Synthesis, characterization, and rotational isomerization of chelate stabilized tungsten(6) alkylidenes, olefin metathes...


Material Information

Synthesis, characterization, and rotational isomerization of chelate stabilized tungsten(6) alkylidenes, olefin metathesis polymerization catalyst precursors
Physical Description:
vii, 240 leaves : ill. ; 29 cm.
Blosch, Laura L
Publication Date:


Subjects / Keywords:
Spectrum analysis   ( lcsh )
Organometallic compounds -- Synthesis   ( lcsh )
Chemistry thesis Ph. D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1993.
Includes bibliographical references (leaves 230-239).
Statement of Responsibility:
by Laura L. Blosch.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001922340
oclc - 30495406
notis - AJZ8152
System ID:

Full Text








My sincere appreciation for his guidance, enthusiasm and personal support is

extended to Dr. Jim Boncella. Jim has a strong commitment to personally mentoring each

of his students, yet he allowed me the flexibility to explore those aspects of chemistry

which were interesting to me but outside of his field of personal experience and

professional expertise, enabling me to develop the invaluable traits of self-confidence and

self-reliance. To Dr. Scott Gamble, I extend a very personal thank you for his invaluable

friendship which has meant so much to me personally and professionally. Additionally,

my gratitude extends to the other present and former members of Dr. Boncella's group, for

stimulating interaction and graduate-student kinship: Dr. Gaines Martin, Dr. Rob

Duttweiller, Dr. Chris Bauch, Will Vaughan, Dan Vanderlende, Larry Villanueva, Chris

Coston, Kevin Cammack, Tegan Eve, Justine Roth and especially Jerrold Miller, for

graciously loaning me his Mac, and Mary Cajigal and Percy Doufou for their selfless hours

of editing assistance and personal support. For their polymer expertise and perspective I

wish to thank Dr. Jasson Patton, Dr. Jim Konzelman, Dr. Dennis Smith and especially,

Dr. Ken Wagener. Finally, I wish to extend my eternal gratitude my husband, Ed, for the

inestimable support, encouragement and understanding he has consistently provided. I am

constantly enriched and delighted by his unique perspectives on life and science.


ACKNOWLEDGEMENTS ........................................................................ ii

ABSTRACT ......................................................................................... iii


1 BACKGROUND AND INTRODUCTION........................................ 1

Olefin Metathesis and Classical Catalysts .......................................... 1
Transition Metal-Carbon Double Bonds ..............................................4
Synthesis of Alkylidenes....................................... .................. 9
Transition Metal-Carbon Triple Bonds...............................................16
Olefin Metathesis Polymerizations...................................................18


Synthesis of Neutral Trispyrazolylborate Alkylidyne and Alkylidene
Synthesis of a Hydrotris(3,5-dimethyl-l-pyrazolyl)borate (Tp')
Neopentylidyne.............................................................. 34
Conversion to a Hydrotris(3,5-dimethyl-l-pyrazolyl)borate (Tp')
Neopentylidyne.............................................................. 35
Synthesis of a Hydrotris(l-pyrazolyl)borate (Tp) Neopentylidyne ..........40
Conversion to a Hydrotris(1-pyrazolyl)borate (Tp) Neopentylidyne........42
Synthesis of a Hydrotris(3,5-dimethyl-l-pyrazolyl)borate (Tp')
Conversion to a Hydrotris(3,5-dimethyl-1-pyrazolyl)borate (Tp')
Benzylidene..................................................................... 45
Synthesis of a Stable Cationic Alkylidene ..........................................48
Attempted Facile Synthesis of Trispyrazolylborate Stabilized
Tungsten Alkylidenes..................................................... ....56
Discussion............... ........................................... ..... 62
Characterization .................................................................62
Synthesis of High Oxidation State Trispyrazolylborate Compounds ...... 71
Conversion to Trispyrazolylborate Oxo Alkylidene and Dioxo Alkyl
Conclusions .........................................................................80

ALKYLIDENES ....................................................................83

Introduction........................................... ...........................83
Metathesis Polymerizations........................................................... 84

Tp'W ()(CHC(CH3)3)Cl Reactivity Studies .......................................90
Attempted Halide Abstraction from Tp'W(O)(CHC(CH3)3)C1................90
Synthesis of Tp' Tungsten Oxo Neopentylidene Methyl Compound.........95
Synthesis of a Tp' Tungsten Oxo Neopentylidene Hydride .............. 101
Conclusions ........................................................................ 108

4 ALKYLIDENE ISOMERIZATION................................................ 114

Background.......................................................................... 114
Solution Phase Conformation Determinations................................... 122
Thermal Isomerization of
[TpW (NAr)(PyrH)(CHC((CH3)2Ph)] [S 3CF3] .......................... 127
Photoisomerizations............................................................... 132
Tp'W(CHC(CH3)3)(O)Cl................................................. ...... 132
TpW(NPh)(CHC(CH3)2)C .................................................... 138
TpW(NAr)(CHC(CH3)2Ph)(OTf) ......................................... 139
Activation Parameters for Anti to Syn Alkylidene Isomerization................ 141
Entropy of Activation............................................................ 142
Statistical Mechanics Calculations of Entropies of Free Internal
Rotation................................................................... 145
Enthalpy of Activation........................................................ 150
Spectroscopic Properties of the Anti Alkylidene Isomers........................ 152
Nuclear Magnetic Resonance ................................................... 152
Infrared Spectroscopy ....................................................... 154
Discussion of Spectroscopic Observations................................... 155
Conclusions ........................................................................... 159
5 EXPERIMENTAL.................................................................... 161
Materials and Methods ............................................................... 161
Syntheses ............................................................................. 162
TpPW (CC(CH3)3)Cl3 ....................................................... 162
[(HB)2(m-N2C5H7)3][(((CH3)CC)WCl2)2(m-Cl)2(m-N2C5H )].......... 163
Tp'W(CPh)Br2 ................................................ 164
TpW(CC(CH3)3)C13 ..................... .............................. 165
Tp'W(O)(CHC(CH3)3)C1 ..................................................... 166
TpW(O)(CHC(CH3)3)Cl ....................................................... 167
TpW(O)(CHPh)Br ......................................................... 168
[Tp'W(CHC(CH3)2Ph)(NAr)(PyrH)] [SO3CF3] ........................... 170
Tp'W(CHC(CH3)2Ph)(NAr)(Pyr) ........................................ 172
Tp'W(O)2C ................................................................... 173
Tp'W(0)2(CH2Ph) ......................... .. .................................. 174
Tp'WC13 and Tp'MoCl3 ........................................................ 175
Tp'W()(C C(C3)HC( 3)3(CH3) ..................................... ........... 176
Tp'W(O)(CHC(CH3)3)(H) ................................................... 177
Crystallographic Studies............................................................ 178
Tp'W(O)(CHC(CH3)3)C1........................................................ 178
[Tp'W(CHC(CH3)2Ph)(NAr)(PyrH)] [S 3CF3] .......................... 179
[(HB)2(m-N2C5H7)3][(((CH3)CC)WC2)2(m-C1)2(m-N2C5H7)].......... 180
Polymerizations....................................................................... 181
Ring Opening Metathesis Polymerization (ROMP) of Cyclooctene......... 181
Ring Opening Metathesis Polymerization (ROMP) of Norbornene......... 182
Experimental Considerations for Kinetics Studies.......................... 182


A REPRESENTATIVE 1H AND 13C NMR SPECTRA ......................... 185

B CRYSTALLOGRAPHIC CHARACTERIZATION.............................. 199

REFERENCES................................................................................ 230

BIOGRAPHICAL SKETCH ................................................................ 240

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


Laura L. Blosch

August, 1993

Chairman: James Boncella
Major Department: Chemistry

The synthesis of a number of trispyrazolylborate stabilized tungsten alkylidynes and
alkylidenes was achieved in moderate to low yield by the addition of the potassium salts of

Tp and Tp' to alkylidyne and alkylidene containing compounds. These materials

represented unprecedented increases in thermal, hydrolytic and oxidative stability for high

oxidation state alkylidyne and alkyidene compounds. They were characterized by 1H
NMR, 13C NMR, MS and IR spectroscopy. Single crystal x-ray studies were performed

for three compounds, including Tp'W(O)(CHC(CH3)3)C1,

[Tp'W(NAr)(CHC(CH3)2Ph)(PyrH)][SO3CF3], and [(HB)2(j-N2C5H7)3]

The trispyrazolylborate tungsten alkylidyne dihalides were readily converted to
trispyrazolylborate tungsten oxo alkylidene monohalides and trispyrazolylborate tungsten

dioxo alkyls by a variety of hydroxide reagents. Other novel compounds prepared included

Tp'W(O)(CHC(CH3)3)(H) and Tp'W(O)(CHC(CH3)3)(CH3).

Addition of one equivalent of a Lewis acid co-catalyst to the trispyrazolylborate

tungsten oxo alkylidenes generates very active ROMP catalysts which operate in moist air

at elevated temperature to produce high molecular weight polymers. The interaction

between Lewis acid and the alkylidene compounds is currently under investigation.

Rotamers of six-coordinate trispyrazolylborate tungsten alkylidenes have been

generated thermally and photochemically. In all cases the thermodynamic isomers were

oriented syn to the x-bonding terminal oxo or imido ligand. Entropy of activation was

correlated with steric interaction in internal rotations which were modeled by statistical
mechanics. A weak, bent, intramolecular hydrogen bond between the a-proton of the

alkylidene and the terminal oxo or imido ligand has been proposed to account for

observations in kinetics studies, NMR and IR spectra of the minor isomers.


The synthesis and characterization of transition metal alkylidene and alkylidyne

compounds is a rapidly expanding,' vital new area of organometallic research. Since such

compounds were first reported less than twenty years ago,2 great strides have been made in

overcoming the difficult synthetic hurdles to preparing these compounds. The synthesis of

alkylidenes and alkylidynes has been strongly linked to the development of olefin

metathesis catalysts,35 and is best understood historically from this perspective. As the

number of transition metal alkylidene and alkylidyne compounds has increased, so has

chemists' understanding of the olefin metathesis reaction and of the fundamental nature of

the transition metal-carbon multiple bond.

Olefin Metathesis and Classical Catalysts

The olefin metathesis reaction can be described as the net breaking of two

carbon-carbon double bonds and the formation of two new carbon-carbon double bonds

with the resulting exchange of olefin substituents,6 Figure 1-1.

R R"
R H H R" R\/R"

+ s+

R' H H R'" H
R' R'"

Figure 1-1. The olefin metathesis reaction.

In addition to olefin, a transition metal catalyst is required for olefin metathesis to

occur under mild conditions.7'8 Most early olefin metathesis catalysts were ill-characterized

mixtures of transition metal halide or oxo halide compounds with Lewis acid co-catalysts,

usually an alkylhalide of aluminum, tin or zinc.5'6 These early catalyst mixtures were

sensitive to temperature and pressure contraints, to traces of impurities, and even to

incompatible functional groups on the olefins which tended to poison the catalysts.5 One

consequence of the sensitivity of most of these systems to functionalities, such as ketones

and amines, was their limitation to hydrocarbon olefin metathesis reactions. Given the

nature of the early catalyst systems, it is not surprising that the mechanism for olefin

metathesis was not well understood initially. Because of superficial similarities between

olefin metathesis catalysts and Ziegler-Natta catalysts,5 comparisons between these catalyst

systems were the major influence on early proposals on the mechanism of olefin


The pairwise mechanism for olefin metathesis was initially suggested by Bradshaw

and coworkers9 in 1964. In this mechanism, two olefins approach the metal center, which

serves as a loose template for the formation of a quasicyclobutane complex, Figure 1-2. As

the quasicyclobutane complex breaks down, the new olefins are formed. This mechanism

was appealing since the metal participation was critical to stabilize the quasicyclobutane

formation (formally a two plus two cycloaddition that is symmetry forbidden),11 yet was

not demanding of a particular functionality on the metal center. Thus, the pairwise

mechanism offered an explanation for the large variety of dissimilar metal compounds and

Lewis acid mixtures that were viable catalysts, yet seemingly had little else in common.5'6

As research on the product ratios and kinetics of the olefin metathesis reaction

progressed, some researchers became increasingly dissatisfied with the pairwise

mechanism. Considerable effort and many elegant mechanistic studies were undertaken in

an effort to definitively describe the mechanism for olefin metathesis.112 The mechanism

that is universally accepted today was first proposed by H6rrison and Chauvin13 in 1970.

C C" C------

C' C'" I

Figure 1-2. The pairwise mechanism for olefin metahtesis with olefin substituents not
shown for clarity.

In this non-pairwise mechanism, a metal-carbon double bond was proposed to be formed
in the catalyst mixtures. This metal-carbon double bond then becomes the active site of the
catalyst. Olefins were proposed to approach the active site singly instead of in pairs, and to
precoordinate to the metal center. Once coordinated, the metal-carbon double bond and the
olefin interact to form a metallacyclobutane which can then decompose in either a
productive fashion, to give a new olefin and new metal-carbon double bond, or in a
degenerate fashion, to regenerate starting materials, Figure 1-3.

R H R R'

[M] H R" M
R'_( R"H"w ^ R' +i R"

R' H

Figure 1-3. The non-pairwise mechanism for olefin metathesis with arrows indicating the
direction of the electron flow for productive metathesis.

Consideration of the components of early olefin metathesis catalyst mixtures
revealed that metal-carbon double bonds could have been formed from the mixture
components.14.15 The typical reaction scheme invoked calls for alkylation of the transition
metal halide compound by the alkylhalide Lewis acid co-catalyst, followed by a-hydrogen

abstraction from the resulting transition metal alkyl to yield a transition metal-carbon double
bond. Figure 1-4 illustrates this scheme for the WOCl4/Et2AlCl2 catalyst system.

Similar schemes are envisioned for other catalyst mixtures. It should be noted that this
scheme is only one of many possible schemes that can be envisioned for generating
transition metal-carbon double bonds from these mixtures.

o +l C1 CH3CH3 I1 HI
C C (AlC4)x C
Cl Cl C. C1
Figure 1-4. One pathway envisioned for double metal-carbon double bond formation in a
classical olefin metathesis catalyst mixture.

Spectroscopic evidence for transitory metal-carbon double bond formation in these
types of mixtures has been presented.5 Tebbe and coworkers were able to isolate a Lewis-
acid stabilized titanium methylidene compound, shown in Figure 1-5, from a mixture of
biscyclopentadienyl (Cp) titanium dichloride and excess trimethylaluminum.16 This
compound was shown to catalyze olefin metathesis and to form Lewis-acid-free
metallacyclobutanes in the presence of olefins, which could themselves be isolated and
used to catalyze olefin metathesis.17 Subsequently, Tebbe's reagent has found considerable
application in the field of organic and organometallic synthesis18 as a carbene donor.

46 CH2 CH3
Ti ^Al
i.ClI CH3

Figure 1-5. Lewis acid stabilized titanium methylidene prepared by Tebbe and coworkers.

Transition Metal-Carbon Double Bonds

The quest for isolable, Lewis-acid-free, transition metal-carbon double bonds was
complicated by the existence of two types of these interactions.1'1'19 The first of these

types is the Fischer carbene. The first example of a Fischer carbene was the six-coordinate
tungsten pentacarbonyl methoxyethylidene,20 Figure 1-6.

(OC)5W 4

Figure 1-6. The first reported Fischer carbene compound.

The distinguishing features of Fischer carbenes include a late transition metal in a

low oxidation state; a c-donor substituent (e.g., a heteroatom or phenyl ring) on the
carbene a-carbon; i-acceptor ancillary ligands (e.g, CO); and electrophilic character for

reactions at the carbene a-carbon. 112'19 Many Fischer carbene compounds with a variety

of transition metals (most common are group 6, 7, 8 and 9 metals) and t-donor

substituents have been prepared.21 The reactivity of Fischer carbenes has also been

extensively reviewed22 and has found application in many organic syntheses, including the
Fischer-Tropsch synthesis19 and cyclopropanation reactions.23 With few exceptions

Fischer carbenes are not active metathesis catalysts, although a few, including

(OC)5WC(OCH3)(CH3) (Figure 1-6), have been observed to undergo stoichiometric olefin
metathesis reactions.

The second type of transition metal-carbon double bond is the Schrock carbene or

alkylidene. Alkylidenes were first discovered at the DuPont Central Research Department

in 1973 by Schrock when he attempted to prepare pentakisneopentyl tantalum.24 Instead of

the projected peralkylated tantalum compound, a four-coordinate tantalum trisneopentyl

neopentylidene compound was isolated,2 Figure 1-7.

The distinguishing features of Schrock alkylidenes include an early transition metal

in its highest (d0) oxidation state; a hydrocarbon or hydrogen substituent on the alkylidene
a-carbon; x-donor ancillary ligands (e.g, imido, oxo, amido, alkoxide, and halide); and

nucleophilic character for reactions at the alkylidene a-carbon.1LZ19 Schrock alkylidene

T T a_H -
T -a IT H 3_ H
( T4Ci C17ll Li ]TaT

Figure 1-7. The first transition metal alkylidene synthesis.

compounds have been reported for zirconium, niobium, tantalum, molybdenum, tungsten,
and rhenium transition metals.1 They have been implicated as important reaction
intermediates or catalysts in many significant transformations including Wittig-like
chemistry,2526 the Shell higher olefi process,12,19 and olefin metathesis, especially olefin
metathesis polymerizations.5,6
There has been some dissention in the literature over the distinction between Fischer
carbenes and Schrock alkylidenes.27 This disagreement has stemmed from the fact that
there are a considerable number of compounds containing transition metal-carbon double
bonds, with group 6 metals in particular, which have somewhat ambiguous oxidation
states. In large part this ambiguity arises from the fact that alkylidenes oxidize the metal
center by two electrons while carbenes are considered to be neutral ligands.1,12 Further,
there are a few compounds which have some of the distinguishing characteristics and/or
reaction proclivities of both types of transition metal-carbon double bonds. For example,
the low-valent osmium nitrosyl bistriphenylphosphine methylidene chloride compound,

Os(NO)(PPh3)2(CH2)C1, was shown to react with both nucleophiles and electrophiles at
the methylidene (or carbene) a-carbon atom,27 Figure 1-8. Understanding the reactivity of

transition metal compounds depends on understanding the electronic factors which
influence the frontier orbitals of these materials.
A number of ab initio studies have been conducted to define the electronic structure
of both carbenes and alkylidenes.28'29 The electronic structure of a Fischer carbene is best
described as consisting of a-donation from the singlet carbene lone pair (sp2 hybrid) to a

metal d-orbital, and n-donation from a filled d-orbital on the transition metal to the empty p-

Ph3 Ph3 Ph3
ON \ O ON \ ON
Os' Os"" / O S=O
Cl1 I C1/p H c1 c
p CH2 C Cl p C
Ph3 Ph3 Ph3 H2
Figure 1-8. A low-valent osmium compound which evidences both Fischer carbene and
Schrock alkylidene reactivity patterns.

orbital on the carbene a-carbon, Figure 1-9. The other lobes of the d-orbital accepting the
a-donation from the carbene singlet have been omitted from Figure 1-9 for clarity. This
type of interaction is very similar to the Dewar-Chatt-Duncanson model1930 of olefin
binding to transition metals.

