Tải bản đầy đủ - 0 (trang)
2 Dimerization, Oligomerization, and Polymerization of Alkenes

2 Dimerization, Oligomerization, and Polymerization of Alkenes

Tải bản đầy đủ - 0trang

Dimerization, Oligomerization, and Polymerization


range of ligands, and so the term single-site catalyst is now also used. The

Phillips catalyst, consisting of Cr supported on Al2O3, behaves similarly.

These catalysts have had a revolutionary impact on the polymer

industry because the variation of L and L′ allows delicate control over

the microstructure of the polymer—how the atoms are connected in

the chains—and over the polydispersity—the distribution of chain

lengths. The catalyst structure controls the physical properties of the

final polymer, affecting how it can be of practical use. Metallocene

polymers can be designed to be very tough, or act as elastomers, or be

easily heat-sealed, or have excellent optical properties, or have easy

processability, and they have therefore displaced higher-cost polymers,

such as polyurethanes, in many applications. Their economic advantage

comes from the low cost of ethylene and propylene. Syndiotactic polypropylene (12.9), unobtainable in pure form before metallocene catalysis, is softer but tougher and optically more transparent than other

forms. It is used in films for food storage and in medical applications.

Catalyst Activation

Kaminsky showed that Cp2ZrCl2 must first be activated with methylalumoxane (MAO, [MeAlO]n), formed by partial hydrolysis of AlMe3.

Initial methylation by MAO gives Cp2ZrMe2, followed by Me− abstraction by MAO to form the active 14e species, [Cp2ZrMe]+, stabilized

by the “noncoordinating” [Me{MeAlO}n]− counterion. Mass spectral

studies have thrown some light on the structure of MAO; one component is [(MeAlO)21(AlMe3)11Me]−.9


Metallocenes produce polyethylene that is strictly linear, without side

branches, termed LLDPE (linear low-density polyethylene). Other processes tend to produce branches and hence a lower quality product. If

shorter chains are needed, H2 can be added to cleave them via heterolysis (Eq. 12.15).


Polypropylene has an almost perfectly regular head-to-tail structure

when produced with metallocenes. The arrangement of the methyl



groups in isotactic polypropylene (12.8) gives the polymer chain a

helical rod structure. The rods are chiral, and catalysts that form isotactic polypropylene are also chiral. Since both hands of the catalyst are

normally present, rods of both left- and right-handed forms are present

in equal amounts.

Syndiotactic polypropylene has no chirality and is formed by catalysts lacking chirality. It tends to adopt a planar zigzag conformation

(12.9) of the main chain.


Dimerization, oligomerization, and polymerization all rely on the

Cossee–Arlman mechanism that consists of repeated alkene

1,2-insertion into the M–C bond of the growing polymer chain (Fig.

12.1).10 The three types only differ in their kg/kt ratio, that is in their

relative rates of chain growth by insertion (kg) to termination by β

elimination (kt). If chain termination is very efficient, kg/kt is small

and we may see dimerization; if kg/kt is somewhat greater, oligomerization, as in the SHOP process discussed later; and if kg/kt is very

large, true polymerization will result, as in Ziegler–Natta and metallocene catalysis. Although discussed separately, they are nevertheless closely related mechanistically (Eq. 12.16).


Unlike the conversion of ethylene to linear polyethylene (PE), propylene polymerization to polypropylene (PP) introduces stereochemical complexity because we can obtain 12.8, 12.9 or a random atactic

product. Surprisingly, selective formation of syndiotactic propylene

(12.9) is seen for many metallocene polymerization catalysts. To see

why, we need to know that d0 [Cp2ZrR]+ is pyramidal (12.12 in Fig.

12.1)11 for much the same reasons that made d0 WMe6 prismatic (Section

3.1). We next have to assume that the pyramidality inverts after each

insertion step, transferring the polymer chain from one side to the other

like a windshield wiper. The nth alkene to insert therefore occupies the

opposite binding site from the (n − 1)th and (n + 1)th alkene—once

Dimerization, Oligomerization, and Polymerization


the insertion takes place, the newly formed M–C bond automatically

finds itself in the other binding site (Fig. 12.1).

