Composites Part B 95 (2016) 1e8

Contents lists available at ScienceDirect

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

Epoxy composites of reduced flammability

Rafal Oliwa*, Maciej Heneczkowski, Mariusz Oleksy, Henryk Galina
Department of Industrial and Materials Chemistry, Rzeszow University of Technology, Pl-35-959 Rzeszow, Poland

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

Article history: The effect of the so called environment friendly hybrid flammability reducing fillers, i.e. the halogen/
Received 21 September 2015 antimony-free ones containing phosphorous, nitrogen and silicon on the utility properties of carbon
Received in revised form reinforced epoxy laminates was studied. The total content of the fillers did not exceed 23 wt.%. The flame
16 February 2016
resistance of the laminates aimed as parts of aircraft construction was determined. The fire retardancy
Accepted 25 March 2016
mechanism of the hybrid fillers was investigated by SEM/EDS examination of partially burned laminates.
Available online 5 April 2016
The best combination of flammability reducing fillers for the epoxy composite matrices consisted of
15 wt.% ammonium polyphosphate, 5 wt.% dipentaerythritol, and 3 wt.% organofilized bentonite. The
Keywords:
A. Flame resistance
synergy of silicon and phosphorus in this composition provided material of UL-94 flammability class V-0.
A. Carbon fibre © 2016 Elsevier Ltd. All rights reserved.
A. Resins
A. Laminates
D. Thermal analysis

1. Introduction reduce flammability of various polymer based materials. These

The excellent strength to density ratio of fiber reinforced poly-
mer composites brought about wide application of these materials
in many branches of technology, in the aviation industry, in
particular. The contemporary planes built up from composites have
their weight reduced by 25e45% and, hence, they reach longer
distances having reduced fuel consumption. Epoxy laminates
reinforced with carbon fibers constitute an important group among
the composites used in the aviation industry [1]. The carbon fibers
are preferred for their high specific tensile modulus and lower
weight compared to glass fibers. The epoxy matrices, on the other
hand, are advantageous because of their high chemical and thermal
resistance and ease of processing. The main disadvantage of the
epoxies limiting their application in building transportation means
(aircrafts, boats, railway cars, buses, motorcars and other vehicles)
is the flammability of this material, often failing to meet safety
regulations [2]. Furthermore, the reinforcing fibers tend to exhibit a
candlewick effect by bringing the fuel into the burning material by
capillary forces thus accelerating the material thermal decompo-
sition and burning process [3]. The improvement of fire resistance
of the advanced composite laminates is therefore a crucial task.
Halogen fire retardants, chlorine and bromine compounds along
with antimony oxides were successfully used for many years to
* Corresponding author.
E-mail address: oliwa@prz.edu.pl (R. Oliwa).
compounds, however, are now replaced by new combinations of
products because in the case of fire the halogen based flame re-
tardants produced highly toxic fumes seriously endangering life or
health of fire casualties [4].
The contemporary efforts aiming at making polymer materials
less flammable focus on seeking for environment friendly and
effective flame retardants containing preferably such elements as
nitrogen, phosphorus, boron and silicon.
There are many phosphorous-containing flame retardants
described in the literature, including those introduced into epoxy
resin structure on the condensation stage through reactions of
epichlorohydrin with phosphorous oxides [5], phosphates [6], or
DOPO (i.e. 9,10-dihydro-9-oxa-10-phosphaphenantrene-10-oxide)
[7,8], by copolymerization of diglycidyl ether of bisphenol A, mostly
with DOPO derivatives [9,10], or by using phosphorous containing
amino curing agents [11e20].
Another group of phosphorous fire retardants constitute the
fillers added to the composite matrix polymer precursor and
thoroughly mixed with it before hardener is added. The most
commonly used products of this group are ammonium poly-
phosphate (APP) [21e29] and melamine polyphosphates (MP)
[22,23]. On fire, these compounds produce foamed coke of multi-
cellular structure by interacting with the intumescent component
of the system. Among these intumescent components there are
acidic or acid forming compounds, e.g. ammonium polyphosphate,
carbonizing agents, e.g. pentaerythritol, and gas releasing com-
pounds, e.g. melamine.

http://dx.doi.org/10.1016/j.compositesb.2016.03.074
1359-8368/© 2016 Elsevier Ltd. All rights reserved.
2 R. Oliwa et al. / Composites Part B 95 (2016) 1e8

