Synthesis of a 9-fluorenone derived β-amino alcohol ligand depicting high catalytic activity and pronounced non-linear stereochemical effects

Article abstract of DOI:10.1055/s-2000-6229

The Jacobsen epoxidation of 9-alkylidenefluorenes 4a (ethylidene), 4b (benzylidene) and 4c (1-naphtylmethylene) with standard (R,R)-manganese salen catalyst has been studied. Both conversion and enantioselectivity depend on the steric bulk of the olefin substituent, best results being recorded with 4b (96% yield, 99% ee). The stereochemical course of the epoxidation of 4a is highly dependent on temperature and solvent, the ee of the resulting epoxide (S)-5a varying from 22% (-18°C, CH₂Cl₂) to 49% (55°C, MTBE). The lithium perchlorate induced ring-opening of 5a with piperidine in MeCN affords a mixture of regioisomeric amino alcohols 3a/2a arising from the amine attack at the more substituted and less substituted carbons, respectively. The 3a/2a ratio can be modulated by the LiClO₄ concentration and the reaction temperature, and varies from 44:56 to 91:9. An independent, completely regiocontrolled synthesis of 3a has been developed involving ring-opening by the more substituted carbon of epoxide 5a with diisopropoxytitanium diazide, reduction of azidoalcohol 10a (H₂, Pd/C) and cycloalkylation (1,5- dibromopentane, K₂CO₃) of the amino alcohol 11a. The amino alcohol 3a exhibits a positive nonlinear stereochemical effect in its action as a ligand for the enantioselective addition of Et₂Zn to benzaldehyde. The use of regiochemically pure 3a of >99% ee has been studied in the addition of Et₂Zn to a representative family of aldehydes [19 examples, 93.6% mean ee]. The use of the directly available 9:1 mixture of 3a and 2a derived from 46-47% ee 5a as a ligand system for the enantioselective addition of Et₂Zn to aldehydes [8 examples, 93.1% mean ee] is also reported.

Full text of DOI:10.1055/s-2000-6229

FEATURE ARTICLE
165
Synthesis of a 9-Fluorenone Derived b-Amino Alcohol Ligand Depicting High
Catalytic Activity and Pronounced Non-linear Stereochemical Effects
Katamreddy Subba Reddy, Lluís Solà, Albert Moyano, Miquel A. Pericàs,* Antoni Riera
Unitat de Recerca en Síntesi Asimètrica, Departament de Química Orgànica, Universitat de Barcelona, c/ Martí i Franquès, 1-11, E-08028
Barcelona, Spain
Fax +34(93)3397878; E-mail: mpb@ursa.qo.ub.es
Received 5 August 1999; revised 10 September 1999
As a practical confirmation of these advantages we have
Abstract: The Jacobsen epoxidation of 9-alkylidenefluorenes 4a
been able to develop optimal ligands for the enantioselec-
(ethylidene), 4b (benzylidene) and 4c (1-naphtylmethylene) with
standard (R,R)-manganese salen catalyst has been studied. Both
conversion and enantioselectivity depend on the steric bulk of the
tive addition of dialkylzinc reagents1 to aldehydes2 and
imines,3 and have extended the methodology for the prep-
olefin substituent, best results being recorded with 4b (96% yield, aration of such ligands to epoxides anchored to Merrifield
and Barlos type resins.4 Only quite recently, we have re-
ported the synthesis of amino alcohols 1a-c, derived from
99% ee). The stereochemical course of the epoxidation of 4a is
highly dependent on temperature and solvent, the ee of the resulting
epoxide (S)-5a varying from 22% (-18°C, CH2Cl2) to 49% (55°C,
enantiomerically pure triphenylethylene oxide, and its use
MTBE). The lithium perchlorate induced ring-opening of 5a with
piperidine in MeCN affords a mixture of regioisomeric amino alco-
as highly efficient ligands for the addition of diethylzinc
to aldehydes.2c-d
hols 3a/2a arising from the amine attack at the more substituted and
less substituted carbons, respectively. The 3a/2a ratio can be mod-
ulated by the LiClO4 concentration and the reaction temperature,
and varies from 44:56 to 91:9. An independent, completely regio-
controlled synthesis of 3a has been developed involving ring-open-
N
OH
OH
ing by the more substituted carbon of epoxide 5a with
diisopropoxytitanium diazide, reduction of azidoalcohol 10a (H2,
Pd/C) and cycloalkylation (1,5-dibromopentane, K2CO3) of the
amino alcohol 11a. The amino alcohol 3a exhibits a positive non-
linear stereochemical effect in its action as a ligand for the enantio-
selective addition of Et2Zn to benzaldehyde. The use of regiochem-
ically pure 3a of >99% ee has been studied in the addition of Et2Zn
to a representative family of aldehydes [19 examples, 93.6% mean
ee]. The use of the directly available 9:1 mixture of 3a and 2a de-
rived from 46-47% ee 5a as a ligand system for the enantioselec-
tive addition of Et2Zn to aldehydes [8 examples, 93.1% mean ee] is
also reported.
Ph
Ph
Ph
Ph
Ph
Ph
OH
N
N
O
1a
1b
1c
An inspection of the commonly accepted transition states
for this reaction5 suggested that blocking the mobility of
the geminal phenyl groups in amino alcohols 1 could help
in the differential recognition of the enantiotopic faces of
a reacting aldehyde, thus leading to increased enantiose-
lectivities. To bring these ideas into practice, we decided
to synthesize the families of amino alcohol ligands 2 and
3, with the characteristics shown in Figure 1. We report in
the present paper on the studies aimed to this goal which
have led to the synthesis of a new 9-fluorenyl derived b-
amino alcohol ligand depicting a high catalytic activity
Key words: addition reactions, epoxidations, amino alcohols,
asymmetric catalysis, ligands
Introduction
Over the last years, we have undertaken a research pro-
gram devoted to the development of purely synthetic ami-
no alcohol ligands starting from readily available
enantiomerically pure epoxides. This approach benefits
from a series of advantages: In the first place, the synthe-
sis of the target ligands is modular in nature, and this al-
lows the easy optimization of catalytic properties through
structural modification for every particular situation. Sec-
ondly, since many structurally different epoxides are
readily available in both enantiomeric forms, a great di-
versity of ligands are made available through this method-
ology with complete specification of stereochemistry.
Finally, the chemistry involved in the synthesis of amino
alcohols from epoxides is easily amenable to parallel and
combinatorial work.
Blocked conformation
Substituents R of variable bulk
of the phenyl rings
could be included to test effects
on enantioselectivity
Y
R
X
Y
X
2
3
OH
N
N
Piperidino determined to be optimal
in 1a; could be modified if desired
OH
Figure 1 Design of amino alcohol ligands containing a 9-fluorenyl
group.
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

166
K. S. Reddy et al.
FEATURE ARTICLE
Biographical Sketches
Miquel A. Pericàs was born Barcelona. After postdoctoral Asymmetric Synthesis) work-
in Palma de Mallorca, Spain, studies at the Spanish research ing on different aspects of
in 1951. He studied Chemical council (CSIC) with Prof. asymmetric synthesis. His ma-
Engineering at the Institut Francesc Camps, he joined the jor research interest is focused
Químic de Sarrià and Chemis- Department of Organic Chem- on the development of modu-
try at the Universitat de Barce- istry of the Universitat de Bar- lar ligands for asymmetric ca-
lona, and obtained his celona in 1980 as an associate talysis, the enantioselective
doctorate in 1979 for work on professor. Since 1991 he is full synthesis of bio-active mole-
acetylene diethers and synthe- professor of Organic Chemis- cules and the study of mecha-
sis of oxocarbons under the try at that institution and coor- nistic and synthetic aspects of
guidance of the late Prof. Fèlix dinates there a research group the Pauson-Khand reaction.
Serratosa at the Universitat de (URSA, for Research Unit on
K. Subba Reddy was born in Professor Goverdhan Mehta, synthesis. At the end of 1998
Katamvaripally,
Andhra- University of Hyderabad, Hy- he joined Professor Frank E.
pradesh, India. He received his derabad, India, on some model McDonald’s group at the Emo-
M.Sc. in Chemistry (Organic studies of natural products ry University, Atlanta, USA,
major) from the S.K. Universi- synthesis. From 1997-1998 he as a Postdoctoral fellow. He is
ty, Anantapure, India and PhD worked as Research Associate currently working on transi-
from the Pondicherry Univer- with Professor Miquel A. Peri- tion-metal-catalyzed
sity, Pondicherry, India, in cas, Universitat de Barcelona, tions, application
reac-
to
1994. He worked, from April Barcelona, Spain, on the de- asymmetric synthesis of natu-
1994-December 1996, as a velopment of highly efficient ral products of biological inter-
CSIR Research Associate with chiral ligands for asymmetric est.
Lluís Solà was born in Vic, studies on the preparation of chemical reactions. He also
Spain, in 1967. He studied ethers from alkenes and on the works with Prof. Pericàs re-
chemistry at the Universitat de Jacobsen epoxidation. Since search group on the develop-
Barcelona (Spain) and ob- 1992 he is in charge of the ment of ligands for asym-
tained his diploma in 1990. He Servei de Calorimetria de Re- metric catalysis.
joined Prof. Pericàs group in acció i Anàlisi Tèrmica at the
1991, and under his guidance, Universitat de Barcelona, in-
he received his doctorate in volved in reaction calorimetry
1998 for work on calorimetric and thermal analysis studies of
Albert Moyano was born and where he was appointed Per- reaction, the preparation of
educated in Barcelona. He re- manent Professor of Organic new metallocene-based li-
ceived his PhD from the Uni- Chemistry (in 1987). His cur- gands for asymmetric cataly-
versity of Barcelona (under rent research work, as a mem- sis, and the enantiocontrolled
the guidance of F. Serratosa). ber of the Research Unit on synthesis of biologically ac-
After a postdoctoral fellow- Asymmetric Synthesis (UR- tive compounds.
ship with A. E. Greene in SA), focuses on the develop-
Grenoble he returned to the ment of enantioselective
University of Barcelona, versions of the Pauson-Khand
Antoni Riera was born in in 1987. After a postdoctoral interests include asymmetric
1957 in Balsareny (Barcelona, stage in the University of synthesis, Pauson-Khand re-
Spain). He graduated from the Pennsylvania with Amos B. actions, conformational analy-
University of Barcelona and Smith III he joined the Organ- sis and small-ring cyclo-
then obtained his PhD in the ic Chemistry Department in alkynes.
same University under the su- the University of Barcelona
pervision of Profs. Fèlix Ser- where he is currently perma-
ratosa and Miquel A. Pericàs nent Professor. His research
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

