Catalysis of the Michael addition reaction by late transition metal complexes of BINOL-derived salens

Article abstract of DOI:10.1021/jo025993t

Salen metal complexes incorporating two chiral BINOL moieties have been synthesized and characterized by X-ray crystallography. The X-ray structures show that this new class of Ni-BINOL-salen catalysts contains an unoccupied apical site for potential coor

Full text of DOI:10.1021/jo025993t

Ca ta lysis of th e Mich a el Ad d ition Rea ction by La te Tr a n sition
Meta l Com p lexes of BINOL-Der ived Sa len s
Venkatachalam Annamalai, Erin F. DiMauro, Patrick J . Carroll, and Marisa C. Kozlowski*
Department of Chemistry, Roy and Diana Vagelos Laboratories, University of Pennsylvania,
Philadelphia, Pennsylvania 19104-6323
marisa@sas.upenn.edu
Received May 30, 2002
Salen metal complexes incorporating two chiral BINOL moieties have been synthesized and
characterized by X-ray crystallography. The X-ray structures show that this new class of Ni-BINOL-
salen catalysts contains an unoccupied apical site for potential coordination of an electrophile and
naphthoxides that are independent from the Lewis acid center. These characteristics allow
independent alteration of the Lewis acidic and Brønsted basic sites. These unique complexes have
been shown to catalyze the Michael reaction of dibenzyl malonate and cyclohexenone with good
selectivity (up to 90% ee) and moderate yield (up to 79% yield). These catalysts are also effective
in the Michael reaction between other enones and malonates. Kinetic data show that the reaction
is first order in the Ni‚Cs-BINOL-salen catalyst. Further experiments probed the reactivity of the
individual Lewis acid and Brønsted base components of the catalyst and established that both
moieties are essential for asymmetric catalysis. All told, the data support a bifunctional activation
pathway in which the apical Ni site of the Ni‚Cs-BINOL-salen activates the enone and the
naphthoxide base activates the malonate.
In tr od u ction
A number of recent studies have shown that multi-
functional ligands possess useful characteristics for
1
catalysis and asymmetric synthesis. For example, ligands
have been reported in which one portion engages a Lewis
acid moiety that coordinates an electrophilic substrate
while another portion of the ligand coordinates to the
nucleophilic substrate partner. The positioning by such
a catalyst of the two reactant partners in close proximity
and with the correct relative geometry facilitates a
reaction in a manner similar to that of some enzyme
catalysts (i.e., ligases). Dual coordination by such as-
semblies further enables the reaction by simultaneously
enhancing the electrophilic character of one partner and
the nucleophilic character of the other partner.2 An
example of this motif can be found in the heterobimetallic
complexes (1) developed by Shibasaki et al. (Figure 1).1
In these complexes, the central lanthanide metal is
proposed to coordinate the electrophile while the hemi-
lable BINOLate oxygens deprotonate the nucleophile.
F IGURE 1. Bifunctional catalysts.
While several bifunctional ligands that operate along
these lines have been discovered, relatively few ex-
1,3
c,d
amples exist in which electronically decoupled sites are
present for individual activation of the electrophilic and
4
nucleophilic substrates. One of the earliest examples of
this latter motif can be found in complex 2 developed by
Hayashi and Ito,4 which catalyzes the aldol reaction
between isocyanoacetates and aldehydes. Complex 2
contains a gold metal site that coordinates both the
a
(
1) (a) Corey, E. J .; Bakshi, R. K.; Shibata, S. J . Am. Chem. Soc.
1
987, 109, 5551-5553. (b) Kitamura, M.; Suga, S.; Niwa, M.; Noyori,
R. J . Am. Chem. Soc. 1995, 117, 4832-4842. (c) Shibasaki, M.; Sasai,
H.; Arai, T. Angew. Chem., Int. Ed. Engl. 1997, 36, 1236-1256. (d)
Kim, Y. S.; Matsunaga, S.; Das, J .; Sekine, A.; Oshima, T.; Shibasaki,
M. J . Am. Chem. Soc. 2000, 122, 6506-6507 and references therein.
(3) For other bimetallic catalysts, see: (Al‚Li) (a) Keller, E.; Veld-
man, N.; Spek, A. L.; Feringa, B. L. Tetrahedron: Asymmetry 1997, 8,
3403-3413. (b) Arai, T.; Hu, Q.; Zheng, X.; Pu, L.; Sasai, H. Org. Lett.
2000, 2, 4261-4263. (c) Manickam, G.; Sundarajan G. Tetrahedron
1999, 55, 2721-2736. (d) Sundararajan, G.; Prabagaran, N. Org. Lett.
2001, 3, 389-392. (Zn‚Li) (e) Nina, S.; Soai, K. J . Chem. Soc., Perkin
Trans. 1 1991, 2717-2720. (Zn‚Zn) (f) Trost, B. M.; Ito, H.; Silcoff, E.
