The synthesis of N,O-ferrocenyl pyrrolidine-containing ligands and their application in the diethyl- and diphenylzinc addition to aromatic aldehydes

Article abstract of DOI:10.1021/jo060894r

(Chemical Equation Presented) A facile route to a series of planar chiral N,O-ferrocenyl pyrrolidine-containing ligands with varying substituents at the nitrogen and oxygen donor atoms is described. The oxygen donor atom was introduced via a diastereoselective ortho-metalation of N-methylpyrrolidinyl and N-allylpyrrolidinyl ferrocene intermediates and was quenched with various ketones. The nitrogen substituent was varied through deallylation and subsequent derivatization of a secondary pyrrolidine. The efficacy of these novel ligands was investigated in the enantioselective addition of diethylzinc and diphenylzinc to aromatic aldehydes. The ligands proved highly effective in the diethylzinc addition to benzaldehyde that resulted in high yields of up to 99% and enantioselectivities (ee's) of up to 95%. The role of planar chirality was explored and the results indicated that the planar chirality, and not the central chirality, of the ferrocenyl ligands was the dominant stereo-controlling element. Employment of a mixed ethyl-phenylzinc reagent in the phenylation of aromatic aldehydes led to a mixture of the two additional products, and the phenylated product was obtained in up to 37% ee.

Full text of DOI:10.1021/jo060894r

The Synthesis of N,O-Ferrocenyl Pyrrolidine-Containing Ligands
and Their Application in the Diethyl- and Diphenylzinc Addition to
Aromatic Aldehydes
Theresa Ahern, Helge M u¨ ller-Bunz, and Patrick J. Guiry*
Centre for Synthesis and Chemical Biology, Conway Institute of Biomolecular and Biomedical Research,
School of Chemistry and Chemical Biology, UniVersity College Dublin, Belfield, Dublin 4, Ireland
p.guiry@ucd.ie
ReceiVed April 28, 2006
A facile route to a series of planar chiral N,O-ferrocenyl pyrrolidine-containing ligands with varying
substituents at the nitrogen and oxygen donor atoms is described. The oxygen donor atom was introduced
via a diastereoselective ortho-metalation of N-methylpyrrolidinyl and N-allylpyrrolidinyl ferrocene
intermediates and was quenched with various ketones. The nitrogen substituent was varied through
deallylation and subsequent derivatization of a secondary pyrrolidine. The efficacy of these novel ligands
was investigated in the enantioselective addition of diethylzinc and diphenylzinc to aromatic aldehydes.
The ligands proved highly effective in the diethylzinc addition to benzaldehyde that resulted in high
yields of up to 99% and enantioselectivities (ee’s) of up to 95%. The role of planar chirality was explored
and the results indicated that the planar chirality, and not the central chirality, of the ferrocenyl ligands
was the dominant stereo-controlling element. Employment of a mixed ethyl-phenylzinc reagent in the
phenylation of aromatic aldehydes led to a mixture of the two additional products, and the phenylated
product was obtained in up to 37% ee.
Introduction
However, the ligand-specific nature of catalytic reactions ensures
that the search for new ligands remains a field of incessant
interest. In comparison to the work on dialkylzinc additions,
the asymmetric addition of diarylzinc reagents to aldehydes is
substantially less developed. Nevertheless, there has been
considerable focus on this area in recent years because chiral
diarylmethanols are important precursors for pharmacologically
The enantioselective formation of C-C bonds is regarded
as one of the most fundamental transformations in asymmetric
1
synthesis. The enantioselective preparation of chiral alcohols
is of particular interest, because they represent valuable structural
units found in many natural products and they are readily
functionalized. The addition of organometallic reagents to
carbonyl compounds has emerged as a key method of preparing
chiral alcohols. Since the pioneering work of Oguni in 1984,
there has been extensive research into the diethylzinc addition
to benzaldehyde, and it is now regarded as one of the benchmark
reactions for understanding the catalytic potential of new ligands.
2
4
,5
and biologically active compounds. To date, several N,O-
6
ferrocenyl ligands, such as the ligands from Schl o¨ gl (1),
3
7
8,9
Butsugan and Watanabe (2), and Bolm (3), have been
successfully applied in the diethyl- and diphenylzinc addition
to aromatic aldehydes.
A literature survey of this test reaction prompted us to prepare
a series of N,O-ferrocenyl ligands in which the substituents at
*
Tel: +353-1-716-2309; Fax +353-1-716-2501.
(1) (a) Ojima, I. Catalytic Asymmetric Synthesis, 2nd ed; VCH: Wein-
heim, Germany, 2000. (b) Noyori, R. Asymmetric Catalysis in Organic
Synthesis; Wiley: New York, 1994. (c) Beller, M.; Bolm, C. Transition
Metals for Organic Synthesis; Wiley-VCH: Weinheim, Germany, 1998.
(4) Meguro, K.; Aizawa, M.; Sohda, T.; Kawamatsu, Y.; Nagaoka, A.
Chem. Pharm. Bull. 1985, 33, 3787.
(5) Toda, F.; Tanaka, K.; Koshiro, K. Tetrahedron: Asymmetry 1991,
2, 873.
