Amino-phosphanes in RhI-catalyzed hydroformylation: Hemilabile behavior of P,N ligands under high CO pressure and catalytic properties

Article abstract of DOI:10.1002/ejic.200500432

The catalytic properties of rhodium complexes containing the α-, β-, or γ-amino-phosphane ligands Ph₂PCH ₂NEt₂ (α,P,N-1), Ph₂PCH(Ar)NHPh [α-P,N-2; Ar = η⁶(o-C₆H₄Cl)Cr(CO) ₃], Ph₂PCH₂NPh₂ (α-P,N-3), Ph₂PCH₂CH(Ph)NHPh (β-P,N), Ph₂PCH ₂(o-C₆H₄-NMe₂) (γ-P,N-1), Ph₂PCH(o-C₆H₄-CH₂NHPh) (γ-P,N-2), and the α,β-diamino-phosphane ligand Et ₂NCH₂P(Ph)CH₂CH(Ph)NHPh (α,β-N,P,N), in styrene hydroformylation have been examined. The results show that the activity increases when the number of backbone carbon atoms linking P and N decreases from 3 to 1. IR and ³¹P HPNMR studies in solution show that all P,N ligands adopt exclusively a κ¹-P coordination mode in rhodium chloride carbonyl complexes under high CO pressure. In the solid state a κ¹-P-α-amino-phosphane coordination has been ascertained by X-ray methods in trans-[RhCl(CO)(γ-P,N-1)₂]. In contrast, an equilibrium between the κ²-P,N and κ¹-P- coordination modes has been observed as a function of the CO pressure for the complex containing the β-P,N ligand. The basicity of the dangling amino group also plays an important role on the catalytic activity and a mechanism involving the nitrogen function in the catalytic cycle is proposed. Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejic.200500432

FULL PAPER
DOI: 10.1002/ejic.200500432
Amino-phosphanes in RhI-Catalyzed Hydroformylation: Hemilabile Behavior of
P,N Ligands under High CO Pressure and Catalytic Properties
Jacques Andrieu,*[a] Jean-Michel Camus,[a] Philippe Richard,[a] Rinaldo Poli,[a,b]
Luca Gonsalvi,[c] Francesco Vizza,[c] and Maurizio Peruzzini*[c]
Keywords: Rhodium / Amino-phosphane ligands / Hemilability / Hydroformylation catalysis
The catalytic properties of rhodium complexes containing the
α-, β-, or γ-amino-phosphane ligands Ph2PCH2NEt2 (α-P,N-
1), Ph2PCH(Ar)NHPh [α-P,N-2; Ar = η6(o-C6H4Cl)Cr(CO)3],
Ph2PCH2NPh2 (α-P,N-3), Ph2PCH2CH(Ph)NHPh (β-P,N),
carbonyl complexes under high CO pressure. In the solid
state a κ1-P-α-amino-phosphane coordination has been as-
certained by X-ray methods in trans-[RhCl(CO)(γ-P,N-1)2]. In
contrast, an equilibrium between the κ2-P,N and κ1-P-coordi-
nation modes has been observed as a function of the CO
pressure for the complex containing the β-P,N ligand. The
Ph2PCH2(o-C6H4–NMe2)
(γ-P,N-1),
Ph2PCH(o-C6H4–
CH2NHPh) (γ-P,N-2), and the α,β-diamino-phosphane ligand
Et2NCH2P(Ph)CH2CH(Ph)NHPh (α,β-N,P,N), in styrene hy- basicity of the dangling amino group also plays an important
droformylation have been examined. The results show that
the activity increases when the number of backbone carbon
atoms linking P and N decreases from 3 to 1. IR and 31P
HPNMR studies in solution show that all P,N ligands adopt
exclusively a κ1-P coordination mode in rhodium chloride
role on the catalytic activity and a mechanism involving the
nitrogen function in the catalytic cycle is proposed.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
was shown that the P,N,P type, Ph2PCH2N(Ar)CH2PPh2,
ligand is much more active than its analogous Ph2P(CH2)3-
PPh2. However, a rationalization for the latter observation
was not proposed.[12]
Introduction
Bifunctional P,N ligands are being intensively investi-
gated because they present advantages compared to com-
mon trialkyl phosphanes in the catalytic applications of
their corresponding Pd, Ni, Ru, and Ir complexes.[1–8] These
ligands can be either κ1-P or κ2-P,N coordinated and each
coordination mode can modify the catalytic properties. For
example, the chelating coordination mode enhances the nu-
cleophilic character of the metal center and subsequently
promotes hydrogen oxidative addition.[7,9,10] A comparative
study between α- and β-amino-phosphanes in rhodium-cat-
alyzed hydroformylation has led to the proposal that a bi-
dentate coordination mode accelerates aldehyde reductive
elimination from RhH2(acyl)(P,N).[11] The authors con-
clude that uncoordinated amino functions do not act as
bases. In contrast, the dangling amine function in a PdII-
coordinated κ1-P-pyridinylphosphane acts as a “proton
messenger” in catalytic alkyne methoxycarbonylation.[1] In
another study of rhodium-catalyzed hydroformylation, it
One of the key questions related to the use of bifunc-
tional ligands, particularly those associating soft and hard
donor atoms, is whether there is reversible association/
dissociation of one labile function during the catalytic cycle
(hemilability concept). In our laboratory, we have developed
different pathways to chiral α- or β-P,N and mixed α,β- and
β,γ-N,P,N ligands and their coordination properties have
been explored in palladium() and rhodium() com-
plexes.[13–16] Particularly, we have reported that β-P,N li-
gands adopt a chelating mode whereas α-P,N ligands are
only P-coordinated in CuI or in RhI complexes.[17,18] Sepa-
rate work has shown that β-P,N ligands in RhI complexes
can adopt a monodentate κ1–P-[19] or a κ2-P,N-coordination
mode.[15,20] Nevertheless, no direct evidence for a hemilabile
character has been reported so far for β-P,N ligands in RhI
complexes. This behavior was only suggested for a few cat-
ionic complexes of type [Rh(COD)(β-P,N)]+ in order to ra-
tionalize a rapid fluxional process.[21] Nor is it known
whether the coordinated N donor in these RhI complexes
and in related complexes with γ-amino-phosphanes is dis-
placed by CO under catalytic conditions. A hemilabile be-
havior for related chiral P,N ligands could explain the low
enantioselectivities observed in asymmetric Rh-catalyzed
hydroformylation.[22,23] On the other hand, an optically
pure RuII β-amino-phosphane complex led to higher ee for
[a] Laboratoire de Synthèse et Electrosynthèse Organométalliques,
UMR 5188 CNRS, Université de Bourgogne, Faculté des Sci-
ences Mirande,
9 avenue Alain Savary, 21078 Dijon, France
[b] Laboratoire de Chimie de Coordination, UPR CNRS 8241,
205 route de Narbonne, 31077 Toulouse Cedex, France
[c] ICCOM-CNR, Istituto di Chimica dei Composti Organometal-
lici, Area di Ricerca CNR di Firenze,
10 Via Madonna del Piano, 50019 Sesto Fiorentino (Firenze),
Italy
E-mail: Jacques.Andrieu@u-bourgogne.fr
Eur. J. Inorg. Chem. 2006, 51–61
© 2006 Wiley-VCH Verlag GmbH & Co. KgaA, Weinheim
51

J. Andrieu, J.-M. Camus, P. Richard, R. Poli, L. Gonsalvi, F. Vizza, M. Peruzzini
FULL PAPER
the asymmetric reduction of ketones, whereas a related Ru-
γ-amino-phosphane complex led to the same product in ra-
cemic form.[6]
We now wish to provide new information on the effect
of a dangling vs. a coordinated nitrogen function of chiral
P,N ligands in rhodium hydroformylation catalysis, thus
helping to clarify a few previously obscure points.[11,12]
Consequently, we have (i) investigated the mono- or biden-
tate coordination mode of P,N ligands in rhodium com-
plexes under CO pressure by 31P HPNMR and IR spec-
troscopy, (ii) examined the catalytic properties of various
RhI–P,N complexes in styrene hydroformylation in order to
establish a relation with the coordination modes, and (iii)
evaluated the effect of the N basicity in P,N ligands on the
formation and properties of the catalytic species.
