Hydrosilylation of cinchonidine and 9-O-TMS-cinchonidine with triethoxysilane: Application of 11-(triethoxysilyl)-10,11-dihydrocinchonidine as a chiral modifier in the enantioselective hydrogenation of 1-phenylpropane-1,2-dione

Article abstract of DOI:10.1039/b205032c

The detailed synthesis and characterization of (-)-11-(triethoxysilyl)-10,11-dihydrocinchonidine (4), a starting material for the immobilization of (-)-cinchonidine on silica based supports, is described. Compound 4 together with its precursors 9-O-(trime

Full text of DOI:10.1039/b205032c

View Article Online / Journal Homepage / Table of Contents for this issue
Hydrosilylation of cinchonidine and 9-O-TMS-cinchonidine
with triethoxysilane: application of 11-(triethoxysilyl)-10,11-
dihydrocinchonidine as a chiral modifier in the enantioselective
hydrogenation of 1-phenylpropane-1,2-dione
1
Anna Lindholm,a Päivi Mäki-Arvela,b,c Esa Toukoniitty,b,c Tapani A. Pakkanen,d
Janne T. Hirvi,d Tapio Salmi,b,c Dmitry Yu. Murzin,b,c Rainer Sjöholma,c and Reko Leino*a
a Department of Organic Chemistry, Åbo Akademi University, FIN-20500 Åbo, Finland.
E-mail: reko.leino@abo.fi
b Laboratory of Industrial Chemistry, Åbo Akademi University, FIN-20500 Åbo, Finland
c Process Chemistry Group, Åbo Akademi University, FIN-20500 Åbo, Finland
d Department of Chemistry, University of Joensuu, P.O. Box 111, FIN-80101 Joensuu, Finland
Received (in Cambridge, UK) 24th May 2002, Accepted 9th October 2002
First published as an Advance Article on the web 7th November 2002
The detailed synthesis and characterization of (Ϫ)-11-(triethoxysilyl)-10,11-dihydrocinchonidine (4), a starting
material for the immobilization of (Ϫ)-cinchonidine on silica based supports, is described. Compound 4 together
with its precursors 9-O-(trimethylsilyl)cinchonidine (2) and 9-O-(trimethylsilyl)-11-(triethoxysilyl)-10,11-dihydro-
cinchonidine (3) were employed as chiral modifiers in the hydrogenation of a prochiral diketone, 1-phenylpropane-
1,2-dione, over a heterogeneous Pt/Al2O3 catalyst using cinchonidine (1) as a reference modifier. The unexpected
enhancement of ee induced by 4, demonstrating the positive effect of distal modifier substitution, is discussed in
the light of molecular modeling and NMR studies.
Introduction
Asymmetric hydrogenation of α-keto esters,1 keto acetals,2 and
diketones3 using cinchona-modified platinum catalysts is one
of the few heterogeneous catalyst systems rivaling the enantio-
meric excesses (ee) obtained with homogeneous hydrogenation
catalysts. In recent years, increasing amounts of papers on this
subject have emerged from both academic and industrial
laboratories. The highest enantioselectivities, approaching 95%
ee, have been obtained in the hydrogenation of methyl and ethyl
pyruvate to the corresponding lactates with cinchonidine†
or 10,11-dihydrocinhonidine modified Pt/Al2O3 catalysts.1 The
Fig. 1 Atomic numbering and structure of (Ϫ)-cinchonidine.
origin of the asymmetric induction, as well as the effect
of structural variations of either the substrate or the modifier
of a flow reactor system for continuous enantioselective hydro-
on enantioselectivity, remain unclear. Certain qualitative
genation of prochiral ketones. We were interested in preparing
trends have been summarized in a recent paper by Blaser and
a cinchonidine derivative suitable for this purpose. Hydro-
silylation of the olefinic double bond of the related cinchona
alkaloid quinine and grafting of the corresponding trialkoxy-
coworkers:1q An extended aromatic system is required to form
an adsorption complex with the Pt surface; in addition, a
chiral amino function capable of interacting with the keto
silyl derivatives to silica for use as chiral HPLC stationary
group of the adsorbed substrate is needed to induce sufficient
phases has been reported previously.4 However, to our know-
enantiocontrol. The sense of asymmetric induction is con-
ledge, the synthesis and characterization of the products and
trolled by the absolute configuration at the C8 and C9 carbon
intermediates have not been described in detail. Furthermore,
cinchonidine/cinchonine as well as quinine/quinidine are pairs
of diastereomers differing markedly in their physical properties,
including their solubilities. General procedures for synthetic
transformations and subsequent purifications cannot be applied
and must be developed separately in each case. Secondly, the
magnitude of ee in the hydrogenation of ethyl pyruvate has
been proposed to be influenced by the conformation of the
atoms of the cinchona modifier (Fig. 1). Substitution of the
C9 hydroxy group with bulky substituents significantly lowers
the enantioselectivity, whereas structural modifications of
the quinuclidine C3 substituent were reported to have only a
moderate effect on the ee in the hydrogenation of ethyl
pyruvate. Notably, enantioselectivities are also significantly
influenced by solvent type and the modifier concentration.
The objective of the present investigation was two-fold. First,
adsorbed modifier, specifically the rotation around the C8–C9
grafting of the cinchona modifier to a silica support through
axis.1h,j,k,q We were thus interested in studying the effect of steric-
covalent bonds would, besides improving the separation and
ally demanding trialkoxysilyl substitution in the cinchonidine
reuse of the chiral modifier,1g potentially allow the construction
quinuclidine moiety on the enantioselectivity of the hydrogen-
ation of prochiral ketones.
While this manuscript was in preparation, a report appeared
describing the hydrosilylation of various cinchona alkaloids,
† The IUPAC name for cinchonidine is α-quinolin-4-yl-5-vinylquin-
uclidine-2-methanol.
DOI: 10.1039/b205032c
J. Chem. Soc., Perkin Trans. 1, 2002, 2605–2612
This journal is © The Royal Society of Chemistry 2002
2605

View Article Online
including cinchonidine, with triethoxysilane in the presence of
the Speier catalyst and the subsequent “one-pot” immobiliz-
ation on a silica surface by sol–gel technology.5 However,
reaction intermediates were not characterized and identific-
ation of the immobilized compounds was based exclusively on
a combination of IR, UV and solid state 13C NMR methods.
