Composites Part A 137 (2020) 105986

Contents lists available at ScienceDirect

Composites Part A
journal homepage: www.elsevier.com/locate/compositesa

Ramie fabric Elium® composites with flame retardant coating: Flammability, T

smoke, viscoelastic and mechanical properties
⁎
Pooria Khalilia, , Brina Blinzlera, Roland Kádára, Per Blomqvistb, Anna Sandingeb,
Roeland Bisschopb, Xiaoling Liuc
a
Department of Industrial and Materials Science, Chalmers University of Technology, 412 96 Gothenburg, Sweden
b
Division Safety and Transport/Safety/Fire Research, RISE Research Institutes of Sweden, 501 15 Borås, Sweden
c
New Materials Research Institute, University of Nottingham Ningbo China (UNNC), Ningbo, 315100, PR China

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

Keywords: This investigation studied the utilization of intumescent thermal resistive mats to provide surface protection to
A. Natural fibres the core natural fibre-reinforced Elium® composite structural integrity. The intumescent mats contained flame
A. Thermoplastic resin retardant (FR) i.e. expandable graphite (EG) with four different expansion ratios and alumina trihydrate (ATH).
E. Vacuum infusion All natural fibre thermoplastic composites were fabricated using a resin infusion technique. The impact of char
Elium®
thickness and chemical compositions on the flammability and smoke properties was investigated. It was found
that surface protection significantly reduced the peak heat release rate, total smoke release, smoke extinction
area and CO2 yield, and substantially enhanced UL-94 rating, time to ignition and residual char network, de-
pending on the EG exfoliation ratio, ATH and mineral wool fibre. The glass transition temperature increased for
the FR composites containing EG with lower expansion ratio. Inclusion of intumescent mats increased the
strength of the composites while it had a negative effect on the modulus.

1. Introduction requirements of manufacturing fibre reinforced polymer (FRP), this

fibre was chosen as a reinforcement for composite components of the
In recent years, with the energy and global environment issues be- airplane secondary structures in the European Union’s Horizon 2020
coming progressively eminent, the awareness of environmental and research and innovation program between the European and Chinses
ecological protection has constantly been strengthened. Natural fibres partners [10]. Ramie is white in color and can be woven easily. The
extracted from plants are considered an environmentally friendly al- fibre is widely used in fabric industry owing to its dyeability, bleach-
ternative to substitute energy-intensive and more expensive synthetic ability and softness. Ramie has the density of 1.5 g/cm3 and the che-
fibres. This is attributed to their attractive features such as availability, mical composition includes 68.6–76.2 wt% cellulose, 13.1–16.7 hemi-
renewability, low cost, low density and abundance, together with cellulose, 0.6–0.7 wt% lignin, 1.9 wt% pectin, 0.3 wt% wax [11]. In
growing environmental concerns for the utilization of carbon and glass terms of modulus, ramie, flax, jute and sisal fibres have the values of
fibres [1–3]. In terms of manufacturing, lignocellulosic fibres reduce 61.4–128 GPa, 27.6 GPa, 13–26.5, 9.4–22.0 GPa, respectively while E-
the abrasiveness of composite part surfaces, decrease damage to glass has the modulus of 70 GPa. The range of tensile strength for
molding and tooling equipment, and provide comparatively better ramie, flax, jute, hemp and sisal is 400–938 MPa, 345–1100 MPa,
surface finishes [4]. Presently, jute, hemp, flax, sisal and coconut fibres 393–773 MPa, 690 MPa and 468–640 MPa, respectively, however, the
are used in the market as replacements for common glass fibre re- strength for E-glass fibre is about 2000–3500 MPa which is relatively
inforcements in automotive parts such as interior panels [2,5–7]. Im- high [2,11,12]. Therefore, amongst natural fibres ramie is considered as
plementation of green composites can also decrease disposal issues. a very suitable choice. It is worth stating that these types of natural
Ramie is known as Chinese grass and is a perennial herbaceous plant fibres possess comparable specific tensile strength with that of glass
of the Urticaceae family [8]. More than 90% of the ramie cultivation is fibre [11].
contributed by China and it can be harvested thrice in a year [8,9]. Due For the choice of polymer, thermosets are used in more high-per-
to the suitable properties of ramie fibres and capability of meeting the formance applications due to their enhanced mechanical properties and

⁎
Corresponding author.
E-mail addresses: pooriak@chalmers.se, pooria.khalili@gmail.com (P. Khalili), brina.blinzler@chalmers.se (B. Blinzler), roland.kadar@chalmers.se (R. Kádár),
per.blomqvist@ri.se (P. Blomqvist).

https://doi.org/10.1016/j.compositesa.2020.105986
Received 23 November 2019; Received in revised form 28 May 2020; Accepted 29 May 2020
Available online 30 May 2020
1359-835X/ © 2020 Elsevier Ltd. All rights reserved.
P. Khalili, et al. Composites Part A 137 (2020) 105986