FEATURE ARTICLE
Synthesis of a 9-Fluorenone Derived b-Amino Alcohol Ligand
167
H
OH
OH
H
R
O
H
R
Jacobsen
Regioselective
Ring-opening
R
H
or
Epoxidation
R
N
N
4a-c
5a-c
2a-c
3a-c
Scheme 1
and a pronounced non-linear stereochemical effect with
practical application in the addition of diethylzinc to alde-
hydes.
6 % mol (R,R) - Jacobsen cat.
aq. NaOCl (pH = 11.3)
20 % mol 4-PPNO
H
R
O
H
R
CH2Cl2; 0 - 5 °C
Results and Discussion
4a-c
5a-c
Scheme 3
A. Initial Screening of Ligands
In order to cover a broad range of steric bulk in the R
group in 2/3, the following substituents were selected:
methyl (a), phenyl (b) and 1-naphthyl (c). Then, the syn-
thesis of these amino alcohols was planned from the cor-
responding olefins 4 via Jacobsen epoxidation followed
by regioselective and stereospecific ring-opening of the
intermediate epoxides 5 (Scheme 1).
Table 2 Jacobsen Epoxidation of Olefins 4a-c
Product
Yield (%)
ee (%)
(crude)
ee (%)
(recryst.)
5a
5b
5c
86
28
99
87
–
>99
>99
96
11a
The starting olefins were readily prepared from fluo-
renone 6, by a two-step process involving Grignard re-
agent addition followed by dehydration of the
intermediate alcohol (Scheme 2 and Table 1).
a Compound 4c was also recovered in 59%.
ing epoxidations of 4b and 4c occurred with high enantio-
selectivity (albeit in rather low yield in the case of 4c). Ep-
oxides 5b and 5c turned out to be crystalline solids that
could be brought to complete enantiomeric purity by re-
crystallization from hexanes.
1) RCH2MgX
H
R
solvent, temp.; aq. NH4Cl
O
2) 20% H2SO4 in AcOH
∆, 60 °C, 60 min
For the preparation of amino alcohol ligands 2a-c, ep-
oxides 5a-c were submitted to ring-opening with
piperidine2c under Crotti’s conditions8 as indicated in
Scheme 4 and Table 3.
6
4a-c
Scheme 2
Olefins 4a-c were next submitted to Jacobsen
epoxidation6 mediated by the standard (R, R) salen cata-
lyst in the presence of 4-phenylpyridine N-oxide (4-PP-
NO) to afford enantio-enriched epoxides 5a-c as shown
in Scheme 3 and Table 2.
Piperidine (10 eq)
LiClO4 (10 eq)
OH
H
N
O
H
R
R
OH
H
+
N
CH3CN (2mL/mmol)
R
As it can be seen, the epoxidation of 4a took place with
disappointingly low enantioselectivity. Moreover, 5a was
obtained as an oil and enantio-enrichment at this stage
proved to be difficult.7 More gratifyingly, the correspond-
5a-c
2a-c
3a-c
Scheme 4
Table 1 Synthesis of Olefins 4a-c
In the case of 5a, the ring-opening reaction proceeded
very fast in acetonitrile under reflux but, in contrast to
what was observed for triphenylethylene oxide,2c the pro-
cess was not regioselective leading in 84% yield to a
30:70 mixture of the regioisomeric amino alcohols 2a and
3a. These compounds turned out to be separable (not
R
X
Solvent Temp (°C) Product Yield (%)
Me
Ph
Cl
Br
THF
Et2O
Et2O
0
23
23
4a
4b
4c
56
73
30
1-naphthyl Cl
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

168
K. S. Reddy et al.
FEATURE ARTICLE
Table 3 Lithium Perchlorate-Induced Ring-Opening of Epoxides
5a-c with Piperidine
HO
H
6 % mol 2 or 3
Et2Zn
Ph-CHO
+
Ph
toluene, 0 °C, 3-4 h
(S)
Products
Conditions
Ratio of
2/3
Yield
(%)
8a
9a
Scheme 5
2a/3a
2b/3b
2b/3b
2c/3c
80°C, 1 h
80°C, 3 h
23°C, 120 h
80°C, 24 h
30:70
100:0
0:100
–
84
10a
10b
c
–
a Compound 7 was isolated in 55% yield.
b Compound 7 was isolated in 60% yield.
c Compound 7 was formed almost exclusively.
Table 4 Addition of Et2Zn to Benzaldehyde (8a) Mediated by 3a,
2b, and 3b
Ligand (% ee) Conv. (%)a
Sel. (%)b
ee 9a (%)c
3a (28)
2b (>99)
3b (>99)
95
80
84
98
82
82
92
17
33
without difficulties) by column chromatography and crys-
tallization from hexanes. A sample of the major regioiso-
mer 3a (28% ee) could be separated in this way.
a Conversion as determined by the integration of residual 8a among the
new products in the gas chromatogram of the crude reaction mixture.
b Selectivity as determined by integration of product 9a (both enantio-
mers) among the new compounds in the gas chromatogram of the
crude reaction mixture.
In the case of epoxide 5b, the reaction required 3 hours for
completion in acetonitrile at reflux. Interestingly, only the
initially sought regioisomer 2b was isolated under these
conditions, albeit in rather low yield (10%). This amino
alcohol was accompanied in the crude reaction mixture by
1-(9-fluorenyl)piperidine (7),9 formed in a substantial
amount (55%), which probably comes from the cleavage
of regioisomer 3b in the reaction medium.
c Determined by GC on a b-DEX 120 column.
The results obtained with the phenyl containing ligands
2b and 3b were mediocre (Table 4) and did not warrant
further scrutiny of the use of these species. On the other
hand, 3a, in spite of its low enantiomeric purity, mediated
the formation of 9a efficiently both in terms of conversion
and stereochemical control. A pronounced non-linear ste-
reochemical effect11 was clearly operating in the reaction
mediated by this ligand. According to these observations,
the regioisomeric amino alcohols 2a and 3a were selected
as main targets of our research, and work was undertaken
aimed at the optimization of their syntheses.
7
N
In an attempt to improve the yield of 2b the reaction was
repeated at room temperature. Under these conditions, the
reaction required five days for completion; surprisingly
the only amino alcohol product was 3b (10% yield), ac-
companied again by a major amount (60%) of 7. With re-
spect to the regiodivergent course of the ring-opening, the
results tend to indicate that at room temperature nucleo-
philic attack takes place exclusively at the more substitut-
ed (doubly benzylic) carbon, whereas at 80°C attack at the
less substituted takes place to some extent. This is based,
of course, on the assumption that 7 arises in all cases from
3b.
B. Studies on the synthesis of 2a/3a
The first problem we wanted to solve was the low enantio-
meric purity of 5a. To this end, we systematically studied
the Jacobsen epoxidation of 4a (Scheme 6). In an initial
set of experiments performed in dichloromethane solu-
tion, the effect of the reaction temperature was analyzed.
In a second stage, the effect of alternative solvents was
studied. The results of all these reactions have been sum-
marized in Table 5.
Finally, the ring-opening of 5c could not be performed.
Although high temperature and long reaction time were
used, the nucleophilic attack did not take place at synthet-
ically useful rates. Not unexpectedly, ring-opening of ep-
oxides 5 appears to be highly dependent on the steric bulk
of the R substituent.
6 % mol (R,R) - Jacobsen cat.
H
O
H
aq. NaOCl (pH = 11.3)
20 % mol 4-PPNO
CH3
CH3
solvent, temp., 30 min
4a
5a
At this stage of the research, the amino alcohols which
had been obtained in regiochemically pure form (3a, 2b
and 3b) were tested as ligands in the enantioselective ad-
dition of diethylzinc to benzaldehyde (8a) leading to (S)-
1-phenylpropanol (9a)10 (Scheme 5).
Scheme 6
From the experiments performed in dichloromethane (En-
tries 1-4) it becomes clear that temperature possesses a
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