This prompted us to report our independent studies on the
hydrosilylation of (Ϫ)-cinchonidine (1) and 9-O-(trimethyl-
silyl)cinchonidine (2) with triethoxysilane in the presence of
hydrogen hexachloroplatinate (Speier catalyst) or platinum
divinyltetramethyldisiloxane (Karstedt catalyst). We describe
here the detailed synthesis and characterization of 11-(trieth-
oxysilyl)-10,11-dihydrocinchonidine (4), a starting compound
for the immobilization of cinchonidine on silica based carriers,
via its precursors 2 and 9-O-(trimethylsilyl)-11-(triethoxysilyl)-
10,11-dihydrocinchonidine (3). All compounds 1–4 were fully
characterized by 1H and 13C NMR spectroscopy and employed
as chiral modifiers in the hydrogenation of 1-phenylpropane-
1,2-dione over
a
heterogeneous Pt/Al2O3 catalyst.6 The
unexpected enhancement of the ee induced by 4, demonstrating
the positive effect of distal modifier substitution, is discussed
in the light of molecular modeling and NMR studies.
Results and discussion
Synthesis and characterization of the chiral modifiers
Direct hydrosilylation of cinchonidine with triethoxysilane in
the presence of the Speier catalyst proved unsuccesful in our
hands. Tertykh and coworkers reported the corresponding
hydrosilylation reaction using a propan-2-ol solution of the
Speier catalyst and equimolar amounts of cinchonidine and
triethoxysilane in a 10 : 0.7 propan-2-ol–acetic acid solvent
mixture.5 However, the product was neither isolated nor char-
acterized and the yield of the hydrosilylation was not reported.
During the initial stages of our investigation we carried out a
Scheme 1 Synthesis of (Ϫ)-11-(triethoxysilyl)-10,11-dihydrocinchon-
idine (4).
mately 90% yield. Further purification was unsuccesful and the
product was used as such in the following step. The 9 : 1 mixture
of 2 and 6 was dissolved in toluene and treated with trieth-
oxysilane at 40 ЊC in the presence of the Karstedt catalyst
and stirred for 3.5 hours at 80 ЊC. According to GC analysis
the hydrosilylation proceeded in approximately 40% yield
under these conditions. Purification by flash chromatography
(methanol–chloroform 1 : 9) gave 9-O-(trimethylsilyl)-11-
(triethoxysilyl)-10,11-dihydrocinchonidine (3) of fair (approx-
imately 75%) purity. Further purification was unsuccesful.
However, after refluxing the obtained fairly pure 3 in methanol
for 20 hours and subsequent washing with pentane, analytically
pure 4 was isolated as an off-white solid in 35% overall yield
based on 2.‡
1
similar reaction in the absence of acetic acid. H NMR and
TLC analyses indicated a complex mixture of different com-
pounds of which the triisopropoxysilyl adduct 5 (Fig. 2) was
NMR Analysis
All compounds 1–5 were fully analyzed by high resolution
1H and 13C NMR spectroscopy. The signal assignments were
based on data obtained using H–H and C–H correlation
spectroscopy (COSY, NOESY, HETCOR, HMBC). In the
following, arguments for the assignments of the signals in the
spectra of cinchonidine (1) are presented. The interpretation of
the spectra of compounds 2–5 was then based on similar
arguments.
Fig. 2 11-(Triisopropoxysilyl)-10,11-dihydrocinchonidine (5).
1
In the H NMR spectrum of cinchonidine (1) the signals of
isolated in 5.3% yield after flash chromatographic (FC) purifi-
cation and identified by HRMS, 1H and 13C NMR analyses (for
analytical data, see Experimental section). Traces of other
transetherification products were detected as well. Based on the
earlier reports on hydrosilylation of quinine derivatives4a,4d we
decided to protect the cinchonidine hydroxy group prior to
reaction with triethoxysilane.
The synthesis of 11-(triethoxysilyl)-10,11-dihydrocinchon-
idine (4) is presented in Scheme 1. 9-O-(Trimethylsilyl)cinch-
onidine (2) was prepared by reacting cinchonidine with TMSCl
and triethylamine in THF. Fairly pure 2 containing approxi-
mately 10% of 9-O-(trimethylsilyl)-10,11-dihydrocinchonidine
(6) resulting from the 10,11-dihydrocinchonidine impurity in
commercial cinchonidine was isolated as a solid in approxi-
H-2Ј and H-9 were easily identified based on their chemical
shifts and they were used as starting points for the inter-
pretation of the 2D spectral data. Centered at δ = 1.70 ppm a
three-proton signal (H-4, 5 and 7) and at δ = 1.36 ppm a two-
proton signal (H-5 and H-7) were observed. The upfield one of
the two signals due to H-7a and H-7b at 1.71 and 1.36 ppm,
respectively, was assigned to H-7a based on its strong NOE
interaction with H-8. The H-5 signal at 1.36 ppm was assigned
to H-5a based on its NOE interaction with H-3. Of the two H-6
signals at δ = 3.43 and 2.45 ppm, respectively, the upfield signal
was assigned to H-6a based on the NOE interaction with H-5a.
‡ Deprotection of 3 using tetrabutylammonium fluoride in THF also
cleaves the silyl ethoxy groups.
2606
J. Chem. Soc., Perkin Trans. 1, 2002, 2605–2612

View Article Online
Scheme 2 Hydrogenation of 1-phenylpropane-1,2-dione.
The signal at δ = 2.92 ppm assigned to H-2a showed a strong
NOE interaction with H-3. The 13C NMR signals were narrow
and well resolved and the assignments were made based
on the 1H NMR signal assignments using HETCOR and
HMBC techniques. The assignments matched those previously
published,7 except for C-3Ј and C-5Ј the shifts of which had to
be interchanged. The spectra of compounds 2–5 were assigned
in a similar manner and were found to be consistent with
1
the proposed structures. Broadening of some of the H and
13C signals (for details, see Experimental section) indicated
hindered rotation around the C-4Ј–C-9 as well as the C-8–C-9
bonds in compounds 2 and 3.
Enantioselective hydrogenation
Fig. 3 Hydrogenation kinetics of 1-phenylpropane-1,2-dione in ethyl
acetate at 25 ЊC. Catalyst: 5 wt% Pt/Al2O3 modified in situ with (Ϫ)-11-
(triethoxysilyl)-10,11-dihydrocinchonidine (4). Symbols: (ꢀ) 1-phenyl-
propane-1,2-dione (7), (᭹) 1-hydroxy-1-phenylpropan-2-one (8a ϩ 8b),
(×) 2-hydroxy-1-phenylpropan-1-one (9a ϩ 9b), (ᮀ) 1-phenylpropane-
1,2-diols (10a ϩ 10b ϩ 11a ϩ 11b).
