Coordination and organometallic chemistry of relevance to the rhodium-based catalyst for ethylene hydroamination

Article abstract of DOI:10.1021/ic201508m

The RhCl₃·3H₂O/PPh₃/nBu ₄PI catalytic system for the hydroamination of ethylene by aniline is shown to be thermally stable by a recycle experiment and by a kinetic profile study. The hypothesis of the reduction under catalytic conditions to a Rh ^I species is supported by the observation of a high catalytic activity for complex [RhI(PPh₃)₂]₂. New solution equilibrium studies on [RhX(PPh₃)₂]₂ (X = Cl, I) in the presence of ligands of relevance to the catalytic reaction (PPh₃, C₂H₄, PhNH₂, X^-, and the model Et₂NH amine) are reported. Complex [RhCl(PPh ₃)₂]₂ shows broadening of the ³¹P NMR signal upon addition of PhNH₂, indicating rapid equilibrium with a less thermodynamically stable adduct. The reaction with Et2NH gives extensive conversion into cis-RhCl(PPh₃)₂(NHEt₂), which is however in equilibrium with the starting material and free Et₂NH. Excess NHEt₂ yields a H-bonded adduct cis-RhCl(PPh₃) ₂(Et₂NH) ·NHEt2, in equilibrium with the precursors, as shown by IR spectroscopy. The iodide analogue [RhI(PPh ₃)₂]₂ shows less pronounced reactions (no change with PhNH₂, less extensive addition of Et₂NH with formation of cis-RhI(PPh₃)₂(NHEt₂), less extensive reaction of the latter with additional Et₂NH to yield cis-RhI(PPh₃)₂(Et₂NH) ·NHEt₂. The two [RhX(PPh₃)₂]₂ compounds do not show any evidence for addition of the corresponding X- to yield a putative [RhX ₂(PPh₃)₂]- adduct. The product of C ₂H₄ addition to [RhI(PPh₃)₂] ₂, trans-RhI(PPh₃)₂(C₂H ₄), has been characterized in solution. Treatment of the RhCl ₃·3H₂O/PPh₃/nBu₄PI/PhNH ₂ mixture under catalytic conditions yields mostly [RhCl(PPh ₃)₂]₂, and no significant halide exchange, demonstrating that the promoting effect of iodide must take place at the level of high energy catalytic intermediates. The equilibria have also been investigated at the computational level by DFT with treatment at the full QM level including solvation effects. The calculations confirm that the bridge splitting reaction is slightly less favorable for the iodido derivative. Overall, the study confirms the active role of rhodium(I) species in ethylene hydroamination catalyzed by RhCl₃·3H₂O/PPh ₃/nBu₄PI and suggest that the catalyst resting state is [RhCl(PPh₃)₂]₂ or its C2H4 adduct, RhCl(PPh₃)₂(C₂H₄), under high ethylene pressure.

Full text of DOI:10.1021/ic201508m

Article
pubs.acs.org/IC
Coordination and Organometallic Chemistry of Relevance to the
Rhodium-Based Catalyst for Ethylene Hydroamination
Aurelien Bethegnies,† Vladislava A. Kirkina,‡ Oleg A. Filippov,‡ Jean-Claude Daran,† Natalia V. Belkova,‡
́
́
Elena Shubina,‡ and Rinaldo Polic*,†
†CNRS, LCC (Laboratoire de Chimie de Coordination), Universite
de Narbonne, F-31077 Toulouse, France
́
de Toulouse, UPS, INP; F-31077 Toulouse, France; 205, route
‡A. N. Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences, Vavilov Street 28, 119991 Moscow,
Russia
S
* Supporting Information
ABSTRACT: The RhCl3·3H2O/PPh3/nBu4PI catalytic system for
the hydroamination of ethylene by aniline is shown to be thermally
stable by a recycle experiment and by a kinetic profile study. The
hypothesis of the reduction under catalytic conditions to a RhI
species is supported by the observation of a high catalytic activity for
complex [RhI(PPh3)2]2. New solution equilibrium studies on
[RhX(PPh3)2]2 (X = Cl, I) in the presence of ligands of relevance
to the catalytic reaction (PPh3, C2H4, PhNH2, X−, and the model
Et2NH amine) are reported. Complex [RhCl(PPh3)2]2 shows
broadening of the 31P NMR signal upon addition of PhNH2,
indicating rapid equilibrium with a less thermodynamically stable
adduct. The reaction with Et2NH gives extensive conversion into cis-
RhCl(PPh3)2(NHEt2), which is however in equilibrium with the
starting material and free Et2NH. Excess NHEt2 yields a H-bonded adduct cis-RhCl(PPh3)2(Et2NH)···NHEt2, in equilibrium with
the precursors, as shown by IR spectroscopy. The iodide analogue [RhI(PPh3)2]2 shows less pronounced reactions (no change
with PhNH2, less extensive addition of Et2NH with formation of cis-RhI(PPh3)2(NHEt2), less extensive reaction of the latter with
additional Et2NH to yield cis-RhI(PPh3)2(Et2NH)···NHEt2. The two [RhX(PPh3)2]2 compounds do not show any evidence for
addition of the corresponding X− to yield a putative [RhX2(PPh3)2]− adduct. The product of C2H4 addition to [RhI(PPh3)2]2,
trans-RhI(PPh3)2(C2H4), has been characterized in solution. Treatment of the RhCl3·3H2O/PPh3/nBu4PI/PhNH2 mixture
under catalytic conditions yields mostly [RhCl(PPh3)2]2, and no significant halide exchange, demonstrating that the promoting
effect of iodide must take place at the level of high energy catalytic intermediates. The equilibria have also been investigated at
the computational level by DFT with treatment at the full QM level including solvation effects. The calculations confirm that the
bridge splitting reaction is slightly less favorable for the iodido derivative. Overall, the study confirms the active role of
rhodium(I) species in ethylene hydroamination catalyzed by RhCl3·3H2O/PPh3/nBu4PI and suggest that the catalyst resting
state is [RhCl(PPh3)2]2 or its C2H4 adduct, RhCl(PPh3)2(C2H4), under high ethylene pressure.
Scheme 1. Products Observed in the Halide-Promoted
Catalyzed Hydroamination of Ethylene by Aniline (MXn/Y−
= PtBr2/Br− or RhCl3·3H2O/I−)
INTRODUCTION
■
An efficient halide-promoted platinum catalyst has recently
been developed in our group for the hydroamination of
ethylene and α-olefins by aniline derivatives.1−6 This reaction
yields N-ethylaniline (1) as major product, accompanied by
traces of the double hydroamination product N,N-diethylaniline
(2) and by minor amounts of quinaldine (3), see Scheme 1.
Synthetic, spectroscopic and solution equilibrium studies5,7,8
assisted by DFT calculations9,10 have elucidated important
mechanistic details of the hydroamination cycle and have
rationalized experimental observations such as the halide
promotion effects, the higher activity for the addition of less
basic aniline derivatives, the strong preference for Markovnikov
addition to higher olefins, and also hinted to the importance of
the substrate and/or halide basicity in the deprotonation of a
Received: July 15, 2011
Published: November 14, 2011
© 2011 American Chemical Society
12539
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
zwitterionic intermediate as a key step leading to catalyst
deactivation.11
validating the concept that the active catalyst is indeed a RhI
species.
A few years ago, it has also been reported within our group
that halides equally promote the aniline addition to ethylene
when catalyzed by RhCl3·3H2O/PPh3.12 Compound
RhCl3·3H2O had previously been shown to efficiently catalyze
the addition of basic secondary amines (Et2NH, piperidine) to
ethylene, albeit under rather forcing conditions, but the
efficiency was greatly reduced for less basic amines or for
higher olefins.13,14 Subsequent work had shown that addition of
PPh3 to this system allows ethylene hydroamination by aniline,
but only with much lower efficiencies.15−17 A few well-defined
complexes were also briefly investigated, giving low productiv-
ities and/or fast decomposition.18−20 The work from our group
has revealed that the same catalytic system, like PtBr2, becomes
much more active for the addition of aniline to ethylene when
promoted by halide ions and the strongest promoting effect for
this system turned out to be associated to iodide; under these
conditions, greater amounts of double hydroamination product
2 were obtained relative to the PtBr2/Br− system, whereas the
quinaldine byproduct 3 was essentially absent.12
The promoting effect of halide ions in homogeneous catalysis
is not unprecedented. For instance, the catalytic activity of a
series of Ir(I) diphosphine complexes in intermolecular
hydroamination (of norbornene by aniline), was shown to be
greatly improved in the presence of fluoride ions,21 whereas
iodide is a recognized promoter for a variety of other catalyzed
processes,22 especially the Rh- and Ir-catalyzed methanol
carbonylation. The origin of its promoting effect, however, is
not always completely understood, being possibly associated to
a number of possible causes, such as its good nucleophilicity, its
redox activity, or simply its better binding to low-valent metal
centers decreasing the tendency of the metal complex to
precipitate and two or more effects may occur in different steps
of the same catalytic cycle.22 For the PtBr2 hydroamination, as
already stated, halides are also promoters and in that case
bromide shows the strongest effect,1 though only marginally
better than iodide and the reasons of this difference are not yet
understood.11
Given the rather reducing conditions imposed by the
ethylene and aniline reagents (catalyst degradation to metallic
rhodium was reported under certain conditions), it is
reasonably assumed that RhCl3·3H2O is transformed to a RhI
complex, probably containing PPh3 and iodide ligands, before
beginning the catalytic cycle. However, the coordination
chemistry of iodide complexes of RhI has been little explored,
particularly with the set of ligands of relevance to this catalytic
reaction. Complex RhI(PPh3)3 (the iodide equivalent of
Wilkinson’s catalyst) has been described by Wilkinson
himself,23 and the dinuclear complex [RhI(PPh3)2]2 is also
known.23−25 With the notable exception of Vaska-type, CO
containing complexes,26,27 adducts with other ligands are scarce
and this coordination chemistry merits renewed attention in
view of this specific catalytic transformation.
