The Use of Iron Carbonyl Complexes in Organic

Home , Cyclobutadiene, Diiron nonacarbonyl, Iron pentacarbonyl, Triiron dodecacarbonyl

THE USE OF IRON CARBONYL COMPLEXES

IN ORGANIC SYNTHESIS

a thesis presented by

GARY DAVID ANNIS

in partial fulfilment of the requirements

for the award of the degree of

DOCTOR OF PHILOSOPHY

OF THE

UNIVERSITY OF LONDON

WHIFFEN LABORATORY CHEMISTRY DEPARTMENT IMPERIAL COLLEGE LONDON SW7 2AY. AUGUST/ 1979. 1.

CONTENTS page

ABSTRACT 3

ACKNOWLEDGEMENTS 5

INTRODUCTION 6

1. CARBONYL INSERTION REACTIONS 8 (a)Sodium Tetracarbonylferrates 8 (b)Sodium Hydridotetracarbonylferrates 13 (c)Lithium Acyl Iron Complexes 14 (d)Magnesium Acyl Iron Complexes 15 (e)Potassium Tetracarbonylferrates 16 (f)Miscellaneous Ferrates 17

CARBONYL INSERTION REACTIONS USING DICARBONYL- PENTAHAPTOCYCLOPENTADIENYL IRON COMPLEXES 20•

CARBONYL INSERTION REACTIONS USING IRON CARBONYLS 25 (a)Reactions of Simple Vinyl Cyclopropanes with Iron Carbonyls 25 (b)The Reactions of More Complex Hydrocarbons with Iron Carbonyls 33 (c)Diene Complexes of Iron Carbonyls 38 (d)The Reaction of Hetero Systems with Iron Carbonyls 46 (e)Coupling of Olefins using Iron Carbonyls 52

2. RING FORMING REACTIONS USING IRON CARBONYLS 55

3. FUNCTIONAL GROUP REMOVAL AND REDUCTION USING IRON CARBONYLS 58

4. ISOMERISATION AND REARRANGEMENTS USING IRON CARBONYLS 61 2.

page

5. OTHER METHODS OF C—C BOND FORMATION USING IRON CARBONYLS 62

6. FUNCTIONAL GROUP PROTECTION USING IRON CARBONYLS 64

7. ACTIVATION OF ALKENES USING IRON CARBONYLS 66

REFERENCES 67

RESULTS AND DISCUSSION 77

Preparation of Lactones from Ferrolactones 78

Mass Spectral and N.m.r. Data of the Ferrolactones 99

Mechanism of Formation of Ferrolactones 104

Mechanism of the Oxidation of Ferrolactones 106

Structure of Iron Carbonyl Complexes 110

Preparation of Lactams from Ferrolactams 115

Mechanism for Formation and Oxidation of Ferrolactams 122

Preparation of NH Lactams 123

Miscellaneous Chemistry 126

EXPERIMENTAL 130

REFERENCES 162 3.

ABSTRACT

A number of ferrolactones have been prepared by the irradiation of vinyl epoxides in the presence of iron pentacarbonyl. In this way, syn- and anti-tricarbonyl{l,1',2'-n-l-vinylcyclopentan-l-yl)-2-oxycarbonyl iron (3) and (4), E- and Z-tricarbonyl-(1,2,3-n-l-nonen-3-yl)-4-oxycarbonyl iron (21) and (22), tricarbonyl-(2,3,4-n-2,3-dimethyl-3-buten-2-yl)-1- oxycarbonyl iron (27), tricarbonyl-(1',2',3'-n-1-propenylcyclohexan-1'- yl)-1-oxycarbonyl iron (35), tricarbonyl-(2,3,4-n-3-cyclohexen-2-yl)- 1-oxycarbonyl iron (41), and 2-tricarbonyl-(2,3,4-n-2,3-(tetramethylene- 3 -buten-2-yl)-1-oxycarbonyl iron (54) were prepared from their respective epoxides. A previously unobserved structural isomerisation has been shown to exist by the separation of syn- (3) and anti- (4).

Oxidation of ferrolactones with ceric ammonium nitrate has been shown to produce S- and 6-lactones. Thus, complexes (3), (4), (21), and (22), (27), (35), (41) and (54) have been converted to 1-vinyl-6-oxabicyclo- 3.2.0 - hept-7-one (6) and 2-oxabicyclo [4.4.0}dec-5-en-3-one (7), (7), trans- and cis-3-pentyl-4-vinyloxetan-2-one (23) and (24), 3-(isopropenyl)-3-methyloxetan- 2-one (28) and 3,6-dihydro-4,5-dimethyl-2-pyrone (29), cyclohexanespiro-4'- (3'--vinyloxetan-3'-one) (37), 7-oxabicyclo14.2.0] oct-2-en-8-one (44), and l 2-methylenecyclohexanespiro-4'-oxetan-2-one (56) and 5,6,7,8-tetrahydroisochroman-3-one (55) respectively.

Ferrolactones (3) and (4), (21) and (22), and (27) have been converted to their corresponding ferrolactams; tricarbonyl-(2',1,2-n-l-ethyl-l-cyclopenten-l'-yl)-2'-benzylaminocarbonyl iron (80), E=tricarbonyl-(2,3,4-n-3-nonen- 2-yl)-benzylaminocarbonyl iron (83), and tricarbonyl-(2,3,4-n-2,3-dimethyl- 3-buten-2-yl)-l-benzylaminocarbonyl iron (77) by the zinc chloride catalysed addition of benzylamine.

By analogy to the ferrolactones, the ferrolactams (80), (83), and (77) were converted to their respective lactam derivatives, 1-benzyl-3-(1'-cyclopent-. enyl)azetidinone (81), cis- and trans-l-benzyl-3-(1'-heptenyl)azetidino ne (84) and 1-benzyl-3-(isopropenyl)-3-methylazetidinone (79) and 1-benzyl-3,6-dihydro- 4,5'-dimethyl-2-pyridone (78) by oxidation with ceric ammonium nitrate in ethanol. V 4.

The lactone (6) and 3-(1-pentenyl-oxetan-2-one (90) have been rearranged to 1-vinyl-3-oxabicyclo[2.2.1hept-2-one (89) and 3-methyl-l-oxabicyclo L3.3.O0 oct-3-en-2-one (91) respectively, by thermolysis in benzene in the presence of zinc chloride. 5.

ACKNOWLEDGEMENTS

I would like to thank Dr. S.V. Ley for his advice, guidance, encouragement, and friendship throughout the course of this work. I thank Mr. K.T. Jones and his staff for the microanalytical service, Mrs. Lee for the mass spectrometry service, and Mrs. Day and Mrs. I. Hamblin for their service at the stores. I would also like to thank my friends (they know who they are) in the laboratory for their help, cooperation, and friendship.

Finally, I would like to thank the Science Research Council for a studentship for the period of this research. 6.

INTRODUCTION

Iron carbonyls are the least toxic and inexpensive transition metal

carbonyls available to the organic chemist. This goes some way in ex-

plaining the enormous interest shown in the reactions of these species

over the past few years. Three of these carbonyls, iron pentacarbonyl 1 (1), diiron nonacarbonyl (2) and the triiron dodecacarbonyl (3) have

found extensive applications in organic synthesis.

o `° -coo 0C—Fei—"CO `CO oc/Ī f-e4ō CO co co

(1) (3)

Elemental iron has 8 electrons in its valence shell (3d6, 4s2). By

coordinating to the equivalent of five 2-electron donors, it can achieve

the electronic configuration of the next inert gas, Krypton (3d10, 4s2,

4p6). This is the electronic configuration adopted by iron in all its

stable complexes, and the iron is said to be coordinatively saturated.

Removal of a 2-electron donor leaves 16 electrons in the valence shell,

and the species is said to be coordinatively unsaturated. Transitory

•reaction intermediates commonly have this configuration. Iron tetra-

carbonyl, the simplest of these species can be generated by thermolysis

-.or photolysis of iron pentacarbonyl=(1), and the thermolysis of diiron a - noncarbonyl (2) and triiron dodecacarbonyl (3). It reacts with a great

variety of organic compounds including olefins, dienes, a,6-unsaturated

carbonyl compounds, vinyl cyclopropanes, strained cyclopropanes and vinyl

epoxides forming pi or sigma bonds to complete its valence shell. 7.

Further ligands may be replacedin this manner to generate highly functionalised complexes. Saturatively coordinated complexes can also react with ligands, either in ligand substitution reactions or in reactions where the adding ligand forces a carbonyl ligand to insert into an iron sigma bond. Because of the property of iron tetracarbonyl to form complexes with electron rich systems, iron carbonyl complexes are finding an increasing number of applications in organic synthesis.3

Many products formed from iron carbonyl complexes are difficult to make efficiently by other methods.

This review will deal with carbonyl insertion reactions of iron complexes in detail. However, in order to demonstrate the potential of iron complexes in synthesis, examples of other important reactions will be highlighted in later sections. 8.

I. CARBONYL INSERTION REACTIONS

CARBONYL INSERTION REACTIONS USING IRON CARBONYL ANIONS

(a) Sodium Tetracarbonylferrates

Disodium tetracarbonylferrate is a nucleophilic reagent which sub- stitutes alkyl halides/tosylates or acyl halides to give two different types of complex, both of which have been characterised.2 The reactions of these complexes are summarised in Figure 1.

FIGURE 1

RH Na t Fe (CO)4 RC HO ,,,...00I1 H+ Fri

0 R'X OC OC — e CO

CO \ fL L (CO or P73) \X21 ROH XZlROFi~ 02 i rho r 02/ HZO ROR RCO H R cod 2

X = Halogen , R'= R= Alkyl

In this way, disodium tetracarbonylferrate can transform alkyl halides/ tosylates and acyl halides to aldehydes4, ketones5'6, acids?, esters7 and amides 7'8 (Figure 2). Of particular importance is the fact that the reagent shows a high selectivity for halides in the presence of other 9.

functional groups. FIGURE 2 1/. a CHO 81% 2/. H+

02Et 1/. a Br 2Z EtI

11 Ts 21. MeI • 79% 99% Optical yield 1/.a NC NC 2/ C F COCQ 715

11. Fe(C0)2-

2tEtI 81 %

1/. a Br 02H Cl 21. IH~ Ōl O-2 2 84%

1/. Fe(C0) 2/. I2I McOH

~Br 1/ 0 NMe2 2/. I2/ Me2NH Na2Fe(CO)4 / CO 80 %

The reagent is also useful for the preparation of polyfluorinated ketones which are difficult to make by conventional methods.6

Br 1/..a Cl 2/ C7F,sCOQCt 10.

There is only one example of an amide being prepared from the addition of an amine to an acyl complex, although a more general method has been reported, in which the acyl complex is added to an alkyl or aryl nitro compound.9 The nitro group is spontaneously reduced to the correct oxidation state during the course of the reaction.

H0 Me CO Fe(CO)4 Na + t NO2 --~ - Me CO NH~

As disodium tetracarbonylferrate is toxic, has a basicity equivalent to the hydroxide anion, and is spontaneously flammable in air, its synthetic use is limited. Initial substrate choice is also limited to primary halides and secondary tosylates, as competing elimination reactions and steric hindrance are important considerations. Allylic halides produce 1,3-diene complexes by elimination of hydrogen bromide. Migratory insertion fails for alkyl groups with adjacent electronegative groups.

For the second substitution reaction substrates are limited to a primary halide, usually (but not always) the iodide.

Acyl iron complexes can also be generated by addition of an anion to iron pentacarbonyl. Thus, treatment of iron pentacarbonyl with an alkoxide followed by alkylation affords esters.10 BrCH2CH2CO2Et CH2CO2Et Na OEt + Fe(CO)5 --~ Et0 CO CH2 48 This method of preparing esters conveniently avoids the use of the more esoteric disodium tetracarbonylferrate.

The use of other ligands, such as ethylene instead of carbon monoxide or triphenyl phosphine,to promote alkyl migration sometimes results in the formation of ethyl ketones.11 Studies using deuterated substrates indicate the formation of an intermediate iron complex (4). Treatment of 11.

complexes which are coordinatively saturated, (i.e. (5)) with carbon monoxide do not react with ethylene to give ethyl ketones. The reaction of alkyl iron complexes with olefins is not limited to simple olefins. Recently a,$-unsaturated carbonyl compounds were shown 12 to act as Michael acceptors when treated with alkyl complexes.

Na Fe(CO)4 0

OTs I I

(4)

A?e(C0) 4 (5)

+ NaFe(CO)2 4

/7 C0 Et 2.5 eq . 02Et 2.

92% A major disadvantage of this pathway as a synthetic method is the requirement for two equivalents of the olefin. Cyclisations have also been brought about in a similar manner, in which the above problem has been overcome, by the addition of acid rather than another equivalent of olefin.13

12.

6 + Na2F e (C04 --~- Fe (C0)4

c>0 2.5 eq.

94 %

OTs 1/. Na2Fe(C0)4

2/. H+ 1

Mechanism

The mechanism of the reactions of disodium tetracarbonylferrate has .14,15 recently been the subject of two studies Earlier work$'16 indica-

ted that disodium tetracarbonylferrate has a basicity equivalent to OH

and exhibits classical SN2 type reactivity. The reactivity of the sub-

strate decreases in the order Me > RCH2 > RR'CH and RI > RBr > ROTs> 17,18 RC1, stereochemical inversion being observed at the carbon site attacked.

.Increase of the solvent polarity gives improved dissociation of the ion

pair and thus results in an increased rate of reaction.

17,18 Migratory insertion is known to occur with retention, However,

the rate of insertion is decreased in more polar solvents, and depends

on the counter cation used. Hence the rate of reaction decreases in the

13.

order Li+ > Na » (Ph3P)2N+. This result seems to indicate that the

reaction species is a tight ion pair.

Final product formation is believed to involve a reductive elimina- 8 tion from a 6-coordinate iron species (6).

e co ,Co oc~ RX or OC--- Fe Fe H+ ~ eco OCA ‘s%'R(H) CO CO (6) Fe(CO)4 + RCR II 0 (b) Sodium Hydridotetracarboulferrate

In contrast to the addition of alkyl complexes to a,s-unsaturated

.compounds, sodium hydridotetracarbonylferrate adds (albeit in low yield)

to give a-substituted carbonyl compounds. Insertion followed by alkyla-

tion of the intermediate complex produces 1,3-dicarbonyl compounds.19

e \— ,, O2Me THF Na H Fe(CO) Na 4 + 30%

~MeI / THF/ NMP =20Me c 67 % 14.

Both alkyl and aryl lithiums readily react with iron pentacarbonyl

to give acyl iron complexes. Subsequent alkylation gave ketones in

moderate yield as expected.20

BzBr Li + + Fe (C0) 5 57%

21 Aryl aldehydes can also be prepared on work up with acetic acid.

Low yields of %-diketones were obtained when lithium acyl complexes were

added to acyl halides.22 However, a far more elegant route to these

compounds involved addition of an acyl anion equivalent to iron penta-

carbonyl. Alkylation of the resulting acyl complex gave a-diketones in 23 moderate to good yields. /XH (CH2 )2\ R CHO XH X=O,S

_/ 1/. nBu Li / 2/. Fe(CO)5 H'/ H2O / 3/. R'I

R` CI

Placing a carbonyl substituent at the 3-position of a pyridine ring

is synthetically difficult to achieve.24 However, when pyridine was

treated with phenyl lithium and the resulting anion treated with iron

pentacarbonyl, aldehydic (7),carboxylic(8) or ester (9) substituents

were obtained depending on the work up conditions used.19 15.

Li Li+. Fe (C 0)5

I21 H20 1RX/H+ 0 D

OR N

(7) 73% (B) 50% (9) 2432%

(d) Magnesium Acyl Iron Complexes

Although lithium alkyls have long been known to add to iron penta- 21 carbonyl, only recently have Grignard reagents been used as the attack-

ing anions. Thus, alkyl or aryl magnesium bromides can be converted to 26 ketones25 or esters. The yields are good, and again the use of the

'pyrophoric disodium tetracarbonylferrate was avoided.

There are very few reports of hetero anions adding to iron penta-

carbonyl. N-magnesium bromides add to iron pentacarbonyl to give an

as yet unidentified intermediate which produces NN'-disubstituted alkyl 27 and aryl ureas,when.treated with nitro compounds, in excellent yield.

This method gives far better yields and is more general than that reported

CbH~ NO2 NH Mg + Fe(CO)5 ---~ Br H+ 16.

by Igbal28, whose method involves the addition of aryl azides to iron pentacarbonyl.

(e) Potassium tetracarboulferrates

The use of potassium tetracarbonylferrates as synthetic reagents has not been investigated to the same extent as other ferrates. The method of preparation is different to that used for disodium tetracarbonylferrate.

The addition of one or two equivalents of potassium hydroxide to a mixture of iron pentacarbonyl and barium hydroxide gives potassium hydridotetracarbonylferrate and dipotassium tetracarbonylferrate respectively.29

Like sodium hydridotetracarbonylferrate, potassium hydridotetracarbonyl- 30 ferrate has been found to add to a,•6-unsaturated carbonyl compounds.

1/. K H Fe(CO) OEt OEt 2l. Ill EtOH O Et

97 3

Et 1/. K H Fe(CO)4 OEt + 2/. I2/ EtOH OEt

89 7

4-. _ Et O Et 4

The regioselectivity of the reaction could be rationalised by the non-reversible addition of the nucleophile to form an intermediate iron complex (10). 17.

{LO C ) Fe H Fe (C0)4 + Et OEt

(10) OH K2Fe(CO)4 OEt Et OH

31 32 Epoxidesp are openedp to give carboxy lated products in 'low yield',competing reactions giving rise to olefins. This

reaction is known to proceed in high yield with cobalt carbonyl analogues 33 on similar substrates.

(f) Miscellaneous Ferrates

for Other cations have been used as counter ions hydridotetracarbo-

nylferrates. For example, the bisltriphenylphosphinel iminium salt

-adds to a variety of substituted acetylenes in excellent yield.34

Methylation of the intermediate ferro-ketone anion (11) followed by

oxidation,' first of the iron c-bond then of the iron it-bond results in

the formation of highly functionalised products.

18.

C Ct ((I)3P)2NHFe(C0)4+ MeOC CO2Me 2 95 o/o Fe (C0)3 e (11) .,-N, /'Me OS 0 F 2 H (CO2Me I H CO2Me 2

Me0 C 1 CO2Me Me0 C ~, i OMe ' 2 2 PPFe(C0)3 Fe (C0) 3

C ein/Me OH\ Me CO2Me H H Me02C CO2Me

. It is uncertain whether the methanol adds in a Michael fashion to

an uncomplexed product, or traps a cation generated during the oxidation.

Another unique reaction is the coupling of acyl halides with diiron non-

acarbonyl to give ketones together with small amounts of aldehyde formed

as a minor by-product.37 The mechanism of this reaction is not yet

nC H COCl + Fe (C0) -H— ~C H CHO + (~ H )2C0 715 2 9 7 15 ~15 4% 54%

clear but remarkably no coupled products are formed when the reaction

is repeated with other metal carbonyls, Fe(CO)5/hv, Fe3(C0)12, Co(CO)8,

or Fe2(CO)6(Pie 2)2. Also, no reaction occurs with alkyl halides. It

could be that a dinuclear iron complex is involved perhaps similar to

that recently reported by Collman.38 This is the only reported carbonyl

insertion by a dinuclear iron complex.

19.

Pth 2 RX P~2 e Nat (CO) Fē Fe(CO) (CO) Fe Fe(CO) 3 3 31 Z 3

P~2

(I)2 \ 1)2 (C 0) ē ~,Fe(C0)3 ~-- (CO) ē(CO)3 2e

P(1)2 P(1)2 •

A major problem with all reactions involving ferrates is the ex-

traction of the products from iron containing species. This problem

can be overcome in certain instances by use of polymer supported iron

speciec.39 The halide counter ion seems to assist the alkyl insertion.

Iron containing products are removed on the resin.

\+ Me 3 H Fe(C O) 4

C0/ fl 7 H15Br

H,15C H O 90 % n C7

Several reviews on iron complexes exist which discuss parts of the 8 40-42 above work. ' 20.

CARBONYL INSERTION. REACTIONS USING DICARBONYLPENTAHAPTOCYCLOPENTADIENYL

IRON COMPLEXES

An enormous amount of information is available about these complexes, the inorganic aspects of which were recently reviewed.43 Carbonyl insertion products can be formed in a number of ways.

Carbon monoxide induced insertion reactions often require the use of extremely high pressures, although identical complexes can bē prepared by treatment of acyl chlorides with the sodium salt of dicarbonylpentā- haptocyclopentadienyl iron.44 Cyanide, isonitriles, olefins45 and

CO Cp Fe (C0)2Me C Fe(CO)2 C O Me

CP = Na (Fe CP(CO)21 + McCOCI certain sulphur compounds are known to cause the insertion of the carbon monoxide ligand whereas, species such as I-, SEt2, 4NH2,p-C1$NH2, 43 C5H5N and C6H11NH2 do not.

CP Fe (CO)2CH2CR2 CH-C H2

Me S CH2 H~.C( Na ECPFe(CO)2 ] (CO)CPF MeS

21'.

Cationic complexes can also react with nucleophilic anions to give carbonyl insertion complexes.46

Na Ō

Perhaps the most important method of inducing carbonyl insertions is 47 by oxidation of alkyl dicarbonylpentahaptocyclopentadienyl iron complexes.

Fe(C0) 2 IV Ce

Me OH

This reaction is quite general and similar transformations result 47 when analogous complexes of Cobalt, Molybdenum and Tungsten are oxidised.

22.

The mechanism of this and other insertion reactions involving

dicarbonylpentahaptocyclopentadienyl iron has been the subject of

detailed investigation. A study of the stereochemistry of the initial

nucleophilic attack of the dicarbonylpentahaptocyciopentadienyl iron 17'18,48 anion, has shown that inversion occurs at the carbon site attacked. 17,18 Alkyl migration can then be induced by either ligand addition or 18'48 oxidation and retention of configuration at the carbon centre

migrating is observed in every case examined. It seems likely that esters • are produced by the displacement of a positively charged iron species,

from the acyl group by nucleophilic attack of the solvent. The yields

of ester produced have been shown to decrease in the order

t t tBu H~ /~DD ~u Bs D~~D . . Inv. Nat FeCO),Cp] H 0~e(CO)C ``1nv. p Ret. teu tBuNC I2/CS2 Me3CNC H Dk )`H R Inv. u I~ Fe(CO) C tBu iD Ce /MeOH 2 P B i D ''~~~H or 02IMe0H ~H Pentane D -~H CQZMe Br Ret.~Cl2! Br /~Ret. ~ 2 COG3 Me OH t u D H Di3/4 COC! CO2Me 23.

/~ NW-KO) C 1 --( )----OSO2(1) 2 P—~ }—~Fe(C0)2CP

Cu 2+ (12) EtOH f

CO Et 2

MeOR > Et0H > Pr©H > tBuOH, for substrates similar to (12), as expected:48

0 A potentially useful application for hydrlazulene synthesis, employing the oxidative insertion reaction and the well known ability of tricarbonyipentahapto iron cations to alkylate, has been reported.49 H F+P Fp

Fe(CO)3 Fe (C0) Fe(C0)3 ' 75% 3 .1 /NaBH4

IV Ce CO -m-- Me OH .- H 61 % Fp = -- Fe (C0)2 Cp

The presence of the bulky iron tricarbonyl group in the dienyl complex results in the attack of the alkylating agent on the opposite face of the ring, preferential cis-closure of the five membered ring resulting in the required stereochemistry.49 Oxidative removal of the 48,49 iron complex from the ring system gives an ester. 24,

The addition of nucleophiles to, olefins coordinated to transition metals is well known and has been recently reviewed.50 Addition of amines to olefins coordinated to dicarbonylpentahaptocyclopentadienyl iron, followed by oxidatively induced carbonyl insertion and elimination has 51 been reported to produce a-lactams,

Bz NH2 Fe (C0)2 ~ Bz Fe (CO) -24°C

THS H2Bz Bu3P /1/.NaOH/ 0°C i 2/. Pb 02 + KCO Fe(C O22 BzNi~ e H Ag0 NH Bz N2 69 % V69 % Ct -78°C 47%

Cyclisations are also possible and some novel compounds have been pre- 51 pared in a similar manner.

80% H3 Fp •BF 4 1/. nBu3N 21. K O tBu Ag2 0 Fp

30 % Overall NH 25.

Cyclisations have also been observed during other ligand induced carbon- 52 yl insertion reactions, although no yields were reported.

Fe(CO)2Na

Fe (C0)2

Insertions of carbon monoxide into Fe-N bonds does not occur by 54 ligand induced53 methods. However, carbamoyl complexes are well known

but little synthetic use has been made of them.

CARBONYL INSERTION REACTIONS USING IRON CARBONYLS

(a) Reactions of Simple Vinyl Cyclopropanes with Iron Carbonyls

One of the earliest reports55 of vinyl cyclopropanes reacting

with iron carbonyls was the treatment of cyclopropyl styrene derivatives

_,with iron pentacarbonyl at 1400. At these high temperatures, dienes

were formed as all the intermediate iron complexes decomposed before

carbonyl insertion could take place. However, photolysis of the

same systems with iron pentacarbonyl gave cyclohexanone derivatives 56 in low yield.

26.

30 %

However, alkyl vinyl cyclopropyl analogues can result in fused ring systems in much better yield.