(R)-(Ϫ)-1-Hydroxy-1-phenylpropan-2-one (phenylacetylcarbi-
nol) is an important intermediate for the synthesis of (1R,2S)-
(Ϫ)-ephedrine and (1S,2S)-(ϩ)-ephedrine, which are major
ingredients of several pharmaceuticals used as anti-asthmatics,
vasoconstricting agents and nasal decongestants.8 Chiral α-
hydroxyketones have also been utilized as building blocks for
the synthesis of other biologically active compounds including
antifungals against AIDS related diseases and complications.9
1-Phenylpropane-1,2-dione (7) can yield enantiomer mixtures
of 1- as well as 2-hydroxyketones upon hydrogenation. Both
regioisomers exist as pairs of enantiomers. The hydroxyketones
may react further to produce the corresponding diols as two
pairs of enantiomers. As shown previously,6 the predominant
product obtained in the hydrogenation of 7 with (Ϫ)-
cinchonidine modified Pt catalysts is the (R)-(Ϫ)-1-hydroxy-1-
phenylpropan-2-one isomer (8a) depicted in Scheme 2, although
variable amounts of other products may be observed as well,
depending on the employed reaction parameters. In the present
study, compounds 1–4 were used as chiral modifiers in the
hydrogenation of 1-phenylpropane-1,2-dione in ethyl acetate
using 5% Pt/Al2O3 as catalyst (Scheme 2). All reactions were
carried out at 25 ЊC and a hydrogen pressure of 5 bar. Enantio-
meric excesses were determined by chiral GC according to the
method described earlier.6
Table 1 Pt/Al2O3 catalyzed hydrogenation of 1-phenylpropane-1,2-
dione in the presence of chiral modifiers 1–4a
Modifier
Conversionb (%)
1
2
3
4
87
76
73
70
a In ethyl acetate at 25 ЊC and 5 bar H2. b Conversion after 20 min
(complete conversion achieved in all experiments).
adsorption for the parent modifier and its derivatives. At the
same time, the modifier structure had a profound effect on the
selectivity. In the case of (Ϫ)-cinchonidine (1) the major
product was 1-hydroxy-1-phenylpropan-2-one (8a/8b). The
ratio between 1-hydroxy-1-phenylpropan-2-one (8a/8b) and
2-hydroxy-1-phenylpropan-1-one (9a/9b) defined as regioselec-
tivity (rs) was approximately 7 : 1, remaining constant with
increasing conversion of the dione (Fig. 4).
Regioselectivities depicted in Fig. 4 show a clear preference
for reduction at the carbonyl group next to the phenyl substi-
tuent for all three modifiers 1, 3 and 4. In the present study,
compound 2 afforded a very low regioselectivity, which in fact
was lower than those obtained in the absence of the modifier,6
whereas the use of modifier 4 resulted in a considerably higher
regioselectivity. Notably, inverse regioselectivity was recently
observed in the asymmetric transfer hydrogenation of 1-phenyl-
propane-1,2-dione using a chiral Ru() catalyst in combination
with formic acid as the hydrogen source.10 Regioselectivities as
high as 9 : 1 resulting in the preferred reduction at the carbonyl
A typical kinetic curve is presented in Fig. 3. The first hydro-
genation step is rapid, whereas further hydrogenation to form
diols is somewhat slower. The rate of 1-phenylpropane-1,2-
dione consumption was observed to be slightly dependent
on the modifier structure (Table 1). Notably, when 1-phenyl-
propane-1,2-dione6 was hydrogenated in the presence and
absence of (Ϫ)-cinchonidine (1) the initial hydrogenation rates
were nearly equal, whereas in the case of butane-2,3-dione3b
a
four-fold rate enhancement in the presence of the modifier was
reported, while for ethyl pyruvate this difference could be
up to two orders of magnitude. Such minor differences in the
reaction rate are thought to be indicative of similar modes of
J. Chem. Soc., Perkin Trans. 1, 2002, 2605–2612
2607

View Article Online
corresponding dihydro compound during the initial stages of
the hydrogenation reaction, at least at higher pressures.6c
Possible differences between the two have thus been neglected
in the context of this work. Data on the enantiomeric excess
obtained with compound 2 are in accordance with previous
findings where C-9 substituents bulkier than methoxy or acetyl
significantly decreased the enantioselectivities in the ethyl
pyruvate case.1q Although the modifier 3 also contains the bulky
C-9 substituent, both regioselectivity and enantioselectivity
remained rather high. A possible explanation relates to the
large quantity of remaining unknown impurities (approx. 25%)
in the employed modifier. Thus, the results obtained here
using compound 3 cannot be taken as definitive evidence of
substituent effects.
Results obtained in the present study indicate not only that
high enantioselectivity requires some modifier–reactant inter-
action, but that such interactions also play a decisive role in
achieving preferential hydrogenation of a particular carbonyl
group, i.e., regioselectivity. In the present system, enantio-
selectivity and regioselectivity are interrelated exhibiting
similar dependence on modifier concentration. The connection
between the two is clearly observable in Figs. 4 and 5. The
modifiers inducing high enantiomeric excess also resulted in
a high regioselectivity (and vice versa), thus also supporting
the involvement of substrate–modifier interactions in the
enhancement of regioselectivity. Further efforts are required to
understand the nature of these interactions. The contribution
of electronic and steric influences of the substrate to the
selectivity will be the subject of forthcoming work.
Fig. 4 Regioselectivity in the Pt/Al2O3 catalyzed hydrogenation of 1-
phenylpropane-1,2-dione in the presence of the chiral modifiers 1–4 in
ethyl acetate at 25 ЊC and 5 bar H2. Symbols: (×) 4, (᭹) 3, (ꢀ) 1, (᭿) (2).
next to the methyl group were reported. In the present case, the
data obtained is consistent with previous observations concern-
ing the origin of regioselectivity. The selectivity enhancement
observed in the presence of a catalyst modifier is a result of
substrate–modifier interactions taking place on the catalyst
surface. In the absence of modifier a lower regioselectivity
was observed, nevertheless, hydrogenation of the benzoylic
position was clearly preferred. Maximally a larger than two-
fold enhancement of regioselectivity could be observed in the
presence of a modifier.
Enantioselectivity was, likewise, observed to depend on the
modifier structure. The enantiomeric excesses (ee) of (R)-1-
hydroxy-1-phenylpropan-2-one (8a) vs. (S)-1-hydroxy-1-phen-
ylpropan-2-one (8b) are reported in Fig. 5 as a function of
Molecular modeling
The experimental activity data can be interpreted using molec-
ular modeling techniques. Catalytic mechanisms have been
proposed, which rely on different cinchonidine conformers. The
lowest energy conformers of a cinchonidine derivative and their
possible interconversion reactions were used to assess their
availability for catalytic reactions. The optimum structures of
cinchonidine (1) have been the subjects of several modeling
studies. Schürch et al.12 and Margitfalvi and Tfirst13 reported
potential energy maps based on molecular mechanics. The
geometry variables in the maps are the dihedral angles between
the quinoline and quinuclidine moieties of cinchonidine. The
minimum energy conformations have been subjected to full
optimization studies with Hartree–Fock and density functional
methods by Bürgi and Baiker.14 They have also reported correc-
tions to obtain relative Gibbs energies. Molecular mechanics
maps are useful for locating minimum energy regions, but the
method fails to give the same energy order for the conformers
as found in the ab initio study by Bürgi and Baiker.14 The
minimum level of ab initio theory that gives the correct order
appears to be HF 6-31G**.