Following the same strategy recently adopted for the study of
the platinum-based catalyst,5 we have decided to investigate
well-defined rhodium complexes in terms of their solution
equilibria in the presence of all available ligands under catalytic
conditions (triphenylphosphine, ethylene, aniline, halide). DFT
calculations have also been carried out to validate the
experimental findings. Finally, additional catalytic tests have
been run in the presence of well-defined RhI iodide complexes,
EXPERIMENTAL SECTION
■
General. Unless otherwise stated, all manipulations were carried
out under an argon atmosphere. Solvents were dehydrated by standard
procedures and distilled under argon prior to use. Compounds
RhCl3·3H2O (Johnson Matthey 41.92%), PPh3 (Aldrich 99%),
nBu4NCl (Fluka >97%) and nBu4PBr (Acros Organics) were used
as received. Tetra(n-butyl)phosphonium iodide was prepared from
nBu3P and nBuI according to the literature4 and stored protected from
light and under argon in a freezer. Et2NH (Fluka), PhNH2 (Acros
Organics; 99% for analysis ACS) and mesitylene (>97%, Fluka) were
distilled and kept under argon in the dark. Ethylene (N25, purity
≥99.5%) was purchased from Air Liquide. Complexes RhCl(PPh3)3,23
[RhCl(PPh3)2]2,23 RhI(PPh3)3,24 and [RhI(PPh3)2]2 were synthe-
24
sized according to literature procedures. Single crystals of RhI(PPh3)3
were grown by diffusion of pentane into a dichloromethane solution.
Instrumentation. The GC analyses were performed on a
Hewlett-Packard HP4890 (FID) chromatograph (HP 3395 integrator)
equipped with a 30 m × 0.320 mm × 0.25 μm HP1 capillary column
(DB-5MS). The NMR investigations were carried out on Bruker DPX
300 and AV300 spectrometers at 298 K operating at 300.13 MHz (1H)
and 121.495 MHz (31P). IR measurements were carried out on Nicolet
6700 spectrometer using CaF2 cell.
Reaction of RhCl(PPh3)3 with PhNH2. In a 5 mm NMR tube
was placed [RhCl(PPh3)3] (10.7 mg, 0.012 mmol) and dissolved in
∼0.7 mL of CD2Cl2. The solution composition was checked by
31P{1H} NMR, revealing the presence of a minor amount of
[RhCl(PPh3)2]2 and free PPh3, with which the starting compound is
in equilibrium. PhNH2 was added by microsyringe and the progress of
the reaction monitored by 31P{1H} NMR. No change in the spectrum
was observed upon introduction of up to 11 equiv of PhNH2, except
for the broadening of the doublet resonance of the dinuclear complex
[RhCl(PPh3)2]2.
Reaction of [RhCl(PPh3)2]2 with Amines. (a). PhNH2. In a 5
mm NMR tube was introduced [RhCl(PPh3)2]2 (10.06 mg, 0.016
mmol) and dissolved in ∼0.7 mL of CD2Cl2. After checking the quality
of the compound, PhNH2 was added by microsyringe and the progress
of the reaction monitored by 31P{1H} NMR, indicating only a weak
affinity of this amine for bonding (see text).
(b). Et2NH. In a 5 mm NMR tube was introduced [RhCl(PPh3)2]2
(5.4 mg, 8.15 × 10−3 mmol) and dissolved in ∼0.7 mL of CD2Cl2.
After checking the quality of the compound, Et2NH was added by
microsyringe and the progress of the reaction monitored by 31P{1H}
NMR, giving evidence for quantitative formation of cis-RhCl-
(PPh3)2(Et2NH) in the presence of excess Et2NH (see text). A
solution prepared separately in a Schlenk tube with the same
concentration and excess Et2NH was set out for crystallization by
diffusion of a pentane layer, affording single crystals of [RhCl-
(PPh3)2]2.
Reaction of [RhI(PPh3)2]2 with Amines. The reaction of
compound [RhI(PPh3)2]2 with both PhNH2 and Et2NH was carried
out in CD2Cl2 inside an NMR tube, following the same experimental
protocol descrived above for the related [RhCl(PPh3)2]2 system.
Reaction of [RhI(PPh3)2]2 with C2H4. Compound [RhI(PPh3)2]2
(10 mg, 0.013 mmol) was suspended in 2 mL of CD2Cl2 in a Schlenk
tube. The argon atmosphere was then replaced with ethylene and the
suspension was stirred at room temperature. After a few minutes, the
precipitate dissolved to yield a clear, yellow solution. 1H NMR
(CD2Cl2): δ 6.0−6.25 (30H, Ph), 4.43 (broad, 4H, C2H4). 31P NMR
1
(CD2Cl2): δ 35.8 (d, JP−Rh = 124 Hz). An analogous experiment was
carried out in toluene-d6, resulting in the same behavior except that the
solubilization of the precursor required stirring for 24 h.
Ethylene Hydroamination Catalysis. (a). Standard Catalytic
Runs. The catalytic experiments were conducted in a 100 mL stainless
steel thermoregulated (electric oven) autoclave with a stirring bar. In a
typical procedure, the autoclave was charged with RhCl3·3H2O (34.2
mg, 0.13 mmol), P(C6H5)3 (68.2 mg, 0.26 mmol) and nBu4PI (3.26 g,
12540
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
8.45 mmol), closed and submitted to several argon-vacuum cycles.
Distilled and degassed aniline (4.1 mL, 45 mmol) was then syringed
into the autoclave. The ethylene pipe was connected to the autoclave,
purged, and the pressure was adjusted to 25 bar at RT. The
temperature was then raised to 150 °C. After 96 h, the autoclave was
allowed to cool down to room temperature and slowly vented. The
reaction mixture was then poured into 120 mL of diethylether and the
resulting suspension was stirred for 2 h and then filtered. The external
standard N,N-di-n-butylaniline, (∼0.15 g) was added to the collected
ethereal phases and the solution analyzed by GC.
(b). Concentration Monitoring Experiments. An autoclave
equipped with a siphon for sample withdrawal was charged with
RhCl3·3H2O, P(C6H5)3, and nBu4PI in similar amounts to the
previous experiment, closed and submitted to several argon-vacuum
cycles. Distilled and degassed aniline (4.1 mL, 45 mmol) and
mesitylene (distilled and degassed, 16 mL) were then syringed into the
autoclave. The dilution with mesitylene was necessary to allow the
withdrawal of a sufficient number of samples. Ethylene was finally
added and the reaction carried out as described above. Samples of
∼0.5 mL were withdrawn are different times through the siphon and
poured into 3.5 mL of diethylether, followed by stirring. The external
standard N,N-di-n-butylaniline, (∼0.010 g) was added to the collected
ethereal phases and filtered before being analyzed by GC.
(c). Consecutive Catalytic Runs. A first catalytic run was carried
out as described above in (a), until after venting the leftover ethylene
from the autoclave. Degassed and distilled diethylether (50 mL) was
syringed into the autoclave under an argon flow and the mixture was
stirred for 2 h. After one hour of settling, the supernatant liquid was
transferred from the autoclave to a collector flask through the siphon.
The operation was repeated several times until the transferred liquid
was colorless. The external standard N,N-di-n-butylaniline (∼0.010 g)
was added to the collected ethereal phases and filtered before being
analyzed by GC. Subsequently, the autoclave was submitted to vacuum
for 2 h and refilled with argon. A new charge of distilled and degassed
aniline (4.1 mL, 45 mmol) was then syringed into the autoclave.
Ethylene was then introduced and the second catalytic run as carried
out as described in a, including the work up procedure.
Table 1. Crystal Data and Structure Refinement
compound
[RhCl(PPh3)2]2
C36H30ClP2Rh
662.90
RhI(PPh3)3
empirical formula
formula weight
temperature, K
wavelength, Å
crystal system
space group
C54H45IP3Rh
1016.62
180(2)
180(2)
0.71073
0.71073
triclinic
orthorhombic
Pna21
P1
̅
a, Å
9.7226(2)
12.6361(3)
13.7002(4)
85.091(2)
83.067(2)
79.384(2)
1638.78(7)
2
19.7950(8)
12.6130(6)
17.7058(8)
90.0
b, Å
c, Å
α, deg
β, deg
90.0
γ, deg
volume, Å3
90.0
4420.7(3)
4
Z
density (calculated), Mg/m3
abs coeff, mm−1
F(000)
1.343
1.180
0.723
1.211
676
1556
crystal size, mm3
0.24 × 0.19 × 0.14
3.60 - 27.48.
35493
0.48 × 0.37 × 0.31
2.99 - 26.37.
24208
θ range, deg
reflns collected
independent reflns (Rint)
completeness, %
abs correction
max./min transmission
refinement method
data/restraints/parameters
GOF on F2
7439 (0.0337)
98.8
8931 (0.0234)
99.8
multiscan
1.0/0.8750
F2
multiscan
0.7053/0.5941
F2
7439/0/361
1.081
8931/1/532
1.049
R1, wR2 [I > 2σ(I)]
R1, wR2 (all data)
absolute structure parameter
residual density, e·Å−3
0.0283, 0.0893
0.0361, 0.0938
0.0236, 0.0558
0.0275, 0.0570
0.008(11)
0.475/−0.753
0.487/−0.325
Reaction of RhCl3·3H2O with PPh3, Bu4PI, and Aniline. In a
Schlenk tube were introduced RhCl3·3H2O (34 mg, 0.13 mmol), PPh3
(68.2 mg, 0.26 mmol), Bu4PI (820 mg, 2.1 mmol). After purging with
Ar, aniline (1 mL, 11.4 mmol) was added, and the resulting mixture
was stirred at room temperature for 15 min, then at 150 °C for 16 h.
The color of the mixture darkened to red-brown during the first 10
min of heating, then it did not further change. After cooling, 0.4 mL of
the solution were transferred into a 5 mm NMR tube together with
∼0.1 mL of CD2Cl2 for a 31P NMR analysis (see text).
X-ray Crystallography. A single crystal of each compound was
mounted under inert perfluoropolyether on the tip of a cryoloop and
cooled in the cryostream of either an Agilent Technologies GEMINI
EOS CCD diffractometer for [RhCl(PPh3)2]2 or an Oxford-
Diffraction XCALIBUR SAPPHIRE-I CCD diffractometer for RhI-
(PPh3)3. Data were collected using the monochromatic MoKα
radiation (λ = 0.71073).
of the data can be obtained free of charge on application to the
Director, CCDC, 12 Union Road, Cambridge CB2 1EZ, U.K. (fax
(+44) 1223-336-033; e-mail: deposit@ccdc.cam.ac.uk).
Computational Details. Calculations were performed with the
Gaussian09 package33 at the DFT/M06 level34 without any ligand
simplification. Effective core potentials (ECP) and its associated SDD
basis set35−38 supplemented with f-polarization functions (SDD(f))39
were applied for rhodium atom and those supplemented with diffuse
and polarization set of exponents (SDD(pd))40 used for iodine atoms.