Figure 1-9. Electronic structure of a Fischer carbene.

For late transition metals, the dx-orbitals are lower in energy than px-orbital of the
carbene fragment, so the metal-carbene x-bond has considerable metal character and the
metal-carbene interaction is polarized in the M--C+ direction.2 The carbocation-like
character of the carbene a-carbon is stabilized by a a-donor substituent in the B-position,
resulting in some single-bond character for the metal-carbon carbene bond and some
double-bond character in the a-carbon-B-substituent bond,2 Figure 1-10.
The electronic structure of a Schrock alkylidene has been described based on ab
initio studies as consisting of a-donation from the triplet alkylidene (sp2 hybrid) to a metal
d-orbital, and x-donation from a triplet metal d, orbital to the px-orbital on the alkylidene
a-carbon.2831 For early transition metals, the metal d-orbitals are higher in energy than the


Figure 1-10. Valence bond description of the resonance structures for a Fischer carbene
stabilized by a 7t-donor substituent (E) in the B-position.

alkylidene p-orbital and, consequently, the metal-alkylidene bond has considerable
alkylidene character. Thus, Schrock-type alkylidenes are polarized in the M+-C- direction
to the extent that the metal dx triplet can be considered to be transferred to the alkylidene
fragment This results in an oxidized metal center and an alkylidene with a negative charge
(-2).1'12 The Schrock alkylidene can then be pictured as a metal-stabilized carbanion which
is both a o- and a 7i-donor to the metal center,28'31 Figure 1-11. The other lobes of the d-
orbital accepting the o-donation from the alkylidene triplet have been omitted from Figure

1-11 for clarity.

Figure 1-11. Electronic structure of a Schrock alkylidene.

It seems reasonable to think of these electronic pictures as limiting extremes12 and to
expect compounds with intermediate properties, or mixtures of both properties to occur
with increasing frequency, as new transition metal-carbon double bond-containing
compounds with increasingly diverse ligand systems and substituents are prepared. As this
perception becomes more prevalent, a number of interesting propositions have arisen. The
most intriguing, and most controversial of these was presented by Rooney, Green and
coworkers who suggested that some Ziegler-Natta catalysts may proceed through

alkylidene intermediates.32 Since insights provided by comparison of Ziegler-Natta and

olefin metathesis catalysts influenced the first mechanistic proposals for olefin

metathesis,9'10 this proposal brings the field of alkylidene metathesis mechanistic studies

full circle. The proposal of Ivin et al. further suggests that it may be possible to synthesize

a catalyst that will facilitate both olefin metathesis and Ziegler-Natta polymerizations,

making it possible to synthesize intriguing new types of block copolymers.

Synthesis of Alkylidenes

Hdrrison and Chauvin's non-pairwise alkylidene mechanism for olefin metathesis

gave considerable impetus to an intensive synthetic investigation directed at transition metal

alkylidene synthesis.1 The quest for isolable alkylidene compounds became inextricably

entangled with the directed synthesis of Lewis-acid-free olefin metathesis catalysts.33 As a

result of this study, however, some general means of synthesizing transition metal

alkylidenes were delineated. It is important to note that not all synthetic routes to

alkylidenes apply to every transition metal. The brief survey of alkylidene synthesis

presented herein describes the most general methods for transition metal alkylidene

synthesis and those methods which apply to tungsten alkylidene synthesis. Many of the

examples given are tantalum reactions, since the earliest and most thorough mechanistic

work was done on tantalum compounds. Analogous tungsten reactions are known for

these cases, but frequently the mechanisms for the tungsten reactions were proposed by

comparison with the analogous tantalum chemistry.

One of the most general methods for synthesizing transition metal alkylidenes is the
a-hydrogen abstraction reaction43437 which resulted in the synthesis of the first alkylidene

compound, previously shown in Figure 1-7. This reaction, sometimes called dealkylation,

requires that the metal center be sterically encumbered. It works best for neopentyl

compounds for that reason. Frequently addition of sterically demanding ligands can induce
a-hydrogen abstraction in polyalkyl transition metal compounds,24'38 Figure 1-12. The

product of the reaction of Figure 1-12, bistrimethylphosphine tungsten neopentyl
neopentylidene neopentylidyne, was the first compound to manifest single, double and
triple transition metal-carbon bonds in the same molecule.39 A brief discussion of transition
metal-carbon triple bonds will be presented later in this chapter.


WA / 2PMe3

< Me,
Figure 1-12. a-Hydrogen abstraction induced by bulky ancillary ligands.

A related synthetic route to transition metal alkylidenes is deprotonation (or net
dehydrohalogenation) of an alkyl compound with a strong base.40 Frequently either direct
deprotonation of an alkyl compound or base coordination followed by a-hydrogen
abstraction can be proposed as the mechanism for a given reaction,1 Figure 1-13.

a-hydrogen Li Ta Ta
abstraction 3 H H

d \7 Li Ta T.
deprotonation [/L3i (7 Ta _H N (H


Figure 1-13. Comparison of the a-hydrogen abstraction and base initiated deprotonation
mechanisms for alkylidene synthesis.

Both the a-hydrogen abstraction and the alkyl deprotonation sythetic routes to
alkylidenes depend on the synthetic accessibility of polyalkyl transition metal compounds.

Synthesis of these polyalkyl compounds is frequently nontrivial. Complications arising
from disproportionation and other reductive processes are well known and frequently result
in low yields for polyalkyl syntheses.41 Some success has been reported in stabilizing the
metal center to these reductive processes by incorporating x-donor ancillary ligands into the
compounds prior to alkylation or by using mild alkylating reagents.4244 Polyalkyl
compounds of tungsten were essentially unknown prior to attempted alkylidene synthesis,
and the few compounds reported were very unstable and sometimes explosive.19 Because
of these complications, the first tungsten alkylidene was synthesized by alkylidene transfer
from a tantalum alkylidene compound, Figure 1-14.

O Me3
OH p
Cl3(PMe3)i2Ta=II H

t-Bu-O-- -0-t-Bu
t-Bu-O O-t-Bu 1/2[Ta(O-t-Bu)4C1]2 Mel
Figure 1-14. Alkylidene transfer from tantalum to tungsten.

The viability of deprotonation of transition metal alkyl compounds as a synthetic
route to alkylidenes suggests that protonation of transition metal alkylidynes might also
yield alkylidenes. This approach to alkylidene synthesis has been successful with a variety
of acids,45,46 Figure 1-15. Related intramolecular proton transfer reactions have also
proven useful for alkylidene synthesis, such as the triethylamine-catalyzed proton transfer
from an amido ligand to an alkylidyne, yielding the imido alkylidene,4146 Figure 1-16.

t-Bu-O X
2 HX t-Bu-O I H
t-Bu-O-W= C-- O Tt- <-
\ -HO-t-Bu t-Bu-OC I
t-Bu-O X
Figure 1-15. Protonation of an alkylidyne, resulting in alkylidene synthesis.

NH NEt3 iC Cl
Cl Cl I \ H

O. \ -0

Figure 1-16. Intramolecular proton transfer catalyzed by triethylamine.

Each of these synthetic routes to transition metal alkylidene compounds, a-

hydrogen abstraction, deprotonation of alkyls, and protonation of alkylidynes, has proven

generally applicable to a number of different compounds and transition metalst,12'19. Three

very intriguing synthetic routes which have so far been successful only for limited tungsten

compounds are the oxidative cleavage of ketones reported by Bryan and Mayer,47 the

Wittig-like chemistry of phosphonium ylide,48 and the oxidative addition of 1,1-

diphenylcyclopropene,49 both reported by Grubbs and coworkers.

The four-electron oxidative addition of cyclopentanone with a tungsten(2)

tetrakismethyldiphenylphosphine dichloride compound is shown in Figure 1-17. The
reaction proceeds through a tungsten bis(Tr2-cyclopentanone) bisphosphine dichloride

compound which is stabilized by excess cyclopentenone.47 Although the reaction depicted

in Figure 1-17 gives a good yield of the cyclopentylidene in 24 hours, other ketones give

very poor yields of unisolated alkylidene products even after weeks. A much lower yield

(45%) of an analogous tungsten para-tolylimido cyclopentylidene compound was obtained

after several days of refluxing the tungsten(2) starting material with N-cyclopentyl-para-


Subsequent Hiickel calculations have revealed that these reactions are symmetry

forbidden and occur due to concomitant rotation of the alkylidene ligand.50 The driving

force for these reactions appears to be the formation of the very strong tungsten terminal

C1 O
L-L -2 L IL /

C1 C1
L = PMePh2

Figure 1-17. Oxidative addition of cyclopentanone to tungsten(2).

oxo bond, as Bryan and Mayer had originally suggested.47 Unfortunately, the oxo or

imido alkylidene products react with excess or unreacted cyclopentanone in a Wittig-like

reaction to give insoluble polyoxotungsten compounds and other uncharacterized products,

limiting the utility of this synthesis.

The tendency for Wittig-like chemistry to occur in these systems was exploited by

Grubbs and coworkers in synthesizing a tungsten alkylidene by transferring the alkylidene

moiety from a phosphonium ylide to tungsten imido tetrachloride,48 Figure 1-18.

Alkylidene transfer from phosphonium ylids had been observed for some tantalum51 and

zirconium52 compounds previously. Oxo transfer reactions from tungsten are virtually

unknown.53 The reaction of Figure 1-18 appears to be general for a variety of substituted

and unsubstituted tungsten imido compounds, but the authors note that the reaction fails

completely in the absence of alkoxide ancillary ligands on the tungsten center prior to the

reduction step.48 The alkylidene moiety also appears to be significant, since the

unsubstituted benzylidene compound coordinated phosphine when synthesized by this

route and rapidly decomposed when the phosphine was removed. The ortho-methoxide on

the phenyl ring of the benzylidene synthesized in the reaction of Figure 1-18 was found (by

nOe studies) to coordinate to the metal center, stabilizing the alkylidene product and

preventing the coordination of phosphine. It would seem that the viability of this pathway

will be limited by the coordination of phosphine which occurs in the absence of an

orthometallated or bischelating alkylidene.4

R-S R' Ph3P=CH- RO .r
a THF--W=C
C Na/Hg *0 -H
RORO C1 Ci -PPh3,-2 NaCl NOR
R = OCCH3(CF3)2 R
R' = Me, H, i-Pr
Figure 1-18. Alkylidene transfer from a phosphonium ylide to tungsten.

The reaction of 1,1-diphenylcyclopropene with ruthenium compounds has been
reported to yield vinylalkylidenes49 (or vinylcarbenes), Figure 1-19. Analogous reactions
with tungsten compounds are being investigated,a as well as an improved synthesis of

Ph3 Ph3
Cl P Ph Cl H
Ru-PPh3 +

/7 Ph 1 Z
Cl p Cl p H Ph
Ph, Ph,
Figure 1-19. 1,1-Diphenylcyclopropene addition to ruthenium(2).

An interesting consequence of these synthetic approaches to Schrock alkylidene
compounds and of the electronic nature of multiply bonded ligands themselves is that many
of these compounds are coordinatively and electronically unsaturated. Since the metal in
Schrock alkylidenes is electron deficient (dO), xt-donor ancillary ligands such as oxos,
imidos, halides, and alkoxides can provide some extra electron density and help to stabilize
the metal center.34 Additionally, the a-hydrogen (or as some researchers have suggested,

the alkylidene-carbon-hydrogen bond) can also donate electron density to the metal center
through an agostic interaction.34'46 This alkylidene distortion has been described in two

a Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. Manuscript in preparation.

ways. Ab initio calculations modeled the interaction as an alkylidene rotation resulting in
a-overlap of the sp2 alkylidene carbon-hydrogen hybrid and an empty metal d-orbital.54

Alternatively, the distortion has been viewed as an organometallic example of

hyperconjugation,55 Figure 1-20. The physical consequences of this interaction include a

shortened metal-carbon alkylidene bond length, an increased metal-carbona-carbong angle
(>150); and a decreased metal-carbona-hydrogen angle (<100).1 These changes can be

detected through single crystal x-ray diffraction studies34.46 and through neutron diffraction

studies.56 Neutron studies are needed in order to observe carbon-hydrogen bond lengths

and metal-carbona-hydrogen angles, because x-ray diffraction studies frequently do not

locate hydrogen atoms in close proximity to transition metals.

H +H
[M =C\ -- [M C-R
Figure 1-20. Hyperconjugation model of the agostic interaction of some tungsten

The spectroscopic consequences of these bond length and angle distortions are quite

distinct The infrared stretch of the agostic alkylidene carbon-hydrogen bond is decreased

as is the 1JCH coupling constant in the gated-decoupled nuclear magnetic resonance (NMR)

spectrum.34'46 Distorted alkylidenes were generated in the protonation of alkylidynes to

alkylidenes reported by Schrock and coworkers.41 Frequently, when rotational

isomerizationc of alkylidenes was reported, one of the two possible alkylidene rotamers

was found to exhibit an agostic interaction with the metal center.57 Strongly distorted

alkylidenes are not active metathesis catalysts.

b Discussion of the spectroscopic characteristics of alkylidenes is deferred to chapter 2 of
this dissertation.
c Rotational isomeriazation of alkylidenes is presented in detail in chapter 4 of this

Transition Metal-Carbon Triple Bonds

Returning briefly to the subject of alkylidynes, the transition metal-carbon triple

bond has been termed a Fischer carbyne or a Schrock alkylidyne using the same criteria
proposed for distinguishing between carbenes and alkylidenes.1'12'9 The distinction
between carbynes and alkylidynes is plagued by the same types of ambiguities that affect
the carbene or alkylidene assignment. As a criteria for distinguishing between carbynes
and alkylidynes, oxidation state is particularly ambiguous since the carbyne is a neutral
ligand while the alkylidyne carries three units of charge (-3), so that the oxidation state of
the compound is critically dependent on the assignment of the type of metal-carbon triple
bond. The situation is further complicated by the apparent ease of interconverting carbynes
and alkylidynes demonstrated in the oxidative conversion of a low-valent carbyne to a do
alkylidyne using bromine (and a similar reaction using oxallylbromide),58 Figure 1-21.

C Br
CO Br2, (CH3OCH2)2 -
OC -WC-CHI-I3 Br---WC--CH3
OC/ --
Br O

Figure 1-21. Oxidative transformation of a Fischer carbyne to a Schrock alkylidyne.

Despite the fact that the conversion of carbyne to alkylidyne was reported by Mayr
nearly ten years ago,58 and the obvious great interest in affecting the oxidation, the
oxidative conversion of carbenes to alkylidenes remains unreported. This observation may
suggest that carbynes and alkylidynes are more closely related than carbenes and
alkylidenes, or that the appropriate reaction conditions to affect the carbene to alkylidene
transformation have not yet been found.
One of the most facile routes to tungsten alkylidynes is the cleavage of a triply
bonded tungsten hexakis-t-butoxide dimer by an alkene,59 Figure 1-22. Unfortunately, this

reaction has not succeeded with any other triply bonded tungsten dimers, owing to

peculiarities in the electronic or steric configuration of the hexakis-t-butoxide dimer which

drives the reaction.

t-Bu-O O-t-Bu O-t-Bu
t-Bu-O-WEW- O-t-Bu + EtC-CEt 2EtCW--O-t-Bu
t-Bu-O O-t-Bu O-t-Bu
Figure 1-22. Metathetical cleavage of a tungsten-tungsten triple bond with

Other synthetic routes to alkylidynes are closely related to those used to generate
alkylidenes. Successive a-hydrogen abstraction from polyalkyls has proven

successful,4146 Figure 1-23. Although the exact mechanism of reaction is not well

understood, the evolution of neopentane was confirmed.41 The alkylidyne is almost
certainly generated by successive a-hydrogen abstractions (or by direct deprotonatation) of

an intermediate nepentyl or neopentylidene compound.

MeO-W-- C+ 6 X7 1MgCl ( =
Figure 1-23. Step-wise a-hydrogen abstractions leading to a tungsten alkylidyne.

Transition metal alkylidynes are not olefin metathesis catalysts; however, they are

frequently alkyne metathesis catalysts.60 The alkyne metathesis reaction, Figure 1-24, is

postulated to proceed through metallacyclobutadiene intermediates for some

compounds.61,62 Support for this proposed mechanism includes isolable

metallacyclobutene compounds which have been shown to metathesize acetylenes.

-R- M R -M MR' +
Figure 1-24. Proposed alkyne metathesis mechanism showing metallacyclobutadiene

Olefin Metathesis Polymerizations

One of the most interesting applications of olefin metathesis is metathesis

polymerization.5,6 There are two distinct types of olefin metathesis polymerizations, ring

opening metathesis polymerization (ROMP) and acyclic diene metathesis (ADMET)

polymerization. The ROMP of olefins, much as the name implies, converts strained cyclic

olefin monomers into linear polymers containing units of unsaturation in the polymer

backbone. ROMP has been practiced industrially with great success for many years.5'6

The ADMET polymerization reaction converts acyclic dienes into polymers, also with
unsaturation in the polymer backbone. ADMET oligomerizations were first reported nearly

20 years ago,63 but successful synthesis of high molecular weight materials by ADMET

chemistry was not achieved until recently.64 Consideration of the mechanisms of ROMP

and ADMET polymerizations underscores the differences between these olefin metathesis


The mechanism for ROMP, Figure 1-25, is well understood and has been

extensively reviewed.5,6'65 A transition metal olefin metathesis catalyst coordinates the

olefin.66 Rapid formation of a metallacyclobutane occurs after coordination of the olefin

monomer. The metallacycle then decomposes in the direction that relieves the ring strain in

the monomer. The relief of ring strain provides the driving force for the reaction. The new

transition metal alkylidene compound can then coordinate more monomer which will form a
metallacycle that decomposes to relieve the ring strain in the monomer, and the reaction

continues by repeating these steps. In well-behaved systems, the catalyst does not

dissociate from the growing polymer chain in a ROMP reaction. Consequently, ROMP is

frequently a chain-growth living polymerization with very narrow molecular weight

distributions and polydispersities near one.6'67 If chain transfer, metathesis of unsaturation

in the polymer backbone, or slow initiation (relative to propagation) occurs, then the

molecular weight distribution and the polydispersity of ROMP polymers increase.



H [M

Figure 1-25. The mechanism for ROMP.

The mechanism for ADMET polymerization is much more involved Figure 1-26.

Figure 1-26 shows the simplest set of steps necessary to produce high molecular weight

polymers by ADMET.68 The first steps of an ADMET polymerization are again monomer
coordination to the metal center and metallacyclobutane formation. Decomposition of the

metallacyclobutane is no longer driven by the relief of ring strain in the monomer.

Consequently, the metallacyclobutane decomposes in productive and non-productive

directions in a statistical distribution.6 The metallacyclobutane decomposition steps in

ADMET polymerizations are equilibrium reactions. When the metallacyclobutane

decomposes in the productive manner (shown by the arrows representing electron flow in
Figure 1-26), a new metal alkylidene and a new olefin, in this case ethylene, is produced.

