Scope and limitations of the 1,2 rearrangement of alkyl groups from carbon to metals to generate carbene complexes

MISSING IMAGE

Material Information

Title:
Scope and limitations of the 1,2 rearrangement of alkyl groups from carbon to metals to generate carbene complexes
Physical Description:
vii, 179 leaves : ill. ; 28 cm.
Language:
English
Creator:
Hanna, Paul K., 1956-
Publication Date:

Subjects

Subjects / Keywords:
Rearrangements (Chemistry)   ( lcsh )
Organometallic compounds   ( lcsh )
Carbenes (Methylene compounds)   ( lcsh )
Rhenium   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (PH. D.)--University of Florida, 1988.
Bibliography:
Includes bibliographical references.
Statement of Responsibility:
by Paul K. Hanna.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001125904
oclc - 20106567
notis - AFM3027
sobekcm - AA00004806_00001
System ID:
AA00004806:00001

Full Text










SCOPE AND LIMITATIONS OF THE 1,2 REARRANGEMENT
OF ALKYL GROUPS FROM CARBON TO METALS
TO GENERATE CARBENE COMPLEXES














By

PAUL K. HANNA


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


UNIVERSITY OF FLORIDA

1988


OJ F LIBRARIES















ACKNOWLEDGEMENTS


I would like to thank William M. Jones for his

patience and help throughout the years.

I would like to thank the people in the lab for their

help and encouragement, especially Yngve Stenstrom,

Su-Min Oon, Wayne Chandler, Nick Conti, Donna Crowther,

Rhonda Trace, Jasson Patton and Laura Quinn.

I would like to thank those people outside the lab

for reminding me (often!) that there is always a light at

the end of the tunnel. There are too many of these

people to mention, but Ted Streleski is not the least of

them.

Finally, a very special thanks to Jeanne Pittari,

Paul Hanna, and Mary Hanna; I couldn't have done it

without them.



















TABLE OF CONTENTS


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

ABSTRACT.................................................vi

CHAPTERS

1 INTRODUCTION....................................... 1

2 RHENIUM............................................. 9

3 IRON.............................................. 104

4 CONCLUSION....................................129

5 EXPERIMENTAL.............................. ... 132

General Procedures............................132

Preparation of 1-Methoxy-l-
thiophenylcyclopropane(51)..................133

Preparation of 1-Methoxycyclopropane-
carboxylic acid(53) ......................... 134

Preparation of 1-Methoxycyclopropane-
carbonylchloride(47) ........................135

Preparation of Pentacarbonyl(1-methoxy-
cyclopropyl-1-carbonyl)rhenium(44)............136

Photolysis of Pentacarbonyl(1-methoxy-
cyclopropyl-1-carbonyl)rhenium (44)..........137


iii








Preparation of 1-Bromo-l-ethoxy-
cyclopropane(65)............................139

Preparation of 1-Ethoxy-l-trimethyl-
siloxycyclopropane(70)......................140

Preparation of Bromotetracarbonyl-
(triphenylphosphine)rhenium(75)...............141

Preparation of Methyltetracarbonyl-
(triphenylphosphine)rhenium(76).............142

Preparation of (1-Ethoxycyclopropyl)-
phenylmethanol(77)..........................143

Preparation of 1-Methoxycyclobutane-
carboxylic acid(86)..........................144

Preparation of Pentacarbonyl(l-methoxy-
cyclobutyl-1-carbonyl)rhenium(78)............146

Photolysis of Pentacarbonyl(1-methoxy-
cyclobutyl-1-carbonyl)rhenium(78)...........147

Preparation of cis-Triethylphosphine-
tetracarbonyl-1-carbonyl(1-methoxy-
cyclobutyl)rhenium(107).....................150

Photolysis of cis-Triethylphosphine-
tetracarbonyl-1-carbonyl(1-methoxy-
cyclobutyl)rhenium(107).....................151

Preparation of Cyclopropylcarbinol(126)...... 153

Preparation of Cyclobutanol(127).............154

Preparation of Cyclobutyl Tosylate(128)......155

Preparation of Cyclobutene Epoxide(123)...... 156

Preparation of 7-Bromobenzo-
cyclobutene(140)............................158

Preparation of 7-Cyanobenzo-
cyclobutene(141)............................159

Preparation of Benzocyclobutene-7-
carboxylic Acid(142).........................160









Preparation of Benzocyclobutene-7-
carbonyl chloride(143) ........................161

Preparation of Benzocyclobutene-7-
carbonylrheniumpentacarbonyl(135)............161

Preparation of benzocyclobutene-7-
rheniumpentacarbonyl(136).....................163

Preparation of Benzocyclobutene-7-
methoxy-7-rheniumpentacarbonyl(130)..........164

Thermolysis of Benzocyclobutene-7-
methoxy-7-rheniumpentacarbonyl(130) ..........166

Preparation of Dicarbonyl-(n5-cyclopenta-
dienyl)(propyl)iron(162) ....................167

Preparation of Dicarbonyl-(n5-cyclopenta-
dienyl)- 2-propenyliron tetra-
fluoroborate(163).............................168

Preparation of Dicarbonyl(n5-cyclopenta-
dienyl)[2-(N,N-dimethylamino)prop
-1-yl]iron(165)...............................169

Photolysis of Dicarbonyl(n5-cyclopenta-
dienyl)[2-(N,N-dimethylamino)prop-l-
yl]iron(165). Isolation and Character-
ization of Carbene Complex 169 and Chelate
Complexes 170.................................171


REFERENCES.................................................175

BIOGRAPHICAL SKETCH................................... 179















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

SCOPE AND LIMITATIONS OF THE 1,2 REARRANGEMENT
OF ALKYL GROUPS FROM CARBON TO METALS
TO GENERATE CARBENE COMPLEXES

By

Paul K. Hanna

December 1988

Chairman: William M. Jones
Major Department: Chemistry


This research investigates the generality of the

rearrangement of alkyl groups from carbon to metal. This

rearrangement had previously been observed only when the

metal was iron and the alkyl group was part of a strained

ring. In this research, systems were synthesized which

contained rhenium instead of iron and an acyclic alkyl

group instead of a strained ring.

The first system synthesized was pentacarbonyl(l-

methoxycyclopropyl-l-carbonyl)rhenium (44). Thermolysis

of this compound in toluene resulted in a complex mixture

which did not contain the desired rearranged product.

Photolysis of 44 in benzene resulted in formation of

vi








Re2(CO)10 and the dimer of the organic acyl fragment,

compound 54. The analogous pentacarbonyl(1-methoxy-

cyclobutyl-1-carbonyl)rhenium (78) was synthesized.

Thermolysis in toluene resulted in a complex mixture

which did not contain the desired rearranged product.

The photolysis of 78 was studied mechanistically. It was

shown to be a radical reaction involving metal hydrides

which generates methoxycyclobutane (88) and 1-

methoxycyclobutene (89) in a 1:1 ratio at room

temperature in benzene. Photolysis at -650C forms 1-

methoxycyclobutanecarboxaldehyde (115) in addition.

The pentacarbonyl(7-methoxybenzocyclobutene)rhenium

system (130) was synthesized through the intermediacy of

pentacarbonyl(7-benzocyclobutenylidene)rhenium tetra-

fluoroborate (137). This is the first known cationic

rhenium pentacarbonyl carbene complex. Thermolysis of

130 resulted in formation of benzocyclobutenone and

pentacarbonyl(7-benzocyclobutene)rhenium (136) with an

activation energy of 21.1 kcal/mol and an entropy of

activation of -2.1 cal/mol-K. The acyclic iron system

dicarbonyl-(n5-cyclopentadienyl)[2-(N,N-dimethylamino)-

prop-1-yl]iron (165) was synthesized and photolyzed. The

desired rearranged product (169) was isolated as

indicated by spectroscopy and chemical reactivity.


vii












CHAPTER 1
INTRODUCTION

The area of research into the chemistry of

organotransition metal carbene complexes was given a

boost when C. P. Casey and T. J. Burkhardt showed for the

first time in 1974 that such a compound (1) could induce

metathesis.1 This was significant in that it allowed the

interconversion of alkenes. One could start with

relatively simple alkenes and transform them into more

complex ones in a single step.




CeHs CH2 OCH3 CH2
(CO), + (CO),W==< +
6Hs C6Hs OCH3 C6H5 CHs CHs
1 1 1 4


The splitting up of alkenes and their recombination is

both interesting from a mechanistic point of view and

practical in that a variety of alkenes may be synthesized

easily such as 1,1 diphenylethylene (4) in the example

above.