(6) Wally, H.; Widhalm, M.; Weissensteiner, W.; Schl o¨ gl, K. Tetrahe-
dron: Asymmetry 1993, 4, 285-288.
(
2) Mukaiyama, T.; Soai, K.; Sato, T.; Shimizu, H.; Suzuki, K. J. Am.
Chem. Soc. 1979, 101, 1455.
3) Oguni, N.; Omi, T. Tetrahedron Lett. 1984, 25, 2823-2824.
(
10.1021/jo060894r CCC: $33.50 © 2006 American Chemical Society
7
596
J. Org. Chem. 2006, 71, 7596-7602
Published on Web 08/25/2006

N,O-Ligands Possessing Central and Planar Chirality
SCHEME 1
the nitrogen and oxygen donor atoms could be readily manipu-
lated. The introduction of the oxygen donor atom, and conse-
quently planar chirality, employed an ortho-directing lithiation
strategy that was developed by Ugi using N,N-dimethyl-1-
ferrocenylethylamine (4).10 Since this pioneering work, a range
SCHEME 2. Synthesis of N-Pyrrolidine 7a
11
12
17
of ortho-directing auxiliaries, including sulfoxides, acetals,
oxazolines,9
,13,14
azepines, sulfoximines, and hydrazones,
15
16
have been employed. Of greater interest to our research was
that ferrocenylpyrrolidine (5) developed by Ganter18 and trans-
19
2,5-disubstituted pyrrolidine (6) from our research laboratories
effected diastereoselective lithiations.
a(a) Ac2O, NEt3, DMAP, rt, 12 h; (b) MeNH2 (2 M solution MeOH) or
C3H5NH2 in MeOH, ∆, 3 h; (c) (i) hot EtOH, (+)-tartaric Acid, (ii) NaOH.
innovative move, the substituents at the nitrogen donor atom
were varied by employing a key N-allylpyrrolidine intermediate
We recently reported the synthesis of N-methylpyrrolidinyl-
ferrocene (7a), which proved to be an efficient ortho-directing
group in the synthesis of the P,N-ferrocenyl ligand (8) (Scheme
7
d and an allylation/deallylation approach. The series of N,O-
ferrocenyl ligands was applied in the diethyl- and diphenylzinc
addition to aromatic aldehydes to determine the effect of
substituents at the donor atoms and the role of planar chirality
in inducing asymmetry in these transformations.
2
0
1
). Herein, we describe the synthetic procedures devised for
novel N,O-ferrocenyl ligands of type 9. An ortho-directed
metalation of 7a and quench with different ketones generated
novel N,O-ferrocenyl ligands with varying degrees of steric bulk
at the hydroxyl-bearing carbon atom of ligand 9. In a previous
attempt to vary the substituents at the nitrogen donor atom, the
synthesis and lithiation of N-pyrrolidinyl analogues 7b and 7c
was investigated. It immediately was apparent that the drawback
to this approach was the erratic nature of the lithiation process.
Although the coordination of the lithiating reagent is strongly
dependent on the N-pyrrolidinyl substituent, each set of lithiation
conditions had to be optimized independently. The optimal
lithiation of pyrrolidine 7b proceeded in 15% yield and 40%
diastereomeric excess (de), whereas the attempted lithiation of
pyrrolidine 7c did not afford any product.20 Thus, in an
Results and Discussion
Synthesis of Pyrrolidinylferrocene. The strategy for the
preparation of ferrocenyl ligands of type 9 involved the synthesis
of pyrrolidine 7 followed by a directed ortho-metalation to
introduce the tertiary alcohol function. The asymmetric synthesis
of pyrrolidine 7, which utilizes the Corey-Bakshi-Shibata
2
1
(CBS) reduction protocol, has been reported in a previous
2
0
communication. The (R)-enantiomer was tentatively assigned
to alcohol 10 based on the known stereochemical outcome of
the CBS reduction of other ferrocenyl ketones and later
confirmed by X-ray crystallography of a compound derived from
alcohol 10. The enantiomeric purity of alcohol (R)-10 was
determined as 93% enantiomeric excess (ee) from chiral HPLC.
Subsequent acetylation of alcohol 10 gave acetate 11, and
treatment with methylamine or allylamine furnished enantio-
enriched N-methylpyrrolidine 7a or N-allylpyrrolidine 7d,
respectively (Scheme 2). It is worth noting that the preparation
of enantioenriched pyrrolidines 7a and 7d proceeded with the
retention of configuration and without the loss of ee, a
(
7) Wantanabe, M.; Hashimoto, N.; Araki, S.; Butsugan, Y. J. Org. Chem.
992, 57, 742-744.
8) Bolm, C.; Hermanns, N.; Hildebrand, J. P.; Mu n˜ iz, K. Angew. Chem.,
Int. Ed. 2000, 39, 3465.
9) Bolm, C.; Mu n˜ iz-Fern a´ ndez, K.; Seger, A.; Raabe, G.; G u¨ nther, K.
J. Org. Chem. 1998, 63, 7860-7867.
10) Marquarding, D.; Klusacek, H.; Gokel, G.; Hoffmann, P.; Ugi, I. J.
Am. Chem. Soc. 1970, 92, 5389.
11) Rebiere, F.; Riant, O.; Ricard, L.; Kagan, H. B. Angew. Chem., Int.
