504

POLYELECTROLYTES

Vol. 7

POLYESTERS, THERMOPLASTIC

Introduction

Polyesters enter our lives in a most ubiquitous manner as textiles, carpets, tire
cords, medical accessories, seat belts, automotive and electronic items, photo-
graphic film, magnetic tape for audio and video recording, packaging materials,
bottles, and so on. Their utility is illustrated by the vast range of their applications.
This article describes the properties, synthesis, manufacture, and raw materials
for the two most widely used thermoplastic polyesters: poly(ethylene terephtha-
late) (PET) [25038-59-9] and poly(butylene terephthalate) (PBT) [26062-94-2]. In
order of volume, PET comes first by virtue of its enormous market tonnage in
polyester fibers and films, as well as the resin for blow-molded bottles, containers,
and food packaging.

Polyesters are linear polymeric molecules containing in-chain ester groups,

formally derived by condensation of a diacid with a diol. The earliest commercial
polyesters were the alkyd resins (qv), nonlinear polymers developed for surface
coatings shortly after World War I. It was not until the 1930s that the classic
studies of Carothers examined main-chain polyesters in a rigorous and sys-
tematic fashion (1). Although he studied a wide range of polyesters, made both
from aliphatic diacids and diols (AA-BB) type and from

ω-hydroxyacids (ABAB

type), for various reasons, Carothers did not pursue polyesters derived from
aromatic diacids and alkylene diols. All his aliphatic polymers were low melting
(

<100

◦

C) and were easily soluble in common organic (dry cleaning) solvents.

They had little utility as textile fibers, which largely explains why Carothers
turned his attention to polyamides. The first successful synthesis of satis-
factory high molecular weight poly(ethylene terephthalate), 2GT, was made in
England in 1942, during the early days of World War II. The inventors were J. Rex

Encyclopedia of Polymer Science and Technology. Copyright John Wiley & Sons, Inc. All rights reserved.

Vol. 7

POLYESTERS, THERMOPLASTIC

505

Whinfield and W. Dickson, working at the Calico Printers’ Association (2,3). Other
polymers pioneered by these workers included poly(1,3-propylene terephthalate),
3GT, poly(1,4-butylene terephthalate), 4GT, and the polyester from ethylene gly-
col and 1,2-bis(4-carboxyphenoxy)ethane, known as CPE-2G or “Fiber-O” (4). Of
these materials, PET was selected for development as a melt-spinnable syn-
thetic fiber, but commercialization was impossible until after the end of World
War II. Eventually, when the various national economies were back on a peace-
time footing, PET polymer and fibers derived from it were put into production.
The whole market-driving force for polyester at this time was in the form of
synthetic fibers. In the United Kingdom, the new material was manufactured
by Imperial Chemical Industries Ltd. under the trade name Terylene, while
DuPont introduced it to the United States in 1953 as Dacron (see P

OLYESTERS

,

F

IBERS

).

In the early 1950s, additional poly(alkylene terephthalate)s were exam-

ined in both United States and Europe for synthetic fiber applications. One
of these, poly(1,4-butylene terephthalate), 4GT or PBT, had (and still has)
some very attractive fiber properties. Notably, it was white and resisted photo-
oxidative yellowing much better than nylon. It accepted disperse dyes much
more readily than PET. It had excellent resilience and elastic recovery. All
these features made it attractive for such fiber end uses as women’s wear,
hosiery, and carpet fiber. However, the twin criteria of pleat-retention and
crease-resistance in apparel fabrics were commercially important at the time
(it was an era when accordion pleats were highly fashionable in women’s wear)
and PET was superior on these counts. Only later was it realized that PBT
polymer was highly suited to injection-molded applications. Not only was it
highly crystalline, it had a high rate of crystallization, so that parts could
be molded in a fully crystallized state and therefore free from distortion or
warping. Unlike nylon, it had a low moisture uptake and was much more di-
mensionally stable to changes in atmospheric humidity. Molding-grade PBT
resins were introduced in 1970 by Celanese Corp. as Celanex and others fol-
lowed.

In contrast, PET was not successful as a molding resin because of its low

rate of crystallization in a cold mold. The mold has to be heated to 130–140

◦

C,

well above the PET glass–rubber transition temperature (T

g

) to obtain adequate

crystallization rates. Satisfactory moldings were still not obtained because of un-
controllable crystal morphology. During the late 1960s, fast-crystallizing grades
of PET were developed which had uniform morphology because of the presence
of specific nucleating agents (5–7). Akzo and DuPont were early entrants in
the field with their Arnite and Rynite range of PET resins; other manufactur-
ers followed and PET molding resins are now widely used, particularly in the
auto industry. However, the biggest single market for moldable PET today is
in blow-molded bottles, which exceeds every other single end use for PET poly-
mer except polyester fibers. In 1990, the annual world production of PET fibers
(8) was about 9 million tons; according to Table 2 (see later) the annual pro-
duction of PET bottle resin in 2000 was 1.4 million ton and growing rapidly. It
is worth noting that poly(ethylene-2,6-naphthalenedicarboxylate) (PEN) was an-
other important polyester also synthesized first by ICI workers in 1948 (9) (see
P

OLY

(

ETHYLENE NAPHTHANOATE

).

506

POLYESTERS, THERMOPLASTIC

Vol. 7

Manufacture of Raw Materials and Monomers

PET and PBT are both made from terephthalic acid and its dimethyl ester. Tereph-
thalic acid (TA) is made by air-oxidation of p-xylene [106-42-3] in acetic acid under
moderate pressure in the presence of catalysts such as divalent cobalt and man-
ganese bromides (10). p-Xylene is the highest melting of three isomeric dimethyl-
benzenes and is separated by fractional crystallization from the C

8

-aromatic frac-

tion (including ethylbenzene) during petroleum refining (11). Alternatively it may
be separated by selective adsorption on a zeolite bed combined with an isomeriza-
tion process (12). For PET fiber production, very pure TA is required and in the
early days, when the oxidation of p-xylene was achieved with 50% aqueous nitric
acid under pressure, the process left some highly undesirable by-products such as
nitroaromatics and carbazoles in the crude TA. Because of the great insolubility
of TA, it was not easily purified by recrystallization and so it was converted to its
dimethyl ester (dimethyl terephthalate, DMT), and the DMT in turn purified by
redistillation under reduced pressure and a final recrystallization. PET was made
by the reaction of DMT with excess ethylene glycol in the presence of catalysts
to promote ester-interchange and polymerization. When the direct air-oxidation
process of p-xylene became the process of choice, the DMT purification route was
still used as an interim process, even though major undesirable impurities had
been eliminated by the new oxidation route.

During the 1967–1972 period, pure TA became available in large quanti-

ties (13) because of the improvements in the process and purification of TA.
Harmful incomplete oxidation products in crude TA such as p-toluic acid and
4-carboxybenzaldehyde (4-CBA) were eliminated by recrystallization from water
under pressure with concomitant hydrogenation. Under these conditions, p-toluic
acid is highly water-soluble and 4-CBA is hydrogenated to toluic acid (14). The
availability of pure TA caused a major change in production methods for PET.
Direct esterification (DE) processes using pure TA superseded the former ester-
interchange (EI) process based on DMT. Interestingly, in recent years DMT has
made something of a comeback, ostensibly because of the recycling of waste PET.
Production of pure DMT derived from the methanolysis and glycolysis of PET
waste is carried out on a significant scale (15,16). Manufacture of TA from p-xylene
has undergone a constant process of improvement over the years and innumer-
able patents have been filed on small improvements, new oxidation catalysts, etc.
A very recent article describes an experimental “Green Chemistry” p-xylene oxi-
dation process using hydrogen peroxide and a manganese dibromide catalyst in
supercritical water as a solvent. Under these conditions, the TA remains in solu-
tion (17).

Turning now to the diol components, ethylene glycol (ethane-1,2-diol) is made

from ethylene in essence by direct air-oxidation to ethylene oxide and ring open-
ing with water to give the 1,2-diol (18). Butane-1,4-diol is still made by the old
Reppe process. Acetylene reacts with formaldehyde in the presence of a catalyst
to give 2-butyne-1,4-diol, which is hydrogenated to butanediol. The ethynylation
step depends on a special cuprous acetylide-bismuth salt catalyst, which mini-
mizes side reactions (19). The hydrogenation step is best done in two stages over
special catalysts (20). Another butanediol route starts from butadiene. In a pro-
cess due to Mitsubishi Chemical Industries, butadiene reacts with acetic acid and

Vol. 7

POLYESTERS, THERMOPLASTIC

507

oxygen in the presence of a palladium catalyst to give 1,4-diacetoxy-2-butene (21),
and the latter is hydrogenated over a special catalyst (22) and finally hydrolyzed
to 1,4-butanediol. It is not known whether this process is in commercial use at
the time of writing. Although Reppe plants still operate, the process has been re-
placed in newer plants by a route based on hydroformylation of allyl alcohol (from
propylene oxide) over a rhodium catalyst (23) to give 4-hydroxybutyraldehyde.
This is reduced to butanediol using a two-stage hydrogenation route (24) to
minimize side reactions. The ethynylation route is still an important one but
new plants use the Arco hydroformylation process, which was purchased by
Lyondell.

