Enhanced fire retardant properties of glass-fiber reinforced Polyamide 6,6 by combining bulk and surface treatments: Toward a better understanding of the fire-retardant mechanism


Jimenez, M.; Duquesne, S.; Bourbigot, S.

Polymer Degradation and Stability 98(7): 1378-1388

2013


Polyamide 6,6 (PA6,6) is usually fire retarded in bulk, using aluminum diethylphosphinate (AlPi). In one of our recent papers, it was shown that good fire-retardant properties (e.g. V0 at the UL94 test) could also be achieved by applying an intumescent varnish on the PA6,6 surface. Excellent fire-retardant properties were even obtained combining 5% of AlPi in the bulk with an intumescent coating applied on the polymer surface. The Glow Wire Flammability Index (GWFI) test was validated at 960 C whereas it was only validated at 750 C without the AlPi. This paper first aims to describe the mechanism of action of an intumescent coating to protect a polymer, as it was never previously studied in the literature. It was evidenced, analyzing with 13 C, 31 P and 27 Al solid state NMR spectra of the materials at different combustion times that during burning one part of the virgin polymer begins to degrade and is mixed with the semi-viscous charring intumescent layer. In some way, due to this mixture of both phases, it is as if the polymer was fire retarded in bulk during burning. The second objective was to investigate the potential synergy between the gas phase fire-retardant mechanism of AlPi and the condensed phase fire protective mechanism of the intumescent coating to explain the enhanced fire-retardant properties. Similarly, using NMR technique, the interest to combine the bulk treatment in low amount and the intumescent coating was evidenced: the AlPi cannot completely sublimate because of the protective coating, and is probably condensed inside the intumescent structure pores. As it is trapped in the condensed phase, it then degrades into aluminophosphates, increasing the heat barrier efficiency of the expanded char layer.

Polymer
Degradation
and
Stability
98
(2013)
1378-1388
Polymer
Degradation
and
Stability
Contents
lists
available
at
SciVerse
ScienceDirect
PLSEVIFR
Polymer
Degradation
and
Stability
journal
homepage:
www.elsevier.com/locate/polydegstab
Enhanced
fire
retardant
properties
of
glass-fiber
reinforced
Polyamide
6,6
by
combining
bulk
and
surface
treatments:
Toward
a
better
understanding
of
the
fire-retardant
mechanism
M.
Jimenez*,
S.
Duquesne,
S.
Bourbigot
Unite
Materiaux
et
Transformations
(UMET),
equipe
Ingenierie
des
Systemes
Polymeres
(ISP),
CNRS-UMR
8207,
ENSCL,
Universite
Lille
Nord
de
France,
131:90108,
59652
Villeneuve
d'Ascq
cedex,
France
ARTICLE INFO
ABSTRACT
Article
history:
Received
20
December
2012
Received
in
revised
form
4
March
2013
Accepted
25
March
2013
Available
online
10
April
2013
Keywords:
Polyamide
6,6
Flame
retardancy
Intumescent
coating
Aluminum
diethylphosphinate
Surface
treatment
Mechanism
Polyamide
6,6
(PA6,6)
is
usually
fire
retarded
in
bulk,
using
aluminum
diethylphosphinate
(AlPi).
In
one
of
our
recent
papers,
it
was
shown
that
good
fire-retardant
properties
(e.g.
VO
at
the
UL94
test)
could
also
be
achieved
by
applying
an
intumescent
varnish
on
the
PA6,6
surface.
Excellent
fire-retardant
properties
were
even
obtained
combining
5%
of
AlPi
in
the
bulk
with
an
intumescent
coating
applied
on
the
polymer
surface.
The
Glow
Wire
Flammability
Index
(GWFI)
test
was
validated
at
960
°C
whereas
it
was
only
validated
at
750
°C
without
the
AlPi.
This
paper
first
aims
to
describe
the
mechanism
of
action
of
an
intumescent
coating
to
protect
a
polymer,
as
it
was
never
previously
studied
in
the
literature.
It
was
evidenced,
analyzing
with
13
C,
31
P
and
27
AI
solid
state
NMR
spectra
of
the
materials
at
different
com-
bustion
times
that
during
burning
one
part
of
the
virgin
polymer
begins
to
degrade
and
is
mixed
with
the
semi-viscous
charring
intumescent
layer.
In
some
way,
due
to
this
mixture
of
both
phases,
it
is
as
if
the
polymer
was
fire
retarded
in
bulk
during
burning.
The
second
objective
was
to
investigate
the
potential
synergy
between
the
gas
phase
fire-retardant
mechanism
of
AlPi
and
the
condensed
phase
fire
protective
mechanism
of
the
intumescent
coating
to
explain
the
enhanced
fire-retardant
properties.
Similarly,
using
NMR
technique,
the
interest
to
combine
the
bulk
treatment
in
low
amount
and
the
intumescent
coating
was
evidenced:
the
AlPi
cannot
completely
sublimate
because
of
the
protective
coating,
and
is
probably
condensed
inside
the
intumescent
structure
pores.
As
it
is
trapped
in
the
condensed
phase,
it
then
de-
grades
into
aluminophosphates,
increasing
the
heat
barrier
efficiency
of
the
expanded
char
layer.
©
2013
Elsevier
Ltd.
All
rights
reserved.
1.
Introduction
Polyamide
6,6
is
used
in
many
industrial
fields
such
as
housing
materials,
transport
or
electrical
engineering
applications
(connectors,
circuit breakers,
electric
distribution,
and
industrial
controls).
Due
to
its
chemical
composition,
this
polymer
is
easily
flammable.
As
far
as
the
main
applications
concern
electrical
and
electronic
equipment
(EEE),
its
flame
retardant
properties
become
an
important
requirement
and
the
materials
have
to
comply
with
severe
fire
retardant
test
methods.
It
includes
Limiting
Oxygen
Index
(LOI),
UL94
and
GWFI
(Glow
Wire
Flammability
Index).
The
aim
is
to
obtain
the
highest
LOI
values,
VO
rating
at
UL94
test
for
1.6
mm
or
0.8
mm
thick
samples
and
a
validation
at
960
°C
for
GWFI
test.
*
Corresponding
author.
Tel./fax:
+33
(0)320337196.
E-mail
addresses:
(M.
Jimenez),
ensc-lille.fr
(S.
Duquesne),
(S.
Bourbigot).
0141-39104
see
front
matter
©
2013
Elsevier
Ltd.
All
rights
reserved.
http://dx.doLorg/10.1016/j.polymdegradstab.2013.03.026
To
reduce
fire
hazards
of
polymeric
materials,
several
ways
can
be
considered:
the
most
common
one
consists
in
incorporating
flame
retardant
additives
into
the
polymer
[1].
This
technique
is
widely
used
and
cost
effective.
Numerous
flame
retardants
are
commercially
available
and
among
them
aluminum
dieth-
ylphosphinate
is
widely
used.
However,
this
approach
leads
to
some
processability
issues
and
to
the
decrease
of
the
intrinsic
properties
of
the
polymer
since
the
loading
is
usually
rather
high
(typically
more
than
15
wt.-%).
Moreover,
some
of
the
commonly
used
additives
(in
particular
some
halogenated-based
compounds)
can
have
ecological
drawbacks
[2-4].
Another
way
consists
in
applying
a
coating
on
the
polymer
surface
after
a
flame
pre-
treatment
[5];
however
the
GWFI
test
might
be
not
validated
in
that
case.
In
a
recent
previous
paper
[6],
an
intumescent
varnish
was
applied
on
a
Polyamide
6,6
(PA6,6)
plate
containing
glass
fi-
bers.
The
LOI
and
UL94
results
were
greatly
improved,
but
the
GWFI
test
did
not
reach
960
°C
for
the
GWFI
test.
A
concept
was
thus
developed,
consisting
in
combining
both
approaches:
a
bulk
treatment
at
low
amount
of
flame
retardants
and
an
intumescent
M.
Jimenez
et
at
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
1379
transparent
varnish.
With
only
5%
of
aluminum
dieth-
ylphosphinate
(AlPi)
in
the
bulk
and
100
gm
of
intumescent
coating
applied,
the
fire
retardant
properties
according
to
cone
calorimeter,
LOI
and
GWFI
were
improved
compared
to
the
use
of
23%
aluminum
diethylphosphinate
in
the
bulk
or
to
the
use
of
the
intumescent
coating
alone.
By
combining
5%
of
AlPi
in
bulk
and
the
100
gm
thick
intumescent
coating
(IC)
on
the
surface,
the
LOI
in-
creases
from
respectively
26
vol.-%
for
PA6,6-GF
+
5%
AlPi
and
52
vol.%
for
PA6,6-GF
+
intumescent
coating
to
63
vol.%
when
both
AlPi
and
the
intumescent
coating
are
combined.
Moreover,
the
sample
is
VO
rated
at
0.8
mm
and
validates
the
GWFI
test
at
960
°C
The
interest
of
this
approach
is
to
combine
two
modes
of
action:
a
condensed
phase
mechanism
through
the
formation
of
an
intu-
mescent
barrier
and
the
gas
phase
action
of
AlPi
which
releases
phosphorus-based
radicals
[7].
