On the question of cyclopropylidene intermediates in cyclopropene-to-allene rearrangements - Tetrakis(trimethylsilyl)cyclopropene, 3-alkenyl-1,2,3-tris(trimethylsilyl)cyclopropenes, and related model compounds

FULL PAPER

On the Question of Cyclopropylidene Intermediates in Cyclopropene-to-Allene

Rearrangements ؊ Tetrakis(trimethylsilyl)cyclopropene, 3-Alkenyl-1,2,3-

tris(trimethylsilyl)cyclopropenes, and Related Model Compounds

[

a]

[a]

[b]

[c][‡]

Armin de Meijere,* Dietmar Faber, Ute Heinecke, Robin Walsh,* Thomas Müller,

and Yitzhak Apeloig*^[

d]

Dedicated to Professor Jürgen Troe on the occasion of his 60th birthday

Keywords: Cyclopropenes / Reactive intermediates / Carbenes / Unimolecular reactions / Silyl group effects / Quantum

chemical calculations

Several tetrasubstituted cyclopropenes have been prepared

and their pyrolyses and photolyses have been investigated.

silylcyclopropene (38) at the QCISD(T)/6-311G*//B3LYP/6-

311G* + ZPVE level of theory, and on 3,3-dimethyl-1-(trime-

Tetrakis(trimethylsilyl)cyclopropene (10), which was ob- thylsilyl)cyclopropene (36b) at the B3LYP/6-311G*//B3LYP/

tained in 25% yield from tris(trimethylsilyl)cyclopropenylium

hexachloroantimonate (9), gave tetrakis(trimethylsilyl)allene

6-311G* + ZPVE level. These calculations provided us with

detailed energy surfaces for the potential pyrolysis pathways.

(12) as the sole product both thermally and photochemically. Although the potential cyclopropylidene species in these re-

Kinetic studies in [D

8

]toluene indicated first-order behavior

arrangements are significantly stabilized, for none of the sys-

tems was this sufficient to permit isomerization via these in-

termediates. 36b is calculated to rearrange via a vinylidene

intermediate to give 3-methyl-1-trimethylsilyl-1-butyne (47),

in agreement with experiment. Comparison of the calcula-

−1

with Arrhenius parameters log(A/s ) = 11.75 ± 1.20 and E

a

=

−

1

(

37.5 ± 2.5) kcal mol . All three new 3-alkenyl-1,2,3-tris(tri-

methylsilyl)cyclopropenes (17a−c, with C -, C -, and C -alk-

enyl groups as tethers, respectively) gave allenes upon irra-

1

2

3

diation, but thermally only two (17a, 17c) gave allenes, whilst

3

tions for 36a and 36b shows that H Si− is a poor model for

1

7b yielded a bicyclo[4.1.0]hept-3-ene derivative 22 as a re- an Me Si− substituent in these rearrangements. When an ap-

3

sult of an intramolecular ene reaction. Photolyses of two fur-

propriate correction is applied, the calculations on disilyl-

ther cyclopropenes (33a,b) bearing 1,2-bis(alkenyldimethyl-

(37) and tetrasilylcyclopropenes (38) are consistent with the

silyl) substituents also gave the corresponding allenes as the experimental findings that the trimethylsilyl-substituted

sole products. For none of these tethered cyclopropenes was cyclopropenes 48 and 10 form allenes 49 and 12, respectively,

a product found that could have originated from intramolecu- via vinylcarbene-type intermediates. These findings consid-

lar trapping of a cyclopropylidene intermediate. Quantum

mechanical (ab initio) calculations have been carried out on

the silyl-substituted cyclopropene model compounds 3,3-di-

methyl-1-silyl- (36a), 3,3-dimethyl-1,2-disilyl- (37), and tetra-

erably extend our understanding of silyl group substituent

effects on the various intermediates involved in cyclopro-

pene rearrangements.

Introduction

irradiation to yield dienes, acetylenes, or allenes.^[¹^Ϫ³^]The

kinetics of these rearrangements, the product compositions,

Due to their inherent ring strain, cyclopropenes readily and even the mechanisms by which these products are

undergo ring-opening reactions upon moderate heating or formed, have been found to depend on the substitution pat-

tern and the nature of the substituents on the cyclopro-

pene ring.^[

2,3]

[

a]

Institut für Organische Chemie, Georg-August-Universität

Göttingen,

The originally suggested ring-opening mode for a cyclo-

propene 2, particularly that of the parent hydrocarbon, in-

volves the formation of a diradical 1, which is a valence

Tammannstrasse 2, 37077 Göttingen, Germany

Fax: (internat.) ϩ 49-(0)551/399475

E-mail: Armin.deMeijere@chemie.uni-goettingen.de

Department of Chemistry, University of Reading,

Whiteknights, P. O. Box 224, Reading RG6 6AD, England

Institut für Anorganische Chemie der Humboldt-Universität

Berlin,

Hessische Strasse 1Ϫ2, 10115 Berlin, Germany

Department of Chemistry and the Lise Meitner Center for

[

b]

c]

[4]

isomer of a vinylcarbene 4, and can be either a singlet or

a triplet species. With appropriate alkyl substituents on 1/

4, the diradical/vinylcarbene can undergo a 1,2- or a 1,4-H

migration to yield dienes 7. However, when at least one of

[

d]

‡]

1

2

the substituents R or R on 2 is a hydrogen atom, it has

Computational Quantum Chemistry, Technion

Institute of Technology,

Ϫ Israel

since been shown^[

3c,5]

that the alkyne product 6 is not

Technion City, 32000 Haifa, Israel

formed via the diradical 1, but rather via a synchronous

Present address: Institut für Anorganische Chemie der Goethe-

Universität, Marie-Curie-Strasse 11, 60439 Frankfurt/Main, process of ring-opening accompanied by 1,3-H migration

Germany

(

extrusion) to yield an alkenylidene (a substituted vinylid-

Supporting information for this article is available on the

WWW under http://www.wiley-vch.de/home/eurjoc or from

the author.

ene) intermediate 3. This is then followed by a 1,2-shift of

2

the substituent R to yield the alkyne 6 (Scheme 1).

Eur. J. Org. Chem. 2001, 663Ϫ680



WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001

1434Ϫ193X/01/0204Ϫ0663 $ 17.50ϩ.50/0

663

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

FULL PAPER

the formation of the allene 8, the best chance of observing

it (or at least of generating the example with the lowest

activation energy for its formation) should be manifested in

the case of tetrakis(trimethylsilyl)cyclopropylidene 11

(

Scheme 2), which has the maximum possible stabilization

by four β-silyl groups. Compounds 17a؊c and 33a,b were

studied with a view to possible intramolecular trapping of

the cyclopropylidene intermediate, if formed, by means of

a tethered double bond. Skattebøl et al.^[showed some 30

years ago that a cyclopropylidene (or at least a cyclopro-

pylidenoid 15) generated from a 2-alkenyl-1,1-dibromocy-

clopropane 14 (n ϭ 1, 2) with methyllithium, is very effici-

ently trapped by the intramolecular double bond to give the

spiropentane 16, especially with a two-carbon tether. This

suggests that 19 would be a reasonable product of 17 if the

cyclopropylidene 18 were involved as an intermediate in its

15]

Scheme 1

When trimethylsilyl substituents are introduced, some in- thermal or photochemical rearrangement (Scheme 3).

triguing variations are observed. If one of the vinylic sub-

stituents is still a hydrogen, the rearrangement proceeds

1

2

3

4

normally. Thus, 2 (R ϭ Me Si, R ϭ H, R ϭ R ϭ Me)

3

[3b]

yields the corresponding alkyne 6 with Ͼ 99% selectivity.

1

2

However, when R ϭ Me Si and R ϶ H, the reaction leads

3

exclusively to allenes 8.^[

3a][3e,6]

Bearing in mind that cyclo-

propylidenes such as 5, when generated from a diazocyclo-

[

7]

[8]

propane or a 1,1-dihalocyclopropane via a carbenoid,

are prone to undergo ring opening to allenes, we conceived

the possibility that in a 1-silyl-substituted cyclopropene 2,

[

9,10]

a fast 1,2-silyl shift

might occur in the first step leading

to a cyclopropylidene 5,^[which could then rearrange to

11]

Scheme 2

the observed allene 8. An example of a 1,2-silyl shift leading

[12]

to a carbene has been reported by Barton et al.,

found that heating of 1,1,3,3-tetramethyl-2,4-dimethylene-

,3-disilacyclobutane results in its rearrangement by consec-

who

1

utive 1,2-silyl and 1,2-H shifts via 1,1,3,3-tetramethyl-2-

methylene-1,3-disilacyclopent-4-ylidene to 1,1,3,3-tetra-

methyl-2-methylene-1,3-disilacyclopent-4-ene. Of course, an

allene 8 would also result from a fast 1,2-silyl shift in the

vinylcarbene 4. The existence of vinylcarbene intermediates

formed by ring opening of substituted cyclopropenes^[has

been proved by very efficient trapping with alkenes in the

4]

[

13a][13b]

case of tetrachlorovinylcarbene,

nyl)carbenes,

(2Ј,2Ј-dialkyloxyvi-

1,2-dihalo-3,3-dimethylvinylcarbenes,^[¹³^d^]

and more recently for 3,3-dialkyloxy-1-(trimethylsilyl)vinyl-

[

13c]

[

13e]

carbenes.

We have mentioned the two possibilities for

the formation of allenes from silyl-substituted cyclopro-

penes in a previous report^[and, in order to pursue them Scheme 3

further, we have now combined an experimental investi-

gation of some specifically designed tetrasubstituted cyclo-

propenes with quantum mechanical calculations on closely

related model compounds.

3e]

For the experimental study, tetrakis(trimethylsilyl)cyclo-

propene 10, a set of three 3-ω-alkenyl-1,2,3-tris(trimethylsi-

Computationally, we have studied the thermal rearrange-

lyl)cyclopropenes, 17aϪc, and a pair of 1,2-bis(ω-alkenyldi- ment of 3,3-dimethyl-1-silylcyclopropene 36a, its trimethyl-

methylsilyl)-3,3-dimethylcyclopropenes 33a,b were de- silyl-substituted analogue 36b, the 3,3-dimethyl-1,2-disilyl-

signed. The logic behind the choice of 10 was based on the cyclopropene 37, and the tetrasilyl-substituted cyclopro-

well-known ability of a β-silyl group to stabilize a carbo- pene 38. In addition to shedding light on the general mech-

[

14]

cation,

which should also apply to a carbene center. We anism of breakdown for each of these compounds, which

reasoned that if a cyclopropylidene 5 were to be involved in will reveal whether or not cyclopropylidene intermediates

664

Eur. J. Org. Chem. 2001, 663Ϫ680

Cyclopropylidene Intermediates in Cyclopropene-to-Allene Rearrangements

FULL PAPER

are involved, the comparison of 36a and 36b allows us to on the relative energies of the transition state and ground

probe the frequently made assumption that a silyl group state of such species. We can infer that the ground state is

can be used to computationally model the experimentally less stabilized (or more destabilized) than the transition

1

2

more convenient trimethylsilyl substituent.

state as compared to the rearrangement of 2 (R ,R ϭ

3

4

SiMe ; R ,R ϭ Me). The steric bulk of the trimethylsilyl

3

groups is almost certainly at least partially responsible for

this, although the kinetic effect alone does not allow any

conclusion to be drawn as to whether the intermediate is a

cyclopropylidene 11 or a vinylcarbene 13. This question was

therefore addressed by computational methods (see below).

For the intramolecular trapping experiments, three 3-alk-

enylcyclopropenes 17a؊c with C -, C -, and C -tethers, re-

spectively (n ϭ 1, 2, 3) were readily prepared from tris(tri-

methylsilyl)cyclopropenylium hexachloroantimonate (9)

and the appropriate alkenylmagnesium bromide in diethyl

ether (Scheme 4). All three compounds 17aϪc were purified

by preparative gas chromatography and were obtained as

colorless oils in yields of 40%, 45%, and 51%, respectively.

Upon irradiation with a low-pressure mercury lamp (λ ϭ

Experimental Studies and Results

To probe the aforementioned hypothesis, the hitherto un-

known cyclopropene 10 had to be synthesized. Several at-

tempts to prepare 10 by reductive silylation of the readily

available tetrachlorocyclopropene with magnesium and

chlorotrimethylsilane in tetrahydrofuran were unsuccess-

1

2

3

[

16]

ful.

HMPA), however, the hexakis(trimethylsilyl)-3,3Ј-bicyclo-

propenyl was formed, as reported by two groups shortly

In the presence of hexamethylphosphoric triamide

(

[

17,18]

thereafter.

Under these conditions, tert-butylchlorodi-

methylsilane and tetrachlorocyclopropene gave, besides the

corresponding bicyclopropenyl, a 1.8% yield of tetrakis-

1

2

3

4

(tert-butyldimethylsilyl)cyclopropene (2, R ,R ,R ,R ϭ Si-

2

54 nm) in n-pentane solution at 20 °C, all three tris(trime-

Me tBu).

2

thylsilyl)cyclopropenes 17aϪc underwent clean rearrange-

ment to the corresponding allenes 21aϪc, which were isol-

ated by preparative gas chromatography (Scheme 4). Even

the C -tethered compound 17b did not give any of the C -

Tetrakis(trimethylsilyl)cyclopropene (10) could, however,

readily be prepared by the reaction of tris(trimethylsilyl)cy-

[

19]

clopropenylium hexachloroantimonate (9)

with tris(tri-

2

methylsilyl)aluminum,^[which has previously been shown

20]

bridged spiropentane derivative 19b, which might have been

expected from a cyclopropylidene intermediate of type 18.

to be an excellent donor reagent of a trimethylsilyl anion

[

21]

equivalent. In this way, cyclopropene 10 was obtained as

a colorless solid (m.p. 196 °C) in 25% yield.

Upon heating at 252 °C, as well as on irradiation with a

low-pressure mercury lamp (λ ϭ 254 nm), 10 underwent

clean and quantitative rearrangement to tetrakis(trimethyl-

silyl)allene (12) (Scheme 2).^[The UV absorption of 10 has

a rather long wavelength maximum at 296.2 nm (ε ϭ 5592),

similar to that observed for tetrakis(tert-butyldimethylsilyl)-

22]

[

16]

cyclopropene,

while the allene 12 has its longest wave-

[22]

length absorption at just 273.3 nm (ε ϭ 36066). The ther-

mal rearrangement of 10 to 12 in solution could easily be

1

monitored by H NMR spectroscopy. Good linear first-

order plots [log (% reactant) versus time] were obtained at

all temperatures in the range 200Ϫ252 °C. Rate constants

were calculated by least-mean-squares fitting (Table 1). The

data were fitted to the Arrhenius equation to give the fol-

Ϫ1

lowing parameters: E ϭ 37.3 Ϯ 2.5 kcal mol , log(A/

a

Ϫ1

s

) ϭ 11.75 Ϯ 1.2.

