Highly enantioselective hydroamination to six-membered rings by heterobimetallic catalysts

Article abstract of DOI:10.1039/c3cc48874h

New bimetallic Zn/Zr salen-type systems were employed as catalysts in the asymmetric intramolecular hydroamination reaction. High enantioselectivity for the formation of piperidines of up to 98% ee were observed. This journal is the Partner Organisations

Full text of DOI:10.1039/c3cc48874h

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Highly enantioselective hydroamination to
six-membered rings by heterobimetallic catalysts†
Cite this: Chem. Commun., 2014,
50, 3862
Lenard Hussein,*a Nibadita Purkait,a Mustafa Biyikal,a Eugenia Tausch,b
Peter W. Roeskyb and Siegfried Blechert*a
Received 20th November 2013,
Accepted 19th February 2014
DOI: 10.1039/c3cc48874h
www.rsc.org/chemcomm
New bimetallic Zn/Zr salen-type systems were employed as catalysts in
the asymmetric intramolecular hydroamination reaction. High enantio-
selectivity for the formation of piperidines of up to 98% ee were observed.
The catalytic addition of an amine N–H bond to unsaturated
carbon–carbon bonds (hydroamination), to give nitrogen containing
molecules in a facile and highly atom-economical way is of great
interest for academic and industrial research.1 Hydroamination has
a high activation barrier. To lower the activation barrier catalytic
systems were employed, which either activate the C–C multiple bond
or the N–H function. Catalytic hydroamination has been intensely
studied in the last decades.1,2 Early results of the intramolecular
asymmetric hydroamination reactions were reported by Marks et al.
using chiral ansa-lanthanocene.3 Significant research activities in the
area of asymmetric inter- and intramolecular hydroamination reac-
tions have been observed for about ten years now.4 Although high
yields with excellent ee values were reported for the intramolecular
hydroamination, the asymmetric hydroamination of alkenes has
remained a challenging task because the excellent results are mainly
Scheme 1 Synthesis of the ligand L-1, Cu-1a and Zn-1.
limited to the formation of pyrrolidines, whereas the ee values Starting from tropolone 1, 2 could be obtained after bromination with
obtained for the formation of piperidines are less impressive.2,5 To Br2 in MeOH followed by a Mitsunobu-reaction with PPh3, DIAD and
the best of our knowledge the highest ee for this reaction was 82%.5e MeOH to give 3. A reaction with optically active 1-phenylethylamine
In the last few years we and others introduced zinc compounds as and subsequent rearrangement reaction with isopropylamine
catalysts for the hydroamination of alkenes and alkynes.6 Within followed by treatment with HCl led to the ortho-aminated salicyl-
these research efforts we reported on the dimeric tetranuclear zinc aldehyde 4. A subsequent condensation with chiral 1,2-diaminocyclo-
complexes derived from a novel type of ortho-aminated salicyl- hexane gave the chiral brominated salen ligand L-1. Reacting L-1 with
aldimine, which showed high reactivity for the intramolecular hydro- M(OAc)2 (M = Cu and Zn) resulted in the corresponding copper and
amination reaction of non-activated alkenes at room temperature.7 zinc complexes Cu-1a and Zn-1. The solid state structure of Zn-1 was
With an aim to generate an enantioselective catalyst for the hydro- established by single crystal X-ray diffraction. Zn-1 crystallizes from
amination, we now introduced chirality into the ligand by synthesiz- d6-DMSO (Fig. 1). The Zn atom is penta-fold coordinated, forming a
ing a salen-type ligand with several stereocenters (Scheme 1). distorted square pyramid. The deprotonated salen ligand coordinates
in the plane while the DMSO molecule is localized in the apex of the
a
¨
¨
Institut fur Chemie, Technische Universitat Berlin, Straße des 17. Juni 135,
square pyramid. The Zn–O and Zn–N bond distances to the
salen ligand are in the expected range of Zn1–N1 2.077(5) Å,
Zn1–N2 2.088(5) Å, Zn1–O3 1.978(5) Å, and Zn1–O4 1.969(4) Å.
