Chiral calcium catalysts for asymmetric hydroamination/cyclisation

Article abstract of DOI:10.1039/c1cc11229e

Calcium complexes supported by chiral 1,2-diamines have been shown to be efficient catalysts for the asymmetric hydroamination of amino-olefin substrates; the calcium complexes [Ca(NN^R){N(SiMe₃) ₂}(THF)] (R = ^tBu, ⁱPr, Ph, 4-C ₆H₄F) give enantioselectivities of up to 26% which marks a significant increase based upon literature precedence. The structure of [Ca(NN^Ph){N(SiMe₃)₂}(py)] has been computed with density functional methods.

Full text of DOI:10.1039/c1cc11229e

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2011, 47, 5449–5451
www.rsc.org/chemcomm
COMMUNICATION
Chiral calcium catalysts for asymmetric hydroamination/cyclisationw
James S. Wixey and Benjamin D. Ward*
Received 1st March 2011, Accepted 22nd March 2011
DOI: 10.1039/c1cc11229e
Calcium complexes supported by chiral 1,2-diamines have
been shown to be efficient catalysts for the asymmetric hydro-
amination of amino-olefin substrates; the calcium complexes
[Ca(NNR){N(SiMe3)2}(THF)] (R = Bu, Pr, Ph, 4-C6H4F)
give enantioselectivities of up to 26% which marks a significant
increase based upon literature precedence. The structure of
[Ca(NNPh){N(SiMe3)2}(py)] has been computed with density
functional methods.
hydrogenation,6 and ring opening polymerisation3,7 catalysis.
Despite the significant advances in this area, the use of
chiral ligands to effect asymmetric catalysis remains much
less developed, particularly in complexes in which strongly
basic amide- or alkyl co-ligands are present.8 Recently
Buch and Harder prepared calcium amide complexes
supported by the ubiquitous bisoxazoline (BOX) ligands,
[Ca(BOX){N(SiMe3)2}(THF)] (Fig. 1b).9 These complexes
were screened in asymmetric hydroamination and hydrosilylation
catalysis; despite being able to prepare and structurally
characterise these complexes, ligand redistribution proved to
be rapid in solution, reforming the achiral homoleptic complex
Ca{N(SiMe3)2}(THF)2 alongside the catalytically inactive
[Ca(BOX)2], thus rendering the enantioselectivities low
(5–10% ee). Such observations have led to the hypothesis
that ligand redistribution must be completely suppressed in
order to generate acceptable enantioselectivities. Even more
recently, Sadow et al. have reported somewhat higher enantio-
selectivities when using trisoxazoline magnesium and calcium
complexes,10 with the highest enantioselectivities being obtained
with the magnesium complexes at elevated temperatures. It is
in this context that we have sought to prepare simple and
readily available chiral ligands that provide greater stereocontrol
in hydroamination catalysis with calcium.
t
i
Whilst the coordination chemistry of the alkaline earth (AE)
metals is well established, the development of well-defined
complexes which can be successfully adapted to catalysis has
only become apparent in the last few years.1 The primary
reason behind this late development of AE metals in catalysis
is their propensity to undergo rapid Schlenk-type ligand
redistribution processes,2 thus rendering catalytically active
species either inactive, or else removing the possibility of regio-
or stereoselectivity. As such, the rich chemistry associated with
the AE metals has been somewhat neglected outside the
context of stoichiometric reagents.
Advances in the last five years have demonstrated that
a further understanding of their coordination chemistry is
likely to propel the catalytic applications of the AE metals,
particularly calcium, into new and exciting directions. In this
regard, remarkable advances in the area have been made,
particularly with b-diketiminato ligands which have proven
remarkably successful in supporting well-defined calcium
complexes (Fig. 1a).3 These, and related, AE complexes have
been shown to be active in hydroamination,4 hydrosilylation,5
A
series of chiral ethylene diamine derivatives
t
i
H2NCH(iPr)CH2NHR (HNNR, R = Bu 1a, Pr 1b, Ph 1c
and 4-C6H4F 1d) have been prepared from L-valine according
to Scheme 1. This route utilises the amidation of an acid
chloride with a series of primary amines or anilines. We have
found that using thionyl chloride is the most convenient
method of converting phthalimide-protected valine into the
acid chloride, and does not give rise to epimerisation of the
chiral acid; the protected valinoyl chloride can be readily
prepared on a 100 g scale and is therefore convenient for the
preparation of large quantities of diamines. The acid chloride
can be converted into an N-substituted amide by adding a
primary amine or aniline in THF. The addition of triethylamine
allows this reaction to proceed under ambient conditions,
although the yields are often much improved with the addition
of catalytic N,N-dimethylaminopyridine (DMAP). The reduction
of the amides to afford diamines was effected on a small scale
with BH3(THF) (using commercial 1 M solution). However,
on increasing the scale of the reaction, this required the use of
excessively large quantities of solvent, causing problems with
heat transfer and unnecessary dilution. The reaction was less
effective with BH3(SMe2) (10 M), whilst this enabled the
Fig. 1 Calcium complexes used in hydroamination catalysis.
