Stereoselective and Stereospecific Reactions of Cobalt Sandwich Complexes: Synthesis of a New Class of Single Enantiomer Bulky Planar Chiral P?N and P?P Ligands

Article abstract of DOI:10.1002/chem.201705113

Starting from (η⁵-acetylcyclopentadienyl)(η⁴-tetraphenylcyclobutadiene)cobalt(I), highly enantioselective (99 % ee) (S)-CBS catalysed ketone reduction followed by stereospecific alcohol-azide exchange, azide reduction and dimethyllation gave (R)-(η⁵-α-N,N-dimethylaminoethylcyclopentadienyl)(η⁴-tetraphenylcyclobutadiene) cobalt(I) (Arthurs’ amine). This underwent highly diastereoselective cyclopalladation to give di-μ-acetate-bis-(R)-[(η⁵-(S_p)-2-(α-N,N-dimethylaminoethyl)cyclopentadienyl, 1-C, N)(η⁴-tetraphenylcyclobutadiene)cobalt(I)]dipalladium, and highly diastereoselective lithiation to give (R)-(η⁵-(S_p)-1-(α-N,N-dimethylaminoethyl)-2-(diphenylphosphino)cyclopentadienyl)(η⁴-tetraphenylcyclobutadiene)cobalt(I) (PPCA) following the addition as electrophile of chlorodiphenylphosphine. This PN-ligand was converted into (R)-(η⁵-(S_p)-1-(α-dicyclohexylphosphinoethyl)-2-(diphenylphosphino)cyclopentadienyl)(η⁴-tetraphenylcyclobutadiene)cobalt(I), a PP-ligand (Rossiphos), by stereospecific amine-phosphine exchange using HPCy₂. These air-stable P?N and P?P complexes are the first examples of a new class of bulky planar chiral ligands for application in asymmetric catalysis.

Full text of DOI:10.1002/chem.201705113

A Journal of
Accepted Article
Title: Stereoselective and Stereospecific Reactions of Cobalt Sandwich
Complexes. Application to the Synthesis of a New Class of Single
Enantiomer Bulky Planar Chiral P-N and P-P Ligands.
Authors: Christopher John Richards, Ross Arthurs, Peter Horton, and
Simon Coles
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201705113
Link to VoR: http://dx.doi.org/10.1002/chem.201705113
Supported by

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
Stereoselective and Stereospecific Reactions of Cobalt Sandwich
Complexes. Application to the Synthesis of a New Class of Single
Enantiomer Bulky Planar Chiral P-N and P-P Ligands.
Ross A. Arthurs,[a] Peter N. Horton,[b] Simon J. Coles[b] and Christopher J. Richards*[a]
Dedication ((optional))
Abstract: Starting from (η5-acetylcyclopentadienyl)(η4-tetraphenyl-
cyclobutadiene)cobalt(I), highly enantioselective (99% ee) (S)-CBS
catalysed ketone reduction followed by stereospecific alcohol-azide
exchange, azide reduction and dimethyllation gave (R)-(η5-α-N,N-
dimethylaminoethylcyclopentadienyl)(η4-tetraphenylcyclobutadiene)
cobalt(I) (Arthurs’ amine). This underwent highly diastereoselective
cyclopalladation to give di-μ-acetate-bis-(R)-[(η5-(Sp)-2-(α-N,N-
dimethylaminoethyl)cyclopentadienyl, 1-C, N)(η4-tetraphenylcyclo-
butadiene)cobalt(I)]dipalladium, and highly diastereo-selective
lithiation, such that subsequent addition of an electrophile
generates a product as predominantly a single diastereoisomer
(Scheme 1).6,7 In addition, the dimethylamino group may be
replaced readily by a nucleophile in a stereospecific substitution
reaction,6 and notable applications of this chemistry include the
synthesis of bisphosphine ligands that are commercially
available and applied widely in asymmetric catalysis.8
Another class of organometallic complexes that broadly
meet the above criteria are derived from (η5-cyclopenta-
dienyl)(η4-tetraphenylcyclobutadiene)cobalt(I).9 Planar chiral
derivatives of this parent sandwich complex have been applied
successfully in asymmetric catalysis,10 and in these cases the
size of the η4-tetraphenylcyclobutadiene moiety is a key factor
leading to high enantioselectivity. However, methodology to
generate planar chiral derivatives remains very limited,11 and
amine 2, a complex from which a wide variety of derivatives
could potentially be synthesised, has to the best of our
knowledge not been reported.
lithiation
to give
(R)-(η5-(Sp)-1-(α-N,N-dimethylaminoethyl)-2-
(diphenylphosphino)cyclopentadienyl)(η4-tetraphenylcyclobutadi-
ene)cobalt(I) (PPCA) following the addition as electrophile of
chlorodiphenylphosphine. This PN-ligand was converted into (R)-(η5-
(Sp)-1-(α-dicyclohexylphosphinoethyl)-2-(diphenylphosphino)cyclo-
pentadienyl)(η4-tetraphenylcyclobutadiene)cobalt(I),
a
PP-ligand
(Rossiphos), by stereospecific amine-phosphine exchange using
HPCy2. These air-stable P-N and P-P complexes are the first
examples of a new class of bulky planar chiral ligands for
application in asymmetric catalysis.
In this paper we describe accessible methodology for the
asymmetric synthesis of 2, and the use of this complex as a key
building
block
to
synthesise
(η5-cyclopentadienyl)(η4-
tetraphenylcyclobutadiene)cobalt(I) complexes displaying both
planar and central chirality. This methodology demonstrates that
highly diastereoselective metallation and stereospecific
substitution are also applicable to these cobalt sandwich
complexes, providing access to a new class of air-stable chiral
organometallics. This is illustrated with the synthesis of new
bulky bidentate ligands for application in asymmetric catalysis.
Introduction
Chiral non-racemic organometallic complexes are employed
extensively in several areas of investigation that include
asymmetric synthesis,1 materials chemistry,2 supramolecular
chemisty,3 and the discovery of metal-based bioactive
compounds.4 This is especially the case if the complexes are air-
stable, are based on inexpensive earth-abundant metals, and
are available readily in enantiopure form. Ferrocene complexes
generally meet these criteria, not least because differentially
disubstituted derivatives display planar chirality. Several
methods have been devised to synthesise planar chiral
ferrocene derivatives,5 the most significant of these being the
use of Ugi’s amine 1. This undergoes highly diastereoselective
[a]
[b]
Dr R. A. Arthurs, Dr C. J. Richards
School of Chemistry
University of East Anglia
Norwich Research Park, Norwich, NR4 7TJ, U.K.
E-mail: Chris.Richards@uea.ac.uk
Dr. P. N. Horton, Prof. S. J. Coles
EPSRC National Crystallography Service, School of Chemistry,
University of Southampton,
Highfield, Southampton, SO17 1BJ, U.K.
Scheme 1. Known and new methods for the asymmetric synthesis of air-
stable sandwich complexes displaying both central and planar chirality.
