Effect of a Type of Triazine on the Properties and

Morphologies of Poly(butylene terephthalate) Composites

Jun Chen, Dongyu Cai, Weixi Jin, Fei Wu, Xianyi Chen
College of Materials Science and Engineering, Nanjing University of Technology, Nanjing 210009 China

Received 22 April 2005; accepted 6 October 2005

DOI 10.1002/app.23404
Published online in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: Composites of poly(butylene terephthalate) ric study revealed that the weight loss of the PBT/TTA
(PBT), 2,4,6-tris(2⬘,4⬘,6⬘-tribromophenoxy)-1,3,5-triazine (TTA), composite was very limited in the temperature range of
and glass fibers were prepared, and the effect of TTA on the 25–300°C, and scanning electron microscopy images of the
properties and morphologies of the composites was studied. blend demonstrated that the TTA particle sizes and their
The results showed that the addition of a suitable amount of distribution in the PBT matrix remained thermally stable
TTA could improve the flame retardancy of PBT composites when the system was heated at 130°C for 3 h. This suggested
reinforced with glass fibers, and good resistance to TTA good compatibility of TTA with PBT. © 2006 Wiley Periodicals,
emigration from the inside of the composites onto their Inc. J Appl Polym Sci 102: 1291–1296, 2006
surfaces was obtained. Fourier transform infrared spectros-
copy analyses of PBT, TTA, and their blend suggested that
there might be no chemical bonds formed on the interfaces Key words: flame retardance; mechanical properties; mor-
between PBT and TTA in the composite; a thermogravimet- phology; polyesters

INTRODUCTION the surface of the products under unfavorable injec-

tion-molding conditions may cause the unwanted con-
The performance and processability of poly(butylene
tamination of the surfaces of the injection mold. After
terephthalate) (PBT) as a kind of engineering plastic
some injection cycles, a layer of thick and somewhat
have been investigated extensively for the purpose of
sticky DBDPO on the mold surfaces can appear; this is
its wide use in electric appliances, electronic connec-
called a juicing phenomenon. The use of a macromo-
tors, and car parts such as lamp holders, switches,
lecular flame retardant is a good way to solve the
circuit breakers, and motor casings that require PBT
problem, but its relatively high price will certainly
composites with a high flame-retardant property.1– 4
affect its wide range of applications. Therefore, many
In many cases, this property of the composites can be
studies have been carried out on the possibility of the
obtained by the incorporation of flame retardants into
replacement of DBDPO with a kind of effective flame
the system. During the last several decades, flame
retardant with a low molecular weight, low price, and
retardants for PBT have been developed, ranging from
excellent antiblooming property.12–15
halogen-containing flame retardants to halogen-con-
A kind of flame retardant with a low molecular
taining polymeric flame retardants and halogen-free
weight, 2,4,6-tris(2⬘,4⬘,6⬘-tribromophenoxy)-1,3,5-tri-
flame retardants.5–11
azine (TTA) has been used for the preparation of
Despite having a number of technical advantages in
fire-resistant polyolefins and some other thermoplas-
applications, low-molecular-weight halogen-contain-
tic composites as well as fibers with reduced bleeding
ing flame retardants such as decabromodiphenyl ox-
of fire retardants.16 However, to the best of our knowl-
ide (DBDPO) may be prone to emigration from the
edge, there has been no report on its application to
inside of the composites onto their surfaces when the
PBT or PBT composites. In this article, we describe the
composites are heated, that is, blooming on the surface
preparation of PBT and PBT/glass-fiber composites
of the composites due to the poor compatibility of
incorporated with TTA and investigate its influence
DBDPO with the matrix, which will directly affect the
on the flame-retardant, antiblooming, and mechanical
properties of the composites, such as the flame-retar-
properties and morphologies of the PBT composites.
dant and electrical properties and adhesion strength,
Moreover, a comparison of TTA and DBDPO is made
thus resulting in difficulty in the practical use of the
with respect to their contributions to the properties
composites. Furthermore, the blooming of DBDPO on
and morphologies of the composites.

Correspondence to: X. Chen (xychen@njut.edu.cn). EXPERIMENTAL

Journal of Applied Polymer Science, Vol. 102, 1291–1296 (2006) The PBT resin (BST L2100), provided by the Yizheng
© 2006 Wiley Periodicals, Inc. Chemical Fiber Group (People’s Republic of China),
1292 CHEN ET AL.