average curing temperature. Thermoplastics in contrast require rela- to lower the smoke release, toxic gases and flame spread. In this work,
tively elevated processing temperatures, which is energy-intensive, but the fire resistive mats were bonded into the composite system in one
they are commonly used for particular applications in automotive, step process i.e. resin infusion. The effect of EG expansion ratio and the
packaging and construction industries. However, due to their recycl- combined impact of ATH and EG particles on the heat release, smoke,
ability [13], good vibration dampening capacities [14], high impact char formation, viscoelastic behavior, mechanical and morphological
tolerance [13] and post-manufacturing formability [15], thermoplastics properties of ramie fabric Elium® composites were investigated for the
are appealing to the manufactures. Arkema Chemicals company has first time. To the best of authors knowledge, the combined effect of
developed a new generation of thermoplastic resin that can be pro- natural fibre, Elium® polymer and FR agents i.e. EG and ATH have not
cessed similar to the thermosets e.g. resin infusion and RTM methods. been investigated prior.
Unlike the available commercial thermoplastics, Elium® can be poly-
merized at ambient temperature and demonstrates mechanical prop- 2. Methodology
erties similar to commercially available epoxy resin systems. Free ra-
dical polymerization initiates with the inclusion of a peroxide hardener, 2.1. Materials
and then the monomer (methyl methacrylate (MMA)) reacts to create
polymer i.e. poly methyl methacrylate (PMMA) [16,17]. The glass Plain-woven (0/90) ramie fabrics were supplied by AVIC Composite
transition temperature (Tg) of natural fibre (NF) Elium® composite Corporation ltd., Beijing, China and had the surface mass of 145 g/m2.
systems were also found to exhibit higher values than NF epoxy based According to the manufacturer, this kind of fabric (model: ZMPW
counterparts [18,19]. 140–1) had the moisture content of 3.4 wt%, and the ultimate tensile
Usage of NF (bamboo, hemp, flax, rice straw and bagasse pulp) load of 793 N and 685 N for the weft and warp, respectively. Elium®
composites has extensively been expanded in maritime structures, 150 thermoplastic resin which is a two-component system was provided
sports equipment, automotive exterior underfloor paneling, floor la- by Arkema company, France. Dibenzoyl peroxide was used a curing
mination, wall insulation, and both window and door frames [20]. A agent. The liquid density and viscosity of the resin was 1.01 g/m3 and
major challenge of the usage of natural fibre composites is their 0.1 Pa·s, respectively. Four different grades of commercial flame-re-
flammability and smoke release [21–24] which confines their use in tardant mats were sponsored by Technical Fibre Products Ltd., UK,
applications where fire safety regulations are strict, for instance namely, E20MI, E11MIL, T6663-03 and T6594-02 mats which mainly
railway, marine and aviation industries [25]. To improve the flame contain expandable graphite (EG) as a flame retardant agent. The
retardancy and reduce smoke and heat release of NF composites, in- particle size of EGs is different in the mats and provide variant thermal
corporation of flame retardant (FR) agents are necessary [26]. Acrylic expansion ratio. E20MI, E11MIL, T6663-3 and T6594-02 include EG
resin (Elium®) is highly flammable even more than NFs as was dis- with the expansion ratio of 20:1 (with the activation temperature of
covered in the previous work [27]. This was seen in the cone calori- more than 190 ˚C), 11:1 (with the activation temperature of more than
metry results and UL-94 burner test analysis, and the FR chemical 190 ˚C), 15:1(with the activation temperature of more than 450 ˚C), and
treatment of fabrics prior to the resin infusion was found to lead to the 9:1 (with the activation temperature of more than 450 ˚C), respectively.
decrease in the strength of the composites. Therefore, it was decided to The typical density (at 10 kPa) of E- and T-based FR mats were about
provide FR coating for the NF composites. To avoid multiple process for 327 kg/m3 and 190 kg/m3, respectively. For E20MI and E11MIL mats,
fabrication of ramie composite laminates and their coatings, insulative PVA (poly vinyl alcohol) of about 20–30 wt% as an organic binder, the
mats containing FRs were used. The incorporation of FR agents in the man-made vitreous silicate (mineral wool)/glass fibres of 20–40 wt%
mat can contribute to the compact and homogenous char residue as, and EG of 4–10 wt% were used to form the mats, and the latter contains
after exposure to the heat source, the particles are constrained to the 10–30 wt% of alumina tri-hydrate as well. For T6663-3, the same
protective mat. Intumescent non-woven mats with active ingredients constituents as the above-mentioned mats were used, but without sili-
can be bonded onto the top and bottom surface of natural fabric com- cate fibre. For T6594-02, alkaline earth silicate (AES) wool was re-
posites using resin infusion and provide a passive fire proofing. In the placed the silicate fibre to make the intumescent mat. The consolidated
event of fire, it is presumed that ignition occurs on the surfaces of thickness of all mats was 0.2 mm which was obtained after their
composite exposed to a heat source, which make the coating even more compaction at the pressure of 5 bar prior to the composite manu-
important. Expandable graphite (EG) is a sort of a layered crystal facturing.
structure intercalated with either an organic acid or inorganic acid i.e.
H2SO4 and HNO3 depending on the temperature at which the exfolia- 2.2. Resin infusion
tion is designed to initiate, which is normally between 150 ˚C to 200 ˚C
[28]. Upon heating, the intercalated graphite flakes decompose re- For fabrication of each composite plate, three layers of ramie fabrics
leasing gaseous products that cause exfoliation in the direction per- measuring 150 mm × 150 mm were cut and dried in a convection oven
pendicular to the carbon layers providing an insulative char barrier. for 24 h at 70 ˚C. The ramie fabrics were placed between two layers of
The generation of this multi-cellular char network hinders the diffusion FR mats, and then were positioned onto a mold, which was already
of oxygen and heat into the bulk of the composite and, therefore in- surface coated by a layer of semiperm® monofilm release spray. The
hibits flames from spreading. Besides, due to the formation of the processing was performed exactly the same as the previous work [27]
porous char network, gas products and combustible volatiles can be where a layer of peel ply was put on the fabrics, and two silicon con-
trapped, and as a result, this contributes to the reduction of the fuel feed nectors, spiral tube, resin feed PVC hose and outlet PVC hose were
into the burning zone and the decrease of the heat release [29]. Alu- connected to the system and sealed with the aid of vacuum bag and
mina trihydrate (ATH), during the course of combustion, undergoes an gum tape, however, as suggested by the manufacturer, in order to
endothermic reaction through generation of water followed by creation achieve void and defect free composite laminates and prevent any
of an insulating mineral layer, known as aluminum oxide (Al2O3), on shrinkage, double bagging system was used. This means that two layers
the bulk of composite. ATH is a well-known FR material for acrylic of gum tape were required around the perimeter of the mold, and two
based resins, and EG acts very promising and more effective at lower layers of vacuum bags i.e. one smaller and the other one larger were
concentration than other phosphorous/nitrogen based and ATH re- adhered to the gum tape. Therefore, −2 bar pressure could be applied
tardants [29,30]. to the composite during fabrication and curing rather than −1 bar with
Very few reports have been published to highlight the coating of only one vacuum bag [31]. In order to compact the fabrics and obtain
natural fabric composites using commercial flame-retardant mats to higher fraction of fibres in the composites, the vacuum was applied to
reduce the heat transfer from the fire source to the main structure and the system and then was released. The process was repeated three times

