A Case Study of Biocomposite Material Use in Automotive Applications

Abstract

Interest in biocomposites is growing worldwide as companies that manufacture high-performance products seek out more sustainable material options. Although there is significant research on biocomposite material options and processing found in the literature from at least the last two decades, there are few experimentally based case studies published to help guide product designers and engineers when considering these materials. This paper discusses the use of biocomposites in the seat of an electric bus. Although it is clear that biocomposite material options are quite limited, the authors eventually settled on three natural reinforcements (cellulose, hemp, flax), two epoxies (one low and the other high viscosity) with high biobased carbon content, and one flax precoated with bioepoxy for consideration. Laminate plates with a 4mm nominal thickness are manufactured using VARTM (low viscosity epoxy only), hand layup as a surrogate for prepregging (high viscosity epoxy only), compression molding, and an out-of-autoclave process called the Pressure Focusing Layer (PFL) method. Permeability of the three reinforcements infused with the high viscosity epoxy and fiber volume fractions are determined experimentally to provide insight into VARTM processing and mechanical performance. The tensile modulus, maximum tensile stress, flexural modulus, and maximum flexural stress are measured for all combinations of reinforcement, resin, and processing using tension testing and three-point bending based on ASTM standards. Basic conclusions are drawn about the specific application and more generally about the process of using biocomposites in commercial products.

Full text

A Case Study of Biocomposite Material
Use in Automotive Applications
DANIEL WALCZYK, RONALD BUCINELL,
STEVEN FLEISHMAN and SHARMAD JOSHI
ABSTRACT1
Interest in biocomposites is growing worldwide as companies that manufacture
high-performance products seek out more sustainable material options. Although there
is significant research on biocomposite material options and processing found in the
literature from at least the last two decades, there are few experimentally based case
studies published to help guide product designers and engineers when considering these
materials. This paper discusses the use of biocomposites in the seat of an electric bus.
Although it is clear that biocomposite material options are quite limited, the authors
eventually settled on three natural reinforcements (cellulose, hemp, flax), two epoxies
(one low and the other high viscosity) with high biobased carbon content, and one flax
precoated with bioepoxy for consideration. Laminate plates with a 4mm nominal
thickness are manufactured using VARTM (low viscosity epoxy only), hand layup as a
surrogate for prepregging (high viscosity epoxy only), compression molding, and an
out-of-autoclave process called the Pressure Focusing Layer (PFL) method.
Permeability of the three reinforcements infused with the high viscosity epoxy and fiber
volume fractions are determined experimentally to provide insight into VARTM
processing and mechanical performance. The tensile modulus, maximum tensile stress,
flexural modulus, and maximum flexural stress are measured for all combinations of
reinforcement, resin, and processing using tension testing and three-point bending based
on ASTM standards. Basic conclusions are drawn about the specific application and
more generally about the process of using biocomposites in commercial products.
Daniel Walczyk, Rensselaer Polytechnic Institute, 110 8th Street, Troy, NY 12180, U.S.A.
Ronald Bucinell, Union College, 807 Union Street, Schenectady, NY 12308, U.S.A.
Steven Fleishman, Light Green Machines, LLC, 7 Stormy View Road, Ithaca, NY 14850, U.S.A.
Sharmad Joshi, Rensselaer Polytechnic Institute, 110 8th Street, Troy, NY 12180, U.S.A.
1645

INTRODUCTION
Despite the global economic recession resulting from the COVID-19 pandemic, the
composite materials market is still expected to reach nearly $40 billion (USD) by 2025
[1]. The growing popularity of advanced composites stems from their inherent
advantages compared to traditional engineering materials including (but not limited to)
high specific stiffness and strength, durability, long service life, dimensional stability,
and tailorable properties and ease of part consolidation that expand design options.
However, one of the main disadvantages with synthetic composites (carbon and glass
fiber, fossil-fuel-based thermoset and thermoplastic resins) is that their end-of-life
options are quite limited, since recycling these materials is difficult, if not impossible
and processing can be energy intensive.
In response, numerous researchers, established manufacturers, and startups have for
at least the last two decades investigated more sustainable composite materials made
from renewable feedstocks, so called biocomposites, that are comprised of bio-based
fiber reinforcement and resins, i.e. biofibers and bioresins, respectively. One of the main
reasons for this interest is that some biofibers (flax, hemp, silk) have specific tensile
strengths and Young’s moduli that rival E-glass [2]. There has been significant new
research and development results on biocomposites published in 100’s of articles in
many respected academic journals [3-19], and the corresponding market is expected to
reach $8.5 billion (USD) by 2025 [20]. Unfortunately, any designer/engineer who wants
to use biocomposites in a product typically lacks familiarity with these materials and
how they are processed, and has limited commercial options for biofiber and bioresin
material systems. The literature has some case studies that describe the process of
designing automotive components (e.g., [21-22]) and aerospace components (e.g., [23],
although they are primarily analytically based.
In this paper, the authors describe a realistic design scenario where biocomposites
are considered for the seat bottom used in an electric bus and assessed experimentally.
The details of material and manufacturing process selection, and mechanical
performance testing are provided along with guidelines that designers and engineers can
use when using biocomposites for structural applications.
DESIGN AND MANUFACTURING REQUIREMENTS
As previously mentioned, the part shape considered in this study is the seat bottom
for a new electric bus being designed and prototyped by Light Green Machines, LLC
(Ithaca, NY). Since the actual part shape is relatively generic, the surrogate shape used
for the material assessment study was a 30.5×30.5cm square flat laminate to simplify
tooling fabrication. Certain aspects of the design and manufacturing requirements were
emphasized by the company including:
 4 mm target thickness
 Domestically sourced biofiber reinforcement
 Domestically sourced thermoset resin with as high a biobased carbon content as
possible
 Short cycle time for production, ideally less than 10 min/part
 Separately stored resin and reinforcement, if possible, to minimize material and
storage costs
1646

