Note: Descriptions are shown in the official language in which they were submitted.
CA 02247239 1998-08-24
CR-9812-A
TITLE
IMPROVED APPARATUS AND PROCESS
FOR A POLYCONDENSATION REACTION
FIELD OF THE INVENTION
An improved process and apparatus for the production of a polyester or
another condensation polymer is disclosed. In particular, polymerization is
conducted in a reaction vessel equipped with a specially designed agitator
that
exposes the polymer melt within the reaction vessel to inert gas floating
through
the vessel. The agitator comprises a plurality of elements that lift a portion
of the
polymer melt in the reaction vessel and generate films of the polymer melt
which
films extend in planes that are parallel to central axis of the agitator and
the flo«~ of
gas through the reaction vessel.
TECHI~TICAL BACKGROUND
Polyester production from aromatic dicarbox5~lic acids or their esters such
1 ~ as dimethyl terephthalate (DMT), and glycols is lno~n. This has been
accomplished by stage-Elise melt polymerization of the dihydroxy ester of the
aromatic dicarboxylic .acid. or low molecular weight oligomers thereof. under
a:. r.. a r f ~.. ~ t; +
Sut;t;eSSi'vety lliglier vaC'u',;uW GvuumlouS: m oruCr Wr ~ilv pOyIT:Ci.~~mOn
~C
continue to the degree needed for most commercial applications. the
condensation
~0 by-products. especially ethylene glycol, must be removed from the reaction
system
at vacuums as high as 133-399 Pa (1-3 mm Hg). Such processes require costly
high vacuum equipment, multistage steam jets to create the vacuum. and N
purged seals and flanges to minimize leakage of air into the system.
Condensate
from the steam jets and organic b~~-products from the system end up as a waste
'?~ water stream that requires treatment and contributes to volatile organic
emissions
to the air. The present invention relates to a less costly polymerization
process that
can be carried out at atmospheric pressure.
Atmospheric pressure processes employing an inert gas have been
disclosed in the prior art. but these suffer from one or more dra~i-backs such
as
30 (1) the quantity of inert gas used is too large to be economical; (2) the
reactor size
might not be feasible for commercial-scale operation; (3) inert-gas velocities
may
be too high to be feasible for commercial-scale production, or (4) contact
between
the inert gas and the polymer melt in tine reactor be inadequate or non-
uniform.
Because of such drawbacks, the processes presently employed for commercial
35 production of polyester continue to be conducted under high vacuum. One
object
of the present invention is to provide further improvement in a process, at
about
atmospheric pressure, for continuous or batch~~ise production of polyesters,
particularly polyethylene terephthalate, of high molecular weight. In another
aspect of the present invention, an improved apparatus that may be employed in
a
1
AM~r!~'~D SY~c~f
CA 02247239 1998-08-24
WO 97/35902 ' PCT/US97/04487
reaction process involving mass transfer of a volatile by-product into an
inert gas,
is disclosed. .
SUMMARY OF THE INVENTION
The present invention is directed to a process for manufacturing polyesters
of aromatic dicarboxylic acids and glycols in a molten state in which an inert
gas
is employed to assist in removing a volatile condensation by-product, wherein
the
improvement comprises, employing a horizontally disposed cylindrical reactor
vessel partly filled with a polymerization reaction mass in the form of a
melt,
which reactor vessel is equipped with the following:
a) a reactor inlet for introducing a polymerizable feed into the
reactor vessel;
b) a gas inlet for introducing an inert gas at or near one end of the
reactor vessel and a gas outlet for removing the inert gas at or near an
opposite
end of the reactor vessel, thereby resulting in gas flow past the reaction
mass in
I S the reactor vessel;
c) means for maintaining the reaction mass in the molten state; and
d) an agitator that rotates on its axis during operation, said agitator
comprising a plurality of elements that are longitudinally disposed to convey
a
portion of the melt as said elements move through the reaction mass, the
elements
being positioned such that said elements generate films, the planes of the
films
being parallel to the central axis of the agitator and the flow of inert gas
which is
predominantly in the axial direction; and
e) a reactor outlet for removing product polymer from the reactor
vessel.
' The agitator according to the present invention is different from agitators
used in conventional vacuum processes, which agitators consist essentially of
rotating disks or screens. Such prior art agitators generate films that are
perpendicular to the axis of the reaction vessel.
In a preferred embodiment of the present process, polymerization is
conducted at atmospheric pressure. A dihydroxy ester of an aromatic
dicarboxylic
acid, or of a low molecular weight polymerizable oligomer thereof, is
polymerized
to a product with a higher degree of polymerization (DP), preferably in the
presence of a polyester polymerization catalyst, wherein by-products of the
polymerization axe removed from the system by means of an inert gas. This
higher degree of polymerization is useful in bottles, fibers and films. This
process
provides an improved method for producing linear aromatic polyesters,
especially
polyethylene terephthalate) (PET). The aromatic dicarboxylic acid used in the
production of PET is terephthalic acid (TPA). The process may involve the
production of polyethylene terephthalate) from terephthalic acid and ethylene
2
- E SUBSTITUTE SHEET (RULE 26)
CA 02247239 1998-08-24
WO 97!35902 ~ ~ PCTIUS97/04487
glycol (EG) by esterification, or from dimethyl terephthalate (DMT) and
ethylene
glycol by a transesterification stage,. followed by polycondensation. The
process
is conducted at about atmospheric pressure or above, thereby avoiding high
vacuum equipment and eliminating possible air contamination that causes
product
decomposition and gel formation. First terephthalic acid is esterified or
dirnethyl
terephthalate is txansesterified with ethylene glycol to produce
bis(2-hydroxyethyl) terephthalate or its low molecular oligomers, which are
then
contacted in melt form with an inert gas. The volatile reaction by-products
are
removed with the inert gas, so that the polymerization is preferably complete
in
less than about 5 hours, more preferably less than 3 hours, of contact time
while
the reactants are kept at a suitable temperature to maintain them in the melt
form
so as to produce polyethylene terephthalate.
The above processes are preferably conducted in the presence of a
polyester polymerization catalyst. However, a catalyst is not needed for the
esterification step if the starting material is terephthalic acid. In a
preferred
embodiment of the invention, a single stream of inert gas is recycled through
a
polymer finishing stage, a polycondensation stage and a stage wherein ethylene
glycol is recovered for reuse in the process.