2
P. Khalili, et al. Composites Part A 137 (2020) 105986

to ensure the compaction of layers. Subsequently, the sealed system was 2.7. Mechanical tests
left for 10 min to ensure the absence of leakage. Elium® and dibenzoyl
peroxide were stirred manually for two min at a stochiometric ratio of The three point flexural performance of composite laminates was
100 (parts by weight) and 1.5 (parts by weight), respectively. The investigated on a Bent Tram A/S, Denmark, test machine according to
mixture was degassed for 3 min to remove the trapped gases from the BS EN ISO 14125. Four rectangular specimens measuring 80 mm
resin right prior to the infusion. After resin infusion, the part was left to (length) × 20 mm (width) with the span length of 64 mm were tested
cure for 24 h at the room temperature prior to the composite demolding for each formulation. The load-cell of the instrument and cross-head
and the thickness of composites was around 1.3 mm. The FR composites speed were 50 kN and 1 mm/min, respectively. The flexural strength
were designated after their commercial mat grades, namely, E20MI, and modulus were obtained, and stress-displacement curves were
E11MIL, T6663 and T6594. The Control composite without the FR plotted.
coating mats was produced for comparison of properties and the fibre The tensile tests were carried out to investigate the trend of modulus
mass fraction was about 45 wt%. It would be worth highlighting that and compare with the obtained bending modulus. Bent Tram A/S test
for the control composite the resin infusion process lasted approxi- instrument was used to perform the tensile tests in accordance with BS
mately 4 min and the procedure took 5 min on average for the com- EN ISO 527. Four rectangular samples of the dimension of 150 mm
posites with two layers of FR coating. Three composite plates were (length) × 20 mm (width) were characterized for each formulation and
produced for each formulation and were cut into specified sample di- the gauge length was 100 mm. Both ends of the samples were attached
mensions using diamond saw machine for different tests. to the tabs of 25 mm long. The cross-head speed was fixed at 1 mm/min
and the load-cell was 50 kN. The tensile modulus and elongation to
break were measured.
2.3. Cone calorimeter test
2.8. SEM
The heat release and smoke production behavior of the composites
were analyzed on a cone calorimeter (Fire Testing Technology, the UK).
Scanning electron microscopy (SEM) imaging of the cross-sectional
Sample dimensions of 100 mm × 100 mm were exposed to a heat flux
fracture surfaces of tensile specimens and an undeformed FR composite
of 35 kW/m2 in accordance with ISO 5660–1. Time to ignition (TTI),
sample were performed using a FEI Quanta200 ESEM, USA, operating
peak heat release rate (PHRR), time to PHRR (tPHRR), total heat release
at 10 kV in the low vacuum mode (0.8 torr) to avoid charging effects.
(THR), the effective heat of combustion (EHC), specific extinction area
All images were acquired in the secondary electron imaging mode.
(SEA), weight loss and total smoke release (TSR) were measured, and
the average values were used for plotting curves and tables. For the
2.9. XPS
sample that demonstrated the best fire retardancy performance, the test
was carried out at a heat flux of 50 kW/m2 as well.
X-ray photoelectron spectroscopy (XPS) tests were carry out using
PHI5000 VersaProbe III electron spectrometer, Japan, to obtain the XPS
2.4. Fourier transform infrared spectrometey (FTIR) spectra of the char residues for FR composites. The X-ray source was
monochromated Al X-ray (E = 1486.6 eV) and Beam size diameter was
A FTIR spectrometer (Thermo Scientific Antaris IGS analyzer, UK) 200 µm.
was connected to the cone calorimeter to measure the concentration of
toxic gases emitted from the samples in accordance with ISO 3. Results
19702:2015. The FTIR instrument had the spectral range of
4800–560 cm−1, resolution of 0.5 cm−1, scan over spectrum of 10 and 3.1. Flammability and smoke properties
time per spectrum of 12 s.
Vertical UL-94 test was performed to evaluate the flammability of
the composite laminates and the UL-94 category of the samples was
2.5. Vertical UL-94 test
chosen. No indication of molten drops was observed from any of the
composite samples during the test. The control sample was found to
In order to characterize flame propagation and dripping behavior of
burn off up to the clamp of the lab stand and rated as ‘’fail’’. Only the
the composites, UL-94 test was performed. A 20 mm height flame is
twisted fabric char remained from the Control samples. The rest of the
applied to the end of samples measuring 130 mm (length) × 13 mm
composite samples containing FR coating were categorized as V-0,
(width), hung vertically. Following is further discussion on the testing
which is the best classification of the vertical rating classification.
procedure. First, the flame is subjected to the bottom end of the sample
E20MI, E11MIL, T6663 and T6594 demonstrated similar phenomenon
for 10 s and the flaming time after withdrawal of the flame is recorded.
by protecting the surface of composite samples through the formation
If the flame extinguishes prior to burning up to the sample holder, the
of expanded graphite after the exposure to the 20 mm flame [29]. No
flame is applied for another 10 s. The flaming time after second removal
after-flame time was also recorded for all four flame retardant com-
of the burner is noted as well. During the test, a piece of cotton is placed
posites. Therefore, further investigation was necessary to distinguish
right below the sample to observe if the flaming drops from the sample
the fire and smoke protection of the composite laminate coated with FR
ignite the cotton.
mats.
The fire and smoke behaviors of NF composites were investigated
2.6. Dynamic mechanical analysis (DMA) test via cone calorimetry. Representative HRR vs. time curves for all the
composites are displayed in Fig. 1 and the derived numerical data are
Dynamic mechanical analysis tests were performed in the three shown in Table 1. Control samples had a relatively short TTI of 22 s and
point bending mode using Rheometrics Solids Analyzer RSA Ⅱ, USA. All pronounced PHRR of 453 kW/m2 while all the FR composites demon-
the samples were conditioned for 24 h at 50% humidity and 23 ˚C be- strated significantly higher TTI by 40–77% and lower PHRR by 58–72%
fore the tests and the sample dimensions were 50 mm than those of Control sample. With regards to the peaks which corre-
(length) × 10 mm (width). The frequency and heating rate were set to sponds to physical process taking place during the pyrolysis due to the
1 Hz and 5 ˚C/min. The temperature was ramped from 30 ˚C to 170 ˚C, exposure, the first sharp peak of each curve is attributed to a rapid heat
and then the storage modulus (E΄) and loss factor (tan δ) were mea- release that originated from the amount of combustible volatiles gen-
sured. erated from heat-induced depolymerization [32]. The first peaks

3
P. Khalili, et al. Composites Part A 137 (2020) 105986

Fig. 1. Heat release rate of the composite laminates.

