Intramolecular reactions of metal carbenoids with allylic ethers: Is a free ylide involved in every case?

Article abstract of DOI:10.1002/chem.201304054

Rhodium-, copper- and iridium-catalyzed reactions of the ¹³C-labelled diazo carbonyl substrates 18* and 19* were performed. Results obtained from copper- and iridium-catalyzed reactions of the ¹³C-labelled α-diazo β-keto ester 19* indicate that either or both of these reactions do not proceed via a free oxonium ylide but instead follow a competing non-ylide route that delivers apparent [2,3]-sigmatropic rearrangement products. In the case of the iridium-catalyzed reaction of α-diazo β-keto ester 19*, results obtained from crossover experiments indicate that the initially formed metal-bound ylide dissociates to give an iridium enolate and an allyl cation, which recombine to form the C-C bond. Differing distributions of ¹³C-labelled rearrangement products in the rhodium-, copper- and iridium-catalyzed reactions of a ¹³C-labelled α-diazo β-keto ester (see scheme, DCE=dichloroethane, acac=acetylacetonate, COD=1,5-cyclooctadiene) revealed that the copper- and iridium-mediated reactions do not proceed via a common free oxonium ylide intermediate.

Full text of DOI:10.1002/chem.201304054

DOI: 10.1002/chem.201304054
Full Paper
&
Rearrangement
Intramolecular Reactions of Metal Carbenoids with Allylic Ethers:
Is a Free Ylide Involved in Every Case?
J. Stephen Clark* and K. Emelie Hansen[a]
Abstract: Rhodium-, copper- and iridium-catalyzed reactions
of the 13C-labelled diazo carbonyl substrates 18* and 19*
were performed. Results obtained from copper- and iridium-
catalyzed reactions of the 13C-labelled a-diazo b-keto ester
19* indicate that either or both of these reactions do not
proceed via a free oxonium ylide but instead follow a com-
peting non-ylide route that delivers apparent [2,3]-sigma-
tropic rearrangement products. In the case of the iridium-
catalyzed reaction of a-diazo b-keto ester 19*, results ob-
tained from crossover experiments indicate that the initially
formed metal-bound ylide dissociates to give an iridium
enolate and an allyl cation, which recombine to form the
CÀC bond.
Introduction
The intramolecular reaction of an allylic ether with a metal car-
benoid and rearrangement of the resulting free oxonium ylide
or metal-bound equivalent offers a potentially powerful cata-
lytic approach to the stereoselective construction of cyclic
ethers and carbocycles.[1,2] Since the publication of seminal
work by of Pirrung and Werner[3a] and Johnson and Roskamp[3b]
in 1986, there have been numerous applications of this reac-
tion to the synthesis of functionalised cyclic ethers and carbo-
cycles.[4–6] Asymmetric variants of the reaction mediated by
chiral catalysts have also been developed, but levels of asym-
metric induction are frequently modest.[7]
In the originally proposed and widely assumed reaction
mechanism, an electrophilic metal carbenoid 2, generated by
reaction of a metal complex with the diazo substrate 1, under-
goes nucleophilic attack by one of the ether oxygen lone pairs
to give a transient metal-bound intermediate 3 (Scheme 1).
Dissociation of the metal complex delivers a highly reactive
oxonium ylide 4, which undergoes a symmetry-allowed [2,3]-
sigmatropic rearrangement to deliver the observed cyclic ether
product(s) 5. Although the products obtained from the reac-
tions are generally consistent with the intermediacy of the
oxonium ylide 4, the possibility of direct rearrangement of the
metal-bound ylide 3 to produce the cyclic ether product(s) 5
cannot be ruled out. The fact that experimental and theoretical
data for analogous intramolecular reactions of carbonyl groups
with electrophilic metal carbenoids show that metal-bound
Scheme 1. Intramolecular reaction of an allylic ether with a metal carbenoid.
carbonyl ylides react directly with dipolarophiles suggests that
direct rearrangement of the metal-bound ylide 3 might be fea-
sible.[8] However, it is extremely difficult to tease apart the pre-
cise mechanistic details, because the overall process involves
four separate reactions and proceeds via three highly reactive
and short-lived intermediates (2–4) that are present in very
low concentrations, precluding spectroscopic detection.