The invention is also directed to the novel apparatus described above for
carrying out polycondensation or other reaction in which a volatile by-product
is
removed by mass transfer from a melt to an inert-gas stream.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 represents a schematic drawing of one embodiment of an
apparatus that is suitable for carrying out the polymerization of the
invention,
wherein material having a lower degree of polymerization is converted to
material
having a high degree of polymerization.
Figure 2 represents a schematic drawing of one embodiment of a rotatable
agitator frame.
Figure 3 illustrates a rotatable agitator frame comprising an additional
inner concentric "cage" formed by another set of agitator elements.
Figures 4a, b, c, and d illustrate in isometric and cross-sectional views of
an agitator employing rectangular screens as agitator elements for the
generation
of film surface.
Figures Sa, b, and c illustrate side and cross-sectional views of an agitator
assembly consisting of concentric cylindrical wire cages.
Figure 6 illustrates concentric octagonal wire cages that can be employed
in an agitator assembly.
These figures are for the purpose of schematic illustration and are not
drawn to scale.
3
SUBSTITUTE SHEET (RULE 26)
CA 02247239 2005-03-23
These figures are for the purpose of scherr~atic illustration and are not
drawn to scale_
DETAILED DI;SCftIPTION OF THE IN~iON
Polyrnerizatian according to the present process can be carried out in one
~ , vessel, or more than one physically distinct vessel in series, wherein the
reaction,
mass is palyao~ttdensed to some degree of polymerization in one vessel and
then
transferred to another vessel for further palymtrization. The number of
vessels '
may depend on mechanical considerations related to handling ofthe polymeric
melt
as its viscosity increases with the degree of polym~:rization, heat input
requirements
to volatilize the by-products of the reaction, and cost. Preferably, a sitagle
vessel
may be employed to covert a prepolymer to a ftnal. product having the des'sred
degree of polymerization (DP).,
The process of the present invention may be carried out batchwise or
continuously. Satchwise production may be preferred for preparing specialty
IS polymers when the production required is not vety~ large and strict quality
control
is required particularly with respect to additives. her large scale production
for
commodity applications, such as molding resin, stf~ple and yarn, it is more
cost
effective to carry oat the above steps cominuousiy~ wherein the reactants are
fed
substantially continuously into the processing vessels and the products are
removed
substantially continuously. The rates of feed sad Irrodnct removal are
coordinated
to maintain a subst.antialIy steady quantity of the reactants in the reaction
vessels
while the inert gas flows countercurretttly to the flew of the melt.
rf two or more vessels are employed in satires for cor~duGting the
polyeondensation, it is preferred that a single stream of inert gas is
employed that
flows countercurrentty to the $ow of the melt in the process, i.e., the inert
gas
leaving a anal stage of polymerization is led throul;h the preceding stage and
fatally
through a stage wherein the ethylene glycol is recovered for reuse attd the
inert gas
is recycled back to the final stage of polymerization.
The preparation of higher molecular weight polyesters by melt
30~ pblymerization from polymerizable monomers and/or aligotners is a well
known ,
process, see for instance N. P. Cheremisino$ F.d_, Handbook of Polymer Science
and Technology, VoI. 1, Marcel lhkkeT, Inc., New York, 1989, pages 8T-90;
H. Mark, .et al., Ed., Encyclopedia of Polymer Science and Technology, 2nd
IFd.,
VoI. 12, John Wiley & Sons, New York, 19$8, pages 43-X16. 130-135 and
2I7-X25, ail of which znay be referred to herein, 'fhe conditions necessary
for these potyrnecizations, which are already generally knovyn to. the
artisan, arc
applicable to the poiymerizatians herein, modified as needed and as described
4
CA 02247239 1998-08-24
herein, to make various polyesters. These known features include items such as
process temperatures and polymerization catalysts (if any).
As an example, polyethylene tereph~halate (PET) is manufactured in this
process by first reacting terephthalic acid (TPA) or dimethyl terephthalate
(DMT)
with ethylene glycol (EG). If DMT is the starting material, a suitable
transesterification catalyst such as zinc or manganese acetate is used for the
reaction. In a preferred process, (trans)esterified DMT/TPA is polymerized as
a
melt at atmospheric pressure or above by contacting the melt with a stream of
inert
gas (for example, but not limited to, Nz or C02) to remove the condensation by-
products, mainly, ethylene glycol. Preferabl5~, the inert gas is preheated to
about
polymerization temperature or above, prior to its introduction into the
polymerization equipment. It is preferred that the inert gas velocity through
the
polymerization equipment be in the range o~ 0.06 to 0.92 m~sec (0.2 to 3
ft/sec),
most preferably 0.09 to 0.46 m/sec (0.3 to 1.5 ft/sec). The vapor leaving the
1 ~ polymerization (containing the ethylene glycol byproduct) may be treated
to
recover the ethylene glycol for recycle to the esterification stage or for
other uses.
The inert gas stream may be then cleaned up and recycled. Thus, the overall
process may operate as a closed loop system which avoids environmental
pollution
and integrates ethylene glycol purification and its recycle into the process.
''0 The quantit5~ of inert gas flow should be sufficient to cam the ethylene
glycol to be removed at a partial pressure of ethylene glycol below the
equilibrium
partial pressure of ethylene glycol with the reaction mass at the operating
temperature. The operating temperature during polycondensation is maintained
sufficiently high so as to keep the reaction mass in a molten state.
Preferably the
2~ temperature range is about 270°C to 300°C. The
pol5~merization equipment is
designed so that the interfacial area between the melt and the inert gas is at
least
1.9 m'- (20 square feet), preferably at least about 2.8 m' (30 square feet),
per
0.028 m' (cubic foot) of the melt and that this surface area is renewed
frequently.
Under these process conditions, the high degree of polymerization useful for
fibers
30 and films and other uses can usually be achieved in less than 5 hours of
residence ,
time, and preferably in less than 3 hours of residence time.
