On the role of triflic acid in the metal triflate-catalysed acylation of alcohols

Article abstract of DOI:10.1016/j.tetlet.2003.11.034

The acylation of alcohols by anhydrides, catalysed by a wide range of metal triflates, is a powerful and mild method for the preparation of a variety of esters. Mechanistic insights demonstrate that triflic acid is generated under these reaction conditions and that, at least, two competing catalytic cycles are operating at the same time: a rapid one involving triflic acid and a slower one involving the metal triflate.

Full text of DOI:10.1016/j.tetlet.2003.11.034

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 825–829
On the role of triflic acid in the metal triflate-catalysed acylation
of alcohols
€
ꢀ
ꢀ*
Raphael Dumeunier and Istvan E. Marko
Deꢀpartement de Chimie, Uniteꢀ de Chimie Organique et Meꢀdicinale, Universiteꢀ Catholique de Louvain, B^atiment Lavoisier,
Place Louis Pasteur 1, B-1348 Louvain-la-Neuve, Belgium
Received 6 October 2003; revised 29 October 2003; accepted 6 November 2003
Abstract—The acylation of alcohols by anhydrides, catalysed by a wide range of metal triflates, is a powerful and mild method for
the preparation of a variety of esters. Mechanistic insights demonstrate that triflic acid is generated under these reaction conditions
and that, at least, two competing catalytic cycles are operating at the same time: a rapid one involving triflic acid and a slower one
involving the metal triflate.
Ó 2003 Elsevier Ltd. All rights reserved.
The acylation of an alcohol by an anhydride, to form the
corresponding ester, is a key synthetic transformation
that can be effected under numerous acidic or basic
conditions.1 Catalytic procedures based upon the use of
metal triflates have recently received a lot of attention.2
These readily available Lewis acids promote the esteri-
fication reaction efficiently and under mild conditions.
During the course of the total synthesis of polycav-
ernoside A,3 a potent marine toxin, the transformation
of the sensitive b-hydroxy sulfone 1 into the corre-
sponding benzoate 2 was required (Scheme 1).
of 5, suggested that residual triflic acid might be the
culprit responsible for the deprotection of the silyl ether.
In order to neutralise these traces of triflic acid, 2,6-
di-tert-butyl-4-methyl pyridine (DTBMP), a highly
hindered organic base that does not interact with the
metal catalyst, was added to the reaction mixture.6
Much to our surprise, no benzoylation of 5 was
observed, even after heating at 50 °C. This observation
strongly suggests that triflic acid plays a prominent role
in these metal-catalysed acylations. In order to shed
some light on the nature of the catalytically active spe-
cies, the benzoylation of the model alcohol 7, using
various metal triflates, in the absence (Table 1) and
presence (Table 2) of DTBMP was investigated.
Initial attempts using base-promoted benzoylation
failed to deliver the desired compound due to the rapid
formation of the elimination product, trienyl sulfone 3.4
The sensitivity of 1 towards basic conditions prompted
us to explore the Sc(OTf)3-catalysed acylation of alco-
hols. Surprisingly, when 10 mol % of Sc(OTf)3 were
added to an acetonitrile solution of b-hydroxy sulfone 1
and benzoic anhydride, immediate deprotection of the
TES substituent ensued, affording quantitatively diol 4.5
In stark contrast, Sc(OTf)3-catalysed benzoylation of
the simpler b-hydroxy sulfone 5 resulted in the smooth
formation of 6 in 96% yield.
As can be seen from Table 1, the conversion of alcohol 7
into benzoate 9 proceeded in general with modest to
good yields, depending upon the nature of the metal
employed. Ytterbium triflate proved to be the best cat-
alyst, affording complete conversion to the desired
product after 25 h (Table 1, entry 1). In stark contrast,
copper triflate barely catalysed the acylation reaction,
providing 9 in only 3% yield after 25 h (Table 1, entry 5).
Scandium, indium and bismuth triflate displayed inter-
mediate behaviour, with scandium being the most
reactive of the three complexes.
The unexpected desilylation of 1 in the presence of the
scandium catalyst, coupled with the successful acylation
Addition of 15 mol % of DTBMP (3 equiv per metal
triflate)7 resulted in a dramatic effect, with the benzoyl-
ation reaction being almost completely suppressed. As
can be seen from Table 2, ytterbium triflate still proved
to be the most active catalyst, generating 9 in 11% yield
Keywords: Triflic acid; Acylation; Metal triflate.