In the study of the relative conformational energies of
modified cinchonidine derivatives, we have calculated potential
energy maps for compounds 1,2 and 4. We chose HF/3-21G as
the level of theory to be used, since molecular mechanics may
not give an accurate description of the energy surface. The
maps were generated by varying the dihedral angles by fixed
increments while optimizing all other variables. The structures
corresponding to the minima found in the maps were subjected
to full optimizations.
Fig. 5 Enantiomeric excess (ee) in the hydrogenation of 1-phenyl-
propane-1,2-dione in the presence of the chiral modifiers 1–4 in ethyl
acetate at 25 ЊC and 5 bar H2. Symbols: (×) 4, (ꢀ) 1, (᭹) 3, (᭿) (2).
the reaction time. The parent modifier (Ϫ)-cinchonidine (1)
produced the major product R-1-hydroxy-1-phenylpropan-
2-one (8a) in an ee of 56% under the employed reaction
conditions. Surprisingly, the enantiomeric excess increased to
70% when the triethoxysilyl dihydroderivative 4 was employed
instead of 1. In previous studies, changes in the quinuclidine
C-3 substituent did not result in significant improvements in
enantioselectivity in the hydrogenation of ethyl pyruvate.1q The
lowest reaction rate and only marginal enantiodifferentiation
(ee = 6%) were observed in the present study with the 9-O-TMS-
substituted modifier 2. As already stated (vide supra), the
regioselectivity towards 1-hydroxy-1-phenylpropan-2-one was
also the lowest with this modifier.
In compound 2 the remaining impurity (approx. 10%)
consists of 9-O-(trimethylsilyl)-10,11-dihydrocinchonidine (6).
In previous studies, hydrogenation of the cinchona alkaloid
10,11-double bond did not significantly influence the enantio-
differentiation.1d,q,11 Commercial (Ϫ)-cinchonidine, commonly
used as such, contains up to 10% of the corresponding 10,11-
dihydro analogue. It is also well known that the 10,11-double
bond of the cinchona modifier is often hydrogenated to the
The HF/3-21G map for cinchonidine (1) is presented in Fig. 6
with ball and stick representations of selected minimum energy
conformations depicted in Fig. 7. The optimizations were also
performed at the HF/6-31G** level starting from HF/3-21G
optimizations. The locations of the minima labeled as Open(n)
or Closed(n) are indicated on the map. The original numbering
scheme of Bürgi and Baiker14 was retained to facilitate direct
comparison of results. The corresponding energies are reported
in Fig. 11. The results for the four lowest conformers agree very
2608
J. Chem. Soc., Perkin Trans. 1, 2002, 2605–2612

View Article Online
Fig. 6 The relative total energy map for 1. The coordinate axes are the
dihedral angles (3Ј–4Ј–9–8) and (4Ј–9–8–N). The energy of the lowest
minimum has been set to zero.
Fig. 8 The relative total energy map for 2. The coordinate axes are the
same as in Fig. 6. The energy of the lowest minimum has been set to
zero.
Fig.
7
Stick representations of the selected minimum energy
conformations of 1.
well with the structures and energies reported earlier.14 Three
new minima were located, Closed(7), Open(8) and Open(9). For
the higher conformers Open(5) and Open(6) we find somewhat
different geometries with lower total energies than in the earlier
study.14 The map also indicates possible pathways for inter-
conversions from one conformation to another. From the
lowest energy conformation Open(3) the next lowest conform-
ation, Closed(1), can be reached via two routes. Both paths,
Open(3)–Open(8)–Open(4)–Closed(1) and Open(3)–Closed(2)–
Closed(7)–Closed(1) have an approximate saddle point energy
of 35 kJ mol1 at the 3-21G level.
Fig. 9 Stick representations of the Open(3) and Closed(1) conform-
ations of 2, 3 and 4.
cinchonidine. A significant difference is, however, that the
Closed(1) conformation has practically the same energy as
the Open(3) (for ball and stick representations, see Fig. 9).
The transformation from Open(3) is more difficult than in 1,
since the Open(3)–Open(4)–Closed(1) path is closed and the
The total energy map for compound 2 is presented in Fig. 8.
The overall outlook of the map is similar to the unsubstituted
J. Chem. Soc., Perkin Trans. 1, 2002, 2605–2612
2609

View Article Online
Open(3)–Closed(2)–Closed(7)–Closed(1) path has an approx-
imate saddle point energy of 44 kJ mol1. One can conclude that
compound 2 can be found in two conformations, but with no
easy interconversion between them.
expected, the conformations behave very similarly to those
of 2. The lowest energy conformation is Closed(1) with Open(3)
marginally higher (Fig. 9). The total energy surface can also be
expected to be similar to the map of compound 2, with a high
interconversion barrier separating the two lowest energy
conformers.
The total energy map for compound 4 is presented in Fig. 10.
The outlook is virtually identical to that of the map of
compound 1. The map demonstrates that the structural differ-
ence in the quinuclidine region does not influence the con-
formational energetics of the cinchonidine frame. The lowest
energy conformers Open(3), Closed(1) and Closed(2) have very
similar energy differences in compounds 1 and 4 also at the
6-31G** level of theory, as seen in Fig. 11. The conclusions
for compound 1 of the transformation from conformation
Open(3) to Closed(1) apply to compound 4 as well. Approx-
imate saddle point energy at the 3-21G level was found to be
34 kJ mol1.
Summary and conclusions
The detailed synthesis and characterization of (Ϫ)-11-(trieth-
oxysilyl)-10,11-dihydrocinchonidine (4) has been reported. This
compound together with its precursors 9-O-(trimethylsilyl)-
cinchonidine (2) and 9-O-(trimethylsilyl)-11-(triethoxysilyl)-
10,11-dihydrocinchonidine (3) were employed as chiral
modifiers in the hydrogenation of
a prochiral diketone
1-phenylpropane-1,2-dione over heterogeneous Pt/Al2O3
a
Compound 3 was the largest molecule to be modeled, since
it has two big substituents. The results of the conformation
analyses of compounds 1 and 4 showed that the substituent in
the quinuclidine region did not have much influence on the
energetics. For compound 3 the 3-21G map was not generated,
instead the minimum energy dihedral geometries of compound
1 were used as the starting points for full geometry optimis-
ations at the 6-31G** level. The results are shown in Fig. 11. As
catalyst. (Ϫ)-Cinchonidine (1) was used as a reference modifier.