During the course of a theoretical study on these

carbene complexes H. Berke and R. Hoffmann2 investigated

the 1,2 migration of a carbon from a carbon to a metal

1








2

resulting in the formation of a carbene complex. The

reaction they specifically studied was the manganese

tetracarbonyl ethyl complex (5) rearranging to the methyl

manganese tetracarbonyl methylidene complex (6). The

calculations predicted that the energy of activation

should be essentially the same as the energy difference

between the two molecules. This indicates that the

activation energy for the rearrangement is nearly zero.

CO CH3 CO
/I /
OC- Mn--CH2-CH3 OC-Mn- =CH2
oc Co oc Co

5 6

In contrast to this predicted facile rearrangement

there was until recently only one reported case in which

there was any evidence that an alkyl group migrated from

a carbon to a metal to generate a carbene complex.3 This

was a case in which the metal was Ni and the presumed

carbene complex (8) was unstable, its existence was

inferred from the resulting chemistry which involved the

trapping of the intermediate carbene (8) with cyclohexene

to generate the norcarane (9).














DD D D

(PPH3)2Ni (PPH3N j

D D CD2


D
0 r>D


There are a great many processes which involve an

analogous rearrangement, the migration of hydrogen to the

metal center instead of carbon.4,5,6


+ CD3
Cp2W
11


PMe2Ph
S PMe2Ph
Cp2W
NCD3


D
+/
Cp2W
12 NCD2


PMezPh
D
CP2W
SCD2-PMe2Ph


(PPh3)3M -CH3 D2


CD4 + CD3H + CD2H2 + CDH3 + CH4


M = Co. Rh

There are, however, many cases known in which an

alkyl group on a metal center migrates to a carbenic


CD3



10


-C2H4












carbon which is on the same metal. This results in the

destruction of the carbene center and the formation of a

new C-C bond.7,8'9


CH2C6H5
;/




CH3CN

+ CH2CsH5
CP2K
NCCH3
12


CH3PMe
+I / 3
Br-I r = CH2
Me3P PMe3

20


PMe3
Br-I .-CH2-CH3
Me3P PMe3

21


CH3

Cp2W

C6Hs


-H
CH3CN


CH2


~C6H5













CH H3 C2 C2H3
Cp Cp2W
CP2 CP2 CP2W
22 CH3 23 CH3 24





CP2W
H
25
All of these facts point to the conclusion that the

predicted facile equilibrium of Berke and Hoffmann2 is

not an equilibrium at all for carbon systems. The

rearrangement normally goes in the direction that leads

to the destruction of the carbene center. The reason for

this "one way equilibrium" has been postulated by N. J.

Cooper and M. L. H. Greene to be one of thermodynamics.3

In an organometallic system in which there is a vacant

coordination site on the metal such a migration is often

possible. When the migration of an alkyl group from a

carbon to a metal center occurs the net result is the

loss of a C-C bond, the formation of a M-C bond, and the

conversion of a M-C bond into a M=C bond. This change

was thought to be thermodynamically unfavorable. The

molecule would invariably find some other, more

energetically favorable, pathway to fill the vacant

coordination site. It is important to note that if this








6

theory is correct, the rearrangement of an alkyl group

from a carbon to a metal generating a carbene complex may

be brought about if the system is perturbed so that the

rearrangement is energetically favorable.

Chemists have long known that the way to shift the

position of an equilibrium is to change the thermodynamic

preference on one side relative to the other. One way to

do this is to decrease the energy of the less favored

side. The other is to increase the energy of the more

favored side.










Energy




While working towards a completely different goal,

Jim Lisko serendipitously employed both of these

techniques and discovered that the rearrangement of alkyl

groups from carbon to metal centers could not only be

made to occur, but to give a stable carbene complex

containing an organic ligand bonded to the metal

center.10

















S 0 light S r OCH3 /i e/OCH3
light light Fe
Co OCH3 light CO.
Co Co cO. I
CO C

26 21




The organic ligand in the starting material contained a

cyclopropane ring (27), which expanded to a 5 membered

ring in the product (28). The ring strain of the

cyclopropane in effect raised the energy of the starting

material (27). A 5 membered ring,as is in the product

(28), is relatively unstrained and does not affect the

relative energies.11 The product (28) contained an

electron donating methoxy group on the carbene carbon.

These are known to stabilize12 carbene complexes of iron.

This, in effect, lowered the energy of the product. The

result of combining these two factors was that the

rearrangement occurred readily. This supported the

previous hypothesis of Cooper and Greene3 that

rearrangement normally does not occur because of

unfavorable thermodynamics.








8

A number of researchers undertook investigations to

delineate the scope of this rearrangement. The initial

efforts focused on a narrow area. The rearrangement was

only studied in systems containing iron bound to highly

strained rings with strongly electron donating groups in

the alpha position (29) and under photolytic

conditions.13






O x O /

co' co coc K H )

(CH) n


N =2.3
X= OR. NR,
The work that will be described here will be

investigations intended to generalize this rearrangement.

It will concern the efforts to initiate such a

rearrangement on a different metal center (which could

undergo the rearrangement thermally) and in an iron based

system in which the rearrangement might occur with no

ring strain in the starting material.













CHAPTER 2
RHENIUM




In an effort to extend this rearrangement to a

different metal system a thorough literature search was

undertaken. There were obviously a great number of

possibilities from which to choose a reasonable

combination of metal and ligands. This search resulted

in the choosing of the rhenium pentacarbonyl group as a

very promising metal system for this rearrangement.

There were a number of factors which helped in this

decision. The first was that rhenium is a third row

transition element. Third row transition elements are

known to form relatively strong single and double bonds

to carbon.12,14 This is generally attributed to their

larger d orbitals, the energies of which allow better

overlap with the carbon orbitals than the first row

elements are capable of. This increases the stability of

the resulting complexes. Indeed there are a wide variety

of stable rhenium acyl complexes (e.g. 31), alkyl

complexes (e.g. 32), and finally there are also carbene

complexes (e.g. 33).15,16,17















(CO)5R/


OCH3
(CO)4Re==
I CH3
Re(CO)s


It was believed that the existence of these closely

analogous complexes would presage the stability of the

complexes that this project would entail making.

The second factor that pointed to this system as the

system of choice was the fact that the rhenium

pentacarbonyl anion (35) is moderately nucleophilic.18





Relative
Nucleophilicity


Cp(CO) 2Fe

PhS"

(CO)5Re-

Cp(CO)3W-

(CO)5Mn-


+ R-X

+ R-X

+ R-X

+ R-X

+ R-X


Cp(CO)2Fe-R

PhS-R

(CO)5Re-R

Cp(CO)3W-R

(CO) 5Mn-R


R = CH3I, C2H5I, C2H5Br, (CH3)2CHBr


(CO)
(CO),Re' )


7x107

2.6x106

2.5x104

5x102

7.7x101









11

It was presumed that this would be advantageous during

the synthesis of starting materials. One possible

synthetic strategy would be to first make the appropriate

organic substrate, then attach the metallic fragment via

a nucleophilic substitution reaction.

The third factor was the ease of synthesis of the

rhenium-containing precursors of the desired

organometallic compounds. The rhenium pentacarbonyl

anion (35), for example, is synthesized by a simple

sodium amalgam reduction of commercially available

dirhenium decacarbonyl (34).19




Na/Hg +
Re2(CO)1o 0- 2 Na Re(CO)5
THF





This could then serve as a nucleophile for reaction with

an organic moiety containing a strained ring (36) that

would generate the organometallic complex 37. This would

later undergo the ring expanding rearrangement to

generate a carbene complex.













OCHi ',"?

Na Re(CO OC-CO),yR-

(C (CH,)n




If one treats dirhenium decacarbonyl (34) with

iodine, one can easily obtain (CO)5ReI (39).20



CHCl
Re2(CO)10 + 12 CHC 2 (CO)sRe-I
60C



In the reverse of the step mentioned above which

generated an organometallic complex (37) via nucleophilic

attack of the rhenium pentacarbonyl anion (35) on an

alkyl halide (36), this metal halide (39) could then be

attacked by an organic nucleophile (40) which contained

the strained ring, resulting in the generation of the

same sigma complex (37).
OCHs
CH' (CO)Re-C
(C O )sR e-I ----
22 (CH,).
\ w^ -'7









13

Both of these methodologies arrive at the same types of

desired products, but from completely different

directions. Thus the choice of the rhenium pentacarbonyl

system provided a back up system for the formation of the

desired organometallic target compounds should our first

attempt fail.

