Unravelling the reaction path of rhodium-monophos-catalysed olefin hydrogenation

Article abstract of DOI:10.1002/chem.201101793

The mechanism of the asymmetric hydrogenation of methyl (Z)-2-acetamidocinnamate (mac) catalysed by [Rh(MonoPhos)₂(nbd)] SbF₆ (MonoPhos: 3,5-dioxa-4-phosphacyclohepta[2,1-a:3,4-a] dinaphthalen-4-yl)dimethylamine) was elucidated by using ¹H, ³¹P and ¹⁰³Rh NMR spectroscopy and ESI-MS. The use of nbd allows one to obtain in pure form the rhodium complex that contains two units of the ligand. In contrast to the analogous complexes that contain cis,cis-1,5-cyclooctadiene (cod), this complex shows well-resolved NMR spectroscopic signals. Hydrogenation of these catalyst precursors at 1 bar total pressure gave rise to the formation of a bimetallic complex of general formula [Rh(MonoPhos)₂]₂(SbF₆)₂; no solvate complexes were detected. In the dimeric complex both rhodium atoms are ligated to two MonoPhos ligands but, in addition, each rhodium atom also binds to one of the binaphthyl rings of a ligand that is bound to the other rhodium metal. Upon addition of mac, a mixture of diastereomeric complexes [Rh(MonoPhos) ₂(mac)]SbF₆ is formed in which the substrate is bound in a chelate fashion to the metal. Upon hydrogenation, these adducts are converted into a new complex [Rh(MonoPhos)₂{mac(H)₂}]SbF₆ in which the methyl phenylalaninate mac(H)₂ is bound through its aromatic ring to rhodium. Addition of mac to this complex leads to displacement of the product by the substrate. No hydride intermediates could be detected and no evidence was found for the involvement at any stage of the process of complexes with only one coordinated MonoPhos. The collected data suggest that the asymmetric hydrogenation follows a Halpern-like mechanism in which the less abundant substrate-catalyst adduct is preferentially hydrogenated to phenylalanine methyl ester. Dimers are forever: Reaction intermediates around the catalytic cycle of the asymmetric hydrogenation of methyl (Z)-2-acetamidocinnamate by [Rh(MonoPhos)₂(nbd)]SbF₆ (nbd=bicyclo[2.2.1]hepta-2,5-diene) catalyst (see scheme) were detected by using ¹H, ³¹P and ¹⁰³Rh NMR spectroscopy and ESI-MS. The asymmetric hydrogenation appears to follow a Halpern-like mechanism.

Full text of DOI:10.1002/chem.201101793

FULL PAPER
DOI: 10.1002/chem.201101793
Unravelling the Reaction Path of Rhodium–MonoPhos-Catalysed Olefin
Hydrogenation**
Elisabetta Alberico,*[a] Wolfgang Baumann,*[b] Johannes G. de Vries,*[c]
Hans-Joachim Drexler,[b] Serafino Gladiali,[d] Detlef Heller,[b]
Huub J. W. Henderickx,[c] and Laurent Lefort[c]
Dedicated to Professor Christian Bruneau on the occasion of his 60th birthday
Abstract: The mechanism of the asym-
metric hydrogenation of methyl (Z)-2-
acetamidocinnamate (mac) catalysed
G
E
ACHUTNGTRNE(NGU MonoPhos)2AHCUTNGTREN{NUGN mac(H)2}]SbF6 in which
the methyl phenylalaninate mac(H)2 is
bound through its aromatic ring to rho-
dium. Addition of mac to this complex
leads to displacement of the product by
the substrate. No hydride intermediates
could be detected and no evidence was
found for the involvement at any stage
of the process of complexes with only
one coordinated MonoPhos. The col-
lected data suggest that the asymmetric
hydrogenation follows a Halpern-like
mechanism in which the less abundant
substrate–catalyst adduct is preferen-
tially hydrogenated to phenylalanine
methyl ester.
by [Rh
Phos:
N
2ACHTUNGTRENNUNG(nbd)]SbF6 (Mono-
ta[2,1-a:3,4-a’]dinaphthalen-4-yl)dime-
1
thylamine) was elucidated by using H,
31P and 103Rh NMR spectroscopy and
ESI-MS. The use of nbd allows one to
obtain in pure form the rhodium com-
plex that contains two units of the
ligand. In contrast to the analogous
complexes that contain cis,cis-1,5-cyclo-
octadiene (cod), this complex shows
well-resolved NMR spectroscopic sig-
nals. Hydrogenation of these catalyst
precursors at 1 bar total pressure gave
rise to the formation of a bimetallic
complex of general formula [Rh-
Introduction
Asymmetric hydrogenation is one of the most attractive
areas of enantioselective homogeneous catalysis.[1] Recently,
it has received new impetus through the discovery that
[a] Dr. E. Alberico
Istituto di Chimica Biomolecolare
Consiglio Nazionale delle Ricerche
trav. La Crucca 3, 07040 Sassari (Italy)
Fax : (+39)079 284 1299
monoACHTNUGRTNEUNG
dentate ligands[2] such as BINOL-based (BINOL=
1,1’-bi-2-naphthol) C1-symmetric phosphonites,[2c] phosphi-
tes[2d] and phosphoramidites[2e] can provide very efficient
catalysts, thus turning around the long-established belief
that only chelating bidentate ligands are properly suited for
ensuring high stereoselectivities (Scheme 1).
E-mail: Elisabetta.Alberico@cnr.it
[b] Priv.-Doz. Dr. W. Baumann, Dr. H.-J. Drexler, Prof. Dr. D. Heller
Leibniz-Institut fꢀr Katalyse e.V. an der Universitꢁt Rostock
Albert-Einstein-Strasse 29A, 18059 Rostock (Germany)
E-mail: Wolfgang.Baumann@catalysis.de
The easier and straightforward synthesis of these mono-
dentate ligands, compared to their bidentate counterparts,
[c] Prof. Dr. J. G. de Vries, H. J. W. Henderickx, Dr. L. Lefort
DSM Resolve and DSM Innovative Synthesis BV
a unit of DSM Pharma Chemicals, P.O. Box 18
6160 MD Geleen (The Netherlands)
E-mail: Hans-JG.Vries-de@dsm.com
[d] Prof. Dr. S. Gladiali
Dipartimento di Chimica, Universitꢂ di Sassari
via Vienna 2, 07100 Sassari (Italy)
[**] MonoPhos: 3,5-Dioxa-4-phosphacyclohepta[2,1-a:3,4-a’]dinaphthalen-
4-yl)dimethylamine.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201101793.
Scheme 1. Binaphthol-based monodentate P-donor ligands.
Chem. Eur. J. 2011, 17, 12683 – 12695
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12683

has opened new possibilities such as the parallel synthesis of
large ligand libraries[3] and the combinatorial screening of
heterocombinations of monodentate ligands in the search
for the best catalyst for a given substrate.[3b,4] Compared to
the state of the art reached with bidentate ligands,[5] readily
accessible and stable monodentate phosphoramidites can
lead to both higher rates and/or higher enantioselectivities
in the rhodium-catalysed asymmetric hydrogenation of a-
and b-dehydroamino acid derivatives, for which ee values
higher than 99% have been achieved (Scheme 2).[6]
studies,[8–10] most mechanistic issues are still waiting for a
convincing definition.[11] A major one is the nature of the
species that are obtained upon hydrogenation of [Rh-
A
(diene)]BF4 (diene=cis,cis-1,5-cyclooctadiene
[2.2.1]hepta-2,5-diene (nbd)). The asymmet-
AHCTUNGTRENNUNG
ric hydrogenation of a-dehydroamino acid derivatives using
rhodium with monodentate phosphoramidite, phosphite or
phosphinites ligands only proceeds with high enantioselec-
tivity if non-protic solvents are used such as CH2Cl2,
ClCH2CH2Cl, EtOAc or THF, in contrast with rhodium/bis-
phosphine catalysts that perform better in protic solvents
such as methanol. In the latter case, hydrogenation of the
catalyst precursor leads to a [Rh(bis-phosphine)ACTHNUTRGNEUNG
(MeOH)2]+
Xꢀ species, which in some cases can be isolated.[12] It re-
mains in question if the use of chlorinated solvents such as
CH2Cl2 can lead to the formation of such bis-solvato species.
Another issue is the uncertainty with regards to the number
of ligands that are ligated to rhodium during the crucial hy-
drogenation step in the catalytic cycle. There are conflicting
reports in the literature about this (vide infra). In this paper,
we report the results of a mechanistic study of the rhodium/
MonoPhos dehydroamino acid hydrogenation, which aid in
shedding some light on these challenging questions.
Scheme 2. Rh/phosphoramidite-catalysed asymmetric hydrogenation of
a-dehydroamino acid esters.
Results and Discussion
The use of Rh–monodentate phosphoramidite catalysts
has been recently commercialised by DSM: the bulky phos-
phoramidite ligand 6 (Scheme 3) in combination with a triar-
ylphosphine 7 has been used in the Rh-promoted hydroge-
nation of the a-isopropylcinnamic acid derivative 4. The
product 5 is an intermediate en route to Aliskiren, an anti-
hypertensive agent commercialised by Novartis, which rep-
resents the first example of an orally active renin inhibitor.[7]
In view of the academic and industrial interest in these
phosphoramidite-based catalysts, we set out to investigate
the mechanism of the asymmetric hydrogenation of methyl
(Z)-2-acetamidocinnamate 2a (Scheme 2) promoted by rho-
dium complexes of 1a (MonoPhos). Despite several earlier
The mechanism of asymmetric hydrogenation catalysed by
rhodium complexes with monodentate phosphorus ligands is
poorly understood, and structural information on the inter-
mediates formed in the catalytic cycle is scarce. Reetz and
co-workers,[8] by using density functional theory (DFT) cal-
culations, found that hydrogenation of itaconic acid dimeth-
yl ester with rhodium monodentate phosphite catalysts pro-
ceeds according to the lock-and-key mechanism, that is, the
prevailing enantiomer of the product originates from the
more abundant catalyst–substrate adduct formed in solu-
tion.[13]
Even if a generalisation cannot be made because the sit-
uation can be different case by case, this is in contrast with
the most credited hydrogena-
tion mechanism that involves
rhodium-chelating diarylphos-
phine catalysts in which it is the
less-abundant catalyst–substrate
adduct that, due to its higher
reactivity, leads to the prevail-
ing enantiomer of the prod-
uct.[14]
The higher lability of mono-
phosphines relative to chelating
bidentate diphosphines may in-
fluence the dissociation equili-
bria in favour of unsaturated
species, therefore one impor-
tant aspect to be addressed con-
Scheme 3. Use of a Rh–monophosphoramidite catalyst for the synthesis of an intermediate for Aliskiren.
cerns the number of monoden-
12684
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12683 – 12695