TABLE I stages were 40, 20, and 1 s, respectively. The pellets

Recipes for PBT Sample Preparation were melted and injected into a family mold that
PBT Glass fiber TTA DBDPO Sb2O3 consisted of bars for tensile testing (ASTM D 638),
Sample (wt %) (wt %) (wt %) (wt %) (wt %) Izod impact testing (ASTM D 256), and flexural testing
PBT1-1 90.0 10.0 0 0 0
(ASTM D 790A). The bars for the Izod impact test
PBT1-2 83.3 10.0 5.0 0 1.7 were then notched according to ASTM D 256.
PBT1-3 76.7 10.0 10.0 0 3.3 The tensile strength, flexural strength, and notched
PBT1-4 70.0 10.0 15.0 0 5.0 Izod impact strength of the samples were measured
PBT2-1 100.0 0 0 0 0 according to related ASTM standards at 23°C. Ten
PBT2-2 93.3 0 5.0 0 1.7
PBT2-3 86.7 0 10.0 0 3.3
specimens were tested, and their results were aver-
PBT2-4 80.0 0 15.0 0 5.0 aged for every datum. The notched Izod impact
PBT3-1 86.7 0 0 10.0 3.3 strength test was carried out with a UJ-4 impact tester
PBT3-2 76.7 10.0 0 10.0 3.3 (Chengde Test Equipment Co., Ltd.); the tensile
PBT3-3 66.7 20.0 0 10.0 3.3 strength and flexural strength measurements were
PBT3-4 56.7 30.0 0 10.0 3.3
made on a CMT5254 tensile tester (Precision Rank 0.5,
Xin San Si Measurement Technology Co., Ltd.). Lim-
iting oxygen index (LOI) tests of the samples were
was in the form of pellets with an intrinsic viscosity of conducted with an HC-2 oxygen index instrument
1.00 dL/g, a carboxyl end group concentration of 30 (Jiangning Analysis Instrument Factory) according to
mmol/kg or less, a melting point of 225°C or higher, a ASTM D 2863; the specimen size was 0.65 cm ⫻ 0.3 cm
color L value of 90 or higher, a color b value of 3.5 or ⫻ 7–15 cm. The melt flow index of the samples was
less, a specific gravity of 1.30 –1.32 g/cm3, a melt flow obtained at 230°C with a load of 2.16 kg (ASTM D
index of 23–28 g/10 min, a tensile strength of 50 – 60 1238) with an XNR-400A melt flow indexer (Chang-
MPa, an elongation value of 200% or higher, a notched chun Second Test Machine Factory). We evaluated the
Izod impact strength of 40 J/m or higher, and a di- blooming performance of the flame retardants on the
electric strength of 20 kV/mm or higher. TTA, a white surface of the PBT composites by rubbing them
powder with a specific gravity of 2.44 g/cm3, a purity against the surface of the samples, which were heated
grade of 95%, and a theoretical bromine concentration in air at 130°C for 3 h, and then observing whether or
of 66.7 wt %, was supplied by the Nanjing Research not an off-white powder appeared.
and Design Institute of Chemical Industry. DBDPO The structures of flame retardant TTA, PBT, and
(DE-83R) was acquired from Great Lake (United PBT composites were analyzed with the help of their
States). The short, cut E-glass fiber was in the form of Fourier transform infrared (FTIR) spectra, which were
chopped strands with a diameter of 13 ␮m and a recorded from either a KBr/TTA powder mixture or
length of 3 mm (183 Fiberglass, Owens–Coring Fiber- films of PBT and its composites with an FTIR spectro-
glass Corp.). photometer (Nexus 670 FTIR, Nicolet Co., United
The recipes for the sample preparation are listed in States). The number of scans per spectrum was 24, and
Table I; the total amount of a sample was 100 parts in the spectral resolution was 2 cm⫺1. Before the FTIR
weight, and the ratio of TTA or DBDPO to Sb2O3 was measurements, the samples were dried in vacuo over-
3:1 because of the practical use of a recipe in industry night. The films of PBT and its composites were pre-
and the synergistic flame-retardant effect. A typical pared via hot pressing at 220°C. The sample morphol-
sample preparation process was as follows. A mixture ogy was observed with scanning electron microscopy
of thoroughly dried materials (PBT, flame retardants, (SEM; JSM-5900, JEOLTD/Japan Electric) at an accel-
and glass fibers) was melt-blended in a contrarotating, erating voltage of 15 kV and a probe current of 20 pA.
twin-screw extruder (TE-35, Keya Plastic Machine Fac- For SEM measurements, the injection-molded bars
tory) with a diameter of 36 mm and a ratio of the were freeze-fractured at the temperature of liquid N2
length to the diameter of 36. The blend was then and then sputter-coated with gold with an JFC-1600
granulated. The screw speed of the extruder was 160 auto fine coater. Thermogravimetric analysis (TGA) of
rpm, and the temperatures of the four bands of the TTA, its composite with PBT, and DBDPO was carried
barrel from the feed section to the die of the extruder out with a Netzsch analyzer between 25 and 300°C in
were 245, 250, 250, 255, and 250°C, respectively. The air at a heating rate of 5°C/min. All samples were
samples for testing the mechanical properties were dried in vacuo at room temperature for 24 h before the
prepared by injection molding. The temperatures of TGA runs.
the three zones in the injection-molding machine
(CJ80M3V, Zhengde Plastic Machine Factory) were
RESULTS AND DISCUSSION
235, 240, and 240°C. The mold temperature was about
80°C, and the injection pressure was maintained at 50 Flame-retardant properties are very important for PBT
MPa. The times for the injection, cooling, and ejection and its glass-fiber-reinforced composites in many
POLY(BUTYLENE TEREPHTHALATE) COMPOSITES 1293