The final factor which was in favor of the rhenium

pentacarbonyl system was the fact that the terminal

carbonyls are thermally labile. This is best exemplified

by the fact that one of the standard methods for forming

rhenium pentacarbonyl sigma complexes (e.g. 42) is

thermolysis of the corresponding acyl compounds (e.g.

41).21




0

HEAT
(CO)Re CH3 B (CO)5Re-CH3
-CO
41 42



The generally accepted mechanism for this transformation

involves initial dissociation of a terminal carbonyl

followed by migration of the alkyl group to the vacant

site on the metal center. The net result is the

conversion of an acyl complex (e.g. 41) into an alkyl









14

complex (e.g. 42) with the loss of one of the terminal

carbonyls.22


o o


(CO)SRe CH3 (CO)4Re CH3 (CO)sRe-CH,
-CO
41 41 42




This is significant in that it provides a

thermal route to a vacant site on the metal center. In

the metal systems which were known to undergo the desired

rearrangement (e.g. the rearrangement of 29 to 30)

photolysis of the starting material was necessary to

expel a CO and provide a vacant site for the

rearrangement. The thermal liability of the COs on the

rhenium pentacarbonyl system opens the possibility of

thermally initiating the desired rearrangement.

Whenever one is expanding one's area of knowledge,

the most rational approach in order to prevent utter

confusion is to change only one variable at a time and

see what effect it has. If this guideline is applied to

the attempt to search for a different metal system to

undergo the rearrangement, then it is obvious that the

organic fragment should be the same as one that is known









15

to rearrange in the original iron system. For this

reason it was decided to make pentacarbonyl(1-methoxy-

cyclopropyl)carbonylrhenium (44).


0

OCH3
(CO)sRe /


44


It was believed that this (44) could be either

thermolyzed or photolyzed to generate the sigma complex

(45). Upon further irradiation or heating this (45)

would undergo the desired rearrangement yielding the

carbene complex (46).



0
oOCH3 OCH3 OCH3
(CO)Re (CO)Re OCH (CO)4Re OCH
S-co I
44 41 0,/





The synthesis of the starting acyl complex (44) was

accomplished without much difficulty. The basic plan was

to react the metal anion (35) with 1-methoxycyclopropane

carboxylic acid chloride (47).












0


OCH3


Na Re(CO)5


0
OCH3

(CO)sRe /

44


This is the same pathway that was followed in the

original iron system (26). The (CO)5Re- (35) was

synthesized easily in the fashion mentioned previously.

The synthesis of the acid chloride (47) was successful

although fairly time consuming. The starting material

was thiophenylcyclopropane (48), which was treated with

n-butyllithium to form the anion (49) and then 12 to form

the iodide (50).23


SC6H5 nBuLi

H


SC6H5


12


SC6Hs

V<


Treatment of this compound (50) with base (K2C03) and

CH30H converted the iodo substituted cyclopropane (50)

into a methoxy substituted compound (51).


SC6Hs

Vt


K2C03
MeOH


SC6H5

OCH3









17

The next step was the conversion of the thiophenyl group

into a carboxylic acid. This was accomplished by

reductive cleavage of the thiophenyl group of (51) with

lithium naphthalenide and then bubbling CO2 into the

reaction mixture.23,24 The methoxycyclopropyl anion (52)

attacks the CO2 to generate 1-methoxycyclopropane

carboxylic acid (53). The yields are generally low and

highly variable. The byproducts (thiophenol and

napthalene) have a wretched stench.




SC6H5 Li Nap .. 1CO2, C02H

CH3 THF CH3 2) H30+ OCH
51 21






When the product (53) was obtained, however, conversion

into the desired acid chloride (47) was easily effected

by treatment with oxalyl chloride. Purification was

accomplished by Kugelrohr distillation.








18

Finally, reaction of this acid chloride (47) with the

rhenium pentacarbonyl anion (35) gave the desired acyl

complex (44) in moderate yields (generally 40% 50%).


o o
CI0 0

C lC


CO2H

OCH3
2.


0


00H3
47


The next step was the decarbonylation of the acyl complex

(44) to form the alkyl complex (45).


O
(COe OCH3
(CO)5 Re /


HEAT OR LIGHT
HEAT OR LIGHT


This was expected to be a simple step, but it turned out

to be the one that brought this pathway to a complete

stop. We were unable to generate the 1-

methoxycyclopropyl sigma complex (45) from the acyl


CCH3

(00)5%->



















,0











,-





0




x
0












a
-41

rz













0 .
S-,


U






0.
AC
SP~











i-4'







s.r1>


tr0
0r
ao Fi


cc
0
0




































































8 ^





8 u


O








-0 r
>1
Ln
0



0
0






r-4
>1













u z
0 0
Y 2 S











o
rl
m I

















oi
N
a)
C








.04
4
C 0> .







0



1-
34


















f








-r- 0
0
OE



35







0



3 U
8


4 4, O





















ON

S.C-
0 -r




OS






S" K







g *"









22

complex (44) synthesized above. Thermolysis in toluene

resulted in gross decomposition. There were a great many

products as shown by thin layer chromatography. All of

these products were in low yield and none could be

characterized adequately. We next turned our attention

to the photochemical decarbonylation of this acyl complex

(44). The photolysis in deuterobenzene was a very clean

reaction. There were only two products. These were

shown to be the dimers of the metallic fragment (34) and

the organic fragment (54).


0
OCH3 0

(CO)5Re OCH Re2(CO)1, + I

O CH30
44




The inorganic dimer (34) was characterized by TLC and

FTIR. The organic dimer (54) was characterized by 1H

NMR, FTIR, and GC/MS.

These results indicate that the main productive

process is homolysis of the Re-C acyl bond in (44) and

subsequent dimerization of the radicals (55 and 56).

















































o O

O 0























0


0













r-4


I


>,
ic -.
4J ,;


4 r-:3
a.,q



4
0 .4






A-
a >
u cr
m 0











O"
&& U

0 1


I-ro



0 0

P4






tO
Po











0

OCH3
CO 4Re


44


(CO)sRe


H0

CH3O


1/2 Re2(CO)Io
M


+ 1/2


OCH3 0



O CH30


Examination of this failure to decarbonylate shows

that it is not without precedent.15 There have been

references in the literature to other cyclopropyl acyl

complexes (57) that could not be decarbonylated to the

corresponding alkyl complexes.


No Reaction


M= CpFe(COz2

(CO)sRe

(CO) Mn


0









25

It is generally acknowledged25 that in photochemical

decarbonylations there are two competing processes: the

first is homolysis of the M C bond; the second is the

photoexcitation and subsequent dissociation of CO from

the metal center. It seems reasonable, therefore, that

if for some reason the migration of our

methoxycyclopropyl group to the metal were proceeding

either slowly or not at all, then the vacant site on the

metal center would have time to pick up a CO from

solution. This would regenerate the starting material

(44) and make the dissociation a non-productive process.

The homolysis to form the radicals 55 and 56 would be the

only observed process.

Further examination of the literature reveals that

the migration of a recalcitrant cyclopropane from an acyl

carbon to a metal center is highly solvent dependant. In

the iron case mentioned above the reaction was later

shown to go cleanly in acetone but only very poorly in

petroleum ether.26













light
pet ether






light
acetone


S 0







CO


yield: 16%







yield: 75%


The photolysis of the rhenium methoxycyclopropyl acyl

complex (44) was consequently tried in acetone. The

results were significantly different. The most obvious

result was that there were a large number of products.

Analysis of the volatiles by capillary gas chromatography

revealed 15 components. Product identification was

attempted by GC/MS. Unfortunately the poorer quality of

the separation on this instrument resulted in the

identification of only 5 of these components: 54, 60, 61,

62, and 63.
















(CO)sRe actoneD6 Re2(CO)o1 +
acetone-D1 4 >Y
O CH30
44
54

0


+ H KQ
OCH3

0



OCH3
H3


OCH3


+ MeOH





The problem seemed to be that we were unable to form

the needed sigma complex (45) that could undergo the

rearrangement. The presence of the carbonyl-containing

compounds (54, 60, and 61) indicates that cleavage is

taking place before the alkyl group shifts to the metal

center. The presence of methoxycyclopropane (62) may be

explained in two ways. The first is that the sigma

complex (45) was formed and then decomposed via radicals









28

55 and 64. The second is that the acyl complex (44)

decomposed to an acyl radical (56) which then lost CO to

generate the methoxycyclopropyl radical (64) and then

picked up the H. We have no evidence to favor either of

these mechanisms.


