On the mechanism of asymmetric allylation of aldehydes with allyltrichlorosilanes catalyzed by QUINOX, a chiral isoquinoline N-oxide

Article abstract of DOI:10.1021/ja711338q

Allylation of aromatic aldehydes 1a-m with allyl- and crotyl- trichlorosilanes 2-4, catalyzed by the chiral N-oxide QUINOX (9), has been found to exhibit a significant dependence on the electronics of the aldehyde, with p-(trifluoromethyl)benzaldehyde 1g

Full text of DOI:10.1021/ja711338q

Published on Web 03/14/2008
On the Mechanism of Asymmetric Allylation of Aldehydes
with Allyltrichlorosilanes Catalyzed by QUINOX, a Chiral
Isoquinoline N-Oxide
Andrei V. Malkov,*,† Pedro Ramírez-López,†,‡ Lada Biedermannová
(née Bendová),†,§ Lubomír Rulíšek,*,| Lenka Dufková,†, Martin Kotora,#
Fujiang Zhu,† and Pavel Kocˇovský*,†
Department of Chemistry, WestChem, Joseph Black Building, UniVersity of Glasgow, Glasgow
G12 8QQ, Scotland, U.K., Institute of Organic Chemistry and Biochemistry and Gilead Sciences
Research Center & IOCB, Academy of Sciences of the Czech Republic, FlemingoVo námeˇstí 2,
16610 Prague 6, Czech Republic, and Department of Organic Chemistry, Charles UniVersity,
HlaVoVa 2030, 12840, Prague 2, Czech Republic
Received January 2, 2008; E-mail: amalkov@chem.gla.ac.uk; lubos@uochb.cas.cz; pavelk@chem.gla.ac.uk
Abstract: Allylation of aromatic aldehydes 1a-m with allyl- and crotyl-trichlorosilanes 2–4, catalyzed by
the chiral N-oxide QUINOX (9), has been found to exhibit a significant dependence on the electronics of
the aldehyde, with p-(trifluoromethyl)benzaldehyde 1g and its p-methoxy counterpart 1h affording the
corresponding homoallylic alcohols 6g,h in 96 and 16% ee, respectively, at -40 °C. The kinetic and
computational data indicate that the reaction is likely to proceed via an associative pathway involving neutral,
octahedral silicon complex 22 with only one molecule of the catalyst involved in the rate- and selectivity-
determining step. The crotylation with (E) and (Z)-crotyltrichlorosilanes 3 and 4 is highly diastereoselective,
suggesting the chairlike transition state 5, which is supported by computational data. High-level quantum
chemical calculations further suggest that attractive aromatic interactions between the catalyst 9 and the
aldehyde 1 contribute to the enantiodifferentiation and that the dramatic drop in enantioselectivity, observed
with the electron-rich aldehyde 1h, originates from narrowing the energy gap between the (R)- and (S)-
reaction channels in the associative mechanism (22). Overall, a good agreement between the theoretically
predicted enantioselectivities for 1a and 1h and the experimental data allowed to understand the specific
aspects of the reaction mechanism.
coordinates to the silicon (Scheme 1).1,4–7 Provided the Lewis
base dissociates from the silicon at a sufficient rate, it can act
Introduction
Allyltrialkylsilanes1 and stannanes1,2 react readily at low
temperature with aldehydes upon activation of the carbonyl
group by a Lewis acid.1–3 By contrast, allyltrichlorosilane (2)
and related reagents require activation with a Lewis base that
as a catalyst (rather than a stoichiometric reagent). Typical Lewis
bases that promote the allylation reaction are the common
dipolar aprotic solvents, such as DMF,4,8 DMSO,4,5 and
HMPA,5,9 and other substances possessing a strongly Lewis
basic oxygen, such as formamides,4,10,11 urea derivatives,12
catecholates,6a and their chiral modifications;1 the allylation with
allyltrifluorosilane can also be promoted by fluoride.6a,c,e
Analogous allyl boranes and boronates are more reactive than
† University of Glasgow.
‡ Current address: Departamento de Productos Naturales, IQOG (CSIC),
Madrid, Spain.
§ Exchange student from the Institute of Organic Chemistry and
ˇ
Biochemistry, AVCR.
|
ˇ
Institute of Organic Chemistry and Biochemistry, AVCR.
Exchange student from Charles University.
(5) Denmark, S. E.; Coe, D. M.; Pratt, N. E.; and Griedel, B. D. J. Org.
Chem. 1994, 59, 6161.
# Charles University.
(6) (a) Kira, M.; Sato, K.; Sakurai, H. J. Am. Chem. Soc. 1988, 110, 4599.
(b) Kira, M.; Kobayashi, M.; Sakurai, H. Tetrahedron Lett. 1987, 28,
4081. (c) Sakurai, H. Synlett 1989, 1. (d) Kira, M.; Sato, K.; Sakurai,
H. J. Am. Chem. Soc. 1990, 112, 257. (e) Kira, M.; Zhang, L. C.;
Kabuto, C.; Sakurai, H. Organometallics 1998, 17, 887.
(7) (a) Malkov, A. V.; Kocˇovský, P. Eur. J. Org. Chem. 2007, 29. (b)
Kocˇovský, P.; Malkov, A. V. In EnantioselectiVe Organocatalysis;
Dalko, P. I., Ed.; Wiley-VCH: Weinheim, Germany, 2007; p 255. (c)
See also: Oestreich, M.; Rendler, S. Synthesis 2005, 1727. (d) Orito,
Y.; Nakajima, M. Synthesis 2006, 1391. (e) Benaglia, M.; Guizzetti,
S.; Pignataro, L. Coord. Chem. ReV. 2007, doi: 10.1016/j.ccr.
2007.10.009.
(1) (a) For leading reviews, see: Denmark, S. E.; Stavenger, R. A. Acc.
Chem. Res. 2000, 33, 432. (b) Denmark, S. E.; Fu, J. Chem. Commun.
2003, 167. (c) Denmark, S. E.; Fu, J. Chem. ReV. 2003, 103, 2763.
(d) Kennedy, J. W. J.; Hall, D. G. Angew. Chem., Int. Ed. 2003, 42,
4732.
(2) (a) Marshall, J. A. Chem. ReV. 1996, 96, 31. (b) Marshall, J. A. In
Organometallics in Synthesis; Schlosser, M., Ed.; J. Wiley and Sons:
Chichester, U.K., 2002; p 399.
(3) (a) For the coordination of the C)O group to Lewis acids (η1 vs η2
fashion), see, for example: Shambayati, S.; Crowe, W. E.; Schreiber,
S. L. Angew. Chem., Int. Ed. Engl. 1990, 29, 256. (b) Lenges, C. P.;
Brookhart, M.; White, P. S. Angew. Chem., Int. Ed. 1999, 38, 552.
