Drop-Weight Impact Studies of GLARE 5 Fiber-Metal Laminates

A. Seyed Yaghoubi, Y.X. Liu, B.M. Liaw

Department of Mechanical Engineering, The City College of New York,
Convent Avenue and 138th Street, New York, NY 10031, USA

ABSTRACT
Impact responses and damage induced by a drop-weight instrument on GLARE 5 fiber-metal laminates with different layup
configurations and geometries were studied. The damage characteristics were evaluated using both the nondestructive
ultrasonic and mechanical sectioning techniques. Only the contour of entire damage area could be obtained using ultrasonic
C-scan whereas more details of damage were provided through the mechanical cross-sectioning technique. It was found that
failure mode changed with varying stacking sequence. GLARE 5 made of unidirectional fibers had the worst impact
resistance; followed by cross-ply and angle-ply configurations, while the quasi-isotropic lay-up showed the best resistance to
impact. Finally, influence of different geometries was considered. The results show that by introducing circular geometry,
damage patterns and impact behaviors were changed. This was especially apparent for panels with the quasi-isotropic layup
configuration.

INTRODUCTION

GLARE 5 is a fiber–metal laminate (FML) made of alternating layers of 2024-T3 aluminum alloy sheets and S2-glass/epoxy
laminates. FMLs combine the beneficial properties of metals and fiber-reinforced polymers, providing superior mechanical
properties in comparison with conventional polymer matrix composites or aluminum alloys [1-3]. Liu et al. [4] investigated
low velocity impact damage on ARALL 3 and various GLARE grades, i.e., GLARE 1, 2 and 3. They found that GLARE 1
with glass-epoxy prepregs possessed higher impact tolerance than ARALL 3 with aramid-epoxy prepregs. They also
concluded that GLARE 3 with [0°/90°] cross-ply glass-epoxy prepregs offered better impact resistance than GLARE 2,
which is made of [Ͳιଶ ] unidirectional glass-epoxy prepregs. Hitchen et al. [5] considered the effect of stacking sequence on
impact damage in a carbon fiber-epoxy composite. They concluded that the stacking sequence influenced both the pre- and
post-compression strengths and affected the impact damage in 16-ply carbon fiber toughened epoxy laminates. They also
reported that the total delamination area was a function of the stacking sequence. Choi et al. [6] concluded that ply orientation
and stacking sequence could significantly affect impact damage. They also found that impact damage was more sensitive to
the change of stacking sequence than of thickness. The influence of lay-up configuration on graphite-epoxy composite
material was studied by Guynn et al. [7]. They concluded that the difference in compression failure strain due to stacking
sequence were small. Cantwell [8] studied geometrical effects in the low velocity impact response of GFRP. He drew a
conclusion that tests on a range of circular and square plates had shown that the degree of damage within the structure was
related to the force generated during the impact event.

In this study, the effect of stacking sequence and geometry were considered to evaluate the low velocity impact behavior of
the GLARE 5 composite materials through drop weight tests. Optical imaging, ultrasonic C-scan and mechanical sectioning
techniques were used to assess the impact damages.

EXPERIMENTAL PROCEDURES

All GLARE panels considered in this study consist of 2024-T3 aluminum alloy with a thickness of 0.305mm (0.012”) per
layer and S2-glass/epoxy laminated layers, each with a thickness of 0.508 mm (0.020”). Each S2-glass/epoxy layer has a
layup orientation according to Table 1. Panels were cut into two different geometrical shapes: square and circular, with

T. Proulx (ed.), Experimental and Applied Mechanics, Volume 6, Conference Proceedings of the Society for Experimental 267
Mechanics Series 9999, DOI 10.1007/978-1-4614-0222-0_33, © The Society for Experimental Mechanics, Inc. 2011
268

dimensions of 101.6 mm x 101.6 mm (4” x 4”) for the square specimens and 101.6 mm (4”) in diameter for the circular
specimens. The typical mechanical properties of constituents are described in Table 2 [9-11]. In Table 1, the term MVF
represents metal volume fraction and is defined as the ratio of the sum of the thicknesses of all aluminum layers over the total
thickness of the fiber-metal laminate [12]:

¦ taluminum
MVF
tFML
(1)

Table 1. GLARE 5 panels tested in this study with different stacking sequences.

