Synthesis of functionalized polycarbosilanes via acyclic diene metathesis (ADMET) polymerization

MISSING IMAGE

Material Information

Title:
Synthesis of functionalized polycarbosilanes via acyclic diene metathesis (ADMET) polymerization
Physical Description:
viii, 113 leaves : ill. ; 29 cm.
Language:
English
Creator:
Church, Anne Cameron, 1975-
Publication Date:

Subjects

Subjects / Keywords:
Silane compounds   ( lcsh )
Polymerization   ( lcsh )
Metathesis   ( lcsh )
Nucleophilic reactions   ( lcsh )
Chemistry thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2001.
Bibliography:
Includes bibliographical references (leaves 104-112).
Statement of Responsibility:
by Anne Cameron Church.
General Note:
Printout.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 028198098
oclc - 49314995
System ID:
AA00014258:00001


This item is only available as the following downloads:


Full Text










SYNTHESIS OF FUNCTIONALIZED POLYCARBOSILANES VIA ACYCLIC
DIENE METATHESIS (ADMET) POLYMERIZATION
















By

ANNE C.-AME RON CHLRCII


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA














ACKNOWLEDGMENTS


Many people have been instrumental in my career at the University of Florida.

First, I would like to thank my advisor, Professor Ken Wagener for all of his good advice

and for having confidence in my abilities as a chemist. Next, I would like to thank Dr.

Jim Pawlow, who worked in our research group as a post-doc for three years. During that

time and since, he has been a tremendous help, from taking the time to teach me

laboratory technique when I was a beginning graduate student to helping edit this

dissertation and being my sounding block on a variety of issues. Now, as my fiancee and

best friend, I cannot wait to start our lives together.

I would like to thank my parents and family for all of their enduring support

during my many years of education. They have made many sacrifices to get me to where

I am today and taught me at a young age what was important in life and how to work

hard to achieve my goals.

There are so many wonderful people on the Butler Polymer Floor that I have had

the good fortune to meet. Dean and Annie Welsh, C.J. and Sonya Dubois, and Carl and

Janet Gaupp have always welcomed me into their homes for many memorable get-

togethers. I have had the opportunity to work in an excellent research environment, and I

would like to thank all the members of the Wagener group, past and present, especially

Jason Smith (a fellow South Carolinian), Krystyna Brzezinska, Fernando Gomez, Pat

O'Donnell, John Schwendeman, and John Sworen and Tim Hopkins (my labmates). I








have to thank the polymer office staff for all of their help, in particular Lorraine

Williams, who keeps the office running so smoothly and who always finds a way to fix a

problem or answer the multitude of questions posed to her by so many.

I also thank Danielle Buchanan, Amy Boone, Jennifer Cottone, Joe McClellan,

and Jim Murphy for their emotional support and friendship and for all the entertaining

times we had during our time here.

Dr. Ion Ghiviriga and the NMR spectroscopic services staff were extremely

helpful and I would like to thank them as well. Finally, I would like to thank my

committee for all of their support and helpful ideas.














TABLE OF CONTENTS
Page


ACKNOW LEDGM ENTS .........................................................................................ii

A B ST R A C T ................................................................... ............................................... vii

CHAPTERS

1. IN TR O D U C TIO N .................................................................... ................................ 1

Historical Perspective ofOlefin Metathesis ................................. ...... ............... 1
The Four Types of Metathesis Reactions..................................................................... 2
Acyclic Diene Metathesis Polymerization............................................... 5
The ADMET Reaction Mechanism................................... ......................... 7
Three Types of Catalyst Systems for ADMET....................................... ............. 8
Silicon Polymers and ADMET Polymerization....................... ...................... 14
Reactive Polycarbosilanes and ADMET ......................... .......... ........... ... 16
Reactive Si-Cl Bonds and ADMET.............................. ........................ 17
Sum m ary ................................................. ......................... .............. .......... 23

2. NUCLEOPHILIC SUBSTITUTION AND ADMET
POLYMERIZATION OF DIALKOXYSILYLDIENES ........................................ 24

Introduction ...................................................................................... ........................... 24
Polymer Properties vs. Pendant Groups........................ .......................... 24
Synthesis of the Parent M onomer................................... ....................... 28
Grignard Chemistry and Halosilanes..................................................... 29
Synthetic A ttem pts............................................................................................ 31
Purification and Spinning Band Distillation........................ .................. ........ 34
Nucleophilic Substitution of the Parent Monomer ................................................. 37
Synthesis of the Dimethoxy Functionalized Monomer ....................................... 38
Using Metal Alkoxides as Nucleophiles........................ ....................... 39
Incorporation of Fluorine Containing Nucleophiles............................................. 39
Phenolic Nucleophiles ............................... ...... .......................... 41
Polymerization of the Substituted Monomers....................................... ............ .... 41
Polymerization of the 2,2,2-Trifluoroethoxysilyl Substituted Diene..................... 42
Polymerization of Di-4-pentenyldiethoxysilane....................................... ............ 45
Polymerization of Di-4-pentenyldimethoxysilane....................................... 47
Polymerization of Di-4-pentenyldiphenoxysilane................................................ 48
Conclusions and Comparisons................................................. 49










3. MACROMOLECULAR SUBSTITUTION USING
ALKYLLITHIUM REAGENTS ............................. ..... ............... 52

Introduction ..................................................................................... ......................... 52
Polymerization of the Parent Monomer......................... .......................... 54
Functionalization with Alkyllithium Reagents....................... ................... 57
Macromolecular Substitution Using Methyllithium .............................................. 58
Synthesis of the Phenyl Substituted Polymer Using Phenyllithium...................... 60
Using Butyllithium as a Nucleophile.............................. ......................... 64
Substitution Reactions of Comparative Polymer Systems........................................ 68
C onclusions.................................... .......................................... .......................... 70

4. HYDROLYTICALLY RESISTANT SILYL ETHERS............................................... 72

Introduction .................................................................................. .......................... 72
Sterically Bulky Aliphatic Nucleophiles ................................... .................... 74
Lithium Reagents as Nucleophiles................................................ 75
Nucleophiles Containing a Phenyl Group................................ ........................ 77
Diols as "Chelating" Nucleophiles........................ ....................... 78
Hydrolytic Stability of Silacycloalkanes .......................... ...................... 84
C onclusions............................................................................ ...... ........................ 85

5. EX PER IM EN TA L .................................................................. ............... ........... 87

M materials .................................................................................. .................. .......... 87
Instrum entation ........................................ ........................................................... 88
Synthesis of Substituted Dialkoxysilane Monomers ................................................. 89
Synthesis of di(4-pentenyl)dichlorosilane (26)..................................... ............. 89
Synthesis of di(4-pentenyl)dimethoxysilane (27)...................................... ............ 90
Synthesis of di(4-pentenyl)diethoxysilane (28)....................................................... 91
Synthesis of di(4-pentenyl)di(trifluoroethoxy)silane (29)................................... 92
Synthesis of di(4-pentenyl)diphenoxysilane (30)................................................... 93
ADMET Polymerization of Substituted Dialkoxysilanes......................................... 94
Synthesis of Poly[(di-4-pentenyl)dimethoxysilane] (31)....................................... 94
Synthesis of Poly[(di-4-pentenyl)diethoxysilane] (32).......................................... 94
Synthesis of Poly[(di-4-pentenyl)di(trifluoroethoxy)silane] (33).......................... 95
Synthesis of Poly[(di-4-pentenyl)diphenoxysilane] (34)........................................ 95
Preparation of Substituted Polymers via a One-Pot, Two-Step
ADMET Polymerization-Macromolecular Substitution............................................ 96
Synthesis of Poly[(di-4-pentenyl)dichlorosilane] (41) .......................................... 96
Synthesis of Poly[(di-4-pentenyl)diphenylsilane] (44)........................................ 96
Synthesis of Poly[(di-4-pentenyl)-n-butylmethylsilane] (43).............................. 97
Synthesis of Poly[(di-4-pentenyl)dimethylsilane] (42) ......................................... 98
Synthesis of Hydrolytically Resistant Substituted Silanes ....................................... 99
Synthesis of di-n-hexyl(di-2-propoxy)silane (46) ................................................ 99
Synthesis of di-n-hexyl(dibenzoxy)silane (49).................................................... 99








Synthesis of di-n-hexyl(diphenoxy)silane (50)................................................. 100
Synthesis of di-n-hexyl(2,4-dimethyl-2,4-pentanedioxy)silane (47)..................... 101
Synthesis of di-n-hexyl(1,3-cyclohexanedioxy)silane (48).................................. 102
Synthesis of di-n-hexyl(di-n-butyl)silane (51) .................................................. 102
Synthesis of di-n-hexyl(diphenyl)silane (52)............................ ................ 103

R EFER E N C E S ........................................................................ ................................ 104

BIOGRAPHICAL SKETCH................................................... ....................... 113














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

SYNTHESIS OF FUNCTIONALIZED POLYCARBOSILANES VIA ACYCLIC
DIENE METATHESIS (ADMET) POLYMERIZATION

By

Anne Cameron Church

December 2001


Chairman: Professor Kenneth B. Wagener
Major Department: Chemistry

A series of polycarbosilanes functionalized with both alkyl and alkoxy pendant

groups were synthesized using a combination of nucleophilic substitution and acyclic

diene metathesis (ADMET) polymerization. This polymerization is a step-growth

polycondensation reaction consisting of the reaction of an a,co-diene with a metal carbene

catalyst and driven by the removal of ethylene. The parent ,co-diene monomer used in

this system contains two highly reactive silicon-chlorine bonds that can be substituted

using a wide range of nucleophiles.

Two different pathways were employed to prepare these functionalized

polycarbosilanes. The first route utilized involved the nucleophilic substitution of both

silicon-chlorine bonds on the diene monomer with a variety of alcohols, creating a series

of functionalized monomers. The substituted dienes then were polymerized via ADMET,

producing a family of substituted poly(dialkoxycarbosilanes) that exhibit a wide range of








properties and variable hydrolytic stability, both of which are determined by the nature of

the pendant group. This route ensures complete replacement of all silicon-chlorine bonds

present, since substitution occurs on the monomer prior to polymerization.

The second synthetic route utilized involves functionalization of a polymer

containing silicon-chlorine bonds in each repeat unit. This reactive poly(dichloro-

carbosilane) is prepared by ADMET polymerization of the parent a,co-diene, which is

then modified via nucleophilic substitution, creating a family of derivative polymers.

The macromolecular substitution reaction of the silicon-chlorine bonds in the polymer

was achieved using a variety of alkyllithium reagents. Although this type of reaction is

more difficult on polymers compared to monomers, quantitative substitution is effected

using these potent nucleophiles, producing stable materials with varying properties.

The results presented demonstrate the dramatic effect that variation of the pendant

group placed along the polymer backbone has on polymer properties. The combination

nucleophilic substitution-ADMET polymerization is shown to be a highly versatile and

flexible route for the synthesis of functionalized polymers, since substitution can be

performed on the monomer or on the polymer. Using this strategy, polymer architectures

can be realized that are difficult or impossible to synthesize by alternative methods.













CHAPTER 1
INTRODUCTION


Historical Perspective ofOlefin Metathesis

In 1967, the expression olefinn metathesis" was first coined by Calderon" and is

defined in Ivin and Mol's seminal text Olefin Metathesis and Metathesis Polymerization

on page 1 as "the (apparent) interchange of carbon atoms between a pair of double

bonds" (Figure 1-1).6


R1 R2
catalyst 3R R2

R3 3 R
R 3 R4 R R4


Figure 1-1. The olefin metathesis reaction.


The advent of olefin metathesis in the mid 1950s7 evolved from the development

of transition metal based catalyst systems for the polymerization of a-olefins and

ethylene.8 Early metathesis catalyst systems were derived from the Phillips type

(heterogeneous supported metal oxides at high temperature) or the Ziegler-Natta type

(homogenous metal halide/alkyl aluminum mixtures at low temperature). These multi-

component catalysts, formed in situ, were ill-defined and difficult to understand.

Because of this, the mechanism and nature of the active catalyst species were not known

for some time. In an effort to study the mechanism of the olefin metathesis reaction,








the Dall'Asta group performed detailed labeling studies proving that olefin double bond

cleavage and subsequent reformation occurred during the reaction.912 In 1970, Herrison

and Chauvin discovered the mechanism of this double bond cleavage.13 They proposed

that metathesis involved the formation of a metallacyclobutane intermediate, formed by

the complexation of a metal carbene (the active catalyst) with an olefin. The formation of

the four-membered ring is a critical step in metathesis chemistry, and it is common to all

metathesis reactions. Although the existence of this species was questioned initially, the

intermediate and the mechanism are now widely accepted by the scientific community.

The first polymers produced by metathesis involved the ring opening of strained

cyclic olefins. For example, in 1963 the Dall'Asta laboratory produced high molecular

weight polypentenamer (2) via the ring-opening metathesis polymerization (ROMP) of

cyclopentene (1) using a catalyst system composed of a mixture of WCI6 with Et2AICI

(Figure 1-2).9-12



n
1 2


Figure 1-2. Dall'Asta's ROMP of cyclopentene.


The Four Types of Metathesis Reactions

Today, olefin metathesis has become an extremely versatile tool for the formation

of new carbon-carbon bonds in both small molecules and polymers. There are four main

types of olefin metathesis reactions, and they are illustrated below in Figure 1-3.

Ring opening metathesis polymerization (ROMP) was the first olefin metathesis

reaction discovered that produced high molecular weight polymers. This reaction is








driven by the thermodynamic release of ring strain present in the cyclic olefin monomer.

Polymerization occurs due to a chain addition type propagation mechanism and is an

exothermic reaction. The polymerization is usually terminated by the addition of a small

amount of ethyl vinyl ether or a carbonyl-containing molecule, which reacts with the

catalyst and destroys its activity.6 Ring opening metathesis polymerization has been

shown to be quite useful for synthesizing a variety of different polymers and

copolymers.'1416











R- ADMET -(-R + n


R1 CM
R, CM R2 +

<^R2


Figure 1-3. The four main types of olefin metathesis reactions.


In 1990, successful ADMET (acyclic diene metathesis) polymerization of a,o-

dienes was first performed by the Wagener group, which emerged as another important

metathesis polymerization pathway.1718 Acyclic diene metathesis is an example of step-

growth polycondensation that involves the reaction of an a,o-diene with an acid-free

transition metal based alkylidene (or carbene) catalyst. As with any polycondensation

reaction, high molecular weight material is produced by the release of a small molecule








during the reaction, driving the monomer-+polymer equilibrium towards polymer

formation.19 In the case of ADMET, ethylene is the small condensate that is removed

during the reaction under reduced pressure. Macrocycle formation occurs to a small

extent in all polycondensation reactions. Therefore, ADMET polymerizations are

typically performed in the bulk (neat) to minimize competing cyclization reactions, and

since this reaction is thermally neutral, no heat sink is necessary. Polymerization of oa,o-

dienes using ADMET is a useful method for synthesizing many unsaturated, linear

polymers and copolymers with a variety of different functionalities that are inaccessible

by other means.20

The interest of olefin metathesis for synthesizing small molecules was limited in

the early literature due to the incompatibility of many functional groups with the early

catalysts. Ring closing metathesis (RCM) involves the intramolecular reaction of two

olefin sites on the same molecule to form a new carbon-carbon bond, producing a cyclic

olefin structure. Many difficult ring closures can be performed by this route due to the

selectivity of metathesis catalysts for olefins over other functionalities present on the

molecule. Ring closing metathesis is a useful tool for natural product synthesis because

many different sizes of rings can be successfully synthesized. With the discovery of the

latest generation metathesis catalysts that illustrate an increased tolerance to functional

groups, interest in olefin metathesis for small molecule synthesis is increasing at a

phenomenal rate.21-28

Cross metathesis (CM) reactions are useful for the synthesis of small molecules as

well. A cross metathesis reaction is different from RCM because it involves the inter-

molecular exchange of the olefinic carbons on two acyclic alkenes with no cyclization.








In CM, two olefins that possess in many cases different functional groups are reacted

with a metathesis catalyst in order to produce a new olefin. This class of reactions has

also found use in higher olefin, pheromone, and other natural products syntheses.29'30


Acyclic Diene Metathesis Polymerization

Acyclic diene metathesis is a step-growth polycondensation reaction that

produces high molecular weight polymers by reacting an a.,o-diene with a metathesis

catalyst and is driven forward by the removal of a small condensate, ethylene. The

polymerization of acyclic olefins by this step-growth polymerization is a challenging

endeavor because it is governed by the same rules as any polycondensation reaction,

requiring high conversions (>99%) and no side reactions in order to produce high

molecular weight materials.9 In the mid-1980s, the Wagener laboratory used styrene as

a probe to determine whether vinyl addition or metathesis chemistry was occurring when

using WCl6/AlEtCI2.17'1 If metathesis occurs, stilbene would be the product, and if vinyl

addition is occurring, then polystyrene would be produced. Polystyrene was produced in

this reaction, which was caused by the Lewis acid present in the catalyst system.