We used the transformation of 2,2-diphenylhex-5-en-1-amine (I)
to 2-methyl-5,5-diphenylpiperidine (II) as a test-reaction (Table 1
and Table S1, ESI†). Since an initial test with in situ generated Zn-1
10623 Berlin, Germany. E-mail: blechert@chem.tu-berlin.de
b
¨
¨
Institut fur Anorganische Chemie, Karlsruher Institut fur Technologie (KIT),
Engesserstr. 15, Geb. 30.45, 76131 Karlsruhe, Germany
† Electronic supplementary information (ESI) available. CCDC 965327–965329.
For ESI and crystallographic data in CIF or other electronic format see DOI:
10.1039/c3cc48874h
3862 | Chem. Commun., 2014, 50, 3862--3864
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
metalloligand Zn-1 in combination with Zr(NMe2)4 showed the same
ee as the in situ generated system but a slightly lower conversion. While
keeping the Zr-source, ZnMe2 was exchanged by Cu(OAc)2, which
resulted in 96% conversion and 77% ee (Table S1, ESI,† entry 6). In
contrast, by using the metalloligand Cu-1a and Zr(NMe2)4 full conver-
sion and 96% ee were obtained (Table S1, ESI,† entries 7 and 8). These
results could be obtained with a catalyst concentration as low as
5 mol% Cu-1a and 5 mol% Zr(NMe2)4. In comparison, by using only
Zr(NMe2)4 with L-1 as a catalyst full conversion but only 77% ee was
observed (Table S1, ESI,† entry 9). This clearly shows the influence of
the bimetallic system. Besides the metals the ligand system also
strongly influences the reactions. By using the Jacobsen-ligand
(Scheme S1, ESI†) instead of L1 no conversion was observed in the
test reaction. Since Zn-1 alone is catalytically inactive we suggest that
the 3d metal just stabilizes the salen ligand by forming a rigid scaffold
(metalloligand), which binds to Zr forming a chiral Zr amido catalyst.
Next we were interested in the influence of the stereocenters on
the selectivity. Cu-1a has four stereocenters (two at the cyclohexyl
ring and two at the phenylethyl moiety), which were systematically
varied leading to the complexes Cu-1b, Cu-1c, and Cu-1d
(Scheme 2). All new complexes were synthesized similar to Cu-1a.
Cu-1a and Cu-1b are both composed of (1S,2S)-1,2-
diaminocyclohexane, but differ in the configuration of the
1-phenylethylamine-group, while Cu-1c and Cu-1d are based
on (1R,2R)-1,2-diaminocyclohexane carrying different stereo-
information at the 1-phenylethylamine-group. These metallo-
ligands in combination with Zr(NMe2)4 were applied in the test
reaction (Table S2, ESI†). The results clearly show that the 1,2-diamino-
cyclohexane-group is responsible for the configuration of the product.
The ligands Cu-1a and Cu-1b bearing a (1S,2S)-1,2-diaminocyclo-
hexane group in the ligand scaffold (Table S2, ESI,† entries 1 and 2)
gave the opposite enantiomer compared to Cu-1c and Cu-1d which are
based on (1R,2R)-1,2-diaminocyclohexane (Table S2, ESI,† entries 3
and 4). Changing the configuration at the 1-phenylethylamine moiety
resulted in only a minor impact on the ee’s. On the other hand we
could show for the related ligand L2 (see below) that a replacement of
the 1-phenylethylamine moiety by other groups is not beneficial.
To investigate the role of the bromine atoms in the ligand a
bromine-free salen ligand (L-2) was synthesized (Scheme 3).
Fig. 1 Solid state structure of Zn-1 (left, only one of two independent
molecules is shown) and Cu-2 (right), omitting the hydrogen atoms.