Table 1 Fig. 2. A 2-D fire safety risk assessment grid in accordance to THR (or duration
Cone calorimetry data for all the composites exposed to 35 kW/m2 and an extra of fire) and PHRR/TTI (propensity to cause a rapidly growing fire) for FR
test for the optimum sample (E20MI) subject to 50 kW/m2, designated E20MI- composites exposed to 35 kW/m2).
50 in the table.
Sample TTI (s) PHRR THR EHC TSR (m2/ SEA (m2/kg) composites with FR coating were observed to delimitate after the
(kW/m2) (MJ/ m2) (MJ/kg) m2 ) completion of the experiments.
Control 22[1] 453[3.8] 19.2[1] 19.5[0.5] 50[2] 60.0[11]
Normally, the THR values, attained from the integration of the area
E20MI 39[4] 125[8.6] 32.5[3.6] 18.2[0.3] 9.5[0.9] 1.6[0.9] under the curves of HRR vs. time, enhanced with the formation of in-
E11MIL 39[0] 161[9.9] 40.5[1.2] 19.1[0.3] 24.9[3.9] 7.8[5.7] sulative char layers as the composites containing FR agents burned for
T6663 34[0] 191[3.5] 30.3[0.4] 19.2[0.1] 76.7[1.8] 44.2[2.7] longer period of time [27,32]; here is the case for the developed com-
T6594 35[2] 163[10.6] 35.1[2] 19.1[0.9] 49.7[2.3] 25.7[2.3]
posites as well. When the FR mats in the composite systems are exposed
E20MI-50 15 159 39.2 18.5 51.6 23.7
to the temperatures of 150–250 ˚C, they undergo thermal decomposi-
*The values in brackets show the standard deviation. tion which leads to production of gaseous species that are responsible
for the exfoliation of physical char. At similar temperature, the acrylic
occurred between 40 s and 48 s of the beginning of the test for all the polymer of the intumescent mat also initiates to degrade thermally to
composites. For the Control, depolymerization of Elium® occurred [27] yield organic gaseous species. EHC of composites coated with in-
while for FR composites, it was depolymerization of both resin and the tumescent mats reduced slightly. Smoke production properties followed
organic binder of the intumescent mats. The sharp decline after the first the same trend as TTI and PHRR for FR coated composites, e.g. a very
peak for the composites with the intumescent mats is explained by the high reduction of TSR by 81% and SEA by 98% for E20 composites as
formation of a protective char layer whereas a second peak for T6594 compared to those of the Control. This is again due to the high ex-
and T6663 and a less distinct second peak for E20 and E11MIL may pansion ratio of EG particles in E20. For E11MIL, the effect of both EG
have been created because of the continuous increase of surface tem- additives and ATH agent was remarkable in the decrease of TSR by 50%
perature resulting in the destruction of the formed char through oxi- and SEA by 87%. T-based composites also showed good properties due
dation. The smaller second peak for E20 and E11MIL is possibly due to to the presence of EG fillers, in particular T6594 owing to the presence
the generation of thicker insulating char layer inhibiting a swift rise in of AES wool. It is worth noting that the average SEA is a parameter
the temperature of the core composite during exposure. These two related to the amount of smoke released during combustion and to the
thicker char layers were capable of providing a more effective thermal opacity. Smoke particles could be trapped in the char network and may
barrier and prolonged the second peak during the test due to char have remained in the condensed phase, thereby lesser release of smoke.
oxidation. The HRR dropped with the depletion in the combustible Conversely, it was revealed that the addition of phosphorous, nitrogen
material volume. and/or carbon based FRs into flax fibre reinforced polypropylene (PP)
The residual char during the combustion process was measured and and polylactic acid (PLA) composites and also into jute fibre reinforced
for the Control no char left after 158 s. However, more than 8 wt% of acrylic polymer composite raised the smoke release [27,33], which
composites with FR coating retained after 620 s of exposure to 35 kW/ demonstrates the effectiveness of used intumescent mats even in terms
m2 heat flux. The char expanded due to the presence of FRs and digital of smoke production.
images in Fig. 3 show the residual char at the end of the tests taken The fire growth rate index (GRI) analyzed as a function of time is a
against a linear scale. E20 revealed the thickest residue, as expected. commonly used parameter to assess the fire safety during the pyrolysis
The fire performance possesses a direct dependency on the depth of of the material and is calculated as the value of HRR divided by the
char. Besides thick char, physical properties of the formed char con- combustion time (kW/m2s). The combustion time is considered the
tributes to the thermal insulation effect. Its structural integrity and time from the commencement of the test. FR composites followed a
thermal conductivity are of important characteristics of an effective similar pattern to that of uncoated composite. However, T-composites
char. The expanded residue acts as a thermal barrier as well as a phy- were found to reduce the GRI by approximately 3.1 time while the
sical insulative layer hindering the propagation of oxygen and heat to magnitude of decrease was about 3.6 and 4.3 for E11MIL and E20,
the combustion zone, hence shielding the underlying fibre polymer respectively, signifying that E20 is safer over the entire period of
laminates in integrated structures. E20 had the EG expansion ratio of burning. The fire behavior of composites can also be investigated by
20:1 which resulted in the thickest residual char, followed by E11MIL plotting the total heat release vs. the values of PHRR divided by TTI, as
and T6594 that had lower EG expansion ratio but with the inclusion of displayed in Fig. 2, to illustrate a 2-D fire risk assessment grid [34]. The
ATH and AES wool in the protective mat, respectively. T6663 showed results of FR composites are displayed in Fig. 2. The effect of both slow
the thinnest expansion amongst FR composites, and as a result less ef- growing fire (x-axes) and short fire duration (y-axes) can obtain a factor
fective, which led to a higher second HRR peak as well. The underlying of lower fire risk. The proximity of the results to the origin demon-
composite got exposed to high degradation temperatures ensuing the strates a greater fire retardancy. E20 and E11MIL had a lower pro-
acceleration of volatilization rate and shorter flame-out time. All pensity to lead to rapidly developing fires and E20, T6594 and T6663

4
P. Khalili, et al. Composites Part A 137 (2020) 105986

T6663
E20 E11MI T6594

T6594
E20 E11MI T6663

Fig. 3. The vertical and the horizontal views of the char of FR composites in the top and bottom rows, respectively, after the cone calorimeter test.

