. .4 S?

, 1‘: fl“ ;

“.mfiamyfi a? .,

Ad. ... ; . 4 .

m. fr... .,,.
a? .

.3
A;
..
.u

l.O
”In.
. c.

is:

2...,

fl

. an:
r53:

"
F?....:\IL
$13..

.. A),

l

3...
f I»: f.
‘huruspnt ;

3.
silk...

(fig h
1... x... ...

a:~ an.

:. :nhw‘lnu

£1- . . 5.1.x. , t- D
:i. liking m...» r.
. x‘ ...: .twri?
9.4.5.4 A.

 

»

"f” myiiiyiiimiyyun
mm LIBRARY

Mlehigan State
University

 

 

 

 

This is to certify that the

dissertation entitled

Oxygen Abstraction
A New Type of Carbene Reactivity and
A New Method to Generate Stable Carbenes

presented by
Dalila G. Kovacs

has been accepted towards fulfillment
of the requirements for

Ph .0. degree in Chemistry

9/“ { 0&"flw
James E. Jackson

I Major professor

 

 

/é' Derac- /39?
Date December 16, 1998

MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M WWW“

Oxygen Abstraction
A New Type of Carbene Reactivity and
A New Method to Generate Stable Carbenes

by

Dalila G. Kovacs

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1998

ABSTRACT

OXYGEN ABSTRACT ION
A NEW TYPE OF CARBENE REACTIVITY AND A NEW METHOD TO
GENERATE STABLE CARBENE

By

Dalila G. Kovacs

A new type of carbene reaction was investigated here, oxygen abstraction by
carbenes. Reactions of three different type of carbene, with suitable oxygen donors were
studied. For methylene, :CHz (triplet ground state), fluorenylidene, Fl: (triplet ground
state but rapidly equilibrating to singlet state), and phenylchlorocarbene, PCC: (singlet
ground state) experimental and computational tools were employed to demonstrate the
carbene ability to abstract oxygen atom and to get some insight into the mechanistic
aspects of this new type of reaction.

This new type of carbenes chemistry was experimentally proved, by isotope
labeling and products detection and analysis. Rates for some of the investigate reactions
of oxygen abstraction reaction were measured. Aspects of the mechanism were revealed
by qualitative Laser Flash Photolysis studies and theoretical calculations. A different
approach toward synthesis of stable carbene by carbodiimide cyclization reaction proved
to be ineffective toward carbene synthesis but a reach subject for mechanistic point of

view.

DEDICATION

Parintilor mei, Silvia si Cornel, din adancul inimii: fara dragostea voastra,

constant?) si neconditionata, nici unul din succesele mele nu ar fi fost posibil.

Fiicelor mele, Linda si Anca: fara voi viata mea nu ar avea sens.

III

AKNOLEDGMENTS

After all this years, finally here!

Dr Jackson, you enriched my professional life by the touch of your mentorship;
Evy and you provide me with kind and friendly support through my personal
struggle. Thank you!

So many people cross their path with mine during this time and what my life
would be without them! I am thankful to all of you for my professional and
personal growth; for your warm shoulder when things did not want to go my way
and for sharing with me happy moments.

My thoughts go to Linda and Anca who struggled through this Ph. D. with me,
taking all the bad moods of an unhappy chemist at work!

And finally, thanks God for give me strength and let me go so far and accomplish
my dreams.

IV

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... VII

LIST OF FIGURES ......................................................................................................... VIII

LIST OF SYMBOLS AND ABREVIATIONS .............................................................. XII

CHAPTER 1

Carbenes and Atom Abstraction Reactions: A Review ....................................................... 1
1.1 Introduction ........................................................................................................ 2
1.2 Atom abstraction reactions in carbene chemistry .............................................. 2
1.3 Oxygen abstraction by Carbon atom as a model for carbenes reactivity .......... 6
1.4 Ylide formation: first step in a possible heteroatom .abstraction .................... 10
1.5 Conclusions ...................................................................................................... 16
1.6 References ........................................................................................................ 17

CHAPTER 2

Carbene-to-Carbene Oxygen Atom Transfer: A New Type of Carbene Reactivity and a

Potential Path to Generate Nucleophilic Carbenes ............................................................ 21
Abstract .................................................................................................................. 22

2.1 Introduction ...................................................................................................... 23

2.2 Carbenes and oxygen donors ........................................................................... 24

2.3 Products detection ............................................................................................ 25

2.4 Rate studies ...................................................................................................... 30

2.5 Mechanism of oxygen transfer reaction .......................................................... 33

2.5.1 Theoretical calculations .................................................................... 37

2.5.2 Laser Flash Photolysis data results ................................................... 53

V

2.6 Experimental .................................................................................................... 58

2.7 Conclusions ...................................................................................................... 61
2.8 References ........................................................................................................ 62
CHAPTER 3

A Theoretical Investigation of Double Bonded Oxygen Abstraction on a Model Reaction

l:CH2 + C02--> CH20 + CO ............................................................................................ 71
Abstract .................................................................................................................. 7 l
3.1 Introduction ...................................................................................................... 73
3.2 Experimental background ................................................................................ 73
3.3 Methods and procedures .................................................................................. 78
3.4 Discussion ........................................................................................................ 79
3.5 Conclusions .................................................................................................... 104
3.6 References ...................................................................................................... 105

CHAPTER 4

An Alternative Approach toward the Synthesis of Nucleophilic Carbenes .................... 111
4.1 Introduction .................................................................................................... 112
4.2 Results and Discussion ................................................................................... 117
4.3 Conclusions .................................................................................................... 140
4.4 Experimental methods .................................................................................... 141
4.5 References ...................................................................................................... 143

APPENDICES ................................................................................................................. 148

VI

 

 

 

v V“?

”1.1

A.\l'

LIST OF TABLE
Table 1.1 Experimental heats of formation ............................................................... 17
Table 2.1 Thermochemistry of Oxygen Donors and Selected Rate Constants for

Table 2.2

Table 2.3
Table 2.4

Table 2.5

Table 2.6
Table 2.7
Table 3.1
Table 3.2
Table 3.3
Table 3.4
Table 3.5
Table 3.6
Table 3.7
Table 4.1

TheIR Reactions with Fluorenylidene ....................................................... 27
PM3 Calculated energies for F1: + O=CX2 -> Fl=0+ :CXz and for
X2C: +TMU —> X2C=O + :C[N(CH3)]2 reactions ................................... 40
AEa (kcal/mol) for the degenerated reactions ............................................ 43
Ab initio calculated energy differences for X2CO + :CX2 exchange
reactions (X=NH2, X2=O) ........................................................................ 45

Ab initio calculated enregy differences for XZCO + :CX2 exchange and

cross reactions (X=H, F) ............................................................................ 49
TS analysis by use of Marcus theory for oxygen transfer via ylide ........... 51
UV-VIS absorbtion of carbenes ................................................................. 55
Overal Thermochemistry ........................................................................... 81
Calculated energies for C02 + :CH2 complexes ........................................ 81
Stationary points Energies (kcal/mol) ....................................................... 85
a-lactone heat of formation AHf ............................................................... 87
Selected calculated IR frequencies ........................................................... 94
Estimate heats of formation for the diradical intermediates ...................... 90
Transition States Energies (kcal/mol) ...................................................... 101

Comparative thermodynamic data computed for overall reactions

1to6 ........................................................................................................ 124

VH

LIST OF FIGURE

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
Figure 1.14
Figure 1.15
Figure 1.16
Figure 1.17

Figure 2.1

Figure 2.2
Figure 2.3
Figure 2.4

Hydrogen and Chlorine atom abstraction .................................................... 2
Oxygen transfer from nitroxides to diphenyl carbene ................................. 4
Oxygen transfer from nitroxides to anthronylidene ..................................... 4
Creege intermediates and dioxiranes ........................................................... 5
Oxygen transfer from oxiranes to carbenes ................................................. 6
The first three electronic states of atomic carbon ........................................ 7
Atomic carbon. generation ............................................................................ 8
Possible reaction routs for atomic carbon .................................................... 8
Heteroatom abstraction by atomic carbon ................................................... 9
Double bonded oxygen abstraction by atomic carbon ................................. 9
Oxygen transfer between atomic carbon and an aldehyde ....................... 10
Ylide resonance structures ........................................................................ 11
Ylides reactivity ......................................................................................... ll
Ylide of tetramethylurea and tetrakis(trifluoro)cyclopentadienylidene ..... l3
Oxadiazolines synthesis and reactivity ..................................................... 13
Ylide synthesis: theoretical studies results ............................................... 14
Ylide cleavage to carbene and carbonyl compound .................................. 15

Fluorenylidene, methylene, phenylchlorocarbene and a carbene-carbonyl

ylide ........................................................................................................... 24
Fl: reactions with various oxygen donors used ......................................... 26
Fl: reaction with TMU ............................................................................. 26
Carbene trapping reactions ........................................................................ 28

VIII

Figure 2.5
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
Figure 2.20
Figure 2.21
Figure 2.22
Figure 2.23
Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4a
Figure 3.4b

Figure 3.4c

Carbene trapping with H20 and CH3OH .................................................. 28

Chemiluminescence setup scheme ............................................................ 29
1Ego-labeling experiment ........................................................................... 30
Competition reaction setup for rate studies ................................................ 31

Stem-Volmer quenching plot for the reaction of PCC: + PNO in
CH3CN ....................................................................................................... 32

Figure 2.10 Stern-Volmer quenching plot for the reaction of PCC: +

MNO in CH3CN ........................................................................................ 32
Possible pathways for oxygen abstraction by carbene .............................. 33
Reversible formation of an ylide ............................................................... 35
A stable carbonyl ylide .............................................................................. 36
Reaction paths considered by theoretical calculations .............................. 38
PM3 calculation at for TMU and Fl: oxygen transfer reaction ................. 41
PM3 IRC calculation at for degenerate reaction, X=F and X = NHz ........ 43
PM3 PES of cross reaction F2C=O + :CHz .............................................. 43
PES of exchange reaction H2C=O + :CHz ............................................... 44
PES of exchange reaction F2C=O + :CFz ................................................ 47
PES of cross reaction H2C=O + :CFz, .................................................... 48
Generic reactions used for Marcus theory applications ............................. 51
LPF transient absorbtions observed for Fl: + oxygen donors .................... 56
LPF transient absorbtions observed for PCC: + oxygen donors .............. 57
Potential intermediates in the :CH2 + C02 reaction .................................. 73
Potential intermediates in the :Cth + C02 reaction ............................... 75
Possible reaction paths ............................................................................... 77
Overall energy profiles at G2 (DB in Kcal/mol) ........................................ 79
Overall energy profiles at MP2/6-31G*(DE in Kcal/mol) ......................... 80
Overall energy profiles at HF/6—31G* (DE in Kcal/mol) .......................... 80

IX

Figure 3.5
Figure 3.6
Figure 3.73
Figure 3.7b
Figure 3.8
Figure 3.93
Figure 3%

Figure 3.10a
Figure 3.10b

Figure 3.11
Figure 3.12
Figure 4.1
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
Figure 4.14
Figure 4.15
Figure 4.16

32anti - 32syn interconversion path ........................................................... 89
T—S direct excitation results for 33p and 33s ............................................. 92
Path a on singlet PBS at MP2/6-31G*//MP2/6-31G* ................................ 95
Path a on singlet PBS at HF/6-31G* .......................................................... 95
Path a on triplet PES .................................................................................. 96
Path b on singlet PBS at G2 ....................................................................... 97
Path b on singlet PBS at I-IF/6-31G* ......................................................... 98
Path c on triplet PBS at G2 ........................................................................ 99
Path c on triplet PBS at HF/6-31G* .......................................................... 99
Energy differences on IRC trajectories from T8] and T82 ..................... 102
Traditional and non-equilibrium surface crossing ................................... 103
First synthesis of nucleophilic carbenes dimers ....................................... 112
l, 3 diadamantyl-imidazolin-ylidene ....................................................... 112
General synthetic methods for preparation of nucleophilic carbene ....... 113
Special cases for synthesis of nucleophilic carbenes in matrix ............... 113
Synthesis of dithiolium carbene dimers ................................................... 114
Benzyne cycloaddition to carbondisulfide ............................................... 114
Cycloheptyne reaction with C82,, ............................................................ 115
Harzler general reaction with the three adducts ...................................... 116
Di-isoPr-carbodiimide reaction with hexafluoro-2—butyne .................... 117
General reaction CD1 + DMAD .............................................................. 121
Processes followed by theoretical means ............................................... 123
General [2+3] cycloaddition of cumulenes .............................................. 127
Calcualted path for [2+3] CDI cycloaddition to acetylene ...................... 129
Calcualted path for [2+3] CS2 cycloaddition to acetylene ...................... 129

The three mechanisms investigated for the addition of DAMD to CD1. 131
HOMO-LUMO gap in ketene and CD1 ................................................... 130

Scheme 4.17 Carbonyl sulfide case ............................................................................... 132
Figure 4.18 Calculated TS for formally [1+2] cycloaddition of CD1 to acetylene ..... 134
Figure 4.19 Calculated path for formally [2+2] cycloaddition of CD1 to acetylene..135

Figure 4.20 Suggested mechanism for 4.3 formation .................................................. 136
Figure 4.21 Product 4.1 X-ray structure ...................................................................... 137
Figure 4.22 Suggested mechanism for 4.1 formation ................................................. 138

Key to Symbols and Abbreviations:

PNO
DAF

F l:
mcxv
TS
=0
TMU
DMI
DMI
BDE

CSE

PCD
PCC:
kx
PES
IRC

NImag

CID
AH!t

laser flash photolysis
pyridine-N-oxide
9-diazofluorene

anthronylidene

Fluorenylidene

carbene philicity parameter
transition state

fluorenone

tetramethylurea
1,3-dimethyl-2-imidazolidinone

1,3-dimethyl-2-imidazolidinone

bond dissociation energy; note that DHf for Oxygen is 59.6 kcal/mol at

298K.

carbene stabilization energy, defined as the energy change for the

following reaction: :CHz + CX2H2 ---> CH4 + :CX2

phenylchlorodiazirine ()
phenylchlorocarbene
rate constants

potential energy surfaces,

intrinsic reaction coordinate calculations

number of imaginary vibrational frequencies

heat of reaction 1,
collisional induced dissociation

activation enthalpy

XII

S—T
DMAD
CDI
DBCDI

TCNE

HOMO
LUMO

Singlet-to-Triplet state energy gap
dimethyl-acetylene-dicarboxylate
carbodiimide

ditertbutyl carbodiimide
tetrahydrofurane

tetracyanoethylene

Maleic Anhydride

Highest Occupied Molecular Orbital

Lowest Unoccupied Molecular Orbital

XIII

' LI!“ . ' .‘JF-Z"

 

CHAPTER 1

Carbenes and Atom Abstraction Reactions: A Review

1.1 Introduction ........................................................................................................ 2
1.2 Atom abstraction reactions in carbene chemistry .............................................. 2
1.3 Oxygen abstraction by Carbon atom as a model for carbenes reactivity .......... 6
1.4 Ylide formation: first step in a possible heteroatom abstraction ..................... 10
1.5 Conclusions ...................................................................................................... 16
1.6 References ........................................................................................................ 17

 

ii

1 . 1 Introduction

Atom abstraction is one of the typical reactions of carbenes. The most familiar
cases involve abstraction of hydrogen or chlorine atoms. In the last decade, a new type of
atom transfer has been investigated, mainly as a result of Scaiano's workl concerning
reactions of triplet diphenylcarbene with nitroxides. More recent work has described
oxygen and sulfur atom abstraction from oxiranes and thiiranes.

The idea of generating stable carbenes by oxygen abstraction reactions from
carbonyl compounds is the focus of the work presented here. The driving force for such a
reaction should be the formation of a thermodynamically stable species as in the
abstraction of a divalent heteroatom such as oxygen or sulfur, bonded by two single 0

bonds or by a double (0 + TC) bond.

1.2 Atom abstraction reactions in carbene chemistry
1.2.1 Monovalent atom abstraction: Hydrogen and Chlorine

Carbenes commonly abstract univalent atoms, such as hydrogen and chlorine, to
yield radical pairs.2 The spin state of the reacting carbene plays a determinant role. It is
commonly accepted that a carbene in its triplet state can abstract hydrogen atoms while
the singlet state abstracts chlorine atoms.3 An early noted exception to this rule is
methylene. Both singlet and triplet methylene may abstract chlorine from suitable donors.

Roth et al. studied the chlorine atom abstraction from chloroform, using CIDNP, for

l
H., H _.
' Cr“ CI

Fl

several types of carbenes (Figure 1.2)

it CHCb a .u
k- ———*H,»—H"" + cI—‘K‘g,

 

R\ Jr CHCb Rh.‘

0. H,” I
or c:
H

 

 

l H'*‘C"’“'*‘gi

Figure 1.1 Hydrogen and Chlorine atom abstraction

 

 

,_
‘rx.

.fi‘
L1H.

.‘x

‘z

I’

The carbene can form either a singlet or a triplet radical pair depending on the its initial
spin state.

Laser flash photolysis (LFP) studies4 of chlorine atom abstraction by
diphenylcarbene, well established triplet ground state, and singlet ground state
phenylchlorocarbene, PCC:, known to be singlet in its ground state, reveal that both
triplet and singlet states may abstract chlorine from a C-Cl bond.5 The rate of chlorine
abstraction by the triplet state of diphenylcarbene is one order of magnitude faster than
the corresponding reaction of singlet PCC:. These studies indicate a transition state with
a considerable amount of carbene-chlorine bond formation and charge development, and
the experimental results6 do not support the intervention of a chloronium ylide in the
reaction of phenylchlorocarbene with halomethanes.

1.2.2 Oxygen and sulfur: divalent atom abstraction

Transfer of a divalent atom, e. g. oxygen or sulfur, to make a closed-shell 1t-
bonded species such as carbonyl compounds, is rare. Only a few donors—
dimethylsulfoxide, nitroxides, amine oxides, Oz, oxiranes, and thiiranes—have been
mentioned in the literature.
1.2.2.1 Dimethylsulfoxide

In 1967 Hayashi et al. reported the oxidation of some carbenes by
dimethylsulfoxide.7 The authors proposed an ylide intermediate which decomposes to
dimethylsulfide and the corresponding carbonyl compound derived from the starting
Carbene. The yield of the carbonyl compound drops rapidly with time, and the authors
Suggested further reaction of the produced carbonyl compound with the methylsulfinyl
CElr‘banion.
1.2.2.2 Organonitrogen oxides

In 1963 Schweitzer8 et al. reported the deoxygenation of pyridine-N-oxide (PNO)
by dichloromethylene. They also investigated the reaction of 9-diazofluorene (DAF) with

PNO in refluxing benzene; 50% of the yield was fluorenone and 35% fluorene-azine.

, .
Acfihpfll

uni-Ni."

(til)?

1

< 3“ ,

tub:
I!

a

9st

~l

These results suggested an ylide intermediate which decomposes to give fluorenone,
Fl=O.

In 1984, Scaiano et al. investigated the reaction of nitroxides with triplet
diphenylcarbene, and found that oxygen atom transfer leads to quantitative yields of
benzophenonel. For 4-hydroxy-2,2,6,6-tetramethylpiperidine—N-oxide the triplet carbene

deoxygenates the nitroxide instead of inserting into the 0-H bond.

r
CH30N
o— -
k = 2.8 x 108 M‘s"

(szc.J h 4 + (PMZCO
CH3CN
O—N OH 'b—wmi’mz
k = 1.9x107M“s"

Figure 1.2 Oxygen transfer from nitroxides to diphenyl carbene

 

 

 

K

Oxygen abstraction is spin allowed, and over 100 kcal/mol exothermic; the attack at the
nitroxide center competes successfully with insertion in O—H bond which is usually
considered to be a singlet carbene reaction.

In 1988 Schuster et al. found that anthronylidene (AN) reacts with PNO to yield

92% anthraquinone.9 The reaction is assumed to occur through the singlet 1AN.

0'6 aQQW-m 0‘

Figure 1.3 Oxygen transfer from nitroxides to anthronylidene

The experiments suggested that in the absence of a good trapping reagent, 3AN converts

t0 lAN which can abstract oxygen from PNO.

ga-

1.2.2.3 Oxygen (02)

Interest in the reaction of carbenes with molecular oxygen gained popularity due
to the general interest in ozone chemistry. Carbenes react with 02 to form carbonyl

oxides, familiar as the Criegee intermediates involved in alkene ozonolysis.

(9 e O
R20-0—0 L‘O
Criegee intermediate Dioxirane

Figure 1.4 Criegee intermediates and dioxiranes

The dioxiranes,10 isomers of the Criegee intermediates, were studied while mapping out
the carbene-02 chemistry. The formation of such intermediates, detected by matrix
isolation techniques, clearly showed the range of possible carbene reactions with
molecular oxygen as well as revealing the novel chemistry of carbonyl oxides and
dioxiranes.

In comparison with the previous study9 3AN is oxidized by molecular oxygen at
a diffusion controlled rate, two orders of magnitude faster than oxidation of the same
carbene by PNO. In general, triplet carbenes react with 02 at or near diffusion control,
whereas carbenes with singlet ground states react much more slowly or not at all. The
presence of products from the reaction with molecular oxygen is usually taken as strong
evidence for triplet state trapping in a given process.
1.2.2.4 Oxiranes and Thiiranes

Fluorenylidene (FI:) abstracts oxygen from cis- and trans-2-butene oxide,11 to
Yield the corresponding alkene. The experimental results indicate that singlet state is
1' E=Sponsible for this process. The reaction is faster than formation of the known Fl:-

aCetonitrile ylide in acetonitrile solvent, but slower than insertion into the 0-H bond of

methanol.

“.C
I.

.;o <
[‘4 g

.r

l

L

 

Flih- 0:(

CH3

Figure 1.5 Oxygen transfer from oxiranes to carbenes

The yields depend on the nature of the oxirane and, from the above cited work,
are: cis-epoxybutane 60%, trans-epoxybutane 50%, styrene oxide 50% and cyclohexene
oxide 30%. An earlier estimate of the rates 12 was confirmed by recent direct
measurements by means of LFP. ‘3 Warkentin et al. give rate constants for oxygen
abstraction by several carbenes from oxiranes and thiiranes. For the wide range of
carbenes studied, rate constants from 104 to 1010 M'lS’l were found. Oxygen or sulfur
abstraction appears to obey a linear free energy relationship with respect to the carbene
philicity parameter mcxy14 with higher reactivity for electrophilic carbenes, lower for
ambiphilic ones and the lowest reactivity for the nucleophilic carbenes such as
dimethoxycarbene. By comparison, the corresponding rate constants for oxygen

abstraction from PNO exceed 109 M'1 3‘1.

1.3. Oxygen abstraction by atomic Carbon as a model for the reactivity of carbenes.
Atomic carbon is, by itself, one of the most fascinating intermediates encountered
in chemistry due to its high energy and its particularly interesting electronic
configuration. Of the fifteen possible electronic states, the triplet ground state C(3P) and
the two low-lying excited metastable singlet states C(lD) and C(18) are thought to be
involved in most of its reactions. Despite the energy differences among these three states
(~30 kcal/mol) all three have to be considered as possible reacting species. Presented in
Figure 1.6 is a schematic representation of the electronic structure of these three states,

and their heat of formation.

,, 23913“ £89

32 kcal/mol
>.-
g E
.5 8 -2—°‘ ‘DV
3 ii
30kcal/mol

 

 

1.13in dig?

Figure 1.6 The first three electronic states of atomic carbon

As with carbenes, carbon atom chemistry presents chemists with the problem of
finding which of the reactive states of the carbon atom is responsible for certain reactions.

Knowledge of the electronic state of the reacting partner and the product(s) are
useful for interpreting the results. To probe for the involvement of the triplet ground state
C(3P), molecular oxygen is often added to the system and its influence on the overall and
particular yields is used as an indication of the participation of the triplet state in the
reaction. Molecular 02 is used as a scavenger for C(3P). The reaction is governed by the
spin conservation rule which states the interaction of triplet carbon with triplet oxygen
molecule. The same rule applies in carbene reactivity studies, where the reaction with
molecular oxygen is a probe for the presence of the triplet state of the carbene in the
given reaction under study.

A general trend from the experimental information available to date is that the
most reactive species seems to be the singlet lD, followed by the triplet 3P and the less
reactive singlet 1S. In this respect the chemistry of atomic carbon parallels the chemistry
of carbenes in which the reaction rates of singlet species are generally higher than the
corresponding rates of the analogous triplet states of a particular carbene.

Atomic carbon maybe generated from diazotetrazole.15 The precursor is prepared
in tetrahydrofuran (TI-IF) solution and the solvent is evaporated while the walls of the

reaction flask are evenly coated with the diazo compound. The substrate may be added in

the initial solution of THF or after evaporation of the solvent, UV irradiation or heating to

800-1000 C, decomposes the diazocompound to N2 and monoatomic carbon in a 3 : l

X
thermal decomposition

N“ / = 3N2 + C
N—N so°-roo° 3 z 1

ratio (Figure 1.7).

 

Figure 1.7 Atomic carbon generation

The advantage of this method is that it produces monatorrric carbon with low kinetic
energy at moderate running temperatures, ~100° C, for the subsequent reaction of the
substrate with carbon.

Due to its unfilled outer electronic shell, atomic carbon acts mainly as an
electrophile. This aspect of its reactivity governs the competition between insertion into 0
or 1: bonds and atom abstraction. Monoatomic carbon presents, in several cases,
unexpected selectivity. In interactions with organic substrates, atomic carbon follows two
possible routes, both typical reactions for carbenes as well: insertion into 0' or it bonds

and atom abstraction (Figure 1.8).16

 

orY X—Z

H or X—C—Y = C

 

 

'CX+-Z

Figure 1.8 Possible reaction routes for atomic carbon

Particularly interesting is the behavior of carbon atoms toward halocarbons.l7
Halogen abstraction is most likely the primary process, generating monovalent CF or CCl
intermediates. 13 Preference for insertion into the C-Cl bond but not into the OP bond
may be explained by considering the strength of the C-halogen bond. This side of carbon

atom chemistry allows further analogies to the chemistry of carbenes.

The reactions with heteroatom—containing moieties are not entirely understood
but they do produce highly reactive carbenes. The presence of heteroatoms with their
nonbonding electrons (as in the case of ethers, thioethers or aziridines), directs the attack
of carbon toward the electron lone pair. The main course of the reaction is atom
abstraction with the formation of carbon monoxide CO, CS, or CNH (Figure 1.9).19 The
reaction is thought to follow a mechanism similar to that of carbene reactions with
oxiranes and thiiranes (Reference 11 and references therein). A carbon atom in its singlet
state is similar to a singlet carbene and the reactions of singlet carbon atoms as an oxygen

abstractor have been studied both experimentally and theoretically.

o<l —> co ll
SQ —’ cs H

C + HNQ —’ CNH H

0—.- Co 2H
001—» co :

Figure 1.9 Heteroatom abstraction by atomic carbon

However, the most interesting cases are the reactions of carbon atoms with
carbonyl or nitroso compounds. In such cases, C atoms abstract a double bonded oxygen
atom. In 1983 Shevlin mentioned for the first time the possibility of excited singlet
methylene formation in the deoxygenation of formaldehyde by atomic carbon.20 The
main products are carbon monoxide and the corresponding carbene or nitrene21 and this

reaction is used generally to generate these species (Figure 1.10).

 

_/
\ \N=0 0—\ /
C0 + N 3 C —-" 3 + CO
Nitrenes Carbenes

Figure 1.10 Double bonded oxygen abstraction by atomic carbon

M

H

Of interest to us is its similarity to the carbene reaction involving abstraction of

double bonded oxygen from suitable donors, especially carbonyl compounds.

D +
’\CH3 H ”3 ’\CH3

 

()3 + \ ”‘6’”
|’\CH3

Figure 1.11 Oxygen transfer between atomic carbon and an aldehyde

Predicted in 1968 by Hoffmann, the formation of l:CH2 was finally completed in 1983 by
Shevlin; the stepwise addition to olefins confirms its spin state.

H20=O +C

 

CD + 1:CH2

Highly energetic species are generated in the carbon atom case, but by carefully directing
a similar reaction from a thermodynamically less stable carbene to a more stable one, the
designed procedure may become a new route to generate thermodynamically stable

carbenes.

1.4 Ylide formation: first step in a possible heteroatom abstraction

The interaction of a carbene with the lone-pair electrons of heteroatoms, such as
nitrogen, oxygen, phosphorus and sulfur, usually leads to zwitterionic species known as
ylides (Figure 1.12). They are the result of interaction of a singlet carbene empty p orbital
with lone electron pairs of the reaction partner.22 Numerous ylides are mentioned in the
literature, involving heteroatoms such as nitrogen (with amines, isoquinoline, 4-picoline,
pyridine and nitriles), oxygen, halogens, phosphorus (mainly triphenylphosphine), sulfur
(sulfonium, sulfoxonium, and thiocarbonyl), arsenium, antimony, bismuth, selenium, and

tellurium.

10

 

 

bl

GQ—H%<: *—-—-> X: C<:

X = Rap, R23, R20

Figure 1.12 Ylide resonance structures

Ylide formation is a certainty well established and ylides of sulfur, nitrogen and
halogens have been detected, isolated and characterized.23 As a consequence, the idea of
ylides as intermediates in carbene reactions has solid experimental support. Generally, the
presence of an ylide as an intermediate in a certain reaction is accepted if characteristic
transformations are observedzo. Figure 1.13 lists reactions where the interrnediacy of an

ylide is generally accepted.

a. a’,B-elimination

</ R? _._A_. X-iPe H2@CH2
H

b. [1,2] sigmatropic rearrangement—Stevens rearrangement

Qi-gfiz A X—CRZ

gHzR H?!R

 

c. carbene insertion in X—H bonds where X=O, N, S

R—x—daz

 

 

R—X—H + 2092 Hfigfie

d. [2,3] sigmatropic rearrangement—allylic rearrangement

8:2 A \ijz

R

 

e. fragmentation of three membered rings

>X + 3992 —" >62ng L h X=CFiz
/

Figure 1.13 Ylides reactivity

ll

,..
if:

 

an

inn!“

~

“Cl

We are interested here in ylide formation as a possible first step in oxygen atom
transfer. Usually, an ylide may follow several different reaction paths (Figure 1.13),
depending on its ability to react with the available partners and on its thermodynamic
stability. Commonly mentioned are the paths a and b in Figure 1.13. In the presence of a
dipolarophile, [2+3] dipolar cycloadditions (path a) are frequently encountered and
specifically used for probing ylide intermediacy in certain processes. In the absence of a
suitable partner, ylides may cyclize to the corresponding oxirane (path b) or, in some
cases, decompose to more stable species (path c). However, if a previously generated
carbonyl ylide has no choice other than to decompose to the carbonyl compound and
carbene, the ylide itself becomes a reliable source for carbenes. Indeed, such a procedure
was pioneered in J. Warkentin's labs at McMaster University and has made its way in the
chemistry world.

Is interesting that it was thought that only carbenes with electron-withdrawing
groups attached to the carbenic center or carbene precursor, frequently a diazocompound,
could generate stable ylides. However, despite popular belief of their involvement in
several processes, carbonyl ylides remained only hypothetical intermediates to the late
70's. In 1978 Bartlett et al. published results for carbonyl-ylide intermediates obtained in
the decomposition of A3-1,3,4-oxadiazolines24 formed through the reaction of
diazoalkanes with ketones. The presence of products from the reactions of 1, 3 dipolar
intermediates was considered strong evidence for the existence of a short lived ylide
intermediate. Literature data are available about the solid-state photolysis of trans-
stilbene oxide which allowed the observation of an intermediate with an ylide-like
structure.25

The first stable carbonyl ylide was reported in literature in the late 803.26 The

thermodynamic stability of this designed ylide is the result of its existence mainly in the

12

 

rm-

zwitterionic polar form, a fact confirmed by the X-ray crystal structure data which clearly

reveals different lengths for the two C—O bonds (Figure 1.14).

F3 104%
F3 0‘a N(CH3)2
F3 CF3

Figure 1.14 Ylide of tetramethylurea and tetrakis(trifluoromethyl) cyclopentadienylidene

At the beginning of 90's, Warkentin published a new method to generate
asymmetric alkoxy—carbenes via thermolysis of A3-1,3,4-oxadiazolines substituted with
different alkoxy groups. The mechanism proposed by the authors involves ylide
formation with N2 loss by a concerted, irreversible, 1,3-dipolar cycloreversion, followed

by the fragmentation of the ylide.27

H

Pb(OAc)4
(CH3)2C=N—NH—COCH3

 

N
\\ 0
N

CH3OH (
cuffs

Hac><OCH3 H3O OCHa HaC\/OCH3

N o A Y ('9 A, C6H6
1}_K CH30H GZ —’
H or CCI4
org 3 crfiHa i
H3O CH3

Figure 1.15 Oxadiazoline synthesis and reactivity

Ylide formation and reactivity is the subject of several theoretical studies. In 1978
a series of results was published by Devaquet et al. and described investigations of the
overall thermal and photochemical behavior of oxiranes, using ethylene oxide as a

prototype. Their results parallel the experimental data.28 For oxiranes bearing substituents

l3

which do not stabilize zwitterionic forms, C-O bond rupture is the most effective primary
process. For the other oxirane derivatives, C-C bond rupture competes and leads to
products less susceptible to reclosure. They can either be trapped by 1,3-dipolar

cycloaddition or fragment into a carbenic entity and carbonyl compounds.29

9
CH2——'-CH2
- 110° \

o \
brig/550% '_ LA ”KGB—Cm /C CH3
110° 1 H

La fi ......

CH2

 

 

 

R2

hv or A R
R >= 0 Z
R7/)\.<R23 R1 + <83

Figure 1.16 Ylide synthesis: theoretical studies results

 

In 1980 Houk et al. published an analysis of the structure and reactivity of
substituted carbonyl ylides.3o Ylides containing amino, cyano and phenyl groups were

considered usingab initio methods (STO-3G, 4-31G basis set with 3 x 3 CI).

14

cor. 3“ H3 H2 H3

7—: —_'
3 I

H3K/DCH3 H3 H3
no H3 CH3

0
.i. Q —~ ~
H3 H3 H3C/\CH3

+ HCI
CC|3

 

 

 

 

 

 

 

 

 

 

H3 CH3
+CDaOD CCI
4
H3CO H3C 03
H3 oco H3 31-! H3
H3C 3 H30 0

Figure 1.17 Ylide cleavage to carbene and carbonyl compound

When X = Y = H the cleavage reaction is calculated to be endothermic by 38 kcal/mol
but for X = H, Y = NH2, it becomes exothermic by 17 kcal/mol. Also, the authors
concluded that thermal fragmentation of a carbonyl ylide from a coplanar ground state is
an orbital forbidden process. The same study concludes that ylides can easily form from
excited state oxiranes.

In 1982 Volatron et al. published also an ab initio study (STO-3G and 4-31G
basis sets plus extensive CI) which analyzes the electrocyclic ring opening of oxiranes.31
In these calculations, a highly energetic point was considered to correspond to an ylide-
like transition state with a significant diradical character. The authors suggest that the best
way to consider the mechanism of such reactions is through an ylide intermediate which
has a planar structure. The transition state (TS), is an orthogonal ylide-like structure
which corresponds to the rotational isomerization of the planar ylide.

