SYNTHESIS AND CHARACTERIZATION
OF A NOVEL CARBOCYCLIC
ALLENE COMPLEX OF IRON
FRANK JOHN MANGANIELLO
A DISSERTATION 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
to my wife and children,
Colleen, Wendy, and Micah,
for their love and understanding
I would like to thank Professor William M. Jones for his support
and help throughout the years. I would also like to express my grati-
tude for the support and encouragement that I received from my parents,
my wife's parents, and my grandparents.
Thanks must go to Dr. M. Radcliffe, Mr. R. Winchester and Mr. S.
Oon for all the 13C and low temperature NMR work. Finally, I would like
to thank Dr. M. Balci, Dr. L. Christinson, Dr. K. Biesiada and especially
Mr. J. Lisko for their help and friendship throughout the years.
Mrs. M. Waby deserves particular thanks for her invaluable assist-
ance in the preparation and typing of this manuscript.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . .
LIST OF TABLES . . .
LIST OF FIGURES . . .
ABSTRACT . . .
I INTRODUCTION . .
II ATTEMPTS TO GENERATE CYCLOPROPYLIDENE
COMPLEXES OF IRON . .
III GENERATION OF THE CARBOCYCLIC ALLENE
Fp COMPLEX . .
IV EXPERIMENTAL . .
Chemicals . .
Instrumentation . .
[FP2] (45) . .
Potassium Dicarbonyl-n -cyclopentadienyl-
ferrate [Fp anion] . .
Acid Chloride (66) . .
Acid Chloride (65) . .
Acid Chloride (55) . .
Acid Chloride (56). . ... 92
carbonyl)iron (51c) . .... .92
hepta-7-ylcarbonyl)iron. . 93
Reaction of Potassium n -Cyclopentadienyl-
dicarbonylferrate with 7-Chlorobicyclo-
[4.1.0]heptyl-7-carboxylic Acid Chloride. .... 94
Reaction of Potassium 5-Cyclopentadienyl-
dicarbonylferrate with 7-Bromobicyclo[4.1.0]-
heptyl-7-carboxylic Acid Chloride ... 95
diethylcyclopropylcarbonyl)iron (53). .. 95
diphenylcyclopropylcarbonyl)iron (52) ...... 96
Reaction of Potassium n5-Cyclopentadienyl-
dicarbonylferrate with Trans-2,3-diphenyl-
1-chlorocyclopropanecarboxylic Acid Chloride. 97
cyclopentadienyliron (62) . 98
iron (28) . . ... .. 98
diethylcyclopropyl)iron (61). . ... 99
diphenylcyclopropyl)iron (60) ... 100
Attempted Synthesis of Dicarbonyl-n5-
heptyl-7-iron (64). . ... 101
Reaction of 7,7-Dibromobicyclo[4.1.0]-
heptane with Potassium n -Cyclopenta-
dienyldicarbonylferrate . ... 102
Reaction of 7,7-Dichlorobicyclo[4.1.0]-
heptane with Potassium n5-Cyclopenta-
dienyldicarbonylferrate . ... 103
methyl)iron (71). . .. 104
hexenecarbonyl)iron (140) . .... 104
hexene)iron (95). . ... 105
hexene)iron Hexafluorophosphate (96). ..... .106
Reaction of Dicarbonyl-n5-cyclopenta-
dienyl(cyclohexene)iron with Tetrafluoroboric
Acid in the Presence of Trimethyl Phosphite 106
1-Bromo-7-methoxycycloheptene (75). ... 107
Method A . . .... .. 107
Method B . . ... .. 108
Method C . . ... .. 108
7-Methoxycycloheptene-l-carboxylic Acid (78). ... 108
Chloride (79) . .... 109
cycloheptene-l-carbonyl)iron (74) ... 110
1-cycloheptenyl)iron (72) . .. 111
Attempted Synthesis of Dicarbonyl-n5-cyclo-
1-Bromo-7-ethoxycycloheptene (109). ... 113
7-Ethoxycycloheptene-l-carboxylic Acid (110). ... 113
Chloride (111). . . ... 114
cycloheptene-l-carbonyl)iron (112). ... 114
Dicarbonyl-n -cyclopentadienyl-n -(7-ethoxy-
1-cycloheptenyl)iron (108). . ... 115
1-Bromo-6-methoxycyclohexene (127). ... 116
6-Methoxycyclohexene-l-carboxylic Acid (129). ... 116
6-Methoxycyclohexene-l-carboxylic Acid Chloride
(130) . .... .. .. .117
cyclohexene-l-carbonyl)iron (131) ... 118
Dicarbonyl-n -cyclopentadienyl-n -(6-methoxy-
1-cyclohexene)iron (126). . ... 119
Reaction of Dicarbonyl-n5-cyclopentadienyl-
with Trimethylsilyl Trifluoromethanesulfonate
Followed by the Addition of Ethanol ... 119
Reaction of Dicarbonyl-n5-cyclopentadienyl-
n -(6-methoxy-l-cyclohexenyl)iron (126) with
Followed by the Addition of Ethanol ... 120
Treatment of Dicarbonyl-n5-cyclopentadienyl-
with Me3SiOSO2CF3 .. . ... 121
Treatment of Dicarbonyl-n5-cyclopentadienyl-
with HBF4Et20. . . ... 124
Treatment of (72) with Me3SiOSO2CF3 Followed
by the Addition of Fp Anion . .... 127
Treatment of (72) with Me3SiOS02CF3 Followed
by the Addition of Tetrabutylammonium Iodide. 128
Reaction of (72) with Me3SiOS02CF3:
Isolation of Allene Complex (35). .
Treatment of (35) with Nal in d6-Acetone. .. .
Decomposition of (35) in the Air. .
Decomposition of (35) in CH2C12 ........
Dicarbonyl-n -cyclopentadienyl-n -(1,2-
cyclononadiene)iron Hexafluorophosphate (23).
Treatment of (126) with Me3SiOSO2CF3 .........
Decomposition of (35) in CH2C12 in the
Presence of Ph3CH . . .
APPENDIX LIST OF REAGENTS PURCHASED FROM
SPECIFIC CHEMICAL SUPPLY HOUSES . .
REFERENCES . . .
BIOGRAPHICAL SKETCH . . ... ..