TABLE II
Comparison of the Blooming Performances of TTA and DBDPO from PBT Composites
Sample PBT1-3 PBT1-4 PBT2-3 PBT2-4 PBT3-1 PBT3-2 PBT3-3 PBT3-4

Blooming F F F F E E E E

E ⫽ blooming; F ⫽ nonblooming.

cases. However, if the flame retardant incorporated and PBT because an ordinary filler normally has a
into the PBT system is not compatible with the matrix negative effect on the mechanical properties of its
and continues to transfer to the surfaces of the prod- composites with polymers. Moreover, the PBT com-
ucts in use, the flame-retardant properties of the prod- posite reinforced with 10 wt % glass fiber showed
ucts will be severely degraded after use for some tensile and flexural strengths of about 80 and 110 MPa,
period of time. Therefore, nonblooming characteristics respectively, exhibiting much better mechanical prop-
of the flame retardant are required to maintain the erties than the PBT/TTA composite to which no glass
permanent flame-retardant properties of the products fiber was added. These results could be easily under-
under their use conditions. In this respect, the bloom- stood with respect to the contribution by the glass
ing performance of TTA from PBT composites and a fiber. However, the notched Izod impact strength of
comparison of this behavior of TTA with that of DB- the composites either with or without glass-fiber rein-
DPO commonly used in industry were investigated; forcement deteriorated with increasing amounts of
the results are summarized in Table II. When flame TTA and Sb2O3; moreover, a higher impact strength
retardants TTA and DBDPO were added to PBT com- was obtained when glass fiber was added to the PBT/
posites, their blooming performances were strikingly TTA composite. This was similar to the case of the
different. For the PBT blends with flame retardant tensile and flexural strengths. Generally, the impact
TTA, no frost phenomenon could be observed on their strength of materials is related to the bond strength of
surfaces even after the composite containing as much molecules and polymer chain movement under cer-
as 15 wt % TTA was heated at 130°C for 3 h. The tain stress, as well as the stress distributions in the
results indicated that no observable number of TTA materials caused by this movement. In other words,
particles could emigrate onto the surface from the the impact strength actually reflects the flexibility of
inside of the composite, probably because of the good the materials. For glass-fiber-reinforced PBT compos-
compatibility of TTA with the matrix; therefore, TTA ites, the good flexibility of the PBT chain makes the
demonstrated an excellent antiblooming ability. How- energy transmission easier on the one hand; on the
ever, off-white frost could be easily and quickly found other hand, the interface of the glass fiber and matrix
on the surface of PBT containing 10 wt % DBDPO is improved after the glass-fiber surface treatment.
when the blend was heated for a very short time at the Under these conditions, the impact load can be easily
same temperature at which the PBT/TTA composite transmitted from the matrix to the glass fiber, leading
was heated. Further experiments showed that DBDPO to the so-called crazing phenomenon in the composite.
blooming could occur even if a very low amount of the
flame retardant was added to the PBT system, sug-
gesting that thermal emigration of the flame retardant TABLE III
from the inside of the composite onto its surface could Mechanical and Flame-Retardant Properties of PBT
easily take place because of the incompatibility of Composites Reinforced with Different Contents
DBDPO with the polymer matrix. of Glass Fibers
The effects of different amounts of PBT and TTA Tensile Flexural Impact Flame
together with Sb2O3 on the mechanical and flame- strength strength strength retardancy
retardant properties of PBT and its composites rein- Sample (MPa) (MPa) (J/m) LOI (%)
forced with glass fibers are demonstrated in Table III PBT1-1 77.6 107.7 57.1 20.2
(entries 1– 8). The tensile strength remained stable for PBT1-2 80.3 114.0 54.3 21.0
either PBT/TTA or PBT/TTA/glass-fiber composites, PBT1-3 80.0 115.5 48.4 25.6
regardless of the amounts of the components in the PBT1-4 83.3 115.7 38.2 31.3
PBT2-1 57.1 83.2 47.4 23.3
range of our study. The addition of TTA and the
PBT2-2 59.1 98.7 44.3 23.5
synergistic flame retardant Sb2O3 initially made the PBT2-3 57.8 100.0 36.4 23.5
flexural strength of the composites in both cases PBT2-4 60.5 100.0 32.3 24.7
slightly better, but the improvement leveled off when PBT3-1 56.5 107.0 33.0 24.1
the concentration of TTA was larger than about 5 wt PBT3-2 64.1 115.8 43.4 25.2
PBT3-3 89.1 134.7 64.9 32.8
%. The tensile and flexural properties of the compos-
PBT3-4 109.3 155.6 81.2 33.5
ites suggested favorable interactions between TTA
1294 CHEN ET AL.