The
concept
consisting
in
combining
two
mechanisms
of
action
has
already
been
developed
by
Braun
et
al.
[8]
and
by
Gallo
et
al.
[9,10]
who
obtained
synergistic
effects
by
combining
in
Poly
(1-4
butylene
terephtalate)
(PBT)
AlPi
respectively
with
melamine
cyanurate
and
with
metal
oxides.
The
combination
of
AlPi
and
intumescent
fire
retardants
in
the
bulk
has
also
recently
been
reported
by
Yi
et
al.
[11]
who
incorporated
in
an
acrylonitrile-butadiene-styrene
copolymer
(ABS)
ammonium
pol-
yphosphate
and
other
additives
such
as
AlPi.
Again,
a
synergistic
effect
was
observed.
In
our
novel
approach,
we
however
apply
the
intumescent
fire
retardants
in
a
coating
on
the
sample
surface
and
not
in
the
bulk
as
in
the
previous
published
papers.
We
were
expecting
the
AlPi
ac-
tion
to
be
inhibited
by
the
intumescent
coating
since
the
presence
of
this
coating
should
limit
the
diffusion
of
the
radical
release
from
AlPi
to
the
flame,
but
it
still
seems
to
play
a
complementary
effect,
particularly
for
GWFI
test:
the
plate
containing
5%
AlPi
is
protected
even
at
960
°C
This
paper
proposes,
using
solid
state
NMR
analyses
at
particular
times
of
combustion
of
different
formulations,
to
explain
the
improvement
of
the
fire
protection
of
the
PA6,6-GF
when
both
intumescent
coating
and
aluminum
diethylphosphinate
are
com-
bined.
Solid
state
NMR
(
27
A1,
13
C,
31
P)
is
in
fact
the
most
accurate
tool
we
can
consider
to
determine
the
mechanisms
of
action
involved
in
the
pyrolysis
of
flame
retarded
polymers
containing
aluminum,
carbon
and/or
phosphorus,
as
it
was
proven
many
times
by
Bour-
bigot
et
al.
[12-16],
who
particularly
developed
the
use
of
27
A1
solid
state
NMR
on
fire
retardant
formulations
incorporating
zeolithes.
The
first
results
presented
in
the
paper
show
the
cone
calorim-
eter
curves
obtained
for
the
formulation
containing
only
AlPi
(PA6,6-AlPi),
for
the
formulation
covered
with
intumescent
coating
(PA6,6-IC)
and
for
the
formulation
combining
both
the
AlPi
(in
bulk)
and
the
intumescent
coating
(on
the
surface).
The
next
parts
respectively
detail
the
13
C,
31
P
and
27
A1
solid
state
NMR
spectra
obtained
for
these
different
formulations
before
burning
and
after
various
combustion
times.
These
results
are
finally
discussed
in
the
last
part
of
the
paper
and
the
mechanism
of
action
of
the
intumes-
cent
coating
on
a
polymer
plate,
as
well
as
the
combined
mecha-
nisms
of
action
of
AlPi
and
intumescent
coating
are
established.
2.
Materials
and
methods
2.1.
Materials
Polyamide
6,6
(PA6,6-GF)
has
been
supplied
by
Rhodia
Engi-
neering
Plastics
(Stabamid
27
AE
1).
Polymer
plates
(100
x
100
x
3
mm
3
,
127
x
12,6
x
1.6
mm
3
and
60
x
60
x
1
mm
3
)
have
been
prepared
with
25%
of
boron-free
glass
fibers
(Advantex
983,
Owens
Corning
Corporation)
via
melt
compounding
using
a
twin-screw
co-rotative
extruder
ZSK40
(rotating
speed
250
rpm,
injection
flow
rate
35
kg/h).
Based
on
a
previous
study
[6],
5%
of
a
fire
retardant
additive
has
been
incorporated
using
the
same
melt
compounding
process:
AlPi
(Exolit
OP1230,
Clariant).
The
intumescent
varnish
was
"Pyroplast
HW
100"
supplied
by
Rfitger
Organics.
It
is
a
transparent
varnish,
mainly
composed
of
carbon,
oxygen,
nitrogen,
phosphorus
and
silica
with
traces
of
metals
such
as
iron
and
magnesium
[6].
It
is
thus
halogen,
boron
and
heavy
metals
free.
2.2.
Surface
treatment:
intumescent
coating
The
waterborne
transparent
intumescent
varnish
has
been
applied
on
the
polymer
plates
after
a
flame
treatment
(IPROS
flame
apparatus).
It
is
possible
on
this
apparatus
to
modify
the
number
of
passes,
the
speed
under
the
flame
and
the
distance
between
the
substrate
and
the
flame.
The
optimized
conditions
are
presented
in
Table
1.
It
was
also
applied
on
a
steel
plate
in
similar
treatment
conditions
in
order
to
analyze
the
behavior
of
the
coating
without
any
interference
with
the
polymer.
The
coating
has
been
sprayed
for
10
s
using
a
semi-industrial
spraying
machine
(CLID
apparatus).
The
thickness
of
the
varnish
measured
using
a
profilometer
Alphastep
IQ
(Scientec)
was
around
100
gm.
2.3.
Fire
tests
Cone
calorimeter
measurements
were
carried
out
using
a
Stanton
Redcroft
Cone
Calorimeter
according
to
the
ASTM
E
1354-
90a
standard
[17].
Samples
(standard
size:
100
x
100
x
3
mm
3
)
were
exposed
under
a
50
kW/m
2
external
heat
flux
which
repre-
sents
the
heat
flux
found
in
the
vicinity
of
solid-fuel
ignition
source.
Conventional
data
such
as
Rate
of
Heat
Release
(RHR)
and
time
to
ignition
(TTI)
can
then
be
obtained.
The
Glow
Wire
Flammability
Index
(GWFI)
was
determined
on
(60
x
60
x
1
mm
3
)
sheets
according
to
the
international
IEC
60695-
2
standard
[18].
The
glow
wire
is
heated
via
an
electrical
resistance.
A
test
specimen
is
held
for
30
s
against
the
tip
of
the
glow
wire
with
a
force
of
1
N.
The
product
is
characterized
by
the
highest
tem-
perature
(for
a
test
specimen),
at
which
one
of
the
following
con-
ditions
are
fulfilled:
flames
or
glowing
of
the
test
specimen
extinguish
within
30
s
after
removal
of
the
glow-wire
and
there
is
no
ignition
of
the
wrapping
tissue
placed
underneath
the
test
specimen;
or
there
is
no
ignition
at
all
of
the
test
specimen.
The
GWFI
test
is
used
to
simulate
the
effect
of
heat
that
might
arise
in
malfunctioning
electrical
equipment.
Test
results
provide
a
way
to
compare
the
ability
of
materials
to
extinguish
flames
and
their
ability
not
to
produce
particles
capable
of
spreading
fire.
2.4.
Solid
state
NMR
In
order
to
investigate
the
mechanism
of
action
of
the
intu-
mescent
paint,
the
cone
calorimetry
experiment
was
stopped
at
characteristic
steps
of
combustion
(Before
ignition,
after
ignition,
at
the
peak
of
Heat
Release
Rate
(pHRR),
before
extinguishment
and
at
the
end
of
the
combustion)
and
the
residues
collected
were
analyzed
by
solid
state
NMR.
This
experiment
is
thereafter
called
"shutter
test".
Table
'1
Optimized
flame
treatment
conditions
[6].
Sample
Number
Treatment
Distance
Adhesion Adhesion
after
of
flaming
velocity
flame-
before
coating
(100
gm)
passes
(mm/s)
substrate
coating
(ASTM
D3359-B)
(cm)
(ASTM
D3359-B)
PA66/PA66
GF
3
200
8
3B
5B
1380
M.
Jimenez
et
at
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
The
samples
were
analyzed
by
31
P,
27
A1
and
13
C
solid
state
NMR
before
ignition,
after
ignition,
at
the
maximum
of
the
peak
of
RHR,
before
extinguishment
and/or
at
the
end
of
the
combustion.
31
1
3
NMR
measurements
were
performed
using
a
Bruker
Avance
II
400
spectrometer
(static
field
9.4
T)
operating
at
a
Larmor
fre-
quency
of
79.22
MHz.
Bruker
probe
heads
equipped
with
4
mm
MAS
(Magic
Angle
Spinning)
assembly
were
used
and
the
spinning
rate
of
the
rotor
was
12.5
kHz.
31
P
NMR
measurements
were
carried
out
using
MAS
(10
kHz)
with
1
H
dipolar
decoupling
(DD)
and
Cross
Polarization
(CP).
A
relaxation
delay
of
5
s
was
used
for
all
samples.
Between
128
and
8192
scans
were
necessary
to
obtain
spectra
with
good
signal
to
noise
ratio.
H3PO4
in
aqueous
solution
(85%)
was
used
as
reference.