Fe(C0)c

hv cc7 Fe (C0)4 70 % (13)

A mechanistically significant minor product was formed during the reaction of the cyclopentane derivative which was shown to be the olefin-tetracarbonyl iron complex (13). Dicyclopropyl vinyl derivatives also give cyclohexanone derivatives.57,58

// `\

Heat or e(C0)3 - + Fe(C0)5 hv

More complex derivatives and photolytic conditions lend greater stability to the iron containing intermediates.57,58

27.

h v

Fe(CO)5

Fe(CO)3 Fe(CO)5

Larger rings have also been synthesised by this method.59

hv Fe(CO)5

A major criticism of the majority of this work is the lack of information given about yields, •and the lack of mechanistic detail, although other workers have investigated similar reactions in greater detail. For example, an excellent piece of investigative chemistry was reported by 60 Aumann. It was shown that the tetracarbonyl iron - olefin complex (14) could not be converted photolytically or thermally to the c,'rr-allyl tricarbonyl iron complex (15). This would seem to imply that the initial reaction site for the formation of the complex (15) is the cyclopropane 28.

hv 1-7 + Fe(CO)5 _ Fe (CO)3 D (14) (15)

CO 20°C* CO / Atm. /20 Atm.

I S\—+ //7) Fe Fe (C O)3 (C0)3 0 76 B0%

ring. The formation of larger amounts of the complex (14) relative to the compound (15) implies that the double bond is the more reactive ligand to iron in this case. The complex (15) decarbonylates reversibly in solution to give a complex (16). High pressure exhaustive carbonylation of complex (15) gives a cyclohexenone. Most compounds containing the unit as in (15) react in this way. If the cyclopropyl ring is fused to a cyclopentane ring, increased strain enhances the reactivity of the cyclopropane ring so that the major product is the insertion product 61 (17). Thermolysis of the carbonyl insertion product results in decarbonylation and formation of the diene complex (18). Thus, this type of complex can be formed either from the olefin-tetracarbonyl iron complex or from a,7r-allyl tricarbonyliron complexes (17).

29.

Fe2(C0)9 + 85% Fe(CO) Fe(C0)4 3 1 10 (17)

Similar results are obtained for the analogous higher homologue.

a-Thujene has also been shown to give a c,ir-allyl carbonyl inser- 62 tion complex with iron pentacarbonyl.

hv + Fe (C0)5 . 0 z Fe(CO)3

55-62%

However, cyclopentadienes with spiro cyclopropane substituents produce , dicarbonylpentahaptocyclopentadienyl iron complexes•63 64

Fe2(C0)9 e(C0)3 0

Fe2(C0)9 35% 30.

Only one example of an insertion reaction involving a vinyl cyclobutane derivative has been reported.65 Thus, a or 0-pinene react under rather vigorous conditions to give insertion products. It seems that decarbonylation of the olefin-tetracarbonyl iron complex occurs when the decomposition pathway to a diene complex is blocked.

Fe(C0)5 -CO

160 °C Fe(CO)4

The coordinatively unsaturated iron complex (19) can then undergo further reaction.

Cyclopropyl acetylenes on reaction with iron pentacarbonyl conversely 66,67 give coupled products.

This type of coupling reaction is well known for unsubstituted acetyl- . 68 nes. 31.

Fe(CO)

28 %

Mechanism

From these simple examples, much can be learned about the mechanism.

An electron rich cyclopropane ring is required for reaction with iron

tetracarbonyl. Thus, 1-phenyl cyclopropane or vinyl cyclopropanes with

nitrile (electron withdrawing) substituents on either the double bond or

the cyclopropane ring do not react.69 A cyclopropane conjugated to a ketone also fails to react. Ring strain plays an important part 60 61 in determining which product will be formed. '

Although a detailed study of the mechanism has not been made, several points become clear (Figure 3). 32.

r FIGURE 3

-\ Fe(CO)4

\T"-\. O Fe(CO)3

Fe(CO)4 (23)

CO4A Fe(CO)4

(20) Fe(CO) \ V 3 L Fe(CO)3

(24) (25)

Direct insertion of the iron tetracarbonyl into a cyclopropane ring could result in an intermediate (21). Metallocycles similar to (21) are 70-72 known and their presence has also been inferred as reaction intermediates. It is impossible to say whether the complex (23) is formed by

Coordination of the olefin to the unsaturated iron in (22) or directly by intramolecular ligand induced carbonyl insertion.

Thermal decarbonylation of (23) followed by a metal catalysed 1,3- hydrogen shift in (25) gives the diene complex.61 The olefin-tetracarbonyl iron complex can be decarboxylated thermally65 and sometimes photolytically

75,76 presumably to give an intermediate (24). This unsaturated species 75,76 then inserts into the cyclopropane ring to give the complex (25). 33.

(b) The Reactions of More Complex Hydrocarbons with Iron Carbonyls

Semibullvalene (26) reacts with iron pentacarbonyl/hv77 or diiron 78,79 nonacarbony l/L to give metal complexes. The 6or-allyltricarbonyl iron carbonyl insertion complex (27) is not stable, in this case and the product isolated is the decarbonylated tricarbonyl iron complex (28).

(26) (27) (28)

A general investigation into the stability of the carbonylated complexes with different bridge substituents has been made.80 Other authors

hv A

CO Fe(CO)3 Fe(CO) e(CO) 5 3 Atm. Ref. 81 prootdūcts-

Fe2(CO)9 n=21 75% Ref. 82 n= 80%

hv CO 0 Fe(CO)5 e(CO)3 Atm. e(C0)3 Ref. 80,83 hv 0 Ref. 84 Fe W 5 e (C0)3 34.

Fe(C0)5 Ref. 85,86 hv -60°C e(C0)3 Fe(CO)4 39 % 20% (29) (30) have reported reactions of analogues of semibullvalene under various conditions. This last example is of particular interest as unusual products were obtained from (29) and (30).

Fe (CO) re(CO)4 (30) (31) (32) -CO

+CO

Fe(C0)3 -CO

Fee(C0)3I

(34)

Fe(CO)3

Fe (CO)3

(36) (37) (38)

COOMe

• e (39J Fe(C0)3 (40) Fe(C0) (41) 3 35.

Photolysis of (30) at -60°C gives (29) in low yield via the usual pathway while pyrolysis of (30) under a carbon monoxide atmosphere at high pressure to prevent thermal decarbonylation) , produces (31) in excellent yield. This compound has also been prepared by direct inser- 72 tion of iron tetracarbonyl into (32).

The a,i-allyl carbonyl insertion complex (29) on thermolysis produces the complex (37). It is known that in structures similar to (33) the bridgehead hydrogen cannot be removed by the iron moiety. For this reason, the usual decomposition pathway - to produce diene complexes - is blocked. Radical cleavage of the intermediate (33) could produce a carbene complex (34). Insertion of carbon monoxide into this complex and rearrangement of bonds produces (35) and (36). Saturated carbonylation of (37) does not produce the usual carbonyl insertion product. Vi- gorous conditions are needed (150 atm. carbon monoxide, 175°C, 12 h), and the product isōlated.is the vinyl cyclopropane (38). This is one of the few. examples where exhaustive carbonylation of a cor-allyl complex regen- erates a vinyl cyclopropane.

Further evidence for the intermediates (33)-(36) is obtained when

(29) is pyrolysed in the presence of methanol. Addition of methanol to the ketene complexes (35) or (36) produces (39) (20%) and (40) (60%). Exhaus- tive carbonylation of (40) gives the expected ketone (41) (67%).

Where both double bonds of semibullvalene are blocked, iron tetra- 70,71 carbonyl has been shown to insert into the cyclopropane ring. Carbon monoxide inserts into the so formed iron complex at atmospheric pressure.

Higher pressures of carbon monoxide regenerate the cyclopropane ring.

This reaction seems to be quite general for strained unconjugated ' 74 cyclopropane rings, 72 and some impressive yields have been reported. 36.

Fe2(C0)9 Fe(CO)4

{0~ 60 Atm.

hv CO

Fe(C0)5 95 % 85 % e(C0)4

Fe (CO)s Fe 86% (CO)1f

The carbon monoxide required for the insertion to take place is generated in the reaction mixture by photolysis of iron pentacarbonyl.

One of the few tetallo-anhydrides was prepared via one of these metallocycles.

Bullvalene

It has been known for some time that builvalene (42) reacts with 37.

iron carbonyls. Recently, the reactions of this hydrocarbon with 87-90 iron carbonyls was the subject of several studies, Photolysis or

thermolysis of iron carbonyls with bullvalene resulted in a whole array

of products. An unusual a,Ir-allyl complex (44) was isolated from the

photolysis reaction. This was formed by a rearrangement of the unsatura-

. ted iron species into the cyclopropane ring (43).

More interesting products were obtained from the thermal reaction, one

of which lends itself to the synthesis of complex hydrocarbon skeletons.

91,92 (Figure 5). This is a good demonstration of the flexibility and

adaptability of iron in the synthesis of hydrocarbons.

Photolysis of the tricyclic hydrocarbon (45) with iron pentacarbony1

gives the usual iron complexe (46)35 However, when this complex was

oxidised with ceric ammonium nitrate, lactones were formed. This reaction

is an oxidatively induced iron extrusion, but the unique feature is

the.insertion of an oxygen atom to form the lactone. It is not yet clear

where this atom comes from.

hv Fe(CO)5 (45)

- 20 % 20%. FIGURE 5

Fe(CO)5 (43) (44) 59% + other A products (42)Fe2(C (CO) Fe~~. 3 (C0) (C0)3 Fe(C0) CO 3 39% 3 26°f° 100Atr. 80°C. /` 89% 3(OC)Fe Fe(C0)3

3(OC)Fe 11. FeCt3 20%

21.55°C. q

CO 100C. 130Atm

95% Fe 110°C (CO) 70/c 3 B F3• Et20 99°l< CO 100 At m. 38.

Iron tricarbonyl complexes of dienes have been found to react

Under a variety of conditions-to give carbonyl insertion products.

For example, treatment of diene complexes with Lewis 93 acids such as aluminium trichloride gives ketones in moderate yields.

However, when the iron complex of cyclooctatetraene was treated in a similar manner, an unusual product (47) was formed. The yield of this product was

Ct3 A- r CO Fe(C01 40 % 3 Fe (CO) 2 9 (4.7) 90%

not increased by the use of a carbon monoxide atmosphere. This compound is identical to that formed by the treatment of barbaralone94'95 with diiron nonacarbonyl. Exhaustive carbonylation of this product affords barbaralone 96,97 in high yield. The mechanism of this reaction is thought to involve initial acylation of the cyclooctatetraene complex. The yield of the

Fe (CO) COT 2

Fe (C0) COT 2 Fe(C0)3 Fe(CO)3 -

COT -cyclooctatetraene(CO)3

39.

intermediate (47) has subsequently been increased to 65 8 by the use of

a carbon monoxide atmosphere so the mechanism will almost certainly need

some revision. The use of a carbon monoxide atmosphere also effects

some interesting transformations in simple diene complexes.98

Fe(C0)3

It is well known that tricarbonyl iron-diene complexes with conjugated

double bonds will react with electrophiles. Protonation of such complexes

leads to stable tricarbonylpentahapto iron cations.99 These complexes can

further react with a nucleophile in one of two ways. Sodium borohydride,

for example can add to give a diene complex or a a,ir-allyl complex, which

can be readily converted to its carbonyl insertion counterpart.100,101

e(C0)3 Fe(C0)3 CO/ 40°C.

e (CO)3 t he formation of these two compounds is easily rationalised if the catio-

nic tricarbonylpentahapto iron is considered as a 2a,ir-allyl system

(48) .

40.

+ C.

+Fe(C01 Fe(CO) `Fē(COl 3 3 3 (48) (49) (50)

Displacement of the iron from an spa carbon breaking one of the a-bonds,

gives a a,n-allyl system (49) which collapses to a diene complex. Addi-

tion of the hydride to one end of the it-allyl system results in a 2a,ir-

system (50), which collapses to a a,ir-allyl system. An elegant synthesis 102 of bridged ketones has been reported using this approach.

Fe(C0)3 lico

CO Fe(C0)3

57 Vo

102 A similar reaction using the tricarbenylcyclooctatetraene complex proceeds by an entirely different pathway. Protonation, followed by reduction with sodium borohydride gave the tricarbonylbicyclo.j 5.1.0] octa-2c4,-diene iron complex (51) as the major product. Thermolysis pf this complex under a carbon monoxide atmosphere gave bicyclo13.3.1l nona-2,6-diene-9-one 41.

Fe(CO) CO 3 Fe(CO) Fe(CO)3 130A m. (51) 3 \FeCE3 900C.

Fe(CO)3

Fe2(CO)9 (55) (54)

Fe(CO)3

0 (53) 92% (52) + (51)

,102 (53) in 92% yield Presumably this was produced via the usual oor-allyl

iron insertion complex (54) and its carbonylated derivative (55). Bicyclo

[5.1.O}octa_24_diene was obtained by ferric chloride oxidation of (51).

When this diene was treated with diiron nonacarbonyl the complex (51) was regenerated together with some tricarbonyl bicyclol4.2.11nona-2,4-diene- i ron 7-onel(52). The formation of this product has prompted further investiga- I tion of the bicyclo [5.1.0, octa-2,4-diene system and its reactions with 103 iron carbonyls. It has been found that this system exists in two con- formations, open and closed, and the complexation of the diene moiety

OPEN CLOSED 42.

results in the formation of two iron complexes, exo (5 6 ) and endo (5 8) from the two forms respectively. The ratio of the exo to endo complexes

.%!N X e(CO) 3 EXO (56)

N X ENDO (58) varies with the substituent X, thus for X = CH2 the ratio is 6:1, similarly for'X = CO the ratio is 3:1, and for X = NCO2Et the ratio is 1:1.

Attempts to convert the diene complexes (56) and (58) to their respective o,n-allyl complexes (57) and (59) often results in other products being isolated due to the lability of these species. However, several a,i-allyl 103 Rhodium analogues of (57) and (59) have been isolated and characterised.

Pyrolysis of the diene complexes (56) and (58) under a carbon monoxide atmosphere, results in the isolation of the exhaustive carbonylation products (60) (61) (62).

The exo complexes (56) have been shown to produce the bicyclo [3.3.1 nona-2,6-diene-9-one system (53). In a similar manner the exo complex

X = NCO2Et gave the.bicyclo[3.3.1] nona-2,6-diene-9-one derivative (60). 43.

X 4(C0)3 4(C0)3 Fe(CO) COIj \co Z

D Fe(C0)3 /CO *C.O\ Fe(CO) CO X _ (60) X=NCO2Et

A similar pathway was proposed for the reaction of the endo complexes with carbon monoxide.

Co 0

e(\0)3 C 0 ~' 0

(61) 44.

For X = NCO2Et the major product isolated was (62), with only smaller amounts of (61) being formed. However, if bicyc.lo[5.l.Olocta-2,4-diene J derivatives were treated with Rhodium complexes the compound (61) was the major product.

Other related systems have also been examined. Ethyl-4-azabicyclo

[5.1.Ojocta_2,5_cilene_4_carboxYlate (63) produces several complexes when 104 irradiated with u.v. light in the presence of iron pentacarbonyl.

RN RN

(63) hv / F e (CO)5 Fe(CO)3

(6 4) ~. hv N Fe(C0)3 Fe(C0)3 "eR (66) CO RN RN

by / \1 (65') (65) Fe(CO)3 Fe(CO)3 A CO RN A)

R= CO2Et

Degenerate valence isomerisms between (64) and (64'), and (65) and (65'), have been shown•to exist by examination of their low temperature n.m.r. - spectra. Complexes (64) and (65) are converted in good yield to ethyl-9= 45.

oxo-2-azabicyclo[3.3. 1]nona_37_2_carboxylate (66) under exhaustive

carboxylation conditions.

Dienophiles have been found to add to cyclooctatetraene iron complex-

es in two ways. The first method involves a 4 + 2 cycloaddition while the

second is more complex, and can result in the formation of carbonyl in- 36, 105 sertion complexes. The addition of tetracyanoethylene to the cyclo-

octatetraene complex (67), and oxidation of the resulting reaction mix-

ture gave (68) as a minor product. It was thought that this may have been

produced from a ar,1r-allyl complex (69) by an oxidatively induced carbonyl

insertion, and iron extrusion. A far more probable mechanism is a

(NC) CeIV (NC) TCNE ~2 (69) 2(NC) TCNE 2(NC)

Fe(C0)3 (68) TCNE (67) je IV or C e

.tetracyanoethylene induced carbonyl insertion, to give a complex (70),

followed by either oxidative iron extrusion or tetracyanoethylene induced

extrusion, analogous to exhaustive carbonylation. A ligand induced

carbonyl insertion and iron extrusion are thought to effect the formation

of the ketone (74), directly from the reaction of N-methyltriazoline- 105 dione with the tricarbonylcyclooctatetraene iron complex (71). The

_ most probable mechanism for both these cycloadditions involves an 46.

electrophilic addition of the dienophile to the cyclooctatetraene complex.

• e

.88

Closure of the intermediate (72) in a known fashion,101 produces a a,ir- allyl complex (73). Ligand induced carbonyl insertion and iron extrusion gives the ketone (74).

(d) The Reaction of Hetero Systems with Iron Carbonyls

Some time ago, it was found that 1-chloro-4-hydroxybut-2-enes, or 106 1,4-dihydroxybut — 2-enes reacted with iron pentacarbonyl photolytically, 107 or diiron nonacarbonyl thermally to give (1,7-allyl carbonyl insertion complexes (ferrōlactones), in low yield. 47.

Fe2(C0)9

21 %

Since then, other workers have found that simple Diels Alder adducts of 108 nitroso compounds, and 1,4-amino alcohols undergo similar reactions.

(Figure 6). It has been shown that the oxazine ring in (75) is first reduced to the amino alcohol (78), which then further reacts'to give the ferrolactone (76).108 The mechanism for the reaction remains obscure, although the presence of water is important to effect the initial reduction.

If saturated oxazinesv/ere used N-phenyl tetramethylene carbamates (85)

N Fe2 (C09) ~~NH I I + 0 H20

08 and Imino alcohols were produced 1

Treatment of ferrolactone complexes (76)and(79)with amines in the presence of alumina gave ferrolactams (77a, b)and(80) in moderate 108, 109 yield. It remains uncertain as to whether the Lewis acid or hydro- scopic properties of alumina are responsible for inducing this reaction.

The bridged oxazine (81) has been reported to react with diiron nonacarbonyl under anhydrous conditions to produce a whole array of 110, 111 products including a S-lactam (82) in low yield. A certain specifically blocked bridged oxazine (83) has been shown to give ferra- lactam complex (84) directly, in contrast to the unbridged oxazine systems lll studied, The ferrolactam (84) is reported to be extremely labile, in

FIGURE 6

Fe2(C0)9 A1 0 2 3 r"-Fe(CO) H2O 3 1I. ()NH 3 2 2/.MeNH (75) (76) 2 (77)R 3/.HNH 4e2(C0)9 2 a L R= 4 0% NH b/.R=Me 45% HU OH 2 c/.R= H 0%

(78)

NH F-Fe(CO) OH 3 A1203 50%

(79) • (80) NH(I) NH(I) Fe2(C0)9 +

(81) (82) 8% 6% . 80/0

"N e (C0 25%

Fe (C0) 2 9 +

(83) 48.

,112 contrast to other reports of similar compounds Thermolysis of this complex yields 6-and S-lactams presumably in low yield as no figure is quoted.

Mechanism No explanation has been given as to why in unbridged oxazines the nitrogen moiety is lost in preference to the oxygen moiety.

Indeed, the ferrolactams may be prepared by treatment of the ferrolactones with an amine in the presence of alumina, or merely by chromatography of the reaction mixture on alumina. The use of the ferrolactone derived from isoprene (79) to demonstrate this reaction reveals 108 a very interesting point. The it-allyl system is attacked by the amine displacing the oxygen moiety, resulting in a possible intermediate of structure (85).108 Closure, and elimination of water from this intermediate

H21:1) ONH2 0 ("11-Fe(CO) --It- ,--Fe(CO) e(CO)3 3 A1 0 2 3 CO 2 (79) (85) (80) gives the ferrolactam (80).

-The reaction of the bridged oxazines (81), (83) with driron nonacarbonyl (under anhydrous conditions), is not thought to involve the production of an amino alcohol as in the simpler oxazines. It is proposed that the electrophile iron.tetracarbonyl inserts into the N-0 bond directly. Insertion of carbon monoxide, followed by elimination of carbon dioxide gives an intermediate (80). Similar intermediates have been reported by Aumann.112 Carbon monoxide insertion into (80) produces the ferrolactam. This does not explain the preferential production of ferrolactones from the unbridged oxazines. 49.

CO Fe(C0)3 e(C0)3

(86)

The most efficient route to ferrolactones yet found is closely ana-

logous to the insertion of iron tetracarbonyl into vinyl cyclopropanes. ofvinyloxiranes in the presence 112 , 113, 114 Photolysisof iron pettacarbonyl (or thermolysis with diiron nonacarbonyl)114 gives ferrolactone complexe in excellent yield. This reaction seems to be limited to iron carbonyls. Vinyl epoxides when treated with analogous Rhodium carbonyls have been reported to produce 115 114 enones} The reaction is stereospecific, as has been shown by the reaction of derivatives of hexa-2,4-diene. The mechanism is probably

(84)%

very similar to that for vinyl cyclopropanes. These complexes decarbonyl.- 112 ate and thermolyse to give similar products to the carbon analogues.

50.

The monoepoxide of cyclooctatetraene produces a ferrolactone, but due 116 its lability further reaction takes place.

+ Fe (C0) 5

e (C0)3

Vinyl epoxides have been converted to S-lactones with carbon mono- 117 xide at high pressures in the presence of an iron catalyst. Presumably a ferrolactone is generated and the lactone produced by exhaustive carbonylation.

CO Fe cat. + 180 Atm. 70° C

Vinyl āziridines have been reported to produce ferrolactams in a similar fashion, though decarbonylation is a very facile reaction.

{OZM e hv Fe(C0)5

Reactive enones118 also give lactones when treated with iron carbonyls.

This reaction is important as is it one of the few reported cases where an isolable olefin-iron tetracarbonyl complex, has been shown to 51.

decarbōnylate photolytically before reaction. Another havily loaded

hv Fe (00)5

2 Days 28% by,/ / 4 Hr... via 45%

119 case has been reported to give ferrolactams of an unusual nature.

One example of a thioketone reacting with iron has been reported.

Thiobenzophenone when treated with diiron nonacarbonyl gives an unusual 121,122 120,122 iron complex (87),120 Lactones and thiolactones can be prepared from this complex in good yield, depending on the conditions used. 52.

( 87)

•

(e) Coupling of Olefins Using Iron Carbonyls

Complex, strained olefins react under thermolytic or photolytic 123--130 conditions with iron carbonyls to produce cyclopentanone derivatives. 124 Detailed analysis of the N.M.R. spectra of these derivatives reveals that the compounds have an exo-trans-exo configuration in nearly every case examined. This is quite remarkable when the total number of different r

53.

74%

A 93% Fe(C0)5

(88)

isomers which could be formed from this reaction is considered. One example has been reported in which the product has an exo-trans-endo configuration (88),129 This is thought to be due to the coordinating effect of the bridging methoxy groups on the iron during the course of the reaction. The mechanism of the reaction has been studied in detail.

125, 128 Addition of the olefin to iron tetracarbonyl gives the expected

+Co + Fe(C0)4 H---Fe(CO) II Fe(C0)3 4 CO (89) (90) II

CO co co ...... _ 0 e(CO)4 Fe(C0)4 (93) + Fe(CO)5 K (92) product (89). Thermal decarbonylation of the complex leads to an unsaturated species (90), which further reacts with another molecule of O(dip) to give a diolefin complex (91). Carbonylation of (91) induces the coupling of 54.

the coordinated olefins to give a metallocycle (92), which can be isolated 125 under certain circumstances. Further carbonylation of the metallocycle

(92), results in carbonyl insertion and iron extrusion to give the cyclo-

pentanone (93). The coupling of olefins is not restricted to complex

hydrocarbons. Addition of methyl acrylate to the methylacrylate tetra- 130 carbonyl iron complex produces a metallocycle (94) in low yield.

Me20 CO2Me Fe(CO)4

Carbon monoxide induced carbonyl insertion and iron extrusion produces

the cyclopentanone derivative (94).

55.

2. RING FORMING REACTIONS USING IRON CARBONYLS

Treatment'of aa'-dibromoketones with diiron nonacarbonyl has long been known to generate an iron stabilised oxyallyl cation, which can add 131 to dienes or olefins to give 7 or 5 membered rings respectively.

R Br Br Br

Recently these, addition reactions have been applied to the synthesis of natural products. Addition of tetra-and tribromo-acetone derivatives to furanoid derivatives has provided a useful route to nezukone (95 ),. ).132 a-thujaplicin (96) and s-thujaplicin (97 Troponoids are known to have antibacterial and fungal activity.

(9S) 56.

Fe•Ln

Br 54%

Fe Ln

Br Br

Similar derivatives from furanoids have been used in the preparation of, •C-nucleoside precursors which have antibiotic, antiviral and anticancer 133 activities. The method allows a variety of bases to be attached to the

sugar ring.

The use of N-substituted pyrroles provides an approach to the tropane 134-136 family of alkaloids, of which atropine and cocaine are members. The

basic skeleton is readily converted to tropine (96), scopine (99), tropane-

diol (100) and teloidine (101). Attempts to use N-methylpyrrole directly 57.

/CO2Me Fe 2 (CO)9

/"Zn/Cu

Di Bal

with dibromoacetone and diiron nonacarbonyl were unsuccessful.

Concerted cycloaddition of iron stabilised allyl cations to mono olefins is a symmetry forbidden process.. The reaction is thus considered as a stepwise process, regioselectivity being controlled by the stability of zwitterionic intermediates. An example of its usefulness has been demons- 137 trated in the synthesis of a-cuparenone.