There are numerous papers dealing with hybrid flame re-
tardants, the components of which would interact and better
protect materials by joint action. The most of these reports
concentrated on a synergetic effect obtained by modifying DOPO
with silanes or siloxanes [30e33] or by introducing both DOPO and
silicon-rich modifiers [34,35]. The latter solution, however, cause
some problems in the modified material processing.
In our studies on the effect of organophilized layered alumino-
silicates on the mechanical properties and flame resistance of
epoxy based glass or carbon fiber reinforced composites [36], we
attempted to reach a synergetic effect in reducing flammability of
these composites by modifying of the epoxy resin with a bentonite
treated with quaternary phosphonium salts and with commercial
flame retardants containing phosphorous and nitrogen.
2. Experimental section
2.1. Materials
The epoxy resin and the curing agent used in this work were
Epidian® 6 (EP6) containing mostly diglycidyl ether of bisphenol A
and triethylenetetramine (Z1) respectively, both commercial grade
products of Ciech-Sarzyna Plant, Poland. Halogen-free commercial
flame retardants used are listed in Table 1. Unidirectional carbon
fabric 200 g/m2 (TC) was purchased from Havel Composites. All
these materials were commercial grade and used without further
purification or modification. The organoclay, bentonite “Specjal”
(Zebiec SA Plants, Poland) was modified with a butyl-
triphenylphosphonium chloride (BSQPS, Xiamen Pioneer Technol-
ogy Inc., China).
2.2. Modification of bentonite with quaternary phosphonium salt
The modification procedure of smectic clays with quaternary
ammonium or phosphonium salts has been previously patented
[37,38] and described in detail in our previous papers [36,39]. It
consists of introducing aqueous solution of quaternary salts to a
suspension of a montmorillonite-rich bentonite in water, removal
of supernatant liquid, drying, grinding and sieving modified
Table 1
Epoxy compositions with flame retardants.
Composition Percentages of component of epoxy compositions [%]
bentonite to obtain organophilized layered aluminosilicate (BSQPS)
of grain sizes below 60 mm.
2.3. Preparation of epoxy compositions
In our previous research on the applications of organophilized
aluminosilicates we have found these products to substantially
reduce flammability of polymers when used as fillers and act as
environment friendly flame retardants [39]. Epoxy resin composi-
tions were prepared containing 3 wt.% of bentonite modified with
quaternary phosphonium salt and commercial flame retardants so
that total amount of filler did not exceed 23 wt.%. Preparation of the
compositions started with introducing modified bentonite to the
resin and homogenizing the mixture. Then, the remaining flame
retardants were introduced and the composition homogenized.
After introducing each component, the four-stage homogenization
procedure was applied [37]. In the choice of type and amount of
fillers, suggestions of producers were taken into account. The
mechanisms of flame retardancy of each component were also
considered in order to find possibly a synergetic effect. The com-
positions selected are listed in Table 1.
2.4. Preparation of carbon fiber/epoxy compositions laminates
The epoxy composition based on EP6, containing mixtures of
flame retardants as shown in Table 1, were used to prepare four-
layer laminates with unidirectional carbon fabric (200 g/m2)
acting as reinforcement. The contact method was used. Successive
sheets of fabric were laid on a Teflon film and impregnated with
epoxy composition containing amine curing agent using a grooved
roller. The top layer, after removing the excess of resin, was covered
with another Teflon film before the sample was placed between
two steel plates of dimensions 200 ! 300 mm which acted as a
mold. The plates were pressed to exert a pressure of about 10 MPa.
The laminates were left to cure at room temperature overnight and
then post-cured in an oven with hot air circulation at 100 " C for 6 h.
The resulting laminates contained ca. 40% of carbon fabric by
weight. The samples were cut from the laminates in the direction
parallel to the direction of fibers with an oscillating cutting disk.
The samples were used in flammability tests as well as to measure
their thermal properties.
2.5. Characterization of carbon fibre/epoxy compositions laminates

