Phosphorus-containing copolyesters: The effect of ionic group and its analogous phosphorus heterocycles on their flame-retardant and anti-dripping performances

Article abstract of DOI:10.1016/j.polymer.2015.01.030

A series of copolyesters containing ionic groups (PETIs-K) were synthesized by melt co-polycondensation using terephthalic acid, ethylene glycol, and potassium salt of bishydroxyethyl ester of 10-hydroxy-10-oxo-10H-phenoxaphosphine-2,8-dicarboxylic acid (DHPPO-K) as monomers. The copolyesters only with analogous phosphorus-containing heterocycles (PETPs) were also synthesized and used as a control. The dynamic rheological behaviors confirmed that the complex viscosity of PETIs-K increased with increasing temperature due to the existence of ionic aggregates, which restricted melt dripping, while PETPs had flow behaviors similar to PET: the complex viscosity decreased with increasing temperature. The ionic groups induced the decomposition of the polymer backbone to form a more stable residue at high temperatures, which made PETIs-K exhibit good flame retardancy even with low phosphorus contents. Cone test showed the peak heat release rate and total smoke production of PETP5 decreased to 34% and 58% of pure PET. Compared with PETP5, these values for PETI5-K were reduced by 31% and 50%, respectively. EDX and ICP-AES results demonstrated that 92% of phosphorus for PETI10-K remained in char residues; however, this value for PETP10 was only 45%. Combined with results of XPS, Raman spectroscopy and Py-GC/MS, it can be proved that the ionization of phosphorus took main effects in condensed phase via promoting carbonization to form stable graphitic char, which avoided the decrease of carbonization due to the competitive effects of phosphorus in gas phase.

Full text of DOI:10.1016/j.polymer.2015.01.030

Polymer 60 (2015) 50e61
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Phosphorus-containing copolyesters: The effect of ionic group and
its analogous phosphorus heterocycles on their flame-retardant and
anti-dripping performances
Yi Zhang, Yan-Peng Ni, Ming-Xin He, Xiu-Li Wang*, Li Chen, Yu-Zhong Wang*
Center for Degradable and Flame-Retardant Polymeric Materials, College of Chemistry, State Key Laboratory of Polymer Materials Engineering, National
Engineering Laboratory of Eco-Friendly Polymeric Materials (Sichuan), Sichuan University, Chengdu 610064, China
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of copolyesters containing ionic groups (PETIs-K) were synthesized by melt co-polycondensation
using terephthalic acid, ethylene glycol, and potassium salt of bishydroxyethyl ester of 10-hydroxy-10-
oxo-10H-phenoxaphosphine-2,8-dicarboxylic acid (DHPPO-K) as monomers. The copolyesters only
with analogous phosphorus-containing heterocycles (PETPs) were also synthesized and used as a control.
The dynamic rheological behaviors confirmed that the complex viscosity of PETIs-K increased with
increasing temperature due to the existence of ionic aggregates, which restricted melt dripping, while
PETPs had flow behaviors similar to PET: the complex viscosity decreased with increasing temperature.
The ionic groups induced the decomposition of the polymer backbone to form a more stable residue at
high temperatures, which made PETIs-K exhibit good flame retardancy even with low phosphorus
contents. Cone test showed the peak heat release rate and total smoke production of PETP5 decreased to
34% and 58% of pure PET. Compared with PETP5, these values for PETI5-K were reduced by 31% and 50%,
respectively. EDX and ICP-AES results demonstrated that 92% of phosphorus for PETI10-K remained in
char residues; however, this value for PETP10 was only 45%. Combined with results of XPS, Raman
spectroscopy and Py-GC/MS, it can be proved that the ionization of phosphorus took main effects in
condensed phase via promoting carbonization to form stable graphitic char, which avoided the decrease
of carbonization due to the competitive effects of phosphorus in gas phase.
Received 20 September 2014
Received in revised form
29 November 2014
Accepted 14 January 2015
Available online 22 January 2015
Keywords:
Phosphorus
Ionomers
Flame retardance
© 2015 Elsevier Ltd. All rights reserved.
1. Introduction
polymerization, or finishing the fabrics/textile in a solution con-
taining flame retardants [5e8]. Among these methods, introducing
Poly(ethylene terephthalate) (PET), one of the most important
commercial polymers, possesses excellent chemical resistance,
thermal stability, spinnability, and good mechanical properties
[1e4]; it has been widely used as films, synthetic fibers, bottles,
and engineering plastics. Unfortunately, PET is flammable and ex-
hibits serious melt-dripping during combustion, which limits its
application in some fields. At present, methods for endowing PET
with flame retardancy are common, and some have been incor-
porated into industrialize processes. Some of these methods are,
for example: adding flame retardants into PET by blending,
incorporating flame-retardant monomers into PET chains via
phosphorus-containing monomers into the PET backbone is an
efficient way for improving PET flame retardancy, as it avoids poor
durability and compatibility problems caused by the other methods
[9e12]. These phosphorus-containing PET copolyesters usually
achieve flame retardancy through removing heat from the burning
surface, which enhances the melt-dripping of PET [13]. Therefore,
how to resolve the conflict between the flammability and melt-
dripping behavior has become a challenge for the development of
flame-retardant PET in real applications.
In our previous work, we found if some unsaturated conjugated
groups (such as azobenzene groups and phenylethynyl groups, etc.)
were introduced into the PET backbone, these copolyesters could
self-crosslink during combustion via chemical cross-linking
[4,14,15]. This self-crosslinking reaction created copolyesters with
high melting viscosity, giving both flame retardancy and anti-
dripping simultaneously. Unfortunately, high cross-linkable-
* Corresponding authors. Tel./fax: þ86 28 85410755.
E-mail addresses: xiuliwang1@163.com (X.-L. Wang), yzwang@scu.edu.cn
(Y.-Z. Wang).
http://dx.doi.org/10.1016/j.polymer.2015.01.030
0032-3861/© 2015 Elsevier Ltd. All rights reserved.