0

OCH3
(CO)sRe


0


(CO)sRe + "
CHO3
51 a


-CO


-CO


OCHa
(CO)sRe-


(CO)sRe


CH30


Since we were unable to obtain the desired sigma

complex through decarbonylation, we decided to attempt to

avoid the intermediacy of the acyl complex (44) and

synthesize the sigma complex (45) directly. There were

two ways we could go about this. We could either use the

rhenium pentacarbonyl anion (35) as a nucleophile and









29

attack an appropriately substituted alkoxycyclopropane

(36 or 65) in the alpha position or we could use a

nucleophilic alkoxycyclopropyl anion (40 or 67) and

displace a leaving group from the rhenium (39).



Br OEt
(CO)5Re- + EX _____ (CO)5Re



OEt


(CO)sRe-I + OU (CO)5Re
32 OEt>




The sequence employing rhenium pentacarbonyl anion

(35) as a nucleophile was attempted first since the

reactions were more familiar and, in any event, the

organic reactant (65) would need to be synthesized anyway

for the sequence employing the cyclopropane (67) as a

nucleophile. It was realized that this sequence was not

on firm ground since the rhenium pentacarbonyl anion (35)

is only moderately nycleophilic18 and it would need to

attack a quaternary carbon.

Both the organic and inorganic reactants were

synthesized by literature procedures.19 The organic









30

synthesis first involved reduction and ring closure of

ethyl 3-bromopropionate (68) followed by trapping with

trimethylsilyl chloride to give l-ethoxy-l-trimethyl-

siloxycyclopropane (70) in 57% yield.27




S a 0 TMSCI OTMS

E20 OEt OEt
Br OEt


68


This (70) was then treated with phosphorous tribromide to

generate l-bromo-l-ethoxycyclopropane (65) in 72%

yield.28




OTMS PBr Br

OEt OEt



As in the case of the acyl complex, the rhenium

pentacarbonyl anion (35) was generated by reduction of

commercially available Re2(CO)10 (34) with sodium

amalgam.19

The results of mixing the rhenium pentacarbonyl anion

(35) with l-bromo-l-ethoxycyclopropane (65) were very









31

disappointing. After periods of time ranging up to two

days, the starting organic material (65) could be

recovered unchanged. The rhenium pentacarbonyl anion

(35) is somewhat unstable, eventually going to a variety

of inorganic compounds, as evidenced by TLC, IR, and 1H

NMR.


Br
(CO)sRe- + ----------- No Reaction
OEt

61



In order to verify that the problem lay with the reaction

and not something else, the reactants were tested

separately. Since we could easily make the cyclopropyl

acyl complex (44) we knew that we were making the metal

anion (35) correctly.

A co-worker reacted the l-bromo-l-ethoxycyclopropane

(35) with the much more nucleophilic dicarbonyl(n5-

cyclopentadienyl)iron anion (71), the result being the

desired sigma complex (72) in 25% yield.28



0 >Br __ 0 Et

co+ OrEt 1.
C1 65
21 --








32

The most likely cause for the lack of reaction, then, is

insufficient nucleophilicity of the rhenium pentacarbonyl

anion (35).

This failure was disappointing, although not entirely

surprising. We next turned our attention to the use of

the 1-ethoxycyclopropyl anion (67) as a nucleophile. The

syntheses of the reactants were again relatively simple.

Treatment of Re2(CO)10 (34) with 12 results in (CO)5Re-I

(39).20 Treatment of l-bromo-l-ethoxycyclopropane (65)

with tert-butyllithium is known to give the metal

exchanged product (67).28 To our surprise, there was no

reaction when the two were mixed. The (CO)5Re-I (39)

could be recovered after reaction times of up to two

days.



(CO)5Re-I + +- No Reaction
OEt







A literature investigation indicated that (CO)5Re-I

(39) had not been the most favorable system to try this

reaction on. It has been shown that when CH3Li is mixed

with (CO)5Re-I (39) displacement of the iodide to give









33

the methyl sigma complex (42) does not occur. The

isolated product is the acetyl-iodo rhenium anion (73).20
co o- Li+
SJOCH3
CH3Li + (CO)5Re-I OC- Re- I
oc/ I
23 co

27





Indeed, if two equivalents of methyllithium are added to

the same reagent (39) and then the product protonated,

the resulting compound is the bis hydroxycarbene (74).30


(CO)5Re-I


1) 2 CH3Li
2) H


CO OH
I H3
OC- Re- I
OC H
HO CH3


The literature also reports that if one of the

terminal COs is replaced by a phosphine then the

substitution of the halide in 75 by a methyl group may be

accomplished easily to give 76.31













CO
/CO
OC- Re-Br
/C I
OC PPh3


CH3Li


CO
I co
I/CO
OC- Re- CH3
OC PPh3


The phosphine substituted rhenium halide (75) was

synthesized according to literature procedures in 21%

yield (literature yield 52%).32


Br2
Re2(CO),o (CO)sRe-Br
CC34


PPh3

CHCI3


(PPh3)(CO)4Re-Br


The reaction with l-lithio-l-ethoxycyclopropane (67) was

carried out. Unfortunately, the results were the same.

The starting rhenium reagent (75) could be recovered

unchanged after two days of reaction.

In order to rule out the possibility that the 1-

lithio-l-ethoxycyclopropane (67) was not being generated









35

at all, the same procedure that had been used previously

to generate it was performed and then benzaldehyde was

added. After protonation and work up the resulting

alcohol (77) was isolated in 67% yield.




B. t-BuLi i t OH

t OD I OEt
alo

22

At this point we were forced to abandon the choice of

cyclopropane as the strained ring to undergo

rearrangement. Our attempts to form the sigma complex

(66) directly were fruitless. Although we were

successful at forming the acyl complex (44), it would not

decarbonylate to form the sigma complex (45) or carbene

complex (46).

The most reasonable alternative to the cyclopropane

system was the cyclobutane system. The reaction scheme

we chose is shown below.











O
(CH OCHH3 OCH3
(CO)5Re / (CO)5Re (CO)4R




There were three reasons for this choice. The first is

that experiments with the analogous dicarbonyl(n5-

cyclopentadienyl)iron system (81) have shown that the

rearrangement of interest can occur in the cyclobutyl

system.13




o 0 O OCH3 OCH
Ire --z 3rz e--
c CO Co Co


81

The second reason was that it was hoped that the

cyclobutyl ring would migrate from the acyl carbon to the

iron to form a sigma complex more easily due to its more

normal ring bonds.

The last reason is that the necessary organic acyl

chloride (87) was known and may be made easily.13














CO2H







NaH
CH3I


1) LDA
2) 02
3) NaHS03


OCH3
- CO2 H


OH
CO2H


85


O OCHO
OCH3
C0 l CI
-- d Cl\


The synthesis of the starting acyl complex (78) was

accomplished in 52% yield without incident.


OCH3
O

+ \
Cl


0
5 / OCH3
(CO)5Re "


(CO)5Re-

23






















J


O


I

o


u

x
o
0



r-4












Or
c




0
c4
**













-.4
04 4-1





t 0
I u



















0'
0
0c
a OA
2 ^
2 10
- X"
ac -1
*


~gu,



























>4
0

0

U
o



4C
-1

0)
E

2 -





.4
O
:c






-P







o,




n
m







\ a!









40












r-I
o











3








0a
o




r- 4
S0Q






4 .C
040 4





0 A





*Gu




a4
mm

c ^-




Qre

: u-OP
:
^ 2,
8 i8 .2









41

Upon thermolyis in toluene at 1050C, however, the

reaction did not go as expected. Thermolysis yielded

several products, all in poor yield. The mixture of

products was very difficult to separate. After many

attempts at column chromatography we gave up trying to

purify the mixture. Some information, however, could be

gleaned from the results. The large amount of gas

evolution observed combined with the observation of peaks

in the vinylic region of the 1H NMR spectrum and the lack

of a characteristic peak very far downfield in the 13C

NMR spectrum indicate that a probable sequence involves

decarbonylation, formation of the desired sigma complex,

and then decomposition via eliminations and ring opening

reactions of the cyclobutane ring. The absence of the

peak far downfield in the 13C spectrum indicated that the

desired carbene complex (80) was not present within the

limits of detection. The 1H NMR spectrum did not

indicate the presence of either methoxycyclobutane (88)

or 1-methoxycyclobutene (89). After this we abandoned

hope of obtaining the carbene (80) via thermolysis. We

turned our attention to photolysis.

It was hoped that the reason for the decomposition

was the large amount of energy available at this high

temperature. It was believed that by doing the









42

decarbonylation photochemically at lower temperature the

sigma complex (79) would not have enough energy to

decompose as it had at higher temperature in the

thermolyses.

The results of the photolyses in deuterated benzene

indicated that these expectations were partially correct.