Table 1 Asymmetrical intramolecular hydroamination reactions
Time Conversion ee(de)
Entry Substrate
Product
(h)
(%)
(%)
1aa
1b
84
18
81
99
98
97
2
3
48
24
55
99
68
22
4
24
99
74
5
6
24
41
99
99
29
77
7
8
24
26
99
99
51(34)
8(27)
0.25 mmol substrate, 10 mol% of L-2, 15 mol% of ZnMe2, 15 mol% of
Zr(NMe2)4, 0.5 ml toluene-d8, 80 1C, sealed NMR-tube, conversions were
a
determined by 1H-NMR, ee were determined by HPLC. 10 mol% of
L-2, 10 mol% of ZnMe2, 9.5 mol% of Zr(NMe2)4 were used.
from L1 and an excess of ZnMe2 did not show any conversion
(Table S1, ESI,† entry 1), we investigated the application of
different bimetallic catalytic systems for the test reactions. The
combination of L1, ZnMe2 and Zr(NMe2)4 and the combination
of Zn-1 and Zr(NMe2)4 resulted in high conversion and 96% ee
(Table S1, ESI,† entries 2 and 5). Zr(NMe2)4 and its derivatives are
Scheme 2 Synthesized metalloligands with different configurations at
known as catalysts for the hydroamination reaction.5b,c,8 The isolated the stereocenters.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 3862--3864 | 3863

View Article Online
ChemComm
Communication
for the formation of a five membered ring resulted in the formation
of 2-methyl-4,4-diphenylpyrrolidine (VI) in 99% conversion and 22%
ee (entry 3). Substituents on the phenyl rings of the substrates have
also a strong influence on the enantioselectivity. Substrates with
p-CH3 (VII) or m-MeO substituents (IX) at the phenyl groups were
cyclized in a selectivity of 74% ee (VIII, entry 4) and 29% ee
(X, entry 5). We suggest that the MeO group may interact as a
Lewis-base with the Lewis-acidic metals and thus cause a significantly
different transition state in the catalytic reaction. In contrast a cyclo-
hexyl-group on the substrate (XI) does not interact with the metals of
the catalysts. 77% ee was observed in the conversion of XI to XII
(entry 6). Besides the enantioselectivity the diastereoselectivity was
also investigated by using substrates XIII and XV. Formation of the six
membered ring XIV by using XIII resulted in 51% ee and 34% de
(entry 7) while the five membered ring XVI was formed with only
8% ee and 27% de (entry 8). In both cases full conversion was
observed.
Scheme 3 Synthesis of the ligand L-2, Cu-2 and Zn-2.
Commercially available 2,6-dibromophenol 5 was protected with
We prepared two new salen-type ligands with four stereocenters.
MOMCl to obtain 6. A Buchwald–Hartwig reaction with optically Reactions of these ligands with appropriate Zn(II) or Cu(II) sources
active (R)-1-phenylethylamine led to 7 and formylation resulted resulted in Zn and Cu salen complexes. These complexes are not
in 8. After deprotection with 6 M HCl compound 9 was obtained. A active as catalysts. The metal only acts as a template by forming
subsequent condensation of 9 with (1S,2S)-1,2-diaminocyclohexane rigid metalloligands. Reacting these metalloligands with Zr(NMe2)4
provided L-2. Reaction of L-2 with M(OAc)2 (M = Cu and Zn) resulted in very active and selective catalysts. Using our catalytic
resulted in the formation of the corresponding copper and zinc system piperidines could be formed by asymmetric hydroamination
complexes Cu-2 and Zn-2. The solid state structures of Cu-2 and in ee’s of up to 98%. These are the highest ee’s reported so far.
Zn-2 were established using single crystal X-ray diffraction (Fig. 1 Currently, we do not know the structure of the catalysts, but we
and Fig. S17, ESI†). The Cu atom is four-fold coordinated forming a anticipate that Zr(NMe2)4 reacts with the NH function of the
square planar arrangement. The Cu–O and Cu–N bond distances to phenylethyl moiety of the metalloligand. By extrusion of Me2NH a
the salen ligand are in the expected range of Cu–O1 1.905(3) Å, Cu– Zr-amido species may be formed in which the zirconium atom is
O2 1.919(3) Å, Cu–N1 1.960(4) Å, and Cu–N2 1.944(4) Å. The solid bound to the metalloligand. Currently, investigations are in progress
state structure of Zn-2 is comparable to that of Zn-1.
to get a deeper insight into the reported bimetallic9 catalytic system.
