Citation
Synthesis and characterization of the first homoaromatic organometallic carbene complexes

Material Information

Title:
Synthesis and characterization of the first homoaromatic organometallic carbene complexes
Creator:
Radcliffe, Marc Dudley, 1953- ( Dissertant )
Jones, William M. ( Thesis advisor )
Battiste, Merle A. ( Reviewer )
Dolbier, William R. ( Reviewer )
Dose, Eric V. ( Reviewer )
Bergeron, Raymond J. ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1982
Language:
English
Physical Description:
vi, 83 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Carbenes ( jstor )
Carbon ( jstor )
Chemical equilibrium ( jstor )
Cots ( jstor )
Electronics ( jstor )
Electrons ( jstor )
Ligands ( jstor )
Photolysis ( jstor )
Protons ( jstor )
Signals ( jstor )
Carbenes (Methylene compounds) ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
Protonation of (n -cyclopentadienylirondicarbonyl) cyclooctatetraene with strong acid produced the novel homoaromatic carbene complex n. - cyclooctatrien-1-ylium-n -cyclopentadienylirondicarbonyl (FpHC0T + ). Examination of the H nmr spectrum of this compound revealed a decrease in the homoaromaticity of the homotropylium ligand when compared with homotropylium itself, consistent with the expectation that significant d-ir pit backbonding interactions into the positively charged ligand by the organoiron occurs. Substitution of triaryl- or trialkyl-phosphine for a carbonyl on the cyclopentadienyliron in the carbene complex sequentially reduces the degree of aromatic derealization in the homotropylium ligand, providing clear evidence that significant increases in the backbonding character of these complexes occurs when the electron density on iron increases, as expected from our earlier work. Furthermore, the three sets of proton coupling constants in the homotropylium ligands reveal increased bond alternation as the other organic ligand is changed from CO to P(Ph) 3 to P(Bu)_. This bond alternation was used to substantiate the relative measure of the aromatic ring current - the difference in proton chemical shift between the homoconjugate exo and endo protons (AS). When the AS of the three homoaromatic carbene complexes was compared with that of several other substituted homotropyl i urns, an approximate measure of the (pi) backbonding ability of the cyclopentadienyliron group relative to normal organic functional groups was obtained. The thermal decomposition of FpHCOT appears to proceed through a 1,2-8 hydrogen shift mechanism. However, the complex has been isolated as the hexafluorophosphate salt and could be manipulated if kept cold. That these carbene complexes having B hydrogens are i sol able at all was postulated as being due to both the stabilizing effect of the homoaromaticity, and the nearly orthogonal orientation of the 6 hydrogen bonds and the carbene (pi) system. Additionally, the isodynamic interconversion processes of three cyclopentadienyliron substituted cyclooctatetraenes were studied by nmr. The initial data appear to indicate that an unusual interaction between the transition metal and the 4n i\ electron ring system may be present.
Thesis:
Thesis (Ph. D.)--University of Florida, 1982.
Bibliography:
Includes bibliographic references (leaves 80-82).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Marc Dudley Radcliffe.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
028900508 ( AlephBibNum )
09617123 ( OCLC )
ABW9321 ( NOTIS )

Downloads

This item has the following downloads:


Full Text











SYNTHESIS AND CHARACTERIZATION OF
THE FIRST HOMOAROMATIC
ORGANOMETALLIC CARBENE COMPLEXES








BY

MARC DUDLEY RADCLIFFE
















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



UNIVERSITY OF FLORIDA















ACKNOWLEDGEMENTS


No acknowledgement can adequately express my gratitude to Dr.

William M. Jones, whose professional leadership and personal encourage-

ment have greatly increased the satisfaction and rewards of my graduate

studies. No past experience has offered greater intellectual freedom,

nor more equitable guidance than found here under his counsel. Moreover,

he has attracted several outstanding individuals to this research group.

William R. Winchester, Frank Manganiello and James Lisko have all contri-

buted significantly to the character of this academic experience and I am

grateful for their stimulating conversations and friendships.















TABLE OF CONTENTS



ACKNOWLEDGEMENTS .............. ........ ii

ABSTRACT . . . . . .. . .. .. . ... . v

CHAPTER

I INTRODUCTION .................... 1

II RESULTS AND DISCUSSION . . . . . . . 9

III EXPERIMENTAL . . . . . . . .. . 68

General . . . . . . . . . . . 68

Cyclopentadienylirondicarbonyl Iodide ...... 69

Tri-n-butylphosphine. . . . . . . . 69

Triphenylphosphine . . . . . . .... 69

Bis(1,2-diphenylphosphino)ethane . . . ... 69

Trifluoroacetic Acid . . . . . . . .. 69

Triphenylcarbenium Hexafluorophosphate . . .. 70

Fluorosulfonic Acid . . . . . . . . 70

Potassium Cyclopentadienylirondicarbonyl . . .. 70

Bromocyclooctatetraene . . . . . . .. 70

Cyclooctatetraene Carboxylic Acid . . . ... 71

Cyclooctatetraene Acid Chloride . . . ... 71

(n5-Cyclopentadienylirondicarbonyl)-
cyclooctatetraene Carbonyl (11) . . . ... 72

Photolysis of 11 .................. .72

(n5-Cyclopentadienylirondicarbonyl)-
cyclooctatetraene (9) . . . . . .... 73









(n -Cyclopentadienylironcarbonyltri-n-butyl-
phosphine)cyclooctatetraene (12) .. . . . 74

(n -Cyclopentadienylironcarbonyltriphenyl-
phosphine)cyclooctatetraene (13) . . . ... 75

(n -Cyclopentadienyliron[bis(1,2-diphenyl-
phosphino)ethane])cyclooctatetraene (19) .... .76

n -Cyclooctatrien-1-ylium-n5 -cyclopentadienyl-
irondicarbonyl Hexafluorophosphate (10) .... .77

1 -Cyclooctatrien-1-ylium-n5-cyclopentadienyl-
ironcarbonyltri-n-butylphosphine Hexafluoro-
phosphate (17) . . . . . . . . 78

n1-Cyclooctatrien-1-ylium-n5-cyclopentadienyl-
ironcarbonyltriphenylphosphine Hexafluoro-
phosphate (14) . . . . . . . . .. 79

REFERENCES . . . . . . . .. . . . 80

BIOGRAPHICAL SKETCH . . . . . . . . ... .. 83















Abstract of Thesis Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



SYNTHESIS AND CHARACTERIZATION OF
THE FIRST HOMOAROMATIC
ORGANOMETALLIC CARBENE COMPLEXES

By

MARC DUDLEY RADCLIFFE

August 1982

Chairman: William M. Jones
Major Department: Chemistry


Protonation of (n5-cyclopentadienylirondicarbonyl) cyclooctatetraene

with strong acid produced the novel homoaromatic carbene complex n -

cyclooctatrien-l-ylium-n5-cyclopentadienylirondicarbonyl (FpHCOT ). Ex-

amination of the 1H nmr spectrum of this compound revealed a decrease in

the homoaromaticity of the homotropylium ligand when compared with homo-

tropylium itself, consistent with the expectation that significant di pi

backbonding interactions into the positively charged ligand by the organo-

iron occurs. Substitution of triaryl- or trialkyl-phosphine for a car-

bonyl on the cyclopentadienyliron in the carbene complex sequentially re-

duces the degree of aromatic delocalization in the homotropylium ligand,

providing clear evidence that significant increases in the backbonding

character of these complexes occurs when the electron density on iron

increases, as expected from our earlier work.









Furthermore, the three sets of proton coupling constants in the

homotropylium ligands reveal increased bond alternation as the other or-

ganic ligand is changed from CO to P(Ph)3 to P(Bu)3. This bond alterna-

tion was used to substantiate the relative measure of the aromatic ring

current the difference in proton chemical shift between the homocon-

jugate exo and endo protons (A6). When the As of the three homoaromatic

carbene complexes was compared with that of several other substituted

homotropyliums, an approximate measure of the T backbonding ability of

the cyclopentadienyliron group relative to normal organic functional

groups was obtained.

The thermal decomposition of FpHCOT appears to proceed through a

1,2-8 hydrogen shift mechanism. However, the complex has been isolated

as the hexafluorophosphate salt and could be manipulated if kept cold.

That these carbene complexes having B hydrogens are isolable at all was

postulated as being due to both the stabilizing effect of the homoaroma-

ticity, and the nearly orthogonal orientation of the B hydrogen bonds

and the carbene r system.

Additionally, the isodynamic interconversion processes of three cy-

clopentadienyliron substituted cyclooctatetraenes were studied by nmr.

The initial data appear to indicate that an unusual interaction between

the transition metal and the 4n i electron ring system may be present.













CHAPTER I
INTRODUCTION


The general class of organometallic compounds known as carbene com-

plexes are thought of as the formal result of allowing an organic carbene

species to approach a ligand substituted transition metal in such a way

that the filled sp2 orbital and unfilled p orbital of the carbene inter-

act with appropriate empty and filled orbitals of the transition metal to

form a compound of the two species. Although no carbene complex has been

produced by this procedure, alternative methods have been developed which

have allowed the isolation and characterization of several of these

species.1'2

One of the most widely used synthetic procedures was developed by

Fisher and Maasbbl.3 Heteroatom stabilized carbene complexes were pro-

duced by organolithium attack on a carbonyl ligand followed by alkylation

of the anionic intermediate.




RLi /R (Me)30BF4

n-0 +
M(CO)n (CO)M _M --C

Li




n(CO)n-M C (CO) M- C
SOMe n-1 OMe









The resulting neutral compound can be characterized as a neutral

metal carbonyl accepting electron density from the filled sp2 orbital

of the "carbene" to form a sigma bond. As a result of this bonding, the

now electron rich metal can release electrons in a d orbital with sym-

metry appropriate for "backbonding" into the unfilled p orbital on the

carbene carbon. Alternatively, the heteroatom substituent can compete

with the metal for bonding into this empty orbital. This or some other

form of additional r stabilization can substantially increase the stabil-

ity of carbene complexes. For example, the methylene carbene complex (1)

is known to have a transitory existence, but cannot be observed spectro-

scopically even at temperatures as low as -80C. However, phenyl


+
H


/ Fe Fe / Fe
CO --H CO -H or CO / C--H

CO R CO H CO




1 2

substitution at the carbene carbon atom stabilizes the complex to such an

extent that the benzylidene complex (2) is isolable.4

Both the above methylene and benzylidene complexes were produced by

protonation of the sigma bonded metal a-methyl ether to remove the

methoxy group as methanol. The remaining positively charged carbon can

be stabilized as indicated. Removal of a good leaving group has become

a versatile method of producing cationic carbene complexes.








A variety of additional carbene complexes have been generated

through other methods, including two general classes of aromatically

stabilized carbene complexes.5'6






(CO)5r + C > ---- (CO)5Cr =








H 6
Fe Fe
CO 03- CO 3C+
CO CO


3



The cationic cycloheptatrienylidene complex (3) was also formed by

removal of a negative leaving group, in this case a hydride, through use

of the hydride abstracting agent triphenylcarbenium ion.

A useful method of forming a cationic carbon site in carbonium ion

chemistry which has not apparently found application in metal carbene

chemistry has been protonation of an alkene. Such a procedure would gen-

erally produce an intermediate which would rapidly decompose by a 1,2

hydrogen shift to reform the alkene as a 7 complex with the metal.7












"D



Fp L//7 H
S Fp= Cyclopentadienylirondicarbonyl



All of these carbene complexes can be represented as resonance hybrids
of three canonical forms


M- M M-

1+ 1



A B C


each of which makes different contributions to the overall structure in
the (several) individual types of carbene complexes. A qualitative in-
dication of the contribution by these resonance forms can be obtained
from the measurement of barriers to rotation about the appropriate
double bonds. For example, a barrier of 12.4 kcal/mol has been observed
for the carbene carbon-oxygen bond rotation in pentacarbonyl[(methoxy)
(methyl)carbene]chromium (0)8 (page 1, M=Cr), as well as a 9.1 and 10.4
kcal/mol barrier in Cp(CO)2Fe=CHC6H5+ (4) and Cp(CO)2Fe=CH(p-CH3C6H4)
(5), respectively, for the Caryl-Ccarbene bond.9














Fp




C?\H
4^ a


&G = 9.1 kcal/mol


AG* = 10.4 kcal/mol


C


For CpFe(PPh2CH2CH2PPh2)=CH2+ (6) the barrier to rotation about the

metal-carbene bond of 10.4 kcal/mol was compared favorably10 with the

calculated electronic barrier.11




+
Fe



PO2

6


While these figures may not be strictly comparable, they do reflect the

fact that contributions by resonance forms B and C can be significant.

When X and Y in these resonance forms are the same and represent a

conjugated (4n +2)r electron ring system, the aromatically stabilized








carbene complexes (e.g. 3) result. Because of the aromatic delocaliza-

tion of the positive charge in the carbene ligand, contributions from

resonance structures having double bond character between the metal and

the aromatic ligand, representing metal dr pr backbonding, should be

substantially reduced.



M: M:









X-ray structures of CpFe(CO)2=C7H6+ (3) and CpFe(CO)2=benzoC7H4 the

aromatic tropylidene complexes produced by Allison and Jones, clearly

indicate a high degree of aromaticity and concurrent reduced backbonding

in these complexes. But when the iron-carbene bond length was compared

with two other model systems, only "some degree of metal-ligand multiple

bonding" was suggested.12

In an effort to better identify the contribution of metal-carbene

backbonding interactions in these aromatic complexes, rotational bar-

riers were obtained for the nonsymmetrically substituted iron carbene

complexes Cp(CO)PBu3Fe=C7H6+ (7) and Cp(CO)PBu3Fe=benzoC7H4+ (8).13
























The rotational barriers observed, 9.6 and 10.4 kcal/mol respectively,

were unexpectedly comparable to the much more strongly a bonding species,

CpFe(DIPHOS)=CH2+ (6), which had been found to have a rotational barrier

of 10.4 kcal/mol.10 Because the X-ray data for compounds 3 and

Fp=benzoC7H4+ indicated weak backbonding, it seemed likely that "the

principle restriction to rotation [in 7 and 8] was steric in origin with

a minor electronic component superimposed."13

At the time that this last work was being done, characterization of

a new carbene complex was being completed, the spectral properties of

which were seen to have significant bearing on the question of the rela-

tive contribution of metal i backbonding in aromatic carbene complexes.

Rather than stabilizing the positive charge in the carbene ligand

by aromatic delocalization, this new complex was thought to achieve a

large measure of its stability through homoaromatic delocalization.

While this is a difference merely of kind rather than of function, a

wealth of information was believed to be available from this system when

compared with the cycloheptatrienylidene complex. In particular, since

the difference in chemical shift of the inner and outer proton of homo-

conjugate CH2 linkages is thought to be a definitive measure of the extent









of aromatic delocalization,14 and since this delocalization will be re-

duced by competitive delocalization of the positive charge through reso-

nance contributions due to metal backbonding, it was recognized that a



H H H H

SM + ,M:




purely electronic measure of metal-carbene a bonding interaction could be

obtained from homoaromatic carbene complexes by the examination of the

difference in chemical shift of the homoconjugate protonsmentioned above.

A further point of interest is the fact that no other homoaromatic

carbene complexes have been reported in the literature.

This work will therefore describe the synthesis and characterization

of the first homoaromatic carbene complexes. Furthermore, a strong corre-

lation will be made between the i bonding ability of cyclopentadienyliron

groups and the + of normal organic functional groups.












q* +1

The Fpbenzocyclobutenylidene complex reported by Giering and Sanders15
is an unlikely candidate for homoaromatic delocalization because of the
required loss of aromaticity and increases in strain energy.














CHAPTER II
RESULTS AND DISCUSSION


The synthesis of the first homoaromatic transition metal carbene

complex was accomplished by the synthetic route shown below.


0 Br Br
Br


KO+


SBr


SFe Fe


CO CO

9 10

*16
Bromocyclooctatetraene was prepared as previously reported.16 Treatment

of the bromide with potassium cyclopentadienylirondicarbonyl7 produced

the (r5-cyclopentadienylirondicarbonyl) cyclooctatetraene (FpCOT) (9).

This reaction is unusual in that the Fp anion is normally used to


Except the dehydrobomination reagent, potassium t-butoxide, was dis-
solved in THF rather than added as the solid.


KFp









displace alkyl halides; however, it was noted that King had successfully

displaced fluoride from fluorobenzenes with Fp anion,8 and similar re-

action with BrCOT seemed feasible. The cyclooctatetraene sigma compleJ9

was formed via this reaction in higher yield and with more convenient

work up than our earlier method using cyclooctatetraenyllithium20 and Fp

iodide. The structural proof of the cyclooctatetraene sigma complex

rests on its well resolved IH nmr.spectrum and the observed conforma-

tional interconversion processes videe infra). Finally, the homoaromatic

carbene complex, n -cyclooctatrien-1-yliden-n -cyclopentadienylirondicar-

bonyl hexafluorophosphate (FpHCOTHFP) (10), was produced by protonation

of the FpCOT with strong acid and subsequent anion exchange using tri-

phenylcarbenium hexafluorophosphate, to yield a very thermally sensitive,

but air stable orange solid in 40% isolated yield from the sigma complex.

Because of the thermal instability of this carbene complex, the

structural proof relies heavily on the 1H nmr which is seen in Figure 1.

The single proton resonance of 8.9 ppm occurs at the characteristic low

field of protons adjacent to the carbene center6 and attests to the un-

symmetrical nature of the ring system in which the homoconjugate linkage

is also immediately adjacent to the carbene center.





H2
Fp


Other possible structures would require two proton resonances at this

characteristic low field, which is not observed. Furthermore, protona-

tion of the FpCOT adjacent to the substitution is expected since the








11


































-- (~
CM





3 *











0*
'- C







Q.
o u




EU
S-C



3 C)



e n
E i




C)5-









1 ^1- E









development of the charge in the transition state would best be stabil-

ized by the B substituent.


H
H+

Fp ---- +Fp




The next four resonances occurring at 7.8, 7.5, and 7.3 ppm were

identified by irradiation studies as H5, H6, and H3 and 4 overlapping,

respectively. These resonances occur upfield of the corresponding

signals centered at 8.5 ppm in homotropyliuml4 itself videe infra) and

reflect both decreased positive charge density and reduction in the

aromatic deshielding in this carbene complex, as would be expected were

backbonding by the metal into the positively charged ring significant,

as will be discussed below.



5.6 1.3 5.1 0.7


58 Fp 6.5 H 8.4
5.8 8.9

7.5 +-I 7.3 8.4 8.6

7.8 7.3 8.6 8.3


The next resonance at 5.8 ppm is coupled only to H6 and the remaining

two split resonances, and so must correspond to H7 since no other proton

can be strongly coupled with more than two protons. The large singlet









resonance at 5.3 ppm is due to the cyclopentadienyl group associated

with the iron. The last two resonances at 5.6 and 1.3 ppm result from

H8 exo and H8 endo, respectively, and are only half the intensity of the

other single proton signals. This was caused by use of FSO3D as the

reagent when forming this sample, and results in the equal population of

both exo and endo sites with deuterium and hydrogen. This equal isotope

population could have been caused by exchange due to bridge flip isomeri-

zation or nonspecific protonation videe infra) and allows spin-spin

coupling of the homoconjugate bridge protons only with H7 (no geminal

coupling) resulting in the observed doublets. Triplets of correct area



D H H D

Fp Fp

H H




due to geminal and adjacent coupling are observed when protic acids are

used.

The assignment of the resonances at 5.6 and 1.3 ppm to H8 exo and

H8 endo respectively rests with the observed cis and trans 3 bond

coupling constants compared with other homotropyliums.21'22'23














H 8.2

9.7 7.3
qJ-Fp

H


H 7.2

9.8
-H


10.8
H' H
9.2 7.7
-OH

H


Furthermore, the

H8 endo can only

periphery of the
exo.


10.5
H H
9.8 7.5
-OMe

TH


extreme difference in chemical shift between H8 exo and

be explained by the positioning of H8 endo within the

aromatic ring current and outside the periphery for H8




H8 endo H8 exo



r-)B


In these positions the bridge protons are strongly influenced by the local

anisotropy of the aromatic ring current caused by the presence of magnet-
ic field B. This anisotropy powerfully shields H8 endo and deshields H8

exo, producing the observed chemical shifts 1.3 and 5.6 ppm respectively.









While the magnitude of the ring current that produces a change in

chemical shift, AS, between the exo and endo protons of 4.3 ppm must be

considerable, a valuable comparison can be made with non-metal substitut-

ed homotropyliums. It has been reported that the A6 of homotropylium

itself is 5.8 ppm, and is due to a ring current of the same order of

magnitude as that in benzene.14

The A6 of FpHCOT is substantially smaller than the Ad of homo-
tropylium and must represent a decrease in the extent of delocalization

or a significant distortion in the geometry. While the steric bulk of a

cyclopentadienyliron is considerably greater than many organic groups,

it will be shown videe infra) that steric changes in the organometal sys-

tem are not the dominant influence in the reduction of A6 and that this

change in A6 corresponds better with the backbonding ability of the cy-

clopentadienyliron substituent.