Results
Coordination, IR, and 31P{1H} HPNMR Studies of Amino-
Phosphanes in RhI Complexes
We first examined whether the nitrogen donor in dif-
ferent P,N ligands (see Scheme 1) is coordinated under cata-
lytic conditions. This investigation completes our recent
studies on the RhI coordination chemistry of P,N li-
gands.[15,18] Indeed, we have already reported that complex
RhCl(COD)(α-P,N-2) (complex 1) adopts a κ1-P coordina-
tion mode both in the solid state and in solution and that
an analogous monodentate coordination mode is also
adopted in the dicarbonyl complex RhCl(CO)2(α-P,N-2)
(IIα, see Scheme 2).[18] The latter was obtained by ligand
exchange from RhCl(COD)(α-P,N-2) (COD = 1,5-cyclooc-
tadiene) under high CO pressure. Lowering the CO pressure
did not result in the formation of a κ2-P, N mononuclear
product (Iα), but rather induces a ligand redistribution to
yield RhCl(CO)(α-P,N-2)2 (IIIα) where both α-P,N ligands
are once again κ1-P coordinated.
Scheme 2.
Figure 1. IR spectra of [Rh(CO)2Cl]2 + P,N ligand mixtures in 1:2
ratio prepared in CH2Cl2 under CO at room temperature, (a) with
the γ-P,N-1 ligand and (b) with the β-P,N ligand.
We then carried out 31P{1H} high pressure (HP) NMR
studies for both dicarbonyl rhodium complexes IIγ and IIβ
in order to confirm the IR results and to obtain additional
Scheme 1.
The behavior of RhI complexes containing the β-P,N and information on their stability. When
a
solution of
γ-P,N-1 ligands was first studied by IR in CH2Cl2 solution [RhCl(COD)(γ-P,N-1)] (complex 2) was pressurized with
under CO (see Figure 1). Irrespective of the amino-phos- CO, its doublet resonance at δ = 29.5 ppm [1J(Rh,P)
=
phane, the IR spectrum shows two absorption bands at 150 Hz] was replaced by a new one at δ = 26.9 ppm [1J(Rh,P)
2092 and at 2010 cm–1, corresponding to the type-II dicar- = 126 Hz] (see Figure 2). The presence of a single complex
bonyl rhodium complex (named IIγ and IIβ, with the γ- is consistent with the IR data. After depressurization, com-
P,N-1 and β-P,N ligands, respectively) where the amino- plex IIγ evolved to complex IIIγ, [doublet at δ = 30.5 ppm
1
phosphane is κ1-P coordinated. However, the IR spectrum with J(Rh,P) = 126 Hz] which became the only detectable
of IIβ shows an additional absorption band at 1986 cm–1 product after flushing the solution for a few minutes with
corresponding to Iβ, as will be discussed in detail below.
a dinitrogen stream (see Figure 2). The assignment of this
52
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 51–61

Amino-phosphanes in RhI-Catalyzed Hydroformylation
FULL PAPER
signal to the monocarbonyl-bis-amino-phosphane rhodium
complex IIIγ was confirmed by an X-ray structure analysis.
Crystals of complex IIIγ (see Figure 3) were obtained upon
attempting to crystallize complex IIγ, and their redissol-
ution in CDCl3 led to a 31P NMR spectrum identical to
spectrum (e) in Figure 2. Therefore, the final product of the
NMR experiment in Figure 2 is unambiguously identified
as a monocarbonyl-bis-amino-phosphane rhodium com-
plex.
The evolution of IIγ to IIIγ must be accompanied by the
formation of undefined organometallic CO-free by-prod-
ucts, as described previously for the analogous evolution of
the α-P,N-2 rhodium complex.[18] The νCO value at
1974 cm–1 for IIIγ is close to that found for
[RhCl(CO)(PPh3)2] (1965 cm–1),[24] thus suggesting similar
electronic properties for the γ-P,N-1 ligand and tri-
phenylphosphane. Table 1 shows quite typical geometrical
features and metric parameters for a complex of type
RhCl(CO)L2 (i.e.: L = trialkyl-, dialkyl-N-pyrrolidinyl-, or
α-amino-phosphanes), namely a nearly ideal square-planar
geometry around the rhodium atom with mutually trans
phosphorus atoms.[18,24]
Table 1. Selected bond lengths [Å] and angles [deg] for IIIγ.
Rh–P
2.3199(7)
1.800(4)
2.3681(9)
1.161(5)
Rh–C(1)
Rh–Cl
C(1)–O
P–Rh–P#
172.83(3)
93.584(17)
180.0
86.416(17)
180.0
C(1)–Rh–P
C(1)–Rh–Cl
P–Rh–Cl
Figure 2. 31P{1H} HPNMR spectra under CO pressure for complex
2 in CDCl3 solution. (a) Under N2, (b) under CO, (c) under 20 bar
of CO, (d) after depressurization, (e) solution after flushing 5 min
with N2.
O–C(1)–Rh
Symmetry transformations used to generate equivalent atoms:
# –x, y, –z + 1/2
Figure 3. ORTEP view of complex IIIγ. Thermal ellipsoids are drawn at the 50% probability level. H atoms are omitted for clarity.
Eur. J. Inorg. Chem. 2006, 51–61
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
53

J. Andrieu, J.-M. Camus, P. Richard, R. Poli, L. Gonsalvi, F. Vizza, M. Peruzzini
FULL PAPER
Worth noting in the structure is that the γ-P,N ligand
does not chelate the rhodium center, contrary to what has
been reported for a variety of Cu, Ru, Pd, and Rh com-
plexes with the related ligands Ph2PCH2[o-C6H4–N(Me)-
CH2CH2OH], Ph2P(o-C6H4–CH2NHtBu), Ph2P(o-C6H4–
CH2NMe2), and PhP[o-C6H4–CH(Me)NMe2]2, respec-
tively.[25–28] All the above results show that the α- and γ-
P,N ligands adopt the same coordination mode in rhodium
chloride complexes and undergo an identical ligand redistri-
bution process (see Scheme 3).
A different situation was observed in the case of rhodium
complexes containing the β-P,N ligand. Indeed, as men-
tioned above, the infrared spectrum recorded under CO
shows an absorption band at 1986 cm–1 (see Figure 1), in
addition to the two carbonyl bands at 2092 and 2011 cm–1
attributed to IIβ. This band corresponds to complex Iβ, as
proven by its independent preparation and characterization
(see Exp. Sect.). The simultaneous presence of Iβ and IIβ
suggests the existence of an equilibrium between the two
complexes under CO. This equilibrium could be followed
by running two dedicated 31P{1H} NMR experiments.
