A study on chiral organocalcium complexes: Attempts in enantioselective catalytic hydrosilylation and intramolecular hydroamination of alkenes

Article abstract of DOI:10.1515/znb-2008-0209

The chiral β-diketimine ligand [(A)-Ph(Me)CH-N=C(Me)]CH₂was prepared by condensation of acetylacetone with the commercially available chiral building block (S)-Ph(Me)CH-NH₂. Reaction of bis(o-Me ₂N-α-Me₃Si-benzy

Full text of DOI:10.1515/znb-2008-0209

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
A Study on Chiral Organocalcium Complexes: Attempts in
Enantioselective Catalytic Hydrosilylation and Intramolecular
Hydroamination of Alkenes
Frank Buch and Sjoerd Harder
Anorganische Chemie, Universita¨t Duisburg-Essen, Universita¨tsstraße 5-7, 45117 Essen, Germany
Reprint requests to Prof. Dr. Sjoerd Harder. E-mail: sjoerd.harder@uni-due.de
Z. Naturforsch. 2008, 63b, 169 – 177; received August 29, 2007
The chiral β-diketimine ligand [(S)-Ph(Me)CH-N=C(Me)]CH2was prepared by condensation of
acetylacetone with the commercially available chiral building block (S)-Ph(Me)CH-NH2. Reaction
of bis(o-Me2N-α-Me3Si-benzyl)calcium with this β-diketimine led to double deprotonation. Reac-
tion of bis(o-Me2N-α-Me3Si-benzyl)calcium with the commercially available chiral bis-oxazoline
(S)-Ph-BOX gave diastereopure [(S)-Ph-BOX](o-Me2N-α-Me3Si-benzyl)calcium which in solution
slowly decomposed with formation of o-Me2N-α-Me3Si-toluene. The corresponding amide complex
[(S)-Ph-BOX]CaN(SiMe3)2·(THF)2 is stable and the crystal structure has been determined. In solu-
tion, this heteroleptic amide is in Schlenk equilibrium with the homoleptic species [(S)-Ph-BOX]2Ca
and Ca[N(SiMe3)2]2·(THF)2. This Schlenk equilibrium can be steered to the heteroleptic side. Use
of the enantiopure calcium amide catalyst for the hydrosilylation of styrene with PhSiH3 or in the
intramolecular hydroamination of aminoalkenes gave good product yields, but only small ee-values
were observed (5 – 10 %). From stoichiometric reactions of the catalyst with the substrates it is con-
cluded that the “true” catalytically active species is mainly present as a homoleptic calcium complex,
which explains the poor enantioselectivities.
Key words: Alkaline Earth Metals, Calcium, Hydrosilylation, Hydroamination
Introduction
Although use of the main group metal calcium in
homogeneous catalysis is relatively rare, reports on
well-defined organocalcium catalysts are on the in-
crease [1 – 5]. The unique feature of such catalysts is
the combination of the relatively high Lewis acidity of
the Ca2+ center with the considerable nucleophilicity
of the ligands. Consequently, these catalysts have been
exploited in living syndioselective styrene polymeriza-
tion [1], dilactide polymerization [2], hydroamination
and hydrophosphination[3], the Tischenko reaction [4]
and in alkene hydrosilylation [5].
In calcium-mediated styrene polymerization a syn-
dioselectivity of up to 93% in r-diades was obtained.
This corresponds to a considerable stereoselectivity
in the chain-end controlled insertion step of ca. 87 %
ee (Eq. 1). Likewise, mixtures of calcium alkox-
ides with chiral bidentate diol or bis-oxazoline lig-
ands have been used for the asymmetric aldol reac-
tion, the Michael addition or the epoxidation of α,β-
unsaturated enones [6]. Although in some cases ee-
values of over 90 % could be obtained, the latter in situ
(1)
terized, and the chemical composition of the catalysts
is unclear.
In the light of these results, we embarked on a search
for well-defined chiral enantiopure calcium catalysts.
We describe the difficulties encountered in the synthe-
ses of these complexes and in the first attempts to intro-
duce chiral Ca-catalysts for enantioselective hydrosily-
lation and intramolecular hydroamination of alkenes.
The rather low ee-values obtained in these particular
catalytic conversions are explained by the results of
an in-depth study of structure and stability of the het-
eroleptic chiral calcium catalysts.
Results and Discussion
Synthesis and characterization of chiral enantiopure
calcium reagents
In this work, we focussed on two strongly coordinat-
prepared Ca catalysts have not been isolated or charac- ing monoanionic bidentate ligands with C2-chirality.
0932–0776 / 08 / 0200–0169 $ 06.00 © 2008 Verlag der Zeitschrift fu¨r Naturforschung, Tu¨bingen · http://znaturforsch.com
Unauthenticated
Download Date | 9/8/18 4:27 AM

