Aluminium(III) trifluoromethanesulfonate as an efficient catalyst for the intramolecular hydroalkoxylation of unactivated olefins: Experimental and theoretical approaches

Article abstract of DOI:10.1002/chem.200501478

The Al(OTf)₃-catalyzed cycloisomerization of unactivated unsaturated alcohols was studied from experimental and theoretical points of view. A series of cyclic ethers was obtained in excellent yields and regioselectivities. This catalyst system provides one of the most straightforward routes to cyclic ethers with Markovnikov-type regioselectivity under mild conditions. Theoretical and NMR studies were carried out in order to better determine the mechanism of this reaction. The NMR studies were in agreement with preferential complexation of Al(OTf)₃ to the oxygen atom of the unsaturated alcohol, but did not exclude complexation to the double bond of the alcohol. Theoretical calculations indicated strong acidification of the hydroxyl proton when Al(OTf)₃ was complexed to the alcohol oxygen atom. A plausible catalytic cycle for the Al(OTf)₃-catalyzed intramolecular hydroalkoxylation of unactivated olefins is proposed.

Full text of DOI:10.1002/chem.200501478

DOI: 10.1002/chem.200501478
Aluminium(iii) Trifluoromethanesulfonate as an Efficient Catalyst for the
A
C
H
T
R
E
U
N
G
Intramolecular Hydroalkoxylation of Unactivated Olefins:
Experimental and Theoretical Approaches
[
a]
[b]
[b]
[a]
Lydie Coulombel, Michel Rajzmann, Jean-Marc Pons,* Sandra Olivero, and
[
a]
Elisabet DuÇach*
Abstract: The Al
A
C
H
T
R
E
U
N
G
(OTf) -catalyzed cy-
Theoretical and NMR studies were car-
ried out in order to better determine
the mechanism of this reaction. The
NMR studies were in agreement with
to the oxygen atom of the unsaturated
alcohol, but did not exclude complexa-
tion to the double bond of the alcohol.
3
cloisomerization of unactivated unsatu-
rated alcohols was studied from experi-
mental and theoretical points of view.
A series of cyclic ethers was obtained
in excellent yields and regioselectivi-
ties. This catalyst system provides one
of the most straightforward routes to
cyclic ethers with Markovnikov-type
regioselectivity under mild conditions.
Theoretical
strong acidification of the hydroxyl
proton when Al(OTf) was complexed
to the alcohol oxygen atom. A plausi-
ble catalytic cycle for the Al(OTf) -cat-
calculations
indicated
preferential complexation of Al(OTf)
A
H
R
U
G
3
AHCTREUNG
3
Keywords: aluminum · cyclization ·
Lewis acids · oxygen heterocycles ·
reaction mechanisms
AHCTREUNG
3
alyzed intramolecular hydroalkoxyla-
tion of unactivated olefins is proposed.
Introduction
due to the strongly electron withdrawing trifluoromethane-
[4]
sulfonate group.
The use of transition-metal catalysis for selective hydroal-
Saturated oxygen heterocycles are important core structures
which are frequently found in polyether antibiotics and
[5]
koxylation of unsaturated hydrocarbons has been reported
[1]
[6]
other biologically active natural products. Great interest is
focused on the development of new and efficient methods
for the synthesis of cyclic ethers. One of the most appealing
approaches to these heterocycles is intramolecular hydroal-
koxylation, in which the cyclic ether is formed by addition
in the cases of CꢀC and activated C=C bonds such as al-
[
7]
[8]
[9]
III
[10]
lenes, 1,3-dienes and Michael acceptors. Ce /NaI,
III
I[11]
I
I
[12]
II
Ru /Ag
and Au /Ag systems, as well as Pt complexes
[13]
in the presence of phosphine ligands, have recently been
used as catalysts for the intramolecular hydroalkoxylation of
unactivated g- and d-hydroxy alkenes to form cyclic ethers.
[2]
of an alcohol to an olefin. Although this reaction is known
to be catalyzed by Brønsted acids in superstoichiometric
amounts, we recently reported the use of tin(iv) triflate as
Herein we report that aluminium(iii) trifluoromethanesul-
A
H
R
U
G
fonate is an efficient catalyst for the highly regioselective in-
tramolecular hydroalkoxylation of unactivated olefins to
afford the corresponding cyclic ethers in excellent yields.
Our study includes experimental results and some mechanis-
tic considerations based on NMR experiments and semiem-
[3]
catalyst for the hydroalkoxylation of unactivated alkenes.
Indeed, the highly Lewis acidic character of this catalyst is
III
[
a] Dr. L. Coulombel, Dr. S. Olivero, Dr. E. DuÇach
Laboratoire de Chimie des MolØcules Bioactives et des Arômes
UMR-CNRS 6001, UniversitØ de Nice-Sophia Antipolis
Parc Valrose, 06108 Nice Cedex 2 (France)
Fax : (+33)492-076-125
pirical calculations. The use of Al compounds for the intra-
molecular cyclization of alcohols has only been reported
[14]
with superstoichiometric amounts of AlCl3.
E-mail: dunach@unice.fr
[
b] Dr. M. Rajzmann, Prof. J.-M. Pons
Laboratoire SYMBIO, UMR-CNRS 6178
FacultØ des Sciences et Techniques, UniversitØ Paul CØzanne
Boîte D12, 13397 Marseille Cedex 20 (France)
Fax : (+33)491-288-841
Results and Discussion
Cycloisomerization of unsaturated alcohols catalyzed by Al-
A
H
R
U
G
A
H
R
U
G
3
E-mail: jean-marc.pons@univ.u-3mrs.fr
able, was investigated as catalyst for the intramolecular hy-
6356
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6356 – 6365

FULL PAPER
Table 1. Hydroalkoxylation of g,d-unsaturated alcohols catalyzed by Al
A
H
R
U
G
3
(5 mol%).
kov-type addition. No five-
Entry Unsaturated alcohol
Conditions
Major product
Yield of iso-
lated product
Regioselectivity five- vs membered cyclic ether was ob-
six-membered cyclic
served.
[
%]
ether
Cycloisomerization of alcohol
1
b (entry 2) bearing a 1,1-di-
CH
2 2
Cl ,
1
93
0:100
100:0
408C, 1 h
substituted double bond gave
the five-membered cyclic ether
2
yield, regiospecifically. In both
cases, the reaction was com-
plete in refluxing dichlorome-
thane after 1 h.
1
1
a
3a
2b
b as the only product in 88%
CH
2 2
Cl ,
2
88
408C, 1 h
b
In the case of alcohol 1c,
with an internal disubstituted Z
double bond (entry 3), cycliza-
tion was successfully carried
out in refluxing nitromethane
and afforded the five-mem-
bered cyclic ether as the major
compound. The selectivity for
five-versus six-membered cyclic
ether was 92:8.
CH
3 2
NO ,
3
4
5
88
87
92:8
100:0
100:0
1018C, 1.5 h
1
1
c
2c
2d
CH
3 2
NO ,
1018C, 1 h
d
CH
3 2
NO ,
>93
1018C, 3 h
1
e
2e
The Al(OTf) catalyst system
A
H
R
U
G
3
ClCH
2
CH
2
Cl,
6
7
91
0:100
was also applied to the intramo-
lecular hydroalkoxylation of
terminal, unactivated olefins
such as 1d or 1e (entries 4 and
5). For both substrates, the re-
action tookplace at the more
848C, 0.5 h
1
1
f
3 f
2g
3 2
CH NO ,
[
a]
88
100:0
1018C, 1.5 h
g
substituted
olefinic
carbon
[
a] cis/trans ratio of 2g was 60:40, as calculated by GC and as determined by NOESY NMR experiments.
atom and led exclusively to the
corresponding tetrahydrofurans
2
d and 2e.
droalkoxylation of unactivated olefins. The influence of the
solvent (e.g., dichloromethane, acetonitrile, dichloroethane
and nitromethane) on the cycloisomerization of two model
substrates 1a and 1c was first examined in the presence of
The Al
ACHTREUNG
(OTf) -catalyzed hydroalkoxylation was also effi-
3
ciently applied to the synthesis of spirobicyclic ether 3 f,
starting from the corresponding doubly unsaturated diol 1 f
(entry 6). Bis-tetrahydropyran 3 f was obtained with com-
plete regioselectivity in quantitative yield.
Al(OTf) (5 mol%). Whereas unsaturated alcohol 1a led
A
C
H
T
R
E
U
N
G
3
quantitatively to tetrahydropyran 3a in all solvents tested,
no reaction occurred with 1c in refluxing dichloromethane,
and only 10% conversion of 1c was reached in refluxing
acetonitrile after 22 h. However, cyclization of 1c was effi-
cient in refluxing nitromethane and afforded the corre-
sponding tetrahydrofuran 2c after 1.5 h with 92% regiose-
lectivity.
As an extension of the reaction, the preparation of g-lac-
tone 2g from the corresponding g,d-unsaturated carboxylic
acid 1g (entry 7) was carried out under the same catalytic
conditions, with 100% regioselectivity in favour of the g-lac-
tone.
The reaction rates observed for the cycloisomerization of
g,d-unsaturated alcohols in the presence of Al(OTf) al-
A
H
R
U
G
3
The cyclization of differently substituted olefinic alcohols
lowed the following reactivity order to be established: tri-
substituted ꢁ 1,1-disubstituted > 1,2-disubstituted > mono-
substituted olefins.
The hydroalkoxylation of olefins was very dependent on
the nature of the catalyst. To highlight the influence of the
counterion on reactivity and selectivity of cycloisomeriza-
tion, we examined the cyclization of substrate (Z)-4-decenol
was examined by using a catalytic amount of Al
5 mol%). The required solvent and reaction temperature
was dependent on the double-bond substitution of the olefin
Table 1). For each substrate the solvent was chosen accord-
A
H
R
N
(
(
ing to its boiling point. However, the solubility of Al
was much higher in nitromethane.
A
H
R
U
G
As shown in Table 1, the hydroalkoxylation of an unsatu-
(1c) in the presence of several aluminum(iii) compounds as
A
H
R
U
G
rated alcohol with a trisubstituted double bond, such as 1a
catalysts.
(
entry 1), led exclusively to formation of the corresponding
Kinetics of the reaction are shown in Figure 1 for Al-
(OTf) , AlBr , AlCl and Al(OiPr) . After 1.5 h in refluxing
tetrahydropyran 3a, the result of cyclization at the more
A
H
R
N
ACHTREUNG
3
3
3
3
substituted carbon atom of the double bond in a Markovni-
nitromethane, the reaction of decenol 1c with 5 mol% of
Chem. Eur. J. 2006, 12, 6356 – 6365
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6357

E. DuÇach, J.-M. Pons et al.
A
C
H
T
R
E
U
N
G
3
Table 2. Cyclization of 1a in the presence of 5 mol% of Al(OTf) or
TfOH and an additive in refluxing dichloromethane.
Entry Catalyst
5 mol%)
Additive
t
Conversion of Yield of isolat-
[
a]
(
[h]
1a [%]
ed 3a [%]
1
2
Al
Al
A
T
E
N
3
3
–
1
1
100
100
93
92
A
H
R
U
G
2,6-lutidine
(
5 mol%)
3
4
TfOH
TfOH
–
0.5
24
100
<5
90
–
2,6-lutidine
(
5 mol%)
5
6
Al
Al
A
T
E
N
3
3
H
2
O
1
1
100
98
92
90
(
5 mol%)
A
H
R
N
2
H O
Figure 1. Kinetics of the hydroalkoxylation of (Z)-4-decenol (1c) cata-
lyzed by aluminium compounds (5 mol%) in refluxing nitromethane (n-
(50 mol%)
[
a] Conversions calculated by GC with n-undecane as internal standard.
undecane used as internal standard); ^: Al
Al(OiPr)
A
H
R
U
G
3 3 3
(OTf) , : AlCl , ”:
&
, ~: AlBr
A
C
H
T
R
E
U
N
G
3
.
In the presence of 2,6-lutidine as a hindered base (Al-
A
H
R
U
G
A
T
E
N
3
AlCl gave a mixture of tetrahydrofuran 2c and tetrahydro-
pyran 3c (Scheme 1) with less than 10% conversion. With
of heptenol 1a was similar to that obtained without added
base (cf. entries 1 and 2). This result indicates that the hin-
dered base does not inhibit cyclization, and proton capture
by the base is not directly involved in the reaction.
3
AlBr the conversion reached up to 40% in favour of 2c,
3
The cycloisomerization of 1a could also be run with
5
mol% of TfOH and afforded the corresponding tetrahy-
dropyran 3a quantitatively (entry 3). However, the reaction
of 1a with 5 mol% of TfOH and a hindered base as additive
(
2
5 mol%) gave a very low conversion of 1a (<5%) after
4 h (entry 4).
Scheme 1.
These observations clearly indicate that when Al(OTf) is
A
C
H
T
R
E
U
N
G
3
but no further evolution of the reaction mixture was ob-
involved, the reaction is indeed induced by the Lewis acid
and not by triflic acid, which could result from the reaction
served. No reaction occurred with Al
A
H
R
U
G
action with 5 mol% of Al(OTf) tookplace quantitatively
after 1.5 h. The Lewis acidity of the catalyst strongly deter-
mines the reactivity and the selectivity.
A
H
R
U
G
between Al
The influence of water on the Al(OTf) -catalyzed cyclo-
3
isomerization of 1a was examined with a view to partial hy-
drolysis of the catalyst. Interestingly, as shown in Table 2
(entries 5 and 6), similar results were obtained with or with-
out added water, that is, the presence of water has minor in-
A
H
R
U
G
3
A
H
R
U
G
Mechanistic studies on the Al(OTf) -catalyzed hydroalkoxy-
A
H
R
U
G
3
lation of olefins: In protic acidic media, the mechanism of
the cycloisomerization of unsaturated alcohols is a well-
fluence on the catalytic activity of Al
A
H
R
U
[15]
known process. It generally involves initial protonation of
the double bond with intermediate formation of the more
substituted carbocation. Subsequent attackof the hydroxyl
group on the more substituted carbon atom of the double
bond forms the ether with Markovnikov-type regioselectivi-
ty. The mechanism of this reaction in the presence of Lewis
acids has not been yet reported.
periments indicate that Al(OTf) is an active species in the
ACHTREUNG
3
intramolecular hydroalkoxylation of unactivated olefins, and
its catalytic activity is similar to that of a superacid.
The Al(OTf) -catalyzed hydroalkoxylation (Table 1) ex-
A
H
R
U
G
3
hibits Markovnikov-type regioselectivity, as is observed for
the same reaction in protic acidic media. This suggests that
the mechanism of Lewis acid catalyzed hydroalkoxylation of
double bonds should involve cationic-type intermediates.
II
In a recent example with Pt in the presence of phosphine
[13]
ligands for the hydroalkoxylation of unactivated olefins,
the observed regioselectivity was similar to that found with
Al(OTf) (Markovnikov-type addition). The proposed
In the hydroalkoxylation of olefins catalyzed by Al(OTf) ,
A
C
H
T
R
E
U
N
G
3
which is a strong Lewis acid, it is assumed that the alumini-
um atom should preferentially coordinate to the hydroxyl
group. However, we were interested to determine whether
coordination to both the hydroxyl oxygen atom and the CÀ
ACHTREUNG
3
mechanism involved platinum(ii) complexation to the olefin,
followed by outer-sphere attackof the pendant hydroxyl
group.
C double bond of the unsaturated alcohol could be consid-
ered.
III
To better determine the role of Al Lewis acid as catalyst
1
for the hydroalkoxylation reaction, the cyclization of sub-
strate 1a was examined in the presence of an added base or
water, and the results were compared to the use of TfOH
with or without an added base. The results are summarized
in Table 2.
To investigate this possibility, we carried out H and
C NMR studies on the complexation of Al(OTf) to alco-
1
3
A
H
R
U
G
3
hol 1c. NMR experiments can highlight complex formation
and have previously been used, for example, to evaluate the
+
[16]
complexation between Ag and alkynes.
6358
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6356 – 6365

Cyclization of Unsaturated Alcohols
FULL PAPER
Several NMR experiments
with decenol 1c were carried
out in CD NO at room temper-
3
2
ature with tetramethylsilane
TMS) as internal standard in
the presence of 6 equiv Al- Figure 3. Possible Al coordination modes to an unsaturated alcohol.
(
III
A
C
H
T
R
E
U
N
G
(OTf) . At the concentrations
3
used, Al(OTf) was not soluble
A
C
H
T
R
E
U
N
G
3
in CDCl or CD Cl .
The H NMR experiments were carried out with 1c and
Theoretical Studies
3
2
2
1
1
Al
A
C
H
T
R
E
U
N
G
(OTf) . A shift of the signal corresponding to protons H
Methodology: All calculations reported in this workwere
3
[17]
of almost 0.1 ppm (Dd = +0.092 ppm, see Figure 2 for
performed with the semiempirical AM1 method available
[18]
in the AMPAC program. Geometries of stationary points
were determined by minimization of energy with respect to
all geometric parameters. Transition states were located by
[
19]
the chain method and characterized by one, and only one,
negative eigenvalue of the Hessian matrix. Finally, the tran-
sitions states found were shown to belong to the studied re-
action by intrinsic reaction coordinate (IRC). The AM1
method was chosen mainly for three reasons: 1) its reliabili-
ty compared to ab initio calculations for reactions involving
[20]
Lewis acids and our experience in the field, 2) the reason-
ably short calculation times, which enabled us to perform
IRC on each transition state, and 3) the possibility to tackle
the real molecules and therefore compare calculations to ex-
periments.
Calculations were run on unsaturated alcohols 1a and 1d,
with the following goals (Scheme 2): 1) Understanding the
action of the Lewis acid and the importance of its nature;
2) Understanding the selectivity for five versus six-mem-
bered cyclization on going from 1 to 2 or 3, respectively.
1
3
Figure 2. C NMR chemical shifts [ppm vs TMS] of alcohol 1c alone and
1
4
5
complexed to Al
CD NO at 208C. [1c]=0.03m and [Al
A
C
H
T
R
E
U
N
G
(OTf)
3
for characteristic carbon atoms C , C and C in
(OTf) ]=0.18m.
3
2
A
H
R
U
G
3
numbering) indicates, as expect-
III
ed, complexation of Al to the
hydroxyl oxygen atom of dece-
nol 1c. A very small variation
of the chemical shift of the eth-
4
5
ylenic protons H and H (Dd =
0.009 ppm) does not allow
Scheme 2.
+
complexation of the double
bond to be concluded.
The same trend, but more pronounced, was observed in
C NMR experiments (Figure 2). The complexation of Al-
Calculations involved only one alcohol molecule, since
preliminary studies showed that the introduction of a
second one had very little influence on the activation ener-
1
3
A
C
H
T
R
E
U
N
G
(OTf) to the double bond and to the oxygen atom of dece-
3
nol 1c was shown by slight shifts of Dd = +0.19 and
gies of the uncatalyzed or Al(OTf) -catalyzed reactions
A
H
R
U
G
3
[21]
À0.28 ppm for the signals corresponding to ethylenic carbon
leading from alcohol 1d to tetrahydrofuran 2d. Moreover,
kinetic data indicated a first-order reaction with respect to
the alcohol.
5
4
atoms C and C , respectively, and a larger shift of Dd =
1
+
0.53 ppm for C bearing the hydroxyl group.
The NMR studies are in agreement with preferential com-
plexation of Al
A
C
H
T
R
E
U
N
G
(OTf) to the hydroxyl oxygen atom of 1c,
Cyclization of 4-penten-1-ol (1d): Calculations were first
run on the cyclization of pentenol 1d in the absence of
Lewis acid, considering both the formation of tetrahydrofur-
an 2d (Table 3, entry 1) and tetrahydropyran 3d (Table 3,
entry 2). As shown by the energies of their transition states,
the competition should be in favour of the five-membered
cyclic molecule 2d, but in both cases, the activation energies
3
but do not exclude some complexation with the isolated CÀ
C double bond.
Our results suggest that in the Al(OTf) -catalyzed hydro-
A
H
R
U
G
3
alkoxylation of unactivated olefins, structures A and B
could both be considered as intermediate species leading to
the cyclic ethers (Figure 3).
À1
are important (>60 kcalmol ). The process is a concerted
Chem. Eur. J. 2006, 12, 6356 – 6365
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6359

E. DuÇach, J.-M. Pons et al.
À1
Table 3. Energies [kcalmol ] for alcohols 1a and 1d, complexes A, A’ and C, transition states TS1–TS12 and ethers 2a,d and 3a,d.
Entry Starting Lewis acid SM energy Complex type Complex Complex energy TS
material
TS energy
E
a
Product
Product energy
1
2
1d
1d
TS1
TS2
+9.7
+19.4
+61.9 2d
+71.6 3d
À62.8
À64.3
À52.2
3
4
5
6
1d
1d
1d
1d
AlCl
AlCl
AlCl
AlCl
3
3
3
3
O
C=C
O
À199.1
À196.6
À200.5
À196.9
TS3
TS4
TS5
TS6
À140.4
À137.8
À129.2
À128.1
+58.7 2d-AlCl
+58.8 2d-AlCl
+71.3 3d-AlCl
+68.8 3d-AlCl
3
3
3
3
À208.9
À212.8
À192.5
C=C
7
8
9
1
1d
1d
1d
1d
Al
Al
Al
Al
A
C
H
T
R
E
U
N
G
(OTf)
(OTf)
(OTf)
(OTf)
3
3
3
3
O
C=C
O
A
C
A
C
À904.8
À901.6
À905.3
À901.6
TS7
TS8
TS9
TS10
À877.6
À845.1
À875.2
À834.1
+27.2 2d-Al
+56.5 2d-Al
+30.1 3d-Al
+67.5 3d-Al
A
T
E
G
3
3
3
3
À913.7
À915.6
A
C
H
T
R
E
U
N
G
ACHTREUNG
À891.0
À910.5
A
C
H
T
R
E
U
N
G
ACHTREUNG
0
A
C
H
T
R
E
U
N
G
C=C
ACHTREUNG
1
1
1a
1b
1a
1a
Al
Al
A
C
H
T
R
E
U
N
G
(OTf)
3
3
O
O
TS11a À904.0
TS11b À895.8
+23.6 2a-Al
+31.8 2a-Al
A
T
E
N
3
3
À928.0
À923.2
A
C
H
T
R
E
U
N
G
(OTf)
ACHTREUNG
A’
À927.6
1
1
2a
2b
1a
1a
Al
Al
A
C
H
T
R
E
U
N
G
(OTf)
3
3
O
O
TS12a À904.0
TS12b À907.8
+23.6 3a-Al
+19.8 3a-Al
A
T
E
N
3
3
A
C
H
T
R
E
U
N
G
(OTf)
ACHTREUNG
mechanism. These results are consistent with the experimen-
tal observation that neither of these two ethers is formed
from 1d in the absence of catalyst.
the cyclization reaction, in agreement with the experimental
results shown in Figure 1.
The results of introducing Al(OTf) in the calculations
A
H
R
U
G
3
We then involved AlCl3 in the calculations with 1d
with alcohol 1d, following the same approach as described
(
Table 3, entries 3–6) and four different approaches were in-
for AlCl , are reported in entries 7–10 of Table 3.
3
vestigated: complexation of AlCl to the oxygen atom of the
Figure 5 presents the energies and geometries for the less
demanding cyclization process corresponding to entry 7.
3
alcohol, as in intermediate A of Figure 3 (entries 3 and 5) or
to the CÀC double bond, as in structure C of Figure 3 (en-
tries 4 and 6). In both cases, the five- versus six-membered
competition leading to 2d and 3d, respectively, was consid-
[22]
ered.
The calculations indicated that the introduction of AlCl3
as catalyst has a very small influence on the activation
energy of the reaction. In all four cases, the reaction remains
concerted and involves a single transition state. The activa-
tion energies involved are still in favour of formation of tet-
rahydrofuran 2d (entry 3), as illustrated in Figure 4. The in-
troduction of AlCl as Lewis acid catalyst is not efficient for
3
Figure 5. Al
A
H
R
U
G
3
(OTf) -catalyzed cycloisomerization of 4-penten-1-ol (1d) to
2
-methyltetrahydrofuran 2d by complexation to the oxygen atom.
Apparently, complexation of Al(OTf) to the oxygen
A
H
R
U
G
3
atom of alcohol 1d allows the reactions towards tetrahydro-
furan 2d and tetrahydropyran 3d to still proceed in a con-
certed manner, with much smaller activation energies of
À1
À1
2
7.2 kcalmol (entry 7) and 30.1 kcalmol (entry 9), re-
À1
spectively. These data compare with 61.9 and 70.9 kcalmol
for the calculations without catalyst (entries 1 and 2), and
À1
5
8.7 and 71.3 kcalmol with AlCl (entries 3 and 5).
3
On the other hand, complexation of Al
double bond of alcohol 1d (entries 8 and 10) shows high ac-
A
H
R
U
G
3
Figure 4. AlCl
3
-catalyzed cycloisomerization of 4-penten-1-ol (1d) to 2-
III
methyltetrahydrofuran 2d by Al complexation to the oxygen atom of
1
d.
tivation energies, similar to those calculated for the uncata-
6360
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6356 – 6365

Cyclization of Unsaturated Alcohols
FULL PAPER
lyzed (entries 1 and 2) and AlCl -catalyzed reactions (en-
to tetrahydropyran 3a (Figure 6). The activation energies
3
tries 4 and 6).
found for the cyclization of 1a with Al
A
H
R
U
G
Calculations on the more favoured O complexation indi-
cated that the competition is in favour of formation of the
ment with those listed in entries 7 and 9 for 1d, and are in
five-membered cyclic ether 2d. A difference of around
À1
3
kcalmol (entries 7 and 9) was found between the transi-
tion-state energies leading to 2d and 3d. The strong reduc-
tion of the activation energy of the process leading to tetra-
hydrofuran 2d (entry 7) as compared to the formation of
tetrahydropyran 3d (entry 9) is in complete agreement with
the experimental results, in which exclusive formation of
product 2d was observed (Table 1, entry 4).
A reaction path involving alkoxyalumination was also in-
vestigated. The addition of the ROÀAl
C=C bond with formation of an R’Al
and subsequent protonation of this organometallic com-
pound (with concomitant regeneration of the Al(OTf) cata-
A
H
E
N
A
H
R
U
G
AHCTREUNG
3
lyst) by triflic acid was considered. Although the first step
leading to the five-membered cyclic intermediate requires
A
H
R
U
G
3
Figure 6. Al(OTf) -catalyzed cycloisomerization of 6-methyl-5-hepten-2-
À1
only 18.4 kcalmol , the protonation step requires a much
ol (1a) to 2,2,6-trimethyltetrahydropyran (3a) through TS12a (note that
TS12b, with lower energy than TS12a and close to TS12a, is not shown in
this figure).
À1
higher activation energy (E =54.5 kcalmol ). Such a mech-
a
anism was therefore unlikely to occur and was not further
investigated.
À1
Cyclization of 6-methyl-5-hepten-2-ol (1a): Experimental re-
the range of 19.8–31.8 kcalmol (entries 11a, 11b and 12a,
sults indicated that the use of Al
the efficient and catalytic cycloisomerization of unsaturated
alcohols. Calculations with 1d were in agreement with a
A
H
R
U
G
12b). Moreover, TS11a and TS12a are quasi-equivalent and
3
1
7
both involve an interesting lengthening of the O ÀH bond
7 5
and an interaction between H and C (Table 4, entries 8
and 10; numbering shown in Scheme 2).
much higher efficiency of Al(OTf) as compared to AlCl ,
A
H
R
U
which indicates a much lower activation energy for the cycli-
zation (Table 3, entry 7). The lowest transition-state energy
Data on the bond lengths and partial charges of the differ-
ent complexes and transition states are presented in Table 4.
The energy difference between the two reaction paths
leading to 2a (TS11a and TS11b) or 3a (TS12a and TS12b)
is in favour of the cyclization steps leading to the six-mem-
was found for Al
atom. Accordingly, for the next calculations concerning the
cyclization of 1a, we only used Al(OTf) , and we investigat-
ed the competition between five- (2a) and six-membered
ring formation (3a) with Al(OTf) complexed to the oxygen
A
C
H
T
R
E
U
N
G
(OTf) complexed to 1d through its oxygen
3
AHCTREUNG
3
5
6
bered 3a. The charge distribution on C and C in TS12b
7
5
A
H
R
U
G
(Table 4, entry 11) induces the transfer of H to C in TS12b,
3
1
6
atom of 1a, as in intermediate A of Figure 3 (Table 3, en-
tries 11 and 12).
For formation of both five- and six-membered cyclic
ethers, the calculated reaction mechanism becomes stepwise
and involves two transition states TS11a and TS11b from 1a
leading to tetrahydrofuran 2a, and TS12a and TS12b leading
as well as O transfer to C .
Calculated data are clearly in agreement with the experi-
mental results. Indeed, the activation energy required to se-
lectively form tetrahydropyran 3a from 1a—the first step is
clearly rate-determining, since the second has a relative acti-
À1
vation energy of À3.8 kcalmol and involves only minor re-
Table 4. Main distances and partial charges of alcohols 1a and 1d, complexes A and A’ and transition states TS7, TS9, TS11a, TS11b, TS12a and TS12b
see also Table 3).
(
Entry
Structures
Distances [Š]
Partial charges [e]
1
5
1
6
1
7
5
6
5
7
6
7
5
6
1
7
O ÀC
O ÀC
O ÀH
C ÀC
C ÀH
C ÀH
dC
dC
dO
dH
1
2
3
4
5
6
7
8
9
1d
3.360
3.240
3.087
2.835
2.702
3.062
3.043
2.613
2.619
2.611
2.647
4.350
4.213
3.598
2.719
2.434
4.147
3.751
3.502
3.296
3.510
2.044
0.965
0.969
0.985
1.495
2.337
0.966
0.984
1.446
1.993
1.446
2.816
1.331
1.331
1.332
1.381
1.421
1.343
1.344
1.403
1.399
1.403
1.483
2.650
2.560
2.340
1.923
1.174
2.484
2.262
1.298
1.511
1.295
1.129
3.557
3.391
2.686
1.323
1.990
3.381
2.872
2.112
1.392
2.116
2.142
À0.173
À0.187
À0.225
+0.141
À0.259
À0.198
À0.254
À0.347
À0.008
À0.348
À0.214
À0.213
À0.202
À0.196
À0.347
+0.299
À0.111
À0.063
+0.235
+0.027
+0.235
+0.364
À0.330
À0.320
À0.332
À0.543
À0.568
À0.327
À0.327
À0.511
À0.557
À0.508
À0.516
0.199
0.231
0.322
+0.398
+0.283
0.202
complex A (1d+AlCl
complex A (1d+Al
TS7
TS9
1a
complex A’ (1a+Al
TS11a
TS11b
TS12a
TS12b
3
on O
1
)
A
H
R
U
G
3
on O
1
)
A
H
R
U
G
3
on O
1
)
0.319
+0.369
+0.335
+0.367
+0.150
1
1
0
1
Chem. Eur. J. 2006, 12, 6356 – 6365
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6361

E. DuÇach, J.-M. Pons et al.
arrangements–-is more than
À1
8
kcalmol lower than the acti-
vation energy necessary for the
formation of tetrahydrofuran
2
a (Table 3, cf. entries 11b and
12a). The calculations for the
cyclization of 1a to 3a are pre-
sented in Figure 6.
On the other hand, compar-
ing the cyclization of substrates
1
d
and 1a, the activation
energy leading to tetrahydro-
pyran 3a (Table 3, entry 12a:
À1
2
3.6 kcalmol ) is lower than
that leading to tetrahydrofuran
2
d (Table 3, entry 7: 27.2 kcal
À1
mol ). These results are also in
agreement with the experimen-
tal results of Table 1: cyclization
of 1a can take place at much
lower temperature than cycliza-
tion of 1d (Table 1, entries 1
and 4).
Figure 7. Proposed mechanism for the concerted and regioselective Al
-penten-1-ol (1d) to 2-methyltetrahydrofuran (2d).
A
H
R
U
G
3
(OTf) -catalyzed cycloisomerization of
4
Discussion
Proposed catalytic cycle: Based
on our catalytic reaction and on
semiempirical calculations, we
can propose a different mecha-
nism for the Al(OTf) -catalyzed
A
C
H
T
R
E
U
N
G
3
cycloisomerization of each of
the unsaturated alcohols 1a and
1
d (Figures 7 and 8).
Theoretical and experimental
studies indicate that AlCl3 is
A
H
R
U
G
3
Figure 8. Proposed mechanism for the stepwise Al(OTf) -catalyzed cycloisomerization of 6-methyl-5-hepten-2-
not active enough to efficiently ol (1a) to 2,2,6-trimethyltetrahydropyran (3a).
effect the cyclization of 1a or
1
d, as opposed to Al
Reaction paths involving the
coordination of Al
(OTf) to the CÀC double bond, as in
structure C, are energetically much more demanding than
A
H
R
U
G
A
C
H
T
R
E
U
N
G
considered. However, theoretical calculations are clearly in
favour of intermediate A as the reactive complex, since it
leads to more stable transitions states TS7 and TS9
(Figure 7 and Table 3, entries 7 and 9).
3
those involving
O
complexation, as in structure
A
(
Figure 3). Therefore, the catalytic cycles for the cycloisome-
rization of 1a and 1d will only be proposed for Al
complexed to the oxygen atom of the alcohols.
A
H
R
U
G
The theoretical data on the cyclization of 1d towards five-
3
1
7
membered cyclic ether 2d indicate that the O ÀH bond
length in 1d (0.965 Š) is larger in complex A (0.985 Š) and
further increases in transition state TS7 (1.495 Š), as shown
in Table 4 (entries 1, 3 and 4).
In the case of pentenol 1d, theoretical calculations provid-
ed a plausible cyclization mechanism involving a single tran-
sition state. We can thus consider a concerted mechanism
with a catalytic cycle (Figure 7). In the first step, coordina-
tion of the catalyst to the oxygen atom of the unsaturated
alcohol gives intermediate complex A. This intermediate
may be in equilibrium with chelate B in which the catalyst is
coordinated both to the C=C bond and to the oxygen atom
of 1d, as suggested by NMR data. Complex C, involving a
complexation only to the double bond of 1d, could also be
7
On the other hand, the partial charge on H of the hy-
droxyl group in 1d (+0.199) increases to +0.322 in complex
A and to at most +0.398 in transition state TS7. As a conse-
1
quence, oxygen atom O becomes more negatively charged
(from À0.330 in 1d to À0.543 in TS7). These data are indi-
cative of a strong increase in the acidity of the OH group on
going from 1d to TS7.
6362
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6356 – 6365

Cyclization of Unsaturated Alcohols
FULL PAPER
5
6
6
7
5
Elongation of the C ÀC bond in transition state TS7 is
shorter than the C ÀH distance, and the charge at C is
5
6
also observed with greater electrophilic character on C as
much more negative than at C , so that selective protonation
of the double bond at C is expected. In the second transi-
6
6
7
5
compared to C . In TS7, the C ÀH distance (1.323 Š) is
5
7
1
6
shorter than the C ÀH distance (1.923 Š), and this indicates
approach of OÀH to the double bond, selectively favouring
formation of five-membered cyclic ether 2d.
tion state TS12b, the O ÀC distance (2.044 Š) is shorter
1 5
than the O ÀC distance (2.647 Š; Table 4, entry 11). Bond
6
formation of the oxygen atom with C leads to 3a with recy-
For the less favoured cycloisomerization of 1d to tetrahy-
dropyran derivative 3d via TS9, the same general tendencies
cling of Al(OTf) in the catalytic cycle.
A
H
R
U
G
3
The combination of Brønsted and Lewis acids has been a
particularly useful tool for explaining several examples of
1
7
5
were observed concerning elongation of the O ÀH and C À
6
1
7
5
6
[23]
C bonds. The charges on O , H , C and C in transition
asymmetric catalysis. The design of such “combined acid
state TS9 are compatible with cyclization to 3d (Table 4,
systems”, which can be classified as Brønsted acid assisted
Lewis acids (BLA), Lewis acid assisted Lewis acids (LLA),
Lewis acid assisted Brønsted acids (LBA) and Brønsted acid
assisted Brønsted acid (BBA), enhance acidity by associa-
tive interaction and provide more organized structures that
lead to more effective asymmetric environments. For exam-
5
7
6
7
entry 5). The C ÀH distance is shorter than the C ÀH dis-
5
tance (not shown), and the OÀC distance is shorter than
6
the OÀC distance, that is, formation of the six-membered
cyclic ether 3d is favoured.
According to the calculated activation energies for forma-
À1
À1
tion of 2d (27.1 kcalmol , Figure 5) and 3d (30.1 kcalmol ,
Table 3, entry 9), the theoretical ratio 2d/3d as predicted by
the Arrhenius equation at 374.15 K (experimental tempera-
ture of 1018C) is of 98.8/1.2. This ratio is in good agreement
with the experimental data of 100/0 (Table 1, entry 4).
Theoretical studies allow a slightly different catalytic
ple, the combination of SnCl with an optically active bi-
4
naphthol derivative increased the acidity of the OH proton
in enantioselective protonation and enantioselective polyene
[24]
cyclization.
To explain the high catalytic activity of Al(OTf) in the
3
A
H
R
N
cycloisomerization of unsaturated alcohols, we propose a
Lewis acid assisted Brønsted acid situation (LBA) in which
cycle for the cycloisomerization of alcohol 1a with Al(OTf)
A
H
R
U
G
3
III
to be proposed, considering the complexation of Al to the
oxygen atom of the substrate. The mechanism involves a
nonconcerted pathway with two more favoured transition
states TS12a and TS12b (Table 3, entries 11, 12; Figure 8).
Cyclization of 1a was examined towards the formation of
the corresponding tetrahydrofuran and tetrahydropyran de-
rivatives 2a and 3a. Semiempirical calculations are clearly
in favour of formation of tetrahydropyran 3a. Since transi-
tion states TS11a and TS12a have the same energy and are
structurally very close, the selected pathway arises from the
the Lewis acid Al(OTf) strongly enhances the Brønsted
A
H
R
U
G
3
acidity of the hydroxyl protons of unsaturated alcohols.
Conclusion
In summary, we have demonstrated, for the first time, the
catalytic efficiency of aluminum(iii) trifluoromethanesulfo-
A
T
E
G
nate for the efficient, regioselective and atom-economic cy-
cloisomerization of unactivated unsaturated alcohols. Substi-
tuted tetrahydrofurans and tetrahydropyrans were selective-
ly and quantitatively obtained from g,d-unsaturated alco-
hols, depending on the substitution of the double bond, with
Markovnikov-type selectivity.
energy difference between transition states TS11b and
À1
TS12b. Indeed, while TS11b is 8.2 kcalmol
higher in
À1
energy than TS11a, TS12b is 3.8 kcalmol lower in energy
than TS12a.
The theoretical selectivity of 3a versus 2a calculated by
the Arrhenius equation at 313.15 K (408C) is strongly in
favour of six-membered cyclic ether 3a, with a 3a/2a ratio
of >99.99/<0.01. This ratio is also in excellent agreement
with experiment, in which tetrahydropyran 3a was exclu-
sively obtained (Table 1, entry 1). Therefore, only the mech-
anism for the cycloisomerization of 1a towards the six-mem-
bered cyclic ether 3a is presented in Figure 8.
The experimental results obtained in the presence of a
hindered base or water indicate that no inhibition is exhibit-
ed by these additives. Al(OTf) plays a particularly active
A
H
R
U
G
3
role as the catalyst in the intramolecular hydroalkoxylation
of unactivated olefins.
Theoretical and NMR complexation studies show that O
complexation to unsaturated alcohols 1a and 1d of Al-
A
H
R
U
G
Accordingly, the first step of the catalytic cycle involves,
as in Figure 8, complexation of the catalyst to the oxygen
atom of 1a to form intermediate complex A’. Complexes B’
and C’, in equilibrium with A’, can also be considered, al-
though theoretical calculations clearly point to complex A’
as the reactive intermediate.
with the double bond. Theoretical calculations indicated
lengthening of the OÀH bond in the transition states when
Al(OTf) was complexed to the oxygen atom of the unsatu-
A
H
R
U
G
3
rated alcohols. This interesting feature is an example of the
enhancement of the “acidity” of the hydroxyl hydrogen
atom by substrate coordination with a Lewis superacid. The
cycloisomerization of unsaturated alcohols catalyzed by Al-
In the first transition state TS12a, an increase in the the
1
7
O ÀH distance of 0.48 Š was calculated compared with 1a,
together with an increase of the positive charge on the hy-
droxyl hydrogen atom from +0.202 to +0.367 (Table 4, en-
tries 6 and 10). These data suggest the generation of a more
A
H
R
U
G
[23]
Lewis acid assisted Brønsted acid (LBA).
The proposed mechanistic cycles for the cycloisomeriza-
tion of alcohols 1a and 1d involve the formation of a com-
plex between the oxygen atom of the unsaturated alcohol
5
7
acidic hydroxyl proton. In TS12a, the C ÀH distance is
Chem. Eur. J. 2006, 12, 6356 – 6365
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6363

E. DuÇach, J.-M. Pons et al.
1
3
and Al
A
C
H
T
R
E
U
N
G
(OTf) with an increase in the acidity of the hydroxyl
6.1 Hz, 3H; CH
3
); C NMR (200 MHz, CDCl
3
): d=75.3, 67.7, 33.1, 25.9,
3
2
1.0 ppm.
proton in the first step. This proposed mechanism is differ-
ent from that recently proposed for platinum-catalyzed cy-
cloisomerization of unsaturated alcohols, in which coordina-
tion of platinum(ii) to the double bond was proposed in the
first step. Our studies constitute the first example of cy-
cloisomerization of unsaturated alcohols catalyzed by an Al-
centred Lewis acid, and include a mechanistic investigation
which supports the increased acidity of the hydroxyl proton
as the key to activating unactivated double bonds.
[
30]
1
2
-Methyl-2,3-dihydrobenzofuran (2e):
d=7.2–7.0 (m, 2H; ArH), 6.8–6.6 (m, 2H; ArH), 4.8 (ddq, J=8.8, 7.7,
.3 Hz, 1H; CHOR), 3.2 (dd, J=15.4, 8.8 Hz, 1H; CHH), 2.7 (dd, J=
5.4, 7.7, 1H; CHH), 1.4 ppm (d, J=6.3 Hz, 3H; CH
H NMR (200 MHz, CDCl ):
3
6
1
13
3
); C NMR
[13]
(200 MHz, CDCl
2
3
): d=159.2, 128.4, 127.3, 125.8, 120.1, 109.2, 79.1, 37.0,
+
+
2.2 ppm; MS (70 eV, EI): m/z (%): 134 (100) [M] , 133 (43) [MÀH] ,
+
+
+
1
19 (61) [MÀCH
3
] , 105 (24) [MÀC
2
H
5
] , 91 (78) [MÀC
3
H
7
] , 77 (24)
+
[
Ph] , 51 (22).
[
31]
1
3
,3,9,9-Tetramethyl-2,8-dioxaspiro
A
H
R
U
G
(3 f):
H NMR
(
200 MHz, CDCl
3
): d=3.45 (d, J=12.0 Hz, 2H; CHHOR), 3.35 (d, J=
2.0 Hz, 2H; CHHOR), 1.7–1.5 (m, 2H; CHHC(CH ), 1.4 (dd, J=6.3,
.6 Hz, 4H; CH ), 1.35–1.25 (m, 2H; CHHC(CH ), 1.12 (s, 6H; CH ),
.10 ppm (s, 6H; CH ); C NMR (200 MHz, CDCl ): d=71.9, 67.6, 32.2,
1.7, 28.2, 27.3, 25.9 ppm. MS (70 eV, EI): m/z (%): 212 (0.4) [M] , 197
] , 179 (4) [MÀCH
trans-3,5-Dimethyldihydrofuran-2-one (2g) (cis/trans 60:40):
1
5
1
3
A
H
R
U
G
3 2
)
2
A
H
R
U
G
3
)
2
3
1
3
3
3
Experimental Section
+
+
+
(
80) [MÀCH
3
5
O] , 138 (54), 109 (87), 67 (100).
Materials and methods: Reagents were obtained from Alfa Aeser and
Aldrich, and were used without further purification. Aluminum(iii) tri-
[32] 1
H NMR
A
H
R
N
(
200 MHz, CDCl
3
): d=4.6 (qt, J=6.4, 5.2 Hz, 1H; CHOR), 2.8–2.5 (m,
), 2.0–1.8 (m, 2H; CH ), 1.3 (d, J=6.4 Hz, 3H; CH ),
); C NMR (200 MHz, CDCl ): d=180.4,
5.0, 37.4, 34.4, 21.4, 16.1 ppm; MS (70 eV, EI): m/z (%): 114 (0.9) [M] ,
fluoromethanesulfonate is commercially available, but it can also be
easily prepared by electrosynthesis from triflic acid and an aluminium
1
1
7
9
H; CHCH
3
2
3
13
.2 ppm (d, J=7.3 Hz, 3H; CH
3
3
[
25]
1
13
rod.
H and C NMR studies were carried out with a Bruker Avance
+
1
1
13
5
00 operating at 500 MHz ( H). H and C NMR spectra of cyclic ethers
+
9 (5) [MÀCH
3
] , 55 (73), 45 (14), 44 (10), 43 (63), 42 (100), 41 (66).
1
were run with a Bruker AC 200 operating at 200 MHz ( H). GC analyses
were performed with a Varian CP 3380.
[
32]
1
cis-3,5-Dimethyldihydrofuran-2-one (2g) (cis/trans 60:40):
200 MHz, CDCl ): d=4.4 (qt, J=6.1, 5.2 Hz, 1H; CHOR), 2.8–2.5 (m,
H; CHCH ), 2.4 (ddd, J=12.0, 8.4, 5.2 Hz, 1H; CHHCH(CH )O), 1.4
ddd, J=12.0, 11.8, 10.4 Hz, 1H; CHHCH(CH )O), 1.3 (d, J=6.1 Hz,
), 1.2 ppm (d, J=6,8 Hz, 3H; CH ); C NMR (200 MHz,
): d=180.0, 75.3, 39.5, 36.8, 21.3, 15.5 ppm; MS (70 eV, EI): m/z
H NMR
(
3
General procedure for Al
A
H
R
U
G
3
-catalyzed hydroalkoxylation of unsatu-
(OTf)
23.7 mg, 0.05 mmol) in distilled nitromethane or dichloromethane
5 mL) was stirred at reflux for 0.5–3 h. The progress of the reaction was
1
3
A
H
R
U
G
3
rated alcohols: A mixture of unsaturated alcohol (1 mmol) and Al
(
(
A
H
R
U
G
3
(
A
H
R
U
G
3
13
3
H; CH
3
3
CDCl
3
monitored by GC analysis. The reaction mixture was quenched with HCl
+
+
(
%): 114 (1.3) [M] , 99 (8) [MÀCH
3
] , 55 (67), 45 (14), 44 (11), 43 (67),
(
1m) and extracted with Et
2
O. The organic layer was washed with HCl
4
2 (100), 41 (69).
(
0.1m), dried with MgSO and the solvent was evaporated. The products
4
were purified by silica-gel column chromatography with pentane/diethyl
1
13
ether 95:5 as eluent and analyzed by H and C NMR spectroscopy and
mass spectrometry, with comparison to data for reported compounds.
The regioselectivities 2:3 were calculated by GC and NMR analysis.
[1] a) G. Cardillo, M. Orena, Tetrahedron 1990, 46, 3321–3408; b) P. A.
Bartlett, Tetrahedron 1980, 36, 2–72; c) T. L. B. Boivin, Tetrahedron
1987, 43, 3309–3362; d) H. Kotsubi, Synlett 1992, 97–106; e) M. C.
Elliot, J. Chem. Soc. Perkin Trans. 1 2000, 1291–1318; f) M. C.
Elliot, E. Williams, J. Chem. Soc. Perkin Trans. 1 2001, 2303–2340.
[2] K. Tani, Y. Katoaka, Catalytic Heterofunctionalization (Eds.: A.
Togni, H. Grützmacher), Wiley-VCH, Weinheim, 2001, p. 171.
[3] L. Coulombel, I. Favier, E. DuÇach, Chem. Commun. 2005, 17,
2286–2288.
For the kinetic studies shown in Figure 1, the reactions were followed by
GC with n-undecane as internal standard. Reactions were repeated twice
with an estimated error of 3%. The ratio of 1c to undecane was 2:1. The
GC response factor of 1c to undecane (relative areas for a 1:1 molar
ratio and a typical FID detection) was 1.95, and that of 2c to undecane
was 1.02. Kinetic data for the cycloisomerization of 1c with Al
A
H
R
U
G
3
(
Figure 1) indicated a first-order reaction with respect to 1c with a rate
À4 À1
2
constant of k=7”10
s
(R =0.987).
[
26]
1
[4] D. D. DesMarteau, Science 2000, 289, 72–73.
2
3
3
,2,6-Trimethyltetrahydropyran (3a):
.7 (dqd, J=12.3, 6.2, 2.0 Hz, 1H; CH), 1.7–1.3 (m, 6H; (CH
H; CH ), 1.19 (s, 3H; CH ), 1.1 ppm (d, J=6.2 Hz, 3H; CH ); C NMR
): d=72.0, 66.7, 36.4, 33.8, 32.4, 23.1, 22.4, 20.5 ppm;
H NMR (200 MHz, CDCl
3
): d=
[
5] For recent reviews, see a) I. Nakamura, Y. Yamamoto, Chem. Rev.
2004, 104, 2127–2198; b) M. Beller, J. Seayad, A. Tillack, H. Jiao,
Angew. Chem. 2004, 116, 3448–3479; Angew. Chem. Int. Ed. 2004,
43, 3368–3398; c) J. Muzart, Tetrahedron 2005, 61, 5955–6008.
2 3
) ), 1.21 (s,
1
3
3
3
3
(
200 MHz, CDCl
3
+
+
MS (70 eV, EI): m/z (%): 128 (0.7) [M] , 113 (69) [MÀCH
3
] , 110 (1)
+
+
[6] a) F. Alonso, I. P. Beletskaya, M. Yus, Chem. Rev. 2004, 104, 3079–
159; b) J. Barluenga, A. DiØguez, F. Rodríguez, F. J. FaÇanµs, T.
[
MÀH
2
O] , 95 (9), 59 (100) [C
-Benzyl-2,2-dimethyltetrahydrofuran (2b):
): d=7.4–7.1 (m, 5H; ArH), 3.9 (dd, J=8.4, 6.7 Hz, 1H; CHHO),
.5 (dd, J=8.4, 7.8 Hz, 1H; CHHO), 2.7–2.5 (m, 1H; CH, CH Ph), 1.9
dd, J=12.3, 6.9 Hz, 1H; CHHC(CH ), 1.5 (dd, J=12.3, 8.3 Hz, 1H;
CHHC(CH ), 1.3 (s, 3H; CH ), 1.2 ppm (s, 3H; CH
200 MHz, CDCl ): d=141.3, 129.0, 128.8, 126.4, 81.3, 72.6, 45.6, 42.1,
3 7
H O] .
3
[
27]
1
4
H NMR (200 MHz,
Sordo, P. Campomanes, Chem. Eur. J. 2005, 11, 5735–5741.
7] A. Hoffmann-Rçder, N. Krause, Org. Lett. 2001, 3, 2537–2538.
8] a) P. W. Jolly, N. Kokel, Synthesis 1990, 771–773; b) M. Utsunomiya,
M. Kawatsura, J. F. Hartwig, Angew. Chem. 2003, 115, 6045–6048;
Angew. Chem. Int. Ed. 2003, 42, 5865–5868.
CDCl
3
(
3
[
[
2
A
H
E
N
3 2
)
1
3
A
C
H
T
R
E
U
N
G
3
)
2
3
3
); C NMR
(
3
3
[
9] a) T. Hosokawa, T. Shinohara, T. Ooka, S.-I. Murahashi, Chem. Lett.
+
9.9, 29.4, 28.6 ppm; MS (70 eV, EI): m/z (%): 190 (2) [M] , 175 (44)
1
1
989, 2001–2004; b) S. Ganguly, D. M. , Roundhill, Organometallics
993, 12, 4825–4832; c) K. J. Miller, T. T. Kitagawa, M. M. Abu-
+
+
+
[
(
MÀCH
3
] , 157 (20), 134 (8) [MÀC
4
H
8
] , 117 (22) [CH
A
T
E
N
(CH
2
)
2
Ph] , 104
+
+
+
+
] .
10) [CHCH
2
Ph] , 91 (66) [CH
2
Ph] , 77 (5) [Ph] , 43
A
H
R
U
G
3
H
7
Omar, Organometallics 2001, 20, 4403–4412.
[
28] 1
2
(
-Hexyltetrahydrofuran (2c):
m, 3H; CH, CH ), 1.9–1.7 (m, 3H; CHHCH
1H; CHHCH(C 13), (CH ), 0.8 ppm (t, J=6,8 Hz, 3H; CH
C NMR (200 MHz, CDCl ): d=79.9, 68.0, 36.2, 32.2, 31.8, 29.8, 26.8,
H NMR (200 MHz, CDCl
3
): d=3.9–3.6
), 1.5–1.1 (m,
);
[
[
10] E. Marotta, E. Foresti, T. Marcelli, F. Peri, P. Righi, N. Scardovi, G.
Rosini, Org. Lett. 2002, 4, 4451–4453.
11] a) K. Hori, H. Kitagawa, A. Miyoshi, T. Ohta, I. Furukawa, Chem.
Lett. 1998, 11, 1083–1084; b) Y. Oe, T. Otha, Y. Ito, Chem.
Commun. 2004, 1620–1621; c) Y. Oe, T. Otha, Y. Ito, Synlett 2005, 1,
2
A
H
R
U
G
6
H
2
1
A
C
H
T
R
E
U
N
G
6
H
2
)
5
3
1
3
3
+
2
6.1, 23.0, 14.5 ppm; MS (70 eV, EI): m/z (%): 156 (0.2) [M] , 138 (2)
+
+
+
] .
[
MÀH
2
O] , 71 (100) [MÀC
6
H
13] , 43 (34) [C
3
H
7
1
79–198.
[
29] 1
2
-Methyltetrahydrofuran (2d):
H NMR (200 MHz, CDCl
m, 3H; CH, CHH), 3.7 (ddd, J=8.1, 7.7, 6.4 Hz, 1H; CHH), 2.1–1.8 (m,
3
): d=4.0–3.8
[12] a) C.-G. Yang, N. W. Reich, Z. Shi, C. He, Org. Lett. 2005, 7, 4553–
4556; b) C.-G. Yang, C. He, J. Am. Chem. Soc. 2005, 127, 6966–
6967.
(
3
H; CHHCHCH
3
, CH
2
3
), 1.5–1.3 (m, 1H; CHHCHCH ), 1.2 ppm (d, J=
6364
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6356 – 6365

Cyclization of Unsaturated Alcohols
FULL PAPER
[
[
[
13] H. Qian, X. Han, R. A. Widenhoefer, J. Am. Chem. Soc. 2004, 126,
536–9537.
14] M. L. Mihailovic, Z. Petrovic, A. Teodorovic, S. Konstantinovic, V.
Andrejevic, C.R. Acad. Sci. Paris 1989, 308, 29–32.
15] J. A. March, Advanced Organic Chemistry: Reactions, Mechanisms
and Structures, 3rd ed., Wiley Interscience, New York, 1985,
pp. 634–636.
from the O complex (Table 3, entries 7 and 9). Indeed, if the reac-
tion is to proceed, the chelate must be cleaved, and due to the rela-
tive strengths of complexation (p complexation vs O complexation),
the Lewis acid remains on the oxygen atom. From these results, we
propose to explain the NMR chemical shifts observed for the ethyl-
enic carbon atoms either as resulting from the formation of this che-
late B or from the C=C complex C. Both complexes can not lead to
the product of the reaction, but since their formation is reversible,
they can exist and therefore be observed by spectroscopy.
[23] H. Yamamoto, K. Futatsugi, Angew. Chem. 2005, 117, 1958–1977;
Angew. Chem. Int. Ed. 2005, 44, 1924–1942.
9
[
[
[
16] a) T. C. Wabnitz, J-Q. Yu, J. B. Spencer, Chem. Eur. J. 2004, 10, 484–
4
93.
17] M. J. S. Dewar, E. G. Zoebish, E. F. Healy, J. J. P. Stewart, J. Am.
Chem. Soc. 1985, 107, 3902–3909.
18] All calculations were performed with AMPAC 8.16, Semichem, Inc.,
PO Box 1649, Shawnee, KS 66222, USA.
[24] H. Ishibashi, K. Ishihara, H. Yamamoto, Chem. Rec. 2002, 2, 177–
188.
[
[
19] D. A. Liotard, Int. J. Quantum Chem. 1992, 44, 723–741.
20] J.-M. Pons, M. Oblin, A. Pommier, M. Rajzmann, D. Liotard, J. Am.
Chem. Soc. 1997, 119, 3333–3338.
[25] a) E. Dunach, I. Favier, D. Hebrault, J-R. Desmurs (Rhodia
Chimie), FR 2818994, 2000; [Chem. Abstr. 2002, 137, 239065]; b) I.
Favier, E. Dunach, Tetrahedron Lett. 2003, 44, 2031–2032.
[26] G. Carr, D. Whittaker, J. Chem. Soc. Perkin Trans. 2 1989, 359–366.
[27] J. Diab, M. Abou-Assali, C. Gervais, D. Anker, Tetrahedron Lett.
1985, 26, 1501–1502.
[
21] The introduction of a second molecule of alcohol 1d slightly in-
À1
creased the activation energies: from 61.9 to 64.7 kcalmol in the
À1
case of the uncatalyzed reaction, and from 27.2 to 28.7 kcalmol in
the case of the Al
22] Since NMR experiments indicated a certain complexation of the CÀ
C double bond with Al(OTf) (as in complex B, Figure 3), we also
A
C
H
T
R
E
U
N
G
(OTf)
3
-catalyzed reaction.
[28] J. Kula, T. B. Quang, M. Sikora, Tetrahedron: Asymmetry 2000, 11,
943–950.
[29] Data compared to commercial sample.
[
A
H
R
U
G
3
examined the possibility of a reaction path starting from a chelate in
which the aluminium atom of the catalyst would be complexed both
to the oxygen atom and the CÀC double bond of the starting mole-
[30] V. H. Grant, B. Liu, Tetrahedron Lett. 2005, 46, 1237–1239.
[31] B. K. Wasson, J. M. Parker, US 3099662, 1963 [Chem. Abstr. 1963,
60, 9784].
[32] J. L. Garcia Gutierrez, F. Jimenez-Cruz, N. R. Espinosa, Tetrahedron
Lett. 2005, 46, 803–805.
cule. In order to do so, we assumed dismutation of Al
A
H
R
U
3
and
+
concomitant formation of Al(OTf)
2
A
H
R
U
G
+
late (1d-Al
A
C
H
T
R
E
U
N
G
(OTf)
2
Received: November 29, 2005
find an original reaction path leading to tetrahydrofuran 2d with a
Revised: March 2, 2006
comparable or even lower activation energy than that found starting
Published online: June 1, 2006
Chem. Eur. J. 2006, 12, 6356 – 6365
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6365