Fulltext 5
Fulltext 5
An overview is presented of the literature on the thermal decomposition and combustion of ther-
moplastic polyesters, especially commercially important poly(ethylene terephthalate) (PET) and
poly(1,4-butylene terephthalate) (PBT). Although the literature is not clear as to whether heterolytic
or homolytic scission of aliphatic fragments is the first step in the thermal decomposition of polye-
sters, in any case volatilization of light aliphatic fragments make polyesters easily ignitable poly-
mers. Despite the presence of benzene groups in the main polymer chain, thermoplastic polyesters
show very limited tendency to char, but instead, aromatic-containing polymer fragments volatilize
and feed the flame. Fire retardant additives, although they usually facilitate decomposition of the
polyesters at lower temperature, also usually promote charring and therefore suppress combustion.
Copyright # 2004 John Wiley & Sons, Ltd.
Scheme 2.
Scheme 3.
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700
Thermal decomposition and combustion of thermoplastic polyesters 693
Scheme 4.
Scheme 5.
Scheme 6.
Scheme 7.
debatable. It may be that accumulation of vinyl ester groups Khemani11 studied the thermal degradation of PET at
allows crosslinking leading to the formation of polyene 2808C over an extended period of time and found a gradual
segments or self-condensation of retained acetaldehyde decrease in the amount of acetaldehyde evolved with time.
produces aldol products which condense and dehydrate to Three different routes generating acetaldehyde: (a) the HO-
form color bodies. end groups; (b) the vinyl end groups; (c) the mid-polymer
Vijaykumar and coworkers10 suggested the formation of chain scission which also generates vinyl end groups were
acetaldehyde, the major pyrolysis product of PET, through discussed. The mechanism of thermal decomposition of
recombination of vinyl-terminated carboxylate and car- chain ends is shown in Scheme 7. Evolution of dioxane in the
boxylic acid-terminated units (Scheme 5). This reaction may low temperature decomposition of PET was also reported by
proceed by the intra- or intermolecular H shift to form vinyl Hovenkamp and Munting.12
alcohol, which isomerizes to give acetaldehyde. The homolytic scission of ester bonds was suggested by
Apart from the formation of acetaldehyde, a considerable Bounekhel and McNeill13 based on their thermal decomposi-
amount of ethylene is also formed. It was suggested that the tion study in vacuum of polyester made from poly(ethylene
vinyl-terminated carboxylate units and the PET units form an glycol) and terephthalic acid. They suggested that initial
eight-membered cyclic transition state to give C2H4 and homolysis can occur in two possible ways, involving either
carboxylic acid-terminated units (Scheme 6). alkyl-oxygen (a) or acyl-oxygen (b) scission (Scheme 8). These
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700
694 S. V. Levchik and E. D. Weil
Scheme 8.
authors suggest that homolytic reactions can explain the
whole range of observed degradation products. Authors
argued that products such as CO and CO2 are formed
throughout the main temperature range of decomposition,
whereas by a non-homolytic route, such products would only
be found at the highest degradation temperatures (in the Scheme 9.
range up to 5008C). They also suggested that radical-trapping
agents would only inhibit a chain-reaction mechanism and reaction of the formed terephthalic acid. (Scheme 10) This
would be ineffective in the proposed decomposition because mechanism was supported by the fact that terephthalic acid
homolysis takes place in a cage in which the pair of radicals yields biphenyl as well as benzene at elevated temperatures
then undergo disproportionation. above 5008C. It is interesting to notice that the formation of
Day et al.14 studied the thermal decomposition of model both benzene and biphenyl decreased as the number of
n-propyl pyrene-1-carboxylate and found propylene and methylene groups in the polymer chain rose.
pyrene-1-carboxylic acid as the major reaction products,
presumably formed via a six-centered intramolecular transi- Poly(1,4-butylene terephthalate) (PBT)
tion state. However, some n-propanol was formed as a result The initiation of the thermal decomposition of PBT is similar
of a radical cleavage, but the contribution of this reaction to to that of PET, however the products of decomposition are
the overall decomposition appeared to be insignificant. somewhat different because of longer aliphatic fragments
Divinyl terephthalate (I) was considered to be a primary in the chain. Several authors10,16,17 report initial polymer scis-
scission product, as found in another study run by Day and sion occurring via the six-membered cyclic transition state.