Polymerization Processes

Thermoplastic polyesters are step-growth polymers which need to be made to
high molecular weight (12,000–50,000) to be useful (25). The first stage is an
esterification or ester-exchange stage where the diacid or its dimethyl ester re-
acts with the appropriate diol to give the bis(hydroxyalkyl)ester and some linear
oligomers. Water or methanol is evolved at this stage and is removed by frac-
tional distillation, often under reduced pressure at the conclusion of the cycle. For
EI, weakly basic metallic salt catalysts are used: the list is extremely long, and
many recipes are proprietary, but such salts as calcium, zinc and manganese ac-
etates, tin compounds, and titanium alkoxides have been widely used (26–28).
Certain EI-catalysts have the undesirable effect of promoting thermal degra-
dation at high temperatures (29) encountered during the latter stages of high
polymerization. A McClafferty rearrangement of the ester unit is often deemed
responsible. To overcome this, EI-catalysts are sequestered at the end of ester-
interchange by adding phosphorus compounds such as triphenyl phosphite, triph-
enyl phosphate, or polyphosphoric acid in stoichiometric amounts (30). Again,
such recipes are often proprietary. Titanium and tin compounds [eg Ti(OR)

4

or

R

2

SnO] act as powerful universal catalysts for both EI and polymerization re-

actions and are left unchanged. They are not suitable for 2G-derived polyesters
because of formation of a yellow color, although recipes are under development
in Japan which claim to obviate this problem. For PBT, ester-exchange using
DMT and a titanium alkoxide catalyst is the route of choice, since butanediol
cyclizes to tetrahydrofuran (THF) in the presence of acids (31). Nevertheless,
the direct polycondensation of butanediol with terephthalic acid using special
reaction catalysts and conditions to minimize THF formation has been described
(32–35).

Batch polymerization is usually done in an autoclave fitted with a powerful

mechanical stirrer to handle the viscous melt, under high vacuum at a temper-
ature above the melting point of the final polymer. During this critical stage, it
is very important to eliminate oxygen and blanket the process with inert gas,
nitrogen, or argon. During the polycondensation stage, the linear oligomers and
the bis-hydroxyalkylterephthalate esters undergo a succession of ester-exchange
reactions, eliminating the diol, which is removed under high vacuum, and thus
increasing the molecular weight steadily. A polymerization catalyst is needed,
whether the initial process is EI-based or DE-based. As already mentioned, tin

508

POLYESTERS, THERMOPLASTIC

Vol. 7

and titanium catalysts are suitable for both ester-interchange and polymeriza-
tion, but for PET, antimony trioxide is the usual polymerization catalyst (36). It
only becomes active at high temperatures and thus can be added at the start of
the reaction along with the other catalysts. More recently, there has been a move
away from heavy metals like antimony, particularly in Europe, where they are
viewed with increasing disfavor on environmental grounds. Even in the United
States, problems can arise with heavy metal contaminants (which include anti-
mony) in waste glycolysis still-bottoms. These cannot be landfilled for environmen-
tal reasons and their safe disposal causes added expense. A less toxic metal would
clearly be advantageous. However, alternatives are not universally satisfactory.
Titanium alkoxides cause unacceptable yellowing of PET, apparently because of
reaction with vinyl ester chain ends. This cannot occur with PBT. Germanium
compounds, either as the dioxide, tetraalkoxide, or glycoloxide, are good catalysts,
nontoxic, and give very white polymers. However, they are usually considered to
be too expensive because of the relative scarcity of germanium. Another problem
is loss of germanium by volatilization. The use of a germanium/titanium mixed
catalyst is disclosed in a patent relating to PET bottles (37). Antimony trioxide
is a robust polymerization catalyst, but in PET it is susceptible to reduction to
antimony metal, which can cause an undesirable gray-blue color in the polymer.

As the polymer molecular weight increases, so does the melt viscosity, and the

power input to the stirrer drive is monitored by a wattmeter so that an end point
can be determined for each batch. When the desired melt viscosity is reached,
the molten polymer is discharged through a bottom valve, often under positive
pressure of the blanketing gas, and extruded as a ribbon or as strands, which are
water-quenched and chopped continuously by a set of mechanical knives. Large
amounts of PET are also made by continuous polymerization processes. PBT is
made by both batch and continuous polymerization processes (38–40).

The polymer is then dried thoroughly and stored for subsequent process-

ing. Whenever a polyester is made by melt polycondensation, a small amount of
cyclic oligomer is formed which is in equilibrium with the polymer. This can be
extracted with solvents from solid polymer but when the extracted polymer is
remelted, more oligomer forms until the equilibrium is reestablished. The level of
such oligomers is about 1.4–1.8% by weight for both PET and PBT. Thus it is im-
possible to completely remove cyclic oligomers from any melt-processed polyesters.
In the case of PET, the main oligomer is a cyclic trimer (41,42) while in the case of
PBT, the oligomers comprise roughly an equal mixture of cyclic dimer and trimer,
together with much smaller amounts of higher oligomers (43). The presence of
such oligomers usually does little harm, but under certain conditions they can ex-
ude to the surface. This is a problem with polyester fibers with their high surface/
volume ratio. For example cyclic trimer can interfere with the fiber dyeing process.
It is also a nuisance in bottle production as the trimer coats the preform injection
molds.

It is often necessary to make polymer of much higher molecular weight than

would be practicable in the melt, either by reason of excessively high melt viscosity
or because degradation reactions would begin to overtake the rate of polymeriza-
tion and limit molecular weight. For this reason, solid-state polymerization (SSP)
is frequently used. Dried polymer chip of moderate molecular weight is heated at
a temperature roughly 20

◦

C below its softening point, either in a high vacuum or

Vol. 7

POLYESTERS, THERMOPLASTIC

509

in a stream of hot inert gas in a device which agitates the solid. Typical devices
might be a twin-cone rotary vacuum drier or a fluidized-bed unit. There are many
types of commercial SSP units available.

One very important practical consideration with PET is that the polymer

chips must be fully crystallized by careful annealing before the solid-phase poly-
merization process begins. Usually the polymer has been water-quenched before
cutting and will be largely amorphous. If not precrystallized, the chips may sin-
ter together on attempted solid-phase polymerization. The difficulties caused by
several tons of polymer setting to a solid mass can well be imagined! It is possi-
ble by careful annealing to raise the T

m

of the PET considerably above the usual

figure (44), thus allowing the solid-phase polymerization to take place at higher
temperatures and shorter reaction times. Various agitation devices (45) and poly-
mer chip treatments (46) have been described to prevent sticking. An integrated
crystallizing and solid-state polymerization process has been described (47) in the
patent literature.

In the solid-state process, the volatile by-products of the polycondensation

reaction (traces of water, methanol, excess diol, etc) escape by vapor diffusion
through the solid chip and are rapidly removed from the chip surface instead of
being limited by viscous diffusion through a bulk melt. Esters of aliphatic diols
and aromatic diacids begin to decompose thermally even in an inert atmosphere
at around 250

◦

C (48). In the polymer melt, local mechanical heating caused by

viscous shear during agitation can cause thermal degradation. SSP eliminates
this and the molecular weight rises within a few hours to the desired figure.

During the polymerization reaction, various by-products are formed. In the

case of PBT, the major one is THF formed by dehydration of butanediol or by
internal cyclization of the C

4

-ester units. This is a harmless by-product, as far

as the polymerization is concerned, since it is nonreactive under polymerization
conditions and quickly volatilizes away. However, its formation is commercially
undesirable because it constitutes an air pollution problem as a volatile organic
emission, and represents the loss of valuable starting material (butanediol). PET,
by contrast, has two major troublesome by-products. One is the generation of di-
ethylene glycol (DEG) units in the chain by dehydration of 2-hydroxyethyl ester
chain ends to form an ether link. This process cannot be entirely prevented, al-
though addition of alkalis can reduce it, and it is further mitigated by restricting
those time/temperature combinations that tend to favor DEG formation. DEG
content is related to the change in softening point (

T

m

) by the empirical relation

(49)

T

m

= ( − 2.2)m(

◦

C)

where m is the molar % concentration of DEG.

In addition to depressing melting point, DEG units adversely affect the crys-

tallinity of the polymer, reduce the strength of both fibers and oriented films,
and increase the susceptibility of the polymer to chemical attack and aqueous
hydrolysis.