R. J . Am. Chem. Soc. 2001, 123, 3367-3368. (g) Yoshikawa, N.;
Kumagai, N.; Matsunaga, S.; Moll, G.; Ohshima, T.; Suzuki, T.;
Shibasaki, M. J . Am. Chem. Soc. 2001, 123, 2466-2467.
(
6
2
e) Gample, M. P.; Smith, A. R. C.; Wills, M. J . Org. Chem. 1998, 63,
068-6071. (f) Yamakawa, M.; Ito, H.; Noyori, R. J . Am. Chem. Soc.
000, 122, 1466-1478 and references therein. (g) Trost, B. M.; Ito, H.
J . Am. Chem. Soc. 2000, 122, 12003-12004. (h) List, B.; Pojarliev, P.;
Castello, C. Org. Lett. 2001, 3, 573-575 and references cited herein.
(2) For reviews on bifunctional catalysis, see: (a) Sawamura, M.;
Ito, Y. Chem. Rev. 1992, 92, 857-871. (b) van den Beuken, E. K.;
Feringa, B. L. Tetrahedron 1998, 54, 12985-13011. (c) Rowlands, G.
J . Tetrahedron 2001, 57, 1865-1882.
1
0.1021/jo025993t CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/11/2003
J . Org. Chem. 2003, 68, 1973-1981
1973

Annamalai et al.
An important structural consideration in the design
of these bifunctional scaffolds was the nature of the
tether connecting the salen to the basic functional group.
This tether must be sufficiently short and/or rigid to
prevent internal complexation of the Brønsted base to
the Lewis acid. Compounds 6 and 7 are inherently rigid
by virtue of the biaryl bonds. We proposed that internal
complexation is not feasible in these complexes; sub-
sequent crystallographic characterization supports this
assertion. In previous work, we have shown the utility
of these BINOL-salen bifunctional catalysts in the Michael
7
reaction of cyclohexenone and benzyl malonate. In this
report we explore the full structural and mechanistic
details of these catalysts and their scope in the Michael
addition of carbon nucleophiles to enones.
Resu lts a n d Discu ssion
Syn t h esis of Bip h en ol- a n d BINOL-Sa len s. Bi-
phenol ligand 10 was easily prepared from biphenol in a
simple four-step procedure as shown in Scheme 1. After
protection of biphenol (8) as the bis-MOM ether, mono-
8
formylation is accomplished via directed lithiation.
Subsequent deprotection yields aldehyde 9, which is
condensed with (R,R)-trans-cyclohexanediamine. The
metalated catalyst 11 is prepared from 10 by heating
2
with Ni(OAc) .
F IGURE 2. Bifunctional salen catalysts.
SCHEME 1a
isocyanoacetate and the aldehyde while the pendant
amine moiety effects the enolization of the isocyano-
acetate. Even so, the potential for catalyst deactivation
by coordination of the pendant amine to the gold center
arises in such a system.
We directed our efforts to the construction of catalyst
systems in which the Lewis acid and Brønsted base
functional groups can be tailored independently. The
catalyst design began from general structure 3 (Figure
2
), which incorporates a basic moiety into a structurally
well-defined and rigid salen complex. The salen scaffold
was chosen because these salens (4-7) are easily made
and their modular nature allows the rapid synthesis of
several analogues. In addition, the functionalized salen
complexes provide an accessible Lewis acid center for
electrophile activation and a basic functional group for
activation of the nucleophile. Finally, the acidic and basic
sites are electronically decoupled and can be indepen-
dently attenuated. In prior work, we found that struc-
tures 4 and 5 function as potent Lewis acid/Lewis base
a
Reagents and conditions: (a) (i) NaH, MOMCl, 100%; (ii)
nBuLi, DMF, TMEDA, 57%; (iii) HCl, THF, 0 °C, 90%. (b) EtOH,
∆, H2N-X-NH2 (c) Ni(OAc)2, EtOH, ∆.
5
catalysts, but were poor Lewis acid/Brønsted base
6
BINOL-salen ligands 13-18 and their corresponding
Ni complexes were prepared in an analogous manner as
shown in Scheme 2.7 Alternatively, a more efficient
preparation of the metallosalens could be accomplished
by simultaneous condensation of the aldehyde with the
corresponding amine and metalation (Scheme 2, reaction
conditions c).
catalysts. Herein, we focus on the evaluation of struc-
tures 6 and 7 as Lewis acid/Brønsted base catalysts.
(
4) (a) Ito, Y.; Sawamura, M.; Hayashi, T. J . Am. Chem. Soc. 1986,
1
08, 6405-6406. (b) Kitajima, H.; Ito, K.; Aoki, Y.; Katsuki, T. Bull.