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
Results and Discussion
heating at 70 °C.17 Acetylation of (rac)-5 was also essentially
quantitative, but subsequent reaction of the product (rac)-6 with
aqueous dimethylamine in methanol (the conditions used for the
synthesis of 1), gave only a low yield of the new amine (rac)-2.
The major by-product was identified as methyl ether (rac)-7, and
following isolation, this was shown not to convert to the amine
under the reaction conditions. Varying the solvent, temperature
and the source of dimethylamine (either a commercial MeOH or
THF solution) revealed that a protic solvent is required for this
reaction to take place, but the solubility of the acetate (rac)-6 in
alcohols is poor.17 The best result was obtained by the use of a
THF solution of dimethylamine in methanol at 66 °C which led to
amine (rac)-2 in a yield of 61%. Under these conditions however
the ether rac-7 is still formed in significant quantities.
The
complex
(η5-cyclopentadienyl)(η4-tetraphenylcyclo-
butadiene)cobalt(I) is readily synthesised in one step from
commercially available starting materials.12 This is in contrast to
pentaphenylferrocene which requires a non-trivial multistep
synthesis,13 and this difference is a key factor in our choosing
the former as the basis for generating bulky complexes and
ligands displaying planar chirality for use in asymmetric catalysis.
A disadvantage, however, of the cobalt over the iron-based
chemistry
is
that
(η5-cyclopentadienyl)(η4-tetraphenyl-
cyclobutadiene)cobalt(I) cannot be directly functionalised by
Friedel-Crafts acylation,14 and lithiation as
a means of
monofunctioalisation only proceeds to a limited extent.11c
Scheme 2. Routes to the synthesis of ketone 3.
As this work required initially an expedient synthesis of
ketone 3, from which to investigate the asymmetric synthesis of
amine 2, the first approach revisited the literature procedure for
the synthesis of 3 involving in situ generation of sodium
acetylcyclopentadienylide followed by complexation [Scheme 2
– Route (a)].15 By changing methyl acetate for ethyl acetate the
previously reported yield of 26% was improved to a more
reasonable 48%. As an alternative procedure, readily
synthesised carboxylic acid 412b was converted into its
corresponding acid chloride followed by addition of one
equivalent of methylmagnesium iodide [Scheme 2 – Route (b)].
This gave ketone 3 in 65% yield without competitive formation of
the tertiary alcohol resulting from the further reaction of 3 with
the Grignard reagent.
Scheme 3. Preliminary investigations into the synthesis of racemic Arthurs’
amine 2.
We chose first to identify an efficient route for the synthesis
of racemic 2 as the basis of a subsequent asymmetric synthesis.
The first reported asymmetric synthesis of Ugi’s amine 1 (it
having been obtained previously as a single enantiomer by
resolution6) involved asymmetric reduction of acetylferrocene,
acetylation of the resulting secondary alcohol, and substitution
with dimethylamine.16 Thus in the same vein ketone 3 was
reduced to (rac)-5 essentially quantitatively using lithium
aluminium hydride (Scheme 3). Use instead of sodium
borohydride in ethanol gave (rac)-5 in a yield of only 48%,
together with a number of byproducts, following prolonged
Scheme 4. Further investigations into the synthesis of racemic Arthurs’ amine
2.
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
As the direct nucleophilic introduction of the dimethylamino
moiety was less than satisfactory, we sought to generate amine
2 by an alternative route. This first involved the introduction of
nitrogen by formation of oxime 8 (Scheme 4), a compound
isolated as a mixture of E and Z isomers. Recrystallisation gave
HPLC (Scheme 5). Use of 5 mol% of the catalyst resulted in a
reduction of the ee to 49%. The absolute configuration was
assigned initially by comparison to literature precedent,21,22 and
was later confirmed by X-ray crystallography (vide infra). As an
initial investigation into the stereochemical integrity of α-
substitution (R)-5 was stirred with 10% acetic acid in methanol
which resulted in formation of methyl ether (R)-7 [Method (a)],
the absolute configuration of which was confirmed following
independent synthesis by oxygen methylation [Method (b)]. Both
samples of (R)-7 were obtained in 99% ee as determined by
chiral HPLC revealing that methanolysis under acidic conditions
proceeds with complete overall retention of configuration via an
intermediate carbenium ion 11, the latter being configurationally
stable under these reaction conditions. This was also the case
when (R)-5 was heated at 60 °C with 10% acetic acid in
methanol to give (R)-7 in 99% yield and 99% ee.17
predominantly the
E
isomer, as determined by X-ray
crystallography,17 and an examination of the material isolated
from the mother liquor by 1H NMR spectroscopy (CDCl3)
revealed a half-life for Z to E isomerisation of approximately 22 h
at room temperature.17 Reduction to primary amine (rac)-9 was
low yielding (due to the formation of unidentified side products),
but the availability of this compound enabled tertiary amine
(rac)-2 to be generated readily from a reductive amination with
formaldehyde.
In seeking a higher yielding route to (rac)-9 we carried out
the reaction of alcohol (rac)-5 with trimethylsilyl azide. On
heating in water/DMF (1 : 1) at 80 °C for 48 hours this resulted in
a 66% yield of azide (rac)-10,17 and this yield was improved to
96% by use of indium tribromide as catalyst and stirring the
reactants at room temperature in dichloromethane for just 30
minutes.18 Thus using the optimum reaction conditions the
amine (rac)-2 was generated in an overall yield of 83% from
ketone 3 using a route applicable to the asymmetric synthesis of
2.
To investigate further the facility of α-substitution via
carbenium ion 11, a competition experiment was performed
between (rac)-5 and (rac)-12 in 10% acetic acid in methanol, to
which was also added THF to ensure complete dissolution
1
(Scheme 6). After 18 h H NMR spectroscopy revealed a 64%
conversion of (rac)-5 to the corresponding methyl ether
compared to a 14% conversion for (rac)-12. Assuming that the
percentage conversion is indicative of the relative stability of the
corresponding α–carbenium ion (formed via rate-limiting loss of
water), this outcome is in agreement with a previous study
comparing a diarylferrocenylmethylium ion to a related cationic
species containing (η5-cyclopentadienyl)(η4-cyclobutadiene)-
cobalt(I). This revealed the cyclobutadienecobalt moiety of the
latter as better able to stabilse the positive charge.23
Scheme 6. Competitive methanolysis of alcohols (rac)-5 and (rac)-12.
This facility for stereospecific α-substitution was extended to
the generation of azide (R)-10 (Scheme 7) for which the
absolute configuration was confirmed by X-ray crystallography.17
Following reduction to (R)-9, subsequent methylation by
reductive amination gave the required amine (R)-2, named here
as Arthurs’ amine after its originator (RAA). Alternatively,
acetylation of (R)-5 followed by reaction of the product (R)-6 with
dimethylamine also furnished amine (R)-2 although is slightly
lower overall yield (59%, 3 steps from 3) than obtained via the
azide route (83%, 4 steps from 3). In addition, following
conversion of the primary amine (R)-9 into Boc derivative (R)-14,
subsequent reduction with lithium aluminium hydride furnished
secondary amine (R)-15, the synthesis of which could be
achieved in a higher yield without isolation of intermediate (R)-
14. Alternatively, quarternarisation of (R)-2 with methyl iodide
led to the quaternary amine (R)-16 that could in turn be
Scheme 5. Asymmetric reduction and stereospecific methanolysis.