Fe2(C0)9 r 58.

3, FUNCTIONAL GROUP REMOVAL AND REDUCTION USING IRON CARBONYLS

The addition of acyl halides to disodium tetracarbonylferrate and sub- 138 sequent hydrolysis of the acyl complex to produce aldehydes is well known.

H+ R COCt + Na2FetC0)4 RCHO

Cyclic anhydrides can also be similarly treated with disodium tetracarbonyl- 139, 140 ferrate to give carboxylic aldehydes. Some selectivity has been

0 1/. Na2Fe(CO)4 CHO

2/. H+ CO2H

observed during the cleavage of acyclic unsymmetrical anhydrides. However, much

more selectivity has been found when carboxylic ethylcarbonic anhydrides are

treated with disodium tetracarbonylferrate, which affords the aldehyde exclu- 141 sively.

H+ nC1ctiziCO2 CO2 Et + N a2 Fe (CO)4_ nC10H CHO 21 67%

The reaction of alkyl halides with disodium tetracarbonyferrate, and

the addition of the acyl complex formed to an acyl halide to produce ketones

has been dealt with in a previous section. However, if an a-haloketone is

treated with disodium tetracarbonylferrate, and the complex formed added to

"an acyl halide an enol ester is the major product formed. No 1,3-dicarbonyl 142 compound was isolatect, The formation of this product is presumably due.

to the formation of an iron enolate rather than an acyl complex, which

undergoes 0-acylation in preference to C-acylation. 59.

71% McCOCH2Fe(C0)4+

Various iron species are also prominent in effecting decarbonylation 143,144 of vinyl acetates, a,f3-unsaturated aldehydes and Knoevenagel cōndensates.

Fe(C0)5 H~ ~ 40% 2"10

36 °/0

Ac 0

HO + CH 2 (COM e) 2 OMe OMe ACH Fe(C0)4 Me OH

CO Me 82 %

Olefins can also be formed by the deoxygenation of epoxides. Both the sodium salt of dicarbonylpentahaptocyclopentadienyl iron and iron tetra-

60.

145, 146 carbonyl have been reported to effect this transformation.

THF + NaFe(CO) Cp 86% 2

1 45°C

TMU Fe(CO)5'

The first of these gives a trans-olefin, from the inversion of stereochemistry 145 by the initial nucleophilic attack and cis-elimination. The most probable mechanism involved in the second transformation, involves insertion into the epoxide ring by iron tetracarbonyl,ligand induced carbonyl insertion, ligand

TMU Fe (C0)4 + e-- 0 C)--Ie (CO)4 4 TMU TMU

(TM U)2Fe (C0)3

+ CO2

146 induced iron extrusion and elimination of carbon dioxide. Iron complexes may reduce many other species including conjugated double bonds, 147 and 148 alkyl halides. 61.

ISOMERISATIONS AND REARRANGEMENTS USING IRON CARBONYLS

Iron carbonyls have been shown to isomerise heteroannular transoid dien- 3,5 es to homoannular cisoid dienes in .steroidal systems. Thus, a A steroid 149 was converted to a A2'4 derivative.

Fe(CO) _ Fe Ct 3 5

nBu20 36 h. Me Fe(CO)3 71 59 Vo

More recently, it has been shown that unconjugated dienes can be re- 150 arranged to conjugated. Thus, the prostaglandin derivative (102)forms a diene complex (103) when treated with triiron dodecacarbonyl. The diene complex is stable to the Wittig reagent used in the transformation (103) } (104). OH

Fe3 (C0) 12

1e(CO)3 (103) // C5H11 '"OTHP (1)113 CH(CH2)3CO2Na

COLLINS '~-/6VA\co2H 02H REAGENT Fe (C0) 3 (104) 5H11 ''OTHP 62.

The protecting iron tricarbonyl moiety is finally removed with Collins reagent, which simultaneously oxidises the cyclopentanol to a cyclopentanone.

OTHER METHODS OF C-C BOND FORMATION USING IRON CARBONYLS

Although aspects of the addition of nucleophiles to dienyl cations has been covered in the synthesis of ketones and esters via carbonyl insertion, certain aspects of this reaction deserve further mention. The addition of the 151 nucleophile to the iron stabilised cation is regio and stereospecific. The methoxy substituent polarises the cation, inhibiting additon of the nucleophile to one end of the dienyl system.

II13COH /Ac20 H BF4 BF4- ___. 2/. -H' Me0 Niel) Fe(CO)3 Fe(C0)3 Fe(C013

~. 1/. CeIV / M e OH 2/. DDQ

M 63.

Many other nucleophiles have been shown to add to these dienyl cations including trialkylalkynylborates , 152 Highly functionalised 4-substituted cyclohexenones can be produced in moderate yields. Other authors have effected

R1 B R2 13F 3 OMe 4 Fe(C013

, Fe(CO)1e3 46% Me3NO R2 R1

0 M Fe(CO)3 Ce Iv /46% R2\Me3N0 46%

153 angular cyanation of intermediate iron stabilised pentadienyl cations,

e(CO)3 e(CO) Me OMe

Fe (CO) Ph3 n 0 5 01. C+B 4 Bu2 43% Fe(C0) ~aCN / H 0 3 OMe 2

80% CN

64.

6. FUNCTIONAL GROUP PROTECTION USING IRON CARBONYLS

The tricarbonyl iron moiety has been used to stabilise many reactive dienes, including. cyclobutadiene derivatives. The tricarbonyl iron moiety is unaffected by a number of reagents, and thus the functional groups on the 154 cyclobutadiene ring may be modified.

Cl Cl 1/. hv/ Pyrex 21. H+ 02Me 3/. CH2N2 O2Me CI Cl Cl

Nat Fe(C0) 40%

COCl so CO2H KOH CO2Me OCt "1"-- 02H McOH CO Me 2 Fe(C0)3 Fe(C0)3 Fe (C0) 3

~BH4 /Di bat B H (BF3 2 6 ECcH20 y HZOH Me FeIC013 lot Me LiAtH4 HCI Fe(C0)3 4 H- Ct IC )I CHZCt Fe (CO)3

Oxidation of the complex with ceric ammonium nitrate generates the cy- • 155 clobutadiene in situ, which may then react further.

Ceiv 50 Vo

65.

Iron complexes in which only one of the double bonds is protected have 156 also been produced.

AgPF6

-78°C.

29%

The dicarbonylpentahaptocyclopentadienyl iron complex has also been used 157 to protect isolated acyclic double bonds.

P BF4 /TFA H /Pd 2

Fp BF4 Nal 66.

ACTIVATION OF ALKENES USING IRON CARBONYLS

Nucleophiles (for amines see carbonyl insertion) have been shown to

add to olefins coordinated to iron. Alkylation occurs from the side opposite

THF ç# ;;.. CFkCO2Et)2 Li+CH(CO2Et)2 + 82% ; - 78°C

AFP

+ 1/MeCN

BF_ 21.1-1,0 z

CHO

to the bulky iron complex, which can be subsequently removed with retention or 158-160 inversion of configuration at the carbon site occupied. 67.

REFERENCES

1. B.F.G. Johnson, J.C.S. Chem. Comm., 1976, 703.

2. W.O. Siegl and J.P. Collman, J. Amer. Chem. Soc., 1972, 94, 2516.

3. E.A.K. Von Gustorf, F.W. Grevels, and I. Fischlēr, 'The Organic Chemistry of Iron, Academic Press, London, 1978. I. Wender and P. Pino, 'Organic Syntheses via Metal Carbonyls', Volume II, Wiley-Interscience , London 1978. H. Alper, 'Transition Metal Organometallics in Organic Synthesis', Volume I, Academic Press, London, 1976.

4. M.P.Cooke,Jr., J.Amer. Chem. Soc., 1970, 92, 6080.

5. J.P. Collman, S.R. Winter, and D.R. Clark, J. Amer. Chem. Soc., 1972, 94, 1788.

6. J.P. Collman and N.W. Hoffman, J. Amer. Chem. Soc., 1973, 95, 2690.

7. J.P. Collman, S.R. Winter, and R.G. Komoto, J. Amer. Chem. Soc., 1973, 95, 249.

8. J.P. Collman, Accounts Chem. Res., 1975, 8, 342.

9. M. Yamashita, Y. Watanabe, T.A. Mitsudo, and Y. Takegami, Tetra- hedron Letters, 1976, 1585.

10. M. Yamashita, K. Mizushima, Y. Watanabe, T. Mitsudo,and Y. Takegami, Chem. Lett., 1977, 1355.

11. M.P.Cooke,Jr.,and R.M.Parlman, J. Amer. Chem. Soc., 1975, 97, 6863.

12. M.P.CookeJr.,and R.M. Pariman , J. Amer. Chem. Soc., 1977, 99, 5222.

13. J.Y. Mērour, J.L. Roustan, C. Charrier, J. Collin, and J. Benaim, J. Organometallic Chem., 1973, 51, C24. 68.

14. J.P. Collman, R.G. Finke, J.N. Cawse, and J.I. Brauman, J. Amer. Chem. Soc., 1978, 100, 4766.

15. J.P. Collman, R.G. Finke, J'.N. Cawse, and J.I. Brauman, J. Amer. Chem. Soc., 1977, 99, 2515.

16. J.P. Collman, J.N. Cawse, J.I. Brauman, and J. Amer. Chem. Soc., 1972, 94, 5905.

17. G.M. Whitesides and D.J. Boschetto, J. Amer. Chem. Soc., 1969, 91, 4313.

18. P.L. Bock, D.J. Boschetto, J.R. Rasmussen, J.P. Demers, and G.M. Whitesides, J. Amer. Chem. Soc., 1974, 96, 2814.

19. T.A. Mitsudo, Y. Watanabe, M. Yamashita, and Y. Takegami, Chem. Lett., 1974, 1385.

20. Y. Sawa, M. Ryang, and S. Tsutsumi, Tetrahedron Letters, 1969, 5189.

21. M. Ryang, I. Rhee, and S. Tsutsumi, Bull. Chem. Soc. Japan, 1964, 37, 341.

22. Y. Sawa, M. Ryang, and S. Tsutsumi, J. Org. Chem., 1970, 35, 1..

23. M. Yamashita, and R. Suemitsu, J.C.S. Chem. Comm., 1977, 691.

24. C.S. Giam and K. Ueno, J. Amer. Chem. Soc., 1977, 99, 3166.

25. M. Yamashita, and R. Suemitsu, Tetrahedron Letters, 1978, 761.

26. M. Yamashita and R. Suemitsu, Tetrahedron Letters, 1978, 1477.

27. M. Yamashita, K. Mizushima, Y. Watanabe, T.A; Mitsudo, and Y. Takegami, J.C.S. Chem. Comm., 1976, 670.

28. A.F.M. Iqbal, Hely. Chimr Acta., 1976, 59, 655.

29. P. Krumholtz and H.M.A.Stettiner,J. Amer. Chem. Soc., 1949, 71, 3035. 69.

30. H. Masada, M. Mizuno, S. Suga, Y. Watanabe, and Y. Takegami, Bull. Chem. Soc. Japan, 1970, 43, 3824.

31. Y. Takegami, Y. Watanabe, H. Masada, and I. Kanaya, Bull. Chem. Soc. Japan, 1967, 40, 1456.

32. Y. Takegami, Y. Watanabe, T. Mitsudo, and H. Masada, Bull. Chem. Soc. Japan, 1969, 42, 202.

33. R.F. Heck, J. Amer. Chem. Soc., 1963, 85, 1460.

34. T.A. Mitsudo, H. Nakanishi, T. Inubushi, I. Morishima, Y. Watanabe and Y. Takegami, J.C.S. Chem. Comm., 1976, 416.

35. K. Hayakawa and H. Schmid, Helv. China. Acta, 1977, 60, 1942.

36. L.A. Paquette, S.V. Ley, S. Maiorana, D.F. Schneider, M.J. Broad- hurst, and R.A. Boggs, J. Amer. Chem. Soc., 1975, 97, 4658.

37. T.C. Floodjand A. Sarhangi, Tetrahedron Letters, 1977, 3861.

38. - J.P. Collman, R.K. Rothrock, R.G. Finke, and F.Rose-Munch, J. Amer. Chem. Soc., 1977, 99, 738,

39. G. Cainelli, F. Manescaichi, A. Umani-Ronchi, and M. Panunzio, J. Org. Chem., 1978, 43, 1598.

40. A.P. Kozikowski and, H.F. Wetter, Synthesis, 1976, 561.

41. H. Takaya and R. Noyori, J. Synth. Org. Chem. Japan, 1977, 35, 615.

42. L.S. Hegedus, J. Organometallic Chem., 1978, 163, 187.

43. A. Wojcicki, Adv. Organometallic Chem., 1973, 87.

44. R.B. King, J. Amer. Chem. Soc., 1963, 85, 1918. 70.

45. M.L.H.Green and M.J. Smith, J. Chem. Soc. A, 1971, 3220.

46. G.M. Reisner, T. Bernai,H. Brunner, and M. Mushiol, Inorg. Chem., 1978, 17, 783.

47. S.N. Anderson, C.W. Fong, and M.D. Johnson, J.C.S. Chem. Comm., 1973, 163.

48. K.M. Nicholas and M. Rosenblum, J. Amer. Chem. Soc., 1973, 95, 4449.

49. N. Genco, D. Marten, S. Raghu, and M. Rosenblum, J. Amer. Chem. Soc., 1976, 98, 848.

50. S.G. Davies, M.L.H. Green, and D.M.P.Mingos, Tetrahedron, 1978, 34 3047.

51. P.K. Wong, M. Madhavarao, D.F. Marten, and M. Rosenblum, J. Amer. Chem. Soc., 1977, 99, 2823.

52. J.L. Roustan, J. Benaim, C. Charrier, and J.Y. Merour, Tetrahedron Letters, 1972, 1953.

53. T. Inglis, and M. Kilner, J.C.S. Dalton, 1976, 563.

54. R.J. Angelici, Accounts Chem. Res., 1972, 5, 335.

55. S. Sarel,R.Ben-Shoshan, and B. Kirson, J. Amer. Chem. Soc., 1965, 87, 2517.

56. R. Victor,R.Ben-Shoshan, and S. Sarel, Tetrahedron Letters, 1970, 4253.

57. R.BenShoshan and S. Sarel, J.C.S. Chem. Comm., 1969, 883.

58. S. Sarel, A. Felzenstein, R. Victor, and J. Yovell, J.C.S. Chem. Comm., 1974, 1025.

59. S. Sarel and M. Langbeheim, J.C.S. Chem. Comm., 1977, 827.

60. R. Aumann, J. Amer. Chem. Soc., 1974, 96, 2630. 71.

61. R. Aumann, J. Organometallic Chem., 1972, 47, C29:

62. S. Sarel and G. Chriki, J. Org. Chem., 1978, 43, 4971.

63. R.M. Moriarty, K-N. Chen, M.R. Churchill, and S.W.Y. Chang, ' J. Amer. Chem. Soc., 1974, 96, 3661.

64. P. Eilbracht, Chem. Ber., 1976, 109, 1429.

65. A. Stockis and E. Weissberger, J. Amer. Chem. Soc., 1975, 97, 4288.

66. R. Victor, R.Ben-Shoshan, and S. Sarel, Tetrahedron Letters, 1973, 4211.

67. R. Victor, V. Usieli, and S. Sarel, J. Organometallic Chem., 1977, 129, 387.

68. E.K. Von Gustorf and F.W. Grevels, Fortsh. Chem. Forsch., 1969, 13, 366.

69. S. Sarel, Accounts Chem. Res., 1978, 11, 204.

70. R.M. Moriarty, K-N. Chen, C--L. Yeh, J.L. Flippen, and J. Karle, J. Amer. Chem. Soc., 1972, 94, 8944.

71. S.W. Tam, Tetrahedron Letters, 1974, 2385.

72 R. Aumann, J. Organometallic Chem., 1974, 76, C32.

73. B.F.G. Johnson, J. Lewis, D.J. Thompson, and B. Heil, J.C.S. Dalton, 1975, 567.

74. R. Aumann and H. Averbeck, Angew. Chem. Internat. Edn., 1976, 15, 610.

75. R. Aumann, H. Wormann, and C. Kruger, Chem. Ber., 1977, 110, 1442.

76. R. Aumann, H. Wormann, and C. Kruger, Angew Chem. Internat. Edn., 1976, 15, 609.

77. R. Aumann, J. Organometallic Chem., 1974, 66, C6.

78. R.M. Moriarty, C.L. Yeh, and K.C. Ramey, J. Amer. Chem. Soc., 1971, 93, 6709. 72.

79. D.Fhnt.holt, A. Rosan, and M. Rosenblum, J. Organometallic Chem., 1973, 56, 315.

80. R. Aumann, Angew. Chem. Internat. Edn., 1972, 11, 522.

81. R. Aumann, J. Organometallic Chem., 1974, 77, C33.

82. R.M. Moriarty, C.L. Yeh, and K.N. Chen, Tetrahedron Letters, 1972, 5325.

83.' R. Aumann, Angew. Chem. Internat. Edn., 1971, 10, 190.

84. H.D. Scharf and J. Maltay, Tetrahedron Letters, 1977, 401.

85. R. Aumann, H. Wormann, and C. Kruger, Chem. Ber., 1977, 110, 1442.

86. R. Aumann, H. Wormann, and C. Kruger, Angew. Chem. Internat. Edn., 1976, 15, 610.

87. R. Aumann, Angew. Chem. Internat. Edn., 1971, 10, 188.

88. R. Aumann, Angew. Chem. Internat. Edn., 1971, 10, 189.

89. G.N. Schrauzer, P. Glockner, K.I.G. Reed, and I.C. Paul, J. Amer. Chem. Soc., 1970, 92, 4479.

90. R. Aumann, Chem. Ber., 1975, 108, 1974.

91. R. Aumann, Chem. Ber., 1977, 110, 1432.

92. R. Aumann, Angew. Chem. Internat. Edn., 1976, 15, 376.

93. B.F.G. Johnson, J. Lewis, and D.J. Thompson, Tetrahedron Letters, 1974. 3789.

94. A. Eisenstadt, Tetrahedron Letters, 1972, 2005.

95, A.H-J. Wang, I.C. Paul, and R. Aumann, J. Organometallic Chem., 1974, 69, 301. 73.

96. V. Heil, B.F.G. Johnson, J. Lewis, and D.J. Thompson, J.C.S. Chem. Comm., 1974, 270.

97. B.F.G. Johnson, J. Lewis, and D.J. Thompson, and B. Heil, J.C.S. Dalton Trans., 1975, 567.

98. B.F.G. Johnson, K.A. Karlin, and J. Lewis, J. Organometallic Chem., 1978, 145, C23.

99. R. Aumann, Chem. Ber., 1976, 109, 168.

100. F.A. Cotton, A.J. Deeming, P.L. Josty, S.S. Ullah, A.J.P. Domingos, B.F.G. Johnson, and J. Lewis, J. Amer. Chem. Soc., 1971, 93, 462 .

101. R. Aumann, Angew. Chem. Internat. Edn., 1973, 12, 574.

102. R. Aumann and J. Knecht, Chem. Ber., 1976, 109, 174.

103. R. Aumann- and J. Knecht, Chem. Ber., 1978, 111, 3927.

104. R. Aumann and J. Knecht, Chem. Ber., 1978, 111, 3429.

105. H. Olsen, Acta. Chem. Scand.B,1977, 1331, 635.

106. R.F. Heck and C.R. Boss, J. Amer. Chem. Soc., 1964, 86, 2580.

107. H.D. Muidock, Hely. Chim. Acta., 1964, 47, 936.

108. Y. Becker, A. Eisenstadt, and Y. Shvo, Tetrahedron, 1974, 30, 839..

109. Y. Becker, A. Eisenstadt, and Y. Shvo, Tetrahedron, 1976, 32, 2123.

110. Y. Becker, A. Eisenstadt, and Y. Shvo, Tetrahedron Letters, 1972, 3183.

111. Y. Becker, A. Eisenstadt, and Y. Shvo, Tetrahedron, 1978, 34, 799.

112. R. Aumann, K. Frohlich, and H. Ring, Angew. Chem. Internat. Edn., 1974, 13, 275.

74.

113. G.D. Annis and S.V. Ley, J.C.S. Chem. Comm., 1977, 581.

114. K N. Chen, R.M. Moriarty, B.G. DeBoer, and M.R. Churchill, and H.J.C. Yeh, J. Amer. Chem. Soc., 1975, 97, 5602.

115. G. Adames, C. Bibby, and R. Grigg, J.C.S. Chem. Comm., 1972, 491.

116. R. Aumann and H. Averbeck, J. Organometallic Chem., 1975, 85, C4.

117. R. Aumann and H. Ring, Angew• Chem. Internat. Edn., 1977, 16, 50.

118. J. Elzinga and H. Hogeveen, J. Org. Chem., 1978, 43, 745.

119. A.N. Nesmeyanov, M.I. Rybinskaya, L.V. Rybin, N.T. Gubenko, N.G. Bokii, A.J. Batsanov,and Y.T. Struchov, J. Organometallic Chem., 1978, 149, 177.

120. H. Alper,•W.G. Root, and A.S.K. Chan, J. Organometallic Chem., 1974, 71, C14.

121. H. Alper and W.G. Root, J.C.S. Chem. Comm., 1974, 956.

122. H. Alper and W.G. Root, J. Amer. Chem. Soc., 1975, 97, 4251.

123. J. Mantzaris and E. Weissberger, Tetrahedron Letters, 1972, 2815.

124. J. Mantzaris and E. Weissberger, J. Amer. Chem. Soc., 1974, 96, 1873.

125. E. Weissberger and P. Laszlo, Accounts Chem. Res., 1976, 9, 209.

126. T. Grandjean, P. Laszlo, and A. Stockis, J. Amer. Chem. Soc., 1974, 96, 1622.

.127. L. Lombardo, D. Wege, and S.P. Wilkinson, Aust. J. Chem., 1974, 27, 143.

128. J. Mantzaris and E. Weissberger. J. Amer. Chem. Soc., 1974, 96, 1880.

129. A. Speert, J. Gelan, M. Anteunis, A.P. Marchand, and P. Laszlo, Tetrahedron Letters, 1973, 2271. 75.

130. F.W. Grevels, D. Schulz, and E.K. Von Gustorf, Angew. Chem. Internat. Edn., 1974, 13, 534.

131. R. Noyori, Accounts Chem. Res., 1979, 12, 61.

132. H. Takaya, Y. Hayakawa, S. Makino, and R. Noyori, J. Amer. Chem. Soc., 1978, 100, 1778.

133. R. Noyori, T. Sato, and Y. Hayakawa, J. Amer. Chem. Soc., 1978, 100, 2561.

134. R. Noyori, S. Makino, Y. Baba, and Y. Hayakawa, Tetrahedron Letters, 1974, 1049.

135. R. Noyori, Y. Baba, and Y. Hayakawa, J. Amer. Chem. Soc., 1974, 96, 3336.

136. Y. Hayakawa, Y. Baba, S. Makino, and R. Noyori, J. Amer. Chem. Soc., 1978, 100, 1786.

137. Y. Hayakawa, F. Schimizu, and R. Noyori, Tetrahedron Letters, 1978, 993. 138. Y. Watanabe, T-A. Mitsudo, M. Tanaka, K. Yamamoto, T. Okajima, and Y. Takegami, Bull. Chem. Soc. Japan, 1971, 44, 2569.

139. Y. Watanabe, M. Yamashita, T--A. Mitsudo, M. Tanaka, and Y. Takegami, Tetrahedron Letters, 1973, 3535.

140. Y. Watanabe, M. Yamashita, T--A. Mitsudo, M. Igami, and Y. Takegami, Bull. Chem. Soc. Japan, 1975, 48, 2490.

141. Y. Watanabe, M. Yamashita, T--A. Mitsudo, M. Igami, K. Tomi, and Y. Takegami, Tetrahedron Letters, 1975, 1063.

'142. T.A. Mitsudo, Y. Watanabe, T. Sakai, H. Nakanishi, M. Yamashita, and Y. Takegami, Tetrahedron Letters, 1975, 3163.

143. S.J. Nelson, G. Detre, and M. Tanabe, Tetrahedron Letters, 1973, 447. 76.

144. M. Yamashita, Y. Watanabe, T. Mitsudo, and Y. Takegami, Tetrahedron Letters, 1975, 1867.

145. M. Rosenblum, M.R. Saidi, and M. Madhavarao, Tetrahedron Letters, 1975, 4009.

146. H. Alper and D. Des Roches, Tetrahedron Letters, 1977, 4155.

147. G.P.Boldrini and A.Umani-Ronchi, Synthesis, 1976, 596.

148. H. Alper, Tetrahedron Letters, 1975, 2257.

149. H. Alper and J.T. Edward, J. Organometallic Chem., 1968, 14, 411.

150. E.J. Corey and G. Moinet, J. Amer. Chem. Soc., 1973, 95, 7185.

151. R.E. Ireland, G.G. Brown,Jr., R.H. Stanford, Jr., and T.C. McKenzie, J. Org. Chem., 1974, 39, 51.

152. A. Pelter, K.J. Gould, and L.A.P. Kane-Maguire, J.C.S. Chem. Comm., 1974, 1029.

153. A.J. Pearson, J.C.S. Chem. Comm., 1977, 339.

154. G. Berens, F. Kaplan, R. Rimerman, B.W. Roberts, and A. Wissner, J. Amer. Chem. Soc., 1975, 97, 7076.

155. R.H. Grubbs, T.A. Pancoast, and R.A. Grey, Tetrahedron Letters, 1974, 2425.

156. A. Sanders and W.P. Giering, J. Amer. Chem. Soc., 1975, 94, 919.

157. K.M. Nicholas,J. Amer. Chem. Soc., 1975, 97, 3254.

158. M. Rosenblum, Accounts Chem. Res., 1974, 7, 122.

159. A.M. Rosan, M. Rosenblum, and J. Tancrede, J. Amer. Chem. Soc., 1973, 95, 3062.

160. K.M. Nicholas and A.M. Rosan, J. Organometallic Chem., 1975, 84, 351.

77.