BSQPSa APPb DPERc MPYPd TEPe EGf ATHg The thermal decomposition of laminates was followed in a
termogravimeter TG 209 (Netzsch, Germany) in temperature range
K0 3 e e e e e e
K1 3 10 10 e e e e
0e800 " C with heating rate 10 deg/min.
K2 3 15 5 e e e e The limiting oxygen index (LOI) for the samples was determined
K3 3 e 10 10 e e e according the standard EN ISO 4589-3 at room temperature using
K4 3 e 5 15 e e e an instrument of Fire Testing Technology Ltd. (United Kingdom).
K5 3 10 5 5 e e e
The flammability tests by using UL-94 method were carried out
K6 3 5 5 10 e e e
K7 3 7.5 5 7.5 e e e in a chamber produced by FTT Ltd. (United Kingdom). The mea-
K8 3 e 10 e 10 e e surements were made according to standard PN-EN 60695-11-10
K9 3 e 5 e 15 e e with vertical sample beam position and methane fed burner of
K10 3 e e e 15 5 e 25 mm height.
K11 3 15 e e 5 e e
K12 3 15 e e e 3 e
The heat release rate (HRR in kW/m2) during sample combus-
K13 3 e e e e e 20 tion as well as other parameter characterizing flammability were
K14 e 17 6 e e e e evaluated for composite plates 100 ! 100 ! 2 mm in size using a
K15 e e 6 17 e e e cone microcalorimeter, a product of FTT Ltd. (United Kingdom),
a
Bentonite modified with buthythriphenylphosphonium chloride [36]. according to standard ISO 13927, by applying the heat flow 25 kW/
b
Aluminum trihydroxide, APYRAL 200 SM, (Nabaltec, Germany). m2 and the distance from ignition source 25 mm.
c
Melamine polyphosphate, Ecoret MPP (Alwernia Poland). To verify the assessments of flame resistance of laboratory
d
Amonium polyphosphate, (Tecnosintesi S.p.A., Italy).
e
Dipentaerythritol, (SigmaeAldrich).
samples made by specialized instruments, direct flammability tests
f
Triethyl orthophosphate, (SigmaeAldrich). were made for reinforcing plates of unmanned plane (DRON) sta-
g
Expanded graphite EG096, (Sinograf, Poland). bilizer. The plates were formed in a vacuum mold from carbon
 R. Oliwa et al. / Composites Part B 95 (2016) 1e8
fabric reinforced laminates by using selected epoxy compositions
presented in Table 2. The test was made using methane fed Mecker
burner providing a 50 mm high flame. The first ignition lasted for
10 s. If the ignited sample had failed to burn, the time of burning
was recorded and the sample ignited for the second time with same
burner for next 10 s. Time of ignition was recorded again after
removing the source of fire. The sample was then weighed and
examined visually.
The samples of partially burned composites were examined by a
scanning electron microscope (SEM, Hitachi S-3400N) equipped
with an elemental microanalyser EDS (energy dispersive spectros-
copy). Detectors of secondary (SE) and back-scattered (BSE) elec-
trons were used at accelerating voltage of 15 kV and spot size below
10 mm.
3. Results and discussion
3.1. Effectiveness of bentonite modification
In order to verify the effectiveness of bentonite modification the
modified samples were compared with unmodified ones by using
differential scanning calorimetry (DSC) Mettler Toledo DSC 822e
and wide angle X-ray scattering (WAXS) Bruker Nanostar. The re-
sults are described in detail in our previous paper [36].
3.2. Limiting oxygen index (LOI)
As can be seen in Fig. 1, the presence of hybrid flame retardants
significantly improves the oxygen indices of laminates. The highest,
almost doubled increase of LOI was found for the sample prepared
using composition K2, containing, by weight, 15% of ammonium
polyphosphate (APP), 5% of dipentaerythritol (DPER) and 3% of
bentonite Specjal modified with quaternary phosphonium salt
(BSQPS). A clear synergetic effect was observed between the com-
mercial flame retardants and the modified aluminosilicate, as the
laminate with K0 matrix, with only a small amount of BSQPS (3 wt.%)
and that with K14 matrix, containing no bentonite had their LOI
higher only by 4 and 11% of O2 compared to the reference sample,
respectively (Fig. 1, EP6). On the other hand, although melamine
polyphosphate (MPYP) belongs to the category of intumescent flame
retardants, similarly as APP, the laminates with K3 and K4 matrices
where APP was replaced by MPYP, had significantly lower LOI values,
namely 27 and 31% O2, respectively. Among the laminates with
matrices containing other hybrid flame retardant systems, the
lowest increase of LOI was noted for those with K8 and K9 matrices
containing triethyl phosphate (TEP), DPER, and BSQPS.
Table 2
The results of UL-94 test for carbon reinforced laminates.
Matrix Burning time, t1 (s) Burning time, t2 (s)
K0 e e
K1 0 10.0
K2 0 2.3
K3 32 e e
K4 5.5 e e
K5 0 5.3
K6 1.4 6.8
K7 0 7.4
K8 e e
K9 e e
K10 e e
K11 0 5.1
K12 0 19.7
K13 e e
K14 0 18.8
K15 e e
3
Fig. 1. Limiting oxygen indices (LOI) of laminates reinforced with unidirectional carbon
fabric prepared using epoxy matrices listed in Table 1.
3.3. Flame resistance by UL-94 method
In the UL-94 flammability test two quantities were determined:
total burning time tc, i.e. the sum of times t1 and t2 of burning the
sample after first and second ignition, respectively, and the mass
loss of the sample. In case the sample all burned out, the material
was classified as the flammable one (Table 2).
The results presented in Table 2 indicate that V-0 flammability
class had only the laminates prepared from compositions con-
taining ammonium polyphosphate. The lowest burned mass was
recorded for the laminate containing composition K2, in agreement
with its LOI. It is worth noticing that laminates prepared from K4
matrix containing 15% of MPYP, 5% of DPER and 3% of BSQPS and
having high LOI value (over 30%) in the vertical UL-94 test did not
perform well and could not be classifies as fire resistant. The reason
might be the poorer thermal stability of ammonium polyphosphate
(APP) compared to that of melamine polyphosphate (MPYP). The
former starts to decompose at 275 " C, whereas the latter, at above
330 " C.
The earlier decomposition of APP flame retardant seems to be
advantageous in this case. The char that forms is reach in phos-
phorous acid. The acid reacts with dipentaerythritol thus
improving thermal stability of the char. Furthermore, released
ammonia acts as the foaming agent and additionally dilutes flam-
mable gases.
Good flame resistance had also the laminates prepared using
matrix K11, containing 15% APP, 5% TEP, and 3% BSQPS. Although LOI
Weight loss (%) Flamability class
e Flammable
2.86 V-0
0.76 V-0
Flammable
Flammable
0.72 V-0
1.18 V-0
0.97 V-0
e Flammable
e Flammable
e Flammable
1.05 V-0
0.98 V-1
e Flammable
2.88 V-1
e Flammable
4 R. Oliwa et al. / Composites Part B 95 (2016) 1e8