Rhodium–MonoPhos
FULL PAPER
tate ligands coordinated to rhodium during catalysis. In the
Rh-catalysed olefin hydrogenation using a monodentate
phosphoramidite derived from a chiral spirodiol, Zhou et al.
postulated the intermediacy of a rhodium substrate complex
that contained only one molecule of the ligand.[9]
groups pointing in the same direction of the coordination
plane and with two non-equivalent P donors. A recent publi-
cation of Pringle and co-workers pointed out that the obser-
vation of two isomers is due to the restricted rotation of the
two monodentate phosphorus ligands caused by the relative-
ly bulky cyclooctadiene ligand.[16] Norbornadiene is less
However, this conclusion is in contrast with the observa-
tion of non-linear effects on the stereoselectivity as observed
by the same authors with their ligand[9] and by other authors
with a different binaphthol-based monodentate phosphora-
midite.[11] Taken together with kinetic and NMR spectrosco-
py data, these facts better fit the hypothesis that two mono-
dentate phosphoramidites are bound to rhodium in the
active catalytic species.[11] This proposal seems to be con-
firmed by recent findings by Gridnev, Fan and Pringle who,
in the hydrogenation with monophosphonite–Rh catalysts,
bulky and thus in the analogous [RhP2ACTHNUTRGNEUNG(nbd)]BF4, rotation is
less hindered and hence resolved signals can be observed at
room temperature. The 31P and 103Rh NMR spectroscopic
data of these complexes are reported in Table 1 for the sake
of comparison with the species that will be discussed later in
this study.
Table 1. 31P and 103Rh NMR data of [Rh{(R)-MonoPhos}
plexes (x=1, 2).[a][15]
xACHTUNGTRENUN(NG cod)]BF4 com-
Compound
Multiplet
d
G
1J
[Hz]
E
J
E
d
(
103Rh)
ACHTUNGTRENNUNG[ppm]
identified an intermediate such as [RhHACTHNUTRGNEUNG
(alkyl)L2]+, which
A
[Hz]
featured two ligands coordinated to rhodium during the cat-
alytic cycle.[10]
[Rh
[Rh
N
E
d
141.4
139.5
133.5
137.4
239.0
242.0
239.1
231.8
–
ꢀ390.9
ꢀ374.2
G
dd
dd
d
39.6
39.2
–
In the case of the rhodium/MonoPhos (1a) system, pre-
liminary investigations by some of us have shown that when
the ligand/Rh ratio is reduced from 2 to 1, the asymmetric
hydrogenation of methyl (Z)-2-acetamidocinnamate (mac;
2a) is faster while the enantioselectivity stays the same.[11]
Although this suggests that a complex that contains one
single unit of ligand might be the active catalytic species, the
authors explained the phenomenon in terms of an equilibri-
um between different [RhLn]+, with n varying from 0 to 4.
Since the formation of inactive [RhL3]+ and [RhL4]+ is dis-
favoured at lower ligand/metal ratios, this could explain the
higher rates observed. The single-ligand hypothesis was defi-
nitely disproven by subsequent work in which Rh complexes
built up in situ from two different monodentate phosphor-
(C1 symmetrical)
[Rh(1a)2A(cod)]BF4
(C2 symmetrical)
[b]
CHTUNGTRENNUNG
[a] Spectra were recorded in CDCl3 at 228 K. [b] This species could not
be observed in the 103Rh NMR spectrum: the corresponding signal in the
31P NMR spectrum is broad, which might make the transfer of magnetisa-
tion and the detection of the signal in the 103Rh NMR spectrum more dif-
ficult.
Because of the inability to isolate pure [RhACHTNUTRGNEN(UG 1a)2ACHTUNGTNER(NUGN cod)]BF4
and in view of its broad 31P NMR spectroscopic absorptions,
the in situ prepared catalyst was not well-suited for mecha-
nistic investigations.[17] We were not able to render the syn-
thesis of the Rh–MonoPhos complexes that contained cod
more selective, but by replacing this diolefin with nbd we
succeeded for the first time in isolating a well-defined cata-
lyst precursor that contained two units of ligand.
ACHTUNGTRENNUNGamidite ligands showed better performances than the coun-
terparts built up from only one of the two phosphoramidi-
tes.[4b] The investigations on this catalytic system are ham-
pered by the fact that when [RhACHTUNTRGNEUNG(cod)2]BF4 and 2.1 equiv of
MonoPhos are mixed in dichloromethane at room tempera-
ture, several species are formed with a ligand to rhodium
ratio that ranges from 1 to 4, this last one being formed in
substantial amounts. However, it is possible to suppress the
formation of the species [RhL3]+ and [RhL4]+ by slow addi-
tion of a solution of the ligand to the solution of the metal
Synthesis of catalyst precursors: When a solution of the
ligand (R)-MonoPhos 1a (2.1 equiv) in either THF or
CH2Cl2 was added dropwise to a solution of the diolefin pre-
ꢀ
ꢀ
ꢀ
cursor [RhACHTUNGTRENGN(U nbd)2]X—with X being either BF4 , SbF6 , PF6
ꢀ
or CF3SO3 —in the same solvent at room temperature, only
one major species 8 was observed (Scheme 4).[18] Complexes
8 share the same spectroscopic features, regardless of the
counteranion. In the 31P NMR spectrum, a single signal
(only slightly broadened) is observed at about d=139 ppm,
which is a doublet as a result of the spin–spin coupling with
the 103Rh nuclide (free MonoPhos d=148.7 ppm in CDCl3).
The coupling to rhodium is comparable for the various
precursor. This allowed the isolation of [Rh
which was contaminated with small amounts of [Rh
ACHTUNGTRENNUNG
(cod)]BF4.[15] Upon 31P NMR spectroscopic analysis, two
ACHUTGTRENN(UG 1a)2ACTHUNGTRE(UNNNG cod)]BF4,
AHCTUNGTRENNUNG
species with two coordinated MonoPhos ligands were identi-
fied that slowly interconverted at room temperature and
showed broad absorptions. By lowering the temperature, it
was possible to freeze out this equilibrium and to observe
resolved peaks for each species. The two complexes are
structural isomers that differ in the relative orientation of
the two coordinated MonoPhos ligands. One of these com-
plexes is C2 symmetric with the dimethylamino groups of
the phosphoramidites pointing in opposite direction, one
above and the other below the plane defined by the metal
and the two magnetically equivalent P donors. The second
complex is C1 symmetric with both the dimethylamino
anions (1J
G
31P NMR spectrum, just one set of signals is observed in the
1H NMR spectrum for the coordinated MonoPhos ligands
(Figure 1). Within each ligand, however, the naphthyl
groups are diasterotopic, and consequently 12 signals, one
for each aromatic proton, are observable and also assignable
(see Table SI3 in the Supporting Information). Four singlets,
each accounting for two hydrogen atoms, are observed for
ꢀ
the coordinated nbd, two for the olefinic CH=, one for the
Chem. Eur. J. 2011, 17, 12683 – 12695
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12685