Y. Zhang et al. / Polymer 60 (2015) 50e61
51
monomer dosage was needed in order to achieve good flame
retardancy, and this destroyed the crystalline properties of PET
seriously.
all obtained from Kelong Chemical Industral Reagent Co. (Chengdu,
China). Tetrabutyl titanate was dissolved into anhydrous toluene to
prepare a 0.2 g/ml solution before use. Other materials were used as
received.
Recently, our research group introduced the ionomer concept to
the field of phosphorus-containing copolyesters, and found the
conflict between the flame retardance and melt-dripping of PET
could be resolved in this manner [16e18]. In our previous work, it
was found that when DHPPO-Na (the sodium salt of 10H-phenox-
aphosphine-2,8-dicarboxylic acid 10-hydroxy-,2,8-dihydroxyethyl
ester 10-oxide) was introduced into PET backone, the melt viscos-
ity of PET was increased greatly via ionic aggregation. Also, the char
of this PET ionomer formed rapidly during decomposition, which
restricted melt-dripping [19]. The thermal behavior, especially the
crystallization ability, of this ionomer was similar with pure
PET, indicating this ionomer with thermo-reversible physical
crosslinking (ionic aggregation) had better processability and
spinnability than the aforementioned chemical cross-linking
copolyesters.
2.2. Preparation of PET copolyesters containing phosphorus
heterocycles (PETPs)
The exact synthesis procedure for PETPs was described as fol-
lows (Scheme 1): PET prepolymer was prepared via esterification
and polycondensation reaction of TPA and EG in a 1-L autoclave.
The mixture of PET prepolymer and DCPPO-Ph (its preparation
procedure and structure characterization are shown in supporting
information, Scheme S1 and Fig. S1(b)), was melted under N2, and
then a quantity of tetrabutyl titanate was added as a catalyst. The
copolyesters were obtained after polycondensation at 240e280 ꢀ
C
under vacuum. 1H NMR (Fig. S2(a), CF3COOD,
d, ppm): 4.76 (s, 4H);
8.08 (s, 4H); 8.31e8.50, 7.36e7.60 (m, 11H). 31P NMR (CF3COOD,
d,
Some literatures reported that different alkaline metal ions had
different flame-retardant effects on the bio-materials (jute, wood
based material etc.) [20e22]. The greater the ionic radius, the lower
the thermal stability for alkali group [23]. And potassium-catalyzed
cellulose pyrolysis considerably increased the char yields and
promoted the generation of phenol and phenol derivatives, which
would form polycyclic aromatic hydrocarbons at high temperature
[24,25].
ppm): 2.68 (s).
2.3. Preparation of PET copolyester ionomers (PETIs-K)
Ionic monomer, DHPPO-K was prepared according to our pre-
vious work (structure characterization shown in Fig. S3) [19].
Poly(ethylene terephthalate)s-based ionomers (PETIs-K) were
prepared in a same way (Scheme 2). 1H NMR (Fig. S2(b), CF3COOD,
Therefore, in this work, we changed our previous ionic mono-
mer DHPPO-Na to DHPPO-K (the potassium salt of 10H-phenox-
aphosphine-2,8-dicarboxylic acid,10-hydroxy-,2,8-dihydroxyethyl
ester, 10-oxide), and introduced it into the PET backbone via melt
polymerization. For better understanding the influence of ionic
groups on flame retardancy, thermal stability, anti-dripping
behavior, and pyrolysis behavior, an analogous, comparative,
phosphorus heterocycles monomer without an ionic group was
prepared and copolymerized with terephthalic acid and ethylene
glycol to prepare another PET copolyesters (PETPs).
d
, ppm): 4.49 (s, 4H); 7.18 (m, 2H); 7.80 (s, 4H); 8.07 (m, 2H); 8.48
(m, 2H). 31P NMR (CF3COOD,
d
, ppm): 8.36 (s).
The basic parameters of PETPs and PETIs-K are shown in Table 1.
2.4. Characterization
Intrinsic viscosities [h] of samples were measured in an Ubbe-
lohde viscometer at 25 ꢀC in phenol/1,1,2,2-tetrachloroethane so-
lution (v/v, 50/50) at a concentration of 5 g polymer/L solvent.
NMR spectroscopy was carried out on a Bruker AVANCE oper-
ating at 400 MHz for 1H and at 161.9 MHz for 31P (DMSO-d6 or
CF3COOD was used as solvent and tetramethylsilane was used as
reference).
2. Experimental
2.1. Materials
The actual phosphorus contents of the purified samples were
determined by oxygen flask combustion-inductively coupled
plasma atomic emission spectroscopy (ICP-AES) method.
The rheological properties were determined with a TA DHR-2
rheometer in air atmosphere. Disk-like samples with a size of
2.5 mm diameter and 1 mm thickness were dried under vacuum at
Terephthalic acid (TPA, fiber grade) was purchased from Jinan
Chemical Fiber Co. Ltd. (Jinan, China). Ethylene glycol (EG, indus-
trial grade) was provided by Xilong Chemical Industries Co. Ltd.
(Chengdu, China). 4,4’-dimethyldiphenylether (AR grade) was
provided by Jiaxing Sicheng Chemicals Co. Ltd. (Jiaxing, China).
Phenylphosphonous dichloride (AR grade), phosphorus trichloride
(AR grade), aluminum trichloride (AR grade), potassium perman-
ganate (KMnO4, AR grade), and tetrabutyl titanate (AR grade) were
120 ꢀC for 12 h. The complex viscosity (
h) was measured as function
of temperature ranging from 250 to 330 ꢀC for PETP3 and PETI3-K,
from 230 to 330 ꢀC for PETP10 and PETI10-K, and from 260 to 330 ꢀC
for neat PET, each at a fixed frequency of 1 Hz.
Scheme 1. Synthesis process of PETP, where m and n denote the mole of DCPPO-Ph and TPA, respectively, not the block length or repeating units.