* Corresponding author. Tel.: +32-10-47-8773; fax: +32-10-47-2788;
e-mail: marko@chim.ucl.ac.be
0040-4039/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.11.034

826
R. Dumeunier, I. E. Markꢀo / Tetrahedron Letters 45 (2004) 825–829
OTES
PhSO2
OTES
OBz
a
b
OTES
OH
3
30%
OH
OH
> 95%
PhSO2
PhSO2
2
1
PhSO2
4
OH
OBz
c or d
PhSO2
PhSO2
5
6
Scheme 1. Reagents and conditions: (a) BzCl, DMAP, Et3N, DCM (1/2/3 ¼ 1:1:1); (b) Sc(OTf)3 10%, Bz2O, CH3CN, 5 min; (c) Sc(OTf)3 (10%),
Bz2O (3 equiv), CH3CN, 5 min, 20 °C––96% yield; (d) Sc(OTf)3 (1 equiv), Bz2O (3 equiv), DTBMP (1 equiv), CH3CN, 24 h, 50 °C––traces.
Table 1. Catalytic benzoylation without added basea
O
Bz2O 8 (3 eqs.)
OH
O
Ph
Cat., MeCN
7 (1 eq.)
9
Entry
Cat. (5%)
Conversion (%)/ha
7
0.67
2.5
25 (h)
1
2
3
4
5
Yb(OTf)3
Bi(OTf)3
In(OTf)3
Sc(OTf)3
Cu(OTf)2
4
8
19
12
15
46
0
51
17
99
40
34
63
3
5
24
51
38
0
Traces
a Measured by capillary GC after calibration of the response for each component.
Table 2. Catalytic benzoylation in the presence of DTBMPa
O
Bz2O 8 (3 eqs.)
OH
O
Ph
DTBMP, Cat., MeCN
7 (1 eq.)
9
Entry
Cat. (5%)
DTBMP (%)
Conversion (%)/ha
0.67
15 (h)
1
2
3
4
5
Yb(OTf)3
Bi(OTf)3
In(OTf)3
Sc(OTf)3
Sc(OTf)3
15
15
15
15
5
Traces
Traces
Traces
Traces
Traces
11
8
3
2
6
a Measured by capillary GC after calibration of the response for each component.
after 15 h (Table 2, entry 1). Scandium triflate, initially
the second best in the absence of base, barely catalysed
the reaction, affording 9 in a meagre 2% yield after 15 h.
In this case, lowering the amount of DTBMP to 5 mol %
led to a marginal improvement in the production of 9
(6% after 15 h). These results are collected in Graph 1.
the benzoylation of 7 was repeated using variable
amounts of TfOH under otherwise identical reaction
conditions. Addition of 8 mol % of triflic acid to an
acetonitrile solution of alcohol 7 and benzoic anhydride
resulted in the formation of ester 9 (20% conversion)
after a few minutes. The reaction then decelerates and
only a 30% conversion of 7 is observed after 5 h. How-
ever, when a second portion (8 mol %) of TfOH was
added, a rapid transformation of alcohol 7 into ester 9
In order to further verify our hypothesis that triflic acid
might be an active catalytic species in these acylations,

R. Dumeunier, I. E. Markꢀo / Tetrahedron Letters 45 (2004) 825–829
827
tonitrile was prepared and added to a mixture of alcohol
7 and benzoic anhydride.9 As expected, only a mediocre
yield of 9 was obtained (21% after 30 min) and the
reaction essentially stopped; no further increase in the
conversion of 7 to 9 was observed after several more
hours.
Sc(OTf)3
100
90
80
70
60
50
40
30
20
10
0
Bi(OTf)3
In(OTf)3
Yb(OTf)3
Sc(OTf)3 / DTBMP (1/3)
Bi(OTf)3 / DTBMP (1/3)
In(OTf)3 / DTBMP (1/3)
Yb(OTf)3 / DTBMP (1/3)
Sc(OTf)3 / DTBMP (1/1)
Taken all together, these results strongly support the
intervention of triflic acid as an active catalytic species in
these benzoylation reactions and demonstrate that water
plays an important role in modulating its activity.10
A
plausible mechanistic rationale is depicted in Figure 1.
