Triarylcarbenium chlorides as catalysts in allylation reaction: A unique type of reaction with negligible intervention of silyl catalysis

Full text of DOI:10.1021/jo9821832

1090
J . Org. Chem. 1999, 64, 1090-1091
Sch em e 1
Tr ia r ylca r ben iu m Ch lor id es a s Ca ta lysts in
Allyla tion Rea ction : A Un iqu e Typ e of
Rea ction w ith Negligible In ter ven tion of Silyl
Ca ta lysis
Chien-Tien Chen* and Shi-Deh Chao
Department of Chemistry, National Taiwan Normal University,
Taipei, Taiwan, Republic of China
Received October 30, 1998
The addition reaction of allylstannanes to carbonyl-
containing compounds by the action of promoters or catalysts
has been recognized as an indispensable tactic for the
synthesis of homoallylic alcohols.1 Despite recent efforts
along this line, few catalytic allylation systems of this
category have been documented.2 It has been more than 10
years since the seminal description of trityl ions as catalysts
in a myriad array of enol silane derivative-mediated addition
reactions.3 However, their actual catalytic roles in such
reactions have not yet been substantiated in view of the
facile, intervening silyl catalysis vividly described in the
literature.4 Recently, we were engaged in clarifying this
issue by directly observing the catalytic behavior of chiral
triarylcarbenium ions in asymmetric Mukaiyama aldol
additions and have identified the possible cause of competing
silyl catalysis as the direct attack of a nucleophile (e.g., silyl
ketene acetal) to the carbenium center of the catalyst with
concomitant release of silyl-X species.5 An intrinsically viable
solution to this dilemma is by stereoelectronic modification
of the trityl ions to increase their reactivity and compatibility
and, in the meantime, to judiciously choose a counterion for
the minimization of silyl catalysis.6 On the basis of these
two considerations, we sought the possible utility of trityl-
type chlorides in catalytic reactions in view of their well-
documented partially ionic characters7 and the extremely
weak Lewis acidic character of TMS-Cl in most solvent
systems.8 We herein describe our preliminary studies toward
their catalytic uses in allylation reactions, which were not
explored previously, and aim at providing more unambigu-
ous evidence as to the role of trityl catalysis.
electron-releasing (Ar ) 4-t-BuC6H4, 4-MeOC6H4) aryl ap-
pendages, the trityl chlorides are stable enough to allow for
full spectroscopic characterizations.11 Their relative ionic
characters were evaluated by the changes in the chemical
shifts of the carbenium carbon (C(5)) between the alcohols
and the corresponding chorides in 13C NMR. With the same
pendant aryl group (i.e., Ar ) 4-t-BuC6H4), the C(5) in the
dibenzosuberane(DBS)-based chloride (X ) (CH2)2) shows
the largest downfield change (∆δ ) +3.3 ppm). On the
contrary, in the fluorene-based chloride (X) (CH2)0) it shows
the largest upfield shift (∆δ ) -9.1 ppm). The template effect
arranged with a decreasing order of ∆δ(δCl - δOH) is as
follows: DBS > xanthene (X ) O) > diphenylmethyl (X )
H, H) > fluorene. On the other hand, the para-substitution
pattern of the aryl appendage exerts a minor inductive effect
on the chemical shift of C(5). The largest chemical shift
changes observed in the DBS-based system suggest that
these trityl chlorides may display better carbenium chloride
attributes and may thus be more Lewis acidic.
Initial attempts in the allylations of benzaldehyde with
allyltri-n-butyltin (1.2 equiv) catalyzed by the DBS-based
trityl chloride 8b (20 mol %) were disappointing. The
addition product was isolated in 34-37% yields when the
allylations were conducted in CH2Cl2 at ambient tempera-
ture for 48 h (entry 4), Table 1. Longer reaction time only
slightly improved the chemical conversion. On the basis of
a mechanistic consideration, we surmised that the poor
conversion may have to do with a sluggish turn-over between
Bu3SnCl and the intermediate tritylated homoallylic alcohol
(vide infra). As expected, the chemical yields of the allylation
products were significantly improved to 87% when either
TMS-Cl12 or TBS-Cl (1.2 equiv) was added to the reaction
media (entries 5 and 6).13 Moreover, the catalyst 8b was
recovered quantitatively as its corresponding alcohol 4b after
hydrolytic workup.