(4) (a) Kobayashi, S.; Nishio, K. Tetrahedron Lett. 1993, 34, 3453. (b)
Kobayashi, S.; Nishio, K. J. Org. Chem. 1994, 59, 6620. (c) Kobayashi,
S.; Nishio, K. Synthesis 1994, 457.
(8) Short, J. D.; Attenoux, S.; Berrisford, D. J. Tetrahedron Lett. 1997,
38, 2351.
(9) Wang, Z.; Xu, G.; Wang, D.; Pierce, M. E.; Confalone, P. N.
Tetrahedron Lett. 2000, 41, 4523.
9
10.1021/ja711338q CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 5341–5348 5341

A R T I C L E S
Malkov et al.
Scheme 1. Allylation of Aldehydes 1 with Allyl and Crotyl
Trichlorosilanes 2–4a
Chart 1. Selected Lewis Basic Catalysts for Allylation of Aldehydes
with 2a
investigation into the mechanism of allylation catalyzed by chiral
phosphoramides, Denmark has found that the reaction follows
a first order kinetics in both silane and aldehyde and a second
order in the monodentate phosphoramide promoter (10).16a,22,23
Hence, not surprisingly, bidentate Lewis bases, such as bispho-
sphoramides (11)16a,22 and bipyridine N,N′-dioxides (12),18,19
proved to be superior in reactivity and selectivity compared to
monodentate catalysts.7
In parallel, we have introduced a family of chiral pyridine
N-monooxides, such as 9 (QUINOX) and 13 (METHOX), acting
as remarkably efficient monodentate catalysts in the allylation
reaction (Scheme 1) with up to 98% ee.24,25 In a preliminary
communication,25 we have demonstrated that, despite the
structural similarity across the whole series, QUINOX (9)
displayed an unusual reactivity pattern, for which a mechanistic
rationale is presented herein.
a For a-m, see Table 1.
their Si and Sn counterparts and do not require an activator.13
Similarly, the chelates, generated from allyltrichlorosilane and
a stoichiometric amount of an amino alcohol, such as pseu-
doephedrine, do not require further activation and have been
shown to produce the corresponding homoallylic alcohols 6 with
high enantioselectivity.14
Asymmetric allylation of aldehydes 1 with allyl- and crotyl-
trichlorosilanes 2-4, catalyzed by chiral Lewis bases (Scheme
1), in particular phosphoramides 10 and 11,15–17 bipyridine N,N-
bisoxides (e.g., 12),18–20 and terpyridine N,N,N-trisoxides21
(Chart 1) has evolved into an efficient method for the synthesis
of enantiomerically enriched homoallylic alcohols 6-8 (Scheme
1).1,7 In general, the reaction displays an excellent diastereo-
control in the case of trans- and cis-crotylsilanes (3/4), sug-
gesting the cyclic transition state (TS) 5. In a detailed
Results and Discussion
In the past few years, we have developed a series of chiral
pyridine N-oxide organocatalysts for the enantioselective ally-
lation of aromatic aldehydes 1 with allyltrichlorosilanes 2–4
(10) (a) Iseki, K.; Mizuno, S.; Kuroki, Y.; Kobayashi, Y. Tetrahedron Lett.
1998, 39, 2767. (b) Iseki, K.; Mizuno, S.; Kuroki, Y.; Kobayashi, Y.
Tetrahedron 1999, 55, 977. (c) For analogous activation of Cl3SiH
by chiral tertiary amides derived from amino acids, see: Iwasaki, F.;
Omonura, O.; Mishima, K.; Kanematsu, T.; Maki, T.; Matsumura, Y.
Tetrahedron Lett. 2001, 42, 2525. (d) Malkov, A. V.; Mariani, A.;
MacDougall, K. N.; Kocˇovský, P. Org. Lett. 2004, 6, 2253.
(11) (a) Kobayashi, S.; Nishio, K. Chem. Lett. 1994, 1773. (b) Kobayashi,
S.; Nishio, K. J. Org. Chem. 1994, 59, 6620. (c) Kobayashi, S.; Nishio,
K. J. Org. Chem. 1994, 59, 6620.
(20) (a) For reviews, see : Chelucci, G.; Thummel, R. P. Chem. ReV. 2002,
102, 3129. (b) Malkov, A. V.; Kocˇovský, P. Curr. Org. Chem. 2003,
7, 1737. (c) Chelucci, G.; Murineddu, G.; Pinna, G. A. Tetrahedron:
Asymmetry 2004, 15, 1373. (d) Kwong, H. L.; Yeung, H. L.; Yeung,
C. T.; Lee, W. S.; Lee, C. S.; Wong, W. L. Coord. Chem. ReV. 2007,
251, 2188. (e) For the most recent example of bipyridine N,N-bisoxide
catalyst, see : Chelucci, G.; Belmonte, N.; Benaglia, M.; Pignataro,
L. Tetrahedron Lett. 2007, 48, 4037.
(21) Wong, W.-L.; Lee, C.-S.; Leung, H.-K.; Kwong, H.-L. Org. Biomol.
Chem 2004, 2, 1967.
(12) Chataigner, I.; Piarulli, U.; Gennari, C. Tetrahedron Lett. 1999, 40,
3633.
(22) (a) Denmark, S. E.; Fu, J.; Coe, D. M.; Su, X.; Pratt, N. E.; Griedel,
B. D. J. Org. Chem. 2006, 71, 1513. (b) For mechanistic studies on
the related aldol reaction, see: Denmark, S. E.; Bui, T. Proc. Natl.
Acad. Sci. U.S.A. 2004, 101, 5439. (c) Denmark, S. E.; Bui, T. J.
Org. Chem. 2005, 70, 10393. (d) Denmark, S. E.; Pham, S. M.;
Stavenger, R. A.; Su, X.; Wong, K.-T.; Nishigaichi, Y. J. Org. Chem.
2006, 71, 3904. (e) For mechanistic studies on Lewis-base catalyzed
opening of epoxides with chlorosilanes, see: Denmark, S. E.; Barsanti,
P. A.; Beutner, G. L.; Wilson, T. W. AdV. Synth. Catal. 2007, 349,
567.
(13) (a) Smith, K. In Organometallics in Synthesis; Schlosser, M., Ed.; J.
Wiley and Sons; Chichester, U.K., 2002; p 514. (b) Chemler, S. R.;
Roush, W. R. In Modern Carbonyl Chemistry; Otera, J., Ed.; Wiley-
VCH: Weinheim, 2000; Chapter 11.
(14) (a) Wang, X.; Meng, Q.; Nation, A. J.; Leighton, J. L. J. Am. Chem.