Depending
on
the
number
of
the
bridging
oxygen
the
phosphate
tetrahedra
can
be
described
as
Q,
where
i
represents
the
number
of
the
bridging
oxygen
and
can
have
a
value
of
0,1,
2,
3.
Such
notation
was
first
used
by
Lippmaa
[19]
and
is
widely
accepted
in
the
literature
for
describing
structures
of
phosphates
and
sili-
cates.
A
framework
built
up
with
Q
3
denotes
a
fully
condensed
structure,
while
Q2
unit
gives
only
a
two-dimensional
structure
based
on
chains
or
rings.
Q
means
two
tetrahedral
connected
by
a
corner
and
Qo
means
isolated
tetrahedra.
Qo
units
are
for
example
characteristic
for
orthophosphate
structures.
13
C
NMR
measurements
were
performed
on
the
same
spec-
trometer
as
above
mentioned
with
MAS
at
10
kHz,
DD
and
CP.
A
relaxation
delay
of
5
s
was
used
for
all
the
samples.
Typically,
1024
scans
were
necessary
to
obtain
spectra
with
good
signal
to
noise
ratio.
Tetramethylsilane
was
used
as
reference.
27
A1
NMR
measurements
were
performed
using
a
Bruker
Avance
II
800
spectrometer.
Bruker
probe
heads
equipped
with
3.2
mm
MAS
(Magic
Angle
Spinning)
assembly
were
used.
27
A1
measure-
ments
were
carried
out
using
MAS
(12.5
kHz)
with
1
H
decoupling
(DD)
and
cross-polarization
(CP).
A
relaxation
delay
of
900
s
was
used
for
all
samples.
256
scans
were
necessary
to
obtain
spectra
with
good
signal
to
noise
ratio.
The
reference
used
was
1
M
solution
of
aluminum
nitrate.
3.
Results
In
a
precedent
paper
[6],
we
showed
that
by
combining
5%
AlPi
and
the
intumescent
coating,
UL94
and
GWFI
test
results
were
improved
in
a
great
manner
(Table
2).
When
5%
AlPi
are
added
in
the
PA6,6-GF,
GWFI
is
validated
at
850
°C
but
the
sample
is
not
rated
at
the
UL-94
test.
When
PA6,6-GF
is
coated
with
the
intumescent
coating,
the
GWFI
is
only
validated
at
750
°C
but
the
sample
is
VO
rated.
The
combination
of
5%
AlPi
and
the
intumescent
coating
allows
validating
the
GWFI
at
960
°C
and
obtaining
a
VO
rating
at
UL94
test.
The
samples
were
also
tested
to
cone
calorimeter.
In
our
previous
paper
the
HRR
curves
obtained
during
cone
calorimeter
test
for
the
PA6,6-GF,
PA6,6-GF
containing
5%
AlPi,
PA6,6-GF
covered
with
100
gm
intumescent
varnish
and
PA6,6-GF
containing
the
5%
AlPi
and
covered
with
the
100
gm
intumescent
varnish
were
compared
[6].
The
use
of
5%
AlPi
in
PA6,6-GF
leads
to
a
decrease
of
the
pHRR
of
the
material
of
around
40%.
On
the
other
hand,
the
100
gm
thick
intumescent
coating
allows
reducing
drastically
the
RI-IR.
Indeed,
Table
2
Fire
test
results
on
the
different
formulations.
UL-94
GWFI
(CC)
PA6,6-GF
PA6,6-GF
+
5%
AlPi
PA6,6-GF
+
intumescent
coating
PA6,6-GF
+
5%
AlPi
+
intumescent
coating
NC
NC
VO
VO
850
750
960
pRHR
decreases
by
around
80%
compared
to
the
PA6,6-GF.
When
both
AlPi
and
the
coating
are
combined,
a
curve
similar
to
the
one
obtained
for
coated
PA6,6-GF
is
obtained,
with
only
a
slight
improvement
(decrease
of
pHRR
of
30
kW/m
2
compared
to
coated
PA6,6
GF).
The
HRR
of
coated
PA6,6-GF
5%
AlPi
shows
a
maximum
value
around
100
kW/m
2
after
500
s.
The
TTI
also
increases:
ignition
is
observed
at
around
220
s
for
coated
samples
versus
54
s
for
non
coated
material
loaded
with
5%
of
AlPi
and
90
s
for
pure
PA6,6-GF.
It
was
thus
demonstrated
through
different
fire
test
results
(L01,
UL94,
GWFI,
Cone
calorimeter)
that
by
combining
the
bulk
treat-
ment
and
the
surface
treatment
a
synergistic
effect
is
obtained.
Thus,
the
aim
of
the
following
part
of
the
paper
is
to
compare
the
mechanisms
of
action
of
AlPi
in
PA6,6-GF,
of
the
intumescent
coating
applied
on
PA6,6
GF
and
of
the
combination
of
AlPi
and
of
the
intumescent
coating
(IC)
in
the
PA6,6,
in
order
to
explain
the
improvement
of
the
fire
protective
effect.
To
do
that,
cone
calorimeter
experiments
were
stopped
at
different
characteristic
steps
of
combustion
and
the
residues
were
then
analyzed
by
13
C,
31
P
and
27
A1
solid
state
NMR.
When
the
intumescent
coating
is
applied
on
the
polymer
plate,
two
parts
can
be
distinguished
before
and
during
burning:
the
polymer
plate
and
the
coating
part.
This
is
why
the
NMR
analyses
of
these
two
parts,
entitled
"plate"
and
"surface
layer"
for
each
shutter
test
will
be
separately
presented
for
the
samples
PA6,6-GF/IC
and
PA6,6-GF/5%
AlPi/IC.
3.1.
13
C
NMR
analyses
Fig.
1
shows
the
13
C
NMR
analyses
at
characteristic
steps
of
combustion
of
the
PA6,6-GF
filled
with
5%
AlPi.
The
spectrum
of
the
virgin
material
is
similar
to
the
spectrum
before
ignition,
so
it
is
not
presented
here.
On
the
spectrum
before
ignition,
the
carbonyl
functions
and
the
aliphatic
carbons
of
PA6,6
are
respectively
observed
at
175
ppm
(Cl)
and
in
the
range
of
15-50
ppm
(C2)
[20].
The
signal
at
8
ppm
(C3)
is
attributed
to
the
CH3
groups
of
the
phosphinates
[7].
The
CH
2
groups
of
the
phosphinates
are
usually
detected
at
20
ppm
(C4),
however
this
peak
is
hindered
by
the
massif
of
peaks
corre-
sponding
to
the
aliphatic
groups
of
the
PA6,6.
During
its
combustion,
the
PA6,6
leads
to
an
extensive
degra-
dation,
resulting
in
the
almost
complete
disappearance
of
carbonyl
resonance
at
175
ppm
[21].
Only
a
broad
massif
of
peaks
centred
around
40
ppm
(C2)
remains
due
to
aliphatic
fragments,
as
well
as
Cl
C5
C2
C4
C3
End
of
combustion
pRHR
Before
ignition
200
100
0
-100
Chemical
shift
(ppm)
Fig
1.
CPMAS
13
C
solid
state
NMR
of
PA6,6-GF-5%
AlPi
at
characteristic
steps
of
the
combustion
process.
End
of
combustion
pRHR
After
ignition
Before
ignition
(a)
End
of
combustion
pRHR
After
ignition
Before
ignition
100
-100
(b)
100
C7
M.
Jimenez
et
al.
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
1381
C6
End
of
combustion
pRHR
Aker
Ignition
Before
ignition
oo
200
100
0
IPPB0
Chemical
shift
(ppm)
Fig.
2.
CPMAS
13
C
NMR
of
the
intumescent
coating
at
characteristic
steps
of
the
combustion
process.
the
small
peak
(C3)
characteristic
of
phosphinates.
This
peak
de-
creases
but
it
has
not
completely
disappeared
at
the
peak
of
RHR.
What
is
interesting
to
mention
is
that
the
spectra
before
and
after
ignition,
until
pHRR,
are
quasi
similar,
meaning
that
the
structure
is
not
modified,
which
confirms
the
gas
phase
action
mechanism
of
phosphinate
[22-24].
For
the
residue
obtained
at
the
end
of
the
combustion,
only
one
broad
band
of
low
intensity
is
recorded
at
130
ppm
(C5).
This
peak
is
characteristic
of
sp2
aromatic
carbons
observed
when
char
is
formed.
The
low-intensity
of
the
13
C
NMR
signal
has
already
been
observed
and
is
generally
attributed
to
radicals
trapped
in
an
aro-
matic
carbonaceous
structure
preventing
the
detection
of
the
car-
bon
species.
Ganapathy
[25]
has
linked
the
decrease
of
the
signal
with
the
paramagnetic
susceptibility
of
aromatic
compounds.
This
relation
has
then
been
confirmed
in
other
fire
retardant
systems
by
electron
paramagnetic
resonance
[26,27].
Some
char
remains
pre-
sent
but
is
however
not
or
slightly
detected
by
13
C
NMR.
To
have
a
better
idea
of
the
behavior
of
the
coating
without
any
interference
with
the
polymer
plate,
the
intumescent
coating
was
applied
on
a
steel
plate
and
tested
under
the
cone
calorimeter
at
characteristic
times.