Chem. Soc. J pn. 1997, 70, 207-217. (c) Sibi, M. P.; Cook, G. R.; Liu,
P. Tetrahedron Lett. 1999, 40, 2477-2480. (d) Hamashima, Y.; Sawada,
D.; Nogami, H.; Kanai, M.; Shibasaki, M. Tetrahedron 2001, 57, 805-
8
14 and references therein. (e) Ooi, T.; Kondo, Y.; Maruoka, K. Angew.
In a screening of azaphilic late transition metals, Ni-
II) was identified as the optimal Lewis acid component.
Chem., Int. Ed. Engl. 1997, 36, 1183-1185. (f) France, S.; Wack, H.;
Hafez, A. M.; Taggi, A. E.; Witsil, D. R.; Lectka, T. Org. Lett. 2002, 4,
7
(
1
603-1605. (g) J osephsohn, N. S.; Kuntz, K. W.; Snapper, M. L.;
Hoveyda, A. H. J . Am. Chem. Soc. 2001, 123, 11594-11599.
(7) DiMauro, E. F.; Kozlowski, M. C. Org. Lett. 2001, 3, 1641-1644.
(
5) (a) DiMauro, E. F.; Kozlowski, M. C. Org. Lett. 2001, 3, 3053-
(8) (a) For a related biphenol directed lithiation see: Rudkevich, D.
3
1
2
056. (b) DiMauro, E. F.; Kozlowski, M. C. J . Am. Chem. Soc. 2002,
24, 12668-12669. (c) DiMauro, E. F.; Kozlowski, M. C. Org. Lett.
002, 4, 3781-3784.
M.; Verboom, W.; Brzozka, Z.; Palys, M. J .; Stauthamer, W. P. R. V.;
van Hummel, G. J .; Franken, S. M.; Harkema, S.; Engbersen, J . F. J .;
Reinhoudt, D. N. J . Am. Chem. Soc. 1994, 116, 4341-4351. (b) For
directed lithiation of MOM-protected phenols see: Townsend, C. A.;
Bloom, L. M. Tetrahedron Lett. 1981, 22, 3923-3924.
(
6) DiMauro, E. F.; Mamai, A.; Kozlowski, M. C. Organometallics
2
003, 22, 850-855.
1
974 J . Org. Chem., Vol. 68, No. 5, 2003

Catalysis of the Michael Addition Reaction
SCHEME 2
reaction substrates (eq 2). Increased enantioselectivity
is observed with increased atomic radius/polarizability
of the naphthoxide counterion, with the Cs complexes
7
yielding the best results. It appears that more basic (less
tightly coordinated) naphthoxides are necessary for
optimal reactivity and selectivity.
Various solvents were examined with use of Ni‚Cs
BINOL-salen catalyst 25a (Table 1). Although catalyst
solubility was improved with increased solvent polarity
(CH
2 2
Cl , THF, DMF), strongly polar coordinating solvents
(DMF) were detrimental to enantioselectivity. A solvent
screen was also done with the more selective Ni‚Cs
BINOL-salen catalyst 25d and various ethereal solvents.
Catalyst solubility was worse in ether, DME, and methyl-
tetrahydrofuran when compared to THF. This correlated
to lower yields and selectivity.
TABLE 1. Solven t Scr een of BINOL-Sa len Ca ta lysts 25a
a n d 25d in th e Mich a el Ad d ition Rea ction (Eqs 1 a n d 2)
catalyst
5a
25a
25a
solvent
DMF
yield (%)a
ee (%)b
The Ni(II) derivatives of salens 13-18 were therefore
2
77
80
66
48
81
79
65
43
66
5
31
4
38
0
71
34
27
21
prepared from Ni(OAc)
2
as shown in Scheme 2. Consis-
THF
tent with formation of the illustrated complexes, a lower
CdN stretch in the IR spectra of the complexes relative
CH3CN
CH2Cl2
toluene
THF
2
2
2
2
5a
5a
5d
5d
2
to that of the free salen was observed. Also, the C -
symmetry element is indicated from the 13C NMR spectra
of the complexes.
2-MeTHF
25d
25d
Et O
2
Bip h en ol- a n d BINOL-Sa len Ca ta lysts in th e
Mich a el Rea ction . Replacement of the two hydrogens
in Ni-BINOL-salen complexes 19-24 with Li, Na, K, or
Cs provides metal complexes containing Lewis acidic and
Brønsted basic sites. From the trends observed in the
crystal structure of the protonated form, interactions
between the naphthoate counterion and the salen oxy-
DME
a
Isolated yields. b Enantiomeric excess determined by HPLC.
Absolute configuration (R) based on optical rotation values.