To find out the best combination of a diarylmethyl tem-
plate and a pendant aryl group to reach a maximal reactivity
of the corresponding trityl-type chloride, we have prepared
various trityl-type alcohols by independent treatment of
xanthone, fluorenone, benzophenone, and dibenzosuberone
with five different aryllithiums of varying electron de-
mands.6,9 The resulting alcohols can be readily converted
to the corresponding chlorides by treatment with SOCl2 (5
equiv) in anhydrous CCl4 at 0 °C or at ambient temperature,
Scheme 1.10 In the cases with the parent (Ar ) C6H5) and
The effects of templates and pendant aryl groups on the
reactivity and compatibility of trityl-type chlorides under
* To whom correspondence should be addressed. Tel: 886-2-2930-9095.
Fax: 886-2-2932-4249. E-mail: chefv043@scc.ntnu.edu.tw.
(1) Yamamoto, Y.; Asao, N. Chem. Rev. 1993, 93, 2207.
(2) For allylations in acid media, see: (a) Yanagisawa, A.; Morodome,
M.; Nakashima, H.; Yamamoo, H. Synlett 1997, 1309. For Sc(OTf)3, see:
(b) Kobayashi, S.; Nagayama, S. J . Am. Chem. Soc. 1997, 119, 10049. (c)
For a review using Ln(OTf)3, see: Kobayashi, S. Synlett 1994, 689. (d) Zr-
(IV): Cozzi, P. G.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C. Synlett 1994,
857. (e) Cationic Rh and Ir complexes: Muss, J . M.; Rennels, R. A. Chem.
Lett. 1993, 197.
(3) Kobayashi, S. Kagaku to Kogyo 1989, 42, 245 and references therein.
(4) (a) Hollis, T. K.; Bosnish, B. J . Am. Chem. Soc. 1995, 117, 4570. (b)
Carreira, E.; Singer, R. A. Tetrahedron Lett. 1994, 35, 4323. (c) Denmark,
S. E.; Chen, C.-T. Tetrahedron Lett. 1994, 35, 4327.
(8) It has been shown that either TrCl or TMSCl was catalytically
inactive unless combined with SnCl2. (a) Mukaiyma, T.; Akamatsu, H.; Han,
J . S. Chem. Lett. 1990, 889. (b) Iwasawa, N.; Mukaiyama, T. Chem. Lett.
1987, 463. (c) Mukaiyama, T.; Kobayashi, S. Chem. Lett. 1987, 491.
(9) Borham, B.; Wilson, J . A.; Gasch, M. J .; Ko, Y.; Kurth, D. M.; Kurth,
M. J . J . Org. Chem. 1995, 60, 7375.
(10) (a) Wilfried, D.; Karl-Heinz, B.; Manfred, P. Ger. Offen. 1811654,
1970. (b) Lew, C. S. Q.; Wong, D. F.; J ohnston, L. J .; Bertone, M.; Hopkinson,
A. C.; Lee-Ruff, E. J . Org. Chem. 1996, 61, 6805.
(11) See the Supporting Information for full spectroscopic characteriza-
tions.
(5) Chen, C.-T.; Chao, S.-D.; Yen, K.-C.; Chen, C.-H.; Chou, I.-C.; Hon,
S.-W. J . Am. Chem. Soc. 1997, 119, 11341.
(6) Chen, C.-T.; Chao, S.-D.; Yen, K.-C. Synlett 1998, 924.
(7) (a) Smith, W. B.; Rao, P. S. J . Org. Chem. 1961, 26, 254. (b) Evans,
A. G.; McEwan, I. H.; Price, A.; Thomas, J . H. J . Chem. Soc. 1955, 3098. (c)
Swain, C. G.; Scott, C. R. J . Am. Chem. Soc. 1953, 75, 246.