Finally, the paper published by Warkentin et al. in 1981 presents experimental

results that relate to Houk's calculations. It shows the fragmentation of alkoxy substituted

15

carbonyl ylides as a thermal process which leads to a carbene and a carbonyl

compound.32

1.5 Conclusions

The precedents presented in this overview of carbene reactivity set the stage for
an exploration of the possibility of generating thermodynamically stable carbenes by
means of oxygen abstraction reactions. We reasoned that by choosing a carbene product
such as diarninocarbene, we should endow the reaction with sufficient driving force. The
known cases of oxygen abstraction from nitroxides or from oxiranes should represent
good reference points in the exploration of this proposed new reaction. The similarity of
carbon atom chemistry with that of singlet carbenes leads to the choice of a , highly
reactive singlet carbene to perform the abstraction of the double bonded oxygen.
However, other species were considered as well. One well known case is the one of
phosphorus atom as it behaves in the Corey-Winter reaction.33 The mechanism of such a

reaction is intriguing and constitutes a rich subject to be studied.

16

 

1.6 References

9

a) Casal, H. L.; Werstiuk, N. H.; Scaiano, J. C. J. Org. Chem. 1984, 49, 5214.
b) Clark, K. B.; Bhattacharyya, K.; Das, P. K.; Scaiano, J. C.; Schaap, A. P. J. Org.
Chem. 1992, 57, 3706.

a) Kirmse, W. "Carbene Chemistry" Second Edition, Academic Press, Inc. New
York, 1971, Chapter 7. b) Jones, M., Jr., Moss, R. A. "Carbenes" John Wiley & Sons,
New York, 1973.

a) Roth, H. D. Acc. Chem. Res. 1988, 21, 236-242. b) Roth, H.D. Acc.Chem.Res.
1977, 10, 85.

a) Closs, G. L.; Rabinow, B. E. J. Am. Chem. Soc. 1976, 98, 8190. b) Scaiano, J.
C. Acc.Chem.Res. 1982, 15, 252. c) Reference Sa and 5b.

a) Platz M. S.; Maloney, V. M. in "Kinetics and Spectroscopy of Carbenes and
Biradicals" M. S. Platz, Ed.; Plenum PresszNew York, 1990; Chapter 8. b) Jackson,
J. E.; Platz, M. in Advances in Carbene Chemistry, U. H.Brinker, J AI Press LTD.;
Greenwich, CT, 1994.

Jones, M B.; Jackson, J. E.; Soundararajan, N .; Platz, M. S. J. Am. Chem. Soc.
1988, 110, 5597.

Oda. R.; Mieno, M.; Hayashi, Y. Tetrahedron Lett. 1967, 25, 2363.
Schweizer, E. E.; O'Neill, G. J. J. Org. Chem. 1963, 28, 2460.

Field, K. W.; Schuster, G. B. J. Org. Chem. 1988, 53, 4000.

10 Criegee, R. Angew. Chem. Int. Ed. Engl. 1975, I4, 745.

11 Shields, C. J .; Schuster, G. B. Tetrahedron Lett. 1987, 28, 853.

12 Kovacs, D.; Lee, M-S.; Olson, D.; Jackson, J. E. J. Am. Chem. Soc. 1996, 118, 8144.

13 Pezacki, J. R; Wood, P. D.; Gadosy, T.; Lusztyk, J.; Warkentin, J. J. Am. Chem. Soc.

1998, 120, 8681.

17

 

14

15

16

l7

18

19

a) Moss, R. A; Acc. Chem. Res. 1989, 22, 15; b) Moss, R. A; Acc. Chem. Res. 1980,
13, 58. The reactivity of carbene intermediates may be predicted using the philicity
scale mCXY; values for mCXY are either determined empirically based on selectivity
of carbene toward olefines or calculated using 012+ and 0'1 values.

a) Shevlin, P. B. J. Am. Chem. Soc. 1970, 97, 1379. b) Shevlin, P. B. Tetrahedron
Lett. 1970, 46, 3987. c) Waali, E. E., Tivakompannarai, S. J. Am. Chem. Soc. 1986,
108, 6059.

The products are short lived species (see Table 1.1, heats of formation in kcal/mol)
and their existence as intermediates may be proved by product analysis of their
subsequent transformations. The intermediates may by a monovalent species known
as methylynes, carbenes or radicals.

Table 1.1 Experimental heats of formation

 

Species AHExp Species A“Exp Species AHExp
C1(3P) 171 CH 142.4 iso-C3H7 22.3
C1(1D) 201 1012 99.8 H2C=CH 63.4

C 108) 233 3CH2 92.3 H2C=CH2CH3 40,0
cz 200.2 CH3 34.8 HO 9.3

C3 196 n-C3H7 16.8 HzN 45.1

 

* from: Steward, J. J. P. J. Comp. Chem. 1989, 2, 221—264.

Blaxell, D., MacKay, C., Wolfgang, R. J. Am. Chem. Soc. 1970, 92, 50.

a) Rahman, M., McKee, M. L., Shevlin, P. B. J. Am. Chem. Soc. 1986, 108, 6296.
b) LaFrancois, C., Shevlin, P. B. J. Am. Chem. Soc. 1994, 116, 9405.

a) Skell, P. S., Plonka, J. H. J. Am. Chem. Soc. 1973, 95, 1547. b) Figuera, J. M.,
Worley, S., Shevlin, P. B. J. Am. Chem. Soc. 1976, 98, 3820. c) Skell, P. S.,
Villaume, J. Am. Chem. Soc. 1972, 94, 3455.

20 a) Rahman, M.; Shevlin, P. B. Tetrahedron Lett. 1985, 26, 2959.

b) Ahmed, S. N.; Shevlin, P. B. J. Am. Chem. Soc. 1983, 105, 6488.

18

 

21 a) Skell, P. S., Plonka, J. H. J. Am. Chem. Soc. 1970, 92, 836. b) Armstrong, B.,
Shevlin, P. B. J. Am. Chem. Soc. 1994, 116, 4071.

22 a) Clark, J. S.; Dossetter, A. G.; Wittingham, W. G.; Tetrahedron Lett. 1996, 37, 605.
b) Olson, D. R.; Platz, M. S. J. Phys. Org. Chem. 1996, 9, 759. c) Padwa, A.;
Austen, D. J.; Hombuckle, S. F. J. Org. Chem. 1996, 61, 63. d) Padwa, A.; Curtis, E.
A.; Sandanayaka, V. P.; J. Org. Chem. 1996, 61 , 73. e) PLatz, M.S.; Olson, D. R. J.
Phys. Org. Chem. 1996, 9, 689. f) Sueda, T.; Nagaoka, T.; Goto, S.; Ochiai, M. J. Am.
Chem. Soc. 1996, 118, 10141. g) Curtis, E. A.; Sandanayake, V. P.; Padwa, A.
Tetrahedron Lett. 1995, 36, 1989. h) Bonneau, R.; Reuter, 1.; Liu, M. T. H. J. Am.
Chem. Soc. 1994, 116, 3145. i) Padwa, A.; Horbuckle, S. F. Chem Rev. 1991, 91, 263.
j) Trost, B. M.; Melvin, L. S. Sulfur Ylieds: Emerging Synthetic Intermediates;
Academic Press: New York, 1975.

23 Nikolaev, V. A.; Korobitsyna, I. K. Zhumal Vses. Khim. Ob-va. Mendeleeva, 1979,
24, 496.

24 Shimizu, N.; Bartlett, P. D. J. Am. Chem. Soc. 1978, 100, 4260.

25 T rozzolo, A. M.; Leslie, T. M.; Sarpotdar, A. 8.; Small, R.D.; Ferraudi, G. J. Pure &
Appl. Chem. 1979, 51, 261.

26 a) Janulis, E. P.; Arduengo III, A. J., J. Am. Chem. Soc. 1983, 105, 3563.
b) Janulis, E. P.; Arduengo III, A. J ., J. Am. Chem Soc. 1983, 105, 5929.

27 Zoghbi, M.; Warkentin, J. J. Org. Chem. 1991, 56, 3214.
28 Huisgen, R. Angew. Chem. Int. Ed. Engl. 1977, 16, 572.

29 a) Bigot, B.; Sevin, A.; Devaquet, A. J. Am. Chem. Soc. 1979, 101, 1095.
b) Bigot, B.; Sevin, A.; Devaquet, A. J. Am. Chem. Soc. 1979, 101, 1101.

30 Houk, K. N .; Rodan, N. G.; Santiago, C.; Gallo, C.; Griffin, G. W. J. Am. Chem.
Soc. 1980, 102, 1504.

31 Volatron, F.; Anh, N. T.; Jean, Y. J. Am. Chem. Soc. 1983, 105, 2359.

19

 

32 a) Bekhazi, M.; Warkentin, J. J. Am. Chem. Soc. 1981, 103, 2473. b) El-Saidi, M.;
Kassam, K.; Pole, D. L.; Tadey, T.; Warkentin, J. J. Am. Chem. Soc. 1992, 114,

8751.

33 a) Corey, E. J.; Winter, J. J. Am. Chem. Soc. 1979, 101, 1095. b) Corey, E. J. Pure
Appl. Chem. 1967, I4, 19. For reviews see Block, E. Org. React. 1984, 30, 457.

20

CHAPTER 2

Carbene—to—Carbene Oxygen Atom Transfer:
A New Type of Carbene Reactivity and a Potential Path to Generate

Nucleophilic Carbenes

Abstract ...................................................................................... 22
2.1 Introduction ............................................................................ 23
2.2 Carbenes and oxygen donors ........................................................ 24
2.3 Product Studies ........................................................................ 25
2.4 Rate Studies ............................................................................ 30
2.5 Mechanism of oxygen transfer reaction ............................................. 33

2.5.1 Theoretical calculations .................................................... 37

2.5.2. Laser Flash Photolysis data .............................................. 53
2.6 Experimental ........................................................................... 58
2.7 Conclusions ............................................................................ 61
2.8 References .............................................................................. 62

21

Abstract:

Fluorenylidene (F|:) abstracts oxygen from pyridine-N-oxide (PNO), to give fluorenone
(Fl=O) in high yield. Other substrates—N-methyl morpholine-N-oxide (MNO),
sulfolane, trimethyl phosphate, tetramethylurea (TMU), 1,3-dimethyl-2-imidazolidinone
(DMI), and dimethyl carbonate (DMC)—also oxygenate Fl:. The latter three, TMU,
DMI, and DMC, should yield carbenes instead of stable deoxygenated byproducts. This
heretofore unexplored carbene-to-carbene exchange has been verified by two observations:
the 18O label is transferred from TMU to FI=O, and tetrakis(dimethylamino)ethylene (the
dimer of the product carbene) is detected by its chemiluminescent reaction with atmospheric
oxygen. Competition between methanol and oxygen donors along with laser flash
photolysis studies in acetonitrile show that Fl: deoxygenates PNO as fast or faster than it
inserts into methanol O-H bonds: kPNO/kMCOH = 1.7 i 0.4; analogous experiments with
TMU and DMI give kTMU/kMeOH = 0.49 i 0.02 and kDMI/koOH = 0.52 i 0.02. These
relative rate data translate into absolute rate constants of 2.6 and 2.8 x 108 M'lS'l,
respectively, in good agreement with an independently determined LFP value of 3 x 108 M'
13'1 for quenching of Fl: by TMU. These measurements represent the first absolute rate

constants measured for carbene-to-carbene oxygen atom transfer processes.

22

2.1. Introduction

We report here evidence for carbene-to-carbene oxygen atom transfer, according to

the reaction:
RzC: + O=CR'2 —> R2C=O + :CR'2 (Reaction 1)

Singlet atomic carbon abstracts oxygen atoms from a wide variety of carbonyl
compounds to produce carbon monoxide and carbenesl; these highly exotherrrric processes
can produce "hot" carbenes with unusual behavior.2 We now report that the reactive
carbenes fluorenylidene (Fl:) and methylene (:CHz) behave analogously if the carbene
product is sufficiently stabilized with electron donor groups (see Table 2.1).3 Besides its
intrinsic interest as an abstraction of a doubly bonded atom,4 this type of reaction
represents a new photochemical pathway to generate nucleophilic carbenes and study their
chemistry.

Carbenes do abstract oxygen atoms from suitable donors such as N -oxides5 ,
nitroxides6, or epoxides.7 Many carbenes react with molecular oxygen to give carbonyl
oxides and their isomeric dioxiranes.8 With simple carbonyl compounds such as
aldehydes,9 ketones,lo esters,ll amides,12 and ureas,l3 electrophilic singlet carbenes
attack the oxygen lone electron pairs to form carbonyl ylides 1 (Figure 2.1). These
intermediates may then cyclize to form epoxides14 or undergo cycloaddition with a second
equivalent of carbonyl compound to give dioxolanes.9'10g A third, almost unexplored
pathway, is fragmentation to a new carbene/carbonyl compound pair. Indeed, Warkentin et
al. have shown15 that explicit synthesis of ylide 1 (R=CH3, R'=OCH3) via oxadiazoline

decomposition leads to acetone and dimethoxycarbene products.

23

H.

PCC:

13’! or:
3-5.1.1
U

PC D

1.4.?
XL‘LM

Fl :CHz PCC: 1

Figure 2.1 Fluorenylidene, methylene, phenylchlorocarbene and carbene—carbonyl ylide

2.2 Carbenes and oxygen donors

In this work, fluorenylidene (Fl:), methylene (:CH2) and phenylchlorocarbene

(PCC:) have been examined as oxygen abstractors. The carbenes were photolytically '5‘

generated from diazofluorene (DAF), diazomethane (DAM), or phenylchlorodiazirine

(PCD), respectively, and their reactions were studied by product analysis and laser flash 1!

photolysis (LFP). 1'
DAF is a relatively stable diazo compound, readily available in two steps from ...

 

fluorenone.l6 Irradiation of DAF with UV-VIS light from a high pressure Hg lamp (500
W) filtered through uranium glass generates Fl:. DAM is synthesized via established
methods and trapped directly in the neat urea used as the oxygen donor, thereby eliminating
ether, the standard solvent used for DAM solutions. The diazirine PCD is synthesized via
Graham's method.17 After purification, it is photolyzed to form PCC: in neat urea or
solutions containing the oxygen donor.

Despite its triplet ground state, Fl: generally shows singlet behavior because of the
high reactivity of its singlet state and its small singlet-triplet gap (1.1 kcal/mol).18 LFP
studies of Fl: via the ylide probe method19 are well described in the literature. In addition,
the oxygenation product fluorenone (F l=O) is easily detected. Similarly, singlet reactivity
dominates the chemistry of :CHz in condensed phases due to the singlet’s high reactivity
and relatively slow rate of intersystem crossing to the triplet ground state.20 The singlet
carbene PCC: has also been extensively studied by LFP both direct and via the ylide
probe method.

24

By adding the PCC: to our study we were able to cover the entire spectrum of
carbene reactivity, from a carbene with typical triplet ground state (:CHz) to one with a
small S-T gap and rapid equilibration (Fl:) to one with a known singlet ground state
(PCC:). Also, the known behavior of PCC: and Fl: in LFP studies offer the solid
foundation needed in exploring a new reaction type such as oxygen atom transfer reaction.

The oxygen donors used here may be classified as: (i) familiar oxygen donors such
as pyridine-N-oxide (PNO), 4-picoline-N-oxide, N-methyl morpholine-N—oxide (MNO)
and cis- and trans-2—butene oxides,7 all expected to follow the previously described Tr
chemistry in the literature; (ii)'poorer' donors, such as dimethyl carbonate, sulfolane, and
trimethyl phosphate, which may undergo the same types of reaction but with slower rates;

and (iii) urea—type donors which are expected to generate stable carbenes substituted by
21 l

 

amino substituents at the carbenic center.
The idea of generating stable nucleophilic carbenes is the core of our study. Such
species are expected to be more thermodynamically stable compared with the starting
carbenes that we plan to use, such as :CH2. We also expect the carbene—to—carbene
oxygen atom transfer to be driven by the favorable thermodynamic outcome of the overall

process.

2.3. Product Detection
2.3.1 Identification

In dry degassed acetonitrile or benzene, F1: is oxygenated by pyridine-N—oxide
(PNO), 4-picoline-N-oxide, N -methyl morpholine-N-oxide and cis- and trans-2-butene
oxides7 to give Fl=O. Minor byproducts are bifluorenyl, bifluorenylidene, and in some

cases, products of reaction with solvent.

25

Fl: 4» O=X —%9—EI—> FI=O + byproducts

\
ms :NC) ND) QESO >1: :JqN] :P(OCH3)3
/ /

Figure 2.2 Fl: reactions with various oxygen donors used

Given the checkered history of the ylide formed from FI: and acetonitrile (see Platz,
M. S., ref. 4d, pp. 285-287 and references therein) we explicitly generated this species by
photolysis of the corresponding azirine precursor in the presence of PNO to verify that
PNO oxygenation of this ylide could not make the F l=O observed. In any case,
subsequent rate studies showed that, in the concentration ranges studied, the acetonitrile

ylide formation could not compete with oxygen atom transfer.22 The poorer oxygen

 

I: 1" ‘_ -qu'i'infi‘fi up.

donors—dimethyl carbonate, sulfolane, and trimethyl phosphate—gave similar results
when used neat. Yields of Fl=O were substantial, ranging from 30-90% based on DAF.

In the key reaction of our study we consider the stability (see Table 2.1) of
diamino3 and dialkoxy carbenes23 as the driving force to completion of oxygen transfer.
We examined tetramethyl urea (T MU), 1,3-dimethylimidazolidin—2-one (DMI), 1,3-di-
tert-buthylimidazolidin—Z-one ( DTBI), imidazolidin-Z-one (1M) and dimethyl carbonate
(DMC) as substrates. Like the more traditional oxidants, they reacted with F1: to give
Fl=O.

mates“

1 it

Figure 2.3 Fl: reaction with TMU

26

Table 2.1 Thermochemistry of oxygen donors and selected rate constants for their
reactions with F l:.

 

x: waxy AHf(XO)a BDEb c313.C kxokaCOH ka(LFP)x10‘8d
1:c: 201 —26 286 51

F1: 156e 13f 202

1H2C: 102 -26 187

we: -45 -153 168 57

(MeO)2C: -35g -139 163 92 1.2i0.2x10’3 0.01
(H2N)2C: 39h -59 158 79

(MezN)2C: 44i -57 160 72 4.9:02xro-1 2.6 (3)
CzH4(NMe)2C: 56i -41 156 52:02 x 10-1 2.8

:co -26 -94 127 120

(MeO)3P: -167 -265 158 1.4202 x 10-2 0.8
C4H3SO -35 -88 113

Z-2—butene —2 ~30 88 3
E-2-butene -3 -31 88 9

C5H5N: 33 14 79 17:04 9 (4.8)
4-Mecjg4N: 25 6 79 1.7204 9

 

3 Unless otherwise noted, these are AHf values at 298 K, from Lias, S. G.; Bartmess, J. E.; Liebman, J. F.;
Holmes, J. L.; Levin, R. D.; Mallard, W. G.; “Gas Phase Ion and Neutral Thermochemistry” J. Phys.
Chem. Refi Data, 1988, 17, Suppl. 1.

b BDE = bond dissociation energy; note that AHf for Oxygen is 59.6 kcal/mol at 298K.

c CSE = carbene stabilization energy, defined as the energy change for the following reaction: :CH2 +
CY2H2 ---> CH4 4» :CYz

d Griller, D.; Nazran, A. S.; Scaiano, J. C. J. Am. Chem. Soc, 1984, 106, 2128-39.

e Li, Y.; Schuster, G. B. J. Org. Chem. 1988, 53, 1273; based on a computational estimate, this can not be
considered an independent value.

f Sabbah, R.; Watik, L. E.; Minadakis, C. Comptes Rendus de l 'Academie des Sciences de Paris, 1988, 307,
Serie II, 239.

g Estimated from the PA (234 kcal/mol) calculated at MP3/6-31+G*//HF/6-3IG* and the known AHf (97
kcal/mol) of (MeO)2CH+. The MP2/6-31G*//HF/6-31G* + scaled ZPE reaction energies for F2C: +
(MeO)2C=O --> F2C=O + (MeO)2C: and (H2N)2C: + (MeO)2C=O --> (HzN)2C=O + (MeO)2C:
give similar results, -31 and -35 kcal/mol respectively. Overall, the agreement between these three

independent calculations leads us to discount the -61 kcal/mol value reported in an earlier
experimental study (see ref 21).

h McGibbon, G. A.; Kingsmill, C. A.; Terlouw, J. K. Chem. Phys. Letter , 1994, 22, 129-34.

i Substituted diaminocarbene AHf values were estimated via experimental values and the MP2/6-
31G*//HF/6-31G* + scaled ZPE reaction energies for the oxygen exchange reactions with urea.

 

27

 

In a parallel set of experiments, :CHz was examined with TMU, DMI and DTBI.
In each case the reaction was run in the neat urea and only products from carbene
dimerization were detected

Benzoyl chloride is obtained as major product from the reaction of PCC: with
TMU and DTBI, by oxygen atom transfer. Quenching the final reaction mixture with
MeOH gave methyl benzoate, which was detected via GC-MS spectrometry.The results
listed in Table 2.1 implicate the corresponding stabilized carbenes as byproducts in the urea
and carbonate deoxygenations.

In the TMU deoxygenation studies we tried to trap the diarrrinocarbene with
alkenes such as cyclohexene, dimethylfumarate, acrylonitrile, norbornadiene or

chloroacrilonitrile, as shown in Figure 2.4.

CN
CN X /
X5 : + E X’hQ-K'
Cl 0'
Carbene 2 expected product

Figure 2.4 Carbene trapping reactions

However, our attempts to trap the bis(dimethylamino)-carbene were unsuccessful and the
only products found were those from the reaction with Fl: and/or DAF. Such highly
stabilized carbenes may not react significantly with simple olefins, preferring to dimerize
instead. In contrast, electron-deficient alkenes, which should be most appropriate for
trapping nucleophilic carbenes, readily undergo 1,3 dipolar cycloadditions with DAF,
destroying the carbene precursor and leading ultimately to fluorene-containing
cyclopropane products.

Figure 2.5 Carbene trapping with H20 and CH3OH

28

These compounds were not isolated since they were generated in amounts detectable only
by GC-MS, and they are hydrolytically sensitive materials which makes their isolation and

characterization challenging.

2.3.2 Chemiluminescence experiments

Like dimethoxycarbene,” the bis-(dimethylamino)-carbene generated by TMU
deoxygenation dimerizes in the absence of traps. The resulting tetrakis(dimethylamino)—
ethylene reacts rapidly with 02 at room temperature to produce TMU (as in our 18O-TMU
synthesis)21 and visible light. This characteristic chemiluminescence verified the presence
of the carbene dimer when air was bubbled through photolyzed samples of DAF in neat
TMU inside a fluorimeter cavity.24 Neither photolyzed samples of TMU without DAF,
nor unphotolyzed DAF-containing samples showed chemiluminescence. The same

analysis demonstrated that :CHz, generated by photolysis of diazomethane, also abstracts

oxygen from TMU.
M92 18 Me
Fl: + '50: 018—» FI=O + 2N».

M 82%:0 02 MezN>C_dNMez J Dimenzatlon
M62N M62N NM92

DETECTED

Chemiluminescence
in VIS region

Figure 2.6 Chemiluminescence setup scheme

This experiments are correlated with detection of 18O-Fl=O via GC-MS and FTIR (vide

infra).

2.3.3. Oxygen isotope (130) labeling experiments
To verify that oxygen transfer occurs as we envisioned, we needed to demonstrate

the specific transfer of oxygen from the donor to the Fl: acceptor to yield Fl=O as a

29

product. Thus we required a urea oxygen donor, labeled with 180. T etramethyurea proved
the most accessible by far. To build the tetramethyl urea structure with the desired isotopic
substitution a rather unconventional synthesis was employed25 involving the reaction of
1802 with tetrakis(dimethylamino)ethylene. After distillation and column chromatography,
labeled 18O—TMU was obtained in near quantitative yield; the content of 18O-TMU in the
final product was 95% (determined by GC-MS).

MezN __ NMe2 1802 M92 ,8

MezN NM92__'2 M62 0

Figure 2.7 18O-labeling experiment

Reaction of Fl: with 18O-labeled TMU was conducted, according to the general
procedure described in the experimental part, and the products were analyzed by GC-
MS.26 The GC retention times were identical with for the labeled and unlabeled F|=O
product while the mass and fragmentation pattern showed that the 180 was incorporated in
the product Fl=O, confirming that oxygen transfer occurs from the carbonyl substrate to

the carbene Fl: (see Appendix).

2.4 Rate Studies

Absolute rate constants for the F1: reaction with oxygen donors were measured by
LFP using the ylide probe method.18 The results are listed in Table 2.1. Competition with
insertion into methanol O-H bonds allowed additional rate constants to be calculated via the

absolute rate constant for F1: quenching with methanol in acetonitrile.27

9-Methoxyiluorene F|=O

Figure 2.8 Competition reaction setup for rate studies

30

 

Ratios of Fl=O to the 9-methoxyfluorene product were determined either by NMR or GC
analysis; the resulting rate constants are included in Table 1. Several absolute concentration
ratios of oxygen donor to methanol were examined to ensure that the observed product
ratios reflected the direct reaction of Fl: with the substrates and not indirect pathways to the
same product. Table 2.1 shows that rate constants obtained directly by LFP compare
reasonably well with those determined in competition experiments; as expected, the
reactivities of oxygen donors with Fl: track inversely with C=O bond strengths.

Unpublished data, previously obtained in our labs, prove the singlet nature of the
O-transfer reactions from closed-shell donors. Diphenyldiazomethane was photolyzed with
PNO and methanol. As in the case of F l:, the rate constant for oxygen transfer from PNO
to diphenylcarbene (DPC:) is essentially the same as for O-H insertion. Thus, the
reactivity of PNO parallels that of methanol for both carbenes, even though DPC:, with its
larger singlet-triplet gap,28 reacts substantially slower than Fl:.29 In contrast, DPC: reacts
at nearly diffusion—controlled rates with open-shell oxygen donors such as 2,2,6,6-
tetramethylmorpholine (TEMPO) (see Reference 4b).

In addition to the values obtained for Fl: (Table 2.1), our rate studies were
extended with an LFP study of PCC: abstracting oxygen from PNO and MNO. Absolute
rate constants for the reaction of oxygen transfer to PCC: from the two N—oxides were
measured by means of UV-LFP and the pyridine ylide method, using the Stern-Volmer
analysis (Reference 31). For more details see Appendix.

Bimolecular rate constants of 2.2 x 109 M'IS‘I and of 2.1 x 109 M‘IS'l were
obtained for PCC: abstracting oxygen atom from PNO and MNO, respectively. Oxygen
abstraction reactions could not be followed in pentane because of the low solubility of the
two N—oxides used here as oxygen donors. Also, the rapid formation of the fluorenyl
radical by hydrogen atom abstraction from the solvent would strongly compete with
oxygen transfer from N-oxide to F1: while the absorption band of Fl-H will have a chance

to overlap with possible short-lived intermediates.

31

 

P5
‘11‘: 1.

l"...
U

21 -. CPC: + Py + PNO .
' CPCz-Py ylide at 475 nm .

 

 

16 i
O //
/
11 — / 7 .
O
6 O . O
O
1 4>/'
r-4 . r r T 1 1 1 I
-0.01 0.01 0.03 0.05 0.07 0.09 0.11 0.13 !
[I’NO] nrol/l 3

Figure 2.9 Stem-Volmer quenching plot for the reaction of PCC: + PNO in CH3CN

 

 

 

 

PCC + MNO + Py
6 l PCC:-Py ylide 475nm 3
5 ~ : //’/’
4 d o :”/ o ’
" ’ 0 ,/’/0
if 3 o //. . o
< . /
2 t //
1 4r/
0 r-- — - -- — — - 2,
O 0 05 0.15

h 0 1
[MNOl mol/l

Figure 2.10 Stern-Volmer quenching plot for the reaction of PCC: + MNO in CH3CN

2.5 Mechanism of oxygen transfer reaction

The mechanism responsible for the oxygen abstraction reaction is intriguing when
the transfer of a double bonded atom from a carbonyl compound to a carbene is involved.
Such a reaction has no precedent in the literature except for the case of carbon atom

chemistry. The pathways we considered in our investigation are depicted in Figure 2.11.

 

R14. R 1 R14. R 2 R1 .13
H2 R R2 H H2 H

Carbene 1 Ylide Carbene 2

I.

e
4 5 e
\———> R1,". .\\R R O\\°“R
Oxirane

 

 

Figure 2.11 Possible pathways for oxygen abstraction by carbene

One attractive pathway is the direct formation of the ylide (Reaction 1), which
depending on the specific 1,3 disubstitution pattern, may directly cleave to generate carbene
2 and the carbonyl compound (Reaction 2). Such a two step route for oxygen atom transfer
from one carbene (1) to another (2) will be dependent upon the stability of the ylide, the
barrier for ylide cleavage in Reaction 2, and competition between the opening of the ylide to
the more thermodynamically favored carbene 2 vs. ylide closing to the corresponding
oxirane (Reaction 3). Another possibility is a single step reaction (Reaction 1 + Reaction 2,
where the ylide is not a minimum but a transition state, TS) in which, after reaching the
corresponding transition state to the ylide, a direct passage to carbene 2 will be found due
to its specific 1,3 disubstitution. If carbene 1 directly adds to the carbonyl C=O bond in a
[1+2] cycloaddition the first step will be oxirane formation (Reaction 4) followed by C-C
oxirane ring opening (Reaction 5) to the ylide. The ylide will cleave to carbene 2 and the
second carbonyl compound.

A solid starting point for our analysis is the available literature concerning the
mechanisms considered above. Literature data about ylide or oxirane formation and their
fate in the cases of :CH2, FI: and (PCC:) are presented and are analyzed in connection
with our findings.

:CHz is the prototype of triplet ground state carbene. Ylide formation in its case

should be the result of fast reaction of the singlet before it can equilibrate to the most stable

33

triplet state. Methylene was found to form a stable ylide with acetonitrile with an absorption
maximum at 280 nm.30 The case of the postulated methylene ylide with water is the matter
of long-standing controversy. Most of the theoretical studies propose a barrierless process
with the intermediate is too short-lived to be experimentally detected. But mass
spectrometric techniques used by Wesdemiotis et al. found the existence of an intermediate
in the gas phase, with the corresponding mass of an H2C+ 0H2 ylide.31 Recent high level
theoretical calculations of Wiberg et al. describe the ylide as a stable species.32 By contrast,
similar high level calculations for the case of dichloromethylene show no evidence for a
stable ylide formed from dichlorocarbene, ClzCz, and H2033

:CHz is proved to be more reactive toward the C=O bond than to the C-H (a) or C-
H(B) of butanone. The relative reactivity was found to be 1:0.08: 0.05.34 Because of the S-
T gap of ~9 kcal/mol, the interception of the singlet methylene by the carbonyl compound
to form an ylide or oxirane should be faster than the intersystem crossing to its triplet state.
High level theoretical calculations for the reaction of ClzC: with formaldehyde reveal the
ylide as a stable species (calculated as a minimum ) with transient character.35 However,
the ylide formation was calculated to be 2.5 times slower than direct addition to the C=O
bond, which is the dominant process.

The case of :CH2 and the acetone ylide was investigated experimentally36 by
photoacoustic calorimetry. The heat of formation for the ylide is estimated to be 4.5i9.9
kcal/mol for the gas phase and 12.4i10.9 kcal/mol in solution. The same study found the
estimated heat of decomposition for the ylide to methylene and acetone to be -45 kcal/mol
exothermic. The methylene ylide with formaldehyde is also controversial. The results
mentioned above show the thermal cleavage of methylene—acetone ylide as ~45 kcal/mol
exothemric and theoretical calculations on methylene and formaldehyde found the cleavage
reaction also exothermic by 38 kcal/mol.37 However, a different communication indicates
reversible formation of the same methylene-formaldehyde ylide at -78° C because of HID

exchange observed in a system containing :CH2 and D2C=O (ref 8a).

34

With its much smaller S-T gap, Fl: presents a different case than :CHz. The
literature on F1: indicates a series of ylides detected as short lived transients by LFP.38 The
florenylidene ylide with acetonitrile absorbs at 400 run while those with carbonyl
compounds absorb in the region of 620-680 nm.

The case of a carbene with well known singlet ground state, PCC:, is discussed
here because of our interest in using this carbene in experimental work. Preliminary
experiments in our lab proved qualitatively the possibility of oxygen transfer from urea—
type oxygen donors to PCC:. The well established behavior of PCC: in LFP experiments
encouraged us to include it in our studies. The experimental detection of ylides from the
PCC: reactions with acetone, acetonitrile or ethyl acetate is controversial. Despite their
controversial nature, we present here the available data from the literature (vide infra).

Substitution in the para position of the phenyl ring with nitro, chloro or
trifluoromethyl groups allows the formation of detectable carbonyl ylides via LFP. 39
Thus, PCC: forms an acetone ylide with Amax at 450 nm (weak) while the p-NOz-PCC:
acetone ylide is seen at 590 nrn.40 In addition, a transient found at 530 nm was attributed to
the ylide of PCC: with benzaldehyde.41 The cyclization of the p—NOg-PCC:O=C(CH3)2
adduct was used to explain the appearance of a new absorption at 365 nm, 42 attributed to
the corresponding oxirane.

Ylides which may be formed in reversible processes have been reported.43 Also,
the ylides from biphenylchlorocarbene and ethers are believed to be responsible for the 1,3
dipolar cycloaddition products experimentally obtained.44 However, LFP experiments only
I"fiveal a 15 nm bathochromic shift in the carbene absorption maximum. A kinetic study on
the same arylcarbene system with various concentrations of the ether and monitoring of the
Carbene absorption allowed determination of the equilibrium constant (Figure 2.13) for the

TeVersible formation of the ylide.
Am,” / _ A93 0/
0" K=0.5M'1.293K C 69

Figure 2.12 Reversible formation of an ylide

 

35

Cases in which the carbonyl partner in the formation of an ylide is of urea type are
scarcely mentioned in the literature. Of specific interest in our case is an electronically
stabilized carbonyl ylide (Figure 2.2.2) formed from tetramethyl urea and tetrakis-
(trifluoromethyl)cyclopentadienylidene13. This ylide was generated from the corresponding
diazocompound of the carbene with TMU in THF and recrystallized from CHCl3/n-pentane
as a yellow solid with mp = 190-193 °C.

F3 £01102
F3 0 N(CH3)2
9
F3 cr=3

Figure 2.13 A stable carbonyl ylide

However, this encouraging idea of TMU's capability to form stable ylides with carbenes
should be reconsidered with respect to the nature of the carbene. If such stable ylides were
formed in our experimental cases with TMU and :CH2, Fl: or PCC:, then the
decomposition of the ylide to the diarrrinocarbene would likely not occur at room
temperature. In this respect, a comment by Platz et al. 45 seems especially important. These
authors note that the ylide of cyclopentadienylidene and TMU was never detected in their
LFP experiments, although several other ylides were generated and observed from
cyclopentadieneylidene.