TABLE 2. Ni‚Cs-Sa len Na p h th oxid e Com p lexes (Eq 2)
in th e Mich a el Ad d ition Rea ction (Eq 1)
yield (%)a
ee (%)b
entry
Ni salen
catalyst
9
gens are likely to be weak. As a result, it should be
1
2
3
4
5
6
19
20
21
22
23
24
25a
25b
25c
25d
25e
25f
80
49
84
79
88
83
31 (R)
31 (R)
54 (R)
71 (R)
42 (R)
21 (R)
possible to modulate the reactivity of the salen Lewis acid
and the naphthoate base independently.
The Ni-BINOL-salen complexes 19-24 were examined
in the Michael reaction10 of cyclohexenone with benzyl
malonate (eq 1). The active catalysts were generated from
the preformed Ni-BINOL-salen complexes 19-24 via
treatment with a base in THF prior to addition of the
a
Isolated yields. b Enantiomeric excess determined by HPLC.
Absolute configuration based on optical rotation values.
The structural basis for stereoinduction was examined
through variation of the configuration and the confor-
mational mobility afforded by the salen diamine linker
(
9) DiMauro, E. F.; Kozlowski, M. C. Organometallics 2002, 21,
454-1461.
10) For recent reviews, see: (a) Comprehensive Asymmetric
1
(
Catalysis; J acobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer:
New York; 1999; Chapter 31. (b) Leonard, J .; Ciez-Barra, E.; Merino,
S. Eur. J . Org. Chem. 1998, 2051-2061.
(Table 2). Exchanging the ethylenediamine linker in the
Ni‚Cs-BINOL-salen catalyst with a diaminobenzene or
J . Org. Chem, Vol. 68, No. 5, 2003 1975

Annamalai et al.
TABLE 3. Ma tch ed Sa len Na p h th oxid e Ca ta lyst 25d in
th e Mich a el Ad d ition Rea ction a
a
0 mol % catalyst. b Isolated yields. c Enantiomeric excess
1
determined by HPLC. Absolute configuration based on optical
rotation values. d Enantiomeric excess determined from the
1
H
NMR spectra of the (S,S)-2,3-pentanediol ketals.
F IGURE 3. Results from the reaction of dibenzyl malonate
with cyclohexenone (eq 1) employing 10 mol % of the indicated
complexes (generated in situ, eq 2).
1, Table 3). Similar enantioselectivity was observed in
the addition of dibenzyl malonate to cycloheptenone (73%
ee, entry 6) compared to cyclohexenone, but much lower
selectivity was observed with cyclopentenone (38-50%
ee, entries 7 and 8). Moderate selectivity could also be
obtained with acyclic enones such as chalcone (72% ee,
entry 10), but catalyst turnover was compromised.
Cr ysta llogr a p h ic Str u ctu r es of th e BINOL-Sa len
Ni Com p lexes. Crystal structures of the nickel com-
plexes 19 and 22 provided definitive evidence that the
proposed salen metal complexes were formed. The struc-
tures of 19 and 22 are remarkably similar (Figures 4 and
5, Table 4), including similar conformational forms in the
unit cells for the two different compounds (19a vs 22a
and 19b vs 22b).
a diphenylethylene linker resulted in a less effective
catalyst (entries 2 and 6 vs entry 1). Switching to a cis-
diaminocyclohexane linker resulted in a more effective
catalyst (entry 5 vs entry 1). The trans-diaminocyclohex-
ane derivatives 25c and 25d (Figure 3) provided the
largest degree of improvement (entries 3 and 4). The
latter proved to be the matched catalyst system, provid-
ing the Michael adduct in 71% ee (entry 4). Upon cooling
the reaction to -40 °C, the Michael adduct was obtained
in up to 90% ee with 10 mol % 25d (Figure 3).11
Interestingly, both of the diaminocyclohexane derived Ni‚
Cs-BINOL-salen catalysts 25c and 25d catalyzed the
Michael reaction with improved selectivity compared to
the ethylenediamine analogue 25a (Figure 3). This
phenomenon may result from the stereochemical influ-
ence and conformational restriction imposed by the
cyclohexyl backbone. Since catalysts 25c and 25d , which
differ in the configuration of the bisimine linker, afford
the same prevailing enantiomer in the product, the biaryl
chirality appears to be the dominant control element
responsible for facial selectivity. This conclusion is sup-
ported by the results from the biphenol-derived complex
Despite their apparent structural similarities, bimetal-
lic catalysts derived from complexes 19 and 22 cause
significantly different levels of enantioselection (31% vs
7
1% ee) in the Michael addition reaction. Although solid-
state crystallographic structures may not be representa-
tive of the solution structures due to crystal packing
forces, several points encouraged us that comparison of
1
9a and 22a would be informative. First, the NMR
spectra of 19 and 22 were consistent with C -symmetric
2
structures as found in 19a and 22a . Second, the core
structures of the two conformational forms found in the
crystal structure of 19 are similar, as are the two
conformational forms of 22, indicating that these core
structures likely persist in different conformational forms
such as those found in solution. Finally, prior reports
indicate that solid-state Ni(II) salen geometries (square
planar, square pyramidal, or octahedral) are retained in
solution.12
2
9 (Figure 3). While complex 29 was an efficient catalyst
for the Michael reaction (90% yield), the product was
obtained in 0% ee. Thus, a rigid biaryl axis is necessary
for asymmetric induction.