Removal of this small molecule side product drives the ADMET polymerization cycle. The
new metal alkylidene then reacts with another equivalent of monomer to form a new
metallacyclobutane which decomposes productively to form a terminal methylidene and a
dimer which is no longer attached to the catalyst. The terminal methylidene reacts with
monomer to form another metallacyclobutane, which decomposes productively with loss of
ethylene and formation of another metal alkylidene; this begins the repetition of the
polymerization cycle.

[M]= HR-.R- [q CHR-.-


[M==/ \

[M)== CH2


or polymer

1 x

Figure 1-26. The mechanism for ADMET polymerizations.



//X/ ,

This ADMET polymerization mechanism represents a step-growth polymerization

in which the polymer chain increases in molecular weight very slowly.67 First, monomer is

converted to dimer then to trimer and then tetramer, and so on. In such a regime, very high

conversions are necessary to produce high molecular weight polymers. Additionally, since

ADMET is a step-type polymerization in which the catalyst dissociates from the growing

polymer chain, wider molecular weight distributions and polydispersities close to two are

observed.6'67 Catalysts that will polymerize cyclic olefins have been known for some time.

Even classical catalyst mixtures of transition metal compounds and Lewis acid catalysts

have demonstrated utility for ROMP.5'6,67

The directed search for Lewis-acid-free metathesis catalysts culminated in the

synthesis of Schrock's catalyst,41 the most active Lewis-acid-free, molecular metathesis

catalyst, Figure 1-27. The many synthetic studies aimed at creating metathesis catalysts

have enabled researchers to make many generalizations on the features which are most

critical for a successful olefin metathesis catalyst Schrock's catalyst exemplifies many of

these key features.

Schrock's catalyst is composed of an early transition metal, and in its highest

oxidation state, both the tungsten and molybdenum69 analogs are known. A it-bonding

spectator ligand, in this case an imido, is present and helps to stabilize the formation of the

metallacycle to competing reductive decomposition pathways.70


11 H

H3 CF3 0
H3 /\
Figure 1-27. Schrock's olefin metathesis catalyst.

Terminal oxo ancillary ligands have also been used in some olefmn metathesis

catalysts.3.42,71,72 However, the oxo ligand cannot be substituted and therefore much steric

and electronic variation is precluded for transition metal oxo alkylidenes. The bulky 2,6-

diisopropylarylimido ligand also serves to stabilize the metal center to bimolecular

decomposition pathways.61,73 The alkylidene substituent is also variable since both t-butyl

and neophyl substituents can be synthesized.41'74 The synthesis of these alkylidenes is
achieved by a-hydrogen abstraction from polyalkyl compounds so that less sterically

demanding alkylidene substituents are not synthetically attainable by these routes. Bulky

alkylidene substituents have been shown to have slower rates of metathesis than their less

hindered counterparts.57 However, the bulky substituents block reductive elimination of

the alkylidene and may hinder the formation of alkylidene-bridged dimers.33

The ancillary alkoxide ligands are also important75 Different functions have been

proposed for these ligands including i-donation to stabilize metallacyclobutane formation,

and electron-withdrawing character to enhance olefin coordination rates.76,77 The

hexafluoro-t-butoxide ligands produce more active catalysts than their trifluoro-t-butoxide

or nonafluoro-t-butoxide analogs.78 The fact that hexafluoro-t-butoxide compounds are

better metathesis catalysts than nonafluoro-t-butoxide compounds seems to preclude a

purely electron-withdrawing function for the alkoxides. Contributing steric effects have

been proposed, although the magnitude of the steric differences between these compounds

is not clear. Alternatively, electronic effects which result in conformational changes of the

molecule affecting metallacyclobutane decomposition have also been proposed.75

Successful olefin metathesis catalyst design is then best characterized as the attainment of a

delicate balance between electronic and steric factors which accelerate the desired olefin

metathesis and allow it to compete favorably with non-metathesis reaction pathways.

One of the less attractive characteristics of Schrock's catalyst include a time-

consuming and expensive synthesis which proceeds in moderate yield at best.43 Improved

synthetic routes41,69 have been developed which have somewhat ameliorated this problem,

although the synthesis is still quite sensitive to purity of the reagents and solvents and to

traces of moisture or oxygen, which dramatically reduce the yields. The catalyst compound

itself is extremely sensitive to moisture, oxygen, and Lewis base compounds (like ketones,

aldehydes and amines) with which it reacts in a Wittig-like fashion to produce tungsten oxo

or imido compounds lacking an alkylidene functionality,7981 Figure 1-28. Interestingly,

the molybdenum analog of Schrock's catalyst has a much greater tolerance for

functionalities74'82-84 (though not for moisture or oxygen) than the tungsten catalyst.

Furthermore, the tungsten analog of Schrock's catalyst has shown a tendency to

metathesize internal olefins more rapidly than terminal olefins,74 while this preference is

reversed for the molybdenum analog. The reasons for these intriguing differences in the

observed activity of the analogous tungsten and molybdenum compounds are not well

[M R
Figure 1-28. Wittig-like chemistry of a transition metal akylidene.

There are a number of well-defined, molecular ROMP catalysts as well as several

very useful catalyst mixtures. The trend in ROMP catalyst research has not truly been

toward more active catalysts, but toward catalysts that are increasingly tolerant of a wide

variety of functionalities.1 A recently developed low-valent ruthenium catalyst that shows

unprecedented stability has recently been reported by Grubbs and coworkers,49 and shows

great promise. This ruthenium complex has been found to catalyze the ROMP of

substituted norbornenes in aqueous,85 alcoholic, or protic media.86 The molybdenum

analog of Schrock's catalyst, although intolerant of even trace amounts of moisture, has the

greatest demonstrated tolerance of functionalities and has been used to prepare a wide range

of polymers. The most common ROMP monomer is norbomene, a cyclic olefin with

considerable ring strain that is metathesized to an unsaturated polymer with cyclopentane in

the backbone,76 Figure 1-29. Numerous substituted norbornenes containing esters, ethers,

ketones, nitriles, and halogens have been polymerized using the molydenum analog of

Schrock's catalyst Even ferrocene derivatives have been successfully polymerized. The

range of specialty polymers produced with Schrock's molybdenum catalyst is no less

diverse and includes liquid crystalline polymers,87"90 conducting polyacetylenes,91 redox

active9293 and optically active block copolymers,94'95 and polymer films incorporating

precious metal9699 or semiconductor nanoclusters.'00

catalyst n

Figure 1-29. The ROMP of norbornene.

The only catalysts that have been shown to produce high molecular weight

polymers by ADMET are the tungsten and molybdenum analogs of Schrock's catalyst.

Classical catalyst mixtures suffer from competing vinyl addition chemistry.64'101 Both

ROMP and ADMET polymerization catalysts must be stable to competing decomposition

pathways, at least for the extent of the metathesis reaction. The ROMP reaction is very

rapid, overwhelming many slower decomposition pathways, and has a strong driving force

in the forward direction only, so that each metallacyclobutane formed decomposes in a

productive manner.6 The ADMET polymerization proceeds through a series of equilibria,

and so must form metallacycles many more times than the ROMP catalyst in order to

produce high molecular weight polymer. This effect results in net slower rates for

productive ADMET polymerizations in comparison with ROMP reactions.64 Also, there is

some evidence that the metallacyclobutanes may be longer lived in ADMET polymerizations

than in ROMP. These factors make competing non-productive metallacyclobutane

decomposition reactions more troublesome in ADMET processes. Consideration of the

ADMET cycle reveals the daunting number of intermediates that an ADMET catalyst must

pass through to generate high molecular weight polymer. A successful catalyst must

incorporate all of these structures in a regime in which no one structure is much more stable

than the others. If one intermediate is much more stable than the others, the polymerization

will fail as the catalyst becomes trapped in a resting state configuration from which it cannot

proceed. Given these constraints, Schrock's catalyst and the ADMET synthesis of

polymers are indeed formidable accomplishments.

One advantage of ADMET polymerizations include a wider monomer base.

Organometallic monomers can be envisioned which could be polymerized by ADMET and

would contain transition metals in the polymer backbone.102 Such polymers might have

unusual magnetic or electronic properties. In the short period of time since the original

reports of high molecular weight polyoctenomer were prepared from 1,9-decadiene via

ADMET, Figure 1-30, a considerable number of acyclic diene monomers have been

polymerized by ADMET, including ethers,03-'06 thioethers,107 carboxylic acids,108'109

esters,110-112 carbonates,113 carbosilanes114 and carbosiloxanes.115 Many of the structure

reactivity relationships for ADMET polymerizations have been delineated in these

studies.116 Random copolymers,117 elastomers118 and telechelomers1I8'109,119 have also

been synthesized by ADMET.

n catalyst

Figure 1-30. ADMET polymerization of 1,9-decadiene.

One intriguing consequence of the equilibrium nature of the ADMET polymerization

is that, under ethylene pressure, it is possible to depolymerize unsaturated polymers via
ADMET.120'121 However, Schrock's catalyst, which is the only catalyst that has been

shown to generate high molecular weight polymers by ADMET, has some characterisitcs

that are less than ideal for ADMET chemistry.

Polymerization reactions are often run in neat monomer. The use of monomer as

solvent maximizes the effective concentration of the monomer and so accelerates the

polymerization reaction. The absence of solvents also circumvents a potential source of

contaminants and an extra expense. A negative consequence of the absence of solvent is

that the growing polymer chain can precipitate, rendering efficient stirring of the oligomer

difficult. In effect, the solution becomes too thick to stir and the catalyst is somewhat

isolated in the precipitated oligomer, so the reaction rate slows significantly. Heating of the

reaction mixtures to the melt temperature of the polymer is one strategy that has been used
to overcome this difficulty.67 The thermal instability of Schrock's catalyst above 60C

precludes this solution for ADMET polymerizations. Additionally, ADMET reactions are

often slow by conventional polymerization standards. Part of the reason for the long

reaction times required to generate high molecular weight materials via ADMET is that the

dissociated oligomers can coordinate to the catalyst via an internal double bond instead of a

terminal double bond. The resulting metathesis reaction of the internal double bond

effectively chops up the growing polymer chain, depolymerizing the oligomer rather than

increasing the molecular weight, Figure 1-31.


H 12 -1122
H2 n n ^ CH n H

Figure 1-31. Metathesis of an internal olefin of a diene dimer.

Each of these difficulties presents an impediment to the further development and

commercialization of ADMET polymerizations. An ideal ADMET catalyst can be

envisioned which would allow for the heating of the polymerization reaction to melt

temperatures, thereby allowing efficient stirring of the growing polymer chain and

diffusion of monomer to the catalyst molecules. This improved catalyst would have the

highest possible olefin affinity in order to increase the rate of olefin coordination and

metallacyclobutane formation.d Cationic alkylidenes, for example, would be expected to be

very electrophilic, and to coordinate olefin rapidly. Additionally, the ideal ADMET catalyst

would have enough steric bulk to decrease the tendency for internal olefin metathesis.

In addition to the benefits conferred upon ADMET polymerization studies by such a

stable molecule, the potential for fundamental studies of the nature of the transition metal

-carbon double bond increases with molecules that are stable to more intensive and invasive

analytical techniques. With these criteria in mind, the present study considers the synthesis

of thermally stable, highly electrophilic, sterically encumbered tungsten(6) alkylidenes.

Because of the success of Schrock's catalyst in both ROMP and ADMET chemistry,

Schrock's catalyst served as the model for the functionalities desired in these new catalysts.

Syntheses were begun using intermediates in published syntheses of Schrock's catalyst

Several strategies for increasing the electrophilicity of Schrock's catalyst or intermediates in

the Schrock's catalyst synthesis have been investigated. These synthetic procedures were

designed to yield cationic tungsten(6) alkylidenes in a few steps from published

compounds. Preliminary synthetic studies revealed that the catalyst systems sought were

highly unstable to loss of the alkylidene functionality.e Thus, in the present study,

chelating ligands, which have long been known to stabilize transition metal

compounds,2'19 have been investigated as ancillary ligands for tungsten(6) alkylidene

synthesis. Several chelating ligands were considered, including substituted

d This statement carries an implied assumption that olefin coordination or
metallacyclobutane formation are rate limiting, a statement that had not been adequately
assessed when this study was first proposed. There is some evidence that metallacycles
may be the resting state for these systems, which would imply that metallacyclobutane
decomposition is rate limiting.
e Blosch, L. L.; Boncella, J. M. unpublished results.

acetylacetonates, tris(l,3,5-hydroxy)cyclohexane, and polypyrazolylborates. The greatest

successes were achieved with the polypyrazolylborates. These results are presented herein.

A brief background of polypyrazolylborate synthesis and chemistry will be presented in

chapter 2.

The synthesis of a series of trispyrazolylborate stabilized tungsten(6) alkylidynes

and alkylidenes is presented. These compounds posesses unprecedented thermal,

hydrolytic and oxidative stability for high oxidation state compounds of this type. The

complete characterization of the compounds includes extensive 1H NMR spectroscopy,

13CMR spectroscopy, infrared spectroscopy, high resolution mass spectroscopy,

elemental analysis and three single crystal X-ray studies, including the first single crystal

X-ray study of a cationic tungsten(6) alkylidene which is stable indefinitely to air, moisture

and elevated temperatures even in solution. The moderate to low yields of these materials

is rationalized by a competing addition and decomposition of the trispyrazolylborate ligand

induced by the Lewis acidity of the tungsten metal center. Decomposition products have

been isolated and characterized in support of this assertion.

The conversion of the neutral trispyrazolylborate tungsten(6) alkylidyne dihalides to

neutral trispyrazolylborate tungsten(6) oxo alkylidene halides and trispyrazolylborate

tungsten(6) dioxo alkyls is described and compared for a variety of reagents. Reactivity

studies of the compound Tp'W(O)(CHC(CH3)3)Cl are presented, including the synthesis

of trispyrazolylborate stabilized tungsten oxo neopentylidene hydride, phenylacetylide and

methyl compounds. Since the trispyrazolylborate tungsten(6) alkylidenes synthesized are

all coordinatively and electronically saturated, they are not active metathesis catalysts in the

absence of Lewis acid cocatalysts. Addition of a Lewis acid cocatalyst to these systems

generates a very active ROMP catalyst which operates in untreated air and at elevated

temperatures, producing high molecular weight polyoctenamer or polynorborene within

minutes. The complex nature of the Lewis acid attack on the compounds prevents

identification of the active species in these catalyst systems to date.


Due in part to the stability of these compounds, a kinetic study on the rotational

isomerization of the tungsten-alkylidene double bond was completed. Rotational

isomerization of the tungsten-alkylidene double bond was achieved photochemically and

thermally for some compounds. The results of these studies indicate a bent intramolecular
hydrogen bonding interaction between the alkylidene a-hydrogen and the cis oxo or imido




Polypyrazolylborates are trischelating, monoanionic ligands which were first
reported in the 1960s by Swiatoslaw Trofimenko at the Central Research Department of E.

I. du Pont.12'23 Since the initial reports of these chelating ligands, the chemistry of

polypyrazolylborate ligands has been extensively developed, encompassing a vast number
of ligands with different electronic and steric properties. Numerous reviews have appeared

in the literature covering the synthesis of a variety of polypyrazolylborate-stabilized

metalloid and transition metal-containing compounds,124-127 including a very recent review

which lists more than 450 papers published through early 1993.128 Nearly every metal or

metalloid in the periodic table has been completed to one or more of the poly-

pyrazolylborate ligands. Polypyrazolylborate ligands have been used extensively with

group 6 transition metals, especially molybdenum and tungsten.128 However, most of the

compounds reported are in low to intermediate oxidation states. Many polypyrazolylborate

ligands have been shown to impart remarkable thermal, oxidative and hydrolytic stability to

reactive organometallic compounds.27 For this reason and the reasons outlined in the

previous chapter of this dissertation, the synthesis of polypyrazolylborate alkylidyne and

alkylidene compounds was undertaken. No high oxidation state polypyrazolylborate

alkylidyne or alkylidene compounds had appeared in the literature prior to this study. A

very brief review of the synthesis of polypyrazolylborate ligands and the types of chelate-

stabilized transition metal compounds that appear in the literature most frequently will be

given. Some of the most common properties of the various polypyrazolylborate ligands

will be presented in order to reveal the rationale behind the choice of polypyrazolylborate

ligands for this study.

Polypyrazolylborate ligands are readily synthesized from borohydride salts and

excess pyrazole which are mixed together as solids and heated to melt temperature under an

inert atmosphere.12 The excess melted pyrazole serves as the solvent. Substitution of

hydrides occurs in a stepwise manner with concomitant evolution of hydrogen gas, Figure

2-1. The degree of hydride substitution is controlled thermally.


xs N1 B
KBH4 Y (H4-n + n/2 H2

Figure 2-1. Polypyrazolylborate synthesis.

Pyrazolylborate ligands containing substituents on the pyrazole carbons are

prepared in the same manner using an appropriately substituted pyrazole.125 The

substituents on the pyrazole moieties can be varied extensively. Small substituents, like

methyl groups, must be symmetrically substituted on the pyrazole (i.e, 3,5-disubstituted as

shown by the Y substituents in Figure 2-1) in order to produce homogeneously substituted

polypyrazolylborate ligands. Polypyrazolylborates with extremely bulky substituents (e.g,
t-butyl, isopropyl, and phenyl) on the pyrazole groups have also been reported

recently.129133 While the bulky substituted polypyrazolylborates have found application

recently in the synthesis of unusually low coordination number transition metal

compounds, bulky substituents on the pyrazole rings tend to block access to the transition

metal very efficiently.128 The most commonly reported transition metal

polypyrazolylborate compounds contain the parent ligands, poly(3,5-dimethyl-l-

pyrazolyl)borate and poly(1-pyrazolyl)borate, which are derived from 3,5-dimethyl

substituted pyrazole and from unsubstituted pyrazole, respectively.

In addition to substitution of the carbons of the pyrazole rings, the number of

pyrazole rings attached to boron is also readily varied. The bis- (n = 2), tris- (n = 3) and

tetrakis- (n = 4) polypyrazolylborates have all been synthesized and used extensively as

ligands with transition metals and with some alkali earth metals. '1225 Each of the ligands

is monoanionic. As noted above, tetrakispyrazolylborates are attainable only with

unsubstituted pyrazoles. Tetrakispyrazolylborates have been observed to be

triscoordinating with one uncoordinated pyrazolide moiety or bischelating with two

uncoordinated pyrazolide moieties.127 The number of coordinated pyrazoles for the

tetrakispyrazolylborate ligands is controlled by the size of the metal center and the steric

demands of the other ligands, with larger metals and bulky ligands giving rise to tris-

coordinated tetrakispyrazolylborates and smaller metals with less sterically demanding

ligands giving rise to bis-coordinated tetrakispyrazolylborates. Tris-coordinated

tetrakispyrazolylborates are often fluctional, with exchange of the uncoordinated pyrazolide

and the transition metal coordinated pyrazolide moieties occurring by a variety of

mechanisms.134 Some bis-coordinated tetrakispyrazolylborate systems have been observed

to bridge two metal centers,127,128 Figure 2-2.


