‘ 5r!

 

lull! \- L . ‘ ‘ , . . . . '.!"’.‘Hii§ll§u'. {-5

{sh .. V . . A u . .,;..c..... I. .u u..?.,,.r .. . «1.,571, o .H....' . .v
=2,er , .. . . , My? “wt 1% pmgfifimfik} E; u

.,
. L. .v...) n I .3 I .

 

:1 2? {3%an ‘

i
‘5... a - ,l.‘u. .
'v

This is to certify that the
dissertation entitled

The Use of Exciton-Coupled Circular Dichroism for the Absolute
Stereochemical Determination of __-Chiral Carboxylic Acids and
Erythro Diols and Aminoalcohols

presented by

Courtney C. Olmsted

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Chemistry

 

 

    

Major Professor’sW“

C/i

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

.C-0-.-.-.-.-.-.-.-.-C-I-.---.-.-.-.-l-O-

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

l mm a 00

{r L‘ [1 ‘AAA
r 1T7 r 4111]“
NHL: U ‘6 J LHJH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/CIRC/DateDue.indd-pr1

 

   

THE USE OF EXCITON-COUPLED CIRCULAR DICHROISM FOR THE
ABSOLUTE STEREOCHEMICAL DETERMINATION OF oc-CHIRAL
CARBOXYLIC ACIDS AND ERYTHRO DIOLS AND AMINOALCOHOLS
By

Courtney C. Olmsted

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Chemistry

2006

ABSTRACT

THE USE OF EXCITON-COUPLED CIRCULAR DICHROISM FOR THE
ABSOLUTE STEREOCHEMICAL DETERMINATION OF oc-CHIRAL
CARBOXYLIC ACIDS AND ERYTHRO DIOLS AND AMINOALCOHOLS
By

Courtney C. Olmsted

Chirality affects all of life, from the very simplest amino acids to pharmaceuticals
used to treat diseases, but there is still much we have to learn about Chirality. We are
interested in the use of Exciton-Coupled Circular Dichroism (ECCD) to probe Chirality in
small molecules. ECCD is a non-empirical method which allows for the determination of
the absolute stereochemistry of chiral compounds using the through—space coupling
between two or more chirally oriented non-conjugated chromophores. ECCD has been
used on compounds without chromophores through interaction of a chiral molecule with
a host chromophoric molecule, causing induced Chirality in the host-guest complex which
is observable by ECCD.

Porphyrins are ideally suited for use in ECCD due to their strong extinction
coefficient, low-energy absorption band, and a central metal core for binding to guest
molecules. As a bisporphyrin tweezer with a 1,5-pentanediol linker, these molecules can
bind with many small chiral molecules and the resultant ECCD can be observed.
However, we noticed in our research that a-chiral carboxylic acids derivatized as amides

did not always show the expected ECCD Sign. Therefore we set out to explore the

rotation of the Cczo-Qll bond in our carboxylic acids derivatized as amides.
Computational modeling of the derivatized carboxylic acids and as zinc
tetraphenylporphyrin complexes have given us insight into the nature of the complexes
formed. We have also explored binding modes beyond our initial models for (1-
halocarboxylic acids, including the possibility of cooperative binding occurring between
the carbonyl oxygen, the halogen, and the metal core of the porphyrin.

To improve the reliability of our ECCD predictions, we have also developed
alternate porphyrin tweezer systems, including tweezers with a diacetylene linkage, a
ferrocene linkage, and a urea linkage. Synthesis of the porphyrin tweezer with the urea
linkage actually generated porphyrin polymers which may have use outside the field of
ECCD.

We wished also to extend the ECCD method to erythro diols and aminoalcohols,
which traditionally have posed difficulties in standard stereochemical determination
methods. We initially pursued the possibility of converting erythro diols into dioxanes
with chromophores. However, we discovered a novel method through the use of cis—
Ru(bpy)2Clz. Addition of erythro diols and aminoalcohols to cis-Ru(bpy)2Clz provides

ECCD spectra which are equal and opposite for enantiomers.

Dedicated to my best friend Mark, who also happens to be my husband

ACKNOWLEDGMENTS

First of all, I would like to express my heartfelt thanks to Babak. He has always
been a willing listener to my often outrageous ideas and has encouraged my pursuit of
answers. Due to him, I have become a better researcher and chemist. Beyond that, he
has always been a friend and a mentor, encouraging me to try new opportunities and
giving me freedom to express my opinions. I have learned much from him outside of
organic chemistry, and I will always cherish the years I have spent in his labs.

Many thanks are due to the Borhan lab members. Qifei started the circular
dichroism project in our labs and was the source of many wonderful discussions. I
especially cherish the discussions of life in China and the many differences and
similarities in US and Chinese culture. Marina has taken over the CD project and is
doing a phenomenal job of extending the scope of this project. Cnacn60 for your help
with Russian and especially the MacCoffee. Thanks also to Chrysoula, Ben, and
Montserrat who started graduate school with me. It was amazing to find people so
different who could get along so well. Additional thanks are also due to past and present
members of the Borhan lab, including Rachael, Radha, Tao, Jennifer, Jun, Stewart, and
Dan and many others who worked as high school and undergraduate researchers. The
laughter and tears I have shared with all of you are a treasured part of graduate school.
Keep up the great work and never let Babak rest.

I would also like to thank my committee members, in particular Prof. Rob
Maleczka and Prof. Ned Jackson. Ned, your insight into the computational aspects of my

research was invaluable and you were always willing to help. Rob, thank you for the

many challenges you have presented during graduate school. You are an amazing
researcher but an even more phenomenal teacher. Thanks for pushing me to become a
better chemist.

A special thanks goes to the many other friends I have made while at MSU. To
name them all would be impossible; yet, I will always remember the wonderful
discussions and time spent discussing life as well as research. You have all broadened
my horizons and my understanding and acceptance of the global world we live in.

Thanks to my family for their continual support through this long process,
especially my mom who introduced me to science at a young age and encouraged my
experimentation as long as I cleaned up the kitchen. Most importantly, thanks to Mark
for never allowing me to give up. I could not ask for a better friend to go through life

with.

vi

TABLE OF CONTENTS

List of Figures ......................................................................................... ix
List of Schemes ...................................................................................... xv
List of Tables ....................................................................................... xvii
List of Abbreviations .............................................................................. xviii

Chapter 1. Introduction
1.1. The Exciton Coupled Circular Dichroism Method (ECCD) ............................ 1
1.1.1. Chirality ....................................................................................... 1

1.1.2 Optical Rotary Dispersion (ORD) and Circular Dichroism (CD) Spectroscopy. ...4

1.1.3. Exciton Coupled Circular Dichroism (ECCD) .......................................... 12
1.1.4 ECCD Based on Quantum Mechanics .................................................... 16
1.2.1 The Use of ECCD to Determine the Absolute Stereochemistry of Chiral

Molecules .................................................................................... 19
1.2.1 Use of Chromophoric Hosts for ECCD of Non-chromophoric Chiral

Molecules .................................................................................... 19
1.2.2 Porphyrins as Chromophoric Receptors ................................................. 22
References ............................................................................................ 32

Chapter 2. Use of ECCD for Absolute Stereochemical Determination of a-Chiral

Carboxylic Acids
2.1. Use of Computational Modeling to Understand the Rotation About the CC_0- Ca
Bond .......................................................................................... 36
2.1.1. Modeling of Derivatized Carboxylic Acids Complexed to Zinc
Tetraphenylporphyrin ..................................................................... 40
2.1.2. Modeling of Derivatized Carboxylic Acids with a Second-Generation
Carrier ........................................................................................ 43
2.1.3. Modeling of Chiral Carboxylic Acids Derivatized as Amides ........................ 45

vii

2.2. Exploration of Possible Binding Modes for Derivatized Carboxylic Acids with

Zinc Tetraphenylporphyrin 1,5-Pentanediol Tweezer ................................. 53
2.3. Exploration of a Novel Binding Model Involving Coordination of an (It—Halogen
and an Amide Carbonyl Oxygen to Zinc ................................................ 59
References ............................................................................................. 70

Chapter 3. Development of Alternate Tweezers for Use with ECCD in the
Absolute Stereochemical Determination of Small Molecules

3.1. Porphyrin Tweezer with a Diacetylene Linkage ........................................ 73
3.2. Porphyrin Tweezer with a Ferrocene Linkage .......................................... 83
3.3. Porphyrin Tweezer with a Urea Linkage ................................................. 91
Experimental materials and general procedures ............................................... 103
References ........................................................................................... 121

Chapter 4. Use of Ru(bpy)2Clz for Determination of Absolute Stereochemistry of
Erythro 1,2-Diols and Aminoalcohols

4.1. Conversion of Erythro 1,2-Diols Into Dioxanes ....................................... 126
4.2. Consideration of cis-Ru(bpy)2C12 as a Chromophoric Host ......................... 133
4.3 Studies of ciS-Ru(biq)2C12 and cis-Ru(bpy)2C12 ...................................... 136
4.3.1. Complexation and Studies of cis-Ru(bpy)2C12 with Erythro 1,2-Diols and -

Aminoalcohols ............................................................................ 140
Experimental materials and general procedures ............................................... 147
References ........................................................................................... 153

viii

LIST OF FIGURES

Note: Images in this dissertation are presented in color.

Figure 1-1.

Figure 1-2.
Figure 1-3.

Figure 1-4.

Figure 1-5.

Figure 1-6.

Figure 1-7.

Figure 1-8.

Figure 1-9.

Figure 1-10.

Figure 1-11.

Figure 1-12.

Figure 1-13.

Examples of chiral molecules, (S)-2-bromobutane (left) and gingkolide B

(right) ................................................................................. 1
Examples of some chiral molecules and their enantiomers .................... 2
Structures of (R)- and (S)-Thalidomide ........................................... 3
The magnetic (blue) and electric (green) components of light, where the
direction of light propagation is to the right ..................................... 5
The circular and linear components of polarized light ......................... 5

Representation of linearly polarized light passing through a chiral
medium, where the left circular component is absorbed to a greater extent
than the right .......................................................................... 6

(a) Positive and (b) negative CD Cotton effects...................................8

Use of the “right hand rule” to determine magnetic transition dipole
moment direction .................................................................... 9

(P)-Hexahelicene and (M)-hexahelicene, showing the direction of current
and the parallel [(P)—hexahelicene] and antiparallel [(M)—hexahelicene]
transition moments ................................................................ 10

The observed ECCD spectrum of a steroidal 2,3-bisbenzoate. . . . . . .....12

Qualitative explanation of ECCD using a dibenzoate ester. (a) Orientation
of the electric transition dipole moment in dibenzoate esters; (b) Preferred
conformation of a 1,2-dibenzoate with a positive dihedral angle, to; (c) In-
phase combination of the electric transition dipole moments; (d) Out-of-
phase combination of the electric transition dipole moments ................ 13

Bisignate ECCD curve expected from the example given in Figure 1-
11 .................................................................................... 15

Possible orientations of chromophores (clockwise and counterclockwise)
and their corresponding ECCD Si gns ............................................ l6

ix

Figure 1-14.

Figure 1-15.

Figure 1-16.

Figure 1-17.

Figure 1-18.

Figure 1-19.

Figure 1-20.

Figure 1-21.

Figure 1-22.

Figure 2-1.

Figure 2-2.

Figure 2-3.

Figure 2-4.

Figure 2-5.

Splitting of the two excited states of non-conjugated chromophores i and j
by exciton interaction. The energy gap Vij is known as Davydov
splitting ............................................................................. 16

A chiral (denoted by an asterisk) guest molecule can complex with an
achiral host chromophore to produce a chiral complex capable of
generating ECCD .................................................................. 19

Complexation between racemic biphenol 1 and chiral diamine 2 leading to
a chiral complex ................................................................... 20

Examples of an achiral porphyrin and an inherently chiral porphyrin. ....23
UV-vis spectra of tetraphenylporphyrin methyl ester and zinc
tetraphenylporphyrin methyl ester showing changes in the Q band upon
metallation .......................................................................... 24

Bismetallated porphyrin tweezer 4 and shorthand representations... . ......25

Example of stereochemical determination for chiral diamines complexed
to zinc porphyrin tweezer 4 using ECCD ...................................... 26

Carrier 5 reacts with (S)-1-cyclohexylethylamine to form conjugate 6,
which has a preferred conformation where the carbonyl is anti to the
pyridine nitrogen and the small group (hydrogen) of the chiral center is

syn to the carbonyl ................................................................. 29
Complexation of conjugate 6 with tweezer 4 leading to the most favored
conformation for the complex and the observed ECCD spectrum... .......30
Bis(tetraphenylporphyrin) tweezer 1 and its simplified representation ..... 36

Complexation of a chiral acid conjugate with zinc bisporphyrin tweezer 1
leading to the most favored conformation for the complex and the
observed ECCD spectrum ........................................................ 38

Carriers A and B used to derivatize chiral molecules with one site of
attachment for use with ECCD (site of derivatization shown in blue). . ...40

Binding of zinc porphyrin to the derivatized carboxylic acid forces
rotation of the C(aryl)-N(amide) bond, causing the near coplanar amide to
become staggered with respect to the aromatic ring ........................... 40

Mnemonic developed for determining the absolute Stereochemistry of
derivatized carboxylic acids using ECCD ..................................... 43

Figure 2-6.

Figure 2-7.

Figure 2-8.

Figure 2-9.

Figure 2-10.

Figure 2-11.

Figure 2-12.

Figure 2-13.

Figure 2-14.

Figure 2-15.

Figure 2-16.

Figure 2-17.

Figure 2-18.

Figure 2-19.

Rotation around the C(CzO)-C(chiral) bond leading to opposite ECCD
predictions .......................................................................... 44

Possible conformational changes due to dipole interactions (shown in
blue) between the carbonyl and the pyridine ring upon binding to the
porphyrin tweezer ................................................................. 45

Relative energies of the lowest energy conformers generated from a

conformational search of (S)-phenylpropionic acid derivatized with carrier
A ..................................................................................... 46

The plane of the carbonyl and the plane of one group on the chiral center
used for coordinate driving ....................................................... 47

Graph of the energy values obtained for the coordinate driving of (2R)-2-
bromopropionic acid derivatized with carrier B ............................... 48

Graph of the energy values obtained from the coordinate driving of (2S)-2-
acetoxypropionic acid derivatized with carrier B ............................. 48

Graph of the energy values obtained from a molecular mechanics
geometry optimization of each molecule from the coordinate driving of
(25)-2-acetoxypropionic acid derivatized with carrier 18 (Newman
projections represent the view from a-carbon to carbonyl carbon)... ......50

Graph of the energy values obtained from a PM3 geometry optimization
of each molecule from the coordinate driving of (2S)-2-acetoxypropionic
acid derivatized with carrier B (Newman projections represent the view
from (it-carbon to carbonyl carbon) .............................................. 51

Examples of 11,2- and 1,3-carriers, from chiral amines and carboxylic
acids, respectively ................................................................. 52

Possible binding modes for carboxylic acids derivatized with carrier A...53

Pseudo-A-1,2- and A-1,3-interactions of the amide nitrogen and carbonyl
with the substituents on the chiral center ....................................... 55

Possible conformation changes due to hydrogen bonding leading to a
switch in the expected ECCD sign .............................................. 57

Possible cooperative binding between the carbonyl oxygen and the a-
halogen with the metal core of the porphyrin ................................. 60

(2R)-2-bromopropionic acid as both diastereomers of a Mosher amide....60

xi

Figure 2-20.

Figure 2-21.

Figure 2.22.

Figure 2-23.

Figure 2-24.

Figure 3-1.

Figure 3-2.

Figure 3-3.
Figure 3-4.
Figure 3-5.
Figure 3-6.

Figure 3.7.

Figure 3-8.

Figure 3-9.

Figure 3-10.

Figure 4-1.

CCD data depicting dihedral angles of a-halocarbonyls ..................... 62

Lowest energy conformation for bromoacetamide based on DFT
calculations ......................................................................... 63

DFI‘ geometry optimization of N-methyl-(2R)-2—bromoacetamide using
implicit solvation leading to a conformation with the bromine anti to the

carbonyl ............................................................................. 65

Geometry optimization of water molecules hydrogen bonded to N-phenyl-
2-bromopropionamide ............................................................ 65

Rotation around the C(C=O)-C(chiral) bond due to coordination of the
carbonyl oxygen with zinc ....................................................... 68

Representation of possible differences in binding between porphyrin
tweezers and guest molecules, leading to opposite ECCD signs. ............72

Indication of bite angle in proposed meta-substituted bisphenylacetylene
linker ................................................................................. 74

Proposed rigid porphyrin tweezer l utilizing a bisacetylene linker. . . . .....75
Red shift in UV-vis absorption of tweezer 1 with 1,5-diaminopentane....77
Alternative bisacetylene porphyrin tweezer 779
Structure of the proposed bisporphyrin ferrocene tweezer 11 ............... 84

Representation of possible binding outside of tweezer core for bis(zinc
porphyrin) ferrocene ll ........................................................... 91

Representation of Takagishi’s dimer showing steric interaction between
the urea carbonyl and the methyl substituents on the phenyl rings... .. ....92

Schematic showing control of helix direction upon addition of a chiral
guest to a porphyrin trimer connected by urea linkages, leading to
ECCD ................................................................................ 93

Spartan molecular model of helix formed from urea-linked porphyrins
(phenyl rings not shown for clarity) ............................................. 93

Newman projections of low energy conformers of (2R, 3R)-2,3-
dihydroxypentane. When derivatized with chromophores (X), the
expected lowest energy conformer would lead to a negative ECCD
bisignate curve ................................................................... 123

xii

Figure 4-2.

Figure 4-3.

Figure 4-4.

Figure 4-5.

Figure 4-6.

Figure 4-7.

Figure 4-8.

Figure 4-9.

Figure 4-10.

Figure 4-11.

Figure 4-12.

Figure 4-13.

Newman projections of low energy conformers of (ZS, 3S)-2,3-
dihydroxypentane. When derivatized with chromophores (X), the
expected lowest energy conformer would lead to a positive ECCD
bisignate curve .................................................................... 124

Newman projections of low energy conformers of (2R, 3S)-2,3-
dihydroxypentane. When derivatized with chromophores, no ECCD is
expected to be observed .......................................................... 125

Newman projections of low energy conformers of (2S, 3R)-2,3-
dihydroxypentane. When derivatized with chromophores, no ECCD is
expected to be observed .......................................................... 125

Transformation of erythro diols into six-membered rings functionalized
with chromophores (X) .......................................................... 126

Newman projections of the product formed from (2R, 3S)-2,3-pentanediol
or (ZS, 3R)-2,3-pentanediol and an (R, R) bis-chromophore ............... 127

Newman projections of the product formed from (2R, 3S)-2,3-pentanediol
and an (R, S) meso bis-chromophore (X) indicating an expected negative
ECCD bisignate curve ............................................................ 128

Newman projections of the product formed from (25, 3R)-2,3-pentanediol
and an (R, S) bis-chromophore (X) indicating an expected positive ECCD
bisignate curve .................................................................... 128

A and A forms of cis-Ru(bpy)2C12 ............................................. 133

Possible method of stereodifferentiation where the chiral erythro diol
binds preferentially to A-ciS-Ru(bpy)2Clz .................................... 134

Transfer of Chirality from the erythro diol to the ruthenium complex,
changing the initial stereochemistry of the ruthenium to match the
diol ................................................................................. 135

Spartan models of Ru(bpy)2Y where Y is (2R,3S)-pentanediol. Note the
curve of the bipyridine in the structure on the right (emphasized with a
blue curved line) ................................................................. 136

UV-Vis studies on RuC13 in acetonitrile (solid line), benzene (large

dashes), CHzClz (small dashes), MeOH (dash-dot-dash), and
methylcyclohexane (dash-dot-dot-dash) ...................................... 139

xiii

Figure 4-14.

Figure 4-15.

Figure 4-16.

Figure 4-17.

Figure 4-18.

UV-Vis studies on cis-Ru(bpy)2C12 in acetonitrile (solid line), benzene
(large dashes), CHzClz (small dashes), MeOH (dash-dot-dash), and
methylcyclohexane (dash-dot-dot—dash) ...................................... 139

ECCD spectra of cis-Ru(bpy)2Clz refluxed with ( IS) l—phenylethane- 1 2,-
diol (solid blue line) and (1R)-1 phenylethane- 1, 2- diol (dashed pink line)
inspectroscopygradeMeOH.... ..142

ECCD spectra of cis-Ru(bpy)2C12 refluxed with (-)-norephedrine (solid
blue line) and (+)—norephedrine (dashed pink line) in spectroscopy grade
MeOH ............................................................................. 143

ECCD spectra of cis—Ru(bpy)2C12 refluxed with L-allo-threonine methyl
ester (solid blue line) and D-allo-threonine methyl ester (dashed pink line)
in spectroscopy grade MeOH144

Representation of pseudo-axial and —equat0rial positions for the possible

complex forming between cis—Ru(bpy)2C12 and erythro compounds
(bipyridine rings not shown for clarity) ........................................ 145

xiv

Scheme 1-1.

Scheme 1-2.

Scheme 3-1.

Scheme 3-2.

Scheme 3-3.

Scheme 3-4.

Scheme 3-5.

Scheme 3-6.

Scheme 3-7.

Scheme 3-8.

Scheme 3-9.

Scheme 3-10.

Scheme 3-11.

Scheme 3-12.

Scheme 3-13.

Scheme 3-14.

Scheme 3-15.

LIST OF SCHEMES

Esterification of binaphthalene 3 with a chiral secondary alcohol... . ......21

Conversion of amino acids and amino alcohols for use as diamine guests

for the host porphyrin tweezer 4 ................................................. 27
Synthesis of (trimethylsilyl)acetylene derivatized porphyrin 4 ............. 75
Synthesis of bisacetylene tweezer l .............................................. 76
Proposed synthesis of alternate tweezer 8 ...................................... 80
Proposed synthesis of tweezer 8 using a functionalized aldehyde .......... 82
Attempted synthetic routes for p-(l-
trimethylsi1ylpropargyl)benzaldehyde .......................................... 83
Reported synthesis of porphyrin monosubstituted ferrocenes. . . . ....84
Synthesis of 5-(p-bromophenyl)-10,15,20-triphenylporphyrin ............. 85
Reported synthesis of bis(tributyltin)ferrocene 12 ............................ 87
Attempted coupling of 5-(p-bromophenyl)-10,15,20-triphenylporphyrin 9
with bis(tributyltin)ferrocene 12 using PdC12(PPh3)2 ......................... 87
Attempted coupling of 5-(p-bromophenyl)—10,15,20-triphenylporphyrin 9
with bis(tributyltin)ferrocene 12 using Pd2(dba)3 .............................. 88
Synthesis of bis(boronic acid)ferrocene 13 ..................................... 89

Coupling of zinc 5-(p-bromophenyl)-10,15,20-triphenylporphyrin with
bis(boronic acid)ferrocene ........................................................ 90

Synthesis of alcohol 15 and benzaldehyde 16 from 4-methyl-3,5—

dinitrobenzoic acid ................................................................. 94
Synthesis of 5-(4-methyl-3,S-dinitrophenyl)-10,15,20-triphenylporphyrin
17 ..................................................................................... 95
Attempted synthesis of 5-(4-methyl-3,5-dinitrophenyl)-10,15,20-

triphenylporphyrin using propionic acid ......................................... 95

XV

Scheme 3-16.

Scheme 3-17.

Scheme 3-18.

Scheme 3-19.

Scheme 3-20.

Scheme 3-21.

Scheme 4-1.

Scheme 4-2.

Scheme 4-3.

Scheme 4-4.

Scheme 4-5.
Scheme 4-6.
Scheme 4-7.

Scheme 4-8.

Scheme 4-9.

Scheme 4-10.

Synthesis of (4—methyl-3,5-dinitrophenyl)dipyrrole 18 and phenyldipyrrole

19 ..................................................................................... 96
Synthesis of 5-(4-methyl-3,5-dinitrophenyl)—10,15,20—triphenylporphyrin
17 from (4-methyl-3,5-dinitrophenyl)dipyrrole 18 ............................ 97
Synthesis of 5-(4-methyl-3,5-dinitrophenyl)-10,15,20-triphenylporphyrin
17 using a pseudo “one pot” procedure ......................................... 99
Reduction of the nitro groups of 17 to give 5-(3,5-diamino-4-
methylphenyl)-10,15,20-triphenylporphyrin 20 ............................... 99

Formation of porphyrin polymer 21 from 5-(3,5-diamino—4-
methylphenyl)-10,15,20-tripheny1porphyrin 20 and triphosgene. . . . ......100

Synthesis of zinc porphyrin polymer 22 from porphyrin polymer 21....102

Synthesis of meso-l,2-dibromo-1,2-(4,4’-dinitrophenyl)ethane 2 from 4-
nitrobenzyl chloride ............................................................. 129

Synthesis of the unexpected product bis-(4-nitrophenyl)acetylene 3.. ...130

Attempted derivatizations of trans-1,2-cyclohexanediol in various
solvents ........................................................................... 131

Attempted derivatization of trans-1,2—cyclohexanediol using HMPA in

ACN ................................................................................ 132
Attempted synthesis of cis-Ru(biq)2C12 ....................................... 137
Synthesis of chirally enhanced cis—Ru(bpy)2Clz .............................. 137
Synthesis of racemic cis-Ru(bpy)2Clz ......................................... 138
Attempted complexation of (S,S)-(-)-hydrobenzoin with cis-
Ru(bpy)2C12 .............................................................................................. 140
Attempted complexation of (S,S)-(-)-hydrobenzoin with cis-Ru(bpy)2C12
using n-BuLi ..................................................................... 141
Control reaction of (S,S)-(-)-hydrobenzoin with n-BuLi .................... 141

xvi

Table 1-1.

Table 2-1.

Table 2-2.

Table 2-3.

Table 2-4.

LIST OF TABLES

Definition of positive and negative exciton Chirality for a binary
system ............................................................................... 18

Predicted and observed ECCD signs for chiral acids derivatized as amides
and complexed with tweezer 1 .................................................... 39

Conformational search results of various chiral carboxylic acids
complexed to carrier A ............................................................ 42

Conformational search results of various chiral carboxylic acids
complexed to carrier B ............................................................ 42

Representative a-halocarboxylic acids derivatized with carrier B showing
unexpected ECCD signs .......................................................... 59

xvii

[a]

Ac
ACN
a.u.
biq
bpy
Boc
bu
CCD
CD
CE
p-chloranil
cm

(1

dba
DFT
DIEA
DMF

DMSO

LIST OF ABBREVIATIONS

angle of rotation

specific rotation

CD amplitude

acetate

acetonitrile

atomic units

2,2’-biquinoline
2,2’-bipyridine
tert-butoxycarbonyl

butyl

Cambridge Crystallographic Database
circular dichroism

Cotton effect

tetrachloro- l ,4-benzoquinone
centimeter

doublet

dibenzylidene acetone
density functional theory
diisopropylethyl amine
N,N-dimethylformamide

dimethyl sulfoxide

xviii

ECCD
equiv

Et

GUI

HMPA

IR

kcal

MALDI
MCH
MeOH
min

mg

MHz

MLCT

mmol

molar absorption coefficient
exciton-coupled circular dichroism
equivalents

ethyl

gram

graphical user interface

hour

Hartree-Fock
hexamethylphosphoramide
infrared

NMR coupling constant
kilocalorie

magnetic transition dipole moment
molar (moles/liter)

matrix-assisted laser desorption ionization
methylcyclohexane

methanol

minute

milligram

megahertz

milliliter

metal-1i gand charge transfer

millimole

xix

MM4
mol

MS

nm
NMP
NMR
NOE
ORD
PC
PCC
Ph
PM3
PPh3
ppm

q. v.

R

rt
Ru(bpy)2C12
s

SNl

8N2

molecular mechanics force field
mole

mass spectrometry

refractive index

nanometer

N-methyl-2-pyrrolidone

nuclear magnetic resonance

nuclear Overhauser effect

optical rotatory dispersion

personal computer

pyridinium chlorochromate

phenyl

Parametric Method Number 3
triphenylphosphine

parts per million

quad vide

rotational strength

room temperature

ruthenium (H) bis(2,2’-bipyridine) dichloride
singlet

nucleophilic substitution, first order
nucleophilic substitution, second order

triplet

XX

TBAF
TEA
TFA
THF
TLC
TMEDA
TMS
TPP
p-TSA
11L

UV-vis

tetrabutyl ammonium fluoride
triethylamine

trifluoroacetic acid
tetrahydrofuran

thin layer chromatography
tetramethylethylenediamine
trimethylsilyl
tetraphenylporphyrin
p-toluenesulfonic acid
microliter

ultraviolet-visible spectroscopy

xxi

Chapter 1
Introduction

1-1. The Exciton Coupled Circular Dichroism Method
1-1.1. Chirality

Mystical and enigmatic, chirality affects all of life, from the very simplest amino
acids to pharmaceuticals used to treat diseases, yet the origin of chirality is still a
mystery. In the past few centuries we have begun to understand the impact of chirality in
our universe. Chirality refers to non-superimposable mirror images, and was named after
the non-superimposable mirror images always with us—our hands—coming from the
Greek word cheir, meaning “hand”.1 In chemistry, chirality refers to the asymmetric
position of atoms in a molecule, such that the mirror images of that molecule are non-
superimposable. Examples of chiral molecules can range from the simplest—a single
atom with four different substituents such as 2-bromobutane—to large natural molecules

with multiple chiral centers, such as ginkgolide B (Figure l-l). The simplest example of

Me

HIM.-
Br

 

Figure 1-1. Examples of chiral molecules, (S)-2-bromobutane (left) and ginkgolide B (right).

a chiral compound is a molecule containing one chiral center; such as a tetrahedral carbon
with four different substituents. Chemists use the letters R (from the Latin “rectus”) and
S (from the Latin “sinister”) to indicate the configuration (arrangement of groups) of the

chiral center, based on priorities set in place by Cahn, Ingold, and Prelog.2

Molecules do not have to have a chiral center to be chiral. Some of the most
fascinating chiral molecules include substituted allenes,3 helicenes,4 and atropisomers,S
which all have no plane of symmetry and therefore do not have superimposable mirror
images. This type of chirality is referred to as axial chirality. Figure 1-2 shows examples

of chiral molecules and their non-superimposable mirror images (enantiomers).

