Aryloxy tungsten-based classical catalytic systems and group 14 metal-containing dienes in acyclic diene metathesis polymerization /

Material Information

Aryloxy tungsten-based classical catalytic systems and group 14 metal-containing dienes in acyclic diene metathesis polymerization /
Gómez, Fernando José, 1973- ( Dissertant )
Wagener, Kenneth B. ( Thesis advisor )
Place of Publication:
Gainesville, Fla.
University of Florida
Publication Date:
Copyright Date:
Physical Description:
x, 157 leaves : ill. ; 29 cm.


Subjects / Keywords:
Alkenes ( jstor )
Catalysts ( jstor )
Dienes ( jstor )
Flasks ( jstor )
Ligands ( jstor )
Metathesis ( jstor )
Molecular weight ( jstor )
Monomers ( jstor )
Polymerization ( jstor )
Polymers ( jstor )
Chemistry thesis, Ph. D ( lcsh )
Diolefins ( lcsh )
Dissertations, Academic -- Chemistry -- UF ( lcsh )
Metathesis ( lcsh )
Organometallic compounds ( lcsh )
Organometallic compounds -- Synthesis ( lcsh )
Polymerization ( lcsh )
Polymers -- Synthesis ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


The design of a methodology for the acyclic diene metathesis (ADMET) polymerization of hydrocarbon and ester dienes using classical catalytic systems composed of aryloxide complexes of tungsten and tin-based activators is presented. The experimental conditions explored afford clean, linear polymers of high molecular weight. The attributes of the tungsten complexes investigated were extrapolated to include new precatalyst structures, which demonstrated their applicability not only in ADMET polymerization, but also in other metathesis processes such as ring opening metathesis polymerization and ring closing metathesis. The nature of the formation of the actual catalytic species in the mentioned classical systems motivated the investigation of a synthetic scheme towards ADMET polycarbostannanes in which stannadienes act as both monomers and activators of the catalytic systems, and its results are also presented. This synthetic approach afforded tin-containing ADMET polymers that exhibit interesting thermal properties such as high ceramic yields and low glass transition temperatures. ADMET polymerization was used in the synthesis of new tin-containing polymers, in an effort to investigate the effect of structure modification on their physicochemical properties. These include unsaturated polymers containing the dimethyl- or diphenylstannane moiety along the polymer backbone. A methodology for the synthesis of organometallic polymers containing oligostannane segments was devised and used in their preparation. Complementing the organometallic polymer structures studied, ADMET polymers were also synthesized from germanadienes, providing access to linear, unsaturated germanium-containing polymers. ( , )
KEYWORDS: ADMET polymerization, metal-containing polymers, olefin metathesis, tin, germanium, catalysis, polymers
Thesis (Ph. D.)--University of Florida, 2000.
Includes bibliographical references (leaves 150-156).
General Note:
General Note:
Statement of Responsibility:
by Fernando José Gómez.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
45839969 ( OCLC )
002566149 ( AlephBibNum )
AMT2430 ( NOTIS )


This item has the following downloads:

Full Text







A mi madre,
por soltar siempre mi mano
para que yo recogiera canelitas.

A la memorial de mi abuelo Mario
quien me cneie', que un hombre debe ser
honest, trabajador y buen amigo.

To my mother,
for she always let my hand go
so I could pick up canelitas.

To the memory of my grandfather Mario,
who taught me that a man should be
honest, a hard worker and a good friend.


Near or far, the infinite confidence and support I found in my family made every

day possible. No words can describe the way in which their hands helped mine to get

things done through optimism, strength, and the promise of better times for all of us.

This learning experience could have not been possible without the constant

scientific and personal interaction with the past and present members of the Butler

Polymer Research Laboratory, especially with those in the Wagener group who kept alive

my interests in our chemistry. Special thanks go to Dr. Mark Watson, Dr. Shane Wolfe,

Dr. Jason Portmess, and Dr. Debra Tindall for turning into friendship their support,

guidance, and acceptance. Very special gratitude is expressed to Dr. Tammy Davidson,

for she was a caring and supportive friend in the hardest of times. The cultural and

racquetball experiences with Gayanga Weeraserekera are also appreciated.

A large group of friends from every corner of Latin America and especially from

my missed Colombia were my home away from home and helped me tremendously

through these years. Our experiences together made me proud of our culture and heritage.

Among these, the support and unconditional friendship of Roberto Bravo, Edwin

Saldarriaga, and Maria and Mauricio Ospina will always be remembered.

Thanks to Claudia Madrigal, for offering me the caring hands, companionship,

and love that made the end of my graduate career much easier.

The contributions to the scientific work described herein by Dr. Andrew Bell (BF

Goodrich), Dr. Jerry Feldman (DuPont), Dr. Kirk Abbey (LORD Corp.), Michael Manak,

Nelson Guzman, Gabriela Feldberg, and Dr. Carlos Ortiz were instrumental for the

execution of this research. Thanks also to Professors Lisa McElwee-White, James

Boncella, William Dolbier, and Christopher Batich for serving on my graduate committee

and for sharing with me their experiences every time I wanted to talk chemistry with

them. The prompt and invaluable help of the spectroscopic services staff, especially Dr.

Khalil Abboud, Dr. David Powell, and Dr. Ion Ghiviriga, is also acknowledged.

Thanks also to the polymer chemistry office staff, and very especially to Mrs.

Lorraine Williams, for always having the answers to my questions.

The financial support provided by LORD Corporation (Cary, NC) and the

National Science Foundation is highly appreciated.

Finally, eternal gratitude is extended to Professor Kenneth B. Wagener, for being

a true mentor, for giving me the appropriate amount of opportunity and advice I needed,

and for sharing with me his exemplary research, teaching, and life philosophy.



A CK N OW LED GM EN T S.. ..................................................................... ................ iii


1 INTRODUCTION ............................ .. ......... .............. .............. .. 1

Addition and Condensation Polymerization, Chain and Step
Polym erization ................................................... ...... .. ........ ... ........... 2
Macromolecule Formation: The Concepts................................ 2
From Small to Large Molecules: Functionality............................ 5
O lefin M etathesis ...................................................................... 7
Exchange M etathesis................................................... 10
R ing C losing M etathesis ............................... ............................. 11
Ring Opening Metathesis Polymerization............................... 12
Acyclic Diene Metathesis Polymerization ............................. 15
M echanism and K inetics ............................... .............. 16
Functionality and ADMET Polymers................................ 19
Architectures and Modeling with ADMET.................... 21

ADM ET POLYM ERIZATION .................. ............................................... 26

Metal Carbenes and Olefin Metathesis Catalysts.................................26
W ell-D efined ADM ET Catalysts .......................................... .............. 28
Olefin Metathesis Classical Catalytic Systems ................................... 32
Transition Metal Halides and Alkoxides................................. 33
Classical Systems and ADMET Polymerization........................ 36
Tungsten Aryloxide Complexes and ADMET Polymerization.............. 39
ADMET Polymerization of 1,9-Decadiene............... .............. 40
ADMET Polymerization of 1,8-Nonadiene ..............................44
The Reaction of Other Hydrocarbon Dienes............................. 46
Investigations with Oxygen Functionalized Dienes................... 47
C onclu sions ...................................................................................... . 49

MODIFICATION OF THE ARYLOXIDE LIGAND ........................................ 50

Aryloxide Tungsten Complexes: The Ligand Electronic Effect ............ 51
Ligand Substituent Effects and Catalytic Activity .............................. 52
N ovel Aryloxy-Tungsten Structures............... ................................... 55
Structural C haracterization........................................ .............. 57
A D M E T C hem istry ................................................... .............. 64
R O M P C hem istry ...................................................... .............. 68
R ing Closing M etathesis ........................................... .............. 71
C onclu sions ...................................................................................... . 73

WELL-DEFINED METATHESIS CATALYSTS................... .................... 74

M etal-C containing Polym ers .................................................. .............. 74
Tin-C containing Polym ers ...................................................... .............. 75
P olystannanes ..................................... ....... ............ .. ........ .. .. 76
P oly carb ostannanes ....................................................................... 79
Polycarbostannanes via ADMET Polymerization...................... 81
Monomer-Activated Metathesis Polymerization....................................... 81
Poly [bis(4-pentenyl)di-n-butylstannane] .... ............... .............. 82
Poly [bis(3-butenyl)di-n-butylstannane]......................................... 88
N ew Polycarbostannanes....................................................... .............. 90
M onom er Synthesis ............... .................................... .............. 9 1
Poly(dim ethyldi-n-butylstannane)............................. ............... 92
Poly(diphenyldi-n-butylstannane) ............................. .............. 97
Oligostannane Segments in ADMET Polymers................................. 101
M onom er Synthesis...................................................... 102
AD M ET Polym erization ....... ......... .................................... 106
C onclu sions ................................................................. . .......... 109


Main Chain Germanium Containing Polymers....................................... 111
P olygerm anes .................................. .. ....... ............ . ... ....... 111
Polycarbogerm anes................................................. .............. 112
Germanium-Containing Polymers via Metathesis ............................ 115
ADMET Polycarbogermanes ........................ 117
M onom er Synthesis...... .... .... .................... 117
ADMET Polymerization ........................... 118
C onclu sions ................................................................. . .......... 125


6 E X P E R IM E N T A L .............................................................................................. 126

Instrum entation and A nalysis....... ... ........................................ 126
M materials and Techniques...... ....... ..... .................... 127
Synthesis and Characterization ....... .......... ..................................... 129
Experim ental for Chapter 2..................................... .............. 129
Polymerization of 1,9-decadiene with
WOCl2(O-2,6-C6H3Br2)2 and tri-n-butyltin hydride .... 129
Polymerization of 1,9-decadiene with
WOCl2(O-2,6-C6H3Br2)2 and tetrabutyltin (general
polymerization procedure) ................. ................. 130
ADMET polymerization of 1,8-nonadiene .................. 130
ADMET polymerization of 9-decenyl
10-undecenoate (37) ....... ................... ................. 131
Experim ental for Chapter 3 ....... ... ................................... 132
Sem iem pirical calculations ................... ................. 132
Synthesis of
(V I) oxychloride ....... ... .................................... 132
Synthesis of
trans-bis(2,4,6-tribromophenoxy)tungsten (VI)
oxychloride ............................. .. .......... . ........... .. 133
Synthesis of 2,6-dibromo-4-(trifluoromethyl)-phenol..... 134
Synthesis of
-tungsten (VI) oxychloride ................. ................. 135
Synthesis of
(V I) oxychloride ....... ... .................................... 136
ROMP of dicyclopentadiene ................. ................. 136
ROM P of norbornene ....... ................... ................. 137
RCM of diethyl diallylmalonate..... ............ 137
Experim ental for Chapter 4..................................... .............. 138
Synthesis of bis(4-pentenyl)di-n-butylstannane (73)....... 138
Synthesis of bis(4-pentenyl)dimethylstannane............. 139
Synthesis of bis(4-pentenyl)diphenylstannane ............ 140
Polymerization of 73 using
(2 6 ) ............................................................................... 1 4 1
Polymerization of 73 using
WOCl2(O-2,6-C6H3-Br2)2 (32).................................. 142
Polymerization of bis(4-pentenyl)diphenylstannane
(83) with
(2 6 ) ............................................................................... 1 4 3

Ring closing metathesis of
bis(3-butenyl)di-n-butylstannane using
WOCl2(O-2,6-C6H3-Br2)2 (32)............................... 143
Synthesis of
6,6,7,7-tetrabutyl-6,7-distanna-1,11-dodecadiene........ 144
ADMET polymerization of 92 ............... ................. 145
Experim ental for C chapter 5 ......................................................... 146
Synthesis of bis(4-pentenyl)-diethylgermanium
(113). (6,6-diethyl-6-germana-1,10-undecadiene) ....... 146
Cognate synthesis of
bis(3-butenyl)-diethylgermanium (114).
(5,5-diethyl-5-germana-1,8-nonadiene)..................... 147
Cognate synthesis of
bis(4-pentenyl)-dimethylgermanium (115).
(6,6-dimethyl-6-germana-1,10-undecadiene)............ 147
ADMET polymerization of diene 113: synthesis of
polym er 116 ....................... .................. ..... ............... 148
Cognate polymerization of diene 114: synthesis of
polym er 117 .............. ............ ...... ...... .... ............ .. 148
Cognate polymerization of diene 115: synthesis of
polym er 118 .................................... ...... ............ .. 149

REFERENCES ............................................... ............................ 150

BIOGRAPH ICAL SK ETCH ................................................................. .............. 157

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Fernando Jose G6mez

May 2000

Chairman: Kenneth B. Wagener
Major Department: Chemistry

The design of a methodology for the acyclic diene metathesis (ADMET)

polymerization of hydrocarbon and ester dienes using classical catalytic systems

composed of aryloxide complexes of tungsten and tin-based activators is presented. The

experimental conditions explored afford clean, linear polymers of high molecular weight.

The attributes of the tungsten complexes investigated were extrapolated to include

new precatalyst structures, which demonstrated their applicability not only in ADMET

polymerization, but also in other metathesis processes such as ring opening metathesis

polymerization and ring closing metathesis.

The nature of the formation of the actual catalytic species in the mentioned

classical systems motivated the investigation of a synthetic scheme towards ADMET

polycarbostannanes in which stannadienes act as both monomers and activators of the

catalytic systems, and its results are also presented. This synthetic approach afforded

tin-containing ADMET polymers that exhibit interesting thermal properties such as high

ceramic yields and low glass transition temperatures.

ADMET polymerization was used in the synthesis of new tin-containing

polymers, in an effort to investigate the effect of structure modification on their

physicochemical properties. These include unsaturated polymers containing the dimethyl-

or diphenylstannane moiety along the polymer backbone. A methodology for the

synthesis of organometallic polymers containing oligostannane segments was devised and

used in their preparation.

Complementing the organometallic polymer structures studied, ADMET polymers

were also synthesized from germanadienes, providing access to linear, unsaturated

germanium-containing polymers.


Synthetic polymeric materials occupy a privileged position in modem society, for

they have become indispensable. From commodity plastics to specialty polymers, the

field of polymer synthesis is challenged everyday with demands for new materials and

efficient polymerization processes. The interaction between the commercial, engineering

and scientific activities involved in polymer technology seems to be shaped as a constant

quest for developments, and as chemists we are responsible for discovery.

Long before our time others set the frame-theory onto which macromolecular

science is inscribed. The fact that a group of substances behaved uniquely when

compared to chemicals known at the time and the fact that the similarities with respect to

known natural polymers were not clear, were instrumental for the birth of a new science:

polymer science. Within this field, nature is the greatest maker, and natural structures

such as polysaccharides, proteins, and even human tissue are its masterpieces of polymer


With functions and properties as essential as those exhibited by natural polymers,

a parallel exploration of sources and targets for synthetic polymers was to be expected at

a point of chemical maturity in the development of scientific knowledge. With a little

help from serendipity, several polymerization schemes towards synthetic polymers were

developed following the acceptance of macromolecular theory, and as a result synthetic

polymers can be found in virtually every aspect of modern life.