E. Alberico, W. Baumann, J. G. de Vries et al.
the negative-ion mode, whereas
for the cation [Rh{(R)-1a}2-
AHCTUNGTRENNUNG
(nbd)]+, a common fragmenta-
tion pattern for all complexes
in the positive-ion mode was
observed which shows the mo-
lecular ion at m/z 913, a small
peak due to loss of nbd at m/z
821 and the base peak that
arises from the loss of a mole-
cule of MonoPhos at m/z 554.
Scheme 4. Preparation of [Rh{(R)-1a}2ACTHUNRGTNE(NUG nbd)]X catalyst precursors 8.
Despite several attempts, it
was not possible to grow crys-
tals of these complexes suitable
Table 2. 31P and 103Rh NMR spectroscopy data of [Rh{(R)-1a}
complexes 8 in CDCl3.
2ACHTUNGTRNE(NUNG nbd)]X
for X-ray analysis. However, based on the NMR spectrosco-
py and ESI-MS results, it can be confidently stated that the
synthesised complexes contain two ligands coordinated to
rhodium and that they possess C2 symmetry. These com-
plexes are fairly stable in the solid state, and the NMR spec-
trum of a CDCl3 solution of the complex [Rh{(R)-1a}2]SbF6
(8a) prepared under argon remained unchanged after three
days with no other species with a different L/Rh ratio or an
unsymmetrical arrangement of the two ligands being ob-
served.
Compound Multiplet
dG
G
dAHCTUNETGRNNUNG(
103Rh) [ppm]
8a
8b
8c
8d
1a
d
d
d
d
s
138.8
138.8
138.4
138.8
148.7
251.1
249.4
253.3
251.1
ꢀ268[a]
[a] 103Rh NMR spectrum was recorded on a sample dissolved in CD2Cl2.
ꢀ
ꢀ
ꢀ
ꢀ
allylic CH and one for the bridging CH2 . Whereas the
coordinated double bonds are equivalent, within each
double bond the two hydrogen atoms are not. In the
1H,103Rh HMQC spectrum of complex [Rh{(R)-1a}2-
In addition to the expected bis-MonoPhos/Rh complex, a
minor byproduct (yieldꢂ5%) was formed, regardless of the
counterion. If no excess amount of ligand was used in the
preparation of the catalyst (2 equiv of 1a instead of 2.1), its
formation could be further minimised. A species that
showed the same spectroscopic features of the byproduct
ACHTUNGTRENNUNG(nbd)]BF4 (8b), the signal due to the rhodium(I) nucleus (d-
(
103Rh)=ꢀ268 ppm) is a triplet (1J
ACHTUNGERTN(NUNG Rh,P)=250 Hz) indica-
tive of the coupling to two equivalent phosphorus atoms
(see Figure SI8 in the Supporting Information).
was selectively prepared by mixing [RhACTHNGUETRNNU(G nbd)2]BF4 with
4 equiv of (R)-MonoPhos in dichloromethane at room tem-
perature. Crystals suitable for X-ray analysis were grown
that identified the complex as [Rh{(R)-1a}4]BF4 (9b):[19] this
is the same species that is formed in much higher amount
The stoichiometry of the [Rh{(R)-1a}2ACTHNUTRGNE(NUG nbd)]X complexes
8 was confirmed by the integrated intensities of the signals
1
in the H NMR spectrum and further corroborated by ESI-
MS. For each complex, the relative anion Xꢀ was probed in
when the catalyst is prepared from [RhACTHNGUETRNNU(G cod)2]BF4 and
Figure 1. 1H NMR spectrum of 8a recorded in CDCl3 at room temperature.
12686
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12683 – 12695

Rhodium–MonoPhos
FULL PAPER
2.1 equiv of MonoPhos 1a.[20] Crystals of 9b were tested as
catalyst in the hydrogenation of test substrate mac 2a: in di-
chloroethane at room temperature under 1 bar total pres-
sure, no conversion of the substrate was observed after 24 h.
Therefore 9b is catalytically inactive under the experimental
conditions of the previously reported kinetic investigations.
This reinforces the notion that its presence in solution could
explain the inverse dependence of the hydrogenation rate
on the MonoPhos/Rh ratio:[11] when the latter is lowered
from 2 to 1, the relative amount of the catalytic inactive spe-
cies 9 formed is reduced and the observed hydrogenation
rate increases as it is experimentally observed.[11]
Kinetic studies: Throughout this study, either complex 8a or
8b was used. These complexes are catalyst precursors from
which the active catalytic species is generated upon hydro-
genation of nbd. The catalytic hydrogenation of nbd in 1,2-
dichloroethane[21] follows the expected behaviour: a clear se-
lectivity for the hydrogenation of the first (nbd to norbor-
nene (nbe)) and the second (nbe to norbornane (nba))
double bond of the diene being observable (Figure 2).[22]
The former is represented by the first linear part of the
curve (reaction zero order in the substrate concentration),
the latter (faster) by the second part. This effect is quite
often found and can be explained with the smaller equilibri-
um constant of the nbe complex compared to that of the
nbd complex (chelate effect), which prevents the complexa-
tion of the monoene, a prerequisite for its hydrogenation, as
long as there is diene present in solution. The pseudo-first-
order rate constant is 2.03 minꢀ1 (compare the in situ cata-
Figure 2. Hydrogen-consumption curves for the catalytic hydrogenation
of norbornadiene promoted by 8a and 8b. The difference between the
curves in the latter stages of the reactions are due to a slight difference
in the amount of diene used in the two experiments.
tions, as probed by NMR spectroscopy and ESI-MS (see the
Supporting Information).
Detection of reaction intermediates under hydrogenation
conditions
Dimer: When nbd is displaced from the catalyst precursor
8a by hydrogenation in deuterated CH2Cl2 ([Rh]ꢃ0.017m),
the 31P NMR spectrum shows that the doublet due to 8a (d,
d=138.78 ppm, 1J
ACTHNUTRGNEU(GN P,Rh)=251.1; Figure 3a) is no longer
present and two sharp doublets of doublets are observed,
which is indicative of the formation of a new species that
shows two distinct phosphorus resonances that are mutually
coupled and also coupled to rhodium (d=146.9 ppm, 1J-
lyst [Rh{(S)-1a}2ACHTUNGTRENNUNG(cod)]BF4 in iPrOH for which a kobs of
0.071 minꢀ1 has been measured).[15] Most importantly, the re-
ꢀ
action rate is independent of the chosen counterion (BF4
ꢀ
and SbF6 ) of the cationic rhodium complexes, because both
A
(P,P)=53 Hz; d=145.0 ppm, 1J
ACHUTGTNRENNUG ACHTUNGTRENNUNG
provide the same intermediates under hydrogenation condi-
ACHTUNGTRENNUNG
1
AHCTUNGTRENNUNG
Figure 3. 31P NMR spectra of reaction intermediates detected during hydrogenation promoted by 8a in CD2Cl2 at 297 K: a) spectrum of 8a; b) after hy-
drogenation for nbd in 8a is complete; c) after sample degassing from H2 and the addition of 1.2 equiv of mac 2a; d) after complete hydrogenation of
mac 2a.
Chem. Eur. J. 2011, 17, 12683 – 12695
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12687

E. Alberico, W. Baumann, J. G. de Vries et al.
catalyst precursor and closely resemble the value found for
the h6-product complex 12 described later.
of the “outer” ring of the naphthyl group (see the Support-
ing Information).[26]
In the 1H NMR spectrum (Figure 4), the appearance of
characteristic high-field signals in the region d=6.0–7.0 ppm
suggests that an arene fragment of the ligand interacts with
the rhodium atom.
By analogy with previously reported species for both Rh–
diphosphines and Rh–monophosphine complexes, it is con-
cluded that a dimeric species 10 (Scheme 5) has formed
upon hydrogenation of nbd in the catalyst precursor.[12,27]
The dimeric formulation was further supported by ESI-MS:
besides a small peak at m/z 1878.7 due to {[(Rh{(R)-1a}2)2]-
To gain more insight into the coordination mode, further
analysis of the NMR spectra was accomplished by correla-
tion spectroscopy including both 1H,103Rh HMQC and
31P, 103Rh HMQC. A single absorption in the 103Rh NMR
spectrum, correlated to both the phosphorus donors, was de-
tected, thus indicating the presence of only one rhodium
species.[23] The chemical shift, d=ꢀ620 ppm (compare with
the chemical shift of d=ꢀ268 ppm for 8b),[24] is in the range
of known values for arene diphosphine Rh complexes.[25]
AHCTUNGTRENNUNG
(SbF6)}+, the base peak is observed at 821.6, this value
being obtained by dividing the molecular mass of the dimer
cation [(Rh{(R)-1a}2)2]2+ by its double charge (Table 3). The
Table 3. Selected ESI-MS results for 8a and 10.
Compound
Fragment
[M]
z
m/z
1
8a
10
[Rh{(R)-1a}2]+
[Rh{(R)-1a}2]2
821
1643
1
2
821
821.5
The observation of cross-peaks in the H,103Rh HMQC spec-
2+
trum for the aromatic protons confirmed the coordination
1
Figure 4. Comparison between the low-field portion of the H NMR spectrum of a) 8a and b) that of 10 in deuterated CH2Cl2.
Scheme 5. Intermediates detected during hydrogenation promoted by 8a.
12688
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12683 – 12695

