Journal of Building Engineering 20 (2018) 114–122

Contents lists available at ScienceDirect

Journal of Building Engineering

journal homepage: www.elsevier.com/locate/jobe

Corrugated cardboard core sandwich beams with bio-based flax fiber T

composite skins
⁎
Aidan McCracken, Pedram Sadeghian
Department of Civil and Resource Engineering, Dalhousie University, Halifax, NS, Canada B3H 4R2

A R T I C LE I N FO A B S T R A C T

Keywords: This paper presents an experimental study on the behavior of sandwich beams made of green materials for both
Corrugated cardboard core and skin components. A unidirectional flax fabric and a partial bio-based epoxy were used to make fiber-
Flax FRP reinforced polymer (FRP) skins and three flute varieties of corrugated cardboards (known as B, C, and BC flutes)
Bio-based polymer with the bulk densities of 170, 127, and 138 kg/m3 were used for the core, respectively. A total of 30 small-scale
Green
sandwich beam specimens were manufactured across six unique beam varieties with dimensions of 50 mm in
Sandwich
width, 25 mm in depth, and 200 and 350 mm in length (150 mm and 300 mm spans) and tested under four-point
bending up to failure. Two failure modes of transverse indentation for the short specimens and longitudinal
crushing of the core and skin for the long specimens were observed. The load-deflection, load-strain, and mo-
ment-curvature behaviors were analyzed to evaluate the strength and stiffness of the sandwich beam specimens.
C flute with the lowest bulk density and the highest availability in the market amongst all the three flutes
exhibited the highest strength and stiffness for sandwich applications. Overall, the corrugated cardboard cores
combining with the flax FRP skins may be considered as a viable, green option for the fabrication of large-scale
structural sandwich panels for building applications.

1. Introduction that this may be acceptable since the core failure is one of common
failure modes in sandwich structures with strong skins [7,8]. However,
Composite sandwich structures made of fiber-reinforced polymer sandwich structure may experience different failure modes (core shear,
(FRP) skins and lightweight, low-density core materials have been indentation, skin crushing/rupture, and skin wrinkling) dependent on
shown to be very effective in reducing weight and increasing strength the ratio of skin thickness to span length and relative core density
and stiffness in a variety of construction and building applications. The [9,10]. Additionally, natural fibers have many economic and environ-
FRP skins resist the tensile and compressive stresses under flexure, like mental advantages compared to synthetic fibers [11–13]. Thus, FRP
the action of the flanges on an I-shaped beam, while the core resists skins made of natural fibers represent a viable structural option for
shear stresses, increases the distance between skins resulting in a higher sandwich structures and are a more environmentally-friendly choice
moment of inertia, and provides insulation for the system. The popu- than synthetically produced fibers. In FRPs, the role of polymer resins
larity of sandwich structures in the form of wall, cladding, roof, and to impregnate and bond the fibers is critical. Synthetic polymers, such
floor panels is growing as engineers look to improve the structural ef- as epoxy and vinyl ester, have been used with natural fibers [14–16].
ficiency and insulation properties of buildings. To be more en- However, numerous studies have been conducted on the use of fully or
vironmentally-conscious, conventional materials and structures need be partially bio-based resins with natural fibers [17–19].
re-evaluated to determine how they can become more sustainable and Beside of skin materials, the core materials of sandwich structures
have a smaller environmental impact during manufacturing, installa- play a major role in the structural (i.e. shear, composite action, and out-
tion, and service. of-plane properties) and insulation properties. Many different core
Although FRP composites made of synthetic fibers, such as glass or materials have been explored for use in sandwich structures. Core
carbon fibers, are often used for the skins of sandwich panels [1–3], the materials that are commonly studied include low-density foam and
concept of using plant-based natural fibers, such as flax and hemp fi- plastic or metal honeycombs [20–22]. To present a more sustainable
bers, has also been explored [4–6]. Although the natural fibers have a option, this study considers corrugated cardboard as the core material.
lower strength than their synthetic counterparts, it has been showed According to the Paper and Paperboard Packaging Environmental

⁎
Corresponding author.
E-mail address: [email protected] (P. Sadeghian).

https://doi.org/10.1016/j.jobe.2018.07.009
Received 6 January 2018; Received in revised form 15 May 2018; Accepted 11 July 2018
2352-7102/ © 2018 Elsevier Ltd. All rights reserved.
A. McCracken, P. Sadeghian Journal of Building Engineering 20 (2018) 114–122