LIST OF TABLES
1 Yields from the reactions of Fp~ with the
corresponding acid chlorides . .... .14
2 Yields from the photolysis of the
corresponding acyl complexes . ... 18
3 1H NMR resonances of the Cp of i-complexes .. 35
4 1H NMR resonances for (106a-c) . .... .47
LIST OF FIGURES
1 1H NMR spectrum of (74) in CDC13 ........... 31
2 1H NMR spectrum of (72) in CDC13 .......... 32
3 1H NMR spectrum of the crude products from the
addition of HBF4 to (72) . ... 34
4 1H NMR spectrum of (90) in d6-acetone ... .. 40
5 1H NMR spectrum of the crude products from the
addition of Me3SiOSO2CF3 to (72) . ... 43
6 'H NMR spectrum of the addition of Nal to the
crude products of Figure 5. . 44
7 100 MHz 1H NMR spectrum of the crude products
from the addition of Me3SiOSO2CF3 to (72) in
CDC13 at -300C . . 49
8 'H NMR spectrum of (23) in CD3NO2 at room
temperature . . ..... 51
9 100 MHz H NMR spectra displaying the fluxional
behavior of (35) . . 52
10 The addition of Nu- to (107) .. ... 54
11 H NMR spectrum of (108) in CDC13 ... 57
12 100 MHz H NMR spectrum of (35) in d6-acetone
at -200C ... .. 60
13 13C NMR spectrum of (35) in CD3NO2 at -200C ... 61
14 1C NMR spectrum of (23) in CD3NO2 at
room temperature . ..... 62
15 100 MHz 'H NMR spectra of (35) in d6-acetone ..... 65
16 (a) 100 MHz 1H NMR spectra of 35) in d -acetone;
(b) irradiation of Ha at -20C; (c) irradiation
at -600C . . 68
(a) 100 MHz 1H NMR spectrum of (35) in d -acetone;
(b) irradiation at Ha at -400C; Tc) irradiation at
-600C. . . .. .
18 H NMR spectrum of (35) and NaI in d6-acetone
at room temperature . .
19 H NMR spectrum of the crude products from the
addition of Ph CH to (35) in d6-acetone .. ..
20 1H NMR spectrum of the addition of Nal to the
products in Figure 19 . .
21 H NMR spectrum of the crude products from the
exposure of (35) to the air . .
22 H NMR spectrum of (131) in CDC. .
23 1H NMR spectrum of (126) in d6-benzene .
24 1H NMR spectrum of the crude products from the
addition of Me3SiOSO2CF3 to (126). .
25 H NMR spectrum of (96) in d6-acetone. .
26 Calibration curve of a) 1,3-cycloheptadiene
and b) cycloheptene . .
27 Calibration curve of cycloheptanone. .
28 1H NMR spectrum of (35) in d -acetone at -200C
after 1 hour at room temperature .
29 An expansion of the 1H NMR spectrum of (35)
in d6-acetone at -200C after 1 hour at
room temperature . .
30 1H NMR spectrum of (35) in CD3NO2 at -20C .
31 Calibration curve of a) 1,3-cyclohexadiene
and b) cyclohexene . .
Abstract of Dissertation Presented to the
Graduate Council of the University of Florida
in Partial Fulfullment of the Requirements for
the Degree of Doctor of Philosophy
SYNTHESIS AND CHARACTERIZATION
OF A NOVEL CARBOCYCLIC
ALLENE COMPLEX OF IRON
Frank John Manganiello
Chairman: William M. Jones
Major Department: Chemistry
The smallest stable carbocyclic allene is 1,2-cyclononadiene. As
one would anticipate as the carbocycli-c allene ring is decreased in
size, the stability is also decreased. In contrast, transition metal
complexes of Pt have been employed to stabilize these strained carbo-
cyclic allenes. Here, it was found that treatment of dicarbonyl-n -
cyclopentadienyl-n -(7-methoxy-l-cycloheptenyl)iron [(7-methoxy-1-
cycloheptenyl)Fp] with trimethylsilyl trifluoromethanesulfonate gave a
71% yield of dicarbonyl-n5-cyclopentadienyl-n2-(1,2-cycloheptadienyl)-
iron trifluoromethanesulfonate as a crystalline solid.
Examination of the 1H NMR, at low temperatures, revealed the
fluxional behavior of the metal, Fp, associated with known Fp allene
complexes. Furthermore, the activation energy, AGt, for the inter-
conversion, [1,2]Fp shift, of the isomeric allene complexes was observed
at 280C with a AGt of 13.9 0.2 kcal/mole. Further proof of the
structure of the carbocyclic allene Fp complex was obtained by the use
of spin saturation transfer which allowed additional examination of the
[1,2]Fp shift at low temperatures.
The 13C NMR, at low temperatures, revealed that the carbocyclic
allene Fp complex shows some metallocyclopropyl character, as seen in
the Pt carbocyclic allene complexes, as characterized by the extreme
upfield shifts of the C1 + C2 carbons of the carbocyclic allene.
Contrary to earlier reports on the addition of MeO~ to acyclic
allene Fp complexes, it was found that addition of EtOH occurred on the
C1 carbon of the n2-allene fragment. This generated the sigma complex,
dicarbonyl-n5-cyclopentadienyl-n1-(7-ethoxy-l-cycloheptenyl)iron, in a
A second attempted approach to the carbocyclic allene complex was
through the generation of the cyclopropanonyl-iron sigma complex.
Previously, it was reported that the parent cyclopropanol complex of Fp
would not decarbonylate to the corresponding sigma complex. Contrary
to these earlier reports, photodecarbonylation of the corresponding acyl
complexes gave excellent yields of the desired sigma complexes.
However, attempts to generate the desired alpha-halo sigma complexes
were unsuccessful, thus not allowing entrance into the allene manifold.
Unfortunately, methoxy abstraction from dicarbonyl-n5-cyclopenta-
dienyl-nl-(6-methoxy-l-cyclohexenyl)iron with Me3SiOSO2CF3 did not
afford the carbocyclic allene complex but rearranged products derived
from carbocation intermediates.
The synthesis, bonding, and structure of transition metal complexes
of carbenes have been investigated and reviewed many times over the last
few years.1-4 The bonding within complexes containing hetero atoms or
n-systems which are alpha to the carbene carbon can be considered in
terms of three principal canonical forms (1, 2, 3).1-3 Resonance form 1
shows no interaction of the vacant pz orbital of the methylene carbon
while resonance form 2 represents the vacant pz carbene orbital interact-
ing with the r-system (4, 5)5,6 or the lone pair on the hetero atom
Fp= I (CO)5Cr-
Fp = C5H5Fe(CO)2
(X or Y) of the ligand (6, 7, 8).7-9 Resonance form 3 represents
/OMe /NR2 + /NR2
(CO)5Cr =C (CO)5Cr =C Fp =C NR2
Me Ph NR2
6 7 8
interaction of the vacant pz orbital with the d orbitals of the metal.