The substrate scope of the matched Ni‚Cs-BINOL-
salen catalyst 25d in the Michael addition reaction was
next investigated (Table 3). The highest yield and enan-
tioselectivity were observed for the reaction between
cyclohexenone and dibenzyl malonate (entries 1-4, Table
In the pseudo-C -symmetric complexes 19a and 22a ,
2
3
). When the reaction temperature is lowered enantiose-
the binaphthyl dihedral angles are compressed to 68.1°
lection is improved at the expense of reactivity. With
dimethyl malonate as the nucleophile, catalyst reactivity
and selectivity are significantly lower (entry 5 vs entry
and 63.2° (for the C2-C17 biaryl) and 65.4° and 64.7°
(for the C15-C27 biaryl), respectively. The binaphthyl
dihedral angle in uncoordinated BINOL is 78.3°. Even
(
C.
11) Reaction time was 48 h at 20, 0, and -20 °C and 5 d at -40
(12) For example: Mockler, G. M.; Chaffey, G. W.; Sinn, E.; Wong,
H. Inorg. Chem. 1972, 11, 1308-1314.
°
1
976 J . Org. Chem., Vol. 68, No. 5, 2003

Catalysis of the Michael Addition Reaction
F IGURE 4. Top and side views for both unit cell structures of Ni-BINOL-salen 19 X-ray structure.
F IGURE 5. Top and side views of both unit cell structures for Ni-BINOL-salen 22 X-ray structure.
so, the naphthoate oxygens simply cannot reach a posi-
tion to bind to the Ni. Rather these naphthoates are in
the proper orientation for delivery of a stabilized anion
to an electrophile bound to the salen metal center.13 An
overlay (Figure 6) shows that complexes 19a and 22a
differ in the positioning of this putatively important
naphthoate with respect to the salen core.
In the case of 19, the ethylenediamine linker responds
to the helical twist imposed by the BINOL portions which
causes the ethylene to cant from lower left to upper right
in the orientation shown in Figure 6b. In the (R,R)-
cyclohexanediamine case, such an orientation would
require the amines of the trans-cyclohexanediamine to
be axial, which is not possible in this structure. When
the amine groups are equatorial, the cant of the diamine
linking portion is reversed (upper left to lower right).
(13) Examination of nickel salen complexes in the Cambridge
Structural Database (The United Kingdom Chemical Database Ser-
vice: Fletcher, D. A.; McMeeking, R. F.; Parkin, D. J . Chem. Inf.
Comput. Sci. 1996, 36, 746) revealed square-pyramidal or octahedral
adducts upon coordination of further ligands. In general, the salen
ligand coordinates in a square-planar array. Square pyramidal: (a)
Elerman, Y.; Kabak, M.; Atakol, O. Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 1993, 49, 1905-1906. (b) Clarke, N.; Cunningham,
D.; Higgins, T.; McArdle, P.; McGinley, J .; O’Gara, M. J . Organomet.
Chem. 1994, 469, 33-40. Octahedral: (c) Gomes, L.; Sousa, C.; Freire,
C.; de Castro, B. Acta Crystallogr., Sect. C: Cryst. Struct. Commun.
2000, 56, 1201-1203.
J . Org. Chem, Vol. 68, No. 5, 2003 1977

Annamalai et al.