Council (PPEC), approximately 85% of corrugated cardboard in Canada Table 2

is recycled and new cardboard is produced with nearly 100% recycled Cardboard flute comparison.
materials [25,22]. Along with being 100% biodegradable, corrugated Cardboard flute Thickness (mm) Flutes per meter Bulk density (kg/
cardboard is a very sustainable as it can be repurposed and produces length m3)
very little waste. Corrugated cardboard is an orthotropic sandwich with
B 2.8 160 170
two paper based facings separated by a lightweight corrugated core
C 4.0 120 127
known as fluting [23]. The system has good thermal insulation prop- BC 6.6 Mix 138
erties and high load‐carrying capacity by being low‐costs and light-
weight [24]. Although studies have been conducted on balsa wood and
cork cores for sandwich composites [4,5,26,27], corrugated cardboard 2.2. Material properties
has not been explicitly studied in the context of a sandwich composites
with natural fibers and bio-based resins. Three unique cardboard flutes were used in the fabrication of the
In this study, flax FRP skins were combined with corrugated card- sandwich beams: B, C and BC flutes. Cardboard flutes are standard in
board cores to manufacture sandwich beam specimens. Unidirectional international packing and are identified with a single capital letter.
flax fabrics were impregnated using a resin with almost 30% bio con- Each flute has a different nominal thickness and density. Table 2
tent. As a result, the sandwich beams produced were constructed using compares the approximate measured dimensions of each flute in this
almost entirely green materials. The aim of the study is to analyze and study. The density measurements were taken after the flute layers had
evaluate the structural performance of corrugated cardboard and flax been combined into a core for the specimens. Thus, this density reflects
FRP sandwich beams as an alternative to conventional construction and the actual density of the core, including the small amount of starch-
building materials. Although flax has previously studied for use in based adhesive used to combine the layers of cardboard.
sandwich beams, the combination of flax FRP with cardboard has yet to Fig. 1 shows a visual comparison between the flutes with both a
be analyzed. This combination of materials represents a structural photo of the flutes as a part of a core as well as a 2D side-view sche-
system that has a minimal impact on the environment as corrugated matic. For the flax FRP skins, a unidirectional flax fabric with a re-
cardboard is readily available and composed almost entirely of recycled ported aerial weight of 275 g/m2 (gsm) was used. In terms of epoxy,
and green material. Super Sap ONE was used, which is a bio-based epoxy (30% bio content)
with a reported tensile strength, modulus and elongation of approxi-
mately 53.23 MPa, 2.65 GPa and 6%, respectively. Betts et al. [26]
2. Experimental program conducted a study on the tensile properties of flax FRP composites
manufactured using the same unidirectional flax fabric and three dif-
This section presents the details of test matrix, material properties, ferent epoxies. For the flax FRP samples made of the bio-based epoxy
specimen preparation, test setup, and instrumentation of the test spe- used in this study, the average tensile strength and initial modulus were
cimens. The experimental program using two loading arrangements reported to be 198.0 ± 9.3 MPa and 17.09 ± 0.63 GPa, respectively
was specifically designed to investigate the properties of three corru- (two-layer thickness = 1.97 ± 0.09 mm). A secondary modulus was
gated cardboard cores. reported as 11.93 ± 0.39 GPa as it was found flax FRPs display an
approximately bi-linear mechanical behavior. The rupture strain of flax
FRP was reported to be 0.0153 ± 0.0006 mm/mm.
2.1. Test matrix

A total of 30 flax FRP and corrugated cardboard sandwich beams 2.3. Specimen preparation
were fabricated to be tested in four-point bending. All specimens were
constructed using one layer of flax FRP skin on either side and a cor- The first step in the fabrication of the sandwich beams was to
rugated cardboard core with a thickness of approximately 25 mm. The construct the cardboard cores. To do this, strips of cardboard (manu-
variables being tested were span length as well as the flute sizes of the facturer: Maritime Paper, Dartmouth, NS, Canada) approximately
corrugated cardboard. Two span lengths of 150 mm and 300 mm as 25 mm in width were cut from larger sheets using a straight edge and a
well as three flute sizes of B, C, and BC were tested. More information sharp blade. The two span lengths being tested were 150 and 300 mm,
concerning the flute sizes can be found in the next section. A complete thus strips were cut to lengths of 200 and 350 mm to provide an
summary of this study's test matrix is shown in Table 1. Note that three overhang of approximately 25 mm on each end of the specimen. To
identical specimens were manufactured and tested per case. All speci- bond the strips together, a small amount of a vegetable starch-based
mens are identified with a specimen identification (ID) which follows adhesive known as Tri-Tex Tribond P-1031 adhesive was used. This
the format X-SY where X identifies the flute size, S stands for span and Y adhesive was provided by the cardboard manufacturer and is the same
identifies the specimens test span in mm. For example, the specimen B- used in the manufacturing of corrugated cardboards. The number of
S150 designates a flax FRP and cardboard sandwich beam constructed strips in the core varied per flute as all cores were manufactured to have
using B flute cardboard with a test span length of 150 mm. an approximate width of 50 mm. Fig. 2 shows the fabrication process of
the cardboard cores.
Once the cardboard cores were completed, the flax FRP skins were
Table 1 applied using the standard wet lay-up method. Sheets of flax fabrics
Test matrix. approximately 300 mm in width and either 200 or 350 mm in length
were pre-cut before the mixing of the epoxy. A sheet of parchment
Case # Specimen ID Cardboard flute Span (mm)
paper was put on the bottom surface and a layer of epoxy was applied.
1 B-S150 B 150 Next, a sheet of flax fabric was applied to the epoxy, then the top side of
2 B-S300 B 300 the fabric was saturated with another layer of epoxy. Each of the five
3 C-S150 C 150 cores per case was placed on the saturated sheet of flax. A piece of
4 C-S300 C 300
particle board was placed on top of the cores while the bottom layer of
5 BC-S150 BC 150
6 BC-S300 BC 300 flax FRP cured. Once the first side of had cured, this process was re-
peated for applying the flax FRP skin to the other side of the cores. This
Three identical specimens per case were prepared and tested. method allowed for the curing FRP to always be below the cardboard

