Lloyd L. Handbook of Industrial Catalysts

Подождите немного. Документ загружается.

328 Chapter 8

the addition of diethylboron ethoxide in the concentration range 2–10 ppm as

co-catalyst. When a similar catalyst is used with the addition of triethylboron in

the same range of concentrations, the density of the polymer decreases. This

suggests that the presence of triethylboron leads to the formation of some α-

olefins and that these are incorporated into the HDPE to produce a polymer with

a lower molecular weight, more analogous to LLDPE.

It has been possible to introduce many different polyethylene varieties with

Phillips catalysts by modifying the silica surface or using co-catalysts, which

will continue to allow a much wider use of chromium catalysts in the future.

8.3.6. Organo-Chromium Catalysts

Catalysts derived from organo-chromium compounds supported on silica gel are

very different to the original Phillips chromium oxide catalyst. In the original

catalysts one molecule of chromium oxide is bound to two silica units, to form

the silyloxy chromate sites. In contrast, organo-chromium compounds are usual-

ly only bound to a single site on the silica support and the other coordination

sites are still occupied by the organic ligands. For example, tris-π-allyl chromi-

um reacts with a hydroxylated support, with the evolution of propylene and the

formation of a bis-π-allyl chromium siloxy derivative.

The support for an organo-chromium catalyst must, therefore, be carefully

calcined to lower the concentration of surface hydroxyl groups, leaving a suffi-

ciently wide separation to avoid the possible linkage of the chromium to two

units of silica.

Excessive calcination has to be avoided as it removes too much

water, leaving too few hydroxyl groups to achieve the required loading of chro-

mium.

A new catalyst in which chromocene was supported on silica gel was de-

veloped by Union Carbide

for use in gas-phase polymerization reactions. A

typical catalyst contained about 2.5% by weight of chromium, and was activated

by the addition of small amounts of alkylaluminium alkoxides, such as diethyl-

aluminium ethoxide. Use of such a catalyst formulation gave rise to a polymer

with much narrower molecular weight distributions. The average molecular

weight could be controlled by the addition of hydrogen, which caused termina-

tion of the growing chain.

Other organo-chromium catalysts include those produced by the impregna-

tion of silica supports with diarene chromium species, chromium allyls, and

bis(triphenyl) silyl chromate. These catalysts give very broad molecular weight

distributions of the polymer. Branching in the polymer chains is often extensive,

because low-molecular weight olefins are formed and can act as co-monomers.

Polymer chain termination is usually by ß-hydrogen abstraction, leaving a ter-

minal double bond in the polymer chain.

Mixed catalysts have been prepared by reaction of an activated chromium

oxide silica catalyst with an organo-chromium compound. Mixed catalysts are

Olefin Polymerization Catalysts 329

very active and produce branched polymers with a low density. The type of

organochromium compound used in the catalyst controls the formation and

isomerization of the olefins incorporated into the polymer chains,

so influenc-

ing polymer density.

8.4. OTHER CATALYSTS

Standard Oil of Indiana used a molybdenum oxide/alumina catalyst that could

be prepared either by co-precipitation or impregnation methods.

The catalyst

was activated by calcination in air at 500°C followed by reduction in hydrogen

(or carbon monoxide) at a temperature in the range 450°–500°C, to give a prod-

uct in which the valency of the molybdenum was less than six. The catalyst

needed to be activated using sodium alkyls, possibly because the catalyst was

difficult to reduce. The original catalyst had a low productivity. Better results

were obtained when the alumina support had been halogenated, and the process

operated in the liquid phase at an operating temperature of 200°–275°C and a

pressure of 35–100 bar. Although other modifications were made to the catalyst

and the process, only three plants seem to have been built and operated. A dif-

ferent catalyst, based on vanadium pentoxide/alumina, was mentioned in some

early publications but it was only slightly easier to activate than the molyb-

denum oxide/alumina catalysts. Both molybdenum and vanadium oxides melt at

higher temperatures than chromium trioxide, so consequently they are not as

mobile on a silica surface.

8.5. POLYMERIZATION PROCESSES

The first commercial process for the manufacture of polyethylene was commis-

sioned in 1939, and required pressures of 2000 bar and temperatures of 200°C.

Free-radical initiators were used, and low-density polymer was formed as highly

branched chains. In the 1950s, new processes which did not require the use of

free-radial initiators, and which operated at low temperatures and low pressures

ranging up to 40 bar, were introduced. Catalysts were used and a higher-density

polyethylene with little or no chain branching was produced.