The decreased delocalization observed in FpHCOT+ can be thought of

as the result of backbonding by the Fp substituent into the homoaromatic

T system and consequent localization of the a electrons and positive

charge. Additional spectral evidence that significant iron carbene

carbon double bonding exists is found in the 13C nmr spectrum of this







Fp- = Fp--



compound (Figure 2). Large low field chemical shifts of the carbene car-

bon have been found in Fp carbene complexes.6 The resonance of the










































II- S-0




= .0 .
u ar- 0







000





7 -J
C 0 C- c




S 0 -0
5 .--












41 Q) r_
u cu cu.

SI O w o








C -4 -
u >j ) U,-0






a (.
S- "
^ I- i"
v =1*r i
S. r 1
r 4- ) *CT
m n0
S ^ u c^ -- 1
: ^ ^ Q- (0 10 t/
2 ? i c c
a Ocn 1
S- t1 *- (3









carbene carbon at 269 ppm for FpHCOT is 147 ppm further downfield than

the comparable resonance at 122 ppm in homotropylium,24 and represents

a greater downfield shift than is observed in Fp+=cycloheptatrienylium,6

consistent with the expectation that a homoaromatic delocalization would





Fp 161

Fp+ 269 24

A = 147 ppm A = 81 ppm



be more readily disrupted by backbonding than a fully aromatic delocali-

zation.

The above evidence clearly supports the description of this material

as a homoaromatic carbene complex, a hitherto unknown compound having im-

portant ramifications in the descriptive chemistry of aromatic carbene

complexes. Two additional carbene complexes of cyclooctatrienylium will

be described,each having a ligand of different electronic and steric

characteristics, in an attempt to better characterize the overall struc-

tural and electronic properties of these homoaromatic complexes.

The three carbene complexes described in this work all have a com-

mon starting material in (n5-cyclopentadienylirondicarbonyl)cyclooctate-

traene, FpCOT. This sigma complex and its phosphinylated derivatives

have some interesting properties involving the conformational intercon-

version processes of cyclooctatetraene which will be discussed here.









A synthetic route from cyclooctatetraene (COT) producing reasonable

yields of FpCOT was described earlier. This substituted COT was found

to have an unusual property apparent from the 'H nmr in Figure 3. The

COT resonances are spread out over a range of nearly one ppm, a remark-

able change from the normal broadened singlet observed for all other

monosubstituted COT's. This is apparently an anisotropic effect due to

local variations in the applied magnetic field near the iron atom com-

bined with conjugative effects, because acyl complex 11, in which the

iron and COT are "insulated" from one another by the carbonyl, does not



0 0

KFp + C Fp



11



show the expansion of the COT resonances, but simply exhibits the normal

broadened COT singlet (Figure 4).

In effect, this sigma bonded complex, FpCOT, contains an internal

shift reagent. This expansion of the COT resonances dramatically eases

structural identification and the observation of the well-known COT in-

terconversion kinetics. Unlike previous studies of COT where examina-

tion of 13C satellites at low temperatures25 and full line shape26

analysis was required to determine proton coupling constants and chemical

shifts, this information can simply be measured from the spectra of the

sigma complex (Figure 5).









19



















CL











cu
cu

CC\




I-

0


C
.-


4-


U
E

















:3
0
s-














U-
01


,. )L












































4 1

a 0

4-1
SI-

U
0 *


C- C

U 2 Q.
,n D.







= C
S.-- S.-
.4,

S---*S



0.
0

s--


a]V





S.-
01 ~
a l





























0_


0

Ln


S- E
E 0-



Ln n
N











0 me
-
ro
*I -

C)






0 9-0
CD LL


Ln 0
0 E -0
S 0 (U
C 4





LO



CD
C Lj

































































t1 t~,









5* ii
4Sr Cr


C
0

J (.)





CO

U


Xo












Uc C=
4 -' ,
i- V5



0


0

CE .





5--1 (- 5
Uo

















05













Fp 3.4 Hz

HH \ 10.9 Hz
11.1 H
2.8 Hz 3.2 Hz




The close match between the reported coupling constants of 11.4 and
25
3.87 for COT,25 and the respective couplings in FpCOT implies that

little steric distortion occurs even though Fp is by far the largest re-

ported COT substituent.

FpCOT, like COT, undergoes isodynamic ring interconversion processes

(Figure 7). These processes were studied by diastereomer exchangel5 and

line shape analysis26 for COT and several monosubstituted derivatives and

were found to have free energy of activation barriers ranging from 12.5

to 14.8 kcal/mol for the Ring Inversion (RI) processes, and 14.9 to 17.4

kcal/mol barriers for Bond Shift (BS) interconversions.27 The measurement

of this Bond Shift activation barrier, in all reported cases in the lit-

erature, is performed by observation of the kinetic exchange of signals

due to R2 and R8. It is necessary to use complicated line shape analy-

sis26 to identify the R2 and R8 resonances when monosubstituted COT's

(R1 f H) are being investigated, although multisubstituted COT's can be

more easily examined by observing exchange of the substituent resonances.





This skew coupling constant has a reported range of ca. 2.5 to 3.87
Hz.28




























2S







0 ,
U)



C4-


0-
Sr-
CO




>0




1 m
H-CO
/ 4-

Cdi
'^ C.o
) = "" 0 e.
\C

















-R R







However, R2 = H2 and R8 = H8 resonances of monosubstituted FpCOT

are easily identified (H2 at 5.68 ppm and H8 at 6.16 ppm in Figure 5)

and the sample may simply be warmed to coalesce these resonances (Figures

8 and 9) to identify-AGBS = 17.9 kcal/mol.

This same procedure was used to measure the Bond Shift activation

barriers for two other substituted sigma complexes. Both of these new

compounds, (n5-cyclopentadienylironcarbonyltri-n-butylphosphine) cyclo-

octatetraene (FppBuCOT) 12, and (n5-cyclopentadienylironcarbonyltri-

phenylphosphine) cyclooctatetraene (FppCCOT) 13, were made in high yield

via photolysis of FpCOT in presence of tri-n-butylphosphine or triphenyl-

phosphine respectively.






The simple coalescence formula28 has been applied to determine rate
constants at the coalescence temperature. Although this equation is
strictly applicable to the exchange of single lines of equal population,
negligible errors will result when applied to coupled lines if J is much
less than 6v as applies here (J/Sv ca. 0.2). A precision of 0.5 kcal/mol
results when the possible range of temperatures at coalescence is ca.
50C.







































C-
U
4> C



C
u -
I-
0'4-
A) C
LL >

E4- 40

rO

5- H- E
S4- a
o *



ci
a Q r
lO0




=e C *-






5-


LU-










27




























Fo
Ln a
o 0

co
4-> -P-
4-




o .0
00




o- o
0n
01






S. -
LI- X






Q- CI-









C-


LL














CO e COT


R = Bu




Fe


PBu3


12


Fe
hP 3 / \COT
3 CO C PR3
v-------P


Fe



3

13


It will be observed that the major spectral difference between FpCOT and

these compounds (Figure 10 and Figure 11) other than the phosphine

ligand signals, is the double multiplicity of the cyclopentadienyl (Cp)

and COT resonances. This is due to the diastereomerism inherent in these

molecules. It is well known that monosubstituted COT's are chiral and it

should be evident that these organoiron moieties are "pseudotetrahedral"

and are thus chiral when substituted with a phosphine ligand. These

chiral metals are optically stable in the temperature range of interest29

but this is not true of the COT ligand. Both the Ring Inversion and





























Lo


j










Lm






L



L
i'








i-


I-
0



4--
CL

E

5-r







"C:
N C,
0
0







L-








30



















Ls



Ln
L -C





o .a

L D

SaOI)


L E
L'
L 3 Cs (





SN'-'
L, : r i















I I


Fe
C R


Bond Shift processes exchange configurations,30 and the rapid exchange
of diastereomers results. This exchange can be observed as the


S Fe
PCO
PR3


CO Fe
PR
PR


COFe /


C PR
PR3


averaging of the nmr Cp resonances due to the diastereomer pairs in each
of these phosphinylated sigma complexes. This behavior is shown in
Figure 12.




































coalescence


I i


320C


Figure 12. 1H nmr of exchange between
diastereomeric cyclopenta-
dienyls in FppBuCOT. The
fine splitting seen in the
low temperature spectrum is
due to the 1.1 Hz 31p-1H
coupling.


490C

460C



45oC'

44C

43C

420C


380C









Because AGBSt is usually substantially greater than AGRI and thus

the rate constant kBS very much smaller than kRI, observation of race-

mization of COT or exchange of diastereomers is generally attributed to

RI processes exclusively.26 (For an exception see reference 30.) This

was assumed when calculating28 the AGRI for FppBuCOT and Fpp0COT as

17.0 and 17.3 kcal/mol respectively.

As discussed previously, the observation of exchange in the COT re-

sonances is a phenomenon due exclusively to Bond Shift processes and is

entirely separate from the above consideration. In fact the activation

energies observed for the coalescence of the COT resonances in these


Table 1. Approximate Activation Barriers


Molecule

FpCOT

FppBuCOT

FppOCOT


AG RI


17.0

17.3


AGt BS

17.9

17.8

18.5


AGt in kcal/mol

*Potentially available from the averaging
in the 13C nmr (Figure 6).
AGt RI 0.2; AGt BS 0.5 videe supra)


of diastereotopic CO resonances


phosphinylated sigma complexes are the only additional values necessary

for a complete solution of equations30 describing the activation barriers

AGRIt and AGBS Unfortunately, these results were obtained with a pre-

cision that could adversely affect the overall results (due to an instru-

mental difficulty) and so this experiment must be redone more carefully.









Nonetheless, it is apparent that AAG the difference in activation

barriers for the Bond Shift and Ring Inversion processes (AGBS AGRI)

is smaller than, by possibly as much as one half, the AAGt for other

reported monosubstituted COT's.27


Table 2. Ring Inversion and Bond Shift Activation
Substituted Cyclooctatetraenes


Substituent

C(OH)Me2

OCHMe2

OEt


FppBu

Fppo


40.8

41.2


AGBst

17.4

15.6

16.2

%17.9

'17.8

M18.5


Barriers for Various


AGRI

14.7

12.7

12.5


'17.0

,17.3


AGt not corrected for small variations due to differing temperatures.





If these results are borne out by further experiment, the unusually small

AAGt seen for iron substituted COT's must be explained.

One possibility is an increase in electron density in the planar

transition state caused by backbonding from iron disturbing the antiaro-

matic delocalization characteristic of the high energy Bond Shift transi-

tion state, and reducing the total energy of the Bond Shift pathway.



















PR3
PR3


PR3


This is feasible since the occurrence of metal i backbonding into aryl

systems should be more disfavored than metal T-antiaromatic system in-

teractions, yet is known to exist.31'32

Alternately, steric interactions are known to increase in going to

the planar transition state and can cause reduction in the AAG.30

While the steric demands of the iron substituent are not high, steric

interactions do occur and can be observed using "fast" spectroscopic

techniques like IR.33 Still, it is unlikely that steric interactions in

this system can be of the same order of magnitude as in 1,2,3-trimethyl-

COT, which is the least crowded COT reported to depress AAGt values to

below 2 kcal/mol.30


3Fe

PR3


Me



Me









Clearly additional work needs to be performed on these systems be-

fore final conclusions may be drawn, but it should be noted that an ex-

planation of the small observed AAGt should be accessible. This is be-

cause the FppBu substituent is smaller than the Fpp0 substituent34 and

also is more electron donating relative to Fpp,; if steric interactions

are more significant in reducing AAG' the larger group, Fpp,, should

have decreased AAGt, while if electronic effects dominate, the more

electron donating group, FppBu, should have decreased AAGt. The latter

effect seems to be indicated from the present data.

This result would be interesting because it would be a rare, if not

exclusive, example of the effect of backbonding by an organometallic sub-

stituent into an antiaromatic ring system. A significant comparison

could then be made between the above effect and its counterpart in the

aromatic carbene complexes.

As discussed earlier, protonation of the FpCOT sigma complex pro-

duced a new homoaromatic carbene complex. Similarly, protonation of

FppCOT leads to n -cyclooctatrien-1-ylidene-n -cyclopentadienyliron-

carbonyltriphenylphosphine (FppgHCOT ) (14) which can be isolated as its

hexafluorophosphate salt. Although this compound is considerably more








Fe. 2) HFP

CO /a CHFP
Da D fHFP









thermally stable than the parent homoaromatic carbene complex, FpHCOT+,

it too is thermally labile and will decompose upon standing at RT for

several hours. Therefore, once again the structural proof relies on the

nmr spectra (Figures 13 and 14). As can be seen in the proton nmr, the

most downfield signal at 8.5 ppm represents one hydrogen adjacent to the

carbene center and identifies the structure as being similar to the

parent FpHCOT The large multiplets at 7.5 and 7.4 ppm are due to the






H
Fpp



triphenylphosphine (P03) ligand, and the next four resonances at 7.1,

6.8, 6.6, and 6.5 ppm are H5, H6, H4, and H3 respectively as identified

by decoupling. The quartet resonance at 5.4 ppm next to the solvent sig-

nal is H7. The Cp resonance occurs at 4.9 ppm. The last two resonances

represent the two geminal protons of the homoconjugate CH2 linkage,

identified as H8 exo at 4.8 ppm and H8 endo at 1.7 ppm by their coupling




7.2 8.2 8.9
H H H H H H
9.8 7.2 10.0 7.3 8.9 7.7
-H -Fp -- Fpp

H H H









38
























C-
Q.
Q-






+ 0)




o-















IC
O


L



L ai
EC


L *
0 ro-



































*0


+ Co
- CJ

4-. O
UX ..

CC

4- QC
O*

S.--
4-, 1-




jo
o-










ST
U-

LL
LL.


D


B









constants, separated in chemical shift once again by the anisotropy of a

substantial ring current.

However, it is expected that this ring current would be reduced re-

lative to the parent carbene, FpHCOT because backbonding by Fppo into

the homoaromatic ring should be more favorable than backbonding by Fp.10

This is because substitution of a phosphine ligand in place of a carbonyl

results in greater electron density on iron. This will allow increased

Fppg backbonding and increased localization of electrons in the cyclo-

octatrienylidene I system. The reduction in the ring current explains







Fe + Fe
CO/co
P03 P03



the decrease in the difference in chemical shift of H8 exo and H8 endo,

A6, to a value of 3.11 ppm, down from 4.30 ppm observed in FpHCOT+ and

5.86 ppm in homotropylium itself.

That the diminuation of A6 is an electronic effect rather than a

simple steric influence is supported by the extreme low field shift of

the carbene carbon resonance in the 13C nmr (Figure 14). The signal

(doubled by phosphorous coupling) occurs at 311.4 ppm, nearly 200 ppm

downfield from the C-1 resonance in homotropylium, and is further down-

field than that observed for the same carbon in FpHCOT This indicates

greater backbonding contributions,3 consistent with the observed














Fpp F 311;: Fp H
3





Fpp Fp=4 H-






changes in the proton nmr A6. Furthermore, the difference in shift of

the homotropylium ligand C-1 carbon between Fpp HCOT+ and FpHCOT+ is 42

ppm, similar to the analogous difference of 37 ppm6,13 observed in the

cyclophetatrienylidene series and reflects similar changes in backbond-

ing due exclusively to ligand substitution.

Another property of this carbene complex gives additional evidence

that greater backbonding occurs than that seen in FpHCOT It was ob-

served that exchange broadening and shifting of the signal due to H8 endo

occurred in the proton nmr for temperatures greater than 00C (Figure 15).

This was originally construed to indicate exchange with excess acid used

to produce the nmr solutions of the carbene. However, disappearance of

the signal was not observed when using deuterated acid; rather, the in-

tegrated signal intensity was seen to remain at one half proton. More-

over, exchange of this signal occurred when the hexafluorophosphate salt

Fpp0HCOT HFP was dissolved in solvent containing no acid. Therefore no

acid exchange process was present.








42














uuu

L
o a 1






K02





L um '
I LO
C40 0 10
v5


a*)- CL
O r- r-












0
L O LO
L, S O


V 0 =
L VI -



I CO



CO N 01 MW E
CD 1- 0)

C (D o
No Stu E


S ? r- L. L


L
L
5 IL









It is known that homotropylium undergoes an exchange process in

which the homoconjugate CH2 unit flips from a position above the approxi-

mately planar235 seven membered ring to a position below it. The



D HH



H D



activation barrier measured by Winstein for this process was 22.3

kcal/mol.36 Substitution of C-1 on homotropylium with a methoxy group

reduces the activation barrier to 19.6 kcal/mol.23 Brookhart interpreted

this reduction as being due to the greater substituent stabilization of

the positive charge on the sp2 rehybridized C-1 of the transition state
3
15 than the substituent stabilization of the charge on the more sp


H H H

H H
R &9R
15 16


hybridized C-I of the ground state 16.23

With this in mind, a re-examination of the kinetic behavior of

Fpp HCOT+ indicated that the signal due to H8 exo may have been obscured

by the Cp resonance in the 100 MHz nmr. Protonation of FppOCOT in d6
benzene shifted the Cp resonance (probably due to Cp-benzene stacking

interactions) to reveal the now relatively unobscured H8 exo resonance










which showed temperature dependent exchange broadening. This same phe-

nomenon was seen in the 300 MHz nmr (Figure 15). The exclusive exchange

of H8 endo and exo signal is taken to indicate that the bridge flip pro-

cess occurs much more rapidly in Fppg substituted homotropylium than in

normal homotropyliums since exchange broadening in 1-hydroxyhomotropylium

is not seen even to 800C.22 Accurate activation barriers have not been

obtained since thermal decomposition occurs rapidly above 60C although

a barrier of ca. 14 kcal/mol was estimated from line shape changes.

However, it was seen that this exchange process in FppgHCOT occurred

more rapidly than in FpHCOT because no exchange broadening was observed

when FpHCOT was rapidly heated to decomposition at 50C. (Unfortunate-

ly, unlike other homotropyliums,233637 no exo-endo specificity is seen

when deuteration of FpCOT is carried out even at-1000C and therefore the

kinetics of signal appearance could not be observed.) This last obser-

vation of faster exo-endo exchange in FppgHCOT than in FpHCOT is con-

sistent with Brookhart's interpretation that the activation barrier for

bridge flip interconversion is reduced when more powerful electron

dontaing substituents are present at C-i, the more powerfully donating

substituent in this case being Fpp .

A caveat is appropriate at this point in that the straightforward

interpretation of the H8 exo- H8 endo broadening presented here is likely

too simplistic to account for more recent spectral observations of this

phenomenon. High field nmr spectra which resolve the H8 exo signal show

that it exchanges and shifts toward low field rather than towards the H8

endo signal at higher field. Speculation as to the nature of this pro-

cess or of the possible superimposition of an additional process upon the

normal bridge flip exchange is currently in progress.









A somewhat more electron releasing substituent would be n5-cyclo-

pentadienylironcarbonyltri-n-butylphosphine, FppBu. However, because

the cone angle of occupied space in ligand PBu3 is only 130 degrees as

opposed to 145 degrees for P03,34 this substituent should be sterically

less demanding than Fppo. If steric factors play a significant role in

the structural and electronic properties of these organoiron substituted

homotropyliums, a reversal in the spectroscopic trends as seen for

FpHCOT and FppgHCOT should be apparent in the spectrum of n -cyclo-

octatrien-l-yliden-n5-cyclopentadienylironcarbonyltri-n-butylphosphine,

FpPBuHCOT+ (17) (Figure 16).

When FppBuCOT is protonated to produce this final new homoaromatic

carbene complex, the general trends are preserved. The difference in

shift of the CH2 protons, A6, is 2.98 ppm, a value slightly less than

the A6 of 3.11 ppm observed for FppHCOT in accord with the slightly

greater electron releasing ability of the FppBu substituent. The chemi-

cal shift of the carbene carbon is 315.7 ppm, again indicative of

slightly better backbonding in this complex relative to Fpp0HCOT+ (the

C-1 of which appears at 311.4 ppm). In fact, all the spectral details

of these two phosphinylated carbene complexes are rather similar, except

for one difference. The chemical shift of the two homoconjugate protons

are both shifted about three quarters of a ppm (0.68 ppm for H8 exo and

0.81 ppm for H8 endo) downfield from their counterparts in Fpp0HCOT.