Tests were also carried out using the 6-311G** basis set for the I
atom,41 leading to no difference for the relative energy of cis and trans
isomers. The carbon and hydrogen atoms of the phenyl rings of PPh3
ligands were described with 6-31G basis set; P atoms and organic
ligands (C2H4, PhNH2, Et2NH) were described with a 6-31G(d,p) set
of basis functions. Scaling factors were not applied to the calculated
frequencies. Nonspecific solvent effects were introduced through SMD
solvation model42 by single-point energy calculations on gas phase
optimized geometries for dichloromethane (ε = 8.93). The GSMD
values account for the solvation free energies, with inclusion of the
The structures were solved by direct methods (SIR97)28 and
refined by least-squares procedures on F2 using SHELXL-97.29 In
compound [RhCl(PPh3)2]2, some residual electron density was
difficult to model. Therefore, the SQUEEZE function of PLATON30
was used to eliminate the contribution of the electron density in the
solvent region from the intensity data, and the solvent-free model was
employed for the final refinement. There is one cavity of 238 Å per
unit cell. PLATON estimated that the cavity contains 35 electrons
which may correspond to a solvent molecule of dichloromethane as
suggested by chemical analyses. All H atoms attached to carbon were
introduced in idealized positions and treated as riding on their parent
atoms in the calculations. The drawing of the molecules was realized
with the help of ORTEP3.31,32 Crystal data and refinement parameters
are shown in Table 1.
solute free energy contributions ΔGSMD = ΔESMD + ΔGgas − ΔEgas
where ΔESMD is the electronic energy plus the solvent entropy.43
,
RESULTS AND DISCUSSION
■
(a). Hydroamination Catalysis. Before addressing the
chemistry of well-defined RhI complexes in solution, we present
catalytic results that give additional information on the catalyst
stability and mode of action. Within a previous PtBr2-catalyzed
investigation,6 it has become obvious to us that the use of a
glass liner, systematically used in previously published work,
was causing an artificial decrease of the measured activity and a
Crystallographic data (excluding structure factors) have been
deposited with the Cambridge Crystallographic Data Centre as
supplementary publication no. CCDC 828152 and 828153. Copies
12541
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
a
Table 2. Hydroamination of Ethylene by Aniline Catalyzed by RhCl3·3H2O
b
run
I−/Rh
I2/Rh
conv./%
TON 1
TON 2
TON 3
CE
1
10
65
150
65
c
13(14)
83(65)
69(65)
57(73)
2
42(47)
50(130)
138(102)
150(28)
6
4(3)
239(93)
100(120)
51(225)
0
49(53)
531(324)
341(352)
251(484)
5
2
1(4)
2(5)
(3)
0
3
4
2
5
6(1)
6(2)
65
73
28
229
1
488
86
37
266
0
570
a
Values in parentheses are those reported in the previous contribution.12 Conditions: RhCl3·3H2O (34.2 mg, 0.13 mmol), PPh3 (68.2 mg, 0.26
b
mmol, 2 equiv), n-Bu4PI, PhNH2 (4.1 mL, 45.5 mmol, 350 equiv), C2H4 (25 bar; ∼770 equiv), 150 °C, 96 h. CE (catalytic efficiency) = TON 1 +
2(TON 2) + 2(TON 3). Same amounts of all reagents as in entry 2, in the absence of RhCl3·3H2O.
c
loss of reproducibility because of the high temperature
conditions and the resulting condensation of aniline vapors in
the catalyst-free zone of the autoclave underneath the liner. We
have therefore repeated a selected number of previously
reported experiments12 without the glass liner in the autoclave.
The results are shown in Table 2. Note that the catalytic
reactions run with the RhCl3·3H2O/I− catalyst require a much
longer time (4 days) than those carried out with the PtBr2/Br−
catalytic systems (10 h) in order to achieve reasonable
conversions, because of the lower TOF of the former catalyst.
The catalytic efficiency (CE) is calculated on the basis of the
number of cycles needed by the catalyst to yield each product
(two cycles for 2 and 3). As can be seen, while certain
experiments gave essentially the same results (e.g., runs 1 and
3), others gave significantly improved conversions and catalytic
efficiencies (run 2). All new results obtained without glass liner
were repeated and found to be reproducible within 6%. The
new results give the highest activity for a 65-fold excess of the
iodide salt. Repetition of the experiment carried out in the
presence of the I2 additive, however, showed lower efficiency in
the absence of the glass liner. The higher activity previously
reported with glass liner has indeed been reproduced by us (an
even greater catalytic efficiency of 542 has been obtained, which
is essentially identical to the activity found without I2 and
without glass liner, entry 2). Hence, while I2 was believed to
have a promoting effect on the catalysis when this was carried
out inside a glass liner, the opposite effect is found when the
solution is in contact with the stainless steel autoclave walls.
The reason for this change is currently not clear, but may be
related to the negative interference, only after reaction with I2,
of unknown species chemisorbed on the autoclave walls. At any
rate, the fact that the activity in the presence of I2 when using a
glass liner12 is not better than that in the absence of I2 without
glass liner (run 2) shows that I2 has no promoting effect on this
catalysis. A control experiment run under the same conditions
but without RhCl3·3H2O gave insignificant amounts of
hydroamination (run 5).
The results are shown in Table 2 (run 6). As clearly seen, a
similar catalytic efficiency is obtained for the first and the
second run, and both are in good agreement with those of the
single run of entry 2. As a matter of fact, the discrepancy
between the activities in the first and second run (the first being
slightly smaller and the second slightly higher than for entry 2)
is a bit greater than the usual reproducibility level mentioned
above. We tentatively attribute this discrepancy to an
incomplete recovery, in spite of several washings, of the
products from the first run (there is a significant dead volume
between the bottom of the autoclave and the siphon nozzle),
which have therefore remained in the autoclave and added to
those produced in the second run, artificially lowering the
results of the first run and increasing those of the second one.
At any rate, the experiment proves that the catalyst retains most
of its activity after 4 days of operations at 150 °C. Hence, this
catalyst is more robust than the PtBr2/Br− system.1,6
Additional information on the progress of the catalytic
reaction was obtained by monitoring the substrate and product
concentrations with time. The conditions used are those of
entry 2 of Table 2, except that a diluent (mesitylene) was added
to allow a sufficient number of sample withdrawals, again
because of the large dead volume below the siphon nozzle. The
results are shown in Figure 1a. The amount of aniline could not
be determined accurately in this case because of peak overlap
with the mesitylene diluent, but the peak separation was
sufficient to demonstrate the essentially complete consumption
of aniline at the end of the catalytic run. The conversion
proceeded smoothly to first yield 1 and subsequently 2,
consistent with the notion that 1 is an intermediate in the
generation of 2. The absence of a notable induction time
indicates a rapid conversion of the precatalyst to the active
species. Quinaldine (3) was formed in small amounts
throughout the reaction and it is not possible to conclude
whether this reaction leading to this byproduct proceeds
through 1 or not. The evolution of the sum of the three
products (1 + 2 + 3) seems linear, suggesting zero-order in
aniline. After 70 h, essentially no aniline is left in the mixture,
and the sum of all products no longer increases, while the
transformation of 1 into 2 continues to take place. The
essentially complete aniline consumption is in line with the
absence of catalyst deactivation, as already indicated above. The
final mass balance is good, the three products accounting for
>85% of the aniline amount charged in the autoclave, according
to the GC calibration. A second reaction monitoring was also
carried out with a double ethylene pressure (50 bar), giving
very similar results, see Figure 1b.
One important question concerns catalyst deactivation, since
this was shown to occur and to limit the TON for the related
PtBr2/Br−−catalyzed reaction: the catalyst turned into a black
deposit (probably metallic Pt) with no residual catalytic
activity.6 To address this question for the Rh-based catalyst,
we ran two subsequent experiments in the same autoclave
under the same conditions as run 2, with the same catalyst
charge, the second run being carried out after removing the
product mixture of the first run, washing the solid catalyst
several times (all operations were carried out inside the sealed
autoclave under argon), and introducing a new charge of aniline
and ethylene reagents (see details in the Experimental Section).
In conclusion, these new catalytic experiments not only
reproduce the basic findings reported in the previous
12542
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
Comparison of the result of run 7 with that of run 2 (Table
2) indicates that [RhI(PPh3)2]2 indeed leads to efficient
catalysis, although the activity is slightly lower than that
obtained from the same amount of RhCl3·3H2O. This
difference is probably related to parallel reactions that occur
during the RhCl3·3H2O catalyst reduction and conversion to
the iodide system, which involve the chloride ions, ethylene,
aniline and perhaps also PPh3, and produce other species that
possibly further promote the catalysis. The water present in the
RhCl3·3H2O compound may also act as a catalysis promoter.
The results strongly support the proposal that the active form
of the hydroamination catalyst is indeed, as initially suspected, a
RhI species, because run 7 was carried out with a RhI precatalyst
and without any I2 additive. Note also that essentially no
activity was obtained when the experiment was carried out in
the absence of the iodide salt promoter (run 8). This results
perfectly parallels that reported for the RhCl3·3H2O/I−
system.12
We now report the results of our investigations of the
coordination chemistry and the relative thermodynamic
stability of various RhI complexes, both in the chloride and
iodide versions, to learn more about the equilibria that may take
place for the catalytic species under the catalytic hydro-
amination conditions.
(c). Chloride System. As known in the literature, complex
RhCl(PPh3)3 (Wilkinson’s catalyst) equilibrates in solution
with the dinuclear complex [RhCl(PPh3)2]2 and free PPh3 to a
different extent depending on solvent and concentration.23
Most of our NMR studies were carried out in CD2Cl2 at ∼10−2
M concentration, where this dissociation was visible albeit not
extensive. A representative 31P spectrum is shown in the
Supporting Information (Figure S1), indicating <10% dissoci-
ation. Hence, the equilibrium shown in eq 1 is shifted to the left
under our conditions, even in the absence of excess PPh3. The
experiments that will be described further down, carried out
starting from the pure dinuclear compound, were inspired by
the observed changes to the small doublet 31P{1H} resonance
of this compound (δ 51.50 JPRh = 196 Hz) when carrying out
reactions on RhCl(PPh3)3.
Figure 1. Kinetic profile of the RhCl3·3H2O-catalyzed reaction
between PhNH2 and C2H4: 1 (◆), 2 (▲), 3 (●), 1 + 2 + 3 (■).