School of Chemistry, Cardiff University, Main Building, Park Place,
Cardiff CF10 3AT, UK. E-mail: WardBD@Cardiff.ac.uk;
Fax: +44 (0)29 208 74030; Tel: +44 (0)29 208 70302
w Electronic supplementary information (ESI) available: Experimental
procedures, characterising data, and computational methods. See
DOI: 10.1039/c1cc11229e
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 5449–5451 5449

View Article Online
structure) for the diamine backbone protons. The presence of
a single bis(trimethylsilyl)amide moiety was also evident from
the relative integration of the signal at ca. 0 ppm. The fluxional
process, presumably arising from the lability of the NNR
ligands, could not be completely ‘‘frozen out’’ using low
temperature NMR analyses, within our instrumental capability
(500 MHz, ꢀ90 1C), however NMR tube scale reactions
carried out in situ demonstrated the elimination of one molar
equivalent of HN(SiMe3)2, thus supporting the formation of a
heteroleptic complex containing one deprotonated diamine
moiety and one remaining bis(trimethylsilyl)amide ligand.
Most noteworthy was the relative intensity of the pyridine,
coordinated to the calcium centre. In each case, the complexes
were dried in vacuo to 4 ꢁ 10ꢀ2 mbar, in order to ensure a
consistent level of base coordination. Remarkably, when the
alkyl-substituted diamines were employed (2a and 2b), the
number of pyridine ligands was less reproducible, ranging
from 0.3–0.6. This suggests that the pyridine is only weakly
bound to the calcium, and so we attempted to prepare the
base-free complex [Ca(NNR){N(SiMe3)2}] by subjecting the
complex to vacuum overnight, but in all cases we were unable
to completely remove the pyridine. In contrast, approximately
one molar equivalent of coordinated pyridine was observed in
complexes 2c and 2d, which decrease less after extended
periods in vacuo. It is unclear as to the observed differences
in pyridine coordination between alkyl and aryl systems,
although we tentatively suggest that the difference may lie in
the difference in steric constraints imposed by the different
N-substituents. Given the different degrees to which pyridine
coordinates to the calcium and the extremity of the NNR
lability, we cannot categorically rule out the possibility that
complexes 2a–2d, existing as dimers/oligomers; we suggest that
the actual structure could be a complex mixture of monomeric
and oligomeric species. Such a phenomenon is not unexpected,
given the propensity of Grignard reagents to undergo such
redistribution processes.2 Attempts to crystallographically
characterise complexes 2a–2d were unsuccessful, yielding only
redistributed calcium species without the NNR ligand.
Scheme 1 Synthesis of chiral diamines 1a–d.
reaction to be scaled up safely, these conditions resulted in
significantly lower yields, in addition to by-products arising
from the degradation of the THF solvent. Conversely, the use
of NaBH4/Me3SiCl was found to give excellent yields,11 with
much more control over the reaction concentration; this
procedure has been demonstrated to be successful on a scale
of up to 12 g in 250 ml solvent. This simple multistep method
allows a series of diamines (HNNR) to be prepared in which
the R substituent can be varied in a modular fashion, without
introducing additional synthetic complexity.
Reaction of the protio-ligands 1a–d with Ca{N(SiMe3)2}2(THF)2
in C6D6 afforded pale yellow precipitates, which although are
suitable precursors for hydroamination catalysis (vide infra),
were difficult to unambiguously characterise since they
were found to be extremely labile in THF-d8, but are insoluble
in hydrocarbon solvents and C6D5Cl, and unstable in
CDCl3 and CD2Cl2. In order to determine the extent of
base coordination, the complexes were prepared using the
pyridine analogue Ca{N(SiMe3)2}2(py)2 in the presence of an
excess of pyridine.12 These complexes are formulated as
[Ca(NNR){N(SiMe3)2}(py)n] (2a–d) (Scheme 2).
