Advanced Composites and Hybrid Materials (2019) 2:339–350

https://doi.org/10.1007/s42114-019-00096-6

ORIGINAL RESEARCH

Adhesiveless honeycomb sandwich structures of prepreg carbon

fiber composites for primary structural applications
Md. Nizam Uddin 1 & Helene T. N. Gandy 1 & Muhammad M Rahman 1 & Ramazan Asmatulu 1

Received: 6 October 2018 / Revised: 11 March 2019 / Accepted: 27 March 2019 / Published online: 2 May 2019
# Springer Nature Switzerland AG 2019

Abstract
This study presents a comparative study of adhesiveless honeycomb sandwich structures of carbon fiber prepreg composites
while assessing the structure with an adhesive film at the skin-to-core interface. Laminate and honeycomb sandwich panels were
fabricated and tested with consistent layup, curing, and testing processes. Adhesive sandwich panels were produced using the
CYCOM 977-2 prepreg system, while adhesiveless sandwich panels were constructed with the MTM45-1 prepreg system.
Laminate panels were also fabricated using the two different prepregs. Specimens from the panels were tested for physical
and mechanical properties as well as moisture-absorption performance. Test results obtained from non-destructive testing and
experimental work confirmed that physical properties of the self-adhesive prepreg mainly met the component and void-content
recommendations for use in primary structures. Mechanical tests were performed at room (24 °C), hot (82 °C), and cold (− 55 °C)
temperatures, as well as under dry (low moisture) and wet (85 ± 2% humidity) conditions. In all environments tested, the
adhesiveless sandwich panel with MTM45-1 exhibited a 14–16% higher tensile strength than the sandwich panel with 977-2
prepreg and AF191 adhesive film, testifying to the stronger bond at the skin-to-core interface. The short-beam shear strengths of
the dry laminates above 58 MPa were also suitable indicators of polymeric composite materials for primary structural purposes in
aircraft. The average short-beam shear strength of the 977-2 laminate was 28% higher than that of the MTM45-1 laminate, hence
displaying a tougher and stronger epoxy resin. These experiments suggest that an aramid honeycomb sandwich structure with a
self-adhesive carbon prepreg system compared to a similar structure using additional adhesive demonstrates the ability of these
configurations for use as primary structures in aircraft. This study also confirms using adhesiveless composite panels for different
industrial applications (e.g., wind turbine, automotive, ship, defense, and space).

Keywords Prepreg carbon fiber composites . Honeycomb structure . Laminate composite . Adhesiveless bonding . Mechanical
properties

1 Introduction functionality. The honeycomb sandwich structure offers high

strength-to-weight and stiffness-to-weight ratios. Skins in the
The development of materials and structures throughout the sandwich structure carry the bending stress, like the flanges of
centuries, from the I-beam to the composite sandwich panel, the beam, while the composite core resists the shear load, like
has shown an improvement in weight reduction and the web. Honeycomb sandwich structures are extensively
used in the aerospace and automotive industries for primary
structures [1–3]. For an aircraft, the primary structures are
those considered critical to flight safety and whose failure
* Ramazan Asmatulu
[email protected] can be catastrophic, such as wings, empennages, control sur-
faces, and their systems, as well as fuselages, engine mount-
Md. Nizam Uddin
[email protected] ings, and landing gear [4]. Honeycomb sandwich panels with
carbon graphite face sheets are commonly used for such sys-
Muhammad M Rahman
tems and components. Because these structures are subjected
[email protected]
to shear forces, the load is transferred from the skin to the core.
1
Department of Mechanical Engineering, Wichita State University, The interface is also subject to shear loads during that transfer,
1845 Fairmount St., Wichita, KS 67260-0133, USA causing a risk of delamination in the structure. Therefore,
340 Adv Compos Hybrid Mater (2019) 2:339–350