52
Y. Zhang et al. / Polymer 60 (2015) 50e61
Scheme 2. Synthesis process of PETI-K, where m and n denote the mole of DHPPO-K and TPA, respectively, not the block length or repeating units.
The thermal stabilities were investigated by NETZSCH TG 209F1
program of the capillary column (DN-1701 FAST 10 m 0.10 mm
0.10 mm) of GC was as following: 2 min at 45 ꢀC, temperature
increased to 280 ꢀC at a rate of 15 ꢀC/min then kept at 280 ꢀC for
5 min. The injector temperature was 280 ꢀC. MS indicator was
operated in the electron impact mode at electron energy of 70 eV,
and the ion source temperature was kept at 180 ꢀC. The detection of
mass spectra was carried out using a NIST library. In a few cases, the
compounds not included in the library were identified on the basis
of both molecular ion m/z values and the ion decomposition pat-
terns for the best fit with the mass spectra.
apparatus in the temperature range from 40 to 700 ꢀC (gas flow:
50 ml/min). The thermal degradation kinetics were analyzed using
the Flynn-Wall-Ozawa method at various heating rates: 5, 10, 20,
and 40 ꢀC/min.
The sample residues at 700 ꢀC in TGA test were collected and
their morphologies were observed by a JEOL JSM 5900LV scanning
electron microscopy (SEM) (JEOL, Japan) at the accelerating voltage
of 20 kV.
The burning behaviors of samples were tested using LOI and
cone calorimeter measurements. The LOI measurements were
performed on the Oxygen Index Flammability Gauge (HC-2C) ac-
3. Results and discussion
cording to ASTM
D
2863-97 with
a
sample size of
120 mm ꢁ 6.5 mm ꢁ 3.2 mm. The cone calorimeter tests were
3.1. Rheological behaviors of PETPs and PETIs-K
carried out using the ISO 5660 standard with a FTT (UK) cone
calorimeter. Specimens were moulded to
a
size of
According to our previous results [19]. the melt viscosity of PET
containing ionomers in the main-chain backbone will increase due
to the thermo-reversible ionic aggregation. In order to demonstrate
this ionic monomer effect for different metal ions, melt rheological
performance was measured. The dependences of complex viscosity
100 mm ꢁ 100 mm ꢁ 3.2 mm on a plate vulcanizing machine. The
heat flux was 50 kW/m2.
Energy dispersive X-ray (EDX) measurements were conducted
on an energy dispersive X-ray spectrometer (EDX, INCA, PENTA-
FETX3, OXFORD) equipped with an SEM (JEOL JSM-5900LV).
The laser Raman spectra were recorded at room temperature
using a LabRAM HR laser Raman spectrometer (HORIBA Co., France)
with excitation wavelength of 532 nm.
(j *j) on temperature for PET, PETPs, and PETIs-K are shown in Fig.1.
h
It can be seen that PETPs and neat PET exhibit shear thinning be-
haviors with increased temperature, consistent with previous re-
ports [4,14]. The rigid heterocycle of DCPPO-Ph reduces the
X-Ray photoelectron spectroscopy (XPS) was measured with a
mobility of polymer chains, leading to the higher j
PETPs compared with that of neat PET [4,14]. In comparison with
the linear decrease for PETPs, the j *j tendency for PETIs-K is
h
*j values of
XSAM 800 (Kratos Co., UK), using Al K
1486.6 eV).
a excitation radiation (hv-
h
The pyrolysis behaviors of samples were investigated with Py-
GC/MS. The pyrolyzer (CDS5200) was full of helium atmosphere,
the relevant samples (300 m
g) were heated from ambient to 500 ꢀC
at a rate of 1000 ꢀC/min, with isothermal steps of 20 s. The pyro-
lyzer was coupled with a DANI MASTER GC-TOF-MS Systems, also
using the carrier gas helium. For the operation, the temperature
Table 1
Basic parameters of PET, PETPs, and PETIs-K.
Sample
DCPPO-Ph
(mol%)a
DHPPO-K
(mol%)b
Pt (%)c
Pa (%)d
[h] (dL∕g)
PET
PETP3
PETP5
PETP10
PETI3-K
PETI5-K
PETI10-K
e
3
5
10
e
e
e
e
e
e
e
3
0
0
0.94
0.94
0.82
0.77
0.84
0.68
0.57
0.46
0.73
1.33
0.46
0.73
1.35
0.41
0.67
1.27
0.41
0.67
1.33
5
10
a
DCPPO-Ph mol content based on TPA.
DHPPO-K mol content based on TPA.
b
c
Theoretical phosphorus content of the copolyester.
Fig. 1. Complex viscosity of neat PET, PETPs and PETIs-K as a function of temperature
at an angular velocity of 1 Hz.
d
Actual phosphorus content of the copolyester determined by ICP-AES.

