Gas-phase reactions of bis(Cyclopentadienyl) methyl zirconium (1+) with unsaturated hydrocarbons and nitriles


Material Information

Gas-phase reactions of bis(Cyclopentadienyl) methyl zirconium (1+) with unsaturated hydrocarbons and nitriles
Physical Description:
x, 154 leaves : ill. ; 28 cm.
Christ, Charles Stewart
Publication Date:


Subjects / Keywords:
Ziegler-Natta catalysts   ( lcsh )
Polymerization   ( lcsh )
Transition metal complexes   ( lcsh )
Zirconium   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1989.
Includes bibliographical references (leaves 144-153).
Statement of Responsibility:
by Charles Stewart Christ, Jr.
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 - 001479116
notis - AGZ1059
oclc - 21071833
System ID:

Full Text







To my Pap, Gerard F. Genellie


This research and dissertation could not have been

possible without help and encouragement from family, friends,

and professorial advisors and their staffs. It is difficult

to identify directly each person that has had an impact on

this work. Therefore, I would like to extend a heartfelt

thanks to all those I have come to know here at the University

of Florida. I believe that each has played a part in my

development as a person and a chemist.

Specifically, I would like to thank my advisor, Dr. David

E. Richardson, who not only provided leadership, guidance,

and patience but also a passion for chemistry and a unique

friendship. Dr. John R. Eyler also deserves a great deal of

thanks for his willingness to share with me his expertise in

the fields of physical chemistry and life. Dr. James Boncella

helped me to a greater understanding of synthesis and

organometallic chemistry. Cliff Watson and Bryan Hearn were

always available to discuss the particulars of the FTMS and

without them my understanding of the system would be greatly



I must thank the entire Richardson research group, each

member with so different a personality and outlook on life,

but all at one time or other helping me to reach this one

goal. My thanks go to Paul Sharpe for those early days

studying for Dr. Ryschkewitsch's advanced inorganic class.

I learned so much and had such a great time. Mehrzad Emad was

always there to listen to my organic questions, answering in

mechanisms and discussing the problems in Iran. Francis

Armitage and her inspirational organization also provided much

insight into life. Cyndy and Jay Bongers taught me about self

sacrifice and gave me an alternate perspective. Matt Ryan,

when things were tough, had a way of helping me to lighten up.

Finally, I can not begin to thank Fran Goodman for all the

time and energy she spent reading my first attempts. I

learned a great deal and found a special friendship that I

will not soon forget.

My family has sacrificed tremendously to make me what I

am and impart to me the anti-quitter mentality that has

brought me to this place. My brothers have also made great

contributions and will always be close to my heart. I could

never leave out my Grandfather, Gerard Genellie. The man is

inspired and inspiring. He has always been there with support

of all kinds (just like a Dutch Uncle). Most of all I want

to thank my wife Anne for not only putting up with my

incredible mood swings and still being my friend but also for

sacrificing her time and effort for our family and

particularly for me while my one track mind has concentrated

almost exclusively on the task at hand.

A second family also deserves significant credit. Both

Dr. and Mrs. Radcliffe and their unforgettable daughter have

certainly added an important ingredient to my life and

education. They have been extremely kind and allowed me the

use of their computers, which is of obvious importance. They

also were very generous when things were tough financially.

I will always remember Irene's words "I've been there Chuck."

I only hope that I will be able to repay this "cosmic dept"

by helping someone else as the Radcliffes have helped me.

Finally, I gratefully acknowledge the National Science

Foundation, the American Chemical Society (Petroleum Research

Fund), and Research Corporation for support.



ACKNOWLEDGEMENTS .................................. iii

KEY TO ABBREVIATIONS ................................ viii

ABSTRACT .................. ........... ................ ix


1. GENERAL INTRODUCTION ......................... 1

Scope ..................................... 1
Ziegler-Natta Catalysis .................. 2
Homogeneous Analogues of Ziegler-Natta
Catalysts ................................. 6
Carbon-Hydrogen Activation ................ 19
Gas-Phase Reactions ....................... 25
Theoretical Studies ...................... 29
Conclusions ............................... 30


Results .................................... 32
Discussion ................................ 48
Conclusions .............................. 71
Experimental ............................. 74

(1+) WITH NITRILES ......................... 77

Introduction ............................... 77
Results ................................... 78
Discussion ...................... .... ..... 95
Conclusions .............................. 115
Experimental .............................. 117

4. EXPERIMENTAL PROCEDURES ...................... 120

Pressure Measurement ..................... 120
Designing the FTICR Experiment .......... 126

5. SUMMARY AND CONCLUSIONS ..................... 133

Relationships to Ziegler-Natta
Catalysis ........................... 133
Carbon-Hydrogen Activation
in the Absence of Solvent ........... 135
Interactions of Cp2ZrCH3+ with Nitrile
Substrates .......................... 136
Conclusion .............................. 137

APPENDIX ............................................... 139

REFERENCES .............................. ........... 144

BIOGRAPHICAL SKETCH ................................... 154



bipy = bipyridyl

Cp = cyclopentadienyl

Cp* = pentamethylcyclopentadienyl

M = metal

Me = methyl

R = hydrocarbyl

THF = tetrahydrofuran


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



Charles Stewart Christ, Jr.

May, 1989

Chairman: David E. Richardson, Ph.D.
Major Department: Chemistry

The solution phase investigation of dO and dOfn electron-

deficient lanthanide, actinide and transition metal

organometallic complexes such as Cp2ZrCH3(CH3CN)n+ (n = 1, 2)

and Cp*2ScCH3 (Cp = pentamethylcyclopentadienyl), has

revealed a high reactivity toward hydrocarbon and nitrogen

containing hydrocarbon substrates. A similar reactivity is

observed in the gas-phase reactions of Cp2ZrCH3 1, (Cp =

Cyclopentadienyl). The present work is a report on the gas-

phase reactions of 1 with dihydrogen and a variety of alkenes,

alkynes, and nitriles. Whenever possible a direct comparison

to the solution reactivity of analogous complexes is

presented. Some thermodynamic information is also given and

labelled compounds were used in mechanistic investigations.

Cp2ZrCH3+ generally reacts with hydrocarbons and nitriles
by the two well known mechanistic sequences, migratory

insertion and carbon-hydrogen activation. Ethylene is

polymerized by I and similar complexes in solution, but

polymerization is not observed in the gas-phase. Reasons for

this difference in reactivity are given and additional

information on the effects of the gas-phase medium on the

reactions of 1 is also offered.

Although a great deal of specific reaction data is

compiled, the main thrust of the research is to illustrate the

usefulness of gas-phase techniques in further understanding

the reactivity of organometallic complexes in solution and the

gas-phase. In the short time of this research all of the

basic reaction sequences, insertion of alkenes, alkynes, and

nitriles, B-hydride shift, B-methide shift, and a number of

C-H activation reactions, have been observed in the reactions

of 1, which indicates the possibility of extending this type

of study to other organometallic systems.



The reactivity of Cp2ZrCH+ in the gas-phase is

directly comparable to the reactivity of analogous dO and

dOfn metal complexes in solution. The initial synthesis and

study of isolated lanthanide, actinide and transition metal

alkyl complexes in condensed phase is in many cases directly

related to the study of Ziegler-Natta polymerization


The objectives of the present research are threefold.

First, the mechanism of Ziegler-Natta polymerization has

been under investigation for over thirty years. Since

cationic metal-alkyls similar to Cp2ZrCH3+ have been

implicated as possible active species, the present work

seeks to add support to this assertion. Second,

investigations of solvent free organometallic complex ions

are not numerous and we believe much can be learned by

gas-phase studies. Therefore we intend to show that the

Fourier Transform Ion Cyclotron Resonance technique presents

a unique opportunity to characterize the intrinsic

reactivity of a well-studied homogeneous ethylene


polymerization catalyst. Finally, this study attempts to

make meaningful comparisons to the reactions of solution

analogues and perhaps present a new and useful perspective

on condensed phase reactions and proposed mechanisms.

The subject of Ziegler-Natta catalysis is discussed in

some detail in the opening portion of this introduction.

This discussion seems appropriate since the vast majority of

publications that mention Ziegler-Natta catalysis assume

that the reader has a working knowledge. Therefore a brief

review of Ziegler-Natta catalysis is prudent and necessary

for a complete understanding of the material to be


The tremendously extensive research literature

associated with the catalytic mixture is so large that no

authors have ever attempted an exhaustive review. One

source indicates that 25 years after the first report

appeared over ten thousand distinct journal articles and

patents regarding Ziegler-Natta catalysts had appeared.1

Thus, the introduction to this manuscript will only seek to

acquaint the reader with some of the salient background

features of Ziegler-Natta catalysis. In addition, a brief

review of the closely related reactivity known as carbon-

hydrogen activation is also included.2,3,4

Ziegler-Natta Catalysis

In the fall of 1953 K. Ziegler discovered a catalytic

system that produced high density polyethylene under mild


conditions. Polyethylene was commercially available;

however, high pressures were required for the

polymerization. During of the following year, G. Natta also

became involved by reporting the formation of isotactic

polypropylene.1 The chain of events resulting from the

development of Ziegler-Natta catalysts has indeed played an

extremely important role in the evolution of organometallic

chemistry.1'5-10'11,12 Its effect on the chemical industry

was profound since not only were new processes being

developed but a range of new products were being synthesized

and the polymer industry kept growing. In fact, one

estimate states that the products arising from this

technology in 1980 were worth nearly 9 billion dollars.13

As in many instances, commercial processes and

applications were largely refined long before a clear

understanding of the nature of the system had been

rendered.9 Even at present the Ziegler-Natta catalyst and

its mode of action are not exhaustively understood.

Nevertheless, polymeric materials permeate our man-made

surroundings and while all cannot be attributed to the

single event of discovery in 1953, the total worth of the

technology spawned from the production of new materials and

compounds and the study of the catalyst systems related to

their synthesis is immense.10

The accepted definition of a Ziegler-Natta catalyst has

itself evolved. As stated by Boor1 and derived from the


original patent definition, the catalyst consists of a

mixture of a main group metal-alkyl of groups I-III and a

transition-metal from the Ti,V,Cr,Mn, and Fe triads,

including Sc,Th and U. A variety of mixtures are only

active under certain conditions, and certain combinations

employing other transition-metals and f-elements are also

active. For example, the chromium based Phillips catalyst

and the Standard of Indiana catalyst are extremely active in

the polymerization of ethylene but do not contain a main

group alkyl and therefore do not fall under the category of


One reviewer suggests that the spirit of the first

Ziegler catalyst patents would even include the numerous

examples of isolated metal-alkyl complexes, which in some

cases are highly active in the polymerization of ethylene.9

Some examples of isolated metal-alkyls show poor activities

in homogenous media but are rather active when supported as

heterogenous catalysts.7

Another method of defining the Ziegler catalyst is

dependent on classification with respect to the mechanistic

mode of action.9 In many instances the literature refers to

the Ziegler-Natta catalyst with terms such as "anionic

coordination polymerization" or simply "coordination

polymerization."14 Natta is apparently responsible in part

for the term "anionic coordination polymerization," which

arises from the proposed coordination of an olefin unit to

the metal center prior to insertion.15 Since polarization

of the metal-carbon bond is likely, the term anionic was

applied. It seems probable that from an organic viewpoint

the term used to describe the polymerization would

correspond to the organic polymer portion of the

coordination catalyst. Thus the partial negative charge on

the carbon atom residing at the origin of polymer growth

yields the term "anionic coordination polymerization".

Indeed the electrophilic polarizing nature of the

lanthanide, actinide and early transition-metal alkyls is of

prime importance in the activation of alkenes toward

polymerization and carbon-hydrogen activating

processes.3'4'16-18 Confusion in terminology arises because

of the diversity in experimental results and proposed

mechanisms given by so many different investigators.

