Improving the Catalytic Activity in the Rhodium-Mediated Hydroformylation of Styrene by a Bis(N-heterocyclic silylene) Ligand

Article abstract of DOI:10.1002/ejic.201700148

For the first time, a significant boost in catalytic activity in the rhodium-catalysed hydroformylation of an alkene by using a bidentate bis(N-heterocyclic silylene) ligand is reported. This is shown by the hydroformylation of styrene at 30 bar CO/H₂ pressure in the presence of [HRh(CO)(PPh₃)₃] with an excess of the ferrocenediyl-based bis-NHSi ligand 4, [({η⁵-C₅H₄{PhC(NtBu)₂}Si})₂Fe], which results in superior catalytic activity, compared with the bidentate diphosphines DPPF (3a) and xantphos (3b). In contrast, the hydroformylation of styrene in the presence of [HRh(CO)(PPh₃)₃] with excesses of the monodentate NHSi ligands [{PhC(NtBu)₂}SiNMe₂] (1) and [{C₂H₂(NtBu)₂}Si:] (2) at 30 bar CO/H₂ pressure revealed considerably slower conversion to the aldehyde products than [HRh(CO)(PPh₃)₃], with or without an excess of PPh₃, showing catalyst deactivation. Surprisingly, the germanium analogue of 4 is shown to be virtually catalytically inactive. The superior activity of 4, compared with the xantphos-containing benchmark system, is rationalized on the basis of solution NMR spectroscopic studies, and the comparative catalyst cycles are elucidated using density functional theory (DFT) methods. The latter quantum-chemical studies explain very well the favourable energy pathway for the hydroformylation of styrene using 4 versus xantphos.

Full text of DOI:10.1002/ejic.201700148

A Journal of
Accepted Article
Title: Improving the Catalytic Activity in Rhodium-Mediated
Hydroformylation of Styrene by a Bis(N-heterocyclic silylene)
Ligand
Authors: Matthias Driess, Marcel Schmidt, Reinhard Schomäcker,
Burgert Bloom, and Tibor Szilvási
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201700148
Link to VoR: http://dx.doi.org/10.1002/ejic.201700148

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
Improving the Catalytic Activity in Rhodium-Mediated
Hydroformylation of Styrene by a Bis(N-heterocyclic silylene)
Ligand
Marcel Schmidt,[a] Burgert Blom,[b],[c] Tibor Szilvási,[d] Reinhard Schomäcker, *[a] and Matthias Driess*[c]
Abstract: For the first time, a significant boost in catalytic activity in
the rhodium-catalyzed hydroformylation of an alkene by using a
bidentate bis(N-heterocyclic silylene) ligand is reported. This is shown
by the hydroformylation of styrene at 30 bar CO/H2 pressure in the
presence of HRh(CO)(PPh3)3 with excess of the ferrocendiyl-based
bis-NHSi ligand 4, [ɳ5-C5H4(PhC(NtBu)2}Si)]2Fe, which results in
superior catalytic activity compared to the bidentate diphosphines
DPPF 3a and XantPhos 3b. In contrast, the hydroformylation of
styrene in the presence of HRh(CO)(PPh3)3 with excess of the
performance and selectivity of the hydroformylation process thus
the synthesis and application of new ligands with novel properties
are the center of interest. Ligands with Group 15 elements as
donor atoms have been widely studied in this regard and it has
been shown that phosphines outperform analogous nitrogen
coordinating ligands in catalytic activity.[5,6] Phosphites have also
been shown to exhibit higher activity than phosphines due to the
fact that they have a stronger π-acceptor character and decrease
the energy barrier of the rate determining step.[7–13] Nowadays,
bidentate phosphine ligands such as XantPhos (Scheme 1, 3b)
are widely used because its specific bite angle can improve the
selectivity and avoid undesired branched products.[14–16] Recently,
N-heterocyclic carbenes (NHC) have also been introduced as
ligands in hydroformylation showing high activity.[17–19] Due to the
stable metal-CII (NHC) bond, the complexes are also robust under
hydroformylation conditions and CO or phosphorus based ligands
do not eliminate the NHC.[20]
To explore NHC analogues ligands bearing heavier divalent
Group 14 elements (Si, Ge, Sn, Pb), we set our focus on N-
heterocyclic silylenes (NHSi), which to our surprise have escaped
testing for hydroformylation catalysis. The synthesis and isolation
of the first N-heterocyclic silylene (NHSi, Scheme 1, 2) by Denk
and West in the early 1990s introduced the possibility of
incorporating isolable silicon-based ligands in catalysis.[21]
monodentate NHSi ligands [{PhC(NtBu)2}SiNMe2]
1
and
[{C2H2(NtBu)2}Si:] 2 at 30 bar CO/H2 pressure revealed considerably
slower conversion to the aldehyde products than HRh(CO)(PPh3)3
with or without excess of PPh3 showing catalyst deactivation.