Low-density polymers produced at high pressure became known as low-

density polyethylene (LDPE), while the higher-density, low-pressure polymer is

now known as high-density polyethylene (HDPE). When

-olefin co-monomers

were used with ethylene, the low-pressure polymer was less dense than HDPE

as a result of regular branching in the polymer chain. The product is now known

as linear low-density polyethylene (LLDPE). The approximate proportions of

the different polyethylene types are shown in Table 8.5.

330 Chapter 8

TABLE 8.5. Relative Proportion of Polyethylene Types.

Year LDPE (%) HDPE (%) LLDPE (%)

1985 55 35 10

1990 45 40 15

1995 40 40 20

2000 30 40 30

Relative proportions of HDPE/LLDPE depend on demand.

Polyethylene is now produced by several processes that are more conven-

ient and do not require the high capital investment and expertise associated with

high-pressure operation. When catalysts were first introduced, the most widely

used processes were based on solution or slurry operation, in which polymeriza-

tion took place on catalyst suspended in an inert liquid hydrocarbon. Since 1968,

many new plants have used a gas phase process in which the catalyst and poly-

mer particles are suspended as a fluid bed, by the circulation of gaseous mono-

mer. An approximate estimate of the market share of each process by 1999 is

given in Table 8.6. A typical plant for the production of polyethylene is shown

in Figure 8.1.

As demand for polyethylene has increased, there has been a continuous ef-

fort to improve catalyst activity. This was important to achieve high productivity

at relatively short residence times. These properties have avoided the need to

remove catalyst residues from the product. It was also important for polypropyl-

ene to have a high isotactic index to eliminate the need for equipment for the

extraction of atactic polymer.

Two features distinguish the use of modern polymerization catalysts. First,

there is no separation of the catalyst from the product; second, the catalyst must

be able to disintegrate, or fragment, during operation, to expose new active cen-

ters. Use of the original low-activity catalysts led to unacceptably high residues

of transition metal, and in the case of Ziegler catalysts, halogen in the product,

which had to be removed before the polymer could be sold.

Reactor design is important and plug flow is essential to prevent both back

mixing of reactants and to maintain a reasonable product quality. This require-

ment led to the use of long tubular reactors, or a number of small vessels operat-

ing in series. Newer gas phase processes are efficient in this respect and opera-

tion can be controlled to remove product of the appropriate size and molecular

weight from the bottom of the fluid bed as it is formed. Another requirement of

the processes is the need to control reactor temperature by removing the heat of

reaction, so preventing the formation of low-molecular-weight polymers. High

temperatures also lead to agglomeration of the product and deactivation of the

catalyst. The reaction exotherm would lead to a rise in temperature of about 16C

for every % of ethylene converted.

Figure

olyole

roduc

onMob

is the

cracke

oxyge

hydro

enter t

Process

Gas pha

Slurry

Solution

8.1. Unipol/Do

ins plant is the

d with permiss

l Chemical and

further essent

removal of t

s. The most c

, water vapo

enation or by

e polymeriza

TABLE 8.

plant for the

Unipol reactor

on from Univa

Dow Chemical.

al feature of

e typical imp

mmon impur

, and sulfur

bsorption usi

ion unit.

6. Relative Use

Polyeth

–

5–

Olefi

production of

which is the l

ion Technolog

peration, parti

rities found

ties are acetyl

ompounds an

g convention

(%) of Comme

lene (%)

–

Polymerization

olyethylene. T

rgest of its typ

es, a Joint Ven

cularly in gas

in olefins der

nes, dienes, c

these are re

l procedures

cial Processes (

Catalysts

he key unit in

in the world.

ture between E

phase process

ved from ste

arbon monoxi

oved either

efore the ole

1999).

Polypropylene

(

–

the

es,

(

–

332 Chapter 8

8.5.1. Slurry Processes

In the slurry process,

a solution of about 2–6 wt% of the monomers in solution

in a hydrocarbon such as isobutane, n-pentane or n-hexane is slurred with the

catalyst, and the polymerization proceeds in a loop reactor achieving conver-

sions 97%. Typical co-monomers with ethylene can be butene-1, hexene-1, 4-

methylpentene-1, and octene-1. The catalyst is suspended in the solution, which

is then pumped as turbulent slurry through relatively narrow tubes. The polymer

is essentially insoluble in the liquid diluent and forms a slurry containing up to

30 wt% of small polymer particles, which are removed continuously.

The operating temperature of the process must be below the melting or sof-

tening point of polymers and, depending on the type of polymer produced and

the diluents used, is generally less than 100°C. A unique loop reactor was de-

signed by Phillips Petroleum Corporation to prevent the deposition of solids

onto the reactor walls. It is reported that loop reactors have operated continuous-

ly for as long as three years without the need for cleaning.