(a) Conditions are the same as for entry 2 of Table 2. (b) Conditions
are the same as for entry 2 of Table 2, except for a double C2H4
pressure (50 bar, ∼1440 equiv).
contribution,12 but also show that higher and more
reproducible results are obtained in the absence of a glass
liner (a negative effect of I2 under this conditions remains to be
rationalized) and demonstrate that the catalyst does not
degrade under these conditions and furthermore give additional
insights into the kinetic profile of the catalyzed transformation.
(b). Hydroamination with [RhI(PPh3)2]2. As stated in the
Introduction, the presence of a reducing environment (ethyl-
ene, aniline) and of the PPh3 ligand presumably reduces the
RhCl3·3H2O precatalyst to a RhI−PPh3 complex. Furthermore,
since the catalytic activity of RhCl3·3H2O is best promoted by
the iodide ion, we initially assumed that the active species is a
RhI−I complex. Indeed, iodide derivatives of RhI can been
prepared by salt metathesis from the chlorido analogues and an
iodide salt.25 To test this hypothesis, we have carried out a
catalytic run with the [RhI(PPh3)2]2 complex as a precatalyst,
(1)
The reaction between complex RhCl(PPh3)3 and ethylene
has already been studied, since it represents a key step of the
olefin hydrogenation process catalyzed by Wilkinson’s cata-
lyst.23 It affords compound trans-RhCl(PPh3)2(C2H4), which
could be isolated, but the latter tends to lose ethylene to
generate [RhCl(PPh3)2]2, with which it is in rapid equilibrium
(the 1H NMR shows only one resonance for free and
coordinated C2H4, the position of which shifts as a function
of the C2H4 excess). These results indicate a relatively
comparable stability for the two sides of eq 2. We recall
these literature results here because we will compare them later
with other equilibria and with the results of computational
investigations of relative thermodynamic stability.
Table 3. Hydroamination of Ethylene by Aniline Catalyzed
a
by [RhI(PPh3)2]2
b
run
I−/Rh
65
conv./%
TON 1
TON 2
TON 3
CE
7
8
74
1
144
2
116
0
0
0
377
1
a
Conditions: [RhI(PPh3)2]2 (98.1 mg, 0.13 mmol of Rh), PhNH2 (4.1
mL, 45.5 mmol, 350 equiv), C2H4 (25 bar; ∼770 equiv), 150 °C, 96 h.
b
CE (catalytic efficiency) = TON 1 + 2(TON 2) + 2(TON 3).
(2)
Exposure of RhCl(PPh3)3 to aniline, up to a large excess
(>10 equiv) yielded no observable changes in the 31P{1H}
NMR spectrum, except for a broadening of the doublet
resonance of [RhCl(PPh3)2]2. For this reason, the interaction
of aniline was studied more in detail starting from the pure
both with and without the addition of excess free iodide in the
form of the tetra-n-butylphosphonium salt. The results are
shown in Table 3.
12543
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
dinuclear complex. The addition of a large excess of PhNH2
(up to 65 equiv) did not reveal any new 31P{1H} resonance but
Information Figure S1; the corresponding doublet of doublets
at δ 31.2, not shown in Figure 3, is also visible in the spectrum).
Resonances of free PPh3 and Ph3PO are also observed in this
spectrum. The 1H spectrum shows, in addition to the large and
unresolved multiplet resonance of the PPh3 ligands in the
aromatic region, broad resonances of the amine CH2 and CH3
protons at δ 3.15 and 1.60. In the presence of excess NHEt2,
these resonances remain distinguished from those of free
NHEt2, indicating slow exchange, but become sharper and
notably the CH2 resonance pattern is more complex than a
binomial quartet, as expected since these protons become
inequivalent (diastereotopic) upon coordination.
A parallel IR study in the NH stretching vibration region
showed the complete consumption of Et2NH (νNH = 3327
Figure 2. 31P{1H} study in CD2Cl2 of the interaction between
[RhCl(PPh3)2]2 (c = 0.027 M) and PhNH2 in CD2Cl2. (a) Without
PhNH2. (b) With 65 equiv of PhNH2. The difference in noise level is
caused by a much larger number of accumulations for spectrum (b).
shows a significant broadening without major shifting of the
resonance of compound [RhCl(PPh3)2]2, as illustrated in
Figure 2, suggesting a fast exchange between the reagents and
the product (eq 3), strongly shifted to the reagents. Estimating
a maximum amount of 1% product, given the conditions of
temperature and concentrations, we can estimate that the
product is at least 5.5 kcal mol−1 higher in free energy than the
dinuclear precursor. On the other hand, the line broadening
(w1/2 ≈ 150 Hz) gives us information on the pseudo-first-order
exchange rate constant, yielding an estimated barrier of 14 kcal
mol−1 for the aniline addition process.
(3)
Given the difficulty in obtaining a stable amine derivative, we
have oriented our investigation toward a more basic amine as a
Figure 4. Infrared study of the reaction of [RhCl(PPh3)2]2 with
NHEt2 in CH2Cl2 (path length = 2.2 mm). (a) Spectra of
[RhCl(PPh3)2]2 (c = 6 × 10−3 M) (dotted line) and NHEt2 (c =
0.05 M, dashed line). (b) Spectra of a 6 × 10−3 M solution of
[RhCl(PPh3)2]2 after addition of NHEt2 with the successive
concentrations (6, 12, 24, 60, and 120) × 10−3 M.
cm−1, Δν1/2 = 45 cm−1 ε = 2 L mol−1 cm−1) at a 1:1 Et2NH/Rh
ratio and the appearance of a new narrower νNH band assigned
to the coordinated amine (νNH = 3279 cm−1, Δν1/2 = 12 cm−1),
see Figure 4. This behavior is similar to that recently described
for the PtBr2(C2H4)(NHEt2) complex.8 Addition of Et2NH
beyond a Et2NH/Rh ratio of 2 reduced the intensity of this new
band in favor of yet another band at 3263 cm−1 (Δν1/2 = 26
cm−1), which can be attributed to the νNH vibrations of both
coordinated and hydrogen bonded amine in a cis-RhCl-
(PPh3)2(Et2NH)·Et2NH adduct on the basis of DFT
calculations (vide infra). This evidence illustrates the presence
of an H-bonding equilibrium, as detailed in eqs 4 and 5.
Figure 3. 31P{1H} NMR spectrum in CD2Cl2 of cis-RhCl-
(PPh3)2(Et2NH), obtained from [RhCl(PPh3)2]2 and Et2NH (5
equiv).
model system. The reaction of [RhCl(PPh3)2]2 with Et2NH
gave rise to a new product, cis-RhCl(PPh3)2(Et2NH). The
31P{1H} spectrum of the product of reaction 4 is shown in
Figure 3. The two resonances in the form of doublets of
doublets clearly indicate the cis stereochemistry, showing a
mutual (PP) coupling of 47.5 Hz and couplings to the Rh
nucleus by 207 Hz for the resonance at δ 55.0 and by 167 for
that at δ 48.3. Note that no starting material remains visible in
this spectrum (cf. Figure 2), and no other product peak is
visible, notably there is no doublet that could indicate the
formation of a trans isomer. The small signals flanking the
doublet of doublets centered at δ 48.3 is the doublet of triplets
of a small amount of RhCl(PPh3)3 contaminant (cf. Supporting
(4)
12544
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
acetate per dimer,47 whereas a crystal containing a disordered
dichloromethane molecule was obtained under our conditions.
A view of the molecule is reported in Figure 5. The structural
details are in good agreement with those of the ethyl acetate
solvate and no further comment on this structure is warranted.
In addition to PPh3, C2H4, and PhNH2, the hydroamination
catalytic mixture also contains a halide salt. The study of the
interaction between RhCl(PPh3)3 or [RhCl(PPh3)2]2 and a
halide salt was limited to chloride, introduced as the n-
butylammonium salt, nBu4NCl, to avoid the complication of
mixed halide systems. The corresponding iodide system will be
examined below. To put the maximum chances on our side, we
studied the addition of Cl− to the thermodynamically less stable
dinuclear complex [RhCl(PPh3)2]2, expecting to observe the
product of equilibrium 6. However, the 31P NMR spectrum of
the solution, in the presence of as much as 10 equiv of nBu4Cl,
still showed the doublet resonance of the dinuclear starting
material as the dominant resonance in the spectrum, unshifted
and unbroadened relative to that of the pure starting material.
Several other minor peaks, besides a peak of Ph3PO, were also
visible (spectrum shown in Supporting Information Figure S2),
but it is impossible to assign any of them with certainty to the
expected product of eq 6. The solution also shows instability, as
additional small resonances became visible upon keeping the
solution under argon for several hours. The resonance of
[RhCl(PPh3)2]2 always remained the dominant one, however.
This experiment allow us to at least conclude that equilibrium
6, if any [RhCl2(PPh3)2]− product is formed at all, must be
heavily shifted to the left-hand side. Note that, on the other
hand, the reaction of [RhCl(CO)2]2 with chloride generates
[RhCl2(CO)2]− quantitatively.48
(5)
The literature is rather poor of reports of CO-free amine
complexes of rhodium(I). A complex similar to cis-RhCl-
(PPh3)2(Et2NH) was obtained by the same strategy (eq 4) with
a substituted pyridine in place of Et2NH, affording again a
single product with cis geometry as confirmed by an X-ray
structural analysis, but no 31P NMR data were reported.44
Complex RhCl(PPh3)2(Et2NSiMe3), obtained by ligand ex-
change from RhCl(PPh3)3, on the other hand, shows only a
single 31P NMR resonance in agreement with a trans
geometry.45 It is possible in principle that the nonequivalence
of the 31P resonances for RhCl(PPh3)2(Et2NH) is caused by
restricted rotation around the Rh−N in a trans structure,
however in this case the RhCl(PPh3)2(Et2NSiMe3) analogue
would also be expected to have restricted rotation, since the
amine is bulkier, leading to inequivalent P nuclei. We prefer to
believe in an electronic effect of the amine ligand (the
preference of a cis structure has also been confirmed by DFT
calculations, vide infra). In further support of the assigned cis
1
2
structure, the two JPRh and the JPP values observed for
RhCl(PPh3)2(Et2NH) are very close to those reported for two
different conformers of RhCl(Ph2PCH2CH2PPh2)[N-
(CH2Ph)C(Me)(C6H4-4-OMe)] where the cis conforma-
tion is enforced by the chelating nature of the diphosphine
(203.5, 166.9, and 42.9 for one; 199.4, ∼171, and 43.5 for the
1
other), with the same pattern (higher JPRh for the lower field
resonance).46 We have attempted to crystallize the Et2NH
adduct and indeed single crystals were obtained from a solution
that contained a large excess of NHEt2 (5 equiv). However, the
(6)
Figure 5. Molecular view of [RhCl(PPh3)2]2 with the atom-labeling
scheme. Displacement ellipsoids are drawn at the 50% probability
level. H atoms have been omitted for clarity. Relevant bond distances
(Å) and angles (deg): Rh−Cl, 2.4309(6); Rh−Cli, 2.3967(6); Rh−P1,
2.2021(6); Rh−P2, 2.2168(6). P1−Rh−P2, 96.05(2); P1−Rh−Cl1,
176.23(2); P1−Rh−Cl1i, 95.24(2); P2−Rh−Cl1, 87.33(2); P2−Rh−
Cl1i, 168.01(2); Cl1−Rh−Cl1i, 81.51(2); Rh−Cl1−Rhi, 98.49(2).