The other major by-product is acetaldehyde, which is produced by thermal

degradation of the PET unit. Random oxygen–alkyl scission of ester units leaves a
vinyl ester end and a carboxyl-ended chain. The vinyl ester reacts with a polymer

510

POLYESTERS, THERMOPLASTIC

Vol. 7

end-group to form a new polymer link and expels acetaldehyde, the tautomer
of vinyl alcohol (50). The vinyl ester end can also thermally polymerize to give
chain-branched and cross-linked products and gel-particles, and further thermal
degradation of these polyvinyl units gives rise to colored polyenes (51,52). Al-
though acetaldehyde is highly volatile, its presence is particularly objectionable
in PET bottle resin used for soda bottles. Its presence may not exceed 3 ppm in
the final container if used for potable substances, as it imparts an off-taste to
popular cola drinks (53). Every time PET is melted during its processing, more
acetaldehyde is generated. One reason for bottle resin undergoing a final solid-
phase polymerization before stretch blow molding is to remove the last traces of
acetaldehyde.

Physical Properties of PET

The full crystal structure of PET has been established by x-ray diffraction (54–57).
It forms triclinic crystals with one polymer chain per unit cell. The original cell pa-
rameters were established in 1954 (54) and numerous groups have reexamined it
over the years. One difficulty is determining when crystallinity is fully developed.
One researcher (57) annealed PET at up to 290

◦

C for 2 years.

Cell parameters are a

= 0.444 nm; b = 0.591 nm; c = 1.067 nm; α = 100

◦

;

β = 117

◦

;

γ = 112

◦

; density

= 1.52 g/cm

3

Thermochemical data depends on the degree of crystallinity in the polymer,

and a very highly annealed polymer sample can have T

m

= 280

◦

C, much higher

than the usual value of 260–265

◦

C (58). The heat of fusion (59) is about 140 J/g,

(33.5 cal/g). The glass–rubber transition temperature (T

g

) depends on both the

method of measurement and the state of the polymer. A solid chip sample as
measured by differential scanning calorimetry (dsc) gives a value around 78

◦

C

(60) but a highly oriented and crystalline-drawn fiber measured by the dynamic
loss method will give values as high as 120

◦

C (61). The specific gravity of undrawn

amorphous PET is 1.33, whereas crystalline-drawn fiber has a value of 1.39 (62).

As a step-growth polymer made under equilibration conditions, PET has

a molecular weight distribution very close to the theoretical value of 2.0. The
Mark–Houwink equation relates the intrinsic solution viscosity [

η] to the molec-

ular weight as [

η] = KM

v

α

, where M

v

is the viscosity average molecular weight

and K and

α are the Mark–Houwink coefficients determined experimentally for

individual solvents. The usual solvents historically used for PET are 60/40 w/w
phenol/tetrachlorethane (P/TCE) and 2-chlorophenol (OCP). Neither solvent sys-
tem is entirely satisfactory and better results have been obtained using either
hexafluoroisopropyl alcohol (HFIP) or a 50/50 v/v mixture of HFIP and pentafluo-
rophenol (PFP) at 25

◦

C (63). Although expensive, these acidic fluorinated solvents

readily dissolve even highly crystalline samples of PET at moderate temperatures,
thus avoiding the degradation problems commonly encountered with the older sol-
vents at high temperatures. Thus the results are more likely to be reliable, since
degradation under normal conditions is minimal. The Mark–Houwink coefficients
for PET for a range of solvent systems are shown in Table 1.

PET is a strong, stiff polymer. High tenacity tire-cord fibers with a tensile

strength (tenacity) of up to 0.9 N/tex (10 g/den) and a tensile modulus of 11 N/tex

Vol. 7

POLYESTERS, THERMOPLASTIC

511

Table 1. Mark–Houwink Coefficients for PET
in Various Solvents at 25

◦

C

Solvent

K, 10

− 4

dL/g

α

OCP

6.31

0.658

P/TCE

7.44

0.648

HFIP

5.20

0.723

PFP/HFIP

4.50

0.705

(125 g/den) are readily obtained. Normal (ie apparel) textile fibers usually have
tenacity around 0.5 N/tex (5.6 g/den) with a modulus of 7 N/tex (80 g/den). The
tensile modulus of PET is significantly higher than that of either PBT (4GT) or
PPT (3GT) polyester fibers.

Applications of PET

The applications of PET are almost too numerous to mention and new uses appear
constantly. Historically, PET was first commercialized in the early 1950s as a
fiber-forming polymer. Staple (noncontinuous) fiber for blending with wool (qv)
and cotton (qv) for increased fabric performance was the first product and as
such was highly successful. Later, continuous filament yarn, frequently bulked
mechanically, for knitwear and especially for women’s apparel was introduced. The
high T

g

and low moisture regain (0.4%) gave excellent wash-wear fabric properties

with significantly reduced tendencies to wrinkle. Industrial high strength fibers
were then developed, notably tire cords for automobile tires. These were usually
made from high intrinsic viscosity (IV) polymer (ca 0.75) and spun as continuous
filament yarn. When automobile seat belts became the norm, they were also made
from polyester because, unlike nylon, it had a high modulus and low tendency to
stretch. The whole topic of polyester fibers is a very large one and for a more
comprehensive treatment, see Ref. 64 and Polyesters, Fibers.

Another market where PET proved of great value was in the form of polyester

films, which were made by extrusion through a slit die, followed by a biaxial
drawing process. They are used for photographic film backing and for magnetic
audiotape and videotape substrates. Another market is drafting film. In all of
these outlets the high strength and high modulus of PET, particularly for biaxially
drawn films, give the desired dimensional stability and low stretch. For further
details of film fabrication and processing, see Ref. 65 and films, manufacture; films,
orientation.

One of the largest uses of PET resin and certainly the most dramatic in

growth during the last 20 years is the stretch blow-molded PET bottle; annual
consumption runs into billions of units in United States alone with corresponding
rapid expansion on a worldwide scale. The advantages of a thermoplastic bottle
are self-evident: it is light-weight, shatterproof, and recyclable. The early market
driver was weight, which reduced transportation costs compared with glass during
a severe energy shortage. Improved product safety associated with shatter-proof
plastic bottles was another factor: recycling did not become a factor until
later.

512

POLYESTERS, THERMOPLASTIC

Vol. 7

The major technical problem facing any thermoplastic bottle manufacturer

is the permeability of the plastic bottle wall to oxygen and carbon dioxide, which
affect the shelf-life of the contents. The average 2-L soda bottle maintains an
internal pressure of CO

2

of ca 520 kPa (75 psi). To stop the product going “flat,”

carbon dioxide pressure must be retained during storage before unsealing for
several weeks. Likewise, oxygen must not diffuse in through the bottle walls to
oxidize the contents, spoiling the flavor of the product. PET is semipermeable
to such gases. At one time, attempts were made to coat PET with impermeable
layers but safety, cost, and recycling difficulties defeated them, although more
recent processes appear promising.

The basic enabling invention in 1973 was the stretch blow-molded PET bot-

tle process (66,67). In this process, the polymer bottle wall is subjected to a rapid
biaxial drawing, which greatly increases the molecular orientation of the bottle
wall. Not only does this increase the mechanical strength of the bottle but it also
reduces the permeability of the walls (68). Blow-molding thermoplastic hollow ar-
ticles is a highly specialized process and for more details the reader should consult
a specialist publication (69). However, the process will be outlined so that the rea-
sons for certain polymer properties will become apparent. A typical process is the
two-stage blow-molding process. A bottle “preform” is molded by a conventional
injection-molding process with thick walls and the neck with screw-cap threads
in place. This preform is amorphous and clear. Multiple cavity dies are used to
increase productivity. In the second stage, the preforms are heated in a mold to a
carefully controlled temperature above the glass–rubber transition temperature,
typically to 90–100

◦

C. The inside of the mold cavity is the size and shape of the

finished bottle. The preform is subjected to a combined axial and radial stretching
process. A hollow metal mandrel passes into the preform and partially elongates
it in the axial direction while dry air at about 345–690 kPa (50–100 psi) blows the
walls of the softened preform outward to fill the mold, giving radial stretching.
The mold opens to allow the bottle to cool. This combined process results in both
radial and axial drawing of the bottle walls, which causes strain-induced crys-
tallization and gives a container superior strength, clarity, and no environmental
stress-cracking.

The usual 2-L bottle preforms weigh about 50 g and the final blown bottle

has a wall thickness of about 0.38 mm (0.015) in. (70). In view of the enormous
number of bottles produced annually, the processes are constantly being modified
to raise throughput rates and are all highly automated. Several large machinery
manufacturers (eg Cincinnati-Milacron, Sidel) specialize in building the complex
blow-molding equipment. Typical cycle times for the actual blow molding are 3–
6 s. This process had one serious drawback in the early days—it gave a bottle
with a hemispherical base, which is clearly not capable of standing upright. A
flat base could not be fully oriented, so gases would permeate out through it. For
several years, PET soda-bottles were fitted with separate flat-bottomed basecaps,
usually molded from high density polyethylene (HDPE) and secured with a hot-
melt adhesive. This meant extra cost because of extra material and processing
steps and interfered with recycling. The invention by the Continental Group of
the so-called petaloid base bottles with a five- or six-lobed pattern base was a
major advance: it made possible a one-piece, fully biaxially oriented bottle which
could stand up by itself (71).