 High fiber volume fraction (Vf) and
 At least one surface with A-side smoothness.
Engineering requirements for structural rigidity and strength were not considered
initially. This is typical of a Design for Manufacture situation where product design and
manufacturing details are developed in parallel.
MATERIALS, EXPERIMENTAL HARDWARE AND PLAN
Materials: Two thermoset resin systems with high biobased carbon content and
three different fiber textiles were used for this study. The two resins are made by
Sicomin Epoxy Systems (France):
 SR GreenPoxy 33 resin (35% biocontent) mixed with SZ4775 hardener in a
100/27 weight ratio as prescribed by the manufacturer (GreenPoxy)
 SR InfuGreen 810 resin (38% biocontent) mixed with SC8824 hardener in a
100/22 weight ratio as prescribed by the manufacturer (InfuGreen).
Greenpoxy is relatively viscous resin intended for wet layup and compression molding,
whereas InfuGreen has much lower viscosity and is intended for resin infusion
processes. The fiber textiles include filament derived from wood and two bast fibers:
 Biomid continuous filaments made of highly crystalline cellulose in a 407 g/m2
(12 oz/yd2) plain weave by Absecon Mills, Inc. in Cologne, NJ (Cellulose)
 Hemp fiber filaments in an 814 g/m2 (24 oz/yd2) plain weave made by Hem
Mills, Inc. in Concord, NC (Hemp)
 EKOA flax fiber in a 366 g/m2 (10.8 oz/yd2) 2x2 twill weave made by Lingrove,
Inc. in San Francisco, CA (Flax).
In addition, EKOA P-SX 11.8 unidirectional flax fiber made by Lingrove, Inc. that has
a 400 g/m2 (11.8 oz/yd2) areal weight and is precoated, not impregnated, with a
proprietary epoxy resin was used (Precoated Flax). Since biocomposite testing included
three-point bending and tension, layups for the cellulose, hemp, and flax textiles, and
the flax precoated were [016], [04], [04], and [04], respectively, so that the preform
thickness was approximately 4 mm. A 0° direction in this case is the warp direction of
the textile. Each of the biocomposite systems evaluated (i.e. combination of textile,
resin, and consolidation/curing method) were fabricated into square plates with 30.5cm
on a side, as previously mentioned.
Plate Manufacturing: Two different processes were used to impregnate dry preform
textile layups for this study: (1) wet layup for the Greenpoxy resin and (2) vacuum-
assisted resin transfer molding (VARTM) for the InfuGreen resin. Wet layup is not
typically a viable process for impregnating composite in a high production setting
because it is slow and manually intensive. However, it was used for this study as a
surrogate process for prepregging, since there are so few biocomposite prepregs
currently available in the U.S. The process involved pouring sufficient GreenPoxy
mixed resin over individual plies, working the resin in using a standard squeegee,
stacking wet plies in the correct orientation, and debulking with the squeegee. For
example, a dry stack of Cellulose textile is shown in Figure 1a. The wet stack was
eventually bagged between two layers of high temperature release film sealed along the
edges with vacuum sealant tape. VARTM involved vacuum bagging a dry preform
stack with 2.5cm wide infusion flow media positioned along opposite edges of the stack
and attaching vacuum and resin inlet hoses, as shown in Figure 1b. Vacuum between
1647