Ed. Engl. 1993, 32, 568.
12) Kagan, H. B.; Samuel, O.; Riant, O. J. Am. Chem. Soc. 1993, 115,
835.
1
(
(
(
(
22
phenomenon previously observed by Ugi. Enantiopure 7a was
obtained from its enantioenriched mixture in 81% yield through
the preparation of diastereomeric salts from tartaric acid in
(
5
(
13) Sammakia, T.; Latham, H. A.; Schaad, D. R. J. Org. Chem. 1995,
1
0
6
0, 10.
ethanol. Although several resolving agents were screened for
the resolution of 7d (e.g., tartaric acid, ditolyl tartaric acid,
dibenzoyl tartaric acid, and mandelic acid, among others) in
various solvent systems, its fractional recrystallization was
unsuccessful and, thus, enantioenriched 7d was used for all
further synthetic steps. It was noted that previous attempts at
(
14) Richards, C. J.; Damalidis, T.; Hibbs, D. E.; Hursthouse, M. B.
Synlett 1995, 74.
15) Widhalm, M.; Mereiter, K.; Bourghida, M. Tetrahedron: Asymmetry
998, 9, 2983.
16) Bolm, C.; Kesselgruber, M.; Muniz, K.; Raabe, G. Organometallics
000, 38, 1648.
17) Enders, D.; Peters, R.; Lochtman, R.; Raabe, G. Angew. Chem.,
Int. Ed. 1999, 38, 2421.
(
1
2
(
(
(
(
(
18) Ganter, C.; Wagner, T. Chem. Ber. 1995, 128, 1157.
19) Farrell, A.; Goddard, R.; Guiry, P. J. J. Org. Chem. 2002, 67, 4209.
20) Meaney, K., Ph.D. Thesis, National University of Ireland, 2002.
(21) Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Am. Chem. Soc. 1987,
109, 7925.
(22) Gokel, G.; Marquarding, D.; Ugi, I. J. Org. Chem. 1972, 37, 3052.
J. Org. Chem, Vol. 71, No. 20, 2006 7597

Ahern et al.
SCHEME 3. Synthesis of N,O-Ferrocenyl Ligands 12-15a
propose that the lower diastereoselectivity observed for 7d is a
result of inferior differentiation in the lithiation transition states
as compared to those for 7a. In the absence of an X-ray crystal
structure of the major diastereomer, (R)-planar chirality was
inferred from the asymmetric lithiation in addition to trends in
optical rotations within this series. Purification by column
chromatography and subsequent recrystallization from pentane
afforded (R,R)-15 with absolute diastereomeric and enantiomeric
purity as determined by chiral HPLC. This was advantageous
to our synthetic scheme as it illustrated that in the case of 15 it
was possible to obtain optically pure material at this late stage.
Selective Preparation of Minor Diasteromeric Ligands.
The selective preparation of the minor diastereomers involved
a
(a) s-BuLi, Et2O, -78 °C for 3 h, 0 °C for 1 h, then (R′)2CO, 0 °C to
rt, 2 h.
temporarily protecting the preferred ortho-position of 7a and
TABLE 1. Synthesis of N,O-Ferrocenyl Ligands 12-15a
25
7
d with a trimethylsilyl (TMS) group (Scheme 4). The
yield (R,R)
yield (R,S)
(%)
de
diastereoselectivity of the TMS-derivative 16a and 16b were
determined from the H NMR as 73 and 70% de, respectively.
b
b
(%)c
entry
ligand
(%)
1
1
2
3
4
12
13
14
15
38
73
55
45
0
12
8
>99
73
76
Prior to subsequent transformations, the mixtures of diastere-
omers were separated by column chromatography on silica gel
with the major diastereomers (R,R)-16a being isolated in 60%
yield and (R,R)-16b being isolated in 55% yield. Subsequent
deprotonation at the remaining ortho-position of 16a and 16b
and quenching with benzophenone or bis(3,5-dimethyl-phenyl)-
methanone yielded trisubstituted ferrocenyl ligands 17, 18, and
22
33
a
Lithiation conditions of - 78 °C for 3 h followed by 0 °C for 1 h then
quenched with ketone. Isolated yield of diastereomer after column
chromatography. Determined by H NMR.
b
c
1
1
9. Treatment of the TMS-containing derivatives with tetrabu-
tylammonium fluoride (TBAF) furnished the (R,S)-diasteromers
3, 14, and 15 in high yields. The predicted (R)-central chirality
the resolution of enantioenriched pyrrolidines 7b and 7c also
failed. It was suggested that the steric bulk on the nitrogen atom
1
23
may inhibit salt formation in the resolution of tertiary amines.
and (S)-planar chirality of ligands 13 and 14 were confirmed
Using a directed ortho-lithiation and quench with the ap-
propriate electrophile, it was possible to generate planar chiral
ferrocene derivatives with different degrees of steric bulk in
the immediate vicinity of the hydroxyl group. In general, the
lithiation procedure can be quite capricious, because its success
is highly dependent on the reaction conditions.2 After a
preliminary optimization of the lithiation conditions, 7a was
treated with s-BuLi (1.2 equiv) in Et2O at -78 °C for 3 h
followed by 1 h at 0 °C to ensure complete lithiation before
the addition of the electrophile. Using this strategy, the lithiated
intermediate of 7a was quenched with three symmetrical
ketones, namely, acetone, benzophenone, and bis-(3,5-dimethyl-
phenyl)methanone, to furnish 12, 13, and 14, respectively
by X-ray crystallographic analysis (Figure 1). The chirality of
(R,S)-15 was similarly inferred. Selected bond lengths, angles,
and torsion angles are presented in Table 2.
