Composites Part B 114 (2017) 101e110

Contents lists available at ScienceDirect

Composites Part B
journal homepage: www.elsevier.com/locate/compositesb

Synergistic of ammonium polyphosphate and alumina trihydrate as

fire retardants for natural fiber reinforced epoxy composite
P. Khalili a, K.Y. Tshai a, D. Hui b, I. Kong a, *
a
Department of Mechanical, Materials and Manufacturing Engineering, The University of Nottingham, Malaysia Campus, Jalan Broga, 43500, Semenyih,
Selangor, Malaysia
b
Department of Mechanical Engineering, University of New Orleans, Lake Front, New Orleans, Louisiana, LA 70138, USA

a r t i c l e i n f o a b s t r a c t

Article history: Various compositions of epoxy/hardener with flame retardant (FR) as additives were used to produce
Received 7 September 2016 natural fiber composite through the resin infusion technique. The effects of natural fiber (NF), ammo-
Received in revised form nium polyphosphate (APP), alumina trihydrate (ATH) and ATH/APP hybrid on flammability, thermal and
11 December 2016
mechanical properties of the composites were investigated. Incorporation of NF into neat epoxy matrix
Accepted 28 January 2017
Available online 31 January 2017
reduced the gross heat of combustion with improved thermal degradation properties. Addition of APP
enhanced the flame resistant properties of the composite with a much reduced total flame time and zero
drip. The 10 wt$% ATH and 5 wt$% APP hybrid showed the most promising flame retardancy with a self-
Keywords:
Hybrid
extinguishing property as well as the lowest gross heat and greatest char residue amongst the various
Polymer-matrix composites (PMCs) formulations investigated. In general, the formulations containing only ATH couldn't provide the effec-
Mechanical properties tive flame resistivity as compared to that of the APP-filled composites. The ATH-filled formulations
Thermal properties reduced the tensile strength, elongation at break and flexural properties while an increase in the flexural
Thermosetting resin strength of APP-loaded formulations was observed. An increase in moduli of the FR-filled composites was
measured as compared to that of the NF/epoxy composite, owing to a more brittle characteristic.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction Unlike thermoplastics, thermoset resin possess low viscosity

prior to crosslink and their fabrication route does not require to be
Over the past decade, the increasing awareness in community operated at high temperature. The merit of thermosetting polymer
health and environment has boosted the utilization of fibres ob- provides ease of fabrication for accommodating constituent
tained from renewable sources. The usage of natural fiber (NF) mixture of NF reinforced composites [6], and it was also found that
composites have been gaining momentum in several applications epoxy resin could exhibit stronger surface interaction than that of
such as the automotive, furniture and food packaging industries the unsaturated polyester. In addition, processing of thermosetting
due to their renewability, availability, low density, low cost and composites can be performed with much lower processing pres-
appropriate mechanical performances [1e5]. However, the fall- sure than that of the thermoplastic polymers [7].
backs such as high flammability behavior of NF, poor compatibility Intumescent flame retardants (IFRs) have been widely used to
of NF with polymeric matrices of hydrophobic nature, and the high reduce the flammability of polymers, with enhanced properties
processing temperature of thermoplastic polymers (which such as low corrosion and toxicity, reduced smoke emission,
adversely affects physio-mechanical performances of NF) [1] increased char formation and anti-dripping flame [8]. IFRs consist
broadly limits their wider application, in particularly the venture of a source of blowing agent, a source of char formation, and an acid
towards aviation industry which governs by an exhaustive set of source which together capable of generating a foamed char barrier
stringent regulatory requirements. that decreases the transport of fuel and heat into the bulk of the
constituents [9e11]. In the event of fire, an intumescent ammo-
nium polyphosphate (APP) agitated by the heat source eliminate
ammonia (NH3) and water while forming ultraphosphate which
* Corresponding author. Department of Mechanical, Materials and Manufacturing
Engineering, The University of Nottingham, Malaysia Campus, Jalan Broga, 43500, could act as a protective layer against fire propagation. Alumina
Semenyih, Selangor, Malaysia. trihydrate (ATH) on the other hand undergoes endothermic
E-mail address: [email protected] (I. Kong).

http://dx.doi.org/10.1016/j.compositesb.2017.01.049
1359-8368/© 2017 Elsevier Ltd. All rights reserved.
102 P. Khalili et al. / Composites Part B 114 (2017) 101e110