80-85 kPa was applied, and resin was pulled from a mixing container. Layup of the
Precoated Flax involved allowing the material to reach room temperature, then stacking
individual plies while debulking after each layer (Figure 1c).
(a) (b)
(c)
Figure 1. (a) 16-ply stack of dry Cellulose textile and squeegee before wet layup, (b) 4-ply stack of dry
Hemp textile that is vacuum-bagged prior to resin infusion via VARTM, and (c) 4-ply stack of
Precoated Flax.
The formed and impregnated layups had to be: (1) consolidated to increase fiber
volume fraction (Vf) and decrease void content (Vv); and (2) cured under the appopriate
time vs. temperature conditions to sufficiently crosslink the resin matrix. The two
consolidation/curing methods investigated for this study were compression molding
(CM) between one rigid and one temperature-controlled tool (Figure 2a), and an out-of-
autoclave alternative called the Pressure Focusing Layer (PFL) process (Figure 2b) from
Vistex Composites [24]. The PFL process involves curing and consolidating a wet
composite layup between a temperature-controlled ‘curing mold’ and a specially
engineered, rubber-coated (40 Durometer A Silicone rubber) ‘compression mold.’ In
both cases, approximately 150 kPa of pressure was applied using a hydraulic press for
consolidation, and the curing mold was a temperature-controlled hot plate. Curing of
the GreenPoxy and InfuGreen laminates required 100°C for 0.5 hours and 80C for 4
hours, respectively, under pressure. The Precoated Flax required tool temperature to be
ramped from 20 to 100C over 1 hour to allow for pre-impregnation to occur, then
ramped up to 120C over 10 min, and finally held at that temperature for a minimum of
30 minutes to allow for full curing.
1648

(a) (b)
Figure 2. (a) Compression molding of 305mm square composite test specimen and (b) PFL molding of a
test specimen.
Test Specimen Preparation: Each plate was marked to identify the warp and weft
(fill) directions. Warp testing specimens were waterjet cut from the panels for tensile
(ASTM D3039) and bend (ASTM D790) testing. Figure 3a shows the typical cutting
pattern used for the tensile and bending specimens. The two diverging lines through the
center of the samples are in the weft (fill) direction and were used to identify the location
of the specimens with respect to the plate.
(a) (b)
Figure 3. (a) Tensile (larger) and bending (smaller) warp specimens cut from a typical composite plate
and (b) experimental configuration used to conduct tensile tests.
The following sections will summarize the tensile and bending experimental
evaluation of the material systems considered in this study.
Tensile Evaluation: ASTM standard D3039 was used as a guide in the evaluation of
the various material systems. The length, width, and thickness dimensions of each
tensile specimen measured prior to testing. The length was measured using a standard
scale (±0.5mm), the width was measured using calipers (±0.01mm), and the thickness
was measured with micrometers (±0.001mm). The specimens were approximately
30.5cm long × 25mm wide, while their thicknesses varied. The order of testing for the
specimens was determined using a Latin Square Randomized Block design [25]. This
strategy was employed to eliminate any data biasing related to the test sequencing.
The tensile testing was conducted on the electro-mechanical load frame shown in
Figure 3b. All tests were conducted using a 45kN loadcell. The specimens were secured
1649

using wedge grips and the top wedge grip was attached to the load cell through a
universal joint. The tests were conducted under stroke (displacement) control and the
crosshead rate was set to 1.3mm/min for all tensile tests.
Both load and crosshead displacement were recorded at a rate of 10Hz by the load
frame controller. The load and crosshead displacement were sent to a Digital Image
Correlation (DIC) system that was being used to collect full field displacements on the
specimen surface at a rate of 1Hz. The DIC system utilized two cameras (stereo). A
55×44mm field of view was calibrated for these tests. After the data was collected, a
40mm virtual extensometer was placed at the center of the specimen in the axial
direction and a 15mm virtual extensometer was placed at the center of the specimen in
the transverse direction. These extensometers were used to calculate axial and transverse
strain as the specimens were loaded.
Axial stress (

) was computed by taking the load recorded by the loadcell (l) and
dividing it by the specimen width (w) multiplied by the specimen thickness (t):
𝜎

⋅
(1)
A typical axial Stress-Strain curve that resulted from using Eqn. 1 for stress and the
axial virtual extensometer is seen in Figure 4a. A least square’s fit of the data from
3-20MPa was used to compute the axial modulus (slope), which is E
x
= 6.89GPa for
this particular specimen. The initial linear range used to compute the axial elastic
modulus varied with each specimen tested.
The transverse virtual extensometer and the stress computed using Eqn. 1 were used
to construct the transverse stress-strain curve, with a typical one shown seen in Figure
4b. Again, a least square’s fit of the data between 3-20MPa was used to compute the
slope of the curve in this region, which is M = 29.7GPa. This slope and the modulus
from before were used to compute Poisson’s Ratio (

xy
):
𝜈






. GPa
. GPa
0.232 (2)
The average elastic modulus, maximum tensile stress, and Poisson’s ratio for each
material system tested is summarized in the next section.
(a)
1650