The energy difference between the metal d orbital and that of the pz
orbital influences the contribution of resonance form 3. This so called
"back bonding" can be assessed with the use of x-ray crystallographic
and spectroscopic data.
The first transition metal complex of a carbene (5) with a (4n+2)-
i-system (n=O) was reported by Ofele6 in 1969. More recently the
synthesis of a series of aromatic transition metal complexes (4 and 9)
with (4n+2)r-electrons (n=l) was achieved by Allison and co-workers.5
In principle cyclic conjugated ligands such as C7H6 could exist in
either a carbene (10) or allene form (11). In the case of such complexes
M +M +M
as(4)and(9)it is clear from the NMR spectra that the ligand is completed
in the form of the carbene rather than an allene;5 even in the case
where a P(C4Hg)3 derivative was prepared the carbene form was favored.11
This is notable because recent evidence (theoretical and experimental)
has pointed to the probability that in nonmetallic compounds the twisted
allene structure (12) is lower energy than the carbene structure (13).10
To date, evidence for an example of a metal-cycloheptatetraene (11) has
not been found. This indicates that in the cases reported to date,
either the carbene complex (10) is of lower energy than the allene (11)
or the method of synthesis leads to a higher energy but kinetically
The question arises, then, can transition metal complexes of carbo-
cyclic allenes (either conjugated or nonconjugated) be prepared? A
review of the literature reveals that carbocyclic allenes are generally
very reactive and rather unstable molecules.12-17 The smallest stable
carbocyclic allene is 1,2-cyclononadiene (14). If the carbocyclic ring
is decreased in size, the strain within the ring is increased; thus 1,2-
cyclooctadiene (15) can only be observed spectroscopically (at low
temperature) and trapped, whereas 1,2-cycloheptadiene (16)18'19 can be
generated and trapped but has never been observed (spectroscopically).
* An idealized (C=C=C) angle of 1800 and torsional angle of 900 dictates
the requirement for relatively large rings.
On the other hand, in 1972, Visser and Ramakers20 synthesized
platinum complexes (17, 18, 19) of a number of carbocyclic allenes and
succeeded in isolated stable crystalline products for cyclic allenes as
few as seven carbons. Preparation of these carbocyclic allene complexes
(20) was achieved by the generation of the carbocyclic allene (21)
+ 4 PtL2
(dehydrohalogenation of the bromocycloalkenes at low temperatures)20 in
the presence of (22). These complexes were presumably formed by addi-
tion to the vacant coordination site of the metal followed by the loss
of ethylene.21 One iron complex of a carbocyclic allene has also been
reported. In this case, Foxman and co-workers22 were able to prepare
a stable carbocyclic allene complex (23) (93%) by the exchange reaction
of (24) with the stable carbocyclic allene (14); the smaller carbocyclic
Fp + A Fp+
24 14 23
allenes would not work here due to the high temperature needed to gen-
erate the metal allene complexes.
Other carbocyclic allenes (i.e., cyclonona-1,2,6-triene and cyclo-
deca-1,2,6,7-tetraene)2324 are known to form 1:1 complexes with silver
nitrate or with cuprous chloride; however these are not applicable here.
Reflecting back on the interconversion of (10) and (11) one sees
that this is a specific example of a much broader question of the
M +M +M
energetic and valence isomerization of the types of metal complexes
(25), (26), and (27). Here the isomeric form (25) shows a sigma bond
M M M
25 26 27
with virtually no back bonding possible, a sigma bond in (26) with ex-
tensive back bonding and a i-bond in (27).
A review of the literature reveals only a few examples of these
kinds of interconversions. Recently Cohen and co-workers25 found that
hydride abstraction from (28) gave the allene complex (29), presumably
+ + II
28 30 29
via (30). Lennon and co-workers26 found that methoxy abstraction from
(31) also gave the allene (29). However, here the iron may assist in
S -Oe +F
31 32 29
the direct loss of the methoxy so that the sigma complex (31) may go
directly to the allene complex (29), bypassing the cation (32), via
an anti Fp assistance.
To study the interconversion of (25), (26), and (27) ideally one
would like to enter a given manifold at one of the higher energy forms
and then observe its collapse to one of the more stable complexes (i.e.,
(30) to (29), the more stable product).
It was the aim of this research to study the interconversion of Fp
complexes of small carbocyclic valence isomers (33, 34, 35) entering a
given manifold at one of the higher forms and observing its collapse to
33 34 35
a more stable complex. Since the platinum complex of 1,2-cyclohepta-
diene (19) has been trapped, we elected to begin with attempts to
generate the cyclopropylidene and the cycloheptenyl Fp complexes (34)
and (35), respectively, to see if they would collapse to the allene
ATTEMPTS TO GENERATE CYCLOPROPYLIDENE COMPLEXES OF IRON
In principle, the simplest method to generate cyclopropylidene
complexes of Fp, (30), would be simple hydride abstraction from appro-
priately substituted sigma Fp complexes. And, indeed, one example of
such a reaction has been reported using an exotic reagent (36) with
the sigma complex (28).25 However, hydride abstraction does not
F + H ;p -
36 28 30 29
appear to be a generally useful method for preparing this kind of
complex. For instance, Cutler and co-workers27 have found that
attempts to abstract hydride from (28) with trityl did not give the
desired complex (30), but rather the addition product (37).
ICFP Fp C33
We therefore turned our attention to the synthesis of sigma cyclo-
propyl Fp complexes in which the alpha-hydrogen has been replaced by a
better leaving group such as a halogen or a methoxy group.
The first synthetic approach chosen for the synthesis of sigma
bonded alpha-halocyclopropyl Fp complexes was patterned after that of
Culter and co-workers,27 who found that the parent complex could be
prepared by either treatment of cyclopropyl bromide (38) with Fp anion
or by the alkylation of FpBr with cyclopropyllithium (39). By analogy,
either the reaction of (40a) or (40b) with Fp anion or the reaction
(41) with Fpl, could result in the formation of (42a) or (42b).