Surprisingly, the germanium analogue of 4 is shown to be virtually
catalytically inactive. The superior activity of 4 compared with the
XantPhos containing benchmark system is rationalized on the basis
of solution NMR studies, and the comparative catalyst cycles are
elucidated using Density Functional Theory (DFT) methods. The latter
quantum chemical studies explain very well the favorable energy
pathway for the hydroformylation of styrene using 4 versus XantPhos.
Introduction
Otto Roelen’s early discovery of the hydroformylation of olefins to
aldehydes in the 1930s has emerged as one of the most important
industrial processes world-wide.[1] Nowadays the total volume of
produced aldehydes via the hydroformylation process overruns
10 million tons per year.[2] Over the last decades mammoth efforts
have been made to optimize the Rh catalysed hydroformylation
reaction.[1–4] The choice of an appropriate ligand is crucial for the
[a]
M. Schmidt, Prof. Dr. R. Schomäcker
Department of Chemistry: Technical Chemistry
Technische Universität Berlin
Straße des 17. Juni 124, 10623 Berlin (Germany)
E-mail: schomaecker@tu-berlin.de
[b]
[c]
Dr. B. Blom
Maastricht Science Program, Faculty of Humanities and Sciences
Maastricht University
P.O. Box 616, 6200 MD, Maastricht (The Netherlands)
Dr. B. Blom, Prof. Dr. M. Driess
Department of Chemistry: Metalorganics and Inorganic Materials
Technische Universität Berlin
Scheme 1. Ligands employed for the rhodium-catalyzed hydroformylation of
styrene in this work.
Straße des 17. Juni 135, 10623 Berlin (Germany)
E-mail: matthias.driess@tu-berlin.de
Dr. T. Szilvási
[d]
Department of Inorganic and Analytical Chemistry
Budapest University of Technology and Economics
Szent Gellért tér 4, 1111 Budapest (Hungary)
For the last twenty years the synthesis and characterization of
new and stable NHSi complexes has received considerable
attention.[22,23] Nowadays there are a range of different stable
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
NHSi ligands with different electronic and steric features. The
potential of NHSi-supported transition-metal compounds as
catalysts is enormous because of their unique properties, in
First, we explored the coordination ability of the ligands 1 and 2
towards RhI in the precursor HRh(CO)(PPh3)3 by NMR
experiments. Scheme 3 illustrates the formed rhodium complexes
with different molar equivalents of 1. As expected, there was a
substitution reaction of the triphenylphosphine (TPP) ligand by the
NHSi 1, when 3 molar equivalents of 1 were added to the
precursor HRh(CO)(PPh3)3. This indicates a strong interaction of
1 to the RhI center due to the stronger σ donor ability of 1 in
comparison to TPP, facilitating the substitution. The 1H NMR
spectrum (in C6D6) shows the clean formation of the tris-
substituted HRh(CO)(1)3 product, consistent with the absence of
31P-coupling for the resonance signal of the hydride ligand at δ =
particular their enhanced -donor/π-acceptor properties.[24–28]
A
relatively large number of transition-metal NHSi complexes have
been synthesized, characterized and tested in different catalytic
and stoichiometric reactions, but compared to their NHC
counterparts, this is still an emerging field, with many avenues still
left open to explore, particularly in the realm of metal complex-
mediated homogeneous catalysis. The donor-stabilized
heteroleptic [{PhC(NtBu)2}(NMe2)Si:] (Scheme 1, 1) silylene is
especially promising because it has been shown that it exhibits a
higher σ-donor strength compared to NHCs and phosphines.[29,30]
In addition, a closely related bidentate analogue has also been
studied (Scheme 1, 4) together with the germanium counterpart
(Scheme 1, 5). Both of them can potentially serve as alternatives
to benchmark diphosphine ligands such as DPPF or XantPhos
(Scheme 1, 3a and 3b) in hydroformylation, inspiring us to
investigate this hitherto unexplored arena.[31]
Herein, we report the first application of NHSi ligands in
hydroformylation, using styrene as a substrate. The remarkable
activity and selectivity of respective Rh complexes are compared
with the current benchmark complexes containing diphosphine
ligands. We observed that the use of the bis-NHSi ligand 4 with a
1,1’-ferrocendiyl backbone outperforms that containing the
diphosphines DPPF 3a and XantPhos 3b in terms of activity.