dehydration upon fire, releasing water vapor which function as copolymer noticed clear interactions between APP and ATH as a FR
diluent for the combustible gases with concurrent formation of a system. The chemical reactions took place over a varying range of
thermally stable ceramic alumina (Al2O3) that serve as an insulating temperatures. The chemical reaction first occurred between APP
layer against heat and mass transfer [12,13]. The chemical struc- (NH4PO3) and hydroxyl groups (
OH) of ATH at a temperature
tures of APP and ATH are shown in Fig. 1. However, increasing the lower than 400  C (Fig. 2). This resulted in the formation of
concentration of FRs in a polymeric formulation often associated aluminum phosphate with generation of water and ammonia (NH3)
with reduction in mechanical performances and inevitably out of ammonium salt (NH4). These reactions kept repeating until
increased material cost, thus the smallest effective FR loading, aluminum metaphosphate ([Al(PO3)3]n) was formed.
which could be up to 20 wt$%, is generally desirable [14]. From the At higher temperature, further reactions took place forming
manufacturers' perspective, it is crucial to derive the optimum aluminum orthophosphate (AlPO4) from the long chain poly-
amount of IFRs to meet the demand for fire retardant system. phosphate as well as releasing ammonia and water (Fig. 3).
Castronvinci et al. [15] investigated the effects of APP/ATH The combined effect of APP/ATH contributed to an improved fire
hybrid on fire retardancy of styrene butadiene rubber (SBR) block extinguishing behavior due to a number of factors. Firstly, the
generation of aluminum phosphate prevents the evaporation of
phosphorous volatility. Secondly, the basicity of ATH that furnish
protons exchange and an easier evolution of water and ammonia
gas. Thirdly, the crystalline form of Al-P-O alters the glassy ultra-
phosphate structure by which gases can readily escape the solid
surface.
Introduction of NF into neat polymers reduced the heat release
and enhanced flame retardant performances relative to those of the
pure matrices [16e18], and incorporation of small loading level of
FR could further reduce the flammability behavior of biocomposite
Fig. 1. The chemical structures of (a) APP and (b) ATH. at the cost of a slight drop in mechanical properties. Few reports

Fig. 2. Reactions between ATH and APP.

Fig. 3. Reactions between ATH and APP at elevated temperatures.

 P. Khalili et al. / Composites Part B 114 (2017) 101e110 103

[19e21] have been published on the enhancement of flammability, gum tape. Resin feed hose was connected to the bag and another
thermal stability and mechanical properties of NF based compos- PVC hose was firmly connected to the bag through vacuum line. An
ites using FRs and no previous studies were conducted to investi- infusion line clamp was used to close the resin feed line and, as a
gate the effect FRs on the fire resistant of the palm empty fruit result, evacuating air from the vacuum system. At the stage where
bunch (EFB) fiber reinforced epoxy based composite. This work full vacuum was achieved, the sealed bag was left for 15 min to
aimed to study the combined effects of APP, ATH and EFB fiber on ensure there isn't any leakage. Prior to resin infusion, FRs were
the flame resistivity and mechanical performance of the resulting added into the epoxy resin and approximately 15 min of ultra-
composite by means of a series of characterizations to obtain an sonication was applied using a UP400S Ultrasonicator from
optimized composition of the constituent loadings. Hielscher Ultrasonics GmbH at a frequency of 24 kHz and 50%
amplitude until a homogeneous dispersion was achieved. Hardener
2. Methodology was added into the sonicated mixture at a weight ratio of 5:1
(epoxy:hardener). The mixture was then infused throughout the
2.1. Materials fiber mat in the mold. The curing took place under room temper-
ature for 24 h at atmosphere temperature (Fig. 4). The combination
Alumina trihydrate (ATH) with median particle size of 1 mm in of neat epoxy and EFB fiber, as control sample, was also prepared
diameter was purchased from J. M. Huber Corporation (Malaysia); for comparison purpose. The formulations of the constituents and
ammonium polyphosphate (APP) powder of grade Exolit AP 422 their abbreviations are depicted in Table 1.
(degree of polymerization > 1000) was provided by Clariant
(Singapore) Pte Ltd; an ambient cure two part epoxide resin system
Ultimeg 2020 was purchased from AEV Asia Sdn. Bhd. (Malaysia). 2.3. Characterization
The viscosity of mixture of epoxy and hardener at 25  C was
measured 0.2e0.6 N s/m2; pulverized palm EFB fiber was pur- 2.3.1. Vertical Bunsen burner test
chased from Szetech Engineering Sdn Bhd. The EFB fiber had an Bunsen burner test was carried out in accordance with Federal
average length in the range of 10e20 mm and bundle diameter Aviation Regulation (FAR) part 25.853-b (equivalent to ASTM F 501,
100e200 mm; a water based formaldehyde-free cross-linked acry- Daimler-Benz Aerospace ATS 1000.001, Douglas DMS 1511, Boeing
late binding agent, Acrodur DS 3530 was supplied by BASF Malaysia BSS 7230 F2, British Aerospace BACM 155A and Airbus ATS
Sdn. Bhd. 1000.001) to determine flammability characteristic of the