The basis of this approach was the expectation that
substitution reactions occurring at an α-stereogenic centre would
proceed stereospecifically with retention of configuration. This is
well established where one of the substituents of a carbon-
based stereogenic centre is ferrocene,19 or another
organometallic structure20 that imparts configurational stability on
the carbenium ion resulting from loss of a nucleofuge. As the
CBS-oxazaborolidine catalyst21 has been applied successfully to
the asymmetric reduction of acetylferrocene (98% ee),22 this
was applied at 30 mol% loading in conjunction with borane
dimethylsulfide to ketone
3
resulting in an essentially
quantitative yield of (R)-5 in 99% ee, as determined by chiral
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
converted into (R)-15 by a stereospecific substitution reaction
with methylamine.
chiral HPLC were unsuccessful, therefore other methods were
devised to confirm the stereochemical integrity of these
compounds. Following monoacetylation of (R)-9, the ee of amide
(R)-17 was determined as 99% by chiral HPLC analysis.
Alternatively, DCC mediated coupling of (R)-9 with (S)-Mosher’s
acid gave amide (R,S)-18 for which the 19F NMR spectrum
contained almost exclusively a single signal (-68.61 ppm), and
only a trace (dr = 166 : 1) of the additional signal observed
following the reaction of (rac)-9 with (S)-Mosher’s acid (-68.88
ppm).24 Finally, a known NMR method to determine quickly the
ee of Ugi’s amine 125 was extended to Arthurs’ amine 2. Addition
of excess (S)-mandelic acid to a CDCl3 solution of the amine,
followed by filtration and recording of the 1H NMR spectrum,
gave clear differentiation between racemic and enantioenriched
amine samples (Figure 1). Under these conditions the
dimethylamino group signals for the two diastereoisomers
arising from (rac)-1 are separated. In contrast, for amine 2,
separation is greater for the signals arising from the methyl
group attached to the stereogenic centre. Irrespective of which
of the two methods was used for the synthesis of (R)-2, the
other enantiomer was not detected using this method.
Figure 1 1H NMR spectra (CDCl3 – ppm scale) of (a) (R)-1/(rac)-1 and (b) (R)-
2/(rac)-2 following addition of (S)-mandelic acid.
Scheme 7. Asymmetric synthesis of Arthurs’ amine (R)-2 and related primary
and secondary amines.
Unlike Ugi’s amine 1, which is an oil at room temperature,
Arthurs’ amine 2 was obtained as a crystalline solid which
enabled the X-ray crystal structure to be determined. The unit
cell contained both P and M configured propeller isomers with
only the latter represented here (Figure 2). In addition to
providing further confirmation of the absolute configuration as R,
this structure of Arthurs’ amine provided a basis for analysing
the outcomes of the diastereoselective metallation reactions that
we explored next.
Scheme 8. Derivatives of amine (R)-9 used for the determination of
enantiomeric excess.
Attempts to determine the enantiomeric excess of azide (R)-
10 and amine (R)-9 (and by extension (R)-2 and (R)-15) by
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
Pd(OAc)2
(R)-2
PhMe
110 oC, 2 h
27%
Me2
N
O
Pd
O
Co
Ph
Ph
Ph
Ph
2
LiAlD4
Et2O
RT, 2 h
99%
Na(hfacac)
H2O/acetone
RT, 16 h
99%
(R,Sp)-19
N
Me2
Pd
O
CF3
O
NMe2
D(70%)
Co
Co
Ph
Ph
Ph
Ph
Ph
Ph
CF3
Ph
Ph
(R,Rp)-2-d-2
(R,Sp)-20
Figure 2. Representation of the X-ray crystal structure of Arthurs’ amine (R)-2
(values quoted are for the represented M configured structure - hydrogen
atoms omitted for clarity). Principal bond lengths [Å] include: C(1)-C(6) =
1.519(4), C(6)-N(1) = 1.470(4), Co(1)-Cp (centre of mass) = 1.6707(14),
Co(1)-Cb (centre of mass) = 1.6930(14). Principal torsions [°] include: C(5)-
C(1)-C(6)-C(7) = -15.3(4), C(5)-C(1)-C(6)-N(1) = 109.3(3).
Scheme 9. Cyclopalladation of Arthurs’ amine (R)-2, ligand substitution and
deuterium introduction.
Cyclopalladation was investigated first as this is
metallation reaction that has been applied successfully to other
substituted derivatives of
(η5-cyclopentadienyl)(η4-
a
tetraphenylcyclobutadiene)cobalt(I) to achieve both high
diastereoselectivity11b,d,e and high enantioselectivity.10d Heating
(R)-2 with palladium acetate in toluene heated at reflux gave a
new complex containing three cyclopentadienyl signals (4.14,
Scheme 10. Lithiation followed by silylation of Arthurs’ amine 2.
1
4.08 and 4.03 ppm) in the resulting H NMR spectrum (CDCl3),
and a methyl singlet (1.60 ppm) indicative of a bridging acetate
ligand. Furthermore, the dimethylamino moiety gave rise to two
diastereotopic methyl singlets, an expected consequence of
palladium coordination.10d The identity of this new complex as
(R,Sp)-19 was confirmed following its conversion into
hexafluoroacetylacetonate (hfacac) monomer (R,Sp)-20 for
which the CD spectrum correlated closely with related Sp
configured planar chiral palladacycles.10d,11e,17 No spectroscopic
evidence was observed for the alternative Rp configured
complex. The highly selective cyclopalladation of Arthurs’ amine
2 is in contrast to the cyclopalladation of Ugi’s amine 1 which
results the same sense of selectivity, but with a diastereoisomer
ratio of just 85 : 15.26
The reaction of acetate-bridged dimer (R,Sp)-19 with lithium
aluminium deuteride resulted in replacement of palladium by
deuterium27 to give a significant quantity of (R,Rp)-2-d-2.28 The
1H NMR spectrum of the product revealed diminution (70%) of
only one of the two diastereotopic α–hydrogen signals enabling
assignment of the pro-Sp and pro-Rp hydrogens in 2 (4.58 and
4.67 ppm respectively).
Table 1. Investigations into the diastereosective lithiation of Arthurs’
amine (rac)-2 in Et2O resulting in (R*,Sp*)-21.
Entry
Base
Temp.
Time
Yield[a]
d.r. [b]
(equivalents)
1
2
3
4
5
6
n-BuLi (1.5)
n-BuLi (10)
s-BuLi (10)
s-BuLi (10)
s-BuLi (10)
s-BuLi (2)
RT
2 h
1 h
2 h
2h
7%
0%[c]
18%
52%
58%[d]
0%[e]
>99:1
-
40 oC
RT
>99:1
>99:1
>99:1
-
40 oC
40 oC
40 oC
3h
24h
[a] Following isolation by column chromatography. [b] Determined by 1H NMR
spectroscopy prior to chromatography. [c] Starting material recovered. [d]
Starting with (R)-2 to give (R,Sp)-21. [e] Observation of (η5-
vinylcyclopentadienyl)(η4-tetraphenylcyclobutadiene)cobalt(I)
in
product
mixture.