demonstrated a shorter fire duration than that of E11MIL. E20 is In order to further investigate the chemical structure of the char
deemed of greater overall fire safety according to the risk assessment residues, the XPS test was performed after the cone calorimeter tests
grid. and the measurements were aligned with carbon peak (C1s) at 284.8 eV
The FTIR results give the amount of toxic gases during composite [37]. The following observations were made, a large amount of C 1 s
combustion, as shown in Table 2. The CO2 and CO emissions and smoke (73–78 atomic % for all FR coated composites) and O 1 s (20–26 atomic
release are key parameters of a fire since the toxic gas production and % for all FR coated composites) were formed on the surface of the char
loss of visibility can inhibit the escape in an event of a fire. The CO2 residue, which is attributed to the exfoliated graphite and the anti-
yield, related to the combustion efficiency, reduced with the addition of oxidation of the char residues, respectively (Fig. 4 (a)). A small quantity
FR mats and the greatest reduction was for E20 by about 12% as of Si 2p (less than 1 atomic % for all FR coated composites) was also
compared to that of the Control. The order of decrease was followed by discovered in the char residues of the composite, which could enhance
E11MIL, T6594 and then T6663. It was discovered that the CO yield the fire protection by generating a compact silica layer [38], as T6594
enhanced for FR composites; similarly, Branda et al. obtained that silica contained the AES wool and glass, E11MIL and E20 had the silicate
treatment of hemp fabric and also addition of ammonia polyphosphate fibres and glass, and T6663 possessed the glass as well. Traces of S 2p
into the epoxy based composite systems led to an increase in CO yield (less than 1 atomic % for all FR coated composites) were found in the
[35]. In another study, a phosphorus based amine type curing agent char residues owing to the concentrated sulfuric acid which was in-
was used as FR and it was concluded that CO yield of hemp fibre epoxy tercalated between the crystalline network of EG particles [29].
composite increased [36]. The FTIR instrument was also calibrated to XPS spectrum of C 1 s and O 1 s for the char residue of E20 com-
measure the amount of nitric oxide (NO), nitrogen dioxide (NO2), sulfur posite sample are shown in Fig. 4 (b) and (c). The test was performed
dioxide (SO2), hydrogen bromide (HBr), hydrogen fluoride (HF), hy- for all four FR composites and the same bonds were obtained. The
drogen chloride (HCl) and hydrogen cyanide (HCN). None of the chemical state analysis of C 1 s XPS spectrum was performed and the
aforementioned gases were detected in the test measurements. It is peaks for different carbon based bonds were detected. The C-C, C-O,
worth highlighting that data on CO2 and CO was obtained from the C = O, COO–, CO32– and CO bonds were discovered in FR composite
cone calorimeter as well and the agreement for CO2 was good, however samples. The peaks located at 284.8 eV, 286.1 eV, 287.3 eV, 288.6 eV,
CO test data was a little variable. The FTIR apparatus was calibrated for 289.8 eV and 291.3 eV can be corresponded to the C-C, C-O, C = O,
CO in the low concentration range (less than 100 ppm) relevant here, COO–, CO32– and CO bonds, respectively [37,39]. For O 1 s peaks, the
therefore the FTIR data was used. C = O, C-O/SiO2, –OH, H2O and CO bonds were detected in the
As E20 demonstrated the best fire retardancy and smoke sup- composite samples, which is referred to the binding energy of 531.3 eV,
pressant performance, the cone calorimeter test was carried out at a 532.6 eV, 533.8 eV, 535.1 eV and 536.9 eV, respectively. Principally,
higher heat flux of 50 kW/m2 as well to evaluate its fire and smoke the presence of C = O bonds in the residues indicates good oxidation
behavior at a greater heat flux magnitude. As tabulated (Table 1 and resistance [40].
Table 2), for E20MI, the TTI was seen to decrease and other cone It could be interesting to compare the fire and smoke performance
parameters i.e. PHRR, THR, TSR and SEA were found to increase with a of natural fibre thermoset and thermoplastic systems with the current
raise in the heat flux. As obtained from the FTIR test data, the CO2 yield developed ramie fabric Elium® composite laminates. The data of the
remained unchanged and CO concentration decreased slightly after the following natural fibre composite systems are shown in Table 3. H/E-
increase in the heat flux. 15APP and HT/E-15APP composites contained 25 wt% hemp fabric and
15 wt% ammonia polyphosphate (APP) [35]; in HT/E-15 APP compo-
site, the fabrics were treated with water–glass. NHF composite had [36]
Table 2 30 wt% woven hemp fabric and epoxy resin with flame retardant curing
The yields of gases released from the specimens as determined by FTIR in- agent i.e. phosphorous containing amine type hardener, which pro-
strument.
vided 2.5 wt% of phosphorous in the composites. Pornwannachai et al.
Sample CO2 (g/g) CO (g/g) [33] made 50 wt% natural fibre reinforced composites using com-
mingled flax/PP and flax/PLA fabrics. The flame retardant used in these
Control 1.920[0.05] 0.0145[0.0015]
two composites was guanylure methylphsphonate. For FR-flax/PLA, the
E20MI 1.69[0.04] 0.0403[0.0065]
E11MIL 1.715[0.01] 0.0395[0.0045] concentration of FR on fabric, phosphorous and nitrogen contents in the
T6663 1.750[0.01] 0.0375[0.0005] composite were 9.9 wt%, 0.8 wt%, and 1.4 wt%, respectively. The
T6594 1.745[0.03] 0.0335[0.0015] quantity of mentioned constituents for FR-flax-PP composite was
E20MI-50 1.680 0.030
11.2 wt%, 1 wt% and 1.6 wt%, respectively. It can be realized that E20
*The values in brackets show the standard deviation.
demonstrates very decent performance in terms of PHRR, THR, TSR,

5
P. Khalili, et al. Composites Part A 137 (2020) 105986

Fig. 4. (a) XPS wide scan spectra of the char residue for FR composites and (b) XPS spectra of E20 char residue for C 1 s on and (c) O 1 s regions.

SEA and CO yield.