The direct formation of oxiranes from carbenes and carbonyl compounds has been
mentioned in literature, 45 but it was generally connected with fluorinated or perfluorinated
carbenes. Also, the pyrolysis of several perfluoropropylene epoxides around 165° C was
found to form exclusively difluorocarbene.47

Considering the range of reactivities observed and the uncertainty surrounding

processes involving carbenes and their capacity to form ylides, one should not expect the

36

investigation of a new type of carbene reaction to be an easy task. We have limited our
study here to only two modest investigations of the oxygen transfer mechanism. One is the
use of theoretical calculations, performed on model systems in order to compare the
carbenes' abilities to form ylides vs. their known tendency to cycloadd to double bonds and
form oxiranes. The other includes experimental investigation by qualitative detection/no
detection of transient species, observable by nanosecond LFP. Both sets of results are

presented in the following sections.

2.5.1 Theoretical calculations

The interaction of carbenes with carbonyl oxygen atoms was confirmed by our
experimental results. The products are a new carbonyl compound and a new carbene.
DAM, DAF and DPC were used as sources for :CH2, F l:, and PCC:, respectively,
while ureas were the oxygen donors.

Semi empirical and ab initio computational methods were used to calculate the
results for comparison to the relevant experimental data available. A primary focus of this
effort was to understand the thermodynamics involved in the oxygen transfer reaction as a
function of carbene stabilization. Exploration of the two possible reaction paths, via oxirane
(Reaction 4 in Figure 2.15) or via an ylide intermediate or ylide-like transition state
(Reaction 1 and 2 in Figure 2.15) were also key goals. In the case where both paths are
available, we looked for the energetically favored path for oxygen atom transfer. It is also
possible that both types of intermediates are present on the path that connects them, as
shown in Figure 2.15.

In our analysis, we looked at two different types of reactions (Figure 2.15):
exchange reactions, where X=Y and cross reactions where X¢Y. The cases of :CXz with
X=H, F, NHz, :C(N(CH3)2)2, Fl:, and X2=CO were studied at the serni-empirical level
while the expense of ab initio methods limited us to X = H and Y = F, two exchange

reactions, :CH2 + H2C=O and FzC: + F2C=O, and one single cross reaction between

37

 

:CHz + F2C=O. An extensive study was carried out on the model reaction C02 + :CH2,
and is the subject of Chapter 3.

By following these cases, we hope that our data analysis would allow comparisons
among the different carbenes' behaviors and insight into "the" mechanism of oxygen
transfer from carbene to carbene. More accurately, each case must be carefully considered

with regard to both the nature of the carbene and the oxygen donor.

9
1 Rh _
/_T R:/GO\ R \

ylide
Y
x... _ ”.xY
Carbene 1 \ 2 Fl 1A3 / Carbene 2
——> R2 R

oxhane

Figure 2.14 Reaction paths considered by theoretical calculations

2.5.1.1 General considerations

The attack of an electrophilic carbene on a carbonyl-containing partner may occur
either at the 7: bond of the carbonyl moiety or at the lone pair electrons of the oxygen atom.
The 1: (direct) attack leads to formation of an oxirane. The side attack at the lone pair leads
to formation of an ylide (Figure 2.15)

If the oxirane is formed initially, its chemistry may involve subsequent C-O or
C-C bond cleavage. The C-C bond cleavage leads to an ylide-like configuration, possible
an intermediate or only a transition state. Cleavage of the ylide releases a new carbene and
the corresponding carbonyl compound. Direct decomposition of oxiranes to ylides and
eventually to carbenes and carbonyl compounds was claimed to occur in the case of
tetrafluoro ethylene oxide.48 There are numerous theoretical studies of oxirane ring
opening. 49 The existence of the ylide formed in such a way was proved by absorption and

emission detection and LPF. 50 Detailed, high level calculations for the C-C ring opening

38

of ethylene oxide have also been published.51 The process is generally described as
opening of the oxirane ring to the ylide followed by isomerization of the ylide and its
reclosure to the oxirane ring. The isomerization of the ylide was found to be the step with
the highest barrier.

Ylide formation is the result of an initial attack on the lone pair of electrons from the
oxygen atom. The ylide is usually trapped by 1,3-dipolar cycloaddition reactions. In the
absence of a suitable partner for cycloaddition, the ylide itself has two paths for generating
a carbene and a carbonyl product: direct decomposition into the new products, and
cyclization to the corresponding oxirane.

Special consideration should be given to cases where diazocompounds are used as
carbene precursors. In such cases, a cycloadduct of the diazocompound with the carbonyl
compound may be formed. It has an oxadiazoline—type of structure and it decomposes,
photochemically or thermally, generating N2 and ylides. The ylides formed through this
path are able to decompose with generation of a carbene and a carbonyl compound.52
Oxadiazolines' formation and ring opening, however, is not the subject of our
investigation.

We describe here a range of molecular orbital calculations run at Semi-Empirical
AM1, PM3 and ab initio RHF/6-31G* and MP2/6-31G*//MP2-6-31G*53 levels, using
Spartan, Mopac54 and Gaussian 9455 programs. Our goal was to find the differences and
similarities among the mechanisms of this reaction and define the method and level of
calculations which best fit the requirements for describing such mechanisms. The strategy
involved finding (i) possible minima, ylides or oxiranes, (ii) transition states related with
these minima, and (iii) paths connecting these two classes of stationary points. Only the

ground state potential energy surfaces, PES's, were considered.

39

2.5.1.2 Results and discussion

Semi-Empirical results

Semi—empirical level calculations were used to evaluate the overall thermodynamic
outcome of the oxygen transfer reaction (AHRcaction) for Fl: and TMU and the results are
presented in Table 2.2. Other carbenes were also considered for the exchange reactions
defined in Reaction 1 (Figure 2.15) where X = H, F, NH2 and X2 = O. The use of :CHz
offers a handle for comparisons with the available experimental and theoretical data. The
substituents in the case of X = F and NH2 should provide information about the role of
substituents on the carbenic and carbonyl centers along with the stabilization or
destabilization of the possible intermediates by electron donating groups. At the same time
F may itself constitute an unpredictable exception.56 The case of CO abstraction of oxygen
atoms from C02 may lead to more complete understanding of CO oxidation and the
reactivity of C02. Meanwhile, it may prove to be a special case of oxygen atom transfer

(see Chapter 3).

Table 2.2 PM3 Calculated energies for Fl: + O=CX2 —> F l=O+ :CX2 and for
X2C: +TMU -> X2C=O + :C[N(CH3)]2 reactions

 

X EFl:+OCX2 ECompl. TSto ylide ylide TSto oxirane onirane
H 76.4 not found 108.6 89.3 120.0 64.3
F 7.4 1.9 not found not found 4. 8 -33.7
NH2 108.5 102.5 73.6 86.8 72.6

N(CH3)2 129.6 not found 102.8 99.1 113.9 92.6

 

E:CX2+TMU Ec6mp1. T810 ylide ylide T310 oxiran onirane

 

H 65.9 not found -l2.4 -17.5 -2.3 -2.2
F -112.9 -118.1 -8l.5 -155.9 -95.2 -104.9
X2=0 -50.6 -53 .2 not found not found 13 .7 -64.2

 

4O

As expected, in all cases considered, the oxygen transfer reaction is calculated to be
exothermic. For the experimental case of F l: + TMU the oxygen transfer is calculated to

be 55.4 kcal/mol exothermic. However, barriers were also computed in all the cases.

/ S to oxirane

,_ ‘\ 18 lo Ylide
: 44 /

~
7

TS to oxirane

        

 

 

 

 

L; ’—‘ it

E 8.

R >

a“: .9

8 8

i a.» ,2

o g -

% =5 0 l“. TS to oxirane _13 X=H

a E; T ‘..--2;~.

b C) ' ‘~

8 TMU + 'CF LE :30_-5 X=Me2N

I“ ' 2 : Fl: + O=CX2 -37
reaction coordinate reaction coordinate

Figure 2.15 PM3 calculation for TMU and Fl: oxygen transfer reaction

Our semiempirical results describe the reaction of :CHz with CH20 (the case of X
= H) with no complex formed between the reactants. A direct path leads to a planar ylide.
The ylide cyclizes to oxirane. No direct path to oxirane was found. For the case of CF20 +
:CFz, X=F, one possible pathway leads to oxirane via a high TS. Another path was found
direct toward the formation of an ylide. The ylide is less stable than the oxirane but its
formation requires less energy than the formation of the corresponding oxirane. In Figure
2.16 below, the energies are plotted with respect to the reaction coordinate. X: (NH2)2,
(NH2)2CO + :C(NH2)2 a similar behavior as in the case of X = F was found. There are
two possible pathways for the oxygen transfer reaction, one through an ylide-like structure,
the other through an oxirane one, with similar appearance as in the case of X = F. Carbon
monoxide (the case of X2 = O), with its singlet ground state, was included here because of
its analogy to a singlet carbene. The :CO+C02 reaction only follows the pathway through
an oxirane structure via a defined transition state. A special type of intermediate was found

with a four-membered cyclic structure. IRC proved that such a pathway is not related with

41

 

oxygen transfer. For these reasons, we concluded that the analogy between a singlet
carbene and CO will not give useful inforrnations about the mechanism investigated here
and no further analysis was performed on this system. Even though the system is of no
interest for the cases investigated here. The CO—to—COz reaction remain an unexplored
theoretical model and experimental challenge.

Two possible reaction pathways on the , one through an ylide-like intermediate with
a slightly lower activation energy (TS), and another one through an oxirane were found. In
the case of :CF2 and :C(NH2)2 the pathway following the ylide-like structure presents a
lower barrier than the one via an oxirane intermediate despite the higher energy of an ylide
vs. an oxirane intermediate. IRC calculations performed at the AMI and PM3 levels show
no relation between the two paths, implying that there is no pathway on the reaction PES

from ylide to oxirane or vice versa.

  
 

   
 

 

 

 

 

:11
g TS to oxirane E i to oxrrane
§ / a
I TS to ylide g TS to iide
g 8 y o
b _" ..... ‘ c x
g ’X 16.5 '- -- ._1_3 F2C'O CF2 g .’ 36 ......... (“le 2C7 C(NHziz
'6 ’ x 91.; \7'6
C 9‘ g (H2N)2C (NH2l2
w 0 FZC-cr2 w o

F20=0 + ICF2 (NH2)2C=O + :C(NH2)2

reaction coordinate 3 reaction coordinate T

Figure 2.16 PM3 IRC calculation at for degenerate reaction, X=F and X = NH2

Even though the ylide, when found, is less stable than the corresponding oxirane, the

activation energy required to generate the ylide is much smaller than is needed for the

oxirane to cleave to carbene (see Table 2.3).

42

Table 2.3 AEa (kcal/mol) for the exchange reactions

 

X,X Ea to ylide Ea to oxirane AEa

 

H no ylide -33-2
F,F 16.5 25.8 9.3
HzN,NH2 36.3 47.7 10.5
=0 no ylide 49.6

 

For the cross reaction F2C=O + :CH2 the ylide path require less energy despite the

fact that the ylide is calculated to be less stable (Figure 2.17).
F2C=O + :CH2

ii

TS from oxirane

/ TS from ylide
‘\ '35
\‘ / H2C=O + :CFZ
— ............ u.- - - - ~ —
-51 / '48 .... - ‘50

complex

 

Energy difference (kcal/mol)

 

reaction coordinate

Figure 2.17 PM3 PES of cross reaction F2C=O + :CH2

Ab initio results

MO ab initio calculations were performed at the HF/6-31G* level for optimization,
vibrational analysis and IRC. Single point calculation and occasionally, geometry
optimization and vibrational analysis were run also at MP2/6-31G*. The energy data
obtained for the stationary points and the corresponding transition states are presented in
Table 2.6 and 2.7 as differences, in kcal/mol.

The case of the methylene and formaldehyde exchange reaction presents the same
general patterns here as found at PM3 level.

H2C=O + :CHz ---> H2C: + O=CH2 AHexp/calc = O kcal/mol

Initial attack at the oxygen lone pair leads, via a complex, directly to a planar ylide

intermediate. The ylide evolves into the oxirane via a TS.

43

TSone step

    

complex

— HF/6-31G*+ZPE oxirane -88-7
....... MP2/6-31G*+ZPE

 

Figure 2.18 PES of exchange reaction H2C=O + :CHz

Discrepancies between the results at the HF and MP2 levels were not surprising as long as
HF level calculations is generally provide poor descriptions of cases like the one under
discussion here.57 HF/6-31G* IRC calculations confirm the connection of the starting
species with the oxirane via TScomplex to ylide- However, single point calculations at the
MP2 or MP4/6-31G* level starting with HF geometries as well as full optimization at the
MP2/6-31G* found no barrier on the path from H2C=O + :CHz to ylide. The barrier for
C-C cleavage is calculated to be extremely high ~ 82 kcal/mol. A direct one step path
through a possible ylide-like TS was found only at HF, over a barrier of 15.5 kcal/mol. A
similar structure could not be found at higher level of calculations

Two possible mechanisms were found for the reaction with X = F:

F2C=O + :CF2 ---> F2C: + O=CF2 AHexp/calc = 0 kcal/mol

Initial attack at the oxygen lone pair leads to a complex ~2 kcal/mol lower in energy than the
starting species. Two different pathways were found for the attack of :CFz on the carbonyl
compound. One is a two-step pathway through an oxirane intermediate, with Ea = 23
kcal/mol at the HF level (TS complex to oxirane) and 10 kcal/mol at MP2 level. The other
pathway, found only at the HF level, is described as an one-step oxygen transfer via an
ylide-like TS (TS complex to ylide): with Ea = ~86 kcal/mol. IRC calculations confirm the
evolution of the reaction in one step, with an overall transfer of an oxygen atom from one
molecule of :CFz to another. The HF level of calculations does not lead to an ylide as a
minimum on the PES. However, while searching for the corresponding TS of the one-step
reaction at the MP2 level, a minimum was found (with a geometry close to the ylide-like TS
found at HF), 34 kcal/mol higher in energy than the starting F 2C: + O=CF2 (Figure
2.20).58

Table 2.4 Ab initio calculated energy differences for exchange reactions

 

 

HF MP2
(HzN)2CO + :C(H2N)2 0.0 0.0
Ylide +26.6
Oxirane +5.6 +13.3
vdW complex -9.5 -1 1.9
TS through ylide +33.6 +885
TS through oxirane +29.9 +79.2
COz+CO 0.0 0.0
vdW complex -0.4 -0.6
Oxirane +49.9 +74.8
TS through oxirane +53.2 +80.2

 

As for the case of X = F, calculations were run only at the HF/6-31G* and MP2/6-
31G*//HF/6-31G* levels for urea—type of oxygen atom donors, where X: NH2. The
energy difference values for X=NH2 and X2=O are summarized in Table 2.4. A full
optimization on the structure of TMU-formaldehyde ylide run at MP2/6-31G* found such

an ylide a minimum. This result, even though it may not be interpreted in terms of the

45

barriers on the PES, indicates that ylides may be short-lived intermediates in the carbene

oxygen atom abstraction reaction.

i

TS compl to ylide

i E (kcal/mol) HF
_ rim-310* + ZPE
- - MP2J6-3IG*//MP2/6-3IG* + ZPE 859 Mp2 not found

   
 

......

 

complex

Figure 2.19 PES of exchange reaction F2C=O + :CFz

For the case of cross reaction, F2C=O + :CH2 ---> FzC: + O=CH2

AHexp = -20 kcal/mol, AI-Icalc = -25 kcal/mol, initial attack at the oxygen lone pair leads

to a complex, ~6 kcal/mol lower in energy than the starting species. From the complex the

reaction follow the path which leads to an oxirane intermediate through a transition state,

TSl

new

which we were unable to locate. . In a following step, oxirane evolves via TS2 to a

complex 6 kcal/mol lower in energy than the products. Finding such a complex

indicates a possible significant interaction between the formaldehyde oxygen atom and the

empty it orbital of the singlet :CFz. The transition state for the oxirane ring opening is

calculated to be ~ 50 kcal/mol higher in energy than the difluorooxirane but ~26 kcal/mol

below the energy of F2C=O + :CHz.

46

i

j E

 

—- HF/6-31G" +ZPE
MP2/e316: + ZPE

TS one step

  
     
 

. -27.6
complexl _ . H200 +:CF2
- . I fi @ ‘\‘"' """ -
.401 '33-6

 

complex2

oxhane

Figure 2.20 PES of cross reaction F2C=O + :CHz

47

 

Seneca-9-3532 a

 

om- Sm- Ram- 38. mean. Sam- Sem- SZN- 8.8.
:. when Sam- weam- comm- Sem- ahmm- 8.8. 283623 £0" + 0%:
3.98 mamm- Nwem- 8.? 8.9.- 8.8- 36368
$8- 3.8- 3%. mmdm- ENN- eSm 8.8 a 9.95 E.
8.3- wm.m~- new 26 a 58.9-8 we
Saw- 2.5- 33. 0.5. 88. scene
36. com. 8.? 36388
E. Amacowaoc «50" + OUNm
cents-9-85352 a
3 3.2 Ga 82. :w 8.8 0.3 seams-fies 2.
8.? 3% 8.3 2%» a: 9-. as we
Sam- :.8. 66mm- 43-N- Efi- 88:8
8.3 8.; 8i
2. can- EN- EN- 8a. :2. .8358
Amacoweob NED" + OUNm
- w
3w- Ew- Nesa- Raw- 8.x:- Nmea- 6.87 8:. SS- 9.2999 98 8%
3.2. omd- NE a- S s- 63- one 8.4 sewers-me; E
No.8. 38. 8a 2%. a: 1.62.-. same
:2- $.8- 32 as 2995 me
33. onww- 8.x:- Nmea- 6.87 5:- 5.9. Ease
~32.- owdr mama 5.3.- 3.8- 2.0 3.2. 62;
3:- 22- 2.8- .35- 2.:- RA- 2:- 5388
Amazowmob ~50“ + DUNE
m
Sm No NEE: EDS-$2 team-£2: $5.3 "HESS ENE: 52-2.5

 

E £33 25:82 805 new owcasoxo NXO + OUNX a8 moo—58:6 xwaoco 2:5 no “023330 Wm use--

48

2.5.1.5 Exploring the ability of Marcus Theory to predict the oxygen transfer barrier
Marcus theory was first applied to electron transfer and, later on, to protons or
other atom or group transfers. We explore here the possibility of applying Marcus theory to
the (gas-phase) carbene—to—carbene oxygen atom transfer reactions.

In the simplest kind of atom transfer, the reaction coordinate involves a concerted

motion forming one bond and stretching the other.

X-A+X _> X----A----X —->X+A-X exchangereactionll

Y-A + Y —-> Y- - - - A - - - -Y —> Y + A-Y exchange reaction 22

X-A + Y ___> x- ---A---- Y —-> X + A-Y cross reaction 12

When the bond breaking—bond forming process is the principal contributor to the reaction
coordinate, the equation below can be used to relate the barrier of a cross reaction to those
of exchange reactions:

E12 = E(l+ ABC/4E) eq 2.5.1
where: E12: the barrier of cross reaction

E=(E11+ E22)/2 where E11, E22: barriers of exchange reactions

ABC: the potential energy change in the cross reaction
Thus, Marcus theory offers a way to predict the activation energy of degenerate exchange
reactions. We sought to check the predicted values against those found by exploration of
the reaction potential energy surfaces.

A series of oxygen transfer reactions involving carbenes with the general structure
XzCz, where X=H, F, Cl, Br, and I were investigated. Because of the heavy atoms
involved and the computer time required, the calculations were performed only at the PM3
Semiempirical level. As detailed above, the two possible paths, through oxirane and ylide-

]ike intermediate were found in all cases.

49

x, Y X,,, ,.--Y

 

"—0 + :-""m ., . .
xf'——- \Y r . + O=\Y
Carbonyl Carbonyl
compound 1 Carbene 1 Carbene 2 compound 2

where: X = F, Cl, Br, I

Figure 2.21 Generic reactions used for Marcus theory applications

Comparison of the computed (Appendix 1) energies with those obtained using
Marcus theory are in fairly good agreement. Our results, calculated PM3 semi—empirical
level are listed in Table 2.7 as differences in energy, in kcal/mol. For easier interpretation
of Table 2.7 the relative energies of carbene involved along with the energies of the
corresponding carbonyl compound are included in the last two column of the table. The last
column in Table 2.7 contains the difference between the calculated at Semi-Empirical and
calculated values using Marcus' theory. The range of differences is very broad, from 23.6
kcal/mol for X = F, Y = Br to only 3.1 kcal/mol for the case of X = H, Y = F. The fact that
for X=H and Y=F, the difference is only 3.1 kcal/mol as it is in the case of X = Br, Y = I,
may be only the cancellation of error effect rather than a better description of the theory
apphed.

Only with this data in hand is difficult to predict the capacity of Marcus theory to
elucidate the barriers in the oxygen atom transfer. First, the level of calculations used,

PM3, may give only a qualitative picture of the ranges of energies involved. more detailed
ab initio data will be required to conclude about the possibility of Marcus's theory

applicability on the oxygen atom transfer.

50

.268 can .090 939 3 demo
”An-E o .2- .098 ”ad: 2 _- .096 .: .3 .o .ommo 930:8 a 89.2.... 9998.8 92092 .28 85228 £99.59. .8 .99. Nmuuo 9 230m 9

deg 2: $0 59? V 3.

émo éau 3H- .98 x23 2: - .Eu 5;: o .90” 932.8 a mamas 8228.8 92092 .83 85928 .8925 .6 98.. Eu” 9 299.9; a

 

 

 

mhoaé mmwcfi ww.m am.NH N69 w.on NS H mm

mNON.w mhwofi wbd NQMH “Nb Wow 907 H HO

mbboéH- mboHdH H.H .v Fwd- 5mm #9 w.mm gm 6

oawH H- ma©.© mhwdH VON.” adH- H.mN- Wm?- H "H

mowmdm- mbmbd- mom.mH vaéN- gem- odo- om- um "H

www.cH- mH H6 moHKH mmfl- Héw- YHQ- N.Nmu HU nH

92: mod- mNodH- wm.¢ Hda méa vdH H H.H

2.99.0- mNom.mH- mmde modH- v.3. Wmm H.NH Hm H.H

mbcod whocfi- mmc.NH- Him ohm VHM Qmm HU H.H

woo-m meMH Vmfi- ahmbH Hawa- H.Nw- H.wc m I

0.0m HMOH 39.: HdNH OHMH H H
H6 c.0- HrNd 0.05 adm- um um
06H- vam- bofi Him YOH HO HO
CH H- H.wH.H- cmdm HawON- v.me- nH ”H
o.o 0.0 Human. v.3. H.0HV HH H.H

NXUHO NXU”

£34 «HHHd waoEEmHloEUm £532 ame NHAxme+xxmc .030 ham-H m>U +ONXU UHmeH. Hcam-H4 V X

 

fimquE «3 comment samba 8H H.505 3832 Ho 3.: E math—£8 WH- m.m 033-

51

2.5.1.5 Summary and Conclusions of Theoretical Results

From the calculations presented here it appears that semiempirical results provide a
reasonable estimate of the heat of reaction and in many cases a qualitative picture of the
overall mechanism of carbene-to—carbene oxygen atom transfer. The findings are similar to
the ab initio data, but in some of the cases, an ylide described as a minimum at the PM3
level becomes a TS at ab initio.

Our calculated data suggest general characteristics for the oxygen transfer reaction
mechanism:

(i) on the singlet PBS, the attack of the carbene on the lone pair of the carbonyl oxygen
predominates (except for the FzCO case).

(ii) oxirane intermediates are located in deep wells on the PES. However, C-C bond
closure requires, in general, traversal of high energy barriers.

(iii) at the level of calculation used here, one-step mechanisms, via ylide-like TSs,
present barriers comparable or higher then those via oxirane.

(iv) when the thermodynamics of the reaction is favorable, a direct path for oxygen
transfer from carbene-to-carbene via an ylide intermediate or T8, with no oxirane
intermediate involved, may be preferred.

As a general conclusion, the oxygen transfer seems as expected to be, heavily
dependent upon the substituents present at the carbenic center. High levels of theory will
likely be required to fully describe the systems under investigation. Our initial survey
results, however, indicate that the oxygen atom transfer from a carbonyl compound to
carbene is most likely to occur via a transient ylide intermediate. If such a path is available

the intermediacy of an oxirane is unlikely.

52

1:1

p D-.

1~r,.—1.
. .
Hui-l...

2.5.2. Laser Flash Photolysis results

Carbenes form ylides when interacting with carbonyl compounds59 and abstract oxygen
from €02.60 Methylene and fluorenylidene abstract oxygen from ureas leading to
formaldehyde or fluorenone, respectively.61

Carbonyl ylides, generated by decomposition of oxadiazolines‘"2 or by photolytic or
thermal C-C cleavage of oxiranesé3 lead, in some cases, to carbenes and carbonyl
compounds. In order to experimentally probe the question of the possible intermediates in
the oxygen abstraction reaction by carbene, we undertook a laser flash photolysis (LFP)
study . 64

LFP studies of PCD and diazoflorene DAF are reported. These compounds, upon
irradiation, afford phenylchlorocarbene PCC: and fluorenylidene F 1:. In addition to our
rate measurements (vide supra) we include here LFP experiments for detection of possible
short-lived intermediates in the reaction of F I: or PCC: with ureas as oxygen donors.
Qualitative data about the presence of transient species in the LFP of Fl: and PCC: in
acetonitrile and pentane in the presence of oxygen donor were obtained. A laser setup
similar to the one used for rate studies was used here (Nd-YAG laser, 355 nm, 10 ns
pulse, 10 mJ). Stock solutions of precursors in acetonitrile and pentane, with a optical
density of 0.3 to 0.7 at the irradiation wavelength (355 nm), were prepared. A large excess
of the oxygen donor was added and each sample was homogenized and degassed with
bubbling Ar for 10 to 15 min.

Control experiments were conducted by irradiation of PCD or DAF alone in the
same solvents to verify the correlation with the available data from the literature. The
following ureas were used as oxygen donors: TMU, DMI, IM, and DTBI. These
compounds also were irradiated alone under the experimental conditions to confirm that no
new absorptions were seen. For Fl: in acetonitrile, the known transient absorptions of
triplet Fl:, at 470 nm, the FlH radical at 500 nm, and the acetonitrile-Fl: ylide at 400 nm

were identified when no urea was present in the photolyzed mixture. Irradiation of PCD

53

afforded the known transient absorption of PCC: at 318 nm with no interference from any
other transient.27

In Figure 2.25 and Figure 2.26 the transient absorptions obtained for both Fl: and
PCC: are presented. The experiments shown evidence for an intermediate for both PCC:
and for Fl:. None of the known absorptions, usually present in the LFP of these two
carbenes, matches the new features. Considering that the control experiments showed no
additional absorptions than those characteristic for each carbene in the given solvent, the
new absorptions have to be the result of processes involving the carbene, PCC: or Fl:,
and the oxygen donor, TMU, DMI, IM or DTBI-

Considering the possible mechanisms depicted in Figure 2.14, the new absorptions
may be attributed to ylide or oxirane intermediates, to the newly generated carbenes or to a
combination of these and the known ylide formed by the carbene and the solvent. Since the
available literature data correlates with our calculated pathways for the oxygen atom
transfer, we believe that the transient absorption in the LFP of our analyzed systems may
belong either to a short-lived ylide intermediate or to the main product of the oxygen
transfer reaction, the diaminocarbene.

Even though the literature of the last two years abounds in data about new
diaminocarbenes with many crystal structures reported, we found no information about
their UV-VIS spectra. Table 2.8 lists the literature for the UV-VIS absorption of some

carbenes which may be compared with our case.

54

V In
OIIA

45..
«till

W t)

~r':

Table 2.7. UV-VIS absorption of carbenes

 

 

Amax (nm)
1:CH2 550-950 gas phasea
:CHF 430-600 gas phaseb
:CHCl 550-820 gas phaseC
:CF2 220-270 matrixd
:CC12 480-560 matrix‘3
:CFCI 360-390 gas phasef
Co 360 MeTI-IF gassg
:CCl(OMe) 3 l 8 matrixh
:CPh(OMe) 290 hexane, -10 °Ci
:C(OMe)2 255 matrixj

:CF(OMe) 240-280 matrixk

 

a) Graham W. H. J. Am. Chem. Soc. 1965, 87, 4369.

b) Merer, A. J.; Travis, D. N. Can. J. Chem. 1966, 44, 525. c) Merer, A. J.; Travis, D. N. Can. J.
Chem. 1966, 44, 1541. d) Smith. C. E.; Jacox, M. E.; Milligan, D.E. J. Mol. Spectroscopy, 1976, 60,
381. e) Jacox, M. E.; Milligan, D.E. Chem. Phys. 1976, I6, 195. f) Tiel, J. J.; Wampler, F. B.; Rice,
W. W. Chem. Phys. Lett. 1979, 65, 425. g) Quinkert, G.; Kaiser, K. H.; Stohrer, W. -D. Angew. Chem.
Int. Ed. Engl. 1974, 13, 198. h) Kesselmayer, M. A.; Sheridan, R. S. J. Am. Chem. Soc. 1986, 108,
99. i) Moss, R. A.; Shen, S.; Hadel, L.; Kmiecik-Lawynowice, G.; Wlostowska, J.; Krogh-Jespersen, K.J.
Am. Chem. Soc. 1987 109, 4341. j) Moss, R. A.;Wlostowski, M.; Shen, S.; Krogh-Jespersen, K.;
Matro, A. J. Am. Chem. Soc. 1988, 110 4443-44. k) ref 23.

55

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.22 LPF transient absorptions observed for F I: + oxygen donors
Fl: + ureas in pentane
o 5 AM .-
2 or!" - - :79..-” .
E -5 . Fl: - DTBl
3-10. .5 W Fl:~TMU
.9 -15-
-2O . . . .
360 410 460 510 560
w avelenghtlnm
H: + ureas in acetonitrile
20 ]
13 -
a l \.
5+ a. .
g o r 5
°\° -5 M Fl: + DTBI
i Fl: * TMU
'10 v 1 ~ r
360 410 460 510 560
w avelenght/nm
Fl: in acetonitrile
0
8
(6
e
O
U)
D
(U
a?
-1 fi . .
360 410 460 510 560
wavelenght/nm

 

 

 

56

 

9....

Figure 2.23 LPF transient absorptions observed for PCC: + oxygen donors

 

 

 

 

 

 

 

 

 

 

PCC: +ureas in acetonitrile
fl)
8
E
O
8
(5
.\° 7 ‘ V..—
wrfa' PCC: + DTBI
T "" ’ ' PCC: + TMU
450 500 550
wavelenght/nm
PCC: + ureas in acetonitrile
8
E PCC:
‘9', PCC: + DTBl
a PCC: + TMU
°\° PCC: + DMl
PCC: + lM
w avelenght/nm

 

 

57

In the case of PCC:, a strong new absorption band is easily noticed in the case of DTBI
+ PCC:, with Amax = 380-410 nm. However, the observation of analogous absorptions
for TMU, DMI and [M is hampered by the remnant signal from the 350 nm laser pulse
and the residual absorption from the carbene itself. In the Fl: case, as may be noticed from
the transient absorption spectra, the presence of the oxygen donor in a photolyzed sample
of Fl: in acetonitrile gives rise to a previously unnoticed broad absorption, with a
maximum at A = 390—450 nm. The position and intensity of this new absorption change
with the urea donors. However, because of the broadness of the band and its overlap with
the absorptions of fluorenyl radical (hmax = 500 nm) and the Fl: - acetonitrile ylide (An-ax
= 400 nm) it is hard to define quantitatively the amount of changes with respect to the ureas
used.

These qualitative results show that a new transient species is formed in the oxygen transfer
reaction, detectable on the timescale of our LFP experiments. The existence of such a
transient, whether it is intermediate or product, may allow direct measurements of rate
constants. We are inclined to attribute these absorptions to the newly formed carbene rather
than to the possible ylide. With no available data in the literature, however further

investigations and different approaches are necessary in order to confirm our assumption.

2.6 Experimental

genera methods. Melting points were determined on a Thomas Hoover capillary apparatus
and are uncorrected. Fourier-transform infrared (IR) spectra were recorded on a Mattson-
Galaxy FT -IR 3020 or Nicolet IR/42 spectrometers. Samples were measured either as thin
layers prepared by evaporating a CH3CN solution on a NaCl plate (liquids) or as KBr
pellets(solids).

58

Electron impact (EI) mass spectra were obtained on a Fission VG trio-1 mass
spectrometer which operates in line with a Helwett Packard 5890 gas chromatograph for
GC-MS measurements.

Routine 1H and 13C NMR spectra were obtained at 300 and 75.43 MHz
respectively using either a Varian VXR—300 Spectrometer or a Varian GEMINI 300 NMR
Spectrometer. The 1H NMR chemical shifts are referenced to the 1H resonance in CDC13
(7.24) and acetonitrile-d3 (1.93). The 13C chemical shifts are referenced to CDC13 (77.0)
or acetonitrile-d3 (broad 118.0 and 1.3 septet).

Solvents were purchased from Aldrich and dried and deoxygenated by standard
procedures (see Vogel in Reference 66). Gravity and flash column chromatography were
performed on E. Merck silica gel (230—400 mesh). Thin-layer chromatography was done
on E. Merck plastic-backed plates silica gel 60, F245, 0.2 mm. High resolution mass
spectra analysis were carried out on a JEOL JMS-HXl 10 high resolution double—focusing

mass spectrometer.

General procedure
Literature procedures were followed for DAF65, DAM66 and PCC:l7 synthesis.

DAF was recrystallized from pentane. DAM was collected directly in neat oxygen donor.
PCD was purified either by distillation or by column chromatography (Reference 17 ) The
physical and spectroscopic data of the carbenes precursors listed above were in agreement
with those reported in literature.

General procedure for irradiation: l to 2 mL solution of carbene precursor (3 to 5 x
10‘3M) in dry acetonitrile was prepared and placed in a standard NMR tube (1 cm x 10
cm). After being purged with dry nitrogen gas for 15 mins, the solution was irradiated for
l h with a 500W high pressure mercury lamp shielded with a uranium glass filter. The
nitrogen purge was continued throughout the process of irradiation. After removal of

solvent, the mixture was separated by flash chromatography over silica gel. Same

59

procedure was used for the case of neat oxygen donor experiments when samples of 1 to 2
mL solutions of carbene precursors in neat oxygen donor were used. The samples were
deoxygenated by purging with nitrogen for 10 to 20 min. before irradiation. The carbenes
were generated in the reaction mixture by irradiation with light from a 500 W Oriel high-
pressure Hg lamp, through an uranium glass filter. Reagents consumption was followed in
all the cases by 1H and 13C NMR. GC-MS spectra were taken for the initial reaction
mixture and exactly the same conditions were applied at the end of the reaction time to
detect the final mixture composition.

l8O-TMU was prepared from tetrakis(dimethylamino)ethylene (TDMAE)
purchased from Aldrich by reaction with molecular 1802. 2.1 mL. (0.0073 mol) TDMAE
was placed in a round bottom flask and attached to the vacuum line. 74 mL 1302 was added
to the flask via vacuum line by heating-cooling cycles, in three portions. The reaction flask
was kept under stirring at the vacuum line for 3 hours, after which time the visible
chemiluminiscence vanished. The crude product was separated by column chromatography
with a solution of isopropanol : acetone in a 3:2 ratio. 18O-TMU was obtained with a 93.1
% yield and a content of 98% 180 by GC-MS. The IR taken on NaCl pellets shows vc=018
shifted to smaller wavenumbers, at 1626, compared with vc=016 at 1658.