Therefore, acid-free catalyst systems are necessary to promote metathesis chemistry and

eliminate vinyl addition.

In 1986 the discovery of a well-defined, acid-free, single-site catalyst by the

Schrock group enabled this polymerization to become a viable route for the synthesis of

unsaturated linear polymers.32"36 These first catalysts discovered were based upon the

early transition metals tungsten and molybdenum. Using these new catalyst complexes,

the Wagener group was able to successfully polymerize 1,9-decadiene (3), producing

polyoctenamer (4), by an ADMET polymerization (Figure 1-4). 1718








[_at----- -ca
-C2H4
3 4

Figure 1-4. The first ADMET polymerization of 1,9-decadiene.


Since ADMET polymerizations are reversible, they are conducted in the bulk

(neat) in order to maximize the molar concentration of monomer. This promotes a shift

in the equilibrium of the reaction from monomer towards unsaturated polymer formation.

Like many other polycondensation reactions, ADMET is performed under reduced

pressure in order to remove the ethylene condensate. Removing the condensate

irreversibly shifts the reaction equilibrium towards polymer formation and dramatically

increases the molecular weight of the product polymer. The same kinetics (second order

in monomer) and molecular weight distribution (Mw/M, n 2.0) are observed for ADMET

that are seen with other step-growth polycondensation reactions.37

The clean mechanistic chemistry of ADMET is a direct result of the well-defined,

single-site catalyst systems used. The only side reactions that occur during ADMET are

cyclization and olefin interchange reactions. Cyclization is generally observed in

polycondensation chemistry; therefore, an a,o-diene monomer usually consists of a ten

atom or longer chain in order to suppress the competing RCM cyclization reaction.

Monomer chains shorter than ten have an increased propensity to form stable five, six,

and seven-membered ring structures.38-4 Internal interchange reactions occur, similar to

that observed with polyester and polyamide condensation reactions.19 In metathesis,

these reactions involve an active catalytic species at the end of a polymer chain

exchanging with an internal double bond on another chain (Figure 1-5). Two new

polymer chains are the result.















+


Figure 1-5. Example of an olefin interchange reaction that can occur during ADMET.


The ADMET Reaction Mechanism

The catalytic cycle of ADMET is illustrated in Figure 1-6.41 A general reaction

scheme involving an a,a-diene with a metal alkylidene initiator is discussed for

simplicity.

Each step in this catalytic cycle is a true chemical equilibrium, and the ultimate

driving force for the reaction is the removal of ethylene. Initiation begins with the n-

coordination [A] of a metal alkylidene with one olefin site of an ta,o-diene, forming B.

An olefin exchange reaction through a retro [2+2] cycloaddition occurs, forming a new

metal carbene C, the first intermediate in the catalytic cycle. The metal-carbene unit is

now on the end of a monomer or growing polymer chain C, which then reacts with

another monomer or polymer chain to form the first metallacyclobutane D. The

distinguishing feature of the ADMET mechanism is the formation of two

metallacyclobutane intermediates D, F, whereas the ROMP cycle only contains one.

Cleavage of D through a retro [2+2] cycloaddition occurs, producing methylidene E

(M=CH2) and a dimer or polymer chain. The active catalyst E then reacts with another

monomer or polymer terminus to form the second metallacyclobutane F in a [2+2]

cycloaddition reaction. Upon productive ring cleavage of metallacyclobutane F, ethylene






8

is generated. A new alkylidene C is formed, bound to the growing polymer chain, which

regenerates the initial intermediate discussed above. At this point, one catalytic cycle is

completed, which is then repeated numerous times to generate high polymer.



__ R R
R R -L4 BL
+ LnM- ,
LnM=_, L J R,



H2C=CH2 LM R R

c



L F D
R R



"R^ y LnM=CH2 R R


Figure 1-6. General ADMET catalytic cycle.


Three Types of Catalyst Systems for ADMET

Three distinct catalyst systems have been used for ADMET polymerization,

ranging from the early, ill-defined catalyst mixtures to the recent early and late transition

metal based well-defined systems.

"Classical" catalysts (5) are ill-defined systems consisting of a mixture of two or

more components which form the active catalyst in situ.6 These systems are typically






9

composed of an early transition metal halide, such as WCl6 or WOCl4, with a main group

alkylating agent, such as Bu4Sn or EtAlC12, functioning as the activator (Figure 1-7).






C, y ,T1C
[ Cl C
+ Bu4Sn



5


Figure 1-7. An example of an ill-defined metathesis catalyst (5).42


These catalysts have limited utility because of the harsh reaction conditions they require.

High temperatures (~100 C) and the use of mild Lewis acids as activators make these

systems incompatible with many functional groups. Nevertheless, these catalyst mixtures

are cheap and easy to prepare, which makes them useful for some industrial applications

and under certain conditions. These classical catalysts have limited utility because the

active catalyst cannot be isolated, and consequently, catalyst concentration cannot be

quantified or monitored during the reaction.

The discovery of well-defined, single-component alkylidene catalysts

revolutionized the field of olefin metathesis. In the late 1980s, the Schrock laboratory

synthesized a series of molybdenum (6)35,36 and tungsten (7)32-34 based alkylidenes which

were found to be highly reactive catalysts for both ADMET and ROMP with no side

reactions detectable (Figure 1-8). The preparation of these complexes provided an

impetus for a significant research effort in metathesis chemistry.












N
CF3 M= Mo 6
FaC O... M=CH M= W 7
CH3 0
F C H3 CH3
FC- 3


Figure 1-8. The Schrock-type metathesis catalysts 6 and 7 used in ADMET.


Extensive study of these alkylidene complexes and their ligand set was completed

in order to optimize their catalytic activity. The tungsten version of this catalyst (7) was

the first catalyst utilized for ADMET.3234 However, more emphasis on the molybdenum

catalyst (6) is shown in the literature, and this is based upon several factors. The

molybdenum catalyst (6) is more prevalent because it is easier to synthesize and more

tolerant to functionalities such as carbonyl groups.43 The high reactivity observed from

these Schrock alkylidene complexes is a result of the high oxidation state of the metal

(+6) and the presence of electron-withdrawing fluorinated alkoxy ligands. The resulting

highly electropositive metal center favors strong olefin-metal interaction, of critical

importance if metathesis is to occur. The n-donor ability of the nitrogen and oxygen

based ligands present stabilizes the metal center, and the high steric bulkiness of these

ligands prevents decomposition through bimolecular reactions. The Schrock-type

alkylidenes are highly active, allowing for metathesis of both terminal and internal

olefins as well as ring-opening polymerization of low strain cyclic monomers.43 Even so,

these catalysts are limited by the substantial oxophilicity of the metal center, requiring

that all reactions performed use strict, anaerobic conditions involving Schlenk line or






11

glovebox manipulations. These catalysts are also intolerant to many different functional

groups, such as acids, alcohols, aldehydes, and ketones for the [W] version (7), and acids,

alcohols, and aldehydes for the [Mo] version (6).43

Researching more air, water, and functionality tolerant complexes, the Grubbs

group discovered a series of carbene complexes based on the late transition metal

ruthenium in the mid-1990s (Figure 1-9).44-50


PCy3

CI.,R=CH
PCy3

Cy= cyclohexyl

8

Figure 1-9. The first generation Grubbs' ruthenium catalyst 8.


Being less oxophilic, ruthenium carbene complexes are more tolerant to many functional

groups and preferentially react with olefins in the presence of acids, alcohols, aldehydes,

esters, and amides.50 The first generation Grubbs' ruthenium catalysts all incorporate two

bulky, basic, and electron donating PCy3 ligands with large cone angles and two chlorine

ligands.46 This combination, in conjunction with the reactive benzylidene moiety bound

to the ruthenium metal center, produces catalyst 8 that is quite reactive in olefin

metathesis. On the other hand, the first generation Grubbs' [Ru] catalyst (8) is slower

than the Schrock's [Mo] catalyst (6) by an order of magnitude, but the advantage of its

functional group tolerance and the ease of handling eclipses this limitation. The








discovery of the ruthenium metathesis catalysts enables olefin metathesis to be a viable

route for the synthesis of both functionalized small molecules and polymers of variable

architectures.

Mechanistic studies of ruthenium catalyzed metathesis show a dissociative

mechanism predominates which involves the dissociation of one PCy3 ligand followed by

olefin coordination.48'50 The presence of a highly active monophosphine intermediate

encouraged many research groups to evaluate other electron rich ligands in an attempt to

further optimize and improve this catalyst system. In 1999, the Herrmann laboratory

discovered a ruthenium catalyst based upon the first generation Grubbs' catalyst (8), but

the catalyst possessed two N-heterocyclic carbene (NHC) ligands instead of two PCy3

ligands (Figure 1-10).51'52 NHC-ligands are strong o-donors and much less labile than

the phosphine ligands.535 Replacement of both PCy3 ligands with NHC ligands

produces catalyst 9 that is much more robust, but no increase in activity was

observed.51 52


f=-\
Cy-N N -Cy
| ,,Cl
Cy Ruc ch

Cy-N N-Cy

Cy= cyclohexyl
9


Figure 1-10. The Herrmann catalyst (9) bearing two NHC-ligands.








Concurrently, the research groups of Grubbs and Nolan discovered in 1999 that

using only one NHC-ligand and retaining one phosphine ligand produces a significantly

improved metathesis catalyst 10 (Figure 1-11).52,56 Since the NHC-ligand is strongly

electron donating and sterically demanding, the presence of only one of these ligands

enhances the dissociation of the other phosphine ligand and also stabilizes the active

catalytic species. Replacing the one phosphine with the NHC-ligand dramatically

increases the reactivity and rate of the catalyst.


Mes-N N-Mes Mes- N N -Mes


/e R C Ru
PCy3 \^j PCy3 __

Mes= mesityl (1,3,5-trimethylphenyl)
10 11


Figure 1-11. The Grubbs second generation ruthenium catalysts 10 and 11.


In terms of activity, this second generation Grubbs catalyst (10) is more active

than the first generation ruthenium catalyst (8),56 and the improved saturated NHC-ligand

(11) is even more active than the unsaturated NHC-ligand (10).576' In fact, both of these

second generation Grubbs catalysts (10, 11) match or exceed, in some cases, the activity

of the Schrock alkylidene catalysts in both ADMET and ROMP reactions.62'63 In

addition, the NHC-ligated Grubbs catalyst is even more tolerant to functionality, air,

moisture, and temperature than the first generation catalyst.5"61 The discovery of this

new series of catalysts has caused an explosion of interest in research involving olefin








metathesis, particularly in the ring closing metathesis of substituted olefins and the

preparation of highly functionalized polymers.23'50

The discovery and use of both the Schrock and Grubbs metathesis catalysts has

been instrumental in the synthesis of various polymer architectures and functionalities via

ADMET.20 Once a desired a,(o-diene containing functionality or branching is prepared,

ADMET polymerization can be performed, which retains the functionality and specific

backbone sequence. Consequently, a broad range of polymers with designed

architectures can be prepared that are difficult or impossible to synthesize via other

routes.


Silicon Polymers and ADMET Polymerization

Polycarbosilanes, polymers with a backbone composed of both silicon and

carbon, do not exist in nature. Linear carbosilane polymers are useful materials due to

their thermal, electronic, and optical properties but are difficult to prepare.6466 The

discovery of the Schrock and Grubbs metathesis catalysts has allowed the silicon

functionality to be incorporated into polymer backbones quite easily. The Wagener

group has demonstrated that ADMET is a viable route for the synthesis of

polycarbosilanes, siloxanes, and other silicon containing polymers.6769 Utilization of

metathesis polymerization opens up a new route for the synthesis of these polymers that

avoids many of the limitations found in earlier synthetic methods.

The first linear unsaturated polycarbosilanes synthesized using ADMET involved

a polydimethylcarbosilane backbone.67 During this work, it was discovered that

methylene spacer groups are necessary between the silicon atom and the double bond

moiety in order for polymerization to occur. Divinyl substituted silanes would not








condense because of unfavorable steric interactions between the dimethylsilyl group

adjacent to the double bond, and the catalyst.67 These steric interactions inhibit formation

of the critical metallacyclobutane intermediate, thereby preventing metathesis from

occurring. Further, vinyltrimethylsilanes will not self-metathesize and produce a dimer,

as illustrated by Schrock.70 The addition of another methylene spacer group between the

silicon atom and the olefin site was required to produce linear, silicon containing

polymers (13) via ADMET polymerization (Figure 1-12).67


H3 7 H3

i [Icat] Si v
S -C2H4 n
CH3 2H4 CH3
12 13

Figure 1-12. The synthesis ofpolydimethylcarbosilane (13) via ADMET.


Organofunctionalized polysiloxanes are another important class of silicon

containing polymers, possessing a unique combination of properties ranging from

hydrophobicity, low-temperature flexibility, and thermal stability.71-74 The presence of

siloxane linkages in polymer chains as well as copolymers containing siloxane and olefin

units enables useful materials to be synthesized because they combine the advantages of

the hydrocarbon segment (increased mechanical properties of the polymer) with the

flexibility and hydrophobicity of the siloxane segment.717 A series of this type of

polymer (15) was synthesized using ADMET by altering both the number of methylene

spacer groups between the double bond and the silicon atom and the length of the

siloxane segment (Figure 1-13), demonstrating the compatibility of metathesis chemistry

with these functionalities.68








CH3 CH3 CH3 CH3
I I [cat] I I
_"Si-O-Si S1 i-O- .i
SI -C2H4 I I .. n
CH3 CH3 CH3 CH3
14 15


Figure 1-13. The synthesis of a polycarbosiloxane (15) via ADMET.


Reactive Polvcarbosilanes and ADMET

An intriguing aspect of metathesis polymerization is that many reactive functional

groups are inert to the catalysts used and remain intact after the reaction. The catalysts

used in ADMET polymerization, the Schrock and Grubbs families of single-site catalysts,

are the reason behind this behavior, for without the discovery of these well-defined

selective catalysts, the synthesis of functionalized polymers via ADMET could not be a

reality.

A class of polymers currently under investigation are recognized as latent reactive

polymers, which possess silicon-methoxy groups.75 The Si-OCH3 bond is inert to both

the [Ru] (8, 10, 11) and [Mo] (6) metathesis catalysts, but this bond is hydrolytically

unstable and reacts slowly over time with atmospheric moisture, crosslinking and

releasing methanol. In order to produce a material that is both flexible and has good

mechanical strength, a monomer containing hydrolytically reactive Si-OCH3 groups (16)

was copolymerized with a hydrolytically stable a,co-diene monomer containing a flexible

siloxane segment (17) (Figure 1-14).

Since each monomer has equal reactivity, a statistical copolymer (18) was

synthesized composed of a "soft" siloxane segment with a "hard" segment containing the

crosslinkable Si-OCH3 groups. Further research effort in order to improve the








mechanical properties of these copolymers is ongoing, and the synthesis of these

polymers via metathesis demonstrates the viability of employing reactive groups on

polymer backbones.


oH3 OCH3 CH3 CH



OCH I
OCH3 CH3 CH3
16 [cat] 17
-C2H4


OCH3 OC\ CH3 CH3
S S ---Si

O0H3

18


Figure 1-14. ADMET copolymerization of a "hard" and "soft" monomer.


Reactive Si-Cl Bonds and ADMET

Even more advantageous is the inherent compatibility of metathesis catalysts with

the highly reactive Si-Cl bond. The Si-Cl bond is highly reactive to nucleophiles and is

easily substituted, and this chemistry is widely utilized in industry to prepare large

quantities of functionalized silanes.76 Incorporation of reactive Si-Cl bonds along a

polymer backbone would serve as "handles" for further functionalization, producing

materials with a wide range of applicability and properties. Previous work by the

Wagener group involving the synthesis of chlorosilane oligomers, has proven that the Si-








Cl bond is inert to Schrock's [Mo] catalyst (6) under ADMET conditions.77'78 Linear,

unsaturated, reactive polydichlorocarbosilanes can be designed and synthesized using this

knowledge.

A suitable a,0o-diene monomer containing reactive Si-Cl bonds was necessary in

order to synthesize dichlorocarbosilane polymers. Molecules of this type are not

commercially available, so several pathways were evaluated to enable the synthesis of a

parent ca,o-diene containing the appropriate dichlorosilane moiety. Ultimately, Grignard

chemistry was successfully utilized to produce the parent monomer. After successful

synthesis of this monomer, ADMET polymerization was performed, producing a

poly(dichlorocarbosilane). This reactive polymer is not useful by itself without further

modification because rapid hydrolysis of the Si-Cl bonds occurs, producing corrosive

HCI gas and forming a crosslinked siloxane network polymer (Figure 1-15). However,

the reactive polymer is useful as a precursor polymer backbone.