While no assumptions have been made concerning the linearity of the

absolute magnitude of the chemical shifts due to the H8 protons as a

function of ring current (an assumption which has been specifically con-

traindicted 36), it is believed that a direct and linear relationship

holds for the comparison of AS and the magnitude of the ring current when










46





















C-






oo


00




4--




C,
Nj




S-
CL
0-

SQm




-C






Do r






w 1

C-








































o+

I-




0



-L
u u




ci
CL
r:
C-,



rr
s-


01
C


Al
(I1a







48

relatively constant geometries are assumed.14'16 While the expected and

observed results of this work are consistent with the latter statement,

it is somewhat bothersome that a slight deviation in the expected pat-

tern of convergence of the H8 shifts exists for FppBuHCOT This




6

'" H H8 exo
4
H8 endo
3 :
6 "-- FppBu
2
2''-F-fpaF-P
1 I I -EP


1 2 3 4 5 6
A5



deviation can be seen above in the convergence of H exo-H endo shifts

with increasing electron donation capability of the 1-substituent in

the homotropylium ligand.

Whatever the source of this deviation, it is apparent that it

results in a slight variation in the geometry of the molecule. This can

be seen by the examination of the complete set of bridge coupling con-

stants for these three systems and the comparison of these data with

other substituted homotropyliums vide infra.

One of the most often cited indications of homotropylium structure

is the two bond proton proton coupling constant for H8 exo H8
21 22,23,38
endo.21222338 When originally formulated Rosenberg, Mahler and
39
Pettit were undecided about the electronic structure of protonated

cyclooctatetraene (HCOT ). Subsequent work led them to strongly support




















the homoaromatic structure.38 However, Deno preferred a description of

the molecule as the bicyclic species.40 Winstein pointed out that the

bicyclic form was unlikely in that the expected proton coupling constants

for a cyclopropylcarbinyl system were not consistent with the observed

couplings in HCOT+41 vide infra.




A B8






JA,B= ca. 4.7 Hz

JA,C= ca. 4.5 Hz

JBC = ca. 8 Hz


As can be seen below, the geminal and three bond cis and trans

coupling constants of the organoiron substituted homotropyliums are

quite similar to those observed in HCOT+ and other substituted homotro-

pyliums.














H 8.2 H

9.7 73


H8~ H
-7J,3


10.8
H H
9.2 7.7
-OH


H 7.2H

9.8 7.2

-H


10.5
H H
9.8 7.5
--OMe

H


H, -H


Clearly these carbene complexes are not distorted toward a closed

cyclopropane type structure. Nor do the coupling constants reveal any

significant departure from a relatively constant geometry, even though

FppBuHCOT couplings reflect a small variation in the trend. What is
exemplified in the above is the marked similarity between the homotro-

pylium carbenes and the homotropyliums substituted with electron





It should be pointed out that the figures for Fpp HCOT were far more
difficult to determine because of overlap of H7 and H8 signals in the
proton nmr, and probably have a larger error than the other carbene data.









donating groups. These similarities give strong support that the

changes observed above are primarily geometrical changes that result

from electronic interactions. That small geometrical changes do occur,

resulting from electronic interactions, can be seen through examination

of the 13C-11 coupling constants at C8. In cyclopropylcarbinyl this

13C-1H coupling is 180 Hz or more.42 In homotropylium, due to the modi-

fied a type interactions between C-1 and C-7, the larger C1-CH2-C7 bond

angle is expected to cause reduction in the s character in the CH bonds,

leading to decreased coupling interactions between the 13C and 1H nuclei.

The observed hydrogen-bridge carbon coupling in homotropylium is 159




H H H H H H




0 R
J = 180 Hz 1J 8,H 159 Hz C8,H= 130- 136 Hz



Hz,21 and the expected reduction in 1,7 bonding in FpHCOT and
+ *
FppgHCOT produces the even smaller couplings, ca. 131 Hz and 136 Hz

respectively.





* +
The somewhat low value for FpHCOT is probably an inaccuracy caused by
the method used to obtain it. The actual value measured was the 13C-
deuterium coupling constant, which was then multiplied by the ratio of H
and 2H magnetogyric ratios to calculate the 1H-13C coupling.









This reductioninl,7bonding due to backbonding by the iron substitu-

ent should also be accompanied by increased localization of the T system.

This effect can be seen in the coupling constants of the protons about

the periphery of the nearly planar seven membered ring portion of these

systems (Figure 18).

These values must be compared with the only other complete set of

coupling constants available,21 those from homotropylium itself. The

differences between these two sets of values arise from the unsymmetrical

backbonding substitution of the homotropylium framework and manifest

themselves as an alternation in the magnitude of the coupling constants

due to apparent i localization and consequent alternation of the double

and single bond character in the ring system. Prior to the recent X-ray

structural verification35 and the impressive theoretical calculations by

Haddon,43 the coupling constant information for homotropylium was the

best evidence that homotropylium was comprised of a distinctly puckered

but symmetrical seven membered ring portion (7MR).21 A modified, but

similar geometry is though to exist videe infra) in the homoaromatic

carbene complexes.

The nonsymmetrical substitution of homotropylium, along with sequen-

tially increasing backbonding by the cyclopentadienyliron series, Fp,

Fpp0 and FppBu, should lead to sequential increases in the localization

of the i electron density and thus the alternation of the coupling con-

stants. The full set of proton coupling constants for these molecules

(Figure 18) are consistent with this and are quite accurate, as can be

seen in the spectral simultions based on the evaluated couplings in

Figures 19, 20 and 21. Interestingly, the asymmetry of the line shapes

in FpHCOT+ (Figure 19) is related to the symmetry of a portion of its















7.2

8.6 .4
-H
10.1 ( hiZ)i
10.1 7.4
8.6


8.2

9.7

9.1 N -
11.4 9.0
6.9


8.9

11.5, /
8.9J Fpp0

11.6 8.5
6.4


7,8 endo = 10.0


J =8.9
7,8 endo 9

7,8 exo 7.7


8.3

11.7
8.9 PPBu

11.5 8.5
5.7


7,8 endo 8.1

7,8 exo 77


Figure 18. Proton-proton coupling constants for homo-
tropylium and the three homoaromatic carbene
complexes. All values reported in Hertz.


7,8 endo =9.8
J exo = 7.2
7,8 exo


7,8 exo

2,5


= 7.3

= 0.9






















cz,
0.






a.





r-








o
0











L -
CO 0 -a





S E




co
L i-













- LL










55



































Lo
Q -
0





.

4 -
4- C0






C 41



4-

08


m
Sn








































L'











:, -' C
E




C +D
U 4-


E >
ou-



t -o
CO

sI
13
S-

U=_









7MR, which causes an unusually large long range coupling of 0.9 Hz be-

tween H2 and H5. Significant through space interactions such as this

usually occur only when the bonds of the interacting nuclei are nearly

colinear. The long range coupling (together with the extensive Tr




H2
H Fp


H5

delocalization) helps to verify the approximate planarity of the 7MR in

this Fp carbene complex.

The more strongly backbonding substituents, Fpp, and FppBu, could

cause an increased distortion of the 7MR from planarity as increases in

the alternating bond character allow torsional changes to occur. That

this is an electronic effect is shown by the correlation of backbonding

strengths (but not steric considerations) and coupling constant alterna-

tion. The difference between adjacent couplings is presented in Table 3.





3 2

4 6
5 6









Table 3. 1H Three Bond Coupling Constants for Monosubstituted Homotro-
pyliums, and Differences in Adjacent Couplings.


2, 3 J3, 4 4, 5 J5, 6 6,7
A A A A


HCOT+ 8.6 10.1 10.1 8.6 7.4
1.5 0 1.5 1.2


FpHCOT+ 9.7 9.1 11.4 6.9 9.0
0.6 2.3 4.5 2.1


FppOHCOT+ 11.5 8.9 11.6 6.4 8.5
2.6 2.7 5.3 2.1


FppBuHCOT+ 11.7 8.9 11.5 5.7 8.5
2.8 2.6 5.8 2.8




What is to be noted here is that the increase in dr pi backbonding (which

restricts the delocalization of the 7MR ir system) as R becomes H, Fp,

Fppo and FppBu causes the difference in the adjacent coupling constants

to generally increase--as would be expected were the difference in the

adjacent bond characters to increase.

This pattern authenticates the proposal that the geometry of homo-

tropylium carbene ligands is more responsive to the electronic behavior

of the cyclopentadienyliron than to the possible steric interactions.

Because of this, the observation that the variation in the difference in

chemical shifts (A6) of the bridge protons corresponds primarily to the

iT electronic interaction capabilities of the organometallic substituent

should stand on a firmer basis.









With this in mind, it is pertinent to note once again that both the

A6 and bridge coupling constants fall in the range of respective values

arising from a variety of 1- substituted HCOT+'s. Thus it seems possi-

ble that a direct comparison with normal organic functional groups may

reasonably be made. There does not appear to be any significant indica-

tion in the previous data or discussion that would argue against such a

comparison.



Table 4. Comparison Between Various R Substituted Homotropyliums.


Ha Hb

-R

H c


Ha,Hb
Ha H

H ,Hc
Hbc'


J
gem

trans

cis


R groups H21 Me38


Fp Fppo FppBu


OH22 OMe22


gem 7.2 8.0 8.2 8.9 8.3 10.8 10.5

Jtrans 9.8 10.0 10.0 8.9 8.1 9.2 9.8

Jcis 7.2 7.5 7.3 7.7 7.7 7.7 7.5


A6 5.8 5.0 4.3 3.1 3.0 3.1 3.1




The thrust of much of this work's discussion has been the valida-

tion of the relationship between the magnitude of the homoaromaticity

as measured by A, and the r donation capability of the cyclopentadienyliron










groups. The analogous measure of the T donating ability of organic func-
+ 44
tional groups is the a .

When the a+ of the substituents of the reported 1-substituted

homotropyliums is plotted against the A6 of that compound a relatively

straight line is produced (Figure 22). Locating the A6 value of the

three carbene complexes FpHCOT FppHCOT and FppBuHCOT on this line

produces a values of -0.43, -0.80, and -0.84 respectively. These num-

bers should at best be considered approximations, and it would be wise

to think of a phosphinylated cyclopentadienylironcarbonyl group (Fpp) as

having about the same electron donating ability as a hydroxy or methoxy

group, and of a cyclopentadienylirondicarbonyl group (Fp) as a substan-

tially weaker donor. Nonetheless, this is the first sensitive indica-

tion of organoiron T donating strength directly comparable to normal

organic groups. Other comparative methods have not been sensitive
45,46
enough to separate carbonyl and phosphine substitution effects,4546





-.2 M ..--Me
4 .. _ -. 6 OMe+ +
-8 FpPu 0 PP
-1
-1.2
1 .4 ,

-1 .8 .- 0
-2 I
0 1 2 3 4 5 6
A6
Figure 22. Least squares plot of A5 vs. o for
R substituted Homotropyliums. a
of 0- estimated from value of a
for NMe2 44,47,48,49









much less the backbonding effects of alkyl versus aryl phosphine sub-

stitution as observed in this work.

This comparison fairly sharply delineates the a donor capability

of Fp and Fpp in the homotropylium system. Certainly some variation in

this parameter would be expected with other carbene ligands having dif-

ferent electronic demands. Nonetheless, an approximate indication of

the relative backbonding or i donating ability of these organoiron

groups has a clear advantage over the activation energy of rotational

barriers as a predictive descriptor in the chemistry of these cationic

carbene complexes. This is especially true in the 6w electron aromatic

series, the members of which have notable steric contributions to their

rotational barriers.13

For the sake of comparison, an attempt was made to measure the bar-

riers to rotation in the three homoaromatic complexes as well, but this

data could not be collected. In the case of FpHCOT any barrier to ro-

tation was expected to be low through analogy with the proposed small

barrier in Fp+=C7H6 (3).13 No changes in the 1H nmr of FpHCOT were

seen between temperatures of -1000C to -100C, above which decomposition

was seen to occur rapidly. (Highly acidic conditions were seen to alle-

viate this thermal sensitivity somewhat, so that the rapid heating to

500C was accomplished in 90% trifluoroacetic acid without spectral changes

other than final decomposition.) Combined with the observation of a

single set of diastereotopic carbonyl resonances in the 13C nmr (Figure

2), the spectral data point to rapid (nmr time scale) rotation about the

iron-carbene bond.

















diastereotopic CO/ "-chiral
-"-CO



On the other hand, the large phosphine group in the FppHCOT 's

could be expected to interact even more strongly with the rotating car-

bene ligand than was seen in FppC7H6+ (7) in which a 9.6 kcal/mol elec-

tronic and steric composite barrier to rotation was observed,3 because

the greater backbonding videe supra) in the homoaromatic complexes would

not only lead to closer proximity of the ligands, and thus increase

steric interactions, but would also increase the electronic contribution

to the rotational barrier. However, no spectral changes were seen from

-1000C to 500C other than the previously mentioned exchange broadening

of H8 exo and H8 endo. In view of diastereomerism inherent in these








O Fe Fe
PR3 PR3


diastereomers









FppHCOT complexes, it is possible that a considerable difference in the

energy of the two diastereomers could prevent the observation of an ex-

change between signals of greatly different intensity in the proton nmr.

This diastereomerism may be the cause of difficulty in obtaining

crystalline salts of the Fpp complexes. In both FppBuHCOT HFP (17) and

FppgHCOT HFP (14) only amorphous solids could be collected after careful

attempts at recrystallization. Since the non-diastereomeric FpHCOT HFP

(10) crystalline solid was ultimately too thermally sensitive for X-ray

analysis, the symmetrical bis(1,2-diphenylphosphino)ethane (DIPHOS) sub-

stituted carbene complex was sought. It was hoped that this compound,

n -cyclooctatrien-1-yliden-n -cyclopentadienyliron[bis(1,2-diphenyl-

phosphino)ethane] hexafluorophosphate (CpFe(DIPHOS)HCOT HFP) (18), would

be more thermally stable than the other HCOT+ carbenes, would crystallize





Fe


P02

18


for X-ray analysis, possess adequate symmetry for rotational barrier de-

tection and provide a final point for comparison of + and 66. However

the starting material n5-cyclopentadienyliron[bis(diphenylphosphino)-

ethane] cyclooctatetraene (19), could not be synthesized, as discussed

below.













CpFe(DIPHOS)



19



Addition of DIPHOS to FpCOT by photolysis occurred with rapid evo-

lution of carbon monoxide, but the mixture of products did not contain

substituted cyclooctatetraene, nor did addition of strong acid to a so-

lution of this material produce any carbene complex signals in the nmr.

Addition of cyclooctatetraenyllithium (LiCOT)20 to cyclopentadienyl-

iron(DIPHOS)chloride was not attempted since the addition of cyclohepta-

trienyllithium to the chloride did not yield a desired product.50

Reaction of LiCOT with [CpFe(DIPHOS)]2N2(HFP)251 produced a very

small yield of a mixture possibly containing the product 19, but this

was not followed to conclusion because of the extremely low yields.

This line of research was abandoned in order that other pertinent

questions might be addressed. For example, the thermal decomposition of

FpHCOT HFP (10) was shown to produce a COT complex of Fp This probably

precedes via a 1,2-8 hydrogen shift7 from the homoconjugate CH2 to pro-

duce the olefin complex (n5-cyclopentadienylirondicarbonyl)-n2-cyclo-

octatetraene (Fp -COT) (20). A possible52 mechanism for this process is
indicated below.














H H


FpO





H


HIj
+Fp

Jreductive
elimination


The structure of this product was established by the release of COT
(as identified by H nmr) from a solution of the thermally decomposed
carbene complex via nucleophillic displacement of the COT from Fp with
sodium iodide.53 Also observed in the products was Fp iodide.


Fp


Nal


+ FpI


This reaction was verified by grinding the thermally decomposed
FpHCOT HFP (10) with KC1 prior to introduction in the mass spectrometer.
A reaction analogous to that seen in solution apparently occurs in the
+
solid state, since COTP was identified (high resolution in accord with
m/e calculated for COT) as well as trace signals attributable to the
modestly volatile FpCl and its fragmentation products.









This decomposition is exactly what is expected of a nonheteroatom

stabilized organometallic carbene complex having B hydrogens present.7

On the other hand, the exhibited stability of these homoaromatic carbene

complexes is more analogous to that of heteroatom stabilized carbene





OMe OMe /OMe

(CO) Cr--C t (CO)Cr =C\ r- (CO)5Cr-C
CH 21 CH3 H




complexes such as 21. Neither the homoaromatic nor heteroatom stabilized

carbene complexes undergo facile 1,2-hydrogen transfer reactions, proba-

bly because the carbene centers are part of the conjugated i systems.

However, pentacarbonyl[(methoxy)(methyl)carbene]chromium (0) 21, is some-

what acidic,54 and thereby indicates the availability of its Tr hydrogens

for 1,2-shift decompositions. This points out another possible stabil-

izing factor in these homotropylium carbenes: the nearly orthogonal or-

ientation of the homoconjugate CH2 bonds to the metal carbene conjugat-

ed T system. Without appropriate orbital orientation, 1,2-hydrogen

shifts cannot occur.

This would imply that the pathway to decomposition is the bridge

flip intermediate, where neither stabilizing effect mentioned above is

as effective.

H









A problem here is the observation that the FppHCOT 's are more

thermally stable than FpHCOT yet with Fpp substitution, this bridge

flip intermediate should be more accessible videe supra), therefore

leading to more rapid decomposition. It is possible, however, that

greater backbonding by Fpp would reduce the charge density on C-1, thus

inhibiting 1,2-hydrogen shifts relative to Fp substitution.

Nonetheless, it is apparent that the ground state homoaromatic

carbene complex is a modestly stable organometallic species having B hy-

drogens present. This is sufficiently rare enough to warrant note,55

but it is clear that the geometrical alteration upon protonation to form

these complexes is quite unique, so that this method of forming carbene

complexes should not be regarded as having general utility.

In conclusion, the synthesis and characterization of the first

three homoaromatic carbene complexes are described in this work and a

strong correlation has been made with the r backbonding ability of cyclo-

pentadienylirondicarbonyl (Fp) and cyclopentadienylironcarbonyltrialkyl

(or aryl)phosphine (Fpp), and the + of normal organic functional groups.

Additionally, three novel organoiron sigma bonded complexes of

cyclooctatetraene (COT) were developed for the above work which dramati-

cally display the known isodynamic conformational processes of COT, and

which may reveal a hitherto undescribed organometallic interaction with

a 4n n electron orbital system.















CHAPTER III
EXPERIMENTAL



General



The melting points were obtained on a Thomas Hoover melting point

apparatus and are uncorrected. Nuclear magnetic resonance spectra were

obtained on either a Jeol FX-100 or Nicolet NT-300 spectrometer, and are

reported in ppm relative to TMS or the deuterated solvent resonance.

Coupling constants for the nmr spectra were either measured directly if

the line was a simple doublet or simulated using the Nicolet 1180 ITRCAL

program. Infrared data were obtained on a Perkin-Elmer 137 spectropho-

tometer. Combusion analyses were performed by Atlantic Microlab Inc.,

Atlanta, Georgia. Mass spectra were obtained on an AEI MS 30 spectro-

meter.

All nmr solvents were degassed by bubbling nitrogen through the

cold solvent for ten minutes and subsequently stored in a Vacuum Atmo-

sphere Inc. recirculating glove box under oxygen-free conditions. All

other solvents were distilled. All solvent and reagent solutions were

swept free of 02 by bubbling N2 through the solution for at least ten

minutes when dealing with air sensitive compounds. THF was distilled

from benzophenone ketyl using Na/K alloy. Methylene chloride was dis-

tilled from P205.









The alumina used was Fisher certified neutral alumina, Brockman

Activity I, to which 3% w/w water was added. After the addition of

water, the alumina was allowed to stand at least 12 hours before being

purged several times with nitrogen immediately prior to use. All reac-

tions were performed under a nitrogen atmosphere and Schlenk type appa-

ratus was used when appropriate.


Cyclopentadienylirondicarbonyl Iodide (Fpl)

This material was used as supplied from Alfa Products.


Tri-n-butylphosphine (PBu3)

This material was used as supplied by Aldrich Chemicals after de-

gassing with N2.


Triphenylphosphine (P03)

This material was recrystallized from 95% ethanol. It was obtained

from Aldrich Chemicals.


Bis(l,2-diphenylphosphino)ethane (DIPHOS)

This material was recrystallized from acetone. It was obtained

from Alfa Products.


Trifluoroacetic Acid (TFA)

This material was distilled and degassed prior to use. d -TFA was

made by hydrolysis of acetic anhydride with D20.









Triphenylcarbenium Hexafluorophosphate (03C HFP)

This material was used as purchased from Aldrich Chemicals.


Fluorosulfonic Acid (FSO3H)

This material was distilled and degassed with nitrogen bubbling

before use. FSO3D was used as purchased from Aldrich.


Potassium Cyclopentadienylirondicarbonyl

This material was made according to the method of Gladysz et a.17


Bromocyclooctatetraene (BrCOT)

To a dry 1000 mL 3 necked round bottomed flask equipped withmechan-

ical stirring, N2 bubbler and a 300 mL addition funnel, and containing

47.2 g (0.45 mol) of freshly distilled cyclooctatetraene (COT) dissolved

in 500 mL dry CH2C12 at -780C was added dropwise 25.0 mL (77.5 g, 0.48

mol)Br2in25 mLdry CH2C12.