As an alternative,

autoclave slurry reactors with efficient stirrers have been operated in batch or

cascade processes. Changes in the partial pressure of the hydrogen moderator

from one stage to another allow a range of different average molecular weights

to be obtained within the product. Loop processes using Phillips chromium cata-

lysts operate at pressures in the range 20–40 bar, whereas autoclave processes

operate with Ziegler catalysts at lower pressures.

Residence times of up to 1 h are typical, with the continuous addition of

monomer, catalyst, and fresh diluent. Productivity can exceed more than 10,000

lb of polymer per pound of catalyst. The use of low concentrations of catalyst

allows the grade of the product to be changed within a few hours. The average

molecular weight of the product can be controlled by the type of catalyst used,

the activation procedure, operating conditions, and by changing the partial pres-

sure of the hydrogen moderator.

8.5.2. Solution Processes

In the solution process, polymerization takes place in solution, in a liquid hydro-

carbon such as cyclohexane or iso-octane, in which both the monomer and pol-

ymer can dissolve. About 0.2–0.6 wt% of a Ziegler catalyst is normally sus-

pended in the solvent, by stirring efficiently in a suitable reactor. The operating

temperature must be above the melting point of the polymer but lower than the

decomposition point of the catalyst. Typical temperatures are in the range 125–

175°C.

The polymerization is carried out at a typical pressure of about 30 bar, and

the solution of monomer and co-monomer is admitted into the reactor with fresh

catalyst. After reaction, the solution containing some 10–20% of polymer is re-

moved from the top of the reactor. The spent catalyst taken from the reactor is

Olefin Polymerization Catalysts 333

largely deactivated and is separated from the solution. In some of the early ver-

sions of the process, the catalyst could be recycled, provided that it still pos-

sessed sufficient activity for this to be worthwhile. The viscosity of the solution

needs to be sufficiently low to avoid circulation and pumping problems, and this

requirement can limit the magnitude of the average molecular weight of poly-

mer.

The residence time in the reactor is only a few minutes and the conversion

of monomers usually exceeds 95%, depending on the concentrations of ethylene

and polymer in the solution. Unused ethylene is recycled. This makes the pro-

cess very adaptable if it is necessary to change the grades of polymer at regular

intervals. It is claimed that more than 500 tonnes of polymer can be produced

per kilogram of titanium in the catalyst.

The use of solution processes was re-

ported to be increasing and by 1990, represented up to 15% of world polyeth-

ylene production.

8.5.3. Gas Phase Process

In the most successful process, the catalyst is fluidized in a large reactor by the

circulation of gaseous olefin monomers at high space velocity.

Polymerization

takes place in the gas phase at pressures less than 20 bar and an operating tem-

perature in the range 80°–100°C, to ensure that the polymer does not soften.

Conversion is limited to about 3% per pass. This allows circulation of the mon-

omers through external coolers to remove the heat of reaction while polymer and

catalyst particles reside in the reactor for an average of 2–3 h. The largest poly-

mer particles gradually fall to the bottom of the fluid bed and are removed when

the appropriate physical properties are achieved. Fresh catalyst is injected, often

as a slurry, to make up for losses. An immediate attraction of the process is that

the concentration of catalyst in the polymer is very low and the product does not

have to be de-ashed. Other gas phase processes operate with vertical reactors,

which use stirrers,

and horizontal vessels which use weirs and paddles.

Gas phase polymerization is flexible and can produce a wider range of pol-

ymer grades than the earlier processes. Capital and operating costs are relatively

low and the process has been very successful. Modern process developments

include the injection of liquid hydrocarbon directly into the fluid bed, rather than

to the recycle gas.

This makes the process more economic by allowing better temperature con-

trol and increased conversion.

This has become known as the condensation

mode and can considerably increase the capacity of existing units. All types of

active catalysts can be used.

334 Chapter 8

8.6. METALLOCENE/SINGLE-SITE CATALYSTS

In the early 1990s a new type of olefin polymerization catalyst was introduced.

Catalyst derived from metallocenes give a polymer with a very narrow molecu-

lar weight distribution with regular inclusion of co-monomers in the growing

chains. This was a significant improvement on the relatively wide molecular

weight distribution and a variable co-monomer content of Ziegler–Natta prod-

ucts.

