Molecular orbital investigations of metal-oxo catalyzed oxidations

MISSING IMAGE

Material Information

Title:
Molecular orbital investigations of metal-oxo catalyzed oxidations
Physical Description:
viii, 98 leaves : ill. ; 29 cm.
Language:
English
Creator:
Cundari, Thomas R., 1964-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 88-97).
Statement of Responsibility:
by Thomas R. Cundari.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001613884
notis - AHN8302
oclc - 23535945
System ID:
AA00003741:00001

Full Text











MOLECULAR ORBITAL INVESTIGATIONS
OF METAL-OXO CATALYZED OXIDATIONS













BY

THOMAS R. CUNDARI


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

UNIVERSITY OF FLORIDA

1990


'UNIVERSITY OF FLORIDA LIBRARIES















This thesis is dedicated to the memory of my father,

Michael V. Cundari Sr. (1925-1988), whose lessons about

hard work and perseverance has more to do

with the completion of this dissertation

than any of the chemistry I ever learned.













ACKNOWLEDGMENTS

No one lives life, or does research, in a vacuum. The

amount of people who must be thanked is enormous. Family

and friends are the most obvious choice. Even for an anti-

social beast such as myself, they are too numerous to

mention. They know who they are; my life, my research and

this dissertation have profited from knowing them.

Colleagues, instructors and mentors are gratefully

acknowledged for the intellectual stimulation of many

fruitful discussions. I owe a very special debt of

gratitude to my research advisor, Professor Russell S.

Drago, for putting up with my harassment and giving it back.

Although it is unfair to single out a single member from the

rest of my Doctoral Committee, it would be even more unfair

not to acknowledge the contribution of Professor Michael C.

Zerner to my chemical education. His seemingly boundless

energy and enthusiasm, as well as encouragement, has

provided much motivation. I would like to thank the rest of

the Department of Chemistry at the University of Florida (in

particular, the Division of Inorganic Chemistry, i.e. the

best, most happening division on campus!), for providing a

great place to live, work and play. Those who are connected

with the running and maintaining of the microVAX (upon which








the majority of the research presented here was carried

out), i.e. Steve Cato, Erik Deumens and the rest of QTP,

etc. are acknowledged.

Some specific acknowledgements for the work contained

herein are necessary. Much of the work was supported in

part by grants from the National Science Foundation and the

United States Army. For our work on epoxidations, we

gratefully acknowledge helpful discussions with K. F.

Purcell (Kansas State University) and a generous grant of

computer time by A. S. Goldstein and G. C. Martin. The

epoxidation work is based on a talk delivered by the author

at the 197th National Meeting of the American Chemical

Society in Dallas. The author wishes to thank the organizer

of the Symposium on Oxygen and Related Group Transfers, J.

T. Groves (Princeton), for organizing a wonderful symposium

and to the other participants for many stimulating

discussions. The work on alcohol oxidation was presented at

the 29th Sanibel Symposium organized by the Quantum Theory

Project of the University of Florida. The organizers of the

Sanibel Symposium are thanked for their hospitality. The

help of Greg Harris in preparing the section on alcohol

oxidations is gratefully acknowledged. Gilda Loew

(Molecular Research Institute) was kind enough to supply

reprints of her work (References 4c and 4d).

From the obvious we must go to the obscure. I would be

remiss in concluding this acknowledgement section without








professing the great sense of gratitude I feel toward those

who have gone before me, i.e. Lewis, Pauling, Mulliken,

Hund, HUckel and the myriad of other who have taken this

still relatively new quantum theory and wrested it from

those who would turn theoretical chemistry into an exercise

in "number crunching." I have always believed, still

believe, and hopefully will always believe that the purpose

of theoretical chemistry is to illuminate, stimulate and

suggest to our brethren, the experimental chemist, the what,

why, where and how of this great science chemistry.














TABLE OF CONTENTS
page

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

ABSTRACT ..... .... ........................... ....... vii

CHAPTERS

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

2 EPOXIDATIONS.................................. 9

Introduction. .................................. 9
Calculations............................. ..... 9
Results and Discussion........................ 12
Conclusions..... ............................... .. 31

3 ALCOHOL OXIDATIONS............................. 34

Introduction........ ......................... 34
Calculations.............. .... .. ..... ........... .. 34
Results and Discussion........................ 35

4 SULFIDE OXIDATIONS............................ 51

Introduction............... ............... ....... 51
Calculations............................ ..... 51
Results and Discussion....................... 52

5 OLEFIN OXIDATIONS BY RU-DIOXO COMPLEXES........ 68

Introduction............................. ........ 68
Calculations................................... 69
Results ....... ... ........ .... ................ 70
Discussion ...... ... .... ..... ............... 80

6 CONCLUSIONS... ............. ................... 84

REFERENCES. ........... .............................. 88

BIOGRAPHICAL SKETCH................................. 98














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

MOLECULAR ORBITAL INVESTIGATIONS
OF METAL-OXO CATALYZED OXIDATIONS

By

Thomas R. Cundari

August, 1990

Chairman: Professor Russell S. Drago
Major Department: Chemistry

Molecular orbital analyses of oxidations by a six-

coordinate, Ru(IV)-oxo complex (cis-[Ru(HN=CH-

HC=NH)2(NH3)(0)]2+) are combined with experimental data and

basic electronic structural arguments in an effort to

understand the nature of the chemical bond in the systems

studied. An analysis of the pathways and intermediates for

the epoxidation of olefins using a six-coordinate, Ru(IV)-

oxo model complex is presented. As in the overwhelming

majority of calculations described herein, INDO/1

calculations were used. As with the more familiar organic

analogues, concerted [1+2] and [2+2] cyclo-additions lead to

unfavorable orbital interactions. A non-concerted [1+2]

process is preferred. The initial interaction between

substrate and catalyst in the oxidation of alcohols by the


vii







same Ru(IV)-oxo complex is studied using the INDO/1 method.

Activation of the C-H bond of the alcohol by the oxo oxygen

of the Ru(IV)-oxo complex is compared and found to be

competitive with inner-sphere coordination shell expansion

of the Ru-oxo catalyst. The preferred pathway combines

aspects of both proposals, i.e. C-H-Oxo activation and

coordination to the metal. The subsequent conversion of the

coordinated methanol to carbonyl products is also discussed.

An INDO/1 investigation of the oxidation of organic sulfides

by the Ru(IV)-oxo complex is presented. Two distinct

reaction pathways were considered: oxygen atom transfer to

the sulfide and pre-coordination of the sulfide to the

metal. Oxygen atom transfer was found to be the most

favorable pathway in agreement with experimental evidence.

The INDO/1 model is employed to analyze the reaction between

olefins and cis- and trans- Ru(VI)-dioxo complexes. The

differences in reactivity of olefins with the cis- and

trans-Ru(VI)-dioxo complexes arise from different preferred

pathways: a non-concerted [1+2] pathway for the trans isomer

and a concerted [3+2] pathway for the cis isomer.


viii













CHAPTER 1
INTRODUCTION

The oxidation of organic substrates catalyzed by

transition metal complexes is an area of great scientific

interest.1,2,3 Biological4,5 and industrial6'7 processes

have spurred this great interest. One example is the

oxidation of organic species to less toxic, more water

soluble products by cytochromes P-450,8 with their iron-

porphyrin active site.9 Another example is the conversion

of cyclohexane to cyclohexanol and cyclohexanone (which are

subsequently converted to adipic acid,10 an important

precursor in the manufacture of Nylon) catalyzed by a

homogeneous Co(II) catalyst.

Two distinct transition metal oxidation systems are the

72 peroxidell,12 and the metal-oxo.13 Much experimental

and theoretical work has gone into understanding these

oxidants. Group VIB r2 -peroxides have received their

greatest use as catalysts for the epoxidation of

olefins.14,15 Two alternate proposals have dominated much

of the discussion as to the mechanism for the epoxidation of

olefins by the 72 peroxo complexes. The so-called Mimoun14

mechanism involves the pre-coordination of the olefin to the

metal followed by a [1+3] dipolar cycloaddition to an MO2







2

fragment to yield the proposed, five-membered ring

intermediate the peroxometallocycle. This intermediate

then undergoes an [1+3] dipolar cycloreversion to produce

the epoxide product. In the Sharpless15 variation there is

direct nucleophilic attack of the olefin on one of the

oxygen atoms of the peroxide moiety. Electronic

rearrangement yields the epoxide product by what amounts to

direct oxygen atom transfer. In both cases the other oxygen

of the peroxide unit becomes an oxo oxygen. An external

oxidant (such as HOOH or an organic hydroperoxide, ROOH) is

necessary to regenerate the active peroxo species. A

theoretical analysis by Jorgensen and Hoffmann16 using the

Extended HUckel method17-19 suggests a mechanism which is a

hybrid of the Sharpless and Mimoun formulations. There is

activation of the olefin by coordination to the electron-

deficient, high-valent metal followed by migration of the

olefin toward a peroxide oxygen to form a species similar to

the Sharpless intermediate. Purcell20 considered the

energetic for decomposition of the most widely proposed

intermediate in epoxidation of olefins by an n2-peroxo

complex, i.e. the peroxometallocycle (a five-membered M-O-O-

C-C ring), to an epoxide for both Mo(VI) and a Rh(III)

systems. Once again, Extended HUckel calculations were used

to delineate the pertinent orbital changes and energetic

associated with the reaction.20







3

Another prevalent substrate for oxidations by n2

peroxides have been organic sulfides, R2S. DiFuria and co-

workers21 have shown MoO(02)2L (L = basic ligand, e.g.

PO(NMe2)3) complexes to be effective sulfide oxidation

catalysts. The products are mainly sulfoxide, R2SO, and

sulfone, R2SO2. Research conducted in our own

laboratories22 has shown metal-acetylacetonoate complexes

(Mo(0)2-(acac)2 and VO(acac)2) to be potent sulfide

oxidation systems in the presence of (CH3)3COOH in

acetonitrile solution. The most plausible active species is

the n2-peroxide. Drago and co-workers23 have also shown

MoO(O2)(H20)(C5H3N(CO2)2) to be an effective sulfide

oxidation catalyst.

The current interest in metal-oxo chemistry was

initiated by the discovery of Groves and co-workers24 that

iodosyl benzene (C6H5-I-O) and Fe(III)tetraphenyl porphyrin

chloride could mimic the reactivity of the class of enzymes

cytochromes P-450 and molecular oxygen. Cytochromes P-450

are a family of monooxygenases responsible for the oxidation

of lipophilic substrates. The active species is thought to

be a ferryl heme complex. Groves and co-workers have

carried out a detailed experimental characterization of the

iron and related P-450 models.25-30 Since then, many

different oxygen atom transfer reagents (such as C6F5-I-031,

Cl-0-32,33, R3N-034,35, etc) and metal species have been

investigated.36-42 Metal-oxo reagents have been used to







4

oxidize a wide range of substrates: olefins to epoxides,43-
52 alcohols to ketones and aldehydes53, sulfides to

sulfoxides and sulfones,54 N-alkyl amines to less

substituted amines,55 hydrocarbons to alcohols,56-61 and

others.62

Theoretical research into the pathways and

intermediates used in metal-oxo mediated oxidations has been

very profitable. Using Extended HUckel MO calculations

Jorgensen63 has investigated the conversion of an Fe-oxo

porphyrin species (Fe[porphyrin](O)(H20)) to an N-oxo

precursor. The proposal of the N-oxo (the N is from one of

the nitrogen ligands, i. e. the pyrrole unit of the

porphyrin) as the active species in the oxidation of organic

substrates by P-450 models was derived from the work by

Bruice and his collaborators34,35 which showed that

nitrogen-oxides (e.g. pyridine N-oxide or morpholine N-

oxide) could act as oxygen atom transfer reagents.

Jorgensen's results indicate that this perturbation is

possible. Jorgensen proceeds to analyze the interaction

between the N-oxo species and ethylene leading to epoxide

formation. Strich and Veillard64 have used ab-initio

wavefunctions to study this isomerization for a similar

species Fe[porphyrin](0)(py). Their results indicate that

this isomerization is quite unfavorable with the N-oxo

intermediate being approximately 100 kcal mol-1 higher in

energy than the ground state Fe-oxo isomer.







5

Goddard and Carter65 have studied metal-oxo diatomic

systems using Generalized Valence Bond Theory66-68 and then

subsequently69 have shown how the properties of the isolated

metal reactive site could be transferred to a larger,

coordinatively saturated complex which contained this same

fragment. Rapp6 and Goddard70,71 have looked at the

proposed intermediates and transition states for the

oxidation of olefins with Group VIB metal-oxo reagents (e.g.