Y. Zhang et al. / Polymer 60 (2015) 50e61
53
Table 2
entirely different. The j *j first experiences a decrease, subse-
h
The thermal stabilities of PET, PETPs, and PETIs-K at different atmosphere.
quently increases to a maximum (a cross-linking reaction occurs in
other words) [4,14,26,27], and afterward decreases again. The in-
Sample
N2
Air
crease of j *j is contributed to the rearrangement of ionic structure
h
T5% (ꢀC) Td max Residue T5% (ꢀC) Td max1 Td max2 Residue
(ꢀC)
(%)
(ꢀC)
(ꢀC)
(%)
and enhanced ionic interaction between polymer chains at higher
testing temperatures, which means the change of metal ions
PET
395
391
393
395
388
389
436
428
430
430
427
428
425
9
18
19
25
20
24
25
337
378
387
387
382
382
373
434
432
420
427
431
430
425
544
562
560
565
538
538
541
1
1
1
4
2
5
5
doesn't alter its self cross-linking nature [19]. The j *j value of
h
PETP3
PETP5
PETP10
PETI3-K
PETI5-K
PETI3-K is always higher than that of neat PET due to ionic in-
teractions of ionomer [28,29]. Because the intrinsic viscosity of
PETI10-K (0.57 dL/g) is much lower than that of neat PET (0.94 dL/
g), and the j *j value drops faster as incorporating more ionic
h
PETI10-K 376
groups, as with other systems [30,31]. The lowest value of j *j
h
observed for PETI10-K is between 260 ꢀC and 296 ꢀC. As we know,
the melt fluidity becomes more difficult, meaning the polymer has
more time for charring [18]. Although, PETIs-K and PETPs have a
similar structure, the ionization for PETIs-K plays a positive role in
anti-dripping through restricting the movement of polymer chains
at the beginning of burning.
significantly from 336.6 ꢀC for pure PET to 387.1 ꢀC for PETP10,
consistent with the improved carbonization in the presence of
phosphorus. This illustrates the initial thermal stability of PET is
enhanced when a phosphorus heterocycle monomer is introduced.
However, The PETIs-K show a different decomposition tendency
compared with PETPs. T5% of the ionomer decreases from 382 ꢀC of
PETI3-K to 373 ꢀC of PETI10-K. This illustrates that the existence of
ionic groups not only promotes the carbonization but also catalyzes
the decomposition of the polymer matrix easily [19]. The polarity of
DHPPO-K is much higher than that of DCPPO-Ph, and the ester bond
breaks faster for ionomers at high temperatures. But the PeC seg-
ments of the ionomer will be oxidized in air sooner and form stable
protective substances.
3.2. Thermal stabilities and thermal degradation kinetics of PETPs
and PETIs-K
In order to compare the effect of the ionic monomer and the
phosphorus heterocycles monomer on copolyester's thermal sta-
bility, TGA curves of PETPs and PETIs-K under nitrogen and air at-
mosphere are shown in Fig. 2. And the corresponding data such as
onset decomposition temperature (T5%), the temperature at
maximum weight loss rate (Tdmax), and char residue are summa-
rized in Table 2. Under nitrogen atmosphere, PETPs and PETIs-K
display a single decomposition process which is similar to other
phosphorus-containing copolyesters [9,12]. The T5% of PETP3 is
slightly lower than that of PET, which is probably due to the weaker
bond energy of PeC [18]. However, the T5% of PETPs increases when
the content of phosphorus-containing heterocycles increases from
3 mol% to 10 mol%. At the same time, we note that the residue of
PETPs increases from 8.9 % to 25.1 % for PETP10, which illustrates
that the introduction of DCPPO-Ph significantly improves the
carbonization. Such a substantial char may well block the volatil-
izable products during thermal degradation, enhancing thermal
stability. The T5% of PETIs-K is lower than that of PETPs and PET due
to the polar ionic groups accelerating the decomposition of the
polymer matrix, however, a more stable char residue forms at high
temperatures.
In air, the T5% of PETI10-K is much higher than that of PET, and
obviously lower than that of PETP10. The char residue for PETIs-K is
also higher than that of PETPs and PET, which will obstruct the
transmission of heat and substance, benefiting for the flame
retardancy.
The SEM images of the residues of PETI10-K and PETP10 after
TGA either in nitrogen or in air are shown in Fig. 3. After TGA test (in
nitrogen), a few bubble structures, formed during the decomposi-
tion, are observed for the residue of PETI10-K (as shown in
Fig. 3(a)). And these bubble structures are easy to be destroyed and
leave macro-pores in PETP10 residues (Fig. 3(a)). When TGA test is
carried out in air, however, their residues show different mor-
phologies. Specifically, the residue of PETP10 shows some discon-
tinuous segments and almost no condensed char left; for PETI10-K,
an apparent firm char can be observed, and its char residue surface
is much more compact than that of PETP10 (as shown in Fig. 3(c)
and (d)). This means that ionic group will enhance the carboniza-
tion ability of PET, which will be further discussed in the following
part.
Under air atmosphere, all pure PET, PETPs, and PETIs-K, show a
two-step decomposition process. The PETPs have a lower peak
temperature (Tdmax1) at the major weight loss stage and a higher
peak temperature (Tdmax2) at the minor weight loss stage compared
to PET. The onset decomposition temperature increases
The Flynn-Wall-Ozawa method (Eq. (1)) is used to further
analysis the thermal degradation behaviors of PETPs and PETIs-K
[32e34].
Fig. 2. TGA thermograms of PET, PETPs, and PETIs-K at a heating rate of 10 ꢀC/min in different atmospheres: (a) nitrogen and (b) air.

54
Y. Zhang et al. / Polymer 60 (2015) 50e61
Fig. 3. SEM micrographs of PETI10-K (a, c) and PETP10 (b, d) residues after TGA test either in nitrogen (a, b) or in air (c, d).
content. This means that in nitrogen atmosphere, the carbonization
of PETIs-K can promote the formation of more stable residue, which
is consistent with previous TGA results.
AE
R
E
RT
lgFðaÞ ¼ lg
ꢂ lgb ꢂ 2:315 ꢂ 0:4567
(1)
where F(
exponential factor, E is the activation energy, R is the gas con-
stant, is the heating rate, and T is the temperature. This method
can determine the activation energy without knowing the reaction
order. The activation energy for different conversion values can be
a
) is the integral function of the conversion, A is the pre-
3.3. The burning behaviors of PETPs and PETIs-K
b
The combustibility of samples was first tested by LOI. In general,
the higher the phosphorus content, leading to better flame
retardancy, makes a higher oxygen concentration necessary to
allow for burning. As shown in Table 4, the LOI value increases from
22.0 vol.% for PET to 29.0 vol.% for PETP10 as the phosphorus
content increases from 0 to 1.27%. When the P content is lower than
0.41% (corresponding to PETP3), the LOI value scarcely increases.
This means only increasing P content to a certain value an increase
of LOI value can be found. Compared with the results of other
copolyesters [10,35], whose LOI value increases with the increase of
P content, this phenomenon is probably due to the partly flame-
retardant activity of PETPs in solid phase. As for as PETIs-K, their
LOI values are enhanced linearly with the increase of ionic mono-
mer content. The LOI values of PETIs-K range from 26.0 vol.% ꢂ33.0
vol.%, which are higher than that of PETPs. Additionally, the LOI
values of PETIs-K are even higher than that of PETIs-Na with the
same ionic monomer content [19], proving the higher flame-
retardant effect of potassium ion for PET. In detail, the combus-
tion processes of PET, PETPs, and PETIs-K are illustrated in Fig. 6.
When the testing oxygen concentration is 26 vol.%, PET is ignited
quickly with rapid flame propagation; while melt-dripping with
fire is observed after 15s. For PETP5, the flame propagation is slow;
and the flame can extinguish with apparent melt-dripping. At an
oxygen concentration of 29 vol.%, PETP10 exhibits no melt-dripping
during the test, which is different from the other phosphors-
containing PET copolyesters [12,13]. This means that phosphorus-
calculated from lg
b versus 1/T plot. The TGA curves of PET, PETPs,
and PETIs-K at different heating rates under nitrogen are shown in
Fig. 4.
As expected, the decomposition temperature increases with
increasing heating rates. Activation energies at 5, 10, 15, 20, 30, 40,
and 50% conversion were studied. Best-fit lines for the plots of lg
b
versus 1/T plot are illustrated in Fig. 5. The calculated activation
energies are given in Table 3. The best-fit lines are nearly parallel
with a correlation coefficient above 0.99, thus indicating the
applicability of this method to our system in the conversion range
investigated. As shown in Table 3, activation energy increases with
increasing phosphorus content for PETPs (the correlation coeffi-
cient of activation energy of samples according to Ozawa's method
is shown in Table S1). The value increases from 194 kJ/mol of pure
PETP3 to 200 kJ/mol of PETP10 at 5% conversion, and these values
increased to 210 kJ/mol and 217 kJ/mol at 50% conversion, respec-
tively. The high activation energy of PETPs means the decomposi-
tion of residue is difficult, in other words, stable residue forms. For
the ionomers, the activation energy firstly increases to 192 kJ/mol
for PETI5-K and then decreases to 169 kJ/mol for PETI10-K at 5%
conversion, which is due to the catalytic decomposition effect of
the ionic group. After 40% conversion, the activation energy of
PETIs-K is higher than that of PETPs with the same phosphorus