Table 3 measured for these laminates reached only 28%, in UL-94 test they
The results of flammability test in the cone microcalorimeter. had flammability class V-0 and low burning times t1 and t2.
Parametersa Matrix On the other hand, the laminates with matrices (K3, K4,
EP6 K0 K2 K14
K8eK10) containing TEP or MPYP (10e15%) turned out to be
2
flammable. Matrices containing slightly less of these flame retar-
HRR [kW/m ] 130 82.3 57.9 66.8
dant, but also APP reached the flammability class V-1 or V-0. These
pHRR [kW/m2] 322 187 122 163
THR [MJ/m2] 18.2 16.7 16.2 15.7 flame retardants have to be considered as the additives acting
TTI [s] 130 130 155 136 synergistically. Particularly useful is TEP which is a liquid reducing
PML [%] 67.2 66.3 51.6 53.2 viscosity of epoxy compositions.
MLR [g/m2s] 0.093 0.066 0.137 0.200
The results also show that replacement of a part of commercial
tpHRR [s] 135 156 220 150
EHC [MJ/kg] 12.3 8.94 9.87 10.6 flame retardants with 3 wt.% of bentonite modified with quaternary
a
phosphonium salt significantly reduces flammability of laminates.
HRR e average heat release rate; pHRR e peak (highest) heat release rate; THR e
total heat released; TTI e time to ignition; PML e percent of mass loss; MLR e mass
Although both the carbon fabric reinforced laminates prepared
loss rate; tpHRR e time to peak heat release rate; EHC e effective heat of using K15 or K2 matrices ceased to burn immediately after
combustion. removing burner, after second ignition, time t2 for the sample with
K14 matrix was by 16 s longer than that for the laminate with K2
matrix containing modified bentonite. Hence the former was
classified as V-1, only. The longer burning times caused also four
times larger mass loss of the sample with K14 matrix, compared to
that with K2 matrix (Table 2).

3.4. Cone microcalorimeter

Selected laminates were tested using a cone microcalorimeter in

Fig. 2. The heat release rate (HRR) curves recorded in a cone microcalorimeter for
carbon fabric reinforced laminates prepared using plain epoxy resin (EP6) and the
resins K0, K2, and K14 modified with flame retardants as shown in Table 1.
order to verify the synergetic action of flame retardants. Since the
best system so far was that with epoxy matrix containing 15% APP,
5% DPER, and 3% BSQPS, the flammability of laminates with this
matrix K2 was examined in details, and compared with those made
using matrices K0 and K14, as well as for laminates prepared with
unmodified epoxy resin (EP6). The results, which are averages from
three runs for each sample, are presented in Table 3 and Fig. 2.
The curves presented in Fig. 2 clearly indicate the differences in
burning kinetics caused by the presence of various flame retardants
in the epoxy matrices. The carbon fabric reinforced laminates with
unmodified epoxy matrix burn quickly and vigorously yielding a
sharp and high peak of heat release rate (HRR). On the other hand,
the laminates with matrices containing hybrid systems of flame

Fig. 3. The photographs of samples after flammability test in cone microcalorimeter of carbon reinforced laminates with matrix: a) EP6, b) K2 and c) K14.
 R. Oliwa et al. / Composites Part B 95 (2016) 1e8 5

Fig. 4. A composite plate reinforced with carbon fiber with matrix: a) K2, b) K14, 10 s after removing the flame source.

retardants burn less violently, their HRR peaks are lower and effect consists of formation of a glassy sintered char in the burning
broader. The time of burning is also prolonged. The laminates with carbon laminate. Consequently, upon the second ignition, the
K0 matrix, containing just 3 wt.% of BSQPS burn almost as quickly as partially burned laminate carbon fabric/K2 matrix kept the flame
the laminates with pristine epoxy resin EP6, but the HRR peak is for only 16 s due to the presence of the sintered char of high
considerably lower than in previous cases with HRR and pHRR thermal stability (Fig. 4a), whereas the composites prepared
reduced by 37% and 42%, respectively. The highest reduction of HRR without the presence of bentonite burned completely (Fig. 4b,
and pHRR, by 55 and 65%, respectively, was observed for the lam- Table 4).
inates prepared using the epoxy matrix containing bentonite
Specjal modified with quaternary phosphonium salt (BSQPS, 3%),
beside the commercial flame retardants (Table 3). Moreover, the 3.6. Thermogravimetric analysis (TGA)
samples prepared using epoxy resin EP6 containing 15% APP, 5%
DPER, and 3% BSQPS had ignition time delayed by 25 s., i.e. 10 s. The thermogravimetric analysis carried out in nitrogen atmo-
longer compared to that for laminates with K14 matrix. The lami- sphere indicates a slight thermal stability decrease of laminates
nates with K0 matrix had ignition time similar to that for those prepared using the modified resin, compared to that prepared from
prepared with plain epoxy. The laminates with K2 and K14 matrices
had considerably reduced mass loss, by 23.5 and 20.8% compared to Table 4
mass loss recorded for laminates with plain EP6. These results The results of Mecker's burner flammability test.
confirm the synergetic effect of the flame retardants used and
Matrix Burning time, t1 [s] Burning time t2 [s] Mass loss [%]
support the conclusions following from LOI and UL-94 experi-
ments. The samples became porous upon burning (Fig. 3). This EP6 e e e
K0 e e e
confirms the intumescent effect caused by releasing gases and the
K2 10 16 0.94
fact that the effect is not hindered by the presence of modified K14 80 e e
bentonite. On the contrary, it helps to form highly porous char. The
effect might be related to different viscosity characteristics of
matrices having the dominant role in build up of porous structure
morphology described also in Zhang et al. papers [31e34] and
carbonization agents action presented in the Bugajny's paper [40].