Cl Cl H :CH3

 

i
t
r
IJ.
git

   

Figure 1-2. Examples of some chiral molecules and their enantiomers.

While stereochemistry is interesting from a visual standpoint, the stereochemistry
of molecules also affects their chemical and biological properties. Much work has been
done to understand chirality and how it relates to our lives, as it has become apparent that
molecules of different chirality have different properties. Many enantiomeric natural
products used as drugs have different actions within the body. One particular molecule
of historical note is thalidomide. Thalidomide was introduced as a commercially
available sedative and anti-nausea drug in Europe and Canada in the mid-1950’s. As
shown in Figure 1-3, thalidomide has a single stereogenic center and can thus exist in two
stereoisomeric forms.2 One of the enantiomers, (R)-thalidomide, has strong sedative

properties, while the other enantiomer, (S)-thalidomide, is highly teratogenic.

Unfortunately, the drug was produced and distributed in racemic form, and women who

were given thalidomide during the first few months of pregnancy had babies with a wide

variety of birth defects.
0 O O O
NH NH
N O N' " O
O O
(R)-Thalidomide (S)-Thalidomide

Figure 1-3. Structures of (R)- and (S)-thalidomide.

It is, therefore, of great importance to know the absolute stereochemistry of chiral
compounds for natural product chemistry, pharmaceutics, and biochemistry. However,
determination of the absolute stereochemistry of molecules is not simple since
enantiomers have the same physical properties such as boiling point. Separation of
isomers originally focused on crystallization of a racemic mixture and separation of the
enantiomeric crystals by hand—a very tedious and time-consuming task. However, this
did allow compounds to be labeled with a stereochemistry based on their crystal
formation. Other molecules that formed crystals of the same handedness were then
assigned the same stereochemistry. Scientists now use methods such as X-ray
crystallography, chiroptical methods such as optical rotatory dispersion (ORD) and
circular dichroism (CD),‘5 nuclear magnetic resonance (NMR) spectroscopy of Mosher
esters and related derivatives,7 and chemical correlation with known chiral compounds
for determination of absolute stereochemistry.l

While X-ray crystallography is the most unambiguous method for absolute
stereochemical determination, suitable crystals can be hard to obtain. Crystallization

often requires large amounts of compound that may not be available for recently

discovered natural products or complicated synthetic endeavors. 1f the compound of
interest is not particularly amenable to crystallization, searching for the correct
crystallization conditions can take months or even years. Thus, other methods have been
developed to help offset the difficulties associated with crystallography.

One of the most common methods for stereochemical determination is the use of
Mosher-ester type derivatives of the chiral molecule in question.7 Analysis using
Mosher’s technique involves derivatization of the compound of interest into two
diastereomers containing an aryl ring. The NMR spectrum of each is obtained and a
comparison is made between the diastereomers and the chemical shift observed for
protons within the molecule. This method requires enough sample to do two
derivatizations and to prepare NMR solutions. To circumvent some of the
abovementioned problems, a new technique has been developed which uses
chromophores to explore the chiroptical properties of molecules. This technique is
known as circular dichroism.

The goal of our research is to develop general and sensitive exciton-coupled
circular dichroic (ECCD) methods for the absolute stereochemical determination of chiral
compounds.

1-1.2. Optical Rotatory Dispersion (ORD) and Circular Dichroism (CD)
Spectroscopy

Optical Rotatory Dispersion (ORD) and Circular Dichroism (CD) are both
methods that take advantage of the different chiroptical properties of enantiomers. In
these methods, light is used to distinguish between enantiomers. Light consists of both

an electric and a magnetic field, which oscillate perpendicular to one another and to the

direction of the propagation of light (Figure 1-4). The direction of the electric field

(Direction of light propagation)

 

Figure 1-4. The magnetic (blue) and electric (green) components of light, where the direction of light
propagation is to the right.

vector defines the polarization of light.8 Light from ordinary light sources, such as the
sun or a light bulb, is unpolarized, since it consists of light waves propagating in all
directions. However, if unpolarized light passes through a polarizing filter, only the light
waves with oscillation parallel to the direction of the filter pass through. The light.
passing through the filter is now aligned in one plane of oscillation, and is defined as

linearly polarized light. Linearly polarized light is composed of two circularly polarized

 

L AK R
polaiizing filter; ‘5
~ - linearly left and right corrrponents
"“901de “gm polarized light of linearly polarized light

Figure 1-5. The circular and linear components of polarized light.

light beams: one left circular component and one right circular component. Figure 1-5
shows the composition of linearly polarized and circularly polarized light.9

The magnitude the electric field of circularly polarized light is constant, but traces
out a helix as a function of time. When linearly polarized light passes through an achiral
compound or a racemic mixture, the velocity and absorbance of the circular components

are affected in the same way, so there is no change in the polarization of the light.

However, if linearly polarized light passes through an optically active medium, the
velocity and absorbance of one circularly polarized component (either the lefi or the
right) is altered to a greater extent as compared to the opposite circularly polarized
component.'0 In Figure 1-6, the representation of linearly polarized light passing through

a chiral medium leads to greater absorption of the left circular component than the right,

x
A light propagation

 

 

SOUI'CC

 

Figure 1-6. Representation of linearly polarized light passing through a chiral medium, where the left
circular component is absorbed to a greater extent than the right.

generating right circularly polarized light. This is referred to as a Cotton effect (CE), in
honor of French physicist Aimé Cotton, who observed both optical rotatory dispersion
(ORD) and circular dichroism (CD) phenomena.

The difference in the velocity of the lefi and right components of circularly
polarized light is known as optical rotatory dispersion (ORD). A different velocity for
left and right circularly polarized light correlates to differences in the respective
refractive indices. In equation (1-1),8 nL and me are the refractive indices for left and right
circularly polarized light, respectively, when passing through an optically active medium.

An=nL—mi¢0 (1-1)
If the left and right circularly polarized light components travel in an optically active
medium at different velocities, the two components become out of phase. The resultant
addition vector of the two light components is rotated by an angle a relative to the

original plane of polarization. The direction of the addition vector is determined by the

direction of the component less affected by the optically active medium. ORD
spectroscopy is the measurement of the specific rotation, [at], of an optically active
medium as a function of wavelength.

The angle of rotation, a, is related to the difference in the refractive indices for
left and right circularly polarized light through equation 1-2:8

a = (m - nR)18001

[10 (1-2)

 

where a is the angle of rotation in degrees, nL and we are the refractive indices for left and
right circularly polarized light respectively, I is the path length in decimeters, and lo is
the wavelength in a vacuum of the light beam in centimeters.

The observed angle of rotation, at, can then be used to calculate the specific
rotation, [at], using the standard equation 1-3:8
lal=§ (1-3)
where a is the angle of rotation, with the unit in degrees, c is the concentration of the
sample in g mL", and I is the path length in decimeters.

All chiral substances exhibit a molecular refraction at almost any wavelength of
irradiation. So in theory, ORD may be detected over all wavelengths, since it is based on
the unequal refractive indices for left and right circularly polarized light. Most ofien,
optical activity is detected and quantified at a single wavelength, such as 589 nm (sodium
D-line). However, as with refractive index, ORD in and of itself does not allow for
absolute stereochemical determination, since it is not possible to know based on the

molecular structure of a compound which circularly polarized light component (right or

left) will be affected more.

Another development related to ORD is circular dichroism (CD), which uses the
absorption of the circular components as opposed to the refraction. When circularly
polarized light passes through an optically active medium, the velocity of the left and
right circularly polarized components are different, as well as the absorbance (equation 1-
4).8 Circular dichroism is a measure of the difference in the molar absorptivity, A5,
where q and SR are the molar absorption coefficients for left and right circularly
polarized light, respectively. Circular dichroism is observed when 61 and 5;; are not
equaL

A£=£L—£7t¢0 (1-4)

Since circular dichroism describes the difference in absorption of the left and right
circularly polarized light, it is to be expected that CD should occur near an absorption
band. Thus, in order to observe a CD spectrum the chiral substance being studied needs
to contain a chromophore. Since CD is related to absorbance, the shape and appearance

+ (a) + (b)

Wavelength (nm)

 

A8 A8

 

Wavelength (nm)

 

 

Figure 1-7. (a) Positive and (1)) negative CD Cotton effects.

of a CD curve closely resembles that of a UV-Vis absorption curve. However, unlike
ordinary UV-Vis absorption curves, CD curves may be positive or negative as they

reflect the difference in absorbance (Figure 1-7). CD curves plot Ag vs. wavelength.8

Circular Dichroism (CD) and Optical Rotatory Dispersion (ORD) are two
different observations of the same phenomenon. Since they are related, the molar
amplitude a of ORD can be approximated by the intensity of a CD curve, A8, for the
same molecule (Equation 1-5): H

a e 40.28Ae (1-5)
Physically, the phenomenon of CD can be explained through the interaction of the
electric and magnetic transition dipole moments. When light is absorbed by a molecule,
electrons are promoted from the ground state to an excited state, creating a momentary
dipole. This dipole is known as the electric transition dipole moment and is denoted by
the vector [1. In an achiral molecule, the overall net redistribution of electrons is planar,
however, in a chiral molecule the electrons are redistributed in a helical fashion. For
chiral molecules, this rotation of electric charge creates a magnetic field whose strength
and direction may be described by the magnetic transition dipole moment, denoted by the
vector m.1 The direction of the magnetic transition dipole moment m can be determined
by the application of the “right hand rule” to the rotation of the electric charge: If the
right hand fingers are curled in the direction of the charge, then the outstretched thumb
points the direction of m (Figure 1-8).'2

m& em

Figure 1-8. Use of the “right hand rule” to determine magnetic transition dipole moment direction.

The sign and strength of a Cotton effect (CE), is given by the theoretical
parameter R, designated the rotational strength. Rotational strength is determined from

the scalar product of the electric and magnetic transition moments (equation 1-6):l

R=pom=|y||m|cosfl (1-6)

where ,u and m are the electric and magnetic transition dipole moments, respectively, and
,6 is the angle between the two transition moments. The Sign of the CE is positive when
the angle is acute (0<,B<90°) or parallel and it is negative when the angle is obtuse
(90°<,6<180°) or antiparallel. So, if R>0, a positive CD curve is observed, and if R<0, a
negative CD curve is observed. No Cotton effect is observed when the electric and
magnetic transition dipole moments are perpendicular.

The interaction of the electric and magnetic dipole transition moments is seen
clearly with hexahelicene (Figure 1—9).13 Upon excitation, the electrons flow in a helical
manner as described above, following the path of the helix. The electric transition dipole

moment ,u is identical for each enantiomer, but the magnetic transition dipole moment

C\ /
/ 090 090 /

(P)—hexahelicene (M)-hexahelicene

Figure l-9. (P)-Hexahelicene and (M)-hexahelicene, showing the direction of current and the parallel
[(P)—hexahelicene] and antiparallel [(M)-hexahelicene] transition moments.

direction m is reversed based on the “right hand rule”. (P)-Hexahelicene has its electric
and magnetic dipole transition moments parallel and (M)-hexahelicene antiparallel. Thus
(P)-hexahelicene has a positive CE according to equation 1-6, and (M)-hexahelicene has
a negative CE, leading to positive and negative CD spectra, respectively.

Both ORD and CD spectroscopy can be recorded by spectropolarimeters known
as circular dichrometers. Circular dichrometers use a double monochromator to polarize

the light beam linearly. This linearly polarized light is then passed through a modulator

10

to generate left and right circularly polarized light. The circularly polarized light is then
passed through the sample, into a photomultiplier tube, and into an amplifier before being
recorded as a CD spectrum. Most CD spectropolarimeters measure differential
absorbance, AA, which is converted to A8 based on Lambert-Beer’s Law (equation 1-7):l

AA = A, — A, = Aacl (1-7)

Although ORD and CD are phenomena only observed in chiral compounds, they
still cannot be used for absolute stereochemical determination. Since the discovery of
ORD and CD, much work has been undertaken to determine empirical rules,14 semi-
empirical rules, such as sector or quadrant rules,15 or quantum chemical calculations‘é’ ‘7
which permit a priori calculation of the ORD or CD of a given structure. However, to
date no empirical or semi-empirical rules have been found to be valid for all molecules.
Although many of the rules developed can be useful for stereochemical analysis, care
must be taken not to apply the rules too liberally, and it must be understood that none of
these rules allow for the determination of absolute stereochemistry. Quantum chemical
calculations allow for the rotation angle of chiral molecules to be determined, and thus
the absolute stereochemistry can be assigned for a given molecule by comparing the
calculations with the experimental result. In one method, the contributions of all of the
atoms of a molecule are calculated and used to determine the specific rotation angle for a
given chiral molecule.“ ‘9 This method has been of great use in assigning the absolute
stereochemistry of large molecules with multiple chiral centers, including natural

20'” as the structure of the compound is related to the Sign and magnitude of the

products,
calculated optical rotation angle.23 However, these methods are still being developed and

it is yet to be shown if they are truly applicable for all situations.

11

1-1.3. Exciton-Coupled Circular Dichroism (ECCD)

Exciton-Coupled Circular Dichroism (ECCD)24 is a method developed out of the
principles of CD. Based on the coupled oscillator theory25 and group polarizability
theory,26 ECCD is a non-empirical method which allows for the determination of the
absolute stereochemistry of chiral compounds using the through-space exciton coupling
between two or more chirally oriented non-conjugated chromophores. As shown in
Figure 1-10, which presents the observed ECCD spectrum of a steroidal 2,3-
bisbenzoate,24 a typical ECCD spectrum is a bisignate curve. The peak (positive or
negative) at longer wavelength is called the first Cotton effect, and subsequently, the
peak (positive or negative) at shorter wavelength is called the second Cotton effect. If the

Sign of the first Cotton effect is positive, the ECCD spectrum is defined as positive, and if

positive couplert

    

 

 

/ A8 A
mosh; i
o H

 

 

syn

Figure 1—10. The observed ECCD spectrum of a steroidal 2,3—bisbenzoate.

the sign of the first Cotton effect is negative, then the Sign of the ECCD spectrum is
designated negative. The amplitude (A) of the ECCD spectrum is defined as A = A8] -
A52 where A61 and A52 are the amplitudes of the first and second Cotton effect,

respectively. The observed sign of the ECCD spectrum is dictated by the relative

orientation of the chromophores in the chiral molecule. As seen in Figure 1-10, a

12

clockwise orientation of the two chromophores (looking from front to back) results in a
positive ECCD curve, and a counterclockwise orientation will show a negative ECCD
curve.24

ECCD results from the interactions of two non-conjugated chromophores, as
mentioned above. When the chromophores are identical, the most significant interaction
occurs at the transitions which occur at the same energy. The through-space coupling of
the electric transition dipole moments of these two chromophores generates one of two
helices, depending upon whether they are in-phase (symmetric) or out-of-phase (anti—
symmetric), so a CD spectrum with two peaks is generated. This can be explained by
examining the results of a 1,2-dibenzoate ester (Figure l-l 1).12

Benzoate esters have a strong UV absorption band at about 230 nm resulting from

(a) O (b) H
0/ . ‘O,,O .771

go

 

 

 

 

 

 

HH
(1)
(C) 69
MD ,
\ ~‘ 9
(9/6) ‘ \ *‘9_ m
(d) 9 A
trauma V 69
fl 8 ” m

Figure 1-1 1. Qualitative explanation of ECCD using a dibenzoate ester. (a) Orientation of the electric
transition dipole moment in dibenzoate esters; (b) Preferred conformation of a 1,2-dibenzoate with a
positive dihedral angle, or; (c) ln-phase combination of the electric transition dipole moments; ((1) Out-
of-phase combination of the electric transition dipole moments.

13

the rt—m“ transition involving the benzene ring and the carbonyl of the ester group.27
The electric transition dipole moment p of each benzoate group is approximately oriented
along the long axis of the benzoate group and nearly parallel to the direction of the C-0
ester bond (Figure 1-1 1a). The conformations of 1,2-dibenzoates are well known and
orient either clockwise or counterclockwise depending on the chirality of the substrate.
Figure 1-1 1b shows a 1,2-dibenzoate with the two benzoate groups oriented clockwise,
leading to a dihedral angle a) which is positive and which correlates roughly to the angle
between the two electric transition dipole moments of the two benzoates. The individual
electric transition dipole moments ,u couple both in-phase (symmetric) and out-of-phase
(anti-symmetric) with the other chromophore’s electric transition dipole moment (Figure
1-1 1c and d, respectively). When the two chromophores couple in-phase (Figure 1-1 1c)
the overall electric transition dipole moment is oriented along the C2 axis of the two
chromophores, and when the two chromophores couple out-of-phase (Figure 1-1 1d) the
overall electric transition dipole moment is oriented perpendicular to the C2 axis. The
helix associated with the charge rotation generated from the transition can be determined
by visualizing each chromophores’ electric dipole transition moment placed on the
outside of a cylinder with the cylinder aligned along the axis of the overall electric
transition dipole moment. Using the right hand rule, m can be determined by following
the helix described around the cylinder by the individual electric transition dipole
moments. For Figure 1-11c, the electric transition dipole moment ,u and the resultant

magnetic transition dipole moment m are oriented antiparallel to each other, and

according to equation 1-6, R = ,u o m = ly||m| cos ,6 , this leads to a negative Cotton effect

and thus a negative CD band. For Figure 1-1 1d, the electric transition dipole moment

14

,u and the resultant magnetic transition dipole moment m are oriented parallel to each
other, leading to a positive Cotton effect and a positive CD band.12 Thus, the two CD
bands generated from the through-space coupling of the electric transition dipole
moments of the dibenzoate have equal magnitude but opposite Sign. The observed
bisignate ECCD curve thus results from the overlapping of the positive and negative CD
bands occurring very close in energy.

Which CD band, positive or negative, occurs at lower energy is determined by the
phasing of the coupling. Repulsion between poles of the same Sign (Figure 1-11c)
creates in—phase coupling that is higher in energy than the out-of—phase coupling (Figure
1-1 1d), where the poles of opposite sign are closest together. Consequently, the CD band
associated with the out-of-phase mode (in this example, positive) is at longer
wavelengths than the CD band resulting from the in—phase mode (in this example,
negative). Since the positive CD band is at the longer wavelength than the negative CD
band, for the absolute configuration shown in Figure l-llb, the bisignate couplet is
designated positive (Figure 1-12).

The above demonstration can be performed for chromophores in a

positive CD band

 

As

 

negative CD band

Figure 1-12. Bisignate EC CD curve expected from the example given in Figure 1-1 1.

15

counterclockwise conformation as well. Thus, the Sign of an ECCD spectrum can be
correlated to the orientation of the interacting chromophores. For chromophores oriented
clockwise, the observed ECCD bisignate will always be positive, and for chromophores
oriented counterclockwise, the observed ECCD bisignate will always be negative (Figure
1-13).28 Thus when the orientation of the chromophores is known the ECCD bisignate

curve can be predicted and vice versa.

9
F

Je

Figure 1-13. Possible orientations of chromophores (clockwise and counterclockwise) and their
corresponding ECCD signs.

1-1.4. ECCD Based on Quantum Mechanics
ECCD can also be explained by quantum mechanical theories. When a molecule
contains two identical non-conjugated chromophores, because of the through-space

interactions, excitation is delocalized between the two chromophores and the excited state

 

 

——f—~:2 2V1; j:: , Excited statea

 

 

Energy

 

 

 

 

 

Ground state 0

 

Chromophore i Delocalized Chromophore j
excnafion

 

 

 

Figure l-l4. Splitting of the two excited states of non-conjugated chromophores l' and j by exciton
interaction. The energy gap V.)- is known as Davydov splitting.

l6

(exciton)29 is split into two states, a and B, by resonance interaction of the local
excitations (Figure 1-14). The energy gap between the two excited states is designated
Zij, and is known as Davydov splitting.24

For a binary system, the rotational strength, R, which is a theoretical parameter
representing the Sign and strength of a CD Cotton effect, can be approximated as

equation 1-8:24
Ra, ,6 = i(%)7z'OORij . (Ilioa X [000) “-8)
where the positive and negative signs correspond to the a- and B- state, respectively, Rij

is the distance vector from chromophore i to j, [1'00 and fljoa are the electric transition
dipole moments for excitation from the ground state to excited state a of chromophores i
and j, respectively, and 00 is the excitation number of transition from the ground state to
excited state a, which can be expressed as: on = v0 / c = (E, — E0)/(hc), where h is
Planck’s constant and c is the speed of light in vacuum. As is evident from equation 1-8,
the Cotton effect of the a- and ,6- state have identical rotational strength of opposite Sign,
leading to a bisignate CD spectrum with two Cotton effect peaks of opposite signs, which
are separated by the energy gap (Davydov splitting) (Figure 1-12).

The Sign and amplitude of the bisignate ECCD spectrum is theoretically defined

in equation 1-9 as:24
Rij . (ll—jive X ll-jjoa)I/ij (1-9)

where Rij is the distance vector from chromophore i to j, fiioa and fljoa are the electric
transition dipole moments for excitation from the ground state to excited state a of

chromophores i and j, respectively, and Vij is the interaction energy between the two

17

chromophores i and j. If equation 1-9 is positive, the observed ECCD spectrum is
defined as positive, and if it is negative, then the observed ECCD spectrum is defined as
negative. Thus it is clear that the Sign of the bisignate curve depends entirely on the

Spatial orientation of the two chromophores (Table 1-1).28’ 3°

 

Qualitative
Definition

69
D - —. — . . .
Rij 0 (Illioa X Moa)I/ij > 0 posmve first and negative

Quantitative Definition Cotton effects

second Cotton effects

O
F - ~ — . . .
R ij 0 (,Ltioa X ,uj'oa)Vij < 0 negative first and posrtlve

second Cotton effects

 

 

 

Table 1-1. Definition of positive and negative exciton chirality for a binary system.

The amplitude (A) of the ECCD couplet, defined as the difference in A8 between
the first and second Cotton effects of the couplet, is affected by the number of interacting
chromophores, the molar absorptivity coefficient of the chromophores, the distance
between the chromophores, and the projection angle between the chromophores.31 The
number of interacting chromophores (x, y, z) affects A in an additive effect, such that A
for the whole molecule is determined by the sum of A of each interacting pair, i.e.,
Amolecule = Axy + Ayz + An. This has proven to be true through experimental and theoretical
testing of systems with multiple chromophores.32'34 Since A is proportional to the square
of the molar absorptivity coefficient (a) of the chromophores, chromophores with strong
absorption lead to greater sensitivity for ECCD. A is also inversely proportional to R2,
where R is the distance between chromophores. Consequently it is beneficial to have the

chromophores oriented close to each other in space. A is at a maximum when the

18

projection angles between chromophores is close to 70°, as a result it is beneficial to use
systems which allow for the orientation of the chromophores close to this angle.
1-2. Use of ECCD to Determine the Absolute Stereochemistry of Chiral Molecules

Since ECCD is a based on quantum mechanical principles as shown in the
previous section, the absolute stereochemistry of chiral molecules which display ECCD
can be determined non-empirically. Due to the sensitivity of the method, ECCD can be
utilized with samples as small as micrograms, making it ideal for stereochemical
determination of rare natural products and newly synthesized compounds. ECCD has
been used for the absolute stereochemical determination of many types of compounds,
including polyols,35 carbohydrates,36 quinuclidines,37 hydroxy acids,38 and others.39
l-2.l. Use of Chromophoric Hosts for ECCD of Non-chromophoric Chiral
Molecules

ECCD is an effective method for assigning the absolute stereochemistry of chiral
compounds in a non-empirical fashion, yet molecules must contain two interacting, non-
conjugated chromophores in order to display ECCD. This limits the use of ECCD to only
a very small subset of chiral compounds. To circumvent this problem and make ECCD
amenable to more chiral molecules, chromophoric molecules have been designed to act
as hosts for chiral molecules without chromophores. The chiral molecule is introduced as

a guest to the host chromophoric molecule. Through covalent or non-covalent binding,

 

Figure 1-15. A chiral (denoted by an asterisk) guest molecule can complex with an achiral host
chromophore to produce a chiral complex capable of generating ECCD.

l9

the host is designed to adopt a chiral conformation based on the chirality of the guest
molecule (Figure 1-15). The chiral conformation of the host chromophoric molecule is
then observed as an ECCD bisignate curve.

Chiral induction is not a new phenomenon, and in fact was used with polypeptides
bound to dyes to probe changes in the helical conformations.40 Induction of chirality
through intermolecular forces between two molecules has also been used for probing the
conformations of macromolecules.“ Chiral induction requires two processes to occur
between the molecules: (1) complex formation and (2) some dynamic process or
processes associated with the formation of the complex, such as conformational changes

4244 If the steric or electronic interactions between two

of the interacting molecules.
molecules are strong enough, chiral information will be transferred from one to the other.
In order to simplify the observed ECCD, most chromophoric hosts used for
ECCD studies are achiral. However, upon binding to a chiral substrate, the chromophoric
host will adopt an induced chirality dictated by the chirality of the substrate. This
induced chirality of the complex can then be monitored by ECCD and used to determine

the absolute stereochemistry of the chiral guest molecule. As mentioned previously,

binding between the host and guest molecules can be either covalent or non-covalent,

Br HO OH Br 'fl ,N\ ,N.’
M ’ HH ‘
R H . i H
O O + Cr. = .. xiv .
"* O O
R = -CH2CH2t-BU
OZN N02

1 2

Figure 1-16. Complexation between racemic biphenol 1 and chiral diamine 2 leading to a chiral
complex.

20

depending on the nature of the complex being formed. One example of non-covalent
complex formation can be seen in Figure 1-16.45 In this example, the racemic biphenol 1
hydrogen bonds to the chiral diamine 2,4648 forming a new chiral complex. In the
absence of chiral induction, biphenol 1 has two chiral conformations which rapidly
interconvert. Formation of the complex causes rotation about the C-C biphenyl bond,
locking it into a single conformation. Although it would appear that the phenyl rings are
conjugated and would not show ECCD, steric interactions between the substituents on the
phenyl rings cause the rings to be nonplanar, and therefore non-conjugated and able to
undergo exciton coupling. Thus the ECCD spectrum that is observed reflects the
absolute stereochemistry of chiral diarrrine 2.

Covalent bonding has been used to form chiral complexes of binaphthalenes to
determine the absolute stereochemistry of chiral secondary alcohols (Scheme 1-1).49 The
chromophoric reagent, 3-cyanocarbonyl-3’-methoxycarbonyl-2,2’-binaphthalene 3 is
used to esterify the chiral secondary alcohol and the resultant complex exhibits induced
ECCD as a result of favoring one atropisomer due to restricted rotation about the CC
phenyl bonds. The chirality of the secondary alcohol favors one atropisomer due to steric
interactions between the methyl ester and the new chiral ester leading to induction of

chirality to the binaphthalene system.49

COOMe COOMe
00 .... 0c ......
R*OH
DO DMAP, ACN 00

Scheme 1-1. Esterification of binaphthalene 3 with a chiral secondary alcohol.

 

3

21

1-2.2. Porphyrins as Chromophoric Receptors

Some of the most widely used chromophoric receptors are porphyrins or
porphyrin derivatives, due to several properties which make them highly suitable for
chiral recognition.50 Porphyrins have an ideally situated absorption band near 418 nm for
their main absorption band (known as the Soret band). This absorption band can vary
depending on the substituents on the porphyrin. Since this absorption band is much
lower in energy than most other chromophores, the chance of unwanted interactions
between the porphyrin and any chromophores in the guest moleclrle is unlikely. Not only
do they absorb at low energy, porphyrins are also strong chromophores with a molar
absorptivity coefficient (a) of 400,000 or higher. Since the amplitude of the observed
ECCD bisignate is proportional to £2, a larger molar absorptivity coefficient for the host
chromophore leads to enhanced sensitivity in ECCD, allowing for the use of smaller
sample amounts. This enhanced sensitivity also makes the porphyrins able to detect
subtle changes in their environment, including the binding of chiral ligands, which can be
followed by UV-Vis, fluorescence, NMR, and resonance Raman spectroscopy.50
Porphyrins have a central core consisting of four nitrogens which may incorporate a
variety of metals with different properties and sizes. This metal core can then also be
used for binding chiral guest molecules. Porphyrins with no metal or small metal cores
are generally planar structures with a well-defined binding pocket in the center, which
can be enhanced by substituents on the outside of the porphyrin ring. There are many
sites that can be derivatized such as the mesa and ,B-positions, the central metal and the
inner nitrogen atoms. Adjusting the substituents on the porphyrin can also control

solubility for use in both polar and non-polar solvents.

22

As with other molecules, porphyrins can be both achiral and chiral due to the
steric effects of outer substituents on the conformation of the porphyrin ring (Figure 1-
l7).51 Achiral porphyrins tend to be planar molecules, while the chiral porphyrins are
usually “saddle” shaped where two sides point up and two sides point down. Achiral

porphyrins have found wide use in stereochemical studies with circular dichroism for

52-54

absolute configurational assignments, stereochemical differentiations of sugars and

55,56 57,58

amino acids, and interactions with bio-macromolecules.