Mechanisms for the polymerization of ethylene by anionic

metal catalyst surfaces have also been asserted.19 However,

the majority of proposed mechanisms describe the

polarization of the metal-carbon bond with a partial

positive charge on the metal center. This agrees well with

implications of high valent transition metal catalytic

active sites.20-23

The term coordination, as applied to the catalyst,

attempts to describe one mechanistic feature of Ziegler

systems and its origin is immediately evident in light of

the abundance of the olefin r-complexes isolated by reaction


of Ni-dialkyls with olefins.24-26 The binding of an olefin

in effect activates the metal-carbon bonds of the nickel

complex resulting in the reductive elimination of the alkyl

ligands.5,8 In the case of nickel-dialkyls, polymerization

does not occur. However, the observed reaction presents

evidence for the initial coordination of olefin prior to

insertion. In several cases the r-complexes were isolated

without reductive elimination at low temperature.

Homogeneous Analogues to Ziegler-Natta Catalysts

The heterogeneous nature of Ziegler-Natta catalysts

render them unsuitable for mechanistic study.6 Therefore,

the large body of information concerning the mechanism by

which they convert alkenes to polymers has been achieved

through the study of homogeneous analogues. The belief that

the transition metal alkyl or hydride was the active species

in these systems lead to the synthesis and study of a

massive number of organometallic compounds and a better

understanding of the two component heterogeneous

catalysts.1, 5-8,11,17,18,28,29 One reason supporting the

notion that the transition metal is the propagating center

was the discovery of transition metal catalysts that

polymerize ethylene without the presence of a main group


The fact that the Ziegler-Natta mixture generated the

active species in situ from an alkylating agent and a

transition-metal salt portended the first approach of the


organometallic chemist. An alkylating agent was generally

used in conjunction with a transition metal salt just as

would be done in preparing the catalyst. Instead of adding

the monomer, attempts were made to isolate the transition

metal alkyls. This was quite difficult because addition of

water was the normal method for destroying excess alkylating

agents and transition metal alkyl were also destroyed in

this process. Great strides in the development of new

synthetic methods occurred yielding a great number of new

complexes.30,31 After isolation the monomer was added and

the reaction monitored by spectroscopic as well as chemical

techniques. The polymerization of ethylene was the standard

test for activity because it is generally most readily


Early efforts produced complexes such as C13TiCH3,

Cl2CrCH3(THF)3, and (bipy)Ni(CH3)2 from Ziegler type

reaction mixtures. In many cases the isolated complexes did

polymerize ethylene, but not with the same activity as the

two component heterogeneous systems.1,5-9

One of the most striking properties of the Ziegler-

Natta catalyst was its function in producing stereoregular

polypropylene,19 particularly, in the TiCl3/AlR3 system.
However, C13TiCH3 did not produce isotactic polypropylene.

Nevertheless, the general approach of isolating transition

metal alkyl and hydride complexes to be studied in

homogeneous reactions continues and has brought a greater

understanding to the mechanism and controlling factors in

Ziegler systems.

A mechanism was proposed by Cossee in 196426,32,33

involving steric control of the incoming propylene monomer

by the structure of the TiCl3 crystal surface. The study

included an in depth molecular orbital discussion and

invoked an octahedral environment surrounding the metal

center in which a vacant coordination site was necessary for

interaction with monomer. However, no universal agreement

has been achieved regarding the exact type of steric

interaction responsible.

From the tremendous effort expended in seeking to

understand how the Ziegler systems operated, several key

conclusions were made. (1) The transition metal alkyl or

hydride bond was necessary for polymerization. (2) One

propagating center per transition metal was active. This

was by no means unequivocally clear due to the complexity of

the systems being studied. (3) The role of aluminum

compounds was at least two-fold.5,6,34 (a) Aluminum alkyls

were responsible for the alkylation of the transition metal,

(b) Aluminum compounds stabilized the transition metal

alkyl, presumably by preventing deactivation by the

interaction of two transition metal centers. (4) The ligand

environment can induce the formation of isotactic polymers

and the growing polymer chain also influences

stereoregularity. (5) Increasing the Lewis acidity of the


metal complex increases activity for polymerization. (6) An

open coordination site available for interaction with

incoming monomer is essential.5,7,26 Several studies

indicate the retardation of polymerization rate by addition

of coordinating ligands such as pyridine. If an open

coordination site is not available, partial dissociation of

a ligand may be the rate determining step for initiation.

In the case of allyl metal complexes conversion from q3 to

71 seems likely for initiation of polymerization.7 (7)

Chain termination occurs through the abstraction of a B-

hydrogen and elimination of the polymer with an unsaturated

endgroup. This process was termed S-elimination and

generates a metal hydride which also undergoes insertion.

(8) Molecular weight control by addition of hydrogen also

yields a metal hydride and a polymer with a saturated

endgroup. (9) Terminal alkenes are polymerized much more

readily than internal olefins. In fact some studies

indicate that certain catalysts isomerize internal olefins

to terminal olefins before polymerization occurs.35

By the mid-1970's the polymer industry began to show

signs of maturity10, yet basic research in the area of

mechanistic study and development of new organometallic

alkyl complexes continues. The debate concerning the active

site in the Ziegler catalyst was ongoing as evidenced by the

proposal that appeared in 1978 and became known as the

Green-Rooney mechanism.36-39


Not only have many isolated transition metal alkyls and

hydrides been shown active in polymerization of olefins, the

lanthanide and actinide metals have also shown outstanding

catalytic potential.18'40-42 The knowledge gained in

earlier studies has provided valuable information, which has

revolutionized the design of new catalytic systems.

Nevertheless, after reading reviews such as Boor one cannot

help realizing the complexity of the Ziegler systems.

Although the idea of a universally applicable mechanism for

Ziegler polymerization of olefins consistent with all

available data is attractive, the body of experimental

evidence indicates that this cannot be achieved. Perhaps

this explains the continuing effort to supply more data to

provide a clearer picture of the catalyst's mode of action.

Several aspects of the mechanism of Ziegler type

polymerization were not completely clarified by the plethora

of reports that had appeared by the mid 1970's.

Experimental evidence was needed to support two basic

mechanistic steps in the polymerization. First, evidence

concerning the formation of an olefin w-complex was lacking.

Secondly, no observation of the sequential insertion of an

olefin into a metal carbon bond had been reported.

As previously mentioned, coordination polymerization

implies that an interaction of the transition metal with an

olefin through a Y-complex occurs prior to insertion. While

examples of this were abundant for the later transition


elements, few reports of such interactions for the early

transition metals had appeared. Only one example of a fully

characterized vanadium ethylene complex had been

published.43 Another study of the equilibrium constants for

the binding of alkenes to zirconium and hafnium complexes

attached to a support had been reported but no structural

information was attained and the complexes were not

isolated.44 In 1983, the first example of a fully

characterized titanium ethylene r-complex was reported by

Cohen et al.45 A few publications of other early transition

metal olefin r-complexes have appeared46 and the first

isolated ethylene bridged complex containing a lanthanide

metal was reported in 1987.47 The lack of d electrons in

electron deficient metals results in low stabilities for

olefin r-complexes, because no back donation is possible.

The generally accepted mechanism for polymerization

indicates that the polymer remains attached to the metal

throughout the polymerization process and olefin units are

added by repeating the coordination of monomer and insertion

steps. No transition metal study has yielded the

observation of the stepwise insertion of an alkene.

However, a highly hindered alkyne was shown to undergo

insertion into the Ti-CH3 bond generated from the soluble

two component Ziegler type system Cp2TiCl2/AlMeCl248 As

a relevant digression, the system CP2TiCl2 and AlEt3 and

related versions of this combination were studied


extensively in the early development of Ziegler catalysts as

a homogeneous analogue.1,21,49 The zirconium analogue was

also investigated and indicates the relevance of the present

study. The proposed cationic nature of the active site is

compared to an isolated zirconium alkyl cation well

characterized in solution and studied in the gas phase.

As early as 1975 lanthanide complexes were shown active

for ethylene polymerization with and without added main

group compounds.50,51 Systems derived from bridged dimeric

species often polymerized ethylene. However, the

possibility that the active species was monomeric could not

be ruled out.51 By a series of complex ligand displacement

reactions a combination that generated a significant amount

of a monomeric methyl species was found.52 The one

component lutetium alkyl was reported by P. Watson and

resulted in an excellent model system for the Ziegler

polymerization (Scheme 1-1).53 This system manifested many

of the proposed transformations for initiation, propagation,

and termination. It rapidly polymerized ethylene to linear

polyethylene. The expected product from the insertion of

propylene into the Lu-CH3 was also observed. Allylic C-H

activation predominates after the initial insertion of

propylene and a second insertion occurs only to a minor

extent. B-hydride shift/alkene elimination54 as well as

B-alkyl shift alkene elimination sequences are also

indicated. In addition, the hydride generated by

Scheme 1-1


Cp*2Lu-Me + H2C=CHMe --- CP*2Lu-CH2CHMe2


Cp*2Lu-CH2CHMe2 + H2C=CHMe -- Cp*2Lu-CH2CH(Me)(R)

R = CH2CHMe2

Chain Termination by B-Hydride Elimination

Cp*2Lu-CH2CHMe2 -- Cp*2Lu-H + H2C=C(Me)2


Cp*2Lu-H + H2C=CHMe --- Cp*2Lu-CH2CH2Me

B-elimination rapidly inserts propylene yielding evidence

for the control of molecular weight through this process.

The lanthanide alkyls investigated also react with

hydrogen to yield alkanes and the Lu-H, again providing

evidence for a mode of molecular weight control by addition

of H2. The addition of coordinating bases to the lutetium

complex strongly inhibits the polymerization of olefins.

Although no evidence for a discrete r-complex was reported,

the inhibition of polymerization by adding coordinating

molecules is consistent with the necessity of an available

site for olefin coordination before insertion. The success

of this catalyst system greatly relies on the formation of a

sterically unhindered coordination sphere for the lanthanide


T. J. Marks and coworkers have also made tremendous

strides in the area of Ziegler model systems using

lanthanide and actinide alkyl complexes. Work concerning

several areas of catalysis have appeared from this group

including alkene polymerization, olefin hydrogenation and C-

H activation.23,41,55-67

The chemistry of several lanthanide alkyl complexes and

their respective reactions with a number of substrates was

the subject of the rare appearance of three consecutive

publications from the same group in the Journal of the

American Chemical Society in 1985.41,55,56 These systematic

studies indicated that processes similar to those observed

in the lutetium system investigated by Watson were also

operative for other lanthanides. This similarity is not

unexpected because the different f orbital electron

configurations are not usually very important for ligand

interactions.16 Therefore, lanthanide metal structure and

reactivity is more often related to electrostatic and steric

factors, which are determined greatly by the ionic radii of

the lanthanide. Reactivities for ethylene polymerization

were shown to decrease with the size of the lanthanide metal

center i.e. La > Nd >> Lu. The rate of polymerization for

Cp*2LaH2 and Cp*2NdH2 reveal that they are two of the most

active homogeneous ethylene polymerization catalysts.

In addition, reactions with propylene, hexene and

cyclohexene are reported. These substrates are not

polymerized. However, insertion followed by C-H activation

yields q3 allyl complexes. Reactions with cyclohexene are

slower and are tentatively reported to yield cyclohexenyl


The most recent development by Marks and coworkers in

the area of olefin polymerization results from the

interaction of Cp2Th(CH3)2 with a MgCl2 support. Magic

angle spinning NMR indicates the formation of a cationic

thorium methyl complex similar to that of Cp2ZrCH3+ which is

observed to polymerize ethylene.23,42 The cationic nature

of the thorium supported complex is inferred from previous

results68'69 and its reactivity toward polymerization is

much greater than the corresponding dialkyl complex.

