacement curves of the bolted and
joints had the lowest stiffness since no tightening torque was applied. hybrid joints under the tightening torque of 10 Nm and 20 Nm. In the
 1138 / JUNE 2015 INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 16, No. 6

case of the bolted joint, the stiffness of the joint increased as the
tightening torque increased. The stiffness increased up to 21% and 67.5%
under a torque of 10 Nm and 20 Nm, respectively.
The stiffness of the hybrid joint increase rate was not bigger than the
bolted one. The stiffness increased to 10.2% under a torque of 20Nm.
The displacement at failure was getting short due to the increase of the
joining torques. Based on what is shown in Fig. 6, the load rebound
after the first drop reduced as the tightening torque increased. There
were few rebounds under the torque of 20 Nm.

3.2 Failure loads and failure modes

Fig. 7 shows the failure load variation of the single-lap joints along
the different tightening torques and widths.
The failure load of the bonded joints increased with an increase in
width, as shown in Fig. 7(a). The failure load increased 122.2% as the
width increased 22.2%.
In contrast, the failure load increases of the bolted and hybrid joints
were 15.5% and 17.2%, under the finger tight condition. For the hybrid
joints, as mentioned before, there were two load peaks, so the failure
load was divided into two values of bond failure and the final failure
load. Under the finger tight condition, there was no big difference in
two peak values. The failure load of the bonded joints showed the
largest standard deviation of 2362 N, while the bolted joints showed the
smallest scatter (standard deviation of 371 N).
In the tightening torque of 10 Nm condition, the failure load increase
of the bolted and hybrid joints (in this case, the bond failure load) were
16.5% and 47.9% as the width increased 22.2%. For the hybrid joints,
the bond failure load dramatically increased as the tightening torque
increased. The load rebounding that occurred in the hybrid joint (as
shown in Figs. 5(b) and (c)) was not observed or was very small after
the bond failure in the hybrid joints, which was more than the width of
36 mm.
In the tightening torque of 20 Nm condition, the increase in failure
load of the bolted and hybrid joints were 11.4% and 42.1% as the width
increased 22.2%.
Although the tightening torque increased from 0 to 20 Nm, the
failure load of the bolted joint did not show a big difference. The reason
why the failure load did not increase was an early net tension failure
of the GFRP part. However, the failure load increased from 7.8% to
30.7% in the hybrid joints. Fig. 7 Failure load for three types of joints under the different joining
The failure sequence of the bonded joints is shown in Fig. 8. Firstly, torque conditions: (a) finger tight, (b) torque=10 Nm, (c) torque=20 Nm
there was partially debonding between the GFRP and Mg adherends
and the crack propagated along the bond layer. Finally, the two adherends
were fully debonded. was sustained by the bolt and the joint was already under the larger
In the case of the bolted joints (Fig. 9), a large bending deformation deformation. Then, the GFRP adherend was fractured by the bending
occurred at the joining region as the applied load increased. Then, the of the bolted joints.
bolt was inclined due to the larger deformation and the crack initiated Fig. 11 illustrates the deformed shape of the single lap joints tested.
at the adjacent region of the bolt hole of the GFRP adherend. Finally, Initially, the applied load generates the shear stress in the joining region
the GFRP adherend was fractured by the bending. The delamination of the single lap joints, as shown in Fig. 11(a). However, the bending
and fiber failure was observed at the failure region. In this joint deformation induced by the anti-symmetric load causes the peel stress
configuration, the GFRP adherend had lower stiffness than the at the end of the joining region as the applied load increases, as shown
magnesium so the GFRP adherend was fractured in all cases. in Fig. 11(b). In the bonded joints of Fig. 11 (c), the interfacial
In the case of the hybrid joints (Fig. 10), there was partial debonding debonding occurs at both ends of the joining area where the peel stress
between the GFRP and Mg adherends and the crack propagated along has maximum value and both cracks propagated to the center of the
the bond layer like the bonded joints. After the full debonding, the joint joint along the bond layer. The joints were finally fractured as the two
 INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 16, No. 6 JUNE 2015 / 1139