Unfortunately, when (40a) was added to a suspension of Fp anion (at
00C) (42a) was not observed; only (43), (44), (45) and unreacted
+ H + Fp2
starting material were recovered. Likewise, when (40b) was treated
with Fp anion, under the conditions as above, only (45), (46)
unreacted starting material were found. Formation of (44)
(46) finds analogy in the report of Marten and co-workers,28
found that the reaction of (47) with Fp anion gave the reduced
product (48). Since the addition of the Fp anion to the dihalo com-
pound gave only the reduction products, an attempt was made to
generate (42b) by the reaction of the anion (41) with Fpl. Unfortunately,
again all attempts failed. Isolation of the products gave only (46),
(49), and (50), the latter presumably arising from the reaction of
L Cl + FpC4Hg
unreacted BuLi with FpI.
Since alpha-substituted Fp complexes could not be prepared by
either of the above methods, a third approach was attempted as illus-
R2C- --Fp hv >
It is well known that metal anions readily react with acid
chlorides to generate the corresponding acyl metal complexes which,
in turn, can normally be decarbonylated (either thermally or photo-
chemically) to the corresponding sigma metal complex.29 Unfortunately,
Bruce and co-workers30 had reported that attempts to decarbonylate the
acyl cyclopropyl metal complexes (51a-d) were unsuccessful. These
a) MLx= Mn(CO)5 b) MLx=Re(CO)5
c) MLx= Fe(CO)2Cp d) MLx= Ir(CO)CVPPh3)2
attempts included thermolysis, photolysis and reaction with
(Ph3P)3RhC3133 (known to be an efficient abstractor of carbon
monoxide). Despite their failure, we noted that their attempts were
limited to the parent complexes and suspected that the substituents
on the cyclopropyl rings of the Fp complexes might cause these com-
plexes to behave differently. We therefore undertook the synthesis
of a series of model alpha-hydrogen substituted cyclopropyl Fp com-
plexes to see if effective decarbonylation conditions could be found.34
The synthesis of the desired substituted acyl complexes (52), (53)35
and (54) was found to be straightforward. Treatment of the substi-
tuted cyclopropyl acid chlorides (55), (56) and (57) with Fp anion
(at 00C) gave poor to excellent yields (Table 1) of the desired metal
complexes (52), (53) and (54).
Table 1. Yields from the reactions of Fp with the corresponding
Reaction of yield % Acyl Complex
Fp and Acyl Holide
H EAc EylF
Et H Et
Ph 0 X Ph 0
H -CIH 77% H --Fp
H X Br 15%
C-CI 20 % C--Fp
Since the unsuccessful attempts to photodecarbonylate the parent
acyl Fp complex (51c) had been carried out in petroleum ether (or neat)
an alternate solvent was sought. Recently, Graham and Heinekey36
reported that the photolysis of (58) in acetone (at -780C) gave the
corresponding sigma complex (59) in a good yield (60%). We therefore
carried out our photodecarbonylation in acetone and were gratified to
find that (52) underwent smooth decarbonylation (at 00C; in 2 hours) to
give an excellent yield (97%) of the desired sigma complex (60).35
H 11 H
I/C-Fp hv 0OC H Fp
IH H d. -acetone I H H
Likewise photolysis of (53)35 and (54), under the same conditions, also
gave the alkyl complexes (61) and (62), albeit the latter in only 15%
As a result of these encouraging results, photodecarbonylation of
the parent acyl Fp complex (51c) was also attempted in d6-acetone.
Indeed, photolysis of (51c), at 00C for 10 minutes, led to the forma-
tion of a new metal complex (a new Cp resonance appeared at 6 4.90).
2C-Fp hv O*C FP
H d -acetone H
After photolysis for one hour and 10 minutes the 1H NMR showed nearly
complete loss of starting material. The solution was worked up at
this time to give the desired Fp complex (28) in a 75% yield.
Since our photolyses were conducted in d6-acetone, while those of
Bruce and co-workers30 had been carried out in petroleum ether (or neat),
we desired to photolyze (51c) and (52) in petroleum ether to see if
this led to reduced yields. Indeed, in both cases even though the
photolysis gave the desired sigma complexes (28) and (60) the yields
in each case were significantly lower than in d6-acetone. The photol-
yses were also carried out in d6-benzene. Again each acyl complex
gave the desired sigma complex but in this case yields were between
those observed in d6-acetone and petroleum ether (Table 2).
* The product (28) may have not been detected by Bruce and co-worker30
due to low yield (16%).
Table 2. Yields from photolysis of the corresponding acyl complexes.
Acyl Photolysis Sigma yield %
Complex Solvent Complex
Ph 0 d ocetone Ph 97%
H -Fp d benzene H eP 64%
H H 5 1 0
H Petroleum ether Ph H 51 %
Et 0 Et
H -- Fp d -acetone H EFo 75 %
H d, -acetone H75 %
SC--Fp d,-benzene HF 52 %
H Petroleum ether H H 16 %
Having demonstrated the viability of the decarbonylation step for
the model system we then turned our attention to the preparation of
alpha-halo acyl Fp complexes.
We first attempted to prepare (63) and (64). Unfortunately, in
each case reaction of the alpha-halo acyl halide (65 and 66) with Fp
(63) and (64),
Br 63 Br
CI 64 CI
not in the formation of the desired metal complexes
but rather the reduced complex, (54).
was also found to be the case for the reaction of (67) with
again only the reduced product (52) could be isolated.
The mechanism of the reduction is not known. However, one
reasonable possibility is shown below.
C-CI _Fp_ C=
In this mechanism the Fp anion induces elimination to give the ketene
(68) which further reacts to give the reduced products.
Since it appeared that alpha-halo acyl chlorides could not be used
to prepare the desired carbene precursors, we turned our attention to
complexes with a methoxy group in the alpha position. We were optimis-
tic about this as a viable approach because Cutler and co-workers37
were able to generate the acyl Fp complex (69) by reacting (70) with Fp
anion; no reduction occurred. In turn, we were able to photodecarbony-
late (69) in d6-acetone (at 00C) to give the known sigma complex (71).
II THF II
Fp + CICCH2OMe Fp- C-CH2OMe
Fp- C-CH2OMe --- Fp-CH2OMe
We expect that methoxy abstraction may have real potential for
the preparation of cyclopropylidene Fp complexes. However, a search
of the literature revealed very few alpha-methoxycyclopropyl carboxylic
acids and we expect their synthesis to be quite a challenge. Further-
more, concurrent with this work, we were also exploring ways to enter
the seven membered ring manifold at the allyl cation (33). Since this
33 34 35
was progressing better than the cyclopropylidene investigation, we
turned our attention to this new area.