After stirring for 15 minutes, 75 g (0.67 mol) potassium t-butoxide

dissolved in 250 mL dry THF was added portionwise to the dibromide

solution over the course of 6 hours, maintaining the temperature between

-60 and -780C. The mixture was stirred overnight at low temperature and

then, after warming to room temperature, poured over 500 mL crushed ice,

swirled gently and placed in a separatory funnel. The organic layer was

separated, and the aqueous layer and emulsion mixed with 50 mL diatoma-

ceous earth, suction filtered and extracted with 3x50 mL CH2Cl2. The

combined organic layers were washed with water, dried over MgSO4 and

flash evaporated at room temperature.









The crude BrCOT was mixed with an equal portion of pentane, filter-

ed through alumina and flash evaporated to obtain a dark yellow liquid

which was carefully distilled between 40 and 450C at 0.5 torr to collect

70 g BrCOT as a yellow oil in 85% yield. The IR and nmr spectral pro-

perties are in accord with those reported in the literature. 20


Cyclooctatetraene Carboxylic Acid (COT-CO2H)

To a stirred solution of 6.4 g (34.6 mmol) BrCOT in 50 mL dry

ethyl ether and 4.0 g clean Mg turnings was added ca. 2 mL of a 1:1 solu-

tion of ethane dibromide in ether. The solution was warmed to initiate

the reaction and then placed in an ice bath with the slow addition of

the remainder of the 1:1 dibromide-ether solution (total ethane dibro-

mide 6.5 g, 34.6 mmol) such that ethylene continued to be evolved with-

out interruption. The solution was strongly stirred for about 3 hours

and then rapidly siphoned over crushed, dry, solid CO2. After warming

to RT, the mixture was extracted with 3x100 mL H20. The aqueous layer

was then acidified to pH 3 to form a cloudy yellow suspension of COT-

CO2H. This was extracted with 3x50 mL pentane, dried over MgSO4 and

rotovapped to an oil which crystallizes to form yellow plates after

standing for several days to produce 3.25 g COT-CO2H in 64% yield.

COT-CO2H: mp 66-69C; 1H nmr (CDCl3)6 5.9 (s, COT), 10.3 (br s, 1H).


Cyclooctatetraene Acid Chloride (COT-COC1)

Cyclooctatetraene carboxylic acid (3.07 g, 20.7 mmol) was placed in

a 50 mL round bottomed flask equipped with magnetic stirring, reflux

condenser and heating mantle. To this was added 6.8 g (4.4 mL, XS)

oxalyl chloride in 10 mL benzene portionwise with much bubbling. The









solution was allowed to stir until CO evolution ceased and then was

gently refluxed for 2 hours. The excess benzene and oxalyl chloride

was removed by flash evaporation, and another 20x2 mL benzene added

and removed to separate the product from residual oxalyl chloride. The

brown oil was then distilled at 520-56C at 0.05 torr to yield 3.25 g

COT-COC1 in 93% yield. COT-COC1: IR (neat) 3010, 2400, 1750, 1625,

1250 cm1.



(n5-Cyclopentadienylirondicarbonyl)cyclooctatetraene Carbonyl (11)

To a stirred solution of 1.3 g (6.2 mmol) KFp in 30 mL dry THF at

-78C was added dropwise 1.1 g (6.6 mmol) cyclooctatetraene acid chlo-

ride in 15 mL dry THF. After completion of the addition, the stirred

solution was allowed to slowly warm to room temperature over the course

of 2 hours. The reaction mixture was then coated onto 10 mL of silica

gel by flash evaporation and chromatographed on a 6 by 1 inch silica

gel column using benzene as eluent. After the brown Fp dimer band was

eluted, the yellow band of acyl complex was carried off the column by

adding up to 50% CHC13 to the benzene eluent. The product, 0.62 g of

11, was collected as a yellow-orange crystalline air stable solid in

36% yield after evaporation. 11: mp 106-110C with decomposition; IR

(CHCI3) 2000, 1600 cm-1; 1H nmr (CDC13)6 4.9 (s, Cp), 5.9 (s, COT).


Photolysis of 11

No products other than highly insoluble solids were obtained after

photolysis of hexane or MeC12 solutions of 11 using either pyrex or

quartz filtered 550 W Hanovia Hg lamps.









(n5-Cyclopentadienylirondicarbonyl)cyclooctatetraene (FpCOT) (9)

To a rapidly stirred solution of 1.0 mL (1.42 g, 7.8 mmol) BrCOT
in 250 mL dry, 02 free THF was added 3.5 g (16 mmol) KFp in THF at room

temperature. The addition of the KFp solution was performed at high

dilution (to reduce Fp dimer formation) by a modified, nonsiphoning

Soxelet extraction of the modestly soluble KFp by refluxing the THF

under reduced pressure (90 120 torr) while maintaining the solution

temperature at room temperature with a heating mantle (BrCOT is thermal-

ly labile). After no further KFp was extracted, the darkened reaction

mixture was vented with nitrogen and stirred overnight. The crude mix-

ture was then coated onto 10 mL of Grade II Fisher neutral alumina by

flash evaporation and chromatographed on a 1 by 4 inch degassed, Grade

II Fisher neutral alumina column using degassed hexane as eluent. A

pale yellow BrCOT band preceded the intense yellow FpCOT band which was

collected under nitrogen to produce 0.94 g FpCOT in 43% yield. The

yield varied between 20 and 43% depending on the dilution of the Fp

anion addition. This yellow, air-sensitive oil should be stored re-

frigerated in solution (do not use CHC13). FpCOT: Anal. calcd. for

C15H13FeO: C, 64.31; H, 4.32. Found: C, 64.07; H, 4.40; IR (film)
2960, 1990, 1925, 1560, 1400, 1005, 909, 827, 802, 714 cm-; 1H nmr

(300 MHz, CD2C12)5 4.81 (s, Cp), 5.22 (dd, H7, J7,8 =11.1 Hz, J6,7=

3.4 Hz), 5.51 (dd, H5, J5,6= 10.9 Hz, J4,5=3.2 Hz), 5.68 (fine d, H2,

2,3 =2.8 Hz), 5.75 (dd, H6), 5.77 to 5.85 (m, H3 and H4), 6.16 (d, H8);
1C nmr (d6-acetone)6 87.2 (Cp), 119.3, 129.2, 132.6 (2C), 136.7, 139.3,

147.7, 151.7 (C1), 217.2 (CO), 217.7 (CO); mass spectrum m/e, 280.01984

(calcd. 280.01850).









An alternative synthesis produces FpCOT in lower yield. To a cold

(-780C) solution of LiCOT, made by the method of Cope and Burg20 from

2.42 g (13.2 mmol) BrCOT, was added 4.50 g (13.2 mmol) Fpl in 20 mL dry,

02 free THF dropwise with stirring. After stirring for 2 hours at

-780C, the solution was allowed to warm to room temperature and stirred

for an additional I hour. The dark mixture was adsorbed onto 10 mL

degassed Grade II neutral alumina and chromatographed as above using a

12 x1 inch column. The product, 0.61 g FpCOT, was collected in 16%

yield.


(n5-Cyclopentadienylironcarbonyltri- n-butylphosphine)cyclooctatetraene
(FppBuCOT) (12)
To a solution of 0.23 g (0.82 mmol) FpCOT in 50 mL dry, 02 free THF

in a tall, narrow Schlenk tube was added 0.30 mL (0.24 g, 1.2 mmol, 1.5

fold XS) tri-n-butylphosphine. Nitrogen was bubbled briefly through the

solution to sweep out residual 02 and the solution was then irradiated

through Pyrex under N2 at RT (H20 cooled jacket) 6 inches from a 550 W

Hanovia Hg lamp. The reaction progress was monitored by taking 0.2 mL

aliquots and evaporating to dryness, redissolving in CHCI3 and observing

the 2000 cm-1 region for disappearance of the two Fp absorptions and

appearance of the single FppBu absorption at 1925 cm-1. Photolysis ap-

peared to be complete after 3 hours but was continued 1 additional hour.

The product, 0.35 g (0.77 mmol) FppBuCOT as a yellow-orange air-sensi-

tive oil, was collected in 94% yield after chromatographing the reaction


This is essentially the same as the procedure reported by Cooke et al19









mixture as described above for FpCOT. This material should be stored
in the same manner as the parent FpCOT. FppBuCOT: IR (CHC13) 2995,
1925 cm1; 1H nmr (d6-acetone)6 0.9 to 1.9 (m, butyl), 4.4 and 4.5

(d's, Cp, diastereomers with JHP= 1.1 Hz), 4.9 (m, H7), 5.3 (m, H5),
5.6 (m, H2, H6, H3, H4), 6.3 (m, H8); 13C nmr (d6-acetone)6 24 to 32
(butyl), 83.6 (d, Cp, diastereomers), 115.7 (d, diastereomers), 127.4
(d, diastereomers), 132.0 (s), 133.3 (d, diastereomers), 137.0 (2C, d
below 00C), 149.9 (d, diastereomers), 167.1 (dd, C1,diastereomers),
221.9 (dd, CO).
Alternatively, FppBuCOT can be produced without isolating FpCOT.
Addition of 2 fold XS (based on an assumed 40% yield of FpCOT) of tri-
n-butylphosphine to the crude rxn mixture of FpCOT in THF followed by
overnight photolysis as above with stirring produced greater than 90%
yields of FppBuCOT after workup.


(n -Cyclopentadienylironcarbonyltriphenylphosphine)cyclooctatetraene
(FppOCOT) (13)
FppCOT was prepared in greater than 80% yield from FpCOT by using
a 1.5 fold XS of triphenylphosphine in the primary synthesis described
above, however photolysis occurred more slowly. To a solution of 0.85

g (3.0 mmol) FpCOT in 100 mL dry, 02 free THF was added 1.44 g (5.5 mmol)
triphenylphosphine. After overnight photolysis under N2 using the pre-
viously described photolysis set up, the reaction mixture was coated
onto 15 mL Grade II neutral alumina by flash evaporation and placed on a
1 by 4 inch column of degassed, Grade II neutral Fisher alumina in hex-
ane. After eluting the XS P03 and unreacted FpCOT with hexane, an
orange band was collected under N2 by eluting successively with 5% and








10% MeC12 (or ether) in hexane. The product, 1.3 g (2.5 mmol, 82%)

FppCOT as an air stable orange glass, remained after removal of the
solvent. This mixture of diastereomers could not be recrystalized even

after repeated attempts. Fpp COT: Anal. calcd. for C32H27FeOP: C,

74.72; H, 5.29. Found: C, 74.49; H, 5.35. IR (CHC13) 2995, 1920 cm-1
H nmr (300 MHz; CD2C12)6 4.45 and 4.49 (d's, Cp, diastereomers with

JHP=1.1 Hz), 4.91 (dd, H7, J7,8=11.11 Hz, J6,7 =3.22 Hz), 4.95 (dd,
H7', J7',8 =11.02 Hz, J6',7 = 3.22 Hz), 5.30 to 5.84 (complex, H2, H5,

H6, H4, H3), 6.10 (dd, H8), 6.29 (dd, H8'), 7.37(br s, O's); 13C nmr

(CD2C12)6 85.4 (d, Cp), 116.6 (d), 130 to 140 (0 and COT), 149.5 (s),

164.4 (dd, C1 of COT), 222.3 (dd, CO). Mass spectrum m/e, 486 (M-CO),

383, 262 (100%), 224,121, 56.


(n5-Cyclopentadienyliron[bis(1,2-diphenylphosphino)ethane])cyclooctate-
traene (CpFe(DIPHOS)COT) (19)

This compound could not be produced by the methods discussed below.
Photolysis of 100 mg FpCOT with 1.2 equivalents DIPHOS in d6-benzene was

performed using all quartz glassware and a 550 W Hanovia Hg lamp. After

5 minutes, CO evolution was observed but the nmr spectrum of the solu-

tion appeared unchanged. After 45 minutes of photolysis, several Cp re-
sonances had replaced the original Cp signal at 4.5 ppm. Chromatography

yielded only mixtures of unidentified products. Similar results were
obtained when the above photolysis was carried out in THF using a pyrex
filtered lamp. An alternative method also did not yield the desired

product. A suspension of the highly insoluble [CpFe(DIPHOS)]2N2(HFP)2
salt produced by the method of Sellman and KleinschmidtS1 (0.237 g,

0.175 mmol in 50 mL THF) was added to 0.5 mmol LiCOT20 in THF at -780C








and stirred overnight at low temperature. No color change was observed

until the solution was warmed to RT, after which the evaporate was

coated onto 5 mL degassed Grade II neutral alumina. Chromatography pro-

duced hydrocarbon products along with a small quantity of possible pro-

duct 19 visible as a small set of signals at 6.5 ppm (COT) and 4.5 ppm

(Cp) in the nmr. Although the low yield of products was probably due to

the insolubility of the nitrogen complex, no other appropriate solvents

were found.


n -Cyclooctatrien-I-ylium-n5-cyclopentadienylirondicarbonyl Hexafluoro-
phosphate (FpHCOT HFP) (10)

Excess anhydrous HC1 carried by N2 was bubbled through a cold

(-400C), magnetically stirred solution of 0.31 g (1.1 mmol) FpCOT in 25

mL degassed CH2C12 for ca. 15 minutes; the bright yellow solution imme-

diately turned deep red due to the presence of the homoaromatic carbene

complex. A two fold excess of degassed fluorosulfonic or trifluoroace-

tic acid in CH2CI2 could be used in place of the HC1. To the deep red

FpHCOT solution was added 0.43 g (1.1 mmol) triphenylcarbenium hexa-

fluorophosphate in 10 mL CH2C12 (anmonium hexafluorophosphate was not

soluble) and the FpHCOT PF6 salt precipitated by slow addition of an

equal volume of cold ethyl ether. The resulting microcrystalline orange

needles were collected by filtration under N2 on a glass frit surrounded

by a dry ice isopropanol cup. This thermally sensitive solid (stable

below -100C in air) was washed by the addition of 3 x2 mL of cold ethyl

ether and stored in the freezer (-300C). A yield of 0.19 g (40%) FpHCOT

was isolated from the FpCOT sigma complex. However, low temperature

protonation of FpCOT in CD2 C2 instantly generated nmr solutions of the








carbene complex quantitatively. These solutions must be rigorously 02
free. FpHCOT HFP: MP upon rapid heating, orange solid turns dark
brown at about 700C, melts at about 800C, and then turns black with ef-
fervescence at 1000-105C; IR (KBr) 2002, 1970, 1440, 1400, 1300, 1275,
1070, 980, 840 cm1; 1H nmr (300 MHz, CD2C12, -200C)6 1.27 (t, H8 endo,

8 endo, 8 exo 8.2 Hz, 7,8 endo =10.0 Hz), 5.26 (s, Cp), 5.57 (t, H8
exo J7,8 exo = 7.3 Hz), 5.78 (t, H7, J6,7 =9.0 Hz), 7.34 and 7.35 (m,
H4 and H3, J3,4=9.1 Hz), 7.47 (t, H6, J5,6= 6.9 Hz), 7.83 (m, H5,
J4,5= 11.4 Hz), 8.86 (d, H2, J2,3= 9.7 Hz, J2,5=0.9 Hz); 13C nmr
(CD2C12, -200C) 62.3 (C8), 89.8 (Cp), 133.2, 133.7, 134.3 (C4, C7, C6),
139.7 (C3), 143.2 (C5), 169.0 (C2), 210.4, 211.5 (CO's), 269.0 (Cl).


n -Cyclooctatrienyl-l-ylium-n -cyclopentadienylironcarbonyltri -n-
butylphosphine Hexafluorophosphate (FppBuHCOT HFP) (17)
This compound was produced from FppBuCOT in the same manner as des-
cribed above for the Fp carbene complex. The initially yellow orange
solution immediately became intense burgundy colored upon addition of
acid. The dark purple hexafluorophosphate salt is substantially more
thermally stable than the parent, but is not crystalline and is subject
to air oxidation. FppBuHCOT HFP: IR (KBr) 2970, 1960, 1485, 1465,1425
1080, 835 cm-1; 1H nmr (300 MHz, CD2Cl2, -200C)6 0.9 to 2.0 (m, Butyl),
2.48 (t, H8 endo, J endo,8exo=8.3 Hz, J7,8 endo= 8.1 Hz), 4.97 (d,
Cp, JHP= 1.2 Hz), 5.48 (m, H8 exo and H7, J7,8 exo = 7.7 Hz), 6.45 (t,
H3, J3,4 = 8.9 Hz), 6.62 (t, H4, J4,5 = 11.5), 6.73 (t, H6, J6,7 = 8.7
Hz), 7.04 (dd, H5, J5,6 = 5.7 Hz), 8.48 (d, H2, J2,3 = 11.7 Hz); 13C
nmr (CD2Cl2, -20C) 13.6 to 29.9 (n-Butyl), 62.7 (C8), 90.3 (Cp), 123.2,
130.1, 134.3, 135.4, 137.4 (C3 to C4), 161.5 (C2), 217.1 (d, CO, JCP =
28.1 Hz), 315.7 (d, C1, JCP = 22.0 Hz).








n1-Cyclooctatrien-1-ylium-n5-cyclopentadienylironcarbonyltriphenyl-
phosphine Hexafluorophosphate (Fpp0HCOT HFP) (14)

This deep purple amorphous salt was produced from Fpp0COT in the
same manner as described earlier for the Fp carbene complex except that
a two fold XS of TFA was used in place of anhydrous HC1. The thermal
stability of this compound is the highest of these three homoaromatic
carbene complexes. Brief exposures to as high as 500C in solution have
been tolerated with no decomposition. After vacuum drying the salt ap-
pears to be air stable indefinitely when kept cold (-10C); FppHCOT
HFP: MP upon rapid heating, dark purple solid turns ash colored at ca.
1000C, then darkens at 1500C without melting; IR (KBr) 2995, 1975, 1490,
1465, 1430, 1095, 845 cm1 ; H nmr (300 MHz, CD2C12, -200C)6 1.67 (t, H8
endo, J8 endo, 8 exo = 8.9 Hz, J7,8 endo = 8.9 Hz), 4.79 (t, H8 exo,

J7,8 exo = 7.7 Hz), 4.87 (d, Cp, JHP = 1.2 Hz), 5.42 (q, H7, J6,7 = 8.5
Hz), 6.51 (t, H3, J3,4 = 8.9 Hz), 6.66 (t, H4, J4,5 = 11.6 Hz), 6.78
(t, H6), 7.07 (dd, H5, J5,6 = 6.4 Hz), 7.37 to 7.77 (m, O's), 8.51 (d,
H2, J2,3 = 11.5 Hz); 13C nmr (CD2C12/TFA, -20C)6 61.7 (C8), 90.8 (Cp),
124 to 138 (Ring C's and O's), 161.5 (C2), 216.4 (d, CO, JCP = 29.3 Hz),
311.4 (d, Cl, JCP = 22.0 Hz).















REFERENCES


1. F.A. Cotton and C.M. Lukehart, Prog. Inorg. Chem. 16, 487 (1972).

2. D.J. Cardin, B. Centikays and M.F. Lappert, Chem. Rev. 72, 545
(1972).

3. E.O. Fisher and A. Maasb61, Angew. Chem. Int. Ed. Eng. 3, 580
(1964).

4. M. Brookhart and G.O. Nelson, J. Am. Chem. Soc. 99, 6099 (1977).

5. K. Ofele, J. Organomet. Chem. 22, C9 (1970).

6. N.T. Allison, Y. Kawada and W.M. Jones, J. Am. Chem. Soc. 100, 5224
(1978).

7. A. Cutler, R.W. Fish, W.P. Giering and M. Rosenblum, J. Amer. Chem.
Soc. 94, 4354 (1972).

8. C.G. Kreiter and E.O. Fisher, Angew. Chem. Int. Ed. Eng. 8, 761
(1969).

9. M. Brookhart and J.R. Tucker, J. Organomet. Chem. 193, C23 (1980).

10. M. Brookhart, J.R. Tucker, T.C. Flood and J. Jensen, J. Am. Chem.
Soc. 102, 1203 (1980).

11. B.E.R. Schilling, R. Hoffman, D.L. Lichtenberger, J. Am. Chem. Soc.
101, 585 (1979).

12. P.E. Riley, R.E. Davis, N.T. Allison and W.M. Jones, J. Am. Chem.
Soc. 102, 2458 (1980).

13. F.J. Manganiello, M.D. Radcliffe and W.M. Jones, J. Organomet. Chem.
228, 237 (1982).

14. C.E. Keller and R. Pettit, J. Am. Chem. Soc. 88, 606 (1966).

15. W.P. Giering, A. Sanders, L. Cohen, D. Kennedy and C.V. Magatti,
J. Am. Chem. Soc. 95, 5430 (1973).

16. J. Gasteiger, G.E. Gream, R. Huisgen, W.E. Konz and U. Schnegg,
Chem. Ber. 104, 2412 (1971).









17. J.A. Gladysz, G.M. Williams, W. Tam and D.L. Johnson, J. Organomet.
Chem. 140, C1 (1977).

18. R.B. King and M.B. Bisnette, J. Organomet. Chem. 2, 38 (1964).

19. This compound was subsequently found to have been made in a manner
similar to our original synthesis by M. Cooke, C.R. Russ and F.G.A.
Stone, J. Chem. Soc., Dalton, 256 (1975).