2.2. Composite fabrication

2.2.1. Fiber mat

Randomly-orientated EFB fibers were dispersed in a water-
binder suspension of 10:1 (v/v) ratio to form a mat in a custom-
ized sieve (180 mm  320 mm). Fiber length of 1 cm was chosen to
avoid fiber entanglement and to encourage even distribution of
fiber during mat fabrication [22]. The sieve was then taken out
where draining of water took place. The fiber mat was then flipped
upside down onto pieces of dry cloth, and another layer of dry cloth
was placed on top of the fiber mat to absorb residual water from
both sides. The bonded mat was shifted to metal plates with blot-
ting papers on top and bottom of the mat. A load of 15 kg (equal to
2.6 kPa) was then applied on top of the fiber mat for 10 min to press
out the remaining water and to produce a mat with an even
thickness of 1 mm. Fully soaked blotting papers were replaced with
new ones and pressed with the same weight until the mat was
visibly dried. The dried fiber mat was subsequently placed within a
convection oven maintained at 100  C for 4 h with an applied top
loading of 4 kg while measurements were taken to ensure the fiber
mat showed no further weight loss due to moisture evaporation.
The dried fiber mat was further kept dry within the oven at a lower
temperature of 70  C for 24 h until resin infusion was to be carried
out. The fiber mat was dried under selected temperature to ensure
that the structure and properties were not damaged by the extreme
temperature.

2.2.2. Resin infusion

A summary of the composite fabrication process is outlined. In
the first stage, a PVC plate was sprayed with silicone release agent.
Then gum tape was applied around the perimeter of the mold and
the EFB fiber mat was placed into the mold. A layer of peel-ply was
positioned on the EFB mat. This was followed by an infusion mesh
that helped the resin to flow and distribute through the mold.
Subsequently an infusion spiral and resin feed silicone connector
was fitted to the vacuum system. Then vacuum line silicon
connector was positioned and the bagging film was stuck to the Fig. 4. The composites left to cure after the resin infusion process.
104 P. Khalili et al. / Composites Part B 114 (2017) 101e110

Table 1 composites' fractured surfaces was used to observe the adhesion,

Formulations of constituents. dispersion and load transfer between the fiber, FRs and epoxy.
Sample name ATH (wt$%) APP (wt$%) EFB fiber (wt$%) Epoxy (wt$%)

E100 e e e 100
3. Results and discussion
E82F18 (Control) e e 18 82
CATH5 5 e 18 77 3.1. Flammability
CATH10 10 e 18 72
CATH15 15 e 18 67
The control sample (E82F18) was fabricated to provide proper
CAPP5 e 5 18 77
CAPP10 e 10 18 72 comparisons with FR-loaded formulations. The results recorded in
CAPP15 e 15 18 67 terms of drip flame time, total flame time and mass loss (wt$%) are
CATH5APP10 5 10 18 67 shown in Fig. 5. Incorporation of ATH into EFB fiber reinforced
CATH10APP5 10 5 18 67
epoxy composite did not reduce the burning resistivity, as
compared to that of the control sample (E82F18), in terms of drip
flame time and total flame time. All of the ATH-filled composites
materials. Three specimens measuring 152.5  37.5  2 mm3, burned intensely and brightly with severe dripping. Adding more
equivalent to half of the standard FAR specified dimension was ATH fillers reduced the total flame time, with increasing of drip
tested for each formulation. The specimens were held vertically via flame time. With increasing loading of APP, the total flame time
a lab stand and flamed for 12 s using methane fuel supplied by slightly reduced. The lower total flame time and no dripping phe-
Bunsen burner placed directly under the specimen. The total flame nomenon were detected for all APP-filled composites relative to
time, mass loss percentage and drip flame time were measured to those of other formulations. This is probably due to the degradation
analyze the flammability properties. In order to meet the FAR re- of APP that leads to the formation of a strong char barrier which
quirements, the average flame time, drip flame time and burn limits heat and mass transfer, with release of ammonia and water
length are required to be below 15 s, 3 s and 152 mm. which dilutes the gaseous phase. Suardana and Lim [23] found that
degradation of APP generates phosphoric acid which could in-