This dissertation can be contemplated within the general field of polymer

synthesis from both of the approaches described above, namely the synthesis of new

materials and the exploration of alternatives to known polymerization processes.

Addition and Condensation Polymerization, Chain and Step Polymerization

Macromolecule Formation: The Concepts

The terms addition and condensation refer to the propagation reaction, and

describe the type of reaction involved in linking molecules for polymer formation. By

classifying a polymerization reaction as addition or condensation, we are making use of

the same concepts used in organic chemistry. Some straightforward examples of addition

reactions are the hydrosilation of olefins, cycloadditions (e.g., Diels-Alder) and

hydrogenation of alkynes. The most relevant characteristic of addition reactions to our

discussion is the fact that all of the fragments present in the reactant molecules can be

found in the product. No formation of byproducts is observed. In a similar way,

recurrence to some organic reactions classified as condensations is a helpful tool, and

reactions such as esterifications and some coupling reactions (e.g., McMurray) exemplify

this concept properly. In this case, a byproduct (condensate) is removed from the intended

product (either by itself or as a conjugate).

Analogously, the terms chain and step also refer to the mode of propagation of a

certain polymerization but describes how macromolecules grow from monomers. In the

literature since as early as 1927 (Carothers 1927), this is arguably the most important

classification of polymerization processes. It is essential for the understanding of not only

the mechanistic details of a given polymerization scheme, but also for the effective

prediction of some of the physical properties and possible applications of polymeric

materials. It is imperative to describe and compare the main features of these two

different propagation modes prior to the discussions of polymerization processes included

in further sections of this work. For a comprehensive presentation of step and chain

polymerization see Odian (1991).

In chain polymerization, an active species (e.g., a cation, a free radical) is

generated in the first reaction called initiation. This active species reacts with the first

monomer unit that gets incorporated, and a new active entity is generated. The reaction of

this new species is the basic propagation phenomenon, and subsequent reactions of the

growing chain with monomers will generate species with similar reactivities to that of the

prior propagating species. For the sake of clarity, other important phenomena in chain

polymerization (chain transfer, termination) will not be discussed.

In step polymerization no initiation step needs to be defined and with few

exceptions, propagation reactions mark the beginning of the polymerization scheme. The

first propagation event involves the reaction between two monomer units to yield a dimer,

a molecule of increased size which contains the same functional groups that were present

in the monomer, available for further reaction. Reactions of this dimer with another

molecule of monomer yields a trimer, and the polymer size increases slowly as the

reaction proceeds. It is safe to assume that all of the species present in the reaction

mixture at any given time exhibit the same functional group reactivity towards each other;

this identical reactivity ensures the growth of the polymer after every propagation step

produces a new molecule. A general scheme for step and chain (initiation and)

propagation reactions is shown in Figure 1-1.

Chain Step

I k 1* 1 A-R-B + A-R-B A-R-R-B

1* + M k, I-M* A-R-R-B + A-R-B A-R3-B

I-M* + M -M* A-R2-B + A-R2-B kp A-R4-B
A-R2-B + A-R2-B A-R4-B

k k
I-M*_ + M I--M* A-RT-B + A-R-B --p A-R- B

Figure 1-1. General schemes for propagation in chain and step polymerization.

The first evident and important difference between the two schemes is the time at

which monomer disappears from the reaction. While in chain polymerization monomer

units are present in the reaction mixture until the polymerization is completed, they

disappear at very early stages in step polymerization. Analysis of a polymerizing mixture

following a chain process reveals that monomer and (usually high molecular weight)

polymer are present at any given time, while in the step polymerization case, the mixture

consists of exclusively molecules of low to medium molecular weight. This is a

characteristic feature of the polymerization mechanisms, and is often used to distinguish

between the two.

The average chain length is another significant difference between chain and step

polymerization. High molecular weight polymer is formed at low % conversion in chain

polymerization, while only until very high functional group conversion can it be found in

a step process. This fact prompts for very specific experimental conditions for each

process, and often has a significant effect on the molecular weight of the final product. It

also suggests that since very high percentages of functional group conversion are needed

in order to reach high molecular weight in step polymers, only a few organic reactions

proceeding almost quantitatively can be used in the synthesis of macromolecules via step


Among the organic reactions used in the synthesis of step polymers, condensation

reactions are very common. In fact, most of the processes employed for step

polymerizations involve condensations in their propagation reactions, and this stands as

the reason why the terms condensation and step polymers are often -and sometimes

carelessly- interchanged. The rationale behind the application of condensations in step

polymerization is simple: condensation reactions are often equilibrium processes, and the

removal of the condensate from the mixture ensures high conversion through shifting the

equilibrium towards product.

From Small to Large Molecules: Functionality

The concept of functionality (f) in polymer chemistry helps understand the

transition from small to large molecules. Functionality can be seen as an index, which

describes the maximum number of possible connections a substrate can make when

involved in the polymerization reaction, whether the mechanism is chain or step

polymerization. In order to yield a linear polymer, the value of f must be at least 2 for all

the monomers in a specific polymerization scheme. As an example consider benzoic acid

and methanol in an esterification reaction. Benzoic acid has only one carboxyl group

which can be involved in the esterification: only one ester linkage could be formed with a

molecule of benzoic acid (f=1). It is the same case for methanol: one is the maximum

number of ester groups that can be formed in its esterification (f=1). As a consequence,

the reaction of these two species will yield a small molecule (methyl benzoate) and not a

polymer. The same analysis can be performed on terephthalic acid (f=2) and methanol (f

=1), and it reveals that the product (dimethyl terephthalate) is not a polymeric structure.

The number of ester linkages was given by the maximum value of f found. However,

when two difunctional molecules (f =2) are reacted (e.g., terephthalic acid and

ethyleneglycol) at very high functional group conversion, the result is a large

macromolecule. This is the case for polymerization of hydroxyacids which contain one

alcohol and one carboxyl functional group in the same molecule, yielding a difunctional

monomer (f =2).

As mentioned before, a molecule's functionality depends on the polymerization

scheme in which it is considered. For addition polymers (e.g., the free radical

polymerization of styrene) the olefin bond has a functionality of 2, since two new

covalentt) bonds can be formed upon its reaction.

From the analysis described above, it is easy to understand the effects of adding

monomers with f values other than 2 to polymerization mixture. For example, the

reaction of a growing chain with a monofunctional molecule causes the chain growth to

stop by incorporating a fragment contained in the monofunctional molecule at the chain

termini. This monofunctional adduct is known as a chain limiter, and the technology

based on its usage is widely used as a molecular weight control procedure. In a similar

way, polymerizing systems with average functionalities greater than 2 causes branching

and the inter-chain linkages are responsible for the formation of network and crosslinked

structures. This phenomenon can be used in a systematic way, and synthetic strategies

towards interesting polymer architectures such as star polymers can be devised. Odian

(1991) presents an interesting and comprehensive treatise on functionality and its effect

on polymer synthesis.

Olefin Metathesis

One of the most attractive transition metal-catalyzed processes available to

today's chemist is olefin metathesis. Modem synthesis strategies of specialty polymers,

fine chemicals, natural products, and even protein mimics reported in the last decade are

just some representative examples of the fields in which olefin metathesis is finding its


Several (almost simultaneous) observations led the way to realizing that olefin

metathesis was the reaction operating in a series of chemical processes reported in the

chemical literature between 1960 and 1964 (Truett 1960, Banks 1964), with precedent

reports as early as 1957 in the patent literature (Eleuterio 1957). The pioneering work of

Calderon (1967,a,b) is attributed to be fundamental to this realization, as well as to

coining the expression olefin metathesis to describe such reaction.

In a general way, olefin metathesis can be described as the apparent exchange of

carbon moieties between a pair of double bonds in two alkene molecules. This is a true

equilibrium process, which in some cases can be attained in seconds. A simple example is

illustrated in Figure 1-2.

\- R
R ) +
1 2 3

Figure 1-2. Olefin metathesis of a terminal alkene.

Although elusive at first, a consistent mechanistic scheme known today as the

carbene/metallacyclobutane mechanism, was proposed by Herrison and Chauvin in 1971

(Herrison 1971). This proposal suggests that the active catalytic species is a metal carbene

(an organometallic species containing a metal-carbon double bond), which mediates the

exchange of the alkylidene portions between the two reactant olefins. The most important

step in Chauvin's proposal, consists in the formation of a metallacyclobutane

intermediate from the reaction of the metal carbene and an olefin. As an illustration, the

mechanism by which the conversion depicted in Figure 1-2 proceeds is shown in Figure


LnM= L M

R 3
4 RR 6

2 R L~M MLn
8 7

Figure 1-3. The carbene/metallacyclobutane mechanism illustrating the metathesis of a
terminal alkene 1.

The mechanism shown in Figure 1-3, begins with the methylidene 4, which is

believed to be the actual catalytic species and is generated in previous steps of the

reaction (not shown). The reaction of a substrate molecule (1) with this alkylidene in a

fashion that resembles a [2+2] cycloaddition forms the metallacycle 6, a key intermediate

present in the mechanism of every olefin metathesis process. (The formation of 6 is

believed to proceed through the 7t-complex 5 (Tallarico 1997, Dias 1997) and the

presence of this intermediate in our mechanistic discussion will be of significant

importance in further sections of this work. In an analogous way, cleavage and formation

of every metallacycle may involve 7t-complexes such as 5, and their formation will be

implied for the rest of our discussion.) The productive cleavage of 6 yields ethylene and

the new alkylidene 7, which upon reaction with a second molecule of substrate forms

metallacycle 8 with a stereoselectivity mainly determined by the steric interactions

between the substituent R and the ligand set. The disproportionation of 8 produces the

new internal olefin and regenerates methylidene 4, which begins another catalytic cycle.

Several features are instrumental to understand both the outcome of metathesis

processes in terms of the final structure and properties of the reaction products and the

implications that both catalyst and substrate have on this mechanism. Ivin and Mol (Ivin

1997) have presented these in depth. Two key factors are of utmost importance in our

discussion: the equilibrium nature of every step has to be considered, and the formation of

the metallacyclobutane intermediates is considered to be the rate-determining step in

every olefin metathesis application.

Exchange Metathesis

The reaction depicted in Figure 1-2 is the most basic example of one type of olefin

metathesis known as exchange metathesis, in which two molecules of a terminal

(monosubstituted) olefin react to yield ethylene and the disubstituted olefin 2. This

reaction proceeds with ease due to the low steric constraints imposed by only one

substituent on the double bond, and it becomes of synthetic importance when the

disubstituted olefin obtained as one of the products is of greater value than the starting

material. The second olefin produced (ethylene) is in most cases treated as a byproduct,

because of its low cost and of being a gas at standard conditions, which facilitates its

removal from the reaction mixture ensuring not only essentially quantitative conversions

but also very high product purity. The stereochemical outcome is in most cases

determined by the type of catalyst used (geometry, ligand set, etc.,) as well as by the

extent of isomerization (also known as cis : trans equilibration) processes that may

follow product formation.

Exchange metathesis can also take place with two disubstituted olefins as

substrates. Its synthetic utility is given by the identity of the species formed in the

equilibrium, as well as by the ease of separation of the olefins formed. This is a field of

growing interest in the fields of fine chemicals and natural products.

The equilibration of more than one type of starting olefin is known as cross

nmelt/hei' Depending of the extent of substitution of the starting materials and

considering all the possible stereoisomers obtainable, this reaction may lead to a very

complex mixture of olefins coexisting at the equilibrium. This phenomenon which

appears to render the reaction of low synthetic value, makes it a valuable tool in the

construction of libraries for combinatorial chemistry (Brandli 1999, Giger 1999). Cross

metathesis is also used when one of the equilibrium components is an olefin of very high

value or otherwise very difficult to synthesize by other means. The exploration of

selectivity issues associated with both sterics and electronics of the reacting olefins is a

field of active research (Randall 1998).

Ring Closing Metathesis

Ring closing metathesis is arguably today's most important application of olefin

metathesis. The advances in the field have been recently reviewed by Armstrong (1998),

Furstner (1997), Grubbs (1998), Schuster (1997), and Wright (1999). It consists of the

reaction of two tethered olefins to yield carbo- and heterocyclic structures, mainly from

terminal dienes. A general scheme is shown in Figure 1-4.

Catalyst Q +

9 10

Figure 1-4. General scheme for ring closing metathesis (RCM)

Two factors of interest should be highlighted. First, in order to ensure an

intramolecular reaction the experimental conditions often include high dilution. Second,

the feasibility of a specific conversion must be analyzed taking into account kinetic and

thermodynamic parameters given by the ring size (and thus ring strain), the effect of

substituents, both from the steric and electronic point of view, and finally the

compatibility between the functional groups present in the starting diene and the catalyst


Ring Opening Metathesis Polymerization

As mentioned in the previous sections, olefin metathesis can be employed in the

synthesis of ring systems of different sizes by the proper choice of experimental

conditions and whenever the ring strain is sufficiently low (or virtually none) so that the

equilibrium shown in Figure 1-4 favors the cyclic species 10. Strained olefins

(represented by 11 in Figure 1-5) are engaged in a different type of chemical equilibrium

when exposed to a metathesis catalyst, and this process is known as ring opening

metathesis polymerization (ROMP), chronologically the first method of polymer

synthesis that involves olefin metathesis. The general scheme for the conversion of a

strained olefin into a polymer is shown in Figure 1-5.

n Catalyst_

__ n

11 12

Figure 1-5. General scheme for ring opening metathesis polymerization (ROMP)

Although an equilibrium in principle, it is the case for most ROMPs to be

considered irreversible. This is the consequence of a very favorable forward reaction due

to the release of ring strain, making the overall conversion an exothermic process. The

product of this chain polymerization is a macromolecule containing a double bond every

repeat unit. Because of its widespread applications and its mechanistic beauty, ROMP has

received a great deal of attention since its discovery in the fifties -in fact, some of the first

examples of olefin metathesis were polymerizations of cycloolefins. Examples of

common ROMP polymers include polynorbornene and polynorbornene derivatives and

polydicyclopentadiene (PDCPD), a thermoset resin of high volume production in the

United States.

The mechanism by which the ring opening metathesis polymerization of a

cycloolefin proceeds is shown in Figure 1-6. First, the interaction between the

cycloalkene 11 and a carbene (e.g., the methylidene 4) yields the fused bicyclic structure

13, which can be visualized as a substituted metallacyclobutane. When this ring cleaves

productively, breakage of the c bonds opens the two rings in a virtually irreversible step.

The ring strain is released and a new carbene is formed, this time at the termini of a linear

unsaturated structure (14). It is at this point where the chain nature of the process should

be noticed, since an active center is attaching at the chain end after the propagation step.