Rhodium–MonoPhos
FULL PAPER
isotopic peak patterns at m/z 821–814 shown in Figure 5
allow us to discriminate between the cationic fragment
[Rh{(R)-1a}2]+ derived from the catalyst precursor 8a and
limits of the sensitivity of the instrument, only one species
was observed in our case at room temperature.
The dimer is very stable in CH2Cl2 within the 0.0174–
0.0043m concentration range with no indication of the for-
mation of solvato species upon dilution.[28] Nor were hydride
complexes detected in the solutions of the dimer under hy-
drogen (1 bar total pressure) at room temperature. That is,
1
no signals were detected in the H NMR spectrum at high
field, beyond d=0 up to ꢀ30 ppm.[29–31]
Rh–substrate adducts: The prochiral substrate, methyl (Z)-2-
acetamidocinnamate (2a), is known to form strong chelate
complexes when treated with various [Rh(diphos-
phine)(S)2]+ (S=solvent).[1b,14a,d,27d,32] When a 1.2-fold excess
amount of mac was added to an argon-blanketed solution
([Rh]=0.017m) of the dimer 10 at room temperature, new
signals, besides those of the dimer, appeared in the
31P NMR spectrum (Figure 3c). The new signals are broad,
yet their multiplicity is consistent with the formation of a
catalyst–substrate adduct [Rh{(R)-1a}2ACHTNUTRGNEU(GN mac)]SbF6 (11).
Under these conditions, however, the formation of 11 is
slow and a gentle heating of the sample was useful for
speeding up the reaction. The variable-temperature spectra
of 11 were run on a more concentrated solution of the
dimer ([Rh]=0.025m) that was gently heated until NMR
spectroscopic monitoring at room temperature showed that
the ratio between 10 and 11 was as high as 1:3 (Figure 6a–
d).
By lowering the temperature, the signals broaden further
and almost disappear in the background at 263 K due to an
intense exchange. Decreasing the temperature further to
221 K, a second set of signals with a similar eight line pat-
tern appears at lower field: the two sets of signals are well
resolved and the following parameters are measured: 11-
Maj d=133.0 ppm (dd, 1J
d=141.1 ppm (dd, 1J(P,Rh)=254 Hz, 2J
min d=136.9 ppm (dd, 1J(P,Rh)=257 Hz, 2J
d=148.7 ppm (dd, 1J(P,Rh)=242 Hz, 2J
(P,P)=69 Hz). At
A
ACHTUNGTREN(NUNG P,P)=62 Hz);
A
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
221 K the two species are present in approximately a 3-to-2
1
ratio. The observed line patterns and coupling constants J-
2
G
N
complexes in which it has been demonstrated that both the
C=C bond, either through the si or re face, and the oxygen
of the amido group are coordinated to rhodium in a chelate
fashion and are therefore ascribed to species 11-Maj and 11-
min (Scheme 5). The occurrence of the chelate coordination
of mac is further supported by 103Rh NMR spectroscopic
evidence (vide infra) and by the characteristic chemical shift
Figure 5. Isotopic cluster of the m/z 821.1 peak of species 8a (top spec-
trum) and superimposition of the isotopic cluster of the m/z 821.1 peak
of species 8a and the isotopic cluster of the m/z 821.6 peak of species 10
(bottom spectrum).
ꢀ
of the coordinated olefinic =CH (see Figure SI20 in the
Supporting Information).
that derived from the dimer 10: because the dimer is doubly
charged, its M+1 peak shows up as a M+0.5 peak, which is
missing in the isotopic pattern of the m/z 821.1 peak due to
the [Rh{(R)-1a}2]+ fragment derived from the catalyst pre-
cursor.
In the dimer, the phosphorus atoms that bear the bridging
naphthyls are chiral, which means that in principle different
diastereomeric dimers may exist.[27e,f] However, within the
This coordination mode renders the two MonoPhos li-
gands non-equivalent, thus giving rise to two distinct phos-
phorus resonances that are mutually coupled and also cou-
pled to rhodium.[33] The formation of the catalyst–substrate
adducts was further corroborated by the ESI-MS spectra in
which the corresponding molecular ion peak is clearly visi-
ble (see the Supporting Information). It is not possible to
Chem. Eur. J. 2011, 17, 12683 – 12695
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12689

E. Alberico, W. Baumann, J. G. de Vries et al.
chiraphos and (S,S)-dipamp,[35]
for which it is well established
that the minor adduct is the
more reactive toward hydrogen
oxidative addition. According
to von Philipsborn, the differ-
ence in dACHTUNGRNENUG(
103Rh) chemical shift
between the two diastereomers
is influenced by, among several
factors, the size of the metal va-
lence orbitals, hrꢀ3i, and this
factor decreases (r increases)
for the major diastereomer be-
cause of a better overlap be-
tween ligand and metal-based
valence orbitals. Because of the
negative sign of the paramag-
netic shielding term sp, this im-
plies a shift to lower frequency.
The increasing shielding of the
metal should therefore indicate
stronger binding of the sub-
strate, which in turn would
render the metal less reactive
Figure 6. Variable-temperature 31P NMR spectra of adducts 11 in deuterated CH2Cl2. Spectra (b–d) have been
recorded after having warmed up the samples to favour formation of 11 from dimer 10.
say why the minor isomer is not observable at room temper-
ature. It can either be present in very low concentration or
its signal may be too broad due to extensive line broaden-
ing. The two diastereomers are in dynamic equilibrium and
their signals show a different line broadening with increasing
temperature, which might be attributed to a different disso-
ciation rate.
Although at 221 K the 31P NMR spectrum of the adducts
1
11 is well resolved, the corresponding H NMR spectrum is
ill-defined due to the presence of the dimer 10 and free
mac. This hampered any attempt to clear their stereochemis-
Figure 7. 2D (31P, 103Rh{1H}) HMQC NMR spectrum (162.0 MHz) of ad-
ducts 11 in deuterated CH2Cl2 recorded at 221 K. Chemical shifts d-
try and to correlate their absolute configuration with that of
the prevailing hydrogenation product enantiomer. Elucida-
ACTHNGUTRNENUG(
103Rh) 11-Maj=324 ppm (X=3.161025 MHz); 11-min=617 ppm (X=
1
tion of the dynamic processes by either H or 31P exchange
3.161951 MHz).
spectroscopy at 221 K proved unsuccessful.[34] However, it
became evident that the ratio between major and minor dia-
stereomers changed immediately upon temperature varia-
tion (see Figure SI19 in the Supporting Information). Be-
cause of this fast interconversion, hydrogenation at low tem-
perature to assess which of the two adducts is preferentially
reduced to afford the prevailing enantiomer and whether
the hydrogenation does or does not follow the Halpern
mechanism was not attempted because its result would have
been inconclusive and even misleading.
toward adding another ligand such as hydrogen gas. This
would suggest that it is the minor diastereomer of the two
substrate complexes that is the more reactive one and that
leads to the product.
The relative stability of the dimer 10 with respect to the
catalyst-substrate adducts might explain the unexpected ki-
netic behaviour observed in the catalytic reduction of mac
with 8a. Two experiments directed towards hydrogen con-
sumption measurements have been run (see Figure SI70 in
the Supporting Information): in the first experiment, the cat-
alyst precursor was prehydrogenated (induction period) to
reduce nbd. Methyl (Z)-2-acetamidocinnamate was then
added and hydrogen consumption was monitored until com-
plete conversion of the substrate. In this case, the measured
rate should reflect the “intrinsic activity” of the catalyst as it
relates to mac hydrogenation, with no influence of the in-
The 2D (31P, 103Rh{1H}) correlation spectrum of the two ad-
ducts at 221 K shows that the d
steromer appears at d=324 ppm, at lower frequency (more
shielded) than the chemical shift d
103Rh) of the minor dia-
ACTHUNGTRNENUG(
103Rh) of the major dia-
AHCTUNGTRENNUNG(
stereomer, d=617 ppm, thereby reflecting an electronic dif-
ference at rhodium between the two species (Figure 7). The
very same trend had been observed for the analogous rhodi-
um complexes that contain the chiral diphosphines (S,S)-
12690
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12683 – 12695