To more reliably produce good quality product of the desired degree of
polymerization, the polymerization should preferably be completed in a
reasonably
short period such as less than 5 hours, preferably less than about 3 hours. By
this
35 is meant the average residence time of the polymerizing mass in the process
is
preferably about S hours or less, more preferably about 3 hours or less. The
polymerization is considered completed when the degree of polymerization (DP)
desired for a particular application is achieved. For most common
applications,
such as fibers, the DP should be at least 50, preferably at least 60, and most
5
~~/J~rtn,r,n f~llp,::~
CA 02247239 1998-08-24
preferably at least 70. By "degree of polymenzation'' is meant the average
number
of repeat units in the polymer, for instance for polyethylene terephthalate),
the
average number of ethylene terephthalate units in a polymer molecule. Exposure
of the polymeric melt to high operating temperatures for prolonged period can
cause chain cleavage and decomposition reactions with the result that the
product
is discolored and/or a high degree of polymerization is not achieved. If the
inert
gas velocities are too lover, polymerization takes longer. If the velocity is
too high
it can lead to entrainment of the reaction mass in the gas. In a continuous
mode of
operating, high inert gas velocities in a countercurrent direction can also
hinder the
flow of the melt through the equipment. Also, higher velocities may require
larger
quantities of gas flow without substantially increasing the effectiveness of
polymerization.
The quantity of inert gas.flow employed to remove the ethylene glycol or
other volatile byproduct that evolves is sufficiently high so that the partial
pressure
1 ~ of ethylene glycol or other byproduct in the gas, at any point in the
process, is
belong, preferably well below, the equilibrium partial pressure of ethylene
glycol
with the melt at this point. Larger quantities of gas flow generally increase
the rate
of polymerization but the increase is not proportionatel5r greater. Therefore,
very
large amounts of gas are not usually necessary or desirable as large
quantities ma~~
increase the size of recycling equipment and the cost. Very large quantities
may
also require larger size polymerization equipment in order to keep the gas
velocity
in the desired range.
In the continuous embodiment of this invention, wherein the inert gas flows
countercurrently to the flow of the molten reaction mass, effective
polymerization
''S rates can be achieved ~~ith about 0.14-0.32 kg (0.3-0.7 pounds) of Nz per
pound of
the melt (equivalent to about ? to ~ moles of inert gas per mole of the
polymer
:. repeat unit) as long as the inert gas velocity is at least about 0.06 m/sec
(0.2 ft/sec),
preferably at least about 0.09 m/sec (0.3 ft/sec). The N~ flow, however,
should
preferably be at least 0.02 kg/l:g (0.2 lb/lb) of polymer (equivalent to 1.~
moles of
inert gas per mole of polymer repeat unit). Larger quantities of gas flow may
however be needed to obtain the preferred gas velocities.
In the process of this invention, the reactant is kept in a molten state,
i.e.,
above its melting point, which for instance is about 260-265°C for PET.
At
temperatures much above 300°C, decomposition reactions often cause
product
discoloration which interferes with the quality of the product. For PET, the
reaction mass should preferablv..be maintained at about 270°C to about
300°C.
For the polycondensation to continue, ethylene glycol or other volatile
byproduct generated must be removed from the reaction mass by the inert gas.
This removal is facilitated if there is a high interfacial area between the
melt and
6
AM~~sp ~p S~E~~
CA 02247239 2005-03-23
the gas please. To complete the polymerization in ~a reasonably short period,
the
surface area should be at least about 1.9 m2!0.028 m3 (20 ft~lft3) of the
melt,
preferably at least about 2.8 m210.028 m3 (30 ft21i~3) of the melt. A higher
surface
area is preferred to increase the fate of polymeriz<<tion. The reaction
equipment for
S contacting the melt and the inert gas should also >:~e designed to
frequently~renew
the interfacial area and mix the polymer melt. This is particularly important
as the
degree'of polymerization increases and the melt becomes more viscous.
The rate of polymerization can also be increased by using a suitable
polymerization catalyst, particularly where a high interfacial area is
provided for
inert gas - melt contact. The increase in the overall rate, however, is not
proportional to the concentration of catalyst as the: removal ofethylene
glycol
starts to limit the overall polymerization rate.
The catalyst may also increase the rates of decomposition reactions. An
effective concentration of catalyst for a set of reaction conditions, such as
temperature, gas flow, velocity and surface area, e s such that it gives the
most
enhancement in the rate of polymerization without substantial decomposition.
The
optimum concentration of catalysts of various species are known in the art, or
caa
be determined by experimentation. It would generally be in the range of a ftw
parts per million parts of the polymer, such as about 5-300 parts per million.
Catalysts far facilitating the polymerization are any one or more polyester
polymerization catalysts known in the prior art to catalyze such
polymerization
processes, such as, but not lznnited to, compounds of antimony, germanium and,
titanium. Antimony trioxide (Sb20g) is an especi~~lly effective catalyst which
may
be introduced, for convenience, as a glycolate solution in ethylene glycol.
Examples of such catalysts are found in U.S. 2,57,3,660, U.S. 2,647,885 and
U.S.
2,789,772, may be referred to herein.
Polymers which can be produced by the present process include those
derived from one or more aromatic dicarboxylic aphids and one or more
aliphatic or
cycloaliphatic glycols. By an aromatic dicarboxylic acid is rrteant a
dicarboxylic
acid in which the two carboxyl groups are each bound directly to a carbon atom
of
an aromatic ring. The aromatic dicarboxylic acid may otherwise be substituted
with orte or more other groups which do not interfere with the polymerization,
such as alkyl groups, chloro groups, alkoxy groups, etc. Examples of useful
aromatic dicarboxylic acids include terephthalic a~:id, isophthalic acid and
2,6-naphthalene dicarboxylic acid.
CA 02247239 1998-08-24
WO 97/35902 ' PCT/US97/04487
ethylene glycol, 1,4-butanediol and 1,3-propanediol, and ethylene is
especially
preferred. .
Preferred combinations of aromatic dicarboxylic acids and glycols are (the
polymer produced is in brackets) terephthalic acid and ethylene glycol
[poly(ethylene terephthalate)], terephthalic acid and 1,3-propanediol ,
[poly(1,3-propylene terephthaiate)], terephthalic acid and 1,4-butanediol
[poly(1,4-butylene terephthalate)], 2,6-naphthalene dicarboxylic acid and
ethylene
glycol [poly(ethylene 2,6-napthoate)], and a combination of terephthalic and
isophthaiic acids and ethylene glycol [copoly(ethylene
isophthalate/terephthalate)]. Polyethylene terephthalate) and/or
poly(I,3-propylene terephthalate) are especially preferred products. The
polymers
can be made by polymerization of various polymerizable ethers and/or oligomers
described herein.