Prepreg Plies Total

Lay-up Configuration MVF
& Orientation Thickness
(m/n)
ª¬0q4 º¼
unidirectional
ª¬0o / 90o / 90o / 0o ¼º
1.930mm
3/2 cross-ply 0.474
(0.076")
ª¬45 /  45 /  45 / 45 º¼
o o o o

angle-ply
ª¬0 / 45o /  45o / 90o º¼
o

quasi-isotropic
The configuration notation m/n means the panel is composed of m
aluminum-alloy layers interlaced with n fiber-reinforced epoxy layers.

All impact tests were conducted using an Instron Dynatup 8250 pneumatic-assisted, instrumented drop-weight impact tester
as shown in Fig. 1. The specimens were clamped circumferentially along a diameter of 76.2 mm (3”) in the specimen fixture
and impacted by a hemispherical steel impactor of diameter 16 mm with a mass of 6.14 kg. A pair of pneumatic breaks was
also used to ensure no multiple strikes during impact test. The contact force history of impact was measured using a load cell
located right above the impactor nose, and the impactor velocity was recorded by a pair of photoelectric-diodes attached to
the base of the test machine. The time histories of the impactor position and the absorbed energy were obtained numerically
by applying Newton’s law and energy balance principle.

After each test, the specimen was carefully removed from the fixture for post-impact damage assessments. Figure 2 illustrates
how the post-impact permanent central deflection was measured. Both nondestructive and destructive evaluation techniques
were used to assess the impact damage inside the impacted panels. The impacted specimens were first scanned through an
UltraPAC immersion ultrasound system for nondestructive damage evaluation, then carefully sectioned into two halves by a
diamond saw through the impact center. Finally cross-sectional optical pictures were taken in order to reveal the more
detailed damage inside the specimens. In this study a pair of 5 MHz ultrasonic transducers (one focused and another flat) in
through-transmission mode was adopted to access the damage in the impacted specimens.
 269

Table 2. Typical values of mechanical properties.

Aluminum UD S2
Mechanical property Glass/FM 94
2024-T3 Epoxy Prepreg

L 455 1900
Tensile ultimate strength (MPa)
T 448 57
L 359 –
Tensile yield strength (MPa)
T 324 –
L 72 54
Tensile modulus (Gpa)
T – 9.4
L 19 3.5
Ultimate strain (%)
T – 0.6
L 303 –
Compressive yield strength (MPa)
T 345 –
L 74 54.4
Compressive modulus (GPa)
T – 11
3
Density (g/cm ) 2.79 2.00
The symbols, L and T, stand for longitudinal (the rolling direction for the metal) and transverse
directions, respectively.
270

(a) (b)
Figure 1. (a) Drop Weight impact tester, (b) The specimen fixture.

Figure 2. Measurement of the permanent central deflection.

RESULTS AND DISCUSSION

• SQUARE SPECIMENS

Stacking sequence effect was studied on GLARE 5 (3/2) specimens with different stacking sequences, namely:
unidirectional, angle-ply, cross-ply and quasi-isotropic. Several drop-weight tests were conducted on GLARE 5 (3/2) with the
cross-ply layup orientation, as a benchmark, in order to choose a proper impact energy to study and compare the impact
responses and damage patterns of the GLARE 5 (3/2) of various stacking sequences. Based on the test results, 40J impact
energy was chosen. Figure 3 shows the histories of absorbed energy, central displacement, contact force and contact stiffness
for the square GLARE 5 (3/2) specimens with cross-plyሾͲ°ȀͻͲ°ሿୱ , unidirectionalሾͲ°ସ ሿ, angle-plyሾ൅Ͷͷ°ȀെͶͷ°ሿୱ and quasi-
isotropicሾͲ°Ȁ൅Ͷͷ°ȀെͶͷ°ȀͻͲ°ሿ layup sequences. The corresponding C-scan results, back-side (non-impacted) views, as well
as the cross-sectional views are shown in Figs. 4 and 5.
 271

Figure 3. Impact responses of the cross-ply, unidirectional, angle-ply and quasi-isotropic GLARE 5 (3/2) square
specimens under 40J impact energy.