In catalyst 12.10, each binding site is locally chiral, but because the

whole molecule has C2 symmetry, both sites have the same local symmetry. The propylene monomers insert in the same way, leading to

isotactic product 12.8. In catalyst 12.11, each binding site is again locally

chiral, but because the whole molecule has a plane of symmetry, each

site has the opposite local symmetry. The propylene monomers insert

in the two possible ways with alternation between the two on each successive insertion, leading to syndiotactic product 12.9.

Computational work indicates the probable structures for the key

intermediate propylene complexes in the two classes of catalyst. In

the chiral isotactic catalyst, 12.10, the methyl group tends to be

located as shown in Fig. 12.2 (upper), so that successive propylenes

enter with the same chiralities and bind via the same face (re in the

figure). In the achiral syndiotactic catalyst, 12.11, in contrast, successive propylenes enter with opposite chiralities and bind via alternating faces (re then si).

FIGURE 12.1  Windshield wiper model for alkene polymerization by metallocene catalysts. The insertion causes the M–C bond to the polymer chain (PC)

to move alternately from one side to the other in the pyramidal [Cp2ZrR]+

intermediate as each insertion occurs. The open box represents a vacant site

in [Cp2ZrR]+ where the next alkene can bind.







FIGURE 12.2  Chiral metallocene catalyst 12.10 (upper) leads to alternate

propylenes (shaded) binding via the same re-face to give isotactic polymer.

The achiral catalyst 12.11 (lower) leads to alternate propylenes binding via

the opposite faces, re then si, to give syndiotactic polymer. Source: From

Brintzinger et al., 1995 [64]. Reproduced with permission of Wiley-VCH.

The Cossee–Arlman mechanism involving C=C insertion into the

M−C bond of the growing polymer chain seems to apply generally. The

insertion is much faster in the Ziegler–Natta catalysts than in many

isolable 18e alkyl olefin complexes because the reaction is strongly

accelerated by coordinative unsaturation in the key intermediate, such

as 16e [Cp2ZrMe(C2H4)]+. The alkyl can become agostic and rotate to

direct the lone pair of the R− ligand toward the alkene, facilitating

insertion (modified Green-Rooney mechanism). Theoretical work has

indicated that in the model intermediate [Cp2ZrMe(C2H4)]+, the CH3

group is agostic (Fig. 12.3, left), as allowed by the formally 16e count

for this species. The principal axis (C3 axis) of the methyl group is

indeed rotated by 40°, turning the CH3 sp3 hybrid orbital toward the

alkene. At the transition state for insertion (Fig. 12.3, right), this value

has increased to 46°.

Dimerization, Oligomerization, and Polymerization


FIGURE 12.3  Structures of a model intermediate [Cp2ZrMe(C2H4)]+ (left),

showing the agostic methyl. The methyl leans over even more at the transition

state (right). The results were obtained by Ziegler and coworkers by density

functional theoretical calculations. Source: From Fan et al., 1995 [65]. Reproduced with permission of the American Chemical Society.

In the f-block metals, successive alkene insertions into a Lu−R bond

can be observed stepwise (Eq. 12.17). Not only do the alkenes insert

but the reverse reaction, β elimination of an alkyl group, as well as the

usual β elimination of a hydrogen, are both seen. For the d block, a β

elimination of an alkyl group would normally not be possible, but the

greater M−R bondstrengths in the f block makes the alkyl elimination

process sufficiently favorable to compete with β elimination of H.


SHOP Oligomerization

Most late d block metals favor β elimination, thus their higher kt often

leads to dimerization or oligomerization, rather than polymerization.

The Shell higher olefins process (SHOP) is based on homogeneous

nickel catalysts (Fig. 12.4) discovered by Keim.12 These oligomerize

ethylene to give 1-alkenes of various chain lengths (e.g., C6–C20). Insertion is therefore considerably but not overwhelmingly faster than β





oligomerization step









isomerization step heterog. catalyst



metathesis step heterog. catalyst



hydroformylation step homog. catalyst





FIGURE 12.4  In the Shell higher olefins process (SHOP), Keim’s nickel

catalyst gives 1-alkenes of various chain lengths. The subsequent steps allow

the chain lengths to be manipulated to maximize the yield of C10–C14 products.

Finally, SHOP alkenes are often hydroformylated, in which case, the internal

alkenes largely give the linear product, as discussed in Chapter 9.