2.3.2. Bomb calorimeter teracts with the hydroxyl group of natural fiber to form phospho-
Bomb calorimeter was used to validate the heat capacity of rous esters. This ester acts as catalyst for dehydration of fibers,
compounds prompted by FRs filled natural fiber reinforced com- leading to the formation of a carbonaceous structure with reduc-
posites. Gross heat of combustion was measured by PARR 6100 tion of vapor phase composition and released energy. CAPP10
bomb calorimeter according to ASTM D-240 standard. 1 mL of showed 5% lesser mass loss than that of CAPP5 and displayed 12%
water was contained in the metal bucket where the specimen was reduction in the mass loss compared to that of CAPP15. Also, similar
placed and gradually bomb charged with O2 to 30 atm (3 MPa) at optimum concentration of APP particle in the range of 3e20 wt$%
ambient temperature. Sample weight of 0.8e1 g was placed in a was achieved at 10e12 wt$% by Castrovinci et al. [15]. The flame
cylindrical crucible and the specimen was linked to the fused wire resistivity of APP at higher loading (>10 wt$%) tends to be lower,
for combustion to occur. At least two samples from each formula- implying that APP can affect either positively or negatively
tion were tested and the average measurements were computed. depending on the optimum loading.
The results attained from the bomb calorimeter were repeatable to APP is much more effective than ATH in flame resisting com-
approximately ±99.85%. posites. It was observed that [15] inclusion of 60 wt$% ATH was
required to achieve similar flame resistivity effect attained with
2.3.3. Thermal stability (TGA) 10e12 wt$% APP. However, composites with higher concentration
The thermal stability was evaluated on a TGA 1 STARe system of ATH were prone to significant drop in mechanical properties.
(Mettler Toledo) Q500 instrument. Specimen of approximately Exceptionally, when 10 wt$% APP and 5 wt$% ATH were intro-
10e15 mg in weight was placed in an open alumina pan under duced into the EFB reinforced epoxy composite, a synergistic effect
nitrogen flow rate of 50 ml/min and heated from ambient (30  C) to took place that created self-extinguishing properties upon the
600  C at a heating rate of 20  C/min. The char residual and thermal removal of Bunsen burner. However, formulation with 10 wt$% ATH
degradation temperature (Td) were calculated. and 5 wt$% APP couldn't establish self-extinguishing behavior but
zero drip flame was still maintained. The synergistic effect of the
2.3.4. Tensile testing combined APP and ATH was discussed earlier in the introduction
Rectangular specimens of 120  20  2 mm3 were tested with section.
the aid of a Llyod Instruments LR50k Testing Machine in accordance
to ASTM D3039 at a crosshead speed of 2 mm/min and the load-cell 3.2. Analysis of residual char
of 5 kN was used. Five specimens of each formulation were tested
using a gauge length 60 mm and the mean value was recorded. Fig. 6 shows the structure of composites containing APP and
APP/ATH upon completion of the flammability tests. The compos-
2.3.5. Flexural testing ites with ATH-based additive burned more vigorously than those
Rectangular specimens of 100  20  2 mm3 were tested using a loaded with APP, and no mass residue was left in the clamp of lab
Llyod Instruments LR50k Testing Machine in a three point bending stand. All the burned parts of ATH loaded composites fell down on
mode with a span length of 60 mm in accordance to ASTM D790. the tray placed and ATH filled composites couldn't retain the
Crosshead speed of 5 mm/min and load-cell of 5 kN were used. Five structure. This is owing to the greater flame resistivity of APP than
specimens of each composition were tested and the mean values that of ATH as this filler could not adequately confine the pyrolysis
were used to analyze the flexural properties of the various process and volatile gases liberation during the fire, expediting the
formulations. combustion progression and increasingly deteriorating the
structure.
2.3.6. Scanning electron microscopy (SEM) As seen in Fig. 6, the bulk structure was maintained in APP filled
SEM was carried out on FEI Quanta 400F FESEM microscope at composites. The hybrid of ATH and APP exhibited its synergistic
the increasing voltage of 20 kV. The SEM micrograph of the effect where inclusion of 5 wt$% APP and 10 wt$% ATH (in
 P. Khalili et al. / Composites Part B 114 (2017) 101e110 105