GENERATION OF THE CARBOCYCLIC ALLENE Fp COMPLEX
In view of our unsuccessful attempts to generate the cyclopropyli-
dene complex (34) alternate approaches into the valence isomer manifold
Fp Fp +
=Fp .. _
33 34 35
(33, 34, 35) were explored. Ideally one would like to enter this mani-
fold at a higher energy form and observe its collapse to a more stable
complex. We therefore undertook an approach aimed at the preparation
of the allyl cation complex (33) to see if it would rearrange to the
carbocyclic allene (35) or even cyclize to the cyclopropylidene complex
As mentioned before, it was reported some years ago that treatment
of the ether complex (31) with HBF4 Et20 gave the stable acyclic allene
complex (29),26 possibly via the allyl cation (32). Applying the same
methodology to the carbocyclic system, preparation of (72) was under-
In principle (72) could be prepared from (73) by employing the
two approaches discussed in Chapter 2 [i.e., either by the alkylation
of Fpl or chlorocarbonylation of (73) followed by conversion to the
sigma complex (72)]. Thus, the synthesis rested on the preparation
of (75) (using the logic shown in the following scheme).
Recently, Balci and Jones38 prepared the alcohol (76) by treatment
r acetone/H90 ,
r AgClO4 A 4
of (40a) with AgCO14 in the presence of H20/acetone in an 80% yield.
With (76) in hand we found that treatment with NaH followed by Mel gave
1) NaH/THF %
the desired ether (75) in a 41% yield. Later a more direct and higher
yield preparation of (75) was achieved by treatment of (40a) with AgNO3
* AgClO4 in MeOH gives a higher yield (97%) but the reaction mixture
The first approach to the synthesis of the carbocyclic sigma
complex (72) is shown below:
Unfortunately, although treatment of (75) with n-Buli, at -780C, gave
a yellowish solution that was apparently (73), treatment of this
solution with FpI at -78C did not give the desired sigma complex
(72), but rather (77) in a 55% yield.
Since the above approach to the desired sigma complex (72) failed,
a second approach was attempted as seen in Scheme 1.
Treatment of (75) with n-BuLi at -110C followed by addition of
CO2 and acid work-up gave the acid (78) in a 66% yield. The acid (78)
was then treated with SOC12 (in benzene) to give the acid chloride (79)
in a 70% yield. Addition of the acid chloride (79) to a suspension of
Fp anion in THF at -780C gave a 27% yield of the desired acyl complex
(74) as an air-sensitive, brownish oil. The 1H NMR spectrum of (74)
is given in Figure 1. Photodecarbonylation of (74) in d6-acetone at
00C gave an air-sensitive, brownish oil (72) in a 53% yield. The 1H
NMR spectrum of (72) is given in Figure 2.
At this point the question arises, will (72), when treated with
HBF4.Et20, behave similarly to the model system (31) as reported by
Lennon and co-workers?26 It was found that treatment of (72) with
O J --
... ..... ... ... r-PfOliii'- l Cz L
HBF4*Et20 at -78C gave an immediate precipitate which, from the low
field Cp resonances (6 -5.65) in the 1H NMR spectrum (Figure 3)
suggested the formation of r-complexes of Fp. This downfield shift
is consistent with known r-complexes of Fp as indicated in Table 3.
It is well known that n-complexes of Fp undergo facile displacement
of olefinic ligand by halide ions, usually I whereas sigma complexes
of Fp are unreactive.3941 The precipitate from above was therefore
treated with Nal in d6-acetone to immediately give a color change
which upon work-up revealed that Fpl had been produced. Further exami-
nation, glpc, revealed that the ligands that had been displaced were
cycloheptanone (80) or its precursor, cycloheptene (81),and trace
amounts of 1,3-cycloheptadiene (82) by GC/MS.
Table 3. H NMR resonances of the Cp of r-complexes.
Cp (in ppm's)
One possible mechanism to rationalize these products is seen in
Addition of HBF4*Et20 to (72) could initially generate two dif-
ferent intermediates (33) and (83). Simple protonation of the double
bond of (72) would generate the carbocyclic carbene (83) which could
rearrange to give the i-complex (84). A closer examination of the 1H
NMR spectrum of the crude protonation product (Figure 3) reveals a
singlet at 3.80 ppm, a triplet at 4.30 and a singlet at 5.70 which is
a good indication of structure (84) since (85) shows similar resonances
at 4.03 ppm, 3.06 and 5.52. It should be noted that the resonances at
C34.30 6 5.08
MeO OMe H CH
-Fp 5.52 Fp 5.70 Fp Fp
H e- 6 3.06 e6 4.3 6 3.75 H
85 84 86 87
3.06 ppm in (85) and 4.30 ppm in (84) are somewhat different, but this
difference is not unexpected since alkyl substitution in (86) vs. the
parent complex (87)42 causes a comparable difference in the chemical
shifts of the protons.
The second proposed reaction is protonation of the methoxy group
to give the allyl cation (33) which abstracts a hydride to generate
* Possible hydride donors include Et20, CH30H (unlikely) or starting
the sigma complex (88). This, in turn, could react with any acid
present to give the carbocyclic carbene (89) which should rearrange
to the 7-complex (90). This is reasonable since Casey and co-workers43
recently found that treatment of (91) with HBF4*Et20 gave the i-complex
- ~- O 92
(92), via the carbene (93). The latter was trapped with P(OMe)3to give
(94). Further evidence for the proposed conversion of (88) to (90) was
obtained by treating the carbocyclic sigma complex (95) with HBF4.Et20,
at -780C, to afford (96) in a 96% yield. Evidence for a carbene
intermediate (97) was obtained by protonation of (95) in the presence
of P(OMe)3 which led to the isolation of (98). Finally, further proof
H+ ) K Fp P(OMe)3
95 97 98
for the formation of (90) comes from the fact that the H NMR spectrum
(Figure 3) reveals a singlet at 6 5.80 (Cp) which can be overlapped
with the known carbocyclic r-complex (90) as seen in Figure 4.
As mentioned above, treatment of (84) and (90) with excess
Nal in d6-acetone should release the corresponding carbocyclic
compounds (99) and (81), respectively. Although it seems unlikely
Fp+ Fp+ Nal
84 90 99 81
that (99) could result from reaction with I- to generate
(80), it is feasible that in the presence of Fpl this
could be possible. Further proof of this can been seen in
the treatment of the known r-complex (86)44awith excess Nal in
d6-acetone to give Fpl and acetaldehyde.44b
Fpl + CH3CHO
Regardless of the mechanism of formation of these products the
desired allene complex (35) was not observed although some reaction
appeared to be taking place on the double bond. Therefore, a search
for a methoxy abstractor which would not react with the double bond
Recently Brookhart and Tucker45 reported that treatment of (100)
with trimethylsilyl trifluoromethanesulfone, Me3SiOS02CF3, (101)
afforded the carbene complex (102) by simple methoxy abstraction.