FEATURE ARTICLE
Synthesis of a 9-Fluorenone Derived b-Amino Alcohol Ligand
169
Table 5 Effect of Temperature and Solvent on the Jacobsen Epoxi-
dation of 4a to 5a
OH
N
H
H
N
ee (%)a Yield (%)b
CH3
Entry Solvent
Temp. (°C)
O
H
OH
excess
+
H
CH3
N
CH3
LiClO4
1
2
3
4
5
6
7
8
9
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
(CH2Cl)2
EtOAc
MTBE
MTBE
TAME
-18
0
22
28
35
38
38
44
38
49
49
47
<10
86
90
90
89
93
92
90
85
95
35
40
83
65
25
55
87
5a
3a
2a
Scheme 7
In a first series of experiments, the ring-opening was per-
formed in excess piperidine as the solvent, in the presence
of increasing amounts of LiClO4 (Entries 1-5). In agree-
ment with our expectations, the 3a/2a regioisomeric ratio
increased from 44:56 (0.1 equiv of added LiClO4, Entry 1)
to 69:31 (15 equiv of added LiClO4, Entry 5). Keeping the
amount of LiClO4 constant, further improvements in re-
gioselectivity could be achieved through the use of MeCN
as a cosolvent (up to 82:18, Entry 6), and by simply low-
ering the reaction temperature from 50 to 25°C (up to
91:9, Entry 7). A 95% yield of 3a/2a can consistently be
obtained in this way. Thus, a simple modification of reac-
tion conditions can direct the ring-opening of 5a towards
the high yield formation of 3a.
10
MTBE/EtOAc (9:1) 50
a Determined by HPLC on a Chiracel-OD column.
b Referred to isolated product.
marked beneficial effect on the enantiomeric excess of the
resulting epoxide. Thus, while the reaction performed at
-18°C affords (S)-5a with 22% ee, the reaction at 40°C
delivers the epoxide with 38% ee. While a parallel effect
has been noted in some instances12 by using a similar cat-
alytic system, the variation of ee with temperature is re-
stricted in those examples to 2-5%. Thus, the present
example represents from the quantitative point of view a
paradigm in positive effect of temperature on enantiose-
lectivity of Jacobsen epoxidation.
The separation of 3a and 2a could be performed by com-
bination of column chromatography and crystallization
from hexanes, as dictated by the initial composition of the
regioisomeric mixture. Very interestingly, the enantio-
meric composition of 3a is also enriched (up to >99.9% ee
when starting from epoxide 5a of 95% ee7) during crystal-
lization. The amino alcohols 3a and 2a of high enantio-
meric purity arising from these separations were
compared as ligands for the addition of diethylzinc to
benzaldehyde 8a in toluene solution. In the case of 3a (2%
mol, >99.9% ee) complete conversion was observed after
3 hours at 0°C, and (S)-1-phenylpropan-1-ol 9a of 96.1%
ee was formed with >98% selectivity. Conversely, amino
alcohol 2a turned out to be much less efficient as a ligand
than expected. Thus, in the presence of a 6% mol of 2a
(95% ee, arising from 95% ee 5a7) conversion was incom-
To test the possibility of a further increase of the enantio-
selectivity at higher temperatures, a selection of solvents
with boiling points higher than dichloromethane were also
tested in the reaction (Entries 5-10). Whereas the use of
1,2-dichloroethane at reflux provokes no increase in the
ee of 5a, a slight improvement is recorded with ethyl ace-
tate at 65°C (44% ee). Most notably the branched ethers
MTBE (tert-butyl methyl ether) and TAME (tert-amyl
methyl ether) allow the formation of 5a of 49% ee. For
practical reasons (solubility of all components of the reac-
tion system) preparative experiments were best performed
in a 9:1 mixture of MTBE/EtOAc at 50°C. In this way, a
95% yield of 5a with 46-47% ee7 can be consistently iso-
lated after a fast reaction.
A final aspect of the synthesis of ligands 2a/3a still requir-
ing improvement was the regiochemical control in the
ring-opening of 5a. With respect to this, we reasoned that
variations in the concentration of LiClO4 would provoke
a parallel variation in the polarity of the reaction medium.
If lithium cation was acting in the reaction medium as a
Lewis acid coordinated to the oxygen atom of the epoxide,
the ionic nature of the aggregate would evolve from oxo-
nium towards carbenium type when the polarity of the re-
action medium increases. According to this, the
regioselectivity of the ring-opening process could be per-
haps controlled by the concentration of LiClO4.
Table 6 Effect of LiClO4 Concentration and Temperature on the Re-
gioselectivity of the Ring-Opening of 5a with Piperidine
Entry
LiClO4
(equiv)
Temp. (°C) Time (h)
Ratio of
3a/2ab
1
2
3
4
5
6
7
0.1
1
2
50
80
80
60
60
50
25
24
1
1
1
1
44:56c
48:52
55:45
61:39
69:31
82:18
91:9d
5
15
15a
15a
1
6
a Acetonitrile (1.5 mL/mmol 5a) was used as a cosolvent.
b By 1H NMR spectroscopy.
To test this hypothesis we studied the ring-opening of 5a
with piperidine in the presence of lithium perchlorate at
different concentrations (Scheme 7). The results of this
study are summarized in Table 6.
c Conversion was 20%.
d A mixture of 3a/2a could be isolated in 95% yield.
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

170
K. S. Reddy et al.
FEATURE ARTICLE
C. Configurational Assignments
plete (78%) under the aforementioned reaction condi-
tions, and 9a was obtained with only 52% ee. The reaction
could be driven to completion (99% conversion) by in-
creasing the amount of 2a to 9% mol, but the enantiomeric
purity of 9a only experienced a minor improvement (63%
ee). In view of these results, we decided to concentrate our
efforts on the use of 3a as a ligand.
Guided by contemporary results in our research group,2d
we decided to dedicate some effort to the development of
a synthetic sequence leading exclusively to this regioiso-
mer. To this end, we studied the ring-opening of epoxide
5a with diisopropoxytitanium (IV) diazide,13 which could
allow the highly regioselective introduction of a nitrogen
substituent at the more substituted carbon in the epoxide
ring. A subsequent reduction of the azide and final alkyla-
tion at nitrogen would then allow the construction of the
piperidine ring present in 3a.
In order to establish the absolute configuration of 3a, this
amino alcohol was converted to a pair of diasteromeric O-
methylmandelate esters (S,S)-12 and (S,R)-12 by reaction
with the enantiomers of O-methylmandelic acid (DCC-
DMAP).14 Analysis of the 1H NMR spectra of the diaste-
reomeric esters according to well established empirical
correlations15 (Figure 2) allowed the assignment of an S-
configuration to amino alcohol 3a.
According to this plan, 5a was treated with
[Ti(OPr-i)2(N3)2] in benzene solution at 65°C for 30 min
(Scheme 8) to afford the azidoalcohol 10a with very high
regioselectivity (only one regioisomer could be isolated)
in 91% yield. Reduction of the azide by hydrogenation
over 10% Pd/C in methanol at room temperature allowed
the isolation of amino alcohol 11a in 97% yield. Finally,
the piperidino ring was easily constructed by reaction with
1,5-dibromopentane in ethanol under reflux in the pres-
ence of potassium carbonate. In this way, 3a was obtained
in 72% yield.
Figure 2 Computer-generated view of the diastereomeric esters
(S,R)-12 and (S,S)-12 in the proposed extended conformation (AM1
geometry optimization). Encircled protons should be more shielded.
Further support to this assignment can be found in the fact
that the enantioselective addition of Et2Zn to benzalde-
hyde mediated by 3a predominantly leads to the S-enanti-
omer of 1-phenylpropan-1-ol, in full agreement with the
prediction resulting from the application of the very gen-
eral empirical rule developed by Noyori.1b
Ti(OiPr)2(N3)2
N3
OH
O
H
C6H6, 65 °C, 0.5 h
H
Taking into account the anti-stereospecificity of the ring-
opening process, it follows that 2a should possess R-con-
figuration. If this is assumed, the signs of the enantiose-
lectivities dictated by Noyori’s rule and observed for the
addition of Et2Zn to benzaldehyde mediated by 2a are in
complete agreement.
CH3
CH3
91 %
5a
10a
H2 (1 Atm)
97% 10 % Pd-C
MeOH, r.t.
From the configurations of 3a and 2a, it follows that the
configuration of 5a should be S. This is also in agreement
with the absolute configuration of the closely related (S)-
1,1-diphenylpropene oxide obtained in the Jacobsen ep-
oxidation of 1,1-diphenylpropene with the same enantio-
mer of the catalyst.6b
Br
Br
N
NH2
OH
OH
K2CO3, EtOH
H
CH3
H
CH3
reflux, 24h
72 %
In agreement with this assignment, 11a and 10a should
also possess S-configuration: whereas 11a leads to (S)-3a
by N-alkylation; 10a comes from (S)-5a through a process
that does not affect the chiral center in the molecule.
3a
11a
Scheme 8
Completely parallel arguments based on both Noyori’s
rule and the expected parallelism between the enantiose-
lectivities in the epoxidation step, allows for the assign-
ment of the configuration of the phenyl substituted species
as follows: (S)-3b, (R)-2b and (S)-5b. Finally, the S-con-
figuration has been assigned to 5c in analogy to 5a and 5b.
It is important to realize that this procedure, besides its
complete regiochemical control, offers the additional ad-
vantage of being applicable to the synthesis of related
amino alcohols by systematic variation of the substituents
at the nitrogen. This characteristic could be relevant when
fine tuning of catalytic properties for a given application
is required.
All over the manuscript, the chiral intermediates and
products have been represented with the configurations
discussed here.
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