1
-10.45 ppm (d, 1H, J (H,Rh) = 9.2 Hz). Accordingly, the 31P{1H}
NMR spectrum (C6D6) shows only ‘free’ PPh3. Moreover, the
29Si{1H} NMR spectrum exhibits a clean doublet resonance signal
1
at δ = 62.8 ppm (d, J (Si,Rh) = 80.4 Hz), further confirming the
fully NHSi-substituted product. The observation of a clean doublet
resonance signal in the 29Si NMR spectrum suggests equatorial
substitution of the NHSi ligands in the trigonal-bipyramidal
coordinated Rh center, but Berry pseudo-rotation is another
possibility to explain the highly symmetrical spectra. Variable
temperature NMR experiments show that HRh(CO)(1)3 is stable
in the presence of PPh3, and no evidence of reversion back to the
HRh(CO)(PPh3)3 precursor could be observed over a range of
temperatures (298 K to 378 K). Strikingly, addition of one molar
equivalent of 1 to HRh(CO)(PPh3)3 precludes the formation of the
expected mono-substituted species HRh(CO)(PPh3)2(1) and a
mixture consisting of HRh(CO)(1)3 and HRh(CO)(PPh3)3 is
formed instead, evidenced by NMR studies. This observation can
be explained by the stronger σ donor capacity of 1, kinetically
labilizing the PPh3 ligands in the intermediate species
HRh(CO)(PPh3)2(1) and resulting in their rapid dissociation,
facilitating the formation of HRh(CO)(1)3. In close analogy,
monitoring the reaction of HRh(CO)(PPh3)3 with ligand 2 by NMR
spectroscopy showed an identical reactivity pattern: Reaction of
HRh(CO)(PPh3)3 with 3 molar equivalents of ligand 2 affords, as
expected, the tris-substituted HRh(CO)(2)3 product with
concomitant PPh3 elimination. The rhodium bound hydride shows
Results and Discussion
To learn about the steering properties of NHSi ligands in the
rhodium-catalyzed hydroformylation we probed styrene as a
substrate. Styrene is often used as model substrate because
isomerisation of the double bond does not occur. Scheme 2
shows the hydroformylation of styrene with a homogenously
catalyzed rhodium complex under synthesis gas (CO/H2)
atmosphere yielding the corresponding linear and branched
aldehyde products.
1
a resonance signal at δ = -9.68 ppm (d, 1H, J (H,Rh) = 5.2 Hz)
with no 31P coupling, evidence again complete substitution of
PPh3. Correspondingly, the 29Si{1H} NMR spectrum exhibits a
clean doublet resonance signal at δ = 106.0 ppm (d, 1J (Si,Rh) =
72.0 Hz) in C6D6. As before, the observation of only one doublet
resonance signal suggests equatorial substitution in the trigonal-
bipyrmaidal geometry around the Rh center, or Berry pseudo-
rotation. The reaction of HRh(CO)(PPh3)3 with only one molar
equivalent of 2 yields a mixture of unreacted HRh(CO)(PPh3)3
with HRh(CO)(2)3, with no sign of the expected intermediate
HRh(CO)(PPh3)2(2). Our efforts to isolate the HRh(CO)(1)3 and
HRh(CO)(2)3 complexes failed, because the eliminated PPh3
could not be completely removed even after several
recrystallization procedures.
Scheme 2. Rh catalyzed hydroformylation of styrene to give the linear (l) and
branched (b) aldehydes.
We employed the commercially available HRh(CO)(PPh3)3
complex as precursor for our investigations. The starting point of
our investigations was focussed on monodentate NHSi ligand
systems 1 and 2 (Scheme 1). We tested the heteroleptic silylene,
[{PhC(NtBu)2}(NMe2)Si:] (1),[29] which exhibits a strong σ donor
capacity and the homoleptic [{C2H2(NtBu)2}Si:] (2),[21] respectively.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
rhodium complex by de-coordination of a NHSi ligand might be
hindered because of the strong dative SiRh bond. On the other
hand the coordination-insertion step of styrene to the rhodium
center, which is the initial step of the catalytic cycle,[32,33] was
found to be rate determining for electron donating monodentate
ligands.[13] Probably, the strong σ-donor capacity of the NHSi
ligands hampers styrene coordination leading to deleterious
catalytic performance. Moreover the sterically demanding tert-
butyl groups in both ligands 1 and 2 might also hinder the styrene
coordination, decreasing the activity in hydroformylation.