Fl=O18 was prepared by irradiating a solution of 5.3 mg of DAF and 0.1 mL of
TMU in 0.8 mL deuterated acetonitrile degassed by 5 freeze—pump-thaw cycles. After the
irradiation, the crude reaction mixture was separated by column chromatography, using a

mixture of hexane : methylene chloride, 5:1 and was analyzed by GC-MS. .

2.7 Conclusions

Carbene-to-carbene oxygen atom transfer has been demonstrated by isotopic labeling and
by observation of products from the newly generated carbene. Dimers of the newly
generated carbene byproduct have been detected by chemiluminescence. Rate constants for
oxygen abstraction by Fl: from various donors, including ureas, have been determined.

Whether these highly exothermic oxygen transfers occur in a single step or via the

60

intermediacy of ylide or oxirane intermediates are questions that we address by theoretical
and experimental (LFP) means. Qualitative LFP transient absorption suggests the presence
of new species. We tentatively suggest that the new species is the diaminocarbene formed
as a result of the oxygen atom transfer. However, the possibility of an ylide intermediate
which may absorb in the same region is not excluded. Our initial studies indicate that LFP
will be a valuable tool in rate studies of oxygen transfer reaction. Theoretical calculations
describe detailed ab initio molecular orbital studies of the remarkably convoluted potential
energy surfaces for carbene-to-carbene oxygen transfer reactions. The degenerate and cross

reactions present the same pattern: the possibility of two competing paths, through an ylide-

like intermediate or transition state and through an oxirane. For the cases of :CFZ and

:C(NH2)2, ab initio calculations find the ylide-type structure to be a TS for oxygen atom

transfer, rather than a stable intermediate.

61

 

2.8 References

l

a) Skell, P. S.; Plonka, J. H. J. Am. Chem. Soc. 1970, 92, 836. b) Dewar, M J. 8.;
Nelson, D. J.; Shevlin, P. B.; Biesiada, K., A. J. Am. Chem. Soc. 1981, 103, 2802.
c) Ahmed, S. N., Shevlin, P. B.J. Am. Chem. Soc. 1983, 105, 6488.

a) Shevlin, P. B.; Wolf, A. P. Tetrahedron Lett. 1970, 3987. b) Rahman, M;
Shevlin, P. B.Tetrahedron Lett. 1985, 26, 2959. c) Fox, J. M; Gillen Scacheri, J.
E.;Jones, K. G. L.; Jones, M Jr.; Shevlin, P. 8.; Armstrong, B. M.; Sztyrbicka, R.
Tetrahedron Lett. 1992, 33, 5021. (1) Armstrong, B. M.; McKee, M. L.; Shevlin, P.
B. J. Am. Chem. Soc. 1995, 117, 3688.

Donor groups strongly stabilize carbenes, even to the point of their being isolable:

a) Wanzlick, H. W. Angew. Chem. Int. Ed. Engl. 1962, 1, 75. b) Arduengo III, A.
J.; Harlow, R. L.; Kline, M J. Am. Chem. Soc. 1991, 113, 361. c) Dixon, D. A.;
Arduengo HI, A. 1.]. Phys. Chem. 1991, 95, 4180. d) Arduengo III, A. J.;Dias, R.
H. V.; Harlow, R. L.; Kline, M. J. Am. Chem. Soc. 1992, 114, 5530.

Carbenes commonly abstract monovalent atoms or groups: broadly, triplet carbenes
abstract hydrogen and singlet carbenes halogen. Such processes yield radical pairs
which may undergo rapid cage coupling, disproportionate, or diffuse apart. a) Kirmse,
W. Carbene Chemistry; 2nd ed.; Academic Press: New York, 1971. b) Roth, H. D.
Ibid 1977, 10, 85. c) Platz, M S. Acc. Chem. Res. 1988, 21, 236. d) Platz, M S.;
Maloney, V. M In Kinetics and Spectroscopy of Cw'benes and Bircdicals; M S. Platz,
Ed.; Plenum Press: New York, 1990; Chapter 8.

62

 

10

a) Field, K. W.; Schuster, G. B. J. Org. Chem. 1988, 53, 4000. b) Schweitzer, E.,
E.;O’Neill, G. J.J. Org. Chem. 1963, 28, 2460.

Triplet diphenylcarbene deoxygenates nitroxides to give quantitative yields of
benzophenone. This process is preferred even when OH groups, usually effective
carbene traps, are present in the substrate. a) Casal, H. L.; Werstiuk, N. H; Scaiano, J.
C. J. Org. Chem. 1984, 49, 5214. b) Clark, K. B.; Battacharyya- K.; Das, P. K.;
Scaiano, J. C. Ibid. 1992, 57, 3706.

Shields, C. J.; Schuster, G. B. Tetrahedron Lett. 1987, 28, 853.

a) Prakash G. K. 8.; Ellis, R. W.; Felberg, J. D.; Olah, G. A. J. Am. Chem. Soc.
1986, 108, 1341. b) Sander, W. Angew. Chem. Int. Ed. Engl. 1990, 29, 344.

L’Esperance, R. R; Ford, T. M.; Jones, M.; Jr. J. Am. Chem. Soc. 1988, 110, 209.

a) Shimizu, N .; Bartlett, P. D. J. Am. Chem. Soc. 1978, 100, 4260. b) Nikolaev, V.
A; Korobitsyna, I. K. Mendeleev Chem. J. 1979, 24, 88. c) Wong, P. C.; Griller,
D.; Scaiano, J. C. J. Am. Chem. Soc. 1982, 104, 6631. (1) Liu, M T. H;
Soundararajan, N.; Anand, S., M; Ibata, T. Tetrahedron Lett. 1987, 1011. e) Ibata,
T.;Liu, M. T. H.; Toyoda, J. Ibr'd. 1986, 27, 4383. f) Liu, M. T. H.; Ibata, T. J.

63

 

ll

12

l3

14

15

16

Am. Chem. Soc. 1987, 112, 774. g) Ibata, T.;Toyoda, J.; Liu, M. T. H.Chem. Lett.
1987, 2135.

Chateauneuf, J. E.;Liu, M.T.H. J. Am. Chem. Soc. 1991, 113, 6585.

a) Padwa, A.; Dean, D. C.; Zhi, L. J. Am. Chem. Soc. 1989, 111, 6451. b) Padwa,
A.; Zhi, L. J. Am. Chem. Soc. 1990, 112, 2037. c) Padwa, A.; Hombruckle, S. F.
Chem. Rev, 1991, 91, 263 and references therein.

a) Janulis, E. P.; Arduengo III, A. J ., J. Am. Chem. Soc. 1983, 105, 3563.
b) Janulis, E. P.; Arduengo III, A. J ., J. Am. Chem. Soc. 1983, 105, 5929.

a) Trozzolo, A. M.; Leslie, T. M.; Sarpotdar, A. S.; Small, R. D.; Feraudi, G. J.;
DoMinh, T.; Hartless, R. L. Pure & App. Chem. 1979, 51, 261. b) Arnold, D. R.;
Kamischky, L. A. J. Am. Chem. Soc. 1970, 98, 1404. See also ref 6b and 10c.

a) Bekhazi, M.; Warkentin, J. J. Am. Chem. Soc. 1981, 103, 2473. b) Bekhazi, M.;
Warkentin, J. J. Org. Chem. 1982, 47, 4870. c) Zoghbi, M; Warkentin, J. J. Org.
Chem. 1991, 56, 3214. d) El-Saidi, M; Kassam, K.; Pole, D. L.; Tadey, T.;
Warkentin, J. J. Am. Chem. Soc. 1992, 114, 8751. e) Wong, T.; Warkentin, J.;
Terlouw, J. K. Int. J. Mass Spectrom. Ion Processes, 1992, 115, 33.

Moss, R. A.; Joyce, Y. Y. J. Am. Chem. Soc. 1978, 100, 4479.

64

 

17

18

19

20

21

22

23

24

Graham, W. H. J. Am. Chem. Soc. 1965, 87, 4369.

Griller, D.; Hadel, L.; Nazran, A. S.; Platz, M. S.; Wong, P. C.; Savino, J. C.;
Scaiano, J. C. J. Am. Chem. Soc., 1984, 106, 2227 and references therein.

a) Jackson, J. E.; Platz, M in Advances in Cm'bene Chemistry, Brinker, U. H. JAI
Press LTD.; Greenwich, CT, 1994, pp. 89-160. b) Moss, R. A.; Ibid. p59.

Platz, M. S. in ref. 4d, pp. 298-300 and references therein.

Herrrnann, W. A.; Kocher, C. Angew. Chem. Int. Ed. Engl. 1997, 36, 2162

and references therein.

Ming-Shee Lee Ph.D. Dissertation, MSU 1993.

Du, X. M.; Fan, H.; Goodman, J. L.; Kesselmayer, M. A.; Krogh—Jespersen, K.;
LaVilla, J. A.; Moss, R. A.; Shen, S.; Sheridan, R. S. J. Am. Chem. Soc., 1990,
112, 1920.

The meajurements were performed using SPEX Triplemate 1877 with 150g/mm
gratings and a PAR OMAY 1024 x 256 detector, LN 2/UV enhanced CCD. The spectra

are the result of 305 accumulations.

65

 

25 a) Wiberg -von, N .; Buchler, J. W. Z. Naturforschg. 19645 2055b) Winberg, H. E.;
Camham, J. E.; Coffman, D. D.; Brown, M. J. Am. Chem. Soc., 1965, 87, 2054.

26 The expected product of 180 transfer, labeled Fl=130, was detected via GC-MS of the
reaction mixture. In the experiments to date, the fluorenone was found as a mixture of
Fl=180 and Fl=160 in a 1:3 ratio. We believe that Fl=16O results from reaction of
traces of 1602 with F]: either in the reaction mixture or (from residual DAF) in the

injector of the GC-MS.

 

I gnaw." '- 1 r

27 Photolyses were run in dry acetonitrile solvent for all oxygen donors except dimethyl

carbonate, sulfolane, and trimethyl phosphate, which were run neat.

28 Singlet -triplet gap for DPC was found to be solvent dependent and estimated at 3
kcal/mol. a) Hadel, L. M.; Platz, M. S.; Scaiano, C. J. J. Am. Chem. Soc. 1984,
106, 283. b) Savino, T. G.; Senthilnathan, V. P.; Platz, M. S. Tetrahedron, 1986,
42, 2167.

29 Analysis of benzophenone/benzhydryl methyl ether product ratios by 1H NMR
indicated that for DPC in acetonitrile, kpNo~ kMeOH = 2.4 x 107 M'IS‘I; see Griller,

D.; Nazran, A. S.; Scaiano, J. C., J. Am. Chem. Soc. 1984, 106, 198.

66

‘5

5
Q
N

fil-
‘

«Us. .

 

30 Turro, N. J.; Cha, Y.; Scaiano, J. C. J. Am. Chem. Soc. 1987, 109, 2101.

31

32

33

34

35

36

37

Wesdemiotis, C.; Feng, R.; Danis, P. 0.; Williams, E. R. McLafferty, F. W. J. Am.
Chem. Soc. 1986, 108, 5847.

Gonzales, C; Restrepo-Cossio, A.; Marquez, M.; Wiberg, K. B. J. Am. Chem.
Soc. 1996, 118, 5408.

a) Pliego, J. P., Jr., DeAlmeida, W. B. Chem. Phys. Lett. 1996, 249, 136.

b) Pliego, J. P., Jr., DeAlmeida, W. B. J. Phys. Chem, 1996, 100, 12410.

c) Pliego, J. P., Jr., Franca, M. A.; DeAlmeida, W. B. Chem. Phys. Lett. 1998,
258, 121.

a) Rose, T. L.; Fuqua, P. J. J. Am. Chem. Soc. 1976, 98, 6988. b) Bradley, J. N .;
Ledwith, A. J. Am. Chem. Soc. 1963, 86, 3480; c) Reference 9 and references

therein.

Pliego, J. P., Jr., DeAlmeida, W. B. J. Chem. Phys. 1997, 106, 3582.

LaVilla, J. A.; Goodman, J. L. Tet. Lett. 1988, 29, 2623.

Houk, K. N.; Rondan, N. G.; Santiago, C.; Gallo, C. J.; Gandour, R. W.; Griffin,
G. W. J. Am. Chem. Soc. 1980, 102, 1504.

67

(.1:

 

38

39

40

41

42

43

45

46

47

48

49

50

51.

See Table 12, p. 142 and Table 14, p. 145 in ref 18.

Soundararajan, N.; Jackson, J. E.; Platz, M. S. Tetrahedron Lett. 1989, 29, 3419.

PCC:-nitro ylide absorbtion in Reference 19a.

Chateauneuf, J. E.; Liu, M.T.H. J. Am. Chem. Soc. 1990, 112, 744.

Chateauneuf, J. E.; Liu, M.T.H. J. Chem. Soc. Chem. Comm. 1991, 1575.

Sueda, T.; Nagaoka, T.; Goto, S.; Ochiai, M. J. Am. Chem. Soc. 1996, 118, 10141.

Oku, A. J. Chem. Soc. Perkin Trans. 2, 1996, 725.

Platz, M. S.; Olson, D. R. J. Phys. Org. Chem. 1996, 9, 689.

Mahler, W. J. Am. Chem. Soc. 1968, 90, 523.

Mahler, W.; Resnik, P. R. J. Fluorine Chemistry, 1973/74, 3, 451.

a) It may be used as clean sources for :CF2 or other perfluorinated carbenes, by

pyrolysis Mahler W., ref 16 & 17 above.

a) Volatron, F. Can. J. Chem. 1984, 62, 1502, and references therein.

a) Griffin, G. W. J. Am. Chem. Soc. 1970, 92, 1402. b) Griffin, W. Angew. Chem.
Int. Ed. Engl. 1971, 10, 537. c) Huisgen, R. Angew. Chem. Int. Ed. Engl. 1977,
16, 572. d) Trozzolo, A. M.; Leslie, T. M.; Sarpotdar, A. 8.; Small, R.D.; Ferraudi,
G. J. Pure & Appl. Chem. 1979, 51, 261.

a) Feller, D.; Erikson, D. J. Am. Chem. S 0c. 1984, 3700. b) reference 57.

68

(H -\ . If!» q f FRI»
o‘. o‘ . 0‘» -\ -\l~

 

52

53

54

55

56

57

58

a) Warkentin, J.; J. Am. Chem. Soc. 1981, 103, 2473. b) J. Org. Chem, 1991, 56,
3214; Can. J. Chem. 1992, 70, 2967; Liebigs Ann. 1995, 1907.

Foresman, J. B.; Frisch, A. Exploring Chemistry with Electronic Structure Methods: A

Guide to Using Gaussian, Gaussian, Inc: Pittsburg, 1993.

Steward, J. J. P. MOPAC, A Semi-Empirical Molecular Orbital Program, QCPE, 455,
1 9 8 3 .

Gaussian 94 (Revision D.3), Frisch, M. J .; Trucks,G. W.; Schlegel, H. B.; Gill, P.
M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Peterson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.;
Ortiz, J. V.; Foresman, J. B.; Cioslowski, J .; Stefanov, B. B.; Nanayakkara, A.;
Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J.
L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.;
Defrees, D. J .; Baker, J .; Steward, J. P.; Head-Gordon, M.; Gonzales, C. and P0ple,
J. A.; Gaussian, Inc.; Pittsburg PA; 1995.

Moss, R. A; Acc. Chem. Res. 1989, 22, 15; b. Moss, R. A; Acc. Chem. Res. 1980,
13, 58. The reactivity of carbene intermediates may be predicted using the philicity
scale mCXY; values for mCXY are either determined empirically based on selectivity

of carbene toward olefines or calculated using O'R+ and 01 values.

Schaefer, H. F. et al. J. Am. Chem. Soc. 1993, 115, 5790.

Experimental data in ref 46 and 47 describe thermal decomposition of tetrafluorooxirane

at ~500 °C to generate :CF2

69

 

59 a). Clark, J. S.; Dosseter, A. G.; Wittingham, W. G.; Tetrahedron Lett. 1996, 37,
605. b). Olson, D.R.; Platz, M. S. J. Phys. Org. Chem. 1996, 9, 759.

60 a) Kistiakowski, G. B. et al. J. Am. Chem. Soc. 1958, 80, 1066. b) Milligan, D. E.;
Jacox, M. E. J. Chem. Phys. 1962, 36, 2911. c) Laufer, A. et al. Chem. Phys.
Letters, 1977, 46, 1151. d).Hsu, D. et al. Intern. J. Chem. Kinetics, 1977, 507.

e) Chapter 3, this work, for a complete review of the reaction of methylene with C02.
61 Chapter 2, sections 2.1 to 2.4, this work.
62 Ref 28.
63 Photolytic C-C cleavage in oxiranes, Reference 17.
64 The laser system was kindly provided by David Modarelli at University of Akron.
65 DAF synthesis in Moss, R. A.; Joyce, Y. Y. J. Am. Chem. Soc. 1978, 100, 4479.

66 DAM synthesis from Vogel, A. Vogel's Practical Organic Chemistry including
qualitaitve organic analysis, 4th ed, John Wiley & Sons, Inc., 605 Third
Avenue, New York, NY 10158, 1978, p289.

70

\

CHAPTER 3:

A Theoretical Investigation of Double Bonded Oxygen Atom Abstraction on
a Model Reaction: CH2 + C02 ---> CH20 + CO

Abstract .................................................................................................................. 70
3.1 Introduction ..................................................................................................... 70
3.2 Experimental background ................................................................................ 72
3.3 Methods and procedures .................................................................................. 77
3.4 Discussion ........................................................................................................ 78
3.5 Conclusions .................................................................................................... 103
3.6 References ...................................................................................................... 104
Abstract:

Ab initio G2 calculated pathways are presented for the reaction CH2 + C02 --->
CH20 + C0, in which net transfer of a double bonded oxygen atom occurs from C02 to
the carbene. Of particular interest are the electronic state of the attacking methylene, the
structure of the possible intermediates and the lowest energy path(s) available for this
reaction. As expected, our results support the assignment of OL—lactone 1 as the
intermediate observed by IR in the matrix isolation experiments of Milligan and Jacox
(Reference 4); analogous reactions involving substituted carbenes have more recently
been reported by Sander et al. We obtain AHf(l) = -43.7 kcal/mol at the G2 level while a
variety of isodesmic reactions point to slightly higher values (-40 to -42 kcal/mol).
Acyclic °CH20(CO)° (methylen—oxycarbonyl) and °CH2C02° (acetoxyl) biradicals 2 and
3, respectively, were also considered on both singlet and triplet potential energy surfaces.

According to the calculations, the singlet reaction proceeds with little or no barrier

to form 1; subsequent ring fragmentation (AHi = 27.4 kcal/mol) yields the products,

71

CH20 + C0. Collision orientation must play a role, however; Wagner et al. have
reported that reaction is only half as fast as collisional deactivation of l:CHz to 3:CH2
which presumably occurs via nonproductive encounter geometries. An activated channel
(AHi' = 24.5 kcal/mol) was also located in which l:CH2 directly abstracts oxygen from
C02 via an ylide-like TS, 12. The lowest energy 3:CH2 + C02 attack forms the triplet
acetoxyl diradical 33 and a higher energy path leads to methylene—oxycarbonyl diradical
32; no path for isomerization of 33 to 32 was found. Barriers for these two processes are
AH?t =20.l kcal/mol and AH1 =58.9 kcal/mol.

Attempts to locate regions on the triplet approach surface where the singlet
crosses to become the lower energy spin state were complicated by the difficulty of
optimizing geometries within the composite G2 model. Preliminary efforts however,
indicate that such crossings occur at geometries higher in energy than separated 1:CH2 +
C02, suggesting that their role should be relatively unimportant in the chemistry of this

reaction.

72

.1.)

0’1?

I“

3.1 Introduction

We present here a mechanistic analysis of the gas phase reaction of methylene,

ICHZ, with carbon dioxide, C02.1

C02 + :CHZ —'_’ C0 + O=CH2 (Reaction 1)

This reaction represents the first reported abstraction of a double bonded oxygen
atom by a carbene. Our own recent studies of oxygen transfer from carbonyl compounds
to carbenes2 focused our attention on the mechanism of such processes and led us to the
present computational effort. Of particular interest is the potential existence and the
electronic nature of a single-step “pluck” reaction3 in which pairs of bonds are
simultaneously broken and formed. The results also shed light on recent suggestions of

ylide formation, H2C:‘---+0=C=O.

3.2 Background

Reaction 1 was first observed in 1958 when Kistiakowsky et al., studying the
reaction of methylene generated from ketene, noted excess C0 production when C02 was
used as an “inert” buffer gas. They proposed that the attack of triplet methylene on C02

forms an intermediate, or—lactone (1), which decomposes to formaldehyde and CO.

0 o /O 5 0 S)\c=o
0A1}? ”2.0/ \9/ ”2‘5/ \C/ Hc/
2

1 32 l2 33

Figure 3.1 Potential intermediates in the :CH2 + C02 reaction

73

A few years later, in 1962, Milligan and Jacox4 photolyzed diazomethane in a
C02 matrix, at ca. 50 K. Monitoring the methylene generated in the matrix, the authors
observed the appearance of a species whose IR absorption bands grow as those from the
initially generated :CH2 decrease. The new IR spectrum shows a carbonyl stretch shifted
toward higher frequencies, over 1900 cm'1.5 The species was tentatively assigned as the
a-lactone 1 in which the carbonyl stretch is shifted due to the ring strain, but the
possibility of acyclic diradical intermediates such as 32 or 33 was not ruled out.

In 1977, almost two decades after Kistiakowski’s initial experiment, Laufer and
Bass6 studied the kinetics of reaction 1 by flash photolysis with CH2N2 and CH2CO as
methylene precursors. The rate constant obtained for reaction 1 was 3.3 x 10'14
cm3molecule'13'1, based on known values for triplet methylene dimerization (k = 5.3 x
10'11 cm3molecule'18'1) and reaction with acetylene (k = 7.5 x 10'12 cm3molecule'IS'1).
Although the latter rate constant has since been revised,7 Laufer’s results correctly
describe a slow reaction between methylene and C02.

In the same year, Hsu and Lin3 determined the vibrational energy of the carbon
monoxide released from reaction 1, monitoring its production with a continuous wave C0
laser. Methylene was generated by photolysis of CH212 (A > 210 nm). The vibrational
state population of CO produced was found to be close to that predicted by statistical
calculations, assuming a long lived CH2C02 intermediate,9 and a total final vibrational
energy of 63 kcal/mol, the exothermicity of the reaction. Such an assumption minimizes
contributions from reactions slower than reaction 1. The authors found that CO is
vibrationally excited up to v = 4, carrying an average of 1.9 kcal/mol. By extrapolation to
earlier reported reaction rates the authors assigned the singlet electronic state to the
reacting methylene but cited Laufer's then unpublished work, indicating that triplet
methylene might be also responsible for some oxygen abstraction.

Sophisticated new techniques allowed Wagner et al.10 in 1990, to perform

detailed kinetic measurements on reaction 1. Methylene, generated by laser photolysis of

74

ketene, was observed by laser induced fluorescence (1CH2) and laser magnetic resonance
(3CH2) allowing quantitative evaluation of intersystem crossing (ISC). The authors found
that roughly 2/3 of l:CH2 + C02 collisions relax l:CHz to 3:CH2; for the remaining 1/3,
three exothermic reaction channels were considered: formation of OCH-CHO (glyoxal),
H2 + 2C0, or H200 + CO. Preliminary experiments in Wagner's lab showed that 3:CH2
+ C02 react very slowly even at higher temperatures11 and the products may be only the
result of collisional activation of 3:CH2 to 1:CH2 which then reacts rapidly with C02.
Wagner concluded that formaldehyde and C0 are "the probable chemical products", and
that a-lactone is a "plausible intermediate".

Based on DeMore’s observation12 that diphenyldiazomethane photolysis in the
presence of 02 generates benzophenone, Sander13 investigated :CPhg as an "oxygen
abstractor" from C02 doped into Ar or Xe matrices. The observed intermediate has a high
frequency carbonyl stretch (1890 cm‘l) and fragments, on UV irradiation (A > 220 nm),
to Ph2CO and CO. The intermediate is assigned as diphenyloxiranone 4. The spin
forbidden reaction of the ground state triplet :CPh2 with C02 to form 4 shows a thermal
barrier, occurring only when the matrix is annealed to 35 K.14 Photoexcited :Cth,
however, reacts rapidly above 10 K to yield the same products, perhaps due to the
availability of a low lying empty p orbital as in the singlet case. No direct evidence was
obtained for diradical or zwitterionic intermediates such as 5 or 6, the diphenyl analogs of

20r 3.

thc

4 5 6

Figure 3.2 Potential intermediates in the :CPh2 + C02 reaction

75

To clarify the carbene characteristics responsible for oxygen abstraction, Sander
et al. used four carbenes, ‘5 two with triplet and two with singlet ground states. 16 The
authors found no indication that the reactivity is influenced by the spin state. Instead, the
"philicity" of the carbene appears to control the oxygen abstraction via rate determining
nucleophilic attack on the C02 carbon. With the data available it was impossible to
determine whether the reaction is concerted or not.

Reaction 1 has also been probed by Chateuneuf 17 in a laser flash photolysis
(LFP) study of :CPh2 in supercritical C02. The :Cth, generated by photolysis from
diphenyldiazomethane forms "most likely" diphenyl a-lactone, though no direct
observation of the product was reported. The state selectivity of the carbene was also
unclear, no differentiation being made between the "spin allowed" concerted addition of
the singlet or the "spin forbidden" formation of the lactone through the triplet manifold,
via a diradical intermediate.

In a theoretical study by Davidson et al. ‘8 exploring gas phase dissociation of
chloroacetyl anion, three intermediates were suggested as products of anion departure: 0t-
lactone (1 ), acetoxyl diradical (33) and a zwitterionic state of the
dioxatrimethylenemethane structure 3. Calculations found the lactone to be the most
stable,19 with the diradical 35 kcal/mol higher in energy; the zwitterion collapses upon
optimization to or-lactone or separated :CH2 and C02. Davidson's calculations yield a
heat of formation for or-lactone 1 of -51 kcal/mol. This value is in reasonable agreement
with the experimental data of Squires et al. 20 obtained by collision-induced dissociation
measurements on several acetate anions; it differs significantly, however, from a more
recent computational estimate (-45.2 i 2.5 kcal/mol via two different isodesmic
reactions)21 and from our G2 value (vide infra). In a theoretical study of or-hydroxy-
carboxylic acid decarboxylation,22 a-lactone is found to fragment to CHzO + CO with an

Ea and TS structure very similar to our own (vide infra). Squires et al. 23 investigated the

76

formation of (Jr-lactone by tandem mass spectrometry and energy-resolved collision-
induced dissociation and determined the its heat of formation to be -47 i 4.7 kcal/mol

To address the question of methylene's spin state in reaction 1, we began our
theoretical investigation by looking at several modes of methylene attack on C02 on both
the singlet 1:CH2 and triplet 32CH2 potential energy surfaces. We envisioned three
possible paths (Figure 3.3):

a. electrophilic attack at the oxygen lone pair of electrons, leading to an ylide-like
intermediate 12 ; the possibility of direct oxygen transfer via a 12—like TS to yield C0 +
H2C0 was investigated along with further cyclization of 12 to 1;

b. ambiphilic attack at a C=0 1t bond, leading to the formation of oc—lactone 1;
fragmentation of 1 generates the final products C0 + HzCO;

c. nucleophilic or radical attack at the C atom, leading to acetoxyl diradical 33;

further transformation of 33 to 1 was considered, it being difficult to envision a direct

path from 33 to C0 + HzCO.
+ 0
H25/ F c/
9 ,1 : 12 or32 ‘
9993'" 1
ll) "98th:: 0 \
:anE ---------- » 02—} ----------- *C0+|'12C=0
' “eff/,0 1
‘ O\c=o

Figure 3.1 Possible reaction paths

In the course of this work other questions were addressed: is there a direct path, concerted
or stepwise, via an open ylide-like structure 12 (transition state or intermediate) to C0 +

H2CO? Is oc-lactone 1 the observed intermediate in the early matrix studies? If so, is it

formed in one step (path b), or via cyclization of other species, such as 12 or 33? If attack

77

occurs at carbon (path c), are there stable acetoxyl—like intermediates 33? If so, would
they close to (rt-lactone 1 or rearrange to diradicals 32? If 32CH2 reacts with C02 to any
significant degree, at what point does intersystem crossing occur en route to closed—shell

products 1 or C0 + H2CO?

3.3 Methods and procedures

All calculations were performed using the Gaussian 94 package24 run on a cluster
of Silicon Graphics computers. All stationary points (NImag=0), were optimized and
characterized by vibrational analysis at the HF/6-31G* and MP2/6-31G* level. Single
point calculations were run at MP4/6-3lG* using HF/6-31G* and MP2/6-3lG*
geometries, respectively. Transition structures (TS, NImag=1) were characterized by one
single imaginary vibrational frequency.25

The connections on the PES among the stationary points and TSs found were
verified by running intrinsic reaction coordinate (IRC) calculations, 26 starting from each
TS at the HF/6-31G*//HF/6-31G* and MP2/6-31G*//MP2/6-3lG* levels. Selected IRC
points were analyzed as single points at MP2/6-31G*//HF/6-31G* and MP4/6-
31G*//HF/6-31G* levels. To better understand the presence or absence of stationary
points on the reaction path from singlet :CHz approaching C02, B3LYP/6-31G* and
QCISD/6-31G* optimizations were also carried out for this reagion of the singlet PES.

Although well—defined economical methods are available for accurate
calculations of carbene S—T gaps,27 their general extrapolation to reaction potential
energy surfaces is nontrivial. In an effort to obtain experimentally relevant energetics, we
computed 02 energies for all stationary points. The G228 error between experimental and
calculated values29 for the singlet—triplet (S-T) gap in methylene is only 2.4 kcal/mol,
much closer than with the other above mentioned methods. Also, because of significant
differences in the results obtained at different levels (vide infra), key points on the

reaction paths were reevaluated using the G2 method.

78

3.4 Discussion

3.4.1 Overall thermochemistry

The experimental heat of reaction 1, AHr = -61 kcal/mol, 30 is compared in Table

1 to data from our ab initio calculations. The overall energy profiles, as calculated here at

the three different levels indicated above, may be followed in Figures 3.4a, 3.4b and,

3.40. Interestingly, G2 and RHF, the most and least sophisticated of the methods used,

approximate the experimental heat of reaction most closely. The overall thermochemistry

suggests also that the reaction should end on the singlet PES.

----- Paths on Singlet PES

 

ll

 

Paths on Triplet PES

AE (kcal/mol) T84

G2
.' 3H2c=0+co
1 ‘ m-
cnz+co2 ' 22.5
—, :f,‘
6. "~~.
o
3 CH2+002

|
\
\

 

1H20=O+CO
L.—
-54

\

Figure 3.4a Overall energy profiles at G2 (AB in kcal/mol)

79

 

A AE (kcanol)
MP2/6-31 G‘l/MP2/6-31 G"
T84
T86 ....... 50.4 . 4o 3 3H2C=O+CO
1:CH CO / ,-.--\-"33 \JX‘T
2+-—2‘\7:7\ 33.4
17.6 31.2 19. 7 193 ‘.\
%~. ‘. T32
0 ~ \ ‘ ‘4 2'17“.
3:CH2+C02
““s 1 a” \\ ‘4 1
‘-' ‘u‘ H2C=O+CO
“34.1 \h

-47.2
Figure 3.4b Overall energy profiles at MP2/6-31G*(AE in kcal/mol)

ll
AE (kcanol)
HF/6-31G*

 

 

\ 1H20=O+CO

-51.5
Figure 3.4c Overall energy profiles at I-IF/6-31G* (AB in kcal/mol)

Though singlet carbene is generally thought to be the initial product formed from all
methylene precursors used,31 collision—induced equilibration to the triplet ground state

occurs readily, raising the possibility of intersystem crossing (ISC) before the CHz-COz

collision or somewhere along the reaction path.32

80

Table 3.1: Overall Thermochemistry

Method Aern Aern
singlet PES triplet PES

 

I-IF/6-31G* 58.7 15.2
MP2/6-3 lG*//HF/6-31G* -69.8 32.7
MP4/6-31G*//HF/6-31G* -69.3 24.5
MP2/6-3 1G*/1MPZ/6-31G* -64.8‘=l 5135‘
G2 -60.5 22.5

Experimental -61 23.8b

 

a Corrected with zero point energy, ZPE. b Calculated from experimental
heat of formation of triplet methylene (ref 30) and triplet formaldehyde
(from G. Hertzberg, “Electronic Spectra of Polyatomic Molecules” van

Nostrand, Princeton, 1967).

3.4.2 Methylene approach

The possible existence of stable van der Waals complexes was investigated for
reactants lC1 (Figure 3.5b), lsz (Figure 3%) and products, le (Figure 3.75b). In the
singlet methylene case, a minimum was found at the HF level, 3.2 kcal/mol lower in
energy than separated species. In structure lC1, (Figure 3.5b) the methylene is complexed
through the empty it orbital to a lone pair of one of the C02 oxygen atoms. IRC
calculations at the HF level find 1C1 to be the starting point for a one-step oxygen atom
"pluck" pathway, which occurs via a substantial barrier, TS3. As expected, this finding
suggests that direct oxygen transfer should begin via electrophilic attack on an oxygen
atom lone pair. Despite extensive searching, no conventional CHg-COZ ylide minimum
was located, casting doubt on a recent report from the group of Oku et al.33 This result is
unsurprising; C02 is a weak Lewis base, as evidenced by its low proton and methyl
cation affinities, compared to CHzO (129.2 and 49.4 vs. 170.4 and 78.5 kcal/mol,34
respectively), so its affinity toward methylene would be expected to be substantially

lower than formaldehyde affinity toward methylene. A second minimum (NImag=0) also

81

found only at the HF level corresponds to perpendicular nucleophilic methylene attack on
the C02 carbon (1C2v in Figure 3.9, path b). The HF/6-31G* IRC calculations connect
this minimum with a-lactone 1 via TS], over a barrier of 3.5 kcal/mol (i.e. 0.2 kcal/mol
above separated 12CH2 + C02). The lsz structure orients the methylene hydrogens
perpendicular relative to the C02 moiety, the CC distance is 3.131 A, 0.7 A longer than
in 1C1, and its energy is 3.3 kcal/mol below l:CH2 + C02, similar to the 3.2 kcal/mol
value for 1C1.