Upon addition of a metal triflate to a mixture of alcohol
14 and benzoic anhydride 8, activation of the anhydride
occurs, leading to the formation of the mixed triflic–
benzoic species 12.11 Acylation of alcohol 14 by 12 then
produces the desired ester 15 and releases triflic acid,
which might react according to two different and com-
petitive pathways. In the first catalytic cycle, triflic acid
displaces the benzoate ligand from the mixed metal
species 10, regenerating the corresponding metal triflate,
which initiates a new catalytic cycle (Cycle 1). Alterna-
tively, triflic acid might react directly with benzoic
anhydride to produce the mixed acylating derivative
12.12 Upon condensation with alcohol 14, TfOH is
regenerated, completing Cycle 2.
Time (h)
0
2
4
6
8
10 12 14 16 18 20 22 24 26
Graph 1.
ensued (82% conversion). After 20 h, complete disap-
pearance of the starting material was observed.
Careful purification of the product afforded the desired
ester 9, albeit in only 68% yield, despite the complete
conversion of alcohol 7. This moderate yield results
from the competitive dehydration of substrate 7 by tri-
flic acid, affording volatile 1,3-hexadiene and water.8
Besides triflic acid and the metal triflate, an important
intermediate in this proposed catalytic cycle appears to
be the acyl triflate 12, which participates in both Cycles
1 and 2. In order to verify the involvement of 12 in this
acylation reaction and, at the same time, provide further
evidence for our postulated mechanism, we decided to
prepare 12 independently13 and employ it as the sole
These observations suggested that water, generated by
this side reaction, might be responsible for the inhibition
of the triflic acid catalytic activity, probably by buffering
its acidity. Therefore, a 2:1 H2O/TfOH solution in ace-
ROH
14
(PhCO)2O
PhCOOR
15
8
O
PhCOOH
Ph OTf
M(OTf)2(OCOPh)
16
12
10
O
TfOH
13
Ph OTf
Cycle 1
Cycle 2
12
ROH
14
(PhCO)2O
M(OTf)3
8
11
M(OTf)2OCOPh
PhCOOR
15
PhCOOH
10
16
Figure 1.
O
BzOTf 12 (8 + 7%)
CH3CN
Bz2O
+
OH
O
Ph
7 (1 eq)
9
64%
Scheme 2.

828
R. Dumeunier, I. E. Markꢀo / Tetrahedron Letters 45 (2004) 825–829
To shed some light on the relative rates between Cycle 2
O
a
and the first part of Cycle 1 (the key steps of this cata-
lytic cycle), the stoichiometric acetylation of 7 was
investigated (Scheme 3).
OH
O
(12%; 10 min)
(17%; 60 min)
(42%; 60 h)
7 (1 eq)
18
In the presence of 3 equiv of acetic anhydride, 1 equiv of
Sc(OTf)3 and 3 equiv of DTBMP, alcohol 7 underwent
slow acetylation, affording acetate 18 in 12% conversion
after 10 min, 17% after 60 min and 42% after 60 h. Since
DTBMP completely inhibits both Cycle 2 and the sec-
ond part of Cycle 1 (by removing triflic acid), the con-
versions obtained during this experiment reflect the
ability of the metal triflate to catalyse the acylation of
alcohols.16
Scheme 3. Reagents and conditions: (a) Sc(OTf)3 (1 equiv), Ac2O 17
(3 equiv), DTBMP (3 equiv), MeCN.
catalyst in the benzoylation of alcohol 7. Under these
conditions, both the yield and rate of acylation of 7
should be identical to those obtained using TfOH
(Scheme 2).
Gratifyingly, and in perfect agreement with our postu-
lated mechanism, the addition of 12 to alcohol 7 and
Bz2O resulted in the smooth formation of the benzoate 9
after 22 h, at room temperature (100% conversion, 64%
yield).14 These results unambiguously establish the
intermediacy of benzoyl triflate 12 in Cycle 2 and pro-
vide yet another entry into this catalytic manifold.
From this experiment and the above-mentioned data,
catalysis by triflic acid appears to be the dominant
pathway in the absence of the hindered base and the
metal triflates act as selective reservoirs of triflic acid,
releasing it rapidly at the onset of the reaction. This is
also evidenced by the shape of the conversion versus time
curves (Graph 1) where a fast rate of reaction is initially
observed (which mimics the activity of triflic acid alone),
followed by a more sluggish reaction, eventually culmi-
nating in the complete conversion of 7 into 9.