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
170
F. Buch – S. Harder · Chiral Organocalcium Complexes
and 2 failed. Reaction of the ligand 1 with (Me2N-
α-Me3Si-benzyl)2Ca·(THF)2 (3) in a 1/1 ratio gave
immediate deprotonation of the ligand already at r. t.
The 1H NMR spectra of the reaction mixtures showed
several products that could not be identified unam-
biguously. The immediate appearance of a red color,
which intensified over time, and the complete conver-
sion of the benzylcalcium functionality into 2-Me2N-
α-Me3Si-toluene indicated a second deprotonation of
the ligand in the benzylic position.
Deprotonation of (S)-Ph-BOX (2) with (Me2N-α-
Me3Si-benzyl)2Ca·(THF)2 (3) was more selective and
gave clean conversion to 4 (Scheme 1). Whereas the
dibenzylcalcium precursor 3 consists of a pair of inter-
converting diastereomers in benzene solution (Tcoal
=
60 ◦C, ∆G‡ = 16.8 kcal mol−1), only one diastereomer
was observed for 4. This underlines efficient commu-
nication between the chiral (S)-Ph-BOX ligand and the
chiral benzylic carbon center.
Although 4 can be prepared in situ, isolation turned
out to be impossible for two reasons. First of all, in so-
lution 4 is in a Schlenk equilibrium with the homolep-
tic species 3 and 5 (Scheme 1; the ratio of heteroleptic
to both homoleptic species is approximately 4/1/1, i.e.
K ≈ 0.06). Similar Schlenk equilibria have also been
observed for in situ prepared heteroleptic BOX-Mg-
iPr complexes [8b]. More importantly, a yellow ben-
zene solution of 4 slowly decomposed completely over
a two day period at r. t. to give an intensely red solu-
◦
tion (at 50 C decomposition is complete within 2 h).
Although the decomposition product is as yet uniden-
tified (due to very broad signals in the NMR spec-
trum), complete conversion of the benzylic anion into
2-Me2N-α-Me3Si-toluene indicates a second (slower)
deprotonation of the Ph-BOX ligand. The intense red
color of the decomposition product strongly suggests a
species in which one chiral benzylic carbon has been
deprotonated (6, Scheme 1).
Scheme 1.
We synthesized a chiral enantiopure β-diketimine (1)
Decomposition reactions of heteroleptic benz-
by condensation of acetylacetone with the readily ylcalcium complexes have been observed earlier [9]
available building block (S)-α-Me-benzylamine.In ad- and are due to the high reactivity of the benzyl-Ca
dition, the commercially available bis-oxazoline (S)- functionality. Therefore, we directed our investigations
Ph-BOX (2) is addressed. The BOX system is an im- to the less reactive calcium amide complexes. Depro-
portant class of C2-symmetric chiral ligands generally tonation of 1 with [(Me3Si)2N]2Ca·(THF)2 in benzene
used in the neutral form [7]. Although it has been needed more forcing conditions, but was essentially
◦
used as an anionic ligand in alkaline-earth metal chem- complete after 16 h at 50 C. No second deprotona-
istry [6e, 8], hitherto no complexes have been isolated tion of the β-diketimine ligand could be observed. The
or structurally characterized.
heteroleptic product 7 is in Schlenk equilibrium with
Attempts to prepare stable well-defined heteroleptic the homoleptic species [(Me3Si)2N]2Ca·(THF)2 and 8
benzylcalcium complexes based on the chiral ligands 1 in an approximate ratio of 4/1/1 (Eq. 2).
Unauthenticated
Download Date | 9/8/18 4:27 AM