coworkers.15 Vinyl acid terephthalate and terephthalic acid According to the actual geometry of the chains, the major
were not detected but were not ruled out since the products of degradation route in PBT would lead to the formation of
decarboxylation of these two acids were detected, namely butadiene (Scheme 11).
vinyl benzoate, benzoic acid, and benzene (Scheme 9). Apart from butadiene a considerable amount of tetrahy-
drofuran (THF) is also obtained in the degradation products
of PBT. When the well-known acyl-oxygen cleavage of the
ester linkages proceeds in the PBT chains intra- or inter-
molecular H shifts can occur, leading to the formation of
hydroxyl-terminated units, which in turn undergo further
degradation to yield THF and the carboxylic acid-terminated
At elevated temperatures the acids from terminal groups chain10 (Scheme 12).
rapidly became converted to the more stable phenyl Another less feasible mechanism for the formation of THF
terminals by elimination of carbon dioxide.16 Detected and butadiene from PBT by the intra- or intermolecular H
biphenyl was most likely formed through the bimolecular shift can form a diol which eliminates a molecule of water to
Scheme 10.
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700
Thermal decomposition and combustion of thermoplastic polyesters 695
Scheme 11.
Scheme 14.
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700
696 S. V. Levchik and E. D. Weil
Scheme 15.
Scheme 16.
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700
Thermal decomposition and combustion of thermoplastic polyesters 697
Scheme 18.
and accompanied by char formation to decrease flammable evolution of a large amount of carbon dioxide (incombustible
products. The observed initial temperature of the evolution of gas), which might be a decomposition product of polycarbo-
SbBr3 and its flame-poisoning effect nearly coincides with nate, dominated in the extinguishment stage.
that of HBr evolution. PET fibers containing commercial flame retardants,
A FT-IR gas cell was used for comparison of thermal namely Dacron 900F (containing ethoxylated tetrabromobi-
decomposition of plain PBT and two formulations containing sphenol A comonomer), TRIS (containing the now-
different highly brominated aromatic compounds as flame discontinued tris(2,3-dibromopropyl) phosphate), AB19
retardants.24 Samples with brominated flame retardant (containing cyclic phosphonate ester, Antiblaze 19, now
degrade primarily to CO2, butadiene and PBT chain Amgard CU), and Trevira T271 (containing phosphonic acid
fragments. One formulation was found to degrade at a comonomer) were compared with non-flame-retarded fibers
significantly higher temperature than another, the first one Dacron T54 and MICH using thermogravimetry.26 It was
only started degrading at 2958C whereas another one concluded that the presence of phosphorus-only flame
released PBT oligomers at 2258C. The thermal stability of retardants (AB19 and T271) caused no large differences in
the stable flame retardant formulation was similar to non- the measured interval of thermal decomposition. In the case
fire-retardant PBT. Plain PBT started to degrade at around of the bromine-containing polyesters (i.e. 900F and TRIS)
3008C, while the gas spectra of the even less stable flame some changes were noted. With the 900F compound the
retardant additive showed degradation above 3608C. The degradation was apparently retarded, whereas with the TRIS
authors concluded that it is unlikely that the stable flame polyester system, the volatilization of gaseous fuel proceeded
retardants have any great influence on the decomposition of at an earlier stage than was encountered with the untreated
PBT. PET.
The model ester n-propyl pyrene-1-carboxylate (II) was Furthermore, the TRIS material at 9008C gave higher
decomposed under pyrolysis conditions reflective of the fire concentrations of highly volatile products, especially
situation to give propylene and pyrene-1-carboxylic acid as methane whose concentration was almost a factor of 10
the major reaction products, presumably via a six-centered larger than obtained with the untreated material.15 The
intramolecular transition state.14 While some n-propanol was formation of radicals helps to explain such minor products of
formed, the contribution of this reaction to the overall pyrolysis as ethane, propylene, acetone, ethanol, toluene,
decomposition was insignificant. The addition of decabro- benzaldehyde, methyl benzoate, p-methylacetophenone,
modiphenyl oxide to the model ester appeared to have no vinyl p-methylbenzoate, biphenyl and diphenylmethane.
significant effect on the solid phase pyrolysis process. Thus, it The 900F polyester system also produces higher concentra-
was concluded that for bromine-containing compounds of tions of the more volatile species than obtained with the
this type to have an influence on reducing the flammability standard polyester.