They indicate that the decarbonylation and rearrangement

to the sigma complex did take place successfully. It did

not react further in the same way that it had under

thermolytic conditions. Instead the sigma complex

reacted further in a very interesting and unexpected way.

The products obtained were characterized by 1H NMR,

FT-IR, GC-MS, and comparison to the reported properties

of the known compounds.33,34 The organic products were

shown to be a nearly 1:1 mixture of methoxycyclobutane

(88) and 1-methoxycyclobutene (89). The inorganic

product was determined to be almost exclusively dirhenium

decacarbonyl (34).


0
OCH3
(O)Re OCH3 light ,----OCH3 +
(CO),-Re------ Re(CO),o .[ +
-CA-
34 89













































' H
8 8
















U


43



















C
4







I




,-4
4 0o o-





W4-41
Oc
(A ( 4








*r4 >i

i U)



t
41.-I
0 u

CO
>i3

.f-4




(*4









44

There were two possible mechanisms that were

considered to explain this observation. They both start

with the decarbonylation of the acyl complex (78) and

subsequent rearrangement to the desired sigma complex

(79). The two mechanisms then diverge. The first

mechanism relies on beta hydrogen abstraction by the

metal to generate the products. The second relies on a

radical pathway to produce them.

The overall scheme of the beta hydride abstraction

mechanism is as follows:


OCH3
OH3 OCH3
(CO)sRe-- -CO (CO)4Re- L L(CO)4Re-H +
S1 91 0
H B2
22
OCH3
(CO)4Re -


79

OCH3
Re2(CO)o--nL + t

22 BA



There are two variants of the radical process, a

chain mechanism and a non-chain mechanism. An overview

of the non-chain mechanism is as follows:
















OCH3
(CO)5Re


light


OCH3
(CO)5Re' +

21


OCH3
(CO)5Re-H + NT
94


OCH3



19


OCH3



21


OCH3


sa


(CO),Re +



I
Re2(CO)io
3A


OCH3

\'I









46

The steps in the chain mechanism are very similar to

those in the non-chain mechanism:


OCH'3
(CO)5Re--


1ight


(CO)5Re
5a


OCH3


9l


OCH3
(CO)5Re 1z


12


OCH3


+ (CO)sRe-H
94


OCH3


22


OCH3



88


(CO)5Re +


(CO)sRe +
55









47

The first step in the beta hydride abstraction is

proposed to be dissociation of a CO from the sigma

complex (79) to form a 16e" intermediate (95). This

intermediate could either pick up a ligand from solution

or abstract a beta hydrogen to give a methoxycyclobutene

pi complex containing a metal hydrogen bond (90 or 98).






OCH3 OCH3 OCH3
(CO)5Re-- H (CO)4Re COX4Re OCH
__ _" (CO)4Re- i
79 H



OCH3 OCH3

L(CO)4Re L(CO)3Re







OCH3

L(CO)3Re--
H


This type of rearrangement has been shown to occur in the

analogous dicarbonyl(n5-cyclopentadienyl)iron system

(99).13















OCHa

_ FeFeJ----
H

99 l00


SFeT1
CO' H

OCH3
Ill


The pi complex could then expel the 1-methoxycyclobutene

(89) and replace it with a ligand from solution to give a

hydride complex (91 or 103).


OCH3

(CO)4Re- -









OCH3

L(CO)3Re

H
102


L











L


L(CO)4Re-H +
21









I2(CO)3Re-H

103


OCH3



BS


OCH3


89









49

This hydrido complex (91 or 103) would then attack a

molecule of 1-methoxycyclobutane sigma complex (104) to

generate methoxycyclobutane (88) and the metal dimer

(106).35


OCH3


"'I'


LM(CO)5.MRe-H +

l1 M=1
103 M=2


OCH3
(CO)NRe-


104 N=4,5


+ Re2(CO)I0.MLM


The other mechanism involves the homolytic cleavage

of the acyl (78) or sigma complex (79) to generate the

rhenium pentacarbonyl radical (55) and the

methoxycyclobutyl radical (93). The rhenium

pentacarbonyl radical (55) could then abstract a hydrogen

from a molecule of the sigma complex (79) to generate

methoxycyclobutene (89), (CO)5Re-H (94), and another

rhenium pentacarbonyl radical (55).













OCH3
(CO)5Re--


light


OCH3


9I


+ (CO)5Re
55


OCH3


B2
"D-


(CO)sRe-H
94


In the non-chain mechanism the rhenium pentacarbonyl

radical (55) abstracts a hydrogen from the

methoxycyclobutyl radical (93) to generate the metal

hydride (94) and 1-methoxycyclobutene (89).


(CO)5Re
51


OCH3
(CO)5Re --h


22


light


OCH3



93


(CO)sRe-H
94

+
OCH3


82


(CO)sRe
51


OCH3



* C 3


OCH3
(CO)SRe I


7922









51

The hydride (94) could then react with another molecule

of the 1-methoxycyclobutyl sigma complex (104)36 or the

methoxycyclobutyl radical (93) to generate

methoxycyclobutane (88).



OCH3 OCH3

(CO)5Re-H + (CO)NRec Re2(CO)+N +

i a
104

OCH3 OCH3

(CO)5Re-H + (CO)5Re +
94 55
93
21 SS





In an effort to determine which scheme is the more

nearly correct, several experiments were carried out.

They fell into two basic areas. The first of these

concerned deuterium incorporation during photolysis in

deuterated solvents. The second involved observation of

the changes induced in the photolysis by the presence of

added reagents.

The products of the photolyses in deuterated benzene

and especially deuterated toluene were analyzed carefully

by 1H NMR and GC-MS and there was no evidence for









52

deuterium incorporation. This was especially important

for the hydrogen on carbon 1 of methoxycyclobutane (88)

since it must come from some intermolecular source. It

would seem reasonable that if this were a radical

mechanism and it were carried out in deuteriotoluene,

there would be some incorporation of the deuterium into

the products. This speaks against the radical mechanism,

but is obviously only negative evidence.

The second area of investigation concerned the

effects of added phosphines on the photolysis. When one

equivalent of triethylphosphine is present during the

photolysis, two significant changes are observed. First,

one can actually observe the presence of a metallohydride

in the 1H NMR spectrum. The hydride that is observed is

the one substituted by two phosphine ligands (105).



CO
I PEt3
OC-Re-- H

oc PEt3





Second, the ratio of methoxycyclobutane (88) to 1-

methoxycyclobutene (89) is no longer roughly 1:1. In the

initial stages of the reaction there is as much as a 5









53

fold excess of 1-methoxycyclobutene (89) over

methoxycyclobutane (88).




0
|H OCH3

0)Re light OCH3 + + (PEt3)2(CO)3Re-H
PEt3 10I


1 : 5



As the photolysis is continued, the amount of the

metallohydride (105) present decreases and the ratio of

methoxycyclobutane (88) to 1-methoxycyclobutene (89)

returns to nearly 1:1. Towards the end of the photolysis

trace amounts of hydride substituted with a single

phosphine are detected. The number of phosphine ligands

present on a metallohydride is easily determined due to

the fact that phosphorus has a spin of 1/2. The net

result is that the presence of one phosphorus on the

metallohydride splits the far upfield peak in the 1H NMR

spectrum (5 to 6 ppm upfield from TMS) from the hydride

into a doublet and the presence of two phosphorus ligands

splits the hydride signal into a triplet.









54

Replacing the triethylphosphine with triphenyl-

phosphine or trimethylphosphite results in significantly

different, although related, results. The

triphenylphosphine acted almost as if it weren't there.

Although detectable, the hydride never built up to a very

high concentration and the ratio between

methoxycyclobutane (88) and 1-methoxycyclobutene (89) was

unchanged as the reaction progressed. The presence of

trimethylphosphite resulted in a very rapid growth of the

peaks due to the hydride and then an equally rapid

decrease. This also left the ratio of methoxycyclobutane

(88) to 1-methoxycyclobutene (89) unchanged. A

reasonable explanation for this is that the

triphenylphosphine is not taken up during the reaction

and the trimethylphosphite is, but the hydride

substituted with two molecules of trimethylphosphite

reacts rapidly to generate the methoxycyclobutane and

metal dimer.

All of the mechanisms presented above can account for

these observations. In the beta hydride abstraction

mechanism a simple way to account for this is to have the

addition of one triethylphosphine be faster than the

rearrangement of the acyl complex (78) to the sigma

complex (79).