Rhodium–MonoPhos
FULL PAPER
duction period necessary to displace nbd. In the second ex-
periment, no catalyst prehydrogenation was performed. All
reagents were placed in the reaction vessel and hydrogen-
consumption monitoring was started immediately after hy-
drogen gas was admitted into the reactor. Surprisingly, the
reaction rate was faster in this latter case, contrary to what
normally observed with similar rhodium-catalyst-containing
diphosphines.[22] One possible explanation might be the for-
mation of the dimer when prehydrogenation of the catalyst
is carried out in the absence of mac. The lower rate of mac
hydrogenation observed in this case might then reflect the
slow formation of the catalyst–substrate adducts 11 from the
dimer 10.
The 103Rh chemical shift d
N
103Rh) at ꢀ849 ppm is then typ-
That the observed doublet in the 31P NMR spectrum is
truly due to the catalyst–product adduct could be confirmed
by the following experiment. When the hydrogenation of
mac is carried out with [Rh{(R)-1a}2ACTHNUTRGNEUNG(nbd)]SbF6, the configu-
ration of the prevailing enantiomer of the hydrogenated
product N-acetylphenylalanine methyl ester is S. Upon addi-
tion of 1.2 equiv of the other enantiomer (R)-N-acetylphe-
nylalanine methyl ester to the above solution, a second dou-
blet appeared in the 31P NMR spectrum with a chemical
1
Rh–product adducts: Hydrogenation of the solution that
contained the catalyst–substrate adducts 11-Maj and 11-min
(1 bar total pressure) at room temperature provided the ex-
pected product N-acetylphenylalanine methyl ester (S)-
mac(H)2 in 93% ee, in accordance with published data.[11]
shift and a coupling constant (d=145.38 ppm (d, J
A
308.0 Hz)) very similar to those of the first one (Scheme 6).
The new doublet is therefore attributed to the diastereo-
meric “catalyst-product” adduct (R,R)(R)-12. The same re-
sults were obtained in separate experiments in which the
catalyst–product adducts were prepared by hydrogenation
of 8a in the presence of N-acetyl-rac-phenylalanine methyl
ester, N-acetyl-(R)-phenylalanine methyl ester and N-acetyl-
(S)-phenylalanine methyl ester, respectively. The difference
in the chemical shift of the two diastereomeric adducts, ap-
proximately d=0.5 ppm, is in line with what previously ob-
served for analogous diastereomeric adducts between {Rh-
The 31P NMR spectrum of the solution showed a doublet at
1
d=145.9 ppm with J
N
ESI mass spectrum a peak at m/z 1041.9, which is supportive
of the presence of a catalyst–product adduct [Rh{(R)-
1a}2{(S)-mac(H)2}]SbF6 (12), was observed. The product is
likely coordinated to rhodium through the phenyl ring as
suggested by the appearance of up-field resonances in the
1H NMR spectrum between d=6 and 7 ppm (Figure 8). Fur-
ther characterisation has been achieved both in dichlorome-
thane and 1,2-dichloroethane solutions (see Figures SI22–
SI25 in the Supporting Information).
ACHUNTGRENUN[G (S,S)-bdpp]ACHTUGNTRENN(UNG CD3OD)2}ClO4 (bdpp=2,4-bis-diphenylphos-
phino)pentane) and N-acetyl-(R)-phenylalanine and N-
acetyl-(S)-phenylalanine.[37] (31P and 103Rh NMR spectro-
scopic data of all detected species are collected in Table 4).
The coordinated product can be very easily displaced
from the catalyst by addition of fresh mac. Five equivalents
are sufficient to shift the equilibrium completely in favour
of the catalyst-substrate adducts 11 as confirmed by the
ESI-MS spectrum of the solution in which only the peak rel-
ative to 11 is present (see Figures SI66–SI68 in the Support-
ing Information).
Hydrogenation of a 200-fold excess amount of mac with
the same solution (final rhodium concentration [Rh]=
0.0004m) at room temperature provided the product N-
acetyl-(R)-phenylalanine methyl ester of the same optical
purity (93% ee) as in the stoichiometric hydrogenation.
1
Figure 8. Section of the H,13C HMQC NMR spectrum of the hydrogena-
Conclusion
tion solution of mac/ACHTUNGTRENNUNG 2ACHUTGNTRENN(GUN nbd)]SbF6, molar ratio 1.2, in [D4]1,2-
[RhI{(R)-1a}
dichloroethane taken after complete conversion of mac. The signals of
In this study, we have been able to observe and characterise
several Rh–phosphoramidite species involved in the asym-
metric hydrogenation of methyl (Z)-2-acetamidocinnamate
(2a). First of all, the pre-catalyst [Rh{(R)-MonoPhos}2-
the h6-coordinated reaction product in complex 12 are shown.
ꢀ
ꢀ
ꢀ
Complex 12 is coordinatively labile. At an Rh/product
ratio of 0.8, only half of the arene is coordinated and the ex-
change with the free product is detectable on the NMR
spectroscopic timescale. The nature of the remaining Rh
complexes was not established (due to strong line broaden-
ing, no other species were visible in the 31P NMR spectrum),
but the presence of the dinuclear complex 10 is most likely.
(nbd)]X 8 (with X being either BF4 , SbF6 , PF6 or
NMR spectroscopy and ESI-MS. Upon hydrogenation of
ꢀ
the catalyst precursor 8a (X=SbF6 ), a new dimeric species
10 was formed. In this species, the monodentate phosphora-
midite now plays the role of a bidentate bridging ligand and
can coordinate a second unsaturated Rh center through one
Chem. Eur. J. 2011, 17, 12683 – 12695
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12691

E. Alberico, W. Baumann, J. G. de Vries et al.
Scheme 6. 31P NMR spectra (CD2Cl2, 297 K) of catalyst–product adduct (R,R)(S)-12 (bottom) and of a mixture of (R,R)(S)-12 and (R,R)(R)-12 (top).
Table 4. Summary of 31P and 103Rh NMR spectroscopy data recorded in
[D2]CH2Cl2 for the hydrogenation intermediates reported in this study.
Rh, can be responsible for all the catalytic steps involved in
the hydrogenation. In addition to previous results (the
hetero-combination of ligand, the presence of non-linear ef-
fects), this study provides conclusive evidence that the
active species in Rh/phosphoramidite contains two mono-
dentate ligands.
Compound
Multiplet
d
G
1J
[Hz]
E
J
E
d
(
103Rh)
ACHTUNGTNER[NUNG ppm]
A
[Hz]
8a
10
d
138.9
145.0
146.9
133.0
141.1
136.9
148.7
145.9
145.3
253.1
224.6
209.8
265.0
254.0
257.0
242.0
308.0
308.0
–
(ꢀ268)[a]
ꢀ620
dd
dd
dd
dd
dd
dd
d
53.0
53.0
63.0
62.0
69.0
69.0
–
From the 103Rh NMR spectrum, the minor adduct 11-min
turns out to be the most kinetically labile of the two diaste-
reoisomers. This fact supports the conclusion that the ab-
sence of visible NMR peaks in the room temperature
1H NMR spectrum of this species is caused by an intense ex-
change rather than to a concentration effect. Since accord-
ing to Gridnev[1b] the catalyst–substrate complex that ex-
changes more intensively is the one that is hydrogenated at
a faster rate, this behaviour is indicative that the relative
ease of the hydrogenation of 11-Maj and 11-min should be
in favour of the minor adduct 11-min.
In summary, although it was not possible to assess by
direct methods the exact stereochemistry of the two Rh–
substrate diastereomers, from these arguments we can confi-
dently conclude that the major adduct 11-Maj may be the
less reactive towards the activation of H2. In which case, the
Rh/phosphoramidite system would follow an anti lock and
key mechanism, opposite to the one claimed for the similar
Rh/phosphite system.[8] Nevertheless, further studies are
needed to clarify this last point for certain.
11-Maj
11-min
324
617
T
ꢀ849
N
d
–
[a] This value refers to complex 8b.
of its naphthyl groups. The dimer is very stable in CH2Cl2
within the 0.0174–0.0043m concentration range with no indi-
cation as to the formation of solvate species upon dilution.
No hydride complexes were detected in the solutions of the
dimer under hydrogen. This is most probably due to the
high p-acidic character of the MonoPhos ligands, which con-
sequently fails to stabilise the octahedral dihydride RhIII
complex.[38]
The addition of mac 2a to the dimeric Rh species led to
the formation of two diastereomeric catalyst–substrate ad-
ducts [Rh{(R)-MonoPhos}2ACTHNUTRGNE(UNG mac)]SbF6 (11-Maj and 11-min),
which were detected and characterised by NMR spectrosco-
py and ESI-MS. The NMR spectra show that two molecules
of MonoPhos are coordinated to rhodium in these crucial
hydrogenation intermediates. The occurrence of chelate co-
ordination of mac in the adducts 11 can be deduced from
Experimental Section
ꢀ
the shift at higher field of the olefinic =CH resonance and,
All manipulations involving air- and moisture-sensitive organometallic
compounds were carried out using standard Schlenk techniques under
argon atmosphere. Dichloromethane and 1,2-dichloroethane were freshly
distilled from calcium hydride, diethyl ether and n-hexane from sodium
benzophenone ketyl. Deuterated dichloromethane [D2]CH2Cl2 and 1,2-
dicholoroethane [D4]DCE used for the detection of intermediates under
hydrogenation conditions were distilled from calcium hydride. Solutions
of crude reaction mixtures were filtered when necessary through a short
even more conclusively, from the rhodium chemical shifts
that are as expected for such a type of chelate complexes.
Addition of H2 to this catalyst–substrate adduct leads to
the formation of the hydrogenated product with the expect-
ed ee value (93%) and configuration (S). The hydrogenated
product remains bound to the metal in a new Rh-product
adduct 12 from which it can be promptly displaced upon ad-
dition of fresh mac. Consequently, we have demonstrated
that different Rh species, all of which bear two ligands per
pad of filter aid Fluka Celite 535. Catalyst precursors [{Rh
A
ꢀ
and [Rh
G
prepared according to published procedures. NMR spectra were recorded
1
using Varian Mercury Plus (400 MHz for H), Bruker AV-300 (300 MHZ
12692
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12683 – 12695