these sandwich structures have commonly been fabricated chemical bonds in the matrix and cause swelling. These
with additional adhesive between the skins and the core in changes lead to the loss of dimensional stability, reduction in
order to maintain continuous support of the skins for amplified ultimate strength and stiffness, and an increase in the residual
consistent stiffness. Elimination of the adhesive film results in stress of the laminate. As determined by Chalkley and
multiple benefits. The proven technical performance of an Chester, the glass transition temperature (Tg) of the epoxy
adhesiveless honeycomb sandwich structure has shown that resin is also reduced by moisture absorption. They quantified
the use of self-adhesive prepregs would result in tremendous that about 1 wt% gain by the composite in moisture absorption
weight savings by eliminating the added adhesive film on the level could cause the resin Tg to drop from 175 to 130 °C [15].
composite materials. Adhesive film alone can represent up to Shafizadeh and Seferis focused their work on bare honey-
10 % of the weight of a sandwich panel based on the skin comb, studying the effect of long-term exposure of bare hon-
closeout and surface area of the core [5]. Other benefits of eycomb cores in water [16]. After 38 weeks of soaking in
using adhesiveless honeycomb sandwich structures would water at 60 °C, aramid cores absorbed the greatest water, the
be time, labor, and cost savings in the layup process, material, highest being Nomex, with 7% water retention without even
and inventory. reaching saturation. A slight decrease in the flexural strength
With continuous improvement in resin toughness and resin of the core, which highly depends on the web strength in the
adhesive capabilities, the new generation of self-adhesive pre- machine direction of the core, was observed. Crocket et al.
pregs used for face sheets has allowed the fabrication of reported that freeze-thaw cycles were not always damaging to
adhesiveless honeycomb sandwich structures. A couple of honeycomb sandwich structures and that the damage mainly
different techniques are used to bond together a sandwich depended on the materials selected for the structure [17].
structure [6]. Throughout the years, improvements in honey- This study focused on a comparative study of adhesiveless
comb sandwich structures have been a balance in engineering honeycomb sandwich structures of carbon fiber prepreg compos-
design and manufacturing decision-making. Developments ites fabricated with adhesive film at the skin-to-core interface.
have come with tradeoffs in cost and sometimes weight. The The laminates and honeycomb sandwich panels were construct-
porosity sometimes seen in the skin and at the core interface ed, and physical and mechanical properties were tested. The
has been corrected by adding layers of thin or thicker adhesive effect of moisture absorption on the mechanical properties of
films in the structure, which does add additional weight to the the laminate and sandwich panels was also studied. The novelty
structures [7]. The development of self-adhesive prepreg is of the present study includes the following: (1) for the first time,
another beneficial achievement that has focused on modifying the adhesiveless method was applied on honeycomb-structured
the formula of the epoxy resin typically used in the prepreg composites that are used primarily for aircraft manufacturing and
with minimal weight addition. The prepreg allows for precise (2) the produced panels have some distinct advantages relative to
control of the amount of resin in the face sheet while keeping weight reduction and cost. The fundamental knowledge gained
an elevated fiber-to-resin ratio, hence superior quality and here can be used to enhance the development of lighter compos-
repeatability in production, which is desired in structural parts ites for aircraft and other industries.
[8–10]. Shafizadeh et al. reported that the self-adhesive pre-
pregs used for laminate skins offered adhesion to the core
without needing additional adhesive film, but these prepregs 2 Experiment
exhibited porosity issues, with a void content of 3 to 5% in the
laminate [11]. Porosity in the prepreg is highly dependent on 2.1 Materials
the impregnation method. Hayes et al. studied the effect of the
prepreg manufacturing process on the level of porosity and the The MTM45-1 prepreg system from the Advanced
overall quality of the laminate after cure by comparing Composites Group (ACG) was used for the adhesiveless sand-
modeled self-adhesive prepregs using the hot-melt method, wich panel. For comparison, two MTM45-1 prepreg systems
the solvent-assisted method, and commercial solvent- with different resin content were employed: one with 36%
produced BMS 8–79 prepreg [12]. resin weight contained equally on either side of the prepreg,
In an investigation of the lifecycle of the honeycomb sand- referred to as MTM45-1 36% RW (resin weight), or MTM45-
wich structure on in-service aircraft, three main sources and 1 normal, and the other with 50% resin weight, of which 20%
their combinations were found to be the cause of water ingres- was on one side of the prepreg and 80% was on the other side,
sion and migration in the core [13]. Crossman et al. explained referred to as MTM45-1 50% RW, or self-adhesive prepreg.
that the susceptibility of epoxy resin to moisture absorption is The prepreg and adhesive films used for the sandwich panel
caused by its hydrophilic polar nature, which allows its sur- with adhesive were the CYCOM 977-2 prepreg system from
face to eventually absorb water [14]. Changes to the carbon Cytec, referred to as 977-2 (resin weight 42 ± 3%), and adhe-
graphite/epoxy laminate depend on how much moisture has sive films EA9686 and AF 163 EA9686 from the 3M
been absorbed. The moisture and thermal increase break Company. Three aircraft using the 977-2 toughened resin
Adv Compos Hybrid Mater (2019) 2:339–350 341

system are the Hawker 4000, Airbus A380, and Boeing 787 The laminate fabrication process using the MTM45-1
[18, 19]. The core used in all cases was the Nomex aramid 36% RW prepreg system followed the same process as
honeycomb from Hexcel. Cell size, thickness, and density of that for the 977-2 resin system (Fig. 1). Sandwich panels
the core material were 0.32 cm, 1.27 cm, and 0.0481 g/cc, were fabricated using two different prepreg systems: with
respectively. Both the MTM45-1 and 977-2 prepreg systems and without adhesive film. The adhesive and honeycomb
were impregnated using the hot-melt method, and reinforce- prepreg sheets were cut and then, using the 977-2 system,
ment was the standard modulus AS4 from Hexcel. laid in four layers, one after the other at 0° to the ribbon
direction of the prepreg. Then one layer of the AF191
adhesive and honeycomb core was added. The same pro-
2.2 Fabrication of laminate and sandwich panels cess was repeated to add another four layers of prepreg
and one layer of adhesive. Then release film (non-
In this study, two laminate panels were constructed: one with perforated) was used on the top and bottom of the sand-
the 977-2 resin system and the other with the MTM45-1 36% wich panel, and potting tape was placed all around the
RW prepreg. Both had 12 layers of prepreg and followed the tool. Last, the breather material, vacuum hook-up, and
exact same layup process to avoid variability. For all prepregs vacuum bag were added. The sandwich panel with the
and adhesive, the materials were completely thawed at room MTM45-1 prepreg system was similar, but no adhesive
temperature to avoid condensation. The layup was performed layer was used in this panel because the final layer of
in a clean room environment. After cutting the prepreg, the ply has a little more resin (self-adhesive). The laminate
plies were laid at a 0° angle to the ribbon direction of the and sandwich panels were cured in an autoclave accord-
prepreg. Then release film (non-perforated) was used on the ing to the manufacturer’s recommendation. During curing
top and bottom of the laminate. Following this step, all four of the laminate and sandwich panels with the MTM45-1
sides of the plate to the release film were taped, while allowing system, the vacuum was run throughout the autoclave,
air to flow through the corners. An adhesive cork dam was whereas with the 977-2 system, the vacuum was vented
also installed around all edges of the panel. In addition, breath- out. The same cure cycle was used for both the MTM45-1
er material, a vacuum hook-up, and a vacuum bag were ap- and the 977-2 laminate panel. Likewise, the cycle used for
plied on the prepreg laminates, effectively sealing the bag the sandwich panel with the self-adhesive prepreg was the
against potting around the tool. same as the sandwich panel with adhesive film.