3.5. Direct flammability test with Mecker's burner

In order to verify the laboratory tests, we have directly exam-

ined flammability of an unmanned plane stabilizer fins (Fig. 4)
made of carbon fabric reinforced with plain epoxy resin EP6,
composition K2, or composition K14. The elements were directly
ignited with a Mecker's burner for 10 s and, if the flame died out,
reignited for another 10 s. The time of burning t1 and t2 were
measured after the first, and second, ignition, respectively. The
results are shown in Table 4.
The results of the Mecker burner test confirm again the syner-
getic effect of the commercial flame retardants and the layered Fig. 5. DTG curves recorded under nitrogen for laminates reinforced with unidirec-
aluminosilicate modified with phosphonium salt. Apparently, the tional carbon fabric prepared using matrix EP6, K0, K2, or K14.
6 R. Oliwa et al. / Composites Part B 95 (2016) 1e8

Table 5 by 10 " C higher than that for the sample prepared using matrix K14.
The results of thermogravimetric analysis under nitrogen. This extra presence of silicon moved up also T5% of K2 laminate,
Parameters measureda Matrix compared to that of K14 one, by 5 " C. The TGA results confirm again
EP K0 K2 K14
the conclusion about the synergetic effect of phosphorous and
silicon in formation of a sintered char improving its thermal sta-
T5% [ C]
"
356.0 344.0 300.6 310.7
bility at elevated temperature. Still, T50% temperatures for the
T50% [" C] 412.2 388.6 405.7 395.8
Tmax [" C] 384.6 370.5 351.8 347.2 composites with K2 or K14 matrices were lower than the value for
Mass loss at Tmax [%] 47.4 32.4 50.3 52.9 laminates with unmodified matrix. This confirms formation of
Residue [%] at 750 " C 30.0 32.0 39.0 38.0 gaseous phosphorus products that on the early decomposition
a
T5%, T50%, and Tmax are the temperatures at 5%, 50%, and at the highest mass loss stages may hinder the flame [41].
rate, respectively. TGA results are consistent with those obtained from the cone
microcalorimeter measurements. Despite of slightly more rapid
decomposition, the laminates prepared using matrices modified
plain epoxy resin. The temperature at which the mass loss of the with both commercial and aluminosilicate flame retardants left
sample reached 5 wt.% (T5%) was in each case lower for the of behind more of solid residue.
modified resins, as compared to the laminate obtained from plain
epoxy (Fig. 5). The lowest decrease of T5% (12 " C) was recorded for 3.7. SEM/EDS analysis of partially burned laminates
the matrix containing only the modified aluminosilicate (K0). The
laminate of flammability class V-0 prepared with matrix K2 started The SEM/EDS analysis of partially burned carbon fabric rein-
to decompose at temperature T5% by 55 " C lower than the un- forced laminates with modified epoxy matrices reveals differences
modified one, but left the highest amount of residue at 750 " C equal in distribution of phosphorous and silicon elements in various parts
39% (Table 5). Clearly, the early start of degradation of this laminate of the char. An enhancement of silicon content was particularly
was due to the presence of thermally unstable ammonium poly- well seen for the sample containing 3 wt.% of BSQPS, where in area
phosphate. The latter decomposes yielding phosphorous-rich res- 2 (Fig. 6b), partially melted in high temperature flame, concentra-
idue which improves stability of the material at higher tion of silicon was 10 times higher than in area 1, more distant from
temperatures as compared to that of the unmodified laminate. This the flame. Also noticed was the doubled concentration of phos-
effect manifested itself particularly for the sample: carbon fabric/ phorous in area 2, compared to the average amount of this element,
K2 matrix, which on top of phosphorous and carbon contained also suggesting formation of silicon and phosphorous rich glassy sin-
silicon. Consequently, the temperature at 50% mass loss, T50%, was tered char (Table 6).

Fig. 6. SEM microphotographs of partly burned laminates with matrix: a) EP6, b) K0, c) K2, d) K14. EDS elemental analyses were made for areas marked 1 through 5.
 R. Oliwa et al. / Composites Part B 95 (2016) 1e8 7

Table 6 One of the major factors improving the flame resistance of the
The results of EDS element analysis of partly burned carbon fiber reinforced lami- laminates was formation of porous and tight char having improved
nates with matrices EP6, K0, K2, or K14.
thermal resistance caused by silicon accumulation as confirmed by
Matrix Number of Element contents [%] SEM/EDS analysis.
area in Fig. 6.
C N O Si P