Table 1. Rate constants for the isomerization of tetrakis(trimethyl-

silyl)cyclopropene (10)

T [°C]

200

0.30

215

1.10

230

3.73

252

15.16

k [10 s^Ϫ¹

Ϫ5

]

Scheme 4

The Arrhenius activation energy for the rearrangement

Ϫ1

of 10 was found to be 6.8 kcal mol lower than that for

Upon heating to 380 °C in cis-decalin solution in a pres-

1

2

3

,3-dimethyl-1,2-bis(trimethylsilyl)cyclopropene 2 (R ,R ϭ surized thick-walled glass ampoule, however, only 17a and

3

4

[3e,23]

SiMe ; R ,R ϭ Me).

Moreover, the rate constant for 17c gave the respective allenes 21a and 21c. On the other

the rearrangement of 10 was evaluated as being ca. 15 times hand, the C -tethered cyclopropene 17b underwent clean

3

2

1

2

3

4

greater than that in the case of 2 (R ,R ϭ SiMe ; R ,R ϭ rearrangement to a new product upon heating to 210 °C in

3

1

13

1

Me) at 500 K. This clearly shows that the replacement of toluene solution. On the basis of its H, C, and H- H

the 3-methyl group by trimethylsilyl has a significant effect NOESY NMR spectra, the structure of the new product

Eur. J. Org. Chem. 2001, 663Ϫ680

665

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

FULL PAPER

was

assigned

as

1,6,7exo-tris(trimethylsilyl)bicy- the alkenylchlorodimethylsilane reagents to yield the bis(si-

clo[4.1.0]hept-3-ene (22) (Scheme 4). The only reasonable lylated) cyclopropenes 33a (45%) Ϫ in which one of the allyl

mechanism to account for its formation would be an intra- groups had apparently undergone a base-catalyzed double-

molecular ene reaction, whereby one of the two allylic hy- bond shift Ϫ and 33b (40%), respectively. Upon irradiation

drogen atoms from the side chain in 17b ends up as the 7- of 33a and 33b in n-pentane with a low-pressure mercury

endo hydrogen atom in 22. Indeed, thermal rearrangement lamp (λ ϭ 254 nm), both were completely isomerized to the

of the labeled cyclopropene 17b-D with two deuterium corresponding allenes 35a and 35b, respectively (Scheme 6).

atoms in the allylic position gave 3,7endo-dideuterio- In neither case could any trace of the expected trapping

1

,6,7exo-tris(trimethylsilyl)bicyclo[4.1.0]hept-3-ene (22-D) products 34a,b be detected, which are akin to compounds

as the sole product. Again, no bridged spiropentane deriv- 19, but have a silicon atom in the chain bridging the spiro-

atives of type 19 were obtained, strongly indicating that pentane unit.

cyclopropylidenes of type 18 are not intermediates in either

the photochemical or thermal reactions of compounds 17.

The labeled compound 17b-D was obtained from 9 and

the Grignard reagent derived from 4-bromo-3,3-dideu-

teriobut-1-ene (29) (Scheme 5), which, in turn, was pre-

pared from the DielsϪAlder adduct 23 of methyl acrylate

[

24]

and anthracene. Reduction of 23 with lithium aluminum

deuteride yielded the deuterated hydroxymethyl derivative

[

24]

2

(

4 (quant.),

85%). Chain extension was achieved by treating the Grig-

nard reagent derived from 25 with methyl chloroformate

52%); reduction of the ester 26 with LiAlH gave the hy-

which was converted into the bromide 25

(

4

droxymethyl derivative 27 (quant.), which was cleaved by

flash vacuum pyrolysis to yield the 3,3-dideuteriobut-1-en-

4

-ol (28) (100%). The latter was then converted into the

deuterated homoallyl bromide (29) (65%). The Grignard re-

agent derived from 29 was finally added to the cyclopropen-

ylium salt 9 to yield 17b-D (46%).

Scheme 6

The absence of spiro compounds of types 19 or 34 might

be explained if they were too unstable and thus undergo

further rearrangement under the applied experimental con-

ditions. Indeed, the parent bridged spiropentane 16 (n ϭ 2)

rearranges relatively easily to give a mixture of products.^[

This would correspond to 19 (n ϭ 2) in our studies, which

cannot be involved because of the favored ene pathway

leading from 17b to 22. For the other cases, nothing is

known about the stabilities of the potential bridged spiro-

27]

Scheme 5

As a second type of model substrate for intramolecular pentane derivatives 19 (n ϭ 1, 3). It would appear extremely

trapping of a cyclopropylidene intermediate, the pair of 1,2- unlikely, however, that, even if thermally unstable, these

bis(silyl)-substituted cyclopropenes 33a and 33b, having compounds would rearrange cleanly to the observed prod-

tethered ω-vinyl groups on the silicon atoms, were straight- ucts 21a and 21c. It is also noteworthy that compounds 19

forwardly prepared (Scheme 6). 1-Lithio-3,3-dimethylcyclo- (n ϭ 1, 2, or 3) are not formed under photochemical condi-

propene (31), generated from 1,1-dibromo-2-chloro-3,3-di- tions either.

methylcyclopropane (30) by treatment with methyllithium

Although no products characteristic of a cyclopropyli-

according to Baird et al.,^[was electrophilically substi- dene intermediate could be detected, the observed forma-

25]

[26]

tuted with allylchlorodimethylsilane

thylsulfonyloxy)dimethylsilane^[

or but-3-enyl(me- tion of 22 from 17b is notable as the first example of an

to give the trisubsti- intramolecular ene reaction involving a cyclopropene. This

26c,26d]

tuted cyclopropenes 32a (80% yield) and 32b (63%), re- constitutes yet another mechanistic variation to add to the

spectively. The latter were deprotonated with lithium diiso- already rich and varied chemistry of cyclopropene re-

propylamide in THF and then substituted once more with arrangements.

666

Eur. J. Org. Chem. 2001, 663Ϫ680

Cyclopropylidene Intermediates in Cyclopropene-to-Allene Rearrangements

FULL PAPER

Computational Studies

Discussion

(

a) General Considerations

Methods

Multi-reference CI calculations by Yoshimine and co-

All calculations were performed with the Gaussian 94

series of programs. Initially, the structures of the molec-

workers^[have shown that the thermal isomerization of

the parent cyclopropene does not proceed via a cyclopro-

pylidene intermediate but rather via vinylmethylene and

propenylidene intermediates, yielding allene and propyne,

33]

[

28]

ules were fully optimized at the HF/6-31G* level of theory

and refined using density functional theory^[

29a]

employing

Becke’s three-parameter hybrid method using the LYP cor-

respectively (propyne being more stable than allene by just

related functional (B3LYP)^[

29bϪ29d]

and the polarized and

Ϫ1

0

.3 kcal mol ). According to the calculations, the allene is

triply split 6-311G* basis set. Frequency calculations at the

B3LYP/6-311G* level were used to characterize the various

stationary points as minima or transition states. Transition

structures were located using the TS routine of the

formed in two steps via a vinylmethylene intermediate 4,

1

2

3

4

R ,R ,R ,R ϭ H (see Figure 1). This reaction path is pre-

Ϫ1

ferred by 11.1 kcal mol (at the MRCI/DZP level of the-

ory) as compared to the alternative path via cyclopropyli-

Gaussian 94 package and the intrinsic reaction coordinate

1

2

3

4

[33]

dene 5, R ,R ,R ,R ϭ H.

A direct interconversion of

30]

(

IRC) method. Schaefer et al.^[have recently demonstrated

allene to propyne by a 1,3-H shift is unlikely in view of the

that the hybrid DFT B3LYP method with a sufficiently

large basis set represents a valuable alternative to elaborate

multi-reference CI calculations for systems related to those

under study here, such as the electrocyclic ring opening of

cyclopropylidene. Energies were also obtained at the

QCISD(T)/6-311G* level of theory^[using the B3LYP/6-

Ϫ1

very high calculated barrier of 94.9 kcal mol (at the SCF/

DZP level).

31]

3

11G* optimized geometries and were corrected using the

B3LYP/6-311G* zero-point energies

QCISD(T)/6-311G*//B3LYP/6-311G*

[denoted

as

ϩ

ZPVE].

QCISD(T) energies are known to give reliable results, even

in difficult cases for which a single-determinant approach

might not be sufficiently accurate.^[Unless stated other-

wise, the relative energies and the energies of the isodesmic

reactions quoted in this paper have been calculated at this

level. All species were calculated in their lowest singlet state.

32]

3 4

Figure 1. Part of the potential energy surface of C H at MRCl/

[33]

DZP according to Yoshimine et al. ; relative energies in kcal

Ϫ1

Results

mol

Computationally, we have studied the thermolysis of four

systems: 3,3-dimethyl-1-silylcyclopropene (36a), its trime-

thylsilyl analogue 36b, 3,3-dimethyl-1,2-disilylcyclopropene

Also by computational methods [at the QCISD(T)/6-

11G*//MP2/6-31G* level], Shimizu and Gordon^[found

34]

3

some interesting differences regarding the thermal isomeri-

zation of 1-silylcyclopropene (Figure 2) compared to that

of the parent cyclopropene (Figure 1). The two-step reac-

tion path by which cyclopropene is converted to allene via

the vinylmethylene intermediate (Figure 1) merges for the

(37), and the tetrasilylcyclopropene (38). While in the calcu-

lations on 36a, 37, and 38 an H Si group served as a model

3

for the Me Si group used in the experiments (e.g. 38 was

3

used as a model for 10), the computational results obtained

for 36b can be directly compared with the previously pub-

lished experimental results for this molecule.^[Moreover,

comparison of the potential energy surfaces (PESs) for the

decompositions of 36a and 36b allows an estimation of the

electronic and steric effects of the trimethylsilyl group in

1

-silyl-substituted cyclopropene into a concerted single-step

3b]

reaction. However, the barrier associated with the cyclopro-

Ϫ1

pylidene path is higher in energy by 11.7 kcal mol for 1-

silylcyclopropene as well (Figure 2). Therefore, the authors

concluded that cyclopropylidenes are probably not interme-

diates in the thermal isomerization of 1-trimethylsilylcyclo-

propenes either. The most stable isomer on the H SiC H

3

6b as opposed to a silyl group in 36a. This comparison is

important since calculations of the PESs of interest and,

in particular, of the transition states for the trimethylsilyl

analogues of 37 (i.e. 48) and 38 (i.e. 10), are still not feasible

due to the size of these molecules. The total energies and

zero-point energies of all the species calculated are available

as Supporting Information.

3

PES is 1-silylpropyne, but silylallene is more stable than 3-

silylpropyne (relative energies: Ϫ8.0, 0.0, and 2.9 kcal

Ϫ1

mol ). The 1,3-silyl migration by which 3-silylpropyne is

converted to silylallene has an associated barrier of just 52.9

Ϫ1

kcal mol , roughly half of the 94.9 kcal mol required

for the allene-to-propyne interconversion.

[35]

In a recent theoretical study, Goldberg et al.

have

1

3

shown that alkyl-substituted cyclopropenes (2, R ϭ R ϭ

2

4

1

2

3

4

H, R ϭ R ϭ Me; 2, R ϭ R ϭ R ϭ R ϭ Me) decom-

Eur. J. Org. Chem. 2001, 663Ϫ680

667

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

FULL PAPER

The calculations show that 39a is stabilized by 8.3 kcal

Ϫ1

mol relative to cyclopropylidene (42) [Equation (1), R ϭ

H] and that 40, with four β-silyl groups, is more stable than

Ϫ1

4

2 by 12.5 kcal mol [Equation (1), R ϭ SiH ]. This rep-

3

resents a substantial stabilizing effect, but, as expected, is

significantly smaller than that in the corresponding cyclo-

propyl cations. Thus, the tetrasilylcyclopropyl cation 45 is

Ϫ1

found to be 37.1 kcal mol more stable than the parent

cyclopropyl cation 46 [Equation (2)].

The occurrence of hyperconjugative stabilization in 40 is

supported by a natural bond orbital (NBO) analysis^[of

the wavefunction, as well as by the optimized geometry of

39]

Figure 2. Part of the calculated potential energy surface of

C H (SiH

3 3 3

) at the QCISD(T)/6-311G*//MP2/6-31G* level accord- the molecule (Figure 3). Thus, the occupancy of the form-

[34] Ϫ1

ing to Shimizu et al. ; relative energies in kcal mol

ally vacant carbenic 2p orbital is substantially higher in 40

than in 42 (0.15 e and 0.05 e, respectively), in agreement

with increased charge transfer from the β-CϪSi bonds to

the carbenic 2p(C) orbital. In the analogous carbocation,

ϩ

pose via the corresponding alkyl-substituted vinylidenes 3.

It was also found that although homolytic bond cleavage,

giving rise to vinylcarbenes 4, is in some cases energetically

preferred over vinylidene formation, it leads to an energetic

dead end, because barriers to subsequent hydrogen shifts

leading to allenes or dienes are all higher in energy than

those associated with the corresponding vinylidene path-

ways. It was concluded that for these systems vinylcarbenes

are not intermediates in the formation of alkynes from

cyclopropenes.

the β-silyl effect is much larger with a calculated 2p(C )

occupancy of 0.36 e in 45 (compared with 0.14 e in 46). As

1

2(3)

a result of the hyperconjugative interaction, the C ϪC

1

bonds (C is the carbenic carbon atom) in 40 are shortened

˚

by 0.031 A compared to those in 42 (1.502 A), while the

2(3)

˚ ˚

C

ϪSi bonds are elongated by 0.016 A and 0.003 A com-

pared to the C ϪSi bonds in cyclopropene 38 (1.890 A).

Concomitant with the shortening of the C ϪC

C ϪC distances, the C ϪC bond in 40 is elongated by

3

˚

1

2(3)

and

1

3

2

3

˚

0

.066 A as compared to that in 42 (see Figure 3).

(b) The β-Silyl Stabilization of Cyclopropylidenes

The question central to this study has been whether addi-

tional silyl substitution would change the energetics of the

competing reaction channels in favor of a cyclopropylidene

intermediate making this the preferred pathway. This qual-

itative expectation was based on the knowledge that silyl

groups at the β-position strongly stabilize electron-deficient

centers, particularly carbenium ions, through hyperconjug-

ative interaction between the σ-(SiϪC) bond and the empty

Figure 3. Calculated geometries (at the B3LYP/6-311G* level) of

cyclopropylidenes 40 (C ) and 42 (C2v

2

)

ϩ

[36]

2

p(C ) orbital.

For β-silyl-substituted carbocations, the

The evidence from the trapping experiments presented

silyl group stabilizing effect is extremely large, of the order

1

2

3

above indicates that the cyclopropylidene 5 (R , R , R ,

Ϫ1

of 30 kcal mol (calculated at the MP2/6-31G*//6-31G*

4

R ϭ Me Si) is probably not an intermediate in the re-

3

[

37]

to 34 kcal mol^Ϫ¹(measured in the gas phase)

[38]

level)

for a secondary carbocation.

arrangement of 10 to 12. Evidently, the hyperconjugative

stabilization of a cyclopropylidene by four silyl groups is

insufficient to make it an intermediate along the lowest-

energy reaction path leading from 10 to 12. As is shown

below, this is also the conclusion that emerges from the

computational results. To determine the preferred reaction

path, we computed the effect of silyl substitution on the

various reaction pathways for the thermal decomposition

of the four silyl-substituted cyclopropenes 36a,b, 37, and

The stabilization of cyclopropylidene by β-silyl groups

can be evaluated by the isodesmic Equation (1), which com-

pares the stabilities of di- and tetrasilyl-substituted singlet

cyclopropylidenes 39, 40 with that of the parent singlet

cyclopropylidene 42.