TABLE 4. Selected Bon d Dista n ces (Å), Bon d An gles (d eg), a n d Dih ed r a l An gles (d eg) for 19 a n d 22
1
9a
19b
22a
22b
Ni-N1
Ni-N2
Ni-O1
Ni-O2
O1-C1
O3-C18
O2-C16
O4-C28
1.832(7)
1.823(6)
1.855(5)
1.846(5)
1.342(9)
1.377(8)
1.345(9)
1.393(9)
1.821(7)
1.847(6)
1.849(4)
1.835(5)
1.332(9)
1.374(8)
1.333(9)
1.356(7)
1.858(8)
1.843(8)
1.831(6)
1.877(7)
1.327(11)
1.382(11)
1.329(10)
1.375(12)
1.876(8)
1.842(8)
1.855(7)
1.848(6)
1.324(11)
1.307(13)
1.345(10)
1.355(11)
N1-Ni-N2
N1-Ni-O1
N2-Ni-O2
O1-Ni-O2
N1-Ni-O2
N2-Ni-O1
86.1(3)
94.9(3)
94.9(2)
84.2(2)
176.8(3)
178.0(3)
86.0(3)
94.0(2)
94.4(2)
85.9(2)
175.4(3)
176.0(3)
86.9(3)
95.2(3)
93.9(3)
84.0(3)
177.8(3)
177.3(3)
85.0(3)
94.8(3)
94.3(3)
86.9(3)
173.3(3)
171.0(3)
C3-C2-C17-C18
C3-C2-C17-C26
C1-C2-C17-C18
C1-C2-C17-C26
C14-C15-C27-C28
C14-C15-C27-C36
C16-C15-C27-C28
C16-C15-C27-C3
-117.09(0.76)
61.98(0.93)
59.63(0.94)
-121.37(0.77)
-116.15(0.73)
59.44(0.85)
-128.09(0.62)
53.42(0.75)
54.33(0.85)
-124.16(0.60)
-73.99(0.70)
106.83(0.63)
112.99(0.66)
-66.19(0.74)
-119.1(1.0)
64.1(1.2)
53.0(1.2)
-123.9(1.0)
-120.8(0.9)
55.2(1.1)
-119.02(1.09)
59.9(1.3)
60.7(1.3)
-120.4(1.0)
-76.7(1.1)
97.6(1.0)
103.2(1.0)
-82.5(1.1)
57.69(0.87)
-126.72(0.66)
56.3(1.1)
-127.7(0.9)
F IGURE 7. Nonlinear effects study with matched Ni‚Cs-
BINOL-salen 25d and its enantiomer ent-25d for the reaction
in eq 1 (THF, 5 d, -20 °C, 10 mol % catalyst).
thoate oxygens are closer to the bound electrophile
(
Figure 6). Since the proposed mechanism involves
coordination of the nucleophile to the deprotonated
naphthoate oxygen, such an arrangement may lead to
higher selectivity. This hypothesis is also consistent with
the lower selectivity of the catalyst derived from 21 (54%
ee) vs 22 (71% ee). In 21, the (S,S)-cyclohexanediamine
would be able to better accommodate the helical twist
caused by the BINOL portions and the type of change
observed from 19 to 22 (shift of the diamine cant) would
not occur.
Rea ction Mech a n ism . To probe the reaction mech-
anism, several studies were undertaken. In the first,
nonlinear effects were measured (Figure 7). A positive
nonlinear effect was observed; however, the formation of
a visible precipitate with the lower enantiomeric excess
catalysts along with their much lower reactivity indicates
the formation of nonreactive aggregates. These results
are consistent with a reservoir effect14 rather than the
intervention of two catalyst molecules in the enantio-
selectivity-determining step.
F IGURE 6. Overlay of 19a and 22a (superposition of Ni, N1,
N2, O1, O2). Half-bond structure ) ethylenediamine Ni
complex 19a ; purple structure ) (R,R)-cyclohexanediamine Ni
complex 22a . (a) Top view (b) Edge view.
Thus, the (R,R)-cyclohexanediamine counters the twist
imposed by the BINOL portions. The net result is that
the dihedral angle between the core salen plane (Ni-
N1-N2-O1-O2) and the ancillary naphthols is more
perpendicular for the (R,R)-cyclohexanediamine deriva-
tive [70.1(1), 57.8(1)°] compared to the ethylenediamine
derivative [45.4(1), 40.7(2)°]. This angle governs the
position of the naphthoxide Brønsted base with respect
to an electrophile bound in the salen apical position1
and is correlated to the higher selectivity of 22 compared
to 19 (Table 2). If an electrophile binds perpendicular to
the core salen plane, then the more perpendicular naph-
2,13
1
978 J . Org. Chem., Vol. 68, No. 5, 2003

Catalysis of the Michael Addition Reaction
F IGURE 8. Kinetic analysis of the catalyst order for salen
naphthoxide catalyst 25d in the Michael addition reaction.
SCHEME 3. P ossible Ca ta lytic Cycle for th e
Ca ta lyzed Mich a el Rea ction
F IGURE 9. Additional catalysts used in the Michael reaction
of benzyl malonate and cyclohexenone (eq 1) to probe the
mechanism (48 h reaction time unless otherwise noted).
the addition, (3) the Lewis acid of the salen complex acts
in concert with a base arising from a separate source (not
an intramolecular base), or (4) the Brønsted base of the
salen complex acts in concert with an acid arising from
a separate source (not an intramolecular Lewis acid). To
determine which of these pathways is most likely, several
additional experiments were performed (Figure 9).