RESULTS AND DISCUSSION

INTRODUCTION

The stereospecific preparation of a,9r-allyl iron complexes from vinyl epoxides is known, although the chemistry of these interesting complexes has not been previously investigated 1'2

Fe2(CO>g

or e(C0)3 R. hv/Fe(CO R

We argued that oxidative extrusion of the iron moiety could lead to coupling of the oxycarbonyl carbon with either end of the allyl system and hence afford lactones.

(0)

1,5 coupling

[0) R 3,5 coupling R

In principle, therefore, the reaction could be viewed as a synthetic equivalent of the addition of carbon dioxide to a diene to afford lactones, a transformation which is difficult to achieve by conventional methods. In suitably substituted examples one could envisage the process for d-lactones being applied to natural product synthesis. a-Vinyl S-propiolactones produced by this method would provide an interesting and relatively new class of compound which could be studied further. We have, therefore, prepared 78.

ā number of dienes from which ferrolactone complexes have been prepared via

intermediate epoxides. The ferrolactones have been oxidised under a variety

of conditions to afford the desired lactones.

Preparation of Lactones from Ferrolactones

Preparation and Oxidation of Syn- and Anti- Tricarbonyl-(1,1',2'-rfi-l-vinyl-

cyclopentan-l-y1)-2-oxycarbonyl iron (3i and (4)

mcpba

(1) (2) (3) (4)

The iron complexes (3) and (4) were prepared from the epoxide (2) which

in turn was derived from the diene (1). Preparation of diene (1) from the

allylic alcohol has been previously reported.3 The method involves the double

I2 (5) (1)

distillation of the allylic alcohol (5) in the presence of iodine under

reduced pressure, to give the diene. However, it was found that only one

distillation was necessary, and provided the temperature of the reaction

. mixture was kept below 100°C, the diene could be obtained in 1v85% yield. 79.

Oxidation of the diene with m-chloroperbenzoic acid gave the mono- (2) epoxide/in moderate yield. The epoxide was distilled, although in order

to prevent substantial decomposition, high temperatures had to be avoided.

The epoxide was readily characterised by examination of its 1H n.m.r. spectrum which shows a signal at 6 3.50 due to the methine proton on the oxirane ring, and signals at 6 5.86 and 6 5.48-5.08 due to the vinylic grouping.

Irradiation of the epoxide in the presence of pentacarbonyl iron produced the ferrolactones (3) and (4) in 79% overall yield. In this case, relatively high concentrations of the epoxide could be used (0.036 mmol/ml) and the reaction could easily be followed to completion by i.r. spectroscopy. Removal of the solvent and excess iron pentacarbonyl by rotary evaporation, and recrystallisation of the residue from ether/petroleum ether gave the iron complexes as white thermally labile crystals. Many refine- ments of this method were found to be necessary for the preparation of other ferrolactones (see later section). The product was found to consist of.two isomers, readily separable by fractional recrystallisation or, better, by chromatography on Florisil. Their structures were determined by analysis of their 100 MHz n.m.r. and i.r. spectra, and by X-ray crystallographic determination.

HA Hc HD H0 ,

syn-(3) 6 4-26 5.34 3.49 264

HA HE H0 H0,

anti -(4) 6 4-32 4.61 350 2.99

The complexes were isolated in the ratio of 6:1, (3) : (4). Detailed analyses of the spectra in relation to the X-ray structuresof the two isomers is discussed in a later section.

The choice of solvent used for the oxidation of these iron complexes with ceric ammonium nitrate was found to be very important. In ethanol the rate of reaction was slow. In acetone solution the complexes were oxidised to an as yet unidentified product containing no carbonyl group. The first successful oxidations were performed on mixtures of the complexes (3) and

(4) with ceric ammonium nitrate in dimethyl sulphoxide/water,.at

70oC, where the 0-lactone (6) was produced in a yield of 36%.

Cely

(3) (6) '(7) Fe(CO)3 el°1 Ce Iv

(4) (7) Later oxidations with ceric ammonium nitrate in acetonitrile have shown that the syn-product (3) produces both g- and 6-lactones (6) and (7) respectively in a ratio of 1.77 : 1 in an overall yield of 79.8%. The anti-complex (4), however, on similar oxidation produces exclusively the 6-lactone (7) in 74% yield._ The two lactones from (3) were readily separated by chromatography, and their structures assigned from their spectral data. For example, the i.r. carbonyl absorptions at 1820 and 1745 cm 1, were typical of a- and 6- lactones respectively. 1H n.m.r. spectroscopy shows characteristic signals due to the vinyl group in the 13-lactone at 66.23-5.48 and 6 5.56-5.15, with the ring junction proton HA as a doublet at 6 4;7. Molecular models indicate

81.

that the proton HA should only be coupled to one a-proton in the compound with a cis-fused ring junction, and therefore, should resonate as a doublet.

The corresponding proton in the S-lactone produces a signal at S 4.32, and also shows only one olefinic proton at S 5.00.

Due to the instability of many of these lactones, satisfactory microanalyses could not be obtained. For this reason they were reduced to diols which were easily analysed.

Thus, reduction of both S- and S-lactones. ((6) and (7)), with lithium aluminium hydride produces the expected dials (8) and (9) in 97% and 70.4% yields respectively.

LiAtH4 OH

(6) (6J

LiAlH4

(7) (9) Hydrogenation of the 6-lactone (7) using Palladium on carbon did not result in the formation of the saturated lactone (10), rather preferential cleavage of the allylic oxygen bond, followed by reduction of the double bond gave the carboxylic acid (11) in 91% yield.

H2/Pait o fLo

4 (7) 82.

Preparation and Oxidation of E- and Z-Tricarbonyl-(1,2,3-n-l-nonen-3-yl)-

4-oxycarbonyl iron (21 ) and (22)

Both E and Z iron complexes were prepared from the diene (12) in the usual way. It was proposed to prepare the initial diene by dehydration of

1-nonen-3-ol (13), which was readily prepared in 80% yield by addition of vinyl magnesium bromide to heptaldehyde. Dehydration of this alcohol proved OH

(13) to be more difficult than anticipated. Mesylate and tosylate derivatives failed to eliminate to give the diene, while treatment of the allylic alcohol with thionyl chloride/pyridine gave resinous products. It had been reported that alcohols could be dehydrated4 using o-methylphenyl chloro- thioformate5 (14). Treatment of the allylic alcohol (13) with this reagent

+ ArOC C! I I OH S (13) (14) 83.

did not result in the isolation of the expected 0-alkyithiocarbonate (15), but instead an S-alkylthiocarbonate (16).was the major product (57%) of the reaction. Presumably this was formed via rearrangement of the 0-alkylthiocarbonate.

Burgess' reagent (17) has been reported to dehydrate allylic alcohols,6 however, slight variations of the conditions can result in rearrangement to

Me 02CNSO2NEt3

OH (13) (17)

C6H,0 CO2Me J

NHCO2Me

(18)

carboxylamines. Treatment of the allylic alcohol (13) produced the carbo- xylamine (18) in low yield as the major product, which has a characteristic absorption in its i.r. spectrum at 1720 cm 1. During the course of this work, it was reported that allylic acetates can be converted to 1,3-dienes by treatment with catalytic amounts of palladium (II) chloride.7 Prepara- tion of the allylic acetate (19) (93%), and subsequent treatment of this compound with palladium (II) acetate afforded the diene (12) in 78.5%. yield. Both E and Z isomers were formed which made the n.m.r. spectrum of the diene quite complex. However, the integral of the olefinic protons at 84.

6 6.83-4.75 relative to the integral for protons in the rest of the molecule was in the correct ratio for the proposed structure of the diene (12).

Oxidation of the diene with peracetic acid gave a mixture of diene

and monoepoxide, which were difficult to separate. However, m-chloroper-

benzoic acid gave the epoxides (20) cleanly in 67.5% yield. Again, the

n n.m.r. spectrum of this compound was complex, but the integrals were in

good agreement with the proposed structure (20).

The irradiation procedure in the presence of iron penta'carbonyl for

these epoxides (20) and all other subsequent examples has undergone exten-

sive modification. Lower concentrations of the epoxide (0.011 mmol/ml) were found to improve the yields of ferrolactone complexes. This could be due to the absorption of the u.v. light by the ferrolactone products, which inhibits further reaction of the epoxide. The use of a Chance OX1 filter has also been found to improve the yield of the ferrolactones. Certain complexes are extremely susceptible to decarbonylation in solution, and thus removal of the solvent under reduced pressure using a rotary evaporator would seem to be inadvisable. If the reaction mixture was frozen and the solvent removed under high vacuum, i.e. freeze dried, the exposure of ferrolactones to vacuum and heat in solution was avoided. Good yields were obtained in many cases where this procedure was adopted. The stability of the ferrolactone complexes in solution was also found to be dependent on the nature of the solvent. Chlorinated solvents and benzene seem to destabilise the complexes, while more polar solvents such as acetonitrile and ether lend greater stability to the iron complex. Thus, crystallisation of the residue after freeze drying from ether/petroleum ether was found to give superior results to the use of benzene/petroleum ether, which had been reported in the literature.1 85.

Ac Cl (13) OH (19) OAc Pd Cl2 mcpba •

(20) by (12) Fe(COS

(21)

Thus, the epoxides (20) were photolysed using the procedure outlined above and the product obtained as an oil (after freeze drying) in 76% yield.

The products were found to be a mixture of two isomers by t.l.c., (presumably from the Z and E isomers of the diene) which could be separated by chromatography. Florisil was found to be the best material for effecting the separation, as decomposition occurred on silica gel even at -78°C. The ratio of the least polar to the more polar product was found to be 1:2.4, and each compound had a typical carbonyl absorption at 1660 cm 1 in their i.r. spectra. Both complexes were extremely unstable in deuterochloroform at concentrations required for 60 MHz 1H n.m.r. spectroscopy, although fortunately at lower concentrations the complexes were sufficiently stable to permit a 100 MHz spectrum to be recorded. The minor product's spectrum

/C e(C0)3 HA HO HC HD HD'

b 4-30 4.96--4.60 3.80 312 q D' HD (21) 86.

could be readily assigned on the basis of previous complexes. (For a detailed discussion of the structure of iron complexes see a later section). The major product had a more complex spectrum.

HB\ ~ %C Fe(C0)3

H0 C N , 5 N11 D (22) H B HC HA HD

4.90-4.54 4.24 - 3.82

Oxidation of the minor complex isomer (21) with ceric ammonium nitrate in acetonitrile gave trans-3-pentyl-4-vinyloxetan-2-one (23) in

68% yield.

Ce"

C5H11 HA (21) (23)

This result nicely demonstrates that the oxidation of these iron complexes is stereospecific, and confirms the proposed structure of the iron complex (21). 87.

The structure of the trans-S-lactone was assigned on the basis of its

carbonyl absorption at 1830 cm-1 and analysis of its 100 MHz 1H n.m.r. spec-

•trum. The protons HA and I produced multiplets at 6 4.48-4.31 and d 3.94-3.80

respectively. Decoupling of the methylene protonsa- to HA revealed a coup-

ling constant of 4 Hz between HA and HB, which is close to the value of

3Hz observed for trans-protons on other oxetanones.8

Similarly, the major isomer (22) gave the cis-S-lactone (24) in 64%

yield. Again,. identification of the stereochemistry of the product confirms

IV Ce Fe(CO)

`C5H11 nA 5H11 (22) (.24)

the stereochemistry of the preceeding iron complex (22). The carbonyl

absorption of this complex occurred at 1830 cm-1 and its 1H n.m.r. spectrum

contained triplets at 6 4.58 and 6 4.36, JAB 6 Hz which may be assigned to

B respectively. The coupling constant is in good agreement with HA and• H 8 the coupling observed for cis protons on other oxetanones.

Formation of the 0-lactones by oxidation of these complexes was in

some ways unfortunate as the formation of a 6-lactone would have provided

a potential route to the natural product massoia lactone (25);

In an attempt to affect the geometry of the transition state and the

nature of the products from the oxidation, a carbon monoxide ligand was

88.

replaced by triphenyl phosphine in both ferrolactones. Oxidation of the resultant complexes again produced the (3-lactones exclusively. Variation of the solvent affected only the rate of reaction and not the nature of the products.

It is hoped that the S-lactone precursor of massoialactone (25) will eventually be prepared by exhaustive carbonylation of the iron complexes

(21) and (22), as there is ample precedent for this type of reaction for 1O hydrocarbon analogues.

Preparation and Oxidation of Tricarbonyl-(2,3,4-n-2,3-dimethyl-3-buten-

2-y1)-1-oxycarbonyl iron (27)

mcpba by

Fe(C0)5 e(CO)3

(26) (27)

The iron complex may be prepared from the diene via the epoxide (26).

Oxidation of 2,3-dimethyl-1,3-butadiene with m-chloroperbenzoic acid produced the epoxide (26) in 59.7% yield as expected. This compound was readily characterised by the signal at S 2.72 in its n.m.r. spectrum due to the methylene protons on the oxirane ring. Irradiation of the epoxide (26) in the presence of iron pentacarbonyl without the use of OXl , filters in the modified procedure gave a 'L 30% yield of the complex (27). However, the use of filters in the modified procedure gave a yield of 65% of the iron complex (27),11 as white crystals. The complex was sufficiently stable in deuterochioroform solution to enable a 60 MHz 1H n.m.r. spectrum to be obtained. 89.

The signals at 6 3.44 and 62.64 were assigned to the protons HA and HB

respectively, the difference between the two is characteristic of a terminal

methylene group on a n-allylic system. The resonance at 6 3.97 was assigned

to the methylene group HC.

Oxidation of the ferrolactone (27) with ceric ammonium nitrate at 60°C

in ethanol/water produced the 6-lactone (29), (carbonyl absorption 1740 cm 1)

in 38% yield. Signals at 6 4.73 and 6 2.96 were assigned to the methylene

e(C0)3 Ce iv

(27) (28) (29)

protons adjacent to the heteroatom and the carbonyl group, respectively. The

signal due to the methyl moieties appeared at 6 1.72. This spectrum was

identical to an authentic sample prepared by a different route.12 When the

ferrolactone was oxidised with ceric ammonium nitrate in acetonitrile solution

at room temperature, however, a mixture of S- and 6-lactones was obtained,

in a combined yield of 60% in a ratio of 2.56 : 1. This seems to imply that

the nature of the solvent affects the course of the reaction, although other

explanations are possible. The products were readily separable by chroma-

. tography, and the 8-lactone was characterised by Its spectral characteristics.

The compound has a typical carbonyl absorption at 1825 cm-1, and a C=C

absorption at 1640 cm-1. The n.m.r. contains a broad singlet at -6 4.93

90.

due to the two olefinic protons, a typical AB quartet at S 4.16 due to the

methylene protons on the oxetanone ring, and singlets at S 1.83 and

S 1.60 due to the two methyl groups, the former showing a small allylic

coupling. Reduction of the S-lactone with lithium aluminium hydride gave

the diol (30) in 87% yield.

Li AL H.

(30)

Preparation and Oxidation of Tricarbonyl-(1',2',3'-n-l-propenylcyclohexan-

1 )-1-ox carbonyl iron (35).

Fe (CO)5

(33) (34) (35)

15,16 The iron complex (35) was prepared from the diene (33) via the

epoxide (34). It is known that reduction of the tetrahydropyranyl derivative 16 (32) with lithium aluminium hydride affords the diene (33) in good yield.

This method requires the distillation of the key intermediate (32). However,

OTHP H0

4. ii (OTHP H . FtMc)Br LiAMH4 D H P cyclohexanone (32) (31) (33) 91.

it was found that distillation of the compound (32) at 0.3 mm Hg (as reported) resulted in decomposition of the product. However, if the distillation was performed at 0.03 mm Hg a good yield of the product was obtained. Spectral properties of the diene obtained were in good agreement with those reported ' 15 16 in the literature. '

The spiro epoxide (34) was formed in 61% yield by treatment of the diene

(33) with peracetic acid. The methine proton on the oxirane ring produced a characteristic signal at S 3.12 in the 1H n.m.r. spectrum. •

Irradiation of the epoxide (34) under the usual conditions afforded only a modest yield (48%) of the iron complex (35). It was thought that the steric hind' ance of the epoxide or deoxygenation of the epoxide to give the diene complex (36) may be responsible for the low yield. The diene complex would be much less crystalline than the ferrolactone (35) and would undoubtedly be

\Fe (CO) 1 3

(36) removed during crystallisation of the product. The iron complex obtained had -1 a characteristic carbonyl absorption at 1670 cm , and gave a good 60 MHz

(35) 92.

n.m.r. spectrum. The methylene protons HD and HD' were readily distinguished

by their chemical shifts and their coupling to HC (J 13 and 9 Hz respectively).

An unusual feature of this spectrum was that HC produced a multiplet at

S 5.14-4.58, and HB a doublet at S 3.90. Normally one would expect the proton

H$ to be at lower field than H .

Oxidation of the iron complex (35) failed to give lactone products in

every solvent except acetonitrile. However, some interesting (but as yet

unidentified) products were obtained from the oxidation when performed in

methanol. The product from the oxidation in acetonitrile was the 3-lactone

CeIv

(35) (37)

(37), in 52% yield. The compound had a typical carbonyl absorption at 1830 cm-1

in its i.r. spectrum and resonances at S 5.68-4.68 due to the vinyl group, and

6 3.26 from the methine protons on the oxetanone ring. Decoupling of this

proton resulted in some simplification of the signal due to the vinyl group.

Reduction of the lactone (37) gave the diol (38) in 72% yield.

(37) (38)

{ 93.

Preparation and Oxidation of Tricarbonyl(2,3,4-ri-3-cyclohexen-2-yl)-1-oxycarbonyl iron (41)

(39) (40) (41)

The iron complex (41) was readily prepared from 1,3-cyclōhexadiene via the epoxide (40). The monoepoxide (40) was obtained in 58.6% yield, upon treatment of the 1,3-cyclohexadiene.with peracetic acid. The coethine protons on the oxirane ring produced characteristic signals at 6 3.53-3.43 and 6 3.28-3.11 in the 1H n.m.r. spectrum.

Irradiation of the epoxide in the presence of iron pentacarbonyl in the usual way produced the ferrolactone (41) in 32% yield. The crude product was impure and the complex (41) failed to crystallise directly. However, chromatography on silica gel gave the complex (41) as yellow thermally labile crystals

The poor yield could be due as in other cases to the deoxygenation of the epoxide, or to the formation of an unreactive anti-olefin tetracarbonyl iron complex (42). The ferrolactone was characterised by its typical carbonyl

(42)

absorption at "1630 cm-1, and its 1H n.m.r. spectrum which was assigned by 1 analogy to the reported spectrum of the cyclopentadiene complex (43).

94.

HA HB HC H0

b 4.70 5.84- 5.44 4-57- . 5.86 4.44

HA HB HC HD

b 3.75 4.80 4.60 3.23 •

(43)

CeIV e (C 0)3

(41) (44)

Oxidation of the ferrolactone in either ethanol or ethanol/water with

ceric ammonium nitrate gave the cis-S-lactone in 86% and 69% yields respect-

ively. This is one of the few examples in which ethanol is a good solvent for

the oxidation reaction. The ferrolactone appears to form a complex with the

ceric ammonium nitrate in solution, which decomposes as the temperature of

the reaction mixture rises above.9 °C, to give the cis-3-lactone. The

a-lactone (44) had a typical carbonyl absorption at 1820 cm-l and signals at

d 6.30-6.00, 5.85-5.53, 4.91-4.73 and 4,10 in its 1H n.m.r. The signal at

d 4.10 was a triplet (J 6 Hz) and was assigned to the proton H Decoupling B. confirmed this assignment as HA simplifies when HB (d 4.10) was irradiated.

Irradiation of HA resulted in the signal from HB reducing to a doublet (J 6 Hz)

95.

as expected. The spectrum was also in agreement with that of a similar 13-lac- 18' 19 tone prepared by a different route.

Reduction of the lactone (44) gave the diol (45) which was not isolated, but converted to and characterised as the ditosylate (46).

LiAtM4 H Ts Cl OTs

(44) (45) (46)

Preparation and Oxidation of Tricarbonyl-(2,3,4-n-2,3-(tetramethylene)-3-buten-

)-i-oxycarbon 1 iron (54)

(52) (54)

The proposed synthetic route to the iron complex (54) required the synthesis of the diene (52). This compound had previously been prepared by.

Blomquist.20 cis-1,2-Cyclohexanedicarboxylic anhydride was reduced to give the diol (47) in 95% yield.Blomquist's method then requires the conversion of the diol (47) to the dibromide (49) with phosphorus tribromide, and sub sequent conversion td the salt (49). Treatment of the dibromide (48) with trimethylamine in a pyrex pressure vessel gave a precipitate of the salt (49).

When precipitation appeared to be complete, after about 2 weeks, the 96.

LiAIH4

H (48)

Ac NMe 3 Ac

H (51)

•. A 1g20 NMe3Br Me30H NMe38r

(52) H (50) H (49)

pressure was released and the white solid dried in vacuo . In practice the salt was always contaminated with the starting material and other products, which had the effect of poisoning the silver oxide, which was added later to replace the bromide counterion with hydroxyl moiety. Pyrolysis of the salt

(50) resulted in the formation of the diene (52) in only 18% overall yield from the dibromide. Another method of preparation of the diene, (which was heavily criticised by Blomquist)was pyrolysis of the diacetate (51),21 which may be prepared in 90% yield from the diol (47). This could then be pyrolysed to give the diene (52) in 'L50% yield. The overall yield for this route, starting from the diol (47) was 45%, which compares favourably to Blomquist's method 20 which gives an overall yield of 18%.

The diene (52) was then subjected to the usual sequence of reactions.

Monoepoxidation with peracetic acid gave the epoxide (53) in good yield 05%).

The signal in the 1H n.m.r. spectrum due to the methylene protons on the oxirane was a singlet (6 2.67). The neighbouring methylene group appeared as two distinct signals at 6 4.87 and 6 4.60.

97.

McCO3H hv

Fe(CO)5 .. T`Fe(CO)3 (52) (53) Ceiv (54)

+•

(55) • (56)

Irradiation of the epoxide in the presence of iron pentacarbonyl in the usual way gave an unusually low yield (29%) of the iron complex (54). This could be explained by the formation of olefin-tetracarbonyl iron complexes such as (57) and (58). Photolytic decarbonylation of the syn-complex (57),

(57)

followed by insertion into the oxirane ring would produce the desired ferrolactone (54). However, the anti-complex (58) would not react in this way, as it has been shown for hydrocarbon analogues that the bond in the cyclo- 22,23 propane ring cleaved is on the same side of the molecule as the iron species.

Another explanation is the lability of the ferrolactone (54). Early prepara- tions of this complex, involving the removal'of the solvent and excess iron pentacarbonyl by rotary evaporation gave very low yields of ferrolactone (54).

Other products were produced which are discussed in a later section.

98.

Oxidation of the iron complex (54) in acetonitrile gave a mixture of B- and 6-lactones (56) and (55) (1:3.66) in 70% yield. This is the only case examined in which the 6-lactone is the predominant product. The 6-lactone (55) had a carbonyl absorption at 1740 cm-1, and its 1H n.m.r. spectrum contained no olefinic signals. Signals at 6 4.63 and 6 2.98 were assigned

to the protons adjacent to the oxygen and carbonyl moieties respectively. -1 The 0-lactone had a typical carbonyl absorption at 1830 cm The ~H n.m.r. spectrum of (56) in deuterochloroform showed the olefinic protons as a singlet at 6 4.90, and methylene protons on the oxetanone ring as a typical

AB quartet centred at 6 4.18 with coupling constants 12 and 5 Hz. In contrast in D6 benzene the olefinic protons produced two signals at 6 4.35 and

4.60, and the ABq was shifted upfield to 6 3.57 (J 20Hz and 5Hz).

The 6-lactone (55)has also been prepared in 52.5% yield by oxidation of the ferrolactone (54) with ceric ammonium nitrate in ethanol/water at room temperature. Any S-lactone that may have been formed would almost certainly have been hydrolysed under these reaction conditions.

Reduction of the lactones (55) and (56) by lithium aluminium hydride gave the diols (60) and (59) in yields of 46.6 and 79% respectively. Both diols'had spectral properties in good agreement with their proposed structures and gave satisfactory microanalyses.

LiAI H4

(56)

LiAII-1

(55) 99.

Mass Spectral and N.m.r. data of the Ferrolactones

Fe(CO)3 11X-Fe(CO)

Anti -(4) Syn -(3)

(35) (41)

e(C0)3

HQ HA H (54) AA 100.

/HC \ A HC Fe(CO)

5 11 / HD HA HD

C5H11HD' (22)

(101)

Ferrolactones seldom produced a molecular ion in their mass spectra.

However, sequential loss of carbon monoxide and/or carbon dioxide from the complex generated a characteristic pattern of peaks in the mass spectrum (Table 1).

Ferrolactones (3) and (4),and (21) and (22) produced identical pairs of'spectra, confirming that they are structural isomers. TABLE 1

M M M M M M M M M M M M Ferrolactone M+ -CO -CO2 -2C0 -CO -3C0 -2C0 -4C0 -3C0 -Fe(CO)3 --Fe(C0)4 -Fe(CO)3 CO2 CO CO2 2 CO2

4 250, " 234 222 206 194 178 166 150 138 110 94

3 250 234 222 206 194 178 166 150 138 110 94

41? , 208 180 152 80

35 234 222 206 194 178

27 266 238 222 210 194 182 166 154 138 126 98 82

54 264 248 236 220 208 192 ` 180 164 152 124 108

21 308 280 264 252 236 224 208 196 180 140 124

22 280 264 252 236 224 208 196 180 140 124

102.