When Ca{N(SiMe3)2}2(py)2 was reacted with two equivalents
of HNNR, two equivalents of HN(SiMe3)2 were eliminated,
t
forming the homoleptic complex [Ca(NNR)2] (R = Bu 3a or
4-C6H4F 3b). Again, the NMR spectra were broad at both
ambient and low temperatures. Although we cannot completely
rule out the possibility that the homoleptic complexes
[Ca(NNR)2] form in solution from the redistribution of ligands
in 2, it is unlikely that this represents a major component since
complexes 2 were found to be catalytically active in the hydro-
amination/cyclisation of amino-olefins, whereas 3a and 3b were
found to be completely inactive.
As with the THF complexes, the pyridine adducts were
found to be only sparingly soluble in benzene and toluene, and
decomposed in chloroform and dichloromethane. The NMR
spectra in THF-d8 were broad at ambient temperature, but
nevertheless showed three clear signals (albeit without fine
In order to gain insight into the structure of the proposed
mononuclear species, and especially to determine which of the
Scheme 2 Preparation of [Ca(NNR){N(SiMe3)2}(Py)n] 2a–d.
c
This journal is The Royal Society of Chemistry 2011
5450 Chem. Commun., 2011, 47, 5449–5451

View Article Online
which the chiral pyrrolidine is formed in 26% ee. This level of
enantioselectivity represents a remarkable increase from the
5–10% obtained with calcium BOX complexes8 and is also
higher than the calcium complexes reported by Sadow et al.10
It is thus far unclear as to why the enantioselectivity in this case
far exceeds the other diamine–substrate combinations tested,
however the phenyl-substituted diamine/phenyl-substituted
substrate combination suggests that p–p stacking may be an
important feature in one of the intermediate steps of the catalytic
cycle. Such interactions may well explain the lack of selectivity in
the entries 2 and 4; as well as the lack of selectivity in any entry
involving the methyl-substituted substrate A. It is clear however
that further investigations into the detailed mechanism are
warranted, for such studies will undoubtedly facilitate further
developments in stereoselective catalysis with calcium complexes.
Chiral 1,2-diamines have been shown to be efficient stereo-
directing ligands in the calcium-catalysed intramolecular
hydroamination of amino-olefins. Whilst there is clearly room
for improvement, the enantioselectivities described herein
represent a significant advance in calcium-mediated stereo-
selective catalysis. Further work in this area, particularly to
probe the structure of these complexes, is currently underway.
We thank the EPSRC (EP/H012109) and the Royal Society
for financial support. Assistance by Dr R. L. Jenkins with
NMR spectroscopic measurements is gratefully acknowledged.
Fig. 2 Calculated structure of [Ca(NNPh){N(SiMe3)2}(py)] 2ccalc
.
amine moieties is likely to be preferentially deprotonated in the
NNR ligands, the structure of the [Ca(NNPh){N(SiMe3)2}(py)]
2ccalc was calculated using density functional methods.w The
computed structure is displayed in Fig. 2. The two possible
structures, with the primary and secondary amines deproto-
nated, were calculated, however the structure with the secondary
amine deprotonated, as shown in Fig. 2, was found to be
significantly more stable, by ca. 13 kcal. In addition, the
structure indicates a significant deviation of the amido nitrogen
from the expected trigonal planar geometry; the consequence is
that the phenyl ring is orientated so as to provide a chiral
environment at the calcium centre, which may well explain the
effective stereocontrol in hydroamination catalysis.
Notes and references
Complexes 2a–2d were tested as precatalysts for the enantio-
selective hydroamination of the amino-olefin substrates A and B
(Scheme 3). The activities and enantioselectivities of these cata-
lytic reactionsz are summarised in Table 1. For each of the
ligands 2a–2c the reaction proceeded significantly slower than
with the homoleptic bis amide complex Ca{N(SiMe3)2}-
(THF)2 (23 h for A and 20 min for B at ambient temperature),
as is expected when adding a sterically demanding spectator
ligand. Interestingly, when the para-fluorophenyl ligand 2d was
employed, no activity was observed at all over a two week period
(entries 7 and 8). The same was true of ligand 2c, although only
with the dimethyl-substrate A (entry 5). Most noteworthy is
catalyst 2c (phenyl N-substituent) with substrate B (entry 6), in
z 70 mmol [Ca(NNR){(N(SiMe3)2}(THF)], 0.7 mmol substrate, 0.5 ml
C6D6, rt. Reaction progress monitored by 1H NMR spectroscopy.