Fig. 1 Step-by-step fabrication of laminate panel (977-2 resin system; top and middle rows) and sandwich panel (977-2 resin system; bottom row)
342 Adv Compos Hybrid Mater (2019) 2:339–350

2.3 Testing of physical properties 2.4 Testing of mechanical properties

2.3.1 Ultrasonic and constituent content test 2.4.1 Flatwise tensile testing of sandwich panels

The ASTM E2580 reference standard—Ultrasonic Testing The flatwise tensile test was performed to measure the
of Flat Panel Composites and Sandwich Core Materials strength of adhesion between the face sheets and the core
Used in Aerospace Applications—was selected to create material. Specimen preparation and testing were accom-
test instructions for the ultrasonic testing, which was per- plished according to ASTM C297/C297M. Because the phys-
formed on all four panels. High-frequency sound beams ical properties of the laminates and core in a sandwich struc-
were used to locate any discontinuity within the fiber/ ture can change based on temperature, this test was performed
resin matrix composite material. The flatbed water-jet at some extreme temperatures to which an aircraft can be
pulse-echo was used for the non-destructive testing. The exposed in real environments: high temperature (82 °C), room
material was set on the flatbed with a water jet running temperature (24 °C), and low temperature (− 55 °C). The
from the probes: one on top and the other on the bottom sandwich specimen was adhered to the aluminum block fix-
of the panel. The output result was a CAD data map of tures using EA9686 adhesive film for high temperature, and
each panel being tested with decibel (dB) readings. Then AF 163 EA9686 adhesive film for room and cold tempera-
the specimens were cut from each panel using a diamond tures. These adhesives were selected based on their properties.
grit table saw and diamond grit wafering machine for Tests were performed using the horizontal united pull tester
microscopic visual inspection. First, all specimens were for specimens in the temperature-controlled chamber and the
cleaned with acetone to remove all markings and dust vertical united pull tester for specimens at room temperature.
from the cutting. A KEYENCE microscope (up to ×200 The ultimate flatwise tensile strength Fz is
magnification) with a camera was used to inspect the
Pmax
specimens for any visual defects. On the sandwich panels, Fz ¼ ð1Þ
special emphasis was placed on the microscopic picture of A
the skin-to-core interface to observe the fillet formed in where Pmax is the ultimate force for failure and A is the cross-
the panels with and without adhesive film. To determine sectional area of the specimen.
the constituent content of the laminates, the ASTM D3171
standard was chosen. In the sandwich specimens, the top
2.4.2 Short-beam shear testing of laminate panels
and bottom (tool side) face sheets were cut off the hon-
eycomb, and faces of the core were smoothed with sand-
The interlaminar shear strength of the laminate was measured
paper to remove the core and fillets on those laminates. In
using the short-beam shear strength test. The reference stan-
the constituent content test, the resin digestion method
dard ASTM D2344 was used to create the test instructions.
was used to dissolve the matrix, keeping only the laminate
Common failure modes visually observed are interlaminar
fibers. The fiber volume, matrix volume, and void content
shear, flexure, and inelastic deformation. This test was per-
percent were then calculated with prior knowledge of res-
formed at room temperature on the MTS 880 material testing
in and fiber densities from the manufacturers.
system, an analysis technique using a hydraulic wedge grip
(referred to as the tool). The short-beam shear strength Fs is
2.3.2 Glass transition temperature Pmax
F s ¼ 0:75* ð2Þ
b*t
The glass transition temperature of the matrix is impor-
tant in the cure cycle, cure characteristics, and mechan- where Pmax is the maximum load observed during the test and
ical properties of the composite. It also provides the b and t are the width and thickness, respectively.
operational limits of the material and assurance that
the service temperature of the end use equipment will 2.5 Moisture absorption
be met. MIL-HDBK-17-1F specifies three testing
methods to determine the T g of the resin [20]. To evaluate the effect of environmental exposure on the phys-
Differential scanning calorimetry (DSC, N2 atmosphere) ical and mechanical properties of the laminate and sandwich
was chosen for this test because the specimens were not specimens, the gravimetric test method was used according to
pure resin but rather prepregs with fibers. Using DSC reference standard ASTM D5229/D5229M. The humidity
instead of the other two mechanical tests ensured that chamber was set at 85 ± 2% humidity and 71 ± 0.34 °C tem-
the fibers had no effect on the results of the resin Tg. perature. The weights of the specimens were measured over
Testing was conducted according to ASTM D3418. time to obtain the percent moisture mass gain. Laminate
Adv Compos Hybrid Mater (2019) 2:339–350 343