3.2. Dynamic mechanical analysis test

Dynamic mechanical analysis (DMA) tests were carried out to study

the impact of FR mats on the viscoelastic properties of ramie fibre Elium
composites (Fig. 5).
At the glassy phase (below 70 ˚C), the storage modulus of all FR
composites demonstrated an increase in comparison with that of the
control and the trend continued till the end of rubbery state region. This
is attributed to the capability of the coating mats to elastically store
more energy. The reduction in storage modulus of the composites in-
tensified at approximately 80 ˚C, associated with the beginning of the
glass transition zone.
It was found that only E11MIL composite reduced the glass transi- Fig. 5. Loss factor versus temperature of the composites.
tion temperature (Tg) by approximately 8% as compared to that of the
control (Fig. 5). E20 composite showed almost the same values as the
control while T6594 and T6663 composites enhanced the Tg by 9% and higher than that of the Control. The lowest enhancement was about 9%
7% relative to that of the Control, respectively. The reduction of Tg by for E20 and E11MIL composites in comparison with that of the Control.
E11MIL mat coating could be attributed to the presence of ATH. It was The increase in tensile strength after inclusion of mats is ascribed to
shown that [18] the addition of ATH into natural fibre composites led good impregnation between the polymer matrix and the fibres in both
to a drop in Tg. The enhancement of Tg for the T-based mat composites sides of the coating.
could be referred to their smaller EG size than E11MIL and E20. The After the tests, cracked samples were examined and it was found out
small particle size, as also shown in the previous work i.e. graphene that no delamination or significant fracture occurred, and the mode of
nano-platelet [27], have the potential of constraining the mobility level fracture was micro-cracks on the tension (bottom) side of the samples.
of micro molecules of polymeric matrix segments [18,41], which infers It is worth noting that the cracks formed at a higher displacement for
the Tg improvement. It was demonstrated prior that EG incorporation the Control samples as compared to the FR composites.
into epoxy and PA-6 polymer composites broadened the peaks [29,42]. The decrease in modulus could be attributed to the low aspect ratio
It is worth noting that the Tg of natural fibre Elium based composites (ratio of the highest to the lowest dimension) of fillers i.e. EG and ATH.
are higher than that of natural fibre epoxy counterparts [18,19]. The ability for fillers to modulate the bending modulus depends on the
particle size and aspect ratio of fillers. Reducing the particle size of
additives can lead to an increase in the modulus on the condition that
3.3. Mechanical properties
the aspect ratio enhances with the reduction of particle size. Here EG
and ATH are not counted as fillers with high aspect ratio; the potential
The flexural strength, modulus and stress–deflection of the Control
of nano fillers such as clay and graphene for enhancement of me-
and composites with FR coatings are shown in Fig. 6 (a). The bending
chanical properties is higher [43–45]. It was discovered in another
strength and modulus were 65.6 MPa and 9.8 GPa for the Control, re-
study [46] that the flexural modulus experienced a drop after the in-
spectively. The incorporation of FR mats was found to increase the
corporation of expanded graphite as a coating layer for a carbon fibre
bending strength whereas the modulus was seen to decrease.
reinforced resin system. Amongst FR composites, the modulus and
The increased flexural strength of the FR composites showed the
strength had the trend of decreasing for the composites containing
good adhesion between polymer and glass fibre and/or wool. The
larger EG particle size and thickness. T6594 composite had the EG with
greatest bending strength (T6594) was recorded approximately 42%

Table 3
Comparative cone calorimetry results of different natural fibre composite systems at a heat flux of 35 kW/m2.
Sample TTI (s) PHRR (kW/m2) THR (MJ/ m2) TSR (m2/ m2) SEA (m2/kg) CO2 (g/g) CO (g/g)

H/E-15APP [1] 46 259 34.4 938 394 0.87 0.05

HT/E-15 APP [1] 44 232 40.1 1230 413 1.02 0.05
NHF [2] – ̴
800 45.8 – 224 0.8948 0.0723
FR-flax/PLA [3] 260 261 9 619 – – –
FR-flax-PP [3] 46 297 87 1736 – – –
E20MIL 39 125 32.5 9.5 1.6 1.69 0.0403

6
P. Khalili, et al. Composites Part A 137 (2020) 105986

Fig. 6. (a) The three point bending properties of the composites and (b) the tensile performances of the composites.

(a)Control (b)E11MI

(c)T6594 (d)T6663

NF layer

(e) (f)
FR coating
1st layer

2nd layer
3rd layer FR coating
FR mat

Fig. 7. SEM micrographs of tensile fracture surface of the (a) Control, (b) E11MIL, (c) T6594, (d) T6663 composites, (e) digital image and (f) SEM micrograph of
undeformed T6663.

lowest expansion ratio of 9:1, which demonstrated higher mechanical modulus with the bending counterpart and it was detected that the
properties, and for the E20 composite with the EG expansion ratio of trend of tensile modulus of the composites agrees with the flexural
20:1, the mechanical properties were at the lowest level modulus (Fig. 6 (b)). The tensile modulus and elongation to break were
The tensile test was also performed to compare the trend of tensile seen to reduce after the incorporation of FR mats. For instance, the