In this way, the impact strength of PBT composites

would be improved by the addition of glass fiber.
Moreover, the effect of glass fiber on the fluidity of
PBT/TTA composite was investigated. The melt flow
index of PBT containing 10.0 wt % TTA and 3.3 wt %
Sb2O3 exhibited a trend of a steady decline with an
increasing amount of glass fiber, indicating that the
addition of glass fiber reduced the fluidity of the com-
posite because it enhanced the inner friction in the
composite, therefore hindering the flow of the melting
composite.
For comparison, measurements of the mechanical
Figure 1 TGA curves for (a) TTA, (b) the PBT/TTA/Sb2O3
properties of PBT/DBDPO composites with and with- (80/15/5 w/w/w) composite, and (c) DBDPO.
out glass-fiber reinforcement were carried out, and the
results are summarized in Table III. The data listed in
entries 3 and 10 and entries 7 and 9 of Table III
indicate that both PBT and its composite with 10 wt % retardants and PBT composite remained thermally
glass fiber, when incorporated with TTA, had proper- stable in the temperature range of our study; there-
ties similar to those when they were incorporated with fore, HBr evolved and the corrosion caused in this
DBDPO, with the exception of the tensile strengths of time was very limited.
the glass-fiber-reinforced composites. The properties of PBT composites certainly depend
According to Table III (column 5), the LOI, that is, on the components of the composites and the struc-
the flame retardancy of the composites, strengthened tures and interfacial actions of the components; there-
with the amount of TTA, and this trend was even fore, FTIR and SEM investigations of PBT composites
obvious when the amount of TTA increased from 5 to were carried out. The FTIR spectra of pure PBT, flame
15 wt % for the glass-fiber-reinforced PBT composite; retardant TTA, and their composite are given in Fig-
however, only a slight increase in the flame retardancy ure 2. In the case of pure PBT, a strong characteristic
could be attained for the composite without glass band at 1713 cm⫺1 was due to the absorption of the
fiber. Very interestingly, when the addition of TTA carbonyl group, and the absorption peaks in the 1457–
was less than around 8 wt %, the PBT/TTA/glass- 1577-cm⫺1 region could be assigned to the stretching
fiber composite did not show as good a flame-retar- vibration of the benzene ring and COH bending of
dant property as the PBT/TTA composite; however, methylene. The broad and strong absorption band
this property of the former composite was improved near 2962 cm⫺1 could be used to characterize the
greatly when the amount of TTA exceeded the critical COH stretching vibration of the benzene ring and
point. There is no doubt that the difference in the OCH2O in the PBT skeleton. Moreover, the absorp-
flame retardancy for the two composites should be tion peaks in the 680 –1000-cm⫺1 region from out-of-
attributed to the glass-fiber contribution. It is believed plane COH bending of the benzene ring could be
that glass fiber self-behaves as a flame retardant; at the observed. For the FTIR spectrum of TTA, strong bands
same time, its interface with the matrix may provide at 1234 cm⫺1 and in the 1360 –1583-cm⫺1 region cor-
more opportunities for flame. Therefore, the total ef- responded to the COOOC vibration of aromatic ether
fect of glass fiber on the flame retardancy of a com- and stretching motions of both benzene and triazine
posite should be dependant on which factor is domi- rings, respectively. The weak peak at 620 cm⫺1 was
nant in the system. Similar results for DBDPO/PBT attributed to the characteristic absorption of COBr.