characteristics of polyesters, they must influence the gas Granzow et al.27 studied the thermal degradation of PET in
phase reactions. the presence of red phosphorus. They assumed that pyrolysis
of PET occurs in two stages: the first step involves the primary
bond breaking of the polymer chain, while the second step
represents the further decomposition of the primary degra-
dation products into volatile species. The close agreement of
the kinetic parameters with and without phosphorus
indicated no inhibition of the initial chain cleavage. There-
fore, it was assumed that the inhibition by phosphorus is
exclusively due to the decrease in the velocity of the second
step. The effect of phosphorus on the degradation of low
molecular products is possibly due to the formation of
Nagasawa et al.25 studied the decomposition gases from a thermally stable phosphorus-oxygen bonds in the pyrolyzing
fire retardant PBT, which was prepared by combining PBT PET, thus preventing the formation of gaseous pyrolysis
mechanically with brominated polycarbonate and antimony products. This explanation is supported by the fact that
oxide. Using a fast thermolysis/FT-IR spectroscopy techni- elemental phosphorus has been shown to be an inefficient
que gaseous antimony tribromide was not observed. The flame retardant for oxygen-free polymers.
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700
698 S. V. Levchik and E. D. Weil
Phosphorus containing flame retardants used in a study by PBT were found in the presence of ammonium polypho-
Bednas et al.15 were Antiblaze 19 (cyclic phosphonate, AB19, sphate.30
Albright & Wilson, now Albemarle Amgard CU) and Trevira Balabanovich and Engelmann31 flame retarded PBT by
271 (reacted-in phosphinate, Hoecsht). When considering the addition of poly(sulfonyldiphenylene phenylphosphonate).
role of the flame-retardant systems on the pyrolysis data Using infrared, it was shown that the polyphosphonate
obtained at 9008C, it was obvious that both phosphorus- changes the degradation pathways of PBT resulting in
containing polyesters behave very similarly to the standard formation of polyarylates (Scheme 20). It was believed that
untreated polyester in terms of the products formed and their polyarylates are formed due to recombination of carbox-
concentrations. In another study28 it was found that yphenylene and phenylene radicals appearing from thermo-
phosphorus- and phosphorus-halogen-containing additives lysis of PBT.
increased the amount of acetaldehyde formed. Significant The presence of phenolic species and associated
decomposition of CH3CHO occurring at 7008C resulted in aromatic ketones was explained by Fries type isomerization
increased amounts of CH4 and CO formed. Phosphorus was (Scheme 21). In the presence of the phosphonate the
considered as promoting larger quantities of gaseous phenolic groups react with the P–O–C bond by a trans-
products compared to those produced by plain PET. esterification mechanism, which is proposed to be a part of
However, it was also shown that phosphorus promotes char the flame retardant effect.
formation and increases the aromaticity of the residue.
Combustion and thermal decomposition of PBT was Kinetics of thermal decomposition
studied in the presence of cyclic phosphonate (Antiblaze Since the mechanism of thermal decomposition of polyesters
1045, a higher molecular weight relative of Antiblaze 19) and is complex, this leads to complexity of kinetics, which depend
melamine.29 Infrared provided evidences that the cyclic on the detection method, heating regimes and atmosphere.
phosphonate with melamine modifies the mechanism of The presence of flame retardant additives affect the kinetics
thermal decomposition of PBT towards formation of amides of the thermal decomposition as well as the composition of
instead of acids (Scheme 19). Dehydration of aromatic the products.
amides, which cools the flame, and fast formation of For example, Lum17 postulated a multistage decomposi-
benzonitriles contributed to the flame retardancy. Similar tion mechanism for PBT involving two major reaction
alterations in the mechanism of thermal decomposition of pathways: initial degradation by an ionic decomposition
Scheme 19.
Scheme 20.
Scheme 21.
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700
Thermal decomposition and combustion of thermoplastic polyesters 699
with evolution of THF and then concerted ester pyrolysis with the conversion. The activation energy of the second
involving an intermediate cyclic transition state and yielding stage was also slightly increasing with conversion, and only
1,3-butadiene. Activation energies of formation for the main the third stage showed a constant activation energy. The
pyrolysis products were determined from the dynamic phosphorus-containing additives did not greatly affect the
measurements of the major ion species in mass spectrometry values of the activation energies, whereas bromine-contain-
and indicated values of 117 kJ/mol for the production of THF ing Dacron 900F increased and phosphorus-bromine contain-
and 208 kJ/mol for the production of butadiene. ing TRIS decreased the activation energy.