Soc. 2002, 124, 10672. (b) Kinnaird, J. W. A.; Ng, P. Y.; Kubota, K.;
Wang, X.; Leighton, J. L. J. Am. Chem. Soc. 2002, 124, 7920. (c)
Kubota, K.; Leighton, J. L. Angew. Chem., Int. Ed. 2003, 42, 946.
(15) (a) Iseki, K.; Kuroki, Y.; Takahashi, M.; Kishimoto, S.; Kobayashi,
Y. Tetrahedron 1997, 53, 3513. (b) Wadamoto, M.; Ozasa, N.;
Yanagisawa, A.; Yamamoto, H. J. Org. Chem. 2003, 63, 5593. (c)
Iseki, K.; Kuroki, Y.; Takahashi, M.; Kobayashi, Y. Tetrahedron Lett.
1996, 37, 5149.
(23) In the allylation, unlike in the related aldol reaction (ref 22d), the
minor, one catalyst manifold could not be characterized (ref 22a).
(24) (a) Malkov, A. V.; Orsini, M.; Pernazza, D.; Muir, K. W.; Langer,
V.; Meghani, P.; Kocˇovský, P. Org. Lett. 2002, 4, 1047. (b) Malkov,
A. V.; Bell, M.; Orsini, M.; Pernazza, D.; Massa, A.; Herrmann, P.;
Meghani, P.; Kocˇovský, P. J. Org. Chem. 2003, 68, 9659. (c) Malkov,
A. V.; Bell, M.; Vassieu, M.; Bugatti, V.; Kocˇovský, P. J. Mol. Catal.
A 2003, 196, 179. (d) Malkov, A. V.; Bell, M.; Castelluzzo, F.;
Kocˇovský, P. Org. Lett. 2005, 7, 3219. (e) For other N-monooxide
catalysts, see: Traverse, J. F.; Zhao, Y.; Hoveyda, A. H.; Snapper,
M. L. Org. Lett. 2005, 7, 3151. (f) Chai, Q.; Song, C.; Sun, Z.; Ma,
Y.; Ma, C.; Dai, Y.; Andrus, M. B. Tetrahedron Lett. 2006, 47, 8611.
(g) Pignataro, L.; Benaglia, M.; Annunziata, R.; Cinquini, M.; Cozzi,
F. J. Org. Chem. 2006, 71, 1458.
(16) (a) Denmark, S. E.; Fu, J. J. Am. Chem. Soc. 2000, 122, 12021. (b)
Denmark, S. E.; Fu, J. J. Am. Chem. Soc. 2001, 123, 9488.
(17) Hellwig, J.; Belser, T.; Müller, J. F. K. Tetrahedron Lett. 2001, 42,
5417.
(18) (a) Nakajima, M.; Saito, M.; Shiro, M.; Hashimoto, S. J. Am. Chem.
Soc. 1998, 120, 6419. (b) Nakajima, M.; Saito, M.; Hashimoto, S.
Chem. Pharm. Bull. 2000, 48, 306.
(19) (a) Shimada, T.; Kina, A.; Ikeda, S.; Hayashi, T. Org. Lett. 2002, 4,
2799. (b) Shimada, T.; Kina, A.; Hayashi, T. J. Org. Chem. 2003, 68,
6329. (c) Kina, A.; Shimada, T.; Hayashi, T. AdV. Synth. Catal. 2004,
346, 1169.
(25) Malkov, A. V.; Dufková, L.; Farrugia, L.; Kocˇovský, P. Angew. Chem.,
Int. Ed. 2003, 42, 3674.
9
5342 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Mechanism of Asymmetric Allylation of Aldehydes
A R T I C L E S
binaphthyl (BINOL),29 which gave the crystalline material
containing (S)-(-)-BINOL and (+)-9 in a 1:1 ratio, while (-)-9
remained in the solution. This cocrystallization, followed by a
chromatographic separation of (+)-9 from (S)-BINOL, furnished
pure (+)-9 of 98% ee (as revealed by chiral HPLC) in 89%
isolated yield.26 The absolute configuration of 9 was found to
be (R)-(+)-9 by crystallographic analysis of the molecular crystal
of (S)-(-)-BINOL·(+)-9 (Figure 1A).25
Scheme 2. Synthesis of QUINOX and its Deoxy Analogue
To asses the electronic role of the methoxy group in QUINOX
(9), we have now synthesized its methyl analogue 19 (Scheme
2). The synthesis was carried out in the same way as that of
QUINOX, namely, via the Suzuki-Miyaura coupling of 1-chloro-
isoquinoline (14)27b with boronic acid 16, which in turn was
prepared from 1-bromo-2-methyl naphthalene by lithiation with
n-BuLi in THF, followed by a reaction with (MeO)3B and
subsequent hydrolysis of the intermediate boronate ester with
HCl.25,27b,30 The coupling itself was carried out by refluxing a
mixture of 14 and 16 in DME with aqueous K2CO3 and
(Ph3P)4Pd as catalyst (3 mol %) for 48 h to afford 18 (71%).
Oxidation of the latter derivative with MCPBA gave rise to the
racemic N-oxide 19 (81%), which was resolved by cocrystal-
lization with (R)-(+)-BINOL from CH2Cl2.29 The crystals
contained the molecular compound (R)-BINOL·(R)-19, as
revealed by single crystal X-ray analysis (Figure 1B). Flash
chromatography of the latter material gave (R)-(-)-19 (33%;
g99.5% ee) and the recovered (R)-(+)-BINOL. The mother
liquor provided the enantiomerically enriched (S)-19, which was
purified by cocrystallization with (S)-(-)-BINOL to produce
(S)-(+)-19 (19%; g99.8% ee).
(Scheme 1). Thus, for instance, the terpene-derived METHOX
(13) and its congeners have been shown by us to work generally
very well and to tolerate a wide range of electron-poor and
electron-rich benzaldehydes, exhibiting little dependence of the
reaction rate and enantioselectivity on the nature of the Ar group
in 1 (96 and 93% ee for 4-MeOC6H4CHO and 4-CF3C6H4CHO,
respectively, with 13 as catalyst).24d However, in contrast to
most of the catalysts reported to date, QUINOX (9), another
member of the N-oxide family, exhibited a significant preference
toward electron-poor aldehydes (96% ee with 4-CF3C6H4CHO),
whereas electron-rich substrates reacted much slower and were
characterized by dramatically lower selectivities (16% ee for
4-MeOC6H4CHO!).25 This intriguing behavior suggests that two
fundamentally different mechanisms operate in these reactions
as a function of the catalyst structure, which prompted us to
embark on a detailed mechanistic study.
Allylation of Aldehydes 1a-m Catalyzed by QUINOX (9).