EP6 1 92.03 3.83 4.14 e e Acknowledgments

2 89.31 4.89 5.80 e e
3 94.30 2.15 3.55 e e
Financial support from Structural Funds in the Operational
4 91.89 3.12 4.99 e e
K0 1 88.27 4.84 6.74 0.06 0.06 Program e Innovative Economy, European Regional Development
2 88.49 4.39 4.97 0.67 0.14 Fund - Project No. POIG.0101.02-00-015/08: Modern Materials
K2 1 57.51 11.37 25.28 0.33 5.38 Technologies for Aviation Industry is gratefully acknowledged.
2 78.63 7.61 11.31 0.17 2.22
3 81.07 e 15.45 0.38 3.05
4 84.79 e 12.70 0.29 2.22 References
5 72.94 e 21.50 0.28 5.28
K14 1 60.51 9.74 25.15 e 4.60 [1] Mouritz AP. Fire safety of advanced composites for aircraft. ATSB research and
2 78.81 e 17.64 e 3.55 analysis report. Australian Transport Safety Bureau. Australian; April, 2006.
3 71.13 e 23.41 e 5.00 [2] Innes JD. Flame retardants and their market applications. Flame retardants
4 88.59 e 9.83 e 1.58 101: basic dynamics, past efforts create future opportunities. Baltimore: Fire
Retardant Chemicals Association; 1996. p. 61.
[3] Zhao CS, Huang FL, Xiong WC, Wang YZ. A novel halogen-free flame retardant for
glass-fiber-reinforced poly(ethylene terephthalate). Polym Degrad Stabil
As a result of using the hybrid flame retardant systems in matrix 2008;93(6):1188e93. http://dx.doi.org/10.1016/j.polymdegradstab.2008.03.010.
K2, consisting of both modified aluminosilicate and commercial [4] Lu SY, Hamerton I. Recent developments in the chemistry of halogen-free
flame retardant polymers. Prog Polym Sci 2002;27:1661e712. http://
products, areas were detected on partially burned samples of the dx.doi.org/10.1016/S0079-6700(02)00018-7.
highest concentration of phosphorous ranging from 2.22 to [5] Jeng RJ, Shau SM, Lin JJ, Su WC, Chiu YS. Flame retardant epoxy polymers
5.38 wt.%. This high amount of phosphorous originated from both based on all phosphorus-containing components. Eur Polym J 2002;38:
683e93. http://dx.doi.org/10.1016/S0014-3057(01)00246-4.
ammonium polyphosphate and phosphonium salt used for
[6] Chen-Yang YW, Lee HF, Yuan CY. A flame-retardant phosphate and
bentonite modification (Fig. 6c, Table 6). This finding was particu- cyclotriphosphazene-containing epoxy resin: synthesis and properties.
larly important since the amount of phosphorous introduced to the J Polym Sci A Polym Chem 2000;38:972e81. http://dx.doi.org/10.1002/(SICI)
1099-0518(20000315)38:6<972::AID-POLA6>3.0.CO;2-N.
laminate with matrix K14 was higher than that with matrix K2.
[7] Liu YL. Phosphorous-containing epoxy resins from a novel synthesis route.
Replacement of 2% of APP with 3% of modified bentonite which J Appl Polym Sci 2002;83:1697e701. http://dx.doi.org/10.1002/app.10092.
contained only ca. 30% of phosphonium salt did not justified the [8] Meenakshi KS, Sudhan EPJ, Kumar AS, Umapathy MJ. Development and
observed phosphorous enrichment. Moreover, for the same sample characterization of novel DOPO based phosphorus tetraglycidyl epoxy nano-
composites for aerospace applications. Prog Org Coat 2011;72:402e9. http://
(laminate with K2 matrix) accumulation of silicon element was dx.doi.org/10.1016/j.porgcoat.2011.05.013.
observed. In area 3 (Fig. 6c), more liquefied due to flame proximity, [9] Wang CS, Lin CH. Synthesis and properties of phosphorus-containing epoxy
concentration of silicone was by 36% higher than in area no. 5, more resins by novel method. J Polym Sci A Polym Chem 1999;37:3903e9. http://
dx.doi.org/10.1002/(SICI)1099-0518(19991101)37:21<3903::AID-
remote from high temperature region. It is worth pointing out that POLA4>3.0.CO;2-X.
even a small amount of silicone on the level of ca. 0.3%, in [10] Perret B, Schartel B, Sto
€ ß K, Ciesielski M, Diederichs J, Do
€ ring M, et al. A new
connection with a substantial phosphorous enrichment, caused the halogen-free flame retardant based on 9,10-dihydro-9-oxa-10-
phosphaphenanthrene-10-oxide for epoxy resins and their carbon fiber
change of flammability class from V-1 to V-0 and increased of composites for the automotive and aviation industries. Macromol Mater Eng
limiting oxygen index from 29.2 to 33.0% in comparison to the 2011;296:14e30. http://dx.doi.org/10.1002/mame.201000242.
sample without silicone presence. [11] Artner J, Ciesielski M, Walter O, Do €ring M, Perez RM, Sandler JKW, et al.
A novel DOPO-based diamine as hardener and flame retardant for epoxy resin
The SEM/EDS analysis supports the previous conclusion about