Symmetry transformations used to generate equivalent atoms: (i) −x
+ 2, −y + 1, −z.
Figure 6. Molecular view of RhI(PPh3)3 with the atom-labeling
scheme. Displacement ellipsoids are drawn at the 50% probability
level. H atoms have been omitted for clarity. Relevant bond distances
(Å) and angles (deg): Rh−I, 2.6840(3); Rh−P1, 2.2303(7); Rh−P2,
2.2937(7); Rh−P3, 2.3239(7). I−Rh−P1, 163.53(2); I−Rh−P2,
86.09(2); I−Rh−P3, 86.76(2); P1−Rh−P2, 97.38(3); P1−Rh−P3,
95.39(3); P2−Rh−P3, 157.89(3).
reversibility of eq 4 and the lower solubility of the dinuclear
products led to the crystallization of the starting material. The
molecular structure of [RhCl(PPh3)2]2 has previously been
reported in a habit containing an interstitial molecule of ethyl
(d). Iodide System. Although compound RhI(PPh3)3 is
known,23 its X-ray structure has apparently never been
reported. The only reported structure with a RhIIP3
coordination environment is apparently that of compound
12545
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
RhI(diop)(PPh3).49 Compound RhI(PPh3)3 crystallizes with
the full molecule in the asymmetric unit in the orthorhombic
space group Pna21. A view of the molecule is shown in Figure 6.
The presence of three sterically encumbering PPh3 ligands
forces the structure to deviate from the preferred square planar
configuration toward a butterfly arrangement (trans angles
around 160°). The same effect was previously noted for
different polymorphs of the related RhCl(PPh3)3 compound.50
The Rh−P distances in the iodide structure compare well with
those in the two different chloride structures, with the two
phosphine ligands trans to each other showing slightly longer
distances relative to that trans to the halide. The latter has a
marginally longer distance than the same ligand in the two
chloride structures [2.225(4) and 2.214(4) Å]. The Rh−I
distance is shorter relative to the above-mentioned diop
complex [2.704(1) Å].49
As already described in the literature, complex RhI(PPh3)3
equilibrates in solution with the dinuclear species [RhI-
(PPh3)2]2 and free PPh3, like the chloride analogue. The
dinuclear species is favored to a greater extent for the iodide
system and indeed the best preparation method of the
dinuclear complex is by PPh3 dissociation under thermal
conditions in a solvent where the dinuclear product is less
soluble and precipitates.24 The literature also shows that PPh3
dissociation is faster for RhI(PPh3)3 than for the corresponding
chloride, since the 31P resonances are broad at room
temperature25 and become sharper upon cooling. All these
observations, reproduced in our hands, suggest that the Rh−
PPh3 interaction is weaker in the iodide system.
Figure 7. 31P{1H} NMR spectra of [RhI(PPh3)2]2 in CD2Cl2 before
(a) and after (b) the addition of 3 equiv of Et2NH. The starred peaks
are due to a trace amount of chlorido derivative, RhCl(PPh3)2(NHEt2)
(cf. Figure 3).
Et2NH, on the other hand, resulted in reaction and formation
of the expected cis-RhI(PPh3)2(Et2NH). The latter compound
is characterized, like the above-described chlorido analogue, by
two doublets of doublets for the two inequivalent PPh3 ligands,
see Figure 7. At variance with the chlorido system, the Et2NH
addition process was not quantitative, resonances of the starting
dinuclear complex remaining visible even after adding an excess
amount of amine. The JPRh of the dinuclear compound in the
two spectra of Figure 7 is the same (190 Hz; the spectra were
taken at two different field strengths).
The IR study of the reaction showed the consumption of the
free amine νNH band, accompanied by the appearance of the
band of the coordinated amine at νNH = 3279 cm−1 (Δν1/2 = 12
cm−1). This frequency is identical to that of the chlorido
The reaction of [RhI(PPh3)2]2 with C2H4 takes place readily
in dichloromethane and more slowly in toluene, to afford trans-
RhI(PPh3)2(C2H4) selectively (eq 7), as shown by the presence
of only one doublet resonance in the 31P NMR spectrum at δ
35.8 (1JP−Rh = 124 Hz) in CD2Cl2. The possibility that a
halogen exchange has occurred with the chlorinated solvent to
yield trans-RhCl(PPh3)2(C2H4), besides being against the
HSAB principle, is categorically excluded because the same
experiment run in toluene-d8 gave a doublet resonance at δ 36.4
(1JP−Rh = 125 Hz), whereas the corresponding experiment using
[RhCl(PPh3)2]2 in place of [RhI(PPh3)2]2 gave a doublet
1
resonance at δ 34.9 (d, JP−Rh 129 Hz). The coupling constant
Figure 8. Infrared study of the reaction of [RhI(PPh3)2]2 (c = 3 × 10−3
M) with NHEt2 in CH2Cl2 (path length = 2.2 mm), after addition of
NHEt2 with the successive concentrations (3, 4.5, 6, 12, 18, and 30) ×
10−3 M.
in the CD2Cl2 spectrum matches well with that obtained for the
iodide system but not with that obtained for the chloride
system in toluene-d8. The 1H spectrum of trans-RhI-
(PPh3)2(C2H4) shows a broad resonance for the ethylene
ligand, like that reported in the literature23 for the
corresponding chlorido derivative. Note that this complex
could not be isolated and an NMR spectrum was not reported
in the original study by Wilkinson et al.,23 where its generation
was attempted from RhI(PPh3)3. Our results and those already
available in the literature suggest that the ethylene binding
constant to [RhI(PPh3)2] is smaller than for the chlorido
analogue.
analogue, showing a negligible effect of the halide nature on the
N−H stretching frequency. However, further addition of
Et2NH resulted in a much less significant change compared
to the chlorido system. The lower frequency νNH band of the
[RhI(Et2NH)(PPh3)2]·Et2NH adduct appeared in the spectrum
only with a 10-fold excess of amine as a weak shoulder at
∼3260 cm−1, see Figure 8. This is in agreement with the
computational results, showing less favorable formation of this
adduct in case of iodide (vide infra).
(7)
Finally, addition of I− (in the form of its Bu4N+ salt, up to 3
equiv per Rh) to complex [RhI(PPh3)2]2, like for the related
experiment on the chloride system, resulted in no significant
spectral change, indicating that equilibrium 8 is heavily shifted
toward the left-hand side. Like for the chlorido analogue, the
corresponding dicarbonyl complex [RhI2(CO)2]− is a well-
known compound.48
Addition of aniline, even in large excess, to [RhI(PPh3)2]2
did not lead to any spectral change and notably, contrary to the
chlorido system, did not significantly broaden the 31P
resonance, suggesting that the energy gap between the reactant
mixture and the putative RhI(PPh3)2(PhNH2) product is
greater than for to corresponding chlorido system. Addition of
12546
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
thermodynamic. In support of this argument, we note that Cl/I
halide exchange processes have been shown to favor the
chlorido species also in other cases when carried out in the
presence of soluble salts (e.g., CpMoI2(PMe3)2 to CpMoICl-
(PMe3)2 and CpMoCl2(PMe3)2 with PPNCl/CH2Cl2;51
MoOI2(PMe3)3 to MoOICl(PMe3)3 and MoOCl2(PMe3)3
with nBu4NCl/acetone52) whereas exchange in the opposite
direction occurs in the presence of partially soluble alkali metal
salts, presumably because of the lower solubility of the alkali
metal chloride (e.g., CpMoCl2(PMe3)2 to CpMoICl(PMe3)2
and CpMoI2(PMe3)2 with NaI/THF;53 MoOCl2(PMe3)3 to
MoOICl(PMe3)3 and MoOI2(PMe3)3 with NaI/acetone;52 and
of stronger relevance to this work, RhCl(PPh3)3 to [RhI-
(PPh3)2]2 and free PPh3 with LiI/toluene25). DFT calculations
have also been conducted to address this point (vide infra). At
any rate, whatever the reason for the lack of halide exchange,
the absence of a notable induction time in catalysis suggests
that [RhCl(PPh3)2]2 is an active precatalyst, without the need
to be converted to the iodide analogue. The results do not
exclude that the real catalyst is one of the very minor and
unidentified byproduct formed in the reaction. However, the
catalytic activity demonstrated independently for [RhI-
(PPh3)2]2 lends credence to the hypothesis that the dinuclear
RhI structure is indeed responsible for generating the active
catalyst. If this is the case, then the promoting effect of iodide
(much better than chloride)12 must operate at the level of
higher-energy catalytic steps, rather than at the level of the
resting state.
(8)
(e). Treatment of RhCl3·3H2O with PPh3, Bu4PI, and
Aniline. A final spectroscopic investigation consisted of the 31
P
NMR study of a mixture prepared with all the catalytic
components (except for ethylene) under conditions mimicking
those of the catalysis to gain further insights into the conversion
of the RhCl3·3H2O precatalyst into the active catalyst.
Excluding ethylene allows the reaction to be carried out in
simple glassware (the boiling point of aniline is 184.13 °C).
Thus, a mixture containing RhCl3·3H2O, PPh3, nBu4PI, and
PhNH2 in a 1:2:16:88 ratio was kept at 150 °C for 16 h, even
though no further change of color was noted after the first 10
min of heating. The nBu4PI/PhNH2 ratio in this experiment
Figure 9. Excerpt of the 31P{1H} NMR spectrum of the final mixture
generated from RhCl3·3H2O, PPh3, nBu4PI, and PhNH2 in a 1:2:16:88
ratio at 150 °C.