Fp '-(H Me3SiOSO2CF3 F
Accordingly, (72) was treated with Me3SiOSO2CF3 at -780C to give an
immediate precipitation which when warmed to room temperature followed
by removal of the solvent afforded a brownish oil (Figure 5). When ex-
cess Nal was added to a d6-acetone solution of this oil an immediate
color change, brown to purple, occurred (Figure 6). Work-up afforded
FpI, cycloheptene (81), 1,3-cycloheptadiene (82) and trace amounts of
cycloheptanone (80) by GC/MS. A possible mechanism for formation of
these products is seen in Scheme 3.
In this scheme addition of Me3SiOSO2CF3 gives the allyl cation
(33) which could rearrange to (90) as the solution is warmed as seen
in Scheme 2. In a competing reaction (33) could lose H+ to give (103).
This in turn could then react with H+ to give the more stable carbene
* The Fp may not be able to assist in the loss of the methoxy group
because this would force the leaving group into the center of the
- S -
^^^. \ -N
^^^ S U
-^^^^ /* -
\ ~ ^
complex (104) followed by rearrangement to give (105). Here again reac-
tion of (90) and (105) with Nal should give FpI, cycloheptene (81) and
1,3-cycloheptadiene (82), as observed.
Since the treatment of (72) with Me3SiOSO2CF3 did not give rise to
any cycloheptanone (80) or any organic compounds with oxygen present,
it appears that indeed the methoxy was clearly abstracted. The crucial
questions are whether the precursor to (90) and (105) is the allyl
cation (33) or the allene complex (35) and the structure of the pre-
cipitate initially formed. To address the latter, we undertook a low
temperature 1H NMR study of this material.
To prepare for this study, a search for suitable models revealed
that Olah and Liang46 had reported the preparation and properties of
(106a-c) in superacid. It was found that in the 1H NMR spectrum, the
3 R R
106(a-c) a=H b=CH3 c=C1
parent cation (106a) showed a doublet at 10.18 ppm (H1 + H3) and a
triplet at 8.34 (H2). Substituted allyl cations show vinyl resonances
slightly upfield from these values as seen in Table 4.
Foxman and co-workers22 reported the synthesis of (23) as a stable
crystalline solid whose 1H NMR revealed resonance at 4.45 ppm (H1),
6.50 (H3) and 5.80 (Cp), which we used as a model for the allene
Table 4. 1H NMR resonances for (106a-c)
R H1 H2 H3
a) H 6 10.18 6 8.34 6 10.18
b) Me 6 8.05 6 9.50
c) Cl 6 8.05 6 9.58
H Fp H
From these models clearly the carbocyclic allene complex (35) or
the carbocyclic allyl cation (33) should have very different and
characteristic H NMR spectra which should make it easy to distinguish
between them. A CD2C12 solution of the methoxy complex (72) was treated
with a slight excess of Me3SiOSO2CF3 at -780C in an NMR tube and the 1H
NMR spectrum recorded. This revealed a singlet at 5.70 ppm (Cp) and
two multiplets at 6.50 (H3) and 4.2 (H1) (Figure 7 ) which corresponds
Fp 2 Fp 2
* No resonances were observed from 7 to 15 ppm and 0 to -13 ppm.
closely with (23) (Figure 8) as reported by Foxman and co-workers.22
Further evidence for the allene structure (35) would be the observation
of fluxional behavior22'47 as the temperature is changed. And indeed,
as the sample was slowly warmed from 00C to 400C, a coalescence of the
multiplets at 6.5 ppm (H3) and 4.2 ppm (H1) was observed (Figure 9)
occurring somewhere between 200 and 400C. After the sample was warmed
to 400C, it could be cooled back to 00C to reveal the multiplets at 6.5
ppm and 4.2 again. -As expected from earlier work, if the sample is
kept at room temperature in solution it slowly decomposes.
A characteristic property of Fp acyclic allene complexes is their
reactions with nucleophiles such as alcohols. We therefore decided to
examine the reaction of (35) with ethanol. Interest in nucleophilic
reactions with acyclic Fp allenes ran rather high about 10 years ago.
Lichtenberg and Wojcicki48 reported that treatment of (107) with most
* (35) Is stable for at least two hours at 0OC.
o 0 0
nucleophiles afforded attack at the terminal carbon with migration of
the metal to the central carbon. One exception was MeO- which prefer-
entially or exclusively attacks at the central carbon (Figure 10).
We found that the reaction of (72) was Me3SiOSO2CF3 at -78C
followed by the addition of a cold (00C) mixture of EtOH/Na2CO3
afforded (108)t in a 54% yield. To prove the structure assigned to
(108) an alternate synthesis was accomplished as seen in Scheme 4.
* Non completed allenes are inert to attack by nucleophiles.
( Addition of the alcohol either two minutes or 2.5 hours after the
initial addition gave the same product.
t This result is not only consistent with addition to the allene
complex but possibly to the allyl cation complex too.
Figure 10. The addition of Nu- to (107).
In this sequence, treatment of (40a) with AgNO3 in EtOH gave (109)
in a 75% yield. Treatment of (109) with n-BuLi at -780C followed by
addition of CO2 and acid work-up gave (110) in a 65% yield. To the
acid (110) was added SOC12 in benzene to give the acid chloride (111).
Addition of the acid chloride (111) to a suspension of Fp anion in THF
at -780C gave the desired acyl complex (112). The acyl complex (112)
was photolyzed for 2 hours at 00C in acetone to give (108) as a brown-
ish, air-sensitive oil, 28% from the acid chloride. The 1H NMR spectrum
of (108) is given in Figure 11, which is identical with that obtained
from the allene complex.
Up to this point, evidence for the carbocyclic allene structure
(35) has been based on the crude 1H NMR, the observation of its flux-
ional behavior and its reaction with EtOH. It would be desirable to
release the allene ligand and detect it as its dimer (113). A solution
of (72) in CD2C12 was therefore treated with Me3SiOSO2CF3 followed by
the addition of Me4N I-, which is soluble in CH2C12. Although Fpl was
formed the dimer (113) was not observed.