The task of establishing the precise reaction mechanism is
also complicated by the possibility that involvement of a free
oxonium ylide is dependent on the presence of stabilizing
groups on the diazo carbonyl substrate (e.g., ketone, ester, di-
ketone, or keto ester), the nature of the catalyst used to gener-
ate the carbenoid and solvent polarity. Indeed, it is conceivable
that the course of the reaction could be highly dependent on
all of these factors and that the reaction proceeds via a free
oxonium ylide in some cases and via a metal-bound ylide in
others. The intermediacy of a free oxonium ylide or a metal-
bound ylide is more than just a question of semantics. If one is
to engage in the rational design of new ligands and catalysts
[a] Prof. Dr. J. S. Clark, Dr. K. E. Hansen
WestCHEM, School of Chemistry
Joseph Black Building, University of Glasgow
University Avenue, Glasgow G128QQ (UK)
Fax: (+44)141-330-6867
E-mail: stephen.clark@glasgow.ac.uk
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304054.
Chem. Eur. J. 2014, 20, 1 – 7
1
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
to improve ratios of diastereomeric products or to render the
reaction highly enantioselective, a full understanding of the
role of the metal complex in the rearrangement reaction is es-
sential.
isomer ratio can be controlled by choice of the catalyst sug-
gests that a free oxonium ylide is not an intermediate in the
reaction.
To account for the catalyst-dependent stereochemical out-
come of the above reactions [Eqs. (1) and (2)] while maintain-
ing the intermediacy of a free oxonium ylide, it would be nec-
essary for one of the diastereotopic lone pairs to undergo se-
lective nucleophilic attack on the electrophilic metal carbenoid
and for the identity of the metal complex to influence the
choice of lone pair directly in each case. Furthermore, the free
ylide mechanism would require the stereogenic centre (*) cre-
ated at the oxonium site to be preserved to a large extent
during the lifetimes of intermediates 3 and 4 (Scheme 1).
The catalyst dependence of the product distribution from
the intramolecular reaction of an allylic ether with a metal car-
benoid has been documented by others. For example, West
and co-workers observed that the outcome of the metal-medi-
ated reaction of diazo ketone 11 is highly dependent on the
catalyst employed.[5e] The reaction promoted by rhodium(II) tri-
phenylacetate produced a 1:5 mixture of the diastereomeric bi-
cyclic ethers 12a and 12b, whereas the reaction mediated by
copper(II) trifluoroacetylacetonate delivered a 30:1 mixture of
diastereomers with compound 12a predominating [Eq. (3)].
West and co-workers have also published several examples of
reactions in which an ylide or metal-bound ylide generated
from a metal carbenoid undergoes apparent [2,3] or [1,2] rear-
rangement and the stereochemical outcome is influenced by
the structure of the catalyst.[5]
In recent years, the conventional view that the intramolecu-
lar reaction of an allylic ether with a metal carbenoid involves
a short-lived oxonium ylide has been called into question by
observations that are difficult to rationalise mechanistical-
ly.[2h,4,5] In the course of detailed studies on the reaction, we
have shown that the nature of the catalyst (both metal and
ligand) has a profound influence on the yield and stereochemi-
cal outcome of individual reactions.[4–6] For example, we ob-
served that reaction of diazo ketone 6 with rhodium(II) acetate
afforded a mixture of cis- and trans-2,5-dialkyldihydro-2H-
pyran-3(4H)-ones 7a and 7b in 21:79 ratio [Eq. (1)].[4f] In con-
trast, copper-catalyzed reactions produced mixtures of isomers
in which the isomer 7a was found to be the major product;
the product ratio was also shown to be highly ligand depen-
dent. Intriguingly, the reaction mediated by [Cu(hfacac)2]
(hfacac=hexafluoroacetylacetonate) produced the by-product
8 (13% yield). This observation is significant because the ether
8 cannot have arisen from a free oxonium ylide; otherwise, it
would have been isolated from reactions performed with the
other catalysts and in the same relative amount. In addition, it
is highly unlikely that compound 8 could have been produced
by concerted migration of the allyl group in the free oxonium
ylide because of geometrical constraints in the requisite cyclic
transition state.
A similar dependence of diastereoselectivity on catalyst
structure has been observed in reactions involving carbenoids
generated from a-diazo b-keto esters. For example, Yakura and
co-workers reported that the reaction of the substrate 13 with
rhodium(II) triphenylacetate produces a mixture of the diaste-
reomeric cyclic ethers 14a and 14b in 91:9 ratio [Eq. (4)].[9]
When the catalyst ligands were changed from triphenylacetate
to acetate, the diastereoselectivity of the reaction was reduced
substantially.
Further recent results from our own work also illustrate the
highly catalyst dependent nature of the reaction. Treatment of
diazo ketone 9 with a suitable catalyst delivers a mixture of
the isomeric bridged bicyclic ethers (Z)-10 and (E)-10, and the
stereochemical outcome of reaction can be tuned by catalyst
selection [Eq. (2), TBS=tert-butyldimethylsilyl].[4n] When rho-
dium(II) triphenylacetate was employed as the catalyst, the less
thermodynamically stable isomer (E)-10 was the major product,
but when the reaction was performed with [Cu(hfacac)2] as the
catalyst, (Z)-10 was obtained as the major product. Density
functional calculations of the energetics of rearrangement of
the free oxonium ylide(s) indicated that the reaction should
proceed with low E/Z selectivity.[4p] The fact that the rearrange-
ment reaction proceeds diastereoselectively and that the
Hashimoto and co-workers have demonstrated that asym-
metric reactions of the a-diazo b-keto esters 15a–c mediated
by the chiral rhodium complex [Rh2(S-PTTL)4] deliver significant
amounts of apparent [1,2]-rearrangement products 17a–c in
addition to the expected apparent [2,3]-rearrangement prod-
&
&
Chem. Eur. J. 2014, 20, 1 – 7
www.chemeurj.org
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
ucts 16a–c [Eq. (5)].[7i] The formation of apparent [1,2]-shift
products in competition with those arising from a lower-
energy symmetry-allowed [2,3]-rearrangement reaction and
the dependence of product distribution on the substituents
adorning the allylic ether is intriguing. Hashimoto and co-work-
ers did not investigate whether the ratio of the products 16
and 17 is dependent on the catalyst used for carbenoid gener-
ation, the solvent or the conditions under which the reaction
is performed.
were selected because their rhodium-catalyzed reactions are
well documented and the products only contain aliphatic hy-
drogen atoms in the migrated group, making NMR analysis
easier. In their seminal publication, Pirrung and Werner report-
ed that diazo substrates 18 and 19 underwent rhodium(II) ace-
tate catalyzed cyclization in dichloromethane at room temper-
ature to afford benzofuranones 20 and 21 in good yield.[3a]
We first conducted a brief survey of copper complexes as
catalysts for the cyclization of substrates 18 and 19; we also
explored the use of [{Ir(cod)Cl}2] (cod=1,5-cyclooctadiene) as
a potential catalyst for carbenoid generation [Eq. (6), Table 1].
Although it is known that diazo carbonyl compounds react
with iridium complexes and that iridium carbenoids can be
generated from sulfur ylides,[12] we are not aware of any exam-
ples of direct generation of an iridium carbenoid from a diazo
carbonyl compound followed by reaction of the presumed iri-
dium carbenoid with an allylic ether.
Important data regarding the reaction mechanism have also
been disclosed by Doyle and co-workers.[6f,7f,10] In early studies,
they reported highly enantioselective intermolecular reactions
of simple allyl methyl ethers with carbenoids generated by the
reaction of ethyl diazoacetate and chiral rhodium complexes.[7f]
Direct rearrangement from a metal-bound ylide was proposed
to explain the very high levels of asymmetric induction
(ꢀ98% ee) obtained from these reactions, and this mechanism
was supported by the finding that allylic iodides underwent
analogous asymmetric reactions, albeit with lower ee values.
More recently, Doyle and co-workers have studied the intramo-
lecular reactions of cyclic ethers with rhodium carbenoids de-
rived from a-diazo b-keto esters to give conformationally re-
stricted bicyclic reactive intermediates that undergo apparent
[1,2] or [2,3] rearrangement.[6f,10] On the basis of the invariance
of both the isomer ratios and the ratios of the apparent [1,2]-
and [2,3]-rearrangement products with catalyst, Doyle and co-
workers concluded that these reactions occur through a free
oxonium ylide, a finding that contradicts conclusions based on
their earlier work.
Table 1. Copper- and iridium-catalyzed reactions of the diazo ketone 18
and a-diazo b-keto ester 19 to give benzofuranones 20 and 21.
Entry
Substrate
Catalyst ([mol%])
Solvent
T
Yield[a] [%]
1
2
3
4
18
18
19
19
[Cu(acac)2] (5)
[{Ir(cod)Cl}2] (1)
[Cu(acac)2] (10)
[{Ir(cod)Cl}2] (1)
CH2Cl2
CH2Cl2
DCE[b]
DCE[b]
reflux
reflux
reflux
reflux
57
9
61
39
[a] Yield after purification by chromatography. [b] 1,2-Dichloroethane.
The results of preliminary studies showed that the a-diazo
carbonyl substrates 18 and 19 react with [Cu(acac)2] (acac=
acetylacetonate) to deliver the expected benzofuranone prod-
ucts 20 and 21 in reasonable yield (Table 1, entries 1 and 3).
The iridium-catalyzed reactions (Table 1, entries 2 and 4) were
much lower yielding than the copper-catalyzed reactions or
the previously reported rhodium-catalyzed reactions, but in
the case of substrate 19 a reasonable yield of the expected re-
arrangement product 21 was obtained.
To probe the reaction mechanism, we employed substrates
18* and 19*, which were 13C-labelled at the terminal position
of the allylic ether. The use of these 13C-labelled substrates was
expected to allow us to detect the formation of apparent [1,2]-
shift products analogous to those (17a–c) obtained by Hashi-
moto and co-workers during the rearrangement of a-diazo
b-keto esters 15a–c,[7h] but without the presence of substitu-
ents that might bias the reaction as a consequence of elec-
tronic or steric factors. In conducting these experiments, we
wished to determine whether the relative amounts of the ap-
parent [2,3]- and [1,2]-rearrangement products were depen-
dent on the catalyst used for carbenoid generation. If all reac-
Finally, evidence for the potential involvement of a metal-as-
sociated intermediate in related carbenoid reactions comes
from recently published work of Davies and co-workers.[11]
They have demonstrated that it is possible to perform analo-
gous enantioselective rearrangement reactions by using chiral
rhodium donor/acceptor carbenoids on allylic and propargylic
alcohols rather than allylic ethers. In these cases, it is difficult
to account for the very high levels of asymmetric induction
that are obtained without the involvement of the chiral metal
complex in the CÀC bond-forming step.
Results and Discussion
We attempted to establish the intermediacy of a free oxonium
ylide or a metal-bound ylide (analogous to 3 in Scheme 1) in
the key rearrangement reaction by investigating the metal-
mediated reactions of simple substrates 18 and 19 to give
benzofuranones 20 and 21 [Eq. (6)]. Substrates 18 and 19
Chem. Eur. J. 2014, 20, 1 – 7
www.chemeurj.org
3
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
tions proceed exclusively through the same metal-free oxoni-
um ylide, the ratio of apparent [2,3]- and [1,2]-rearrangement
products should be fixed and independent of the catalyst used
to generate the oxonium ylide. Thus, any variations in product
distribution with catalyst would indicate that some or all of
these reactions proceed through a sequence in which rear-
rangement with concomitant CÀC bond formation takes place
directly from a metal-associated ylide rather than a free oxoni-
um ylide [Eq. (7)].
scrambling of the label as a consequence of formation of
a symmetrical p-allyl iridium complex.
Attention was then focussed on reactions of the labelled
a-diazo b-keto ester 19* [Eq. (8), Table 3]. The results obtained
Table 3. Metal-catalyzed reactions of labelled a-diazo ketone 19*.
Entry
Catalyst
Yield [%]
Ratio 21*i:21*t
1H NMR
Ratio 21*i:21*t
13C NMR
1
2
3
4
5
6
7
Rh2(OAc)4
76
20
34
43
75
63
37
>99:1
96:4
95:5
84:16
86:14
88:12
44:56
98:2
97:3
96:4
85:15
87:13
90:10
47:53
Rh2(O2CCF3)4
Rh2(O2CCPh3)4
[Cu(acac)2]
[Cu(hfacac)2]
Cu[(MeCN)4]PF6
[{Ir(cod)Cl}2]
Reactions of substrate 18*, which has a 13C label at the ter-
minus of the allyl group, to give products with the label at
either the internal position (20*i) or the terminal position
(20*t) were explored [Eq. (7)]. The relative amounts of 20*i
with this substrate were more conclusive than those obtained
from the reactions of the a-diazo ketone 18*. When a rhodium
complex was employed as the catalyst, a very small but detect-
able quantity of the compound 21*t containing the label at
the terminal position was obtained in addition to the expected
product 21*i, in which the label is located at the internal posi-
tion of the allyl chain (Table 3, entries 1–3). The ratio of labelled
compounds 21*i:21*t was found to be dependent on the
nature of the rhodium complex employed as catalyst; of the
three catalysts studied, rhodium(II) triphenylacetate gave the
highest proportion of compound 21*t. A much higher propor-
tion of 21*t was obtained from the copper-catalyzed reactions
than from the rhodium-catalyzed reactions, and again the
21*i:21*t ratio was found to be dependent on the specific
copper complex used (Table 3, entries 4–6). Finally, the use of
[{Ir(cod)Cl}2] as catalyst produced almost equal amounts of the
labelled products 21*i and 21*t, indicating that complete
scrambling had occurred (Table 3, entry 7).
1
and 20*t were determined by H and 13C NMR spectroscopy,
and there was very close agreement between the ratios mea-
sured with each technique. Analysis of the product distribution
in each case showed that a small but discernible amount of
ketone 20*t containing the 13C label at the terminal position of
the alkene was generated alongside the expected formal [2,3]-
rearrangement product 20*i, in which the label is positioned
at the internal (allylic) position. There were small variations in
product distribution in the case of rhodium- or copper-cata-
lyzed reactions (Table 2, entries 1–4), but these variations are
Table 2. Metal-catalyzed reactions of labelled a-diazo ketone 18*.
Entry
Catalyst
Yield [%]
Ratio 20*i:20*t
1H NMR
Ratio 20*i:20*t
13C NMR
1
2
3
4
5
Rh2(OAc)4
26
62
25
46
>99:1
98:2
95:5
98:2
52:48
98:2
98:2
97:3
99:1
53:47[b]
Rh2(O2CCF3)4
Rh2(O2CCPh3)4
[Cu(hfacac)2]
[{Ir(cod)Cl}2]
[a]
–
[a] Very small amounts of the product were obtained. [b] 13C NMR analysis
was performed on unpurified product.
The dependence of the 21*i:21*t ratio on the catalyst used
to generate the carbenoid is inconsistent with the reaction
proceeding exclusively through a free ylide. If 21*t had been
formed by [1,2]-rearrangement of a common free oxonium
ylide, the proportion of 21*t would be completely independ-
ent of the catalyst used to generate the oxonium ylide; the re-
sults presented in Table 3 show that this is not the case.
within experimental error, and so the results were inconclusive.
However, in the case of the iridium-catalysed reaction (Table 2,
entry 5), almost complete scrambling of the label occurred
during the reaction leading to equal amounts of products 20*i
and 20*t. This result shows that when the iridium catalyst is
used to generate the carbenoid, the reaction does not proceed
by a pathway involving concerted [2,3]-sigmatropic rearrange-
ment of a free oxonium ylide, even though unlabelled 20 is
the product expected from rearrangement of a free oxonium
ylide. At this stage there seemed to be two possible explana-
tions for the result of the iridium-catalyzed reaction: the metal-
bound ylide undergoes dissociation to produce an iridium eno-
late and allylic cation, which recombine by CÀC bond forma-
tion, or metal-assisted allyl transfer from O to C occurs with
The 21*i:21*t ratios could be calculated accurately by inte-
1
gration of appropriate signals in the 13C and H NMR spectra.
As shown in Figure 1, it was easy to determine the product
1
ratio by integration of peaks in the region of the H NMR spec-
trum where signals for the allylic methylene group appear. In
the case of unlabelled benzofuranone 21 (Figure 1a), the pro-
tons appear as two apparent ddt centred at d=3.07 and
1
2.83 ppm in the H NMR spectrum. In the case of the fully la-
&
&
Chem. Eur. J. 2014, 20, 1 – 7
www.chemeurj.org
4
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
nal position of the allyl group, and 19*d, having a 13C label on
the carbon atom bearing the diazo group, were treated with
a suitable copper or iridium catalyst. In the event that the allyl
group is transferred intramolecularly, only products 21*i, 21*t
and 21*d should be observed. However, if dissociation of the
allyl group occurs prior to CÀC bond formation, one would
expect to observe formation of the unlabelled crossover prod-
uct 21 along with the doubly labelled crossover products
21**t and 21**i. Compound 21**i would be easy to identify in
the 13C NMR spectrum, because 13C–13C coupling would be evi-
dent. In addition, the ratio of crossover to non-crossover prod-
ucts could be determined from the relative amounts of 21**i
and 21*i.
When a mixture of 19* and 19*d was treated with [{Ir-
(cod)Cl}2] in 1,2-dichloroethane (0.31m) at reflux, a significant
amount of crossover product 21**i was observed. The com-
pound exhibited a 13C–13C J value of 38 Hz, which is typical of
coupling between adjacent sp3-hybridized 13C centres.[13] The
13C NMR data indicated a 21*i:21**i ratio of approximately 2:1.
The amount of doubly labelled product 21**i obtained was
also found to be concentration-dependent, and significantly
less of this compound was formed when the reaction was per-
formed at lower concentration (0.041m). When the same cross-
over experiment was repeated with [Cu(hfacac)2] in 1,2-di-
chloroethane (DCE, 0.31m), only traces of crossover product
21**i (<4%) were obtained.
Conclusion
We have provided important mechanistic insights into the
course of the intramolecular reaction of carbenoids with allylic
ethers. The results obtained from the copper- and rhodium-
catalyzed reactions of the 13C labelled diazo ketone 18*
[Eq. (7)] are consistent with a mechanism involving a free oxo-
nium ylide, but do not rule out involvement of the metal
centre in the rearrangement reaction. In contrast, results of the
copper- and iridium-catalyzed reactions of 13C labelled a-diazo
b-keto ester 19* [Eq. (8)] indicate that these reactions do not
proceed via a free oxonium ylide but instead follow major
competing non-ylide routes that deliver apparent [2,3]-sigma-
tropic rearrangement products. Furthermore, crossover experi-
ments with substrates 19* and 19*d (Scheme 2) suggest that,
Figure 1. 1H NMR spectra (d=2.6–3.3 ppm) for a) unlabelled 21, b) fully la-
belled 21*i and c) a mixture (47:53) of 21*i and 21*t.
1
belled compound 21*i (Figure 1b), H–13C coupling splits the
original signals to give four sets of apparent ddt centred at
d=3.20, 2.96, 2.93 and 2.69 ppm. This means that if mixtures
of 21*i and 21*t are obtained (Figure 1a) there is clear base-
line separation for the signals corresponding to products 21*i
and 21*t. Integration of these signals gives ratios that can
then be correlated with product ratios obtained from the
13C NMR spectra.
Having established that there was significant label scram-
bling during the reaction of 19* when both copper and iridi-
um catalysts were used, we sought to establish whether this
resulted from internal transfer or complete dissociation of the
allyl group. To do this we designed a crossover experiment in
which equal quantities of 19*, having a 13C label at the termi-
Scheme 2. Crossover experiment with labelled substrates 19*and 19*d.
Chem. Eur. J. 2014, 20, 1 – 7
www.chemeurj.org
5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
setter, A. J. Blake, W.-S. Li, W. G. Whittingham, Chem. Commun. 1999,
749; f) J. S. Clark, G. Whitlock, S. Jiang, N. Onyia, Chem. Commun. 2003,
2578; g) J. S. Clark, T. C. Fessard, C. Wilson, Org. Lett. 2004, 6, 1773;
h) J. S. Clark, C. A. Baxter, J. L. Castro, Synthesis 2005, 3398; i) J. S. Clark,
T. C. Fessard, G. A. Whitlock, Tetrahedron 2006, 62, 73; j) J. S. Clark, S. B.
Walls, C. Wilson, S. P. East, M. J. Drysdale, Eur. J. Org. Chem. 2006, 323;
k) J. S. Clark, C. Guꢁrot, C. Wilson, A. J. Blake, Chem. Commun. 2007,
4134; l) J. S. Clark, S. T. Hayes, C. Wilson, L. Gobbi, Angew. Chem. 2007,
119, 441; Angew. Chem. Int. Ed. 2007, 46, 437; m) J. S. Clark, C. A. Baxter,
A. G. Dossetter, S. Poigny, J. L. Castro, W. G. Whittingham, J. Org. Chem.
2008, 73, 1040; n) J. S. Clark, R. Berger, S. T. Hayes, L. H. Thomas, A. J.
Morrison, L. Gobbi, Angew. Chem. 2010, 122, 10063; Angew. Chem. Int.
Ed. 2010, 49, 9867; o) J. S. Clark, D. Vignard, A. Parkin, Org. Lett. 2011,
13, 3980; p) J. S. Clark, R. Berger, S. T. Hayes, H. M. Senn, L. J. Farrugia,
L. H. Thomas, A. J. Morrison, L. Gobbi, J. Org. Chem. 2013, 78, 673.
[5] a) T. H. Eberlein, F. G. West, R. W. Tester, J. Org. Chem. 1992, 57, 3479;
b) F. G. West, T. H. Eberlein, R. W. Tester, J. Chem. Soc. Perkin Trans.
1 1993, 2857; c) F. G. West, B. N. Naidu, R. W. Tester, J. Org. Chem. 1994,
59, 6892; d) F. P. Marmsꢂter, F. G. West, J. Am. Chem. Soc. 2001, 123,
5144; e) F. P. Marmsꢂter, J. A. Vanecko, F. G. West, Tetrahedron 2002, 58,
2027; f) F. P. Marmsꢂter, G. K. Murphy, F. G. West, J. Am. Chem. Soc. 2003,
125, 14724; g) F. P. Marmsꢂter, J. A. Vanecko, F. G. West, Org. Lett. 2004,
6, 1657; h) G. K. Murphy, F. P. Marmsꢂter, F. G. West, Can. J. Chem. 2006,
84, 1470; i) C. Stewart, R. McDonald, F. G. West, Org. Lett. 2011, 13, 720.
[6] a) M. C. Pirrung, W. L. Brown, S. Rege, P. Laughton, J. Am. Chem. Soc.
1991, 113, 8561; b) N. P. Karche, S. M. Jachak, D. D. Dhavale, J. Org.
Chem. 2001, 66, 6323; c) T. Yakura, W. Muramatsu, J. Uenishi, Chem.
Pharm. Bull. 2005, 53, 989; d) T. M. Weathers Jr., Y. Wang, M. P. Doyle, J.
Org. Chem. 2006, 71, 8183; e) D. M. Hodgson, D. Angrish, S. P. Erickson,
J. Kloesges, C. H. Lee, Org. Lett. 2008, 10, 5553; f) D. M. Jaber, R. N.
Burgin, M. Hepler, P. Zavalij, M. P. Doyle, Chem. Commun. 2011, 47,
7623; g) D. M. Hodgson, S. Man, Chem. Eur. J. 2011, 17, 9731.
[7] a) N. McCarthy, M. A. McKervey, T. Ye, M. McCann, E. Murphy, M. P.
Doyle, Tetrahedron Lett. 1992, 33, 5983; b) T. Ye, C. Fernꢃndez-Garcꢄa,
M. A. McKervey, J. Chem. Soc. Perkin Trans. 1 1995, 1373; c) N. Pierson,
C. Fernꢃndez-Garcꢄa, M. A. McKervey, Tetrahedron Lett. 1997, 38, 4705;
d) M. P. Doyle, D. G. Ene, D. C. Forbes, J. S. Tedrow, Tetrahedron Lett.
1997, 38, 4367; e) J. S. Clark, M. Fretwell, G. A. Whitlock, C. J. Burns,
D. N. A. Fox, Tetrahedron Lett. 1998, 39, 97; f) M. P. Doyle, D. C. Forbes,
M. M. Vasbinder, C. S. Peterson, J. Am. Chem. Soc. 1998, 120, 7653; <
lit g>D. M. Hodgson, M. Pertoliagi, Tetrahedron: Asymmetry 2001, 12,
877; h) S. Kitagaki, Y. Yanamoto, H. Tsutsui, M. Anada, M. Nakajima, S.
Hashimoto, Tetrahedron Lett. 2001, 42, 6361; i) N. Shimada, S. Nakamura,
M. Anada, M. Shiro, S. Hashimoto, Chem. Lett. 2009, 38, 488.
[8] a) D. M. Hodgson, P. A. Stupple, C. Johnstone, Tetrahedron Lett. 1997, 38,
6471; b) H. Suga, H. Ishida, T. Ibata, Tetrahedron Lett. 1998, 39, 3165;
c) S. Kitagaki, M. Anada, O. Kataoka, K. Matsuno, C. Umeda, N. Wata-
nabe, S. Hashimoto, J. Am. Chem. Soc. 1999, 121, 1417; d) F. Rodier, M.
Rajzmann, J.-L. Parrain, G. Chouraqui, L. Commeiras, Chem. Eur. J. 2013,
19, 2467.
in the case of the iridium-catalyzed reaction, the initial metal-
bound ylide dissociates to give an iridium enolate and allylic
cation, and these recombine by CÀC bond formation. In the
case of the copper-catalyzed reaction, intramolecular transfer
of the allyl group occurs with little dissociation of the inter-
mediate metal-bound ylide.
Experimental Section
Standard procedure for cyclization reactions of substrates
18 and 19
A solution of diazo ketone in appropriate solvent (1 mL per
0.1 mmol of diazo ketone) was added dropwise to a solution of
catalyst (1–10 mol%) in an equivalent volume of the same solvent
at reflux, and the reaction was monitored by TLC until completion.
The mixture was then cooled to room temperature and the reac-
tion was quenched by the addition of aqueous K2CO3 solution. The
phases were separated and the aqueous phase was extracted with
dichloromethane. The combined organic extracts were washed
with water followed by brine, dried (MgSO4) and concentrated in
vacuo. The residue was purified by flash column chromatography
on silica gel to give the product(s).
Acknowledgements
WestCHEM and the University of Glasgow are gratefully ac-
knowledged for providing studentship support for K.E.H. We
thank Dr. Alistair Boyer (University of Glasgow) for assistance in
gathering NMR data.
Keywords: carbenoids
rearrangement · ylides
· cyclization · isotopic labeling ·
[1] For early examples of intermolecular generation of oxonium ylides or
ylide-like intermediates from carbenoids, see: a) H. Nozaki, H. Takaya, R.
Noyori, Tetrahedron 1966, 22, 3393; b) H. Nozaki, H. Takaya, S. Moruiti,
R. Noyori, Tetrahedron 1968, 24, 3655; c) W. Ando, S. Kondo, K. Nakaya-
ma, K. Ichibori, H. Kohoda, H. Yamato, I. Imai, S. Nakaido, T. Migita, J.
Am. Chem. Soc. 1972, 94, 3870; d) M. P. Doyle, J. H. Griffin, M. S. Chinn,
D. van Leusen, J. Org. Chem. 1984, 49, 1917.
[2] For reviews on the formation of oxonium ylides or ylide-like intermedi-
ates from carbenoids, see: a) A. Padwa, S. F. Hornbuckle, Chem. Rev.
1991, 91, 263; b) T. Ye, A. M. McKervey, Chem. Rev. 1994, 94, 1091;
c) M. P. Doyle, M. A. McKervey, T. Ye in Modern Catalytic Methods for Or-
ganic Synthesis with Diazo Compounds, Wiley, New York, 1998, chapt. 7,
pp. 355–421; d) D. M. Hodgson, F. Y. T. M. Pierard, P. A. Stupple, Chem.
Soc. Rev. 2001, 30, 50; e) J. S. Clark, in Nitrogen, Oxygen and Sulfur Ylide
Chemistry, A Practical Approach in Chemistry (Ed.: J. S. Clark), Oxford Uni-
versity Press, Oxford, 2002, Ch. 1, pp. 1–113; f) A. G. H. Wee, Curr. Org.
Synth. 2006, 3, 499; g) J. B. Sweeney, Chem. Soc. Rev. 2009, 38, 1027;
h) G. K. Murphy, C. Stewart, F. G. West, Tetrahedron 2013, 69, 2667.
[3] a) M. C. Pirrung, J. A. Werner, J. Am. Chem. Soc. 1986, 108, 6060; b) E. J.
Roskamp, C. R. Johnson, J. Am. Chem. Soc. 1986, 108, 6062.
[9] T. Yakura, K. Matsui, K. Matsuzaka, M. Yamashita, Heterocycles 2009, 79,
353.
[10] D. M. Jaber, R. N. Burgin, M. Helper, P. Y. Zavalij, M. P. Doyle, Org. Lett.
2012, 14, 1676.
[11] a) Z. Li, H. M. L. Davies, J. Am. Chem. Soc. 2010, 132, 396; b) Z. Li, B. T.
Parr, H. M. L. Davies, J. Am. Chem. Soc. 2012, 134, 10942; c) Z. Li, V. Boy-
arskikh, J. H. Hansen, J. Autschbach, D. J. Musaev, H. M. L. Davies, J. Am.
Chem. Soc. 2012, 134, 15497.
[12] I. K. Mangion, I. K. Nwamba, M. Shevlin, M. A. Huffman, Org. Lett. 2009,
11, 3566.
[13] S. D. Yeoh, C. E. Skene, J. M. White, J. Org. Chem. 2013, 78, 311.
[4] a) J. S. Clark, Tetrahedron Lett. 1992, 33, 6193; b) J. S. Clark, S. A. Krowiak,
L. J. Street, Tetrahedron Lett. 1993, 34, 4385; c) J. S. Clark, G. A. Whitlock,
Tetrahedron Lett. 1994, 35, 6381; d) J. S. Clark, A. G. Dossetter, W. G.
Whittingham, Tetrahedron Lett. 1996, 37, 5605; e) J. S. Clark, A. G. Dos-
Received: October 16, 2013
Published online on && &&, 0000
&
&
Chem. Eur. J. 2014, 20, 1 – 7
www.chemeurj.org
6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
&
Rearrangement
J. S. Clark,* K. E. Hansen
&& – &&
Intramolecular Reactions of Metal
Carbenoids with Allylic Ethers: Is
a Free Ylide Involved in Every Case?
Differing distributions of 13C-labelled
rearrangement products in the rhodi-
um-, copper- and iridium-catalyzed reac-
tions of a 13C-labelled a-diazo b-keto
ester (see scheme) revealed that the
copper- and iridium-mediated reactions
do not proceed via a common free oxo-
nium ylide intermediate.
Chem. Eur. J. 2014, 20, 1 – 7
www.chemeurj.org
7
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