The hydrogenation rates showed only slight dependence on the
modifier structure. The carbonyl group in the position closer
to the phenyl ring was preferentially reduced, forming mainly
(R)-1-hydroxy-1-phenylpropanone. Compound 4 induced an
unexpected enhancement in regioselectivity and ee thus
demonstrating the positive effect of distal modifier substi-
tution. In contrast, considerably lower values of regioselectivity
and ee were obtained with 9-O-(trimethylsilyl)cinchonidine (2).
As a summary of the cinchonidine derivative modeling one
can conclude that potential energy maps at the HF/3-21G level
and the refined geometry optimizations at the HF/6-31G**
level give a reliable qualitative picture of the energetics of the
lowest energy conformers. The O-TMS substituent stabilizes
another conformation of the cinchonidine frame. The inter-
conversion between the lowest energy conformers becomes at
the same time less likely, possibly contributing to the low enan-
tioselectivities and poor overall performance of compound 2 as
a chiral catalyst modifier. Additionally, the unfavorable con-
formation together with the bulky 9-O-TMS-substituent may
hinder the effective adsorption of this modifier on the platinum
surface via the π-system of the quinoline moiety in the preferred
flat orientation.15 Substituents at the quinuclidine moiety have
only a minor influence on the relative energies of the con-
formers. Thus assuming that compounds 4 and 1 adsorb on the
platinum surface in a similar Open(3)-type π-bonded15 lowest
energy conformation, the higher enantio- and regioselectivities
obtained with compound 4 are likely to result from its more
favorable interaction or complexation with the substrate
diketone.
Fig. 10 The relative total energy map for 4. The coordinate axes are
the same as in Fig. 6. The energy of the lowest minimum has been set to
zero.
Fig. 11 The relative total energies of compounds 1–4. For each compound the energy of the lowest minimum has been set to zero.
J. Chem. Soc., Perkin Trans. 1, 2002, 2605–2612
2610

View Article Online
(ur, 1H, H-7b), 2.57 (ddd, 1H, J = 14.3 Hz, 11.0 Hz, 5.2 Hz,
Experimental
H-6a), 2.12 (m, 1H, H-3), 2.49 (dm, 1H, J = 13.1 Hz, H-2b),
1.68 (m, 1H, H-4), 1.60 (m, 1H, H-5b), 1.38 (m, 1H, H-5a), 1.34
(m, 1H, H-7a), Ϫ0.07 (s, 9H, Si-CH3). 13C NMR (CDCl3, δ):
148.89 (C-2Ј), 147.96 (C-8aЈ), 147.40 (C-4Ј), 140.97 (C-10),
129.45 (C-8Ј), 127.82 (C-7Ј), 125.45 (C-6Ј), 124.45 (C-4aЈ),
122.01 (br, C-5Ј), 117.63 (br, C-3Ј), 113.05 (C-11), 72.60 (br,
C-9), 60.50 (C-8), 56.35 (C-2), 41.98 (br, C-6), 39.08 (C-3), 26.97
(C-4), 26.54 (C-5), 20.37 (br, C-7), Ϫ0.85 (3C, Si-CH3).
General
All reactions with air-sensitive reagents were carried out under
argon atmosphere using standard Schlenk, vacuum or glove
box techniques. THF was dried and distilled under argon from
sodium–benzophenone prior to use. Toluene was dried over 4 Å
molecular sieves. (Ϫ)-Cinchonidine (Aldrich, 96%), triethoxy-
silane (Gelest) and platinum–divinyltetramethyldisiloxane
complex (Karstedt catalyst, 2.1–2.4% Pt concentration in
xylene, Gelest) were used as received. Melting points were
determined in open glass capillaries and are uncorrected.
Electron impact high resolution mass spectra (HRMS) were
obtained with a Fisons ZabSpec mass spectrometer at 70 eV.
NMR spectra were recorded at 303 K in CDCl3 (ca. 0.15 M
solutions) using JEOL JNM-L 400 or JNM-A 500 NMR spec-
trometers and referenced against tetramethylsilane (TMS). The
chemical shifts are expressed in ppm downfield from TMS.
Signal multiplicities and coupling constants are given in paren-
theses [br = broad unresolved multiplet (1H NMR) or signal
broadening (13C NMR); ur = unresolved multiplet without
broadening (1H NMR)]. Polarimetric measurements were
carried out using a Perkin-Elmer 241 Polarimeter with a cell
volume of 1 mL and a cell length of 10 cm. Optical rotations
are given in units of 10Ϫ1 deg cm2 molϪ1. Microanalysis was
conducted at the Department of Microanalytics, University of
Groningen, the Netherlands.
Synthesis of 9-O-(trimethylsilyl)-11-(triethoxysilyl)-10,11-
dihydrocinchonidine (3)
To a solution of the 2–6 mixture (1.0 g from the previous
reaction, approx. 2.4 mmol of 2) in toluene (3 mL) was added
the Karstedt catalyst (0.2 mL of a 2.1–2.4% Pt solution in
xylene) and triethoxysilane (0.5 mL, 3.0 mmol) at 40 ЊC. The
reaction mixture was stirred for 3.5 hours at 80 ЊC. Purification
by flash chromatography (methanol–chloroform 1 : 9) gave 0.79
1
g of fairly pure (approx. 75% by H NMR) 3 as a yellowish
amorphous material. HRMS (calcd/found): 530.2996/530.2996.
1H NMR (CDCl3, δ): 8.79 (d, 1H, J = 4.6 Hz, H-2Ј), 8.06 (dd,
1H, J = 8.5 Hz, 1.4 Hz, H-8Ј), 8.05 (br, 1H, H-5Ј), 7.63 (ddd,
1H, J = 7.7 Hz, 6.9 Hz, 1.4 Hz, H-7Ј), 7.49 (ddd, 1H, J = 7.7 Hz,
6.7 Hz, 1.2 Hz, H-6Ј), 7.42 (br, 1H, H-3Ј), 5.60 (br, 1H, H-9),
3.64 (q, 6H, J = 7.0 Hz, SiOCH2CH3), 3.32 (ur, 1H, H-6b), 2.98
(dd, 1H, J = 13.6 Hz, 10.1 Hz, H-2a), 2.91 (br, 1H, H-8), 2.58
(ddd, 1H, J = 13.0 Hz, 10.8 Hz, 4.6 Hz, H-6a), 2.26 (ur, 1H,
H-2b), 1.73–1.61 (ur, 3H, H-4, H-5b, H-7b), 1.36 (ur, 1H, H-3),
1.34 (m, 1H, H-10), 1.22–1.17 (ur, 2H, H-5a, H-7a), 1.04 (t,
9H, J = 7.0 Hz, SiOCH2CH3), 0.42 (m, 1H, H-11), Ϫ0.04 (s, 9H,
Si-CH3). 13C NMR (CDCl3, δ): 148.87 (C-2Ј), 148.37 (C-4Ј or
C-8aЈ), 147.36 (C-4Ј or C-8aЈ), 129.36 (C-8Ј), 127.92 (C-7Ј),
125.32 (C-6Ј), 124.43 (C-4aЈ), 122.04 (br, C-5Ј), 117.63 (br,
C-3Ј), 72.30 (br, C-9), 60.20 (C-8), 57.72 (C-2), 57.27 (3C,
SiOCH2CH3), 42.15 (br, C-6), 38.40 (C-3), 27.33 (C-5), 26.54
(C-10), 24.41 (C-4), 20.00 (br, C-7), 17.17 (3C, SiOCH2CH3),
7.20 (C-11), 0.19 (3C, Si-CH3).