On searching the literature it was found that this result is not
inconsistent with the allene structure because Lichtenberg and
Wojcicki49 found that treatment of (107) with +NR41- gave Fpl as the
major product; however no free allene was observed. An unstable
material (114) was characterized by IR which was believed to
S Fpl + ?
spontaneously decompose to Fpl and a tarrish material at room tem-
perature. It has also been reported that (14) was released from (23)
with Fp anion.22 We therefore treated (35) with Fp anion in THF at
-780C but again no trace of the dimer (113) could be detected even
though Fp2 was isolated.
* A similar decomposition has been noted by Lukas and co-workers50 for
a Pt complex.
Since our attempts to release the carbocyclic allene were un-
successful, an attempt to isolate and further characterize the carbo-
cyclic allene complex (35) was next approached.
One of the major problems isolating (35) was its instability in
solution at room temperature; however, if the solution was kept at 00C
the allene complex seemed to be somewhat stable for at least two hours.
Therefore, after the initial generation of (35) at -780C the reaction
mixture was warmed to 00C at which time the solvent was removed in
vacuo. This afforded a tan solid (71%) which gave the 1H NMR spectrum
shown in Figure 12. This material,which is believed to be the allene
complex (35), appears to be stable indefinitely in the solid phase at
-780C under a N2 atmosphere; in fact it was sufficiently stable to
obtain a correct C,H analysis.
To further prove that indeed the new complex (35) had the allene
structure a 13C NMR spectrum (Figure 13) was taken at -200C. In this
spectrum the resonance for the allene carbons C1-C3 appeared at 150 ppm
(C2), 125.8 (C3) and 43.6 (C1). These compare quite well with the
resonances for the 1,2-cyclononadiene complex (23) which shows similar
resonances for carbons (C1-C3) at 148.5 ppm (C2), 122.3 (C3) and 51.3
(C1) as seen in Figure 14. This is strong evidence for the structural
similarity of complex (23) and (35).
+ Fp 2
Fp 2 1
------ \ ---
i- I I
In both cases the upfield shift of the C1 and C2 carbons as
compared with uncomplexed allenes suggests significant metallocyclo-
propyl character as depicted in (115) and (116). Furthermore, in
(35) the C1 resonance is ca. 8 ppm upfield from the C1 resonance in
(23) which indicates slightly more cyclopropyl character. This is
not surprising since the geometrical constraints opposing incorpora-
tion of the cyclopropyl ring in a structure such as (118) are less
than for a double bond as in (16). This relief of strain can be seen
best in the fact that (118)12,51,52 can be isolated whereas the more
strained allene (16) has been generated at low temperatures, but has
not been observed as the free molecule.17
At this point since the pure carbocyclic allene complex (35) was
in hand a more detailed examination of its fluxional behavior was
attempted. As discussed earlier the low temperature, at -39C, 1H NMR
spectrum of (35) exhibits two multiplets at 6.55 ppm (Ha) and 4.45
(Hb). As the temperature was raised the point of coalescence was ob-
served at 290C as seen in Figure 15. Here exchange of Ha and Hb became
rapid on the NMR time scale. Using Equation 1,53 the free energy of
activation (AGt) was obtained. In this equation, Tc represents the
AGt = R Tc(loge Tc/6v + 22.96)
temperature of coalescence and Sv is the distance between Hb and Ha in
Hz. The barrier for interconversion of isomeric allene complexes
(35a 35b) was calculated to be 13.9 0.2 kcal/mole with 6v equal to
215.6 Hz at -78C. This result corresponds to the activation barrier
E oO 0
I m m fcW a n Co 0)
I C C'3CUCUCU t- m
of known acyclic allenes22 of 23.1, 18.0 and 16.3 kcal/mole for (107a),
(119) and (120), respectively.
Further support for the structure of (35) comes from an extremely
useful NMR technique known as spin saturation transfer,54-57 an NMR
technique which can be applied to systems that are in rapid equili-
brium. Since Ha and Hb havedifferent chemical shifts (6 6.50 and 4.45,
respectively; Figure 12) then either Ha or Hb could be specifically
saturated by applying a strong rf field at the appropriate frequency.
If the exchange rate between Ha and Hb is comparable to the relaxation
rate at both sites, then complete saturation of one resonance will
result in partial or complete saturation of the other resonance. If
the exchange process becomes sufficiently slow at low temperatures, then
saturation at Hb (or Ha) should result in no apparent saturation of Ha
(or Hb) since the relaxation process should take place before exchange
occurs. Conversely, if the exchange process is rapid at higher tempera-
tures then saturation of Hb should result in partial or complete satura-
tion of Ha since the relaxation process does not take place before ex-
Accordingly, the 1H NMR spectrum of (35) reveals a multiple at
6 6.55 (Ha) and 4.45 (Hb) as seen in Figure 16a. When a strong rf field
was applied at Ha (6 6.55 at -20C) the exchange process between Hb and
Ha apparently takes place without relaxation occurring in the transfer
process (Figure 16b) because the intensity of Hb was observed to de-
crease. However, at -60C the exchange process is significantly slower
than relaxation so complete relaxation occurs before exchange, Hb to Ha.
As a result, the intensity of Hb did not diminish when Ha was saturated.
At -400C, saturation of Ha results in partial saturation of Hb,(Figure
17b) thus resulting in an exchange process somewhat slower than that at
-200C.(Figure 16b) but faster than the exchange at -600C,(Figure 17c).
This is feasible for (35) since one would expect at high temperatures
(in this case -200C) a fast [1,2]Fp shift allowing Ha and Hb to exchange
-^v ---.- 35-
/ i .-
J -^ j |
\ ^\ -f3 i
"' ? ^
u ja 1
whereas at lower temperatures the exchange is slowed down so that the
exchange between Hb and Ha is shown according to the NMR time scale.
At this point, we decided to reexamine some unanswered questions.
As mentioned before, addition of Nal to a crude reaction mixture of
(35) did not give the dimer (113) although Fpl was generated. Since the
allene complex (35) could now be obtained as a pure solid we decided to
reexamine this reaction.
A sample of the solid allene complex (35) was dissolved in a solu-
tion of Nal in d6-acetone at -78C. This mixture was then allowed to
stir for 2 hours at -780C followed by 4 hours at room temperature. An
examination of the 1H NMR spectrum (Figure 18) again revealed the pre-
sence of Fpl. A closer examination of Figure 18 also reveals the pre-
sence of cycloheptene (81) and cycloheptadiene (82). Analysis of the
mixture showed the amounts of (81) and (82) to be 20% and 49%,
respectively. From this result we conclude that the complex rearranges
to the corresponding r-complexes (90) and (105) faster than the allene
is released. The i-complexes then immediately react with Nal to give
the observed olefins (81) and (82).