The absorbed energy history revealed that under the same impact condition, unidirectional panel released less energy whereas
quasi-isotropic panel released the most among the other type of panels. The peak value of central displacement history was
the highest for the unidirectional panels while the lowest was for the quasi-isotropic panels. For the GLARE 5 (3/2) cross-ply
and angle-ply, this peak was the same. Hence, the quasi-isotropic panels offered more resistance to the indentation than any
other type of panels. The unidirectional panels were the worst panel to sustain the impact. By comparing the contact force
histories, the unidirectional panels had the lowest peak contact force and the longest contact time duration. The quasi-
isotropic panels offered the highest peak contact force among the other panel types. It is worth noting that although the
patterns of contact force history for the GLARE 5 (3/2) panels with various layup configurations were similar, the damage
patterns were different. From Fig. 3(d), it could be observed that contact stiffness curves started to deviate at time 1.16 ms,
corresponding to 4.1 mm of the central deflection.

Mechanical-sectioning technique revealed that the major types of damage for the GLARE 5 (3/2) cross-ply panels were
through-specimen crack, fiber breakage/splitting and delaminations (Figs. 4 and 5). The major delamination located between
the non-impacted side aluminum layer and the adjacent glass-epoxy layer. Some small regional delaminations around the
middle aluminum layer could also be found. A diamond-shaped damage zone was induced with major diagonal along 90q.
For the unidirectional panels, long visible through-the-thickness cracks occurred, in which the cracks extended near the
clamped edge. There was not much delamination besides fracture of aluminum layers and the splitting of the glass-epoxy
272

layers for the unidirectional panels. Since there was no bending stiffness mismatch in a unidirectional prepreg layer, no
delamination was expected in the interface of the unidirectional prepreg layer. The fracture of aluminum in the non-impacted
side resulted from bending while the stress concentration induced the crack in the impacted face. The damage zone observed
from C-scan looked like a lip-shaped with the major axis aligned along the fiber direction. The damage pattern for the angle-
ply panels was very close to the cross-ply panels except the damage contour for the angle-ply panels was smaller than the
cross-ply panels (Fig. 4). Furthermore, from the C-scan results, the major diagonal, i.e. perpendicular to the fiber direction,
was noticeably shorter for the angle-ply panels compared to the cross-ply panels. This means that the GLARE 5 (3/2) panels
with the angle-ply stacking sequence resisted more to impact compared to the cross-ply lay-up with the same panel
configuration. The reason for this is believed to be due to the fact that the angle-ply panels offered more shear stiffness than
the cross-ply panels. For the quasi-isotropic panels, major damage were crack in the non-impacted side aluminum, damage in
the bottom ሾͲ°Ȁ൅Ͷͷ°ȀെͶͷ°ȀͻͲ°ሿ prepreg layer as well as significant delamination between the í45° and 90° layers. An
approximately circular-shaped damage zone could be detected through ultrasonic C-scan for the quasi-isotropic panels. It is
worth noting that under 40J impact energy, a visible crack appeared in all specimens on the non-impacted side. The crack
was straight for the cross-ply, unidirectional and angle-ply GLARE 5 (3/2) specimens whereas it was not straight for the
quasi-isotropic specimens.

Based on the discussion above, GLARE 5 (3/2) made of the unidirectional layup offered the worst impact resistance,
followed by cross-ply and angle-ply layups, while quasi isotropic layup showed the best resistance to the impact. The reason
for this is that the specimen stiffness decreased as layup configuration changed from quasi-isotropic layup to unidirectional.
In other words, propagation of damage was alleviated and delayed as the bending stiffness mismatch decreased.
 273

Figure 4. Back-side (non-impacted) view and corresponding C-scan view of the GLARE cross-ply, unidirectional,
angle-ply and quasi-isotropic GLARE 5 (3/2) square specimens under 40J impact energy.
274

Figure 5. Cross-sectional view of the cross-ply, unidirectional, angle-ply and quasi-isotropic GLARE 5 (3/2)
square specimens under 40J impact energy. The symbol means that the fiber direction is perpendicular to the
cross-sectional view picture.