It was noted that the precursor to the minor diastereomers,
namely, trisubstituted ferrocene derivatives 17, 18, and 19, are
also potential ligands for asymmetric catalysis. Previous reports
suggest that the TMS group can influence the conformation of
4
the other substituents on the ferrocene ring leading to augmented
20,26
enantiocontrol of the reaction.
We decided to investigate
this potential TMS effect by comparing the catalytic results of
R,R)-14 and (R,R)-17. To fully ascertain the effect of the TMS
(
group, it was necessary to prepare its diastereomer (R,S)-17 for
comparative purposes with (R,S)-14. One approach taken to
prepare (R,S)-17 involved protecting the hydroxyl group of
(
Scheme 3). Similarly, the lithiated intermediate of 7d was
quenched with benzophenone to generate N,O-ferrocenyl ligand
5.
For the acetone-derived ligand 12, the lithiation of 7a was
(R,R)-14 and then lithiating at the ortho-position and quenching
1
with trimethylsilyl chloride. Unfortunately, several attempts to
27
temporarily protect the hydroxyl group of (R,R)-14 failed.
completely diastereoselective (Table 1) with the sole diastere-
omer being isolated in 38% yield. The lithiation of 7a and
quenching with benzophenone and bis-(3,5-dimethyl-phenyl)-
methanone furnished ligands 13 and 14, respectively. By
However, the required ligand (R,S)-17 was prepared in 63%
yield through lithiation of (R,S)-16a (the minor diastereomer
in the preparation of TMS-containing derivative 16a) and quench
with benzophenone (Scheme 5).
As outlined above, the variation of the N-pyrrolidine sub-
stituent prior to the lithiation step was problematic in its
resolution and in the individual optimization of the lithiation
conditions. We investigated an alternative route in which
pyrrolidine (R,R)-15 was employed to access a series of closely
related ligands with varying N-substituents (Scheme 6). The
deallylation procedure involved the treatment of 15 with
palladium(tetrakistriphenylphosphine) and N,N-dimethyl barbi-
1
comparing the N-Me signals in the crude H NMR spectrum,
the diastereomeric ratios were determined as 73 and 76% de,
respectively. The mixures of the diastereomers were separated
by column chromatography. Crystals grown from n-hexane of
the major diastereomers of 12 and 13 were suitable for X-ray
diffraction, and analysis confirmed the (R)-central chirality of
the pyrrolidine and demonstrated the (R)-planar chirality of the
ferrocene backbone (Figure 1). Similarly, the lithiation inter-
mediate of 7d was quenched with benzophenone to furnish 15
1
in 33% de as determined from the H NMR spectrum. We
(25) Richards, C. J.; Mulvaney, A. W. Tetrahedron: Asymmetry 1996,
7
, 1419.
(23) Vries, T.; Wynberg, H.; van Echten, E.; Koek, J.; ten Hoeve, W.;
(26) Patti, A.; Nicolosi, G.; Howell, J. A. S.; Humphries, K. Tetrahe-
Kellogg, R. M.; Broxterman, Q. B.; Minnaard, A.; Kaptein, B.; van der
Sluis, S.; Hulshof, L.; Kooistra, J. Angew. Chem., Int. Ed. 1998, 37, 2349.
dron: Asymmetry 1998, 9, 4381-4394.
(27) (a) s-BuLi, Et2O, TMSCl, -78 °C.; (b) Imidazole, DMF, TMSCl,
50 °C.; (c) HMDS, TMSCl, n-hexane, 50 °C.; (d) DDQ, MeOH, 40 °C.
(24) See Supporting Information.
7598 J. Org. Chem., Vol. 71, No. 20, 2006

N,O-Ligands Possessing Central and Planar Chirality
2
FIGURE 1. X-ray crystal structure of (R,R)-12, (R,R)-13, (R,S)-13, (R,R)-14 (H atoms have been omitted for clarity except the C -pyrrolidine and
hydroxyl proton). Displacement ellipsoids are shown at 50% probability.
SCHEME 4. Synthesis of Minor Diastereomeric Ligands (R,S)-13-15a
a(a) s-BuLi, ether, -78 °C for 3 h then 0 °C for 1.5 h, then TMSCl, 1 h; (b) s-BuLi, ether, -78 °C for 3 h then 0 °C for 1 h, then (R′)2CO, THF, -78
C to rt, 2 h; (c) TBAF (1 M soln THF), ∆, 2 d.