and its glass-fiber composite (Table III) could offer The FTIR spectrum of the PBT/TTA composite shows
strong support for this explanation. Thus, it can be that the addition of TTA neither changed the charac-
concluded that both the mechanical properties and teristic bands of pure PBT nor produced new absorp-
flame retardancy of PBT composites will be affected tion bands, suggesting that there might be no new
by the addition of TTA and glass fiber, and the chemical bonds formed between TTA and PBT in the
amounts of the addition should be optimized with composite. However, the flame retardant indeed had
respect to the comprehensive properties of the com- an excellent antiblooming ability, as discussed previ-
posites. ously; therefore, the good compatibility of TTA with
The thermal resistance of TTA, the PBT/TTA/Sb2O3 PBT, which could lead to a stable dispersion of TTA in
(80/15/5 w/w/w) composite, and DBDPO is demon- the matrix, may make it understandable why no ob-
strated in Figure 1. The weight losses of TTA, its servable thermal emigration of TTA from the inside of
composite with PBT, and DBDPO were 1.4, 2.7, and the matrix onto its surface could occur.
1.7%, respectively, quite similar if the errors of the The morphology of the PBT composites with both
measurements were considered. Obviously, the fire TTA and DBDPO was investigated with SEM. Figure
POLY(BUTYLENE TEREPHTHALATE) COMPOSITES 1295

Figure 2 FTIR spectra of TTA, PBT, and their composite.

3 shows the results of fractured surfaces of the PBT/ stable even after the sample was heated at 130°C for
TTA (90/10 w/w) composite with different thermal 3 h [Fig. 3(b)]. A close inspection of the SEM images
histories. As shown in Figure 3(a), TTA particles were with relatively high magnification for the same sam-
well dispersed in the PBT matrix before heating. ples indicated that the particle sizes, their distribution,
Moreover, the dispersion of TTA particles remained and the dispersion of the fire retardant remained sta-
ble when the composite was under different thermal
conditions; this provides convincing evident for the
good compatibility of TTA and PBT. The combination
of these SEM results and the FTIR spectra should give
a reasonable explanation for the excellent antibloom-
ing property of TTA in the PBT composites because no
obviously observable amount of TTA emigrated from
the inside of the composite onto its surface according
to the thermally stable sizes, size distribution, and
dispersion of TTA particles in the PBT composite.
The SEM micrographs of fractured surfaces of the
PBT/DBDPO (90/10 w/w) blend under different ther-
mal environments are illustrated in Figure 4(a,b).
Small DBDPO particles were relatively well dispersed
in the PBT matrix before heating [Fig. 4(a)]. However,
when the composite was heated for a short time, the
DBDPO particles were prone to agglomerate in the
inside of the blend; therefore, many large DBDPO
particles were detected [Fig. 4(b)]. Moreover, a poor
TTA distribution and obvious changes in the particle
size distribution in the composite were observed be-
cause of the agglomeration of DBDPO particles and
their thermal emigration from the inside of the com-
posite onto its surface to reduce the surface energy of
the particles and increase their stability in the PBT
matrix. Therefore, the phenomenon of whitening ap-
pearing on the surface of the composite inevitably
occurred, and this was in good agreement with our
Figure 3 SEM micrographs of PBT/TTA composites observations. These results suggested that the compat-
(90/10 w/w): (a) before heating and (b) after heating at ibility of DBDPO with PBT should not be as good as
130°C for 3 h. that of TTA with PBT.
1296 CHEN ET AL.