• CIRCULAR VERSUS SQUARE SPECIMENS

Figure 6 shows histories of absorbed energy, central displacement, contact force, contact stiffness for the circular GLARE 5
(3/2) specimens with the cross-plyሾͲ°ȀͻͲ°ሿୱ , unidirectionalሾͲ°ସ ሿ, angle-plyሾ൅Ͷͷ°ȀെͶͷ°ሿୱ and quasi-isotropicሾͲ°Ȁ൅Ͷͷ°Ȁ
െͶͷ°ȀͻͲ°ሿ layup sequences. Figures 7 and 8 depict corresponding C-scan results, back-side (non-impacted) views, as well
as cross-sectional views of specimens of those configurations. Tables 3 summarizes experimental results of the impact-
induced crack lengths measured on the impacted side (the top aluminum sheet) and the non-impacted side (the bottom
aluminum sheet), respectively, along with the post-impact permanent central deflection in GLARE 5 (3/2) composite
materials with different stacking sequences and geometries subjected to 40J impact energy. Comparing these results to those
of the square specimens, the following remarks can be found.

The energy history curves were in a very close pattern to those of the square specimens. The central deflection curves for the
circular unidirectional and quasi-isotropic specimens increased noticeably in comparison with the corresponding panels of
 275

square specimens. This enhancement was very negligible for the cross-ply and angle-ply panels. The peak force values for
the circular specimens were lower than those of the corresponding square specimens. The contact time was also raised by
changing the geometry from square to circular. This difference was quite noticeable for the unidirectional and quasi-isotropic
panels but it was relatively negligible for the cross-ply and angle-ply panels. The contact stiffness decreased by changing the
geometry from square to circular. The effect of the above mentioned differences could be revealed by comparing the cross
sectional, C-scan and back side views of the specimens with the ones for the square specimens. Comparing the cross-
sectional views of the two different geometries for the cross-ply and angle-ply specimens, the overall damage patterns were
similar except in circular geometry there was no debonding between the non-impacted aluminum and the adjacent prepreg
layer. For the circular unidirectional specimen the induced damage pattern was different from the corresponding square
specimen. Prepreg damages among the aluminum layers were noticeably increased. Also unlike the unidirectional square
specimens, in which the prepreg close to the non-impacted side were broken into several pieces by through-the-thickness
cracks, it did not happen for the circular specimen. The major differences in damage pattern were obvious for the circular
quasi-isotropic specimen. By evaluating the cross-sectional views for the quasi-isotropic square and circular specimens, a
dramatic change could be seen in the induced damages. Unlike the square specimen, which only the bottom aluminum
damaged, all the aluminum layers failed for the quasi-isotropic circular specimen. It is worth noting that the impacted-side
aluminum failed in two places. Another important difference was that the delamination was negligible for the circular
specimens whereas it was relatively notable for the square specimens.

Fig.6. Impact responses of the cross-ply, unidirectional, angle-ply and quasi- isotropic GLARE 5 (3/2) circular
specimens under 40J impact energy.
276

Fig.7. Back side (non-impacted) view and corresponding C-scan view of the cross-ply, unidirectional, angle-ply
and quasi-isotropic GLARE 5 (3/2) circular specimens under 40J impact energy.

The C-scan results revealed that the damage contour decreased for the cross-ply and angle-ply circular specimens. For the
unidirectional panels of the two geometries, the C-scan shapes were almost similar to each other with the exception that for
circular geometry, the length of the damage along the 0° fiber direction was longer (Table 3, Figs. 4 and 7). In fact, the
induced crack in the unidirectional circular specimens passed the grip area of the specimen. For the quasi-isotropic
specimens, the crack at the non-impacted side of the circular specimen was noticeably longer than the square specimen. By
transition from square to circular geometry, the damage contour changed from almost a circular shape to an elliptical shape
with major axis along the 0° fiber direction. Due to change of the geometry, the central deflections were higher for the
circular specimens compared to those of square specimens.
 277

Fig.8. Cross-sectional view of the cross-ply, unidirectional, angle-ply and quasi-isotropic GLARE 5 (3/2) circular
specimens under 40J impact energy. The symbol means that the fiber direction is perpendicular to the cross
sectional view picture.
278

Table 3. Crack lengths and permanent deflections of the cross-ply, unidirectional, angle-ply and quasi-isotropic
GLARE 5 (3/2) under 40J impact energy.