specimens were placed in the humidity chamber with the intent Fig. 2. The area in blue represents the actual panel, indicating
of reaching full saturation, followed by the short-beam shear that about 80 dB were transmitted from the top to the bottom of
strength test at room temperature. On the other hand, the sand- the panel as the water jet ran through it. The area where the two
wich specimens from both panels were edge-filled with circular foam stickers were placed for reference orientation is
Flexane 15350, a brushable urethane, with the goal of partially clearly visible in the scan because the sound transmitted was
sealing the edges. This was done to allow a random amount of reduced to 20–40 dB. The purple area represents plain air with
moisture absorption in the honeycomb as well as the face full sound transmittal because the top and bottom water jets are
sheets. The intent was to orchestrate a random range of the touching. The ultrasonic inspection did not reveal porosity on
water-absorbed specimens for flatwise tensile testing at room this panel. The sandwich panel inspection results were also
temperature. The average moisture content is determined as comparable, with a sound transmission of about 50 dB
(Fig. 3). This lower number is not an indication of porosity
W c −W i
M % ¼ 100* ð3Þ in the skin or interface, but rather the increased thickness with
Wi respect to the laminate panel and the presence of the core.
where Wcand Wiare the present and initial specimen mass, re- The purpose of ultrasonic and microscopic inspection was
spectively. After moisture absorption, sandwich panels were pre- to ensure the uniformity of sound transmission within each
pared for flatwise tensile testing. In this test, the two-part room- sandwich panel as a sign of void absence. An inspection of
temperature adhesive EA9359.3 was used to adhere the speci- the sandwich panel is not as accurate as that of the laminate:
mens to the aluminum block fixtures. Short-beam shear testing of standard deviations here were 17.13 dB with a mean of
the moisture-absorbed laminate panels was also performed. 65.5 dB for the adhesiveless panel, and 18.46 dB with a mean
of 63.7 dB for the adhesive-film panel. Except for the foam
locations and the edges, as expected, the ultrasound level was
consistent throughout each sandwich panel. The total area
3 Results and discussion scanned included the free air perimeter around the panel and
the jets on the reference circular foam. Nevertheless, the stan-
3.1 Physical properties dard deviation (3.84 dB for MTM45-1 and 3.24 dB for 977-2)
is still a good indication of the proximity of the data points to
3.1.1 Ultrasonic and microscopic inspection the mean (84 for both panels). As indicated by non-destructive
testing, all four panels were well fabricated and, therefore,
The amplitude C-scan of the ultrasonic inspection of the lam- good candidates for test specimens. Figure 4 shows the mi-
inate panel from two different prepreg systems is shown in croscopic images of both MTM45-1 and 977-2 laminate

Fig. 2 Amplitude C-scan display of laminate panels. a MTM45-1. b 977-2

344 Adv Compos Hybrid Mater (2019) 2:339–350

Fig. 3 Amplitude C-scan display of sandwich panels. a Adhesiveless with MTM45-1 prepreg. b Adhesive film with 977-2 prepreg

panels. One observation made in the microscopic inspection resin that connects the fillet on either side of the cell. This
was in the width of the tow used for both prepregs measured provides additional adhesive properties at the interface. In
on the laminate panels. The MTM45-1 panel was woven with the same area of this figure, an air pocket can be seen in the
tows of the same width, about 0.15 cm, whereas the 977-2 fillet, which could be a sign of volatiles that did not escape to
panel appeared to be using tows of two different thicknesses, the surface before gelation of the resin.
0.23 cm and 0.15 cm. The thicknesses of the laminate speci- This is a case where the air was trapped between the adhe-
mens as viewed under the microscope were measured to be sive that flowed from the prepreg, although the plies were de-
0.22 cm for the MTM45-1 panel (Fig. 4a) and 0.25 cm for the bulked, and a vacuum was well applied in the layup process.
977-2 panel (Fig. 4c). No porosity was observed in the skins of both sandwich
Figure 5 displays the profile of the sandwich panel with and panels and laminate panels. However, fillet size is often cor-
without adhesive film. The skin-to-core interface of the two related to bond quality at the interface. The strength of the
specimens from the two different panels shows a considerable bond depends heavily on the fillets formed in the curing pro-
deposit of resin between the walls of the core cell and the skin, cess by the flow of resin. The sandwich panels with and with-
as indicated by arrows in Fig. 5c,d. The middle cell of the out the adhesive film had desirable filleting at the skin-to-core
adhesiveless panel in Fig. 5c appears to show a buildup of interface. Again, no significant physical difference was seen