Fig. 5. The results of vertical Bunsen burner testing for the control sample and the composites filled with FR.

CATH10APP5 formulation) could prevent the drip of flame. When polymeric formulations, the greatest flame resistivity was observed
the replacement was changed to the optimum value of 10 wt$% APP in formulations containing 10e12 wt$% APP. This signifies the dual
and 5 wt$% ATH (in CATH5APP10 formulation), self-extinguishing flame retardancy impact of APP on polymeric composite that can be
properties were observed. This composite burned for < 1s upon either enhancing or dropping depending on the amount of APP.
removal of Bunsen burner after 12 s of continuous exposure to The incremental introduction of ATH could further slowdown
direct flame and created zero drip flame and burn length of less the heat release of combustion. This FR decomposes during com-
than 1 cm with merely 3% mass loss. bustion, generating water in the vapor state through pyrolysis and
It would be noteworthy to add that pieces of control sample alumina (Al2O3). The water leads to an endothermic process, which
broke off on the burning edge and subsequently dropped while this removes the heat from the underlying layers. The Al2O3 layer also
phenomenon didn't occur for all the composites containing ATH acts as a shielding barrier against the transfer of heat and mass.
and APP. This shows the effectiveness of APP and the hybrid of APP/ The hybrid APP and ATH tested (CATH5APP10 and CATH10APP5)
ATH on the NF-reinforced epoxy composites. recorded the lowest GH, 13.9% drop relative to that of the control
sample, implying their synergetic effect and the optimal weight
3.3. Bomb calorimetry study ratio as evaluated in this work, in agreement with results from
Bunsen burner analysis. The hybrid of APP and ATH resulted in a
The bomb calorimetry testing was used to investigate the heat of thermally stable surface shielding with aluminum orthophosphate
combustion of the reaction between the composites and oxygen Al-P-O4 crystalline compounds rather than the ultraphosphate, and
measured based on the changes in temperature of the water. The more complete elimination of ammonia and water which enhance
gross heat (GH) of combustion evaluated by the bomb calorimetry the thermal surface shielding of composites.
exhibits the combustibility of the composites. Pure epoxy burned
wildly when subjected to combustion, with GH of 35.36 MJ/kg, as 3.4. Thermal properties
shown in Table 2. When EFB fiber was introduced into pure epoxy, a
large drop in the GH, 11.02% was observed, which is in agreement The TGA testing was conducted in nitrogen atmosphere, as
with those reported in literature [24]. This can be attributed to illustrated in Table 3 and Fig. 7. For the pure fiber, the initial thermal
lesser gross heat of combustion of EFB fibre than that of the neat degradation (50e130  C) occurred with a slight mass loss which
epoxy. can be ascribed to the evaporation of water content. A subsequent
The incorporation of ATH or APP into epoxy/EFB fiber led to a sharp drop took place in the temperature range 220e315  C, mainly
reduction in GH. APP demonstrated 3.83% reduction in combustion due to the weight loss of hemicellulose followed by the pyrolysis of
performance of EFB fiber reinforced epoxy composite at low cellulose at temperature regions 315e400  C. The degradation of
loading (5 wt$%). Upon the addition of 10 wt$% APP, the GH reduced lignin took place over a large temperature range ~160e600  C. The
by 6.91% relative to that of the control sample, which is due to lesser low molecular weight protolignin degrades first but at a slower rate
heat of combustion of APP particles. APP generates water and than other constituents [25,26]. The poor linkage was ruined be-
ammonia which mitigates the gaseous state. The flame resistant tween 170 and 230  C whereas the debonding of stronger struc-
effect for the composites containing APP reached optimum when tures in aromatic rings occurred at elevated temperature. It's worth
10 wt$% APP was added. The same trend of heat release was noting that lignin content and thermal stability possess an inverse
recorded [15] when 3 to 20 wt$% of APP was incorporated into relationship, i.e. higher concentration of lignin increases the
106 P. Khalili et al. / Composites Part B 114 (2017) 101e110

Fig. 6. Digital photos of burnt samples of (a) CAPP5, (b) CAPP10, (c) CAPP15, (d) CATH5APP10 and (e) CATH10APP5 composites after vertical Bunsen burner experiment.

Table 2
The gross heat of combustion of E100, control and composites filled with FR from bomb calorimeter experiment.

Formulations Gross Heat (MJ/kg) Decrement of GH from the control composite (%)

E100 35.36 11.02

Control 31.85 0.00
CATH5 30.55
4.08
CATH10 30.00
5.81
CATH15 28.56
10.34
CAPP5 30.63
3.83
CAPP10 29.65
6.91
CAPP15 31.50
1.09
CATH5APP10 27.41
13.95
CATH10APP5 27.40
13.96

Table 3 releases water upon thermal degradation [27,28] and contributed

TGA results of the EFB fibre, control and formulations loaded with APP and ATH. to a premature degradation of epoxy as a result of hot water hy-
Formulations Char Residual at T 600 C (wt$%) Td ( C) at 50 wt$% mass loss drolysis. Besides, the second reason of reduced thermal degrada-
EFB fiber 30.72 346.8
tion temperature is owing to the presence of phosphate groups,
E100 13.59 380.0 which decomposes earlier to generate thermal char [29,30]. Similar
Control 15.56 385.8 fashion of lower Td was observed when APP was incorporated with
CATH5 20.47 383.3 epoxy [31], and into PP [32]. The same trend of higher post com-
CATH10 21.20 388.5
bustion mass residual with increasing APP loading was reported in
CATH15 22.16 389.3
CAPP5 21.41 376.7 the work of Dorez et al. and Wen et al. [1,8], respectively. In general,
CAPP10 22.20 358.0 the thermal decomposition mechanism of NF filled composites can
CAPP15 19.50 376.0 be categorized into two stages: the first stage corresponds to the
CATH5APP10 29.18 368.2
fiber phosphorylation and its degradation at low temperature
CATH10APP5 24.43 368.4
[33,34] while the subsequent stage corresponds to the matrix
decomposition. APP interacts with both the polymer matrix and
thermal stability. The solid charred residual up to 400  C can be natural fiber, and mainly fiber lignin content [35], leading to the
attributed to the pyrolysis of lignin and the cellulose. formation of shielding charred barrier for mass and heat transfer.
The composites containing FR started to break down at The existence of phosphate groups in the structure of APP relatively
approximately 300  C owing to the thermal degradation of epoxy lowers the thermal decomposition temperature and subsequently
resin. The major weight loss started at approximately 315  C and forms a protective charred layer which retards the epoxy degra-
sustained for at least another 100  C. Incorporation of APP was dation [29,30,36].
observed to enhance the char residue content, where a hybrid of With increased loading of ATH, the mass residue and thermal
10 wt$% APP and 5 wt$% ATH resulted in the highest recording. This stability improved when heated, which can be attributed to its
increment can be ascribed to the degradation of APP and ATH that decomposition into water vapor due to endothermic reaction and
leads to the formation of aluminum orthophosphate and aluminum conversion to the thermally stable aluminum oxide [2Al(OH)3 /
metaphosphate. However, APP addition to the composites reduced Al2O3 þ 3H2O] that demonstrates a very high melting temperature
the onset of thermal degradation temperature, Td since APP of 2054  C [37].
 P. Khalili et al. / Composites Part B 114 (2017) 101e110 107

Fig. 7. The degradation behaviors of the selected formulations containing APP and ATH.