No G2 calculations were carried out for 1C1 , T8], or 1C 2v because no
corresponding minima could be located on the MP2/6-31G* PES. In all three cases,
reoptimization at the MP2 level simply falls onto the monotonically exothermic trajectory
of a concerted barrierless attack by l:CH2 on the 1: system of C02 to form or-lactone 1.
Similar behavior was found on the B3LYP/6-31G* PES, but at the QCISD/6-31G* level,
true stationary points for 1C1 and T8] were found, 4.3 and 3.6 kcal/mol below the energy
of l:CH2 + C02. Interestingly, for the 1C1 geometry, the triplet single point energy is 19
kcal/mol higher than the reactants on the triplet surface. Considering that methylene's
QCISD/6-31G* S-T gap (optimized geometries) is 16 kcal/mol and assuming two
monotonically parallel channels for singlet and triplet approach toward C02, with the
triplet complex 19 kcal/mol higher in energy (same geometry) we may consider that even
though the reaction starts on the triplet PES it may have crossed to the singlet PES before
the methylene gets close enough to the reaction partner, C02.

For the triplet case, calculations at HF, MP2, B3LYP and QCISD/6-31G* levels
yield structures for complex 3C1 with C-C distances ranging from 3.5 to 3.2 A, and
energies 0.6 to 1.1 kcal/mol below 3:CH2 + C02. G2 calculations starting from the MP2
optimized geometry find an energy only 0.8 kcal/mol lower than the separated species.
IRC data connect this species via T84 (Figure 3.4a) over a 40 kcal/mol barrier to

diradical 32.

82

Table 3.2: Calculated energies for C02 + :CHz complexes

 

AE (kcal/mol)a HF MP2 QCISD B3LYP
Singlet 1C2v -3.1 -3.3

Singlet 1C1 -3.3 -4.3

Triplet 3C1 -0.6 -1.1 -l.1 -1.1

 

3 Calculated at the same level of theory, in respect with C02 + 1:CH2 or 31CH2.

According to our calculations, the nature of the attack depends on the electronic state of
carbene as expected from the findings of Wagner et al. and seemingly at odds with the
philicity—dominated suggestion of Sander et al. However, since philicity is defined in
terms of a carbene's substrate selectivities rather than the absolute reactivity (i.e. rate
constants) of the carbene's more reactive singlet state, this may be a less significant
discrepancy than it appears at first. The singlet prefers electrophilic attack on the n
system of C02 with the formation of a-lactone 1 in a barrierless process, whereas the
triplet prefers to attack at the carbon atom as a radical and follows the lowest energy path

toward the acetoxyl diradical 33.

3.4.3 Possible intermediates

Among potential intermediates in this reaction, we have examined or-lactone 1,
ylide 12, diradical 32 and acetoxyl diradical 33 (Figure 3.1). The singlet zwitterionic state
of structure 3 was not considered. Although early experimental studies in solution
invoked such a dipolar species formed by ring opening of substituted a-lactones to
explain formation of polyester products,35 previous calculations of Davidson et al. (see
ref 18) found the corresponding structure for the parent Ot-lactone to be higher in energy
than the corresponding diradicals, and to collapse without barrier to 1.36 Table 3.3 lists
energies in kcal/mol relative to 3:CH2 + C02 for each intermediate. Vibrational frequency
data are presented in Table 3.5 and a complete list of the total energies and cartesian

coordinates for each species considered here are available in the Appendix 3.

83

3.4.3.1 Ylide 12 (= TS3 )

No minimum was found for singlet l2 , the hypothetical ylide. Only a TS is found
in this region of the PES, at all level of calculations used here. If the attack of singlet
methylene is directed toward oxygen, the reaction path goes through this species (named
here TS3 or 12) and effects an one step oxygen "pluck". The barrier found in our
calculation for direct one-step oxygen abstraction via TS3 is 62.5 chmol at the HF
level but significantly lower at MP2 or G2 (see Figure 4.3) with respect to 1:CH2 + C02.

The geometry of 12 changes dramatically when calcualted at the HF and MP2
levels (Figure 3.7). The geometry at I-IF/6-31G* looks like a late TS, with the
formaldehyde part almost formed, with an HCH angle close to 1200 (1240). The MP2/6-
31G* structure suggests an early TS with the methylene part almost unmodified from its
starting geometry. At the same time, the C2—O3—C4 distances are located in the bonding
range, indicating strong interactions. IRC calculations, both at HF and MP2 levels,
confirm the connectivity between the reagents and the products in one single step, via
TS3 on the singlet PES of the reaction.

A single point energy calculation for vertical excitation of the 32 diradical from
the triplet to singlet electronic structure finds the singlet with the exact same geometry as
32, 15.2 kcal/mol higher in energy than 12 at the HF level but only 6.6 kcal/mol higher at
MP4/6-31G*. Considering that the initial S-T gap in methylene at HF is ~30 kcal/mol,
the difference between the two calculated structures, 32 diradical and ylide-like 12 is
much smaller.

In our endeavor to find a minimum with an ylide-like structure we calculated the
open shell version of an ylide starting from both the singlet transition state 12 and the
triplet minimum 32 geometries. In both cases, using UHF/6-31G* and UMP2/6-31G*
wavefunctions, the calculations found the same minimum with an open shell singlet

structure.37 The minimum is lower in energy than the singlet TS3 and its geometry is

84

essentially that of the triplet 32 geometry. The extent of spin contamination in these
calculations supports this idea, being so great that we believe the result simply

corresponds to the triplet diradical.

Table 3.3 Energy (kcal/mol)3 of stationary points

 

Ab initio method 1:CH2 1 31 12 32a 325b 33 p 33 5
C02

NImagc 0 1 0 0 0 1

HF/6-31G* 30.8 -6.9 53.4 89.8 22.0 19.9 1.2 10.0

HF/6-31G* with zpe corr. 30.5 0.1 58.7 93.0 25.2 24.2 8.1 13.3
MP2/6-31G*//HF/6-31G* 21.0 —40.6 55.4 31.2 5.2 4.7 14.3 22.1
MP4/6-31G*//HF/6-31G* 17.2 —40.0 50.0 27.7 3.6 3.4 8.4 16.5

MP2/6-31G*//MP2/6-31G* 20.9 -40.3 40.1 15.8 15.5 17.0 24.0
MP2/6-31G* with zpe corr. 17.6 -34.1 43.4 19.7 16.6 21.2 27.0
MP4/6-31G*//MP2/6-31G* 17.1 -39.7 36.2 14.2 14.2 11.7 18.8
MP2(full)/6-31G* 19.0 -40.9 39.8 15.8 15.5 16.9 24.1
QCISD(T,4ET)/6-311G(dp) 12.5 -36.7 36.8 13.3 13.3 9.2 18.1
G2 (0 K) 6.6 -39.9 31.1 12.5 13.2 9.15 17.9

 

a Energies are reported as difference value in respect with 3:CH2 + C02; at each corresponding level of
theory, as follows: I-IF/6-31G*, -226.55568; with zpe correction, -226.52450; MP2/6-31G*//MP2/6-31G*,
-227.1 1114 with zpe correcton -227.08l65; MP4/6-3lG*//MP2/6-31G* -227.152407; MP2(full)/6-31G*,
-227.1258165; QCISD(T, 4ET)/6-311G(d,p), -227.269984; G2, -227.4303l. b Acetoxyl staggered structure
optimized at HF/6-3 1G* level goes to alpha-lactone (see text). C NImag = the number of imaginary
frequencies obtained by vibrational analysis at bot HF and MP2 levels; structures with Nimg = 0 represent

stationary points on PESs, whereas NImag =1 indicates a transition state.

3.4.3.2 (rt-lactone 1
Most often mentioned as a possible intermediate, (rt-lactone 1 is characterized here

by geometry optimization and vibrational analysis (see Tables 3.3 and 3.4 and Figure

3.9).38 The first estimated value available in the literature for the heat of formation of or—

lactone,39 based on bond additivity, is -31 kcal/mol. Davidson‘s ab initio data18 suggest

85

AHf (1) = ~51 kcal/mol. This value was successfully used by Squires to interpret
collisional induced dissociation (CID) measurements on acetyl anion.20 More recently
Rodriguez et al. combined QCISD(T) calculations with known heats of formation of
other species in two different isodesmic reactions and estimated a heat of formation for
a-lactone of 45.4 i 2.4 kcal/mol. Recent measurements of Squires et al.23 found AHf (1)
= -47.3 d: 4.7 kcal/mol. Our own estimates, calculated using GZ reaction energies and
experimental40 heats of formation for known species, are shown in Table 3.4.

The ketene + oxirane isodesrrric reaction we used (see the fourth reaction in Table
3.4) suggests, from calculated heats of reaction at the G2 level, combined with the
experimental AHf of ethene, oxirane and ketene, AHf (l) = -41.5 kcal/mol. Heats of
formation directly calculated for ketene and oxirane from G2 atomization energies yield
values higher than those experimentally measured. In contrast, for ethene the calculated
value is 0.3 kcal/mol lower. This observation suggests that heat of formation for 0t-
lactone 1 may be slightly higher than the value calculated here, -43.7 kcal/mol, and

certainly within the error limits suggested by Rodriquez et al. (-45 i 2.4 kcal/mol).

86

00200, 00003000 NO 9 00020 0:: 05 008:0 03 £03088 0: 000: 0:200— :00 0.005000.— UBB 0.690 05 E 030=0>0 0030> 05 89m 0

.02.— .memowm- .mewhHm- .Nmeng-m- H00w0 nodm 003-an- .H-vQ-aéb- .momwo.E-- UO 3.? .H-oowmd- €080.0- .m.o-HHH 50:508: :2 5:05.098
56:88:00 05 0:0 Mwom :0 00:00:00 300:0 N0 .M0 :0 00:00:00 300:0 NO 00505 30:0 N0 56:88:00 05 m0 890 :000 :8 320p 0000:

00 M mom .0 00850.0 mo 380 .3 £53 5805.5: 505.0 .H. .<.H.<QOU :85 38300 .00: 0:0 300:0 0:8: 0:00 55 8:00:00 0.890% 55:50:00 0:0
800: .9: .33 50:0 0.le .H. .-H 50.00102 .2 500305—20 ”.4. 50¢ ”90. 0:20.022 0:0 HNNH. .00 .3a 50:0 .034 .H. .< H. 0391.3 .0 00—00;. ”.VH
.t050>0:w0m ”.< ...H .0550 :55 3.05:0 NO 55 8:002:90 00 800: 56:88:00 05 80¢ :3 0000—00—00 3 .0000.0& ”SH £802-00 00:52 00090000
00:0:0H0M 0:00:00m .092 0:0 .2096 N H .wwan S05 0% 50:0 0.me .H. :xbfifiozoopcofi 05:02 0:0 :3 00.0.3 000.. ._0 .0 0.05 Bob .54 H050Etoaxm H0

 

00.003 2. £0:- 8980 00:00 AI 0 + :0 + 00
2.0903:- 20.2- 3.893 04:00 Al 0 + E. + 00
$0: 8.2 35:80 £00 Al 5 + 00
£0..- 3882-80 80:00 AI 00 + :0 + 00
on:

H.0Hm.moH- H.Hv- 9m?- hoév me N: A V All OHL :

/o\

I
0.0.3.00- 0.9.- new 0.00 3.9. . 0 1| 010 9:
9 mm /o\

00.0: ”20000.2- 0:.- 0. z.- 80- 8.0- " JV ll.- 01 "0.: fl

2.008.00- nvodflmvom- 0.00- 000- 0.2 00.2 $080028 mommmu AI GUN: + 00
28.093.8- 20.00 5.0. 0.00 0.:4 0.00- 3889.3 80:00 AI N8 + £9
08.008.8- ” ES 0:1 30 0.0? 0.0..- 388980 80:00 Al N8 + 0:9

$900: No 808 3.05:: 608
90 054 034 034 NO 5:34 S:H2 :03 088030 0:08:00 9 :00: 0:03000m

 

8:005:80 00 800: 0:0 08:000.: 0:50:09 80.: 00900.8 :2 in 0E0H.

87

Of the three intermediates considered here, 1 is the most stable and is formed
through attack of singlet methylene on a C=O 1: bond of C02 in a barrierless process.
Ring fragmentation, 1 --> C0 + H2CO, is calculated (GZ) to occur via T82 with AHi =
28.4 kcal/mol and AHrxn = -14.1 kcal/mol, similar to the 32.2 and -12.2 kcal/mol MPZ/6-
31G** values obtained by Domingos et al.22 in their theoretical study of gas-phase 0t-
hydroxyacid decarboxylation. These Aern values differ from Liebman’s -23.4 kcal/mol
estimated exothermicity39 as expected, given the above adjustments in the estimated AHf
of the (It-lactone. Ring-opening (path b) to zwitterionic intermediates has been proposed
to explain polyester products in solution studies of a-lactone-forming reactions (ref. 35
and references therein). However, as noted above, all singlet biradical structures we
examined fragmented or collapsed to 1 upon optimization. Thus, 1 represents a fairly
deep potential energy well, and should be easily observed at sufficiently low
temperatures. The IR data attributed by Milli gan and Jacox to 1 show 12C=O and l3C=O
carbonyl stretches at 1967 and 1933 cm‘l, respectively, and as shown in Table 3.5, scaled
HF/6-31G* and MP2/6-31G* IR frequencies match experiment rather well both for l and
for CHZO and CO.

The triplet a-lactone excited state, 31, was examined as part of our exploration of
the triplet PES and ISC possibilities. The HF/6-31G* structure lies 60 kcal/mol higher in
energy than the corresponding singlet and shows substantial pyramidalization at the
carbonyl carbon; the MP4/6-31G*//HF/6-31G* energy difference is even larger, at 90
kcal/mol. Upon reoptimization at MP2/6-3IG*, the ring opens via C-C cleavage to form
diradical 32 (anti isomer), 52.4 kcal/mol higher in energy than 11 at the GZ level. These
results indicate that 31 is energetically out of reach and therefore unlikely to play any role

in the title reaction.

88

3.4.3.3 Diradical 32

Infrared studies of the reaction of HO- radicals with CO 41 indicate the existence
of linear H—O—CO in two conformations, syn and anti. As expected, our calculations find
two analogous conformations for the triplet OHZC—O—COO diradical 32 (Figure 3.5). The
energy differences between the two conformers are small at all levels (from 1.2 kcal/mol
at HF/6-BIG* to 0.6 kcal/mol at 62). On the conformational potential energy surface; for
328, (Figure 3.5), barriers of 7 and 9 kcal/mol were found for anti-->syn conversion and

for rotation of the CH2 group of 328, respectively.

0 Carbon

0 Oxygen
0 Hydrogen

AE (kcal/mol)
HF/6-31G'
MP2/6-31 G'l/M P2/6-31 G

  

 

 

Figure 3.5 32a - 32$ interconversion path

IRC calculations connect 32a with separate 3CH2 + C02 via TS4. Thus, the 12.5 kcal/mol
endothermic addition to form 32a would occur with a barrier of 59 kcal/mol; subsequent
cleavage to triplet formaldehyde and CO, via TSS (Figure 3.7) is again endothermic by
10 kcal/mol, over a 17 kcal/mol barrier (Figure 3.8).

The matrix experiments of Milligan and Jacox located the carbonyl stretches for

anti and syn HO—CO° species at 1883 cm‘1 and 1793 cm"1 respectively. Our calculations

89

(Table 3.5) found the corresponding vibrations for 32a at 1811 cm'1 (corrected, MP2 see
footnote to Table 3.5) and for 323 at 1781 cm‘1 (corrected, MPZ) at the MP2/6-31G*
level. These frequencies are substantially lower than those for a-lactone and reflect the
acyclic connectivity in 32. Similarly, our values of 126.3° and 129.9° for the O—C—O
angles in 32a and 325 compare well with the 126.7° and 130.l° for HOC0- calculated by
Rauk et al.42 As shown in Table 3.6, AHf(3za) was calculated to be 9.6 kcal/mol at the
G2 level, 3.8 kcal/mol lower than the value obtained by summing the separate enthalpy
changes for loss of hydrogen atoms from methyl formate to make a "noninteracting"
diradical. This energy lowering can be understood as the stabilization due to interaction

between the two unpaired electrons in 32a.

Table 3.6 Estimated heats of formation for the diradical intermediates

 

AHf (kcal/mol) Experimentala Calculated Calculatede

 

H3C—CO-OH -1032 103.2c
H3c—co—o- -45.4c
-H2C—CO—OH -58.1c

33 -H2C—CO—O- -0.3d 5.8
H—CO—OCH3 -85.0 4387*)
oco—OCH3 -35.3b
H—CO—OCHg- -40.ob

32 -CO—OCH2- 13.4d 9.6

 

aAHf from Lias et al. in Reference a, Table 3.4. bAHf of diradicals calculated from their heat of
atomization using 62 enthalpy (see same ref as in Table 3.4) and experimental AHf atoms corrected with
zpe and heat capacity, CODATA, J. Chem. Thermodynamics 1978, 10, 903. cAHf from Yu, D.; Rank, A.;
Armstrong, D. A. J. Chem. Soc. Perkin Trans. 2 1994, 2207-2215; the value calculated with our data for
acetic acid is 1.6 kcal/mol higher than our value. dAHf calculated considering the loss of hydrogen atom
enthalpy as an additive property. eAHf of diradicals calculated from their heat of atomization (*CHz—O—
CO- --> 2C + 2H + 20, AHf= 9.6 kcal/mol and OCHz-CO—O- --> 2C + 2H + 20, AHf = 5.8 kcal/mol)
using 02 enthalpy and experimental AHf atoms corrected with zpe and heat capacity 43of elements, at 298
K. Note: The difference between the AHf of the two radicals, 32 and 33 is 3.4 kcal/mol calculated with 02
total energy vs. 3.8 kcal/mol with G2 enthalpy.

9O

3.4.3.4 Acetoxyl diradical 33

Addition of 3CH2 to the carbon of C02 leads to 33, the acetoxyl diradical, which
has been computationally analyzed (ref. 18) in a somewhat different context. Using the 6-
31G* basis set (compared with 6-31+G* from ref 18), we also found two structures,
planar 33p and staggered 338. At the MP2 and HF levels, only the planar diradical is a
minimum (NImag = O). The planar and staggered structures of 33 are higher in energy
than 3:CH2 + C02, by 9.2 kcal/mol and 17.9 kcal/mol (G2), respectively. Geometrical
data compare well with reported MP2/6-31G* structures of monoradicals °CH2-COOH
and CH3—CO—O0; energies from the same source (see Rauk et al. in ref 42) were used in
Table 3.6 to calculate the AHf of diradical 33. Unlike the case of 32a, the value predicted
assuming additivity, -O.3 kcal/mol, is substantially lower than the 5.8 kcal/mol calculated
from G2 heats of atomization, suggesting that the two unpaired electrons interact to raise
the energy of the entire entity.

Of the two possible 3CH2 + C02 adducts, 33p and 32a, the former is both lower in
energy (by 3.3 kcal/mol) and accessible via a lower barrier (20.1 vs. 58.9 kcal/mol via
T86 vs. TS4, respectively, Fig 3.4a). However, completion of the oxygen transfer to form
3CH20 and CO would require isomerization over a high barrier (via TS7 found only at
HP level) to 32a, before the cleavage (via TSS) which itself has a barrier of 17.9 kcal/mol
(G2). The above connectivities on the reaction PES were all confirmed through IRC
calculations. According to these results, 3:CH2 + C02 could react at thermal energies to
form acetoxyl diradical 33, but this intermediate's principal choice would be to
redissociate or undergo ISC to the singlet surface, followed by barrierless closure to 1. As
expected, optimization on the singlet surface starting from the geometry of 333 (CS
symmetry) led directly to a-lactone 1, as found by Davidson.18 Singlet optimization (Cs

symmetry) from the analogous planar geometry 33p leads to a transition structure for

91

0H2C—COz- bond rotation (NImag=l), 17.9 kcal/mol (GZ) higher in energy than 3:CH2 +
C02.

The acetoxyl diradical was recently studied23 by tandem mass spectrometry and
by energy resolved collisional induced dissociation. From the appearance energy
measurements, a gap of 2 kcal/mol was estimated between the triplet and its lowest
singlet state. The triplet lowest energy path is decomposition to 3:CH2 + C02 while the
open shell singlet is thought to easily evolve to "hot" a-lactone which, in its turn,
decomposes to C0 + HzCO, species detected in the NRMS experiments; their presence is
explained only by ring opening of a-lactone.

Comparing the available singlet and triplet data, the region of the PES
surrounding 3 appears well suited for ISC. Single point energy calculations for possible
singlets with 3p and 3s geometries show the singlet l3s much lower in energy than the

corresponding triplet and even lower than the energy of starting compounds on the singlet

PES (Figure 3.6).

 

1
E" “c’ 43.4
3% 52
8 338 g 333
3 1795 g C02+1ICH2 3318.1
002+1:CH2 JR. .5 — —L 13
>_‘ 12.5 92 3
— +- ' _
6.6 9'15 55 COZ+3:CH2 7.6
0 E: 0
COz+3ICH2 5’2
0
C
1 1
_ -36.7
-40.0

 

Figure 3.6 T—S direct excitation results for 33p and 33s
Mote: At QCISD/6—31G* the singlet spin calculations indicate 3’35 lower in energy than both 33p and 33s
and even lower than C02 + 12CH2 on the singlet PBS.

92

Rotation from the planar acetoxyl diradical 33p to the staggered 33s, described as a TS
by the calculation, offers a possible channel for ISC to the singlet 133 which may "roll"

downhill into (at-lactone, on the singlet PES.

3.4.4 Vibrational analysis from IR calculated data

IR vibrational frequency calculations contain known systematic errors, due to the
neglect of electron correlation and anharrnonicity. The results give, generally, an
overestimate of ~10%-12%. The values reported here in Table 3.5 are corrected by a
factor of 0.89 44 and 0.95445 for the HF and MP2 values respectively. The most intense
vibration of the experimental IR spectrum is the band at ~1967 cm'1 for 12C or 1933 cm“
1 for 13C with the isotopes in C02. This band is assigned to the carbonyl stretch in l with
the abnormally high frequency attributed to the strain in the a—lactone ring. The deviation
of our computed values from the experimental ones is much smaller than in the case of
formaldehyde.46 From all the possible intermediates considered here, only 1 has a very
strong vibration located close to the experimental detected value. The other candidates, 32
and 33, also show carbonyl stretch vibrations, but at lower frequencies (see Table 3.5).
The only other available experimental data are for the diphenyl a—lactone. Three
vibrations attributed to the C=O stretch were detected”, with the strongest located at
1890 cm'1 and shifting to 1837 cm'1 when 13C was used. Our SE/PM3 calculations locate
the carbonyl vibration for diphenyl oc-lactone at 1990 cm'1 (uncorrected) for 12C.

These vibrational analyses support the assignment of the observed bands in the

matrix experiments to oc-lactone 1.

93

r-

19

II
.1“

L?

«.1.

Table 3.5 Selected calculated IR frequencies

 

 

Compound VC=O (Gm'l) VC-O (cm'l) HF
di-Methyl or-lactonea 1900

di-(Trifluoro) Ot-lactonea 1975

di-n-Butyl-OL-lactonea 1895 1 163

l (or-lactone)b 1974 HI:; 1967 MP2 1210

3211b 1879 HI3; 1811 MP2 1210

33pb 1667 HF; MP2 1225

HzCO (experimental) 1746

HZCO (calculated) 1804 HF

 

a From Chapman, 0. L.; Wojtkowski, P. W.; Adam, W.; Rodriquez, O.; Rucktaschel, R. J. Am. Chem. Soc.
1972, 94, 1365-1367. b Values corrected with 0.89 and 0.954 for HP and MP2, respectively.

3.4.5 Possible reaction paths

The connectivities among the found intermediates were established through the
various transition structures found, and are presented in Schemes 5 to 8 below.47 For the
geometry data and the active vibrations see Appendix 3. Their energies with respect to
the triplet methylene and C02 are presented in Table 3.7. The energy profiles reflect the
tremendous differences among the calculated paths at HF/6-3lG*, MP2/6-31G*/fMP2/6-
31G*, and G2 levels, and are presented separately in Fig 3.4.

path a, singlet PES: This is the direct one—step path for an oxygen atom "pluck"
(Figure 3.7). It takes place via TS3, a structure proposed for ylide 12 (for which no
minimum was found). The energy barrier is 62.5 kcal/mol at HF/6-31G*//HF/6-31G"‘ but
drops dramatically to 24.5 kcal/mol at the G2 level. IRC calculations confirm the
connection from TS3 to separated reactants and to products CO + CHZO on both HF and

MP2 PESs.

94

. Carbon
0 Oxygen
0 Hydrogen
1
AE (kcaI/mol)
MP2/l6-31G'

 

..
O
O
O
O
O
C
O
O
O
O
'
C
O
O
I
C
O
O
G
C
O
O
O
O
O
'0
G

 

 

‘ mac-mea

Q
Q
\
—

29.6
(42.2)

Figure 3.7a Path a on singlet PBS at MP2/6-31G*//MP2/6-31G*

0 Carbon

0 Oxygen
I Hydrogen

AE (kcal/mol)

 

 

HF/eaie'
‘ZCHTOCO;
'Ci
so '——-‘
6 26.7
o a.
0
e
‘H.c-0+co
-. 1c, ,—
—
1.
-52.1 '5 5

   

 

Figure 3.7b Path a on singlet PES at HF/6-31G*

95

Attack

Clem

0131.1

CH;-

(.J-J
KC)
~_.

path a, triplet PES: This is a two-step process via diradical 32 (Figure 3.8).
Attack of triplet methylene on an oxygen atom of C02 leads to 32 via TS4; subsequent
cleavage via TSS leads to C0 + 3CH20. The high barrier and substantial endothermicity
of the initial attack (58.9 and 12.5 kcal/mol, respectively) are expected, reflecting the
difficulty of breaking a C20 double bond. Two rotation isomers (anti and syn) of 32 were
found, which interconvert via TSrot (Figure 3.5) with a lower barrier than either of the
fragmentation channels. With high barriers and an excited state product (3CH20), this
overall path, as described, is endothermic and unlikely to play any significant role in the
CH2 + C02 reaction.

. Carbon
0 Oxygen
0 Hydrogen

ll
AE (kcal/mol)
GZ

 

’ :cmof"

 

 

Figure 3.8 Path a on triplet PES

path b, singlet PES: This is the two-step oxygen transfer via or-lactone 1 (Figure
3.9). At the HF/6-31G* level, the computed reaction path begins in the shallow CH2 +

C02 van der Waals complex 1C1 and surmounts the 3.5 kcal/mol barrier of TSl to arrive

at a-lactone 1, the most stable CszOz intermediate found by our calculations. At the

MP2/6-31G*//MP2/6—31G* level, 1C1 and T8] are no longer stationary points, allowing

96

separate CH2 + C02 to collapse without barrier to a-lactone 1. Fragmentation of 1 via
TSZ then yields C0 + CHzO with a G2 barrier of 36.1 kcal/mol (MP2) or 27.3 kcal/mo]
(G2). Because of the difference in the results at different levels of theory we further
examined the two structures, 1C1 and T81, using DFT and QCI methods.48 At the
B3LYP/6-3lG* level, both structures collapsed directly to (Jr-lactone 1, but the QCISD/6—
3 1G* model found 1C1 to be a minimum 4.3 kcal/mol lower in energy than 1CH2 and
C02; TSl was then a transition state only 1.3 kcal higher than 1C1 but 3 kcal/mo] lower
than the starting species. The S—T methylene gap at QCISD is 16 kcal/mol. Because of

these contradictory results, the region around TSl on the singlet PES appears to be a

possible site for ISC (vide infra).

A E (kcanol)
62 1
:CHvCO,

 

 

Figure 3.9a Path b on singlet PES at G2

97

path c. I
We calm
3TH: all
diradlcii]
it isome
diradical
157, of
higher 1:
theory u
Would re

Wham

4
’AE (kcal/mol)
HF/6-31G'

 

 

 

 

Figure 3% Path b on singlet PBS at HF/6-31G*
path c, triplet PES: This is a three—step process via acetoxyl diradical 33 (Figure 3.10).
We calculate it to be the lowest energy path for 3:CH2 attack on C02. At the G2 level,
3CH2 attacks C02 at carbon via T86 with a 20.1 kcal/mo] barrier to form planar acetoxyl
diradical 33p. However, 33p can not complete the oxygen transfer of interest here, unless
it isomerizes to 32 or closes to or-lactone 1. An isomerization path from 33p to the
diradical 32 was found only at the HF/6-31G* level and involves passage over a barrier,
TS7, of 86.1 kcal/mo]. Despite substantial effort, no TS7-like structure was found at
higher levels, suggesting that T87 may be an artifact due to the limitations of the HF
theory used. Even if such a path could be followed, completion of the oxygen transfer

would require passage over the additional barrier, T85 already presented in the context of

path a on the triplet PBS.

98

 

%
ll '
AE (kcal/mol)
62 ,fl,
/ 20.1 ‘

 

3:Cl-l2 4602/"
—"

 

0
AE (kcal/mol)
HF/6-31G*

 

 

   

Figure 3.10b Path c on triplet PBS at HF/6-31G*

Direct closure of acetoxyl diradical to or-lactone 1 may occur only if ISC connects the

triplet 33p with the corresponding singlet structure. Optimization of a singlet starting at

99

the ram

those 01

3.4.6 In

direct c
.\1P4 It
singlet
min.
regon
the ex
cones
lht [\l

probu'

the geometry of 33p evolves directly into a-lactone 1. In this respect our results parallel

those of Davidson et al. 18

3.4.6 Intersystem crossing.

Selected IRC points describing all the above reaction paths were subjected to
direct excitation calculations.49 Single point energies were calculated at the MP2 and
MP4 levels using the HF optimized structures from the IRC calculations both from
singlet and triplet PESs. We switched each structure's spin state and looked for energy
variations along the paths. No crossing regions were found, with the exception of the
region around T81 on the singlet PBS and T86 on the triplet PES. A plot (Figure 3.11) of
the energy differences between selected points from both the singlet and the
corresponding triplet IRC trajectories of T81 and T86 show possible ISC in the region of
the two described transition states. Our way of approaching intersystem crossing is

probably verifiable with nontrivial procedures.

100

.II; r A\.::\:...v.¢» z >E-_.u:...~ 21:27.. .2.:...J:.:..~. \l N ..~u~...~-

$2605 05 5 0353958 ho 28m 2558—0 05 3033032 95382 05 5 2.3.80 05 .8 88m 05 @2365 m was fl
39882:; :23 N30” + «CO H m .m ”main... 8. a Bo: 2: E 829 com .mmonm + NOD mo 35:0 28m 95on 05 55 .0238 E .9538— E .oocobtfi 35cm a

 

SN .1; 3m in 0.2- No

839:

02 mam 08 men 3. 659.6860

4.8 a: 26 3m 3. *oseaggz

mhuEeow NU

3N 3A N2 3m 3w- *ozeagroaééz

Q: 3m EN 6% 0% 3m 2:- 6882:; .0398:
08 EN 0% 9:. 36 4.2. no- b39§é¢9n£§

hams—cow 9:2

a.: CNN 2% Nd“ mom mac v.5 as- a.» *O_m-c\mw§*0~m-o\§2
wwfi oém vow YNN wdm ch mgm Qm- 2: *Ofim-o\mm\\*o_m-c\mm2
v.3 ~.m~ mom mdm mom m.mw mow wém wdm .too 0% £5 #39033
“.mm o.wm m.vw mgm own.“ mew wdw wdm mém *Otmbim
baufieow m:

31? «Minn 31x 9...? Sims ml: 94 75 @8880

 

Hecmhm amhm hmhm emhm mmhm vmhm m9: NW: “ma;
€583: gm 8.3 8283 E 2.3.

101

Figure 3.11 Energy differences on IRC trajectories from T 81 and T82

S-T Energy Difference

Path c: HF‘, MP2’, MP4’; Path b: HF“, MP2“, MP4“, ocrso7

118T

 

 

>0

2!

+_. ._ _——._——— .;__-- H-.-

98 i
78 i l‘
.. 0
2 g 0
g 58 O x x
.2 0 X 00
., «l—
8 38 a )gfim
5 m"
2 t i x
I X
-22 _+____. +——»~—~--—-—-+ -———-—a———--- -+ —»— —~+~-—~~——- ._._-,
45.5 -4.5 -2.5 -o.5 1.5 3.5 5.5
Reaction coordinate (IRC data)
S-T Energy Differencea
Path c: HF‘, MP2’, MP4’; Path b: HP‘, MP2‘, MP4“, ocuso7
118 ‘r
|
98 i
78 l 3"
— ' O
2 . 0
T; 58 1’ 0 X X
TL“ ~ X 0
2:” 38 i 0 Q‘ § 8
18 " !
l 'i a 6 e a i
-2 T x
l X
-22 4.-.__._____. ~+-—-—a ‘ .
-25 -1.5 -05 0.5 1.5

Reaction coordinate (IRC data)

2.5

 

9 Seriest
O SeriesZ
Ammo
X Series4
X SeriesS
" SeriesG
,0 Series] j

 

 

0 Se ries1
0 SeriesZ
A Series3
X Se ries4
X SeriesS
" SeriesG l
{D Seriesll

 

 

 

“ Expanded plot around T81 and T86 PES's region (reaction coordinate = 0)

102

Wit

ETC

ll'

,
Pa

The direct excitation calculations mentioned above along with the monotonically down-
hill process on the singlet PES are analogous to the well known case of non-equilibrium
intersystem crossing for the :CPh2 reaction with methanol.50 If the singlet PES is
described as having a low or no barrier, as in the case of path b, the crossing with the
triplet PES which is going monotonically up—hill, as it is in the case of path c, may occur
before T86 on the triplet path is reached. The reaction starts on the triplet PBS and may
“roll“ down toward the products on the singlet PES (Figure 3.12) in a so called

nonequilibrium surface crossing.

CHaoH + ‘CPha CH30H +10%

002 + 30H2
Path c triplet
CO +3CH
CH OH 3Ph 2 2
3 + 2 CHgOH + 3th Path ctripl t
H H
P11230013 PhZCKOCHgt (fl
A B C

A: Traditional surface crossing; B: Non-equilibrium surface crossing in the reaction of diphenyl carbene
With methanol, adapted from Platz, M. S. in ref. 48, Fig 21, page 327. C: Possible non-equilibrium surface

crossing for the case of methylene reaction with CO2.

Figure 3.12 Traditional and non-equilibrium surface crossing

We propose here a possible intersystem crossing in the intersection region of path c with

path b before the state described as T81 at HF/6-31G* level, on path b, is reached.

103

3.5 Conclusions

Abstraction of oxygen by l:CH2 from CO2 appears to occur via stepwise
processes. Singlet methylene attack on the 1: bond of C02 to form (at-lactone was found to
be the most favorable process on the singlet potential energy surface, much like the
familiar barrierless concerted cycloaddition of methylene to alkenes. The net O-transfer is
completed by the lactone's fragmentation into CO from CHzO. A one-step oxygen atom
“pluck” pathway was also found, but with its 24 kcal/mol G2 activation barrier, this
process is unlikely to play a significant role at common reaction temperatures. The lowest
energy triplet attack is directed toward the carbon atom of C02 in a nucleophilic manner.
If the reaction remains on the triplet PES, the acetoxyl diradical 33 is formed but there is
a substantial barrier for this process and two further steps (via 32) are required to
complete the overall endothermic oxygen transfer on the triplet PES.