From this point of view, it is also useful to visualize the carbene used (e.g., 4) as an

initiator and not as a catalyst since its original structure is not regenerated after the

propagation cycle: a totally new chemical species is formed.


11 13


12 14

Figure 1-6. Mechanism of the ring opening metathesis polymerization (ROMP) of a
strained cycloolefin.

In summary, ROMP is a chain polymerization that involves the olefin metathesis

reaction in every propagation event and a metal carbene as the initiator. In depth

discussion of this topic can be found in Odian (1997), as well as in reviews by Grubbs

and Khosravi (Grubbs 1999), and Schrock (1990, 1993).

Acyclic Diene Metathesis Polymerization

ADMET (acyclic diene metathesis) polymerization is another method that

provides access to linear, unsaturated hydrocarbon polymers and materials containing a

wide variety of functionalities along the polymer backbone. Two recent reviews provide a

comprehensive view of the ADMET chemistry (Tindall 1998, Davidson 1999). In

general, ADMET polymerization can be described as the condensation of terminal dienes

into unsaturated polymers, yielding a molecule of ethylene in every propagation step.

Based on this description, ADMET falls in the category of step-propagation

condensation-type polymerization (Odian 1991).

R Catalyst R

Figure 1-7. General scheme for acyclic diene metathesis polymerization.

Both hydrocarbon and functionalized polymers can be produced following the

same synthetic scheme. As suggested by Figure 1-7, the incorporation of the functionality

takes place at the monomer design step, making of ADMET a valuable tool in polymer

synthesis. In addition, the rules that govern functional group tolerance as well as

structure-reactivity relationships have been explored throughout the years and while

obeyed, make useful monomers of a variety of terminal dienes. These have been used in

the construction of interesting macromolecular architectures through grafting,

hydrogenation, copolymerization and the condensation of telechelic dienes.

Mechanism and Kinetics

ADMET polymerization is an extension of an exchange metathesis reaction to

terminal dienes, in which they are perfectly difunctional monomers (f=2). Because of

being a step polymerization, ADMET requires high monomer purity and a single

mechanistic event operating throughout the reaction in order to achieve the high

conversion levels. Experimentally the polymerization is often run under bulk conditions

in order to maximize the concentration of reactive functional groups, an essential

requirement for the formation of high molecular weight polymer. These experimental

conditions suggest that high monomer conversion is often associated with a dramatic

increase in viscosity and with changes in the physical state of the polymerizing mixture

which, in turn, affect the kinetic profile of the reaction.

With these two features in mind, we can analyze the proposed mechanism by

which acyclic dienes are converted into unsaturated polymeric structures, illustrated in

Figure 1-8. Again, for the sake of simplicity, we have illustrated the alkylidene in its most

abbreviated form where the site LnM refers to a moiety containing M, a transition metal,

and a series of ligands (Ln). The initiation steps of the mechanism begin with the

interaction between a metal alkylidene (16, top right corner) and an olefin site within the

diene monomer (14). Consequently, a 7t-complex (17 or 18) forms between the metal

alkylidene and the olefin site, which then collapses to a metallacyclobutane; both steps

being reversible and thermoneutral in essence. Because of the two possible types of

addition (ua, in which the metal attaches to the terminal carbon of the double bond; or f3,

in which the metal attaches to the internal carbon of the olefin) two possible

metallacyclobutanes (19 and 20) can be formed. Next, either metallacyclobutane

LnM //\R/ 14 R6


L~.. \\J


21 R R

LnM=CH2 14 7


R LnM v 14 LnM 23 (
14 RR 14



R \ 1
9\ 10

Figure 1-8. Mechanism for acyclic diene metathesis polymerization of a diene 14.

ring disproportionate and in the case of oa-addition, the new alkylidene is the

methylidene 4, eliminating part of the original complex as a new diene (21). In the case of

P-addition, the metallacycle 20 cleaves to eliminate a new alkene containing a portion of

the original complex (22), and placing the transition metal on the end of a propagating

monomer structure (23).

It is at this point that the step condensation cycle (in the dotted inset) begins where

another diene monomer (14) interacts with either alkylidene as has just been previously

described: a 7t-complex is formed, rearranges into a metallacyclobutane, disproportionate

eliminating either the condensate (3) or the propagating polymer chain (15), at the same


LnM{ T c

Ln -_7 Rc

flR 20

time regenerating either structure 4 or 23, which continues the cycle. As shown in Figure

1-8, ethylene is ultimately evolved as the condensate in this reaction when cL,c0-dienes are

used, and polymer growth continues in a stepwise manner.

There is ample evidence for the validity of this mechanism. For example, the

kinetics have been shown to be second order in monomer as is often observed in step

polymerization chemistry (Wagener 1997b). Further, the molecular weight distribution

that is obtained is in keeping with that statistically expected, i.e., a molecular weight

distribution of 2.0 is often observed. In addition, steric factors play an important role, for

if the olefin involved in 7t-complexation is sterically crowded, then the kinetics of the

reactions are either dramatically slowed or polymerization is not observed at all. Finally,

the presence of functional groups can alter the rates of reaction, again for reasons related

to both the modification of the electronic properties of the transition metal found on the

propagating monomer (i.e. structure 23 in Figure 1-8) and to inter- as well as

intramolecular direct interactions between such functionality and the metal center (mostly

of Lewis acid-base type).

From an enthalpic point of view (and as implied by the overall reaction shown in

Figure 1-7) ADIIMET polymerization is essentially thermoneutral, making of this process

an entropy-driven transformation. The observed kinetic parameters of ADMET

polymerization strongly support the accepted mechanistic description. The reaction is

second order in monomer resembling in this way other polycondensation reactions such

as nylon or polyester synthesis (Wagener 1997b).

Functionality and ADMET Polymers

A wide variety of functionalities can be incorporated into ADMET polymers; and

monomer design plays a crucial role in determining not only the feasibility of ADMET

polymerization, but also the properties of the final product.

Ideally, any terminal diene can be homo- or copolymerized using ADMET

chemistry. However, all of the transition metal complexes used in metathesis catalytic

systems are inherently Lewis-acidic, and acid-base interactions may be of significant

importance in terms of catalyst deactivation. Because of this, compatibility between the

catalyst and the functional group contained in the monomer must be ensured through the

exploration of different experimental conditions.

Another important factor to be considered is what is known as the negative

neighboring group effect, which can be visualized as differences in reactivity exhibited by

monomers that differ exclusively in the spacing between the reactive olefin and the

functional group. This effect has been studied for a number of functionalities as well as

quantified through kinetic studies (Wagener 1997b).

Last, of significant importance in any step polymerization is monomer purity.

Besides the described effect of basic functionalities in catalyst decomposition, the

presence of impurities containing groups that may become involved in metathesis

chemistry (e.g. monoenes, alkynes) is detrimental to achieving high molecular weight

polymer, and may lead to ill-defined structures.

Ph Ph
N, R
0 0" ,

O~ O O Ar

S r

Figure 1-9. Representative structures of ADMET polymers synthesized to date.

The study of hydrocarbon monomers, both aliphatic and aromatic, has led to the

understanding of many features of ADMET chemistry. Pure hydrocarbon monomers such

as 1,9-decadiene are readily available and have been used as the standard for reactivity

and compatibility mainly because of the absence of basic functionalities and their ease of


Among the most interesting hydrocarbon ADMET polymers is 1,4-polybutadiene,

obtained through the condensation of 1,5-hexadiene. ADMET polymerization offers the

possibility of molecular weight control as well as easy end functionalization of this useful

material. Other hydrocarbon ADMET polymers of interest include

poly(p-phenylene)octylene and fully conjugated oligomers such as polyacetylene (Tao

1994) and poly(1,4-paraphenylenevinylene) (PPV) (Thorn-Csanyi 1999).

Several oxygenated functionalities have been incorporated in unsaturated

polymers via ADMET. Among these, the synthesis and light induced crosslinking of

polyethers, as well as the direct synthesis of polyalcohols via the condensation of

hydroxydienes are of special importance from the materials point of view. ADMET

polymerization can also provide novel synthetic approaches to other conventional step

polymers such as polycarbonate and polyesters. The utility of ADMET in the synthesis of

otherwise hard to synthesize polymeric structures has been demonstrated in the synthesis

of polymers bearing ketone and very especially, acetal functional groups. The liquid

crystal properties of ADMET polyester-ethers containing mesogenic groups along the

main chain have been explored, and constitute a very promising field for applications of

ADMET polymers (Walba 1996).

The efficient incorporation of other heteroatoms along the backbone of

unsaturated polymers via ADMET has led to a diversity of materials. For example,

boronate- and amine-containing dienes have been efficiently converted to ADMET

polymers. Polymers that contain silicon have been thoroughly explored, and constitute

today an area of very active research towards useful materials. These include

polycarbosilanes and polycarbosiloxanes, the latter being silicones with amount of

hydrocarbon content along the backbone that are variable through monomer design,

copolymerization and polymer modification. Another type of silicon-containing polymers

synthesized using ADMET chemistry are polycarbochlorosilanes. This type of polymers

are of special interest because of the possibility of accessing a variety of materials

through polymer modification, mainly using macromolecular nucleophilic substitution.

Architectures and Modeling with ADMET

As mentioned earlier, monomer design plays a pivotal role in ADMET

polymerization. Some structural features, polymer properties, and possible applications

may be inferred as early as monomer conception and synthesis. Such predictability allows

the use of ADMET chemistry in the synthesis of interesting macromolecular

architectures, as well as in the rational preparation of models for other polymerization


Among the different types of known copolymer structures, segmented copolymers

are easily accessed using step-propagation chemistry. ADMET polymerization has been

used in the preparation of this type of materials, which may exhibit unique properties

such as microphase separation, ampiphilic behavior or increased blend compatibility. The

synthetic scheme involves the design of diene-terminated oligomers (diene telechelomers)

which can be homo- or copolymerized with other terminal dienes, either mono- or

oligomeric. Representative examples of diene telechelomers include cL,c0-diene

terminated polytetrahydrofuran and polyisobutylene, two segments based on polymers

with low glass transition temperatures. These telechelomers have been efficiently

incorporated as the soft phase in copolymers with hard polyalkenylene, polyester and

polyurethane segments (Wagener 1997a, Tindall 1999).

Another application is the synthesis of polyalkenylene telechelics by end

functionalization of ADMET polymers through the addition of a chain limiter. Using this

strategy, chlorosilane- and methoxysilane-terminated polyoctenylene has been

synthesized and successfully copolymerized with hydroxy-terminated silicones yielding

an ABA triblock copolymer (Brzezinska 1999). Along the same lines, acetoxy-terminated

polyoctenylene and 1,4-polybutadiene has been synthesized using an unsaturated ester for

functionalization (Nubel 1994). This material is a precursor for both ester- and

hydroxy-terminated polyethylene, a building block of special importance in step

polymerization (and very especially in polyester technology) which can be isolated upon

hydrogenation (and hydrolysis) of the parent unsaturated polymer.

R + -V! G


L J n

z Grafting

An Z


x X X

Other polymerization methods,
olefin end-functionalization


, Grafting


x x!

Figure 1-10. Examples of polymer architectures using ADMET polymerization.

Finally, the use of functionalized dienes as initiators for other type of

polymerizations yields diene macromonomers which upon homo- or copolymerization,

produce graft copolymers via the combination of two mechanisms. The ultimate

composition and some of the architectural properties of such copolymers can be predicted

by the choice of comonomer, the feed ratio, and the structure present in the diene


Block Copolymers


Graft Copolymers

Polyoctenylene, the ADMET polymer produced from 1,9-decadiene, is a linear

structure that contains a double bond every repeat unit separated from each other by six

methylene groups. Hydrogenation of this polymer provides access to perfectly linear

polyethylene. Structures of this type are step polymers that can be used to model polymers

produced by chain chemistry.

Modeling polyolefin crystallization behavior using ADMET chemistry is currently

an area of extensive research. The polymerization of symmetrical dienes containing an

alkyl substituent (i.e. methyl) and its subsequent hydrogenation yields polyethylene with

alkyl branches separated by a specific number of methylene groups. This strategy allows

the side chain and the spacing between branches to be defined by the monomer structure.

The case of methyl-substituted dienes is of significance since it is a model for a

poly(ethylene-co-propylene) in which the frequency of propylene incorporation is


Along the same lines, ADMET polymerization serves as a route to precise models

of copolymers of ethylene with a variety of olefins such as vinyl chloride, styrene, methyl

acrylate and vinyl acetate, through the design and polymerization of dienes containing the

corresponding substituent. Furthermore, the polymerization of dienes containing a

reactive group such as hydroxyl or carboxylate facilitates the synthesis of graft

copolymers with precise placement of the grafted segments.

As depicted in Figure 1-7, ADMET polymerization is an overall equilibrium. This

suggests that once the experimental conditions are found, exposure of an unsaturated

polymer to an excess olefin in the presence of a metathesis catalyst (ensuring 100 %

olefin conversion) should afford a low molecular weight diene. In the case of ethylene

being used as the depolymerizing olefin, the theoretical product is the monomer.

S R +Z{Z
S R n

Metathesis Z = H, CH2-SiMe3

Figure 1-11. ADMET depolymerization.

ADMET chemistry is, in fact, an opportunity of recycling unsaturated polymers.

Several studies have demonstrated the feasibility of ADMET depolymerization, leading

to both functionalized telechelics or oligomeric terminal dienes. When this

depolymerization scheme is applied to polyalkenylenes such as 1,4-polybutadiene, it

becomes a new route towards telechelic polyolefins (e.g., polyethylene). Since the same

reactivity rules apply for the reverse reaction, depolymerization of functionalized

unsaturated polymers can also take place. ADMET depolymerization has been observed

with allylsilanes and ethylene as depolymerizing agents (Watson 1999, Marmo 1997)


Metal Carbenes and Olefin Metathesis Catalysts

Organometallic complexes containing a metal-carbon double bond are referred to

as metal carbenes. The classification of metal carbenes is based on the identity of the

atoms attached to the carbene carbon, which in many cases has a dramatic influence on

the electronic properties (and thus the reactivity) of the complex itself. If such carbon

bears a heteroatom, the name metal carbene or Fischer carbene is given to the complex.

On the other hand, if only carbon or hydrogen is attached, the complex is designated as a

metal alkylidene or a Schrock carbene Figure 2-1. Although the reactivity of the complex

is determined by a collection of factors such as the electronic properties of the ligands and

the type of metal-ligand interaction they exhibit, it is often observed that the carbene

ligand in a Fisher carbene shows electrophilic behavior, while nucleophilic behavior is

observed in Schrock carbenes. Today these features are well known and the reactivity of a

determined complex can usually be rationally modified by the alteration of its ligand set

(Elsenbroich 1992). A detailed comparison of the properties of Fisher and Schrock's

carbenes is given in Crabtree (1994).