(b)
Figure 4. (a) Axial stress-strain and (b) transverse stress-strain curves for a tensile specimen used in
this study.
Bending Evaluation: ASTM standard D790 was used as a guide in the evaluation of
the various material systems in bending. The length, width, and thickness dimensions
of each tensile specimen were measured prior to testing using the same instruments as
with the tensile evaluation. The specimens were approximately 15cm long × 12 mm
wide, while their thicknesses varied. The order of testing for the specimens was
determined using a Latin Square Randomized Block design similar to the tensile
evaluation.
The bend testing was conducted using a three-point bending configuration and the
electro-mechanical load frame shown in Figure 5a. All tests were conducted using a
1.1kN loadcell. The lower supports were set at a span of 6.1cm, which corresponds to
at least a 16:1 span-to-average thickness ratio, and the upper support was centered on
the span. An extensometer was placed directly under the center span to record beam
deflection. The tests were conducted under stroke (displacement) control and the
crosshead rate was set to 1.5mm/min for all bend tests. The terminal deflection was
determined to be 7.6mm for all specimens. The load, crosshead displacement, and
extensometer displacement were recorded by the load frame at a rate of 10Hz.
(a)
1651

(b)
Figure 5. (a) Three-point bending configuration used to test the samples in this study and (b) outer
fiber stress vs. outer fiber strain plot from data generated in the three-point bending tests.
The recorded load data was converted into outer fiber stress (

) using the following
ASTM D790 equation: 𝜎
⋅⋅
⋅⋅

(3)
where P is the load recorded from the load cell, L is the span between supports, w is the
width of the specimen, and t is the specimen thickness. The outer fiber strain (ε) is
calculated from the extensometer deflection (D) using another ASTM D790 equation:
𝜀
⋅⋅


(4)
The outer fiber stress (

) from Eqn. 3 is plotted versus the outer fiber strain (ε) from
Eqn. 4, and a typical curve is found in Figure 5b. The Flexural Modulus (E
f
) for this
specimen is calculated by performing a Least Squares Linear Regression through the
initial linear portion of the curve and determining the slope of the fit line. For the linear
portion of approximately 10-50 MPa in Figure 5b, E
f
= 11.1MPa. A summary of the
average bending modulus and maximum stress in the outer fibers for all the samples in
this study is found in the next section.
Experimental Plan: As summarized in TABLE I, six specimens each for tensile and
three-point bending evaluation (Figure 3a) were tested for each combination of textile,
resin, and consolidation/curing method. In addition, permeability for all of the
reinforcement types infused with InfuGreen and GreenPoxy were measured to better
understand the VARTM processing characteristics.
TABLE I. BIOCOMPOSITE PLATES MANUFACTURED AND TESTED FOR THIS STUDY
Plate
#
Reinforcement Number of
Plies, n
Resin/Impregnation
Process
Consolidation/Curing
Process
1
Cellulose 16
GreenPoxy/Hand Layup CM
2 PFL
3 InfuGreen/VARTM CM
4 PFL
5
Hemp 4
GreenPoxy/Hand Layup CM
6 PFL
7 InfuGreen/VARTM CM
8 PFL
1652

9
Flax 4
GreenPoxy/Hand Layup CM
10 PFL
11 InfuGreen/VARTM CM
12 PFL
13 Precoated CM
14 PFL
To measure permeability for the reinforcement, resin, and layup combinations in
this study, rectangular strips of reinforcement were laid up as indicated in TABLE II
(5.112.7cm for the more viscous GreenPoxy and 5.125.4cm for the less viscous
InfuGreen). After flow mesh was added to the beginning and end of the layups to
facilitate uniform vacuum and resin distribution across the layup width, they were
vacuum bagged and fitted with inlet and outlet hoses. Vacuum sealant tape was used to
create all vacuum seals. The outlet hose was connected to a vacuum pump, while the
inlet hose’s free end was immersed in a container of either InfuGreen or GreenPoxy
two-part resin mixed in their proper resin-to-hardener ratio. One centimeter intervals
were marked on the top vacuum bag to indicate x position during infusion, and the time
at which each interval is reached was recorded.
TABLE II. EXPERIMENTAL PARAMETERS FOR EACH FIBER/RESIN COMBINATION
Reinforcement Layup Width, w
(cm)
Length, L (cm) p
(kPa)
µ (Pas)
InfuGreen GreenPoxy InfuGreen GreenPoxy
Hemp [04]
5.1 25.4 12.7 84 0.2 1.3
Cellulose [016]
Flax [04]
PROCESS MODELING
Estimating Fiber Volume Fraction: Fiber volume fraction, Vf, of a bicomposite is a
critical material parameter for understanding the role of reinforcement and resin in
overall mechanical properties. Unfortunately, using standard digestion, ignition, or
carbonization (ASTM D2584-11, ASTM D3171-15 Test Method I) is impractical, since
the fibers undergo complete decomposition along with the resin in all of the methods
[26]. Hence, Vf of a flat biocomposite laminate can be estimated using ASTM D3171-
15 Test Method II based on geometry and properties of the dry reinforcement plies and
fiber and by assuming that the void volume fraction Vv = 0. The fiber mass within the
laminate is simply 𝑚𝑛𝜌𝐴, where n = number of plies in the laminate preform,