 

Figure l-17. Examples of an achiral porphyrin and an inherently chiral porphyrin.

Porphyrins are used extensively in UV-Vis and CD spectroscopy due to their red-
shifted absorption band and their strong molar extinction coefficient (a). The most

intense band around 415 nm in the near ultraviolet is the B band, usually called the Soret

23

band. Additional bands (N, L, M bands) in the ultraviolet region may be observed but
these bands are usually much weaker. Substituents on the outside ring of the porphyrin
usually have a small effect on the energy of the absorbance of the Soret band, depending
on the type and position of the substituents. If the molar absorption coefficient is known
for a given porphyrin, the concentration of the solution can be calculated using the
observed absorbance and Beer-Lambert’s law, A = ecl.

Porphyrin free base generally has four absorption bands (Q bands) in the visible

region due to 7t-rt* transitions. Upon metallation of the porphyrin core, the four Q bands

 
    
  

0'4 Soret band

418nm

 

 

 

 

 

 

 

Wavelength (nm)
0.4 Soret band
419 um i ’ ‘ ,\ expanded‘o’betid'
0'2 450 500 550 i
As J
0 0 A
350 525 700
Wavelength (nm)

 

002MB

Figure 1-18. UV-vis spectra of tetraphenylporphyrin methyl ester and zinc tetraphenylporphyrin methyl
ester showing changes in the Q band upon metallation.

24

in the visible region collapse into essentially two bands due to higher symmetry, while
the Soret band may remain in the usual range or shift to higher or lower energy (Figure 1-
18). The molar absorption coefficient (a) of the Soret band also changes with the
insertion of metals into the porphyrin core. For example, zinc porphyrin methyl ester
(i5'=550,000)59 has a higher molar absorption coefficient than the corresponding porphyrin
free base ($440,000)60 and the position of this band is also slightly red shified (Figure l-
18).

Although useful for UV-Vis and CD studies, single porphyrins are not useful as

host chromophores for ECCD due to the lack of a second chromophore that will absorb at

 

B]—
B‘

 

 

Figure l-l9. Bismetallated porphyrin tweezer 4 and shorthand representations.

25

a similar energy. However, recently bismetalloporphyrin systems have been developed
specifically for use with ECCD (Figure 1-19). These compounds are achiral and present
an ideal environment for chiral guest molecules to bind, creating a chiral complex that
can then be monitored by ECCD and used to determine the absolute stereochemistry of
the guest moleculesém’6

Zinc porphyrin tweezer 4 (Figure 1-19) was originally designed for use in the
absolute stereochemical determination of chiral diamines. This host molecule can
undergo host-guest complexation through the zinc centers of the porphyrins. The

porphyrin tweezer then adopts a chirally oriented conformation leading to a bisignate

ECCD spectrum revealing the helical nature of the chromophores. This helicity can then

be related back to the structure of the guest chiral diamine. As shown in Figure 1-20,

| I
I

 

Zinc W‘Phy‘l“ Zn <— NH2 H2N ->Zn

M6020, H __ %\ tweezer
ANA/we HZN NH2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 00
426 nm (benzene)
50 _ (+ 90)
0 ~—— —A A v' V-
l:[> A8
-50 -
-100 " 434 nm
(-1 l7)
-1 50 1 J
350 400 450 500

negative ECCD positive ECCD

wavelength (nm)

Figure 1-20. Example of stereochemical determination for chiral diamines complexed to zinc porphyrin
tweezer 4 using ECCD.

26

when zinc porphyrin tweezer 4 is added to a solution of chiral diamine (S)-methyl-2,4-
diaminobutyrate, the chiral diamine binds to the zinc centers of the porphyrin tweezer
through Zn-N coordination to form a complex where the chiral guest is held in between
the two porphyrin chromophores.61 The porphyrin which is farther away from the chiral
center does not interact with the chiral center and merely coordinates to the primary
amine. However, the porphyrin coordinated to the amine closest to the chiral center
interacts with the groups on the chiral center. Due to steric interactions, the porphyrin
closest to the chiral center prefers to slide away from the large group and cover the small
group (Figure 1-20a). This conformation is more stable than the alternate (Figure 1-20b),
in which the porphyrin has strong steric interactions with the large group. The more
stable conformation (0), leads to the two porphyrin planes adopting a counterclockwise
orientation and a negative ECCD couplet is predicted, which matches the observed
ECCD (Figure 1-20, center).

The diamine approach has been extended to determine the absolute
stereochemistry of chiral amino acids and amino alcohols, which have a primary amino
group directly attached to the stereogenic center.“ Although not able to act as guest

molecules since they only contain one amine group, simple chemical modifications

1. BocHN/V ”Hz 0

 

 

H
COOH A H NH
La/ 7 N/\/ 2
NHBOC 2. Deprotection L NH2H
CO H O
H 1. BocHN/V 2
A H
L )'/\0H 7 VOMNHZ
NHBOC 2. Deprotection L NHz

Scheme 1—2. Conversion of amino acids and amino alcohols for use as diamine guests for the host
porphyrin tweezer 4.

27

convert these compounds into diamines (Scheme 1-2), making them suitable for use with
the method described above.

As was evident with the above examples, molecules with only one site of
attachment are not suitable for use with zinc porphyrin tweezer 4 since the porphyrins
need to be aligned such that they can interact through space for exciton coupling to occur.
To overcome this problem, molecules have been designed to convert compounds with
only one site of attachment to molecules possessing two sites of attachment. These
molecules are known as carriers, and are achiral molecules which can be used to create
derivatives of the chiral substrates which have two sites of attachment. In order to be
effective, carriers need to have a firnctional group that can react with the chiral
compounds to form a derivative, i.e., a carboxylic acid can be coupled with a chiral
amine or alcohol. The carrier molecules should also be rigid, since flexible structures
will lead to a molecule with multiple possible conformations which could create
difficulties in the interpretation of the ECCD data. The carriers should be achiral to avoid
the formation of diastereomers which will also interfere with interpretation of the ECCD
data. Finally, the derivatized molecules should be able to bind to zinc porphyrin tweezer
4 or another desired host in a 1:1 ratio.62

Ideally, each guest molecule will bind with the host porphyrin tweezer in only one
conformation. However, this is most often not the case, and the lowest energy
conformation for the complex must be determined in order to accurately determine the
stereochemistry of the guest molecule. The conformation of the complex is dependent on
the conformation of the guest molecule, thus an ideal carrier generates a rigid and

predictable conformation with any derivative chiral molecules. Too many degrees of

28

[1% O H ¢ HN MCH?’
N COZH H2N »
CH3

 

\\ H
I \ N\ MACH3 N\ O
/ > I / H 2) l
/
NHBoc NH2 NH
5 6 2

Figure 1-21. Carrier 5 reacts with (S)-l-cyclohexylethylamine to form conjugate 6, which has a
preferred conformation where the carbonyl is anti to the pyridine nitrogen and the small group
(hydrogen) of the chiral center is syn to the carbonyl.

freedom in the derivatized compounds will lead to multiple ECCD signals due to varying
interactions of the porphyrins with the groups on the chiral center. Because of this
necessity for rigidity, carrier 5 was designed and synthesized in order to determine the
absolute stereochemistry of chiral monoamines complexed with zinc porphyrin tweezer 4
(Figure 1-21).62 This carrier can be derivatized with chiral monoamines through coupling
of the carboxylic acid functionality to form an amide bond. After removal of the Boc
protecting group, the derivatized monoamine has two nitrogens available for binding to
zinc porphyrin tweezer 4. Based on the knowledge of amide bond rotation, the
conformation of the derivatized monoamine can be predicted with relative confidence.
This conjugate can bind to the zinc porphyrin tweezer to form a complex, the
stereochemistry of which can be assigned based on the observed ECCD spectrum.
Derivatization of (S)-1-cyclohexylethylamine with carrier 5, followed by deprotection,
leads to conjugate 6 with two nitrogen atoms, the pyridine nitrogen and the free amino
nitrogen, able to bind to both zinc centers of porphyrin tweezer 4. Molecular modeling of
conjugate 6 suggests that it is a rigid molecule which has a predictable low energy

conformation.62 In the most favorable conformation, the free rotations around the single

29

bonds CAryl-Cc=o, Cc=o-NH, and NH-Cchiml are restricted. Molecular modeling has also

shown that conjugate 6 favors a coplanar orientation of the carbonyl group to the pyridine
ring, yet due to electrostatic repulsion, the conformation with the carbonyl oxygen anti to
the pyridine nitrogen is preferred by ~30 kJ/mol. It is also well known that in most
amides, the small group prefers to be syn to the carbonyl oxygen due to pseudo A-1,3
interactions between the carbonyl oxygen and the groups on the carbon attached to the
nitrogen.

Since the conformation of conjugate 6 is known with relative certainty, the

orientation of the porphyrins in the host-guest complex can be similarly determined. The

(M)
(L) H3C Wt“;

1N CH3 XHN 2_’Zn/
zn ‘_:
NH2 6
O 0WD 0 O 0W0 O
“ [1
200 432
.135. f; a»

A8 0 <1:
w
’ (-108)

-200 . . . i
410 420 430 440

wavelength (nm)

 

 

 

 

 

 

 

 

 

 

 

Figure 1-22. Complexation of conjugate 6 with tweezer 4 leading to the most favored conformation for
the complex and the observed ECCD spectrum.

30

three substituent groups at the chiral center are assigned S (small), M (medium), and L
(large) according to their size,1 with hydrogen being the S group, methyl being the M
group and cyclohexyl being the L group (Figure 1-22). Upon binding to zinc porphyrin
tweezer 4, the zinc porphyrin’s approach to the pyridine nitrogen will favor approaching
the side of the M group in order to avoid unfavorable steric interactions with L. Thus, the
two porphyrin planes of the tweezer orient in a clockwise way, and dictate a positive
ECCD spectrum. This prediction matches the experimental observation (Figure 1-22).
The previous work that has been done has introduced the concept of using
porphyrin tweezers for absolute stereochemical determination with ECCD. However, the
compounds that this method has been extended to are not diverse. Ideally, we would like
to develop a system that will allow for the absolute stereochemical determination of any
molecule using ECCD. Our work presented in this dissertation has focused on extending
the previous work done with porphyrin tweezer 4 to include other porphyrin tweezers as
well as determining the conformations of the complexes formed between porphyrin
tweezers and their guest molecules. We have focused on using molecular modeling to
probe the dynamics of these complexes and explore the effects of different carriers for
molecules with one site of attachment. We have also explored the synthesis and ECCD
of novel porphyrin tweezers with various linkers. Finally, we have probed the possibility
of using metal complexes, specifically Ru(bpy)2Cl2, for complexation with erythro diols

to facilitate the use of ECCD with these molecules.

31

References:

1. Eliel, E. L.; Wilen, S. W.; Mander, L. N., Stereochemistry of Organic Compounds.
Wiley: New York, 1994.

2. Bruice, P. Y., Organic Chemistry. 3rd ed.; Prentice—Hall, Inc.: Upper Saddle River,
New Jersey, 2001.

3. Rossi, R.; Diversi, L. Synthesis 1973, 25.
4. Laarhoven, W. H.; Prinsen, W. J. C. Top. Curr. Chem. 1984, 125, 63.

5. Eliel, E. L., Stereochemistry of Carbon Compounds. McGraw-Hill: New York, 1962.

6. Humpf, H.-U.; Berova, N.; Nakanishi, K.; Jarstfer, M. B.; Poulter, C. D. J. Org.
Chem. 1995, 60, 3539.

7. Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512-519.

8. Lambert, J. B.; Shurvell, J. F.; Lightner, D. A.; Cooks, R. G., Organic Structural
Spectroscopy. Prentice Hall: Upper Saddle River, 1998.

9. Solomon, T. W. G., Organic Chemistry. John Wiley & Sons. Inc.: New York, 1978.

10. Velluz, L.; LeGrand, M.; Grosjean, M., Optical Circular Dichroism. Academic Press
Inc.: New York, 1965. '

ll. Mislow, K.; Glass, M. A. W.; O'Brien, R. E.; Rutkin, P.; Steinberg, D. H.; Weiss, J .;
Djerassi, C. J. Am. Chem. Soc. 1962, 84, 1455.

12. Snatzke, G. Angew. Chem. Int. Ed. Engl. 1979, 18, 363.

13. Nogrady, M., Stereochemistry. Pergamon Press Ltd.: Oxford, 1981.

14. Moffitt, W.; Woodward, R. B.; Moscowitz, A.; Klyne, W.; Djerassi, C. J. Am. Chem.
Soc. 1961, 83, 4013.

15. Snatzke, G., Optical Activity and Chiral Discrimination. Reidel, Dordrecht, and
Neatherlands: 1979.

16. Polavarapu, P. L. Molec. Phys. 1997, 91, 551.
17. Polavarapu, P. L.; Chakraborty, D. K. J. Am. Chem. Soc. 1998, 120, 6160.
18. Kondru, R. K.; Wipf, P.; Beratan, D. N. Science 1998, 282, 2247.

19. Kondru, R. K.; Lim, S.; Wipf, P.; Beratan, D. N. Chirality 1997, 1997, 469.

32

20. Kondru, R. K.; Chen, C. H.; Curran, D. P.; Beratan, D. N.; Wipf, P. Tetrahedron:
Asymmetry 1999, 10, 4143.

21. Kondru, R. K.; Beratan, D. N.; Friestad, G. K.; Smith, A. B.; Wipf, P. Org. Lett.
2000, 2, 1509.

22. Ribe, S.; Kondru, R. K.; Beratan, D. N.; Wipf, P. J. Am. Chem. Soc. 2000, 122, 4608.
23. Kondru, R. K.; Wipf, P.; Beratan, D. N. J. Phys. Chem. A 1999, 103, 6603.

24. Harada, N.; Nakanishi, K., Circular Dichroic Spectroscopy: Exciton Coupling in
Organic Stereochemistry. University Science Books: Mill Valley, 1983.

25. Kuhn, W. Transactions of the Faraday Society 1930, 26, 293.

26. Kirkwood, J. G. J. Chem. Soc. 1937, 5, 479.

27. Wiesler, W. T.; Vazquez, J. T.; Nakanishi, K. J. Am. Chem. Soc. 1987, 109, 5586.
28. Harada, N.; Chen, S.-M. L.; Nakanishi, K. J. Am. Chem. Soc. 1975, 97, 5345.

29. Chen, S.-M. L.; Harada, N.; Nakanishi, K. J. Am. Chem. Soc. 1974, 96, 7352.

30. Harada, N.; Nakanishi, K. Acc. Chem. Res. 1972, 5, 257.

31. Weckerle, B.; Schreiber, P.; Humpf, H.-U. J. Org. Chem. 2001, 68, 8160.

32. Dong, J .-G.; Akritopoulou-Aanae, 1.; Guo, J .; Berova, N.; Nakanishi, K.; Harada, N.
Enantiomer 1997, 2, 397.

33. Liu, H. W.; Nakanishi, K. J. Am. Chem. Soc. 1981, 103, 7005.

34. Lichtenthaler, F. W.; Sakakibara, T. Carbohydr. Res. 1977, 59, 47.

35. Rele, D.; Zhao, N.; Nakanishi, K.; Berova, N. Tetrahedron 1996, 52, 2795.
36. Wiesler, W. T.; Nakanishi, K. J. Am. Chem. Soc. 1989, 111, 9205.

37. Zhao, N.; Berova, N.; Nakanishi, K.; Rohmer, M.; Mougenot, P.; Jurgnes, U. J.
Tetrahedron 1996, 52, 2777.

38. Gimple, O.; Schreiber, P.; Humpf, H.-U. Tetrahedron: Asymmetry 1997, 8, 1 l.

39. Nakanishi, K.; Berova, N., Circular Dichroism. Principles and Applications. VCH
Publishers: New York City, 1994.

40. Blout, E. R.; Stryer, L. Proc. Natl. Acad. USA 1959, 45, 1591.

33

41. Stryer, L.; Blout, E. R. J. Am. Chem. Soc. 1961, 83, 1411.
42. Beer, P. D.; Rothin, A. S. J. Chem. Soc, Chem. Commun 1988, 52.

43. Hayashi, T.; Asai, T.; HoKazono, H.; Ogoshi, H. J. Am. Chem. Soc. 1993, 115,
12210.

44. Mizutani, T.; Yagi, S.; Honmaru, A.; Ogoshi, H. J. Am. Chem. Soc. 1996, 118, 5318.

45. Mizutani, T.; Takagi, H.; Hara, 0.; Horiguchi, T.; Ogoshi, H. Tetrahedron Lett.
1997, 38, 1991.

46. Hanessian, S.; Gomtsyan, A.; Simard, M.; Roelens, S. J. Am. Chem. Soc. 1995, 117,
7630.

47. Kramer, R.; Zundel, G. J. Chem. Soc, Faraday Trans. 1990, 86, 301.

48. Bell, C. L.; Barrow, G. M. J. Chem. Phys. 1959, 36, 5.

49. Hosoi, S.; Kamiya, M.; Ohta, T. Org. Lett. 2001, 3, 3659.

50. Ogoshi, H.; Mizutani, T. Acc. Chem. Res. 1998, 31, 81.

51. Mizutani, T.; Ema, T. J. Am. Chem. Soc. 1994, 116, 4240.

52. Matile, S.; Berova, N.; Nakanishi, K. Enantiomer 1996, I, l.

53. Matile, S.; Berova, N.; Nakanishi, K. J. Am. Chem. Soc. 1995, 117, 7021.
54. Arimori, S.; Takeuchi, M.; Ahinkai, S. J. Am. Chem. Soc. 1996, 118, 245.

55. Crossley, M. J.; Hambley, T. W.; Mackay, L. G.; Try, A. C.; Walton, R. J. Chem.
Soc, Chem. Commun. 1995, 1077.

56. Crossley, M. J.; Mackay, L. G.; Try, A. C. J, Chem. Soc., Chem. Commun. 1995,
1925.

57. Suenaga, H.; Arimori, S.; Shinkai, S. J, Chem. Soc, Perkin Trans. 2 1996, 607.

58. Foster, N.; Singhal, A. K.; Smith, M. W.; Marcos, N. G.; Schra, K. J. Biochim.
Biophys. Acta. 1988, 950, 118.

59. Matile, S.; Berova, N.; Nakanishi, K.; Fleischhauer, J.; Woody, R. W. J. Am. Chem.
Soc. 1996, 118, 5198.

34

60. Rickman, B. H.; Matile, S.; Nakanishi, K.; Berova, N. Tetrahedron 1998, 54, 5041.

61. Huang, X. F.; Rickman, B. H.; Borhan, B.; Berova, N.; Nakanishi, K. J. Am. Chem.
Soc. 1998, 120, 6185.

62. Huang, X. F.; Borhan, B.; Rickman, B. H.; Nakanishi, K.; Berova, N. Chem-Eur. J.
2000, 6, 216.

63. Huang, X. F.; Fujioka, N.; Pescitelli, G.; Koehn, F. E.; Williamson, R. T.; Nakanishi,
K.; Berova, N. J. Am. Chem. Soc. 2002, 124, 10320.

64. Huang, X. F .; Nakanishi, K.; Berova, N. Chirality 2000, 12, 237.

65. Kurtan, T.; Nesnas, N.; Li, Y. Q.; Huang, X. F.; Nakanishi, K.; Berova, N. J. Am.
Chem. Soc. 2001, 123, 5962.

66. Proni, G.; Pescitelli, G.; Huang, X. F .; Quraishi, N. Q.; Nakanishi, K.; Berova, N.
Chem. Commun. 2002, 1590.

35

Chapter 2
Use of ECCD for Absolute Stereochemical Determination of

a-Chiral Carboxylic Acids

2-1. Use of Computational Modeling to Understand the Rotation About the Cc=o-C..
Bond
The idea of being able to “see” a molecule and know what its structure looks like

is not a new one—various kinds of modeling have been used for centuries. Beginning

 

Q.—
B‘

01'

 

 

I— —1

Figure 2-1. Bis(tetraphenylporphyrin) tweezer 1 and its simplified representation.

36

with simple stick-and-ball type models, we have progressed to the point of being able to
visualize molecules in silico through the use of supercomputers and even our own PCS.
Now, with the advent of faster, even more powerful computers, it has become practical to
model molecules in such a way as to be able to predict stable conformations both in
solution and gas phase. Small molecules can be modeled with good accuracy at a small
cost (referring to computing time and resources), but macromolecules pose special
problems in that there are usually a large number of degrees of freedom, leading to many
possible low-energy conformers. Although a molecule may exist in several forms in
solution, it is often useful to know if one conformer is preferred, particularly for chiral
compounds.

Initial work in the field of absolute stereochemical determination of derivatized
carboxylic acids using porphyrin tweezer 1 (Figure 2-1) and ECCD provided a mnemonic
that allowed for the prediction of an ECCD Sign for a given chiral monoamine derivatized
as an amide (Figure 2-2). However, in expanding this method to chiral molecules with
other functional groups, we determined that the usual method of predicting the Sign of the
CD couplet does not always work for chiral carboxylic acids derivatized as amides.l It
was noticed that several chiral carboxylic acids produced the opposite ECCD Sign as
expected (Table 2-1). It was not immediately obvious why some chiral acids produced
the opposite ECCD Sign as expected. To probe the possible conformational issues that
might be affecting the observed ECCD Sign, we decided to use molecular modeling, as
crystals of derivatized porphyrins can be difficult to grow and obtaining a crystal

structure of a host-guest complex poses even more problems.

37

 

N
l ‘ N H’N
/ H CH3
\
Z" Z” <—: .. —>zn
/ NH2
NH
W W
O O O O O O O O

 

 

200
432 (+135) F (L)

 

 

424 (-108)

 

 

-200 l l l l
410 42? 4?0 440
wave ength nm)

Figure 2-2. Complexation of a chiral acid conjugate with zinc bisporphyrin tweezer leading to the most
favored conformation for the complex and the observed ECCD spectrum.

 

 

 

 

To perform molecular modeling on our systems, we chose to initially use
Macromodel, a suite of visualization and computational tools for building, manipulating,
and calculating various properties of molecules.2 Monte Carlo, a conformational search
program applicable to macromolecules with multiple degrees of freedom, is contained
within Macromodel. Monte Carlo methods work through molecular mechanics, a low-
level minimization technique.3 Although higher level calculations such as semi-empirical

and ab initio generate more accurate results, these types of calculations are prohibitively

38

It)

 

1
\
I j]: :) ECCD
/
N N R -
H chlrd
Chiral acid Predicted Observed
H CO H
2 (S) X 2 negative not observed
Ph OCH3
H CO H
3 (S) X 2 negative negative
Ph OAc
H H
4 (S) X 002 negative negative
H30 OAc
H 002H
5 s X '
( ) H30 OPh negative positive
5 (R) HXCOZH positive negative
PhO CH3
H CO H
HR) X 2 sitive ne ative
H3C Ph po 9
H H
8 (3) 3,002 negative negative
H30Hzc CH3
H CO H
9 (R) - 2 negative negative

H 30 "CHZPh

Table 2-1. Predicted and observed ECCD signs for chiral acids derivatized as amides and complexed
with tweezer 1.

expensive (in terms of computer time) for medium to large molecules (i.e., over ~30
atoms). Monte Carlo conformational searches can obtain several low-energy
conformations of a chiral molecule with or without chromophores, which we expected to
help predict and understand our CD spectra. From these structures, the position of the
chromophores and their electric transition dipole moments can be determined and then
used to predict or explain ECCD spectra. This works well provided that the lowest

energy conformations are CD active. If the CD-active conformations of our substrates

39

are not the lowest energy conformations, we would be limited in our use of molecular

modeling.
2-l.l Modeling of Derivatized Carboxylic Acids Complexed to Zinc
Tetraphenylporphyrin

Our initial research was focused on the use of a carrier molecule with two sites of

attachment for the porphyrin tweezer. This derivatization was necessary to use the

NH

NH2

2

\
/

N NH2
A

Figure 2-3. Carriers A and B used to derivatize chiral molecules with one site of attachment for use
with ECCD (site of derivatization shown in blue).

ECCD method for stereochemical determination on molecules with only one site of
attachment. Our first carrier molecule, carrier A (Figure 2-3), was used to derivatize
chiral carboxylic acids as amides. Table 2-1 shows the carboxylic acids used for these

studies. Initially, we developed a simplistic model of each molecule’s conformation

 

 

8A + Zn-TPP

Figure 2-4. Binding of zinc porphyrin to the derivatized carboxylic acid forces rotation of the C(aryl)-
N(amide) bond, causing the near coplanar amide to become staggered with respect to the aromatic ring.

40

based on a Monte Carlo conformational search of the derivatized carboxylic acid. Our
expected ECCD signs were dictated by placing the medium group syn to the amide
hydrogen, which in turn would lead to a preferred helicity of the bound porphyrin
tweezer as a result of avoiding the steric interactions with the large group. This led to a
number of predictions of ECCD which were not upheld by our experimental data. We
believed this to be due to the lack of consideration of possible conformation changes in
the derivatized carboxylic acid upon binding to the zinc porphyrin tweezer.
Conformational analyses of various derivatized carboxylic acids coordinated to one zinc
porphyrin suggested a flaw in the design of our carrier A. It became apparent that the
coplanar amide-phenyl arrangement was not maintained due to steric crowding (Figure 2-
4). Rotation of the Cal-Namide bond upon binding of the derivatized carboxylic acid to
the porphyrin tweezer led to many unpredictable conformers of conflicting ECCD sign.
As can be seen from Table 2-2, Monte Carlo conformational searches of various
carboxylic acids derivatized with carrier A yielded a large number of conformers within a
small energy range. In many cases, the number of conformers with a predicted ECCD
Sign had an equal number of conformers with the opposite predicted sign. After
evaluating the conformers, it became apparent that carrier A had too many freely rotating
bonds, leading to a large number of low energy conformers. At this realization, we
considered a more rigid carrier, carrier B (Figure 2-3). As opposed to carrier A,
conformational searches of carrier B not only produced fewer low energy conformers, but
all conformers for a given acid predicted the same ECCD Sign (Table 2-3). From this
information and the observed ECCD data, we were able to create a new mnemonic using

carrier B (Figure 2-5). However, from the data obtained from the conformational

41

No. of E range Conformers No. of conformers No. of conformers
conformers (kcal/moi) w/in 1 kcal/mol predicting (+) ECCD predicting (—) ECCD

 

2A-TPP 37 8.2 5 5 0
3A-TPP 73 9.2 7 4 3
4A-TPP 34 7.4 6 3 3
6A-TPP 63 9.2 9 6 3
7A-TPP 59 9.5 14 9 5
8A-TPP 24 9.5 6 l 5
9A-TPP 71 9.9 14 7 7

Table 2-2. Conformational search results of various chiral carboxylic acids complexed to carrier A.

No. of E range Conformers No. of conformers No. of conformers
conformers (kcal/moi) w/in l kcal/mol predicting (+) ECCD predicting (-) ECCD

 

ZB-TPP 64 9.9 7 0 7
3B-TPP 97 1 0 4 O 4
4B-TPP 6 l 9. 8 4 0 4
6B-TPP 69 9. 9 5 5 0
7B-TPP 2 l 9. 8 2 O 2
8B-TPP 14 7.9 l 0 1
9B-TPP 22 9. 8 3 0 3

Table 2-3. Conformational search results of various chiral carboxylic acids complexed to carrier B.

searches with carboxylic acids derivatized with carrier B, we observed that upon binding
to the zinc porphyrin, an unexpected conformation of the amide moiety was
predominating in which the large group is perpendicular to the amide carbonyl. We were
interested in probing this conformation in order to better predict ECCD signs for novel

compounds.

42

 

 

 

 

 

/ , :1“: . . J i I) e ECCD
531 ‘“ ‘ 1 _, g L M
i /1"..-v-..«-- fun-‘1‘.” \ S
. I? \ :>® ECCD
‘1ELI7tvr' “4 L
l "v\ /5) 0
‘~ g ' H H
I ‘ ‘ \\ All " Q79
/ . > Ph =
Ph OCH
ArHN OCHs 3

 

 

 

 

 

‘ =/UV\¥ .- 1
i l l I 0
375 400 425 450 475
Wavelength (nm)

Figure 2-5. Mnemonic developed for determining the absolute stereochemistry of derivatized
carboxylic acids using ECCD.

2-1.2. Modeling of Derivatized Carboxylic Acids with a Second-Generation Carrier

From initial modeling and geometry optimization of porphyrin tweezer 1
complexed with various derivatized carboxylic acids, the structures generated would be
anticipated to Show both the expected (from previous mnemonics) and unexpected ECCD
couplet signs. As an initial hypothesis, we thought that perhaps sterics were playing a
significant role in the interaction of the guest molecules with the tweezer. In considering
the conformation of amides, hydrogens alpha to the carbonyl are often depicted as

parallel to the carbonyl; in fact, semi-empirical calculations in our lab on carboxylic acids

43

derivatized as amides indicate that this is the most favored conformation. However,
when bound to a porphyrin tweezer, it may not be favorable to have the medium and
large groups both pointing towards the porphyrin face. Instead, it is possible for rotation
to place the medium group nearly parallel to the carbonyl with the small and large groups
facing the porphyrins which reverses the Sign of the ECCD couplet (Figure 2-6).
Geometry optimization and energy minimization using Spartan v.5. l .3 indicate that these
amide conformers (with the medium group syn to the carbonyl) are close in energy to the
previously expected conformer and may be preferred. Another possible hypothesis takes
into consideration the change in the dipoles of this molecule as the pyridine nitrogen
binds to the zinc. Normal dipole interactions would place the dipole of the carbonyl
pointing opposite to that of the pyridine ring. However, once the nitrogen is bound, the
dipole of the pyridine ring reverses, which may affect how the carbonyl is positioned in
relationship to the pyridine ring (Figure 2-7). Semi-empirical studies were undertaken to

explore this hypothesis. It appears that, regardless of basis set used, geometry

HZN
\
I N/ r“tug/M
negative ECCD positive ECCD

Figure 2-6. Rotation around the C(C=O)-C(chiral) bond leading to opposite ECCD predictions.

optimization which allows for a structure with the carbonyl dipole opposite to the

44

pyridine dipole generates structures lower in energy than structures with the carbonyl and

pyridine dipoles aligned.