6- X 6+ R
M in a low oxidation state L, M=C+ L M=/C_ M in high oxidation states
X = amines, alkoxides L \H \H R = alkyl, aryl, silyl

24 25

Figure 2-1. Structures of a Fischer (24) and a Schrock (25) carbene.

Early examples of olefin metathesis reactions employed ill-defined catalyst

systems (Ivin 1997). Although these catalysts demonstrated very early their utility in

metathesis chemistry, the mechanism was not only poorly understood but also difficult to

study, mainly because of the heterogeneous nature of such systems. In 1971, Herrison and

Chauvin were the first to propose metal carbenes as intermediates in the olefin metathesis

mechanism, the same mechanistic proposal that is widely accepted today (Herrison 1971).

A number of transition metal carbenes are known to be active catalysts for olefin

metathesis (Ivin 1997). These catalytic systems have been classified as well-defined or

classical catalysts based on whether the carbene moiety is present from the beginning of

the reaction or formed in situ by the reaction of aprecatalyst with an activator. While all

of the well-defined systems are soluble and have been thoroughly characterized, several

examples of heterogeneous classical systems exist in the literature. These systems have

been used extensively in the metathesis of olefins at elevated temperatures, and although

much less is known about the mechanism by which they are converted to a metal carbene,

supported metathesis catalysts continue to be the best alternative in industrial applications

because of their robustness and reusability, albeit the side-reactions often induced by

these ill-defined systems (Ivin 1983). However, since we are concerned with the issues

surrounding the applicability of classical catalysts to ADIIMET chemistry, we will

concentrate our discussion on soluble classical systems. Because of the mechanistic and

kinetic features already described (Chapter 1), only the most active and tolerant

metathesis catalysts can be used in acyclic olefin metathesis and hence ADIMIET

polymerization, but more importantly, only systems capable of polymerizing dienes in a

clean fashion can be considered as alternative ADIIMET catalysts.

Well-Defined ADMET Catalysts

It was noticed early that the electronic properties of Fischer carbenes

(coordinatively and electronically saturated metals in most cases) were not correct in

order to induce most metathesis reactions, and not until the advent of the Schrock

carbenes did metathesis chemistry using well-defined catalysts prove to be so remarkably


The isolation and characterization of the first alkylidene complexes led to early

studies on olefin metathesis catalysis by well-defined systems. However, very poor

reactivities were found, and the steady development of very active alkylidenes occurred

parallel to major discoveries in the area of classical systems. In fact, the Schrock catalysts

26 and 27 illustrated in Figure 2-2 are the result of a lengthy period of research examining

both steric and electronic issues (Schrock 1988, 1990a, Feldman 1991). motivated in part

by the rapid evolution of alkylidyne complexes (Schrock 1983, Sharp 1981). While

several transition metals have been explored, the molybdenum and tungsten versions of

Schrock catalysts have received the most attention and, in fact the use of Schrock

alkylidenes led to the first successful ADMET reaction (Wagener 1991).

F30 II
F3C O '"M=CH
H3C 0 CH3
F3C ICH3 26 M = Mo
CF3 27 M = W

Figure 2-2. Two examples of Schrock carbenes, active catalysts for olefin metathesis.

Several features contribute to the high reactivity of complexes such as 26 and 27

towards olefin metathesis. First, the metal is in a high oxidation state (+6) and bears

electron withdrawing ligands such as fluorinated alkoxides. This combination results in a

very electropositive metal, which favors the metal-olefin interaction previously described

as imperative (Chapter 1). At the same time, stabilization of the metal is a consequence of

the 7t-donor ability of oxygenated and nitrogenated ligands, and especially of the

bulkiness of all the ligands present, which prevents the decomposition of the catalyst by

for example, bimetallic pathways. Finally, the complex is coordinatively unsaturated (a

pseudo-tetrahedron) which allows the first event in the mechanism: olefin coordination.

The ligand set on complexes 26 and 27 is the product of a careful exploration of the effect

each ligand has on the complex activity, the results of which have set the guidelines for

olefin metathesis catalysts design.

While Schrock's catalyst presented the opportunity to be definitive in terms of

structure reactivity relationship for a number of ADMET monomers, it is an air-sensitive

compound that requires handling under stringent conditions. Consequently the

opportunity existed for the creation of catalyst structures more amenable to

polymerization and workup procedures.

It was with this in mind that the development of ruthenium-based catalyst systems

moved from their use in more classical catalyst arrangements to well-defined ones. In the

early 1990's Grubbs and coworkers were able to generate well-defined ruthenium catalyst

structures such as those illustrated in Figure 2-3 (Nguyen 1993, Schwab 1996, Dias


PCy3 PCy3
C Y3 I Cl,,. I u
CIuCH Ru-=

PCy3 C1 PCy3

28 29

Figure 2-3. Two examples of ruthenium carbenes, well-defined catalysts for olefin

Note that these indeed are well-defined metathesis catalyst structures where again

tuning appropriate molecular features leads to higher degrees of reactivity. More

importantly, these catalyst structures are active under conditions requiring less care in

handling. These catalysts can actually be handled in air, and though kinetically they are

approximately an order of magnitude slower than those observed for the Schrock

alkylidenes, they still are quite useful systems.

The activity of the ruthenium system 28 in the metathesis of acyclic olefins can be

explained following a different rationale than that of the Mo and W alkylidenes. The

ruthenium systems are different in terms of their electronic properties: they are not in high

oxidation states, and only halogens have been included as electron withdrawing ligands.

Also, two phosphine ligands -excellent y-donors- are present in the coordination sphere

of the complex. A closer look at the trigonal-bipyramidal complex 28 reveals that the two

phosphine ligands are coordinated in the apical positions, and that they are responsible for

the steric bulk around the metal center.

In an effort to fine-tune the activity of ruthenium based systems, Grubbs and

coworkers have carried out comprehensive studies of the effect of the ligand identity on

the catalytic activity of a series of complexes, which revealed very interesting features of

these systems. (Dias 1997, Schwab 1996) For example, activities decrease when moving

from Cl to Br to I, an observation explained in terms of the trans effect of the halides as

well as of the steric effect on olefin coordination and geometric rearrangement. It was

also found that modification of the electronic properties of the phosphine ligand had a

more dramatic effect than steric modifications; and even more surprising, that the activity

towards acyclic alkene metathesis increased parallel to the electron donating ability of the

phosphines (Dias 1997). The mechanistic proposal for ring closing metathesis derived by

Grubbs reveals intimate details of these systems that can be extrapolated to ADMET

polymerization, and constitutes the beginning of a new school of thought within the area

of ruthenium carbenes as catalysts for ADMET.

In summary, two types of well-defined catalysts have been successfully employed

in ADMET polymerization: the Schrock-type alkylidenes (26 and 27), and the

Grubbs-type Ru benzylidene 28. Both types are useful ADMET catalysts and each type

exhibits unique properties that make it suitable for specific monomer or polymer systems.

The advantages of Schrock's systems over complexes like 26 are their higher activity

(faster polymerization kinetics), their higher reactivity towards substituted olefins, and the

endless possibilities for fine-tuning through ligand modification. At the same time,

Grubbs' carbenes are synthesized in a very convenient and straightforward manner, which

makes them the system of choice for many large-scale applications in polymer as well as

in organic synthetic chemistry. Furthermore, these systems display high air and moisture

resistance, as well as a superb tolerance to polar functionalities.

Olefin Metathesis Classical Catalytic Systems

The first olefin metathesis reactions were performed by the use of ill-defined

catalysts, and the term classical has been commonly given to systems that fall into this

category. By classical catalytic system we designate the mixture of a transition metal

complex with an activator (also known as cocatalyst), which react to form a metal

carbene in situ. These types of systems have been used in many applications of olefin

metathesis, using either cyclic or acyclic substrates (Ivin 1997). The ease of synthesis and

the robustness often exhibited by the transition metal complexes used as precatalysts

make the classical systems of special interest, and some of these features are often

considered advantageous for certain types of metathesis processes, especially at an

industrial level where stringent conditions are undesired.

Transition Metal Halides and Alkoxides

Among the wide variety of soluble classical catalysts, halides and alkoxides of W,

Mo, Ta, Nb, and Re, activated by main group alkyls constitute a subgroup of special

interest. In fact, one of the most common catalytic systems consists of binary and ternary

mixtures containing WCl6, aluminum or tin-based activators, and sometimes other

organic additives. Today, numerous of applications of metathesis chemistry are possible

because of its availability, low cost and high efficiency. Some examples include the ring

opening metathesis polymerization of dicyclopentadiene, carried out at a commercial

scale using EtAlCl2 as the activator, as well as the synthesis of polyacetylene derivatives

involving Me4Sn and other tin alkyls as cocatalysts (Balcar 1994, 1995, Masuda 1986,


In order to arrive at an agreement between the olefin metathesis catalytic cycle,

and the identity of the ill-defined systems described, the activation mechanism by which

metal halides are converted into metathesis catalyst should lead to the formation of a

metal carbene. Figure 2-4 depicts a plausible activation route for the WCl6/Me4Sn system,

as proposed by Thorn-Csanyi (1985). The first event in the mechanism is the double

alkylation of the tungsten occurring through a transmetallation reaction, yielding a

transition metal dialkyl complex (30). The existence of such an intermediate (although

not always isolated or even observed) is in concert with the experimental data (Ivin

1987), and the extent of this reaction reflects in many cases the nucleophilicity (or

alkylating ability) of the alkyl groups present in the main group metal complex (Ivin

1983). In the second step, the dialkyl complex 30, is believed to undergo a (usually

thermally induced) ce-abstraction reaction, eliminating the metal carbene 31 and an alkane

molecule. This second part of the mechanism is in good agreement with the well known

use of dialkyl complexes as precursors to alkylidenes, elegantly demonstrated and

extensively supported in the literature (Schrock 1974, 1975, Feldman 1991, Elsenbroich


Cl Cl
ClI,, I \CI Cli,, I \CH3
WC C + 2 (CH3)4Sn C- CH3 + 2 (CH3)3SnCI

CI/, I \CH3 cil,,

30 31

Figure 2-4. Proposed mechanism for the conversion of WCl6 into a tungsten carbene.

A few characteristics of complex 31 should be mentioned. It not only contains the

carbene moiety but also the tungsten atom in its highest oxidation state (+6). The latter

fact and the available coordination site ensure the feasibility of the first mechanistic event

in the olefin metathesis catalysis cycle: olefin coordination.

As noted by Ivin (1987) only a portion of the initial precatalyst is converted into a

carbene, and only part of the formed carbene is responsible for the olefin metathesis

catalysis. This finding was extrapolated from the calculated degree of polymerization of a

sample of polynorbornene obtained when a 1:2:1 mixture of WC16/(13CH3)4Sn/NBE was

reacted in an NMR experiment. A low extent of activation (as low as 0.7%) was

observed, and this behavior can be expected for other similar systems.

In turn, organic compounds used as additives play a significant role in ternary

catalytic systems. In some of the early examples the presence of a Lewis base

(phosphines, ethers, nitrogenated bases) as the third component of the catalytic mixture

resulted in increased stabilities with respect to the binary systems known, and this effect

can be rationalized from an electronics point of view (Ivin 1997). The protection of the

carbenes formed seems to have an observable effect on the apparent extent of conversion

to (or the half-life of the) active catalyst (Masuda 1992).

The observed improvements in catalyst stability resulted in the introduction of

protic Lewis bases, among which alcohols and phenols are well known examples.

However, the reactivity of early transition metal halides with alcohols and water suggests

that the formation of a new type of catalytic species was indeed taking place. Reasonable

structures for the precatalysts formed include metal alkoxides (and phenoxides),

generated though halide substitution reaction schemes.

MX, + y ROH Xn-yM(OR)y + y HX

Figure 2-5. Metal alkoxides from metal halides.

The generation of alkoxide (or phenoxide) species then was proposed, and soon

after the pioneering work of Dodd (1982) became the first report on olefin metathesis

catalysis by systems generated from (tungsten) phenoxide precursors, thereby opening a

door to systems with a many advantages with respect to their parent metal halides.

Among these, phenoxide derivatives of WCl6, WOCl4, and MoOCl4 have found utility in

the metathesis of acyclic olefins, both terminal (Dodd 1988) and internal (Quignard 1985,

1986), ring opening metathesis polymerization (Bell 1992, Barnes 1994, Dietz 1993), the

polymerization of substituted acetylenes to yield 7t-conjugated structures (Nakayama

1993, 1996) and the ring closing metathesis of a variety of dienes to yield carbo-and

heterocyclic structures (Nugent 1995). A summary of these applications is presented in

Table 2-A.

The most notable consequences of halide substitution by an aryloxide group are -

besides the already mentioned increase in stability/half-life- the modification of the

electronic properties of both the precursor complex (and the aimed catalytic species

synthesized therefrom) as well as an increased compatibility with organic substrates, an

especially important feature in solvent-free processes. The third effect is related to the

electronics of the complex, since the electron withdrawing properties of the aryloxide

group are often different than those of halide ligands.

Classical Systems and ADMET Polymerization

As noted earlier, the original attempts to polymerize dienes by polycondensation

chemistry met with difficulty primarily due to the fact that the classical catalysts that had

been chosen (WCl6/EtAlCl2) resulted in more than one mechanistic event (e.g., Lewis

acid-induced crosslinking) (Lindmark-Hamberg 1987, Wagener 1990). This has been

attributed to the high Lewis acidity of the studied cocatalyst (EtAlCl2) and/or of the actual

catalytic species. Figure 2-6 illustrates a case of crosslinking via vinyl addition catalyzed

by a Lewis acid (LA) in an attempted ADMET polymerization.

Table 2-A. Some examples of metathesis reactions catalyzed by metal aryloxide-based
classical systems.

Transformation Conditionsa Reference

C006 6 C6 Et3Al2CI3 or SnR4, Dodd 1988
20 to 140 C

C1 02 1 1 0^C1 + 2 "^C2 WCl4(OAr)2, Quignard
and some fatty acid esters R4M, b 85 OC 1986

> _n ~WOC12(OArO), Dietz 1993
/ Et2AlC1, 0 OC Barnes 1994

/ ______ Crosslinked WOCl2(OAr)2,
polydicyclopentadiene Bu3SnH, 80 oC Bell 1992

---- Mt'Cl(dmp)y' Nakayama
Rt -A1Et3,
R -20 to 60 oC 1993

/i X WOCl2(dbp)2,c Nugent 1995
/+ / Et4Pb, 85 C Nugent1995

a) OAr represents any monodentate, substituted phenoxide, while OArO represents a
bidentate aryloxide (e.g., binaphtol derivatives). For specific examples, see the
corresponding references.
b) A series of tin and lead-based activators were explored.
c) dmp=2,6-dimethylphenoxide, dbp=2,6-dibromophenoxide.