Y. Zhang et al. / Polymer 60 (2015) 50e61
55
Fig. 4. TGA thermograms of PET, PETP5, PETP10, and PETI10-K at different heating rates in nitrogen.
containing heterocycle of DCPPO-Ph has a partly flame-retardant
activity promoting charring in the solid phase [36], which gives
PETP10 an anti-melt-dripping property.
After ionization, the melt-dripping is further suppressed and the
carbonization is accelerated by ionic groups. Charring residues are
quickly generated covering the burning zone, and no melt-dripping
occurs for PETI5-K and PETI10-K during the LOI test. At the same
time, the burning time of PETI5-K and PETI10-K is only 24 s and 21 s
respectively, which is much shorter than that of PETP5 and PETP10.
From the above results, we can draw a conclusion that physical
cross-linking networks can effectively reduce the fluidity of
burning melt and induce carbonization quickly, thus endowing
Fig. 5. Ozawa plots of PET, PETP5, PETP10, and PETI10-K at the following fractional weight losses (a): 0.05, 0.10, 0.15, 0.20, 0.30, 0.40 and 0.50.

56
Y. Zhang et al. / Polymer 60 (2015) 50e61
Table 3
information such as heat release rate (HRR), total heat release
Activation energy of samples according to Ozawa's method.
(THR), smoke production rate (SPR), and residue mass [38,39]. The
HRR and SPR as a function of burning time are plotted in Fig. 7 and
the exact data are summarized in Table 4. The peak heat release rate
(PHRR) significantly decreases for PETPs and PETIs-K, with PHRR of
PETIs-K much lower than that of PETPs with the same 3 mol%
monomer content. This indicates that both ionic monomer and
phosphorus heterocycles can reduce PET fire danger. The HRR
curves of PETPs and PETIs-K presents a platform around the
maximum, which are different from the sharp peak as PET exhibits.
This is attributed to the barrier effect of intumescent char.
Compared with pure PET, the residue mass of PETPs and PETIs-K
increases because of the increased carbonization effect. And a
decreased trend in THR is observed for PETPs and PETIs-K.
Conversion
a
(%) Activation energy E (kJ/mol)
PET PETP3 PETP5 PETP10 PETI3-K PETI5-K PETI10-K
5
179 194
181 195
182 197
182 198
183 204
184 208
190 210
196
198
199
199
204
210
212
200
203
204
206
216
217
217
180
194
199
203
205
212
213
192
210
215
216
216
222
223
169
189
199
204
213
218
225
10
15
20
30
40
50
Table 4
The data of PET, PETPs, and PETIs-K obtained from Cone test.
As we know, smoke suppression is a crucial factor in real fire,
which directly protects people from death caused by suffocation
and/or inhalation of the toxic gases on site [40,41]. Total smoke
production (TSP) and smoke production rate (SPR) are here used to
quantify the smoke yield for evaluating fire hazard. For PETPs, the
TSP and SPR firstly increase and then decrease with increased
phosphorus content. This indicates the increased phosphorus-
containing heterocycle group content is able to enhance the
partly flame-retardant activity in solid phase, which is consistent
with TGA results. While PETIs-K exhibit a different trend in TSP and
SPR, the obviously decreased TSP and SPR for PETIs-K show the
Sample
TTIa
(s)
PHRR
THR
TSP
Residue
(wt%)
LOI
(vol.%)
(kW/m2)
(MJ/m2)
(m2)
PET
PETP3
PETP5
PETP10
PETI3-K
PETI5-K
PETI10-K
52
45
48
49
46
46
45
879
435
426
280
341
295
203
66
44
43
42
60
56
49
12
13
14
9
7
7
13
25
26
36
18
22
32
22.0
22.0
26.0
29.0
26.0
29.0
33.0
5
a
Time to ignition.
PETIs-K with better performances in flame retardance and anti-
dripping [26,27,37].
Cone calorimetry is considered as one of the most useful tests to
simulate real-word fire condition, which offers abundant
Fig. 7. Cone calorimetric results of PET, PETPs, and PETIs-K: (a) heat release rate (HRR)
Fig. 6. Combustion process of PET, PETPs, and PETIs-K.
and (b) smoke production rate (SPR).