This mechanism also rationalises the powerful inhibi-
tory activity of DTBMP. Indeed, addition of DTBMP
to the reaction mixture results in the removal of triflic
acid––which holds a cardinal position at the junction
between the two catalytic cycles––and the concomitant
suppression of Cycle 2 and the second part of Cycle 1
(regeneration of the metal triflate catalyst). Only the
initial two steps of Cycle 1 are still operative under these
conditions.15
In summary, we have shown that the metal triflate
promoted acylation of alcohols appears to be catalysed
mainly by triflic acid, released at the onset of the reac-
tion. Two catalytic cycles compete in this process. Cycle
2, which involves directly triflic acid, appears to pre-
dominate at the beginning of the reaction. Care must
therefore be exercised when using acid sensitive sub-
strates and when attempting to rationalise the particular
selectivities often observed using metal triflates.
In order to verify that these mechanistic explanations
were not limited to the benzoylation reaction––a rela-
tively slow process––alcohol 7 was acetylated using
acetic anhydride and Sc(OTf)3 in the absence and pres-
ence of DTBMP (Table 3).
Acknowledgements
Due to the greater reactivity of Ac2O, a blank experi-
ment was initially performed. Reassuringly, less than 2%
of acetate 18 was formed after 16 h. In the presence of
1 mol % of Sc(OTf)3, complete conversion of 7 to 18
occurred in 5 min. Gratifyingly, addition of 3 mol % of
DTBMP completely suppressed the acetylation and less
than 5% of 18 was observed after 20 h. As predicted, the
use of 3 mol % of TfOH resulted in the quantitative
transformation of 7 to 18 after 5 min.
ꢀ
Financial support of this work by the Universite Cath-
olique de Louvain and the Fond National de la
Recherche Scientifique (FNRS) for R.D. (Aspirant
FNRS) is gratefully acknowledged. I.E.M. is thankful to
Merck & Co. for its continuous support and for the
Merck Young Investigator Award. We are also indebted
to Professor C. Le Roux and to Rhodia for providing us
Table 3. Catalytic acetylation of alcohol 7a
O
O
Ac2O 17 (3 eqs.)
Cat., MeCN
OH
7 (1 eq.)
18
Entry
Cat. (5%)
DTBMP
Conversion (%)a
Time
1
2
3
4
––
––
––
3%
––
<2
>99
<5
16 h
1% Sc(OTf)3
1% Sc(OTf)3
3% TfOH
5 min
20 h
5 min
>99
a Measured by capillary GC after calibration of the response for each component.

R. Dumeunier, I. E. Markꢀo / Tetrahedron Letters 45 (2004) 825–829
829
(b) Brown, H. C.; Kanner, B. J. Am. Chem. Soc. 1966, 88,
986–992; (c) Brown, H. C.; Kanner, B. J. Am. Chem. Soc.
1953, 75, 3865.
with enlightening suggestions and samples of metal tri-
flates. I.E.M. is grateful to Rhodia for the 2001 Rhodia
Outstanding Award.
7. Since up to three equivalents of triflic acid could theoret-
ically be generated per metal triflate, we have decided to
use a 1:3 ratio of metal complex/DTBMP. However, even
one equivalent of the hindered base is sufficient to
significantly reduce the rate of these acylation reactions
(Table 2, entry 5).
8. Competitive elimination has also been observed in the
metal triflate acylation of several allylic and tertiary
alcohols (see Refs. 2a, g, i, j).
References and notes
1. (a) Green, W.; Wuts, P. G. M. In Protective Groups in
Organic Synthesis, 3rd ed.; John Wiley: New York, 1999; p
150; (b) Larock, R. C. Comprehensive Organic Transforma-
tions; VCH Publishers: New York, 1989; p 980; (c) Zhdanov,
R. I.; Zhenodarova, S. M. Synthesis 1975, 222–245.
2. (a) Orita, A.; Tanahashi, C.; Kakuda, A.; Otera, J. J. Org.
Chem. 2001, 66, 8926–8934; (b) Mohammadpoor-Baltork,
I.; Aliyan, H.; Khosropour, A. R. Tetrahedron 2001, 57,
5851–5854; (c) Orita, A.; Tanahashi, C.; Kakuda, A.;
Otera, J. Angew. Chem., Int. Ed. 2000, 39, 2877–2879; (d)
Chauhan, K. K.; Frost, C. G.; Love, I.; Waite, D. Synlett
1999, 1743–1744; (e) Saravanan, P.; Singh, V. K. Tetra-
hedron Lett. 1999, 40, 2611–2614; (f) Zhao, H.; Pendri, A.;
Greenwald, R. B. J. Org. Chem. 1998, 63, 7559–7562; (g)
Procopiou, P. A.; Baugh, S. P. D.; Flack, S. S.; Inglis, G.