0

CI ( Ii
Si H20 0 O
SI hydrolysisis) I


O



Figure 1-15. Hydrolysis of the reactive poly(dichlorocarbosilane), producing a
crosslinked siloxane network polymer.






19

Obtaining an understanding of the substitution chemistry of the Si-Cl bonds is the

first step in the preparation of functionalized polymers. A model study was performed on

a substrate that would mimic a polymer repeat unit, where various nucleophilic reagents

were reacted with a long chain dialkyldichlorosilane in order to evaluate the nucleophilic

substitution reaction on a R2SiCI2 site.79'80 The second substitution of the Si-Cl bond is

kinetically much slower than the first Si-Cl substitution. Additionally, both steric and

electronic factors play a significant factor in the substitution reaction and determine

whether mono or disubstitution is observed. Small, sterically undemanding nucleophiles

such as methanol and ethanol are able to be disubstituted, whereas bulky nucleophiles

such as t-butanol only allow one Si-Cl bond to be substituted. In the case of weaker

phenolic nuclephiles, the presence of electron-donating groups allows disubstitution to

occur, while electron-withdrawing groups promote only monosubstitution. A detailed

study of this nucleophilic substitution reaction and the chemistry entailed has been

discussed elsewhere.79

Knowledge of the nucleophilic substitution chemistry of Si-Cl bonds learned

during the model study was instrumental in the synthesis of substituted polycarbosilanes.

The goal of this research was to produce a series of polycarbosilanes with differing

pendant groups by exploiting the reactivity of Si-Cl bonds in nucleophilic substitution

reactions. By replacing the chlorine atom with different groups, both the stability and the

overall properties of the final functionalized polymer will be altered, but the parent

polymer backbone remains the same.








The initial route studied to produce substituted polycarbosilanes involved the

nucleophilic functionalization of the monomer, followed by ADMET polymerization

(Figure 1-16).


Cl
I

Si
substitution i n polymerization
Nuc Cl
I -I

NNuc u IC
Nuc
polymerization / substitution
Si

Nuc n /


Figure 1-16. Two routes to produce substituted polycarbosilanes using both nucleophilic
substitution and ADMET polymerization.


The advantage of this route is that quantitative substitution of each Si-Cl bond is not

necessary because the monomer is purified prior to polymerization. This step ensures

complete replacement of all Si-Cl bonds. Using this strategy, alcohol nucleophiles were

first employed to replace the Si-Cl bonds with Si-OR bonds, and then reactive polymers

containing Si-OR bonds were synthesized by ADMET polymerization of the

functionalized diene monomer. The rate of hydrolysis of the resulting polymer pendant

groups is governed by the identity of the OR group chosen.

Such materials would be useful for a variety of applications. For example, use of

a reactive alkoxy group would enable slower hydrolysis to occur, producing a stable

crosslinked network material after polymer shaping and processing. This process, known








as room temperature vulcanization (RTV), is widely used to produce a variety of

industrially important materials.81

Functionalization of the dichlorosilyl group was first performed on the parent

a,o-diene monomer using several different alcohols. Since the ruthenium metathesis

catalysts are inert to and compatible with the functional groups substituted on the diene,

ADMET polymerization was utilized in the preparation of these substituted polymers.

The details are discussed in Chapter 2. A series of polymers with designed architectures

was produced with very different properties and resulting hydrolytic stabilities.

The second route utilized for producing substituted polycarbosilanes was

macromolecular substitution of a preformed, reactive polymer backbone containing two

Si-Cl bonds per repeat unit videe supra). The advantage of this route is that an entire

series of polymers can be synthesized from the exact same parent polymer backbone,

creating a family of polymers with different properties. A parent dichlorocarbosilane

polymer backbone was first synthesized, and then substitution reactions upon the polymer

backbone were performed. Macromolecular substitution reactions can often be quite

difficult because of limited accessibility of the Si-Cl bonds present in the polymer. It is

critical to substitute each and every Si-Cl bond on the polymer backbone in order to

produce a soluble, useful material. Any residual Si-Cl bonds left on the polymer

backbone will hydrolyze upon exposure to the atmosphere, causing crosslinking of the

polymer to occur, and an insoluble network will be formed. Using alkyllithium reagents

as nucleophiles, successful comprehensive macromolecular substitution was performed,

producing a series of stable, linear, unsaturated polycarbosilanes by a combination

ADMET-macromolecular substitution route. These studies are detailed in Chapter 3.








The materials produced are hydrolytically stable because carbanion nucleophiles were

utilized, replacing Si-Cl bonds with stable Si-C bonds.

As discussed above, impeding the hydrolysis of alkoxy functionalized polymers

would allow for useful polymers as-is, as well as providing more time to process these

materials before any crosslinking could occur. Hydrolysis could still be beneficial;

however, the rate of which would be slowed considerably. Chapter 4 details the synthesis

of model compounds with improved hydrolytic stability versus the substituted monomers

employed in Chapter 2. These model compounds utilize nucleophiles that are sterically

bulky or form stable rings, producing Si-OR bonds that more resistant to hydrolysis

(Figure 1-17).



SOH OH
+Si + \ /0
"x + amine Si



Figure 1-17. The synthesis of silacycloalkanes.


Alcohols such as isopropanol, phenol, and benzyl alcohol, as well as certain diols, were

shown to effectively react with both Si-Cl bonds of a dichlorosilyl group. Certain

reagents containing two OH groups per molecule were able to react with both Si-Cl

bonds on the same silicon atom, producing stable cyclic structures.

All of the nucleophilic systems studied produce products with improved stability

and consequently, can be worked up in the atmosphere. Utilization of these functional

groups as pendant groups would presumably produce polymers with increased stability

and considerably different properties than the polymers synthesized in this work.








Summary

A series of unsaturated, functionalized polycarbosilanes were synthesized using

ADMET polymerization. The flexibility of this synthetic route demonstrates not only the

inherent compatibility of modem metathesis catalysts used in the polymerization, but the

ability of ADMET to prepare polymers with designed architectures and functionalities

that are difficult or impossible to prepare by other means. Tolerance of the molybdenum

(6) and ruthenium (8, 10, 11) catalysts to both silicon-chlorine and silicon-alkoxy bonds

respectively, allow the utilization of both of these groups in ADMET. Exploitation of

known chlorosilane substitution chemistry enabled successful, quantitative substitution to

be realized, producing fully substituted polymers. Two different routes were successfully

employed to produce these functionalized materials: polymerization followed by

functionalization and functionalization prior to polymerization. Few other polymer

synthesis routes are described in the literature that approach the same level of flexibility.

The results described in this work demonstrate that ADMET polymerization is not only

an extremely versatile route to functionalized polycarbosilanes, but polymer properties

can be significantly modified by the nature of the polymer pendant group.














CHAPTER 2
NUCLEOPHILIC SUBSTITUTION AND ADMET POLYMERIZATION OF
DIALKOXYSILYLDIENES


Introduction

Alkoxysilanes are a class of compounds that enjoy a wide variety of applications

ranging from biomedicine to lubricants and coatings.82 89 Even so, their applications are

dependent on the hydrolytic stability of the alkoxy group incorporated. For example,

materials such as thin shell casings, resins, coatings, and low heat glasses require

hydrolysis of the alkoxysilane incorporated in the final processing step." Other materials

require the alkoxysilane groups to remain intact, and these alkoxysilanes are useful for

materials such as lubricants, hydraulic fluids, and diffusion-pump oils.90


Polymer Properties vs. Pendant Groups

The relationship between pendant group and product application can also be

correlated in polymeric systems. The best example of this in the literature involves the

Allcock group, who have studied a family of polymers based upon a reactive

phosphazene backbone (20) (Figure 2-1).94'00


CI ci
5 c/ R
N P'N 2500C R /
CI-pN- cI N- N4 NaOR N=P
CI OR
19 20 21

Figure 2-1. Macromolecular substitution ofpoly(phosphazene).








The parent poly(phosphazene) (20) contains two P-Cl bonds per repeat unit, and the

resulting dichloropolymer is hydrolytically unstable.92 However, once functionalized,

poly(phosphazenes) (21) become stable materials that can be useful for a wide variety of

applications. The pendant group utilized governs the application of the resulting

material. A myriad of pendant groups and combinations of different pendant groups have

been successfully attached to this polymer backbone, forming a series of materials with a

wide range of properties.91-94 In this system, functionalization via macromolecular

substitution must occur after the ring opening polymerization of the parent monomer has

been complete. The reason for this is that substitution of the cyclic hexachloro-

phosphazene trimer (19) prior to polymerization prevents ring opening from occurring

and no high polymer forms.9

Table 2-1 illustrates that by changing the pendant group of the polymer, the

properties of the polymer produced can be vastly different.91'92 Many other pendant

groups have been studied as well.


Table 2-1. Relationship of polymer properties vs. pendant group in poly(phosphazenes).

Pendant Group Properties TM/Tm (C)
OCH2CH3 elastomer -84 / a

OC6H5 hydrophobic, -8 /+390
microcrystalline thermoplastic

OCH2CF3 hydrophobic, -66/ +242
microcrystalline thermoplastic

O(CH2)20(CH2)20CH3 water-soluble elastomer -84 / a

NHCH3 glass, water-soluble +14/ a

a) no Tm observed






26

For both of these reasons, an effort was directed towards the synthesis of a series

of analogous alkoxy-substituted polycarbosilanes in order to study the effects the

variation of a particular pendant group would have on the properties of this type of

polymer. Each material synthesized would contain the exact same backbone but with

different reactive alkoxy pendant groups. The polymer properties and atmospheric

stability will be strongly influenced by the individual pendant group chosen. Such a

series of polymers would have potential uses for a variety of applications such as hoses,

belts, and conformable seals for gas masks (Figure 2-2).



9



O= reactive functional group


Figure 2-2. General structure of a functionalized polycarbosilane.

The degree of hydrolytic stability of the polymer is completely dependent upon

the pendant group. The parent silane monomer that is used contains two highly reactive

Si-Cl bonds. After polymerization, the material produced is not very useful because

rapid crosslinking of the hydrolytically sensitive Si-Cl bonds present in each repeat unit

occurs. This hydrolysis reaction produces a siloxane polymer network which is insoluble

and cannot be shaped or processed. In addition, corrosive HCI gas is liberated, which is

highly undesirable. Because of this phenomenon, functionalization of the Si-Cl bonds

with stable alkoxy or alkyl groups is necessary in order to produce a useful material.






27

As seen in previous model studies,79'80 small, sterically unhindered alkoxy groups

such methoxy and ethoxy, undergo facile substitution of Si-Cl bonds. These groups are

commonly used in a variety of alkoxysilane polymers. The 2,2,2-trifluoroethoxy group

was evaluated because fluorinated alkoxy groups have been shown to produce materials

that are highly water-repellent and resist swelling by petroleum fluids.92 94 Further,

phenoxy groups attached to silicon also tend to be more resistant to hydrolysis as well.

This phenomenon has been discussed in the literature, but no one has offered a

reasonable explanation based on either sterics or electronic effects.10' These pendant

groups were chosen for this study because of the differences in polymer properties they

would impart, but also they would produce polymers with different hydrolytic stabilities.

After substitution of the Si-Cl bonds of the parent diene with the appropriate

nucleophile, the functionalized monomers were polymerized using ADMET to produce

high molecular weight materials of increased stability. Depending upon its ultimate use

and choice of pendant group, the resulting polymer could be hydrolytically stable or

crosslink slowly. By synthesizing completely substituted monomers prior to

polymerization, all the reactive Si-Cl bonds have been functionalized, eliminating

incomplete substitution of the polymer backbone. Incomplete substitution would be

detrimental to the material because any residual Si-Cl bonds would crosslink when

exposed to the atmosphere, producing an insoluble, intractable, brittle material.

Metathesis catalysts such as Grubbs' ruthenium catalysts (8, 11) are not affected by the

functionalities present on these pendant groups, thus allowing the functionalized






28

monomers to be easily polymerized. This inherent flexibility is a significant advantage to

ADMET polymerization, since functionalization is not limited to macromolecular

substitution reactions.


Synthesis of the Parent Monomer

In order to produce functionalized monomers, a synthetic route to an ADMET-

capable a,co-diene containing highly reactive Si-Cl bond(s) was desired. Therefore, the

synthesis of this type of monomer was undertaken. There are two practical routes to

produce new Si-C bonds: hydrosilation and Grignard chemistry.8s5'02'10 Initial attempts

to synthesize the parent monomer 23 involved hydrosilation chemistry.04,'05 Dichloro-

silane gas was condensed and reacted with excess 1,5-hexadiene (22) in the presence of

Speier's catalyst (H2SiC16) (Figure 2-3).1"o102


Cl
H2PtCl6 I
+ H2SiCI2
"Speier's catalyst"
22 CI
+ cyclics and internal isomers formed
23


Figure 2-3. Hydrosilation of a chlorosilane and a diene.


Although platinum catalysts are well known for their efficiency in hydrosilation

chemistry, it is not a route without complications when using dienes.o3 The hydro-

silation reaction can produce cyclic species from an intramolecular reaction as well as

low molecular weight polymers from intermolecular addition. In addition, internal olefin

bond formation is facile, a direct consequence of using a late transition metal catalyst,








which is known to promote olefin migration via p-hydride elimination.'03 All of these

side reactions were observed in the reaction of 1,5-hexadiene (22) with H2SiCI2.79 A

mixture of dialkenyldichlorosilane isomers were separated from the cyclics and polymer

by distillation. Due to the trivial differences in their boiling points, the desired a,o-diene

(23), which was produced in approximately 10% selectivity, could not be isolated from

the other isomers. Varying reaction conditions and using alternative platinum hydro-

silation catalysts such as Kardstedt's, Pt(divinyltetramethyldisiloxane) in xylenes, were

ineffective at completely suppressing this isomerization reaction. Therefore, this

synthetic route was abandoned as being impractical.


Grignard Chemistry and Halosilanes

Since ADMET polymerization is a step-growth polycondensation reaction,

absolute monomer purity is required in order to achieve high molecular weight

material.1'19 Because of the difficulty of purification of the desired a,co-diene using

hydrosilation chemistry, Grignard chemistry was evaluated in order to produce a useful

monomer. The molecule targeted for synthesis via this route was a ten carbon, a,c-diene

containing a dichlorosilyl moiety in the center, di-4-pentenyldichlorosilane.

The Grignard reaction of a halosilane is quite different in reactivity than its

carbon analog. It is well known that a Grignard reaction involving chlorosilanes is rarely

stepwise and that multiple substitution of the substrate is favored in most instances.108

The reason for this increased reactivity is not based upon the differences in bond

strengths, because a Si-Cl bond is stronger than a C-Cl bond as shown in Table 2-2.

However, a significant difference lies the in length of both the Si-CI and the Si-C bonds,








which are both longer than its carbon analog, causing the silicon atom to be more

exposed to an attacking nucleophile.103


Table 2-2. Bond strengths and bond lengths for various silicon and carbon analogs.83'109

Bond (Si-X or C-X) Bond Strengths (kJ/mol) Bond Lengths (A)

Si-Cl 472 2.05

C-C1 335 1.78

Si-C 369 1.87

C-C 334 1.53



The two major disadvantages of this Grignard route are the general inability to

produce clean, integral, partial substitution and that the removal of undesired corrosive,

hydrolytically sensitive by-products is not trivial.'08 The advantage of this route is that it

produces silanes with the terminal olefins intact with absolutely no isomerization.

Depending upon whether complete or partial substitution is desired, the sequence of

addition of reagents becomes significant.'0 If full substitution of a halosilane with a

Grignard reagent is desired, a "normal" addition is advantageous. A "normal" addition

involves addition of the halosilane directly to a solution of Grignard reagent. On the

other hand, if partial substitution is necessary, then a "reverse" addition is utilized. This

method involves the addition of a desired number of equivalents of Grignard reagent to a

halosilane solution.

The other important factor of this reaction involves the miscibility of the Grignard

reagent with the substrate solution. This is of critical importance because the reaction of

the halosilane with the Grignard reagent must be slowed down in order to have a chance






31

of producing useful amounts of the desired partially substituted product. The less soluble

the Grignard reagent is with respect to the solution of halosilane, the better the chance of

partial substitution. Rosenburg et al. have demonstrated this by using heptane as the

reaction solvent for the Grignard reaction involving arylchlorosilanes.'1 In this work,

they found that for the reaction of 2.1 equivalents of phenylmagnesiumbromide with

SiC14, 77% of disubstituted product was produced. The selectivity is the highest seen in

the literature to date. This methodology was a key factor in the successful synthesis of

di-4-pentenyldichlorosilane using Grignard chemistry.