20. A.C. Cope and M. Burg, J. Am. Chem. Soc. 74, 168 (1952).

21. P. Warner, D.L. Harris, C.H. Bradley and S. Winstein, Tetrahedron
Lett. 46, 4013 (1970).

22. M. Brookhart, M. Ogliaruso and S. Winstein, J. Am. Chem. Soc. 89,
1965 (1967).

23. M.S. Brookhart and M.A. Atwater, Tetrahedron Lett. 43, 4399 (1972).

24. G.A. Olah, J.S. Staral and G. Liang, J. Am. Chem. Soc. 96, 6233
(1974).

25. F.A.L. Anet, A.J.R. Bourn and U.S. Lin, J. Am. Chem. Soc. 86, 3576
(1964).

26. J.F.M. Oth, Pure Appl. Chem. 25, 573 (1977).

27. G.I. Fray and R.G. Saxton, The Chemistry of Cyclooctatetraene and
Its Derivatives, Cambridge University Press, New York, 1978.

28. R.J. Abraham and P. Loftus, Proton and Carbon-13 NMR Spectroscopy,
Heyden and Son Ltd., London, 1978.

29. T.C. Flood, P.J. DiSanti and D.L. Miles, Inorg. Chem. 15, 1910
(1976).

30. L.A. Paquette, Pure Appl. Chem. 54, 987 (1982).

31. J.D. Duncan, J.C. Green, M.L.H. Green and K.A. McLauchlan, Chem.
Commun., 721 (1968).

32. M.D. Rausch, A.K. Ignatowicz, M.R. Churchill and T.A. O'Brien, J.
Am. Chem. Soc. 90, 3242 (1968).

33. K. Stanley and M.C. Baird, J. Am. Chem. Soc. 97, 4292 (1975).

34. C.A. Tolman, Chem. Rev. 77, 313 (1977).

35. R.F. Childs, A. Varadarajan, C.J.L. Lock, R. Fagginani, C.A. Fyfe
and R.E. Wasylishen, J. Am. Chem. Soc. 104, 2452 (1982).

36. S. Winstein, C.G. Kreiter and J.I. Brauman, J. Am. Chem. Soc. 88,
2047 (1966).









37. H.D. Kaesz, S. Winstein and C.G. Kreiter, J. Am. Chem. Soc. 88,
1319 (1966).

38. C.E. Keller and R. Pettit, J. Am. Chem. Soc. 88, 604 (1966).

39. J.L. von Rosenberg, J.E. Mahler and R. Pettit, J. Am. Chem. Soc.
84, 2842 (1962).

40. N.C. Deno, Prog. Phys. Org. Chem. 2, 157 (1964).

41. S. Winstein, H.D. Kaesz, C.G. Kreiter and E.C. Friedrich, J. Am.
Chem. Soc. 87, 3267 (1965).

42. G.A. Olah, D.P. Kelly, C.L. Jeuell and R.D. Porter, J. Am. Chem.
Soc. 92, 2544 (1970).

43. R.C. Haddon, J. Org. Chem. 44, 3608 (1979).

44. H.C. Brown and Y. Okamoto, J. Am. Chem. Soc. 80, 4979 (1958).

45. R.P. Stewart and P.M. Treichel, J. Am. Chem. Soc. 92, 2710 (1970).

46. E.S. Bolton, G.R. Knox and C.G. Robertson, Chem. Commun., 664 (1969).

47. R.W. Taft, J. Phys. Chem. 64, 1805 (1960).

48. J. Hine, J. Am. Chem. Soc. 82, 4877 (1960).

49. R.W. Taft, E. Price, I.R. Fox, I.C. Lewis, K.K. Anderson and G.T.
Davis, J. Am. Chem. Soc. 85, 3146 (1963).

50. F. Manganiello, M.S. Thesis, University of Florida (1980).

51. D. Sellman and E. Kleinschmidt, Angew. Chem. Int. Ed. Eng. 14, 571
(1975).

52. P.J. Davidson, M.F. Lappert and R. Pearce, Chem. Rev. 76, 219 (1976).

53. K.M. Nicholas and A.M. Rosan, J. Organomet. Chem. 84, 351 (1975).

54. C.P. Casey and R.L. Anderson, J. Am. Chem. Soc. 96, 1230 (1974).

55. R.R. Schrock and G.W. Parshall, Chem. Rev. 76, 243 (1976).















BIOGRAPHICAL SKETCH


Marc Dudley Radcliffe was born in Schnectady, New York, in February

1953. After spending his early childhood on the coasts of New England,

he and his parents moved to Titusville, Florida. There he watched the

dawn of the space age and became interested in science. He attended

college at Miami University in Oxford, Ohio, Worcester Polytechnic In-

stitute in Worcester, Massachusetts, and finally at New College in

Sarasota, Florida, before finding employment in the semiconductor in-

dustry, where he participated in the development of the early digital

watch technology.

He returned to college at the University of Florida in 1975 and sub-

sequently received the Summer Undergraduate Fellowship in Chemistry. He

began his graduate work there in 1977 with Dr. William M. Jones, received

a supplemental graduate fellowship during his studies there, and has re-

cently accepted a postdoctoral position with Dr. Kurt Mislow at Princeton

University.









I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.






William M. Jones chairman n
Professor of Chemikry




I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.






Merle A. Battiste
Professor of Chemistry




I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.





William R. Dolbier, Jr. /
Professor of Chemistry









I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.





Eric V. Dose
Assistant Professor of Chemistry





I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.





Raymond J. Be geron
Associate Professor of Medicinal
Chemistry





This dissertation was submitted to the Graduate Faculty of the
Department of Chemistry in the College of Liberal Arts and Sciences and
to the Graduate Council, and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.

August 1982


Dean for Graduate Studies andResearch




Full Text

PAGE 1

SYNTHESIS AND CHARACTERIZATION OF THE FIRST HOMOAROMATIC ORGANOMETALLIC CARBENE COMPLEXES BY MARC DUDLEY RADCLIFFE A THESIS PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1982

PAGE 2

ACKNOWLEDGEMENTS No acknowledgement can adequately express my gratitude to Dr. William M. Jones, whose professional leadership and personal encouragement have greatly increased the satisfaction and rewards of my graduate studies. No past experience has offered greater intellectual freedom, nor more equitable guidance than found here under his counsel. Moreover, he has attracted several outstanding individuals to this research group. William R. Winchester, Frank Manganiello and James Lisko have all contributed significantly to the character of this academic experience and I am grateful for their stimulating conversations and friendships. n

PAGE 3

TABLE OF CONTENTS ACKNOWLEDGEMENTS ii ABSTRACT v CHAPTER I INTRODUCTION 1 II RESULTS AND DISCUSSION 9 III EXPERIMENTAL 68 General 68 Cyclopentadienylirondicarbonyl Iodide 69 Tri-n-butylphosphine 69 Triphenylphosphine 69 Bis(l,2-diphenylphosphino)ethane 69 Trifluoroacetic Acid 69 Tri phenyl carbenium Hexafluorophosphate 70 Fluorosulfonic Acid 70 Potassium Cyclopentadienylirondicarbonyl 70 Bromocyclooctatetraene 70 Cyclooctatetraene Carboxylic Acid 71 Cyclooctatetraene Acid Chloride 71 c (n -Cyclopentadienylirondicarbonyl )cyclooctatetraene Carbonyl (11) 72 Photolysis of 11 72 c (n -Cyclopentadienyl irondi carbonyl )cyclooctatetraene (9_) 73 m

PAGE 4

(n -Cyclopentadienyli roncarbonyl tri -n-butylphosphine)cyclooctatetraene (_12) 74 (n -Cyclopentadienyli roncarbonyl tri phenyl phosphine)cyclooctatetraene (_13) 75 c (n -Cycl opentadienyli ron[ bis (1,2-di phenyl phosphino)ethane])cyclooctatetraene (19) 76 1 5 n -Cyclooctatrien-l-ylium-n -cyclopentadienylirondicarbonyl Hexafluorophosphate (H)) 77 1 5 n -Cyclooctatrien-l-ylium-n -cyclopentadienyli roncarbonyl tri -n-butyl phosphi ne Hexaf 1 uorophosphate (17) 78 1 5 n -Cyclooctatrien-l-ylium-n -cyclopentadienyli roncarbonyl tri phenyl phosphi ne Hexaf 1 uorophosphate (14) 79 REFERENCES 80 BIOGRAPHICAL SKETCH 83 TV

PAGE 5

Abstract of Thesis Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SYNTHESIS AND CHARACTERIZATION OF THE FIRST HOMOAROMATIC ORGANOMETALLIC CARBENE COMPLEXES By MARC DUDLEY RADCLIFFE August 1982 Chairman: William M. Jones Major Department: Chemistry Protonation of (n -cyclopentadienylirondicarbonyl) cyclooctatetraene with strong acid produced the novel homoaromatic carbene complex n. cyclooctatrien-1-ylium-n -cyclopentadienylirondicarbonyl (FpHC0T + ). Examination of the H nmr spectrum of this compound revealed a decrease in the homoaromaticity of the homotropylium ligand when compared with homotropylium itself, consistent with the expectation that significant d-ir pit backbonding interactions into the positively charged ligand by the organoiron occurs. Substitution of triarylor trialkyl-phosphine for a carbonyl on the cyclopentadienyliron in the carbene complex sequentially reduces the degree of aromatic derealization in the homotropylium ligand, providing clear evidence that significant increases in the backbonding character of these complexes occurs when the electron density on iron increases, as expected from our earlier work.

PAGE 6

Furthermore, the three sets of proton coupling constants in the homotropyl i urn ligands reveal increased bond alternation as the other organic ligand is changed from CO to P(Ph) 3 to P(Bu)_. This bond alternation was used to substantiate the relative measure of the aromatic ring current the difference in proton chemical shift between the homoconjugate exo and endo protons (AS). When the AS of the three homoaromatic carbene complexes was compared with that of several other substituted homotropyl i urns, an approximate measure of the tt backbonding ability of the cyclopentadienyliron group relative to normal organic functional groups was obtained. The thermal decomposition of FpHCOT appears to proceed through a 1,2-8 hydrogen shift mechanism. However, the complex has been isolated as the hexafluorophosphate salt and could be manipulated if kept cold. That these carbene complexes having B hydrogens are i sol able at all was postulated as being due to both the stabilizing effect of the homoaromaticity, and the nearly orthogonal orientation of the 6 hydrogen bonds and the carbene tt system. Additionally, the isodynamic interconversion processes of three cyclopentadienyliron substituted cyclooctatetraenes were studied by nmr. The initial data appear to indicate that an unusual interaction between the transition metal and the 4n i\ electron ring system may be present. VT

PAGE 7

CHAPTER I INTRODUCTION The general class of organometallic compounds known as carbene complexes are thought of as the formal result of allowing an organic carbene species to approach a ligand substituted transition metal in such a way that the filled sp orbital and unfilled p orbital of the carbene interact with appropriate empty and filled orbitals of the transition metal to form a compound of the two species. Although no carbene complex has been produced by this procedure, alternative methods have been developed which have allowed the isolation and characterization of several of these 1 2 species. ' One of the most widely used synthetic procedures was developed by 3 Fisher and Maasbb'l. Heteroatom stabilized carbene complexes were produced by organolithium attack on a carbonyl ligand followed by alkylation of the anionic intermediate. RLi /R (Me) 3 0BF 4 M(C0) n _^i_> (CO) ^=5=^ —— n-1 v •0 + Li (CO). ,M=C' R n _ 1 u * » (co) n _ lM : — c + x 0Me """-OMe

PAGE 8

The resulting neutral compound can be characterized as a neutral metal carbonyl accepting electron density from the filled sp 2 orbital of the "carbene" to form a sigma bond. As a result of this bonding, the now electron rich metal can release electrons in a d orbital with symmetry appropriate for "backbonding" into the unfilled p orbital on the carbene carbon. Alternatively, the heteroatom substituent can compete with the metal for bonding into this empty orbital. This or some other form of additional tt stabilization can substantially increase the stability of carbene complexes. For example, the methylene carbene complex (Jj is known to have a transitory existence, but cannot be observed spectroscopically even at temperatures as low as -80°C. However, phenyl Z^> + H CO / ^C — H CO R C — H substitution at the carbene carbon atom stabilizes the complex to such an extent that the benzylidene complex (2^) is isolable. Both the above methylene and benzylidene complexes were produced by protonation of the sigma bonded metal a-methyl ether to remove the methoxy group as methanol. The remaining positively charged carbon can be stabilized as indicated. Removal of a good leaving group has become a versatile method of producing cationic carbene complexes.

PAGE 9

A variety of additional carbene complexes have been generated through other methods, including two general classes of aromatically stabilized carbene complexes. :C0) 5 Cr -2 CI CI -* (C0) 5 Cr Fe CO //V^\ CO u The cationic cycloheptatrienylidene complex [3) was also formed by removal of a negative leaving group, in this case a hydride, through use of the hydride abstracting agent tri phenyl carbenium ion. A useful method of forming a cationic carbon site in carbonium ion chemistry which has not apparently found application in metal carbene chemistry has been protonation of an alkene. Such a procedure would generally produce an intermediate which would rapidly decompose by a 1,2 hydrogen shift to reform the alkene as a it complex with the metal.

PAGE 10

"^ . Fp ur V Fp — C+ H FP + -|Fp= Cyclopentadienyl i rondicarbonyl All of these carbene complexes can be represented as resonance hybrids of three canonical forms A C each of which makes different contributions to the overall structure in the (several) individual types of carbene complexes. A qualitative indication of the contribution by these resonance forms can be obtained from the measurement of barriers to rotation about the appropriate double bonds. For example, a barrier of 12.4 kcal/mol has been observed for the carbene carbon-oxygen bond rotation in pentacarbonyl[(methoxy) Q (methyl )carbene]chromium (0) (page 1, M=Cr) , as well as a 9.1 and 10.4 kcal/mol barrier in Cp(C0) 2 Fe=CHCgH 5 + (4) and Cp(C0) 2 Fe=CH(p-CH 3 C 6 H 4 ) + (5), respectively, for the C aryl -C carbene bond. 9

PAGE 11

&G =9.1 kcal/mol Fp AG = 10.4 kcal/mol Me G For CpFe(PPh 2 CH 2 CH 2 PPh 2 )=CH„ (6) the barrier to rotation about the metal-carbene bond of 10.4 kcal/mol was compared favorably with the calculated electronic barrier. While these figures may not be strictly comparable, they do reflect the fact that contributions by resonance forms B^ and C_ can be significant. When X and Y in these resonance forms are the same and represent a conjugated (4n +2)tt electron ring system, the aromatically stabilized

PAGE 12

6 carbene complexes (e.g. 3>) result. Because of the aromatic delocalization of the positive charge in the carbene ligand, contributions from resonance structures having double bond character between the metal and the aromatic ligand, representing metal dir pir backbonding, should be substantially reduced. o — X-ray structures of CpFe(CO) 2 =C 7 Hg + (3) and CpFe(C0) 2 =benzoC 7 H 4 + , the aromatic tropylidene complexes produced by Allison and Jones, clearly indicate a high degree of aromaticity and concurrent reduced backbonding in these complexes. But when the iron-carbene bond length was compared with two other model systems, only "some degree of metal -ligand multiple 12 bonding" was suggested. In an effort to better identify the contribution of metal -carbene backbonding interactions in these aromatic complexes, rotational barriers were obtained for the nonsymmetrical ly substituted iron carbene complexes Cp(C0)PBu 3 Fe=C ? H 6 + (7) and Cp(C0)PBu 3 Fe=benzoC 7 H 4 + (8). 13

PAGE 13

Fe CO X> PB» 3 The rotational barriers observed, 9.6 and 10.4 kcal/mol respectively, were unexpectedly comparable to the much more strongly it bonding species, CpFe(DIPHOS)=CHp (6J , which had been found to have a rotational barrier of 10.4 kcal/mol. Because the X-ray data for compounds 3^ and Fp=benzoC 7 H. indicated weak backbonding, it seemed likely that "the principle restriction to rotation [in 7_ and 8] was steric in origin with 13 a minor electronic component superimposed." At the time that this last work was being done, characterization of a new carbene complex was being completed, the spectral properties of which were seen to have significant bearing on the question of the relative contribution of metal it backbonding in aromatic carbene complexes. Rather than stabilizing the positive charge in the carbene ligand by aromatic derealization, this new complex was thought to achieve a large measure of its stability through homoaromatic derealization. While this is a difference merely of kind rather than of function, a wealth of information was believed to be available from this system when compared with the cycloheptatrienylidene complex. In particular, since the difference in chemical shift of the inner and outer proton of homoconjugate CFL linkages is thought to be a definitive measure of the extent

PAGE 14

of aromatic delocalization, and since this derealization will be reduced by competitive delocalization of the positive charge through resonance contributions due to metal backbonding, it was recognized that a purely electronic measure of metal-carbene it bonding interaction could be obtained from homoaromatic carbene complexes by the examination of the difference in chemical shift of the homocon jugate protons mentioned above. A further point of interest is the fact that no other homoaromatic * carbene complexes have been reported in the literature. This work will therefore describe the synthesis and characterization of the first homoaromatic carbene complexes. Furthermore, a strong correlation will be made between the i\ bonding ability of cyclopentadienyliron groups and the a of normal organic functional groups. V, * + 15 The Fpbenzocyclobutenylidene complex reported by Giering and Sanders is an unlikely candidate for homoaromatic delocalization because of the required loss of aromaticity and increases in strain energy.

PAGE 15

CHAPTER II RESULTS AND DISCUSSION The synthesis of the first homoaromatic transition metal carbene complex was accomplished by the synthetic route shown below. Br. K0+ KFp 10 Bromocyclooctatetraene was prepared as previously reported. f 16 17 Treatment of the bromide with potassium cyclopentadienylirondicarbonyl produced the (n -cyclopentadienylirondicarbonyl) cyclooctatetraene (FpCOT) {9). This reaction is unusual in that the Fp anion is normally used to * Except the dehydrobomi nation reagent, potassium J>butoxide, was dissolved in THF rather than added as the solid.

PAGE 16

10 displace alkyl halides; however, it was noted that King had successfully 18 displaced fluoride from fluorobenzenes with Fp anion, and similar reaction with BrCOT seemed feasible. The cyclooctatetraene sigma complex was formed via this reaction in higher yield and with more convenient 20 work up than our earlier method using cyclooctatetraenyllithium and Fp iodide. The structural proof of the cyclooctatetraene sigma complex rests on its well resolved H nmr, spectrum and the observed conformational interconversion processes ( vide infra ). Finally, the homoaromatic 1 5 carbene complex, n -cyclooctatrien-1-yliden-n. -cyclopentadienylirondicarbonyl hexafluorophosphate (FpHCOT HFP) (10) , was produced by protonation of the FpCOT with strong acid and subsequent anion exchange using triphenylcarbenium hexafluorophosphate, to yield a wery thermally sensitive, but air stable orange solid in 40% isolated yield from the sigma complex. Because of the thermal instability of this carbene complex, the structural proof relies heavily on the H nmr which is seen in Figure 1. The single proton resonance of 8.9 ppm occurs at the characteristic low field of protons adjacent to the carbene center and attests to the unsymmetrical nature of the ring system in which the homocon jugate linkage is also immediately adjacent to the carbene center. Other possible structures would require two proton resonances at this characteristic low field, which is not observed. Furthermore, protonation of the FpCOT adjacent to the substitution is expected since the

PAGE 17

11 4 — C\J tJ I— CD 1— 03

PAGE 18

12 development of the charge in the transition state would best be stabilized by the B substituent. Fp J^ The next four resonances occurring at 7.8, 7.5, and 7.3 ppm were identified by irradiation studies as H5, H6, and H3 and 4 overlapping, respectively. These resonances occur upfield of the corresponding 14 signals centered at 8.5 ppm in homotropylium itself ( vide infra ) and reflect both decreased positive charge density and reduction in the aromatic deshielding in this carbene complex, as would be expected were backbonding by the metal into the positively charged ring significant, as will be discussed below. 5.6 1.3 5.1 "0.7 7.8 7.3 8.6 8.3 The next resonance at 5.8 ppm is coupled only to H6 and the remaining two split resonances, and so must correspond to H7 since no other proton can be strongly coupled with more than two protons. The large singlet

PAGE 19

13 resonance at 5.3 ppm is due to the cyclopentadienyl group associated with the iron. The last two resonances at 5.6 and 1.3 ppm result from H8 exo and H8 endo, respectively, and are only half the intensity of the other single proton signals. This was caused by use of FSOgD as the reagent when forming this sample, and results in the equal population of both exo and endo sites with deuterium and hydrogen. This equal isotope population could have been caused by exchange due to bridge flip i somen* zation or nonspecific protonation ( vide infra ) and allows spin-spin coupling of the homoconjugate bridge protons only with H7 (no geminal coupling) resulting in the observed doublets. Triplets of correct area D -Fp H due to geminal and adjacent coupling are observed when protic acids are used. The assignment of the resonances at 5.6 and 1.3 ppm to H8 exo and H8 endo respectively rests with the observed cis and trans 3 bond 21 22 23 coupling constants compared with other homotropyliums. '

PAGE 20

14 9.7 H 8-2 H 7.3 T> 10.8 9.2 1 7.7 _^0H 10.5 9.8 7.5 — OMe Furthermore, the extreme difference in chemical shift between H8 exo and H8 endo can only be explained by the positioning of H8 endo within the periphery of the aromatic ring current and outside the periphery for H8 exo. H8 endo. •H8 exo GOIn these positions the bridge protons are strongly influenced by the local anisotropy of the aromatic ring current caused by the presence of magnetic field B. This anisotropy powerfully shields H8 endo and deshields H8 exo, producing the observed chemical shifts 1.3 and 5.6 ppm respectively.