TABLE 2

CHEMICAL SHIFT SIGNAL COUPLING CONSTANT Ferrolactone H H HB .11C HD HD A A'

4.33 4.61 3.50 2.99 4 dd,6,10 dd,9,12 dd,2,9 dd,3,12

3 4.26 5.34 3.49 4,2.64 tr,6 - dd,8,12 dd,2,8

4.70 5.84-5.68 5.44 4.57-4.44 41 tr,6 m t,6 m

3.90 5.14-4.58 3.51 3.00 35 d,8 m d,9 d,13

4.42 4.64 3.63 73 3.09 dd,4,10 dd,9,14 dd,2,9 dd,2,13

4.41 4.76 3.71 3.00 74 dd,5,12 dd,8,13 dd,2,8 dd,1,13

3.97 3.44 2.64 27 d,2 d,2 d,2

4.03 - 3.23 4,2.E 54 -ABq,12,20 d,3 d,3

4.30 4.92-4.60 3.70 3.12 21 d,5 m d,8 d,12

4.18-3.82 4.54-4.88 4.18-3.82 22 m m m

4.37 4.16 4.77 4.04 101 s s d,4 dd,4,11

4.06 4.92 3.70 3.03 102 s s d,6 d,12 103.

Although there is tremendous variation in the chemical shifts of

protons in various positions on these complexes, several conclusions can

be drawn from analysis of these spectra. The ranges of chemical shifts

of various protons are shown in Table 3.

TABLE 3

NATURE OF PROTON PROTON CHEMICAL SHIFT RANGE

Methine HA HA, 4.02-4.07

Methylene HA HA, 3.97-4.06

HB 3.90-4.92

HC 4.60-5.44

Methine HD 4.51

Methylene HD .4.02-3.23

Methine HD, 4.30

Methylene' HD, 2.64-4.02

This may appear confusing as there are several overlapping regions.

However, in a given spectrum several relationships between'the protons

H HD exist. HD, is always at higher field than HD, both of which are A at higher field than HA or HA,. HAA- are always at higher field than

HB or HC, while HB and H vary in these relative positions. Thus, the

order may be expressed as: HBHC < HA < HHB < HD,. -INCREASING FIELD STRENGTH 104.

In a sterically crowded system such as the ferrolactone derived from cyclohexadiene (41) or the ferrolactone (22), the protons tend to be shifted into groups of HBHC and HAHDHD,. For this reason, the complex (22) can be assigned syn-stereochemistry and the complex (21) anti-stereochemistry.

The coupling constants of the rr-allylic system also fall into a pattern:

JAB " 6 BC 6-9

JCn 6-9 JrQ, 12-13

JDD' 2-3

As can be seen, the coupling constants JBC and JBD are very similar, hence assignment of protons solely by observed coupling constants should be avoided.

Mechanism of the Formation of Ferrolactones

The mechanism of the reaction between iron tetracarbonyl and vinyl 1 epoxides has not been investigated, but presumably follows a similar pathway to that proposed for the analogous vinyl cyclopropanes, which has-been discussed in the review. 105.

Fe(CO)4 or /~ Fe(C0)

(61) (62). •\ /co

(63) Insertion of the iron tetracarbonyl into the most electron rich bond

of the oxirane ring gives a metallocycle (61). Spontaneous insertion of

a carbonyl group, followed by coordination of the double bond to the coor-

dinatively unsaturated iron in (62) gives the ferrolactone (63). Alterna-

tively,coordination of the double bond to the iron in this species (61)

induces a carbonyl insertion into the Fe-0 .bond. It is possible that

olefin-tetracarbonyl complexes (64) may be formed, which may or may not

decarboxylate photolytically to give a coordinatively unsaturated iron

complex (65). This complex could subsequently insert into the oxirane

ring to give the a,ir-allyl complex (66). Insertion of carbon monoxide would

then lead to the ferrolactone (63). Alternatively, it is possible that

direct insertion of the iron tetracarbonyl into the oxirane ring may

.give the ferrolactone.

Thermal decarboxylation of simple olefin tetracarbonyl iron complexes

with adjacent cyclopropane rings is known to result in the formation of

diene-tricarbonyl iron complexes. Thus, to attempt to generate the un-

saturated species (65) by thermolysis of the complex (64) would be

inadvisable as similar pathways may be open to the heterocyclic system_

Iron tetracarbonyl has also been reported to deoxygenate epoxides 106.

in tetramethylurea at h1gh. temperatures t 0 g1ve . 0 1 e f'1ns. 24 (AI though

other iron species are known to deoxygenate epoxides at lower tempera- , 25 tures.) The 16-electron intermediate (67) is similar to an intermediate

proposed in the high temperature 'deoxygenation reaction. It seems quite

"reasonable that this intermediate could eliminate carbon dioxide to

+ CO "'/; 2 (67) / 3(O()Fe~ (68) produce tricarbonyl irondiene complexes (68). This alternative reaction

, cou~d go some way to explain the variation of yields experienced in the

preparation of ferrolactones.

Mechanism of the Oxidation of Ferrolactones

In every cas~ (except two) examined the oxidation of the ferro lactone

complexes resul ted in the preferential formation of !3-lactones (Table 4)

TABLE 4

Ferrolactone Ratio !3 Ratio 0

(41) 1 0 e(CO)3 (C(Fe((O~

(35) , 1 0 107.

Table 4/continued...

Ferrolactone Ratio a Ratio S

Fe(C0)3 1 0 H (21) or(22) 511

Fe(CO)3. .

Q Q 1.82 1 (3)

e (C0)3

2.56 1 (27)

1 3.66 (54) e(C0)3

,0,Fe(C0)3

0 (4) 1

The mechanism must thus take account of this fact.

A o,zr-allyl complex may be regarded as two contributing 2c,tr-complex-

.es (69) and (70). Oxidation of (69) will lead to S-lactones. In general

the oxidation of ferrolactones seems to result in the preferential forma-

tion of S-lactones, which seems to imply that the 5-ring form (69) is

favoured. The X-ray data for the syn.- and anti-complexes (3) and (4)

108.

e(C0)3

shows that the C2 atom has a high proportion of sp3 character, indicating that the predominant complex in the equilibrium is'(69). The ratio of

-S- to 6-lactones from the oxidation of the syn-complex (3) confirms this observation. Thus, the implication is that where oxidation of complexes leads to the formation of 8-lactones exclusively, the C2 atom has a completely tetrahedral nature.

Substituents of the 2-position would tend to destabilise the

5-membered ring, due to steric interactions with the bulky iron moiety, thus the 7-membered ring structure would be favoured. Conversely, a bulky substituent at position 4- would favour the 5-membered ring structure 109.

(69). This explanation appears to fit the data well. In complexes

(27), (3) and (54) increasing amounts of (5-lactone are formed as the substituent at the 2-position varies from a methyl group to a cyclohexane ring system: The anti-ferrolactone complex (4) would appear to be a complete exception to the rule. However, analysis of the X-ray structures of the syn- and anti-complexes (3) and (4) reveal that the cause of this apparent anomaly is structural rather than mechanistic.

An alternative mechanism which may be considered is the single electron oxidation of the iron complex to give an allylic cation. The

-e

stability of the respective allylic cations determines the composition of the products formed. This fits the data reasonably well, with the 110.

exception of the complex (54), in which the major product, the

d-lactone is the product from the closure of a primary allylic cation.

The tertiary allylic cation would produce the minor product, the

6-lactone. This could be explained by the steric hinderance of

the cation (71), which thus leads to the formation of the d-lactone.

Fe (CO)3 -e I ~0` K 0 • Fe (54) (C0)3 (71) (CO) 3

Structure of the Iron Carbonyl Complexes

The formation of ferrolactones by irradiation of epoxides with iron 1,2,26 pentacarbonyl is claimed to be stereospecific. Thus, the formation of two ferrolactones from the epoxide ( 2 ) was unexpected. These compounds were separable by fractional crystallisation or better by chromatography, and had similar but different 1H n.m.r. spectra. The oxidation of these complexes individually gave lactone products in C0) ..-Fe(C0) 3 3 Ō

(2) (4) (3)

completely different ratios. The minor component (the less polar ferro-

lactone by chromatography) gave the (S-lactone (7), while the major com-

ponent (more polar by t.l.c.) gave both the d-lactone (7) and the

S-lactone (6) in a ratio of 1:1.77.

(6)

A careful study of the X-ray structures (Figure 1) of these ferro-

lactones revealed the difference between the two isomers. In the major

isomer (3) the oxygen moiety and the tricarbonyl iron moiety are on the

same side of the ring structure and is therefore denoted syn- (3). The

minor isomer (4) shows the oxygen moiety on the opposite side to the iron and

is consequently denoted anti- (4). The anti- arrangement is a novel fea-

. ture for ferrolactone complexes and has not been observed by other

workers.

The differences in bond length between various component atoms, can

be used to explain most of the spectral differences between the syn- and

anti- isomers. FIGURE 1

(4) (3) 112.

C-Fe BOND DISTANCE R

C1 C2 C3 C4

SYN-isomer (3) 2.17 2.11 2.13 1.99

ANTI-isomer (4) 2.14 2.03 2.12 2.06

DIFFERENCE 0.03 0.08 0.01 0.07

Carbonyl groups are bonded to transition metals by a and Tr bonds. The

Tr bonds are formed by the overlap of the d orbitals of the transition metal

and Tr* orbitals of the carbonyl group. Donation of electron density into

the Tr* orbitals has the effect of weakening the C=0 bond. Thus the oxy-

carbonyl group of the syn-isomer which has a shorter, stronger C-Fe bond

than the anti-isomer, should have a weaker C=0 bond. This was observed

to be the case. The oxycarbonyl carbonyl absorption of the syn-isomer is -1 -1 at 1660 cm , i.e. some 20 cm lower than the similar absorption for the

anti-isomer.

Protons or ligands attached to iron moieties are shielded and there-

fore resonate at higher field than in the uncomplexed state. This effect

is thought to be due to the diamagnetic anisotropy of the iron atom rather

than to any major net flow of electron density onto the ligand. The

•difference in chemical shift between the inside and outside protons on the

allylic system can also be explained by the difference in proximity of

the magnetically anisotropic iron moiety. This principle can also be used

to explain the difference between the 1H n.m.r. spectra of the syn- and

anti-ferrolactones (3) and (4). In the anti-isomer the iron moiety is

0.08 Ā closer to the C2 atom in the allylic system. Consequently, the

proton on this atom is closer to the iron than in the syn-ferrolactone, 113.

and is more shielded. Hence the signal from this proton on C2 in the anti-isomer appears at S 4.61 compared to S 5.34 for the same proton in the syn-ferrolactone. Similarly, the iron moiety is 0.03 Ā closer to C1 in the anti-isomer (4). This in turn implies that the distance between the iron moiety and the syn-proton is increased, and hence deshielded. The proton resonates atS 2'99 in the anti-ferrolactone (4) compared to S 2.64 for the same proton in the syn-ferrolactone (3). This proton is likely to be far more sensitive to the proximity of the iron moiety than the anti- proton, which is further away from the iron in both complexes. This proton resonates at tid3.5 in both cases. The proton on C4 might have been expected to resonate differently in the syn- and anti-ferrolactones (3) and

(4). In fact, there is only 0.05 ppm difference between the signals in the respective spectra. Presumably the proton is unaffected as it is not within the coordination sphere of the iron moiety.

Although oxidation of the syn-isomer produces both S-'and &-lactones, a possible explanation for the preferential formation of the S-lactone can b derived from the observation that C3 is some.0.25 Ā out of the plane relative to C2, C4 and C5 and consequently has high sp3 character. The tetrahedral nature of this carbon is undoubtedly due to the contribution of a form in which there is significant a- iron bonded character. Although

Fe(CO) 3

the carbon atom C3 in the anti-isomer (4) has high sp3 character, carbon- carbon coupling on oxidation would lead to the highly strained trans-S- lactone and its formation is, therefore, precluded. Other work in these laboratories27 has shown that the steroidal epoxide (72) produces two 114.

iron complexes when irradiated with iron pentacarbonyl, in relative pro- portions of 1:1.

3(0C)

3( O C) Fe- ~_) - (74)

Once again, oxidation of the syn-ferrolactone (74) gives an mixture of S- and 6-lactones, (75).and (76), while oxidation of the anti-

1 ferrolactone yields the 6-lactone (76) exclusively. The 100 MHz H n.m.r. spectra of these complexes show similar patterns of chemical shift to the cyclopentane analogues. The proton HA has a similar chemical shift at

6 4.41 and d 4.42 in both cases as expected.

The proton Hc is 0.13 ppm further upfield, and HD' 0.09 ppm further downfield in the anti-complex, also as expected. The proton HD is 0.07 ppm further upfield in the anti-complex than in the s n-complex. This shift 115.

was expected but not observed in the case of the cyclopentane derivatives, as the iron is presumably slightly nearer to C1 in the anti-isomer. These minor differences can be readily accounted for by variations in the con- formation of the 6-ring versus the 5-ring systems.

Preparation of Lactams from Ferrolactams

Ferrolactam complexes have been prepared in the literature by a

variety of methods. The first and most direct of these routes involves

irradiation of a vinylaziridine1 in the presence of iron pentacarbonyl.

hv 100F 02Me --~. kvNCO2Me Fe (C0)5

,A more general method, however, involves the addition of amines to ferro- 28,29 lactone complexes in the presence of alumina. Considerable variation

e (C0)3 R N H 2 A1203 e (C0)3 11G.

in the yield obtained was observed for many of these ferrolactam complexes.

Other low yielding methods employ irradiation or thermolysis of iron carbo-

nyls in the presence of hindered oxazins. Although other workers have

Fe2(C0) Fe(C0)3 9 -~

(1)

thermolysed a few ferrolactams to obtain very low yields of S-lactams, the 3O,31 method is not at all synthetically useful. We argued that oxidation

offerrolactams could result in high yields of lactams by analogy with the

previous work with lactones. For this reason, a number of ferrolactam

complexes were prepared and their oxidation studied.

Preparation and Oxidation of Tricarbonyyl-(2,3,4-n-2,3-dimeth l-3-buten-2-

. yl)-1-benzylaminocarbonyl iron (77)

w4e(C0)3 ZnCl2 Ce" BzNH 2 e(CO)3

As aluminium oxide was found to be a poor catalyst to effect the

formation of ferrolactam complexes, several other Lewis acids, aluminium

chloride, zinc iodide, and zinc chloride were investigated. The best of

these was found to be zinc chloride, which gave the ferrolactam•(77) in

75% yield. The products were routinely isolated by chromatography on 117.

silica gel. However, the grade of the silica gel used was important, as

decomposition of the ferrolactam resulted if too fine a grade was used.

The ferrolactam had a characteristic carbonyl absorption at 1575 cm-1,

in its i.r. spectrum and an 1H n.m.r. spectrum similar to the corresponding

ferrolactone analogue, although as expected the protons adjacent to the

nitrogen moiety appear at higher field, i.e. at S 3.14 compared to S 3.97

in (27). The terminal allylic protons occur at d 3.42 (HA) and 2.44(HB).

Oxidation of the ferrolactam complex (77) with ceric ammonium nitrate

in ethanol gave the 8-lactam (79) and the 6-lactam (78) in 89,8% combined

yield and in a ratio of 1:1.68 respectively. The lactams (78) and (79) 53 were readily separated by chromatography on silica, and characterised

by their carbonyl absorptions at 1740 and 1640 cm-1 respectively. The 1H• n.m.r. spectrum of the 8-lactam was in good agreement with the proposed

structure (79), showing an AB quartet centred at S 3.10 for the protons

a to nitrogen and olefinic resonances at S 5.02 and S 4.90.The 1H n.m.r.

spectrum of the 6-lactam (78) showed structurally significant absorptions S at S 3.73 and/3.03 due to the protons adjacent to the nitrogen moiety and

-carbonyl group respectively. 118.

Preparation and Oxidation of Tricarbonyl (2',1,2-fl-l-ethyl-l-cyclo~enten-

-11 -y1)-2'-benzylaminocarbony1 iron (80)

e (C0)3 0 )3zNH2 Ce IV

Z_n 0.2

(3) & (4) (80) (81)

Treatment of the ferrolactones (3) and (4) with benzylamine and zinc

chloride gave the ferrolactam (80) in 48% yield. The structure of the

ferrolactam was assigned on the basis of its i.r. carbonyl stretching fre-

quency at 1580 cm-1, and analysis of its 1H n.m.r. spectrum.

(8 0)

Thus, the signal at 4.29 with an integral corresponding to 3 protons was

assigned to the proton HA and the benzyl methylene protons. The signal at

6 3.27 with an integral corresponding to 2 protons was assigned to the meth-

ylene protons HC, while the allylic proton HB is lost in the ring proton

.envelope. Oxidation of the ferrolactam complex (80) with eerie ammonium

nitrate in ethanol gave the 0-lactam (81) in 84% yield. This compound

had a typical carbonyl absorption at 1740 cm , and. its 1H n.m.r. spectrum

contained one olefinic resonance S 5.67 and three signals from theazetidi-

none ring protons at S 4.0,53.34 and S 3.09. It is known that on- oxetan-3-one

analogues the cis coupling constant for adjacent protons on the four membere

ring is 'L 6 Hz, whereas the trans coupling contant is 3 Hz. Hence the

triplet at S 3.34 (J 6 Hz) can be assigned to HB, and is a product of gemina 119.

coupling to H (6 Hz) and cis-coupling to HA (6 Hz). The quartet 6 3.09 c (J = 6 Hz and 3 Hz) is therefore assigned to HC, and is the product of

(6 Hz) and trans-coupling to HA (3 Hz). The mul- geminal coupling to HB tiplet at 6 4.0 is an unresolved quartet and can be assigned to the proton HA. The observation of the formation of this iron complex (80) rather than the regioisomer (82) by nucleophilic attack of the amine on 22,23 the allylic system is in agreement with other reports.

!/FeO)3

Bz \ HB ~Bz

(82)

Preparation and Oxidation of Tricarbonyl (2, 3,,4-n-3-nonen-2-y1)-1-benzyl- aminocarbonx1 iron (83)

\e(CO)3 BiNH Z 3(OC )Fe \I

ZnCt2 C5H11 11 5 11 (21) (22) (83)

N Bz

C5H11 ($4)

When a mixture of the complexes (21) and (22) were treated with benzylamine in the presence of zinc chloride a single product was formed by t.l.c. After chromatography, the ferrolactam (83) was isolated as white crystals in 82% yield. This compound had a characteristic carbonyl stret- 120.

-1 1H ching frequency at 1590 cm The 60, 100 and 220 H Hz n.m.r. spectra of this compound were difficult to assign and did not establish the homo- geneity of this compound. However, a scale expanded decoupled 360 MHz 32 1H n.m.r. spectrum provided an easily assignable first order spectrum.

Detailed analysis of this spectrum gives much information about the structure of the ferrolactam (83)

NCH20 F

FROWN CHEMICAL SHIFT

HA HB HC HD HE HF HG BENZYL PHENYL

1.727 2.277 3.780 4.625 4.169 5.261 3.091 4.311 4.155 730°- 7120 m m m dd dd dd d d d jp

JA' - * 4 - - - - J * - 10 - - - - - B JC 4 10 - 12 - - - JD - - 12 - 9 - -

JE - - - 9 - 6 -

JF - - - - 4 - 13

JG - - - - - 13 -

The trans-relationship between HD and He is clearly shown by the

12 Hz coupling constant. Similarly, protons HD and HE .have a coupling constant of 9 Hz, indicating a cis-relationship. The protons HF and

HG are non equivalent. The proton at lower field (HF J 6 Hz) couples to HE while the proton H which is at higher field shows no coupling to G 121.

HE thus implying an angle of ^490° between the two protons. Protons HA and H are coupled (4 Hz), while HB at lower field is coupled to H by a very small constant. These results suggest a cis-relationship between o HA and H and an angle of close to 90 between HB and HC.

The ferrolactam (83) was the sole product from the reaction of the ferrolactones (21) and (22) with benzylamine, presumably because- the less thermodynamically stable Z-ferrolactone (22) was inverted to an E-form during the course of the reaction. Oxidation of the ferrolactam complex

(83) produced a 0-lactam (84) in 63.7% yield with a characteristic -1 carbonyl stretching frequency at 1745 cm . Analysis of the H1 n.m.r. spectrum of this compound revealed that the oxidation had produced both cis- and trans- geometries about the double bond. The signals due to the olefinic protons 6 6.0-5.0 were very complex and did not simplify when adjacent protons were decoupled. The difference in the environment of the methylene benzyl protons resulted in an unusual. signal at 6 4.37 as a quintet. Subsequent addition of shift reagent (Eu(FOD)2) showed this to be a combination of a triplet and a AB quartet. The constitution of the azetidinone ring was determined by the observation of signals at

62.92 and 6 3-32 due to the methylene protons ,on the ring system. These showed the characteristic cis- plus geminal and trans- plus geminal and a triplet (5Hz) couplings to give a quartet (J 5Hz and 3 Hz)/respectively. Thus, although the oxidation of the iron complexes generally proceeds with retention of configuration, double bond geometries can be isomerised presumably due to the generation of cationic intermediates during the course of the oxidation, which can rotate freely. It is thought that the oxidation may be more stereospecific if performed at a lower temperature. Unfortunately, all other ferrolactones treated with benzylamine in this way have so far failed to produce significant amounts of ferrolactams. 122.

Mechanisms for the Formation and Oxidation of Ferrolactams

The mechanism of the addition of amines to ferrolactones has not been investigated in detail, but certain features of the reaction have 28,29 become apparent. Other workers have proposed that amines add regioselectivity to the allylic system of a ferrolactone, displacing the oxycarbonyl moiety followed by the formation of a new C—N bond.

HZN R e(C013 RNH2 R ~~- Fe (CO) CO2 Zn CI or Alt 03

The process is catalysed by the presence of a Lewis acid such as alumina or zinc chloride. The regioselectivity of the addition of the amine to the carbon skeleton of the ferrolactone was confirmed by the use of unsy- metrical carbon frameworks. However, other possible mechanisms cannot be excluded.

33 It is interesting to note an observation in a paper by Angelici on carbamoyl complexes. Carbonyl ligands with force constants greater than 17.2 m dyn/Ā readily form carbamoyl complexes. Those with values between 16.0 and 17.0 form reversible equilibria, and < 16.0 do not react.

In other words, carbonyl ligands with absorption below about 2000 cm-1 do not form carbamoyl complexes. All the ferrolactones so far treated -1 with benzylamine have a sharp carbonyl absorption above 2000 cm , 123,

but a much stronger, broader absorption below this value. This would seem to indicate that one of the carbonyl groups in a given position on the• iron moiety is available for the formation of a carbamoyl bond. Thus, if this group is sterically hindered, ferrolactam complex formation could be inhibited.

This hypothesis could explain the diversity of the reactivity of the complex es examined with benzylamine.

It is known that the ferrolactone complexes can be regarded as two contributing 2a,x-complexes, of which the five ring form is favoured (as shown by the X-ray data). Oxidation of each respective form leading to

S- and S-lactones. It seems likely that a similar situation exists in the ferrolactam complexes. Substituents at the 2-position would also be expected to affect the observed geometry of complexes. Model studies

indicate that this effect would be more pronounced in the ferrolactam complexes due to the greater steric congestion about the 5-membered ring.

In the only case examined where there is a substituent in the 2-position, complex (77), the ratio of lactam products 5:8 is 1.68:1, (compared to a ratio of 0.39:1 for the lactone analogues) which is in accord with this prediction.

PreQaration of N-H Lactams

Treatment of ferrolactones with amines in the presence of zinc 124.

chloride produces ferrolactams. By analogy N - unsubstituted ferrolactams

should be produced by the reaction of ammonia with ferrolactones in the

presence of zinc chloride. Tricarbonyl-(2,3,4-n-2,3-dimethyl-3-buten-2-

yl)-l-oxycarbonyl iron (27) when treated with ammonia under pressure gave

3,G-dihydro-4,5-dimethyl-2-pyridone (85) in 38% yield. S-Lactam was

not detected. The lactam (85) had a carbonyl absorption at 1650 cm,

(27) (85)

which was rather weak, probably due to some of the enol form. The n.m.r. spectrum contained signals at 6 4.03 due to the methylene protons adjacent

to the carbonyl group, 6 3.18 due to the NH proton and 6 3.05 due to

the methylene protons adjacent to the heteroatom. This ferrolactone (27) was the only one which reacted in this manner. Unfortunately, all other complexes treated in a similar way have so far failed to give isolable yields of .lactams. New methods of introducing the NH group are being 51 studied in these laboratories. Owing to the poor results obtained above, in an effort to produce NH 0-lactams, we decided to briefly assess the use of vinyl aziridines as suitable precursors. It was decided initially to prepare the vinyl aziridines by Stogryn's method,34 from the readily available vinyl epoxides. However, owing to the instability of certain of the vinyl epoxides in aqueous media, the route was not very successful.

Similar observations had been noted by other workers.35 Nevertheless several vinyl aziridines ( (86), (87) and (88)) were prepared in this way (albeit in very low yield), and were treated with iron pentacarbonyl in the usual way. A complex mixture of products was usually obtained • and further chemistry was therefore abandoned. 125.

iE\R

(86) (87) (88)

43, 44 Other routes to vinyl aziridines were investigated without success.

126.