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
F. Buch – S. Harder · Chiral Organocalcium Complexes
171
˚
Table 1. Selected bond lengths (A) and angles (degrees) for
[(S)-Ph-BOX]CaN(SiMe3)2·(THF)2 (9).
Ca-N1
2.404(2)
2.381(2)
2.348(2)
78.84(7)
131.00(7)
96.97(7)
88.15(7)
88.70(7)
88.96(7)
77.94(8)
Ca-O3
2.410(2)
2.417(2)
3.191(4)
88.15(7)
88.70(7)
88.96(7)
163.73(9)
113.42(9)
81.65(8)
(2)
Ca-N2
Ca-N3
Ca-O4
Ca···C23
N2-Ca-O4
N3-Ca-O3
N3-Ca-O4
C23···Ca-N1
C23···Ca-N2
C23···Ca-O3
N1-Ca-N2
N1-Ca-N3
N2-Ca-N3
N1-Ca-O3
N1-Ca-O4
N2-Ca-O3
C23···Ca-O4
mother liquor gave crystals of a well-defined product
with composition [(S)-Ph-BOX]CaN(SiMe3)2·(THF)2
(9).
The crystal structure of 9 is shown in Fig. 1 and
geometrical data are summarized in Table 1. The
BOX ligand chelates the calcium metal via both ni-
trogen atoms and retains approximate C2-symmetry
(Fig. 1b). The oxazoline rings are slightly tilted with
respect to each other: the dihedral angle between the
least-squares planes through both rings is 13.4(2)◦.
˚
The Ca-N(SiMe3)2 bond of 2.348(2) A is consid-
erably longer than the terminal Ca-N(SiMe3)2 bond
˚
of 2.275(7) A in [((Me3Si)2N)2Ca]2 [10]. In addition,
an agostic Si-Me···Ca2+ interaction is evident from
˚
a short Ca···C23 distance of 3.191(4) A (sum of the
˚
van-der-Waals radii of Ca and C is 3.49 A) and corre-
˚
sponding H···Ca distances of 3.03(4) and 3.06(4) A.
This agostic Me···Ca interaction causes tilting of
the amide ligand with respect to the Ca-N3 axis:
the Ca-N3-Si1 angle of 112.4(1)◦ is significantly
smaller than the Ca-N3-Si2 angle of 121.3(1)◦. Like-
wise, the N3-Si1-C23 angle of 108.4(1) is squeezed
somewhat with respect to an ideal tetrahedral angle.
Consequently, other N-Si2-C angles in the complex
are widened (111.9(2)◦ – 115.1(1)◦). The coordination
sphere for Ca is completed by two THF ligands.
An interesting comparison can be made with
the heteroleptic achiral calcium amide 10 [11].
The CH{(CMe)(2,6-iPr2C6H3N)}2 ligand (= DIPP-
nacnac) in this complex coordinates in a bidentate
˚
fashion with Ca-N bonds of 2.352(1) and 2.370(1) A,
i.e. slightly shorter than those in 9. Also the Ca-N
Fig. 1. (a) Crystal structure of [(S)-Ph-BOX]CaN(SiMe3)2
· (THF)2 (9); hydrogen atoms omitted for clarity. Selected
bond lengths and angles are given in Table 1. (b) View along
the approximate C2 axis of the BOX ligand.
˚
(SiMe3)2 bond of 2.313 A in 10 is slightly shorter than
that in 9. A similar agostic interaction with compara-
ble features as in 9 is also observed for 10. The larger
spatial requirement of the DIPP-nacnac ligand in 10 al-
Deprotonation of the (S)-Ph-BOX ligand 2 with one lows coordination of only one THF ligand. Therefore,
equivalent of [(Me3Si)2N]2Ca·(THF)2 was complete the steric bulk of the DIPP-nacnac ligand equals that
after 30 min at r. t. Concentration and cooling of the of one (S)-Ph-BOX ligand and a THF ligand.
Unauthenticated
Download Date | 9/8/18 4:27 AM