FEATURE ARTICLE
Synthesis of a 9-Fluorenone Derived b-Amino Alcohol Ligand
171
induce the enantioselectivities reported in Table 7, in the
case of aliphatic and a,b-unsaturated aldehydes (8l-s) a
6% molar amount of the ligand is more conveniently used
in order to bring the reactions to completion in 4 hours.
D. Use of Amino Alcohol 3a as a Ligand for the High-
ly Enantioselective Ethylation of Aldehydes
We have already mentioned that enantiopure 3a (2% mol)
efficiently controls the addition of diethylzinc to benzal-
dehyde (8a) at 0°C, (S)-1-phenylpropan-1-ol (9a) being
formed in >96% ee. As a preliminary step to the use of 3a
as a ligand for the ethylation of aldehydes of different
structural types, other reaction temperatures were tested
for the addition to 8a.
3-6% mol Ligand
OH
(S)
Et2Zn
RCHO
R
Toluene, 0°C, 3-4 h
8a-s
9a-s
To this end, a 3% molar amount of 3a was used in the
enantioselective addition of diethylzinc to benzaldehyde
at room temperature producing 9a in over 94% ee. On the
other hand, performing the reaction at -20°C resulted in
some enhancement in the ee of 9a (97.5%). In view of
these results, the most convenient temperature of 0°C was
selected for the enantioselective ethylation of a variety of
aromatic and aliphatic aldehydes (Scheme 9 and Table 7).
While ortho-substituted aromatic aldehydes (8b-d), 1-
naphthaldehyde (8j) and linear aliphatic aldehydes (8l-n,
8p) undergo the addition reaction with generally good lev-
els of enantioselectivity (albeit slightly lower than for
conformationally unrestricted analogous ligands), excel-
lent enantiomeric excesses (94-98%) are recorded for
meta- and para-substituted benzaldehydes, (8e-i), 2-
naphthaldehyde (8k), branched aliphatic aldehydes (8o,
8q-r) and a-substituted a,b-unsaturated aldehydes (8s).
Although a 3% molar amount of the ligand is sufficient to
Scheme 9
E. Non-Linear Stereochemical Effects in the Enan-
tioselective Addition of Et2Zn to Benzaldehyde Medi-
ated by 3a
As we have already mentioned, a positive non-linear ste-
reochemical effect (NLE) was observed when 3a with
28% ee was used as a ligand for the addition of Et2Zn to
benzaldehyde. The availability of samples of 3a of differ-
ent enantiomeric compositions, arising from the different
preparations and enrichments of 5a, allowed a systematic
study of enantiomeric composition of the resulting 1-phe-
nylpropan-1-ol 9a as a function of the enantiomeric com-
position of 3a. The reactions were performed in toluene
solution at 0°C for 4 hours. The results of this study are
shown in Table 8 and graphically represented in Figure 3.
An interesting observation is that samples of ligand 3a
Table 7 Catalytic Enantioselective Addition of Et2Zn to Aldehydes 8a-s Leading to Alcohols 9a-s Mediated by Ligand 3a and by the Ligand
System URSA99
Starting Aldehyde
Product
Ligand 3a (> 99% ee)
Ligand URSA99 (46-47% ee)
Con.
(%)a
Sel.
ee
Con.
(%)a
Sel.
ee
(%)b
(%)c
(%)b
(%)c
Benzaldehyde (8a)
o-Chlorobenzaldehyde (8b)
o-Tolualdehyde (8c)
o-Methoxybenzaldehyde (8d)
m-Chlorobenzaldehyde (8e)
m-Tolualdehyde (8f)
m-Methoxybenzaldehyde (8g)
p-Fluorobenzaldehyde (8h)
p-Tolualdehyde (8i)
1-Naphthaldehyde (8j)
2-Naphthaldehyde (8k)
Hexanal (8l)
Heptanal (8m)
Nonanal (8n)
Isovaleraldehyde (8o)
3-Phenylpropanal (8p)
Cyclohexanecarbaldehyde (8q)
2-Ethylbutyraldehyde (8r)
(E)-a-Methylcinnamaldehyde (8s)
Mean value
(S)-9a
(S)-9b
(S)-9c
(S)-9d
(S)-9e
(S)-9f
(S)-9g
(S)-9h
(S)-9i
(S)-9j
(S)-9k
(S)-9l
(S)-9m
(S)-9n
(S)-9o
(S)-9p
(S)-9q
(S)-9r
(S)-9s
99
99
99
100
100
100
100
100
100
87
99
99
99
85
95
99
98
99
98
94
94
97
99
98
99
99
100
97
98
98
99
98
99
99
100
96
98
96
91
96
82
97
96
96
97
100
100
98
93
96
96
100
100
98
99
96
96
97
90
96
91d
91d
90d
95e
91e
98e
94d
94d
93.6
99
100
98
99
89d
89d
99
65
97
94
90e
71
93d
93.1
a Conversion as determined by integration of residual 8 among all new products in the gas chromatogram of the crude reaction mixture.
b Selectivity as determined by integration of 9 (both enantiomers) among all new compounds in the gas chromatogram of the crude reaction
mixture.
c Determined by GC on a b-DEXTM 120 column unless otherwise specified.
d The acetate derivative was analyzed by GC on a b-DEXTM 120 column.
e Determined by GC on a a-DEXTM 120 column.
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

172
K. S. Reddy et al.
FEATURE ARTICLE
Table 8 Non-linear Stereochemical Effect in the Enantioselective
Addition of Et2Zn to Benzaldehyde Mediated by 3a.
ate, high levels of enantioselectivity can be achieved in
the reactions for which the ligand has been designed.
3a [mol %]
ee of 3a [%]
ee of 9a [%]
In the present case, the important consequence is that 3a
arising from readily available 46% ee 5a could be directly
employed as a ligand without enantiomeric enrichment.
8
6
6
6
3
3
8.1
34.2
83.5
89.2
92.3
96.2
96.3
16.8
22.6
28.0
46.0
99.9
F. URSA99: A Practical Ligand System for the Ethyl-
ation of Aldehydes
This observation has ultimately led to the development of
a most practical system for the use of 3a as a ligand. As
we have already seen, the amino alcohol mixture 3a/2a
can be prepared in a very efficient manner through the
ring-opening of 5a with piperidine. Starting from the di-
rectly available epoxide of 46-47% ee, the amino alco-
hols can be obtained in 95% yield as a 91:9 mixture of
regioisomers. The catalytic results obtained with the indi-
vidual regioisomers which have already been discussed in
this paper, suggested that this mixture could be directly
used as a ligand system (which we call URSA99) without
the need of any regioisomeric separation, nor of enantio-
meric enrichment.16
with an enantiomeric purity of at least 46% catalyze the
addition of diethylzinc to benzaldehyde with the maximum
observed enantioselectivity.
Thus, both 2a and 3a induce the same sign of enantiose-
lectivity in the addition of diethylzinc to aldehydes. More-
over, although the additions mediated by 2a take place
with lower enantioselectivity they are also substantially
slower than those mediated by 3a. These observations, to-
gether with the consideration of the NLE associated to the
use of 3a, suggested that the presence of small amounts of
2a would not perturb the catalytic activity and enantiose-
lectivity of 3a.
Figure 3 Non-linear stereochemical effects in the enantioselective
addition of Et2Zn to benzaldehyde mediated by 3a.
To test this hypothesis, URSA99 (6% molar amount) was
used to induce the addition of diethylzinc to a representa-
tive set of aldehydes. The results of this study have been
included in Table 7 for comparison with those obtained
with the more difficultly available regioisomerically pure
3a of 99% ee. Very interestingly, no difference in catalyt-
ic activity is observed between 3a and the regioisomeric
mixture URSA99. From the point of view of enantioselec-
tivity, when the performance of both catalytic systems is
analyzed on a basis of 8 different aldehydes where direct
comparison is available, the mean enantiomeric excess of
the resulting alcohols is only decreased by 0.8% (from
93.9 to 93.1%) in going from regio- and enantiomerically
pure 3a to URSA99.
It is generally accepted that the observation of non-linear
stereochemical effects in the addition of Et2Zn to alde-
hydes mediated by amino alcohols arises from the differ-
ent stability of the homo- and heterochiral dimers of the
ethylzinc alkoxide derived from Et2Zn and the amino al-
cohol ligand.11 Many different examples of this behavior
are known, and the matter has been recently reviewed.11c-d
From a quantitative point of view, 3a exhibits this behav-
ior to a very high degree. In any case, the observation of
NLE with purely synthetic ligands like 3a possesses an
additional, qualitative interest. Thus, if we consider
ligands derived from naturally occurring, enantiomerical-
ly pure products, the interest of their NLE is purely aca-
demic except for the phenomena of chiral aplification; i.e.
the enantiomerically pure ligand is directly available
whereas the scalemic samples have to be synthesized by
mixing known amounts of their enantiomers. On the con-
trary, when dealing with ligands of synthetic origin, like
in the present case, the enantiomeric purity of the ligand
is limited by the efficiency of the synthetic process lead-
ing to it. It is thus of primordial importance that, even
when the enantiomeric purity of the ligand is only moder-
Conclusions
In summary, the investigations reported here have re-
vealed some unique properties of the 9-ethylidenefluo-
rene system: First, the enantioselectivity of the Jacobsen
epoxidation of the olefin is unusually strongly dependent
on temperature and solvent, the ee of the resulting epoxide
varying from 22% at -18°C in dichloromethane to 49% at
55°C in tert-butyl methyl ether. Second, the regioselectiv-
ity of the ring-opening of the epoxide with piperidine can
be controlled to an important degree by the polarity of the
reaction medium and, to some extent, by the reaction tem-
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