Fig. 4 Microscopic images of laminate panels. a MTM45-1. b 977-2

Adv Compos Hybrid Mater (2019) 2:339–350 345

a b

Scale: 20/1 Scale: 20/1

c d

Scale: 100/1 Scale: 100/1

Fig. 5 Microscopic images of sandwich panel. a MTM45-1 adhesiveless panel. b 977-2 with adhesive film and interface of sandwich panel, c MTM45-1
adhesiveless panel. d 977-2 panel with adhesive film

between the skin of the sandwich panel with self-adhesive MTM45-1 systems in the skin of the adhesiveless panel
prepreg and the sandwich panel with adhesive film. and by the additional adhesive film in the sandwich panel
with the 977-2 bleeding into the prepreg. The low resin
3.1.2 Constituent content weight fraction can also be justified by the flow of resin
at the interface to form the fillet. The average sample den-
The average resin weight percentages for each of the sities exhibited in Fig. 6b were higher for both the laminate
panels are depicted in Fig. 6a. Both the laminate with and sandwich skin specimens with the 977-2 prepreg. Yuan
MTM45-1 prepreg and the laminate with 977-2 prepreg et al. fabricated solvent solution-coated self-adhesive pre-
had average resin weight percentages within the tolerance pregs, and then cured sandwich panels using hot-press
band published by the manufacturers for those prepregs, molding while varying the temperature ramp rate, pressure,
that is, 36 ± 3% for MTM45-1 and 42 ± 3% for 977-2. and application time of pressure, in order to determine the
Skins from the sandwich panels with and without adhesive best combination of parameters [21]. The optimum speci-
film each had an average resin weight fraction of about men with less than 0.4% porosity had a bottom skin RW
30%. These values from the skins were not expected to that was nearly 50% higher than that of the top skin. Even
accurately define this characteristic of the prepreg. The focusing on the RW alone, the bottom skins of the best
additional influence was driven by the mixture of the two specimen were about 25% higher than the top skins.

Fig. 6 Laminate and sandwich 50%

panels. a Resin weight. b Sample
density MTM 45-1 1.60 MTM 45-1
40%
977-2
977-2
Resin Weight (%)

1.20
Sample Density (g/cc)

30%

20% 0.80

10% 0.40

0% 0.00
Laminate Sandwich Laminate Sandwich

(a) (b)
346 Adv Compos Hybrid Mater (2019) 2:339–350

Fig. 7 Skin resin weight 40% top skin

40% top skin
percentages. a Adhesiveless bottom skin
bottom skin
sandwich specimen (MTM45-1).
30%
b Adhesive film specimen (977- 30%
2), hot-melt/autoclave cured

Resin Weight ( %)
Resin Weight (%)
20%
20%

10% 10%

0% 0%
1 2 3 4 5 6 1 2 3 4 5 6
Specimens Specimens

a) b)

In comparison, the RW of the top and bottom skins of the 2) and the adhesive film in the sandwich panel. The face
adhesiveless sandwich specimens, as well as those of the sheets of a sandwich panel have minimal porosity, which is
sandwich with adhesive film specimens, were consistently critical for the mechanical properties of the whole sandwich
close to equal, as is needed in honeycomb sandwich structures structure. For primary aircraft structures with honeycomb
(Fig. 7). This emphasizes the choice of the hot-melt impreg- sandwich panels and as indicated by the manufacturer’s pre-
nation method and autoclave curing medium for better control preg data, the maximum void content allowed in the skin
and an ideal process for honeycomb sandwich structures. laminate is 2%. Yuan et al. recommended a 1% maximum
The average fiber volume fraction obtained for the void content for external surfaces because of their aerodynam-
MTM45-1 laminate was 56.43%, which was within the range ics requirement in service [21]. In addition, both laminate
provided by the manufacturer for the MTM45-1 prepreg panels and skins from both sandwich panels have a void con-
(52.2–56.6%). Furthermore, for the 977-2 laminate, the fiber tent well below 1%. The MTM45-1 prepreg used in the
volume fraction was 50.79%, which is consistent with the adhesiveless sandwich panel had a nearly zero void of only
manufacturers provided data (47–57%). The MTM45-1 had 0.04%, whereas the sandwich panel using the 977-2 prepreg
a lower resin volume fraction than the 977-2. The sandwich with adhesive film had a void content of 0.13%. The void
panels had a higher fiber volume fraction, 59.70% for the content from both sandwich panel skins was nearly equal,
adhesiveless sandwich panel with MTM45-1, and 62.55% 0.52% for the adhesiveless panel and 0.56% for the panel with
for the adhesive film sandwich panel with 977-2. However, adhesive film. Acid digestion test methods were applied to
the composition of the epoxy in each of the constituents could measure the void content of the test coupons. In the specimen
result in the different bleeding behavior of the resin during descriptive statistic data for average density, resin weight per-
curing. For both panels, some of the resin flowed into the cent, and volume fractions from all panels, the standard devi-
interface to form the fillet, with the resin flowing from the ation also showed a very small deviation from the mean,
prepregs (MTM45-1 36% RW and the one layer 50% RW) confirming the accuracy of the testing and the consistency of
for the adhesiveless panel and the resin from the prepreg (977- the specimen properties.