3

8. Compounds 37 and 38 constitute structurally close

[3e]

models for the experimentally studied systems 48 and 10,

respectively (the Me Si groups in the experimental systems

3

are replaced by H Si in the calculations). On heating, both

3

4

8 and 10 undergo isomerization to exclusively afford the

corresponding allenes 49 and 12, respectively. In contrast,

the monosilyl-substituted cyclopropene 36b isomerizes al-

most quantitatively to the 1-trimethylsilylalkyne 47.^[

3b]

668

Eur. J. Org. Chem. 2001, 663Ϫ680

Cyclopropylidene Intermediates in Cyclopropene-to-Allene Rearrangements

FULL PAPER

1

2

Table 2. Relative energies ∆E of C R

(SiR ) molecules

3

4Ϫx

3 x

1

2

Ϫ1 [a]

(

R ,R ϭ H, Me) [kcal mol

]

Compound B3LYP/

QCISD(T)/

6-311G*

-311G*^[

b]

6-311G*

[b]

6

[

b][c]

ϩ∆ZPVE

3

5

6a

23.1

71.3

54.5

62.2

Ϫ1.4

13.6

17.6

64.2

55.7

60.8

Ϫ5.5

6.4

17.4

63.0

54.6

59.0

Ϫ4.6

6.1

5

0a-Si

0a-H

3a

4a

[a]

1a

0

2a

Ϫ8.0

65.0

56.0

68.9

64.9

65.0

63.4

64.5

75.2

82.9

80.0

59.1

74.5

Ϫ9.6

63.0

55.7

71.5

63.7

66.2

54.8

57.5

68.4

76.3

74.7

57.3

74.0

Ϫ9.0

61.1

54.5

68.3

61.5

62.8

53.4

54.9

66.0

73.3

73.2

55.8

70.4

TS1a(H)

TS1a(Si)

TS2a(H)

TS2a(Si)

The calculated relative energies of all the computed spe-

cies, at several levels of theory, are summarized in Table 2.

The calculated reaction paths for the isomerization of cyclo- TS3a

TS4a(Si)

propenes 36a,b, 37, and 38 are also shown schematically in

Figures 4Ϫ8.

TS4a(H)

TS5a(Si)

TS5a(H)

TS6a

(c) Calculations on the Thermal Rearrangement Pathways

TS7a(Si)

TS8a(H)

for Silyl-Substituted Cyclopropenes

.........................................................................................................

[

d]

(

i) Monosilyl-Substituted Cyclopropenes: First, the iso- 36b

22.7

55.7

Ϫ2.5

14.6

22.5

54.9

5

0b-Si

3b ϵ 47

merization of 36a will be considered (Figure 4 and Fig-

[

d]

Ϫ2.4

ure 5). The most stable isomer on the C H Si potential en-

[d]

5

10

54b

14.4

[

a]

[a]

ergy surface is the butadiene 52a, which is 9.0, 4.4, and 51b

0

[

d]

Ϫ1

52b

TS1b(Si)

Ϫ7.4

57.0

68.1

65.9

67.8

64.0

Ϫ6.9

1

5.1 kcal mol lower in energy than the allene 51a and the

[d]

55.9

65.0

62.0

66.8

61.5

[

d]

isomeric 1-silyl- and 3-silylbutynes 53a and 54a, respect- TS2b(H)

TS3b

ively.

TS4b(Si)

There are three possible reaction channels for the thermal

[d]

TS4b(H)

isomerization of 36a. The first of these is a two-step process .........................................................................................................

involving the isomeric vinylcarbene intermediates 50a-H 37

19.8

71.5

5.8

14.7

62.9

0.6

14.6

62.2

1.0

6

5

0

9

and 50a-Si, which leads to the allene 51a (Path A, Figure 4).

[a]

The calculations show that both vinylcarbenes, but in par- 58

0

TS1c

TS4c

TS5c

62.8

61.1

72.0

77.9

51.4

62.2

52.0

63.8

71.4

50.7

60.5

51.1

62.9

70.1

49.8

ticular the α-silyl-substituted vinylcarbene 50a-Si, lie in

relatively flat energy minima, so that the formation of allene

5

1a from 36a along path A is almost concerted with an TS6c

Ϫ1

TS7c

overall barrier (via 50a-H) of 44.1 kcal mol ; the barrier

is higher, at 50.9 kcal mol , via 50a-Si. The α-silyl-substi-

.........................................................................................................

Ϫ1

3

8

32.3

76.9

7.8

27.5

69.7

0.8

26.8

68.6

0.9

tuted vinylcarbene 50a-Si^[is the key intermediate en route 40

40]

6

2

1

to the formation of butadiene 52a. The 1,4-hydrogen shift

[a]

0

in 50a-Si leading to 52a is subject to a relatively small bar-

TS1d

72.6

67.7

81.0

50.3

68.8

59.6

71.0

84.4

46.3

66.8

58.3

69.5

82.9

45.6

Ϫ1

rier of just 8.2 kcal mol , but the overall barrier for the TS4d

Ϫ1

TS5d

TS6d

TS7d

formation of 52a from 36a is 45.4 kcal mol , which is 1.3

Ϫ1

kcal mol higher than the barrier associated with allene

formation via 50a-H. The formation of alkyne 53a from

vinylcarbene 50a-Si through a 1,2-H shift [TS8a(H)] across

[

a]

Absolute energy of 51a: B3LYP/6-311G*//B3LYP/6-311G*:

Ϫ1

Ϫ486.06098 H, ZPVE: 80.1 kcal mol ; QCISD(T)/6-311G*//

B3LYP/6-311G*: Ϫ484.93003 H. Absolute energy of 51b: B3LYP/

the double bond is also associated with a high overall bar-

Ϫ1

rier of 53.0 kcal mol . Thus, the lowest barrier for the iso- 6-311G*//B3LYP/6-311G*: Ϫ604.07101 H, ZPVE: 134.3 kcal

Ϫ1

mol . Absolute energy of 58: B3LYP/6-311G*//B3LYP/6-311G*:

merization of 36a along path A amounts to 44.1 kcal

Ϫ1

Ϫ776.78441 H, ZPE: 89.8 kcal mol ; QCISD(T)/6-311G*//

Ϫ1

mol , leading via 50a-H to the allene 51a.

B3LYP/6-311G*: Ϫ775.13602 H. Absolute energy of 61: B3LYP/6-

Ϫ1

3

11G*//B3LYP/6-311G*: Ϫ1279.58435 H, ZPE: 74.0 kcal mol ;

[b]

Reaction path B (Figure 5) involves simultaneous CϪC

QCISD(T)/6-311G*//B3LYP/6-311G*: Ϫ1277.14590 H. Ϫ Using

bond breaking and a 1,2-vinylic H or SiH migration across

[c]

3

B3LYP/6-311G* geometries. Ϫ

Zero-point energies (ZPE) are

3

[d]

the CϪC bond to the adjacent sp -hybridized carbon atom, unscaled. Ϫ At B3LYP/6-311G*//B3LYP/6-311G* ϩ ∆ZPVE.

yielding a vinylidene intermediate. By both experimental

[

5c,e]

[33Ϫ35]

means

and calculations,

vinylidenes have been es- TS4a(Si)] are not minima, but are transition states that con-

tablished as being intermediates in the thermolysis of cyclo- nect 36a directly to the silyl-substituted butynes 53a and

propene and of alkyl-substituted cyclopropenes. For 36a, 54a. The barriers associated with these isomerizations are

Ϫ1

we find that the vinylidene structures [TS4a(H) and relatively low, amounting to 37.5 and 36 kcal mol , re-

Eur. J. Org. Chem. 2001, 663Ϫ680

669

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

FULL PAPER

Figure 4. Part of the calculated potential energy surface of C

3

HMe

2

3

(SiH ) [at the QCISD(T)/6-311G*//B3LYP/6-311G* ϩ ZVPE level];

Ϫ1

relative energies in kcal mol ; the lowest barrier is highlighted by a box

Figure 5. Part of the calculated potential energy surface of C

3

HMe

2

3

(SiH ) [at the QCISD(T)/6-311G*//B3LYP/6-311G* ϩ ZVPE level];

Ϫ1

relative energies in kcal mol ; the lowest barrier is highlighted by a box

Ϫ1

spectively, with formation of the 3-silyl-substituted alkyne 1,2-shifts are relatively high, amounting to 48.6 kcal mol

Ϫ1

5

4a being preferred by 1.5 kcal mol

.

[TS5a(Si)] and 55.9 kcal mol [TS5a(H)] for the silyl and

The third possible mechanism is a two-step process invol- the H shift, respectively. Furthermore, the second step, i.e.

ving migration of either a hydrogen or a silyl group across the electrocyclic ring opening of 55, is even higher in energy

Ϫ1

the CϭC bond, yielding cyclopropylidene 55 as an interme- than TS5a(Si), being 55.8 kcal mol above the precursor

diate (path C, Figure 5). The barriers associated with these 36a.

670

Eur. J. Org. Chem. 2001, 663Ϫ680

Cyclopropylidene Intermediates in Cyclopropene-to-Allene Rearrangements

FULL PAPER

A direct interconversion between allene 51a and 1-silyl- in full agreement with the experimental findings. Further-

-butyne 53a seems unlikely as all attempts to locate the more, the calculated activation energy of 39.0 kcal mol^Ϫ¹

1

transition state for the required 1,3-H shift failed and led for the process 36b Ǟ 53b is in reasonable agreement with

Ϫ1 [3b]

instead to TS8a(H), which connects the vinylcarbene 50a- the experimental data (37.2 Ϯ 0.4 kcal mol ).

Path A,

Si with the butyne 53a. In contrast, the isomerization of leading to the formation of butadiene 52b via the interme-

allene 51a to the terminal butyne 54a proceeds through a diate vinylcarbene 50b-Si, is calculated to be only 0.5 kcal

Ϫ1

concerted 1,3-silyl shift in TS7a(Si) (Path D). The calcu- mol

less favorable than the 36b Ǟ 53b process. This

Ϫ1

lated activation energy of 49.7 kcal mol for the reverse theoretical result is also in good agreement with the experi-

reaction (i. e. 54a Ǟ 51a) is in excellent agreement with the mental detection of ca. 1% of 52b as the only thermolysis

experimentally measured barrier of 49.9 Ϯ 0.6 kcal mol^Ϫ¹product besides 53b. The process leading to the allene 51b

Ϫ1

for the isomerization of 3-trimethylsilylpropyne (56) to tri- [via TS2b(H)] is a further 3.0 kcal mol higher in energy

methylsilylallene (57).^[

41]

(Figure 6) and, in agreement with this theoretical result, 51b

is not observed experimentally. Thus, the calculations and

the gas-phase experiments^[on the thermolysis of 36b are

in very good qualitative and quantitative agreement.

Comparison of the calculations on 36a and 36b reveals

that the substitution of H Si by an Me Si group has a very

3b]

3

significant effect. As expected, this is especially important

for reaction steps for which the transition states are

In summary, the calculations on 36a clearly show that the sterically congested. Thus, the activation energy required

reaction channel via the cyclopropylidene intermediate 55 for the 1,2-shift of the silyl group across the CϪC bond is

Ϫ1

(

path C) can be excluded as this pathway is much higher in 8.3 kcal mol higher for an Me Si group [TS4b(Si), E ϭ

3 a

Ϫ1

energy than the other possibilities. The lowest-energy path- 44.3 kcal mol ] than for an H Si group [TS4a(Si), E ϭ

3

a

Ϫ1

way for the thermal isomerization of 2,2-dimethyl-1-silylcy- 36.0 kcal mol ]. As expected, the effect is smaller for the

clopropene (36a) is ring opening via a vinylidene-like trans- analogous 1,2-H shift and the activation energy of TS4b(H)

Ϫ1

ition state TS4a(Si) (path B, Figure 5), yielding the terminal is only 1.5 kcal mol higher than that of TS4a(H). As a

3

-silyl-substituted butyne 54a. This computational result result of these Me Si vs. H Si effects, the relative energy

3 3

apparently contradicts the experimental finding that pyro- difference between the two vinylidene-like transition states

Ϫ1

lysis of the corresponding trimethylsilyl-substituted cyclo- leading to isomeric butynes decreases from 1.5 kcal mol

Ϫ1

propene 36b leads almost exclusively to the 1-trimethylsilyl- for H Si [TS4a(H) vs. TS4a(Si)] to Ϫ5.3 kcal mol for

3

butyne 47, along with just 1% of butadiene 52b as the only Me Si [TS4b(H) vs. TS4b(Si); TS4b(H) favored]. Entropy

3

identifiable by-product.^[However, the calculations were also disfavors the highly crowded transition state for Me Si

3b]

3

performed for an SiH substituent (i.e. on 36a) while the migration, TS4b(Si), relative to the other transition states.

3

experiments were carried out with an Me Si-substituted Thus, the differences (at 298 K) between the corresponding

3

Ϫ1

cyclopropene (i.e. with 36b). The steric requirements are free energies G are 2.4, 3.8, and 3.3 kcal mol

[for

greater in TS4a(Si), where the migrating group closely ap- TS4b(H), TS2b(H), and TS3b, respectively] larger than the

3

proaches the gem-dimethyl group at C , than in TS4a(H) corresponding differences between the enthalpies calculated

leading to the 1-silyl-substituted butyne 53a) and this effect at 0 K.^[On the other hand, the replacement of H Si by

43]

(

3

may reverse their relative stabilities, especially in view of Me Si stabilizes the α-silyl-substituted vinylcarbene 50b-Si

3

Ϫ1

the fact that the calculated energy difference between the relative to 50a-Si by 4.8 kcal mol , probably on electronic

transition states leading to the isomeric 54a and 53a is only grounds. As the Me Si vs. H Si substitution has opposite

3

Ϫ1

1

.5 kcal mol . Adcock et al. have recently pointed out that effects on the α-silylvinylcarbenes and on the TS for silyl

the common practice of theoretical modeling of Me Si by group migration along path B, substitution of H Si by

3

the use of H Si to reduce the complexity of the computa- Me Si has a large effect on the relative energies of TS4(Si)

3

tional problem can be inadequate.^[⁴²^]

Ϫ1

vs. 50-Si. While ∆E[TS4a(Si) Ϫ 50a-Si] is Ϫ1.2 kcal mol

To test the adequacy of the modeling of Me Si by H Si [TS4a(Si) being lower in energy], the corresponding ∆E

3

Ϫ1

in the present case, the competing low-energy reaction path- with the Me Si group is 11.9 kcal mol , with 50b-Si being

3

ways A and B for the experimentally studied system 36b lower in energy. This large effect resulting from the substitu-

were calculated (see Figure 6). Comparison of the calcula- tion of an H Si by an Me Si group must be taken into con-

3

tions for 36a and 36b indeed shows that the Me Si substitu- sideration whenever an H Si group is used as a model for

3

ent considerably destabilizes TS4b(Si) compared to the trimethylated analogue. This is exemplified here by 37

TS4b(H), so that the one-step isomerization of 36b to 53b and 38, for which, due to the size of the systems, we could

through H-migration becomes the most favorable reaction not carry out calculations on the Me Si derivatives used in

3

channel. Thus, while TS4a(Si) is lower in energy than the experimental study.