Buxbaum3 surveyed early publications on the kinetics of Based on the TGA data in inert atmosphere, Granzow
thermal degradation of PET, the rates being measured in et al.27 measured the activation energy of 155 kJ/mol for plain
terms of products formed, rate of change of molecular weight PET and of 163 kJ/mol for PET with 2 wt% red phosphorus. It
(as measured by the melt viscosity or the intrinsic viscosity) is evident that the activation energy for the decomposition of
or rate of change in concentration of end groups. Activation PET is only slightly increased in the presence of phosphorus.
energy calculated based on intrinsic viscosity change was However, in air, an activation energy of 126 kJ/mol was
261 kJ/mol, whereas activation energies calculated from found for pure PET and 151 kJ/mol for PETþ2% phosphorus
change of melt viscosity were 134 and 201 kJ/mol (two was obtained. Contrary to the results obtained in an inert
different sources). The activation energy found from the atmosphere, the activation energy for thermal oxidation
kinetics of total gas evolution was 160 kJ/mol. The kinetics strongly increased in the presence of phosphorus. The
based on increase of carboxyl end groups revealed an Arrhenius plots for decomposition in an inert atmosphere
activation energy of 175 kJ/mol and based on total end and in air intersect at temperatures of 4708C, indicating that
groups, 246 kJ/mol. above this temperature, the thermal oxidation is accelerated
Later the rates of degradation of PET were measured in in the presence of phosphorus.
terms of weight loss of volatile degradation products,
decrease of inherent viscosity and increase in concentration Combustion
of carboxyl end groups.32 Thermal-oxidative degradation Although polyesters have aromatic structures in the polymer
was investigated by differential scanning calorimetry (DSC). chain they are very combustible materials. The ease of release
Calorimetric results showed that thermal-oxidative degrada- of highly combustible aliphatic gases, e.g. acetaldehyde,
tion of PET is an exothermic reaction, with an apparent ethylene, dioxane, THF, butadiene, and the low tendency to
activation energy of 117 kJ/mol. Melt temperature, melt crosslink and char are the main reasons for the high flamm-
residence time, melt environment, and drying environment ability. The rate of heat release measured at 300 sec combus-
have all been found to affect the degradation of PET. Based on tion time in cone calorimetry at a heat flux of 50 kW/m2 for
decrease of inherent viscosity, activation energies of 117 and PET was 270 kW/m2 and for PBT was 470 kW/m2.36 The oxy-
159 kJ/mol have been calculated for air-dried and vacuum- gen index of PET is 25, whereas PBT has an oxygen index of
dried samples respectively. 22.36 The logarithm of the oxygen index for PET was propor-
A kinetic study of the thermal decomposition of PET and tional to the reciprocal of the concentration of carboxylic
PBT has been performed by controlled-rate thermogravime- chain ends.37 This was related to the idea that the main step
try (CRTG) and measurement of evolved gas components, of PET thermal degradation was hydrolytic and that this was
obtained from a simultaneous TG-MS system.33 Significant carboxylic acid catalyzed.
advantages of CRTG were found for studying the thermal Stuetz et al.38 used oxygen index apparatus and graded the
decomposition of polymers, because of better separation of flammability of various polymers based on the time interval
overlapping decomposition steps. For PBT the activation during which a sample continued to burn at a given oxygen
energy for the main step of weight loss was found at the level concentration (OC) after removal of the ignition flame. This
of 160 kJ/mol, and for PET at the level of 200 kJ/mol. time interval was called the self-quenching time (SQT). The
Based on the thermal decomposition of PET under resulting data were combined in SQT versus OC plots in
isothermal conditions, it was shown34 that the degradation order to find the critical oxygen concentration (COC) at which
in nitrogen probably occurs by a random chain scission the sample does not ignite. It was shown that COC depends
mechanism, while the degradation in air is more complex. In on the sample diameter, but burning in the presence of glass
the latter case, the exothermic character of the process wick after bottom ignition begins at an OC of 15.3%,
suggested that the chain degradation reaction is overlapped independent of specimen diameter. This COC for the onset
by oxidative processes. of geometry-independent burning in the presence of a wick
The thermal oxidative decomposition of PET has been was referred to as COC-BW. The similar value for the top
studied using thermogravimetry at several heating rates ignition was referred as COC-TW. PBT gives the following
between 0.1 and 1008C/min.35 The degradation was found to values COC-BW ¼ 15.4, COC-TW ¼ 17.5 and PET COC-
be a complex process composed of at least three overlapping BW ¼ 15.8, COC-TW ¼ 18.8. Although PET is less flammable
stages with the apparent activation energies 100–120, 185– than PBT both polymers need to be flame retarded, especially
200, and 140–145 kJ/mol. Phosphorus-containing flame for electrical and electronic applications.