The Marks research group has also been active in the

area of thermochemistry.70,71 The paucity of thermochemical

data pertaining to metal-carbon bonds leads to difficulty in

understanding the nature of reactions. Therefore, studies

in this area are particularly useful in comprehending the

extent to which reactivity is determined by kinetic and

thermodynamic factors. A report of thermodynamic parameters

for polymerization of a number organic monomers in solution

has appeared.72

Recently, highly electrophilic scandium alkyl and

hydride compounds (isoelectronic, neglecting f electrons,

with the lutetium alkyl complex) were synthesized and found

to be extremely reactive toward polymerization of

ethylene.73'74 The highly electrophilic character of this

complex is considered extremely important in producing its

reactivity, which is similar to the lanthanides mentioned


R. F. Jordan and coworkers synthesized the first

cationic mononuclear transition metal ethylene

polymerization catalyst.75 Many zirconium, titanium and

hafnium alkyl complexes had been synthesized; however, their

activity in polymerization is not nearly as high as other

species such as the lanthanide and scandium complexes

previously cited.76 The zirconium complex, Cp2ZrCH3(THF)+,

polymerizes ethylene rapidly under mild conditions in non-

coordinating solvent.77 This breakthrough holds strong

implications for the mechanism and active species in Ziegler

polymerization. While this compound could not be isolated

as an unsolvated cationic species, early work yielded

evidence implicating the unsolvated Cp2ZrCH3+ species as the

site of activity.77

The complex was originally isolated as an acetonitrile

adduct Cp2Zr(CH3)(CH3CN)+. Recrystallization from THF

solution yields a THF adduct Cp2ZrCH3(THF)+. Interestingly,

a significant w-component was found for the interaction of

the oxygen lone pair electrons of THF with the zirconium

center which orients the THF molecule in a sterically

unfavorable position. The electronic stability associated

with the structure overshadows the steric interactions of

the THF with the Cp rings. In THF solution the zirconium

cation did not polymerize ethylene. Yet in a non-

coordinating CH2C12 solvent formation of polyethylene occurs

rapidly under mild conditions. The observations indicate

that dissociation of THF is necessary for polymerization to

take place. The dissociation is more readily achieved in

non coordinating solvent. Since the THF adduct is labile,

dissolving the complex in CH2C12 decreases the amount of THF

adduct present due to simple equilibrium considerations.

Analysis of the polymer produced in the zirconium

cationic system gives evidence for chain termination by B-

hydride shift alkene elimination.20 This chain termination

sequence is consistent with that of Ziegler Natta

polymerization systems.

A cationic zirconium benzyl complex Cp2Zr(CH2Ph)(THF)+

has also been synthesized.78 The conformation of the benzyl

ligands in electrophilic metal benzyl complexes have been

shown to have a unique structure,6,7 i.e. the benzyl ligand

adopts an q2 benzyl structure. Jordan and coworkers

reasoned that this interaction might stabilize the

unsolvated benzyl species and allow its observations in

solution. Indeed, when synthesized the benzyl cation does

have the n2 type benzyl zirconium interaction and H1NMR

evidence indicates that the structure is retained in

solution. The CH3CN adduct Cp2Zr(CH2Ph)(CH3CN) does not

polymerize ethylene (another indication of the necessity of

an available coordination site). Synthesis of the benzyl

cation in THF yields the THF adduct Cp2Zr(CH2Ph)(THF)+ and

in CH2Cl2 the THF adduct does polymerize ethylene apparently

because a shift in the equilibrium increases the

concentration of the unsolvated benzyl cation. NMR spectra

of a dilute solution of the THF adduct in CH2Cl2 yields

resonances corresponding to the "naked" benzyl species.

Reactivity of Cp2ZrCH3+(L) (L = PMe3, THF, CH3CN)

cations has also been investigated in reactions with H2.

The nature of L has a profound effect on the rates of

hydrogenation.79 The bis acetonitrile adduct


Cp2ZrCH3(CH3CN)2+ is found in CH3CN solution and does not
react with H2 significantly. The rate of hydrogenation for

the THF adduct is faster in CH2C12 than in THF. The PMe3

adduct can be produced and rapidly exchanges with free PMe3

in THF and CH2Cl2. Correspondingly, the rate of

hydrogenation is very rapid.

The results are consistent with significant interaction

of H2 with a vacant LUMO on the zirconium center.80,81 This

is also the case for the zirconocene alkyl hydride and is

consistent with the process of chain termination during

polymerization by addition of H2. The zirconium cation

system is therefore another excellent model system for the

study of Ziegler type polymerizations.

In addition, studies have also appeared regarding a

similar catalyst system derived from Cp2ZrCl2 and methyl

aluminoxanes82,83. The conclusion that the active species in

the system are the cations Cp2ZrCH3+ and Cp2ZrH+ is

supported by x-ray photoelectron spectroscopy results.82

The indication of an active site comprised of a similar

cationic titanium complex had been proposed over 20 years

ago.84,85 Very recently a cationic chromium complex has

been synthesized and readily polymerizes both ethylene and

propylene under suitable conditions.86

Carbon-Hydrogen Activation
The proposition of converting alkanes and methane into

other useful molecules is indeed exciting, especially since


methane is a relatively abundant feedstock. Unfortunately,

these compounds, comprised of only carbon and hydrogen, are

notoriously unreactive. The inert nature of hydrocarbons

makes functionalization a difficult task.17

The first step in achieving the goal of synthesizing

useful molecules from saturated hydrocarbon substrates is to

find complexes that can "activate" C-H bonds. The term C-H

activation simply means to generate a more reactive metal-

carbon bond from a less reactive C-H bond.4 There are two

general ways to accomplish C-H activation. One is by using

a later transition element which normally reacts with a

hydrocarbon substrate via oxidative addition.27,17,3

Another way to achieve C-H activation is with an

electrophilic lanthanide, actinide or early transition

element. The latter situation is most applicable to

previously discussed compounds and will be the focus of this


The relationship between the author's work and the

objectives of C-H activation is necessarily fundamental.

The main thrust of the present work is to present

information regarding the factors which cause the C-H bond

of alkanes, alkenes, aromatics, and alkynes to be broken and

concomitantly form other C-H bonds. The opportunity to

investigate features of this reactivity in the gas-phase are

valuable not only because results may further the


understanding of C-H activation but because the effects of

solvation on this process may also be studied.

Electrophilic dO and dOfn alkyl complexes are not only

excellent model systems for the study of Ziegler catalysis

but are also highly reactive toward the activation of C-H

bonds.2-4,17,18,23,41,56,73,74,87-89 Reviews concerning C-H

activation by these compounds are scarce. However, many

examples of both intermolecular and intramolecular C-H

activating reactions have been observed for these complexes.

In many instances these reactions have been studied by the

same research groups with the same systems previously cited

in some detail with respect to Ziegler polymerization.

Therefore the objective in this section is only to alert the

reader to the fundamental mechanistic aspects of C-H

activation as they relate to the present study and not to

present a detailed review.

P.L. Watson was the first to observe C-H activation for

an electrophilic metal.90 The reaction was observed for

Cp*2LuCH3 and 13CH4 as shown in equation 1-1. The proposed

Cp*2Lu-CH3 + 13CH4 --- Cp* 2Lu-13CH3 + CH4 (1-1)

mechanism for this reaction involves a four-membered

transition state similar to that proposed for insertion of

olefins into M-C bonds (Figure 1-1). This process is

thermoneutral and still occurs at a significant rate

CH e


Figure 1-1. The four membered transition state for
activation of the C-H bond of methane.


emphasizing the reactivity accessible for such unsaturated

electrophilic complexes.

The study of the Cp*2Sc-H and Cp*Sc-CH3 reactivity

toward hydrocarbons73,74 is didactic when considering the

reactivity of Cp2ZrCH3+ in the gas-phase. Activation of C-H

bonds in alkenes other than ethylene by lanthanide complexes

are proposed to occur at the weakest C-H bond. At least two

systems are reported to undergo reactions with18,55

propylene in which allyl complexes are formed by allylic C-H

activation (Figure 1-2).

In contrast to these results, Thompson et al.74 have

made a strong case for a kinetic ordering of reactivity for

hydrocarbon substrates in which the activation of vinylic C-

H bonds in alkenes occurs more rapidly than allylic C-H

activation. The study coined the phrase "a-bond metathesis"

to describe the process because the fundamental

transformation observed is the exchange of the original R

group on the metal with R' in R'H substrates as shown in

equation 1-2. A four-membered transition state analogous to

that shown in Figure 1-1 was invoked to describe the

transformation (Figure 1-2).

Cp*2Sc-R + R'-H -- Cp*2ScR' + R-H (1-2)

The concept developed by Thompson et. al. indicates

that the main interaction responsible for the kinetic




S-Cp*2Ln< H


--- Cp 2Ln )CH + RH


Figure 1-2. Activation of allylic C-H bond of
propylene by lanthanide alkyls. (Ln =
La, Nd, Lu and R = CH3)


effects observed arises from the s orbital character of the

hydrocarbon bond being activated. Arguments presented are

based mainly on orbital overlap considerations in the

framework of the four-membered transition state. The rate

of reaction for a particular substrate is not governed by

thermodynamic considerations but by the ability of the

reacting molecules to assume efficient orbital overlap in

the transition state. Thus, in a valence bond picture, the

rate of reaction correlates with the amount of s character

in the C-H bond. For example, alkyne C-H bonds are

activated more rapidly than vinylic C-H bonds. In valence

bond terms the order of reactivity for C-H bonds of R in

equation 1-2 is R = R'= H >> R = alkyl, R'= H >> R = alkyl,

R'= sp C-H > R = alkyl, R' = sp2 C-H > R = alkyl, R'= sp3 C-


Detailed kinetic data for the lanthanide systems is not

available but as previously mentioned orbital considerations

are rarely necessary to explain reactivity for these metal

complexes. This provides an explanation for the observed

activation of allylic C-H bonds.

Gas-Phase Reactions

In the past 20 years the development of gas-phase metal

ion chemistry has progressed greatly. Both ion beam and ion

cyclotron resonance mass spectrometry techniques have proven

excellent means for probing the chemistry of bare metal ions


and other ionic organometallic species. Two recent reviews

deal explicitly with bare metal ion chemistry.91'92

Most work in this area is based on reactions of bare

metal singly charged cations. Although Freiser and

coworkers have also published work concerning reactions of

bare metal alloy dimers and doubly charged metal ions.93,94

The target substrates in bare metal ion studies are a

variety of organic molecules including alkanes, alkenes,

alkyl halides, alcohols, aldehydes and ketones.91 By far,

the most studied organic compounds are alkanes.

The observation of the insertion of bare metal ions

into both C-H and C-C bonds has created an entire area of

chemistry that is continually being investigated with

increasingly complex experimental techniques.92'95 The

relationship of the gas-phase organometallic chemistry to

catalysis of both alkene polymerization and C-H activation

is obvious, since insertion of unsaturated hydrocarbons and

C-C and C-H activation were observed as the main processes

occurring in gas-phase systems. A truly formidable number

of publications have resulted from efforts in this

relatively new area of organometallic chemistry. For

example, the Freiser research group published fourteen

papers in 1985 alone in the Journal of the American

Chemistry Society. However, the usefulness of this work for

the solution phase organometallic chemist is dubious,

mainly because of the lack of solution phase analogues to


gauge the reactivity of the gas-phase ions and effect a

change in the way the solution phase organometallic chemist

views gas-phase studies.

Another drawback to gas-phase organometallic chemistry

results from the lack of direct structural probes. At

present, collision induced dissociation (CID), infra-red

multiphoton dissociation (IRMPD), and reactivity results are

the main tools for determining the structures of gas-phase

ion molecule reaction products. The lack of structural

probes leads to speculation rather than proven ion

structures, which often produces highly speculative reaction

mechanisms. For instance, the mechanistic process invoked

nearly exclusively for transformations at bare metal ion

centers and explained in depth by Allison91 is metal

insertion (into C-C or C-H bonds)/8-hydride

shift/competitive ligand loss. This sequence has proven

extremely useful in explaining a variety of bare metal

ion/molecule reactions but sometimes is not used

judiciously. Until recently, the process of oxidative

addition followed by reductive coupling has been used

commonly and favored over other processes such as multi-

centered concerted activation of C-H bonds.95 However,

recent theoretical treatments indicate that d orbitals on

metal-ions may facilitate the latter.96

Several studies relate directly to the catalytic

Ziegler polymerization.97-102 TiCl4 and TiC13CH3 were used


to study reactions of alkenes and some insight regarding the

Ziegler systems could be gleaned from the results but

ethylene was not polymerized in the gas-phase. Insertion of

ethylene into the Ti-Me bond has been observed. However,

the loss of H2 generates an allyl species which does not

undergo additional insertion.