The first experiment was carried out under CO on a
CDCl3 solution containing [RhCl(CO)2]2 and the β-P,N li-
gand. The 31P{1H} NMR spectrum showed two doublets
at δ = 56 ppm [1J(Rh,P) = 168 Hz] and at δ = 21 ppm [1J(Rh,P)
= 122 Hz] in ca. 1:2.5 ratio. The signal at δ = 56 ppm corre-
sponds to complex Iβ, as verified by comparison with an
authentic sample. Following the evidence provided by the
Scheme 3.
IR spectrum (vide supra), the second resonance is assigned IIβ and no signal at δ = 56 ppm [see spectra (a) and (b),
to the dicarbonyl complex IIβ. The second experiment was Figure 4].
carried out under a higher CO pressure (20 bar) on a
After keeping the NMR solution pressurized with CO
CDCl3 solution containing complex [RhCl(COD)(β-P,N)], for 14 h, it was surprising to observe a new broad doublet
1
3, readily formed in situ from [RhCl(COD)]2 and β-P,N. In at δ = 19 ppm [with J(Rh,P) = 168 Hz], named IIЈβ, [see
this case, we observed only the doublet related to complex spectrum (c), Figure 4]. Tentatively, we attribute this broad
Figure 4. 31P{1H} HPNMR spectra under CO pressure of complex 3 in CDCl3 solution. (a) Under N2, (b) under 20 bar of CO, after
3 min, (c) under 20 bar of CO, after 14 h, (d) after depressurization. The asterisk denotes a small amount of a unidentified rhodium
compound.
54
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 51–61

Amino-phosphanes in RhI-Catalyzed Hydroformylation
FULL PAPER
resonance to a pentacoordinate RhI complex resulting from of amino-phosphane ligands (see Scheme 5) in styrene hy-
a further rhodium carbonylation of IIβ at this high CO droformylation tests are shown in Table 2.
pressure, although we have no further spectroscopic evi-
dence to support this assignment at the moment. Depres-
surization of the NMR tube and purging the solution from
dissolved CO with a dinitrogen stream removed the signal
due to IIЈβ and restored the equilibrium between complexes
Iβ and IIβ [see spectrum (d), Figure 4]. The above IR and
NMR observations are summarized in Scheme 4.
Scheme 5.
All catalytic runs were monitored by periodical with-
drawal of aliquots from the autoclave. The conversion was
found to be essentially linear with time, without any induc-
tion period. Thus, each reported TOF is obtained from a
plot obtained by using several data points and is affected
by a small error (ca. 5%, see experimental section for de-
tails). All investigated catalysts afforded complete chemose-
lectivities (no product other than aldehydes are observed)
and essentially the same regioselectivity, namely ca. 9:1 in
favor of the branched product, which is standard for styrene
hydroformylation with rhodium catalysts.[29]
A blank experiment carried out with [RhCl(COD)]2 and
PPh3 (run 1) showed no activity whatsoever under our mild
catalytic conditions. As previously reported,[30] the transfor-
mation of the chloride precursor to the catalytically active
hydride complex requires an external base to trap the HCl
equivalent which forms as a by-product during the initial
H2 activation step. The addition of one equivalent of
Ph2NH or NEt3 as an external base to the previous PPh3-
based catalyst led to a weak catalytic activity (run 2). When
Scheme 4.
Both experiments confirm the occurrence of an equilib-
rium between Iβ and IIβ depending on the CO pressure as
well as the hemilabile character of the β-P,N ligand in RhI
complexes. The combined IR and HPNMR experiments
unambiguously prove the preference for keeping the amino
group uncoordinated, under high CO pressure, in both β-
and γ-P,N RhI complexes, as previously shown for the re-
lated α-P,N RhI complexes, resulting in a selective κ1-P co-
ordination mode. We now wish to examine the effects of
the dangling amine group on the catalytic properties of the
Rh-P,N hydroformylation.
Catalytic Styrene Hydroformylation with Rh-Amino-
Phosphane Complexes
The catalytic properties of different systems obtained an amino group, on the other hand, is incorporated into
from either [RhCl(COD)]2 or [Rh(acac)(CO)2] and a variety the tertiary phosphane ligand (runs 3 to 7), a similar or
Table 2. Activity of RhI-P,N complexes in styrene hydroformylation catalysis.
Run[a]
Catalytic precursor
Ligand (1 equivalent)
TOF[b] [h–1]
b/l ratio[c]
1
[RhCl(COD)]2
[RhCl(COD)]2
[RhCl(COD)]2
[RhCl(COD)]2
[RhCl(COD)]2
[RhCl(COD)]2
[RhCl(COD)]2
[RhCl(COD)]2
[RhCl(COD)]2
PPh3
PPh3 + Ph2NH
PPh3 + NEt3
Ph2P(o-C6H4–CH2NHPh) (γ-P,N-2)
Ph2PCH2 (o-C6H4–NMe2) (γ-P,N-1)
Ph2PCH2CH(Ph)NHPh (β-P,N)
Ph2PCH2NPh2 (α-P,N-3)
Ph2PCH(Ar)NHPh[d] (α-P,N-2)
Ph2PCH2NEt2 (α-P,N-1)
Ͻ1
4
10 0.5
4
–
2a
2b
3a
3b
4
5
6
0.2
88:12
87:13
91:9
92:8
91:9
89:11
89:11
91:9
0.2
0.4
1
1
2
8
13
15
42
7
230 12
[a] Conditions: [RhCl(COD)]2 (2.62×10–2 mmol) or [Rh(acac)(CO)2] (5.23×10–2 mmol), styrene/Rh = 1000, CHCl3 (35 mL), 55 °C,
p(syngas) = 600 psi. [b] TOF calculated from the slope of the conversions as a function of time. [c] Regioselectivity determined after
complete conversion or from initial conversion rate for the faster catalytic reactions. [d] Ar = η6(o-C6H4Cl)Cr(CO)3.
Eur. J. Inorg. Chem. 2006, 51–61
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
55

J. Andrieu, J.-M. Camus, P. Richard, R. Poli, L. Gonsalvi, F. Vizza, M. Peruzzini
FULL PAPER
higher (quite substantial in some cases) increase of catalytic ligands display an uncoordinated amine function under
activity resulted. Although the formation of a very active catalytic conditions, as demonstrated above. Additionally,
phosphane-free carbonyl species could be envisaged during the presence of the free internal base is not devoid of conse-
the ligand scrambling process for the γ-amino-phosphane quence for the catalytic efficiency because the whole system
complexes IIγ (see Scheme 3), the lower activities observed becomes more efficient as the P–N distance decreases,
with the complexes (runs 3a,b) indicate that either this respectively going from γ-, β- to α-P,N ligands, and as its
scrambling process does not take place under the catalytic basicity increases.