Fig. 8 Tg values of different 250

Glass Transition Temperature (°C)

composite panels used in present

study
200

150

100

50

0
Laminate Laminate Bagside Bagside with Toolside Toolside with
MTM45-1 977-2 adhesiveless adhesive film adhesiveless adhesive film
MTM45-1 977-2 MTM45-1 977-2
Adv Compos Hybrid Mater (2019) 2:339–350 347

3.1.3 Glass transition temperature strength, 14 to 16% higher than a panel with adhesive film.
Besides at room temperature, the composite panels exhibited
The glass transition temperature of epoxies for high- higher strength than at low and high temperatures. However,
temperature applications is typically less than 260 °C [22]. higher-temperature environments are more detrimental to the
Some aircraft manufacturers use a minimum Tg of 160 °C mechanical properties of these structures.
for the resin system in primary sandwich structures. The Tg The panel with MTM45-1 without adhesive consistently
obtained from different panels is illustrated in Fig. 8. In all had greater strength, 14 to 16% higher than the panel with
cases, the Tg was above 175 °C, which is highly acceptable for adhesive film. In addition, at room temperature, the strength
aircraft applications. was greater than at low and elevated temperatures. The short-
It is important to note that the manufacturer’s published beam strengths were all above 58 MPa, which is considered
maximum temperature for the Nomex core and AF191 adhe- acceptable for these types of materials for use in primary
sive film was 176.67 °C. These materials are commonly used structures [22]. The laminate with 977-2 prepreg had a greater
in the aircraft industry for primary structures. The laminate short-beam shear strength of about 28% higher than MTM45-
specimens with MTM45-1 prepreg had a greater Tg than that 1, as illustrated in Fig. 9b. Constituent content test results of
of the 977-2 prepreg (215.28 °C for MTM45-1 and 179.48 °C the laminates showed that the laminate with 977-2 prepreg
for 977-2). A higher Tg is more desirable because the material had a greater resin volume fraction (49.07%) than the laminate
can be used at higher temperatures. The high Tg of the with MTM45-1 prepreg (43.55%), which could also amelio-
MTM45-1 was attributed to the nature of the resin system, rate a greater resistance of the matrix in the 977-2 laminate.
the composition of which is only known to the manufacturer. The fracture surfaces of the laminates from the short-beam
Similar results obtained from the sandwich panels are also shear and tensile tests are shown in Fig. 10. The interlaminar
shown in Fig. 8. The Tg of the laminate with MTM45-1 pre- shear failure is clearly visible from the fracture surface, as
preg (adhesiveless panels) was higher than that of the 977-2 shown in Fig. 10a,b. Both laminates from the MTM45-1 and
prepreg (adhesive film panels). During the curing of a sand- 977-2 prepregs exhibit predominant microcracks in the resin,
wich panel, heat comes through the top (bag side) first, which with delamination as well as minor flexural failure on the
should provide a higher Tg than what is obtained on the tool edge, showing fiber microbuckling. Overall, specimens from
side. It takes time for the heat to dissipate, with the core adding both laminate panels displayed a combination of failure
an insulating effect [5, 22–24]. In this work, the Tg of the bag modes, but the inelastic deformation was not identified. The
side was not consistently higher for all specimens, especially failure mode of a sandwich panel in tensile testing is complete
for the 977-2 material. It is possible that the adhesive film later core failure, which demonstrates that the strength of the adhe-
comingled with the resin in the 977-2 prepreg, both of which sion at the laminate and core interface is greater than the
have different epoxy compositions, thus causing inconsisten- strength of the core itself, as indicated in Fig. 10c,d.
cy from one specimen to the other and hence a lower Tg. Adhesion in the adhesiveless sandwich panel at the interface
as well as with the adhesive film meet this structural require-
3.2 Mechanical properties of composites ment. Other failure modes at the interface, such as cohesive
failure, adhesive failure, and facing tensile failure, which in-
The flatwise tensile strength and short-beam shear strength of volves delamination of the plies, are all acceptable, if the min-
laminate and sandwich panels are depicted in Fig. 9. The panel imum flatwise tensile strength set by the equipment designer
with MTM45-1 without adhesive consistently had greater is reached prior to failure.

Fig. 9 Dry laminate and 3 Adhesiveless sandwich MTM45-1

sandwich panels. a Flatwise Sandwich with adhesive film 977-2 100

tensile strength. b Short-beam 2.5
Flatwise Tensile Strength (MPa)

Short Beam Shear Strength (MPa)

shear strength 80
2

60
1.5

40
1

0.5 20

0 0
24°C - 55°C 82°C Laminate MTM45-1 Laminate 977-2
Temperature (°C)

a) b)
348 Adv Compos Hybrid Mater (2019) 2:339–350

a b

c d

Fig. 10 Fracture surface of specimens from short-beam shear tests: a Laminate with prepreg MTM45-1 and b Laminate with prepreg 977-2. Tensile
fracture specimens without adhesive sandwich panels: c with MTM45-1 prepreg and d with 977-2 prepreg

3.3 Moisture absorption of laminate and sandwich MTM45-1 had already reached its saturation by day 26.
specimens Saturation was validated for each specimen by ensuring that
the percent weight gain was at the most 0.02% for at least
In comparing the laminate panels with MTM45-1 and 977-2 4 days. Validation was performed within a 9-day window with
prepregs, the average percent weight gain of each panel was data sets. During that timeframe, all specimens remained be-
very similar for the first 12 days, as shown in Fig. 11a. After low the 0.02% weight set as saturation. MTM45-1 samples
9 days, the 977-2 panel started gradually increasing in weight, did not gain any more weight in that timeframe, except for the
up to the equilibrium weight gain at around day 36, while the 0.005%, which could be due to the testing and measurement

1.2
12
1
10
0.8
8
% Weight Gain
% Weight Gain

0.6
6
MTM 45-1
0.4 MTM 45-1: adhesive-less
4
977-2
977-2: with adhesive film
0.2 2