The most important conclusions to be drawn from this
experiment is that the resting state of the catalytic cycle must be
either [RhCl(PPh3)2]2 or trans-RhCl(PPh3)2(C2H4), because it
is known from independent experiments reported here and in
the literature that the reaction of [RhCl(PPh3)2]2 (or the
iodido analogue) with ethylene is equilibrated but more
favorable than the additions of aniline or halide ions.
was identical to that of the catalytic runs, but the Rh and PPh3
concentrations were ∼4 times higher in order to facilitate the
31P NMR measurement of the reaction products in the
presence of the large excess of nBu4P+ ion. The nBu4P+
resonance was indeed by far the most intense one in the
spectrum, but was accompanied by two major and relatively
broad peaks, which account for most of the PPh3 intensity, a
stronger one at δ 52.0 and a smaller one at δ 50.6 (see Figure
9). The spectrum also showed a large number of other very
weak resonances revealing the formation of many other
products, none of which could be identified except Ph3PO.
Comparison of chemical shift and line width indicates that
the major product (δ 52.0) is [RhCl(PPh3)2]2 (cf. Figure 2).
The presence of the large excess of aniline is responsible for the
resonance broadening as discussed above. The dinuclear iodide
analogue, [RhI(PPh3)2]2, does not appear to be present, since it
should generate an unbroadened resonance centered at δ 48.8
(cf. Figure 7). It seems therefore reasonable to assume that the
small resonance at δ 50.6 belongs to a mixed-halido species,
Rh2(I)(Cl)(PPh3)4, broadened by exchange with aniline like
the chlorido dimer and unlike the iodido dimer.
The results of this experiment confirm therefore that
RhCl3·3H2O is reduced to RhI under catalytic conditions in
the presence of PPh3, aniline, and iodide ions but reveal that no
extensive halide exchange has taken place and the chlorido
dinuclear species [RhCl(PPh3)2]2 is the major reaction product,
at least after cooling the mixture back to room temperature.
The lack of extensive halide exchange may result from
thermodynamic (the relative amount of iodide used here is 4
times less than in the catalytic run, but still in quite large excess
relative to Rh) or kinetic factors. We believe that the reason is
(f). DFT Calculations. No ligand simplification was made
and the calculations were full QM, at the M06 level. The
Table 4. Comparison between DFT Optimized and
Experimental Geometries (Distances in Å, Angles in
Degrees)
RhCl(PPh3)3
RhI(PPh3)3
a
b
DFT
exp
DFT
exp
Rh−X
Rh−P(trans to X)
Rh−P(trans to P)
2.440
2.258
2.340
2.361
2.404(4)
2.225(4)
2.304(4)
2.338(4)
2.760
2.278
2.344
2.374
2.6840(3)
2.2303(7)
2.2937(7)
2.3239(7)
163.53(2)
86.09(2)
X−Rh−P(trans)
X−Rh−P(cis)
164.9
166.7(2)
161.5
82.4
85.1
84.5(1)
85.3(1)
159.1(2)
97.7(1)
96.4(2)
85.3
87.6
86.76(2)
P−Rh−P(trans)
P−Rh−P(cis)
156.5
101.6
95.7
157.7
99.3
157.89(3)
97.38(3)
94.1
95.39(3)
a
b
Data from ref 50. This work.
geometries of the dinuclear [RhX(PPh3)2]2 systems (X = Cl, I)
and their ligand adducts [RhX(PPh3)2(L)] with L = PPh3, X−,
PhNH2, Et2NH and C2H4 have been optimized by DFT
calculations. The optimized geometries are in good agreement
with the available experimental structures, namely, those of
RhCl(PPh3)3 (orange polymorph) available in ref 50 and
RhI(PPh3)3 reported in the present work. A comparison is
12547
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
available in Table 4. The largest deviation in the bond lengths is
seen for the Rh−I bond (0.07 Å), the others being within 0.05
Å, whereas the angles are all calculated within 2° from the
experimental values.
They are calculated with respect to the precursor complex
[RhX(PPh3)2]2 and free ligands L according to eq 9. The direct
comparison with the experimental data is complicated by
several approximations, which are well-known in the area of
computational chemistry applied to condensed phase thermo-
dynamics, the most important one being probably the use of
unquenched translational and rotational modes for the
estimation of the condensed phase entropy. Ion pairing effects
(neglected in our calculation) may also affect the results for the
ionic species.
The calculated geometry of [RhCl(PPh3)2]2 agrees with the
experimental ones (both the previously reported ethyl acetate
solvate47 and the dichloromethane solvate reported herein) in
terms of bond lengths and angles, except for the bending at the
Rh(μ-Cl)2Rh bridge, whereas the experimental structures have
a flat core. As a matter of fact, the preference for a bent or flat
geometry for edge-sharing square planar d8−d8 dimers rests on
a delicate energetic balance, as previously discussed in detail.54
Very similar structures of [RhCl(PR3)2]2 complexes have been
shown to adopt either a flat (e.g., PiPr3,55 in addition to the
above-mentioned PPh3 structures) or bent (e.g., PMe3,56
t B u 2 P H , 5 7 ( C 2 F 5 ) 2 P C H 2 C H 2 P ( C 2 F 5 ) 2 , 5 8 a n d
iPr2PCH2CH2PiPr259) geometry, and the related [RhF-
(PPh3)2]2 compound adopts either a flat60 or bent61 geometry
(9)
Solvent effects have also been included in the calculations
using the SDM solvation model42 in the CH2Cl2 medium (ε =
8.93) and this turns out to be important for the rationalization
of certain trends. In particular, while solvation has either a
negligible effect or favors the dimer splitting reaction by a few
kcal/mol when adding a neutral ligand, it strongly disfavors the
addition of X−. This is attributable to the higher solvation
energy of the free X− ions relative to the neutral species (see
Table 5, footnote c). Note that free X− is not better solvated
than the [RhX2(PPh3)2]− ion, but the balance is shifted more in
Table 5. Various Energetic Parameters (in kcal/mol) of
RhX(PPh3)2(L), Relative to Those of 1/2[RhX(PPh3)2]2 + L
(X = Cl, I; L = PPh3, X−, PhNH2, Et2NH, and C2H4)
a
b
c
compound
ΔE
ΔG
ΔESMD ΔGSMD,
Gsolv
(a) chloride system
RhCl(PPh3)3
−11.7
−6.8
−4.4
−9.2
−3.8
−9.1
−10.8
−3.7
−12.7
8.2
−4.7
5.8
−43.8
−73.6
−70.6
−36.7
cis-[RhCl2(PPh3)2]−
trans-[RhCl2(PPh3)2]−
13.4
−8.9
7.1
cis-
−3.4
RhCl(PPh3)2(PhNH2)
trans-
3.2
−8.6
5.2
6.4
−3.1
9.2
1.1
−13.7
−1.8
4.3
−8.2
2.3
−39.0
−34.4
−36.2
RhCl(PPh3)2(PhNH2)
cis-
RhCl(PPh3)2(Et2NH)
trans-
RhCl(PPh3)2(Et2NH)
cis-RhCl(PPh3)2(C2H4)
−4.7
−6.2
−0.1
−3.3
−8.8
−10.4
−4.2
−7.4
−33.8
−33.8
trans-
RhCl(PPh3)2(C2H4)
(b) iodide system
RhI(PPh3)3
−9.3
−0.9
−0.6
−6.4
3.3
−0.8
−3.8
−5.5
−1.3
6.4
−9.5
8.9
−1.0
6.0
−42.8
−69.1
−66.9
−35.6
−37.9
cis-[RhI2(PPh3)2]−
trans-[RhI2(PPh3)2]−
cis-RhI(PPh3)2(PhNH2)
11.5
−5.3
2.1
6.5
Figure 10. Relative ΔESMD and ΔGSMD (in parentheses) in kcal/mol
for geometry optimized [RhX2(PPh3)2]− and RhX(PPh3)2L (L =
PPh3, PhNH2, Et2NH and C2H4), relative to [RhX(PPh3)2]2.
−0.2
5.3
trans-
RhI(PPh3)2(PhNH2)
cis-RhI(PPh3)2(Et2NH)
−5.8
6.3
0.5
−10.8
0.0
−4.6
4.2
−34.1
−35.4
trans-
RhI(PPh3)2(Et2NH)
10.4
favor of the reactants side by the solvent, because the
combination of [RhX(PPh3)2]2 and 2X− is better solvated
than 2[RhX2(PPh3)2]−. As expected, solvation free energy is
higher for the systems containing the smaller Cl atom. The
experimentally observed higher stability of the Et2NH adduct
relative to that of the PhNH2 adduct is also reproduced by the
calculation only when the solvent effect is included. The ΔESMD
and ΔGSMD parameters are graphically summarized in Figure
10.
cis-RhI(PPh3)2(C2H4)
−2.3
−4.5
2.0
−5.4
−8.6
−1.1
−6.6
−32.6
−33.6
trans-RhI(PPh3)2(C2H4)
−2.5
a
b
Gas-phase at T = 298.15 K. ΔESMD + (ΔG − ΔE) at 298.15 K.
c
Solvation free energy calculated by the SDM model for the given
compound (values for the reagents: [RhCl(PPh3)2]2, −56.5; [RhI-
(PPh3)2]2, −56.1; PPh3, −14.6; Cl−, −60.3; I−, −50.8; PhNH2, −8.7;
Et2NH, −1.0, C2H4, −1.4).
Note that for all systems except the anionic ones, ΔGSMD
>
in different crystal habits. The calculations suggest preference
for a bent geometry for an isolated [RhCl(PPh3)2]2 molecule,
at least at this level of theory, thus the experimental observation
of a flat structure is probably the consequence of crystal packing
effects.
The relative energies (both electronic and Gibbs free
energies) for all RhX(PPh3)2(L) systems (L = PPh3, X−,
PhNH2, NHEt2 and C2H4; X = Cl, I) are reported in Table 5.