0
OCH3
- (CO)4Re


106
PEt3
Faster

SPE


OCH3
slower (CO)5Re
,- (CO)SRe-


The second phosphine replaces the 1-methoxycyclobutene

(in 108) to form the diphosphine metallohydride (105)

which then slowly reduces the sigma complex (79) to give

the metal dimer (106) and methoxycyclobutane (88).


OCH3

PEt3(CO)3Re-

H


OCH3


PEt
PE3 _. (PEt3)2(CO)3Re-H

105


OCH3








56

Either radical mechanism can account for this simply by

having the rhenium pentacarbonyl radical (55) substitute

two phosphines, the second more rapidly than the first.

There is literature precedent for the rapid exchange of

phosphines on radical centers, with different phosphines

substituting at different rates and with differing

amounts of mono- and di-substitution.37



(CO)sRe + 2 PEt3 (PEt)(CO)4Re------- (PEt3)2(CO)3Re





To distinguish between radical and non radical

mechanisms two samples of the l-methoxycyclobutyl-l-

carbonylrheniumpentacarbonyl (78) were photolyzed in

benzene in the presence of triphenylmethane and p-methoxy

phenol, respectively. It was hoped that these compounds

would donate hydrogen atoms to any 1-methoxycyclobutyl

radicals (93)present, thus increasing the ratio of

methoxycyclobutane (88) to 1-methoxycyclobutene (89).

The triphenylmethane had no effect. The ratio remained

at roughly 1:1. In contrast, photolysis in the presence

of 5 fold excess of p-methoxy phenol resulted in a large

increase in the amount of methoxycyclobutane (88). The

final ratio was 4.3:1 in favor of methoxycyclobutane









57

(88). This is very strong evidence in favor of the

presence of the methoxycyclobutyl radical in the

solution, supporting the radical mechanism. There was no

reaction of the p-methoxyphenol with the acyl complex

(78) in the absence of photolysis. In addition, it will

be shown later that p-methoxyphenol does not react with a

complex that is very similar to sigma 79, the only

difference being that the cyclobutane ring is

benzannelated (i.e. complex 130). This indicates that p-

methoxyphenol is not a strong enough acid to form

methoxycyclobutane (88) in the absence of a radical

reaction.




0
OCH3
(CO)ReH light OCH3 i+

OH


OMe 4.3 :



Qualitatively the rate of the reaction was not

affected by the presence of p-methoxyphenol. If the

chain mechanism were operating, one would expect a

decrease in the rate of reaction.









58

This would come about as a result of the trapping of

the rhenium pentacarbonyl radical (55) to give the

hydride (94). Since the radical (55) is the chain

propagating agent, the chain would necessarily be

broken. The lack of an observable rate decrease speaks

strongly against a chain mechanism.

It has been shown by Heinekey and Graham38 that 7-

carbonylcycloheptatrienerhenium pentacarbonyl (111) may

be decarbonylated to the corresponding sigma complex

(112) by photolysis in deuterated acetone at low

temperature.




0

Light
(CO)5Re (CDO-- (CO)Re-
650C

112
111




We decided to follow his example and also expand on

it. We carried out photolyses of the methoxycyclobutyl

acyl complex (78) not only in deutero acetone at -650C,

but also at room temperature and in deuterated toluene at

-650C to try to synthesize the sigma complex (79). The










59

products were characterized by 1H NMR and also GC-MS. At

room temperature the organic products of the photolysis

in deuterated acetone were characteristic of a free

radical mechanism.


H
H-OCH3




+
OH
CH30

C-- CD3

CD3
113
+


OCH3

+ + MeOH

S2 fil


OCH3


OCH3


Light
(CD3)2CO
Room temp.








60

Since this did not result in the desired product, we

cooled the sample to -650C during the photolysis of the

acyl complex (78) in deuterated acetone. This gave only

three organic products. These products were 1-

methoxycyclobutene (89), methoxycyclobutane (88), and 1-

methoxycyclobutanecarboxaldehyde (115) in a 1:0.56:0.40

ratio, again products typical of a free radical reaction.




SH OCHOCH O

(CO)5Re OCH3 Ligt z OCH3 + + PH
(CO)Re(CD3)2CO
-650C

2a 0.56 : 1 :0.40








The low temperature (-650C) photolysis was repeated

except in deuterated toluene' The products were the

same. The amount of methoxycyclobutane (88)could not be

determined due to overlap of its methoxy signal with

unreacted starting material (78). The ratio of aldehyde

(115) to alkene (89) was 0.5:1.












X H OCHH OCH O
OCH3 Light OCHH
(CO)sRe OC Toluene-D8 + P H
-650C
S9 115

1 : 0.5



A reasonable explanation for the observation of the

aldehyde (115) at low temperature is that at room

temperature after absorption of the first photon and loss

of CO the migration of the cyclobutyl group to the metal

center is relatively fast. By the time the photon comes

which converts the molecule to radicals, the molecule is

already a sigma complex (79). The radicals, hence, are

the rhenium pentacarbonyl radical (55) and the 1-

methoxycyclobutyl radical (93). At low temperature,

however, the migration is not nearly so fast. After the

ejection of the CO the methoxycyclobutyl group does not

migrate to the metal immediately. This allows time for

the ejected CO to return to the vacant site. The low

temperature would also aid this process due to the

decrease of rates of diffusion near the vacant site,

increasing the length of time the CO spends in the

vicinity of the metal and hence increasing the

probability of return. This decreases the efficiency of









62

the decarbonylation and increases the probability that a

relatively inefficient process that depends on the

concentration of the acyl complex (78) (i.e. homolysis to

generate radicals 116 and 55) will become significant.


0
OCH3
(CO)5Re /


(CO)sRe +

ff


0
OCH3
(CO)4Re /


106


-CO

+CO


(CO)4Re

117


OCH3




116


Hi

116


+CO


(CO)sRe +

55


OCH3 0




116


H.-1









63

In addition, the photolysis of the cyclobutyl acyl

(78) was performed under 6.3 atmospheres of CO at room

temperature in CgD6. There were no observed changes in

either the rate of reaction or the products (i.e.,

methoxycyclobutane (88) and 1-methoxycyclobutene (89))

compared to an identical sample photolyzed side-by-side

with the CO pressurized sample except under 1 atmosphere

of N2. This is consistent with the product forming

reaction being faster than the reaction of 106 with CO.

At this point we felt we had enough information to

make a reasonable postulate as to what was happening.

Upon photolysis the vast majority of the photons which

were absorbed resulted in the ejection of a molecule of

CO. At room temperature the cyclobutyl group migrated

rapidly to the metal center. At low temperature there

was only partial migration. The next productive reaction

was the homolysis of the Re-C bond through absorption of

another photon to generate the metal radical (55) and

either the methoxycyclobutyl radical (93) (at room

temperature) or the methoxycyclobutyl acyl radical (116)

(at low temperature). This brings up an interesting

point: why (at room temperature) is the first productive

reaction upon absorption of light the ejection of CO and

the second productive reaction upon absorption of light









64

the loss of the alkyl radical? It seems unlikely that

the electronic structure of the rhenium pentacarbonyl

moiety is so drastically different when it has an alkyl

ligand attached compared to when it has an acyl ligand

attached that there is no loss of CO in the alkyl case

and all photons cause homolysis to the radicals. The

upshot is that there is probably loss of CO from the

alkyl case, but it comes back in to the vacant site.

This is important since it points out that the cyclobutyl

ring did not rearrange to form a carbene even though

there was a vacant coordination site present to permit

it. This indicates that the rearrangement was either

endothermic or had a high energy of activation. We

decided to lower the energy of the desired carbene

complex. If the lack of rearrangement was due to the

rearrangement being endothermic, this change might induce

the formation of the desired carbene. If the problem lay

in the activation energy, this change might help also due

to the Hammond postulate.

Just as an electron-donating group on a carbene

carbon stabilizes the complex, it has been shown that

replacement of a CO on a metal complex by a phosphine

also stabilizes the complex.39 We decided to replace one

of the CO's on the cyclobutyl acyl complex (78) with a









65

triethylphosphine group and compare its (107) reactivity

with that of the analogous unsubstituted complex (78).

This substitution was accomplished by simply heating

the unsubstituted acyl (78) with 1 equivalent of

triethylphosphine in benzene for 6 hours at 700C.




o o
SOCH3 70C OCH3
(CO)sRe + PEt3 C cis PEt3(CO)4

118






The reactivity, unfortunately, was exactly the same

as before the substitution. Thermolysis resulted in a

very complex mixture of products. The 13C NMR spectrum

did not show the presence of the characteristic low field

peak indicative of a carbene carbon. The photolysis in

benzene produced a mixture of methoxycyclobutane (88) and

1-methoxycyclobutene (89) in a roughly 1:1 ratio.