FEATURE ARTICLE
Synthesis of a 9-Fluorenone Derived b-Amino Alcohol Ligand
173
um. The residual solid was purified by chromatography to afford
5.06 g (93%) of the intermediate carbinol which was subsequently
dehydrated as described for 4a to afford 3.55 g (73% overall yield)
of 4b as a white solid; mp 75-77°C (Lit.17 mp 76-77°C).
perature. Third, the major regioisomer arising from the
ring-opening process, 3a, which can be also prepared in a
completely selective manner through the ring-opening of
epoxide 5a with diisopropoxytitanium diazide followed
by reduction and cycloalkylation, behaves as a very active
and enantioselective catalytic ligand and depicts a pro-
nounced non-linear stereochemical effect in the addition
of diethylzinc to aldehydes.
1H NMR (200 MHz, CDCl3): d = 7.00-7.80 (m, 14 H).
9-(Naphtylmethylene)fluorene (4c)
Reaction was conducted as described for 4b. Starting from Mg
(1.22 g, 50 mmol), 1-(chloromethyl)naphthalene (8.2 mL, 55 mmol)
and 9-fluorenone (6; 7.2 g, 40 mmol), olefin 4c (3.6 g, 30%) was ob-
tained as a white solid; mp 146°C.
As an important practical consequence of the different
catalytic activity of regioisomers and non-linear ste-
reochemical effects, the regioisomeric mixture of amino
alcohols (3a/2a = 91:9) directly available from epoxide
5a in moderate enantiomeric purity (46-47% ee) can be
used as a catalytic system with practically the same effi-
ciency as the regio- and enantiomerically pure amino al-
cohol 3a. Research directed to the identification of related
regioisomeric alcohols giving rise to similar or even coop-
erative effects are underway in our laboratories.
1H NMR (200 MHz, CDCl3): d = 6.80-8.20 (m, 15 H).
13C NMR (50 MHz, CDCl3): d = 120.2 (CH), 120.2 (CH), 121.0
(CH), 125.3 (CH), 125.7 (CH), 125.8 (CH), 126.0 (CH), 126.8
(CH), 126.9 (CH), 127.3 (CH), 127.6 (CH), 127.7 (CH), 128.9
(CH), 129.1 (CH), 132.2 (C), 134.2 (C), 134.8 (C), 137.3 (C), 138.5
(C), 139.7 (C), 139.9 (C), 141.7 (C). Two (CH) signals were not ob-
served.
Jacobsen Epoxidation of 4a to (15S)-15-Methylspiro[fluorene-
9,3’-oxirane] [(S)-5a]
(i) In CH2Cl2 at 0°C: Buffered bleach (100 mL, 72 mmol, pH 11.3)
at 0°C was added dropwise during 60 min to a solution of 4a (10.0
g, 51.5 mmol), (R,R)-Jacobsen catalyst [(R,R)-(-)-N,N'-bis(3,5-di-
tert-butylsalicylidene)-1,2-cyclohexanediaminomanganese(III)
chloride] (2.0 g, 3.2 mmol) and 4-phenylpyridine N-oxide (1.76 g,
10.3 mmol) in CH2Cl2 (100 mL) at 0°C. After 30 min the two phases
were separated, the aqueous phase was extracted with CH2Cl2
(2 î 75 mL) and the combined organic phases were washed with
H2O (2 î 150 mL), dried and concentrated under vacuum. The re-
sidual oil was purified by chromatography to afford 9.32 g (86%) of
(S)-5a of 28% ee as an oil. Conditions of HPLC analysis: Chiralcel
OD column, 25 cm, 30 °C; eluent: hexane/propan-2-ol (95:5); flow
rate, 0.5 mL/min; R-isomer, tR:16.0 min and S-isomer tR:17.4 min.
[a]D -44.2 (c = 1.57, CDCl3, 95% ee).
IR (KBr): n = 3040, 2960, 1447, 886, 741 cm-1.
1H NMR (200 MHz, CDCl3): d = 1.69 (d, J = 5 Hz, 3 H), 3.87 (q,
J = 5 Hz, 1 H), 7.22-7.48 (m, 6 H), 7.78-7.88 (m, 2 H).
13C NMR (50 MHz, CDCl3): d = 14.6 (CH3), 61.8 (CH), 66.6 (C),
120.0 (CH), 120.4 (CH), 121.3 (CH), 124.3 (CH), 126.9 (CH),
127.4 (CH), 128.9 (CH), 139.6 (C), 140.2 (C), 141.8 (C), 142.6 (C).
One (CH) signal was not observed.
Optical rotations were measured at r.t. (23°C) (concentration in g/
100 mL). Melting points were determined in open capillary tubes
and are uncorrected. IR spectra were recorded as film between NaCl
plates or by KBr pellet techniques. 1H NMR were recorded at 200,
300 or 500 MHz (s = singlet, d = doublet, t = triplet, m = multiplet,
br = broad). 13C NMRwere recorded at 50.3 MHz, 75.4 MHz or 125
MHz. TMS was used as an internal reference in NMR measure-
ments. Carbon multiplicities have been assigned by distortionless
enhancement by polarization transfer (DEPT) experiments. Chro-
matographic separations were carried out using Et3N pre-treated
(2.5% v/v) silica gel (70-230 mesh), eluting (unless otherwise stat-
ed) with hexanes/EtOAc mixtures of increasing polarity.
9-Ethylidenefluorene (4a)17
A 1 M THF solution of EtMgBr (34.4 mL, 34.4 mmol) was added
to a solution of 9-fluorenone (6; 5.00 g, 27.5 mmol) in THF (20 mL)
at 0°C. After 2 h, satd aq NH4Cl solution (100 mL) and CH2Cl2 (100
mL) were added. The phases were separated, the aqueous phase ex-
tracted with CH2Cl2 (3 î 100 mL) and the combined organic phas-
es were dried and concentrated under vacuum. The residual solid
was purified by chromatography to afford 4.20 g (73%) of the inter-
mediate alcohol which was subsequently dissolved in a 2:8 mixture
of H2SO4/AcOH (17 mL) and heated at 60°C. After 1 h, the reaction
mixture was cooled to r.t. and Et2O (75 mL) and H2O (75 mL) were
added. The aqueous phase was extracted with Et2O (2 î 75 mL)
and the combined organic phases were extracted with satd aq
NaHCO3 solution (75 mL), H2O (75 mL), brine (75 mL) and H2O
(75 mL), dried and concentrated under vacuum. The residual solid
was purified by chromatography to afford 2.94 g (77%) of 4a as a
white solid; mp 52-54°C (Lit.17 mp 53-54°C).
MS (CI, NH3): m/z (%) = 209 (C15H12O·H+, 10), 226
+
(C15H12O·NH4 , 100).
Anal. Calcd for C15H12O (208.3): C, 86.51; H, 5.81. Found: C,
86.66; H, 5.81.
(ii) In MTBE/EtOAc (9:1): Buffered bleach (110 mL, 78 mmol,
pH 11.3) was added dropwise at 50°C to a solution of 4a (10.7 g,
55.7 mmol), (R,R)-Jacobsen catalyst (2.12 g, 3.34 mmol) and 4-phe-
nylpyridine N-oxide (1.905 g, 11.14 mmol) in a 9:1 mixture of
MTBE and EtOAc (100 mL). After 30 min the reaction was cooled
to r.t. and the phases were separated. The aqueous phase was ex-
tracted with EtOAc (2 î 100 mL) and the combined organic phases
were washed with H2O (2 î 150 mL), dried and concentrated un-
der vacuum. The residual oil was purified by chromatography to af-
ford 11 g (95%) of (S)-4a of 47% ee as an oil.
1H NMR (200 MHz, CDCl3): d = 2.40 (d, J = 8 Hz, 3 H), 6.86 (q,
J = 8 Hz, 1 H), 7.20-8.00 (m, 8 H).
9-(Phenylmethylene)fluorene (4b);17 Typical procedure
Benzyl bromide (3.56 mL, 30 mmol) was slowly added to a suspen-
sion of Mg (0.73 g, 30 mmol) in anhyd Et2O (100 mL) at a rate that
generated a gentle reflux and stirred at that temperature for an addi-
tional 2 h. The mixture was cooled to r.t. and a solution of 9-fluo-
renone (6; 3.6 g, 20 mmol) in Et2O (100 mL) was added dropwise.
After 2 h, satd aq NH4Cl solution (100 mL) was added and the mix-
ture was stirred for 0.5 h. H2O (100 mL) and Et2O (100 mL) were
then added, the phases separated and the aqueous phase was extract-
ed with Et2O (3 î 100 mL). The organic phase was washed with
H2O (100 mL), brine (100 mL), dried and concentrated under vacu-
(15S)-15-Phenylspiro[fluorene-9,3’-oxirane] [(S)-5b]
Buffered bleach (24 mL, 17 mmol, pH 11.3) at 0°C was added drop-
wise to a solution of 4b (3.0 g, 11.8 mmol), (R,R)-Jacobsen catalyst
(0.45 g, 0.71 mmol) and 4-phenylpyridine N-oxide (0.40 g, 2.4
mmol) in CH2Cl2 (24 mL) at 0°C. After 90 min the reaction mixture
was treated as described for 5a to afford 3.00 g (96%) of (S)-5b as
white crystals of 99% ee. A single recrystalization from hexanes af-
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