115
A. McCracken, P. Sadeghian Journal of Building Engineering 20 (2018) 114–122

Fig. 1. Visual comparison of cardboard sheets with B, C, and BC flutes.

core to help ensure that unwanted resin did not seep down into the
cardboard. Fig. 3 shows the application process of the second side of
flax FRP. Applying larger sheets of flax fabric allowed for a quicker
fabrication process. Once both sides fully cured, a bandsaw was used to
cut the beams to their approximate width of 50 mm and a rotary sander
was used to smooth the edges of the flax composite and ensure it was in
line with the sides of the core. A completed sandwich beam is shown in
Fig. 4.

2.4. Test setup and instrumentation

The test setup was designed based on ASTM D7250 [29] to test each
type of sandwich beams with two different loading arrangements to-
wards obtaining the flexural stiffness and shear stiffness. All specimens
were tested under four-point bending per with a loading span propor-
tional to the supporting spans of 150 and 300 mm. The loading span (L)
Fig. 2. Cardboard core fabrication: (a) cutting; (b) applying adhesive; (c) was selected to be equal to (2/11) of the supporting span (S). A sche-
combining into one core; and (d) completed cores for 150 mm span. matic of the four-point bending setup is shown in Fig. 5, where P is the
total applied load. In terms of instrumentation, a strain gauge was ap-
plied on either side of the sandwich beam, centered in the longitudinal
direction to measure the tensile and compressive strains. Additionally,
two linear potentiometers (LPs) were setup in the middle of the beam's

Fig. 3. Specimen fabrication: (a) applying epoxy; (b) saturating flax fabric; (c) placing cardboard cores on saturated fabric; (d) first side complete; and (e) both sides
complete.

116
A. McCracken, P. Sadeghian Journal of Building Engineering 20 (2018) 114–122

Fig. 4. Sandwich beam specimens before and after strain gauging.

3.1. Failure modes

As expected, the failure of the core was the initial source of failure
in both the 150 and 300 mm span sandwich beams. Due to their higher
stiffness, the 150 mm span specimens did not flex very much, only
deflecting and average of 3.37 mm at peak load. All three tested
150 mm specimens failed by indentation crushing of the core due to the
transverse loads. This was followed by indentation of the top skin.
However, the 300 mm span specimens reached a significantly lower
peak load and failed by longitudinal crushing of the core due to com-
pressive bending stresses. Once the corrugated cardboard begun
crushing longitudinally, this created a noticeable increase in compres-
sive strain on the top of the beam, which caused the top skin to be
crushed after the peak load. Images of these two failure modes can be
seen in Fig. 6.

3.2. Load-strain behavior

The longitudinal strains of top and bottom skins at mid-span of each

specimen were recorded. Curves comparing the load-strain behavior of
the specimens are shown in Fig. 7. Each curve is the average of three
Fig. 5. Test setup and instrumentation: (a) four-point bending schematic; (b)
identical specimens of each group of specimens. The 300 mm span
testing machine and data acquisition system; and (c) specimen ready for testing.
specimens experienced much larger compressive strain compared to
tensile strain. This was caused by the longitudinal crushing of the
span to measure an average mid-span deflection. These values, along cardboard near the top skin of the sandwich. The 150 mm span samples
with the applied load, measured every 0.1 s, were collected for data experienced comparable tensile and compressive strains until the in-
processing. All tests were completed using a 100 kN universal testing dentation crushing of the cardboard core. Overall, the specimens with B
machine and were displacement controlled using a fixed rate of 2 mm/ and C flutes showed better performance than the specimens with BC
min. flute. The load-strain behavior of the 150 mm span specimens with B
and C flutes were very close at the tension side until around the peak
load. However, at the compression side, the strain of specimens with B
3. Test results and discussion flute went more non-linear. As the tensile rupture strain of flax FRP was
about 0.015 mm/mm based on tensile coupon tests, it can be seen that
This section presents the derails of failure mode, load-strain, mo- only the 150 mm span specimens with B and C flutes reached to that
ment-curvature, load-deflection, and flexural and shear stiffness of the level of strain, however the specimens failed by indentation of com-
test specimens. A summary of the test results as well as the modes of pression side under the load. In the 300 mm span specimens, the strain
failure of the specimens is shown in Table 3. of tensile skins of B and C flutes reached to 0.013 and 0.009 mm/mm,