Fig. 11 Deformation process of the typical single lap joints

Fig. 8 Failure sequence of the bonded joints

Fig. 12 Fracture joint specimen; (a) bonded joint, (b) bolted joint

concentrated around the area contacted by the bolt head and nut.
Finally, the net tension failure occurred at this region.
The hybrid joints behaved as a combination of the previous joints.
Fig. 9 Failure sequence of the bolted joints
Fig. 12 shows the fracture joint specimen of the bonded and bolted
single lap joints. In the bonded joints, the magnesium part was
plastically bent and there was interfacial failure between the bond layer
and magnesium adherends. The magnesium part of the bolted joints
was plastically deformed like the bonded one and the bolt hole was
extended by the bearing load. In the GFRP part, the net tension failure
occurred due to the excessive bending moment at the center of the bolt
hole. In addition, a small bearing failure was observed in front of the
bolt hole of the GFRP part.
For the hybrid joints (Fig. 13), the magnesium part was plastically
bent and a bolt hole extension was observed like the bolted joints. The
bolt hole extension was getting smaller as the width increased. In the
GFRP, the net tension failure was a main failure mode while there was
severe delamination by the bearing load compared to the bolted joints.
In the bolted joint configurations, the GFRP adherends failed before
the bolt fully supported the applied load so that there was a small-sized
bearing damage in the bolted joints. In contrast to the bolted joint
Fig. 10 Failure sequence of the hybrid joints configurations, the applied load was supported by both the adhesive
and bolt in the hybrid joint configuration so the bolt was able to support
much more load before the failure of the GFRP adherend. Therefore,
cracks met the center of the joint. a larger-sized bearing damage was observed. From the test results, the
In the bolted joints, both adherends were separated at both ends by hybrid joints showed more displacement to final failure compared with
the bending moment, as shown in Fig. 11(d). Then, the entire load was the adhesive joint. It means the delay of the final failure. In addition,
 1140 / JUNE 2015 INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 16, No. 6

Journal of Mechanical Science and Technology, Vol. 27, No. 10, pp.
2917-2921, 2013.

2. Watarai, H., “Trend of Research and Development for Magnesium

Alloys,” Science & Technology Trends, Vol. 18, No. 1, pp. 84-97,
2006.

3. Matsuzaki, R., Shibata, M., and Todoroki, A., “Improving

Performance of GFRP/Aluminum Single Lap Joints using Bolted/
Co-Cured Hybrid Method,” Composites Part A: Applied Science
and Manufacturing, Vol. 39, No. 2, pp. 154-163, 2008.

4. Kelly, G., “Load Transfer in Hybrid (Bonded/Bolted) Composite

Single-Lap Joints,” Composite Structures, Vol. 69, No. 1, pp. 35-43,
2005.

Fig. 13 Fracture joint specimen of the hybrid joint 5. Kelly, G., “Quasi-Static Strength and Fatigue Life of Hybrid
(Bonded/Bolted) Composite Single-Lap Joints,” Composite
Structures, Vol. 72, No. 1, pp. 119-129, 2006.
it showed more reliable behavior than adhesive joint because it has
6. Pakdil, M., “Failure Analysis of Composite Single Bolted-Joints
smaller deviation of the final failure load than adhesive joint.
Subjected to Bolt Pretension,” Indian Journal of Engineering and
Materials Sciences, Vol. 16, No. 2, pp. 79-85, 2009.

4. Conclusion 7. Camanho, P. P. and Matthews, F., “Stress Analysis and Strength

Prediction of Mechanically Fastened Joints in FRP: A Review,”
In this study, the joining performance of three different types of Composites Part A: Applied Science and Manufacturing, Vol. 28,
GEP224 glass/epoxy-to-AZ31B magnesium alloy single-lap joints was No. 6, pp. 529-547, 1997.
evaluated. The main findings from this study are outlined below:
8. Ban, C.-S., Lee, Y.-H., Choi, J.-H., and Kweon, J.-H., “Strength
(1) The bonded joint showed the highest failure load increase rate
Prediction of Adhesive Joints using the Modified Damage Zone
as the width increased.
Theory,” Composite Structures, Vol. 86, No. 1, pp. 96-100, 2008.
(2) The bolted joint failure load showed a decreasing trend as the
torque increased due to the early net tension failure of CFRP 9. Wang, Z. Y, Wang, L., Guo, W., Deng, H., Tong, J. W., and
parts, while the failure load increased as the torque increased in Aymerich, F., “An Investigation on Strain/Stress Distribution
case of the hybrid. Around the Overlap End of Laminated Composite Single-Lap
(3) The main failure modes of the bonded joint were the interfacial Joints,” Composite Structures, Vol. 89, No. 4, pp. 589-595, 2009.
debonding and delamination of the GFRP parts.
10. Da Silva, L. F. and Adams, R., “Techniques to Reduce the Peel
(4) In the case of the bolted joint, the main failure mode was the
Stresses in Adhesive Joints with Composites,” International Journal
bending failure of GFRP parts due to the large bending
of Adhesion and Adhesives, Vol. 27, No. 3, pp. 227-235, 2007.
deformation.
(5) The hybrid joint showed the combined failure modes of the 11. Lee, C.-J., Lee, S.-H., Lee, J.-M., Kim, B.-H., Kim, B.-M., and Ko,
previous two joints. It showed the delay of the final failure D.-C., “Design of Hole-Clinching Process for Joining CFRP and
compared with the adhesive joint. In addition, it showed more Aluminum Alloy Sheet,” Int. J. Precis. Eng. Manuf., Vol. 15, No. 6,
reliable behavior than adhesive joint. pp. 1151-1157, 2014.

12. Amancio-Filho, S., Bueno, C., Dos Santos, J., Huber, N., and Hage,
E., “On the Feasibility of Friction Spot Joining in Magnesium/Fiber-
REFERENCES Reinforced Polymer Composite Hybrid Structures,” Materials
Science and Engineering: A, Vol. 528, No. 10, pp. 3841-3848, 2011.
1. Yoon, J., Lee, Y., and Huh, H., “Investigation of Deformation and
Collapse Mechanism for Magnesium Tube in Axial Crushing test,”

View publication stats