Dihydroxy esters of various aromatic dicarboxylic acids may be used in the
processes described herein. These are monomeric compounds that can polymerize
to a polymer. Examples of such compounds are bis(2-hydroxyethyl)
terephthalate, bis(3-hydroxypropyl)terephthalate, bis(4-hydroxybutyl)
terephthalate, bis(2-hydroxyethyl) napthoate, bis(2-hydroxyethyl)
isophthalate,
bis[2-{2-hydroxyethoxy)ethyl] terephthalate, bis[2-(2-hydroxyethoxy)ethyl]
isophthalate, bis[(4-hydroxyrnethylcyclohexyl)rnethyl] terephthalate,
bis[(4-hydroxymethylcyclohexyl)methyl] isophthalate, and a combination of
bis(4-hydroxybutyl) terephthalate and their oligomers. Mixtures of these
monomers and oligomers may also be used to produce copolymers.
By a "polymerizable oligomer" is meant any oligomeric material which
can polymerize to a polyester. This oligomer may contain low molecular weight
polyester, and varying amounts of monomer. For example, the reaction of
dimethyl terephthalate or terephthalic acid with ethylene glycol, when carried
out
to remove methyl ester or carboxylic groups usually yields a mixture of
bis(2-hydroxyethyl) terephthalate, low molecular weight polymers (oligomers)
of
bis(2-hydroxyethyl) terephthalate and oligomers of mono(2-hydroxyethyl)
terephthalate (which contains carbonyl groups). This type of material is
referred
to herein as "polymerizable oligomer".
The process may be used to produce various polyesters such as
polyethylene terephthalate), polypropylene terephthalate}, poly(1,4-butylene
terephthalate), poly{ethylene napthoate), polyethylene isophthalate), poly(3-
oxa-
1,5-pentadiyl terephthalate), poly(3-oxa-1,5-pentadiyl isophthalate),
poly[1,4-bis(oxymethyl)cyclohexyl terephthalate] and poly[1,4-bis(oxymethyl)-
cyclohexyl isophthalate]. Poly{ethylene terephthalate) is an especially
important
commercial product.
8
SUBSTITUTE SHEET (RULE 26}
CA 02247239 1998-08-24
The process may be used to produce v,xrious polycateis such as
polyethylene terephthalate), polypropylene terephthalate), poly(1,4-butylene
terephthalate), polyethylene napthoate), polyethylene isophthalate), poly(3-
oxa-
1,5-pentadiyl terephthalate), poly(3-oxa-1,5-nentadiyl isophthalate),
poly[1,4-bis(oxymethyl)cyclohexyl terephthalate] and poly[I,4-bis(oxymethyl)-
cyclohexy~1 isophthalate]. Polyethylene terephthalate) is an especially
important
commercial product.
The process avoids high vacuum polymerization processes characteristic of
the conventional art. Advantages of the process often are a simpler flo~~
pattern.
I O and/or lower operating costs and/or the avoidance of steam jets, hot wells
and
atmosphere emissions. The process also has environmental advantages due to the
elimination of volatile organic emissions and waste water discharge.
Furthermore,
polymerization is conducted in an inert environment. Therefore, there is often
less
decomposition and gelLformation which results in better product quality.
Ethylene
_ 1 ~ glycol and inert gas (e.g., N~ or CO~) may be recycled continuously.
In a preferred embodiment of the process for making PET, an oligomer
exiting the esterifier is prepolymerized to a degree of polymerization (DP) of
about
1.5-30 and this prepolymer is fed to a finisher in order to polymerize it
further to a
higher DP of between about ~0 and I ~0, preferably about 60 to about 1?0 and
20 more preferably about 70 to about 90. The f nisher is maintained at a
temperature
greater than about 260°C but not too high to cause polymer
decomposition. A
temperature range of about '_'70°C to 300°C is preferred. The
polymerization
product is continuously removed from the finisher. An inert gas, preferably
nitrogen, is heated in a heater to a temperature of from about 280°C to
320°C and
is introduced into the finisher to flow countercurrent to the direction of
polymer
flow in order to remove volatile reaction by-products, primarily ethylene
glycol.
Preferably; the nitrogen is employed in a closed loop and all processing
equipment
for cleaning and recycling the nitrogen is operated at atmospheric pressure
(or
above, as is necessary to ensure the flow of nitrogen through the equipment in
the
30 loop). The quantity of inert gas introduced into the system is sufficient
so that the
partial pressure of the by-products is maintained below the equilibrium
pressure of
the by-products with the melt in order to provide for the continuous
polymerization. Theyuantity of inert gas may be as small as about 0.14-0.32 kg
-
(0.3-0.7 pounds) for each kg (pound) of PET produced.
35 FIG. 1 illustrates one embodiment of a reactor or finisher that is suitable
for
carrying out the polymerization of the invention, especially for producing
high
viscosity polymers having a degree of polymerization encountered in a
finisher. ~.
The reactor comprises a horizontal, agitated cylindrical reaction vessel 1.
The
reactor housing 2 is conveniently constructed with a cylindrical body (shell)
and
9 ,
~,r~,,,~,- J ~h~~'-ET
t~...
CA 02247239 1998-08-24
end plates 4 and 6 that close off the ends of the cylindrical body. A reactor
jacket 8 through which a heat transfer material is passed surrounds the
cylindrical
body. An exemplary heat'-transfer material is Dowtherm~ heat transfer fluid,
commercially available from Dow Chemical (Michigan). Other methods of
heating known in the art may be used, such as hot oil heat, high pressure
steam or
electrical heating. A reactor inlet 14 for introducing a prepolymer feed into
the
reactor is showJn at one end of the reactor, a reactor outlet 16 for
discharging
product from the reactor vessel is shown at the opposite end of the reactor.