Ϫ1

TS4a(H) by 1.5 kcal mol , the order is reversed for 36b,

(ii) Disilyl- and Tetrasilyl-Substituted Cyclopropenes: The

Ϫ1

TS4b(H) being 5.3 kcal mol

TS4b(Si). This leads to the computational prediction that less complex than that of C Me H(SiH ), due to the higher

lower in energy than C Me (SiH ) potential energy surface (PES) (Figure 7) is

3

2

3 2

3

2

3

5

3b should be the major product upon thermolysis of 36b, symmetry of the precursor 37. The global minimum on this

Eur. J. Org. Chem. 2001, 663Ϫ680

671

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

FULL PAPER

3 2 3

HMe (SiMe ;

) (at the B3LYP/6-311G* ϩ ∆ZPVE level); relative energies in kcal mol^Ϫ¹

Figure 6. Calculated potential energy surface of C

the lowest barrier is highlighted by a box

PES is the allene 58, which is 1.0 kcal mol^Ϫ¹more stable corresponding Me Si analogue 48, which has been studied

3

than the alkyne 59. This stability order is different from experimentally (see below).

that for the monosilylated system, where the alkyne 53a is

The overall topologies of the PESs for the tetrasilyl- and

Ϫ1

more stable than the allene 51a by 4.6 kcal mol . The di- disilyl-substituted systems are very similar (see Figure 8).

silylcyclopropene 37 is found to be higher in energy than The allene 61 is more stable than the propyne 62 by 0.9

Ϫ1

the allene 58 by 14.6 kcal mol . Here, path A has become kcal mol . However, the cyclopropene 38 is strongly desta-

Ϫ1

merged into a single-step process leading to 58 via a vinyl- bilized, being 26.8 kcal mol higher in energy than 61.

Ϫ1

methylene-type transition state (TS1c), with an associated This difference is 12.2 kcal mol larger than that between

Ϫ1

barrier of 45.9 kcal mol (Figure 7). The cyclopropylidene 37 and 58. The cyclopropylidene 40 lies in a very shallow

6

0 is a minimum at the B3LYP/6-311G* level, but the bar- potential well, but here again reaction via this intermediate

rier for its back reaction to cyclopropene is extremely small (path C) is significantly higher in energy (requiring 56.1

Ϫ1

(

0.7 kcal mol ) and thus path C also practically merges kcal mol ) than the other two competing reaction chan-

into a single step, the highest point being TS6c, which lies nels. The concerted reaction via a vinylcarbene-type trans-

Ϫ1

5

5.5 kcal mol above 37 Ϫ the highest energy among the ition state (path A via TS1d) requires 40.0 kcal mol . The

investigated reaction channels for 37. The lowest-energy re- lowest-energy path (path B), leading to the propyne 62 via

action channel is path B, leading through a concerted pro- the vinylidene carbene-type transition state (TS4d), requires

Ϫ1

cess to the alkyne 59, with a barrier of 36.5 kcal mol

.

an activation energy of only 31.5 kcal mol . Due to the

Thus, for 37, the cyclopropylidene path can again be ex- significant gem-disilyl destabilization of the precursor 38,

cluded with confidence, as the barriers along this path are the activation energies for paths A and B are lower by 5.9

Ϫ1

significantly higher than those for the other reaction chan- and 5.0 kcal mol , respectively, than in the case of the 3,3-

nels. According to the calculations, the preferred isomeri- gem-dimethyl-substituted cyclopropene 37. Thus, on going

zation path for 37 is along path B, yielding exclusively the from 3,3-dimethyl substitution to 3,3-disilyl substitution,

butyne 59. Even in the gas phase, where 59 is generated the reaction rate is predicted to be significantly enhanced.

in a vibrationally excited state, it is unlikely that 59 would In agreement with this theoretical prediction, it has been

isomerize to the allene 58 under normal experimental con- found experimentally that replacement of the 3,3-dimethyl

ditions since the activation energy for this slightly exo- substituents in 48 by 3,3-bis(trimethylsilyl) substituents as

Ϫ1

thermic process, requiring a 1,3-silyl shift, is 48.8 kcal in 10 lowers the activation energy by 6.7 kcal mol

mol (path D). Thus, the calculations predict that the ther- (Table 3).

Ϫ1

mal rearrangement of 37 should yield exclusively the alkyne

The calculations reveal interesting differences between

5

9. As stated above, this is not necessarily the case for the the C H and the C H (SiH ) potential energy surfaces.

3

4

3

4Ϫn

3 n

672

Eur. J. Org. Chem. 2001, 663Ϫ680

Cyclopropylidene Intermediates in Cyclopropene-to-Allene Rearrangements

FULL PAPER

Figure 7. Calculated potential energy surface of C

3

Me

2 3 2

(SiH ) [at the QCISD(T)/6-311G*//B3LYP/6-311G* ϩ ∆ZPVE level]; relative

Ϫ1

energies in kcal mol ; the lowest barrier is highlighted by a box

3 3 4

Figure 8. Calculated potential energy surface of C (SiH ) [at the QCISD(T)/6-311G*//B3LYP/6-311G* level]; relative energies in kcal

Ϫ1

mol ; the lowest barrier is highlighted by a box

One important feature is that increasing silyl substitution HF/6-31G* level, the tetrakis(trimethylsilyl)propyne 63 is

Ϫ1

energetically favors the allene compared to the isomeric al- found to be 12.0 kcal mol higher in energy than the iso-

kynes or the precursor cyclopropene. Increased steric de- meric allene 12. This compares to an energy difference of

Ϫ1

mand at the propargylic carbon atom also increases the just 5.7 kcal mol for the tetrasilyl-substituted analogues

alkyneϪallene energy difference. Thus, at the HF/6-31G*// 62/61. Because of the size of these systems, these calcula-

Eur. J. Org. Chem. 2001, 663Ϫ680

673

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

FULL PAPER

Table 3. Calculated and experimental Arrhenius activation energies (E

a

) [kcal mol^Ϫ¹] of thermal cyclopropene isomerizations

Reaction barrier via path/product^[

B (vinylidene)

a]

Starting

material

A (vinylcarbene)

C (cyclopropylidene)

Ref.

3

6a

6b

7

calcd. 44.1/51a

calcd. 42.5/51b

calcd. 45.9/58

calcd. 40.0/61

calcd. 36.0/54a

calcd. 39.0/53b

calcd. 36.5/59

calcd. 31.5/62

calcd. 55.9/51a

this work

calcd. 55.5/58

calcd. 56.1/61

8

[

3b]

3e]

3

4

1

6b

8

0

exp. 37.2Ϯ0.4/53b

exp. 44.1Ϯ0.5/49

exp. 37.3Ϯ2.5/12

this work

[

a]

See text for definition of reaction paths.

tions were only possible at this relatively low level of theory. rearrangement to the acetylene 62. In contrast, it is found

Furthermore, even at this level it is not practical to search experimentally that the analogous tetrakis(trimethylsilyl)cy-

the PES, particularly with a view to locating the various clopropene (10) yields exclusively the allene 12. How can

transition states.

these differences be rationalized? As discussed above, the

Even more important is the finding that the direct in- calculations on the Me Si-substituted cyclopropene 36b vs.

3

terconversion of silylated alkynes and the corresponding all- the H Si-substituted cyclopropene 36a showed that replace-

3

enes can compete energetically with multi-step isomeri- ment of H Si by Me Si increased the reaction barrier asso-

3

zation via a cyclopropene. The barrier for the 1,3-silyl shift ciated with the pathway leading to the corresponding acet-

Ϫ1

interconverting alkynes and allenes is significantly lower ylene by ca. 8 kcal mol compared with that for the path-

Ϫ1

(

45Ϫ50 kcal mol ) for alkynes 54a, 59, and 62 as com- way leading to the corresponding allene [TS2a(Si) vs.

Ϫ1

pared with the 95 kcal mol barrier for H-migration in TS4a(Si) and TS2b(H) vs. TS4b(Si)]. This effect is expected

the parent cyclopropene. These computational results are to be even more pronounced for 38, due to the steric crowd-

supported by the experimental finding that the isomeri- ing of three large Me Si groups in TS4d as compared to

3

zation of 3-trimethylsilylalkyne (56) to trimethylsilylallene two smaller methyl groups and one trimethylsilyl group in

(

57) proceeds with a relatively low barrier of 49.9 Ϯ 0.6 kcal 36b. Even a conservative estimate of this additional desta-

Ϫ1 [41]

mol

.

An analogous rearrangement of tris(trimethylsi- bilization of TS4d compared with TS1d of just 3Ϫ4 kcal

Ϫ1

lyl)propyne (64) to tris(trimethylsilyl)allene (65) occurs mol would be sufficient to reverse the energy balance be-

upon prolonged heating of neat 64 at 110 °C, giving an tween the transition states leading to the acetylene (TS4d)

equilibrium mixture of 65 and 64 in a ratio of 60:40.^[

22]

and to the allene (TS1d) so as to favor the latter (the calcu-

lated difference between these two reaction channels in the

Ϫ1

H Si system 38 is 8.5 kcal mol ). Thus, taking into consid-

3

eration the effect of the Me Si groups, the calculations and

3

the experiments for 10 reach agreement. We hope that in

the future, with additional computational resources, we will

be able to explicitly calculate the thermal isomerization

channels for 10 and hence test this analysis.

One may therefore envisage an overall isomerization of

cyclopropenes 37, 38 to allenes 58, 61 via butynes 59, 62 as

intermediates (paths B and D, Figure 7 and 8). In principle,

when formed, 59 and 62 have sufficient internal energy to

undergo the subsequent isomerization to the corresponding Conclusions

allene. While in the gas phase the importance of the sub-

sequent reaction depends on the lifetime of the vibrationally

Our experimental and theoretical results clearly show

excited species, in practice, under normal pressures (10Ϫ50 that cyclopropylidenes are not involved as intermediates in

Torr) and for large molecules such as these, collisional sta- the isomerizations of silylated cyclopropenes. Thus, while

bilization will be fully effective. Of course, in the condensed the experimental activation energies for the isomerization

Ϫ1

phase, the excess energy will likewise be completely trans- of these systems are in the range of 37Ϫ44 kcal mol , the

ferred to the medium. If the rearrangement of an initially calculations show that isomerization via cyclopropylidenes

Ϫ1

formed alkyne to an allene were to have been the reaction would require much higher energies of ca. 55 kcal mol

path, then the alkynes 63 and Me SiCϵCC(SiMe )Me (66) (see Table 3). It is difficult to imagine that this striking dis-

3

2

should have been detected chromatographically or spectro- crepancy could be wholly attributed to the methyl groups

scopically. Since not even the slightest indication of the on the silyl residues, which were omitted in the calculations.

formation of alkyne 63 has ever been found in the studies Furthermore, the calculations point to two alternative

of the solution isomerization 10 Ǟ 12, this possibility can lower-energy rearrangement channels, with activation ener-

be excluded with confidence.

gies of the order of those found experimentally. Although it

The calculations predict that upon thermolysis the tetra- is difficult to make a definitive statement about the precise

silyl-substituted model system 38 should undergo exclusive mechanisms of these rearrangements on the basis of the

674

Eur. J. Org. Chem. 2001, 663Ϫ680

Cyclopropylidene Intermediates in Cyclopropene-to-Allene Rearrangements

FULL PAPER

calculations, the overall evidence seems to indicate that the enes via vinylcarbene-like transition states. This difference

thermolysis of 1,2-disilyl- and tetrasilyl-substituted cyclo- in mechanism, according to the calculations, is mainly due

propenes yields the corresponding substituted allenes in a to steric congestion of the Me Si groups in the vinylidene-

3

1

3

single-step process involving cleavage of the C ϪC bond in type transition states. Furthermore, the computations sug-

the cyclopropene and proceeding via a vinylcarbene-type gest that H Si and Me Si substitution of cyclopropenes

3

TS.

should lead to different pathways. We hope that it will be

As already discussed, the agreement between calculation possible to test this experimentally in the future.

and experiment is only good when the calculations include

the trimethylsilyl groups actually present in the experiments.

Thus, for 36b there is a good agreement between the ob-

served and the theoretically predicted product distribution

and activation energies (Table 3) and hence the mechanism

of the thermal rearrangement of 36b can be confidently

concluded to proceed via a vinylidene-type transition state.

The situation is more complex for 10 and 48. Calculations

of the potential energy surfaces for these systems at a suffi-

ciently high level of theory are unfortunately beyond our

computational means. Thus, we have to make do with the

computational results for the model compounds 37 and 38,

Scheme 7

where the Me Si groups used experimentally are replaced

3

by SiH groups. Calculations on the rearrangements of 37

3

and 38 predict a concerted ring opening via a vinylidene-

like transition state in each case, to give the alkynes 59 and Experimental Section

6

2, respectively, with relatively low associated activation

Ϫ1

General Remarks: Unless otherwise indicated, all reactions were

performed in oven- or flame-dried glassware under argon. Solvents

barriers of 36.5 and 31.5 kcal mol . This is inconsistent

with both the measured activation barriers (which for both

were distilled immediately prior to use: Et

dium/benzophenone ketyl, toluene and acetonitrile from LiAlH

and CH Cl from CaH . Ϫ Thin-layer chromatography was per-

2

O and THF from so-

1

0 and 48 are 6Ϫ8 kcal mol^Ϫ¹higher than calculated)

4

,

(Table 3) and with the observed allenic products 12 and

2

4

9, respectively.

formed on MachereyϪNagel Alugram Sil G/UV254 Fertigfolien

and spots were visualized under UV light or by exposure to p-

The theoretical conclusions change, however, if we take

into consideration the fact that the vinylidene-like trans- anisaldehyde or molybdatophosphoric acid. Ϫ NMR spectra were

ition states TS4e and TS4f (see Scheme 7), which involve recorded with Bruker AM 250 or Varian VXR 500 S spectrometers

silyl group migration, are considerably destabilized by the with samples in CDCl , C D , or [D ]toluene solution. Coupling

3 6 6 8

constants are given in Hz. Ϫ IR spectra were recorded with a

Bruker IFS 66 FT-IR spectrometer. Ϫ UV spectra were measured

with a Varian Cary 219 spectrophotometer. Ϫ Melting points were

measured with a Büchi melting point apparatus; values are uncor-

rected. Ϫ The pure compounds were isolated by preparative GC

with an Intersmat 131 instrument using a thermal conductivity de-

tector. The column used was 3.5 m ϫ 10 mm, 15% SE 30 on 60Ϫ80

steric requirements of the trimethylsilyl group as compared

to the vinylcarbene-like transition states TS1e and TS1f.