retardant also showed these three stages, although the very Day et al.39 studied the flammability limits of pyrolysate
early stage had a lower activation energy and the second gases generated from polyester fabric and polyester/cotton
stage was displaced to higher conversions.26 The halogen- blend using an infrared heating system with 13.27 kW/m2
containing additives did not show the uniform first stage but heat flux. Every 30 sec a spark was discharged to test for
instead, the apparent activation energy increases steadily ignitability. The polyester spun woven fabric showed the
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700
700 S. V. Levchik and E. D. Weil
following results: heating time to ignition ¼ 9 min, flamm- aromatic hydrocarbons. Because of high flammability hazard
ability limits (weight loss %) lower ¼ 10, higher ¼ 42, flamm- both PET and PBT need to be flame retarded in some textile as
ability range (weight loss %) ¼ 31. Polyester/cotton blend well as electrical and electronic applications.
50/50 showed the following results: heating time to
ignition ¼ 4.75 min, flammability limits (weight loss %) low-
er ¼ 47, higher ¼ 73, flammability range (weight loss %) ¼ 26.
REFERENCES
It is seen that PET fibers have a very small lower flammability
limit, indicating ease of ignition since only a relatively small 1. Jadhav JY, Kantor SW. Encyclopedia of Polymer Science and
Engineering, vol. 12. Wiley: New York, 1988; 217–256.
weight loss of 10% will give a flammable pyrolysate-air 2. Kinoshita R, Teramoto Y, Yoshida H. J. Thermal. Anal. 1993;
mixture. However, in comparison to the cellulosic blended 40: 605.
material a larger amount of heat energy has to be fed into the 3. Buxbaum LH. Angew. Chem. Int. Ed. 1968; 7: 182.
4. Macooll A. J. Chem. Soc. 1958; 3398.
system (i.e. longer heating period) before ignition occurs. The 5. Lorente J, Ellis G, Marco C, Gomez MA, Fatou JG. Eur.
observed high lower flammability limit for polyester/cotton Polym. J. 1994; 30: 621.
blend 47% is totally unexpected, especially in the view of 6. Pohl HA. J. Am. Chem. Soc. 1951; 73: 5660.
7. Adams R. J. Polym. Sci., Polym. Chem. 1982; 20: 119.
studies made with blended cotton/polyester fabrics which 8. Luderwald I, Urrutia H. Makromol. Chem. 1976; 177: 2079.
have revealed these blends being more flammable than the 9. Edge M, Wiles R, Allen NS, McDonald WA, Mortlock SV.
individual components. Some of the mystery was solved by Polym. Degrad. Stab. 1996; 53: 141.
10. Vijayakumar CT, Ponnusamy E, Balakrishnan T,
Miller et al.40 who showed that the blend gave off much more Kothandaraman H. J. Polym. Sci., Polym. Chem. 1982; 20: 2715.
combustible gases, notably ethylene and acetylene, which the 11. Khemani KC. Polym. Prepr. 1999; 40(1): 625.
individual polymers did not. 12. Hovenkamp SG, Munting JP. J. Polym. Sci. Part A-1 1970; 8:
670.
Studies have shown41 that the composition of gases 13. Bounekhel B, McNeill IC. Polym. Degrad. Stab. 1995; 49: 347.
emitted during PET combustion depends strongly on the 14. Day M, Ho K, Wiles DM. J. Appl. Polym. Sci. 1980; 25: 2943.
degradation temperature. At lower temperatures (200– 15. Bednas ME, Day M, Ho K, Sander R, Wiles DM. J. Appl.
Polym. Sci. 1981; 26: 277.
3008C), the emission of smoke was not observed. During 16. Sugimura Y, Tsuge S. J. Chromatographic Sci. 1979; 7: 269.