Synthetic Attempts

Early attempts to synthesize this molecule used Et20 as the reaction solvent for

both the formation of the Grignard reagent and for its reaction with the halosilane. The

magnesium turnings used to form the Grignard reagent were kept free of oxide and

moisture by storage in a vacuum oven before use. The magnesium was activated by a

process called entrainment, involving the addition of a small amount of 1,2-dibromo-

ethane."',"2 Apparently, this reactive organic halide cleans the surface of the

magnesium metal and activates it, forming MgBr2 (the activator) and ethylene, which is

released in the process. After formation of the Grignard reagent, which involved the

reaction of 5-bromo-l-pentene (24) with the Mg turnings, titration of RMgBr (25) was

performed according to the method described in Vogel's to ensure the correct number of

equivalents was added to the SiC14 (Figure 2-4).'13 Typically, between 80-85% yield of

Grignard reagent was formed in this reaction. In diethyl ether solvent, reacting SiC4

with the alkenyl Grignard reagent (25) using reverse addition produced about 50%

selectivity for disubstitution.









r + Mg Et20 ,,,-, MgBr
A
24 25
Cl
solvent
-..,..-MgBr + SiCl4 A i

2.1 equivalents Cl
25 26


Figure 2-4. Synthesis of di-4-pentenyldichlorosilane (26) via Grignard chemistry.


Gas chromatography of the reaction mixture illustrated a peak shoulder directly

after the product peak. Carbon-13 NMR analysis of this species showed a resonance at

64.0 ppm, which is characteristic of a carbon atom bonded to an oxygen atom (Figure 2-


* = by-product peaks


1604014 120 100 80 60 40 20 ppm


Figure 2-5. The 13C NMR spectrum of 26, when diethyl ether is used as the solvent
(post-distillation).






33

The 'H NMR spectrum displayed a quartet at 3.87 ppm and a triplet at 1.65 ppm, which

can be attributed to ethoxy group protons. Obviously, all of this evidence pointed to the

formation of an undesirable by-product, an ethoxy-substituted silane.

Occurrence of by-product formation is attributed to diethyl ether reacting with

SiC14 during the Grignard reaction. Literature precedence for this phenomenon is over

100 years old"4 and has been observed by others as well.I"5"7 In addition to the Si-OEt

bonds being formed, as proven by the NMR spectra discussed above, Si-Et bonds can

also be produced. This can occur because EtCI by-product, created from the Si-OEt bond

formation, can react with any residual magnesium."t7 This produces EtMgCI, which then

reacts with the Si-C1, forming new Si-Et bonds. Whitmarsh and Interrante have proven

this phenomenon by preparing their chlorocarbosilane polymers in deuterated Et20 and

performing 2H NMR analysis."17 Neither by-product, ethoxy nor ethyl substituted diene,

can be effectively removed from its chloro analog by distillation. Hence, the use of

ethereal solvent for this reaction is not desired.

Because side reactions occur when using Et20 as the reaction solvent, it was

necessary to replace Et20 with heptane as the reaction solvent. Heptane does not react

with Si-Cl bonds and form any undesirable products, and in addition, greater selectivities

result when using this solvent. Using methodology modeled on Rosenburg's system,

excess (7.27 g, 0.30 mol) Mg turnings and 5-bromo-l-pentene (40.50 g, 0.27 mol) were

reacted in 175 mL Et20, producing 82% yield of Grignard reagent (0.22 mol). Reverse

addition of the Grignard reagent solution (2.1 equivalents) dropwise over 6 hours to SiC14

(16.86 g, 0.10 mol), dissolved in an excess of 500 mL heptane (2:1 heptane/ether),

followed by refluxing overnight under Ar, produced the desired product. Complete






34

suppression of the formation of any Si-OEt and Si-Et bonds occurred, evidenced by GC,

NMR, and elemental analyses. Using GC analysis, 73% selectivity towards di-4-

pentenyldichlorosilane was observed, with 24% trisubstituted product and 3%

monosubstituted product formed. No tetrasubstituted product was observed.


Purification and Spinning Band Distillation

Absolute isolation of the diene monomer from its reaction by-products is of

critical importance if the synthesis of linear, soluble polymer is to be achieved. Any

residual monoene, if reacted under ADMET conditions, would act as a chain termination

agent in an ADMET polymerization, and any triene or tetraene present would react at

every terminal olefin site, forming an insoluble network polymer (Figure 2-6).

During the dialkenyldichlorosilane monomer synthesis, three different alkylation

products were produced videe supra). Mono-, di-, and trisubstituted products were

formed and separation of these compounds is not a trivial task.


Cl
S chain terminator
Si -CI I

Cl


Cl
I network

Si


SS Si-Cl




Cl


Si


y


Figure 2-6. The outcome of mono and trisubstituted olefins in the reaction mixture.








Chlorosilanes are corrosive and hydrolytically unstable materials, and in addition,

vacuum distillation must be employed to distill these particular compounds due to their

high boiling points. The boiling point of the monoene is around 220 OC at room

temperature, the diene 260 C, and the triene is slightly above 300 C. As the distillation

pressure is lowered, the efficiency of all columns, hence separation, is significantly

reduced, often as much as 50-70%. Generally, the number of plates required to separate

two different compounds is equal to 120/AT, where AT is the difference in boiling point

between the two compounds.'13 Since the difference at 4 Torr between mono-, di-, and

trisubstituted dienes is approximately 15 C per species, a minimum of eight distillation

plates would be required to separate these components. Traditional laboratory Vigreux

columns have three to six distillation plates, and it was not possible to separate the

mixture. Therefore, a spinning band distillation column equipped with a Teflon band was

employed in order to purify these materials."8"12 The Teflon band does not react with

chlorosilanes, and the apparatus employed was rated at 30 distillation plates at ambient

pressure. Even with reduced efficiency at lower pressures, the column is capable of

separating this mixture. Since maintenance of pressure is crucial during vacuum

distillation, a manostat containing dibutylphthlate was connected to the system in order to

keep the pressure constant at 4 Torr during the entire distillation.'22 Constant heating was

controlled using a thermocouple monitored heating mantle connected to a Variac constant

voltage regulator. Teflon boiling stones were used to order to prevent any superheating

or "bumping" of the mixture.







36

After three hours of equilibration, consisting of constant reflux (30 drops/min) at

the boiling point, the spinning band distillation column demonstrated its superior

efficiency in purifying the monomer (Figure 2-7).


Cl
2 4

1 3 5
CI


...... ... .. .. .. .. .. ....... ..- .......................- -,. .. .. iL LIl


160 W40 120


100 B0 60 40 20 PPI


Figure 2-7. The 3C NMR spectrum of di-4-pentenyldichlorosilane (26).


Using a reflux ratio of 15:1, separation of the monomer, di-4-pentenyldichlorosilane (26),

was achieved (BP: 90-92 OC/ 4 Torr), yielding 38% disubstituted product. Calculated

yields by gas chromatography of the disubstituted product were much higher (73%

selectivity), and the reduced yields obtained are attributed to distillation losses,

hydrolysis polymerization of the chlorosilane from slight traces of moisture present, and






37

discarding of intermediate distillation fractions. The successful separation of the parent

monomer is illustrated by the absence of any impurity peaks in the 'H, "C, and 29Si NMR

spectra and in the GC trace.


Nucleophilic Substitution of the Parent Monomer

After the di-4-pentenyldichlorosilane parent monomer (26) was successfully

isolated, four functionalized monomers were synthesized via nucleophilic substitution of

both Si-Cl bonds. The nucleophiles utilized were judiciously chosen based upon

information discovered in a previous model study.79'80 Four nucleophilic alcohols

(MeOH, EtOH, CF3CH20H, and PhOH) that would easily substitute both Si-Cl bonds

and produce polymers that exhibit significantly different thermal properties and

differences in hydrolytic stability were chosen. Based upon previous experience, the

anticipated hydrolytic stability of the resulting products should range from dimethoxy-

silane (least stable) to diphenoxysilane (most stable). All four monomers were

synthesized using two similar methodologies, by reaction of a dichlorosilane (26) with

excess equivalents of an alcohol/amine or metal alkoxide nucleophile (Figure 2-8).


Cl OR

/ Si Et2O ~ -~tSi
r'3fI N + ROH/Et3N or KOR A- asI
Cl OR

R= OCH3 27
OCH2CH3 28
OCH2CF3 29
OC6H5 30

Figure 2-8. Functionalization of the dienes via nucleophilic substitution.








Synthesis of the Dimethoxy Functionalized Monomer

The dimethoxysilane containing a,co-diene (27) was synthesized by a one-step

reaction of di-4-pentenyldichlorosilane (26) with two equivalents of both MeOH and

Et3N per Si-Cl bond (Figure 2-9).



OCH3
2 4 6

1 3 5
OCH3
6





5
3






91 7 6 5 4 3 2 PP


Figure 2-9. 'H NMR spectrum of di-4-pentenyldimethoxysilane (27).


Triethylamine serves only as a proton acceptor, as it is sterically inhibited enough to be

considered non-nucleophilic, and it is not a strong enough base to deprotonate the

alcohol. Diethyl ether was used as the solvent because the triethylamine hydrochloride

salts produced are reasonably insoluble in Et20, making purification by Schlenk filtration

easier. After refluxing overnight under an inert atmosphere and subsequent work-up, the

monomer was purified by vacuum distillation in 58% yield.








Using Metal Alkoxides as Nucleophiles

The diethoxysilane functionalized monomer (28) was synthesized by using KOEt

as the nucleophile. Potassium ethoxide is a much stronger nucleophile than EtOH/NEt3;

however, either system is effective for this substitution reaction. The reason that KOEt

was used in this reaction was to illustrate the versatility of the nucleophilic substitution

reaction on these types of compounds and the viability of alternative reagents. The

amount of KOEt utilized was two equivalents per Si-Cl bond, and this reaction required

the use of THF as the solvent due to the solubility of the KOEt in this medium and its

insolubility in Et2O. The substitution of both Si-Cl bonds occurred readily, and this

monomer was isolated after vacuum distillation at 0.005 Torr in 44% yield. Only one

peak at -7.9 ppm was observed in the 29Si NMR spectrum, which verifies formation of a

dialkoxydialkenyl silane and that complete substitution has occurred (Figure 2-10).123'124


OCH2CH3



OCH2CH3





30 20 10 0 -10 -W PK

Figure 2-10. The 29Si NMR spectrum of di-4-pentenyldiethoxysilane (28).


Incorporation of Fluorine Containing Nucleophiles

The di(2,2,2-trifluoroethoxysilyl) functionalized monomer (29) was synthesized

because of the hydrophobicity of the 2,2,2-trifluoroethoxy entity. Further, this pendant








group has exhibited an increased resistance to hydrolysis in other reactive polymer

systems.92-94 Although the nucleophilic strength of 2,2,2-trifluoroethanol is weaker than

the other two nucleophiles discussed previously, substitution of both of the Si-Cl bonds in

di-4-pentenyldichlorosilane still occurs. A fourfold excess of the 2,2,2-trifluoroethanol

and Et3N in Et20 was reacted with di-4-pentenyldichlorosilane, and after reaction

workup, the monomer was purified by vacuum distillation (57% yield). Nuclear

magnetic resonance spectroscopy, 1H, "C, 9F, and 29Si, was completed, verifying the

correct disubstituted product was synthesized. Figure 2-11 illustrates the "C NMR

spectrum of di-4-pentenyldi(2,2,2-trifluoroethoxy)silane. The presence of OCH2CF3

groups in the product is confirmed by the two quartets in the '3C NMR spectrum due to a

one bond C-F coupling ('J= 280 Hz) and P two bond C-F coupling (2J= 40 Hz).



OCH2CF3
2 4

1 3 5
OCH2CF3 4
6 7


Figure 2-11. 13C NMR spectrum of di-4-pentenyldi(trifluoroethoxy)silane (29).








Phenolic Nucleophiles

The diphenoxysilyl functionalized monomer (30) was prepared in a similar

fashion to the other monomers discussed; however, the substitution of the Si-Cl bonds did

not occur without the addition of a small amount of DMAP (4-dimethylaminopyridine).

It is likely that both the decreased nucleophilic strength of phenol and the increased steric

bulk play a significant factor in its inability to effectively substitute two Si-Cl bonds. The

use of DMAP and imidazole has been well established in the literature for assisting more

difficult substitution reactions by acting as a good leaving group.125'126

Di-4-pentenydichlorosilane (26) was added to an ethereal solution of two

equivalents of phenol and triethylamine per Si-Cl bond and a catalytic amount of DMAP.

After reflux overnight under Ar and standard workup procedures, it was observed that the

resulting diphenoxysubstituted monomer (30) exhibited an increased resistance to

hydrolysis. This is clearly illustrated by the fact that the monomer was purified by silica

gel column chromatography using hexanes/methylene chloride as the mobile phase (yield

53%). NMR analysis confirmed the correct product was isolated and purified. It is

important to note that none of the other three substituted dialkoxysilyl monomers

prepared could be exposed to the atmosphere, or especially an acidic silica gel column

(pH 5), for periods of time without significant degradation or hydrolysis to a siloxane.


Polymerization of the Substituted Monomers

After all four substituted monomers were synthesized, purified, and analyzed, the

monomers were polymerized by standard ADMET methodology using the 2"d generation

Grubbs' ruthenium metathesis catalyst (11) (Figure 2-12).








R OR

'Si cat] Si
-3 3 -C2H4 \ l 3 ')"
OR OR

27-30 OR= OCH3 31
OCH2CH3 32
OCH2CF3 33
OC6H5 34


Figure 2-12. ADMET polymerization of the functionalized monomers.


Polymerization of the 2.2,2-Trifluoroethoxvsilvl Substituted Diene

Di-4-pentenyldi(2,2,2-trifluoroethoxy)silane (29) was chosen as an ADMET

monomer because trifluoroethoxy pendant groups have been shown in the literature to

resist swelling in petroleum fluids and to impart hydrophobicity to the resulting

polymer.9294 The best example of this lies in polyphosphazene systems, where a stable

polymer was prepared containing two trifluoroethoxy groups bonded to a phosphorous

atom per repeat unit. This functionalized polymer both repelled water and possessed a

low T, and high a Tm, which is characteristic of a microcrystalline thermoplastic

material.994 It was hoped that similar stability and physical properties would result from

a polycarbosilane analog.

The polymerization of di-4-pentenyldi(trifluoroethoxy)silane (29) was first

attempted using Schrock's molybdenum-based metathesis catalyst (6). This catalyst is

highly active and well known for its utility in ADMET polymerization using low

loadings. Under ADMET conditions, Si-Cl bonds were found to be inert to this

metathesis catalyst.78'79 The monomer was carefully dried and degassed before

combining with the [Mo] catalyst (6) in an Ar atmosphere glovebox in a 250:1








monomer:catalyst ratio (0.4% catalyst). No bubbling of ethylene was evident after the

catalyst was mixed with the monomer, which is highly atypical. Usually, instantaneous

reaction occurs when exposing a diene to Schrock's catalyst. The reaction tube was then

sealed, removed from the glovebox, and placed on a high vacuum line. Under high

vacuum (1 x 103 Torr), the evolution of ethylene was very slow and after 3 days, the

reaction mixture had become viscous and gel-like. The polymerization procedure was

terminated, and the solubility of the resulting product was tested. The polymer was

determined to be insoluble in all solvents tested, including CHCI3, THF, benzene, and

boiling chlorobenzene. It is believed that some interaction between the trifluorosilyl

groups and the catalyst produces a crosslinked gel.

Since a crosslinked polymer was obtained when using Schrock's catalyst (6),

polymerization using Grubbs' 2nd generation ruthenium catalyst (11) was attempted under

identical conditions. The metathesis catalyst was added to a flask containing the

monomer inside of an Ar atmosphere glovebox. The monomer to catalyst ratio used was

400:1 (0.25% catalyst), typical for ADMET polymerizations involving ruthenium

catalysts. No bubbling was evident, which is normal using this catalyst system, as it is

slower and usually has an induction period of several minutes. The reaction flask was

sealed, removed from the glovebox, and placed under reduced pressure on a high vacuum

line. As the temperature was raised slowly to 45 C using an oil bath, bubbling (ethylene

evolution) became evident. As the reaction mixture was stirred overnight at constant

temperature and at 1 x 103 Torr, ethylene slowly evolved, concurrent with an increase in

solution viscosity. After 24 hours, the reaction mixture became too viscous to stir and the

oil bath temperature was slowly raised to 60 C. Stirring and ethylene evolution








continued for another six hours, after which both ceased. At this point, the

polymerization procedure was stopped and a sample was removed for NMR analysis.

The resulting viscous polymer (33) was soluble in chloroform. Subsequent NMR

analysis indicated that ADMET polymerization occurred with no detectable side

reactions. Nevertheless, it was observed that rapid crosslinking occurs with this polymer

even when stored under an Ar atmosphere. This infers that even trace quantities of

moisture, from the initial aliquot being removed, are able to hydrolyze the Si-OCH2CF3

bonds. It appears that the presence of the trifluoro group weakens the silicon-alkoxy

bond towards hydrolysis instead of promoting increased stability via a hydrophobic

effect.