Binuclear titanium chloride ions were also observed as

ion/molecule reaction products. So two pertinent

observations can be made from these results. First,

insertion of ethylene does occur although propagation by

additional insertion is precluded by formation of an allyl

species. Second, binuclear ions were also observed

indicating that the possibility of deactivation of the

catalyst by formation of binuclear species does occur in the


Recently, Freiser and coworkers have investigated the

reactions of FeCH3+ and CoCH3+ generated from ion molecule

reactions of Fe+ of Co+ with CH3.99 Reported reactivity of

these species with unsaturated hydrocarbons is rather

pertinent to the present study, but comparisons must be made

very carefully because several extremely important

differences exist. First, both FeCH3+ and CoCH3+ do not

contain dO metal centers. Therefore, the process of

oxidative addition is viable. Secondly, they are both

highly unsaturated, where as the Cp2ZrCH3+ complex cation is

unsaturated but to a much lesser degree. Thirdly, the


greater degree of unsaturation in the FeCH3+ and CoCH3+ ions

removes steric interactions almost completely. Both the

metal and the CH3 group are easily accessible to an incoming

substrate molecule. Finally, the thermodynamics of bond

formation are greatly changed. Not only are interactions

such as back donation for alkene substrates possible,

yielding more stable r-complexes, but the fundamental

ordering of CH3 and H bond strengths is reversed.

One important commonality for the two systems is that

they are both studies in the gas-phase under similar

conditions. Thus, careful comparisons yield insight into

the effects of the differences on chemical reactivity.

Theoretical Studies

The importance of Ziegler-Natta polymerization and C-H

activation have not escaped the theoretical chemist.80,103-
110 Unfortunately, the complexity of the catalyst prevents

facile high level calculations. The general approach of the

theoretical chemist has been to model somewhat simplified

one-component systems. Rabba and Hoffmann have produced

excellent discussions using extended Huckel theory. Since

detailed information is given regarding the molecular

orbitals available for bonding interactions these

publications are particularly helpful when evaluating

possible structures for ions.104 The theoretical articles

are also useful for obtaining general background information


because they rely upon experimental results to gauge their


Rappe et al. have also treated these catalytic

processes for electron deficient complexes at the ab initio

level. The higher level calculations in these studies offer

some thermodynamic information.106-108 The calculations

have added significance for the gas-phase chemist because

solvation thermodynamics are not included in the treatment.

Therefore comparisons can be made directly.


An attempt has been made to set the stage for the

reader to evaluate both the results and conclusions of the

present study by including several areas of background

information in this introduction. The majority of the

discussion deals with solution organometallic chemistry

rather than gas-phase organometallic chemistry. The

reasoning behind this representation derives from the

author's conviction that gas-phase studies can truly be

complementary and important to condensed phase

investigations. This conviction does not belittle the

significance of pure gas-phase investigations because

clearly an in depth understanding of any type of reactivity

does not need to be widely applicable or industrially

feasible to be an asset in the search for a more clear

picture of fundamental chemical transformations.

Nevertheless, the lofty goal of producing information


relevant to more than a single chemical community is a

worthwhile goal. Therefore, significant effort has been put

forth to bring the relatively unconnected areas of gas-phase

and condensed phase organometallic chemistry together by

studying, in the gas-phase, reactions of a species well

characterized in solution. The main differences are the

lack of solvation and the restrictions of the gas-phase




We applied FTICR Mass Spectrometry11l-113 to study the

reactions of Cp2ZrCH3+, I with H2, HD, D2, and a number of

alkenes and alkynes. Cp2ZrCD3 2, was also employed to

provide further insight into the mechanisms of these


The product ions are reported as m/z values in this

section and correspond to isotopic species having the most

abundant 90Zr isotope. Zirconium has six principal

isotopes. Contributions from isotopes of carbon and

hydrogen result in the observation of seven isotopic species

in the mass spectra of compounds containing zirconium,

carbon, and hydrogen. A typical mass spectrum for

Cp2Zr(CH3)2 (ionization energy = 12 eV) is shown in Figure

Empirical formulae of the form CP2ZrCxHy+ are given

for product ions in this section, and no structure is

implied other than the CP2Zr unit. Any eliminated neutrals

are taken as the most stable species possible with the

correct molecular formula. Detailed structures of the

Cp2Zr+ and Cp2ZrCH3+ generated by El of Cp2Zr(CH3)2 VBM -12V


225 230




Figure 2-1.

Mass spectrum of Cp Zr(CH3)2. Electron
impact ionization at 12 eV.




Cp+Zr+ Cp2ZrCH,+

-- ----. 1 0

0 I


zirconium product ions are presented in the Discussion


Determination of Product Ion Distribution

Some reactions of 1 yield more than one product ion.

In cases where the intensity of a m/z value has

contributions from more than one product ion, it is

necessary to quantify the contributions of the different

product ions to a single m/z value. Theoretical isotope

ratios for each product ion were calculated using known

isotopic abundances for the elements. In every case the

intensity of at least one m/z value results from a single

isotopic species; therefore, its contribution to the

observed intensity can be subtracted and the product

distribution determined.

Product ion distributions were calculated from at least

five to as many as thirty measurements at several different

reaction thermalization times and substrate pressures.

Constant product distributions measured under various

conditions suggest that the reactions are studied under

thermal conditions. The resulting distributions are

reported as percentages of the total product ion intensity

and include 95% confidence limits.

Experiments yielding products with overlapping peaks

were conducted without ejection of Cp2Zr This precaution

prevents any extraneous distortion of Cp2ZrCH3+ isotope

ratios that might arise due to undesired excitation during

ejection of Cp2Zr+ and affect product ion abundances.

Reactions of Cp2Zr+ with substrate were also done to ensure
that the presence of Cp2Zr+ did not affect the
interpretation or measurement of product ion intensities for

reactions of 1.
Reaction of Cp2ZrCH3+ with H2 and HD

We previously reported the reaction of Cp2ZrCH3+ with
D2 and the corresponding second order rate constant kD = 3.9
0.5 x 109 M-1s-1 (6.5 0.8 x10-12 cm3s-1) (eq 2-1).114

Cp2ZrCH3+ + D2 Cp2ZrD+ + CH3D (2-1)
(m/z 235) (m/z 222)

We have measured the analogous rate for reaction of I with
H2, which yields kH = 7.9 1.8 x109 M-is-1 (1.3 0.3
x10-11 cm3s-1). These two rate constants give a deuterium

kinetic isotope effect, kH/kD 2.
Reactions of Cp2ZrCHII+ with C-24 and C2D4

The reaction of I with C2H4 proceeds with the
formation of m/z 261 and loss of H2 (eq 2-2). Similarly, 2
reacts with C2D4 to produce m/z 266 and D2 (eq 2-3).

Cp2ZrCH3+ + H2C=CH2 Cp2ZrC3H5+ + H2 (2-2)

(m/z 235) (m/z 261)

Cp2ZrCD3+ + D2C=CD2
(m/z 238)

Cp2ZrC3D5 + D2
(m/z 266)

H/D scrambling is observed for reaction of I with C2D4

and 2 with C2H4 (eqs 2-4, 2-5). The H2, HD and D2

scrambling ratios for the reactions in eqs 2-4 and 2-5 were

obtained by first measuring the isotopic abundances of all

the isotopes for the products of the reactions in eqs 2-2

and 2-3, respectively. The measured isotope abundances for

m/z 261, and m/z 266 show little variation and often differ

slightly from the calculated abundances; therefore, the most

accurate product distributions are determined by using

experimentally measured abundances.

Cp2ZrCH3+ + C2D4
(m/z 235)

Cp2ZrCD3+ + C2H4 --
(m/z 238)

Cp2ZrC3HD4 + H2
(m/z 265)

Cp2ZrC3H2D3++ + HD (2-4)
(m/z 264)

Cp2ZrC3H3D2+ + D2
(m/z 263)

Cp2ZrC3H2D3+ + H2
(m/z 264)

Cp2ZrC3H3D2 + HD (2-5)
(m/z 263)

Cp2ZrC3H4D+ + D2
(m/z 262)


We previously reported the relative amounts, 15% H2,

74% HD, and 11% D2 elimination for the reaction of Cp2ZrCH3

with C2D4 and 23% HD and 77% H2 elimination for reaction of

Cp2ZrD+ with C3H6. For completeness, the product
distribution was determined for reaction of 2 with C2H4.

Thus for the reaction shown in eq 2-5, 7% D2, 58% HD, and

35% H2 are eliminated. The results for these reactions

indicate that significant H/D scrambling occurs prior to

elimination of dihydrogen. If H2, D2 and HD were eliminated

statistically 14.3% D2, 57.1 % HD and 28.6% H2 loss is

expected. The measured product distributions for the

reactions indicated in eqs 2-4 and 2-5 show that the amount

of D2 loss is less than statistically predicted in both

cases. The results for the reaction of Cp2ZrD+ and C3Hg are

much closer to the statistically predicted ratio; however,

no D2 can be eliminated in this reaction.

Reaction of Cp2ZrCH3+ and C2ZrCD3+ with Alkenes

The product distributions for reactions of 1 with a

variety of a-olefins are summarized in Table 2-1. For all

but one substrate at least one neutral molecule is

eliminated. The exception is allene, which yields three


Reactions of Cp2ZrCD3+ involving H2 and CH4 elimination
show unique behavior. In all reactions resulting in H2

elimination scrambling is observed, and H2, HD and D2

elimination occurs, except for the reaction of 2 with allene

Table 2-1.

li i4 t4 -

Tlv Ftw l

Dnri t-4 ibti %IV4- \

wwwwwwww m na e on ormu a s r u n











a-Methyl Styrene













Cp2ZrC7H13 +












87 5

13 5

52 7

48 7




62 4

6 3

31 2

in which no D2 elimination is observed. In contrast, for

reactions where CH4 is eliminated no scrambling of the

deuterium label is detected, i.e., only elimination of CH3D

is observed.

For example, in the reaction of I with allene three
products are concomitantly produced. Double resonance

experiments indicate that each is formed directly from i and

allene (eq 2-6).

Investigation of the allene reaction by using Cp2ZrCD3+

reveals exclusive loss of CD3H producing Cp2ZrC3H3+

(m/z 259), and, for the loss of H2, significant H/D

scrambling results in the elimination of H2, and HD;

however, no D2 elimination is observed (eqs 2-7, 2-8).

Cp2ZrC3H3+ + CH4
(m/z 259)

Cp2ZrCH3 + C3H4 Cp2ZrC4H7 (2-6)

(m/z 275)

Cp2ZrC4H5+ + H2
(m/z 273)

Cp2ZrCD3+ + C3H4 Cp2ZrC3H3+ + CD3H (2-7)
(m/z 238) (m/z 259)


--Cp2ZrC4H3D2 + + HD

(m/z 275)

Cp2ZrCD3+ + C3H4 (2-8)
(m/z 238)
Cp2ZrC4H2D3+ + H2

(m/z 276)

Isobutene and a-methyl styrene react more slowly than

the other terminal olefins investigated. A comparison of

the half lives for the reactions of 1 with ethylene and

isobutene indicate that the ethylene reaction is

approximately seven times faster than the isobutene

reaction. In contrast to other alkenes studied, the neutral

product for reaction of i with isobutene and a-methyl

styrene is methane rather than dihydrogen (eqs 2-9, 2-10).

Cp2ZrCH3+ (m/z 235) is also formed in the reaction of
Cp2ZrCD3+ with isobutene (eq 2-11). Only trace C2ZrCH3+

and Cp2ZrPh+ are formed in the reaction of 2 with a-methyl


Cp2ZrCD3+ + H2C=CMe2 Cp2ZrC4H7+ + CD3H (2-9)
(m/z 238) (m/z 275)

Cp2ZrCD3 + H2C=C(Me)(Ph) Cp2ZrC3H4Ph+ + CD3H (2-10)
(m/z 238) (m/z 275)

Cp2ZrCD3+ + H2C=CMe2 Cp2ZrCH3+ + C4H5D3 (2-11)
(m/z 238) (m/z 235)

The reactions of I with fluoroethylene and 1,1,1-

trifluoropropene were also investigated. The single product

ion in both reactions is CP2ZrF+ (eqs 2-12, 2-13).