PAGE 21

15 While the magnitude of the ring current that produces a change in chemical shift, A6, between the exo and endo protons of 4.3 ppm must be considerable, a valuable comparison can be made with non-metal substituted homotropyliums. It has been reported that the A6 of homo t ropy 1 i urn itself is 5.8 ppm, and is due to a ring current of the same order of 14 magnitude as that in benzene. The A6 of FpHCOT is substantially smaller than the AS of homotropylium and must represent a decrease in the extent of derealization or a significant distortion in the geometry. While the steric bulk of a cyclopentadienyliron is considerably greater than many organic groups, it will be shown ( vide infra ) that steric changes in the organometal system are not the dominant influence in the reduction of A<5 and that this change in A6 corresponds better with the backbond! ng ability of the cyclopentadienyliron substituent. The decreased derealization observed in FpHCOT can be thought of as the result of backbonding by the Fp substituent into the homoaromatic it system and consequent localization of the tt electrons and positive charge. Additional spectral evidence that significant iron carbene 13 carbon double bonding exists is found in the C nmr spectrum of this cS: compound (Figure 2). Large low field chemical shifts of the carbene carbon have been found in Fp carbene complexes. The resonance of the

PAGE 22

16 ^ *

PAGE 23

17 carbene carbon at 269 ppm for FpHCOT is 147 ppm further downfield than 24 the comparable resonance at 122 ppm in homotropylium, and represents a greater downfield shift than is observed in Fp =cycloheptatrienylium, consistent with the expectation that a homoaromatic derealization would A = 147 ppm be more readily disrupted by backbonding than a fully aromatic delocalization. The above evidence clearly supports the description of this material as a homoaromatic carbene complex, a hitherto unknown compound having important ramifications in the descriptive chemistry of aromatic carbene complexes. Two additional carbene complexes of cyclooctatrienylium will be described, each having a ligand of different electronic and steric characteristics, in an attempt to better characterize the overall structural and electronic properties of these homoaromatic complexes. The three carbene complexes described in this work all have a common starting material in (n. -cyclopentadienylirondicarbonyl)cyclooctatetraene, FpCOT. This sigma complex and its phosphinylated derivatives have some interesting properties involving the conformational interconversion processes of cyclooctatetraene which will be discussed here.

PAGE 24

18 A synthetic route from cyclooctatetraene (COT) producing reasonable yields of FpCOT was described earlier. This substituted COT was found to have an unusual property apparent from the H nmr in Figure 3. The COT resonances are spread out over a range of nearly one ppm, a remarkable change from the normal broadened singlet observed for all other monosubstituted COT's. This is apparently an anisotropic effect due to local variations in the applied magnetic field near the iron atom combined with conjugative effects, because acyl complex 11, in which the iron and COT are "insulated" from one another by the carbonyl , does not FP'' -T) ii show the expansion of the COT resonances, but simply exhibits the normal broadened COT singlet (Figure 4). In effect, this sigma bonded complex, FpCOT, contains an internal shift reagent. This expansion of the COT resonances dramatically eases structural identification and the observation of the well-known COT interconversion kinetics. Unlike previous studies of COT where examination of C satellites at low temperatures and full line shape analysis was required to determine proton coupling constants and chemical shifts, this information can simply be measured from the spectra of the sigma complex (Figure 5).

PAGE 25

19 Q. — O UCD

PAGE 26

20 r—

PAGE 27

21 CL

PAGE 28

22 »0 i 4 • =3 ii in — +J o

PAGE 29

23 The close match between the reported coupling constants of 11.4 and 3.87 for COT, and the respective couplings in FpCOT implies that little steric distortion occurs even though Fp is by far the largest reported COT substituent. FpCOT, like COT, undergoes isodynamic ring interconversion processes (Figure 7). These processes were studied by diastereomer exchange and line shape analysis for COT and several monosubstituted derivatives and were found to have free energy of activation barriers ranging from 12.5 to 14.8 kcal/mol for the Ring Inversion (RI) processes, and 14.9 to 17.4 27 kcal/mol barriers for Bond Shift (BS) interconversions. The measurement of this Bond Shift activation barrier, in all reported cases in the literature, is performed by observation of the kinetic exchange of signals due to R„ and R . It is necessary to use complicated line shape analy2 o 2fi sis to identify the R ? and R ft resonances when monosubstituted COT's (R, f H) are being investigated, although multisubstituted COT's can be more easily examined by observing exchange of the substituent resonances. Hz This skew coupling constant has a reported range of ca. 2.5 to 3.87 28

PAGE 30

24
PAGE 31

25 However, R ? = H2 and R R = H8 resonances of monosubstituted FpCOT are easily identified (H2 at 5.68 ppm and H8 at 6.16 ppm in Figure 5) and the sample may simply be warmed to coalesce these resonances (Figures 8 and 9) to identify AG BS t 17.9 kcal/mol This same procedure was used to measure the Bond Shift activation barriers for two other substituted sigma complexes. Both of these new compounds, (n -cyclopentadienylironcarbonyltri-n-butylphosphine) Cycloid octatetraene (Fpp B COT) 12, and (n -cyclopentadienylironcarbonyltriphenylphosphine) cyclooctatetraene (Fpp,,C0T) JJ3, were made in high yield via photolysis of FpCOT in presence of tri-n-butylphosphine or tri phenyl phosphine respectively. * 28 The simple coalescence formula has been applied to determine rate constants at the coalescence temperature. Although this equation is strictly applicable to the exchange of single lines of equal population, negligible errors will result when applied to coupled lines if J is much less than 6v as applies here (J/Sv ca. 0.2). A precision of 0.5 kcal/mol results when the possible range of temperatures at coalescence is ca. ±5°C.

PAGE 32

26 *-* o I in 1 Q. OJOO (/) o •

PAGE 33

27 E-r=> 3

PAGE 34

28 //\ot -JtU /A 12 13 It will be observed that the major spectral difference between FpCOT and these compounds (Figure 10 and Figure 11) other than the phosphine ligand signals, is the double multiplicity of the cyclopentadienyl (Cp) and COT resonances. This is due to the diastereomerism inherent in these molecules. It is well known that monosubstituted COT's are chiral and it should be evident that these organoiron moieties are "pseudotetrahedral" and are thus chiral when substituted with a phosphine ligand. These chiral metals are optically stable in the temperature range of interest but this is not true of the COT ligand. Both the Ring Inversion and 29

PAGE 35

29

PAGE 36

30

PAGE 37

31 Z^> Fe .// \, PR 30 Bond Shift processes exchange configurations, and the rapid exchange of diastereomers results. This exchange can be observed as the CO / PR, CO .// 1 // /<— PR. CO / PR -A averaging of the nmr Cp resonances due to the diastereomer pairs in each of these phosphinylated sigma complexes. This behavior is shown in Figure 12.

PAGE 38

32 coalescence 38°C 32°C Figure 12. H nmr of exchange between diastereomeric cyclopentadienyls in Fpp Bu C0T. The fine splitting seen in the low temperature spectrum is due to the 1.1 Hz 31 P1 H coupling.

PAGE 39

33 Because AG BS ' is usually substantially greater than AG RI * and thus the rate constant k R( , very much smaller than k RI , observation of racemization of COT or exchange of diastereomers is generally attributed to RI processes exclusively. (For an exception see reference 30.) This 28 £ was assumed when calculating the AG RI T for Fpp B COT and Fpp COT as 17.0 and 17.3 kcal/mol respectively. As discussed previously, the observation of exchange in the COT resonances is a phenomenon due exclusively to Bond Shift processes and is entirely separate from the above consideration. In fact the activation energies observed for the coalescence of the COT resonances in these Table 1. Approximate Activation Barriers Molecule AG^ RI AG' BS FpCOT * Fpp Bu C0T 17.0 Fpp COT 17.3 17

PAGE 40

34 Nonetheless, it is apparent that AAG', the difference in activation barriers for the Bond Shift and Ring Inversion processes (AGpo* AG RT ') is smaller than, by possibly as much as one half, the AAG* for other 27 reported monosubstituted COT's. Table 2. Ring Inversion and Bond Shift Activation Barriers for Various Substituted Cyclooctatetraenes Substituent

PAGE 41

35 CO PR, This is feasible since the occurrence of metal it backbonding into aryl systems should be more disfavored than metal /T-antiaromatic system in31 32 teractions, yet is known to exist. ' Alternately, steric interactions are known to increase in going to the planar transition state and can cause reduction in the AAG T . While the steric demands of the iron substituent are not high, steric interactions do occur and can be observed using "fast" spectroscopic 33 techniques like IR. Still, it is unlikely that steric interactions in this system can be of the same order of magnitude as in 1,2,3-trimethylCOT, which is the least crowded COT reported to depress AAG* values to below 2 kcal/mol 30 Me PR, Me Me' o.

PAGE 42

are more significant in reducing AAG' the larger group, Fpp,,, should 36 Clearly additional work needs to be performed on these systems before final conclusions may be drawn, but it should be noted that an explanation of the small observed AAG* should be accessible. This is be34 cause the Fpp„ substituent is smaller than the Fpp substituent and also is more electron donating relative to Fpp„,; if steric interactions int have decreased AAG' , while if electronic effects dominate, the more electron donating group, Fpp„ , should have decreased AAG*. The latter effect seems to be indicated from the present data. This result would be interesting because it would be a rare, if not exclusive, example of the effect of backbonding by an organometallic substituent into an antiaromatic ring system. A significant comparison could then be made between the above effect and its counterpart in the aromatic carbene complexes. As discussed earlier, protonation of the FpCOT sigma complex produced a new homoaromatic carbene complex. Similarly, protonation of 1 5 Fpp-COT leads to n -cyclooctatrien-1-ylidene-n -cyclopentadienylironcarbonyltri phenyl phosphine (Fpp^HCOT ) ( 14 ) which can be isolated as its hexafluorophosphate salt. Although this compound is considerably more L CO HFP

PAGE 43

37 thermally stable than the parent homo aroma tic carbene complex, FpHCOT , it too is thermally labile and will decompose upon standing at RT for several hours. Therefore, once again the structural proof relies on the nmr spectra (Figures 13 and 14). As can be seen in the proton nmr, the most downfield signal at 8.5 ppm represents one hydrogen adjacent to the carbene center and identifies the structure as being similar to the parent FpHCOT . The large multiplets at 7.5 and 7.4 ppm are due to the Fppt triphenylphosphine (P0,) ligand, and the next four resonances at 7.1, 5.8, 5.6, and 6.5 ppm are H5, H6, H4, and H3 respectively as identified by decoupling. The quartet resonance at 5.4 ppm next to the solvent signal is H7. The Cp resonance occurs at 4.9 ppm. The last two resonances represent the two geminal protons of the homocon jugate CFL linkage, identified as H8 exo at 4.8 ppm and H8 endo at 1.7 ppm by their coupling 7.2 8.2 H-. H 10.0 "5 7.3 -Fp — H 8.9 H' H 8.9 7.7 " Fpp fl

PAGE 44

38 CLr— Q. (O M O Q. nz cmq. C_5 00

PAGE 45

39 F

PAGE 46

40 constants, separated in chemical shift once again by the anisotropy of a substantial ring current. However, it is expected that this ring current would be reduced relative to the parent carbene, FpHCOT , because backbonding by Fpp~ into the homoaromatic ring should be more favorable than backbonding by Fp. This is because substitution of a phosphine ligand in place of a carbonyl results in greater electron density on iron. This will allow increased Fppg backbonding and increased localization of electrons in the cyclooctatrienylidene tt system. The reduction in the ring current explains the decrease in the difference in chemical shift of H8 exo and H8 endo, A6, to a value of 3.11 ppm, down from 4.30 ppm observed in FpHCOT and 5.86 ppm in homotropylium itself. That the diminuation of A6 is an electronic effect rather than a simple steric influence is supported by the extreme low field shift of 13 the carbene carbon resonance in the C nmr (Figure 14). The signal (doubled by phosphorous coupling) occurs at 311.4 ppm, nearly 200 ppm downfield from the C-l resonance in homotropylium, and is further downfield than that observed for the same carbon in FpHCOT . This indicates 13 greater backbonding contributions, consistent with the observed

PAGE 47

41 Fp + ^y25T^) FPP B ^V279 F P=rf42 161 changes in the proton nmr AS. Furthermore, the difference in shift of the homotropyl i urn ligand C-l carbon between FppgHCOT and FpHCOT is 42 fi 11 ppm, similar to the analogous difference of 37 ppm ' observed in the cyclophetatrienylidene series and reflects similar changes in backbonding due exclusively to ligand substitution. Another property of this carbene complex gives additional evidence that greater backbonding occurs than that seen in FpHCOT . It was observed that exchange broadening and shifting of the signal due to H8 endo occurred in the proton nmr for temperatures greater than 0°C (Figure 15). This was originally construed to indicate exchange with excess acid used to produce the nmr solutions of the carbene. However, disappearance of the signal was not observed when using deuterated acid; rather, the integrated signal intensity was seen to remain at one half proton. Moreover, exchange of this signal occurred when the hexafluorophosphate salt FpppjHCOT HFP was dissolved in solvent containing no acid. Therefore no acid exchange process was present.

PAGE 48

42

PAGE 49

43 It is known that homotropylium undergoes an exchange process in which the homoconjugate CH^ unit flips from a position above the approxi21 35 mately planar ' seven membered ring to a position below it. The £S activation barrier measured by Winstein for this process was 22.3 36 kcal/mol. Substitution of C-l on homotropylium with a methoxy group 23 reduces the activation barrier to 19.6 kcal/mol. Brookhart interpreted this reduction as being due to the greater substituent stabilization of 2 the positive charge on the sp rehybridized C-l of the transition state 3 15 than the substituent stabilization of the charge on the more sp H H R 15 23 £5 hybridized C-l of the ground state 16. With this in mind, a re-examination of the kinetic behavior of Fpp HCOT indicated that the signal due to H8 exo may have been obscured by the Cp resonance in the 100 MHz nmr. Protonation of Fpp^COT in d g benzene shifted the Cp resonance (probably due to Cp-benzene stacking interactions) to reveal the now relatively unobscured H8 exo resonance

PAGE 50

44 which showed temperature dependent exchange broadening. This same phenomenon was seen in the 300 MHz nmr (Figure 15). The exclusive exchage of H8 endo and exo signal is taken to indicate that the bridge flip process occurs much more rapidly in Fpp substituted homotropyl i urn than in normal homotropyl i urns since exchange broadening in 1-hydroxyhomotropylium 22 is not seen even to 80°C. Accurate activation barriers have not been obtained since thermal decomposition occurs rapidly above 60°C although a barrier of ca. 14 kcal/mol was estimated from line shape changes. However, it was seen that this exchange process in Fpp HCOT occurred more rapidly than in FpHCOT because no exchange broadening was observed when FpHCOT was rapidly heated to decomposition at 50°C. (Unfortunatepo "if. 07 ly, unlike other homotropyl i urns , ' ' no exo-endo specificity is seen when deuteration of FpCOT is carried out even at -100°C and therefore the kinetics of signal appearance could not be observed.) This last obser+ + vation of faster exo-endo exchange in Fpp^HCOT than in FpHCOT is consistent with Brookhart's interpretation that the activation barrier for bridge flip interconversion is reduced when more powerful electron dontaing substituents are present at C-l, the more powerfully donating substituent in this case being Fpp . A caveat is appropriate at this point in that the straightforward interpretation of the H8 exoH8 endo broadening presented here is likely too simplistic to account for more recent spectral observations of this phenomenon. High field nmr spectra which resolve the H8 exo signal show that it exchanges and shifts toward Jow field rather than towards the H8 endo signal at higher field. Speculation as to the nature of this process or of the possible superimposition of an additional process upon the normal bridge flip exchange is currently in progress.

PAGE 51

45 A somewhat more electron releasing substituent would be n -cyclopentadienylironcarbonyltri-n-butylphosphine, Fpp D . However, because bu the cone angle of occupied space in ligand PBu 3 is only 130 degrees as opposed to 145 degrees for P0 3 , this substituent should be sterically less demanding than Fpp . If steric factors play a significant role in the structural and electronic properties of these organoiron substituted homotropyl i urns , a reversal in the spectroscopic trends as seen for FpHCOT and Fpp HCOT should be apparent in the spectrum of n -cyclo5 octatri en1-yl i den-ri -cycl opentadi eny 1 i roncarbonyl tri -n-butyl phosphi ne , Fpp Bu HC0T + (J7) (Figure 16). When Fpp„ COT is protonated to produce this final new homoaromatic carbene complex, the general trends are preserved. The difference in shift of the CHp protons, A
PAGE 52

46 — CM UO

PAGE 53

47 A •

PAGE 54

48 relatively constant geometries are assumed. ' While the expected and observed results of this work are consistent with the latter statement, it is somewhat bothersome that a slight deviation in the expected pattern of convergence of the H8 shifts exists for Fpp D HCOT . This DU 6" 54-3-I1 '^Bu Ep 2 3 A6 • H8 exo H8 endo deviation can be seen above in the convergence of H exo-H endo shifts with increasing electron donation capability of the 1-substituent in the homotropyl i urn ligand. Whatever the source of this deviation, it is apparent that it results in a slight variation in the geometry of the molecule. This can be seen by the examination of the complete set of bridge coupling constants for these three systems and the comparison of these data with other substituted homotropyl i urns vide infra . One of the most often cited indications of homotropylium structure is the two bond proton proton coupling constant for H8 exo H8 21 22 23 38 endo. ' ' ' When originally formulated Rosenberg, Mahler and 39 Pettit were undecided about the electronic structure of protonated cyclooctatetraene (HCOT ). Subsequent work led them to strongly support

PAGE 55

49 €3 -30 the homoaromatic structure. However, Deno preferred a description of the molecule as the bi cyclic species. Winstein pointed out that the bi cyclic form was unlikely in that the expected proton coupling constants for a cyclopropylcarbinyl system were not consistent with the observed couplings in HCOT vide infra . J A>B = ca. 4.7 Hz J. c = £a. 4.5 Hz J D = ca. 8 Hz D ,U As can be seen below, the geminal and three bond cis and trans coupling constants of the organoiron substituted homotropyliums are quite similar to those observed in HCOT and other substituted homotropyliums.

PAGE 56

50 9.7 H 8-2 H 7.3 ~2 10.8 7.7 _,0H -J 10.5 9.8 7.5 ^OMe 8.9 8.9 1 7.7 F PP„ Clearly these carbene complexes are not distorted toward a closed cyclopropane type structure. Nor do the coupling constants reveal any significant departure from a relatively constant geometry, even though + * Fpp„ HCOT couplings reflect a small variation in the trend. What is exemplified in the above is the marked similarity between the homotropylium carbenes and the homotropyliums substituted with electron It should be pointed out that the figures for Fpp R HCOT were far more difficult to determine because of overlap of H7 an8 u H8 signals in the proton nmr, and probably have a larger error than the other carbene data.

PAGE 57

51 donating groups. These similarities give strong support that the changes observed above are primarily geometrical changes that result from electronic interactions. That small geometrical changes do occur, resulting from electronic interactions, can be seen through examination 13 1 of the CH coupling constants at C8. In cyclopropylcarbinyl this CH coupling is 180 Hz or more. In homotropylium, due to the modified tt type interactions between C-l and C-7, the larger C1-CH 2 -C7 bond angle is expected to cause reduction in the s character in the CH bonds, TO 1 leading to decreased coupling interactions between the C and H nuclei. The observed hydrogen-bridge carbon coupling in homotropylium is 159 l J = 180 Hz ^g H = 159 Hz J C8 H = 130136 Hz Hz, and the expected reduction in 1,7 bonding in FpHCOT and + * Fpp-HCOT produces the even smaller couplings, ca. 131 Hz and 136 Hz respectively. The somewhat low value for FpHCOT is probably an inaccuracy caused by the method used to obtain it. The actual value measured was the 13qdeuterium coupling constant, which was then multiplied by the ratio of H and 2 H magnetogyric ratios to calculate the *H-l3c coupling.