Rhodium–MonoPhos
FULL PAPER
for 1H) and Bruker AV-400 (400 MHz for 1H) spectrometers. Chemical
shifts d are reported in ppm relative to 85% H3PO4 for 31P and the sol-
vent resonance for 1H and 31C. 103Rh chemical shifts are given relative to
SI14 and SI15 in the Supporting Information). ESI-MS (HRMS, positive
ion mode): m/z calcd for C88H72N4O8P4Rh: 1539.335 [M+]; found:
1539.3391.
XACHTUNGTRENNUNG(
103Rh)=3.16 MHz. Coupling constants, J, are given in Hz. Mass spec-
Detection of reaction intermediates under hydrogenation conditions
trometric analyses were performed using an LCQ Deca XP ion trap MS
(ThermoFisher, Bremen, Germany).
Complex 10: A solution of 8a (20 mg, 0.017 mmol) in CD2Cl2 (1 mL) was
prepared in a 5 mm YOUNG NMR spectroscopy tube under argon. H2
gas (p=1 bar) was then admitted into the sample, which was intensively
shaken manually until hydrogenation at room temperature of coordinat-
ed nbd was complete. During the course of the hydrogenation, the solu-
tion turned intensively dark red. 1H and 31P NMR spectra were recorded
before and after degassing the sample from hydrogen. No signals relative
to hydride species were detected. 31P NMR (161.89 MHz, [D2]CH2Cl2,
Standard procedure for the preparation of complexes 8: A solution of
(R)-MonoPhos (0.36 mmol, 2.1 equiv) in freshly distilled THF or CH2Cl2
(8 mL) was added dropwise to
a stirred solution of {[RhACHTUNTGNERUG(N nbd)2]X}
(0.27 mmol) in freshly distilled THF or CH2Cl2 (2 mL). The mixture was
stirred at room temperature for 2 h. It was then concentrated under re-
duced pressure and the complex precipitated by addition of diethyl ether.
The solid was repeatedly washed with diethyl ether and then dried in
vacuum to give the complex as an orange-yellow solid. All complexes
show similar 1H, 13C and 31P NMR spectra. 1H and 13C NMR chemical
298 K): d=145.05 (dd, 1J
(dd, 1J
(P,Rh)=209.8 Hz,
A
ACHTUNGERTN(NUNG P,P)=52.6 Hz), 146.95 ppm
A
JACHTUNGTRENNUNG
[D2]CH2Cl2, 297 K): d=ꢀ620 ppm; ESI-MS (positive ion mode): m/z
shifts are reported only for [RhI
ACHTUNGTRENUNG(nbd){(R)-MonoPhos}2]BF4 in CDCl3 and
CD2Cl2 (for the latter solvent, see Table SI1 in the Supporting Informa-
tion).
(%): 1878.7 (14) ({[Rh{(R)-MonoPhos}2]
([(Rh{(R)-Mono-Phos}2)2]2+).
(SbF6)}+), 821.1 (100)
2ACHTUNGTRENNUNG
Complexes 11-Maj and 11-min: After removal of hydrogen through
three consequent cycles of freezing, pumping and warming, (Z)-2-acet-
ammidocinnamic methyl ester 2a (mac, 4.6 mg, 0.021 mmol, 1.2 equiv)
was added to the above prepared sample under argon. The sample was
shaken manually and gently warmed to speed up formation of the cata-
lyst substrate adducts. 31P NMR (400 MHz, [D2]CH2Cl2, 218 K): 11-Maj
Complex 8a: Isolated yield: 89%. 31P NMR (161.89 MHz, CDCl3,
298 K): d=138.30 ppm (d, 1J
ACHTUNGTRENNUNG
(84) ([RhACHTUNGTRENNUNG
554 (100) ([Rh
ACHTUNGTRENNUNG
(nbd){(R)-MonoPhos}]+), 234.9 (100) (SbF6ꢀ), 236.9 (75)
(SbF6ꢀ).
d=133.0 (dd, 1J
(P,Rh)=254.0 Hz,
257.0 Hz, J
N
ACHTUNGTRENNUNG
Complex 8b: This complex was prepared in CH2Cl2. Isolated yield:
1
ꢀ
ꢀ
E
J
E
ACHTUNGTRENNUNG
89%. H NMR (400 MHz, CDCl3, 298 K): d=1.71 (brs, 2H; CH2 nbd),
2.73 and 2.74 (overlapping d, 3J
(H,P)=5.4 Hz, 12H; NCH3), 4.11 (brs,
ꢀ
N
G
ACHTUNGTRENNUNG
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
2H; CH CH=nbd), 5.25 (brs, 2H; CH CH=nbd), 6.09 (brs, 2H;
ꢀ
CH CH=nbd), 7.00 (brd, 2H; HAr, J=7.3 Hz), 7.24–7.27 (m, 4H; HAr),
7.29–7.33 (m, 4H; HAr), 7.38 (d, J=8.8 Hz, 2H; HAr), 7.44 and 7.46 (over-
lapping dd, 1J=5.4 Hz, 2J=2.4 Hz; 1H; HAr and 1J=5.8 Hz, 2J=2.9 Hz,
1039.9 (22) ([Rh((R)-MonoPhos)2ACTHNUTRGNEUNG
(mac)]+), 821.1 (100) ([Rh((R)-Mono-
Phos)2]+).
1
2
1H; HAr), 7.52 and 7.54 (overlapping dd, J=5.4 Hz, J=2.4 Hz; 1H; HAr
and 1J=5.4 Hz, 2J=2.4 Hz, 1H; HAr), 7.86 (d, J=8.8 Hz, 2H; HAr), 7.92
(d, J=8.3 Hz, 2H; HAr), 7.98 (d, J=8.8 Hz, 2H; HAr), 8.05 ppm (d, J=
8.3 Hz, 2H; HAr); 13C NMR (100.56 MHz, CDCl3, 298 K): d=38.05 and
ꢀ
Complex ACHTNUTRGNE(NUG R,R)(S)-12: The above sample was degassed from argon by
three consequent cycles of freezing, pumping and warming. H2 gas (P=
1 bar) was then admitted into the sample, which was intensively shaken
until NMR spectroscopy showed hydrogenation of mac to N-acetylpheny-
lalanine methyl ester mac(H)2 was complete. 31P NMR (400 MHz,
38.10 (overlapping d, JACHTUNGTRENNUNG(C,P)=5.4 Hz, 2C; NCH3 and JACHTUNGTRENN(UGN C,P)=6.1 Hz,
ACTHNUTRGNEUNG
[D2]CH2Cl2, 298 K): d=145.9 ppm (d, 1J(P,Rh)=308.0 Hz); 103Rh NMR
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
2C; NCH3), 55.47 (s, 1C; CH2 nbd), 72.20 (brs, 2C; CH CH=nbd),
ꢀ
ꢀ
93.36 (brs, 2C; CH CH=nbd), 120.94 (s, 2C; CAr), 121.07 (s, 2C; CAr),
122.48 (s, 2C; CAr), 123.13 (s, 2C; CAr), 125.99 (s, 2C; CAr), 126.26 (s, 2C;
CAr), 126.95 (s, 2C; CAr), 127.15 (s, 2C; CAr), 127.26 (s, 2C; CAr), 127.50
(s, 2C; CAr), 127.58 (s, 2C; CAr), 128.82 (s, 2C; CAr), 128.96 (s, 2C; CAr),
130.97 (s, 2C; CAr), 131.69 (s, 2C; CAr), 131.78 (s, 2C; CAr), 131.91 (s, 2C;
CAr), 132.57 (s, 2C; CAr), 147.93 (brs, 2C; CAr), 148.74 ppm (brs, 2C;
CAr); 31P NMR (161.89 MHz, CDCl3, 298 K): d=138.70 ppm (d, 1J-
(12.6 MHz, [D2]CH2Cl2, 297 K): d=ꢀ849 ppm; ESI-MS: m/z (%): 1041.9
(60) ([RhI{(R)-MonoPhos} (macH2)]+), 821.1 (100) ([Rh{(R)-Mono-
2ACHNUTTGRENNUNG
Phos}2]+). GC analysis of this sample shows that it contains (S)-N-acetyl-
phenylalanine methyl ester in 93% ee.
NMR spectra, ESI-MS analyses, GC chromatograms and hydrogen con-
sumption curves are to be found in the Supporting Information.
ACHTUNGTRENNUNG
(P,Rh)=250.8 Hz); 103Rh NMR (12.6 MHz, [D2]CH2Cl2, 297 K): d=
ꢀ268 ppm; ESI-MS: m/z (%): 913 (84) ([Rh
(nbd){(R)-MonoPhos}2]+,
ACHTUNGTRENNUNG
821 (12) ([Rh{(R)-MonoPhos}2]+), 554 (100) ([Rh
ACTHNUTRGNE(NUG nbd){(R)-Mono-
Phos}]+); 87 (100) (BF4ꢀ), 86 (24.8) (BF4ꢀ).
Acknowledgements
Complex 8c: Isolated yield: 88%. 31P NMR (161.89 MHz, CDCl3,
298 K): d=138.70 (d, 1J
(P,Rh)=251.8 Hz), ꢀ144.13 ppm (J(P,F)=
(nbd){(R)-MonoPhos}2]+),
A
ACHTUNGTRENNUNG
Collaborative research carried out by S.G. and E.A. within the frame-
work of the COST D40 EU project. E.A. acknowledges partial financial
support from the Regione Autonoma della Sardegna, L.R. 7 Agosto
2007, n. 7. The skilful technical assistance of Cornelia Pribbenow of the
Leibniz-Institut fꢀr Katalyse e.V. an der Universitꢁt Rostock in running
hydrogen-consumption measurements is gratefully acknowledged.
712.6 Hz); ESI-MS: m/z (%): 913 (84) ([RhACTHUNGTRENNUNG
ACTHNUTRGNEUNG
821 (12) ([Rh{(R)-MonoPhos}2]+), 554 (100) ([Rh(nbd)(R)-MonoPhos]+
); 144.9 (100) (PF6ꢀ).
Complex 8d: Isolated yield: 86%. 31P NMR (161.89 MHz, CDCl3,
298 K): d=138.69 ppm (d, 1J
(P,Rh)=250.8 Hz); ESI-MS: m/z (%): 913
(84) ([Rh
(nbd){(R)-MonoPhos}2]+), 821 (12) ([Rh{(R)-MonoPhos}2]+),
554 (100) ([Rh
(nbd){(R)-MonoPhos}]+); 149 (100) (CF3SO3ꢀ).
Complex 9b: {[Rh(nbd)2]BF4}(35 mg, 0.093 mmol) and (R)-MonoPhos
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
[1] a) Handbook of Homogeneous Hydrogenation, (Eds.: J. G. de Vries,
C. J. Elsevier), Wiley-VCH, Weinheim, 2007; b) I. D. Gridnev, T. Im-
amoto, Chem. Commun. 2009, 7447–7464, and references therein.
[2] a) F. Lagasse, H. B. Kagan, Chem. Pharm. Bull. 2000, 48, 315–324;
b) I. V. Komarov, A. Bçrner, Angew. Chem. 2001, 113, 1237–1240;
Angew. Chem. Int. Ed. 2001, 40, 1197–1200; c) C. Claver, E. Fernan-
dez, A. Gillon, K. Heslop, D. J. Hyett, A. Martorell, A. G. Orpen,
P. G. Pringle, Chem. Commun. 2000, 961–962; d) M. T. Reetz, G.
Mehler, Angew. Chem. 2000, 112, 4047–4049; Angew. Chem. Int.
Ed. 2000, 39, 3889–3890; e) M. van den Berg, A. J. Minnaard, E. P.
Schudde, J. van Esch, A. H. M. de Vries, J. G. de Vries, B. L. Feringa,
J. Am. Chem. Soc. 2000, 122, 11539–11540.
(138 mg, 0.38 mmol, 4.1 equiv) were placed in a Schlenk flask. Freshly
distilled dichloromethane (10 mL) was added to give a clear yellow solu-
tion, which was stirred for 2 h at room temperature. The solution was
concentrated under reduced pressure and the product precipitated by ad-
dition of n-hexane. The solid was repeatedly washed with n-hexane and
then dried in vacuum to give the complex as a light yellow solid. Isolated
yield: 130 mg, 85%. Crystals suitable for X-ray analysis were obtained by
recrystallisation from dichloroethane/n-hexane. The X-ray investigation
confirmed the structure already published.[11] NMR spectroscopy data
were collected on a solution of the recrystallised sample: no attempt was
made to interpret the rather complicated 31P NMR spectrum (see Figures
Chem. Eur. J. 2011, 17, 12683 – 12695
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12693