If the active portion of the catalyst was simply the
salen Lewis acid, then the protonated form of the complex
2
5a should give rise to similar enantioselectivity and
reactivity. When this form (19) was employed, no reaction
was observed (Figure 9). Further evidence for the neces-
sity of a Brønsted base can be found in the lack of
reactivity observed upon replacement of both phenolic
protons with a divalent metal such as Zn, Mg, or Ca
leading to general structure 32 (Figure 9). Presumably
a stable tetrahedral complex forms between the four
BINOL oxygens, rendering the naphthoates ineffective
as bases. These results indicate that a component is
needed in addition to the Lewis acid or that the Lewis
acid of the salen does not play any role. To rule out the
latter possibility, a reaction was performed employing the
Pd‚Cs-BINOL-salen catalyst 30. We anticipated that the
Pd(II) derivative, which is less Lewis acidic than the
isosteric Ni(II) derivative, would give rise to lower
reactivity/selectivity if the salen metal acted as a Lewis
acid. Since reaction with the Pd derivative 30 was slower
in a side-by-side trial and the selectivity was lower
compared with that of Ni‚Cs-BINOL-salen 25a , we
conclude that the salen metal does play a role as a Lewis
acid. Thus, it appears that the salen Lewis acid is a
critical component and that it acts in concert with a
second moiety.
The reaction mechanism for the Michael addition of
dibenzyl malonate to cyclohexanone with 25d was further
studied by measuring pseudo-first-order rates over a
range of catalyst concentrations (Figure 8). A plot of rate
vs [catalyst] indicates that the reaction is first order in
2
5d . On this basis, it appears that only one molecule of
the salen catalyst is involved in the rate-determining
step.
Using this information a potential catalytic cycle for
the Michael reaction is depicted in Scheme 3. Deproto-
nation of the dibenzyl malonate by the naphthoate anion
would yield a malonate anion held in proximity to the
salen via coordination through the naphthoate counterion
(i.e., Cs). Coordination of the cyclohexenone to an apical
site of the Ni-salen complex would activate the electro-
phile in close proximity to this nucleophile allowing the
Michael addition to occur. Proton transfer from the
naphthol to the enolate intermediate would then allow
product release and catalyst regeneration.
However, several other mechanisms are also possible,
including the following: (1) the Lewis acid of the salen
complex acts alone to catalyze the addition, (2) the
Brønsted base of the salen complex acts alone to catalyze
Further experiments were directed at determining if
the Brønsted base in the Ni‚Cs-BINOL-salen complexes
2
5a -d was solely responsible for the catalytic activity.
(
14) For reviews on nonlinear effects, see: (a) Blackmond, D. G. J .
The high rates obtained with Cs-naphthoxide 33 indicate
that a reaction in which the naphthoxide portion of the
catalyst acts alone as a Brønsted base is feasible.
However, a BINOL-derived chiral Cs-naphthoxide base,
Am. Chem. Soc. 1997, 119, 12934-12939. (b) Guillaneux, D.; Zhao,
S.; Samuel, O.; Rainford, D.; Kagan, H. B. J . Am. Chem. Soc. 1994,
16, 9430-9439. (c) Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed.
998, 37, 2922-2959.
1
1
J . Org. Chem, Vol. 68, No. 5, 2003 1979

Annamalai et al.
such as 34, alone is not a sufficient determinant for good
enantioselectivity. Rather, the entire Ni‚Cs-BINOL-salen
assembly is required for selectivity in line with a bifunc-
tional catalyst comprised of a Lewis acid salen metal
acting in concert with a naphthoxide Brønsted base.
The third possibility, that the salen metal acts as a
Lewis acid and a separate Brønsted base such as Cs
CO or CsHCO
(see below)15 is involved, was tested by
employing the methylated Ni-BINOL-salen 31 in the
presence of Cs CO . This complex displays reduced
2
-
3
3
2
3
reactivity compared to 25a -d and the absence of any
selectivity points away from the salen Lewis acid acting
in concert with a separate Brønsted base.
The fourth possibility, that the salen complex acts as
a Brønsted base while a separate species acts as a Lewis
acid, seems unlikely in light of the correlation between
the Lewis acidity of the salen metal and the selectivity
(
see above). In addition, the only other species present
that could act as a Lewis acid is the Cs counterion. To
rule out the direct involvement of the Cs CO or CsHCO
present in the in situ catalyst preparations obtained by
treatment of 10 mol % 22 with 20 mol % Cs CO (see
F IGURE 10. Experiments for studying competing pathways
in the Michael addition reaction (eq 1, room temperature).
2
3
3
2
3
below), we undertook a preparation of Ni‚Cs-BINOL-
salen 25d in pure form (eq 3). A solution containing 10
mol % Ni-BINOL-salen 22 was treated with 10 mol %
Cs
formation of Ni‚Cs-BINOL-salen 25d by evolution of H
CO as H O and CO . When 25d prepared in this manner
was employed in the reaction of dibenzyl malonate with
cyclohexenone, the results were very similar (58% conv,
2 3
CO and was then heated to reflux to promote
2
-
3
2
2
F IGURE 11. Non-Arrhenius behavior of the salen naphthox-
ide catalyst 25d in the Michael addition reaction (eq 1).