This work was supported by the Deutsche Forschungsge-
To compare the bromine free ligand L-2 with the bromine
containing system L-1, Cu-2 and Zn-2 and in situ prepared Zn-2 (from meinschaft, the transregional collaborative research center
L-2 and ZnMe2) were applied in the test reaction. In all cases SFB/TRR 88 ‘‘Cooperative effects in homo- and hetero-metallic
Zr(NMe2)4 was used to mediate the reaction (Table S3, ESI†). The complexes (3MET)’’, and the Fonds der Chemischen Industrie.
combination L-2 with ZnMe2 and Zr(NMe2)4 (97% ee; Table S3, ESI,†
entry 3) is the most efficient system by a small margin for the test
reaction. It is also slightly better than the combination of L-1 with
ZnMe2 and Zr(NMe2)4 (Table S1, ESI,† entry 2). With the new bromine
free system in hand we then optimized the ligand to metal ratio.
Notes and references
1 T. E. Mu¨ller, K. C. Hultzsch, M. Yus, F. Foubelo and M. Tada, Chem.
Rev., 2008, 108, 3795–3892.
2 A. L. Reznichenko and K. C. Hultzsch, Organometallics, 2013, 32,
In the test reaction the highest ee’s (98%) were obtained by using
10 mol% of L-2, 10 mol% of ZnMe2 and a slight deficiency of
9.5 mol% of Zr(NMe2)4 (Table 1, entry 1a). Since after 84 h only a
conversion of 81% was obtained, a ratio of 10 mol% of L-2, 15 mol%
of ZnMe2 and 15 mol% of Zr(NMe2)4 was tested (Table 1, entry 1b,
and Table S4, ESI†, entry 4). The slightly worse enantioselectivity
1394–1408.
3 M. A. Giardello, V. P. Conticello, L. Brard, M. R. Gagne and T. J. Marks,
J. Am. Chem. Soc., 1994, 116, 10241–10254.
4 (a) N. Nishina and Y. Yamamoto, in Topics in Organometallic Chemistry,
ed. V. P. Ananikov and M. Tanaka, Springer, Berlin, Heidelberg, 2013,
vol. 43, pp. 115–143; (b) A. Reznichenko and K. Hultzsch, in Topics in
Organometallic Chemistry, ed. V. P. Ananikov and M. Tanaka, Springer,
Berlin, Heidelberg, 2013, vol. 43, pp. 51–114.
(97%) in combination with a moderate catalyst loading, higher 5 (a) Y. Zhang, W. Yao, H. Li and Y. Mu, Organometallics, 2012, 31,
4670–4679; (b) K. Manna, S. Xu and A. D. Sadow, Angew. Chem., Int. Ed.,
2011, 50, 1865–1868; (c) K. Manna, W. C. Everett, G. Schoendorff, A. Ellern,
T. L. Windus and A. D. Sadow, J. Am. Chem. Soc., 2013, 135, 7235–7250;
conversion (99%) and shorter reaction times (18 h) convinced us to
use this ratio for an extended substrate screening (Table 1). For the
test reaction we observed the highest ee’s which were, to the best of
our knowledge, ever observed for the hydroamination/cyclization of a
six-membered ring. Although the other substrates were not converted
with such an impressive enantioselectivity some of the observed ee’s
are still higher than any reported in the literature.1 Compared to I
substrate III in which both phenyl groups are in 3-position gave only
(d) D. V. Gribkov, K. C. Hultzsch and F. Hampel, J. Am. Chem. Soc., 2006,
128, 3748–3759; (e) P. D. Knight, I. Munslow, P. N. O’Shaughnessy and
P. Scott, Chem. Commun., 2004, 894–895.
6 J. Jenter, A. Lu¨hl, P. W. Roesky and S. Blechert, J. Organomet. Chem.,
2011, 696, 406–418.
¨
7 M. Biyikal, K. Lohnwitz, P. W. Roesky and S. Blechert, Synlett, 2008, 3106–3110.
8 (a) D. C. Leitch, R. H. Platel and L. L. Schafer, J. Am. Chem. Soc., 2011, 133,
15453–15463; (b) K. Born and S. Doye, Eur. J. Org. Chem., 2012, 764–771.
68% ee (entry 2). Using substrate V which has a similar scaffold than I 9 S. Matsunaga and M. Shibasaki, Chem. Commun., 2014, 50, 1044–1057.
3864 | Chem. Commun., 2014, 50, 3862--3864
This journal is ©The Royal Society of Chemistry 2014