Addition of allyltrichlorosilane (2) to benzaldehyde (1a) (Scheme
1), carried out in the presence of (R)-(+)-9 (5 mol %) at -40
°C for 12 h, produced (R)-(+)-6a of 87% ee (Table 1, entry 1).
Lowering the catalyst load to 1 mol % slowed the reaction but
had no effect on the enantioselectivity (entry 2).
The solvent effect was investigated briefly. In MeCN, the
reaction was slower and rather less enantioselective than that
in CH2Cl2 (compare entries 1 and 5). Further decrease of
enantioselectivity was observed for CHCl3 (63% ee, entry 6),
but the reaction proved to be much faster (30 min at -40 °C).
On the other hand, in toluene, the reaction became very sluggish
owing to poor solubility of the catalyst (entry 7), although
moderate asymmetric induction (61% ee) was still observed.
Introduction of electron-withdrawing substituents into the
aromatic ring of the aldehyde resulted in a notable increase in
both reactivity and enantioselectivity. Thus, p-nitrobenzaldehyde
1b afforded the corresponding product in better yield and with
slightly higher ee than benzaldehyde (compare entries 1 and
8); the halo derivatives 1c-f gave >90% ee (entries 9–12). The
highest conversion and enantioselectivity (96% ee) was attained
with the p-trifluoromethyl derivative 1g (entry 13).
By contrast, the electron-rich p-methoxybenzaldehyde (1h)
afforded an almost racemic product (entry 16) and its o-isomer
1i also showed low selectivity (entry 19); deceleration was
observed in both cases. On the other hand, m-methoxybenzal-
dehyde 1j (entry 20), where the methoxy group serves as a weak
acceptor, again showed a good level of selectivity (80% ee),
although the reaction was still quite slow. 3,5-Dimethylbenzal-
Synthesis of QUINOX (9) and its Methyl Analogue (19). The
synthesis of QUINOX 9, briefly described in our preliminary
communication (Scheme 2),25,26 was inspired by the initial steps
toward QUINAP.27 Thus, the Suzuki-Miyaura coupling of
1-chloro-isoquinoline (14)27b with boronic acid 15,27b,28 carried
out in refluxing DME in the presence of Cs2CO3 and (Ph3P)4Pd
(3 mol%) overnight, afforded the biaryl derivative 1727 (95%),
whose treatment with m-chloroperoxybenzoic acid provided
the racemic N-oxide (()-9 (99%). The latter racemate was
resolved by cocrystallization with (S)-(-)-2,2′-dihydroxy-1,1′-
(26) (+)-QUINOX has been first prepared by Nakajima, including the
same Suzuki-Miyaura coupling and resolution. However, the
synthesis was mentioned as a footnote without revealing the con-
ditions, and the absolute configuration was not determined. Nakajima,
M.; Saito, M.; Uemura, M.; Hashimoto, S. Tetrahedron Lett. 2002,
43, 8827.
(27) (a) Brown, J. M.; Hulmes, D. I.; Layzell, T. P. J. Chem. Soc., Chem.
Commun. 1993, 1673. (b) Alcock, N. W.; Brown, J. M.; Hulmes, D. I.
Tetrahedron: Asymmetry 1993, 4, 743.
(29) For the previous use of BINOL as a resolving agent, see refs 18, 19a,
26, and Nakajima, M.; Saito, M.; Hashimoto, S. Tetrahedron:
Asymmetry 2002, 13, 2449.
(28) Vyskocˇil, Š.; Meca, L.; Tišlerová, I.; Císarˇová, I.; Polášek, M.;
Harutyunyan, S. R.; Belokon, Y. N.; Stead, R. M. J.; Farrugia, L.;
Lockhart, S. C.; Mitchell, W. L.; Kocˇovský, P. Chem. Eur. J. 2002,
8, 4633.
(30) Pathak, R.; Nhlapo, J. M.; Govender, S.; Michael, J. P.; van Otterlo,
W. A. L.; de Koning, C. B. Tetrahedron 2006, 62, 2820.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5343

A R T I C L E S
Malkov et al.
Figure 1. (A) Disposition of the molecules in the crystal containing (R)-(+)-9 (right) and (S)-(-)-BINOL (left) as shown by X-ray crystallography. (B)
Molecules in the crystal of (R)-(-)-19 (right) with (R)-(+)-BINOL (left). Both diagrams illustrate the N-O· · ·H-O hydrogen bonding.
Table 1. The Allylation of Aldehydes 1a-m with Allylsilane 2, Catalyzed by 9 and 19 (Scheme 1)a
entry
catalyst
aldehyde
Ar
solvent
time (h)
t (°C)
yield (%)b
ee (%)c,d
1
2
3
4
5
6
7
8
(R)-(+)-9
(R)-(+)-9e
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(S)-(-)-9
(S)-(-)-9
(S)-(-)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(+)-9
(R)-(-)-19g
(R)-(-)-19g
(R)-(-)-19g
1a
1a
1a
1a
1a
1a
1a
1b
1c
1d
1e
1f
1g
1g
1g
1h
1h
1h
1i
Ph
Ph
Ph
Ph
Ph
Ph
Ph
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
MeCN
CHCl3
12
12
12
2
12
0.5
12
2
2
2
2
2
-40
-40
-20
18
68
55
76
40
60
79
<5
73
65
72
49
79
85
88
65
70
45
10
73
40
68
86
71
19
26
7
87 (R)
87 (R)
84 (R)
65 (R)
72 (R)
63 (R)
61 (R)
89 (R)
93 (R)
91 (R)
95 (R)
91 (R)
96 (R)
91 (R)
89 (R)
16 (S)f
72 (S)f
45 (S)f
37 (R)
80 (R)
81 (R)
51 (R)
55 (R)
64 (S)
71 (S)
7 (S)
-40
-40
-40
-40
-40
-40
-40
-40
-40
-20
0
toluene
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
4-NO2-C6H4
4-Cl-C6H4
2-Cl-C6H4
3-Cl-C6H4
4-F-C6H4
4-CF3-C6H4
4-CF3-C6H4
4-CF3-C6H4
4-MeO-C6H4
4-MeO-C6H4
4-MeO-C6H4
2-MeO-C6H4
3-MeO-C6H4
3,5-Me-C6H3
PhCHdCH
PhCHdC(Me)
Ph
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
2
2
2
18
18
18
12
12
16
12
12
12
12
12
-40
-20
0
-40
-40
-40
-40
-40
-40
-40
-40
1j
1k
1l
1m
1a
1g
1h
4-CF3-C6H4
4-MeO-C6H4
a The reaction was carried out at 0.4 mmol scale with 1.1 equiv of 2, in the presence of 9 (5 mol %, g98% ee) or 19 (5 mol %, g99% ee) as
catalyst, and (i-Pr)2NEt (1 equiv) as base, unless stated otherwise. b Isolated yield. c Determined by chiral HPLC or GC. d The configuration of the
products 6 was established by the comparison of their optical rotations (measured in CHCl3) and their GC and HPLC retention times with the literature
data and with the behavior of authentic samples (refs 24 and 25). e With 1 mol % of the catalyst. f Note that this experiment was carried out with
(S)-(-)-9. g Note that because of the change of substituent preferences in the CIP nomenclature, (R)-19 actually corresponds to (S)-9.
dehyde 1k exhibited good reactivity and enantioslelectivity
(entry 21), while cinnamyl derivatives 1l and 1m produced
modest results (entries 22 and 23).
deceleration effect in the case of 1h (entries 16–18), reflected
in a decreased yield (and an increased amount of the unreacted
aldehyde recovered).