An investigation into the lithiation of Arthurs’ amine initially
used the same reaction conditions that have been employed
frequently for the diastereoselective lithiation of Ugi’s amine 1.
Thus addition of n-butyllithium to an ether solution of (rac)-2,
followed by the addition of trimethylsilyl chloride, resulted in
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
substitution of the cyclopentadienyl ring and formation of what
appeared to be a single diastereoisomer, albeit in just 7% yield
(Scheme 10, Table 1 – entry 1). The relative configuration of the
product 21 was assigned initially as R*,Sp*, by analogy with the
lithiation diastereoselectivity displayed by Ugi’s amine under
similar conditions, and the configuration of the new element of
planar chirality was later confirmed by chemical correlation and
X-ray crystallography (vide infra). Increasing the quantity of n-
butyllithium to ten equivalents with a reaction temperature of
40 °C (i.e. the reaction heated at reflux) resulted in no product
formation (entry 2), but use of ten equivalents of s-butyllithium at
room temperature resulted in an 18% yield (entry 3). This
increased to 52% on heating at 40 °C for 2 hours (entry 4), and
58% with a three hour reaction time (entry 5), the latter reaction
starting with (R)-2 to give (R,Sp)-21. In all cases only a single
diastereoisomer of the product was observed by 1H NMR
spectroscopy. Product 21 was not observed when 2 equivalents
of s-BuLi were used at a reaction temperature of 40 °C for 24
into account the percentage of (R*,Rp*)-2-d-1, and ignoring the
possible presence of (R*,Sp*)-2-d-1, the value of kH/kD is
approximately 12, i.e. less that the primary kinetic isotope effect
value of ~20 determined from similar experiments on
a
ferrocenyloxazoline,29 and significantly less than the value of
>50 observed for some non-diastereoselective lithiation
reactions.32
1
hours. Instead the H NMR spectrum of the unpurified reaction
mixture following work-up revealed the presence of a vinyl
substituted sandwich complex derived from the starting material
by an elimination reaction (entry 6). Thus it appears that the
lithiated intermediate is itself not stable for prolonged periods
under the conditions required for its formation, and a large
excess of s-BuLi is required to achieve lithiation relatively quickly.
It is of note that the conditions optimised with the use of
trimethylsilyl chloride as electrophile (58% yield) resulted in
essentially complete conversion to product with the other
electrophiles (vide infra).
Scheme 11. Use of a deuterium blocking group to lower the lithiation
diastereoselectivity of Ugi’s amine.
To bias further the reaction in favour of (R*,Rp*)-5-d-23
requires reaction conditions for the lithiation of non-deuterated
When the minor isomer from a diastereoselective reaction
can not be observed there is a degree of doubt as to the
1
Ugi’s amine that result in
a reduction in the 14 : 1
selectivity observed as, for example, the H NMR signals of the
diastereoselectivity. Several experiments were performed with
this aim, and although the selectivity could be reduced
significantly, this was always accompanied by a very low
reaction yield which precluded further application.17 This is
unfortunate as higher yielding access to what is normally the
minor diastereoisomer could be coincident with those of the
major diastereoisomer. Thus we set out to synthesise the minor
diastereoisomer using a deuterium blocking group to disfavour
the formation of the major diastereoisomer, a technique we have
employed previously to reverse the diastereoselectivity of
ferrocenyloxazoline lithiation.29 First we investigated the
application of this technique to Ugi’s amine (Scheme 11).
Addition at room temperature of n-butyllithium to an ether
solution of (rac)-1 followed by trapping with trimethylsilyl chloride
resulted in a 14 : 1 ratio of diastereoisomers (R*,Sp*)-22 and
(R*,Rp*)-23. Several repetitions of this reaction gave the same
selectivity, and it is of note that this ratio is less than the 22 : 1
ratio of isomers reported in the first description of this
chemistry.30
Adding instead deuterated methanol following lithiation gave
predominantly (R*,Rp*)-2-d-1 as 1H NMR spectroscopy revealed
approximately 90% deuterium incorporation at the 2-position.31
When this deuterated material was lithiated under the same
conditions, followed by the addition of trimethylsilyl chloride, the
crude product mixture contained a 1 : 0.72 : 0.53 ratio of starting
material to (R*,Sp*)-22 to (R*,Rp*)-5-d-23 respectively. The latter
silylated product was isolated as a single diastereoisomer by
column chromatography, albeit in low yield. Thus following
introduction of a deuterium substituent the ratio of the two
silylated diastereoisomers changed from 14 : 1 to 1.4 : 1. Taking
minor discarded diastereoisomer of
a lithiation/electrophile
quench sequence with Ugi’s amine could potentially provide
many additional ligands for testing in asymmetric catalysis.33
Furthermore, subjecting (R*,Rp*)-5-d-23 to the standard
conditions for lithiation/trimethylsilyl chloride quench gave
predominantly (R*,Rp*)-5-d-24 as a result of 1’-lithiation, with
both deuterium and the trimethylsilyl group preventing lithiation
of the substituted cyclopentadienyl ring (Scheme 12). Non-
deuterated (R*,Rp*)-24 is unavailable using existing
methodology.
Scheme 12. Use of deuterium and trimethylsilyl blocking groups to favour 1’-
lithiation.
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
product.6,35 This is represented as TS-B(i) in Figure 3, and
contrasts with TS-A(i) leading to the major diastereoisomer.
Subsequent DFT calculations support this basic model,36 and
give values for the difference between the two transition states
of between 10 and 13 kJ mol-1 depending on the degree on n-
butyllithium aggregation and solvent (Et2O) coordination to
lithium. In all of the calculated transition states for the formation
of the major diastereoisomer the methyl group of the α–
stereogenic centre is essentially in the plane of the
cyclopentadienyl ring. The X-ray crystal structure of 2 reveals a
similar arrangement (torsion angles -15.3(4) and -20.8(4)o for the
two structures in the unit cell, see Figure 2) with the methine
hydrogen pointing towards the bulk of the sandwich complex.
Thus a similar minimum energy conformation in solution requires
little change in rotation about the C(Cp)-C(α) bond to access the
transition state TS-A(ii) leading to the only observed
diastereoisomer, in accordance with the principle of least
motion.37 Furthermore, the transition state TS-B(ii) that would
lead to the minor diastereoisomer contains a more significant
repulsion than is the case for TS-B(i) due to the larger size of
the Co(η4-C4Ph4) moiety. A >99.9:0.1 diastereoselectivity as a
result of kinetic control in this way requires an energy difference
between the two transition states of >18 kJ mol-1 (at 40 oC). The
diastereoselectivity of cyclopalladation is presumably the result
of a similar transition state discrimination following amine
coordination to palladium, and participation of an acetate ligand
in a concerted metallation-deprotonation step.38
Scheme 13. Attempted replacement of the of pro-Sp hydrogen of Arthurs’
amine 2.