ar = fabric mass/fabric area = areal density of the reinforcement fabric, and A =
laminate area. The fiber volume in the laminate is then found by 𝑚𝜌
𝑛𝜌𝐴𝜌
,
where

f = density of the solid fiber. The total laminate volume is ℎ𝐴, where h =
laminate thickness. Finally, the fiber volume fraction is found by dividing the fiber
volume by the total volume to get:
𝑉

 
. (5)
1653

Vacuum Infusion Study: To characterize the permeabilities of all fiber
reinforcements that would be impregnated using VARTM and better understand the
infusion times, a vacuum infusion study was conducted. The prevailing theory used for
modeling infusion of liquid resin through reinforcement fibers is Darcy’s Law. In its
original form, this law describes the one-dimensional flow rate (Q) of a fluid through a
porous medium, as shown in Figure 6, by:
𝑄𝑣𝐴󰇛𝑑𝑥 𝑑𝑡
⁄󰇜
𝐴𝑘𝐴𝑝 𝜂𝑥
 , (6)
where 𝑣𝑑𝑥 𝑑𝑡
⁄ is the flow front velocity, K = in-plane permeability of the fiber
reinforcement preforms, A = the constant cross-sectional area of the flow, µ = viscosity
of the resin, and


= the pressure drop over some distance. Rearranging Eqn. 6 and
solving the differential equation for time yields a relationship that can be used to
approximate elapsed time (t) for a resin flow front to move an impregnation depth
distance (x) within a preform due to a differential pressure (p) as:
𝑡𝑑𝑡


󰇡𝜇𝑥 𝐾𝑝
󰇢𝑑𝑥


µ𝑥

2𝐾𝑝
. (7)
Eqn. 7 can be rearranged to isolate permeability as:
𝐾



. (8)
If µ and p are held constant, the value of K for a particular reinforcement, preform layup,
and resin is found simply by dividing the slope of the line for a plot of x
2
vs. t by the
factor 2𝑝 𝜇
.
Figure 6. Geometry with process parameters indicated for Darcy’s Law applied to one-dimensional
flow.
EXPERIMENTAL RESULTS AND DISCUSSION
Measuring Thickness and Estimating V
f
: The average value of thickness, h, of each
plate measured for 10 specimens using a digital caliper is provided in TABLE III. The
target thickness of 4 mm was achieved with four hemp plies. Additional plies would be
required for Cellulose (2-3), Flax (only 1), and Flax Precoated (2-3) to reach this
thickness.
1654

Fiber volume fractions are estimated for the 14 plates in Table 3 based on Eqn. 5,
where

ar is provided by the manufacturer,

f is found in Ref. [27], and h measured as
previously discussed. Values of Vf for both the Cellulose and Hemp textiles range from
0.55 to 0.64, which are typical for advanced composite applications. The Flax textile Vf
values are surprisingly low (0.31-0.34), specifically about half that of Cellulose and
Hemp, whereas the Precoated Flax values are in between these two extremes.
Table III. ESTIMATES OF Vf FOR ALL BIOCOMPOSITE PLATES
Plate # n

ar
(kg/m3)

f
(kg/m3)
Measured h Vf
mm m
1 16 0.203 1500 3.57 0.00357 0.607
2 16 0.203 1500 3.53 0.00353 0.613
3 16 0.203 1500 3.37 0.00337 0.643
4 16 0.203 1500 3.33 0.00333 0.650
5 4 0.814 1450 3.73 0.00373 0.602
6 4 0.814 1450 4.03 0.00403 0.557
7 4 0.814 1450 4.09 0.00409 0.549
8 4 0.814 1450 4.06 0.00406 0.553
9 4 0.367 1450 3.00 0.00300 0.337
10 4 0.367 1450 3.08 0.00308 0.329
11 4 0.367 1450 3.16 0.00316 0.320
12 4 0.367 1450 3.30 0.00330 0.307
13 4 0.400 1450 2.41 0.00241 0.458
14 4 0.400 1450 2.46 0.00246 0.449
Resin/Reinforcement Permeability: Examples of the experimental setup and graphs
of x2 vs. t for all six reinforcement/resin combinations (see TABLE I) are provided in
Figures 7 and 8, respectively. Approximate fill times and the calculated permeabilities
based on the Figure 8 data are provided in TABLE IV. Four plies of flax had the highest
permeability and lowest overall fill times. Hemp, which consisted of a denser weave
than flax, had longer fill times and about half the permeability. Cellulose was a very thin
weave and required 4X as many layers as both flax and hemp. Not surprisingly, the fill
times were nearly an order-of-magnitude higher than the other two and permeabilities
were nearly an order-of-magnitude lower. It should be noted that because cellulose fill
times were so long, the x2 vs. time curves for InfuGreen and GreenPoxy (Figures 4c and
4d) became non-linear beyond 17 min and 23 min, respectively. This is due to the resins
starting to exotherm in their mixing containers and thicken up due to partial cross-
linking. Only the linear portion of these curves were used in calculating Keff, since one
can assume that in-line mixing would be used for production.
1655