Because this polymer crosslinks rapidly, molecular weight analysis using gel

permeation chromatography (GPC) was unable to be performed. The level of moisture

present in GPC solvents such as THF and CHCl3 is sufficient to form an insoluble

network. Instead, end-group analysis of the 'H NMR spectrum was used to calculate the

number average molecular weight (Mn) of the polymer (Figure 2-13).

Since the end groups of ADMET polymers are terminal olefins, integration of

these resonances versus internal olefin resonances produced by metathesis enables the

average degree of polymerization to be calculated. Multiplication of the average degree

of polymerization by the molecular weight of the polymer repeat unit calculates the Mn.

Using this method, the M. was determined to be 12,000 g/mol. Integration of the

quantitative 13C NMR spectrum resonances indicated a trans olefin content of

approximately 80%, which is typically observed in ADMET polymerizations using well-

defined ruthenium or molybdenum-based metathesis catalysts.17'18'31









1
OCH2CF3

SiO2 4
I s! -;5 -
OCH2CF3 36










5


4 3


Figure 2-13. The 'H NMR spectrum of the di(trifluoroethoxy) functionalized
polycarbosilane (33).


Differential Scanning Calorimetry (DSC) was used to thermally analyze this

polymer. The DSC analysis of polymer 33, which was most likely partially crosslinked,

revealed a glass transition temperature (Tg) of-63 C and no discernable melting

temperature (Tm) when scanning from -95 C to 70 C These thermal analysis results

are typically observed for amorphous elastomers.


Polymerization of Di-4-pentenvldiethoxvsilane

The 2nd generation Grubbs' ruthenium catalyst (11) was used to polymerize di-4-

pentenyldiethoxysilane (28). Monomer and catalyst were mixed together in a flask under

an Ar atmosphere using a 200:1 monomer:catalyst ratio (0.5% catalyst). It is important to

note that once again, evolution of ethylene did not occur until the reaction mixture was


PPA








placed under high vacuum (1 x 103 Torr) and the external oil bath temperature had

reached at least 35 C. This is in agreement with results observed by others when using

this particular catalyst. Over the next 3 days, the temperature of the oil bath was slowly

raised to 65 OC to facilitate stirring. Throughout this time period, slow evolution

(bubbling) of ethylene was observed concurrently with an increase in the solution

viscosity of the mixture. After stirring ceased, the polymerization procedure was

terminated and the resulting polymer (32) was taken into an Ar atmosphere glovebox,

where all further manipulations were performed. This included preparation of the sample

for DSC analysis, thus minimizing the possibility of unintentional crosslinking. Polymer

32 was soluble in CHCI3, and subsequent 'H NMR analysis revealed ADMET

polycondensation chemistry had occurred. The molecular weight characterization via

end group analysis revealed a Mn= 15,000 g/mol (Figure 2-14).


1 2
OCH2CH3

I Si 3 4H 5
6
OCHCH3 6







6


5
4


Figure 2-14. 'H NMR spectrum of functionalized diethoxypolycarbosilane (32).






47

Carbon-13 NMR analysis indicated a typical cis/trans olefin ratio of the product polymer.

Molecular weight analysis was attempted using GPC, but the polymer crosslinked during

sample preparation in the CHCI3 solvent. Most likely the moisture levels present in

standard chromatographic solvents are sufficient to promote hydrolysis of the ethoxysilyl

group at a rate that prevents analysis by GPC.

Thermal analysis by DSC revealed a very low Tg (-83 OC) for this polymer and no

observed Tm when scanning a temperature range from -95 C to 70 C. Multiple DSC

scans of this polymer sample and quenching with liquid N2 were utilized to eliminate any

previous thermal history and for repetitive data verification. The temperature of analysis

was purposely limited to 70 C in order to avoid any thermal crosslinking of the polymer.


Polymerization of Di-4-pentenvldimethoxvsilane

Di-4-pentenyldimethoxysilane (27) was polymerized in the same manner (1 x 103

Torr) as the diethoxy functionalized monomer, using the 2nd generation Grubbs' catalyst

(11) in a monomer to catalyst ratio of 200:1 (0.5% catalyst). Over 5 days, the oil bath

temperature was slowly increased to 60 C during the course of the polymerization. After

this time period, the polymerization procedure was terminated and the product was taken

into an Ar atmosphere glovebox. It is well known that methoxy substituted silanes

hydrolyze rather quickly upon exposure to atmospheric moisture and this is accelerated

by both acidic and basic media, making handling difficult due to the presence of reactive

groups in each repeat unit of the polymer. Consequently, this methoxy functionalized

polycarbosilane (31) had to be kept under an inert atmosphere to prevent hydrolysis and

gelation. Like its diethoxysubstituted analog, polymer 31 was also too hydrolytically

sensitive for GPC analysis to be performed. However, the polymer was soluble in CHCI3








and 'H NMR end-group analysis indicated a Mn of 11,000 g/mol. Typical ADMET

condensation chemistry occurred, no side products or reactions were detected, and typical

stereoselectivity was observed. Preparing the sample pan and performing the analysis

under an inert nitrogen atmosphere enabled DSC analysis to be performed. The sample

was cold-quenched prior to scanning, in a similar fashion to the other dialkoxy polymers.

The scan rate and temperature range were also identical to that of the ethoxy-

functionalized polymer. The Tg of this methoxy-functionalized polymer (31) was -80 C,

and no Tm was observed.


Polymerization of Di-4-pentenvldiphenoxvsilane

The 2"d generation Grubbs' catalyst (11) was also used for the ADMET

polymerization of di-4-pentenyldiphenoxysilane in an analogous fashion to the previous

examples, using a monomer to catalyst ratio of 200:1 (0.5% catalyst). During this

reaction, ADMET polymerization occurred, producing a viscous, oily polymer (34). No

side reactions were detected via NMR analyses, and the trans olefin content was

determined to be approximately 80%. Since it was discovered during the synthesis and

purification of the monomer that this molecule is resistant to hydrolysis, the analysis of

polymer 34 was not performed under the same rigorous inert conditions that the previous

three polymers required. The resulting unsaturated polymer was readily soluble in

chloroform, and GPC analysis was performed in this solvent using dual UV/RI detectors.

The molecular weight analysis indicated a Mn of 18,000 g/mol and a polydispersity index

(Mw/Mn) of 1.8, typical for step-growth ADMET polymers. The DSC sample was not

prepared in an inert atmosphere; however, the sample was treated identically to that of

the other samples in terms of temperatures scanned (-95 C to 70 OC), the number of






49

scans, and the initial cold-quench of the sample pan. Figure 2-15 illustrates the DSC scan

of the phenoxy substituted polymer. This thermal analysis proved interesting; the glass

transition temperature of the diphenoxy-substituted polymer was -29 oC, an increase of

over 50 degrees compared to the other alkoxy-substituted analogs. No Tm was observed

for this polymer within the range of temperatures scanned.


20-

delta CP= 0.5 J/g C
Tg (onset)= -28.8 C
18-



o 16-

0

S14-
)tt


12- .. 0
-80 -60 -40 -20 0 20 40

temperature (C)

Figure 2-15. DSC scan of the phenoxy-substituted polymer (34).


Conclusions and Comparisons

The goal of this study was to examine the effects of varying the pendant group on

a common carbosilane backbone. Four derivative monomers were synthesized from the

same parent monomer containing two highly reactive bonds. The ancestral a,co-diene 26

is functionalized via substitution chemistry by reaction with excess nucleophile, in this

case, an alcohol (or alkoxide). Complete substitution at both sites is controlled by careful

product purification and isolation. All of these functionalized monomers were






50

polymerized using the same methodology, ADMET polycondensation, and four different

polymers with well-defined microstructures and similar molecular weights were

produced. Since the polymer backbone is stable, the atmospheric stability of these

materials is directly determined by the identity of the alkoxy pendant group. The pendant

group incorporated has also demonstrated an influence in the physical properties of the

resulting polymer, particularly the glass transition temperature. A variation of 54 degrees

is a result of simply changing the pendant group from ethoxy to phenoxy.

Interestingly, the glass transition temperatures of the substituted polycarbosilanes

produced by ADMET polymerization are comparable to their poly(phosphazene) (21)

and poly(silylenemethylene) analogs when the same pendant groups are used, in spite of

having vastly different polymer backbones as illustrated in Table 2-3.


Table 2-3. Comparison of three different polymers with identical pendant groups.

Pendant Group Tgs of poly- Tgs of poly- Tgs of polysilylene
carbosilanes via phosphazenes91,92 methylenes27
(OR) ADMET [NP(OR)2]n [Si(CH3)(OR)CH2]n

OMe -80 OC -76 OC a

OEt -83 C -84 OC -79 C

OCH2CF3 -63 oC -66 OC -51 C

OC6H5 -29 C -8 OC -18 C

a) not determined


The viability of substituting silicon-containing monomers prior to polymerization

was evaluated in this chapter. The ADMET polymerization of substituted monomers

using functionality tolerant metathesis catalysts has been shown to be an effective route






51

for producing functionalized polymers of this type. The advantage of this route is that all

of the functionality is guaranteed to be in place prior to polymerization. This assures that

quantitative functionalization of the backbone is a reality, since substituting a polymer

backbone can be much more difficult and often leads to incomplete functionalization.

The ADMET polymerization route possesses a significant advantage compared to

other systems, where prior monomer functionalization can lead to catalyst deactivation or

thermodynamically unfavorable polymerization conditions. Further, this methodology is

versatile and flexible, since one parent monomer can be substituted rather simply using

various nucleophiles, creating a large family of derivative polymers with vastly different

properties. The alternative route to polymer functionalization, macromolecular

substitution, where a reactive polymer backbone is derivitized, will be discussed in

Chapter 3.














CHAPTER 3
MACROMOLECULAR SUBSTITUTION USING ALKYLLITHIUM
REAGENTS


Introduction

Macromolecular substitution is a synthetic approach that involves the

functionalization of a pre-formed polymer backbone containing reactive groups with an

appropriate reagent. Using this strategy, variation of the side chains or pendant groups

can be accomplished. A related family of polymers could then be synthesized from the

same parent polymer backbone using this methodology. Since polymer properties can be

heavily influenced by the properties of its pendant group, a variety of different materials

with a wide range of properties and applications can be synthesized from only one parent

polymer.

Often, a significant drawback of this concept is the requirement that the

functionalization reaction be quantitative. This can be difficult and challenging to

accomplish, due to a variety of factors. Many chemical reactions are not quantitative and

produce undesirable by-products. This severely limits the choice of synthetic tools

available. In addition, the reactivity of functional groups on a polymer backbone is often

reduced compared to small molecules due to limited accessibility and steric hindrance. In

many instances, it is difficult for a reagent (nucleophile) to have enough encounters with

the polymer backbone in order to quantitatively substitute all of the reactive groups along

the backbone.








In spite of these limitations, there are several successful examples of macro-

molecular substitution in the literature involving polymers containing hydrolytically

sensitive phosphorus-chlorine and silicon-chlorine bonds. The Allcock group has

quantitatively substituted a poly(dichlorophosphazene) backbone containing two reactive

P-Cl bonds per repeat unit with a myriad of different nucleophiles, producing a family of

stable polymers with differing properties.92 A more detailed discussion of this polymer

system is found in Chapter 2.

Another pertinent series of polymers prepared using macromolecular substitution

are the poly(silylenemethylenes), studied by the Interrante laboratory.101,127-132 The

parent polymer, poly(dichlorosilylenemethylene) (36) was synthesized by the ring

opening polymerization of 1,3-dichloro-l,3-dimethyl-l,3-disilacyclobutane (35) using a

platinum-based catalyst (H2PtCI6).'28 A series of derivative polymers (37) were then

produced through the nucleophilic substitution of the Si-Cl bonds present each repeat

unit. Several different nucleophilic reagents were utilized, including alcohol/amine, an

effective system discussed previously in Chapter 2 (Figure 3-1). 01o,29.l30


CH3 CH3
HI ROHIE,3N
(CH3)(CI)Si Si(CI)(CH3) PC 16 or NaR Si --CH
I\ '/ or NaOR \ |I /
CI OR
35 36 37


Figure 3-1. Ring opening polymerization of a silylenemethylene, followed by
macromolecular substitution.


The West group has also successfully utilized macromolecular substitution in the

preparation of completely functionalized polysilanes (40).133'134 The parent reactive






54

polymer was synthesized via a ring opening polymerization of 1,1,2,2,3,3,4,4-octachloro-

cyclotetrasilane (38).135136 The resulting polysilane backbone (39) contained two Si-Cl

bonds per repeat unit. Substitution was enabled using a variety of different alcohol

nucleophiles in conjunction with an amine base as a proton acceptor.35"36 A series of

functionalized polysilanes (40) with interesting electronic and physical properties were

the result (Figure 3-2).


l CI
ClSi--Si --CI Cl OR
ii, ROH/amine J 1,
SSi Si
SCl C n I n
CI Cl Cl OR
38 39 40


Figure 3-2. Ring opening polymerization of a cyclotetrasilane, followed by
macromolecular substitution.


Polymerization of the Parent Monomer

To demonstrate the synthetic flexibility of the carbosilane backbone prepared via

ADMET polymerization, macromolecular substitution of the reactive parent

poly(dichlorocarbosilane) was performed. This polymer was synthesized by the ADMET

condensation polymerization of the a,to-diene monomer di-4-pentenyldichlorosilane

(26); the preparation of which was discussed in Chapter 2. Since it has been established

that Si-Cl bonds are inert to metathesis catalysts, two highly reactive bonds per repeat

unit of the polymer remain available for further reaction after polymerization is

complete.7778 The parent polymer itself is not useful due to its extreme hydrolytic

sensitivity, but the reactive functionalities present can be converted to more stable








moieties via simple nucleophilic substitution reactions. As a result, one basic polymer

can be converted into a whole family of derivative polymers with identical molecular

weight and molecular weight distribution.

Di-4-pentenyldichlorosilane (26) was polymerized using standard procedures for

ADMET polymerization. The monomer was carefully degassed using three freeze-

pump-thaw cycles and stored in an anhydrous environment before polymerization was

attempted. These steps were performed because Schrock's [Mo] metathesis catalyst (6)

was used to polymerize this monomer, and this early transition metal alkylidene is highly

sensitive to both moisture and oxygen. In an Ar atmosphere glovebox, di-4-

pentenyldichlorosilane was added to a reaction flask, followed by Schrock's [Mo]

catalyst (6) with a monomer to catalyst ratio of 500:1 (0.25% catalyst) (Figure 3-3).


CI CI

S[Mo] Si
3 -CH4 3 3
CI CI
26 41


Figure 3-3. ADMET polymerization of di-4-pentenyldichlorosilane (26) using Schrock's
[Mo] catalyst (6).


Upon mixing catalyst and monomer, bubbling (ethylene evolution) occurred instantly,

evidence of the high activity and high rate of reaction of this catalyst. The flask was

sealed, removed from the glovebox, and attached to a high vacuum line. The contents of

the reaction mixture were stirred at room temperature under intermittent vacuum, and

after a few hours, stirring had ceased due to increased solution viscosity. The reaction

flask was immersed in an oil bath, and the bath temperature was slowly increased to








40 C over the course of a day, enabling stirring to continue. As the temperature was

raised, full dynamic vacuum was applied. Slow bubbling of the contents continued,

resulting in the gradual formation of a thick, viscous oil. The polymerization procedure

was continued for another 48 hours, after which stirring ceased. The reaction flask was

not opened to the atmosphere in order to terminate the polymerization and destroy the

active catalytic species, which is common practice for ADMET polymerization using the

[Mo] catalyst (6), but instead was taken into a glovebox.

Solubility tests of the resulting polymer 41 were then performed using carefully

dried and degassed solvents. Only low molecular weight oligomers were soluble in polar

solvents such as chloroform, but complete solubility of the polymer was observed in

lower polarity hydrocarbon solvents such as benzene and toluene. Spectroscopic analysis

by NMR was utilized to confirm that the desired polymer had formed and no side

reactions were detected. Since polymer 41 is extremely moisture sensitive due to the

presence of a large number ofhydrolytically unstable groups along the backbone,

molecular weight determination by GPC analysis could not be performed. This polymer

crosslinks rapidly when exposed to the level of moisture present in chromatographic

solvents. End-group analysis by integration of the 'H NMR resonance signals was not a

viable method to determine the molecular weight either, as catalyst residue peaks were

present in the same region of the vinylic end groups at 5.0 ppm and 5.8 ppm, thus

preventing accurate integration of these resonances. However, after conversion of the Si-

Cl bonds to more stable moieties, molecular weight data can be obtained. Thermal

analysis using DSC of this parent polymer (41) was also performed. Careful

arrangements were made to avoid exposure to any moisture, and the sample was prepared






57

in an inert atmosphere. The sample was scanned from -90 C to 70 C, revealing a Tg at

-57 C and no discernable Tm, indicative of an amorphous elastomer. Considering the

differences between the phosphazene and carbosilane backbones, it is surprising that their

glass transition temperatures are quite similar. Poly(dichlorophosphazene) exhibits a Tg

of -66 OC, only a difference of 9 OC.92


Functionalization with Alkvllithium Reagents

After the solubility of the parent dichlorocarbosilane polymer had been

determined, it was functionalized with several different alkyl groups through nucleophilic

substitution reactions using alkyllithium reagents (Figure 3-4).