Thus, TS4b(Si) is destabilized relative to the corresponding

Ϫ1 [44]

vinylcarbene 50-Si by 13.3 kcal mol

when H Si is re-

3

placed by Me Si (Figures 4Ϫ6). When the barriers associ-

3

ated with the Me Si migration are corrected accordingly by

3

increasing the energies of the vinylidene-type transition

2

mesh Chromosorb W-AW-DMCS; carrier gas was H . Ϫ The

states leading to formation of alkynes 59, 62 by ca. 10 kcal photolyses were carried out with a low-pressure mercury UV lamp

Ϫ1

mol , path A via TS1e (which is less sensitive to steric (254 nm, 250 W) from Gräntzel, Karlsruhe. Ϫ Mass spectra were

effects), leading exclusively to allenes, becomes the lowest- obtained with a Finnigan MAT 95 spectrometer. High-resolution

mass spectra were obtained with a Varian MAT 311 A spectro-

energy reaction path. Support for this interpretation is the

meter. Molecular compositions were determined by high-resolution

good agreement between the activation energies computed

mass spectrometry with pre-selective ion peak matching at R ഠ

for path A in the cases of 37 and 38 (see Table 3) and those

determined experimentally for 48 and 10, respectively. In

1

0000 within Ϯ 2 ppm of the exact masses. Ϫ Microanalyses were

performed by the staff of the Mikroanalytisches Laboratorium, In-

stitut für Organische Chemie der Georg-August-Universität

Göttingen.

contrast, the calculated barriers in the H Si-substituted

3

model compounds (Figure 7 and 8) for the reaction along

path B are too low.

Tetrakis(trimethylsilyl)cyclopropene (10): To a suspension of the

In summary, the combined experimental and theoretical

study described herein allows us to confidently conclude

that cyclopropylidenes are not intermediates in the thermal

isomerizations of silyl-substituted cyclopropenes. The com-

putations predict that 1-trimethylsilylcyclopropene should

isomerize to trimethylsilylpropyne via a vinylidene-type

cyclopropenylium salt 9 (354 mg, 0.6 mmol) in Et

Ϫ78 °C, a solution of tris(trimethylsilyl)aluminum (675 mg,

.7 mmol) in Et O (5 mL) was added over a period of 20 min. The

reaction mixture was stirred at Ϫ78 °C for 9 h and then diluted

with Et O (25 mL) and quenched with H O (10 mL). The organic

layer was separated, washed with brine (10 mL), and dried with

2

O (13 mL) at

2

transition state, whereas the bis- and tetrakis(trimethylsilyl)- Na SO . The solvent was evaporated under reduced pressure. Puri-

2

4

cyclopropenes should isomerize to the corresponding all- fication of the residue by preparative gas chromatography gave 49

Eur. J. Org. Chem. 2001, 663Ϫ680 675

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

FULL PAPER

mg (25%) of 10 as a colorless solid (m.p. 196 °C). Ϫ IR (KBr): ν˜ ϭ

28 mg (93%) of the allene 21a as a colorless oil. Thermolysis of

17a (20 mg, 0.07 mmol) in cis-decalin (2 mL) at 380 °C for 24 h

gave, after purification by preparative gas chromatography, 14 mg

2

5

957, 2900, 1721, 1406, 1249, 1020, 979, 758, 683, 641, 623, 611,

Ϫ1

1

3

22, 505 cm . Ϫ H NMR (250 MHz, CDCl ): δ ϭ 0.20 (s, 18

13

H), Ϫ0.11 (s, 18 H). Ϫ C NMR (62.9 MHz, CDCl

2.3, 0.2, Ϫ0.4. Ϫ C15 (328.8): calcd. C 54.80, H 11.04; found

C 54.76, H 10.10.

): δ ϭ 135.7, (70%) of the allene 21a. Ϫ IR (film): ν˜ ϭ 2956, 2898, 1894, 1734,

3

Ϫ1

1

2

H36Si

4

1248, 1019, 900, 839, 759, 690, 638 cm . Ϫ H NMR (300 MHz,

): δ ϭ 5.82Ϫ5.96 (m, 1 H), 4.94Ϫ5.06 (m, 2 H), 2.64Ϫ2.68

6 6

C D

1

3

(

m, 2 H), 0.18 (s, 18 H), 0.13 (s, 9 H). Ϫ C NMR (300 MHz,

): δ ϭ 208.0, 138.4, 114.6, 79.6, 75.9, 33.2, 0.6, Ϫ1.0. Ϫ

HRMS: m/z calcd. for C15 296.1812; found 296.1811.

General Procedure for the Preparation of 3-Alkenyl-1,2,3-tris(trime-

thylsilyl)cyclopropenes: To a suspension of 1 equiv. of the cyclopro-

6 6

C D

H

32Si

3

penylium salt 9 in Et

of the alkenylmagnesium bromide in Et

ture for 3 h at room temp., the reaction was quenched by the addi-

tion of saturated aq. NH Cl solution. The organic layer was

4

washed with H O and brine and dried with MgSO . The solvent purification of the residue by preparative gas chromatography gave

2

O at 0 °C was added a solution of 10 equiv.

2

O. After stirring the mix-

1,1,3-Tris(trimethylsilyl)-1,2,6-heptatriene (21b): Irradiation of 17b

(30 mg, 0.10 mmol) in n-pentane (30 mL) for 3 h with a low-pres-

sure mercury lamp (λ ϭ 254 nm), evaporation of the solvent, and

4

2

was evaporated in vacuo. Purification of the residue by preparative

gas chromatography gave a colorless oil in each case.

26 mg (87%) of the allene 21b as a plastic mass. Ϫ IR (film): ν˜ ϭ

2956, 2899, 1894, 1641, 1402, 1248, 903, 839, 760, 689, 620, 522

Ϫ1

1

cm . Ϫ H NMR (250 MHz, CD

3

CN): δ ϭ 5.79Ϫ5.97 (m, 1 H),

3-Allyl-1,2,3-tris(trimethylsilyl)cyclopropene (17a): Addition of a

4

.89Ϫ5.07 (m, 2 H), 2.09Ϫ2.21 (m, 2 H), 1.30Ϫ1.43 (m, 2 H), 0.08

solution of allylmagnesium bromide (3.39 mmol) in Et

2

O (4 mL)

O (5 mL)

13

(

s, 18 H), 0.04 (s, 9 H). Ϫ C NMR (62.9 MHz, CD

07.4, 140.5, 115.2, 82.1, 77.9, 36.1, 23.8, 0.7, Ϫ1.1. Ϫ HRMS: m/z

calcd. for C16 310.1968; found 310.1968.

3

CN): δ ϭ

to cyclopropenylium salt 9 (200 mg, 0.34 mmol) in Et

2

gave 40 mg (40%) of 17a. Ϫ IR (film): ν˜ ϭ 3074, 2956, 2897, 1710,

H

34Si

3

Ϫ1

1

636, 1406, 1248, 991, 908, 839, 754, 693, 628, 515 cm . Ϫ

NMR (250 MHz, CDCl ): δ ϭ 5.60 (ddt, J ϭ 7, 10, 17 Hz, 1 H), 1,1,3-Tris(trimethylsilyl)-1,2,7-octatriene (21c): Irradiation of 17c

.79Ϫ4.90 (m, 2 H), 2.05 (dt, J ϭ 1, 7 Hz, 2 H), 0.20 (s, 18 H),

H

3

4

(30 mg, 0.09 mmol) in n-pentane (30 mL) for 2 h with a low-pres-

13

Ϫ0.09 (s, 9 H). Ϫ C NMR (125 MHz, CDCl

3

): δ ϭ 141.0, 138.9, sure mercury lamp (λ ϭ 254 nm), evaporation of the solvent, and

1

14.5, 44.3, 16.4, Ϫ0.4, Ϫ1.2. Ϫ HRMS: m/z calcd. for C15

H

32Si

3

purification of the residue by preparative gas chromatography gave

296.1812; found 296.1811.

27 mg (90%) of the allene 21c as a colorless oil. Thermolysis of 17c

(

20 mg, 0.06 mmol) in cis-decalin (2 mL) at 380 °C for 22 h gave,

3-(4Ј-But-1Ј-enyl)-1,2,3-tris(trimethylsilyl)cyclopropene (17b): Addi-

after purification by preparative gas chromatography, 15 mg (75%)

tion of a solution of 4-but-1-enylmagnesium bromide (4.24 mmol)

in Et O (3 mL) to cyclopropenylium salt 9 (250 mg, 0.42 mmol) in

Et

O (7 mL) gave 59 mg (45%) of 17b. Ϫ IR (film): ν˜ ϭ 2956, 1708,

640, 1406, 1249, 906, 840, 755, 693, 631, 516 cm^Ϫ¹. Ϫ H NMR

250 MHz, CDCl ): δ ϭ 5.65Ϫ5.85 (m, 1 H), 4.82Ϫ4.98 (m, 2 H),

.68Ϫ1.80 (m, 2 H), 1.34Ϫ1.47 (m, 2 H), 0.17 (s, 18 H), Ϫ0.10 (s,

H). Ϫ ¹³C NMR (62.9 MHz, CDCl

): δ ϭ 142.5, 141.0, 115.3,

8.4, 34.9, 16.8, Ϫ0.3, Ϫ1.1. Ϫ C16 (310.7): calcd. C 61.85,

of the allene 21c. Ϫ IR (film): ν˜ ϭ 2956, 1893, 1641, 1402, 1248,

2

Ϫ1

1

019, 905, 838, 760, 689, 620, 522 cm . Ϫ H NMR (250 MHz,

CDCl ): δ ϭ 5.81 (ddt, J ϭ 7, 10, 17 Hz, 1 H), 4.91Ϫ5.05 (m, 2

H), 1.43Ϫ2.14 (m, 6 H), 0.09 (s, 18 H), 0.05 (s, 9 H). Ϫ C NMR

250 MHz, CDCl ): δ ϭ 207.0, 138.9, 114.3, 79.0, 76.4, 33.8, 29.3,

7.1, 0.5, Ϫ1.2. Ϫ HRMS: m/z calcd. for C17 324.2125;

found 324.2124.

2

3

1

(

1

9

3

13

3

(

3

2

H36Si

3

H

34Si

3

H 11.03; found C 61.75, H 10.62.

1,6,7exo-Tris(trimethylsilyl)bicyclo[4.1.0]hept-3-ene (22): Thermo-

lysis of 17b (20 mg, 0.06 mmol) in toluene (2 mL) at 210 °C for 14

h gave, after purification by preparative gas chromatography, 17

mg (85%) of the bicycloheptene 22 as colorless plastic crystals. Ϫ

IR (KBr): ν˜ ϭ 3026, 2953, 2898, 1406, 1249, 1066, 945, 892, 836,

3-(5Ј-Pent-1Ј-enyl)-1,2,3-tris(trimethylsilyl)cyclopropene (17c): Ad-

dition of solution of 5-pent-1-enylmagnesium bromide

4.24 mmol) in Et O (3 mL) to cyclopropenylium salt 9 (250 mg,

O (7 mL) gave 70 mg (51%) of 17c. Ϫ IR (film):

a

(

2

0

.42 mmol) in Et

2

Ϫ1

1

7

55, 687, 637, 617, 440 cm . Ϫ H NMR (500 MHz, C

6 6

D ): δ ϭ

ν˜ ϭ 2956, 1700, 1684, 1653, 1558, 1540, 1506, 1457, 1248, 908, 838,

5

.46Ϫ5.48 (m, 2 H), 2.16Ϫ2.28 (m, 4 H), 0.60 (s, 1 H), 0.21 (s, 9

Ϫ1

1

7

1

53 cm . Ϫ H NMR (250 MHz, C

6

D

6

): δ ϭ 5.82 (ddt, J ϭ 8,

13

H), 0.19 (s, 18 H). Ϫ C NMR (125 MHz, C

8.9, 16.0, 2.2, 1.4. Ϫ HRMS: m/z calcd. for C16

found 310.1968. Ϫ C16 34Si (310.7): calcd. C 61.85, H 11.03;

found C 60.38, H 11.10.

6

D

6

): δ ϭ 122.6, 30.5,

0, 17 Hz, 1 H), 4.94Ϫ5.08 (m, 2 H), 1.95Ϫ2.07 (m, 2 H),

.47Ϫ1.56 (m, 2 H), 1.23Ϫ1.38 (m, 2 H), 0.21 (s, 18 H), 0.07 (s, 9

1

H34Si

3

310.1968;

H

3

13

H). Ϫ C NMR (62.9 MHz, C

6

D

6

): δ ϭ 142.0, 139.5, 114.8, 38.7,

3

1

5.0, 28.6, 17.7, Ϫ0.1, Ϫ0.8. Ϫ HRMS: m/z calcd. for C17

H36Si

3

24.2125; found 324.2124. Ϫ C17

H

36Si

3

(324.7): calcd. C 62.88, H 9,10-Dihydro-9,10-(bromodideuteriomethylethano)anthracene (25):

At 0 °C, a solution of triphenylphosphane (15.1 g, 58 mmol) in

1.17; found C 62.18, H 10.91.

[

24]

CH

2

Cl

2

(30 mL) was slowly added to a solution of alcohol 24

(22.9 g, 69 mmol) in CH Cl (170 mL).

General Procedure for the Photolyses of 10 and 17a؊c: Under ar-

gon, a degassed solution of 10 or one of 17a؊c (30 mg) in n-pent-

ane (30 mL) was irradiated with a low-pressure mercury UV lamp

(

9.3 g, 39 mmol) and CBr

4

2

After stirring overnight, the mixture was diluted with pentane until

a white precipitate separated. The mixture was filtered, and the

filtrate was concentrated to dryness. The crude product was puri-

fied by column chromatography (light petroleum ether, b.p. 40Ϫ80

(λ ϭ 254 nm) for 3 h in a ring reactor (quartz glass).

General Procedure for the Thermolyses of 10 and 17a؊c: A degassed

solution of 10 or 17b (20 mg) in toluene (2 mL) or of 17a or 17c

°

f

C) on 270 g of silica gel (R ϭ 0.20) to yield 10.0 g (85%) of

bromide 25 as a colorless solid; m.p. 126 °C. Ϫ IR (KBr): ν˜ ϭ

(20 mg) in cis-decalin (2 mL) was heated in an oil or sand bath until

3

7

021, 2958, 2940, 2859, 1457, 1284, 1173, 1020, 984, 961, 901, 795,

72, 743, 684, 609, 546 cm . Ϫ H NMR (250 MHz, CDCl ): δ ϭ

3

the starting material was no longer detectable by analytical GC.

Ϫ1

1

,1,3-Tris(trimethylsilyl)-1,2,5-hexatriene (21a): Irradiation of 17a

7.08Ϫ7.42 (m, 8 H), 4.48 (d, J ϭ 2.8 Hz, 1 H), 4.28 (t, J ϭ 2.8 Hz,

1 H), 2.29Ϫ2.40 (m, 1 H), 2.08 (ddd, J ϭ 2.8, 9.7, 12.5 Hz, 1 H),

1.19 (ddd, J ϭ 2.8, 4.9, 12.5 Hz, 1 H). Ϫ C NMR (125 MHz,

(30 mg, 0.10 mmol) in n-pentane (30 mL) for 2 h with a low-pres-

1

3

sure mercury lamp (λ ϭ 254 nm), evaporation of the solvent, and

purification of the residue by preparative gas chromatography gave

CDCl

3

): δ ϭ 143.5, 143.3, 143.1, 139.5, 126.2, 125.9, 125.79,

Eur. J. Org. Chem. 2001, 663Ϫ680

676

Cyclopropylidene Intermediates in Cyclopropene-to-Allene Rearrangements

FULL PAPER

1

2

25.78, 125.6, 123.8, 123.4, 123.2, 47.1, 44.2, 40.8, 37.4 (quint, J ϭ under reflux. This Grignard solution (3 mL, 2.63 mmol) was then

0 Hz), 34.7. Ϫ HRMS: m/z calcd. for C17

Br 300.0482; added to a suspension of the cyclopropenylium salt 9 (155 mg,

0.26 mmol) in Et

O (5 mL) at 0 °C. After stirring at room temp.

overnight, the reaction was quenched by the addition of saturated

aq. NH Cl solution (3 mL). The organic layer was washed with

O (3 mL) and brine (3 mL) and dried with MgSO . The solvent

13 2

H D

found 300.0482.