MoO2Cl2 and CrO2C12) using GVB calculations. Their analysis

indicates that a four-membered M-O-C-C ring structure, i.e.

a oxometallocycle, is the most favorable intermediate and

that the driving force for the reaction (to produce epoxide)

is the strengthening of the second metal-oxygen bond from an

approximate double bond to a triple bond as the first oxygen

atom is being transferred.70 Rapp6 and Goddard also studied

models for molybdenyl and chromyl species supported on

inorganic oxides (alumina, silica, etc.).71 Goddard carried

out a detailed analysis of the ground and excited states of

NiO and commented at length as to the similarity of this

species with molecular oxygen with respect to the low energy

excited states which are generated by electron correlation

within the Ni dr 0 pr MOs.72 Bauschlicher and co-workers

have compared PdO and NiO, MoO and CrO, and CuO and AgO with

the aim of delineating the crucial differences between first

and second transition series with respect to their

electronic structure and chemical bonding.73 Krauss and







6

Stevens have employed pseudo-potential methods to compare

RuO and FeO.74 In both cases the ground state possesses a

dioxygen-like v4,"2 electronic configuration. Bagus has

investigated FeO using all-electron ab-initio

calculations.75 He concludes on the basis of his

calculations that the ground state of FeO is not 5Z+ as has

been proposed.75 Goddard and Carter have utilized the GVB

method to study typical early (VO+) and late (RuO+)

transition metal oxo species.65 Their results indicate that

the early transition metal oxo species are similar to carbon

monoxide, i.e. a strong, approximately triple bond, and that

the later transition metal oxo complexes are more like

dioxygen, i.e. covalent I system. This rationalizes why the

early transition metal-oxo systems, e.g. vanadyl and

molybdenyl, do not readily engage in direct oxygen atom

transfer while later transition metal-oxo systems, e.g. Ru-

oxo, can. Loew and others 76,77,78 have employed semi-

empirical MO calculations (of the NDO type79-84) in their

studies. Much of the work of Loew and her co-workers has

been concerned with the electronic structure of iron

complexes which are germane to P-450 model systems. Loew

has, however, published an MO analysis of the oxidation of

aromatic hydrocarbons,85a the conversion of ethylene and

methane to ethylene oxide and methanol, respectively,85b and

the oxidation of propylene (allylic oxidation versus

epoxidation).85b The P-450 model studied in these papers







7

was triplet oxygen (3p). These studies show that 0 (3p) is

a good qualitative model for the active oxidizing species in

cytochrome P-450. Using 0 (3p) to model the active species

in cytochromes P-450 neglects effects brought about by the

metal and ligands.85c Blyholder, Head and Ruette have

analyzed Fe-dioxo species using a semi-empirical MINDO model

and have shown it to describe the chemistry of these species

quite well.86 Their main goal was to understand the

interaction between molecular oxygen and atomic iron.

Zerner and co-workers87,88 (using the spectroscopic, semi-

empirical INDO/1 model developed in his laboratories) and

Yamamoto and co-workers89 (using ab-initio calculations)

have presented detailed analyses of the ground and low

energy excited states of the iron-porphyrin reactants

(Fe(II)[porphyrin]87 and Fe(III)[porphyrin]C188) and a model

for the proposed active species, Fe(IV)O(porphyrin)(py),

respectively. The ab-initio CASSCF (complete active space,

self-consistent field) study by Yamamoto et al. 89 has

shown that for the Fe[porphyrin](py)(0) model system that

the ground state is largely ionic, Fe+4-O02, with this

configuration or "resonance structure" comprising over 80%

of the ground state wavefunction. The greater covalency of

comparable second transition series bonds versus those for

the first transition series (i.e. the metals belong to the

same group of the Periodic Table) has been appreciated by

Bauschlicher and his group.73,90 The study of Bauschlicher







8

et al. deserves special mention since it is the only

detailed comparison of the nature of the chemical bond for

analogous first and second row transition metal species that

has been published. The Ru-oxo bond is much more covalent

than the Fe-oxo bond as noted by Goddard and Carter.65

Theoretical research on the nature of the bonding in

the reacting systems has been sparse.91-93 Sevin has used

Extended HUckel wavefunctions as a starting point for the

construction of Valence Bond (VB) wavefunctions. The

reaction of interest was the Fe-oxo mediated oxidation of

ethylene to ethylene oxide. The model complex studied was

FeH402-.92,93 Sevin concludes that the preferred pathway is

a non-concerted formation of the bound epoxide. This study

suffers in two respects the use of Extended HUckel to

describe a polar system and the modelling of a porphyrin by

four hydrides. Yamaguchi and co-workers have looked at the

interaction of ethylene with various diatomic metal-

oxides.91 Their results show that the preferred pathway

depends strongly on the particular metal. As mentioned

previously, the ligands play an essential role.













CHAPTER 2
EPOXIDATIONS

Introduction

The mechanism of epoxidation by transition metal-oxo

systems has been intensively studied by experimentalists,

particularly for iron-oxo as it is the proposed active

species in cytochrome P-450 catalyzed oxidations.24 The

approach taken in the present work is to use INDO/1

calculations to explore the potential energy surface (PES)

for the interaction of ethylene with the metal-oxo reactant

and analyze the results using the language of the Woodward-

Hoffmann rules.94 A Ru(IV)-oxo complex, Figure 2-1, was

chosen since it is a well characterized complex43 which is

similar to the biologically significant Fe-oxo species.24



Calculations

The calculations were carried out on a microVAX using a

semi-empirical INDO/1 program written by M. C. Zerner and

his group. The UHF formalism of Pople and Nesbet95 was used

since most of the species involved are open-shell molecules.

To reduce the overall basis set size, and computer time,

glyoxal diimine was used in the calculation to model the

experimentally employed nitrogen ligands, Figure 2-1.












+2








NRu N

















Figure 2-1: Geometry of the Ru-oxo model complex upon which
the calculations were performed.







11

Bipyridine parameters (bond lengths and angles) were used

for glyoxal diimine96 and free ammonia values96 for the

ammonia ligand in the model complex, Figure 2-1. Gradient

driven geometry optimizations97 were carried out on various

Ru-oxo/ethylene species to yield stationary points on the

PES. The M-L bond lengths were obtained by INDO/1 geometry

optimization (Ru-O = 1.96A, Ru-N = 2.06A) and agree well

with the experimental values.98 Convergence to a minimum on

the PES was assumed when the gradient was less than 5.0 x

10-3 Hartrees/Bohr.

The atomic bond index used in the MO analysis is

defined as

on A on B

BAB = Z Puv*Pu (1)

u v

where the summation is over all atomic orbitals Ixu> and

IXv> on atoms A and B, respectively and Puv is the bond

order or density matrix element between Ixu> and Ixy>. The

atomic bond index yields a value of 1 for a single bond, 2

for a double bond, etc. Puv is defined as

occ.

Puv = Z cui cvi (2)
i

where Cui and cvi are the AO coefficients of IXu> and Jxv>

in MO lTi> using the LCAO-MO approximation.







12

Results and Discussion



Electronic Structure of the RuN502+ Complex

The frontier MO diagram for the d4 RuN502+ complex is

that of a octahedron (i.e. energy(t2g) < energy(eg*)) with

the dxz and dyz components of the t2g being perturbed to

higher energy by antibonding mixing with the lower energy 0

2pr AOs. The RuN502+ complex is similar to oxygen in that

it contains a filled Ru di 0 pi bonding level and a half-

filled Ru dr 0 pr antibonding level. The n and the v* are

comprised largely of Ru dn and 0 pn AOs. As with 02 there

are low energy singlet and triplet states.99 The ground

state triplet (~1 a t*2 a) is calculated to be zleV below the

lowest energy singlet state (n a Tl*B). Note that the

formal oxidation state on the metal is +4, yielding a d4

metal.



Formation of an N-oxo Intermediate

The transformation of the metal-oxo species to an N-oxo

species has been proposed for the iron-oxo porphyrin

system.63,100 A potential energy surface, triplet spin

state, was constructed for the distortion mode that converts

the Ru-oxo complex to a species with increased N-0 bonding.

The Neq-Ru-O angle, 8, was compressed from 900, to allow

comparison with the results of Veillard100 and Jorgensen.63

A soft potential energy surface resulted up to 8 = 500,







13

after which the total energy increases markedly. Up to e =

500 there is no N-0 bonding, as evidenced by a maximum

atomic bond index of 0.06 and a minimum N-O distance of

1.67A on the shallow portion of the potential energy

surface. Stabilization of the HOMO (RuO v*) of the Ru-oxo

system, due to a decrease in the unfavorable overlap of the

Ru dw O pr antibonding combination, is counteracted by a

destabilization of the Ru dr O pr bonding combination

(caused by a decrease in the favorable overlap). The

doubly occupied bonding combination dominates the singly

occupied antibonding combination with respect to their

effects on the overall energetic. These considerations

lead to a high energy barrier for N-oxide formation and

reduce the possibility of the N-oxo mechanism for ruthenium.

The results for cis-[Ru(HN=CH-CH=NH)2(NH3)(0)]2+ (hereafter

referred to as Ru(IV)N502+) are in agreement with the ab-

initio analysis of Strich and Veillard100 on

Fe[porphyrin](0)(py) (py = axial pyridine) who found an N-

oxo intermediate to be unfavorable, but opposite to the

Extended HUckel results of Jorgensen63 on

Fe[porphyrin](0)(H20).


Initial Formation of an Ion-Molecule Complex

At separations greater than RC-Oxo 2.7A, energy

minimization leads to the formation of a weakly bound (by "

9 kcal mol-1) ion (RuN502-6)-molecule (C2H4+6) complex.







14

This ion-molecule complex is interesting in that it is

similar to the charge-transfer complex which has been

proposed by Bruice31,101 in the epoxidation of olefins by

oxo-chromium and oxo-iron porphyrins. One must be careful

about extrapolating the results of "gas-phase" calculations

to solution-phase experiments. The ground state for this

complex is a triplet. Covalent C-Oxo bonding is virtually

nonexistent (Atomic Bond Index = 0.), and the Ru-Oxo/Olefin

interaction is largely electrostatic.

Proceeding from the shallow ion-molecule minimum toward

covalent C-Oxo distances incurs a small barrier, < 3 kcal

mol-1, at RC-Oxo a 2.5A. Geometry optimization after this

barrier has been passed leads to local minima which are

covalently bound. The minima are the bound epoxide; the

oxometallocycle was never observed as a stationary point on

the potential energy surface for the epoxidation of ethylene

by the Ru-oxo complex.



[2+21 Cycloaddition Pathway

A [2+2] cycloaddition, to yield a four-membered M-O-C-C

ring (referred to as an oxometallocycle), has been proposed

by several workers.46,102 Formation of an oxometallocycle

by a [2+2] pathway is allowed due to the lowered Cs

symmetry. Geometry optimization of structures in which the

Ru-O and C-C bonds are parallel was carried out. The Ru-C

bond was initially set to 2.3A (slightly greater than the







15

Ru-C bond distance of 2.21A in ruthenocene96). The

resulting minimum showed almost no ethylene-metal oxo

interaction: the C-C and Ru-O bonds do not change from their

individually optimized bond lengths (RCC = 1.32k, RRu =

1.96A); the H atoms maintain a planar arrangement around the

ethylene C's; and the metal complex maintains an octahedral

geometry. These calculations were repeated with the

pentammine analogue, as it possesses more freedom in its

bending modes and can distort out of the way to accommodate

an oxometallocycle. Similar results to those presented

above were obtained. These results suggest that the

formation of an oxometallocycle intermediate in a concerted

manner is not likely.

The unfavorable, planar [2+2] addition pathway results

from repulsion between the occupied RuO Ty and C2H4 i MOs.

The diabetic correlation diagram shows the correlation of

the ethylene t with a high energy a* orbital in the

products, Figure 2-2, that would result if an

oxometallocycle was formed. This two-center, four-electron

interaction is repulsive and leads to the unfavorable

concerted [2+2] pathway observed, although the symmetry

requirement is relaxed.

In RuN502+ it is the RuO moiety which largely

determines the reactivity. Removing the ligands from the

RuN502+ complex amounts to ignoring the small contribution















RUC aCUO


nT*c -
7tTr't A------

7r ccU






R

0
vow R I a #





II



R



Figure 2-2: Schematic diabetic orbital correlation diagram
showing the evolution of the active orbitals as a function
of R for the concerted [2+2] pathway.







17

of the ligand orbitals to the frontier orbitals. These

frontier orbitals are largely Ru 4d in nature. This

approach is similar to that used by Sevin and Fontecave in

their Extended HUckel study of Fe-O catalyzed

epoxidations.92,93 The removal of five ligands results in

the 6* MO being lowered in energy. Thus, two electrons must

be added to the Ru-oxo moiety in order to yield a species

with the necessary v4 *2 electronic configuration. Ab-

initio calculations performed on RuO and RuO+ by ourselves

and others74 yield ground and low energy excited states with

a 4T,*2 electronic configuration over a range of Ru-O bond

lengths. Our INDO/1 calculations describe these species

quite well. An analysis of the INDO calculations for the

concerted [2+2] pathway using RuO (RRuO = 1.96A) as the

metal-oxo reagent reveals that the crossing of the bonding

combination of the RuO i* and the higher energy ethylene n*

with the antibonding combination of vCC and RuO v occurs at

RCOxo and RRu-C = 2.5A, Figure 2-2, as evidenced by a

transfer of electron density from the olefin to the metal-

oxo fragment. These bond lengths are greater than typical

C-O or Ru-C covalent bonds making the concerted [2+2]

pathway unfavorable. The non-concerted pathway will be

discussed in greater detail below. The geometries of the

ethylene and Ru-oxo fragments were kept constant in this

explorations of the Ru-oxo/olefin PES.









[1+21 Cycloaddition Pathway

Continuing along the concerted [1+2] pathway, after

passing the small barrier from the ion-molecule minimum

leads to a large repulsion between the occupied ethylene w

MO and the oxo lone pair (largely pz). As with the

formation of cyclopropane by the least motion pathway from

ethylene and methylene,103 the concerted [1+2] addition of

ethylene to the metal-oxo is disfavored, Figure 2-3. Once

again using the Ru-O moiety and ethylene the orbital

correlation diagram for the [1+2] pathway was constructed,

Figure 2-3. A crossing between the antibonding combination

of the oxo lone pair and the ethylene v MO with the bonding

combination of the RuO and ethylene occurs at RC-Oxo

z 2A. This distance is, of course, greater than a covalent

C-O bond, Figure 2-3.



Non-concerted [1+2] and [2+21 processes

Non-concerted [1+2] and [2+2] pathways were considered

next, as the concerted pathways were shown to possess

unfavorable two-orbital, four-electron repulsions. The

concerted [1+2] and [2+2] pathways are connected by these

non-concerted pathways. Structures on the non-concerted

[1+2] and [2+2] pathways differ mainly by the Ru-O-C angle.

Structures on the non-concerted [1+2] and [2+2] pathways

possess Ru-O-C angles of = 1800 and = 900, respectively.