Arthurs’ amine (rac)-2 was lithiated as before and quenched
with deuterated methanol to give deuterated Arthurs’ amine for
1
which the H NMR signal at 4.67 ppm had vanished, and the
signal for the diastereotopic hydrogen at 4.58 ppm was
unchanged (Scheme 13). This formation of (R*,Rp*)-2-d-2
confirmed that lithiation is highly diastereoselective, and also
revealed the selectivity for the new element of planar chirality
(i.e. by comparison to (R*,Rp*)-2-d-2 previously synthesised by
cyclopalladation/deuterium replacement). A second compound
was isolated by column chromatography and identified as
deuterated (R*,Rp*)-2-d-7, the relative configuration of which
was confirmed following the quaternization of (R*,Rp*)-2-d-2 and
reaction with methanol.17 As apart from (rac)-2 no non-
deuterated methyl source was used in the reaction or workup,
the methyl group of ether (R*,Rp*)-2-d-7 is presumably derived
from the dimethylamino group of the starting material by an as
yet unknown mechanism. Application of the same lithiation
conditions to (R*,Rp*)-2-d-2 followed by addition of trimethylsilyl
chloride gave only (R*,Sp*)-21, with 1H and 13C NMR
spectroscopy revealing no indication of the presence of a
diastereoisomer in the crude product mixture. Thus even
following introduction of
a deuterium blocking group, the
transition state leading to the alternative diastereoisomer is
inaccessible, pointing to a diastereomeric ratio of >99.9 : 0.1 for
the lithiation of 2. When (R*,Sp*)-21 was subjected to the same
lithiation/silylation sequence, recovered starting material was
Figure 3. Representation of the diastereomeric transition states for α–lithiation
of amines 1 and 2.
obtained predominantly, together with
demethylated amine (R*,Sp*)-25 (ratio ~3 :1) In this instance the
unavailability of the alternative cyclopentadienyl lithiation
a
quantity of
A
number of other planar chiral compounds were
reaction allows for an alternative reaction pathway. This is in
synthesised from Arthurs’ amine (R)-2 by utilisation of the
optimised lithiation conditions (Scheme 14). Addition as
electrophile of DMF gave an excellent yield of (R,Sp)-26
displaying a single aldehyde proton signal at 9.43 ppm in the 1H
NMR spectrum. Utilising 1,2-dibromotetrachloroethane as an
electrophile led to (R,Sp)-27, albeit in a reduced yield due to
difficulties separating the product from remaining (R)-2. Use as
contrast to the analogous reaction with (R*,Sp*)-22 where the
cyclopentadienyl ring may be functionalised readily by
a
lithiation/electrophilic quench sequence.34
It has been argued that the lithiation diastereoselectivity
observed with 1 is a result of repulsion between the methyl
group of the α–stereogenic centre and the iron-cyclopentadienyl
group of ferrocene in the transition state leading to the minor
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
the electrophile of chlorodiphenylphosphine oxide or
chlorodiphenylphosphine resulted, respectively, in the formation
of phosphineoxide (R,Sp)-28 and phosphine (R,Sp)-29. The latter
was readily separated from a small quantity of (R,Sp)-28
presumably formed on work-up. In addition, the use of
chlorodicyclohexylphosphine as an alternative phosphine
electrophile yielded (R,Sp)-30. All five planar chiral products
were formed as single diastereoisomers, and for (R,Sp)-28,
(R,Sp)-29 and (R,Sp)-30 this was confirmed additionally by 31P
NMR spectroscopy.
very significant intermediate for the synthesis of several
extensively utilised ferrocene-based bis-phosphine ligands. By
analogy an acronym suggested for (R,Sp)-29 is PPCA.40 The X-
ray crystal structure of this new ligand was determined, this
further confirming both the relative and absolute configuration as
R,Sp (Figure 4). The conformation displayed by the two
cyclopentadienyl substituents in the solid state is similar to that
displayed by PPFA in the same state,39 the largest difference
being the orientation of the methyl group attached to the
stereogenic centre away from the η4-tetraphenylcyclobutadiene
moiety [C(5)-C(1)-C(6)-C(7)
=
-1.1(4)° compared to the
corresponding torsion angle of -16.32°in PPFA].
Figure 5. Ferrocene-based PPFA (R,Sp)-31 and Josiphos (R,Sp)-32 ligands,
and palladium complex (R,Sp)-33.
Scheme 14. Synthesis of planar chiral derivatives from (R)-2.
As α-substitution reactions of (η5-cylopentadienyl)(η4-
tetraphenylcyclobutadiene)cobalt(I) complexes proceed readily,
and with retention of configuration [e.g. (R)-5 into (R)-10], it was
anticipated that PPCA (R,Sp)-29 could be used for the synthesis
of a related bis-phosphine ligand. Heating (R,Sp)-29 with
dicyclohexylphosphine in acetic acid resulted in clean
conversion into a new bisphosphine formed as a single
diastereoisomer (Scheme 15), as revealed by the two signals in
the resulting 31P NMR spectrum [(MeCN-d3) at 11.90 (PCy2) and
1
-33.53 (PPh2) ppm]. It was observed that the H NMR spectrum
of (R,Sp)-34 in CDCl3 was complex, and varied from sample to
sample, an outcome attributed to residual acid present in this
solvent.
The identity of the bisphosphine as (R,Sp)-34 was confirmed
by determination of the X-ray crystal structure of (R,Sp)-35,
obtained
by
reaction
of
the
ligand
with
bis(acetonitrile)dichloropalladium(II) (Figure 6). The new
bisphosphine ligand, dubbed ‘Rossiphos’, is similar with respect
to the electronic properties of the two coordinating phosphine
moieties to the Josiphos ligand (R,Sp)-328a (Figure 5) [(CDCl3)
15.7 (PCy2) and -25.8 (PPh2) ppm]. However, comparison of the
X-ray crystal structure of (R,Sp)-35 to that of the (η3-
Figure 4. Representation of the X-ray crystal structure of (R,Sp)-PPCA [(R,Sp)-
29] (hydrogen atoms omitted for clarity). Corresponding values for PPFA in
parenthesis.39 Principal bond lengths [Å] include: C(1)-C(6) = 1.518(4) [1.480]
C(6)-N(1) = 1.468(4) [1.468], C(2)-P(1) 1.833(2) [1.803]. Principal torsions [°]
include: C(5)-C(1)-C(6)-C(7) = -1.1(4) [-16.32], C(5)-C(1)-C(6)-N(1) = 127.1(3)
[112.42], C(1)-C(2)-P(1)-C(10) 91.0(2) [89.0], C(1)-C(2)-P(1)-C(16) -168.9(2) [-
166.01], C(2)-P(1)-C(10)-C(15) 17.0(3) [12.90], C(2)-P(1)-C(16)-C(17) -65.5 [-
80.12].