Enantiomeric excesses determined by 1H NMR spectroscopy after the
addition of R-(ꢀ)-O-acetylmandelic acid. See ref. 13.
1 Recent reviews: (a) M. Westerhausen, Z. Anorg. Allg. Chem., 2009,
635, 13; (b) S. Harder, Chem. Rev., 2010, 110, 3852;
(c) M. Westerhausen, Coord. Chem. Rev., 2008, 252, 1516;
(d) J. D. Smith, Angew. Chem., Int. Ed., 2009, 48, 6597.
2 D. Seyferth, Organometallics, 2009, 28, 1598.
3 M. H. Chisholm, J. Gallucci and K. Phomphrai, Inorg. Chem.,
2004, 43, 6717.
4 (a) M. R. Crimmin, I. J. Casely and M. S. Hill, J. Am. Chem. Soc.,
2005, 127, 2042; (b) S. Datta, H. W. Roesky and S. Blechert,
Organometallics, 2007, 26, 4392; (c) A. G. M. Barrett,
M. R. Crimmin, M. S. Hill, P. B. Hitchcock, G. Kociok-Kohn
¨
and P. A. Procopiou, Inorg. Chem., 2008, 47, 7366; (d) J. R. Lachs,
A. G. M. Barrett, M. R. Crimmin, G. Kociok-Kohn, M. S. Hill,
¨
M. F. Mahon and P. A. Procopiou, Eur. J. Inorg. Chem., 2008, 4173;
(e) A. G. M. Barrett, C. Brinkmann, M. R. Crimmin, M. S. Hill,
P. A. Hunt and P. A. Procopiou, J. Am. Chem. Soc., 2009,
131, 12906; (f) M. Arrowsmith, M. S. Hill and G. Kociok-Kohn,
¨
Organometallics, 2009, 28, 1730; (g) P. Horrillo-Martinez and
K. C. Hultzsch, Tetrahedron Lett., 2009, 50, 2054.
Scheme 3 Hydroamination catalysis (R0 = Me A or Ph B).
5 J. Spielmann and S. Harder, Eur. J. Inorg. Chem., 2008, 1480.
6 J. Spielmann, F. Buch and S. Harder, Angew. Chem., Int. Ed.,
2008, 47, 9434.
7 M. G. Cushion and P. Mountford, Chem. Commun., 2011, 47, 2276.
8 Group 2 metals have been used in asymmetric Lewis acid catalysis:
S. Kobayashi and Y. Yamashita, Acc. Chem. Res., 2011, 44, 58.
9 F. Buch and S. Harder, Z. Naturforsch., B, 2008, 63, 169.
10 S. R. Neal, A. Ellern and A. D. Sadow, J. Organomet. Chem., 2011,
696, 228.
11 S. W. Coghlan, R. L. Giles, J. A. K. Howard, L. G. F. Patrick,
M. R. Probert, G. E. Smith and A. Whiting, J. Organomet. Chem.,
2005, 690, 4784.
12 D. C. Bradley, M. B. Hursthouse, A. A. Ibrahim, K. M. A. Malik,
M. Motevalli, R. Moseler, H. Powell, J. D. Runnacles and
¨
Table
1
Catalytic hydroamination of amino-olefins using
[Ca(N2NR){N(SiMe3)2}(THF)]z
Entry
Ligand
Substrate
Timea
Conv.a %
eeb
%
1
2
3
4
5
6
7
8
1a
1a
1b
1b
1c
1c
1d
1d
A
B
A
B
A
B
A
B
7 d
24 h
5 d
1 h
21 d
3 d
90
499
499
499
0
80
0
0
0
6
12
5
—
26
—
—
14 d
14 d
a
Determined from 1H NMR spectra when no further conversion
A. C. Sullivan, Polyhedron, 1990, 9, 2959.
13 G. Zi, F. Zhang, L. Xiang, Y. Chen, W. Fang and H. Song, Dalton
Trans., 2010, 39, 4048.
observed. Determined by 1HNMRusingR-(ꢀ)-O-acetylmandelicacid.13
b
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 5449–5451 5451