The electronic energies of the open structures [RuN5OC2H4]2+







19

and [RuOC2H4] are calculated to be within 10 kcal mol-1 of

each other depending on the exact geometry (Ru-O-C angle,

orientation of the H's about the ethylene carbons, etc.).

Based on the lower repulsions expected between any

substituents on the olefin with the ligands of the complex

for angles closer to 900, i.e. the non-concerted [2+2]

pathway, the non-concerted [1+2] trajectories would be

expected to be more favorable for steric reasons. Also, all

open structures optimized to epoxides regardless of Ru-C

distance, even in those cases in which the Ru-C distance is

less than the C-O distance. The Ru-C distance was changed

by changing the Ru-O-C angle. Greater attention was focused

on the non-concerted [1+2] pathways for these reasons. In

the open structures the C-C bond distance is roughly between

that of a single and double bond (Rcc = 1.4A); the

coordination about the terminal ethylene carbon is trigonal

planar; the coordination about the non-terminal ethylene

carbon is approximately tetrahedral; and the C-O bond length

is close to a typical covalent value (RCO = 1.45A).96

Single-point INDO/1 calculations were performed for the

concerted pathways, as well as the two geometric

perturbations which characterize the non-concerted pathways.

The RuO fragment (RRuO = 1.96A) was used to model the

oxidant. Due to the importance of electron correlation in

bond-making/bond-breaking processes an exact, quantitative

description of the potential energy surface is difficult.













CTC -

7ri7rRuo t


7rc


t7rL


.cO
CO


;R

WC=C "


Figure 2-3: Schematic diabetic orbital correlation diagram
showing the evolution of the active orbitals as a function
of R for the concerted [1+2] pathway.


A a,







21

However, general trends are expected to be well described.

The two geometric perturbations are shown in Figure 2-4. $,

describes the motion of the ethylene off-center, and was

varied from 00 to 900, for R = 3.0k, 2.5A, 2.0k, 1.5k, and

1.0k, Figure 2-4. 9, which yields information about the

process of canting, was also varied from 00 to 900 for the

same distances, Figure 2-4. The PESs constructed ($ x R x E

and 8 x R x E) on the triplet manifold (all multiplicities

gave similar results) are shown in Figures 2-5 and 2-6. As

with the well known organic examples, a non-concerted

trajectory results from a high barrier in the least motion,

concerted pathway. Translation of the ethylene off-center

(i.e. the center of the C-C bond is no longer co-linear with

the Ru-oxo bond) from the non-concerted [1+2] trajectory

decreases repulsion between the oxo lone pair and the

ethylene r MO and allows for back bonding between the

ethylene t* and the O lone pair (these MOs have zero overlap

otherwise). As evidence for decreased repulsion between the

oxo lone pair and the ethylene i, the calculated splitting

between the bonding and antibonding combinations of these

two MOs decreases by 0.009 hartrees for a 100 perturbation

off-center from the concerted [1+2] pathway at RC-Oxo = 2A.

At relatively large separations on the concerted [1+2]

pathway, R = 2.5A, perturbation of the ethylene off-center

causes the total energy of the reactants to decrease.














Z


^C0

Ru

x I














Figure 2-4: The two geometric perturbations which
characterize the non-concerted pathways. The values in
parenthesis are the 4 and 0 values, respectively. Thus,
(00,00) is the concerted [1+2] pathway and (900,00) is the
concerted [2+2] pathway.












RuO + C2H4: (Ph L x R x E)


*s


^l
39
I


'


Figure 2-5: PES (0 x R x E) for the coordinates discussed
in the text and shown in Figure 2-4. This PES describes the
perturbation of the ethylene fragment off-center of the
vector which describes the Ru-O bond.
















RuO + C2H4: (The te x R x E)

















'<*
"I

















Figure 2-6: PES (e x R x E) for the coordinates discussed
in the text and in Figure 2-4. This PES describes the
canting process of the ethylene fragment.







25

For example, at R = 2.5k, 8 = 00, moving from 0 = 0 to 100

decreases the total energy by 53 millihartrees (a 32 kcal

mol-1). As R approaches covalent values, R z 1.5A,

perturbation off-center causes an increase in the total

energy. This is consistent with the notion of stretching a

covalent bond from its equilibrium bond length (C-O in this

case). At R = 1.5A on the concerted [1+2] pathway ( 8 = =

00), moving from 0 = 0 to 100 increases the total energy by

15 millihartrees (= 9 kcal mol-1). Canting (i.e. rotation

about the C-C axis such that the molecular plane of the

ethylene is not perpendicular to the vector which describes

the Ru-O bond) leads to a more stable pathway due to

increased overlap of the ethylene v symmetry orbitals with

the T symmetry orbitals of the Ru-oxo fragment. For non-

canted pathways the T symmetry AOs of ethylene are

orthogonal to at least one set of the RuO v orbitals.

Canting removes this orthogonality, and allows interaction

between the ethylene w symmetry AOs and both sets of RuO i

orbitals. In the concerted [2+2] pathway the interaction of

the rCC with the rRuO is not symmetrical due to the

polarized nature of the latter MO (r(RuO-INDO) = 0.42 RuT +

0.91 Or compared to v(RuO-ab-initio) = 0.40 Ruv + 0.92 Or).

The "RuO is concentrated to a greater extent on the more

electronegative oxygen atom. Thus, C-O interaction is

stronger than C-Ru interaction and provides the driving

force for the observed non-concerted process. The decrease








26

in the repulsion between the 0 lone pair and the ethylene v

provides the driving force for a non-concerted [1+2] pathway

as it does for the analogous CH2 +C2H4 --> cyclo-C3H6

reaction.103

Geometry optimizations of RuN5OC2H42+ species reveal

that a non-concerted process takes place once C-O bonding

dissymmetry is induced, i.e. one C-0 bond is nearly fully

formed, while the second is virtually non-existent. In

fact, starting with the ethylene on the concerted [1+2]

pathway will lead to non-concerted pathways. As we saw for

the shallow minimum at RC-Oxo 2.7A videe supra), there was

already a slight dissymmetry in C-0 bonding.

An on-center trajectory in which the vector that

describes the Ru-0 bond bisects the C-C bond and is

perpendicular to the molecular plane of ethylene has

approximately C2v symmetry. Using the symmetry labels

pertinent to the C2v point group, these MOs transform under

the following irreducible representations: C2H4 7 (al), 7T

(bl); RuO ,x (b2), "y (bl), ix* (b2), 7y* (bl). Thus, the

ethylene may interact with the RuO vy and y* but the

ethylene v may not. Since the ethylene v and n" MOs are of

different symmetries they may not interact by a second-order

process. Moving the ethylene off-center of the Ru-O bond

reduces the symmetry to Cs and the pertinent MOs to the

following irreducible representations: C2H4 T (a), t* (a);

RuO nx (b), 7y (a), rx* (b), 7y* (b). Interaction may now








27

occur between the a symmetry orbitals. In an on-center

geometry the MO interaction diagram in Figure 2-7 applies.

The ethylene v MO does not interact with the RuO ry and y

MOs for the reasons given above. Moving the ethylene off-

center allows the doubly occupied ethylene n MO to interact

with the MO (singly occupied) formed from the bonding

combination of the RuO n and vCC* (Y* in Figure 2-7). The

doubly occupied bonding combination places more orbital

density on the non-terminal ethylene carbon, i.e. C-O

bonding is strengthened, Figure 2-8. An analysis of the PES

for the RuO moiety and ethylene shows this effect. Even at

the relatively far distance of R = 2.5k (S(Cpv, Opf) = -

0.01) the orbital described above changes from Y* + CC

0.92 Olp (lp = lone pair) + 0.27 Cnw + 0.27Ctr (B = t = 00)

to Y* + tCC = 0.97 Olp + 0.18 Cnr + 0.14Ctr (0 = 00, t =

100). Cn is the non-terminal ethylene carbon; Ct is the

terminal ethylene carbon. As R decreases to this effect

becomes more pronounced. The singly occupied antibonding

combination places more orbital density on the terminal

ethylene C, i.e. the terminal C acquires radical-cation

character, Figure 2-8. Translation of the ethylene from e =

00, t = 0 to e = 00, $ = 100 at R = 2.0A increases the spin

density on the terminal (un-bound) carbon from -0.02 to

+0.15 electrons. With a large amount of orbital density on

the Ct in this frontier MO, interaction with the orthogonal

RuO can, upon further reaction, lead to oxometallocycle







28

or bound epoxide. As stated previously, the perturbation of

the ethylene fragment to an off-center geometry increases

mixing by the ethylene i MO. Chemically, the ethylene

donates more electron density to RuN502+ and becomes more

electrophilic. The second-order mixing process places

greater partial positive charge on the terminal carbon.

This build up of positive charge on the terminal carbon

makes attack at the electron-rich oxo, to yield epoxide,

preferred in relation to attack on the high valent Ru, to

yield oxometallocycle. These open structures are analogous

to those species which were seen on the non-concerted [2+2]

pathway videe supra). The minima for the formation of a

bound epoxide gave approximately octahedral L-M-L angles,

RRuo = 2.1A (compared to 2.07A and 2.03A for the compounds
trans-[RuCl4(OH)(NO)]2- and trans-[Ru(NH3)4(OH)(NO)]2+ 96)

RCC = 1.50A(1.47A in ethylene oxide96), RCO = 1.43K (1.44A

in ethylene oxide96).



Comparison of the [1+21 and [2+21 Intermediates

The preference for formation of a bound epoxide over an

oxometallocycle intermediate can be analyzed by using basic

frontier orbital concepts. The oxometallocycle arises as a

result of the formation of aCO and aMC bonds and two O

lone pairs. Formally, two electrons are transferred from

the ethylene substrate to the Ru atom of the RuN502+
















E *




,lL YLL
Ly*



d4-LsRuO C2H4




Figure 2-7: The MO interaction diagram between the T
symmetry MOs for an on-center, C2v geometry.













Q0
40




ob
04


QO


Figure 2-8: The mixing of the ethylene 7 MO that occurs as
a result of the perturbation off-center from a concerted
[1+2] pathway. The bonding combination reinforces C-O
bonding. The antibonding combination (singly occupied)
increases the orbital density on the terminal carbon.







31

With the INDO calculations showing the M-C bond to be

approximately covalent and the M-O a bond to be largely

ionic, a d5, Ru(III) complex results. The ground state of

the oxometallocycle as derived from the INDO/1 calculation

corresponds to the limiting case [Ru(III)N5(OC2H4-)]2+ with

a 14d62 4dv2 (4da-Ca)(aB-Ba)j electronic configuration. The

oxometallocycle is a seven-coordinate structure.104 Seven-

coordinate geometries have been shown to possess a

characteristic frontier orbital splitting of two non-bonding

d orbitals below three antibonding d-orbitals making a d5,

seven-coordinate Ru(III) disfavored in relation to the bound

epoxide which is a stable, low spin d6, octahedral geometry.

The INDO/1 calculation yields a Jd y2 d 2 dxz2 dz2 dx2-

y20 electronic configuration for the bound epoxide. In
fact, geometry optimization of an oxometallocycle with a

starting structure in which the Ru-C and Ru-O bond lengths

were 2.12A and the C-0 and C-C bond lengths were 1.50A

yields a bound epoxide minimum.



Conclusions

The theoretical results described here are in good

accord with available experimental evidence. First, a

charge-transfer complex, as has been proposed for

epoxidations by ferryl and chromyl porphyrin systems,31,101

is found as a minimum on the PES for RuN502+'C2H4 Second,

least motion [1+2] and [2+2] pathways are found to be







32

unfavorable, as for organic analogues, so that non-concerted

processes are induced. The perturbation off-center causes a

dissymmetry in the C-O bonding, leading to a non-concerted

process. As the first C-O bond is formed, C-C t bonding is

decreased, and the rotational barrier is lowered. For an

open structure an 8.8 kcal/mole rotational barrier was

calculated (3 kcal mol-1 for ethane, 65 kcal mol-1 for

ethylene). In the epoxidation of cis-stilbene 5% isomerized

product is found.43 This suggests somewhat hindered

rotation about the C-C bond; and that the rate of formation

of the second C-O bond is competitive with C-C bond

rotation. If this barrier were lower as in ethane more

isomerized product (trans-stilbene oxide) would be expected

in the epoxidation of cis-stilbene.

The slower rate for cis-stilbene (observed to be one

order of magnitude less than that for the trans isomer) is

consistent with a preference for canted pathways.

Experimental and theoretical evidence105 shows that trans-

stilbene is planar while cis-stilbene is not. Thus, for the

preferred canted approach modes the cis isomer is expected

to have less favorable steric interactions.

In simple chemical terms, this reaction is a

nucleophilic attack of the olefin directed at the oxo

oxygen. For the reasons that we have already discussed in

detail the oxygen atom transfer step proceeds in a non-

concerted manner, i.e. one C-0 bond forms and then the







33

second. Since this is a nucleophilic attack, a positive

charge builds up on the terminal (non-bound) carbon of the

substrate. Simple charge considerations lead one to

predict, and INDO/1 calculations support, that C6+-O6-

interaction will be more favorable, and lead to bound

epoxide, than C6+-Ru6+ interaction. The bound epoxide is a

low spin, d6 pseudo-octahedral complex and therefore, quite

stable. The metal does not interact with the substrate

directly but as an "electron sink" for the two electrons on

being reduced from Ru(IV) to Ru(II). Exploring the

interaction of the Ru-oxo complex with a substrate that

possesses good donor ability, e.g. alcohols, will be

interesting too see if the metal can play a direct role in

oxidation.