Table 3
Summary of test results.
Specimen group ID Peak load (N) Initial stiffness (N/mm) Peak moment (N m) Curvature at peak (1/km) Flexural stiffness (N m2) Failure mode

AVG SD AVG SD AVG SD AVG SD AVG SD

B-S150 3587 279 2671 9.0 110.1 8.6 1003 164 257 22 Indentation
B-S300 1493 13 252 12.1 91.6 0.8 1163 136 173 14 Crushing
C-S150 3985 297 2209 154.6 122.3 9.1 953 188 231 10 Indentation
C-S300 1715 201 387 27.1 105.2 12.4 1168 208 215 21 Crushing
BC-S150 1881 297 2089 336.4 57.7 9.1 564 254 209 29 Indentation
BC-S300 917 161 329 26.8 56.3 9.9 677 72 175 10 Crushing

AVG = average; and SD = standard deviation.

117
A. McCracken, P. Sadeghian Journal of Building Engineering 20 (2018) 114–122

Fig. 6. Failure modes: (a) indentation; (b) longitudinal crushing; and (c) detail
of top skin crushing.

Fig. 8. Moment-curvature behavior of specimens: (a) 150 mm span; and (b)

300 mm span (note: each curve is the average of three identical specimens).

3.3. Moment-curvature behavior

The moment in the central loading span of each specimen was

calculated at any load level. The curvature of each specimen was also
calculated based on the slope of strain profile at mid-span. The strain
profile was assumed linear and obtained using the tensile and com-
pressive strains of skins discussed earlier. As expected, the moment-
curvature behavior was similar between the two spans that were
compared. Graphs comparing the moment-curvature behavior of the
specimens are shown in Fig. 8. The specimens with 150 mm span and B
and C flutes exhibited very close moment-curvature behavior with al-
most the same ultimate curvature. Both the specimen groups, showed
significantly better performance than the specimens with BC flute. As
the specimens with 300 mm span failed due to flexure, moment-cur-
vature curves are more applicable explaining the behavior of these
specimens. As shown in Fig. 8, the specimens with C flute exhibited a
higher flexural stiffness and strength than those with B and BC flutes.
Flexural stiffness of the specimens was calculated based on the slope of
the linear region of each curve and presented in Table 3. The results will
be discussed further in Section 3.5.
Fig. 7. Load-strain behavior of specimens: (a) 150 mm span; and (b) 300 mm
span (note: each curve is the average of three identical specimens).
3.4. Load-deflection behavior

which are slightly less than the tensile rupture strain of 0.0153 mm/ The mid-span deflection of each specimen was obtained by aver-
mm. This shows that the two cardboards core were strong and stiff aging the displacement data from the two LPs installed on the bottom
enough to develop that level of tensile strains. At the compression side, skin. Each load-deflection curve in Fig. 9 represents the average load-
the flax FRP skins of both B and C flutes experienced a maximum deflection curves of three identical specimens of each group. The figure
compressive strain of almost 0.014 mm/mm before the load dropped. indicates that both 150 and 300 mm span specimens experienced a
Overall, B and C flutes showed compatible strength and stiffness to the short linear behavior followed with a long non-linear region up to the
flax FRP skins. peak load. The initial stiffness K of each specimen was calculated base
don the slop of the linear region. As shown in Table 3, the initial
stiffness of the specimens with different flutes are close to each other,

118
A. McCracken, P. Sadeghian Journal of Building Engineering 20 (2018) 114–122

Table 4
Flexural and shear stiffness of cardboard sandwiches.
Cardboard Flexural stiffness, D (N m2) Shear Shear
flute stiffness, U modulus, G
ASTM Moment- (kN) (MPa)
method curvature
slope

B 127.1 215.3 192.4 127.0

C 242.9 223.3 182.9 121.9
BC 197.0 191.9 271.6 194.9

span length (i.e. two equations for each flute), the two parameters of D
and U can be calculated [5,29]. It should be highlighted that the ASTM
methods are limited to sandwich beams exhibiting linear force-deflec-
tion response. As a result, the initial linear region of load-deflection
curves was used to determine the initial stiffness K of each specimen.
Table 4 shows the calculated values for D and U as well as an experi-
mental value of D based on moment-curvature behavior. It should be
noted that the flexural stiffness based on moment-curvature presented
in Table 4 is the average from the short and long specimens.