Vol. 7

POLYESTERS, THERMOPLASTIC

513

PET Bottle Resin.

Stretch blow-molding is a mechanically severe opera-

tion with very high deformation rates. The consumer wants a glass-clear bottle:
any opacity caused by strain-induced crystallization is highly undesirable. PET
bottle resins are usually made to high molecular weight (M

n

= 22,000–40,000; IV

0.75–0.90 dL/g) so as to withstand the severe blow-molding operation. Such an
intrinsic viscosity is too high for melt-polymerization, and solid-stage polymer-
ization is required. The process has been reviewed (72). PET bottle resin is made
in a continuous melt-polymerization plant either by direct esterification using
TA/glycol, or by EI using purified DMT, often recovered from recycled PET bot-
tles. The process to be described is a direct esterification process. Firstly, the acid
and glycol are thoroughly mixed to a paste and catalyst and stabilizers are added.
The paste is pumped to the esterifiers where water is driven off. The molten mix-
ture of low polymer and oligomers passes through various polymerization stages:
a prepolymerizer, an intermediate polymerizer, and finally the melt arrives at the
high polymerizer or finishing stage where the final IV is about 0.65 dL/g. During
the various successive stages, the melt grows increasingly viscous and high vac-
uum is applied at the finisher to complete the reaction. The agitators used in the
polymerizers are designed to disengage volatiles without high shear-rates and to
obviate dead spots where polymer melt could stagnate and undergo degradation.
A series of DuPont patents (73–75) describes a process whereby no vacuum is
needed to build molecular weight to the desired level. By use of specially designed
mechanical agitators and an ingenious reactor design, which continually gener-
ates thin films of polymer melt, a brisk stream of inert gas (nitrogen or CO

2

) is

sufficient to cause efficient disengagement of the excess glycol and thus drive the
reaction to the desired IV. The entrained glycol can be recovered from the exit gas
stream and the inert gas recycled. Clearly, the absence of vacuum equipment re-
duces capital costs and, it is claimed, the process can be operated both batchwise
and continuously and is suitable for several commercial polyesters including PET,
PEN, and PBT.

Returning to the original process description, at the exit from the high poly-

merizer, the molten polymer is extruded into a water-bath and continuously diced
to small chips about 2.5–3.5 mm across. The chips are dried and passed to the
crystallizers where they are annealed in the solid state above T

g

, gradually ris-

ing to a temperature close to the point of maximum crystallization rate (approx.
40–160

◦

C). They are slowly agitated to prevent them from sintering together as

they move to the solid-state polymerizer. During the crystallization stage, the chip
density increases from 1.33 to about 1.37 g/cm

3

. Finally the chips pass into the

solid-phasing towers. Here they descend slowly under gravity in a plug-flow mode
through a long hot zone under a countercurrent flow of inert gas to sweep away
the volatile by-products. The speed of descent is controlled so that an optimum
time/temperature profile is maintained for the desired final IV. The scale of oper-
ation is impressive: typical melt-continuous polymerizers run at 10 tons/h, which
implies a capacity of approximately 70 million kg/year; a large manufacturer may
have several such units and each one provides enough polymer for well over one
billion 2-L soda bottles. Newer plant capacities are even more remarkable: a new
plant in South Carolina has a capacity of 182 million kg/year in a single line.

The number-average molecular weight for typical bottle resin is between

24,000 and 31,000 Da/molecule. One of the most objectionable by-products of

514

POLYESTERS, THERMOPLASTIC

Vol. 7

PET polymerization is acetaldehyde, which affects the taste of cola drinks at
concentrations as low as 60 ppb. The specification for acetaldehyde in the fi-
nal product must not exceed 3

µg of acetaldehyde per liter of headspace. The

bulk of the acetaldehyde produced in the polymerization process is removed dur-
ing the final solid-phase polymerization stage. Since blow molding is carried out
well below 200

◦

C, only minute amounts of fresh aldehyde are formed by thermal

degradation at the last stage. Originally, bottle preforms weighed 60–70 g but
this has now been reduced to about 50 g. The lighter weight bottles have thin-
ner walls so that during biaxial drawing, excessive strain-crystallization (opac-
ity) becomes a problem. Most manufacturers (76,77) have introduced copolymers
of PET containing minor amounts (2–5 mol%) of such comonomers as isoph-
thalic acid (PET-A) or cyclohexanedimethanol (PET-G) to reduce the polymer
melting point by about 4–12

◦

C with a lower tendency to crystallize. Polyesters

hydrolyze very rapidly at 280

◦

C in the melt. Rigorous polymer-drying to a chip

moisture content below 50 ppm moisture is necessary before the injection mold-
ing of bottle preforms. (PET stored under ambient conditions can easily have a
moisture content of 2000 ppm). Melt-dyed polymers using FDA-approved dyes
are used to mold preforms for colored bottles where the customer requires
them.

In recent years, more and more potables have been packed in PET bottles—

one of the most rapidly growing areas, first in Europe but now in the United States
as well, is bottled drinking water, both noncarbonated and name brand spa wa-
ters. Interestingly, for such potables, acetaldehyde must be even more stringently
excluded than for cola drinks. A major market prize would be a bottle sufficiently
impermeable to oxygen for packaging beer and wine without spoiling the flavor.
This is a much more difficult task than packaging carbonated soft drinks. Nu-
merous designs and bottle coatings have been explored but recycling must not
be impaired. Plasma-coating of PET bottles with inorganic materials such as sil-
ica and glass to bring about a fivefold reduction in oxygen permeability has been
disclosed (78,79). Sidel Corp. (80) has gained FDA approval for a process which
plasma-coats amorphous carbon on the insides of 600-mL PET beer bottles at a
rate of 10,000 bottles/h. The bottles, which are amber-colored, are said to be fully
recyclable.

Other PET Packaging Products.

In addition to blow-molded bottles, a

packaging market has appeared in recent years for other types of PET polymer.
One is amorphous PET (APET) which uses a comonomer (isophthalic acid is one)
to minimize crystallinity. Such polymers are used in clear bubble and blister packs
for a variety of pharmaceutical products and medical devices (81). Food packaging
is also a rapidly growing market for APET and clear heat-sealable thermoformed
packages are widely used. By use of blowing agents, foamed PET with a spe-
cific gravity of as little as 0.05 g/cm

3

can be made. Markets include baking trays

and the very light foams can be used in food packaging. PET bottle resin is the
feedstock and the foams can be recycled. One ingenious process uses the hydro-
carbon heptane at 249

◦

C as a PET blowing agent (82). These products are largely

spin-offs from the vast and rapidly growing bottle resin process and provide new
opportunities and markets for resin manufacturers. PET is regarded as a premium
packaging material and its recyclability is both a market and an environmental
advantage (83).

Vol. 7

POLYESTERS, THERMOPLASTIC

515

Table 2. World Nonfiber Polyester Resin Consumption Patterns
1999–2000,

a

,b

10

3

t

PET

Country or region

Bottles

Film and sheet

Molded

Total

PBT

USA and Canada

1299

136

423

c

2132

67

1389

148

454

c

2317

71

Western Europe

1211

79

d

NA

1520

116

1352

83

d

NA

1703

129

Japan

370

242

20

632

119

380

250

20

650

130

China

NA

NA

NA

NA

36

NA

NA

NA

NA

37

a

Ref. 84.

b

The values in bold are for the year 2000.

c

Compounded resin.

d

Does not include recording film.

PET Molding Resins.

It is difficult to establish accurately the world mar-

ket size for PET as a molding resin. The total tonnage is small compared with the
vast amounts used for fibers and bottle resin. However Table 2 shows estimates
for 1999–2000 consumption of nonfiber PET and PBT resin, broken down into
end uses in several major world markets. The bulk of the U.S. market for PET
molding resin is the auto industry. Its T

m

and heat distortion temperature are

higher than the corresponding values for PBT, and its low moisture uptake and
dimensional stability with respect to changes in humidity make it superior to
nylon. Both PET and PBT engineering resins have good resistance to chemicals
and being crystalline, do not suffer from the solvent stress-cracking problems that
plague amorphous materials such as polycarbonate. Polyesters are only attacked
by powerful acidic or phenolic solvents, hot strong aqueous alkali, and certain
bases such as hydrazine.

In the unfilled state, PET is not a good molding resin and all commercial

grades are filled with either chopped glass strand (typically 3–4 mm long) or
mineral fillers (usually mica) or a mixture of the two. Various proprietary nucle-
ating agents added are often sodium salts of various organic carboxylic acids (see
Ref. 5). Some manufacturers supply fire-retardant (FR) polymer grades as well.
Such formulations often involve a synergistic mixture of an aryl halide with anti-
mony oxide. One difficulty with flame-retardant PET polymer is that recipes which
contain antimony trioxide can suffer severe polymer degradation at molding tem-
peratures around 280–290

◦

C. As has been discussed earlier, antimony trioxide is

a polymerization catalyst and can also act as prodegradant at high temperatures.
This is a less serious problem with FR grades of PBT because of lower processing
temperatures (240–250

◦

C). To mitigate the effect on PET, pentavalent antimony

compounds such as sodium antimonate is used in some FR formulations. The halo-
genated species is often a ring-brominated polystyrene (85–90). These materials

516

POLYESTERS, THERMOPLASTIC

Vol. 7

are approved for Underwriters Laboratory V-0 classification and are nonfugitive
and less likely to be environmentally harmful than are small molecule additives
like decabromodiphenyl oxide. In Europe there has recently been a move to elimi-
nate organohalogen compounds from FR formulations for environmental reasons.
As a result, organophosphorus derivatives such as triarylphosphine oxides are
being introduced (91).

