xx ‘ .. ..(
‘13:... .
11...}...
\. ;..«

L. 3...;

(‘16....

 

 

 

‘ \lilllllll\ll\lll\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

 

 

 

 

This is to certify that the

thesis entitled

lVblecular Magnets from Stable Organic Free Radicals:
An Ion—Binding Approach

presented by

Sei—Hum Jang

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Chemistry

QM filo

Major prof or

 

Date (O//717/q}

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

 

 

 

_ LIBRARY
Michigan State '
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

"33.04.1225,

L , I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity lnditution
, , 7 , , Warn-9.1

 

 

 

Molecular Magnets from Stable Organic Free Radicals:
An Ion-Binding Approach

by

Sei — Hum Jang

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1993

 

 

ABSTRACT

Building Organic Magnets from Stable Organic Free Radicals:
An Ion-Binding Approach

by

Sei - Hum Jang

Structural studies of the propeller-shaped radical 1 and related
triaryl-X species show pairs of binding sites comprising tripods of ether
oxygens. The preorganized nucleophilic pockets have enough flexibility to
accept a range of metal ion sizes. By coordination with small alkali metal
cations the radicals may stack to make extended chains of interacting
electron spins as shown below. In fact, mixtures of tris(2,6-
dimethoxyphenyl)methyl 1 and LiBF4 form solids with weak
ferromagnetic behavior and solids from 1 and CdC12 show clear

antiferromagnetic behavior.

nX'

nX'

 

 

 

nX'

M = u 2 (RMR dimer) 3 (RW oligomer)
x = BF4‘, BPh4’, r, 0104‘, SCN'

NMR, pulsed EPR, UV—Vis, and electrochemical studies of radicals
and diamagnetic analogs show behavior consistent with ion binding to form
complexes. In addition, by varying substituents on the aryl rings of 1, it is

 

 

 

 

 

 

 

possible to adjust redox and binding properties and thus the radical's
electronic and structural nature. Such adjustments are significant tools for
fine tuning radicals for effective ion binding and electron coupling.
Extended oligomeric arrays are difficult to characterize, so model
compounds that form well-defined molecular complexes were devised to
separate and characterize the metal-radical, metal-radical-metal, and
radical—metal-radical relationships. Because magnetism is a bulk property
that arises from a three-dimensional structure, potential chain-coupling di-
and tri-radical building blocks were devised by utilizing m-phenylene as a
high spin coupling unit.

Magnetic measurements with a Superconducting Quantum Interference
Device (SQUID) magnetometer show that, unlike l-LiX, some of
monomeric complexes show simple paramagnetism, with independent
electron spins. Control over magnetic properties can be achieved by
varying the metal cation, thus complex 1-LiX shows weak ferromagnetism
while complex 1-CdC12 shows antiferromagnetic coupling.

Electron Spin Echo Envelope Modulation (ESEEM) studies of complexes
in the radical ionophoric system 28 in frozen matrices yield the number of
metal ions bound and their distances to the radical center. Molecular
mechanics and semiempirical molecular orbital (MO) calculations coincide
remarkably well with the spectrosc0pic findings. Complexation studies in
solution support these ideas while Ab Initio calculations on a linear
H3C-Li+-CH3 model show that high-spin (ferromagnetic) electron
coupling is favored for a wide range of radical-Li+ distances. Model Ab
Initio MCSCF computations on the same system offer an explanation for
high—spin complexation-induced magnetic coupling of radicals.

Attempts to assemble and characterize a series of triarylmethyls and their
metal ion complexes were made. Triaryl-X frameworks are new to the ion
binding field, so the complexation abilities of 1 and its congeners were
studied by UV-Vis, electrochemistry, NMR and EPR methods. ESR,
SQUID, and X-ray studies on these systems were used to asses inter-
electron communication. Besides addressing structural and theoretical
issues of magnetic coupling, the requirements for the design of molecular
solids with long-range structure induced by ion complexation were probed.

 

 

 

 

To my Parents

 

 

 

ACKNOWLEDGMENTS

My sincere thanks are given to Professor James E. Ned Jackson
whom I believe to be one of the great chemists. It was my honor to be his
first student and is going to remain as a happy memory in my heart.
Numerous discussions with him on my research and others made me realize
the meaning of learning how to learn.

I would like to thank Professor Bart E. Kahr, Professor John McCracken,
and Hong-In Lee for their help in my research and Professor James L.
Dye, Professor William Reusch, and Professor Daniel G. Nocera for
serving on my committee.

I am grateful to the National Science Foundation, the United States Air
Force, and the Michigan State University Center for Fundamental Materials
Research for financial support in the form of teaching and research
assistantships. I would like to thank to the Department of Chemistry for its
excellent research environment. I also owe special thanks to many friends
in the Chemistry and Physics departments for their friendships.

Last, I would like to thank my wife and family for love and patience
throughout my study.

 

 

 

TABLE OF CONTENTS

Chapter page
LIST OF TABLES ......................................................................... viii
LIST OF FIGURES .......................................................................... x
INTRODUCTION ......................................................................... l
1. Macromolecular Chemistry and Molecular Engineering ................... 4
2. Stable Organic Radicals ................................................................ 6
2.1. Tri- and Diphenylrnethyl Radicals
Historical Background .................................................... 7
2.1.1. Gomberg's Triphenylrnethyl Radical ............................... 7
2.1.2. Schlenk, Chichibabin, and Thiele's Hydrocarbon ............... 9
2.1.3. Ballester's Stable Triaryhnethyl Radicals ......................... 11
2.1.4. Jackson and Kahr's Lithium Complex of Triarylrnethyl ..... 13
2.2. Nitrogen Centered Radicals ............................................ 15
2.2.1. Hydrazyl Radicals .......................................................... 15
2.2.2. Nitroxyl Radicals ........................................................... 17
2.3. Aroxyl Radicals ............................................................. 21
3. Magnetism ................................................................................... 23
3.1. Diamagnetism ............................................................... 23
3.2. Paramagnetism .............................................................. 24
3.3. Curie-Weiss Law ........................................................... 28
3.4. Pairwise Magnetic Exchange ........................................... 30
3.5. One-Dimensional or Linear Chain Systems ....................... 31
3.6. Alternating Linear Heisenberg Chain ............................... 35
4. Molecular Magnets ....................................................................... 37
4.1. Molecular Magnet models ............................................... 37
4.1.1. McConnell Model .......................................................... 37
4.1.2. Mataga Model ............................................................... 39
4.1.3. Ovchinnikov Model ....................................................... 41
4.1.4. Breslow Model .............................................................. 42
4.1.5. Torrance Model ............................................................ 44
4.1.6. Kahn's Proposal ............................................................ 44

V

 

 

 

4.2. Metal Based Molecular Magnets ...................................... 46

4.2.1. Single Atom Bridged Bimetals ........................................ 46
4.2.2. Molecule Bridged Bimetals ............................................. 48
4.2.3. Metal and Organic Radical Based Systems ........................ 49
4.2.4. Charge Transfer Complex Based Systems ......................... 54
4.3. Galvinoxyl and N itroxyl Radical Based systems ................ 55
4.3.1. Galvinoxyl Radicals ....................................................... 55
4.3.2. Nitroxyl Radicals ........................................................... 58
4.4. Altemant Hydrocarbon Based Polyradicals ....................... 63
4.4.1. Triplet Diradicals .......................................................... 66
4.4.2. Perchlorinated Triphenylmethyl Radicals ......................... 68
4.4.3. Polyphenyl Carbenes ...................................................... 72
5. Molecular Magnets by Macromolecular Chemistry .......................... 75
5.1. Long-range Magnetic Interactions in Organic Radicals ...... 77
5.2. Through Space Magnetic Interactions in Organic Radical

Complexes .................................................................... 81
5.3. Design of Molecular Magnets by Ion-Binding ................... 83
RESULTS AND DISCUSSION .......................................................... 90
CHAPTER I. Triaryhnethyl Radicals ................................................. 90
1. Structure of Triaryhnethyl Radical ................................................ 90
1.1. Structure of T riaryhnethyl Radical .................................. 92
1.2. Structure of Triaryhnethyl Cation ................................... 95
1.3. Structure of Triaryhnethyl Borane .................................. 96

1.4. Hexachloro Triphenylmethyl Radical and Borane .............. 100

2. Substituent Effects ........................................................................ 102

2.1. Hexamethoxy T riphenylmethyl Radicals ........................... 104

2.2. The Cyclized Xanthenol ................................................. 120

2.3. Tetramethoxy Triphenylmethyl Radicals .......................... 121

2.4. Rotational Barriers ........................................................ 131

3. Ion-Binding .............................................................................. 137

3.1. Borane ......................................................................... 137

3,2. Double faced Paramagnetic Ionophore ............................. 141

3.3. Tetramethoxy Triphenyl amines ...................................... 153

3.4. Ammonium Complexes by Hydrogen bonding. ................. 160

3.5. Metal Ions ..................................................................... 165

vi

 

a o n

c i

 

CCSILZAJ.

E123

 

 

 

CHAPTER II. Magnetism of Triaryhnethyl Radical Complexes ............ 167

1. Calculations on Pairwise Interaction of Methyl Radicals ................... 167
2. Antiferromagnetic Interaction and Dimer of The Radical ................. 169
3. Magnetic Behavior of Radicals and Complexes ................................ 171

3.1. Diamagnetism of Methane and Metal Salts ........................ 173

3.2. Paramagnetism of Radicals and Radical Complexes ........... 175

3.3. Ferromagnetism of Complexes ........................................ 184

3.4. Antiferromagnetism of Radical Complexes ....................... 187
4. Design of Molecular Magnets by Other Ion-Binding ........................ 192
5. Introduction of Cross-Linkers ....................................................... 195
CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK ............ 199
Conclusions ...................................................................................... 199
Suggestions for future work .............................................................. 201
1. Aromatic Linked Biradicals .......................................................... 201
2. Magnetic Coupling through Hydrogen Bonds .................................. 202
3. Ion-Binding in Various N itroxyl Radicals ....................................... 204
EXPERIMENTAL ......................................................................... 208
1. General Procedures ...................................................................... 208
2. Solvents and Chemicals ................................................................. 209
3. Equipment and Procedures ............................................................ 209

CV Measurements .................................................................... 209

VT—EPR Measurements ............................................................ 210

SQUID Measurements .............................................................. 210
4. Synthesis ..................................................................................... 212
REFERENCES ............................................................................... 242

vii

 

 

Ta

 

 

 

LIST OF TABLES

lble p age

Aryl ring twists in tripod binding sites ....................................... 99

Crystal structure determination and refinement data for tris
(2,6-dimethoxyphenyl)methyl 1, tris(2,6-dirnethoxyphenyl)methyl
tetrafluoroborate 62, tris(2,6-dimethoxyphenyl)borane 63 ......... 99

Summary of electrochemical data for derivatives of
tris(2,6-dimethoxyphenyl)methyl 1 ............................................ 105

Correlation of E1 /2 and o-values for derivatives of
tris(2,6-dimethoxyphenyl)methyl 1 ............................................ 108

Summary of hyperfine constants for derivatives of
tris(2,6-dimethoxyphenyl)methyl 1 ............................................ 116

Summary of electrochemical data for derivatives of
phenyl-bis(2,6-dimethoxyphenyl)methyl 77 ................................ 123

Correlation of E1 /2 and o—values for derivatives of
* phenyl—bis(2,6-dimethoxyphenyl)methyl 77 ................................ 125

Summary of hyperfine constants for derivatives of
phenyl-bis(2,6-dimethoxypheny1)methyl 77 ................................ 126

Summary of VT-NMR results for tris(2,6-dimethoxy-
phenyl)methyl ammonium tetrafluoroborate 72, tris(2,6-
dimethoxy-3,5-dichlorophenyl)methyl ammonium
tetrafluoroborate 75, tris(3,5-dichloro-2,6-dimethoxy

phenyl)methane 76 .................................................................. 133
Complexation data from ESEEM for tris(2,6-di(2-methoxy
ethoxy)phenyl)methyl 86*2MX ................................................. 1 50

viii

 

 

 

Summary of electrochemical data for tris(2-(methoxy
ethoxy)-6-methoxyphenyl)methyl 85 and tris(2,6—di-
(methoxyethoxy)phenyl)methyl 86 ............................................ 151

. Summary of electrochemical data for bis(2,6-dimethoxyphenyl)
-4-chlorophenyl amine 87, bis(2,6-dimethoxyphenyl)
-3,5-dimethoxypheny1 amine 88, tris(3,5-dimethoxyphenyl)

amine 89 ................................................................................ 155

. Crystal structure determination and refinement data for tris(2,6-
dimethoxyphenyl)methyl ammonium tetrafluoroborate 72

and tris(3,5-dichloro-2,6-dimethoxyphenyl)methyl

ammonium tetrafluoroborate 75 ................................................ 162

ix

 

 

ll

 

;ure

 

LIST OF FIGURES

page
Temperature dependence of the paramagnetic susceptibility ........ 25
The molar susceptibility for ferro- and antiferromagnets ............. 29
Bleaney Bowers susceptibilities compared with the Curie law ....... 31
Ising susceptibilities compared with Curie law ............................ 34
Ising "perpendicular" susceptibilities .......................................... 34
Heisenberg linear chain susceptibilities ...................................... 36
Illustration of the oligomeric chain of
tris(2,6—dimethoxyphenyl)methyl 1 with M+ ions ....................... 87
MM2 optimized structure of an RMR dimer 2 calculated ............ 87
Stereo view of the X-ray structure of the radical
tris(2,6-dimethoxyphenyl)methyl 1 ............................................ 94
Space-filling views of the binding site for tris(2,6-
dimethoxyphenyl)methyl 1 and tris(2,6-dimethoxyphenyl)
methyl tetrafluoroborate 62 ...................................................... 96
Ball & Stick representations of the binding site for
tris(2,6-dimethoxyphenyl)methyl l and tris(2,6-dimethoxy
phenyl)borane 63 .................................................................... 97
Stereo view of the packing diagram of tris(2,6-dimethoxy
phenyl)borane 63 .................................................................... 98
MM2 calculated structure of tris(3,5-dichloro-2,6-dimethoxy-
phenyl)borane 64 dimer with Li+ ............................................. 101

Cyclic voltammogram of tris(3,5~dichloro-2,6-dimethoxyphenyl)
borane 64 in methylene chloride ............................................... 101

X

 

 

 

20. '

22.

23.

24.

25.

26.

27.

28.

 

5. Cyclic Voltammogram of tris(2,6-dimethoxyphenyl)

- \

 

 

 

methyl 1 in THF ...................................................................... 103
. AE1/2(V) vs 6 in THF for HMTP methyl radicals ....................... 108
'. EPR spectrum of tris(2,6-dimethoxyphenyl)methyl 1

in THF with simulations ........................................................... 117
. EPR spectrum of tris(3,5-dichloro-2,6-dimethoxyphenyl)

methyl 65 in THF with simulations .118
. EPR spectrum of tris(4—chloro-2,6—dimethoxyphenyl)

methyl 66 in THF with simulations 119
. The X-ray structure of cyclized tris(2,6-dimethoxy

phenyl)methyl carbinol 120

AE1/2(V) vs 6 in THF for TMTP methyl radicals ....................... 124

EPR spectrum of phenyl-bis(2,6-dimethoxyphenyl)
methyl 77 in THF with simulations ...... 127

 

EPR spectrum of phenyl-bis(2,6-dimethoxyphenyl)
methyl 77 in toluene with simulations ........................................ 128

EPR spectrum of bis(2,6—dimethoxyphenyl)-4—chlorophenyl
methyl 78 in THF with simulations ..... 130

 

 

EPR spectrum of bis(2,6-dimethoxyphenyl)-3,5-dimethoxy
phenyl methyl 84 in THF with simulations ................................. 130

Stereo view of the X-ray structure of tris(Z,6-dimethoxyphenyl)
methane 73 ............................................................................. 131

VT 300 MHz 1H NMR spectrum of tris(2,6—dimethoxy-
3,5—dichlorophenyl)methyl ammonium tetrafluoroborate 75 ........ 134

VT 300 MHz 1H NMR spectrum of tris(2,6-dimethoxy-
3,5—dichlorophenyl)methane 76 135

 

xi

 

31

3f

36

37

38

39

40.

41.

 

UV-Vis spectra showing ion-binding in
tris(2,6-dimethoxyphenyl)borane 63 139

 

Mass Spectrum of a dimer 2 of tris(2,6-dimethoxy-
pheny)methyl 1-Na+ol ............................................................ 140

EPR spectrum of tris(2,6-di(methoxyethoxy)phenyl)
methyl 86 with simulations ....................................................... 143

ESEEM spectrum of tris(2,6-di(methoxyethoxy)phenyl)
methyl 86-2LiBF4 .................................................................. 145

ESEEM spectrum of tris(2,6—di(methoxyethoxy)phenyl)
methyl 86°2NaBPh4 ............................................................... 147

Summary of ESEEM and MNDO results of tris(2,6—
di(methoxyethoxy)phenyl)methyl 86-2LiBF4 ............................ 149

Stereo view of the MNDO calculated structure of
tris(2,6-di(methoxyethoxy)phenyl) methyl 86-2LiBF4 ............... 149

1H NMR spectra for ion-binding of bis(2,6—dimethoxyphenyl)—
4-chlorophenyl amine 87 with Lil ............................................. 156

13C NMR spectra for ion—binding of bis(2,6—dimethoxyphenyl)—
4-chlorophenyl amine 87 with Lil.... ....157

 

1H NMR spectra for ion-binding of bis(2,6-dimethoxyphenyl)-
3,5-dimethoxyphenyl amine 88 with Lil .................................... 158

13C NMR spectra for ion-binding of bis(2,6-dimethoxyphenyl)-
,5-dimethoxyphenyl amine 88 with Lil .................................... 159

tereo views of the X—ray crystal structure of tris(2,6-
imethoxyphenyl)methyl ammonium BF4 72 and tris(3,5-

ichloro2,6—dimethoxyphenyl)methyl ammonium BF4 75 ............ 161

H NMR spectrum of complex of tris(2,6~dimethoxyphenyl)

orane 63-NH4I in CD3CN ...................................................... 163
xii

 

 

42.

43. 1

44. 1

 

 

1H NMR spectrum of complex of tris(2,6-dimethoxy—
3,5—dichlorophenyl)borane 64-NH4I in CD3CN ......................... 164

Qualitative MO diagram for orbital interactions in MeLi+Me ....... 167
Curie—Weiss behavior of tris(2,6-dimethoxyphenyl)methyl 1 ...... 169
The X-Ray structure of head-to-tail peroxydimer 91 .................. 170

Plot of magnetization M vs H for tris(2,6-dimethoxyphenyl)
methyl 1 and 1-LiBF4 ...172

 

Plot of magnetization M vs H for tris(2,6-dimethoxyphenyl)
methane 73 and CdC12 ............................................................ 174

Plot of l/x vs T and xT vs T for tris(2,6—dimethoxyphenyl)
methyl 1 ................................................................................ 176

Plot of M vs. H and l/X vs T for tris(2,6-di(methoxyethoxy)
phenyl) methyl 86-2LiBF4 . ................................................... 179

Plot of ueff vs T for tris(2,6-dimethoxyphenyl)methyl 1-2LiI ....180

Plot of 1/x vs T for tris(2,6-dimethoxyphenyl)methyl 1-NH4I...181

 

 

Plot of Brillouin functions calculated for S=l/2 to 7/2 ................. 182
The saturation magnetization of the complex of
tris(2,6-dimethoxyphenyl)methyl l-LiI at 1.8K .......................... 184
lot of magnetization M vs H for tris(2,6-dimethoxyphenyl)methyl
~LiBF4 and 1-ZnC12 .................. 186
lot of x vs T for tris(2,6-dirnethoxyphenyl) l-CdC12
ith Bleaney-Bowers fit ........................................................... 189
ltemating dimer and linear chain Heisenberg fit for
is(2,6—dimethoxyphenyl) l-CdClz .......................................... 190
-ray powder patterns of 1 and l-CdClz .............................. 191
xiii

_____A

 

0V1

an:

int:

sta‘

the

are

INTRODUCTION

Organic materials science is in a stage of infancy compared to its
inorganic sibling. The difficulties are associated mainly with the lack of
control of structure in the organic solid state. Since the Nobel prize in
19871, interest in macromolecules and their self-assembly has increased
;lramatically. Attention has focused not only on the fundamental molecular
'ecognition processes of the macromolecules but also on the design of
nolecular devices using various intermolecular interactions such as
:lectrostatic, hydrogen bonding, van der Waals forces, etc. Organic
materials chemistry requires the resources of traditional molecular
hemistry combined with an understanding of non-covalent interactions so
3 to form macromolecular entities that possess structures as well defined as

rose of molecules themselves.

ne of the greatest challenges to materials science is the design of
olecular magnetic materials. Simple paramagnetic organic compounds—
:e radicals,2 triplet carbenes,3 biradicals4,5 —have been well studied
er the last 30 years, and their magnetic characteristics can be adequately
ilyzed in terms of current theory.6 Molecular magnets represent a
damentally new class of materials in which a defined magnetic exchange
:raction has been imposed on adjacent paramagnetic entities in the solid
6.7 In spite of significant growth in the field of molecular magnetism,
mechanisms of colligative magnetic properties such as ferromagnetism

incompletely understood.8

 

 

 

Interest in organic and molecular magnetism has grown so much in
the last ten years that special ACS symposia on the subject were held in
1989 and 1992,9 yet to date there are only a few molecular magnets that

clearly exhibit long—range ferromagnetic interactions.10

  
 

AF Coupler
F Coupler

 
 

. + i:> open shell singlet state

+

‘ —-4— :9 open shell triplet state

11 an isolated atom or molecule, a pair of electron spins may be coupled to
'ield closed-shell singlet, open-shell singlet, or triplet states. In other
lOI‘dS, we can relate the magnetic interaction between two organic radicals
) their chemical bonding interactions. Strong interactions between two
{dicals will result in electrons pairing up in a closed shell singlet state (0—
nd), while very weak interactions between two radicals will result in a
ir of open shell doublet states (isolated paramagnets). We can view an

tiferromagnetic coupling as a weak bonding interaction between radicals.

extended odd-electron systems, the three modes above correspond to
agnetic (closed-shell singlet), antiferromagnetic (open-shell singlet),
ferromagnetic coupling (triplet), respectively.“ Several mechanisms
6 been discussed by which high-spin states of molecular materials may
stabilized.12,13 However, there is still no predictive theory that can
urately forecast the magnitude or even the sign of magnetic coupling in

:eneric chemical system of known structure. Magneto-structural

 

 

 

 

correlations are best described among structural homologs, and even then,
experts do not always agree on the mechanisms of coupling. This thesis
describes an approach to this problem that uses structurally well defined

systems made of stable organic radicals.

The principles developed in selective ion binding studies of polyether
ligands14 suggest that the methoxy groups in tris(2,6-dimethoxy—
phenyl)methyl radical can serve as binding sites for metal cations. Two
radicals may "sandwich" a cation of appropriate size between them in a
distorted octahedral pocket,15,16 fixing their relative orientation. The
spatial relationships in such controlled radical aggregation can, in turn,

dictate intermolecular electron coupling in "interrupted O'- bonds"———radical

 

pairs or oligomers in which electron interactions are mediated by metal
cations. Extended stacks formed by this complexation mechanism would
ave at their cores a linear array of one-electron carbon-centered p-

rbitals interacting through metal ions.

In writing the first dissertation on this project in the Jackson group, I

all have a rather extended introductory chapter to give a detailed picture

f current molecular magnetism research, a brief overview of the theory of
agnetism, and a historical review of stable organic radical chemistry. I
all give the rationale behind our research, followed by results and
scussion. Results and discussion will be divided into three parts: building
uctural bases for the formation of complexes; magnetic studies on
icals and complexes; and attempts at the introduction of linkers for

tended magnetic structures.

 

 

 

 

to
en
co

en

de
sel
im

an.

mo
p05
of
ent
tha'
latt
C00

con

 

1. Macromolecular Chemistry and Molecular Engineering

In contrast to molecular chemistry, which is predominantly based on
the covalent bonding of atoms, supramolecular chemistry is based on
intermolecular interactions. Supramolecular interactions are the foundation
for highly specific biological processes, such as substrate binding by
enzymes or receptors, formation of protein complexes, intercalation
complexes of nucleic acids, and immunology. An exact knowledge of the
energetic and stereochemical characteristics of these non-covalent, multiple
intermolecular interactions within defined structural motifs should allow the
design of artificial enzymes or receptors, which bind substrates strongly and
selectively.17 There are many examples of such receptor or enzyme models
including cyclodextrins,18 and cyclophanes based on bipheny119, terphenyl,
nd triphenylmethyl20 frameworks, and cryptophanes.21

Molecular materials are characterized by being made up of discrete
olecules. The structural properties of molecular materials offer many
ossibilities to modulate the bulk electrical, magnetic, and optical properties

the material by choosing appropriate molecules as the macromolecular
tities. At the same time, it is a challenge to develop synthetic strategies
at allow the control of the spatial distribution of the molecules in the
ttice. The bulk properties of materials are always determined by
operative interactions between the constituting molecules, which
-nsequently must be assembled in the lattice in such a way as to maximize

3 bulk response.22

 

 

 

re
311

00

 

Beyond the basic problem of establishing correlations between structure
and a given property, molecular materials appear promising for
development of new characteristics through the combination and linking of
their properties. The recognition of inter- and intramolecular interaction in

macromolecules is the basis for the design of molecular materials.7

An organic crystal is perhaps the most precise example of molecular
recognition, since molecular chemistry blends with intermolecular forces to
propagate a periodic molecular array.23 The almost perfect alignment of
molecules in an organic crystal usually results in highly regular physical
and chemical properties of the molecules, which in turn justify efforts in
:rystal engineering. Of course, phenomena such as hydrogen bonding and
nolecular complexation have been known for decades. However, ideas

:onceming molecular recognition have only been refined in recent years.

The research on organic materials with nonlinear optical properties
d the studies on organic conductors and superconductors are some of the
ost active areas in the field. Much of the motivation for the increased
search activity in conducting materials has been based on the potential for
chnological applications. Conducting, light-weight, soluble, and
ermoplastic materials are of great industrial and economic interest.
ning the past ten years, there has been a very significant international
search effort as a result of the search for higher Tc organic
)erconductors, much of it stimulated by the discovery of the high Tc

)per oxide ceramic superconductors.24a25

 

 

 

 

 

 

 

Til
CC
In:

3V

 

2. Stable Organic Radicals

A free radical is an atom, molecule or complex which contains one or
more unpaired electrons. Free radicals can be obtained in a variety of
ways; those which can be isolated and exist as stable, pure substances may be
prepared by conventional chemical methods but the more reactive radicals
are produced by thermal or photochemical bond cleavage, decomposition,
irradiation, mechanical degradation, or by electron transfer reactions. The
majority of free radicals obtained by these methods are highly reactive and
have lifetimes measurable in terms of micro— or milliseconds unless they are

stabilized by trapping in some inert matrix or on the surface of a solid.

The concept of stability in organic radicals depends on the chemical
:ircumstances; for example, the methyl radical can be stabilized indefinitely
an argon matrix at very low temperatures although it disappears
eversibly on warming the matrix. Diphenylpicrylhydrazyl (DPPH) can
ist permanently as a free radical in the solid state at room temperature; it
n be crystallized, isolated and treated as a normal organic compound but
'hen exposed to other organic radicals it usually reacts rapidly to give
amagnetic products. The word "stable" should only be used to describe a
dical so persistent and so unreactive to air, moisture, etc., under ambient
nditions that the pure radical can be handled and stored in the lab with no
we precautions than would be used for the majority of commercially

lilable organic chemicals.26,27

 

 

2.1. Tri- and Diphenylrnethyl Radicals: Historical Background

2.1.1. Gomberg's Triphenylmethyl Radical

In 1900 Gomberg discovered triphenylmethyl, the first organic free
radical. The historical significance of the discovery was reflected in "The
History of Organic Chemistry in The United States, 1875—1955" by D. S. &
A. T. Tarbell.28

" Without question, the single most important piece of work from the
United States in the beginning of the century and the one which aroused the

rnost interest abroad was the discovery by Moses Gomberg of structures

 

which contained trivalent carbon. Gomberg was an example of the best
rind of American success story. " " Coming to this country from Russia as
child with his penniless father, he eventually worked his way through the
niversity of Michigan at Ann Arbor and received a doctorate in chemistry
ere in 1894." "During a year of study with Victor Meyer at Heidelberg in
97, he synthesized tetra-phenyl methane, and after his return to Michigan,
3 studied the preparation of the hexaphenylethane by the action of various
etal salts and metals on triphenylmethyl chloride. He obtained, instead of
e expected hexaphenylethane, a compound whose composition and

emical properties corresponded to the peroxide."
mberg suggested in his first paper that the highly reactive substance from

reaction of silver metal with triphenylmethyl chloride contained

alent carbon. Gomberg's interpretation of the triphenyl methyl was

g

 

against the classical structural theory of tetravalent carbon. From the same

reaction under nitrogen, he isolated the first product and showed that it had
the composition corresponding to hexaphenylethane by elemental analysis.
However, the structure of the first product was not proven by Gomberg to
be hexaphenylethane. The elegance and serendipity of Gomberg's
experiments in 1900 have made the discovery familiar to many organic
chemists. The significance of the claim of an isolable trivalent carbon
molecule was immediately recognized, and leading chemists of the day

entered the debate on the nature of Gomberg's hydrocarbon 4.

QiHiQ

Gomberg Dimer

he hypothesis of an equilibrium between triphenylmethyl radicals and
hexaphenylethane had been accepted as certainly as that of tetrahedral
:arbon until the discovery by Lankarnp, Nauta, and MacLean in 1968 that
hexaphenyl ethane" in fact had the unsymmetrical quinoid structure 5

vhich Jacobson had proposed for it in 1904. 29

o o
—e
00

So far, Gomberg's hexaphenylethane has not been synthesized except

 

the derivative hexakis(3,5-di—t—butyl-4-biphenyl) ethane by Winter in
78 and hexakis(3,5-di-t-butylphenyl) ethane by Mislow and Kahr in

 

 

1986.30 The realization that trivalent carbon could exist as a relatively

stable structure made chemists more ready to accept free radicals as
intermediates in reactions in solutions even when the individual radicals

could not then be identified.
2.1.2. Schlenk, Chichibabin, and Thiele's Hydrocarbon

On dehalogenating 1,3—bis(diphenylchloromethyl)benzene Schlenk
and Brauns obtained a colorless product whose elemental analysis accorded
with the molecular formula of Schlenk biradical 6. Investigation showed

the biradical to be associated into a polymer.31

Ph Ph

Ph Ph

 

Schlenk Biradical 6

Sfimmermann produced the monomeric Schlenk biradical by
electrochemical reduction of the corresponding dication with a rotating
platinum electrode and assigned the triplet EPR spectrum with the zero-
field splitting parameters of D =0.0079 cm'l, E S 0.0005 cm'1 for Am=1.

The zero-field splitting parameters D and E can be determined directly
from the triplet EPR spectrum; D and E values provide information on the
distance between radical centers and the molecular symmetry of the triplet,
respectively. The three levels of a triplet EPR spectrum (according to x, y,
and z axis) are split even in the absence of an external magnetic field. This

ero field splitting (zfs) arises from a dipolar coupling of the two spins that

 

 

creates an internal magnetic field in the molecule, which splits the energy

levels.

Siimmermann could not detect the half field transition, a hallmark of a high
spin state, from the spectrum of Schlenk biradical.32 The "half—field
spectrum" corresponds to Am=2 transitions in a triplet EPR spectra. This
transition is quantum-mechanically forbidden, but in most organic -
diradicals, the zero field splitting relaxes the selection rule and the
transition can be observed. It is a critical spectral feature because its

presence unambiguously signals that one is observing a triplet state.

Shortly after Gomberg's preparation of the triphenylmethyl radical in

1900, Chichibabin attempted to synthesize the analogous diradical 7.33 He
obtained a blue-violet compound that reacted rapidly with oxygen, yielding
polymeric peroxide. Early studies on Chichibabin's hydrocarbon 7 led to
peculation as to whether it existed as a singlet, a triplet, or a mixture of the
wo spin states.34 Numerous attempts to observe triplet spectra in solution
ave failed; doublet spectra were found, however. Platz discussed existing
xperimental inconsistencies and assigned a singlet ground state to
hichibabin's hydrocarbon. Montgomery reported a crystal structure of
hichibabin's hydrocarbon and concluded that it has a singlet ground state
ith an unusually large amount of "diradical character" by comparing each
nd length in the molecule with di-p-xylylene framework.35 He only
corded a singlet EPR spectrum even with the single crystal of the

adical used for the X-ray structure determination.

 

 

 

 

1 1
Ph Ph Ph Ph
Chichibabin‘s Biradical 7 Thiele's Hydrocarbon 8

Thiele synthesized the first isolable derivative of p-xylylene 8 in
1904, after attempts to prepare the unsubstituted compound failed.36
Thiele‘s hydrocarbon 8 reacted with oxygen, but could be manipulated
without elaborate precautions. The Thiele's hydrocarbon has been
examined by a variety of physical methods, and there seems to be little
doubt that the molecule has a singlet ground state. Montgomery reported a
crystal structure of Thiele's hydrocarbon and concluded that it has a singlet

ground state.
2.1.3. Ballester's Stable Triarylmethyl Radicals

Since Gomberg's discovery of the first free radical a great number of
stable organic radicals have been detected and isolated. Although carbon
centered free radicals are highly reactive species, some of them show
emarkable stability, the best example being Ballester's
erchlorotriphenylmethyl radical series shown below.37 Some of these
adicals have half-lives of decades in the air, and withstand typical radical
eagents like nitric oxide, hydroquinone, quinone, and even highly
ggressive chemical species like concentrated sulfuric, nitric acids, sodium
ydroxide, or halogen, with little or no alteration. Also, they possess a
markably high thermal stability, up to 300 °C. They are all completely

'sassociated in solution.

 

12

R Ballester‘s PMT radicals R

   

R. R'. R"=C1. C6C15 R =H. Br. 1. CH3, cozn, C02CH3, co,NH4

These radicals have distorted D3 propeller structures (C2) in the solid state,
and their high stabilities have been traced predominantly to the shielding of
the trivalent carbon atom by their phenyl rings with a propellerlike
:onformation and the six ortho chlorines. Thermal and chemical stabilities
f free radicals are governed by two main factors: resonance and steric
indrance.38 The inhibition of resonance and steric hindrance frequently
ause high thermal stability and chemical inertness of the aromatic

lorocarbons.39 The perchlorotriphenylrnethyl radical has twist angles of

.3, 53.4, and 538° between the phenyl groups and the reference plane of
e central carbon in the solid state. The effect of the absence of ortho or
eta chlorines on the twist angles of the rings, and therefore on their
bility, has been assessed by direct comparison of X-ray structures of
rious derivatives of Ballester's radical by Veciana and coworkers. The
ults were consistent with the steric shielding of chlorine atoms being the

in reason for the observed stabilities of these radicals.

 

 

—M‘ m ‘A 7"; V: ‘2

13
2.1.4. Jackson and Kahr's Lithium Complex of Triarylmethyl

Tris(2,6—dimethoxyphenyl)methyl 1, originally synthesized by J. C.
Martin, is a remarkably stable organic free radical.40 It can be synthesized
efficiently using the procedure shown below. It is air-stable, presumably
due to the fact that in its D3 propeller conformation the central trivalent
carbon is protected from above and below by six ortho-methoxy groups
similar to Ballester‘s radicals. Martin reported detailed EPR studies of the
radical 1 and its 2D and 13C (central carbon) derivatives with the
correlation of hyperfine coupling constants and the average twists of aryl

rings of 47° in solution.

 

Li
0/ E5 0 O/
MeOJLOMe
——> ————->~ HOC

O O 3

\ / \ /
O 0

11+ +C CI'Clz C

O 3 O 3
\ \1

In 1985 Jackson and Kahr treated solutions of the radical 1 with solutions of
lithium salts in order to study the possibility of complex formation between
the radical 1 and metal cations. Precipitates formed upon mixing that
requently gave unusual, intense ESR spectra. Some of the spectra were
ery broad, others had considerable fine and hyperfine structure, and often
—values deviated markedly from the free electron value.41 During the

ollowing year they unsuccessfully tried to establish a structural basis for

44

14
the surprising magnetic behavior of their precipitates by growing single

crystals from the precipitated solids. Results at that time suggested
colligative magnetic properties as a possible explanation for their puzzling
and highly variable ESR spectra.

Their proposed model did not include any of the structural elements
suggested by available strategies for achieving ferromagnetic coupling of
spins between carbon centered radicals.42,43 The general line of research
was set aside because, at the time of the initial discoveries, the principle
investigators were graduate students whose chief responsibilities were
unrelated to the present subject or the chemical system described. However,
in 1989, Professor Jackson reexamined precipitates of 1-LiBF4 and
l-LiClO4 with a Superconducting Quantum Interference Device (SQUID)
magnetometer in the Department of Chemistry (Professor Dye's SQUID) at
Michigan State University.

 

While a polycrystalline sample of 1 behaved as a paramagnet, the
susceptibility of the "salted" precipitates showed a ferromagnetic field—
dependent hysteresis, signatures of colligative magnetic behavior. Surprised
by these new results Professor Jackson revived investigations of organic
magnetism with the aim of establishing a rational basis for the observed
magnetic phenomena. We are now pursuing this research with the
expectation that a deeper understanding of electron coupling between simple

aramagnets will result.

 

15

2.2. Nitrogen Centered Radicals

2.2.1. Hydrazyl Radicals

The discovery and early investigations of hydrazyl radicals were
made mainly by Goldschmidt and coworkers during the period of 1920 to
1929.44 From his original observation that treatment of triphenylhydrazine

with lead dioxide in ether gave a deep blue solution which rapidly changed

 

to green and then a red—brown, he proposed that the labile blue intermediate

was triphenylhydrazyl 9 and the final product hexaphenyltetrazene 10.

Q
QNN $0.14
2:30 (1,6)

9 10

 

e equilibrium between radical and dimer was further confirmed by the
ependence of the blue color on the polarity and temperature of the solvent.
oldschmidt also prepared and characterized various hydrazyl radicals
cluding the remarkably stable, 2,2-diphenyl-1-picrylhydrazyl (DPPH);
hown on the next page which shows no tendency to dimerize in the solid
ate or in solution even at room temperature.45 Because of its applications,
terest in hydrazyl chemistry has been focused on DPPH. It has been used
nce 1950 as a radical scavenger in polymer chemistry and radiolysis and

a lesser extent in synthetic organic chemistry.46

 

 

No2

DPPH

Very early magnetic susceptibility measurements established that under
zarious conditions DPPH was essentially 100% free radical.47 It has been
ised as a reference radical frequently in connection with EPR spectroscopy
)ecause of its availability and stability. The EPR spectrum of crystalline
;amples of DPPH was measured and found to consist of one narrow line
tbout 1.7 gauss wide, but in a dilute benzene solution it shows a five line
spectrum which was originally attributed to equal coupling of the unpaired
:lection with the two nitrogen atoms, aN1=aN2= 8.9 gauss. More refined

alysis of the spectrum revealed that the two nitrogens in DPPH were not
quivalent and had coupling constants of 9.35 and 7.85 gauss. The larger of

ese two constants was shown to be associated with the picryl nitrogen by
5N labeling studies.48 Apparently, for some reason, there are no studies
n these radicals as base radicals for organic magnetic systems. They

ould have significant potential for such applications in comparison with

er nitro functionalized organic radicals.

 

 

 

l7

Oxidation of the dihydrazine 11 with lead dioxide yields unstable
pentaphenyl picryl mono hydrazyl (PPPMH) 12 which has hyperfine
coupling constants similar to other meta-substituted picryl hydrazyls .
Further oxidation gives the very unstable "pentaphenyl picryl dihydrazyl
(PPPDH)" 13, which is considered to be diamagnetic below -30 °C from
EPR and NMR measurements. Some workers claim to have obtained the '

EPR spectrum of the biradical form shown below.49

2.2. Nitroxyl Radicals

Porphyrexide 14 was the first organic nitroxyi radical to be isolated
.d characterized even though it was formulated as a "derivative of
iadrivalent nitrogen.50 A number of diarylnitroxides 15 were
bsequently prepared and characterized as radicals which do not dimerise

the solid state or in solution (R=OMe). 51

 

 

 

18
r 9
NH R : ‘R
HN R=H,OMe
14 15

2,2,6,6-tetramethylpiperidine-N-oxy1 (TEMPO) 16 has been studied
extensively, and its Lewis basicity has been studied mainly by Drago.52
EPR spectra of TEMPO in solution have been reported to be solvent
sensitive. Drago tried to explain the facts by the correlation of changes in
:he hyperfine coupling constant on adduct formation with enthalpies of
interaction and with changes in the infrared spectra of the hydrogen-
aonding acid upon interaction. The nitrogen hyperfine coupling constant is
found to increase with the enthalpy of adduct formation. Quite a few
Iariations on this sterically hindered nitroxyl have been reported, mainly

'rom France and Russia including 4-hydroxy TEMPO (TEMPOL) 17.53

O' Q
{)4 N
OH
TEMPO 1 6 TEMPOL 1 7

. K. Hoffmann reported a series of t—butyl nitroxides prepared by
ectrochemical or sodium reduction of nitrobutane in glyme, with a
oposed mechanism shown below, in 50% yield. The method was not
neral for the preparation of other di-t-alkyl nitroxides, but, t—butyl-2,6-
nethoxyphenyl nitroxide 18 is prepared by hydrolysis of the reaction

 

 

 

19
mixture obtained from 2,6-dimethoxyphenyl-lithium and nitrobutane in

17.8% yield.

0' t Bu
/ t-B ' \ ,O' t-B
t-Bu—NOZ & t-Bu—N‘“ Na+ u N+ Na+ fl» u\N—O'
\ / \O- B x
O' t—Bu t‘ u

Ar -
\ O]L1+ H20 Ar\

R/ \O R’N—O.
0' \O ('3'
#14 £er
0
l
18

Diphenylnitroxide 15 (R=H) is only stable in dilute solution. In
:oncentrated solution or in the solid state it behaves like t-
)utylphenylnitroxide 19, decomposing spontaneously. Thus, the increase in
.npaired electron delocalization, which occurs as a result of replacing a t-
utyl group by a phenyl group, merely increases the number of reactive

ites in the molecule and not its stability.

'0. ,0.
©le CCU
19
s in the t-alkyl aryl series, however, stable diarylnitroxides are produced
len C—O dimerization is prevented by the presence of blocking groups in

= para—positions, thus, di—para-anisyl nitroxide 20 which is believed to be

i first organic radical studied by single crystal X—ray and 4,4'-

 

 

 

 

 

20
linitrodiphenyl nitroxide 21 are stable for months. In this respect the

:hlorine substituted system is apparently inadequate since it decomposes
'apidly. Di-(p-tolyl)nitroxide (di—p-methyl) is sufficiently stable to be
solated but decomposes within hours.54

0 (I)
1'“ ONO
\0/0 00/ N02 N02
2 0 2 1

)rtho-substituents in diaryl nitroxide stabilize the radical probably by
'orcing the phenyl groups out of conjugation with the radical center,
hereby reducing unpaired electron density at the para-positions.55 Of
>articular interest is bis-(2,6-dimethoxyphenyl) nitroxide 22 shown below.
t has a very interesting structure not only by the similarity to our
riarylmethyl systems, but also for the natural extension of our rationale in

esigning molecular magnets by ion-bindings which I will discuss later.

\0 O/

\O

22

 

 

21

2.3. Aroxyl Radicals

It has long been appreciated that aroxyl radicals are intermediates in
many phenol oxidations, and interest in these radicals has been enhanced by
the fact that they play a role in many biological oxidations. Since the
discovery of 2,4,6-tri-t-butylphenoxyl radical 23, attention has been given
largely to hindered phenoxyl radicals such as galvinoxyl 24, some of which

have been isolated and are sufficiently stable to be used as reagents.56

O' t-Bu t—Bu
t-Bu t—Bu O' I I O
t-Bu / t-Bu
t-Bu
2 3 Galvinoxyl 24

Of special interest are a number of stable aroxyl biradicals. Oxidation of
the quinone methide with lead dioxide gives first the mono-radical 25, its
EPR spectrum shows interaction of the unpaired electron with the four
equivalent ring protons of the conjugated rings (aH=1.3 gauss). On further
oxidation Yang's biradical 26 is obtained in which there is delocalization
between the three rings giving seven lines from the six ring protons

(aH=0.86 gauss).57

 

 

 

22

 

2 5 Yang's Biradical 2 6

Yang's biradical 26 is the best characterized stable biradical with high spin

multiplicities among conjugated systems.58

 

 

 

 

23

3. Magnetism
3.1. Diamagnetism

If a sample is placed in a magnetic field H, the observed field with a
sample will generally differ from the free space value. If the observed field
with a sample is reduced, the substance is said to be diamagnetic.59 The
molar susceptibility of a diamagnetic material is negative, and rather small
{-1 to -100 x 10-6 emu/mol). Diamagnetic susceptibilities do not depend on
field strength and are independent of temperature. Since organic
:ompounds with unpaired electrons also have a number of filled shells, they
nave a diamagnetic contribution to their susceptibility. Diamagnetic
susceptibilities of atoms in molecules are largely additive, and this provides
1 method for the estimation of the diamagnetic susceptibilities of ligand

itoms and counter ions in a complex like 1-MX.