(a) (b) (c)
(d) (e) (f)
Figure 7. Resin infusion experimental runs for (a) hemp/InfuGreen, (b) hemp/GreenPoxy, (c)
cellulose/InfuGreen, (d) cellulose/GreenPoxy, (e) flax/InfuGreen, and (f) flax/GreenPoxy.
(a) (b)
(c) (d)
1656

(e) (f)
Figure 8. Plots of x
2
vs. time data and linear fits of permeability experiments for (a) Hemp/InfuGreen,
(b) Hemp/GreenPoxy, (c) Cellulose/InfuGreen, (d) Cellulose/GreenPoxy, (e) Flax/InfuGreen, and (f)
Flax/GreenPoxy.
TABLE 4. FILL TIMES AND EFFECTIVE PERMEABILITY FOR EACH FIBER/RESIN
COMBINATION
Reinforcement Layup
Fill Time (min) K
eff
(m
2
)
InfuGreen
(L=25.4cm)
GreenPoxy
(L=12.7cm)
InfuGreen GreenPoxy
Hemp [0
4
] 7 13 1.71

10
-10
1.4210
-10
Cellulose [0
16
] 36 120 4.17

10
-11
4.0710
-11
Flax [0
4
] 3 9 3.89

10
-10
2.4210
-10
Laminate Surface Texture: Images of surface textures for composite plates 1, 2, 5,
6, 9, 10, 13, and 14 after consolidation/curing and before cutting into text specimens are
provided in Figure 8. It should be noted that wrinkling on the surface of Plate 1 in Figure
8 is due to wrinkling in the release layer and is not indicative of what would happen
with real production tooling. Since the flax has a smaller thread count (i.e. larger tows
used in weave) and courser weave than both the hemp and cellulose, this resulted in the
highest surface roughness (measured qualitatively) and lowest permeability of all the
woven reinforcements (see Table 3). Compression molding provides equivalent A-side
and B-side surface textures, both smooth; whereas the Vistex PFL process provides a
smooth A-side and rough B-side surface texture due to the compliant rubber tooling
used.
1657

(1) (2) (5)
(6) (9) (10)
(13) (14)
Figure 8. Top surface texture close-ups for the composite plates 1, 2, 5, 6, 9, 10, 13, and 14.
Mechanical Properties: Six specimens from each of the 14 plates were tested in
tension and flexure to measure each material system’s inherent mechanical properties.
Raw data is shown in TABLE V. Bar graphs of elastic modulus, maximum tensile stress,
flexural modulus, and maximum flexural stress are provided in Figures 8a, 8b, 9a, and
9b, respectively.
Cellulose has significantly higher elastic modulus, roughly by 50%, compared to all
other reinforcement fibers. There is also a distinct benefit to compression molding over
PFL for cellulose only. Elastic modulus values for all other plate combinations are
around the same value, and consolidation process and choice of resin shows no clear
effect. Flax shows surprisingly high modulus even though its Vf is about half that of all
1658

other systems. One reason for cellulose’s higher modulus may be that the polymer
filaments are synthesized from wood, i.e. it is not a naturally occurring fiber.
Cellulose also exhibits significantly higher tensile stress at failure than all other fiber
types, while hemp is the next strongest. Flax strength was generally lower than either
cellulose or hemp, although its lower Vf must be taken into account.
Cellulose with GreenPoxy show roughly double the flexural modulus of all other
biocomposite material systems tested. All plates made with InfuGreen resin exhibited
significantly lower Ef than those with GreenPoxy suggesting that the former does not
bond as well to natural fibers for transferring shear during bending and/or it has lower
inherent stiffness than the latter. Flax had generally lower values as compared to hemp,
although this may be attributable to the lower Vf. Precoated Flax modulus was on par
with all other materials other than Cellulose/GreenPoxy.
Once again, Cellulose with GreenPoxy exhibited the highest flexural stress at
failure; roughly 2-3X higher than all other material combinations. InfuGreen generally
had lower values than GreenPoxy.
TABLE V. MECHANICAL PROPERTIES OF ALL COMPOSITE MATERIAL SYSTEMS
INVESTIGATED IN THIS STUDY
Plate #
Tension Test 3-Point Bending Test Vf
(from
Table 2)
Elastic
Modulus
(GPa)
Maximum
Stress
(MPa)
Poisson’s
Ratio
Flexural
Modulus
(GPa)
Maximum
Stress
(MPa)
1 17.1 159 0.13 23.8 350 0.607
2 15.8 153 0.14 27.4 374 0.613
3 15.7 152 0.21 9.1 185 0.643
4 14.8 149 0.16 11.4 204 0.650
5 9.7 106 0.27 14.7 229 0.602
6 10.2 119 0.28 15.1 226 0.557
7 10 116 0.32 11.0 174 0.549
8 9.5 110 0.32 10.1 165 0.553
9 10.1 98 0.14 9.3 127 0.337
10 10.2 92 0.14 10.6 145 0.329
11 9.7 94 0.15 6.4 100 0.320
12 10.9 108 0.15 5.0 84 0.307
13 9.6 93 0.11 8.7 126 0.458
14 9.4 71 0.21 4.6 80 0.449
1659