Development has been slower than forecast owing to operational prob-

lems and the high cost of both metallocenes and the first methyl aluminoxane

activator. However, the high level of industrial interest was demonstrated by

350 patent applications by the end of 1993 increasing to 1500 by 1997. Compe-

tition led to litigation over patent rights, followed by a series of joint ventures

between the important producers. By 2000, up to 2% of the total worldwide pol-

yethylene production was based on the use of metallocene, or single-site cata-

lysts, as they became more accurately described.

This corresponded to about 1

million tonnes of polymer.

Exxon Chemicals tested their range of Exxpol catalysts in a demonstration

plant at Baton Rouge, Louisiana, in 1991.

The unit had a capacity of 15,000

tonnes of polyethylene per year using an autoclave type of reactor. The plant

was operated in the temperature range 100°–300°C and a pressure in the range

1000–2000 bar, using the monomer as a supercritical fluid solvent. These condi-

tions led to the inclusion of a co-monomer into the polyethylene, over a wide

range of molecular weights. A second, gas phase plant, was then retrofitted at

Mont Belvieu, Texas. It was found that plant capacities could be increased with

the new catalysts.

The Exxon catalyst is a derivative of cyclopentadiene, indene, or fluorene

which may be bridged, and which may also be substituted with up to five differ-

ent atoms or radicals. This organic group is bonded to a metal atom from Group

IV b of the Periodic Table, titanium, zirconium or hafnium, which has been acti-

vated by addition of a strong Lewis acid, to give a cationic structure.

In 1993, Dow Chemicals started operation using its Insite catalyst for a so-

lution process, in Freeport, Texas

, producing a copolymer of ethylene contain-

ing octene-1 co-monomers as well as plastomers. A second plant in Tarrogona,

Spain, began operation in 1996. The Dow catalyst consists of a Group-IV transi-

tion metal (titanium, zirconium, hafnium) covalently bonded to a substituted

monocyclopentadienyl group bridged with a heteroatom such as nitrogen. The

components form a “constrained cyclic structure with the metal center.” Other

companies, including BASF, Hoechst, and Phillips, were also testing single-site

catalysts by 1995.

Olefin Polymerization Catalysts 335

8.6.1. Early Development

Although single-site catalysts were only used industrially beginning in

1991/1993, they were not really new. In the late 1960’s and early 1970’s, the

Ballard group

at ICI Corporate lab studied a range of π-ally derivatives

transition metals, such as zirconium and chromium, and benzyl derivatives of

titanium and zirconium as potential candidates for olefin polymerization cata-

lysts. Relatively high activities were found, particularly when the π-ally or ben-

zyl derivative were reacted with a support, though all of the derivatives showed

significant activity as pure compounds in hydrocarbon solutions. This work

showed significant promise, but in the end, was never commercialized. Even

earlier, about 1953, both Natta

(Montedison) and Breslow

(Hercules) had

investigated the catalytic potential of metallocenes during mechanistic studies

on ethylene insertion. At the time they used a typical activator such as diethyl

aluminium chloride to produce the metallocene cation intermediate in two ion-

forming steps:

TiCl

+ AlR

Cl ↔ [Cp

TiR

. AlRCl

−

] (8.2)

[Cp

TiR

. AlRCl

−

] ↔ Cp

TiR

+ AlRCl

−

(8.3)

Unfortunately the forward reaction equilibrium was unfavorable and the catalyst

had a low activity.

In later experiments, it was shown that the metallocene catalysts had higher

activity when traces of water were present during the preparation of the catalyst.

This is in direct contrast with the use of the original Ziegler Natta catalysts, for

which the rigorous exclusion of water was essential. This observation was linked

by Breslow to the formation of aluminoxanes by the hydrolysis of the alky alu-

minium derivative.

The activation of zirconacenes by high molecular weight aluminoxanes de-

rivatives led to the development of active and stable catalysts for the polymeri-

zation of ethylene. High yields of polymers with molecular weights in the range

to 1.5x10

, depending on operating temperature, are now produced.

The

high-molecular weight methyl aluminoxane alkylates the zirconacenes and con-

verts the product into a cationic complex:

ZrRX + MAO + Olefin → Cp

ZrR(Olefin)

.MAOX (8.4)

A 200-fold excess of MAO (methyl aluminoxane) was originally needed to

provide a reasonable active catalyst. By the late 1990s, this proportion had been

reduced to about 20. It was later realized that the active species in the catalyst

formulation was a cation. This led to the replacement of the aluminoxanes spe-

336 Chapter 8

cies by compounds that could provide a source of a suitable non-coordinating

anion, such as tetrakispentafluorophenylborate:

ZrMe

+ PhN

HB(C

)

→ Cp

ZrMe

B(C

)

+ PhNMe

+ CH

(8.5)

The first metallocene catalyst (Cp

MCl

.MAO) was completely nonstereo-

specific for the polymerization of propylene and the polymer was atactic. It was

then shown that when the catalyst structure was modified by replacing the halo-

gen atoms with phenyl groups (Cp

MPh

.MAO) the catalyst became stereospe-

cific at a temperature of –30°C.