\OI’ Li+ : -1.0 x 10-66emu/g atom
C104 : -32.0 x 10 emu/g atom
(Q g (3104‘ C25 :-6.0 x 10'6-2mu/g atom x 25
H27 : -2.93 x 10 emu/g atom x 27
O 06 : -4.61 x 10.6 emu/g atom x 6
/ § -2.89 x 104 emu/mol

 

he Pascal constants provide an empirical method for this procedure. One
Ids the atomic susceptibility of each atom, as well as the constitutive
rrection to take account of such factors as n-bonds as shown above for the

mplex l-LiClO4. This procedure is only of moderate accuracy, and the

lues given could change from compound to compound. Greater accuracy

 

24
can be obtained by the direct measurement of the susceptibility of a

diamagnetic analogue of the paramagnetic compound.

3.2. Paramagnetism

A paramagnet concentrates the lines of force provided by an applied -

magnet and thereby moves into regions of higher field strength. This
results in a measurable gain in weight in Gouy or Faraday balances.
Paramagnetic susceptibility is generally independent of the field strength,
but it is temperature dependent. To a first approximation at high
temperature, the susceptibility x varies inversely with temperature, which is

the Curie Law:

:2
Zr

C is called the Curie constant, and T is the absolute temperature. Since x '1
= T/C, a plot of x '1 vs. T is a convenient procedure for the determination
of the Curie constant; note that the line goes through the origin for

paramagnetic materials in Figure 1.

Since the magnitude of magnetic susceptibility at room temperature is an
inconvenient number, it is common among chemists to report the effective
magnetic moment, ueff , which is defined as

l

l l _
3k — ~ _ 2 2
pay =<—fi)2<xr)2 -[8 5‘5 +12] #3

 

hr
1
1‘

 

 

25
Here, k is the Boltzmann constant, 1.38 x 10‘16 erg/K, N

is Avogadro's
number, 6.022 x 1023 /mol and MB

= lei/27 2mc = 9.27 x 10'21 erg/G is the
Bohr magneton. (the units of treff is 1413).

X(emu/mol)

 

 

 

0 50 100 150 2(X) 250 300

Temperature (K)

Figure l. The molar susceptibility (26mm), a, and the

reciprocal susceptibility(x'1m01), b, as a function of
temperature for g=2, S=1/2.

Vhen an isolated radical is placed in an external magnetic field H ( the
Iceman perturbation), the field will resolve the degeneracy of the various
:ates according to the magnetic quantum number m S, which varies from -S
t S in steps of unity. The energy of each of the sub revels in field becomes
=m,guBH. where g is a Landé constant, characteristic of each system,

hich is equal to 2.0023 when S=1/2 and L=0 (total angular magnetic
lantum number).

 

 

 

 

26

 

 

 

H

Thus, at a given external field, the distribution (population) of radicals

among the various states can be calculated from the Boltzmann relation,

 

fix exr>( 571.5")

N]. k
5Ei being the energy level separation between the level i and the ground
state j. This model provides a basis for a simple derivation of the Curie
Law, which states that a magnetic susceptibility varies inversely with
emperature. For an organic free radical (S=1/2) in zero field, the two

evels ms =i'l/2 are degenerate, but split when a field Hz is applied.
'he magnetic moment of a radical in the level n is given as
“n = 5Ian/5H = -msgilB
1e molar macroscopic magnetic moment M is obtained as the sum over

agnetic moments weighted according to the Boltzmann factor and it can

approximated as following:

 

 

 

in

fol

Heft

 

 

27

M ~ N gZuBZHz/4kT

Static molar magnetic susceptibility is defined as x mol = Mmol/I‘I which is

in the form of the Curie law where the Curie constant can be expressed as
following when S=1/2.

 

C = Ngztth /4k = 0.125g25(s +1)cm3K/mo/

This is a special case of the more general spin-only formula, where

Heffz = gZuBzS(S+l) is the square of the "magnetic moment".

 

Ng2u32S(S +1) Ate/)2
z = \ = N
3kT 3kT

ft is interesting to examine the behavior of M in the other limit, of very
arge fields and very low temperatures, the magnetization becomes
ndependent of field and temperature, and becomes the maximum or
aturation magnetization Msat which the spin system can exhibit. This

ituation corresponds to the complete alignment of magnetic dipoles by the
)plied field.

M50! = Ng/‘lBS

 

 

28
3.3. Curie-Weiss Law

The Curie law is the magnetic analog of the ideal gas law, which is
expressed in terms of the variables p, V, T. For magnetic systems, one uses
H, M, T and the thermodynamic relations derived for a perfect gas can be

translated to a magnetic system by replacing p by H and V by 1/M.

There are many situations in which the Curie law is not strictly obeyed.
One source of the deviations can be the presence of an energy level whose
population changes significantly over the measured temperature interval
(magnetic phase transition); another source is the magnetic interactions

which can occur between paramagnetic centers.

To the simplest approximation, this behavior is expressed by a small
modification of the Curie law, to the Curie-Weiss law, where the correction

term, 6, has the units of temperature.

 

120w)

When 6 is negative in sign it is called antiferromagnetic; when 9 is
positive, it is called ferromagnetic. The constant, 0, characteristic of any
particular sample, is best evaluated when T>10K , as curvature of x-1

usually becomes apparent at smaller values of T.

 

 

 

29

800

0.1125

X (emu/moi) 0.075

400 l/x (moi/emu:

0.0375

 

O 50 100 150 200 250 300

Temperature (K)

 

.b)

 

 

XT(emu/mol) he =+10K

 

 

 

 

 

 

 

0.5 ,

‘ ‘_ e§=-10K

(D

 

l i i i
( 40 so 120 160 2 )0

Temperature (K)

Figure 2. a)The reciprocal susceptibility, x'l, extrapolated from
the high temperature data; b) xT, as a function of temperature.
These plots represent a system in which g=2, S=1/2, as well as
ferromagnetically coupled ( 9 =+10K) and antiferromagnetically
coupled ( 6 = -10K) systems.

 

 

 

 

 

 

 

 

 

 

 

3.4.

Blea
Bleai

For r.
by se
at J/i
susce
This

zero ‘
positi
the C
10w V1
2 whi

mOIe .

C0llplc
the sa:
bIOad’

 

30

3.4. Pairwise Magnetic Exchange

The magnetic susceptibility per mole of dimers was calculated by
Bleaney and Bowers via the application of Van Vleck‘s equation (the

Bleaney—Bowers equation).

_ (2Ng2/132 /kT)
3 + exp(—2] / kT)

For negative J (antiferromagnetic), x has a maximum. This may be found
by setting dx /dT=0. With the definitions used here, the maximum occurs
at J/kTm ~ -4/5. For ~J/k <<T (or T >> Tm)(at high temperature), the
susceptibility follows a Curie—Weiss law, x = 3/[4(T- 0 )] with 6 = J/2k.
This is a specific instance of the more general connection between a non-
zero Weiss constant and the presence of exchange interaction. When J is
positive, the susceptibilities calculated according to the above equation and
the Curie law for spin S=1/2 do not differ greatly until temperatures very
low with respect to 2.] /k are achieved; this is because of the extra factor of
2 which enters when comparing a mole of uncoupled S=1/2 spins with a

mole of dimers.

he susceptibilities for a pair of S=1/2 radicals antiferromagnetically
oupled are discribed in Figure 3. While the low temperature behavior is
he same, the temperature of maximum x increases with decreasing 0. A
road, featureless peak is observed because of the continuous population

ecrease of the various levels as temperature decreases.

 

311

(hassi
dhner
thh)

behav

Then:
descn
SySlen

than“

 

31

0.04

0.035

0.025

X (emu/mmol) 0.02

0.015

    

40
Temperature (K)

Figure 3. The Bleaney Bowers susceptibilities compared with
the Curie law; plots for J=0, -2.5, —5, -7.5, —10, -l2.5 cm'l.

  
  
 
 
 
   
   

3.5. One-Dimensional or Linear Chain Systems

Much of what has been covered so far could come under the
classification of short-range order. The physical meaning implied by low
dimensional magnetism is that radicals are assumed here to interact only

with their nearest neighbors in a particular spatial sense. This magnetic

 

behavior follows directly from the structure of the various compounds.

 

ere are good theories and extensive experimental data available which
escribe the thermodynamic properties of one-dimensional magnetic
ystems, at least for S=1/2; for systems of metal ions linked into uniform

hains. In order to discuss exchange at the atomic level, one must

 

 

 

 

 

 

 

 

 

 

 

 

 

intrt

Hart

This
mod

whet

The I
of th.
Situai
the I
Heisc

pract.

There
for a
Chara
in 1D
exten

S)lSter

Since
Calcu]

Pledic

 

‘3' ’ A :‘z—V-mé—‘v'

32
introduce quantum mechanical ideas and we need to use the following

Hamiltonian.

212—2st,- cs]-

This is an isotropic Hamiltonian and is often referred to as the Heisenberg
model. A very anisotropic expression, called the Ising model is obtained

when 7 is set equal to zero in the following equation.

H = —2JZ[y(51,52, + 31,52 , ) + 51,522]

The Heisenberg model, which corresponds to using all the spin-components
of the vectors 81 and 82, is obtained when 7:1. This appears as an artificial
situation, and yet the Ising model is important, in part, because solutions of
the Ising Hamiltonian are far more readily obtained than those of the
Heisenberg Hamiltonian. Both of these Hamiltonians are restricted in their

practical application to nearest—neighbor interactions.

There are no exact or closed-form solutions for the Heisenberg model, even
for a S=1/2 one—dimensional system. But, calculations are available which
characterize Heisenberg behavior to a high degree of accuracy, particularly
in 1D systems. Furthermore, one-dimensional short-range order effects are
extended over a much larger region in temperature for the Heisenberg

systems than for the Ising systems.

 

Since the Heisenberg model is an isotropic one, the susceptibility is
calculated to be isotropic. A broad maximum in the susceptibility is

predicted for a linear chain antiferromagnet. For an infinite linear chain,

 

 

33
Bonner and Fisher calculate that the susceptibility maximum will have a

value xmax/(N gZuBZ/IJI) ~ 0.07346 at the temperature kTmax/IJI ~ 1.282.

The best example of an antiferromagnetic Heisenberg chain compound is
[N(CH3)4]MnCl3, TMMC. The crystal consists of chains of J =5/2 manganese
atoms bridged by three chloride ions. The antiferromagnetic 71D behavior

in this nearly ideal Heisenberg system has been reviewed.60

The zero—field susceptibilities have been derived by Fisher for the S=l/2
Ising chain.

Na .211 2
95,), = —%iu / 2kT)exp(J / kT),

Ng 21132 2
ape, = ——p"l——[tanh(J/2kT)+(J /2kT)sech (J/2kT)].

Many susceptibility measurements are made on powdered paramagnetic
samples, so that only the average susceptibility, <x>, is obtained.

The average susceptibility <x> is defined as
< Z >= (Xpar + 2;rper)

3

e meaning of the symbols "parallel" and ”perpendicular" refer to the

xtemal magnetic field direction with respect to the direction of spin—

uantization or alignment within the chains, rather than to the chemical or

tructural arrangement of the chains. The zero-field susceptibilities for xpaI

d xper are plotted in Figures 4 and 5.

 

 

 

 

 

34
0.05
004 '
: J=0
. J=-5
003 '
X (emu/mol) ' J='7'5
I J=-10
002 F
: J=-12.5
001
h '+ +_+
. 3" ,
0 llllllJlllllllllll
O 20 40 60 80 100

Temperature (K)

Figure 4 The Ising "parallel" susceptibilities compared with
Curie law; plots for J: -2.5, -5, -7.5, -10, -12.5 cm‘l.

 

=~20

)-
\-

   

X (emu/moi)
0.08

    

,II'III[ll11I

 

 

 

ll11‘lirll

l I LPAJJJLLL

o
l.
1'
f

 

20 4o 60 80 100
Temperature (K)

0

Figure 5 The Ising "perpendicular" susceptibilities for
J: -20, ~30, -40, -50 cm-l.

44

inl

alt

int

6X

fer

SUE

0n
alt:
alt:
the

(let

 

 

 

35

3.6. Alternating Linear Heisenberg Chain

The models discussed so far are uniform linear chain systems. The
intrachain exchange constant has been assumed not to vary with position

along the chain, and it is assumed that there is no interchain interaction.

H = —212[s,-_1 . s,- + as, . [+1]
.-

An alternating chain is defined by letting 0t be less than one. For a
Heisenberg alternating linear chain, -2] is then the exchange interaction
between a spin and one of its nearest neighbors, and -2(XJ is the exchange
constant between the same spin and the other nearest neighbor in the chain.
When 0t=0, the model reduces to the dimer model with pairwise
interactions (the Bleaney—Bowers model). The alternating linear chain
antiferromagnet with S=1/2 has been studied in detail both theoretically and
experimentally.61 In 1955 Oguchi developed a general theory of
ferromagnetism and antiferromagnetism, based on the Heisenberg model of
two—spin clusters. Ohya-Nishiguchi later extended this theory to the

susceptibility of interacting spin-pair systems.

One of the important susceptibility results for an antiferromagnetic
alternating Heisenberg linear chain is a broad maximum for all values of the
alternation parameter or. The susceptibility curves decrease exponentially as
the temperature approaches zero.62 The zero—field susceptibilities have been
derived by Fisher for the S=l/2 linear Heisenberg chain where u=(coth K-
1/K) with K=3J/2kT.

 

Each
susce]
derive
4' is I
SUSCeI

chain

36

1 h 1—
_23Ns#2 ”(“5317 T) /(23/<T)1

4kT [ 23JT 3] J
1— til—+1 ——
co (2kT) /( kT)

 

J=-10
/

. f \ J=-20
0.008 4 \ / I an;

x (emu/moi) 0.006 -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 20 40 60 80 100
Temperature (K)

Figure 6. The Heisenberg linear chain susceptibilities for
J = ~10, -20, —30, -40, -50 cm-l.

Each of the magnetic models described above (Heisenberg and Ising chain
susceptibilities) can be modified by the addition of a mean-field correction
derived by Carlin where Z is the number of neighboring dimers or chains,
I ’ is the coupling constant between dimers or chains, and x is the molar

‘usceptibility. The Bleaney—Bowers model turns into the alternating linear

 

hain model with the mean-field correction when 2:2.

)0:— ' X 7
1— an /Ng'#32 )x

 

4. Molecule
4.1. Molect
4.1.1. McC

In 1963
on the studies
pr0posal was
hydrogen mo]
free radicals (
negative atom:
radicals may 5
01 positive sp
density in nei
interaction far

molecules,64

If We Conside

inleflCthI] bet

the Hamiltonia

 

37

4. Molecular Magnets
4.1. Molecular Magnet models
4.1.1. McConnell Model

In 1963 McConnell proposed a model for organic ferromagnets based
on the studies by Edelstein and Mandel,63 of solid organic free radicals. His
proposal was based on the Heitler-London spin exchange principle for a
hydrogen molecule. He pointed out that in certain aromatic and olefinic
free radicals (especially odd-altemant radicals) there are large positive and
negative atomic rc-spin densities, and it is possible that in special cases these
radicals may stack on top of one another in the crystal lattice so that atoms
of positive spin density are exchange coupled to atoms of negative spin
density in neighboring molecules, resulting in a ferromagnetic exchange
interaction favoring parallel total spin angular momentum on neighboring

molecules.54

If we consider only two center exchange integrals, then the exchange
interaction between two aromatic radicals A and B can be approximated by

:he Hamiltonian

 

 

AB_ AB A B
H “21,1. 5,- as,
U

 

where 81A is t

electron spin 1

in the form

where S A and
are the rt-spin
McConnell a1
coupling in so

are usually p0

for aromatic 1r

He late.
Superexchange
"If an ionic m1
Whose neutral
traIlSlcer to lea

molecules. T
acceptor mole

POSSibility."

38
where SA is the n-electron spin on atom i of molecules A, and SjB is the 1:-

electron spin on atom j of molecule B. The equation above can be written
in the form

HA3 2 —SA .SBZJU ABpiA .ij
’7

where S A and SB are the total spin operators for A and B, and piA and ij
are the n-spin densities on atoms i and j of molecules A and B.
McConnell attributed the dominant tendency toward antiparallel spin

coupling in solid free radicals to the fact that the spin densities piA and ij
are usually positive, and the largest exchange integrals JO. AB are negative

for aromatic molecules that are stacked on top of one another in crystals.

He later proposed another recipe for molecular magnets using
superexchange in charge transfer complexes. 65 He stated in this paper,
"If an ionic molecular crystal D+A— could be formed with a donor molecule
whose neutral ground state was a triplet, then one would expect back charge
transfer to lead to ferromagnetic coupling of the spins on adjacent ion
molecules. The same effect could be achieved if, instead, the neutral
acceptor molecule A had a triplet ground state. Thus, ferromagnetic
coupling of the spins of free radicals in certain ionic molecular crystals is a

possibility."

 

 

 

—4

13+

if

A'

lhe result of r
Spins on two 1
transfer can le.
such charge m
be parallel to '
parallel to both
10 a ferromagn
could be two-dj

“’2' Matag.

In 1968,
80m extended

slitting Whi Ch
tOpologlcal nan]
the mechanism
Possible lollies
“em page, are

number of 1h C g:

39

—+ +H++ —

D+ A‘ D A

+4 +«—»++ 4+

D+ A D

The result of such charge transfer will be to favor a system in which the
spins on two neighboring species D” and A" are parallel, so that charge
transfer can lead to the favored triplet state of one of the partners. Since
such charge transfer can go in either direction, the spin on a given A" will
be parallel to the spins on both neighboring D“, which in turn must be
parallel to both other neighboring A". Such parallel spin correlation leads
to a ferromagnetic domain; if the charge transfer is not only linear there

could be two-dimensional or three-dimensional domains.
4.1.2. Mataga Model

In 1968, Mataga proposed several possible ferromagnetic states of
some extended hypothetical hydrocarbons with conjugated rt-electron
systems which will show ferromagnetic spin alignment due to the
topological nature of the MOS. Qualitative discussions have been given for
the mechanism of couplings in the "ferromagnetic state", as well as some
possible routes for their preparation.66 These compounds, shown on the
next page, are all altemant hydrocarbons (even or odd according to the

number of the groups and atoms).

 

 

These are poly
and they are a
assumed that
participates in
remains as a "1

these compoun

These meta-gut
to the HiiCliel IV
orbitals (NBM
electrons as ind:
If the Orbital E
sufficiently Sma

40

:C::C::C:: Czl

gogczCCECK

c: C: C: Cz’
These are polymeric carbenes or extensions of the triphenylmethyl radical,
and they are all meta-substituted compounds. In the case of carbenes, it is
assumed that one of the unpaired n-electrons of the divalent carbon
participates in conjugation with the ring it-electrons while the other of them

remains as a "n-orbital." For the sake of simplicity, Mataga assumed that

these compounds are all co-planar.

these meta-substituted hydrocarbons have non—bonding n-MO'S according
o the Hiickel MO theory and the number of such n-Non Bonding Molecular
)rbitals (NBMO) is equal to the number of carbon atoms with unpaired
lectrons as indicated in the structure.

fthe orbital energy differences between the non-bonding orbitals are

rfficiently small, all unpaired electrons will have parallel spin according to

 

 

 

Hund's rule

extended over
relatively cor
electronic stru
like, and the '

electron spins
4.1.3. 0vch

In 1978
multiplicity of
coIljllgated bor
the class of al
1W0 EFOUps are
than zero. He
shOWIl below 1

electrons in a I)

gives Very little

41
Hund's rule (extended to a molecule). Although the n-NBMO's are

extended over the whole molecule, the electron density in a rc-NBMO is
relatively concentrated on each trivalent atom. In this situation, the
electronic structure on each carbon with the n-orbital may be rather atom-
like, and the "intra-atomic" interactions probably induce alignments of the

electron spins in all the n-orbitals.
4.1.3. Ovchinnikov Model

In 1978 Ovchinnikov proposed an organic magnet model based on the
multiplicity of the ground state of large altemant organic molecules with
conjugated bonds which are similar to Mataga's model. He showed that for
the class of altemant hydrocarbons in which the numbers of atoms in the
two groups are unequal (i.e., for the odd altemants) the total spin is more
than zero. He presented some possible planar and linear molecules such as
shown below having the full spin proportional to the number of unpaired
electrons in a molecule.67 This model is basically similar to Mataga's and it

gives very little insight into how or why high-spin is expected.

QWQQQ‘
QIQIOIOC

 

 

 

4.1.4. Bresl

In 1982
ions as buildiI
pr0posa1 (char
conlugated sys
states. If stat
crYStallize as r
be used to p“

and t0 the fern

Over the year
Which have ap
shown to exist
been shown to
salt of this cat
Showed a typir
an E Value of

being examine

 

42

| l l
/C- B C°\
| l I I
/B\©/B\©/B\ :; /B\
/B\ /C°\ /B\

4.1.4. Breslow Model

In 1982 Breslow proposed a series of stable high spin molecules or
ions as building block for molecular magnets based on McConnell's second
proposal (charge transfer model).68 He suggested that cyclic 4n n-electron
conjugated systems with C3 or greater symmetry should have triplet ground
states. If stable examples of such systems could be prepared they might
crystallize as molecular magnets. He outlined a theory of how triplets could
be used to prepare ferromagnets. Synthetic approaches to stable triplets,

and to the ferromagnetic system were described.

Over the years, the Breslow group has examined a number of systems
which have appropriate symmetry and electron count and which have been
shown to exist to some extent as triplets. The simplest of these, which has
been shown to exist as a ground state triplet is cyclopentadienyl cation. A
salt of this cation 27, prepared from 5-bromocyclopentadiene and SbF5,
showed a typical triplet EPR spectrum, with a D value of 0.1865 cm'1 and
an E value of O. The zero value of E indicates that the triplet molecule

being examined has a plane polygonal geometry, as expected by symmetry.

 

 

The rather lar‘

average fairly

Unfortunately.
for the grount
simple ideas
thIIlperature dc
0f Pentapheny
state lying of
energy gap tl
equilibrium wj
unsubstituted (
down to 4K, sr

MeO

MeOJ

Although the c'
triplet EPR SP1

rellhirements r

43
The rather large D value indicate that the two unpaired electrons are on the

average fairly close.
Br Br

Br + Br

Br

27
Unfortunately, even in a symmetrical 4n-7r-electron molecule, it is possible
for the ground state to be a singlet, in contrast to expectations based on
simple ideas about orbital degeneracy and Hund's rules. From the
temperature dependence of the EPR signal it was seen that the ground state
of pentaphenylcyclopentadienyl cation is actually a singlet, with a triplet
state lying 0.35 to 1.15 kcal/mole higher in energy. With such a small
energy gap there was a detectable population of triplets in thermal
equilibrium with the singlet at essentially all temperatures. By contrast, the

unsubstituted cyclopentadienyl cation triplet followed the Curie—Weiss law

J

down to 4K, so it is a ground state triplet.

,/\N

 

“N7 29

\lthough the dications (28 and 29) of the triphenylene shown above have
riplet EPR spectra and the charge—transfer solid appears to have met the

equirements of Breslow's model for ferromagnetism, these dicatlon

 

 

systems were

dependence st

4.1.5. Torra

In 198‘
containing stg
with iodine at
phase to form
The reaction
reproducible.
Structure Shov
analyses. }

femMagnetic

4'1‘6‘ Kahn

Kahn ha
MCCOHHEH m‘

 

 

 

44
systems were also found to be ground-state singlets based on temperature

dependence studies with a Faraday balance.69

4.1.5. Torrance Model

In 1987 Torrance proposed a model for organic ferromagnets
containing stacks of radical ions.7O Symmetrical triaminobenzene reacts
with iodine at room temperature in a variety of solvents and in the vapor
phase to form a black, insoluble polymer of radical cations of aryl amines.

The reaction is complex and the resulting polymer was not very

 

reproducible. He proposed a possible precursor of "unoxidized" polymeric
structure shown below which was consistent with IR data and chemical

analyses. However, the extreme lack of reproducibility of the

 

ferromagnetic material has made characterization troublesome.

N‘N/ \N‘N/

’

N N
I l

.1.6. Kahn's Pr0posal

Kahn has recently presented a careful and detailed discussion of the

cConnell mechanism for the possible mixing of ground and charge-

transfer confi
states. He Sht
transfer mixin-
for the triplet
stabilized trip
energies and(
organic magn
success. In ad
stacks is suggt
might be obt;

molecules in tl

 

 

45
transfer configurations of a donor—acceptor pair for both singlet and triplet

states. He showed that in contrast to McConnell's assumption, more charge-
transfer mixing terms are available for the stabilization of the singlets than
for the triplet in the charge-transfer complexes. For the superexchange-
stabilized triplet to be favored, a delicate balance of magnetic orbital
energies and overlaps must be achieved, which suggests that approaches to
organic magnets based 0n the above strategy have little probability of ‘
success. In addition, a modification of McConnell's mechanism for uniform
stacks is suggested. Kahn proposed that a triplet ground state of the dimer
might be obtained only for a highly symmetrical arrangement of the

molecules in the stack.71

 

 

4.2. Metal

It is m
magnetism it
appeared rece

charge transfe

Studies on ma
and Bowers

COpper(H) ace
established a t
magnetic susce
the energy par
ions in dimerit
were Synthes:
Comparing the
and more reset
1“ he light of.
and magnetic

Understanding (

4'11. Single

 

46

4.2. Metal Based Molecular Magnets

It is my intention to summarize the current state of molecular
magnetism in this section even though many excellent reviews have
appeared recently from the leaders in their areas-mainly bimetallic and

charge transfer systems.72

Studies on magnetic polynuclear complexes started in 1952, when Bleaney
and Bowers demonstrated that the magnetic and EPR properties of
copper(II) acetate were due to the dimeric nature of the molecule. They
established a theoretical expression (the Bleaney Bowers equation) for the
magnetic susceptibility of such a system as a function of the temperature and
the energy parameters J characterizing the interaction between the Cu2+
ions in dimeric copper (II) acetates.73 Numerous polynuclear complexes
were synthesized, and their coupling parameters J determined by
comparing the observed magnetic data with the theoretical values. More
and more researchers then attempted to rationalize observed magnetic data
in the light of structural data. This allowed correlations between structural
and magnetic properties to be made, and provided a step toward an

understanding of the mechanisms of interaction between metal centers. 74
.2.1. Single Atom Bridged Bimetals
The single atom bridged bimetallic systems have been followed by

ahn et al.75 and Drillon et al,76 who used metal ions bridged by organic

igands. In this case, one-dimensional ferrimagnets (incomplete cancellation

 

 

of spins by th
which in sorr

highest ferron

Bimetallic cor
compared to

coupled). The
based molecui
atom bridging
have been attr
magnetic orbit

0f ”men and

interactions in

 

Anderson diSCl
interactions am

as electron der

Increase in eler

Coupling. This
°°W0rkers, 7a

 

47
of spins by the differences in magnitude of interacting spins) were obtained,

which in some cases ordered ferromagnetically at low temperature. The

highest ferromagnetic phase transition temperature was reported to be 14K.

Bimetallic complexes exhibiting a spin-triplet ground state are very rare as
compared to those with spin-singlet ground state (antiferromagnetically
coupled). There are three known cases of ferromagnetically coupled metal
based molecular magnetic systems, two of which are bridged by a single
atom bridging ligand.77 The ferromagnetic properties of these compounds
have been attributed either to accidental or to strict orthogonality of their
magnetic orbitals or to spin-polarization effects using Hund's rule for both
of oxygen and nitrogen ligands. Thus, it is very rare to have ferromagnetic

interactions in single atom bridged bimetallic systems.

 

ferromagnetic antiferromagnetic

nderson discussed the influence of ligand substituents on superexchange
teractions and concluded that the antiferromagnetic interaction decreases
s electron density is removed from the bridging atoms. Conversely, an
crease in electron density raises the ligand levels and enhances the AF
upling. This conclusion was also reached by Hodgson, Hatfield, and their

workers. 78

4.2.2. Mole

The at
interaction be

demonstrated

 

The Symmetrir

was understo<
value of -193 r

tetraamineeth;
However, due
carbOxylato 1

exhibited a J/k

TSipiS rePOrte
Cum) binuck
proposed two
goes through
Very few bim.

magnetic 1mm.

 

48
4.2.2. Molecule Bridged Bimetals

The ability of the carboxylate group to mediate the exchange
interaction between two Cu(II) ions separated by more than SA has been well

demonstrated in the case of (oxalato) copper (H) compounds shown below.79

   

oxalate bridge dicarboxylate bridge

The symmetric orientation of magnetic orbitals in the bimetallic complexes
was understood as a reason for a large antiferromagnetic J /k coupling
value of -193 cm'1 in [(tetraamineethylenediamine(H20)Cu(C204)Cu(H20)
tetraamineethylene diamine](ClO4)2 with a Cu—Cu separation of 5.14 A.
However, due to the unfavorable magnetic orbital configuration of its

arboxylato bridge, the [Cu(NH3)2(CH3COO)Br] polymeric complex
xhibited a J/k value of -4.3K. 80

sipis reported magnetic studies on a ferromagnetic phthalato-bridged
u(II) binuclear complex 30 with its Cu(II) ions separated by 6.7 A. He
roposed two possible pathways for spin-orbital couplings, one of which
oes through two carboxylate and aromatic at-orbitals. This is one of a
cry few bimetallic molecular bridged systems showing very long-range

agnetic interactions.81

 

 

Least-squares
susceptibility .
energy gap of
bridged Cu(II
obtained for a
orbital interpr
ferromagnetic

4.2.3. Metal

One of
magnets is has
and Stable Org:
radicals haVe
chemishy has ‘
has Weak Lew
metal ions,

Paramagnetic r

 

 

49

‘\ O o
\_ ’ 0
ENth 0 O \C .N
N N "UN:
\___1 Ni {

30 5
Least—squares fitting of the variable-temperature (4.2-295K) magnetic
susceptibility data to the Bleaney-Bowers equation indicated a singlet-triplet
energy gap of 80 i 10 cm'l, g=2.10, and J = -l.81 cm—1 for the phthalato-
bridged Cu(II) binuclear complex. However, no direct EPR evidence was
obtained for a triplet state (half field transition). From the standpoint of
orbital interpretation of the coupling, this is a new situation leading to
ferromagnetic interaction between magnetic centers separated by multi-

atom-bridging units.

23. Metal and Organic Radical Based Systems

One of the most unique approaches to the design of molecular
nagnets is based on a model of directly interacting paramagnetic metal ions
and stable organic radicals, such as nitroxides, by Gatteschi et a1.82 These
adicals have been widely used as spin probes, and their coordination
hemistry has been reviewed previously.83 The oxygen atom of the radical
as weak Lewis basicity, and can serve as a ligand toward many different
aetal ions. When a nitroxide binds through its oxygen atom to a

aramagnetic metal ion, the spins can orient anti—parallel to each other or

 

pandl

coupl

The s
easfly
(nbha
mgnn’
onhog

hlthe
The]:
radica
has b
Antife
Where
tibia
cOUpli

Inagne

 

50
parallel to each other, resulting in antiferromagnetic or ferromagnetic

coupling.

The sign of the coupling of a nitroxide directly bound to a metal ion is
easily predicted on the basis of orbital overlap considerations.84 If the
orbitals containing the unpaired electrons on the metal and the radical have

significant overlap, the coupling is antiferromagnetic, while if they are

orthogonal to each other, the coupling is ferromagnetic.
R R

 

ferromagnetic antiferromagnetic

In the ferromagnetic case, the two orbitals are orthogonal to each other.
The latter situation occurs in a copper(II) nitroxide complex, where the
radical is in an axial coordination site, and experimentally the triplet state
has been found to lie 10 - 70 cm'1 below the singlet state.
ntiferromagnetic coupling occurs in copper(II)-nitroxide complexes
here the radical coordinates in an equatorial coordination site. If the M-
-N angle is not 180°(very rare case), substantial overlap can occur and the
oupling becomes antiferromagnetic. This expectation was confirmed by

agnetic measurements showing singlet ground states for these systems.

Cu

m2

CO]

the

fer

pa

er]

ad

the

the

pai

ten

brt

altt

av

Slit

51
It has recently been demonstrated that the coordination polymer

Cu(hfac)2-TEMPOL 31, shown below, behaves like an alternating linear
magnetic chain.85 At high temperature, the magnetic behavior of the
complex shows characteristics of two independent spin S=1/2 moieties. As
the temperature of the sample is decreased, the spins interact
ferromagnetically, leading to a triplet ground state for the copper nitroxyl

pair.

OH
F3C

3i
The exchange constant, based on an analysis using the Bleaney-Bowers
expression, was found to be 2] /k = 19 i 7K. The average g value for the
adduct was found to be 2.08. At a temperature of 4.2K and below, one
therefore expects the system to consist of spin S=1 units with g=2.08, and
the question was asked whether there is a weak magnetic link between the
pairs via the saturated ligand and the "hydrogen bond" At lower
temperatures the magnetic susceptibility data increase and go through a
broad maximum at about 80mK; these data have been explained as an

ltemating Heisenberg linear chain with g=2.2 and 2J'/k= -78mK, which is

 

very weak exchange constant but is reasonable given the proposed lengthy

uperexchange path.

with t
differ-
yieldh
two (1:
bind t

nitrox

Accor
three-
report
Comp:
orden
and e
20K \
transi‘
COaxh
of me
Coils

 

52

Nitronyl nitroxides show very interesting coordination chemistry
with transition metal ions.86 It has been found that they can bind in several
different ways: they can bind via one oxygen atom to one metal ion,
yielding mononuclear complexes; they can bind via the two oxygen atoms to
two different metal ions to form magnetic chains. In these cases the radicals
bind to the metal ions, keeping their radical nature; the 7e* orbital of the

nitroxide has a weak overlap with the magnetic orbitals of the metal ions.

W H
nnuO" NYNi O'IIIIMIIIIIO" NYN‘ O‘IIMIIIH
R R

R = Ph, Me, Et, Py, etc.,

According to this model, ferrimagnetic chains have been obtained and
three-dimensionally ordered ferromagnets have been reported. Gatteschi
reported the magnetic properties of the Mn(hexafluoroacetoacetate)2
complex of nitronitronyl nitroxide shown above, which is ferromagnetically
ordered at 8K, and of the [Mn(pentafluorobenzene)2]2 complex of methyl
and ethyl nitronyl nitroxide, which show bulk ferromagnetism at 24 and
20K with a Weiss constant of 25K. He characterized the magnetic phase
transition using ac susceptibility measurements ( sample is placed within a
coaxial set of coils and a low—frequency signal is applied with a few gausses
of measuring (ac) magnetic field; the change in mutual coupling between the
coils is proportional to magnetic susceptibility) and reported low

emperature EPR spectra that show apparent short-range order effects,

providing inf

he was not ab

The 4-

acetylaceton:
acetoacetate).
the basis of :
complexes or
groups as sh<
assignment w
pyridine nitro
might form a

order ferroma

 

53
providing information on the preferred spin arrangements. Unfortunately,

he was not able to get the X-ray structure of the dimetal complexes.87

The 4-Pyridyl nitronyl nitroxide reacts with metal hexafluoro
acetylacetonates forming compounds of formula Mn2M(hexaf1uoro
acetoacetate)6(4-pyridylnitronyl nitroxide)2 32, with M = Mn, Co, Ni. On
the basis of magnetic and IR evidence, Gatteschi reported that all these
complexes exist as chains of radicals bridged to metal ions through NO
groups as shown below. Unfortunately, he was not able to confirm this
assignment with the X-ray structure of the complex. He proposed that the
pyridine nitrogen donors bind the metal M to connect different chains and
might form a ladder-like two dimensional structure. These compounds

order ferromagnetically below 10K when M and M' are Mn.

I llll O’/ “.0. "II-M ".0 '/N N+“O IIIMIIII
/ / i
\N I \N
/N I /N I
\ \

 

4.2.4. Char

The ch:
second mole
synthesized a
ferromagnetic
calculated usi:
is suggested 1
mechanism of

of the compor

The Charge t1
him much
thoroughly fez
Proposed to a
instead of th
magnetic bei
templex 0 f

metailocenilm

the Cp rings

 

I ‘ . -
—f—"'f Mm... -__-. _ - -

54
4.2.4. Charge Transfer Complex Based Systems

The charge transfer complex based systems motivated by McConell's
second molecular magnet model have been studied by Miller who
synthesized a compound of formula [Fe(Me5Cp)2](TCNE) 33 which orders
ferromagnetically at 4.8K and has a coupling constant of J = 19 cm'1 as
calculated using an empirical linear chain Heisenberg model. In this case, it
is suggested that the parallel alignment of the spins is achieved through a
mechanism of virtual charge transfer between the two different spin centers

of the compound.88

‘9‘

33

The charge transfer complex superexchange model of Miller which had

I
"'0. I ...\\
\
’I'O‘ -— ,'.‘

driven much of the synthetic activity in the field has recently been
thoroughly reanalyzed and questioned by Kahn, as mentioned earlier.89 He
roposed to apply McConnell's first magnetic model of spin polarization
nstead of the second charge transfer model to explain the observed
agnetic behavior in the decamethylferrocenium tetracyanoethenide
omplex of Miller. He argued that the unpaired electrons of the
etallocenium ions localized on the metal induce a negative spin density on

he Cp rings which interacts antiferromagnetically with the unpaired

 

 

electron in

polarization '

4.3. Galvii

Stable
mwmmMi
this country.
magnet was a
stable organi.
density induc
hydrocarbons
Possible to b1
many eXperts
Japan and Mt:

4'3°1- Galr

 

55
electron in the 75* LUMO of the acceptor molecule TCNE by spin

polarization to give ferromagnetic interactions.

4.3. Galvinoxyl and Nitroxyl Radical Based systems

Stable organic radicals such as galvinoxyls and nitroxyls have been -

used in building organic magnetic systems by several groups in Japan and in
this country. As discussed earlier, McConnell's first proposal for organic
magnet was a simple suggestion for the possibility of magnetic interaction in
stable organic radical in the solid state by the interaction of negative spin
density induced by spin polarization between radical centers in altemant
hydrocarbons. He did not propose or explain in detail "how or why" it is
possible to build such a system. Galvinoxyl radical is a stable radical that
many experts in the field pursued including Awaga, Kinoshita, and Mukai in

Japan and Miller in this country.

4.3.1. Galvinoxyl Radicals

Galvinoxyl, (4-[[3,5-bis(1,1-dimethylethyl)-4-oxo-2,5-cyclohexadien-
l-ylidene]methyl]—2,6-bis(l,l-di-methyl)phenoxyl 24, is ferromagnetic by
itself. Awaga and Kinoshita reported evidence for ferromagnetic
'ntermolecular interaction in crystalline galvinoxyl and proved that this
nteraction extends one-dimensionally with an exchange constant J of about

cm‘l. The condition of the ferromagnetic couplings have been discussed

the context of spin-unrestricted INDO calculations. 90

t-Bu/

The calculatir
at room tem;
hydrogen att
allows them
explained the
from charger
and B'Spin or
(NHOMO) ar.
Sllggested tha
stabilized rela
In Order to 0
transfer conf;
interacting 0,
Which the fer
that the form
(”bin (80]

intramolecula

Awaga
neasurements

 

56

 

The calculation was based on the molecular structure determined by X-ray
at room temperature, except that all the t-butyl groups are replaced by
hydrogen atoms.91 The nearly planar molecular structure of Galvinoxyl
allows them to separate the tt-type and the (Hype orbitals and they
explained the observed ferromagnetism by a dimeric interaction derived
from charge transfer between radicals. The large splittings between the 0t-
and B—spin orbitals in the non bonding highest occupied molecular orbital
(NHOMO) and non bonding lowest unoccupied molecular orbital (NLUMO)
suggested that the energies of triplet charge transfer configurations are well
stabilized relative to those of singlet ones by spin polarizations.

In order to obtain the magnitudes of the transfer integrals for the charge
transfer configurations, the intermolecular overlap integrals between the
interacting orbitals were calculated based on one-dimensional stacks, in
which the ferromagnetic interactions were found to work. They concluded
that the ferromagnetic interactions result from singly occupied molecular
orbital (SOMO) -NHOMO and SOMO-NLUMO interactions and

intramolecular spin polarizations.

Awaga and Kinoshita reported another interesting magnetic
measurements which was later confirmed by Miller. From the
measurements of the temperature dependence of the magnetic susceptibility,

it was found that the ferromagnetic intermolecular interactions, which are

lost below 8
maintained (1:
hydrogalvino:
bmmma
offering mec
measured the

transitions in '

For the purpr
SQUID at M
measurement
Chemical Co.
any problem
hydrogalvino:
and a Weiss c
to 157 K f0

Miller gave a]

"NO aPltarent

of these Stru.
“meme obt:
StnICture of
diSintegration
suggGStS a In:

of the Crystal

6
1

 

 

57
lost below 85K in pure galvinoxyl because of a phase transition, are

maintained down to 2K in mixed crystals of 6:1 and 4:1 galvinoxyl and
hydrogalvinoxyl 33. They reported Weiss constants of 19, 13, 14, and 13K
in the high temperature region in 19:1, 9:1, 6:1, and 4:1 mixtures without
offering mechanisms for the observed magnetic behaviors. They also
measured the temperature dependence of the EPR intensities of half-field

transitions in those mixed crystals.

For the purpose of comparing magnetic data obtained by others with our
SQUID at Michigan State University , I ran the temperature dependence
measurements on a polycrystalline sample of galvinoxyl from Aldrich
Chemical Co. at 200G and 5006 and found a Weiss constant of 19K without
any problem of phase transitions, presumably due to contamination of some
hydrogalvinoxyl in the crystals. Miller reported a value for J of 8.7 cm-l,
and a Weiss constant 0 of 15.2K for pure galvinoxyl; he also found 0 = 12.6
to 15.7 K for a mixture of crystals in proportion le-x (x=0.85i-0.02).

Miller gave an interesting conclusion in his paper as follows:92

"No apparent change in the crystal structure was detected when the results
of these structures are compared to each other and the earlier reported
structure obtained at room temperature. Attempts to obtain the crystal
structure of hydrogalvinoxyl-free sample were unsuccessful due to
isintegration of the crystals as the samples were cooled through Tmax. This
uggests a major transformation occurs which leads to the loss of integrity

f the crystal; thus, this mysterious phase transition remains unsolved."

 

 

One m«
analogue nitr
butyl-4-oxoc;
contrast to gal
antiferromagr

of a phase trar

It has been sr
are paired up
triPlot separar
Was observer”
Structure of f
instead, it has

gaiViIlOXyl an

4'3-2~ Nitr

it
CXtensively a

nitronyl nitrc

because they

interacfi0n tl

c0nSidered as

 

58
One more related magnetic measurement is that of the structural

analogue nitrogen substituted galvinoxyl radical, 2,6-di-t-butyl-4(3,5-di-t-
butyl-4-oxocyclohexa-2,5-dienylidene amino) phenoxyl (BIP) 34.93 In
contrast to galvinoxyl, the magnetic susceptibility of the BIP radical exhibits
antiferromagnetic behavior with a broad maximum at 54K and no evidence

of a phase transition.
t-Bu t—Bu

/
t-Bu N t-Bu

34
It has been suggested that the unpaired electrons on neighboring molecules
are paired up by exchange interactions to form dimers and the singlet-
triplet separation was estimated to be 60.5 cm'l. No half-field transition
was observed over the accessible temperature range. The single crystal
structure of BIP was reported recently not to be coplanar like galvinoxyl;
instead, it has significant twist between the two aromatic rings so that the

galvinoxyl and amino galvinoxyl are not crystallographically isostructural.

4.3.2. Nitroxyl Radicals

As mentioned earlier, nitronyl nitroxides have been used
extensively as precursors for molecular magnets. Nitroxides, especially
nitronyl nitroxides are good candidates as molecular magnet precursors
because they show significant spin polarization and provide possibilities for
interaction through conjugation. 94 The Nitroxyl functionality may be

:onsidered as an odd electron bonded system with a o-bond and a 3-electron

 

a-bond betwe

small dipole i

For organic
convenient, a
nitrogen atom
the odd electr
the nitrogen
electron arrar
effects which

Part in prever

Oar.