°
SCHEME 5. Synthesis of Ferrocene (R,S)-17a
Diethylzinc Addition to Aromatic Aldehydes. A series of
experiments were conducted to determine the efficacy of N,O-
ferrocenyl ligands 12, 13, 14, 15, 17, 20, and 21 in the
diethylzinc addition to aromatic aldehydes. The reactions were
typically carried out in toluene on a 1 mmol scale of aldehyde
substrate employing 5 or 10 mol % ligand (Scheme 7). After a
general optimization of the reaction conditions, our investigation
focused on three main inquiries: first, the role of planar chirality
in determining the ee and stereochemical outcome of the
alkylation process; second, the influence of steric bulk directly
on the hydroxyl-bearing carbon atom on the ee; and third, the
steric and electronic influence of varying the N-pyrrolidinyl
substituent.
a(a) s-BuLi, Et2O, -78 °C for 3 h, 0 °C for 1 h, then Ph2CO, THF, -78
C-rt overnight.
°
turic acid (NDMBA) to furnish pyrrolidine (R,R)-20 in 87%
yield.28 The secondary pyrrolidine (R,R)-20 was derivatized in
the presence of potassium carbonate and benzylbromide to
furnish N-benzylpyrrolidine (R,R)-21 in 76% yield. The minor
diastereomer (R,S)-21 was prepared in a similar procedure.
The optimal reaction conditions with regard to temperature,
catalyst loading, and solvent were determined using ligands
(R,R)-13 and (R,S)-13 and benzaldehyde as the substrate (Table
(28) Garro-Helion, F.; Merzouk, A.; Guib e´ , F. J. Org. Chem. 1993, 58,
6
109.
3). It was immediately apparent that the planar chirality plays
J. Org. Chem, Vol. 71, No. 20, 2006 7599

Ahern et al.
TABLE 2. Selected Bond Lengths, Angles, and Torsion Angles of (R,R)-12, (R,R)-13, (R,S)-13, and (R,R)-14
bond length (Å)
bond angle (°C)
C3-C4-N
torsion angle (°C)
C3-C2-C1-O C2-C3-C4-N
ligand
O-H
N-H
C2-C1-O
O-H-N
(R,R)-12
(R,R)-13
(R,S)-13
(R,S)-14
0.71
0.79
0.86
0.88
2.16
2.00
1.82
1.82
111.80
111.51
111.79
112.30
110.94
102.99
115.83
115.90
155
154
174
179
19.1
23.1
-46.7
-45.1
51.3
44.6
19.0
33.4
SCHEME 6. Synthesis N-Benzylpyrrolidine 21a
a
(a) NDMBA, Pd(PPh3)4, CH2Cl2, 30 °C, 5 h; (b) benzylbromide, K2CO3, THF, 24 h, rt.
SCHEME 7. Diethylzinc Addition to Benzaldehyde
TABLE 4. Diethylzinc Addition to Benzaldehyde Catalyzed by
N,O-Ferrocenyl Ligands 12-14a
entry
ligand
time (h)
yield (%)
ee (%)b
configc
1
2
3
4
5
6
7
(R,R)-12
(R,R)-13
(R,R)-14
(R,R)-17
(R,S)-13
(R,S)-14
(R,S)-17
36
36
48
48
36
48
48
97
87
80
94
98
95
54
67
84
82
95
92
75
67
R
R
R
R
S
TABLE 3. Diethylzinc Addition to Benzaldehyde Catalyzed by
R,R)-13 and (R,S)-13a
(
planar
chirality
ligand
(mol %)
temp
(°C)
yield
(%)
ee
(%)
S
R
b
configc
entry
1
2
3
(R)
(R)
(R)
(R)
(S)
(S)
(S)
(S)
5
10
5
5
5
10
5
5
-20
-20
0
-20
-20
-20
0
87
96
99
94
98
99
99
99
84
83
78
85
92
91
90
92
R
R
R
R
S
S
S
S
a
The reactions were carried out in toluene in the presence of 5 mol %
b
of ligand at -20 °C. Determined by HPLC analysis using a Chiracel OD
column. c Determined by comparison with literature data (optical rotation
d
4
and HPLC retention times).
5
6
7
donor atom and the role of planar chirality, the series of
ferrocenyl ligands was applied in the diethylzinc addition to
benzaldehyde under standard conditions (Table 4).
The appropriate steric bulk of ligands can be difficult to
balance because excessively large substituents may adopt
conformations in the transition state that lower the efficiency
of the reaction. The increase in steric bulk at the hydroxyl-
bearing carbon atom from a methyl group in (R,R)-12 to a
phenyl group in (R,R)-13 results in an increase in the ee (67 to
e
8
-20
a
The reactions were carried out in toluene unless otherwise stated.
b
c
Determined by HPLC analysis on Chiracel OD column. Determined by
comparison with literature data (optical rotation and HPLC retention times).
n-Hexane used as solvent. 1:2 ratio of n-hexane/toluene used as solvent.