HzNj H2N H2N
N . H G) N . H (+) L

H L’M I H L’M 1 HM
H

Figure 2-7. Possible conformational changes due to dipole interactions (shown in blue) between the
carbonyl and the pyridine ring upon binding to the porphyrin tweezer.

2-1.3. Modeling of Chiral Carboxylic Acids Derivatized as Amides

Modeling the host-guest tweezer complexes provided us with many clues as to
possible conformations that would lead to ECCD. However, it became of great interest to
determine the actual conformations of the guest molecules alone. Information about the
conformation of our derivatized carboxylic acids could lead to the development of better
carrier molecules for derivatization, or even other tweezers that would take better
advantage of conformational changes within our guest molecules. We chose to evaluate
the conformation of the amide guests before they bind to the tweezer, and how these
conformations are influenced upon binding with a single porphyrin in greater depth than
we had previously explored with molecular mechanics.

An initial conformational search was undertaken to explore possible low-energy
conformers for the derivatized carboxylic acids. Using Macromodel, a search was
performed on (S)-phenylpropionic acid derivatized with 2-amino-4-
(methylamino)pyridine. Of the 38 conformers within 8 kcal/mol of the global minimum,

17 occur with the methyl group syn to the carbonyl, 13 occur with the hydrogen syn to the

45

carbonyl, and three occur with the phenyl group syn to the carbonyl. The remaining
conformers consist of conformations in which the carbonyl bisects the angle between two
groups on the chiral center. Two conformers, one with the hydrogen syn to the carbonyl
and one with the methyl group syn to the carbonyl (i.e., the two smallest substituents by
A-values) were exported to be used in ab initio studies in Spartan. Calculated energies
using the Hartree-Fock (HF) method with the 3-21G* basis set gave energies of -
757.4280119 a.u. and -757.4313368 a.u., respectively, which translates to a difference in
energy of 2.09 kcal/mol in favor of the conformer with the methyl group syn to the
carbonyl (Figure 2-8). This was somewhat surprising since methyl is a larger group than
hydrogen, but as we considered the molecules it became apparent that there were possible
interactions between the hydrogen of the amide nitrogen and the substituents on the chiral
center. Realizing that we needed to model these interactions more extensively, we turned
to coordinate driving. However, we continued to use conformational searches in
Macromodel to predict initial low energy conformers at which a minimum might be

occurring.

H N H N H
2 Conformational 2 2N
search

I \ 3 I \ I \
N/ NKCHa N/ NkH N/ NkCHS
H ” H ”
Ph H [CHaph H Ph H

Relative energy 2.09 kcal/mol 0 kcanol
H3

H?H.§gh T»). H

Figure 2-8. Relative energies of the lowest energy conformers generated from a conformational
search of (S)-phenylpropionic acid derivatized with carrier A.

 

 

 

 

 

 

 

 

 

 

46

Initially, we believed that the production of the opposite CD spectrum than
predicted could be due to rotation about the C-C bond between the carbonyl and the
chiral center. This could occur due to steric interactions between the large and medium
groups of the chiral center and the porphyrin binding to the pyridine nitrogen. Rotation
could occur such that the small group is pointing toward the porphyrin face instead of syn
to the carbonyl, as normally portrayed. This would slide the large and medium groups
towards the carbonyl such that the carbonyl would more or less split the angle between
the two groups (Figure 2-6). This would occur if the steric interactions of the large (and

potentially medium) group were greater with the porphyrin face than the barrier of

«p

«.../<7

ll

I

'5
| I

Figure 2-9. The plane of the carbonyl and the plane of one group on the chiral center used for
coordinate driving.

 

 

 

 

rotation past the carbonyl. To explore this possibility, coordinate driving studies were
started using Spartan v. 5.0. Coordinate driving “drives” the molecule in question under
the constraints that are set by the user. Molecules can be driven based on angles, reaction
coordinates, etc. After each “drive,” the molecule is reminimized with the driven
coordinates constrained and the energy of the molecule is calculated at the level of theory
chosen by the user. In this case, we were interested in driving the dihedral angle between
the plane created by the carbonyl and the plane created by the chiral carbon and one of

the groups on the chiral center (Figure 2-9).

47

-10 H
-10.5 4

-11.5 4

Energy (kcal/mol:
R3

 

'14 r f f r 1
O 20 4O 6O 80 100 120
Molecule

.4

Figure 2-10. Graph of the energy values obtained from the coordinate driving of (2R)-2-
bromopropionic acid derivatized with carrier B.

-87

Energy (kcal/mol:

 

 

 

-97 4e - T - , —- ‘ we .......
o 30 60 90 120

Molecule

Figure 2-11. Graph of the energy values obtained from the coordinate driving of (257)-2-
acetoxypropionic acid derivatized with carrier B.

Initially, all calculations were performed using the plane created by the chiral
carbon and the small group as the second plane. Calculations were done at the semi-
empirical level with the PM3 basis set. This allowed for a relatively high level of
calculation with a minimal use of computer resources and time. Initial calculations were

done over 360° with 180 steps, creating drive angles of 2°. This produced extremely

48

large data files, as an individual file with complete energy information is created and
saved for each drive step. After discovering this, the drive angle was changed to 3° (120
steps) to decrease the amount of computational time and resources needed. The results
were plotted as a graph of energy (kcal/mol) versus molecule, where molecule is the
degrees in units of 2° each from the starting dihedral angle between the two planes.
Some of the graphs display expected energy diagrams with three minima that correlate to
the various groups on the chiral carbon being syn to the carbonyl (e.g., the graph from the
coordinate driving of (2R)-2-bromopropionic acid, Figure 2-10). However, as the groups
on the chiral carbon change, the energy diagrams become more complex and less easily
explained (e.g., the graph from the coordinate driving of (ZS)-2-acetoxypropionic acid,
Figure 2-11). These graphs often show only one minimum, with unusual peaks and
plateaus in the energy plot which cannot be explained through conventional steric
interactions.

To try to determine if these results are accurate and significant, a second study
was initiated in which each molecule from the coordinate driving was subjected to a
molecular mechanics geometry optimization. The idea is that normal geometry
optimization will optimize a molecule to the nearest local minimum. In this case,
separate geometry optimization of each step of the coordinate driving should tell us how
many minima we have and the conformation of each of those minima, since each
molecule should minimize to a local minimum. As can be seen from Figure 2-12, this
was performed on the derivatized (25')-2-acetoxypropionic acid. Indeed, five distinct
minima are observed through coordinate driving of the Cc:o-CCl bond. The minima

correspond to the carbonyl syn to the hydrogen on the chiral carbon, the carbonyl

49

splitting the methyl and hydrogen, the carbonyl syn to the methyl, the carbonyl splitting
the methyl and acetate, and the carbonyl splitting the hydrogen and the acetate, in order
from left to right along the x-axis (the last apparent minimum is due to the carbonyl
splitting the methyl and hydrogen with the hydrogen on the nitrogen pointing in the
opposite direction as the other minimum of the same designation due to the nitrogen not

being completely planar).

“We . “2“
(a) (b) 0 ,CH3 H CH3
Nirn 2A00®CH3 N _ H => ®

"I OAC
H A00 CH3 H OAc
HzN O H2N 0
H3 H3C OAc
u ”’1 H OM fl CH3 H
H OAC H
H2N
O
... O 4 We“
4 _ u l OAc CH3
CH3

Energy (kcal/mol'
O

 

 

 

 

O 20 4O 60 80 100 120

Molecule

Figure 2-12. Graph of the energy values obtained from a molecular mechanics geometry optimization
of each molecule from the coordinate driving of (2.5')-2-acetoxypropionic acid derivatized with carrier B
(Newman projections represent the view from a-carbon to carbonyl carbon).

50

To explore these minima and determine if these are actually the lowest energy
minima, each minimum conformation from the coordinate driving calculation (at a
molecular mechanics level) was subjected to a semi-empirical geometry optimization at
the PM3 level of calculation (Figure 2-13). Interestingly enough, the results of this
calculation do not entirely match those from the mechanics calculation. There are five
distinct minima evident here, corresponding to the carbonyl splitting the hydrogen and

HZN 0 A00 H HZN H
(”O/“WSW ®H3 (”OK/H: Q9
Ll! OAc:> N

I ACO CH3

H2N
CH H30 OAc
‘9 ON%(CH3 => 3 (d) WEI—”O (\OAc :, 65
H

" I H OAC
H OAc

 

 

_92 _ AcO CH3
1
g C
~93 4
20‘ l
g -94 >1
- l
g -95 J d
> J
g -96 4' b
I.I.I Ia
~97 4 - e
l
-93 .1 ____ __- _ _.__,__._____..____r_ _______r____ *_,_ _ ”.1
0 20 40 60 80 100 120
Molecule

Figure 2—13. Graph of the energy values obtained from a PM3 geometry optimization of each
molecule from the coordinate driving of (28)-2-acetoxypropionic acid derivatized with carrier B
(Newman projections represent the view from a-carbon to carbonyl carbon).

51

the acetate, the carbonyl syn to the hydrogen, the carbonyl syn to the methyl, the carbonyl
splitting the acetate and the carbonyl, and the carbonyl syn to the hydrogen, from left to
right across the x-axis. Differences in the energy of the two minima where the carbonyl
is syn to the hydrogen are due to changes in the position of the hydrogen on the nitrogen
(as described previously), which slightly changes the steric interaction of that hydrogen
with other groups in the molecule.

It is known based on the calculations described above that the difference in
energy between minima is much higher for 1,2-carriers (i.e., carboxylic acids derivatized
as amides) than for 1,3-carriers (i.e., amines derivatized as amides) (Figure 2-14), with
the carrier numbers referring to the distance between the aromatic ring and the carbonyl.
This would indicate that it does appear feasible for conformation changes to be occurring
for the 1,3-carriers which have not been observed in the 1,2-carriers, in which the lowest
energy conformer is not actually the best conformation when bound to the tweezer. This
would be expected to occur where the energy difference between conformations is low
enough to allow for rotation about the C(;=o-CCl bond, when the energy required for this
conformational change is offset. A rotation to relieve steric strain between the groups on
the chiral carbon and the porphyrin ring as the guest molecule approaches and binds to
the host tweezer would possibly release more energy than is required to move past the
rotational energy barriers. As a result, there are more conformational possibilities

available to the derivatized carboxylic acids than we had previously thought. These

1,2-carrier 1,3-carrier
Figure 2-14. Examples of 1,2- and 1,3-carriers, from chiral amines and carboxylic acids, respectively.

52

additional conformations need to be evaluated to determine which conformation is most
likely for an individual molecule, which conformations occur when complexed to a
porphyrin, and which conformations would be expected to lead to an ECCD spectrum
when complexed.

2-2. Exploration of Possible Binding Modes for Derivatized Carboxylic Acids with

Zinc Tetraphenylporphyrin 1,5-Pentanediol Tweezer

(b)

 

Figure 2-15. Possible binding modes for carboxylic acids derivatized with carrier A.

In trying to determine why we cannot predict with any certainty the CD results of
carboxylic acids derivatized with 2-amino-4-(methylamino)pyridine, we have explored a
number of conformational options but no one conformational explanation correlates with
all compounds. With this in mind, we thought to explore differences in binding options
between the guest and host molecules. There are three binding options for the derivatized

guest to the porphyrin tweezer (Figure 2-15): (a) The carrier is binding at the pyridine

53

nitrogen and the primary amine. (b) The carrier is binding at the pyridine nitrogen and
the carbonyl of the amide. (c) The carrier is binding at the primary amine and the
carbonyl of the amide (the pyridine ring is flipping so that the pyridine nitrogen is syn to
the carbonyl). Option (b) we see as least likely, as we would expect the free primary
amine to bind more readily to the zinc than the carbonyl oxygen. Allowing for various
binding possibilities, we need to determine what is happening conforrnationally at the
chiral (Jr-carbon of the amide. According to studies done by Wiberg and Rush,4’5 amides
in the Z-conformation (i.e., where the carbonyl oxygen and the amide hydrogen are anti)
prefer the substituents on the a-carbon to be staggered. Amides in the E—conformation
(the carbonyl oxygen and the amide hydrogen are syn) prefer the substituents on the a-
carbon to be eclipsed with the carbonyl. Their studies, however, were only carried out
with a methyl group as the a-carbon, so it is unclear how groups other than hydrogen
would behave in terms of orientation with respect to the carbonyl.°'9 A recent study by

1.,10 calculated structure, rotational barriers, vibrational spectra, and dipole

Allinger, et a
moments for ketones using a molecular mechanics (MM4) force field (for reference see
also the study of aldehydes by the same authors”). Their findings indicate that for all
ketones studied, the lowest energy conformer is not the one with an tit-hydrogen eclipsed
with the carbonyl. Instead, a significantly lower conformer is found where either a
methyl group eclipses the carbonyl (for methyl ketones) or the carbonyl bisects the
methyl groups (in isopropyl ketones). If this is true for ketones, it may be even more so

in amides, where the substituents on the a-carbon could be in psuedo-A-l,3 interactions

with the amide hydrogen (for the Z amide). It could be possible that the pseudo-A-1,2

54

 

interactions between the carbonyl and the (it-substituents are less important sterically than

the A-l,3 interactions between the amide hydrogen and the a-substituents (Figure 2-16).

: r3 Pseudo-A-1,2

N .,,
H I
Pseudo-A- l ,3 V

Figure 2-16. Pseudo-A-1,2- and A-l,3-interactions of the amide nitrogen and carbonyl with the
substituents on the chiral center.

To explore some of these issues and determine low energy conformations for our
system, we have done several different types of calculations. To begin with, single-point
energy calculations using density functional theory were done on all of the molecules that
we have explored experimentally. Density functional theory (DFT) allows for electron
correlation—taking into account the theory that electrons will react to and move away
from each other in a molecule—without expensive (in terms of computer resources)
calculations.‘2 Ab initio and semi-empirical methods do not use electron correlation
except in an averaged sense, and this approximation can introduce error. Using the
B3LYP force field with different basis sets gave mixed results. The molecules surveyed
included derivatives of (R/S)-2-bromopropionic acid, (R/S)-2-phenylpropionic acid, and
(R/S)-2-phenoxypropionic acid. Calculations were done on each molecule in three
different conformations: one with the small group syn to the carbonyl, one with the
medium group syn to the carbonyl, and one with the large group syn to the carbonyl (size
of the groups was based on reported A values”). Each conformation was subjected to
geometry optimization with the 3-21G* basis set, and then the resulting conformer was
optimized again with the 6-31G* basis set. To represent possible binding through the

pyridine nitrogen, the N-methyl version of each molecule was also optimized with the 3-

55

21G* basis set to observe possible changes in conformation due to the reversed dipole of
the pyridine group. Each molecule gave distinctly different results, although in all cases
the energy difference between the different conformations was no more than 6 kcal/mol.
However, it still was not possible to correlate the calculations with experimental ECCD
results. Although some of the molecules optimized to a single conformation, some did
not, and even the ones that optimized to a single conformation were not consistent in the
orientation of the a-carbon substituents relative to the carbonyl. Although potentially
useful in evaluating and comparing conformers, none of the information from this study
alone allowed for an explanation of the ECCD data. Thus, more computational studies
were undertaken.

Previous coordinate driving studies on Spartan indicated that some of the
molecules might not have well defined energy maxima and minima, perhaps suggesting
more free rotation around the Cc=o-C.1 bond than expected. If there is more free rotation
than expected around this bond, the conformation leading to ECCD results may not be
the lowest energy conformer, but may instead be locked into a specific conformation
when the carrier binds to the porphyrin tweezer, making it difficult to predict which
conformation for each molecule is actually leading to the ECCD results observed.
Depending on how the guest molecule is binding to the porphyrin tweezer, there could be
multiple conformations that would lead to the observed ECCD. It came to our attention
that perhaps calculations should be done on the derivatized chiral carboxylic acids taking
into consideration other binding possibilities.

All of the previous calculations focused on the theory that the carboxylic acids

derivatized with carrier A were binding through the primary amine and the pyridine

56

nitrogen (Figure 2-15a). However, it is a possibility that binding is occurring through the
carbonyl present in the amide functionality. It has been suggested that the amide group
does not actually exist as the traditional resonance forms normally used to explain amide

5 In computational studies on the amide group, Wiberg, et al., have found a

behavior.“
significant difference in how the electron density is distributed, where the majority of the
electron density is located on the carbonyl oxygen as expected in one of the normal
amide resonance forms. They propose, however, that this may be more characteristic of
the ground state than previously believed. If this is true and a nearly negative charge is
present on the oxygen, it would provide some basis for why we might see binding
occurring through the amide carbonyl instead of the pyridine nitrogen (Figure 2-15c).l4
Another possibility is hydrogen bonding of one of the chiral substituents to the

amide N-H (Figure 2-17). Although this is a possibility, it does not account for all of the

H2N HzN

negative EOCD positive ECCD

Figure 2-17. Possible conformational changes due to hydrogen bonding leading to a switch
in the expected ECCD sign.

experimental data (i.e., conformations expected by hydrogen bonding do not always
predict the correct ECCD sign). It appears that multiple effects of the substituents on the
chiral carbon may be influencing the outcome of the ECCD, and those effects may

include sterics and electronic effects, as well as hydrogen bonding. We continued our

57

modeling studies using Maestro, an improved graphical user interface (GUI) for
Macromodel. We performed a number of conformational searches using chiral
carboxylic acids derivatized with carrier B. To minimize the amount of data to only what
is useful we analyzed each conformation. We only recorded data for the lowest energy
conformer and then any others which would be expected to show ECCD when placed in
the host system. This allowed us to evaluate which conformers would show ECCD and
how they compare energetically to the lowest energy conformer (which may or may not
be expected to show ECCD). In all molecules studied, the lowest energy conformer
would not be expected to show ECCD based on our binding models. None of the
conformers that would be expected to show CD have the medium group anti to the
carbonyl, which was one of the proposed binding models. We also performed two
studies in Spartan to look for the feasibility of hydrogen bonding between the hydrogen
of the amide group and one of the groups on the chiral carbon. Using density functional,
which takes electron effects into account, we were able to obtain conformations in which
the phenoxy group has a weak hydrogen bond to the amide hydrogen. This was done
through constraining the distance of the phenoxy oxygen to the amide hydrogen.
However, we still would not expect this conformation to show much if any ECCD,
because there is a group syn to the N-H where we expect the porphyrin face to approach.
Another possibility that was proposed is H-rt stacking between the amide hydrogen and
the phenyl ring of the phenyl or benzyl groups. This is very difficult to model with our
software and this type of hydrogen binding does not seem to be supported within the
methods we are using. NMR studies by another member of our group did not lead to

conclusive evidence for H-7t stacking.15

58

2-3. Exploration of a Novel Binding Model Involving Coordination of an a-Halogen
and an Amide Carbonyl Oxygen to Zinc

As we have progressed in our understanding of the conformations of a-chiral
carboxylic acids derivatized as amides, we have continued to pursue the use of ECCD to
determine the absolute stereochemistry of a-chiral carboxylic acids with a variety of
functional groups on the alpha carbon. We have been able to develop mnemonics which
work for many compounds. However, carboxylic acids with halogens on the chiral
center have completely thwarted our efforts to explain the observed ECCD. Based on our
initial mnemonic (Figure 2-5), we would expect the exact opposite ECCD for every (1-
halocarboxylic acid we have tested (Table 2-4). A-values do not predict the observed
ECCD. We were most intrigued by the possibility of cooperative binding to zinc by the
carbonyl and halogen of a-haloamides when they are acting as bound guests in our zinc
porphyrin tweezer system. This would place the halogen syn to the carbonyl (Figure 2-

18). It is very difficult to model this type of cooperative bonding since most molecular

 

Mnemonic Expected ECCD Actual
r
R positive negative
H 3C H
r
S negative positive
H
Br
S GY negative positive
H

r
S H Ph negative positive

Table 2-4. Representative a-halocarboxylic acids derivatized with carrier B showing unexpected
ECCD signs.

59

modeling programs do not account for non-covalent, non-hydrogen bond interactions. It

is possible to model these systems by placing restraints on the bond lengths, etc., between

    

HN
2 O 0| i
H

H

Figure 2-18. Possible cooperative binding between the carbonyl oxygen and the a-halogen with the
metal core of the porphyrin.

the ligating atoms and the zinc, but this imparts false information into the calculations
making them less than reliable. Due to the difficulties in modeling these systems, we
looked into spectroscopic methods aided by molecular modeling for determining whether
or not zinc is cooperatively binding to the halogen of the haloamides.

Another member of our group derivatized some of the chiral carboxylic acids to
compounds similar to Mosher’s amides, where the derivatizing amine contained a chiral

center with an aromatic group that would be expected to produce different NMR

0 H O H
H Agra-13 H Agrph
Hadfifi Ph +15%” CH3
Br Br

Figure 2-19. (2R)-2-bromopropionic acid as both diastereomers of a Mosher amide.

shielding patterns on the substituents of the carbonyl a-carbon (Figure 2-19).15 This
method would allow us to determine the average conformation for free a-haloamides in
solution, which we hoped to be able to relate to the bound systems. To determine
expected shielding for the substituents, these compounds were first submitted for
conformational searches in order to determine the most favored low energy conformers.

By comparing the angle of the aromatic plane with the position of the substituents on the

60

a-carbon, we had hoped to be able to predict which substituent would be shielded in each
diastereomer.

The derivatized enantiomers were submitted for a Monte Carlo molecular
mechanics conformational search in Spartan. The conformers were evaluated by relative
energy and conformation of the substituents on the chiral centers. Each compound
studied showed all three conformers (hydrogen, bromine and methyl syn to the carbonyl)
within 1 kcal/mol of the lowest energy conformer. For all conformers under 1 kcal/mo]
from the lowest energy conformer, the medium group of the other chiral center (of the
derivatizing group) was syn to the N-H bond. The problem occurs with the aromatic
group that is expected to change the shielding for the chiral center of the derivatized
carboxylic acid. The angle of the aromatic plane makes it difficult to predict if the
substituents would be shielded or deshielded. This, along with the similar energy of
different conformations on the Ca chiral center, makes it difficult to predict which groups
should be shielded and deshielded, a necessary determination to interpret the NMR data
appropriately. We also modeled different derivatizing amines to try to determine if larger
aromatic or other substituents would assist in locking the conformation of the aryl ring
into a favorable position. It does not appear that an increase in the size of the other
substituents increases the likelihood of being able to determine the average conformation.
The normal Mosher amide-type NMR studies thus did not appear to be useful to us in
determining the conformation of the a-haloamides, since we had no way of interpreting
the chemical shifts that we observed.

To probe expected conformations for the a-haloamides, we were interested in

determining the typical conformation for a-haloamides as reported in the Cambridge

61

Crystallographic Database (CCD). Using Conquest, we searched for a-haloamides in
which the carbonyl was bonded to a metal. We did not restrict the search through the
type of oxygen-metal bond or type of metal. This search produced 274 positive hits;
however, the majority of these compounds were groups with a CF3 group alpha to the
carbonyl. Only three compounds returned by the search were a—monohalogenated
carbonyls. It cannot be determined from this small amount of data if variations in the
O=C-C-X dihedral angles are due to the different types of halogens, different metals, or if
there is simply not much preference for a specific dihedral angle or range of angles. We
then simplified the search and looked for a-haloamides without metal to determine

typical O=C-C-X dihedrals. This search returned 252 hits, but we discarded lactams (due

   

1

O
[x
‘_
I
‘_
(D
._

O=C-C-X dihedral angle

Figure 2-20. CCD data depicting dihedral angles of a—halocarbonyls.

to their locked conformation) and compounds with multiple a—halogens. Figure 2-20
shows the results from this study in the form of a bar graph. Here it is clear to see that
the different halogens do have preferences (at least in crystalline form) for the O=C-C-X

dihedral. Fluorine is shown to have the majority of its conformations with the halogen

62

anti to the carbonyl. Bromine and iodine favor conformations in which the halogen is
perpendicular to the carbonyl, while chlorine favors conformations in which the halogen
is either syn or anti to the carbonyl, with several exceptions. However, there is still
enough diversity of conformations that it is not clear what we could expect with our
systems merely based on reported crystal structures.16

It seemed that an in-depth study of a simple a—haloamide might be useful to check
effects of dipole and charge distribution. We began a series of DFT single point energy
calculations on bromoacetamide to calculate the energy, dipole moment, atomic charges,
and charge density at specific O=C-C-Br dihedrals. Based on these results, the most
favorable conformation still seems to be that in which the bromine is nearly anti (O=C-C-
Br dihedral = 150°) to the carbonyl (Figure 2-21). With this conformation, the energy is
minimized and the charge distribution places most of the negative charge on the bromine
and the oxygen. This conformation corresponds to the one we had originally expected,
which would not be expected to show the ECCD bisignate signs we have observed. It
seems to be clear from our results and others16 that a free a-haloamide would be expected
to exhibit an average conformation where the bromine is anti to the carbonyl and the
plane of the carbonyl bisects the other two substituents. However, it appears that this is
not the conformation that is present when the u-haloamides are guests in our zinc
porphyrin tweezers, as the ECCD bisignates observed cannot be correlated with the

predicted conformation of the free amides.

,er

Br

Figure 2-2 1. Lowest energy conformation for bromoacetamide based on DFT calculations.

63

We became aware of a series of papers published by Taniguchi, et al.,17'19 in
which they observed an apparent hydrogen bonding chain between a carbonyl and an a-
halogen in a-haloacetones. In their system, they believe a 2- to 3-molecule water chain
forms a network between the carbonyl and the halogen, with both acting as hydrogen
bond acceptors. They were able to observe this through Raman spectroscopy in which
the position of the C-Br stretching band shifts upon addition of water. This led us to
consider the possibility that bonding of the porphyrin zinc is occurring with both the
carbonyl oxygen and the a-halogen of the amides—a sort of cooperative bonding which
reverses the normal conformation of the a-haloamides and positions the halogen syn to
the carbonyl as opposed to the expected anti conformation (Figure 2-18). It was unclear
whether binding could be occurring solely through the bromine (i.e., no binding between
the carbonyl oxygen and the zinc) or as a cooperative system involving the halogen, the
metal, and the oxygen. However, based on previous NMR studies of the complexes,15 it
is relatively clear that the oxygen of the carbonyl is probably involved in some type of
binding. It then became necessary to demonstrate cooperative binding in which the
halogen is also in some way coordinated to the metal. We began by trying to reproduce
Taniguchi’s work in silico by modeling a-haloamides with water molecules in Spartan.
Only semi-empirical solvation models are available in Spartan for calculating solvation
without explicitly including water molecules. Initially we performed a DFT geometry
optimization with the AMlaq solvation model on N-methyl-2(R)-bromopropionamide.
The starting geometry had the (rt—hydrogen syn to the carbonyl. The final structure placed
the bromine anti to the carbonyl with a solvated energy nearly 10 kcal/mol lower than the

gas phase (Figure 2-22). This is not improbable since the polar solvent would be

64

 

l

DFT geometry
H Optimization ‘HCH
\ — \ ~
N N 3
H H

'I'Br
CH3 Br

Figure 2-22. DFT geometry optimization of N-methyl-2(R)-bromoacetamide using implicit
solvation leading to a conformation with the bromine anti to the carbonyl.

expected to stabilize the dipoles associated with the carbonyl and the carbon-halogen
bond. We tried a few more calculations with variations in the solvation model and/or
level of calculation with similar results for all calculations. Unfortunately, solvation
studies alone were not assisting in determining if water molecules could be linked
between the halogen and the carbonyl oxygen due to the implicit nature of the models.
We decided to try modeling the water molecules explicitly, hoping hydrogen
bonding in the calculations would account for the possibility of hydrogen bonding with a

halogen. We began the calculation in Spartan as a DFT geometry optimization on N-

Figure 2-23. Geometry optimization of water molecules hydrogen bonded to N-phenyl-Z-
bromopropionamide.

methyl-2(R)—bromopropionamide, placing a single water molecule between the carbonyl
oxygen and the bromine with no constraints or partial bonds. After the end of the first
optimization, the water molecule had moved to the side of the carbonyl at an Oczo-Hmo
distance of 2.079A, far enough away from the bromine that no hydrogen bonding would

be possible. For the second optimization, we inserted a second water molecule in

65

between the oxygen and the bromine as before, with the initial water molecule in the
same position it had at the end of the first minimization. At the end of the second
optimization, the second water remained in position between the oxygen and the bromine
with an 090-an0 distance of 1.961A and a Br-HHzo distance of 3.26A (Figure 2-23).
This distance is a bit too long to be considered reasonable for a hydrogen bond between
water and bromine. To determine if the calculation could even consider bromine as a
hydrogen bond acceptor, we did a third calculation which lefi the first two water
molecules in the same position they had at the end of the second calculation. We then
placed a third water molecule ~lA away from the bromine and resubmitted the molecule
for geometry optimization. Here, the water molecule moved away from the bromine to a
distance of 2.78A, still rather far to be considered a hydrogen bond for most systems
given the van der Waals radii of bromine and hydrogen as 1.85A and 1.2A, respectively.
However, with the large electronic size of bromine, this may not be an unreasonable
distance for hydrogen bonding.