For this ma,

llnPaired elec
While Others

bEIieVed to b

20% ()n the 0}

 

Ifi———=TM~ aw i 2%

59
n-bond between nitrogen and oxygen. Such a structure is supported by the

small dipole moment, the bond length, and the infrared spectrum.95

For organic purposes the resonance hybrid representation is most
convenient, and nitroxides with aryl or conjugated groups attached to the
nitrogen atom are best described by a series of resonance forms in which
the odd electron is delocalized over the conjugated system as well as over
the nitrogen and oxygen atoms. An inherent stability comes from an
electron arrangement around the nitrogen and oxygen atoms, and steric
effects which arise from the groups attached to the nitrogen atom play little

part in preventing dimerization.

630 ._. 030+» OED

For this reason, there is a controversy concerning the exact amount of
unpaired electron density on nitrogen or oxygen. Some claim pN~ 0.3,
while others claim pN ~ 0.9. 96 In the case of TEMPO the electron is
believed to be delocalized approximately 80% on the nitrogen atom and

20% on the oxygen atom.

 

Awaga
dependence c
nitrophenyl)-

35 in several

The temperat
the Curie-We
ttHd Weiss or
respectively,
more rapid 8
Presence of 1
Shtnvn that
t6niaeratrre

The radical ;
interesting pl
N” and 0- I
“Sing an M1
domilnted b:

60
Awaga and Kinoshita reported the temperature dependence and field

dependence of the magnetic susceptibility of the organic free radical, 2-(4-
nitrophenyl) -4 ,4,5 ,5 -tetramethy1—4,4-dihydro-1H—imidazolyl-1-oxy 3 -oxide

35 in several crystalline phases.

0'/N /N+‘o- 0°/N; :N *0-
E NO2
N02

para-NITNPh meta-NI'INPh
3 5

The temperature dependence of the susceptibility of 32 was found to follow
the Curie—Weiss law over a temperature range of 2 to 250K, and the Curie
and Weiss constants are determined to be C=0.375 emuK/mol and 0.9K,
respectively. The magnetization curves of 35 at 2, 2.9, and 4.5K exhibit a
more rapid saturation than that of a S=1/2 spin entity, which supports the
presence of ferromagnetic intermolecular interaction. Furthermore, it was
shown that magnetization comes to saturation more rapidly as the
temperature decreases. The structural isomer, meta nitroxide exhibits

antiferromagnetic intermolecular interactions.

The radical 35 has several crystal phases; the best characterized and most
interesting phase was found to be the two-dimensional network linked by
N+ and 0' reported by Awaga. MO calculations have been carried out
using an MNDO RHF-doublet method and they suggest a mechanism

dominated by the Coulombic attractions. The results of the calculation also

indicated that
side of the 11
ring, wherca:

nitrophenyl r.

They suggest
much Smaller
NLUMO at t
interactions

coupling, T1
to the SOMO

61
indicated that the unpaired electron occupying the SOMO is localized on the

side of the nitronyl nitroxide and has little population in the nitrophenyl
ring, whereas the NHOMO and NLUMO are distributed mainly in the

nitrophenyl ring.

s—e

N+

\O‘uuuuN+ IMHO" N N
II 'I _

IIO.’N / ‘O'IIIIIIIINTIII
II I
O O—

IIIIN+II m'O/N /N+\O
[I'—
O

/N+-O- .......... ’N+"”'O/N /N+\O' """"" I’Nj'lll
ONT o’"-'o Y

They suggested that the intermolecular overlap between SOMOs could be
much smaller than those between SOMO-NHOMO and/or between SOMO-
NLUMO at the contact points in the 2-D network and the intermolecular

IIIIIIII N+IIIIIO”N /N+‘O‘IIII
I ',
o' o-

interactions between SOMOs would mainly result in antiferromagnetic
coupling. Thus, the ferromagnetic interactions of NITNPh were attributed

to the SOMO—NHOMO and/or SOMO-NLUMO interactions.

I“ H N H. W \‘H N H' W
‘ I‘. + a". -.'- + 0“
I ‘0', N N‘ O;- I ‘00, N / N~ 0:..-
I" ' H H“ '-_ H H“.
H .
I. c“. H H.- '- ..-“.‘ H / --.
.00 \ N \ Nt'o‘l-~ i “.9 \ N \ NTO‘x \
" H N H' H ‘ H N

H

The SC
have been evz
radical showr
shown to be
"hydrogen-b0
between the S

frontier orbitz

The ferroma
nitroxide (34
Weiss consta
relationship v
in the cage of
diagram wen

though they d

 

 

62
The SOMO-NHOMO and/or SOMO-NLUMO interaction arguments

have been evaluated again by Awaga in the para-pyridyl nitronyl nitroxide
radical shown above. The para-pyridyl nitronyl nitroxide 36 has been
shown to be a ferromagnetic linear chain with J/k==0.27 K through
"hydrogen-bonding" interactions. There is little overlap in this radical
between the SOMOs, while the overlaps between the SOMO and the other

frontier orbitals are rather large.

N H, _. T r ‘
P. —T
Hi -
.‘ ‘.‘Hh
- \

 

 

 

 

 

 

 

 

 

 

O.\ N AN-wO t n-conjugation
Ti .
A L i‘ - H N 4+

36

The ferromagnetic intermolecular interaction in 3-quinolyl nitronyl
nitroxide (3-QNNN) 37 has been reported by Kinoshita.97 The reported
Weiss constant was 0.27K and the crystal structure was described and the
relationship with the intermolecular interaction is discussed as being similar
to the case of radical 36. The main interactions according to their packing
diagram were again through hydrogen bond between the radicals, even

though they did not take account of the hydrogen bond explicitly.

 

 

4.4. Altern

Advanl
allowed the :
past 30 year:
understood i1
basis of Hun
themselves 1;
different SpaI
{W0 non bom
the ground St
HOMO-wk
electrons int

Singlet will b

The queSlio;
degenerate p

 

4.4. Alternant Hydrocarbon Based Polyradicals

Advances in the observation of transient open—shell species have
allowed the study of numerous types of n-conjugated diradicals over the
past 30 years. Many EPR spectra of n—conjugated diradicals have been

‘ understood in terms of Hund's rule. Electron repulsion is the underlying
basis of Hund's rule. Two non-bonding electrons will prefer to distribute
themselves in different degenerate orbitals, since they will then occupy
different spatial regions and minimize their electrostatic repulsion. If the
two non bonding orbitals are split by a small amount, the triplet can still be
the ground state because of its minimal electron repulsion. However, if the
HOMO-LUMO splitting is large, the energy benefit of placing both
electrons in the HOMO will outweigh the electron repulsion factor, and the

singlet will be the ground state.

The question remains whether all diradicals with two electrons in a

degenerate pair of NBMOs would prefer triplet high spin state. Berson

 

 

proposed that
criterion. Me

triplet ground

There are St
polyradieal s
(EHMO) tret
questionable
there is no

molecule.

In 1936, Hi
coefficients
junction link;
at the HMO

0f Hund's rul

In 1977 and
the concepu
Violate Hum
molecule, th

rePlllsion th;

The In‘Quir
triplet by a
around a be

earlier for S

 

 

 

64
proposed that having such orbitals is neither a necessary nor a sufficient

criterion. Meta-Xylylene and meta-quinomethane have been shown to have

triplet ground states even through they lack degeneracy at the HMO level.

There are several controversies on theoretical treatments in organic
polyradical systems including the Extended Hiickel Molecular Orbital
(EHMO) treatment. Most importantly, the applicability of Hund's rule is
questionable for heteroatom substituted altemant hydrocarbons in which
there is no symmetry-induced degeneracy of NBMOs and even for a

molecule.

In 1936, Huckel pointed out that some hydrocarbons with zero HMO
coefficients at the joint (disjoint) do not follow Hund's rule. Since the
junction links two sites with zero coefficients, the exchange energy vanishes
at the HMO level of approximation, and the normal basis for the application

of Hund's rule vanishes with it.

In 1977 and 1978, Borden and Davidson's theoretical studies strengthened
the conceptual basis for identifying which n-conjugated molecules might
violate Hund's rule. If the NBMOs are confined to separate regions of the
molecule, they are said to be disjoint. This property erases the Coulombic

repulsion that usually destabilizes the singlet in n-conjugated biradicals.

The m-Quinodimethane (m—xylylene) biradical has been predicted to be a
triplet by all levels of theory and the quinodimethyl biradical topology
around a benzene ring (0, m, or p) produces different results as discussed

earlier for Schlenk, Thiele, and Chichibabin's hydrocarbons.

 

There is no st
degenerate no
M0 (HMO) ti
only if the

degeneracy gt

Since these ar
models can b
meaning that
Starf€d~such
Avariety of 5
been derived
COUpIing thl‘t
POWerful para

In recent yea]
been reportel
building bloc]
are based on
Units When th

of all gleam

 

65

There is no stable Kekulé structure from meta-Xylylene. There is a pair of
degenerate non bonding molecular orbitals (NBMOs) at the level of Hiickel
MO (HMO) theory. At higher levels of theory, the NBMOs are degenerate
only if the system has a threefold symmetry axis, although a near
degeneracy generally exists in the absence of such symmetry.

U U U

Since these are planar 1: systems, HMO theory and a variety of other simple
models can be applied. Most of the systems are altemant hydrocarbons,
meaning that the atoms can be divided into two sets—starred and non
starred—such that no two atoms of the same set are connected to each other.
A variety of simple rules concerning orbital energies and coefficients have
been derived for altemant hydrocarbons. The realization that meta
coupling through a benzene ring is always high spin has provided a

powerful paradigm for the design of very high spin molecules.

In recent years, a series of organic molecules with spin as high as S=9 has
been reported, in which meta coupled di-phenyl carbenes provide the
building blocks toward ever higher spin states.98 These delocalized systems
are based on a design which predicts high spin coupling between adjacent
units when the contiguous 7t topology of the molecule precludes the pairing

of all electrons into bonds (non-Kekulé structures, Mataga model).99

4.4.1. Triple

The tt-
heteroatoms a
Berson and l
xylylene and i
tetra radical

semiempirical

m-Benzoquin
high level of
state about 1(
experimental
MNDO-UHF
fixed geomet

311d TCl'clllVe e

Lahti g
boom,
phenyhlllmng

66

4.4.1. Triplet Diradicals

The n-topology of organic diradicals and diradicaloids with
heteroatoms allows one to construct building blocks for organic magnets.
Berson and Lahti reported properties of organic diradicals such as meta-
xylylene and its heteroatom diradicaloid analogues 38; they built the quintet
tetra radical shown below and analyzed its electronic structure using
semiempirical molecular orbital theory. 100

O' 0'

COO .

38

C'

m-Benzoquinodimethane (meta-xylylene) has been treated at a relatively
high level of SCF—MO-CI ab initio theory, by which it has a triplet ground
state about 10 kcal/mol below the singlet state in qualitative agreement with
experimental EPR results.101 Berson and Lahti found the combination of
MNDO-UHF and INDO-CISD (semiempirically parameterized methods at
fixed geometries) computations to be effective in giving both geometries

and relative energies for spin states of diradical and diradicaloid systems.

Lahti and Iwamura independently reported a unique approach to the
design of building blocks for molecular magnets through the linkage of two
phenylnitrene units by n-conjugated linker groups.105,102 These open-shell

n—conjugated systems with the general structure of :N:—Ph-X-Ph-:N: have

been studied '
linkers and co
For 1,l-bis(4
with zero-fie]
0.0029 cm'1
with zfs para]
results suppo:
qualitative co

hydrocarbon

a

N

Dough
between tw<
Combining

triInethylene

radical 41 b)

 

67
been studied by variable temperature EPR with various cross—conjugating

linkers and connectivity types.

For l,l-bis(4-nitrenophenyl)ethene 39, a quintet ground state was found
with zero-field splitting (zfs) parameters ID/hcl = 0.151 cm-1 and IE/hcl =
0.0029 cm'l. For 4,4'-dinitrenobenzophenone 40, a quintet ground state
with zfs parameters of lD/hcl = 0.156 cm'1 and IE/hcl = 0.0046 cm-l. These
results support the idea that minor heteroatom substitution does not reverse
qualitative connectivity-based exchange coupling effects (extended altemant

hydrocarbon model).

Dougherty reported an alternative ferromagnetic coupling mechanism
between two triplets to give the higher spin state quintet (S=2). By
combining derivatives of the classic non-Kekulé diradical,
trimethylenemethane (TMM), he was able to build and study the tetra
radical 41 by EPR.103

“r E1>=§§=<jj

TMM

4.4.2. Perc

As mer
Ballester (Bal
called an im
triphenyl me
properties hat
magnetic pro;
they provide

according to .

For examplt
interactions 1
of acid and
bonding inlet
and .1 to

intefIIIolecu]
add and este

for ~C1, ‘Br.

filnCliOnaliz<

68
4.4.2. Perchlorinated Triphenylmethyl Radicals

As mentioned earlier, the perchlorinated triphenyl methyl radical of
Ballester (Ballester's radical) is a very stable organic radical which has been
called an inert radical.43 A series of functionalized polychlorinated
triphenyl methyl radicals have been synthesized and their magnetic
properties have been analyzed.104 Among the various systems reported, the
magnetic properties of monofunctionalized radicals are interesting because
they provide a collection of possible dimeric magnetic interactions

according to their functional groups.

 

R =H, Br, 1, OH, CH3, COZH. CO2CH3. COZNH4

For example, we can expect to see the effects of hydrogen-bonding
interactions in the groups -C02H, COZNH4. Also, by comparing the cases
of acid and ester, we can estimate the contribution of each hydrogen
bonding interaction involved and compare the differences between -Cl, -Br,
and -l to find any trends to follow in determining the possible
intermolecular interactions. As expected, there are differences between
acid and ester in Weiss constants of 0.9 and -3.8K and 2.1, -3.4, and -4.4 K
for —Cl, -Br, and -l , respectively. The Weiss constant for the COZNH4

functionalized radicals has been reported to be OK, in contrast to most of

them being at
without havir
radical has a
metal salts of

phenomena.“

The -21
studied by X-
has been disc
term of McC.
been assignet
crossed inter;

pentachlorOpl

The PerChlor
fluorenyl 44
10 have Wei

Structural 6X]

69
them being antiferromagnetic. This result is unique and difficult to interpret

without having the single crystal structure. The hydroxy functionalized
radical has a problem of forming quinoidal structures and various alkali
metal salts of this radical have been shown to have interesting association

phenomena. 105

The -2H substituted perchloro triphenylmethyl radical 42 has been -
studied by X-ray analysis and its observed antiferromagnetic susceptibility
has been discussed and related to molecular packing and spin densities in
term of McConnell's theory.106 The interaction sites between radicals have
been assigned to be the two para—carbons and two meta-chlorines through a
crossed interaction pattern with closest distances of 3.64 A between two

pentachlorophenyl groups.

 

AI' = C5C15 42

The perchloro-9—pheny1fluoreny1 43 and dodecachloro-3-methoxy—9-phenyl
fluorenyl 44 radicals shown below have been studied by EPR and reported
to have Weiss constants 0 of —l4.5 and —l7.5K, respectively, without any

structural explanation of the magnetic interactions.

 

C1.

Triple
reported by ‘
strategies, 1
meta-xylylen
stable solids
and enantio:
reported zfs .
new and ID,
suggest Sign

although it is

C1

C1 ‘

Magnetic 8

diastereoiS or

 

 

Triplet biradical homologues of Ballester's radical have been
reported by Veciana and Riera using n-conjugated meta-xylylene diradical
strategies. 107 The biradical 45 and its corresponding monoradical of
meta-xylylene diradical derivatives have been synthesized and isolated as

stable solids at ambient condition. The biradical forms a mixture of mesa

 

and enantiomeric dl forms which can be differentiated by EPR. The
reported zfs parameters were lD/hcl = 0.0152 and lE/hcl = 0.0051 cm-1 for
meso and lD/hcl = 0.0085 and lE/hcl S 0.003 cm“1 for the d] isomer which
suggest significantly stronger coupling in the mesa form than in the dl,

although it is not clear why this should be true.

 

mesa symmetry 4 5 one of d,[ symmetry

iVlagnetic susceptibility and magnetization measurements of the

iiastereoisomers in solid state show "quasi-ideal S = 1 paramagnetic

 

 

behavior" dos
multiplicities
nmnno
interactions.

and 46) shor

biradical.

Althou
Spin organi
dimensional
mechanism :
radical systei
uIlderstandil

based 0n t

0Vehiflnikos

A seri
high Spin 1T.

methyl rat

peiyradical:

 

 

 

71
behavior" down to 4.2K. They suggest that the prediction of ground state

multiplicities in any non-disjoint altemant hydrocarbon are true regardless
of the lack of planarity and changes in the symmetries in intramolecular
interactions. They extended these strategies to the tri- and tetra radicals (45
and 46) shown below, with results consistent with the findings for the

biradical.

 

Although the above approach is successful for construction of high
spin organic molecules, it is inappropriate for assembly of three
dimensional structures which would show ferromagnetism unless there is a
mechanism for intermolecular interaction. Most studies of reported poly
radical systems have concentrated on building molecules themselves without
understanding the nature of their intermolecular interactions. They are
based on the proposed organic magnet mechanism of Mataga and

Ovchinnikov which were discussed earlier.

A series of polyradicals has been assembled by Rajca108 in building
high spin molecules using similar strategies. Instead of using Ballester's
radical, he employs homologous derivatives of t-butyl substituted triphenyl
methyl radicals. Although there is little structural data on these

polyradicals, the crystal structure and magnetic property of the related

 

 

tris(3,5-di-t-b1
expected, tris(
intermolecula:
with S=l/2.

Rajca was al
above diradit
dependences
measurement
REi-PIO) ac(
without 3110‘
BleaneY‘Bov
dimeric inte

Pebtadicals

4.4.3. P01.

IWalm
is Contmue

Incorporatet

72
tris(3,5—di—t-butylphenyl)methyl radical has been reported by Kahr.109 As

expected, tris(3,5-di-t-butylphenyl)methyl radical does not show significant
intermolecular magnetic interaction and behaves as a simple paramagnet
with S=l/2.

 

47

Rajca was able to correlate the sign of the exchange interactions in the
above diradicals 47 in the solid state. He reported temperature and field
dependences of these diradicals using SQUID measurements. The
measurements showed anti-ferro or ferromagnetic interactions (when R=H,
R'=i-Pro) according to interactions which he claimed to be intermolecular
without showing the mechanism. He was unsuccessful in applying the
Bleaney-Bowers equation to their data which means that it is not a simple
dimeric interaction. More recently, he reported an extension of these

polyradicals to hepta- and decaradicals with S=7/2 and 8:5, respectively.

4.4.3. Polyphenyl Carbenes

Iwamura reported a novel dicarbene system in which spin multiplicity

is controlled by the overlap modes of spin-distributed benzene rings

incorporated in the rigid [2,2] paracyclophane framework.110 According

 

to McConnell
interaction be
product of sp
negative in si

organic radicz

Among three
the pseudo-c
Spin density
Spill distribu
°YC10phanes,
0f the dicarl

the tt-electro

Iwamura ant
series of hig
pelicarbene

known (dete

 

73
to McConnell's theory on intermolecular magnetic interactions, exchange

interaction between two aromatic-radicals can be ferromagnetic when the
product of spin densities at two interacting sites on different molecules is
negative in sign, since the exchange integral is usually negative between

organic radicals at the van der Waals contact distance .

 

 

 

 

Ph .. " —Ph
+ — - \- —- + + — -/
_/ \+ +/ \_ _/ \+
\ / \ / \ /
E + _ -\ j E - - + j E + — - j
' '— + 00 —Ph ‘ _ + ' — +
+/ \_ __| +/ \_ +/ \_
\ / \ / \ /
/'-+ /‘_+ /'—+
Ph .. ph .. ph ..
pseudo—ortho pseudometa pseudo-para

Among three isomers of bis(phenylrnethylenyl)[2.2]para-cyclophanes only
the pseudo-ortho and pseudo—para isomers are expected to have negative
spin density at each interacting site between the two benzene rings and the
spin distribution is expected to result in quintet ground states for these
cyclophanes, as confirmed by EPR. These differences in spin multiplicity
of the dicarbenes are the first experimental examples of overlap control of

the n-electron magnetic interactions in layered aromatic radicals.

Iwamura and Itoh extended the meta—xylylene n-conjugation topology into a
series of high spin poly carbenes of 8:4 and Iwamura reported 2-D extended
polycarbenes 48 of 8:9 which is highest spin molecular system currently

known (determined by saturation magnetization experiments).111

 

0.

lwamura alsc
the expected
obeyed a Cut

1

It is clear th;
from alterna
intramolecul

available, 11

 

 

 

74

Iwamura also reported the dinitrenes (49 and 50) shown below which show

 

 

 

the expected high spin multiplicities (quintet) and magnetic behaviors which

obeyed a Curie law in the temperature range 12-85 K.112
H = H = =
1}! 49 13 so

It is clear that successful strategies to design magnetic molecular materials
from alternate hydrocarbons will require a far more precise knowledge of

intramolecular and intermolecular exchange interactions than is presently
available. 113

 

 

 

5. Molecul:

As we
present in the
different fror
radicals surr
depend on tl
radicals thr01
either a spin
said to be 2
interactions 1

involved (3-f

The nature a
by carefully
311d by the s
radicals and

be utilized f

The design
in a geome

collpler.

4

ferremElgne
based on til

SYStem,

 

 

75
5. Molecular Magnets by Macromolecular Chemistry

As we have seen in many examples, when two organic radicals are
present in the same molecular entity, the magnetic properties can be quite
different from the sum of the magnetic properties of two such mono
radicals surrounded by their nearest neighbors. These new properties
depend on the nature and the magnitude of the interaction between the
radicals through the coupling unit. The molecular state of lowest energy is
either a spin singlet or a spin triplet. In the former case, the interaction is
said to be antiferromagnetic, in the latter case ferromagnetic. These
interactions are often termed superexchange because of the large distances

involved (3-5 A) between the radical centers.114

The nature and the order of magnitude of the interaction can be engineered
by carefully choosing the interacting organic radicals and the coupling unit
and by the symmetry and the delocalization of the orbitals centered on the
radicals and occupied by the unpaired electrons. The same strategy could

be utilized for designing molecular ferromagnets.

The design of extended magnetic structures requires the alignment of spins
in a geometrically well defined environment, where the prOper coupling
unit not only plays a role as a physical linkage but also as a ferromagnetic
coupler. As noted by Dougherty, the design of an extended molecular
ferromagnet can be naively envisioned by the following simple diagram

based on the spin sources and the magnetic coupling units in a linear chain

system.

 

@

There are sev
system: Tl.
characterized
The couplir
themselves, 5
Spin sources
to form a 11'

molecule inc

We have sel
a11011161211 ca
0f building

through Spa
0Tganic di:
interactions
induce throl
Crown ether
these diIneri
Where the

maemlmiu

 

 

76

Coupling UnCoupling Unit iNW
WCoupling Unitr~§~WEouphng [WCoupling UM

There are several points that we should address in building such a magnetic ‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

system: The spin sources should be stable and magnetically well
characterized, and they might form a linear chain with the coupling units.
The coupling units, which could be diamagnetic or paramagnetic
themselves, should be able to couple spin sources ferromagnetically. The
spin sources could be any stable organic radical with proper functionalities
to form a linear chain, and the coupling units could be any atom or

molecule including stable organic radical and poly radical.

We have selected tris(2,6-dimethoxyphenyl)methyl radical 1 as a spin base
and metal cations as the magnetic coupling unit. I will introduce our design
of building organic magnets followed by examples of through bond and
through space magnetic interactions in organic radicals: examples of
organic diradicals showing long-range (superexchange) magnetic
interactions through o-bonds; examples of stable organic radicals used to
induce through-space magnetic interactions by ion—binding of spin labeled
crown ethers. Our design of molecular magnets will be given by extending
these dimeric interactions in organic radicals into a linear chain and further

where the geometrical control and magnetic coupling are achieved by

macromolecular chemistry.

 

5.1. Long-r

As me
the first org:
were studied
been examin
bond and t1
suggested in
ferro- or an
through-bon
general pher
transfer prc
through-spar
magnetic in

very useful 1

In 19.
MOIllellgate
below,115
System Sho
diradicals 1i
Orbital on t
for the OXyI
Was greater
1’ 2* 3, 4) t

 

77
5.1. Long-range Magnetic Interactions in Organic Radicals

As mentioned earlier, Chichibabin and Schlenk hydrocarbons were
the first organic diradicals in which intramolecular magnetic interactions
were studied. Since then, numerous examples of polyradical systems have
been examined by many renowned chemists. The orbital nature of through-
bond and through-space interactions for magnetic coupling has been
suggested initially by Hoffmann and many examples of diradicals showing
ferro- or antiferro-magnetic interactions have been reported. Still, the
through-bond and through-space magnetic interactions are not accepted as
general phenomena in organic radicals even though electron and energy
transfer processes are often analyzed in terms of through-bond and
through-space interactions. It is my intention to show here that long-range
magnetic interactions in organic radicals are not only possible but can be

very useful tools in building molecular magnets.

In 1956, Sloan and Vaughan reported stable organic biradicals with
rt-conjugated and nonconjugated magnetic coupling linkers such as shown
below.115 They found that the unpaired electrons in the oxygen-linked
system shown below are less insulated from one another than in the
diradicals linked by simple aliphatic chains because of the existence of a it-
orbital on the oxygen. To their surprise, they obtained an EPR spectrum
for the oxygen linked diradical 52 with a g -value of 2.003li0.0004 which
was greater than the 200253200004 for the aliphatic linked diradicals (n:
1, 2, 3, 4) 51.

 

 

o .
O

The EPR spe
the relative
coupling cor
interact witl
situations: T
lines with se
Jis much so
if each elect:
constant) is:
separation 0
highly sens
(hYPerfine c

In general, -
or held rigi
three iines 2
“9 riaid fr

dinilloxides

 

 

?0

. O O . O

O O

52

The EPR spectrum obtained from a biradical in solution depends largely on
the relative magnitudes of the exchange interaction and the hyperfine
coupling constants. For a bisnitroxyl in which the two unpaired electrons
interact with two equivalent nitrogen atoms, there are three possible
situations: The biradical may behave like two mono radicals and show three
lines with separation of a (hyperfine constant) when the exchange coupling
J is much smaller than the hyperfine coupling. The biradical may behave as
if each electron spends half of its time on each nitrogen when a (hyperfine
constant) is much smaller than J and five line spectra are then obtained with
separation of a (hyperfine constant)/2. More complicated spectra which are

highly sensitive to solvent and temperature are obtained when J and a

(hyperfine constant) are about the same in magnitude .

In general, when the two nitroxide groups in nitroxyl radicals are far apart
or held rigidly apart then the exchange interaction tends to be small and
three lines are observed, while if the linking group is neither especially long
nor rigid five lines or more appear in the spectrum. There are several

dinitroxides which belong to latter category.116

 

The EPR sp
150 °C. Sin
of biradicals

from interra.

 

79

0 q
o N—O‘
ouN \N/fl\o
b O
5 3
o
\w/OJKO/QLO.
o
OnN
b O
5 4

The EPR spectrum of diradicals 53 and 54 show five line spectra at 140-
150 °C. Similarly, as the number of methylene groups increase in the series
of biradicals 55 the intensities of the additional spectral lines which come

from interradical interactions in the corresponding spectra decrease.

o 11 o
ouN N—O'
o o

n=2t08
55

The biradicals shown below all gave spectra with five lines in which aN =
7.4 gauss, which is half of that in an analogous mononitroxide,

corresponding to the situation of J » a.

O. -NHN-O. 0. -Nb N_NfiN— O.

 

Ullma:
geometrical t
Bulk magne
which meas
reference pe
By compari
showed "75
Ulhnan clai
spinsrrs

r
\

/

Very recen
by Rey,

measureme
U11man's bi
1 Via the B1
Laht‘

1 mpg
polyradica

 

 

Ullman reported a conjugated nitronyl-nitroxide biradical 56 with
geometrical evidence which showed particularly strong dipolar interactions.
Bulk magnetic susceptibility measurements by Evans' NMR method117
which measures the amount of paramagnetic shifts of NMR peaks to a
reference peak such as TMS gave 49 % triplet character for the radical 56.
By comparison, the unconjugated trimethylene-bis-nitronyl nitroxide 57
showed "75% triplet character" with a half field transition spectrum.

Ulhnan claimed that the value correspond to a weakly interacting free

spins.118
O' O‘ O
I \ /
>-< 11%in
N N N
\ I \
O' O’ 0
Ullman's biradical n = l or 20
5 6 5 7

Very recently Ulhnan's biradical 56 and its derivatives have been studied
by Rey. He reported X-ray structures and magnetic susceptibility
measurements showing antiferromagnetic intramolecular interaction for the
Ullman’s biradical. The singlet-triplet energy gap estimated was as -311 cm-

1 via the Bleaney-Bowers equation.

Lahti reported computational modeling studies of various n-conjugated

polyradicals using the AMI semiempirical molecular orbital method with

 

 

configuration
altemant hyd:
polyradical n
variety of cor
groups as in

Nitrene.1 19

5.2. Thro

Complexes

A ser
constructed
nitroxyl, an
radicals as;
use of Spin
SuIIlmarize:
complexes
interactions
E100mplex
binding pr
eempiex wr
below, If fl
centers, tilt
sPeeifica1

ferr(magn:

81
configuration interactions which give theoretical basis for the heteroatom

altemant hydrocarbons. He has been interested in synthesis of n-conjugated
polyradical model systems based upon coupling of phenoxyl radicals by a
variety of connecting spacer groups, such as heteroatoms and olefinic spacer
groups as in the general model of Phenoxy-X-Phenoxy and Nitrene-X-

Nitrene. 1 19

5.2. Through Space Magnetic Interactions in Organic Radical

Complexes

A series of spin labeled crown ethers (lS-crown-S) have been
constructed by Mukai using stable organic radicals such as galvinoxyl,
nitroxyl, and verdazyl (58, 59, 60, and 61) with the goal of using these
radicals as probes for biological systems.120 There are good reviews on the
use of spin labeled organic molecules for a biological probes and I will
summarize mainly the systems with magnetic interactions. The alkali metal
complexes of these radical crown ethers show interesting magnetic
interactions in solution EPR .121 When the spin labeled crown ethers form
a complex with alkali metal salts in solution according to the general ion-
binding properties of crown ethers, lS-crown—S can only form a 1:1
complex with sodium salts but a 2:1 complex with potassium salts; as shown
below. If there are significant magnetic interactions between the two radical
centers, the EPR spectra of 2:1 complexes would reflect the interactions;
specifically, a triplet spectrum would result when they are

ferromagnetically coupled.

warm

Triplet spec
galvinoxyl
observed in
system whit
instead of p
characterist:
verdazyl rar
triplet sper
complexes
systems wi
potential it
great exten
\

0'-N

I

H

 

 

 

Triplet spectra have been reported for 2:1 complexes of nitroxyl 58,
galvinoxyl 59, and phenoxyl radicals 60. Half field transitions were
observed in all complexes except for the verdazyl labeled 15-crown-5
system which was reported as triplet coupled when sodium salts were used
instead of potassium salts which is in contrast to the general ion-binding
characteristics of lS-crown-S. The question remains whether the triphenyl
verdazyl radical 61 itself forms a 2:1 complex with sodium cation to give a
triplet spectrum. Furthermore, there is the possibility of forming
complexes with different conformations (syn or anti) even in the above
systems with triplet EPR spectra. Finally, the role of the oxy radicals as

potential ion-binding sites for alkali metals has not been explored to any

great extent.

00—b0 t—Bu

w on ob&oo

O 5 8 O 003 t-Bu 5 9 O 003
K/Ov’ K/Od

O' t-Bu

 

 

 

5.3. Desigr

The t1
traditionally '
it has been r
ions in bimet
interaction,
overlap inte‘

orbitals.

More recent
such analyse
Species, witl
earlier. Int
SUperexchar
were made
metal cente
01 tt-electro
ttflk betwee
energies be
extended E
btlnetaflic r
twe‘eiectrc
of the calc

observe in

 

 

83
5.3. Design of Molecular Magnets by Ion-Binding

The theoretical interpretation of superexchange interactions has
traditionally been based on ideas developed for infinite solid lattices. Since
it has been realized empirically that the bridging atoms between the metal
ions in bimetallic systems determine the sign and magnitude of the exchange
interaction, these qualitative treatments focus on the various types of -
overlap interactions between the ligand atomic orbitals and the metal d

orbitals.

More recently there have been theoretical treatments which seek to extend
such analyses to the cases involving molecular, rather than atomic bridging
species, with special interest in molecular bimetallic complexes as discussed
earlier. In this context a broader theoretical framework for the analysis of
superexchange interactions has been proposed by Hoffmann.122 Attempts
were made to show a connection between antiferromagnetically coupled
metal centers and the phenomenon of through bond coupling of lone pairs
or rt-electron systems in organic molecules. 123 Hoffmann established the
link between antiferromagnetic exchange interactions and the difference in
energies between degenerate MO's using the orbital energies obtained from
extended Hiickel calculations (the simplest all valence-electron model) in
bimetallic complexes. Although these calculations do not explicitly include
two-electron interactions in organic radicals, he suggested that the behavior
of the calculated orbital energies are expected to reflect what one would

observe in more sophisticated calculations.

 

The effect r
exchange in
geometricalr
bond angle r
same symme
the metal orl

earlier in bir

Photoelectrr
pair and re.
interactions
complexes

affects the

these Splitt
metal-ligan

respond to

 

 

84
The effect of the metal-single-atom-ligand-metal bridge angle on the

exchange interaction has been studied extensively by Kahn.124 Such
geometrical considerations lead one to expect a large AF coupling for a 180°
bond angle when the metal orbitals can interact with a ligand orbital of the
same symmetry, and a ferromagnetic coupling for a 90° bond angle when
the metal orbitals are interacting through orthogonal ligand orbitals, as seen

earlier in bimetallic systems.

Photoelectron spectroscopy has provided evidence for the splitting of lone
pair and n-levels as a consequence of through-space or through-bond
interactions such as shown below. The bridging groups in dimetal
complexes provide orbitals of a certain symmetry type, and this in turn
affects the splitting of the metal orbitals. The interaction which leads to
these splittings has a strong conformational dependence (180o angle for
metal-ligand-metal) and metal centers coordinated to such systems should

respond to the energy splitting by showing sizable magnetic coupling.

_.1_N/ \N_1_ _.L_N// \\N_1_
| H | | \.=/ |

 

 

 

 

 

Along the sa
possible mag
the electron
between org

substituents i

As (1
Organic free

carbon, the
bulk of the
ion binding
can serve
dtlnethoxy
Size betWe

Orientation

The Spatia

dietate int

 

 

85
Along the same line we might expect the systems shown below to have a

possible magnetic coupling via bridging diamagnetic metals. Furthermore,
the electron density on the metal may affect the degree of interaction
between organic radicals as we have seen in reviewing the effect of

substituents on bridging ligands in dimetal complexes.

 

As discussed earlier, tris(2,6-dimethoxyphenyl)methyl 1 is a stable
organic free radical and in its D3 propeller conformation around the central
carbon, the unpaired electron is protected from above and below by the
bulk of the six methoxy groups.125 The principles developed in selective
ion binding studies of polyether ligands suggest that these methoxy groups
can serve as binding sites for metal cations. Two of the tris(2,6-
dimethoxyphenyl)methyl radicals can "sandwich" a cation of appropriate
size between them in a distorted octahedral pocket, fixing their relative

orientation.

The spatial relationships in such controlled radical aggregation can, in turn,

dictate intermolecular electron coupling126,127 in "interrupted 0- bonds",

 

 

 

radical pairs
metal cation
would have

orbitals inter
with Kahn's

"The orbitai
easier to ac
respects the
up-down ty]
space intera
weak excep

each other.‘

We are wo
higher olig
enforced b-
might acco
cations to
them para
eXelusivel

WOIklng n

 

86
radical pairs or oligomers in which electron interactions are mediated by

metal cations. Extended stacks formed by this complexation mechanism
would have at their cores a linear array of one-electron carbon-centered p-
orbitals interacting through metal ions.128 This idea happens to coincide

with Kahn's recent suggestions for designing molecular magnets.129

"The orbital approaches based on the spin polarization effect might be
easier to achieve. This approach is particularly attractive. Indeed, it
respects the strong tendency of nature to favor local spin interactions of the
up-down type." " To conclude, we would also like to stress that the through-
space interactions on which we have focused in this Account are generally
weak except when p atomic orbitals belonging to adjacent molecules point to

each other."

We are working to exert control over electron coupling in radical pairs or
higher oligomers designed so that electron interactions are mediated and
enforced by metal cations with varying electron densities. Various metals
might accomplish this task, from diamagnetic alkali metal and alkaline earth
cations to those of transition metals in various oxidation states, some of
them paramagnets in their own right. Our efforts have focused almost

exclusively on diamagnetic salts. The following scheme illustrates our

working model:

 

 

 

 

 

 

X=E

Figur
assoc

Figr
M01
radii

Figr
eltgomers
Stttletures
trtPOd bir.
radical dir

 

nX'

    

O O mung-mm!

   

 

 

nX'

M = Li, Na etc. 2 (RMR dimer) 3 (RMn oligomer)
X = BF4‘. BPh4', 1', C104‘, SCN'

Figure 7. Schematic illustration of the dimeric and oligomeric
association of tris(2,6-dimethoxyphenylmethyl) 1 with M+ ions.

 

 

Figure 8. Optimized structure of an RMR dimer 2 calculated by the
Molecular Mechanics methods in PCModel. Here the individual
radicals adopt a homochiral, staggered conformation.

 

Figure 7 portrays these "interrupted o-bonds" in radical pairs 2 or
oligomers 3. CPK models, molecular mechanics calculations, and X-ray
structures (vide infra) of related species support the metal ion-radical ether
tripod binding. Figure 8 displays an illustration of a radical-metal ion-

radical dimeric such as 2 as calculated by molecular mechanics.

 

 

The a
magnetic p:
paramagneti
straightforv
rings and ir
paramagnet
stoichiomet
2:1, or 1:2
examples c
molecule r

possibility

Thee
Structure c
electrons 1
mechanic:
electron.
Of structu
Organic 1
aPillicatio
tor develr
layer a p
model sy;

gaVe [he

88

The above strategy targets rationally engineered materials whose
magnetic properties result from collective interactions between simple
paramagnetic centers. The basic structure of radical 1 allows for a range of
straightforward modifications at the meta and para-positions of the aryl
rings and in the group bound to the ether oxygen. By careful assembly of
paramagnetic radical ligands, metal ions, counterions, and their respective
stoichiometries, a variety of novel complexes maybe accessible. Simple 1:1,
2:1, or 1:2 complexes between metal ions and 1 in solution are monomeric
examples of the desired types. Ultimately, extended chains in which each
molecule of 1 is coordinated to two metal ions and vice versa offer the

possibility of a designed molecular solid with unique magnetic properties.

These would be fundamentally new materials, whose composition and
structure can be easily to vary their bulk properties. With their unpaired
electrons residing on light atom centers, organic free radicals are quantum
mechanically simple with S=l/2 and g values close to that of the free
electron. Our work may therefore shed light on the fundamental interplay
of structure and magnetic character in materials.130 Since the problem of
organic magnetism is essentially a problem in electronic structure,
application of ab initio and semiempirical computational methods is essential
for developing a detailed understanding of the electronic interactions that
favor a positive exchange integral. Preliminary ab initio calculations on a

model system CH3Li+CH3 have been conducted by Professor Jackson which

gave the triplet as a ground state.

 

 

 

 

The X-ray s
in building
ligand assoc
Visible spec
radical 1 W]
are though
interaction
synthesized
We have pu
of our syst
SQUID me
markedly
susceptibii

aniferroma

Detailed r
Single Cry:
Structure 2

robust ma

1 Will pm:
the Stmc
Propertiei
terms of
interactio
at introdr

coildllsk

 

 

89
The X-ray structures of related monomers provided structural information

in building ligands for better ion-binding. Determination of the metal-
ligand association processes in solutions are evidenced by NMR, EPR, UV-
Visible spectrophotometry and cyclic voltammetry. The redox properties of
radical 1 which can be adjusted by varying substituents of the aryl rings,
are thought to be important in superexchange-mediated magnetic
interactions. Thus, various substituted-triarylmethyl radicals were
synthesized and studied electrochemically to determine the redox properties.
We have pursued SQUID and EPR studies to examine the magnetic behavior
of our systems as a function of temperature and magnetic field strength.
SQUID measurements on various complexes of triarylrnethyl radicals show
markedly nonlinear Curie-Weiss behavior and high paramagnetic

susceptibilities, while polycrystalline powders of l-CdClz shows

aniferromagnetic behavior.

Detailed magnetic characterization of well-defined materials, ideally as
single crystals, will enhance our understanding of the relationships between
structure and odd-electron coupling. We hope these efforts will lead to a

robust materials with controllable magnetic properties.

I will present the results and discussion in the following order: i) building a
the structural basis for complex formation; ii) studies on electronic
properties of substituted radicals; iii) evidence for complex formation in
terms of ion—bindings; iv) theoretical basis for pairwise magnetic
interactions; v) magnetic behavior of radicals and complexes; vi) attempts

at introduce magnetic cross couplers in building extended structures; vii)

conclusions and suggestions for future studies.

 

Illiuiilu 1”

 

CHAPTER

1. Struct

Unde
is critical ti
materials (1
properties.
prediction
analyzing ‘
determiner

molecular

The effect
triphenyln
estimated
from HP]
carbon. T
at the 1:
MeCOHne
for all a

gaussBl

 

 

90

RESULTS AND DISCUSSION

CHAPTER I. Triarylmethyl Radicals

1. Structure of Triarylmethyl Radical

Understanding the structures of organic compounds in the solid state
is critical to the engineering of physical properties. The search for organic
materials demands understanding of the forces that determine structure and
properties. Such understanding should lead to improved methods for the
prediction and design of organic crystal structures. The major challenge in
analyzing the structure of molecular organic solids is that the structures are
determined by summation of the many contributing weak intra and inter

molecular forces.

The effects of ortho or meta substituents on the twist angles of the rings in
triphenylmethyl radicals, and therefore on their stability, have been usually
estimated by comparison of the experimental hyperfine coupling constants
from EPR spectra with theoretically calculated Spin densities on para
carbon. The comparison can be achieved by the calculation of spin density
at the para carbon from the hyperfine coupling constants using
McConnell's equation, where Q is a factor that is approximately constant

for all aromatic hydrocarbons and has an empirical value of -22.5

gauss.131

 

 

 

The twist at
between the
where [3 is t
aromatic rir
rings and t
densities ar
from hyper
only 1H hy

A systema
electronic
X-ray stru
while crys
ions have
Of the ch
cempared

by Veciar

To
tris(2,6_d
trustrater
02 toy
crystal g
settletura
dttfractii

SttttCturz

91

The twist angle is introduced by using B cos 0 for the resonance integral
between the central carbon atom and the adjacent carbon atom in each ring,
where B is the resonance integral between neighboring carbon atoms in in
aromatic ring and 0 is the angle between the plane of one of the phenyl
rings and the reference plane of the central carbon and its bonds. Spin
densities are then obtained as functions of 0. But, these calculated angles
from hyperfine coupling constants are not reliable enough, at least when

only 1H hyperfine coupling constants are compared.

A systematic study of the influence of substitution on geometries and
electronic structures of chlorinated triphenylmethyl radicals based on their
X-ray structures has been started by Veciana and coworkers in 1987132
while crystal structure determinations of many triphenylmethyl carbonium
ions have been reported since 1965.133 The hyperfine coupling constants
of the chlorinated triphenylmethyl radicals have been calculated and
compared with those obtained in isotropic solution using INDO calculations

by Veciana and coworkers.

To date we have not obtained single crystals of metal complexes of
tris(2,6-dimethoxyphenyl)methyl radical 1. Our efforts have in part been
frustrated by the fact that Li+ catalyzes the oxidization of the radicals by
02 to yield the cation and the persistent cation in solution hampers the
crystal growth process which I will discuss later. Consequently, our
structural studies have focused on characterizations of the binding site via

diffraction studies of the uncomplexed monomeric species and some close

structural analogues and pulsed EPR studies.