S N NNN )'

Figure 2-2. Tetrakispyrazolylborate bridging two metal centers, bis-coordinated to
each metal.

Bispyrazolylborate stabilized compounds have also been reported.

Bispyrazolylborates with two hydrides remaining on the boron have been demonstrated to

undergo hydride transfer reactions.124 This tendency can be overcome by using a

bisalkylborohydride salt as the starting material for ligand synthesis so that a bischelating

bispyrazolyl-bisalkylborate ligand is obtained,12 Figure 2-3. These bispyrazolyl-

bisalkylborate ligands are more difficult to prepare and to purify than the bispyrazolyl-

dihydrideborates. Transition metal compounds with either bispyrazolyl-bisalkylborate or

bispyrazolyl-dihydrideborate ligands are also more difficult to prepare and are less stable

than other transition metal polypyrazolylborate compounds.124,125

SNC6J (R( N- )
KR2BH2 + xs+ H2
ND 2

Figure 2-3. Dialkyl bispyrazolylborate synthesis.

Of all of the polypyrazolylborate ligands, the trispyrazolylborate ligands have been

used most extensively with transition metals.128 Trispyrazolylborate ligands have been

frequently compared to cyclopentadienyl ligands. Both are five electron, monoanionic

donors which take up three coordination sites at a metal center. However, cyclopentadienyl

ligands are x-bonding, electron withdrawing ligands while trispyrazolylborate ligands are
o-donating ligands. Frequently, trispyrazolylborate transition metal compounds are more

stable than their cyclopentadienyl analogs.135'136

The popularity of the trispyrazolylborate ligands is based on the ease of synthesis

and preparation of both the ligands and the transition metal compounds containing the

trispyrazolylborate ligands.126'28 Further, transition metal trispyrazolylborate compounds

generally do not exhibit the undesirable complications associated with the other transition

metal polypyrazolylborate compounds. The remaining hydride on the boron in the

trispyrazolylborate ligand has not been observed to be reactive as a hydride source.a

Fluxional processes involving the trispyrazolylborate ligands on transition metal

compounds are rare. Additionally, trispyrazolylborate containing transition metal

compounds tend to be extraordinarily stable to elevated temperature, oxidation and

hydrolysis, far in excess of the stability of bispyrazolylborate transition metal compounds.

For these reasons, the trispyrazolylborate ligands were targeted as the polypyrazolylborate

ligands most likely to produce high yields of readily characterized, monomeric,

polypyrazolylborate stabilized high oxidation state tungsten alkylidynes and alkylidenes.
Since access to the metal center would be a requirement for an active metathesis catalyst,

the parent ligands, hydrotris(3,5-dimethyl-l-pyrazolyl)borate (Tp') and hydrotris(l-

pyrazolyl)borate (Tp) would be investigated first.

Synthesis of Neutral Trispyrazolylborate Alkylidyne and Alkylidene Compounds

Synthesis of a Hydrotris(3.5-dimethyl-l-pyrazolyl)borate (Tp') Neopentylidyne

Addition of the trischelating monoanionic ligand hydrotris(3,5-dimethyl- 1-

pyrazolyl)borate124 (Tp') to a cold, stirring solution of W(CC(CH3)3)C13(DME)41 yielded

the Tp' tungsten neopentylidyne dichloride 1,137 Figure 2-4. Compound 1 was

recrystallized from toluene, producing small, fine purple needles in 49% yield. The 1H and
13C NMR spectra of 1 were consistent with the proposed structure, with the alkylidyne ca-

carbon resonance appearing 335 ppm downfield of tetramethylsilane in the region

characteristic for tungsten(6) alkylidynes.46 Remarkably, compound 1 was found to be

completely and indefinitely air and moisture stable in both the solid phase and in solution.

No change in the physical appearance, 1H NMR spectrum, or IR spectrum of compound 1

a Alkyl-trispyrazolylborate ligands are also synthetically attainable. Alkyl substituents on
the boron can be used to vary the steric and electronic characteristics of the
trispyrazolylborate ligands.

could be detected after three months storage of the solid in air or three weeks stirring in air
of a benzene solution.


\ [K] H /
a J0- N NN Vc N



Figure 2-4. Synthesis of Tp'W(CC(CH3)3)C12

No evidence of decomposition was observed (by 1H NMR) when compound 1 was
heated to 2750C in an open capillary tube. Due in part to the extreme, unprecedented
thermal stability of compound 1, a very clean fragmentation pattern and an intense
molecular ion envelope (nominal parent ion mass 620.16 amu) were detected for compound
1 by electron impact high-resolution mass spectroscopy. Although protonation of
alkylidyne compounds has been shown to yield alkylidenes in some cases,45,46 compound
1 proved to be stable to protonation with triflic acid, tetrafluoroboric acid and hydrochloric
acid. In attempting to purify compound 1 on an alumina column, a yellow band was
observed to develop slowly on the column. 1H NMR analysis of the yellow band revealed
resonances consistent with a tungsten(6) alkylidene compound.

Conversion to a Hydrotris(3.5-dimethyl-1-pyrazolvylborate (Tp) Neopentvlidene

In light of the results obtained from column chromatography of compound 1 on an
alumina column, compound 1 was slurried with neutral, activity 1 alumina and was
quantitatively converted to a Tp' tungsten oxo neopentylidene monochloride,137 Figure 2-5.
After filtration from alumina compound 2 was pure (by 1H NMR) and was dried under

reduced pressure, giving a bright yellow, air-stable powder. Like compound 1, compound
2 can be stored indefinitely in air as a solid and heated to 2200C in an open capillary tube

without evidence of decomposition.

\I I I
N N aluminaN N

1 > 2

1 2
Figure 2-5. Alumina mediated conversion of Tp'W(CC(CH3)3)C12 to

Compound 2 also has remarkable thermal stability since it can be heated to reflux in
cyclooctene solution (1450C) overnight in the air with no evidence of decomposition
observed by 1H NMR spectroscopy. The IH NMR a-proton resonance of the alkylidene
was detected at 10.4 ppm (2JHW = 3 Hz), and the 13C NMR alkylidene a-carbon

resonance was found at 304 ppm (1Jcw = 158 Hz). Both resonances were in the expected

region for tungsten(6) alkylidenes.34 All other 1H NMR and 13C NMR resonances were in

agreement with the structure shown. The composition of compound 2 was further verified

by elemental analysis, chemical ionization high-resolution mass spectroscopy (nominal
M+1 603.19 amu), and infrared spectroscopy (vOW = 973 cm-1), as well as by a single-

crystal X-ray diffraction study.137

Slow, room temperature evaporation of a diethyl ether solution of compound 2
yielded small, single, bright yellow crystals. The structure of compound 2 revealed the
expected pseudooctahedral coordination geometry at tungsten as shown in the ORTEP
drawing in Figure 2-6.


Orbital constraints associated with multiple multiply-bonded ligands as well as
geometric constraints imposed by the Tp' ligand require that the oxo, alkylidene and
chloride ligands are mutually cis138.139 with the O-W-C(16) angle 99.8 (4). The high

energy tungsten-oxo stretch observed in the infrared spectrum of compound 2 suggested
considerable triple bond character for this bond.1'40 The WO= bond length of 1.685 (8) A
confirmed the strong x-donating character of the oxo ligand for this compound Both

previously crystallographically characterized six-coordinate tungsten oxo alkylidenes,

W(O)(CHC(CH3)3)C12(P(CH2CH3)3)242,141142 and

W(O)(CHC(CH3)3)C12(P(CH3)Ph2)2,47 have identical formal electron counts at tungsten
and have similar WO= bond lengths of 1.697 and 1.708 A, respectively. Also of interest
was the W=C(16) bond length of 1.949 (8) A and the W-C(16)-C(17) angle of 136.0 (6),

both of which are characteristic of undistortedd tungsten alkylidene linkages and are
comparable with the values reported for W(O)[CHC(CH3)3]C12(P(CH2CH3)3)2 (1.986 A,
142.20) and W(O)(CHC(CH3)3]C12(P(CH3)Ph2)2 (1.980 A, 132.20). The alkylidene was

oriented with the t-butyl group bent slightly away from the oxo ligand presumably to
minimize repulsions between the oxo lone pairs and the t-butyl moiety. The N-W distances
vary from 2.174 (5) A to 2.359 (9) A and are consistent with a decreasing trans influence
of the ligands oxo > alkylidene > chloride.'
Having firmly established the identity of the product of the reaction of

Tp'W(CC(CH3)3)C12 with alumina, investigation of the conversion process was
undertaken. The use of alumina to generate low-valent group 6 trispyrazolylborate oxo
compounds has been previously reported.143 Alumina had not been previously used to
generate a terminal oxo ligand on a compound with a high-valent alkylidene functionality.

b In transition-metal-oxo compounds, if the formal electron count of the metal is less than
18 electrons, the oxo-metal bond length is appropriately short, and a x-symmetry orbital in
the appropriate orientation is available, a triply bonded "WO=" is the preferred valence
bond description of the metal-oxo interaction.
c The electron count at tungsten for compound 2 is 18 electrons if the oxo is counted as a
six electron donor (2-), the alkylidene is counted as a four electron donor (2-), and the Tp'
ligand is counted as a six electron donor (1-).

A series of reaction conditions for the alumina-assisted conversion were employed in an

effort to gain an understanding of the conversion process. Although all types of

commercially available alumina were found to convert compound 1 to compound 2, basic

alumina effected the conversion more than twice as rapidly as acidic alumina. Neutral

activity 1 alumina and adsorption alumina had a rate which was just slightly slower than
basic alumina.

Solvent effects on the rate of conversion were much greater. When diethylether

was used as a solvent, the solution above the alumina bed was much more intensely

colored and the alumina was nearly colorless. When hexanes were used as a solvent, the

solution above the alumina bed was very pale blue and the alumina bed was intensely

colored. Conversion rates were up to five times greater in hexanes than in other solvents.

Further, technical grade hexanes facilitated the conversion significantly, resulting in a

nearly two-fold increase in conversion rate as compared to dried and distilled reagent grade


Other means of affecting the transformation of Tp'W(CC(CH3)3)C12 to

Tp'W(O)(CHC(CH3)3)Cl were also sought The reagents investigated were triethylamine
with one equivalent of water, wet coordinating solvents, and lithium hydroxide. The rate

of converting 50 mg of Tp'W(CC(CH3)3)C12 to Tp'W(O)(CHC(CH3)3)Cl using each of

the investigated reagents is summarized in Table 2-1. Percent conversion was measured by

1H NMR. In all cases, no products other than Tp'W(O)(CHC(CH3)3)C1 were observed.

For reported conversions of less than 100%, the remainder of the mixture was unreacted

Tp'W(CC(CH3)3)C12. Alumina was found to be the most efficient reagent for effecting the

conversion of Tp'W(CC(CH3)3)C12 to Tp'W(O)(CHC(CH3)3)CI. Although the rate of the

conversion on alumina was strongly solvent dependent, the slowest conversion rate using

alumina as the reagent (and diethylether as solvent) was more than 24 times faster than any

other reagent investigated. The fastest conversion rate for an alumina reagent (and technical

grade hexanes as solvent) was more than 120 times faster than the other reagents. One

equivalent of triethylamine and one equivalent of water was the best of the non-alumina

methods of conversion, but only 34% conversion of Tp'W(CC(CH3)3)C12 to

Tp'W(O)(CHC(CH3)3C1 was observed with this reagent after five days. Wet solvents

were also a poor means of effecting the conversion. Only coordinating solvents had any

utility for the conversion. Acetone and acetonitrile had better conversion rates than

diethylether. No reaction was observed with lithium hydroxide after two days. Use of

anhydrous lithium hydroxide to effect this conversion was undoubtedly hindered by the

poor solubility of lithium hydroxide in tetrahydrofuran.

Table 2-1. Percent conversion of compound 1 to compound 2 as a function of reagent and

reagent time % conversion
alumina/tech. hexanes 3 hours 100

alumina/reagent hexanes 7 hours 100

alumina/wet diethylether 15 hours 95

NEt3/H20/THF 5 days 34

wet acetone 5 days 20

wet diethylether 21 days trace

LiOH/THF 2 days 0

Synthesis of a Hydrotris(1-pvrazolvl)borate (TpD Neopentvlidvne

Addition of hydrotris(l-pyrazolyl)borate24 (Tp) to a cold, stirring solution of

W(CC(CH3)3)C13(DME)41 yielded the Tp tungsten alkylidyne dichloride 3, Figure 2-7.
Compound 3 is entirely analogous to compound 1 with the exception that the 3- and 5-

positions of each of the pyrazole rings of the pyrazolylborate ligand are unsubstituted.

Compound 3 was extracted from the crude reaction mixture into toluene and then dried

under reduced pressure, yielding a dark blue solid in 82% yield. The solid was pure by 1H

NMR without further workup.

ov H -B
a o N

Figure 2-7. Synthesis of TpW(CC(CH3)3)C12.

The 1H and 13C NMR spectra of compound 3 were consistent with the proposed
structure, with the alkylidyne a-carbon resonating 338 ppm downfield of tetramethylsilane

in the region characteristic for tungsten(6) alkylidynes.46 Like its Tp' analog compound 1,

compound 3 was found to be completely air and moisture stable in the solid phase. In

solution, compound 3 was slightly less stable than its Tp' analog, converting to Tp

tungsten oxo neopentylidene in moist solvent or in dry solvents exposed to moist air over

several days. The thermal stability of compound 3, although diminished from that of

compound 1, was still quite remarkable when compared to other alkylidyne compounds.

No evidence of decomposition was observed (by 1H NMR) when compound 3 was heated
to 234C in an open capillary tube. A very clean fragmentation pattern and an intense

molecular ion envelope (nominal parent ion mass 536.07 amu) were detected for compound

3 by electron impact high-resolution mass spectroscopy. Compound 3 also proved to be

stable to protonation with triflic acid, tetrafluoroboric acid and hydrochloric acid. Based on

the reactivity of compound 1 with alumina, investigation of the interaction of compound 3

and alumina was undertaken.

Conversion to a Hydrotris(1-pyrazolyvborate (Tp) Neopentvlidene

Compound 3 was slurried with neutral, activity 1 alumina and was converted to a
mixture of two products, a Tp tungsten oxo neopentylidene monochloride
(TpW(O)(CHC(CH3)3)Cl, compound 4) and a Tp tungsten dioxo neopentyl species
(TpW(0)2(CH2C(CH3)3), compound 5) Figure 2-8. Both species are yellow, although
the Tp tungsten dioxo neopentyl species is less intensely colored than the Tp tungsten oxo
neopentylidene chloride compound.


C 4 Cl HH
_\ /I /I
N\ N)N alumina \ N N N \)

,4 N

Figure 2-8. Alumina mediated conversion of TpW(CC(CH3)3)C12 to
TpW(O)(CHC(CH3)3)Cl and TpW(02)(CH2C(CH3)3)

Compound 4 can be stored for weeks in air as a solid, but slowly decomposes to
compound 5 in moist air and decomposes rapidly when heated in an open capillary tube.
Compound 5 can be stored indefinitely in air as a solid and heated to 1200C in an open
capillary tube without evidence of decomposition. In solution, compound 4 decomposes to
compound 5 within days when exposed to moist air.
The 1H NMR a-proton resonance of the neopentylidene of TpW(O)(CHC(CH3)3C1
was detected at 10.7 ppm (2JHW= 4 Hz), and the 13C NMR neopentylidene a-carbon
resonance was found at 303.5 ppm (1Jcw = 158 Hz). Both resonances were in the
expected region for tungsten(6) alkylidenes.34 All other 1H NMR and 13C NMR

resonances of compound 4 were in agreement with the structure shown. The composition

of compound 4 was further verified by elemental analysis and infrared spectroscopy

(vow = 973cm-1).

For TpW(O)2(CH2C(CH3)3), the 1H NMR a-proton resonance of the neopentyl

was detected at 2.1 ppm (2JHw= 8 Hz), in the expected region for tungsten(6) alkyls. All

other 1H NMR and 13C NMR resonances of compound 5 were in agreement with the

structure shown. The composition of compound 5 was also verified by elemental analysis
and infrared spectroscopy (vOW = 960 cm-1 and 917 cm-l).

Since the reaction of compound 3 with alumina was shown to be distinct from the

chemistry of its Tp' analog, compound 1, investigation of the process of converting

compound 3 to compounds 4 and 5 was also undertaken. There were two products of this

conversion on alumina, consequently, product distribution as well as the rate of conversion

was monitored. The reagents investigated were alumina, triethylamine with one equivalent

of water, and wet coordinating solvents. Lithium hydroxide was not investigated as a

reagent for the conversion of compound 3 because of its poor showing in the previous

investigations. The efficiency of the various means of converting compound 3 to

compounds 4 and 5 are summarized in Table 2-2, which lists the rates of converting 50 mg

of compound 3 and the product distribution achieved with each method. The percent

conversion was measured by 1H NMR. In all cases, no products other than compounds 4

and 5 were observed. For reported conversions wherein the sum of percent conversions to

compounds 4 and 5 are less than 100%, the remainder of the mixture was unreacted

compound 3.

Again, alumina was shown to be the fastest method of converting compound 3 to a

mixture of compounds 4 and 5. There was little selectivity (63:37 4:5) with this method if

the reaction of compound 3 on alumina was stopped just after all traces of blue starting

material had been converted to yellow products. Longer reaction times correlated with

greater percent conversion to compound 5 (and less of compound 4), however, compound

5 was difficult to remove from alumina, in part because of its low solubility in general, and

also likely due to strong interactions between the alumina and the two oxo groups of

compound 5. Triethylamine and water effected the conversion of compound 3 48 times

more slowly than alumina, but the Tp tungsten oxo neopentylidene monochloride

compound 4 was the major product of the conversion by this route. Wet solvents were a

very slow means of effecting the conversion. Only strongly coordinating, polar solvents

dissolved compound 3 enough to have reasonable rates of conversion. Although the

solvents used for this conversion were wet, they were thoroughly degassed by bubbling

argon through the solution for 20 minutes. The reactions were then run under an argon

atmosphere. These precautions were taken since compound 4 was known to decompose to

compound 5 in solution in moist air. In THF, 21 days were required to effect 81%

conversion of compound 3 to compound 4 and a trace of compound 5.

Table 2-2. Percent conversion of compound 3 to compounds 4 and 5 as a function of
reagent and time.

reagent time % 4% 5
alumina/tech. hexanes 2 hours 37 63

NEt3/H20/THF 4 days 73 13

wet THF 21 days 81 trace

Synthesis of a Hydrotris(3.5-dimethyl-1-pyrazolvl)borate (Tp') Benzylidvne

Addition of the trischelating monoanionic ligand hydrotris(3,5-dimethyl-1-

pyrazolyl)borate12 (Tp') to a cold, stirring solution of W(CPh)Br3(DME)58 yielded the Tp'

tungsten alkylidyne dibromide 6,144 Figure 2-9. Compound 6 was extracted from the

crude reaction mixture with toluene, washed repeatedly with pentane, then dried under

reduced pressure to yield a dark blue powder in 13% yield. The other tungsten containing

products of the reaction were not identified.