conditions (high CO pressure) or that the species formed
by such a process does not possess a very high catalytic
activity. The conversion rate became progressively greater
Rhodium Acetylacetonate Catalyst
as the number of carbon atoms in the ligand backbone
bridging the P and N donor ends decreased from 3 to 1. A
from the chloride catalyst precursor and H2 is scavenged by
jump in activity was particularly evident on going from the
γ- (3 carbon atoms) to the α-P,N ligand (1 carbon atom),
with essentially identical Brønsted basicities for the dan-
gling amino groups (runs 3a, 4, and 6). The effects of the
Under the assumption that the HCl molecule originating
the dangling amine function, the latter would be converted
into an ammonium function. Thus, the active catalyst
would turn over in the ammonium form, rather than in the
free amine form. In order to check the effect of this proton-
amine basicity on the catalyst activity were then investi-
ation process on the catalytic activity, experiments have also
gated.
been carried out using [Rh(acac)(CO)2] as the catalytic pre-
cursor. This allows the uncoordinated amine to remain in
the neutral form during the catalytic cycle, since the active
hydride species forms with elimination of acetylacetone
Nitrogen Basicity
rather than HCl. These experiments were only carried out
A comparison between runs 5 and 7 shows that when the
with the α-P,N ligands, and the results are collected in
uncoordinated NPh2 moiety in the α-P,N ligand is replaced
Table 3. Generally, [Rh(acac)(CO)2] is known to be a better
by the more basic NEt2 unit, the activity (TOF) increases
catalytic precursor than [RhCl(COD)]2 and, indeed, a com-
dramatically from 15 to 230 h–1. A similar, albeit more
parison between runs 1 (Table 2) and 8 (Table 3) shows that
this is true also in our case.
The activity in the presence of PPh3 (see run 8, Table 3)
remains nevertheless lower than in the presence of the α-
moderate, effect is also observed for a similar change in γ-
P,N ligands in which the amino group is farther from the
metallic center (see run 3a and 3b). These observations
highlight the fact that the presence of an uncoordinated ba-
P,N ligands (see runs 9, 10, and 11, Table 3), highlighting
sic function in proximity to the metal center has a beneficial
the beneficial effect of the free amine on the catalysis rate-
effect on the overall catalytic activity. A similar activity/ba-
determining step. We then ran an additional catalytic ex-
sicity relationship for an uncoordinated amine function was
periment which consisted of the addition of one equivalent
pointed out by Reetz et al.,[12] who reported that the activity
of NEt3 into the α-P,N-3 rhodium chloride catalytic system.
of rhodium cationic complexes in styrene or octene hydro-
This led to better conversions (TOF = 220 h–1, run 12) than
formylation increased significantly when the dppp ligand
the NEt3-free system (TOF = 15 h–1, run 5), and rather
[dppp = Ph2P(CH2)3PPh2] was replaced by an amino-di-
close to those obtained with the Rh-acetylacetonate cata-
phosphane ligand [Ph2PCH2N(Ph)CH2PPh2], in which the
nitrogen atom remains uncoordinated. It is to be noted that
lytic precursor (TOF = 285 h–1, run 10).
this conclusion is exactly opposite to that of Abu-Gnim and
Amer, although both the ligands and conditions employed
Discussion
by them are apparently quite similar to ours.[11] It is pos-
sible that such a divergence comes from Amer’s conclusion,
The new results reported in this contribution show that
which is based on the TON values. Such data are obviously all P,N ligands are κ1-P coordinated to RhI under hydrofor-
less significant than the related TOF and, in addition, have mylation conditions, no matter the distance between the N
been calculated from only one catalytic experiment per li- and P atoms, and that the catalytic activity is enhanced by
gand used, with different reaction times as a function of the the proximity of the amine function and by its basicity. The
ligand.[11] Our straightforward interpretation of the activity role of the free amine group can be proposed at two dif-
results presented in Table 3 is based on the fact that all P,N ferent levels. First, we recall that no induction time is ob-
Table 3. Activity of α-P,N-based RhI complexes in styrene hydroformylation catalysis.
Run[a]
Catalytic precursor
Ligand (1 equivalent)
TOF[b] [h–1]
b/l ratio
8
9
10
11
12
Rh(acac)(CO)2
Rh(acac)(CO)2
Rh(acac)(CO)2
Rh(acac)(CO)2
[RhCl(COD)]2
PPh3
130
170
285 13
325 16
220 11
7
9
93:7
Ph2PCH(Ar)NHPh[b]
Ph2PCH2NPh2
Ph2PCH2NEt2
Ph2PCH2NPh2 + NEt3
88:12
90:10
92:8
[c]
90:10
[a] Conditions described in Table 2. [b] Ar = η6(o-C6H4Cl)Cr(CO)3. [c] 1 Equivalent relative to the rhodium complex.
56 Eur. J. Inorg. Chem. 2006, 51–61
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Amino-phosphanes in RhI-Catalyzed Hydroformylation
FULL PAPER
Scheme 6.
served for the generation of the catalytically active species. experiments, that the hydrogenolysis is rate determining.
In order to generate the active Rh–H species, rhodium() Details of these studies are provided in a separate paper[40]
chloride precursors typically show induction periods in the In addition, no evidence for the accumulation of a hydride
catalytic process because of the need to eliminate HCl after intermediate (as required if the olefin insertion step were
the dihydrogen oxidative addition to rhodium. The addition rate determining) was obtained by HPNMR (vide infra).
of an external amine favors this process and reduces or even
We must now attempt to rationalize two different obser-
eliminates the induction time.[30] In our own hands, vations: (i) a proximal free amine function has a beneficial
[RhCl(COD)]2 still affords a very active catalyst in the ab- effect (i.e. see the experiments carried out with the acetyl-
sence of any added phosphane ligand under the same con- acetonate catalyst or with the chloride in the presence of an
ditions given in Table 3, but only after quite a long induc- external base, cf. runs 12 and 2b); (ii) the beneficial effect
tion period (Ͼ3 h). It seems therefore reasonable to us that is observed even when the dangling amine function is likely
the dangling nitrogen function plays an important role in turned into an ammonium derivative (i.e. see all the experi-
assisting the rapid formation of the Rh–H active species. ments carried out with the chloride catalyst, without the
We see two possible ways by which this “assistance” effect use of an external base, runs 3–7).
can operate. A first possibility involves, after the initial H2
Concerning the first point, we assume that the action of
oxidative addition, the N-assisted promotion of HCl elimi- the uncoordinated amine function on the rds is similar to
nation from the intermediate RhIII dihydride (see Scheme 6, that proposed above for the activation of the chloride pre-
path a). On the other hand, an alternative mechanistic path- cursor, see Scheme 7. The greater activity could in principle
way could involve heterolytic cleavage of H2 (see Scheme 6, be interpreted on the basis of two possible mechanistic
path b). Indeed, many authors have proposed a base-as- routes. The first one (path a, Scheme 7) involves H2 oxidat-
sisted H2 heterolytic splitting in relation to other catalytic ive addition followed by aldehyde reductive elimination. On
processes (i.e. involving Ru,[31–34] Ir,[35] Pd,[36] and Rh[37]
)
the basis of this mechanism, we would argue that the free
over the last few years. Heterolytic splitting of H2 is also amine ligand somehow helps either one or both of these
shown by computational methods to be kinetically and elementary steps. However, the lack of interaction between
thermodynamically more favorable than the classical dihy- the rhodium-acyl species and the dangling nitrogen moiety
drogen oxidative addition for the interaction with the rho- does not allow an obvious rationalization of the observed
dium complex Rh(PH3)2(HCO2) in the presence of NH3.[38]
Consequently, a similar H2 heterolytic activation process
seems a reasonable possibility also for the rhodium hydro-
formylation catalytic systems here described, in particular
with the basic Ph2PCH2NEt2 ligand (see Scheme 6). In
either case, the HCl produced during the activation step is
likely to be trapped by the nearby amine function. However,
this activation step cannot be enhancing the activity, be-
cause the chloride precursor is not the resting state of the
catalytic cycle.