Investigation of the effects of temperature on the enantiose-
lectivity of allylation catalyzed by QUINOX (9) revealed that
benzaldehyde 1a and its electron poor p-trifluoromethyl deriva-
tive 1g behaved in an ordinary fashion, showing a slight erosion
of the ee at increased temperatures (entries 1, 3, and 4 for 1a
and 13–15 for 1g). At the same time, the electron-rich
p-methoxybenzaldehyde 1h exhibited a significant temperature
dependence (entries 16–18), known as isoinversion effect,31
where the highest value of 72% ee was achieved at -20 °C.
Furthermore, the increase in temperature had a significant
Allylation Catalyzed by Methyl-QUINOX (19). The reactiv-
ity of the N-oxide 19, a methyl analogue of QUINOX, was
explored with the aid of three representative aldehydes 1a,g,h
(Table 1, entries 24–26). Compared to QUINOX, the reactions
turned out to be much slower and the enantioselectivities lower.
Nevertheless, the same trend was observed, that is, the electron-
poor 1g was the most reactive of the series, while the worst
results were obtained with the electron-rich aldehyde 1h. This
behavior clearly shows that the methoxy group in QUINOX
(9) not only serves as steric element preventing the rotation
about the chiral axis but also must play an additional, presum-
ably electronic, role.
(31) (a) For a discussion of isoinversion principle, see: Schmid, R.; Sapunov,
V. N. Non-formal Kinetics; VCH: Weinheim, Germany, 1982. (b)
Gypser, A.; Norrby, P.-O. J. Chem. Soc. Perkin Trans. 2 1997, 939.
9
5344 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Mechanism of Asymmetric Allylation of Aldehydes
A R T I C L E S
Table 2. The Allylation of Aldehydes 1a,g,h with Allylsilanes 3 (E/Z g 98:2) and 4 (E/Z e 2:98) Catalyzed by (S)-(-)-9 (Scheme 1)a
% ee (configuration)c,d
anti, syn
entry
aldehyde
Ar
silane
yield %b
7:8 (anti:syn)
1
2
3
4
5
6
1g
1a
1h
1g
1a
1h
4-CF3-C6H4
Ph
4-MeO-C6H4
4-CF3-C6H4
Ph
3
3
3
4
4
4
75
65
40
85
78
50
96:4
95:5
83:17
1:99
1:99
4:96
92 (1S,2S), 90 (1S,2R)
66 (1S,2S), 77 (1S,2R)
25 (1S,2S), 12 (1R,2S)
nd, 94 (1S,2R)
69 (1S,2S), 79 (1S,2R)
40 (1R,2R), 60 (1S,2R)
4-MeO-C6H4
a The reaction was carried out at 0.4 mmol scale with 1.1 equiv of 3/4, in the presence of (S)-(-)-9 (5 mol %, 98% ee) as catalyst and (i-Pr)2NEt (1
equiv) as base in CH2Cl2 at –40 °C for 24 h. b Isolated yield. c Determined by chiral GC. d The absolute configuration of 7/8 was determined by
comparison of their optical rotations (measured in CHCl3) and their GC retention times with the literature data (refs 16b, 33) and with the behavior of
authentic samples.
Crotylation Catalyzed by QUINOX (9). The crotylation
reaction was found by us to be highly diastereoselective,25 which
is consistent with the cyclic chairlike transition state 5. Previ-
ously, we reported that addition of crotyltrichlorosilane 3 (trans/
cis ratio 87:13) to benzaldehyde (1a), catalyzed by QUINOX
(9), produced a ∼70:30 mixture of 7a and 8a, suggesting a
partial loss of diastereoselectivity.25 We have now repeated these
experiments with the isomerically pure 3 (trans/cis g 98:2)32
and 4 (trans/cis e 2:98)32 and found that the reaction, in fact,
was highly diastereoselective for both 1a and 1g (Table 2, entries
1, 2, 4, and 5);33 only with the electron-rich 1h (entries 3 and
6) it was less stereoconvergent. Significantly, cis-crotylsilane 4
exhibited a higher rate of allylation compared to the trans isomer
3, which represents a reversal of the trend commonly observed
with other N-oxide catalysts,24 in particular METHOX (13).24d
On the basis of these results, the seeming loss of diastereocon-
Figure 2. Reaction order in 2.
trol, reported by us previously for 3 with the 87:13 trans/cis
ratio,25 can now be attributed to a higher reactivity of the minor
(Vo) versus the [silane]/[catalyst] ratio (Figure 2) consists of two
cis isomer 4, thus showing a kinetic preference of QUINOX
distinct areas. At the silane/catalyst ratio up to 15:1 (lg value
(9) toward cis-crotylsilane (in the excess of the geometrically
1.18), the order in silane is unity (1.02 for both 1g and 1h),
impure reagent). Erosion of diastereoselectivity in the case of
1h was accompanied by low enantioselectivities (entries 3 and
6). Furthermore, the absolute configuration at C(1) of the minor
product proved to be opposite to that of the major diastereoi-
whereas above that point the reaction slows down and the order
in silane turns negative.35 At low silane-to-catalyst ratios, the
kinetic data obtained for aldehyde 1g and silane 2 mirror those
reported by Denmark for phosphoramides.16a,22 However, the
somer (entries 3 and 6), demonstrating an inversion of the facial
order in QUINOX (9), determined at 10–25 mol % loadings
(with 1 equiv of 1g and 1.5 equiv of 2), was found close to
unity (0.82), which contrasts with the second order observed
for monodentate phosphoramides. Taking into account the
reversible nature of the coordination of QUINOX (9) to silane
2 and the absence of a nonlinear effect (in the case of 9),
combined with the observations that enantioselectivity is not
affected by catalyst loading in the range of 1–10 mol %,25 it
appears very likely that only one molecule of the catalyst is
involved in the rate and selectivity determining step.
attack at the CdO bond from si to re, which may originate
from participation of a twisted boatlike transition structure.34
A probe into the possible nonlinear effect in the allylation of
1a with 2, catalyzed by (R)-(+)-9 of 29%, 50%, and 75% ee,
respectively, has demonstrated a fully linear relationship between
the enantiopurity of the catalyst and the product. In these
experiments, the resulting alcohol 6a was of 27%, 45%, and
71% ee, respectively.