Here the presence of a mixed catalyst system probably allowed cationic vinyl

addition chemistry at a rate competitive with metathesis polycondensation chemistry.

Thus, it was initially thought that classical catalyst systems would never serve much of a

purpose in condensing dienes to their polymers.








-4---- -'-^-^



Figure 2-6. A plausible mechanism for the Lewis acid-induced crosslinking of an
unsaturated polymer.

However, in 1994 Nubel, Lutman, and Yokelson demonstrated that a mixture of

WCl6 and Me4Sn in the presence of propyl acetate would produce linear

1,4-polybutadiene through the clean ADMET polymerization of 1,5-hexadiene (Nubel

1994). In this case the use of a Lewis base like propyl acetate was pointed as the key to

linear polymers, reducing the side reactions that could lead to crosslinked product, and

this condensation became the first example of a clean ADMET polymerization catalyzed

by a classical catalytic system.

WCl6/ Me4Sn / PrOAc
Tol. reflux


Figure 2-7. ADMET polymerization of 1,5-hexadiene.


Nubel's experiment prompted the exploration of other classical systems in order

to develop a flexible polymerization methodology, especially in light of the possibility of

applying such to the synthesis of high temperature ADMET polymers.

Tungsten Aryloxide Complexes and ADMET Polymerization

Three known aryloxide tungsten complexes were chosen for our initial studies on

the ADMET polymerization of hydrocarbon dienes: WOCl2(O-2,6-C3H6Br2)2 (32)

WCl4(O-2,6-C3H6Ph2)2 (33) and WCl4(O-2,6-C3H6Br2)2 (34), whose structures are shown

in Figure 2-8.

Br z -" Z
..W -,,,, 0
Cl o? i

32 Z
32 / 33 Z= Ph
34 Z = Br

Figure 2-8. Structure of complexes 32-34 studied as precursors to classical catalysts for
ADMET polymerization.

The polymerization studies were conducted using Bu4Sn, Me4Sn or Bu3SnH,

cocatalysts used in previous applications of systems 32-34 (Quignard 1986, Bell 1992).

ADMET Polymerization of 1,9-Decadiene

The exploration of tungsten aryloxide-based systems began in our hands with the

bulk polymerization of 1,9-decadiene, and the first binary system studied was a 1:2

mixture of complex 32 and Bu3SnH. Experimentally, a solution of complex 32 was

heated to 65 C under an argon atmosphere, and the hydride addition was performed

using a microsyringe. A distinctive color change of the reaction mixture, from dark red to

dark orange, was evident soon after the tungsten complex came in contact with the tin

hydride, and the release of a gas was suggested by small bubbles formed in the reaction

flask. As shown in Figure 2-4, one mole of alkane is produced in the ce-abstraction step

when the cocatalyst used is a tetraalkyltin complex. However, based on the differences of

the reacting species, a different activation mechanism for the activation using metal

hydrides has been proposed by Kelsey and coworkers (Kelsey 1997), and no efforts were

made for the identification of this initial gaseous byproduct. The steady gas evolution

under constant stirring can then be attributed to the release of ethylene, which is evidence

for the propagation stages of the polymerization. After ca. 4-6 hours of reaction, the

mixture was exposed to high vacuum in order to remove the condensate, and the reaction

was continued for ca. 2 days. After 48 hours of reaction, the viscosity of the mixture was

high enough to hinder magnetic agitation. The polymerization was now complete, and

crude solid polyoctenylene was obtained upon cooling of the reaction flask. The study of

other activators shows that a mixture of complex 32 and Bu4Sn catalyzes the

polymerization of 1,9-decadiene at 85-95 C, and a significant viscosity change can be

observed in a shorter period of time than when Bu3SnH is used as the activator. In a

similar fashion, tetramethyltin was also capable of activating complex 32 to generate an

active catalyst and the polymerization of 1,9-decadiene in the presence of 32 and Me4Sn

at 110-120 C also yielded polyoctenylene (35).

32, activator 4


Figure 2-9. Synthesis of polyoctenylene from 1,9-decadiene.

Soluble, high molecular weight polyoctenylene can be isolated after the workup

procedure in all cases, and no evidence for addition chemistry can be observed in any of

the characterization steps for these polymers. The quantitative 13C-NMR of the

precipitated polymers (e.g., Figure 2-10) shows the expected signals for polyoctenylene

and matches exactly the 13C-NMR of a sample of the same polymer synthesized using

either Schrock's molybdenum (26) or Grubbs' ruthenium (28) well-defined systems.

The experimental conditions described are the result of a series of reactions

carried out in order to establish the optimum polymerization conditions, varying

activation temperature and pressure conditions. The summarized results are shown in

Table 2-B. The determination of molecular weight was performed by gel permeation

chromatography (polystyrene standards) and by NMR end group analysis through the

integration of the olefin signals.

b d

a c

a (trans)

a (cis)

b (trans)

130 120 110 100 90 80 70 Ppm 50 40 30

Figure 2-10. 13C-NMR of polyoctenylene synthesized using complex 32 and Bu4Sn.

From comparative experiments, it was found that among the three complexes

studied Bu4Sn is apparently the most active cocatalyst for complex 32 in ADMET

polymerization. For example, high molecular weight polyoctenylene has been obtained

using this catalytic system in the polymerization of 1,9-decadiene, even in

polymerizations run at atmospheric pressure. (Table 2-B) However, the difference in

reaction times cannot be directly associated with a difference in activity of the catalytic

species reported in this study because of two factors. First, the three cocatalysts exhibit

different activation temperatures with respect to the same precatalyst (e.g., several failed

attempts to initiate the polymerization of 1,9-decadiene with complex 34 and Bu4Sn at

temperatures below 80 C confirmed this fact.) Secondly, this observation is based only

in the apparent changes in viscosity of the polymer obtained at different temperatures,

which does not constitute a reliable indication of the percent conversion.

Table 2-B. Experimental conditions for the polymerization of 1,9-decadiene with
complexes 32-34 and organotin activators.


32 + Bu3SnH

32 + Bu3SnH

32 + Bu4Sn

32 + Bu4Sn

32 + Bu4Sn

32 + Me4Sn

32 + Me4Sn

33 + Bu3SnH

33 + Bu4Sn

34 + Bu4Sn


65 C, 48 h

65 C, 72 h, 10-3 atm

65 oC, >100 h

85 oC, 48 h

85 oC, 72 h, 1 atm Ar

65 oC, > 72 h

120 oC, 36 h

100 C, 48h

85 oC, 24 h

85 oC, 24 h










1H-NMRc 13



No reaction observed



No reaction observed





a) Monomer/Precatalyst/Cocatalyst ratio = 500:1:2-2.5, pressure = ca. 1 x 10-5 atm.
b) Solvent: THF, calibrated using polystyrene standards.
c) Determined by integration of the terminal vinyl and internal olefin signals.
neg = no end groups observable.
d) High yields (75-97%) observed in all cases. Deviations from quantitative conversion
are due to monomer loss at high temperatures and are not reproducible.










ADMET Polymerization of 1,8-Nonadiene

Although several hydrocarbon and functionalized dienes have been polymerized

via ADMET chemistry, the polymerization of 1,8-nonadiene has not been reported

previously. In order to expand the scope of this polycondensation, we decided to study

the reaction of this diene not only with the classical systems 32-34 but also with the

well-defined system 26 under bulk conditions.

When 1,8-nonadiene is heated to 85 C in the presence of a 1:2 mixture of (e.g.,)

33 and Bu4Sn under an argon atmosphere, ethylene evolution is also observed for a

period of 8 h, and continues for ca. 48 h under reduced pressure. Again, the formation of

high molecular weight polymer hinders magnetic stirring of the reaction mixture and the

polymerization is stopped. Linear, soluble polyheptenylene (36) can be isolated in good

yields (77-87%) after dilution in toluene followed by precipitation into cold, stirring

acetone (Figure 2-11). This behavior resembles that exhibited by the polymerization

performed by Schrock's system 26, which also effectively catalyzes the polymerization of

1,8-nonadiene in a much faster reaction. Solid polyheptenylene can be isolated after 0.5 h

of reaction at room temperature followed by 1 h at 60 C, after which agitation becomes

impossible. After similar purification, a polymer with the same characteristics as 36 is


S b d In


7 6 5


c. d

1 ppm

Figure 2-11. 1H-NMR of polyheptenylene synthesized using complex 32 and Bu4Sn.

In the characterization of the polymers synthesized using the aryloxy-tungsten

systems 32-34, we find that the trans/cis ratio approaches 80:20 in all cases (determined

by integration of the olefin signals at e.g., 130.3 and 129.8 ppm in the quantitative

13C-NMR of polymer 35). This also is observed in the ADMET polymerization catalyzed

by the well-defined complexes 26-28 and represents the expected equilibrium distribution

of geometrical isomers, as a consequence of long reaction times. Similarly, the

polydispersity of all the polymers synthesized with systems 32-34 ranged between 1.7 and

2.3, a range in concert with the step propagation-type ADMET mechanism.

The Reaction of Other Hydrocarbon Dienes

The reaction of other unsubstituted terminal dienes with complexes 32 and 33 in

the presence of tetrabutyltin was studied in order to explore the behavior and selectivity

of the reaction towards cyclization in the bulk.

The reaction of 1,7-octadiene with a 1:2 mixture of complex 32 and Bu4Sn was

initially investigated under different experimental conditions. Although Feldman and

co-workers reported the synthesis of substituted cyclohexenes via the RCM of substituted

octadienes catalyzed by a mixture of complex 32 and Et4Pb in solution (Nugent 1995),

the attempted cyclization in the bulk could not be driven to completion. However,

analysis of the 1H- and 13C-NMR of the reaction mixture suggests not only the formation

of cyclohexene but also of low molecular weight oligomers; combined conversions are

<20 % in most of the cases. The highest % conversion (ca. 56 %) was observed with

complex 34 and Bu4Sn, which yielded a mixture consisting of ca. 55 % cyclohexene, 30

% linear oligomer and 10 % unreacted starting material.

On the other hand, 1,5-hexadiene has been previously utilized in the synthesis of

linear 1,4-polybutadiene under different experimental conditions that involve not only

well-defined metathesis catalysts but also classical systems (Brzezinska 1996, Nel 1989,

Nubel 1994). In a similar fashion, all of the attempts to polymerize 1,5-hexadiene were

unsuccessful, and no conversion to oligomeric product was evidenced in any case. In the

majority of cases, unreacted starting material could be isolated from the mixtures by

vacuum transfer. This can be attributed to an incomplete activation of the W species

because of the intramolecular coordination of the hexadiene. This phenomenon has been

proposed as the justification of differences in the polymerization kinetics of

1,9-decadiene and 1,5-hexadiene) and indirectly evidenced by Snapper and co-workers

through the isolation of a ruthenium alkylidene exhibiting intramolecular olefin

complexation (Tallarico 1997).

Investigations with Oxygen Functionalized Dienes

Feldman and coworkers' successful cyclization of dienes containing a variety of

functionalities using complex 32 activated by Et4Pb appeared promising for the synthesis

of functionalized ADMET polymers using classical systems (Nugent 1995). Based on

these results, the ADMET polymerization of two oxygen containing cL,co-dienes,

3-butenyl 4-pentenoate (36) and 9-decenyl 10-undecenoate (37) was investigated.

0 0

36 37

Figure 2-12. Structures of the two ester dienes 36 and 37.

Our studies of the ADMET polymerization of the esters 36 and 37 reinforce the

concept of negative neighboring group effect, introduced in Chapter 1 (Wagener 1993,

O'Gara 1993). While the 1H-NMR in the case of ester 36 shows no conversion to internal

olefin after being subjected to the general polymerization conditions, the ester 37 can be

efficiently converted to a polymer with a Mn of ca. 10,800 g/mol (determined by

end-group analysis from the integration of the signals corresponding to the terminal vs.

internal olefin in the 'H-NMR). GPC of the same sample in THF vs. polystyrene

standards yields a Mn of 22,000 g/mol) and this marked difference in reactivity can be

explained by the deactivation of the catalytic entity via intramolecular coordination of one

of the oxygens in the ester moiety in the case of ester 36 forming stable 5- or 6-membered

rings with the metal. Such deactivation has been previously observed in the case of

well-defined alkylidenes after productive metathesis with an acrylate ester (Feldman

1989) and has been suggested for complex 32 with other esters (Nugent 1995). However,

when we exposed the ester 37 to the general polymerization conditions described

elsewhere in this paper, polymerization takes place suggesting that chelation (and

deactivation derived therefrom) does not occur when enough methylene spacers are

placed between the basic functionality and the reactive double bond. The 1H-NMR of the

polymer produced in the ADMET polymerization of ester 37 is shown in Figure 2-13.


f b,e

7 6 5 4 3 2 I ppm

Figure 2-13. H-NMR of poly(carbonyloxy-9-octadecene) (38) synthesized using
complex 32 and Bu4Sn.


The scope of the ADMET methodology for the synthesis of unsaturated polymers

has been expanded by the exploration of aryloxy tungsten-based classical catalytic

systems. The most important feature of these systems is their ability to produce linear,

soluble polymers in a clean w/i',/the'i' process under bulk conditions, in comparison to

the classical systems previously studied (Lindmark-Hamberg 1987, Wagener 1990).

Three complexes were successfully used as the precatalytic entities while tetrabutyltin,

tetramethyltin and tri-n-butyltin hydride were used as activators. Tetrabutyltin was the

most efficient activator to complexes 32-34 but no significant difference in activity was

found among these three tungsten complexes. The experimental conditions explored in

this study not only allow high conversion to polymer, but make this methodology a very

promising alternative for the polymerization of novel monomer structures.

In fact, this polymerization scheme has already been suggested as the method of

choice in the synthesis of poly(p-phenylenebutylene), a semicrystalline polymer which

requires high polymerization temperatures (Steiger 1999).


A conscious analysis of the mechanism by which transition metal halides (and

derivatives) are activated by main group metal alkyls described in Chapter 2 suggests that

strong alkylating agents are required to activate the precatalytic species to a significant

extent. In fact, early homogeneous bicomponent systems included lithium, magnesium,

and aluminum organyls as cocatalysts, and their presence motivated research in more

practical cocatalysts at the expense of a lower extent of activation (Ivin 1997). The

introduction of tin and lead-based activators constituted a significant advance in terms of

portability and applicability of the olefin metathesis reaction to environments in which

very stringent conditions are undesired.

It is also reasonable to assume that an increase in the extent of the transmetallation

reaction depicted in Figure 2-4 may also be consequence of an increase in the

electrophilicity of the transition metal complex. This modification would result in higher

concentrations of the alkyl complexes that may in turn yield higher "true catalyst"


This chapter describes the efforts to modify the electronic properties of a series of

tungsten phenoxides through a rational approach. It also describes the search for useful

descriptors of the correlation between key ligand properties and the effect on the

transition metal complexes synthesized therefrom.