The second level where the external base is likely to have
an effect concerns the rate-determining step of the catalytic
cycle (rds), namely the hydrogenolysis of the rhodium acyl
intermediate. Detailed kinetic and mechanistic studies car-
ried out by van Leeuwen et al. have shown that the rate-
determining step in rhodium catalyzed hydroformylation
may be either the olefin insertion into the Rh–H bond or
the Rh-acyl hydrogenolysis, depending on the ligand nature
and on the conditions.[29,39] In the cases at hand, we have
collected preliminary evidence, via H/D isotopic exchange Scheme 7.
Eur. J. Inorg. Chem. 2006, 51–61
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
57

J. Andrieu, J.-M. Camus, P. Richard, R. Poli, L. Gonsalvi, F. Vizza, M. Peruzzini
FULL PAPER
effect. The second pathway (path b, Scheme 7) involves a hydride species B. The most active species C would then
base-assisted heterolytic H2 splitting, leading to a zwitter- form only in minor proportion, leading to a low catalytic
ionic rhodium monohydride intermediate. The latter would activity. According to Scheme 8, the amount of C could be
be stabilized by a P,N-ligand bearing a uncoordinated am- increased only upon addition of a strong external base. On
monium end. According to this mechanism, increasing the the other hand, the catalytic formation of aldehyde was im-
amine basicity could lower the activation barrier leading to mediately observed after the styrene introduction into the
the zwitterionic intermediate, thereby accelerating the rds. HPNMR tube, although the hydride resonance remained
The aldehyde would then be formed by protonolysis of the undetected. The latter result is consistent with the proposi-
rhodium-acyl bond by the dangling ammonium group, tion that the olefin insertion step is not the rds for this sys-
rather than by a reductive elimination process.
tem.
Clearly, the accelerating effect just discussed cannot be
provided when the amine function is protonated. However,
it is possible to envisage a proton exchange process,
whereby the ammonium hydride derivative B formed during
the catalyst activation step may equilibrate with the amine–
chloride precursor A to yield an inactive ammonium chlo-
ride species D and an active amine–hydride complex C,
identical to that obtained from the acetylacetonate precur-
sor (see Scheme 8). Consequently, the different accelerating
effect that is observed for the chloride systems in the ab-
sence of external base (runs 3–7) may be attributed to the
presence of the equilibrium species D, which would again
operate through the same mechanism of Scheme 7, the ex-
tent of the effect being a function of the equilibrium
amount of that species. Acyl complexes are known to give
hydrogenolysis even without the assistance of an internal
amine function. However, the presence of an internal base
increases the activity of the rhodium catalyst, to a growing
extent as its basicity increases (see runs 8 through 11) and
could then promote the hydrogenolysis step through a
mechanism encompassing the H2 heterolytic splitting.
Conclusions
The IR and 31P HPNMR studies reported in this contri-
bution have shown that the coordination properties of a
variety of amino-phosphane ligands in RhI complexes de-
pend on the length of the P,N backbone. The γ-P,N ligands
behave as κ1-P-α-amino-phosphanes, whereas the β-P,N
amino-phosphane leads to a solution equilibrium between
κ2-P, N and κ1-P-coordination modes, depending on the CO
pressure. Under hydroformylation conditions, however, all
the investigated Rh-amino-phosphane precatalysts exist as
monodentate [RhCl(CO)2(κ1-P-P,N)] complexes with a
noncoordinated nitrogen function. The reported catalytic
runs demonstrate that the dangling amino group favors the
formation of the Rh–H active species when using
[RhCl(COD)]2 as the catalytic precursor and that both the
basicity of the nitrogen end and its distance from the metal
center have a crucial effect on the turnover frequency. The
hydride formation could occur either by H2 heterolytic
splitting (assisted by the nearby amino group) or by a classi-
cal oxidative addition (with the amino group assisting the
subsequent HCl reductive-elimination). Theoretical calcula-
tions are currently in progress in order to elucidate the pre-
ferred mechanistic pathway. We also suggest that the dan-
gling amino group could take an additional mechanistic
function as promoter of the hydrogenolysis step through a
second heterolytic H2 splitting. Isotopic H/D exchange
studies under catalytic conditions using chiral P,N ligands,
which are reported in a separate contribution,[40] provide
additional mechanistic details in hydroformylation catalysis
via rhodium amino-phosphane systems.
Experimental Section
All manipulations were carried out under purified argon using
standard Schlenk techniques. All solvents were dried and deoxy-
genated prior to use by standard methods. Standard pressure NMR
measurements [1H, 13C{1H}, and 31P{1H}] were carried out with a
Bruker DRX300 spectrometer in CDCl3 at room temperature. The
peak positions are reported with positive shifts in ppm downfield
of TMS as calculated from the residual solvent peaks [1H and
13C{1H}] or downfield of external 85% H3PO4 (31P). The high-
pressure NMR (HPNMR) spectra were recorded by using a stan-
Scheme 8.
In order to obtain further mechanistic information, we
have carried out a 1H HPNMR study of the [RhCl(COD)]2-
Ph2PCH2NPh2 mixture in CDCl3 under syngas (20 atm),
approaching the experimental conditions of the catalytic
runs. No Rh–H species could be observed under these con-
ditions. This observation suggests that the A/B equilibrium
1
dard 10-mm probe tuned to 31P and H nuclei on a BRUKER AC
200 spectrometer operating at 81 MHz. The HPNMR experiments
were performed in a 10-mm sapphire tube (Saphikon Inc., NH)
proposed in Scheme 8 is only weakly shifted to the Rh–H charged with a solution of precatalyst (10 mg of P,N ligand, plus
58 Eur. J. Inorg. Chem. 2006, 51–61
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Amino-phosphanes in RhI-Catalyzed Hydroformylation
FULL PAPER
3
2
0.5 equiv. of Rh precursor) in CDCl3 (2 mL) under nitrogen, after
which the first spectrum (reference) was recorded at room temp.
The tube was then pressurized with CO 20 atm at room temp. The
variable temperature experiment was started on a Bruker AC 200
spectrometer, recording a sequence of spectra up to 80 °C. After a
chosen time, the probe was cooled to room temperature and an-
other spectrum was recorded. The last spectrum was taken after
venting the tube and flushing with nitrogen. The coordinated 1,5-
COD (1,5-cyclooctadiene) C and H nuclei are labeled from 1 to 8,
C1 and C2 being trans to P.