Fig.
2
shows
the
13
C
NMR
spectra
of
the
intu-
mescent
coating
applied
on
a
steel
plate
and
analyzed
at
charac-
teristic
steps
of
the
combustion.
The
intumescent
coating
has
an
unknown
composition,
it
is
thus
difficult
to
interpret
exactly
the
spectrum
obtained
before
ignition.
However,
the
peak
at
160
ppm
(C6)
is
characteristic
of
the
carbonyl
functions
of
the
resin
used
(probably
an
acrylic
resin).
In
the
massif
of
peaks
between
40
and
80
ppm,
we
assume
that
aliphatic
carbons
are
detected
(signal
named
C7),
but
no
difference
will
be
made
in
the
paper
between
the
carbons
of
the
resin
and
the
carbons
of
the
(unknown)
fire
retardant
fillers
incorporated
in
the
resin
(e.g.
pentaerythritol,
etc.).
At
the
opposite
of
that
observed
previously,
the
spectrum
of
the
material
after
ignition
is
very
different
compare
to
the
one
before
ignition,
demonstrating
a
chemical
modification
of
the
structure
of
the
coating.
Thus
it
is
observed
that
the
intumescent
system
degrades
rapidly
and
forms
a
protective
char
upon
heating,
thus
protecting
the
underlying
substrate.
According
to
the
13
C
spectra
collected
during
combustion
(after
ignition
and
at
pHRR),
only
broad
signals
with
low
intensity
are
observed
around
130
ppm
(C5
signal
cor-
responding
to
sp2
aromatic
carbons)
and
around
40
ppm
(signal
C2),
probably
corresponding
to
aliphatic
carbons
(as
in
the
PA6,6-
GF
degradation
process).
At
the
end
of
combustion,
no
signal
is
detected.
However,
char
formation
is
observed.
The
same
expla-
nation
as
above
is
given
for
this
absence
of
signal.
Fig.
3
shows
the
13
C
NMR
spectra
obtained
at
different
times
of
combustion
for
PA6,6-GF
covered
with
the
intumescent
coating.
Both
the
plate
(Fig.
3a)
and
the
surface
layer
corresponding
to
the
intumescent
coating
(Fig.
3b)
were
analyzed.
Before
ignition,
in
the
plate
(Fig.
3a),
only
the
peaks
character-
istic
of
PA6,6
are
observed
(peak
Cl
of
carbonyl
function
resonance
and
collection
of
bands
C2
of
aliphatic
carbons).
After
ignition,
the
PA6,6-GF
degrades
and
the
Cl
signal
decreases
until
complete
disappearance
at
the
end
of
combustion.
The
C5
signal
appears,
corresponding
to
the
formation
of
aromatic
compounds.
Indeed,
it
is
well
established
that
PA6,6
is
a
charring
material
and
thus
partially
condensed
when
degraded.
The
C2
signal
in
the
plate
layer
surprisingly
does
not
disappear
and
remains
quite
constant
until
the
end
of
the
combustion.
Only
a
slight
broadening
is
observed.
This
is
an
important
change
compared
to
the
usual
spectra
of
C5
C2
C7
Cl
C5
C2
C6
C5
C2
Chemical
shift
(ppm)
Chemical
shift
(ppm)
Fig.
3.
CPMAS
13
C
NMR
of
the
PA6,6-GFIIC
at
characteristic
steps
of
the
combustion
process
:
(a)
plate
and
(b)
surface
layer
(*spinning
side
bands).
1382
M.
Jimenez
et
al.
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
degradation
of
PA6,6-GF
(not
shown
in
the
paper),
in
which
both
the
carbonyl
and
aliphatic
signals
disappear
around
the
maximum
of
HRR.
At
the
end
of
the
combustion,
the
C5
signal,
characteristic
of
the
PA6,6
carbonization
is
still
clearly
visible.
The
C2
signal,
char-
acteristic
of
aliphatic
carbons,
is
still
present
but
is
broader,
meaning
that
the
aliphatic
chains
of
PA6,6-GF
partially
remain
at
the
end
of
the
experiment.
This
clearly
shows
that
the
intumescent
coating
plays
its
barrier
effect,
protecting
the
underlying
material
from
a
complete
degradation.
In
the
surface
layer
(Fig.
3b),
only
the
characteristic
signals
of
the
intumescent
coating
are
observed
(signals
C6
and
C7).
A
similar
behavior
to
the
one
observed
analyzing
the
IC
on
the
steel
plate
is
observed.
It
could
however
be
noted
that
after
ignition,
the
degradation
of
the
IC
is
not
total,
at
the
opposite
to
what
observed
with
a
steel
substrate.
Moreover,
at
the
end
of
the
test,
some
signals
are
detected
whereas
it
was
not
the
case
with
the
steel
substrate.
It
could
thus
be
concluded
that
the
substrate
partially
interferes
within
the
intumescent
process
of
the
coating.
Fig.
4
compares
the
spectra
obtained
during
combustion
of
PA6,6
GF
containing
5%
AlPi
and
coated
with
the
intumescent
coating.
The
aim
is
to
examine
the
potential
influence
of
AlPi
on
the
degradation
of
the
materials.
Again,
the
plate
and
surface
were
analyzed
separately
and
the
results
are
respectively
presented
in
Fig.
4a
and
b.
Before
ignition,
the
spectrum
of
the
plate
(Fig.
4a)
is
logically
the
same
as
that
obtained
for
the
formulation
PA6,6-GF
AlPi.
The
Cl
(carbonyl),
C2
(aliphatic)
and
C3
(CH3
phosphinates)
signals
are
identified,
the
C4
signal
(CH
2
phosphinates)
being
hindered
by
the
C2
signal.
When burning,
the
Cl
peak
completely
disappears
and
the
C5
signal
(aromatic
species)
increases
after
ignition
and
remains
pre-
sent
until
the
end
of
the
combustion.
The
collection
of
bonds
C2
also
remains
present
until
the
end
of
combustion
and
the
C3
signal,
characteristic
of
the
phosphinates
remains
present
until
the
end
of
the
combustion.
At
the
end
of
combustion,
C2
and
C5
signals
can
also
be
observed.
It
is
thus
observed
that
the
main
difference
when
the
spectra
of
PA6,6-GF
AlPi
and
PA6,6
GF
AlPi
IC
are
compared
is
the
presence
of
the
C3
signal
until
the
end
of
the
experiment
and
the
maintaining
of
the
aliphatic
part
of
PA6,6
as
observed
when
PA6,6-GF
was
coated
with
IC.
It
thus
suggests
that
AlPi
is
still
trapped
in
the
partially
degraded
polymer
at
the
end
of
the
experiment,
proving
again
that
the
intumescent
coating
prevents
the
complete
polymer
degradation.
Dealing
with
the
analyses
of
the
surface
(Fig.
4b),
the
spectrum
obtained
before
ignition
is
the
same
as
that
obtained
for
the
intu-
mescent
coating.
The
peak
of
C6
(carbonyl)
and
C7
collection
of
bonds
(aliphatic)
can
be
distinguished
on
the
spectra.
During
burning,
the
C2
(aromatics)
and
C5
(aliphatic
fragments)
signals
appear
and
remain
until
the
end
of
combustion.
No
particular
dif-
ferences
are
visible
compared
to
the
PA6,6-GF
IC
spectra
presented
in
Fig.
2b
except
at
the
end
of
combustion:
the
spectrum
is
of
lower
resolution.
A
hypothesis
could
be
that
the
paramagnetic
suscepti-
bility
of
the
compounds
formed
is
increased
when
both
AlPi
and
IC
are
combined.
It
was
thus
demonstrated
that
when
the
aluminum
phosphinate
and
the
intumescent
coating
are
combined,
no
particular
effects
on
the
13
C
NMR
spectra
are
visible.
The
polymer
plate
is
not
completely
degraded
after
combustion,
this
being
due
to
the
barrier
effect
of
the
intumescent
coating.
At
the
end
of
the
combustion,
some
phosphinates
also
remain
in
the
bulk,
probably
in
the
less
degraded
part
of
the
polymer.
Another
hypothesis
is
possible:
the
sublimation
of
phosphinates
could
be
prevented
by
the
protective
coating,
thus
favoring
the
condensation
of
these
species
in
the
pores
of
the
structure.
This
will
be
discussed
later
looking
at
the
31
P
and
27
A1
NMR
results.
32.
31
P
solid
state
NMR
analyses
The
different
samples
were
also
analyzed
by
31
P
NMR.
Fig.
5
presents
the
spectra
obtained
for
the
formulation
containing
PA6,6-GF
and
5%
AlPi.
The
analysis
has
been
carried
out
in
the
bulk
of
the
sample
at
characteristic
combustion
times.
Before
ignition,
the
spectrum
presents
one
multiplet
between
42
and
38
ppm
(P1
and
P2
signals).
These
signals
are
characteristic
of
Q0
sites
in
phosphinates
(CH3—CH2—P(0)—O—)
and
are
shifted
because
of
Al
in
the
second
sphere
of
coordination
of
P.