NMR data of (؊)-cinchonidine (1)
1H NMR (CDCl3, δ): 8.61 (d, 1H, J = 4.9 Hz, H-2Ј), 7.95 (dd,
1H, J = 8.5 Hz, 1.2 Hz, H-8Ј), 7.85 (dd, 1H, J = 8.5 Hz, 1.2 Hz,
H-5Ј), 7.51 (ddd, 1H, J = 8.2 Hz, 6.9 Hz, 1.4 Hz, H-7Ј), 7.47 (d,
1H, J = 4.3 Hz, H-3Ј), 7.20 (ddd, 1H, J = 8.4 Hz, 7.0 Hz, 1.4 Hz,
H-6Ј), 5.59 (ddd, 1H, J = 17.2 Hz, 10.4 Hz, 7.8 Hz, H-10), 5.56
(br d, 1H, J = 3.8 Hz, H-9), 4.84 (dt, 1H, J = 17.2 Hz, 1.4 Hz, H-
11-Z), 4.79 (dt, 1H, J = 10.4 Hz, 1.3 Hz, H-11-E), 3.43 (dddd,
1H, J = 13.3 Hz, 10.5 Hz, 4.5 Hz, 2.4 Hz, H-6b), 2.96 (ddd, 1H,
J = 9.6 Hz, 7.6 Hz, 3.2 Hz, H-8), 2.92 (dd, 1H, J = 13.6 Hz, 10.0
Hz, H-2a), 2.52 (ddd, 1H, J = 13.7 Hz, 4.7 Hz, 2.5 Hz, H-2b),
2.45 (m, 1H, H-6a), 2.14 (m, 1H, H-3), 1.71–1.69 (m, m, 3H,
H-4, H-5b, H-7b), 1.36 (m, m, 2H, H-5a, H-7a). 13C NMR
(CDCl3, δ): 149.90 (C-2Ј), 149.90 (C-4Ј), 147.96 (C-8aЈ), 141.74
(C-10), 129.97 (C-8Ј), 128.93 (C-7Ј), 126.49 (C-6Ј), 125.48
(C-4aЈ), 122.93 (C-5Ј), 118.23 (C-3Ј), 114.27 (C-11), 71.57 (C-9),
60.34 (C-8), 56.92 (C-2), 43.15 (C-6), 39.85 (C-3), 27.87 (C-4),
27.54 (C-5), 21.26 (C-7).
Synthesis of 11-(triethoxysilyl)-10,11-dihydrocinchonidine (4)
The fairly pure 3 from the previous step was refluxed in meth-
anol (40 mL) for 20 hours. The crude product was washed with
pentane (40 mL) and filtered to leave 0.40 g (0.9 mmol, 35%
yield based on 2) of 4 as an analytically pure off-white solid.
HRMS (calcd/found): 458.2600/458.2601; mp 178–179 ЊC;
[α]2D4 = Ϫ76.9 (c = 10.7 in EtOH); elemental analysis calcd (%)
for C25H38N2O4Si (458.3): C 65.47, H 8.35, N 6.11; found
1
C 65.52, H 8.30, N 6.17%. H NMR (CDCl3, δ): 8.70 (d, 1H,
Synthesis of 9-O-(trimethylsilyl)cinchonidine (2)
J = 4.6 Hz, H-2Ј), 7.95 (dd, 1H, J = 8.4 Hz, 1.1 Hz, H-8Ј), 7.82
(d, 1H, J = 8.4 Hz, H-5Ј), 7.52 (d, 1H, J = 4.5 Hz, H-3Ј), 7.48
(ddd, 1H, J = 8.4 Hz, 6.8 Hz, 1.2 Hz, H-7Ј), 7.14 (ddd, 1H,
J = 8.4 Hz, 6.8 Hz, 1.2 Hz, H-6Ј), 5.69 (s, 1H, H-9), 5.20 (br, 1H,
OH), 3.63 (q, 6H, J = 6.9 Hz, SiOCH2CH3), 3.55 (m, 1H, H-6b),
2.99 (m, 1H, H-8), 2.97 (dd, 1H, J = 13.4 Hz, 10.0 Hz, H-2a),
2.54 (ddd, 1H, J = 13.4 Hz, 10.4 Hz, 4.6 Hz, H-6a), 2.31 (ddd,
1H, J = 13.4 Hz, 4.6 Hz, 2.4 Hz, H-2b), 1.76–1.70 (m, m, m, 3H,
H-4, H-5b, H-7b), 1.41 (m, 1H, H-3), 1.35 (m, 1H, H-5a),
1.27 (m, 1H, H-7a), 1.19 (m, 1H, H-10), 1.03 (t, 9H, J = 6.9 Hz,
SiOCH2CH3), 0.40 (m, 1H, H-11). 13C NMR (CDCl3, δ): 149.00
(C-2Ј), 148.27 (C-4Ј), 147.04 (C-8aЈ), 129.07 (C-8Ј), 127.92
(C-7Ј), 125.58 (C-6Ј), 124.44 (C-4aЈ), 121.89 (C-5Ј), 117.31
(C-3Ј), 70.03 (C-9), 59.17 (C-8), 57.30 (3C, SiOCH2CH3),
57.23 (C-2), 42.39 (C-6), 37.10 (C-3), 26.77 (C-5), 26.48 (C-10),
24.25 (C-4), 19.55 (C-7), 17.18 (3C, SiOCH2CH3), 7.18 (C-11).
To an ice-cooled solution of (Ϫ)-cinchonidine (1) (2.0 g, 6.8
mmol) in THF containing triethylamine (1.1 mL, 7.9 mmol)
was added dropwise trimethylchlorosilane (1 mL, 7.9 mmol).