(12) For the uses of TMS-X to assist in turnover of metal-catalyzed
reactions, see: (a) Gong, L.; Streitwieser, A. J . Org. Chem. 1990, 55, 6235.
(b) Yu, C.-M.; Choi, H.-S.; Tung, W.-H.; Lee, S.-S. Tetrahedron Lett. 1996,
37, 7095. (c) Whitesell, J . K.; Apodaca, R. Tetrahedron Lett. 1996, 37, 3955.
(13) Acetyl chloride and TMS-Br can also promote the turnover of
Tr-Cl, albeit with poorer efficiency.
10.1021/jo9821832 CCC: $18.00 © 1999 American Chemical Society
Published on Web 01/29/1999

Communications
J . Org. Chem., Vol. 64, No. 4, 1999 1091
Ta ble 1. Effects of Silyl P r om oter s a n d Tem p la tes on
th e Tr ityl Ch lor id e-Ca ta lyzed Allyla tion s of
Ben za ld eh yd e w ith Allytr i-n -bu tyltin
Ta ble 2. Effects of P en d a n t Ar yl Gr ou p s a n d Solven ts
on th e Allyla tion s of Ald eh yd es w ith Allytr i-n -bu tyltin
Ca ta lyzed by 8
entry
R
TrCl (R′)
solvent
yield,a
74/84b
%
1
2
3
4
5
Ph
Ph
Ph
Ph
Ph
8a (H)
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
8b (t-Bu)
8c (OCH3)
8d (CF3)
8e (F5)d
87 (18)c/93b
77/79b
entry
TrCl (X)
promotera
yieldb (9a ), %
yield,b %
<10
46
1
2
3
4
5
6
5a (H, H)
6b (O)
7b ((CH2)0)
8b ((CH2)2)
8b ((CH2)2)
8b ((CH2)2)
TMSCl
TMSCl
TMSCl
65
56
58
37
87
85
quant (10)
quant (11)
quant (12)
0 (13)
0 (13)
0 (13)
6
7
8
9
10
11
Ph
Ph
Ph
Ph
Ph
Ph
8b (t-Bu)
8b (t-Bu)
8b (t-Bu)
8b (t-Bu)
8b (t-Bu)
8b (t-Bu)
CCl4
toluene
Et2O
THF
EtNO2
CH3CN
47 (0)c
20 (0)c
41 (5)c
26 (4)c
33 (22)c
39 (34)c
TMSCl
TBSCl
a
b
In all cases, 1.2 equiv of R3SiX was added. Isolated yields.
this allylation protocol was next examined. The triphenyl-
methyl-, fluorene-, and xanthene-derived chlorides, 5a and
6-7b, were also found to be capable of mediating the
catalytic allylations. The homoallylic alcohol 9a was pro-
duced in moderate (56-65%) yields. However, these cata-
lysts were all completely consumed by direct nucleophilic
allylation of allylstannane to their respective carbenium
centers, Table 1. Presumably, the compatibility of 8b with
the nucleophilic allylstannane arises from the steric encum-
brance of the carbenium center imposed by the pseudoaxial
hydrogens of the two-carbon linkage in the DBS backbone.14
The reactivity of the DBS-based trityl chlorides was
governed by the inductive, electronic stabilization of the
incipient carbenium ions. The chlorides bearing the electron-
withdrawing 4-trifluoromethyl- (R′ ) CF3) and pentafluo-
rophenyl (R′ ) F5) groups led to the least productive
allylations (<10 and 46%, respectively), entries 4 and 5 in
Table 2. The chlorides bearing electron-releasing aryl groups
(i.e., R′ ) OCH3 and t-Bu) behave slightly more efficient than
the parent chloride (R′ ) H) in the allylation (entries 1-3).