 

 

1.1. Strut

Tris(2
is a remark
traced main
rings whicl
six ortho r
coupling Cr
rings in t
comparing
triphenylr
nitropheny
Martin est
dimethox
triphenyln
1. He p

Substanti:
SUbstitutir
delocaliza
the pmpr
methoxy

We were
SOiution j
allgles 01

“Sing Var

 

92

1.1. Structure of Triarylmethyl Radical

Tris(2,6-dimethoxyphenyl)methyl 1, originally reported by Martin,
is a remarkably stable organic free radical. The high stability has been
traced mainly to the shielding of the trivalent carbon atom by the phenyl
rings which are strongly twisted in a propellerlike conformation and the
six ortho methoxy groups. Martin correlated observed EPR hyperfine
coupling constants of the radical 1 with the twist angles of the aromatic
rings in tris(2,6~dimethoxyphenyl)methyl radical in solution. By
comparing the calculated twist angle of 30° for Gomberg's
triphenylmethyl (aHpara=2.77), the X-ray structure of tris(4-
nitrophenyl)methyl (0=30°), and a Hiickel Molecular Orbital calculation,
Martin estimated the increase in angle of twist to be 17° for tris(2,6-

dimethoxy-phenyl)methyl (aHpara=2.26) over that of Gomberg's

triphenylmethyl resulting in an estimated twist angle of 47° for the radical
1. He pointed out that the reduction in apara from 2.77 to 2.26 is
substantially larger than would be expected for simple methoxy
substitutions in triphenyhnethyl radicals and that the decrease in spin
delocalization should therefore be attributed to the increase in the pitch of
the propeller conformation of the radical 1 induced by the six ortho

methoxy groups.

We were able to directly compare the solution and solid structures by using
solution EPR and the X-ray structure of the radical 1. The estimated twist
angles of the radical 1 in solution were nearly identical to Martin's values

using various experimental conditions and solvents such as THF or diethyl

 

 

 

ether (solut
initially by
Michigan S
were obtair
radicals adr
twisted out
two rings a
the way I
pathwayt
unpreceder

triaryhnetl

 

Increased
SUbstituen
temPeratu:
Sterically ,
tutti mini
kctti/mol t
We belies
stable so]

 

 

93
ether (solution EPR). The X-ray structure of radical 1 was determined

initially by Professor Kahr at Purdue University and repeated here at
Michigan State University later. Brick-red monoclinic prisms of radical 1
were obtained by fast evaporation of an ether solution under argon. The
radicals adopt an unusual conformation in the solid state. One aryl ring is
twisted out of the central methyl carbon plane by only 12° while the other
two rings are twisted by 61°. This structure represents a point well along
the way to the transition state for the two-ring flip racemization
pathway134 and its large deviation from the D 3 ground state is
unprecedented for triaryl-X propellers, especially in a perorthosubstituted

triarylrnethyl radical.

 

 

 

Increased racemization barriers are consistent with large ortho-
substituents; perchlorotriphenylrnethyl does not racemize on the room
temperature time scale.135 However, methoxy groups are substantially less
sterically demanding than chlorine atoms, and AMl calculations on 1 at the
fully minimized and X-ray twist angles show an energy cost of less than 3
kcal/mol for this distortion.

We believe that the X-ray structure of the radical 1 represent a kinetically
stable solid state structure which may be far from the relaxed solution

conformation. Because of the dissymmetric twists of the rings, the ether

 

 

tripod bindi

radical.

The 1
average dis
But, as note

of energies

Figr

Small moi

or Other ]
results in
the chara
been con
Removal

radical.

Pr
the Li+

methoxy

 

94
tripod binding site is poorly represented by the crystal structure of the

radical.

The crystal structure of the radical l is shown in Figure 9. The
average distance between the oxygen atoms of a tripod pocket is 3.7 A.
But, as noted above, this distance can vary substantially over a small range

of energies as a function of the three aryl rings' rotations.

 

Figure 9. Stereo view of the X-ray structure of the radical 1.

In pure solution, radical 1 is air stable; this suggests that even such a
small molecule as 02 is sterically excluded. However, addition of LiBF4
or other lithium salts to air-saturated THF or ether solutions of radical 1
results in nearly instantaneous oxidation of the red radical solution to give
the characteristic purple color of the triarylrnethyl cation. This process has
been confirmed by Jackson and Kahr to be inhibited by excess l2-crown-4.
Removal of air and reduction with acidic Cr(II) salts regenerates the

radical.

Professor Jackson proposed several possible interpretations. First,
the Li+ ion may bind, as shown in the Scheme below, to the tripod of

methoxyl oxygens in radical l, inducing a conformational change which

 

 

 

opens up a
binding site
02 itself, :
molecule. i
simply acti‘
02 radical:

 

Asr

conceminp

the sensiti
1‘2. Str

A r
cryStal st
deviates
PrOpefler
and 3101
(MMpz)

 

 

95
opens up access for 02 to the radical center. Alternatively the fourth

binding site on the (normally tetrahedrally coordinated) Li+ ion may bind
02 itself, serving as a bridge between the radical center and the 02
molecule. Finally, the presence of the Li't' ion in an organic solvent may

simply activate electron transfer (ET) by strongly stabilizing the product
02 radical anion via ion pairing.

 

 

 

 

As mentioned above, our attempts to obtain structural information
concerning complexes of radical 1 with metal salts have been hampered by

the sensitive redox chemistry of this system.

1.2. Structure of the Triarylmethyl Cation

A more realistic picture of the binding site than 1 is provided by the
crystal structure of the related cation tetrafluoroborate salt 62 which
deviates from D3 symmetry in the lattice but can still be described as a
propeller. The comparison of the X-ray structure (determined by Kahr
and Blount at Princeton University) of the cation 62 and a calculated

(MMP2) structure of the radical 1 is shown in Figure 10.

 

 

 

As a consec
groups are
pockets wit?

cavity.

 

Figr
state
(fror

13. Stru

Sim
avoid son
circumve
ettmplexa
dimethox
boron rep
tettowinp
handled i

Structure

 

96
As a consequence of the propeller conformations of l and 62 the methoxy

groups are arranged in such a way that they form a pair of nucleophilic
pockets with the oxygens' lone pairs projecting toward the center of a small

cavity.

 

Figure 10. Space-filling views of the binding site: MMP2 ground
state conformation of radical 1 (left) and X-ray structure of 62
(from 62 ~BF4, right).

1.3. Structure of Triarylmethyl Borane

Since small metal ions catalyze oxidation of 1 by 02, it is difficult to

avoid some oxidation in all our preparations of complexes. Wishing to
circumvent these difficulties while still gaining structural data on
complexation, we have synthesized the diamagnetic compound tris(2,6-
dimethoxyphenyl)methyl borane 63, a structural analogue to 1 in which
boron replaces the central carbon atom. This compound is easily prepared
following a modified Martin's procedure as shown below; it is easily
handled in the absence of Brbnsted acids, it crystallizes easily, and its X-ray

structure is shown in Figure 11, along with an MMP2 calculated structure

 

 

of radical

shown in I

This neutrt

in the solir
barrier for
structure r
twists in tr
crystal st
dimethox:

Fig
ray
001']

97
of radical 1 for comparison. A stereo view of the packing diagram is

shown in Figure 12.

\

. \0 /

“L1,
\\\\ ”"1
/ 0\ 131:3-an 3b
———>

O o
O
, \

This neutral isostructural analogue of 1 gives an approximate D3 structure
in the solid state. From this structure and AMl calculations (rotational
barrier for aryl rings; 4-6 kcal/mol),136 we conclude that the X—ray
structure of the radical 1 is anomalous. The comparison for 'aryl ring
twists in tripod binding sites are summarized in Table 1 and a summary of
crystal structure determination and refinement data for tris(2,6-
dimethoxyphenyl)methyl 1, tris(2,6—dimethoxyphenyl)borane 63, and

tris(2,6-dimethoxyphenyl)methyl cation 62 are summarized in Table 2.

 

Figure 11. Ball and Stick representations of the binding site: X-
ray structure of borane 63 (left), and MMP2 ground state

conformation of radical 1 (right).

 

 

Figur

It was $1
trimesitylb
identical. 1
stmctural c
cempound
haem
trierI’Ibor;
difference;
only a sir.-
Significan

Should 110‘

 

 

98

 

 

 

 

 

 

 

 

Figure 12. Stereo view of packing diagram of borane 63.

It was shown by Power that the solid state conformations of
trimesitylborane and the trimesitylborane radical anion are virtually
identical.137 The conclusion of this comparative study was that the

structural consequences of the addition of one electron into the LUMO of a

compound of the type Ar3X are trivial. While this conclusion is supported
by a tendency toward D3 symmetry in many reported structures of
triarylboranes, triarylmethyls, and triarylamines, there are gross
differences in the crystal conformations of l and 62, which also differ by
only a single electron. Similarly, the isoelectonic species 62 and 63 are
significantly different. These results caution that simple generalization

should not be drawn from a comparison of only two crystal stuructures.

 

Tabl

aTw
soul
solu

Tal
tris
bor

99

Table 1. Aryl ring twists in tripod binding sites (in degrees)3

 

1 12.3
1b 47
1C 45
63 62.8
62 32.6

61.0 61.0
47 47
45 45
64.2 64.2
46.1 48.9

C2 Axis

D3

D3

C2 Axis
General site

 

aTwists out of central atom plane; coplanar = 00° bDetermined in
soultion by Martin using EPR measurements.C Determined in
solution by Ishizu and Mukai using ENDOR measurements.138

Table 2. Crystal structure determination and refinement data for
tris(2,6-dimethoxyphenyl)methyl 1, tris(2,6-dimethoxyphenyl)
borane 63, and tris(2,6-dimethoxyphenyl)methyl cation 62.

 

 

Crystal data 6 3 6 2 1

Space group C2/c Pl P2/n

2* 4 2 2
Temperature 295 1 10 295

a (A)** 11.076 7.214 10.405

b** 20.839 12.931 9.429
c** 9.944 13.633 11.767
or (°)** 90 83.13 90

B** 98.40 77.70 102.120
y** 90 80.56 90

u (cm-1)*** 0.80 (Mo Kor) 9.7 (Cu Koc) 0.83 (Mo Kor)

 

*number of molecules in the unit cell, ** Cell dimensions, ***X-ray

source.

 

 

1.4. Hex

Tris
dimethoxy
various mt
groups wl
sterically ]
position i
substituen
ether oxy

ligand's bi

Berane 6
63 with t
methoxy
Borane i
ferrocen
Voltamr
tetrafluo
Potentiai

00nditi0

 

 

 

100

1.4. Hexachloro Triphenylmethyl Radical and Borane

Tris(2,6-dimethoxy-3,S-dichlorophenyl) borane 64 and tris(2,6-
dimethoxy-3,5-dichlorophenyl)methyl radical 65 were synthesized with
various modifications. The chlorine substituents are vicinal to the methoxy
groups which form the ether tripod binding sites; they may therefore
sterically perturb the methoxy groups, twisting them out of their preferred
position in the aryl ring plane. These sites also position a vicinal
substituent to withdraw or donate electron density by resonance with the
ether oxygens, potentially modifying their Lewis basicity and hence the

ligands binding abilities, without perturbing the radical center.

   

Borane 64 show remarkable stability toward protic acid compare to borane
63 with better preorganization for ion binding because of the twisting of
methoxy groups out of the binding cavities as shown in Figure 13.

Borane 64 shows a reversible redox potential of -2.14V (referenced to

 

ferrocene oxidation) for the process of B B" in cyclic

 

voltammetry in methylene chloride with tetrabutylammonium
tetrafluoroborante as a supporting electrolyte. Determination of the redox

potential for borane 63 was unsuccessful under a variety of experimental

conditions.

 

 

 

Figr
with
ion-
twis

Fi
so
PC

 

101

 

Figure 13. MM2 calculated structure of borane 64 dimer
with lithium cation showing improved preorganization for a
ion-binding over that in borane 63 by twisting methoxy group
twistings out of conjugation with the aryl rings.

 

0.000 I

 

r trio‘s

2.000

1.000

 

 

_____ -__._-1.-_.__._I_.__-.L__-_L --_L__-._‘

 

”.093 J l l l j
.000 -1.600 ‘1300 -0.000 -0.400 -O 000

Figure 14. Cyclic voltammogram of 64 in methylene chloride
solution of tetrabutylammonium tetrafluoroborate. The redox
potential was referenced to the oxidation of ferrocene.

 

 

2. Subst

It is
scheme, tl
the centra
magnetic j
therefore

substitute

The cyclir
conveys 1
consists r
unstirred
this worl
Saturated
(Ag/AgC
solvents

Cttiomel

0t whic‘,

OUT rest
and met
OXidatir
Voltamr
exeminr
cumin

 

102

2. Substituent Effects.

It is reasonable to suppose, based on our paramagnet coupling
scheme, that variations in the odd electron's availability (spin density) at
the central carbon and methoxy oxygen will affect the formation and
magnetic properties of crystallized complexes of these radicals. We have
therefore synthesized and examined the redox properties of various

substituted analogues of the radical 1 using cyclic voltammetry (CV).

The cyclic voltammogram is analogous to a conventional spectrum in that it
conveys inforrnations as a function of an energy scan. Cyclic voltammetry
consists of cycling the potential of an electrode, which is immersed in an
unstirred solution, and measuring the resulting current. The potential of
this working electrode is controlled by a reference electrode such as a
saturated calomel electrode (SCE) or a silver/silver chloride electrode
(Ag/AgCl). The accessible potential ranges for common electrochemical
solvents have been studied and found to be -3.0 to 1.4 V vs SCE (Saturated
Calomel Electrode ) for THF and -l.8 V to 1.9 V vs SCE for CHZCIZ, both

of which are wide enough for our measurements.139

Our results show reversible oxidation and reduction waves for 1 in THF
and methylene chloride as shown in Figure 15, indicating that all three
oxidation states are stable, at least on the time scale of our cyclic
voltammetry scans (50-200mV/sec). Various supporting electrolytes were
examined, including tetraalkylammonium and alkali metal salts, in order to

examine their effects on redox potential for ion-bindings.

 

/t

Fig
kin
rad
res
sec

One stre‘
Cempone
binding
accessib
mediates
Valuable
Choice (
electron
focused

 

103

0.400

 

 

 

 

 

Figure 15. Cyclic Voltammogram of 1 in THF. Note that
kinetically reversible electron transfer was found for both
radical/cation and anion/radical couples (-0.49 V and -2.05V
respectively vs. ferrocene oxidation), at scan rates of 200mV
sec-1.

One strength of our approach to magnetic materials is that the organic
components can be modified in synthesis to adjust the redox and ion-
binding properties of these multidentate radicals. As the range of
accessible oxidation states is thought to be related to superexchange-
mediated magnetic coupling in stacked materials, such control may prove
valuable.1‘10,141 This tuning of the paramagnetic ligands, combined with
choice of metal couplers, promises broad control over the magnetic and
electronic characteristics of our extended complex materials. I have
focused first on simple modifications to make derivatives of radical l, and

then progressed to the more complicated systems.

 

 

2.1. Her

Mar
made mir
approach
targets, 1
procedure
withdraw
reagents.
devoted :

the desire

Systemt
Oxidatir
detiyati
Based

cheract

 

104

2.1. Hexamethoxy Triphenylmethyl Radicals.

Martin's synthesis of 1 is simple and economical. Although we have
made minor modifications in building derivatives of radical 1, our
approach is basically unchanged from the original work. In selecting new
targets, I initially focused on substituents which can tolerate the synthetic
procedures. However, many of the most potentially interesting electron
withdrawing substituents cannot survive treatment with organolithium
reagents. Such substituents are of particular interest and we have therefore
devoted substantial effort to developing new syntheses in order to access

the desired range of substituted radicals.

       

\
O
1. PhLi or
n-BuLi Cr2+ or
—_>
2. (EtO)2CO . Zn/H30T or
Vitamin C
O
/
6 2 1

Electrochemical studies on simple para-substituted triphenylmethyl
systems have shown that the substituents can substantially alter their
oxidation and reduction potentials.142,143 Functionalization of l to make
derivatives similarly varies the redox properties of this family of radicals.
Based on simple resonance pictures, the electron donor or acceptor

character of 4-substituents in aromatic rings should strongly affect the

 

 

 

energy 0

perturbin

 

 

105
energy of the singly occupied M0 (the SOMO) without significantly

perturbing the oxygens of the methoxy groups.

1 :x =x'=x" = H

66 :x =X’=X" = Cl

67 : x =X'=X" = OCH3, OBu
68 : X =X'=X" = CH3

69 : x =X'=X" = Ph

70 : x = CH3

 

Table 3. Summary of electrochemical data for derivatives of
radical 1 in THF a) and CH2C12 b).

 

 

 

 

 

a)

x X' x" R+ _ R' R' ..__—- R'
H H H —0.49 V -2.05 V
C1 C1 Cl -0.30 V -1.84 V
OMe OMe OMe -0.89 V -2.29 V
Ph Ph Ph 050 V -1.90 V

THF Solvent; Reference: Ferrocene oxidation

b)

H H H -0.44 V -2.07 V
C1 C1 Cl -0.34 V «1.84 V
OMe OMe OMe -0.93 V -2.31 V
CH3 CH3 CH3 -0.72 V ~2.19 V
Ph Ph Ph 053 V -1.88 V
CH3 H H -0.61 V -1.99 V

 

CH2C12 Solvent; Reference: Ferrocene oxidation

 

 

 

By using
oxidation
property.
orbital en
might fin
helpful.

and 70, s
radicals

while elr
expected.
are only

condense

The redr
between
0f the 31
Where it

(meta 01

P iS a r
Peculia
eorrelat

Where ]

unsubsr
144

 

 

 

106
By using electron withdrawing substituents in 1, we may produce

oxidation-resistant derivatives of 1 which still retain 1's ion-binding
property. . On the other hand, if triplet coupling is promoted by better
orbital energy matching between radical and lithium cation orbitals, we
might find donor or resonance delocalized substituents to be especially
helpful. As a start on these studies, we have synthesized 66, 67, 68, 69,
and 70, simple derivatives of 1. It is clear that electron donor substituted
radicals 67 , 68, and 69 are more vulnerable to air oxidation than is 1,
while electron acceptor substituted radical 66 is less so. As might be
expected, yields vary from good to poor as some of the substituents studied
are only marginally tolerant of the ortho lithiation conditions required to

condense the intermediate triarylrnethyl carbinol.

The redox potentials in Table 3 can be examined for a possible relation
between half-wave potential (for the one-electron reduction) and the nature
of the substituent. One such relation is expressed in the Hammett relation
where k and k° represent rate or equilibrium constants for substituted

(meta or para) and unsubstituted compounds, respectively.

 

log k =p0'
k0

p is a constant characteristic of the reaction series, and o is a constant
peculiar to the substituents; cm is generally set equal to zero. The

correlations of electrochemical data are made using the following equation

where E1/2R and E1/2H are the half-wave potentials for the substituted and

unsubstituted compounds. Note that in the equation, p is expressed in volts.
144

 

 

 

The intr

electrode
many or
wave pc
reductio
cation/r:

couples.

Examin
graphic
(5 valur
values
becaus
involvr
correla
correla
cation
induct
contri'
001161:
Worse

times

 

 

107
AEI :EIR __E1H =p0'

2 2 2
The intrinsic difficulties of attempting E1/2 - o correlations when the
electrode reactions are irreversible have been stated.”5 Unfortunately,
many organic reductions and oxidations are irreversible and hence half-
wave potentials cannot have thermodynamic significance. However, the
reduction potential measured in this work are generally reversible for the
cation/radical couples and reversible or nearly so for the radical/anion

couples.

Examination of a relation between half-wave potential and 0 may be done
graphically merely by linear plotting of one against the other using known
6 values.”6 The correlations between half-wave potential and sigma
values are not so good as we can see in the least-squares fit in Table 4
because of the complexities of reactions and conformational effects
involved. It has been suggested that the cation stability should be
correlated with 0+ values, whereas anion and radical stabilities are to be
correlated with G or (5' values. 147 It appears as if substituents affect the
cation with resonance and/or hyperconjugation terms in addition to
inductive and polarizability terms, whereas only the polarizability
contributes substantially to radical and anion stability. An attempt to
correlate our experimental data with 6+ or (5‘ was unsuccessful and gave
worse correlation parameters than with o (the values were summed three

times for the trisubstituted radicals for the correlation).

 

 

 

 

 

NV) ,

I93,

 

 

 

 

EtV)

I I
-2 II -2-1 I .

108

 

 

E (V) 1

I
' Anion
0 Cation

 

 

 

 

 

 

 

Figure 16. AE1/2(V) vs 6 in THF for HMTP methyl radicals a) in
THF and b) in CH2C12.

Table 4. Least-squares fit of Correlation for half-wave potential
and sigma values for HMTP methyl radicals. AE1/2= pX-E'1/2

 

 

 

P E'1/2 #3 Rb
(S (Cation)C 0.3969 -0.53 4 0.965
(S (Anion)C 0.3019 -2.00 4 0.852
G (Cation)d 0.3912 -054 6 0.905
(S (Anion)d 0.3177 -2.00 6 0.805

 

 

a Number of experimental points used. b Correlation coefficient. C
THF solution. d CH2C12 solution.

 

 

Bank pro
radical sy
simple su
first of t]
triphenyl
balanced
monomer
are repor
expect si
function
stabilizir
maximal
stabilizir
unsubsti

methyl 2

Bank s1
System
Withdra
effect;
SUbStitu

exactly

Direct
Variety

commc
1imited

 

 

109
Bank proposed that substituent effects detected in CV of triaryhnethyl

radical systems are not likely due to conformational changes but rather to
simple substituent contributions. He recorded the largest change for the
first of the three methyl substitutions for H in the para position of the
triphenylmethyl system. Steric and conjugative effects are said to be
balanced in each case to provide the optimum stabilities. For para
monomethylated triarylrnethyl anion, the twist angles for the phenyl rings
are reported to be 197°, 306°, and 448°, respectively.148 Thus, we can
expect significant differences in the degree of conjugation of aryl rings as a
function of substitution. For the electron-deficient cation the greater
stabilizing effect can occur when the donor methyl group is in the
maximally conjugated ring . Similarly for the electron rich anion better
stabilizing effects are achieved when the maximally conjugated ring is
unsubstituted. This analysis is based on the electron donor properties of

methyl and the differential ring conjugation.

Bank suggested that the preferred ring conjugation of the triarylrnethyl
system should be determined by various substituents. In fact an electron-
withdrawing group should reverse the pattern of sequential substituent
effect; thus, for the anion, greater stabilization would result when a
substituted ring is maximally conjugated. For the cation the prediction is

exactly the opposite.

Direct application of Martin's original synthesis of radical 1 calls for a
variety of 4- and 5-substituted 1,3-dimethoxybenzenes. The range of
commercially available dimethoxybenzenes is adequate if somewhat

limited. In cases where the desired substituents could not tolerate Martin's

 

 

procedure
substituer
assembler
modified
systems.
functiona
acid-stab
breaking

It has her
amines c
7 2, do
dimetho
modifiee
The desi
With sod

a

Ihates

COITeSp

111016 1'

electroi

 

 

1 10
procedure, an alternative strategy would be desirable which would allow to

substituents to be added after the triarylrnethyl framework was fully
assembled. The resulting intermediate could then be converted to a new
modified radical, opening the way to much broader class of substituted
systems. With this goal in mind, we have explored the possibility of
functionalization after assembly of the triarylrnethyl framework via some
acid-stable derivatives in which the central carbon is sp3 hybridized,

breaking up the extended rt-system.

It has been found that ammonia, as well as various primary and secondary
amines can add into the central carbon in cation 62. The ammonia adduct,
72 , does not have the proton acid sensitivities of tris(2,6-
dimethoxyphenyl)methyl carbinol 71. This ammonium salt can be
modified with a wide range of halides by simple electrophilic reactions.
The desired radicals can be than generated readily by a simple diazotization

with sodium nitrite in acidic water.

 

In a less generally successful strategy, carbinol 71 can be reduced to the
corresponding triarylrnethane 73 using sodium cyanoborohydride which is
more robust toward both treatment with organometallic reagents and

electrophilic aromatic substitutionsl‘t9 The corresponding hexachloro

 

 

methane

methylene
yet found
Although
umnodifir
evidently
at deprot
pathways

 

 

 

111
methane 76 was synthesized by a simple chlorination with S02C12 in

methylene chloride. The products are well characterized, but we have not
yet found a clean conversion back to the corresponding radical or cation.
Although oxidation with p-chloranil is effective at converting the
unmodified methane 73 to the cation, the hexachlorinated analogue 76 is
evidently too deactivated to undergo the analogous oxidation. All attempts
at deprotonation were also met with failure. A variety of less general

pathways were also introduced for specific substitution patterns.

 

 

 

 

71
‘ HBF4
\
0
(3+ NaBH3CN
Ceric
Ammonium
/ 3 Nitrate

 

62

 

There ar
can be u
could ne
availablr
scheme.
mainly

carbona

for furtl

 

 

112

 

There are several more reactions worth mentioning here. The methane 76
can be used for building several substituted triphenylmethyl radicals which
could not be built by the standard procedure because of side reactions or an
available precursors. One example is illustrated in the following reaction
scheme. The methane 76 can undergo ortho-chlorine directed lithiation
mainly by the induction effect followed by quenching with dimethyl
carbonate to give various substituted triaryl methanes which can be used

for further functionalization of aromatic rings in the system.

 

 

 

W
useful rr
halogen
unsatur:
These r
derivati

radicals

Electro
have 1
precurr
dirneth

reactio

113

H-C

   

While building these substituted radicals, two very interesting and
useful reactions were somewhat serendipitously introduced: Electrophilic
halogenations of resorcinol derivatives; and condensation of or, B-
unsaturated ketone and dimethyl malonate followed by aromatization.
These reactions, especially the electrophilic halogenation of resorcinol
derivatives, provided several key transformations in building these

radicals.

Electrophilic halogen substitution and alkylation of resorcinol derivatives
have been studied for decades. While pursuing the preparation of
precursors for radicals, we were able to build a variety of halogenated 1,3-
dirnethoxy benzenes which we found later in the literature.150 Some of the

reactions are summarized in the following reaction schemes.

/ O\ /O O\
2 Brg
_____y
Br Br
X X
X: H, Cl. 1, OMe, Ph

/ O O\ / O\
m
C1 C1
X

X
X: H, C1, 1. OMe, Ph

 

Most of 1
An ether
amount
ovemigh
usually
dimetho:

dimethor

The 00:
followe
phenyl.
overall
aldehy.

aromat'

 

114

Most of the reactions are nearly quantitative with easily purified products.

An ether solution of precursors was combined with a stoichiometric

amount of halogenating agent at ~78 °C followed by slow warm up

overnight and filtration of crystalline products. Washing with (1
usually gave clean crystals of the products except
dimethoxybenzene which we could distill under vacuum

dimethoxybenzene was synthesized with aqueous NBS.

’0 0~ ,o 0‘
Ct sozcr. Cr
—*’ Cl
0 0‘
’ U NBS ”U“
——->
. Br
,0 o, ,o 0‘
U 4728‘ U
C Cl Br

iethyl ether
4-chloro

. 4-Bromo

The condensation of 0t,B-unsaturated ketones with dimethylmalonate

followed by aromatization was a key reaction in the preparation of 4-

phenyl substituted triphenyl methyl radical 70. With a good efficiency for

overall reaction and wide commercial availability of a range of aryl

aldehydes, these reactions have great potential for building various

aromatic group substituted triphenyl methyl radicals.

 

 

 

 

All of
variah
been i
Figur
deterr

Tabb

 

OO

 

O
D H A. NaOH MeoU OMe Naorat
V

 

 

 

 

All of the radicals generated have been studied by room temperature and
variable temperature X-band EPR. Some ion-binding studies also have
been carried out by EPR. Some of the typical EPR spectra are shown in
Figures 17, 18, and 19 along with simulations. The experimentally

determined coupling constants ai and the spin density pi are summarized in

Table 5.

 

4—‘AJH-f‘

 

11
f0
ra

Kivelr
tfiphm
father
subsd
the m
Garbo
ava‘n;
donat
may :

CatiOr

 

116

Table 5. A summary of 1H hyperfine constants ai (in Gauss).
for coupling with meta and para proton in HMTP methyl

radicals in THF.

 

 

ameta (#)2 apara (#)2 Ipmetal1 tpparat1
1 1.06 (6) 2.26 (3) 0.047 0.100
14 1.07 (6) 2.35 (3)
65 2.3 (3) 0.102
663 .1.0 (6) 0.044

 

1 lpcxpl = la /22.5| 2numbers of protons in each positions.

3radicals 67 , 68, and 69 show similar results. EPR spectra
simulated with the ESRa program written by A. K. Rappé and C.
J. Casewit, Calleo Scientific Software, Colorado State University.
4Determined by Ishizu and Mukai using ENDOR.138

Kivelson reported that chloride substitution in the para position in

triphenyl methyl increases the spin density at the ortho carbon slightly and

rather more at the meta position.151 He suggested that methoxy

substitution in the para position in triphenyl methyl radical does not affect

the meta carbon spin density but increases the spin density on the ortho

carbon considerably. We can follow the same arguments in estimating

available spin densities on methoxy oxygen in our radicals with electron-

donating para-substituents. The spin densities appear to be small but they

may play a significant role in the magnetic interactions in our radical metal

cation complexes.

 

 

 

 

117

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 17. EPR spectrum of l in THF with simulation (top).

 

 

 

118

 

—.

Figure 18. EPR spectrum of 65 in THF with simulation (top).

 

 

119

 

‘AA
W v—v ' —A A.‘
W wfi '

Figure 19. EPR spectrum of 66 in THF with simulation (top).

2.2. Ti

Sr
undergo
acidic c:
the eye
Figure

cyclizer

 

 

 

120

2.2. The Cyclized Xanthenol

Some of the cations which are intermediates in generating radicals
undergo a cyclization to form a xanthenyl X (= OH, NH3+, N+H2R ) under
acidic conditions by way of the mechanism outlined below. We believe that
the cyclization occurs when two of the rings are forced to be coplanar.

Figure 20 shows a stereo view of the X-ray crystal structure of the

cyclized xanthenyl from carbinol 71.

 

 

. \
o __\o o o o
O C 70 GO :0 CD
{<5 9* n 0
/\ ‘
H+

 

Figure 20. Stereo view of cyclized xanthenol from X-ray structure.

 

 

 

2.3. T

0
build lir
tripod e
ion-bin
molecu
packin;
forms '.
the twr

a roug

the so]

0
4/9

It 0
stab‘
com
the

dist

 

 

 

121

2.3. Tetramethoxy Triphenylmethyl Radicals

Our purpose in using complexes of radical 1 with metal cations is to
build linear chains of radicals by complexation of the metal ion with the
tripod ether oxygens in the radical. Various ion-binding studies suggest that
ion-binding can occur with as few as four methoxy groups in a given
molecule. Furthermore, careful examination of the three-dimensional
packing diagram of radical 1 shown below reveals that the radical itself
forms linear chains without a metal cation. The four methoxy groups on
the two more twisted rings are already placed in such a way that they form
a roughly square planar ligand environment between adjacent radicals in

the solid state.

/

, 2

.J. 79»

;“m/24664
/////

‘/

 

It occurred to us that if a radical with only four methoxy groups would be
stable enough to be handled in the usual way and could form a linear chain
complex with metal cations between radicals, it should be possible to study
the magnetic behavior of the resulting coordination chemistry. As I

discussed earlier, the anomalous solid state structure for l is troublesome

 

 

 

 

even tho
anount
thought
defined

altemati

A seri
methy
modif
usual
derive
dimet'
Mukt
Specu
65°, ;
const

Show

const

 

122
even though various calculations show that it can be explained with a small

amount of crystal lattice energy (3-6 kcal/mol by AMlcalculations). We
thought that if the tetramethoxy radical could also give a structurally well
defined complex, and magnetic interactions, this new complex might be an

alternative building block for molecular magnetic materials.

 

A series of para-substituted phenyl bis-(2,6-dimethoxyphenyl)phenyl
methyl radicals was prepared as shown in next page by a simple
modification of Martin's procedure. All of the radicals were studied as
usual by EPR and CV as with the hexamethoxy triphenylmethyl
derivatives; the results are summarized in Tables 6 and 7. The di-(2,6-
dimethoxyphenyl)phenyl methyl radical 77 has been studied by Ishizu and
Mukai using the Electron Nuclear Double Resonance (ENDOR)
spectroscopy and was found to have C2 symmetry with twist angles of 65°,
65°, and 30° in toluene. They reported almost identical hyperfine coupling
constants to our values for the radical 77 in toluene.152 This spectrum
shows significant solvent dependence as seen from the hyperfine coupling

constants in THF (Table 8).

 

 

 

 

 

 

123

 

77: X =X'= H

78: X = C1

79: X = OCH3

80: X = CH3

81: X = OH

82: x = COzH

83: X = N02

84: x = H, x: OCH3

Table 6. Summary of electrochemical data for derivatives of radical
19 in CH2C12.

 

 

4-X 3,5-X2' R+ = R R = R
H H -0.32 V -1.90 V
C1 H -0.30 V -1.84 V
OCH3 H -0.53 V -2.04 V
CH3 H -0.42 V -1.93 V
OH H -0.58 V -1.82 V
C0211 H -0.22 V -1.34 V
N02 H -0.25 V -1.80 V
H OCH3 -0.34 V -1.87 V

 

CH2C12 Solvent; Reference: Ferrocene oxidation

 

 

 

 

As with the l
half-wave p01
(TMTP) met]
correlations

l complexities
of hydrogen
81 and 82 sh:
be explained
It.

Flgur
in CH

 

 

 

124

As with the hexamethoxy triphenymethyl radicals, correlations between

half-wave potential and literature sigma values of tetramethoxy triphenyl

(T MTP) methyl radicals are poor, as shown in the least-squares fit of

correlations in Tables 7; presumably, such difficulties reflect the
complexities of reactions and conformational effects involved. The effects

of hydrogen bonding in CV are not completely understood. The radicals

 

81 and 82 show the most dramatic departure from the series, which might

be explained by the effects of hydrogen bonding and n-conjugation through

it.

 

 

 

 

 

‘ o
O
-0.5'. o
E(V)
U
-1.5"
I
I E] a
I
-25 v 1 f I ' I ' I fi I
O4 02 00 02 04 06 08

Figure 21. AE1/2(V) vs 0' in THF for TMTP methyl radicals
in CH2C12.

 

Table
potenti

Bin

0 (Cat
0 (Am'
a Nun

c CH2

Some of the

With mm

the Spin der

125

Table 7. Least-squares fit of Correlation for half-wave
potential and sigma values for derivatives of 77. AE1/2= pX-

 

 

E'1/2

P E'1/2 #a Rb
0 (Cation)C 0.2957 -0.4029 8 0.790
G (Anion)C 0.2741 -1.8480 8 0.265

 

a Number of experimental points used. b Correlation coefficient.

0 CH2C12 solvent.

Some of the typical EPR spectra are shown in Figure 22, 23, 24 and 25

with simulations. The experimentally determined coupling constants ai and

the spin density pi are summarized in Table 8.

 

 

Table 8. A summary of 1H hyperfine constants ai (in Gauss) for

126

coupling with meta and para proton in TMTP methyl radicals.

 

 

aortho (#) ameta (111)2 apara (#)2 lporthI1 lpmeta|1 Ipparall
77 0.91 (4) 1.00 (2) 0.040 0.044
4 4.30 (2) 1.50 (2) 4.96 (1) 0.191 0.067 0.220
5 4.09 (2) 1.51 (2) 4,67 (1)
78 0.9 (4) 1.0 (2) 0.040 0.044
4 4.2 (2) 0.7 (2) 0.187 0.031
843 0.8 (4) 1.1 (2) 0.035 0.049
4 4.8 (2) 3.2 (1) 0.187 0.142

 

 

 

1 Ipexpl = |(a/22.5)| 2numbers of protons in each position.

3Radicals 79, 80, and 83 shows similar results. 4Asymmetric
ring protons coupling constants in TMTP methyl radicals.
5Determined by Ishizu and Mukai using ENDOR in toluene.138

128

1 l O

 

 

Figure 23. EPR spectrum of 77 in toluene with simulation (top).

129

 

Figure 24. EPR spectrum of 78 in THF with simulation (top).

 

 

 

 

131
2.4. Rotational Barriers

Triaryl-X propellers, both planar and pyramidal, are well known to
exhibit barriers to aryl ring rotation. Some of the triaryl systems
constructed in this work and related studies show pairs of methoxy proton
resonances at room temperature which are interpreted to indicate ring
rotation that is slow on the NMR timescale. We thought that it would be
useful to determine rotational barriers for relevant systems to have a better
understanding of the structures of our systems in solutions. The rotation
barriers of the aromatic rings in the 71 and 73 have been studied by
Rieker and Kessler using 1H NMR and found to be less than 8.2 kcal/mol
for carbinol 71 and 11.1 kcal/mol for methane 73.153 The silane 74 in
which the central carbon in methane 73 is replaced by Si crystallizes on a
three-fold axis so that all three rings are equivalent. There are two
independent molecules in the unit cell and they have twist angles of 30 and
33° respectively. It is worth adding that the methane 73 crystallizes in the
same space group as the silane and has similar ring twists (27°). Due to
its single intramolecular hydrogen bond(one of the methoxy groups is
twisted 75° out of the ring plane because of the hydrogen bond), carbinol

71 has a lower symmetry in the solid state (Pbca space group).

 

Figure 26. X-ray structure of methane 73.

 

 

 

 

 

132
Variable temperature 1H NMR studies have been conducted over the

temperature range of -90°C to +100°C using the Varian VXR 300 NMR
instrument in toluene-d8 or CDC13 solvent to observe coalescence of two
methoxy resonances in systems with doubled methoxy proton peaks at
room temperature and therefore restriction of free rotation of rings. The
results are consistent with a site exchange process resulting from the rapid
flipping of the aryl rings on the NMR time scale. The value of the
rotational barrier has been determined by applying the Gutowsky-Holm

approximation to the observed methoxy group site exchange at the

coalescence temperature (TC in K).154

For an exchange process between two nuclei A and B with a mutual

coupling J AB with peak separation of Av in Hz, the rate constant kC at the

coalescence temperature To is given by:

 

kc = 2.227sz + 6J2”

From the Eyring equation, one can determine the free energy of activation
by using the following equations. To calculate the activation barrier, we
need a pair of kC and Tc. k3 is the Boltzmann constant, K is a transmission
coefficient usually assumed to be 1, and h is Planck's constant.

@e—Ac‘mr

k=K

AGf = 4.58TC (10.32 + log—£90m] / mol

C

AGC¢ =19.14TC(10.32 + log £9” / mol

C

 

 

 

133
The activation barriers, AGi , for aryl ring rotation of tris(2,6-

dimethoxyphenyl)methy1 ammonium tetrafluoroborante 72, tris(2,6-
dimethoxy-3,5- dichlorophenyl) methylammonium tetrafluoroborante 75,
and tris(2,6-dimethoxy-3,5- dichlorophenyl) methane 76 were 12.67i0.3
kcal/mol-l, 15 .7i0.8 kcal/mol, and 17.3i0.8 kcal/mol, respectively.

Table 9. Variable Temperature NMR results for tris(2,6-
dimethoxyphenyl)methy1 ammonium tetrafluoroborate 72,
tris(2,6-dimethoxy phenyl)methane 73, tris(2,6-dimethoxy-
3,5-dichloro phenyl) methylammonium tetrafluoroborate 75,
and tris(2,6-dimethoxy-3,5-dichlorophenyl) methane 76.

 

 

 

72 73a 75 76
ASH (CH3O, ppm) 0.63 0.254 0.331
(at T/K) 233 --- 263 253
AGi (kcal mol'l) 12.67 11.1 15.68 17.27
Tc (K) 273 216 323 358
kC (8'1) 419 —-- 169 220
AV (Hz) 189 33 76.2 99.3

 

 

3Determined by Rieker and Kessler in CDC13.

 

134

 

 

.____.. .__, ,,_ -\ A

.__- h”--. -_- _._.—-. -_—-__

 

 

 

17'"[W'IllllI—rlllllll'IlYlllllll'lTllITlllllllTlllllllllllllll’llll‘llT—q
4.4 4.: 4.- 4.1 4.0 71.9 3.8 3.7 3.b 3.5 3.4 3.3 .3,; 00m

Figure 27. 300 MHz 1H NMR signals of the methoxy protons in
hexachloro tris(2,6-dimethoxyphenyl)methyl ammonium BF4 75 ,
recorded at different temperature. For slow rotation two peaks are
obtained, whereas for fast exchange there is only one. In the
intermediate range the signals are broadened. The coalescence
temperature Tc is 50°C. (-10, 20, 30, 50 °C from the bottom)

 

 

 

 

 

135

 

 

_ ._»\ ._ , .4»w~»—'§_—-W‘r (x A ’-._’ 'V 7 ‘ d“

WM

.IVWVIY11117111[TllllllTTfTllTlll'llllllllT‘IIlllTITllllllllllIllT—Tlllll111111111111111171'111

.- 4.0 3.8 30 34 3.2 30 2.8 2.6 2.4

 

Figure 28. 300 MHz 1H NMR signals of the methoxy protons in
hexachloro tris(2,6-dimethoxyphenyl)methane 76, recorded at
different temperatures. For slow rotation two peaks are obtained,
whereas for fast exchange there is only one. In the intermediate range
the signals are broadened. The coalescence temperature Tc is 100°C
for the methane. (20, 40, 60, 80, 100 °C from the bottom)

 

 

 

136
Unlike 72, 7 5, and 76, HMTP carbinol 71, and HMTP silane 74 do not

show splitting of methoxy 1H resonances at 300MHz even at -90 °C. These
results are consistent with a faster flipping of the aryl rings than NMR
timescale even in carbinol 71 in which we believe that there is a hydrogen-
bond between one of the methoxy oxygens and the hydroxy proton at the
temperature. We may understand why the carbinol 71 has a lower barrier
for rotation than the methane 73 using "the two-ring-flip mechanism" of
Mislow. The formation of the hydrogen-bond would fix the twist angle of
ring in a "vertical" orientation, making it easier to rotate the other two rings
appropriately in carbinol 71. By assuming the methane 73 as a worst case of
all the triaryl methyl systems studied as far as rotation barrier of rings
(except the case of 72, in which the methoxys are twisted), it would be safe to

say that the activation barriers for the rotation in these systems are smaller
than 11.1 kcal/mol in solution of CDC13 and toluene-d8.

The rotation of rings in 71 and 72 would result a cleavage of the hydrogen-
bonding interactions. We expected that the rotation barrier for methoxy
groups in these systems will reflects the strength of these interactions. The
carbinol 71 does not show any splitting of the methoxy proton signals at
-90°C in toluene-d8. The difference between the rotation barrier of 72 and
73 is only 1.6 kcal/mol which might reflect the strength of these interactions.
It is difficult to understand why they show such a small difference when 72
has three hydrogen—bonding interactions in the solid state compared to 73

which does not have any.

 

 

 

 

 

137

3. Ion-Bindings

Operating on the supposition that the relaxation of 7Li nuclei

(S=3/2) would be considerably shorter in the complex l-LiBF4 than in

uncomplexed mixture, Jackson and Kahr compared the line widths of the
solid state 7Li-NMR spectra for the precipitated solids with pure LiBF4
salts. The 7Li line widths in the reported complexes were 16 kHz at half
height as opposed to 500 Hz in the pure salts. Jackson and Kahr interpreted
these changes in terms of a specific binding interaction between radical l
and the 7Li ions in the precipitated solids; however these results do not

exclude other nonspecific broadening mechanisms.

3.1. Borane

Small metal cation binding studies of borane 63 by UV—Vis

 

 

spectrophotometry suggest ion-binding, consistent with the notion of stack
formation. Similar UV—Vis studies of 1 are hindered by the Li+ catalyzed
oxidation reaction which yields the intensely colored cation 62.

\
o O/

LiX UK
a a
B 0 Hz0
<

O
I

 

 

 

 

138
UV-Visible spectrophotometry provides a useful probe of the

binding of small metal cations by radical 1 and borane 63. The long-

wavelength maxima in each of these species in ether, THF, CH2C12, and
CH3CN are markedly diminished in intensity by addition of solutions of
lithium salts (LiBF4, LiClO4, LiSCN). Washing the complex solutions

with water recovers the original species' spectra. Analogous spectral
changes are not seen in solutions of anisole or m-methoxyanisole upon
treatment with lithium salts; the only observed changes correspond to
simple dilution of the sample solutions. We interpret these observations in
terms of the binding of small metal cations to the triaryl propellers l and
63, which alters their absorption spectra. These ligands are apparently not

such efficient ion binders that they can compete with bulk H20; hence the

observed reversibility upon washing as shown in Figure 29.