MISCELLANEOUS CHEMISTRY

Rearrangements of a-Vinyl-S-lactones and Thermolysis of Ferrolactones

(7) Zn Ct2

(6)

(89)

Attempts to convert the f3-lactone (6) to the 8-lactone (7) by a

1,3-acyl shift by either irradiation, thermolysis, treatment with acid or boron trifluoride diethyl etherate resulted in.•failure. However, an unusual product was isolated from the reaction of the 8-lactone (6) with zinc chloride in benzene. The product was obtained in 40% yield and -1 showed a carbonyl absorption in its i.r. spectrum at 1790 cm , which is characteristic of a five-membered ring lactone. Its 1H n.m.r. spectrum contains signals characteristic of a vinyl group IS 6.22 and S 5.48-5.12, and a methine ring junction a to the oxygen atom at S 4.90. The signals are 1 all slightly downfield of their counterparts in the 0-lactone H n.m.r. spectrum. The bicyclic lactone (89) is therefore proposed as the structure which best fits all the spectroscopic and analytical data. A mechanism to account for its formation is outlined below. This type of rearrangement involving the formation of y-lactones from a-vinyl-S-lactones seems to be quite general. An isomeric 3-lactone (90)36 when treated under similar

127.

conditions gave the butenolide (91) in 20% yield. The butenolide produced characteristic carbonyl absorption in its i.r. spectrum at 1765 cm 1, and

Zn Cl2

(9Ō) (91)

signals in the H n.m.r. spectrum at d 4.50 and 2.16 due to the ring junction proton HA and the methyl group respectively. Although the yields for both these rearrangements are low at present, variation of the reaction conditions is expected to lead to a substantial improvement.

Pyrolysis of the iron complex (41) has been reported to produce the 1 hydroxy diene complex (92). This transformation is directly analogous

Fe(CO)3

(41) (92) to the thermolysis of hydrocarbon c,rr-allyl complexes, discussed in an earlier section.

128.

Thermolysis of the complex (54) produces an intermediate complex, presumably (93), which on oxidation gives the enal (94) in 60% yield.

Fe(CO)3

(54) i to) / Fe(CO) 3

(93)

-1 The enal (94) has characteristic carbonyl absorptions at 1665 cm (C°0) -1 and 1635 cm (C=C) in its i.r. spectrum. The 1H n.m.r. spectrum of this compound contains a signal at Ō 2.13 due to the 7-allylic protons and a signal at S 1.83-1.45 from the four non-allylic ring protons.

37 Approaches toward the Synthesis of Boschnia Lactone (100)

In an earlier section it was reported that 1,2-dimethylenecyclohexane can be converted to an isochromanone derivative via ferrolactone complexes. It seemed reasonable, therefore, to attempt the synthesis of the natural product Boschnialactone from the intermediate spiroepoxide (99).

CHZ SMez

(98) (99) (100) 129.

•The preparation of (98) should be straightforward from the readily available

2-carbethoxy-cyclopentanone (95). Selective methylation38 of (95) gave

(96) in 80% yield. Formation of the Mannich base (97) by Poulter's method

with an analogue of Eschenmoser's salt (101) proceeded smoothly in 85%

CO2Et CO2Et

(97) • (95) (96)

yield.39 Initial attempts to decarboxylate this material by standard

methods, such as hydrochloric or sulphuric acids, gave complex mixtures

of products.40 Trimethylsilyl iodide also failed to produce the desired

product,41 although we are hopeful that after suitable modification, the

route will be successful and it is being continued by other workers in 51 these laboratories. 130.

EXPERIMENTAL

Melting points were determined on a Kofler block and are uncorrected.

• I.R. spectra were recorded with a Perkin-Elmer 237 grating spectrometer.

Spectra of solids were obtained from nujol mulls, and oils from thin films.

N.m.r. spectra were recorded with a Varian E.M. 360A, XL 100, or Bruker-360.

Apparent J values are quoted throughout. Spectra were obtained for solutions

in deuterochloroform with a tetramethylsilane internal standard unless other-

wise stated. Mass spectra were recorded with a V.G. Micromass 7070 spectro-

meter. All reaction solvents were purified before use. Petroleum ether was

redistilled, acetonitrile was dried over calcium hydride and redistilled,

benzene was dried over sodium wire and degassed overnight with argon, and

ether was dried over sodium wire. All other reagents were used as received.

Product solutions were dried over sodium sulphate unless otherwise stated.

Merck-Kieselgel 60 (0.040-0.063 mm) and B.D.H. Florisil (200-300 U.S. mesh)

were used for the chromatography of products.

GENERAL METHODS

A. Preparation of Epoxides

m-Chloroperbenzoic acid was added in one portion to a stirred slurry

of the diene and sodium carbonate in methylene chloride at 0°C. After all

the peracid had been consumed (as shown by starch/iodide paper), the mix-

" ture was filtered and the precipitate -washed with methylene chloride. Removal

of the solvent by either distillation or rotary evaporation from the combined

extracts gave the crude product as an oil. 131.

B. Preparation of the Ferrolactones

A stirred solution of the epoxide and iron pentacarbonyl in

benzene, under an argon atmosphere was irradiated with either a 450 W

Hanovia medium pressure Hg discharge tube or 2 x 740 W Hanovia medium pressure discharge tubes, through a Chance OX1 filter. The concentra-

t,on of the ferrolactone was maximised (as shown by i.r.), and the reaction mixture frozen. Freeze drying of the mixture in vacua gave the ferrolactone. The crude product was stirred with ether and filtered

through a plug of celite. The solution was reduced to a small volume

and cooled, Addition of petroleum ether precipitated the ferrolactone

as white crystals.

C. Oxidation of the Ferrolactones

The ferrolactone was added in one portion to a stirred slurry

of ceric ammonium nitrate at -5°C in acetonitrile. After all the

ferrolactone had been oxidised (as shown by t.l.c.) the solvent was removed under reduced pressure. The residue was redissolved in the

minimum amount of water, and the resultant solution extracted with ether.

The combined extracts were dried (MgSO4) and the solvent removed

under reduced pressure to give the lactone or lactones as an oil. The

crude product or products were purified by chromatography on Florisil. 132,

D. Reduction of the Lactones

A solution of the lactone in tetrahydrofuran was added dropwise to a stirred slurry of lithium aluminium hydride in .tetrahydrofuran.

After all the lactone had been reduced (as shown by t.l.c.), saturated sodium sulphate was added to the reaction mixture until a white precipitate was produced. The mixture was filtered and the precipitate was washed with ether. The combined organic extracts were dried, and

the solvent removed under reduced pressure to give the crude diol as an oil,

E. Preparation of the Ferrolactams

Zinc chloride was added to a stirred mixture of the ferrolactone and benzylamine in ether. After all the ferrolactone had reacted

(as shown by t.l.c.), the mixture was filtered, and the precipitate washed with ether. The combined extracts were dried, and the solvent removed under reduced pressure to give the crude product as an oil.

Chromatography of the crude product or silica gave the ferrolactam.

F. Oxidation of the Ferrolactams

Ceric ammonium nitrate in ethanol was added to a stirrred solution of the ferrolactam in ethanol maintained between -2° and

+2°C. After the additon was complete, the mixture was allowed to attain room temperature and was stirred until all the ferrolactam had 133.

been oxidised (as shown by t.l.c.). The solvent was removed under reduced pressure by rotary evaporation and the residue redissolved in water, The solution was extracted with ether, and the combined extracts dried and evaporated under reduced pressure to give the lactam or lactams.

Preparation of 1-Vinylcyclopentanol (5)

Cyclopentanone (21 g, 0.25 ml) in tetrahydrofuran (20 ml) was added dropwise to a solution of vinyl magnesium bromide,42 [from magnesium (5.37 g, 0.224 mol) and vinyl bromide (25.7 g, 0.24 mol), , at such a rate so as to maintain gentle boiling of the solvent. The reaction mixture was allowed to cool and was added to water (200 ml).

The aqueous suspension was extracted with ether (3 x 100 ml) and the combined organic phases dried (MgSO4). Removal of the solvent under reduced pressure gave 1-vinylcyclopentanol (5) (25.65 g, 95%) as an oil,

b.p. 75°C at 35 mmHg, S 6.0 (1H, dd, J 10 and 18 Hz), 5.36-4.76 (2H, m),

and 2.13-1,50 (11H, m, 1 proton exch. D20).

Preparation of 1-Vinylcyclopentene3 (1)

Iodine (0.5 g, 1.96 mmol) was added to 1-vinylcyclopentanol (5)

(17,1 g, 0.152 mol). The pressure was reduced to 75 mmHg and the tempera- 0 ture of the reaction mixture increased to 100 C. The products which 134.

distilled were collected in a flask, cooled to -78°C. The crude product was allowed to warm to room temperature and dried. Filtration gave 1-vinylcyclopentene (1) as an oil (12.26 g, 35%), b.p. 58° at 140 mmHg; S 6.56

(1H, dd, J 10 and 18 Hz), 5.70-5.10 (2H, m), 4.87 (1H, s), and 2.6-1.5 (6H, m) .

Preparation of 1,2-E•ox'-1-vin lc clopentane (2)

• m-Chloroperbenzoic acid (5.87 g, 34 mmol) was added in small portions to a stirred mixture of 1-vinylcyclopentene (1) (3.2 g, 34 mmol), dichloromethane (240 ml) and saturated sodium bicarbonate solution (60 ml) at 0°C.

After 3 h, the mixture was allowed to warm to room temperature, and the organic phase removed and dried. Removal of the solvent under reduced pressure and distillation of the residue gave 1,2-epoxy-l-vinylcyclopentane (2) (2.1 g, 59%) as an oil, b.p. 50° at 10 mm Hg; 5.86 (1H, dd, J 10 and 17 Hz), 5.48-5.08 (2H, m), 3.35 (1H, s), and 2.35-1.15 (6H, m).

Preparation of Syn- and Anti-Tricarbonyl-{1,l',2'-n-l-vinylcyciopentan-l-yl)-

2-oxycarbonyl iron (3) and (4)

A mixture of 1,2-epoxy-l-vinylcyclopentane (2) (2g, 18.2 mmol) and iron pentacarbonyl (15 g, 76.5 mmol) in benzene (500 ml) under an Argon atmosphere was irradiated (Phillips 125 HPK mercury discharge tube) for 4 h.

After this time the solvent and excess pentacarbonyl iron were removed under reduced pressure to give the crude product as a brown oil. This oil was redissolved in the minimum amount of ether, filtered through a plug of celite, and cooled to 0°C. Addition of petroleum ether (b.p. 30-40°) precipitated syn- and anti-tricarbonyl-(1,1',2'-n-l-vinylcyclopentan-l-yl)-2-oxycarbonyl iron (3) and (4) (4 g, 79%) as white crystals. Chromatography of the mixture on Florisil (eluted with ether/petroleum ether 1:6) gave (a) anti-

135.

tricarb n-l-vinylpentan-l-yl)-2-oxycarbonl iron (4) (0.54 g,

10.6%) as white crystals, m.p. 88-89°C (decomp.); vmax 2925, 2075, 2000,

1680, 1455, 1390, 1135, 1115, 1072, 995, 940, and 670 cm 1; 6 4.61 (1H, dd, J 12 Hz and 9 Hz), 4.44-4.22 (1H, br m), 3.5 (1H, dd, J 9Hz and 2Hz), 2.99

(1H, dd, J 12 Hz and 3 Hz), 2.76-2.44 (1H, br m), 2.42-1.96 (4H, m), and 1.64-

1.20 (1H, br m); m/e 250, 234, 222, 206, 194, 178, 166, 150, 148, 138, 122,

110, 95, and 94 (Found: C, 47.35; H, 3.55. C11H1005Fe requires C, 47.5; H, 3.6%) and (b) syn-tricarbonyl-(1,1',2'-n-l-vinylpentan-l-y1)-2-oxycarbonyl iron (3) (3.24 g, 64.4%) as white crystals, m.p. 96-97° (decomp); vmax 2925, 2075, 2000, 1990, 1655, 1465, 1455, 1375, 1095, 1085, 1010, 990, 945, and

670 cm 1; 6 5.34 (1H, dd, J 8Hz and 12 Hz), 4.26 (1H, t, J 6 Hz), 3.49 (1H, dd, J 8Hz and 2Hz), 2.78-2.42 (3H, m), and 2.38-1.52 (4H, m); m/e

250, 234, 222, 206, 194, 178, 166, 150, 148, 122, 110, and 94 (Found:

H1505Fe requires C, 47.4; H, 3.6%). C, 47.5; H, 3.9. C11

Pre.aration of 1-Vinyl-6-oxabic clo 3.2.0 he.t-7-one (6) and 2-Oxabicyclo-

[4.4.0] dec-5-en-3-one (7) from (3)

Syn-tricarbonyl-(1,1',2'-n-l-vinylcyclopentan-l-yl)-2-oxycarbonyl- iron (3) (910 mg, 3.27 mmol) in acetonitrile (50 ml) was oxidised with ceric ammonium nitrate (10.67 g, 19.47 mmol), and worked up in the usual way. Chromatography of the crude product on Florisil, (eluted with ether/ petroleum ether, 1:6, then ether) gave 1-vinyl-6-oxabicyclo [3.2.O]hept-7- 2900, 1820, 1635, 1130, 1100, 1020, one (6) (230 mg, 51%) as an oil, vmax -1 960, 910, 875, and 820 cm ; 6 6.23-5.48 (1H, m), 5.56-5.15 (2H, m), 4.7

(1H, d, J 3Hz), and 2.41-0.63 (6H, m); and 2-oxabicyclo[4.4.0]dec-5-en-3- one (7) (137 mg, 28.8%) as an oil, Vmax 2930, 1745, 1300, 1220, 1210, 1130,

1090, and 1065 cm-1; 6 (C6D6) 5.08-4.86 (1H, m), 4.32 (1H, br s), 2.66- 2.35 (2H, m), and 2.20-0.73 (6H, m) (Found: C, 69.35; H, 7.55. 0 C8H10 2 136.

requires C, 69.55; H, 7.30%).

Alternative Preparation of 1-Vinylbicyclo S.2.0 6-oxahept-7-one (6)

Syn- and anti-tricarbony1-(1,1',2'-rfi-1-vinylcyclopentan-l-yl)-2- oxycarbonyl iron (3) and (4) (0.5 g, 1.8 mmol) were added in one portion to a stirred mixture of eerie ammonium nitrate (10 g, 18.2 mmol) in dimethyl sulphoxide (40 ml) and water (20 ml). The mixture was warmed to 70°C over a 2 h. period, allowed to cool to room temperature, and added to water (250 ml). The aqueous solution was extracted with ether (6 x 50 ml), and the combined organic extracts dried, and reduced to a small volume (20 ml) under reduced pressure. The concentrated extracts were washed with water

(2 x 20 ml), dried and evaporated under reduced pressure to give the 1-vinyl-

6-oxabicyclo[3.2.0]hept-7-one (6) (90 mg, 36.2%), identical with the previously prepared sample (by t.l ,., i.r. and1H n.m.r.).

^NJ

Preparation of 2-Oxabicyclo~4.4.01 -dec-5-en-3-one (7) , rom (4)

} Anti-tricarbonyl-(1,1',2'-r)-1-vinylcyclopentan-1-y1)-3-bxycarbonyl iron (4) (20 mg, 0.072 mmol) in acetonitrile (3 ml) was oxidised with eerie ammonium nitrate (0.235 g, 0.43 mmol) and worked up in the usual way, (ter drying over Na2SO4/Na2CO3) to give 2-oxabicyclo[4.4.O]dec-5_en-3-one (7)

(7.4 mg, 74.5%) identical by t.l.c., i.r. and 1H n.m.r'. to the previously prepared sample.

Preparation of 2-14ydroxyy72-vinylcyclopentanol (8)

1-Vinyl-6-oxabicyclo[3.2.0]hept-7-one (6) (150 mg, 1.1 mmol) in 137.

tetrahydrofuran (5 ml) was reduced with lithium aluminium hydride (200 mg,

5.2 mmol) in tetrahydrofuran (10 ml) for 3 h, and worked up in the usual way to give 2-methylhydroxy-2-vinylcyclopentanol (8) as an oil, vmax 3325,

2925, 1640, 1410, 1050, 1040, 1010, and 910 cm 1;45.86 (1H, q, J 10 Hz and

18 Hz), 5.25-4.96 (2H, m), 4.25-4.11 (1H, m), 3.69 (2H, s), 3.05-2.25 (2H, m), and 2.15-1.18 (6H, m).

Preparation of 2-(CiS-3'7hrydroxypropylidyl)cyclopentanol (9) •

2-Oxabicyclo [4.4.0] dec-5-en-3-one (7) (138 mg, 1 mmol) in tetra- . hydrofuran (5 ml) was reduced with lithium aluminium hydride (200 mg,

5.2 mmol) in tetrahydrofuran (10 ml), and worked up in the usual way to give 2-(syn-3'-hydroxyyrop lidyl)cyclopentanol (9) (100 mg, 70.4%) as white crystals, m.p. 52.5-53°; vmax 3350, 3280, 2950, 2920, 1440, 1065, 1030,

970, 945, 920, 900, 875, and 865 cm 1; d 5.40 (1H, t, J 8Hz), 4.50 (1H, s),,4.06 (2H, s), 3.87-3.16 (2H, m), and 2.74-0.80 (8H, m) (Found: C,

67.6; H, 10.2. C8H1402 requires C, 67.55; H, 10.2%).

Preparation of 4-Cyclopentylbutanoic acid (11)

2-Oxabivyclo [4.4.0] dec-5-en-3-one (7) (50 mg, 0.36 mmol) in ethanol

(10 ml) was hydrogenated at ambient temperature over 10% palladium-charcoal

(10 mg) for 4 h. At the end of this time, the mixture was filtered through a plug of celite and the solvent removed under reduced pressure to give the 4-cyclopentenylbutanoic acid (11) (47 mg, 91%), vmax 3100 (br), 2950, -1 1715 (br), 1260, and 1185 cm ;8 2:53-2.13 (2H, m), and 2.03-0.73 (14H, m). 138.

Preparation of 1-Nonen-3-ol (13)

Heptaldehyde (5.5 g, 48 mmol) in tetrahydrofuran (10 ml) was added to vinyl magnesium bromide42 (6.5 g, 50 mmol) [from magnesium (1.2 g, 50 mmol) and vinyl bromide (6.1 g, 57.5 mmol)] in tetrahydrofuran (30 ml), so that the mixture boiled gently. After the addition was complete, the mixture was stirred for a further 3 h at room temperature. Water was added dropwise to the mixture until a white precipitate formed, which was removed by filtration, and washed with ether (3 x 30 ml). The combined extracts were dried and the solvent removed under reduced pressure to give the 1-nonen-3-ol (13)

°C at 12 mmHg; 3300, 2900, 2850, 1640, (5.5 g, 80%) as an oil, b.p. 89 max 1455, 990, and 925 cm-1, S 6.13-5.53 (1H, m), 5.26-4.84 (2H, m), 3.92 (1H, m) and 1.78-0.64 (14H, m).

Preparation of 3-Acetox~y-1-nonene (19)

Acetyl chloride (35 g, 0.446 mol) was added dropwise to a stirred solution of 1-nonien-3-ol (13) (13.0 g, 91.5 mmol) and pyridine (10.86 g,

0.137 mol) in ether (100m1). After 7 h the mixture was added to excess sodium bicarbonate solution. The ethereal layer was separated, washed with dilute hydrochloric acid (50 ml, 10%) sodium bicarbonate solution (saturated,

50 ml), and dried. Removal of the solvent under reduced pressure gave the

3-acetoxy-l-nonene (19) (16.7 g, 93%) as an oil, 6 6.04-5.4 (1H, m), 5.32-

4.85 (2H, m), 1.98 (3H, s), and 1.8-0.7 (9H, m), which could be used without any further purification.

4 Preparation of trans$-non-2-eny1-0-4-methylphenylthiocarbonate (16)

1-Nonen-3-ol (13) (0.76 g, 5.3 mmol) in dichloromethane (5 ml) 139.

was added dropwise to a stirred solution of 0-4-methylphenylchlorothioformate

(16)5 (lg, 5.3 mmol) in pyridine (5 ml) at 0°C. After stirring for lh, water

(50 ml) and dichloromethane (20 ml) were added. The organic phase was separated washed with hydrochloric acid (10%, 3 x 10 ml), dried and the solvent removed under reduced pressure to give the crude product as a black oil. Chroma- tography of this material on silica gel (eluted with ether /petroleum ether

1:1)gave trans-5-non-3-eny1-0-4-methylphenylthiocarbonate (16) (1.0 g, 57%) as a red oil, Vmax 2900, 1725, 1500, 1270, 1210, 1190, 1165, 1100, 1020, 960, -1 and 800 cm , 6 7.16-6.83 (4H, m), 5.66-5.40 (1H, m);.3.43 (2H,d,6Hz), 2.33

(3H, s), 1.97 (2H, m), and 1.70-0.70 (11H, m).

Preparation of (Carboxysul honyl)triethylammonium Hydroxide inner Salt

Methyl Ester (17)6

Methanol (4.9 g, 0.15 mol) in benzene (5 ml) was added dropwise to ciilorosulphonyl isocyanate (17.0 g, 0.12 mol) in benzene (50 ml)-at

10°C. Removal of the solvent and excess methanol under reduced pressure gave the carbomethoxysulphamoyl chloride (20.2 g, 97%). Carbomethoxysuiphamoyl chloride (6.94 g, 40 mmol) in benzene (100 ml) was added to triethylamine

(9.2 g, 90mmol) in (950 ml) at room temperature under a nitrogen atmosphere.

After 1 h, the triethylamine hydrochloride removed by filtration under nitrogen, and the solvent removed under reduced pressure, gave the (carboxysul-w phonyl)triethylammonium hydroxide inner salt methyl ester (17) (8.5 g, 89.2%), m.p. 67-69°C (lit.,6 m.p. 71-72°), 6 3.73 (3H, s), 3.50 (6H, q, J-$ Hz and 16Hz) and 1.45 (9H, t, J 7Hz).

Preparation of 1,3-Nonadiene (12)7

A stirred solution of 3-acetoxy-l-nonene (19) (15.66 g, 0.126 mol), triphenyl phosphine (2.24 g, 8.5 mmol) and Palladium (II) acetate (78 mg, 140.

0.3 mmol) in dioxan (85 ml) was warmed to 100°C for 24 h. At the end of this time, the mixture was poured into a mixture of water (300 ml) and petrol (200 ml). The petrol layer was separated, washed with water (3 x

100 ml), dried and the solvent removed by distillation. The product was removed from the residue by distillation under reduced pressure to give the diene (12) (8.38 g, 78.5%) as an oil, b.p. 96° at 14.3 cm Hg; S 6.83-

4.75 (5H, m) and 2.4-0.64 (71H, m).

Preparation of 2,3-Epoxy-l-nonene (20)

Oxidation of 1,3-nonadiene (19) (3.66 g, 2.95 mmol) with m-chloro-

perbenzoic acid (6.0 g, 85%, 2.99 mmol) in the presence of sodium carbonate

(5 g, 4.7 mmol) in dichloromethane (50 ml) in the usual way, gave 2,3-epoxy-

1-nonene (20) (2.8 g, 67.5%) as an oil, b.p. 66° at 13 mm Hg; S 5.74-4.96

(3H, m) and 3.12-0.66 (15H, m).

Preparation of E-and Z-Tricarbonyl-(1,2,3-11-l-nonen-3-y1)-4-oxycarbonyl Iron

(21) and (22)

A stirred solution of 2,3-epoxy-l-nonene (20) (0.8 g, 5.7 mmol)

and iron pentacarbonyl (8 g, 48.8 mmol) in benzene (500 ml) was irradiated

and worked up in the usual way. Chromatography of the crude product (1.37 g,

77%) on florisil gave E-tricarbonyl-(1,2,3-p-l=nonen-3-y1)-4-oxycarbonyl

iron (21) (240 mg, 17.5%) as white crystals, m.p. 97° (decomp.); vmax

2900, 2050, 1992, 1665, and 1455 cm-1, 5 4.96-4.60 (2H, m), 4.30 (1H, d,

J 5Hz), 3.70 (1H, d, J 8Hz), 3.12 (1H, d, J 12 Hz), 1.70-1.10 (8H, m), and

0.88 (3H,br s); mfe 308 (M+), 280, 264, 236, 224, 208, 196, 180, 140, and

124 (Found: C, 50.95; H, 5.2. Fe05 requires C, 50.65; H, 5.2%), C13H16 141.

and Z-tricarbonyl.( 1,2,3-n-l-nonen-3-y1)-4-ox~ycarbonyl iron (22) (575 mg, 42%)

as yellow crystals, m.p. 41-43°; 2850, 2050, 1990, 1660, 1460, 1370,1110, vmax 1070, 1040, 980, and 660 cm 1; 5 4.90-4.54 (2H, m), 4.18-3.82 (3H, m), 2.4-

2.08 (1H, m), 1.90-1.14 (8H, m) and 0.92 (3H,br s);m/e 280, 264, 252, 236, 224,

05Fe requires 208, 196, 180, 140, and 124 (Found: C, 50.85; H, 5.3. C13H16 C, 50.65; H, 5.25).