It is important to note that the allylations in these three
cases (entries 1-3) can even be completed at 0 °C in 72 h
with equal performances (79-93%).15 Among seven different
solvents examined, the allylation utilizing the best DBS-
based trityl chloride 8b (R′ ) t-Bu) proceeds most smoothly
in CH2Cl2 (87%). The chemical yields of the homoallylic
alcohol 9a fell within a range of 20-47% when reactions
were run in other solvent systems. More importantly, control
experiments in the absence of trityl chloride 8b revealed that
TMS-Cl-mediated allylations were negligible in most sol-
vents under similar reaction conditions except in C2H5NO2
(22%) and CH3CN (34%).16 The optimal allylation procedure
was applicable to both aromatic and aliphatic aldehydes.
With the six representative aldehydes examined, the ho-
moallylic alcohols were all furnished in good (77%, R ) Ph-
(CH2)2) to excellent yields (97%, R ) 2-Naph).
12
13
14
15
16
2-Naph
8b (t-Bu)
8b (t-Bu)
8b (t-Bu)
8b (t-Bu)
8b (t-Bu)
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
97 (9b)
91 (9c)
87 (9d )
85 (9e)e
77 (9f)
4-MeOC6H4
4-NO2C6H4
4-HC(O)C6H4
Ph(CH2)2
a
Isolated yields. b The isolated yieds for the reactions performed
at 0 °C. c The values in parentheses correspond to allyaltions
without 8b. The pending aryl group is C6F5. e Bis(homoallylic
d
alcohol) was obtained.
Sch em e 2
tri-n-butyltin chloride generates the tritylated homoallylic
alcohol 14. Metathesis of the trityl ether with Bu3SnCl would
generate the corresponding stannylated ether with release
of the trityl chloride. Since this exchange process is ham-
pered by the steric bulk of Bu3SnCl, the use of TMS-Cl or
TBS-Cl can facilitate this turnover process, thus completing
the catalytic cycle. We have further confirmed the feasibility
of this exchange process by treatment of an independently
prepared tritylated homoallylic alcohol 14a (Tr ) Ph3C) with
TMS-Cl (1.2 equiv) in CD2Cl2. The exchange occurs readily
1
with a half-life (t1/2) of 30 min as monitored by H NMR.
In conclusion, we have developed a new type of nucleo-
phile-tolerant triarylcarbeium chlorides based on the diben-
zosuberane scaffold. We have provided the first unambiguous
evidence claiming their catalytic roles in the allylation reac-
tions with minimal intervention of silyl catalysis. Efforts
aimed at the development of an asymmetric variant of this
new catalytic system are currently underway.
A mechanistic picture converged from the studies is
delineated in Scheme 2. Initial attack of allyltri-n-butyltin
to the trityl chloride-activated aldehyde with extrusion of
(14) A control experiment by direct equimolar mixing of catalyst-8b with
allyltributyltin in CH2Cl2 at ambient temperature for 3 days led to the
formation of 13 in <5% yield.
(15) For the allylations catalyzed by 5a , 6b, and 7b at 0 °C, the product
yield can be improved from 56% to 83% only in the case of 6b (X ) O)
although these catalysts were all still gradually consumed at this temper-
ature. Apparently, the direct nucleophilic allylation of 6b was effectively
slowed down at 0 °C.
(16) Increased Lewic acidity of TMSCl in CH3CN has been documented;
see: (a) Wang, D.-K.; Dai, L.-X.; Hou, X.-L. Tetrahedron Lett. 1995, 36,
8649. (b) Yasuda, M.; Fujibayashi, T.; Shibata, I.; Baba, A.; Matsuda, H.;
Sonoda, N. Chem. Lett. 1995, 167.
Ack n ow led gm en t. We are grateful to the National
Science Council of the Republic of China for generous
support of this research.
Su p p or tin g In for m a tion Ava ila ble: Preparation and full
spectroscopic characterization of 1a , 2b, 3b, 4a -e, 5a , 6b, 7b, 8a -
c, 9a -f, 10-13, and 14a as well as the optimal procedures for
the allylations catalyzed by 5-8 are provided.
J O9821832