174
K. S. Reddy et al.
FEATURE ARTICLE
fords optically pure material. Conditions of HPLC analysis: Chiral-
cel OD-R column, 25 cm, 30°C; eluent: MeOH; flow rate: 0.5 mL/
min; R-isomer, tR: 18.5 min and S-isomer tR: 20.2 min; mp 77-
78°C; [a]D -200.5 (c = 1.34, CDCl3).
IR (KBr): n = 1451, 739, 724 cm-1.
1H NMR (200 MHz, CDCl3): d = 4.96 (s, 1 H), 6.51 (d, J = 11 Hz,
1 H), 6.88-6.97 (m, 1 H), 7.20-7.50 (m, 9 H), 7.60-7.80 (m, 2 H).
13C NMR (75 MHz, CDCl3): d = 65.7 (CH), 67.6 (C), 120.1 (CH),
120.2 (CH), 121.7 (CH), 124.1 (CH), 126.9 (CH), 127.0 (CH),
127.6 (CH), 128.2 (CH), 128.3 (CH), 129.1 (CH), 129.3 (CH),
135.2 (C), 138.7 (C), 140.6 (C), 141.6 (C), 141.8 (C).
125.2 (CH), 125.8 (CH), 126.5 (CH), 127.2 (CH), 128.2 (CH),
128.3 (CH), 140.7 (C), 141.9 (C), 142.5 (C), 144.8 (C).
MS (CI, NH3): m/z (%) = 294 (C20H23ON·H+, 100).
Anal. Calcd for C20H23NO (293.4): C, 81.87; H, 7.90; N, 4.77.
Found: C, 82.11; H, 7.92; N, 4.66.
(ii) Isolation of (R)-2a from a 48:52 Regioisomeric Mixture (S)-3a/
(R)-2a: Epoxide (S)-5a (0.208 g, 1 mmol, 95% ee) and LiClO4
(0.106 g, 1 mmol) were dissolved in piperidine (1 mL) and the re-
sulting mixture was heated at 80°C for 1 h. After workup as above,
the crude product was submitted to careful column chromatography
using a high silica gel to substrate ratio and slowly increasing the
polarity of the eluent. From the first fractions of the eluate, a highly
enriched sample of 2a was isolated. A recrystallization from hex-
anes afforded regiochemically pure 2a; mp 97°C; [a]D -80 (c = 1.0,
CDCl3, 95% ee).
MS (CI, NH3): m/z (%) = 271 (C20H14O·H+, 1), 288 (C20H14O·NH4 ,
+
100).
Anal. Calcd for C20H14O (270.3): C, 88.86; H, 5.22. Found: C,
88.79; H, 5.18.
IR (KBr): n = 3208, 2931, 1451, 735 cm-1.
1H NMR (200 MHz, CDCl3): d = 0.44 (d, J = 7 Hz, 3 H), 1.51-1.59
(m, 2 H), 1.70-1.80 (m, 4 H), 2.40-2.61 (m, 2 H), 2.78-2.90 (m, 2
H), 3.05 (q, J = 7 Hz, 1 H), 5.98 (br s, 1 H), 7.20-7.61 (m, 8 H).
13C NMR (50 MHz, CDCl3): d = 8.2 (CH3), 24.6 (CH2), 26.9 (CH2),
53.7 (CH2), 69.0 (CH), 83.3 (C), 119.5 (CH), 119.7 (CH), 123.0
(CH), 124.3 (CH), 126.9 (CH), 127.5 (CH), 128.2 (CH), 128.5
(CH), 140.6 (C), 147.8 (C), 148.7 (C). One (C) signal not observed.
(15S)-15-Naphthylspiro[fluorene-9,3’-oxirane] (5c)
Buffered bleach (17 mL, 11.9 mmol, pH 11.3) at 0°C was added
dropwise to a solution of 4c (2.42 g, 7.95 mmol), (R,R)-Jacobsen
catalyst (0.31 g, 0.49 mmol) and 4-phenylpyridine N-oxide (0.27 g,
1.59 mmol) in CH2Cl2 (15 mL) at 0°C. After 2 h the mixture was
treated as described for 5a to afford 0.28 g (11%) of (S)-5c as an oil
in 87% ee. A recrystallization from hexanes affords 5c of >99% ee.
Conditions of HPLC analysis: Chiralcel-OD-R column, 25 cm,
30 °C; eluent: MeOH; flow rate: 0.5 mL/min; R-isomer, tR: 21.9
min and S-isomer tR: 30.9 min; mp 173-174°C; [a]D -109.1
(c = 1.03 in CDCl3).
MS (CI, NH3): m/z (%) = 294 (C20H23ON·H+, 100).
Anal. Calcd. For C20H23NO (293.4): C, 81.87; H, 7.90; N, 4.77.
Found: C, 81.81; H, 7.92; N, 4.73.
IR (KBr): n = 3054, 1451, 768 cm-1.
1H NMR (200 MHz, CDCl3): d = 5.31 (s, 1 H), 6.42-6.46 (m, 1 H),
6.72-6.80 (m, 1 H), 7.16-8.00 (m, 13 H).
13C NMR (50 MHz, CDCl3): d = 64.9 (CH), 67.7 (C), 120.0 (CH),
120.3 (CH), 121.7 (CH), 123.1 (CH), 123.4 (CH), 124.6 (CH),
125.1 (CH), 125.9 (CH), 126.4 (CH), 126.9 (CH), 127.6 (CH),
128.4 (CH), 128.6 (CH), 129.0 (CH), 129.4 (CH), 130.8 (C), 131.5
(C), 133.1 (C), 138.6 (C), 140.8 (C), 141.2 (C), 141.4 (C).
Ring-Opening of (15S)-15-Phenylspiro[fluorene-9,3’-oxirane]
(5b) with Piperidine
(i) Reaction at 23°C: A mixture of 5b (405 mg, 1.5 mmol), LiClO4
(1.58 g, 15 mmol) and piperidine (1.5 mL, 15 mmol) in MeCN (3
mL) was stirred at r.t. for 5 d under N2 and treated as described for
5a to afford, after column chromatography, 227 mg of 1-(9-fluore-
nyl)piperidine 79 and 52 mg (10%) of 3b as an oil; [a]D -18.5
(c = 0.59, CDCl3).
MS (CI, NH3): m/z (%) = 321 (C24H16O·H+, 14), 338
(C24H16O·NH4 , 100).
IR (KBr): n = 3480, 2923, 1449, 733 cm-1.
1H NMR (200 MHz, CDCl3): d = 1.30-1.80 (m, 6 H), 2.40-2.90
(m, 4 H), 5.52 (s, 1 H), 6.40-6.60 (m, 2 H), 6.60-6.80 (m, 3 H),
7.20-7.40 (m, 6 H), 7.60-7.80 (m, 2 H).
13C NMR (50 MHz, CDCl3): d = 24.9 (CH2), 27.1 (CH2), 48.8
(CH2), 73.8 (CH), 78.6 (C), 119.2 (CH), 119.7 (CH), 126.0 (CH),
126.1 (CH), 126.2 (CH), 126.2 (CH), 126.3 (CH), 126.8 (CH),
127.0 (CH), 128.0 (CH), 128.4 (CH), 137.9 (C), 140.1 (C), 142.1
(C). Two (C) signals not observed.
+
Ring-Opening of (15S)-15-Methylspiro[fluorene-9,3’-oxirane]
(5a) with Piperidine
(i) Isolation of (S)-3a from a 91:9 Regioisomeric Mixture (S)-3a/
(R)-2a: Epoxide (S)-5a (0.90 g, 4.3 mmol, 47% ee) was added to a
suspension of LiClO4 (6.9 g, 65 mmol) in MeCN (6.5 mL) and the
resulting mixture was stirred at r.t. until a homogeneous solution
formed. Piperidine (6.4 mL, 65 mmol) was added dropwise over 30
min keeping the reaction temperature below 35°C. After 6 h at
25°C, Et2O (50 mL) was added, the ethereal layer was washed with
H2O (2 î 50 mL), dried (Na2SO4), and concentrated under vacu-
um. The residual solid was purified by column chromatography to
afford 1.20 g (95%) of a 91:9 mixture of (S)-3a and (R)-2a. Recrys-
tallization from hexane (3 times) afforded 0.38 g (30%) of regio-
chemically pure (S)-3a of 99% ee. Conditions of HPLC analysis:
Chiralcel-OD column, 25 cm, 30°C; eluent: hexane/propan-2-ol
(94:6); flow rate: 0.5 mL/min; S-isomer, tR: 8.3 min and R-isomer
tR: 9.5 min; mp 136 °C; [a]D -133 (c = 1.0, CDCl3, 99% ee).
IR (KBr): n = 3334, 2996, 1447, 754 cm-1.
1H NMR (200 MHz, CDCl3): d = 0.45 (d, J = 6 Hz, 3 H), 1.30-1.62
(m, 6 H), 2.36-2.78 (m, 4 H), 4.61 (q, J = 6 Hz, 1 H), 7.20-7.37 (m,
5 H), 7.57-7.70 (m, 3 H).
13C NMR (50 MHz, CDCl3): d = 15.9 (CH3), 24.9 (CH2), 27.0
(CH2), 48.5 (CH2), 67.2 (CH), 77.9 (C), 119.5 (CH), 119.8 (CH),
MS (CI, NH3): m/z (%) = 355 (C25H25ON+, 100).
(ii) Reaction at 80°C: A mixture of 5b (400 mg, 1.48 mmol), LiClO4
(1.58 g, 15 mmol) and piperidine (1.5 mL, 15 mmol) in MeCN (3
mL) was heated at 80°C under N2 during 3 h and treated as de-
scribed for 5a to afford, after column chromatography, 203 mg of
1-(9-fluorenyl)piperidine 79 and 55 mg (10%) of 2b as an oil;
[a]D +6.6 (c = 0.34, CDCl3).
IR (KBr): n = 3407, 2930, 1451, 739 cm-1.
1H NMR (200 MHz, CDCl3): d = 1.30-1.45 (m, 2 H), 1.50-1.80
(m, 4 H), 2.20-2.40 (m, 2 H), 2.80-3.00 (m, 2 H), 4.07 (s, 1 H),
6.55-6.62 (m, 2 H), 6.90-7.60 (m, 11 H).
13C NMR (50 MHz, CDCl3): d = 24.2 (CH2), 26.9 (CH2), 54.7
(CH2), 79.5 (CH), 83.0 (C), 119.5 (CH), 119.9 (CH), 124.0 (CH),
124.8 (CH), 127.2 (CH), 127.4 (CH), 128.6 (CH), 128.7 (CH),
129.4 (CH), 133.9 (C), 140.3 (C), 141.1 (C), 147.5 (C), 148.0 (C).
Two (CH) signals not observed.
MS (CI, NH3): m/z (%) = 356 (C25H25ON·H+, 100).
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