We thought we might be able to use the same methods as Taniguchi and use
Raman spectroscopy to see a change in the C-Br stretch upon binding of the zinc. We
initially tried a solid sample of N-phenyl-2(R)-bromopropionamide in an available
Raman instrument. The acquired spectrum was reasonable although the peaks were of
very low intensity. A peak was apparent around 550 cm", which we thought could be
either a phenyl C-H out-of-plane bending or a C-Br stretch. This was a little low for a C-
Br stretch, which should come around 660-700 cm".

Since we were not able to definitively identify the peaks in the Raman spectrum,

we turned to NMR as a possible method of determining conformation. We continued

66

using the model compound of N-phenyl—2(R)-bromopropionamide. Initially, we had
thought to take NOEs of the amide without and with zinc. We believed that we would be
able to see a change in NOEs from the amide hydrogen to the substituents on the alpha
carbon upon addition of zinc if the conformation was changing as we believed. However,
upon initial NMR testing, we saw changes in the spectra upon addition of zinc without
doing NOEs. This suggested that we might be able to probe the conformation changes
upon addition of zinc through NMR.

A large number of changes are seen in the NMR spectrum for N-phenyl-2-
methylpropionamide at room temperature when zinc bromide is added. The amide N-H
proton moves downfield approximately 2 ppm, possibly due to increased shielding as the
amide N—Cc:0 bond has greater double bond character with the carbonyl oxygen binding
to the zinc. The aromatic proton para to the amide is shifted downfield by 0.03 ppm, also
in agreement with changes due to the carbonyl binding. The methine hydrogen is shifted
downfield by 0.2 ppm. This could be due to changes in the electronic nature of the
carbonyl as a result of binding, or due to a conformational change in which the hydrogen
is no longer shielded as it was before the addition of the zinc. We would expect that,
with no other effects, the medium group on the alpha carbon should be pseudo-anti to the
carbonyl. This would place one of the two methyl groups pseudo-syn to the carbonyl
prior to addition of zinc. The zinc is expected to approach the carbonyl from the same
face as the a—carbon, as that face is less sterically hindered than the opposite face with the
bulky N-phenyl group. The a-methyl would be expected to slide away from the bulky
zinc, leading to a conformation in which the carbonyl bisects the proton and one methyl

group, with the other methyl group anti to the carbonyl (Figure 2-24). This can possibly

67

be verified through the changes in the proton shifts for the methyl groups upon addition
of zinc.

Initially, there were two doublets at 1.22 and 1.14 ppm corresponding to the two
methyl groups. After addition of zinc, these two doublets both shifted downfield 0.04
ppm and a new doublet appeared farther upfield at 1.1 ppm. The two doublets indicate
the initial conformation being the one in which the methyl group is syn to the carbonyl
and therefore shielded. Upon addition of the zinc, the methyl group moves out of the

shielding cone of the carbonyl and shifts downfield. However, the methyl group which

’Zn

anl’z I
[:L Njg;:i3——> O ier3
N H
CH3 H CH3

Figure 2-24. Rotation around the C(C=O)—C(chiral) bond due to coordination of the carbonyl
oxygen with zinc.

was previously in the deshielding cone of the carbonyl is moved anti to the carbonyl, thus
reducing the influence it feels from the carbonyl and causing it to resume a chemical shift
more expected for methyl hydrogens. The same reasoning can be applied to the methine
hydrogen, which shifts downfield 0.2 ppm upon addition of zinc. The hydrogen, which
was outside of the influence of the carbonyl deshielding cone, is now brought within the
deshielding region and shifted downfield.

Based on the modeling and NMR information, it does seem possible that there is
cooperative binding occurring between the carbonyl oxygen, the a-halogen and the zinc
core of the porphyrin tweezer. This conformation would be expected to lead to the
observed ECCD spectra. However, it is unclear at this time if coordination can occur

with other fiinctional groups besides halogens, such as methoxy, phenoxy, or acetoxy.

68

Work is continuing on the synthesis and testing of cyclic a-haloamides and a-chiral
amides with other functional groups to probe possible cooperative binding. These
molecules do not allow for rotation around the Cc=o-Cch,m1 bond, thus eliminating

multiple conformations.

69

References:
1. Yang, Q.; Olmsted, C.; Borhan, B. Org. Lett. 2002, 4, 3423.

2. Schrodinger, Inc. Macromodel, 7.0; 1999.

3. Pillardy, J. C., C.; Wedemeyer, W.; Scheraga, H. Helv. Chim. Acta 2000, 83, 2214.
4. Wiberg, K. 8.; Rush, D. J. J. Org. Chem. 2002, 67, 826.

5. Wiberg, K. B.; Rush, D. J. J. Amer. Chem. Soc. 2001, 123, 2038.

6. Deng, S. T., J. J. Amer. Chem. Soc. 2002, 2002.

7. Radom, L. R., N. V. Australian J. Chem. 1982, 35, 1071.

8. Schnur, D. M. Y., Y. H.; Dalton, D. R. J. Org. Chem. 1989, 54, 3779.

9. Avalos, M. B., R.; Bameto, J. L.; Bravo, J. L; Cintas, P.; Jimenez, J. L.; Palacios, J.
C. J. Org. Chem. 2001, 66, 7275.

10. Langley, C. H.; Lii, J.-H.; Allinger, N. J. Computational Chem. 2001, 22, 1426.
11. Langley, C. H.; Lii, J .-H.; Allinger, N. J. Computational Chem. 2001, 22.

12. Foreman, J. B. F ., A., Exploring Chemistry with Electronic Structure Methods. 2nd
ed.; Gaussian, Inc.: Pittsburgh, 1996.

13. Eliel, E. L.; Wilen, S. H., Stereochemistry of Organic Compounds. John Wiley and
Sons, Inc.: New York, 1994.

14. Wiberg, K. B.; Rablen, P. R.; Rush, D. J.; Keith, T. A. J. Amer. Chem. Soc. 1995,
11 7, 4261.

15. Yang, Q. The Use of Exciton Coupled Circular Dichroism for the Absolute
Stereochemical Determination of Chiral Carboxylic Acids. Ph. D. Dissertation, Michigan
State University, East Lansing, 2004.

16. Structures were taken from the Cambridge Crystallographic Database. Structures
used include: HALTOG, HALVOF, HALWOG, HAMDOO, HAMGOR, HAMLAI,

HAMLUC, HAMMEN, HAMMIR, JIRSIM, NOCWAD, ZITMIY, and ZITWOE (CCD
six-letter codes).

17. Kato, M.; Nanba, Y.; Taniguchi, Y. Chem. Phys. Lett. 1998, 289, 30.

18. Shiratori, Y.; Kato, M.; Taniguchi, Y. Spectrochimica Acta Part A 1999, 55, 2659.

70

19. Shiratori, Y.; Kato, M.; Taniguchi, Y. Spectrochimica Acta PartA 2000, 56, 1693.

71

Chapter 3
Development of Alternate Tweezers for Use With ECCD in the Absolute

Stereochemical Determination of Small Molecules

In the process of derivatizing various carriers with chiral substrates and
monitoring ECCD as discussed in the previous chapter, we observed various ECCD
signs, often conflicting. Depending on the carrier used, the sign of the ECCD curve
would reverse for the same chiral substrate of the same stereochemistry. In other words,
(R)—2-bromopropionic acid would show a positive ECCD sign when derivatized with one
carrier and a negative ECCD sign when derivatized with a different carrier.I This
suggests that there are either different modes of binding between the porphyrin tweezer
and different carriers, dependent on the structure of the carriers, or different
conformations of the derivatized carriers as discussed previously (q.v., Chapter 2).
Determining the mode of binding for each carrier can be problematic, and potentially
impossible without a crystal structure. Part of this difficulty lies in the large amount of
rotational freedom present in the 1,5-diester linker joining the porphyrin chromophores

studied. Because there is so much flexibility in the linker, the porphyrins can wrap

 

Figure 3-1. Representation of possible differences in binding between porphyrin tweezers and guest
molecules, leading to opposite ECCD signs.

72

around the guest molecules in various ways, as shown in Figure 3-1.

Molecular modeling indicates there are numerous possible low energy
conformations which differ in how the tweezer wraps around the guest (q. v., Chapter 2).
In light of this, we were interested in designing a tweezer with a more rigid linkage. This
would assist us in a number of ways, including making the prediction of the ECCD sign
easier, rigidifying the host molecule to allow us to predict conformation with more
confidence and easing molecular modeling costs in terms of number of conformations
found and time for calculations.

3-1. Porphyrin Tweezer with a Diacetylene Linkage

The simplest linker that would afford rotation about an axis perpendicular to the
porphyrins would consist of a diacetylene linker.2 Rotational freedom would be afforded
through the single bond connecting the triple bonds. If the acetylenes were connected to
the para position of the porphyrins, as with the diester linkage, the porphyrins would be
aligned at 180°, thereby showing no ECCD, and in fact unable to bind to diamines due to
the lack of flexibility in the linker. However, if the acetylenes are connected through the
meta positions of the phenyl substituents on the porphyrins, the porphyrins will have a
fixed angle of approximately 120° relative to the acetylene as shown in Figure 3—2.
Although the porphyrins are not necessarily face-to-face when joined by a bisacetylene
linker, when a compound binds to the two zinc cores of the porphyrins, the porphyrins
will swing around to face one another. Chirality of the guest substrate will influence the
host porphyrin tweezer to adopt either a clockwise or counterclockwise orientation of the
chromophores as viewed from front to back, leading to an ECCD bisignate curve. With

this goal in mind, we began the synthesis of porphyrin tweezer 1 (Figure 3-3).

73

 

 

 

 

Figure 3-2. Indication of bite angle in proposed meta-substituted bisphenylacetylene linker.

A simple retrosynthetic analysis reveals that the most logical way to construct the
desired tweezer is by homocoupling of an acetylenic porphyrin. In light of this, 3-
[(trimethylsilanyl)ethyny1]benzaldehyde 3 was synthesized from 3-bromobenzaldehyde
and (trimethylsilyl)ethyne 2 through a palladium-catalyzed coupling.3 This was then
used as the building block with benzaldehyde and pyrrole for the synthesis of the
monosubstituted porphyrin to give 5,10,15-triphenyl-20-(3-
trimethylsilanylethynylphenyl)porphyrin 4 (Scheme 3-1).

Removal of the TMS group by TBAF yielded the deprotected acetylene porphyrin

5. Initial homocoupling reactions of 5 using standard Glaser-Hay coupling conditions did

74

 

 

Figure 3-3. Proposed rigid porphyrin tweezer l utilizing a bisacetylene linker.

not proceed to give the desired product 1.4 The first reaction attempted was performed on
a non-metallated (free base) porphyrin system, which bound to the copper used in the
reaction, preventing the coupling reaction from occurring.5 To prevent this, the
porphyrin was zincated with Zn(OAc)2 to give zinc porphyrin 6. The coupling was then

performed under an atmosphere of oxygen using Cu(OAc)2-H20 in pyridine, and after

HO 2 HO
E—TMS g
Pd(OAc)2, PPh3, TEA
Br 100 °C, 10 h, 77% \\

 

1) benzaldehyde, pyrrole,
BF3‘OEt2,- it, 12 h

‘

2) p-chloranil, 45 °C, 2 h,
44%

 

Scheme 3-1. Synthesis of (trimethylsilyl)acetylene derivatized porphyrin 4.

75

TBAF (1 M in TH F)
THF, rt, 24 h, 98%

 

 

 

 

Zn (OAc)2 Cu(OAc)2+|20 _
CHzclz, rt, pyridine, 02'
12 h, 77% 40 °C, 24 h, 88%

 

 

 

Scheme 3-2. Synthesis of bisacetylene tweezer 1.

preparatory thin layer chromatography the homocoupled tweezer l was isolated in 88%

yield (Scheme 3-2).3‘ 5 UV studies show an absorbance at 414nm (n-hexane; 416 nm in

76

methylcyclohexane), within the region expected for this type of metalloporphyrin
tweezer.

To explore binding of this system, diamines of different lengths were used
including 1,2-diaminopropane; 1,3-diaminopropane; 1,5—diaminopentane; L-ornithine;
and L—lysine methyl ester. Initial titrations of achiral diamines were only somewhat
successful with the diacetylene tweezer. 1,2-Diaminopropane and 1,3—diaminopropane
showed no binding. 1,5-Diaminopentane showed a 7 nm red shift with addition of one
equivalent of the diamine. The shift occurred from 416 nm to 423 nm (Figure 3-4).

Nearly complete conversion to bound tweezer is seen at one equivalent, with a small

 

 

 

300 350

400 450 500
Wavelength (nm)

Figure 3-4. Red shift in UV-vis absorption of tweezer 1 with 1,5-diaminopentane.
shoulder still present at 416 nm. Attempts were made at inserting a chiral diamine as the
guest in the tweezer host. Both L-lysine methyl ester (a 1,5-diamine) and L-omithine (a
1,4-diamine) were used in UV studies. Most titrations with L-omithine did not show any
binding, even with 100 equivalents of the diamine. However, in two titrations a shift was
observed with the addition of 100 equivalents. At 100 equivalents, two peaks of equal
intensity are observed, one at 416 nm and one at 424 nm. With the addition of another

100 equivalents (total 200 equivalents) nearly complete conversion to the second peak at

77

424 nm is observed, with a small shoulder still present at 416 nm. There are a couple of
possibilities for a shift occurring with such large equivalents of diamine. The shift could
be due to residual ethanol still present from the deprotonation of the diamine, or it could
be due to coordination of amine to the zinc outside of the tweezer pocket, i.e., not
bidentate diamine binding. A titration of the tweezer with ethanol was undertaken to
determine if residual ethanol could create the same shift as that seen with L-ornithine.
Upon titration with ethanol, a shift is seen from 416 nm to 423 nm, leading us to believe
that it is possible that residual ethanol is binding to the tweezer instead of L-omithine.
The L-ornithine may still be binding in some capacity to the tweezer, but it was unclear at
this point exactly what type of binding, if any, was occurring. Another possibility is that
the lack of binding and unusual binding pattern might be due to the carboxylic acid
present in the amino acid. Synthesis of the methyl ester was undertaken but was
unsuccessful using trimethylsilyldiazomethane.6

In testing other chiral diamines, L-lysine methyl ester did show binding with the
bisacetylene tweezer with a 8 nm shift from 416 nm to 424 run through addition of 15
equivalents, however, as discussed above no other size of diamines were found to show
any binding to the bisacetylene tweezer 1. Unfortunately, we were unable to observe
any ECCD curves when L-lysine methyl ester was coordinated to the host porphyrin
tweezer. This could be due to the rigidity of the system, which may prevent optimum 1:1
binding of the guest diamine to the host tweezer. After viewing molecular models of this
compound with various size diamines, it appears that the structure of the tweezer is too
rigid to permit binding to diamines of various sizes. Although the tweezer can be

modeled with different size diamines, a large amount of strain is seen in the linkage

78

between the porphyrins when large diamines are used (LS-diamines and larger). Similar
strain is observed with smaller diamines, as the porphyrins cannot flex enough through
the linker to facilitate binding.

To possibly help alleviate binding problems due to the rigidity of the system,
another synthesis was attempted to create porphyrin tweezer 7 (Figure 3-5). This tweezer
would maintain the diacetylene linker but give more flexibility to the system through two
methylene groups that attach to the linker in the para position of the porphyrin phenyl

group. Two synthetic schemes were proposed such that convenient and inexpensive

 

 

Figure 3-5. Alternative bisacetylene porphyrin tweezer 7.

starting materials could be used. One route involved building the zinc porphyrin around
the bromobenzaldehyde first. Grignard transformation followed by reaction with 1-
bromo—3-trimethylsilylpropyne would lead to the trimethylsilylpropargylporphyrin. After
addition of zinc and deprotection of the trimethylsilyl group, a Glaser-Hay coupling
would lead to the desired porphyrin tweezer (Scheme 3-3). A second route would create
a protected bromobenzaldehyde that would then be transformed into a Grignard reagent

to react with l-bromo-3-trimethylsily1propyne (Scheme 3-4). The porphyrin would then

79

HO 1. pyrrole (4 equiv),
benzaldehyde (3 equiv),
BF3°OEt2

2. p-chloranil
Br

 

  

1. Zn (OAc)2

 

 

r

 

2. TBAF
3. CU(OAC)2! 02

 

Scheme 3-3. Proposed synthesis of alternate tweezer 8.

80

be built off of the deprotected aldehyde, the trimethylsilyl group deprotected, and a
coupling reaction performed.

Following the first method described (Scheme 3-3), bromoporphyrin 9 was
synthesized using p-bromobenzaldehyde, benzaldehyde, and pyrrole with BF3°EtzO and
p-chloranil. Separation was effected using 20% CHZCIZ/hexanes. The first band from the
column was collected and dried to give bromoporphyrin 9 in 98% yield. Reconsidering
that this synthesis had the potential for loss of a large amount of starting material during
the addition of the propargyl group, we decided to build the porphyrin onto the aldehyde
with the propargyl group functionality already present on the aldehyde and did not pursue
this route any further.

Following the second method described above (Scheme 3-4), formation of the
protected bromobenzaldehyde 10 through the use of p-toluenesulfonic acid and ethylene
glycol proceeded smoothly in 74% yield. The first attempt at attaching the
trimethylsilylpropargyl group was by transforming the protected bromobenzaldehyde into
a Grignard reagent through the use of activated magnesium. The reaction was heated to
40 °C and then refluxed for 12 h after addition of all reagents. Although some starting
material was still present, a large mixture of products was seen, none of which was the
desired product by NMR. The second attempt involved using t-butyl lithium with the
protected bromobenzaldehyde to effect a lithium-halogen exchange. The reaction
occurred with violent sputtering, and after work—up no starting material or desired
product was evident. The third reaction followed literature precedent for alkylation of o-
bromobenzaldehyde7 and started by forming the lithiated compound of the protected p-

bromobenzaldehyde. Reaction with magnesium bromide was expected to give the

81

 

 

 

 

/ \ O O
H O 0
Br
B
r Br Q
TMS
" ”i O Q ‘ W”
2. pyrrole, benzaldehyde, 2. TBAF
BF3.OEt2 3. CU(OAC)2, 02

3. p-chloranil

TMS

 

 

Scheme 3-4. Proposed synthesis of tweezer 8 using a functionalized aldehyde.

Grignard, to then react with the trimethylsilylpropargyl bromide. Once again no starting

material or desired product was evident (Scheme 3-5) with a large number of peaks

evident by 1H NMR and several spots by TLC chromatography. At the same time,

synthesis was attempted for 3-bromo-1-(trimethylsilyl)propyne starting from

82

1. Mg°, THF, 40 °C

2. propargyl bromide
O

H
O HO\/\OH_ X 1.t-BuLi,Et20,0°C

 

cat. p-TSA. 2. propargyl bromide
benzene,

120 °C, 12 h
3’ Br 1. seoBuLi, MgBr2

-40 °C, Etao r
2. propargyl bromide

 

Scheme 3-5. Attempted synthetic routes for p-(I-trimethylsilylpropargyl)benzaldehyde.

(trimethylsilyl)acetylene and methylene bromide.8 NMR spectra indicated no reaction
after a number of attempts. After the difficulties encountered in synthesizing the
bispropargyl porphyrin tweezer 8, this route was abandoned in favor of other promising
linkers.
3-2. Porphyrin Tweezer with a F errocene Linkage

In pursuit of a more rigid tweezer with greater bite angle flexibility than the
bisacetylene tweezer, we decided to synthesize tweezers with a ferrocene-based linker
(Figure 3-6). Although the rigidity of the linker again could pose issues for the ability of
the tweezer to bind diamines of various sizes, the nature of the ferrocene linker permits
more flexibility than covalently bound linkers, as the bond from the pentadienyls to the
iron core can expand and contract slightly. A bisporphyrin ferrocene has not previously
been reported, although the monosubstituted compounds have been synthesized. The
synthesis of the monoporphyrin ferrocene is accomplished through the use of a ferrocene
monoaldehyde which is then reacted with a dipyrrole and benzaldehyde to form the
desired porphyrin on one of the ferrocene rings (Scheme 3-6).9 However, use of this

method for a bisaldehyde ferrocene could be problematic, since it will be very difficult to

83

 

HO 1. TFA, 01-12012, rt

2. DIEA

 

4 . Zn (OAC)2

   

3. p-chloranil l
R

 

 

 

 

 

Scheme 3-6. Reported synthesis of porphyrin monosubstituted ferrocenes.

84

control the reactions with multiple reactive aldehydes in solution. In light of this, we
attempted the synthesis of a bisporphyrin ferrocene through coupling of a ferrocene
derivative to a porphyrin derivative.

Our first attempt at synthesis of 11 involved palladium coupling of
bis(tributyltin)ferrocene and monobromoporphyrin. Monobromoporphyrin 9 was
synthesized using p-bromobenzaldehyde, benzaldehyde, and pyrrole as shown in Scheme
3-7. However, we realized that the integrations for the proton NMR were not calculating
correctly even though it appeared we only had one product by thin layer and column
chromatography. We submitted the product from one reaction for mass spectrometry

(MS) analysis. MS indicated a mixture of tetraphenylporphyrin; 5-(p-bromophenyl)-

HO HO H
N 1. BF3°OEt21 rt! 7 h
+ 3 + 4 U 2, p-chloranil, rt,
Br

 

2 h, 98%

 

Scheme 3-7. Synthesis of 5-(p-bromophenyl)-10,15,20-triphenylporphyrin 9.

10,15,20-triphenylporphyrin; 5,10-di(p-bromophenyl)-15,20-diphenylporphyrin; and
5,10,15-tri(p-bromophenyl)-20-phenylporphyrin products. Unfortunately, thin layer
chromatography under several solvent systems only shows one spot (not unexpected
since there should not be much difference in polarity between porphyrins with different

number of bromines attached). We tried to separate the compounds by slow column

85

chromatography, extractions, and recrystallizations, but no separation was apparent by
NMR. Since additional brominated compounds will complicate the coupling of 9 and 13,
we needed to find a way to make only monobrorrrinated porphyrin. We decided to try
synthesizing the 5-(p-bromophenyl)-10,15,20-triphenylporphyrin 9 using an excess of
benzaldehyde, which should increase the production of tetraphenylporphyrin (TPP) and
decrease the production of all of the brominated compounds, hopefully leading to
production of only monobrominated product 9 plus TPP. TPP will not react in the
coupling, so we do not need to separate it from the desired product. We ran the same
brominated porphyrin synthesis reaction with nearly 7 equivalents of benzaldehyde (as
compared to 3 for a typical synthesis of monosubstituted porphyrin), with the normal
ratio of 4:1 for pyrrole:4-bromobenzaldehyde. After column chromatography,
benzaldehyde was still present in the product so the desired product was recrystallized
from CHCl3/hexanes. NMR of the recrystallized product showed no benzaldehyde and
matched the expected integrations for the desired monosubstituted product 9, possibly
contaminated with a small amount of tetraphenylporphyrin. Since TPP will not react in
the coupling reaction, we have used the 5-(p-bromophenyl)-10,15,20-triphenylporphyrin
9 produced in this reaction without further purification.

Bis(tributyltin)ferrocene 12 was synthesized through a reported procedure
(Scheme 3-8).1°’ H While the syntheses of 9 and 12 proceeded in good yield, the coupling
was more problematic. We started by trying to couple the bromoporphyrin 9 and
bis(tributyl)tin ferrocene 12 using PdC12(PPh3)2 in dry DMF (Scheme 39).” '3 The
bis(tributyltin)ferrocene did not dissolve in DMF when added but the bromoporphyrin 9

was added to the suspension to yield a purple solution. After workup, we obtained only

86

 

hexane, rt, 12 h
Fe =
2. SHBU3CI,

1. n-BuLi, TMEDA, ESQ; Sn(Bu)3
0 °C" rt, 2 h, 63%

Fe
12
Scheme 38 Reported synthesis of bis(tributyltin)ferrocene 12.

nBua nBu3
PdCl2(PPh3)2

+ Fe = X
dry DMF, 120 °C, 3 h

 

 

Scheme 39 Attempted coupling of 5-(p-bromophenyl)-10,15,20-triphenylporphyrin 9
with bis(tributyltin)ferrocene 12 using PdC12(PPh3)2.

starting material, most likely due to lack of solubility of the bis(tributyltin)ferrocene 12 in
the reaction mixture.

We next tried a coupling using Pd2(dba)3 and AsPh3 (Scheme 3-10).14 This
reaction was run per literature precedent. After workup, the residue was
chromatographed on silica gel with a 1:1 mixture of CH2C12 and hexane. The first band
was yellow, the second orange, and the third purple. The third band was dried and
recrystallized with CHCl3/hexanes to give only starting material. Since after repeated
trials the palladium couplings were not working, we decided to switch coupling reactions.
Due to the difficulties in installing multiple functionalities on the porphyrin, it seemed
ideal to keep the monobromoporphyrin 9 as a starting material and change the ferrocene

functionality.

87

 

THF, 55 °C, 5 d

nBU3 nBu3
Pd2(dba)3, ASPh3
+ Fe ¢ X
12

 

Scheme 3-10. Attempted coupling of 5-(p-bromophenyl)-10,15,20-triphenylporphyrin 9
with bis(tributyltin)ferrocene 12 using Pd2(dba)3.

We decided to synthesize the ferrocene tweezer using bis(boronic acid) ferrocene
13 and 5-(p-bromophenyl)—10,15,20-triphenylporphyrin 9 (Scheme 3-12). Arylation of
ferrocene derivatives through the boronic acids has been reported.‘5 However, formation
of the bis(boronic acid) ferrocene 13 using TMEDA, n-butyl lithium, and tributoxyborate
was initially problematic. The reported procedure was done on a 10 g scale, but the first
time we did the reaction (on a 1 g scale), we did not obtain any product. We were
concerned that perhaps the reaction had not been dry enough, so the ferrocene was dried
in the oven for 3 h prior to attempting the reaction again. After reaction completion, we
obtained an orange solid, but this decomposed upon evaporation of the solvent under
vacuum (possibly due to high temperature) and turned dark brown while 1H NMR
showed no product.

On the third attempt we dried the ferrocene in the oven overnight, during which
time it sublimed and formed crystals. This reaction proceeded as before but using

triisopropyl borate instead of the reported tributyl borate, and upon addition of the

88

ferrocene dianion to the boronic ester the solution turned orange. The solution
precipitated yellow solid upon acidification, and this was filtered and washed with ether
to give a brown solid of very low yield (2.2%) which was paramagnetic. The 1H NMR
peaks from this compound match those reported,15 but there is no mention in any of the
references about the desired compound being paramagnetic.

After some discussion, the TMEDA was distilled directly before using it in the
synthesis of the ferrocene bis(boronic acid) 13. Following the same procedure as before
but using tributyl borate, the synthesis worked and we obtained 13 in 36% yield (Scheme
3-11). Although the yield was low, the reaction was done on a large scale and no

optimization was pursued since over 6 g of material was obtained.

1. n-BuLi, TMEDA, 3(0le 3(OH)2

G ‘0 E120 , rt, 12 h
Fe > Fe
2. (BuO)3B,

-78 °C->rt, 24 h
3. 10% aq. KOH, 13
36%

 

Scheme 3-11. Synthesis of bis(boronic acid)ferrocene 13.

We next proceeded to couple the monobrominated porphyrin 9 with the
bis(boronic acid) ferrocene 13 (Scheme 3-12). Following a literature procedure,IS
coupling of 13 and 9 yielded the bis(tetraphenylporphyrin) ferrocene 14 in 95% yield. It
was a little surprising that the coupling worked in such good yield, so we submitted the
product for MALDI (necessary due to the high molecular weight). Major peaks were
found at 1545, 1466, 1387, 1307 with a matrix of N-bromoacetamide (138 g/mol). The
product compound is expected to have a molecular weight of 1411.47. Based on the peak
at 1545, subtracting out the matrix gives a molecular ion peak of 1407, which would be

the expected ion peak if the product lost four N-H protons. This could occur due to the

89

B(OH)2 B(0H)2

 

1. Pd(PPh3)4, 2 M Na2003
o + .. =
toluene, 95 °C, 5 d, 74%
13 2. Zn(OAC)2. CHZCIZ, rt,
3 d, 60%

 

Scheme 3-12. Coupling of zinc 5-(p-bromophenyl)-10,15,20-triphenylporphyrin with bis(boronic acid)
ferrocene.

basic workup. This led us to believe we had formed the desired compound. Zincation of
this compound followed the coupling to give bis(zinc tetraphenylporphyrin) ferrocene 11.
We then took 11 and subjected it to UV-Vis testing. All compounds tested (n-
butylamine, 1,3-diaminopropane, l,4-diarninobutane, and 1,10—diaminodecane, L—lysine,
and L-omithine) show binding and approximately 10 nm shifts from 416.5 nm to 427 nm

using from 1 to 5 equivalents of diamine. However, ECCD of the ferrocene porphyrin

9O

   

  

 

 

 

\S
r
PSL H2
h” H2N’-N -
HzN-NHz
° ' i s
l l
v Fe ‘4

Fe
11

Figure 3-7. Representation of possible binding outside of tweezer core for bis(zinc porphyrin)
ferrocene 11.

tweezer with chiral substrates has not been obtained using L-lysine or L-omithine. It is
possible that this tweezer suffers from a similar problem as the diacetylene in that the
binding pocket of the tweezer is too small to permit guests into the pocket, thus
preventing observance of ECCD. Binding may occur on the outside of the tweezer
pocket as shown in Figure 3-7. The iron core of the ferrocene may also be interfering
with the substrate binding leading to unexpected or unpredictable conformations that are
not ECCD active.
3-3. Porphyrin Tweezer with a Urea Linkage

It has recently been reported in literature that porphyrins with a urea linker can be
made to maintain a face-to-face orientation. Takagishi, et al., showed that they were able
to make a porphyrin tweezer where both porphyrins are held face-to-face through the use
of bulky methyl groups which interact with the tweezer linker (i.e., steric interactions
between the methyl groups on the phenyl ring and the urea carbonyl prevent rotation
around the urea bond, Figure 3—8).16 We thought this might be useful to allow us to form
porphyrin oligomers or polymers with urea linkages. Steric interactions between the

phenyl substituents on the porphyrins prevent them from stacking directly on top of one

91

 

Figure 3-8. Representation of Takagishi’s dimer showing steric interaction between the urea carbonyl
and the methyl substituents on the phenyl rings.

another, leading to a helical structure. The addition of a chiral guest could then be used
to influence the direction of the helix (P or M), which in turn could be monitored through
ECCD (Figure 3-9). The electric dipole interaction between a series of porphyrins should
lead to an amplified ECCD signal with very small amounts of material. It could be
possible then to control the number of porphyrins (or polymeric units) to change
solubility, ECCD amplitude, etc. Based on molecular modeling, these systems form
beautiful helices as expected (Figure 3-10). The synthesis of these systems appeared to
be very simple and straightforward with a minimum of steps.