PET destruction at the temperature range of 400–7008C, large 17. Lum RM. J. Polym. Sci., Polym. Chem. 1979; 17: 203.
amounts of less-volatile compounds were emitted, forming a 18. Dzieciol M, Trzeszczynski J. J. Appl. Polym. Sci. 1998; 69:
2377.
pale yellow smoke. The analyses have shown that the emitted 19. Buxbaum LH. Polym. Prepr. 1967; 8: 552.
smoke is a complex mixture of aromatic compounds: 20. Yoda K, Tsuboi A, Wada M, Yamadera R. J. Appl. Polym. Sci.
terephthalic acid and its esters, benzoic acid, p-substituted 1970; 14: 2357.
21. Nearly DI. J. Polym. Sci., Part A-1 1971; 9: 2063.
benzoic acid, low oligomers of PET, and acetylated and 22. Spanniger PA. J. Polym. Sci., Polym. Chem. 1974; 12: 709.
polycyclic aromatic hydrocarbons. 23. Sato H, Kondo K, Tsuge S, Ohtani H, Sato N. Polym. Degrad.
Stab. 1998; 62: 41.
24. Pellow-Jarman M, Hetem M. Polym. Degrad. Stab. 1995; 47:
CONCLUSIONS 413.
25. Nagasawa Y, Yamamoto M, Ozawa K. Thermochim. Acta
Thermal decomposition of polyesters usually starts with the 1995; 267: 223.
26. Cooney JD, Day M, Wiles DM. J. Appl. Polym. Sci. 1984; 29:
scission of the polymer chain through a six-membered ring 911.
transition stage. It is believed that this scission is mostly het- 27. Granzow A, Firrillo RG, Wilson A. J. Appl. Polym. Sci. 1977;
erolytic and not a free-radical process, however this is still 21: 1687.
28. Avondo G, Vovelle C, Delbourgo R. Combust. Flame 1978;
under discussion in the literature. Secondary reactions 31: 7.
mostly involve the formation of vinyl ester chain ends which 29. Balabanovich AI, Levchik GF, Levchik SV, Engelmann J.
lead to polyene structures, crosslinking, but also evolution of J. Fire Sci. 2002; 20: 71.
30. Balabanovich AI. J. Fire Sci. 2003; 21: 285.
light volatile products. In air polyesters undergo crosslink- 31. Balabanovich AI, Engelmann J. Polym. Degrad. Stab. 2003;
ing, a relatively minor process in an inert atmosphere. Flame 79: 85.
retardants have been often reported to modify the mechan- 32. Jabarin SA, Lofgren EA. Polym. Eng. Sci. 1984; 24: 1056.
33. Arii T, Ichihara S, Nakagawa H, Fujii N. Thermochim. Acta
ism of thermal decomposition of polyesters. Phosphorus- 1998; 319: 139.
containing products, surprisingly, often increased the 34. Birladeanu C, Vasile C, Schneider I-A. Makromol. Chem.
concentration of combustible light gases, while promoting 1976; 177: 212.
35. Coney JD, Day M, Wiles DM. J. Appl. Polym. Sci. 1983; 28:
charring. Kinetic analysis has shown that the thermal decom- 2887.
position of polyesters is a complex process with activation 36. Joseph P, Ebdon JR. In Fire Retardant Materials, Horrocks
energies varying from 100 to 250 kJ/mol. AR, Price D (eds). CRC Press: Boca Raton, 2001; 220–263.
37. Pearce EM. In Contemporary Topics in Polymer Science, vol. V,
Although thermoplastic polyesters contain aromatic struc- Vanderburg EJ (ed). Plenum Press: New York, 1984;
tures in their chain, they are very combustible. If ignited from 401–413.
the bottom both PET and PBT can burn even in an oxygen lean 38. Stuetz DE, DiEdwardo AH, Zitomer F, Barnes BP. J. Polym.
Sci., Polym. Chem. 1980; 18: 987.
atmosphere (15%). PET fibers are very easily ignitable, 39. Day M, Suprunchuk T, Wiles DM. J. Appl. Polym. Sci. 1983;
especially when they are blended with cotton fibers. Upon 28: 449.
combustion or high temperature decomposition PET gen- 40. Miller B, Martin JR, Meiser CH Jr, Gargiullo M. Text. Res.
J. 1976; 46: 530.
erates some smoke composed of terephthalic and benzoic 41. Dziecinol M, Trzeszczynski J. J. Appl. Polym. Sci. 2001; 81:
acid derivatives, low oligomers of PET, and polycyclic 3084.
Copyright # 2004 John Wiley & Sons, Ltd. Polym. Adv. Technol. 2004; 15: 691–700