CHAPTER 3
ALCOHOL OXIDATIONS

Introduction


Most of the proposed mechanisms for oxidations by P-450

models have concentrated on attack of the substrate on the

oxo moiety, followed by oxidation, and subsequent removal of

the oxidized product. Our previous INDO/1 investigation of

epoxidation, using the model complex of this study,

indicates that attack of the olefin directed at the oxo is

the most favorable pathway and that bonding of the olefin to

the metal does not occur. In this chapter we wish to

investigate the oxidation of a substrate, i.e. alcohols,

with good donor ability to see if the metal can play a

direct role in oxidation or merely act as a reservoir for

electrons. Six-coordinate Ru(IV)-oxo complexes have been

shown by Meyer and co-workers53,106 to be active alcohol

oxidations catalysts. Two pathways are considered:

activation of the alcohol C-H bond by the oxo and

coordination of the alcohol to the metal.



Calculations

A molecular orbital analysis of different modes of

interaction between cis-[Ru(HN=CH-HC=NH)2(NH3) (0)]2+







35

(hereafter referred to as RuN502+) and methanol was

performed using the INDO/1 method.80,81 The calculation of

both open- and closed-shell molecules was carried out using

the Unrestricted Hartree Fock formalism95 so as to allow for

the direct comparison of energies. Geometry optimizations

were of the gradient driven type.97 Convergence of the

gradient to stationary points on the potential energy

surface was assumed when the maximum component of the

gradient was less than 1.0 x 10-4 hartrees/bohr in all

cases. This led to energies which minimized within 1.0 x

10-2 kcal mol-.



Results and Discussion



Electronic Structure of the Methanol Fragment

The electronic structure of the alcohol needs no in-

depth discussion: the two lone pairs of valence orbital

theory are combined in MO theory to yield two non-bonding

MOs (NBMOs): a symmetric or a non-bonding MO and an anti-

symmetric or f NBMO. The a and T labels for the NBMOs are

with respect to the mirror plane of the alcohol (which

contains the hydroxyl H, O and the alpha C). Thus, alcohol

ligands may act as a w donor (albeit not to the extent of a

chloride ligand) to the metal complex in a coordination mode

in which Ca, O, H and Ru are co-planar.







36

C-H Activation at the Oxo Moiety

The relative energetic of C-H activation by the oxo of

the Ru-oxo complex were compared to those of coordination of

the alcohol to the ruthenium. Two limiting trajectories for

the interaction of the C-H bond of MeOH to the oxygen of the

Ru-oxo moiety were considered a triangular pathway (i.e.

the C-H bond of the methanol forms a triangle with the oxo

and approaches in a perpendicular fashion), and a linear

pathway (i.e. the Oxo-H-C angle is 1800).

A potential energy curve for both of the C-H activation

pathways, i.e. linear and triangular, was constructed as

follows. First, the geometries of the reactant fragments

(i.e. MeOH and RuN502+) were kept constant. Second, single

point INDO/1 calculations were used to evaluate the energy

and other properties at various separations of the oxidant

and substrate (as measured by the Oxo-H internuclear

distance, Roxo-H) from Roxo-H = 3.2A to Roxo-H = 0.94A in

approximately 0.1A increments.

The triangular trajectory was calculated to be

unfavorable at all internuclear distances with RC-Oxo < 2A.

At points on the potential energy surface (PES) in which the

C-Oxo distance is less than a typical covalent C-0 bond

(=1.4A96), but greater than an O-H covalent bond (=0.96A96),

the potential energy surface is extremely repulsive.














RuNsO2+ + MeOH


10.00 -



5.00



0.00



-5.00



-10.00


ODOD0 T(MeOH)
A_ V CH)
0.0.0.0.0 V(RuO)


-15. 5 1. 05 1.. ... 45 ..... 1 .... '85
0.85 1.05 1.25 1.45 1.65 1.85


Figure 3-1: Potential Energy Curve for the translation of
the entire methanol fragment relative to the Ru-oxo fragment
in a linear orientation and with C-H and Ru-0 stretching
modes imposed upon the ion-molecule minimum.







38

The potential surface for the linear C-H-Oxo trajectory

provides an energetically more favorable orbital interaction

and less steric repulsions, Figure 3-1. The energy for

RuN502+MeOH at ROxo-H 1.7A is equal to the sum of the

energies of isolated RuN502+ and MeOH fragments. At ROxo-H

z 1.4k, the sum of the ionic radii of 02- and H+,107 a

slight barrier of 3.1 kcal mol-1 to further translation of

the methanol, Figure 3-1, is found originating from a

repulsive two-orbital, four-electron interaction between the

Ru-O a non-bonding MO (predominantly O 2pz) and the

methanol C-H a bond. Evidence for this is a polarization

of the Ru-oxo active space such that electron density (= 1

electron as determined by a Mulliken Population Analysis) is

transferred from the O 2pz to the largely unperturbed RuO .*

MOs. At ROxo-H w 1.2k, Figure 3-1, there is a minimum which

is calculated to be 10.6 kcal mol-1 below the separated

fragments. Shortening of ROxo-H below 1.2A causes a large

increase in the total energy of the Ru-oxo/alcohol system,

Figure 3-1.

The linear complex resembles a three-center, four-

electron hydrogen bond. The atomic bond index between the

oxo and the hydrogen at the minimum is 0.25 while that for

the C-H bond is 0.65 (0.90 in "free" methanol) indicating a

weak oxo-H bond. The RuN502+MeOH interaction is

significant, but it should be remembered that this is a

calculation on a gas phase system and in solution this







39

interaction has to compete with the solvent shell around the

complex (water or alcohol). Formation of this weakly bound

species would be difficult in solution.

Assuming the linear ion-molecule complex can occur, two

further geometric perturbations from the linear ion-molecule

minimum, Figure 3-1, were explored with the intent of

studying the energetic of closing the oxo-H bond to

covalent distances, ROxo-H = 0.96A.96 First, a C-H

stretching mode, VCH, was calculated. The total energy

decreases108 for expansion of the C-H bond by about 0.04A

(and thus closing ROxo-H by a like amount), after which the

total energy rises, Figure 3-1. Second, the Ru-0 bond

stretching mode, VRuO, was explored. This yielded a

decrease in the potential energy curve for 6R(Ru-Oxo) a

+0.06A, after which the energy rises steeply, Figure 3-1.

Thus, direct formation of a covalent oxo-H bond by either of

these processes alone is not indicated. The third

possibility is a concerted stretching along the C-H and Ru-O

bond distortion modes. The difference in ROxo-H between the

ion-molecule complex (ROxo-H 1.20A) and a covalent oxygen-

hydrogen bond(ROH = 0.96-96) is 0.24A. Stretching the C-H

and Ru-O bonds each by 0.12A, in the ion-molecule complex,

is calculated to increase the total energy by 52 kcal mol-1.

This energy is quadruple the experimental activation

energy.53107







40

Possible Seven Coordinate Intermediates

In their treatment of seven-coordinate complexes,104

Hoffmann and co-workers noted a definite pattern of two low

and three higher energy frontier orbitals (all largely metal

d in character). This is significant, in the present case,

as it confers relative stability (1st and 2nd order Jahn-

Teller stability) to d4 seven-coordinate transition metal

complexes. The greater size of second-row transition metals

lends further weight to the possibility of a seven-

coordinate intermediate or transition state. Thus, a seven-

coordinate Ru(IV) complex might be expected to be a

relatively stable intermediate (or transition state).

Experimentally,53,107 there is a large, negative AS* for the

oxidation of MeOH by [Ru(bpy)2(py)(0)]2+ indicating that a

strongly associated complex exists in the rate determining

step. Also, seven-coordinate Ru(IV) complexes have been

isolated and characterized by Pignolet and co-workers.109

Based on these observations, an inner coordination-sphere

expansion pathway could be proposed as a viable alternative

or competing pathway to the C-H activation scheme.

Extended HUckel calculations104 on d4 ML7 (L = a donor,

7 donor, or 7 acceptor) indicated that the pentagonal

bipyramid (PBP7) is the preferred geometry over the capped

octahedron (C07) and the capped trigonal prism (CTP7) but

only by a slight amount.








41

Seven-coordinate complexes can be obtained from an

octahedral molecule by distortion to form the seven-

coordinate precursor, followed by coordination of the

seventh ligand. The term precursor, as used in the rest of

this paper, refers to a structure in which the six-

coordinate Ru-oxo complex has been distorted towards an

arrangement which will allow for coordination of a seventh,

i.e. ROH, ligand to produce the types of seven-coordinate

complexes studied. Three seven-coordinate structures were

investigated the pentagonal bipyramid (PBP7), the capped

octahedron (C07), and the capped trigonal prism (CTP7). The

CTP7 possesses a higher calculated activation energy (15.6

kcal mol-1) than does the PBP7. The C07 was discounted due

to significant steric hinderance.

The PBP7 is formed by the expansion of the equatorial

plane to accommodate one extra ligand (i.e. 90 L-M-L angles

are compressed to 720). Geometry optimization of the six-

coordinate Ru-oxo complex, on the triplet manifold, was

performed. A minimum with a Ru-oxo bond length of 1.96A, an

average Ru-N bond length of 2.06A, and a pseudo-octahedral

metal environment was obtained. These results agree very

well with the experimental determinations of Che98 which

yield Ru-O and Ru-N bond lengths of 1.77A to 1.86A and 2.07K

to 2.16A, respectively, for Ru(IV)-oxo complexes with

nitrogen ligands. The perturbation of the equatorial N-Ru-N

angles to a structure with 720 L-M-L angles (PBP7 precursor)







42

was calculated to cost 9.6 kcal mol-1. The ground state

remains a triplet.

Due to the lower calculated activation energy for the

formation of the pentagonal bipyramidal complex, along with

the structural and electronic similarity of seven-coordinate

complexes (as evidenced by their fluxionality110), only the

coordination of MeOH to the PBP7 precursor was investigated.

The calculations indicate no activation barrier for

coordination. The total energy decreases as RMO decreases,

up to a metal-oxygen (from ROH) distance of 2.3 2.4k.

Covalently bound seven-coordinate species are formed; the

calculated binding energies are 30.0 kcal mol-1 (triplet)

and 30.3 kcal mol-1 (singlet). Re(Ru-O) is 2.40A for the

triplet RuN50(ROH)2+ and 2.35A for the singlet analogue; the

singlet state is 18.2 kcal mol-1 below the triplet state at

R(Ru-o) = 2.35A for RuN50(ROH)2+

If we take the calculated minimum of the metal-

coordination route and rotate the methanol ligand about an

axis perpendicular to H-O-Ca plane and passing through the

oxygen, Figure 3-2, a further 38 kcal mol-1 of stabilization

is calculated. Although this amount seems spurious it does

point to a stabilization from increasing C-H"**Oxo

interaction. Thus, we are left with a cyclic association

complex and one which seems to combine the attributes of

both proposals, Figure 3-2.







43

Further Reaction of a Seven-Coordinate Complex

While the main goal of this research was to compare

initial substrate/oxidant interaction, it is interesting to

probe the nature of any subsequent reaction of the Ru(IV)-

oxo/alcohol complex to produce the Ru(II)-OH2 and carbonyl

products. The mechanism which Meyer and co-workers have

proposed after an exhaustive mechanistic study53,106 is

given in Equations 1 3.


N5Ru-O2+ + HC(OH)R2 N5Ru-O2+,HC(OH)R2 (1)

N5Ru-O2+,HC(OH)R2 [N5Ru-O"-H-C(OH)R2 2+

N5Ru-OH+,C(OH)R2+ (2)

N5Ru-OH+,C(OH)R2 -+ N5Ru(OH2)2+ + O=CR2 (3)



This mechanism entails association, Eqn. 1, "hydride

transfer", Eqn. 2, followed by a proton transfer, Eqn. 3.

The most favorable conformation for a donor with one a and

one filled n orbital (a single-faced donor104) in a PBP7 is

with the n orbital of the ligand in the equatorial plane of

the complex. Maximum overlap of the ROH a NBMO with the

virtual dS (the vacant hybrid orbital which points in the

direction of the missing seventh ligand) is provided.

Destabilizing two-center, four-electron repulsions are

minimized by keeping the i functions orthogonal. In this

conformation the alkyl or hydroxyl hydrogens of the alcohol

are ideally positioned to transfer to the oxo oxygen. One

of the orbitals that the alcohol is using to transfer







44

electron density to the metal is almost a pure p orbital (Ci

= 0.92 in an INDO/1 calculation) and the other is

approximately 67% p-character ( = sp2 hybridization), p

orbital density is transferred to the metal, leaving

proportionately more s electron density on the oxygen. In

accord with Bent's Rule111 and in comparison to simple

hydrocarbons as the percentage of s character increases the

angles involved, the C--O-H angle of the alcohol in the

present case, increase. This has the effect of moving the

hydrogens alkyll and hydroxyl) closer to the Ru-O moiety of

the RuN502+ complex. At this point two things can occur:

the two hydrogens (i.e. those which convert the alcohol to

the carbonyl) transfer in two separate and discrete steps or

these two hydrogen atoms can transfer at the same time in

one concerted process. Assuming the two-step process takes

place, a complex with the formula [RuN5(HOCH2)(OH)]2+ will

be formed. The nature of this reactive moiety is subject to

much speculation,53,106 as is the nature of the H species

which is transferred two limiting configurations would

seem to be reasonable, [Ru(IV)N5(HOCH2-)(OH-)]2+ and

[Ru(III)N5(HOCH2")(OH-)]2+. For a seven-coordinate pathway

to be favorable the former configuration would have to make

a significant contribution to the ground state wavefunction.

The second H transfers to form a Ru(II)OH2-carbonyl species.













0


A


Ru-
R u <-------

H


Figure 3-2: Structure of "cyclic" intermediate which
combines C-H-Oxo and M-O pathways.