Since the early 1990s in the United States, there has been pressure on man-

ufacturers from the automotive industry, led notably by the Ford Motor Co., to use
at least 25 wt% of post consumer recycled (PCR) material in their resins rather
than 100% virgin polymer. In the case of PET, this is readily achievable, because
of the large volume of recovered PET bottle polymer chip now available in the
United States and in Europe. A recent article (82) states that during 1998–1999,
273,000 t of PET waste was recycled in the United States. Some suppliers use
100% recycled polymer in their compounded PET resins. The PCR recovery pro-
cess and subsequent melt-compounding and reformulation reduce the initial IV
of the original resin from around 0.75–0.85 dL/g to values around 0.62–0.65 dL/g,
but this is still in the range of virgin melt-polymerized resin (as opposed to SSP
resin). In the automobile industry, components are often pigmented black, so that
differences in feedstock PCR chip color are minimized.

Initially, PET moldings were used in small components, typically electrical

connectors and covers for fuses, etc (see Figs. 1 and 2). This is still the case, even

Fig. 1.

Electrical automotive connectors molded from CelanexR PBT, Grote & Hartmann

(Germany). Courtesy of Ticona, A business of Celanese AG.

Vol. 7

POLYESTERS, THERMOPLASTIC

517

Fig. 2.

Appliance timer housings molded from CelanexR PBT, Mallory (USA). Courtesy

of Ticona, A business of Celanese AG.

Fig. 3.

Healthcare inhaler for asthma patients molded from CelanexR PBT, Astra Tur-

bohaler(tm). Courtesy of Ticona, A business of Celanese AG.

more so with the growing complexity of the modern automobile with its numerous
on-board electronics, but recently the trend has been to use PET moldings more
and more for non-load-bearing structural parts, such as radiator grille supports
and headlamp mountings (Fig. 4). Glass/mineral filled PET moldings do not have

518

POLYESTERS, THERMOPLASTIC

Vol. 7

Fig. 4.

Automotive, electroplatable grill-opening retainer molded from ImpetR PET, Mer-

cury Sable. Courtesy of Ticona, A business of Celanese AG.

a smooth enough surface for exterior body-parts. However, they are very suit-
able for internal structural components. This trend is part of the drive to reduce
automobile weight to improve gas-mileage or boost the range and performance
of electric cars. Some moldings are dimensionally quite large, weighing well over
5 kg per shot. Improvements in mold design and better understanding of melt-flow
behavior in molds, brought about by increasing use of computer-aided design and
flow-simulation programs, have helped to make these large moldings possible on a
routine production basis. Figures 5 and 6 show an experimental concept car where
the bulk of the body/chassis pan is molded in a PET formulation. As stated, PET

Fig. 5.

Chrysler composite concept vehicle (CCV) with large body parts molded from

ImpetR PET, Daimler Chrysler. Courtesy of Ticona, A business of Celanese AG.

Vol. 7

POLYESTERS, THERMOPLASTIC

519

Fig. 6.

Large body parts for Chrysler composite concept vehicle (CCV) molded from Im-

petR PET, Daimler Chrysler. Courtesy of Ticona, A business of Celanese AG.

does not crystallize well in the unoriented state even in a hot mold unless nucle-
ating agents and/or plasticizers (qv) are added. Commercial PET molding-grade
polymers are nearly always filled. Typical compounded polymer properties are
shown in Table 3.

Table 3. Typical Properties of PET Molding Resins

a

Glass

ASTM

(mineral

Property

method

30%

45%

35%

filler)

Specific gravity

D792

1.58

1.70

1.60

1.60

Tensile strength, MPa

b

D638

166

197

97

103

Elongation at break, %

D638

2.0

2.0

2.2

2.1

Flexural strength at 5%, MPa

b

D790

245

310

148

152

Flexural modulus, GPa

c

D790

9.66

14.5

9.66

9.66

Notched Izod, J/m

d

D256

80.1

107

58.7

58.7

Heat deflection

D648

224

229

202

216

Temperature at 1.82 MPa

b

,

◦

C

Flammability

e

UL-94

HB

HB

HB

HB

Dielectric strength, V/25

µm

D149

5.2 mm

565

540

500

450

1.6 mm

904

631

550

575

0.8 mm

975

951

810

860

Volume resistivity at 23

◦

C, 50% rh,

·cm

D257

3.0

× 10

15

1.0

1.0

Dielectric constant

ε, H

2

D150

10

3

3.2

3.5

3.8

3.8

10

5

3.1

3.4

3.6

3.7

a

Ref. 92.

b

To convert MPa to psi, multiply by 145.

c

To convert GPa to psi, multiply by 145,000.

d

To convert J/m to ft

·lbf/in., divide by 53.38.

e

HB

= Brinell hardness.

520

POLYESTERS, THERMOPLASTIC

Vol. 7

Economic Aspects of PET

The total world market for nonfiber PET polymer is still growing because of the
demand for bottle resin. This grew worldwide from 1993 to 2000 at about 15%/year,
although overproduction led to a falling-off recently. A published article (93) stated
that the global consumption of PET grew at 19% during 1992 to a world total of
1,719,000 ton (3.79 billion lbs). Adding up the bottle resin figures from Table 1
given an estimated total of 3,121,000 ton, a 181% increase in 7 years. The world
total devoted to beverage bottles is astonishing, but it is significant that Europe
had by the year 2000 almost caught up with the United States in total bottle resin
tonnage. There are interesting regional variations, notably the strong demand for
packaging for mineral waters in Europe and the large hot-fill market in Japan.
Looking ahead, the future growth of PET looks to be assured, its balance of prop-
erties and ability to be recycled being extremely favorable factors. New outlets for
PET packaging are constantly appearing.

One significant trend in recent years has been the complete disappearance of

some well-known major producers of PET polymer and the emergence of new ones.
In the late 1990s, several major chemical companies sold off commodity chemical
and polymer businesses to concentrate on other more profitable market areas
such as pharmaceuticals and life-sciences. In 1998–1999, Hoechst Celanese split
into Celanese AG and Aventis, a pharmaceutical company. It sold its large PET
interests to the Koch-Saba group, now called Kosa, and headquartered in Houston,
Tex. Celanese’s remaining engineering polymer interests were taken over by its
subsidiary Ticona. During 1998–2000, Shell gave up its polymer business and sold
some 300,000 ton of PET capacity to the Mossi & Ghisolfi Group of Milan, Italy
(94). ICI sold its PET plants in Europe to DuPont, and Dow in 1996 became a
major PET manufacturer by acquiring 270,000 ton of plant capacity. Dow is also
in the process of buying Union Carbide. Allied-Signal became Honeywell and were
then almost acquired by GE until (2001) the European Union prevented the sale
on antitrust grounds. Table 4 shows major world manufacturers of PET resin in
late 1998, a situation that has now changed again with the removal of Shell and
Hoechst Celanese as PET producers.

Table 4. Major World PET Producers

a

Manufacturer

Nameplate capacity, 10

3

t

Eastman

1556

Kosa

840

DuPont

758

Shell

750

Nan Ya

540

Wellman

490

Rhodia-Ster

280

Hoechst-Celanese

250

Sunkyong

240

Far Eastern

200

Total

5904

a

Ref. 83.

Vol. 7

POLYESTERS, THERMOPLASTIC

521

Commercial PET Engineering Resins.

The first company to introduce

nucleated PET molding resins was Akzo Plastics BV with their Arnite

®

range,

later acquired by DSM NV. DuPont introduced their Rynite

®

fast-crystallizing

materials in 1978, followed by other manufacturers. The present North
American market is still dominated by DuPont, but there are other notable
suppliers like Honeywell (Allied-Signal) with Petra

®

and Ticona with Impet

®

.

Eastman also markets PET injection-molding grades under the tradename
Thermx. In Europe the situation with PET is confined primarily to two suppli-
ers, DSM NV and DuPont, while in Japan the main suppliers are Teijin, Toray,
and Toyobo. The market prices for PET molding resins at the time of writing in the
United States are $2.90–3.15/kg for 30% glass-filled PET, $3.26–3.41/kg for 55%
glass-filled PET, and $3.23–3.45/kg for 30% glass-filled FR grade PET (95). These
compare with a current commodity price for PET bottle resin of $1.36–1.50/kg (see
E

NGINEERING

T

HERMOPLASTICS

).