Ion binding studies by NMR on the borane using various Li and Na salts in
various solvents do not show significant changes except in the case of
Lil-Borane which is still small (0.02 ppm changes under very anhydrous

conditions) presumably because of the very low binding constant involved.

~—-r1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

139
1. 500 .37
\ /\
l ,«" \..
1‘ ,-' \
ABS //"*\\/" ‘
\J’ \
\3
3.000 A—C—uu __________ a
ABS
3. 000 fl , j _%
\
c)
ABS
0.000 , , , 4 l
rm 240 mo 320 360 400

Figure 29. UV-Vis spectra illustrating reversible Li+ binding in
CHZClz. (a) Borane 63 alone; (b) Mixture after addition of LiBF4

solution; (c) Organic layer after H20 wash. .

 

140
The FAB Mass spectrum give one more piece of evidence for a notion of

binding of radical 1 with metal cation. It shows RMR peaks in the case of

 

NaBPh4 complex of radicals 1 and 68 which means that the complex
formed by simple mixing of salts with radical is strong enough to remain
intact through the FAB ionization process even if it only shows very little

changes in other spectroscopy.

J502180005 Scan 2 HT=0:38 100%:410000 um 19 Feb 93 14:54
Compacted SLRP +EI SA70(NBA)1

 

 

 

 

 

 

 

   

 

 

+ NaBPh4
100— 446.0
151.0 R'——>
‘— R'Na‘
523.0
04 -AL _. i A 5"0
ultrrftllilluriirrrrr
150 200 250 300 350 '400 450 500 550 600 650
1007
.1 I R'Na‘R’
”0.2
m2 ' 0311.2 1045.2 1130.5 1193.5
o—J1110721llIllllllll’llllrlll‘llljj
700 800 900 1000 1100 1200

Figure 30. Mass Spectrum of dimer 2 of l-Na+-1 (869) in
m-nitrobenzylamine matrix.

 

141

3.2. Double faced Paramagnetic Ionophore

Substitution of Z for the methyl groups shown below in 1 allows for
a whole range of new radicals to be synthesized. By controlling the size,
electron withdrawing/releasing properties, and ligand capabilities of Z, the
ion binding properties of 1 can be varied. Synthetically, the Z-sites are
certamly the easiest to substitute; resorcinol (3-hydroxyphenol) and 3-
methoxyphenol are commercially available starting materials, and
Williamson ether synthesis provides a straightforward means for attaching

many possible 23.

The theme of two radicals sandwiching a metal ion can be inverted in a

complex with two metals fixed around a single radical center.

To evaluate the validity of our electron-coupling strategy, it would be
useful to know the dimensions and the spin density on the bound metal
cation. From such data it should be possible to estimate limiting

magnitudes for the exchange coupling which could occur through the metal

ion couplers in the hypothetical structure.

 

 

 

142

 

86-2LiBF4 85-1.5LiBF4

We have recently reported155 the radical 86, a double "octopus"
analogue of 1.156 This work addresses the spin density questions and
simultaneously provides structural insight into the novel two-faced binding
properties of the hexamethoxy-triaryl complexants. In parallel with the
studies of 86, radical 85 has also been built, with the aim of making a
radical with one face closed to chain formation.

Radicals 85 and 86 were synthesized by straightforward modification of
Martin's original procedures. Unlike 1, compounds 85,86, and their
related triarylrnethanol precursors are oils. Purification in this system is
most conveniently achieved by crystallizing tetrafluoroborate salts of the
corresponding triarylrnethyl cations; these are subsequently reduced to
make the neutral radicals. CW-EPR spectra collected for 86 at -70 °C
showed a 13- line pattern with the splitting between consecutive lines being

1.0 gauss. Figure 29 shows the experimental and simulated CW-EPR

spectra of 86.

 

 

143

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UJU

Figure 31. EPR spectrum of 86 with simulation (top).

 

 

 

 

 

144
Our purpose in synthesizing 86 was to build an analogue of 1 which

could not form extended chains by pairing around metal cations as in dimer
2. Radical 86 offers a bound Li+ ion a full coordination sphere of six
ether oxygens, effectively "capping" the radical on both faces. Thus, the
radical center in 86-2LiBF4 is encapsulated and should behave as a
completely isolated paramagnet. Like 1, radicals 85 and 86 show some
tendency toward oxidation on treatment with small metal cations in air.
This behavior is evidenced by the appearance of the blue color of the

corresponding triarylrnethyl cation in each case.

The electron spin echo envelop modulation (ESEEM) technique of pulsed
EPR spectroscopy has been used to measure weak electron-nuclear
hyperfine coupling between the radical centers and coordinated metal
cations in 86 in collaboration with Professor McCracken and Mr. Hong-In
Lee. The ESEEM method provides structural details regarding radical-
metal ion distance, the number of metal ions held near a given radical
center, and an estimate of the unpaired electron spin density transferred to

the metal.

 

 

 

145

 

‘0
A

 

f
A

 

U I
A

U
L

 

 

1

 

 

J L l 1 J

0246,8101214161820
froquoncy 0.1:)

 

Figure 32. ESEEM spectrum of 86-2LiBF4 in 1:1 THFzToluene

at 4 K. The spectra are fourier transforms of time domain data
collected under the following conditions: microwave frequency

* 9.346 GHz; magnetic field strength 3340 G; microwave pulse power
50W; pulse widths 16ns FWHM; two pulse excitation.

 

L_“

 

 

 

 

 

 

146
Addition of excess (10 equivalents) of LiBF4 and NaBPh4 to 10'6 M

THF solutions of 86 results in complex 86-2LiBF4 and 86-2NaBPh4.

Because no new hyperfine couplings could be resolved in the CW-EPR
spectrum after this treatment, the ESEEM technique was used to check for
possible weak 7Li hyperfine coupling to the paramagnetic center.157
ESEEM data were collected on a home-built spectrometer described
elsewhere.158 A two-pulse (900-1-1800) ESEEM spectrum collected at a
microwave frequency of 9.35 GHz and a magnetic field strength of 3350

gauss is shown in Figure 32.

The prominent doublet centered at 5.6 MHz shows a splitting of 1.0 MHz.
When ESEEM data were collected at 8.52 GHz with a field strength of
3052 gauss, the center of the doublet shifted to 5.0 MHz, consistent with its
assignment to 7Li. Also present in the spectrum of Figure 32 are 7Li sum
and difference combination frequencies centered at 11.2 and 1.2 MHz,
respectively, and a peak near 14.3 MHz due to weakly coupled protons. An
analysis of modulation frequencies and depths using the formalism of
Shuben and Diknov showed that this coupling arose from two 7Li+ ions
coupled to the radical center at an effective dipole-dipole distance. of 3.5 A
and having an isotropic hyperfine coupling of 0.4 gauss.159 Because the
line width of the sum combination peak is independent of hyperfine
anisotropy, the amplitudes and damping factors for the fundamental peaks

(5.0 and 6.2 MHz) relative to that of the sum combination line can be used

to determine the dipole-dipole distance.

 

 

 

 

 

 

 

147

 

 

 

 

% k A L L A‘ .A A L

é 4 6 a :0
frequency (MHz)

Figure 33. ESEEM spectrum of 86-2NaBPh4 in 1:1

THFzToluene at 4 K. The spectra are fourier transforms of
time domain data collected under the following conditions:
microwave frequency 8.90 GHz; magnetic field strength 3195
G; microwave pulse power 25W; pulse widths 16 ns FWHM;
and a 1: value 147ns for a stimulated echo excitation sequence.

 

 

148
The absolute intensities of the lines can be used to determine the number of

nuclei that give rise to the coupling. No coupling is seen to 10B, 11B, or

19F nuclei of the BF4" counterions; thus the splittings do not simply arise

via nonspecific coupling to magnetic nuclei near to the radical.

Analogous studies of 86-2NaBPh4 and 86-2Nal complexes gave similar
findings; a stimulated echo spectrum of 86-2NaBPh4 is shown in Figure
33. Two peaks centered at 3.6 MHz, the 23Na larmor frequency, and split
by approximately 2.5 MHz are observed. The lack of a pronounced sum
combination peak in the two pulse ESEEM data precludes independent
determinations of dipole-dipole distance and the number of coupled Na+
ions. Computer simulations of these ESEEM data are consistent with an
isotropic hyperfine coupling of 2.4 MHz, and an effective dipolar distance
of 3.8 A. Values for reff are calculated assuming that two Na+ cations are
coupled. If only a single Na+ nucleus is coupled the calculated dipole-
dipole distance decreases to 3.3 A, a distance that seems unlikely based on
CPK models and MNDO studies. Assuming that the 23Na hyperfine
coupling constants can be compared with their atomic values, unpaired spin
densities of approximately 0.3% can be estimated for both Li and
Na.160The Li+-radical center distances from ESEEM analysis using the
point dipole-dipole approximation are in good agreement with those
obtained from MNDO calculations. These results are summarized in

Figures 34, 35, and Table 10.

A larger splitting of 6.0 MHz is also observed in the Na salt-treated

samples of 86. This result strongly suggests a complex with only one Na+

 

 

 

 

 

 

149
bound, inducing pyramidalization at the radical center and hence decreased

distance and increased overlap between radical center and N a+.

 

Figure 34. Summary of ESEEM spectroscopic and MNDO
calculated results.

CPK models yield similar geometrical
characteristics.

       

   
   
 

   
      
      
   
   

/
/% 6%

     
   

)- i' ‘I
rig $7.}; 051
y/j/x. ./,,. .

Figure 35. Stereo view of the MNDO calculated structure;
the BF4‘ counterions were left out of the calculations.

 

 

 

Table 10. ESEEM Determined Complexation Data for 86 with
Various Salts.

 

 

Salt AisoaVle) rem/1) IMNDO #bound M+ Spin Density on W
LiBF4 0.9 3.5i0.2 3.45 2 0.3%!
LiI 1.0 3.4i0.2 3.45 2 0.3%a
NaBPh4 2.4 38:20.20 b 2 0.3%
NaI 2.4 3.8i0.2C b 2 0.3%

 

a. MNDO calculations used starting geometries generated from
a simpler MNDO calculation on 1-2LiF by "growing" the
ether atoms in the expected lowest energy conformation, as
evaluated using the Chem 3D molecular mechanics package.
Optimization was carried out in D3 symmetry, as suggested by
calculation on 1, the above complex, and related systems.
MNDO calculates a spin density of 0.22% on the bound 7Li
ions, in reasonable agreement with experiment.
Unfortunately, without suitable reference systems, we cannot
judge the significance of this value.

b. MNDO parameters are not available for Na so these
structures were not calculated. Attempts made using
"sparkles" which represent Na+ as a positively charged hard
sphere gave unreasonable structures.

 

 

 

 

151
Electrochemical characterizations of 85 and 86 were accomplished

via cyclic voltammetry (CV)161. The salts 85+BF4— and 86+BF4-(1 x 10-4

M) were studied in THF and methylene chloride solution using
tetrabutylammonium tetrafluoroborate as a supporting electrolyte.

Reversible reduction of each cation to the corresponding radical was readily

measured; a second reduction wave (radical to anion) could also be 1
observed when THF solutions of radicals were used instead of the cation

salts. Table 11 compares the observed reduction potentials for l, 85, and
86.

Table 11. Summary of electrochemical data for radicals 1,
85, and 86 with data for radical 1 for comparison. in THF a)

 

 

 

and CH2C12 b).
a)
R+ :=——‘ R. R. 2 R-
l -0.49 V -2.05 V
85 -0.47 V -2.10 V
86 -0.48 V -2.04 V

 

THF Solvent; Reference: Ferrocene oxidation

 

b)
1 -0.44 V -2.07 V
85 -0.50 V -1.98 V
86 -0.48 V -l.97 V

 

CH2C12 Solvent; Reference: Ferrocene oxidation

 

 

 

 

152
As expected, the redox potentials for 1, 85, and 86 are nearly identical,

which shows that the added ether extensions in 86 have essentially no effect
on the redox properties of the Ar3C nucleus. The separations between
cathodic and anodic peaks were large compared to the theoretical limit for
Nemstian behavior (0.06 V for peak to peak distance); the wave for
ferrocene/ferrocene+, a well-behaved reversible redox couple, showed

similar separations. Thus, we believe electron transfers from an electrode

to Ar3C substrates are rapid and reversible.

In principle, electrochemistry should be a useful probe of ion
binding in triaryl-X systems such as 1 and 86. Placement of a metal cation
close to the center of radical 86 would be expected to broaden and shift
both reduction waves to more positive voltages. These expectations are
qualitatively borne out, but the voltammograms obtained are so poor and
irreversible that we cannot identify well-defined reduction potentials for
the metal cation-containing complexes. In support of the idea that specific
binding is occurring, we note that no changes are observed in the cyclic
voltammograms when salts of larger cations such as Na+ or K+ are

substituted for the Li+.

 

 

 

 

 

 

 

 

 

153

3.3. Tetramethoxy Triphenyl amines

For the purpose of ion-binding studies of derivatives of radical 77,
structurally analogous amines in which the central carbon is replaced by
nitrogen have been synthesized. In NMR ion-binding studies, it has been
found that the use of metal salts with non-reducing counter anions such as
BF4- and C104; generate significant amounts of radical cations of triaryl
amine which can complicate data analysis by broadening NMR lines.
However, Scott Stoudt has demonstrated that use of I' salts effectively
protects against the oxidation process of the triarylamines; I- is easily

oxidized and evidently is sacrificed to make 12 instead of allowing the

amine to be oxidized to make the persistent amine radical cations. Thus, in
the case of LiI we were able to study ion-binding in CDC13. Furthermore,
iodide also appears to be a strong enough ligand in organic solvents that it
"caps" the tripod-bound lithium cation, filling out its tetrahedral
coordination geometry instead of giving way to allow a second tripod ether

to make an octahedral cavity.

The amines 87 and 88 shown below are isostructural to radicals 78 and
84. They were synthesized by a modification of Fréchet's method for
building triarylamines as shown in next page.162 These amines were used
in ion-binding studies with metal cations to demonstrate the abilities of four

methoxy groups as coordination ligands.

 

 

 

 

 

 

 

 

154
\o
X' 0—
‘ 1
N X \.. N
X' 0 0—
’0
872 X = C1, X'= H 89
88: X = H, X'= OCH3
X I X' \0
X' X'
D ’ O\ c K (:0 ‘ —
u. o L
+ 2 I, N((CH2)20(CH2)20Me)3’ X D N 0
NHz X' IO 2

They show interesting binding properties which can be detected by NMR.
Similar to the ion-binding experiments with borane 63 and radical 1, these

triaryl amines can be used for more quantitative studies. Addition of salts

to CDC13 solutions of amines in NMR tube results in significant chemical

shift changes (0.2 ppm); D20 washing regenerates the original spectrum of
triaryl amines. Figures 36 and 37 show 1H and 13C NMR spectra for

such experiments using amine 87.

Other salts including Na and Cd did not induce changes in NMR spectra.
Figure 38 shows an ion-binding study of amine 88 by 1H NMR which has
internal reference built into it. Our expectation for ion—binding of
tetramethoxy derivatives and dimethoxy benzene suggest that the two
methoxy groups in the 3,5 positions should not participate in ion-binding.
To our surprise, even those two methoxy groups show chemical shift

changes after adding Lil to the CDCl3 solution of amine 88. These

 

;—

 

 

 

 

 

 

 

 

155
findings may reveal conformational changes by ion-binding and suggest

that the tetramethoxy amines in solution may not be preorganized for ion-
binding. The possibility of participation of the central nitrogen atom's

lone pair electrons is still an unresolved question.

The radical cations of these triarylamines are isoelectronic to radicals 79
and 84 and show interesting redox properties. Those radical cations might
offer alternative building blocks for magnetic chains with the required
stability and structures. Their redox properties were therefore examined
by cyclic voltammetry. Electrochemical data are summarized in Table 12;
Attempts to conduct ion-binding studies by CV were unsuccessful due to

the fact that the oxidations of these amines were not reversible.

Table 12. Summary of electrochemical data for oxidation of
——__¥ +. . . .
N ‘———— N in amrne 87, 88, and 89 1n CH2C12.

 

8 7 +0.50 V
8 8 +0.44 V
8 9 +0.57 V

 

CH2C12 Solvent; Reference: Ferrocene oxidation

 

 

 

 

156

3CL40ME AMINE +LII 1 020

—7.2400
16.5628
15.5351

JUL _ -___L -- -- L

_ 11. 10 a - -1,
. in 15' t ,

TTTTTTT TI T I I T I T I I Y I T I I. I Y I V . I I T
6.0 5.5 5.0 4.5 4.0 3.5 PPM

 

 

 

 

 

 

' 1
3.0

Figure 36. 1H NMR spectra for the complexation of 87
with Lil in CDC13 showing chemical shifts of methoxy and

corresponding aromatic protons followed by D20 wash.

 

157

 

W

 

 

 

 

rmr" " '1" """ 1"7'I'T'Yfirer'V‘l'V'l VIII I'II Irrv vva IIIrrIIII 1111 1'11 rITI n” 1+1...fir
160 151) 140 130 1211 110 l 101) I 911 811 I 70 I 56 9911
Figure 37. 13C NMR spectra of 87 with Lil in CDC13

showing chemical shifts of methoxy and corresponding
aromatic carbons.

158

 

 

 

 

 

 

 

 

<— 3,5-OMe
2,6—OMe —>-
L A LL (1

. .- - - .. _ - -_.. __.-__ -.._... --... L..._______ L__

_ LA ii _ 1L___ __ L_

[7W v , . . r f . - ‘7 H
I 1 l "'1»"'i""1"7 I‘ll‘fl—YTIY1‘. ;r: 11111 .fit.III‘.I1111.‘V.1'Itr!'v.‘..
10 9 8 7 6I 5 4' 31 I 2' I 1l PPM

Figure 38. 1H NMR spectra for the complexation of 88
with Lil in CDCl3 showing chemical shifts of methoxy and

corresponding aromatic protons.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

———— 7 F
W| .. . . I * ‘-.vr|r*rw.;rfrrl7rIrrrT1I IITfifi.Y—fl..firIII—r—frlf.IIj—frrfi—i—rf7—flfi7—'fi‘.fi'—;—r I
800 100 160 140 120 100 80 60 4o 20 PPM 0

Figure 39. 13C NMR spectra of 88 with Lil in CDC13

showing chemical shifts of methoxy and corresponding
aromatic carbons.

 

 

 

 

 

 

160
3.4. Ammonium Complexes by Hydrogen bonding

As a way to construct desirably substituted radicals like
perchlororinated triarylrnethyl radicals, we were working on the route for
making the radical through tris(2,6-dimethoxyphenyl)methyl ammonium
tetrafluoroborate 72 which is not as sensitive to protonic acids as the
corresponding carbinol and can be converted into the radical by a simple
diazotization reaction. The X-ray structure of the given ammonia adduct
72 shown in Figure 40 shows three hydrogen-bonding interactions,
which hint at the hydrogen bonding ability of the ether tripods in radical 1.
An attempt has been made to determine the ability of radical l to hydrogen

bond with ammonium salts by using the borane 63 as a model compound in
CD3CN as a solvent.

 

To our suprise, after adding the salts into a solution of the borane 63 in a
NMR tube, the characteristic proton peaks (equal-intensity triplet from
coupling to N) of NH4I appeared which corresponded to eight protons per
molecule of borane from the integration of the peaks. Intrigued by this
fact, we tried a similar experiment on hexachloroborane 64, resulting in
similar behavior except that in this case we can assign only four ammonium
protons per molecule of the borane 64. The X-ray structure of the

hexachlorinated ammonium salt 75 has recently been solved; the aryl ring

 

 

 

 

 

 

 

 

161
twists have turned out to be 37, 51, and 51°. As we expected, the methoxy

groups in 75 are twisted down out of the aryl ring planes, improving
preorganization for binding; this twisting (the average of 67°) can be seen

in the following stereo view of the X-ray crystal structure of 75.

a)

b)

 

Figure 40. Stereo views of crystal structure of a) tris(2,6-
dimethoxyphenyl)methyl ammonium tetrafluoroborate 72
b)tris(2,6-dimethoxy-3,5-dichlorophenyl)methyl ammonium
tetrafluoroborate 7 5. Note that at the time of writing, the
ammonium H atom positions had not been finally refined.

 

 

 

 

 

 

 

 

 

162

Table 13. Crystal structure determination and refinement data
for tris(2,6-dimethoxyphenyl)methyl ammonium tetrafluoroborate
72 and tris(2,6-dimethoxy-3,5—dichlorophenyl)methyl ammonium
tetrafluoroborate 75.

 

 

Crystal data 7 2 7 5
Space group P2/n P21 /n
Z 4 4

Temperature 296 K 298K
a (A) 15.292 12.03
b 10.796 24.07
c 15.731 12.1 1
[3 109.99 94.38

11 (cm-1) 9.7 (Cu K01) 0.8 (Mo K01)

 

 

 

 

 

 

163

14321
.:;;.223g

- 1.9299

.-351
..217

‘,
b54911;
L_
r”
\L 1 o

 

 

 

 

 

 

 

 

I 'l;"% 1 1 T I [W 117 I'fiIT
‘rrVV'fiIFV 1
‘0 9 H I s 5 4 3 2
.J
115 26.0 ”'9 P";

Figure 41. 1H NMR spectrum of complex of 63°NH4I in
CD3CN showing characteristic triplet peaks after the addition

of the salt which was not present before.

164

33801—

(3.
CI)
I

9299

0

It.
rx
1
—- — —-~—-—————-—-— - -~»—_--l-h—n-—J -
IIIIIIIjIIWW'IIII I II IIIIIIIIII
8 7 2

1"filr"' m I I WTYUTI'fi'j—Iljlmn‘rtiUT'IWT[YrT_Y_
6 5 4 _3 19PM

'—1 5—3——J '
14.1 I .4 1%.!

—1

 

0"I‘ITTIIIIIrrTrI

Figure 42. 1H NMR spectrum of complex of 64-N H41 in
CD3CN showing characteristic triplet peaks after the addition
of the salt which was not present before.

 

 

165

3.5. Metal Ions

A variety of new salts have been investigated in reaction with 1 and
the related radicals described above. We have examined a variety of salts,
varying both metal cations and their anion partners. Solid phase SQUID
and EPR studies at low temperatures were performed on all complexes,
giving detailed information about the magnetic properties of these
materials. As with X-ray diffraction, single crystal studies would be
particularly valuable, as they would provide data on the magnetic
anisotropies in such crystals. Howerver, among the magnetically

interesting materials only polycrystalline solids were obtained.

Dirners and oligomers of types 2 and 3 can obviously be envisioned
with metal ions other than lithium. In addition to the few salts that we
have already investigated, we have examined various transition metal ions,
including Cr3+, Co3+, Cd2+, and Rh3+, all ions with strong affinities for
ethers and with coordination spheres of appropriate size to bind to the
ether tripods in 1. In a qualitative survey of roughly 20 alkali metal,
alkaline earth, and transition metal salts, we examined the color changes
due to radical oxidation in the presence of metal salts and Oz. A clear
trend emerged, with metal ions of small radius (S 1.1 A)163 promoting the
rapid color change and larger ones failing. These observations are
consistent with the observed size of the binding pockets in 1 and 63 and we
expect such survey data to be useful in helping to select new salts for

further study.

 

 

 

166

The counter anions of the salts used in our complexation studies may
be critical to the successful formation and isolation of crystals for X-ray
studies. Crystallization of the extended (radical-metal)n stacks that we
envision may be constrained by the requirements of fitting together such
infinite charged structures in an efficiently packed lattice. Furthermore,
the metal ions' charges are spaced at fixed intervals along such stacks or
chains, as determined by the cation's radius and the aryl ring twist angles in
the radicals. To maintain charge neutrality, the counterions must find
appropriate sites with the right spacing to allow the chains to pack together
in one of the relatively few simple patterns available to them. Counterion
size, shape, and charge may therefore play a key role in determining which

systems of salts and radicals will lead to well-formed crystals.

 

 

 

 

167
CHAPTER II. Magnetism of Triarylmethyl Radical Complex

1. Calculations on the Metal-Mediated Pairwise Interaction of
Methyl Radicals

Professor Jackson has examined and compared the total and one-

 

electron frontier orbital energies of calculated singlet and triplet states for
the simplest model system, H3C-Li+-CH3. The linear interaction of two
carbon 2p-orbitals on radical centers, communicating via the 2s and 2p
orbital set of a lithium cation, can be described by a simple qualitative

interaction diagram, shown in Figure 43.

 

. A10
H3C° LI+ °CH3

Figure 43. Qualitative MO diagram schematically indicating the

coincidental near-degeneracies from mixing a pair of sp3 hybrid
orbitals with a Li2pz orbital and a LizS orbital.

   

 

 

168
Preliminary work on the metal-mediated radical pair H3C~Li+oCH3

(3-21G and 6-31G*), suggests a triplet ground state for this model system.
Geometry differences between singlet and triplet optimized structures are
small, and vibrational analysis on the triplet shows that the linear form is a
local minimum. The fact that the calculated total energies for the two
spin states are close is surprising, and we need to pursue more complete
wave functions which will determine the ordering and separation of the
two Spin states as a function of C-Li distance. In conjunction with our
molecular mechanics and semiempirical molecular orbital modeling of the
actual radicals under study, we expect these higher-level calculations
ultimately to lead to a set of generalizations which will help guide the

design of new paramagnetic complexants and extended structures.

Techniques such as MNDO, AMl, which are widely available and
easy to use, provide a reasonable middle-ground for attempts at studying
large systems such as radical 1 and its complexes. Reliable absolute
predictions of preferred electronic couplings might not be possible using
these methods, but they should help to predict trends and guide our choices
of modified systems. Professor Jackson has fully Optimized the structure
of radical 1 by MNDO and AMl within D3 symmetry, and its structure is
nearly identical with that obtained by the molecular mechanics methods
using PCModel. The program PCMode1164 offers molecular mechanics
calculations based on MM2 force field. Calculations using the program
have been shown to work well in predicting geometries for the triaryl-X
monomers whose structures we have studied by X-ray. As a rapid source

of reasonable estimates for geometries and strain energies, this program is

a valuable tool.

 

 

 

 

 

 

169

2. Antiferromagnetic Interaction and Peroxy dimer of The
Radical 1

Magnetic susceptibility measurements on crystals of radical 1 show
very weak antiferromagnetism according to Curie-Weiss behavior. It is
not clear what is the mechanism responsible for the coupling, but careful
examinations of the three-dimensional packing of the radical show a closest
distance of 4.20 A between para carbons of the flattened conjugated aryl

rings of the radical.
Curie-Weiss Fitting of HMTP Methyl Radical

 

200

y = 8.807] + 3.1449x R"2 = 1.000

0 =-2.8K
J=2k0 = -7.73e-23 J
=-3.88 cm-l

“X 1001

 

 

 

 

- ' I ' I '
0 10 20 30 40 50
TEMP (K)

Figure 44. Curie-Weiss fitting of HMTP methyl 1 with Curie-
Weiss constant of -2.8K (J = -3.88 cm'l) suggesting weak anti
ferromagnetic interactions between radicals.

Radical 1 reacts slowly in the air to form peroxide 90 (shown on
next page) as determined by 1H NMR, FAB mass spectrometry, and

elemental analysis. The formation of peroxide can be explained by the

 

 

 

 

 

170
reaction of a spin-rich para-carbon in the radical 77 with oxygen.

According to McConnell's approximation there is a significant amount of
unpaired electron spin density on that para-carbon which might be
available for the reaction. Since it is known that oxyradicals add to p-
positions of triarylrnethyl radicals, this reaction is the point of initial
oxygen addition. The tetramethoxyphenyl methyl radical 77 shows similar
dimerization to form a peroxy dimer with an interesting head-to-tail
connectivity which is different from the tail-to-tail mode found for radical
1. With only four methoxy groups for steric blockage, radical 77 does not
have enough steric protection against the penetration of oxygen to the
central carbon so that it forms a peroxy dimer linked between a central
carbon and the para carbon in that asymmetric third ring. We are in the
process of determining the first step of the reaction by reaction of radical
77 in a hydrogen donating solvent; The preliminary results suggest that
the first step is the attack of oxygen at the para carbon. The X-ray
structure of the peroxy dimer (J acobson's ) 91 from radical 77 is shown in

stereo view in Figure 45.

 

Figure 45. Stereo View of X-Ray structure of head-to-tail
peroxydimer 91.

 

 

 

 

171

3. Magnetic Behavior of Radicals and Complexes

We believe that we have made magnetic materials in polycrystalline
form from 1 and various lithium salts. I studied these blue-black powders
by standard methods (SQUID, EPR) for magnetic susceptibility
determination as a function of temperature and magnetic field strength. In
our experiments, ferromagnetic behavior165 was reproducibly observed in
the blue-black powders obtained when solutions of 1 and LiBF4 were
mixed and evaporated to dryness under an argon atmosphere. Studies of
magnetic field dependence in these materials showed hysteresis (Figure
46) a classic signature of bulk magnetism with far less spin counts than
theoretically expected. The tetrafluoroborate anion was replaced by the
similar-sized perchlorate ions without qualitatively affecting the field
dependent magnetic properties. This result indicated that the bulk behavior
is not due to chemistry involving the counterion. Radical 1 alone, a red
powder, showed weak antiferromagnetism, in contrast to the salt-treated
samples; lithium salts alone showed diamagnetic behavior, suggesting that

the observed ferromagnetism must somehow derive from the combination

of the radical with the lithium salt.

 

 

 

 

 

172

 

6O

40"

20"

  

M(emu/mol) 0 q

   

1-LiBF4

-20 ‘

-40 -

 

 

 

‘60 j I ' l I 1 I I
-6000 4000 -2000 O 2000 4000 6000
H (G)

Figure 46. Plot of magnetization M (in EMU) vs. magnetic

field strength H (in Gauss) at 5K for Radical 1 alone and for its
complex with LiBF4.

Iron from the laboratory environment is a frequent ferromagnetic
contaminant in studies of magnetic materials. In light of the reproducible
SQUID results above and our fear that contamination with iron or other
ferrous metals might be causing the observed hysteresis, we devised a
metal-free synthesis of radical 1. Martin's original synthesis used Cr(II)
ion as the reducing agent; we showed that zinc works well too. However,
we sought to avoid metals, such as Cr or Zn, that could carry ferrous
contaminants. Ultimately, we discovered that ascorbic acid (vitamin C) and
Iodide salts are both able to reduce salts of the cation in THF; the neutral
radical then migrates into a nonpolar organic solvent layer, where it is

easily separated and isolated. Various attempts were made to eliminate

 

 

 

 

173
potential iron contamination in our SQUID samples; careful filtration of

radical and salt solutions before mixing, use of glass distilled solvents and
high purity salts, chromatography of the starting radical solution over
silica gel, etc. A series of SQUID samples were analyzed by atomic
emission photometry which showed some iron at the level of the
background, regardless of their SQUID behavior in all samples and
controls. However, the levels found would be more than sufficient to yield

the ferromagnetism observed, if the iron was in a ferromagnetic form.

3.1. Diamagnetism of Triaryl Methane 73 and Metal Salts

The Pascal constants provide an empirical method for estimation of
diamagnetic corrections. For the complex 1-LiBF4 the diamagnetic
contribution is calculated to be -2.89 x 10'4 emu/mol. Greater accuracy
can be obtained by the direct measurement of the susceptibility of a

diamagnetic analogue of the paramagnetic compound which is of interest.

For example, a diamagnetic correction for radical 1 can be achieved more
accurately by a subtraction of magnetic measurement data from the
corresponding triaryl methane 73 discussed earlier (section 2.4 of chapter I)
which has one more hydrogen than the radical. Furthermore, by this
process of diamagnetic correction we were able to eliminate the error
factors inherent in the SQUID measurement which come from the method of
sample handling, etc. The observed diamagnetic correction for radical 1
using the methane 73 gave similar diamagnetic contributions at the high
temerature region. Figure 47 shows a typical field dependence of methane

73 at 2K showing characteristic diamagnetic behavior.

 

 

 

 

 

 

 

 

 

174

 

4-1

M (emu/moi) 0'l

-2-1

 

 

‘6 ' I ' 1 I fl I
-6000 -40 -2000 0 2000 4000 6000
H (G)

 

 

f I '7

1))

 

0.4

0.21

M (emu/mol) 0.0-

-0.2 "

 

 

 

-0'4 —I 1 f If If r r7 I I '

-6000 -4000 -2000 0 2000 4000 6000
H (G)

Figure 47. Field dependence of a) methane 73 at 2K and b)
CdC12 at 5K. The induced molar magnetic moments decreases

as the field strength increases which is the characteristic field
dependence (inversely proportional) of a diamagnetic material.

 

 

 

 

 

 

 

 

 

 

 

175

3.2. Paramagnetism of Radicals and Radical Complexes

As I discussed in the introduction, a paramagnet concentrates the
lines of force provided by an applied magnetic field and thereby moves
into regions of higher field strength. Paramagnetic susceptibility is
generally independent of the field strength, but shows temperature

dependence according to the Curie Law.

Radical 1 shows a paramagnetic temperature dependence over a wide
temperature range; it is weakly antiferromagnetic at low temperature,
although the coupling is not a big one. We can show the paramagnetic
properties of the radical in two different ways using the plot of 1/x vs T
and xT vs T. Figure 48 show such plots with a slight antiferromagnetic
downturn at low temperature. Most of HMTP methyl radicals behave

similarly to radical 1 in their temperature dependencies.

 

 

 

 

 

 

 

176

 

1000

800 '

600'

1/X (emu/mo!)

400'

200 "

 

 

 

100 200 300
Temperature (K)

5' OT‘

 

0.4

0.3 ‘

xT(emu/mol) 0,2 J

 

0.1 "

 

 

 

0.0 - . fi —r f .
0 100 200 300
Temperatue (K)

Figure 48. The temperature dependence of radical 1
analyzed in a) l/x vs T and b) xT vs T at 5000G with Curie
constant of 0.34 emu/mol which corresponds to 91% of ideal
S=l/2 radical. Presence of the diamagnetic peroxy dimer 91
reduces the total spin count.

 

 

_ 177
The radicals 86 which I discussed earlier (section 3.2 in chapter I)

were synthesized to build an analogue of 1 which could not form extended
chains by pairing around metal cations. Radical 86 offers a bound Li+ ion
a full coordination sphere of six ether oxygens, effectively "capping" the

radical on both faces. Thus, the radical center in 86-2LiBF4 is

encapsulated and should behave as a completely isolated paramagnet.

\o \o‘
[O (0.99
(Q 3: 2 LiBFg (Q 3 . 2 BF;
0 0‘
ea 9..
86 86°2LiBF4

As shown by CV data earlier, the radicals 1 and 86 are respectable
reductants with nearly identical redox potentials. Fears about impurities in
our studies of radical 1 with salts led us to speculate that even though no
ferromagnetism was found in our starting materials, something
ferromagnetic might be generated by reduction of a diamagnetic impurity.
Radical 86 should be just as likely to effect the hypothetical reduction as 1,
but magnetic measurements shows that the complex 86-2LiBF4 is a
textbook paramagnet. In fact, these samples are among our best instances
of isolated S=1/2 radical behavior. Thus, blocking chain formation in 1
significantly alters the nature of the LiBF4 complex products, consistent
with the notion that such chains or at least the open faces are important to

the complexation chemistry of radical 1. We have verified this

 

 

 

 

 

178
expectation by magnetic susceptibility studies of powders of the complex.

Figure 49 shows plots of magnetic susceptibilities as a function of applied

magnetic field and temperature. As expected for an isolated radical, a

Curie plot (l/x vs. T) shows ideal paramagnetic behavior with a ueff of
1.56 1.113. We attribute this slightly low value to uncertainty in the exact
radicalzLiBF4 ratio, and to a small amount of radical oxidation during
preparation. There is no indication that adding LiBF4 to 86 puts the
radical centers into communication. This finding is in contrast to the
situation with 1. Like 1, radical 86 shows some tendency toward oxidation
on treatment with LiBF4 under the air. This behavior is evidenced by the
appearance of the blue color of the corresponding triarylrnethyl cation in
each cases. This result reinforces the notion that the ferromagnetism seen
from a lithium complex of radical l is unique to a radical which can bind
metals on both open faces to form chains, and is not simply due to an
artifact such as iron particle production by reduction of adventitious iron

salts.

 

 

 

179

 

400

200 7

M (emu/mol) 0 '

-200 ‘

 

 

 

-400 I I I r I I F I I I
-6000 -4000 -2000 0 2000 4000 6000
H (G)

r

b)

 

1000

800 ‘

600"

l/x

400 ‘

200‘

 

 

 

0 a T 4 a 4
0 100 200 300
TEMP (K)

Figure 49: (a) Plot of M vs. H for 86-2LiBF4 at 5K; (b)
Plot of 1/x vs. T for 86-2LiBF4 at 5000G showing

paramagnetic behavior.

 

180

In the course of the NMR ion-binding studies, we developed a more
detailed understanding of one counterion's role as a partner and ancillary
ligand to Li+ bound in the tripod ether's pocket. These insights have
translated into an explanation of the simple S=1/2 paramagnetic behavior
found in magnetic susceptibility studies of 1-2LiI complexes. If the radical
is capped by Lil subunits, it is sensible that it should behave as an isolated

paramagnetic center like complex 86c2LiBF4. Figure 50 shows

paramagnetic behavior of complex 1-2LiI at 5000G.

 

1.50

1.00 -
ueff ( 11B )

0.50 ‘

 

 

 

0.00 . . ‘ ' ' '
0 100 200 300
Temperature (K)

Figure 50. Temperature dependence of complex l-2LiI at
5000G showing simple paramagnetism of the complex (81% of

spins are counted).

 

 

 

 

 

181
By the same token, if the radical is capped by NH4I subunits,

complex l-N H41 should behave as an isolated paramagnetic center like the

case above. The temperature dependence of this complex is shown in

Figure 51.

 

l/x

 

 

 

 

O 100 200 300
Temperature (K)

Figure 51. Temperature dependence of complex l-NH4I at 500G

showing simple paramagnetism of the complex by effective capping
of the linear chain formation.

At very high external fields and very low temperatures, the

magnetization M becomes independent of field and temperature, and

approaches the maximum or saturation magnetization Msat which the spin

system can exhibit. This situation corresponds to the complete alignment

of magnetic dipoles by the field (Msat=5.58 x 103 erg/Os mol for g=2,
S=1/2).

 

 

 

 

182

 

 

 

 

8
7 b
: ...............U
I ..
6 _- .0.
O AAAAAAA‘9446AA”
5 - o. .e‘“
C A‘ +++++++++
: A ++++++
M/NuB 4 '. ' .
E 0 A +++ .0000000009"'”
I A + ....
3 r 04 + .0. 000000000
' i ° 00000000
A O
2 -o * ° 0000 ...o
:°+o 00 °°°°°°°°°°°°
1:-‘”o°
000 ......... °°°'
Ito.° °°°°°°°°
'§:2°l°.. l l l
o l llllllllllllllllllll
0 05 1 1.5 2 25 3
71

Figure 52. Plot of magnetization per radical, expressed in Bohr
magnetons, against 11 for S = 1/2 to 7/2 (bottom to top).

It is convenient to calculate the magnetic moment of a magnetic system

with arbitrary spin-quantum number. The average magnetic moment of

one radical can be calculated using the Brillouin function B301) which is
defined below for N non-interacting atoms.

Msat = ngBSBS(n)

8113 E;

n=kT

1 1 1 1
35(1)) = —§[(S+ -2—)coth(S + -2-)n — Ecom—g]

N is Avogadro's number, g is Landé's constant, S is the spin-quantum

number.

 

183
The mean magnetic moment <uz>, or magnetization can be written as

below. A plot of magnetization per radical expressed in Bohr magnetons

11B, against 1] (dimensionless parameter directly proportional to

fieldstrenth) for S: 1/2 to 7/2 is shown in Figure 52.
< #2 >= gflBSBSU?)

M = N < 0, >= NgrtBSBsW)

Most of our radicals do not show saturation magnetization even at
55000G at 1.8K which are the field and temperature limits of our SQUID
at Michigan State University. However in the case of the complex l-LiI, a
near perfect fit is found to the Brillouin function. (Figure 53 ). Many of
our radicals might show saturation at higher than 55000G using a pulsed
magnet. In some of the organic radicals such as the DPPH complex with
benzene which forms a linear chain in the solid state, Gerven used a pulsed

magnetic field to achieve saturation of the complex at 13.4 T, more than

twice of the field strength of our SQUID! 166

 

 

 

184

 

M/NuB

 

 

 

Figure 53. The saturation magnetization of the complex l-LiI
at 1.8K. The data were fitted to Brillouin function using
Kaleidagraph 3.0 with S=0.62 and R=0.9999.

3.3. Ferromagnetism of Complexes

Ferromagnetic behavior was observed in the precipitates obtained
when solutions of 1 and LiBF4 were mixed under argon atmosphere.
Radical 1 alone, a red crystal, showed a weak anti ferromagnetism, in
contrast to the salt-treated samples; lithium salts alone showed diamagnetic
behavior, suggesting that the observed ferromagnetism must somehow

derive from the combination of the radical 1 with the lithium salts. Along

the same line, the complex of radical 1 with ZnC12 shows a hysteresis

curve with less saturation than complex l-LiBF4 as shown in Figure 54.

 

 

 

 

185

Our results are reproducible despite variations in synthesis of radical
1, careful control experiments on starting materials, and filtration to
remove any magnetic particulates. Atomic emission experiments yielded
inconclusive results for the presence of ferromagnetic impurities. An ac
susceptibility measurement, which is now available with the new SQUID at
Michigan State University might also give better handles when we have

structurally well defined materials.

 

 

 

186

 

 

 

 

 

 

 

M (emu/mol) 0

4

/

 

1

 

 

/

 

 

-3000

——v

 

-1500
b)

 

0
H (G)

 

 

1500 3000

 

15

10'

5-

M (emu/mol) 0-

-5-

—10 ‘

 

 

 

-15

-5500

f—I

-1500

f m

-3500

500
H (G)

f —v

2500 4500

Figure 54. The field dependence of a) the complex l-LiBF4 at 5K
and b) the complex loZnClz showing characteristic ferromagnetic

hysteresis curve.

 

 

 

 

187

 

3.4. Antiferromagnetism of Radical Complexes

According to our original scheme for building magnetic chains,
control over magnetic properties can be achieved by varying the metal
cation, and thus, the distance between radical centers. The first
consideration in selecting metal salts was the size of the metal cation and
the availability of unpaired electrons in the metal. The complex of triaryl
methyl radical and a paramagnetic metal cation might give us better model
to study the linear combination of hetero Open shell environment. Various
alkali metal salts were the logical choice at first and some first row
transition metal salts like CuC12 and CrCl3 were tested for possible
magnetic chain formation. None of the sodium or magnesium complexes
with radical 1 showed any interesting magnetic behavior except simple
paramagnetism which suggests that the sodium cation is unable to induce a
magnetic interaction to occur. The complex of l-CrCl3 again showed
simple paramagnetic behavior with far less spin count than expected. The
significance of this observation is not clear. Perhaps the metal caps the
radical center as in the case of l-LiI and loN H41 and there are
antiferromagnetic interactions between the unpaired electrons of the metal
and the radical. It is not clear what kind of interactions are involved

without knowing the solid state structure of the complex.

The complex 1-CdC12 shows classical antiferromagnetic temperature
dependencies with xmax at 25K which may be explained with an alternating

linear chain Heisenberg model. Attempts were made to fit the temperature

dependence data with various models. The coupling constant J varies from

 

 

 

 

 

-10.2 to -19.7 cm-1 depending on the model used (Figures 55 and 56).
The Bleaney-Bowers dimer model with paramagnetic impurity correction
gave an inter-radical coupling constant of J = -l3.5 cm-1 (-19.4 K) with
5% paramagnetic impurity (R =0.9998). Calrin‘s mean field correction
of the Bleaney-Bowers model (described in the introductions) with
paramagnetic impurity correction gave J = -13.6 cm'1 (-19.5 K), J'Z= 11.4
cm“1 (16.5 K) where Z is number of neighboring dimers with 5.6%
paramagnetic impurity (R = 0.9998). This alternating dimer model (chain
of dimers) can be explained as an alternating Heisenberg linear chain when
Z is 2 with an interdimer coupling constant of J ’= 5.7 cm-1 (8.3 K). It is

still not clear which model is the best for the complex l-CdC12 without

having the solid state structure of the complex.

The anti-ferromagnetic behavior of the complex l-CdC12 was reproduced

several times; this combination is unique in the sense that the complexes of

radical 1 with CdBr2 and Cdlz only show simple paramagnetic
temperature dependencies, as do those with NaBPh4. Numerous attempts
to grow a single crystal were unsuccessful, with only polycrystalline

materials Obtained. X-ray powder patterns (Figure 57) of the radical

and the CdClz complex show clear differences but do not provide real

structural information for this unique system.