S+ H2C=CHF Cp2ZrF + H2C=CH(CH3) (2-12)
(m/z 239)

S+ H2C=CH(CF3) Cp2ZrF+ + H2C=CH(CF2CH3) (2-13)
(m/z 239)

The product distributions for reaction of 1 and a

number of internal olefins are listed in Table 2-2. The

rates for reactions of I with internal olefins including

cyclohexene and 1,5-cyclooctadiene indicate that gis olefins

generally react more slowly than trans olefins, e.g. half

lives for reaction of I with cis and trans-2-pentene show

that the gci isomer reacts approximately three times more

slowly than the trans isomer. In addition, only small

amounts of CH4 are eliminated for gci olefins in contrast to

nearly exclusive loss of CH4 for the trans isomers.

Reaction of CP2ZrCH-3 and CP2ZrCD3+ with Alkynes
The reactions of I with several alkynes were

investigated and are listed in Table 2-3. Most of the

Table 2-2.

lu"ctra- +-

li i t dT A ..,+ ,

wv nOr~. be A. m..lJIIIm Cn. e



Ion Formula
































78 +

22 +

_ m J

Table 2-3.


v1 .i r a 4-A^


Ion Formula


Distribution( *





















* trace amount

61 3

39 3


+ 3

+ 5


90 4

10 4

Rnhr~rrrwrtY-rL~I L tal~i

reactions were also repeated with Cp2ZrCD3+ as the reactant


Acetylene reacts with 1 to yield Cp2ZrC3H3+ (m/z 259)

and eliminates H2 (eq 2-14). The product ion, Cp2ZrC3H3 ,

does not react detectably with acetylene substrate. The

second order rate constant for formation of Cp2ZrC3H3+,

determined from six measurements, is 5.6 1.9 x1010 M-1s-1

(9.3 3.2 x10-11 cm3s1).

1 + HCmCH Cp2ZrC3H3+ + H2 (2-14)

(m/z 259)

The reaction of propyne is more complex than that of

acetylene. Initially two products are observed, Cp2ZrC3H3+

(m/z 259), and Cp2ZrC4H5+ (m/z 273). Double resonance

experiments indicate both products are formed from 1 and

propyne directly (eq 2-15).

Cp2ZrC3H3+ + CH4

(m/z 259)

1 + CH3CaCH (2-15)

Cp2ZrC4H5+ + H2

(m/z 273)


In addition, CP2ZrC3H3+ (m/z 259) reacts further with

propyne to produce Cp2ZrC6H7+ (m/z 299), and no neutral
molecule is eliminated (eq 2-16). CP2ZrC6H7+ (m/z 299) is

inert to subsequent detectable reaction with propyne.

Cp2ZrC3H3 + HCC-CH3
(m/z 259)

(m/z 299)

Product ion Cp2ZrC3H3+ is produced with exclusive loss

of CD3H in the reaction of propyne with CP2ZrCD3+ (eq 2-17).

The formation of Cp2ZrCH3+ is also observed along with the

products from elimination of H2, HD, and D2 (eqs 2-18,


Cp2ZrCD3+ + HCaC-CH3
(m/z 238)

Cp2ZrCD3+ + HCmC-CH3
(m/z 238)

Cp2ZrC3H3 + CD3H
(m/z 259)


Cp2ZrCH3+ + HCsC-CD3 (2-18)
(m/z 235)

Cp2ZrCD + + HCmC-CH


(m/z 274)

(m/z 275)

(m/z 276)

+ HD (2-19)

+ H2



The reaction of i with phenylacetylene is analogous to

the reaction with propyne, however, two additional products

are observed (eq 2-20, 2-21).

Cp2ZrCH3+ + HC=C-Ph Cp2ZrC3H4Ph (2-20)
(m/z 235) (m/z 337)

Cp2ZrCH3+ + HCwC-Ph CP2ZrPh+ + CH3CCH (2-21)

(m/z 235) (m/z 297)

Cp2ZrC3H4Ph+ is inert to subsequent detectable
reaction, and the small amount of Cp2ZrPh+ does not yield

observable reaction products.

Collision Induced Dissociation (CID)112 of Cp2ZrCH3,

CID experiments on Cp2ZrCH3+ (m/z 235) show the loss of

CH3 (eq 2-22). At higher fragmentation energies m/z 194 and

192 are observed. Only trace m/z 155 is produced in the CID

experiment. In addition, m/z 194, 192 and 155 are produced

in the CID of m/z 220.

Cp2ZrCH3+ Cp2Zr + "CH3 (2-22)
(m/z 235) (m/z 220)

CID of Product Ions

Collision induced reactions or rearrangements such as

the loss of H2 or unsaturated hydrocarbons are observed as


well as more direct fragmentation processes.115 In all

cases the fragment Cp2Zr+ (m/z 220) is observed. Product

ions with a given m/z value may be formed from different

substrates. A comparison of CID spectra for ions of common

m/z value often provides evidence for more than one distinct

structure. For example Cp2ZrC4H7+ (m/z 275) is produced

from propylene, allene, isobutene, trans-2-butene, and cis-

2-butene (eqs 2-23 2-27).

IiL + 03H6

1 + C3H4

1 + iso-C4H8

1 + trans-2-C4Hg

1 + cis-2-C4Hg

(m/z 275)

(m/z 275)

(m/z 275)

(m/z 275)

(m/z 275)

CID of the product ions with m/z 275 in eqs 2-23 2-25

produce m/z 235 and m/z 220 (eq 2-28) at approximately the

+ H2

+ CH4

+ CH4

+ CH4






same excitation energy and in roughly the same proportions,

whereas CID of m/z 275 in eq 2-26 produces m/z 221 and m/z

220 (eq 2-29). CID of m/z 275 in eq 2-27 could not be

obtained due to its low abundance.

-Cp2ZrCH3' + C3H4

Cp2ZrC4H7 (m/z 235) (2-28)
(m/z 235)
Cp2ZrC4H7+ (2-28)

(m/z 275)
-z 2) Cp2Zr-+ + C4H7

(m/z 220)

Cp2ZrH+ + C3H6
(m/z 221)

Cp2ZrC4H7+ (2-29)
(m/z 275)

Cp2Zr+ + C4H7
(m/z 220)


The reactions of I with olefins and alkynes can be best

described in the context of two major reaction sequences:

the migratory insertion of unsaturated hydrocarbons into the

zirconium-methyl bond, and the activation of C-H bonds via

a-bond metathesis, (Scheme 2-1, 2-2 respectively). Both

reaction sequences are well precedented in the solution


reactions of similar neutral and cationic, electron

deficient dOfn metal complexes.73,74,23,41,52-63 Metathesis

in this study is indicated by reactions of Cp2ZrCD3+ with

hydrocarbons that produce CD3H as the sole eliminated


Scheme 2-1

H3C +

P2 3"CHR\ 1


- Cp2Zr2+ CHR

Scheme 2-2

Cp2ZrCD3+ + R-H

- Cp2Zr 2H

SCp2ZrR+ + CD3H

An attempt is made here to establish that insertion and

C-H activation processes are operative in the gas-phase

reactions of Cp2ZrCH3+ with unsaturated hydrocarbons.

However, it is necessary to identify certain aspects of the

gas-phase medium that differ significantly from the

condensed-phase before arguments concerning gas-phase

reactivity can be made effectively.

One important difference in the two media arises from

the frequencies of third body collisions. In solution,


reacting species are nearly always in contact with solvent,

which acts as an energy sink by continually removing excess

energy generated during a reaction. In contrast, third body

collisions occur much less frequently in the low pressure

regime of the FTICR experiment (10-8-10-5 torr).

Scheme 2-3 depicts a general reaction sequence with two

pathways to products for an exothermic gas-phase

ion/molecule reaction.116-119 The following discussion is

simply an abbreviated overview of the principles governing

formation of products as they apply to the present system.

Scheme 2-3

kdec +
E + F

A+ + B -i (C*)+ --
k-1 C+

When reactants A+ and B interact in a bimolecular

exothermic reaction under low pressure FTICR conditions, the

energy of the reaction cannot be released to solvent and

therefore resides in (C*)+ as internal energy. The excess

internal energy of (C*)+ (< 40 kcal/mol) is randomized into

all vibrational modes.

Two pathways are available for product formation. The

activated complex (C*)+ may undergo unimolecular

decomposition or stabilization via third body collision to


produce C Infrared photon emission is another process

that results in the elimination of excess internal energy

from a chemically activated species, but will not be

discussed further. Unimolecular decomposition produces E+

and F or reactants A+ and B. The only observable

unimolecular process is the forward reaction yielding E+ and

F. The elimination of F in a translationally and

vibrationally excited state is likely and also provides a

means for removal of excess energy from the chemically

activated complex (C*)+. This type of product stabilization

is not limited to the elimination of a single neutral

molecule. If E+ remains vibrationally excited after neutral

elimination from (C*)+, further stabilization may occur by

additional neutral loss. The second pathway for

stabilization of the activated complex forms product ion C+.

In the strong collision limit,119 a single collision of M

with (C*)+ removes the excess energy necessary for

stabilization and observation of C+, i.e. after a third body

collision C+ no longer has the internal energy necessary to

undergo unimolecular decomposition at a significant rate.

Products analogous to (C*)+ are likely to form in the

gas-phase and decompose to form the reactants, if kdec is

small and the time between third body collisions is large in

comparison to the lifetime of the collision complex, (C*)+.

In order to observe product formation a stabilizing third

body collision is necessary unless energy can be removed by


the elimination of a neutral product. For example, no

product ion is observed in the reaction of CP2ZrH+ with

ethylene (eq 2-30). However, the reaction of Cp2ZrD+ and

Cp2ZrH+ + H2C=CH2 no products observed (2-30)

ethylene shows the formation of Cp2ZrH+ (eq 2-31). This is

explained by an insertion/deinsertion/elimination process.

Cp2ZrD+ + H2C=CH2 Cp2ZrH+ + H2C=CHD (2-31)

The ethyl insertion product,(C*) is not observed

indicating that kdec is small in comparison to k_1 and it

decomposes to reactants more rapidly than it is stabilized

by a third body collision. Even at the highest pressures

available in the FTICR experiment, kc[M] is small compared

to k-1.

Another important difference for the gas-phase FTICR

experiment results because the system is not closed, i.e.

reagent gases are leaked into and pumped through the high

vacuum chamber continually. Any neutral molecule formed in

an ion molecule reaction is rapidly pumped out of the

system. The removal of neutral products in this fashion

completely prevents the formation of products from neutrals

eliminated in previous reactions. Thus product ions are

produced only from neutral substrates admitted through inlet


leak valves. In addition, the number of ions present in the

ICR cell is in the range of 106.120,121 Therefore the

pressure of neutral molecules generated from an ion/molecule

reaction is extremely small (= 10-13-10-12 torr) compared to

the overall substrate pressure (10'-710-6 torr).

Structure of Cp2ZrCH3

The structure of I is of obvious importance when

considering its reaction chemistry. The crystal structure

of unligated Cp*2ScCH3,74 which is isoelectronic with cation

I indicates a pseudo-trigonal planar geometry around the

metal center, with the Cp ligands bent back. It is not

likely that 1 is a methylene hydride isomer, since such a

structure requires further oxidation of the dO zirconium

center. In addition, CID of i initially shows the loss of

CH3, (eq 2-22) affirming the metal-methyl structure and no

fragments corresponding to Cp2ZrCH2+ or Cp2ZrH+ are


In considering the reactions of I with hydrocarbons it

is important to characterize the extent of involvement of

the Cp ligand hydrogens. Complications in the mechanistic

interpretation of observed reactions and H/D scrambling

would result from participation of ring hydrogens in the

elimination of neutral molecules. In a prior publication we

assumed the "Cp2Zr" unit retains its integrity throughout

reactions with H2, ethylene and propylene; no direct

evidence supporting this assumption was presented.114


Evidence for the retention of the "Cp2Zr" unit and lack

of Cp ligand hydrogen involvement in observed reactions is

provided by the absence of elimination of HD or H2 for the

reaction of Cp2ZrCD3+ with C2D4. The reaction of Cp2ZrCD3+
with C2D4 yields Cp2ZrC3D5 and D2. The analogous reaction

of 1 with C2D4 indicates significant H/D scrambling (eq 6).