Y. Zhang et al. / Polymer 60 (2015) 50e61
57
quantity of volatile components during burning decreases. This
means the ionic group has a stronger effect on solid phase flame
retardance, which shows a better smoke suppression for reducing
the toxic smoke hazard.
3.4. Cone char residue of PETPs and PETIs-K
In order to investigate the flame-retardant effect for
phosphorus-containing heterocycle monomer and ionic monomer,
EDX analysis for char residues after cone calorimeter test was car-
ried out. Table 5 gives the results of EDX analysis for char residues of
PET, PETP10, and PETI10-K. As shown in Table 5, the C/O value of
PETP10 is bigger than that of neat PET on the inner surface of char
residue, proving the carbonization effect of phosphorus-containing
groups. While the smaller C/O of PETI10-K on the inner surface of
char residue is probably a result of the catalytic effect of the ionic
group on decomposition of the matrix. On the outer surface of char
residue, PET has a higher C/O value than that of PETP10 and PETI10-
K, which is ascribed to the unstable oxidized neat PET products
forming gas or soot.
As for P content, P content (1.73 wt%) in PETP10 outer surface is
much lower than that of PETI10-K (8.64 wt%), which illustrates that
a portion of PETP10 enters the gas phase to capture the free radicals
leading to a much lower P content. And the enrichment of P content
for PETI10-K verifies most ionic monomers have a solid flame-
retardant effect. That is to say, when PETI10-K burns, ionic groups
induce PET oxidative decomposition and form more stable oxidized
products remaining in solid phase, which raises the O and P con-
tents [42].
Furthermore, the P content in the residue after cone test was
also determined by ICP-AES, and the results are shown in Table 5.
The maximum theoretical phosphorus content (Pmcotne) in the res-
idue is calculated according to the following equation:
cone
P
¼ ðPt=residueÞ ꢁ 100%
(2)
mt
where Pt is theoretical phosphorus content of the copolyester. R is
defined as the ratio of Pmcotne, where Pacone is actual phosphorus
content in char residue after cone test. The R value of PETP10 is only
45%, indicating that a considerable amount of phosphorus volatil-
ized into the air. While, the R value of PETI-K is 92%, meaning that
the majority of phosphorus remained in char residue. The different
R values suggest the different flame retardance mechanisms be-
tween PETPs and PETIs-K.
Raman spectroscopy was also used to investigate the graphitic
structure of char residue after cone test, and the corresponding
spectra are shown in Fig. 8. The char residue of each sample dis-
plays over lapped D and G peaks. The D band (around 1360 cm-1
)
corresponds to disordered graphite or glassy carbons and the G
band (around 1580 cm-1) is associated with the vibration of carbon
atoms in graphite layers, representing organized graphitic structure
Table 5
The data of EDX analysis and P content analysis of the residues after cone test.
Fig. 8. Raman spectra of char residue of PET, PETI5-K, PETP10, and PETI10-K.
cone
Sample
Inner surface content
(wt%)
Outer surface content
(wt%)
P
Pacone
R
mt
(%)a (%)b (%)c
[43e45]. The graphitization degree of char residue can be esti-
mated by the relative intensity ratio of the D and G bands (ID/IG),
where ID and IG are the integrated intensities of the D and G bands,
respectively. Generally, the lower ratio of ID/IG means the higher
graphitization degree of char. The ID/IG value follows the sequence
of PET (2.94) ＞ PETI5-K (2.63) ＞ PETP10 (2.44) ＞ PETI10-K (2.08).
This illustrates that with more phosphorus and ionic groups
C
O
P
C
O
P
PET
95.24
PETP10 97.53
PETI10-K 90.90
4.76
1.47
6.72
e
99.39
96.16
0.61
e
e
e
e
1.00
2.38
2.11 1.73 3.69 1.65 45
77.72 13.64 8.64 4.22 3.90 92
a
The maximum theoretical phosphorus content in char residue after cone test.
Actual phosphorus content in char residue after cone test.
The ratio Pacone=P
b
c
cone
.
mt

58
Y. Zhang et al. / Polymer 60 (2015) 50e61
Table 6
Compounds identified in the pyrograms of PET, PETP10, and PETI10-K.
Peak Retention
time (min)
M
Assigned structure
Relative area (%)
PET PETP10 PETI10-K
15
1
4.26
6.00
94
e
e
8
148
122
6
6
6
2
6.83
26
30
35
3
4
7.52
8.10
136
150
1
1
1
1
<1
<1
5
6
8.36
8.55
154
148
2
2
2
1
2
1
16
7
8.55
8.69
170
166
e
e
2
1
1
1
17
18
9.44
9.56
164
166
e
e
e
e
2
2
Fig. 9. P2p spectra of char residue of PETP10 and PETI10-K.
introduced into PET the formation of thermally stable graphitic
structure is promoted [46e49].
8
9.86
218
4
4
5
In order to determine the chemical state of phosphorus after
burning, XPS analysis is used for the PETP10 and PETI10-K char
residue. As shown in Fig. 9, the binding energy of P2p in char res-
idue shifts from 133.7eV for PETP10 to 134.9 eV for PETI10-K after
ionization. Usually, the binding energy between 134 eV and 135eV
can be assigned to PeOeC and/or PO-3 groups in phosphate species
and/or P2O5, respectively [50]. The increase of the binding energy
illustrates the formation of P2O5 during burning [51,52]. All the
results demonstrate this phosphorus-containing ionic group
9
10.41
10.61
194
198
17
18
7
2
19
e
e
10
11
11.18
12.36
220
192
4
3
4
3
<1
6
20
12.40
214
e
e
<1
21
12
13.64
13.94
294
270
e
e
1
3
5
2
13
14
16.42
16.75
298
292
9
5
1
3
e
e
Fig. 10. Pyrograms of PET, PETP10, and PETI10-K obtained at 500 ꢀC.