PAGE 58

52 This reduction in 1,7 bonding due to backbonding by the iron substituent should also be accompanied by increased localization of the tt system. This effect can be seen in the coupling constants of the protons about the periphery of the nearly planar seven membered ring portion of these systems (Figure 18). These values must be compared with the only other complete set of 21 coupling constants available, those from homotropylium itself. The differences between these two sets of values arise from the unsymmetrical backbonding substitution of the homotropylium framework and manifest themselves as an alternation in the magnitude of the coupling constants due to apparent tt localization and consequent alternation of the double and single bond character in the ring system. Prior to the recent X-ray 35 structural verification and the impressive theoretical calculations by 43 Haddon, the coupling constant information for homotropylium was the best evidence that homotropylium was comprised of a distinctly puckered 21 but symmetrical seven membered ring portion (7MR). A modified, but similar geometry is though to exist ( vide infra ) in the homoaromatic carbene complexes. The nonsymmetrical substitution of homotropylium, along with sequentially increasing backbonding by the cyclopentadienyliron series, Fp, Fpp and FpP Bu > should lead to sequential increases in the localization of the tt electron density and thus the alternation of the coupling constants. The full set of proton coupling constants for these molecules (Figure 18) are consistent with this and are quite accurate, as can be seen in the spectral simultions based on the evaluated couplings in Figures 19, 20 and 21. Interestingly, the asymmetry of the line shapes in FpHCOT (Figure 19) is related to the symmetry of a portion of its

PAGE 59

53 J 7,8 endo = 9 ' 8 J 7,8 exo = 7 2 8.2. 9.7

PAGE 60

54 CL Q. .

PAGE 61

55 m CD

PAGE 62

56 m

PAGE 63

57 7MR, which causes an unusually large long range coupling of 0.9 Hz between H2 and H5. Significant through space interactions such as this usually occur only when the bonds of the interacting nuclei are nearly colinear. The long range coupling (together with the extensive it H2 d^" Ftl derealization) helps to verify the approximate planarity of the 7MR in this Fp carbene complex. The more strongly backbonding substituents, Fpp,. and Fpp R , could cause an increased distortion of the 7MR from planarity as increases in the alternating bond character allow torsional changes to occur. That this is an electronic effect is shown by the correlation of backbonding strengths (but not steric considerations) and coupling constant alternation. The difference between adjacent couplings is presented in Table 3.

PAGE 64

58 Table 3. H Three Bond Coupling Constants for Monosubstituted Homotropyliums, and Differences in Adjacent Couplings.

PAGE 65

59 With this in mind, it is pertinent to note once again that both the A6 and bridge coupling constants fall in the range of respective values arising from a variety of 1substituted HCOT 's. Thus it seems possible that a direct comparison with normal organic functional groups may reasonably be made. There does not appear to be any significant indication in the previous data or discussion that would argue against such a comparison. Table 4. Comparison Between Various R Substituted Homotropyliums. H a^/ H b -\ J H ,H K = J gem a b J — J H ,H = J trans a c J H. ,H = J cis b c R groups ,21 Me 38 Fp FPPr: Fpp Bu OH 22 OMe 22 gem

PAGE 66

60 groups. The analogous measure of the it donating ability of organic func+ 44 tional groups is the a . When the a of the substituents of the reported 1-substituted homotropyliums is plotted against the A6 of that compound a relatively straight line is produced (Figure 22). Locating the A6 value of the three carbene complexes FpHCOT , Fpp^HCOT , and Fpp R HCOT on this line produces o values of -0.43, -0.80, and -0.84 respectively. These numbers should at best be considered approximations, and it would be wise to think of a phosphinylated cyclopentadienylironcarbonyl group (Fpp) as having about the same electron donating ability as a hydroxy or methoxy group, and of a cyclopentadienylirondicarbonyl group (Fp) as a substantially weaker donor. Nonetheless, this is the first sensitive indication of organoiron tt donating strength directly comparable to normal organic groups. Other comparative methods have not been sensitive 45 46 enough to separate carbonyl and phosphine substitution effects, ' . 2 . 4 . 6 . 8 1 -1.2 -1.4 -1.6 -1.8 -2 V
PAGE 67

61 much less the backbonding effects of alkyl versus aryl phosphine substitution as observed in this work. This comparison fairly sharply delineates the tt donor capability of Fp and Fpp in the homotropylium system. Certainly some variation in this parameter would be expected with other carbene ligands having different electronic demands. Nonetheless, an approximate indication of the relative backbonding or it donating ability of these organoiron groups has a clear advantage over the activation energy of rotational barriers as a predictive descriptor in the chemistry of these cationic carbene complexes. This is especially true in the 6tt electron aromatic series, the members of which have notable steric contributions to their 13 rotational barriers. For the sake of comparison, an attempt was made to measure the barriers to rotation in the three homoaromatic complexes as well, but this data could not be collected. In the case of FpHCOT , any barrier to rotation was expected to be low through analogy with the proposed small barrier in Fp + =C ? H 6 (3). No changes in the H nmr of FpHCOT were seen between temperatures of -100°C to -10°C, above which decomposition was seen to occur rapidly. (Highly acidic conditions were seen to alleviate this thermal sensitivity somewhat, so that the rapid heating to 50°C was accomplished in 90% trifluoroacetic acid without spectral changes other than final decomposition.) Combined with the observation of a 13 single set of diastereotopic carbonyl resonances in the C nmr (Figure 2), the spectral data point to rapid (nmr time scale) rotation about the iron-carbene bond.

PAGE 68

62 diastereotopic chiral On the other hand, the large phosphine group in the FppHCOT 's could be expected to interact even more strongly with the rotating carbene ligand than was seen in FppC 7 H fi (_7) in which a 9.6 kcal/mol elec13 tromc and steric composite barrier to rotation was observed, because the greater backbonding ( vide supra ) in the homoaromatic complexes would not only lead to closer proximity of the ligands, and thus increase steric interactions, but would also increase the electronic contribution to the rotational barrier. However, no spectral changes were seen from -100°C to 50°C other than the previously mentioned exchange broadening of H8 exo and H8 endo. In view of diastereomerism inherent in these & A CO 25 PR. diastereomers

PAGE 69

63 FppHCOT complexes, it is possible that a considerable difference in the energy of the two diastereomers could prevent the observation of an exchange between signals of greatly different intensity in the proton nmr. This diastereomerism may be the cause of difficulty in obtaining crystalline salts of the Fpp complexes. In both Fpp g HCOT HFP (17) and Fpp-HCOT HFP (14) only amorphous solids could be collected after careful attempts at recrystallization. Since the non-diastereomeric FpHCOT HFP (10) crystalline solid was ultimately too thermally sensitive for X-ray analysis, the symmetrical bis(l,2-diphenylphosphino)ethane (DIPHOS) substituted carbene complex was sought. It was hoped that this compound, 1 5 n -cyclooctatrien-1-yliden-n -cycl opentadienyli ron[bis(l,2-di phenyl phosphino)ethane] hexafluorophosphate (CpFe(DIPHOS)HCOT HFP) (18), would be more thermally stable than the other HCOT carbenes, would crystallize for X-ray analysis, possess adequate symmetry for rotational barrier detection and provide a final point for comparison of a and A6. However 5 the starting material n -cyclopentadienyliron[bis(diphenylphosphino) ethane] cyclooctatetraene (19), could not be synthesized, as discussed below.

PAGE 70

64 CpFe(DIPHOS) f 19 Addition of DIPHOS to FpCOT by photolysis occurred with rapid evolution of carbon monoxide, but the mixture of products did not contain substituted cyclooctatetraene, nor did addition of strong acid to a solution of this material produce any carbene complex signals in the nmr. 20 Addition of cyclooctatetraenyl lithium (Li COT) to cyclopentadienyliron(DIPHOS)chloride was not attempted since the addition of cyclohepta50 trienyl lithium to the chloride did not yield a desired product. Reaction of LiCOT with [CpFe(DIPHOS)] 2 N 2 (HFP) 2 51 produced a s/ery small yield of a mixture possibly containing the product lj), but this was not followed to conclusion because of the extremely low yields. This line of research was abandoned in order that other pertinent questions might be addressed. For example, the thermal decomposition of FpHCOT HFP (10) was shown to produce a COT complex of Fp . This probably procedes via a 1,2-6 hydrogen shift from the homoconjugate CH^ to pro5 2 duce the olefin complex (n -cyclopentadienylirondicarbonyl)-n -cyclo+ 52 octatetraene (Fp -COT) (20). A possible mechanism for this process is indicated below.

PAGE 71

65 £) reductive el imination The structure of this product was established by the release of COT (as identified by H nmr) from a solution of the thermally decomposed carbene complex via nucleophillic displacement of the COT from Fp with 53 sodium iodide. Also observed in the products was Fp iodide. Fp o Nal Fpl This reaction was verified by grinding the thermally decomposed FpHCOT HFP (10) with KC1 prior to introduction in the mass spectrometer. A reaction analogous to that seen in solution apparently occurs in the solid state, since COT* was identified (high resolution in accord with m/e calculated for COT) as well as trace signals attributable to the modestly volatile FpCl and its fragmentation products.

PAGE 72

66 This decomposition is exactly what is expected of a nonheteroatom stabilized organometallic carbene complex having B hydrogens present. On the other hand, the exhibited stability of these homoaromatic carbene complexes is more analogous to that of heteroatom stabilized carbene y 0Me + OMe ^ (C0) c Cr— / «-* (C0) 5 Cr=C ^ (C0) 5 Cr— C 5 CH 3 21 CH 3 h+ CH 2 complexes such as 2_1. Neither the homoaromatic nor heteroatom stabilized carbene complexes undergo facile 1,2-hydrogen transfer reactions, probably because the carbene centers are part of the conjugated tt systems. However, pentacarbonyl[(methoxy) (methyl )carbene]chromi urn (0) 21^, is some54 what acidic, and thereby indicates the availability of its tt hydrogens for 1,2-shift decompositions. This points out another possible stabilizing factor in these homotropyl i urn carbenes: the nearly orthogonal orientation of the homoconjugate CH ? bonds to the metal carbene conjugated tt system. Without appropriate orbital orientation, 1,2-hydrogen shifts cannot occur. This would imply that the pathway to decomposition is the bridge flip intermediate, where neither stabilizing effect mentioned above is as effective.

PAGE 73

67 A problem here is the observation that the FppHCOT 's are more thermally stable than FpHCOT , yet with Fpp substitution, this bridge flip intermediate should be more accessible ( vide supra ) , therefore leading to more rapid decomposition. It is possible, however, that greater backbonding by Fpp would reduce the charge density on C-l, thus inhibiting 1,2-hydrogen shifts relative to Fp substitution. Nonetheless, it is apparent that the ground state homoaromatic carbene complex is a modestly stable organometallic species having 6 hy55 drogens present. This is sufficiently rare enough to warrant note, but it is clear that the geometrical alteration upon protonation to form these complexes is quite unique, so that this method of forming carbene complexes should not be regarded as having general utility. In conclusion, the synthesis and characterization of the first three homoaromatic carbene complexes are described in this work and a strong correlation has been made with the it backbonding ability of cyclopentadienylirondicarbonyl (Fp) and cyclopentadienylironcarbonyltrialkyl (or aryl)phosphine (Fpp), and the a of normal organic functional groups. Additionally, three novel organoiron sigma bonded complexes of cyclooctatetraene (COT) were developed for the above work which dramatically display the known isodynamic conformational processes of COT, and which may reveal a hitherto undescribed organometallic interaction with a 4n tt electron orbital system.

PAGE 74

CHAPTER III EXPERIMENTAL General The melting points were obtained on a Thomas Hoover melting point apparatus and are uncorrected. Nuclear magnetic resonance spectra were obtained on either a Jeol FX-100 or Nicolet NT-300 spectrometer, and are reported in ppm relative to TMS or the deuterated solvent resonance. Coupling constants for the nmr spectra were either measured directly if the line was a simple doublet or simulated using the Nicolet 1180 ITRCAL program. Infrared data were obtained on a Perkin-Elmer 137 spectrophotometer. Combusion analyses were performed by Atlantic Microlab Inc., Atlanta, Georgia. Mass spectra were obtained on an AEI MS 30 spectrometer. All nmr solvents were degassed by bubbling nitrogen through the cold solvent for ten minutes and subsequently stored in a Vacuum Atmosphere Inc. recirculating glove box under oxygen-free conditions. All other solvents were distilled. All solvent and reagent solutions were swept free of Op by bubbling N„ through the solution for at least ten minutes when dealing with air sensitive compounds. THF was distilled from benzophenone ketyl using Na/K alloy. Methylene chloride was distilled from P 2 5 . 68

PAGE 75

69 The alumina used was Fisher certified neutral alumina, Brockman Activity I, to which 3% w/w water was added. After the addition of water, the alumina was allowed to stand at least 12 hours before being purged several times with nitrogen immediately prior to use. All reactions were performed under a nitrogen atmosphere and Schlenk type apparatus was used when appropriate. Cycl opentadi eny 1 i rondi carbonyl Iodi de (Fp I ) This material was used as supplied from Alfa Products. Tri-n-butylphosphine (PBu 3 ) This material was used as supplied by Aldrich Chemicals after degassing with N„. Tri phenyl phosphine (P0 3 ) This material was recrystallized from 95% ethanol . It was obtained from Aldrich Chemicals. Bis(l,2-di phenyl phosphino)ethane (DIPHOS) This material was recrystallized from acetone. It was obtained from Alfa Products. Trifluoroacetic Acid (TFA) This material was distilled and degassed prior to use. <2 3 -TFA was made by hydrolysis of acetic anhydride with D^O.

PAGE 76

70 Tri phenyl carbeni urn Hexaf 1 uorophosphate (0 3 C HFP) This material was used as purchased from Aldrich Chemicals. Fluorosulfonic Acid (FSOgH) This material was distilled and degassed with nitrogen bubbling before use. FSCLD was used as purchased from Aldrich. Potassium Cyclopentadienylirondicarbonyl This material was made according to the method of Gladysz et al. Bromocycl ooctatetraene (BrCOT) To a dry 1000 mL 3 necked round bottomed flask equipped with mechanical stirring, N 2 bubbler and a 300 mL addition funnel, and containing 47.2 g (0.45 mol) of freshly distilled cycl ooctatetraene (COT) dissolved in 500 mL dry CH 2 C1 2 at -78°C was added dropwise 25.0 mL (77.5 g, 0.48 mol)Br 2 in25mLdry CH 2 C1 2 . After stirring for 15 minutes, 75 g (0.67 mol) potassium t-butoxide dissolved in 250 mL dry THF was added portionwise to the di bromide solution over the course of 6 hours, maintaining the temperature between -60 and -78°C. The mixture was stirred overnight at low temperature and then, after warming to room temperature, poured over 500 mL crushed ice, swirled gently and placed in a separatory funnel. The organic layer was separated, and the aqueous layer and emulsion mixed with 50 mL diatomaceous earth, suction filtered and extracted with 3x50 mL ChLCl,,. The combined organic layers were washed with water, dried over MgSO. and flash evaporated at room temperature.

PAGE 77

71 The crude BrCOT was mixed with an equal portion of pentane, filtered through alumina and flash evaporated to obtain a dark yellow liquid which was carefully distilled between 40 and 45°C at 0.5 torr to collect 70 g BrCOT as a yellow oil in 85% yield. The IR and nmr spectral properties are in accord with those reported in the literature. Cyclooctatetraene Carboxylic Acid (COT-CO^H) To a stirred solution of 6.4 g (34.6 mmol) BrCOT in 50 mL dry ethyl ether and 4.0 g clean Mg turnings was added ca. 2 mL of a 1:1 solution of ethane dibromide in ether. The solution was warmed to initiate the reaction and then placed in an ice bath with the slow addition of the remainder of the 1:1 di bromide-ether solution (total ethane dibromide 6.5 g, 34.6 mmol) such that ethylene continued to be evolved without interruption. The solution was strongly stirred for about 3 hours and then rapidly siphoned over crushed, dry, solid CO,,. After warming to RT, the mixture was extracted with 3 x 100 mL HL0. The aqueous layer was then acidified to pH 3 to form a cloudy yellow suspension of C0TC0 ? H. This was extracted with 3x50 mL pentane, dried over MgSO, and rotovapped to an oil which crystallizes to form yellow plates after standing for several days to produce 3.25 g C0T-C0 ? H in 64% yield. C0T-C0 2 H: mp 66°-69°C; *H nmr (CDC1 3 )6 5.9 (s, COT), 10.3 (br s, 1H). Cyclooctatetraene Acid Chloride (C0T-C0C1) Cyclooctatetraene carboxylic acid (3.07 g, 20.7 mmol) was placed in a 50 mL round bottomed flask equipped with magnetic stirring, reflux condenser and heating mantle. To this was added 6.8 g (4.4 mL, XS) oxalyl chloride in 10 mL benzene portionwise with much bubbling. The

PAGE 78

72 solution was allowed to stir until CO evolution ceased and then was gently refluxed for 2 hours. The excess benzene and oxalyl chloride was removed by flash evaporation, and another 20x2 mL benzene added and removed to separate the product from residual oxalyl chloride. The brown oil was then distilled at 52°-56°C at 0.05 torr to yield 3.25 g C0T-C0C1 in 93% yield. C0T-C0C1 : IR (neat) 3010, 2400, 1750, 1625, 1250 cm" 1 . (n -CyclopentadienylirondicarbonyDcyclooctatetraene Carbonyl ( 11 ) To a stirred solution of 1.3 g (6.2 mmol) KFp in 30 mL dry THF at -78°C was added dropwise 1.1 g (6.6 mmol) cyclooctatetraene acid chloride in 15 mL dry THF. After completion of the addition, the stirred solution was allowed to slowly warm to room temperature over the course of 2 hours. The reaction mixture was then coated onto 10 mL of silica gel by flash evaporation and chromatographed on a 6 by 1 inch silica gel column using benzene as eluent. After the brown Fp dimer band was eluted, the yellow band of acyl complex was carried off the column by adding up to 50% CHC1 3 to the benzene eluent. The product, 0.62 g of 11 , was collected as a yellow-orange crystalline air stable solid in 36% yield after evaporation. _U: mp 106°-110°C with decomposition; IR (CHC1 3 ) 2000, 1600 cm" 1 ; l U nmr (CDC1 3 )6 4.9 (s, Cp), 5.9 (s, COT). Photolysis of 11 No products other than highly insoluble solids were obtained after photolysis of hexane or MeCl 2 solutions of lj^ using either pyrex or quartz filtered 550 W Hanovia Hg lamps.

PAGE 79

73 5 (n -Cycl opentadi enyl i rondi carbonyl ) c.ycl ooctatetraene (FpCOT) (_9) To a rapidly stirred solution of 1.0 mL (1.42 g, 7.8 mmol) BrCOT in 250 mL dry, 2 free THF was added 3.5 g (16 mmol) KFp in THF at room temperature. The addition of the KFp solution was performed at high dilution (to reduce Fp dimer formation) by a modified, nonsiphoning Soxelet extraction of the modestly soluble KFp by refluxing the THF under reduced pressure (90 120 torr) while maintaining the solution temperature at room temperature with a heating mantle (BrCOT is thermally labile). After no further KFp was extracted, the darkened reaction mixture was vented with nitrogen and stirred overnight. The crude mixture was then coated onto 10 mL of Grade II Fisher neutral alumina by flash evaporation and chromatographed on a 1 by 4 inch degassed, Grade II Fisher neutral alumina column using degassed hexane as eluent. A pale yellow BrCOT band preceded the intense yellow FpCOT band which was collected under nitrogen to produce 0.94 g FpCOT in 43% yield. The yield varied between 20 and 43% depending on the dilution of the Fp anion addition. This yellow, air-sensitive oil should be stored refrigerated in solution (do not use CHC1,). FpCOT: Anal, calcd. for C 15 H 13 Fe0: C ' 64 31 ; H ' 4 32 Found: c > 64.07; H, 4.40; IR (film) 2960, 1990, 1925, 1560, 1400, 1005, 909, 827, 802, 714 cm" 1 ; l H nmr (300 MHz, CD 2 C1 2 )6 4.81 (s, Cp), 5.22 (dd, H7, J ? g-11.1 Hz, J g , = 3.4 Hz), 5.51 (dd, H5, J g 6 = 10.9 Hz, J 4 5 = 3.2 Hz), 5.68 (fine d, H2, J 2 3 = 2.8 Hz), 5.75 (dd, H6), 5.77 to 5.85 (m, H3 and H4) , 6.16 (d, H8); 13 C nmr (d 6 -acetone)6 87.2 (Cp), 119.3, 129.2, 132.6 (2C), 136.7, 139.3, 147.7, 151.7 (CI), 217.2 (CO), 217.7 (CO); mass spectrum m/e, 280.01984 (calcd. 280.01850).