Br NN N 1N

Br r CN \ I N T
Br N NN Br'"Z

Figure 2-9. Synthesis of Tp'W(CPh)Br2.

The 1H and 13C NMR spectra of 6 were consistent with the proposed structure,
with the benzylidyne ct-carbon resonance appearing 327 ppm downfield of

tetramethylsilane in the region characteristic for tungsten(6) alkylidynes.4 Compound 6
was found to be air-stable in the solid phase for months. In contrast to the analogous Tp'
tungsten neopentylidyne compound, 1, compound 6 is slowly converted to the Tp'
tungsten oxo benzylidyne bromide in solution in moist air. Compound 6 is also less
thermally stable than its Tp' neopentylidyne analog, decomposing at 2180C when heated in

an open capillary tube.

Conversion to a Hydrotris(3.5-dimethyl-1-pyrazolyl)borate (Tp') Benzylidene.

Compound 6 was slurried with neutral, activity 1 alumina and was converted to a
mixture of two products, a Tp' tungsten oxo benzylidene monobromide

(Tp'W(O)(CHPh)Br, compound 7) and a Tp' tungsten dioxo benzyl species

(Tp'W(0)2(CH2Ph), compound 8), Figure 2-10. The Tp' oxo benzylidene monobromide
is a dark orange crystalline compound and the Tp' dioxo benzyl is cream-colored solid.
Crystals of compound 8 were grown from dichloromethane, however the crystals
lost solvent slowly when dried. Compound 7 can be stored for months in air as a
crystalline solid, but as a powder, it slowly decomposes to compound 8 over several
weeks in moist air. In solution, compound 7 decomposes to compound 8 within days


N alumina N N
-=N ON- ON- N N--. + N\ N-
Br Br
0BC 05 05-" f
ob .b
)7 8

Figure 2-10. Alumina mediated conversion of Tp'W(CPh)Br2 to Tp'W(0)(CHPh)Br and

when exposed to moist air. Thermal decomposition of compound 7 ensues at 2100C in an
open capillary tube. Compound 8 can be stored indefinitely in air as a solid and heated to
1800C in an open capillary tube without evidence of decomposition.
The 1H NMR a-proton resonance of the neopentylidene of compound 7 was
detected at 11.0 ppm (2JHW < 3 Hz, unobserved), a second alkylidene resonance was
found for compound 7 at 11.4 ppm (2JHW = 12 Hz). The second alkylidene resonance
was attributed to a rotational isomer of the tungsten-benzylidene double bond which results
in the observed coupling constant (2JHW) difference for the two alkylidenes.144 Also,
elemental analysis of the mixture corresponded to that calculated for pure compound 7.
The minor isomer alkylidene 1H NMR resonance was removed by heating the mixture to
600C and was present only in very minor quantities (the maximum observed ratio of major
to minor isomers was 6:1) at room temperature. This behavior is consistent for a system of
two rotamers with a large barrier to interconversion.d The 13C NMR benzylidene a-carbon
resonance for the major isomer was found at 290.2 ppm (1JCW = 151 Hz). The minor
isomer 13C NMR resonance was not observed due to the low concentrations of minor
isomer. These 1H NMR and 13C NMR resonances were in the appropriate regions for

d The rotational isomerization of tungsten (6) alkylidenes will be presented in detail in
chapter four of this dissertation.

tungsten(6) alkylidenes.34 All other 1H NMR and 13C NMR resonances of compound 7

were in agreement with the structure shown. The composition of 7 was further verified by

elemental analysis and electron impact high-resolution mass spectroscopy (nominal parent

ion mass 666.11 amu).
For Tp'W(O)2(CH2Ph), compound 8, the 1H NMR a-proton resonance of the

benzyl was detected at 3.2 ppm (2JHW = 12 Hz), and the 13C NMR neopentylidene a-

carbon resonance was found at 60.10 ppm. Both resonances were in the expected region

for tungsten(6) alkyls. All other 1H NMR and 13C NMR resonances of compound 8 were

in agreement with the structure shown. The composition of 8 was further verified by

elemental analysis, chemical ionization high-resolution mass spectroscopy (nominal M+1 =
605.20 amu), and infrared spectroscopy (vOW = 957 cm-1 and 916 cm1 ).

Other means of affecting the transformation of compound 6 to compounds 7 and 8

were also considered. The reagents investigated were triethylamine with one equivalent of

water, wet coordinating solvents, and lithium hydroxide. As expected from the earlier

studies, all of these methods for converting compound 6 to compounds 7 and 8 were very

slow compared to the reaction on alumina. The efficiency of various means of converting

50 mg of compound 6 to compounds 7 and 8 by each method are summaried in Table 2-3.

The percent conversion was measured by 1H NMR. In some cases, products other than

compounds 7 and 8 were observed, but not identified. For those reactions the percent

conversions given are percent of the identifiable products. Purification of the compounds

was not facile so absolute yields were not calculated.

Again, reaction on alumina was the most rapid method of converting compound 6

to a mixture of compounds 7 and 8, but not an extremely selective method. Triethylamine

and water, and wet THF had the second fastest rate of conversion. Both of these methods

were much slower than alumina, but both had better selectivity with wet tetrahydrofuran

giving a nearly completely selective conversion to the Tp' tungsten oxo benzylidene

monobromide, compound 7. No other products were generated with this reaction.

Triethylamine and water was very selective for the formation of the Tp' tungsten dioxo

benzyl, compound 8, however, unidentified products were also generated. Finely-ground

lithium hydroxide showed a modest conversion to compound 7 and a trace of compound 8

after one day, but unidentified products were also generated by this method of conversion.

Table 2-3. Percent conversion of compound 6 to compounds 7 and 8 as a function of
reagent and time.

reagent time %7 %8
alumina/tech. hexanesa 1.5 hours 25 75

NEt3/H20f/TFa 4 days trace 99

wet THF 4 days 99 trace

LiOHa 24 hours 11 trace

aUnidentified products were observed.

Synthesis of a Stable Cationic Alkylidene

Cationic tungsten(6) alkylidenes were projected to be potentially extremely active

olefin metathesis catalysts since the increased electrophilicity associated with a cationic do

metal center should result in enhanced rates of coordination of olefins to the metal center.

Synthetically, cationic high oxidation state materials are challenging to produce and very

few are well characterized and reported in the literature.4243 Chelating ligands offered the

potential of synthesizing cationic alkylidenes with greater stability than known cationic

materials, and yet potentially enhanced reactivity compared to neutral chelate-stabilized

alkylidene compounds. Synthesis of a cationic tungsten alkylidene compound was

projected from the addition of trispyrazolylborate ligands to the known arylimido

tungsten(6) neophylidene bistriflate compounds, (DME)W(NAr)(CHC(CH3)2Ph)(OTf)2.74

The ensuing reaction, shown in Figure 2-11, was pictured as resulting in the displacement

of both triflate anions and the neutral bischelating DME ligand with the monoanionic

trischelating trispyrazolylborate ligand. The reaction would take advantage of the bulkiness
of the 2,6-diisopropylarylimido ligand, the neophylidene, and the trispyrazolylborate
ligands (especially Tp') to generate not only to generate a cationic alkylidene, but a
coordinatively unsaturated (five-coordinate) alkylidene as well.

OTf = 03SCF3


N N N-


- H [OTf]

N--C, HI
.P H
Ar- I

Figure 2-11. Projected synthesis of a five-coordinate cationic tungsten alkylidene.

Addition of one equivalent of KTp' to a cold stirring THF solution of
(DME)W(NAr)(CHC(CH3)2Ph)(OTf)2 74 gave a low yield of the six-coordinate cationic
Tp' tungsten imido neophylidene pyrazole compound (9),145 Figure 2-12. The pyrazabole
(10) side product was identified by comparison of its 1H and 13C NMR spectra with
literature spectra.46

OTf = S CF
OTf = SO3CF3




Figure 2-12. Synthesis of a cationic Tp' tungsten(6) alkylidene with concomitant Tp'

As stated earlier, proton catalyzed decomposition of pyrazolylborates to pyrazaboles

has been previously reported.124,147 The last step in the synthesis of the starting alkylidene

compound, (DME)(TfO)2W(NAr)(CHC(CH3)2Ph)), for the reaction pictured in Figure 2-

12 involves the loss of one imido group from a bis-imido bis-neophyl tungsten compound,

(NAr)2W(CH2C(CH3)2Ph)2.4 This reaction, shown in Figure 2-13, generates an
equivalent of anilinium triflate salt which is very difficult to remove completely, and which
might be acidic enough to catalyze the decomposition of KTp'.

I /H DME, 3 HOTf / H
O--N= w---C W -.W0-= c
X |-H3NAiOTf C T Xh
Sh -H3 CC(CH3)2)Ph O OTf Ph
H / Ph OTf = 03SCF3
DME = H3CO(CH2)20CH3

Figure 2-13. Synthesis of bis-triflate starting material and anilinium triflate byproduct.

To test this hypothesis a sample of (DME)(TfO)2W(NAr)(CHC(CH3)2Ph)) was
recrystallized repeatedly until it was anilinium-free by 1H NMR. Addition of KTp' to the
anilinium-free (DME)(TfO)2W(NAr)(CHC(CH3)2Ph)) resulted in the same product shown
in Figure 2-12 in the same yield. Performing the reaction shown in Figure 2-12 in the
presence of an excess of other coordinating ligands, such as pyridine or
trimethylphosphine, gave only the cationic Tp' tungsten pyrazole-coordinated product,
compound 9.
Eight high oxidation state, stable, cationic, Lewis-acid-free tungsten alkylidenes
have previously been reported, including one four coordinate oxo neopentylidene dication.
In each case, the cationic alkylidene was generated by halide abstraction from an oxo or
imido neopentylidene dihalide (or halide alkyl) stabilized by phosphine ligands. The
resulting five coordinate cationic oxo alkylidenes42 were all characterized as extremely
sensitive to Lewis base solvents, such as Et20 and THF, as well as thermally unstable
above 250C. The cationic imido neopentylidene analogs43 were more thermally stable
(600C in CDC13) but still extremely sensitive to air and moisture. By contrast, compound

9 was found to be completely stable indefinitely both in the solid state and in solution in
moist air. No 1H NMR evidence of thermal decomposition was observed when
compound 9 was heated as a solid to 180C in air.

The IH NMR spectrum of compound 9 was consistent with the proposed structure
assuming free rotation of the neophylidene phenyl, but no rotation about the C-N bond of

the arylimido ligand at room temperature. The alkylidene proton resonated at 11.34 ppm

and the remaining resonances were consistent with the structure given for compound 9.
High temperature 1H NMR gave no evidence for thermally induced dissociation of the
coordinated pyrazole or for rotation of the arylimido ligand at temperatures up to 1600C. A

minor rotational isomer of the neophylidene of compound 9 was observed to grow into the
spectrum beginning at 600C. At 900C, an equilibrium ratio of 1.9:1 (major rotamer.minor

rotamer) was rapidly established. Coalesence of the alkylidene rotamer signals was not

achieved prior to decomposition, but a rotational barrier of greater than 19 kcal/mol was
calculated. At 1600C, decomposition of both rotamers of compound 9 ensued by reaction

with the solvent, bromobenzene-d5. The 13C NMR spectrum was entirely consistent with
the proposed structure for compound 9, with the alkylidene a-carbon resonance appearing

at 297.8 ppm (1Jcw= 155 Hz, 1JCH= 115 Hz). Compound 9 was further characterized by
elemental analysis and infrared spectroscopy (VNH= 3348 cm-1), as well as by a single

crystal x-ray diffraction study.145

The structure of compound 9 consisted of well separated molecules with

pseudooctahedral coordination geometry at tungsten as shown in the ORTEP drawing in

Figure 2-14. Orbital constraints associated with multiple multiply-bonded ligands as well

as geometric constraints imposed by the Tp' ligand require that the imido, alkylidene and

pyrazole ligands be mutually cis.138'139 No other cationic tungsten(6) alkylidene has been

previously crystallographically characterized. The W=N and W=C bond distances of

1.752(6) and 1.964(9) A, respectively, are well within the normal range for six-coordinate
imido alkylidenes.1 The W-N distances for the Tp' ligand vary from 2.179(6) to 2.311(6)

A and are consistent with a decreasing trans influence of the ligands imido > alkylidene >

pyrazole.1 Although the W-C(21)-C(22) angle of 149.0(6)0 is somewhat large, the 1H-13C

coupling constant (1JCH= 115 Hz) is normal1'3448 and gives no evidence for interaction







between the tungsten center and the alkylidene a-hydrogen. Since the formal electron

count at tungsten is 18 electrons, such an interaction is not expected. Steric interactions of

the arylimido isopropyl groups, Tp' 3-methyl groups, and the neophyl group may account

for the large W-C(21)-C(22) angle.

Attempts to increase the yield of compound 9 by adding a second equivalent of
KTp' resulted in formation of compound 11,145 the neutral pyrazolide analog of compound

10, as shown in Figure 2-15.


[K H N N 4N
+ /N 2 N N
Oh N N _N
OTf = 03SCF3 N-N Ph

Figure 2-15. Synthesis of the neutral analog of compound 9.

Compound 11 was distinguished by the absence of a pyrazole proton in the 1H

NMR spectrum and a N-H stretch in the infrared spectrum, as compared to the spectra of

9. The alkylidene resonance was shifted to 10.79 ppm in the 1H NMR spectrum and to

292.1 ppm (1Jcw= 166.9 Hz, 1JCH= 118.2 Hz) in the 13C NMR spectrum. Like

compound 9, compound 11 exhibited remarkable air, moisture and thermal stability and

was characterized by elemental analysis and infrared spectroscopy.145 The interconversion

of compounds 9 and 11 was readily achieved by addition of n-BuLi to compound 9 and

protonation of compound 11 with triflic acid,145 Figure 2-16. This acid-base chemistry
was clean only when stoichiometric amounts of acid and base were used. In the presence
of excess triflic acid, the neutral compound 11 decomposes to products which do not have

1H NMR resonances in the alkylidene region. A number of resonances are present in the
alkyl region of the 1H NMR spectrum, suggesting that excess acid may protonate the
alkylidene moiety. This inference seems likely since the tungsten-alkylidene bond has been
predicted to be the highest occupied molecular orbital (HOMO) for such compounds.

H + CF3SO3- H

-N N N n-BuLi I N N

AMO.N N C- H 'CF3S03H Ar'-N-H
H N-N +Ph N-.N Ph
Ar..N c-- j /

9 11
Figure 2-16. Acid-base interconversion of compounds 9 and 11.

The analogous Tp chemistry was undertaken by a coworker, S. Gamble. In
contrast to the reactivity observed with KTp', the reaction of equimolar amounts of KTp
and (DME)W(NAr)(CHC(CH3)2Ph)(OTf)2 resulted in high yield (90%) formation of
TpW(CHC(CH3)2Ph)(OTf),145 Figure 2-17. No evidence of decomposition was of Tp
was detected in the crude reaction product


N [KqJ H. N N
ToH / 1 + )x
O-W= C / -DMEI H

I OTf Ph N OTf
S. N
OTf = O3SCF3 Ph Ar
Figure 2-17. Synthesis of TpW(NAr)(CHC(CH2)Ph)(OTf).

Attempted Facile Synthesis of Trispvrazolvlborate Stabilized Tungsten Alkvlidenes

One of the most significant impediments to the study of high oxidation state
alkylidenes has always been the sensitivity and synthetic inaccessibility of these
compounds. Alternate means of generating chelate-stabilized high oxidation state
alkylidenes from readily available or easily synthesized materials was also explored in this
study. Several potential starting materials were readily accessible. The reaction of
equimolar amounts of KTp' and W(O)2C12 results in the high yield synthesis of
Tp'W(0)2C1,144 compound 12, which can be readily alkylated with a variety of Grignard
reagents and organolithium reagents to produce Tp' tungsten dioxo alkyl complexes,44
Figure 2-18.


xr H I1
B, I-

8 R = benzyl
13 R = neopentyl

Figure 2-18. Synthesis of Tp' tungsten dioxo halides and Tp' tungsten dioxo alkyls.

Benzyl (Tp'W(0)2(CH2Ph) compound 8)144 and neopentyl

(Tp'W(0)2(CH2C(CH3)3 compound 13) Tp' tungsten dioxo compounds were synthesized


and compared to the analogous compounds generated in the conversion of
trispyrazolylborate alkylidyne dihalide compounds to trispyrazolylborate tungsten
alkylidene oxo monohalide. The compounds were spectroscopically identical whether
synthesized from the trispyrazolylborate tungsten alkylidyne dihalides or from Tp' tungsten
dioxo chloride.e
These Tp' tungsten dioxo alkyl compounds are, at least superficially, very similar
to some intriguing pentamethylcyclopentadienyl (Cp*) tungsten dioxo alkyls extensively
studied by Legzdins and coworkers.49-151 The Cp* tungsten dioxo alkyls have been
shown to be readily converted to Cp* tungsten oxo alkylidene alkyls by a reaction sequence
involving the addition of hydrochloric acid, PC15 or TMSC1 followed by two equivalents of
Grignard reagent,149,151 Figure 2-19.

or PC15 RMgCl -H2R
< or PCl5 rR

Re O R\ R R R H

Figure 2-19. Halogenation of Cp* tungsten dioxo alkyl.

When the same reaction sequence was performed on the Tp' tungsten dioxo alkyls
or on Tp' tungsten dioxo chloride, no reactivity with any of the halogenating reagents was
observed even when six equivalents of halogenating reagent were employed. The lack of
reactivity for the trispyrazolylborate compounds is somewhat surprising given the reactivity
of the Cp* compounds. There are, however, numerous differences between
cyclopentadienyl and trispyrazolylborate ligands. Sterically, the trispyrazolylborates tend
to be much more demanding than all but the most bulky substituted cyclopentadienyl
ligands. Additionally, trispyrazolylborate ligands are considerably more donating than

e Since these compounds were first synthesized in this laboratory, two independent reports
of the synthesis of tungsten and molybdenum trispyrazolylborate dioxo alkyls have
appeared in the literature.152,153

cyclopentadienyl ligands, which tend to be withdrawing. Further cyclopentadienyl ligands

bond to the metal center through t-symmetry orbitals, while the trispyrazolylborates
interact with transition metals through o-symmetry orbitals. At this point, it is unclear

whether the observed lack of reactivity is a kinetic or a thermodynamic phenomenon, or

some combination of steric and electronic effects.