Aryloxide Tungsten Complexes: The Ligand Electronic Effect

As described in Chapter 2, several "new classical catalytic systems" have been

devised and synthesized through the replacement of a halide atom with a phenoxide (or in

a more general way, an aryloxide) ligand, particularly from WCl6 and WOCl4. As

mentioned before, this replacement brings about a series of advantages which include an

increased compatibility with organic substrates and solvents, higher solubility than the

parent halides, and an increased chemical stability mainly arising from the bulkiness of

the ligand.

Without a doubt, the most significant effects are those observed in the electronics

of the complex. When comparing the two ligands as anions (e.g., Cl vs. a generic ArO ),

the first evident difference is their basicity and nucleophilicity. Not only are the halide

anions much less basic than the phenoxides but also less nucleophilic, two clear features

derived from the negative charge stabilization by the two species.

These characteristics are also expressed in their reactivity in organometallic

chemistry: halides are precursors to alkoxide and aryloxide complexes since they are

easily substituted not only by the anions as nucleophiles, but also by the parent alcohol or

phenol. Once substituted and when alkoxides and halides are compared, it is easy to

observe that the electron withdrawing ability of these oxygenated ligands is often lower

than the parent halide. In some cases the participation of the oxygen p-electrons has

significant effects on bonding and reactivity. Because of their decreased electron

withdrawing abilities, the metal becomes less electrophilic when the substitution takes

place, this -arguably- being the most important electronic effect.

Ligand Substituent Effects and Catalytic Activity

The development of aryloxide tungsten complexes as precatalysts for olefin (and

alkyne) metathesis occurred parallel to understanding the effect that ligand substitution

has on the complex properties. The pioneering work of Quignard and co-workers set the

guidelines for further exploration of such effect (Quignard 1986). In their study, a series

of bis(aryloxide) derivatives of WCl6 synthesized as shown in Figure 3-1 (Quignard




39 Y = Me
40 Y = Ph
41 Y=F
42 Y = CI
43 Y = Br

+ WC16

o Y
CIW" O ",
Cl Y

44 Y = Me
45 Y= Ph
46 Y = F
47 Y = CI
48 Y= Br

Figure 3-1. Synthesis of a series of bis(aryloxide) tungsten complexes (44-48) from WCl6
and the parent phenols (39-43).

+ 2 HCI

The catalytic properties of complexes 44-48 in the metathesis of linear, acyclic

olefins were studied as a function of the substituent identity, the activator used, and the

activation time. These experiments revealed interesting details of aryloxy tungsten-based

catalysts and provided further mechanistic evidence in support of the existing proposal.

One of the reactions used by Quignard and coworkers is the equilibration

metathesis of cis-2-pentene in chlorobenzene, as shown in Figure 3-2.

WCl4(OAr)2 / MR4
85 C, PhCI

Figure 3-2. Metathesis of cis-2-pentene as studied by Quignard (1986). Bu4Pb, Bu4Sn and
Me4Sn constitute the group of activators studied.

Several features of the catalytic systems based on complexes 44-48 should be

highlighted. First, for any given precatalyst (44-48) the activity of the system based on the

activator employed decreased in the order Bu4Pb > Bu4Sn > Me4Sn, and this trend can be

correlated with an increase in the M-R bond dissociation energy.

Second, there is a rough correlation between the pKa of the parent phenol and the

catalytic activity of the corresponding catalysts, which increases parallel to the acidity of

the phenol. This observation can be explained using the electron withdrawing effect that

the substituent Y exerts on the complex, and opens a number of possibilities for ligand

modification. Table 3-A summarizes some of Quignard's experimental data observed for

the reaction shown in Figure 3-2.

Table 3-A. Activities of the two component systems 44-48 / MR4 in the metathesis of
cis-2-pentene expressed as equilibration times under standardized conditions

Me4Sn Bu4Sn Bu4Pb

44 >48h >48h 5h

45 > 48 h 5 h 3 h

46 >48 h 5 h 75 min

47 24 h 3 h 45 min

48 2 h 15 min 15 min
a) Reactions carried out in chlorobenzene at 85 C.
Substrate/precatalysts/activator = 50:1:2.

Similar conclusions were arrived at by Bell (1992) while studying the bulk

polymerization of dicyclopentadiene using tungsten oxychloride derivatives activated by

Bu3SnH, synthesized as shown in Figure 3-3.

+ WOCl4 2-

+ 2HC

Z,Y = Me
Z,Y = Ph
Z,Y = CI
Z,Y = Br
Z,Y = 'Pr
Z,Y = OMe
X,Y = CI, Z=Me

Z,Y = Me %:
Z,Y = Ph
Z,Y = CI
Z,Y = Br
Z,Y = 'Pr
Z,Y = OMe
X,Y = CI, Z=Me

Figure 3-3. Synthesis of a series of bis(aryloxide) tungsten complexes (32,52-57) from
WOCl4 and the parent phenols.

Although the mechanism by which tungsten aryloxides are activated by tin

hydrides is still unclear (for a recent discussion, see Kelsey 1997), Bell's contribution

demonstrated that the mechanistic implications of the ligand substitution discussed above

also apply to this type of catalytic systems. Furthermore, Bell demonstrated that two

parameters, namely the AMI-calculated charge density on the phenoxide oxygen and the

experimental reduction potential of the complex correlated effectively and could be used

as quantitative indicators of the catalytic activity of the catalysts synthesized therefrom.

Two significant advantages were introduced in this study. First, instead of using a

metathesis transformation as an amplification of the catalytic system properties, the

experimental measurement of the reduction potential was chosen as an indication of the

electropositive character of the metal, and correlated with the expected trend. Second, the

semiempirical calculation of the charge on the oxygen atom of the corresponding

phenoxide anion correlated well for the series, making of this property an attractive guide

for further development of olefin metathesis catalysts. Among the explored set and in

agreement with Quignard's observations, derivatives synthesized from 2,6-dibromophenol

(the most acidic phenol in the series) exhibited the highest activities.

Novel Aryloxy-Tungsten Structures

During our initial experiments with classical systems as catalysts for ADMET

polymerization, we explored aryloxide derivatives of tungsten (VI) oxychloride and

tungsten (VI) chloride 32-34. Based on our early observations we started a study on the

former group, following the pioneering approaches of Basset and Bell, through the

modeling of a complementary series of phenoxides which can be used in the synthesis of

systems of the type WOCl2(OAr)2. The appended results are shown in Figure 3-4.




Expected Increase in -0.4365
- Catalytic Activity -0.4364
-0.4101 D-



E Bell, 1992
SOur work

-0.40 -0.45 -0.50 -C
Calculated Oxygen Charge (e-)

Figure 3-4. AMI-calculated charge on the oxygen atom for a series of phenoxides.

Within the modeled set the patterned columns represent the derivatives of

2,6-dibromophenoxide bearing a substituent on the para position. These potential ligands

were chosen in light of the possibility of further correlation of the complexes properties

with the substituent Hammett constant, cy. Thus it is reasonable to state that complexes

derived from more electron withdrawing phenoxides should exhibit high activity as

metathesis catalysts as well. This observation encouraged us to synthesize complexes

63-67, and study their behavior as precatalysts in olefin metathesis classical systems.

Br-_ Br B r
2 + WOC14 II + 2 HCI
,1 ^Tol, 110 C Br W"II
Cl 0
X Br
Br /

58 X = F 63 X = F
59 X = Br 64 X = Br X
60 X = CF3 65 X = CF3
61 X=CN 66 X=CN
62 X = NO2 67 X = N02

Figure 3-5. Synthesis of a new series of bis(aryloxide) tungsten complexes (63-67) from
WOC14 and the parent 2,6-dibromophenol derivatives.

Complexes 63-67 were synthesized in a straightforward manner following the

general procedure reported by Nugent (1995) which consists of refluxing WOC14 with the

corresponding phenol in toluene (see Chapter 6). Essentially quantitative yields of crude

material can be isolated following the removal of the solvent in vacuo, and this crude

complex is usually of adequate purity to be used in metathesis chemistry. Complexes

63-66 can be purified through recrystallization, whereas complex 67 cannot. It is a black,

insoluble powder, which has not been included in our comparative experiments.

Structural Characterization

A detailed structural analysis of the parent 2,6-dibromophenoxide complex by

Nugent, Feldman and Calabrese revealed the asymmetry of complex (32). A non-bonding

interaction between one of the bromine atoms and the metal renders the two aryloxide

ligands inequivalent (Nugent 1995).

Single crystal X-ray analysis shows similar features among complexes 63, 64 and

32. Inequivalent phenoxide ligands are present due to a presumably weak interaction

between Bri and the tungsten atom, and such geometry distortion is (in both cases) also

reflected in the inequivalent W-O-C bond angles. (Figure 3-6, Figure 3-7). The possibility

of exploring its extent (through comparison of interatomic lengths and angles) as a

quantitative parameter for predicting activity in the series was envisioned.

A detailed observation of structure 63 (see Table 3-C) reveals that the

incorporation of the fluorine substituent in the para position causes an increase of the

W-Brl distance (3.205 A in 63 vs. 3.120 A in 32). The same decreased interaction is also

evidenced by the wider bond angle on 02 (139.28 in 32 to 142.18 in 63). These

observations can be justified by appealing to a larger extent of p donation from the

oxygen orbitals of the aryloxy ligand to the empty metal orbitals, a phenomenon well

documented elsewhere (Kershner 1989a,b)

At the same time, a comparison of complexes 64 and 63 suggests a tighter

non-bonding W-Br interaction revealed in a shorter interatomic distance (3.149 in 64 vs.

3.205 A in 63). This feature as well as the angle about 02, (140.5 vs. 142.1 in 63) again

implied a correlation between the electronics on the ligand and the solid state structure.

No significant differences were found when comparing the phenoxide W-0 bond

distances. A summary of structural data for complex 64 is shown in Table 3-E.

Table 3-B. Crystal data and structure refinement for complex 63.

Empirical formula
Formula weight
Crystal system
Space group

Unit cell dimensions

Density (calculated)
Absorption coefficient
Crystal size
Theta range for data collection
Index ranges
Reflections collected
Independent reflections
Absorption correction
Refinement method
Data / restraints / parameters
Goodness-of-fit on F2
Final R indices [I>2((I)]
R indices (all data)
Extinction coefficient
Largest diff. peak and hole

173(2) K
0.71073 A
a= 15.1172(2) A c= 900
b= 7.9837(1)A 3 =112.7430(10)0
c= 16.6186(1) A y= 900
1849.77(4) A3
2.903 Mg/m3
15.201 mm-1
0.26 x 0.22 x 0.17 mm3
2.49 to 27.490
-18 11725
4230 [R1nt = 0.0280]
Full-matrix least-squares on F2
R1 = 0.0240, wR2 = 0.0598 [3878]
RI = 0.0283, wR2 = 0.0625
1.343 and -1.452 e. A-3

F2 C-224 02

C25 C28 C11 12


Figure 3-6. Molecular structure of 63, with 50% probability ellipsoids, showing the atom
numbering scheme.

Table 3-C. Selected bond lengths (A) and angles (0) for complex 63.

Bond lengths (A)

W-01 1.689(3)

W-02 1.879(3)

W-03 1.856(3)

W-Cll 2.3370(10)

W-C12 2.3395(10)

W-Brl 3.205(7)

02-C11 1.337(5)

03-C21 1.352(5)

Bond angles (0)

C11-02-W 142.1(3)

C21-03-W 162.5(3)

01-W-03 102.4(2)

01-W-02 99.6(2)

02-W-03 157.64(13)

01-W-Cll1 101.99(11)

Cll-W-C12 155.9(4)

Table 3-D. Crystal data and structure refinement for complex 64.

Empirical formula
Formula weight
Crystal system
Space group

Unit cell dimensions


Density (calculated)
Absorption coefficient
Crystal size
Theta range for data collection
Index ranges
Reflections collected
Independent reflections
Absorption correction
Refinement method
Data / restraints / parameters
Goodness-of-fit on F2
Final R indices [I>2((I)]
R indices (all data)
Extinction coefficient
Largest diff. peak and hole

173(2) K
0.71073 A
a = 7.8459(5) A ca
b = 8.8504(5) A P
c = 14.2994(9) A y
984.64(10) A3


3.138 Mg/m3
18.320 mm-1
0.16 x 0.19 x 0.23 mm3
1.43 to 27.50.
-10 < < <9, -10 < k< 11, -17 < < 18
4417 [Rn, = 0.0181]
Full-matrix least-squares on F2
4417 / 0/218
RI = 0.0258, wR2 = 0.0628 [3826]
RI = 0.0323, wR2 = 0.0647
1.537 and -1.000 e. A-3

Figure 3-7. Molecular structure of 64, with 50% probability ellipsoids, showing partial
atom numbering scheme.

Table 3-E. Selected bond lengths (A) and angles (0) for complex 64.

Bond lengths (A)

W-01 1.684(4)

W-02 1.888(3)

W-03 1.859(3)

W-Cll 2.3104(13)

W-C12 2.3302(13)

W-Brl 3.149(3)

02-C1 1.349(6)

03-C7 1.345(6)

Bond angles (0)

C1-02-W 140.5(3)

C7-03-W 159.2(3)

01-W-03 102.9(3)

01-W-02 99.4(8)

02-W-03 157.3(6)

01-W-Cll1 101.5(0)

Cll-W-C12 157.2(6)

02-W-Brl 75.3(3)

In contrast, crystallization of complex 65 turned out very difficult due to the high

solubility of the complex in non-protic, non-coordinating solvents. A single crystal of low

quality was grown from a mixture of CH2C12/toluene at -30 C, and although the amount

of disorder found in the structure refinement was very high, X-ray diffraction data reveals

interesting structural information. In contrast with complexes 32, 63 and 64, the ligand

arrangement appears to be symmetrical in complex 65, and no W-Br interaction can be

observed. A view along a Cl-W bond shows the symmetrical arrangement of the ligand

set, as well as the square-based pyramidal geometry. Apparently, the effect of the

trifluoromethyl group on the complex electronics is not the only significant variant

influencing its conformation in the crystal.

1C1 C12

Figure 3-8. Platon view of complex 65.

These collected observations suggest that the properties of the complexes in the

solid state (e.g., bond angles, lengths) cannot be used as an inference parameter for the

catalytic activity of the alkylidenes generated therefrom. No trend could be observed

within the structural data.