Synergistic effect was detected when the combined ATH/APP control sample. The Young's modulus of the composites with
was used, where enhanced mass residue was recorded with 10 wt$% and 15 wt$% ATH loadings increased relative to that of the
approximately 87.53% and 57% for CATH5APP10 and CATH10APP5 control sample. Similar fashion was observed in the tensile prop-
relative to that of the control sample, respectively. The Td, which erties of ATH-filled thermoset polymer by Petersen et al. [40].
was computed at 50 wt$% mass loss, wasn't affected due to the Incorporation of FRs generally provides higher modulus than that
presence of APP in the mixtures. The same synergetic effect was of the control sample. Ramazani et al. [41] observed that the
observed in the Bunsen burner and bomb calorimetry analyses. modulus was higher or slightly lesser with increasing ATH loading
in the composites than that of the control sample. For the hybrid
3.5. Tensile properties composites, the tensile strength and strain reduced while the
Young's modulus increased relative to that of the control sample.
The tensile strength, strain and modulus of composites con-
taining FR and the control sample (E82F18) are shown in Fig. 8.
Intumescent flame retardants (IFR) considerably influence the 3.6. Flexural properties
mechanical properties of polymeric composites. The incorporation
of IFR significantly reduced the tensile strength of the composite Fig. 9 shows the flexural properties of the formulations. APP-
relative to that of the control sample. FRs acted as a nucleating filled composites were observed to possess greater flexural
agent, reducing tensile strength, and decreasing elongation at break strength than that of the control. The flexural strength was the
except those with lower FR loadings. It has been previously shown highest at 5 wt$% APP loading, with 34% increment as compared
that APP develops moisture absorption characteristic in wood to that of the control, followed by 10 wt$% APP and 15 wt$% APP. The
reinforced composites, resulting in reduced mechanical properties decline in flexural strain and strength with increasing FR loading
[38]. Incorporation of APP has been associated with decreased was caused by the poor interfacial adhesion between the FRs and
tensile strength of formulations in numerous work [8,9,34,39] as epoxy, which became increasingly significant at higher FR
well as reduction in the elongation at break [34,39], which was concentration.
attributed to the poor compatibility between APP and polymer Lim et al. [36] attained similar trend in which increasing APP
matrices. In the current work, the formulation with 10 wt$% APP loading resulted in a drop of the flexural performance of glass fiber/
loading showed higher tensile strength than that of 5 and 15 wt$% epoxy APP loaded composite, and flexural strengths of APP filled
APP filled composites. The control sample demonstrated an composites were all higher than that of the control sample. The
approximately 13% greater tensile strength than that of the control sample showed greater flexural strength and strain than
formulation with 10 wt$% APP. Compared to that of the control those of the ATH-filled composites and increasing addition of ATH
sample, the elastic modulus of APP-based formulations was loading decreased the flexural properties of composites. The flex-
enhanced, generally owing to the reduction in elongation to break ural strength of CATH10APP5 was slightly lesser than the strength
and an increasing brittleness of the composites. of CAPP5 and the flexural strength of CATH5APP10 further dropped
The average tensile stress of composites with low loading of relative to that of CATH10APP5. The strain values of the hybrid
ATH, i.e. 5 wt$% concentration, was 19.1% lower than that of the composites were lesser than that of CAPP5.
108 P. Khalili et al. / Composites Part B 114 (2017) 101e110

Fig. 8. Tensile strength, Young's modulus and elongation at break of APP and ATH loaded EFB/epoxy composites.

Fig. 9. Flexural properties of ATH and APP filled EFB/epoxy composites.

 P. Khalili et al. / Composites Part B 114 (2017) 101e110 109

Fig. 10. Morphology of (a) control, (b) CATH5, (c) CAPP10, (d) CAPP15, (e) CATH5APP10 and (f) CATH10APP5 at 600 .