In an attempt to determine if the precursor to the observed U-
complexes (90) and (105) is the allyl cation (33), we decided to see if
the addition of Ph3CH, a good hydride donor, would change the relative
amounts of (81) and (83). An excess amount of Ph3CHwas therefore added
to a solution of (35) in CH2C12 at -78C followed by warming for 3 hours
at room temperature. This gave a brownish oil after the removal of the
solvent in vacuo. The 1H NMR spectrum of this oil is seen in Figure 19.
Addition of Nal (Figure 20) followed by work-up gave Fpl, cycloheptene
(34%) and cycloheptadiene (51%). Decomposition under identical
^^-^S:^ '' .ol
^\ ~ m S
1 \ n e
____ ^----- Ic
----------- i- r -
conditions in the absence of Ph3CH gave 1,3-cycloheptadiene (40%) and
a trace amount of cycloheptene, from this it appears that the presence
of a hydride donor increases the yield of cycloheptene (81) suggesting
that indeed the cycloheptene originates from a hydride abstraction
The stability of the allene complex (35) is also of some import-
ance. While it was found that the complex (35) is stable under a N2
atmosphere as a solid at -780C for at least 6 weeks, exposure to the
air at ambient temperatures for 30 minutes affords a brownish oil.
The 1H NMR spectrum (Figure 21) of this oil in d6-acetone reveals a new
Cp resonance at 5.7 ppm. Furthermore, addition of Nal to this solution
followed by work-up afforded Fpl, 83% of cycloheptanone (80) and trace
amounts of cycloheptene (81) and 1,3-cycloheptadiene (82) by GC/MS.
One possible mechanism to rationalize this product is shown in
In this scheme the allene complex (35) reacts with H20 in the air
to afford the sigma complex (121) which can further react with H+ to
generate the carbocyclic carbene (122). This could then rearrange to
give the i-complex (123). Although formation of r-complexes of vinyl
alcohols58-60 are known, i.e. (124), it has not been reported whether
or not Nal would release the n-complex to give the aldehyde or ketone.
Since the vinyl ether complex (86) is released by I- one would expect
the same to occur here.
Since it is clear that the carbocyclic allene complex (35) has
been generated we decided to explore whether or not the carbocyclic
allene complex (125) could be prepared. Therefore, the synthesis of the
sigma complex (126) was undertaken. The approach here is similar to
the approach used in the synthesis of (72) and is shown in Scheme 6.
6-Methoxy-l-bromocyclohexene (127) was prepared by the treatment
1) NaH/THF 00C
2) Mel 00C
of (128) with NaH followed by Mel. Treatment of (127) with n-BuLi at
-780C followed by addition of CO2 and acid work-up gave the acid (129)
1) n-BuLi THF -78C
21 CO- -780C
in a 33% yield.
The acid (129) was then treated with SOC12 in benzene
to give the acid chloride (130). Addition of the acid chloride (130)
- 0 COCi
to a suspension of Fp anion in THF at -78C gave an 8% yield of the
desired acyl complex (131) as an air-sensitive, brownish oil. The 1H
NMR spectrum of (131) is given in Figure 22. Photodecarbonylation of
(131) in d6- acetone at 00C gave an air-sensitive, brown oil (126) in a
70% yield. The 1H NMR spectrum of (126) is given in Figure 23.
Now with (126) in hand the question of whether or not it would
behave similarly to the reaction of (72) with Me SiOSO2CF3 remained to
Here it was found that treatment of (126) with Me3SiOSO2CF3 at
-780C followed by warming to room temperature afforded a brownish oil
(Figure 24) after removal of the solvent in vacuo. When Nal was added
to a d6-acetone solution of this oil an immediate color change, brown
to purple, occurred. Work-up afforded Fpl, 30% of cyclohexene (132)
and 10% of 1,3-cyclohexadiene (133). A possible mechanism for forma-
tion of these products, which is similar to the C7 manifold, is seen
in Scheme 7.
In this mechanism addition of Me3SiOSO2CF3 to (126) gives the
allyl cation (134) which immediately hydride abstracts and rearranges
to the ir-complexes (96) and (137) via the same intermediate as seen in
the C7 manifold.
---------- -- -- -C 4- -
A further examination of the 1H NMR spectrum (Figure 24) reveals
resonances at 6 5.7 (Cp) and 5.45 (CH=CH) which is a good indication
of structure (96) since the resonances overlap with the known T-complex
(96) (Figure 25). The ir-complex (137) was not isolated; however, as
mentioned above addition of Nal to the reaction mixture (Figure 24)
afforded Fpl, cyclohexene, and 1,3-cyclohexadiene which is consistent
with previous work.
Although the products obtained here were similar to the products
obtained by the treatment of (72) with Me3SiOSO2CF3, we decided to see
if addition of Me3SiOSO2CF3 to (126) followed by addition of EtOH/Na2CO3
1) Me3SiOSO2CF3 -780C
2) EtOH/Na2CO3 -78C
would generate (138). In this case, and unlike (72), addition of
Me3SiOS2CF3 to (126) did not give (138) but rather (95) in a 15% yield.
1) Me3SiOSO2CF3 -780C
2) EtOH/Na2CO3 -78C
i o k
S -o -o -- O *- -. -
- ... ..
-- .. E I
-g_-.-- s------------------------- w
This is, of course, the proposed cyclohexene precursor in Scheme 7.
From this result it would appear that in the six member ring (126)
very rapidly gives (95) which is protonated in a slow reaction. Thus,
addition of base stopped the reaction at this point. To further prove
the structure (95) an alternate synthesis was accomplished as seen
COC1 COFp Fp
Fp THF hv O C
139 140 95
Addition of the acid chloride (139) to a suspension of Fp anion
at -780C gave the acyl complex (140) in a 69% yield. The acyl complex
(140) was then photolyzed for 2.5 hours at 00C in d6-acetone to give a
yellow, air-sensitive solid (95) in a 97% yield, which is consistent
with the above work.
From these experiments it would appear that the allyl cation (134)
is generated; however, if it rearranges to the allene complex (125) as
seen in the C7 manifold, the latter must very rapidly convert back to
the cation to give the observed products.
In conclusion, the characterization and synthesis of the seven
membered carbocyclic allene complex (35) has been described within this
work. Further, it appears that this allene complex rather readily
converts to the cation to give the carbocation products. As expected if