7
P. Khalili, et al. Composites Part A 137 (2020) 105986

tensile modulus of T6594 reduced by 15% as compared to that of Supervision. Per Blomqvist: Validation, Resources, Formal analysis.
Control composite. As observed in Fig. 7 (a), the Control composite had Anna Sandinge: Resources. Roeland Bisschop: Resources. Xiaoling
good interfacial adhesion between the ramie fibres and the Elium® Liu: Resources.
matrix. The fibre breakage could be detected in the image, the fibres
were covered in the matrix and no indication of delamination was seen. Acknowledgments
The tensile tests are more fibre dependent and therefore, this re-
sulted in smaller tensile properties for the FR composite samples, which The financial support for this project is provided by Chalmers Area
is explained by the absence of the plain-woven natural fabrics in the top of Advance: Materials Science. The work was performed by the support
and bottom coatings. Therefore, the load does not transfer from the of All Wood Composites Platform based in Chalmers University of
main composite part to the coating sides, however, the bonding be- Technology and the fire tests were sponsored by RISE. Special thanks to
tween the ramie fibre composite and FR mats as a coating material Arkema company and Technical Fibre Products Ltd. for sponsoring the
could be considered suitable (Fig. 7 (b), (c) and (d)). The fibre thermoplastic resin and commercial intumescent mats. The authors
breakages are highlighted by red arrows in the micrographs and the would like to thank Malo Hedouin and Chirag Gurumurthy for the
interface between the main natural fibre composite part and coating technical support in the samples’ preparation and thank Georgia
material is indicated in black rectangular. No sign of delamination was Manika and Ahmet Semih Ertürk for helping with the SEM sample
witnessed in the micrographs. Fig. 7 (e) and (f) illustrates the digital preparation.
and SEM images of undeformed sample where NF layers and FR mat can
be observed. References
Amongst FR composites, EG particles, in particular the larger ones
i.e. in E11MIL and E20, disturbed the load transfer from matrix to fibres [1] Joshi SV, Drzal LT, Mohanty AK, Arora S. Are natural fiber composites en-
and this resulted in the reduced tensile properties. Besides, Nie et al. vironmentally superior to glass fiber reinforced composites? Compos A Appl Sci
Manuf 2004;35(3):371–6. https://doi.org/10.1016/j.compositesa:2003.09.016.
reported that addition of ATH, which is present in one of the for- [2] Saba N, Jawaid M, Alothman OY, Inuwa I. Hassan AJPfAT. A review on potential
mulations (E11MIL), had a negative effect on the toughness of poly development of flame retardant kenaf fibers reinforced polymer composites.
(methyl metacrylate) (PMMA) [47]. It is worth noting that the tensile 2017;28(4):424–34.
[3] Saba N, Jawaid M, Paridah M. Al-Othman OJPfAT. A review on flammability of
strength of Control was measured 66 MPa; however, as expected the epoxy polymer, cellulosic and non-cellulosic fiber reinforced epoxy composites.
incorporation of FR mats were found to reduce the strength of T6594 by 2016;27(5):577–90.
about 35% and of other formulations by ca. 50%. [4] Monteiro SN, Lopes FPD, Ferreira AS, Nascimento DCOJJ. Natural-fiber polymer-
matrix composites: cheaper, tougher, and environmentally friendly.
2009;61(1):17–22.
4. Conclusions [5] Marsh GJMt. Next step for automotive materials. 2003;4(6):36–43.
[6] Holbery J, Houston DJJ. Natural-fiber-reinforced polymer composites in automotive
applications. 2006;58(11):80–6.
The NF thermoplastic composites coated with the FR mats were
[7] Zah R, Hischier R, Leão AL. Braun IJJocp. Curauá fibers in the automobile in-
successfully produced using a resin infusion method. The non-woven dustry–a sustainability assessment. 2007;15(11–12):1032–40.
intumescent mats were seen to prolong the TTI by 55–77%, lead to V-0 [8] Liu F, Liu Q, Liang X, Huang H, Zhang SJGR, Evolution C. Morphological, anato-
rating, reduce the PHRR by 58–72%, SEA by 27–98%, GRI by about mical, and physiological assessment of ramie [Boehmeria Nivea (L.) Gaud.] toler-
ance to soil drought. 2005;52(5):497–506.
68–77% and the CO2 yield by 9–12% for all the composites with FR [9] Wang B, Peng D-X, Liu L-J, Sun Z, Zhang N, Gao S-MJBS. An efficient adventitious
coating when compared to those of the Control. TSR also reduced by shoot regeneration system for ramie (Boehmeria nivea Gaud) using thidiazuron.
81% and 50% for E20 and E11MIL composites respectively as compared 2007;48(2):173–80.
[10] Dong S, Xian G. Yi X-SJA. Life cycle assessment of ramie fiber used for FRPs.
to that of the Control. The residue was more than 8 wt% for FR coated 2018;5(3):81.
composites after 620 s exposure and according to the 2-D fire safety [11] Mohanty A, Misra MA, Hinrichsen GJMM. Engineering, Biofibres, biodegradable
assessment grid, E20MI was found to demonstrate the lowest fire risk. polymers and biocomposites: An overview. 2000;276(1):1–24.
[12] Kim N, Lin R, Bhattacharyya D. Flammability and mechanical behaviour of poly-
Char thickness was found to enhance proportional to the expansion propylene composites filled with cellulose and protein based fibres: A comparative
ratio of EG particles. The presence of ATH and wool fibre could also study. Compos A Appl Sci Manuf 2017;100:215–26.
contribute to the char formation for E11MIL and T6594, respectively. [13] Offringa AR. Manufacturing. Thermoplastic composites—rapid processing applica-
tions. 1996;27(4):329–36https://www.sciencedirect.com/science/article/pii/
When the ramie fibre Eium® composites with FR coating were com- 1359835X95000487.
pared with the natural fibre epoxy based composite counterparts, the [14] Chou PJ, Ding D. Chen W-HJJorp, composites. Damping of moisture-absorbed
enhanced properties in terms of PHRR, TSR, SEA and CO yield could be composite rackets 2000;19(11):848–62.
[15] Sadighi M, Rabizadeh E. Kermansaravi FJJompt. Effects of laminate sequencing on
highlighted. The Tg values of T6663 and T6594 were higher than for the
thermoforming of thermoplastic matrix composites. 2008;201(1–3):725–30.
Control. A large amount of C 1 s was discovered by XPS analysis on the [16] Rijswijk KV, Bersee HE. Manufacturing. Reactive processing of textile fiber-re-
char resides due to the presence of exfoliated graphite, and the C-C, C- inforced thermoplastic composites–An overview. 2007;38(3):666–81https://www.
O, COO–, CO, CO32– and C=O bonds were observed in the char residues sciencedirect.com/science/article/pii/S1359835X06002247#!.
[17] Bhudolia SK, Perrotey P, Joshi SC. Optimizing Polymer Infusion Process for Thin Ply
of the FR composites, which the latter reveals good oxidation re- Textile Composites with Novel Matrix System. 2017;10(3):293.
sistance. Inclusion of intumescent mats onto the NF composites not only [18] Khalili P, Liu X, Tshai KY, Rudd C, Yi X, Kong I. Development of fire retardancy of
enhanced the flame retardancy and smoke properties of the composite natural fiber composite encouraged by a synergy between zinc borate and ammo-
nium polyphosphate. Compos B Eng 2019;159:165–72.
laminates, but also had a positive impact on the bending strength of [19] Nissilä T, Hietala M, Oksman K. A method for preparing epoxy-cellulose nanofiber
developed materials; however, the modulus was lower than that of the composites with an oriented structure. Composites Part A: Applied Science and
Control. It was observed that the SEM micrographs of fracture samples Manufacturing; 2019. p. 105515.
[20] Pickering KL, Efendy MGA, Le TM. A review of recent developments in natural fibre
displayed fibre breakages. The developed composites demonstrate good composites and their mechanical performance. Compos A Appl Sci Manuf
properties for interior applications and our work continues to establish 2016;83:98–112.
better performances for the bio-based fibre composites. [21] Kandola BK. Flame Retardant Characteristics of Natural Fibre Composites. Natural
Polymers; Volume 1. Composites: The Royal Society of Chemistry United Kingdom
2012:86–117.
CRediT authorship contribution statement [22] Azwa Z, Yousif B, Manalo A, Karunasena W. A review on the degradability of
polymeric composites based on natural fibres. Mater Des 2013;47:424–42.
[23] Shah AUR, Prabhakar M, Song J-IJIJoPE, Technology M-G. Current advances in the
Pooria Khalili: Conceptualization, Visualization, Formal analysis,
fire retardancy of natural fiber and bio-based composites–A review.
Investigation, Methodology, Supervision, Writing - original draft. Brina 2017;4(2):247-62.
Blinzler: Funding acquisition, Validation, Resources, Supervision. [24] Kim N, Dutta S, Bhattacharyya DJCS. Technology. A review of flammability of
Roland Kádár: Funding acquisition, Validation, Resources, natural fibre reinforced polymeric composites 2018;162:64–78.