2

9

,10-[Dideuterio(methoxycarbonyl)methylethano]-9,10-dihydro-

4

anthracene (26): A Grignard solution was generated by adding

bromide 25 (5.12 g, 17 mmol) in Et O (30 mL) to Mg powder (0.46

g, 19 mmol) in Et O (10 mL) under gentle reflux. After a further

h, the solution was cooled to room temp. At 0 °C, this Grignard

solution (40 mL, 17 mmol) was then slowly added to a solution of

methyl chloroformate (3.2 g, 34 mmol) in Et O (20 mL). After stir-

H

2

4

2

was evaporated in vacuo. Purification of the residue by preparative

2

gas chromatography gave 38 mg (46%) of the cyclopropene 17b-D

2

1

as a colorless oil. Ϫ H NMR (250 MHz, CDCl

3

): δ ϭ 5.76 (dd,

J ϭ 10, 17 Hz, 1 H), 4.83Ϫ4.97 (m, 2 H), 1.49 (s, 2 H), 0.18 (s, 18

2

1

3

H), Ϫ0.09 (s, 9 H). Ϫ C NMR (125 MHz, CDCl ): δ ϭ 141.2,

ring for 0.5 h at room temp., the mixture was heated under reflux

for a further 5 h. The reaction was then quenched by the addition

1

39.8, 113.5, 37.7, 32.1 (quint, J ϭ 20 Hz), 16.6, Ϫ0.4, Ϫ0.7.

of saturated aq. NH

ganic layer was washed with H

dried with MgSO . The solvent was evaporated and the crude prod-

uct was purified by column chromatography [light petroleum ether

b.p. 40Ϫ80 °C)/Et O, 10:1] on 160 g of silica gel (R ϭ 0.15) to

yield 2.47 g (52%) of the methyl ester 26 as a colorless solid (m.p.

4

Cl solution (20 mL), and the separated or-

3,7endo-Dideuterio-1,6,7-tris(trimethylsilyl)bicyclo[4.1.0]hept-3-ene

(22-D): A degassed solution of 17b-D (30 mg, 0.10 mmol) in tolu-

ene (2 mL) was heated to 210 °C for 14 h under argon. After puri-

fication by preparative gas chromatography, 25 mg (83%) of bi-

cycloheptene 22-D was obtained as white plastic crystals. Ϫ IR

(KBr): ν˜ ϭ 2955, 1408, 1250, 1096, 1048, 911, 835, 756, 687, 636,

2

O (20 mL) and brine (20 mL) and

4

(

2

f

Ϫ1

1

21 °C). Ϫ IR (KBr): ν˜ ϭ 3021, 2947, 2866, 1734, 1458, 1434, 1377, 440 cm . Ϫ H NMR (250 MHz, C D ): δ ϭ 5.53 (s, 1 H), 2.23

1

6 6

263, 1224, 1018, 747, 619, 565 cm^Ϫ¹. Ϫ ¹H NMR (250 MHz,

): δ ϭ 7.08Ϫ7.42 (m, 8 H), 4.27 (d, J ϭ 2.8 Hz, 1 H), 4.21 δ ϭ 123.2, 30.27, 30.20, 18.50, 18.46, 2.2, 1.4. Ϫ HRMS: m/z calcd.

t, J ϭ 2.8 Hz, 1 H), 3.67 (s, 3 H), 2.35Ϫ2.48 (m, 1 H), 2.12 (ddd,

for C H D Si 312.2094; found 312.2094.

J ϭ 2.8, 9.8, 12.4 Hz, 1 H), 1.23 (ddd, J ϭ 2.8, 4.7, 12.4 Hz, 1 H).

13

(s, 4 H), 0.21 (s, 9 H), 0.19 (s, 18 H). Ϫ C NMR (75 MHz, C D ):

6

CDCl

3

(

1

6

32

2

3

13

Allylchlorodimethylsilane: To magnesium turnings (5.35 g,

.220 mol), which had been activated with a crystal of iodine, in

Ϫ

3

C NMR (125 MHz, CDCl ): δ ϭ 172.8, 143.62, 143.59, 142.9,

0

1

4

40.0, 125.9, 125.7, 125.53, 125.45, 125.3, 123.4, 123.3, 123.0, 51.2,

8.4, 43.9, 39.5 (quint, J ϭ 20 Hz), 34.5, 34.2. Ϫ HRMS: m/z calcd.

diethyl ether (30 mL) was added a solution of freshly distilled allyl

chloride (15.31 g, 0.200 mol) in diethyl ether (30 mL) at such a rate

that the temperature remained between Ϫ5 and 0 °C. The mixture

was stirred at ambient temperature for 12 h, then added in small

portions to a solution of freshly distilled dichlorodimethylsilane

for C19

16 2 2

H D O 280.1432; found 280.1432.

9

,10-[(1Ј,1Ј-Dideuterio-2Ј-hydroxyethyl)ethano]-9,10-dihydro-

anthracene (27): A mixture of the methyl ester 26 (2.28 g, 8.1 mmol)

and LiAlH

reflux for 12 h. After cooling to room temp., the reaction was

quenched by the addition of saturated aq. NH Cl solution (15 mL),

and the separated organic layer was dried with MgSO . Evapora-

4

(0.31 g, 8.1 mmol) in THF (25 mL) was heated under (36.2 mL, 0.300 mol). After stirring the reaction mixture at ambient

temperature for 18 h, it was filtered through Celite. The solvent

was distilled off through a 50-cm Vigreux column. Distillation of

the residue gave 11.72 g (44%) of allylchlorodimethylsilane (b.p.

4

1

tion of the solvent gave the alcohol 27 in quantitative yield (2.04

g) as a colorless solid; m.p. 114 °C. Ϫ IR (KBr): ν˜ ϭ 3235, 2932,

85Ϫ95 °C) as a colorless liquid sensitive to hydrolysis. Ϫ H NMR

3

(250 MHz, CDCl ): δ ϭ 5.79 (ddt, J ϭ 8.0, J ϭ 9.5, J ϭ 17.5 Hz,

3

Ϫ1

4

2

901, 2858, 1653, 1457, 1322, 1261, 1055, 1025, 799, 749, 550 cm

.

1 H, CH ϭCH), 4.93Ϫ5.02 (m, 2 H, CH ϭCH), 1.83 (ddd, J ϭ

2

1

4

3

Ϫ H NMR (250 MHz, CDCl

3

2 3 2

): δ ϭ 7.03Ϫ7.41 (m, 8 H), 4.28 (d, 1.0, J ϭ 1.0, J ϭ 8.0 Hz, 2 H, SiCH ), 0.42 [s, 6 H, Si(CH ) ]. Ϫ

1

3

J ϭ 2.6 Hz, 1 H), 4.15 (t, J ϭ 2.6 Hz, 1 H), 3.66 (m, 2 H),

C NMR (62.9 MHz, CDCl ): δ ϭ 132.0 (ϩ, CH ϭCH), 115.3

3

2

13

2

.02Ϫ2.16 (m, 2 H), 1.15Ϫ1.29 (m, 1 H). Ϫ C NMR (125 MHz, (Ϫ, CH ϭCH), 26.4 (Ϫ, SiCH ), 1.0 [ϩ, Si(CH ) ].

2 2 3 2

3

CDCl ): δ ϭ 144.3, 143.8, 143.3, 140.7, 125.8, 125.60, 125.57,

1

-(Allyldimethylsilyl)-3,3-dimethyl-1-cyclopropene (32a): To a solu-

tion of 1,1-dibromo-2-chloro-3,3-dimethylcyclopropane (30)

11.9 g, 45.4 mmol) in diethyl ether (60 mL) a 1.61  solution of

methyllithium in diethyl ether (62.0 mL, 99.9 mmol) was added

1

2

25.4, 125.3, 123.4, 123.3, 123.0, 61.0, 49.0, 44.3, 38.4 (quint, J ϭ

0 Hz), 34.7, 34.4. Ϫ MS: m/z ϭ 252, 221, 178.

(

2,2-Dideuteriobut-3-ene-1-ol (28): Flash vacuum pyrolysis of the al-

cohol 27 (1.86 g, 7.4 mmol) under argon at 350 °C gave butenol 28 dropwise at Ϫ78 °C. After 5 min at Ϫ78 °C, the mixture was al-

in quantitative yield (0.55 g) as a colorless liquid. Ϫ ¹H NMR

lowed to warm to room temp. and stirred for a further 1 h. It was

(

250 MHz, CDCl

3

): δ ϭ 5.80 (dd, J ϭ 10, 17 Hz, 1 H), 5.10Ϫ5.19

then cooled to Ϫ78 °C once more, whereupon allylchlorodimethyl-

m, 2 H), 3.68 (s, 2 H). Ϫ ¹³C NMR (62.9 MHz, CDCl

): δ ϭ silane^[²⁶^b^](5.09 g, 37.8 mmol) was added. After 5 min, the mixture

3

134.8, 117.6, 61.2.

was allowed to warm to room temp. once more. After 2 h, H O

2

(

80 mL) was added. The aqueous phase was extracted with diethyl

ether (3 ϫ 30 mL), and the combined organic phases were dried

MgSO ). The solvent was evaporated at Ϫ10 °C under reduced

4

-Bromo-3,3-dideuteriobut-1-ene (29): To a stirred solution of but-

enol 28 (549 mg, 7.41 mmol) in pyridine (174 mg, 2.20 mmol) at

°C was added PBr (812 mg, 3.00 mmol). After allowing the mix-

(

4

0

3

pressure (15 Torr), and the residue was subjected to bulb-to-bulb

distillation at room temp. (0.1 Torr) to yield 5.05 g (80%) of 32a

as a yellow liquid. For characterization, a fraction was isolated by

preparative-scale gas chromatography in Ͼ 99% purity. Ϫ IR

ture to warm to room temp., the homoallylic bromide 29 was dis-

tilled off into a cooled (Ϫ78 °C) receiving flask. Yield 657 mg

1

(65%). Compound 29 was used without further purification. Ϫ H

NMR (250 MHz, CDCl

.10Ϫ5.19 (m, 2 H), 3.41 (s, 2 H). Ϫ ¹C NMR (62.9 MHz, CDCl

δ ϭ 135.1, 117.5, 70.3.

3

): δ ϭ 5.80 (dd, J ϭ 10, 17 Hz, 1 H),

Ϫ1

1

(

neat): ν˜ ϭ 3079, 2961, 2929, 2857, 1669, 1630 cm . Ϫ H NMR

3

5

3

):

3

(

3

250 MHz, CDCl ): δ ϭ 7.96 (s, 1 H, 2-H), 5.80 (ddt, J ϭ 8.1,

3

4

2

J ϭ 10.3, J ϭ 16.8 Hz, 1 H, 2Ј-H), 4.90 (ddt, J ϭ 1.2, J ϭ 2.2,

3

4

2

3

-[4Ј-(3Ј,3Ј-Dideuterio)but-1Ј-enyl]-1,2,3-tris(trimethyl-

silyl)cyclopropene (17b-D): A Grignard solution was first generated

by adding bromide 29 (360 mg, 2.63 mmol) in Et O (2 mL) to Mg

O (1 mL) and heating for 1 h

J ϭ 16.8 Hz, 1 H, 3Ј-H), 4.88 (ddt, J ϭ 1.2, J ϭ 2.2, J ϭ

4

3

10.3 Hz, 1 H, 3Ј-H), 1.66 (ddd, J ϭ 1.2, J ϭ 1.2, J ϭ 8.1 Hz, 2

H, 1Ј-H), 1.14 [s, 6 H, C(CH

NMR (62.9 MHz, CDCl

1

3

2

3

)

2

], 0.16 [s, 6 H, Si(CH

3

)

2

]. Ϫ

C

powder (64 mg, 2.63 mmol) in Et

2

3

, DEPT): δ ϭ 138.1 (ϩ, C-2Ј*), 134.4 (ϩ,

Eur. J. Org. Chem. 2001, 663Ϫ680

677

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

3 2

C-2*), 132.9 (Cquat, C-1), 113.4 (Ϫ, C-3Ј), 28.9 [ϩ, C(CH )

], 23.5 to maintain a gentle reflux. The mixture was then heated under

FULL PAPER

(

Ϫ, SiCH

2 3 2

), 16.1 (Cquat, C-3), Ϫ3.1 [ϩ, Si(CH )

]. Ϫ MS (70 eV): reflux for a further 5 h. Subsequently, this solution was slowly ad-

ϩ

m/z (%) ϭ 151 (17) [M Ϫ CH

3

], 125 (24) [M Ϫ C

3

H

5

], 99 (10)

ded to a solution of dichloromethylsilane (6.30 mL, 52.3 mmol) at

ϩ

[Si(CH

3

)

2

C

3

H

5

], 98 (100) [Si(CH

3 2

)

C

3

H

4

]. Ϫ C10

H

18Si (166.3): room temp. The resulting mixture was stirred for 18 h at room

temp., then filtered through Celite, and the solvent was distilled off

through a 15-cm Vigreux column. The residue was purified by bulb-

calcd. C 72.21, H 10.91; found C 72.07, H 10.86.

2

-(Allyldimethylsilyl)-1-[dimethyl-(1E)-propenylsilyl]-3,3-dimethyl-

cyclopropene (33a): To a solution of diisopropylamine (0.51 mL,

.6 mmol) in THF (10 mL) a 2.36  solution of n-butyllithium in

to-bulb distillation to yield 1.19 g (31%) of but-3Ј-enylchlorodime-

thylsilane as a colorless liquid, which proved sensitive to hydrolysis.

The spectroscopic data are identical to those reported in the literat-

3

n-hexane (1.17 mL, 2.77 mmol) was added dropwise at Ϫ78 °C.

After 5 min at Ϫ78 °C, the reaction mixture was allowed to warm

to room temp. and stirred for a further 45 min. Then, a solution

of 32a (0.50 g, 3.00 mmol) in THF (2 mL) was added, the reaction

[26c]

ure.

1,2-Bis(3Ј-butenyldimethylsilyl)-3,3-dimethyl-1-cyclopropene (33b):

To a solution of diisopropylamine (0.47 mL, 3.3 mmol) in THF

(10 mL) a 2.36  solution of n-butyllithium in n-hexane (1.17 mL,

2.77 mmol) was added dropwise at Ϫ78 °C. After 5 min at Ϫ78 °C,

the reaction mixture was allowed to warm to room temp. and

stirred for a further 45 min. It was then cooled to Ϫ78 °C once

more, whereupon a solution of 32b (0.50 g, 2.77 mmol) in THF (2

mL) was added. The resulting mixture was stirred at room temp.

mixture was stirred at room temp. for 12 h, and subsequently added

dropwise to a solution of allylchlorodimethylsilane^[

26b]

(0.486 g,

3

.61 mmol) in THF (10 mL) at Ϫ78 °C. The resulting mixture was

stirred at Ϫ78 °C for 1 h and at room temp. for 2 h, then poured

into H O (30 mL). The aqueous phase was extracted with n-pent-

ane (3 ϫ 15 mL), the combined organic phases were dried

MgSO ), and the solvent was evaporated under reduced pressure.