H








46

This transfer of the second hydrogen can be envisioned as

occurring by two different processes: the second hydrogen

transfers directly to the Ru-OH or it may be abstracted by a

solvent molecule which then reacts with Ru-OH to yield Ru-

OH2. After the second hydrogen transfer two things happen:

1) The strong alkoxide/alkoxy radical ligand is

replaced by a weaker 0 bound carbonyl ligand.

2) The very strong oxide (02-) ligand is replaced with

the weaker OH2.

Thus, the electron density on the metal increases (also by

donation from the alcohol). Since seven-coordinate

complexes are averse to d orbital counts greater than 4, the

carbonyl jettisons and the six-coordinate RuN5(OH2)2+ is

obtained.

Our present results deal most directly with the nature

of the association complex in Equation 1. What light can we

shed upon the subject? We proceed by looking at the

arguments against a seven-coordinate intermediate. They

are: no precedent for coordination sphere expansion, rate of

180 exchange (between RuO and 180H2) is slow, and no

spectrophotometric evidence exists for such an

intermediate.53,106

As we have seen there are precedents which allow us to

propose a seven-coordinate association complex. First,

simple frontier MO considerations point to stability for d4

seven-coordinate complexes104 Second, d4 complexes of the








47

iron triad have been characterized by Pignolet109 and an

[Os(IV)(bpy)(PPh3)2H2(CO)]2+ by Meyer112 (note that the

ionic radius of the hydride is quite large, 208pm113).

We would expect water to form association complexes

since it is a comparable ligand to alcohols and is

sterically less demanding. However, once this association

complex is formed the water is at a disadvantage with

respect to scrambling of the 180 label for various reasons.

First, the homolytic bond strength of the O-H bond (119 kcal

mol-1) in water is significantly larger than the O-H bond in

alcohols (= 103 kcal mol-1) and the C-H bond in alcohols (

92 kcal mol-1). Experimentalll14115 and theoretical74

analyses of Ru-0 systems have pointed to the similarity of

the RuO moiety to 02 and the radical nature of RuO

complexes. Thus, we would expect the "hydride transfer" to

entail hydrogen atom abstraction followed by intramolecular

electron transfer. Given the much greater strength of the

O-H bond in water, H atom abstraction would be difficult.

The second barrier to H atom transfer from water arises from

the fact that the H atoms of water would be pointing away

from the RuO f MOs (the orbitals where unpaired electron

density resides to the greatest extent). If this scrambling

arises from proton transfer the dihydroxy species formed

will contain two symmetry inequivalent hydroxyl groups.

This may be the cause of the observed lack of 180

scrambling. The lack of spectrophotometric evidence for a







48

seven-coordinate association complex may be due to there

only being a very small, but kinetically important,

concentration of this complex.

There are other benefits of the combined C-H'*'Oxo/M-0

interaction pathway. In the linear C-H"*'Oxo pathway a

repulsive two-orbital, four-electron interaction between the

MOs which corresponds to the C-H a bond and the Oxo lone

pair must necessarily arise. Bringing the C-H bond from the

side allows for greater interaction between the C-H bond to

be activated and the RuO T system the orbitals which

essentially define the reactivity of these and most metal-

oxo complexes.13 In a linear C-H'**Oxo pathway the C-H bond

to be activated is orthogonal to RuO v MOs making direct

interaction impossible. In this "cyclic" intermediate the

coupling of electron transfer and RuN502+ vibrational modes

that Meyer has proposed53'106 may occur. Incidentally,

Groves116 has proposed for C-H activation of alkanes by

ferryl porphyrin systems just such an approach (i.e. the C-H

bond approaches the metal-oxo moiety from the side and

directed at the oxo). Also, coordination of the alcohol to

the metal provides for activation of the alcohol by transfer

of electron density from the alcohol to the high-valent Ru

atom. Our INDO/1 results indicate a significant

acidification of the hydroxyl proton upon coordination to

the metal. Thus, proton abstraction is facilitated by pre-

coordination to the metal.









Summary

The INDO/1 results indicate that for the oxidation of

alcohols by a six-coordinate Ru(IV)-oxo complex, alcohol

coordination is comparable with, if not more favorable than,

a C-H activation mechanism. The linear interaction of a C-H

bond in methanol with the oxo oxygen leads to a loosely

bound ion-molecule interaction with a oxo-H distance of

1.2A. The formation of a covalent oxo-H bond,

corresponding to hydrogen transfer, is costly in terms of

the total energy of the system for various modes. The

pentagonal bipyramid is calculated to present the best

combination of favorable electronic and steric interactions.

The calculated barrier for the formation of the PBP7 complex

is 9.6 kcal mol-1. Further interaction after the seven-

coordinate association complex is formed increases C-H"'Oxo

interaction. Thus, the preferred pathway combines both

proposals, i.e. pre-coordination of the alcohol and C-H''O

interaction.

This study of alcohol oxidations by a six-coordinate,

high-valent metal-oxo complex with nitrogen ligands (i.e. a

P-450 model) shows that when the substrate has a good donor

ability direct metal participation in the oxidation pathway

is plausible. Most proposed mechanisms for oxidation by P-

450 models involve the metal only indirectly as a receptor

for two electrons. Our results suggest that metal

participation may be important; in the oxidation of








50

substrates that possess good donor ability, i.e. alcohols,

amines, and perhaps sulfides, the substrate may ligate to

the metal in the course of the oxidation. Unfortunately,

the Ru-oxo complex which we have studied has not been

reported to be active for amine dealkylations or the

hydroxylation of alkyl amines. This six-coordinate Ru-oxo

complex has been shown to oxidize sulfides54 which are,

therefore, good subjects for further theoretical scrutiny.














CHAPTER 4
SULFIDE OXIDATIONS

Introduction

In the present research attention has been concentrated

on the oxidation of sulfur to sulfoxides and sulfones. The

main purpose of this chapter is to continue our

investigations into the oxidation of organic substrates

using the same model cis-[Ru(HN=CH-HC=NH)2(NH3)(0)]2+

catalyst. Interaction of the sulfur with the oxygen and

metal atoms of the Ru-oxo active site is considered. The

organic sulfides are good donors and acceptors and should

provide an interesting case for comparison with the other

substrates (olefins and alcohols). Furthermore, the same

moiety of the substrate (the S atom) is interacting with

either end of the metal-oxo oxidant. The effect of the 3d

orbitals on the bonding interactions of the sulfur are of

interest and discussed where pertinent.



Calculations

The majority of calculations performed in the present

work were of the semi-empirical INDO/180'81 variety, unless

otherwise noted. All INDO/1 calculations were run two

times: with and without 3d orbitals on the sulfur atoms. It








52

was found in the INDO/1 calculations that excluding 3d

orbitals had the same effect as employing a severely

contracted (exponent3d = 100.) Slater-type function for the

radial portion of the atomic orbital. An exponent of 1.731

for the S 3d orbitals was chosen in those INDO/1

calculations in which the sulfur atom had a 3d basis. Ab-

initio calculations were run using the HONDO-5 program.117

Ab-initio and semi-empirical geometry optimizations were of

the gradient-driven type and included electron correlation

(Moller-Plesset perturbation theory to second-order118) for

the ab-initio optimizations.



Results and Discussion



Electronic Structure of the Ru-oxo model.

The electronic structure of the six-coordinate Ru(IV)-

oxo complex has been discussed in detail elsewhere and need

not be repeated here.



Electronic Structure of the Sulfide Fragment

The proper modeling of the subtle effects of adding

electronegative oxygen atoms to sulfur and

excluding/including 3d orbitals is crucial for this study.

To ascertain the suitability of the INDO/1 method for this

research, geometry optimizations of SH2, SOH2, SO2H2 where

carried out and compared with experiment, where available,







53

or with ab-initio calculations. The results of the INDO/1

calculations are summarized in Table 4-1, and are, in

general, quite good. The electronic structure of C2v,

hydrogen sulfide needs no in-depth discussion.119

The effect of the 3d orbitals is always of interest

when discussing the electronic structure of the second-row

main group elements. The 3d orbitals may mix, in a second-

order way (hybridization in valence bond terminology), into

the correct symmetry sp-basis molecular orbitals of hydrogen

sulfide. The 3d orbitals transform as 2a, (z2,x2-y2) + a2

(xy) + b, (xz) + b2 (yz) under the C2v point group. Since

the 3d orbitals are higher in energy than the 3s and 3p AOs,

their effect will be to mix into the MOs which contain these

AOs in a bonding manner,120 and lower their energies. An

INDO/1 calculation indicates this orbital energy lowering to

be minimal, with values ranging from 2 to 25 millihartrees

(1 millihartree = 0.6 kcal/mole). An INDO/1 geometry

optimization of hydrogen sulfide was run with and without

the 3d orbitals. The theoretical results are summarized in

Table 4-1. The inclusion of the 3d orbitals gives a lower

energy (expected on the basis of the variational principle),

as well as shorter bond lengths and smaller bond angles,

indicating that the 3d orbitals are strengthening the S-H

bonds. The bond lengths and angles are closer to

experimental values for the basis which includes 3d AOs.













Table 4-1
Effect of 3d AOs on the Geometry of Sulfur Species


Method

INDO/1

INDO/1

expt.

INDO/1

INDO/1

ab-initioc

INDO/1

INDO/1

ab-initiof


3d exponent

1.731

100.0b



1.731

100.0

?d

1.731

100.0

?


RSHa

1.37

1.36

1.33

1.37

1.37

1.34e

1.37

1.38

1.35


RSO e(HSH)

93.39

96.12

-- 92.2

1.52 90.7

1.73 95.8

1.47 100.

1.52 104.1

1.74 103.3

1.44 ?


e(OSO)



--

--








118.3

127.0

121.6


a Bond lengths (A); bond angles (0); energies (a.u.).

b The value of 100.0 was used for the S 3d AO exponent
to replace the default value of 1.731.

c See Reference 121.

d A question mark indicates that the specified quantity
was not reported by the authors.

e A value of 1.34A for the S-H interatomic distance was
assumed and kept constant.

f See Reference 122.


Species

H2S

H2S

H2S

H2SO

H2SO

H2SO

H2SO2

H2SO2

H2SO2








55

Electron withdrawing groups on the sulfur will increase d

orbital participation by d orbital contraction and energy

lowering of these orbitals. Thus, H2SO and H2SO2 will have

a larger 3d orbital "effect." This is seen in the

calculations which give a total 3d orbital population of

0.028, 0.403 and 0.767 electrons for isolated SH2, SOH2 and

H2SO2, respectively. The INDO/1 method gives reasonable

values as compared to the ab-initio results in terms of

geometry optimizations. For SOH2 and SO2H2, the effect of

the 3d AOs on the S-H bond lengths is minimal while S-0 is

much reduced. The S-0 bond changes from an approximately

single bond value (the average of X-X in HXXH, X = S and 0,

is 1.77A) to double bond value (the S-0 bond length in SO2

is 1.43A).96 The calculated Atomic Bond Indices are

consistent with this interpretation, changing from 1.87 to

0.65 upon the severe contraction of the 3d orbitals. The

effect of 3d orbital participation in analogous compounds is

the subject of two excellent papers by Hoffmann and co-

workers123 and by Van Wazer and Absar.124



Direct Sulfur-Oxygen Interaction: Linear Coordination Mode

The interaction of the sulfur atom of the hydrogen

sulfide molecule with the oxo oxygen of cis-[Ru(HN=CH-

HC=NH)2(NH3)(0)]2+ was investigated in two distinct modes:

linear, with the Ru, O, S and H atoms (from SH2) co-planar

with the Ru-oxo bond and bent with respect to this plane.








56

This affords two different modes of interaction. In the

linear interaction mode there is both a a and r interaction

with the two lone pair combinations. In the bent

interaction mode there is only a a interaction with one pair

while the other remains essentially non-bonding.

The linear interaction between the sulfur and oxo

decreases the total energy of the Ru-oxo/sulfide system

until a covalently bound minimum is reached, regardless of

the sulfur basis (sp or spd) or multiplicity, Figure 4-1.

At 2.07A, spd basis and triplet spin state, there is a

slight polarization of the Ru-oxo fragment which entails the

transfer of 0.25 electrons from the RuO a to t manifold.

This transfer of electron density from the a to v is

reminiscent of the polarization of RuO by the C-H bond of

methanol in the C-H activation mode of alcohol oxidations.

However, unlike that case there is no activation barrier

present. This is a result of the polarization of the Ru-oxo

MOs being much less in the sulfide case than in the alcohol

case (0.25 vs. 1.0 electron). A Mulliken Population

Analysis125 indicates that the electron density comes from

the O 2pz (the major component of the oxo lone pair). The

S-H atomic bond index changes little (0.95 to 0.93)

indicating that the orbital responsible for this

polarization is the non-bonding, molecular orbital. The sum

of the 3d orbital populations remains at 0.02 electrons.

However, the removal of the sulfur 3d orbitals increases the








57

polarization of the RuO caused by the hydrogen sulfide

significantly. The dipole moment decreases upon inclusion

of the 3d orbitals from 3.58 to 1.98 Debyes. Thus, the

effect of the 3d orbitals is mainly electrostatic.

The minima for linear substrate/oxidant interaction

occur at RSO = 1.48 to 1.88A depending on the multiplicity

and S basis, Figure 4-1. The sulfur 3d AOs are important in

the formation of covalently bound S-0 species. The S-O bond

distances in SO2 and SO3, which have some double bond

character, are 1.43A and 1.43A, respectively.96 The

increase in energy on the PE curve is caused by too large a

polarization of the O lone pair by a pseudo-symmetry

orbitals on sulfur. For example, on the triplet curve, sp

sulfur basis, the translation of the sulfide from the

minimum energy structure is accompanied by a large increase

in transfer of electron density from the RuO a to T manifold

and a jump in the positive charge of the sulfide (+0.261 at

the minimum to +0.629 at RSO = 1.58k). The 3d orbitals act

to lower the dipole moment of the sulfide and thus this

excessive polarization of the RuO moiety occurs at a smaller

S-O interatomic distance. In both cases there is a shift

from a triplet to singlet ground state which allows the S-0

bond length to shorten further. This is a result of two

electrons being transferred from the sulfide to the Ru-oxo

fragment to produce a d6 pseudo-octahedral complex.

