FEATURE ARTICLE
Synthesis of a 9-Fluorenone Derived b-Amino Alcohol Ligand
175
(1S)-1-(9-Azidofluoren-9-yl)ethanol [(S)-10a]
1H NMR (500 MHz, -20°C, CDCl3): d = 0.57 (d, J = 6 Hz, 3 H),
1.24-1.42 (m, 6 H), 1.86 (m, 2 H), 2,31 (m, 2 H), 3.41 (s, 3 H), 4.81
(s, 1 H), 5.9 (q, J = 6 Hz, 1 H), 7.25-7.46 (m, 10 H), 7.65 (d, J = 7
Hz, 1 H), 7.69 (d, J = 7 Hz, 2 H).
13C NMR (75 MHz, 25°C, CDCl3): d = 14.8 (CH3), 25.0 (CH2), 26.8
(CH2), 48.4 (CH2), 57.5 (CH3), 71.3 (CH), 77.2 (C), 82.8 (CH),
119.2 (CH), 119.8 (CH), 125.1 (CH), 126.4 (CH), 126.8 (CH),
127.2 (CH), 127.4 (CH), 128.2 (CH), 128.3 (CH), 128.4 (CH),
128.5 (CH), 136.4 (C), 141.3 (C), 141.5 (C), 144.0 (C), 145.7 (C),
170.3 (C).
A solution of 5a (92% ee) (500 mg, 2.40 mmol) in anhyd benzene
(25 mL) was added to a suspension of [Ti(OPr-i)2 (N3)2] (800 mg,
3.2 mmol) in benzene (20 mL) under argon at 65°C. After stirring
for 30 min, the mixture was allowed to cool to r.t. and benzene was
removed under reduced pressure. To the crude product in Et2O (15
mL) was added 5% H2SO4 (20 mL) and the resulting mixture was
stirred until two clear phases formed (ca 1 h). Phases were separat-
ed, the aqueous one was extracted with Et2O (4 î 15 mL) and the
combined ethereal extracts were dried (Na2SO4) and concentrated
under vacuum. The residual solid was purified by column chroma-
tography to afford 550 mg (91%) of azidoalcohol 10a as an oil;
[a]D -52.5 (c = 2.37, CDCl3).
MS (CI, NH3): m/z (%) = 442 (C29H31O3N·H+, 100).
(S)-Methoxyphenylacetic Acid (S)-1-(9-Piperidin-1-yl-9H-fluo-
ren-9-yl)ethyl Ester [(S,S)-12]
IR (KBr): n = 3452, 2103, 1451, 756 cm-1.
IR (neat): n = 2935, 1744, 1109, 737 cm-1.
1H NMR (500 MHz, CDCl3): d = 0.65 (d, J = 10 Hz, 3 H), 2.67 (d,
1H NMR (500 MHz, -20°C, CDCl3): d = 0.62 (d, J = 5 Hz, 3 H),
1.06-1.10 (m, 2 H), 1.20-1.30 (m, 4 H), 1.87-1.90 (m, 2 H), 2.17-
2.20 (m, 2 H), 3.50 (s, 3 H), 4.81 (s, 1 H), 5.97 (q, J = 6 Hz, 1 H),
7.23-7.26 (m, 2 H), 7.33-7.41 (m, 8 H), 7.50 (d, J = 7 Hz, 1 H),
7.59 (d, J = 7 Hz, 1 H), 7.65 (d, J = 7 Hz, 1 H).
13C NMR (75 MHz, 25°C, CDCl3): d = 15.3 (CH3), 24.9 (CH2), 26.5
(CH2), 48.3 (CH2), 57.5 (CH3), 71.0 (CH), 76.9 (C), 83.2 (CH),
119.2 (CH), 119.8 (CH), 125.1 (CH), 126.4 (CH), 126.8 (CH),
127.3 (CH), 127.6 (CH), 128.1 (CH), 128.2 (CH), 128.6 (CH),
136.4 (C), 141.2 (C), 141.6 (C), 143.8 (C), 145.5 (C), 170.3 (C).
J = 4 Hz, 1 H), 4.35 (m, 1 H), 7.34-7.50 (m, 5 H), 7.68-7.76 (m, 3
H).
13C NMR (125 MHz, CDCl3): d = 16.1 (CH3), 72.4 (CH), 77.9 (C),
120.1 (CH), 120.3 (CH), 123.7 (CH), 126.0 (CH), 127.9 (CH),
128.0 (CH), 129.6 (CH), 129.7 (CH), 140.6 (C), 140.9 (C), 141.3
(C), 142.9 (C).
MS (CI, NH3): m/z (%) = 252 (C15H13ON3·H+, 1), 269
+
(C15H13ON3·NH4 , 100).
(1S)-1-(9-Aminofluoren-9-yl)ethanol [(S)-11a]
MS (CI, NH3): m/z (%) = 442 (C29H31O3N·H+, 100).
A solution of 10a (92% ee) (500 mg, 1.99 mmol) in MeOH (20 mL)
was added to a suspension of 10% Pd/C (70 mg, 14%) in MeOH (20
mL) under H2. The suspension was vigorously stirred for 5 h at r.t.,
filtered through a pad of Celite and concentrated under vacuum. The
crude product was purified by column chromatography eluting with
a 19:1 mixture of CHCl3/MeOH to afford 435 mg (97%) of 11a as
a white solid; mp 94°C; [a]D -32.4 (c = 1.4, CDCl3).
IR (KBr): n = 3351, 3160, 1449, 735 cm-1.
1H NMR (200 MHz, CDCl3): d = 0.55 (d, J = 6 Hz, 3 H), 1.80-2.20
(br s, 2 H), 4.23 (q, J = 6 Hz, 1 H), 7.26-7.40 (m, 5 H), 7.65-7.70
(m, 3 H).
Enantioselective Amino Alcohol-Catalyzed Addition of Dieth-
ylzinc to Aldehydes; General Procedure
To a solution of the chiral ligand (0.03-0.06 mmol, 3-6 mol%) in
toluene (2 mL), was added the aldehyde (1 mmol) at r.t. The mixture
was stirred for 20 min and then cooled to 0°C. To this mixture was
added dropwise diethylzinc (2.2 mL of a 1 M hexanes solution, 2.2
mmol). The mixture was stirred for the corresponding reaction time
under N2. The reaction was quenched by the addition of a satd aq
NH4Cl solution (10 mL). The mixture was then extracted with
CH2Cl2 (3 î 10 mL) and the combined organic extracts were dried
and concentrated under vacuum. Conversion, selectivity and enan-
tiomeric purity of the resulting alcohols were determined from the
crude mixture by GC or HPLC analysis. For 9a-d and 9f-s condi-
tions of analysis and retention times of the R- and S-isomers have
been reported elsewhere.2c-d For 1-(3-chlorophenyl)propan-1-ol
(9e): Supelco b-DEXTM 120 column, 30 m length, 0.25 mm inner
diameter, 135°C, He as carrier gas (2.4 mL/min), tR R-isomer 67.1
min, tR S-isomer 60.6 min. For 9a-d and 9f-s the establishment of
the absolute configuration of the major enantiomer resulting from
the reaction has been performed as already reported.2c-d For 1-(3-
chlorophenyl)propan-1-ol (9e) the absolute configuration of the ma-
jor enantiomer resulting from the reaction was assumed to be S.
13C NMR (50 MHz, CDCl3): d = 16.6 (CH3), 68.5 (C), 72.9 (CH),
119.7 (CH), 119.9 (CH), 122.6 (CH), 125.0 (CH), 127.5 (CH),
127.6 (CH), 128.2 (CH), 128.3 (CH), 140.0 (C), 140.1 (C), 147.2
(C), 148.8 (C).
MS (CI, NH3): m/z (%) = 225 (C15H15ON+, 100).
Anal. Calcd for C15H15NO (225.3): C, 79.97; H, 6.71; N, 6.22.
Found: C, 80.00; H, 6.72; N, 6.11.
(1S)-1-(9-Piperidylfluoren-9-yl)ethanol [(S)-3a]
A mixture of amino alcohol 11a (170 mg, 0.76 mmol), 1,5-dibro-
mopentane (350 mg, 1.52 mmol), K2CO3 (210 mg, 1.52 mmol) and
EtOH (2 mL) was heated at reflux for 24 h under N2. The mixture
was allowed to cool to r.t., H2O (10 mL) and CH2Cl2 (15 mL) were
added and phases were separated. The aqueous phase was extracted
with CH2Cl2 (2 î 15 mL), and the combined organic extracts were
dried and concentrated under vacuum. The residual solid was puri-
fied by column chromatography to afford 210 mg (72%) of (S)-3a
as a white solid.
Acknowledgement
Financial support from DGICYT (PB95-0265) and CIRIT
(1998SGR-00005) is gratefully acknowledged. Katamreddy Subba
Reddy thanks MEC for a fellowship under the program: Estancias
Temporales de Científicos y Tecnólogos Extranjeros en España.
References
Diastereomeric O-Methylmandalate Esters of (S)-3a
Diastereomeric O-methylmandelate esters (S,S)-12 and (S,R)-12
were prepared from amino alcohol (S)-3a and the enantiomers of O-
methylmandelic acid by a standard procedure with DCC-DMAP.14
(1) For reviews, see:
(a) Soai, K. Chem. Rev. 1992, 92, 833.
(b) Noyori, R. Asymmetric Catalysis in Organic Synthesis;
Wiley: New York, 1994; pp 255-297.
(R)-Methoxyphenylacetic Acid (S)-1-(9-Piperidin-1-yl-9H-fluo-
ren-9-yl)ethyl Ester [(S,R)-12]
(2) (a) Vidal-Ferran, A.; Moyano, A.; Pericàs, M. A.; Riera, A. J.
Org. Chem. 1997, 62, 4970.
IR (neat): n = 2935, 1744, 1109, 737 cm-1.
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York