E. Alberico, W. Baumann, J. G. de Vries et al.
[3] a) L. Lefort, J. A. F. Boogers, A. H. M. de Vries, J. G. de Vries, Org.
Lett. 2004, 6, 1733–1735; b) J. G. de Vries, L. Lefort, Chem. Eur. J.
2006, 12, 4722–4734; c) C. Jꢁkel, R. Paciello, Chem. Rev. 2006, 106,
2912–2942.
[4] a) M. T. Reetz, T. Sell, A. Meiswinkel, G. Mehler, Angew. Chem.
2003, 115, 814–817; Angew. Chem. Int. Ed. 2003, 42, 790–793; b) D.
PeÇa, A. J. Minnaard, J. A. F. Boogers, A. H. M. de Vries, J. G. de
Vries, B. L. Feringa, Org. Biomol. Chem. 2003, 1, 1087–1089.
[5] D. PeÇa, A. J. Minnaard, A. H. M. de Vries, J. G. de Vries, B. L. Fer-
inga, Org. Lett. 2003, 5, 475–478.
[6] a) A. J. Minnaard, B. L. Feringa, L. Lefort, J. G. de Vries, Acc.
Chem. Res. 2007, 40, 1267–1277, and references therein; b) L. Eber-
hardt, D. Armspach, J. Harrowfield, D. Matt, Chem. Soc. Rev. 2008,
37, 839–864; c) D. J. Ager, L. Lefort, J. G. de Vries, ACS Symp. Ser.
2009, 1009, 239–250.
[7] J. A. F. Boogers, U. Felfer, M. Kotthaus, L. Lefort, G. Steinbauer,
A. H. M. de Vries, J. G. de Vries, Org. Process Res. Dev. 2007, 11,
585–591.
[21] Dichloroethane has been used in place of dichloromethane because
the latter is a poorly suited solvent for running kinetic investigations
at 1 bar total pressure due to its high vapour pressure. In dichloro-
ethane, the same intermediates as found in dichloromethane have
been detected by NMR spectroscopy.
[22] D. Heller, A. H. M. de Vries, J. G. de Vries in Handbook of Homo-
geneous Hydrogenation, Vol. 3 (Eds.: J. G. de Vries, C. J. Elsevier),
Wiley-VCH, Weinheim, 2007, pp. 1483–1516.
[23] a) J. M. Ernsting, S. Gaemers, C. J. Elsevier, Magn. Reson. Chem.
2004, 42, 721–736; b) W. von Philipsborn, Chem. Soc. Rev. 1999, 28,
95–105.
[24] Although the complexes have a different counterion, the compari-
son is still appropriate as it has been shown that non-coordinating
anions affect very weakly the 103Rh chemical shift of the correspond-
ing complexes: A. Fabrello, C. Dinoi, L. Perrin, P. Kalck, L. Maron,
M. Urrutigoity, O. Dechy-Cabaret, Magn. Reson. Chem. 2010, 48,
848–856.
[25] D. Heller, H.-J. Drexler, A. Spannenberg, B. Heller, J. You, W. Bau-
mann, Angew. Chem. 2002, 114, 814–817; Angew. Chem. Int. Ed.
2002, 41, 777–780.
[26] The possibility for phosphoramidite ligands to act as multidentate li-
gands through dangling aromatic substituents (h2 or h6) and phos-
phorus has been previously described, see T. Osswald, H. Rꢀegger,
A Mezzetti, Chem. Eur. J. 2010, 16, 1388–1397, and references
therein.
[8] M. T. Reetz, A. Meiswinkel, G. Mehler, K. Angermund, M. Graf, W.
Thiel, R. Mynott, D. G. Blackmond, J. Am. Chem. Soc. 2005, 127,
10305–10313.
[9] Y. Fu, X.-X. Guo, S.-F. Zhu, A.-G. Hu, J.-H. Xie, Q.-L. Zhou, J.
Org. Chem. 2004, 69, 4648–4655.
[10] I. D. Gridnev, C. Fan, P. G. Pringle, Chem. Commun. 2007, 1319–
1321.
[27] Similar dimeric species have been reported with the following phos-
phine ligands: DPPE: a) J. Halpern, D. P. Riley, A. S. C. Chan, J. J.
Pluth, J. Am. Chem. Soc. 1977, 99, 8055–8057; b) D. P. Fairlie, B.
Bosnich, Organometallics 1988, 7, 936–945; c) D. P. Fairlie, B. Bos-
nich, Organometallics 1988, 7, 946–954; DIPAMP: d) A. Preetz, W.
Baumann, C. Fischer, H.-J. Drexler, T. Schmidt, R. Thede, D.
Heller, Organometallics 2009, 28, 3673–3677; CYCPHOS: e) D. P.
Riley, J. Organomet. Chem. 1982, 234, 85–97; DIPH: f) D. G. Allen,
S. B. Wild, D. L. Wood, Organometallics 1986, 5, 1009–1015; CHIR-
APHOS: g) S. H. Bergens, P. Noheda, J. Whelan, B. Bosnich, J. Am.
Chem. Soc. 1992, 114, 2121–2128; h) R. W. Barnhart, X. Wang, P.
Noheda, S. H. Bergens, J. Whelan, B. Bosnich, Tetrahedron 1994, 50,
4335–4346; SKEWPHOS: i) K. Mikami, Y. Yusa, M. Hatano, K.
Wakabayashi, K. Aikawa, Chem. Commun. 2004, 98–99; BIPHEP:
j) K. Mikami, S. Kataoka, Y. Yusa, K. Aikawa, Org. Lett. 2004, 6,
[11] M. van den Berg, A. J. Minnaard, R. M. Haak, M. Leeman, E. P.
Schudde, A. Meetsma, B. L. Feringa, A. H. M. de Vries, C. E. P.
Maljaars, C. E. Willans, D. Hyett, J. A. F. Boogers, H. J. W. Hender-
ickx, J. G. de Vries, Adv. Synth. Catal. 2003, 345, 308–323.
[12] A. Preetz, C. Fischer, C. Kohrt, H.-J. Drexler, W. Baumann, D.
Heller, Organometallics, DOI: 10.1021/om200486t.
[13] a) D. A. Evans, F. E. Michael, J. S. Tedrow, K. R. Campos, J. Am.
Chem. Soc. 2003, 125, 3534–3543; b) H.-J. Drexler, W. Baumann, T.
Schmidt, S. Zhang, A. Sun, A. Spannenberg, C. Fischer, H. Bausch-
mann, D. Heller, Angew. Chem. 2005, 117, 1208–1212; Angew.
Chem. Int. Ed. 2005, 44, 1184–1188; c) for a recent discussion of the
Major/Minor concept, see: T. Schmidt, Z. Dai, H.-J. Drexler, M.
Hapke, A. Preetz, D. Heller, Chem. Asian J. 2008, 3, 1170–1180.
[14] a) A. S. C. Chan, J. J. Pluth, J. Halpern, J. Am. Chem. Soc. 1980, 102,
5952–5954; b) J. Halpern, Science 1982, 217, 401–407; c) J. Halpern
in Chiral Catalysis, Vol. 5 (Ed.: J. D. Morrison), Academic Press, Or-
lando, 1985, pp. 41–69; d) C. R. Landis, J. Halpern, J. Am. Chem.
Soc. 1987, 109, 1746–1754; e) B. McCulloch, J. Halpern, M. R.
Thompson, C. R. Landis, Organometallics 1990, 9, 1392–1395; f) T.
Schmidt, W. Baumann, H.-J. Drexler, A. Arrieta, D. Heller, H.
Buschmann, Organometallics 2005, 24, 3842–3848; g) T. Schmidt, Z.
Dai, H.-J. Drexler, W. Baumann, C. Jꢁger, D. Pfeifer, D. Heller,
Chem. Eur. J. 2008, 14, 4469–4471.
3699–3701; PACHTNUGTRENUNG(pTolyl)3: k) P. Marcazzan, M. B. Ezhova, B. O. Pat-
rick, B. R. James, C. R. Chim. 2002, 5, 373–378; Monophosphonite:
reference [10].
[28] In the case of binaphthol-based monophosphonites (see Ref. [10]) in
dichloroethane, the dimer is in equilibrium with the solvato species,
the relative concentrations being dependent on the total rhodium
concentration in solution.
[29] A different behaviour is expected in a coordinating solvent (see
Ref. [27k]): when hydrogenation of 8a is carried out in MeOH
under otherwise identical conditions the dimer is not observed. The
31P NMR spectrum shows a broad absorption centred around d=
154 ppm, which might point to rapidly interconverting solvato or hy-
dride species.
[15] M. van den Berg, Ph.D. Thesis, University of Groningen (The Neth-
erlands),
2006;
http://dissertations.ub.rug.nl/faculties/science/
2006m.van.den.berg/.
[16] P. Jankowski, C. L. McMullin, I. D. Gridnev, A. G. Orpen, P. P. Prin-
gle, Tetrahedron: Asymmetry 2010, 21, 1206–1209.
[17] N. B. Johnson, I. C. Lennon, P. H. Moran, J. A. Ramsden, Acc.
Chem. Res. 2007, 40, 1291–1299.
[30] R. R. Schrock, J. A. Osborn, J. Am. Chem. Soc. 1976, 98, 2134–
2143: Rhodium complexes [Rh(L)2ACHTNUGTRENUNG(nbd)]BF4 that contain mono-
phosphines such as PPh3, PMePh2 and PMe2Ph upon hydrogenation
in a coordinating solvent at room temperature give rise to stable di-
hydrides that could be isolated and characterised. According to Hal-
pern (Ref. [27a]) the stability of such dihydrides might be ascribed
to the fact that, because of the presence of monophosphines, in the
dihydride neither hydrogen ligand needs to be trans to a phosphine
(assuming a cis disposition of the two hydrogen atoms).
[18] R. R. Schrock, J. A. Osborn, Inorg. Chem. 1971, 10, 2397–2407.
[19] The X-ray structure of this species has been already reported in
Ref. [11], so it is not included in this paper. However, the X-ray
analysis served to unambiguously assign its structure as the solution
NMR spectra are fairly complicated.
[20] Even with BisP* ligands (BisP*=1,2-bis(alkylmethylphosphino)-
ethanes, alkyl=tert-butyl-1-adamantyl, 1-methylcyclohexyl, 1,1-di-
ethylpropyl, cyclopentyl, cyclohexyl, isopropyl), the relative amount
of rhodium complexes that contain four-coordinated phosphorus
donors prevails if the catalyst precursors are prepared from [Rh-
[31] a) J. M. Brown, P. A. Chaloner, P. N. Nicholson, J. Chem. Soc. Chem.
Commun. 1978, 646–647; b) J. M. Brown, P. A. Chaloner, A. G.
Kent, B. A. Murrer, P. N. Nicholson, D. Parker, P. J. Sidebottom, J.
Organomet. Chem. 1981, 216, 263–276; no stable hydrides were de-
tected when using ligand (R)-(o-methoxyphenyl)phenylmethylphos-
phine, which probably acts as a chelating ligand through the ortho-
methoxy group.
ACHTUNGTRENNUNG(cod)2]BF4 relative to [RhAHCUTNGTRENN(GUN nbd)2]BF4: I. D. Gridnev, Y. Yamanoi, N.
Higashi, H. Tsuruta, M. Yasutake, T. Imamoto, Adv. Synth. Catal.
2001, 343, 118–136.
12694
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12683 – 12695