 

 

 

 

 

 

 

189

 

 

0.03

“emu/mol)

 

0.02

   
 

1

  

0.01 r-Cdcn

 

 

 

0 50 100 150 200 250 300 350 ,7,
Temperature (K) '

 

0.01

0.008

0.006

X (emu/mol)

0.004

 

 

 

O [JUIJULIUUIJULJILHLILILIIIU

0 50 100 150 200 250 300 350

 

Temperature (K)

Figure 55. The temperature dependence of a) the radical 1 and
complex 1-CdC12 at 500G; b) with Bleaney-Bowers dimer model

with J = -13.5 cm'1 (-19.4 K), 5% paramagnetic impurity, and R
==0.9998.

 

 

 

 

190

 

0.01

0.008

0.006

X (emu/mol)

0.004

 

 

0.002

 

 

0 LUIJJIJJIJJIILIILIIIJIIJWLIJJU
O 50 100 150 200 250 300 350

 

Temperature (K)

 

0.008 - 000
0.006
x (emu/mol)

0.004

0.002

 

b

 

 

0 j[JIJLI”DULJIJIHILILILILILILIHI

0 50 100 150 200 250 300 350
Temperature (K)

 

Figure 56. The temperature dependence of the complex 1-CdC12

at 500G a) with alternating dimer model with J = -13.6 cm'1 (-l9.5
K), J'= 3.8 cm-l (5.5K), Z=3, 5.6% paramagnetic impurity, and R =
0.9998; b) with linear chain Heisenberg model J = -19.7 cm:1 (-
28.5K) and R = 0.9884.

 

 

 

 

191

1800 % El)

o—«v 1‘

1200

600

 

 

..—...——.-._..~_.—._’-. -—

    

5 1o 15 20 25 30

 

 

 

2 THETA RISAKU / USA
8000 -
' 1
I
4000 4 3 4 7 'o
e .
6
' 9
5
__ 4 AA A A A A‘
o fiff vrjri—va—vy—v—Vfifivvyvv —v rvIfifiv —rfw V—fj
5 10 15 20 25 30
2 THETA HIGAKU / USA
4
1500 * C)
'. 2
1200 + s
600

 

 

   

 

o t—‘raT—vfifi—va—‘~Vfifv—r’f. .yv .‘ .v v v 1
5 10 15 20 25 30
2 THETA RIGAKU / USA

Figure 57. X-ray powder patterns of a)CdC12, b) the radical
1, and c) the complex l-CdClz.

 

 

192
An attempt was made to find information on the electronic coupling of

Cd(II) and the radical center by ESEEM using radical 86. Although
significant new features were observed, the spectra were too complex to
yield any meaningful answers. The experiment is complicated by the fact
that there are two different magnetic Cd isotopes (111 and 113), roughly
equal in natural abundance, and a third non-magnetic isotope (112) that
makes up the majority (~75%) of the element's composition. It would be
very useful to have the structure of the complex in order to understand the

mechanism of the magnetic coupling between CdC12 and 1. There are two

 

possible coupling routes to form an alternating dimeric chain: a) the
originally intended linear chain formation through metal ion-binding
oxygen pockets to form a chain; b) simple Cd-mediated linkage from para-
carbon to para-carbon in two radicals as shown below; or both. The
magnetic data from CdBr2 and Cd12 suggest that the mechanism shown
below may not be the right one. The magnetic measurements on complexes
of other para substituted radicals such as 78 may give more

straightforward answers to this question.

\ /

0 o
f“ $1
@o O ‘1“

0 C1

()0
l \

 

 

4. Design of Molecular Magnets by Other Ion-Binding

Recently, we found that treatment of TEMPO with ammonium or

lithium salts leads to high-spin coupled materials. EPR spectra for these

 

 

 

193
systems Show strong half-field transitions, diagnostic for high spin-

coupling. No structural details are yet available, but the Lil complex of
TEMPO forms crystalline long needles, and we are optimistic that it might
provide a probe for the long overdue question of pairwise radical coupling

directly through alkali metal cations.

< N—OIIIM+IllO—N >
< N—O”""H— +-HIII||IO'—N >

It is tempting to envision pairwise radical interactions through a metal

unnflj—Z—munu

cation in the case of TEMPO and DPPH. In contrast, the resonance
delocalized structures of galvinoxyl and nitronyl nitroxides suggest that
these subunits may be able to participate in forming extend structures.

Special attention should be paid to the search for triplet or higher spin

transitions in the EPR spectra and susceptibility measurements using
SQUID.

 

 

 

+ O -O--M+- O-N O--M+~
H H
£1
----- H—N+ ---H O~<::E\I-O-m-H--N+ H"- -0- $70.12? -H—N+ -H-----

Furthermore, other commercially available stable organic radicals such as
TEMPOL can be used to probe this problem including the effects of the
counter anions of the coupling salts or the formation of extended structures

such as those shown above.

 

195
5. Introduction of Cross-Linkers

To obtain materials which exhibit bulk ferromagnetism, high-spin
coupling of the constituent paramagnets must occur in three dimensions.
Our picture of stacked radicals sandwiched around metal ions provides a
one-dimensional coupling scheme, but how can the expected stacks be made
to communicate magnetically ? In the few known molecular ferromagnets,
this issue has simply resolved itself without explicit action being taken to
control it. However, to understand the requirements for stack-stack
coupling in detail, we planned to examine both intra- and interstack types
of electron-electron coupling by explicitly building biradical systems to

model and/or induce these interactions.

By isolating the functional diradical unit 2 of our proposed polymeric
chains, the intrastack electronic coupling can be studied. We planned to
build two series of these diradical complexants; the triarylrnethyl radicals
linked through alkyl chains and meta-xylylene linked poly radicals. An
attempt to build the alkyl chain linked biradical shown below using

monomer 92 was unsuccessful.

"(OOHO cho3 __
2 6 BrMBr
HO
1

 

 

 

 

196
To avoid chain-chain cancellation of magnetic moments, an

interchain coupling mechanism is necessary. An approach to this
requirement is to build electronically and physically cross linked
monomers into the chains, enforcing interchain communication. In
contrast to 92, aromatic ring linked diradicals should be fairly rigid, with
little opportunity to vary the radical-radical distance, whether or not a
metal ion is present in the internal cavity. There are many possible model

systems which could be used in building three dimensional structures.

 

Among numerous possible aromatic linked biradicals, several phenyl
linked ones, such as diradicals 93 and 94 are synthetically accessible by
simple modification of usual procedure. According to the n-topology of
altemant hydrocarbon discussed in the introduction, the diradical 93
should be high-spin coupled and the diradical 94 should be low-spin
coupled. Radical 93 shows two reversible redox waves in the

cation/radical couple of -O.12 and -O.41V, and only one redox wave in the

 

 

 

197
anion/radical couple of -2.03 V in CV (Referenced to ferrocene oxidation;

Measured in CH2C12 solution of (n-Bu)4NBF4 ). The two reversible redox
waves in the diradical 93 presumably correspond to the radical cation and
the diradical which is characteristic evidence for the formation of a
diradical. The CW-EPR spectrum of the diradical 93 in THF was too
complex to analyze for hyperfine coupling constants at present.167 The

half-field transition of diradical 93 was also not detected under various

conditions.
Ar Ar
° ° Ar Ar
Ar Ar
Ar Ar
93 (Ar=2,6-dimethoxyphenyl) 9 4

A diradical such as 95 could also be used in building three-
dimensional structures with proper ion-binding ability. This system shows
a reversible redox potential of -O.44 V for the cation/radical couple and
-1.93 V for the anion/radical couple (Referenced to ferrocene oxidation;
Measured in CH2C12 solution of (n-Bu)4NBF4 ), and a CW-EPR spectrum
which is comparable to those of other para substituted TMTP methyl
radicals such as 78, 79, 80, and 83 which means this radical is not

antiferromagnetically coupled strongly as predicted by theory. A half-field

transition was not detected.

95 (Ar=2,6-dimethoxyphenyl)

 

 

 

198
A triradical such as 96 could also be used in building extended

structures by using two-dimensional frameworks extended into three-
dimensional structures by ion-binding. Characterization of such systems is
complex and the already detected half—field transition spectra of 96 at 77K
has a problem of interpretation due to the fact that it might also be the half-

field transition of the diradical cation or a diradical as shown below.

 

 

 

 

 

199
CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK

Conclusions

We have begun a new line of investigation into molecular magnetic
materials, which may ultimately extend our understanding of molecular
magnetism. Magnetic measurements (SQUID) suggest that we can induce
ferro~ and antiferro-magnetic behavior in complexes of stable organic
radicals with various metal salts. Cyclic voltammetry of HMTP methyl and
TMTP methyl radicals demonstrates our ability to adjust their electronic
character by variation of substituents. X—ray data for radicals 1,
triarylrnethyl cation 62, triarylborane 63, triaryl carbinol 71, triaryl
methane 73, triarylsilane 74, and ammonium adduct 72, and the
hexachloro ammonium adduct 75 have built a structural basis for
preorganized ion-binding. Complexation studies of 63, 64, 87 , and 88 by
UV—Vis spectrophotometry and NMR suggest complexations. ESEEM
studies of the truncated complex 86-2LiBF4, 86-2LiI, 86~2NaBPh4 , and
8602NaI show that two metal ions are bound at the expected distance with
spin densities of 0.3% on the metal cations, and molecular mechanics and
semiempirical molecular orbital calculations agree on the structure of the
complex. Ab Initio molecular orbital calculations offer an explanation for

the high-spin coupling in radical metal radical systems.

Two critical issues have focused our efforts in this research. First,
structural characterization of metal-radical complexes is essential to our

understanding of the rules governing long-range organization in these new

 

materials. Such an understanding is required for the rational construction

 

 

200
of organic magnetic substances. By using X-ray crystallography, we have

studied monomers with the binding sites of interest; we have also obtained
a certain amount of indirect (spectroscopic, electrochemical) evidence for
ion binding. Now more detailed geometrical insights into the complexation
phenomena are needed.

The second focal issue is the fundamental description of the metal
ion-mediated magnetic interactions between organic paramagnets.
Assuming our structural pictures of the extended complexation are correct,
how is a lone radical such as 1 perturbed by metal ion binding? More
importantly, how do two radicals 1, magnetically independent in the
absence of salts, couple around a metal ion to make an interrupted G-bond?
What is the sign and magnitude of the exchange integral between the two
paramagnetic centers? Can it be properly understood in terms of simple
molecular orbital pictures?

We have begun with magnetic, chemical, and structural
characterization of various triaryl methyl radicals and their complexes with
different metal cations. There is enormous structural flexibility in our ion-
binding approach to building molecular magnet, where the triaryl-X
platforms provide triether binding pockets for metal ions. There are
several controllable variables in the basic triaryl-X structure, and in terms
of the number of triether binding faces present. All of these variables
should be further investigated in order to optimize the selection of the best

candidates for assembly of stable extended complexes.

X-ray structures for complexes of radicals and boranes with metal salts are
desperately needed. Detailed magnetic characterization of well-defined

materials, ideally as single crystals, will enhance our understanding of the

 

 

 

 

201
relationships of structure and odd—electron coupling. By systematically

building up an understanding of the ion binding and magnetic properties of
these systems, we hope to establish this new approach as an effective
paradigm for the assembly of 1— and ultimately of 3-dimensional molecular

magnetic materials.

Suggestions for future work

1. Aromatic Linked Biradicals

We planned to attach a second hexamethoxytriphenyhnethyl radical
unit at the para position of an aryl ring of 1 to make diradical 97. Studies
on Chichibabin's hydrocarbon168 suggest that for the highly twisted 97 ,
open-shell paramagnetic electronic structures will be important, despite the
availability of closed-shell classical structures.169 In hopes of extending
our studies to the rational assembly of metal-organic complex solids with
bulk magnetism, the examination of the complexation and chain-forming

behavior of 97 , 98, and their derivatives will be very important.

 

 

 

 

202

   

98

2. Magnetic Coupling through Hydrogen Bonds

Numerous examples of macromolecular chemistry have been based
on frameworks of hydrogen-bonds.”0 Extensive studies have been
reported of the structural and electronic properties of such systems. The
detailed descriptions of orbital interactions through hydrogen-bonds lead us

to consider a strategy in building organic magnets.

Electron transfer and energy transfer through space and through bonds,
especially through hydrogen bonding have been pursued extensively in
recent years, and theoretical understanding of the interactions involved can

be explained by many theories.”1

Magnetic interactions through bonds and through space have been proposed
by Hoffmann and Gleiter in organic diradicals and in inorganic bimetallic
bridging ligand systems and the basic theory has been accepted for more
than 20 years.172 Magnetic interactions through hydrogen-bonds were
suggested by Carlin in 1985 in an attempt to explain magnetic susceptibility
data on a 2:1 complex of nitroxyl radical and Cu (11) complexes. In an

attempt to explain the structural and magnetic behavior of a TCNQ

 

 

 

 

 

'8'

 

203
complex of radical cation of 1,4-hydrazine systems shown below, Koch

reported at the 1993 Denver ACS meeting the first example of magnetic
interactions in an infinite organic chains in which the ferromagnetic

coupling was mediated by hydrogen-bonding interactions.

The possibilities of designing magnetic interaction through hydrogen
bonding are endless in organic radicals as illustrated by the numerous
macromolecular structures already reported. Any stable organic radicals
with proper functions for hydrogen bonding can be considered to be a
target for testing the proposed magnetic coupling. Simple derivatives of
TMTP radical such as radicals 82, 99, and 100 shown below (which we

have synthesized already) can serve as pairwise model systems in building

organic magnets.

 

 

 

204

 

3. Ion-Binding in Various Nitroxyl Radicals

As discussed in the introduction, the study of nitroxyl radicals is one
of the most active areas of molecular magnet research. The nitroxyl
radicals, especially diaryl nitroxyl radicals, offer logical extensions of our
strategy in building molecular magnets by ion-binding. The 4,4'-
dimethoxy diphenyl nitroxide shown below has never been studied for
molecular magnets even though the X-ray structure with a packing diagram

has been known for more than 40 years. Magnetic measurements on a

 

 

205
single crystals of the radical might give insight on the mechanism of inter-

radical interactions in nitroxyl radicals.

O,© :O

There are several advantages of using nitroxyl radicals compared to triaryl
methyl radicals: The Lewis basicity of nitroxyl as demonstrated by
Drago in TEMPO; spin polarizability of nitroxyls; inherent stability of
diaryl nitroxyls when they are protected against CO dimerizations by para-
substitution and twisting of rings by ortho-substituents. Possible linear
complexes of bis-(2,6-dimethoxyphenyl nitroxyl and bis(2,6-pyrimidyl)

nitroxyl with metal ions are shown on the next page.

\0 (I) O/ (I).
N N\ N N\
a?) or To

 

 

 

 

 

206

 

 

As we have seen earlier, the para-pyridyl nitronyl nitroxide is
ferromagnetic in the solid state and forms linear chains through
"hydrogen-bonding" interactions. The para-pyridinium nitronyl nitroxide
might crystallize to give a magnetic structure with a hydrogen~bonding
framework shown below. There would be little overlap in this arrangement
between the SOMOs, while maintaining overlaps between other frontier

orbitals which is one of the requirements for ferromagnetic coupling.

\ N+ H'm "0° °’N /+~N~O .....
).>—<3

Iz‘i'I-m". "0’
I
2/ 2-0m 'I-Z+

\

""" Of‘N-F’ N'O."" 'H— N+_ \ ;élt

 

 

 

 

207
Ullman's biradica1130 is another logical precursor for an extension of our

non-covalent interaction strategy. A prOposed Ullman's biradical complex

with NH4+ is shown below as a stereo view. There will be little overlap in

this diradical between the SOMOs because of the orthogonality of the

orbitals, while maintaining overlaps between other frontier orbitals.

H H
N+
I \

1-1 H
0* 0'
I \
N+ N

H
\
N N“
\ I
Q' 9'
H\ .H
N4-
H H'

:
~

 

 

 

 

 

 

 

208

 

EXPERIMENTAL

1. General Procedures

All 1H NMR and 13C NMR spectra were obtained with a 300MHz Varian
Gemini or a 300 MHz Varian VXR-300 instrument. IR spectra were
recorded using Perkin-Elmer 599 IR or Nicolet IR/42 spectrometers. UV
Spectra were recorded on a Shimadzu UV-l60 or Hitachi U-2000 spectro-
photometers. Gas Chromatography was conducted on a Perkin-Elmer
8500 using a 25-m (0.32 mm Diameter) bonded methyl silicone column
with an FID detector. HPLC was conducted on a Perkin-Elmer Binary LC
Pump 250 and LC-235 Diode Array Detector. Mass spectra were recorded
on a Finnigan 4000 GC/MS or on a VG Trio-1 5890 GC/MS, using a 30m
SE-54 Altech Capillary column. Electron Spin Echo Envelope Modulation
(ESEEM) studies by pulsed EPR esperiments were conducted on an
instrument build by Professor John McCracken, MSU.

Powder X-ray pattern were recorded on a Phillips XRG 3000 computer—
controlled powder diffractometer, operating at 40kV, 35 mA. Single
crystal X-ray diffraction peaks were recorded on a Nicolet P3F
diffractometer or Enraf-Nonius CAD-4 diffractometer using graphite
monochromated Cu Ka radiation and an w-20 scan by Professor Kahr at
Purdue University. Absorption corrections were applied using the
program DIFABS (Walker & Stuart, 1983), SHELXS86 (Sheldrick, 1986),
and MULTAN82 (Main, Fiske, Hull, Lessinger, Germain, Dec lercq, and

 

 

 

 

 

 

 

209
Woolfson, 1982). In each case computations were performed on a VAX

computer using SDP/V AX software (Frenz, 1978).

2. Solvents and Chemicals

All chemicals used were from Aldrich Chemical Co. or Lancaster Synthesis
Inc., unless otherwise noted. Some of the metal salts were from Alpha
Chemical Co. Solvents were from Fisher Scientific Co., MTM Research
Chemicals, J. T. Baker Chemical Co., Johnson Matthey Chemical Co, and
Mallinckrodt, Inc.

THF, diethyl ether, and benzene were freshly distilled from sodium and

benzophenone under argon. Dry acetonitrile and other solvents were

distilled from CaH2 under argon as needed.

3. Equipments and Procedures

CV Measurements

CV spectra were recorded on a EG & G Princeton Applied Research
Scanning Potentiostat Model 273 and 362 using Bioanalytical Systems
Ag/AgCl as a reference electrode. A potentiostat Model 273 was interfaced
to a PC which controls the potentiometer, including data collection and
plotting. The construction of the experimental cell has been described
elsewhere. (P. T. Kissinger, J. Chem. Educ. 60, 702, 1983.)
Measurements were made under argon on a 1x 10*4 to 1x 10*5 M sample of

substrate in 7 to 8 ml of freshly distilled dry THF or spectroscopic grade

 

 

 

 

210
methylene chloride containing 0.1 M tetra-n-butyl ammonium

tetrafluoroborate. A platinum button was the working electrode and a
platinum coil was the counter electrode. The reference electrode was a
silver wire immersed in a 3M NaCl solution, separated from the bulk
solution by a Vycor tip. The Vycor tip gave some water leakage problems,
but it was satisfactory in determining redox potentials with ferrocene as an
internal reference to determine the potentials. Pulsed differential
voltammograms were obtained with a slight modification in operating
programs in the PC. The scan rates were 50 to 400 mV per second and a
typical one was 200 mV per second. In all cases reversibility was checked
by verifying the equal height of the oxidation and reduction waves in cyclic
voltammetry, and by examining peak to peak distances of the forward and

I'CVCI‘ SC scans.

VT-EPR Measurements

CW-EPR spectra were recorded using a Varian E4 spectrometer with
variable temperature control. The temperature range was 25 to —140 °C
(temperature was controlled by a flow of nitrogen gas running through a
cooling coils in liquid nitrogen). Most of the spectra for HMTP methyl
derivatives and TMTP methyl derivatives gave strong enough signals
without phasing problems, except in the range of -70 to -90 °C, which was
evidently inherent to the spectrometers. Various conditions were studied

with scan windows of 10G to 4000G, modulation amplitude of 1 x 10'2 G

to 1G and receiver gain of l x 101 to 5 x 103 with microwave power of l
to 5 mW.

 

 

 

 

 

SQUID Measurements

 

Magnetic susceptibility measurements in the temperature range from 1.8K
to 350 K at -55kG to 55kG were performed on a MPMS Quantum Design
SQUID magnetometer. Powder or polycrystalline samples were dried,

weighed and placed into a capsule with a minimum amount of cotton on the

 

top. A ventilating hole was made in the capsule with a fine needle, the
capsule was placed in a normal drinking straw, and it's position was fixed
securely with threads up and down. The sample straw was attached to a
sample rod with general centering according to the length of the sample

from the t0p of the rod with the mass of the sample, and inserted into the

 

evacuation chamber. The sample was lowered into an experiment chamber
slowly followed by three cosecutive evacuations of air.

The sample was oriented in the 200 G to 1000 G external magnetic field
according to the centering scheme built into the machine. Magnetic
susceptibility measurements were conducted according to temperature and
field dependence sequences. A general sequence consisted of field
dependence measurements over -5000 G to 5000 G at 2 and 5 K,
magnetization measurements from 0 G to 50000 G at 1.8 K and
temperature dependence experiments from 0 to 320 K at 200, 500, and
5000 G field strengths. The data collected were extracted into a data file in
IBM PC ASCII format consisting of temperatures in K, external field
strength in gauss, induced magnetic moments in emu, and their standard
deviations. The data files were converted as Macintosh format and

analyzed and plotted as molar susceptibilities.

 

 

L_—

 

212
4. Synthesis

General Procedures for Generation of Radicals

Method A: The cation tetrafluoroborate salts (200 mg) were reduced by
10 ml of ~5 M CrC12 solution (prepared with CrCl3-6H20 and Zn/Hg in
10% HCl solution), extracted with 10 m1 of methylene chloride, washed
with water twice, and dried with magnesium sulfate. Flash column
chromatography with diethyl ether over silica gel gave a solution of the

radical.

Method B: The cation tetrafluoroborate salts (200 mg) were reduced by
the addition of an excess (5 equivalents) solid CrC12 into 10 ml of
methylene chloride solution of the cation and water (1 :1 in volume) and the
radical was extracted into methylene chloride, washed with water twice,
and dried with magnesium sulfate. Flash column chromatography with
diethyl ether over silica gel gave a chromatographically pure solution of

radical.

Method C: The cation tetrafluoroborate salts (100 mg) were reduced by
the addition of an excess iodide salt (3 equivalents) (NaI, Lil, NH4I, and n-

Bu4NI) solution in 20 ml of dry THF and the radical was flash
chromatographed using diethyl ether over silica gel as in Method B.

Method D: The cation tetrafluoroborate salt 4 (100 mg) was reduced to
radical 1 by the addition of vitamin C (200 mg) in 20 ml of dry THF

 

 

 

 

 

 

213
followed by stirring overnight under argon at room temperature. The

radical solution was flash chromatographed as above.

Method E: Tris-(2,6-dimethoxy-3,5-dichlorophenyl)methyl ammonium
tetrafluoroborate 17 was converted into a radical by diazotization. To a
precooled solution of the salt (100-200 mg) in conc. HCl, a stoichiometric
amount of saturated solution of NaN02 was added dr0pwise and the
mixture was stirred until it warmed up to room temperature. The radical
was extracted with 30 ml of diethyl ether, washed with water twice, dried
with magnesium sulfate. Flash column chromatography using diethyl ether

over silica gel gave a chromatographically pure solution of the radical.

nrlPr d r frPrifiation f tion

A solution of the triaryl carbinol (200 to 500 mg.) in diethyl ether was
treated with a small amount (0.1 to 0.5 ml. ) of cone. HBF4, and the dark

blue precipitate was filtered, washed with diethyl ether, and dried to give

cation tetrafluoroborate salt as a crystalline solid.

 

 

 

 

nrlPr r frPr rinof rinl
1/3 JUL
0 ; Method A:
1/2 JL / 0H
Ar 0 Jr
t Ar AIAI Method B:
0
Ar AT A Method C:

 

 

 

 

214
Method A: To a solution of 2,6-dimethoxyphenyl lithium (0.218 mol.) in

200 m1. of dry diethyl ether, 0.3 equivalent amount (0.720 mol) of
dimethyl carbonate in 300 ml. of benzene was added dr0pwise (using a
pressure equalized dropping funnel) under argon, and the mixture was
refluxed for 3 days. The reaction mixture was poured into 600 ml. of
water and the organic phase was concentrated. Crystallization from ether

or methylene chloride in n-hexane usually gave crystalline carbinols.

Method B: To a solution of 2,6-dimethoxyphenyl lithium (0.218 mol.) in
200 ml. of dry diethyl ether, 0.5 equivalent amounts (0.109 mol) of the
esters in 400 ml. of benzene were added dropwise through a pressure
equalized dropping funnel under argon and the mixture was refluxed for 3
days. The reaction mixture was poured into water and the organic phase
was concentrated. Crystallization from ether or methylene chloride in n-

hexane usually gave crystalline carbinols.

Method C: To a solution of 2,6-dimethoxyphenyl lithium (0.05 mol.) in
60 ml. of dry diethyl ether, an equivalent amount of the ketone (0.05 mol)
was added dropwise through a pressure equalized dropping funnel under
argon and the mixture was refluxed for 3 days. The reaction mixture was
poured into 100 ml of ice water and the organic phase was concentrated to

yield a gray residue. Crystallization from ether or methylene chloride in

n-hexane gave carbinols.

 

 

 

 

 

 

215
2,6,2',6',2",6"-Hexamethoxytrip_henyl methyl, HMTP 1:

2,6-dimethoxymethylbenzoate: 2,6—dimethoxybenzoylchloride (100 g.) in
1 liter of methanol with 5% 2,6-dimethoxybenzoic acid was refluxed
overnight under argon with stirring. (Commercially available 2,6-
dirnethoxybenzoyl chloride contains 5% 2,6-dimethoxybenzoic acid and it
has been used as a catalyst for the reaction.) The solvent was evaporated
and the residue was extracted with diethyl ether, washed with 5% NaHCO3
solution, dried with magnesium sulfate, and crystallized to give the
theoretical amounts of ester as white crystals. 1H NMR (CDC13, 299.95
MHz): 3.80 (s, 6H), 3.89 (s, 3H), 6.54 (d, J=8 Hz, 2H), 7.26 (t, J=8 Hz, 1H)
13C NMR (75.43 MHz): 52.04, 55.58, 77.02, 103.43, 130.68, 156.86.

di-(2,6-dimethoxyphenyl)methanone: A solution of phenyl lithium was
prepared by the addition of bromobenzene (39.2g., 0.250 mol.) in 125 ml.

of ether to lithium wire (4.0 g., 0.577 mol.) and 60 ml. of ether.
Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added and the reaction
mixture was allowed to stand at room temperature under nitrogen for 60
hours. The reaction mixture was poured into 43.75 g. (218 mmol.) of 2,6-
dimethoxy benzoilchloride in 400 ml. of cold diethyl ether and the reaction
mixture was stirred for 3 hours at -78°C. The reaction mixture was
poured into 300 ml. of ice water and the organic phase was dried and
concentrated to yield a gray residue. Recrystallization from ether gave 25
g. of the product (38%). 1H NMR (CDCl3, 299.95 MHz): 3.67 (s, 12H),
3.50 (d, J=8 Hz, 4H), 7.21 (t, J=8 Hz, 2H) 13C NMR (75.43 MHz): 56, 104,
111, 131, 158 m/z 302 (M+), found 302.

 

 

 

 

 

216
Tris-(2.6-dimethoxyphenyl)methyl carbinol 71: A solution of phenyl

lithium was prepared by the addition of bromobenzene (39.2g., 0.250
mole) in 125 ml. of ether to lithium wire (4.0 g., 0.577 mol..) and 60 ml.
of ether. Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added and
the reaction mixture was allowed to stand at room temperature under
nitrogen for 60 hours. Dimethyl carbonate (6.49 g., 0.072 mole) in 400
ml. of benzene was added and the reaction mixture was refluxed for 3 days
under nitrogen. The reaction mixture was poured into 600 ml. of water
and the organic phase was concentrated to yield a gray residue.
Recrystallization from ether gave 22.4 g. of the carbinol. (71 %) 1H N MR
(CDCl3, 299.95 MHz): 3.42 (s, 18H), 6.46 (d, J=8 Hz, 6H), 6.82 (s, 1H),
7.02 (t, J=8 Hz, 3H) 13C NMR (75.43 MHz): 56.21, 78.22, 105.98,
126.11, 126.85, 158.49, m/z 440 (M+), found 440.

A solution of the radical in diethyl ether was prepared according to

general Method A.

Tris-(2,6-dimethoxyphenyl) borane 63: A solution of phenyl lithium was
prepared by the addition of bromobenzene (39.2 g., 0.250 mole) in 125 ml.

of ether to lithium wire (4.0 g., 0.577 mol..) and 60 ml. of ether.
Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added and the reaction
mixture was allowed to stand at room temperature under nitrogen for 60
hours. 10.22 g. (8.85 ml., 72 mmol.) of BF3-OEt2 in 300 ml. of benzene
was added and the reaction mixture was refluxed for 3 days under
nitrogen. The reaction mixture was poured into 400 ml. of water and the

organic phase was concentrated to yield a gray residue. Recrystallization
from diethyl ether gave 11 g. of the borane (36%). 1H NMR (CDCl3,

 

 

 

 

 

217
299.95 MHz): 3.43 (s, 18H), 6.46 (d, J=8 Hz, 6H), 7.16 (t, J=8 Hz, 3H) 13C

NMR (75.43 MHz): 56.52, 105.09, 130.17, 162.54 m/z 422 (M+), found
422.

Tris-(2.6-dimethoxynhenyl-3.5-dichlorophenyl) borane 64: To a
precooled solution of 3.7 g. (8.77 mmol.) of tris-(2,6-dimethoxyphenyl)
borane in dry diethyl ether (to act as a base) and methylene chloride
solution (10 ml. ether, 30 ml. methylene chloride), 4.5 ml. (7.1 g., 55
mmol.) of sulfurylchloride was slowly added through syringe at -78 oC and
refluxed overnight with stirring under argon. The reaction mixture was
washed with water twice, dried with magnesium sulfate, concentrated, and
chromatographed with 10% ethyl acetate in n-hexane to give white crystals
(3.0 g., 54%). 1H NMR (CDCl3, 299.95 MHz): 3.42 (s, 18H), 7.42 (s, 3H)
13C NMR (75.43 MHz): 61.09, 129.87, 133.00, 156.20 m/z 629 (M+),
found 629.

Tris-(4-chloro—2,6-dimethoxyphenyl)methanol 66: To a cooled solution
of 5.0 g. (29 mmol.) of 3-chloro-dimethoxybenzene in 50ml of diethyl

ether, 11.6 ml. (29 mmol.) of 2.5M n-BuLi in n-hexane was added slowly
at -78°C under argon with stirring. The mixture was stirred at the room
temperature for 2 days, 862 mg. (9.6 mmol.) of dimethyl carbonate in 50
m1. of benzene was added, and the reaction mixture was refluxed for 3

days under nitrogen. The reaction mixture was poured into 100 ml. of

water and the organic layer was separated, dried with sodium sulfate, and
the solvent was evaporated (200 mg., 4%). Cation: 1H NMR (CDC13,

299.95 MHz): 3.64 (s, 18H), 6.47 (s, 6H) 13C NMR (75.43 MHz): 56.90,
77.05, 105.63, 149.16, 162.30.

 

 

 

 

218

 

Tris-(2.4,6-trimethoxyethoxyphenyl)methanol 67: To a cooled solution of
20 g. (119 mmol.) of trimethoxybenzene in 100ml of diethyl ether, 48.4
ml. (121 mmol.) of 2.5 M n-BuLi in n-hexane was added slowly at -78°C
under argon with stirring. The mixture was stirred at the room
temperature for 2 days, 3.57g (39.6 mmol.) of dimethyl carbonate in 200
ml. of benzene was added, and the reaction mixture was refluxed for 3
days under nitrogen. The reaction mixture was poured into 300 ml. of
water and the organic layer was separated, dried with sodium sulfate, and
the solvent was evaporated (13.4g, 64%). 1H NMR (CDC13, 299.95 MHz):
3.44 (s, 18H), 3.75 (s, 9H), 6.04 (s, 6H), 6.66 (s, 1H). 13C NMR (75.43
MHz): 54.65, 77.02, 91.94, 157.77, 158.66. Cation: 1H NMR (CDCI3,

299.95 MHz): 3.57 (s, 18H) 3.97 (s, 9H) 6.03 (s, 6H).

Tris-12,6-dimethoxy-4-methylphenyl)methanol 68: A solution of phenyl
lithium was prepared by the addition of 11.85 g. (75.4 mmol.) of

bromobenzene in 60 ml. of ether to 1.21 g. (174.1 mg. atom) of lithium
and 30 ml. of ether. 10 g. (65.8 mmol.) of dimethoxy toluene was added
and the reaction mixture was allowed to stand at room temperature under
nitrogen for 60 hours. 1.95 g. (21.7 mmol.) of dimethyl carbonate in 200
ml. of benzene was added dropwise and the reaction mixture was refluxed
for 3 days under nitrogen. The reaction mixture was poured into 200 ml.
of water and the organic phase was concentrated to yield a gray residue.
Recrystallization from ether gave 5.9 g. of the carbinol (56 %). 1H NMR
(CD3CN, 299.95 MHz): 2.23 (s, 9H), 3.37 (s, 12H), 6.32 (s, 6H) 6.54 (s,
1H)13C NMR (75.43 MHz): 21.37, 56.67, 107.67, 118.30, 125.30, 136.78,
159.32.

 

 

 

 

 

 

219

Tris-g2,6-dimethoxy-4-phenylphenyl)methyl 69:

2-hydrox3_L—4-oxo-6-phenyl-2-cyclohexenecarboxylate : To a solution of
3.5 g. of NaOH in 150 ml. of methanol, 9.8 g. of dimethyl malonate (75
mmol.) was added and the mixture stirred for 30 min. at room temperature
under argon. 10 g. (68 mmol.) of trans-4-phenyl-3~buten-2-one was added
portion wise with stirring. The reaction mixture was refluxed for 3 hours
under argon and allowed to cool to room temperature. The solvent was
distilled under reduced pressure. The oily residue was treated with acidic
(pH=5) water and extracted with dichloromethane. The solution was dried
over magnesium sulfate, the solvent was evaporated, and the product
mixture was dried under vacuum. The crude mixture was submitted to the

next reaction without purification.

5-Phenyl Resorcinol : 20 g. (81.3 mmol.) of crude phenylcarboxylate 2~
hydroxy-4-oxo-6-pheny1-2-cyclohexenecarboxylate was dissolved in 150
ml. of DMF and 4.13 ml. (12.8g, 80 mmol.) of bromine was added
dropwise with dry ice cooling under argon. The mixture was heated for
three hours at about 80 °C until the evolution of C02 ceased followed by
continued refluxing for 2 days. The solvent was distilled off under
reduced pressure and the mixture was extracted with 300 ml. of 5% NaOH
solution twice. The combined solution was treated with conc. HCl carefully
with cooling and the 5-Phenyl resorcinol mixture was extracted with 500
ml. of methylene chloride. The solution was dried with magnesium sulfate

and the solvent was evaporated under vacuum to give crystalline product.
(5.6g, 43%) 1H NMR (CD3CN, 299.95 MHz): 6.36 (t, J=2 Hz, 1H), 6.61

 

 

 

 

 

220
(d, J=2 Hz, 2H), 7.31 (m, 1H), 7.40 (m, 2H), 7.55 (m, 2H), 8.35 (s, 2H),

13C NMR (CD3CN, 75.43 MHz): 102.48, 106.32, 107.11, 127.56, 128.10,
129.50, 144.05, 159.77.

3,5-dimethoxy-phenyl benzene : 10 g. (53.76 mmol.) of 5-
phenylresorcinol was dissolved in 300 ml. of 5% NaOH solution, 12.72 ml.

(16.95g, 134.4 mmol.) of dimethylsulfate was added using a separatory
funnel, the mixture was mixed vigorously with shaking, and left for 30
min. followed by an extraction with 200 ml. of diethyl ether. The organic
layer was dried, the solvent was evaporated, and the mixture was
chromatographed over a silica gel using 30% ethyl acetate in hexane to
give NMR clean product (8.9g, 78%).1H NMR (CDCl3, 299.95 MHz): 3.84
(s, 6H), 6.46 (t, J=2 Hz, 1H), 6.73 (d, J=2 Hz, 2H), 7.34 (m, 1H), 7.42 (tm,
2H), 7.56 (dm, 2H) 13C NMR (75.43 MHz): 54.99, 98.83, 105.01, 126.77,
127.13, 128.27, 140.76, 143.01, 160.59.

Tris-(2,6-dimethoxy-4-phenyl phenyl)methanol : To a cooled solution of 2
g. (9.345 mmol.) of 5-phenyl—dimethoxybenzene in 100ml of diethyl ether,

4.11 ml. of 2.5 M n-BuLi in n-hexane was added slowly at -78°C under
argon with stirring. The mixture was stirred at the room temperature for
2 days, 279 mg. (3.10 mmol.) of dimethyl carbonate in 100 ml. of benzene
was added, and the reaction mixture was refluxed for 3 days under
nitrogen. The reaction mixture was poured into 100 ml. of water, the
organic layer was separated, dried with sodium sulfate, and the solvent was

evaporated (4.4g, 71 %). 1H NMR (CDC13, 299.95 MHz): 3.54 (s, 18H),
6.75 (8, 6H), 6.95 (S, 1H), 7.32 (m, 3H), 7.41 (m, 6H), 7.59 (m, 6H) 13C

 

 

 

 

 

 

221
NMR (75.43 MHz): 56.43, 56.48, 105.02, 125.85, 126.83, 127.12, 128.64,

139.23, 141.40, 157.00, 158.79 m/z 668 (M+), found 668.

A solution of carbinol in diethyl ether was treated with a small amount of
conc. HBF4 and the blue precipitates were filtered and dried to give the
cation tetrafluoroborate salt as crystalline solids. Cation: 1H NMR (CDC13,
299.95 MHz): 3.71 (s, 18H), 6.78 (s, 6H), 7.52 (m, 9H), 7.75 ((1, 1:7 Hz,
6H) 13C NMR (75.43 MHz): 56.48, 102.97, 126.95, 128.87, 129.94,
138.43, 192.29.

A solution of the radical in diethyl ether was prepared according to

general Method A.

Di-

 

A solution of phenyl lithium was prepared by the addition of 19.6 g. (125
mmol.) of bromobenzene in 60 ml. of ether to 2 g. (289 mol.) of lithium
and 30 ml. of ether. 15g (109.6 mmol.) of 1,3-dimethoxybenzene was
added and the reaction mixture was allowed to stand at room temperature
under nitrogen for 60 hours. 20.3 g. (96.8 mmol.) of 2,6-dimethoxy,5
methyl benzoate in 200 ml. of benzene was added and the reaction mixture
was refluxed for 3 days under argon. The reaction mixture was poured
into 200 ml. of water and the organic phase was concentrated to yield gray
residues. Recrystallization from ether gave a crude carbinol. A solution

of the carbinol in diethyl ether was treated with a small amount of cone.

HBF4 and the blue precipitate was filtered and dried to give the cation

tetrafluoroborate salt as crystalline solids (12g, 27%). Cation: 1H NMR
(CDC13, 299.95 MHZ): 2.48 (S, 3H), 3.54 (s, 12H), 3,59 (S, 6H), 6.36 (S,

 

 

 

 

 

 

222
2H), 6.48 (d, J=9 Hz, 4H), 7.41 (t, J=9 Hz, 2H) 13C NMR (75.43 MHz):

24.65, 56.60, 57.14, 104.75, 106.21, 138.03, 152.20, 160.63, 164.83,
165.71.

Tris-12,6-dimethoxyphenyl) methyl ammonium tetrafluoroborate 72: To a
10 m1. THF solution of tris-(2,6-dimethoxyphenyl)methanol (500 mg., 1

mmol.), 0.2 ml. of cone. HBF4 was added followed by the addition of conc.
aqueous ammonia until the color of the mixture turned red. 30 m1. of
diethyl ether was added and the mixture was left to give red precipitates
which were filtered and dried to give NMR (and X-ray) clean crystals of
the product (380 mg., 72%). 1H NMR (CDC13, 299.95 MHz): 3.44 (s,
18H), 6.58 (d, J=8 Hz, 6H), 7.21 (t, J=8 Hz, 3H) 8.45 (s, 3H) 13C NMR
(75.43 MHz): 56.25, 105.93, 106.02, 117.5, 129.06, 145.50, 157.50.

Tris-(2,6-dimethoxyphenyl)methane 73 : To a solution of 5 g. (6.82
mmol.) of tris-(2,6-dimethoxyphenyl)methanol in 100 m1. of THF, 30 ml.

of 10% HCl solution, 496.4 mg. (7.90 mmol.) of solid NaBH3CN was
added slowly and the mixture was stirred for 2 hours at the room
temperature. 100 ml. of water was added to the mixture and the product
filtered, washed with water, and dried under vacuum to give 2.7 g. of clean
products (93.4%). 1H NMR( CDCl3, 299.95 MHz): 3.41 (s, 18H), 6.45 (s,
1H), 6.48 (d, J=8 Hz, 6H), 7.02 (t, J=8 Hz, 3H) 13C NMR (75.43 MHz):
56.82, 106.37, 124.00, 125.64, 159.51 m/z 424 (M+), found 424.

Tris-12,6-dimethoxyphenyl) silane 74: A solution of phenyl lithium was
prepared by the addition of bromobenzene (39.2g., 0.250 mole) in 125 ml.

of ether to lithium wire (4.0 g., 0.577 mol..) and 60 ml. of ether.

 

 

 

 

 

223
Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added and the reaction

mixture was allowed to stand at room temperature under nitrogen for 60
hours. 11.8 g. of triethoxysilane (0.072 mole) in 400 ml. of benzene was
added and the reaction mixture was refluxed for 3 days under nitrogen.
The reaction mixture was poured into 600 ml. of water and the organic
phase was concentrated to yield gray residues. Recrystallization from
diethyl ether gave 13.4 g. of the silane (42%). 1H NMR( CDC13, 299.95
MHz): 3.46 (s, 18H), 5.52 (s, 1H), 6.44 (d, J=8 Hz, 6H), 7.19 (t, J=8 Hz,
3H) 13C NMR (75.43 MHz): 55.41, 77.1, 103.80, 103.80, 129.69, 164.90.

Tris-( 2,6-dimethoxyphenyl-3 ,5-dichlorophenyl )methyl ammonium tetra-

fluoroborate 75: To a solution of 5 g. of tris-(2,6-dimethoxyphenyl)-
methyl ammonium tetrafluoroborate in 10 ml. of diethyl ether and 50 ml.
of dichloromethane, 10 ml. of sulfurylchloride was added and the mixture
was refluxed under argon overnight. The organic layer was separated,
dried with magnesium sulfate, the solvent was evaporated, and
chromatographed over silica gel using methylene chloride to yield 2.75 g.
of hexachloro-ammonium tetrafluoroborate as white crystals (37.5 %). 1H

NMR (CDC13, 299.95 MHz): 3.38 (s, 9H), 3.66 (s. 9H), 7.53 (s, 3H), 8.69
(s, 3H).

Tris-12,6-dimethoxyphenyl-3,5-dichlorophenyl) methane 76: To a
solution of the methane 73 (2 g., 4.71 mmol.) in 10 ml. of diethyl ether

and 50 ml. of dichloromethane, 33 mmol. of sulfurylchloride was added
and the mixture was refluxed under argon overnight. The organic layer
was separated, dried with magnesium sulfate, and the solvent was

evaporated. Chromatography over silica gel using dichloromethane gave

 

 

 

224
hexachloromethane as white crystals (2.2g, 75%). 1H NMR (CDC13,

299.95 MHz): 3.07 (s, 9H), 3.73 (s, 9H), 6.65 (s, 1H), 7.31 (s, 3H) 13C
NMR (75.43 MHz): 59.34, 61.00, 122.43, 123.94, 129.43, 133.32, 153.05,
156.93.