Kinetic Experiments. To shed light on the mechanism, a series
of kinetic experiments were carried out. The order in each
component was determined by using the method of initial rates.
A sampling technique was employed, in which the aliquots taken
at certain intervals were analyzed by GC to monitor the product
formation. Although the high sensitivity of the reaction to traces
of moisture complicated the data collection, acceptable repro-
ducibility was achieved up to the 20% conversion. The reaction
followed a first-order kinetics in aldehyde (1.12 for 1g), whereas
the order in silane exhibited an unusual behavior, irrespective
of the nature of the aldehyde. The lg/lg plot of the initial rates
The rate limiting step (RLS) was elucidated by using
Hammett correlation. The plot in Figure 3 shows that electron-
withdrawing substituents accelerated the reaction (F ) 1.18),
indicating that it is the electrophilicity of the carbonyl carbon
(rather than the Lewis basicity of the carbonyl oxygen) that
influences the reactivity. This observation is consistent with the
rate determining C-C bond formation rather than pre-equilibra-
tion.36 This F value also suggests that the reaction center in the
(32) Iseki, K.; Kuroki, Y.; Takahashi, M.; Kishimoto, S.; Kobayashi, Y.
Tetrahedron 1997, 53, 3513.
(35) Analysis of the reaction mixtures by 1H NMR revealed that both
catalysts 9 and 19 remained intact for the duration of the process; no
reduction of the N-oxide group or catalyst decomposition was observed
under the reaction conditions.
(33) (a) For comparison, see ref 16b and the following: Hackman, B. M.;
Lombardi, P. J.; Leighton, J. L. Org. Lett. 2004, 6, 4375. (b) McManus,
H. A.; Cozzi, P. G.; Guiry, P. J. AdV. Synth. Catal. 2006, 348, 551.
(34) For a similar effect in the related aldol reaction, ref 22d.
(36) In the allylation, the C-C bond formation was assumed to be a rate-
limiting step (ref 22a) but not confirmed by kinetic experiments.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5345

A R T I C L E S
Malkov et al.
participation of the tightly packed transitions state 22 can be
expected to lead to higher selectivities. The low F value, together
with the observed higher reaction rates and better selectivities
attained in nonpolar solvents, is consistent with the reaction
pathway via the neutral, octahedral transition state 22. The
observed maximum on the plot of enantioselectivity versus
reaction temperature in the case of p-methoxybenzaldehyde 1h
(the inversion effect) is indicative of either a temperature-
dependent shift in the RLS (from C-C- bond formation to
pre-equilibrium) or competition between different mechanism;31
however, at this point the kinetic data do not provide a definite
answer.
To obtain further insight into the deactivation of QUINOX
(9) by a large excess of allyltrichlorosilane (2), for which
coordination of Si to the MeO group could be suggested (21 in
Scheme 3), we have investigated the methyl catalyst 19, where
the latter interaction cannot exist. Indeed, in the allylation of
p-trifluomethylbenzaldehyde (1g), increasing the silane/catalyst
ratio from 10:1 to 50:1 (while the concentration of aldehyde
1g was kept constant) resulted in the proportional (∼4-fold)
increase in the reaction rate; that is, no catalyst deactivation
was taking place. However, it is pertinent to note that at the
silane/catalyst ratio 10:1, QUINOX (9) reacted nearly 3.5 times
faster than its methyl analogue 19. Hence, the MeO group of
QUINOX can be regarded as the key feature to attain sufficient
reaction rates and high enantioselectivities (with the exception
of electron-rich aldehydes, such as 1h).
Figure 3. Hammett plot for the allylation of 1 with 2 catalyzed by 9.
Scheme 3. Allylation of Benzaldehyde 1a Catalyzed by QUINOX
(9)
Quantum Chemical Calculations
To get a better insight into the mechanism of the allylation
catalyzed by QUINOX (9), a computational analysis of the reaction
coordinate and the transition states (TS) was carried out using
benzaldehydes 1a and 1h as model substrates.
TS is nonionic, weakly polar,37 implying that the Lewis acidity
of the silicon in 5 is just sufficient to allow binding of aromatic
aldehydes but without affecting their electrophilicity. At higher
σ values, deviation from linearity was observed,38,39 indicating
a gradual shift in the RLS toward precoordination. These data
were corroborated by the measurement of the secondary isotope
effect (SIE), namely by comparing the relative reactivities of
ArCDdO vs ArCHdO. In the case of 1h, the SIE value of
0.70 indicates a late transition state, closely resembling the sp3
hybridized product. The sp2-character in the TS gradually
increased as evidenced by the increase in the SIE to 0.81 (for
1a) and further to 0.86 (for 1g), confirming a shift toward
precoordination RLS, which is reflected in the first order kinetics
in 1g.
On the basis of the kinetic data, the following mechanistic
scenario can be proposed (Scheme 3). Allyltrichlorosilane (2)
and QUINOX (9) reversibly form the reactive complex 20,
which is deactivated by a large excess of the silane, presumably
via binding a second molecule of 2 to generate a catalytically
inactive complex, such as 21 (vide infra). The negative order
in silane 2 at higher concentrations arises from the need to
release the extra silane molecule to return to 20. The next
reaction step, binding the aldehyde, can proceed via associative
(22) or dissociative (23) mechanisms.22a Since the octahedral
complex 22 is more crowded than the trigonal bipyramidal 23,
Computational Details. All density functional theory (DFT)
calculations were carried out using the program Turbomole 5.7.40
The Perdew-Burke-Ernzerhof (PBE) functional41 has been used
throughout. The calculations were expedited by expanding the
Coulomb integrals in an auxiliary basis set, the resolution-of-identity
(RI-J) approximation.42,43 All geometry optimizations were carried
out using the 6-31G(d) basis set,44 whereas the single point energies
were recomputed using the TZVP basis set.45 To account for the
description of dispersion forces (which are not described by the
currently used DFT functionals), the empirical dispersion parameters
were added to both the energy and gradients during geometry
optimization and the calculations of the single point energies
[DFT(+D) method].46,47 Here, default parameters47 were used,
except for the C6 parameters for Si, where the values for P were
used, and Cl, where we used the values of Neumann and Perrin.48
This method has been shown to essentially yield the data with ∼1
kcal mol-1 acuracy in comparison with the reference CCSD(T)
values.47
(40) Ahlrichs, R.; Bär, M.; Häser, M.; Horn, H.; Kölmel, C. Chem. Phys.
Lett. 1989, 162, 165.
(41) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1996, 77,
3865.