3J(Ha,Hb) = 14, J(Ha,Hc) = 4 Hz, 1 H, PCHa], 2.86 [dt, J(P,Hb)
=
3J(Hb,Ha) = 14, J(Hb,Hc) = 2 Hz, 1 H, PCHb] ppm. 31P{1H} NMR: δ
3
= 56.4 [d, J(Rh,P) = 168 Hz] ppm. 13C{1H} NMR: δ = 188.5 [dd,
1
1J(Rh,C) = 74, J(P,C) = 18 Hz according to CO cis to P,[45] 1 C,
2
2
C=O], 145.7–125.5 (m, 24 C, arom.), 8.51 [d, J(P,C) = 8.6 Hz, 1 C,
NCH], 38.8 [d, J(P,C) = 26 Hz, 1 C, PCH] ppm. IR: ν = 1986 cm–1
˜
(vs, νCO) and νNH not observed. C27H24ClNOPRh·0.5CH2Cl2
(590.28): calcd. C 55.90, H 4.24, N 2.37; found C 55.64, H 4.38, N
2.54.
trans-[RhCl(CO)(γ-P,N-1)2]
(IIIγ):
[RhCl(CO)2]2
(39 mg,
CAUTION: All manipulations involving high pressure are potentially
hazardous. Safety precautions must be taken at all stages of NMR
studies involving high pressure tubes.
0.098 mmol) and γ-P,N-1 (125 mg, 0.392 mmol) gave 113 mg of
1
IIIγ, yield 71%. H NMR: δ = 8.39–6.93 (m, 28 H, arom.), 4.40 (t
not resolved, 4 H, PCH2), 2.30 (s, 6 H, NCH3). 31P{1H} NMR: δ
1
= 31.20 [d, J(Rh,P) = 126 Hz] ppm. 13C{1H} NMR: δ = 188.0 [dt,
All IR spectra were recorded in CH2Cl2 solution with a Bruker
IFS 66 V spectrophotometer with KBr optics and the absorption
vibration bands are given in cm–1. Elemental analyses were carried
out by the analytical service of the L.S.E.O. with a Fisons Instru-
ments EA1108 analyzer. The commercial compounds PPh3,
Ph2NH, [RhCl(COD)]2, [Rh(acac)(CO)2], and [Rh(CO)2Cl]2 were
used as received. The ligands Ph2PCH2NEt2, Ph2PCH2NPh2,
2
1J(Rh,C) = 76, J(P,C) = 14 Hz according to CO cis to both P,[45]
1
C, C=O], 154.2–120.1 (m, 36 C, arom.), 45.2 (s, 4 C, NCH3). 26.8
[t, J(P,C) = 27 Hz, 2 C, PCH ] ppm. IR: ν = 1974 (vs, νCO) cm–1.
˜
2
C43H44ClN2OP2Rh (805.13): calcd. C 64.15, H 5.51, N 3.48; found
C 64.46, H 5.55, N 3.34.
Formation of the Unstable cis-[RhCl(CO)2(γ-P,N-1)] (Iiγ). Method
A: A mixture of [RhCl(CO)2]2 (19.5 mg, 0.050 mmol) and γ-P,N-1
(32 mg, 0.098 mmol) was dissolved in CDCl3 (0.7 mL). After
10 min, the mixture was analyzed by 1H and 31P{1H} spectroscopy
showing IIγ as the major complex. 1H NMR: δ = 7.75–7.03 (m, 14
Ph2PCH(Ar)NHPh {Ar
= o-C6H4Cl or o-C6H4Cl[Cr(CO)3]},
Ph2PCH2CH(Ph)NHPh, Ph2PCH2(o-C6H4–NMe2) (γ-P,N-1), and
Ph2P(o-C6H4–CH2NHPh) (γ-P,N-2) were prepared according to lit-
erature.[13,14,17,41–44] The synthesis of complex [RhCl(COD)(α-P,N)]
1 with α-P,N=Ph2PCH(o-C6H4Cl[Cr(CO)3])NHPh has previously
been described.[18]
2
H, arom.), 4.27 [d, J(P,H) = 12 Hz, 2 H, PCH2], 2.34 (s, 6 H,
1
NCH3) ppm. 31P{1H} NMR:
δ
=
31.1 [14%, d, J(Rh,P)
=
1
125 Hz] ppm for IIIγ and 27.4 [86%, d, J(Rh,P) = 126 Hz] ppm for
IIγ. After ca. 0.75 h and 12 h, the 31P{1H} NMR gave 1:4.5 and
1:2 ratios respectively for the IIIγ/IIγ mixture complexes. The above
procedure was repeated in CH2Cl2 instead of chloroform. The at-
mosphere was then purged with CO and monitored by infrared.
After 2 min, the pure dicarbonyl was formed. No 13C NMR spec-
trum was recorded due to the increase of the ratio IIIγ/IIγ with
time. The above procedure was repeated in CH2Cl2 and a slow dif-
fusion of pentane to the yellow solution afforded poor quality yel-
low crystals which were identified as the complex IIIγ.
Synthesis of [RhCl(COD)(P,N)] Complexes
The syntheses of the [RhCl(COD)(P,N)] complexes were generally
carried out as follows. A mixture of [RhCl(COD)]2 and P,N ligand
were dissolved in CH2Cl2 (5 mL). The yellow solution obtained
was stirred for 1 h. Addition of pentane afforded yellow or orange
microcrystals which were isolated and dried in vacuo.
[RhCl(COD)(γ-P,N-1)] (2): [RhCl(COD)]2 (84.4 mg, 0.171 mmol),
γ-P,N-1 (109 mg, 0.341 mmol) gave 144 mg of 2, yield 75%. 1H
NMR: δ = 8.81–7.12 (m, 14 H, arom.), 5.61 (s, 2 H, C8H12, olefinic
2
protons, trans to P), 4.21 [d, J(P,H) = 11.6 Hz, 2 H, PCH2], 2.93 (s,
Method B: CO was introduced into a solution of [RhCl(COD)(γ-
P,N-1)], 1a (35 mg, 0.061 mmol) in CH2Cl2 (4 mL). Under CO and
after 2 min, the IR spectrum showed only the pure dicarbonyl com-
plex IIγ at 2092 and 2010 cm–1 (both vs, νCO), and this had not
changed after two days under CO.
2 H, C8H12, olefinic protons, cis to P), 2.48 (s, 3 H, NCH3), 2.49–
1.86 (m, 8 H, C8H12, aliph. H) ppm. 31P{1H} NMR: δ = 30.1 [d,
1J(Rh,P) = 150 Hz] ppm. 13C{1H} NMR: δ = 154.1–120.2 (m, 18 C,
arom.), 104.3 (m, 2 C, C1,2), 70.9 (s, 1 C, C5), 70.7 (s, 1 C, C6), 45.3
(s, 2 C, NCH3), 33.3 (s, 2 C, C3,8), 29.3 (s, 2 C, C4,7), 28.5 [d, 2J(P,C)
= 21 Hz, 1 C, PCH2] ppm. C29H34ClNPRh (565.92): calcd. C 61.55,
H 6.05, N 2.48; found C 61.09, H 5.66, N 3.09.
Formation of the Unstable cis-[RhCl(CO)2(β-P,N)] (Iiβ). Method A:
CO was introduced into a solution of [RhCl(COD)(β-P,N)] 1b
(42 mg, 0.071 mmol) in CH2Cl2 (4 mL). An IR spectrum was re-
corded after 2 min and showed a mixture of monocarbonyl com-
plex Iβ at 1986 cm–1 and dicarbonyl complex [RhCl(CO)2(κ-P-
P,N)] IIβ at 2011 and 2092 cm–1. The ratio between the two com-
plexes did not change after two days under CO.