H

Light OCH3

250C


1


OCH3



89


1


The photolysis in deuterated acetone at -650C gave

l-methoxycyclobutene (89), methoxycyclobutane (88), and

l-methoxycyclobutanecarboxaldehyde (115) in a 1:0.42:0.47

ratio.


H OCH3 OCH3
OCH3 Light -OCH3 + + -
(CD3)2CO +H
> -650C


0.42 : 1.0 : 0.47


A mechanism which incorporates all of the data is as

follows:


PEt3((













0
5 / OCH3
(CO)sRe


OCH3
(CO)sRe -


Light

PEt3
-CO


(CO)5Re

51


(PEt3)2(CO)3Re
-io


OCH3




29


PEt3
Fast


OCH3

+ E

291




OCH3
+


(PEt3)2(CO)3Re-H

io5


OCH3




22


OCH3



la


Re2(CO)s(PEt3)2
119















OCH3
(CO)5Re-


Light

-co


(CO)5Re

5s













(CO)5Re-H

24


Re2(CO)lo +

34


OCH3
















OCH3






OCH3




93


OCHa




88














0
OCH3
(CO)5Re


OCH3

Light (CO)5Re i- (CO)5Re + O

-Co


I OH


OMe


+ (CO)5Re-H
24










+ (CO)5Re
55


(CO)5Re +
51


OCH3


tCuS


OCH3



91


OCH3

t +


(CO)5Re +
55


(CO)5Re-H
94


OCH3



HI
T3







OCH3













OCH3
Light (CO)Re Light (CO)sRe
Acetone-De ast
-650C


(CO)sRe-H
24


OCH3




22


Re2(CO)io
34


OCHR
(CO)SRe


+ (CO)sRe +

116 5


OCH3


+


Light
Acetone-D6
-650C


(CO)5Re +
f5


OCH3 0




116


OCH3


88


(CO)5Re-H
24

+
OCH3



131


OCH3 O


H---"
H

115









71

The direct methods of forming a carbene complex from

a cyclobutyl complex were obviously not giving the

desired results. An indirect method was sought which

might allow a system to undergo a rearrangement of the

type under study. An effort was made to put the

possibility of beta hydride abstraction to our advantage.

It has been shown in the iron system that the trans beta

methoxy cyclobutyl sigma complex (120) will, upon loss of

CO, isomerize to the alpha methoxy sigma complex (99) and

then rearrange to the carbene complex (83).13


- 7

OCH3
120


OCH3

- Fe-|
co'' I
H


OCH3
1i.


It was hoped that the analogous rhenium pentacarbonyl

complex (121) would undergo similar rearrangements.













OCH3
(CO)Re- (C)4Re (CO)4Re-

OCH3 OCH3 H
122
121
OCH3 OCH3
(CO)4Re- (CO)4He






There is one major advantage to using this indirect

pathway. It stems from the fact that we were unable to

isolate any alpha methoxy sigma complex (79) from the

previous work. This indicates that it (79) is either a

highly reactive species which decompose to something

other than the desired carbene complexes or was never

formed. The mechanistic results detailed previously

(e.g. the formation at room temperature of

methoxycyclobutane (88) and 1-methoxycyclobutene (89) but

no 1-methoxycyclobutanecarboxaldehyde (115)) argue that

the sigma complex (79) is formed. If sigma complex 79 is

formed it decomposes to compounds other than the desired

carbene complex (80). The advantage of the above scheme

is that once the alpha methoxy sigma complex (95) is

formed in the indirect pathway there is already a vacant








73

coordination site on the rhenium. It was hoped that this

would increase the probability of the rearrangement to a

carbene (80) happening as opposed to decomposition

pathways.

The synthetic strategy involved the reaction of

cyclobutene epoxide (123) with rhenium pentacarbonyl

anion (35) followed by trapping by CH3I.




o
CHOI
(CO)sRe- + (CO),Re 0- CH31 (CO)sRe-

123 124 OCH3
121


The cyclobutene epoxide (123) was synthesized by

literature methods. The first step involved the

reduction of commercially available cyclopropane

carboxylic acid (125) with LiAlH4 to the alcohol (126).40




LiAIH4
-CO H L ->- "CH2OH
125 126

The second step was isomerization of the cyclopropyl

carbinol (126) to cyclobutanol (127) via acid

catalysis.39 This compound (127) was then converted to









74

the tosylate (128) by reaction with tosyl chloride in

pyridine.39





H+ OH TsCI -- OTs
P_--__HOH ------- --- I
H2H H20 Pyridine
127 128



The tosylate (128) was then treated with potassium

t-butoxide in DMSO to form cyclobutene (129). The

cyclobutene (129) was collected and bubbled through

a -200C solution of m-chloroperoxybenzoic acid in CH2Cl2.

This mixture was then stirred for 3 days at 0C. The

solvent was carefully removed by Vigreaux distillation.

The final purification of the cyclobutene epoxide (123)

was accomplished by preparative gas chromatography.41



0
OTs KOtBu mCPBA
DMSO CHzC I
128 -200C


The results of the addition of rhenium pentacarbonyl

anion (35) to the cyclobutene epoxide (123) followed by

trapping with methyl iodide were very disappointing. The

only isolated organometalic product was (CO)5Re-CH3, and









75

this was only in trace amounts. Apparently the

nucleophilicity of the rhenium pentacarbonyl anion (35)

is not sufficient to open up the epoxide (123). The

anion (35) simply does not react and slowly decomposes.

The small amount that remained eventually attacked the

methyl iodide.

A short summary of the major points of this project

is in order. The first attempt involved the use of 1-

alkoxy substituted cyclopropane (44) to achieve the

desired rearrangement. The requisite sigma complex (45)

could not be formed from the acyl complex (44). The 1-

ethoxycyclopropyl sigma complex (66) could not be formed

either via nucleophilic attack on a rhenium halide (39,

76) or from nucleophilic attack of a rhenium

pentacarbonyl anion (35) on l-bromo-l-ethoxycyclopropane

(65).

The decarbonylation of the 1-methoxycyclopropyl acyl

complex (44) resulted in homolytic cleavage of the Re-C

bond. This was attributed to the reluctance of the

cyclopropane to migrate to vacant coordination sites on

the metal.













0
O H3 OCH3
(CO)SRe ,- (O)s Re


44




The attempt to solve the problem of the lack of

migration from acyl to alkyl by going to the 4 membered

ring seems to have been successful.



o
OCH3 OCH3
(CO)sRe/ (CO)5Re-






This brought with it a different problem. The metal

seemed to be removing a beta hydrogen from the sigma

complex (79) during radical decomposition.

Before we had conclusively found that the mechanism

of decomposition was via radicals, we searched for a

system which would be immune to the problems encountered

previously. It must not involve a cyclopropane, it must

have a methoxy in the alpha position, and either have no

beta hydrogens or beta hydrogens which, if they









77

rearrange, would give a pi complex that would not

dissociate.

The system which was chosen satisfied all of these

requirements. This system was the 7-methoxy-7-

rheniumpentacarbonyl-benzocyclobutene system (130).



OCH3
(CO)5Re



130



There is obviously no cyclopropane. Molecular mechanics

calculations42 indicate that opening the 7-methoxy-

benzocyclobutene system can relieve more strain (48.5

kcal/mol) than the opening of either the methoxy-

cyclopropane (28.1 kcal/mol) or methoxycyclobutane (26.1

kcal/mol) systems. The problem of beta hydrogens is

solved elegantly. The solution to the problem lies in

the fact that if a beta hydrogen should be removed then

the product would be a methoxycyclobutadiene derivative.

If the hydrogen were removed via a beta hydrogen

abstraction then the methoxycyclobutadiene would be

completed to rhenium (131).













LuL3 CH30
CO) Re (CO)4Re


10 -131








Although cyclobutadiene complexes are known,43 the

cyclobutadiene moiety is completed tightly to the metal.

We expected that the hydride ligand on the same metal

would react with the cyclobutadiene ligand to regenerate

a sigma complex (130) which would then rearrange to a

carbene complex. This would be completely analogous to

the iron pi complex (100) to sigma complex (99) to

carbene (83) series of rearrangements shown earlier.13

If the beta hydrogen were removed by a radical

process, then, assuming the same reactivity as in the

non-benzannelated case, the cyclobutadiene would be the

free molecule (132).