Safety and Environmental Factors

PET.

PET polymer is safe and poses no threat to animals or humans. PET

fibers have been in use for nearly 50 years and PET has U.S. Food and Drug Ad-
ministration (FDA) approval for use as a food-packaging material. PET fibers have
been used in internal arterial prostheses. The only significant hazard in handling
PET resins is the dust hazard associated with mineral or glass fillers during chip
grinding or compounding operations. Appropriate protective equipment must be
worn. All extruders or machinery handling molten polymer should be properly
ventilated to remove harmful fumes from the decomposition of molten polymer.
Molten PET can cause serious contact thermal burns: it has a high heat capacity
and sticks to the flesh. Adequate protection must always be worn when handling
hot polymer.

PBT.

PBT resins are not harmful or hazardous when handled at room tem-

perature under normal conditions according to their Materials Safety Data Sheets.
No problem with contact with the pellets has been encountered under normal con-
ditions. Glass fines can however cause skin irritation, and if glass-filled resins are
being ground or reground, due precautions must be taken. Inhalation of dust
must be guarded against, as is true for grinding any glass-filled resin. During the
molding, the temperature must not exceed 520

◦

F and never over 550

◦

F as decom-

position with the evolution of harmful vapors can occur. As with all thermoplastics,
adequate ventilation must be provided around injection-molding machines.

PBT Molding Resins

Poly(butylene terephthalate) is historically the oldest of the crystalline ther-
moplastic polyester molding resins, having been introduced by Celanese Corp.
in 1970 under the trade name Celanex. General Electric Co. then brought out
their own version Valox

®

, and today there are numerous suppliers including

BASF, Bayer AG, and DSM. Celanese have recently become Ticona. In Japan,

522

POLYESTERS, THERMOPLASTIC

Vol. 7

the major manufacturers are Polyplastics, Toray, Teijin, and Mitsubishi (see
E

NGINEERING

T

HERMOPLASTICS

). As already explained, unlike PET, PBT has the

ability to crystallize very rapidly (like nylon-6,6) even in a cold mold. It gives tough,
distortion-free moldings without special additives or nucleants. Although the un-
modified polymer has very good flow properties, and is used in electrical connec-
tors and fiber-optical cable buffer tubes, it performs even better if reinforced with
inorganic fillers, notably

∼3 mm (1/8 in) chopped glass fiber. Additional mineral

fillers are incorporated for special applications where high heat-deflection temper-
ature and stiffness are important. These mineral fillers include mica, talc, wollas-
tonite, and even barium sulfate, the latter for special applications such as counter-
tops.

The filled grades of PBT are tougher, stiffer, and stronger materials and

they have improved notched Izod impact strength, since unfilled PBT is notch-
sensitive (96,97). Even when unfilled, the plastic has good strength, rigidity and
toughness, low creep, minimal moisture absorbance, and does not undergo di-
mensional changes with changes in humidity. It is characterized by excellent
electrical and dielectric properties and a high surface finish. It has found wide
acceptance in a variety of end uses where precision molding and a high quality
finish are required. Typical are the electrical and electronic markets where it is
widely used in such parts as connectors, plugs, switches, typewriter and computer
keyboard components, printed circuit boards, and small electric motor compo-
nents. PBT is widely used in the automotive industry for electrical components
such as distributor caps, coil-formers, rotors, windshield wiper arms, headlight
mountings, and other fittings. In the auto market, “under the hood” components
have to maintain their dimensional stability at elevated temperatures as well
as resist various automotive fluids. Other uses for PBT are home appliances,
such as food mixers, hair dryers, coffee makers, toasters, and camera parts. PBT
is used in industrial machinery, for example in molded conveyor-belt links and
in medical devices such as nasal sprays and nebulizers. An appreciable quan-
tity is used in polymer alloys and blends with other polymers. PBT is marketed
in both standard and flame-retardant grades, the latter being essential in the
United States to meet Underwriters’ Laboratory 94V-0 standards in thin-walled
sections.

Properties of PBT

Physical Properties.

Unlike PET, the polymer PBT exists in two poly-

morphs, the

α and β forms, which have distinctly different crystal structures. The

two forms are interconvertible under mechanical stress (98–100). Both crystal
forms are triclinic and the crystal parameters are shown in Table 5.

The change in the two forms mainly involves the c lattice dimension, which

lengthens from 1.174 to 1.300 nm. It is believed that the relaxed

α-form has a

gauche–trans-gauche conformation of the three C C bonds in the C

4

-glycol unit,

and the extended

β-form exists in an all-trans form (62). The melting point of PBT

is 222–224

◦

C, depending on the degree of crystallization and annealing conditions.

The heat of fusion is about 140 J/g (101) and the T

g

is usually quoted at about 45

◦

C,

although this depends on the physical nature of the sample (102). PBT crystallizes

Vol. 7

POLYESTERS, THERMOPLASTIC

523

Table 5. Crystal Parameters for the Two Forms of PBT

Cell parameter

α form (unstretched)

β form (stretched)

a, nm

0.482

0.469

b, nm

0.593

0.580

c, nm

1.174 (1.165)

a

1.300

α, degrees

100 (98.9)

a

102

β, degrees

115.5 (116.6)

a

120.5

γ , degrees

111

105

Volume, nm

3

0.260

0.267

density, g/cc

1.41

1.37

a

Results given in Ref. 100.

Table 6. Mark–Houwink Parameters for PBT

Solvent

K, dL/g

α

P/TCE (at 30

◦

C)

1.17

× 10

− 4

0.87

OCP (at 25

◦

C)

6.62

× 10

− 5

0.915

very readily from the melt and it is difficult to obtain a truly amorphous sample.
Its crystallization kinetics are thus not easy to determine (103). The density of the
annealed crystalline unfilled polymer is 1.33 g/cm

3

whereas the amorphous mate-

rial has a value of 1.26 g/cm

3

(104). Like PET, PBT is made to various molecular

weights, the M

n

values being in the 20,000–50,000 range. Intrinsic viscosities are

usually measured in o-chlorophenol (OCP) or a phenol/tetrachlorethane mixture.
The Mark–Houwink parameters are shown in Table 6 (105,106).

Flame-retardant grades of PBT usually consist of synergistic mixtures of

antimony trioxide with various halogenated (brominated) aromatic compounds.
A typical recipe for PBT might be 10 wt% decabromodiphenyl oxide and 5 wt%
antimony oxide. Recently the trend has been to use polymeric or oligomeric bromi-
nated additives. A typical additive is an end-capped polycarbonate derived from
tetrabromobisphenol-A [94334-64-2]; another is a mixture of epoxy oligomers de-
rived from the diglycidyl ether of tetrabromobisphenol-A [68928-70-1]. The bromi-
nated polystyrenes (loc. cit.) have only limited usefulness in PBT as they have a
low melt compatibility (107).

Chemical Properties.

PBT is highly crystalline and does not suffer from

solvent stress corrosion cracking as do amorphous materials. It is resistant at
room temperature to most common chemicals and solvents, lubricants, greases,
and automotive fluids. Ketones will attack it at elevated temperatures. Parts
made from PBT are dishwasher-safe, but will not withstand repeated steam au-
toclaving. PBT is attacked by aqueous alkali and other strong bases and by di-
lute acids, particularly at elevated temperatures. PBT has very good resistance
to weathering, and black pigmented grades with uv-stabilizers have excellent
outdoor stability. Like all polyesters, PBT is susceptible to hydrolytic attack by
moisture in the melt. Injection-molding screw temperatures are usually about
250

◦

C and IV drop is rapid unless the polymer chip is dried to below 50 ppm

moisture content and kept dry. Inadequate drying is the cause of most molding

524

POLYESTERS, THERMOPLASTIC

Vol. 7

Table 7. Mechanical Properties of PBT

a

30% glass

Unfilled, grade,

General

Flame

High

Property

low mol. wt.

purpose

retardant

impact

Specific gravity

1.31

1.54

1.66

1.53

Tensile strength, MPa

b

57

135

135

97

Tensile modulus, GPa

c

2.5

9.7

11.7

8.3

Elongation, %

5

2

1.5

3.1

Flexural strength, MPa

b

85.5

193

193

152

Flexural modulus, GPa

c

2.5

8.3

10.3

6.9

Notched Izod, J/m

d

37.4

90.7

69.4

160

Unnotched Izod, J/m

d

1228

240

214

641

HDT at 1.82 MPa

b

,

◦

C

51

206

208

191

Volume resistivity,

·cm

10

15

10

16

5

× 10

15

4

× 10

14

Dielectric strength, V/25

µm

420

560

490

500

Dielectric constant

ε, 100 Hz

3.2

3.7

3.9

4.3

Flammability UL94, at 0.8 mm

HB

HB

V0

HB

a

Ref. 108.

b

To convert MPa to psi, multiply by 145.

c

To convert GPa to psi, multiply by 145,000.

d

To convert J/m to ft

·lbf/in., divide by 53.38.

problems. As the PBT melt is quite fluid, drooling from nozzles is sometimes a
problem.