The reaction mixture was stirred for 20 hours at room temper-
ature and then for two hours at 60 ЊC. The product was
extracted with chloroform (50 mL) and washed with water
(3 × 50 mL). The water layer was extracted with additional
chloroform (50 mL) and the combined organic extracts were
dried over sodium sulfate. Evaporation of the solvents left 2.3 g
(90%) of solid 2 containing approximately 10% of 9-O-
(trimethylsilyl)-10,11-dihydrocinchonidine (6) as determined by
1H NMR and identified by HRMS [6: (calcd/found): 368.2288/
368.2284]. Analytical data of 2: HRMS (calcd/found):
1
366.2127/366.2127. H NMR (CDCl3, δ): 8.79 (d, 1H, J = 4.6
Hz, H-2Ј), 8.04 (dd, 1H, J = 8.6 Hz, 1.2 Hz, H-8Ј), 8.01 (br, 1H,
H-5Ј), 7.58 (ddd, 1H, J = 8.2 Hz, 6.4 Hz, 1.2 Hz, H-7Ј), 7.45
(ddd, 1H, J = 8.2 Hz, 7.0 Hz, 1.2 Hz, H-6Ј), 7.39 (br, 1H, H-3Ј),
5.60 (br, 1H, H-9), 5.58 (ddd, 1H, J = 17.1 Hz, 10.1 Hz, 7.6 Hz,
H-10), 4.80 (ddd, 1H, J = 17.1 Hz, 1.5 Hz, 1.2 Hz, H-11-Z),
4.74 (br d, 1H, J = 9.8 Hz, H-11-E), 3.31 (ur, 1H, H-6b), 2.97
(dd, 1H, J = 13.6 Hz, 10.1 Hz, H-2a), 2.92 (br, 1H, H-8), 2.70
Analytical data of 11-(triisopropoxysilyl)-10,11-dihydrocinchon-
idine (5)
FC eluents: benzene–diethylamine–diethyl ether 20 : 12 : 5.
HRMS (calcd/found): 500.3072/500.3070. 1H NMR (CDCl3, δ):
J. Chem. Soc., Perkin Trans. 1, 2002, 2605–2612
2611

View Article Online
8.68 (d, 1H, J = 4.6 Hz, H-2Ј), 7.98 (d, 1H, J = 8.2 Hz, H-8Ј),
7.82 (d, 1H, J = 8.6 Hz, H-5Ј), 7.52 (t, 1H, J = 8.2 Hz, H-7Ј),
7.49 (d, 1H, J = 4.6 Hz, H-3Ј), 7.19 (dd, 1H, J = 8.2 Hz, 7.0 Hz,
H-6Ј), 5.63 (d, 1H, J = 3.0 Hz, H-9), 4.88 (br, 1H, OH), 4.08
(sept., 3H, J = 6.1 Hz, SiOCH ), 3.43 (br, 1H, H-6b), 2.96 (br,
1H, H-8), 2.92 (dd, 1H, J = 13.1 Hz, 10.1 Hz, H-2a), 2.48 (ddd,
1H, J = 14.4 Hz, 10.7 Hz, 4.3 Hz, H-6a), 2.28 (dm, 1H, J = 15.6
Hz, H-2b), 1.74–1.71 (br, m, m, 3H, H-4, H-5b, H-7b), 1.40 (br,
1H, H-3), 1.39–1.35 (br, 2H, H-5a, H-7a), 1.18 (br, 1H, H-10),
1.00 (d, 18H, J = 6.1 Hz, SiOCCH3), 0.36 (br, 1H, H-11). 13C
NMR (CDCl3, δ): 149.02 (C-2Ј), 148.66 (C-8aЈ), 147.12 (C-4Ј),
129.13 (C-8Ј), 127.92 (C-7Ј), 125.52 (C-6Ј), 124.59 (C-4aЈ),
121.95 (C-5Ј), 117.23 (C-3Ј), 70.90 (C-9), 63.87 (3C, SiOC),
59.09 (C-8), 57.54 (C-2), 42.35 (C-6), 37.76 (C-3), 27.22 (C-5),
26.81 (C-10), 24.45 (6C, SiOCCH3), 24.38 (C-4), 19.96 (C-7),
8.83 (C-11).
References
1 (a) Y. Orito, S. Imai and S. Niwa, J. Chem. Soc. Jpn., 1982, 137;
(b) M. Garland and H.-U. Blaser, J. Am. Chem. Soc., 1990, 112,
7048; (c) H. U. Blaser, Tetrahedron: Asymmetry, 1991, 2, 843; (d )
H. U. Blaser, H. P. Jalett, D. M. Monti, A. Baiker and J. T. Wehrli,
Stud. Surf. Sci. Catal., 1991, 76, 147; (e) H. U. Blaser, M. Garland
and H. P. Jallet, J. Catal., 1993, 144, 569; ( f ) G. Wang, T. Heinz,
A. Pfaltz, B. Minder, T. Mallat and A. Baiker, J. Chem. Soc., Chem.
Commun., 1994, 2047; (g) H. U. Blaser, H.-P. Jalett, M. Müller
and M. Studer, Catal. Today, 1997, 37, 441; (h) A. Baiker, J. Mol.
Catal. A, 1997, 115, 473; (i) P. J. Collier, T. J. Hall, J. A. Iggo,
P. Johnston, J. A. Slipszenko, P. B. Wells and R. Whyman, Chem.
Commun., 1998, 1451; (j) M. Bartók, K. Felföldi, B. Török
and T. Bartók, Chem. Commun., 1998, 2605; (k) P. B. Wells,
K. E. Simons, J. A. Slipszenko, S. P. Griffiths and D. F. Ewing,
J. Mol. Catal. A, 1999, 146, 159; (l ) C. LeBlond, J. Wang, J. Liu,
A. T. Andrews and Y.-K. Sun, J. Am. Chem. Soc., 1999, 121, 4920;
(m) D. Ferri, T. Bürgi, K. Borszeky, T. Mallat and A. Baiker,
J. Catal., 2000, 193, 139; (n) T. Bürgi and A. Baiker, J. Catal., 2000,
194, 445; (o) J. M. Bonello, R. M. Lambert, N. Künzle and
A. Baiker, J. Am. Chem. Soc., 2000, 122, 9864; (p) S. P. Griffiths,
P. Johnston and P. B. Wells, Appl. Catal. A, 2000, 191, 193;
(q) H. U. Blaser, H. P. Jalett, W. Lottenbach and M. Studer,
J. Am. Chem. Soc., 2000, 122, 12675; (r) R. Wandeler, N. Künzle,
M. S. Schneider, T. Mallat and A. Baiker, Chem. Commun., 2001,
673; (s) M. von Arx, T. Mallat and A. Baiker, J. Catal., 2001, 202,
169.