Y. Zhang et al. / Polymer 60 (2015) 50e61
59
with relatively high molecular mass, it is hard to assign their
structures. For PETI10-K, diphenyl ether, phenol, and their de-
rivatives (Peak 15-21) can be identified as decomposition products
of DHPPO-K, which is different from that for PETP10. In fact, there
are still unidentified pyrolysis segments, but without evidence of
phosphorus-containing products in the program of PETI10-K. As
shown in Fig. 12, the scission of different bonds (marked D, E, F, G,
H, J, L, M, and N, respectively) and new combinations of the seg-
ments produce the compounds (Peak 15-21). From the results, it is
noted that the breaking of P-Ph bonds (marked G, H in Fig. 11) in
PETIs-K is easier than that in PETPs, giving the retention of
phosphorus in the condensed phase.
Considering the results of EDX, ICP-AES, and XPS in char resi-
dues, we can conclude that the phosphorus of PETPs have flame
retardance activity both in gas and solid phases, while the phos-
phorus of PETIs-K just have the main flame retardance activity in
solid phase.
The detailed flame retardation process for PETPs may be
described as follows. During burning, PETPs will release the
phosphorus-containing heterocycle segments (Peak 14) to capture
radicals in air [55]. Part of the phosphorus remains in the
condensed phase to promote the carbonization. For PETIs-K, the
ionic aggregates play a key role in flame retardation. When heating,
the physical cross-linking networks caused by ionic aggregates are
enhanced and are able to restrict the movement of the melt. After
burning, the P-Ph bonds (marked G and H in Fig. 12) first break
leaving most of the phosphorus in residues. Then the phosphorus
fragments migrate to the surface and are soon oxidized to form
KH2PO4, which will convert to potassium polyphosphate (xK2O-
$yP2O5) at high temperatures [57,58]. At last, the cross-linking takes
place and the matrix char is induced.
The flame retardation process of PETIs-K in condensed phase is
illustrated in Fig. 13 [19,59]. Additionally, phenol and its derivatives
as decomposition products of PETIs-K are also radical scavengers
[60,61] and may arrest the free radicals or form polycyclic aromatic
hydrocarbons [25] in fire which benefit flame retardance.
Furthermore, multiple aromatic radicals caused by the cleaving of
DHPPO-K have some contributions to char formation (graphitiza-
tion) as well [62].
Fig. 11. Mass spectrum assigned to 10-phenylphenoxaphosphine-10-oxide (peak 14),
as a pyrolysis fragment of DCPPO-Ph in PETP10 chains.
promotes carbonization at high temperatures, which leaves the
majority of phosphorus in the condensed phase.
3.5. Pyrolysis behaviors of PETPs and PETIs-K
In order to investigate the detailed decomposition processes
based on the gaseous pyrolysis products, the pyrolysis experi-
ments of PET, PETP10, and PETI10-K were performed using Py-GC/
MS. The pyrograms are shown in Fig. 10, and the corresponding
peaks and their possible assignments are listed in Table 6. The
main pyrolysis products of PET are vinyl benzoate, benzoic acid, p-
methylbenzoic acid, p-ethylbenzoic acid, biphenyl, p-vinylbenzoic
acid, divinyl terephthalate, dimethyl terephthalate, and their de-
rivatives which are almost the same as that reported before
[53,54]. Small products such as CO, CO2, and H2O are not observed
because their molecular mass is under the detection limit. The
introduction of DCPPO-Ph does not significantly change the py-
rolysis products of backbone. However, PETP10 produces a greater
number of pyrolysis products due to the influence of DCPPO-Ph -
especially, a new compound (Peak 14) is detected and shown in
Fig. 11, which confirms a similar decomposition process like DOPO
and DPPO [55,56]. Peak 14 further proves the gaseous flame-
retardant activity of PETP10. As to the other pyrolysis products
4. Conclusions
Potassium-based ionomers and anaglous copolyesters with
phosphorus heterocycle integrated into the PET main chain were
successfully synthesized via random polycondensation. Although
they had a similar chemical structure, the ionomers (PETIs-K) and
copolyesters (PETPs) showed different rheological behaviors, ther-
mal stability, flame retardancy, and pyrolysis behaviors. The com-
plex viscosity of PETPs exhibited a linear decrease like neat PET;
while that of PETIs-K increased with the increase of temperature
due to the physical cross-linking from ionic aggregation, which was
beneficial to anti-dripping. Thermogravimetric analysis and ther-
mal degradation kinetic results showed that the incorporation of
DCPPO-Ph increased the thermal stability, and the incorporation of
DHPPO-K accelerated the decomposition of backbone to form sta-
ble residue char at high temperatures. Although both PETPs and
PETIs-K showed good flame retardancy in LOI and cone calorimetry
tests, ionization endowed the PETIs-K with better self-
extinguishing properties and restricted melt-dripping perfor-
mance than PETPs. More importantly, PETIs-K overcame the
shortcoming of the high smoke release during the combustion of
phosphorus-containing PET copolyesters. Furthermore, it was
found that phosphorus in PETPs and PETIs-K had quite different
flame retardance activities. Part of the phosphorus in PETPs vola-
tilized to the gas phase in a similar pyrolysis process as DPPO-ring.
Fig. 12. Proposed pyrolysis scheme of DHPPO-K in PETI10-K chains.

60
Y. Zhang et al. / Polymer 60 (2015) 50e61
Fig. 13. Proposed flame-retardant process of DHPPO-K in PETIs-K.
[4] Jing XK, Wang XS, Guo DM, Zhang Y, Zhai FY, Wang XL, et al. J Mater Chem A
2013;1:9264.
[5] Li YC, Mannen S, Schulz J, Grunlan JC. J Mater Chem 2011;21:3060.
[6] Carosio F, Alongi J, Malucelli G. J Mater Chem 2011;21:10370.
[7] Li YC, Mannen S, Morgan AB, Chang S, Yang YH, Condon B, et al. Adv Mater
2011;23:3926.
[8] Chen L, Ruan C, Yang R, Wang YZ. Polym Chem 2014;5:3737.
[9] Lin CH, Huang CM, Wang MW, Dai SA, Chang HC, Juang TY. J Polym Sci Part A
Polym Chem 2014;52:424.
[10] Jing XK, Guo DM, Zhang JB, Zhai FY, Wang XL, Chen L, et al. Ind Eng Chem Res
2013;52:4539.
However, most of the phosphorus in PETIs-K remained in the
condensed phase via ionization so as to promote the formation of
stable intumescent barriers to reduce the heat and smoke release.
Acknowledgments
This work was financially supported by the National Natural
Science Foundation of China (50933005 and 51121001), and the
Excellent Youth Foundation of Sichuan (2011JQ0007). The authors
would also like to thank the Analysis and Testing Center of Sichuan
University for the NMR and Raman measurements.
[11] Wang LS, Wang XL, Yan GL. Polym Degrad Stab 2000;69:127.
[12] Chang SJ, Sheen YC, Chang RS, Chang FC. Polym Degrad Stab 1996;54:365.
[13] Sato M, Endo S, Araki Y, Matsuoka G, Gyobu S, Takeuchi H. J Appl Polym Sci
2000;78:1134.
[14] Zhao HB, Chen L, Yang JC, Ge XG, Wang YZ. J Mater Chem 2012;22:19849.
[15] Zhao HB, Liu BW, Wang XL, Chen L, Wang XL, Wang YZ. Polymer 2014;55:
2394.
Appendix A. Supplementary data
[16] Jenkins JE, Seitz ME, Buitrago CF, Winey KI, Opper KL, Baughman TW, et al.
Polymer 2012;53:3917.
[17] Ge XG, Wang C, Hu Z, Xiang X, Wang JS, Wang DY, et al. J Polym Sci Part A
Polym Chem 2008;46:2994.
[18] Wang JS, Zhao HB, Ge XG, Liu Y, Chen L, Wang DY, et al. Ind Eng Chem Res
2010;49:4190.
[19] Zhang Y, Chen L, Zhao JJ, Chen HB, He MX, Ni YP, et al. Polym Chem 2014;5:
1982.
The preparation procedure of DCPPPO-Ph, the NMR spectra of
DPPO-Ph, DCPPO-Ph, DHPPO-K, PETPs, and PETIs-K, and the table of
the correlation coefficient of activation energy of samples according
to Ozawa's method are provided in supporting information.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.polymer.2015.01.030.
[20] Mehta RD, Hoque AKMM. Textile Res J 1982;52:607.
[21] DeGroot WF, Shafizadeh F. J Anal Appl Pyrolysis 1984;6:217.
[22] Kleen M, Gellerstedt G. J Anal Appl Pyrolysis 1995;35:15.
[23] de Britto D, Assis OBG. Thermochim Acta 2009;494:115.
[24] Nowakowski DJ, Jones JM. J Anal Appl Pyrolysis 2008;83:12.
[25] Hajaligol MR, Sharma RK. J Anal Appl Pyrolysis 2003;66:123.
[26] Song P, Liu H, Shen Y, Du B, Fang Z, Wu Y. J Mater Chem 2009;19:1305.
References
[1] Eslami H, Müller-Plathe F. Macromolecules 2009;42:8241.
[2] Ronkvist ÅM, Xie W, Lu W, Gross RA. Macromolecules 2009;42:5128.
[3] Wang YZ, Chen XT, Tang XD, Du XH. J Mater Chem 2003;13:1248.