0 0
1 5 12 19 26 33 36 40 42 5 12 19 26 33 36 40 46 50
Days Days
(a) (b)
Fig. 11 Moisture absorption of specimens. a Laminate panels. b Sandwich panels
Adv Compos Hybrid Mater (2019) 2:339–350 349

Fig. 12 Moisture-absorbed 2

Flatwise Tensile Strength (MPa)

100
sandwich and laminate panels. a

Short Beam Shear Strength (MPa)

Flatwise tensile strength. b Short- 1.6
beam shear strength 80

1.2
60

0.8
40

0.4
20

0
Adhesiveless Sandwich Sanswich with Adhesive 0
MTM45-1 Film 977-2 Laminate MTM45-1 Laminate 977-2

a) b)

process, whereas for the 977-2, at least two specimens fluctu- observed with the dry specimens, the 977-2 specimens con-
ated in percent weight during that time. The overall average sistently had a greater short-beam shear strength than the
percentage weight for MTM45-1 was 0.77%, which was MTM45-1 specimens, as illustrated in Fig. 12b. In the fracture
much lower than for 977-2 at 1.05%. Weight gain does show surface of the short-beam shear test, gouges were observed;
the hydrophilic nature of epoxy in the prepreg, as explained by these lower the short-beam shear strength of both type lami-
Crossman et al. [14]. Both sandwich panels with and without nate panels. Due to moisture, the average short-beam shear
adhesive film have close average percent weight gains strength was reduced, by 7% for the panel with 977-2 prepreg
throughout the 50 days in the humidity chamber, as shown and by 11% for MTM45-1 panel. Degradation of the short-
in Fig. 11b. The intent moisture absorption in the sandwich beam shear strength due to moisture was much lower than that
specimen is to orchestrate a random range in the water seen for flatwise tensile strength (20% on 977-2 and 28% on
absorbed by specimens for the flatwise tensile test, with an MTM45-1). Sandwich specimens are more susceptible to deg-
overall percentage weight gain between 4 and 15%. radation because they have a greater amount of material with
which to absorb moisture: two skins containing both fiber and
3.4 Mechanical properties of moisture-absorbed resin, the interface fillet which contains resin, and the core
laminate and sandwich panels being another composite of its own. The laminate specimens
tested for short-beam shear strength had only fiber, hence less
The flatwise tensile and short-beam shear test of the moisture- moisture absorption. The loss of strength in mechanical prop-
absorbed specimens was performed at room temperature. The erties seen in this experiment was significant, with the simu-
flatwise tensile and short-beam shear strengths of the moisture- lation of the damaged material allowing more moisture ab-
absorbed laminate and sandwich panels from two different pre- sorption in the honeycomb.
preg system are depicted in Fig. 12a. All specimens exhibited
failure in the core. Water was visually observed in all the spec-
imens after testing. Moisture absorption reduced the flatwise 4 Conclusions
tensile strength in both categories of specimens. The
adhesiveless panel with MTM45-1 had a strength only 5% Sandwich panels with and without adhesive film were fabri-
greater than the panel with the adhesive film and 977-2, 16% cated and tested for physical and mechanical properties. In the
greater compared to the dry specimens at room temperature. experiments performed, all four fabricated panels were shown
This implies that this panel was more affected by moisture to have a consistent thickness within each panel. Ultrasonic
absorption and lost more strength as a result. In comparing inspection did not reveal any porosity and confirmed that the
the different testing environments, it is important to note that panels were well-fabricated and good candidates for test spec-
the moisture-absorbed specimens tested at room temperature imens. Furthermore, microscopic inspection confirmed the
had lower strength than the dry specimens tested at cold tem- tight layering in the laminates and sandwich skins. Fillets at
perature, yet higher than the dry specimen tested in hot temper- the skin-to-core interface in the sandwich panel with the
ature. This observation validates the hot-wet environment com- MTM45-1 self-adhesive prepreg were as well-formed as the
monly known as the most unfavorable to the mechanical prop- fillets in the sandwich panel with the 977-2 prepreg with AF
erties of composite honeycomb sandwich structures. 191 adhesive film. Relative to the mechanical properties test-
The fracture surfaces of both panels prepared from the two ed, the flatwise tensile test of dry specimens at room, cold, and
different prepregs MTM45-1 and 977-2 showed microcracks high temperatures successfully resulted in core failures for
in the resin, which is a typical inter-laminar failure. As both groups of sandwich panels. The adhesiveless sandwich
350 Adv Compos Hybrid Mater (2019) 2:339–350