The esterified DMT or TPA, or low molecular weight oligomers or
prepolymers thereof, is continuously introduced as stream 3 at one end of the
reaction vessel. A preheated inert gas, such as nitrogen, is continuously
introduced
as stream 7 at the other end, so as to provide flow countercurrent to the
polymer
floe. The nitrogen stream.9 carrying reaction by-product vapors, mostly
ethylenqe
glycol, leaves the reaction vessel as stream 11. The reaction mass flows as
the
~ . 1 ~ polymer melt stream 5. The polymerized product, polyethylene
terephthalate, is
removed as stream 1 ~. The flow rates of streams 3 and 1 ~ are coordinated to
be
equivalent to each other and controlled so as to provide the desired hold up
of the
melt in the finisher, usually about 1 to 3 hours, ~~hich is equivalent to a
melt level
at about 1!4 to 1!3 of the diameter of the vessel. The quantity of nitrogen
introduced into the system is sufficient so that the partial pressure of the
evolving
reaction by-products is maintained at less than the equilibrium pressure of
the by-
products in the, for example, polyethylene) terephthalate (PET) melt. so as to
provide adequate driving force to remove ethylene glycol from the melt into
the
has stream. The diameter of the vessel is designed so that the superficial
velocity
of the inert gas stream is in the desired range.
In one embodiment of the process, use of Dow-therm~ heat transfer fluid or
other heating means is eliminated by employing the preheated nitrogen stream
itself for heating. In this embodiment the nitrogen stream is first led
through the
heating jacket 8 in Fig. 1 to maintain the reactor wall above the melting
point of
the reaction mass, and is then fed as stream 7 to the reaction vessel.
The reaction vessel in Fig. 1 is equipped with an agitator 20 attached via
drive shaft 18 to a drive 22 so that the agitator can be rotated at a
controlled speed.
The mechanical design of the agitator is such that
(a) the walls of the vessel are wiped;
(b) a large interfacial area of at least 1.9 m2/0.028 m3 (20 ft2/ft3) of
the melt preferably greater then 2.8 m2/0.028 m3 (30 ft2/ft3) of the melt is
created:
(c) the surface area is renewed frequently; and
_,,,;.
-~;
~_ ru ~-v W'
.J ..
CA 02247239 1998-08-24
WO 97/35902 PCTIUS97104487
predominantly axial along the longitudinal central axis of the reaction
vessel. In
one embodiment, shown in Figure.2, the agitator 20 comprises paddles 21
attached to rotatable ends 23 and 24 that rotate during use, to form a
rotatable
frame.
S A central axle attached to the agitator in the reactor vessel may extend
outside the housing of the reactor vessel where it is attached to a motor or
drive
for providing rotation of the agitator at a suitable rate.
The frame may comprise at least two sets of a plurality of arms that
radially extent from the longitudinal axis of the reaction vessel. Each pair
of arms
can support a wiper, which is suitable as an elongated paddle. The edge of the
paddle may be beveled to better wipe the internal surface of the reaction
vessel.
The wipers, or wiper blades, may be set at a suitable angle to move a suitable
amount of melt as they move through the melt pool, so shedding of the melt for
generating films can last through most of the rotation outside the pool. If
the
angle is such that the space between the wiper and the cylindrical wall is too
narrow, the wipers will carry only a small amount of melt which may become
quickly depleted by running through the clearance between the wipers and the
cylindrical wall, and not enough left to generate films or be wiped on the
inside
cylindrical wall. If the angle is too large, on the other hand, more melt
unnecessarily will be carried around.
The number of wiper blades and the number of arms attached to them at a
point along the length of the reaction vessel may vary. Large diameter vessels
would generally have more wipers. Also, there may be more wipers near the feed
end, where the melt is less viscous, and less near the product end where the
melt is
very viscous. The wipers may , for example, be 2 to 32 in number, preferably 4
to
12 in number.
The wiper-frame assembly is of mechanically strong construction to
withstand the torque required to move through the viscous polymer melt and
carry
it. In one embodiment, cross rods are attached between the wiper blades for
mechanical reinforcement.
As the wipers or paddles move out of the pool of reaction mass, they shed
the polymer melt as films that last for a short distance as the surface
tension starts
to gradually pull the melt film together into thicker streams that have much
less
surface area. It has been observed that the films last for about 1/2 inch when
the
DP is about 30-40, about 1 inch at about 50 DP and about 2 inch at 60-80 DP.
Therefore, to maximize the surface axea, additional longitudinal elements are
placed under the wipers, at suitable distances, over which the melt can fall
and
continue to shed as films. It is advantageous to maintain the spacing between
the
elements narrow near the feed end where the melt is quite fluid and easily
spreads
11
' - . Y_ . ,.,~ ~+ A'~.,.~: ~~,~~,~,.~- SUBSTITUTE SHEET {RULE 26)
CA 02247239 1998-08-24
surface area. It has been observed that the films last for about 1.3 cm (1/2
inch)
when the DP is about 30-40, about 2.5 cm (1 inch) at about 50 DP and about
5.1 cm (2 inch) at 60-80 DP. Therefore, to maximize the surface area.
additional
longitudinal elements are placed under the wipers, at suitable distances, over
which
the melt can fall and continue to shed as films. It is advantageous to
maintain the
spacing between the elements narrow near the feed end where the melt is quite
fluid and easily spreads into thin films, and to increase the spacing towards
the
product end where the melt is very viscous and flows as thick films. If the
spacing
is too narrows, the ~~iscous melt would stagnate between the elements and not
generate the desired surface area. Thus, the spacing may be as small as 1.3 cm
1 /2 inch) near the feed end and ~ . l -10.2 cm (2-4 inch) near the product
end.
Spacing can be optimized for a given diameter reaction vessel and speed of
rotation. The lonDitudinal elements may be rectangular bars, rods, wires,
meshed
screen or sheets of metal punched out or cut to form grids of desired spacing.
1 ~ These may be arranged to form a "cage" or a plurality of concentric
"cages" as
shown in Figure 3.
Alternatively, as sho~m in Figure 4, the elements may be arranged in a
rectangular geometry, these rectangles being parallel to each other and
extending
longitudinally, again keeping the spacing larger at the viscous end and
smaller at
the less ~~iscous end. The agitator may be thus built in sections that appear
like a
"stack" or a "sand~Tich" of rectangular assemblies. These sections may be
installed
in the agitator frame staggered, e.g., the plane of one section may be
perpendicular
to those of the next section to as to keep the inert gas well distributed and
to
minimize by-passing (running through) of the melt by making the path more
torturous.
In Figure 4, the elements are meshed screens, but these could be of other
configurations such as rods or punched sheets of metal. In this type of
agitator, the
melt picked up by the wipers during their travel through the pool at the
bottom,
and thereafter shed by the wipers, flows along the rectangular elements to
generate
surface area.