systems. Macromol Mater Eng 2008;293:503e14. http://dx.doi.org/10.1002/
the importance of silicon accumulation caused by the flame. This mame.200700287.
accumulation in connection with the high temperature of burning [12] Schartel B, Braun U, Balabanovich AI, Artner JB, Ciesielski M, Do € ring M, et al.
Pyrolysis and fire behaviour of epoxy systems containing a novel 9,10-
laminate reaching 1500 " C and high carbon concentration may
dihydro-9-oxa-10-phosphaphenanthrene-10-oxide(DOPO)-based diamino
cause formation of silicone carbide [42e44]. The resulting com- hardener. Eur Polym J 2008;44:704e15. http://dx.doi.org/10.1016/
pounds improve the thermal resistance of material by forming a j.eurpolymj.2008.01.017.
tight and coherent char of good mechanical stability [45]. It also [13] Jirasutsakul I, Paosawatyanyong B, Bhanthumnavind W. Aromatic phosphor-
odiamidate curing agent for epoxy resin coating with flame-retarding prop-
limits delamination process that usually reduces the flame resis- erties. Prog Org Coat 2013;76:1738e46. http://dx.doi.org/10.1016/
tance of materials. j.porgcoat.2013.05.009.
[14] Toldy A, Szolnoki B, Marosi G. Flame retardancy of fibre-reinforced epoxy
resin composites for aerospace applications. Polym Degrad Stab 2011;96:
4. Conclusions 371e6. http://dx.doi.org/10.1016/j.polymdegradstab.2010.03.021.
[15] Perez RM, Sandler JKW, Altsta €d V, Hoffmann T, Pospiech D, Artner J, et al.
Novel phosphorus-containing hardeners with tailored chemical structures for
Among the flame resistance systems applied to protect carbon epoxy resins: synthesis and cured resin properties. J Appl Polym Sci 2007;105:
fabric reinforced epoxy laminates the best results (flammability 2744e59. http://dx.doi.org/10.1002/app.26537.
class V-0 according to UL-94 test) was obtained for the laminates [16] Preeti J, Choudhary V, Varma IK. Phosphorylated epoxy resin: effect of
phosphorus content on the. Properties of laminates. J Fire Sci 2003;21:5e16.
with the matrix composed of 77% of liquid epoxy resin EP6, 15% of http://dx.doi.org/10.1177/0734904103021001001.
APP, 5% of DPER and 3% of BSQPS cured with triethyltetramine. [17] Liu W, Varley RJ, Simon GP. Understanding the decomposition and fire per-
By replacing a part of commercial flame retardants with 3 wt.% formance processes in phosphorus and nanomodified high performance
epoxy resins and composites. Polymer 2007;4:2345e54. http://dx.doi.org/
of bentonite modified with quaternary phosphonium salt (BSQPS)
10.1016/j.polymer.2007.02.022.
the flame resistance of carbon reinforced laminates was signifi- [18] Gordon KL, Thompson CM, Lyon RE. Flame retardant epoxy resins containing
cantly improved. The BSQPS acted synergistically with the com- aromatic poly(phosphonamides). High Perform Polym 2010;22:945e58.
mercial flame retardants and changed the flammability class of the http://dx.doi.org/10.1177/0954008310363164.
[19] Agrawal S, Narula AK. Curing and thermal behaviour of a flame retardant
laminates from V-1 for the system carbon fabric/K14 matrix to V- cycloaliphatic epoxy resin based on phosphorus containing poly(-
0 for those with the matrix K2 (containing BSQPS). amideeimide)s. J Therm Anal Calorim 2014;115:1693e703. http://dx.doi.org/
8 R. Oliwa et al. / Composites Part B 95 (2016) 1e8
10.1007/s10973-013-3391-x.
[20] Wang CS, Shieh JY. Synthesis and properties of epoxy resins containing bis(3-
hydroxyphenyl) phenyl phosphate. Eur Polym J 2000;36:443e52. http://
dx.doi.org/10.1016/S0014-3057(99)00091-9.
[21] Wang X, Song L, Pornwannchai W, Hu Y, Kandola B. The effect of graphene
presence in flame retarded epoxy resin matrix on the mechanical and flam-
mability properties of glass fiber-reinforced composites. Compos Part A
2013;53:88e96. http://dx.doi.org/10.1016/j.compositesa.2013.05.017.
[22] Biswas B, Kandola BK. The effect of chemically reactive type flame retardant
additives on flammability of PES toughened epoxy resin and carbon fiber-
reinforced composites. Polym Adv Technol 2011;22:1192e204. http://
dx.doi.org/10.1002/pat.1960.
[23] Jiao C, Dong J, Chen X, Li S. Influence of T31 content on combustion and thermal
degradation behaviors on flame-retardant epoxy composites. J Therm Anal Cal-
orim 2013;114:1201e6. http://dx.doi.org/10.1007/s10973-013-3057-8.
[24] Liu L, Zhang Y, Lib L, Wang Z. Microencapsulated ammonium polyphosphate
with epoxy resin shell: preparation, characterization, and application in EP