The
same
signals,
of
lower
intensity,
can
be
observed
after
ignition
and
at
the
pHRR.
At
the
end
of
combustion
the
residue
of
the
polymer
plate
ex-
hibits
a
broad
band
at
around
—30
ppm
(P3
signal)
characteristic
of
Q
3
sites,
which
can
correspond
to
PO
4
units
in
aluminophosphate
structures.
Indeed,
it
was
previously
reported
in
the
literature
that
(a)
C1
CS
C2
C4
C3
End
of
combustion
pRHR
After
ignition
(b)
v\INA/
14
C5
C2
End
of
combustion
Before
extinguishment
pRHR
C7
C6
After
ignition
Before
ignition
Before
ignition
100
•100
Chemical
shift
(ppm)
160
100
60
.60
Chemical
shift
(ppm)
Fig.
4.
CPMAS
13
C
NMR
of
the
PA6,6-GF/5%
AlPi/IC
at
characteristic
steps
of
the
combustion
process
:
(a)
bulk
and
(b)
surface.
M.
Jimenez
et
al.
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
1383
P3
P1
P2
End
of
combustion
pRHR
After
ignition
Before
Ignition
*
100
60
0
-50
Chemical
shift
(ppm)
Fig.
5.
CPMAS
31
P
solid
state
NMR
of
PA6,6-GF-5%
AlPi
at
characteristic
steps
of
the
combustion
process
(*Spinning
side
bands).
degradation
of
AlPi
leads
to
the
formation
of
aluminophosphate
171.
This
will
be
confirmed
analyzing
the
27
A1
spectra
in
the
corre-
sponding
part
of
the
paper.
The
intumescent
coating
was
applied
on
a
steel
plate
and
tested
under
the
cone
calorimeter
at
characteristic
times.
The
31
P
NMR
spectra
of
the
intumescent
coating
(not
shown
in
the
paper)
applied
on
a
steel
plate
were
obtained
at
characteristic
steps
of
the
combustion.
The
virgin
intumescent
coating
is
characterized
by
one
peak
at
around
5
ppm
(P4
signal)
characteristic
of
Qo
sites
in
orthophosphate
groups
128].
During
burning
an
additional
band
appears
around
—8
ppm
(P5
signal):
this
band
can
be
attributed
to
Qi
sites
corresponding
to
pyrophosphate
groups
linked
with
alkyl
groups
129].
The
coated
plate
was
then
investigated
by
31
P
NMR
at
the
same
characteristic
times
of
combustion.
Fig.
6
presents
the
31
P
NMR
analyses
of
the
formulation
PA6,6-GF
IC
at
different
steps
of
the
combustion
in
the
bulk
(Fig.
6a)
and
in
the
surface
layer
(Fig.
6b).
The
P4
signal
is
not
detected
before
ignition
in
the
plate
(no
31
P
signal),
whereas
it
logically
appears
in
the
coating
(Fig.
6b).
How-
ever,
after
ignition,
this
band
(P4),
as
well
as
the
band
at
-8ppm
(P5),
characteristics
of
the
degradation
of
the
intumescent
coating,
are
detected
both
in
the
polymer
plate
and
in
the
char
of
the
intumescent
coating.
At
the
end
of
the
combustion,
an
additional
signal
(P6)
at
0
ppm
appears
in
the
bulk
(but
not
in
the
coating
char),
indicating
the
formation
of
Qo
sites
characteristic
of
phosphoric
acid
which
might
have
been
formed
by
char
hydro-
lysis
during
the
waiting
time
between
cone
calorimeter
experiment
and
NMR
analysis.
During
burning,
it
seems
that
the
products
of
degradation
of
the
coating
are
mixed
or
diffuse
within
the
products
of
degradation
of
the
polymer.
This
point
will
be
further
discussed
in
the
last
part
of
the
paper.
Fig.
7
shows
the
spectra
at
characteristic
times
of
the
combus-
tion
process
of
the
material
combining
PA6,6-GF,
5%
AlPi
and
the
intumescent
coating.
Again,
the
analyses
have
been
done
on
the
bulk
(polymer
plate)
and
the
surface
(intumescent
coating)
of
the
material.
In
the
bulk
figure
(Fig.7a),
the
spectrum
of
AlPi
is
pre-
sented
as
reference.
P1
and
P2
signals
characteristic
of
aluminum
dieth-
ylphosphinate
(38
and
42
ppm)
are
detected
in
the
polymer
plate
(Fig.
7a)
after
ignition
and
at
pRHR,
but
never
in
the
surface
layer.
As
in
the
case
of
the
PA66/GF/AlPi,
these
peaks
disappear
at
the
end
of
combustion.
The
peak
characteristic
of
the
intumescent
coating
(P4)
around
3ppm
and
characteristic
of
its
degradation
(P5)
at
—8
ppm
appear
as
well
in
the
plate
after
ignition,
confirming
pre-
vious
observations
of
possible
diffusion
or
mixture
of
both
phases
during
burning
(Fig.
6a).
An
additional
peak
(P3)
appears
during
combustion
at
—25
ppm,
corresponding
to
the
formation
of
alu-
minophosphates
(tetra-coordinated
phosphorous)
probably
in
an
amorphous
structure
(broad
band),
as
observed
at
the
end
of
combustion
for
PA6,6-GF-5%
AlPi.
At
the
end
of
combustion,
only
very
low
intense
signals
between
0
and
—30
ppm
are
detected,
which
might
possibly
correspond
to
orthophosphates
or
poly-
phosphoric
acid.
In
the
surface
layer,
the
signal
characteristic
of
Qo
species
in
the
intumescent
coating
is
detected
before
ignition
at
5
ppm.
After
ignition,
the
signal
at
—8
ppm
appears,
corresponding
to
Qi
species.
An
additional
signal
(P3)
appears
around
—25
ppm,
which
did
not
appear
in
the
Fig.
6b,
corresponding
to
the
formation
alumi-
nophosphates
(as
in
the
bulk
part).
This
peak
disappears
at
the
end
of
the
combustion
and
only
Qo
sites
at
0
ppm,
characteristic
of
phos-
phoric
acid,
and
Qi
sites,
characteristic
of
pyrophosphates,
remain.
13.
27
A1
solid
state
NMR
analyses
The
27
Al
spectrum
of
AlPi
(not
in
the
paper),
shows
at
ambient
temperature
one
peak
at
—12
ppm.
This
signal
can
be
attributed
to
A1
3
+
cations
having
an
octahedral
geometry
and
phosphorous
atoms
in
their
second
sphere
of
coordination
130].
(a)
P5
P4
P6 P5
(b)
End
of
combustion
pRHR
End
of
combustion
p
RHR
After
ignition
After
ignition
No
signal
before
ignition
Chemical
shift
(ppm)
*
Before
ignition
efore
ignition
100
50
0
-50
-100
Chemical
shift
(ppm)
Fig.
6.
CPMAS
31
P
solid
state
NMR
of
PA6,6-GFIIC
at
characteristic
steps
of
the
combustion
process:
plate
(a)
and
char
(b).
1384
M.
Jimenez
et
al.
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
(a)
P4
P5
P3
: .
End
of
combustion
(b)
P1
P2
Max.
RHR
After
ignition
P4P6
PS
P3
Ei
End
of
combustion
Mu.
NH*
OP1230
Int
coating
Before
ignition
103
Chemical
shift
10:I
co
Chemical
shift
•10
4i0
(ppm)
(ppm)
Fig.
7.
CPMAS
31
P
solid
state
NMR
of
PA6,6-GF15%
AINIC
at
characteristic
steps
of
the
combustion
process
(a)
bulk
and
(b)
surface
layer.
Fig.
8
compares
the
27
A1
NMR
spectra
of
the
PA6,6/GF/5%
AlPi
formulation
analyzed
at
characteristic
combustion
steps.
The
27
A1
spectrum
of
the
sample
before
ignition
presents
two
bands
assigned
Al
(58
ppm)
and
A2
(-12
ppm).
The
band
Al
is
attributed
to
the
glass
fibers
present
in
the
sample
(tetrahedral
A104
units
in
a
silicate
network).
The
peak
A2
at
—12
ppm
is
char-
acteristic
of
aluminum
phosphinates
and
is
attributed
to
Al
3
+
cat-
ions
of
AlPi
having
an
octahedral
geometry
1301.
At
higher
temperatures,
the
relative
amount
of
octahedrally
coordinated
Al
species
decreases,
and
thus
the
relative
amount
of
A10
4
groups
increases.
At
the
end
of
combustion,
no
A106
groups
remain,
and
the
dominating
signal
is
the
signal
Al,
corresponding
to
tetrahe-
drally
coordinated
A104
units
in
a
silicate
network.
This
signal,
in
this
context,
could
also
be
attributed
to
A10
4
units
in
a
highly
condensed
amorphous
aluminum
phosphate
network
[7].
Furthermore,
a
minor
band
A3
at
about
40
ppm
is
observed
at
the
end
of
the
combustion.