CI

I1 r
SiRM Si
S3 3 n 3 3 n
CI R
41 R= CH3 42

CH3(CH2)3 43

C6H5 44

Figure 3-4. Macromolecular substitution using alkyllithium reagents.


The rationale behind this approach is twofold. First, the parent polymer is soluble in non-

polar solvents, which are compatible and miscible with alkyllithium reagents. Second,

alkyllithium reagents are more nucleophilic than the alcohol systems discussed in

Chapter 2, and are well known for their ability to react with Si-Cl bonds with high

conversion.80 An additional advantage of utilizing carbanion nucleophiles lies in the

stability of the resulting bonds formed. Silicon-carbon bonds are similar in bond strength

to that of carbon-carbon bonds and most importantly, are stable to hydrolysis.82,83 After








functionalization of all of the silicon-chlorine bonds present, the polymer would be

stable, with the exception of slow oxidation of the unsaturated internal double bonds on

the polymer. This can be inhibited by integration of a small percentage of antioxidants

within the polymer matrix.81


Macromolecular Substitution Using Methyllithium

After the parent dichlorosilane polymer (41) was prepared by ADMET

polymerization videe supra), methyllithium was employed for the macromolecular

substitution reaction. The dichlorocarbosilane polymer (41) was dissolved in an excess

of freshly distilled, rigorously dried benzene in a carefully dried and argon-purged

Schlenk flask. The polymer solution was cooled to 0 C using an ice bath, and a fourfold

excess of 1.6M methyllithium in ether (2 equivalents per Si-Cl bond) was slowly added.

Under a continuous argon stream, the reaction mixture was allowed to warm slowly to

room temperature and stirred overnight. A cloudy white precipitate of LiCI was observed

in the polymer solution, indicating that substitution of the Si-Cl bonds had occurred.

After no additional precipitate formation was observed, the polymer solution was

cannulated into a fivefold excess of rigorously dried, cold methanol, precipitating out a

clear, viscous, oily material. This procedure also enabled the excess methyllithium to be

destroyed by the protic non-solvent, as well as dissolving all of the LiCI precipitate. The

product, poly(di-4-pentenyl)dimethylsilane (42) was then dried in vacuo overnight.

Polymer 42 was readily soluble in chloroform and no gel formation was observed.

When the polymer was exposed to the atmosphere, it remained soluble without gelation.

This evidence strongly suggests that quantitative substitution of the Si-Cl bonds occurred.

Any residual Si-CI bonds will react rapidly with atmospheric moisture, producing stable








Si-O-Si bonds and creating a crosslinked network polymer. An insoluble network will

form with only a few crosslinks present.8' Since this did not occur, all of the chlorine

atoms were replaced with methyl groups. This was corroborated by 'H, "C, and 9Si

NMR spectroscopic evidence. The 29Si NMR spectrum illustrated only one peak at 2.3

ppm, indicative of a silicon atom bonded to four alkyl groups, and the lack of a signal at

33 ppm from the parent dichlorocarbosilane polymer (41) (Figure 3-5).

CH3

SSi- -

CH3
TMS




35 30 25 A0 15 10 5 0 -5 PPR



Figure 3-5. The 29Si NMR spectrum of the methyl functionalized polycarbosilane (42).


The 'H NMR spectrum illustrated that no side-reactions, crosslinking, or alkylation of the

polymer backbone occurred as a consequence of using the alkyllithium reagent. This is

in contrast to results observed by Rushkin and Interrante in the poly(silylenemethylene)

system.26 Vinylic end groups were not observed in the 'H NMR spectrum, evidence of

the formation of a moderately high molecular weight polymer (Figure 3-6). The 3"C

NMR spectrum illustrated that the cis/trans olefin content (~ 80% trans) was not affected

and in agreement with other ADMET polymers synthesized.









1
CH3
/I 2 4-6,
i 1
l\ ;~ 3 74
CH3



6
2 4 5






Figure 3-6. 1H NMR analysis of the methyl functionalized carbosilane polymer (42).


Molecular weight determination for polymer 42 was easily accomplished using

GPC because it was hydrolytically stable. Using a refractive index (RI) detector, the

number average molecular weight (Mn) was determined to be 16,000 g/mol with a

(Mw/Mn) of 2.3, a range commonly observed for ADMET polymers. Thermal analysis by

DSC was also performed, indicating a Tg of-89 oC. This is the lowest Tg observed for

the whole series of polymers produced and illustrates a high degree of chain flexibility in

the polymer backbone combined with a small pendant group.


Synthesis of the Phenyl Substituted Polymer Using Phenyllithium

Using the parent dichlorosilane polymer (41), a polycarbosilane substituted with

two phenyl pendant groups per repeat unit was synthesized using analogous methodology

to that of the dimethyl substituted polymer. Di-4-pentenyldichlorosilane (26) was

polymerized via ADMET using Schrock's [Mo] catalyst (6) and dissolved in excess dry

benzene. Four equivalents ofphenyllithium (1.8M in cyclohexane-ether) per silicon








atom was added in a similar fashion to that discussed above, and after stirring overnight

under argon, two equivalents of methyllithium was added to the reaction mixture.

Methyllithium was added to the reaction mixture to assure that any residual Si-Cl bonds

remaining on the polymer backbone would be substituted, since MeLi is sterically

unencumbered and was shown to quantitatively substitute Si-Cl bonds. During the

reaction, precipitation of LiCI salts was observed in this reaction as well, indicating that

substitution chemistry was occurring. After the reaction was complete, the polymer was

cannulated into an excess of carefully dried, cold methanol, precipitated out of solution,

and dried in vacuo, giving an off-white, tacky, elastomeric solid.

Polymer 44 was readily soluble in chloroform and did not become insoluble upon

exposure to atmospheric moisture, indicating complete substitution of the Si-Cl bonds

occurred. Spectroscopy was used to monitor the percent of phenyl and methyl

substitution of the backbone. Since silicon-methyl groups resonate at frequencies very

close to the common NMR reference standard tetramethylsilane (TMS), deuterated

solvents containing no TMS had to be employed for both 'H and 13C NMR analysis to

ensure accurate quantification of the substitution. For this macromolecular substitution

using sequential addition ofphenyllithium and methyllithium, no Si-Me peaks were

observed in either the 'H or 13C NMR spectra. Therefore, the substitution with

phenyllithium was determined to be quantitative. The 29Si NMR spectrum illustrated

only one peak at -6.6 ppm, confirming a R2SiPh2 backbone structure.

The 3C NMR of the phenyl substituted derivative polymer (44) is illustrated in

Figure 3-7, showing typical cis/trans olefin stereochemistry and providing evidence for a

well-defined polymer microstructure.










1-4






55 6 8
Si






1-4



6 7
8



2. ..50 .0 50 .0 PP



Figure 3-7. 3C NMR spectrum of the phenyl substituted polycarbosilane (44).


This spectrum shows an additional peak at 32 ppm, which was suspected to be the

adjacent allylic carbon present as a cis olefin linkage on the polymer backbone. In order

to prove the type of carbon that was present, both APT (Attached Proton Test) and DEPT

(Distortionless Enhancement by Polarization Transfer) NMR experiments were

performed. Both experiments concluded that the peak in question was from a CH2

carbon. The APT test separates the types of carbons present in a structure by showing

CH and CH3 peaks pointing downward and CH2 peaks pointing upward. As seen in

Figure 3-8, the APT experiment shows the peak at 32 ppm pointing upward.

The DEPT analysis separates all the carbons on the basis of the number of protons

attached, giving four separate spectra, one for each type of carbon: quaternary, CH, CH2,

and CH3. The peak at 32 ppm shows up as a CH2 peak in the DEPT analysis as well. In

addition, the fact that there is a resonance at 130 ppm, attributed to the cis internal olefin









upfield from the trans internal olefin resonance, which is also pointed downward,

confirms the supposition that the peak at 32 ppm is a result of roughly 20% cis olefin

present in the resulting polymer backbone and is not due to main chain alkylation or

proton abstraction by the alkyllithium reagent.


2-4

1 6-8




7

1. 8




140 120 100 80 60 40 20 0 ppa


2-5








Figure 3-8. APT NMR spectrum ofpoly(di-4-pentenyl)diphenylsilane (44).


Since polymer 44 was readily soluble in chloroform and was not prone to

crosslinking, molecular weight determination was conducted using GPC analysis. The

molecular weight was found to be 22,000 g/mol with a polydispersity index (Mw/Mn) of

2.3 using tandem RI and UV detectors (X=254 nm). The DSC analysis illustrated a much

higher Tg than that observed for all the other functionalized polymers, most likely due to







64

rigid, more bulky phenyl substituents directly bonded to the silicon atom. The polymer

possessed a Tg= -2 C and no Tm, demonstrated in Figure 3-9.



16.0-

delta Cp= 0.4 J/gC
S15.5- T, (onset)= -1.9 oC
E


S15.0-
C

0=
14.5-


14.0
-60 -40 -20 0 20 40
temperature ( C)


Figure 3-9. DSC scan of the phenyl substituted polycarbosilane (44).


Using Butvllithium as a Nucleophile

Macromolecular substitution chemistry was employed to functionalize the

dichlorosilane polymer (41) using n-butyllithium. Poly(di-4-pentenyl)dichlorosilane (41)

was once again synthesized using Schrock's [Mo] catalyst (6) using the method discussed

previously. Polymer 41 was dissolved in freshly distilled, dry benzene and the solution

was chilled to 0 C with an ice bath. Under argon with vigorous stirring, two equivalents

of 1.6M n-BuLi in hexanes per Si-Cl bond was slowly added, and the precipitation of

LiCI salts began immediately. The reaction mixture was allowed to slowly warm to room

temperature and stirred overnight. Excess methyllithium was then added in order to react

with any residual Si-Cl bonds present and stirred overnight. After this time period, the









polymer was precipitated by cannulation of the reaction mixture into a large excess of

ice-cold, freshly distilled, dry methanol. The solvent was removed by filter cannulation,

and the polymer was dried at room temperature under vacuum for 24 hours. The

resulting polymer (43) was a clear, viscous oil which was readily soluble in chloroform.

Using NMR analysis, it was discovered that butyl group incorporation was not

quantitative in this case. The presence of methyl groups on the polymer was observed in

both the 'H and '3C NMR spectra. Additional resonance peaks at -0.10 ppm in 'H NMR

and at -5.1 ppm in the 3"C NMR are indicative of silicon-methyl groups. Figure 3-10

shows the 1H NMR spectrum possessing a Si-CH3 peak.










CHs3
1 5


3 4







S 6 5 3 3 1 0 PP,


Figure 3-10. 1H NMR spectrum of the butyl-methyl functionalized polycarbosilane (43).


After it was confirmed that methyl groups were attached to the polymer

backbone, quantification of the methyl incorporation was determined. Quantitative 13C






66

NMR integration, performed by using a pulse sequence that allows complete 3C nuclear

relaxation, determined the amount of methyl content in the polymer to be 18%. In

addition to the amount of methyl groups present, the substitution pattern at each silicon

atom of the polymer needed to be determined. There are three possibilities for the

nucleophilic substitution: butyl-Si-butyl, butyl-Si-methyl, and methyl-Si-methyl. The

percentage of methyl-Si-methyl formation was assumed to be trivial since the overall

percent of methyl group substitution was not large. The 29Si NMR spectra for this

polymer illustrated two peaks at 3.0 and 3.1 ppm, illustrated in Figure 3-11.











TMS




$0 5 0 pol10 PPA


Figure 3-11. 29Si NMR analysis of the resulting polymer (43) from the butyl and
methyllithium substitution.


Since there are only two types of silicon environments present, one of the three

possibilities outlined above did not occur. In order to distinguish between the methyl-Si-

methyl and butyl-Si-methyl substitution pattern possibilities for the second peak, this

peak was compared to the resonances observed for the fully methyl substituted polymer.

The completely methyl functionalized polymer contains a peak at 2.3 ppm in the 29Si








NMR, whereas both the peaks present for the butyl polymer were around 3.0 ppm. In

addition, the methyl-Si-methyl possibility should also be ruled out because of the large

percentage of butyl groups that were on the polymer backbone. Further, it would appear

logical that excess n-BuLi would easily be able to substitute at sites containing two Si-Cl

bonds. Therefore, it is believed that the NMR evidence, along with the order of reagent

addition and the low amount of methyl substitution observed, rules out the methyl-Si-

methyl possibility for the substitution pattern observed.

The butyl functionalized polycarbosilane (43) was also stable in the atmosphere, allowing

for molecular weight determination by GPC analysis. Using this technique, the polymer

Mn was determined to be 12,000 g/mol with a Mw/Mn of 1.9. The polydispersity index

observed for this sample is characteristic of polymers produced by step-growth

polymerization. Although the molecular weights of the series of polymers produced by

macromolecular substitution are not identical, it is not a function of pendant group

variation. Each macromolecular substitution reaction was performed on a different batch

of parent dichlorocarbosilane polymer (41). The molecular weight discrepancies

observed are attributed to variations of experimental conditions encountered during each

polymerization of parent polymer, particularly the difficulties of stirring the polymer

solution as its viscosity increases. Thermal analysis of this amorphous elastomer (43)

illustrated a low Tg of -76 oC with no observed Tm, when scanning the sample from -95

C to 70 C (Figure 3-12). The low T, is attributed to both the flexibility of the butyl side

groups and the resulting irregular polymer structure, which is due to the presence of a

small amount of methyl groups placed randomly along the polymer backbone.









15.5 delta Cp= 0.4 J/gC
T. (onset)= -75.6 C

15.0





140 -

S13 5-



-80 -70 060 -50 40 -30 -20 -10

temperature (C)


Figure 3-12. DSC trace of the butyl-methyl functionalized polymer (43).


Substitution Reactions of Comparative Polymer Systems

Comparisons of ADMET unsaturated polycarbosilanes with Interrante's

substituted poly(silylenemethylenes) (37) are interesting. The glass transition

temperature of ADMET polymer 43 is -76 C, comparable to the poly(silylene-

methylene) containing both butyl and methyl pendant groups prepared by ring-opening

polymerization. In spite of significant differences between the polymer backbones and

method of preparation, the Tg of the latter polymer is -63 OC, only a 13 C difference.'30

At the beginning of this chapter, it was discussed that macromolecular

substitution was quantitative on the silylenemethylene polymers when utilizing an

alcohol/amine nucleophilic system, which reacted readily with the one Si-Cl present in

each repeat unit.10t129 However, when substituting with alkyl groups (carbanions) on this

backbone, a quantitative reaction did not occur. A methylene group flanked by two

silicon atoms in the a-position is more acidic and shows a greater propensity for proton








abstraction by alkyllithiums or other strong bases.137"140 Proton abstraction on the

bridgehead carbon was indeed observed by Rushkin and Interrante using alkyllithium

reagents, producing an insoluble, crosslinked polymer.'t1 This crosslinking reaction was

attributed to carbanion formation on the polymer backbone, which then reacts further

with Si-Cl bonds present in other polymer chains (Figure 3-13).101 This reaction forms

new stable Si-C bonds between two or more polymer chains, producing an irreversible

crosslink site.



S-C RLi i i--CH croslnked
n O |H2 -( |n product
CH3 CH3 CH3
-(-Si--CH2 -

HCH3

CH3

Figure 3-13. Proton abstraction of the bridgehead carbon, producing a crosslinked
polymer.


Consequently, the strongest nucleophiles able to used in the poly(silylenemethylene)

system were Grignard reagents. These are several orders of magnitude weaker in

nucleophilic strength than alkyllithium reagents.141 Butylmagnesium chloride was

reacted with poly(chloromethylsilylenemethylene) and incomplete substitution occurred.

Only about 70% of the sites available were substituted with butyl groups.'0' Earlier

reports in the literature show that when using Grignard reagents, substitution of a fourth

group to a sterically hindered chlorotrialkylsilane is difficult.140








Conclusions

Three functionalized polycarbosilanes were produced from the same parent

dichlorocarbosilane polymer, synthesized by metathesis polycondensation. All of these

alkylated polymers were tacky, elastic materials that were stable under atmospheric

conditions (Table 3-1). This stability is a result of the complete replacement of

hydrolytically sensitive Si-Cl bonds with stable Si-C bonds during the macromolecular

substitution reaction. Phenyl and methyl group disubstitution of the polymers proved to

be quantitative, whereas butyl group substitution was only 82%, and exhaustive

substitution completed using methyllithium. The inability of n-BuLi to quantitatively

substitute all of the Si-Cl bonds present along the polymer backbone may be the result of

aggregation of the nucleophile (n-BuLi) with benzene.