Reactions of Cp2ZrCH 3 with H2. D2, and HD

The observed kinetic isotope effect in the reactions of

H2 and D2 is not large. In the reaction, CH3 + D2/H2 -

CH3D/H + D/H, kH/kD = 4 for the linear transition state.122

A non-linear transition state leads to a lower kH/kD

value.120 Thus, the kH/kD value for reaction of I with

H2/D2 is consistent with a a-bond metathesis transition

Reactions of CP2ZrCH3+ with Terminal Olefins

Two general types of terminal olefins and allene are

discussed in this section. The first type has at least one

hydrogen in the 2 position. The results in Table 1-3

indicate that loss of H2 as the neutral product is observed

in every reaction of i with substrates in this group.

The polymerization of these a-olefin substrates might

be expected since Cp2ZrCH3(THF)+ and a number of complexes

similar to i polymerize ethylene.18,40,42'55'74'77 However,

we have pointed out insertion of ethylene into the Zr-CH3

bond generates a chemically activated [Cp2Zr(n-prop)+]*

complex, which undergoes an intramolecular rearrangement


resulting in the loss of H2 through a six-membered

transition state.114 We have further proposed that the

Zirconium product ion is an n3-allyl complex. The proposed

transition state may also be drawn in analogy to the

transition state for intramolecular a-bond metathesis, 3. A

similar transition state has been proposed for the reaction

of Sc+ with n-butane124 and Cp*2Ln-R with propene and


C2Zr- -CH


Conversion of the n3-allyl complex to an q1 isomer

before coordination of monomer has been proposed for

polymerization of ethylene by isolated transition metal

allyls.6 The isomerization of the n3-allyl to q1 cannot be

promoted by the associative coordination of solvent in the

gas-phase. Therefore, lack of further insertion is also

consistent with r3-allyl formation. Isomerization is

expected to be endothermic and have a significant barrier

even with olefin r-bond formation to the q3-allyl complex,

because conversion of the n3-allyl to 7l is accompanied by

loss of resonance stabilization and charge delocalization.


A distinct inhibition of insertion is associated with

structures that do not readily allow the coordination of

unsaturated substrates.77,53,125 Structures A and 5 are the

two most probable for the allyl cations resulting from the

loss of H2 in reactions of I with terminal alkenes larger

than propylene.

-+ -+
\ 2

Cp2Zr )C-R Cp2Zr )C-H


Cp*2ScCH3+ reacts with alkenes other than ethylene by

elimination of methane rather than insertion presumably

because of the sterically restricting Cp* ligands.74

Crowding is less pronounced in cation i and insertion is

apparently preferred. A scandium hydride complex with a

less crowded ligand sphere has been shown to insert and

isomerize a variety of pentadienes.126

The dehydrogenation of allyl ions generated from alkene

substrates greater than C4 has also been observed in the

gas-phase reactions of CoCH+ and FeCH3+ 127 In addition,

two molecules of H2 are eliminated in some reactions of bare

metal ions with alkanes in the gas-phase.128,129 The loss

of two molecules of H2 for reactions of I with alkenes is

most likely driven by the exothermicity of the insertion and

formation of charge delocalized metallocyclobutenes, 6, Z.

Several titanacyclobutene complexes have been isolated

and characterized.130,131 Clearly, our present observations

-+ -+

Cp2Zr )C-H Cp2Zr C-R

6 7

cannot be explained by successive oxidative additions/H2

eliminations, because oxidation of the dO zirconium center

is energetically prohibitive. Thus, oxidative addition/

reductive elimination mechanisms are not required to explain

our observations and the application of this mechanistic

sequence is open to question.

Hydrogen/deuterium scrambling is also observed in all

cases for reactions with CP2ZrCD3 Scrambling of the

deuterium label has been reported previously for the

reaction of C2D4 with 1 and C12TiCH3 .98 Scrambling is also

observed in the reaction of the highly coordinatively

unsaturated CoCD3+ with ethylene.127

The discussion of H/D isotope scrambling for Cp2ZrCH3+

and C2D4 is also applicable to the scrambling process for


all terminal alkenes studied (eq 2-32)

-- Cp2ZrC3HD4 + H2

(m/z 265)

1 + D2C=CD2 Cp2ZrC3H2D3 + HD (2-32)

(m/z 264)

Cp2ZrC3H3D2 + D2

(m/z 263)

Insertion of terminal alkenes other than isobutene and

a-methyl styrene yields migratory insertion products with at

least one B-hydrogen. The B-hydrogen is apparently rapidly

and reversibly transferred before elimination of H2 (eq 2-


H\ /H
Cp2Zr CCH3 Cp2Zr CH(CH3) (2-33)
CH2 \CH3 C2

This reversible B-hydride abstraction in conjunction with

1,2 hydrogen shifts along the carbon framework provide a

mechanism for H/D scrambling.

Schemes 2-4 and 2-5 indicate two pathways for H/D

scrambling and loss of D2. D2 elimination was chosen as an

example because it requires the greatest amount of

Scheme 2-4

Cp2Zr CH3


Cp 2Zr D


Cp2Zr ,






Scheme 2-5

Cp2 r


/ D D
Cp Zr v
2 D


Cp2 r/D H H



Cp Zr
H H -D


Cp2 Zr


Cp2Zr H




rearrangement, and HD and H2 elimination may be similarly


Both schemes require a B-hydride shift and the

formation of a carbonium ion on the carbon chain. The

energy required to generate the primary carbonium ion in

Scheme 2-4 should not be in excess of 16 kcal mol-1 (the

value of the energy difference for the 10 and 20 carbonium

ions generated from propane).132 The possibility of a

favorable interaction of the charge on the carbon framework

with the metal center and the stabilization generally

associated with charge dispersal in larger ions indicate

that the energy required to generate the 10 carbonium ion

intermediate is less than 16 kcal mol-1. Approximately 23

kcal mol-1 is associated with formation of CP2Zr(n-prop)+

from I and ethylene and subsequent B-hydride shift reduces

the excess energy of the complex to 5-20 kcal mol-1.107,108

Thus formation of the 10 carbonium ion may not be

thermodynamically feasible.

Scheme 2-5 also requires the formation of a carbonium

ion, but in this case the charge is centered in the B-

position with respect to zirconium and thus is stabilized by

the polarization of the metal center. The same polarization

is apparently important in lowering the energy of the

transition state for migratory insertion and B-hydride



Scheme 2-5 indicates formation of a zirconium-carbon

bond in the 2 position of the three carbon chain. This

intermediate is reminiscent of the insertion product that

would be produced from anti-Markovnikov addition of

propylene to CP2ZrH The reaction of I with isobutene
indicates that Markovnikov addition is strongly preferred

but, the same result does not necessarily follow for the

reaction of Cp2ZrH+ with propylene.

Schemes 2-4 and 2-5 are consistent with experimental

results but involve intermediates that may be energetically

unfavorable, therefore neither mechanism is clearly favored

to explain the observed scrambling.

The second category of terminal olefins is represented

by isobutene and a-methyl styrene. Both substrates react

with I to eliminate methane and apparently form 73-allyl

complexes (.,9 R = CH3, Ph) (eq 2-34).

/CHe3 H2CC |
1 + H2C=C Cp2Zr H2C-R (2-34)

The assignment of allyl structures for isobutene and a-

methyl styrene reaction products is based on CID spectra

that indicate these ions are identical to the products

formed in the reaction of 1 with propylene and styrene


Qualitatively these reactions occur more slowly than

reactions of other terminal alkenes. Both substrates react

with Cp2ZrCD3+ to eliminate CD3H exclusively. An analogous

reactivity is seen in solution for Cp*2ScCH3 in reactions

with several alkanes, alkenes and alkynes.74 This process

has been termed a-bond metathesis and occurs rather than

insertion because of the steric interaction of the reacting

substrate with the ligand environment. The ligand imposed

steric requirements for insertion of alkenes other than

ethylene prevent the facile, proper orientation of the C=C

double bond necessary for migratory insertion.

In the reaction of CP2ZrCH3+ with isobutene and a-

methyl styrene, both vinylic and allylic C-H activation are

possible as depicted in Schemes 2-6 and 2-7. Vinylic C-H

activation of isobutene and a-methyl styrene implies the

Scheme 2-6

+ H
+ +H
Cp2ZrCD3+ CD3 Cp2Zr -C=C-CH3

+ CpZr >H + R
2 3R R

formation of zirconium vinyl species, which apparently

rearranges by a 1,3 hydride shift to generate _,9.


Generally 1,3 shifts are only probable when the origin and

terminus of the shifting hydride are in close proximity.133

However, examples of isomerization of vinyl to allyl

complexes are known134 and the driving force for the

formation of the allyl products is likely to be significant.

Only vinylic C-H activation can occur in the reaction of L

with allene and it is competitive with insertion for that


Allylic C-H activation would yield the r3-allyl cation

directly. The reactivity of CoCH3+ and FeCH3+ indicates

that C-H activation occurs at the allylic position because

of the ca. 20 kcal mol-1 lower bond energy.127 The

mechanism proposed for reactions of CoCH3+ and FeCH3

involves initial insertion of the metal ion into the allylic

C-H bond followed by the reductive elimination of methane.

Oxidative addition is not invoked in the reactions of dO

metal complexes since further oxidation of the metal center

Scheme 2-7
+ +
-+ H2C I
Cp2ZrCD3+ /CD3 CP2Zr )C-R
/ \ H2C
+ Cp2Zr( )H
/CH3 / +
2 R


is recognized as energetically unreasonable.74 The present

results suggest that oxidative addition mechanisms are not

the only possible pathways for elimination of neutral

molecules in analogous bare-metal ion reactions.

Cp2ZrCH3+ is produced in the reaction of Cp2ZrCD3+ with
isobutene and provides evidence for insertion in a

Markovnikov fashion. Evidently, insertion produces a

zirconium-neopentyl species, 10, in which all three methyl

groups are equivalent. Thus the reverse reaction,

-H3C+ -+CH
H3C\ 7 CH2
Cp2Zr2 /,C--CH3 Cp Zr- CH3

B-alkyl shift followed by isobutene elimination results in

the formation of 1. 8-alkyl shift has been observed in

solution for dO and dOfn metal complexes in reactions with

unsaturated hydrocarbons.54,126 H2 is not eliminated

because no B-hydrogens are available to initiate the

rearrangements necessary for H2 loss (Scheme 2-4, 2-5).

Anti-Markovnikov addition of isobutene would produce ion 11,

which has eight B-hydrogens and would be expected to undergo

facile H2 elimination videe infra).

Apparent insertion/deinsertion of isobutene without H/D

scrambling is further evidence for the necessity of the B-

hydride shift sequence in scrambling rearrangements. One

might expect scrambling through intermediates analogous to

those presented in Scheme 2-4 and 2-5, (i.e., 12, U1).

However, 1,2 hydride shifts would not accomplish the

interchange of hydrogen and deuterium.

CH3 D /CH3
Cp2Zr. +C-CH3 Cp2Zr /C CH3
H2C 'C
H2C+ CD3

12 13

The reaction of 1 with allene yields three products.

Two of the products are proposed to result from an insertion

process and the third from a-bond metathesis. CID spectra

of the Cp2ZrC4H7+ produced from allene are identical to

those for products observed in the reaction of propylene and

isobutene; therefore, insertion of allene evidently forms

the methylallyl cation, 1, directly. A small portion is

dehydrogenated to produce Cp2ZrC4H5+ the structure of which

is most likely a charge delocalized metallocyclobutene

analogous to that given previously for 7 (R = Me).

The product of a-bond metathesis, Cp2ZrC3H3+, is also
proposed to be a metallocyclobutene with structure 2 (R =

H). CID of Cp2ZrC3H3 yields uninformative spectra similar


to those obtained for the product of the same mass generated

in the reaction of I with acetylene. None of the ions react

further with allene implying that they are not a-complexes.