Attempts to remove a terminal oxo ligand from Tp' tungsten dioxo chloride or Tp'

tungsten dioxo alkyl compounds with excess Grignard, benzoin, trialkylphosphine,

halotrialkylphosphonium halides or phosphoniumylides also failed, Figure 2-20. In the

case of excess Grignard addition (left, Figure 2-20) the coupled organic products of the

Grignard reagents were observed with methyl magnesium chloride and neopentyl

magnesium chloride, however, unreacted Tp' tungsten dioxo chloride or Tp' tungsten

dioxo alkyl starting material was recovered. Benzoin (up, Figure 2-20), trimethylphosine

(lower left, Figure 2-20) and halotriphenylphosphonium halide (lower right, Figure 2-20)

all exhibited no evidence of reactivity with the Tp' tungsten dioxo chloride or Tp' tungsten

dioxo alkyl compounds, even at elevated temperatures. The benzoin and trialkylphosphine

reactions would result in reduction of the metal by two electrons (tungsten(6) to

tungsten(4)) if the predicted oxo abstraction reactions were to occur. Consequently, these

reaction pathways are not optimal for the purpose of generating dO tungsten alkylidenes,

since the tungsten(4) species would then have to be reoxidized to tungsten(6) as the

alkylidene functionality is introduced.

The halotrialkylphosphonium halide reaction has been demonstrated to effectively

remove a terminal oxo ligand from similar tungsten species and replace the oxo with two

halides (tungsten remains in the +6 oxidation state), however, a seven-coordinate Tp'

compound would be generated by this route. Seven-coordinate trispyrazolylborate

compounds are rare, the vast majority of transition metal trispyrazolylborate (Tp and Tp')

compounds reported in the literature are six-coordinate129



-Ns N N






Me3POM'1* R

0 R'

R = C1, Bz, Np



I I 1

0 X

Figure 2-20. Unsuccessful oxo atom transfer reactions from Tp'WO2C1.



The phosphoniumylide reaction (right, Figure 2-20) resulted in a rapid color and

solubility change indicating that a reaction did take place, however, an alkylidene proton

could not be found spectroscopically. Similar reactions have been found to yield

trialkylphosphine stabilized alkylidenes,154156 Figure 2-21. Preliminary spectroscopic data

suggest that such a product might be formed from the reaction of Tp' tungsten dioxo alkyl

and Ph3P=HBz or Ph3P=CHC(CH3)3, however, these products were not pursued.


Figure 2-21. A trialkylphosphine stabilized alkylidene.

The analogous molybdenum compound, Tp'Mo(O)2C1 was also of interest, since

oxygen atom transfer has already been reported for that compound,157 Figure 2-22.

Despite the oxygen atom transfer reported, when Tp'Mo(0)2C1 was synthesized by the

published procedure and exposed to phosphoniumylides no reactivity was observed.

Undoubtedly, the accessibility of the +5 oxidation state for molybdenum is a substantial

driving force for the reaction shown in Figure 2-22. Formation of a strong molybdenum

chloride bond would also help drive the reaction. Since the phosphoniumylide reaction

would produce molybdenum(6) species instead of molybdenum(5) species, and a relatively

weak molybdenum-carbon double bond, the driving force for this reaction is presumably

significantly diminished to the point where no reactivity is observed.

Another potential route into chelate-stabilized tungsten alkylidenes is the synthesis

of Tp' tungsten(4) trihalides which might then be oxidized to tungsten(6) materials. Direct

synthesis of TpWCl3 was undertaken by addition of KTp' to WCl4 with little success due

to the poor solubility of the WCl4. Alternatively, the anionic Tp' tungsten compound

[NEt4][TpW(CO)3] is well known and can be oxidized using thionyl chloride,143.158

Figure 2-23. The resulting Tp' tungsten trichloride is an air stable, paramagnetic

tungsten(6) species with C3v symmetry.


Figure 2-22.


11 /

oc# II -co

Oxo atom transfer from Tp'MoO2CI.

00-- -

14 M = W

15 M = Mo
Figure 2-23. Oxidation of TpW(CO)3 anion with SOC12.

Although the syntheses of Tp'WCl3, compound 15, and Tp'MoCl3, compound

16, have been previously reported, the communication does not present exact experimental

details or extensive characterization. And, although the NMR (no resonances were detected

in the 1H NMR spectrum oc compound 16) and infrared parameters measured for the

compounds are consistent with the proposed products, there were no reported spectra with

which to verify the authenticity of the products of these reactions. Consequently, the purity

of the materials was difficult to assess. Elemental analyses were not satisfactory and the

baseline of the NMR spectrum of compound 15 exhibited considerable roll. Rolling

baselines in the NMR spectra of paramagnetic compounds are frequently observed.

Several attempts to oxidize TpWC13 to a Tp' tungsten oxo trichloride using alumina,

iodobenzene, and N-methylmorpholine-N-oxide have met with little apparent success. The

resulting compound would be seven coordinate and might be fluctional on the NMR

timescale, however attempts to cool the green species produced in each of the oxidation

attempts to a temperature at which the 1H NMR would be indicative of the desired

compound have not succeeded. As stated earlier, there are few seven-coordinate

trispyrazolylborate compounds reported in the literature,128 especially with ligands as

sterically demanding as three chlorides and an terminal oxo. Kinetic inaccessibility of the

metal center may be a significant barrier to reactivity in these compounds. Direct synthesis

of TpW(O)C13 by addition of KTp' to W(O)Cl4 also produced a green material which

remains uncharacterized to date.



The most challenging and critical aspect of any synthetic study is the thorough and

conscientious characterization of the compounds that have been generated. No single

technique can unambiguously characterize the products of a reaction completely, however,

one of the most useful and practical tools for the characterization of organometallic

compounds is NMR spectroscopy. A great deal of information concerning the symmetry,

connectivity, and rigidity of the compounds as well as the identity and hybridization of the

functional groups within the compounds can be extracted from the numbers of resonances

observed, their chemical shifts, relative integrated areas, multiplicities and coupling

constants to other spin active nuclei.159.160 Because of the ease and relatively low cost of

performing most NMR experiments and the large amount of useful data generated, NMR

was the first analysis performed on the compounds generated in this study.

The presence of a trispyrazolylborate ligand in the compounds synthesized was

easily ascertained by 1H NMR spectroscopy since coordination of a trispyrazolylborate

ligand to tungsten results in a number of NMR features which are quite distinct161 These

features will be briefly discussed. In the interest of clarity in the discussion to follow, the

numbering scheme for a pyrazole residue in a trispyrazolylborate ligand is described in

Figure 2-24. For all of the Tp' tungsten(6) compounds synthesized, the singlet resonances

for the 3,5-methyls of the pyrazole rings of the trispyrazolylborate ligand are found in the

region from 1.8 to 3.2 ppm in 1H NMR spectra.

[B I 5


Figure 2-24. Numbering scheme for a pyrazole residue in a trispyrazolylborate
containing tungsten compound.

The resonances for the methyl groups at the 3-position of the pyrazole rings are

generally found downfield of the 5-position methyl resonances and are more dispersed

since the 3-methyls are closer to the metal center and experience a more diverse magnetic

environment. The Tp 1H NMR resonances are generally found in the region from 7.1 to

9.0 ppm. Fine splitting occurs frequently in these resonances due to coupling of the
aromatic ring protons. For both Tp' and Tp compounds, the 4-position proton resonances

of the pyrazole rings are found in the region from 5.0 to 7.0 ppm. Because boron is a
quadrapolar nucleus and trispyrazolylborate compounds do not have tetrahedral symmetry

at the boron, the resonance for the hydride directly attached to boron is frequently not

observed due to extreme broadening.159 In concentrated, extremely clean 1H NMR

samples of these trispyrazolylborate tungsten(6) compounds, a very broad resonance can

sometimes be observed in the baseline of the spectra in the region from 3.5 to 4.5 ppm.

This resonance integrates to one proton and is attributed to the borohydride.

Upon addition of potassium trispyrazolylborate to the tungsten alkylidyne trihalide

compounds described, the expected pyrazolylborate resonances are immediately apparent

and the alkylidyne resonance is displaced to a new chemical shift. For the tungsten
alkylidyne compounds, the a-carbon resonance of the alkylidyne is expected in the region

from 200 to 400 ppm46 which is generally free of other resonances, but overlaps with the
tungsten(6) alkylidene a-carbon range of 220 to 320 ppm.34 Alkylidynes and alkylidenes

are easily distinguished from one another on the basis of 1H NMR spectra, in which
alkylidenes have a-proton resonances in the region from 9.0 to 14.0 ppm.1,34

Additionally, the 1Jcw coupling constant can be used to distinguish between alkylidynes
and alkylidenes. The diagnostic value of coupling constants of spin active nuclei to

tungsten will be addressed separately, later in this chapter. For the trispyrazolylborate

tungsten alkylidyne dihalide compounds synthesized, 13C NMR resonances were observed

in the expected region, as shown in Table 2-4, and no 1H NMR resonances were observed

for these compounds in the alkylidene region.

Table 2-4. 13C NMR resonances for Tp and Tp' tungsten alkylidyne dihalides.

Compound 8 alkylidyne a-carbon JCW (Hz)
Tp'W(CC(CH3)3)C12 335 212
TpW(CC(CH3)3)C12 338 211
Tp'W(CPh)Br2 327

It is immediately apparent from the number of pyrazole resonances and the 2:1

pattern of those resonances in the 1H and 13C NMR spectra that a plane of symmetry

equates two of the three pyrazole rings of the trispyrazolylborate ligands of these alkylidyne

compounds. This plane of symmetry would be expected in a transition metal alkylidyne
dihalide compound. It would contain the alkylidyne ligand and bisect the angle between the

two halide ligands, Figure 2-25, resulting in a molecule with C2v symmetry.

Figure 2-25 shows the generalized, pseudo-octahedral structure of the
trispyrazolylborate alkylidyne dihalide compounds and a Newman projection of these
compounds looking down the vector which passes through the center of the face of the
octahedron defined by the two halides and the alkylidyne a-carbon atom. In both of the

representations shown, the plane of symmetry in the molecule is perpendicular to the page
and parallel to its long axis.


x Cx C
Figure 2-25. Valence bond representation and Newman projection showing the pseudo-
octahedral coordination geometry and the symmetry of the trispyrazolylborate tungsten
alkylidyne dihalides.

The conversion of trispyrazolylborate tungsten alkylidyne dihalide compounds to
trispyrazolylborate tungsten alkylidene monohalide compounds was signalled
spectroscopically by the appearance of a new 1H and 13C NMR resonance in the previously
defined expected regions, Table 2-5, and by the loss of the plane of symmetry which had
equated two of the three pyrazole rings of the trispyrazoleborate ligand.

Table 2-5. 1H NMR resonances for Tp and Tp' tungsten alkylidenes.
Compound gH C 2JHW 1JCH 1JCW
(ppm) (ppm) (Hz) (Hz) (Hz)
Tp'W(O)(CHC(CH3)3)Cl 10.41 304 3 123 158
TpW(O)(CHC(CH3)3)CI 10.67 310 4 160
Tp'W(O)(CPh)Br 10.97 290 151
Tp'W(NAr)C(HC(CH3)2Ph)(Pyr) 11.34 298 5 115 155
[Tp'W(NAr)(CHC(CH3)2Ph)(PyrH)]+ 10.79 293 118 157

The 1:1:1 pattern observed for each of the 3, 4, and 5- positions of the pyrazole

resonances of the trispyrazolylborate ligand reveals that a chiral metal center has been

generated in these compounds, Figure 2-26. The displacement of one of the two halides of

the trispyrazolylborate alkylidyne dihalides accounts for the loss of symmetry. Figure 2-26

shows the generalized, pseudo-octahedral structure of the trispyrazolylborate alkylidene

halide compounds and a Newman projection of these compounds looking down the vector

which passes through the center of the face of the octahedron defined by the halide, the

terminal oxo, and the alkylidene a-carbon atom.





Figure 2-26. Valence bond representation and Newman projection showing the pseudo-
octahedral coordination geometry and the symmetry of the trispyrazolylborate tungsten oxo
alkylidene monohalides.

The 1JCH coupling constant of alkylidene a-proton and carbon atoms is another

informative 1H NMR parameter. Although similar to the values expected for organic sp2

hybridized compounds, the 1JCH coupling constants for alkylidenes tend to be consistently

lower than predicted from organic compounds. There is an excellent correlation between

the 1JCH coupling constant for transition metal alkylidenes and the metal-carbona-carbonp

angle.1 As the metal-carbona-carbonp angle increases, the 1JCH coupling constant

decreases. This correlation can be rationalized as follows: increasing the metal-carbona-

carbonp angle results in a increase in the s-character of the hybrid-orbitals for the tungsten-

carbona and carbona-carbonp bonding (going toward a linear sp hybrid) while the

carbona-proton s-character decreases (going toward a bond to the a-carbon p-orbital). The

increase in p-character of the a-carbon-hydrogen bond results in the lowered 1JCH

coupling constant The alkylidene metal-carbona-carbonp angle is sensitive to both steric

and electronic factors.1 The smallest alkylidene metal-carbona-carbonp angle reported is
1270.1 Larger metal-carbona-carbonp angles are found for more sterically encumbered

molecules and for molecules with many very strongly withdrawing ancillary ligands.

Strongly withdrawing ancillary ligands decrease the electron density at the metal center,

which is compensated for by an increase in electron donation from the tungsten-carbon

alkylidene bond (again going toward an sp hybridized bond with some triple bond

Also, since 183W is 14% abundant (with I = 1/2), one and two bond coupling

constants of spin active nuclei can often be observed. These coupling constants provide a

great deal of information concerning the bonding interactions and hybridization of the
atoms. The IJCW coupling constant of a tungsten(6) alkylidyne ranges from 200 to 225 Hz

and from 145 to 160 Hz for a tungsten(6) alkylidene.1'346 Within a similar series of

compounds, when other possible effects have been carefully excluded, the magnitude of

the 1JCW coupling constant can be taken as a indication of the relative strength of the

tungsten-alkylidyne or alkylidene bond.
The 1JCH and 2JHW coupling constants of an alkylidene can also be very

instructive. The 1JCH coupling constant has been proposed as a means of distinguishing

which of the two possible alkylidene rotamers is present in solution.57,74,78 Additionally,
the 1JCH and 2JHW coupling constants and the alkylidene a-proton chemical shift are a

reliable non-crystallographic means of detecting a metal-alkylidene-a-hydrogen agostic

interaction.1 For compounds in which an agostic interaction occurs, the alkylidene a-

hydrogen (or the alkylidene carbon-hydrogen bond) donates electron density to an empty
orbital at the metal center,34" Figure 2-27. Such an interaction brings the alkylidene a-

proton in close to the metal center where it is significantly shielded (compared to non-

agostic alkylidenes which are significantly deshielded).1 Additionally, the overlap of the
metal and alkylidene orbitals results in an increased 2JHW coupling constant and a
decreased 1JCH coupling constant commensurate with a metal-hydrogena interaction in
which the alkylidene a-hydrogen acts as a donor and the metal acts as an acceptor.34,54
The chemical shifts and coupling constants for the trispyrazolylborate tungsten alkylidene
monohalide compounds were collected in Table 2-5 and provided no evidence for such a
interaction. Further, there was no evidence for an agostic interaction (W-carbona-carbonp
< 150') in the single crystal x-ray diffraction studies of Tp'W(O)(CHC(CH3)3Cl or
[TpW(NAr)(CHC(CH3)3Ph)(PyrH)][SO3CF3]. Such an interaction would not be
expected in these compounds since they are both coordinatively (six-coordinate) and
electronically (18 electron) saturated.



Figure 2-27. An agostic interaction between the alkylidene a-H and the tungsten (d0) metal

In addition to the NMR evidence for the trispyrazolylborate tungsten compounds
synthesized, infrared spectroscopy can provide useful information on the nature of the
compounds synthesized. The trispyrazolylborate ligand has a number of characteristic
infrared active stretches.161 Table 2-6 lists the most readily distinguished infrared active
stretches associated with the trispyrazolylborate tungsten compounds synthesized.

Table 2-6. IR stretches of some of the trispyrazolylborate moieties of some Tp and Tp'
tungsten(6) compounds.

Compound VB-H VC-Hpyr VC=Npyr
(cm-1) (cm-1) (cm-1)

Tp'W(CC(CH3)3)Cl2 2527 3148 1547
Tp'W(O)(CHC(CH3)3)CI 2543 3149 1543
TpW(CC(CH3)3)C12 2519 3112 1504
TpW(O)(CHC(CH3)3)CI 2496 3113 1503
TpW(0)2(CH2C(CH3)3) 2498 3115 1504
Tp'W(CPh)Br2 2512 3106 1502

Tp'W(0)2(CH2Ph) 2550 3159 1543
Tp'W(NAr)(CHC(CH3)2Ph)(Pyr) 2577 3157 1542
[Tp'W(NAr)(CHC(CH3)2Ph)(PyrH)]+ 2552 3158 1546
Tp'W(O)2C1 2555 3157 1543

Tp'W(0)2(CH2C(CH3)3) 2545 3154 1544
Tp'WC13 2563 3152 1542

The non-trispyrazolylborate infrared stretches of the tungsten alkylidyne, alkylidene
and alkyl species can be extremely informative as well. The terminal oxo stretch for
tungsten(6) compounds is both intense, due to the large change in bond dipole associated
with vibrations about the tungsten-oxo bond, and sharp, since the stretching vibrations of
terminal oxos are not strongly coupled to other stretches. Additionally, the terminal oxo
stretch generally appears in the region from 900 to 1100 cm-1 which is frequently not
crowded with the stretching modes of other ligands.40 Because the terminal oxo stretch is
not coupled to other stretches, it can frequently be diagnostic for the presence of a terminal
oxo and can be a valuable gauge of the tungsten-oxo bond strength. The oxo stretches of
the trispyrazolylborate tungsten alkylidene oxo halides and trispyrazolylborate tungsten

dioxo alkyls are collected in Table 2-7. Each of the stretches in Table 2-7 confirmed the
presence of terminal oxos on the compounds which was suggested by the reduced
symmetry revealed in the NMR spectra. Further, the high frequency of the terminal oxo
stretches suggested a strong x-donating interaction between the terminal oxo ligands and
the metal center, suggesting a tungsten-oxo triple bond valence.1'140

Table 2-7. IR stretches for some mono- and dioxo Tp and Tp' tungsten(6) compounds.

compound V(w-o) V(W-o)
(cm'1) (cm-')
Tp'W(O)(CHC(CH3)3)Cl 973

TpW(O)(CHC(CH3)3)Cl 982
TpW(0)2(CH2C(CH3)3) 960 917
TpW(0)2(CH2Ph) 958 916
TpW(O)2CI 950 908

TpVW(0)2(CH2C(CH3)3) 959 914

Although the infrared stretches of the terminal oxos provide a wealth of information

about the identity of the ligands and the strength of the tungsten-oxo interaction, the
assignment of other infrared stretches (e.g, organoimido, tungsten-alkylidyne, tungsten-
alkylidene and alkylidene-proton) is problematic at best. There has been considerable

debate in the literature as to the assignment and significance of organoimido stretches.1
Frequently, these stretches and the tungsten-alkylidyne, tungsten-alkylidene and
alkylidene-proton are strongly coupled to other stretches (e.g, carbon-nitrogen, alkylidene
R group, organoimido R group etc.) and are proposed to occur in regions where infrared
active bands are abundant The alkylidene a-hydrogen stretch is often obscured by other

CH stretches.1 There are no high oxidation state alkylidyne stretches reported in the
literature. For the trispyrazolylborate compounds synthesized, the infrared bands of the

ligand complicate the spectra so that definitive assignment of these stretches was not
possible without resorting to labelling experiments. Although the infrared stretches of
these functionalities would be of interest, the ambiguity of their assignment in the infrared
spectra was not critical, since these functionalities are all NMR active unlike the major
isotope of oxygen. Since the synthesis of the trispyrazolylborate compounds was not
achieved in high yield or with great control, labelling experiments were highly impractical.