Mechanical Properties.

Properties of typical grades of PBT, either as un-

filled neat resin or glass-fiber filled, and FR grades are set out in Table 7. This
table also includes impact-modified grades which incorporate dispersions of elas-
tomeric particles inside the semicrystalline polyester matrix. These dispersions
act as effective toughening agents which greatly improve impact properties. The
mechanisms are not fully understood in all cases. The subject has been discussed
in detail (109), and the particular case of impact-modified polyesters such as PBT
has also been discussed (110,111).

Economic Aspects of PBT

According to information published between 1999 and 2000 the market for PBT
in the United States and Canada increased from 67,000 to 71,000 t (147.6–156.6
million lbs), as shown in Table 2. The European and Japanese markets also use
large amounts of PBT. The main outlets are electrical, electronic, and automotive
components. The major manufacturers of PBT in North America, Europe, and
Japan are shown in Table 8.

A more detailed analysis of end uses for PBT in North America, Europe, and

Japan are shown in Table 9.

The future prospects for economic growth for PBT look good. It is still displac-

ing thermosets from some markets and its versatility, excellent flow properties,
and ease of molding will assure it a prominent place in the world for years to

Vol. 7

POLYESTERS, THERMOPLASTIC

525

Table 8. Principal World Manufacturers of PBT According to Region

USA

Europe

Japan

BASF – ULTRADUR

®

Bakelite AG

General Electric

DuPont – CRASTIN

®

BASF

Mitsubishi Eng. Plastics

General Electric – VALOX

®

Bayer AG

Polyplastics

a

Ticona – CELANEX

®

DSM NV

Teijin

a

DuPont

Toray

General Electric

Toyobo

a

These two companies merged in 2000 to form Wintech which will produce both PET and

PBT molding resins.

Table 9. End Uses of PBT for 1999–2000, ton

Market

1999

2000

United States and Canada

a

Appliances

7,020

7,392

Consumer/recreation

2,267

2,267

Electrical/electronic

14,286

15,420

Industrial

7,165

7,392

Transportation

33,559

35,827

Others

2,494

2,721

TOTAL

66,971

71,019

Western Europe

b

Appliances

11,000

12,500

Consumer/recreation

9,000

9,900

Electrical/electronic

36,700

39,700

Industrial

8,700

9,800

Transportation

43,300

48,200

Other

7,300

8,900

TOTAL

116,000

129,000

Japan

c

Automotive and vehicle

30,000

33,000

Electrical/electronic

38,000

44,000

Other

19,000

19,000

Total domestic

87,000

96,000

Exports

32,000

34,000

TOTAL

119,000

130,000

a

Ref. 112.

b

Ref. 113.

c

Ref. 114.

come. Quoted prices (Nov. 1995) for PBT resins were $3.61–3.85/kg for unfilled
resin, $3.74–4.18/kg for 30% glass-filled FR grades, and $4.29–4.51/kg for high
impact grades (115).

BIBLIOGRAPHY

“Polyesters, Thermoplastic” in EPST 1st ed., Suppl. Vol. 1, pp. 444–467, by J. R. Caldwell,
W. J. Jackson Jr., and T. F. Gray Jr., Tennessee Eastman Co., Division of Eastman Kodak

526

POLYESTERS, THERMOPLASTIC

Vol. 7

Co.; “Polyesters, Thermoplastic” in EPSE 2nd ed., Vol. 12, pp. 217–256, by J. Y. Jadhav and
S. W. Kantor, University of Massachusetts. “Polyesters” in EPSE 2nd ed., Vol. 12, pp. 1–75,
by I. Goodman.

1. H. Mark and G. S. Whitby, eds., The Collected Works of W H Carothers on High

Polymeric Substances, Interscience Publishers Inc., New York, 1940.

2. U.K. Pat. 578,079 (June 14, 1946), J. R. Whinfield and J. T. Dickson (to ICI Ltd.).
3. U.S. Pat. 2,465,319 (Mar. 22, 1949), J. R. Whinfield and J. T. Dickson (to E. I. du Pont

de Nemours & Co., Inc.).

4. J. Shimizu and co-workers, J. Polym. Sci., Polym. Phys. Ed. 21, 275–284 (1983).
5. U.K. Pat. 1,239,751 (July 21, 1971), A. J. Dijkstra, J. A. W. Reid, and I. Goodman (to

ICI Ltd.).

6. U.S. Pat. 4,393,178 (July 12, 1983), R. M. H. Legras and co-workers, (to Imperial

Chemical Industries plc).

7. U.S. Pat. 4,380,621 (Apr. 19, 1983), E. Nield, D. E. Higgins, and M. W. Young (to

Imperial Chemical Industries plc).

8. S. M. Hansen and P. B. Sargeant, Kirk-Othmer Encyclopedia of Chemical Technology,

4th ed., Vol. 10, John Wiley & Sons, Inc., New York, Figure 8, p. 679.

9. Br. Pat. 604,073 (1948), J. G. Cook, H. P. W. Hugill, and A. R. Low (to I.C.I. Ltd.).

10. W. S. Witts, Chem. Process. Eng. 51, 55 (1970).
11. U.S. Pat. 3,467,724 (Sep. 16, 1969), S. A. Laurich (to Chevron Research Co.).
12. U.S. Pat. 3,636,180 (Jan. 18, 1972), D. B. Broughton (to Universal Oil Products

Co.).

13. U.K. Pat. 994,769 (June 10, 1965) (to Standard Oil Co.).
14. U.S. Pat. 3,584,039 (June 8, 1971), D. H. Meyer (to Standard Oil Co.).
15. J. Milgrom, Outlook for Plastics Recycling in the United States, Decision Resources

Inc., 1993, pp. 25–27.

16. K. A. Domeshek, in D. Brunnschweiler and J. Hearle, eds., Tomorrow’s Ideas and Prof-

its: Polyester: Fifty Years of Achievement, The Textile Institute, Manchester, England,
1993, pp. 288–290.

17. M. Poliakoff and co-workers, Chem. Eng. News 30, (July 16, 2001).
18. G. O. Curme and F. Johnston, Glycols (ACE Monograph No. 114), Reinhold Publishing

Corp., New York, 1952, Chapt. “2”.

19. U.S. Pat. 3,920,759 (Nov. 18, 1975), E. V. Hort (to GAF Corp.).
20. U.S. Pat. 3,950,441 (Apr. 13, 1976), S. Rudolf and W. DeThomas (to GAF Corp.).
21. U.S. Pat. 3,922,300 (Nov. 25, 1975), T. Onoda and co-workers, (to Mitsubishi Chemical

Industries Ltd.).

22. U.S. Pat. 4,010,197 (Mar. 1, 1977), M. J. Toriya and K. Shiraga (to Mitsubishi Chemical

Industries Ltd.).

23. U.S. Pat. 4,064,145 (Dec. 20, 1977), P. D. Taylor (to Celanese Corp.).
24. U.S. Pat. 4,356,125 (Oct. 26, 1982), N. A. de Munck and J. F. Scholten (to Stamicarbon

BV).

25. W. L. Hergenrother and C. J. Nelson, J. Polym. Sci., Polym. Chem. Ed. 12, 2905

(1974).

26. R. E. Wilfong, J. Polym. Sci. 54, 388 (1961).
27. Fr. Pat. 1,169,659 (1959) (to ICI Ltd.).
28. Can. Pat. 573,301 (1959), W. K. Easley and co-workers, (to Chemstrand Corp.).
29. H. Zimmerman and N. T. Kim, Polym. Sci. Eng. 20, 680–683 (1980).
30. U.S. Pat. 4,501,878 (Feb. 26, 1985), L. J. Adams (to Eastman Kodak Co.).
31. U.S. Pat. 2,822,348 (1958), J. H. Haslam (to E. I. du Pont de Nemours & Co.,

Inc.).

32. U.S. Pat. 4,439,597 (Mar. 27, 1984), H. K. Hall and A. B. Padias (to Celanese

Corp.).

Vol. 7

POLYESTERS, THERMOPLASTIC

527

33. U.S. Pat. 4,824,930 (Apr. 25, 1989), M. L. Doerr (to Celanese Fibers Inc.).
34. U.S. Pat. 4,511,708 (Apr. 16, 1985), T. Kasuga and co-workers, (to Polyplastics Co.

Ltd.).

35. U.S. Pat. 4,656,241 (Apr. 7, 1987), H. Iida, K. Azuma, and M. Hayashi (to Toray In-

dustries Inc.).

36. C. M. Fontana, J. Polym. Sci., Part A-1 6, 2343–2358 (1968);

Ref. 10 in this paper

lists numerous U.S. patents relating to antimony oxide as a polymerization catalyst.

37. U.S. Pat. 4,820,795 (Apr. 11, 1989), S. Hirata and Y. Watanabe (to Toyo Seikan Kaisha).
38. U.S. Pat. 4,499,261 (Feb. 12, 1985), H. Heinze and co-workers, (to Davy McKee

AG).