Reactions of Cp2ZrCH 3 with Internal olefins

Cis and trans isomers of 2-butene and 2-pentene as well

as cyclohexene and 1,5-cyclooctadiene have been investigated

(Table 2-2). The trans isomers react almost exclusively by

a-bond metathesis resulting in the loss of CH4. This

conclusion is supported by reactions of Cp2ZrCD3+ which show

no H/D scrambling and loss of CD3H only. Cia isomers yield

products from both insertion/elimination and a-bond

metathesis. The major product for cis-2-butene and cis-2-

pentene results from the insertion sequence. No a-bond

metathesis is observed for either of the two cycloalkenes


The influence of the isomeric gci and trans structures

for internal alkenes leads to striking differences in

reactivity. Since either isomer yields identical insertion

products, the reactivity difference must arise from the

initial orientation of the alkene as it interacts with the

zirconium center. Trans isomers are generally bound more

weakly than their gci counterparts in r-complexes with

transition metals135 and this may explain the lack of

insertion for the trans alkenes. This explanation hinges on

the assertion that coordination of the alkene as a r-complex

seems necessary for insertion to occur. In the present


scenario there is obviously a vacant coordination cite in

cation 1. However, it is not obvious by using molecular

models that cis coordination is significantly less

sterically hindered than trans coordination. Thus the

mechanistic feature responsible for the large difference in

reactivity observed for gis and trans isomers presently

defies adequate description.

The cycloalkenes offer additional information because

these substrates only show products from insertion/

elimination reactions. One might expect similar products

for reactions of cis-2-butene and cyclohexene. The

controlling factor in this case is apparently the stability

of the resulting products from a-bond metathesis. The

product of the reaction of cis-2-butene and 1 is proposed to

generate an allyl complex, because subsequent reaction with

cis-2-pentene is not observed. Allylic and vinylic C-H

activation are possible in reactions of 1 and 2 with

internal olefins. For cyclohexene the formation of an q3-

allyl complex produced from a-bond metathesis is probably

precluded by an unfavorable steric interaction of the C6

ring with the Cp ligands. The n3-allyl species generated by

insertion and elimination of H2 is not as sterically

unfavorable as the r3-allyl formed from the metathesis


Reaction of C2ZrCH13+ with Fluorinated Alkenes

The absence of products attributable to insertion for

reactions of I with fluorinated substrates is explained by

polarization considerations in a four-membered four-electron

transition state, 12 (R = F, CF3). The barrier for methyl

migration is significantly increased due to destabilization

of the transition state caused by the electron withdrawing

inductive effect of the fluoro, and trifluoromethyl


-- +
6+ ,3 +
Cp2ZrI I 'C

1. R = F, CF3

Reactions of CP2ZrCH3+ with Alkynes

Reactions of I with four alkynes were investigated and

products proposed to result from both a-bond metathesis and

insertion are observed. Acetylene, the simplest alkyne

reacts readily with 1 to produce Cp2ZrC3H3+, 6 (R = H), and

eliminate H2. Reaction of 2 with acetylene indicates H/D

scrambling; however, the elimination of H2 is not observed

in significant amount. Since no further insertion occurs

the structure of Cp2ZrC3H3+ is most likely a charge

delocalized metallocyclobutene analogous to (R = H). No


products from a-bond metathesis are observed in the reaction

of I with acetylene.

A large portion of the products in the reaction of 1

with diphenylacetylene do not eliminate a neutral molecule.

Insertion followed by 1,3 hydride shift is proposed to yield

the allyl complex 14. Elimination of H2 also takes place,

probably producing a metallocycle similar to 14. No further

reaction is observed for either product ion.

-- +

Cp2Zr CPh


Propyne and phenylacetylene yield products from both a-

bond metathesis and insertion. The metathesis route gives

products that continue to react with substrate implying that

they are a-complexes rather than delocalized allyls or

metallocycles. Metathesis product structures are given as

15 and J1 for propyne and phenylacetylene.

Cp2Zr-CmC-CH3+ Cp2Zr-CmC-Ph+

15 16


A second insertion yields products Cp2ZrC6H7 and

Cp2ZrC4HPh2+, which are inert to further insertion
indicating that these species are not simple a-complexes but

also occupy an additional zirconium coordination cite.

Cp*2ScCH3 was shown to catalyze the dimerization of

propyne to the gem-enyne H2C=C(CH3)CmCCH3.74 The proposed

mechanism involves initial a-bond metathesis generating a

propynyl complex which undergoes insertion followed by a-

bond metathesis, eliminating the gem-enyne product. Two of

the three steps in this process are observed in the gas

phase reactions of 1 with propyne and phenylacetylene (eqs

2-35, 2-36). However, the final a-bond metathesis reaction,

eliminating H2C=C(CH3)CmCCH3 does not occur (eq 2-37). This

may simply be due to a significantly lower rate for the

elimination process and can be explained by the interaction

of the alkyne portion of the insertion product and the

zirconium center. In solution this interaction can be

removed by coordination of solvent which would make the

final a-bond metathesis event energetically more favorable.

Several structures are possible for Cp2ZrC6H7+, and

Cp2ZrC4HPh2 CID experiments show loss of a single propyne
and phenylacetylene unit respectively. Structures 12 and I8

are consistent with the observed results and the reaction of

Cp*2ScCH3 with propyne. The similarity in the reactions of

propyne and phenylacetylene indicates that this reactivity

may be common to a variety of a-alkynes.

Cp2ZrCH3+ + HCmC-R Cp2Zr -C=C-R + CH4

p+, 1C

+ H\ /R
Cp2Zr+-CNC-R + HCOC-R C2Zr-C=C

15, 16 ,C9

12, 18 (R = CH3, Ph)




+ HC=C-R

17, 18

-- + (2-37)


15, 16


Reaction of I and 2 with alkenes and alkynes via methyl

migratory insertion is supported by several experimental

observations. Reaction of 2 with isobutene and propyne

yield 1, which is explained by an insertion/deinsertion

process in which the proposed insertion product decomposes

by B-alkyl elimination. In addition, B-hydride elimination

is proposed to explain the formation of Cp2ZrH+ in the

reaction of Cp2ZrD+ with ethylene.




Insertion of terminal alkenes other than isobutene and

a-methyl styrene yield products with B-hydrogens. Since

insertion is highly exothermic, decomposition of the

apparent insertion products is not unexpected. B-hydride

shift provides the initial step for the elimination H2 and

H/D scrambling in reactions with labelled reactants.

Reaction of I with isobutene via Markovnikov addition does

not yield an alkyl product with a B-hydrogen. Therefore,

elimination of H2 is not expected and not observed.

The product from insertion is observed in the reaction

of I with allene. CID spectra of this product is identical

to that of the product generated in the reaction of I and

propylene, which is also consistent with Markovnikov


The strongest evidence for a-bond metathesis is the

exclusive elimination of CD3H for methane extrusion

reactions of 2 with alkene substrates. The absence of

scrambling for methane elimination reactions and the

elimination of CH3D in the reaction of I with D2 are

consistent with a four membered transition state proposed

for a-bond metathesis reactions in solution.

In addition, a-bond metathesis kinetic studies in

condensed phase indicate alkynes undergo a-bond metathesis

more rapidly than alkenes.74 This kinetic trend is

apparently upheld for reactions of 1 with terminal alkenes

and alkynes in the gas-phase. Since the rate for insertion


of alkenes and alkynes is similar, the kinetic trend for a-

bond metathesis is indicated by an increase in the relative

rate of a-bond metathesis compared to insertion for alkynes.

Thus, insertion predominates for most alkene substrates,

whereas a-bond metathesis is competitive with insertion for

several alkynes.

The mechanistic interpretation of the reactions of I is

simplified because of the absence of accessible zirconium

oxidation states. Thus, oxidative addition and reductive

coupling, often invoked to explain the reactions of bare

metal ions, are highly improbable mechanistic sequences in

the reactions of 1. This suggests that mechanisms other

than oxidative addition/reductive elimination may also be

applied to the reactions of bare metal ions in the gas-


We have demonstrated that processes observed in the

solution reactions of d0fn organometallic complexes are also

observed for the CP2ZrCH3+ in the gas-phase. The gas-phase

investigation of cation I provides a unique opportunity to

characterize the a-bond metathesis and insertion reaction

sequences in the absence of solvent, especially because the

results can be compared to the same processes for analogous

complexes in solution. The present results also suggest

that information pertaining to the synthesis of a variety

organometallic complexes and possible side reactions may be

obtained by gas-phase studies.


Results for ion/molecule reactions were obtained by

using Fourier Transform Ion Cyclotron Resonance Mass

Spectrometry. Two different mass spectrometers were

employed for reaction studies. Data used in determination

of rate constants were acquired with a Nicolet FTMS 1000

(3.0 tesla). Product distributions were determined mainly

on the second system, which is constructed around a 2.0

tesla magnet. The second system is very similar to the FTMS

1000; the main difference is a larger vacuum chamber and

increased pumping capability. A check of product

distributions for the same reactions determined on the two

systems show good agreement within experimental error.

Electron impact (EI) (11-12 eV) on Cp2Zr(CH3)2 (p

10-8 -10-7 torr) yields predominantly Cp2ZrCH3+ and Cp2Zr+

Cp2Zr+ any ionized substrate are ejected from the ICR cell
isolating Cp2ZrCH3+ for reaction studies. Substrates are

generally present at pressures ca. 10-7-106 torr. Argon or

krypton are used as buffer gases at pressures in the 10-6-

10-5 torr range. A thermalization time precedes the

acquisition of data and permits Cp2ZrCH3+ multiple

collisions with buffer gas to ensure reactions are studied

under thermal conditions.

The formation of binuclear zirconium ions limits the

study of ion/molecule reactions to those with rate constants

above 109 M-1s-1 (10-12 cm3s-1). However, these limits can

be augmented with pressures of Cp2Zr(CH3)2 in the low 10-8

torr range.

Water reacts with 1 rapidly, even present as a slight

impurity in a substrate (eq 2-38). Therefore, drying of

substrates is imperative.

Cp2ZrCH3+ + H20 Cp2ZrOH+ + CH4 (2-38)

The method used for determination of rate constants was

described previously and yields values estimated to be

within 30% of the absolute rate constants. Product

distributions and rate constants are reported in the text

with 95% confidence limits.

Cp2Zr(CH3)2 and Cp2Zr(CD3)2 were synthesized according
to a literature preparation136 and analyzed by NMR and mass

spectrometry. Deuteromethyllithium was synthesized in dry

diethyl ether from CD3I purchased from Aldrich. Mass

spectrometric analysis of Cp2Zr(CH3)2 indicates > 99%


Liquid substrates were purchased from Aldrich in high

purity and used after drying over CaH2. C2D4 and D2 were

purchased from Cambridge Isotope Labs and used as received.

All other gases were purchased from commercial sources in

high purity and used without further purification.



Several reports have recently appeared in which

electrophilic organometallic, lanthanide, actinide, and

early transition metal alkyls and hydrides coordinate and

insert nitriles.137-146 In some instances it is possible to

isolate simple coordination complexes79 while in other cases

insertion takes place rapidly and an intermediate nitrile

adduct is not observed.125,140 Certain azomethine insertion

products react with a second equivalent of nitrile yielding

an isolable metalloheterocycle.125,143 Nitrile insertion is

also known for aluminum alkyls and hydrides147-149 and

normally occurs when the nitrile adducts are heated.150'151

Reactions of the higher alkyls of aluminum indicate that

oxidation of the alkyl ligands to alkenes competes with

insertion. Many of the same processes reported for the

aforementioned complexes are also observed in the gas-phase

reactions of CP2ZrCH3+, 1, with nitriles.

We have a continuing interest in developing relevant

comparisons of the reactivity of organometallic complex ions

in the gas-phase and solution.114,152'153 Our current

objective is to investigate and understand the reactivity of

the electrophilic Cp2ZrCH3+ cation, which presents an

opportunity to study transformations with several nitrile

substrates without interference from solvent.

Reactions are studied by using Fourier Transform Ion

Cyclotron Resonance mass spectrometry (FTICR).111-113 The

unfortunate absence of direct structural probes in the

present configuration of the FTICR is compensated somewhat

by the nature of Cp2ZrCH3 This zirconium (IV) alkyl

cation is formally dO and therefore cannot reasonably

undergo oxidative addition, which limits the number of

structures and mechanistic permutations available for the

various reactions described. The present work is a report

of the reactions of Cp2ZrCH3+ with several alkyl nitriles

and benzonitrile. Attempts are made to compare gas-phase

and solution reactivity whenever possible and some

thermodynamic data are also included.