Only two imido compounds were synthesized in this study, the cationic compound 9 and
its neutral analog, compound 11. For the reasons described above, assignment of the
tungsten-imido stretch was not possible, however, the pyrazole proton stretch was
unambiguously identified at 3348 cm-1 for the cationic compound 9.

Synthesis of High Oxidation State Trispvrazolylborate Compounds

Addition of the potassium salts of hydrotris(3,5-dimethyl-l-pyrazolyl)borate (Tp')
and hydrotris(1-pyrazolyl)borate (Tp) to tungsten alkylidyne trihalide compounds resulted
in formation of trispyrazolylborate stabilized alkylidyne compounds. The yields of the Tp'
addition reactions ranged from 50% for Tp'W(CC(CH3)3)C12 to 13% for Tp'W(CPh)Br2.
Because of the low yields obtained in these reactions, the fate of the tungsten unaccounted
for as trispyrazolylborate compounds was of interest. For the reaction of KTp' with

(DME)W(CC(CH3)3)13, further analysis of the crude reaction mixture revealed a second
tungsten containing product, a dimeric neopentylidyne tungsten(6) monoanion, [(HB)2(g-

N2C5H7)3][(((CH3)CC)WC12)2(j-Cl)2(.-N2C5H7)] 16 depicted in Figure 2-28.
The dimeric salt was identified by its 1H and 13C NMR (8 alkylidyne a-carbon 336

(1JCW = 223 Hz, a-alkylidyne carbon)) spectra. The proposed structure of compound 16
was confirmed by a single crystal x-ray study as shown in the ORTEP drawing, Figure 2-
29. This material accounted for up to 35% of the yield (based on

(DME)W(CC(CH3)3)C13)). 1H NMR evidence suggests that the other substitutional
isomers of compound 16, [(HB)2(L-N2CSH7)3I[(((CH3)CC)WC12)2(I-Cl)(g-N2C5H7)2]

and [(HB)2(-N2C5H7)3][(((CH3)CC)WCl2)2(P-N2C5H7)3], may also be generated in

low yield.


N NN-N /

N N Cl Cl Cl

Figure 2-28. Monoanionic tungsten(6) alkylidyne dimer produced as a byproduct of the
reaction of KTp' and (DME)W(CC(CH3)3)C13.

Because of the possibility of generating transition metal pyrazole and pyrazolide

compounds, the KTp' synthesized for these reactions was carefully cleaned and all traces
of uncoordinated pyrazole were sublimed from the crude KTp'. Careful 1H NMR analysis
of the KTp' used in these reactions revealed no trace of pyrazole impurity, consequently,

the only possible source of pyrazolide in the dimeric product was the trispyrazolylborate

ligand itself. Proton catalyzed decomposition of pyrazolylborates to pyrazaboles has been

previously reported.124'147 There is no reasonable source of protons in these reactions
(which are all run in a non-protic solvent, THF), however, the electron decifient (d0)

tungsten centers of the alkylidyne trihalides should be excellent Lewis acids. It seems

reasonable that the metal centers could both decompose and coordinate trispyrazolylborate

ligands in a competitive fashion. Although there is no evidence bearing on the mechanism

of trispyrazolylborate salt addition to transition metals, both concerted and step-wise

schemes can be proposed. Both possibilities may operate in solution with the orientation of

the approaching trispyrazolyl moiety dictating the mechanism of addition. A step-wise
mechanism can easily be pictured as leading to decomposition of the trispyrazolylborate
ligand to a metal-coordinated pyrazolide and a neutral hydrobis(l-pyrazolyl)boron


0 -

fragment, Figure 2-30. The hydrobis(l-pyrazolyl)boron fragment could then react with

another metal center or another hydrobis(l-pyrazolyl)boron fragment Abstraction of

pyrazolide from a hydrobis(l-pyrazolyl)boron fragment followed by coupling of the
hydromono(l-pyrazolyl)boron cation thus generated with a hydrobis(l-pyrazolyl)boron
fragment would then lead to the cationic component of compound 16. The true concerted

attack, in which all three pyrazolyl nitrogens approach the transition metal simultaneously
would seem less likely to result in decomposition of the trispyrazolylborate ligand,

although decomposition to pyrazolides could still be envisioned by this route. Similarly,
step-wise coordination of each of the three pyrazole nitrogens of a trispyrazolylborate
ligand need not necessarily result in decomposition of the ligand.

jXN N\l NN'_e F N '1N
N Cl Cr "
Cl C4 Cl

Figure 2-30. Proposed mechanism for Tp' decomposition.

Although the other products of the reaction of KTp' and (DME)W(CPh)Br3 were
not investigated, it seems likely that the low yield of this reaction is also due to competing

decompostion of the trispyrazolylborate ligand and concomitant formation of pyrazole or

pyrazolide tungsten species. It is unclear whether the drastically reduced yields of this

reaction compared with the neopentylidyne reaction are due to electronic or steric effects.
The tungsten bromide bond is both weaker and more polarizable than the tungsten chloride

bond, rendering substitution reactions more facile. The benzylidyne moiety is less
sterically demanding than the neopentylidyne moiety as well, which could result in
enhanced kinetic access to the metal center. Either effect could conceivably result in the

lower yields observed for the reaction of (DME)W(CPh)(Br)3 with trispyrazolylborate

ligands. It is interesting to note that addition of the less sterically demanding, unsubstituted

Tp ligand to the same metal centers resulted in a much higher yield of Tp alkylidyne

species. Presumably this effect is due to the decrease in steric demands of the

unsubstituted ligand as compared to the 3,5-disubstituted ligand. Less steric crowding as

the ligand approaches the metal center could favor coordination of the trispyrazolylborate
lignad over the competing decomposition pathways. These effects on the synthesis of the

trispyrazolylborate stabilized compounds are also reflected in the stabilities of the materials.

The Tp' compounds were consistently more thermally, hydrolytically, and oxidatively

stable than their Tp analogs. Presumably this enhanced stability for Tp' compounds in

comparison with their Tp analogs is due to better kinetic stabilization of the metal center by

the bulky 3,5-dimethyl substituted Tp' ligand.

Bispyrazolylborates were also very interesting since they might result in reduced

coordination numbers at the metal center while retaining the same oxidation state as the

trispyrazolylborate compounds. Preliminary experiments by coworkers with bispyraolyl-

bisalkylborate compounds resulted in the expected complications. Namely a low yield of

the ligand synthesis and difficulties in purifying the ligand. Further, the transition metal

bispyraolyl-bisalkylborate compounds obtained were difficult to characterize. They
exhibited sensitivity to air and moisture and were not thermally stable above 500C.

Additionally, more hindered trispyrazolylborates were considered. Addition of the

very bulky ligands potassium tris(3-phenyl-l-pyrazolyl)borate,133 potassium tris(3-

isopropyl, 4-bromo-l-pyrazolyl)borate,132 and potassium tris(3-(2'-thienyl)-

lpyrazolyl)borate130 to W(CC(CH3)3)C13(DME) resulted in extremely low yields (< 5% by

1H NMR) of the corresponding substituted trispyrazolylborate tungsten alkylidyne
dichloride compounds and resonances indicated significant amount of substituted

trispyrazolylborate ligand decomposition. Because of the extremely low yields of these

compounds, they were not pursued further.

Conversion to Trispyrazolylborate Oxo Alkvlidene and Dioxo Alkvl Compounds

The conversion of the trispyrazolylborate tungsten alkylidyne compounds to

trispyrazolylborate tungsten oxo alkylidenes and, in some cases, to trispyrazolylborate

tungsten dioxo alkyls is an interesting phenomenon. The conversion of trispyrazolylborate

tungsten alkylidyne dihalide compounds to trispyrazolylborate tungsten oxo alkylidene

halide compounds can be thought of as the net protonation of the alkylidyne moiety and

displacement of one halide ligand by a terminal oxo ligand. The conversion of the

trispyrazolylborate tungsten oxo alkylidene halide compounds to trispyrazolylborate

tungsten dioxo alkyl compounds carries this process one step further, corresponding to net

protonation of the alkylidene and displacement of the remaining chloride with another

terminal oxo ligand. A one-step conversion from trispyrazolylborate alkylidyne dihalide to

trispyrazolylborate dioxo alkyl seems extremely unlikely. The metal center maintains its

coordinative and electronic saturation, as well as its formal oxidation state (+6) in each of

the new products. Each of the trispyrazolylborate alkylidyne compounds synthesized

underwent conversion to the corresponding trispyrazolylborate oxo alkylidene halide under

a variety of conditions. However, only Tp'W(CC(CH3)3)C12, the most stable of the

trispyrazolylborate alkylidynes synthesized, was converted solely to the Tp' tungsten oxo

neopentylidene chloride, TpW(O)(CHC(CH3)3)Cl, with no trace of the corresponding Tp'

tungsten dioxo neopentyl, TpW(0)2(CH2C(CH3)3), under any reaction conditions. Each

of the other compounds synthized, both the Tp' and the unsubstituted Tp compounds,

underwent further conversion to the corresponding trispyrazolylborate tungsten dioxo alkyl

compounds. The rate and product ratio of the conversion reaction was found to be

dependent upon the reactants used to achieve conversion.

These conversions reactions are thermodynamically driven by the formation of a

tungsten terminal oxo triple bond. Although there is little exact data concerning the bond

strengths of transition metal carbon multiple bonds, the third bond of the tungsten

alkylidyne linkage would be expected to be much weaker than the tungsten-oxygen triple

bond formed. A tungsten-halide single bond is also lost in each step of the conversion

process as well. The strength of this tungsten-halide bond would be expected to influence

the conversion process. This expectation appears to be born out in the observation that the

Tp' tungsten benzylidyne dibromide compound, 6, undergoes conversion to the Tp'

tungsten oxo benzylidene, 7, more rapidly than the Tp' tungsten neopentylidyne

dichloride, 1, regardless of the reagent used to achieve the conversion. Further, the Tp'

tungsten benzylidyne dibromide undergoes conversion to the Tp' tungsten dioxo benzyl

readily while the Tp' tungsten neopentylidyne dichloride has not been observed to generate

Tp' tungsten dioxo neopentyl under any of the reaction conditions employed, despite that

fact that Tp' tungsten dioxo alkyls were readily synthesized independently from Tp'

tungsten dioxo chloride. It must be noted, however, that the dichloride and dibromide

alkylidynes are not entirely analogous since the dichloride compounds contain a

neopentylidyne moiety while the dibromide compounds contain a benzylidyne moiety. The

substituent on the alkylidyne functionality results in both steric and electronic differences

between the two compounds. Since the alkylidyne functionality is cis to the halides, steric

differences at this site could be expected to have significant ramifications on the kinetic

barriers to halide displacement as well. Since synthesis of the Tp' tungsten benzylidyne

dichloride and Tp' tungsten neopentylidyne dibromide was not undertaken, the steric

versus electronic effects of the substituents of these compounds on their conversion to Tp'

tungsten oxo alkylidene halide and Tp' tungsten dioxo alkyl compounds cannot be fully


The unsubstituted Tp tungsten neopentylidyne dichloride was found to undergo

conversion to Tp tungsten oxo neopentylidene monochloride much more rapidly than its

Tp' analog.f Enhanced reactivity would be expected for Tp compounds on the basis of the

f The Tp tungsten benzylidyne dibromide compound was synthesized and characterized by
1H NMR spectroscopy, but satisfactory 13C NMR spectroscopy, mass spectropscopy, and
elemental analysis were not obtained for this compound. The lack of complete
characterization for this compound was attributed to a combination of factors including

greatly enhanced access to the metal center for Tp compounds as compared to Tp'

compounds.126 Thus, kinetic enhancement of the Tp tungsten alkylidyne dihalide

conversion reactions (in comparison to the analogous Tp' tungsten alkylidyne reactions) is

anticipated. It has been demonstrated that there are minimal electronic differences between

analogous Tp and Tp' compounds with Tp' compounds being slightly more donating than

Tp compounds, as measured by infrared spectrsocopy.126 The small electronic differences

measured, however, would not seem to account for the large differences in reactivity.

Enhanced kinetic accessibility of the metal center must also account for the fact that the Tp

tungsten neopentylidyne dichloride, 3, was observed to be converted to the Tp tungsten

dioxo neopentyl compound, 5, by a variety of means while the analogous Tp' tungsten

dioxo neopentyl compound was not observed under any reaction conditions. Probably,

both steric and electronic effects play a role in the conversion process as in the synthesis of

the alkylidyne compounds.

Reaction of trispyrazolylborate tungsten alkylidynes with alumina was found to be

the most facile and rapid method for effecting conversion of each of the compounds

synthesized. Solvent effects for the alumina mediated conversions were found to be

significant Solvents in which the trispyrazolylborate tungsten alkylidyne dihalide

compounds were not soluble effected the conversions most rapidly. Observation of the

colors of the slurry components suggests that the trispyrazolylborate tungsten alkylidyne

dihalides were strongly bound to the alumina bed only when they were not significantly

solvated. The alumina bed was strongly colored during reactions run in solvents in which

the tungsten compounds were not soluble, while the solution above the alumina bed was

strongly colored during reactions in which the tungsten compounds were soluble. Binding

of tungsten alkyidynes and alkylidenes to alumina has been studied, but is not well

understood. This observation, coupled with the observation that alumina-mediated

conversion rates were so much faster than all the other methods attempted, suggests that

decreased stability of the Tp tungsten benzylidyne dibromide compound and increased
byproducts in the synthesis of this compound which rendered purification difficult.

there is a binding interaction between the alumina the organometallic compounds which

facilitates the conversion reactions.

Conversions on alumina were also found to be affected by the pre-treatment of the

alumina to some degree. Basic alumina was found to effect the conversion process slightly

more rapidly than neutral alumina, but considerably more rapidly than acid-washed

alumina. Because of the observed stability of all of the trispyrazolylborate compounds to a

variety of strong acids, and the lowered rate of conversion on acid-washed alumina, the

conversion is not believed to be acid catalyzed. Alumina, however, even acid-washed

alumina, has a significant number of hydroxide ions available on its surface. One possible

mechanism for the conversion of trispyrazolylborate tungsten alkylidyne dihalide

compounds to trispyrazolylborate tungsten oxo alkylidene chloride compounds involves

displacement of a chloride ligand by a hydroxide, followed by a rapid intramolecular proton
shift from the hydroixde to the a-carbon of the alkylidyne,144 Figure 2-31. Each of the

other means of converting trispyrazolylborate tungsten alkylidyne dihalide compounds to

trispyrazolylborate tungsten oxo alkylidene halide or trispyrazolylborate tungsten dioxo

alkyl compounds studied had a source of hydroxide ions as well. Triethylamine and one

equivalent of water was the second most rapid reagent for effecting the conversion.




a CB // H

Figure 2-31. Proposed mechanism for the conversion of Tp'W(CC(CH3)3)C12 to

For compounds which could be converted to a mixture of trispyrazolylborate

tungsten oxo alkylidene halide and trispyrazolylborate tungsten dioxo alkyl, even one

equivalent of water resulted in some conversion to the trispyrazolylborate tungsten dioxo

alkyl compounds. This result suggests that the displacement of the halide from the

trispyrazolylborate tungsten oxo alkylidene halide proceeds at a rate which is competitive

with the displacement of a halide from the trispyrazolylborate tungsten alkylidyne dihalide

using this reagent. It is not possible to make this judgment concerning the alumina-

mediated conversions since the stoichiometry of the hydroxide ion was not controllable.

Both wet coordinating solvent and lithium hydroxide were relatively slow means of

effecting the conversions. The low effectiveness of lithium hydroxide in this reaction is

very likely due to low solubility of anhydrous lithium hydroxide in organic solvents.

Finely ground and heated lithium hydroxide reactions were found to exhibit enhanced rates

of conversion with these compounds, however, even under these conditions the reactions

were very slow compared to all other methods. It is interesting to note that the reagents

which promoted the fastest rates of conversion corresponded to the greatest percentages of

trispyrazolylborate tungsten dioxo alkyl product, while the slower reactions resulted in

greater percentages of trispyrazolylborate tungsten oxo alkylidene halide.


The synthesis of a number of trispyrazolylborate stabilized tungsten alkylidynes and

alkylidenes was achieved in moderate to low yield by the addition of the potassium salts of

Tp and Tp' to alkylidyne and alkylidene containing compounds.137'144,145 These materials

represented unprecedented increases in thermal, hydrolytic and oxidative stability for high

oxidation state alkylidyne and alkyidene compounds. The low yields of these compounds

were found to be due to competing addition and decomposition reactions of the

trispyrazolylborate ligands. The decomposition reactions were proposed to be due to the

strong Lewis acid character of the dO tungsten center of the starting materials, which

resulted in cleavage of a pyrazolide moiety from the trispyrazolylborate anion. Products
containing abstracted pyrazolide or pyrazole moieties were isolated in support of this
proposal, including pyrazabole, [(HB)2(-N2C5H7)3][(((CH3)3CC)WCl2)2(pj-C)2(-

N2CsH7)], [Tp'W(NAr)(CHC(CH3)2Ph)(PyrH)][SO3CF3], and
Tp'W(NAr)(CHC(CH3)2Ph)(Pyr).145 The amount of decomposition that occurred was
found to be far greater for the more sterically demanding Tp' ligand than for the
unsubstituted Tp ligand, resulting in higher yield for Tp tungsten compounds than for Tp'
tungsten compounds. In some cases, as in the addition of KTp' or KTp to

(DME)W(NAr)(CHC(CH3)2Ph)(OTf)2, different products were obtained based on the
trispyrazolylborate ligand employed and the stoicheometry of the reaction.145 The Tp'
ligand decomposition led to pyrazole or pyrazolide coordinated compounds that were not
isolated for the analogous Tp reactions.
Due in part to the extraordinary stability of the trispyrazolylborate tungsten
alkylidynes and alkylidenes, complete characterization of the compounds was achieved.
Further characterization included the 1H NMR, 13C NMR, and infrared spectra of the
compounds.137,144,145 The characteristic distinguishing features of trispyrazolylborate
tungsten alkylidyne and alkylidene compounds were delineated for these analytical
methods. The use of the trispyrazolylborate ligands as NMR probes of the symmetry of
the molecules was presented. Single crystal x-ray studies were performed for three
compounds, including Tp'W(O)(CHC(CH3)3)CI,137 [Tp'W(NAr)(CHC(CH3)2Ph)(PyrH)]
[SO3CF3],145 and [(HB)2(j-N2C5H7)3][(((CH3)3CC)WCl2)2(t-Cl)2(-N2C5H7)]
Each of the trispyrazolylborate compounds synthesized was found to have pseudo-
octahedral geometry at the tungsten center in keeping with the propensity of
trispyrazolylborate ligands to enforce octahedral geometry at the metal center,162 and was
therefore, coordinatively saturated. No evidence for an agostic interaction of the oa-

hydrogen of the trispyrazolylborate tungsten alkylidene monohalides was found for any of
the compounds synthesized.