ΔESMD (the Gibbs parameter is more positive or less negative
than the electronic energy), as expected from the reduction in
molecularity (entropy decrease). The singular behavior for the
anionic X− adducts can be ascribed to the increase of rotational
entropy for this system, because no rotational (or vibrational)
modes are associated to the monatomic X− reagent. Roughly
speaking, since the major entropic contribution to the free
12548
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
Table 6. Energies (in kcal/mol, Relative to RhX(PPh3)2(Et2NH) (I) + 1/2 (Et2NH···NHEt2)) and Calculated N−H Stretching
Vibrations (in cm−1)
a
b
c
d
compound
ΔE
ΔG
ΔESMD
ΔGSMD,
ν(NH)
ν(NH)
(a) chloride system
I + 1/2 (Et2NH···NHEt2)
II + 1/2 (Et2NH···NHEt2)
III
IV
V
0.0
1.1
0.0
1.8
7.4
5.6
8.7
8.9
0.0
0.0
2.1
5.8
6.3
9.3
9.4
3491
3392
3496
3479
3514
3430
1.4
−2.2
−3.0
−1.1
−1.3
−3.9
−2.4
−0.5
−0.8
3499
3448
3434
3500
VI
(b) iodide system
I + NHEt2
0.0
1.4
0.0
1.0
6.9
4.7
7.9
8.1
0.0
2.4
0.0
2.0
4.5
4.9
8.2
3503
3389
3505
3473
3508
3442
II + NHEt2
III
IV
V
−1.5
−1.8
0.4
−3.9
−1.6
0.8
3488
3446
3434
3496
VI
0.3
−0.1
7.7
d
a
b
c
Gas-phase at T = 298.15 K. ΔESMD + (ΔG − ΔE) at 298.15 K. Coordinated Et2NH ligand. H-bonded Et2NH ligand.
energy change is given by changes in the translational and
rotational modes, the reactions with the bidimensional (C2H4)
or tridimensional (PPh3, PhNH2) ligands entail the disappear-
ance of 3 translational and 3 rotational modes whereas the
reaction with zero-dimensional X− entails on one hand the
disappearance of 3 translational modes and on the other hand
the generation of 3 rotational modes. This effect, predicted by
the gas phase calculations, will obviously be attenuated in a
condensed phase and further modulated by ion pairing.
The addition of L to [RhX(PPh3)2]2 results in energetic
stabilization except for the addition of X− (the solvent has an
important effect on this equilibrium, as already pointed out
above). Indeed, we were unable to observe any significant
conversion of [RhX(PPh3)2]2 in the presence of excess X−. For
the neutral ligands, the relative stability increases in the order
PhNH2 < C2H4 < PPh3 < Et2NH for each X on the ΔESMD
scale. On the ΔGSMD scale, on the other hand, the order is
PhNH2 < PPh3 < Et2NH < C2H4, in better agreement with the
experimental evidence. Indeed, the RhCl(PPh3)2(C2H4)
complex is sufficiently stable to be isolated, although it readily
looses the ethylene ligand unless kept under a protecting
ethylene atmosphere,23 whereas we have shown above that the
corresponding Et2NH adduct also exists in solution but is not
sufficiently stable to be isolated.
The energy change is less negative (or more positive) when
X = I for all systems. In other words, [RhI(PPh3)2]2 has a
greater relative stability with respect to all its ligand adducts
than the corresponding chloride system. This result is in
agreement with all the available experimental evidence. Indeed,
the more facile dissociation of PPh3 from RhI(PPh3)3 or C2H4
from RhI(PPh3)2(C2H4) relative to the corresponding chloride
systems has previously been discussed.23 The new experiments
reported in the present contribution also illustrate the lower
stability of the other ligand adducts for the iodide system (vide
supra).
isomer. The reason for the different stereochemical preference
may be traced to the competition between the different π
acceptors and π donors that are present in the coordination
sphere. Thus, C2H4 is a stronger π acceptor and prefers to be
located trans to a π donor halide ligand, leading to a preferred
trans geometry, whereas the two more weakly π acceptor PPh3
ligands avoid competing with each other in the cis geometry for
the amine and X− adducts.
Given the infrared evidence for a secondary interaction
between the RhX(PPh3)2(Et2NH) complex and additional free
Et2NH, a calculation was also carried out for the RhX-
(PPh3)2(Et2NH)···NHEt2 adducts. Four different minima could
be located, the same ones for both halide systems. Adduct (III)
derives from the lower energy RhX(PPh3)2(NHEt2) molecule
(I) and the other three (IV, V and VI) from a slightly higher
energy minimum (II). Figure 11 shows a view of all the
molecules for the Cl system (those of the I system are given in
the Supporting Information). The relative energies and the
calculated NH stretching frequencies are reported in Table 6.
In order to take into account the H-bond breaking in the
NHEt2 reagent, the H-bonded dimer (Et2NH···NHEt2) has
been chosen as a reference point.
Structures I and II differ in the orientation of the Et2NH
ligand relative to the rest of the molecule; the NH vector points
toward one of the Ph rings of the cis-PPh3 ligand in the lower
energy I and toward the X ligand in the slightly less stable II.
Structure I can only establish an H bond as a proton acceptor
via the X ligand with the NH bond of the external NHEt2
molecule, leading to III. Structures IV and V feature the same
X···H−N interaction but also an additional N−H···N bond
involving the coordinated amine as proton donor and the outer
sphere amine as proton acceptor, whereas VI contains only the
N−H···N interaction with the coordinated amine providing the
proton to the external amine. The lowest energy structure is III
when the solvation effects are taken into account, whereas IV is
slightly favored in the gas phase. Note, however, that the energy
difference between these two structures is small. The H-bonds
are shorter in IV.
Another notable result of the computational study concerns
the relative stability for each pair of stereoisomers: the trans
geometry is preferred for the C2H4 adduct, whereas all other
ligand adducts show a preference for the cis geometry
(moderate for X−, strong for the amines). This is in perfect
agreement with the experimental evidence, according to which
the ethylene adduct is trans and the Et2NH adduct reported in
this contribution is cis, without the detection of the trans
Structures I and IV also give the best agreement between the
calculated νNH frequencies and the experimental data.
Interestingly, the two NH vibrations in IV (for the coordinated
and for the hydrogen bonded Et2NH) appear at similar
frequencies, slightly red-shifted from that of I. Considering the
12549
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
by treatment with excess nBu4P+I−. The equilibrium has been
analyzed in the presence of the free ions (no cation) as well as
with ion pairs modeled by Me4N+X−, Me4P+X−, and Na+X− and
the relevant results are presented in Table 7. The results in the
absence of cation are interesting since they show how much the
right-hand side is favored by solvation with respect to the gas
phase. Introduction of the Me4N+ cation indicates that the
cation−anion association has a similar effect to that of the
solvent model, since the gas phase parameters are essentially
identical to those in solution. Optimization of the ion pair
geometry in the presence of the solvent model does not greatly
affect the results, relative to those obtained by correcting the
gas phase energy by solvation effects, and the effect of Me4N+
and Me4P+ is essentially equivalent, whereas that of Na+ differs
significantly, stabilizing the right-hand side of the equilibrium
by more than 10 kcal/mol. Clearly, the Na+ results are not
quantitatively relevant because NaCl and NaI are essentially
insoluble in CH2Cl2. However, the stronger Na+Cl− association
relative to Na+I−, which is expected from first principles, is well
reflected by the computed energy values, and the smaller
solubility of NaCl than NaI (not known for CH2Cl2 solution,
but confirmed in most polar aprotic organic solvents such as
acetone and MeCN)62 will in fact further accentuate the
preference of equilibrium (10) for the right-hand side.
Quantitative agreement with experiment is not achieved,
since the equilibrium appears to be displaced to the left-hand
side in the presence of soluble halide sources, however the
calculated values are not too far from the expected outcome,
considering that the calculation of free energy differences in
condensed phases, especially when ionic species are involved, is
one of the most challenging areas of computational chemistry,
and the trend is as expected.
Figure 11. View of the optimized geometries of RhCl(PPh3)2(Et2NH)
(I and II) and RhCl(PPh3)2(Et2NH)···NHEt2 (III−VI).
expected difference between the frequencies computed in the
gas phase and those experimentally measured in solution, these
results allow the assignment of the experimentally observed
band at 3263 cm−1 to a combination of the two νNH bands in
IV. Finally, the data in Table 6 indicate a more favorable H-
bond formation for the Cl system on the E scales (both gas
phase and solution), whereas the process is more favorable for
the I system on the Gibbs free energy scales. The experimental
IR data are clearly demonstrating an easier formation of the H-
bonded adduct for the Cl system, suggesting that the gas phase
entropy correction is not adequate to address this phenomen-
on. Hence, the present system is yet another useful example
pointing out to the care that one has to take when using
computed gas phase thermodynamic data (even when including
the solvation effects) for the prediction of solution behavior.
(10)
CONCLUSION
■
The present investigation has provided strong evidence that the
recently reported12 RhCl3·3H2O/I− catalyst for the hydro-
amination of ethylene by aniline is activated by reduction to a
RhI active species. This catalyst is more robust than the PtBr2/
Br− system, no significant loss of activity after 4 days at 150 °C
being observed in a recycle experiment, whereas the Pt-based
catalyst is totally deactivated by the catalytic medium.6,11 The
catalytic activity of the Rh-based catalyst, however, is much
lower than that of the Pt-based catalyst, as shown by the
reaction kinetic profile. The solution studies of the [RhI-
(PPh3)2]2 system and of the chlorido analogue at room
temperature show ligand addition equilibria to yield mono-
nuclear RhX(PPh3)2(L) species (L = C2H4, PPh3, PhNH2, X−,
and the model Et2NH amine) that are of comparable stability
(marginally more stable according to DFT calculations, for
C2H4, PPh3, and Et2NH; or less stable, for PhNH2 and X−).
The reaction of RhCl3·3H2O in the presence of PPh3, I−, and
PhNH2 leads to [RhCl(PPh3)2]2 as the major species, without
notable halide exchange, thus suggesting that iodide is not
needed at the level of the catalyst resting state. The effect of
high temperature and pressure conditions on the equilibria
cannot be easily predicted, but it may be speculated that either
RhCl(PPh3)2(C2H4) or [RhCl(PPh3)2]2 are the rate-determin-
ing intermediates (resting state) of the catalytic cycle. An ionic
[RhClI(PPh3)2]− or [RhI2(PPh3)2]− complex certainly does
Table 7. Various Energetic Parameters (in kcal/mol) for
Equilibrium (10), Depending on the Nature of the
Countercation
a
cation
none
ΔE
ΔG
ΔESMD
ΔGSMD
b
14.4
14.4
−4.1
−5.9
−5.2
−4.9
−15.6
−4.1
−4.8
c
b
Me4N+
Me4N+
Me4P+
−6.8
−5.8
d
−2.0
−1.2
−14.9
d
d
Na+
a
b
Gas-phase at T = 298.15 K. ΔESMD + (ΔG − ΔE) at 298.15 K.