25.00




-25.00




-75.00




-125.00




-175.00 -
1.00


S-- --------- -


A A"'4 "


I





I I4 D~oDDo~

'ci,


sp-basis/singlet
sp-basis/triplet
spd-basis/triplet
spd-bosis/singlet


2.00 3.00 4.00 5.00
2.00 3.00 4.00 5.00


Figure 4-1: PE curves for interaction of Ru-O + SH2 (spd
or sp S basis and singlet or triplet multiplicity) in the
linear mode.


V v







59

In all cases of interaction directed at the oxo the bonding

between the sulfur and the ruthenium is minimal with atomic

bond indices close to zero.

In summary, a linear interaction between the sulfide

substrate and the Ru-oxo catalyst is favorable up to

covalent S-O distances. The 3d orbitals play an important,

if indirect, role. They serve to lower the dipole moment of

the sulfide by making the S-H bonds more covalent or

equivalently by reducing the amount of charge separation in

these bonds. This lowering of the dipole moment helps what

is largely an electrostatic interaction be energetically

feasible by lowering the amount of electronic reorganization

which must occur within the Ru-oxo fragment before covalent

S-O bond formation. The n symmetry non-bonding MO has a

minimal effect on the bonding.



Direct Sulfur-Oxygen Interaction: Bent Coordination Mode

The geometric perturbation of the hydrogen sulfide

fragment from a linear to a bent coordination mode serves to

alter one major factor. The two lone pairs of H2S can be

combined to yield a and T NBMOs. In the linear interaction

mode R2S is a a and T donor to the Ru-oxo moiety.

Distortion to the bent mode diminishes r donor ability.

Thus, the distortion from a linear to bent orientation

reveals the effect of the f-donor ability of the substrate

on its interaction with cis-[Ru(NH=CH-CH=NH)2(NH3)(0)]2+







60

The formation of the bent structure is of particular

interest since this allows the local OSH2 moiety to attain

the geometry of an O-bound sulfoxide, i.e. approximately

tetrahedral geometry about the sulfur.

The results for this geometric distortion with various

spin multiplicities and sulfur basis sets are given in

Tables 4-2 to 4-5. A comparison of the potential energy

surfaces, Tables 4-2 to 4-5 for the singlet and triplet spin

states shows that the effect of the 3d AOs is similar to

that for the linear orientation. The 3d orbitals allow the

favorable region of the potential energy surface to be

extended to covalent S-0 distances. This is due, once

again, to a diminution of the sulfide dipole moment by

mixing of the S 3d AOs into the S-H bonding orbitals in a

bonding way. A comparison of the potential surfaces with a

sulfur spd basis indicates that as the interaction becomes

more covalent the singlet and triplet manifolds approach

each other. This is due to a transfer of electron density

from the sulfide fragment into the Ru-oxo fragment videe

supra).

There is for all spin states and sulfur bases

considered, a stabilization upon the conversion from a

linear to bent sulfide orientation. The origin of this

phenomenon is the stabilization of the S-0 T* upon bending.












Table 4-2
Conversion from Linear To Bent SH2(spd)


Orientation(3)a


R(S-O)A 1.08 1.58 2.07 2.57 3.07


1800

1750

1600

144.750

1300

1150

109.50


1.629C

1.630

1.615

1.599

1.583

1.580

1.583


0.056

0.055

0.056

0.035

0.014

0.002

,Oi,"d


0.068

0.068

0.069

0.062

0.058

0.054

0.054


0.083

0.083

0.107

0.085

0.087

0.090

0.092


0.087

0.087

0.087

0.090

0.091

0.092

0.093


a The first number in parenthesis of the title for Table
4-2 through Table 4-5 refers to the basis set,
principal quantum number = 3, used for the sulfur atom;
the second number in parenthesis refers to the overall
spin multiplicity of the interacting Ru-oxo/SH2
systems.

b 4 is the angle between the SH2 molecular plane and the
vector which describes the Ru-O bond.

c The energies are given in hartrees relative to the
calculated minimum.

d "O" = -131.234 hartrees; other energies are relative to
the "0" value and are in hartrees.












Table 4-3
Conversion from Linear To Bent SH2(sp) Orientation(3)


R(S-O)A

ta 180.

1750

1600

144.750

1300

1150

109.50


1.08

1.635a

1.638

1.596

1.672

1.590

1.587

1.590


1.58

0.071

0.071

0.062

0.045

0.028

0.017

0.015


2.07

0.011

0.011

0.009

0.005

0.003

0.001

,Ob


2.57

0.023

0.023

0.026

0.025

0.027

0.030

0.031


See footnotes in Table 4-2.

"0" = -130.965 hartrees.


3.07


0.027

0.027

0.028

0.029

0.030

0.032

0.033













Conversion from Linear


Table 4-4
To Bent SH2(spd)


Orientation(1)


R(S-O)A 1.08 1.58 2.07 2.57 3.07


1800

1750

1600

144.750

1300

1150

109.50


1.203a

1.204

1.188

1.226

1.146

1.144

1.148


0.081

0.081

0.064

0.037

0.015

0.002

",nbb


See footnotes in Table 4-2.

"0" = -131.273


0.266

0.267

0.273

0.254

0.247

0.244

0.243


0.316

0.316

0.317

0.318

0.320

0.323

0.324


0.329

0.322

0.329

0.330

0.332

0.333

0.335













Table 4-5
Conversion from Linear To Bent SH2(sp) Orientation (1)


R(S-O)A

ta 1800

1750

1600

144.750

1300

1150

109.50


1.08

1.629a

1.630

1.615

1.599

1.583

1.580

1.583


1.58

0.056

0.055

0.056

0.035

0.014

0.002

0o"bb


2.07

0.068

0.068

0.069

0.062

0.058

0.054

0.054


2.57

0.083

0.083

0.107

0.085

0.087

0.090

0.092


3.07

0.087

0.087

0.087

0.090

0.091

0.092

0.093


See footnotes in Table 4-2.

"0" = -130.996







65

This orbital drops in energy below the antibonding Ru a and

provide a low-energy orbital (non-bonding) for the two

electrons contained there.



Geometry Optimization of the O-bound Sulfoxide

The minimum calculated for the translation of the

sulfide fragment relative to the Ru-oxo fragment does not

correspond to the global minimum. There are degrees of

freedom which have not been allowed to relax. These are

mainly metal-ligand in nature and arise as a result of the

change in the oxidation state of the metal. To more closely

compare our results with the available experimental

information,126 a geometry optimization of all 3N-6 degrees

of freedom was carried out on the minimum obtained by

assuming no intra-fragment relaxation. The S-0 bond changed

slightly from 1.58A to 1.54K, as expected. This is in good

agreement with the 1.56A found experimentally for the 0-

bound sulfoxide in [Ru(II)C12(S-DMSO)3(O-DMSO)].126 The Ru-

O bond lengthens from 1.96A to 2.13A, once again in good

agreement with the value of 2.14A obtained for the above

compound. There was a discrepancy in the Ru-O-S angle

(103.80 calcd; 120.20 exptl). However, an analysis of the

isomerization of the 0- to S-bound sulfoxide revealed a soft

potential energy surface over a range of 500 for distortion

of this angle.










Coordination-Sphere Expansion

The seven-coordinate Ru(IV) complexes of Pignolet

possess sulfur ligands.109 For the sake of completeness

seven-coordinate structures were investigated. The

pentagonal bipyramidal (PBP7) structure alone was studied

due to the observed 0-0 steric interference observed for

alcohols (expected to be larger for the larger sulfur atom),

the greater calculated activation energy for the formation

of the capped trigonal prism precursor versus the pentagonal

bipyramid, 16 and 10 kcal/mole, respectively and the

energetic similarity of various seven-coordinate polytopes.

The interaction studied initially was a linear (a and

7) interaction in the equatorial plane of the PBP7. The

isolobal analogy127 dictates that the removal of a ligand

from a d4, ML7 complex will result in a vacant orbital

pointing in the direction of the missing ligand. This MO

corresponds to the sp3d3 hybrid of valence bond theory. The

a interaction will be a result of donation from the sulfide

a symmetry non-bonding MO into this vacant hybrid. The 7

orbitals are perpendicular in the least sterically hindered

conformation and are also expected to be perpendicular as

per Hoffmann's treatment of d4, PBP7 complexes with a single

faced v donor.104

The exclusion of the 3d AOs from the sulfur basis

resulted in sulfides which are not covalently bound, RMS is

> 3A at the calculated minimum. The inclusion of the 3d AOs







67

does not significantly increase the strength of the metal-

sulfur interaction as evidenced by the nearly flat minima at

2.5A 2.6A. The perturbation to a bent geometry will

increase the steric repulsions.



Summary

A molecular orbital analysis of the oxidation of

sulfides to sulfoxides was carried out. Direct oxygen atom

transfer was found to be the most favorable. Unlike alcohol

oxidation there is no energy barrier to the formation of a

covalently bound substrate/oxidant species. The formation

of the seven-coordinate precursors entails the expenditure

of energy. The effect of the sulfur 3d AOs was

investigated. The interesting result obtained was that the

3d AOs influenced the reaction in a largely electrostatic

manner. The inclusion of the sulfur 3d basis decreased the

dipole moment significantly by making the S-H bonds more

covalent. The larger dipole moment increased the

polarization of the Ru-oxo fragment, causing a

destabilization for the substrate/oxidant interaction to

occur at distances much larger than covalent S-O distances.

Without 3d AOs the formation of covalently bounded species

was not observed.














CHAPTER 5
OLEFIN OXIDATIONS BY RU-DIOXO COMPLEXES

Introduction

Since the discovery that the reaction of an Fe(III)-

porphyrin with iodosyl benzene(C6H5-I-0) could mimic the

reactivity of the 02/Cytochrome P-450 system,24 there has

been much research on metal-oxo species.13,128,129

Experimentally, a growing amount of work has been focused on

metal-dioxo, M(0)2, compounds;41,42,130 theoretically, less

attention has been paid to the electronic structure and

reactivity of these complexes.131 Griffith has derived

qualitative rules for predicting which geometric isomer is

the most stable as a function of d orbital population for

poly-oxo complexes.131b Jorgensen and Hoffmann have studied

the interaction of ethylene with OsO4 using the Extended

HUckel method.132 Rapp6 and Goddard have used Generalized

Valence Bond calculations to study the oxidation of ethylene

by various dO molybdenyl and chromyl species.70,71

Dioxo complexes are of particular interest for various

reasons. First, they are plausible intermediates in the

activation of 02 by a single metal center,130a,d Equation

1.86 Ru(II) species have been shown130a,d to regenerate the

active species in the presence of 02 without the need for







69

expensive and difficult to handle oxygen atom transfer

reagents.133

LnMq + 02 -----> LnM(O)2 (1)



Second, there is an interesting difference in the product

distributions for the cis- and trans-Ru(VI)(0)2 complexes;

the cis complexes130d lead to predominantly carbonyl

products formed by the cleavage of the C=C double bond; the

trans complexes41,130a yield epoxides as the major product.

Third, many interesting pathways may be envisioned for

oxidations involving these dioxo systems arising from the

similarity of the cis-RuO2 moiety with ozone.

It is the aim of the present research to get a better

understanding of the Ru(VI)-dioxo/organic substrate

interaction, to explore comparisons between the Ru(VI)-dioxo

and Ru(IV)-oxo complexes and investigate the causes of the

difference in reactivity between the cis- and trans-isomers.



Calculations

The calculations were carried out using the semi-

empirical INDO/1 method.80,81 The Restricted Open-Shell

Hartree-Fock (ROHF) formalism134 was used for open-shell

species. Gradient driven geometry optimizations97 were

carried out on various structures to yield stationary points

on the potential energy surface. Convergence to a minimum

on the PES was assumed when the gradient was less than 5.0 x







70

10-3 hartrees/bohr. The total energy of the species

converged well within 0.1 kcal mole-1 in all cases.


Results

Electronic Structure of Ru(VI)-dioxo complexes

The model complexes used in this present study are cis-

and trans-[Ru(VI)(NH=CH-CH=NH)2(0)2 2+ (hereafter referred

to as cis- and trans-RuO22+, Figure 5-1. Cis- and trans-

Ru(VI)O2 complexes have been used by various workers in the

oxidation of different organic substrates.41,130a,d

Optimization of the Ru-O bond lengths in the cis- and trans-

complexes show the equilibrium value to lie near 1.71A, in

excellent agreement with the values of 1.718A and 1.705A

reported by Che98 for Ru(VI)022+ complexes with nitrogen

ligands. Analogies have been made between the Ru-oxo

moiety65,74 and 02 (with respect to the covalency of the n

system and the presence of low energy singlet and triplet

states99) then the Ru(VI)-dioxo complexes can be considered

to be analogous to ozone.

The frontier orbital splitting diagrams for the cis-

divacant and square planar ML4 complexes are well known.

The electronic structure of cis- and trans-dioxo complexes

has also received attention in the literature.131b With

these simple models predictions about the electronic

structure of d2 dioxo complexes have been made. There are

serious deficiencies in these models, e.g. ignoring the







71

effects of mixing in the 5s and 5p AOs, the fact that these

models are one-electron, single-determinant in nature, etc.

The trans isomer is predicted to be more stable due to the

lower energy of the HOMO (dxy).131b An INDO/1 calculation

places the cis isomer lower in energy than the trans by 34

kcal mol-1. The reasons for this discrepancy are unclear;

it may be attributed to either a deficiency in the model

complex, the INDO/1 method or a single determinant

representation for the ground state of the dioxo complexes.