Dimerization, Oligomerization, and Polymerization


elimination. The C10−C14 fraction is a desirable feedstock; for example,

hydroformylation gives C11–C15 alcohols that are useful in detergent

manufacture. The broad chainlength distribution from SHOP means

that there is a big non-C10–C14 fraction with longer (>C14) and shorter

waste by design via isomerization and metathesis steps that manipulate

the chain lengths so as to produce more C10–C14 material from the

longer and shorter chains. The fact that internal C10–C14 alkenes are

formed does not matter because hydroformylation gives linear alcohols

even from internal alkenes, as discussed in Section 9.4. Homogeneous

catalysts were strong contenders for the isomerization and metathesis

steps of SHOP, but in practice, heterogenized catalysts were adopted.

Several plants are now operating with a production of >107 tons/y.

Another commercially important reaction is du Pont’s synthesis

of 1,4-hexadiene. This is converted to synthetic rubber by copolymerization with ethylene and propylene, which leaves the polymer

with unsaturation. Unsaturation is also present in natural rubber, a

2-methylbutadiene polymer 12.13, and is necessary for imparting

elastomer properties and permitting vulcanization, a treatment with

S8 that cross-links the chains via C–S–C units and greatly hardens

the material.

The 1,4-hexadiene is made by codimerization of ethylene and butadiene, with a RhCl3/EtOH catalyst (Eq. 12.18). The catalyst is about

80% selective for trans-l,4-hexadiene, a remarkable figure considering

all the different dimeric isomers that could have been formed. The catalyst

is believed to be a rhodium hydride formed by reduction of the RhCl3

with the ethanol solvent (Section 3.2). This must react with the butadiene to give mostly the anti-methylallyl (crotyl) intermediate, which

selectively inserts an ethylene at the unsubstituted end. The cis/trans

ratio of the product probably depends on the ratio of the two isomers

of the crotyl intermediate. Adding ligands such as HMPA to the system

greatly increases the selectivity for the trans diene. By increasing the

steric hindrance on the metal, the ligand probably favors the syn isomer

of the crotyl ligand over the more hindered anti isomer. The rhodium

hydride is also an isomerization catalyst, and so the 1,4-hexadiene is

also converted to the undesired conjugated 1,3 isomers. The usual way

around a problem like this is to run the reaction only to low conversion,



so that the side product is kept to a minimum. The substrates, which are

more volatile than the products, are easily recycled.




Most organic commodity chemicals are currently made commercially

from ethylene, a product of oil refining. In the next several decades, we

may see a shift toward other carbon sources for these chemicals. Either

coal or natural gas (CH4) can be converted with steam into CO/H2

mixtures called “water–gas” or “synthesis gas” and then on to methanol

or to alkane fuels with various heterogeneous catalysts (Eq. 12.19).

In particular, the Fischer–Tropsch reaction converts synthesis gas to

a mixture of long-chain alkanes and alcohols using heterogeneous



Water–Gas Shift

The H2:CO ratio in synthesis gas depends on the conditions of its formation, but the initial ratio obtained is often ∼1 : 1, insufficiently high

for a number of applications. For example, conversion of CO to CH3OH

requires a 2 : 1 H2:CO ratio. If so, we can change the ratio via the water–

gas shift reaction (Eq. 12.20), catalyzed either heterogeneously (Fe3O4

or Cu/ZnO) or by a variety of homogeneous catalysts, such as Fe(CO)5.

The reagents and products in Eq. 12.20 have comparable free energies

so the reaction can be run in either direction but H2 production from

CO and H2O is the usual goal.14

Activation of CO And CO2




In the mechanism proposed for Fe(CO)5 (Eq. 12.21), CO bound to Fe

becomes activated for nucleophilic attack by OH− at the CO carbon.

Decarboxylation of the resulting metalacarboxylic acid probably does

not take place by β elimination because this would require prior loss

of CO to generate a vacant site; instead, deprotonation may precede

loss of CO2, followed by reprotonation at the metal to give [HFe(CO)4]−.