Synthesis of the diamine porphyrin necessary for formation of the amide linkage
can be obtained through known literature procedures. Most syntheses begin with a

”"9 so it was necessary to obtain 4-methyl-3,5-

nitroaldehyde to form aminoporphyrins,
dinitrobenzaldehyde 16 which would allow us to make the desired monosubstituted

porphyrin. This aldehyde is not commercially available, so there were two routes that we

looked at using to get this compound. One was nitration of the methylaldehyde, but this

92

 

 

Figure 3-9. Schematic showing control of helix direction upon addition of a chiral guest to a porphyrin
trimer connected by urea linkages, leading to ECCD.

o
.0

. O

0.:

\ r «.0
\9 .

Figure 3-10. Spartan molecular model of helix formed from urea-linked porphyrins (phenyl rings not
shown for clarity).

 

 

 

 

 

 

 

 

 

 

led to concerns of ovemitration and potential generation of explosive products. We opted
for a second route starting from 4-methyl-3,5-dinitrobenzoic acid, which is inexpensive
and readily available. Reduction of the acid to the alcohol 15 proceeded smoothly with

BH3°THF solution, which is reported to be a mild reduction technique that does not

,0
BHgTHF
THF, rt. CH2CI2, rt,
02 12h, 80 0/0 02N NO 02 12h 50% 02N N02

16

 

Scheme 3-13. Synthesis of alcohol 15 and benzaldehyde 16 from 4-methyl-3,5-dinitrobenzoic acid.

reduce nitro groups, even at room temperature and extended reaction times.20 Yields of
this reaction were good for both small and large scale with average yields around 80%.
Oxidation of the alcohol to the aldehyde was accomplished with pyridinium
chlorochromate (PCC) (Scheme 3-13). Column chromatography and recrystallization
gave pure 4-methyl-3,S—dinitrobenzaldehyde 16 in yields of 50% from first-crop
recrystallization.

Once the aldehyde 16 was in hand, attempts were made to synthesize the 5-(4-
methyl-3,5-dinitro)-10,15,20-triphenylporphyrin 17 (Scheme 3-14). Use of the standard
’ monosubstituted porphyrin synthesis (1:3:4, nitroaldehyde:benzaldehyde:pyrrole) gave
porphyrin products with some 4-methyl-3,5-dinitrophenyl substituents. Integration of the
proton NIVIR peaks indicated that we had formed at least some of the disubstituted
product, so the compound was submitted for mass spectroscopy. Analysis of the mass
spectra indicated that the product was a mixture of compounds comprised mainly of

disubstituted and trisubstituted porphyrin, with some monosubstituted also present.

94

Separation of these compounds appeared straightforward based on TLC, but column
chromatography was run repeatedly with very little separation. Adjustment of the
reactant ratios (0.5:3.5:4, nitroaldehyde:benzaldehydezpyrrole) also produced mono-, di-,
and trisubstituted porphyrins, although with less trisubstituted than previously.

Separation by column chromatography was still unsuccessful.

 

/O
1. pyrrole, benzaldehyde,
BF3'OEt2, CHZCIZ, rt, 4 h
2. p-chloranil, rt, 1 h ,
OZN N02
16

 

+ disubstituted and
trisubstituted products

Scheme 3-14. Synthesis of 5-(4—methle3,5-dinitrophenyl)-10,15,20—triphenylporphyrin l7.

”"9 we tried

Pursuing other reported methods for nitroporphyrin synthesis,
formation of the monosubstituted porphyrin using propionic acid. This reaction is

reported to give very low yields of the monosubstituted product, but when we tried it

HO HO H
N propionic acid
+ 3 + 4U 60°C,30m x
16

Scheme 3-15. Attempted synthesis of 5-(4-methyl-3,5-dinitro)-10,15,20-triphenylporphyrin using
propionic acid.

 

(1:3:4, nitroaldehyde:benzaldehydezpyrrole) no porphyrin was formed in the crude

product mixture by NMR (Scheme 3-15). A second reaction (2:324,

95

nitroaldehyde:benzaldehydezpyrrole) gave a very small amount of porphyrin, but the
product mixture was mostly insoluble black solid.
Other monosubstituted porphyrin syntheses in our lab have been successful using

21. 22

dipyrroles to generate disubstituted porphyrins, so we thought perhaps this would be

a way to also obtain the monosubstituted compound we desired. We began by reacting

 

/O
OzN NO2 \ / “330%“,
HO
H TFA

 

+ \/ rt,10m,
4

/o

19

Scheme 3-16. Synthesis of (4-methyl-3,5-dinitrophenyl)dipyrrole 18 and phenyldipyrrole 19.

the dinitroaldehyde with excess pyrrole in the presence of trifluoroacetic acid to give (4-
methyl-3,5-dinitrophenyl)dipyrrole 18 in 90% yield (Scheme 3-16). We thought that
using two dipyrrole units might give a better yield than using one dinitrodipyrrole plus
pyrrole and benzaldehyde, thus phenyldipyrrole 19 was also formed by reacting excess
pyrrole with benzaldehyde. This reaction was less successful (probably due to the
decreased reactivity of benzaldehyde as compared to 4-methyl-3,5-dinitrobenzaldehyde)
but gave 49% yield of the desired product after purification (Scheme 3-16).

While searching for references on the use of dipyrroles, we came across a paper in
23

which the authors had used various salts to increase the yield of porphyrin reactions.

An addition of 0.1 equivalents of sodium chloride is reported to give the best results, so

96

we attempted the use of sodium chloride in the synthesis of (4-methyl-3,5-
dinitrophenyl)dipyrrole 18. Using freshly distilled pyrrole, this reaction yielded >99% of
the desired product after purification.

Coupling of the two dipyrroles 18 and 19 was attempted using benzaldehyde and
TFA. However, when p-chloranil was added, the solution turned purple immediately and
then brown. Workup of the products gave only insoluble compounds and a very small
amount of soluble product, which was not enough for NMR. The same reaction
conditions were tried again with dinitrodipyrrole 18, pyrrole, and benzaldehyde (1:223).
Once again no porphyrin products were present in any of the bands separated by column
chromatography. Reactions catalyzed with trifluoroacetic acid (TFA) seem to be much
faster than those catalyzed by BF3°OEt2, and the additional reactivity of the
dinitrodipyrrole may be causing the formation of unwanted side products in the TFA
reactions. The reaction of dinitrodipyrrole 18, pyrrole, and benzaldehyde (1:3:2) was
attempted again but with BF3-OEt2 instead of TFA. After oxidation and column

chromatography, the desired 5-(4-methyl-3,5-dinitrophenyl)-10,15,20-triphenylporphyrin

 

/ NH HN \
/ \ 1. pyrrole, benzaldehyde,
BF3'OEt2, CHZCIZ, rt, 4 h
2. p-chloranil, rt, 1 h, 17 %
02N N02
18

 

17

Scheme 3-17. Synthesis of 5-(4—methyl-3,5-dinitrophenyl)-10,15,20-triphenylporphyrin 17 from (4-
methyl-3,5-dinitrophenyl)dipyrrole 18.

97

17 was obtained in 17% yield (Scheme 3-17). Reported yields in literature range from 4-
9% using various other methods (such as the ones initially used here, i.e., propionic acid,

' 19 so thrs method is a Significant improvement.

etc.),

In reviewing these procedures, it seemed to be superfluous to workup the
dipyrrole reaction before formation of the porphyrin. Avoiding the workup would
eliminate the need to remove excess pyrrole after the dipyrrole reaction before adding
more pyrrole for the porphyrin formation. These two reactions were then attempted
consecutively, with no intermediate workup. Following similar procedures as before,
dinitrobenzaldehyde 18, sodium chloride, and 15 equivalents of pyrrole were combined
in a small flask. After stirring for 5 minutes, TFA was added and the solution was stirred
for another 5 minutes. This solution was then diluted, and benzaldehyde and additional
solvent added. This solution was stirred under nitrogen for 30 minutes, then shielded
from light upon addition of BF3-OEt2. After stirring at room temperature overnight and
successive column chromatographic separations, the desired product 17 was isolated in
3% yield. While this was a very low yield, it did indicate that the reaction could be done
without an intermediate workup and the yield was not significantly worse than those
previously reported in literature using other conditions.21 As we had obtained a
significant amount of tetraphenylporphyrin (T PP) from the reaction, we thought changing
the reactant ratios might lead to increased product and we initially tried increasing the
benzaldehyde. The same reaction as above was performed but with 6 equivalents of
benzaldehyde, giving the product in 10% yield, but with an increased amount of TPP

formed, as well as formation of phenyldipyrrole 19. The same reaction again with 1

equivalent of benzaldehyde led to isolation of the product in 14% yield with almost no

98

1. pyrrole, TFA, rt, 10 m

2. benzaldehyde, CHZCIZ,
02N N02 BF3'OEt2, rt, 12 h
3. p-chloranil, rt, 2 h, 14 %
16

 

Scheme 3-18. Synthesis of 5-(4-methyl-3,5-dinitrophenyl)-10,15,20-triphenylporphyrin 17 using a
pseudo “one pot” procedure.

TPP formed (Scheme 3-18). This brings us close to the same yield obtained after doing
the reaction in two separate steps, indicating this is a simpler way for forming 5-(4-
methyl-3,S-dinitrophenyl)-10,15,20-triphenylporphyrin 17 as it decreases the time needed
for intermediate workup and reduces the use of excess pyrrole and chromatography

solvents.

1. HCI, SHC|2°H20,
CHQCIZ, rt, 12 h

2. aq. NaHCOa, 7 %

 

17

Scheme 3-19. Reduction of the nitro groups of 17 to give 5-(3,5-diamino-4-methylphenyl)-10,15,20-
triphenylporphyrin 20.

99

The next step in the synthesis of the urea-linked tweezer is reduction of the nitro
groups to amines. Initial attempts at reducing the nitro groups of 17 with H2 and
palladium on carbon24 led to reduction of only one nitro group. Reduction using various
other literature methods was also attempted.25 Reported methods such as SnClz°HzO and
HCI26 did not work as either no reduction occurred or only one of the nitro groups was
reduced. Reduction using a protocol developed in our lab for reduction of aryl nitro
groups was also attempted. This method uses zinc dust and CaClz in ethanol and water to
reduce nitro groups.27 However, the starting porphyrin 17 is not soluble in ethanol, and
attempts to use other solvent systems (such as tetrahydrofuran) did not succeed, as only
starting material was recovered in all cases. Modification of the San procedure with a
basic workup finally yielded 5-(3,5-diamino-4-methylphenyl)-10,15,20-
triphenylporphyrin 20, although in low yield (7% yield, Scheme 3-19). Coupling of the
diamine with one equivalent of triphosgene produced multiple products (Scheme 3-20).28
Separation of the products by column chromatography led to the isolation of three bands.

1H NMR of each band revealed an interesting phenomenon. The first band obtained from

 

triphosgene, TEA
CHZCI2, rt, 12 h

 

Scheme 3-20. Formation of porphyrin polymer 21 from 5-(3,5-diamino-4-methylpheny1)-10,15,20-
triphenylporphyrin 20 and triphosgene.

100

column chromatography showed broader peaks than normally observed for substituted
porphyrins but in the expected region for porphyrin protons, with a peak at 8.81, 8.17,
and 7.73 ppm. The second compound obtained showed only broad singlets, still at 8.81,
8.18, and 7.73 ppm. The third compound also showed very broad peaks at 8.81, 8.18,
and 7.74 ppm, but with a contamination showing distinct splitting with a doublet at 8.056
(J=7.2Hz). and two triplets at 7.561 (J=7.5Hz) and 7.422 ppm (J=7.2Hz). The first band
obtained was thought to be the desired dimer, but this was difficult to determine by
NMR, so each band was submitted for MALDI mass spectrometry. MALDI revealed
that the three bands were higher order polymers, instead of the dimer, trimer, and
tetramer we had anticipated. While waiting for MALDI results, the suspected dimer was
submitted for Zincation with an excess of zinc acetate, which proceeded in good yield
(Scheme 3-21). UV-Vis of the zincated compound showed a peak at 428 nm. Binding
with chiral substrates did not lead to any observed ECCD; however, this compound
should be investigated further for use in stereochemical determination as well as a
possible useful molecule for other applications such as electron transport and energy
transfer.29

Although we have not yet found the “ideal” porphyrin tweezer, the work that has
been done to date has provided insight into the types of tweezers necessary for ECCD-
active host-guest complexes. Too many degrees of freedom in the tweezer linker will
generate complexes with many possible conformations leading to ambiguous ECCD
spectra. However, tweezers that are too rigid do not permit incorporation of a wide
enough range of guests to be practically useful. Of course, all of these conclusions

should be accepted cautiously as the structures of compounds 1, 4, 5, 6, 11, 14, 19, 21,

101

 

 

Scheme 3-21. Synthesis of zinc porphyrin polymer 22 from porphyrin polymer 21.

and 22 are based solely on 1H NMR data. An ideal tweezer therefore is one that has
minimal degrees of freedom in the linker but can accommodate guests of varying size.
Further work with metal complex tweezers (similar to the ferrocene-based tweezer) or
tweezers with linkers that can act as rigid or flexible (such as conjugated dienyl linkers)

may hold the key to an ideal host bisporphyrin tweezer.

102

Experimental materials and general procedures:

Anhydrous CHzClz was dried over CaHz and distilled. The solvents used for CD
and UV-Vis measurements were purchased from Aldrich and were spectroscopy grade.
All reactions were performed in dried glassware under nitrogen unless otherwise noted.
Column chromatography was performed using SiliCycle silica gel (230-400 mesh). 1H-
NMR spectra were obtained on Varian Inova 300 MHz or 500 MHz instruments and are
reported in parts per million (ppm) relative to the solvent resonances (5), with coupling
constants (J) in Hertz (Hz). IR studies were performed on a Galaxy series FTIR 3000
instrument (Matteson). UV-Vis spectra were recorded on a Perkin-Elmer Lambda 40
spectrophotometer, and are reported as km” [nm]. CD spectra were recorded on a
J ASCO J -810 spectropolarimeter, equipped with a temperature controller (Neslab 111),

and are reported as A [nm] (At-zmax [1 mol'1 cm"]).
—:_TMS

(Trimethylsilyl)ethyne30 (2). Ethynyl magnesium bromide (0.5 M solution in THF, 30.0
mL, 0.015 mol, 1.07 equiv) was cooled to 0 °C and trimethylsilyl chloride (1.78 mL,
0.014 mol, 1 equiv) was added dropwise. The solution was allowed to warm to room
temperature and sit for 48 h without stirring. The reaction mixture was extracted with
cold saturated aqueous NH4CI and the organic layer dried over NaZSO4. The solution
was filtered to yield (trimethylsilyl)ethyne solution in THF (36.74 g, 73%). 1H NMR

(300 MHz, CDCl3, ppm): 8 1.84 (s, 1H), 0.08 (s, 9H).

103

HO

%
TMS

3-(Trimethylsilanylethynyl)benzaldehyde3 (3). (Trimethylsilyl)ethyne 2 (3.8 g, 0.0387
mol, 1.8 equiv), 3-bromobenzaldehyde (4 g, 0.0216 mol, 1 equiv), Pd(OAc)2 (56 mg,
0.25 mmol, 0.01 equiv), triphenylphosphine (112 mg, 0.43 mmol, 0.02 equiv), and dry,
deaerated triethylamine (21 mL) were combined in a flask and brought to quick reflux
under argon by placing the flask in an oil bath preheated to 100 °C. The solution was
refluxed for 10 h then cooled to room temperature. The reaction mixture was filtered and
the filtrate extracted with saturated NaHCO3 and CHzClz. The organic layers were
combined and the solvent removed to give a brown oil. This oil was purified by column
chromatography (5% EtOAc/hexanes) to give 3-(l-trimethylsilylethyne)benzaldehyde
(3.342 g, 77%). 1H NMR (300MHz, CDC13, ppm): 5 9.87 (s, 1H), 7.97 (s, 1H), 7.81 (d,

1H, 7.2 Hz), 7.72 (d, 1H, 7.8 Hz), 7.59 (m, 1H), 0.08 (s, 9H).

104

 

5-(3—TrimethylsilanylethynyIpheny|)-10,15,20-triphenylporphyrin (4). 3-
(Trimethylsilanylethynyl)benzaldehyde 3 (20.2 mg, 0.1 mmol, 1 equiv), pyrrole (26.8
mg, 0.4 mmol, 4 equiv), and benzaldehyde (31.8 mg, 0.3 mo], 3 equiv) were combined
in a flask with CHzClz (15 mL) and the solution purged with N2 for 15 minutes.
BF3°OEt2 (5.07 uL, 0.04 mmol, 0.1 equiv) was added, the reaction vessel shielded from
light, and stirred at room temperature overnight. p-Chloranil (98 mg, 0.4 mol, 4 equiv)
was added and the flask placed in a preheated oil bath at 45 0C. The solution was
refluxed for 2 h, then cooled to room temperature and the solvent volume reduced under
vacuum. The reaction mixture was purified by column chromatography (1:2,
CHzClzzhexanes) to give 5-(3-trimethylsilanylethynylphenyl)-10,15,20-
triphenylporphyrin (30.9 mg, 44%). 1H NMR (300 MHz, CDC13, ppm) 5 8.9 (m, 8H), 8.4

(s, 1H), 8.2 (m, 7H), 7.9 (m, 2H), 7.8 (m, 9H), 0.92 (s, 9H), -2.7 (s, 2H).

105

 

5-(3-Ethynylphenyl)-10,15,20-triphenylporphyrin (5). 5-(3-
Trimethylsilanylethynylphenyl)-10,15,20-triphenylporphyrin 4 (88.5 mg, 0.1355 mo], 1
equiv) was dissolved in THF (5 mL) with stirring. Tetrabutylammonium fluoride (1 M
solution in THF, 0.4 mL, 0.4 mol, 3 equiv) was added and the solution stirred at room
temperature for 24 h. The reaction mixture was extracted with H20 and CHzClz, the
organic layer dried with NazSO4, and the solvent removed under vacuum to give 5-(3-
ethynylphenyl)—10,15,20-triphenylporphyrin (77 mg, 98%). 1H NMR (300 MHz, CDCl3,
ppm) 5 8.9 (m, 8H), 8.4 (s, 1H), 8.2 (m, 7H), 7.9 (m, 2H), 7.8 (m, 9H), 3.2 (s, 1H), -2.8

(s, 2H).

106

 

S-(3-Ethynylphenyl)-10,15,20-triphenylzincporphyrin (6). 5-(3-Ethynylphenyl)-
10,15,20-triphenylporphyrin 5 (18 mg, 0.0282 mmol, 1 equiv) and Zn(OAc)2-2H20 (93
mg, 0.423 mmol, 15 equiv) were combined in CHzClz and stirred at room temperature
overnight. The color of the solution changed from dark purple to reddish purple during
the course of the reaction. The reaction mixture was purified with column
chromatography (CHzClz) and the purple band collected. The solvent was evaporated to
yield 5-(3-ethynylphenyl)-10,15,20-triphenylzincporphyrin (16.6 mg, 76.8%). lH NMR
(300 MHz, CDCl3, ppm) 5 8.9 (m, 8H), 8.4 (s, 1H), 8.2 (m, 7H), 7.9 (m, 2H), 7.8 (m,

9H), 3.2 (s, 1H).

107

 

 

Bis[5-(3-ethynylphenyl)-10,l5,20-triphenylzincporphyrin] (1). 5-(3-Ethynylphenyl)-
10,15,20-triphenylzincporphyrin 6 (90 mg, 0.1397 mol, 1 equiv) was combined with
Cu(OAc)20H20 (235 mg, 1.174 mmol, 8.4 equiv) and pyridine (12 mL). The solution
was stirred at 40 °C under an atmosphere of 02 for 24 h. The solution changed colors
from purple to blue-green over the course of the reaction. The reaction mixture was
extracted with H20 and NaHCO3 until the aqueous layer was clear. The organic layer
was dried with Na2SO4, filtered and dried to solid under vacuum to yield bis[5-(3-
ethynylphenyl)-10,l5,20-triphenylzincporphyrin] (79 mg, 88%). 1H NMR (300 MHz,

CDC13, ppm) 5 8.9 (m, 8H), 8.4 (s, 1H), 8.2 (m, 7H), 7.95 (m, 2H), 7,7 (m, 12H).

108

 

5-(4-Bromophenyl)-10,15,20-triphenylporphyrin (9). 4-Bromobenzaldehyde (119 mg,
0.643 mmol, 1 equiv), benzaldehyde (469 mg, 4.42 mmol, 6.9 equiv), and pyrrole (145
mg, 2.16 mmol, 3.36 equiv) was dissolved in CH2Cl2 (150 mL) with stirring. The
reaction mixture was stirred under N2 for 30 minutes, then BF3-OEt2 (24 uL, 0.193 mmol,
0.3 equiv) was added. The reaction mixture was stirred for 7 h at room temperature while
shielded from light. Excess p-chloranil added and the reaction stirred open to air at room
temperature for 2 h. The solution was then evaporated on a rotary evaporator and washed
with MeOH. The reaction mixture was purified with column chromatography (25%
CHCl3/hexanes) and the first band collected and dried under vacuum to give 5-(4-
bromophenyl)-10,15,20-triphenylporphyrin (500mg, 98%). 1H NMR (300 MHz, CDC13,
ppm): 5 8.83-8.81 (m, 8H), 8.23-8.20 (m, 6H), 8.08 (dd, 2H, 8.9/8.1Hz), 7.88 (dd, 2H,
8.4/7.8Hz), 7.75-7.73 (m, 9H), -2.82 (s, 2H). 13C NMR (300MHz, CDCl3, ppm): 5 141.8,

141.7, 135.5, 134.3, 134.2, 129.6, 127.5, 127.4, 126.4, 120.0.

109

3

0

Br

2-(4-Bromophenyl)-1,3-dioxolane31 (10). 4-Bromobenzaldehyde (12 g, 0.065 mol, 1
equiv), ethylene glycol (4.3 g, 0.07 mol, 1.07 equiv), and p-toluenesulfonic acid (2 mg,
0.01 mmol, 0.0001 equiv) dissolved in benzene (15 mL) and refluxed at 120 °C with a
Dean-Stark trap for 12 h. The reaction mixture extracted with brine and H20, the organic
layer dried with Na2SO4, and the solvent removed under vacuum to give 2-(4-
bromophenyl)-1,3-dioxolane (11.08 g, 74.5%). 1H NMR (300 MHz, CDC13, ppm) 5 7.6-

7.3 (dd, 4H), 5.75 (s, 1H), 4.1-3.9 (m, 4H).

SUBU3 nBu3

Fe

Bis(tributyltin)ferrocene32 (12). Hexane (5 mL) was added to ferrocene (1.01 g, 5.43
mmol, 1 equiv) with stirring to give a reddish-orange suspension. The mixture was
stirred under N2 for 10 minutes to help break up the ferrocene clumps. n-Butyl lithium
(1.7 M solution in hexane, 7.0 mL, 12 mmol, 2.2 equiv) was added quickly, then
TMEDA (1.8 mL, 11.9 mmol, 2.2 equiv) was added to the mixture dropwise over 30
minutes via a syringe pump. The solution was stirred at room temperature overnight to
give an orange solution. The reaction mixture was cooled to 0 °C and SnBu3Cl (3.7 mL,
14 mmol, 2.5 equiv) was added dropwise over 1 h via a syringe pump. The solution was
allowed to warm to room temperature, stirred for 1 h, and poured into ice water (100

mL). The reaction mixture was extracted with water and the aqueous layer extracted with

110

hexane until the organic washings were nearly clear. The organic layers were combined,
dried with Na2SO4, and the solvent removed under vacuum to give a dark amber oil.
Ferrocene and (tributyltin)ferrocene were distilled off to yield pure
bis(tributyltin)ferrocene (2.6 g, 62.7%). 1H NMR (300 MHz, CDCl3, ppm): 5 4.23 (m,

4H), 3.96 (m, 4H), 1.9-1.0 (m, 54H).

3(0H)2 3(0H)2

Fe

Ferrocene bis(boronic acid)15 (13). TMEDA was distilled prior to use. To a stirred
suspension of ferrocene (10.1 g, 54.2 mol, 1 equiv) in Et2O (200 mL) at room
temperature was added a mixture of n-butyl lithium (50.2 mL, 120 mmol, 2.22 equiv),
TMEDA (18.2 mL, 120 mmol, 2.22 equiv), and B20 (100 mL). The solution turned
light orange and was stirred at room temperature overnight. The reaction mixture was
cooled to 0 °C and added dropwise over 1 h to a -78 °C solution of tributoxyborate (34.0
g, 131 mmol, 2.41 equiv) in B20 (200 mL). The resulting solution was stirred at -78 °C
for 1 h then allowed to warm to room temperature. After 14 h, the reaction mixture was
hydrolyzed with 10% aqueous KOH (150 mL) to give a light tan solution. The organic
layer was extracted with more 10% aqueous KOH (400 mL) and the aqueous layers
combined and cooled in an ice bath. The aqueous layers then acidified with 10% H2804
to pH 2 and filtered to give ferrocene bis(boronic acid) (6.050g, 36%). 1H NMR (300

MHz, DMSO, ppm): 5 7.46 (s, 4H), 4.33 (s, 4H), 4.16 (s, 4H).

lll

 

Ferrocene bis[tetraphenylporphyrin] .(14). 5-(4-Bromophenyl)-10,15,20-
triphenylporphyrin 9 (107 mg, 0.154 mol, 2 equiv), ferrocene bisboronic acid 13 (21
mg, 0.077 mol, 1 equiv), and Pd(PPh3)4 (0.8mg, 6.9x104 mmol, 0.009 equiv) were
dissolved in freshly distilled toluene (2 mL). Na2CO3 (2 M aqueous solution, 2 mL) was
added and the reaction was refluxed for 5 days. After cooling, toluene was added and the
reaction mixture was extracted with water. The organic layers were combined, dried with
Na2SO4, and the solvent removed under vacuum to give a purple solid. The residue was
purified by column chromatography (15% CH2Cl2zhexane) and the fourth band collected
(starting materials preceded the desired compound). The solvent was removed under
vacuum to give ferrocene bis[tetraphenylporphyrin] (80 mg, 74%). 1H NMR (300 MHz,
CDC13, ppm): 5 8.88-8.84 (m, 16H), 8.22 (d, 12H, 5.5Hz), 8.07 (d, 4H, 5.4Hz), 7.87 (m,

4H), 7.76 (m, 20H), 7.37 (m, 8H), —2.8 (m, 4H).

112

 

Ferrocene zinc bis(tetraphenylporphyrin) (11). Ferrocene bis(tetraphenylporphyrin)
(80 mg, 0.057 mmol, 1 eq) was dissolved in CH2C12 (10 mL) with stirring. Zn(OAc)2
(208 mg, 1.13 mmol, 20 eq) was added and the solution was stirred at room temperature
for 3 d. The reaction was filtered through celite and washed with CHC13 to give 52 mg
(60% yield) of product. 1H NMR (300 MHz, CDCl3, ppm): 5 8.92-8.88 (m, 16H), 8.19
(m, 12H), 8.04 (m, 4H), 7.86 (d, 4H, 8.1Hz), 7.74 (d, 20H, 6.9Hz), 7.33 (s, 8H).

0H

02N NO2

4-Methyl-3,5-dinitrobenzylaleohol20 (15). 4-Methyl-3,5-dinitrobenzoic acid (5 g, 22.1
mmol, 1 equiv) was dissolved in freshly distilled THF (11 mL) and BHg-THF (1M

solution in THF, 29 mL, 28.7 mmol) was added slowly under nitrogen with rapid

113

stirring, generating a large evolution of H2. The reaction mixture was stirred vigorously
overnight then water was carefully added to destroy excess BH3. Saturated brine was
added to the reaction mixture and the organic layer was separated. The aqueous layer
was washed with 320 (3x25mL) and all organic fractions were combined, dried with
Na2SO4, and filtered. The solvent was removed under vacuum to give 4-methyl-3,5-
dinitrobenzylalcohol (3.771 g, 80.4%). 1H NMR (300 MHz, CDC13, ppm): 5 8.02 (s,
2H), 4.70 (s, 2H), 2.48 (s, 3H).