2

(

4

for 12 h and then added dropwise at Ϫ78 °C to a solution of but-3Ј-

enylchlorodimethylsilane^[

27b]

(0.412 g, 2.77 mmol) in THF (10 mL).

The crude product (950 mg) was filtered through 20 g of silica gel

eluting with n-pentane and purified by preparative-scale gas chro-

matography to yield 358 mg (45%) of 33a as a colorless oil. Ϫ IR

The reaction mixture was stirred at Ϫ78 °C for 1 h and at room

temp. for 2 h, then poured into H O (30 mL). The aqueous phase

was extracted with n-pentane (3 ϫ 15 mL), the combined organic

2

Ϫ1

1

(

neat): ν˜ ϭ 2958 cm , 2914, 2852, 1697, 1621. Ϫ H NMR

250 MHz, CDCl

): δ ϭ 6.13 (dq, ³J ϭ 6.1, ³J ϭ 18.4 Hz, 1 H,

H, CH

3

phases were dried (MgSO ), and the solvent was evaporated under

ϭCH, SiCHϭCH), reduced pressure. The residue (560 mg) was filtered through 15 g

4

SiCHϭCH), 5.64Ϫ5.86 (m,

2

4

3

4

.83Ϫ4.91 (m, 2 H, CH ϭCH), 1.83 (dd, J ϭ 1.5, J ϭ 6.1 Hz, 3

2

of silica gel, eluting with n-pentane, and then purified by prepara-

4

3

H, CH

.10 [s, 6 H, C(CH

Si(CH

]. Ϫ ¹³C NMR (62.9 MHz, CDCl

quat, C-2**), 147.6 (Cquat, C-1**), 143.5 (ϩ, CH

ϩ, SiCHϭCH*), 129.0 (ϩ, SiCHϭCH), 113.1 (Ϫ, CH

3

), 1.64 (ddd, J ϭ 1.1, J ϭ 1.1, J ϭ 8.1 Hz, 2 H, SiCH

2

),

tive-scale gas chromatography to yield 325 mg (40%) of 33b as a

Ϫ1

1

3

)

2

], 0.19 [s, 6 H, Si(CH ) ], 0.15 [s, 6 H, colorless oil. Ϫ IR (neat): ν˜ ϭ 3077 cm , 2957, 2919, 1771, 1696,

3 2

1

3

)

2

3

, DEPT): δ ϭ 149.1 1640. Ϫ H NMR (250 MHz, CDCl ): δ ϭ 5.89 (ddt, J ϭ 6.4,

3

4

2

(

C

2

ϭCH*), 134.6

ϭCH),

), 18.3 (Cquat, C-

]. Ϫ MS (70 eV): m/z 0.73 (m , 4 H, 1Ј-H), 0.15 [s, 12 H, Si(CH ) ]. Ϫ C NMR

J ϭ 10.3, J ϭ 16.8 Hz, 2 H, 3Ј-H), 4.99 (ddt, J ϭ 1.3, J ϭ 1.5,

3

4

2

3

2

J ϭ 16.8 Hz, 2 H, 4Ј-H), 4.91 (ddt, J ϭ 1.3, J ϭ 1.5, J ϭ

2

3

9.0 [ϩ, C(CH

), Ϫ1.9 [ϩ, Si(CH

3

)

2

], 23.7 (Ϫ, SiCH

2

), 22.6 (ϩ, CH

3

c 3 2

10.3 Hz, 2 H, 4Ј-H), 2.08 (m , 4 H, 2Ј-H), 1.10 [s, 6 H, C(CH ) ],

1

3

)

2

], Ϫ2.9 [ϩ, Si(CH

3

)

2

c

3 2

ϩ

(%) ϭ 264 (2) [M ], 249 (2) [M Ϫ CH

3

], 223 (43) [M Ϫ C

3

H

5

], (62.9 MHz, CDCl , DEPT): δ ϭ 148.2 [C_q_u_a_t, C-1(2)], 141.6 (ϩ, C-

3

ϩ

1

81 (17) [M Ϫ C

6

H11]. Ϫ C15

H28Si

2

(264.6): calcd. 264.1729 (cor-

3Ј), 112.8 (Ϫ, C-4Ј), 29.1 [ϩ, C(CH ) ], 28.0 (Ϫ, C-2Ј), 17.9 (C_q_u_a_t,

3

2

rect HRMS).

C-3), 14.9 (Ϫ, C-1Ј), Ϫ2.4 [ϩ, Si(CH ) ]. Ϫ MS (70 eV): m/z (%) ϭ

3

2

ϩ

2

92 (Ͻ1) [M ], 277 (Ͻ1) [M Ϫ CH

3

3 2 4 7

], 113 (67) [Si(CH ) C H ].

1

-(But-3Ј-enyldimethylsilyl)-3,3-dimethyl-1-cyclopropene (32b): To a

solution of 1,1-dibromo-2-chloro-3,3-dimethylcyclopropane (30)

10.0 g, 38.1 mmol) in diethyl ether (100 mL) a 1.61  solution of

2

Ϫ C17H32Si (292.6): calcd. C 69.78, H 11.02; found C 69.88, H

1

0.98.

(

methyllithium in diethyl ether (52.0 mL, 83.8 mmol) was added

dropwise at Ϫ78 °C. After 5 min at Ϫ78 °C, the mixture was al-

lowed to warm to room temp. and stirred for a further 1 h. It was

General Procedure for the Photochemical Isomerization of Cyclopro-

penes: A solution of the appropriate cyclopropene (50Ϫ120 mg) in

n-pentane (free from unsaturated impurities; 1 L) was irradiated

with a low-pressure mercury lamp (λ ϭ 254 nm). The progress of

each reaction was monitored by gas chromatography. When the

starting material could no longer be detected, the solvent was evap-

orated under reduced pressure.

then cooled to Ϫ78 °C once more, whereupon but-3-enyl(methyl-

sulfonyloxy)dimethylsilane^[

27a]

(7.22 g, 34.6 mmol) was added.

After 5 min at Ϫ78 °C, the mixture was allowed to warm to room

temp., stirred for 2 h, and then treated with H O (100 mL). The

aqueous phase was extracted with diethyl ether (3 ϫ 30 mL), and

the combined ethereal extracts were dried (MgSO ). The solvent

2

1-(Allyldimethylsilyl)-1-[dimethyl-(1E)-propenylsilyl]-3-methyl-

4

was evaporated at Ϫ10 °C under reduced pressure (15 Torr) and

the residue was subjected to bulb-to-bulb distillation (0.1 Torr) to

yield 3.94 g (63%) of 32b as a yellow liquid. Ϫ IR (neat): ν˜ ϭ 2964

1,2-butadiene (35a): Photolysis of 33a (50 mg, 0.19 mmol) for

1

35 min gave 39 mg (78%) of 35a. Ϫ H NMR (250 MHz, CDCl

3

):

2

δ ϭ 6.00Ϫ6.18 (m, 1 H, SiCHϭCH), 5.60Ϫ5.88 (m, 2 H, CH ϭ

Ϫ1

1

4

cm , 2922, 2866, 1667. Ϫ H NMR (250 MHz, CDCl

3

): δ ϭ 7.92

2

CH, SiCHϭCH), 4.80Ϫ4.97 (m, 2 H, CH ϭCH), 1.86 (dd, J ϭ

3

4

(

s, 1 H, 2-H), 5.91 (m

, 2 H, 2Ј-H), 1.19 [s, 6 H, C(CH

.19 [s, 6 H, Si(CH

]. Ϫ ¹³C NMR (50.3 MHz, CDCl

δ ϭ 141.6 (ϩ, C-3Ј*), 137.6 (ϩ, C-2*), 133.3 (Ϫ, C-1), 112.8 (Ϫ, δ ϭ 212.6 (Cquat, C-2), 142.7 (ϩ, CH

C-4Ј), 28.9 [ϩ, C(CH ], 28.0 (Ϫ, C-2Ј), 16.8 (Ϫ, C-3), 14.8 (Ϫ,

CH*), 130.7 (ϩ, SiCHϭCH), 112.6 (Ϫ, CH

SiCH ), Ϫ2.6 [ϩ, Si(CH

]. Ϫ MS (70 eV): m/z (%) ϭ 180 (Ͻ 1) 3*), 79.8 (Cquat, C-1*), 23.0 (Ϫ, SiCH ), 22.6 (ϩ, CH

].

C(CH ], Ϫ1.3 [ϩ, Si(CH ], Ϫ1.7 [ϩ, Si(CH ].

c

, 1 H, 3Ј-H), 4.99Ϫ5.11 (m, 2 H, 4Ј-H), 2.15 1.5, J ϭ 6.1 Hz, 3 H, CH

3

), 1.65 [s, 6 H, C(CH

), 1.1, J ϭ 1.1, J ϭ 8.1 Hz, 2 H, SiCH ), 0.16 [s, 6 H, Si(CH

, APT): 0.08 [s, 6 H, SiC(CH , DEPT):

]. Ϫ ¹³C NMR (62.9 MHz, CDCl

ϭCH*), 135.3 (ϩ, SiCHϭ

ϭCH), 85.2 (Cquat, C-

3

)

2

], 1.59 (ddd, J ϭ

4

3

(m

c

3

)

2

], 0.74Ϫ0.89 (m, 2 H, SiCH

2

3 2

)

],

0

3

)

2

3

)

2

3

2

)

3 2

2

)

2

3

), 19.1 [ϩ,

2

3

ϩ

[

M ], 165 (14) [M Ϫ CH

3

)

3 2

3

)

2

3 2

)

But-3Ј-enylchlorodimethylsilane:^[²⁶^d^]To magnesium turnings

1,1-Bis(3Ј-butenyldimethylsilyl)-3-methyl-1,2-butadiene

(35b):

(

0.63 g, 26.0 mmol) in diethyl ether (10 mL) was added 4-chloro-1-

Photolysis of 33b (50 mg, 0.17 mmol) for 45 min yielded 45 mg

1

3

butene (2.36 g, 26 mmol) in diethyl ether (10 mL) at such a rate as

3

(90%) of 35b. Ϫ H NMR (250 MHz, CDCl ): δ ϭ 5.90 (ddt, J ϭ

678

Eur. J. Org. Chem. 2001, 663Ϫ680

Cyclopropylidene Intermediates in Cyclopropene-to-Allene Rearrangements

FULL PAPER

Pays-Bas 1986, 105, 462Ϫ464. Ϫ ^[R. Walsh, S. Untiedt, M.

3b]

3

4

2

6

1

0

(

3

.4, J ϭ 10.3, J ϭ 16.8 Hz, 2 H, 3Ј-H), 4.97 (ddt, J ϭ 1.3, J ϭ

3

4

2

3

Stohlmeier, A. de Meijere, Chem. Ber. 1989, 122, 637Ϫ642. Ϫ

.5, J ϭ 16.8 Hz, 2 H, 4Ј-H), 4.89 (ddt, J ϭ 1.3, J ϭ 1.5, J ϭ

0.3 Hz, 2 H, 4Ј-H), 2.02 (m , 4 H, 2Ј-H), 1.60 [s, 6 H, C(CH ],

.65 (m , 4 H, 1Ј-H), 0.08 [s, 12 H, Si(CH

]. Ϫ ¹³C NMR

, DEPT): δ ϭ 212.4 (Cquat, C-2), 141.9 (ϩ, C-

Ј), 112.6 (Ϫ, C-4Ј), 85.2 (Cquat, C-3*), 79.3 (Cquat, C-1*), 28.1 (Ϫ,

C-2Ј), 18.9 [ϩ, C(CH ], 15.9 (Ϫ, C-1Ј), Ϫ1.5 [ϩ, Si(CH ].

[3c]

R. Walsh, C. Wolf, S. Untiedt, A. de Meijere, J. Chem. Soc.,

)

2

[3d]

c

3

Chem. Commun. 1992, 421Ϫ422. Ϫ

R. Walsh, C. Wolf, S.

c

3

)

2

Untiedt, A. de Meijere, J. Chem. Soc., Chem. Commun. 1992,

[3e]

4

22Ϫ423. Ϫ

R. Walsh, S. Untiedt, A. de Meijere, Chem.

62.9 MHz, CDCl

3

Ber. 1994, 127, 237Ϫ245.

[

4]

U. Mißlitz, A. de Meijere, in: Carbene(oide), Carbine (Eds.: M.

Regitz, M. Hanack), Houben-Weyl, Thieme, Stuttgart, 1989,

vol. E 19b, p. 664 ff, and references cited therein.

3

)

2

3 2

)

Kinetics of the Rearrangement of Tetrakis(trimethylsilyl)cycloprop-

ene (10) in Solution: A solution of 10 (3 mg) in [D ]toluene (0.5 mL)

[5] [5a]

H. Hopf, A. Plagens, R. Walsh, J. Chem. Soc., Chem. Com-

[5b]

8

mun. 1994, 1467Ϫ1468. Ϫ

Liebigs Ann. 1996, 825Ϫ835. Ϫ

Schulenburg, R. Walsh, Angew. Chem. 1997, 109, 415Ϫ417;

H. Hopf, A. Plagens, R. Walsh,

[5c]

in an NMR tube was carefully degassed by three freeze-pump-thaw

cycles and then the tube was sealed. The tube was immersed in a

well-stirred silicone oil bath for the appropriate length of time and

then the reaction was quenched by immersing the tube in ice/water.

The temperature was monitored by means of a quartz thermo-

meter. The variation in the temperature during any run was always

less than 0.5 °C (Table 4). The ratio of allene 12 to cyclopropene

0 was estimated by H NMR spectroscopy by integrating the sig-

nals of the relevant methyl groups of these compounds. As an

internal standard, the CD H signal of [D ]toluene was used. The

average values of four integrations were used in the ensuing calcula-

tions.

H. Hopf, W. Graf v. d.

[

5d]

Angew. Chem. Int. Ed. Engl. 1997, 36, 381Ϫ383. Ϫ

Likhotvorik, D. W. Brown, M. Jones, Jr., J. Am. Chem. Soc.

I. R.

[5e]

1

994, 116, 6175Ϫ6178. Ϫ

W. Graf v. d. Schulenburg, H.

Hopf, R. Walsh, Chem. Eur. J. 2000, 6, 1963Ϫ1979.

Under flash-vacuum-pyrolysis conditions, 1-halo-, 1,2-dihalo-,

and 1,2,3,3-tetrahalocyclopropenes also cleanly rearrange to

give only allenes, cf.: W. E. Billups, R. E. Bachman, Tetrahed-

ron Lett. 1992, 33, 1825؊1826.