Dim-dimethoxyphenyl)-_9Lhenyl methanol 77 : A solution of dimethoxy-
phenyllithium was prepared by the addition of 34.8 ml. (86.9 mmol.) of
2.5 M n-BuLi in n-hexane to a solution of 10 g. (72.4 mmol.) of
dimethoxybenzene in sodium-dried diethyl ether at -78°C and the mixture
was stirred at room temperature under argon for 48 hours. 3.13 g. (23.0
mmol.) of methyl benzoate in 200 ml. of benzene was added and the
reaction mixture was refluxed for 3 days under argon. The reaction
mixture was poured into 200 m1. of water, the organic phase was washed
with 200 ml. of water twice, dried, and concentrated to yield carbinol as
gray residues. The carbinol was crystallized in dichloromethane and hexane
to give 9 g. of NMR clean carbinol (65%). A solution of carbinol (500
mg., 1.31 mmol.), in 150 ml. of diethyl ether was treated with 0.1 m1. of
cone. HBF4 and the blue precipitates were filtered and dried to give the
cation tetrafluoroborate salt as a crystalline solid quantitatively. 200 mg. of
the cation tetrafluoroborate salt in 20 ml. of dichloromethane was reduced
by 20 ml. of 5M CrC12 solution (prepared with CrCl3-6H20 and Zn/Hg in
10% HCl solution) and extracted with dichloromethane and washed with
water twice, dried with magnesium sulfate, and chromatographed with

diethyl ether over silica gel to give a chromatographically pure solution of
the radical (35%, based on SQUID measurement, neff). 1H NMR
(CDC13, 299.95 MHz): 3.38 (s, 12H), 6.43 (s, 1H), 6.54 (d, J=8 Hz, 4H),

7.11 (t, J=8 Hz, 2H), 7.13 (m, 1H), 7.21 (m, 2H), 7.46 (m, 2H) 13C NMR

 

 

 

 

 

 

225
(75.43 MHz): 56.50, 78.50, 106.74, 125.37, 126.31, 126.50, 126.63,

127.28, 149.50, 158.00, mp. =104 -105°C.
Cation: 1H NMR (CDC13, 299.95 MHz): 3.54 (s, 12H), 6.63 (d, J=8 Hz,
4H), 7.44 (m, 4H), 7.61 (m, 1H), 7.79 (t, J=8 Hz, 2H) 13C NMR (75.43
MHz): 56.83, 105.36, 125.00, 129.09, 134.59, 137.02, 145.90, 163.79.
191.58 m/e=364.44116 mp. =142 —144°C.

Di-(2,6-dimethoxyphenyl)(4-chlorophenyl)-methanol 78: A solution of

phenyl lithium was prepared by the addition of bromobenzene (39.2g.,
0.250 mole) in 125 ml. of ether to lithium wire (4.0 g., 0.577 mol.) and 60
ml. of ether. Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added
and the reaction mixture was allowed to stand at room temperature under
nitrogen for 60 hours. 14.8 g. (86.8 mmol.) of methyl 4-chlorobenzoate in
400 ml. of benzene was added and the reaction mixture was refluxed for 3
days under nitrogen. The reaction mixture was poured into 600 ml. of
water and the organic phase was concentrated to yield gray residues.
Recrystallization from methylene chloride in n-hexane gave 31.3 g. of
NMR clean crystals of carbinol (87%). 1H NMR (CDC13, 299.95 MHz):
3.39 (s, 12H), 6.49 (s, 1H), 6.53 (d, J=8 Hz, 4H), 7.12 (t, J=8 Hz, 2H), 7.15
(dq, J=8 Hz, 2H), 7.37 (dq, J=8 Hz, 2H) 13C NMR (75.43 MHz): 56.29,
79.12, 106.625, 125.73, 126.64, 127.53, 128.11, 130.88, 148.01, 157.93.
Cation 1H NMR (CDCl3, 299.95 MHz): 3.55 (s, 12H), 6.62 (d, J=8 Hz,
4H), 7.40 (dd, 4H), 7.78 (t, J=8 Hz, 2H) 13C NMR (75.43 MHz): 56.87,
105.42, 124.40, 129.50, 135.54, 142.65, 144.04, 146.23, 163.71, 189.04.

 

 

 

 

 

 

226
Di-(2,6-dimethoxy phenyl)(4-methoxyphenylzmethanol 79: A solution of

phenyl lithium was prepared by the addition of bromobenzene (39.2 g.,
0.250 mole) in 125 ml. of ether to lithium wire (4.0 g., 0.577 mol.) and 60
ml. of ether. Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added
and the reaction mixture was allowed to stand at room temperature under
argon for 60 hours. 14.4 g. (98.6 mmol.) of methyl 4-methoxybenzoate in
400 ml. of benzene was added and the reaction mixture was refluxed for 3
days under argon. The reaction mixture was poured into 600 ml. of water
and the organic phase was concentrated to yield gray residues.
Recrystallization from methylene chloride in n—hexane gave 24.6 g. of
NMR clean crystals of carbinol (61%). 1H NMR (CDC13, 299.95
MHz):3.38 (s, 12H), 3.76 (s, 3H), 6.36 (s, 1H), 6.53 (d, J=8 Hz, 4H), 6.75
(dq, J=8 Hz, 2H), 7.10 (t, J=8 Hz, 2H), 7.34 (dq, J=8 Hz, 2H) 13C NMR
(75.43 MHz): 55.12, 56.47, 79.17, 106.88, 111.96, 126.65, 127.20, 127.73,
141.60, 157.40, 158.03, mp=105 0C.

A solution of carbinol in diethyl ether was treated with a small amount of
cone. HBF4 and the blue precipitates were filtered and dried to give the
cation tetrafluoroborate salt as a crystalline solid. The cation
tetrafluoroborate salt was reduced by CrClz solution (prepared with
CrC13-6H20 and Zn/Hg in 10% HCl solution) and extracted with
dichloromethane, washed with water twice, dried with magnesium sulfate,
and chromatographed using diethyl ether over silica gel to give a

chromatographically pure solution of radical (76%).

 

 

 

 

The radical solution was left under the air for 3 days to give crystals of
bis-(2,6-dimethoxy phenyl)methyl quinone methide shown above. 1H
NMR (CDC13): 3.62 (s, 12H), 6.29 (dq, J=8 Hz, 2H), 6.56 (d, J=8 Hz, 4H),
7.20 (dq, J=8 Hz, 2H), 7.24 (t, J=8 Hz, 2H) 13C NMR (75.43 MHz): 55.98,
104.17, 117.5, 127.51, 130.24, 139.56, 150.00, 158.36, 188.00.

Di- 2 6-dimethox hen l 4-meth 1 hen lmethanol : A solution of
phenyl lithium was prepared by the addition of bromobenzene (39.2g.,
0.250 mole) in 125 ml. of ether to lithium wire (4.0 g., 0.577 mol.) and 60
ml. of ether. Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added
and the reaction mixture was allowed to stand at room temperature under
nitrogen for 60 hours. 14.8 g. (98.6 mmol.) of methyl 4-methylbenzoate in
400 m1. of benzene was added and the reaction mixture was refluxed for 3
days under nitrogen. The reaction mixture was poured into 600 ml. of

water and the organic phase was concentrated to yield gray residues.

Recrystallization from methylene chloride in n-hexane gave 32 g. of NMR
clean crystals of carbinol (83%). 1H NMR (CDCl3, 299.95 MHz):2.29 (s,
3H), 3.38 (s, 12H), 6.38 (s, 1H), 6.53 (d, J=8 Hz, 4H), 7.00 (d, J=8 Hz,
2H), 7.10 (t, J=8 Hz, 2H), 7.32 (d, J=8 Hz, 2H) 13C NMR (75.43 MHz):
21.00, 56.46, 79.38, 106.88, 126.47, 126.61, 127.19, 127.37, 134.65,
146.23, 158.07.

 

 

 

 

 

 

228

Di-g 2,6-dimethoxy phenyl)(4-hydroxyphenyl)methanol 81: Bis-(2,6-
dimethoxy phenyl)methyl quinone in THF was treated with conc. HBF4 and

crystallized by slow evaporation to give needles of di-(2,6-dimethoxy
phenyl)(4-hydroxyphenyl)methyl cation tetrafluoroborate (60% by nmr).
1H NMR (CDCl3): 3.56 (s, 12H), 6.56 (d, J=8 Hz, 4H), 7.17 (dq, =8 Hz,
2H), 7.46 (t, J=8 Hz, 2H), 7.82 (dq, J=8 Hz, 2H).

Di-(2,6-dimethoxy phenyl)(4-nitrophenyl)methanol 83: A solution of
dimethoxyphenyllithium was prepared by the addition of 70 ml. (175

mmol.) of 2.5 M n-Butyl lithium in n-hexane to a solution of 20 g. (145
mmol.) of dimethoxybenzene in sodium-dried diethyl ether at -780C and
the mixture was stirred at room temperature under argon for 48 hours.
6.56 g. (36.2 mmol.) of methyl 4-nitrobenzoate in 300 ml. of benzene was
added and the reaction mixture was refluxed for 3 days under nitrogen.
The reaction mixture was poured into 300 m1. of water and the organic
phase was dried and concentrated to yield brown residues. A solution of
carbinol in diethyl ether was treated with a small amount of conc. HBF4
and the blue precipitates were filtered and dried to give the cation
tetrafluoroborate salt as a crystalline solid. (<1%) Cation: 1H NMR
(CDCl3, 299.95 MHz):3.56 (s, 12H), 6.63 (d, J=9 Hz, 4H), 7.40 (dd, J=8
Hz, 4H), 7.79 (t, J=9 Hz, 2H) 13C NMR (75.43 MHz): 56.88, 105.43,
124.41., 129.52, 135.55, 142.66, 144.01, 146.25, 163.73, 189.05.

Di-(2,6-dimethoxy phenylH3,5-dimethoxyphenyl)methanol 84: A

solution of phenyl lithium was prepared by the addition of bromobenzene
(39.2g., 0.250 mole) in 125 ml. of ether to lithium wire (4.0 g., 0.577
mol..) and 60 ml. of ether. Resorcinol dimethyl ether (30.0 g., 0.218

 

 

 

 

 

 

 

229
mole) was added and the reaction mixture was allowed to stand at room

temperature under nitrogen for 60 hours. 17 g. (98.6 mmol.) of methyl
3,5-dimethoxybenzoate in 400 ml. of benzene was added and the reaction
mixture was refluxed for 3 days under nitrogen. The reaction mixture was
poured into 600 ml. of water and the organic phase was concentrated to
yield gray residues. Recrystallization from methylene chloride in n-hexane
gave 24 g. of NMR clean crystals of carbinol. (55%). 1H NMR (CDCl3,
299.95 MHz): 3.40 (s, 12H), 3.70 (s, 6H), 6.27 (t, J=2 Hz, 1H), 6.46 (s,
1H), 6.52 (d, J=8 Hz, 4H), 6.69 (d, J=2 Hz, 2H), 7.09 (t, J=8 Hz, 2H), 13C
NMR (75.43 MHz): 55.17, 56.43, 97.75, 105.25, 106.76, 127.28, 152.21,
158.06, 159.50.

Tris-1 6-methoxy-2-g 2-methoxyethoxy )phenyl) methanol 85:

1-(2-methoxyethoxy)-3-methoxy benzene : A solution of 50 g. (400
mmol.) of methoxyphenol, 38.12 g. (400 mmol.) of chloromethoxyethane,
38.12g (400 mmol.) of chloromethoxyethane in 700ml of acetone was
refluxed overnight under argon. The solution was filtered, evaporated,
extracted with diethyl ether, dried with magnesium sulfate, concentrated,
and chromatographed using 10% ethyl acetate in n-hexane to give the
product as a clear oil (60%). 1H NMR (CDCl3): 3.42 (d, J=2 Hz, 3H), 3.70
(d, J=2 Hz, 2H), 3.74 (d, J=2 Hz, 3H), 4.07 (d, J=2 Hz, 2H), 6.50 (m, 3H),
7.16 (t, J=2 Hz, 1H) 13C NMR (75.43 MHz): 54.48, 58.45, 66.55, 70.37,
100.59, 106.01, 129.45, 159.75, 160.53. m/z 182 (M+), found 182.

To a cooled solution of 14.82 g. (81.4 mmol.) of methoxyethoxy benzene
in 100 ml. of diethyl ether, 39.1 ml. (97.7 mmol.) of 2.5 M n-BuLi in n-

 

 

 

 

 

230
hexane was added slowly at -78°C under argon with stirring. The mixture

was stirred at room temperature for 2 days and 2.42 g. (26.86 mmol.) of
dimethyl carbonate in 200 m1. of benzene was added and the reaction
mixture was refluxed for 3 days under nitrogen. The reaction mixture was
poured into 300 ml. of water and the organic layer was separated, dried

with sodium sulfate, and the solvent was evaporated.

A solution of carbinol in diethyl ether was treated with a small amount of
conc. HBF4 and the blue precipitates were filtered and dried to give the
cation tetrafluoroborate salt as a crystalline solid. The cation
tetrafluoroborate salt was reduced by CrC12 solution (prepared with
CrCl3-6H20 and Zn/Hg in 10% HCl solution), extracted with
dichloromethane, washed with water twice, dried with magnesium sulfate,
and chromatographed with diethyl ether over silica gel to give a
chromatographically pure solution of the radical.

1H NMR (CDCl3): 3.17 (t, 6H), 3.19 (s, 9H), 3.39 (s, 9H), 3.81 (t, J=2 Hz,
6H), 6.47 (dd, 6H), 6.63 (s, 1H), 7.04 (t, J=2 Hz, 3H) 13C NMR (75.43
MHz): 55.68, 58.44, 68.27, 70.31, 105.93, 107.18, 125.78, 157.09, 158.57.
Cation:1H NMR (CDC13): 3.14 (s, 9H), 3.21 (s, 6H), 3.55 (s, 9H), 3.93 (s,
6H), 6.50 (d, J=2 Hz, 6H), 7.57 (t, J=2 Hz, 3H) 13C NMR (75.43 MHz):
56.3, 58.4, 68.5, 69.7, 105.3, 121.0, 142.0, 156.6.

Tris- 2 6-di 2-methox ethox hen l methanol 8 :

l,3-di(2-methoxyethoxy2benzene : A solution of 70 g. (636 mmol.) of
resorcinol, 185.2 g. (1340 mmol.) of potassium carbonate, 126.2 g. (1340

mmol.) of 2-chloroethyl methyl ether, 20 g. (120 mmol.) of potassium

 

 

 

231
iodide in 700 ml. of acetone was refluxed overnight under argon. The

solution was filtered, evaporated, extracted with diethyl ether, dried with
magnesium sulfate, concentrated, and chromatographed with 10% ethyl
acetate in n-hexane to give the product as a clear oil (42.4%). 1H NMR
(CDC13, 299.95 MHz) 3.42 (s, 6H), 3.72 (m, 4H), 4.07 (m, 4H), 6.51
(m, 3H), 7.13 (m, 1H) 13C NMR (75.43 MHz): 58.75, 66.74, 70.54,
101.24, 106.65, 129.34, 159.48. m/z 226 (M+), found 226.

Tris-1 2,6-di( 2-methoxyethoxy )phenyl )methanol: 1,3 -di(2-methoxyethoxy)-

benzene was prepared by the reaction of resorcinol and 2-chloroethyl
methyl ether with K2CO3 in refluxing acetone. A solution of 1,3-di(2-
methoxyethoxy) phenyl lithium was prepared by the addition of n-Butyl
lithium in n-hexane to a solution of 1,3-di-(2-methoxyethoxy)benzene in
sodium~dried diethyl ether at —78°C and the mixture was stirred at room
temperature under argon for 48 hours. Dimethyl carbonate in benzene was
added and the reaction mixture was refluxed for 3 days under argon. The
reaction mixture was poured into ice water and the organic phase was
concentrated to yield an an oil (11.9%). 1H NMR (CDC13, 299.95 MHz):
3.15 (s, 18H), 3.20 (m, 12H), 3.92 (m, 12H), 6.50 (d, J=9 Hz, 6H), 7.56
(t, J=9 Hz, 3H) 13C NMR (75.43 MHz): 58.37, 68.39, 69.80, 105.35,
111.50, 141.19, 161.82.

A solution of the carbinol in diethyl ether was treated with a small amount
of cone. HBF4 and the blue precipitates were filtered and dried to give the

cation tetrafluoroborate salt as a crystalline solid. The cation tetrafluoro-

borate salt was reduced by CrC12 solution (prepared with CrCl3-6HzO and

 

 

 

 

 

 

 

 

232
Zn/Hg in 10% HCl solution), extracted with dichloromethane, washed with

water twice, dried with magnesium sulfate, and chromatographed with

diethyl ether over silica gel to give a chromatographically pure solution of
the radical 86.

Bis-(2,6-dimethoxyphenyl)-4-chlorophenyl amine 87: The mixture of 950
mg. (7.6 mmol.) of 4-chloroaniline, 5 g. (18.9 mmol.) of 2,6—dimethoxy-

iodobenzene, 6.9 g. (50 mmol.) of potassium carbonate, 2.54 g. (40 mg.
atom) of copper, and 248 mg. (0.767 mmol.) of tris-(2-methoxyethoxy)
amine was refluxed in 50 ml. of o-dichlorobenzene for 48 hours under
argon. Inorganic salts and copper were removed by filtration of the hot
reaction mixture. The solvent was distilled off under reduced pressure and
the residue freed from the impurities by passing through a short silica gel
column. Chromatography over silica gel using 20% ethyl acetate in hexane
gave the product as white crystals (2.4g, 78%). 1H NMR (CDC13, 299.95
MHz): 3.58 (s, 12H), 6.40 (d, J=8 Hz, 2H), 6.56 (d, J=8 Hz, 4H), 6.97 (d,
J=2 Hz, 2H), 7.20 (t, J=8 Hz, 2H) 13C NMR (75.43 MHz): 56.16, 104.00,
105.51, 111.5, 115.69, 125.75, 127.79, 157.22.

Bis-12,6-dimethoxyphenyl)-3,5-dimethoxyphenyl amine 88: The mixture
of 1.16 g. (7.6 mmol.) dimethoxyaniline, 5 g. (18.9 mmol.) 2,6-

dimethoxy-iodobenzene, 6.9 g. (50 mmol.) of potassium carbonate, 2.54 g.
(40 mmol.) of copper, and 248 mg. (0.767 mmol.) of tris-(2-
methoxyethoxy) amine was refluxed in 50 ml. of o-dichlorobenzene for 48
hours under argon. Inorganic salts and copper were removed by filtration
of the hot reaction mixture. The solvent was distilled off under reduced

pressure and the residue freed from the impurities by passing through a

 

 

 

 

 

 

233
short silica gel column. Chromatography over silica gel using 20 % ethyl

acetate in hexane gave the product as white crystals. (2.6lg, 81%). 1H
NMR (CDC13, 299.95 MHZ): 3.59 (s, 12H), 3.60 (S, 6H), 5.80 (d, J=2 Hz,
2H), 5.87 (t, J=2 Hz, 1H), 6.53 (d, J=8 Hz, 4H), 7.15 (t, J=8 Hz, 2H) 13C
NMR (75.43 MHZ): 55.00, 56.50, 90.82, 93.66, 105.50, 125.70, 157.37,
160.78.

Tris-g3,5-dimethoxyphenyl)amine 89: The mixture of 460 mg. (3.80
mmol.) of 3,5—dimethoxyaniline, 2.51 g. (9.50 mmol.) of 2,6-dimethoxy-

iodobenzene, 4.21 g. (30.54 mmol.) of potassium carbonate, 970 mg.
(15.30 mmol.) of copper, and 248 mg. (0.767 mmol.) of tris-(2-
methoxyethoxy) amine was refluxed in 20 ml. of o-dichlorobenzene for 24
hours under argon. Inorganic salts and copper were removed by filtration
of the hot reaction mixture. The solvent was distilled off under reduced
pressure and the residue freed from the impurities by passing through a
short silica gel column. Chromatography over silica gel using 20 % ethyl
acetate in hexane gave the product as white crystals (880 mg., 55%). 1H
NMR (CDC13, 299.95 MHz): 3.68 (s, 18H), 6.14 (t, J=2 Hz, 3H), 6.24 (d,
J=8 Hz, 6H) 13C NMR (75.43 MHz): 54.8, 54.9, 95.0, 102.6, 148.6, 160.7
m/z 425 (M+), found 425.

Jacobson's (head-to-tail) peroxy dimer 91: A fresh solution of radical 77

was allowed to stand for 3 days under air to give the peroxy dimers. The
peroxy dimer was crystallized by slow evaporation of a diethyl ether
solution of the radical (65%, based on SQUID measurement, ueff) (the

peroxydimer can be made almost quantitatively with an 02 saturated

 

 

 

 

 

 

 

 

234
solution of the radical in methylene chloride). 1H NMR (CDC13, 299.95

MHz): 3.40 (s, 12H), 3.58 (d, J=3 Hz, 12H), 4.82 (m, 1H), 5.54 (dd, J=12,2
Hz, 2H), 6.12 (dd, J=10,2 Hz, 2H), 6.44 (dd, 1:82 Hz, 8H), 6.55 (m, 2H),
7.03 (t, J=8 Hz, 4H), 7.05 (m, 1H), 7.11 (t, J=3 Hz, 2H), 7.68 (d, J=8 Hz,
2H). mp=108-111 °C.

Martin's (tail~to-tail) peroxy dimer 90: A fresh solution of radical 1 was
allowed to stand for 3days under air to give tail-to-tail peroxy dimer. (7%
yield, based on SQUID measurement of the best radical, neff) (20%, based
on average SQUID measurement of the radical, neffxwe can increase the
yield of the peroxydimer significantly with 02 saturated solution of the
radical in methylene chloride.) 1H NMR (CDC13, 299.95 MHz): 3.13 (s,
12H), 3.51 (d, J=3 Hz, 24H), 5.05 (d, J=2 Hz, 4H), 5.33 (t, J=2 Hz, 2H),
6.39 (dd, J=2,2 Hz, 8H), 7.00 (t, J=8 Hz, 4H). mp=>300 °C.

Di-g2,6-dimethoxyphenyl)(2-hydroxyphenylzmethyl carbinol 92: A

solution of phenyl lithium was prepared by the addition of bromobenzene
(19.6g., 0.12 mole) in 75 ml. of ether to lithium wire (2.0 g., 0.288 mol..)
and 30 ml. of ether. Resorcinol dimethyl ether (15.0 g., 0.109 mole) was
added and the reaction mixture was allowed to stand at room temperature
under nitrogen for 60 hours. 8.5 g. (56 mmol.) of methyl salicylate in 300
ml. of benzene was added and the reaction mixture was refluxed for 3 days
under nitrogen. The reaction mixture was poured into 400 ml. of water
and the organic phase was concentrated to yield a gray residue.
Recrystallization from methylene chloride in n-hexane gave 10.1 g. of

NMR clean crystals of carbinol (45%). 1H NMR (CDC13, 299.95 MHz):
3.40 (S, 12H), 6.58 (d, J=8 Hz, 4H), 6.64 (m, 2H), 6.86 (dd, J=8,2 Hz, 1H),

 

 

 

 

 

235
7.09 (m, 1H), 7.16 (t, J=8 HZ, 2H), 9.15 (s, 1H) 13C NMR (75.43 MHZ):

56.62, 81.37, 107.00, 115.74, 118.51, 123.89, 127.21, 127.93, 128.00,
131.73, 157.37, 158.25.

 

2,6-dimethoxyphenyl lithium was prepared by the addition of 34.8 ml.
(86.9 mmol.) of 2.5 M n-BuLi in n-hexane to a solution of 10 g. (72.4

mmol.) of dimethoxy benzene in sodium-dried diethyl ether at ~780C, and
the mixture was stirred at room temperature under argon for 48 hours.
3.10 g. (16.0 mmol.) of dimethylisophthalate in 200 ml of benzene was
added and the reaction mixture was refluxed for 3 days under argon. The
reaction mixture was poured into 200 ml. of ice water and the organic
layer was separated, dried with sodium sulfate, and concentrated to yield a
gray residue. Recrystallization from ether gave 2.23 g. of the dicarbinol as
white crystals (21%). 1H NMR (CDC13, 299.95 MHz): 3.26 (s, 24H), 6.37
(s, 2H), 6.30 (d, J=8 Hz, 8H), 7.02 (t, J=8 Hz, 4H), 7.08 (m, 1H), 7.32 (dd,
J=8,2 Hz, 2H), 7.44 (t, J=2 Hz, 1H) 13C NMR (75.43 MHz): 55.77, 79.42,
106.23, 123.56, 124.53, 126.37, 126.53, 146.88, 157.60.

Bis-1,4-(tetra-(2,6-dimethoxyphenyl)methyl dicarbinol 94: A solution of
2,6-dimethoxyphenyl lithium was prepared by the addition of 34.8 ml.

(86.9 mmol.) of 2.5 M n-BuLi in n-hexane to a solution of 10.0 g. (72.4
mmol.) of dimethoxy benzene in sodium-dried diethyl ether at -78°C and
the mixture was stirred at room temperature under argon for 48 hours.
3.10 g. (16.0 mmol.) of dimethyl terephthalate in 200ml of benzene was
added and the reaction mixture was refluxed for 3 days under argon. The

reaction mixture was poured into 200 ml. of ice water and the organic

 

 

 

 

 

 

 

 

 

236
layer was separated, dried with sodium sulfate, and concentrated to yield a

gray residue. Recrystallization from ether gave dicarbinol as white
crystals. (4g, 37 %). 1H NMR (CDC13, 299.95 MHz): 3.33 (s, 24H), 6.18
(s, 2H), 6.48 (d, J=8 Hz, 8H), 7.06 (t, J=8 Hz, 4H), 7 .16 (m, 2H), 7.34 (m,
4H) 13C NMR (75.43 MHz): 56.22, 79.50, 106.73, 124.50, 125.02, 126.95,
128.00, 128.50, 158.03.

 

Ox -4 4' dicarbinol : A solution of phenyl lithium was prepared by the
addition of bromobenzene (39.2g., 0.250 mol.) in 125 ml. of ether to
lithium wire (4.0 g., 0.577 mol.) and 60 ml. of ether. Resorcinol dimethyl
ether (30.0 g., 0.218 mole) was added and the reaction mixture was
allowed to stand at room temperature under nitrogen for 60 hours. 14.3 g.
( 50 mmol.) of 4-oxy-dimethylbenzoate in 400 ml. of benzene was added
and the reaction mixture was refluxed for 3 days under nitrogen. The
reaction mixture was poured into 600 ml. of water and the organic phase
was concentrated to yield a gray residue. Recrystallization from methylene
chloride in n-hexane gave 12.8 g. of NMR clean crystals of the dicarbinol
(33%). 1H NMR (CDC13, 299.95 MHz): 3.40 (s, 24H), 6.38 (s, 2H), 6.53
(d, J=8 Hz, 8H) 6.83 (d, J=8 Hz, 4H), 7.10 (t, J=8 Hz, 4H), 7.34 (d, J=8 Hz,
4H) 13C NMR (75.43 MHz): 56.45, 79.24, 106.84, 116.96, 126.48,
127.32, 127.89, 143.96, 155.41, 158.09.

 

 

 

 

 

 

 

237
Tris-1,3,5-(bis-2,6-dimethoxyphenyl)hydroxymethyl benzene 26: A

solution of phenyl lithium was prepared by the addition of bromobenzene
(3.92g., 0.0250 mole) in 12.5 ml. of ether to lithium wire (0.4 g., 0.058
mol..) and 10 ml. of ether. Resorcinol dimethyl ether (3.0 g., 0.0218
mole) was added and the reaction mixture was allowed to stand at room
temperature under nitrogen for 60 hours. 1.5 g. (5.9 mmol.) of 1,3,5-
trimethylbenzoate in 400 ml. of benzene was added and the reaction
mixture was refluxed for 3 days under argon. The reaction mixture was
poured into 100 ml. of water and the organic phase was concentrated to
yield 1.2 g. of gray residue (21%). 1H NMR (CDC13, 299.95 MHz): 3.20
(s, 36H), 6.35 (d, J=8 Hz, 12H), 6.95 (t, J=8 Hz, 6H), 7.33 (s, 3H), 7 .45 (s,
3H).

4 4'-bi hen llinked dicarbinol 97:

l-(3,5-dimethoxyphenyl135-dimethoxybenzene : A solution of 2 g. (7.6
mmol.) of 3,5-dimethoxy-iodobenzene, Mg (7.6 mmol.), CuI (7.6 mmol.),

dibromoethane (0.01 mmol.) in 30 ml. of THF was refluxed for 3 hours
under argon. Iodo-2,6-dimethoxybenzene was added and the reaction
mixture was refluxed further overnight. The solution was filtered,
evaporated, extracted with methylene chloride, dried with magnesium
sulfate, concentrated, and chromatographed with 10% ethyl acetate in n-
hexane to give 870 mg. of white solid (83%). 1H NMR (CDC13, 299.95
MHz): 3.82 (s, 12H), 6.45 (t, J=8 Hz, 2H), 6.69 (d, J=8 Hz, 4H) 13C NMR
(75.43 MHz): 54.99, 99.02, 105.06, 143.00, 160.51 m/z 274 (M+), found
274.

 

 

 

238
1-( 3 ,5-dimethoxyphenyl )-3 ,5-dimethoxy dimethylbenzoate: 1-(3,5-di-

methoxypheny1)-3,5—dimethoxybenzene was lithiated with n-BuLi in dry
THF and quenched with 10 fold excess of dimethyl carbonate and washed
with water twice and extracted with diethyl ether followed by drying and
evaporation of solvent. Chromatography over silica gel using 20 % ethyl
acetate in hexane gave the product as white solids (23%). 1H NMR (CDC13,
299.95 MHz): 3.87 (s, 12H), 3.91 (s, 6H), 6.64 (s, 4H) 13C NMR (75.43
MHz): 52.51, 56.21, 103.45, 144.76, 157.49, 166.72.

1,3-benzenedicarboxaldehyde : A solution of 1,3-phenylene diamine (50g,
379 mmol.), hexamethylene tetraamine (188.3g, 1.345 mol.) in 90 ml. of
conc. HCl and 600 m1. of 50% acetic acid in a 2 liter round bottom flask
was refluxed for 3 hours with stirring. The hot amber reaction mixture is
poured into a flask and a solution of 56.2 g. of NaOH in 750 ml. of water
was slowly added with stirring. The mixture was covered and allowed to
stand overnight at 5°C. The product was collected and washed with 10 ml.
of cold water and dried (37.2g, 73 %). 1H NMR (CDC13, 299.95 MHZ):
7.72 (t, J=3 Hz, 1H), 8.14 (dd, J=4,1 Hz, 2H), 8.36 (m, 1H), 10.12 (s, 2H)
13C NMR (75.43 MHz): 129.50, 130.60, 134.22, 136.54, 190.65.

4- 3- 3-oxo-1-buten l hen l -3-buten-2-one: To a stirred solution of 10
g. (57.4 mmol.) of 1,3-benzenedicarboxaldehyde in 200 m1. of water and

200 m1. of diethyl ether, 100 ml. of acetone and 5 g. of NaOH were added
and the mixture was stirred at the room temperature for 2 days. The
organic layer was separated by adding 200 ml. of diethyl ether and 100 ml.

of acetone and dried with magnesium sulfate. The combined organic

 

 

 

 

239
solution was evaporated and chromatographed over silica gel using 20%

ethyl acetate in n—hexane to give the diketone as white crystals. (5.5 g.,
45%). 1H NMR (CDC13, 299.95 MHz): 2.37 (s, 6H), 6.73 (d, J=l6 Hz, 2H),
7.46 (m, 2H), 7.54 (m, 3H), 7.67 (s, 1H) 13C NMR (75.43 MHz): 27.70,
127.86, 127.93, 129.58, 129.80, 135.19, 142.22, 198.05 m/z 214 (M+),
found 214.

2-hydroxy-6-(3-13—hydroxy-2-methoxycarbonyl-5-oxo-3-cyclohexenyl)

phenyl)-4-oxo-2-cyclohexenecarboxylate: To a solution of sodium
methoxide (2.3g, 100 mg. atom) in 200 ml. of methanol, dimethyl

malonate (12.36g, 93.63 mmol.) was added and the mixture stirred for 30
min at room temperature under argon. 4-(3-(3-0xo-1—butenyl)phenyl)-3-
buten—2—one (6.68g, 31.21 mmol.) was added portion wise with stirring.
The reaction mixture was refluxed for 3 hours under argon and allowed to
cool to room temperature. The solvent was distilled under reduced
pressure, the oily residue was treated with acidic (pH=5) water, and
extracted with methylene chloride. The solution was dried over
magnesium sulfate, solvent was evaporated, and the product mixture was
dried under vacuum. The crude mixture was crystallized with methanol

(10.7g, 83%) m/z 414 (M+), found 414.

1,3-Bis-( 3 ,5-dimethoxyphenyl )benzene:
1H NMR (CDC13, 299.95 MHz): 3.84 (s, 12H), 6.47 (t, J=2 Hz, 2H), 6.75

(d, J=2 HZ, 4H), 7.50 (m, 3H), 7.74 (s, 1H).

Di- 2 6-dimethox hen l - 4- rid l methanol 99: A solution of phenyl
lithium was prepared by the addition of bromobenzene (39.2g., 0.250

 

240
mole) in 125 ml. of ether to lithium wire (4.0 g., 0.577 mol.) and 60 ml.

of ether. Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added and
the reaction mixture was allowed to stand at room temperature under
nitrogen for 60 hours. 13.5 g. ( 98.6 mmol.) of methyl isonicotinate in 400
m1. of benzene was added and the reaction mixture was refluxed for 3 days
under nitrogen. The reaction mixture was poured into 600 ml. of water
and the organic phase was concentrated to yield a gray residue.
Recrystallization from methylene chloride in n-hexane gave 14 g. of NMR
clean crystals of carbinol (37 %). 1H NMR (CDC13, 299.95 MHz): 3.39 (s,
12H), 6.52 (d, J=8 Hz, 4H), 6.65 (s, 1H), 7.13 (t, J=8 Hz, 2H), 7.36 (dd,
J=8,2 Hz, 2H), 8.40 (dd, J=8,2 Hz, 2H) 13C NMR (75.43 MHz): 56.02,
78.80, 106.30, 121.55, 124.41, 127.82, 148.31, 157.74, 158.50.

Di-(2,6-dimethoxy phenylz-(3-pyridyl) methanol 100: A solution of

phenyl lithium was prepared by the addition of bromobenzene (39.2g.,
0.250 mole) in 125 ml. of ether to lithium wire (4.0 g., 0.577 mol.) and 60
ml. of ether. Resorcinol dimethyl ether (30.0 g., 0.218 mole) was added
and the reaction mixture was allowed to stand at room temperature under
nitrogen for 60 hr. 13.5g ( 98.6 mmol.) of methyl nicotinate in 400 m1. of
benzene was added and the reaction mixture was refluxed for 3 days under
nitrogen. The reaction mixture was poured into 600 ml. of water and the
organic phase was concentrated to yield a gray residue. Recrystallization
from methylene chloride in n—hexane gave 15.5 g. of NMR clean crystals
of the carbinol (41 %). 1H NMR (CDC13, 299.95 MHz): 3.40 (s, 12H),
6.51 (d, J=8 Hz, 4H), 6.63 (s, 1H), 7.12 (t, J=8 Hz, 3H), 7.70 (dt, J=8,2 Hz,
1H), 8.34 (dd, J=6,2 Hz, 1H), 8.64 (d, J=2 Hz, 1H) 13C NMR (75.43 MHz):

 

 

 

 

241
13C NMR (75.43 MHZ): 56.16, 78.20, 106.36, 121.73, 124.82, 127.76,

133.73, 144.46, 146.39, 148.93, 157.83.

Di-(2,6-dimethoxy phenyl)(2-pyridyl) methanol 101: A solution of

phenyl lithium was prepared by the addition of bromobenzene (39.2g.,
0.250 mole) in 125 ml. of ether to lithium wire (4.0 g., 0.577 mol..) and
60 ml. of ether. Resorcinol dimethyl ether (30.0 g., 0.218 mole) was
added and the reaction mixture was allowed to stand at room temperature
under nitrogen for 60 hours. 13.5 g. ( 98.6 mmol.) of methyl picolinate in
400 ml. of benzene was added and the reaction mixture was refluxed for 3
days under nitrogen. The reaction mixture was poured into 600 ml. of
water and the organic phase was concentrated to yield a gray residue.
Recrystallization from methylene chloride in n-hexane gave 23 g. of NMR
clean crystals of the carbinol (61 %). 1H NMR (CDC13, 299.95 MHz):
3.39 (s, 12H), 6.49 (d, J=8 Hz, 4H), 6.56 (s, 1H), 7.00(m, 1H), 7.09 (t, J=8
Hz, 2H), 7.50 (m, 2H), 8.45 (m, 1H) 13C NMR (75.43 MHz): 56.30, 79.44,
106.24, 120.45, 121.70, 127.32, 134.73, 146.71, 158.07, 168.46.

 

 

 

 

 

 

242
REFERENCES

   

 

1. C. J. Pedersen, J. Am. Chem. Soc. 89, 7017, 1967.D. J. Cram, Angew. Chem.
Int. Ed. Engl., 25, 1039, 1986. D. J. Cram, Science, 240, 760, 1988. J.-M.
Lehn, Angew. Chem. Int. Ed. Engl., 27, 89, 1988.

2 H. G. Viehe, R. Merényi, L. Stella, Z. Janousek, Angew. Chem. Int. Ed. Engl.
18, 917, 1979; M. Ballester, Acc. Chem. Res. 18, 380, 1985.

3 For carbenes with particular relevance to organic ferromagnets, see: A. M.
Trozzolo, R. W. Murray, G. Smolinsky, W. A. Yager, E. Wasserman, J. Am.
Chem. Soc. 85, 2526, 1963; K. Itoh, Chem. Phys. Lett. 1, 235, 1967; T.
Takui, K. Itoh, Chem. Phys. Lett. 19, 120, 1973. A. Izuoka, S. Murata, T.
Sugawara, H. Iwamura, J. Am. Chem. Soc. 107, 1786, 1985; Y. Teki, T. Takui,
K. Itoh, H. Iwamura, K. Kobayashi, J. Am. Chem. Soc. 108, 2147, 1986; I.
Fujita, Y. Teki, T. Takui, T. Kinoshita, K. Itoh, F. Miko, Y. Sawaki, H.
Iwamura, A. Izuoka, T. Sugawara, J. Am. Chem. Soc. 112, 4074, 1990.

 

 

4 W. T. Borden, Ed.,Diradicals; John Wiley &' Sons, Inc., 1982; G. L. Closs, R.
J. Miller, 0. D. Redwine, Acc. Chem. Res. 18, 196, 1985.; M. Platz (Ed.)
Symposia-in~Print No. 4, Tetrahedron, 38, 733-867; J. A. Berson, Acc. Chem.
Res. 11, 446, 1978.; C. Doubleday, Jr.; N. J. Turro, J.~F. Wang, Acc. Chem.
Res. 22, 199, 1989.

5 For an excellent discussion of exchange phenomena in bimetallic systems see: R.
D. Willett, D. Gatteschi, O, Kahn Ed.;"Magneto—Structural Correlations in
Exchange-Coupled Systems," NATO ASI Series C, V. 140, D. Reidel, Boston,
1985.; O. Kahn, Structure and Bonding, 68, 91, 1987.; A. Bencini, D. Gatteschi,
Electron Paramagnetic Resonance of Exchange Coupled Systems, Springer-Verlag,
Berlin, 1990.

 

6 V. Bonacic'—Koutecky', J. Koutecky’, J. Michl, Angew. Chem. Int. Ed. Eng].
26, 170, 1987.

7 H. McConnell, J. Chem. Phys. 39, 1910, 1963.; Proc. Robert A Welch Found.
108,11,1967.

8 C. M. Hurd Contemporary Physics 23, 469 1982.

9 197th National American Chemical Society meeting in Dallas, Texas, April 10-14,
1989. Proceedings of the Symposium on Ferromagnetic and High Spin Molecular
Based Materials in Mol. Cryst. Liq. Cryst. 176, 1-562 (1989); J. S. Miller; D. A.
Dougherty, Angew. Chem. Int. Ed. Engl., Adv. Mater. 28, 961 (1989).

10 w. E. Broderick, J. A. Thompson, E. P. Day, B. M. Hoffman, Science, 249, 402
(1990); J. S. Miller, A. J. Epstein, in: Z. Yoshida, T. Shiba, Y. Oshiro (Eds)
New Aspects of Organic Chemistry I, VCH, Weinheim, 1989; p 237; J. S.

¥

 

 

 

243

 

11

12

l3

14

15

16

17

18

19

Miller, A. J. Epstein, W. M. Reiff, Chem. Rev. 88, 201 (1988). J. S. Miller, A.
J. Epstein, W. M. Reiff, Science 240, 40 (1988). Chem. Eng. News, p. 5 (Mar.
30, 1987) and p. 28 (Nov. 9, 1987). J. S. Miller, J. C. Calabrese, H.
Rommelmann, S. R. Chittipeddi, J. J. Zhang, W. M. Reiff,'A. J. Epstein, J. Am.
Chem. Soc. 109, 769 (1987). J. S. Miller, A. J. Epstein, J. Am. Chem. Soc.
109, 3850 (1987).

P. W. Selwood, Magnetochemistry, Interscience, New York, 1943; A. P.
Cracknell, Magnetism in Crystalline Materials, Pergamon Press, Oxford, 1975;
R. L. Carlin, Magnetochemistry, Springer-Verlag, Berlin, 1986.

For early musings on organic ferromagnetism see ref. 9 and: N. Mataga, Theoret.
Chim. Acta 10, 372 (1968); A. A. Ovchinnikov, Theoret. Chim. Acta, 47, 297
(1978); A. L. Buchachenko, Doklady Nauk Acad. SSSR 244, 1146 (1979); I. A.
Misurkin, A. A. Ovchinnikov, Russ. Chem. Rev. (Engl. Trans.) 46, 967 (1977);
R. Breslow, B. Jaun, R. Q. Klutz, C.-Z. Xia, Tetrahedron, 38, 863 (1982).

For some more recent magnetic models see: H. Fukutome; A. Takahashi; M.
Ozaki, 133, 34 (1987); L. Dulog, J. S. Kim, Angew. Chem. Int. Ed. Engl., 29,
415 (1990).; K. Yamaguchi, Y. Toyoda, T. Fueno, Synthetic Metals, 19, 81
(1987).; K. Yamaguchi, Y. Toyoda, M. Nakano, T. Fueno, ibid., 19, 87
(1987).; K. Awaga, T. Sugano, M. Kinoshita, J. Chem. Phys., 85, 2211
(1986). K. Awaga, T. Sugano, M. Kinoshita, Chem. Phys. Lett., 141, 540
(1987). P. M. Lahti, A. S. lchimura, Mol. Cryst. Liq. Cryst., 176, 125 (1989).
P. M. Lahti, A. S. lchimura, J. A. Berson, J. Org. Chem., 54, 958 (1989).; K.
M. Chi, J. C. Calabrese, J. S. Miller, Mol. Cryst. Liq. Cryst., 176, 173 (1989).
K. M. Chi, J. C. Calabrese, J. S. Miller, S. I. Khan, Mol.Cryst. Liq. Cryst.,
176, 185 (1989).

DJ. Cram, Angew. Chem. Int. Ed. Engl., 27, 1009 (1988).; J.-M. Lehn, Angew
Chem. Int. Ed. Engl. 27, 89 (1988).

CPK models are particularly faithful for ion binding ligands and their interactions
and have consistently provided insights into complexant design. In the present
case, two CPK models of radical l neatly encapsulate a marble of appropriate size

to represent a Li+ ion.

The methoxy groups in the analogous tris(2,6~dimethoxyphenyl)phosphine
participate in the ligation of a variety of metal species. See: K.R. Dunbar, S. C.
Haefner, L. E. Pence, J. Am. Chem. Soc., 111, 5504 (1989).

J.-M. Lehn, Angew. Chem. Int. Ed. Engl., 27, 89, 1988.

E. Weber, Top. Curr. Chem, 140, 1, 1987. S. Kamitori, K. Hirotsu and T.
Higuchi, J. Am. Chem. Soc. , 109, 2409, 1987. W. Saenger, Angew. Chem,
92, 343, 1980, F. Vogtle and W. M. Miiller, Angew. Chem. Int. Ed. Engl., 18,
628, 1979.

K. Odashima and K. Koga, in Cyclophanes, Vol. 2. Academic Press, New York,
629, 1983. K. Odashima, A. Ita, Y. Litaka and K. Koga, J. Org. Chem, 50,
4478, 1985.

 

 

 

 

 

244

 

20

21

22

23
24

25

26

27

28

29

30

31

32

33

F. Vogtle and W. M. Muller, Angew. Chem. 96, 711, 1984, F. V'dgtle and W.
M. Miiller Angew. Chem. Int. Ed. Engl., 25, 567, 1986.