HO

OZN N02

4-Methyl-3,5-dinitrobenzaldehyde33 (16). 4-Methyl-3,5-dinitrobenzylalcohol (3.771 g,
17.8 mmol, 1 equiv) was dissolved in freshly distilled CH2Cl2 (100 mL) with stirring.
Pyridinium chlorochromate (PCC) (5.76 g, 26.7 mmol, 1.5 equiv) was added and the
solution was stirred overnight at room temperature. Diethyl ether was added and the
solution filtered and the filtrate was evaporated to give a light brown solid. The residue
was purified by column chromatography (CHCl3) and the first band was collected and
dried to give an off-white solid. This solid was recrystallized by dissolving it in CHC13
and gradually adding hexanes while the solution was stirred over a steam bath to give 4-
methyl-3,5-dinitrobenzaldehyde (1.884 g, 50%). 1H NMR (300 MHz, CDC13, ppm): 5

10.05 (s, 1H), 8.45 (s, 2H), 2.65 (s, 3H).

114

     

\NHHN /

(4-Methyl-3,5-dinitrophenyl)dipyrrole (18). 4-Methyl-3,5—dinitrobenzaldehyde (256
mg, 1.22 mmol, 1 equiv) was dissolved in pyrrole (2.5 g, 37.3 mmol, 30 equiv) with
stirring. The reaction mixture was stirred under nitrogen for 5 rrrinutes then
trifluoroacetic acid (9.2 uL, 0.119 mmol, 0.1 equiv) was added. The reaction mixture
was stirred under nitrogen for another 5 minutes, then quenched with NaOH (0.2 M) and
extracted with EtOAc and H20. The organic layers were combined, dried with Na2SO4,
and the solvent was removed under vacuum to give a light brown oil. Bulb-to-bulb
distillation at 200 °C at 0.5 torr gave (4-methyl-3,S-dinitrophenyl)dipyrrole (259 mg,
49.4%) as a yellow powder. 1H NMR (300 MHz, CDCl3, ppm): 5 8.05 (s, 2H), 7.79 (s,
2H), 6.74 (s, 2H), 6.17 (q, 2H), 5.86 (s, 2H), 5.56 (s, 1H), 2.52 (s, 3H); 13C NMR (300

MHz, CDCl3. ppm): 5 151.8, 143.7, 129.8, 127.5, 125.8, 118.4, 109.0, 108.1, 42.9, 14.4.

    

\NH HN

2,2’-(Phenylmethylene)bispyrrole (Phenyldipyrrole, 19). Benzaldehyde (253 mg, 2.36
mmol) was dissolved in pyrrole (3.96 g, 58.9 mmol, 25 equiv) with stirring. The mixture
was stirred under nitrogen for 5 minutes, then trifluoroacetic acid (18.2 uL, 0.236 mmol,

0.1 equiv) was added. The reaction mixture was stirred under nitrogen for another 5

115

minutes, then quenched with NaOH (0.2 M) and extracted with EtOAc and H20. The
organic layers were combined, dried with Na2SO4, and the solvent removed under
vacuum to give a light brown oil. Bulb-to-bulb distillation at 230 °C at 0.5 torr gave
phenyldipyrrole (259 mg, 49.4%). 1H NMR (300 MHz, CDCl;;, ppm): 5 7.86 (s, 2H),

7.35 (m, 3H), 7.25 (d, 2H), 6.69 (s, 2H), 6.22 (m, 2H), 5.96 (s, 2H), 5.47 (s, 1H).

 

5-(4-Methyl-3,5-dinitrophenyl)-10,15,20-triphenylporphyrin (17). (4-Methyl-3,5-
dinitrophenyl)dipyrrole (50 mg, 0.15 mol, 1 equiv), benzaldehyde (48 mg, 0.45 mol,
3 equiv) and pyrrole (20.2 mg, 0.3 mmol, 2 equiv) were dissolved in CH2C12 (60 mL) and
stirred at room temperature under N2 for 30 min. The reaction vessel was shielded from
light and BF3°OEt2 (5.7 pL, 0.045 mmol, 0.3 equiv) was added. The reaction mixture
was stirred at rt for 12 h, during which time the solution became dark red. p-Chloranil
(75 mg, 0.30 mmol, _2 equiv) was added to the reaction mixture and the solution stirred
for 15 min. The solvent was removed under vacuum and the residue was purified by
column chromatography (50% CH2Cl2/hexanes) to give 5-(4-methyl-3,5-dinitrophenyl)-

10,15,20—triphenylporphyrin (18.8 mg, 17%). 1H NMR (300 MHz, CDCl3, ppm): 5 8.95

116

(d, 2H, 5.1Hz), 8.86 (d, 6H, 6.6Hz), 8.76 (d, 2H, 4.8Hz), 8.21 (d, 6H, 7.2Hz), 7.78 (m,

9H), 2.93 (s, 3H), -2.8 (s, 2H). Mass (DCI): found 718.3, expected 718.7.

5-(4-Methyl-3,5-dinitrophenyl)-10,15,20-triphenylporphyrin (l7) (one-pot
procedure). 4-Methyl-3,S-dinitrobenzaldehyde (50 mg, 0.238 mmol 1 equiv) was
dissolved in pyrrole (244 mg, 3.637 mmol, 15.3 equiv). After stirring the solution for 5
minutes under nitrogen, trifluoroacetic acid (1 uL, 0.013 mmol, 0.05 equiv) was added
and the reaction mixture stirred a further 5 minutes under nitrogen. Then the solution
was transferred to a 2 L flask and diluted with CH2C12 (1 L). Benzaldehyde (254 mg,
2.34 mmol, 10 equiv) was added and the solution stirred for 30 minutes under nitrogen.
The reaction vessel was shielded from light and BF3-OEt2 (3 uL, 0.024 mmol, 0.1 equiv)
was added. The solution was stirred at room temperature for 3.5 h. p-Chloranil (1.5 g,
6.1 mmol, 25.6 equiv) was added and then the solution stirred for 30 minutes. The
solvent was removed under vacuum and the residue was purified by column
chromatography (1:3 CHCl3/hexanes). The second band was collected and the solvent
removed under vacuum to give 5-(4-methyl-3,5-dinitrophenyl)-10,15,20-
triphenylporphyrin (27 mg, 15.8%). Analytical results matched that of the product from

the multistep reaction.

117

 

5-(4-Methyl-3,5-diaminophenyl)-10,15,20-triphenylporphyrin (20). 5-(4-Methyl-3,5-
dinitrophenyl)-10,15,20—triphenylporphyrin 17 (994 mg, 1.383 mol, 1 equiv) was
dissolved in CH2Cl2 (25 mL) with stirring. Concentrated HCl (25 mL) was added and the
solution turned green. SnCl2-2H2O (1.997 g, 8.851 mmol, 6.4 equiv) was added and the
closed solution was stirred at room temperature overnight. Water was added and the
solution very slowly basified with aqueous NaHCO3. The solution was extracted with
CH2Cl2 and the organic layers collected and dried with Na2SO4. The solvent was
removed under vacuum to give 5-(4-methyl-3,5-diaminophenyl)-10,15,20-
triphenylporphyrin (67 mg, 7.4%) as a purple solid. 1H NMR (300 MHz, CDCl3, ppm):
5 8.95 ppm (d, 2H, 5.1Hz), 8.86 (d, 6H, 6.6Hz), 8.76 (d, 2H, 4.8Hz), 8.21 (d, 6H, 7.2Hz),

7.78 (m, 9H), 2.93 (s, 3H), -2.8 (s, 2H). Mass (DCI): found 658.3, expected 658.28.

118

 

 

Polymer (21). To a stirred solution of 5-(4-methyl-3,5-diaminophenyl)-10,15,20-
triphenylporphyrin 20 (67 mg, 0.102 mmol, 1 equiv) and triethylamine (130 mL, 0.94
mmol, 9.2 equiv) in CH2CI2 (20 mL) was added triphosgene (15 mg, 0.051 mmol, 0.5
equiv) in CH2CI2 (5 mL). The reaction mixture was stirred under N2 for 16 h. The
solvent was removed under vacuum and the residue purified by column chromatography
(25% CHCl3/hexanes) to give porphyrin polymer (11 mg). 1H NMR (300 MHz, CDCl3,

ppm): 5 8.81 ppm (b), 8.17 (b), 7.73 (b), 2.0 (s, 3H) -2.8 (s, 2H).

119

 

Zinc polymer (22). Polymer 21 (11mg, 0.008 mmol, 1 equiv) was dissolved in CH2C12
(2 mL) and Zn(OAc)2 (15 mg, 0.082 mmol, 10 equiv) was added. The reaction mixture
was stirred overnight at room temperature and then filtered through a celite plug and

washed with CH2Cl2. The solvent from the filtrate was removed under vacuum to give
zinc polymer (16 mg, 99%). 1H NMR (300 MHz, CDCl3, ppm): 5 8.9 (b), 8.2 (b), 7.7

(b), 2.0 (s, 3H).

120

References:

1. Yang, Q. The Use of Exciton Coupled Circular Dichroism for the Absolute
Stereochemical Determination of Chiral Carboxylic Acids. Ph. D. Dissertation, Michigan
State University, East Lansing, 2004.

2. Nakash, M.; Clyde-Watson, Z.; Feeder, N .; Teat, S. J .; Sanders, J. K. M., Chemistry:
A European Journal 2000, 6, 2112.

3. Austin, W. B.; Bilow, N.; Kelleghan, W. J .; Lau, K. S. Y., J. Org. Chem. 1981, 46,
2280.

4. Twyman, L. J .; Sanders, J. K. M., Tetrahedron Lett. 1999, 40, 6681.

5. Lindsey, J. S.; Prathapan, S.; Johnson, T. E.; Wagner, R. W., Tetrahedron 1994, 50,
8941.

6. Itaya, T.; Mizutani, A.; Watanabe, N ., Chem. Pharm. Bull. 1989, 37, 1221.
7. Knobloch, K.; Eberbach, W., Org. Lett. 2000, 2, 1117.
8. Lan, Y. F; Hammond, G. B., J. Org. Chem. 2000, 65, 4217.

9. Gryko, D. T.; Zhao, F.; Yasseri, A. A.; Roth, K. M.; Bocian, D. F.; Kuhr, W. G.;
Lindsey, J. S., J. Org. Chem. 2000, 65, 7356.

10. Cunningham, A. F., J. Amer. Chem. Soc. 1991, 113, 4864.
11. Guillaneux, D.; Kagan, H. B., J. Org. Chem. 1995, 60, 2502.

12. Liu, C. M.; Chen, B. H.; Liu, W. Y.; Wu, X. L.; Ma, Y. X., J. Organomet. Chem.
2000, 598, 348.

13. Liu, C. M.; Guo, Y. L.; Wu, X. L.; Liang, Y. M.; Ma, Y. X., J. Organomet. Chem.
2000, 612, 172.

14. Shi, X. L.; Amin, S. R.; Liebeskind, L. S., J. Org. Chem. 2000, 65, 1650.
15. Knapp, R.; Rehahn, M., J. Organomet. Chem. 1993, 452, 235.

16. Yagi, S.; Yonekura, 1.; Awakura, M.; Ezoe, M.; Takagishi, T., Chem. Comm. 2001,
557.

17. Collman, J. P.; Gagne, R. R.; Reed, C. A.; Halbert, T. R.; Lang, G.; Robinson, W. T.,
J. Amer. Chem. Soc. 1975, 97, 1427.

18. Rose, E.; Kossanyi, A.; Quelquejeu, M.; Soleilhavoup, M.; Duwavran, F.; Bernard,
N.; Lecas, A., J. Amer. Chem. Soc. 1996, 118, 1567.

121

19. Landrum, J. T.; Grimmett, D.; Haller, K. J .; Scheidt, W. R.; Reed, C. A., J. Amer.
Chem. Soc. 1981, 103, 2640.

20. Yoon, N. M.; Pak, C. S.; Brown, H. C.; Krishnamurthy, S.; Stocky, T. P., J. Org.
Chem. 1973, 38, 2786.

21. Lee, C. H.; Lindsey, J. S., Tetrahedron 1994, 50, 11427.

22. Littler, B. J.; Miller, M. A.; Hung, C. H.; Wagner, R. W.; O'Shea, D. F.; Boyle, P.
D.; Lindsey, J. S., J. Org. Chem. 1999, 64, 1391.

23. Li, F. R.; Yang, K. X.; Tyhonas, J. S.; MacCrum, K. A.; Lindsey, J. S., Tetrahedron
1997, 53, 12339.

24. Gowda, D. C.; Mahesh, B., Synth. Comm. 2000, 30, 3639.

25. Arai, T.; Inudo, M.; Ishimatsu, T.; Akamatsu, C.; Tokusaki, Y.; Sasaki, T.; Nishino,
N .,J. Org. Chem. 2003, 68, 5540.

26. Gosztola, D.; Yamada, H.; Wasielewski, M. R., J. Amer. Chem. Soc. 1995, 117,
2041.

27. Sivakumar, M., Michgan State University, unpublished results, 2002.

28. Collman, J. P.; Wang, Z.; Straumanis, A., J. Org. Chem. 1998, 63, 2424.

29. Girolami, G. S.; Hein, C. L.; Suslick, K. S., Angew. Chem. Int]. Ed. 1996, 35, 1223.
30. Penney, J. M.; Miller, J. A., Tetrahedron Letters 2004, 45, 4989.

31. Lee, Y.; Silverrnan, R. B., Org. Lett. 2000, 2, 303.

32. Cunningham, A. F.,J. Amer. Chem. Soc. 1991, 113, 4864.

33. Remy, D. C.; King, S. W.; Cochran, D.; Springer, J. P.; Hirshfield, J ., J. Org. Chem.
1985, 50, 4120.

122

Chapter 4
Use of Ru(bpy)2Cl2 for Determination of Absolute Stereochemistry of Erythro 1,2-

Aminoalcohols and Diols

Erythro diols have long posed a problem in determination of their absolute
stereochemistry. For their counterparts, threo diols, it is a simpler matter to determine
the absolute stereochemistry, and there have been many methods developed to facilitate
the process. Use of Mosher’s ester derivatives or variations allows for the determination
of stereochemistry through 1H NMR.1 However, both enantiomers of the threo diol must
be derivatized in order to ensure correct assignment of the absolute stereochemistry.
Exciton-coupled circular dichroism has also been used over the past several decades to
determine the absolute stereochemistry of threo diols.2‘1 Due to the arrangement of the
alcohols in threo diols, it is evident from a Newman projection that there should be one
favored low energy conformation. For example, consider threo-2,3-dihydroxypentane.

f“ . r‘ (a .3
HO OH HO H OH H
H

U

RE H )H J
O'CL/

J H

 

9m
x

‘0 H 22> negative ECCDI
H -

 

 

Figure 4-1. Newman projections of low energy conformers of (2R, 3R)-2,3-dihydroxypentane. When
derivatized with chromophores (X), the expected lowest energy conformer would lead to a negative
ECCD bisignate curve.

123

 

As can be seen in Figure 4-1, the (2R, 3R) isomer has two conformers with three gauche
interactions (indicated by the blue arrows), and only one conformer where there are two
gauche interactions, thus this conformer would be expected to be the predominant low-
energy conformation. If the hydroxyl groups are derivatized with an appropriate
chromophore, interaction between the chromophores would lead to an expected negative
ECCD bisignate curve. Comparing the enantiomer of the previously analyzed molecule,
the (ZS, 3S) isomer also has only one conformer where there are two gauche interactions
(Figure 4-2). However, since the hydroxyl groups have the opposite configurations as

before, this conformation would be expected to lead to a positive ECCD bisignate curve

HQ H H at! HF HOW /E$:H
K E H C‘ H L- H

L. H \JOH

 

 

2) positive ECCD I

 

Figure 4-2. Newman projections of low energy conformers of (ZS, 3S)-2,3-dihydroxypentane. When
derivatized with chromophores (X), the expected lowest energy conformer would lead to a positive
ECCD bisignate curve.

when derivatized with chromophores. As shown in these examples, threo 1,2-diols can
be analyzed by ECCD to provide an absolute stereochemical determination.

However, similar analysis of erythro diols shows a fatal flaw. As before, a simple
comparison of the Newman projections of ewthro—2,3-dihydroxypentane gives three
possible conformations. The (2R, 3S) isomer shown in Figure 4-3 has one conformation

with two gauche interactions and two conformations with three gauche interactions.

124

However, the expected lowest energy conformation (the conformation with only two

gauche interactions) would not be expected to show ECCD, since the chromophores

:1 m. r”)

HH_H OH H HO
.. D . J

H

J ottJ H

 

 

 

fl derivatization with chromophores

 

 

 

positive ECCD no ECCD I negative ECCD

 

Figure 4-3. Newman projections of low energy conformers of (2R, 3S)-2,3-dihydroxypentane. When
derivatized with chromophores, no ECCD is expected to be observed.

(N kw
HHE/ffiort H H HPHH
L .. H L H
H L,

 

0“ L.
U derivatization with chromophores fl
6

 

 

 

 

 

positive ECCD no ECCD I negative ECCD

 

Figure 4-4. Newman projections of low energy conformers of (ZS, 3R)-2,3-dihydroxypentane. When
derivatized with chromophores, no ECCD is expected to be observed.

would be in an antiperiplanar position. The other two conformations, which would both
be expected to show ECCD when derivatized, are expected to be similar in energy due to
the same number of gauche interactions. Thus, the ECCD signals (one negative, one
positive) expected from the two higher-energy conformations would cancel out leading to
no observed ECCD. This can be demonstrated to be also true for the enantiomer, the (2S,
3R) isomer, as shown in Figure 4-4. As before, there is one conformation with only two

gauche interactions, but it would not be expected to show ECCD due to the orientation of

125

the chromophores (antiperiplanar). The other two conformations should be similar in
energy and the ECCD bisignate curves will cancel out leading to no observed ECCD.

In light of this, we hoped to design a system that would allow for the simple
determination of absolute stereochemistry for erythro diols using ECCD. Initial
consideration of the problems inherent with the erythro diols led us to believe that a
system in which the diol was in a locked conformation provided the best opportunity for
facile analysis of possible conformations and would allow us to predict with some
certainty the position of the chromophores. Since it is logical that in order to determine
the exact conformation it is beneficial to rigidify the system, we wanted to devise a
system in which the diol becomes part of a six-membered ring, with two chromophores
attached to the ring (Figure 4-5). If we know the most stable conformation of the system

HO OH XIOIR
x o R,

R, R

 

Figure 4-5. Transformation of erythro diols into six-membered rings functionalized with chromophores

(X).
(presumably a lowest energy chair formation), we should be able to predict the ECCD for
the possible erythro 1,2-diol stereochemistries. Hence, we would be able to derivatize a
small amount of diol sample, subject it to ECCD, and from the observed sign determine
the absolute stereochemistry of the diol. To this end, we initially considered converting
the erythro diols into dioxanes.
4.1. Conversion of erythro-1,2-diols into dioxanes

To determine what stereochemistry was necessary for the six-membered ring
system, we studied all of the Newman projections for the possible conformations. We

started by looking at 2,3-pentanediol as the erythro diol and an (R,R) bis-chromophore

126

(Figure 4—6). Both possible stereochemistries [(ZR, 3S)-2,3-pentanediol and (2S, 3R)-2,3-

pentanediol] for the erythro diol were examined as Newman projections after

(2R, 3S)-2,3-pentanediol (ZS, 3R)-2,3-pentanediol
H H H 6
9063101631“ eCfiarC’EMpE‘
X 0 Me X 0 H
H Et H H
ring flip I
X H X t
”$0251” “299%“
H 0 Et H O H
X H

X Me

Figure 4-6. Newman projections of the product formed from (2R, 3S)-2,3-pentanediol or (2S, 3R)-2,3-
pentanediol and an (R, R) bis-chromophore.

 

 

 

 

 

 

derivatizations with the bis-chromophore to form a dioxane (Figure 4-6). When these
two are combined to give a six-membered ring, it becomes apparent that both systems
would predict the same ECCD sign when the chromophores are in the equatorial position
(when the two chromophores are both axial, there would be no ECCD sign expected
since they are antiperiplanar). So a chiral bis-chromophore of this stereochemistry [either
(R, R) or (S, S)] will not work.

We then looked at the conformations for a six-membered ring obtained from a
mesa chromophoric system. In this case, there are several possible conformations that
need to be considered. Using the same diol as before—including both possible
stereochemistries—we compared the various Newman projections using an (R, S) mesa
bis—chromophore, including the two different configurations depending on whether the
diol reacts with the bis-chromophore from the top or the bottom. Here, the (2R, 3S) diol

would give either a positive or negative ECCD sign depending on which conformation is

127

 

 

H H H ‘ X X H H
H$O$ 162.76 kcal/moi $0$ 15034 kcal/moi
O . .
3:} 0 5, M9 positive ECCD H H Et Me pOSItIve ECCD
1L ring flip 1L ring flip
K}
H H 0 H H X 0 H H
a 160.68 kcal/moi 153.33 kcal/moi
X 0 Et H 0 Et
\JX Me negative ECCD H Me negative ECCD

 

 

 

 

 

 

Figure 4-7. Newman projections of the product formed from (2R, 3S)-2,3-pentanediol and an (R, S)
mesa bis-chromophore (X) indicating an expected negative ECCD bisignate curve.

 

 

 

 

 

 

Xfiolfia 160.68 kcal/moi H$0$Et 158.36 kcal/moi
O C
H H H H positive ECCD (f X 0 H H posrtive ECCD

 

 

1L ring flip 1L ring flip

C X t Me H H t M
$0 162.76 kcal/moi goé ° 160.84 kcal/moi
H H O H H negativeECCD 0 O H H negativeECCD

Figure 4-8. Newman projections of the product formed from (2S, 3R)-2,3-pentanediol and an (R, S) bis-
chromophore (X) indicating an expected positive ECCD bisignate curve.

considered. Since it was unclear as to which conformation would be favored, analysis via
molecular modeling was performed to see if there was enough difference in energy to
favor one conformation over another (Figure 4-7).5 In both configurations, the lowest
energy structure by approximately 2 kcal/mol is the one in which the chromophores are
oriented counterclockwise from front to back. This can be explained due to the
preference for the larger group of the diol (in this case, ethyl) to be equatorial, since in
both cases the chromophores (presumably larger than the substituents of the diol) have
one chromophore oriented axial and one equatorial. Thus, based on molecular mechanics

calculations (Spartan v.5.0), the (2R,3S) diol would be expected to give a negative ECCD

128

sign. We then performed the same analysis on the (25,3R) diol, again with an (R,S) bis-
chromophore (Figure 4-8). Here the lowest energy conformers—again by approximately
2 kcal/mol—are oriented clockwise from front to back, leading to the expectation of a
positive ECCD sign. As such it appears that converting erythro 1,2-diols into dioxanes
substituted with mesa chromophores could be a valid way of determining the absolute
stereochemistry of erythro 1,2-diols.

At this point, it became necessary to make a starting chromophoric system which
would be simple to use as a derivatizing agent for diols. Although porphyrins are
excellent chromophores, they are not necessarily a good initial chromophore as it could
be difficult to synthesize the needed compounds, so we chose to start with (R,S)-1,2-
dibromo-l,2-diphenylethane, where the phenyl groups would act as the requisite

chromophores for ECCD analysis. However, a substitution reaction of a diol displacing

 

 

Cl
N02 Br Br
KOH : \ O Brz, CH3N02
EtOH/H20 (4:1) 0 . 0' O 0
N02 1 h. 66% 1 OZN 2 N02

Scheme 4-1. Synthesis of meso-l,2-dibromo-1,2-(4,4’-dinitrophenyl)ethane 2 from 4-nitrobenzyl
chloride.

the bromines is more likely to go through an SN] reaction since the carbocation will be
stabilized by resonance with the aromatic rings. To destabilize this reaction pathway, we
opted to use para substituted phenyl rings where the substituents are strongly electron
withdrawing nitro groups, as shown in 1 (Scheme 4-1). Synthesis of this molecule could
be achieved through bromination of trans-4,4’-dinitrostilbene, which is easily

synthesized from 4-nitrobenzyl chloride.6

129

Synthesis of trans-4,4’-dinitrostilbene proceeded as reported in literature with
66% yield on a 4g scale.6 Allylic bromination reactions are typically performed in the
darkf’9 so the first time the reaction was run the reaction flask was shielded from light
and the concentration of the reaction was controlled to favor the mesa product. However,
only 8% yield of the desired product was obtained, with the remaining material being
decomposed starting material and small amounts of chiral bromination products. The
reaction was then run in the darkroom to ensure absence of ambient light, but this
reaction again gave chiral bromination products. We then found a literature procedure
which described the bromination proceeding entirely to the mesa product if the reaction
was run under radical conditions; i.e., light.9 The reaction was run in ambient light and a
2:1 mixture of starting material and desired mesa product was obtained. We attempted to
crystallize the product out but were unsuccessful. The reaction was run again using a
superbright light. The solution went from orange-yellow to pale yellow as the bromine
was consumed over a period of five days. After workup, the product was recovered
through filtration in 20% yield. The solvent from the filtrate was removed under reduced
pressure and the solid obtained was washed with EtOAc and filtered to obtain an
additional 10% pure mesa-1,2-dibromo-l,2-(4,4’-dinitrophenyl)ethane.

After obtaining the desired bis-chromophore starting material, the next step was

derivatization of a diol. The first substitution reaction performed was with racemic p-
Br Br
EOIOH + "3°” ONH — Q ~02
2 __
0 OH O O THF: rt. 2 h
02N N02

2 3
Scheme 4-2. Synthesis of the unexpected product bis-(4-nitrophenyl)acetylene 3.

 

130

dioxane-2,3-diol (Scheme 4-2). After dissolving the dial in THF, sodium hydroxide was
added dropwise and then the bis-chromophore was added. Although the bis-
chromophore was not very soluble in THF, the reaction changed color gradually from
colorless to pale yellow, at which point the reaction was stopped. No substitution product
was obtained, and the major product by 1H NMR was the double elimination product of
the bis-chromophore, bis-(4-nitrophenyl)acetylene. We then decided to try a variety of
reaction conditions using racemic trans-1,2-cyclohexanediol to try to determine optimal
conditions for the substitution. Since 8N2 is favored with polar aprotic solvents, we set

up three reactions using dimethylsulfoxide (DMSO), acetonitrile (ACN), and

Br Br

DMF, DMSO, or ACN

OH
0’ + = No reaction
"IOH l1, 2 h
02N N02

(racemic) 2

 

Scheme 4-3. Attempted derivatizations of trans-l,2-cyclohexanediol in various solvents.

dimethylforrnamide (DMF). All of these reactions used no base; we added the diol and
the bis-chromophore and stirred them together at room temperature (Scheme 4-3). Both
compounds dissolved easily in DMSO and DMF, but there was considerable cloudiness
in the ACN reaction mixture. The DMF reaction mixture showed no reaction by NMR
after 2 hours of stirring and workup. DMSO was difficult to remove from the reaction
mixture but the reaction contained only starting material from TLC. The ACN reaction
mixture was rotavapped and 1H NMR taken in DMSO. The reaction showed small shifts
in 1H NMR, but it was unclear if the peaks were from the individual starting materials or
from the product as there should not be much difference in the chemical shifts. ‘3 C NMR

would be expected to be more conclusive, as there should be a significant shift in the

131

carbon that was attached to the bromine (normally appearing around ~53ppm) which
becomes bonded to an oxygen as an ether (normally appearing around ~85ppm). A total
of eight peaks were observed in the 13C NMR of the product in DMSO. The carbons in
question in the product have peaks at 73.830 ppm (C-OH) and 51.930 ppm (C-Br), as
compared to 73.875 ppm (C-OH) and 51.826 ppm (C-Br) in the starting materials. Thus
it was apparent that none of the desired product had been formed and no reaction had
occurred.

In a reference on sources of nucleophilic oxygen, Hutchins and Taffer report the
use of HMPA or N-methyl—Z-pyrrolidone (NMP) to increase the nucleophilicity of

1° However, in their chiral

alcohols for use in substitution reactions to form ethers.
examples they do note considerable racemization (80% inversion for NMP, 70%
inversion for HMPA), even with a strong base to neutralize the reaction. As our previous

attempts to synthesize dioxanes using diols had failed, we decide to increase the

nucleophilicity of the oxygens using HMPA (Scheme 4-4). Initially, we tried to run the

Br Br

0” ACN/HMPA
+ O O i No reaction
OH m1zh
OzN N02

(racemic) 2

 

Scheme 4—4. Attempted derivatization of trans-1 ,2-cyclohexanediol using HMPA in ACN.

reaction in HMPA alone, but the bis-chromophore was not soluble so ACN was added to
improve solubility. Workup and a pipet column to remove the HMPA revealed no
reaction had taken place.