[7] [7a]

[6]

1

W. M. Jones, M. H. Grasley, W. S. Brey, J. Am. Chem. Soc.

[

7b]

2

7

1963, 85, 2754Ϫ2759. Ϫ

S. I. Kozhushkov, T. Haumann, R.

Boese, A. de Meijere, Angew. Chem. 1993, 105, 426Ϫ429; An-

gew. Chem. Int. Ed. Engl. 1993, 32, 401Ϫ403.

[

8]

9]

H. Hopf, in: The Chemistry of Ketenes, Allenes, and Related

Compounds (Ed.: S. Patai), part 2, Wiley, New York, 1980, p.

Table 4. Rearrangement of cyclopropene 10 to allene 12: propor-

tion of 10 (%) as a function of time t [min] and temperature T [°C]

7

79Ϫ901, and references cited therein.

C. W. Spangler, Chem. Rev. 1976, 76, 187Ϫ217.

[10]

6

A trimethylsilyl group is known to migrate 10 times faster

than a hydrogen atom in a [1,5]-sigmatropic rearrangement: A.

J. Ashe, J. Am. Chem. Soc. 1970, 92, 1233Ϫ1235.

T ϭ 200

T ϭ 215

T ϭ 230

T ϭ 252

t

10

t

10

t

10

t

10

[11]

By the principle of microscopic reversibility, a cyclopropylidene

should also be able to undergo a 1,2-shift (or 1,2-insertion) to

yield a cyclopropene. At least one such case has been reported:

M. S. Baird, C. M. Dale, J. R. Al Dulayymi, J. Chem. Soc.,

Perkin Trans. 1 1993, 1373Ϫ1374.

0

3

7

1

2

99.36

0

65.73

61.07

52.61

48.03

42.73

38.56

34.78

0

79.85

30

84.54

00

80

96.20

89.07

82.00

78.58

71.83

65.73

120

300

420

600

780

990

60

72.52

66.23

58.76

50.89

41.36

31.42

27.17

21.28

14.69

11.04

60

68.31

49.42

34.99

27.59

23.09

120

180

270

360

510

90

140

590

920

280

120

150

180

[

12]

T. J. Barton, J. Lin, S. Ijadi-Maghsoodi, M. D. Power, X.

Zhang, Z. Ma, H. Shimizu, M. S. Gordon, J. Am. Chem. Soc.

1

995, 117, 11695Ϫ11703.

5

6

7

8

70

60

50

70

13] [13a]

W. Weber, A. de Meijere, Angew. Chem. 1980, 92, 135Ϫ136;

[13b]

Angew. Chem. Int. Ed. Engl. 1980, 19, 138Ϫ140. Ϫ

Kostikov, A. de Meijere, J. Chem. Soc., Chem. Commun. 1984,

R. R.

[13c]

1

528Ϫ1529. Ϫ

D. L. Boger, C. E. Brotherton, J. Am. Chem.

[

13d]

Soc. 1986, 108, 6695Ϫ6713. Ϫ

M. S. Baird, S. R. Buxton,

[

13e]

J. S. Whitely, Tetrahedron Lett. 1984, 25, 1509Ϫ1512. Ϫ

H.

Tokuyama, T. Yamada, E. Nakamura, Synlett 1993, 589Ϫ591.

Ϫ ^[For a review on cyclopropene to vinylcarbene ring-open-

ing and intermolecular trapping, see ref.^[

13f]

Acknowledgments

This work was supported by the Niedersächsisches Ministerium für

Wissenschaft und Kunst, the Deutsche Forschungsgemeinschaft

4]

[14] [14a]

E. Colvin, Silicon in Organic Synthesis, Butterworths, Lon-

[14b]

don, 1981. Ϫ

Synthesis, Springer, Berlin, 1983.

15] [15a]

W. P. Weber, Silicon Reagents for Organic

(

SFB 357, Project A14), and the Fonds der Chemischen Industrie.

[

The research at the Technion was partially supported by the Fund

for Promotion of Research at the Technion, the German Federal

Ministry for Science, Research, Technology, and Education

BMBF), and the Minerva Foundation in Munich. T. M. is grateful

to the Fonds der Chemischen Industrie and the Deutsche For-

schungsgemeinschaft for scholarships. The authors are grateful to

Dr. N. Goldberg for providing a preprint of ref. , to Dr. P. R.

Schreiner for helpful discussions, and to Dr. B. Knieriem for his

careful reading of the manuscript and critical remarks.

L. Skattebøl, Chem. Ind. (London) 1962, 2146Ϫ2147; J.

[15b]

Org. Chem. 1966, 31, 2789Ϫ2794. Ϫ

Nilsen, F. Myhren, Acta Chem. Scand., Ser. B 1986, 40,

82Ϫ790.

L. Skattebøl, N. O.

7

(

[16]

S. Untiedt, Diplomarbeit, Universität Hamburg, 1990.

[17] K. Sakamoto, T. Saeki, H. Sakurai, Chem. Lett. 1993,

1675Ϫ1678.

[

18] [18a]

[

35]

G. K. S. Prakash, S. Quaiser, H. A. Buchholz, J. Casanova,

[18b]

G. A. Olah, Synlett 1994, 113Ϫ114. Ϫ

H. A. Buchholz, D. Deffieux, G. A. Olah, J. Org. Chem. 1994,

9, 7532Ϫ7533.

G. K. S. Prakash,

5

[

19] [19a]

G. Maier, D. Volz, J. Neudert, Synthesis 1992, 561Ϫ564.

Ϫ ^[

19b]

A. de Meijere, D. Faber, M. Noltemeyer, R. Boese, T.

[

1] [1a]

J. J. Gajewski, Hydrocarbon Thermal Isomerization, Aca-

[1b]

Haumann, T. Müller, M. Bendikov, E. Matzner, Y. Apeloig, J.

Org. Chem. 1996, 61, 8564Ϫ8568.

demic Press, New York, 1981. Ϫ

M. S. Baird, in: Car-

bocyclic Three- and Four-Membered Ring Compounds (Ed.: A.

de Meijere), Houben-Weyl, Thieme, Stuttgart, 1997, vol. E 17d,

pp. 2781Ϫ2790.

[

20]

21]

L. Rösch, G. Altnau, J. Organomet. Chem. 1980, 195, 47Ϫ53.

B. M. Trost, J. Yoshida, M. Lautens, J. Am. Chem. Soc. 1983,

105, 4494Ϫ4496.

R. West, P. C. Jones, J. Am. Chem. Soc. 1969, 91, 6156Ϫ6161.

Because of the low vapor pressure of tetrakis(trimethylsilyl)cy-

clopropene (10), the kinetics of its rearrangement could not be

determined with a higher degree of accuracy in the gas phase.

Apart from the larger error limits, however, the solution phase

should not significantly affect the energetics of the thermal re-

2] [2a]

I. M. Bailey, R. Walsh, J. Chem. Soc., Faraday Trans. 1978,

[2b]

[22]

[23]

7

4, 1146Ϫ1158. Ϫ

Chem. Soc. 1980, 102, 1210Ϫ1212. Ϫ

holz, R. Walsh, Chem. Ber. 1985, 118, 3579Ϫ3587. Ϫ

Hopf, G. Wachholz, R. Walsh, J. Chem. Soc., Perkin Trans. 2

1986, 1103Ϫ1106.

3] [3a]

H. Hopf, H. Priebe, R. Walsh, J. Am.

[2c]

H. Hopf, G. Wach-

[2d]

H.

[

M. A. Kirms, H. Primke, A. de Meijere, Recl. Trav. Chim.

Eur. J. Org. Chem. 2001, 663Ϫ680

679

A. de Meijere, D. Faber, U. Heinecke, R. Walsh, T. Müller, Y. Apeloig

FULL PAPER

[

30]

31]

arrangement of a non-polar compound such as 10, cf.: D.

Kaufmann, H.-H. Fick, O. Schallner, W. Spielmann, L.-U.

Meyer, P. Gölitz, A. de Meijere, Chem. Ber. 1983, 116,

H. F. Bettinger, P. R. Schreiner, P. v. R. Schleyer, H. F. Schaefer

III, J. Phys. Chem. 1996, 100, 16147Ϫ16154.

J. A. Pople, M. Head-Gordon, K. Raghavachari, J. Chem.

Phys. 1987, 87, 5968Ϫ5975.

587؊609.

[

24]

The procedures for the first two steps in the synthesis of 29

[32]

[33]

T. J. Lee, A. P. Rendell, P. R. Taylor, J. Phys. Chem. 1990,

were adapted from a published synthesis of 3,3-dideuterioprop-

9

4, 5463Ϫ5468.

[

24a]

1

-en-3-ol:

P. D. Bartlett, F. A. Tate, J. Am. Chem. Soc.

M. Yoshimine, J. Pacansky, N. Honjou, J. Am. Chem. Soc.

[24b]

1

953, 75, 91Ϫ95. Ϫ

J. E. Baldwin, M. Bradley, N. J. Turner,

1

989, 111, 4198Ϫ4209.

R. M. Adlington, A. R. Pitt, H. Sheridan, Tetrahedron 1991,

[

34]

35]

H. Shimizu, M. S. Gordon, Organometallics 1995, 14,

3827Ϫ3831.

N. Goldberg, W. Graf v. d. Schulenburg, Chem. Commun.

1998, 2761Ϫ2762.

4

7, 8203Ϫ8222.

[

25]

M. S. Baird, H. H. Hussain, W. Nethercott, J. Chem. Soc.,

Perkin Trans. 1 1986, 1845Ϫ1854.

26] [26a]

Allylchlorodimethylsilane has been reported previously,

36] [36a]

[36b]

J. B. Lambert, Tetrahedron 1990, 46, 2677؊2689. ؊

Y.

but without spectroscopic characterization: Z. Y. Wei, D.

Apeloig, in: The Chemistry of Organic Silicon Compounds, vol.

1 (Eds.: S. Patai, Z. Rappoport), Wiley, Chichester, 1989, p.

Wang, J. S. Li, T. H. Chan, J. Org. Chem. 1989, 54, 5768Ϫ5774.

[

26b]

Ϫ

For the preparation of allylchlorodimethylsilane from

[36c]

5

7Ϫ225. Ϫ

H.-U. Siehl, T. Müller, in: The Chemistry of

dichlorodimethylsilane, we preferred to use allylmagnesium

chloride (K. V. Baker, J. M. Brown, N. Hughes, A. J. Skarnulis,

A. Sexton, J. Org. Chem. 1991, 56, 698؊703) rather than the

Organic Silicon Compounds, vol. 2 (Eds.: Z. Rappoport, Y. Ape-

[36d]

loig) Wiley, Chichester, 1998, p. 595. Ϫ

J. B. Lambert, Y.

[

26c]

Zhao, R. W. Emblidge, L. A. Salvador, X. Liu, J.-H. So, E. C.

bromide. ؊

R. Calas, P. Bourgeois, J. Dunogues, F. Pisci-

Chelius, Acc. Chem. Res. 1999, 32, 183Ϫ190.

otti, B. Arreguy, Bull. Soc. Chim. Fr. 1974, 11, 2556Ϫ2558. Ϫ

[

26d]

[37]

But-3Ј-enylchlorodimethylsilane could not be obtained

from but-3Ј-enyl(methylsulfonyloxy)dimethylsilane in reason-

Y. Apeloig, T. Müller, in: Dicoordinated Carbocations (Eds.: Z.

Rappoport, P. J. Stang), Wiley Interscience, New York, 1997,

chapter 2, p. 9 ff.

X. Li, J. A. Stone, J. Am. Chem. Soc. 1989, 111, 5586Ϫ5592.

See also: W. Zhang, J. A. Stone, M. A. Brook, G. A. McGib-

bon, J. Am. Chem. Soc. 1996, 118, 5764Ϫ5771.

[

26c]

able yield according to ref.

and was therefore prepared from

[

38]

but-3-enylmagnesium chloride and dichlorodimethylsilane in

close analogy to the preparation of allylchlorodimethylsilane.

27] [27a]

[

H. M. Frey, R. G. Hopkins, L. Skattebøl, J. Chem. Soc. B

[

27b]

1971, 539. Ϫ

W. R. Roth, D. Wollweber, R. Offerhaus, V.

[39] [39a]

A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988,

Rekowski, H.-W. Lennartz, R. Sustmann, W. Müller, Chem.

Ber. 1993, 126, 2701.

[39b]

8

8, 899Ϫ926. Ϫ

A. E. Reed, R. B. Weinstock, F. Weinhold,

[39c]

J. Chem. Phys. 1985, 83, 735Ϫ746. Ϫ

E. D. Glending, A.

28]

M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G.

Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A.

Petersson, J. A. Montgomery, K. Raghavachari, M. A. Al-La-

ham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslow-

ski, B. B. Stefanov, A. Nanayakkara, M. Challacombe, C. Y.

Peng, P. Y. Ayala, W. Chen, M. W. Wong, J. L. Andres, E. S.

Replogle, R. Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley,

J. S. Defrees, J. Baker, J. P. Stewart, M. Head-Gordon, C. Gon-

zalez, J. A. Pople, Gaussian 94, Revisions C.2ϪE.2, Gaussian,

Inc., Pittsburgh PA, 1995.

29] [29a]

E. Reed, J. E. Carpenter, F. Weinhold, NBO version 3.1.

[40]

[41]

For a review on a-silyl-substituted carbenes see: G. Maas, in:

The Chemistry of Organic Silicon Compounds, vol. 2 (Eds.: Z.

Rappoport, Y. Apeloig), Wiley, Chichester, 1998, p. 703.

J. Slutsky, H. Kwart, J. Am. Chem. Soc. 1973, 95, 8678Ϫ8685.

W. Adcock, C. I. Clark, C. H. Schiesser, J. Am. Chem. Soc.

[

42]

1

996, 118, 11541Ϫ11547.

[

43]

The influence of entropy was calculated by comparing the rel-

ative electronic free energy G obtained at the B3LYP/6-311G*

level for 0 K with the sum of the thermal and electronic free

energy ∆G, calculated at 298 K.

[

R. G. Parr, W. Yang, Density-Functional Theory of Atoms

and Molecules, Oxford University Press, New York, 1989. Ϫ

[

29b]

[44]

C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785.

This is the change in the relative energy between TS4a(Si)/50a-

Ϫ ^[

29c]

B. Miehlich, A. Savin, H. Stoll, H. Preuss, Chem. Phys.

Si and TS4b(Si)/50b-Si.

Received July 14, 2000

[O00358]

[

29d]

Lett. 1989, 157, 200. Ϫ

A. D. Becke, J. Chem. Phys. 1993,

98, 5648Ϫ5652.

680

Eur. J. Org. Chem. 2001, 663Ϫ680

Article Doi

On the question of cyclopropylidene intermediates in cyclopropene-to-allene rearrangements - Tetrakis(trimethylsilyl)cyclopropene, 3-alkenyl-1,2,3-tris(trimethylsilyl)cyclopropenes, and related model compounds

DOI: 10.1002/1099-0690(200102)2001:4<663::AID-EJOC663>3.0.CO;2-E

Source and publish data:

Authors:

Article abstract of DOI:10.1002/1099-0690(200102)2001:4<663::AID-EJOC663>3.0.CO;2-E

Full text of DOI:10.1002/1099-0690(200102)2001:4<663::AID-EJOC663>3.0.CO;2-E

Products guided by the article

R&D Labs maybe for 3721-17-3

Relevant to this article

Hot Product