Ansa metallocenes

were the first soluble, stereospecific, single-site cata-

lysts to be produced and were based on indenyl groups

[racemic tetrahydroeth-

ylene (1-indenyl)

ZrCl

]. By variations in the substitution of the indene group it

was possible, with a MAO activator, to match the performance of the third-

generation magnesium chloride–supported Ziegler–Natta catalysts.

In quantita-

tive terms, this was equivalent to the production of about 1 tonne of polypropyl-

ene per gram of transition metal, with an isotactic index greater than 99% and

molecular weight up to at least 500,000 g.mol

−1

8.6.2. Early Development

A single-site catalyst is made up of three components:

• The catalyst framework is usually a pair of organic ring compounds or

ligands, linked by a bridging group. The rings can be substituted with var-

ious organic groups to modify the electronic properties of the catalyst

sites to produce the specific grades of polymer as required. Common lig-

ands include radicals from compounds such as cyclopentadiene, indene,

fluorene and benzidene. Other ligands include imido-compounds, and

some organo-phosphine derivatives, with bridges formed from organosil-

icon groups and a wide range of organic compounds such as ethylene.

The ligand pairs are held in the form of an asymmetric V-shape, de-

scribed as the ansa form (ansa is Latin for bent handle)

and are able to

form stereospecific catalyst sites.

• A transition metal halide from the sub-group titanium, zirconium, and

hafnium is coordinated between the ligand pairs. This provides a rigid,

potentially stereo-selective framework that controls the access of mono-

mers to the active site.

• The use of a co-catalyst to generate the active catalytic cationic centers is

required. Derivatives of methyl aluminoxanes, salts containing the

An ansa metallocene is one in which ligand pairs are attached to the form of an asymmetric V-

shape.

Olefin Polymerization Catalysts 337

tetrakis-pentafluorophenylborate anion, other non-coordinating anions or

trispentafluorophenylaluminium can be used.

Some processes are based upon soluble catalysts but gas phase and slurry pro-

cesses use about 1–2 wt% catalyst impregnated onto a suitable form of silica.

Changes to the catalyst structure can modify catalytic activity, the molecular

weight of the product, and the incorporation of co-monomers in the polymeriza-

tion of ethylene. In the production of polypropylene, structural changes within

the catalyst also direct the formation of isotactic, syndiotactic, and non-

crystalline atactic polymers.

Better control of the polymerization process was

therefore possible, as shown in Table 8.7.

The following general observations have been made:

• The presence of silicon bridges within the catalyst leads to an increase in

the molecular weight, and a small increase in stereospecificity of the pol-

ymer. Longer bridges decrease both molecular weight and stereospecifici-

ty.

• Phenyl substitution within the bridge gives a further increase in molecular

weight.

• In the case of catalysts based on indene, substitution of the both the in-

denyl group and the bridge gives an increase in molecular weight.

• Substitution of the indene ring with phenyl or naphthyl groups, and sub-

stitution of the bridge with alkyl groups leads polymer with a high mo-

lecular weight and stereospecificity.

Asymmetric single-site catalysts have been produced from bridged fluo-

renylcyclopentadienyl rings with alkyl substitution in both the bridge and the

cyclopentadiene ring. These modifications can lead to control over the alterna-

tion of isotactic and syndiatactic sequences in the polymer and, consequently,

TABLE 8.7. Stereospecific Metallocene Catalysts.

Catalyst

Activity Molecular weight Tacticity

Ethylene (indenyl-1)

ZrCl

180 2.4 × 10

Dimethyl silylene

(2-Me, indenyl-1)

ZrCl

99 1.9 × 10

Dimethyl silylene (2-Me,

5-benzene, indenyl-1)

ZrCl

550 7.3 × 10

Dimethyl silylene (2-Me,

5-naphthalene, indenyl-1)

ZrCl

880 9.2 × 10

Many other catalysts have been introduced for the polymerization of propylene to produce a range

of different structures ranging from isotactic, syndiotactic, to hemi-isotactic and isotac-

tic/syndiotactic sequences. These include catalysts based on a bridged cyclopentadiene/ﬂuorene

methyl-substituted ligand framework that can alternate the substitution of ethylene and propylene in

copolymers.