panel with MTM45-1 had a 14–16% higher flatwise tensile 8. Rion J, Leterrier Y, Manson JE (2008) Prediction of the adhesive
fillet size for skin to honeycomb core bonding in ultra-light sand-
strength than the sandwich panel with 977-2 and AF191 ad-
wich structures. Compos Part A 39(9):1547–1555
hesive film, in all conditions tested, testifying to the stronger 9. Zhou L, Christou, P D (2003) Self-adhesive prepreg system for
bond at the skin-to-core interface. The short-beam shear sandwich panel applications, US patent # 6508910. http://www.
strengths of dry laminates, all above 58 MPa, were also suit- freepatentsonline.com/6508910.html
10. Hedges WL, Buyny RA, Boyle MA, Martin CJ, and Kevin D
able indicators of polymeric composite materials for primary
(2004) Development of a self-adhesive prepreg system for sand-
structural use. The average short-beam shear strength of the wich panel applications. SAMPE, Long Beach, CA
977-2 laminate was 28% higher than that of the MTM45-1 11. Shafizadeh JE, Hayes BS, Seferis JC (1997) Fundamentals of
laminate, hence displaying a tougher epoxy resin. Overall, the breather technology in prepreg consolidation. J Adv Mater
29(11):42–49
moisture-absorbed specimens from both laminate and sand-
12. Hayes BS, Seferis JC, Edwards RR (1998) Self-adhesive honey-
wich panels exhibited mechanical strengths acceptable for pri- comb prepreg systems for secondary structural applications. Polym
mary structures and within the parameters tested. Self- Compos 19(1):54–64
adhesive prepregs such as MTM45-1 offer the physical and 13. Shafizadeh JE, Seferis JC, Chesmar EF, Geyer R (1999) Evaluation
of the in-service performance behavior of honeycomb composite
mechanical properties necessary for adhesiveless honeycomb
sandwich structures. J Mater Eng Perform 8(11):661–668
primary structures. Eliminating the use of adhesive at the skin- 14. Crossman FW, Mauri RE and Warren WJ (1979) Hygrothermal
to-core interface reduces the cost of material and results in a damage mechanisms in graphite-epoxy composites. NASA
weight reduction that translates to fuel savings, hence signif- Contractor Report, No. 3189
15. Chalkley PD, Chester RJ (1990) The influence of tropical exposure
icant cost savings overall.
on the durability of graphite/epoxy laminates. Theor Appl Fract
Mec 13(2):137–144
Acknowledgments The authors gratefully acknowledge Wichita State 16. Shafizadeh JE, Seferis JC (2000) Effects of long-time water expo-
University and the National Institute for Aviation Research for technical sure on the durability of honeycomb. International SAMPE
and financial support of these research studies. Symposium and Exhibition (Proceedings). 45:380–388
17. Crocket A, Loken H, and Tomblin J (2008) Fluid ingression dam-
Compliance with ethical standards age mechanism in composite sandwich structures. 4th Annual Joint
Advanced Materials and Structures Center of Excellence (JAMS)
Technical Review
Conflict of interest The authors declare that there is no conflict of
18. Jumahat A, Soutis C, Jones FR, Hodzic A (2010) Fracture mecha-
interest.
nisms and failure analysis of carbon fibre/toughened epoxy com-
posites subjected to compressive loading. Compos Struct 92(2):
295–305
References 19. Advanced Composites Group, ACG’s MTM®45-1 prepreg used in
Antarctic ice sheet mapping mission, (2012)
20. Lubin G (2013) Handbook of composites. Van Nostrand Reinhold
1. Xu Y, Zhu J, Wu Z, Cao Y, Zhao Y, Zhang W (2018) A review on
Company, New York
the design of laminated composite structures: constant and variable
21. Yuan C, Li M, Zhang Z, Gu Y (2015) Experimental investigation
stiffness design and topology optimization. Adv Compos Hybrid
on the co-cure processing of honeycomb structure with self-
Mater 1:460–477
adhesive prepreg. Appl Compos Mater 15(1):47–59
2. Kumar SSA, Uddin MN, Rahman MM, Asmatulu R (2019)
22. Gandy HTN and Asmatulu R (2012) Adhesiveless composite struc-
Introducing graphene thin films into carbon fiber composite struc-
tures with carbon fiber prepregs for aircraft primary structural ap-
tures for lightning strike protection. Polym Compos 40:E517–E525
plications. ASME International Mechanical Engineering Congress
3. Uddin MN, Le L, Zhang B, Nair R, Asmatulu R (2019) Effects of
and Exposition, Houston, TX
graphene thin films and nanocomposite coatings on fire retardancy
23. Das TK, Ghosh P, Das NC (2019) Preparation, development, out-
and thermal stability of aircraft composites. J Eng Mater Technol
comes, and application versatility of carbon fiber-based polymer
141:031004–1–7. https://doi.org/10.1115/1.4042663
composites: a review. Adv Compos Hybrid Mater:1–20. https://
4. Federal Aviation Administration, Electronic Code of Federal
doi.org/10.1007/s42114-018-0072-z
Regulations. http://ecfr.gpoaccess.gov/, December (2011)
24. Kashfipour MA, Mehra N, Zhu J (2018) A review on the role of
5. Gandy HTN (2012) Adhesiveless honeycomb sandwich structure interface in mechanical, thermal, and electrical properties of poly-
with carbon graphite pre-preg for primary structural application: a mer composites. Adv Compos Hybrid Mater 1(3):415–439
comparative study to the use of adhesive film. M.S. Thesis, Wichita
State University
6. Yuan C, Li M, Gu Y, Zhang Z (2008) Experimental study of the
cocure process with vacuum bag method for honeycomb sandwich
structure. Acta Materiae Compositae Sinica 25(2):57–62 Publisher’s note Springer Nature remains neutral with regard to
7. Fogarty JH (2010) Honeycomb core and the myths of moisture jurisdictional claims in published maps and institutional affiliations.
ingression. Appl Compos Mater 17(3):293–307