176
K. S. Reddy et al.
FEATURE ARTICLE
(b) Vidal-Ferran, A.; Moyano, A.; Pericàs, M. A.; Riera, A.
Tetrahedron Lett. 1997, 38, 8773.
(c) Solà, L.; Reddy, K.S.; Vidal-Ferran, A.; Moyano, A.;
Pericàs, M. A.; Riera, A.; Alvarez-Larena, A.; Piniella, J.-F. J.
Org.Chem. 1998, 63, 7078.
(11) (a) Kitamura, M.; Suga, S.; Oka, H.; Noyori, R. J. Am.Chem.
Soc. 1998, 120, 9800, and references cited therein.
(b) Guillaneux, D.; Zhao, S.; Samuel, O.; Rainford, D.;
Kagan, H. B. J. Am.Chem. Soc. 1994, 116, 9430.
(c) Avalos, M.; Babiano, R.; Cintas, P.; Jiménez, J. L.;
Palacios, J. C. Tetrahedron Asymmetry 1997, 8, 2997.
(d) Girard, C.; Kagan, H. B. Angew. Chem. 1998, 110, 3088;
Angew. Chem. Int. Ed. Engl. 1998, 37, 2922.
(12) Hamada, T.; Fukuda, T.; Imanishi, H.; Katsuki, T.
Tetrahedron 1996, 52, 515.
(13) Caron, M.; Carlier, P.R.; Sharpless, K.B. J. Org. Chem. 1988,
53, 5185.
(14) Hassner, A.; Alexanian, V. Tetrahedron Lett. 1978, 4475.
(15) (a) Trost, B. M.; Belletire, J. L.; Godleski, S.; McDougal, P.
G.; Balkovec, J. M. J. Org. Chem. 1986, 51, 2370.
(b) Latypov, S. K.; Seco, J. M.; Quiñoá, E.; Riguera, R. J.
Am.Chem. Soc. 1998, 120, 877.
(16) For a related study involving the use of a diastereomeric
mixture of ligands, see:
Bolm, C.; Muñiz, K.; Hildebrand, J. P. Org. Lett. 1999, 1, 491.
(17) Neuenschwander, M.; Vögeli, R.; Fahrni, H. P.; Lehmann, H.;
Ruder, J. P. Helv. Chim. Acta 1977, 60, 1073.
(d) Reddy, K.S.; Solà, L.; Moyano, A.; Pericàs, M. A.; Riera,
A. J. Org. Chem. 1999, 64, 3969.
(3) Jimeno, C.; Vidal-Ferran, A.; Moyano, A.; Pericàs, M. A.;
Riera, A. Tetrahedron Lett. 1999, 40, 777.
(4) Vidal-Ferran, A.; Bampos, N.; Moyano, A.; Pericàs, M. A.;
Riera, A.; Sanders, J. K. M. J. Org. Chem. 1998, 63, 6309.
(5) (a) Yamakawa, M.; Noyori, R. J. Am.Chem. Soc. 1995, 117,
6327.
(b) Goldfuss, B.; Houk, K. N. J. Org. Chem. 1998, 63, 8998.
(6) (a) Jacobsen, E. N. In Catalytic Asymmetric Synthesis; Ojima,
I., Ed.; VCH Publishers, Inc.: New York, 1993; pp 159.
(b) Brandes, B. D.; Jacobsen, E. N. J. Org. Chem. 1994, 59,
4378-4380.
(7) If desired, the enantiomeric purity of 5a can be readily
upgraded to 95% ee on the basis of the preferential
crystallization of the racemic from hexanes, but at the
expenses of substantial yield losses.
(8) Chini, M.; Crotti, P.; Flippin, L. A.; Gardelli, C.; Giovani, E.;
Macchia, F.; Pineschi, M. J. Org. Chem. 1993, 58, 1221.
(9) Bordwell, F. G.; Cheng, J-P.; Seyedrezai, S. E.; Wilson, C. A.
J. Am. Chem. Soc. 1988, 110, 8178.
Article Identifier:
1437-210X,E;2000,0,01,0165,0176,ftx,en;M02199SS.pdf
(10) The configuration of 1-phenylpropan-1-ol (9a) for ligands 3a,
2b and 3b is in agreement with Noyori’s empirical rule: see
Ref 1b, pp 265-266.
Synthesis 2000, No. 1, 165–176 ISSN 0039-7881 © Thieme Stuttgart · New York