d
e
a dominant role in determining the stereochemical outcome of
the catalytic reaction, (R)-planar chiral (R,R)-13 furnished the
R)-product, and (R,S)-13 furnished the (S)-product. For (R,R)-
3, an increase in temperature from -20 °C to 0 °C resulted in
an increased yield of 1-phenylpropanol product (from 87 to
9%) with a concomitant reduction in the ee (from 84 to 78%
ee) (Table 1, entries 1 and 3). In a parallel study, an increase in
temperature had a negligible effect when (R,S)-13 as the ligand
Table 1, entries 5 and 7) was used. Increasing the ligand loading
from 5 to 10 mol % (R,R)-13 increased the yield (87 to 96%)
(
1
8
4%) and a decrease in the yield of (R)-1-phenylpropanol (97
to 87%) (Table 4, entries 1 and 2). The diarylhydroxymethyl
moiety is often referred to as the “magic group” in catalyst
design and has been used with increasing frequency in recent
9
30
years. A further increase in steric bulk in the form of the 3,5-
dimethylphenyl analogue (R,R)-14 resulted in a slight decrease
in the ee (84 to 82%), and a longer reaction time of 48 h was
required (Table 4, entries 2 and 3). For the (S)-planar chiral
ligands, the increase in steric hindrance when changing from a
phenyl group (R,S)-13 to a 3,5-dimethylphenyl derivative (R,S)-
(
without affecting the ee (Table 1, entries 1 and 2). Higher yields
(
94 versus 87%) and comparable ee’s (84 and 85% ee) were
obtained in hexane as solvent instead of toluene (Table 1, entries
and 4). As per the effect of temperature and solvent, the effect
1
4 had a large negative effect on the ee (92 to 75%) (Table 4,
entries 4 and 5).
1
of catalyst loading was less pronounced using (R,S)-13 as chiral
ligand (Table 3, entries 5, 6, and 7). Ferrocenyl ligand (R,S)-13
was only slightly soluble in n-hexane and thus a toluene/
n-hexane mixture was used as its solvent (Table 3, entry 8).
As a result of these studies, the standard reaction conditions
of 5 mol % ligand in toluene at -20 °C were chosen to test the
efficacy of our other ligands. For amino alcohols, the substit-
uents on the hydroxyl-bearing carbon atom, along with those
on the nitrogen atom, significantly influence the outcome of
catalysis.2 To investigate the steric influence at the oxygen
The introduction of a TMS group to (R,R)-13 to generate
R,R)-17 had a pronounced effect in catalysis as the ee increased
(
from 84 to 95% (Table 4, entries 2 and 4). Alternatively, the
introduction of the TMS group to (R,S)-13 to generate (R,S)-17
had a detrimental effect on the ee. This suggests that the bulky
nature of the TMS group in ligand (R,S)-17 led to the pyrrolidine
ring taking up a conformation which led to poor differentiation
(
29) Soai, K.; Niwa, M. Chem. ReV. 1992, 92, 833.
9
(30) Braun, M. Angew. Chem., Int. Ed. Engl. 1996, 35, 519.
7600 J. Org. Chem., Vol. 71, No. 20, 2006

N,O-Ligands Possessing Central and Planar Chirality
TABLE 5. Diethylzinc Addition to Benzaldehyde Catalyzed by
Diastereomeric Ligands 13, 15, 20, and 21a
TABLE 6. Diethylzinc Addition to Various Aldehydes Catalyzed
by (R,R)-13 and (R,S)-13a
entry
ligand
R
yield (%)
ee (%)b
config.c
planar
entry chirality
yield
(%)
ee
(%)
b
configc
aldehyde
benzaldehyde
4-chlorobenzaldehyde
4-methoxybenzaldehyde
trans-cinnamaldehyde
1-naphthaldehyde
1
2
3
4
5
6
7
8
(R,R)-13
(R,R)-15
(R,R)-21
(R,R)-20
(R,S)-13
(R,S)-15
(R,S)-21
(R,S)-20
Me
allyl
Bn
H
Me
allyl
Bn
H
87
61
33
94
99
99
28
56
84
37
57
66
92
74
48
21
R
R
R
S
S
S
1
2
R
R
R
R
R
R
S
99
78
96
99
37
16
99
92
99
99
5
78
69
71
65
55
39
90
49
69
56
41
35
R
R
R
R
R
R
S
3
4
5
6
2-naphthaldehyde
benzaldehyde
S
R
7
8
S
4-chlorobenzaldehyde
4-methoxybenzaldehyded
trans-cinnamaldehyde
1-naphthaldehyde
S
a
9
S
S
The reactions were carried out in toluene for 48 h at -20 °C using 5
b
10
11
S
S
mol % ligand. Determined by HPLC analysis using a Chiracel OD column.
c
S
S
Determined by comparison with literature data (optical rotation and HPLC
retention times).
12
S
2-naphthaldehyde
18
S
a
The reactions were carried out in toluene for 48 h at 0 °C in the presence
b
of 5 mol % ligand. Determined by HPLC analysis using a Chiracel OD
of the enantiotopic faces of benzaldehyde. Of the oxygen donor
atoms investigated, the diphenylmethanol moiety provided the
optimal yields and ee’s in the diethylzinc addition to benzal-
dehyde. It has been documented that the presence of alkyl groups
on the nitrogen donor atom favors stable catalytic complexes
and promote very rapid reactions, whereas bulky alkyl and aryl
substituents may provide higher ee’s because of better dif-
column. c Determined by comparison with literature data (optical rotation
d
and HPLC retention times). n-Hexane used as solvent.
SCHEME 8. Diphenylzinc Addition to Aromatic Aldehydes
3
1
ferentiation of the enantiotopic faces of the aldehyde substrate.