The agitator is rotated at a rate (rpm) that maximizes the generation of
surface area and_provides frequent surface renewal. Faster surface renewal is
advantageous for increasing the coefficient of transfer of volatile products
from the
reaction melt to inert gas but rotation that is too fast can result in the
viscous
polymer melt being held as "globs" between the elements and, in fact, decrease
the
surface renewal rite. For attaining a reasonably good transfer coefficient it
is
preferred that the surface be renewed at least once per minute. The agitator
speed
is also important to surface area generation. If the rotation is too slow,
su~cient
melt is not lifted from the pool, or is shed too early, and all the elements
do not
12
AMENDED SHEET
CA 02247239 1998-08-24
generate films. If the rotation is too fast the melt may be caught up as
"globs" and
does not flow effectively to generate surface area. The rate of transfer of
the
volatile by-products, and hence the rate at which the polymer DP increases is
proportional to both the transfer coefficient (k) and the surface area (a).
The rate
of rotation, or revolutions per minute (rpm) for a given agitator geometry and
vessel diameter may be optimized to maximize the product k x a. Preferably,
the
agitator is rotated at about 1 to 60 rpm, more preferably at about 1-30 rpm
and
most preferably at about 2-18 rpm.
To illustrate a "cage" type construction in detail. one embodiment of an
agitator is shown in Figure ~ in v~~hich the elements are wires. the
circumferential
spacing of m~hich varies along the length of the reactor vessel. The spacing
is
narrower at the feed end and wider at the discharge end. The agitator is
divided
into sections. and a pluralit5r of concentric "cages" can exist in each
section. the
number of v~~hich may vary from section to section.
1 ~ Surface area in this type of configuration is generated in two wa5~s,
first by
filming of the melt circumferentially over the "cafe" and, second. by drainage
of
the melt from the elements of one "cage" down to a smaller diameter "cage"
belong. The spacing and rpm are optimized so as to obtain good
circumfereritial
coverage and drainage at all points along the length of the agitator. The
carwin' of
melt "lobs" is minimized as discussed earlier. At the preferred ?-1'' rpm. the
spacing near the feed end may be as narrow as 1.3 cm (l; 2 inch) and. near the
product end. it may be ~.I-7.6 cm (2-3 inches). Thus. it is preferable to have
more
concentric "cages" near the feed end and less at the discharge end. The
surface
area generated per unit length is. therefore. greater near the feed end and
decreases
along the length towards the product end as the number of "cages" decreases.
To
compensate for this, sections of larger spacing can be made proportionately
longer.
In this manner. the surface generated at each spacing. and hence the increase
~in DP
at each spacing, is about the same.
The surface area created in the reactor equals the sum of (A) the wiped
surface on the inside v~~all of the reactor, (B) the surface area of the melt
pool. (C)
the surface area of the agitator elements and those of the melt films
generated as
the agitator rotates. The area of the film is to be multiplied by 2 to account
for the
surface area available for mass transfer from both the sides of the films.
As the reactor size increases, contributions to the surface area from (A) and
(B) decreases in relation to that from (C). Thus, for large, commercial scale
finishers. most of the surface area is from the films generated by the
agitator
elements and the area due to (A) and (B) may be neglected for design purposes.
For example, a 2.1 m (7 ft.) diameter X 8.9 m (29 ft.) long reactor, designed
to
generate
13 ~ ~~cf-(
pMENO._'~ ,.
CA 02247239 1998-08-24
15,000 square ft. of surface area, the contribution from (A) and (B) is less
than
4%.
In calculating the surface area, that could be generated ~~i'th an agitator
assembly being considered, it is first assumed that an optimum combination of
agitator RPM and element spacing is selected to maximize films generation,
e.g., in
the screens and v~lire "cage" type agitators, the screens and circumferential
area of
the "cages" are completely covered with melt. The film surface area is twice
the
covered area to account for the two sides of the films. Preferably, the
reactor is
designed for a higher area to compensate for less than complete coverage
during
operation under sub-optimal conditions.
The overall agitator for the reactor is conveniently built in sections or
"spool pieces" that may be fastened together by suitable means. Fabricating
the
agitator in spool pieces offers the flexibility of providing different
spacings or
other variations depending on the particular application or conditions of use.
1~ Such sectionalized fabrication of the agitator also allows the insertion of
baffles, for example discs and donuts which contribute to the distribution of
inert
gas and improves contact between inert gas and the reaction mass. This also
compartmentalizes the reactor longitudinally so that c~~hen it is operated
continuously- it acts like a number of reactors in series and the performance
?0 approaches that of a plug flow or a batch reactor.
The length and spacing of each Section can be conveniently determined by
the following equations in which L is the total length of the agitator. I~T is
the
number of sections desired. The length of the first section (at the feed end)
is
given by the following equation:
~_ 1 N
- Ll =L/CN+N-11 ~ (n-I)
.. n=I
adhere ~= number of folds increase in the DP which is equal to DP of
product/DP of the feed.
30 For subsequent sections, the nth section length is preferably defined as
follows:
Ln=pl . L1
35 ,wherein pn is the pitch or spacing of the «Aires in the nth ~:ection and
pI is
the spacing in the first section. The parameter pn is related to p1 by the
following
equation.
14
~\1L1~~ ~'.,r'T
~(~II~~ _,._
CA 02247239 1998-08-24
pn (X-1 )(n-1 )
p1-1+ N_1
For concentric "cages" in a given section. the spacing between the
consecutive cages equals the pitch.
The length of each section as calculated above may be rounded to a
convenient figure for fabrication. such that:
Ll +L?+L3 .... LN=L
The wires selected for this construction are of suitable gauge and have
adequate mechanical strength to withstand the shear stresses of the viscous
polymer melt. The wires may be 1.6 mm (1/16") diameter near the feed end and
of
thicker gauge, for example. 4.8 mm (3/16") diameter, near the viscous product
end.
_. l~ Cross wires may be welded circumferentially at suitable distances, e.g.,
3 to
~ times the pitch or wire spacing, for mechanical strength.
For ease of fabrication. a long rectangular wire mat of desired pitch and
cross-wires distance may' be first constructed and then rolled into a
"spiral". instead
of constructing individual "cafes," while keeping the separation between
'0 consecutive winding of the spiral about the same as the distance between
consecutive concentric "cages." i.e., about equal to the wire spacing.