3.6. Core shear modulus

Based on the shear stiffness U, the shear modulus G of each core can
be calculated as follows:

U (h−2t )
G=
(h − t )2b (2)

where b is the width of the sandwich beam, t is the skin thickness, and h
Fig. 9. Load-deflection behavior of specimens: (a) 150 mm span; and (b)
is the height of the entire sandwich beam. All dimensions have units of
300 mm span (note: each curve is the average of three identical specimens).
mm for further calculations. As shown in Table 3, the shear modulus of
cores with B, C, and BC flutes was calculated 127, 122, and 195 MPa.
however the specimens with BC flute failed at much lower load than Sadeghian et al. [5] previously studied the properties of sandwich
those with B and C flutes. The conventional deflection limit of span/180 beams manufactured with GFRP skins and 25.75 mm thick poly-
was drawn with a vertical line in the graphs of Fig. 8 to ensure the propylene (PP) honeycomb core with the bulk density of 91 kg/m3
specimens are not in a significant non-linear region at service loads. It before applying resin. The PP honeycomb was shown to have the core
also indicates that the ultimate capacity of the specimens is well beyond shear modulus G of 11.4 MPa. In this study, the cardboard B, C, and BC
of the service load. The load-strain and load-deflection curves are, of flutes with the bulk density of 170, 127, and 138 kg/m3 (before ap-
course, linear up to a point. However, a non-linearity shows up after the plying resin) displayed the shear modulus of 127, 122, and 195 MPa;
linear portion. The main source of non-linearity in the material non- respectively. The results indicate that the corrugated cardboard cores
linearity of flax FRP facings in tension and compression, which will be used in this study had an average shear modulus of almost 13 times that
discussed more in-depth in the analytical section. Overall, the speci- of the PP honeycomb, however the average bulk density of the card-
mens with B and C flutes meet the criteria and those with C flute are the boards was only 60% more than that of the PP honeycomb. The higher
most promising configuration for further studies. Per Table 2, C flute stiffness of the corrugated cardboard core may be due to the fact that it
has the lowest bulk density amongst all three flutes. This means C flute is stretch-dominated and the honeycomb core is bending-dominated. It
is the best configuration with the lowest weight and the highest has been shown that cellular cores can deform by either the bending or
strength and stiffness. It should be noted that C flute is the most stretching of the cell walls [30]. In bending-dominated cells, like hon-
common cardboard in the packing industry, which makes it the best eycombs, the walls of a cell bend due to in-plane loading, however in
option for sandwich applications. stretching-dominated cells, like the corrugated cardboards, cells are
triangular shape and the walls of a cell form a strut-tie system stretching
3.5. Flexural and shear stiffness the walls in tension and compression.

Per ASTM D7250 [29], by comparing the initial stiffness of two span 4. Analytical study
lengths of each group of specimens, the flexural stiffness D and the
shear stiffness U can be calculated by the equation below where K is 4.1. Theoretical flexural stiffness
initial stiffness in N/mm, S is the span length in mm, and L is the
loading span in mm as follows [5]: A theoretical calculation of the flexural stiffness of the sandwich
(2Si3−3Si Li2 + Li3) (Si − Li ) beams tested earlier is presented in this section. It is assumed that all
Ki + Ki =1 three layers are perfectly bonded together so the sandwich beam acts
96D 4U (1)
under full-composite behavior. The cross-section has a width b and total
where i = 1 denotes the parameters to the short-span specimens thickness h. Each skin has thickness t and the two skins are separated by
(S = 150 mm), and i = 2 to the long-span specimens (S = 300 mm). a relatively thick core of thickness c. Therefore, its flexural stiffness D is
The first term in the equation is related to flexural deformation and the the sum of the flexural stiffness of both skins and the core, measured
second term to shear deformation. With having an equation for each about the centroidal axis of the cross-section as follows [31,32]:

119
A. McCracken, P. Sadeghian Journal of Building Engineering 20 (2018) 114–122

Fig. 10. Comparison of theoretical and experimental flexural stiffness of Fig. 11. Comparison of theoretical flexural stiffness of sandwich beams with
sandwich beams with cardboard B, C, and BC flutes. cardboard C flute (150 mm wide and 75 mm thick core) with experimental
results of similar foam cores and flax FRP skins.
bt 3 btd2 bc3
D = Ef + Ef + Ec
6 2 12 (3)

where Ef and Ec are the modulus of elasticity of skin and core, respec-
tively, and d is the distance between the center lines of the upper and
lower skins. The theoretical flexural stiffness of each sandwich group
with different cardboard flute was calculated and presented in Fig. 10.
The theoretical values are compared with the experimental values ob-
tained using both ASTM method and moment-curvature method. As the
ASTM method uses deflections and the moment-curvature method uses
strains, they do not necessarily result in the same values. The theore-
tical method is also independent of the experimental values. The figure
indicates that the theoretical method can predict the flexural stiffness of
the sandwich beams made of the corrugated cardboard core and flax
FRP skins. The agreement also indicates that shear deformation of the
sandwich specimens in comparison with flexural deformation is negli-
gible. As the shear deformation of typical sandwich composites with
soft cores has negative effect on structural properties, failure load, and Fig. 12. Proposed parabolic stress-strain behaviour of flax FRPs in comparison
design criteria; having a lightweight core such as the cardboards stu- with coupon test data adopted from Betts et al. [28].
died in this research with limited negative effect of soft cores is a very
promising option for strong, stiff, and lightweight sandwich structures. multiple studies [5–7,28]. In this study, a parabolic stress-strain curve
for flax FRPs in tension is proposed. The model is verified against tests
4.2. Parametric study data presented in Fig. 12. With the assumption of linear and symmetric
strain profile with respect to the centroidal axis, the strain of top and
In this section, a parametric study is performed to evaluate the bottom skins is predicted at any given load. Fig. 13 compares the top
structural properties of large-scale sandwich beams made of the cor- and bottom skin strains obtained from proposed non-linear model and
rugated cardboard core and flax FRP skins. The second author has test data of long and short specimens made of C flute cardboard sheets.
previously used foam cores with bidirectional flax FRP skins [7]. The Also, a linear model based on initial elastic modulus of flax FRPs are
sandwich beams are considered with 75 mm thick and 150 mm wide presented in the figure.
cores. The flax FRP skins are made of 1, 2, and 3 layers of the bidir- As shown in Fig. 13(a), both linear and non-linear models can
ectional flax FRP with an elastic modulus of 7500 MPa. Fig. 11 presents predict the linear regions of the load-strain curves, however after a load
the comparison of the theoretical flexural stiffness of sandwich beams about 30% of the peak load, only the non-linear model is able to predict
with C flute (150 mm wide and 75 mm thick core) with the experi- the experimental behavior. The non-linear model can predict the tensile
mental results of similar sandwich beams made of foam cores and flax strain at bottom skin very well, however the experimental compressive
FRP skins. The figure indicates that the potentials of the cardboard strain at top skin deviates from the non-linear model at a load about
cores. The next stage of the current research program will be testing of 60% of the peak load. This might be explained by a higher level of non-
the large-scale sandwich beams with cardboard cores under static and linearity if flax FRPs in compression. This is compatible with the fact
impact loading. The results will be used for development of a design that the sandwich specimens failed under crushing of top skins.
procedure for cardboard core sandwich panels for walls and roofs of Fig. 13(b) shows the same trend for short specimens, however both
building under static loadings and dynamic effects due to the impact of linear and non-linear models slightly under predict the initial stiffness
a flying object during high wind, hurricane, and tornado. of the sandwich beams. This might be related to the fact that very short
specimens do not behave according to the ordinary bending theory. For
future studies, a higher-order beam theory is recommended to capture
4.3. Non-linear modeling the actual behavior sandwich composites made of cardboard core and
flax FRP skins.
In this section, a non-linear cross-sectional analysis is presented to
predict the load-strain behavior of the sandwich specimens. The non-
linearity is a material non-linearity of flax FRP skins as shown in

120
A. McCracken, P. Sadeghian Journal of Building Engineering 20 (2018) 114–122

those with B and BC flutes.

• C Flute with the lowest bulk density and the highest availability in
the market amongst all three flutes was selected as the best con-
figuration with the highest strength and stiffness for sandwich ap-
plications.
• The corrugated cardboard cores used in this study showed an
average shear modulus of almost 13 times that of a polypropylene
honeycomb with the same thickness and only 40% lower bulk
density.
• The analytical model based on full-composite action was able to
predict the flexural stiffness of the sandwich beams made of the
corrugated cardboard core and flax FRP skins with a very good
agreement with the test results. It indicated that the shear de-
formation of the core was negligible indicating that the cardboards
are a viable option for strong, stiff, and lightweight sandwich
structures.
• A non-linear cross-sectional model was developed based on non-
linearity of flax FRPs. The model was able to predict the non-line-
arity of load-strain curves of bottom skin obtained from the ex-
perimental program. A more in-depth modeling considering non-
linearity of flax FRPs in compression and higher-order beam theory
is recommended for the future.
• Although more research must be conducted on large-scale speci-
mens under static and dynamic loading, the all-natural flax FRP and
corrugated cardboard sandwich beams displayed encouraging
structural behavior and may prove to be a sustainable and structu-
rally efficient building material.