A major area of interest is the synthesis of trispyrazolylborate stabilized tungsten

alkylidenes and alkylidynes by a more facile route. Attempts to generate alkylidynes or

alkylidenes from the readily accessible compounds Tp'W(0)2R (R = Cl, neopentyl,

benzyl) and Tp'WC13 have not proven successful to date. A facile route to some Tp

tungsten imido alkyl alkylidene compounds has recently been developed.g Although these

compounds can be synthesized in essentially one pot, the yields are modest (22 to 38%)

and the synthesis is limited to imido compounds and Tp ligands. A high yield synthetic

route, and a facile preporatory scheme would greatly advance the field, making large

amounts of stable tungsten alkylidynes and alkylidenes readily available for further study.

Continued studies on facile routes to trispyrazolylborate tungsten alkylidynes and

alkylidene are currently underway.

s Gamble, S. A.; Boncella, J. M. Organometallics, manuscript accepted for publication.



The synthesis of trispyrazolylborate stabilized high oxidation state tungsten

alkylidynes and alkylidenes has been demonstrated.137'144'145 These syntheses opened up

a new class of tungsten(6) alkylidyne and alkylidene compounds with thermal, oxidative

and hydrolytic stabilities far in excess of most other non-chelate stabilized tungsten(6)

alkylidynes and alkylidenes which have been previously reported.1 Frequently, however,

there is a trade-off between stability to undesired reaction pathways, such as thermolysis

and hydrolysis, and activity for desired reactions. For example, polypyrazolylborates with

extremely large substituents on the pyrazole carbons are often extraordinarily stable, but they

are also inert to all but the most extreme reactions conditions.128 The less bulky

polypyrazolylborate ligands, like Tp and Tp' have been used in a variety of ways.
Stabilization of unusual geometries, electronic configurations, or oxidation states, are some

of the most popular uses of these ligands, however, trispyrazolylborate ligands have also

been employed in enzymatic model compounds and other reactive species.128 Frequently,

but not always, the reactivity of such compounds is limited to reactions at the ligands rather

than at the metal center itself. The question then becomes what kind of reactivity will these

remarkably stable trispyrazolylborate tungsten alkylidene compounds evidence. In

particular, what type of metathesis activity will these trispyrazolylborate tungsten oxo and

imido alkylidene compounds exhibit. Can the observed reactivity be correlated with the

steric properties of the trispyrazolylborate ligands, or will the steric and electronic effects of

the other ancillary ligands dominate the chemistry of these compounds. This question of

the reactivity of trispyrazolylborate stabilized tungsten alkylidenes has just begun to be


Metathesis Polymerizations

The olefin metathesis activity of the trispyrazolylborate stabilized tungsten

alkylidenes synthesized was of great interest. None of the trispyrazolylborate tungsten

alkylidenes evidenced any metathesis activity with cyclic olefins (norborene or

cyclooctene) or acyclic dienes (1,9-decadiene) even in refluxing monomer.137'144,145 This

lack of reactivity was not surprising since each of the trispyrazolylborate tungsten

alkylidenes is coordinatively (six-coordinate) and electronically (18 electrons) saturated.

Seven-coordinate trispyrazolylborate tungsten compounds have been reported in the

literature, but they are quite rare and generally consist of sterically undemanding ligands

such as carbonyl and hydride or proton.128 The neopentylidene or neophylidene ligands in

the trispyrazolylborate tungsten alkylidene compounds are much more sterically demanding

co-ligands and would be expected to significantly reduce access to the tungsten metal

center. The trispyrazolylborate ligands themselves effectively prevent access to the metal

center from the other side of these compounds. Under these conditions, it is not likely that

the monomer olefin could coordinate to the metal center, a condition that has been put forth

as a requirement for olefin metathesis activity.1,565,66 Addition of a Lewis acid cocatalyst

has been shown to generate active metathesis catalysts from many transition metal

compounds which were electronically or coordinatively saturated.5,6,14,42,66,163,164

Lewis acid addition to the trispyrazolylborate tungsten alkylidene compounds was

undertaken in an effort to generate a vacant coordination site.

When one equivalent of aluminum trichloride or gallium tribromide was added to
Tp'W(O)(CHC(CH3)3)Cl in cyclooctene at 250C, rapid ROMP of cyclooctene occurred,

producing high molecular weight, solid polyoctenamer within 15 minutes,137 Figure 3-1.


AlCl3 X
A1CI3 x

Figure 3-1. ROMP of cyclooctene with Tp'W(O)(CHC(CH3)3)CI and AIC13.

Cyclooctene polymerization reactions were performed neat at room temperature

with a 500:1 monomer.catalyst ratio. The polymerization may be performed either in the air

or under an inert atmosphere producing polymer with essentially the same molecular weight

and polydispersity under either set of conditions, Table 3-1. The polymerization catalyst

remained active after the initial reaction mixture solidified since subsequent additions of

monomer to the solid polymer-catalyst mixture resulted in dissolution of the polymer

followed by resolidification as the polymerization reaction continued. The precipitated

polymer-catalyst mixture could also be heated to melt temperatures and stirred for several

minutes until higher molecular weight polymer again precipitated, indicating that the

catalyst mixture was active at least for several minutes at low melt temperatures.

Table 3-1. Molecular weight and polydispersity of polyoctenamer prepared with
compound 2 and AIC13.

polymer characterization in Air in N2
Mn 49000 59000

Mw/Mn 1.3 1.3

Norbomene was also polymerized by a mixture of compound 2 and one equivalent

of aluminum trichloride (Figure 3-2) in toluene solution either in the air or under an inert

atmosphere, producing polymers with similar though not identical properties,137 Table 3-2.

Polynorbomene was prepared by weighing equimolar quantities of compound 2 and

aluminum trichloride into a test tube in a drybox followed by dissolution of the mixture in

toluene in air or under an inert atmosphere. It is significant that this polymerization can be

carried out in air since this result demonstrates that the catalyst that is generated by mixing

compound 2 and aluminum trichloride is active in solution in the air for at least 15 minutes

(the time of the reaction). It is possible that the active species in this polymerization is

simply more reactive toward monomer than oxygen or water. This catalyst is not expected

to have the long-term air stability observed for compound 2, especially since aluminum

trichloride degrades rapidly in moist air. Studies are currently underway to find a more

suitable cocatalyst

C1 X

Figure 3-2. ROMP of norborene with Tp'W(O)(CHC(CH3)3)Cl and AICd3.

Table 3-2. Molecular weight and polydispersity of polynorbornene prepared with
compound 2 and A1C13.

polymer characterization in Air in N2
Mn 67000 76000
MwMn2.2 1.8

In an attempt to observe the catalytically active species in these polymerizations, the

reaction of compound 2, aluminum trichloride, and three equivalents of cyclooctene was
carried out in an NMR tube in dichloromethane-d2. At -600C, signals assignable to three

new alkylidene species (9.6, 9.9, 10.2 ppm)a were observed in the 1H NMR spectrum as

were signals due to polyoctenamer (5.4 ppm) and possibly a metallacyclobutane species

(6.4, 0.5 ppm).131 The fact that polyoctenamer rather than trimer was generated from three

equivalents of monomer indicated that only a small amount of the active catalyst species had

a The resonance at 10.2 ppm can now be definitively assigned as the rotational isomer of
the starting alkylidene of compound 2. Rotational isomerization of alkylidenes is presented
in detail in chapter 4 of this dissertation.

been generated under these conditions. The small amount of active catalyst and the
complexity of the mixture prevented assignment of the active species.
As previously mentioned the addition of Lewis acid cocatalysts to similar systems
has succeeded in generating olefin metathesis catalysts previously, but these catalyst
systems do not operate in air. In many cases, the catalyst mixtures thus generated are ill
defined,5,6 however, Osborn and coworkers have studied the site of Lewis acid interaction
in several tungsten systems. For example, with the compound W(O)(CH2C(CH33)3)Br,
which does not contain an alkylidene functionality but was proposed as a model for oxo
alkylidene Lewis acid interaction studies, gallium tribromide was shown to coordinate to
the terminal oxo ligand.165 In a similar imido compound, W(N(CH3))(CH2C(CH3)3)3)C1,
gallium tribromide was shown to coordinate to the halide ligand.166 And in many
compounds which do not contain oxo or imido functionalities, Lewis acid was shown to
interact with halide functionalities, often in a weak equilibrium interaction so that large
excesses of Lewis acid (up to six equivalents) were required to efficiently interact with one
equivalent of tungsten halide.66'163.164 In other systems, such as Tebbe's reagent, Lewis
acid has been shown to interact with the alkylidene functionality.16
The active catalyst species for the Tp' tungsten oxo alkylidene monohalide species

was originally proposed to be a five-coordinate cationic alkylidene generated by attack of
aluminum trichloride at the halide functionality,137'144'145 Figure 3-3.

H H [AlC4
+ AICi 3

N\ A N N
C1 Sr

Figure 3-3. Five-coordinate cationic alkylidene originally proposed as active ROMP
catalyst in TpW(O)(CHC(CH3)3)CI and aluminum trichloride mixtures.

This proposal was based on Osborn's previous studies. Attack at the oxo

functionality cannot be excluded, however, Lewis acid coordination to the oxo ligand

would not generate a vacant coordination site for incoming monomer. Further, since the

oxo ligand would remain coordinated to the metal center and is proximal to the alkylidene

active site for metathesis reactivity, Lewis acid coordination at the terminal oxo

functionality would serve to further sterically encumber the metal center, rendering

monomer coordination more difficult rather than more facile. Lewis acid attack at the

alkylidene functionality suffers from the same difficulties as attack at the oxo ligand, and is

generally not proposed except for very sterically unencumbered alkylidenes, such as

terminal methylidenes.

Although it was clearly demonstrated that Tp'W(O)(CHC(CH3)3)Cl was not an

active ROMP catalyst in the absence of a Lewis acid cocatalyst, it was conceivable that a

seven-coordinate trispyrazolylborate tungsten oxo alkylidene halide olefin compound might

be initially generated in the ROMP reaction if less bulky ancillary ligands were employed.

The formation of a seven coordinate compound should be significantly more facile with

compound 4 than with the its bulky 3,5-dimethyl substituted Tp' analog, compound 2, due

to reduced steric constraints of the Tp ligand as compared with the Tp' ligand. In order to

investigate this possibility, the metathesis activity of the unsubstituted Tp analog of

compound 2, TpW(O)(CHC(CH3)3)C1, was studied.

In the absence of Lewis acid, compound 4 was not found to be an active metathesis

catalyst. Although it is less sterucally encumbered at the metal center than compound 2,

compound 4 is also both electronically and coordinatively saturated. Addition of an

equivalent of aluminum trichloride to a cyclooctene slurry of compound 4 generated an

active ROMP catalyst Interestingly, the catalyst generated with compound 4 and

aluminum trichloride was less active than the Tp'W(O)(CHC(CH3)3)C/Lewis acid system.

One possible explanation for this unexpected effect is the poor solubility of compound 4 in

the hydrocarbon monomers. Additionally, compound 4 is not as stable in solution as its

Tp' analog, so some decomposition may be occurring as well. Polar monomers might help

to facilitate the metathesis reaction if they solvate compound 4 to a greater extent, however,

potential difficulties with decomposition to the known Tp tungsten dioxo alkyl species,

compound 5, or other oxo coordinated intermediates could result if oxo containing

functionalities are present in the monomers or in polar solvents used in the polymerization


The metathesis activity of TpW(O)(CHPh)Br was also investigated.144 Since

bromide should be a better leaving group than chloride, Lewis acid interaction with

compound 7 would be expected to be stronger. Consequently, compound 7 might be

expected to be a better olefin metathesis catalyst than compounds 2 and 4 on the basis of

the more efficient interaction of Lewis acid with the halide functionality. As expected, in

the absence of Lewis acid, compound 7 showed no olefin metathesis reactivity even in

refluxing monomer. The catalyst generated by addition of a Lewis acid to compound 7

was an extremely poor ROMP catalyst, generating only low molecular weight oligomers

after days of stirring with cyclooctene. Insoluble solids generated in these ROMP reactions

indicated that decomposition of the catalyst mixture was occurring even in an inert


Similarly the olefin metathesis activity of [TpW(NAr)(CHC(CH3)2Ph)(PyrH)]-

[SO3CF3] and TpW(NAr)(CHC(CH3)2Ph)(Pyr) were also investigated.145 The cationic
compound 9 showed no evidence of metathesis activity even in the presence of Lewis acid

cocatalysts. Compound 9 also showed no evidence of solubility in the hydrocarbon

monomers employed in this study. Undoubtedly the insolubility of compound 9 in

monomer contributed to the observed lack of metahtesis activity. Additionally, compound

9 is the most sterically encumbered trispyrazolylborate stabilized tungsten alkylidene

ccompound generated in this study, and monomer coordination may be curtailed for steric

reasons. Since compound 9 was produced in such low yields, further metathesis studies

with this compound were not pursued. The neutral analog of compound 9, compound 11,

was found to have similar metathesis activity to compound 7, producing only low

molecular weight oligomers after days of reaction with neat cyclooctene. Since compound

11 was soluble in hydrocarbon monomers and was demonstrated to be extremely stable to

the polymerization conditions, the observed lack of metathesis activity for this compound

was attributed to steric impediments to monomer coordination. Compound 11 should not

be significantly less sterically encumbered than compound 9, since the two species differ

by a single proton on the ancillary coordinated pyrazole residue.

TpW(O)(CHC(CHjl)C1 Reactivity Studies

Since the metathesis activity of the compounds was of critical interest, each of the

trispyrazolylborate tungsten alkylidene compounds presented in chapter 2 of this

dissertation were evaluated as metathesis catalysts. However, because of the low yields

and difficult synthetic routes to several of the compounds reported in chapter 2 of this

dissertation, the reactivity studies of trispyrazolylborate alkylidene compounds performed

to date have focused primarily on the trispyrazolylborate tungsten alkylidene species which

has been prepared in the highest yield and greatest purity, TpW(O)(CHC(CH3)3)C1.

Reactivity studies have so far have emphasized species which might give some evidence as

to the nature of the active species in the metathesis reactions of the trispyrazolylborate

alkylidene compounds with Lewis acid cocatalysts, and on species with other potential

catalytic activity.

Attempted Halide Abstraction from TpW(O)(CHC(CH3~i)C1

In an attempt to isolate a five-coordinate cationic alkylidene like that originally

proposed as the active catalyst species for the olefin metahtesis polymerizattion of

cyclooctene or norbornene with trispyrazolylborate tungsten oxo alkylidene halides and

Lewis acid cocatalysts, halide abstraction from TpW(O)(CHC(CH3)3)Cl was undertaken

using a variety of reagents. Silver salts have a strong affinity for chloride and were first

studied as potential halide abstraction reagents with compound 2. Silver triflate reacted

very slowly with compound 2 to give a mixture of at least three products by IH NMR.

Silver tetrafluoroborate and silver hexafluorophosphate both reacted with compound 2 over

48 hours but complex mixtures of products were generated, even in the presence of Lewis

bases, probably due to fluoride abstraction and other decomposition reactions which have

been demonstrated for similar reactions with tungsten halides.36

Lewis acids were also employed as halogen abstraction agents with mixed results.

These studies compliment the low temperature TpW(O)(CHC(CH3)3)CI/AlCl3/cyclooctene

study presented earlier in this chapter, investigating the interaction of Lewis acids with

compound 2 in the absence of olefins. Weak Lewis acids, such as zinc dichloride and the

somewhat stronger Lewis acid trisperfluorophenylboron (Bpfp)167 exhibited only moderate

reactivity with compound 2 over weeks at room temperature. In 48 hours, no evidence of

reaction was detected by 1H NMR with either Lewis acid. After five days, with Bpfp, the

reaction mixture was noticeably darker, with a purple tinge, but only minimal evidence of

reactivity was observed by 1H NMR and the products of the interaction could not be

identified. After three weeks, all evidence of the starting alkylidene resonance of
compound 2 was lost and multiple resonances in the Tp' region were apparent. These

resonances may correspond to pyrazaboles or other ligand decomposition products. A

broad resonance at 9.9 ppm suggested the presence of free 3,5-dimethyl pyrazole.

Addition of authentic 3,5-dimethyl pyrazole to the reaction mixture resulted in enhancement

of the broad resonance suspected to correspond to the pyrazole proton and a slight

downfield shift of that resonance 10.0 ppm, confirming this assignment.

Addition of strong Lewis acids such as aluminum trichloride or gallium tribromide

to compound 2 resulted in the rapid loss of the stating material alkylidene 1H NMR

resonance which was replaced by a new alkylidene resonance at 9.9 ppm and several very

broad resonances (almost disappearing in the baseline) in the region from 11.2 to 9.6 ppm.

Multiple resonances assignable to 3,5-dimethyl pyrazole, unidentified tungsten Tp'

compounds or pyrazaboles were also present.

Low temperature studies were undertaken to determine the initial products of the

reaction of aluminum trichloride with compound 2. At -60C, the new alkylidene

resonance at 9.9 ppm was already apparent as was a broad resonance at 9.5 ppm. As the

sample was gradually warmed to room temperature, the starting compound alkylidene

resonance broadened into the baseline and slowly diminished in intensity as three broad

resonances grew into the spectrum. The complexity of the mixture indicates multiple

products formed either by attack of the Lewis acid at multiple sites or subsequent reactions

after attack at a few initial sites. The broad resonances in the 11.2 to 9.6 ppm region

further suggest that the Tp' ligand has decomposed, probably to pyrazole and coordinated

tungsten pyrazole compounds. The complexity of the 1H NMR of the mixture precluded

further analysis.

Mass spectroscopy studies of the trispyrazolylborate stabilized tungsten alkylidene

compounds further support the contention that multiple kinetic sites of Lewis acid attack are

present on these molecules. The mass spectra of these molecules is complicated by the

number of isotopes of tungsten (5 isotopes) boron (2 isotopes) and carbon (2 isotopes) that

occur in large enough abundance to appear in a mass spectrum. These isotope distributions

result in mass envelopes for the compounds and their fragments rather than in discrete

mass/charge peaks, Figure 3-4. The assignment of mass envelopes is readily accomplished

by computer modeling of the isotope distribution of a given parent ion or fragment ion and

comparison of the calculated and observed mass/charge envelopes. Excellent agreement

between the calculated and observed mass envelopes for the trispyrazolylborate tungsten

compounds presented in this study were be achieved by this method, Figure 3-5. Matching

of calculated and observed mass/charge envelopes significantly reduces the possibility of

erroneous assignment of a mass/charge envelope.

%D un
o m
H rM


) rr
ir foo

iCl 71044


I- m LZ-----

z- z 0

In =

N _______

U I __

N 0



H- -