The "cages" need not be necessarily cylindrical. For ease of fabrication.
these may be of geometries such as hexagonal, octagonal, etc. Figure 6 shows
an
octagonal assembly of z~~ire cages as viewed from the end of the agitator.
?, Rectangular sections of wire mats 30 are attached to the radial arms 33 of
a
rotatable end. Such geometries allow the ~~ire mats to be cut or made in
rectangular sections that can be welded to the radial arms.
The reaction vessel and the agitator is constructed from a suitable material
of construction having the adequate mechanical strength at the operating
30 temperature and which material, in order to produce a quality product, is
not easily
corroded or reactive with the reaction mass so as to contaminate the product.
Stainless steel is one suitable material having the requisite properties.
The surface area needed to achieve a given degree of polymerization (DP)
can be estimated, as a first approximation, by using the follov~~ing simple
equation
35 which has been found to hold when polymerization is conducted under batch
or
plug flow conditions and a large quantity of inert gas is employed:
' DP-DP°=kat
' ~ P.~y~~sJ~~~J
CA 02247239 1998-08-24
In this equation:
DP = the desired product DP
' DP° = DP of the feed prepolymer or oligomer
a = surface area in square feet
t = residence time or hold up time in hours
k = overall transfer coefficient for transfer of the volatile
condensation by-products, mostly ethylene glycol. from the
melt to the insert gas. The units are ft/hr.
The transfer coefficient, k, depends upon several factors, such as
temperature. surface rene~~al rate, catalyst concentration and inert gas
velocity.
Under the conditions of Example 1, its value alas found to be about 0.24 m/hr
(0.79 ft/hr).
Thus. for polymerizing a prepol5~mer of 20 DPyto a product of 80 DP in
2 hours of residence time. the surface area required, using this value for k,
can be
1 ~ calculated as:
_ DP - DP°
a kt
_ 80-20 _
0.79 x '? -''~ m~/0.0?8 m' (38 ft'-/ft') of melt
For continuous polymerization, the reactor is preferably designed to
prop°ide a larger surface area. such as 4.7-7.0 m~/0.0?8 m' (~0-75
ft'/ft~) of melt
for the above example, to compensate for using less inert gas floe, e.g.,
0. I4-0.3? kg (0.3-0.7 Ib). N~/lb. of melt. and for deviations of the melt
floe from
the ideal plug flow. The higher than calculated surface area also permits
operating
flexibility._ If the reactor has less area, the hold up time would need to be
proportionately longer than 2 hours. The agitator configurations described
herein
can provide the required high surface areas.
For running the polymerization reaction continuousl5r, it is desirable that
the residence time distribution of the melt flow be narrow, i.e., it is closer
to plug
flow, and by-passing is prevented. By-passing can potentially occur around the
~, straight paddles and agitator elements, particularly when the melt is not
highly
viscous.
The reactor may also be divided longitudinally into a number of
compartments by introducing baffles such that melt flows from one compartment
' to the next and the reactor thus performs like several smaller reactors in
series.
One convenient way to achieve this is to insert along the length of the
agitator
rings or donuts with an outside diameter equal to that of the agitator. The
inside
16 _
f,~':h'a''~ ~r' ~;Lli vT
--t~~'v:.:r ._ _L t
CA 02247239 1998-08-24
.
diameter of the donuts is such that the reactor operatC~ at tim desired level.
The
inside diameter may be about 0.7 times the outside diameter. Disks may also be
inserted in between the donuts to form a donut-disk-donut~pattern, to keep the
inert
gas flow well distributed and improve contact with melt by forcing it to go
through
the donut and then around the disk, and so on. The baffles are sized such that
the
velocity of gas through, around or between them is not too high to cause
entrainment or push melt in the direction of the inert gas flow.
Similarly, another embodiment of the agitator comprises partial disks or
partial rings installed such that the inside edges are staggered at
I80°, i.e.,
alternate baffles face in opposite directions, so that inert gas will zigzag
as well as
swirl, creating greater turbulence and more effective contact with the melt,
as these
are rotated.
The process of this invention may also be carried out for batchwise
preparation of polyester wherein a batch of low molecular weight oligomer is
1 ~ charged to the polymerization equipment and contacted v<~ith the inert gas
as
described until the desired high degree of polymerization is achieved. The
oligomer is prepared by esterification as described except that it may also be
prepared batchwise either in a separate vessel or in the poh~merization vessel
itself.
The gas and melt contacting equipment may be similar to that described for the
~0 continuous embodiment of this invention except that it is not necessary to
vary the
spacing between the agitator elements along the length of the vessel. Also,
compartmentalization to approach plug flov~~ is not required. The spacing of
a;itator elements should be chosen to accommodate the viscosity and flow
characteristics of the final high molecular weight product. For batchwise
2~ preparation it is advantageous to adjust the speed of the agitator as the
viscosity of
the melt increases. Initially, when the viscosity is lo~~, the agitator may
operate at
as high as ~l 00 rpm but toward the completion of polymerization a love speed
of
about 1 to 20 rpm, preferably about 2-12 rpm is desirable. Batchwise
production is
suitable for economic reasons when relatively small quantities of polyester
are to
30 be prepared or when a strict control of additives concentrations is
required for ,
product quality considerations. When the quantities to be prepared are very
small,
it may be more economical to not provide equipment for recycling the inert
gas, or
the ethylene glycol, and discharge it to the atmosphere after rendering it
harmless
to the environment by known methods such as scrubbing it thoroughly with water
35 and disposing off the water in an environmentally safe manner.
The invention can also be conducted in a semi-bath fashion wherein the
polymerization equipment is fed intermittently, reaction mass is polymerized
to a
higher degree, and the product is discharged intermittently.
17 ~ ., .;: ~ '=rii=G'
r, . ~~y~~.:_
CA 02247239 1998-08-24
EXAL~IPLIr 1
This examples illustrates polymerization on a pilot scale in a
polymerization reactor according to the present invention. The reactor
consisted of
a nominal 1 ~.2 cm (6 inch) diameter glass tube of 61 cm (2 foot) length. It
was
held inside an 20.3 cm (8 inch) diameter glass tube of similar length with the
help
of end plates so as to form an even annular space around the reactor and
served as
the heating jacket. The heating medium was air heated to 29~-300°C
~~hich was
introduced into the annular space at one end and flowed out from the other
end.