Acknowledgements

The authors of this paper would like to acknowledge the efforts of

Fig. 13. Comparison of top and bottom skin strains obtained from proposed Dillon Betts, MASc Student at Dalhousie University, who aided in the
non-linear model and test data: (a) long specimens; (b) short specimens made of manufacturing and provided guidance throughout the project as well as
C flute cardboard sheets. Dalhousie's technicians, Jesse Keane and Brian Kennedy, who helped
immensely with the setup and testing process. In addition, the authors
5. Conclusion acknowledge the National Science and Engineering Research Council of
Canada (NSERC) for the Undergraduate Student Research Award
In this study, three different flutes of corrugated cardboard cores, (USRA) for the first author, Dalhousie University for supplemental
namely B, C and BC flutes, and flax FRP skins and were used to man- funding through the Innovation-Themed Undergraduate Research
ufacture composite sandwich beams. A total of 30 small-scale sandwich Funding, and Maritime Paper (Dartmouth, NS, Canada) for providing
beam specimens were manufactured across six unique beam varieties the corrugated cardboard and adhesive.
with dimensions of 50 mm in width, 25 mm in depth, and 200 and
350 mm in length (150 mm and 300 mm spans) and tested under four- References
point bending up to failure. The load-deflection behavior, load-strain
[1] E.M. Reis, S.H. Rizkalla, Material characteristics of 3-D FRP sandwich panels,
behavior, and moment-curvature behavior of the sandwich beam spe-
Constr. Build. Mater. 22 (6) (2008) 1009–1018.
cimens were employed to study the effect of each cardboard flute to the [2] T. Sharaf, W. Shawkat, A. Fam, Structural performance of sandwich wall panels
strength and stiffness of the system. The flexural and shear stiffness of with different foam core densities in one-way bending, J. Compos. Mater. 44 (19)
the sandwich beams were quantified and the results were used to cal- (2010) 2249–2263.
[3] A. Manalo, T. Aravinthan, A. Fam, B. Benmokrane, State-of-the-art review on FRP
culate the shear modulus of the cardboards. Also, an analytical study sandwich systems for lightweight civil infrastructure, J. Compos. Constr. 21 (1)
was successfully implemented to predict the flexural stiffness of the (2016) 04016068.
sandwich specimens. The following conclusions can be drawn from this [4] M.A. Dweib, B. Hu, A. O’donnell, H.W. Shenton, R.P. Wool, All natural composite
sandwich beams for structural applications, Compos. Struct. 63 (2) (2004) 147–157.
study: [5] P. Sadeghian, D. Hristozov, L. Wroblewski, Experimental and analytical behavior of
sandwich composite beams: comparison of natural and synthetic materials, J.
• Two failure modes of transverse indentation and longitudinal Sandw. Struct. Mater. 20 (3) (2018) 287–307.
[6] L. CoDyre, K. Mak, A. Fam, Flexural and axial behaviour of sandwich panels with
crushing of the core were observed. All short specimens failed by bio-based flax fibre-reinforced polymer skins and various foam core densities, J.
core indentation, which was followed by indentation of the top skin. Sandw. Struct. Mater. (2016) (Dec 5:1099636216667658).
However, the long specimens reached a significantly lower peak [7] D. Betts, P. Sadeghian, A. Fam, Structural Behaviour of Sandwich Panels
Constructed of Foam Cores and Flax FRP Skins, Canadian Society for Civil
load and failed by longitudinal crushing of the core due to com-
Engineering (CSCE) Annual Conference, Vancouver, BC, Canada, 2017.
pressive bending stresses, which caused the top skin to be crushed [8] H. Mathieson, A. Fam, Static and fatigue behavior of sandwich panels with GFRP
after the peak load. skins and governed by soft-core shear failure, J. Compos. Constr. 18 (2) (2013)

• B and C flutes performed much better than BC flute in terms of load 04013046.
[9] T.C. Triantafillou, L.J. Gibson, Failure mode maps for foam core sandwich beams,
carrying capacity. Also, the B and C flutes showed enough strength Mater. Sci. Eng. 95 (1987) 37–53.
and stiffness in shear to develop a significant level of tensile and [10] A. Petras, M.P. Sutcliffe, Failure mode maps for honeycomb sandwich panels,
compressive strains close to those of flax FRP skins. The specimens Compos. Struct. 44 (4) (1999) 237–252.
[11] P. Wambua, J. Ivens, I. Verpoest, Natural fibres: can they replace glass in fibre
with flute C exhibited a higher flexural stiffness and strength than reinforced plastics? Compos. Sci. Technol. 63 (9) (2003) 1259–1264.