Impact Crack length in outer layer (mm) Permanent

Stacking Sequence &
Energy Deflection (mm)
Geometry Impacted side None impact side
(Joule)
Cross-ply 40 9 18 6.75
Unidirectional 40 69 75 7.65
Angle-ply 40 7 16 6.65
Quasi-isotropic 40 0 12 5.30
Cross-ply 40 8.5 16 6.8
Unidirectional 40 83 83 7.7
Angle-ply 40 8.5 16 6.8
Quasi-isotropic 40 7.5 20 5.80

CONCLUSIONS

This study presents an experimental investigation on the impact response of GLARE 5 (3/2) composite materials considering
the effects of stacking sequence and geometry through using drop weight impact tester. The following remarks can be
concluded from this study.

x GLARE 5 made of unidirectional fibers had the worst impact resistance; followed by cross-ply and angle-ply
configurations, while the quasi-isotropic lay-up showed the best resistance to impact.
x By introducing circular geometry, damage patterns and impact behaviors were changed. This was very obvious for the
panels with quasi-isotropic layup configuration, i.e. ሾͲ°Ȁ൅Ͷͷ°ȀെͶͷ°ȀͻͲ°ሿ.

x Only the profile of damage zone could be detected through the ultrasonic C-scan. The mechanical- sectioning technique
must be adopted in order to get the details of damage inside the fiber-metal laminates. The drop-weight induced damage
included indentation around impact center, delamination between aluminum and glass-epoxy composite, cracks in
aluminum layers, and damage in composite layers. More severe damage occurred on the non-impacted side of fiber-
metal laminates.

ACKNOWLEDGMENTS

This study was sponsored by NASA Faculty Award for Research (FAR) under Grant No. NAG3-2259 and by PSC-CUNY
under Grants 61368-00 39 and 62209-00 40. Dr. Kenneth J. Bowles and Dr. John P. Gyekenyesi were the Technical Monitors
of the NASA grant. Part of the equipment used in this investigation was acquired through Army Research Office Grant No.
DAAD19-99-1-0366.

REFRENCES

1. Vlot, A., Glare-History of the development of a New Aircraft Material. Kluwer, Dordrecht: Kluwer Academic
Publishers, 2001.
2. Vermeeren C.A.J.R., Beumler T., De Kanter, J.L.C.G., Van Der Jagt, O.C., Out, B.C.L., Glare design aspects and
philosophies. Appl. Compos. Mater., 10, 257–276, 2003.
 279

3. Lawcock G., Ye L., Mai Y.W., Novel fiber reinforced metal laminates for aerospace applications — a review: Part I —
background and general mechanical properties. SAMPE Journal, 31, 23–31, 1995.
4. Liu Y., Liaw B., Effect of constituents and lay-up configuration on drop-weight tests of fiber-metal laminates. Appl.
Compos. Mater., 17 (1), 43-62, 2010.
5. Hitchen S.A., Kemp R.M.J., The effect of stacking sequence on impact damage in a carbon fiber/epoxy composite.
Composites, 26, 207-214, 1995.
6. Choi H.Y., Wang H.S., Chang F.-K., Effect of laminate configuration and impactor’s mass on the initial impact damage
of graphite/epoxy composite plates due to line-loading impact. Compos. Mater., 26 (6), 804-827, 1992.
7. Guynn E.G., O’Brien T.K., The influence of lay-up and thickness on composite impact damage and compression
strength. AIAA-85-0646, Proceedings, 26th AIAA/ASME/ASCE/AHS Structures, Structural Dynamics, and Materials
Conference. Orlando, FL., 1, 187-196, April 1985.
8. Cantwell W.J., Geometrical effects in the low velocity impact response of GFRP. Compos. Sci. and Technol., 67, 1900-
1908, 2007.
9. Data sheets. Aviation Equipment Structures, Inc., Costa Mesa, California, 1998.
10. QA Reports B0319B-2, B1008B-1, B0904A-3. Structural Laminates Company, New Kensington, Pennsylvania, 1994.
11. Alloy 7475 Plate and Sheet, ACRP-053-B. Alcoa Mill Products, Bettendorf, Iowa.
12. Vlot A., Gunnink J.W. (Eds.). Fiber Metal Laminates, an Introduction. Dordrecht, the Netherlands: Kluwer Academic
Publishers, 2001.