[RhCl(COD)(β-P,N)] (3): [RhCl(COD)]2 (85 mg, 0.172 mmol) and
β-P,N (131 mg, 0.344 mmol) gave 152 mg of 3, yield 80%. 1H
3
NMR: δ = 8.09–6.72 (m, 20 H, arom.), 7.02 [d, J(P,H) = 7 Hz, 1
H, NH, exchange with D2O], 4.78 (m, 1 H, NCHc), 3.36 [dt, 2J(P,Ha)
3
2
=
3J(Ha,Hb) = 14, J(Ha,Hc) = 4 Hz, 1 H, PCHa], 2.64 [t, J(P,Hb)
=
3J(Hb,Ha) = 14 Hz, 1 H, PCHb], 2.38–1.75 (m, 6 instead of 8 H,
C8H12, aliph. H) ppm. The olefinic protons were not observed at
room temp. due to an olefin rotation dynamic process. 31P{1H}
Method B: A mixture of [RhCl(CO)2]2 (25.6 mg, 0.066 mmol) and
β-P,N (54 mg, 0.131 mmol) was dissolved in CDCl3 (1 mL). After
10 min, CO was introduced into the solution and the mixture was
1
analyzed by H and 31P{1H} NMR spectroscopy. The NMR spec-
NMR: δ = 21.7 [d, J(Rh,P) = 150 Hz] ppm. 13C{1H} NMR: δ =
1
1
147.1–114.9 (m, 24 C, arom.), 105.0 (s, very br., 2 C, C1,2 of COD),
71.0 (s, very br., 2 C, C5,6 of COD), 57.1 (s, 1 C, NCH), 37.7 [d,
J(P,C) = 22 Hz, 1 C, PCH], 31.3 (br. s, 4 C, C3,4,7,8 of COD) ppm.
tra show a mixture of IIβ and Iβ in 1:2.5 ratio. H NMR of IIβ: δ
3
= 8.12–5.63 (m, 28 H, arom.), 5.78 [d, J(P,H) = 6 Hz, 2 H, NH,
exchange with D2O], 4.14 (m, 2 H, NCH), 3.58 (m, 2 H, PCHa),
2.45 (m, 2 H, PCHb) ppm. 31P{1H} NMR of IIβ: δ = 21.11 [d,
1J(Rh,P) = 122 Hz] ppm. No 13C{1H} NMR spectrum was recorded
due to the low solubility of IIβ in CDCl3 and (CD3)2CO.
IR: ν = 3324 (m, νNH) cm–1. C34H36ClNPRh·0.5CH2Cl2 (670.45):
˜
calcd. C 61.75, H 5.52, N 2.09; found C 62.15, H 5.77, N 2.23.
[RhCl(CO)(β-P,N)] (Iβ): [RhCl(CO)2]2 (36.1 mg, 0.093 mmol) and
β-P,N (71 mg, 0.186 mmol) gave 76 mg of Iβ, 57% yield. 1H NMR: Crystal Structure Determination of Complex IIIγ: Intensity data
3
δ = 8.22–6.83 (m, 20 H, arom.), 6.84 [d, J(P,H) = 10.8 Hz, 1 H,
were collected with a Nonius Kappa CCD at 230 K. The structure
was solved by the heavy atom method and refined by full-matrix
NH, exchange with D2O], 3.91 (m, 1 H, NCHc), 3.58 [dt, J(P,Ha)
=
2
Eur. J. Inorg. Chem. 2006, 51–61
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
59

J. Andrieu, J.-M. Camus, P. Richard, R. Poli, L. Gonsalvi, F. Vizza, M. Peruzzini
FULL PAPER
least-squares methods[46] with the aid of the WINGX program.[47]
Non-hydrogen atoms were anisotropically refined. Hydrogen atoms
were included in calculated positions and refined with a riding
model. Crystallographic data for compound IIIγ are reported in
Table 4.
mental error on the TOF determination was carried out from the
TOF values obtained from three successive runs using three times
the same solution for two different catalytic systems (run, 5 and
10). The experimental error for other catalytic runs has been esti-
mated on the basis of the determination as above.
Table 4. Crystal data and structure refinement for IIIγ.
Acknowledgments
Compound
IIIγ
This work was supported by COST D29 (WG009/03) and by the
EC through the contracts HPRN-CT-2002–00176 (HY-
DROCHEM) and MRTN-CT-2003–503864 (AQUACHEM) and
the contribution is kindly acknowledged. We are also grateful to
the Ministère de la Recherche, the Centre National de la Recherche
Scientifique (CNRS) and the Conseil Régional de Bourgogne for
support of this work. J.-M. C. thanks the Ministère de la Recherche
for a Ph. D. fellowship. HPNMR measurements have been possible
through the NMR service of “FIRENZE HYDROLAB”: a project
sponsored by a generous grant by Ente Cassa di Risparmio di Fir-
enze.
Formula
M
T [K]
Crystal system
Space group
a [Å]
C43H44ClN2OP2Rh
805.1
230(2)
monoclinic
C2/c
13.9791(4)
17.7355(8)
17.4564(8)
110.320(2)
4058.6(3)
4
1664
1.318
Enraf–Nonius KappaCCD
mixture of φ rotations and ω scans
0.71073
0.599
0.6×0.5×0.5
0.65
h: –17; 18
b [Å]
c [Å]
β [°]
V [Å3]
Z
F(000)
Dcalc [g/cm3]
Diffractometer
Scan type
λ [Å]
[1] E. Drent, P. Arnoldy, P. H. M. Budzelaar, J. Organomet. Chem.
1993, 455, 247–253.
µ [mm–1
]
[2] K. N. Gravilov, A. I. Polosukhin, Russ. Chem. Rev. 2000, 69,
Crystal size [mm3]
661–682.
sin(θ)/λ max. [Å–1
]
[3] P. Denis, A. Jean, J. F. Crizy, A. Mortreux, F. Petit, J. Am.
Chem. Soc. 1990, 112, 1292–1294.
Index ranges
[4] H. Yang, M. Alvarez, N. Lugan, R. Mathieu, J. Chem. Soc.
Chem. Commun. 1995, 1721–1722.
k: –22; 22
l: –18; 22
SCALEPACK
11009
[5] J.-X. Gao, T. Ikariya, R. Noyori, Organometallics 1996, 15,
Absorption correction
RC = reflections collected
IRC = independent RC
IRCGT = IRC and [I Ͼ 2σ(I)]
Refinement method
1087–1089.
[6] C. Standfest-Hauser, C. Slugovc, K. Mereiter, R. Schmid, K.
Kirchner, L. Xiao, W. Weissensteiner, J. Chem. Soc., Dalton
Trans. 2001, 2989–2995.
[7] T. B. Rauchfuss, J. L. Clements, S. F. Agnew, D. M. Roundhill,
Inorg. Chem. 1977, 16, 775–778.
[8] D. M. Roundhill, R. A. Bechtold, S. G. N. Roundhill, Inorg.
Chem. 1980, 19, 284–289.
[9] J. P. Collman, Acc. Chem. Res. 1968, 1, 136–143.
[10] F. Fache, E. Schultz, M. L. Tommasino, M. Lemaire, Chem.
Rev. 2000, 100, 2159–2231.
4587 [R(int) = 0.05]
3262
full-matrix least squares on F2
4587 / 0 / 228
Data / restraints / parameters
R for IRCGT
[a]
[b]
R1 = 0.0444, wR2 = 0.0934
[a]
[b]
R for IRC
R1 = 0.0779, wR2 = 0.1044
1.021
Goodness-of-fit[c]
Largest diff. peak and hole [e·Å–3
]
0.47 and –1.02
[a] R1 = Σ(||Fo| – |Fc||)/Σ|Fo|. [b] wR2 = {Σw(Fo – Fc2)2/Σ[w(Fo ) ]
}
2
2 2 1/2
where w = 1/[σ2(Fo2) + (0.053P)2], P = [max(Fo2,0) + 2×Fc ]/3.
2
[11] C. Abu-Gnim, I. Amer, J. Mol. Catal. 1993, 85, L275–L278.
[12] M. T. Reetz, S. R. Waldvogel, R. Goddard, Tetrahedron Lett.
1997, 38, 5967–5970.
2
2 2
[c] Goodness of fit = [Σw(Fo – Fc ) /(No – Nv)]1/2
.
[13] J. Andrieu, C. Baldoli, S. Maiorana, R. Poli, P. Richard, Eur.
J. Org. Chem. 1999, 3095–3097.
[14] J. Andrieu, J.-M. Camus, J. Dietz, P. Richard, R. Poli, Inorg.
Chem. 2001, 40, 1597–1605.
[15] J. Andrieu, P. Richard, J.-M. Camus, R. Poli, Inorg. Chem.
2002, 41, 3876–3885.
CCDC-250351 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
Catalytic Runs: The hydroformylation reactions were carried out in
a 300-mL stainless-steel Parr autoclave equipped with a magnetic
drive, an internal glass vessel, and an immersion tube connected to
a valve for solution withdrawals under pressure. The temperature
was controlled by a rigid heating mantle and a single loop cooling
coil. The autoclave was purged three times under vacuum/argon
before introducing the catalytic solution. The 1:1 CO/H2 mixture
was prepared by mixing the pure gases in a 500-mL stainless-steel
cylinder before introduction into the autoclave at the desired pres-
sure (see Table 1 and Table 2). The zero time for kinetic runs was
considered as the time at which the desired pressure was reached
inside the autoclave. In order to maintain temperature and pressure
conditions as constant as possible during each kinetic run, only a
few mL of catalytic solution mixture were carefully withdrawn each
time from the autoclave. The CHCl3 or THF solvent was then ro-
tary evaporated at room temperature and the yellow-orange residue
[16] J. M. Camus, J. Andrieu, P. Richard, R. Poli, Eur. J. Inorg.
Chem. 2004, 1081–1091.
[17] J. Andrieu, J. Dietz, R. Poli, P. Richard, New J. Chem. 1999,
23, 581–583.
[18] J. M. Camus, J. Andrieu, P. Richard, R. Poli, C. Baldoli, S.
Maiorana, Inorg. Chem. 2003, 42, 2384–2390.
[19] X. Liu, A. H. Eisenberg, C. L. Stern, C. A. Mirkin, Inorg.
Chem. 2001, 40, 2940–2941.
[20] L. Dahlenburg, R. Götz, J. Organomet. Chem. 2001, 619, 88–
98.
[21] M. Valentini, K. Selvakumar, M. Wörle, P. S. Pregosin, J. Or-
ganomet. Chem. 1999, 587, 244–251.
[22] A. Aghmiz, A. M. Masdeu-Bulto, C. Claver, D. Sinou, J. Mol.
Catal. A: Chem. 2002, 184, 111–119.
[23] C. Saluzzo, J. Breuzard, S. Pellet-Rostaing, M. Vallet, F. Le Gu-
yader, M. Lemaire, J. Organomet. Chem. 2002, 643, 98–104.
[24] M. L. Clarke, G. L. Holliday, A. M. Z. Slawin, J. D. Woollins,
J. Chem. Soc., Dalton Trans. 2002, 1093–1103.
1
was analyzed by H and 13C{1H} NMR spectroscopy. The experi-
60
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 51–61

Amino-phosphanes in RhI-Catalyzed Hydroformylation
FULL PAPER
[25] J. Andrieu, B. R. Steele, C. G. Screttas, C. J. Cardin, J. Fornies,
Organometallics 1998, 17, 839–845.
[26] P. Crochet, J. Gimeno, S. Garcia-Granda, J. Borge, Organome-
tallics 2001, 20, 4369–4377.
[37] D. Sellmann, G. H. Rackelmann, F. W. Heinemann, Chem. Eur.
J. 1997, 3, 2071–2080.
[38] F. Hutschka, A. Dedieu, J. Chem. Soc., Dalton Trans. 1997,
1899–1902.
[27] G. A. Luinstra, P. H. P. Brinkmann, Organometallics 1998, 17,
5160–5165.
[39] S. C. van der Slot, P. C. J. Kamer, P. van Leeuwen, J. A. Iggo,
B. T. Heaton, Organometallics 2001, 20, 430–441.
[40] J. Andrieu, J. M. Camus, C. Balan, R. Poli, Eur. J. Inorg. Chem.
2006, 62–68; following article.
[41] K. Isslieb, R. Kümmel, H. Oehme, I. Meisser, Chem. Ber. 1968,
101, 3612–3618.
[28] I. Yamada, M. Ohkouchi, M. Yamaguchi, T. Yamagishi, J.
Chem. Soc., Perkin Trans. 1 1997, 1869–1873.
[29] P. W. N. M. Van Leeuwen, C. Claver (Eds.), Rhodium Catalyzed
Hydroformylation, Kluwer, Dordrecht, 2000, vol. 22.
[30] D. Evans, J. A. Osborn, G. Wilkinson, J. Chem. Soc. (A) 1968,
3133–3142.
[42] A. L. Balch, M. M. Olmstead, S. P. Rowley, Inorg. Chim. Acta
1990, 168, 255–264.
[31] H. S. Chu, C. P. Lau, K. Y. Wong, W. T. Wong, Organometallics [43] T. B. Rauchfuss, F. T. Patino, D. M. Roundhill, Inorg. Chem.
1998, 17, 2768–2777.
[32] T. Ohkuma, D. Ishii, H. Takeno, R. Noyori, J. Am. Chem. Soc.
2000, 122, 6510–6511.
1975, 14, 652–656.
[44] D. Hedden, D. M. Roundhill, Inorg. Chem. 1985, 24, 4152–
4158.
[33] K. Abdur-Rashid, M. Faatz, A. J. Lough, R. H. Morris, J. Am.
Chem. Soc. 2001, 123, 7473–7474.
[45] E. Rotondo, G. Battaglia, G. Giuseppe, C. F. Priolo, J. Or-
ganomet. Chem. 1993, 450, 245–252.
[34] L. Dahlenburg, R. Götz, Eur. J. Inorg. Chem. 2004, 4, 888–905. [46] G. M. Sheldrick, SHELX97 (Includes SHELXS97 and
[35] A. J. Lough, S. Park, R. Ramachandran, R. H. Morris, J. Am.
Chem. Soc. 1994, 116, 8356–8357.
[36] P. Pelagatti, A. Bacchi, M. Carcelli, M. Costa, A. Fochi, P.
Ghidini, E. Leporati, M. Masi, C. Pelizzi, G. Pelizzi, J. Or-
ganomet. Chem. 1999, 583, 94–105.
SHELXL97), Release 97–2, Programs for Crystal Structure
Analysis, University of Göttingen, Göttingen, Germany, 1998.
[47] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837–838.
Received: May 12, 2005
Published Online: November 22, 2005
Eur. J. Inorg. Chem. 2006, 51–61
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
61