H. J. Schneider and T. Blatter Angew. Chem. 100, 1211, 1988. H. J. Schneider
and T. Blatter Angew, Chem. Int. Ed. Engl., 27, 1163, 1988. H.—J. Schneider
and I. Theis, Angew. Chem, Int. Ed. Engl., 28, 753, 1989. H.-J. Schneider, D.
Guttes and U. Schneider, J. Am. Chem. Soc. 110, 6449, 1988. F. Diederich and
K. Dick, Angew. Chem. Int. Ed. Engl. 22, 715, 1983. F. Diederich and K.
Dick, Angew. Chem. Int. Ed. Engl. 23, 810, 1984. F. Diederich and D. R.
Carcanague, Angew. Chem. Int. Ed. Engl. 29, 769, 1990.

G. R. Desiraju "Crystal Engineering: the Design of Organic Solids” Materials
Science Monographs, 54, Elsevier, New York, NY. 1989.

G. M. J. Schmidt, Pure Appl. C hem, 27 , 647, 1971. and references therein.
H. R. Ott Nature 363, 56, 1993.

M. A. Beno, H. H. Wang, K. D. Carson, A. M. Kini, G. M. Frankenbach, J. R.
Ferraro, N. Larson, G. D. McCabe, J. Thomson, C. Purnama, M. Vashon, J. M.
Williams, D. Jung, M. -H. Whangbo., Mol. Cryst. Liq. Cryst. 1990, 181, 145.
M. A. Beno, H. H. Wang, A. M. Kini, K. D. Carson, U. Geiser, W. K. Kwok,
J. E. Thompson, J. M. Williams, J. Ren, M. -H. Whangbo, Inorganic. Chem.
1990,29, 1599.

D. Griller and K. U. Ingold Acc. Chem. Res. 9, 13, 1976.

S. W. Benson "Thermochemical Kinetics", Wiley, New York, N.Y., 1968, D.
M. Golden and S. W. Benson, Chem. Rev. 69, 125, 1969, H. E. O'Neal and S.
W. Benson in ”Free Radicals”, Vol II. J. K. Kochi, Ed., Wiley, New York, NY.
197 3.

"T he History of Organic Chemistry in the United States, J 875 -I 955 " D. S. Tarbell
and A. T. Tarbell, K & S Press, Nashville, TN, 1986.

H. Lankarnp, W. Th. Nauta, C. MacLean, Tett. Lett. 249, 1968. P. Jacobson,
Ber. Dtsch. Chem. Ges. 38, 196, 1905.

M. Stein, W. Winter, and A. Rieker Angew. Chem. Int. Ed. Engl. 17, 692,
1978. B. Kahr, D. V. Engen, and K. Mislow J. Am. Chem. Soc. 108, 8305,
1986.

W. Schlenk and M. Brauns, Ber. Dtsch. Chem. Ges. 48, 661, 1915.

G. Kothe, K.-H. Denkel, W. Sfimmermann Angew. Chem. Int. Engl. Ed.,9, 906,
1970.

A. E. Chichibabin, Chem. Ber. 40, 1810, 1907.

 

 

 

 

 

 

245

 

34

35

36

37

38
39
40

41

42

43

44

45
46

47

48

M. S. Platz, Diradicals, W. T. Borden, Ed., Wiley, New York, 1982, Chapter 5.

L. K. Montgomery, J. C. Huffman, E. A. Jurczak, M. P. Grendze, J. Am. Chem.
Soc. 108, 6004, 1986.

J. Thiele, H. Balhorn, Chem. Ber. 37, 1463, 1904.

M. Ballester, C. Molinet, Chem. Ind. (London) 1290, 1954. M. Ballester, C.
Molinet, J. Castaner, J. Am. Chem. Soc. , 82, 4254, 1960, M. Ballester, R.
Riera, J. Am. Chem. Soc., 86, 4505, 1964.

M. Ballester, Pure Appl. Chem. 15, 123, 1967.
M. Ballester, J. Castaner, J. Riera, J. Am. Chem. Soc. 88, 957, 1966.

M. J. Sabacky, C. S. Johnson, Jr., R. G. Smith, H. S. Gutowsky, J. C. Martin,
J. Am. Chem. Soc. 89, 2054 (1967).

While EPR spectra observed by Jackson and Kahr remain unassigned at least one
feature is clear; the spectral dispersion—as large as 800 Gauss in some cases—is
too great to be due solely to hyperfine interactions and must be fine structure
resulting from long-range magnetic coupling of pairs of radicals. The temperature
dependence of the signal intensity indicates that its carrier might be species with a
high Spin ground state. The Spectra were unchanged when LiBF4 was replaced by

LiClO4 or 99% 61.113134.

For the status of research on molecular magnetism in 1985, see H. Iwamura, T.
Sugawara, K. Itoh, T. Takui, Mol. Cryst. Liq. Cryst.125, 251 (1985) R.
Breslow, Mol. Cryst. Liq. Cryst. 125, 261 (1985)

For a valuable enumeration of these models, see D. Dougherty and J. S. Miller in
Mol. Cryst. Liq. Cryst. 176, 25 (1989).

S. Goldschmidt, Ber. 53, 44, 1920, S. Goldschmidt and K. Euler Ber. 55, 616,
1922. S. Goldschmidt and K. Renn, Ber. 55, 628, 1922. S. Goldschmidt, and
F. Graef, Ber. 61, 1858, 1928. R. H. Poirier, E. J. Kahler, and F. Benington, J.
Org. Chem. 17, 1437, 1952.

L. S. Singer, and C. Kikuchi, J. Chem. Phys. 23, 1738, 1955.
G. Fraenkel and PD. Bartlett, J. Am. Chem. Soc. 81, 5582, 1959.

F. L. Allen and S. Sugden, J. Chem. Soc. 440, 1936. W Duffy, J. Chem. Phys.
36, 490, 1962, W. Duffy and D. L. Strandburg, J. Chem. Phys. 46. 457, 1967.

C. A. Hutchison, R.C. Pastor, and AG. Kowalsky J. Chem. Phys. 20, 534,
1952. N. W. Lord and S. M. linder, J. Chem. Phys. 34, 1693, 1961. R. M.
Deal and W. S. Koski, J. Chem. Phys. 31, 1138, 1959. R. W. Holmberg, R.
Livingston, and W. T. Smith, J. Chem. Phys. 33, 541, 1960. R. 1. Walter, J.

 

 

 

 

 

 

 

 

 

246

 

49

50

51

52

53

54

55

56

57

58

59

60

61

Am. Chem. Soc. 88, 1923, 1966, R. 1. Walter, J. Am. Chem. Soc. 88, 1930,
1966.

J. Heidberg and J. Weil, J. Am. Chem. Soc.86, 5173, 1964. A. L.
Buchachenko, "Stable Radicals", Consultants Bureau, New York. 1965.

O. Piloty and BC. Schwerin. Ber. 34, 1870, 1901. O. Piloty and B. G.
Schwerin Ber. 34, 2354, 1901.

K. H. Meyer and H. G. Bilbroth Ber. 52, 1476, 1919. K. H. Meyer and W.
Reppe, Ber. 52, 1476, 1919.

Y. Y. Lim and R. S. Drago, J. Am. Chem. Soc. 93, 891, 1970. Y. Y. Lim and
R. S. Drago, Inorganic Chem. 11, 1334, 1972.

R. Briere, H. Lemaire, and A. Rassat, Bull Soc. Chim. France, 3273, 1965. L.
Chapelet, H. Lemaire, and A. Rassat, Bull. Soc. Chim. F ranmce, 3283, 1965. E.
G. Rozantsev, Bull. Acad. Sci. USSR. 2085, 1964.

A.M. Feldman and A. K. Hoffmann Chem. Abst. 61, 13289, 1964. A. W.
Hanson, Acta Cryst. 6, 32, 1953.

K. H. Meyer and W. Reppe Ber. 54, 327, 1921. A. K. Hoffmann, A. M.
Feldman, and E. Gelblum, J. Am. Chem. Soc. 86, 646, 1964.

C. D. Cook, J. Org. Chem. 18, 261, 1953. C. D. Cook and R. C. Woodworth,
J. Am. Chem. Soc. 75, 6242, 1953.

N. C. Yang and A. Castro, J. Am. Chem. Soc. 82, 6208, 1960. E. A. Chandross
J. Am. Chem. Soc. 86, 1263, 1964.

F. A. Neugebauer, H. Tischmann, and M. Jenne, Angew. Chem. Int. Ed., 6, 362,
1967. E. A. Chandross, J. Am. Chem. Soc. 86, 1263, 1964. K. Mukai and J.
Sakamoto J. Chem. Phys, 68, 1432, 1978.

A. Earnshaw, Introduction to Magnetochemistry, Academic Press, New York,
1968. R. L. Carlin , Magnetochemistty, Springer—Verlag, New York, 1986.

B. Morosin and E. J. Graeber. Acta Cryst. 16, 1176, 1963. L. J. de Jongh and
A. R. Miedema, Adv. Phys. 23, 1, 1974. D.W. Hone and P. M. Richards, Ann.
Rev. Material Sci. 4, 337, 1974. R. Dingle, M. E. Lines, and S. L. Holt, Phys.
Rev. 187, 643, 1969.

T. Oguchi, Prog. Theor. Phys., 13, 148, 1955. H. Ohya-Nishiguchi, Bull.
Chem. Soc. Japan. 52, 3480, 1979. J. C. Bonner, M. E. Fisher, Phys. Rev.
135, A640, 1964. J. C. Bonner, H. W. J. Blote, J. W. Bray, and I. S. Jacobs,
J. App. Phys. 50, 1810, 1979. M. Verdaguer, A. Gleizes, J. P. Renard, and J.
Seiden, Phys. Rev. B. 29, 5144, 1984. W. Duffy, Jr. and K. P. Barr, Phys.
Rev. 165, 647, 1968.

 

 

 

 

 

 

 

 

247

 

62

63
64

65

66

67

68

69

70

71

72

73

74

75

M. E. Fisher, J. Phys. 32, 343, 1964. A. van der Bilt, K. O. Joung, and R. C.
Carlin, Phys. Rev. B. 24, 445, 1981. M. M. Olmstead, W. K. Musker, L. W.
Ter Haar, and W. E. Hatfield. J. Am. Chem. Soc. 104, 6627, 1982. J. W. Hall,
W. E. Marsh, R. R. Weller, and W. E. Hatfield, Inorg. Chem. 20, 1033, 1981.,
W. E. Hatfield, R. R. Weller, and J. W. Hall, Inorg. Chem. 19, 3825, 1980., W.
H. Crawford and W. E. Hatfield, Inorg. Chem. 16, 1336, 1977. A. Bencini, C.
Benelli, D. Gatteschi, and C. Zanchini, J. Am. Chem. Soc. 106, 5813, 1984. C.
Benelli, D. Gatteschi, D. W. Carnegie, Jr., and R. L. Carlin, J. Am. Chem. Soc.
107, 2560, 1985.

A. S. Edelstein, M. Mandel, J. Chem. Phys. 35, 1130, 1961.
H. M. McConnell J. Chem. Phys. 37, 1910, 1963.

H. McConnell, Proc. of the Robert A. Welch Foundation C onf. on Chem. Res. 11,
144, 1967 .

N. Mataga, Theoret. Chim. Acta (Berl.) 10, 372, 1968.

A. A. Ovchinnikov, Theoret. Chim. Acta. (Berl.) 47, 297, 1978. A. A.
Ovchinnikov, V. N. Spector Synthetic Metals, 27 , B615, 1988.

R. Breslow Pure & Appl. Chem. 54, 927, 1982. R. Breslow, B. Jaun, R. Q.
Kluttz, C.—Z. Xia Tetrahedron, 38, 863, 1982.

T. LePage, R. Breslow, J. Am. Chem. Soc. 109, 6412, 1987. R. Breslow.
Mol. Cryst. Liq. Cryst., 176, 199, 1989.

J. B. Torrance, S. Oostra, A. Nazzal, Synthetic Metals, 19, 709, 1987.

C. Kollmar, O. Kahn, J. Am. Chem. Soc. 113, 7987, 1991, C. Kollmar, M.
Couty, and O. Kahn, J. Am. Chem. Soc. 113, 7994, 1991.

C. Kollmar and O. Kahn, Acc. Chem. Res. 26, 259, 1993. A. Caneschi, D.
Gatteschi, and R. Sessoli, Acc. Chem. Res. 22, 392, 1989, J. S. Miller, A. J.
Epstein, and W. M. Reiff, Acc. Chem. Res. 21, 114, 1988, J. S. Miller, A. J.
Epstein, and W. M. Reiff, Chem. Rev. 88, 201, 1988. O. Kahn, Angew. Chem.
Int. Ed. Engl. 24, 834, 1985.

B. Bleaney, K. D. Bowers, Proc. R. Soc. London A 214, 451, 1952.

D. Gatteschi, O. Kahn, R. D. Willett (eds): Magneto Structural Correlations in
Exchange Coupled Systems, D. Reidel, Dordrecht 1984.

M. Verdaguer, M. Julve, A. Michalowicz, and O. Kahn, Inorg. Chem. 22, 2624,
1983. Y. Pei, O. Kahn, J. Sletten., J. Am. Chem. Soc. 108, 3143, 1986., Y.
Pei, M. Verdaguer, O. Kahn, J. Sletten, and J. P Renard, J. Am. Chem. Soc.
108, 7428, 1986. Y. Pei, M. Verdaguer, O. Kahn, J. Sletten, and J. P. Renard,
Inorg. Chem. 26, 138, 1987. O. Kahn, Angew. Chem. Int. Ed. Engl. 24, 834,
1985. K. Nakatani, O. Kahn. C. Mathoniere, T. Pei, and C. Zakine, New. J.

 

 

 

 

 

 

 

248

 

76

77

78

79

80

81

82

83

84

Chem. 14, 861, 1990. O. Kahn, Comments Inorg. Chem. 3, 105, 1984. M.
Julve, M. Verdaguer, A. Gleizes, M. P. Levisalles, adn O. Kahn. Inorg. Chem.
23, 3808, 1984.

 

 

 

D. Beltram, E. Escriva, and M. Drillon, J. Chem. Soc., Faraday Trans. 2, 78,
1773, 1982. M. Drillon, E. Coronado, D. Beltran, J. Curely, R. Georges, P. R.
Nugteren, L. J. de Hongh, and J. L. Genicon, J. Magn. Magn. Mater, 54-57,
1507, 1986. M. Drilon, E. Coronado, A. Fuertes, D. Beltran, and R. Georges,
Chem. Phys. 79, 449, 1983. E. Coronado, M. Drillon, A. Fuertes, D. Beltran,
A. Mosset, and J. Galy, J. Am. Chem. Soc. 108, 900, 1986.

O. Kahn Comments Inorg. Chem. 3, 105, 1984. R. D. Willett, D. Gatteschi, O.
Kahn Eds., Magnetostructural Correlations in Exchange Coupled Systems (NATO
Advanced Study Institute Series), D. Reidel, Dordrecht, The Netherlands, 1984.

D. Y. Jeter, D. L. Lewis, J. C. Hemplel, D. J. Hodgson, and W. E. Hatfield,
Inorga. Chem. 11, 1958, 1972.

 

M. Julve, M. Verdaguer, A. Gleizes, M. P. Levisalle, and O. Kahn, Inorg. Chem.
23, 3808, 1984. Y. Pei, K. Nakatani, J. Sletten, J. P. Renard, Inorg. Chem. 16,
3170, 1986.

L. R. Carlin, A. Kopinga, O. Kahn, and M. Verdaguer, Inorg. Chem. 25, 1786,
1986.

S. K. Shakhatreh, E. G. Bakalbassis, I. Brfidgam, H. Hartl, J. Mrozinski, and C.
A. Tsipis, Inorg. Chem. 30, 2801, 1991.

A. Bencini, C. Benelli, D. Gatteschi, and C. Zanchini, J. Am. Chem. Soc. 106,
5813, 1984. C. Benelli, D. Gatteschi, and C. Zanchini, Inorg. Chem. 23, 798,
1984. C. Benelli, D. Gatteschi, D. W. Carnegie, Jr., and R. Carlin, J. Am.
Chem. Soc. 107, 2560, 1985. A. Caneschi, D. Gatteschi, J. Laugier, P. Rey, R.
Sessoli, and C. Zanchini, J. Am. Chem. Soc. 110, 2795, 1988. A. Caneschi, D.
Gatteschi, J. Laugier, P. Rey, and C. Zanchini, Inorg. Chem. 28, 1969, 1989.
A. Caneschi, D. Gatteschi, J. P. Renard, P. Rey, and R. Sessoli, Inorg. Chem.
28, 1976, 1989. A. Caneschi, D. Gatteschi, J. P. Renard, P. Rey, and R.
Sessoli, Inorg. Chem. 28, 3314, 1989. A. Caneschi, D. Gatteschi, and R.
Sessoli, Mol. Cryst. Liq. Cryst. 176, 329, 1989. A. Caneschi, D. Gatteschi, J.
P. Renard, P. Rey, and R. Sessoli, J. Am. Chem. Soc. 111, 785, 1989. P. Rey,
J. Laugier, A. Caneschi, D. Gatteschi, Mol. Cryst. Liq. Cryst. 176, 337, 1989.
A. Caneschi, F. Ferraro, D. Gatteschi, P. Rey, and R. Sessoli, Inorg. Chem. 29,
4217, 1990. A. Caneschi, D. Gatteschi, R. Sessoli, and P. Rey, Acc. Chem. Res.
22, 392, 1989. A. Caneschi, D. Gatteschi, P. Rey, and R. Sessoli, Chem. Mater,
4, 204, 1992.

 

S. S. Eaton, G. R. Eaton, Coord, Chem, Rev. 26, 207, 1978. S. Eaton, S. S. G.
R. Eaton, Coord, Chem. Rev. 83, 29, 1988.

A. Caneschi, D. Gatteschi, A. Grand, J. Laugier, L. Pardi, and P. Rey, Inorg.
Chem. 27 , 1031, 1988. A. Caneschi, D. Gatteschi, J. Laugier, L. Pardi, P. Rey,
and C. Zanchini, Inorg. Chem. 27, 2027, 1988.

 

 

 

 

 

249

 

85

86

87

88

89

90

91

92

93

94

95

A. Bencini, C. Benelli, D. Gatteschi, C. Zanchini, J. Am. Chem. Soc. 106, 5813,
1984. C. Benelli, D. Gatteschi, and C. Zanchini, Inorg. Chem. 23, 798, 1984.

A. Caneschi, D. Gatteschi, J. Laugier, and P. Rey, "Organic and Inorganic Low
Dimensional Crystalline Materials", P. Delhaes, M. Drilon, Eds.; Plenum Press,
New York, London, p109, p389, 1987.

A. Caneschi, D. Gatteschi, P. Rey, R. Sessoli, Inorg. Chem. 27, 1756, 1988, A.
Caneschi, D. Gatteschi, J. P. Renard, P. Rey, adn R. Sessoli, J. Am. Chem. Soc.
111, 785, 1989.

J. S. Miller, J. C. Calabrese, H. Rommelmann, S. R. Chittipeddi, J. H. Zhang,
W. M. Reiff, and A. J. Esptein, J. Am. Chem. Soc. 109, 769, 1987, J. S. Miller
and A. J. Epstein, J. Am. Chem. Soc. 109, 3850, 1987. J. S. Miller, Acc.
Chem. Res., 21, 114, 1988. J. S. Miller, A. J. Epstein, W. M. Reiff, Chem.
Rev., 88, 201, 1988. J. S. Miller, A. J. Epstein, and W. M. Reiff, Science,240,
40, 1988. A. Chakraborty, A. J. Epstein, W. N. Lawless, and J. S. Miller, Phys.
Rev. B, 40, 11 422, 1989. J. S. Miller and A. J. Epstein, Phil Trans. R. Soc.
Lond. A , 330, 205, 1990.

C. Kollmar, O. Kahn, J. Am. Chem. Soc. 113, 7987, 1991, C. Kollmar, M.
Couty, and O. Kahn, J. Am. Chem. Soc. 113, 7994, 1991. C. Kollmar and O.
Kahn, Acc. Chem. Res. 26, 259, 1993. , C. M. Hurd, Contemporary Physics
23, 469, 1982.

K. Mukai, H. Nishiguchi, and Y. Deguchi, J. Phys. Soc. Japan, 23, 125, 1967,
K. Awaga, T. Sugano, and M. Kinoshita, Solid State Comm. 57, 453, 1986. K.
Awaga, T. Sugano, and M. Kinoshita, Chem. Phys. Lett. 128, 587, 1986. K.
Awaga, T. Sugano, and M. Kinoshita, J. Chem. Phys. 85, 2211, 1986. K.
Awaga, T. Sugano, and M. Kinoshita, Chem. Phys. Lett. 141, 540, 1987. M.
Kinoshita, Mol. Cryst. Liq. Cryst. 176, 163, 1989.

D. E. Williams, Mol. Phys. 16, 145, 1969.

K. M. Chi, J. C. Calabrese, J. S. Miller, adn S. I. Khan, Mol. Cryst. Liq. Cryst.
176, 185, 1989.

K. Mukai, Bull. Chem. Soc. Jpn. 42, 40, 1969.

K. Awaga and Y. Maruyama, Chem. Phys. Lett. 158, 556, 1989. K. Awaga, T.
Inabe, U. Nagashima, and Y. Maruyama, J. Chem. Soc. Chem. Comm. 1617,
1989. M. Kinoshita, P. Turek, M. Tamura, K. Nozawa, D. Shiomi, Y.
Nakazawa, M. Ishikawa, M. Takahashi, K. Awaga, T. Inabe, and Y. Maruyama,
Chem. Lett. 1225, 1991. P. Turek, K. Nozawa, D. Shiomi, K. Awaga, T. Inabe,
Y. Maruyama, and M. Kinoshita. Chem. Phys. 180, 327, 1991.

L. Pauling "The Nature of the Chemical Bond", Cornell University Press, New
York, p343, 1960. J. W. Linnett, J. Am. Chem. Soc. 83, 2643, 1961.

 

 

 

 

250

 

96

97

98

99

100

101

102

103

104

105

106

107

108

Y. Y. Lim and R. S. Drago, J. Am. Chem. Soc. 93, 891, 1970. A. M.
Vasserman and A. L. Buchachenko, J. Struct. Chem. 7, 633, 1966.

T. Sugano. M. Tamura, M. Kinoshita, Y. Sakai, and. Y. Ohashi, Chem. Phys.
Lett. 200, 235, 1922.

H. Iwamura, 197th National Meeting of the American Chemical Society, Dallas,
TX, April 9—14, 1989.

W. T. Borden, Diradicals; W. T. Borden, Ed.; Wiley, New York, 1982, W. T.
Borden and E. R. Davidson, J. Am. Chem. Soc. 99, 4587, 1977.

E. E. Seeger, P. M. Lahti, A. R. Rossi, and J. A. Berson, J. Am. Chem. Soc.
108, 1251, 1986.

S. Kato, K. Morokuma, D. Feller, B. Davidson, W. T. Borden, J. Am. Chem.
Soc. 105, 1791, 1983. B. B. Wright, M. S. Platz, J. Am. Chem. Soc. 105, 628,
1983. J. L. Goodman, J. A. Berson J. Am. Chem. Soc. 106 , 1867, 1984, M
Rule, A. Matlin, D. E. Seeger, E. F. Hilinski, and D. A. Dougherty Tetrahedron
38, 787, 1982.

T. Matsumoto, T. Ishida, N. Koga, and H. Iwamura. J. Am. Chem. Soc. 114,
9952, 1992.

A. J. Novak, R. Jain, and D. A. Dougherty, J. Am. Chem. Soc. 111, 7618,
1989. J. A. Berson, Acc. Chem. Res. 11, 446, 1978, P. Dowd., Acc. Chem.
Res. 5, 242, 1972, R. Jain, M. B. Sponsler, F. D. Coms, D. A. Dougherty, J.
Am. Chem. Soc. 110, 1356, 1988. J. A. Berson, Acc. Chem. Res. 11, 446,
1978.

M. Ballester, J. Castaner, J. Riera, A. Ibanez, and J. Pujadas, J. Org. Chem. 47,
259, 1982. M. Ballester, J. Riera, J. Castaner, A. Rodriguez, C. Rovira, and J.
Veciana, J. Org. Chem. 47 , 4498, 1982. M. Ballester, J. Castaner, J. Riera, J.
Pujadas, O. Armet, C. Onrubia, and J. A. Rio, J. Org. Chem. 49, 770, 1984. M.
Ballester, Acc. Chem. Res. 18, 380, 1985.

J. Veciana, A. D. Martinez and O. Armet, Rev. Chem. Intermediates, 10, 35,
1988, J. Veciana and A. Duran. J. Inclus. Pheno. 5, 173, 1987.

O. Armet, J. Veciana, C. Rovira, J. Riera, J. Castaner, E. Molins, J. Rius, C.
Miravitlles, S. Olivella, and J. Brichfeus. J. Phys. Chem. 91, 5608, 1987.

J. Veciana, C. Rovira, M. I. Crespo, O. Armet, V. M. Domingo, and F. Palacio,
J. Am. Chem. Soc. 113, 2552, 1991. J. Carilla, L. Julia, J. Riera, E. Brillas, J.
A. Garrido, A. Labarta, and R. Alcala. J. Am. Chem. Soc. 113, 8281, 1991, J.
Veciana, C. Rovira, N. Ventosa, M. I. Crespo, and F. Palacio J. Am. Chem. Soc.
115, 57, 1993.

A. Rajca, J. Am. Chem. Soc. 112, 5889, 1990, A. Rajca, J. Org. Chem. 56,
2557, 1991, A. Rajca, S. Utamapanya, and J. Xu, J. Am. Chem. Soc. 113,

 

 

 

 

 

 

 

 

 

 

251

 

109

110

111

112

113

114

115

116

117

118

119

120

9235, 1991. S. Utamapanya and A. Rajca, J. Am. Chem. Soc. 113, 9242,
1991., A. Rajca, S. Utamapanya, and S. Thayumanavan J. Am. Chem. Soc.
114, 1884, 1992.

Crystal Structure and Magnetic Susceptibility of a Hydrocarbon Free Radical: Tris
(3,5-di-t-butylphenyl)methyl B. Kahr, D. V. Engen, and P. Gopalan, submitted for
publication.

A. Izuoka, S. Murata, T. Sugawara, and H. Iwamura, J. Am. Chem. Soc. 107,
1786, 1985., H. Iwamura, Pure & Appl. Chem. 58, 187, 1986. A. Izuoka, S.
Murata, T. Sugawara, and H. Iwamura, J. Am. Chem. Soc. 109, 2631, 1987.

K. Itoh, Chem. Phys. Lett. 1, 235, 1967., Y. Teki, T. Takui, K. Itoh, H.
Iwamura, and K. Kobayashi J. Am. Chem. Soc. 108, 2147, 1986., K. Itoh, T.
Takui, Y. Teki, and Takamasa Mol. Cryst. Liq. Cryst.176, 49, 1989. N.
Nakamura, K. Inoue, H. Iwamura, T. Fujioka, and Y. Sawaki, J. Am. Chem.
Soc. 114, 1484, 1992. H. Iwamura and N. Koga, Acc. Chem. Res. 26, 346,
1993.

S. Murata and H. Iwamura, J. Am. Chem. Soc. 113, 5547, 1991. K. Inoue, N.
Koga, and H. Iwamura, J. Am. Chem. Soc. 113, 9803, 1991.

C. Ling, M. Minato, P. M. Lahti, and H. van Willigen, J. Am. Chem. Soc. 114,
9959, 1992.

M. Kato, H. B. Jonassen, and J. C. Fanning, Chem. Rev. 64. 99, 1964.

G. J. Sloan and W. R. Vaughan J. Org. Chem. 22, 750, 1957, G. J. Sloan and
W. R. Vaughan J. Chem. Phys. 25, 697, 1956.

E. G. Rozantsev, V. A. Golubev, M. B. Neiman, adn Y. V. Kokhanov, Bull.
Acad. Sci. USSR. 559, 1965. E. G. Rozantsev, and V. A. Golubev, Bull. Acad.
Sci. USSR. 695, 1965. R. M. Dupeyre, H. Lemaire, and A. Rassat J. Am.
Chem. Soc. 87, 3771, 1965. E. G. Rozantsev, V. A. Golubev and V. A.
Golubev Bull. Acad. Sci. USSR. 382, 1965.

D. F. Evans, J. Chem. Soc., 2003, 1959.

D. G. B. Boocock, R. Darcy and E. F. Ullman, J. Am. Chem. Soc. 90, 5445,
1968, D. G. B. Boocock and E. F. Ullman J. Am. Chem. Soc. 90, 6873, 1968.
E. F. Ullman and D. G. B. Boocock Chem. Comm. 1161, 1969, P. W. Kopf, R.
Kreilick, D.G. B. Boocock, and E. F. Ullman, J. Am. Chem. Soc. 92, 4531,
1970. F. Alies, D. Luneau, J. Laugier, and P. Rey, J. Phy. Chem. 97, 2922,
1993.

P. M. Lahti and A. S. Ichimura Mol. Cryst. Liq. Cryst. 176, 125, 1989.

O. H. Griffith and A. S. Waggoner, Acc. Chem. Res. 2, 17, 1969, H. Dugas
Acc. Chem. Res. 10, 47, 1977. E. Ullman, J. H. Osiecki, D. G. B. Boocock.
and R. Darcy J. Am. Chem. Soc. 94, 7049, 1972.

 

 

 

 

 

252

 

121

122

123

124

125

126

127

128

129

130

K. Mukai, T. Yano, and K. Ishizu, Tett. Lett. 22, 4661, 1981. K. Ishizu, H.
Kohama, and K. Mukai, Chem. Lett. 227, 1978, K. Mukai, N. Iida, Y.
Kumamoto, H. Kohama, and K. Ishizu, Chem. Lett. 613, 1980, K. Mukai, M.
Yamashita, K. Ueda, K. Tajima, and K. Ishizu, J. Phys. Chem. 87, 1338, 1983.
K. Mukai, M. Tanii, Y. Yurugi, K. Tajima, and K. Ishizu Bull. Chem. Soc. Jpn.,
58, 322, 1985.

P. J. Hay, J. C. Thibeault, and R. Hoffmann, J. Am. Chem. Soc. , 97, 4884,
1975.

R. Hoffmann, Acc. Chem. Res. , 4, 1, 1971, R. Gleiter, Angew. Chem. , 86,
770, 1974.

O. Kahn, Angew. Chem. Int. Ed. Engl. 24. 834. 1985. J. B. Goodenough, ”
Magnetism and the Chemical Bond" , Interscience, New York, N.Y., 1963. J.
Kanamori, Phys. Chem. Solid, 10, 87, 1959, P. W. Anderson, Solid State
Phys., 14, 99, 1963.

See also perchlorotriphenylmethyl: M. Ballester, J. Riera, J. Castaner, C. Badia,
M.J. Monsé, J. Am. Chem. Soc. , 93, 2215, 1971.

The ferromagnetic coupling of a pair of nitroxyl substituted benzo-lS-crown-S-
ethers was accomplished with potassium ions. See: K. Ishizu, H. Kohama, K,
Mukai, Chemistry Letters, 227, 1978. See also: K. Mukai, N. Iida, Y.
Kumamoto, H. Kohama, K. Ishizu, Chem. Lett. , 613 1980; K. Mukai, M.
Yamashita, K. Ueda, K. Tajima, K. Ishizu, J. Phys. Chem. 87, 1338 , 1983.; J.
F. W. Keana, J. Cuomo, L. Laszlé, S. E. Seyerdrezai, J. Org. Chem. 48, 2647,
1983; K. Mukai, M. Tanii, Y. Yurugi, K. Tajima, Bull. Chem. Soc. Jpn. 58,
322, 1985.

In related work, glasses of t1is(biphenyly1)methy1 gave triplet coupling, presumably
due to stacking of radicals pairwise to form D3 n-complexes. See: W. Broser, H.
Kurreck, W. Niemeier, Tetrahedron, 32, 1183, 1976.

For a review of one-dimensional ferromagnetism in inorganic solids see: R. D.
Willett, R. M. Gaura, C. P. Landee in J. S. Miller (Ed.) Extended Linear Chain
Compounds, V. 3, Plenum, New York, 1983; p143.

C. Kollmar and O. Kahn Acc. Chem. Res. 26, 259, 1993.

K. Nath, P. L. Taylor, Mol. Cryst. Lig. Cryst., 180B, 389 , 1990.; K.
Yamaguchi, H. Namimoto, T. Fueno, Mol. Cryst. Lig. Cryst., 176, 151 1989.;
H. Iwamura, S. Murata, Mol.Cryst. Liq. Cryst., 176, 33 , 1989.; J. Veciana, C.
Rovira, O. Armet, V. M. Domingo, M. I. Crespo, F. Palacio, Mol. Cryst. Liq.
Cryst., 176, 77, 1989.; D. J. Klein, S. A. Alexander, M. Randic, Mol. Cryst.
Liq. Cryst., 176, 109 1989. T. Hughbanks, M. Kertesz, Mol. Cryst. Liq. Cryst.,
176, 115 , 1989.; J. S. Miller, A. J. Epstein, Mol. Cryst. Liq. Cryst., 176, 347 ,
1989.

 

 

 

 

 

253

 

131

132

133

134

135

136

137

138

139

140

141

142

143

144

H. M. McConnell, J. Chem. Phys., 24, 764, 1956., D. J. E. Ingram, "Free
Radical as Studied by Electron Spin Resonance,” Butterworths, p.111, 1958. H.
Jukeikis and D. Kivelson J. Am. Chem. Soc. 84, 1132, 1962.

O. Armet, J. Veciana, C. Rovira, J. Riera, J. Castaner, E. Molins, J. Rius, C.
Maravitlles, S. Olivella, and J. Brichfeus J. Phys. Chem. 91, 5608, 1987.

A. G. de Mesquita, C. M. Gillarvy, and E. Eriks, Acta Cryst. 18, 437, 1965; P.
Andersen and B. Klewe Acta Chem. Scand. 21, 2599, 1967; A. Port, M. M.
Olmstead, and P. P. Power J. Am. Chem. Soc. 107, 2174, 1985; J. Veciana, J.
Riera, J. Castaner, and N. Ferret, J. Organomet. Chem. 297, 131, 1985.

K.M. Mislow, Acc. Chem. Res. 9, 26 (1976).

M. Ballester, J. Riera, J. Castaner, C. Badia, and J. M. Monso, J. Am. Chem.
Soc. 93, 2215, 1971. K. S. Hayes, M. Nagamo, J. F. Blount, and K. Mislow,
J. Am. Chem. Soc. 102, 2773, 1980.

"Comparison of Twists in Isosteric Propellers: X-ray Structures of Tris(2,6-
dimethoxy-phenyl)borane, Tris(2,6-dimethoxyphenyl)methyl cation, and Tris(2,6-
dimethoxyphenyl)methyl Radical”; B. E. Kahr, J. E. Jackson, D. Ward, S. -H.
Jang, and J. Blount, Acta Cryst., B48, 324, 1992. "Aryl Ring Twists in
Tris(2,6-dimethoxyphenyl)-Z Tripod Ethers: Why is the Radical Different ?" S. J.
Stout, P. Gopalan, J. E. Jackson, and B. Kahr, Submitted for publication.

M. M. Olmstead and P. E. Power, J. Am. Chem. Soc. 108, 4235, 1986.

K. Ishizu, K. Mukai, A. Shibayama, and K. Kondo, Bull. Chem. Soc. Jpn. 50,
2269, 1977.

R. Terrence, O. Toole, N. Janet, B. Younathan, P. Sullivan, and T. J. Meyer
Inorg. Chem. 28, 3923, 1989.

See F. A. Cotton, G. Wilkinson, "Advanced Inorganic Chemistry" Fifth Ed.,
Wiley, NY. 1988, pp. 1288 ff., 1388 ff.

T. P. Radhakrishnan, Z. G. Soos, H. Endres, L. J. Azevado, J. Chem. Phys. 85,
1126 (1986).

H. Volz, W. Lotsch, Tetrahedron Lett., 27, 2275, 1969. S. Bank, C. L. Ehrlich,
J. A. Zubieta, J. Org. Chem., 44, 1454, 1979. S. Bank, C. L. Ehrlich, M.
Mazur, J. A. Zubieta, J. Org. Chem, 46, 1243, 1981.

For a general overview, see "Organic Electrochemistry," 2nd Ed., M. M. Baizer,
H. Lund, Eds. Marcel Dekker, New York, 1983, Chap. 2.

R. W. Brockman and D. E. Pearson, J. Am. Chem. Soc. 74, 4128, 1952, P.
Zuman, Chem. Listy, 47, 1234, 1953. I. R. Fox, R. W. Taft, Jr., and I. C.
Lewis, J. Am. Chem. Soc. 81, 5343, 1959.

 

 

 

 

 

'lll

 

 

 

 

254

 

145

146

147

148
149

150

151

152

153

154

155

156

157

T. Berzins and P. Delahay, J. Am. Chem. Soc. 75, 5716, 1953. W. H.
Reinmuth, L. B. Rogers and L. E. I. Hummelstedt, J. Am. Chem. Soc. 81, 2947,
1959.

C. D. Ritchie, W. F. Sager Prog. Phys. Org. Chem. 2, 323, 1964. C. G. Swain,
S. H. Unger, N. R. Rosenquist, and M. S. Swain, J. Am. Chem. Soc. 105, 492,
1983, C. Hansch, A. Leo, S. Unger, K. H. Kim, D. Nikaitani, and E. Liem, J.
Med. Chem. 16, 1207, 1973, R. Popielarz and D. R. Arnold, J. Am. Chem. Soc.
112, 3068, 1990.

C. D. Ritchie, "Physical Organic Chemistry", Marcel Dekker, New York, 102,
1975. T. H. Lowry and K. S. Richardson, ”Mechanism and Theory in Organic
Chemistry", Harper & Row, New York, 149, 1987.

J. J. Brooks and G. D. Stucky J. Am. Chem. Soc. 94, 7333, 1972.
M. M. Kreevoy and D. C. Johnston, Croatica Chemica Acta 45, 511, 1973.

D. Masilamani and M. M. Rogic J. Org. Chem. 46, 4486, 1981, W. D. Watson,
J. Org. Chem. 50, 2145, 1985, U. Wriede, M. Fernandez, K. F. West, D.
Harcourt, and H. W. Moore, J. Org. Chem. 52, 4485, 1987. E. Bay, D. A. Bak,
P. E. Timony, and A. Leone-Bay, J. Org. Chem. 55, 3415, 1990. E. Iiehlmann
and R. W. Lauener Can. J. Chem. 67 , 335, 1989. R. A. Eade, F. J. McDonald
and H. P. Pham, Asust J. Chem. 31, 2699, 1978. D. C. Schlegel, C. D. Tipton,
and K. L. Rinehart, Jr. J. Org. Chem. 35, 849, 1970, T. L. Davis and V. F.
Harrington J. Am. Chem. Soc. 56, 129, 1934.

H. Jukeikis and D. Kivelson J. Am. Chem. Soc. 84, 1132, 1962.

K. Ishizu, K. Mukai, A. Shibayama, and K. Kondo, Bull. Chem. Soc. Jpn. 50,
2269, 1977.

A. Rieker and H. Kessler, Tett. Lett. 1227, 1969; H. Kessler, A. Moosmayer,
and A. Rieker, Tetrahedron, 25, 287, 1969.

H. S. Gutowsky, C. H. Holm, J. Chem. Phys. 25, 1228, 1956., D. Kost, E. H.
Carlson, and M. Raban, J. Chem. Soc. Chem. Comm. 656, 1971, J. Sandstrom,
Dynamic NMR Spectroscopy; Academic Press, New York, 97, 1982. H.
Freibolin, Basic One- and T wo- Dimensional NMR Spectroscopy, VCH, New
York, 263, 1991.

Pulsed EPR studies of Ion Binding in a Double-faced Paramagnetic IonOphore:
Tris(2,6-dimethoxyethoxy)phenyl)methyl Radical S. —H. Jang, H. —1. Lee, J.
MaCracken, and J. E. Jackson. Accepted for publication in. J. Am. Chem. Soc.

H. G. Lohr and F. Vogtle Acc. Chem. Res. 18. 65, 1985.

L. Kevan, in "Time Domain Electron Spin Resonance” L. Kevan & R. N.
Schwartz, edit. Chapter 8, Wiley-Interscience, New York, 1979.

 

 

 

 

 

 

 

255

 

158
159
160

161

162

163

164

165

166

167

168

169

170

171

J. MaCracken, D.-H. Shin, and J. L. Dye, Appl. Magn. Reson. 3, 305, 1992.
A. A. Shuben and S. A. Dikanov, J. Magn. Reson. 52, 1, 1983.
W. Weltner, "Magnetic Atoms and Molecules”, Dover, New York, 1983.

M. M. Baizer and H. Lund. ”Organic Electrochemistry", 2nd Edition, Chapter 2,
Marcel Dekker, New York, 1983. S. Bank, C. L. Ehrlich, M. Mazur, and J. A.
Zubieta J. Org. Chem, 46, 1243, 1981. G. Kothe, W. Summermann, H.
Baumgartel and H. Zimmermann, Tett. Lett. 2185, 1969.

S. Gauthier and J. M. Fréchet Synthesis , 383, 1987.

See for example F. A. Cotton, G. Wilkinson, "Advanced Inorganic Chemistry" 5th
Ed., Wiley Interscience, NY, 1988, p. 1385. J. E. Huheey, "Inorganic
Chemistry; Principles of structure and reactivity" 2nd E. Harper & Row, NY,
1978, p71.

"PCModel Molecular Modeling Software", K. E. Gilbert, J. J. Gajewski, M. M.
Midland; Serena Software, Bloomington, IN 47402, (1987).

There have been a number of reports of organic ferromagnetism. The more
substantial accounts are those of the well-characterized crystalline metallocene
derivatives cited. For other examples of purported organic ferromagnets in less
well—characterized polymeric materials see: H. Tanaka, K. Tokuyama, T. Sato, T.
Ota, Chemistry Letters, 1813, 1990 ; J. B. Torrance, S. Oostra, A. Nazzal Syn.
Met. 19, 708 ,1987 ; Y. V. Korshak, T. V. Medvedeva, A. A. Ovchinnikov, V.
N. Specktor, Nature, 326, 370 1987 ; Y. Cao, P. Wang, Z. Hu, S. K. Zhang, J.
Zhao, Syn. Met. 27, B625, 1988 ; M. Ota; S. Otani; K. Kobayashi, Chem. Lett.
1179 , 1989.

P. Grobet, L. V. Gerven, A. V. D. Bosch, and J. Vansummeren, Physica 86-868,
1132, 1977. W. Boon and L. V. Gerven, Physica B 155, 437, 1989.

We have been unable to obtain any commercially available simulation programs for
multiradical CW -EPR spectra.

A careful structure determination with a review of the history of Chichibabin's
hydrocarbon appeared recently: Montgomery, J. C. Huffman, E. A. Jurczak, M. P.
Grendze, J. Am. Chem. Soc. 108, 6004, 1986.

Although the singlet—triplet splitting in Chichibabin’s hydrocarbon has not been
unambiguously determined, it is clear that the (singlet) wavefunction of the
crystalline material studied by Montgomery has a significant biradical component.

G. A. Jeffrey and W. Saenger in "Hydrogen Bonding in Biological Structures ”
Springer-Verlag, Berlin Heidelberg New York, 1991 and references theirin.

J. Herbich, J. Karpikuk, Z. R. Grabowski, N. Tamai, and K. Yoshihara J.
Luminescence 54, 165, 1992. H. Tamiaki and K. Maruyama J. Chem. Soc.

 

 

 

 

 

 

 

256

 

 

Perkin. Trans. 2431, 1992. C. Turro, C. K. Chang, G. E. Leroi, R. I. Cukier,
and D. G. Nocera, J. Am. Chem. Soc. 114, 4013, 1992. A. Harriman, Y.
Kubo, J. L. Sessler, J. Am. Chem. Soc. 114, 388, 1992. D. N. Beratan, J. N.
Betts, and J. N. Onuchic Science, 252, 1285, 1991. Y. Aoyama, M. Asakawa,
Y. Matsui, and H. Ogoshi, J. Am. Chem. Soc. 113, 6233, 1991.

172 R. Hoffmann Acc. Chem. Res. 4, 1, 1971, P. J. Hay, J. C. Thibeault, and R.

Hoffmann, J. Am. Chem. Soc. 97, 4884, 1975. R. Gleiter, Angew. Chem. 86,
770. 1974.

 

{Erilitx

LIBRQRIES

IHHHI

 

l

I

l

STQTE U IV

I

l

.4}.

 

:x».
.1...

 

Fulani
1.: . .:

 

i333
F2529“ :3