Our extensive work with INDO/1 and Ru-oxo catalyzed

oxidations leads us to believe that the latter problem is

the dominant factor. It is interesting to note that

including electron correlation crudely (i.e. correlating the

4 highest occupied and the 4 lowest unoccupied MOs using

configuration interaction double excitations) reverses the

energy ordering the trans isomer is now lower by 35 kcal

mol-1. The greater correlation contribution for the trans

isomer is a result of its frontier orbitals (the highest

occupied and lowest unoccupied MOs have significant Ru dr -

O pr character) being closer in energy. This leads to lower

energy excited configurations and more extensive mixing.

The fact that both cis- and trans-Ru(VI)02 and Os(VI)O2

complexes have been described implies that the two d2, dioxo

geometries are close in energy and that factors such as

steric repulsions may play an important role in determining

the ground state geometry.




















CIS-IRU(V1(h-,d)(OflZ


Figure 5-1: Structure of the cis- and trans-Ru(VI) (NH=CH-
CH=NH)2(0)22+ model complexes used in the present study.







73

Cis-RuO22+/Ethvlene Reaction Pathways

The delocalized nature of the i bonding in the RuO2

moiety, and the analogy with ozone, readily suggests a

variety of pathways and intermediates. Three concerted

pathways (and their non-concerted variants) [1+2], [2+2]

and [3+2] were considered; they lead to the oxo/epoxide

(OE), oxo/metallaoxetane (OM) and dioxometallacycle (DM)

intermediates, respectively.

Geometry optimization yields weakly bound minima on the

potential energy surface (PES) similar to those obtained for

the Ru(IV)-oxo complexes. The average distance from the

center of the C-C bond of ethylene to the center of the line

which connects the two oxos is = 2A 2.5k. The orientation

of the ethylene relative to the metal complex does not

effect the strength of the binding (z 10 kcal mol-1) as

would be expected from a non-directional electrostatic

interaction. The amount of electron density transferred

(<0.1 electrons) is small, coming chiefly from the olefin n

MO. Perturbation of either reactant from their isolated

optimized geometries is insignificant. These systems differ

from outer-sphere, caged-radicals in that much less than one

electron is transferred. This small amount of electron

density transfer keeps the double bond intact and does not

lead to a loss of stereochemistry at this point in the

reaction.










r1+21 Pathway

Bringing the ethylene along the concerted [1+2] pathway

eventually leads to a large repulsion between the ethylene a

MO and the oxo lone pair. INDO/1 calculation show that even

as far out as R = 2.0~ 2.5k moving the ethylene off the

concerted [1+2] pathway leads to a lowering of the total

energy by decreasing the two-orbital, four-electron

repulsion between the oxo lone pair and C2H4 7 MO. This

stabilization gets larger (as a result of greater overlap

and hence greater interaction) as R approaches covalent

values. Deviation of the ethylene from the concerted [1+2]

pathway causes it to approach the [2+2] and [3+2] pathways.

At large separations there is no preference for either

trajectory. Once this perturbation from the concerted [1+2]

pathway is induced, formation of the oxo/epoxide can occur

in a non-concerted manner. An oxo/epoxide intermediate was

found, by geometry optimization, on the PES. The ground

state is a singlet. The Ru-O bond lengths were 1.84A (oxo)

and 2.18A epoxidee) consistent with experimental values of

1.77A 1.86A for Ru(IV)-O complexes98 and 2.03K and 2.07A

for trans-[RuCl4(NO)(OH)]-2 and trans-[Ru(NH3)4(NO(OH)]+2

respectively.96



[2+21 Pathway

As mentioned above, at large separations between the

ethylene and cis-RuO2 translation of the ethylene from a








75

[1+2] trajectory toward either a [2+2] or [3+2] pathway is

equally favorable. However, as R approaches covalent values

the [2+2] pathway becomes increasingly disfavored in

relation to the [3+2] pathway. Simple Woodward-Hoffmann

considerations state that a concerted, planar [2+2]

cycloaddition is unfavorable for a 4 electron system. The

Ru-O v bond, as noted above, is highly covalent. Thus, a

lowering of the activation energy which causes the

"forbiddeness" of the [2+2] pathway is not expected. A

series of geometry optimizations along the [2+2] pathway

were carried out and did not isolate an intermediate (the

oxo/metallaoxetane, OM). An artificially constructed OM is

not an optimized minimum on the [2+2] pathway and undergoes

further geometric relaxation (and energy lowering) to arrive

at a dioxometallacycle intermediate (singlet spin state).



r3+21 Pathway

A concerted [3+2] cycloaddition is an allowed reaction

for a six electron system. The correlation diagram is shown

in Figure 5-2. Notice the correlation of the olefin a MO

with a n symmetry MO (denoted schematically by a p orbital

but predominantly 4dr in character) on the Ru. This leads

to the formal two-electron reduction of the Ru(VI)-dioxo to

yield the d4 -O-Ru(IV)O-CH2CH2- intermediate; the bond

lengths are Ru-O = 1.94A, C-O = 1.41A and C-C = 1.52A.








/M M
A- o 0
S- -
C-C C C

00
C -A



6 0
O0 Ss




M A





t \ I, jM. $M




Figure 5-2: Orbital correlation diagram for the [3+2]
addition (or 1,3 dipolar addition) of ethylene to a cis-
RuO22+ fragment.







77

These values are in good agreement with the Ru-O bond

distances of 2.03A and 2.07A quoted above for the two

ruthenium-hydroxy complexes and the C-0 and C-C bond lengths

in tetrahydrofuran, C-0 = 1.43A0.03A and C-C =

1.54A0.02A.96 The O-Ru-O angle is 810

The dioxometallacycle is roughly 90 kcal mol-l lower in

energy than the oxo/epoxide intermediate. As the OM does

not represent a stationary point on the potential energy

surface it is not possible to compare the energy of the

optimized DM and OE minima with an "artificially"

constructed OM. The energies of the two minima, DM and OE,

can be rationalized. If the Ru-O bonds are considered to

approximately cancel each other out, then the difference in

energy between the DM and the OE corresponds to the

difference in energy between the C-0 bonds formed. In the

DM two unstrained C-0 bonds are formed (BE z 85 kcal mol-1
111); in the OE two strained, epoxidic C-0 bonds are formed

(BE = 50 kcal mol-1 135). Thus, a rough bond energy

estimate is 70 kcal mol-1, in qualitative agreement with the

INDO/1 result.



The Interaction of Ethylene with trans-fRu(VI)(0)2(N-N)22+

The study of the trans RuO2 species is motivated by the

desire to understand the epoxidation system Ru(TPP)(0)2.130a

As with the mono oxo complex two pathways are envisioned for

trans-RuO22+ a [1+2] and [2+2] cycloaddition. It was







78

originally hoped that the effect of the trans ligand on

stabilizing the two intermediates bound epoxide and

oxometallocycle could be studied. However, as with every

other Ru-oxo species a [2+2] "intermediate" is not a stable,

stationary point on the potential energy surface. Once

again, the strength of the Ru-oxo n bond should be regarded

as the chief obstacle standing in the way of the [2+2]

pathway. A bound epoxide intermediate is formed

preferentially for the trans-Ru(0)2 and trans-Ru(O)(THF)

complexes. The only interesting difference between the two

intermediates is the planar (with respect to the Ru-O-C-C

atoms) epoxide coordination mode in the former and the non-

planar coordination mode in the latter. Since the d4

Ru(O)(bound epoxide) has a vacant dr set, the planar

coordination mode of the bound epoxide allows for maximum

interaction with the epoxide a symmetry non-bonding MO. In

the d6 Ru(THF)(bound epoxide) complex both dr orbitals are

occupied and the bent coordination mode of the epoxide

arises from the desire to keep the repulsive Ru dv-Opi

interaction at a minimum.



Geometric Rearrangement of cis-BE to DM

The interaction of cis-RuO2 with ethylene yielded two

distinct covalently bound minima the oxo/epoxide (OE) and

the dioxometallacycle (DM). The question arises as to

whether the formation of the OE is part of competing pathway







79

or a "pit stop" on the way to the more stable

dioxometallacycle? A look at the HOMO is informative,

Figure 5-3. This HOMO may be stabilizedl6 by increasing

the in-phase overlap between the Ru d orbital and the 0a p

orbital as well as the in-phase overlap between the Cb p

orbital and Ob p orbital. To accomplish a closing of the

Ob-Ru-Oa angle (i.e. an IR wagging motion) would be the

simplest motion. Construction of a potential energy curve

for the distortion coordinate which converts the cis-

oxo/epoxide into the dioxometallacycle was carried out. The

O-Ru-O angle was closed from approximately 90 while other

geometric parameters were kept the same. The INDO/1 results

indicate an activation barrier of roughly 6 kcal mol-1. An

analysis of the wavefunctions indicates a process involving

the concerted breaking of one C-0 bond (epoxidic) with the

concomitant formation of a stronger C-O bond. Thus, it

seems plausible to propose the dioxometallacycle, and the

pathway leading to it, as the main route for the oxidation

of olefins by a cis-Ru(VI)02 complex. Groves' report135 of

cis/trans epoxide isomerization by a Ru(II)-porphyrin, i.e.

cist-stilbene oxide is isomerized to trans-stilbene oxide,

is consistent with the notion that the bound epoxide is not

"inactive" once it is formed as well as with the concept of

C-O bond breaking in the bound epoxide being somewhat

facile.







80

Discussion

The stated goals of the present research were: to get a

better understanding of the metal-dioxo/organic substrate

interaction, to explore comparisons between the metal-dioxo

and metal-oxo complexes and investigate the causes of the

difference in reactivity between the cis- and trans-isomers.

The results of this research, as with our previous

study of Ru(IV)-oxo complexes, confirm that the experimental

results are easily explained in simple frontier-orbital or

Woodward-Hoffmann terms. The trans-dioxo complex is quite

similar to the mono-oxo complex despite the change in formal

oxidation state. The most favorable pathway is directly

analogous to that for the mono-oxo case a non-concerted

[1+2] cycloaddition to form a Ru(IV)(0)(bound epoxide)

intermediate. The lowest energy pathway for cis-RuO2/olefin

interaction is quite different a [3+2] cycloaddition or

1,3-dipolar addition. In the cis isomer both oxygens are

accessible to the same olefin and thus different pathways

for olefin oxidation are encouraged as a result.

Finally, how can the difference in reactivity of the

cis- and trans-Ru(VI)(0)2 complexes be accounted for? The

trans complex is more selective for epoxidations and is very

stereospecific.41,42,130a The cis complex, on the other

hand, leads to a large amount of cleavage products.130d

















j ~0


z





Figure 5-3: Schematic showing the composition of the HOMO
of the cis-Ru-oxo/epoxide.







82

The larger selectivity for epoxidation pathways and

high stereo-specificity for the trans isomer is consistent

with the most favorable pathway being the formation of the

trans-Ru(O)(epoxide) by a non-concerted [1+2] addition

similar to that for Ru(IV)-oxo catalyzed epoxidations.43

The small loss of stereochemistry which is observed is

consistent with a process in which the formation of the

second C-O bond is comparable to rotation about the C-C bond

of the erstwhile olefin. Formation of carbonyl products by

a side reaction which entails trans-cis-RuO2 isomerization

is not indicated. The theoretical analysis by Griffith131b

points to the greater thermodynamic stability in the trans

geometry for d2 dioxo complexes. Experimental correlation

for this is provided by the observation of Meyer et al. that

cis-Os(bpy)2(0)22+ is unstable with respect to the trans

isomer in solution.130b

For the cis-isomer the greater amount of carbonyl

product arises from the fact that conversion of the

dioxometallacycle (DM) to Ru(IV)-oxo plus epoxide is a

highly endothermic process. The weak epoxidic C-O bonds

allow no driving force for this pathway. The strengthening

of the C-O single bonds in the DM to the double bonds of the

cleavage carbonyl products is evidently sufficient enough to

overcome the energy that must be spent in breaking the C-C

bond of the DM. The well known cleavage of C=C double bonds

by RuO4 by way of a ruthenate ester, 02Ru-O-CH2-CH2-O,







83

supports this view.137 Any small amount of epoxide that

results in the oxidation of olefins by cis-RuO22+ has

several plausible origins. First, the higher energy cis-

oxo/epoxide may be solvated or disproportionate before it

can be converted into the DM. The calculated activation

barrier of =6 kcal mol-1 might allow the cis-oxo/epoxide a

long enough lifetime for this process to occur. Second,

isomerization of the cis-RuO2 complex to the trans isomer

followed by epoxidation may occur. Finally, there may be

present some partly oxidized Ru(IV)-oxo species present in

solution.













CHAPTER 6
CONCLUSIONS

The theoretical results described for olefin oxidation

by the six-coordinate Ru(IV)-oxo complex and the six-

coordinate Ru(VI)-dioxo complexes show that these reactions

can be understood in a logical manner using molecular

orbital theory. Often it is thought that the "rules" of

chemistry break down when transition metal species are

involved. First, a charge-transfer complex, as has been

proposed for epoxidations by ferryl and chromyl porphyrin

systems,31,101 is found as a minimum on the PES for

RuN502+.C2H4. Second, least motion [1+2] and [2+2] pathways

are found to be unfavorable so that non-concerted processes

are induced. A detailed analysis of the wavefunctions as

the ethylene proceeds along these pathways reveals that the

causes of the unfavorability is reminiscent of the more

familiar organic analogues, i.e. the least-motion, C2v

addition of methylene to ethylene is analogous to the

concerted [1+2] pathway and the least-motion, D2h

dimerization of ethylene is analogous to the concerted [2+2]

pathway. Once again, as with the organic analogues non-

concerted pathways may arise from the unfavorability in the

concerted pathway. Perturbation of the ethylene off-center

causes a dissymmetry in the C-O bonding, i.e. one C-0 bond

84







85

is longer than the other, leading to a non-concerted

process.