Table 3-1. Molecular weight and thermal data for the three polymers synthesized via
ADMET-macromolecular substitution.

Pendant Group Mn (g/mol) Mw/Mn Tg (C)

CH3 16000 2.3 -89

CH3(CH2)3 / CH3 12000 1.9 -76

C6H5 22000 2.3 -2



In contrast with polymers containing alternating [-Si-CH2-] backbones, proton

abstraction on the main chain was not observed during the substitution reactions

performed on this series of polymers. Since these polymers are not susceptible to

crosslinking when exposed to alkyllithium reagents as nucleophiles, a great deal of

flexibility in the choice of macromolecular substitution reagents is permitted. This opens






71

the potential to synthesize a whole series of polymers using either alkyllithium or

Grignard reagents, leading to a wide range of stable polymers with the incorporation of

various functionalities.














CHAPTER 4
HYDROLYTICALLY RESISTANT SILYL ETHERS


Introduction

Nucleophilic substitution has long been exploited as a flexible route of modifying

the properties of reactive silanes. Highly reactive silicon-halide bonds are readily

attacked by a variety of nucleophiles, producing compounds that can be utilized as

protecting groups and reagents in organic synthesis8286 or as pendant groups on inorganic

polymers. 6566 Silyl ethers, molecules containing a reactive silicon-alkoxy bond, are

susceptible to hydrolysis, a reaction that is kinetically dependent upon the nature of the

substituents adjacent to this bond, chemistry which is quite old and useful.85 If hydrolysis

is not desired in the product, then an alternative route of functionalization of the reactive

silicon-chlorine bonds is through alkylation reactions. Formation of a silicon-carbon

bond via Grignard or alkyllithium substitution chemistry has been shown to be quite

straightforward due to both the high reactivity of both reagent and substrate, and the

thermodynamic stability of the resulting product.103

The substitution reaction of Si-Cl bonds has been determined to proceed in a SN2-

like fashion by mechanistic studies.142144 Therefore, the limitations of SN2 reactions still

apply, particularly sterics, to the substitution of chlorosilanes. In the case of dialkyl-

dichlorosilanes, the substitution of the first Si-Cl bond is quite facile. The second

Si-Cl bond is more difficult to substitute, and the rate of reaction is considerably slower.






73

Previous studies show that small, linear alcohol nucleophiles react with both Si-Cl bonds

adequately. On the other hand, as the steric bulk of the nucleophile is increased,

substitution at the second Si-CI bond becomes considerably more difficult and is often

completely inhibited. If two chlorine atoms can be replaced with bulky substituents,

significant improvements to hydrolytic stability can be achieved, especially in the case of

dialkoxysilanes. An in-depth study of the nucleophilic substitution reaction on

dialkyldichlorosilanes using sterically unencumbered nucleophiles has been discussed

elsewhere.79,80

Different methods of synthesizing substituted carbosilane polymers containing

various pendant groups via ADMET polymerization have been discussed in Chapters 2

and 3. Several of the polymers discussed in Chapter 2 were shown to be hydrolytically

unstable, with the chemistry and rate of hydrolysis determined by the identity of the

pendant group. Slowing or preventing the hydrolysis of the pendant groups of these

polymers would be advantageous, allowing for characterization or processing before any

crosslinking begins, as well as providing easier handling. This chapter will focus on the

synthesis of compounds that both model the carbosilane polymer repeat unit and are

hydrolytically stable or undergo slow hydrolysis. Since the carbon-silicon backbone is of

a set architecture, this goal can be accomplished by either increasing the steric bulk of the

nucleophiles or by producing cyclic structures using chelating nucleophiles. Both routes

were explored in this work. A logical extension of these studies would be to produce

ADMET polymers of increased stability using the knowledge acquired from these

approaches.








Di-n-hexyldichlorosilane was chosen as a model substrate to study the

nucleophilic substitution reaction on two Si-Cl bonds that would produce more

hydrolytically stable products. This silane contains two linear alkyl chains and two

chlorine groups bound to a central silicon atom. It serves as a plausible model because of

the similarities to the repeat unit of the polymers synthesized in Chapters 2 and 3. The

environment around the silicon atom is virtually identical to the analogous ADMET

capable a,co-diene, which should allow the results of this model study to be easily applied

to polymers.


Sterically Bulky Aliphatic Nucleophiles

Increasing both the steric bulk of the substrate (groups bound to the silicon atom)

as well as the attacking nucleophiles decreases the hydrolysis or degradation of the

resulting silyl ether.140,145 To test this hypothesis in the case of dialkyldichlorosilanes,

isopropyl alcohol and t-butyl alcohol were evaluated as nucleophiles by reacting them

with di-n-hexyldichlorosilane (45), using Et3N as the proton acceptor (Figure 4-1).



Y
Cl 0


CI

45 .
46

Figure 4-1. Reacting di-n-hexyldichlorosilane (45) with isopropanol.


The reaction of isopropyl alcohol successfully leads to a disubstituted silane (46)

in 32% yield, which could be worked up in the atmosphere, and exposed to solvents that








were not anhydrous without significant hydrolysis. Elemental analysis and NMR

confirmed the identity of di-n-hexyldiisopropoxysilane (46). The 13C NMR spectrum is

shown below (Figure 4-2). Tertiary-butyl alcohol was reacted using similar conditions to

that of the isopropyl alcohol, but only the monosubstituted product was produced. It is

believed that the steric bulk of the t-butyl group is too great to effect the second

substitution.



r2
0
3I
"4 i 4 1,4-7
1 4-7




2 3,8




'a 71 Ii 50 40 31 21 11 ppI


Figure 4-2. The '3C NMR spectrum of di-n-hexyldiisopropylsilane (46).


Lithium Reagents as Nucleophiles

Since alcohols are comparatively weak nucleophiles, significantly stronger bulky

alkyllithium reagents were studied. The rationale for using bulky alkyllithium reagents

as nucleophiles was to determine if the inability of t-butyl alcohol to substitute both Si-Cl

bonds can be overcome with a stronger nucleophile of similar bulk. Additionally, any

successful reaction would produce a product containing new Si-C bonds and be








completely stable to the atmosphere.85 However, steric hindrance still dominates the

results of this reaction, as both t-butyllithium and (trimethylsilylmethyl)lithium were not

able to produce the desired disubstituted product. In addition, the reaction of t-butyl-

lithium created a product that was lithiated at the ct carbon to the silicon atom. This

lithiation reaction is a result of both steric hindrance and the presence of chlorine atoms

bound to the silicon.146-149 Chlorine atoms increase the acidity of the methylene groups

adjacent to the silicon atom, making them more susceptible to lithiation. There are

competing reactions occurring when the lithium reagent is added to the chlorosilane:

metalation versus coupling (Figure 4-3).147 Previous work shows that if the molecule is

sterically bulky, making the substitution difficult, then lithiation predominates. Coupling

(substitution) will occur at the less hindered sites. The reaction of t-butyllithium

produces a mixture of the lithiated and the substituted product, which is attributed to the

differences in the environment around the silicon atom before and after the first Si-Cl

bond is reacted.


Me3SiC(CH3)3
Me3SiCI + t-BuLi (coupling)

LiCH2SiMe2CI Me3SiCI Me3SiCH2SiMe2CI
(metalation)

Figure 4-3. Competition between metalation and coupling.147


It is likely that the first Si-CI bond is substituted in this reaction, and at the second Si-Cl

bond, a mixture of substitution and metalation occurs. The steric bulk of both the

trisubstituted silane substrate and the nucleophilic agent allow for lithiation to occur

instead of quantitative nucleophilic substitution of the Si-Cl bond.






77

In contrast, a-metalation has not been observed when reacting n-butyllithium or

phenyllithium with di-n-hexyldichlorosilane.150 These reagents act as nucleophiles and

not as bases, and no a-lithiation of the C-H bond adjacent to silicon is observed.

Disubstituted product is formed in 79% yield when reacting di-n-hexyldichlorosilane (45)

with n-butyllithium, and in 68% yield using phenyllithium. Sec-butyllithium was not

studied since it has been shown in the literature to pose an a-lithiation hazard similar to t-

butyllithium.1t5


Nucleophiles Containing a Phenvl Group

Phenyl group containing nucleophiles were studied to in order to produce

materials that exhibit an increased resistance to hydrolysis. Although not considered to

be extremely bulky groups, phenoxysilanes have been shown in the literature to be more

hydrolytically stable.10' Initially, attempts to substitute both Si-Cl bonds in di-n-hexyl-

dichlorosilane (45) with the nucleophilic systems phenol/Et3N or sodium phenoxide

failed. The addition of a small amount ofDMAP (4-dimethylaminopyridine) as a

substitution promoter enabled a diphenoxydialkylsilane to form, which was discussed

earlier in Chapter 2. The phenoxy group significantly increased the hydrolytic stability of

resulting polymer.

Benzyl alcohol would be an interesting probe to evaluate the effect of a methylene

group between the phenyl group and the oxygen atom. Benzyl alcohol is not a phenol,

and it would be interesting to examine if the resulting silyl ether linkage is susceptible to

hydrolysis. Upon reaction with di-n-hexyldichlorosilane (45) in the presence of Et3N,

disubstitution occurred readily in a yield of 62%. This reaction does not require the








addition of DMAP like phenol, yet the resulting product was rather stable to hydrolysis,

as it was purified on a silica gel column in the atmosphere without substantial

decomposition.


Diols as "Chelating" Nucleophiles

After examining the possibility of increased hydrolytic resistance using bulky

monoalcohol reagents, the effect dialcohols have on the substitution reaction and the

stability of their resulting products was studied. Diols would be interesting to study

because one molecule would react with both Si-Cl bonds present, forming a cyclic

alkoxysilane structure under the right conditions (Figure 4-4). Depending on the size of

the ring formed and the steric environment around the silicon atom, hydrolysis may be

slowed considerably. This approach is used in the formation of silane protecting

groups.82,84,152 Polymers with such pendant groups might have interesting properties and

improved resistance to hydrolysis.



I R R
"+ R + Et3N / DMAP __ o
I OH OH \/
4 4

Figure 4-4. Using chelating nucleophiles to produce more stable cyclic silicon containing
structures.


Several diols, including diphenylsilanediol, were employed as nucleophiles and

reacted with di-n-hexyldichlorosilane (45) under conditions identical to that used for

linear alcohols. Although silicon and carbon are in the same group on the periodic table,

the chemistry of the two elements is quite different, and the effect on the acidity of








hydroxyl groups is no exception. Silanol protons are considerably more acidic than

alcohol carbon analogs.82 This increased acidity reduces the nucleophilic strength of its

conjugate base, meaning that a silanol is a weaker nucleophile than an alcohol. In spite

of this, the resulting siloxane (Si-O-Si) bond of a Si-Cl substitution reaction is thermo-

dynamically favored and highly stable, which may allow disubstitution to occur.

Unfortunately, the reaction of diphenylsilanediol with di-n-hexyldichlorosilane (45) did

not produce a cyclic monomer and instead produced a low molecular weight siloxane

oligomer with a molecular weight of approximately 700 g/mol. This occurred

presumably because the cyclic species that would have formed from an intramolecular

reaction would have been a highly strained four-membered ring. Therefore, after the first

Si-Cl bond is substituted, the other Si-OH bond of diphenylsilanediol undergoes an

intermolecular reaction with another molecule of di-n-hexyldichlorosilane (Figure 4-5).






Cl OH Si
S/ C" /0 O /Cl
5 5 C /CI Si
Si 5 5 5 5



Figure 4-5. Intermolecular reaction between the monosubstituted product with another
di-n-hexyldichlorosilane.


Aromatic based diols such as catechol (1,2-dihydroxybenzene) were also

evaluated in this substitution study. If the desired product di-n-hexyl(1,2-

phenylenedioxy)silane could be produced, it should have excellent stability to hydrolysis,

and any polymer derivatives synthesized bearing these substituent groups should have






80

interesting physical properties. Nevertheless, disubstitution did not occur when using this

nucleophile. The reason behind this failure may be unfavorable ring strain. A five-

membered ring would have been the result in this reaction, but this cyclic structure was

most likely too strained to form. The substitution chemistry of both dimethyl-

dichlorosilane and diphenyldichlorosilane has been evaluated with catechol.153'154 The

requirement for any 1,3-dioxa-2-sila-substituted five-membered ring is that the average

bond angles are <108".'53 This requirement cannot be met unless some of the bond

angles in the cyclic structure are compressed below that of a tetrahedral bond angle of

109.50. Since the bond angles between the aromatic carbon and hydroxyl group on

catechol are typical of bond angles in aromatic compounds and are approximately 120,

this ring compression is magnified between the other bond angles in the ring. Therefore,

the increased ring strain prevents the reaction from occurring. It has been shown in other

systems that the addition of methylene spacers between the hydroxyl group and the

aromatic ring is required for the disubstituted cyclic monomers to form.154,155

Since the aromatic diol, catechol, did not produce a cyclic monomer, aliphatic

cyclic diols were evaluated as nucleophiles instead. The bond angles around the

hydroxyl groups would not be as large and would be easier to compress, resulting in a

less strained system. Two commercially available cyclohexanediols were studied, trans-

1,2-cyclohexanediol and a mixture ofcis- and trans-1,3-cyclohexanediol. Trans-1,2-

cyclo-hexanediol was reacted with di-n-hexyldichlorosilane (45) in the presence of Et3N

and did not produce a disubstituted product. A crystalline material was isolated from the

reaction, which was believed to be a dimer or oligomer. A small amount of DMAP was

added in an attempt to promote substitution of both of the Si-CI bonds, but the results








were the same; dimer or oligomer formation had occurred. Five-membered cyclic

alkoxysilanes were not able to be synthesized by this reaction. After the first substitution

occurs, it is easier for the second -OH group present to react with another di-n-hexyl-

dichlorosilane molecule, producing a linear species, instead of wrapping around the same

molecule and forming a strained cyclic entity.

Since the formation of five-membered rings from the reaction of di-n-hexyl-

dichlorosilane (45) with geminal and vicinal diols resulted in dimer and oligomer

formation, diols containing an additional methylene spacer between the two hydroxy

groups were studied. These diols would form six-membered rings upon disubstitution,

which possess low ring strain and high stability. The first nucleophile chosen was 2,4-

dimethyl-2,4-pentanediol. The interesting characteristic of this nucleophile is the fact

that both hydroxyl groups are bonded to tertiary carbons, a similar steric environment to

t-butanol. These tertiary alcohol sites should increase the hydrolytic stability of the

resulting compound and sterically protect both alkoxysilane bonds. As noted earlier, t-

butanol was unable to effectively disubstitute di-n-hexyldichlorosilane. However, when

di-n-hexyldichlorosilane (45) was reacted with 2,4-dimethyl-2,4-pentanediol in the

presence of the proton acceptor Et3N and DMAP, both Si-Cl bonds were substituted by

the same diol, forming a stable six-membered ring (47) (Figure 4-6).




S^ ++ -1 + Et3N / DMAP---
I 4 OH OH Si _A-
Cf 4 4
45 47

Figure 4-6. Reaction of di-n-hexyldichlorosilane (45) with 2,4-dimethyl-2,4-pentanediol.








The resulting product was quite stable under atmospheric conditions and reaction

workup. After purification by distillation, the product was isolated in 41% yield, and

NMR analysis was used to confirm the structure of product 47. Figure 4-7 illustrates the

29Si NMR spectrum with one peak at -5.7 ppm, indicative of a R2SiOR2 environment at

the silicon atom.











TMS



20 10 0 10 -20 PPI

Figure 4-7. The 29Si NMR spectrum of di-n-hexyl(di-2,4-dimethyl-2,4-
pentanedioxy)silane (47).


A mixture of cis- and trans-l,3-cyclohexanediol was studied for several reasons.

First, the cyclohexyl 1,3-diol groups should be more flexible than the analogous 1,2-diol

groups, making it more likely to chelate around the two Si-Cl bonds and form a six-

membered ring. In addition, this diol mixture would determine if one isomer would be

conformationally preferred over the other, due to the different spatial orientation of the

cis isomer versus the trans isomer. Di-n-hexyldichlorosilane (45) was reacted with a

mixture of cis- and trans-1,3-cyclohexanediol under identical conditions as the 2,4-

dimethyl-2,4-pentanediol (Figure 4-8). A disubstituted cyclic product (48) was observed








which was resistant to atmospheric hydrolysis during workup. Both a viscous oil and a

small amount of solid material were the result. The viscous liquid was purified by

vacuum distillation and isolated in 45% yield, but the solid material could not be distilled.