Preparation of cis-3-Pentyl-4-vinyloxetan-2-one (24)

Z-Tricarbonyl(1,2,3-n-l-nonen-3-yl)-4-oxycarbonyl. iron (22) (240 mg,

0.78 mmol) in acetonitrile (25 ml) was oxidised with ceric ammonium nitrate

(2.56 g, 4.67 mmol) at -30°, in the usual way. Chromatography of the crude

product on Florisil (eluted with petroleum ether/ether, 10:1) gave the cis-

3-pentyl-4-vinyloxetan-2-one (24) (84 mg, 64%)as an oil, vmax 2925, 2850,

1830,.1640, 1460, 1275, 1130, 1110, 990, 940, and 875 cm-1; 6 6.04-5.66

(1H, m), 5.54-5.30 (2H, m), 4.58 (1H, t, J 7 Hz), 4.36 (1H, t, J 7Hz), 1.90-1.10 02 requires .GBH, m), and 0.9 (3H,t, J 7Hz) (Found: C, 71.15; H, 9.8. C10H16 C, 71.4; H, 9.6%).

Preparation of trans-3-Pentyl-4-vinyloxetan-2-one (23)

E-Tricarbonyl-(1,2,3,-n-l-nonen-3-y1)-4-oxycarbonyl iron (21) (120 mg,

0.39 mmol) was oxidised with ceric ammonium nitrate (1.28 g, 2.33 mmol) in

acetonitrile (15 ml) at -30°C and worked up in the usual way to give trans-

3-pentyl-4-vinyloxetan-2-one (23) (44.3 mg, 68%) as an oil, vmax 2925, 2850,

1830, 1640, 1460, 1380, 1130, 990, 935, and 875 cm 1; 6 6.13-5.78 (1H, m)

5.46-4.24 (2H, m), 4.47-4.31(1H, m), 3.88 (1H, dd, J 4 Hz and 8Hz),

1.98-1.1 (8H, m), and 0.9 (3H, t, 7Hz). 142.

Preparation of 2,3-Dimethyl-1, 2-epoxy-3-butene (26)

2,3-Dimethy1-1,3-butadiene (1.4 g, 17 mmol) was treated with m- chloroperbenzoic acid (3.25 g, 17 mmol, 85%) in dichloromethane (200 ml) in the usual way to give 2,3-dimethyl-1,2-epoxy-3-butene (26) (1.0 g, 59.7%)

S 5.10 (1H, s), 5.00-4.91 (1H, m), 2.72 (2H, s), 1.74 (3H, s), and 1.45

(3H, s).

Preparation of Tricarbonyl{ 2,3,4-n-2,3-dimethyl-3-buten-2-yl)-1-oxycarbony1 11 Iron (27)

A stirred solution of 2,3-dimethyl-1,2-epoxy-3-butene (26) (0.54 g,

5.7 mmol) and iron pentacarbonyl (8.0 g, 48.8 mmol) in benzene (500 ml) was irradiated and worked up in the usual way to give the tricarbonyl{ 2,3,4-n-

2,3-dimethyl-3-buten-2-yl)-1-oxycarbonyl iron (27) (0.96 g, 65%) as white

° (decomp.) (lit.,11 m.p. 89-90° 3000, 2130, 2050, crystals, m.p. 104 ); vmax 1680, 1480, 1400, 1080, and 1010 cm-1; S 3.97 (2H, d, J 2Hz), 3.44 (1H, d, J 2Hz), 2.64 (1H, d, J 2Hz), 2.14 (3H, s), and 2.00 (3H, s); mie 238,

222, 210, 194, 182, 166, 154, 138, 98, and 82. (Found: C, 45.05; H, 3.6.

Fe requires C, 45.1; H, 3.75%). - C10H1005

Preparation of 3-(isormropeny1)-3-methyloaetane-2-one (28) and 3,6-Dihydro-

4,5-dimethyl-2-pyrone (29)

Tricarbonyl{ 2,3,4,-n-2,3-dimethyl-3-buten-2-yl)-1-oxycarbonyl iron

(27) (0.5 g, 1.88 mmol) was oxidised with eerie ammonium nitrate (6.18 g,

11.3 mmol) in acetonitrile (50 ml) for 4 h and worked up in the usual way.

Chromatography of the crude product on Florisil (eluted with petroleum ether/ ether 6:1, then ether) gave 3-(isopropenyl)-3-methyloxetan-2-one (28) 7 143.

(98 mg, 41.5%) as an oil, v 2940, 2900, 1825, 1640, 1450, 1375, 1150, 1085, max 930, 890, and 880 cm -1 ; S 4.96 (2H, d, J 2Hz), 4.17 (2H, ABq, J 12 Hz and

5Hz), 1.83 (3H, br s), and 1.6 (3H, s) (Found: (M+ + 1) 127.0761. C 7H1002 13,14 + 1 requires 127.0759), and 3,6-dihydro-4,5-dimethyl-2-pyrone (28)

(38.4 mg, 16.2%) as an oil, (lit.,14 m.p. 38°), vmax 2875, 1740, 1620, 1460,

1440, 1400, 1380, 1260, 1210, 1110, and 1060 cm 1, S 4.72 (1H, s), 2.96

(1H, s), and 1.7 (6H, s); m/e 126, 98, and 82.

•

Alternative Preparation of 3,6-Dihydro-4,5-dimethyl-2-pyrone (29)

Ceric ammonium nitrate (0.5 g, 0.9 mmol) was added to a stirred

mixture of ethanol (7 ml) and water (7 ml). Immediately the solution

became homogeneous tricarbonyl-(2,3,4-r1-2,3-dimethyl-3-butene-2-yl)-1-oxy-

carbonyl iron (27) (50 mg, 0.19 mmol) was added in one portion. After all

effervescence had ceased a further portion of eerie ammonium nitrate (0.5 g,

0.9 mmol) was added and the reaction mixture stirred overnight at 60°.

After cooling to room temperature, the mixture was added to water (100 ml)

and the resulting solution extracted with ether (3 x 50 ml). The combined

organic extracts were dried, and the solvent removed under reduced pressure

to give 3,6-dihydro-4,5-dimethyl-2-pyrone (29) as an oil (9 mg, 38%), whose

spectral properties (i.r. and n.m.r.) were identical to those of the lactone

prepared previously.

Preparation of 2-(Isopropeny1)-2-methy1-1,3-Propandiol (30)

3-(Isopropenyl)-3-methyloxetan-2-one (28) (130 mg, 1.03 mmol) was

reduced with lithium aluminium hydride (120 mg, 3.15 mmol) in tetrahydro-

furan and worked up in the usual way to give 2-(isopropenyl)-2-methyl-1,3-

propandiol (30) (117 mg, 87%) as white crystals, m.p. 56.5-57,5°, 144.

3350, 2950, 1640, 1460, 1030, and 895 cm-1, 6 5.0-4.93 (1H, m), 4.85 vmax (1H, s), 3.75 (4H, d, J 2Hz), 2.85 (2H, 8), 1.79 (3H, br s), and 1.00 (3H, s) (Found: C, 64.75; H, 10.8. C7H1402 requires C, 64.6; H, 10.85%).

Preparation of the (Tetrahydro-2-pyranyloxy)propyne (31) 45

Concentrated hydrochloric acid (3 drops) was added to a stirred mixture of dihydropyran (83 g, 0.937 mol) and propargyl alcohol (55.3 g,

0.937 mol). To the cool mixture sodium carbonate (1.0 g) was added and the resultant slurry filtered. Distillation of the liquid gave the (tetrahydro-

2-pyranyloxy)propyne (31) (106.9 g, 76%) as an oil, b.p. 92-102° at 50 mmHg;

6 (CC14) 4.67 (1H, s), 4.10 (2H, d, J 2Hz), 3.94-3.16 (2H, m), 2.23 (1H, t,

J 2Hz), and 1.92-1.26 (6H, methylene envelope).

`Preparation of 1-(Tetrahydro-2 pyranyloxy)propynyl)cyclohexanol (32)16

The (tetrahydro-2-pyranyloxy)propyne (31) (109 g, 0.756 mol) in tetrahydrofuran (50 ml) was added to a solution of ethyl magnesium bromide From ethyl bromide (94 g, 0.869 mol) and magnesium (20.6 g, 0.869 mol)J in tetrahydrofuran (200 ml) at 20°C. After lh cyclohexanone (69.45 g, 0.71 mol) was added dropwise over a 30 min period. 3 h After the addition was complete, the solution was added to a mixture of ice and sulphuric acid

(0.326 mol). The organic layer was separated and the aqueous layer extracted with petroleum ether (3 x 100 ml). The combined extracts were washed with saturated sodium carbonate solution (100 ml),saturated ammonium carbonate solution (100 ml), and were dried (Na2CO3•and Na2SO4). Removal of the solvent under.reduced pressure gave 1-((tetrahydro-2-pyranyloxy)propynyl)cycLohexanol (32) (138.9 g, 82%) as an oil, b.p.135-14 at 0.1 mmHg 145.

16 0 (lit., b.p. 124-127 at 0.3 mmHg; S 4.83 (1H, s), 4.34 (2H, s), 4.34

(2H, s), 4,14-3.27 (2H, m), 2.34 (1H, s, exch. D20), and 2.05-0.7 (1H, m).

Preparation of the 2-Propylidenecyclohexane (33)16

1-(Tetrahydro-2-pyranyloxy)propynyl cyclohexanol (138.9 g, 0.61)

(32) in tetrahydrofuran (200 ml) was added to a stirred slurry of lithium aluminium hydride (69.5 g, 1.83 mol) in tetrahydrofuran (1.8 1). After the addition was complete the mixture was refluxed for 3h, cooled to room temperature, and cautiously added to a mixture of ice and sulphuric acid (89.2 g, 0.91 mol). The resulting slurry was extracted with petroleum ether (3 x 500 ml), the combined extracts dried, and the solvent removed by distillation at atmospheric pressure. Distillation of the residue gave the product 2-propylidenecyclohexane (33) (48.5 g, 65%) as an oil, oil, b.p. 81° at 56 mmHg, (lit.,16 b.p. 63° at 45 mmHg), S 6.66-6.20

(1H, m), 5.70 (1H, d, J 10Hz), 5.16-4.73 (2H, m), 2.53-1.36 (10H, m).

47 Preparation of the CvclohexanesDlro-2'-(3'-vinyloxirane)(34)

Peracetic acid (32.8 g, 38%, 0.164 mol) was added dropwise to a rapidly stirred slurry of 2-propylidenecyclohexane (33) (20 g, 0.164 mol), sodium carbonate (43 g, 0.400 mol) and dichloromethane (200 ml) at 25°C.

The mixture was stored in a fridge overnight. The precipitate was removed by filtration and washed with dichloromethane (100 ml). The combined extracts were evaporated under reduced pressure and distillation of the residue gave cyclohexanes.iro-2'-(3-vinyloxirane) (34) (13.8 g, 61%) as an oil, b.p. 72°C at 18 mmHg, S 6.1-5.3 (3H, m), 3.12 (1H, d, J 5Hz) and 146.

-.97-0.7 (10H, m) (Found: C, 78.25; H, 10.2. C9H140 requires C, 78.2;

H, 10.2%).

Preparation of Tricarbonyl-(1',2',3' - n - 1-propenylcyclohexan-1'-yl)-1- oxycarbonyl Iron (35)

A solution of iron pentacarbonyl (8 g, 40.8 mmol) and cyclohexanespiro-2'-(3-vinyloxirane) (34) (0.8 g, 5.8 mmol) in benzene (500 ml) was irradiated for 7 h and worked up, in the usual way to give tricarbonyl-(1',

2',3'-n-l-propenylcyclohexan-l'-yl)-1-oxycarbonyl iron (35) (0.86 g, 48.6%) as white crystals, m.p. 96° (decomp.); 'umax 2925, 2870, 2050, 1990, 1670, -1 1450, 1440, 1165, 985, 960, and 875 cm ; S 5.14-4.58 (1H, m), 3.90

(1H, d, J 8Hz), 3.51 (1H, d, J 9Hz) 3.0 (1H, d, J 13 Hz) and 2.13-1.18

(10H, m); m/e 234, 222, 206, 194, 178, 138, and 122 (Found: C, 50.8;

H, 4.6. Fe05 requires C, 51.0; H, 4.6%). C13H14

Preparation of Cyclohexanespiro-4'-(3'-vinyloxetan-3'-one) (37)

Tricarbonyl-(1',2',3'-n-l-propenylcyclohexane-l'-yl)-1-oxycarbonyl

iron (35) (430 mg, 1.4 mmol) was oxidised with ceric ammonium nitrate (4.58 g,

8.35 mmol) in acetonitrile (35 ml) and worked up in the usual way. Chroma-

tography of the crude product on Florisil (eluted with ether/petroleum ether,

1:6), gave cyclohexanespiro-4'-(3'-vinyloxetan-3'-one) (37) (121 mg, 52%)

as an oil, vmax 2940, 2850, 1830, 1640, 1450, 1280, 1195, 1020, 885, 820, -1 and 810 cm ; S 5.68-4.68 (3H, m), 3.26 (1H, d, J 6Hz) and 1.7-0.64 (10H,

m). 147.

Preparation of 3-Methylhydroxy-l-vinylcyclohexanol (38)

Cyclohexanespiro-4'-(3-vinyloxetan-3-one) (37) (121 mg, 0.73 mmol) was reduced with lithium aluminium hydride (200 mg, 5.2 mmol) in the usual way to give 2-methylhydroxy-l-vinylcyclohexanol (38) (90 mg, 72%) as an oil,

3370, 2940, 2860, 1640, 1450, 1420, 1160, 1080, 1050, 1015, 1000, 970, vmax and 920 cm 1; S 6.18-5.50 (1H, m), 5.30-4.90 (2H, m), 3.80 (2H, d, J 6Hz) and 2.83-2.11 (3H, m), and 1.73-1.26 (10H, m) (Found: C, 70.4; H, 10.9.

02 requires C, 70.55; H, 10.65%). C10H18

Preparation of 1,3-Cyclohexadiene (39)17

Sodium hydride (24.6 g, 50%, 0.51 mol) was added in small portions to a mixture of triglyme (250 ml) and isopropyl alcohol (180 ml), under a nitrogen atmosphere. After the addition was complete the mixture was stirred for 10 min and the excess isopropyl alcohol removed under reduced pressure and a stream of nitrogen. The mixture was warmed to 110°C and the pressure further reduced to 170-130 mmHg. 1,2-Dibromocyclohexane (121 g, 0.51 mol)46 was added dropwise to the reaction mixture, at such a rate as to maintain the temperature of the reaction mixture when the heat was removed. The product was distilled off as it was formed, and collected in a flask cooled in an acetone bath at -78°. The crude product was washed with water. (3 x

50 ml) and dried to give 1,3-cyclohexadiene (39) (32.1 g, 76%) as an oil, b.p. 82° (lit.,17 78-80°); d 5.77 (4H, s) and 2.14 (4H, s).

•

47 Preparation of 1,2-epoxy-3-cyclohexene (40)

1,3-Cyclohexadiene (39) (15 g, 0.18 mol) was oxidised with pera- 148.

cetic acid (38.8 g, 38%, 0.19 mol) in dichloromethane (200 ml) in the presence of sodium carbonate (50 g). The solvent was removed from the combined extracts by distillation. When the temperature of the residue had risen to 65°C, the epoxide was distilled from the reaction mixture under reduced pressure, and collected in a flask cooled to -78°. Redis- tillation gave 1,2-epoxy-3-cyclohexene (40) (10.3 g, 58.6%) as an oil, b.p. 77-78° at 93 mm of Hg; vmax 2900 and 1640 cm 1, d 4.93 (2H, d, J

4Hz), 3.53-3.43 (1H, m); 3.28-3.11 (1H, m) and 2.41-1.43 (4H, m).

Preparation of Tricarbonyl-(2,3,4-11-3-cyclohexen-2- 1)-1-ox carbonyl Iron

(41)

1,2-Epoxycyclohex-3-ene (40) (1 g, 10.4 mmol) and iron pentacarbonyl (10 g, 51.0 mmol) in benzene (400 ml) were irradiated until the product concentration was maximised (by i.r.). The benzene was removed under reduced pressure, and the residue purified by chromatography on silica

(eluted with ether), to give the tricarbonyl-(2,3,4-11-3-cyclohexen-2-yl)- l-oxycarbonyl iron (41) (0.86 g, 32%) (lit.,160%)) as yellow crystals, m.p.

° (decomp.) (lit.l 80 2900, 2100, 1990, 1650, 1460, 1375, 1340, 1325, 74 °), vmax 1060, 1010, 995, 970, and 665 cm-1; d 5.84-5.68 (1H, m), 5.44 (1H, t ,

J 6Hz), 4.70 (1H, t, 6 Hz), 4.57-4.44 (1H, m), 2.64-1.92 (2H, m), 1.78-

1.50 (1H, m), and 1.46-1.08 (1H, m); m/e 208, 180, and 152 (Found: C,

45.35; H, 3.15. C10H$5 Fe requires C, 45.5; H, 3.05%).

Preparation of 7-Oxabicyclol4.2.0Joct-2-en-8-one (44)

A. Ceric ammonium nitrate (3.61 g, 6.6 mmol) was added to tricarbonyl

(2,3,4,-11-3-cyclohexen - 2-yl)-1-oxycarbonyl iron (41) (290 mg, 1.09 mmol) in ethanol (15 ml) at -5°C. After all of the iron complex had been 149. oxidised (as monitored by t.l.c,) the mixture was added to ether (100 ml) and water (100 ml). The aqueous layer was further extracted with ether

(4 x 25 ml) and the combined organic phases dried. Removal of the solvent under reduced pressure and chromatography of the residue on silica gel

(eluted with ether/petroleum, ether, 1:1) gave 7-oxabicyclo [4.2.O]oct-2- en-8-one (44) (116.8 mg, 86%); "max 2900, 1815, 1440, 1420, 1360, 1310, -1 1280, 1225, 1125, 1065, 1025, 970, 880, 850, 795, 745, 690, and 675 cm ;

6 6.30-6.00 (1H, m) 5.85-5.53 (1H, m), 4.91-4.73 (1H, m); 4.10 (1H, t, J

6Hz), and 2.45-0.98 (4H, m). •

B. Tricarbonyl-(2,3,4-p-3-cyclohexen - 2-yl)-l-oxycarbonyl iron (41)

(0.5 g, 1.9 mmol) was added in one portion to a stirred solution of ceric ammonium nitrate (5.0 g, 9.1 mmol) in water (25 ml) and ethanol (25 ml) at room temperature. After all the initial effervescence has ceased the reaction mixture was poured into water (200 ml), and the resulting aqueous solution extracted with dichloromethane (3 x 100 ml). The combined organic extracts were dried, and the solvent was removed under reduced pressure to give 7-oxabicyclo[4.2.01oct-2-en-8-one (44) (162 mg, 69%), identical by i.r., t.l.c. and 1H n.m.r. to the previously prepared sample.

Preparation of 2-Methylhydroxycyclohex-3-enol ditosylate

7-0xa,bicyclo{4.2.0J oct-2-en-8-one (44) (89 mg, 0.71 mmol) was reduced with lithium aluminium hydride (1 g, 27 mmol) in tetrahydrofuran

(10 ml) in the usual way. The crude product from the reaction was dissolved in pyridine (10 ml) and toluene-4-sulphonyl chloride (2g) was 150.

added to the mixture at 0°C. After 12 h at 0°C the mixture was added to

a solution of dilute hydrochloric acid (400 ml) and extracted with methylene chloride (2 x 100 ml). The combined extracts were washed with water (2 x 100 ml), dried and the solvent removed to give a mixture

of the crude product and tosyl chloride as an oil. Chromatography of the

mixture on silica gel (eluted with methylene chloride) gave the 2-methyl-

hydroxycyclohex-3-enol ditosylate (46) (41 mg, 13.3%) as an oil, S 7.74

(4H, d, J 8Hz), 7.30 (4H, d, J 8Hz), 5.90-5.55 (1H, m), 5.43-5.13 (1H,

m), 5.00-4.7 (1H, m), 3.9 (2H, d, J 7Hz), 2.41 (6H, s), and 2.12-1.00

S2 requires C, 57.75; (4H, m); (Found: C, 57.7; H, 5.41; S, 15.05. C21H2406 H, 5.55; 8, 14.7%).

Preparation of cis-1,2-Dihydroxymethylcyclohexane (47)

cis-1,2-Cyclohexanedicarboxylic anhydride (15.4 g, 0.1 mol) in

tetrahydrofuran (50 ml) was added dropwise to a stirred slurry of lithium

aluminium hydride (8.5 g, 0.22 mol) in tetrahydrofuran (150 ml) so that the

tetrahydrofuran boiled gently. When the addition was complete the mixture

was stirred for a further 2 h, and water was added to destroy excess lithium

aluminium hydride. The reaction mixture was added slowly to a stirred solution

of.sulphuric acid (300 ml, 10%), and the resulting solution extracted with

ether (3 x 150 ml). The combined organic extracts were dried and the solvent

removed under reduced pressure. Nitrogen was then blown through the hot diol

to remove traces of solvent whereupon crystallisation was induced. By this

method cis-1,2-dihydroxymethylcyclohexane (47) was obtained (13.7 g, 95%)

43-43.5°); v 3400, 2975, 1465, as white crystals, m.p. 40-41° (lit.,21 max 1115, 1050, and 980 cm 5 3.65 (6H, s), 1.84 (2H, s), and 1.43 (10H, s). 151.

20 Preparation of cis-1,2-Di (bromethyl)cyclohexane mo (48)

Phosphorus tribromide (26.8 g, 0.1 mole) was cooled to -10°C and

cis-1,2-dihydroxymethylcyclohexane (47) (14.4 g, 0.1 mol) was added in

small portions over a 3 h period. The reaction mixture was allowed to

warm to room temperature, and after a further 5 h warmed to 90-100°. After

remaining at this temperature overnight, the reaction mixture was cooled

in an ice bath, water (30 ml) added, and the aqueous solution extracted

with methylene chloride (3 x 30 ml). The combined organic extracts were

washed with saturated sodium carbonate solution (3 x 100 ml), dried,

and the solvent removed under reduced pressure to give the cis-1,2-di(bromo-

methyl)cyclohexane (48) (24 g, 89.5%) as an oil, which needed no further

purification, d 3.37 (4H, d, J 8Hz) and 2.47-1.6 (10H, m).

Preparation of 1,2TDimethylenecyclohexane (52)

cis-1,2-Di(bromomethyl)cyclohexane (32 g, 0.12 mol) (48) and metha-

nol (3 ml) were mixed with trimethylamine (44 g, 0.74 mol) at -78°C. The

reaction vessel (250 ml) was sealed and allowed to warm to room temperature.

After 2 weeks precipitation appeared to be complete and the reaction vessel

was placed under high vacitum. The white solid produced was dissolved in

the minimum amount of water and silver oxide was added until no further

precipitation of the yellow solid occurred. The solution was filtered, and

the filtrate distilled under reduced pressure (4 cm Hg). When most of the

water had been removed, and the temperature of the residue had reached 140°,

the diene began to distil , and was collected in a flask cooled in a

cardite/acetone bath. The crude product was washed with hydrochloric acid

(1 x 10 ml, 10%), water (3 x 20 ml), and was dried (Na2SO4) to give

1,2-dimethylenecyclohexane (52) (2.39 g, 18.4 %) as an oil which required 152.

no further purification; A 2900, 1630, 1440, 950, 895, 860, and 825 cm 1; max S 4.87 (1H, d, J 2Hz), 4.60 (1H, s), 2.20 (4H, br s), and 1.60 (4H, m).

Preparation of 2-Methylenecyclohexanespiro-2'-(oxirane) (53)

1,2-Dimethylenecyclohexane (17.6 g, 0.16 mol) (52) was oxidised

with peracetic acid (32 g, 38%, 0.16 mol) in the presence of sodium carbonate

(42 g) in dichloromethane (300 ml), as described for the epoxide (53). The

crude product obtained was distilled to give 2-methylenecyclohexanespiro-2'-

(oxirane) (53) (11.2 g, 55%) as an oil, b.p. 75°C at 47 mm Hg; S 4.88

(1H, s), 4.73 (1H, s), 2.67 (2H, s), and 2.48-0.70 (8H, m).

Preparation of Tricarbonyl-(2,3,4-n-2,3-(tetramethylene)-3-buten-2-y1)-1-

oxycarbonyl iron (54)

2-Methylenecyclohexanespiro-2'-oxirane (53) (0.8 g, 6.45 mmol) was

irradiated with iron pentacarbonyl (8 g, 40.8 mmol) in benzene (400 ml).

After the usual work up, tricarbonyl-(2,3,4-n-2,3-(hexamethylene)-2-buten-

2 11)-1-oxycarbonyl iron (54) (560 mg, 29%) was obtained as white crystals,

° (dec.);A 2900, 2050, 1990, 1650, 1465, 1450, 1380, 1045, m.p. 83-84 max 995, 950, and 660 cm l; S 4.03 (2H, ABq, J 20 Hz and 12 Hz), 3.20 (1H,

d, J 3Hz), and 2.97-1.5 (9H, m), m/e 264, 248, 236, 220, 208, 192, 180,

164, 152, and 124. (Found: C, 49.3; H, 4.05. C 05Fe requires C, 49.35; 12H12 H, 4.15%). 153.

Pre aratinn of 5,6,7 8-Tetmhydroisochromans-3-one (55) and 2-Meth lenec clo hexanespiro-4'-0xetan-2-one (56)

Tricarbonyl-(2,3,4,-n-2,3-(hexamethylene)-2-buten-2-yl)-1-oxycarbonyl- iron (54) (207 mg, 0.71 mmol) was oxidised by ceric ammonium nitrate (20.6 g,

37.6 mmol) in acetonitrile (15 ml) for 4 h and the reaction mixture worked up in the usual way. Chromatography of the crude product on Florisil (petroleum ether/ether, 6:1, then ether) gave 2-methylenecyclohexanespiro-4'-oxetan-2- one (560 (16.1 mg, 15%) as an oil, v 2925, 2850, 7830, 1640, 1450, 1285, max 1140, 1110, 1095, 1080, 1060, 980, 920, and 900 cm-1; 6 4.90 (2H, s), 4.18

(2H, ABq; J 12 Hz and 5 Hz), and 2.35-0.74 (8H, m); 6 C6D6 4.85 (1H, s),

4.60 (1H, s), 3.57 (2H, ABq, J 17 Hz and 5 Hz), and 2.30-0.74 (8H, m)

(Found: (M+ + 1) 153.0918. C9H1202 + 1 requires 153.0916), and 5,6,7,8- tetrahydroisochroman-3-one (55) (59.2 mg, 55%) as an oil, vmax 2900, 1740, 1635, 1225, and 1045 cm l; 6 4.63 (2H, br s), 2.88 (2H, br s), and

2.05-1.45 (8H, m).