39. U.S. Pat. 4.973,655 (Nov. 27, 1990), G. Pipper and co-workers, (to BASF AG).
40. U.S. Pat. 5,064,935 (Nov. 12, 1991), R. Jackson, D. J. Lowe, and C. A. Stewart (to E. I.

du Pont de Nemours & Co., Inc.).

41. I. Goodman and B. F. Nesbitt, J. Polym. Sci. 48, 423–433 (1960).
42. S. Shiono, J. Polym. Sci., Polym. Chem. Ed. 17, 4123 (1979).
43. G. C. East and M. Girshab, Polymer 23, 323–324 (1982).
44. U.S. Pat. 3,718,621 (Feb. 27, 1973), W. K. Wilson (to Goodyear Tire & Rubber Co.).
45. U.S. Pat. 3,816,377 (June 11, 1974), Y. Okuzumi (to Goodyear Tire & Rubber

Co.).

46. U.S. Pat. 4,064,112 (Dec. 20, 1977), H. J. Rothe and co-workers, (to Zimmer AG).
47. U.S. Pat. 4,161,578 (July 17, 1979), D. J. Herron (to Bepex Corp.).
48. L. H. Buxbaum, Angew. Chem., Int. Ed. 7, 182–190 (1968).
49. Ref. 1, p. 22.
50. H. Zimmerman and N. T. Kim, Polym. Eng. Sci. 20, 680–683 (1980).
51. E. P. Goodings, Soc. Chem. Ind. London., Chem. Eng. Group (Monograph 13), Society

of Chemical Industry, London, 1961, 2112.

52. K. Tomita, Polymer 18, 295 (1977).
53. A. von Hassell, Plast. Technol. 70 (Jan. 1979).
54. R. de P. Daubeny and C. W. Bunn, Proc. R. Soc. London, Ser. A 226, 531 (1954).
55. S. Fakirov, E. W. Fischer, and G. F. Schmidt, Makromol. Chem. 176, 2459

(1975).

56. E. Bornschegl and R. Bonart, Colloid Polym. Sci. 258, 319 (1980).
57. Y. Kitano, Y. Kinoshita, and T. Ashida, Polymer 36, 10, (1995).
58. M. Droscher, Makromol. Chem. 181, 789 (1980).
59. H. W. Starkweather, P. Zoller, and G. A. Jones, J. Polym. Sci., Polym. Phys. Ed. 21,

295 (1983).

60. J. Menczel and B. Wunderlich, J. Polym. Sci., Polym. Lett. Ed. 19, 261 (1981).
61. K. H. Illers and H. Breuer, J. Colloid Sci. 18, 1–31 (1963);

J. H. Dumbleton, T.

Murayama, and J. P. Bell, Kolloid Z. Z. Polym. 228, 54–58 (1968).

62. D. B. G. Jaquiss, W. F. H. Borman, and R. W. Campbell, Kirk-Othmer Encyclopedia of

Chemical Technology, 3rd ed., Vol. 18, p. 552.

63. S. Berkowitz, J. Appl. Polym. Sci. 29, 4353–4361, (1984).
64. Ref. 8, pp. 662–685.
65. K. J. McKenzie, in Ref. 8, pp. 761–787.
66. U.S. Pat. 3,733,309 (May 15, 1973), N. C. Wyeth and R. N. Roseveare (to E. I. du Pont

de Nemours & Co., Inc.).

67. U.S. Pat. 3,778,214 (Dec. 11, 1973), N. C. Wyeth and co-workers (to E. I. du Pont de

Nemours & Co., Inc.).

68. R. W. Tock and co-workers, Adv. Polym. Technol. 4, 307–315 (1984).
69. E. A. Muccio, Plastics Processing Technology, Materials Information Society, 1994,

Chapt. “5”.

70. A. von Hassell, Plast. Technol., 74 (Jan. 1979).

528

POLYESTERS, THERMOPLASTIC

Vol. 7

71. A. von Hassell, Plast. Technol., 76 (Jan. 1979).
72. D. G. Callander, Polym. Eng. Sci. 25, 453–457 (1985).
73. U.S. Pat. 5,552,513 (Sept. 3, 1996), K. K. Bhatia (to E. I. du Pont de Nemours & Co.,

Inc.).

74. U.S. Pat. 5,677,415 (Oct. 14, 1997), K. K. Bhatia (to E. I. du Pont de Nemours & Co.,

Inc.).

75. U.S. Pat. 5,865,423 (Jan. 5, 1999), K. K. Bhatia (to E. I. du Pont de Nemours & Co.,

Inc.).

76. U.S. Pat. 4,578,437 (Mar. 25, 1986), R. R. Light (to Eastman Kodak Co.).
77. P. A. Aspy and E. E. Denison, Mod. Plast. 74–76 (Aug. 1983).
78. nant Surface Technology SAcos, Mod. Plast. 12 (June 2000).
79. res AGcos, Mod. Plast. 12 (Dec. 2000).
80. il Corpcos, Mod. Plast. 10 (Sept. 2000).
81. Modern Plastics Encyclopedia, 2000, p. A-196.
82. Plast. Technol. 100 (Sept. 2000).
83. Mod. Plast. 58 (Jan. 1999).
84. Mod. Plast. 42–49 (Feb. 2001).
85. U.S. Pat. 4,074,032 (Feb. 14, 1978), H. Naarman and co-workers, (to BASF AG).
86. U.S. Pat. 4,143,221 (Mar. 6, 1979), H. Naarman and co-workers, (to BASF AG).
87. U.S. Pat. 4,200,703 (Apr. 29, 1980), K. Diebel and co-workers, (to Chemische Werke

Huls AG).

88. U.S. Pat. 4,360,455 (Nov. 23, 1982), G. Lindenschmidt and co-workers, (to BASF AG).
89. U.S. Pat. 4,352,909 (Oct. 5, 1982), H. J. Barda and S. L. Gray (to Ferro Corp.).
90. U.S. Pat. 4,879,353 (Nov. 7, 1989), D. C. Sanders and co-workers, (to Great Lakes

Chemical Corp.).

91. Eur. Pat. 664296 (July 26, 1995), V. Carletti and co-workers (to Great Lakes Chemical

Italia).

92. Technical Data, IMPET Resins, Hoechst-Celanese Corp., Summit N.J., 1994.
93. A. K. Mitchell, Mod. Plast. 71(12), B48–B50 (1994).
94. Mod. Plast. 14 (Mar. 2000).
95. L. M. Sherman, Plast. Technol. 77 (May 1995).
96. Ger. Offen. 2,042,447 (Apr. 15, 1971) and U.S. Appl. (Aug. 29, 1969), D. D. Zimmerman

and R. B. Isaacson, (to Celanese Corp.).

97. Ger. Offen. 2,533,358 (Feb. 26, 1976) and U.S. Appl. 496,607 (Aug. 12, 1974), F. M.

Berardinelli (to Celanese Corp.).

98. E. Bornschegl and R. Bonart, Colloid Polym. Sci. 258, 319 (1980).
99. I. H. Hall and M. G. Pass, Polymer 17, 807 (1976).

100. R. Jakeways, I. M. Ward, and M. A. Wilding, J. Polym. Sci., Polym. Phys. Ed. 13,

807–809 (1973).

101. K. H. Illers, Colloid Polym. Sci. 258(2), 117 (1980).
102. S. Z. D. Cheng, R. Pan, and B. Wunderlich, Makromol. Chem. 189, 2443 (1988).
103. C. F. Pratt and S. Y. Hobbs, Polymer 17, 12 (1976).
104. K. Tashiro and co-workers, Macromolecules 13, 137 (1980).
105. W. F. H. Borman, J. Appl. Polym. Sci. 22, 2119 (1978).
106. Ho Nan Sung, Hoechst Celanese Corp., Summit N.J., private communication, 1994.
107. J. C. Gill, Plast. Compd., 77–81 (Sept.–Oct. 1989).
108. Ref. 92, Table 8.
109. C. B. Bucknall, Toughened Plastics, Applied Science, London, 1977.
110. C. B. Bucknall and C. J. Page, J. Mater. Sci. 808 (1982).
111. D. J. Hourston, S. Lane, and H. X. Zhang, Polymer 32, 2215–2220 (1991).
112. Mod. Plast. 43 (Feb. 2001).
113. Mod. Plast. 46 (Feb. 2001).

Vol. 7

POLYIMIDES

529

114. Mod. Plast. 49 (Feb. 2001).
115. L. M. Sherman and J. DeGaspari, Plast. Technol., 71 (Nov. 1995).

A

NTHONY

J. E

AST

Consultant

POLYETHERETHERKETONES (PEEK).

See E

NGINEERING

T

HERMOPLASTICS

.

POLYETHYLENE.

See E

THYLENE

P

OLYMERS

.

POLY(ETHYLENE-NORBORNENE).

See E

THYLENE

-N

ORBORNENE

C

OPOLYMERS

.

POLY(3-HYDROXYALKANOATES).

See Volume 3.