We applied FTICR mass spectrometry to study the

reactions of 1 with several n-alkyl nitriles, (CH3)3CCN and

benzonitrile. Reactions were monitored as a function of

time until no net change in the relative abundance of the

product ions occurred.

Zirconium product ions of the form CP2ZrCxHYNz+ yield

seven observed isotopic species. For clarity, the product

ions are reported as a single (m/z) value corresponding to

isotopic species containing the most abundant 90Zr isotope.

In this section product ions are indicated by an empirical

formula Cp2ZrCHyN~ + and no structure is implied other than

the Cp2Zr unit. The eliminated neutral is taken to be the

most stable species possible with the correct molecular

formula. Detailed structures of the Zr product ions will be

presented in the Discussion section.

Collision induced dissociation (CID)115,154-156 of 1
shows only the loss of 'CH3, strongly implying a metal

methyl structure. Reaction of Cp2ZrCD3+ with CD3CN and CID

of the resulting products shows an absence of H/D scrambling

even in the high energy CID experiment, indicating

involvement of the Cp ligand hydrogens is insignificant.
Reaction of Cp2ZrCH3+ with CH3CN

The interaction of I with CH3CN results in two
sequential reactions producing Cp2ZrC3H6N+ (m/z 276) and

Cp2Zr2C5H9N+ (m/z 317) (eq 3-1, 3-2). However, the reaction
represented in eq 3-2 does not proceed to completion (Fig

Cp2ZrCH3+ + CH3CN Cp2ZrC3H6N+ (3-1)

Cp2ZrC3H6N+ + CH3CN Cp2ZrC5H9N2 (3-2)

3-1). The final ratio of m/z 276 : m/z 317 is approximately

0.6 0.1 and shows little variation as the internal energy

of i is increased by augmenting the El beam voltage from 11









+J00000 c


Cp2ZrCH3 (m/z 235)
Cp2ZrC3H6N (m/z 276)
Cp2ZrC5H9N2+ (m/z 317)

Time (s)

Figure 3-1.

Ion abundance vs. time for the reaction of
I with CH3CN. Data are normalized to the
total ion abundance and fit with a cubic




eV to 20 eV. The final ratio is also invariant in the

presence of argon or cyclohexane buffer gas.

Experiments on the effect of increasing nitrile

pressure indicate that the reaction shown in eq 3 is not an

equilibrium. Ejection of m/z 276 after a final ratio has

been established shows no back reaction to form m/z 317;

likewise ejection of m/z 317 proves no subsequent forward

reaction occurs. However, unreacted m/z 276 can be

converted to m/z 317 in the following manner. Low energy RF

irradiation of m/z 276 causes an increase in its

translational energy without inducing collisional

fragmentation. The effect of this translational excitation

followed by collisions with buffer gas and reaction with

CH3CN is measured as an increase in the abundance of m/z

317, i.e., a portion of the unreacted Cp2ZrC3H6N+ (m/z 276)

is converted to CP2ZrC5H9N2+ (m/z 317) by increasing the

total energy of m/z 276 and allowing time for reaction with


In normal CID experiments on m/z 317, the formation of

m/z 276 is observed at lower energies (eq 3-3) and as the

CID energy is increased m/z 235 and m/z 220 are also

detected (eq 3-4, 3-5).

(m/z 317)


Cp2ZrC3H6N+ + CH3CN (3-3)

(m/z 276)

Cp2ZrCH3 + 2CH3CN (3-4)

(m/z 235)
Cp2Zr' + 2CH3CN + *CH3 (3-5)

(m/z 220)

CID of m/z 276 before the final ratio is established is

consistent with CID of m/z 317 (eq 3-6, 3-7). Furthermore,


--- Cp2ZrCH3 + CH3CN

(m/z 235)

Cp2Zr-+ + CH3CN + *CH3

(m/z 220)

the same product fragments are observed for CID of m/z 276

after the final ratio is attained.

Reactions of I with CD3CN are exactly analogous to

those for CH3CN, however this substrate allows the detection

of a product previously indistinguishable from the reactant

cation 1. Cp2ZrCD3+ (m/z 238) is formed in the early stages

of the reaction (eq 3-8). Complete conversion of Cp2ZrCH3

to Cp2ZrCD3+ does not occur before addition of CD3CN forms



Cp2ZrCH3+ + CD3CN
(m/z 235)

Cp2ZrCH3+ + CD3CN

(m/z 235)

S Cp2ZrCD3+ + CH3CN

(m/z 238)

(m/z 279)

m/z 279 (eq 3-9). The production of CP2ZrCD3+ can be

completely suppressed by using a low pressure of CD3CN

(1x10-7 torr) and a high pressure of cyclohexane buffer gas

(5x10-6 torr).

Reaction of Cp2ZrH3 ith CH313CC

Reaction of I with (CH3)3CCN is similar to reaction

with CH3CN in that two sequential additions of nitrile occur

producing Cp2ZrC6H12N+ (m/z 318) and CP2ZrC11H21N2 (m/z

401) (Fig 3-2) (eq 3-10, 3-11).

Cp2ZrCH3+ + (CH3)3CCN
(m/z 235)

Cp2ZrC6H2N+ +(CH3)3CCN

(m/z 318)

(m/z 318

(m/z 401)

In analogy to the CH3CN system the reaction depicted in eq

3-11 does not proceed to completion. The final ratio m/z

318 : m/z 401 is 5.5 0.1. Low power RF irradiation of







*~ 60.00

c 40.00 i

-0 20.00


Figure 3-2.


**.. Cp2ZrCH,+ Qm/z 235)
OOODL Cp2ZrNCsH12 +(m/z 318)
AAAAA Cp2ZrN2CIIH21 (m/z 401)

TIME (s)

Ion abundance vs. time for the reaction of
I with (CH3)3CN. Data are normalized to
total ion abundance and fit with a cubic


m/z 318 after the system attains the constant final ratio is
effective in promoting the conversion of m/z 318 to m/z 401.
The results of CID experiments on Cp2ZrC11H21N2 and

Cp2ZrC6H12N+ resemble those for the products of the CH3CN
reaction. Apparent loss of (CH3)3CCN occurs from both m/z
318 and m/z 401 (eq 3-12, 3-13). Infra-red multi-photon

Cp2ZrC11H21N2 CP2ZrC6H12N+ + (CH3)3CCN (3-12)
(m/z 401) (m/z 318)

Cp2ZrC6H12N Cp2ZrCH3 + 2(CH3)3CCN (3-13)

(m/z 318) (m/z 235)

dissociation (IRMPD)157'158 of Cp2ZrC6H12N+ by using a
continuous wave CO2 laser (1090 cm-1) is consistent with
results of CID experiments producing the Cp2ZrCH3+ fragment
Reaction of Cp2ZrCHI3 with PhCN
Unlike the reaction of I with (CH3)3CCN in which no

Cp2ZrC(CH3)3+ is observed, nearly 90% of I is converted to

Cp2ZrCH3+ + PhCN Cp2ZrPh+ + CH3CN (3-14)
(m/z 235) (m/z 297)

Cp2ZrPh+ (m/z 297) (eq 3-14). Cp2ZrPh+ reacts with PhCN to
produce Cp2ZrCPh2N+ (m/z 400) (eq 3-15). Addition of PhCN

to Cp2ZrCPh2N+ yields Cp2ZrC2Ph3N2+ (m/z 503) (eq 3-16), but

Cp2ZrPh+ + PhCN
(m/z 297)

Cp2ZrCPh2N+ + PhCN
(m/z 400)

(m/z 400)

(m/z 503)

does not convert all m/z 400 to m/z 503. The final ratio,

m/z 400 : m/z 503, is 0.3 0.1. CP2ZrCH3+ also adds PhCN
to produce Cp2ZrC2H3PhN+ (m/z 318) (eq 3-17). Addition of
PhCN to CP2ZrC2H3PhN+ yields Cp2ZrC3H3Ph2N2+ (m/z 441)

Cp2ZrCH3+ + PhCN
(m/z 235)

Cp2ZrC2H3PhN+ + PhCN
(m/z 318)

(m/z 318)

(m/z 441)

(eq 3-18). However, a reliable measurement of the final
ratio m/z 338 : m/z 441 is unattainable due to the low
intensity of these product ions.





CID of Cp2ZrC2Ph3N2+ (m/z 503) yields Cp2ZrCPh2N+ (m/z
400), which is the reverse of eq 3-16. Increasing CID
energy causes the loss of two PhCN units. However, Cp2Zr'+
is not observed even at high CID energies.
Reaction of Cp2ZrCH3+ with C25CN
The reaction of 1 with C2H5CN gives several products
and is summarized in Scheme 3-1. Approximately 80% of the
total product ion intensity is accounted for in a single
reaction pathway involving the production of CP2ZrC6H11N2
(m/z 331) (Fig 3-3).
CID experiments on Cp2ZrC6H1N2+ show fragmentation to
form Cp2ZrC3H6N+ (m/z 276) (eq 3-19). As CID energy is

Cp2ZrC6H11N2+ Cp2ZrC3H6N+ + C2H5CN (3-19)
(m/z 331) (m/z 276)

increased Cp2ZrH+ (m/z 221) is observed in a nine fold
excess over Cp2ZrC2H5+ (m/z 249). IRMPD of m/z 331 is in
agreement with the CID results represented in eq 3-19, but
no other fragments are detected. CID of CP2ZrC3H6N+ (m/z
276) produces Cp2ZrH+ in higher abundance than Cp2ZrC2H5+
resulting in CID spectra similar to that recorded for high

energy CID of Cp2ZrC6H11.N2. Finally, Cp2ZrH+ is the only
fragment observed in the CID of CP2ZrC2H4N+ (m/z 262) (eq 3-

Scheme 1-1

Cp2ZrCH3+ + C2H5CN
(m/z 235)


(m/z 290)

Cp2ZrC2H5+ + CH3CN

(m/z 249)

Cp2ZrNC2H4 + C2H4

(m/z 262)


(m/z 345)



Cp2ZrNC3H6 + CH3CN
(m/z 276)


(m/z 331)

The following ions are observed in less than 1% of the

total ion intensity: Cp2ZrNC5H10+ (m/z 304) is produced by
addition of C2H5CN to m/z 249 and Cp2ZrN2C5H9+ (m/z 317) is
produced by addition of C2H5CN to m/z 262.

100.00 R






Figure 3-3.


**** Cp2ZrCH3+ (m/z 235)
C(QQ= Cp2ZrNC2H4 (m/z 262)
AAAA Cp2ZrNCaHe* (m/z 276)
ooo= Cp2ZrN2C6H,, (m/z 331)


TIME (s)

Ion abundance vs. time for the major
reaction pathway in the reaction of 1 with
C2H5CN. For clarity all product ions are
not represented. Data are normalized to
the total ion abundance and fit with a
cubic spline.


Cp2ZrC2H4N+ Cp2ZrH+ + C2H5CN (3-20)
(m/z 262) (m/z 221)

Approximately 18% of the product ions occur in an
addition pathway similar to that observed for CH3CN and

(CH3)3CCN (Fig 3-4). The final ratio of m/z 290 : m/z 345
is 0.8 0.1 which is much closer to that measured in the

CH3CN system. CID of these product ions is precluded by

their low intensity.

Reaction of CP2ZrCH13+ with n-Cn 2n+CN (n = 3.4)

The reactions of Cp2ZrCH3+ with n-C3H7CN and n-C4HgCN

are summarized in Scheme 3-2. The general reactivity of

these two n-alkyl nitriles is similar to that observed for

C2H5CN. For example, one series of reactions comprises over

75% of the final product ion intensity in the reaction of I
with n-C4H9CN (Fig 3-5). However, a new reaction pathway is

available to n-alkyl nitriles with carbon chains longer than

that of C2H5CN (eq 3-21). Product ions from this reaction

also add n-CnH2n+lCN as indicated in Scheme 3-2.

Cp2ZrCH3 + n-CnH2n+ICN Cp2ZrCnH2n-+ + NC2H5 (3-21)

The single common reaction sequence observed for all
nitriles investigated leads to the products Cp2ZrCn+2H2n+4+

(m/z 304, 318) and Cp2ZrC2n+3H4n+5N2+ (m/z 373,401). The