Y. Zhang et al. / Polymer 60 (2015) 50e61
61
[27] Song P, Zhao L, Cao Z, Fang Z. J Mater Chem 2011;21:7782.
[46] Feng C, Zhang Y, Liu S, Chi Z, Xu J. Polym Degrad Stab 2012;97:707.
[47] Han J, Liang G, Gu A, Ye J, Zhang Z, Yuan L. J Mater Chem A 2013;1:2169.
[48] Tai Q, Hu Y, Yuen RKK, Song L, Lu H. J Mater Chem 2011;21:6621.
[49] Ma H, Fang Z. Thermochim Acta 2012;543:130.
[50] Bourbigot S, Bras ML, Delobel R, Gengembre L. Appl Surf Sci 1997;120:15.
[51] Bourbigot S, Bras ML, Gengembre L, Delobel R. Appl Surf Sci 1994;81:299.
[28] Greener J, Gillmor JR, Daly RC. Macromolecules 1993;26:6416.
[29] Gao Y, Choudhury NR, Dutta NK. J Appl Polym Sci 2012;124:2908.
[30] Spencer MW, Wetzel MD, Troeltzsch C, Paul DR. Polymer 2012;53:555.
[31] Spencer MW, Wetzel MD, Troeltzsch C, Paul DR. Polymer 2012;53:569.
[32] Ozawa T. Bull Chem Soc Jpn 1965;38:1881.
[33] Wang DY, Wang YZ, Wang JS, Chen DQ, Zhou Q, Yang B, et al. Polym Degrad
Stab 2005;87:171.
[34] Chen HB, Dong X, Schiraldi DA, Chen L, Wang DY, Wang YZ. J Anal Appl Py-
rolysis 2013;99:40.
[52] Tsai KC, Kuan CF, Chen CH, Kuan HC, Hsu SW, Lee FM, et al. J Appl Polym Sci
2013;127:1084.
[53] Balabanovich AI, Pospiech D, Haußler L, Harnisch C, Doring M. J Anal Appl
Pyrolysis 2009;86:99.
€
€
[35] Chen HB, Zhang Y, Chen L, Zhao HB, Liu Y, Wang YZ. Ind Eng Chem Res
2010;49:7052.
[36] Shau MD, Wang TS. J Polym Sci Part A: Polym Chem 1996;34:387.
[37] Ryu BY, Emrick T. Angew Chem Int Ed Engl 2010;49:9644.
[38] Morgan AB, Bundy M. Fire Mater 2007;31:257.
[54] Yoshioka T, Handa T, Grause G, Lei Z, Inomata H, Mizoguchi T. J Anal Appl
Pyrolysis 2005;73:139.
€
€
[55] Schafer A, Seibold S, Lohstroh W, Walter O, Doring M. J Appl Polym Sci
2007;105:685.
€
[56] Fischer O, Pospiech D, Korwitz A, Sahre K, Haußler L, Friedel P, et al. Polym
[39] Schartel B, Hull TR. Fire Mater 2007;31:327.
Degrad Stab 2011;96:2198.
[40] Hull TR, Stec AA. Polymers and fire. In: Hull TR, Kandola BK, editors. Fire
retardancy of polymers-new strategies and mechanisms. Cambridge, U.K:
Royal Society of Chemistry; 2009. p. 8.
[41] Yan YW, Chen L, Jian RK, Kong S, Wang YZ. Polym Degrad Stab 2012;97:1423.
[42] Eklund PC, Holden JM, Jishi RA. Carbon 1995;33:959.
[43] Chen Z, Zhao L, Wang Z. Polymer 2013;54:5182.
[57] Banach M, Makara A. J Chem Eng Data 2011;56:3095.
[58] Osterheld RK, Audrieth LF. J Phys Chem 1952;56:38.
[59] Huang N, He J, Du J, Wang J. J Fire Sci 2012;30:256.
[60] Le HT, Flammang R, Gerbaux P, Bouchoux G, Nguyen MT. J Phys Chem A
2001;105:11582.
[61] Wright JS, Johnson ER, DiLabio GA. J Am Chem Soc 2001;123:1173.
[62] Patel P, Hull TR, McCabe RW, Flath D, Grasmeder J, Percy M. Polym Degrad
Stab 2010;95:709.
[44] Pimenta MA, Dresselhaus G, Dresselhaus MS, Canҫado LG, Jorio A, Saito R.
Phys Chem Chem Phys 2007;9:1276.
[45] Yang W, Sun WJ, Chu W, Jiang CF, Wen J. Chin Chem Lett 2012;23:363.