PAGE 80

74 * An alternative synthesis produces FpCOT in lower yield. To a cold 20 (-78°C) solution of LiCOT, made by the method of Cope and Burg from 2.42 g (13.2 mmol) BrCOT, was added 4.50 g (13.2 mmol ) Fpl in 20 mL dry, 0„ free THF dropwise with stirring. After stirring for 2 hours at -78°C, the solution was allowed to warm to room temperature and stirred for an additional i hour. The dark mixture was adsorbed onto 10 mL degassed Grade II neutral alumina and chromatographed as above using a 12x1 inch column. The product, 0.61 g FpCOT, was collected in 16% yield. (n -Cycl opentadi enyl i roncarbonyl tri n_butyl phosphi ne )cycl ooctatetraene (Fpp Bu C0T) (12) To a solution of 0.23 g (0.82 mmol) FpCOT in 50 mL dry, 0~ free THF in a tall, narrow Schlenk tube was added 0.30 mL (0.24 g, 1.2 mmol, 1.5 fold XS) tri -ji-butyl phosphi ne. Nitrogen was bubbled briefly through the solution to sweep out residual 0„ and the solution was then irradiated through Pyrex under N ? at RT (H„0 cooled jacket) 6 inches from a 550 W Hanovia Hg lamp. The reaction progress was monitored by taking 0.2 mL aliquots and evaporating to dryness, redissolving in CHC1, and observing the 2000 cm region for disappearance of the two Fp absorptions and appearance of the single Fpp R absorption at 1925 cm" . Photolysis appeared to be complete after 3 hours but was continued 1 additional hour. The product, 0.35 g (0.77 mmol) Fpp^COT as a yellow-orange air-sensitive oil, was collected in 94% yield after chromatography the reaction This is essentially the same as the procedure reported by Cooke et al . 19

PAGE 81

75 mixture as described above for FpCOT. This material should be stored in the same manner as the parent FpCOT. Fpp B COT: IR (CHC1-) 2995, 1925 cm" 1 ; X H nmr (d g -acetone)6 0.9 to 1.9 (m, butyl), 4.4 and 4.5 (d's, Cp, diastereomers with J Hp = 1.1 Hz), 4.9 (m, H7), 5.3 (m, H5), 5.6 (m, H2, H6, H3, H4), 6.3 (m, H8); 13 C nmr (d g -acetone)6 24 to 32 (butyl), 83.6 (d, Cp, diastereomers), 115.7 (d, diastereomers), 127.4 (d, diastereomers), 132.0 (s), 133.3 (d, diastereomers), 137.0 (2C, d below 0°C), 149.9 (d, diastereomers), 167.1 (dd, CI, diastereomers), 221.9 (dd, CO). Alternatively, Fpp„ COT can be produced without isolating FpCOT. Addition of 2 fold XS (based on an assumed 40% yield of FpCOT) of trin-butyl phosphine to the crude rxn mixture of FpCOT in THF followed by overnight photolysis as above with stirring produced greater than 90% yields of Fpp g COT after workup. 5 (n -Cyclopentadienylironcarbonyltri phenyl phosphine)cyclooctatetraene (Fpp COT) (13) Fpp COT was prepared in greater than 80% yield from FpCOT by using a 1.5 fold XS of tri phenyl phosphine in the primary synthesis described above, however photolysis occurred more slowly. To a solution of 0.85 g (3.0 mrnol) FpCOT in 100 mL dry, 2 free THF was added 1.44 g (5.5 mmol) tri phenyl phosphine. After overnight photolysis under N ? using the previously described photolysis set up, the reaction mixture was coated onto 15 mL Grade II neutral alumina by flash evaporation and placed on a 1 by 4 inch column of degassed, Grade II neutral Fisher alumina in hexane. After eluting the XS P0-, and unreacted FpCOT with hexane, an orange band was collected under N 2 by eluting successively with 5% and

PAGE 82

76 10% MeCl 2 (or ether) in hexane. The product, 1.3 g (2.5 mmol , 82%) Fpp-COT as an air stable orange glass, remained after removal of the solvent. This mixture of diastereomers could not be recrystalized even after repeated attempts. Fpp COT: Anal, calcd. for C 32 H 27 FeOP: C, 74.72; H, 5.29. Found: C, 74.49; H, 5.35. IR (CHClg) 2995, 1920 cm -1 ; *H nmr (300 MHz; CD 2 C1 2 )6 4.45 and 4.49 (d's, Cp, diastereomers with J Hp = 1.1 Hz), 4.91 (dd, H7, J ? g = 11.11 Hz, Jg ? = 3.22 Hz), 4.95 (dd, H7 1 , J ?1 8 , =11.02 Hz, J 6 , ?1 =3.22 Hz), 5.30 to 5.84 (complex, H2, H5, H6, H4, H3), 6.10 (dd, H8) , 6.29 (dd, H8'), 7.37(br s, 0's); 13 C nmr (CD 2 C1 2 )6 85.4 (d, Cp), 116.6 (d), 130 to 140 (0 and COT), 149.5 (s), 164.4 (dd, CI of COT), 222.3 (dd, CO). Mass spectrum m/e, 486 (M-CO), 383, 262 (100%), 224,121, 56. (n -Cycl opentadienyli ron[ bis (1,2-di phenyl phosphi no) ethane] )cyclooctate traene (CpFe(DIPHOS)COT) (19) This compound could not be produced by the methods discussed below. Photolysis of 100 mg FpCOT with 1.2 equivalents DIPHOS in d g -benzene was performed using all quartz glassware and a 550 W Hanovia Hg lamp. After 5 minutes, CO evolution was observed but the nmr spectrum of the solution appeared unchanged. After 45 minutes of photolysis, several Cp resonances had replaced the original Cp signal at 4.5 ppm. Chromatography yielded only mixtures of unidentified products. Similar results were obtained when the above photolysis was carried out in THF using a pyrex filtered lamp. An alternative method also did not yield the desired product. A suspension of the highly insoluble [CpFe(DIPHOS)] 2 N 2 (HFP) 2 51 salt produced by the method of Sellman and Kleinschmidt (0.237 g, 0.175 mmol in 50 mL THF) was added to 0.5 mmol LiC(Tr u in THF at -78°C

PAGE 83

77 and stirred overnight at low temperature. No color change was observed until the solution was warmed to RT, after which the evaporate was coated onto 5 mL degassed Grade II neutral alumina. Chromatography produced hydrocarbon products along with a small quantity of possible product 19 visible as a small set of signals at 6.5 ppm (COT) and 4.5 ppm (Cp) in the nmr. Although the low yield of products was probably due to the insolubility of the nitrogen complex, no other appropriate solvents were found. 1 5 n -Cyclooctatrien-1-ylium-n -cyclopentadienylirondicarbonyl Hexafluorophosphate (FpHCOT HFP) (10) Excess anhydrous HC1 carried by N„ was bubbled through a cold (-40°C), magnetically stirred solution of 0.31 g (1.1 mmol) FpCOT in 25 mL degassed CH^Cl^ for £a. 15 minutes; the bright yellow solution immediately turned deep red due to the presence of the homoaromatic carbene complex. A two fold excess of degassed fluorosulfonic or trifluoroacetic acid in CHpClp could be used in place of the HC1 . To the deep red FpHCOT solution was added 0.43 g (1.1 mmol) tri phenyl carbeni urn hexafluorophosphate in 10 mL CH ? C1„ (ammonium hexafluorophosphate was not soluble) and the FpHCOT PF fi salt precipitated by slow addition of an equal volume of cold ethyl ether. The resulting microcrystalline orange needles were collected by filtration under N„ on a glass frit surrounded by a dry ice isopropanol cup. This thermally sensitive solid (stable below -10°C in air) was washed by the addition of 3 x 2 mL of cold ethyl ether and stored in the freezer (-30°C). A yield of 0.19 g (40%) FpHCOT was isolated from the FpCOT sigma complex. However, low temperature protonation of FpCOT in CD^Clp instantly generated nmr solutions of the

PAGE 84

78 carbene complex quantitatively. These solutions must be rigorously (L free. FpHCOT HFP: MP upon rapid heating, orange solid turns dark brown at about 70°C, melts at about 80°C, and then turns black with effervescence at 100°-105°C; IR (KBr) 2002, 1970, 1440, 1400, 1300, 1275, 1070, 980, 840 cm" 1 ; l W nmr (300 MHz, CD 2 C1 2 , -20°C)6 1.27 (t, H8 endo, J Q , =8.2 Hz, J 7 Q „, =10.0 Hz), 5.26 (s, Cp), 5.57 (t, H8 8 endo, 8 exo 7,8 endo r/ exo, J. Q „ = 7.3 Hz), 5.78 (t, H7, J c 7 = 9.0 Hz), 7.34 and 7.35 (m, / ,o exo o, / H4 and H3, J 3 4 = 9.1 Hz), 7.47 (t, H6, J g 6 = 6.9 Hz), 7.83 (m, H5, J 4 5 =H.4 Hz), 8.86 (d, H2, J 2 3 = 9.7 Hz, J 2 5 = 0.9 Hz); 13 C nmr (CD 2 C1 2 , -20°C) 62.3 (C8) , 89.8 (Cp), 133.2, 133.7, 134.3 (C4, C7, C6), 139.7 (C3), 143.2 (C5) , 169.0 (C2), 210.4, 211.5 (CO's), 269.0 (CI). 1 5 n -Cyclooctatrienyl-l-ylium-n -cyclopentadienylironcarbonyltri -nbutyl phosphi ne Hexaf 1 uorophosphate (Fpp Bij HC0T HFP) (17) This compound was produced from Fpp„ COT in the same manner as described above for the Fp carbene complex. The initially yellow orange solution immediately became intense burgundy colored upon addition of acid. The dark purple hexaf 1 uorophosphate salt is substantially more thermally stable than the parent, but is not crystalline and is subject to air oxidation. Fpp Bu HC0T HFP: IR (KBr) 2970, 1960, 1485, 1465,1425, 1080, 835 cm" 1 ; l H nmr (300 MHz, CD 2 C1 2 , -20°C)6 0.9 to 2.0 (m, Butyl), 2.48 (t, H8 endo, Jg endo>8 exo = 8 ' 3 Hz ' J 7,8 endo = 8A Hz ^ 4 ' 97 (d ' Cp, J Hp = 1.2 Hz), 5.48 (m, H8 exo and H7, J 7j8 eXQ = 7.7 Hz), 6.45 (t, H3, J 3 4 = 8.9 Hz), 6.62 (t, H4, J^ 5 = 11.5), 6.73 (t, H6, J 6j7 = 8.7 Hz), 7.04 (dd, H5, J g 6 = 5.7 Hz), 8.48 (d, H2, J^ = 11.7 Hz); 13 C nmr (CD 2 C1 2 , -20°C) 13.6 to 29.9 (n-Butyl), 62.7 (C8) , 90.3 (Cp), 123.2, 130.1, 134.3, 135.4, 137.4 (C3 to C4) , 161.5 (C2), 217.1 (d, CO, J cp = 28.1 Hz), 315.7 (d, CI, J cp = 22.0 Hz).

PAGE 85

79 1 5 n -Cyclooctatrien-l-ylium-n -cyclopentadienylironcarbonyltri phenyl phosphine Hexafluorophosphate (Fpp HCOT HFP) (14) This deep purple amorphous salt was produced from Fpp-COT in the same manner as described earlier for the Fp carbene complex except that a two fold XS of TFA was used in place of anhydrous HC1. The thermal stability of this compound is the highest of these three homoaromatic carbene complexes. Brief exposures to as high as 50°C in solution have been tolerated with no decomposition. After vacuum drying the salt appears to be air stable indefinitely when kept cold (-10°C); Fpp HCOT HFP: MP upon rapid heating, dark purple solid turns ash colored at ca. 100°C, then darkens at 150°C without melting; IR (KBr) 2995, 1975, 1490, 1465, 1430, 1095, 845 cm" 1 ; l H nmr (300 MHz, CD 2 C1 2 , -20°C)6 1.67 (t, H8 end0 ' J 8 endo, 8 exo = 8 ' 9 Hz ' J 7,8 endo = 8 ' 9 Hz >' 4 ' 79 <*• H8 exo > J 7,8 exo = 1J Hz) ' 4 ' 87 (d ' Cp ' J HP = l ' 2 Hz) ' 5 ' 42 (q ' H7 ' J 6,7 = 8 ' 5 Hz), 6.51 (t, H3, J 3 4 = 8.9 Hz), 6.66 (t, H4, 4 g = 11.6 Hz), 6.78 (t, H6), 7.07 (dd, H5, J g = 6.4 Hz), 7.37 to 7.77 (m, 0's), 8.51 (d, H2, J 2 3 = 11.5 Hz); 13 C nmr (CD 2 C1 2 /TFA, -20°C)6 61.7 (C8), 90.8 (Cp), 124 to 138 (Ring C's and 0's), 161.5 (C2), 216.4 (d, CO, J cp = 29.3 Hz), 311.4 (d, CI, J cp = 22.0 Hz).

PAGE 86

REFERENCES 1. F.A. Cotton and CM. Lukehart, Prog. Inorg. Chem. 16, 487 (1972). 2. D.J. Cardin, B. Centikays and M.F. Lappert, Chem. Rev. 72, 545 (1972). 3. E.O. Fisher and A. Maasbb'l , Angew. Chem. Int. Ed. Eng. 3, 580 (1964). 4. M. Brookhart and G.O. Nelson, J. Am. Chem. Soc. 99, 6099 (1977). 5. K. Ofele, J. Organomet. Chem. 22, C9 (1970). 6. N.T. Allison, Y. Kawada and W.M. Jones, J. Am. Chem. Soc. 100, 5224 (1978). 7. A. Cutler, R.W. Fish, W.P. Giering and M. Rosenblum, J. Amer. Chem. Soc. 94, 4354 (1972). 8. C.G. Kreiter and E.O. Fisher, Angew. Chem. Int. Ed. Eng. 8, 761 (1969). 9. M. Brookhart and J.R. Tucker, J. Organomet. Chem. 193, C23 (1980). 10. M. Brookhart, J.R. Tucker, T.C. Flood and J. Jensen, J. Am. Chem. Soc. K)2, 1203 (1980). 11. B.E.R. Schilling, R. Hoffman, D.L. Lichtenberger, J. Am. Chem. Soc. 101 , 585 (1979). 12. P.E. Riley, R.E. Davis, N.T. Allison and W.M. Jones, J. Am. Chem. Soc. 102, 2458 (1980). 13. F.J. Manganiello, M.D. Radcliffe and W.M. Jones, J. Organomet. Chem. 228 , 237 (1982). 14. C.E. Keller and R. Pettit, J. Am. Chem. Soc. 88, 606 (1966). 15. W.P. Giering, A. Sanders, L. Cohen, D. Kennedy and C.V. Magatti , J. Am. Chem. Soc. 95, 5430 (1973). 16. J. Gasteiger, G.E. Gream, R. Huisgen, W.E. Konz and U. Schnegg, Chem. Ber. 104, 2412 (1971). 80

PAGE 87

81 17. J. A. Gladysz, G.M. Williams, W. Tarn and D.L. Johnson, J. Organomet. Chem. L40, CI (1977). 18. R.B. King and M.B. Bisnette, J. Organomet. Chem. 2, 38 (1964). 19. This compound was subsequently found to have been made in a manner similar to our original synthesis by M. Cooke, C.R. Russ and F.G.A. Stone, J. Chem. Soc, Dalton, 256 (1975). 20. A.C. Cope and M. Burg, J. Am. Chem. Soc. 74, 168 (1952). 21. P. Warner, D.L. Harris, C.H. Bradley and S. Winstein, Tetrahedron Lett. 46, 4013 (1970). 22. M. Brookhart, M. Ogliaruso and S. Winstein, J. Am. Chem. Soc. 89, 1965 (1967). 23. M.S. Brookhart and M.A. Atwater, Tetrahedron Lett. 43, 4399 (1972). 24. G.A. Olah, J.S. Staral and G. Liang, J. Am. Chem. Soc. 96, 6233 (1974). 25. F.A.L. Anet, A.J.R. Bourn and U.S. Lin, J. Am. Chem. Soc. 86, 3576 (1964). 26. J.F.M. Oth, Pure Appl . Chem. 25, 573 (1977). 27. G.I. Fray and R.G. Saxton, The Chemistry of Cyclooctatetraene and Its Derivatives , Cambridge University Press, New York, 1978. 28. R.J. Abraham and P. Loftus, Proton and Carbon-13 NMR Spectroscopy , Heyden and Son Ltd., London, 1978. 29. T.C. Flood, P.J. DiSanti and D.L. Miles, Inorg. Chem. 1J5_, 1910 (1976). 30. L.A. Paquette, Pure Appl. Chem. 54, 987 (1982). 31. J.D. Duncan, J.C. Green, M.L.H. Green and K.A. McLauchlan, Chem. Commun. , 721 (1968). 32. M.D. Rausch, A.K. Ignatowicz, M.R. Churchill and T.A. O'Brien, J. Am. Chem. Soc. 90, 3242 (1968). 33. K. Stanley and M.C. Baird, J. Am. Chem. Soc. 97, 4292 (1975). 34. C.A. Tolman, Chem. Rev. 77, 313 (1977). 35. R.F. Chi Ids , A. Varadarajan, C.J.L. Lock, R. Fagginani, C.A. Fyfe and R.E. Wasylishen, J. Am. Chem. Soc. H)4, 2452 (1982). 36. S. Winstein, C.G. Kreiter and J.I. Brauman, J. Am. Chem. Soc. 88, 2047 (1966).

PAGE 88

82 37. H.D. Kaesz, S. Winstein and C.G. Kreiter, J. Am. Chem. Soc. 88, 1319 (1966). ~~ 38. C.E. Keller and R. Pettit, J. Am. Chem. Soc. 88, 604 (1966). 39. J.L. von Rosenberg, J.E. Mahler and R. Pettit, J. Am. Chem. Soc. 84, 2842 (1962). 40. N.C. Deno, Prog. Phys. Org. Chem. 2, 157 (1964). 41. S. Winstein, H.D. Kaesz, C.G. Kreiter and E.C. Friedrich, J. Am. Chem. Soc. 87, 3267 (1965). 42. G.A. Olah, D.P. Kelly, C.L. Jeuell and R.D. Porter, J. Am. Chem. Soc. 92, 2544 (1970). 43. R.C. Haddon, J. Org. Chem. 44, 3608 (1979). 44. H.C. Brown and Y. Okamoto, J. Am. Chem. Soc. 80, 4979 (1958). 45. R.P. Stewart and P.M. Treichel, J. Am. Chem. Soc. 92, 2710 (1970). 46. E.S. Bolton, G.R. Knox and C.G. Robertson, Chem. Commun., '664 (1969). 47. R.W. Taft, J. Phys. Chem. 64, 1805 (1960). 48. J. Hine, J. Am. Chem. Soc. 82, 4877 (1960). 49. R.W. Taft, E. Price, I.R. Fox, I.C. Lewis, K.K. Anderson and G.T. Davis, J. Am. Chem. Soc. 85, 3146 (1963). 50. F. Manganiello, M.S. Thesis, University of Florida (1980). 51. D. Sellman and E. Kleinschmidt, Angew. Chem. Int. Ed. Eng. 14, 571 (1975). 52. P.J. Davidson, M.F. Lappert and R. Pearce, Chem. Rev. 76, 219(1976). 53. K.M. Nicholas and A.M. Rosan, J. Organomet. Chem. 84, 351 (1975). 54. C.P. Casey and R.L. Anderson, J. Am. Chem. Soc. 96, 1230 (1974). 55. R.R. Schrock and G.W. Parshall, Chem. Rev. 76, 243 (1976).

PAGE 89

BIOGRAPHICAL SKETCH Marc Dudley Radcliffe was born in Schnectady, New York, in February 1953. After spending his early childhood on the coasts of New England, he and his parents moved to Titusville, Florida. There he watched the dawn of the space age and became interested in science. He attended college at Miami University in Oxford, Ohio, Worcester Polytechnic Institute in Worcester, Massachusetts, and finally at New College in Sarasota, Florida, before finding employment in the semiconductor industry, where he participated in the development of the early digital watch technology. He returned to college at the University of Florida in 1975 and subsequently received the Summer Undergraduate Fellowship in Chemistry. He began his graduate work there in 1977 with Dr. William M. Jones, received a supplemental graduate fellowship during his studies there, and has recently accepted a postdoctoral position with Dr. Kurt Mislow at Princeton University. 83

PAGE 90

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William M. Jones a Chairman Professor of Chemi^ry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Merle A. Battiste Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William R. Dolbier, Jr. Professor of Chemistry

PAGE 91

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Eric V. Dose Assistant Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Raymond J. BeVgeron Associate Professor of Medicinal Chemistry This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Liberal Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August 1982 Dean for Graduate Studies andResearch

PAGE 92

UNIVERSITY OF FLORIDA 3 1262 08553 1662