(42) Eichkorn, K.; Treutler, O.; Öhm, H.; Häser, M.; Ahlrichs, R. Chem.
Phys. Lett. 1995, 240, 283.
(43) Eichkorn, K.; Weigen, F.; Treutler, O.; Ahlrichs, R. Theor. Chim. Acta
1997, 97, 119.
(37) A similar value (F ) 1.18) has been reported for the allylation of
aldehydes with allysiliconates: Hosomi, A.; Kohra, S.; Ogata, K.;
Yanagi, T.; Tominaga, Y. J. Org. Chem. 1990, 55, 2415.
(38) For an analogous nonlinear behavior at high σ values, observed for
the silicon-directed aldol condensation, see: Myers, A. G.; Widdowson,
K. L.; Kukkola, P. J. J. Am. Chem. Soc. 1992, 114, 2765.
(39) The plot incorporating all the data points showed lower statistical
significance of R2 ) 0.87.
(44) Hehre, W. J.; Radom, L.; Schleyer, P. I. v. R.; Pople, J. A. Ab Initio
Molecular Orbital Theory; Wiley-Interscience: New York, 1986.
(45) Schäfer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829.
(46) Grimme, S. J. Comput. Chem. 2004, 25, 1463.
ˇ
(47) Jurecˇka, P.; Cerný, J.; Hobza, P.; Salahub, D. J. Comput. Chem. 2007,
28, 555.
(48) Neumann, M. A.; Perrin, M. A. J. Phys. Chem. B 2005, 109, 15531.
9
5346 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008

Mechanism of Asymmetric Allylation of Aldehydes
A R T I C L E S
Figure 4. The equilibrium geometries of the most stable reactant complexes (RC), transition states (TS), and product complexes (PC) along the reaction
coordinate for the associative pathway of allylation of benzaldehyde (1a) catalyzed by (R)-(+)-QUINOX (9). The calculated values for ∆G were obtained
at the RI-PBE(+D)/TZVP//RI-PBE(+D)/6-31G(d) level; all distances are in Å.
To account for solvation effects, the conductor-like screening
model (COSMO)49,50 was employed with the dielectric constant
corresponding to acetonitrile (ꢀr ) 36.6). Gibbs free energy was
then calculated as the sum of these contributions (equation 1):
in step 1 and optimizing all other degrees of freedom by the
DFT(+D)/COSMO method. The difference between the DFT(+D)/
COSMO energy at the in vacuo TS geometry and the relaxed
geometry is denoted as the solvent relaxation term and is regarded
to represent an accurate approximation to the true transition state
in solution, calculated with the DFT(+D) method.
G ) Eel(+D) + Gsolv + EZPE - RTln(qtransqrotqvib
)
(1)
Finally, for benchmarking purposes on the model uncatalyzed
reaction, the CCSD(T) method was used53 with the aug-cc-pVDZ
basis set.54 These calculations were carried out using the MOLPRO
program.55,56
where ∆Eel(+D) is the in vacuo energy of the system [at the RI-
PBE(+D)/TZVPP level, with the geometry optimized at the RI-
PBE(+D)/6-31G(d) level (vide supra)], ∆Gsolv is the solvation free
energy [at the RI-PBE(+D)/6-31G(d) level] and the EZPE - RT
ln(qtrans qrot qvib) term is the zero-point energy, thermal corrections
to the Gibbs free energy, and entropic term (obtained from a
frequency calculation with the same method and software as for
the geometry optimizations at the RI-PBE/6-31G(d) level, 298 K,
and 1 atm pressure, using an ideal-gas approximation51).52
Transition state (TS) optimizations were performed as follows:
First, the TS initial geometry was obtained by the constrained
geometry optimization on the interpolated reaction coordinate,
defined by the Si(1)-C(2) and C(4)-C(5) distances (Figure 4),
which were assumed to attain the discrete values between the
reactant and product with all the remaining internal coordinates
optimized. The maximum on this approximate reaction coordinate
was then taken as the initial guess for the saddle-point optimization
along the eigenvector corresponding to the vibration with the
imaginary frequency. Because of the lack of the analytic second
derivatives in both DFT(+D) and COSMO solvation methods, the
approximate method was used to obtain the transition states in
solution: (1) the in vacuo optimization of transition state according
to the above procedure; (2) constraining the Si(1)-C(2) and
C(4)-C(5) distances to the values corresponding to the TS acquired
Results of Computational Analysis
Approximately twenty structural variants for each of the
reaction mechanisms using (R)-(+)-9 were examined. Notably,
the chelate arising by a bidentate coordination of QUINOX (9)
to allyltrichlorosilane through the N-oxide and methoxy groups
was found to be less stable by nearly 30 kcal mol-1 than the
monocoordinated species 20, which also rules out any contribu-
tion of the bidentate mode in the transition state.57 The results
covering the most stable structures are summarized in Table 3;
the structures of the reactant and product complexes and
transition states for the (R)- and (S)-reaction channels in the
associative mechanism are shown in Figure 4. According to this
analysis, the reaction commences with the formation of a
ˇ
ˇ
(53) (a) Cížek, J. J. Chem. Phys. 1966, 45, 4256–4266. (b) Cížek, J. AdV.
Chem. Phys. 1969, 14, 35–89. (c) Raghavachari, K.; Trucks, G. W.;
Pople, J. A.; Head-Gordon, M. Chem. Phys. Lett. 1989, 157, 479.
(54) Woon, D. E.; Dunning, T. H., Jr. J. Chem. Phys. 1993, 98, 1358.
(55) Werner, H.-J.; et al. MOLPRO,. a package of ab initio programs,
version 2002.6, 2006.
(49) Klamt, A.; Schuurmann, G. J. Chem. Soc., Perkin Trans. 2 1993, 799.
(50) Schäfer, A.; Klamt, A.; Sattel, D.; Lohrenz, J. C. W.; Eckert, F. Phys.
Chem. Chem. Phys. 2000, 2, 2187.
(51) Jensen, F. Introduction to Computational Chemistry; John Wiley &
Sons: 1999.
(52) The free energy calculated according to equation 1 is regarded as a
good approximation to ∆G in diluted solution.
(56) Hampel, C.; Peterson, K.; Werner, H.-J. Chem. Phys. Lett. 1992, 190, 1.
(57) In fact, this value was obtained by constraining the complex geometry
to preserve the bidentate binding mode of QUINOX, because this
structure is not a minimum on the potential-energy surface of the
complex and its geometry optimization converges to a monodentate
structure.
9
J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008 5347

A R T I C L E S
Malkov et al.
simple pyridine N-oxide by 3–5 kcal mol-1 60
.