system. Polym Adv Technol 2011;22:2403e8. http://dx.doi.org/10.1002/
pat.1776.
[25] Katsoulis C, Kandare E, Myler P, Kandola BK. Compos Part A 2012;43:
1389e99. http://dx.doi.org/10.1016/j.compositesa.2012.03.009.
[26] Tang Q, Wang B, Shi Y, Song L, Hu Y. Microencapsulated ammonium poly-
phosphate with glycidyl methacrylate shell: application to flame retardant
epoxy resin. Ind Eng Chem Res 2013;52:5640e7. http://dx.doi.org/10.1021/
ie302591r.
[27] Wang X, Hu Y, Song L, Xing W, Lu H, Lv P, et al. Effect of a triazine ring-
containing charring agent on fire retardancy and thermal degradation of
intumescent flame retardant epoxy resins. Polym Adv Technol 2011;22:
2480e7. http://dx.doi.org/10.1002/pat.1788.
[28] Wang JS, Liu Y, Zhao HB, Liu J, Wang DY, Song YP, et al. Metal compound-
enhanced flame retardancy of intumescent epoxy resins containing ammo-
nium polyphosphate. Polym Degrad Stab 2009;94:625e31. http://dx.doi.org/
10.1016/j.polymdegradstab.2009.01.006.
[29] Lim WKP, Mariatti M, Chow WS, Mar KT. Effect of intumescent ammonium
polyphosphate (APP) and melamine cyanurate (MC) on the properties of
epoxy/glass fiber composites. Compos Part B 2012;43:124e8. http://
dx.doi.org/10.1016/j.compositesb.2011.11.013.
[30] Yu D, Liu W, Liu Y. Synthesis, thermal properties, and flame retardance of
phosphorus-containing epoxy-silica hybrid resins. Polym Compos 2010;31:
334e9. http://dx.doi.org/10.1002/pc.20809.
[31] Zhang W, Li X, Yang R. Blowing-out effect and temperature profile in
condensed phase in flame retarding epoxy resins by phosphorus-containing
oligomeric silsesquioxane. Polym Adv Technol 2013;24:951e61. http://
dx.doi.org/10.1002/pat.3170.
View publication stats
[32] Zhang W, Li X, Yang R. Blowing-out effect in epoxy composites flame retarded
by DOPO-POSS and its correlation with amide curing agents. Polym Degrad
Stab 2012;97:1314e24. http://dx.doi.org/10.1016/
j.polymdegradstab.2012.05.020.
[33] Zhang W, Li X, Jiang Y, Yang R. Investigations of epoxy resins flame-retarded
by phenyl silsesquioxanes of cage and ladder structures. Polym Degrad Stab
2013;98:246e54. http://dx.doi.org/10.1016/j.polymdegradstab.2012.10.005.
[34] Zhang W, Li X, Yang R. The degradation and charring of flame retarded epoxy
resin during the combustion. J Appl Polym Sci 2013;130:4119e28. http://
dx.doi.org/10.1002/app.39689.
[35] Hsiue GH, Liu YL, Tsiao J. Phosphorus-containing epoxy resins for flame
retardancy V: synergistic effect of phosphorusesilicon on flame retardancy.
J Appl Polym Sci 2000;78:1e7. http://dx.doi.org/10.1002/1097-
4628(20001003)78:1<1::AID-APP10>3.0.CO;2e0.
[36] Oliwa R, Heneczkowski M, Oleksy M. Kompozyty epoksydowe do zastosowan "
w przemy"sle lotniczym. Polimery 2015;3:167e78. http://dx.doi.org/
10.14314/polimery.2015.167.
[37] PL395821Galina H, Heneczkowski M, Oleksy M, Oliwa R, Mossety-Leszczak B.
Kompozycja epoksydowa o zmniejszonej palno"sci i podwyzszonej_ odporno"sci
termicznej i sposo"b jej otrzymywania. 11.07.2013.
[38] PL397541 Galina H, Heneczkowski M, Oleksy M, Oliwa R, Marciniec B,
Maciejewski H, et al. Sposo " b otrzymywania samogasna˛ cych kompozyto "w
zywic
_ chemoutwardzalnych. 31.07.2014.
[39] Oleksy M, Oliwa R, Heneczkowski M, Mossety-Leszczak B, Galina H, Budzik G.
Kompozyty zywicy
_ epoksydowej z modyfikowanymi bentonitami dla potrzeb
przemysłu lotniczego. Polimery 2012;57:228e35.
[40] Bugajny M, Le Bras M, Noel A, Bourbigot S. Use of thermoplastic polyurethanes
as carbonisation agents in intumescent blends. Part 3: modification of the
dynamic properties of polypropylene/thermoplastic polyurethane/ammo-
nium polyphosphate formulations with heat and stress. J Fire Sci 2000;18:
104e29. http://dx.doi.org/10.1177/073490410001800203.
[41] Salmeia KA, Fage J, Liang S, Gaan S. An overview mode of action and analytical
methods for evaluation of gas phase activities of flame retardants. Polymers
2015;7:504e26. http://dx.doi.org/10.3390/polym7030504.
[42] Aseeva RM, Zaikov GE. Flammability of. Polymeric materials. Adv Polym Sci
1985;75:172e229.
[43] Hench LL, West JK. The sol-gel process. Chem Rev 1990;90:33e72. http://
dx.doi.org/10.1021/cr00099a003.
[44] Andrievski RA. Nano-sized silicon carbide: synthesis, structure and properties.
Rus Chem Rev 2009;78:821e31. http://dx.doi.org/10.1070/
RC2009v078n09ABEH0.04060.
[45] Levchik S, Wilkie ChA. Char formation. In: Grand AF, Wilkie ChA, editors. Fire
retardancy of polymeric materials. New York: Marcel Dekker Inc; 2000.