This
chemical
shift
is
typical
for
A1PO
4
structure,
where
the
tetrahedrally
coordinated
Al
is
coordinated
with
phosphate
groups
sharing
the
oxygen.
This
apparition
of
an
A3
signal
is
in
accordance
with
the
31
P
lineshape
(Fig.
5):
at
the
end
of
Ail
A3
A2
End
of
combustion
Max.
RHR
Before
ignition
100
0
-100
ppm
Chemical
shift
(ppm)
Fig.
8.
CPMAS-DD
27
A1
solid
state
NMR
of
PA6,6-GF-5%
AlPi
at
characteristic
steps
of
the
combustion
process.
the
combustion,
a
broad
peak
between
0
and
-30
ppm
character-
istic
of
aluminophosphates
is
observed.
Our
results
corroborate
the
mechanism
of
degradation
of
pure
AlPi
established
in
the
literature
[16]
:
between
425
and
500
°C,
the
decomposition
of
aluminum
phosphinates
begins.
At
the
pHRR
three
signals
are
observed
at
37,
6
and
—16
ppm.
The
signal
at
—16
ppm
is
characteristic
of
octahedral
aluminum
(A106
units)
connected
by
A1-0—P
bonds.
The
central
band
at
6
ppm
is
attrib-
uted
to
pentahedral
aluminum
atoms.
The
last
band
observed
at
37
ppm
can
be
assigned
to
A10
4
units
in
an
aluminophosphate
structure.
At
800
°C,
a
single
peak
at
38.4
ppm
(corresponding
to
signal
Al)
is
observed
and
is
attributed
to
tetra-coordinated
aluminum
in
A1PO4.
Our
results
thus
confirm
the
literature
assumptions
that
report
the
formation
of
aluminum
phosphate
as
a
degradation
product
of
AlPi
116,271.
Fig.
9
presents
the
27
A1
NMR
spectra
for
the
formulation
con-
taining
AlPi
and
covered
with
the
intumescent
coating.
Fig.
9a
shows
the
analyses
of
the
bulk
material
and
Fig.
9b
shows
the
an-
alyses
of
the
surface
layer.
In
the
bulk,
similar
spectrum
is
obtained
than
the
one
obtained
for
the
formulation
PA6,6/GF/5%
AlPi.
A
higher
intensity
of
A3
assignment
at
the
pHRR
is
however
observed.
It
thus
demon-
strates
that
the
AlPi
cannot
completely
sublimate
because
of
the
protective
coating,
and
that
it
is
probably
condensed
inside
the
intumescent
structure
pores.
As
it
is
trapped
in
the
condensed
phase,
it
then
degrades
into
aluminophosphates.
At
the
end
of
the
combustion
however,
no
peaks
corresponding
to
A2
and
A3
are
detected.
It
is
surprising
not
to
detect
any
A3
signal
anymore
at
the
end
of
the
combustion.
However,
Al
signal
increases.
This
signal
can
partly
correspond
to
A104
units
in
a
highly
condensed
amorphous
aluminum
phosphate
network:
it
was
then
assumed
that
the
crystalline
structure
of
the
aluminophosphates
changed
during burning
and
that
they
are
still
present
at
the
end
of
the
combustion.
In
the
surface
layer
(intumescent
coating),
before
ignition,
no
signal
corresponding
to
aluminum
containing
species
is
detected.
This
confirms
that
the
intumescent
coating
does
not
contain
any
aluminum
or
only
little
traces
that
could
not
be
detected
using
this
technique.
At
pHRR,
a
low
intensity
band
Al
and
a
higher
intensity
band
A2
are
detected,
meaning
that
AlPi
and/or
its
degradation
products
diffuse
in
the
intumescent
coating.
This
hypothesis
is
Al
A3
A2
(a)
End
of
combustion
(b)
AlA3
A2
End
of
combustion
pRHR
Before
ignition
pRHR
A10
6
A104
in
Silicate
network
Less
A10
6
More
A10
4
Char
Traces
of
Al
[OP(0)(C
2
H
5
)
2
7
3
alumino-phosphates
Et
2
PO2
A104
H
NCO
M.
Jimenez
et
al.
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
1385
' I I
300
200
100
-100
-
200
ppm
300
200
100
-100
-200
PPm
Chemical
shift
(ppm)
Chemical
shift
(ppm)
Fig.
9.
CPMAS-DD
27
A1
solid
state
NMR
of
PA6,6-GF15%
AlPi/IC
at
characteristic
steps
of
the
combustion
process
(a)
bulk
and
(b)
surface
layer.
confirmed
by
the
fact
that
relative
higher
intensity
bands
A2
and
A3
are
observed
at
the
end
of
combustion,
respectively
assigned
to
A1PO
4
and
A106
structures.
These
two
peaks
were
not
observed
after
combustion
of
the
PA6,6/GF/AlPi
formulation,
as
AlPi
is
sub-
limated
in
the
gaseous
phase.
The
intumescent
coating
acts
as
a
barrier
for
these
gaseous
products,
which
are
trapped
in
the
char,
condensed
and
degraded.
4.
Discussion
In
flame
retardancy,
the
most
significant
chemical
reactions
interfering
with
the
combustion
process
can
take
place
in
the
condensed
and/or
gas
phases.
In
condensed
phase
different
types
of
reactions
can
take
place.
Among
them,
the
flame
retardant
can
lead
to
the
formation
of
an
isolating
carbon
layer
(char)
on
the
polymer
surface.
Polymers
flame
retarded
by
intumescence
is
essentially
a
special
case
of
a
condensed
phase
mechanism.
The
activity
in
this
case
occurs
in
the
condensed
phase
and
radical
trap
mechanism
in
the
gaseous
phase
appears
to
not
be
involved.
In
a
gas
phase,
such
as
reported
in
the
case
of
the
AlPi,
the
radical
mechanism
takes
place:
the
highly
energetic
radicals
(W,
OW)
are
trapped
by
different
radicals
(e.g.
PO")
in
the
case
of
phosphorus
based
compounds.
The
exothermic
processes
which
occur
in
the
flame
are
thus
stopped,
the
system
cools
down,
the
supply
of
flammable
gases
is
reduced
and
eventually
completely
suppressed.
The
idea
in
this
paper
was
thus
to
combine
a
mechanism
in
condensed
phase
(intumescent
coating)
and
a
mechanism
in
gaseous
phase
(AlPi).
First
of
all,
the
mechanisms
of
both
systems
were
investigated
separately.
The
last
part
of
the
discussion
pro-
poses
a
possible
mechanism
of
action
when
the
AlPi
is
combined
with
the
intumescent
coating.
The
mechanism
of
action
of
aluminum
diethylphosphinate
in
PA6,6-GF
has
been
already
detailed
by
U.
Braun
et
al.
17].
They
established
that
both
thermal
and
hydrolytic
polymer
de-
compositions
took
place
during
burning,
and
that
the
phosphinates
vaporize
during
both
steps
of
degradation
of
the
PA6,6-GF,
without
showing
particular
interaction
in
condensed
phase
with
the
PA6,6-
GF.
Our
NMR
results
are
consistent
with
this
decomposition
model.
At
the
end
of
the
combustion,
the
13
C
NMR
confirms
that
a
char
is
obtained.
31
P
NMR
shows
the
degradation
of
the
phosphinates
during
combustion,
with
the
apparition
of
some
alumi-
nophosphates
in
amorphous
form
at
the
end
of
the
combustion.
The
27
A1
spectrum
confirms
these
observations,
with
a
very
low
signal
corresponding
to
aluminophosphates
appearing
at
the
end
of
the
combustion.
Fig.
10
summarizes
the
decomposition
mechanism
of
the
PA6,6-GF
AlPi.
1/2
CO
2
NH
3
wCN
H
o
tr'R
AT
H
+
H
2
O
(no
residue)
CO
2
H
NH
3
Fig.
10.
Decomposition
mechanism
of
PA66-GF
5%
AlPi.
1386
M.
Jimenez
et
al.
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
On
the
other
hand,
it
was
shown
that
when
PA6,6-GF
plate
was
coated
with
the
intumescent
coating,
a
protection
of
the
substrate
occurs.
Indeed,
13
C
NMR
analyses
show
that
PA6,6
is
partially
degraded
during
the
cone
calorimeter
experiment
since
the
signals
corresponding
to
the
aliphatic
fragments
of
PA6,6-GF
are
retrieved
in
the
bulk
part
at
the
end
of
the
combustion,
which
is
not
the
case
for
non
protected
PA6,6-GF.
However,
the
signals
obtained
during
degradation
of
both
the
coating
and
the
polymer
plate
being
similar,
it
is
not
really
possible,
using
only
13
C
NMR,
to
determine
if
some
interactions
exist
between
the
polymer
and
the
coating.
On
the
other
hand,
31
P
NMR
demonstrates
that
the
intumescent
coating
and
the
polymer
are
"mixed"
together
during
burning,
since
the
phosphorus
species
are
detected
both
in
the
bulk
and
in
the
surface
layer.
Indeed,
the
31
P
NMR
signals
of
the
coated
PA6,6
during
burning
present
the
same
signals
(Q.0
and
Qi
sites)
for
both
the
bulk
and
the
surface.