The agitator consisted of two end pieces each ~~ith four arms in the shape
of a cross. Each pair of arms held an approximately 50.8 cm (20") long, ~.~ cm
( 1 ") wide paddle or a wriper. Two rings were mounted inside this frame, each
a
few inches inside from the ends to hold four more such blades, such that the 8
blades formed a "cage" of slightly smaller diameter than the 1 ~.2 cm (6")
diameter
of the reactor so it could be freely rotated inside the reactor. Shafts were
attached
16 to the two cross end pieces which could be rotated inside bearings provided
in the
center of each end plate of the reactor. The agitator was rotated by use of a
motor
having a variable speed gear reducer attached to the shaft on one end of the
reactor.
The temperature of the polymer melt and the inert gas was monitored by placing
thermocouples inserted into the reactor from each of its two ends.
The reactor c~~as charged with 41 kg (9 pounds) of a prepolymer of about
20 DP obtained from a commercial plant where it was made by esterifying TPA
with ethylene glycol and prepoltrmerizing it to a DP of about 20. It contained
about 200 ppm antimony as catalyst. The charging was done b}~ feeding the
solid
prepolymer through a melt extruder which melted the prepolymer and heated it
to
'?~ about 280-29~°C. The agitator was rotated at 12 rpm. and N?
preheated to about
29~°C was flowed through the reactor at a velocity of 0.17 m/sec (0.a7
ft/sec)
based on an empty cross-section of the reactor. Since the reactor was about
30%
filled with melt the contact velocity was about 0.28 m/sec (0.82 ft/sec). The
Nz
was introduced at one end and was discharged to the atmosphere from the other
end. Thus, the reactor was essentially at atmospheric pressure. The
temperature of
the reaction mass was maintained at about 280°C by controlling the
temperature of
the hot air in the annulus. Polymerization was continued under these
conditions
for two hours. Samples of the polymer were taken every half hour and analyzed
for DP by gel permeation chromatography (GPC). The number average DP was
found to be approximately 36, 52, 68, and 80, after 1/2, l, 1-1/2 and 2 hours
of
polymerization, respectively. These DP values when plotted against time fit a
straight line:
DP-DP°=(ka)t
with a slope = k a of 30 hr-1.
18
AMENDED SHEET
CA 02247239 1998-08-24
The reactor was estimated to provide cn the werab~ C.43 m2 (4.58 ft~) of
film area which for 4.1 kg (9 1b.) of the melt translates to a value of "a"
equal to
3.5 m2/0.028 m3 (38 ft'/ft3) of melt. The value of k was thus 30/38 = 0.24
m/hr
(0.79 ft/hr).
Initially, when the melt was at 20 DP, it was shed from the agitator blades
as streamlets but after a few minutes it started becoming viscous and falling
as
films that extended about 0.6-1.3 cm (1/4-1/2") from the edges of the blades.
As
the polymerization proceeded to higher DP's the filming became more
pronounced.
The melt extended as films 1.9-2.~ cm (3/4-1 ") from the blades and towards
the
end the shed films extended 2.~-3.8 em (1-1-1/2"). Thus, larger surface area
could
have been generated if additional elements had been placed in the agitator,
under
the blades, over which the melt could fall and drop further as films. Also,
instead
of using hot air in the annular heating jacket, preheated N~ could have been
first
passed through the annulus and then fed to the reactor.
EXAMPLE 2
This example illustrates a design of a prototype finisher according to the
present invention to be operated continuously at a rate of 4~ to 68 kg/hr (
100 to
' 1 ~G Ib.W r). It will be continuousl~~ fed with a prepolymer of 20 DP
prepared 1I2 all
upstream esterifier and a prepolymerizer. The reactor is designed to produce
?0 product PET of about 80 DP useful for spinning into fibers or producing
flakes.
The reactor is 2.8 m (9 ft) long and has a diameter of 46 cm ( 18 inches). It
has a
heating jacket heated with Dowtherm~ vapor. It is fitted with a 2.3 m (7.~-ft)
long
agitator to leave about ?3 cm (9 inches) of space on each end for feed and
discharge nozzles. The agitator has end pieces or plates that each extend to 8
arms
which are attached to the shafts for rotation. To each pair of arms is
attached a
3.8 cm (1-1/2") wide wiper blade positioned at a 4~° angle to the
inside wall of the
reactor. Held inside this eight wiper frame are spools of concentric "cages"
of
varying pitches fabricated from stainless steel wires such that, starting from
the
feed end, there is a 23 cm (9 inch) long section of 1.3 cm (1/2") pitch (and
spacing
30 between the consecutive concentric "canes"), then 46 cm (18") length of
"cages" of
2.5 cm ( 1 ") pitch, followed by 6.9 cm (27") length "cages" that are at 3.8
cm
(1-1/2") pitch and finally 0.92 m (36") length concentric cages of 5.1 cm (2")
pitch
adhere the concentric "cages" are ~.l cm (2") a~urt. Donuts and disks are
inserted
alternatively between the spool pieces so the reactor is compartmentalized to
act as
35 4 reactors in series. The agitator can be rotated at 3-12 rpm. Spools of
each of the
1.3 cm, 2.5 cm, 3.8 cm and 5.1 cm (1/2", 1", 1-1/2" and 2") pitches can
provide
,-
about 5.3 m2 (57 ft2) of surface area for a total of 21.2 m2 (228 ft2) of
surface area.
The reactor is operated with about 136 kg (300 1b.) or 0.11 m3 (4 ft3) of melt
hold
up. This translates to an average surface area of 5.3 m2/0.028 m3 (57 ft2/ft3)
of the
19 '
AMFND~~ SH~E'T
CA 02247239 1998-08-24
melt which is about 50% more than what would be required if it perW rmed as an
ideal plug flow reactor. N~ is flowed countercurrently to the flow of the melt
at 41
'' to 54 kg/hr (90 to 120 lb./hr). The superficial gas velocity under the
operating '
condition of about 98 kPa (1 atmosphere) pressure and 285°C, based on
an empty
cross-section, is 0.11 to 0.1 ~ m/sec (0.36 to 0.48 ft/sec).
.. ~ ,
19A
AMEN' ~ ~h4ET