Experimental thermochemical data show reaction 1 (Table 1) on the singlet PES
to be 60 kcal/mol exothermic. The oc-lactone intermediate, with its high frequency (~1900
cm'l) carbonyl stretch, has also been experimentally observed in matrix isolation
experiments. These facts are in excellent agreement with our calculations. But the
correspondence between experiment and theory is less satisfactory for the available
kinetic data. Singlet methylene is the primary photolysis product from all the precursors
studied to date, but with C02 it is collisionally deactivated to the triplet state twice as fast
as it reactslov 11 The reaction of triplet methylene with C02 is slow, in agreement with
our calculated path c. Considering these experimental findings together with our
calculated results we propose here three possible parallel channels for the reaction of
methylene with CO2, all starting from the initially generated singlet methylene.

On one of the channels, part of the initially generated 1:CH2 forms, in a barrierless
process, the most stable intermediate, or-lactone 1, with its excess vibrational energy (the
reaction is 46.6 kcal/mo] exothermic), more than enough to allow fragmentation to

products under low pressure conditions. In a parallel channel, part of l:CH2 may be

104

(A;

-J

1")

energetically “hot” enough to overcome the barrier of one—step oxygen "pluck" process.51
However, most of the 1:CH2 will equilibrate by collisional deactivation with C02 to
3:CH2. In the third of channel, the reaction starts from the 3:CH2 by nucleophilic attack of
32CH2 at the carbon atom of CO2 (path c) to form acetoxyl diradical 33. On the triplet
PBS, the lowest energy path available to 33 is to fragment back to C02 and 3:CH2, but
ISC from path c to path b is a reasonable candidate for completion of the oxygen transfer

reaction. Once on the singlet PES, barrierless ring closure makes oc—lactone 1.

Supporting information available in Appendix. It contains a complete summary with
cartesian coordinates, calculated total energies and frequencies for all the species

discussed

 

3.6 References

p—n

a) Kistiakowski, G. B.; Sauer, K.J. Am. Chem. Soc. 1958, 86, 1066.

2 Kovacs, D; Lee, M-S.; Olson, D.; Jackson, J. E. J. Am. Chem. Soc. 1996, 118 , 1844.

b»)

a) Jackson, J. E.; Mock, G. B.; Tetef, M. L.; Zheng, G.-X.; Jones, M. Jr. Tetrahedron
1985,41, 1453-1464. b) Jackson, J. E.; Misslitz, U.; Jones, M. Jr.; De Meijere, A.
Tetrahedron 1987, 43, 653.

4 Milligan, D. E.; Jacox, M. E. J. Chem. Phys. 1962, 36, 2911.

5 For CH2N2 + CO2 l967cm‘1, CD2N2 + CO2 1961cm'1. CH2N2 + 13CO2 l933cm'1 "

6 Laufer, A. H.; Bass, A. M. Chem. Phys. Lett. 1977,46, 151.

7 According to recent measurement, this reaction has un upper limits k=10'14

105

10

 

10

11

12

13

14

15

16

17

18

19

20

21

cm3moleC'l x s", Darwin, D. C.; Moore, C. B. J. Phys Chem. 1995, 99, 13467.
Hsu, D. 8.; Lin, M. C. Int. J. Chem. Kinetics 1977, IX, 507.
Shortridge, R. G.; Liu, M. T. J. Chem. Phys. 1976, 64, 4076.

Koch, M.; Temps, F.; Wagener, R.; Wagner, H. Gg. Ber. Bunsenges. Phys. Chem,
1990, 94, 645.

Bohland, T. et al., unpublished results, cited in ref 10.
DeMore, W. B.; Pritchard, H. 0.; Davidson, N. J. Am. Chem. Soc. 1959, 81, 15874.
Sander, W. W. J. Org. Chem. 1989, 54, 4265.

Sander, W. W.; Patyk, A.; Bucher, G. J. Molec. Structure 1990, 222, 21.
Wierlacher, 8.; Sander, W.; Liu, M. T. H. J. Org. Chem. 1992, 57, 1051.

Triplet: diphenylcarbene and cyclohexadienylidene; singlet: phenyl chlorocarbene

and p-nitrophenyl chlorocarbene.
Chateauneuf, J. E. Res. Chem. Intenned. 1994, 20, 159.
Antolovic, D.; Shiner, V. J .; Davidson, E. R. J. Am. Chem. Soc.1988, 110, 1375.

The calculations reported in Reference 18 were performed using 3-21G+d+sp and

D95+d+sp wavefunctions.

Graul, S. T.; Squires, R. R. Int. J. Mass Spectrometry Ion Processing, 1990, 100,
785.

Rodriquez, C. F.; Williams, I. H. J. Chem. Soc., Perkin Trans. 2. 1997, 953.

106

1-

~\

If

)1

5’

 

22

23

24

25

26

27

28

Domingo, R. L.; Andres, J.; Moliner, V.; Safont, V. S. J. Am. Chem. Soc. 1997, 119,
6415.

Schroder, D.; Goldberg, N .; Zummack, W.; Schwarz, H.; Poutsma, J. C.; Squires, R.

R. Int. J. Mass Spectrometry Ion Processes 1997, I65, 71.

Gaussian 94 (Revision D.3), Frisch, M. J.; Trucks,G. W.; Schlegel, H. B.; Gill, P. M.
W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Peterson, G. A.;
Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J.
V.; Foresman, J. B.; Cioslowski, J .; Stefanov, B. B.; Nanayakkara, A.; Challacombe,
M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E.
8.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. 8.; Defrees, D. J.; Baker, J.;
Steward, J. P.; Head—Gordon, M.; Gonzales, C. and P0ple, J. A.; Gaussian, Inc.;
Pittsburg PA; 1995.

Detailed geometry and energy data for the species considered are available as

supporting informations in Appendix.

Foresman, J. B.; Frisch, A. Exploring Chemistry with Electronic Structure Methods:

A Guide to Using Gaussian, Gaussian, Inc: Pittsburgh, 1993.

Carter, E. A.; Goddard, W. A. J. Phys. Chem. 1988, 88, 1752.

Experimental S-T gap is 9 kcal/mo], as in Leopold, D. G.; Murray, K. K.; Miller, A.
E. S.; Lineberger, W. C. J. Chem. Phys, 1985, 83, 4849, or Brunker, P. R.; Jensen, P.;
Kraemer, W. P.; Beardsworth, R. J. Chem. Phys, 1986, 85, 3724, and references
therein; our calculated values at HF/6-31G* and MP2/6-31G* are 30 kcal/mo] and

21 kcal/mo], respectively.

107

‘1‘»

1‘

 

29

30

31

32

33

34

35

36

Curtis, L. A.; Raghavachari, K.; Trucks, G. W.; Pople, J. A. J. Chem. Phys. 1991, 94,
7221.

Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W.
G. “Gas Phase Ion and Neutral Thermochemistry” J. Phys. Chem. Ref Data 1988,
17, Suppl. 1.

Yamamoto, N.; Bemardi, F.; Bottoni. A.; Olivucci, M.; Robb, M. A.; Wilsey, S. J.
Am. Chem. Soc 1994, 116, 2064 and references therein.

"Kinetics and Spectroscopy of Crbenes and Birdicals" 1990 Plenum Press, New York,
Platz, M. 8, Chapter 8, pp 320.

In a recent communication Naito, I., Mizo, T., Kobayashi, S., Matsumoto, T. and
Oku, A. reported laser flash photolysis of (biphenyl-4-yl)chlorodiazirine in C02
saturated acetonitrile; a short-lived species with UV absorption at ca. 510 nm is

suggested to be the ylide formed by the attack of carbene on one of the oxygen atoms

of CO2.

Hunter, E. P.; Lias, S. G. “Proton Affinity Evaluation” in NIST Chemistry
Webbook, NIST Standard Reference Database Number 69, Eds. W. G. Mallard
and P. J. Lindstrom, March 1998, National Institute of Standards and Technology,
Gaithersburg, MD, 20899 (http://webbook.nist.gov).

a) Wheland, R.; Bartlett, P. D. J. Am. Chem. Soc. 1970, 92, 6057. b) Adam, W.;
Rucktaschel, R. J. Am. Chem. Soc. 1971, 93, 557. c) Chapman. O. L.; Wojtkowski, P.
W.; Adam, W.; Rodriquez, R.; Rucktaschel, R. J. Am. Chem. Soc. 1972, 94, 1365.

Experimental data about solution chemistry of substituted or-lactone are cited as

Doctoral Dissertation, McMullen D. F., Indiana University, 1982 in ref 18; the use

108

L; J

45

 

37

38

39

4O

41

42

43

44

45

of water and ethanol/water as solvents allowed isotope effect measurements in the 3-

methyl-or—lactone ring opening and indicate a zwitterionic intermediate which easily

decomposes into ethylene and C02.

Considering the limitations of UMP2 method usage strict interpretation of this results
may be misleading. For a discussion of the use of UMP methods on diradicals see

also Jarzecki, A. A.; Davidson, B. R. J. Phys Chem. A, 1998, 102, 4742.

For a review about stable a—lactones see L'abee G. Angew. Chem. Int. Ed. Engl.

1980, I9, 276 and references therein.
Liebman, J. F.; Greenberg, A. J. Org. Chem. 1974, 39, 123.

For a similar T82 a-lactone ring opening see ref. 22, Table 6, p 6420; C-C distance in
ref 22 is 1.506 vs. 1.500 Angstroms in the one optimized here at MP2/6-31G*.

Milligan, D. B.; Jacox, M. E. J. Chem. Phys 1971, 54, 927.
Yu, D.; Rauk, A.; Armstrong, D. A. J. Chem. Soc. Perkin Trans. 2 1994, 2207.
Pople, J. A.; Krishnan, R.; Schlegel, H. B.; DeFrees, D.; Brinkley, J. 8.; Frisch, M. J .;

Whiteside, R. F.; Hout, R. F.; Hehre, W. J. Int. J. Quantum Chem. Symp. 1981, 15,
269.

Hehre, W.; Radom, L.; Schleyer, P. R. v; Pople, J. A in "Ab Initio Molecular Orbital
Theory", John Willey and Son, 1986, p 233.

For comparison we used the formaldehyde carbonyl stretch. Computed at
HF/6-31G* it is 2027.6 cm‘1 which scaled by 0.89 gives 1804 cm'l; it presents a

deviation of Av = 58 cm'1 from the experimental value, 1746 cm'1 (see ref 14).

109

l-

l-

 

46

47

48

49

50

51

The active vibration is the one corresponding to Nimagz-l; the arrows clearly
indicate the evolution of each TS toward the products or intermediates with which it

is connected on PES (see the energy profile schemes).
CODATA in J. Chem. Thermodynamics 1978, 10, 903.

We used here B3LYP/6-31G* model from the DFT package, Parr, R. G.; Yang, W.
“Density-functional theory of atoms and molecules” Oxford Univ. Press: Oxford,
1989. For QCISD options see Pople, J. A.; Head-Gordon, M.; Raghavachari, K.;
Trucks, G. W. Chem Phys. Lett. 1989, 164, 185.

“Direct excitation” here is used to describe a state with exactly the same geometry of
a point found on the triplet PES calculated as the excited singlet state (Frank-Condon
principle) for comparison between the energies; the same procedure is applied for the

structures described as singlets which are excited to the corresponding triplet states.

“Kinetics and Spectroscopy of Carbenes and Biradicals” Platz, M. S. Plenum Press,

New York, NY, 1990, pagesz320.

For gas phase experiments Wagner et al. generated lCH2 from CH2N2 using a 337
nm laser pulse. The corresponding photon energy is 84.6 kcal/mol. Assuming an even
repartition of the energy at the formation of CH2 and N2 from diazomethane in
respect with their molecular masses (momenta) and considering the experimental AHf
data for CH2N2, CHzCO, 1CH2, N2, CO (51.3, -11.4, 101, 0, -26 in kcal/mo], from
ref 34) one may rationalise that, in the gas phase, with the corresponding energy from
the laser photons, singlet methylene will posses enough energy to overcome the 24.4

kcal/mo] (our calculated G2 value) for a one-step oxygen ‘pluck’ from C02.

110

CHAPTER 4

An Alternative Approach Toward Nucleophilic Carbenes Synthesis

Abstract ................................................................................................................ 1 l 1
1 Introduction .................................................................................................... l 12
2 Results and Discussion ................................................................................... 117
3 Conclusions .................................................................................................... 140
4.4 Experimental methods .................................................................................... 141
4.5 References ...................................................................................................... 143

111

1 Introduction

In the early '603 the idea that a carbene substituted with O, N, S or other electron-
donor groups would be more stable than the elusive carbene species known at that time
gained substantial attention. Wanzlich tried to synthesize such compounds by
deprotonation of the corresponding imidazolium ions, but only the formal products from
reaction of the desired carbenes with alcohols and aldehydes or the carbene dimers of the
carbenes were identified (Figure 1).1 It was almost thirty years later when a stable
carbene itself was first synthesized, isolated 2 and structurally characterized by X-ray

diffraction (Figure 2).3
Ph Ft
|
~ x”
E >2... 3
in lPh / N \N
| |
N 150° N Ph Ph
T -CHCI3 T W lPh
Ph Ph N
( >H —oca,
N
|

 

1960 Wanzlick

Ph

Figure 1 First synthesis of nucleophilic carbene dimers.

Ad 1991 Arduengo

Figure 2 1, 3-diadamantyl-imidazolin-ylidene.

112

Since 1991 several different groups around the world4 have contributed to the area of
stable, isolable carbene chemistry. The successful synthetic routes described so far
(Figure 3) for synthesis of stable carbenes involve either deprotonation of imidazolium or
thioimidazolium salts by strong bases,5 or desulfurization of imidazolthiones with
potassium.6 The former procedure proved effective even for synthesis of such species as
the first acyclic stable carbene derived from the N, N, N', N'-tetraisopropyl-

formamidinium ion by deprotonation.7

 

 

I R.
F1
| l
R N
N 1991
(9... x- Newest“: 9’ I}
R
N
| |
R' R.
l" l"
R N\ K,THF,A A R I N : 1993
I /C::S ' N Kuhn &Kratz
N Fl
R I I
R' ‘1'

Figure 3 General synthetic methods for preparation of nucleophilic carbene

In addition, special procedures were employed (Figure 4.4) for gas-phase and matrix

isolation.8

 

R' R'
|
- N 1997
I N\C —<O hv ‘ E \C: Maier
S / 0 Ar matrix ' S/ Terlow & Schwartz

Figure 4.4 Special cases for synthesis of nucleophilic carbenes in matrix

Dithiacarbenes, with two sulfur atoms stabilizing the carbenic center, were also thought

to be responsible for the formation of tetrathiafulvalenes, the corresponding dimers of the

113

1,3-dithiacarbenes.9 The path to the 1,3-dithiacarbene dimers generally involves the
reaction of carbon disulfide with an electron-deficient substituted acetylene (Figure 4.5).
Usually, most reactions of CS2 proceed from an initial nucleophilic attack at the carbon.
In contrast, the chemistry leading to 1,3-dithiacarbenes involves either electrophilic attack
of an electron deficient species on the sulfur atom of the C82 or a concerted

cycloaddition.

CF3

',' + l—-—» Cilia-<15;

 

3
CF3

Figure 4.5 Synthesis of dithiacarbene dimers.

The role of the triple bond may be extended to other highly reactive species.
Benzyne reacts in gas phase with C82 and the corresponding dimer of 1,3-dithiacarbene,
was identified by mass spectrometry.10 N akayama generated benzyne from different
precursors in the presence of carbon disulfide and methanol. 11 The product of the liquid
phase reaction, dithiacarbene, inserts into the H-0 bond of methanol, leading to 2-
methoxy-1,3-benzodithiole as the main product. N akayama's experiments demonstrated

the higher reactivity of benzyne toward C82 than methanol (Figure 4.6).

G)
.l + CS2—>[ ‘RC: 0 3\CC?
s/ S/
I + CH30H

0 V
s/ \OCH3

Figure 4.6 Benzyne cycloaddition to carbondisulfide.

 

 

114

Similarly, triple bonds from the highly strained tetramethylcycloheptyne and its sulfur
derivative, react with C82, at room temperature, to generate dimers of dithiacarbenes

(Figure 4.7).12

C82

OI
S G)
.I \C: I 5\Ce:
3/ s
S
or c=c< to

Figure 4.7 Cycloheptyne reaction with C82

t

 

<2”

The behavior of benzyne and cyclic alkynes, which do not possess electron—withdrawing
substituents, focuses attention on a possible direct cycloaddition reaction mechanism as
an alternative to electrophilic attack at the sulfur atom followed by rearrangement to the
carbene.

Hartzler thoroughly investigated the reaction of hexafluoro-2-butyne and
dimethylacetylenedicarboxylate (DMAD) with C82. At 100 °C, C82 added to the triple
bond to from three products (Figure 4.8). Tetrathiafulvalene, the dimer of the
dithiacarbene, is formed quantitatively only in the presence of a strong acid such as

trifluoroacetic acid. With no acid present, a 1:3 mixture of acetylene and C82 forms only

115

iii.

the

Ed.

"Ucl

l1 he

2% of the dimer, while the major product is an interesting adduct containing a 3:2 ratio of
the initial species obtained in 60% yield. A 1:6 ratio of acetylene to CS2 yields a different
major product, a 4:4 adduct in 60% yield while tetrathiafulvalene is obtained only in 8%
yield. In the presence of acid, the initially formed dithiacarbene probably does not react
further with the acetylene due to protonation of the dithiolium ion. It appears that it is not
the classical reaction between two singlet carbenes, but the reaction between the carbene

and its conjugated acid partner that forms the dimer.

F 3 S /8 F3
I 5>=C\s I 2 : 2 adduct
01:3

43:; £312.:
ll s>_ 3:2adduct
cr=3
F3 F3 $F3/ /S-S¥ Fag? S F3
| S>c= —c=c -i— —c’ ||
3-3/ \3 F3

F3
4 :4 adduct

Figure 4.8 Harzler general reaction with the three adducts.

The 1,3-dithiacarbenes were also intercepted with alcohols to form 2-alkoxy-1,3-
dithioles, phenols aldehydes and ketones. The expected product of carbene insertion into
the O-H bond of the acid was not observed. Propiolate esters also undergo these
reactions, proving that the presence of only one electron withdrawing group is sufficient
to activate the acetylene toward reaction with C82.

Beside C82, other cumulenes—carbonyl sulfide and diisopropylcarbodiimide—
were briefly tested for the ability to undergo the same type of chemistry and generate
nucleophilic carbene. No reaction was observed between C08 and hexafluoro-2-butyne;

when methanol was present in the reaction mixture, a slow reaction indicated the

116

presence of an initially formed zwitterion which is trapped by methanol. The case of
diisopropylcarbodiimide presents no evidence for chemistry similar to carbon disulfide
(Figure 4.9). Only the product of a [2+2] cycloaddition followed by a tautomerization

was observed.
1

F3 | F0
3
l r—’ EL —~ It;
N "aN
1 Y .. H;

Figure 4.9 Reaction of diisopropylcarbodiimide with hexafluoro-Z-butyne.

The brief investigation of carbodiimide by Hartzler has not extended to date. In terms of
simple bond energy calculations, the carbodiimide reaction with acetylenes should be less
energetically demanding than that of CS2. Considering the results obtained by Hartzler
with C82 two decades ago, and in the light of new examples demonstrating the stability
and reactivity of 1,3-diaminocarbenes from the last six years, we tried to extend the
investigation of carbodiimides as possible 411: donors as a pathway to new stable
nucleophilic carbenes. If the reaction proved to be efficient, a completely new route to

stable carbenes, involving reactions of carbodiimides with acetylenes might be found.

2 Results and Discussion
To find a new and effective route to the synthesis of nucleophilic carbenes, and to
better characterize and probe their physical and chemical properties, we investigated here

the reaction of DMAD with 1,3-di-tert-buty1carbodiimide (DBCDI)

117

 

2.1. Reagents

As cumulated double bond systems, we used 1,3-diisopropyl- and 1,3-di-tert-
butylcarbodiimide. Both appear to be reasonable candidates for generating 1,3-
diaminocarbenes. Diaminocarbenes with similar structures are already known13 as stable
species, in the absence of quenchers such as air, water, and alcohols. The only direct
precedent is the investigation performed by Harzler involving diisopropylcarbodiimide.
The isolated product seems to be the result of a [2+2] cycloaddition followed by
tautomerization made possible by the presence of the an or hydrogen atom from the
isopropyl group linked to the nitrogen atom (Figure 4.9). The 1,3-di-tert-
butylcarbodiimide was chosen for the present investigation because it has no hydrogen
atom in an 0t position. Thus, the possibility of a hydrogen shift or tautomerization, as in
the diisopropylcarbodiimide case, is eliminated. For economical reasons as much as for
the ease of manipulation, dimethyl-acetylenedicarboxylate, DMAD, was used as the
acetylene partner. In addition, brief investigations involving benzyne, maleic anhydride
(MA) and tetracyanoethylene (TCNE) were performed. However, the double bond
proved to be ineffective in the reaction with cumulenes. Neither maleic anhydride or
tetracyanoethylene reacts efficiently with carbodiimide in the manner under investigation
here. The only products isolated and characterized were obtained from the reaction of

carbodiimide with DMAD.

2.2. Investigation of reaction conditions
2.2.1. Energy requirements

According to the literature, the reaction of C82 with hexafluoro—Z-butyne or
dimethylacetylenedicarboxylate substituted acetylenes was slow, requiring heating at
100°C for several days to yield the products of the 1,3-dithiacarbene intermediate. In
contrast, benzyne reactions were fast at room temperature. In the liquid phase; their

dimeric products were formed within 30 minutes of the addition of CS2 to a solution of

118

 

benzyne prepared in situ. In the same manner, cycloheptyne reacts rapidly at room
temperature with carbon disulfide affording a 2:2 adduct, the corresponding dimer of the
carbene.

We investigated the energy requirements of the reactions involving CDI with
DMAD by using initiation methods such as light, heat, microwave or ultrasound. The
evolution of the reaction was followed by 1H and 13C NMR. The photochemical path
seems unlikely at best for the expected cycloaddition. Irradiation with UV light from a
500 W high pressure Hg vapor lamp shielded for wavelengths shorter than 350 nm with
an uranium filter or even unshielded proved ineffective. The reaction mixture remained
unchanged, by NMR and GC analysis, after 2 hours of continuos irradiation. Reactions
performed in an ultrasound water bath, were also unchanged after 2 hours. Heating in an
oil bath at 80 to 100 °C for several days led to partial consumption of the reagents, and to
drive the reaction to completion required heating for as long as 11 days. In contrast, a
sample exposed to microwaves in a sealed NMR tube required only half an hour for total
consumption of the reagents but afforded a more complicated mixture of products.

2.2.2 Solvents

The reactions were run in solvents, such as tetrahydrofuran, acetonitrile or ethyl
acetate, the media used in the initial experiments of Hartzler. Experiments with neat
liquid reagents are were also run, but a higher amount of polymeric products were
obtained and separation of the products was more laborious. The three solvents proved
equally effective in homogenizing the reaction mixture and no differences were noticed in
the course of the reaction. The use of deuterated THF or acetonitrile allowed the reaction
to be followed by NMR. Most characterized diaminocarbenes had their 13C NMR spectra
taken in THF, so it should be possible to make reasonable guesses about the expected

resonance position of the carbenic carbon of our expected species in the same solvent.

119

2.2 3 Traps for carbenes

To detect the presence of the carbene as a short lived intermediate, we tried to trap
it with some of the commonly known fast reacting species such as water, methanol or
benzaldehyde. Products with the appropriate masses were observed via GC-MS but
further confirmation of their structure was not possible because of the small amounts of
product.
2.3 Products

Reaction of DBQDI with DMAD= l to 2 mL of 103 M solution of DBCDI and
DMAD in THF-d8, EtOAc or CD3CN in NMR tube were degassed by freeze-pump-thaw
cycles and sealed. After heating for 4 to 11 days in an oil bath at 80-100°C’ with the
reagent consumption followed by 1H and 13C NMR, the tube was opened. In all cases
pressure developed proving the formation of a gaseous product at RT, which was
identified as isobutylene. 14 The product mixture was separated by flash column
chromatography over silica gel using a set of hexane: CH2C12 solvent mixtures with 5:1
to 1:3 ratios. Besides unreacted starting materials and some di-tert-butylurea byproduct,
EtzO trituration yielded the dimethyl ester of 2-N-tert-butylamino-3-cyano—2(E)-butene-
dioic acid and the tetramethyl ester of s-trans-1-N-tert-butylamino-l,3-butadiene-1,2,3,4-

tetracarboxylic acid (E, E), 1 and 2, respectively, in Figure 10.

Y Y Y

COOMe N

MeOOC NH
n Meooc NH I
| 0 ———~ I MeOOC l

I COOMe
NC COOMe

COOM" k H COOMe

4.1 4-2

Figure 10 General reaction CD1 + DMAD.

 

120

 

With no traps present product 1 and 2 were obtained from the reaction of 1,3-di-tert-
butylcarbodiimide with DMAD. With the equimolar ratios of the two reagents or with
slight excess of CD1, 1 is the major one obtained. When an excess of DMAD is used, 2 is
obtained in higher yield, eventually becoming the major product with a large excess of
DMAD.

When water and D20 water was added to the reaction mixture, along with 1 and

2, small amounts of lzlzladducts of CD1, DMAD and either H20 or D20 were detected
by GC-MS spectrometry. Methanol adducts of the same type were detected when MeOH

or MeOD were added as traps for the potentially formed carbene.

Reaction of DBCDI with Benzyne. Freshly prepared benzyne15 was heated in neat CD1
or in a 10-3 M solution of CD3CN. No [2+3] cycloadition products were identified.

Identified in the reaction mixture were of antranilic acid and unreacted CDI.

Reaction of DBQDI with TCNE. Sample (1 mL 10’3M) of DBCDI and TCNE in EtOAc
and CHC13 respectively, were degassed by 4 freeze-pump-thaw cycles and sealed in an
NMR tube. Heating for 4 days in an oil bath at 80-100°C gave a viscous mixture. After
extraction with ether and column separation of the unreacted reagents, GC-MS allowed
identification of traces of a 1:1 adduct (M=282) of CD1 and TCNE with a fragmentation
pattern attributable to a [2+3] cycloaddition product. The only other compound identified
in both cases was 1,3-di-tert-butylurea. Attempts to further optimized the yields and

investigate the adduct were unsuccessful.

Reaction of DBCDI with Maleic Anhydride (MA). A 1 mL sample of 10'3M of DBCDI
and maleic anhydride in CH3C1 was degassed by 4 freeze-pump—thaw cycles and sealed
in NMR tube. Heating for 4 days in oil bath at 80°C a polymeric reaction mixture was

generated. Extraction with ether and column chromatography allowed identification of

121

mm

 

 

the unreacted starting materials and traces of 1:1 adduct of CD1 and MA detectable via
GC-MS with a fragmentation pattern which may be explained 8 [2+3] cycloaddition
product. The only product separated from the unreacted starting materials cases was di-

tert—butylurea. Attempts to further optimized the yields and investigate the adduct were

unsuccessful.

Reaction of DBCDI with diphenyl-acetylene. After heating for 4 days a sample of CD1

and acetylene in CD 3CN , degassed by 4 successive freeze-thawed-pump cycles and

sealed under vacuum, no reaction was noticed. GC-MS analysis of the mixture showed

only starting materials and solvent.

Reaction of QEQDI with DMAD, 1 ml solutions of DPCDI and DMAD in EtOAc, THF-

d8 or CD3CN were degassed by 4 freeze-pump-thaw cycles and sealed under vacuum.
Samples were heated for 4-5 days in oil bath. Only the starting materials along with some

di -i sopropylurea were separated from the reaction mixture by column chromatography.

N o 2-azetine-like product from a presumable [2+2] cycloaddition as in the case of

hex afluoro-2-butyne was identified. When CF3COOH was added as catalyst, the main

product was the di-isopropylurea and some polymeric mass probably due to the self

condensation of DMAD.

Rea ti n fDP D wihT AlmL10'3M sample ofDPCDI and TCNE in EtOAc

Was degassed by 4 successive freeze-pump-thaw cycles, sealed in NMR tube and heated

for 4 days in oil bath at 80-100°C. Extraction of the reaction mixture with ether and
Column separation of the unreacted starting materials, GC-MS allowed identification of
traCes of 1:1 adduct (M=254) of DPCDI and TCNE with a fragmentation which may

6x131 ain a [2+3] cycloaddition product. Attempts to further optimize the yields and

1“\Nestigate the adduct were unsuccessful.

122

2.3. Discussion

2.3.1 Thermodynamic data

Simple bond energy calculations allow us to predict that the reaction of carbodiimide
with an acetylene should be energetically favored over reactions involving carbon
disulfide. From this starting point, semiempirical and ab initio molecular orbital
calculations were performed in order to compare the energetics of the reactions of C82
and of carbodiimide with acetylene. The computed systems used here are simplified
approximations of the real systems used in the lab, since they include unsubstituted
acetylene, C82 and unsubstituted carbodiimide (Scheme 11). The overall
thermodynamics of the two reactions shows that the reaction of acetylene with CD] is
more exothermic than the one with C82 and, consequently, more thermodynamically
favored. These simple estimations of the overall reaction energy pointed to a possible

new route to synthesize diaminocarbene.

 

H H

N

H N
Hc—:—CH s=C=s —3> [3:

S

r—--NH
HCECH HN= C=NH .1... I
NH

\NH KN

Figure 11 Processes followed by theoretical means

123

Table 1: Comparative thermodmamic data computed for overall reactions 1 to 6

 

Aern PM3 HFa MP2b MP4C MP2d
1 48.1 40.0 40.0 -38.7 -38.8
2 -l8.9 -26.0 -32.3 -29.7 -32.6
1+2 -67.0 -63.8 -72.3 -68.4 -71.4
3 -11.3 -3.0 -0.2 —0.3 -O.6
4 -240 -16.6 -19.2 -18.8 -18.4
5 -7.0 -10.6 -7.6 -8.9 —8.2
6a -20.8 -31.9 -334 -315 -332
6b -20.9 -30.6 -30.6 -28.3 -31.5
4+5+6a -51.8 -59.1 -60.1 -59.2 -59.7
4+5+6b -51.8 -58.9 —57.4 -56.0 -58.3

 

al-IF/6-3 lG*//HF/6-3lG*; bMP2/6-3 lG*//HF/6-3IG*;
CMP4/63 1G*/[HF/6-3 1G*; lefPZ/6-3 lG*//MP2/6-3 1 G*.

The overall thermodynamics show the reaction of C32 with acetylene, (Reaction 3
in Table 1) is almost thermoneutral (-3.0 kcal/mo] at HP and only 0.6 kcal/mol at the
MP2/6-3 10* level). In contrast, both reactions involving CDI are calculated to be
exothermic. The formation of the four-membered ring, (Reaction 4) as in the experiments

0f Hartzler, is 16.6 or 18.4 kcal/mo] exothermic at HF or MP2, respectively. However,
the [2+3] cyclization to diaminocarbene, (Reaction 1), is by far the most
thermodynamically favored process of the three considered here, with Aern = -37.8 or
”3 8-8 kcal/mo] at HF or MP2 respectively.

To calibrate the reliability of the thermodynamics calculated here, we compare

the“) with the available experimental data. Using the experimental heats of formation of

itTliclazole and acetylene and with our best estimate for the heat of formation of CD1,16

124

we calculated an overall Aern(1+2) = ~60.8 kcal/mo] which may be compared with the

calculated -63.8 or -71.4 at HF or MP2 level, respectively.

We considered subsequent reactions of the carbene, which may account for the
detected products formation. From diaminocarbene, the 1,2-H shift to lH-imidazole,
(Reaction 2) is strongly exothermic, as expected. The 2-azetine may undergo the typical
electrocyclic ring opening to the corresponding imino—ketimine (Reaction 5). In order to
explain the formation of a product such as 1, one needs an additional step, (Reaction 6a
or 6b) a net 1,5 H shift. If this process is intramolecular, it should initially lead to the cis
product with an overall exothermicity of ~ -60 kcal/mo]. Our experimentally obtained
product, 1, has a trans arrangement which may result from a 1,5 shift followed by
i somerization, or an intermolecular process.

The data available indicate the formation of the diaminocarbene as the most
thermodynamically favored product. In contrast, the experiments, both in Hartzler's case
and in ours, yield different products. The explanation must be sought in the actual

mechanism and the heights of the barriers involved in each step of the possible processes.

2-3.2 Mechanisms suggested in literature
The most common type of reaction mentioned in the literature for cumulated

double bonded systems are the [2+2] cycloadditions.l7 The reactions involving reagents
Such as C32, COS, ketenimines or carbodiimiides are [2+2] cycloadditions followed, in
most cases, by subsequent thermodynamically or kinetically driven chemistry toward the
mOSt stable product or products. Various [2+2] cycloadditions between carbodiimides or
phel‘iylisocyanate and acetylenes are known. However the reactions require a catalyst
SuCh as iron pentacarbonyl18 or similar complexes with cobalt19 or nickel.20

The existing experimental literature regarding possible [2+3] cycloadditions

luvolves only the C82 cases and there are no previous mentions of CD1 involvement in a

125

cycloaddition of this type. The experimental data available to date about cycloadditions
of a cumulated system with acetylene offer two different points of views on this type of
reaction. On the one side are the fast reactions of benzyne and tetramethylcycloheptyne
with C82. The assumption of a concerted [2+3] cycloaddition in these cases is supported
by the similar behavior of the two acetylenes, benzyne and strained
tetramethylcycloheptyne, which are neither strong nucleophiles nor electrophiles. In the
benzyne case, CSz reacted even faster than MeOH. Only the product of dithiacarbene

i nsertion in the O—H bond of MeOH was obtained, which shows that benzyne reacts
much faster with C82 than with MeOH and only afterward does the 1,3-dithiacarbene
insert into the O—H bond of MeOH (Figure 4.6).

On the other side, Hartzler's thorough investigation of the reaction of C82 with
acyclic acetylenes showed that acetylenes need to have at least one electron-withdrawing
group next to the triple bond in order to react with C82 and generate dithiacarbene.
Unsubstituted acetylene, such as 2-butyne, diphenylacetylene or vinylacetylene are
completely unreactive toward C82 while hexafluoro-Z—butyne or DMAD proved to be
effective and led to the expected products. But the reaction needs long times and high
temperatures (at least 100 °C) and goes better if it is acid catalyzed. These facts are
consistent with a step—by—step mechanism, as initially suggested by Hartzler. Such a

mechanism implies a nucleophilic attack on the acetylene to generate initially a

zwitterion. This ionic species has never been trapped and is thought to cyclize to the

carbene faster than to accept a proton.