Optimized in the gas phase and SMD calculation at the frozen
geometry. Optimization with SMD.
c
d
A final computational study has addressed the halide
exchange equilibrium (10), of interest because of the somewhat
unexpected lack of iodide incorporation into [RhCl(PPh3)2]2
12550
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
(20) Hahn, C.; Spiegler, M.; Herdtweck, E.; Taube, R. Eur. J. Inorg.
Chem. 1999, 435−440.
not appear to play a significant role, thus the promoting effect
of iodide on the catalytic activity must be rationalized through
its effect on the rate-determining transition state, which is still
currently unknown. Future work will address the catalytic
mechanism through computations.
(21) Dorta, R.; Egli, P.; Zurcher, F.; Togni, A. J. Am. Chem. Soc. 1997,
119, 10857−10858.
(22) Maitlis, P. M.; Haynes, A.; James, B. R.; Catellani, M.; Chiusoli,
G. P. Dalton Trans. 2004, 3409−3419.
(23) Osborn, J. A.; Jardine, F. H.; Young, J. F.; Wilkinson, G. J. Chem.
Soc. A 1966, 1711−1732.
ASSOCIATED CONTENT
* Supporting Information
■
S
(24) Hartley, F. R.; Murray, S. G.; Potter, D. M. J. Organomet. Chem.
1983, 254, 119−26.
NMR spectrum of [RhCl(PPh3)2]2, Cartesian coordinates of all
optimized geometries (25 pages). This material is available free
of charge via the Internet at http://pubs.acs.org. Crystallo-
graphic data (excluding structure factors) have been deposited
with the Cambridge Crystallographic Data Centre as
supplementary publication no. CCDC 833545. Copies of the
data can be obtained free of charge on application to the
Director, CCDC, 12 Union Road, Cambridge CB2 1EZ, UK
(fax: (+44) 1223−336−033; e-mail: deposit@ccdc.cam.ac.uk).
(25) Colebrooke, S. A.; Duckett, S. B.; Lohman, J. A. B.; Eisenberg,
R. Chem.Eur. J. 2004, 10, 2459−2474.
(26) Hughes, R. P. Rhodium. In Comprehensive Organometallic
Chemistry; Wilkinson, G., Stone, F. G. A., Abel, E. W., Eds.; Pergamon:
Oxford, U.K., 1982; Vol. 5, pp 277−540.
(27) Sharp, P. R. Rhodium. In Comprehensive Organometallic
Chemistry II; Abel, E. W., Stone, F. G., Wilkinson, G., Eds.; Pergamon:
Oxford, U.K., 1995; Vol. 8, pp 115−302.
(28) Altomare, A.; Burla, M.; Camalli, M.; Cascarano, G.;
Giacovazzo, C.; Guagliardi, A.; Moliterni, A.; Polidori, G.; Spagna, R.
J. Appl. Crystallogr. 1999, 32, 115−119.
AUTHOR INFORMATION
Corresponding Author
*Fax: (+) 33-561553003. E-mail: rinaldo.poli@lcc-toulouse.fr.
■
(29) Sheldrick, G. M. Acta Cryst. A 2008, 64, 112−122.
(30) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7−13.
(31) Burnett, M. N.; Johnson, C. K. ORTEPIII, Report ORNL-6895;
Oak Ridge National Laboratory: Oak Ridge, TN, U.S., 1996.
(32) Farrugia, L. J. J. Appl. Crystallogr. 1997, 32, 565.
(33) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J., J. A.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
ACKNOWLEDGMENTS
■
We are grateful to the ANR (Agence Nationale de la
Recherche, Grant No. NT09_442499), the IUF (Institut
Universitaire de France), the CNRS and the RFBR (Russian
Foundation for Basic Research; France-Russia bilateral grant
No. 08-03-92506) for financial support, as well as the CINES
́
(Centre Informatique National de l’Enseignement Superieur)
and the CICT (Centre Interuniversitaire de Calcul de
Toulouse, project CALMIP) for granting free computational
time.
REFERENCES
■
(1) Brunet, J. J.; Cadena, M.; Chu, N. C.; Diallo, O.; Jacob, K.;
Mothes, E. Organometallics 2004, 23, 1264−1268.
(2) Brunet, J. J.; Chu, N. C.; Diallo, O. Organometallics 2005, 24,
3104−3110.
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
revision A.1; Gaussian, Inc.: Wallingford, CT, 2009.
(34) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−241.
(35) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123−141.
(36) Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.;
Schwerdtfeger, P.; Pitzer, R. M. Mol. Phys. 1993, 78, 1211−1224.
(37) Kuchle, W.; Dolg, M.; Stoll, H.; Preuss, H. J. Chem. Phys. 1994,
100, 7535−7542.
(38) Leininger, T.; Nicklass, A.; Stoll, H.; Dolg, M.; Schwerdtfeger, P.
J. Chem. Phys. 1996, 105, 1052−1059.
(39) Ehlers, A. W.; Bohme, M.; Dapprich, S.; Gobbi, A.; Hollwarth,
A.; Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 111−114.
(3) Brunet, J.-J.; Chu, N.-C.; Rodriguez-Zubiri, M. Eur. J. Inorg. Chem.
2007, 4711−4722.
(4) Rodriguez-Zubiri, M.; Anguille, S.; Brunet, J.-J. J. Mol. Catal. A
2007, 271, 145−150.
(5) Dub, P. A.; Rodriguez-Zubiri, M.; Daran, J.-C.; Brunet, J.-J.; Poli,
R. Organometallics 2009, 28, 4764−4777.
(6) Dub, P. A.; Rodriguez-Zubiri, M.; Baudequin, C.; Poli, R. Green
Chem. 2010, 1392−1396.
(7) Dub, P. A.; Filippov, O. A.; Belkova, N. V.; Rodriguez-Zubiri, M.;
Poli, R. J. Phys. Chem. A 2009, 113, 6348−6355.
(8) Dub, P. A.; Daran, J.-C.; Levina, V. A.; Belkova, N. V.; Shubina, E.
S.; Poli, R. J. Organomet. Chem. 2011, 696, 1174−1183.
(9) Dub, P. A.; Poli, R. J. Mol. Catal. A 2010, 324, 89−96.
(10) Dub, P. A.; Poli, R. J. Am. Chem. Soc. 2010, 132, 13799−13812.
(40) Check, C. E.; Faust, T. O.; Bailey, J. M.; Wright, B. J.; Gilbert, T.
M.; Sunderlin, L. S. J. Phys. Chem. A 2001, 105, 8111−8116.
(41) Glukhovtsev, M. N.; Pross, A.; McGrath, M. P.; Radom, L. J.
Chem. Phys. 1995, 103, 1878−1885.
́
(11) Dub, P. A.; Bethegnies, A.; Poli, R., Organometallics accepted.
(12) Baudequin, C.; Brunet, J.-J.; Rodriguez-Zubiri, M.
Organometallics 2007, 26, 5264−5266.
(13) Coulson, D. R. Tetrahedron Lett. 1971, 429−430.
(14) Coulson, D. R. U.S. Patent 3,758,586 1973.
(15) Diamond, S. E.; Mares, F.; Szalkiewicz, A. Fundam. Res.
Homogeneous Catal. 1979, 3, 345−358.
(16) Diamond, S. E.; Szalkiewicz, A.; Mares, F. J. Am. Chem. Soc.
1979, 101, 490−491.
(17) Diamond, S. E.; Mares, F. U.S. Patent 4,215,218, 1980.
(18) Steinborn, D.; Taube, R. Z. Chem. 1986, 26, 349−359.
(19) Krukowka, E.; Taube, R.; Steinborn, D. Patent DD 296,909
1991.
(42) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B
2009, 113, 6378−6396.
(43) Braga, A. A. C.; Ujaque, G.; Maseras, F. Organometallics 2006,
25, 3647−3658.
(44) Dowerah, D.; Radonovich, L. J.; Woolsey, N. F.; Heeg, M. J.
Organometallics 1990, 9, 614−620.
(45) Petrucci, M. G. L.; Kakkar, A. K. Organometallics 1998, 17,
1798−1811.
(46) Ball, G. E.; Cullen, W. R.; Fryzuk, M. D.; Henderson, W. J.;
James, B. R.; MacFarlane, K. S. Inorg. Chem. 1994, 33, 1464−8.
12551
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552

Inorganic Chemistry
Article
(47) Curtis, M. D.; Butler, W. M.; Greene, J. Inorg. Chem. 1978, 17,
2928−2931.
(48) Vallarino, L. M. Inorg. Chem. 1965, 4, 161−165.
(49) Brunner, H.; Zettler, C.; Zabel, M. Z. Anorg. Allg. Chem. 2003,
629, 1131−1135.
(50) Bennett, M. J.; Donaldson, P. B. Inorg. Chem. 1977, 16, 655−
660.
(51) Poli, R.; Owens, B. E.; Linck, R. G. Inorg. Chem. 1992, 31, 662−
667.
(52) Mata, J. A.; Maria, S.; Daran, J.-C.; Poli, R. Eur. J. Inorg. Chem.
2006, 2624−2633.
(53) Linck, R. G.; Owens, B. E.; Poli, R.; Rheingold, A. L. Gazz.
Chim. Ital. 1991, 121, 163−168.
(54) Aullon, G.; Ujaque, G.; Lledos, A.; Alvarez, S.; Alemany, P.
Inorg. Chem. 1998, 37, 804−813.
(55) Binger, P.; Haas, J.; Glaser, G.; Goddard, R.; Kruger, C. Chem.
Ber. 1994, 127, 1927−1929.
(56) Wang, K.; Goldman, M. E.; Emge, T. J.; Goldman, A. S. J.
Organomet. Chem. 1996, 518, 55−68.
(57) Boettcher, H.-C.; Mayer, P. Z. Naturforsch. B 2008, 63, 1035−
1039.
(58) Schnabel, R. C.; Roddick, D. M. Inorg. Chem. 1993, 32, 1513−
1518.
(59) Oster, S. S.; Jones, W. D. Polyhedron 2004, 23, 2959−2965.
(60) Grushin, V. V.; Marshall, W. J. J. Am. Chem. Soc. 2004, 126,
3068−3069.
(61) Marshall, W. J.; Aullon, G.; Alvarez, S.; Dobbs, K. D.; Grushin,
V. V. Eur. J. Inorg. Chem. 2006, 3340−3345.
(62) Burgess, J. Metal Ions in Solution; Ellis Horwood: New York,
1978.
12552
dx.doi.org/10.1021/ic201508m | Inorg. Chem. 2011, 50, 12539−12552