In
that
case,
as
the
virgin
PA6,6-GF
does
not
contain
any
phosphorus,
it
is
clear
that
during
burning
one
part
of
the
virgin
polymer
begins
to
degrade
and
mix
with
the
semi-
viscous
intumescent
system.
In
some
way,
it
may
be
assumed
that
the
polymer
is
fire
retarded
in
bulk
thanks
to
the
intumescent
coating
applied
on
it.
31
P
NMR
analyses
also
give
some
clues
about
the
mechanism
of
degradation
of
the
intumescent
coating.
Indeed,
even
if
the
IC
is
a
commercial
product,
it
was
demonstrated
that
when
degraded
formation
of
ortho
and
pyrophosphate
(both
aliphatic
and
aromatic)
should
be
suspected.
This
is
often
observed
when
studying
degradation
of
intumescent
coatings
1311.
Fig.
11
summarizes
the
supposed
mechanism
of
action
of
the
intumescent
coating
on
the
PA6,6-GF.
Finally,
the
main
objective
of
the
paper
was
to
explain
the
improvement
of
the
fire
retardant
properties
of
the
material
when
both
the
bulk
and
the
surface
treatments
are
combined.
Indeed,
in
that
case,
the
GWFI
test
shows
great
improvement:
GWFI
is
vali-
dated
at
750
°C
for
PA6,6-GF
IC,
at
850
°C
for
PA6,6
GF
5%
AlPi
whereas
it
is
validated
at
960
°C
when
both
approaches
are
com-
bined.
The
AlPi
clearly
has
an
influence
on
the
GWFI
test
as
with
only
5%
AlPi
in
polymer
bulk
the
GWFI
is
validated
at
850
°C
These
differences
can
first
be
explained
by
the
test
configuration:
the
test
specimen
is
held
for
30
s
against
the
tip
of
the
glow
wire
with
a
force
of
1
N.
It
means
that
the
specimen
penetrates
inside
the
intumescent
coating
layer
and
when
it
has
crossed
this
barrier
it
penetrates
the
polymer.
If
the
polymer
is
not
intrinsically
protected
it
will
burn
and
the
intumescent
coating
will
no
longer
be
efficient.
This
is
a
first
explanation
but
we
wanted
to
demonstrate
if
inter-
action
between
IC
and
AlPi
occurs
or
not.
That
is
the
reason
why
the
chemistry
of
the
system
was
investigated
during
combustion.
The
13
C
NMR
analyses
show
that
similarly
to
what
observed
for
PA6,6-GF
coated
with
the
intumescent
coating,
a
protective
behavior
of
the
intumescent
coating
is
obtained.
Indeed,
the
PA6,6-
GF
also
partially
degrades
during
the
cone
calorimeter
experiment
since
aliphatic
structures
are
maintained
in
the
material,
up
to
the
end
of
the
test.
The
31
P
NMR
spectra
however
show
some
differences
in
pres-
ence
of
AlPi:
the
peaks
characteristic
of
AlPi
are
as
expected
observed
in
the
bulk
material
but
an
additional
peak
appears
after
ignition
at
—30
ppm,
corresponding
to
the
presence
of
alumino-
phosphates.
This
peak
only
appeared
at
the
end
of
the
combus-
tion
in
the
PA6,6-GF
5%
AlPi
spectrum.
It
could
thus
be
assumed
that
the
presence
of
the
intumescent
coating
leads
to
the
conden-
sation
of
the
sublimated
AlPi
in
the
pores
of
the
structure
that
allow
its
further
degradation
up
to
the
formation
of
aluminophosphates.
This
phenomenon
is
confirmed
by
the
27
A1
spectra:
at
the
end
of
combustion,
signals
corresponding
to
A1PO4
structures
are
observed
in
the
bulk.
Moreover,
it
is
interesting
to
observe
that
after
ignition,
peaks
characteristic
of
the
31
P
species
coming
from
the
intumescent
coating
are
observed
in
the
bulk
materials.
It
thus
confirms
that
similarly
to
what
observed
when
PA6,6-GF
is
used
as
a
substrate
both
materials
are
mixed
together
when
burning.
The
hypothesis
we
thus
make
is
that
the
intumescent
coating
acts
as
a
barrier
for
the
gaseous
products
produced
from
the
degradation
of
AlPi.
Aluminophosphates
are
formed
during
burning,
these
species
migrate
and
are
trapped
inside
the
coating
char.
The
presence
of
aluminophosphates
stabilizes
the
char
and
might
allow
a
better
protection
of
the
underlying
substrate.
Fig.
12
summarizes
the
different
mechanisms
involved.
The
interest
to
combine
this
kind
of
bulk
treatment
and
the
intumescent
coating
would
then
be
double:
the
gases
released
by
orthophosphates
pyrophosphates
II
II
orthophosphates
0
1 1
7 7
R
2
-0—
I
I
—O—R
1
R.
R
0
4
I
pyrophosph
t
tes
0
R
3
-
0 -
P
- 0
.-
P
- 0
-
R
3
R
orthophosphates
.
R4
II
R
2
-0—P—O—R
1
II
R.
Aliphatic
carbons
R
2
-0—P—O—R
1
Carbonyl
Aromatic
carbons
A104
A106
(phosphinates)
0
Carbonyl
Aliphatic
carbons
Aromatic
carbons
R,
A104
Ft,
Aliphatic
carbons
Aluminophosphates
ignition
111MirliEFTTligiN1
PA6.6
GF
swelling
-
mor
pyrophosphates
A10
6
(phosphinates)
III
orthophosphates
R
2
-0—P—CY—P-0-1:1
1
0
orthophosphates
11
2
0
P
O-R,
A10
4
in
a
Silicate
network
I I
O
0
Aliphatic
carbons
R
3
Aromatic
carbons
R,
114
Carbonyl
R
I
s
Aliphatic
carbons
A104
A104
Aromatic
carbons
A106
(phosphinates)
Aluminophosphates
Fig.
11.
Mechanism
of
action
of
the
intumescent
coating
on
the
PA6,6-GF.
O—R
i
0
P
R
3
orthophosphates
0
II
0
R
3
pyrophosphates
o
o
II
II
0—P-0--
P—O
R,.
11
3
R4
Carbonyl
Aliphatic
carbons
Aromatic
carbons
orthophosphates
11
2
o
Aliphatic
carbons
Aromatic
carbons
M.
Jimenez
et
al.
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
1387
orthophosphates
0
R
2
-0—P-0-13
1
0
Carbonyl
R
3
Aliphatic
carbons
n
umescen
coa
ing
PA6
6
GF
pyrophosphates
0 0
I
I
13,-0—P-0--P
0
R,
0
orthophosphates
R
3
Carbonyl
0
Aliphatic
carbons
Aromatic
carbons
ignition
pyrophosphates
0 0
II
II
R
2
-0—P-0--P
0
R,.
I
I
0
Rorthophosphates
R
3
R4
R
2
0
P
O—R,
Aliphatic
carbons
0
Aromatic
carbons
R
3
R
4
O
Az
°
P
R,
Fig.
12.
Mechanism
of
action
of
the
intumescent
coating
on
the
PA6,6-GF
containing
5%
AlPi.
the
AlPi
are
trapped
in
the
intumescent
coating
and
this
could
favor
the
swelling
but
also
lead
to
a
stabilization
of
the
intumescent
coating
since
aluminophosphate
are
formed
and
thus
the
protec-
tive
barrier
is
more
stable.
5.
Conclusions
The
first
objective
of
this
paper
was
to
describe
the
mechanism
of
action
of
an
intumescent
coating
on
PA6,6.
It
was
evidenced
characterizing
the
material
by
13
C,
31
P
and
27
A1
solid
state
NMR
at
different
combustion
times
of
cone
calorimeter
test
that
during
burning
one
part
of
the
virgin
polymer
begins
to
degrade;
this
viscous
degraded
polymer
is
then
mixed
with
the
semi-viscous
charring
intumescent
layer.
In
some
way,
the
polymer
is
fire
retarded
in
bulk
thanks
to
the
intumescent
coating
applied
on
it.
The
second
objective
was
to
investigate
the
potential
synergy
be-
tween
the
gas
phase
fire-retardant
mechanism
of
AlPi
and
the
condensed
phase
fire
protective
mechanism
of
the
intumescent
coating
to
explain
the
enhanced
fire-retardant
properties
(e.g.
GWFI
validated
at
960
°C).
Using
the
same
NMR
technique,
the
interest
to
combine
the
bulk
treatment
in
low
amount
and
the
intumescent
coating
was
evidenced
to
be
double:
the
AlPi
cannot
completely
sublimate
because
of
the
protective
coating,
and
is
probably
condensed
inside
the
intumescent
structure
pores.
As
it
is
trapped
in
the
condensed
phase,
it
then
degrades
into
alumi-
nophosphates,
increasing
the
thermal
stability
and
thus
the
heat
barrier
efficiency
of
the
expanded
char
layer.