The generally accepted explanation is that the dominance of cyclization vs. proton
abStraction, in the presence of acid is not the result of kinetic control via faster cyclization
but due to the thermodynamic drive toward the additional stabilization by the formation

of an aromatic structure, such as the 1,3 dithiacarbene. In the absence of suitable traps,

I‘Ial'tzler identified the corresponding dimer of dithiacarbene.

126

Knowing that dimerization in solution is less probable than any other mode of
carbene reaction, Hartzler explained the dimer formation, rather than through coupling of
two singlet carbene molecules, by initial protonation of the carbene to dithiolium ion and
subsequent reaction of the ion with another carbene molecule to form the dimer. The
formation of the dimer is a consequence of a long—lived carbene. For the case of interest
here, Hartzler's investigation showed no similarity in the behavior of C82 and
carbodiimide (see Figure 4.9 and corresponding references from the text).

2.3.3 Possible mechanisms
Given the previously suggested mechanisms and the results of our investigation we

consider here three possible paths for the reaction of the involved species (Figure 11).

a. [2+3] cycloaddition
b. [2+2] cycloaddition

c. Stepwise reaction.

a. [2+3] cycloaddition. For a cumulated a-b-c system free of formal charges, an
electron pair at a and c will create two allyl anion systems perpendicular to each other.
Charge migration during cycloaddition produces a cyclic allyl cation with an additional
anionic charge at b. Among the three resonance structures of the cyclic product formed as
the result of the [2+3] cycloaddition, one is free of formal charges but possesses an

electron sextet on atom b. If the system is C02, CD1 or C82, a carbon atom will play the

role of b and the neutral cyclic species is a carbene.

+ —* Li H ._./ he...

d=e

Figure 12: General [2+3] cycloaddition of cumulenes

127

There is no mention of such a process in the published literature involving C02 and only
one investigation involving CD1. Bis[bis(diisopropylamino)phosphino]carbodiimide is
the one case in the literature in which a CD1 is involved in [3+2] cycloaddition. But even
in this case, it is postulated that the bisphosphino substitution is responsible for [2+3]
cycloadduct formation with DMAD in 92% yield, because the three pieces include only
C and one N from the CD1 moiety along with one P atom from the phosphino part of the
molecule.21
By contrast, the experiments with C82 as cumulene are at the starting point of a
vast literature pool about tetrathiafulvalenes, their synthesis, chemical reactivity and
practical use. In the C82 case, the available data about its reaction with
tetramethylcycloheptyne or benzyne support the idea of a concerted cycloaddition given
that d-e is a triple bond (Figure 12). The reduced reactivity toward open chained
acetylenes, without electron-withdrawing substituents next to the triple bond of the
alkynes along with the requirement of an acid catalyst do not allow one to exclude an
initial electrophilic attack at the heteroatom, S or N, and a step by step mechanism.22
Our calculations for a concerted [2+3] cycloaddition Reaction found a barrier of
63 kcal/mo] at HF and much lower at MP2, 24.5 kcal/mo]. By contrast the C82 case
shows a barrier of 52.1 kcal/mol at the HF level (Figure 13 and Figure 14).

128

      

TSl.2H

   
  

  
 
  
   

0
‘ Q
a @fi‘a
shift
19.6

,una‘
I

\
5.9 ‘.

’IIII.‘

,'24.5

 

   
 

     
 

0.0 ‘s
Csz '1' CNsz

H‘\\ -63.8

mm- MP2/6826 //MP2/6-31G -76.4

_ HF/6-31G'

 

Figure 13 Calculated path for [2+3] CD1 cycloaddition to acetylene

The two values for the concerted addition at HF for C82 and CD1, respectively, support
the idea that, C82 should add more readily to to acetylene because its barrier is 12
kc al/mol lower than that for CD1. However, the height of the barrier is still significant

even for C82 case.

A TS [2+3]

   

0.0
Csz + CS2

@
— HF/6-31G' " o

 

Figure 14 Calculated path for [2+3] C82 cycloaddition to acetylene

129

The easier cycloaddition of C82 may be the result of the difference in the HOMO-LUMO
gap between the two cumulenes,23 as well as the difference in size between N and S
atoms. With sulfur bigger than nitrogen, access to the central carbon may be encumbered
and interaction with either of the sulfur atoms in a stepwise reaction or with both in a
concerted [2+3] cycloaddition would lead to the experimentally obtained dithiacarbene.
In addition, the C=S bond, being longer than C=N bond may allow easier bending for
C82 molecule compared with CD1 and as a consequence, better overlap of both S atoms

with the acetylene. Also the substituents on N atom will complicate the situation

compared with the S atom.
A possible thermal rearrangement of the carbodiimide to the corresponding nitrile

imine followed by a fast reaction of the nitn'l imine with DMAD may be another
explanation for the main product obtained.24 Theoretical studies of this type of

re arrangement at different ab initio levels were published.25 The G2(MP2) level shows
iminonitril only 2 kcal/mo] more stable than the carbodiimide. The rearrangements is
postulated to take place through a l H—diazirine intermediate which is calculated to be,

when the same level of computational method is used, 59 kcal/mol higher in energy than

the carbodiimide.26

b- [2+2] cycloaddition is a common reaction among heterocumulenes.27 Since concerted
thfil‘mal [21ts+ 211:5] cycloadditions are not orbital symmetry allowed, the mechanisms
in vol ve an allowed concerted [21ts+ 21:3] reaction leading to a four membered ring
hf-Eterocycle.” Usually, the olefin is the 1:25 component. The reasoning behind the high
re ac ti vity of heterocumulenes such as ketene in [2+2] cycloadditions is the favorable
SecOndary interaction between the HOMO of the olefin and the vacant orthogonal C=O
“1* antibonding orbital of ketene.

one may imagine that all cumulenic systems should undergo concerted cycloadditions

S‘nce they all have an orthogonal vacant 1t* antibonding orbital. However, this is not true

130

 

for all cumulenes. It is the energy difference between the 71* orbital of the cumulene and

the HOMO of the olefin which dictates the possibility of a concerted reaction rather than
a stepwise reaction. In the ketene case, there is an usually small gap between its HOMO
and LUMO orbitals, (Figure 16). The calculated difference in the carbodiimide case is

1 6.7 (SCF, with 3-ZIG* basis set) or 10.96 eV (MNDO).29 More theoretical analyses of

carbodiimide structure and stability are available in literature.30

Ketene CD'MNDO CDISCF

1,,
W LUMO (eV) 3.8 1.52 5.70
M HOMO (eV) -12.4 -9.44 -1 .52

Figure 16 HOMO-LUMO gap in ketene and CD1

131

 

HW/I mEOOOnm
zo _
//Z\LHM 4

62000

 

 

o.2000

 

/m\

o __
__

z

.7
yr 2..
20 m o m z m
_ II V -
12H 2 m J\

\\
\

 

AGO 2 9375 co 5:63 2: he mEmEmcooE 038mg 8:: 2; .m. :4 2:0ch

132

Such a mechanism is used to explain the results in the experimentally known CD1
cases. The initial product of [2+2] cycloaddition, 2—azetine, lacks the normal stabilization
of an amidine because of the unfavorable azabutadiene resonance structure in its dipolar
form. It is also known that azetines readily undergo electrocyclic ring opening and

subsequent rearrangements.

c. The stepwise reaction would imply a zwitterionic intermediate, in complete
disagreement with the previously cited rapid room temperature reactions of benzyne and
tetramethylcycloheptyne with C82. The existence of such an intermediate has never been
experimentally proven. There is only one case known, that of carbonylsulfrde reacting
with hexafluoro-Z-butyne, in which an initially formed zwitterion was proposed to exist
in equilibrium with the separated species (Figure 15). Cyclization of the zwitterion
to form the corresponding five membered ring would than occur via nucleophilic attack at
the carbonyl moiety. If there is a suitable trap, the equilibrium would be shifted toward

the zwitterion which should react further with alcohols, as in the case C82. Only trans

addition to the acetylenes was observed.

R—E—R g _ Rgzfi C_a__.H 0“ Mtg
<'> 9%,, moo/0%

Scheme 17. Carbonyl sulfide case

Our calculations found no transition state for a concerted [2+2] cycloaddition in
the model reaction between CD1 and acetylene. At the approach of the acetylene to the
CD1 moiety two other possible paths were found, besides the [2+3] cycloaddition route

ClesCribed above. One implies the interaction of only one N atom from CD1 with the

acetylene, in a concerted [1+2] cycloaddition (Figure 17). There are known cases in the

133

literature which describe the isonitrile reaction with DMAD or hexafluoro-Z-butyne to
lead to products with similar structure to the one calculated here.31 The reaction is

thought to follow a stepwise path, via a zwitterionic intermediate or a very polar TS.

 

Figure 18 Calculated TS for formal [1+2] cycloaddition of CD1 to acetylene

Our calculations found also a polar TS, corresponding to a barrier of 79.8 kcal/mo] at the

PIP level (Figure 17). From the TS, the system evolves to azirine and isocyanide. The
IRC calculations describe the reaction path from the TS to separated molecules as
oriented in such a way to form a favorable H bond between the isonitrile and the N in the
azirine ring. This path is of no interest for the case discussed here and, with its high
barrier, is probably not competitive, at least under the conditions of our experiments.
The other path is described as a step by step interaction via a zwitterionic
intermediate. Published theoretical investigations of similar reactions, such as ketene
addi tion to double bonds, reveal the complexity of the problem. Bemardi et al. used an
IVICSCF wavefunction with STO-3G and 4-31G* basis sets for the reaction of ketene
With ethylene.32 Both the perpendicular and the parallel approach of the ketene to the
ethylene lead to a short lived diradical intermediate. No reaction path for a concerted
[27115-1- Zna] was found in contradiction with the assumption of antarafacial addition of

ketene to the suprafacial olefin. Houk et al. studied the same reaction theoretically,

134

adding MP2 correlation energy.33 They found the reaction pericyclic with the bond
formation asynchronous.

Our calculations found a TS ~57 kcal/mol higher in energy than the starting
species. An IRC calculation from this TS leads to a highly unstable zwitterionic
intermediate, 55.1 kcal/mo] above the acetylene and CD1 starting point, at the HF/6-31G*
level. The zwitterion easily closes to 2-azetine, the product of a formal [2+2]
cycloaddition. The optimization of the same species at the MP2/6-3lG*//MP2/6-3IG*
level starting with the HF geometry, also closes to 2-azetine. The fact that the zwitterion
is a minimum only at the HF suggests that there is no barrier for the closing of the
zwitterionic transition state into the 2-azetine ring. IRC calculations starting with the

same TS show the direct connection between the acetylene and CD1 to the four

membered ring product (Figure 18).

  

TSto-azetine

C2H2 + CNZHZ

 

"""" MP2/6-326'//MP2/6-31 G'
— HF/6-31G'

 

Figure 19 Calculated path for formal [2+2] cycloaddition of CD1 to acetylene

135

The [2+2] concerted cycloaddition is unlikely to compete with a step by step reaction via
a zwitterionic transition state and/or intermediate which cyclizes to 2-azetine.

The route from 2-azetine to the product analogous to 1 is 10.5 kcal/mo] exothermic and
should arise via a Woodward-Hoffman allowed34 electrocyclic ring opening to an imino-
ketimine intermediate in a concerted conrotatory fashion. The torquoselectivity rules-2’5
force the donor substituent to go outward to minimize the interaction of its filled orbitals
with the electron pair of the breaking 0 bond. Geometry and energy predictions via
computational methods for the cyclobutene ring system, less complex than our case here,
are contradictory.36 Barriers for the cyclobutene cases were calculated to be in the range
of 30 kcal/mol.

In order to reach a product analogous to 3, the cis isomer of l, a 1,5 H shift
process was computed. We found a TS for this direct process with a barrier of 64
kcal/mo] at PIP/6610*. However, our calculations are on a system which only
approximates the experimental case, as long as the product obtained is mainly trans. Its
formation via intramolecular rearrangement has no relevance to the experimental findings
from substituted CDIs.

In the CD1 cases studied here, the reactions were very slow. In most of the cases,
after 3-4 days of heating at reflux, more than 50% of the starting materials were found
unreacted. The reaction mixtures contained tars, which makes the work—up difficult. Both
Carbodiimide” and DMAD self dimerize or even oligomerize to products of higher
molecular mass. In the case of DMAD, tetramers are thermally generated in aged samples
0f DMAD or by simply heating at 100-120°C for several hours.33 The length of the
reaCtion time and the relatively high temperature probably allow subsequent reactions of
the CD1 or DMAD dimers which are responsible for the polymeric materials formed. In

all the cases pressure was developed in the reaction vessel (always noticed while opening

the sealed reaction tube). This pressure development was attributed to isobutylene,

136

obtained as byproduct of 1 formation reaction. Besides I and 2 described earlier in this
chapter as the two main products present in all the trials, one additional secondary
product, 3, was obtained in some runs. Because of its appearance in the GC-MS close to
the 1 signal and its fragmentation path showed in the MS spectrum, we conclude that 3 is
the cis isomer of 1.

Following the reaction mixture evolution by 1H and 13C NMR39 showed no traces
of free carbene at any time.40 Products of the carbene's subsequent reactions with solvents
or the added traps were also sought. No new products were formed with benzaldehyde
but in the case of water and methanol, GC-MS revealed small amounts of products with
masses appropriate to the carbene + trap adducts.

Compound 1 was analyzed by IR, GC-MS, 1H and 13C NMR and X-ray
crystallography (see experimental part). One possible explanation for this product
formation as the major one in all the cases where the molar ratio of DBCDI to DMAD is
close to 1, is the initially preferred formal [2+2] cycloaddition to form an azetine, similar
to one case described in the literature (Reference 9). Subsequent ring opening to an
irnine-ketimine structure followed by H—transfer coupled with isobutylene elimination

may afford the final product 3 (Figure 19).

—_’

N
‘
S

MeOOC\ KNQ'H Meoo: ,&5”
M9 ‘3 Kg Me I C§N JL

Figure 20 Suggested mechanism for product 3 formation

M6033”); M9000 X
I

 

M90

137

However, X-ray analysis shows that the two carbomethoxy groups are trans oriented with
respect to the double bond (Figure 4.21).

Figure 4.21 Product 1 x-ray crystal structure

73.

”cm

C1~C2
C2-C3
C3-OA
C3-05
OS-Cé
C2-C 7
C7-N8 5

C1-C9 1.52
C9-OIO 1.20
ClO-Oll 1.33
()ll-ClZ 1.45
C1-N13 1.32
Nl3-C141.50

  

d-‘ddddd
' I u 9 o a -

138

W as the major product may be explained only by a different mechanism

than the one suggested for the formation of 3.

ll

s-cis s-trans

$5.4

 

s-trans s-trans

Figure 22 Suggested mechanism for 1 formation

W : The minor product 3 is tentatively assigned as the cis-isomer of 1. Its
eXistence was detected only by GC-MS and the small proportion obtained did not allow
iSolation and purification for complete characterization. The mass spectrum
fragmentation pattern is the only indication that it is the cis isomer of 1. If the structure
assignment is correct, its existence in the GC-MS analysis, points to a high barrier for
c=c bond rotation in 1, and makes the thermal isomerization of 3 to 1 unlikely.
W- This is the second major product in all trials in which the molar ratio
of DBCDI and DMAD is close to 1. When the DMAD is in excess, self condensation

With of the CD1 leads to more 2 formation. We explain the formation of 2 as the result of

139

subsequent reactions of the initially formed dimer of DMAD with DBCDI or their
initially formed 1:1 adduct, which reacts further with a new molecule of DMAD.
Elimination of an isobutylene molecule from the 2: 1 adducts leads the final 2. Similar

2:1 adducts of DMAD ester with isonitriles are reported in literature.41

3 Conclusions

The reaction of dialkyl-CDIs with acetylenes and electron-withdrawing
substituted alkenes were studied. The purpose of the study was to investigate the potential
for addition of alkynes or alkenes to the cumulated double bonds in carbodiimides to
generate five-membered cyclic structures, a potential new route to nucleophilic carbenes.
The results which were consistently obtained under various conditions indicate that
carbodiimide is not an effective partner for [2+3] cycloaddition with electron-poor
acetylenes to give nucleophilic carbenes. Although traces of compounds were found with
masses consistent with carbene formation and trapping, these products could also have
arisen in many other ways. Thus, the [2+3] cycloaddition of dialkyl-CDIs and acetylenes
does not appear to represent a useful new route to nucleophilic carbenes. Even though
some products of carbene trapping were detected, the reaction was not proved a useful
Procedure for synthesizing and study nucleophilic carbenes.

However, despite its failure to produce detectable carbenes, the [2+2] reaction of
CD15 with acetylenes has proven interesting and deserves to be further discussed and
investigated. For instance, the reactions of CDls with strained alkynes, such as benzynes
01' Cycloheptynes certainly merit further investigation. Such reactions may follow a
completely different route than the one found here and may yet represent a reasonable
pathway to diaminocarbenes. With the availability of powerful theoretical tools, a better
understanding of these systems may be achieved and strategies for control of the reaction

may be found.

140

Overall, the work here opens a new perspective on the field of carbodiimides as partners

in cyclization reactions.

4.4 Experimental methods

W- Melting points were determined on a Thomas Hoover capillary
apparatus and are uncorrected. All air sensitive reactions were performed in oven-dried
glassware using standard syringe/cannula techniques. Solvents were purchased from
Aldrich and dried and deoxygenated by standard procedures. Gravity and flash column
chromatography were performed on E. Merck silica gel (230—400 mesh). Thin-layer
chromatography was done on E. Merck plastic—backed plates (silica gel 60, F245, 0.2
mm).

Routine 1H and 13C NMR spectra were obtained at 300 and 75.43 MHz
respectively using either Varian VXR-300 Varian GEMINI 300 NMR Spectrometers.
The 1H NMR chemical shifts are referenced to the residual lH resonance in the
deuterated solvents used: CDCl3 (5 7.24), acetonitrile-d3 (5 1.93). The 13C chemical
shifts are referenced to the one of deuterated solvents: CDC13 (8 77.0) and acetonitrile-d3
(5 broad118.0 and 1.3 septet).

Fourier-transform infrared (IR) spectra were obtained on a Mattson-Galaxy FT-IR
3020 or N icolet IR/42 spectrometers. Samples were measured either as thin layers
Prepared by evaporating a CH3CN or CHCl3 solutions on a NaCl plate (liquids) or as
KBr pellets (solids).

Electron impact (EI) mass spectra were obtained on a Fison VG tn'o-l mass
Specti'ometer which operates in line with a Hewlett Packard 5890 gas chromatograph for
GC‘MS measurements. High resolution mass spectra were obtained on a JEOL JMS-
HXl 10 high resolution double—focusing mass spectrometer.

Pyrex or uranium glasses were used as filters for photolysing the samples with

light from a 500 W Oriel hi gh-pressure Hg lamp.

141

General prxedures. In a standard NMR tube, 1-1.5 mL solution of DBCDI and DMAD
in a l :1 to 1:3 molar ratio was placed. The sample was degassed using 3 to 5 freeze-
pump-thaw cycles and sealed under vacuum. Alternatively, sparging with dry N2 or Ar
for up to 1/2 hours was used for degassing procedure.

The sample tube was transferred to an oil bath and heated at 60 to 80 °C, from 1 hour to
11 days, with shaking every 1 to 4 hours to homogenize the solution. For ultrasound and
microwave experiments, the reaction mixture was prepared as above, in standard NMR
tube and was transferred to a standard ultrasound water bath or to a lab microwave oven.
After the competition of the reaction, each sample was brought to room temperature and
the tube was opened. The samples were extracted twice with ether and separated from the
insoluble polymeric mass. The etheral solution was evaporated and the residue was
separated by gravity or flash chromatography over silica gel with hexane and CH2C12
mixture in ratios 5:1 to 1:3 ( in volumes).

The main products 1 and 2, the unreacted starting materials and di-tert-butylurea were
separated from the reaction mixture.

Reagent consumption was followed in all the cases by 1H and 13C NMR. GC-MS data
were taken for the initial and final reaction mixture. IR spectra were taken in thin film on
NaCl pellets.

Pl'Oduct l isolation.

The major product of DBCDI and DMAD reaction was obtained following the general
Procedure described above. After chromatography, the product was separated as a
CryStalline solid, and was recrystallized from hexane. mp=82—83° C. 1H NMR (300 MHz,
CDC13) 6 1.36 (9H, s), 3.76 (3H, s), 3.94 (3H, s), 9.8 (1H, broad); 13C NMR (75.5 MHz,
CDC13)8 30, 52, 53, 56, 100, 117, 160, 162, 168. IR cm'1 3130, 3206, 2210 (CN), 1729,
1672- MS (EI) m/e 240(M+), 241(M++1), 184, 153, 125. High resolution MS Molecular
Weight 240.1107.

142

Product 2 isolation.

The product of DBCDI and DMAD reaction was obtained following the general

procedure described above. After one initial chromatography, the product was separated

from a later fraction from an oily mixture by trituration with ether.
1H NW (300 MHz, CDC13) 5 1.29 (9H, s), 3.62 (3H, s), 3.63 (3H, s), 3.70 (3H, s), 3.72
(3H, s), 6.72 (1H, s), 9.10 (1H, broad); 13C NMR (75.5 MHz, CDC13) 5 30.4, 51.1, 51,9,

52.4, 52.7, 56, 129.0, 140.0, 151.9, 150.9, 164.0, 166.0, 168.0, 167.0. IR cm“1 1741 (C=O

vs) 1724 (C=O vs), 1666 (C=C), 1593. MS (EI) m/e 357(M+), 301 (C12H15N08), 342,

242 (C10H12N06), 210, 178, 150.

 

4.4 References

l

a) Wanzlick, H. —W.; Schikora, E. Angew. Chem 1960, 72, 494. di-
phenylimidazolidinylidene; b) Wanzlick, H. -W.; Kleiner, H. -J. Angew. Chem 1963,
75, 1204. benzothiazolylidene. c) Wanzlick, H. -W.; Schonherr, H. Angew. Chem
1968, 80, 154. N, N-di-phenylimidazolylidene. d) Wanzlick, H. -W.; Schonherr, H.
Liebigs Ann. 1970, 731, 5530.

Arduengo, A. J. 111,; Harlow, R. L.; Kline, M. J. Am. Chem. Soc. 1991, 113, 361.

Dixon, D. A.; Arduengo, A. J. 111,; Harlow, R. L.; Kline, M. J. Phys. Chem. 1991, 95,
4 l 80.

a) Herrmann, W. A.; Kocher, C. Angew. Chem. Int. Ed. Engl. 1997, 36, 2162-2187
and references therein. b) Regitz, M. Angew. Chem. Int. Ed. Engl. 1991, 30, 674.

Arduengo, J. H1 US Patent [19]: [11] Patent number 5, 077,414.and [45], Dec. 31,
199L

I’{Uhm N .; Kratz, T. Synthesis 1993, 561-562.

Alder, R. W.; Allen, P. R.; Murray, M.; Orpen, A. G. Angew. Chem 1996, 108, 1211.

143

 

8

10

11

12

13

14

15.

16

17

a) Matrix experiments: Maier, G.; Enders, J .; Reisenaur, H. P. Angew. Chem. Int. Ed.
Engl. 1997, 36, 1709; b) Gas-phase: McGibbon, G. A.; Hrusak, J .; Lavorato, D. J .;
Schwartz, J, K.; Terlow, J. K. Chem. Eur. J. 1997, 3, 232.

Hartzler, H. D. J. Am. Chem. Soc. 1970, 92, 1412.
Field, E. K.; Meyerson, S. Tetrahedron Lett. 1970, 629.

Nakayama J. J. Chem. Soc., Chem Commun. 1974, 166; Nakayama J. Synthesis,
1975, 38.

a) Krebs, A.; Kimling, H. Angew. Chem . Int. Ed. Engl. 1971, 10, 509. b) Krebs,
A.; Kimling, H. Liebigs Ann. Chem. 1974, 2074.

a) Regitz, M. Angew. Chem. Int. Ed. Engl. 1991 30, 674-676; b) Regitz, M. Angew.
Chem. Int. Ed. Engl. 1996, 35, 725-728.; c) Herrmann, W. A.; Kocher, C.; Goossen,
L. J.; Artus, G. R. J. Chem. Eur. J. 1996, 2, 1627-1636.

1H NMR at -10 °C was run on the crude reaction mixture before the tube was cut
open. The signals of isobutylene's hydrogen were detected, as follows:

For benzyne synthesis we use Hart, H.; Oku, A. J. Org. Chem. 1972, 37, 4269.
Other procedures: a) Stiles, M.; Miller, R. G.; J. Am. Chem. Soc. 1960, 82, 380.
b) Stiles, M.; Miller, R. G.; Bruckhardt, V. J. Am. Chem. Soc. 1963, 85, 1792.
c) Chapman, 0. L.; Chang, C.-C.; Kolc, J .; Rosenquist, N. R.; Tomioka, H. J.
Am. Chem. Soc. 1975, 97, 6586.

With G2 energies for CD1, calculated by us here as -148.555524 and the G2 energies
of atomic C, -37, 78432, N, -54.51798 and H, -0.5, Aern of CN2H2 --> C+ 2H +2N
was calculated. With the obtained value and the experimental heat of formation of C,
N and H, 171.3, 112, and 52.1 respectively, we estimate AHf CD1 2 40.1 kcal/mol.

a) Huisgen, R. Angew. Chem. Int. Ed. Engl. 1963, 2, 565. b) Williams, A.;
Ibrahim, I. T. Chem. Rev. 1981, 81, 589. c) Kurzer, F.; Douraghi-Zadeh, K. Chem.
Rev. 1967, 67, 107.

18 Ohshiro, Y.; Kinugasa, K.; Minami, T.; Agawa, T. J. Org. Chem. 1970, 35,2136.

144

 

19 Kinugasa, K.; Agawa, T. J. Organomet. Chem. 1973, 51, 329.
20 Hoberg, H.; Burkhart, g.; Synthesis 1979, 525.

21

22

23

24

25

26

27

28

Venetiani, G.; Reau, R.; Dahan, F.; Bertrand, G. J. Org. Chem. 1994, 59, 5927.

Although C82 is not a strong Bronsted base and the reaction was going well with

hexafluoro-Z-butyne, we do not need to induce acid reactions with esters.

The UV-VIS absorbtion of C82 vs N, N'-disubstituted carbodiinrides, reveal an
absorption shifted to higher wavenumbers for the sulfur case, which is an indication
of a smaller HOMO-LUMO gap in the two cumulenes; Behring, H.; Meier, H. Liebig
Ann.Chem. 1957, 67, 607.

Bertrand, G.; Wentrup, C. Angew. Chem. Int. Ed. Engl. 1994, 33, 527.

a) Thomson, C.; Glidewell, C. J. Comput. Chem. 1983, 4, 1. b) Guimon, C.; Kahayar,
S.; Gracian, F.; Begtrup, M.; Pfister-Guiilouzo, G.Chem. Phys. 1989, 138, 159.

c) Moffat, J. B. J. Molec. Structure, 1979, 52, 275. d) Leung-Toung, R. unpublished
results, cited in Bertrand's review in reference 39.

Fischer, 8.; Wentrup, C. J. Chem. Soc. Chem. Commun. 1980, 502.

[2+2] cycloadditions for allenes: a) Broggini, G.; Zecchini, G. Gaz. Chim. Ital. 1996,
216, 479.

Isothiocynate as a close model for the case of interest here: b) Sukumaran, K. B.;
Angadiyavar, C. 8.; George, M. V. Tetrahedron, 1972, 28, 3987.

Carbodiirrride: ref 16; c) Vovk, M. V.; Samarai, L. 1. Russian Chemical Review,
1992, 61, 548. d) Gordestov, A. S.; Kozyukov, V. P.; Vostokov,1. A.;
Sheludyakova, S. V.; Dergunov, Yu. 1.; Mironov, V. F. Russian Chemical Review
1982, 51 , 848. e) Rzepa, H. S.; Molina, P.; Alajarin, P.; Vidal. A. Tetrahedron, 1992,
48, 7425. f) Rzepa, H. S.; Wylie, W. A. J. Chem. Soc. Perkin Trans 2, 1991, 939.

g) Hunt, P. A.; Rzepa, H. S. J. Chem. Soc. Chem. Commun. 1989, 623.
Ketenimines: h) Marchand-Brynaert, J .; Ghosez, L. J. Am. Chem. Soc. 1972, 94,
2870.

[n23+n2a] mechanism in Woodward, R. B.; Hoffmann, “The Conservation of
Orbital Symmetry”, Academic press, New York, NY., 1969.

145

 

29

3O

31

32

33

34
35

36

37

38

a) Guimon, C. ; Khayar, S.; Gracian, F.; Begtrup, M.; Pfister-Guillouzo, G. Chem.
Physics 1989, 138, 157. b) Guimon, C. ; Pfister-Guillouzo, G. Can. J. Chem.
1986, 64, 1467.

a) Hart, B. T. Aust. J. Chem. 1973, 26, 461. b) Moffat, J. B. J. Molec. Structure,
1979, 52, 275.

a) Oakes, T. R.; David, H. G.; Nagel, F. J. J. Am. Chem. Soc. 1969, 91, 4761.
b)Winterfeld, B.; Schumann, D.; Dillinger, H. J. Chem. Ber. 1969, 102, 1656.

Bemardi, F.; Bottoni, A.; Robb, M. A.; Venturini, A. J. Am. Chem. Soc. 1990, 112,
2106.

Wang, X.; Houk, K. N. J. Am. Chem. Soc. 1990, 112, 1754.

Woodward, R. B.; Hoffmann, R. J. Am. Chem. Soc. 1965, 87, 395.
Dolbier, Jr. , W. R.; Koroniak, H.; Houk, R. N. Acc. Chem. Res. 1996, 29, 471.

a) Houk, K. N.; Beno, B. R.; Nendel, M.; Black, K.; Yoo, H-Y.; Wilsey, S.; Lee, J.
K. J. Molec. Structure (Theochem) 1997, 398-399, 169. b) Wiest, 0.; Montiel,
D. C.; Houk, R. N. J. Phys. Chem. 1997, 101, 8378.

a) Ulrich, H., C ycloaddition Reactions of Heterocumulenes, Academic Press, New
York, 1967. b) Carruthers, W. Cycloaddition Reactions in Organic Synthesis,
TetrahedronOrganic Chemistry Series, Volume 8, Pergamon Press, 1990.

a) LeGoff, B.; LaCont, R. B. Tetrahedron Lett. 1967, 2333. b) Winterfeldt, B.;
Giesler, G. Angew. Chem. 1966, 78, 588. c) Winterfeldt, B.; Giesler, G. Chem. Ber.
1968, 101, 4022. d) Kauer, J. C.; Simmons, H. E. J.Org. Chem. 1968, 33, 2720.

e) Gericke, R.; Winterfeldt. E. Tetrahedron, 1971, 27, 4109. f) Tsutsui, M.; Hrung,
Y.; Francis, J. N.; Harasawa, K. Chemistry Letters, 1973, 557. g) Gilchrist, T. L.;
Pearson, P. J. J. Chem. Soc. Perkin Trans 1, 1976, 1257.

39 From the existent literature data about 13C NMR signal of carbene with similar

structure, we expected to get a signal in the region of 240-260 ppm.

146

 

 

40 Considering the relatively high temperature and the length of the reaction time while
the samples are heated, is highly improbably that the most stable carbene, even

formed, will remain unreacted in the reaction mixture!

41 Winterfeldt, B.; Schumann, D.; Dillinger, H-J. Chem. Ber. 1969, 102, 1656.

147

APPENDIX

A Chapter 2

A1 PM3 Calculated energies for exchange reactions ...................................................... 149
A2 Calculated energies for CH20 + CH; exchange reaction ........................................ 150
A3 Calculated energies for CF20 + :CFz exchange reaction ......................................... 151
A4 Calculated energies for CF20 + :CH2 cross reaction ............................................... 152
A5 Rate studies for PCC: + PNO and PCC: + MNO .................................................... 153
A6 Oxiranes data for Marcus theory aplication ............................................................... 157
A7 Ylide data for Marcus theory aplication .................................................................... 158
A8 TSs Ylide like data for Marcus theory aplication ...................................................... 159
B Chapter 3

Bl Calculated energies for C02 + :CH2 complexes ........................................................ 160
B2 Seleced IR frequencies (nm) of the possible intermediates ...................................... 160
B3 Experimental IR data existent in the literature ........................................................... 160
B4 Intermediates geometry data ...................................................................................... 161
BS Calculated energies of intermediates ......................................................................... 162
B6 Calculated transition sates energies ........................................................................... 163
B7 Cartesian coordinates and total energies for the species in Chapter 3 ...................... 164
C Chapter 4
C 1 Calculated energies for the species in Chapter 4 ........................................................ 170
C2 Mass fragmentation pattern for products 4.1 and 4.3, cis and trans isomers ............. 171

148

Al PM3 calculated energies for exchange reactions

 

 

Compound AMI PM3 HF/6-3 1G*
(kcal/mol) (kcal/mol) (hartree)
Hzco 41.5 44.2 413.863312
H2C: 110.6 113.3 -28.8723704
Ylide 22.7 452.7542324
Oxirane -8.9 -8.1 452.867356
TS through oxirane 46.1 46.1 452.7232732
H2CO + :CH2 79.4 79.1 452.7387016
F2CO -146.1 441.5 411.6153061
F2C: -67.9 -67.2 -236.607441
Ylide -2003 484.9
Oxirane -203.8 -2057 -548.3011706
vdW complex -215.4 -l91.8 -548.2784265
TS through oxirane -184.8 -182.4* -548.2415770
TS through ylide 499.0 472.7 -548.1152010
F2CO +:c1=2 -2141 -208.3 -548.2760502
(H2N)2co 44.9 47.0 -223.9846922
(H2N)2C: 20.6 32.6 4490413289
Ylide 10.8 12.2 473.0048292
OXirane 0.4 -92 473.03449209
vdW Complex -32.5 -24.0 ~373.03449291
TS through ylide 11.8 19.2
TS through oxirane 23.6 15.5 472.8850031
(H2N)2co + :C(H2N)2 -244 44.4 4730260211
00 4.7 49.7 412.737877
C02 -79.8 45.0 487.634176
VdW complex —84.8 405.2 400.3730608
OXiI‘ane 47.9 44.9 400.2529079
TS through oxirane 46.2 -51.6 400244234
CO2~|~C0 -85.5 404.8 400372053
\

 

149

 

.
. .1 £8.91.- ...-... .-.. .. . .. "1‘.

 

owmmm.mm_-
@394me

330mm _ -
oemwmodm-
~owmm.v_ T

vmmom.mm—- www.c— .mm _ -

Neg—@2-
hEVWmmT Shaman—-
woNeN.mm—- memo—£2-
ofimgdfl- 046:me

Smmfimfl- cemofimmfi.
ohmamwm- omwmmoom-
Nomenvz- 448244:-

mefimmT Eco—62-
vmw~N.mmT

oooofimfl-

wmmom.mm_- wwwmmdfl-
owmmm.mm_- onemmdmd-
25:me emcom.mm-

Nwhmfimmg- wmmwfimm?
:oog.wm- engag-
mEbfiVZ- Nomad:-

2858

35242- 2.4842- 3.222- 22. -oo__wm%
5.232- 8842- 83842- 288%..
2832. 5222- go 2.