OCH3
(CO)5 Re CH30

(CO)5Re 132

55
+


(CO)5Re-H 94

+

(CO)5Re 55











We would not expect either of these processes to take

place due to the large amount of ring strain that would

be introduced due to the formation of the cyclobutadiene

ring. Molecular mechanics calculations42 were used to

quantify this. They indicated that converting

methoxycyclobutane (strain = 26.1 kcal/mol) to 1-

methoxycyclobutene (strain = 33.0 kcal/mol) involves an

increase of 5.9 kcal/mol in strain energy. They also

indicated that converting 7-methoxybenzocyclobutene

(strain = 48.5 kcal/mol) to 7-methoxybenzocyclobutadiene









80

(strain = 66.0 kcal/mol) involves a 17.5 kcal/mol

increase in strain energy.

Unfortunately the synthesis of the target molecule

(130) was very difficult. There were two major pathways

which were undertaken. The first involved pre-forming

the entire methoxybenzocyclobutene fragment (133) before

attaching it to the metal center.




OOCHCH3
OCH3 OCH3

(CO)5Re" + 0 (CO)sRe _


133 130






The second pathway involved forming the benzocyclobutene

fragment (134), attaching it to the metal, then putting

the methoxy group in place.














(CO)5Re


(CO)5Re" +

35


;CO)5Re



136


The former procedure was attempted first for two reasons.

The first is once a metal is involved in a reaction our

experience has shown that a wide variety of unexpected

reactions may take place. The former sequence involves

fewer steps once the metal is attached. The second

reason is that the cationic rhenium pentacarbonyl carbene

complex (137), which is critical to the sequence, is

unknown. More broadly, there are no references in the








82

literature to any cationic rhenium pentacarbonyl carbene

complexes. The analogous dicarbonyl(n5-

cyclopentadienyl)iron compound (138) has been known for

some time, however.44




CO









The overall synthetic sequence that was begun for the

former plan is as follows:








:CBr2 Br -CN CN


1 14 141




















-OH f C02 H

H20

142


-CO2H


Base
CH31


1) Base
2) 02
3) H20











0 0
--A/
c1 > c1


*COzH










OCHs


The steps up through the carboxylic acid (142) are all

known from the literature. The bromobenzocyclobutene

(140) was synthesized by heating cycloheptatriene (139)

and bromoform in the presence of a base (K2CO3) and a

catalyst (18-crown-6).45















O140 S-r---Br


1400 140
+ CHBr3 + K2CO + 18-C-6





This bromide (140) was then treated with KCN in DMSO

to replace the bromine with a nitrile (141).46 Finally,

the nitrile was hydrolyzed to the acid (142) by treatment

with KOH/EtOH/H20.47




Br KCCN KOHCH
DMSO EIOH
H20
790C
140 141 142



The sequence from cycloheptatriene (139) to the acid

(142) was very interesting from a pragmatic point of

view. As one might expect, the first step was a very

messy reaction. There are several impurities in the same

fraction of the distillate as the bromobenzocyclobutene

(140). After repeated attempts we were able to re-

distill the mixture and obtain pure product (140).

Unfortunately a great deal of the product was lost during









85

the purification. In order to avoid this loss, we tried

to carry the impurities along to the nitrile stage (141)

and separate them then. This was unsuccessful. The best

method by far was to carry the impurities all the way

through to the acid stage (142). All of the impurities

are considerably more volatile than the acid (142).

Simply applying a vacuum (0.01 torr, 250C) removes all of

them, leaving the acid as a white solid.

The rest of the planned synthetic sequence was

completely analogous to that of the non-benzannelated

system (87). The introduction of the methoxy group would

be accomplished by treatment of the acid with two

equivalents of strong base followed by oxygen and work-up

to give the alcohol which could be methylated by

treatment with base and CH31.

The introduction of the alcohol to give 143, gave an

inseparable, complex mixture. After several attempts at

direct purification we decided to carry all the compounds

through to the methoxy stage (144) and purify the mixture

then. The crude mixture was treated with NaH/CH3I to

convert the alcohol into a methoxy group. This gave a

product mixture that was partially resolved by column

chromatography. It gave 4 mixtures of 2 major components

each. Examination by FT-IR and 13C APT spectroscopy








86

revealed that none of the 8 compounds could be the

desired alpha methoxy benzocyclobutene carboxylic acid

(144).

At this point we reexamined the second strategy.

This involved making the benzocyclobutene acyl complex

(135), decarbonylating it to the alkyl complex (136),

converting it to the benzocyclobutenylidene carbene

complex (137), and then adding methoxide to give the

alpha methoxybenzocyclobutene sigma complex (130).

0
(CO)5Re H (CO)sRe




136

135




(CO +Re' BF4 OCH3
PhCBF" ( (CO)sRe




There are two obvious problems associated with this

strategy. The first is that it requires a successful

decarbonylation from an acyl complex (135) to a sigma

complex (136) and in our hands none of our attempts had

been successful on rhenium pentacarbonyl systems. The









87

second is the intermediacy of the benzocyclobutenylidene

carbene complex (137) as mentioned above. Not only is

this not a known compound, there are to my knowledge no

known cationic rhenium pentacarbonyl carbenes. A

previous attempt to form methylidene rhenium

pentacarbonyl cationic complex was unsuccessful.48



Ph3C+BF4-
(CO)sRe-CH3 Ph3C-BF4- (CO)sRe-F-BF3






Finally the sequence requires that not only must the

benzocyclobutenylidene complex (137) be synthesized and

be at least moderately stable, but it must react with

methoxide to give the addition product (130).

Taking this into account we started the sequence.

The benzocyclobutene carboxylic acid (142) was treated

with oxalyl chloride to generate the corresponding acid

chloride (143). This went smoothly. Treatment of the

acid chloride (143) with rhenium pentacarbonyl anion (35)

(generated as before) gave the desired acyl (135) in a

clean reaction.









88



0 0

] CO2 C2H



142
142


(CO)sRe


(CO)5Re-
35


The acyl complex (135) was dissolved in toluene-D8 and

heated to 1050C. The thermolysis was monitored by 1H

spectroscopy. This showed that the decarbonylation went

smoothly and was complete in 90 minutes. The sigma

complex (136) decomposes only very slowly at this

temperature. This degradation is indicated by a slow

growth of peaks in the vinylic region of the 1H NMR

spectrum.


1050C
Toluene
-OD


H

(CO)sRe-









89

The next steps were more difficult. Addition of

trityl tetrafluoroborate to a methylene chloride solution

of the sigma complex (136) produces a deep red color.

This color gradually changed until after approximately

one hour the solution had become dark green. If one

follows this reaction by looking at the absorptions of

the terminal carbonyls in the IR spectrum of the mixture,

one sees a shift to higher wavenumbers after 15 minutes.

This is then followed by a change in the relative

intensities of these peaks, but no general, overall

shift. The addition of methoxide/methanol to this dark

red solution immediately changes the color to light

yellow.

Column chromatography revealed 4 types of compounds.

The first to be eluted was triphenylmethane (146). This

comes from the abstraction of a hydride by the trityl.

The yield of this is very high (e.g. 91%).




H
(CO)5Re
(CO)Re + Ph3C+ "BF4 Ph3C-H + ?

146













90

















-- -




















o,' -- --






I
--NNNNNNNNNNNNN2



NO

0
~-Q-00 n~rr~0----- OT ->



















C






O 0)
a -nr-nnIo
2 e











N U









34L | a ^~~~~- F.-JIs Mo
~~ i *
\ =l w
^ '~-~^ "
2 ^=-<







Y -- = = = = = -
^ ------
s c; 5







33N~iiinSNVYlZ













































--n- ---'



.~mn---rn~- 0
flWQofl-0-4QC
uee'.tw00 DJW.

~, 0
rll~n*I.-


SC It ,z i


-U


*.l
C






-,









.0




CO


40
541
0













4J








(11 Q


0






Oj






.-C
2

u

a J


.3


oit ig 968 L tog 0 L t'OS CsI SC
33MWVlIrSNVMlZ


aC 6it


tic 6t









92

The second compound to elute is trityl methyl ether

(147). This comes from the reaction of unreacted trityl

tetrafluoroborate with methanol.




Ph3C*'BF4 + MeOH Ph3C-OMe + HBF4

142


The third is the desired sigma complex of alpha

methoxybenzocyclobutene (130). This comes from the

reaction of the benzocyclobutenylidene carbene (137) with

either methanol (followed by deprotonation by methoxide)

or methoxide.



+
CO)5 Re' BF4 HOMe
MeOH (CO)sRe


148
137

'OMe OMe
\^


(CO)sI
















0-C












x



I -



Cm

0
oo
- |--o












-4
Sr






C I

-41
M r4
Ln




fX4^


w ~

ac
U_