Alternative Preparation of 5,6,7,8-Tetrahydroisochroman-3-one (55)

Tricarbonyl-(2,3,4-n-2,3-tetramethylene-3-buten — 2-yl)-1-oxycarbonyl iron (54) (300 mg, 1.0 mmol) was added in one portion to a stirred solution of ceric ammonium nitrate (5 g, 9.1 mmol) in ethanol (16 ml) and water

(13 ml). After all effervescence had ceased, the reaction mixture was added to water (100 ml) and the solution extracted with ether (6 x 50 ml). The combined extracts were dried and the solvent removed under reduced pressure to give 5,6,7,8-tetrahydroisochroman-3-one (55) (82 mg, 52.5%) as an oil, whose i.r., t.l.c., and1HU n.m.r. were identical to those of previous samples. 154.

Preparation of 1,1-Dimethlhydroxy-2-methxlenecyclohexane (60)20,21,50

2-Methylenecyclohexanespiro-4'-oxetan-2-one (56) (16.1 mg, 0.10 mmol) was reduced with lithium aluminium hydride (100 mg, 2.6 mmol) in tetrahydro-

furan (20 ml) and the reaction mixture worked up as before to give 1,1-dime--

thylhydroxy-2-methylenecyclohexane (60) (7.7 mg, 46.6%) as white crystals, m.p.

68.5°C (from petroleum ether); vmax 3320, 3080, 2930, 2855, 1635, 1475, 1440, 1372, 1100, 1047, 1035, 1020, 1000, and 892 cm 1; S 4.90 (1H, s), 4.75 (1H,

s), 3.70 (4H, d, J 4Hz), 2.50-2.00 (2H, m), and 1.54 (6H, s) (Found: C, 69.15;

H, 10.35. C9H1602 requires C, 69.2; H, 10.3%).

Preparation of 2-(2-Hydroxymethyl-l-cyclohexenyl)ethanol (59)

5,6,7,8-Tetrahydroisochroman-3-one (55) (54.9 mg, 0.36 mol) was reduced with lithium aluminium hydride (150 mg, 3.9 mmol), in tetrahydrofuran

(6 ml), and the reaction mixture worked up as before. Chromatography of the crude product on Florisil (eluted with ether) gave the 2-(2-hydroxymethyl-

-1-cyclohexenyl)ethanol (59) (44.6 mg, 79%) as an oil, v 3340, 2930, max 2880, 2830, 1665, 1435, 1045, 1015, and 100 cm 1; S 4.05 (2H, s), 3.66

(2H, t, J 6 Hz), 3.90-3.70 (2H, m, exch. D20), and 2.5-1.35 (10H, m);

(Found: M+ 156.1151. C9H1602 requires 156.1150)(Found: C, 69.1; H, 10.25.

C9H1602 requires C, 69.2; H, 10.3%).

Pre aration of Tricarbon 1-(2 3 4-r)-2 3-dimeth 1-3-buten-2- 1)-1-bent lamino- carbonyl iron (77)

Tricarbonyl-(2,3,4-n-2,3-dimethyl-3-buten-2-yl)-1-oxycarbonyl iron

(27) (100 mg, 0.37 mmol) was treated with zinc chloride (100 mg, 0.735 mmol) 155.

and benzylamine in ether (6 ml) for 90 min and worked up in the usual way

to give tricarbony1-1,3,4-n-2,3-dimethyl-3-buten-2-yl)-1-benzylaminocarbony1

° (decomp.); v 2900, 2025, 1990, 1570, 1440, 1185, iron (77), m.p. 98-100 max and 705 cm-1; 6 7.28 (5H, s), 4.28 (2H, s), 3.42 (1H, d, J 2Iz), 3.14,

(2H, s), 2.44 (1H, d, J 2Hz), 2.08 (3H, s), and 1.09 (3H,-s) (Found: C, 57.55;

H, 4.8; N, 3.9. N04Fe requires C, 57.5; H, 4.8; N, 3.95%). C17H17

Preparation of 1-Benzyl-3-(isopropenyl)-3-methylazetidinone (79) and 1-Benzyl-

3,6-dihydro-4,5-dimethyl-2-pyridone (78)

Oxidation of tricarbonyl-(2,3,4-fl-2,3-dimethyl-3-butene-2-yl)-1-

benzylamino carbonyl iron (77) (0.32 g, 0.9 mmol) with eerie ammonium nitrate

(2.28 g, 4.00 mmol) in ethanol (40 ml) in the usual way gave 1-benzyl-3-

(isopropenyl)-3-methylazetidinone (79) and 1-benzyl-3,6-dibydro-4,5-dimethyl-2-

pYridone, (78) as an oil. Chromatography of the crude product on silica gel

(eluted with ether) gave 1-benzyl-3-(isopropenyl)-3-methylazetidinone (79)

2860, 1740, 1640, 1400, and 700 cm-1, 6 7.34 (65 mg, 33.5%) as an oil; vmax (5H, s), 5.02 (1H, s), 4.9 (1H, s), 4.4 (2H, s), 3.10 (2H, ABq, J 16 Hz and

NO requires 6 Hz), 1.84 (3H, s), and 1.50 (3H, s) (Found: M 215.1310. C14H17 215.1309) and 1-benzyl-3,6-dih_ydro-425-dimethyl-2 pyridone (78) (109 mg,

56.3%) as white crystals, m.p. 78-80° (after recrystallisation from petroleum

ether); Vmax 2900, 1640, 1495, 1450, 1320, 1265, 725, and 715 cm 1;

6 7.40 (5H, s), 4.73 (2H, s), 3.73 (2H, br s), 3.03 (2H, br s), and 1.70

N0 requires C, 78.15; (6H, s) (Found: C, 78.0; H, 8.15; N, 6.35. C14H17 H, 7.9; N, 6.5%). 156.

Preparation of E-Tricarbonyl-(2, 3,4-n-3-nonen-2-y1)-1-benzylaminocarbonyl iron (83)

E and Z-Tricarbonyl-(1,2,3-11-1-nonen-3-y1)oxycarbonyl iron (21 )

and (22) (0.46 g, 1.48 mmol) were treated with zinc chloride (200 mg, 1.47

mmol) and benzylamine (0.75 g, 0.7 mmol) in ether (10 ml) in the usual way.

Chromatography of the crude product on silica gel (eluted with petroleum ether/

ether, 2:1) gave E-tricarbonyl-( 2,3,4-n-3-nonen-2-y1)-1-benzylamino iron (83)

as an oil (0.5 g, 82%), which solidified on standing to give crystals, m.p.

80-86°; 2900, 2850, 1995, 1980, and 1590 cm-1, S 7.25 (3H, m), Vmax 7.13 (2H, d, J 6Hz), 4.64 (1H, q, J 12 Hz and 8 Hz), 4.31 (1H, d, J 14 Hz), 4.17 (1H, dd, J 14 Hz and 2 Hz), 4.15 (1H, d, J 14 Hz), 3.78 (1H, m), 3.26

(1H, dd, J 12 Hz and 6 Hz), 3.09 (1H, d, J 12 Hz), 2.27 (1H, m), 1.72 (1H, m), 1.64-1.41 (2H, m), 1.35 (4H, m), and 0.914 (3H, t, J 6Hz) (Found: C, 60.5;

H, 5.8; N, 3.5. CO N04Fe requires C, 60.6; H, 5.9; N, 3.5%). 23

Preparation of cis- and trans-1-Benzyl-3-(1'-heptenyl)azetidinone (84)

E-Tricarbonyi-(2,3,4-n-3-nonen-2-y1)-1-benzylaminocarbonyl iron

(83) (0.63 g, 1.58 mmol) was oxidised with ceric ammonium nitrate (4.1 g,

7.5 mmol) in ethanol (10 ml) at -5°C and worked up in the usual way. Chroma- tography of the crude product on silica gel (eluted with petroleum ether/

ether 3:1) gave the cis- and trans-l-benzyl-3-(1-heptenyl)azetidinone (84) as 2900, 1745, 1390 cm 1;87.3(5H, m), 6.0-5.0 an oil (0.26 g, 63.7%), vmax (2H, m), 4.37 (2H, quin), 3.9-3.4 OH, m), 3.32 (1H, t, J 5 Hz), 2.92

(1H, q, 5 Hz, and 3Hz) and 2.25-0.6 (17H, m) (Found: M+ 257.1772. C17H25N0 requires 257.1779). 157.

Preparation of Tricarbonyl-(2',1,2-n-1-ethyl-l-cyclopenten-1'-yl)-2'-benzyl-

aminocarbonyl iron (80)

Syn- and anti-tricarbonyl-(1,1',2'-n-l-vinytcyclopentan-l-y1)-3-

oxycarbonyl iron(3)&(14)(1.0 g, 3.60 mmol)were treated with zinc chloride

(1.0 g, 7.35 mmol) and benzylamine (5.0 g, 46.7 mmol) in ether (70 ml) for

3 h, in the usual way. Chromatography on silica gel (eluted with ether/

petroleum ether 1:1) afforded tricarbonyl-(2',1,2-n-l-ethyl-1-cyclopenten-

l-yl)-2'-benzylaminocarbonyl iron (80) (0.5 g, 38%)(48% based on recovered star

° (decomp.); v 2900, 2050, ting material) as white crystals, m.p. 98-100 max 3.27(2H, d, T 5 Hz)) 1580, 1180, and 680 cm— 1; S 7.39-7.09 (5H, m), 4.29 (2H, d, J 4Hz),I and 3.12-

1.06 (7H, m)(Found: C, 58.6; H, 4.55; N, 3.75. C18H17N04Fe requires C, 58.9;

H, 4.55; N, 3.8%).

Preparation of 1-Benzyl-3-(l'-Cyclopentenyl)azetidinone (81)

Tricarbonyl-(2',1,2-n-l-ethyl-l-cyclopenten-l'-y1)-2'-benzylamino-

carbonyl iron (80) (190 mg, 0.517mmol) in ethanol (5 ml) was oxidised with

eerie ammonium nitrate (1.35 g, 2.38 mmol) in ethanol (18 ml) and worked up

in the usual way, to give 1-benzyl-3-(1'-cyclopentenyl)azetidinone (81)

(99 mg, 84%) as an oil, vmax 2860, and 1740 cm-1; S 7.25 (5H, s), 5.67 (1H, s),

4.45 (2H, s), 4.00 (1H, m), 3.34 (1H, t, J 6Hz), 3.09 (1H, dd, J 6Hz and

NO 3 Hz), and 2.63-1.47 (6H, m);' (Found: C, 79.0; H, 7.7; N, 6.2. C15H17 requires C, 79.25; H, 7.55; N, 6.15%).

Preparation of 3,6-Dihydro-4,5-dimethyl-2-pyridone (85)

Tricarbonyl--(2,3,4-n-2,3-dimethyl-3-buten-2-yl)-1-oxycarbonyl iron

(27) (200 mg, 0.64 mmol), zinc chloride (450 mg, 3.31 mmol), ammonia (2 g, 158.

117.5 mmol) and ether were stirred together in a pressure bottle (200 ml

volume) at room temperature for 2 days. The ammonia pressure was released,

the mixture filtered, and the precipitate washed with ether (160 ml). Removal

of the solvent under reduced pressure and washing of the white solid

with a little cold ether gave 3,6-dihydro-4,5-dimethyl-2-pyridone (85) (38 mg,

3150, 2900, 1650, and 1450 cm-1; 41%) as a white solid, m.p. 128-133°, vmax S 4.03 (2H, s), 3.18 (1H, s), 3.05 (2H, s), and 1.78 (6H, s) (Found: C, 58.83;

H, '8.97; N, 9.51. C7H11N0. H2O requires C, 58.72; H, 9.15; N, 9.78%).

Preparation of 1-Vinyl-3-oxabicyclo [2.2.11hept-2-one (89)

A stirred mixture of 1-vinyl-6-oxabicyclo[3.2.0)hept-7-one (6)

(84 mg, 0.6 mmol) and zinc chloride (3 g, 22 mmol) in benzene (40 ml) was

refluxed overnight. After this time the mixture was allowed to cool to room

temperature, and water (40 ml) was added. The organic phase was separated,

dried; and the solvent removed under reduced pressure to give the crude

product as an oil. Chromatography of this oil on Florisil (eluted with

petroleum ether/ether, 3:1), gave starting material (10 mg, 11.9%) and

1-vinyl-3-oxabicyclo [2.2.1] hept-2-one (89) (29.4 mg, 35%) as an oil;

2950, 1775, 1642, 1345, 1175, 1140, 1115, 1080, 1060, 1045, 990, 942, vmax 992, 882, and 662 cm-1; S 6.22 (1H, q, J 10 Hz and 18 Hz), 5.48-5.12 (2H,

69.45; H, 7.55 0 m), 4.9 (1H, s), and 2.35-0.65 (6H,m) (Found: C, .C8H10 2 requires C, 69.55; H,. 7.3%).

Preparation of 3-Methyl-l-oxabicyclo [3.3.01oct-3-en-2-one (91)

3-(1-Pentenyl)-oxetan-2-one (90) (95.9 mg, 7 mmol) and zinc chloride

(2.25 g, 16.5 mmol) were refluxed in benzene (50 ml) for 6 h. The solvent

was removed under reduced pressure and the residue partitioned between ether • 159.

(20 ml) and water (20 ml). The organic layer was separated, dried, and the solvent removed under reduced pressure. Chromatography of the residue

on Florisil (eluted with petroleum ether/ether, 4:1) gave 3-methyl-l-oxabicyclo [3.3.0, oct-3-en-2-one (91) (25.-4 mg, 26.5%), vmax 2960, 2930, 2870,

2860, 1765, 1640, 1450, 1350, 1170, 1100, and 980 cm-1; 6 4.50 (1H, m),

2.16 (3H, s), and 2.13-1.16 (6H, m).

Preparation of 1-Formyl-2-methyl-l-cyclohexene (94)

Tricarbonyl-(2,3,4-n-2,3-tetramethylene-3-buten-2-yl)-1-oxycarbonyl iron (54) (200 mg, 0.68 mmol) in ether (20 ml) was heated until all the starting material had disappeared (as shown by t.l.c.). The reaction mixture was cooled to room temperature, and acetone (20 ml) added. Ferric chloride (900 mg, 5.5 mmol) was added in one portion and the mixture stirred overnight. The reaction mixture was added to water (100 ml), the organic layer . separated and the aqueous solution extracted with ether (3 x 50 ml).

The combined organic extracts were dried, and the solvent removed under reduced pressure to give the 1-formyl-2-methyl-l-cyclohexene (94) (50 mg,

60%) as an oil, 2850, 1660, 1635, 1450, 1380, 1360, 1275, 1240, 1140, vmax 1060, and 740 cm-1; 6 10.08 (1H, s), 2.13 (7H, s), and 1.83-1.45 (4H, m).

48,49 Preparation of 2-Carbethoxycyclopentanone (95)

A. Cyclopentanone (4.2 g, 0.05 mol) in benzene (10 ml) was added to a boiling mixture of diethyl carbonate (11.8 g, 0.1 mol) and sodium hydride

(3.4 g, 0,14 mol) in benzene (90 ml) over a 3 h period. The reaction mixture was cooled, and glacial acetic acid (10 ml) added dropwise. The pasty white solid which separated was dissolved in ice cold water (40 ml). The aqueous layer was separated and extracted with benzene (3 x 50 ml). 160.

The combined organic extracts were washed with water (2 x 50 ml), dried, and the solvent removed under reduced pressure to give 2-carbethoxycyclopentanone

(95) (3.0 g, 37%) as an oil, d 4.25 (2H, q, J 14Hz and 8Hz), 3.13 (1H, t,

J 8Hz), 2.6-1.64 (6H, m), and 1.35.(3H, t, J 6Hz).

B. Diethyl adipate (30.3 g, 0.15 mol) was added to a refluxing mixture of sodium (4.98 g, 0.21 mol) in benzene (150 ml). Ethanol (1 ml) was added to initiate the reaction. After refluxing overnight, the reaction mixture was cooled and ice (60 g) was added under a nitrogen atmosphere. The white solid produced by the reaction slowly dissolved. Concentrated hydrochloric acid was added dropwise to the stirred solution until the mixture was acidic.

The aqueous layer was separated and extracted with benzene (3 x 100 ml), the combined organic extracts washed with water (3 x 50 ml), dried and the solvent removed under reduced pressure to give 2-carbethoxycyclopentanone (95) (14.1 g, 60%) as an oil.

Preparation of Dimethyl(methylene)ammonium Chloride (Eschenmoser's Salt)

(101)

Bis(Dimethylamino)methane (10.2 g, 0.1 mol) in ether (50 ml) was added to a stirred solution of acetyl chloride (7.85 g, 0.1 mol) in ether

(50 ml) under an atmosphere of dry nitrogen. The mixture was filtered under nitrogen and the white precipitate washed with cold ether (25 ml), and dried in vacuo to give dimethyl(methylene)ammonium chloride (101) (8.41 g, 92%).

Preparation of 2-Carbethoxy-2-methylcyclo2entanone (96k

Potassium 0).75 g, 19 mmol) was dissolved in t-butanol (25 ml), (2.69 g, 19 mmol) and 2-carbethoxycyclopentanone (95)/added in one portion to the boiling 161.

solution. Methyl iodide (2.98 g, 20 mmol) in t-butanol (10 ml) was added dropwise over a 30 min period. After the addition was complete, the mixture was allowed to cool to room temperature, and water was added dropwise to the cooled reaction mixture until no more precipitation occurred. Filtration followed by the evaporation of the filtrate under reduced pressure gave the crude product as a cloudy oil. This was redissolved in ether and dried over sodium sulphate. Removal of the ether under reduced pressure gave 2-carbethoxy-2-methylcyclopentanone (96) (2.6 g, 81%) as an oil, S 4.20 (2H, q,

J 14Hz and 8Hz), 2.77-1.59 (6H, m), and 1.43-1.1 (6H, m).

Preparation of 2-Carbethoxy-2-methyl-5-(dimethylamino)methylcyclopentanone

(97)

2-Carbethoxy-2-methylcyclopentanone (96) (3.0 g, 19.2 mmol) in tetrahydrofuran (10 ml) was added dropwise to a stirred slurry of potassium hydride (1.04 g, 26.6 mmol) in tetrahydrofuran (20 ml) under an argon atmosphere, at -10°C. When effervescence had ceased, the reaction mixture was transferred via syringe to a cooled dropping funnel maintained at -78°C.

The solution was then added dropwise to a stirred slurry of dimethyl(methylene) xr(1.87 g, 20 mmol) ammonium chloride(in tetrahydrofuran (10 ml) at -78°C. After the addition was complete, the mixture was allowed to warm to room temperature over the course of 1 h, and was stirred for a further lh. Water was added dropwise

'Co the reaction mixture until the salts in suspension precipitated. Filtra- tion of the mixture, and removal of solvent from the filtrate under reduced pressure gave 2-carbethoxy-2-methyl-5-(dimethylamino)methylcyclopentanone

(97) (3.4 g, 85%) as an oil, v 2900, 1710, 1460, 1370, and 1040cm 1; S 4.2 max (2H, q, J 15Hz and 7Hz), 2.0-1.55 (13H, m), 1.50-1,10 (6H, m). 162.

REFERENCES

1. R. Aumann, K. Frohlich, and H. Ring, Angew. Chem. Internat. Edn., 1974,13, 275.

2. K-N. Chen, R.M. Moriarty, B.G. DeBoer, M,R. Churchill, and H.J.C. Yeh, J. Amer. Chem. Soc., 1975, 97, 5602.

3. R.P. Thummel and W. Nutakul, J. Org. Chem., 1977, 42, 3

4. L.A. Paquette, S.V. Ley, S.G. Traynor, J.T. Martin, and J.M. Geckle, J. Amer. Chem. Soc., 1976, 98, 8162.

5. A.F. McKay, D.L. Garmaise, G.Y. Paris, S. Geblum, and R.J. Ranz, Canad. J. Chem., 1960, 38, 2042.

6. E.M. Burgess, H.R. Penton, and E.A. Taylor,Jr., J. Org. Chem., 1973, 38, 26.

7. J. Tsuji, Y. Yamakawa, M. Kaito, and T. Mandrai, Tetrahedron Letters, 1978, 2075.

8. W.T. Brady, and L. Smith, J. Org. Chem., 1971, 36, 1637.

9. L. Crombie, J. Chem. Soc., 1955, 2535; L. Crombie, J. Chem. Soc., 1955, 1007.

10. R. Aumann, Chem. Ber., 1977, 110, 1432.

11. H.D. Murdoch, Hely. Chim. Acta., 1964, 47, 936.

12. Thanks to R. Bonjouklian for supplying spectral data of the authentic material.

13. A.A. Ruden and. R. Bonjouklian, J. Amer. Chem. Soc., 1975, 97, 6892.

14. R.A. Ruden and R. Bonjouklian, J. Org. Chem., 1977, 42, 4095. 163,

15. R. Baudouy and J. Gore, Tetrahedron, 1975, 31, 383.

16. A. Claesson and C. Bogentoft, Acta. Chim. Scand., 1972, 26, 2540.

17. J.P. Schaefer and L. Endries, Org. Synth. Coil. Volume 5, 285.

18. W.E. Barrett and L. Needham, J. Org. Chem., 1975, 40, 2843.

19. G.W. Holbert, L.B. Weiss, and B. Ganem, Tetrahedron Letters, 1976, 4435.

20. A.T. Blomquist and D.T. Longone, J. Amer. Chem. Soc., 1957, 79, 3916.

21. W.J. Bailey and H.R. Golden, J. Amer. Chem. Soc., 1953, 75, 4780.

22. R. Aumann, H. Wormann, and C. Kruger, Chem. Ber., 1976, 109, 174.

23. R. Aumann and J. Knecht, Chem. Ber., 1978, 111, 3927.

24. H. Alper and D. Des Roches, Tetrahedron Letters, 1977, 4155.

25. M. Rosenblum, M.R. Saidi, and M. Madhavaroa, Tetrahedron Letters, 1975, 4009.

26. G.D. Annis and S.V. Ley, J.C.S. Chem. Comm., 1977, 581.

27. Dr. R. Sivaramakrishnan, personal communication.

28. Y. Becker, A. Eisenstadt, and Y. Shvo, Tetrahedron, 1974, 30, 839.

29. Y. Becker, A. Eisenstadt, and Y. Shvo, Tetrahedron, 1976, 32, 2123.

30. Y. Becker, A. Eisenstadt, and Y. Shvo, Tetrahedron, 1978, 34, 799.

31. Y. Becker, A. Eisenstadt, and Y. Shvo, Tetrahedron Letters, 1972, 3183.

164.

32. Thanks to Shell Limited, Sittingbourne for recording a 360 MHz spectrum.

33. R.J. Angelici, Accounts Chem. Res., 1972, 5, 335.

34. E.L. Stogryn and S.J. Brois, J. Amer. Chem. Soc., 1967, 89, 604.

35. A. Mishra, S.N. Rice, and W. Lwowski, J. Org. Chem., 1968, 33, 481.

36. Thanksto Mr. C.R. Self for supplying a sample of this compound.

37. K. Nakanishi, T. Goto, S. Ito, S. Noyori, and S. Nozoe, '.Natural Product Chemistry', Volume 1, Academic Press, New York, p. 48.

38. W.G. Dauben, J.W. McFarland, and J.B. Royan, J. Org. Chem., 1961, 26, 297.

39. J.L. Roberts, P.S. Borromeo, and G.D. Poulter, Tetrahedron Letters, 1977, 1621.

40. R. Mayer and W. Foerst, 'Newer Methods of Preparative Organic Chemistry', Volume 2, Academic Press, New York, 1963, p. 101.

41. M.F. Jung and M.A. Lyster, J. Amer. Chem. Soc., 1977, 99, 968.

42. D. Seyferth, M.S. Newman, and S. Ramachandra, Org. Synth. Coll. Vol. 4, 258.

43. R. Chaabouri and A. Laurent, Synthesis, 1975, 464.

44. G. Alverhe and A. Laurent, J. Chem. Research (S), 1978, 28,

45. J.S. Cowie, P.D. Landon, and S.R. Landor, J.C.S. Perkin I, 1973, 720.

- 46. H.R. Snyder and L.A. Brooks, Org. Syn. Coll. Vol. 2, 171.

47. M. Korachi, D.R. Nielsen, and W.H. Rideout, Org. Syn., 1962, 42, 50.

48. A.P. Krapcho, T. Diamanti, C. Cayen, and R. Bingham, Org. Syn. Coll. Vol. 5, 198. 165.

49. P. Pinkey, Org.. Syn. Coll. Vol. 2, 110.

50. G.P. Newsoroff and S. Sternhell, Aust. J. Chem. 1972, 25, 1669.

51. E.M. Hebblethwaite, personal communication, Imperial College.

52. J.R. Malpass and N.J. Tweedle, J.C.S. Perkin I, 1977, 874.

53. P.K. Wong, M. Madhavarao, D.F. Marten, and M. Rosenblum, J. Amer. Chem. Soc., 1977, 99, 2823.