7
2
2% ee) relative to the form of 25d prepared in situ from
0 mol % Cs CO (79% conv, 71% ee). Thus, the involve-
CO or CsHCO in
more basic than bicarbonate,16 8 mol % of Ni‚Cs-BINOL-
2
3
1
7
salen 25d , 1 mol % of the monodeprotonated 35, 1 mol
ment of a Cs Lewis acid from the Cs
2
3
3
%
%
of Ni-BINOL-salen 22, 2 mol % of Cs
2 3
CO , and 18 mol
the enantioselective reaction pathway is unlikely. Over-
all, both the Lewis acid and the Brønsted base were
determined to be essential for catalysis and the preceding
sets of experiments are all consistent with the bifunc-
tional mechanism outlined in Scheme 3.
of CsHCO are estimated to be present at the time the
3
reaction is initiated (Figure 10). Regardless of the precise
breakdown, there will be 20 mol % total of CsHCO
Cs CO present. On the basis of the results collected in
3
+
2
3
Figure 10, both of these species alone are very effective
catalysts for the Michael reaction and give rise to
nonselective processes. Paradoxically, the rate of reaction
with the Ni‚Cs-BINOL-salen catalyst 25d is slower than
these nonselective reactions, yet the enantioselectivity
Ba ck gr ou n d Rea ction s a n d Syn er gistic Effects.
Additional experiments were conducted to gain insight
into potential competitive catalytic pathways (Figure 10).
When 25d was prepared as described in eq 3 (no residual
Cs
was observed as when 25d was prepared in situ with 2
equiv of Cs CO (Figure 10). From this result, it is likely
that the same chiral catalyst is responsible for the
enantioselective reaction in each preparation. However,
other species (CsHCO
present in the in situ preparation of 25d by treatment
of 22 with 2 equiv of Cs CO (eq 3 vs Figure 10).
Assuming that naphthol is approximately one pK unit
2 3 3
CO and CsHCO ), the same level of enantioselection
(
16) The distribution of species indicated in Figure 10 upon the
treatment of 22 with Cs CO is roughly estimated from the relative
pK values of CsHCO and naphthol (∆pK
∼ 0.5-2). While the
absolute pK values
2
3
2
3
a
3
a
a
values change substanially with solvent, the ∆pK
a
of these delocalized oxygen acids are about the same in water, DMSO,
3 2 3
and/or Cs CO ) will also be
and THF (see: Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463
and references therein). If ∆pK ) 1, the values in Figure 10 are
a
obtained. This analysis is not quantitative, but clearly indicates that
a small amount of Cs CO (<25d ) and a large amount of CsHCO
(>25d ) would be present in the catalyst mix.
17) Since the nondeprotonated adduct (22) is not a catalyst, it seems
2
3
2
3
3
a
(
reasonable that only the half of 35 containing the naphthoate portion
can act as a catalyst.
(15) Present from preparation of the catalyst.
1
980 J . Org. Chem., Vol. 68, No. 5, 2003

Catalysis of the Michael Addition Reaction
is good with 25d from the in situ preparation. In other
words, the nonselective background reactions are sup-
pressed in the presence of 25d , but it is not clear how
this suppression occurs. On the basis of the non-Arrhe-
nius behavior (Figure 11) of 22d in the Michael reaction
of dibenzyl malonate and cyclohexenone, these nonselec-
tive background reactions appear to account for a greater
portion of the reaction at higher temperatures.
chiral diamine linker and the chiral biaryl, with the latter
playing a larger role. The activity of nonselective com-
petitive catalyst species appears to be attenuated in the
presence of these chiral catalysts. While the mechanism
of this attenuation is unclear, the implication is that the
rates of catalyzed reactions do not necessarily need to
be more rapid than those of the background reactions
(i.e., in the absence of catalyst). This family of compounds
has promise in reactions requiring Lewis acid/Brønsted
base catalysts.
Con clu d in g Rem a r k s
In conclusion, several novel salen complexes have been
synthesized and characterized. These complexes are
effective in catalyzing the Michael reaction between
various enones and malonates. The preponderance of
evidence from mechanistic studies of the Ni‚Cs-BINOL-
salen catalysts supports a bifunctional activation mech-
anism in the Michael reaction. There is sufficient spatial
and electronic separation between the Lewis acid and
Brønsted basic sites such that these catalyst elements
can be independently altered. The major elements influ-
encing stereoselection in this family of catalysts are the
Ack n ow led gm en t. Financial support was provided
by the National Institutes of Health (GM-59945) and
Merck Research Laboratories. E.F.D. thanks the Uni-
versity of Pennsylvania for an SAS Dissertation Fel-
lowship.
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails, characterization of all new compounds, and X-ray data
for 19. This material is available free of charge via the Internet
at http://pubs.acs.org.
J O025993T
J . Org. Chem, Vol. 68, No. 5, 2003 1981