obviously improve the flame retardancy of the com-

posites and keep good resistance to its emigration
from the inside of the composites onto their surface
when the composites are heated. The excellent anti-
blooming property of TTA can be attributed to its
good compatibility with the PBT matrix because the
structure analysis by FTIR suggests that there are no
chemical bonds formed on the interfaces of the TTA
particles and PBT matrix in the PBT/TTA blend, and
morphology images by SEM indicate that the particle
sizes and their distribution in the matrix remain ther-
mally stable. The tensile and flexural strengths of PBT
composites remain stable, but the impact strength de-
teriorates with an increasing amount of TTA. In con-
trast, DBDPO exhibits a great capacity for blooming
on the surface of PBT composites. SEM studies reveal
that DBDPO particles are prone to aggregate in DB-
DPO/PBT blends, which are supposed to be heated
for a short time, suggesting that the compatibility of
DBDPO is not as good as that of TTA with PBT.

References

1. Jeng, M.-C.; Fung, C.-P.; Li, T.-C. Wear 2002, 252, 934.
Figure 4 SEM micrographs of PBT/DBDPO composites 2. Liang, J. Z.; Tang, C. Y.; Lee, W. B. J Mater Process Technol 1995,
(90/10 w/w): (a) before heating and (b) after heating at 48, 707.
130°C for 10 min. 3. Takashi, A.; Angola Juan, C.; Hiromu, S.; Takashi, I.; Yasushi, N.
Polymer 1999, 40, 3657.
The reasons for the great difference in the compat- 4. Jiang, X.-M.; Zhao, S.; Chen, X.-Y. Zhongguo Suliao (in Chinese)
ibilities of TTA and DBDPO with PBT are not known. 2001, 15, 32.
The following factors can be considered. Obviously, 5. Balabanovich, A. I.; Engelmann, J. Polym Degrad Stab 2003, 79,
85.
TTA and DBDPO are different in terms of components
6. Balabanovich, A. I.; Levchik, G. F.; Levchik, S. V.; Engelmann, J.
and structure, and the molecular weight and spherical J Fire Sci 2002, 20, 71.
hindrance of TTA are definitely larger than those of 7. Levchik, S. V.; Bright, D. A.; Alessil, G. R. Polym Degrad Stab,
DBDPO. Moreover, it seems that DBDPO particles in 2002, 77, 267.
the PBT matrix are spherical and cylinder-like before 8. Pellow-Jarman, M.; Hetem, M. Polym Degrad Stab 1995, 47, 413.
and after agglomeration, respectively, with a rela- 9. Asrar, J.; Berger, P. A.; Hurlbut, J. J Polym Sci Part A: Polym
tively smooth surface morphology. The interfaces be- Chem 1999, 37, 3119.
10. Wang, J. L.; Favstritsky, N. A. Polym Mater Sci Eng 1997, 77,
tween the DBDPO particles and PBT matrix is clear.
511.
However, compared with DBDPO, TTA particles are 11. Jpn. Pat. JP 22003055434.
smaller, more irregular, and nearly monodisperse; 12. Jpn. Pat. JP 22002080731.
their surfaces appear to be somewhat uneven. 13. Balabanovich, A. I.; Balabanovich, A. M.; Engelmann, J. Polym
Int 2003, 52, 1309.
CONCLUSIONS 14. Ger. Pat. DE 10218902.
15. Levchik, G. F.; Grigoriev, Y. V.; Balabanovich, A. I.; Levchik,
A type of triazine (TTA), when incorporated into S. V.; Klatt, M. Polym Int 2000, 49, 1095.
PBT/glass-fiber composites as a flame retardant, can 16. Jpn. Pat. JP 2004099780.