(a)
(b)
Figure 8. Average values of (a) elastic (tensile) modulus and (b) maximum tensile stress for all
biocomposite material systems tested in this study.
1660

(a)
(b)
Figure 9. Average values of (a) flexural modulus and (b) maximum flexural stress for all biocomposite
material systems tested in this study.
CONCLUSIONS
The section is organized into two parts – the first draws general conclusions about
the materials and processes considered for the transportation application and the second
provides an assessment of how well the company’s preliminary design/manufacturing
requirements were met or not.
Part 1: As mentioned in the introduction, biocomposites is a growing market in the
engineering material space, and there exists a wealth of knowledge, over two decades
worth, about material properties and processing of various natural fibers paired with
synthetic resins and bioresins. Despite this, finding commercially available fibers and
bioresins suitable for an advanced composite part and their material properties proved
to be a difficult task due to the limited number of material suppliers. Two different
epoxy resins with a moderate level of biobased carbon content and three woven textiles
(Cellulose, Hemp, Flax) derived from non-fossil-fuel-based resources were eventually
settled on for this study. Two methods of resin impregnation were investigated –
1661

VARTM and hand layup – although the latter was intended as a surrogate for using
prepreg. In addition, a commercially available unidirectional flax precoated with B-
staged epoxy was tested. All impregnated and debulked laminates were eventually
consolidated/cured using both compression molding and an out-of-autoclave alternative
called the PFL process.
Physical and mechanical properties of the biocomposite panels made varied
significantly. Fiber volume fraction for the Cellulose and Hemp was typical for
advanced composites (~60%), Flax was about half that (~30%), and Precoated Flax was
in between these two extremes. Flax had the highest permeability for resin infusion
(consistent with low Vf) followed by Hemp and Cellulose. In fact, Flax/InfuGreen is the
only material combination with the potential for a VARTM fill time of less than 10 min.
Surface finish resulting from either consolidation/curing method were generally
acceptable for this application, although each reinforcement provides distinctly different
surface textures and aesthetic qualities. Finally, Cellulose/GreenPoxy generally
exhibited the best mechanical performance, but Flax could have comparable or better
performance if a higher Vf were achieved.
Part 2: The challenge for a design engineer considering biocomposites is how to
choose a sustainable combination and commercially available combination of
reinforcement, bioresin, and processing that provides parts with known mechanical
properties and low variability. Assessment of the biocomposite material systems and
processing that were investigated are discussed below.
 4 mm target thickness – this target value can be achieved within a certain
tolerance, although it is vitally important that the composite layup have an even
number of plies and be symmetrical about the laminate midplane. Otherwise, an
unbalanced layup has unwanted bend/twist coupling effects.
 Domestically sourced biofiber reinforcement – there are a moderate number of
material suppliers that can provide biofiber reinforcement in flax, hemp, and
cellulose, although many are located in Europe. The hemp textile used in this
study is actually intended for furniture applications, not composites.
 Domestically sourced thermoset resin with as high a biobased carbon content as
possible – the best commercially available bioresins had <40% biocontent and
are manufactured in Europe.
 Short cycle time, ideally less than 10 min/part when the process reaches a MRL
7 level – the Flax/InfuGreen combination would allow infusion in less than 10
min, but the bending properties are lower and curing takes hours.
Cellulose/GreenPoxy has the best mechanical properties of all the material
combinations considered, although 4-5X more ply layup operations are
required. The best combination may be a Flax reinforcement with higher Vf
coupled with GreenPoxy mixed with a fast-cure hardener.
 Separate resin and reinforcement, if possible, to minimize material and storage
costs – separate resin and reinforcement requires the use of a resin infusion
method. VARTM is too slow, but High Pressure Resin Transfer Molding (HP-
RTM) would significantly reduce cycle time. The best approach investigated in
this study was the use of prepreg material (i.e. represented by hand layup
method) with a fast-cure resin.
 High fiber volume fraction (Vf) – fiber volume fractions of ~60% were achieved
with Cellulose and Hemp reinforcements.
1662