8
P. Khalili, et al. Composites Part A 137 (2020) 105986

[25] Gurunathan T, Mohanty S, Nayak SK. A review of the recent developments in [37] Krawczyk P. Effect of ozone treatment on properties of expanded graphite. Chem
biocomposites based on natural fibres and their application perspectives. Compos A Eng J 2011;172(2):1096–102.
Appl Sci Manuf 2015;77:1–25. [38] Shi Y, Wang GJASS. The novel silicon-containing epoxy/PEPA phosphate flame
[26] Subasinghe A, Bhattacharyya D. Performance of different intumescent ammonium retardant for transparent intumescent fire resistant coating 2016;385:453–63.
polyphosphate flame retardants in PP/kenaf fibre composites. Compos A Appl Sci [39] Jiang L, Zhang J, Xu X, Zhang J, Liu H, Guo Z, et al. Characterization and appli-
Manuf 2014;65:91–9. cation of expanded graphite modified with phosphoric acid and glucose for the
[27] Khalili P, Blinzler B, Kádár R, Bisschop R, Försth M, Blomqvist PJM. Flammability, removal of Ni (II) from aqueous solution. 2015;357:2355–63.
Smoke, Mechanical Behaviours and Morphology of Flame Retarded Natural Fibre/ [40] Zhao W, Liu J, Peng H, Liao J, Wang XJPD. Stability. Synthesis of a novel PEPA-
Elium® Composite. Materials 2019;12(17):2648. substituted polyphosphoramide with high char residues and its performance as an
[28] Rassmann N, Gmb HJPA. Compounding. Expandable graphite used as fire barrier in intumescent flame retardant for epoxy resins. 2015;118:120–9.
plastics. 2000;2(6):12. [41] Pedrazzoli D, Pegoretti A, Kalaitzidou K. Synergistic effect of exfoliated graphite
[29] Khalili P, Tshai K, Kong I. Natural fiber reinforced expandable graphite filled nanoplatelets and short glass fiber on the mechanical and interfacial properties of
composites: Evaluation of the flame retardancy, thermal and mechanical perfor- epoxy composites. Compos Sci Technol 2014;98:15–21.
mances. Compos A Appl Sci Manuf 2017;100:194–205. [42] Uhl Fawn M, Yao Qiang, Nakajima Hiroyoshi, Manias E, Wilkie Charles A.
[30] Khalili P, Tshai K, Hui D, Kong I. Synergistic of ammonium polyphosphate and Expandable graphite/polyamide-6 nanocomposites. Polym Degrad Stab
alumina trihydrate as fire retardants for natural fiber reinforced epoxy composite. 2005;89(1):70–84. https://doi.org/10.1016/j.polymdegradstab.2005.01.004.
Compos B Eng 2017;114:101–10. [43] Saba N, Paridah M, Abdan K, Ibrahim N. Effect of oil palm nano filler on mechanical
[31] Hou T, Jensen B. Evaluation of double-vacuum-bag process for composite fabrica- and morphological properties of kenaf reinforced epoxy composites. Constr Build
tion. 2004. Mater 2016;123:15–26.
[32] Kandare E, Chukwudolue C, Kandola Bl. The use of fire-retardant intumescent mats [44] Saba N, Jawaid M, Asim M. Recent advances in nanoclay/natural fibers hybrid
for fire and heat protection of glass fibre-reinforced polyester composites: Thermal composites. Nanoclay reinforced polymer composites. Springer; 2016. p. 1–28.
barrier properties. Fire and Materials: An International Journal. 2010;34(1):21–38. [45] Saba N, Jawaid M. Chapter 10 - Functionalized Graphene Reinforced Hybrid
[33] Pornwannachai W, Ebdon J, Kandola B. Fire-resistant natural fibre-reinforced Nanocomposites and Their Applications. In: Jawaid M, Bouhfid R, Kacem Qaiss Ae,
composites from flame retarded textiles. Polym Degrad Stab 2018. editors. Functionalized Graphene Nanocomposites and their Derivatives: Elsevier;
[34] Schartel B, Braun U, Schwarz U, Reinemann S. Fire retardancy of polypropylene/ 2019. p. 205-18.
flax blends. Polymer 2003;44(20):6241–50. [46] Miller S, Papadopoulos D, Heimann P, Inghram L, McCorkle L. Graphite sheet
[35] Branda F, Malucelli G, Durante M, Piccolo A, Mazzei P, Costantini A, et al. Silica coating for improved thermal oxidative stability of carbon fiber reinforced/PMR-15
treatments: a fire retardant strategy for hemp fabric/epoxy composites. Polymers. composites. Compos Sci Technol 2007;67(10):2183–90.
2016;8(8):313. [47] Nie S. A micromechanical study of the damage mechanics of acrylic particulate
[36] Szolnoki B, Bocz K, Sóti PL, Bodzay B, Zimonyi E, Toldy A, et al. Development of composites under thermomechanical loading: State University of New York at
natural fibre reinforced flame retarded epoxy resin composites. Polym Degrad Stab Buffalo; 2005.
2015;119:68–76.