This value is
Table 3. The Calculated Thermochemical Data for the Reaction of
1a with 2 Catalyzed by (R)-(+)-QUINOX (9)a
comparable with the strength of the aromatic interaction between
benzaldehyde and the catalyst expected in the TS arrangement.
Moreover, the difference in the attractive π-π interactions
between the (R)- and (S)-transition states also constitutes the
largest contribution to the enantiodifferentiation since the
calculated difference in the dispersion energy stabilizations
between the (R)- and (S)-enantiomers amounts to 1.1 kcal
product
configuration
∆G, reactant ∆G‡, transition ∆G, product
entry
mechanism
complex
state
complex
1
2
3
4
associative
associative
dissociative
dissociative
R
S
R
S
23.4
23.8
19.3
19.8
23.9
25.9
26.0
24.5
1.1
1.4
1.1
1.4
mol-1 61,62
.
a All values are in kcal mol-1
.
Conclusions
transient reactant complex (RC). For the associative mechanism
(22), this complex is higher in energy than the assembly of the
isolated reactants and the catalyst by 23.4/23.8 kcal mol-1 (Table
3, entries 1 and 2) and for the dissociative mechanism (23) by
19.3/19.8 kcal mol-1 (entries 3 and 4); note the greater stability
of the RC with cationic trigonal bipyramidal structure formed
in the dissociative pathway. The calculated transition state
The kinetic and computational data indicate that the allylation
of aldehydes 1 with allyltrichlorosilanes 2–4, catalyzed by
QUINOX (9), is likely to proceed via an associative pathway
involving neutral, octahedral silicon complex 22. In contrast to
catalysis by chiral monodentate phosphoramides, such as 10,
only one molecule of QUINOX (9) is involved in the rate- and
selectivity-determining step. The reaction presumably proceeds
via a cyclic chairlike63 transition state 5 and is characterized
by high enantio- and diasteroselectivity. A drop in enantiose-
lectivity observed in the case of electron-rich substrates appar-
ently originates from narrowing the energy gap between the
(R)- and (S)-reaction channels in the associative mechanism (22).
barriers are ∆G‡
) 23.9 (R) and 25.9 (S) kcal mol-1; and
assoc
∆G‡dissoc ) 26.0 (R) and 24.5 (S) kcal mol-1 for the associative
and dissociative mechanism, respectively.58 Significantly, a
narrow energy gap between the RC and TS in the associative
mechanism correlates well with the experimentally observed
borderline position of the RLS between precoordination of the
substrate aldehyde and C-C bond formation. The overall
reaction thermodynamics, after the catalyst has dissociated from
the product complex (PC), is ∆G ) -10.3 kcal mol-1. The
associative mechanism predicts the formation of (R)-6a in 97%
ee (at 233.15 K) for an enantiopure catalyst, which is in a good
agreement with the experimental value of 87% ee (Table 1, entry
1) attained with the catalyst of 98% ee. Significantly, the
dissociative route favors the formation of the opposite, that is,
(S)-enantiomer.59 Assuming that both mechanisms operated
concurrently, the calculations would predict the formation of
(R)-6a in 53% ee.
Acknowledgment. The paper is dedicated to Professor Victor
Sniecˇkus on the occasion of his 70th birthday. We thank the
EPSRC for Grant No. GR/T27051/01, the Ministry of Education
and Science of Spain for a fellowship to P. R.-L., the Ministry
ˇ
of Education of the Czech Republic (MŠMT CR) for Grant No.
LC512, the Grant Agency of the Czech Republic for Grant No.
203/08/0350, the European Socrates-Erasmus exchange program
for scholarships to L.B. and L.D., Dr. Alfred Bader for an
additional support, and Dr. DeLiang Long for the X-ray structure
determination. We also thank Profs. Guy C. Lloyd Jones and
Pavel Hobza for helpful discussions.
In the case of p-methoxybenzaldehyde 1h, the associative
mechanism is even more favored with the TS barriers ∆G‡
Supporting Information Available: Complete ref 55, general
experimental methods, representative allylation and crotylation
procedures, NMR spectra of new compounds, all kinetic data
and the equilibrium structures, and energies of all the compu-
tationally studied species. This material is available free of
charge via the Internet at http://pubs.acs.org.
assoc
) 26.1 (R) and 26.9 (S) kcal mol-1, while the corresponding
values for the dissociative pathway are ∆G‡
) 29.3 (R)
dissoc
and 29.7 (S) kcal mol-1. Significantly, the energy gap between
the (R)- and (S)-reaction channels in the associative mechanism
for 1h is reduced to 0.8 kcal mol-1 and predicts the formation
of (R)-6h in 62% ee (at 273.15 K; in a very good agreement
with 45% ee observed experimentally at this temperature; Table
1, entry 18). Note that a higher TS barrier for the formation of
(R)-6h compared to (R)-6a correlates well with the experimen-
tally observed higher reaction rates of 1a versus 1h.
JA711338Q
(60) The TS barrier calculated for the reaction of benzaldehyde and
allyltrichlorosilane catalyzed by pyridine N-oxide is ∆G‡ ) 27.5 kcal
mol-1
.
The calculations further show that in the TS the methoxy-
naphthalene unit of QUINOX (9) and the phenyl ring of
benzaldehyde (1a) are arranged in a parallel orientation, which
makes the TS for QUINOX lower in energy than that for the
(61) Decomposition of the energy differences in the free energy barriers
in the associative mechanism between the (R) and (S) enantiomers
(in kcal mol-1) are as follows: in vacuo energies ∆Egp ) 0.3; solvation
free energies ∆Gsolv ) -0.7; in vacuo entropic terms ∆(-T∆S)gp
-0.5; and dispersion energy stabilizations ∆Edisp ) -1.1.
)
(62) Computational analysis of π-π interactions on the truncated system
[benzene-naphthalene] using the TS(R) and TS(S) geometries (Fig.
4), where benzene and naphthalene represented respective interacting
parts of benzaldehyde and QUINOX, produced the interaction energies
∆E ) -2.8 (R) and -2.6 (S) kcal mol-1. In the real system, the
difference in arene-arene interactions energies for (R)- and (S)-
manifolds is expected to further increase.
(58) Lowering the barrier by 4.2 kcal mol-1 for the (R) reaction channel
(compared to the uncatalyzed reaction) corresponds to a ∼103 increase
of the reaction rate, which is in agreement with the experimental
observations.
(59) There is an uncertainty in the translational entropy of the Cl- ion,
which has to be taken into account in the dissociative reaction
mechanism; however, this factor is unlikely to affect the overall picture
significantly.
(63) According to the calculations carried out for 1a and 1h, the chair
conformation is preferred over the twisted boat for both (R)-and (S)-
channels by about 1.5–2.0 kcal mol-1
.
9
5348 J. AM. CHEM. SOC. VOL. 130, NO. 15, 2008