(
R,S)-13 gave the products with (S)-configuration. This sub-
Soai demonstrated that substituents on the nitrogen atom of the
chiral catalyst affect the formation and stability of the transition
stantiates our hypothesis that the planar chirality plays a
dominant role in determining the configuration of the addition
product. Selected aliphatic aldehydes (heptanal and cyclohexa-
nal) also were tested as substrates but proved unreactive under
the testing conditions using (R,R)-13 and (R,S)-13.
Diphenylzinc Addition to Aromatic Aldehydes. The suc-
cessful application of our N,O-ferrocenyl ligands in the dieth-
ylzinc addition to aromatic aldehydes encouraged us to apply
them in the more challenging diaryl transfer to aldehydes.8
We were interested in determining if these N,O-ferrocenyl
ligands were as compatible with the diaryltransfer to aldehydes
using either diphenylzinc or a modified phenylzinc reagent as
the phenyl source. A series of experiments was conducted to
determine the efficacy of N,O-ferrocenyl ligands (R,R)-13 and
3
2
state. We had a range of N-substituted ligands in which to
study the importance of N-variation employing the chosen
conditions of 5 mol % ligand 13, 15, 20, and 21 in toluene at
-
20 °C for 48 h (Table 5). In general, it was noted that the
replacement of the N-methyl substituent on the pyrrolidine ring
with an allyl or a benzyl moiety led to a reduction in the ee
Table 5, entries 1, 2, and 3). This trend also was observed when
replacing the N-methyl group in (R,S)-8 with an N-allyl or
N-benzyl moiety (Table 5, entries 5, 6, and 7).
(
,36,37,38
More importantly, it was observed that the planar chirality
continued to play a dominant role in determining the configu-
ration of the 1-phenylpropanol product. In general, N-substituted
ferrocenyl ligands favored induction of the (R)-enantiomer. In
contrast, secondary pyrrolidine (R,R)-20 afforded the opposite
chiral induction under the same reaction conditions (Table 5,
entries 4 and 8). The change in the stereochemical outcome of
the diethylzinc addition product using N-substituted and N-H
(
R,S)-13 in the diaryltransfer to aromatic aldehydes. Two
protocols for the enantioselective phenyl transfer to aldehydes
were employed. Method A consisted of the addition of diphen-
ylzinc to aromatic aldehydes in the presence of 10 mol %
ferrocenyl ligand at -20 °C for 48 h in toluene. Method B
involved the in situ preparation of an ethyl-phenylzinc reagent
by stirring a 2:1 ratio of the diethylzinc/diphenylzinc solution
in toluene for 30 min. The ferrocenyl ligand and aldehyde were
added and the reaction was stirred at 10 °C for 48 h in toluene
amino alcohols was previously observed by Soai, Wang, and
others.3
2,33
The diminished catalytic potential of amino alcohols
bearing an N-H group was in agreement with the results
previously reported.34,35
With the conditions optimized for benzaldehyde, we explored
the scope of ligands (R,R)-13 and (R,S)-13 in the asymmetric
addition of diethylzinc to several substituted aromatic aldehydes
(
Scheme 8).
Ferrocenyl ligands (R,R)-13 and (R,S)-13 were employed in
the two methods described above (Table 7). Initially, ferrocenyl
ligand (R,R)-13 was investigated in the diphenylzinc addition
to anisaldehyde and 4-chlorobenzaldehyde using method A.
Unfortunately, the yields (27-48%) and ee’s (3-7%) were poor
even after 48 h (Table 7, entries 1 and 2). The low ee’s were
due in part to the rapid uncatalyzed background reaction, which
became more prominent over long reaction times. Similar to
the diethylzinc addition to aldehydes, the ferrocenyl ligand with
(
Table 6). High chemical yields and moderate to good ee’s were
observed. However, it was noted that any substitution on the
aromatic ring lowers the ee relative to the parent benzaldehyde.
More interesting to our investigation is that (R,R)-13 furnished
the secondary alcohol products with the (R)-configuration and
(31) Rocha Gonsalves, A. M. d. A.; Serre, M. E. S.; Murtinho, D.; Silva,
V. F.; Beja, A. M.; Paixao, J. A.; Silva, M. R.; Alte da Veiga, L. J. Mol.
Catal. A: Chemical 2003, 195, 1-9.
(R)-planar chirality furnished the (R)-addition product, and (S)-
(32) Kitamura, M.; Okada, S.; Suga, S.; Noyori, R. J. Am. Chem. Soc.
planar chirality furnished the (S)-product.
1
989, 111, 4028-4036.
(33) Yang, X.; Shen, J.; Da, C.; Wang, R.; Choi, M. C. K.; Yang, L.;
Wong, K.-Y. Tetrahedron: Asymmetry 1999, 10, 133-138.
(36) Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem., Int.
Ed. Engl. 1995, 34, 1159-1171.
(37) Bolm, C.; Rudolph, J. J. Am. Chem. Soc. 2002, 124, 14850-14851.
(38) Kim, J. G.; Walsh, P. J. Angew. Chem., Int. Ed. 2006, 45, 4175.
(
34) Itsuno, S.; Frechet, J. M. J. J. Org. Chem. 1987, 52, 4140-4142.
(
35) Bastin, S.; Agbossou-Niedercorn, F.; Brocard, J.; P e´ linski, L.