3.7. Morphology The incorporation of APP was found to reduce the degradation
temperature, however APP depicted the capability of an enhanced
The SEM micrographs of tensile fracture surface of the control, charring, as revealed in TGA, ascribed to the formation of NF
5 wt$% ATH and APP-loaded and ATH/APP hybrid formulations are phosphorylation, zero drip flame time and reduced GH. Introducing
exhibited in Fig. 10. Homogenous distribution of flame retardants FRs, either ATH or APP alone, showed much lower flame retardant
and no obvious void content were observed in the SEM micro- behaviors, lower mass residue and higher gross heat of combustion
graphs. The mechanical properties can be affected by the physical than those of the ATH/APP coupled systems. This depicted the
characteristics of the polymer, types of additive, additive concen- establishment of synergistic effect, which can be attributed to the
tration and filler dimension, distribution and shape. The surface effect of ATH and APP mixture upon combustion that besides water,
interaction between filler and polymer matrix plays a key factor in forms two different shielding systems including a ceramic metal
the fracture mechanism and crystalline behavior of the polymer, oxide and aluminum orthophosphate with more complete elimi-
both of which directly influence the mechanical properties of the nation of ammonia and water in the vapor phase. Generally, the
formulations [42]. addition of FRs reduced the tensile strength and strain of com-
No additive is present in the control sample (Fig. 10(a)) whereas posites and the incorporation of ATH fillers decreased the flexural
the addition of ATH particles (CATH5) is visible as shown in strength and strain of composites as compared to those of the
Fig. 10(b). For the composites containing 10 wt$% and 15 wt$% APP control sample.
contents, the particles can be observed in Fig. 10(c) and (d),
respectively. The FR fillers are visible in both CATH5APP10 and Acknowledgments
CATH10APP5 formulations (Fig. 10(f) and (g)).
This work was supported in part by the Ministry of Science,
4. Conclusions Technology and Innovation (MOSTI) Malaysia under Grant 03-02-
12-SF0212 as well as the Faculty of Engineering at The University of
The effective amount of IFRs required to provide the best flame Nottingham Malaysia Campus.
retardant behaviors of EFB-epoxy based composites was examined.
The hybrid of 10 wt$% APP and 5 wt$% ATH was the formulation
References
which could meet the aircraft interior FAR requirements in terms of
vertical Bunsen burner test. Other formulations that were investi- [1] Dorez G, et al. Thermal and fire behavior of natural fibers/PBS biocomposites.
gated failed to meet the Bunsen burner test requirements. It was Polym Degrad Stab 2013;98(1):87e95.
observed that ATH typically required a greater concentration to [2] Yao F, et al. Thermal decomposition kinetics of natural fibers: activation en-
ergy with dynamic thermogravimetric analysis. Polym Degrad Stab
create an acceptable FR-based systems, but increasing ATH con- 2008;93(1):90e8.
centration evidently deteriorates the mechanical characteristics. [3] Hesser F. Environmental advantage by choice: ex-ante LCA for a new Kraft
110 P. Khalili et al. / Composites Part B 114 (2017) 101e110

pulp fibre reinforced polypropylene composite in comparison to reference 2012;72(13):1479e86.