Although the dioxanes seemed like a promising solution to the problem of

absolute stereochemical determination for erythro diols, the difficulties in synthesis led

132

us to believe there might be other methods that would be more amenable to diols without
resorting to unusual or harsh synthetic conditions. Development of the dioxane method is
still an option, but we would prefer to have a method that is facile and rapid for general
use. We began exploring other possibilities, and one of the most captivating was the idea
of using metal complexes as achiral hosts for our chiral diols.
4.2. Consideration of cis-Ru(bpy)2Clz as a chromophoric host

As the formation of dioxanes from erythro diols seemed unlikely to yield a viable
method of creating the necessary chromophoric system, we considered the possibility of
using a metal center to coordinate the erythro diols. A metal center with two or more
chromophores could be forced to adopt a chiral configuration upon coordination to a
chiral diol. It was necessary that the metal in question be well-characterized with a
known ability to form chiral complexes. The cis-Ru(bpy)2X2 series fit these
qualifications well, where X may be any other ligand. cis-Ru(bpy)2X2 is known to form

two enantiomers, the A isomer and the A isomer (Figure 4-9).11 At room temperature in a

 

Figure 4-9. A and A forms of cis-Ru(bpy)2Clz.

compound with no differentiation between the X2 ligands (e.g., C12), cis-Ru(bpy)2X2

exists as a racemic mixture, and a single pure isomer (A or A) will racemize over time in

12

solution. Both forms are CD active and show opposite signals. However, it has been

133

shown that one isomer can be preferred with the addition of a ligand that can induce
chirality, such as chiral sulfoxide ligands.l3 With this precedent, we believed it might be
possible to do a similar type of conformational preference with a chiral diol. If we could
add a chiral erythro dial to the RubpyZClz, displacing the chlorines, we could potentially
force the Rubpyzxz to exist in one form, or if not, perhaps increase the enantiomeric
excess of one form versus the other. There are two ways we envisioned this occurring: 1.
The incoming chiral guest binds preferentially to one form or the other (A and A). As the
guest binds to, for example, A—RubpyzClz, that ruthenium is locked into the A
conformation. All of the remaining mixture of cis-Ru(bpy)2C12 (A > A) racemizes. As
this process continues, a large excess of ruthenium with the chiral guest (in this example
as the A form) accumulates and this will be evident in the ECCD (Figure 4-10). 2. The

incoming chiral guest changes the stereochemistry at the ruthenium center to fit with

 

 

Figure 4-10. Possible method of stereodifi‘erentiation where the chiral erythro diol binds preferentially
to A-cis-Ru(bpy)2C12.

itself. For example, if the lowest energy conformation the ruthenium plus guest would
adopt would be A (dependent on the conformation of the chiral diol), the incoming guest
can bind to any ruthenium molecule, but when it binds the conformation of the bipyridine

groups around the ruthenium will change to, or maintain, the A conformation (Figure 4-

134

 

11). With these possibilities in mind, we began trying to form complexes of chiral diols
and cis-Ru(bpy)2C12.

Molecular modeling in Spartan5 indicates that there would be significant strain in
one conformation versus the other upon binding to a chiral guest. As seen in Figure 4-12,
Ru(bpy)2Y [where Y is (2R,3S)-pentanediol] shows different levels of strain depending

on which conformation of Ru(bpy)2Clz it complexes with. Notice in particular the strain

 

HO OH
H M
L M : L
\
H
H

 

Figure 4-11. Transfer of chirality from the erythro dial to the ruthenium complex, changing the initial
stereochemistry of the ruthenium to match the diol.

in the bipyridines in the structure on the right. In this conformation, the ethyl and methyl
groups are pointing towards the bipyridine rings, causing them to buckle to try to avoid
steric interactions. In the conformation on the left, both groups are pointing away from

the bipyridine rings, leading to a less strained system. However, strain energy values for

135

these two conformations are very close in energy (within 1 kcal/mol) which makes it
difficult to accurately determine which conformation is preferred.

We were unsure how well Spartan could model metal centers, specifically
ruthenium, and when we were running conformational searches we required the program

to look at the conformation of cyclic systems involving a metal (the metal to the dial

    

/'

Figure 4-12. Spartan models of Ru(bpy)2Y where Y is (2R,3S)-pentanediol. Note the curve of the
bipyridine in the structure on the right (emphasized with a blue curved line).

Steric interaction

forming a 5-membered ring). It was unclear whether Spartan was correctly modeling the
changes that can occur in a cyclic system (where one atom is a metal) that would relieve
or create strain. Modeling with a molecular model kit seemed to indicate that there
would be a difference in preferred conformation for each enantiomer due to steric
interactions, so we continued to pursue this method.
4.3. Studies of cis-Ru(biq)2Clz and cis-Ru(bpy)2Clz

Initially, we pursued the synthesis of cis-Ru(biq)2C12 (where biq refers to
biquinoline) since it was expected to have a larger electric transition dipole moment and
thus would show stronger ECCD spectra. Most literature references make cis-

Ru(biq)2Clz using similar methods as those for bipyridine. We chose to follow the most

136

 

C _O ascorbic acid, LiCl
Rqu + / \ = No reaction
_ \ / DMF, 120 °C. 1.5 h

Scheme 4-5. Attempted synthesis of cis-Ru(bicDZC12.

 

N N— ascorbic acid, LiCl
— DMF, 120 °C,1.5 h

73%

 

Scheme 4-6. Synthesis of chirally enhanced cis-Ru(bpy)2C12.

recent literature reference and attempted formation of cis-Ru(biq)2C12 from RuCl3, 2,2’-
biquinoline, LiCl and ascorbic acid (Scheme 4-5).14 However, multiple attempts at this
reaction generated only starting material. With more circular dichroism and other
spectral information readily available on cis-Ru(bpy)2C12 compounds, we began synthesis
of this compound instead. Using the same method as described for formation of cis-
Ru(biQ)2C12, formation of cis-Ru(bpy)2C12 proceeded without difficulty through reaction
of RuCl3 with 2,2’-bipyridine in the presence of LiCl and ascorbic acid (Scheme 4-6).
Literature reports were conflicting in the recommended workup, with some references
filtering the reaction mixture and using the crude solid without further purification, while
one reference submitted the solid to a Soxhlet extraction for three days. We chose to
attempt both workups to determine if there was a difference in purity of the final
compound. We filtered the crude reaction mixture and washed the solid obtained with

water. The aqueous filtrate from this step was then extracted with CH2C12 to give trans-

137

Ru(bpy)2C12 in a 4:1 mixture transzcis by 1H NMR. The solid from the filtration was
submitted to a modified Soxhlet extraction for 24 hours, reserving a small sample of the
crude solid. After 24 hours, 1H NMR was taken of the initial crude solid, the solid from
the Soxhlet extraction, and the filtrate from the extraction. The product was determined
to be racemic cis-Ru(bpy)2Clz with no difference, including impurities, evident by 1H
NMR between the crude solid and the solid and filtrate from the Soxhlet extraction, and
the product showed no CD spectrum. Since there appeared to be no benefit in extracting
the crude solid and using the filtrate, all further reactions were done with the crude solid
obtained from simple filtration of the reaction mixture. However, after a couple of weeks
of storing the crude compound in a closed vial protected from light at room temperature,
a very weak ECCD spectrum was observed. Since any inherent chirality could prevent

use of this method for erythro diols, it was imperative that the cis-Ru(bpy)2C12 start and

/ N N— LiCl
RuCl3 + \ \ / r (racemic)
— DMF.120°C,1.5h

79%

 

Scheme 4-7. Synthesis of racemic cis-Ru(bpy)2C12.

remain racemic. We remade the cis-Ru(bpy)2C12 and once again weak ECCD was
observed for the crude product. This indicates that either the A or A form of the cis-
Ru(bpy)2C12 is enantio-enhanced, probably due to the use of ascorbic acid. Synthesis of

cis-Ru(bpy)2C12 was performed again without ascorbic acid (Scheme 4-7), leading to a

138

 

 

 

 

Figure 4-13. UV—Vis studies on Rqu in acetonitrile (solid line), benzene (large dashes), CH2C12
(small dashes), MeOH(dash-dot-dash), and methylcyclohexane (dash-dot-dot-dash).

0.15 a

 

 

 

 

 

Figure 4-14. UV-Vis studies on cis-Ru(bpy)2Clz in acetonitrile (solid line), benzene (large dashes),
CH2C12 (small dashes), MeOH(dash-dot-dash), and methylcyclohexane (dash-dot-dot-dash).

pure product with no ECCD spectrum. To our knowledge, enhancement of one
conformation of cis-Ru(bpy)2Clz using ascorbic acid has not been previously reported.

To determine the optimal solvent for use in ECCD studies of cis-Ru(bpy)2Clz and
related compounds, we carried out a series of UV studies to observe differences in the
UV spectra with varying solvents. As a reference, we first did UV studies on RuCl3,
using acetonitrile (ACN), benzene, dichloromethane (CH2C12), methanol (MeOH), and
methylcyclohexane (MCH). Since RuCl3 is not soluble in many solvents, we first made a
stock solution of RuC13 in ACN at lxlO'2 M. One uL of stock solution was diluted to 1

mL in the desired solvent to give a 1x10'5 M solution which was used for obtaining UV

139

spectra. As can be seen in Figure 4-13, the observed UV absorption for RuCl; is very
solvent dependent. UV studies of cis-Ru(bpy)2C12 in the same solvents as Rqu show
the same solvent dependency (Figure 4-14). In all solvents except methylcyclohexane
(MCH), Ru(bpy)2C12 shows two absorption peaks, one near 375 nm and one near 550-
575 nm, corresponding to the ligand 1t-1r* transition and the metal-ligand charge transfer
band (MLCT). As acetonitrile and methanol are common solvents for UV studies of
ruthenium bipyridine complexes, we chose these solvents to continue our studies.
4.3.1. Complexation and studies of Ru(bpy)2Clz with erythro-1,2-diols, -diamines,
and -aminoalcohols

To test the possibility of diol coordination to the cis-Ru(bpy)2C12, we first
attempted to coordinate the ruthenium with an inexpensive and readily available threo

diol, (S,S)-(-)-hydrobenzoin (Scheme 4-8). After refluxing cis-Ru(bpy)2Clz overnight

HQ OH
' EtOHzHZO (2:1)

1' = No reaction
O O

 

 

Scheme 4-8. Attempted complexation of (S, S)-(-)-hydrobenzoin with cis-Ru(bpy)2Clz.

with an excess of diol in EtOHzHZO (2:1), the solution was filtered, but NMR and UV of
the solid and the filtrate showed only unreacted cis-Ru(bpy)2C12 and dial. Unsure if the
problem was complexation with the diol or the procedure, we attempted to add base to
deprotonate the diol and force complexation. Use of n-BuLi generated trace amounts of

benzaldehyde and starting materials, but no evidence of unreactive complexation

140

n'BULI + starting

t mat 'al
0 O THF,-78°C’rt 6" s

 

   

n-BuLi
= No reaction
THF, -78 °C>rt

 

 

Scheme 4-10. Control reaction of (S, S)-(-)-hydrobenzoin with n-BuLi.

(Scheme 4-9). A control reaction run with (S,S)-(-)-hydrobenzoin and n-BuLi showed no
reaction (Scheme 4-10), indicating the ruthenium was oxidizing the diol to benzaldehyde.
Since the complexation was not proving as facile as we had hoped, we searched for a
similar complexation in literature and found an example of complexation between cis-
Ru(bpy)2C12 and 3,5-di-tert-butylcatechol.ls This literature protocol involves refluxing
cis-Ru(bpy)2Clz for an hour in methanol before adding a substrate to coordinate.
Coordination of a terminal chiral 1,2-diol to cis-Ru(bpy)2Clz was attempted by
refluxing cis-Ru(bpy)2C12 in methanol for one hour, then adding (lS)-l-phenylethane-
1,2-diol to the reaction mixture. After filtration, the crude solid of the reaction mixture
was dissolved in acetonitrile and diluted with spectroscopy grade methanol and the
ECCD spectrum was measured. The spectrum showed a few small peaks in the MLCT
band region and one strong peak at 398 nm. While slightly asymmetric, the ECCD

bisignate curve centers at 297 nm, with a positive peak appearing first at 309 nm and a

141

 

2.5

 
   
 

l
l
. l
W l
/ ’h. \
"REM,“ I}

A _ A Mmmrwa-Www

  

 

-2, 5 f . r . A

250 350 450 550 650

 

Figure 4-15. ECCD spectra of cis-Ru(bpy)2C12 complexed with (lS)-l-phenylethane-1,2-diol (solid
blue line) and (1R)-l-phenylethane-l ,2-diol (dashed pink line) in spectroscopy grade MeOH.

negative peak following at 291 nm for the R diol, and the reverse is observed for the S
diol (Figure 4-15). This indicates that the chiral diol is influencing the racemic
ruthenium molecule and causing the formation of a new chiral ruthenium species. The
chirality of the diol influences the arrangement of the bipyridine groups around the
ruthenium core in such a way that a specific chiral conformation is preferred. However,
it is not possible at this point to determine the nature of the influence of the chiral
molecule on the ruthenium, as no definitive data has been obtained which supports the
formation of a complex other than the ECCD data. As was anticipated, using diols of
opposite stereochemistry leads to ECCD spectra of opposite sign. Although initially we
had hoped to create a system that would allow for the stereochemical determination of
erythro diols, we have found one that is facile and rapid for the determination of chiral
terminal diols.

To try to extend our new method, we chose to test coordination of aminoalcohols.

Accordingly, (-)-norephedrine, a readily available erythro-1,2-aminoalcohol, was added

142

to cis-Ru(bpy)2C12 previously refluxed in methanol. The solution was refluxed overnight,
then cooled and filtered. While cis-Ru(bpy)2C12 shows only 8 proton peaks in the NMR
due to the symmetry of the molecule, the solid obtained from the reaction of
norephedrine and cis-Ru(bpy)2C12 showed multiple peaks in the lH NMR, indicating
potential coordination of the (-)-norephedrine. A CD scan of the compound from 700-
200nm in spectroscopy grade MeOH showed an unusual series of peaks. There were four
peaks at 646 nm (positive), 553 nm (negative), 489 nm (positive), and 417 nm (negative),
ascribed to the ligand charge transfer bands of the metal complex. These peaks are rather
weak and difficult to characterize as ECCD bisignate curves. However, a bisignate curve

was observed centered at 297 nm, with a positive peak appearing first at 301 nm and the

 

 

 

 

200 300 400 500 600 700

Figure 4-16. ECCD spectra of cis-Ru(bpy)2Clg complexed with (-)-norephedrine (solid blue line) and
(+)-norephedrine (dashed pink line) in spectroscopy grade MeOH.

negative peak following directly afier at 293 nm due to the interaction of the electric
transition dipole moments of the bipyridine ligands. Testing of the product formed in the
same manner from (-)-norephedrine’s enantiomer, (+)-norephedrine, was rapidly

performed to show a nearly exact opposite CD curve (Figure 4-16).

143

Testing of another available compound, D-alla—threonine, was performed after
converting the carboxylic acid to an ester and complexing the product with cis-
Ru(bpy)2Clz. A CD scan of the compound from 700-200nm, also in spectroscopy grade
MeOH, showed a similar spectrum as that for norephedrine. Again weak peaks were

observed at 500 nm (positive) and 375 nm (negative), as well as a bisignate curve

 

 

     

 

l
i
l
l
l
l
l
l

 

—3# v T I r T T I

250 300 350 400 450 500 550 600

Figure 4-17. ECCD spectra of cis-Ru(bpy)2C12 complexed with L-alla-threonine methyl ester (solid
blue line) and D-alla-threonine methyl ester (dashed pink line) in spectroscopy grade MeOH.

centered at 297 nm, with a negative peak appearing first at 300 nm and a positive peak
appearing at 293 nm (Figure 4-17). ECCD analysis of the enantiomer, L-alla-threonine
methyl ester, reacted with Ru(bpy)2C12 again revealed a nearly perfectly opposite ECCD
spectrum. These two examples indicate that this method might be a viable way for
distinguishing between enantiomers of l,2-erythra-aminoalcohols.

Analysis of the potential complexes that may be forming has been performed
using Spartan v.7.0. All calculations were done using molecular mechanics.
Interestingly enough, the strain energy values make no sense when comparing

enantiomers, and do not predict the observed ECCD sign. The differences in energy are

144

so large as to be ridiculous. However, when the models are viewed from the perspective
of A-values, new light is shed on a possible conformational preference. If the possible
complex conformations are evaluated in terms of which side of the erythro diol or
aminoalcohol is larger (by A-values), then it is a simple matter of determining which
conformation allows for the large group to sit equatorial or pseudo-equatorial. Although

the ring formed by the anticipated coordination of the aminoalcohol to the ruthenium is

I

O

. eq
,Ru:.\--<2 a,
U /N.
H H ax

Figure 4-18. Representation of pseudo-axial and —equatarial positions for the possible complexes
forming between cis-Ru(bpy)2C 12 and erythro compounds (bipyridine rings not shown for clarity).

not a 6-membered ring, the Ru-O and Ru-N bonds are long enough to allow more
conformational freedom with respect to the two carbons in the ring than would normally
be observed in a purely organic 5-membered ring. This allows the two groups attached to
the aminoalcohol to take up positions pseudo-axial or pseudo-equatorial, where the large
group prefers to be in the pseudo-equatorial position to avoid steric interactions with the
bipyridines also attached to the ruthenium (Figure 4-18).

However, even considering the large group to be pseudo-equatorial, there are still
two possibilities for binding to the ruthenium, similar to a top and bottom face approach.
Here, too, sterics play a large role, in that the conformation is favored in which the large
group is pseudo-equatorial and pointing away from the bipyridine rings. This presents

the large group as pointing into an empty quadrant.

145

This work is still in progress in our lab and other compounds are currently
undergoing testing. More data is being obtained to support the proposed complexes as
described above. At this point, it does seem that reaction of cis-Ru(bpy)2C12 with erythro
1,2-diols, -diamines, -aminoalcohols, and perhaps even other functional groups such as
thiols might make ECCD a viable tool for absolute stereochemical determination of these
types of compounds. Only a small amount of sample is needed for the reaction, and the
reaction conditions are mild. Once a mnemonic has been sufficiently tested, only one
enantiomer will need to be reacted to be certain of the stereochemistry. Further
development of this method will include screening of other ruthenium ligands,
particularly ligands with stronger electric dipole transitions and bulkier ligands which

may increase preference for one conformer.

146

Experimental materials and general procedures:

Anhydrous CH2C12 was dried over CaHz and distilled. The solvents used for CD
and UV-Vis measurements were purchased from Aldrich and were spectroscopy grade.
All reactions were performed in dried glassware under nitrogen unless otherwise noted.
Column chromatography was performed using SiliCycle silica gel (230-400 mesh). 1H-
NMR spectra were obtained on Varian Inova 300 MHz or 500 MHz instruments and are
reported in parts per million (ppm) relative to the solvent resonances (5), with coupling
constants (J) in Hertz (Hz). IR studies were performed on a Galaxy series FTIR 3000

instrument (Matteson). UV-Vis spectra were recorded on a Perkin-Elmer Lambda 40
spectrophotometer, and are reported as kmax [nm]. CD spectra were recorded on a

JASCO J-810 spectropolarimeter, equipped with a temperature controller (Neslab 111)

for low temperature studies, and are reported as 71. [nm] (At:max [1 mol'1 cm'1]).

00

02N
Trans-4,4’-dinitrostilbene6 (1). 4-Nitrobenzylchloride (4 g, 23.34 mmol, 1 equiv) was
heated to 55 °C in EtOH ( 12 mL). KOH (1.4 g, 25.06 mmol, 1.07 equiv) was dissolved
in a mixture of water (1.2 mL) and EtOH (4.8 mL) and the solution was added dropwise
to the 4-nitrobenzylchloride solution. Upon beginning the addition, the resultant solution
immediately turned brilliant green then red. As the remaining KOH solution was added a
precipitate appeared and the solution turned brown with yellow precipitate. Vigorous

stirring was required as there was a significant amount of precipitate. The temperature of

147

the sand bath was raised to 70 °C. The solution was stirred for 1 h then cooled overnight
and filtered. The residue was washed with ethanol to give trans-4,4’-dinitrostilbene
(4.1405 g, 65.7%) as a bright yellow solid. 1H NMR (300 MHz, CDC13, ppm): 5 8.25 (d,

4H, 8.7112), 7.68 (d, 4H, 8.4112), 7.29 (s, 2H).

Br Br

00

Mesa-l,2-dibromo-1,2-(4,4’-dinitrophenyl)ethane9 (2). Dinitrostilbene (1 g, 4 mol, 1
equiv) was dissolved in nitromethane (100 mL) with stirring. Bromine (840 mg, 5.26
mmol, 1.4 equiv) was added to additional nitromethane (100 mL) and this solution added
to the dinitrostilbene solution. The resultant dark yellow-brown solution was stirred at
room temperature under a superbright light for 5 d. The solution became pale red then
very pale yellow. The reaction mixture was quenched with aqueous NaHSO; and filtered
to give the product (323.5 mg) as an off-white solid. The filtrate evaporated and the
yellow solid remaining was washed with EtOAc and filtered to recover additional product
(169.1 mg). These products were combined to yield mesa-1,2-dibromo-l,2-(4,4’-
dinitrophenyl)ethane (492.6 mg, 31%) as an off-white solid. 1H NMR (300 MHz,

DMSO, ppm): 8 8.31 (d, 4H, 9.3112), 7.98 (d, 4H, 8.7Hz), 6.38 (s, 2H). 13C NMR (300

MHz, DMSO, ppm): 5 147.4, 147.2, 129.5, 124.0, 51.9.

148

 

cis-Ruthenium(bipyridine)dichloride (chirally enhanced, 4). RuCl; (1.0 g, 4.821
mmol), LiCl (1.310 g, 30.90 mmol, 6.4 equiv), 2,2’-dipyridine (1.523 g, 9.751 mmol, 2
equiv), and L-ascorbic acid (1.046 g, 5.939 mmol, 1.1 equiv) were dissolved in dry DMF
( 12 mL). The solution was refluxed for 1 hour then cooled to room temperature.
Acetone (30 mL) was added and the solution stored in the freezer overnight. The next
morning, the solution was filtered to give a black solid, which was washed with cold
water (5 mL x 2), then diethyl ether (5 mL x 2). The solid was dried under vacuum to
give cis-ruthenium(bipyridine)dichloride (1.70 g, 73% yield). Used without fiirther
purification. 1H NMR (300 MHz, DMSO, ppm) 5 9.96 (d, 2H, 5.7112), 8.63 (d, 2H,
8.1Hz), 8.47 (d, 2H, 8.1Hz), 8.05 (t, 2H, 7.8112), 7.76 (t, 2H, 6.0Hz), 7.67 (t, 2H, 7.5Hz),

7.50 (d, 2H, 5.4112), 7.09 (t, 2H, 6.3Hz).

cis-Ruthenium(bipyridine)dichloride (racemic, 5). RuC13 (1.0 g, 4.821 mmol), LiCl
(1.310 g, 30.90 mmol, 6.4 equiv), and 2,2’-dipyridine (1.523 g, 9.751 mmol, 2 equiv)
were dissolved in dry DMF (12 mL). The solution was refluxed for 1 hour then cooled to
room temperature. Acetone (30 mL) was added and the solution stored in the freezer

overnight. The next morning, the solution was filtered to give a black solid, which was

149

washed with cold water (5 mL x 2) then diethyl ether (5 mL x 2). The solid was dried
under vacuum to give racemic cis-ruthenium(bipyridine)dichloride (1.70 g, 73% yield) of
product. Used without further purification. l11 NMR (300 MHz, DMSO, ppm) 8 9.96 (d,
2H, 5.7Hz), 8.63 (d, 2H, 8.1Hz), 8.47 (d, 2H, 8.1Hz), 8.05 (t, 2H, 7.8Hz), 7.76 (t, 2H,

6.0112), 7.67 (t, 211, 7.5112), 7.50 (d, 211, 5.4112), 7.09 (t, 211, 6.3Hz).

Representative procedure for reacting Ru(bpy)2Clz with chiral diols and
aminoalcohols. cis-Ru(bpy)2C12 (100 mg, 0.21 mmol) was dissolved in MeOH (3 mL)
and refluxed for 1 hour. D-alIa-threonine methyl ester (28 mg, 0.21 mmol, 1 equiv) in
MeOH (2 mL) and NaOH (16.8 mg, 0.42 mol, 2 equiv) in MeOH (4 mL) were added
and the solution refluxed overnight to give a colored solution. The solution was cooled to
room temperature and filtered. The precipitate was washed with MeOH (2 mL x 2) to
give a dark colored solid. The solid was subjected to UV and ECCD studies without

further purification.

(1S)-1-Phenylethane-l,2-diol product. Ru(bpy)2Clz (100 mg, 0.21 mmol) was
dissolved in MeOH (3 mL) and refluxed for 1 hour. (lS)-l-phenylethane-1,2-diol (29
mg, 0.21 mmol, 1 equiv) in MeOH (2 mL) and NaOH (16.8 mg, 0.42 mol, 2 equiv) in
MeOH (4 mL) were added and the solution was refluxed overnight to give a colored
solution. The solution was cooled to room temperature and filtered. The precipitate was
washed with MeOH (2 mL x 2) to give a purple solid. ECCD (nm): 308.2 (1.9), 290.9 (-

0.6).

150

(lR)—1-pheny|ethane-1,2-diol product. Ru(bpy)2C12 (100 mg, 0.21 mmol) was
dissolved in MeOH (3 mL) and refluxed for 1 hour. (1R)-1-phenylethane-l,2-diol (29
mg 0.21 mmol, 1 equiv) in MeOH (2 mL) and NaOH (16.8 mg, 0.42 mol, 2 equiv) in
MeOH (4 mL) were added and the solution was refluxed overnight to give a colored
solution. The solution was cooled to room temperature and filtered. The precipitate was
washed with MeOH (2 mL x 2) to give a purple solid. ECCD (nm): 310.8 (-2.4), 289.9

(1.25).

(-)—Norephedrine product. Ru(bpy)2C12 (100 mg, 0.21 mmol) was dissolved in MeOH
(3 mL) and refluxed for 1 hour. (-)-Norephedrine (32 mg, 0.21 mmol, 1 equiv) in MeOH
(2 mL) and NaOH (16.8 mg, 0.42 mmol, 2 equiv) in MeOH (4 mL) were added and the
solution was refluxed overnight to give a colored solution. The solution was cooled to

room temperature and filtered. The precipitate was washed with MeOH (2 mL x 2) to

give a black solid. ECCD (nm): 302.2 (7.7), 294.7 (-3.4).

(+)-Norephedrine product. Ru(bpy)2C12 (100 mg, 0.21 mmol) was dissolved in MeOH
(3 mL) and refluxed for 1 hour. (-)-Norephedrine (32 mg, 0.21 mol, 1 equiv) in MeOH
(2 mL) and NaOH (16.8 mg, 0.42 mmol, 2 equiv) in MeOH (4 mL) were added and the
solution was refluxed overnight to give a colored solution. The solution was cooled to
room temperature and filtered. The precipitate was washed with MeOH (2 mL x 2) to

give a black solid. ECCD (nm): 300.1 (-6.9), 288.0 (1.4).

151

D-alla-Threonine methyl ester product. Ru(bpy)2C12 (100 mg, 0.21 mmol) was
dissolved in MeOH (3 mL) and refluxed for 1 hour. D-alla-threonine methyl ester (28
mg, 0.21 mol, 1 equiv) in MeOH (2 mL) and NaOH (16.8 mg, 0.42 mol, 2 equiv) in
MeOH (4 mL) were added and the solution was refluxed overnight to give a colored
solution. The solution was cooled to room temperature and filtered. The precipitate was
washed with MeOH (2 mL x 2) to give a purple solid. ECCD (nm): 299.8 (-1.5), 292.1

(1.4).

L-alla-Threonine methyl ester product. Ru(bpy)2C12 (100 mg, 0.21 mmol) was
dissolved in MeOH (3 mL) and refluxed for 1 hour. L-alla-Threonine methyl ester (28
mg, 0.21 mmol, 1 equiv) in MeOH (2 mL) and NaOH (16.8 mg, 0.42 mol, 2 equiv) in
MeOH (4 mL) were added and the solution was refluxed overnight to give a colored
solution. The solution was cooled to room temperature and filtered. The precipitate was
washed with MeOH (2 mL x 2) to give a purple solid. ECCD (nm): 300.1 (2.2), 290.4 (-

2.3).

152

References:

1. Dale, J. A.; Masher, H. S. J. Am. Chem. Soc. 1973, 95, 512.
2. Adam, W.; Diaz, M. T.; Saha-Moller, C. Tetrahedron: Asymmetry 1998, 9, 589.

3. Hoch, U.; Humpf, H.-U.; Schreier, P.; Saha-Moller, C.; Adam, W. Chirality 1997, 9,
69.

4. Nakanishi, K.; Schooley, D. A.; Koreeda, M.; Miura, I. J. Am. Chem. Soc. 1972, 94,
2865.

5. Wavefunction, Inc. Spartan SGI, 7.0.1; 1998.

6. Harder, T.; Wessig, P.; Bendig, J.; Stosser, R. J. Am. Chem. Soc. 1999, 121, 6580.
7. Bellucci, 6.; Bianchini, R.; Chiappe, C.; Marioni, F. J. Org. Chem. 1990, 55, 4094.
8. Butcher, T. 8.; Zhou, F .; Detty, M. R. J. Org. Chem. 1998, 63, 169.

9. Hartshorn, M. P.; Opie, M. C. A.; Vaughan, J. Australian J. Chem. 1973, 26, 917.
10. Hutchins, R. 0.; Taffer, I. M. J. Org. Chem. 1983, 48, 1360.

11. Hua, X.; von Zelewsky, A. Inorganic Chem. 1995, 34, 5791.

12. Knof, U.; von Zelewsky, A. Angew. Chem. Inter. Ed. 1999, 38, 302.

13. Hesek, D.; Inoue, Y.; Everitt, S. R. L.; Kunieda, M.; Ishida, H.; Drew, M. G. B.
Tetrahedron: Asymmetry 1998, 9, 4089.

14. Keyes, T. E.; Vos, J. G.; Kolnaar, J. A.; Haasnoot, J. G.; Reedijk, J .; Hage, R. Inorg.
Chim. Acta 1996, 245, 237.

15. Haga, M.-a.; Dodsworth, E. S.; Lever, A. B. P. Inorg. Chem. 1986, 25, 447.

153

IIIIIIIIIIIIIIIIIIIIIIIII

1111111111119111111111 1111 1111111111

3 1 3 02845 8671