ADMET Chemistry

The bulk polymerization of 1,9-decadiene was studied in order to explore the

polymerizing ability of complexes 63-67 and in search of a useful probe for the

comparison of their catalytic properties. It was found that complexes 63-66 efficiently

catalyze the condensation polymerization of this and other hydrocarbon dienes such as

1,8-nonadiene, complementing the range of aryloxide tungsten-based classical catalysts

for ADMET polymerization. Linear metathesis polymer can be obtained following the

general methodology described in Chapter 2, and the combined results are shown in Table


Table 3-F. Acyclic diene metathesis polymerization of 1,9-decadiene

Complex [W] mol %' Mn (NMR)b Mn (GPC)c PDIc

63 0.2 1300-1800 3000 1.94

64 0.2 2300 6400 2.71

65 0.2 neg 11,300 2.11

66 0.4 1500

a) Concentration with respect to monomer. Bulk polymerization at 85 C,
b) Calculated via end-group analysis from IH-NMR data, by integrating the
signals corresponding to the terminal vinyl groups vs. internal olefin.
neg = no end groups were observable.
c) Determined by gel permeation chromatography in chloroform (see
experimental) using polystyrene standards.

In concert with the previously described experiments (Chapter 2), Bu4Sn proved

again to be a better activator when compared to Bu3SnH. In fact, the use of Bu3SnH did

not lead to an active catalyst when used with complexes 65 and 66, which consistently

decomposed upon exposure to the standard polymerization conditions. Different results

were found among the studied complexes, and lower molecular weights were found in the

reactions catalyzed by complexes 63 and 66, both by NMR end-group analysis and by gel

permeation chromatography (Table 3-F). The IH-NMR spectrum of a sample of

polyoctenylene synthesized using complex 63 and Bu4Sn is shown in Figure 3-9, and

shows vinyl resonances attributed to terminal groups (4.9-5.0 and 5.8 ppm.)

a c

b d n


6.0 5.5 5.0 4'.5 4.0 .5 3.0 2'.5 2.0 1.5 ppm

Figure 3-9. H-NMR of a sample of polyoctenylene obtained using complex 63 and

In contrast to the behavior shown by complexes 63 and 66, complex 65 appears to

yield the highest molecular yield polymer in good yields, probably as a consequence of

the influence of its particular substituent, the trifluoromethyl group. The 'H and 13C-NMR

spectra of a sample of polyoctenylene synthesized using complex 65 and Bu4Sn are

shown in Figure 3-10 and Figure 3-11 respectively. As demonstrated before for

complexes 32-34, high molecular weight polymer is obtained from this polycondensation

reaction, and this fact is evidenced in the absence of resonances attributed to vinyl end

groups. Furthermore, analysis of the NMR data reveals that no isomerization of the olefin

groups has taken place, a competing reaction observed in other metathesis applications of

analogous complexes (e.g., 5) (Vosloo 1997, van Schalkwyk 1998).

a c

b d

7.0 6.0 5.0 4.0 3.0 2.0 1. .0 ppm

Figure 3-10. H-NMR of a sample of polyoctenylene obtained using complex 65 and

Although no quantitative kinetic measurements were made on the metathesis

reactions catalyzed by systems based on complexes 63-66, qualitative observations reveal

two interesting features of the studied systems. The first characteristic is the apparent

activity increase in agreement with the expected behavior for complexes (32-)64-65 as

ADIVIMET catalysts, which seems to be especially evident for complex 65. This is observed

in a shorter induction period prior to ethylene release from the reaction mixture which can

be attributed to the activation step as proposed by Quignard and co-workers (1986). In

addition, the identity of the substituent X in the series affects not only the activity, but

also seems to have a dramatic effect on the solubility properties of these complexes.

While complex 65 is very soluble in conventional solvents, complex 67 is virtually

insoluble, limiting its usability not only in ADIIMET but also in other metathesis

applications. Although limited solubility has also been observed for complex 66 at room

temperature, the complex is useful at the experimental polymerization conditions. The

solubility behavior of the complexes described may also account for a low % conversion

in the case of complexes 64 and 66, for which faster decomposition kinetics with respect

to complex 65 prevents the formation of high molecular weight polymer.

Also noteworthy, especially while considering the application of this methodology

to other metathesis processes, is the fact that along with an increased activity in olefin

metathesis, a parallel effect on the reactivity towards polar functionalities is also

apparent. This is evidenced in a poor air and moisture-resistance shown in all the

complexes studied, a phenomenon that is remarkably evident in complexes 66 and 67.

a c
b(trans) c,d
b d



140 130 120 110 100 90 80 70 60 50 40 30

Figure 3-11. 13C-NMR of a sample of polyoctenylene obtained using complex 65 and

ROMP Chemistry

Complementary to the study of complexes 63-66 in ADMET polymerization, the

possibility of using complexes 63-66 as precatalysts for ring opening metathesis

polymerization (ROMP) was also explored. Processes as important as the industrial

production of polydicyclopentadiene thermosets are grounded on the use of classical

systems (Kelsey 1997). Heating bulk dicyclopentadiene (DCPD) in the presence of

complexes 63-66 and Bu4Sn yields crosslinked PDCPD after a short induction time.

Although no efforts were made to determine the amount of residual monomer or the

magnitude of the polymerization exotherm, the exploration of the effect of monomer to

tungsten complex (M/W) molar ratio reveals that fast polymerization occurs at ratios as

high as 5000:1 M/W both under solvent and bulk conditions.

At the same time, linear polynorbornene is obtained upon exposure of a 1 M

toluene solution of freshly sublimed norbornene to a mixture of complexes 63-66 and

Bu4Sn in toluene at 85 C. The features of other aryloxide systems in the polymerization

of norbornene have been previously reported (Barnes 1994, Diets 1993).


6.5 -

6.0 -

-5.5 /


< 3.5 -


2.5 -

3.5 4.0 4.5 5.0 5.5 6.0
Log (MW)

6.5 7.0 7.5 8.0

Figure 3-12. GPC traces of four polynorbornene samples synthesized using complexes
63-66. The traces are named based on the ligand substituent

Quantitative conversion of norbornene is required in order to use its ring opening

metathesis polymerization as a probe for comparative catalytic activity among the

systems studied. In the case of solution polymerizations run for extended times (at W/M

ratios of 100:1), catalyst decomposition occurs before the monomer is totally consumed,

and lower yields of polymer can be obtained (71-84% purified polymer). Although it is

safe to overlook this fact while studying ROMP chemistry, the difference in molecular

weights (Mp) can be attributed to differences in either the propagation rates or the extent

of activation, and the designed experiments do not allow the determination of which

factor is more significant for the studied series. While stronger alkylating agents such as

aluminum or lead alkyls are needed for the conversion of analogous systems bearing less

electron withdrawing phenoxide ligands, milder cocatalysts (e.g., Bu4Sn) can be used to

activate systems in which the metal is made more electropositive (Nugent 1995, Quignard

1986, Kelsey 1997). The same was anticipated videe infra) for systems 63-66 which,

having a similar affinity for norbornene differ in the amount of actual catalyst formed, not

in their propagation kinetics. This would determine the number of growing chains and

have a clear effect on the molecular weights observed among the series. In order to test

this assumption, the formation of polynorbornene at short polymerization times should

reveal features of the initiation (or activation in our case) and propagation.

The GPC traces shown in Figure 3-12 belong to the solution polymerization of

norbornene stopped after 15 minutes, i.e. before catalyst decomposition becomes apparent

and before any secondary processes take place. Monomodal molecular weight

distributions were found in all cases, and the polydispersities found ranged from 1.8 to

3.3. An increasing PDI value in the series 65-64-63 suggests that the availability

(concentration and rate of formation) of initiating species varies based on the ligand

substituent, affecting the properties of the polymers obtained. The largest PDI value was

found in the case of complex 66; this observation can be attributed to an inefficient

initiation step, an observation in agreement with other metathesis processes. It is

noteworthy that if a constant extent of activation and similar initiation kinetics were to be

assumed for all four complexes, then the overall propagation constants extrapolated from

differences in the molecular weights would still fall within the same order of magnitude.

The effect of substitution on the conversion kinetics has been subject to further

exploration using a more sensitive probe based on ring closing metathesis chemistry.

Ring Closing Metathesis

The behavior and applicability of the complexes reported herein as precatalysts

for ring closing metathesis (RCM) was also explored.

EtO2C CO2Et 63-69 / Bu4Sn EtO2C CO2Et
+ C2H4
Tol, 90 C

Figure 3-13. Ring closing metathesis of diethyl diallylmalonate.

The results obtained indicate that while complexes 63-65 effectively catalyze the

RCM of diethyl diallylmalonate, complex 66 decomposes before any significant

conversion of the starting material occurs (Table 3-G). The applied experimental

methodology consists in the introduction of the olefin after 5 minutes of interaction of the

W complex with Bu4Sn at the activation temperature (90 oC), and monitoring the reaction

by GC and NMR spectroscopy.

Table 3-G. Ring closing metathesis of diethyl diallylmalonate.

Complex Conditions' %C 90 minb %C Finalb

63 [S]=l M, [S]/[W]=25 51(46) 86 (88)

64 [S]=l M, [S]/[W]=25 37 (-) 54 (49)

65 [S]=l M, [S]/[W]=25 57 (50) 89 (90)

65 [S]=0.25 M, [S]/[W]=12.5 74 (71) 94 (92)

66 [S]=1M, S/W=25 <10 (-) 29 (19)

a) Toluene, 85 C, [Bu4Sn]=2.5x[W], 5 min activation time.
b) Determined by gas chromatography (see experimental) using ca-ionone as
internal standard. Values in parenthesis were extracted from 1H-NMR data by
integrating the signals corresponding to the allylic methylene group of product
and diene.

The results shown in Table 3-G are in concert with the observed behavior of

complexes 63-66 in both ADMET and ROMP chemistry, for they suggest that the identity

of substituent X seems to affect not only the electronics of the complex (both in activity

and chemical stability) but also its solubility. At the same time, the increased chemical

sensitivity of complexes 63-66 with respect to their parent, non substituted complex 32,

can be observed in the substrate to precatalyst ratio (S/W) required to drive the reaction to

excellent yields. While complex 32 has been reported to yield the cyclic diester in 86 %

after 1 h at 90 C (Nugent 1995), the use of complex 65 only afforded the product in

similar yields (89 %) when reacted for longer periods of time (ca. 4 h). Carrying out the

reaction at lower S/W ratio and lower concentration (0.25 M) resulted in very high

percent conversion (Table 3-G, entry 4). Attempts to carry out the reaction using the same

conditions reported for 32 and Et4Pb (Nugent 1995) derived in a 47 % conversion (NMR)

after 60 min, and 50 % (NMR) after 90 min. Although this decreased conversion could be

attributable to a faster decomposition of the alkylidene formed upon activation, it is more

likely to reflect the lower extent of activation achieved with Bu4Sn when compared to



A series of tungsten aryloxide complexes were designed, synthesized, and their

properties investigated in the catalysis of various olefin metathesis processes. Complexes

63-66 catalyze the ADMET polymerization of hydrocarbon dienes, the ring opening

polymerization of norbornene and dicyclopentadiene, and -to some extent- the ring

closing metathesis of diethyl diallylmalonate. An exploration of catalysis conditions in

the above mentioned processes reveals that substitution of the aryloxide ligand causes a

modification of the electronics of the complexes, evidenced not only in qualitative

differences in catalytic activity, but also in the chemical stability and the solubility

behavior exhibited by the studied complexes. Although initially considered as a possible

tool in the prediction of the physicochemical behavior of complexes 32,63-67, structural

analysis show that the observation of properties in the solid state cannot be associated

with features such as catalytic activity.


As described in Chapter 1, the rules governing the relationship among monomer

structure, reactivity and catalytic systems in ADMET polymerization have been the focus

of active research efforts. Complementing the research described in the previous chapters

which was mainly concentrated on novel synthetic methodology and catalytic systems,

investigations on new routes towards tin containing polymers using ADMET chemistry

are herein described.

Metal-Containing Polymers

Although relatively new, the synthesis of organometallic polymers has become an

extensively investigated and fruitful new discipline and has emerged as a possible route

towards materials exhibiting unique properties arising from both the metallic and the

organic components (Manners 1996, Rehahn 1998). Interest in metallopolymers has

grown constantly not only because of the possible applications predicted for some of

these systems, but also because of the variety of synthetic tools and obtainable structures

that a number of organometallic reactions and substrates offer to the polymer chemist.

The four general groups in which organometallic polymers have been classified are

shown in Figure 4-1.

b)c) d)

Figure 4-1. General representation of a) ionic, b) 7t-complex, c) coordination and d)
covalent type metal containing polymers.

Among the numerous combinations of structures and properties available today,

metallopolymers obtained through the polymerization of monomers containing a covalent

metal-carbon bond still represent a relatively small portion (Pomogailo 1994). This is a

clear representation of the marked stability of chelate and -in general- coordination

complexes with respect to covalent structures, as well as the many possibilities offered by

coordination chemistry in terms of oxidation states, ligand combinations, and metal

incorporation. Such advantages have been creatively explored in strategies where the

propagation reaction involves coordination phenomena (Pomogailo 1994).

Tin-Containing Polymers

The polymerization of monomers containing tin as a substituent yields in most

cases polymers with tin-containing side chains, and several examples of organometallic

polymers of tin belong to this category. The polymerization of olefins such as tin

acrylates is the most common route to tin-containing polymers, and this specific example

has become a widely use approach towards polymeric biocides.

As demonstrated by Cummins (1991), ring opening metathesis polymerization can

also be a route towards organometallic polymers. Microphase-separated block

copolymers possessing the structure shown in Figure 4-2 could be synthesized from

norbornene and norbornene-type metal chelates.

x v

R Sn R

Figure 4-2. Microphase-separated block copolymers containing Sn(VI) moieties
synthesized via ROMP.


Polymetallanes is the name given to the family of organometallic polymers

containing backbones exclusively made of metals connected by CY bonds. The vast

majority of the research on polymetallane structures has been concentrated on the

development of polysilanes, due to their optical and electrical properties (Miller 1989,

West 1986, Reichl 1996). Along with their silicon and germanium analogues,

polystannanes have attracted a great deal of attention in recent times, and synthetic routes

towards this type of materials are in constant development. One of the earliest approaches

to polystannanes consists in the coupling of dialkyltin dihalides, and a variety of

experimental procedures have been reported since as early as 1964. Representative

examples of reaction conditions are shown in Table 4-A.

Table 4-A. Experimental conditions used in the synthesis of polystannanes from
dialkyltin dichlorides.

Cl-Sn-CI ClntCi
68 69
Conditions Product Reference

Na, Toluene reflux R = Bu, Mn = 2100 Neumann 1964a

Na, 15-Crown-5, Toluene reflux R = Bu, Mn = 13,800 Zou 1992

Na, 15-Crown-5, Toluene reflux, 6 Devyder 1996
R = Bu, Mn = 1 x 106 Devylder 1996
light exclusion / sonication, 1-5 h

R = Et, 74 % M = 3980
SmI2, THF/HMPA, 23 C, 24 h R = E 4 M 3980
PDI = 1.21 Yokoyama 1995,

R = Et, 76 % Mn = 3570 Mochida 1998a
Sml2, THF, 65 5 h
m PDI= 1.15

R = Bu, Oct
Electrochemical, R flu, Oct
lectrochemica Mn = 6000 11,000 Okano 1998
Cu electrode, DMEpI1.3.
PDI = 1.3 2.6

A serious disadvantage found in the procedures listed above (except for the

electrochemical method developed by Okano) is the use of large amounts of alkali metals

or SmI2, and alternative methodologies are desired in view of the predicted applications

of these materials.