 At least one surface with A-side smoothness – both compression molding and
PFL provide at least one suitable A-side surface.
ACKNOWLEDGEMENTS
The authors would like to acknowledge the New York State Energy Research
Development Authority (NYSERDA) for sponsoring this research.
REFERENCES
1. 2020. Growth Opportunities in Global Composites Market. Lucintel, Aug. 2020.
2. Pickering, K.L., M.G. Aruan-Efendy, and T.M. Le. 2015. “A review of recent developments in
natural fibre composites and their mechanical performance,” Compos. Part A Appl. Sci. Manuf.,
83:98-112.
3. Academic Journal of Polymer Science. https://juniperpublishers.com/ajop/. Accessed on 15 June
2021.
4. Composites Part A: Applied Science and Manufacturing.
https://www.journals.elsevier.com/composites-part-a-applied-science-and-manufacturing. Accessed
on 15 June 2021.
5. Composites Part B: Engineering. https://www.journals.elsevier.com/composites-part-b-
engineering. Accessed on 15 June 2021.
6. International Journal of Biobased Plastics.
https://www.tandfonline.com/action/journalInformation?show=aimsScope&journalCode=tbbp20.
Accessed on 15 June 2021.
7. Journal of Applied Polymer Science. https://onlinelibrary.wiley.com/journal/10974628
8. Journal of Biobased Materials and Bioenergy. http://www.aspbs.com/jbmbe.html. Accessed on 15
June 2021.
9. Journal of Cleaner Production. https://www.journals.elsevier.com/journal-of-cleaner-production.
Accessed on 15 June 2021.
10. Journal of Composite Materials. https://journals.sagepub.com/home/jcm. Accessed on 15 June
2021.
11. Journal of Composites Science. https://www.mdpi.com/journal/jcs. Accessed on 15 June 2021.
12. Journal of Natural Fibers. https://www.tandfonline.com/loi/wjnf20. Accessed on 15 June 2021.
13. Journal of Polymers and the Environment. https://www.springer.com/journal/10924. Accessed on
15 June 2021.
14. Journal of Polymer Research.
https://www.springer.com/journal/10965?gclid=EAIaIQobChMIjbLB8ZXC8QIVAHFvBB2fwwF-
EAAYASAAEgJIXPD_BwE. Accessed on 15 June 2021.
15. Journal of Reinforced Plastics and Composites. https://journals.sagepub.com/home/jrp. Accessed
on 15 June 2021.
16. Journal of Renewable Materials. https://www.techscience.com/journal/jrm. Accessed on 15 June
2021.
17. Nature Materials. https://www.nature.com/nmat/. Accessed on 15 June 2021.
18. Polymer Composites. https://onlinelibrary.wiley.com/journal/15480569. Accessed on 15 June 2021.
19. RSC Advances. https://www.rsc.org/journals-books-databases/about-journals/rsc-advances/.
Accessed on 15 June 2021.
20. 2021. Global Natural Fiber Composites Industry. Global Industry Analysts. April 2021.
21. Shaharuzaman, M.A., S.M. Sapuan, M.R. Mansor, and M.Y.M. Zuhri. 2020. “Conceptual Design
of Natural Fiber Composites as a Side-Door Impact Beam Using Hybrid Approach,” J. Renew.
Mater., 8(5):549-563.
22. Park, K., C. Kong, and H. Park. 2015. A Study on Structural Design of Natural Fiber Composites
Automobile Body Panel Considering Impact Load,” Compos. Res., 28(5);291-296.
23. Scarponi, C. 2015. “Hemp fiber composites for the design of a Naca cowling for ultra-light
aviation,” Compos. B. Eng., 81:53-63.
24. Vistex Composites. https://www.vistexcomposites.com/technology. Accessed on 15June2021.
1663

25. Cochran W.G. and C.M. 1992. Experimental Design: Chapter 4 Completely Randomized,
Randomized Block, and Latin Square Designs. 2nd ed. Wiley Classics Library.
26. Amikhosravi A., M. Pishvar, Y.K. Hamidi, and M.C. Altan. 2020. “Accurate characterization of
fiber and void volume fractions of natural fiber composites by pyrolysis in a nitrogen atmosphere.”
AIP Conference Proceedings, 2205, 020032.
27. Lilholt H. and J.M. Lawther. 2000. Natural organic fibres. In: Comprehensive composite materials,
editors: Kelly A. and Zweben C. Vol. 1. New York: Pergamon Press.
1664