Acknowledgments
The
authors
would
like
to
thank
Rhodia
Company
for
expertise
and
financial
support,
Bertand
Revel
for
NMR
experiments
and
Helene
Gallou
for
sample
preparation
and
fire
tests.
References
[1]
Lewin
M.
Le
Bras
M,
Camino
G,
Bourbigot
5,
Delobel
R,
editors.
Fire
retardancy
of
polymers
the
use
of
intumescence.
Cambridge,
UK:
The
Royal
Society
of
Chemistry;
1998.
p.
3-32.
[2]
Nelson
GL
Nelson
GL,
editor.
Proceedings
of
the
American
chemical
society
symposium,
fire
and
polymers
IL
Washington,
DC:
American
Chemical
Society;
1995.
p.
1-24.
[3]
Weber
LWD,
Greim
H.
The
toxicity
of
brominated
and
mixed-halogenated
dibenzo-p-dioxins
and
dibenzofurans:
an
overview.
Journal
of
Toxicology
and
Environmental
Health
Part
A
1997;50:195-215.
[4]
Camino
G,
Luda
MP,
Costa
L
In:
Casal
J,
editor.
Chemical
industry
and
envi-
ronment,
general
analysis
risk,
11993.
p.
221-7.
[5]
Jimenez
M,
Duquesne
5,
Bourbigot
S.
Fire
protection
of
polypropylene
and
polycarbonate
by
intumescent
coatings.
Polymers
for
Advanced
Technologies
2012;23:130-5.
[6]
Jimenez
M,
Gallou
H,
Duquesne
5,
Jama
C,
Bourbigot
5,
Couillens
X,
et
al.
New
routes
to
flame
retard
polyamide
6,6
for
electrical
applications.
Journal
of
Fire
Sciences
2012;30:535-51.
[7]
Braun
U,
Schartel
B,
Fichera
MA,
Jager
C.
Flame
retardancy
mechanisms
of
aluminium
phosphinate
in
combination
with
melamine
polyphosphate
and
zinc
borate
in
glass-fibre
reinforced
polyamide
6,6.
Polymer
Degradation
and
Stability
2007;92:1528-45.
[8]
Braun
U,
Schartel
B.
Flame
retardancy
mechanisms
of
aluminium
phosphinate
in
combination
with
melamine
cyanurate
in
glass-fibre-reinforced
poly(1,4-
butylene
terephthalate).
Macromolecular
Materials
and
Engineering
2008;293:206-17.
[9]
Gallo
E,
Braun
U,
Schartel
B,
Russo
P,
Acierno
D.
Halogen-free
flame
retarded
poly(butylene
terephthalate)
(PBT)
using
metal
oxides/PBT
nanocomposites
in
combination
with
aluminium
phosphinate.
Polymer
Degradation
and
Stability
2009;94:1245-53.
[10]
Gallo
E,
Schartel
B,
Braun
U,
Russo
P,
Acierno
D.
Fire
retardant
synergisms
between
nanometric
Fe
2
O
3
and
aluminum
phosphinate
in
poly(butylene
terephthalate).
Polymers
for
Advanced
Technologies
2011;22:2382-91.
[11]
Yi
J,
Liu
Y,
Cai
X.
The
synergistic
effect
of
adjuvant
on
the
intumescent
flame-
retardant
ABS
with
a
novel
charring
agent.
Journal
of
Thermal
Analysis
and
Calorimetry
2012:1-9.
[12]
Bourbigot
S,
Le
Bras
M,
Delobel
R,
Breant
P,
Tremillon
JM.
4a
zeolite
synergistic
agent
in
new
flame
retardant
intumescent
formulations
of
polyethylenic
polymers
study
of
the
effect
of
the
constituent
monomers.
Polymer
Degradation
and
Stability
1996;54:275-87.
[13]
Bourbigot
5,
Le
Bras
M,
Delobel
R,
Decressain
R,
Amoureux
JP.
Synergistic
effect
of
zeolite
in
an
intumescence
process:
study
of
the
carbonaceous
structures
using
solid-state
NMR.
Journal
of
the
Chemical
Society
Faraday
Transactions
1996;92:149-58.
[14]
Bourbigot
S,
Vanderhart
DI,
Gilman
JVV,
Awad
WH,
Davis
RD,
Morgan
AB,
et
al.
Investigation
of
nanodispersion
in
polystyrene-montmorillonite
nano-
composites
by
solid-state
NMR.
Journal
of
Polymer
Science,
Part
B:
Polymer
Physics
2003;41:3188-213.
[15]
Bourbigot
5,
Vanderhart
DI,
Gilman
JW,
Bellayer
S,
Stretz
H,
Paul
DR.
Solid
state
NMR
characterization
and
flammability
of
styrene-acrylonitrile
copol-
ymer
montmorillonite
nanocomposite.
Polymer
2004;45:7627-38.
[16]
Samyn
F,
Bourbigot
S.
Thermal
decomposition
of
flame
retarded
formula-
tions
PA6/aluminum
phosphinateimelamine
polyphosphateiorganomodified
1388
M.
Jimenez
et
al.
/
Polymer
Degradation
and
Stability
98
(2013)
1378-1388
day:
interactions
between
the
constituents?
Polymer
Degradation
and
Stability
2012;97:2217-30.
[17]
ASTM
E
1354-90aStandard
test
method
for
heat
and
visible
smoke
release
for
materials
and
products
using
an
oxygen
depletion
calorimeter
1990.
[Philadelphia].
[18]
International
Standard
IEC
60695-2-12
fire
hazard
testing
part
2-12:
glowing/hot-wire
based
test
methods
glow
wire
flammability
test
method
for
materials.
[19]
Lippmaa
E,
Samoson
A,
Magi
M.
High-resolution
27A1
NMR
of
aluminosili-
cates.
Journal
of
the
American
Chemical
Society
1986;108:1730-5.
[20]
Weeding
TI„
Veeman
WS,
Angad
Gaur
H,
Huysmans
WGB.
Structural
inves-
tigation
of
polyamide-6
and
polyamide-6
composites
using
13C
cross
polar-
ization/magic
angle
spinning
NMR.
Macromolecules
1988;21:2028-32.
[21]
Jahromi
S,
Gabrielse
W,
Braam
A.
Effect
of
melamine
polyphosphate
on
thermal
degradation
of
polyamides:
a
combined
X-ray
diffraction
and
solid-
state
NMR
study.
Polymer
2003;44:25-37.
[22]
Camino
G,
Costa
L,
Luda
di
Cortemiglia
MP.
Overview
of
fire
retardant
mechanisms.
Polymer
Degradation
and
Stability
1991;33:131-54.
[23]
Schartel
B,
Braun
U.
Fire
retardancy
mechanisms
of
phosphorus
in
thermo-
plastics.
Polymeric
Materials:
Science
and
Engineering
2004;91:52-3.
[24]
Braun
U,
Bahr
H,
Schartel
B.
Fire
retardancy
effect
of
aluminium
phosphinate
and
melamine
polyphosphate
in
glass
fibre
reinforced
polyamide
6.
E-Polymers
2010.
[25]
Ganapathy
5,
Bryant
RG.
Carbon-13
spectral
line
broadening
in
paramagnetic
solids.
Journal
of
Magnetic
Resonance
1969;1986(70):149-52.
[26]
Levchik
SV,
Camino
G,
Costa
I,
Luda
MP.
Mechanistic
study
of
thermal
behaviour
and
combustion
performance
of
carbon
fibre-epoxy
resin
com-
posites
fire
retarded
with
a
phosphorus-based
curing
system.
Polymer
Degradation
and
Stability
1996;54:317-22.
[27]
Le
Bras
M,
Bourbigot
5,
Delporte
C,
Siat
C,
Le
Tallec
Y.
New
intumescent
for-
mulations
of
fire-retardant
polypropylene
discussion
of
the
free
radical
mechanism
of
the
formation
of
carbonaceous
protective
material
during
the
thermo-oxidative
treatment
of
the
additives.
Fire
and
Materials
1996;20:
191-203.
[28]
Nocuti
M.
Structural
studies
of
phosphate
glasses
with
high
ionic
conductivity.
Journal
of
Non-Crystalline
Solids
2004;333:90-4.
[29]
Vaghefi
MM.
In:
Phosphorus
NMR
chemical
shifts
of
nucleotide
derivatives
and
some
related
organophosphorus
compounds,
nucleoside
triphosphates
and
their
analogs.
Taylor&Francis:
CPC
Press
Book;
2005.
p.
373-4.
Appendix
1.
[30]
Miller
DBG,
Grunze
I,
Ladwig
G,
Hallas
E,
Haubenreisser
U.
Solid
state
high-
resolution
27A1
nuclear
magnetic
resonance
studies
of
the
structure
of
Ca0-
A1
2
0
3
-P
2
0
5
glasses.
Physics
and
Chemistry
of
Glasses
1983;24:37-42.
[31]
Jimenez
M,
Duquesne
5,
Bourbigot
S.
Intumescent
fire
protective
coating:
toward
a
better
understanding
of
their
mechanism
of
action.
Thermochimica
Acta
2006;449:16-26.