Protonation of this anionic hydride liberates H2 and regenerates the


Monsanto Acetic Acid Process

Over 8 million tons of acetic acid derivatives a year are produced in

>99% selectivity by carbonylation of methanol with a Rh(I) catalyst,

[RhI2(CO)2]− (Eq. 12.22).15 The process is 100% atom economic since

all the reactant atoms appear in the acetic acid. The net effect is the

cleavage of the methanol H3C–OH bond and insertion of a CO. The

methanol substrate requires activation with HI to produce an equilibrium concentration of MeI, which can oxidatively add to the metal in

the turnover limiting step (Fig. 12.5).


Once the rhodium methyl is formed, migratory insertion with CO gives

an acetylrhodium iodide. Reductive elimination of the acyl iodide is

followed by hydrolysis to give acetic acid and HI, which is recycled. The

Monsanto process for making acetic acid is replacing the older route

that goes from ethylene by the Wacker process to acetaldehyde, followed by oxidation to acetic acid in a second step. An improved process

based on iridium (Cativa process) has been developed by BP-Amoco,15

and a biological analog of this reaction is discussed in Section 16.4.


















CO –


ox. addn.





CO –



red. elim.












FIGURE 12.5  Catalytic cycle proposed for the Monsanto acetic acid process

that converts MeOH and CO to MeCOOH with a Rh catalyst.

CO2 Activation

A related process, CO2 activation, has attracted much attention in the

hope of producing useful chemicals from a cheap starting material.16,17

CO2 is so thermodynamically stable, however, that few potential products can be made from CO2 by exothermic processes. With ∼10 12 tons

of excess CO2 already in the atmosphere and ∼2.4  ×  109 tons being

added per year,16b CO2 conversion to chemical products cannot have a

significant impact on mitigating the climate change problem, but it at

least goes in the right direction.

Catalytic reduction of CO2 with H2 to give HCOOH involves CO2

insertion into M–H bonds. Although this is “uphill” thermodynamically (ΔG  =  +8  kcal/mol), the reaction becomes favorable under

Activation of CO And CO2


gas pressure or in the presence of base to deprotonate the formic

acid. One of the best homogeneous catalysts to date is 12.14, which

gives 150,000 turnovers per hour at 200°.18 As an 18e catalyst, a

hydride is likely to attack an outer sphere CO2 to give HCOO− ion

that can now coordinate to Ir via O, so this is an unusual type of

insertion, greatly favored because a hydride trans to another hydride

is particularly hydridic, consistent with the sdn model of Section 1.8.

This step is followed by RE of HCOOH and OA of H2 to close the

cycle (Eq. 12.23).


Formic acid can easily be further converted, for example, to

CH3OH + CO2 by disproportionation using [Cp*Ir(dipy)(OH2)][OTf]219

or to H2 + CO2 with [{P(CH2CH2PPh2)3}FeH] as catalyst.20

Carbon–carbon bond formation from CO2 is illustrated by the Pd

catalyzed conversion of CH2=CHCH2SnR3 to CH2=CHCH2CO2SnR3

by a series of [(η3-allyl)PdL(OOCR)] complexes (L  =  phosphine or

NHC). The stannane transfers the substrate allyl to Pd, followed by the

attack of the resulting η1-allyl terminal =CH2 group on CO2 in the key

C–C bond-forming step.21

The most important CO2 activation process is photosynthesis in

green plants, in which solar photons drive a reaction that would otherwise be uphill thermodynamically: the reduction of CO2 to carbohydrates coupled to water oxidation to O2. Many metalloenzymes are

involved in these processes, such as ribulose diphosphate carboxylase

that “fixes” CO2 via nucleophilic attack on an enolate anion from a

sugar. Artificial photosynthesis22 takes the natural version as inspiration

and seeks to photochemically reduce CO2 to fuels such as MeOH.

Naturally, a catalyst is needed–Re(CO)3(bpy)X holds promise in this

regard by converting CO2 to CO and HCOOH.23

Assigning mechanisms in electrocatalysis is hard, as illustrated by

what was initially considered a “metal-free” electroreduction of CO2 to

HCOOH with pyridinium ion as the electrocatalyst, where direct interaction of the 1e-reduced [C5H5NH] radical with CO2 was proposed.

An alternative mechanism involving the Pt electrode, thought to form

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

2 Dimerization, Oligomerization, and Polymerization of Alkenes

Tải bản đầy đủ ngay(0 tr)