2 (a) B. Török, K. Felföldi, K. Balázsik and M. Bartók, Chem.
Commun., 1999, 1725; (b) M. Studer, S. Burkhardt and H.-U.
Blaser, Chem. Commun., 1999, 1727.
3 (a) W. A. H. Vermeer, A. Fulford, P. Johnston and P. B. Wells,
J. Chem. Soc., Chem. Commun., 1993, 1053; (b) J. A. Slipszenko,
S. P. Griffiths, P. Johnston, K. E. Simons, W. A. H. Vermeer and
P. B. Wells, J. Catal., 1998, 179, 267; (c) M. Studer, V. Okafor and
H. Blaser, Chem. Commun., 1998, 1053.
4 (a) H. W. Stuurman, J. Köhler and G. Schomburg, Chromato-
graphia, 1988, 25, 265; (b) P. N. Nesterenko, V. V. Krotov and
S. M. Staroverov, Russ. J. Phys. Chem., 1991, 65, 1415; (c) P. N.
Nesterenko, V. V. Krotov and S. M. Staroverov, J. Chromatogr., A,
1994, 667, 19 . See also: C. Pettersson and C. Gioeli, J. Chromatogr.,
1987, 398, 247; (d ) M. Lämmerhofer and W. Lindner,
J. Chromatogr., A, 1996, 741, 33; (e) S. Schefzick, M. Lämmerhofer,
W. Lindner, K. Lipkowitz and M. Jalaie, Chirality, 2000, 12, 742 and
references cited therein.
5 V. A. Tertykh, V. V. Yanishpolskii, L. V. Bereza, J. J. Pesek
and M. Matyska, J. Therm. Anal. Calorim., 2000, 62, 539.
6 (a) E. Toukoniitty, P. Mäki-Arvela, A. N. Villela, A. Kalantar
Neyestanaki, T. Salmi, R. Leino, R. Sjöholm, E. Laine, J. Väyrynen,
T. Ollonqvist and P. J. Kooyman, Catal. Today, 2000, 60, 173;
(b) E. Toukoniitty, P. Mäki-Arvela, A. Kalantar Neyestanaki,
T. Salmi, A. Villela, R. Leino, R. Sjöholm, E. Laine, J. Väyrynen and
T. Ollonqvist, Stud. Surf. Sci. Catal., 2000, 130, 3363; (c)
E. Toukoniitty, P. Mäki-Arvela, M. Kuzma, A. Villela, A. Kalantar
Neyestanaki, T. Salmi, R. Sjöholm, R. Leino, E. Laine and
D. Y. Murzin, J. Catal., 2001, 204, 281.
Hydrogenation of 1-phenylpropane-1,2-dione
The platinum catalyst (148 mg) containing 5% Pt on Al2O3
(Strem) was activated at 400 ЊC in a 100 mL Sotelem batch
reactor under hydrogen flow for two hours. The chiral modifier
(molar amount corresponding to 20 mg of cinchonidine) was
dissolved in 50 mL of a 0.05 M solution of 1-phenylpropane-
1,2-dione in ethyl acetate. The solution was flushed under a
hydrogen flow for 10 minutes prior to injection into the reactor.
The hydrogenation was carried out at 25 ЊC and 5 bar hydrogen
pressure using a stirrer speed of 1800 rpm. The reaction was
followed by GC analysis. For determination of the enantio-
meric excesses, microsamples (V = 200 µl) were collected during
the hydrogenation experiments and analyzed using a Varian
3300 GC equipped with a chiral β-dex column. Identification of
the product enantiomers and calibration of the GC analysis
was carried out utilizing previously prepared enantiomerically
pure model compounds 8a and 10a. The minor products, enan-
tiomers 9a/9b and 10b/11b could not be properly separated.
Details of the GC standard synthesis, identification of the
product enantiomers and details of the GC analysis procedure
have been reported in a previous paper.6 All hydrogenations
were carried out in duplicate and resulted in reproducible ee
values.
Computational details
Geometry optimizations were carried out in two steps for com-
pounds 1, 2 and 4. In the first step a two-dimensional con-
formational energy map was calculated. The starting point was
the experimental crystal structure of 1.16 The dihedral angles
(3Ј-4Ј–9–8) and (4Ј–9–8–N) were scanned simultaneously with
10 or 20Њ steps and at each point all other geometry variables
were optimised at the HF/3-21G level. In the minima regions a
smaller angle increment was applied. In the grid positions of
minimum energy a full optimization was first performed at the
HF/3-21G level and subsequently at the HF/6-31G** level.
Optimizations were also carried out for compound 1, with
initial geometries corresponding to the structures reported by
Bürgi and Baiker.14 The minimum energy conformations for
compound 3 were derived from HF/6-31G** optimizations. In
the initial geometries the dihedral angles were taken from the
optimum values for compound 1.
7 (a) C. G. Moreland, A. Philip and F. I. Carroll, J. Org. Chem., 1974,
39, 2413; (b) O. Were, M. Benn and R. Munavu, Planta Med., 1997,
63, 90.
8 (a) See, for example: P. M. Subramanian, S. K. Chatterjee and
M. C. Bhatia, J. Chem. Technol. Biotechnol., 1987, 39, 215; (b)
P. Iwan, G. Goetz, S. Schmitz, B. Hauer, M. Breuer and M. Pohl,
J. Mol. Catal. B: Enzym., 2001, 11, 387.
9 D. Gala, D. J. DiBenedetto, J. E. Clark, B. L. Murphy,
D. P. Schumacher and M. Steinman, Tetrahedron Lett., 1996, 37,
611.
10 T. Koike, K. Murata and T. Ikariya, Org. Lett., 2000, 2, 3833.
11 V. Morawsky, U. Prüße, L. Witte and K.-D. Vorlop, Catal.
Commun., 2000, 1, 15.
12 M. Schürch, O. Schwalm, T. Mallat, J. Weber and A. Baiker,
J. Catal., 1997, 169, 275.
13 J. L. Margitfalvi and E. Tfirst, J. Mol. Catal. A: Chem., 1999, 139,
81.
14 T. Bürgi and A. Baiker, J. Am. Chem. Soc., 1998, 120, 12920.
15 D. Ferri and T. Bürgi, J. Am. Chem. Soc., 2001, 123, 12074.
16 B. J. Oleksyn, Acta Crystallogr., Sect. B, 1982, 38, 1832.
Acknowledgements
We thank Mr Markku Reunanen for his kind help with
the HRMS analyses and Mattias Roslund for assistance in
recording some of the NMR spectra.
2612
J. Chem. Soc., Perkin Trans. 1, 2002, 2605–2612