allyl)palladium
complex
(R,Sp)-33,41b
which
itself
is
representative of other Josiphos transition metal complexes,41
revealed significant differences in the conformation of the
palladium coordinated ligands. Due to the size of the η4-
tetraphenylcyclobutadiene moiety the C(2)-C(1)-P(1)-C(8)
torsion angle in (R,Sp)-35 is -159.4(12)°, compared to the
corresponding angle of -107.94° in (R,Sp)-33, as a result of the
proximity of a phenyl group to the η5-cyclopentadienyl group of
ferrocene. This phenyl group in the ferrocene derivative (R,Sp)-
33 can be regarded as pseudo-axial with respect to the six-
Phosphine (R,Sp)-29 is the (η5-cyclopentadienyl)(η4-
tetraphenylcyclobutadiene)cobalt(I) equivalent to PPFA (R,Sp)-
31 (Figure 5), a chiral bidentate ligand in its own right and a
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
C(3)-C(2)-C(6)-P(2) 164.3(10), C(2)-C(1)-P(1)-C(14) 92.8(13), C(2)-C(1)-P(1)-
membered chelate ring, with the other phenyl group attached to
phosphorus being pseudo-equatorial. In the cobalt complex
(R,Sp)-35 these assignments are reversed due to the significant
conformational restriction imposed by the bulky sandwich
complex. These differences are highlighted by an overlay of the
two X-ray crystal structures (Figure 7). Other manifestations of
this restriction include the C(2)-C(6)-P(2)-Pd(1) torsion angle in
(R,Sp)-35 of 30.2(12)° compared to the corresponding torsion
angle in (R,Sp)-33 of 72.46°, and the Co-Pd distance of 4.895(2)
Å compared to the Fe-Pd distance of 5.237 Å. In complex
(R,Sp)-35 one side of the palladium-centred square-plane is
C(8) -159.4(12), C(1)-P(1)-C(14)-C(19) 3.8(14), C(1)-P(1)-C(8)-C(13)
-
89.9(12), C(2)-C(6)-P(2)-C(26) -98.2(11), C(2)-C(6)-P(2)-C(20) 151.1(10),
C(2)-C(6)-P(2)-Pd(1) 30.2(12), C(2)-C(1)-P(1)-Pd(1) -26.0(14).
The facility with which ligand (R,Sp)-34 was formed in a
reaction proceeding with complete retention of configuration is
significant as it reveals that complexes such as (R,Sp)-29
[PPCA] have the potential to be precursors to a range of ligands
based upon the sandwich complex (η5-cyclopentadienyl)(η4-
tetraphenylcyclobutadiene)cobalt(I). Thus the planar chiral
bidentate PN and PP complexes reported herein are the first
examples of a new and potentially diverse class of bulky ligands
with the potential to be applied in several areas of asymmetric
catalysis.
largely covered by
a
phenyl group of the η4-
tetraphenylcyclobutadiene moiety.
(R,Sp)-29
HPCy2
5% Ac2O/AcOH
80 oC, 3 h.
74%
Cl
Pd
Cy2
P
Cl
P
PCy2
Ph2
Ph
PdCl2(NCMe)2
PPh2
Ph
Co
Co
Ph
Ph
Ph
Ph
CH2Cl2
RT, 18 h.
57%
Ph
Ph
(R,Sp)-34
(R,Sp)-35
Scheme 15. Synthesis of ‘Rossiphos’ ligand (R,Sp)-34 and its palladium
chloride complex (R,Sp)-35.
Figure 7. Overlap of the X-ray crystal structures of palladium complexes
(R,Sp)-35 (blue) and (R,Sp)-33 (orange - counter ion removed).
Conclusions
Asymmetric synthesis methodology, that has been exploited
extensively for the synthesis of a number of widely employed
and commercially available
ferrocene-based ligands, is
applicable to the synthesis of complexes based on substituted
cyclopentadienyl derivatives of the sandwich complex (η5-
cyclopentadienyl)(η4-tetraphenylcyclobutadiene)cobalt(I).
Specifically,
the
synthesis
of
(R)-(η5-α-N,N-dimethyl-
aminoethylcyclopentadienyl)(η4-tetraphenylcyclobutadiene)
cobalt(I) (R)-2 (Arthurs’ amine), showed that CBS-catalysed
methyl ketone reduction is highly enantioselective (99% ee), that
substitution reactions at the resulting α-stereogenic centre
proceed with complete retention of configuration, and that
palladation and lithiation are highly diastereoselective (>99: 1 dr).
For the lithiation reaction, the alternative diastereoisomer could
not be detected despite attempting to exploit the high value of
Figure 6. Representation of the X-ray crystal structure of (R,Sp)-35 (hydrogen
atoms and solvent of crystallisation omitted for clarity). Principal bond lengths
[Å] include: C(2)-C(6) = 1.552(19), C(6)-P(2) 1.830(14), C(1)-P(1) 1.800(14),
P(1)-Pd(1) 2.238(4), P(2)-Pd(1) 2.296(4), Pd(1)-Cl(1) 2.354(4), Pd(1)-Cl(2)
2.364(4). Principal angles [°] include: P(1)-Pd(1)-P(2) 93.76(14), Cl(1)-Pd(1)-
Cl(2) 87.38(15). Principal torsions [°] include: C(3)-C(2)-C(6)-C(7) 34.2(17),
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
kH/kD for lithiation following introduction of a deuterium blocking
group. In contrast, application of this technique to Ugi’s amine,
the ferrocene-based analogue, changed the lithiation
diastereoselectivity from 14 : 1 to 1.4 : 1. The three key aspects
of the chemistry of (η5-cyclopentadienyl)(η4-tetraphenylcyclo-
butadiene)cobalt(I) derivatives demonstrated in this work
(enantioselective reduction, stereospecific substitution and
diastereoselective lithiation) enabled the synthesis of first in
class P-N ligand (R,Sp)-29 (PPCA) and P-P ligand (R,Sp)-34
(Rossiphos). Despite the superficial similarity of the new cobalt
sandwich complex based ligands to the equivalent ferrocene-
based structures, the X-ray crystal structure of the palladium
Swansea) and Myles Cheesman (UEA) for assistance with and
use of the UEA circular dichroism facility.
Keywords: cobalt • diastereoselectivity • enantioselectivity •
ligands • sandwich complexes
[1]
[2]
[3]
[4]
New Frontiers in Asymmetric Catalysis (Eds.: K. Mikami, M. Lautens)
Wiley,Hoboken New Jersey 2007.
Molecular Organometallic Materials for Optics (Eds: H. Le Bozec, V.
Guerchais) Springer-Verlag, Berlin Heidelberg 2010.
Supramolecular Organometallic Chemistry (Eds.: I. Haiduc, F. T.
Edelmann) Wiley-VCH, Weinheim, 2008.
Bioorganometallic Chemistry: Applications in Drug Discovery,
Biocatalysis, and Imaging (Eds.: G. Jaouen, M. Salmain) Wiley-VCH,
Weinheim, 2015.
chloride complex (R,Sp)-35 reveals both
a
significant
conformational difference, and a steric influence on palladium of
the η4-tetraphenylcyclobutadiene moiety. The impact of these
features on the effectiveness of the new ligands applied to metal
catalysed asymmetric reactions is currently under investigation.