121
A. McCracken, P. Sadeghian Journal of Building Engineering 20 (2018) 114–122

[12] L. Yan, N. Chouw, K. Jayaraman, Flax fibre and its composites–a review, Compos. applications, Compos. Struct. 52 (3) (2001) 441–452.
Part B: Eng. 56 (2014) 296–317. [23] Z. Aboura, N. Talbi, S. Allaoui, M.L. Benzeggagh, Elastic behavior of corrugated
[13] D. Hristozov, L. Wroblewski, P. Sadeghian, Long-term tensile properties of natural cardboard: experiments and modeling, Compos. Struct. 63 (1) (2004) 53–62.
fibre-reinforced polymer composites: comparison of flax and glass fibres, Compos. [24] A. von der Heyden, J. Lange Assessment of the Utilisation of Corrugated Cardboard
Part B: Eng. 95 (2016) 82–95. as a Core Material for Sandwich Panels. ce/papers, 2017 Sep 1,1(2–3), pp.
[14] D. Ray, B.K. Sarkar, R.K. Basak, A.K. Rana, Thermal behavior of vinyl ester resin 1716–1725.
matrix composites reinforced with alkali‐treated jute fibers, J. Appl. Polym. Sci. 94 [25] Paper and Paperboard Packaging Environmental Council. Factsheet 03-2017.
(1) (2004) 123–129. 〈http://www.ppec-paper.com/pdfFiles/factsheets/2017/Packaging/FS03-2017.
[15] O. Faruk, A.K. Bledzki, H.P. Fink, M. Sain, Biocomposites reinforced with natural pdf〉.
fibers: 2000–2010, Progress. Polym. Sci. 37 (11) (2012) 1552–1596. [26] E. Kandare, P. Luangtriratana, B.K. Kandola, Fire reaction properties of flax/epoxy
[16] L. Wroblewski, D. Hristozov, P. Sadeghian, Durability of bond between concrete laminates and their balsa-core sandwich composites with or without fire protection,
beams and FRP composites made of flax and glass fibers, Constr. Build. Mater. 126 Compos. Part B: Eng. 56 (2014) 602–610.
(2016) 800–811. [27] M. Osei-Antwi, J. De Castro, A.P. Vassilopoulos, T. Keller, Shear mechanical char-
[17] A.K. Mohanty, M. Misra, L.T. Drzal, Sustainable bio-composites from renewable acterization of balsa wood as core material of composite sandwich panels, Constr.
resources: opportunities and challenges in the green materials world, J. Polym. Build. Mater. 41 (2013) 231–238.
Environ. 10 (1) (2002) 19–26. [28] D. Betts, P. Sadeghian, A. Fam Tensile properties of flax FRP composites, in:
[18] C. McSwiggan, A. Fam, Bio-based resins for flexural strengthening of reinforced Proceedings of the 6th Asia-Pacific Conference on FRP in Structures (APFIS 2017)
concrete beams with FRP sheets, Constr. Build. Mater. 131 (2017) 618–629. 2017 Jul 19., Singapore.
[19] M.L. Loong, D. Cree, Enhancement of mechanical properties of bio-resin epoxy/flax [29] ASTM D7250, Standard Practice for Determining Sandwich Beam Flexural and
fiber composites using acetic anhydride, J. Polym. Environ. 26 (1) (2018) 224–234. Shear Stiffness, ASTM International, West Conshohocken, PA, USA, 2012.
[20] N. Gupta, E. Woldesenbet, Characterization of flexural properties of syntactic foam [30] V.S. Deshpande, M.F. Ashby, N.A. Fleck, Foam topology: bending versus stretching
core sandwich composites and effect of density variation, J. Compos. Mater. 39 (24) dominated architectures, Acta Mater. 49 (6) (2001) 1035–1040.
(2005) 2197–2212. [31] A. Zinno, C.E. Bakis, A. Prota Mechanical characterization and structural behavior
[21] H.N. Wadley, N.A. Fleck, A.G. Evans, Fabrication and structural performance of of composite sandwich structures for train applications, in: Proceedings of the 8th
periodic cellular metal sandwich structures, Compos. Sci. Technol. 63 (16) (2003) International Conference on Sandwich Structures (ICSS8), Berlin: Springer, 2008
2331–2343. May 6, Vol. 1, pp. 570–581.
[22] J.F. Davalos, P. Qiao, X.F. Xu, J. Robinson, K.E. Barth, Modeling and character- [32] H.G. Allen, Analysis and Design of Structural Sandwich Panels, Pergamon Press,
ization of fiber-reinforced plastic honeycomb sandwich panels for highway bridge Oxford, UK, 1969.

122