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
172
F. Buch – S. Harder · Chiral Organocalcium Complexes
Table 2. Catalytic conversion of styrene and phenylsilane into
PhCH(SiH2Ph)Me.
Entry mol-% cat.
1 [5] 2.5 %
T(◦C) t(h) conv.(%)
20 < 0.1 > 98
(3)
3
2
2.5 % [(Me3Si)2N]2Ca 50
1
> 98
· (THF)2
3
4
5
a
2.5 %
5 %
5 %
9
50
50
50
16 > 98 (S(-), 5 % ee)
16 > 98 (S(-), 9 % ee)
16 > 98 (S(-), 9 % ee)
5/9 (1/1)a
7/8 (1/1)a
In contrast to 4, complex 9 is stable against a sec-
ond deprotonation of the bidentate ligand. However, 9
mol-% catalyst calculated on the basis of the active Ca-N(SiMe3)2
functionality.
1
is not stable against ligand exchange: the H NMR
control over the stereochemistry is therefore highly
desirable.
spectrum of crystals of 9 dissolved in C6D6 shows a
Schlenk equilibrium with the two homoleptic calcium
complexes [(Me3Si)2N]2Ca·(THF)2 and 5 (Eq. 3) in a
ratio of 4.2/1/1. The α-CH2 protons of the THF ligands
in 9 are diastereotopic and give separate signals in the
1H NMR spectrum, whereas the more distant β-CH2
protons display a single resonance. This indicates that
also in solution the BOX ligand and both THF ligands
Since Ca-mediated hydrosilylation of alkenes is
limited to activated alkenes, e. g. conjugated alkenes,
styrene was chosen as an appropriate prochiral sub-
strate. Hydrosilylation of styrene with phenylsilane us-
ing the benzylcalcium catalyst 3 was shown to be fast
and completely regioselective (Table 2, entry 1) [5].
Use of chiral calcium amides as catalysts requires the
catalytic activity of the calcium amide functionality.
We found that [(Me3Si)2N]2Ca·(THF)2 is also an ef-
ficient, but somewhat less active, catalyst for this reac-
tion giving exclusively one regio-isomer (Table 2, en-
try 2). Catalysis with the chiral amide catalyst 9 gave
overnight essentially full conversion of the substrates
(Table 2, entry 3). Oxidation of the chiral product ac-
cording to Tamao-Fleming, which is known to pro-
ceed with retention of configuration at the chiral benz-
ylic carbon atom [14], gave the corresponding 1-Ph-
ethanol, however, a very low enantiomeric excess of
ca. 5 % ee was observed. As [(Me3Si)2N]2Ca·(THF)2
is an effective catalyst for this reaction, this could orig-
inate from insufficient control of the Schlenk equi-
librium. Runs with a 1/1 mixture of 9 and 5, which
is largely free of homoleptic [(Me3Si)2N]2Ca·(THF)2,
gave only slightly improved results. Likewise, hydrosi-
lylation of styrene with a 1/1 mixture of 7 and 8 gave a
similarly low ee of 9 % (Table 2, entry 4).
are bonded to Ca2+
.
The Schlenk equilibria observed for enantiopure
calcium amides 7 and 9 in solution could seriously
hamper the stereoselectivity in catalytic conversions
(assuming that [(Me3Si)2N]2Ca·(THF)2 is also catalyt-
ically active). Both Schlenk equilibria in Eqs. 2 and 3,
however, can be steered to the heteroleptic species 7
and 9, simply by adding the homoleptic species 8
and 5, respectively, which are not catalytically ac-
tive in hydrosilylation or hydroamination catalysis. A
1/1 mixture of 7 and 8 in benzene only contains small
amounts of [(Me3Si)2N]2Ca·(THF)2 (< 3 %). Simi-
larly, only minor amounts of [(Me3Si)2N]2Ca·(THF)2
(< 2 %) could be detected in a 1/1 mixture of 9 and 5
in benzene.
Catalytic hydrosilylation and hydroamination with
chiral calcium amide complexes
We recently introduced the catalytic hydrosilyla-
tion of alkenes with early main group metal cata-
These remarkably low ee values indicate that
lysts [5] for which we proposed a mechanism similar the “true” catalytically active species might be achiral.
to that for organolanthanide catalysts [12] (Scheme 2, The proposed chiral catalyst (L∗CaH in Scheme 2) is a
left). Although these catalysts can not compete with heteroleptic calcium hydride for which recently an ex-
the well-established class of highly active transi- ample has been isolated [15]. Although Schlenk equi-
tion metal catalysts [13], they exhibit some remark- libria for the catalyst precursors (L∗CaN(SiMe3)2 in
able features. Complete regiocontrol of the hydrosi- Scheme 2) can be directed quite well to the heterolep-
lylation reaction, which in some cases could be tic side, Schlenk equilibria for L∗CaH could mainly be
switched by either metal or solvent choice, was ob- shifted to the homoleptic side. In this case catalytic
served. Moreover, use of catalysts based on the much conversion of styrene into PhCH(SiH2Ph)Me would
cheaper and biocompatible calcium could certainly largely proceed via in situ generated CaH2. This pre-
be of interest for potential applications. Additional sumption is enforced by the observation that stoichio-
Unauthenticated
Download Date | 9/8/18 4:27 AM