Rhodium–MonoPhos
FULL PAPER
[32] a) J. M. Brown, P. A. Chaloner, Tetrahedron Lett. 1978, 19, 1877–
1880; b) A. S. C. Chan, J. J. Pluth, J. Halpern, Inorg. Chim. Acta
1979, 37, L477–L479; c) A. S. C. Chan, J. Halpern, J. Am. Chem.
Soc. 1980, 102, 838–840; d) J. M. Brown, P. A. Chaloner, J. Am.
Chem. Soc. 1980, 102, 3040–3048; e) J. M. Brown, P. A. Chaloner, J.
Chem. Soc. Chem. Commun. 1980, 344–346; f) I. Ojima, T. Kogure,
N. Yoda, J. Org. Chem. 1980, 45, 4728–4739; g) J. M. Brown, D.
Parker, J. Org. Chem. 1982, 47, 2722–2730.
1987, 1583–1588; d) J. A. Ramsden, T. D. W. Claridge, J. M. Brown,
J. Chem. Soc. Chem. Commun. 1995, 2469–2471; e) R. Kadyrov, T.
Freier, D. Heller, M. Michalik, R. Selke, J. Chem. Soc. Chem.
Commun. 1995, 1745–1746; f) D. Heller, R. Kadyrov, M. Michalik,
T. Freier, U. Schmidt, H. W. Krause, Tetrahedron: Asymmetry 1996,
7, 3025–3035.
[35] B. R. Bender, M. Koller, D. Nanz, W. von Philipsborn, J. Am. Chem.
Soc. 1993, 115, 5889–5890.
[33] In the case of the analogous adducts of rhodium–monophosphonite
complex, no indication of the formation of a chelated substrate com-
plex was gained from NMR spectra, and only a doublet was ob-
served in the 31P NMR spectrum. By the use of labelled a-13C
methyl a-benzoyloxyethenephosphonate, a strongly chelating sub-
[36] I. D. Gridnev, N. Higashi, K. Asakura, T. Imamoto, J. Am. Chem.
Soc. 2000, 122, 7183–7194.
[37] J. M. Buriak, J. A. Osborn, J. Chem. Soc. Chem. Commun. 1995,
689–690.
[38] a) C. R. Landis, P. Hilfenhaus, S. Feldgus, J. Am. Chem. Soc. 1999,
121, 8741–8754; b) A. L. Sargent, M. B. Hall, M. F. Guest, J. Am.
Chem. Soc. 1992, 114, 517–522.
[39] G. Giordano, R. H. Crabtree, Inorg. Synth. 1990, 28, 88–90.
[40] R. R. Schrock, J. A. Osborn, J. Am. Chem. Soc. 1971, 93, 3089–
3091.
strate to [Rh(diphosphine)ACHTNUGRTNEUNG
(Solv)2]+ fragments, it was proven instead
that, if the ligand is tBuP(binaphthoxo), this substrate is only weakly
monocoordinated through the phosphoryl oxygen (see Ref. [10]).
[34] a) H. Bircher, B. R. Bender, W. von Philipsborn, Magn. Reson.
Chem. 1993, 31, 293–298; b) J. M. Brown, P. A. Chaloner, G. A.
Morris, J. Chem. Soc. Chem. Commun. 1983, 664–666; c) J. M.
Brown, P. A. Chaloner, G. A. Morris, J. Chem. Soc. Perkin Trans. 2
Received: June 11, 2011
Published online: September 28, 2011
Chem. Eur. J. 2011, 17, 12683 – 12695
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12695