22842- 8842- £232- o2- 95:8
3:32- 8242- 2832- SN 62;
2832- 4832- 3:22- xofleoo
mo”

83:me maonT mammfimfl- ms:- 2- +OUNE
wwmoo.wm- Nmewfim- bmmhwem- NM: NS NED“
£52.37 thde- cmmow.m:- New- om- DUNE

 

NO

222:2 ooh +82

*0_ m-QNEZ EVA:

"HERE 6% +22 .0842: 22 oxm

 

5:88 uwcocoxo NEON + ONIO he“. mawooco 322330 N<

150

 

2858 Q

 

 

envovdfi- oceanovm- «Macmdvm- omvovdvn Nmavmdvn NoENme- wflvmwvm- 2N2- EEoooom:
omomdem- mwmomdfi- $42 2.2%. $23de- >.N:- swoesmh
Romvdfi- 9 392me- vawsvdg- $524.an- : 23.me- nwom- 28:8
omwdeVm- :wimdg- 844$. 023
commode? acmmodem- $324de- goomdvm- mewomdvm- @32- quaEOo
mu”
033.3%- vomvdem- geomag- VQ :4.on- Sfimdem- Nvmmmdvm- £55.me- h.wom- + Comm
Sofie-mm- >342 .x-mm- one _ mumm- mman-mm- gage-mm- 033.98. 3.8063. «.5- 54. who”
abound" m- 5.33.3 m- Kmmmdm m- aomnm.m~ m- ESQQ m- Nmmmmg 3”- Ems; 2m- 2:;- mm 7 CON-m

SEES 6% +82 82.28: 2:82 2:82 6% +82 82.2.2 go 92

8:82 £0” + OED 65 8o. mflwooco boom—330 m<

151

 

«RUI-

 

Eoimm- 2822- 8222- N882...- mmommamm- 2.882- 284.com- mommomm- :- com:

323.:- 88222- 2.822.:- E2....:- 883.:- anoz- 82.22- 8.8.2.22- Nem- om- 82

8832. 22.2.5- 8322- 2:22- 25.5.8- 8832- 8832. $838. «so. 2- N8.

2.82.22- 22.8mm- 22822- 8822- 2.282- Emmomm- 4.2- 29:8

Qua

8222- 2.882.. mwoowfim- 2882. 3282- 884.com- 483.com- noo- Mumm-

8832- 2822- 2232.2- 2882- 28- ozomo

882.22- 823.22.- 282.22- 2522- 2322- 32- 88.8

$832- Nwomvomm- 824.com- :38

":03

$822. mzoafim- 888mm- mmoomamm- 8:22- 23..on- woaoomm- ME..- 2. 88

822.8- osmooom- 882.2- :882. 2832- 2828. 3.2.8. 8848. N22 82 £0.

2.284:- ooosmdm- 224:- :83:- oomafim- 22.3:- Nmoomdm- :23:- 22- m2- 8.2
8 8282 68... +82 82-282 .8482 "E82 6% +8: 88.222 28 Sm

 

corona moo-6 NED” + Club .28 flew-Eco vogue—«U 3..

152

A5 Rate studies for PCC: + PNO and PCC: + MNO

Upon irradiation of PCD in acetonitrile with a Nz-YAG laser puls (355 nm, 10 ns
pulse, 10 m1) a long-lived absorption was observed, centered at 315 nm and which
decays on the us timescale. It was previously assigned to PCC:.1 When irradiation of
PCD in acetonitrile in the presence of pyridine is performed, the PCC:-pyridine ylide
long-lived absorption stable on the ms time scale and centered at 475 nm was observed.
This on is also a. This one was previously assigned to pyridinium ylide (ref 18). We
found that the intensity of this absorption was inversely proportional to the concentration
of the oxygen donor, with constant pyridine and diazirine concentration. The rate of
oxygen abstraction may be measured either by the decay of the carbene signal or by
Stem-Volmer quenching of the pyridinium ylide signals as described below. In our case
the latter was found more convenient due to the relative intensity of the signals. Also, an
intermediate absorption was noted, centered around 300 nm which may overlap with the
carbene signal.

The ratios kx/kpyr[pyridine] where kx are the bimolecular rate constants for the
reactions of FCC with the oxygen donor, and kpyr is the bimolecular rate constant for the
reaction of FCC with pyridine were determined by linear least-squares analysis of the
ratio Aoylide/Aylide versus oxygen donor concentration with the y-intercept defined as 1,

according to Stem-Volmer relation2 in equation 3 derived from eq 1 and 2.

kobs = k0 + kpyr[pyridine] + kx[oxygen donor] eq 1
(I) = kpyr[pyridine] /(kpyr[pyridine] + kx[oxygen donor]) eq 2
(Do/(D = Aoylide/Aylide = l + kx[oxygen donor]/kpyr[py1idine] eq 3

1 FCC absorbtion reference

2 Platz, M. S.; Modarelli, D.A.; Morgan, 8.; Toscano, J. P. Progress in Reation Kinetics;
Rodgers, M. A., Ed.; Elsevir: Oxford, 1994; Vol. 19 , p 93 and reference therein.

153

 

hv 355 nm 0

MeCN " Oxygen donor + Py or
————___> _ - '
32 —"——*_N2 @- k, A. NM°M°'P“°""9
h l '

1PCC:
A = 318 nm

1Q

C.>'~‘©

A = 475 nm

pyridine N-oxide (PNO) 0-—¢\© -
Oxygen donor< .
i

I".
N-Me-morpholine N-oxiddVlNO) Oe—PO

Figure 2.10 Rate measurements by ylide probe method for PCC: + N-Oxides donors

 

In eq 2 and 3, (D is the quantum yield of pyridinium ylide formation in the absence of a
second trap and (I) is the quantum yield of the of pyridinium ylide formation in the
presence of oxygen donors. Aoylide is the intensity of absorption of pyridinium ylide in
the absence of the oxygen donor while Aylide is the intensity of absorbance of the same
species in the presence of oxygen donors. The plot for the reaction of PCC: with
pyridine-N-oxide is shown in Figure 2.1.x. The slope of the line is equal to
kx/kpy,[pyridine]. The values for kx were than calculated by using the known value for

kpyr = 3.5:... x 108 M'lS’l (from ref 18), and known concentrations of pyridine, following

eq 5 and 6 derived from eq 3.
kx[0xy gen donor]: [(Aoylide/Aylide) ' llkpyrwyridine] eq 4
kx = [(A°y|idJAyndc) - l] kpyr[pyridine]/[oxygen donor] eq 5

Bimolecular rate constants of 2.2-L. x 109 M‘IS'l and of 2.1 x 109 M'13‘1 were obtained
for PCC: abstracting oxygen atom from PNO and MNO, respectively.
Alternatively, kx can be obtained as the slope of the plot of observed rate constant of

Either the decay of the carbene using the kinetic expression in eq 4 or the growth of

154

pyridine ylide versus oxygen atom donor concentration using the kinetic expression in eq
1.
kobs = k0 + kx[oxygen donor] eq 6

Such a plot constructed from the pseudo-first order rate constants for the decay of the
PCC versus [oxygen donor] in acetonitrile yielded a value of kx = 2.2 x 109 M’IS’l for N-
Me-morpholine N-oxide which is exactly teh same with the one obtained by Stem-
Volmer quenching of the yield of ylide.

Fl: rate constants for the oxygen atom transfer from N-oxides to spin-equilibrated
fluorenylidene Fl: have been measured by Stem-Volmer quenching of the acetonitrile
ylide of Fl: in acetonitrile solution. After irradiation of DAF in Ar saturated acetonitrile

solution three transients are observable, centered at 400, 470, and 500 nm.3

 

MeCN
3Fl: ' PM"
A: 470 nm A = 500 nm
hv 355 nm U
MeCN 0 en do or
DAF———> 1Fl:-flJ-——n> Fl=O + Py or
-N2. N-Me-Morpholine
MeCN
Fl <--NC-Me
A = 400 nm

Figure 2.11 Rate measurements by ylide probe method for Fl: + N-Oxides donors

 

3 a).Zupancic, J. J.; Schuster, G. B.; J. Am. Chem. Soc. 1980, 102, 5958.

b).Senthilnathan, V. P.; Platz, NM. 8.; J. Am. Chem. Soc. 1980, 102, 7637.

c) Wong, P. C.; Griller, D.; Scaiano, J. C. J. Am. Chem. Soc. 1981, 103, 5934.

d) Zupancic, J. J.; Schuster, G. B.; J. Am. Chem Soc. 1981, 103, 944. e) Griller, D.; Montgomery,
C. R.; Scaiano, J. C.; Platz, M. 8.; Hadel, L. M.; J. Am. Chem. Soc. 1982, 104, 6813. f) Griller, D.;
Hadel, L. M.; Nazaran, A. S.; Platz, M. S.; Wong, P. C.; Savino, T. G.; Scaiano, J. CJ. Am. Chem. Soc.
1982, 106, 2227.

g) Scaiano, J. C.; McGimpsey, W. G.; Casal, H. L. J. Am. Chem Soc. 1985, 107, 7204. h) Barcus,
R. L; Wright, B. B.; Leyva, B.; Platz, M. S. J. Phys. Chem. 1987, 9], 6677.

155

The half life time of spin equilibrated Fl: in acetonitrile is ~17 us which corresponds to a
rate constant of 2.1 x 106 M’lS'l, 4 which is assumed to be the rate constant for the
formation of the F l: acetonitrile ylide. Using Stern -Volmer methods similar to that
described above for pyridine, we measured the rate constant of oxygen abstraction from
pyridine N-oxide and N-Methylmorpholine N-oxide. Fresh prepared deoxygenated
solution of DAF were used for each concentration of oxygen donor. The rate constant for
oxygen abstraction was determined to be kx = 2.2 x 109 M'15'1 for N-Me-morpholine N-

oxide.

4 Jackson, J. E.; Soundararajan, N .; Platz, M. 8.; Doyle, M. P; Liu, M. T. H. Tetrahedron Lett.
1989, 30. 1335.

156

 

 

22.2 22.2- 232- 2.2- 2.8 22 2.2 22 2.2- 2 5
222 22.3.- 22.2- 22- 2.2 2.8. 35- 22 a9.- 2 6
22.2- 22:.- 22? 23.2- 22 E2- 2.2 we 2.2.- Hm 6
23- 22.2- 22:- 22- :22.- 2;...- 22- NS 2:- H m
28.:- 222- 8.2- 2.2.- 2.2- 8.2- 2.2- 22 23. E m
22.2- 22.2- SE- 2.2- 2.2:- 2.2- 22.2- 22 NE- 5 m
2.2 2.2- 22.2- .1.9.- 29. 2.2 8.8 NS 2.2- H m
222 22.2- 320- 2.2- 3 2.2 3.2 we 22- 2m m
2:; 23.22- :..B- 3.3.- 8.2- 2.2 2.2 2.2 22- 6 m
22.2. 22.2- 2.22.- 23:- 222- So- ga- 8- 2.2- m m
o
2.2- :..2 232 NS 2.2- H 2
2.2- 2.2 3.2 22 2.2- Hm Hm.
2.2- 2.2- 22 2.2 2.2.- 6 6
2 32- 22- $- 2.:- m m
2.2- 2- 2.2 :2 2.2- m m
anew—2m 3.838 N
beam-2E5 ape—am 22.221me .28 225cm ego 28222”; C525 96 086 w x

 

.085 2632 mo 8: .3 warns“ 5mm 0:820 o<

157

 

2:2 22.8- 2.2- 22- 82 22 2.22 2.2 22- 28.3 2.8 8.22 2.2- 2m
8:.- 23- 23- 28.8- 22- 2.2- 8 22m 22- 2.8 3.2 8.5 2.2.- So
28.:- 222- 8.8- 88.8- 82.2- 2.2 8.2 2.2 2.2- 322 8.2 2.2 2.3.- .m 8
22- 22 2.8 2.2 2.8- 2.8- 8.8- 2.8- 22- 2.2- 2.2- 8.22 2:- 2 m
88.:- 282- M2- 3- 2.2- 8- 222- 2.8- 2.2- 2.8- 8.2- 2.2 2:- .m m
222- 228- $.22 222 23:- 22- 2.8:- 28- 2.2.- 222- 2.8- 2.2 2:- 8 m
28.2 23:- 2.8- 2.2- 222- 2.2 2.2: 22: 2.2- :82 8.8 8.22 22- H m
222 28.2- 2.87 2.9.- 222. 2.2 8.2 mo: 22- 222. 3.2 2.2 n8- 8 m
28.2 23.2- 2.22- 28.2- 22.2- 2.2 8.8 .8: <9.- 23: 2.2 2.2 8.8- 8 m
22; 23 2.8- 2.2- 22.: :8 8.2- 3: NE- 28.2- 2.2- 2.8- 8.8- m m
82 o ~32 8.5 2.2- 222 2.22 8.5 2.2- H H
2.2- o 8.2 2.2 2.2- 8.8 8.2 2.2 2.2- as
2;- o 22 2.2 $2.- :.22 2.2 2.2 3...- 8 8
2.2 o 2.22- 2.8- 2:- 2.2- 2.22- 28- 2:- m m
2.8- o 2.2 3: 22- 22 2.2 mo: n8- m :
goEEMQ Ammm+_ de NVU
26mm $5223 2. *3 .28 2mm Ema +96 38 096 082 +038 36 0&6 > x

 

.985 2882 no 8: .3 28328 Sam 02; S.

158

 

 

159

8.8 2.2 8.8 8.8H 8.88 8.82 8.2 H E
8.8 8.8 8.8 8.8H 8.82 8.88 8.2. H U
8.2. 2: 8.8 8.8- 2.88 8 8.88 88H 6
8.2- 2.8 8.8H 8.8- 8.8H- H88- 2.88- H ”H
8.88- 2.8- 8.2 8.88- 888- 8.88- 88- 8m ”H
8.8- H.8 8.: 8.2- H.H-8- 8H8- 8.88- 6 --H
8.2 H8- 8.2. 3. 28 8.88 8.8H H HH
8.8- 8.8H- 8.8H- 8.8H- 8.82 8.88 H.8H 5 HH
H8 2.8- 8.8H- 8.8 8.88 38 8.88 8 HH
H.8 8.8H 8.8- 8.2 2.88- H88. H88 8 HH
8.: 32 8.H8H H H
8.8 8.2 8.82 88H 5
H8 8.8 8.2 Ho 6
8.88 2.888- 8.88H- m "H
8.88- 8.2 H88 HH m

guaEmw—aohflz
-288. 88m 88381888 .23 88m 36 +0888 82.882 m4 8 x

 

boofi 8382 8o 08: 3 8.83858 ”v-53 o2; w<

Bl Calculated energies for C02 + :CH2 complexes.

 

 

AE(kcal/m01)a HF MPZ QCISD B3LYP
Singlet 1C2v -3.1 -3.3

Singlet 1C1 33 -4.3

Triplet 3C -0.6 -1.1 --1.1 -l.l

 

a Calculated at the same level of theory, in respect with C02 + 1:CHz or 3ICH2.

B2 Seleced IR frequencies (nm) of the possible intermediates.
1 12 323 323b 33 p
1210 1168 1210 vs 1204 1225 vs

1456 1424 1425 1407 1408
1974 vsa 1949 1879 vs 1540 vs 1667 vs

2967 2991 2973 2987 2989
3058 3142 3103 3120 3099

 

 

a very strong

B3 Experimental IR data existent in the literature.

VC=o (cm'l) Vc—o (cm‘l)

 

di-Methyl oc-lactonea 1900
di-(Trifluoro) (at-lactonea 1975
di-n-Butyl-oz-lactonea 1895 l 163
1 (oz-lactone) 1974 HP; MP2 1210
328 1879HF; MP2 1210
331) 1667 HP; MP2 1225

 

a From Chapman, 0. L.; Wojtkowski, P. W.; Adam, W.; Rodriquez, 0.; Rucktaschel, R. J. Am. Chem. Soc.
1972, 94, 1365-1367.

160

B4 Intermediates geometry data

 

alfa-lactone ylide diradical diradical diradical
l 12 323 325 33p
Bond Length
O(l)-C(2) l.170;1.200 1.148;1.190 1.161;].193 1.199;1.168 1.191;1.208
C(2)-O(3) 1.302;].342 1.225;1.220 1.322;1.355 1.335;1.304 1.337;1.341
O(3)-C(4) l.480;1.539 1.353;1.460 l.376;1.389 1.410;].392 2.329;2.304
C(4)-H(5) 1.074;l.085 1.070;].090 1.070;1.078 1.080;l.080
C(4)-H(6) 1.074;l.085 1.069;1.090 1.073;].080 1.080;].071
C(2)-C(4) 1.453;1.448 2.309;2.324 2.335;2.323 1.453;1.476
Bond Angle
O(l)-C(2)-O(3) 139.5;138.9 146.4;1520 128.1;126.4 129.9;130.8 121.1;122.6
O(l)-C(2)-C(4) 155.2;154.2
O(3)-C(2)-C(4) 65.3;66.8
C(2)-O(3)-C(4) 61.7;59.8 143.1;147.0 117.6;115.7 116.5;1189
C(2)—C(4)-O(3) 53.0; 53.3
C(2)-C(4)-H(5) 120.7;120.9
C(2)-C(4)-H(6) 120.7;120.9
O(3)-C(4)-H(5) 113.8;113.2 113.0 lll.8;lll.8 116.5;115.7
O(3)—C(4)-H(6) ll3.8;113.2 110.3 116.8;117.1 116.1;115.7
H(5)-C(4)-H(6) 116.7;116.6 124.4;114.0 121.5;123.0 123.9;122.7 120.8;120.9
Dihedral
O(l),C(2),O(3),C(4) l80.0;180.0 130.6;1340 l78.6;177.9 0.0;0.0
O( 1),C(2),C(4),O(3) 180.0; 1 80.0
O(l),C(2),C(4)H(5) 81.9;82.7
O(l),C(2).C(4).H(6) -81.9;-
82.7
O(3),C(2),C(4),H(5) -98. 1 ;97.3
O(3),C(2),C(4),H(6) 98.1;-97.3
C(2),O(3)-C(4),H(5) 62.7 174.8;167.9 79.3;76.98
111.5;112.2
C(2),O(3)-C(4)-H(6) 111.5;112.2 152.9 28.2;18.2 79.3;76.98

 

a values optimized at HF/6-31G* level
b values optimized at MP2/63 lG*//MP2/6-3 10* level

161

883858 H2285;

3 35830 8265328 bacmeEH Ho 83:5: 05" $252 8H 5on 83 2582-828 OH 800» Ho>oH tug-OE: H8 HooNHEHHnHo 23258 Hoe-8388 3823888 a

 

 

HoocHov. hm H hm H v. ovmmmcw. vmmoHv. 2.2-own. ogamavfimm- NO

HmmcmNN. omcova. commmm. cm H 33. moowwem. meow-H Hm. mmHVwmmHNN- H.H-V9900
Nwmmmwc. H mE-wo. omowwoo. NoooHoH. mbmoooH . manage. £583 .83. *0 H méEHEVNnHHZ
NO

@3va H. H H m MNNH . 83.3 H. @833. H HwH-QNH . wmomvmo. 53m H NHNN- *OH m-o\NnHH>H\\*OH méhkuz
©3988. 3 £8. mbwtho. 383. 953. 235. $882 .bmm- $8 38 53, *OH WEN;
ommgoo. Nwmwmho. bmHOVwo. mmchmo. H Hmmmwo. 23H Duo- amwo. 2.1mm: .hmm- *DH m-o\N8HH2\\*0H Wag
899.8% mm:

1582. oocowH H. gonH. mHawomH. mummamH. NmooHoH. mmcH. Hogaomfimm- *OHm-o\nHwH\\*OHm-o\v8HH\/H
cwmmcmo. ooomono. SH ammo. womohoc. chNH-oo. mowmmo. NH- H o. 2.85: BNN- *OH 8.6855an méRmSH
oH avomv. wvvmmom. thH Hm. cvwmwv. whmgv. omchm. came. mmvvmmdmm- ES 0% 5H3 *0H mbEHH
cmmwmnv. wwmbomm. H Howmm m . uncammm. Seaman. 0382.. 828. 05.000ch *OH méEH-H
EoEoow "HE

H H o o o H o n OSZHZ

8 an H 8mm .38 8N8 «~8- mmh u NH H8 HH H8052: oHHHE n<

 

8onSHHo «EH—HHS 6383280 Hm.m< oHnm-H-

162

 

 

H emwcv. wow H mm. mamcmm. Shown. hmvomv. NO
HE-oom. mmommm. NmoooH. oomH Hm. 85R. H.H-V9500
@322. w H ammo. megs. owmmoo. ovcmmH. *OH Wokzsbmfiz
ND
5809 H. MVNNH H on NNH. mom: H. mH mwvo. vomvao. oocoH. *OH m-c\NmH>S*OH @9va
oH oH mo. w H NHo. mamvo. 895. 38ch mmm H oo. 03%. 68 can 5:5 *OH m-c\mnHSH
bmwmvoo. mpg. $55. 888. mmw Hoo. wcH go. abomH. *OH m-o\mg\*OH TEN;
Eofioom NnHH>H
moot; H. mwomH H. awe: H. 38H H. ong. 823. mohmH. hmmHg-NN- *OHm-c\nHH-H\\*0Hm-o\v&2
mmmwocc. Swag. H 3% no. CNH-co. vm H 8a. mowmmo. HovH H. owwofimm- *OH m-o\nHH-H\\*OH mA-VRAHSH
mvH mg. mmvomv. $2 wv. mcmhv. cmomwm. amchm. 38858. 5898.8” .888 0&8 5H3 *0H m-o\nHHH
mH oomom. mm H 8898. H mwwH m . 500%. NH 39. omeHv. N H mom. N H cmdmml *OH méEH-H
bHoEoow EH
am8IMHm mmlam Honm 8N8I8Nm N8I~H8 nHHImH 8H HIH a HIM:
wmh-m mmhm mmhm .8um vmhm nmPH NmBH HmHH

 

82325 8238 282858.:- HooHaHHHonu cm

163

B7 Cartesian coordinates and energies for the species in Chapter 3

‘CH, singlet at GZ:
Energy GZ=-39.0583994; NImag=0
H 0.862245 0.000000 0.523019
C 0.000000 0.000000 0.174340
H 0.862245 0.000000 0.523019

3CH2 triplet at GZ:
Energy GZ=-39.069004l; NImag=0
H 0.9823994481 0. 03313065703
C 0. 0. 0.1104355234
H 09823994481 0. 03313065703

CO singlet at G2:
Energy G2=-113.1774964; Nlmag=0
O 0. 0. 0.4929593158
C 0. 0.-0.6572790877

CO2 singlet at G2:
Energy GZ=-188.3613201; NImag=0
O 0. 0. 1.1432643
C 0. 0. 0.
O 0. 0.-1.1432643
'HzCO singlet at G2:
Energy GZ=-l 14.3389107; NImag=0
H 0. 09341725696 -l.1246980027
C 0. 0. 05364297749
0 0. 0. 0.6834968319
H 0. 0.9341725696 -l.1246980027
3HzCO triplet at GZ:
Energy GZ=-114.2170052; NImag=0
H 0.961359272 0. -1.0956690211
C 0. 0. 06060668323
0 0. 0. 0.7284673795
H 0961359272 0. -1.0956690211

lOl-Laetone singlet at G2:
Energy G2=-227.4939909; NImag=0
0,08959712153,-0.5939595636,-0.1256176739
C,0.5837534797,-1.0105930933,-0.2137323168
H,0.891985264,-1.6711811933,0.5908347229

164

C,0.072856026,0.3 155835795,0.0667433906
H,0.89l985264,-1.2888680055,-1.2168613377
0,0.18051777,l48522284880314]126953

3Ol-Lactone triplet at HF/6-31G*

Energy HF =-226.4705866; NImag=0
0,-0.8242034733 ,-0.03 591 89632,-0.63 70710623
C,-0.7831445652,-0.0538373395,0.78l145306
C,0.3508133382,-0.4006542725,-0.0623133832
H,-l .3444676895,-0.8437410168,] .2450167176
H,-0.8304012471,0.9156196955,] .2435122308
0,1.4203105106,0.3678028374,-0.2131 189983

’Zanti triplet intermediate
Energy GZ=-227.410324; NImag=0
0,3
H,-2.5023236365,-0.4823244738,0.1644361944
0,-0.45806831 l 1,-0.4655363305,0.0093331549
C,-1.6849812127,0.1818084037,-0.065561153
C,0.6331566642,0.3367752059,-0.01 18725154
0,1.7722841397,-0.0177147775,0.0072681338
H,-1.700455701,l .2368316795,0.l673555057

32syn triplet intermediate
Energy G2:-2274093249; NImag=0
0,3
0,1.4253821599,0.4630913079,-0.007l 175146
C,0.751 1984437,-0.5270083266,0.0074967265
0,-0.5832349449,-0.6654934857,0.01 14303727
C,-1.336936391,0.5055555979,-0.0624886475
H,-0.836562224,l .4267137209,0.1973 556742
H,-2.3861 878121,0.3212200734,0.098092987

33planar at G2
Energy GZ=-227.4157187; NImag=0
0,3
0,-0.375442133,l.2846508313,0.
0,1.2943980737,-0.2016641477,0.
C,-0.0018625352,0.13588989730
C,-O.8374173368,-1.0784320707,0.
H,-1.9083137909,-0.938l470621,0.
H,-0.4076545035,-2.0704933663,0.
'3planar optimized at HF
0,1

165

HF=-226.5667565 (or-lactone)

33staggered at G2
Energy GZ=-227.4016901; NImag=1
0,3
0,-134195800170.10555315330.
0,0.6293 799304,] . 1 559303 504,0.
C,0.0000595279,0.1 1 108638650.
C,0.6102690829,-125243225690.
H,1.0193264527,—1.6218964033,0.9305417839.
H,1.0193264527,-1.6218964033,-0.93054l7839.

3C, triplet complex
Energy QCISD=-227.1295055; Nimag=0
0,3
0,0.8756760345,l .1886337984,-0.0270843777
C,0.89024236l8,0.0157910339,-0.0284213522
0,0.917907612,-1.15681 1884,00301776463
C,-2.352934073 8,-0.041949024900763475906
H,-2.801982751,0.9429507083,0.084251884
H,-2.770536l489,-1.0405780772,0.086286877

Energy B3LYP= Nimag=0
Energy MP2= Nimag=0
0,3

0,0,0.8539094176,1.1845488299,0.0018501063
C,0,0.891 8255514001 169419180001454723
0,0,0.942720152,-1.1607354612000201 14075
C,0,-2.3756958447,-0.0303045853,-0.0667543975
H,0,-2.7592109412,0.951464433801820340007
H,0,-2.7106038559,-1 03030902230 1788719354

lC2V singlet complex at
Energy HF = -226.511438965 NImag=0
0,1
C,0,-0.8664454162,-00007494801,-0.0043767l94
0,0,-0.901926071 1,-0.0037133389,].1388563473
0,0,-0.8904147086,0.0021623787,—1.1479086167
C,0,2.2729097566,0.001 975 73 56,001 14728044
H,0,2.9506799921 ,-0.85 83 854299,0.01 26823019
H,0,2.9492602034,0.8634355788,0.0171593429

Energy MP2=-227.0830709 NImag=O
0,1

166

C,0,0.,0.,2.] 881324482
C,0,0.,0.,-0.8360577795
0,00,-1.1790122496,—0.867104503
0,00,] .1790122496,-0.867104503
H,0,0.8646349464,0.,2.8806l20176
H,0,-0.8646349464,0.,2.88061201 76

1Cl singlet complex at

HF

Energy HF = -226.51 14387; Nimag=0
0,1

C -0.0569508 0.0013784 0.8629406
0 0.9029210 0.6165304 0.9550486
0-1.0207462 0.6136905 0.8305619
C 0.1492875 0.0035842 -2.2639478
H 0.6560849 -0.7306272 -2.9074249
H -0.2675024 0.7211421 -2.9714159

'TSl singlet at HF/6-31G*
Energy HF=~226.5012017; NImag=l
0,1
0, 0.301752, 1.007470, 0.000018

, 1.370051, 0699051,0000038

C, 0480738, 0.127541,-0.000013

H, 1.951778, 0.449479, 0.879895

H, -1952125, 0449226, 0879668

0, -1.456724, 0466499, 0000008

0

lT82 singlet at G2

Energy G2=-227.4504276; NImag=1

0,1

0, 1.0392416675, 07871004626, 0.0000239426
C, 0.9280945472, 0.5364713846, 0.0000163445
C, 05694607556, 0.3557411034, 00000123948
H, 1.2145872439, 1.1227218537, 08977236483
H, 1.2145517058, 1.1227292398, 0.8977628512
0, -1.6118593799, 01627402901, 00000318053

'TS3 singlet at G2
Energy 02=-227.3807067; NImag=l
0,1
0, 0240963, 0.096460, 0.417699
C, 0337004, 0.263084,1.830369
C, 0.304502, 0.077635, 0662286

167

H, 0.142866, 0.685541, 2.325076
H, -1.364973, 0.607747, 1.913069
0, 0.453820, 0.149355, -1.823529

’TS4 triplet at G2
Energy GZ=-227.3363951; NImag=1
0,3
0,-1.5945]84263,0239260071700593604838
004296289217,00514933084,—0.0948726008
0,0.466431723608913497836,-0.0030400483
C,1 .3589]6897,05273222367,-0.046422l 882
H,1.18591 18274,-] .6038979562,-0.0095036595
H,2.2630539427,01399264683,0.4067089096

3TSS triplet at G2
Energy G2:-2273818684; NImag=1
0,3
0,0.3707816098,00124371966,-l .8868089525
C,0.0015124485041836094]8,-0.8833650152
0,0.1258129012,0511084552707155526436
C,05185459254,0.1290582855,1.72403529
H,06321331266,1.2030216962,1.643425476
H,02384220999,02993630649,2.6826033468

3TS6 triplet
Energy GZ=-227, 381868; NImag=1
0,3
0,0,0.1356665204,1.2486699258,01728921383
C,0,0.301072972500420102471001 58506409
0,0,1.0731065215,-0.8595885815,0.2161984005
C,0,-1.3321 14l239,-0.3664039992,-0.094549789
H,0,-2.1006874071,0.2782145297,0.2983021066
H,0,-].383250019,-1.444502771 1,01725573147

3TS7 triplet
Energy HF=-226.5100422 NImag=l
0,3
0,0.5 106279,-0.5643490667,-0. 14254
C,1.5111329,0.3602559333,0.092737
H,1.2187859,1.39371193330074726
H,2.4383529,0.0543259333,-0.346296
C,-0.699625 1 ,-0.2594940667,0.3503 98
0,-1.5764011,0.3077729333,-0.155865

168

0,3
O,-0.666874376608869762300001585899
H,-l .5427672282,-0.883307246,—0.9213850893
H,-1.5426362837,-0.883518177809214104737
C,-1.0557021867,-0.5607374519,0.000015085
C,0.297309263,-0. 1 2823 63 79,-0.00003 83922
0,1.6213445084,-0.1493926788,-0.0001442825

3'I‘Santi triplet at MP2
Energy MP2=-227.070205; NImag=l
0,3
H,-1.5474173587,-1.0004476722,0.8225714339
0,0520877236606087875765,00928277669
C,-1.4946880499,-0.387360041,00692186924
C,0.716961599,0.2086494615,0.3935135451
0,1.595608974,0340685056,01791558531
H,-2.3840778348,-0.0721090141,05924715899
HF=-226.5064849\MP2=-227.0702055\PUHF=-226.51 18838
PMP2-0=-227.0741958\S2=2.039\SZ-1=2.019\82A=2.001\RMSD=4.

3TSsyn triplet at MP2
Energy MP2=—227.0787589
0,3
0,-1.3824336307,-0.52778087620.
C,-0.7664381956,0.50185350020.
0,0.5528755691,0.7109712084,0.
C,1.3721 14\6553,-0.4367343504,0.
H,1.5012028674,-0.928l 187787,-0.9531 168462
H,1.5012028674,-0.9281 187787,0.9531 168462

169

 

 

8.88 2828.8 888882.888 898588282
8H8 88888.82 88888.888 8.8: H8882
8888888888 8988282
888828.888- 88828.8 28888888- H2: 80888882
88282888 2.88 8+88H.
8888.888 888828.888 88:8888 8.8 8888.8 888888.888 88888.2
828888.888 888888.888 8888.88 8.88 888288 2888.888 8.88 8886282282
88888888 8.88 88288.8 888888.888 8.88H 88828882
8.88 88288.8 2888.888 8.88H 8888858882
888888882 2.88H- goo-825825
828888.888 8.82 H8825
8828888888 8.88H- 995
8.88 882888.888 888828.888 288:8.888 8H8 88888.8 8888888888 38
888.888 8.88 288288.888 88888888 888888.888 8.88 8.8 888888.888 88.88 888888828
828888.888 888888.888 882888.888 8.88 88828.8 288.888 88.88 88:
888288.888 888888.888 888888.888 8.88 88828.8 888882.888 H28 85888280
828888.888 88888888 88888888 88882888 H82 9888888.
88.8
888.88 8882888 8888.888 8.88 8828.8 888888882 8.8 88.88 8882::
88888.88 8888888 228888.88 888882.888 8.82 88888888
88288888 888888.888 888888.888 8.88 88228.8 288822.888 8.88 858.82
82888888 28288.888 888888.888 2.88 2882888 888888.888 8.88 8.8: 880
888888.83 888888.83 2888888: 8.88 88888 888882: 8.28 Bo
88288H.22 88288.22 88888.22 88888.22 8.8H 88888.8 828828.82 2.88 H88 moo:
98880 888 6888882 88 882 88 882 88 888 58-882: 828 88m 8888

 

8 229850 :H 828mm 85 9c 8on5:0 H88H8H=oH80 H.888 035.

170

C2 Mass fragmentation for 1 and 3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 57.9 152.9 184 .1 M9000 NH
109 I
NC COOMe
41.1 12' ’-
/
l
zrs«
“tr-1
124.1
{53.3 ‘ 1185':
59.2 126.2
I /
55.2
\ 68.181'193-1 16 .1 1, '9
9,, . ”‘11 IIJL .11111Lu'1911:1 .1 .'1. ‘. v 11. . . '1 I . . 221': I
HI: 4 so so 199 129 149 160 199 299 220 2: o
7 o
1139- 9
.\1Bu
1" MeOO RH
MeOO C§N
'l
‘J
ZFS<
1s .9
12 2
I332:- 124 1
’ 69.9 82.193-9 ‘1 139.0 )54'2 ”T“ 22?.124 .1
a-LL ”'- -'11 .111. . I'. - .I . . J“. . . - - L
n/z so no 199 129 1413 .169 1913 2130 220 240

 

 

 

171

HICHIGRN S

lWWM
3‘12

1

|
9

7an UNIV. LIBRARIES
IWWWWWWWWW
018017339

1
3