materials. Compos Part B Eng 2015;79:197e203. [23] Suardana NPG, Ku MS, Lim JK. Effects of diammonium phosphate on the
[4] Scarponi C, Messano M. Comparative evaluation between E-Glass and hemp flammability and mechanical properties of bio-composites. Mater Des
fiber composites application in rotorcraft interiors. Compos Part B Eng 2011;32(4):1990e9.
2015;69:542e9. [24] Chapple S, Anandjiwala R. Flammability of natural fiber-reinforced compos-
[5] Rwawiire S, et al. Development of a biocomposite based on green epoxy ites and strategies for fire retardancy: a review. J Thermoplast Compos Mater
polymer and natural cellulose fabric (bark cloth) for automotive instrument 2010;23(6):871e93.
panel applications. Compos Part B Eng 2015;81:149e57. [25] Marcovich NE, Reboredo MM, Aranguren MI. Modified woodflour as ther-
[6] Thakur VK, Thakur MK. Processing and characterization of natural cellulose moset fillers: II. Thermal degradation of woodflours and composites. Ther-
fibers/thermoset polymer composites. Carbohydr Polym 2014;109(0): mochim Acta 2001;372(1e2):45e57.
102e17. [26] Yang H, et al. Characteristics of hemicellulose, cellulose and lignin pyrolysis.
[7] Anbukarasi K, Kalaiselvam S. Study of effect of fibre volume and dimension on Fuel 2007;86(12):1781e8.
mechanical, thermal, and water absorption behaviour of luffa reinforced [27] Reti C, et al. Flammability properties of intumescent PLA including starch and
epoxy composites. Mater Des 2015;66(Part A):321e30. lignin. Polym Adv Technol 2008;19(6):628e35.
[8] Wen P, et al. One-pot synthesis of a novel s-triazine-based hyperbranched [28] Le Bras M, et al. Intumescent polypropylene/flax blends: a preliminary study.
charring foaming agent and its enhancement on flame retardancy and water Polym Degrad Stab 2005;88(1):80e4.
resistance of polypropylene. Polym Degrad Stab 2014;110(0):165e74. [29] Gao M, Wu W, Yan Y. Thermal degradation and flame retardancy of epoxy
[9] Fox DM, et al. Flame retarded poly(lactic acid) using POSS-modified cellulose. resins containing intumescent flame retardant. J Therm Analysis Calorim
2. Effects of intumescing flame retardant formulations on polymer degrada- 2009;95(2):605e8.
tion and composite physical properties. Polym Degrad Stab 2014;106(0): [30] Wu CS, et al. Thermal stability of epoxy resins containing flame retardant
54e62. components: an evaluation with thermogravimetric analysis. Polym Degrad
[10] Horacek H, Pieh S. The importance of intumescent systems for fire protection Stab 2002;78(1):41e8.
of plastic materials. Polym Int 2000;49(10):1106e14. [31] Zhang W, et al. The influence of the phosphorus-based flame retardant on the
[11] Jimenez M, Duquesne S, Bourbigot S. Intumescent fire protective coating: flame retardancy of the epoxy resins. Polym Degrad Stab 2014;109:209e17.
toward a better understanding of their mechanism of action. Thermochim [32] Deng C-L, et al. An intumescent flame retardant polypropylene system with
acta 2006;449(1):16e26. simultaneously improved flame retardancy and water resistance. Polym
[12] Zhang L, Li J, Ding X. Research of the properties of flame-retardant flexible Degrad Stab 2014;108:97e107.
PVC. Am J Mater Am J Mater Res 2014;1(1):20e5. [33] Shumao L, et al. Influence of ammonium polyphosphate on the flame
[13] Walker, K., A. Isarov, and T. Chen, Fire performance synergies of metal hy- retardancy and mechanical properties of ramie fiber-reinforced poly(lactic
droxides and metal molybdates in antimony-free flexible PVC. acid) biocomposites. Polym Int 2010;59(2):242e8.
[14] Arao Y, et al. Improvement on fire retardancy of wood flour/polypropylene [34] Zhang ZX, et al. Effect of flame retardants on mechanical properties, flam-
composites using various fire retardants. Polym Degrad Stab 2014;100(0): mability and foamability of PP/woodefiber composites. Compos Part B Eng
79e85. 2012;43(2):150e8.
[15] Castrovinci A, et al. Ammonium polyphosphateealuminum trihydroxide [35] De Chirico A, et al. Flame retardants for polypropylene based on lignin. Polym
antagonism in fire retarded butadieneestyrene block copolymer. Eur Polym J Degrad Stab 2003;79(1):139e45.
2005;41(9):2023e33. [36] Patrick Lim WK, et al. Effect of intumescent ammonium polyphosphate (APP)
[16] Gallo E, et al. Tailoring the flame retardant and mechanical performances of and melamine cyanurate (MC) on the properties of epoxy/glass fiber com-
natural fiber-reinforced biopolymer by multi-component laminate. Compos posites. Compos Part B Eng 2012;43(2):124e8.
Part B Eng 2013;44(1):112e9. [37] Hapuarachchi T, Peijs T. Aluminium trihydroxide in combination with
[17] Boccarusso, L., et al., Hemp fabric/epoxy composites manufactured by infusion ammonium polyphosphate as flame retardants for unsaturated polyester
process: improvement of fire properties promoted by Ammonium Poly- resin. eXPRESS Polym Lett 2009;3(11):743e51.
phosphate. Composites Part B: Engineering. [38] Kurt R, MENGELOGLU  F. Utilization of boron compounds as synergists with
[18] Lau K-t, Hoi-yan Cheung K, Hui D. Natural fiber composites. Compos Part B ammonium polyphosphate for flame retardant wood-polymer composites.
Eng 2009;40(7):591e3. Turkish J Agric For 2011;35(2):155e63.
[19] El-Sabbagh A, et al. Processing parameters and characterisation of flax fibre [39] Umemura T, et al. Synergy effects of wood flour and fire retardants in flam-
reinforced engineering plastic composites with flame retardant fillers. Com- mability of wood-plastic composites. Energy Procedia 2014;56(0):48e56.
pos Part B Eng 2014;62:12e8. [40] Petersen MR, et al. Mechanical properties of fire-retardant glass fiber-
[20] Jeencham R, Suppakarn N, Jarukumjorn K. Effect of flame retardants on flame reinforced polymer materials with alumina tri-hydrate filler. Compos Part B
retardant, mechanical, and thermal properties of sisal fiber/polypropylene Eng 2015;78(0):109e21.
composites. Compos Part B Eng 2014;56:249e53. [41] Ramazani SAA, et al. Investigation of flame retardancy and phys-
[21] Chen X, et al. New glass fiber/bismaleimide composites with significantly icalemechanical properties of zinc borate and aluminum hydroxide propyl-
improved flame retardancy, higher mechanical strength and lower dielectric ene composites. Mater Des 2008;29(5):1051e6.
loss. Compos Part B Eng 2015;71:96e102. [42] Lu M, Zhang S, Yu D. Study on poly(propylene)/ammonium polyphosphate
[22] Lee K-Y, et al. Hierarchical composites reinforced with robust short sisal fibre composites modified by ethylene-1-octene copolymer grafted with glycidyl
preforms utilising bacterial cellulose as binder. Compos Sci Technol methacrylate. J Appl Polym Sci 2004;93(1):412e9.