Cl OH

i + + Et3N / DMAP -- -4 4
"4 | 4 0 ( \ 0
Cl OH H-L H

45 48


Figure 4-8. Reaction of cis and trans 1,3-cyclohexanediol with di-n-hexyldichlorosilane.


Product 48 was analyzed by NMR spectroscopy and several interesting

observations were made. Most importantly, the 3C NMR spectrum of the product

elucidates pertinent structural information of the product because only one peak at 68.5

ppm is observed at the frequency where an oxygen atom is bonded to a carbon atom

(Figure 4-9). In the starting material, there are two peaks from both the cis and trans

isomers present at 69.0 ppm and 66.6 ppm.156 Therefore, since the product is

symmetrical, it is believed that only one of the isomers reacts to form the product.

The product observed in this reaction is likely due to the reaction of the cis

isomer; the trans isomer is effectively prevented from forming a cyclic species. This is

caused by the spatial conformation of the -OH groups on the cyclohexane ring (Figure 4-

10). During the reaction, a cis or trans hydroxyl group has equal reactivity with one Si-

Cl bond. Once this bond is formed, the -OH group on the cis isomer is able to reach

around for an intramolecular attack and form a six-membered ring. The trans isomer








-OH group is spatially inaccessible and therefore must initiate an intermolecular attack

on another chlorosilane, forming an oligomeric species.




Od,,o


Figure 4-9.


i70 0 50 4 0 20 PP

The 'C NMR spectrum of the cis 1,3-cyclohexanediol chelated product (48).


HcHO O H

1.3-cis


OH HO

1,3-trans


Figure 4-10. Conformations of the cis- and trans-l,3-cyclohexanediols.


Hydrolvtic Stability of Silacvcloalkanes

The hydrolytic stability of diisopropyl and di-t-butyl substituted dioxosila-

cycloalkanes has been evaluated in the literature.152 It was found that for the diisopropyl

substituted silanes, five-membered ring species are easily cleaved during silica gel thin

layer chromatography (mildly acidic environment, pH = 5).'52 Even when using bulky


.......Y .. _.I ..-..L---- r------ _. _-._ .I_..._IL1. L-.-r- LL r__-l_ld.L-






85
substituents such as di-t-butyl groups, five-membered dioxosilacycloalkanes are rapidly

hydrolyzed under basic conditions. In contrast, the six-membered ring analogs are

hydrolytically stable for hours in environments with pH levels varying from 4-10.152

Similar stability is also observed on six-membered 1,3-dioxo-2-silacycloalkanes with

smaller substituents bonded to the silicon atom, demonstrating the superior hydrolytic

resistance of this ring structure.


Conclusions

Slowing or inhibiting hydrolysis of alkoxysilane bonds on silicon-containing

polymers can be beneficial because it allows for easier manipulation and processing of

materials without significant hydrolytic crosslinking. It is well known that increasing the

steric bulk of the substituents on the silicon atom slows or inhibits hydrolytic bond

cleavage. On a linear polycarbosilane, the choice of substituents may be limited by the

requirements of the polymer backbone. In an effort to improve the hydrolytic stability of

the series of substituted carbosilane ADMET polymers prepared in this work, only two

reactive sites on the silicon atom can be modified. Substitution of reactive Si-Cl bonds

using alkyllithium reagents produces a stable alkylated material, but careful

considerations must be made to avoid competing lithiation reactions. Use ofphenoxy

groups was found to increase hydrolytic stability. However, tertiary alcohols were unable

to substitute both Si-Cl bonds; secondary alcohols such as isopropanol are the most

sterically hindered nucleophiles able to substitute both sites. These show significant

improvements in hydrolytic stability compared to primary alkoxysilanes. Cyclic species

were also evaluated in an effort to study the effect that a ring structure would have on

moisture resistance. Four and five-membered rings from geminal and vicinal hydroxy






86

groups did not form, due to excessive ring strain of the potential product. Six-membered

rings form readily and show good stability under atmospheric conditions.'52

It would be very interesting to apply the knowledge learned in the experiments

described in this chapter for the preparation of new substituted polycarbosilanes via

ADMET polymerization. Since these substituents all display reduced tendencies to be

hydrolyzed under atmospheric conditions, the resulting polymers should show the same

properties. Not only would these polymers possess interesting architectures and a broad

range of physical properties, but also the rate of and ability to be crosslinked can be

selectively chosen by judicious choice of the pendant group.













CHAPTER 5
EXPERIMENTAL


Materials

Di-n-hexyldichlorosilane was purchased from Gelest, distilled under reduced

pressure (bp = 113 C/6 mmHg), and stored over activated 4A molecular sieves under

argon. The 5-bromo-l-pentene was purchased from Aldrich and stored over activated 4A

molecular sieves. Silicon tetrachloride (Aldrich) was transferred to a dried Kontes flask

via cannula and stored in an Ar atmosphere glovebox. The 1,2-dibromoethane (Aldrich)

was distilled from CaH2 and stored over activated 4A sieves. Anhydrous methanol and

2-propanol were purchased from Aldrich, further dried over Mg/I2, and stored over

activated 3A molecular sieves. Purification of 2,2,2-trifluoroethanol (Aldrich) was

accomplished by distillation under Ar and was stored over activated 3A sieves. Benzyl

alcohol (anhydrous) and t-butyl alcohol (anhydrous) were purchased from Aldrich and

stored over activated 4A sieves. Phenol (Aldrich) was azeotropically distilled from

benzene. A 50/50 mixture of cis- and trans-1,3-cyclohexanediol (Aldrich) and 2,4-

dimethyl-2,4-pentanediol were used as received. Magnesium turnings (Aldrich) were

dried in a vacuum oven at 100 C before use. The lithium reagents CH3Li [1.4M in Et2O]

(Acros) and [1.6M in ether] (Aldrich), CrHsLi [1.8M in cyclohexane-ether] (Aldrich),

and n-BuLi [1.6 M in hexanes] (Aldrich) were used as received and titrated according to

the method by Suffert.'57 Potassium ethoxide and 4-dimethylaminopyridine (DMAP)






88
(Aldrich) were used as received. Triethylamine was purchased from Aldrich and dried by

distillation over CaH2. N-phenyl-1-naphthylamine was dried at 56 C under vacuum

using an Abderhalden apparatus. Diethyl ether and pentane (Fisher) were dried and

distilled over NaK-benzophenone ketyl. Benzene (Aldrich) and THF (Fisher) were dried

and distilled over K-benzophenone ketyl. Heptane, toluene, and o-xylenes (Aldrich)

were dried by distillation over Na-benzophenone ketyl. Hexanes, chloroform, and

CH2C12 (Fisher) were used as received. Deuterated solvents, C6D6 and CDCl3,

(Cambridge Isotope Laboratories) were stored over activated 4A sieves. The NHC-

ligated 2"d generation Grubbs' [Ru] catalyst (11) Cl2Ru(IMes)(PCy3)[=CHPh] was

synthesized by literature procedure.56 Schrock's [Mo] catalyst (6),

[Mo=CHCMe2Ph(=N-C6H3-i-Pr2-2,6)(OCMe(CF3)2)2], was prepared using a literature

procedure.35'36


Instrumentation

All NMR spectra, 'H (300 MHz), "C (75 MHz), '9F (282 MHz), and 29Si (60

MHz) were conducted on either a Varian VXR, Gemini, or Mercury series super-

conducting spectrometer system and referenced to residual CsH6 or CHCI3 solvent

signals. Fluorine-19 NMR spectra were internally referenced to CFCI3. For the 29Si

NMR spectra, a heteronuclear gated decoupling pulse sequence with a pulse delay of 30 s

was used, with a 1% internal TMS reference added. For quantitative 13C NMR

experiments, a gated decoupling pulse sequence was used with a delay of 10 s. Gas

chromatography was performed on a Shimadzu GC-17A gas chromatograph equipped

with a 15m Restek RTX-5 crossbonded 5% diphenyl-95% dimethyl siloxane column

using He as the carrier gas and a FID detector. Gel permeation chromatography was








performed using two 300mm Polymer Laboratories gel 5pm mixed-C columns. The

GPC instrument consisted of a Rainin SD-300 pump, Hewlett-Packard 1047-A RI

detector, Kratos Spectroflow 757 UV detector (254 nm), TC-45 Eppendorfcolumn heater

set to 30 C, and a Waters U6K injector. The solvent used was CHCI3 at a flow rate of

1.0 mL/min and the peaks were referenced to polystyrene standards from Polymer

Laboratories (Amherst, MA). Differental scanning calorimetry measurements were

taken using a Perkin Elmer DSC 7 instrument equipped with TAC 7/DX controller and a

CCA7 cooling accessory. The samples were scanned from -95 C to 70 C at a heating

rate of 10 C per minute. Liquid N2 was used as the coolant. Spinning band distillation

was performed using a B&R Instruments Model 8T regulated with a dibutylphthlate-

filled manostat. Elemental analyses were performed by Atlantic Microlab (Norcross,

GA).


Synthesis of Substituted Dialkoxvsilane Monomers


Synthesis of di(4-pentenvl)dichlorosilane (26)

A 500 mL three-necked flask equipped with an addition funnel, condenser, and

stir bar was flame-dried under vacuum. The apparatus was then flushed with Ar and 7.27

g Mg turnings (0.30 mol) were weighed and added to the flask. Freshly distilled dry

diethyl ether (175 mL) was added to the flask via cannula, followed by the addition of 0.2

mL of 1,2-dibromoethane. Bubbling ensued, the gray mixture was stirred at room

temperature for I h, and 5-bromo-l-pentene (40.50 g, 0.27 mol) was added dropwise via

addition funnel over 2 h. The reagent was refluxed overnight, followed by removal of

insoluble material by filter cannulation into a 250 mL volumetric flask fitted with an air-






90

free adapter. This gray solution was filled to the volume mark with additional dry diethyl

ether via cannula. The 4-pentenyl magnesium bromide formed was titrated by the method

described in Vogel's (0.22 mol, 81.5%)."' Following titration, a 1000 mL three-neck

round bottom flask equipped with a condenser, addition funnel, and stir bar was flame

dried under vacuum. After purging the flask with Ar, 500 mL of dry heptane was added

via cannula. In a separate 25 mL pear-shaped flask and added to the round bottom flask,

SiCL (16.86 g, 0.10 mol) was diluted with 10 mL dry heptane via cannula. The Grignard

solution was transferred to an addition funnel via cannula and added dropwise over 6 h.

The solution was refluxed overnight, forming a white precipitate. The product was

separated from magnesium salts by filter cannulation into a 1000 mL Schlenk flask; the

residual salts were washed with 2 x 100 mL portions of pentane, and all organic solutions

were combined. All solvent was removed in vacuo, giving a crude yellow liquid product.

The product, a clear, colorless liquid, was purified by vacuum spinning band distillation.

Boiling point: 90-92 OC (4 mmHg), % yield (GC): 73%; isolated 8.80 g, 38%. 'H NMR

(5, CDCI3): 1.10 (Si-CH2)(m, 4H), 1.59 (CH2)(m, 4H), 2.10 (CH2CH)(m, 4H), 4.98

(CH2CH)(m, 4H), 5.74 (CHCH2)(m, 2H). "C NMR (5, CDCI3): 19.7 (Si-CH2), 21.7

(CH2), 36.2 (CH2CH), 115.5 (CHCH2), 137.7 (CHCH2). 29Si NMR (5, C6D6): 33.3

(R2SiCl2). Elemental Anal. for CloHlsSiCl2: Calcd (Found) C 50.84 (50.79) H 7.68

(7.78).


Synthesis of di(4-pentenvl)dimethoxvsilane (27)

Into a 250 mL three-necked round bottom flask equipped with a condenser and

stir bar, 150 mL of dry diethyl ether was added via cannula. Using a syringe, 5.3 mL

(0.038 mol) of dry Et3N was added to the reaction flask, followed by 1.6 mL (0.038 mol)






91

of dry MeOH. Di(4-pentenyl)dichlorosilane, (2.26 g, 0.0095 mol), was diluted with 10

mL dry diethyl ether and added to the reaction flask via cannula. Precipitation of

Et3NH+CI salts was immediate. The reaction mixture was refluxed overnight, followed

by removal of all volatiles in vacuo. Dry pentane was used to wash the product mixture

(3 x 75 mL) and to precipitate any dissolved salts. The crude product was isolated as a

yellow liquid and then purified by vacuum distillation. The final product was a clear,

colorless liquid. Boiling point: 50 C (0.005 mmHg), % yield (GC) 98%; isolated

1.26 g, 58%. 'H NMR (8, CDC13): 0.62 (Si-CH2)(m, 4H), 1.45 (CH2)(m, 4H), 2.05

(CH2CH)(m, 4H), 3.52 (OCH3)(s, 6H), 4.98 (CHCH2)(m, 4H), 5.81 (CHCH2)(m, 2H).

"C NMR (8, CDCI3): 11.4 (Si-CH2), 22.1 (CH2), 37.3 (CHCH2), 50.3 (OCH3), 114.8

(CH2CH), 138.5 (CHCH2). 29Si NMR (5, C6D6): -4.2 (R2SiOR2). Elemental Anal. for

Ci2H24SiO2: Calcd (Found) C 63.12 (63.47), H 10.60 (10.65).


Synthesis of di(4-pentenl)diethoxvsilane (28)

A 250 mL three-necked round bottom flask equipped with a stir bar was taken

into a glovebox and charged with 1.57 g (0.019 mol) ofKOEt. Di(4-pentenyl)-

dichlorosilane, (2.00 g, 0.0084 mol), was weighed out in the glovebox into a separate 25

mL flask. Both flasks were removed from the glovebox and an addition funnel and

condenser were attached. Dry THF was added to the three-necked flask using a cannula

(150 mL); the chlorosilane was diluted with 10 mL dry THF. The chlorosilane solution

was transferred to the addition funnel via cannula and added dropwise over 30 min.

Precipitation of white salts occurred immediately and the reaction mixture was refluxed

overnight under Ar. Residual THF was removed in vacuo, and the product was washed

with 3 x 75 mL of dry pentane. Any remaining inorganic salts precipitated out by the







wash were removed via filter cannulation. All volatiles were removed under vacuum,

giving a crude orange liquid product, which was subsequently purified by vacuum

distillation. The distilled product was a clear, colorless liquid. Boiling point: 47-50 C

(0.005 mmHg), % yield (GC) 88%; isolated 0.95 g, 44%. 'H NMR (8, CDCI3): 0.65 (Si-

CH2)(m, 4H), 1.13 (CH2CH3)(t, 6H), 1.58 (CH2)(m, 4H), 2.05 (CH2CH)(m, 4H), 3.68

(OCH2)(q, 4H), 5.07 (CHCH2)(m, 4H), 5.78 (CHCH2)(m, 2H). 13C NMR (6, CDCI3):

13.0 (Si-CH2), 19.1 (OCH2CH3), 23.3 (CH2), 38.1 (CHCH2), 58.6 (OCH2), 115.3

(CH2CH), 139.3 (CHCH2). 29Si NMR (6, CsD6): -7.9 (R2SiOR2). Elemental Anal. for

C14H28SiO2: Calcd (Found) C 65.58 (65.92), H 11.02 (10.98).


Synthesis of di(4-pentenvl)di(trifluoroethoxy)silane (29)

Using an analogous procedure to that of 27, 4.7 mL (0.034 mol) of Et3N and 2.5

mL (0.034 mol) of CF3CH2H were added to the reaction flask containing diethyl ether.

Addition of 2.00 g (0.0084 mol) ofdi(4-pentenyl)dichlorosilane to the reaction mixture

was performed via cannula. Identical workup and reaction conditions were used as in

the preparation of 27, and the isolated product was a cloudy, yellow suspension that was

purified by vacuum distillation. The final isolated product was clear and colorless.

Boiling point: 50-51 C (0.005 mmHg), % yield (GC) 98%; isolated 1.76 g, 57%. 'H

NMR (8, CDCI3): 0.71 (Si-CH2)(m, 4H), 1.50 (CH2)(m, 4H), 2.10 (CH2CH)(m, 4H),

4.03 (OCH2)(q, 4H), 5.03 (CHCH2)(m, 4H), 5.78 (CHCH2)(m, 2H). '3C NMR (8,

CDCI3): 11.3 (Si-CH2), 21.5 (CH2), 36.9 (CHCH2), 60.3, 60.8, 61.3, 61.8 (OCH2), 115.3

(CH2CH), 118.0, 122.0, 125.7, 129.5 (CF3), 138.0 (CHCH2). 29Si NMR (6, C6D6): -7.7

(R2SiOR2). '9F NMR (8, CFCI3): -77.2 (t, CF3). Elemental Anal. for Ci4H22SiO2F6:

Calcd (Found) C 46.14 (46.41), H 6.09 (6.13).




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EKACYD524_UO1VCO INGEST_TIME 2013-09-28T00:40:44Z PACKAGE AA00014258_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES