Aluminium triflate: A remarkable Lewis acid catalyst for the ring opening of epoxides by alcohols

Article abstract of DOI:10.1039/b508924g

Al(OTf)₃ was found to be an extremely effective catalyst (at ppm levels) for ring opening reactions of epoxides using a range of alcohols. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b508924g

View Article Online / Journal Homepage / Table of Contents for this issue
C O M M U N I C A T I O N
Aluminium triflate: a remarkable Lewis acid catalyst for the ring
opening of epoxides by alcohols†
D. Bradley G. Williams* and Michelle Lawton
Department of Chemistry, University of Johannesburg, P.O. Box 524, Auckland Park, 2006.
E-mail: dbgw@rau.ac.za; Fax: +27(0)11 489 2819; Tel: +27(0)11 489 3431
Received 23rd June 2005, Accepted 5th August 2005
First published as an Advance Article on the web 17th August 2005
Al(OTf)3 was found to be an extremely effective catalyst (at
ppm levels) for ring opening reactions of epoxides using a
range of alcohols.
of aryl-substituted epoxides to aldehydes and ketones.3 Yttrium
triflate, on the other hand, was used to catalyse the addition of
lithium enolates to epoxides to form c-hydroxy ketones,4 and
ytterbium(III) triflate catalyses highly regioselective allylation of
aromatic epoxides using allyltributyltin in THF.5
From our review of the literature, we found that very little
work has been done using aluminium triflate. This study,
therefore, involved an investigation into the efficacy of Al(OTf)3
as a Lewis acid for the catalysed ring opening of epoxides by
selected alcohols.
Initially, we conducted our experiments using styrene oxide.
Using the protocol set out by Green in a 1987 patent,6 six
equivalents of methanol or ethanol (with no added solvent) and
0.1% of Al(OTf)3 were added to styrene oxide and the mixture
was heated under reflux for one hour. The reaction afforded
excellent yields of the glycol ether where the nucleophile had
attacked at the more hindered carbon atom of the epoxide ring
(Scheme 1).
Epoxides are important molecules in organic synthesis and,
because of their versatility and reactivity with a range of nucle-
ophiles, are often used as starting materials and intermediates.1
In general, ring opening reactions of epoxides can be catalysed
under basic or acid conditions. The former takes place through
an SN2 type reaction and attack of the nucleophile typically
occurs at the less hindered carbon.2 Under acidic conditions,
the ring opening of the epoxide resembles an SN1 reaction,
sometimes called a borderline SN2 reaction.2 As a result, the
nucleophile tends to attack at the more highly substituted carbon
atom of the epoxide ring, which bears more of the positive
charge and resembles a stable carbocation species.2 Therefore,
the formation of one product is highly favoured under acidic
conditions.2
Recently, metal triflates have received much attention for their
role as Lewis acids in a number of reactions. Bismuth triflate
has been used as a catalyst in the ring opening of epoxides
by aromatic amines under aqueous conditions, and the corre-
sponding b-amino alcohols were afforded in excellent yields.1
Bismuth triflate has also been employed in the rearrangement
† Electronic supplementary information (ESI) available: Characterisa-
tion data. See http://dx.doi.org/10.1039/b508924g
Scheme 1
Table 1 Yields (%) of products obtained from reactions with selected epoxides in various alcoholsa
% Yield
0.0005% cat.
% Yield
0.001% cat.
% Yield
0.002 cat.
% Yield
0.003% cat.
% Yield
0.004% cat.
Entry
Product
b
1
2
3
4
5
6
7
8
2 R = Me
2 R = Et
0
94
93
91
—
—
—
—
1
95
97
92
14
0
77
41 (34)
98
94
—
—
—
96 (4)
—
—
–
—
—
—
—
92 (8)
85
81
—
2 R = nPr
2 R = iPr
2 R = nBu
2 R = 2-Bu
2 R = tBu
92 (4)c
—
97
51
77
—
—
9
—
—
31 (24)
11 (7)
—
—
—
—
—
—
10
a Products were characterised using IR, MS, 1H-, 13C- and 2D NMR and GC-FID analysis. See reference 9 for experimental details. b Reactions not
performed. c Yields in parentheses refer to the other regioisomer.
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 2 6 9 – 3 2 7 2
3 2 6 9
©

View Article Online
Table 2 Yields (%) of products obtained from reactions with selected epoxides in various alcoholsa
% Yield
0.001% cat.
% Yield
0.002% cat. 0.003% cat.
% Yield
% Yield
0.004% cat. 0.01% cat.
% Yield
% Yield
0.02% cat. 0.04% cat.
% Yield
Entry Product
1
b
85
77
55
21
—
—
—
—
86
83
88
42
89
45
3
–
—
—
—
—
—
—
—
7
—
—
—
—
—
—
18
14
—
—
—
—
—
—
—
21
—
—
—
—
—
—
—
49
2
3
4
5
6
7
8
—
—
62
—
—
—
—
—
9
—
—
—
—
—
—
63
a Products were characterised using IR, MS, 1H-, 13C- and 2D NMR and GC-FID analysis. See reference 9 for experimental details. b Reactions not
performed.
Satisfied that we were able to repeat the experimental proce-
dures set out in the patent literature, we changed the reaction
conditions in order to determine the catalytic ability of Al(OTf)3,
the effects of different alcohols and, finally, the effect of different
epoxides on the reaction.
Firstly the effect of temperature on the outcomes of the
reactions was investigated. Reactions with styrene oxide and 6
equivalents of MeOH were carried out at 0 ◦C, room temperature
and under reflux for 24 hours and samples were taken and
analysed by GC-FID at regular intervals. Al(OTf)3 was added
at 0.001%, 0.002%, 0.003% and 0.004%,7 respectively. In each
case, for benchmark purposes, the reactions were also performed
without added catalyst. The results demonstrated selectivity for
the isomer where attack had occurred at the more hindered
carbon, in agreement with prior work.2 (The formation of
what is believed to be an oligomer was also seen.8) At 0 ◦C,
0.004% of Al(OTf)3 was needed before a 75% conversion to
product 2 was observed. Under these conditions, 7% of the
putative oligomer was formed after 24 hours. In contrast,
at room temperature, 0.003% of catalyst was required before
85% conversion to 2 was seen and 6% oligomer was formed
after only 18 hours. The best conversion observed was under
reflux, at which temperature 98% conversion to product 2
was accompanied with only 2% oligomer formation. This was
achieved with only 0.002% of the catalyst. The benchmark
reactions performed with no catalyst yielded only 5% of the
product together with 0.6% of the oligomer after 24 hours under
reflux.
3 2 7 0
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 2 6 9 – 3 2 7 2

View Article Online
As discussed above these reactions tended to form the
oligomer by-product and, in order to prevent this, subsequent
reactions with styrene oxide were carried out using twenty equiv-
alents of alcohol. This change minimised oligomer formation to
only trace amounts.
In general, the reactions with the bulkier alcohols required
more Al(OTf)3 catalyst to be added before appreciable yields
of products were obtained. Nonetheless, acceptable yields were
achievable for all of the reactions using only ppm levels of the
triflate.
We also investigated the effects of different alcohols, epoxides
and how different catalyst concentrations would influence these
reactions (Table 1).9 In all cases, except for those performed in
the presence of methanol, reactions of styrene oxide required
less catalyst to be added than the analogous reactions with
other epoxides in order for an appreciable amount of product
to be formed. The reactions with styrene oxide were remarkably
regioselective, with the formation of only a single regioisomer
being detected, namely that where attack of the nucleophile had
taken place solely at the more hindered position of the epoxide.
In these reactions, electronic effects dominate over steric effects:
the intermediate carbocation at the secondary carbon atom
is stabilised through resonance with the phenyl ring, thereby
promoting nucleophilic attack at this position.
Triflic acid is often found in residual amounts in the alu-
minium triflate used. If triflic acid is present in these reactions,
it may be capable of catalysing the ring opening of the epoxide.
To test this potential, reactions were performed using triflic acid
as the only catalyst, and the results showed that the added triflic
acid had a negligible influence. The experimental procedure
previously employed was followed, but using triflic acid instead
of Al(OTf)3. After 24 hours under reflux in the presence of
0.005% triflic acid only 4% of product 2 had formed. We
could therefore conclude that the catalytic activity observed was
ascribable to the Al(OTf)3.
In order to test Al(OTf)3 on epoxides containing different
structural motifs, glycidyl ethers also were used (Scheme 2), in
ethanol, with some surprising results (Table 3).
In contrast, butylene oxide provided an almost 50 : 50 ratio of
isomers, where the nucleophile had attacked at either end of the
epoxide. In these reactions neither steric nor electronic effects
played a more dominant role.
In contrast to the previous results discussed, the main product
formed in these reactions was the 2◦ alcohol analogue, where
the nucleophile strongly favoured attack at the less hindered
carbon atom of the epoxide ring. Presumably, this is because
the Al(OTf)3 formed a chelate structure with the two oxygen
atoms of the gylcidyl ether (Fig. 1). This would decrease the
Lewis acidic (electron-withdrawing) effect that the metal would
usually have on the internal carbon atom, allowing steric effects
Cyclohexene oxide required only 0.001% of the catalyst
(0.0005% failed to promote the reaction) to be added to afford
an 85% or 77% yield of the monoglycol ether in methanol
and ethanol, respectively (Table 2). However, the reactions
with cyclododecane epoxide required forty times the amount of
Al(OTf)3 to be added before a 49% conversion to the monoglycol
ether was seen when the reaction was performed in ethanol. We
believe that this was due to the large, flexible cyclododecane
molecule folding over on itself (hydrophobic effects) in the polar
solvent (ethanol) making nucleophilic attack difficult. In order
to test this hypothesis we repeated the reaction using the less
polar solvent 1-butanol, which improved the yield to 63% after
one hour under similar conditions.
Fig. 1
Scheme 2
Table 3 Yields (%) of products obtained from reactions with glycidyl ethers and ethanola
% Yield
0.002% cat.
% Yield
0.004% cat.
% Yield
0.005% cat.
% Yield
0.01% cat.
% Yield
0.02% cat.
% Yield
0.04% cat.
Entry
Product
b
1
2
3
4
4 R = allyl
5 R = allyl
4 R = tBu
5 R = tBu
4 R = Ph
5 R = Ph
6
22
1
–
5
18
10
28
—
—
—
14
49
18
54
7
41
35
19
69
21
63
10
60
26
—
—
—
—
12
80
—
—
—
—
0
0
30
4
—
—
—
5
—
—
—
24
36
60
a Products were characterised using IR, MS, 1H-, 13C- and 2D NMR and GC-FID analysis. See reference 9 for experimental details. b Reactions not
performed.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 2 6 9 – 3 2 7 2
3 2 7 1

View Article Online
to play a more dominant role. The nucleophile therefore attacked
at the less hindered carbon atom and the 2◦ alcohol was formed.
The glycidyl ethers required more Al(OTf)3 to be used in
the reactions, when compared with those of styrene oxide and
cyclohexene oxide, to observe similar rates of reactions. This is
also probably due to the decreased Lewis acid effect of the metal
because of the chelate structure that formed. At lower catalyst
concentration the reactions with 1,4-butanediol glycidyl ether
yielded a significant proportion of the mono glycol ether. Only
when 0.04% Al(OTf)3 was added was the diglycol ether formed
as the major product.
3 K. A. Bhatia, K. J. Eash, N. M. Leonard, M. C. Oswald and R. S.
Mohan, Tetrahedron Lett., 2001, 42, 8129–8132.
4 P. Crotti, V. Di Bussolo, L. Favero, F. Macchia and M. Pineschi,
Tetrahedron Lett., 1994, 35, 6537–6540.
5 P. R. Likhar, M. P. Kumar and A. K. Bandyopadhyay, Tetrahedron
Lett., 2002, 43, 3333–3335.
6 M. J. Green, US pat. 4,687,755, 1987.
7 In all of the reactions the low levels of Al(OTf)3 were readily obtainable
making use of stock solutions in either DME or the alcohol in question
and then adding the correctly calculated volume of the solution to the
reaction mixture. It was found that there was no difference in the
activity of the stock solution due to a solvent effect (DME stock
solution vs. alcohol stock solution). In addition, the solutions were
found to be stable for over six months, providing identical reactivity
after having stood for that period of time.
8 The oligomer was tentatively identified as a trimer of the ring-opened
product, containing one methoxy residue and three incorporated
styrene oxide units, on the basis of 1H- and 13C NMR studies, together
with MS analysis.
9 Typical experimental procedure; 2 mL (17.5 mmol) of styrene oxide
were mixed with 20 eq. of the alcohol in question (0.35 mol). The
desired amount of Al(OTf)3 stock solution (see ref. 7) was then added
and the mixture was allowed to stir at the reflux temperature of
the mixture or 100 ◦C (oil bath temperature), whichever was the
lower, for one hour, after which the solution was filtered through
silica and analysed by GC-FID. Products were isolated using column
chromatography and/or vacuum distillation.
In summary, Al(OTf)3 was found to be a highly effective Lewis
acid catalyst (at ppm levels) for the ring opening of a variety
of epoxides by a range of alcohols, providing the anticipated
products in very high yields with often remarkable selectivity.
Acknowledgements
We gratefully acknowledge generous funding from SASOL,
NRF South Africa, and THRIP.
Notes and references
1 T. Ollevier and G. Lavie-Compin, Tetrahedron Lett., 2004, 45, 49–52.
2 R. E. Parker and N. S. Isaacs, Chem. Rev., 1959, 737.
3 2 7 2
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 2 6 9 – 3 2 7 2