G. A. J. Org. Chem. 1998, 63, 2342–2347; (h) Mikami, K.;
Kotera, O.; Motoyama, Y.; Sakaguchi, H.; Maruta, M.
Synlett 1996, 171–172; (i) Ishihara, K.; Kubota, M.;
Kurihara, H.; Yamamoto, H. J. Org. Chem. 1996, 61,
4560–4567; (j) Ishihara, K.; Kubota, M.; Kurihara, H.;
Yamamoto, H. J. Am. Chem. Soc. 1995, 117, 4413–4414.
3. (a) Fujiwara, K.; Murai, A.; Yamashita, M.-Y.; Yasu-
moto, T. J. Am. Chem. Soc. 1998, 120, 10770–10771; (b)
Paquette, L. A.; Barriault, L.; Pissarnitski, D. J. Am.
Chem. Soc. 1999, 121, 4542–4543; (c) White, J. D.;
Blakemore, P. R.; Browder, C. C.; Hong, J.; Lincoln, C.
M.; Nagornyy, P. A.; Robarge, L. A.; Wardrop, D. J. J.
Am. Chem. Soc. 2001, 123, 8593–8595.
9. The 2:1 ratio of (H2O/TfOH) was deduced from the ratio
of acylated adduct versus elimination product.
10. Triflic acid has been used, and sometimes demonstrated to
be the active species, in several catalytic processes: (a)
Kotsuki, H.; Arimura, K.; Ohishi, T.; Maruzasa, R. J. Org.
Chem. 1999, 64, 3770–3773; (b) Kotsuki, H.; Ohishi, T.;
Inoue, M.; Kojima, T. Synthesis 1999, 603–606; (c)
Kotsuki, H.; Arimura, K.; Araki, T.; Shinohara, T.
ꢀ
Synlett 1999, 462–464; (d) Repichet, S.; Le Roux, C.;
Dubac, J.; Desmurs, J.-R. Eur. J. Org. Chem. 1998, 2743–
2746; (e) Laurent-Robert, H.; Le Roux, C.; Dubac, J.
Synlett 1998, 1138–1140; (f) Tsuchimoto, T.; Tobita, K.;
Hiyama, T.; Fukuzawa, S.-I. J. Org. Chem. 1997, 62,
6997–7005, see also Ref. 2g.
11. For reasons of clarity, only one benzoate is shown on the
metal. However, it is quite possible that several species,
formed by the exchange of one, two or three triflate
ligands might be involved in these catalytic processes.
12. For a catalytic version of the Friedel–Crafts acylation
using triflic acid and proceeding via the acyl triflate
intermediate, see: Effenberger, F.; Epple, G. Angew.
Chem., Int. Ed. 1972, 11, 300.
13. Brown, L.; Koreeda, M. J. Org. Chem. 1984, 49, 3875–
3880.
14. As with TfOH, a second aliquot of 12 (7 mol %) had to be
added after 18 h to reach full completion.
4. See, for example: Keck, G. E.; Savin, K. A.; Weglarz, M.
A. J. Org. Chem. 1995, 60, 3194–3204.
5. The deprotection of various silyl ethers using Sc(OTf)3, in
the presence of H2O, has been reported previously:
Oriyama, T.; Kobayashi, Y.; Noda, K. Synlett 1998,
1047–1048; For pKa studies, see: Castellani, C. B.; Carugo,
O.; Giusti, M.; Leopizzi, C.; Perotti, A.; Invernizzi, A. G.;
Vidari, G. Tetrahedron 1996, 52, 11045–11052.
15. It is important to note that the triflate salt of DTBMP
does not catalyse the acylation reaction.
16. The results of the stoichiometric acylation appear to
suggest that one triflate ligand is released initially from the
metal and that catalysts such as MeCOOSc(OTf)2 are
poorly active. Indeed, if only one triflate is exchanged, the
maximum yield of acetate 18 should be 33%.
6. (a) Balaban, T. S.; Unuta, C.; Gheorghiu, M. O.;
Balaban, A. T. Tetrahedron Lett. 1985, 26, 4669–4672;