[5]
[6]
Chiral Ferrocenes in Asymmetric Catalysis: Synthesis and Applications
(Eds.: L.-X. Dai, X.-L. Hou), Wiley-VCH, Weinheim, 2010.
D. Marquarding, H. Klusacek, G. Gokel, P. Hoffmann, I. Ugi, J. Am.
Chem. Soc. 1970, 92, 5389.
[7]
[8]
T. Hayashi, K. Yamamoto, M. Kumada, Tetrahedron Lett. 1974, 4405.
Josiphos: (a) A. Togni, C. Breutel, A. Schnyder, F. Spindler, H.
Landert, A. Tijani, J. Am. Chem. Soc. 1994, 116, 4062. BoPhoz: (b) N.
W. Boaz, S. D. Debenham, E. B. Mackenzie, S. E. Large, Org. Lett.
2002, 4, 2421. Walphos: (c) T. Sturm, W. Weissensteiner, F. Spindler,
Adv. Synth. Cat. 2003, 345, 160. TriFer: (d) W. Chen, P. J.
McCormack, K. Mohammed, W. Mbafor, S. M. Roberts, J. Whittall,
Angew. Chem. Int. Ed. 2007, 46, 4141. ChenPhos: (e) W. Chen, F.
Spindler, B. Pugin, U. Nettekoven, Angew. Chem. Int. Ed. 2013, 52,
8652. Also synthesised using similar approach is Mandyphos: (f) L.
Schwink, P. Knochel, Tetrahedron Lett. 1996, 37, 25.
Experimental Section
Preparation of (R)-(η5-(Sp)-1-(α-N,N-dimethylaminoethyl)-2-(diphenyl-
phosphino)cyclopentadienyl)(η4-tetraphenylcyclobutadiene)cobalt(I)
(R,Sp)-29 (PPCA)
(R)-2 (0.200 g, 0.36 mmol) was added to a flame dried Schlenk tube
under an inert atmosphere and dissolved in dry diethyl ether (10 mL). To
this was added sec-butyllithium (1.4 M in hexanes) (2.59 mL, 3.63 mmol)
and the sealed reaction mixture was heated in an oil bath at 40 °C with
stirring for 3 hours. The reaction was cooled to room temperature, where
upon freshly distilled chlorodiphenylphosphine (0.74 mL, 4.00 mmol) was
added and the reaction stirred for 30 mins. The reaction was then
quenched with saturated sodium hydrogen carbonate and the resulting
yellow suspension was dissolved in dichloromethane, separated with
water and dried over potassium carbonate. Column chromatography
(SiO2, 20 % EtOAc in hexane) yielded the product as a yellow crystalline
solid (0.21 g, 80 %). The sample for X-ray analysis was obtained by
[9]
A. Nakamura, N. Hagihara, Bull. Chem. Soc. Jp. 1961, 34, 452.
[10] Review: (a) H. Nomura, C. J. Richards, Chem. Asian J. 2010, 5, 1726.
See also: (b) H. V. Nguyen, D.C.D. Butler, C. J. Richards, Org. Lett.
2006, 8, 769. (c) M. Ismail, H. V. Nguyen, G. Ilyashenko, M. Motevalli,
C. J. Richards, Tetrahedron Lett. 2009, 50, 6332. (d) D. J. Cassar, G.
Ilyashenko, M. Ismail, J. Woods, D. L. Hughes, C. J. Richards, Chem.
Eur. J. 2013, 19, 17951.
[11] (a) M. Uno, K. Ando, N. Komatsuzaki, T. Tsuda, T. Tanaka, M. Sawada,
S. Takahashi, J. Organomet. Chem. 1994, 473, 303. (b) A. M. Stevens
and C. J. Richards, Organometallics 1999, 18, 1346. (c) R. G. Arrayás,
O. G. Mancheño, J. C. Carretero, Chem. Commun. 2004, 1654.(d) R. S.
Prasad, C. E. Anderson, C. J. Richards, L. E. Overman,
Organometallics 2005, 24, 77. (e) D. J. Cassar, H. Roghzai, D. Villemin,
P. N. Horton, S. J. Coles, C. J. Richards, Organometallics 2015, 34,
2953.
evaporation of the column fractions and trituration with hexane:42 Rf 0.65
(30 % EtOAc in hexane); M.p. 243 - 244 °C; [α]D
23°C
= -436 (c = 0.2,
chloroform); 1H NMR (500 MHz, CDCl3) δ = 7.25-7.21 (12H, m, m+p-
PhH), 7.18-7.15 (8H, m, o-PhH), 7.15-7.08 (6H, m, o,m+p-PhH), 7.04 (2H,
brt, J = 7.3, m-PhH), 6.74 (2H, brt, J = 7.2, o-PhH), 4.83 (1H, d, J = 1.4,
CpH), 4.75 (1H, t, J = 2.4, CpH), 4.53 (1H, brs, CpH), 3.80 (1H, dq, J =
6.7, 1.7, CH), 1.52 (6H, s, N(CH3)2), 0.51 (3H, d, J = 6.7, CH3); 13C NMR
(125 MHz, CDCl3) δ = 138.6 (ipso-PhC), 136.1 (ipso-PhC ), 136.0 (d, J =
22.7, o-PhC), 135.9 (ipso-PhC ), 132.3 (d, J = 20.5, o-PhC), 129.2 (o-
PhC), 128.4 (p-PhC), 128.1 (d, J = 7.5, m-PhC), 128.1 (m-PhC), 128.0
(p-PhC), 127.3 (d, J = 7.4, m-PhC), 126.3 (p-PhC), 107.6 (ipso-CpC),
89.6 (d, J = 17.4, ipso-CpC), 87.2 (d, J = 6.3, CpC), 84.6 (CpC), 81.8 (d,
J = 3.7, CpC), 74.5 (CbC), 56.1 (d, J = 4.6, CH), 38.7 (N(CH3)2), 7.6
(CH3); 31P NMR (202 MHz, CDCl3) δ = -32.37 (PPh2). IR (film): ṽ = 3055,
2963, 2927, 2853, 2817, 2773, 1600, 1498 cm-1; HRMS (AS) m/z calcd
for C49H44CoNP: 736.2543 [M+H]+; found 736.2543.
[12] (a) C. E. Anderson, L. E. Overman, C. J. Richards, M. P. Watson, N.
White, Org. Synth. 2007, 84, 139. (b) H. V. Nguyen, M. R. Yeamine, J.
Amin, M. Motevalli, C. J. Richards, J. Organomet. Chem. 2008, 693,
3668.
[13] (a) S. McVey, P. L. Pauson, J. Chem. Soc. 1965, 4312. (b) D. C. D.
Butler, C. J. Richards, Organometallics 2002, 21, 5433.
[14] M. D. Rausch, R. A. Genetti, J. Org. Chem. 1970, 35, 3888.
[15] Y. Ortin, K. Ahrenstorf, P O’Donohue, D. Foede, H. Müller-Bunz, P.
McArdle, A. R. Manning, M. J. McGlinchey, J. Organomet. Chem. 2004,
689, 1657.