Tetrahedron: Asymmetry 2001, 12, 2399-2408.
J. Org. Chem, Vol. 71, No. 20, 2006 7601

Ahern et al.
TABLE 7. Phenyl Transfer to Aldehydes Catalyzed by (R,R)-13
quantities of the ethyl-addition product observed for our
sterically demanding ferrocenyl ligands. To rationalize these
results, the currently available computational methods need to
be extended to incorporate parameters for both phenyl transfer
and bulky ligands, such as those containing ferrocene moi-
a
and (R,S)-13
planar
entry chirality method
yield
ee
b
(%)c configd
aldehyde
anisaldehyde
4-chlorobenzaldehyde 48
(%)
1
2
3
4
5
6
R
R
R
R
S
A
A
B
B
B
B
27
3
7
R
R
R
R
S
39,40
eties.
anisaldehyde
anisaldehyde
anisaldehyde
anisaldehyde
49(3) 30(53)
37(9) 18(39)
79(12) 33(61)
36(3) 37(48)
e
Conclusion
e
S
S
We have reported a convenient synthesis of novel N,O-
ferrocenyl ligands that allows for a facile modification of the
substituents at the oxygen and nitrogen donor atoms. The oxygen
donor group was introduced via a distereoselective lithiation to
generate two diastereomers in up to 99% de. The selective
preparation of the minor diastereomer employed a trimethylsilyl
protecting group strategy. The N-pyrrolidinyl substituent was
varied accordingly using an allylation/deallylation strategy. We
have shown that these ferrocene compounds act as efficient
ligands in the diethylzinc addition to aromatic aldehydes. Several
ligand structure-activity relationships were investigated, pre-
dominantly, the role of planar chirality in determining the
stereochemical outcome of the catalytic reaction. It was found
that the planar chirality governed the stereochemical outcome
in catalysis. In general, (R)-planar chirality furnished the (R)-
product, and (S)-planar chirality furnished the (S)-product with
the exception of the secondary pyrrolidine, which gave the
product of opposite configuration. Further investigations on the
application of these ferrocenyl compounds as chiral ligands in
other catalytic reactions will be reported in due course from
these laboratories.
a
The reactions were carried out in toluene in the presence of 10 mol %
chiral ligand. Method A: ZnPh2, -20 °C for 48 h. Method B: 2:1 ratio of
b
ZnEt2/ZnPh2, - 20 °C for 48 h. Isolated yield of phenyl addition product,
c
yields in parentheses are for the ethyl-addition product. ee’s for phenyl
product as determined by HPLC using a Chiracel OD and AD column,
ee’s in parentheses are for the ethyl-addition product. Determined by
comparison with literature data (HPLC retention times). Reaction carried
out at 10 °C.
d
e
Subsequently, we focused on the asymmetric phenyl transfer
using a modified phenylzinc reagent that was prepared in situ
from a 2:1 mixture of diethyl- and diphenylzinc. It has been
documented that the use of this mixed zinc reagent improves
8
the ee of the catalytic process. We also observed this trend
using anisaldehyde and ferrocenyl ligand (R,R)-13; 3% ee was
obtained using method A and 18-30% ee using method B
compare entries 1 with 3 and 4 in Table 7). It has been recorded
that the yields for this mixed zinc reagent are lower than when
diphenylzinc alone is used. Surprisingly, in this instance, the
(
8
yields were higher than those achieved using Method A,
although a mixture of products was recovered. Ligand (R,S)-
1
3 furnished the highest yield with 79% phenyl addition product
and 33% ee (Table 7, entry 5). Reducing the temperature to
20 °C was detrimental to the yield (79-36%), although the
Acknowledgment. We thank Enterprise Ireland for the
award of a Research Scholarship and UCD for a Research
Demonstratorship to T.A. We are grateful to Eli Lilly Ireland
Ltd. (Kinsale), CSCB, and Cork County Council for financial
support. Dr. Raymond Bronger is kindly acknowledged for a
critical reading of this manuscript.
-
ee increased marginally (33-37%) (Table 7, entries 5 and 6).
Several research groups have employed the mixed ethyl-
phenylzinc reagent with complete selectivity for the phenyl
8
transfer. To the best of our knowledge, this is the first example
Supporting Information Available: X-ray crystallographic files
of the use of a mixed phenyl-ethylzinc reagent that results in a
mixture of the two addition products. However, Norrby postu-
lated that while selective phenyl transfer over ethyl transfer is
the prevailing reaction mechanism for the majority of ligands,
sterically congested ligands may favor the less demanding ethyl-
transfer route.39 This could possibly account for the small
1
13
for 12, 13, and 14, full spectral characterization ( H and C NMR
spectra) for all new compounds, HPLC characterization for fer-
rocenes 7d and 15, and experimental details concerning the diethyl-
and diphenylzinc additions to aromatic aldehydes. This material is
free of charge via the Internet at http://pubs.acs.org.
JO060894R
(
39) Rudolph, J.; Bolm, C.; Norrby, P.-O. J. Am. Chem. Soc. 2005, 127,
(40) Rasmussen, T.; Norrby, P.-O. J. Am. Chem. Soc. 2001, 123, 2464-
2465.
1
548-1552.
7602 J. Org. Chem., Vol. 71, No. 20, 2006