Another elegant approach towards polystannanes consists in the transition

metal-catalyzed dehydropolymerization of dialkyl (or diaryl) tin dihydrides, an extension

of a synthetic methodology employed in the synthesis of polysilanes from silicon

dihydrides (Tilley 1993, Imori 1994). This reaction has been shown to proceed in the

presence of organometallic complexes of a variety of transition metals, summarized in

Table 4-B. Among the different experimental conditions reported, the activity, product

yields and high molecular weights found by Imori (1995) are remarkable.

Table 4-B. Experimental conditions used in the synthesis of polystannanes from
dialkyltin dihydrides.

H-Sn-H H n+H
70 71
Conditions' Reference

A variety of Zr and Hf complexes, Imori 1995,
mainly metallocenes. Lu 1996
Zr and Hf metallocene dichlorides
or W, Mo, Cr hexacarbonyls in Woo 1997
the presence of Al hydrides

Pd/Fe dinuclear complexes Braunstein 1998

During the last decade, the interest on the properties of macrostructures built with

group 14 metals led Sita and coworkers (1994) to the exploration of a controlled route to

linear polystannanes (Sita 1992, Shibata 1998). In this approach, the oligostannane chains

are built one atom at a time, which although considered a limitation for the extension of

this methodology to higher macromolecules, has given valuable insight into the

fundamentals of polymetallane structures (Figure 4-3).

1/2 R2SnH2
1/2 R2SnH2 HC OEt R2OEt LiNMe2 R2Sn OEt
+ ^ R2SnHCI R2Sn R2Sn
1/2 R2SnCI2 AIBN CI NMe2 72

HNMe2 SnR'3

R'3Sn--Sn--Sn-H < R2Sn" -
LR n R Several SnR'3 nR3

Figure 4-3. Sita's strategy towards pure, linear polystannane oligomers.

The extensive research on linear polystannanes can be easily justified when their

physical properties are considered. Polydialkylstannanes display a CY-CY* transition at

390-400 nm which represents a red shift of ca. 70 nm with respect to their silicon

analogues, while polydiarystannanes seem to exhibit cy,7i-conjugation and even lower

bandgaps (430-440 nm) have been observed (Lu 1996). In addition, Imori and coworkers

(1995) observed thermochromic behavior in some polydialkylstannanes as well as fast,

irreversible photobleaching initiated by UV or visible light. These properties are very

promising in terms of possible applications in sensor devices and in semiconductor

technology. In fact, although only at a very preliminary stage, conductivities in the order

of 0.3 S cm-1 were found in polydialkylstannanes doped with SbF5 (Imori 1995).


Polymers that contain covalently bound carbon and tin atoms along the backbone

are known as polycarbostannanes. The first examples of polycarbostannane polymers

were synthesized via step polymerization of diolefins (or diacetylenes) and tin dihydrides,

as shown in Figure 4-4 (Leusink 1964, Noltes 1961, 1962).

R" R"
R2SnH2 +

R R" R"
LR \n

R2SnH2 + R' ---,- H-Sn S


R2SnH2 + R' --Sn
+R n

Figure 4-4. General scheme for the synthesis of step addition polymers between dialkyltin
dihydrides and diolefins or acetylenic compounds.

Another interesting example of polymers containing tin atoms incorporated along

the backbone is the family of structures shown in Figure 4-5. In this case, the polymers

are synthesized from tin-containing oL,co-diynes and metal or aryl dihalides leading to

materials in which the effect of O,7t-conjugation on the optical properties of the polymers

was investigated (Hagihara 1994).

- CH3 1
I C 3Sn-Art
CH3 n

CH3 PR3 -

L;H3 PR3 31

Figure 4-5. Structures of tin-containing G,7t-conjugated structures (Hagihara 1994). M
Pt, Pd.

Polycarbostannanes via ADMET Polymerization

As presented in Chapter 1, ADMET polymerization is a valuable tool for the

synthesis of unsaturated polymers containing a variety of functional groups (Figure 1-7).

In 1997, Wolfe demonstrated that Schrock's molybdenum catalyst 26 could be used in the

polymerization of the stannadienes 73 and 74 to yield high molecular weight poly[bis(4-

pentenyl)di-n-butylstannane] (75) as well as linear (76) and cyclic (77) oligomers of

poly[bis(3-butenyl)di-n-butylstannane]. The studies performed by Wolfe demonstrated

the feasibility of synthesizing tin-containing polymers via ADMET polymerization, as

well as the compatibility between the metathesis catalyst 26 and the alkyltin moiety.

Bu, ,Bu Bu., Bu
S n, S

73 75

Bu, Bu 26 Bu4 Bu Bu 4.Bu
S n l n S

74 76 77

Figure 4-6. ADMET polymerization of bis(alkenyl)di-n-butylstannanes by Wolfe (1997).

Monomer-Activated Metathesis Polymerization

By 1997 ADMET polymerization had become a new route towards tin-containing

polymers, as demonstrated by Wolfe (1997). At the same time, the availability of a new

methodology for ADMET polymerization using soluble aryloxide tungsten-based

catalysts (Chapter 2) and the fact that tetraalkyltin complexes were used as activators of

these and other classical catalytic systems (Nubel 1994, Ivin 1997) allowed the

consideration of bis(alkenyl)dialkyltin compounds as both monomers and activators for

ADMET polymerization. The same stannadienes studied by Wolfe (1997) were

considered as possible monomers, namely bis(3-butenyl)di-n-butylstannane (74) and

bis(4-pentenyl)di-n-butylstannane (73).


Similar to the behavior observed in the bulk polymerization of hydrocarbon

monomers (Chapter 2) by systems composed of a tungsten aryloxide complex and a

tetraalkyltin activator, the polymerization of diene 73 proceeds with ease. When a

mixture of 100 equivalents of diene 73 and 1 equivalent of complex 32 was heated to the

activation temperature (85-90 C) under an argon atmosphere, a slight color change was

accompanied by gas evolution. During the first three hours of reaction the viscosity

increased significantly and ethylene evolution was observed. Exposure of this mixture to

vacuum accelerates the bubbling, and the reaction is completed after ca. 7 h under

reduced pressure. The observed polymerization carried out without the use of an external

activator suggests that the activation takes place through the mechanism discussed in

Chapter 2 and represents the first case in which a reagent serves as both the monomer

and the activator in a wtniciltn i polymerization (see below).

Contrary to the observed results in the case of hydrocarbon monomers,

purification of the title polymer under normal conditions did not proceed as expected.

After the first precipitation of the polymer from warm chloroform into methanol an

insoluble gel was obtained suggesting partial crosslinking of the product. In order to

present an explanation to this phenomenon, the mechanism of catalyst activation for this

particular polymerization needs to be analyzed in detail.

Proposed activation mechanism
In agreement with the mechanism detailed in Chapter 2, the active catalyst is

formed in a two-step sequence. The first step (shown in Figure 4-7) consists of the

alkylation of 1 equivalent of complex 32 by 2 equivalents of diene 73 yielding a dialkyl

complex (78) and the tin halide 79.

0 Br




2 /

Figure 4-7. First step in the proposed mechanism for the formation of an active catalyst
from complex 32 and the stannadiene 73: alkylation.

The second step is the thermally induced ca-abstraction reaction shown in Figure

4-8, which leads to the formation of the oxocarbene 80 and a molecule of n-alkane

(butane in the example).

0 Br + C4H10
B Br r Br
B\ B Br Br

78 80

Figure 4-8. Second step in the proposed mechanism for the formation of an active catalyst
from complex 32 and the stannadiene 73: ca-abstraction.

The properties of the product obtained can be explained by the incorporation of tin

chloride moieties in the polymer. As depicted in Figure 4-7, two equivalents of the

stannane chloride 79 are formed in the transmetallation reaction and this is a difunctional

monomer that can be incorporated in the polymer chain. Isolation and purification of the

polymer using wet solvents and exposure of the polymer sample to the atmosphere may

cause partial hydrolysis of the Sn-Cl bonds and lead to the formation of tin hydroxides,

which can in turn crosslink the polymer. A general scheme for this phenomenon is shown

in Figure 4-9. Although no direct evidence exists for this reaction, modification of the

purification conditions to eliminate the moisture sources affords linear, soluble polymer.

In turn, when the polymerization is performed at higher monomer to precatalyst ratio

(250:1) no gelation is observed upon precipitation of the polymer under normal

conditions, which can be attributed to a lower concentration of Sn-Cl bonds in the

polymer chains. Furthermore, samples of the same polymer synthesized using the

molybdenum complex 26 by Wolfe (1997) do not crosslink upon exposure to moisture,

an observation that supports the proposed crosslinking route.

R H20 I
-,,Sn,- -,, I -,,,,Sn-,,





Figure 4-9. Proposed mechanism for chemical crosslinking of the polymer obtained from
the ADMET polymerization of diene 73 using complex 32 (100:1).

Optimization of the experimental conditions led to the isolation of polymer 75

through the clean polymerization of diene 73 using the tungsten complexes 32 or 33.

Bu4 NBu


[ Bu4 Bu
32 or 33 Sn

Figure 4-10. ADMET polymerization of bis(4-pentenyl)di-n-butylstannane using
complexes 32 or 33 and no external activator (250:1).

Polymer 75 has the same physical properties described by Wolfe (1997). It is a

viscous liquid, soluble in halogenated and aromatic solvents, but insoluble in polar

solvents such as methanol from which it can be precipitated. High molecular weight was

observed in all cases (GPC and 13C-NMR end-group analysis). The summarized

characterization data is presented in Table 4-C.

g h

+Ab d' n

a (c)





140 130 120 110 100

90 80 70 60 50 40 30 20 10 ppm

Figure 4-11. 13C-NMR of polymer 75 synthesized using complex 33 and no external
activator (250:1).

Several interesting features can be observed in the NMR spectral characteristics of

polymer 75. First, the polymer contains mainly trans linkages (80% from quantitative

13C-NMR data), a figure in agreement with other ADMET polymers. Second, three

distinct signals can be observed in the 119Sn-NMR of the polymer samples analyzed

(Figure 4-12) and their relative intensities correspond to the calculated statistical

distribution of tin atoms between olefins with different stereochemical arrangements

(Table 4-C).



F -13.22

-12.6 -12.8 -13.0 -13.2 -13.4 ppm
63.48 31.46 5.07

Figure 4-12. 119Sn-NMR of polymer 75 synthesized using complex 32 and no external
activator (250:1). The signals correspond to Sn atoms between trans-trans, trans-cis and
cis-cis double bonds.

One of the most important characteristics of polymer 75 is its thermal behavior.

When heated under a nitrogen atmosphere above the onset decomposition temperature

elimination of the organic components takes place, and the weight of the residual material

accounts for the theoretical amount of metal introduced in the polymer. Similarly,

decomposition under an air atmosphere shows a slightly higher ceramic yield, attributable

to partial oxidation to tin oxide. This behavior had not only been observed in the samples

synthesized using the well-defined molybdenum catalyst 26 (Wolfe 1997) but also in

other tin-containing polymers (Imori 1995). This suggests that possible applications of

these materials could be based on this behavior, making of ADMET polycarbostannanes

precursors to ceramic materials and metallic thin films. However, the physical state of the

polymers synthesized to date impedes the casting of free-standing films, which limits

their applicability. Efforts to overcome this difficulty are discussed in a later section.

Table 4-C. Polymerization of stannadiene 73 using a well-defined catalyst (26)
(Wolfe 1997) and tungsten precatalysts 32 and 33.

M. transwcis
Cata. Experimental n Intensities
Conditions rat13 n13 ratie
Conditions GPC C-NMR (3C-NMR) Calcd. Obs.

26a RT-60 C, 48 h 36000 17000 79:21 64:32:4 65:30:5

32 85-100 C, 36 h 17000 9300 79:21 64:32:4 64:31:5

33 90-100 C, 48 h 30000 16000 80:20 64:32:4 63:32:5

a) Wolfe (1997), appended for comparison.


The second stannadiene studied was bis(3-butenyl)di-n-butylstannane, which

yielded both cyclic and linear oligomers with the well-defined molybdenum complex 26

(Wolfe 1997). Because of the removal of a methylene group between the olefin and the

tin atom, the metal content of the polymer increases (ca. 38 vs. 35 % Sn in polymer 75).

Complexes 32 and 33 are also activated when heated to 85-90 C in the presence

of diene 74 and form carbene species able to catalyze ADMET polymerization. However,

even if the same experimental observations described for polymer 75 (ethylene evolution

and increase in the mixture viscosity) are present, complexes 32 and 33 are unable to

drive the conversion yield above 75 %. Less than 20 % functional group conversion is

observed in the case of complex 32, whereas after 36 h of reaction complex 33 yields low

molecular weight oligomer, as well as ring closing metathesis product. The significant

decrease in activity observed upon reduction of the spacing length between the olefin and

the tin moiety with both well-defined (Wolfe 1997) and classical catalysts is still an

unexplained phenomenon. The ring closing metathesis product present in the described

reaction mixture, can be isolated as the main product when the reaction is carried out in

the presence of solvent ([diene] = 0.1 M in toluene, 100:1 substrate/precatalyst) (Figure


BuSnBu Bu Bu
33,85 0C Sn

74 81

Figure 4-13. Self-activated ring closing metathesis of bis(3-butenyl)di-n-butylstannane
(74) yielding the 7-membered ring 81.

Spectroscopic analysis of the resulting product (after catalyst and solvent removal

reveals the formation of the cyclic product. In turn, mass spectrometry reveals a

difference of 28 units with respect to the base peak in the CI-MS of the starting material,

which accounts for the molecule of ethylene released in the reaction. The yield based on

IH-NMR is ca. 83 %, and the incomplete conversion is evidenced by the starting material

signals present in both 1H- and 13C-NMR. This conversion constitutes the first example of

a substrate-activated ring closing iwiti/heii The 13C-NMR of the reaction product is

shown in Figure 4-14

a b
c Sn

i 140 ' ' ' '20 ' ' 'I ' ' 8'0 I 60 40 20 ppm

Figure 4-14. 13C-NMR of compound 81. The following impurities were present in the
crude preparation (after solvent removal) = Residual toluene, o = starting material.

New Polycarbostannanes

The synthesis of dibutyl derivatives by Wolfe and reported herein motivated the

search for new polycarbostannane structures. An extrapolation of the thermal behavior

observed for polymers 75 and 76 (Wolfe 1997) suggests that polymers designed with

higher metal content should lead to higher % of metal deposition (higher ceramic yields).