[16] (a) N. W. Boaz, Tetrahedron Lett. 1989, 30, 2061. See also: (b) Y.
Matsumoto, A. Ohno, S.-J. Lu, T. Hayashi, N. Oguni, M. Hayashi,
Tetrehedron: Asymmetry 1993, 4, 1763. (c) M. Woltersdorf, R. Kranich,
H.-G. Schmalz, Tetrahedron 1997, 53, 7219.
Acknowledgements
[17] Complete details are provided in the supporting information.
[18] P. Vicennati, P. G. Cozzi, Eur. J. Org. Chem. 2007, 2248.
[19] (a) G. Gokel, P. Hoffmann, H. Klusacek, D. Marquarding, E. Ruch, I.
Ugi, Angew. Chem., Int. Ed. Engl. 1970, 9, 64.
The AI-Chem Channel and the EPSRC (EP/N019393/1) are
thanked for financial support. We also thank the EPSRC
National Mass Spectrometry Centre (University of Wales,
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
[20] (a) Ruthenocene: T. Hayashi, A. Ohno, S. Lu, Y. Matsumoto, E. Fukuyo,
K. Yanagi, J. Am. Chem. Soc. 1994, 116, 4221. (b) Osmocene: M. Goto,
Y. Yamazaki, K. Honda, A. Masaki, Bull. Chem. Soc. Jp. 1995, 68,
1305. (c) Cyrhetrene: C. Bolm, L. Xiao, L. Hintermann, T. Focken, G.
Raabe, Organometallics 2004, 23, 2362. (d) (η4-C4H4)Fe(CO)3: A.
Pfletschinger, U. Schneider, J. Lex, H.-G. Schmalz, Eur. J. Org. Chem.
2007, 3991. (e) (η6-C6H6)Cr(CO)3:Y. Hayashi, H. Sakai, N. Kaneta, M,
Uemura, J. Organomet. Chem. 1995, 503, 143.
[32] (a) D. Hoppe, M. Paetow. F. Hintze, Angew. Chem., Int. Ed. Engl. 1993,
32, 394. (b) J. Clayden, J. H. Pink, N. Westlund, F. X. Wilson,
Tetrahedron Lett. 1998, 39, 8377.
[33] There are a few examples of ligands derived from (R,Sp)-22 (or its
enantipode) where a SiMe3 group was used as a removal blocking
group following the introduction of PPh2 at the other position adjacent to
the α–N,N-dimethylaminoethyl substituent. See for example: (a) T.
Hayashi, T. Mise, M. Fukusima, M. Kagotani, N. Nagashima, Y.
Hamada, A. Matsumoto, S. Kawakami, M. Konishi, K. Yamamoto, M.
Kumada, Bull. Chem. Soc. Jp. 1980, 53, 1138. (b) X. Hu, H. Chen, H.
Dai, X. Hu, Z. Zheng, Tetrahedron: Asymmetry 2003, 14, 2073. (c) X.-P.
Hu, Z. Zheng, Org. Lett. 2004, 6, 3585. (d) X.-P. Hu, H.-L. Chen, Z.
Zheng, Adv. Synth. Catal. 2005, 347, 541.
[21] E. J. Corey, R. K. Bakshi, S. Shibata, J. Am. Chem. Soc. 1987, 109,
5551.
[22] J. Wright, L. Frambes, J. Organomet. Chem. 1994, 476, 215.
[23] R. Gleiter, H. Schimanke, S. J. Silverio, M. Büchner, G. Huttner,
Organometallics 1996, 15, 5635.
[24] Clear differentiation between the (R,S) and (S,S) diastereoisomers was
also observed in the 1H NMR spectrum. See supporting information for
more details.
[34] I. R. Butler, W. R. Cullen, S. J. Rettig, Organometallics 1986, 5, 1320.
[35] I. R. Butler, W. R. Cullen, F. G. Herring, N. R. Jagannathan, Can. J.
Chem. 1986, 64, 667.
[25] N. A. Kataeva, Y. K. Grishin, V. V. Dunina, Russ. Chem. Bull. 2001, 50,
1323.
[36] A. Skvorcová, R. Šebesta, Org. Biomol. Chem. 2014, 12, 132.
[37] J. Hine, Adv. Phys. Org. Chem. 1977, 15, 1.
[26] (a) V. I. Sokolov, L. L. Troitskaya, O. A. Reutov, J. Orgamomet. Chem.
1977, 133, C28 (b) T. K. Hollis, L. E. Overman, Tetrahedron Lett. 1997,
38, 8837.
[38] D. Lapointe, K. Fagnou, Chem. Lett. 2010, 39, 1118.
[39] F. W. B. Einstein, A. C. Willis, Acta Cryst. 1980, B36, 39.
[40] PPFA
=
(R)-N,N-dimethyl-1-[(S)-2-(diphenylphosphino)ferrocenyl)-
[27] M. E. Günay, C. J. Richards, Organometallics 2009, 28, 5833.
[28] In this paper, compounds that differ only by the presence or
absence of deuterium have the same number. Position 1 of the
substituted cyclopentadienyl ring of amines 1 and 2, and derivatives
thereof, is attached to the α-dimethylaminoethyl substituent.
[29] R. A. Arthurs and C. J. Richards, Org. Lett. 2017, 19, 702.
ethylamine. Thus PPCA describes by comparison ligand 29 with F
(ferrocenyl) replaced by C (cobalt).
[41] (a) C. Gambs, G. Consiglio, A. Togni, Helv. Chim. Acta 2001, 84, 3105.
(b) A. Togni, C. Breutel, M. C. Soares, N. Zanetti, T. Gerfin, V. Gramlich,
F. Spindler, G. Rihs, Inorg. Chim. Acta 1994, 222, 213.
[42] CCDC 1578881 (8), 1578882 (10), 1578883 (32), 1578884 (27) and
1578885 (2) contain the supplementary crystallographic data for this
paper. These data are provided free of charge by The Cambridge
Crystallographic Data Centre.
[30] The ratio
of
diastereoisomers
was
determined following
chromatographic separation of (R,Sp)-21 (57.3%) and (R,Rp)-22 (2.6%).
See reference 6.
[31] The percentage deuterium incorporation at the other position (i.e. of
(R*,Sp*)-2-d-1) could not be determined as this signal is obscured by
other ferrocenyl signals.
This article is protected by copyright. All rights reserved.

10.1002/chem.201705113
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Who ate all the pies? A cobalt
sandwich complex of π groups and
substituents provides a bulky
framework for the synthesis of a new
class of chiral bidentate ligands.
These result from the facility of highly
enantioselective ketone reduction,
stereospecific α–substitution
Ross A. Arthurs, Peter N. Horton, Simon
J. Coles and Christopher J. Richards*
Page No. – Page No.
Stereoselective and Stereospecific
Reactions of Cobalt Sandwich
Complexes. Application to the
Synthesis of a New Class of Single
Enantiomer Bulky Planar Chiral P-N
and P-P Ligands
reactions, and highly
diastereoselective metallation
(lithiation and palladation) to generate
the element of planar chirality.
This article is protected by copyright. All rights reserved.