In simple chemical terms, this reaction is a

nucleophilic attack of the olefin directed at the oxo

oxygen. For the reasons that we have already discussed in

detail the oxygen atom transfer step proceeds in a non-

concerted manner, i.e. one C-0 bond forms and then the

second. Since this is a nucleophilic attack, a positive

charge builds up on the terminal (non-bound) carbon of the

substrate. Simple charge considerations lead one to

predict, and INDO/1 calculations support, that C6+-6-

interaction will be more favorable, and lead to bound

epoxide, than C6+-Ru6+ interaction. The bound epoxide is a

low spin, d6 pseudo-octahedral complex and therefore, quite

stable. The metal does not interact with the substrate

directly but as an "electron sink" for the two electrons on

being reduced from Ru(IV) to Ru(II).

The INDO/1 results show that for the oxidation of

alcohols by a six-coordinate Ru(IV)-oxo complex, alcohol

coordination is comparable with, if not more favorable than,

a C-H activation mechanism. The linear interaction of a C-H

bond in methanol with the oxo oxygen leads to a loosely

bound ion-molecule interaction with a oxo-H distance of

1.2k. The formation of a covalent oxo-H distance,

corresponding to hydrogen transfer, is costly in terms of

the total energy of the system for various modes. The







86

pentagonal bipyramid is calculated to present the best

combination of favorable electronic and steric interactions.

Further interaction after the seven-coordinate association

complex is formed increases C-H''Oxo interaction. Thus,

the preferred pathway combines both proposals, i.e. pre-

coordination of the alcohol and C-H-'O interaction.

This study of alcohol oxidations by a six-coordinate,

high-valent metal-oxo complex with nitrogen ligands (i.e. a

P-450 model) shows that when the substrate has a good donor

ability direct metal participation in the oxidation pathway

is plausible. Most proposed mechanisms for oxidation by P-

450 models involve the metal only indirectly as a receptor

for two electrons. Our results suggest that metal

participation may be important; in the oxidation of

substrates that possess good donor ability, i.e. alcohols,

amines, and perhaps sulfides, the substrate may ligate to

the metal in the course of the oxidation.

A molecular orbital analysis of the oxidation of

sulfides to sulfoxides was carried out. Direct oxygen atom

transfer was found to be the most favorable. Unlike alcohol

oxidation there is no energy barrier to the formation of a

covalently bound substrate/oxidant species.

The main goals of the Ru(VI)-dioxo/olefin research were

to get a better understanding of the metal-dioxo/organic

substrate interaction and to explore comparisons between the

Ru-dioxo and Ru-oxo complexes and investigate the causes of







87

the difference in reactivity between the cis- and trans-

isomers. The results of this research, as with our previous

study of Ru(IV)-oxo complexes, confirm that the experimental

results are easily explained in simple frontier-orbital or

Woodward-Hoffmann terms. The trans-dioxo complex is quite

similar to the mono-oxo complex despite the change in formal

oxidation state. The lowest energy pathway for cis-

RuO2/olefin interaction is quite different a [3+2]

cycloaddition or 1,3-dipolar addition. In the cis isomer

both oxygens are accessible to the same olefin and thus

different pathways for olefin oxidation are encouraged as a

result.














REFERENCES


1. Sheldon, R. A.; Kochi, J. K. "Metal Catalyzed
Oxidations of Organic Compounds;" Academic: New York,
1981.

2. Groves, J. T. in "Metal Ion Activation of Dioxygen;"
Spiro, T. G. Ed., Wiley: New York, 1980, 125.

3. "Activation and Functionalization of Alkanes;" Hill, C.
L., Ed., Wiley: New York, 1989.

4. (a) Dawson, J. H. Science 1988, 240, 433.
(b) Guengrich, F. P.; Macdonald, T. L. Acc. Chem. Res.
1984, 17, 9.
(c) Loew, G. H.; Collins, J. R.; Luke, B.; Waleh, A.;
Pudzianowski, A. T. Enzyme 1986, 36, 54.
(d) Loew, G. H.; Collins, J. R. J. Biol. Chem. 1988,
263, 3164.

5. Coon, M. J.; White, R. E. in "Metal Ion Activation of
Dioxygen;" Spiro, T. G., Ed. Wiley: New York 1980.

6. Masters, C. Homogeneous Transition Metal Catalysis;"
Chapman and Hall: London, 1981.

7. Landau, R.; Sullivan, G. A.; Brown, D. CHEMTECH 1979,
602.

8. "Cytochrome P-450: Structure, Mechanism, and
Biochemistry;" Ortiz de Montellano, P. R., Ed. Plenum
Press, New York, 1985.

9. (a) Poulos, T. L.; Finzel, B. C.; Gunsalus, I. C.;
Wagner, G. C.; Kraut, J. J. Biol. Chem. 1985, 260,
16122.
(b) Poulos, T. L.; Finzel, B. C.; Howard, A. J.
Biochemistry 1986, 25, 5314.

10. Mimoun, H. in "Comprehensive Coordination Chemistry;"
Pergamon: Oxford, 1984, p. 21.

11. Gubelmann, M. H.; Williams, A. F. "Structure and
Bonding;" vol. 55, Springer-Verlag: Berlin, 1983.








89

12. Lyons, J. E. in "Aspects of Homogeneous Catalysis;"
vol. 3, Ugo, R., Ed. Reidel: Dordrecht, 1977.

13. Holm, R. H. Chem. Rev. 1987, 87, 1401.

14. Mimoun, H.; Seree de Roch I., Sajus, L. Tetrahedron
1970, 26, 37.

15. Sharpless, K. B.; Townsend, J. M.; Williams, D. R. J.
Am. Chem. Soc. 1972, 94, 295.

16. Jorgensen, K. A.; Hoffmann, R. Acta Chem. Scanda.
1986, B40, 411.

17. Hoffmann, R. J. Chem. Phys. 1963, 39, 1397.

18. Hoffmann, R.; Lipscomb, W. N. J. Chem. Phys. 1962, 36,
2179.

19. Wolfsberg, M.; Helmholz, L. J. Chem. Phys. 1952, 20,
837.

20. Purcell K. Organometallics 1985, 4, 509.

21. Campestrini, S.; Conte, V.; DiFuria, F.; Modena, G.;
Bortolini, D. J. Org. Chem. 1988, 53, 5721.

22. Cundari, T. R.; Drago, R. S.; Goldstein, A. S.; Hage,
J. P.; Patton, D. E. Short Term Analytical Services
(STAS) Report 1989.

23. Drago, R. S.; Cundari, T. R. unpublished results

24. Groves, J. T.; Nemo, T. E.; Myers, R. S. J. Am. Chem.
Soc. 1979, 101, 1032.

25. Groves, J. T.; Nemo, T. E. in "Coordination Chemistry
of Metalloenzymes;" Drago, R. S., Ed. Reidel: Dordrecht
1982, p. 329.

26. Groves, J. T.; Nemo, T. E. J. Am. Chem. Soc. 1983, 105,
5786.

27. Groves, J. T.; Subramanian, D. V. J. Am. Chem. Soc.
1984, 106, 2177.

28. Groves, J. T.; Myers, R. S. J. Am. Chem. Soc. 1983,
102, 5791.

29. Groves, J. T.; Kruper, W. J.; Haushalter, R. C. J. Am.
Chem. Soc. 1980, 102, 6735.








90

30. Groves, J. T.; Kruper, W. J.; J. Am. Chem. Soc. 1979,
101, 7613.

31. Ostovic, D.; Bruice, T. C. J. Am. Chem. Soc. 1989, 111,
6511 and references therein.

32. Burrows, C. J.; Yoon, H. J. Am. Chem. Soc. 1988, 110,
4087.

33. Collman, J. P.; Brauman, J. I.; Meunier, B.; Hayashi,
T.; T. Kodadek; Raybuck, S. A. J. Am. Chem. Soc. 1985,
107, 2000.

34. Shannon, P.; Bruice, T. C. J. Am. Chem. Soc. 1981, 103,
4580.

35. Nee, M. W.; Bruice, T. C. J. Am. Chem. Soc. 1982, 104,
6123.

36. Samsel, E. G.; Srinivasan, K.; Kochi, J. K. J. Am.
Chem. Soc. 1985, 107, 7606.

37. Zheng, W.; Loebach, J. L.; Wilson, S. R.; Jacobsen, E.
N. J. Am. Chem. Soc. 1990, 112, 2801.

38. VanAtta, R. B.; Franklin, C. C.; Valentine, J. S.
Inorg. Chem. 1984, 23, 4120.

39. Hill, C. L.; Smegal, J. A.; Henley, T. J. J. Orq. Chem.
1983, 48, 3277.

40. Srinivasan, K.; Michaud, P.; Kochi, J. K. J. Am. Chem.
Soc. 1986, 108, 2309.

41. Che, C. -M.; Wong, K. -Y.; Mak, T. C. J. Chem. Soc..
Chem. Commun. 1985, 988.

42. Perrier, S.; Kochi, J. K. Inorg. Chem. 1988, 27, 4163.

43. Dobson, J. C.; Seok, W. K.; Meyer, T. J. Inorg. Chem.
1986, 25, 1513.

44. Traylor, T. G.; Marsters, J. C.; Nakano, T.; Dunlap, B.
E.; J. Am. Chem. Soc. 1985, 107, 3735.

45. Nappa, M. J.; Tolman, C. A. Inorg. Chem. 1985, 24,
4711.

46. Collman, J. P.; Kodadek, T.; Raybuck, S. A.; Brauman,
J. I.; Papazian, L. M. J. Am. Chem. Soc. 1985, 107,
4343.







91

47. Maskiko, T.; Dolphin, D.; Nakano, T.; Traylor, T. G. J.
Am. Chem. Soc. 1985, 107, 5537

48. Castellino, A. J.; Bruice, T. C. J. Am. Chem. Soc.
1988, 110, 158.

49. Castellino, A. J.; Bruice, T. C. J. Am. Chem. Soc.
1988, 110, 7512.

50. (a) Garrison, J. M.; Bruice, T. C. J. Am. Chem. Soc.
1989, 111, 191.
(b) Garrison, J. M.; Ostovic, D.; Bruice, T. C. J. Am.
Chem. Soc. 1989, 111, 4960.

51. Sharpless, K. B.; Teranishi, A. Y.; Backvall, J. E. J.
Am. Chem. Soc. 1977, 99, 3120.

52. Miyaura, N.; Kochi, J. K. J. Am. Chem. Soc. 1983, 105,
2368.

53. Roecker, L.; Meyer, T. J. J. Am. Chem. Soc. 1987, 109,
746.

54. Roecker, L.; Dobson, J. C.; Vining, W. J.; Meyer, T. J.
Inorq. Chem. 1987, 26, 779.

55. Lindsay-Smith, J. R.; Mortimer, D. N. J. Chem. Soc.,
Chem. Commun. 1985, 64.

56. Hill, C. L.; Schardt, B. C. J. Am. Chem. Soc. 1980,
102, 6374.

57. Lindsay-Smith, J. R.; Sleath, P. R. J. Chem. Soc..
Perkin Trans. II 1983, 621.

58. Tabushi, I.; Koga, N. Tetrahedron Lett. 1979, 3681.

59. Guilmet, I.; Meunier, B. Tetrahedron Lett. 1980, 4449.

60. DePoorter, B.; Ricci, M.; Bortoline, O.; Meunier, B. J.
Mol. Catal. 1985, 31, 221.

61. DePoorter, B.; Meunier, B. J. Chem. Soc.. Perkin II
1985, 1735.

62. Roecker, L.; Meyer, T. J. J. Am. Chem. Soc. 1986, 108,
4066.

63. Jorgensen, K. A. J. Am. Chem. Soc. 1987, 109, 698.

64. Strich, A.; Veillard, A. Theor. Chim. Acta 1981, 60,
379.








92

65. Goddard, W. A.; Carter, E. A. J. Phys. Chem. 1988,
92, 2874.

66. Goddard, W. A.; Dunning, T. H.; Hunt, W. J.; Hay, P. J.
Accs.Chem. Res. 1973, 6, 368.

67. Goddard, W. A.; Ladner, R. C. J. Am. Chem. Soc. 1971,
93, 6750.

68. Goddard, W. A.; Blint, R. J. Chem. Phys. Letters 1972,
14, 616.

69. Goddard, W. A.; Carter, E. A. J. Phys. Chem. 1988,
92, 5679.

70. Rapper, A. K. ; Goddard, W. A. J. Am. Chem. Soc. 1982,
104, 448.

71. Rapper, A. K.; Goddard, W. A. J. Am. Chem. Soc. 1982,
104, 3286.

72. Walch, S. P.; Goddard, W. A. J. Am. Chem. Soc. 1978,
100, 1338.

73. Bauschlicher, C. W.; Nelin, C. J.; Bagus, P. S. J.
Chem. Phys. 1985, 82, 3265.

74. Krauss, M.; Stevens, W. J. J. Chem. Phys. 1985, 82,
5584.

75. Bagus, P. S.; Preston, H. J. T. J. Chem. Phys. 1973,
59, 2986.

76. Loew, G. H.; Kert, C. J.; Hjelmeland L. M.; Kirchner,
R. F. J. Am. Chem. Soc., 1979, 99, 3534.

77. Loew, G. H.; Herman, Z. S. J. Am. Chem. Soc. 1980, 102,
6173.

78. Hanson, L. K.;Chang, C. K.; Davis, M. S.; Fajer J. J.
Am. Chem. Soc. 1981, 103, 663.

79. Pople, J. A. ; Beveridge, D. L. "Approximate Molecular
Orbital Theory;" McGraw-Hill: New York, 1970.

80. Bacon, A. D. ; Zerner, M. C. Theor. Chim. Acta 1979,
53, 21.