Scheme 3. Reactivity of the NHSi ligand 1 towards HRh(CO)(PPh3)3. Analogous
products were obtained for the NHSi ligand 2 by means of multinuclear NMR
spectroscopy.
Clearly, the monodentate NHSi ligands decelerate the reaction
rate of the rhodium catalyzed hydroformylation, in part due to the
full substitution observed in the NMR studies (see above). To
avoid the formation of a tris-substituted and kinetically inert NHSi
rhodium complex, we next envisaged the use of a bidentate NHSi
ligand. We reasoned that the use of such a ligand might preclude
tris-substitution of the PPh3 ligands as observed with the
We compared the activity and selectivity of the in situ generated
NHSi rhodium complexes HRh(CO)(1)3 and HRh(CO)(2)3 with
that of the phosphine modified rhodium precursor
HRh(CO)(PPh3)3 in the hydroformylation of styrene at 30 bar
syngas pressure at 50 and 80 ºC respectively. In Figure 1 the
conversion X of the hydroformylation reaction is plotted against
the reaction time reflecting these catalyst runs. At both reaction
temperatures the unmodified precursor complex HRh(CO)(PPh3)3
is more active in comparison to the NHSi-modified rhodium
complexes HRh(CO)(1)3 and HRh(CO)(2)3, respectively.
However, the l:b selectivity ratio of all catalysts are comparable.
Additionally, we tested the unmodified Rh(CO)2(acac) catalyst at
50 °C and 30 bar syngas pressure. As expected, the unmodified
catalyst shows the highest conversion because there is no ligand
preventing the styrene coordination. According to the literature,
the l:b ratio of the unmodified catalyst is rather low (l:b=16:84) and
comparable to that of the modified ones.[32]
monodentate ligands, affording
a
complex of the type
HRh(CO)(PPh3)(κ2-L2) (L2
=
bis-NHSi ligand). This should
enhance the catalytic activity, since the higher σ-donor capacity
of the bidentate ligand, and, in turn, facilitate dissociation of the
remaining PPh3 ligand, and thereby promote coordination of the
substrate. For this purpose, we decided to employ the bidentate
ligand 4 (Scheme 1), linked with a ferrocendiyl bridge, which was
synthesized previously in our group.[31] To compare the catalytic
activity, we decided to include the analogous bis-germylene
ligand (bis-NHGe) 5 (Scheme 1) in our investigations, and to
compare them with the ability of the bidentate phosphine
XantPhos (3b in Scheme 1), typically employed in Rh-mediated
hydroformylation. Furthermore, we tested DPPF (3a in Scheme
1) as ligand because its backbone is the same compared to the
ligands 4 and 5.
The coordination behaviour of diphosphines is well known and
has been investigated in detail.[34] To understand the coordination
behaviour of the related bis-NHSi and -NHGe rhodium complexes,
we performed several NMR experiments as before with the
monodentate ligands 1 and 2. These studies categorically show
that the bidentate ligands 4 and 5 form the desired and expected
complexes HRh(CO)(PPh3)(κ2-4) and HRh(CO)(PPh3)(κ2-5) via a
substitution process shown in Scheme 4.
Even with an excess of the ligands 4 and 5, the same products
were afforded with unreacted bidentate 4 and 5 present. In the
case of HRh(CO)(PPh3)(κ2-4) an NMR experiment with a ratio of
HRh(CO)(PPh3)3 : 4 = 1:1 showed the clean formation of
HRh(CO)(PPh3)(κ2-4) with a characteristic signal of the hydride
ligand in the 1H NMR spectrum (THF-d8) at δ = -9.43 ppm, which
appears as a triplet resonance signal (two overlapping doublets;
1J (H,Rh) and 2J (H,P) = 11.4 Hz), in accordance with the
expectations for HRh(CO)(PPh3)(κ2-4). Moreover, the 31P{1H}
NMR spectrum shows a doublet resonance signal at δ = 44.7 ppm
(d, 1J (P,Rh) = 98.7 Hz), providing further evidence of the fact that
one PPh3 is still bound to the Rh center. In close analogy, NMR
experiments of HRh(CO)(PPh3)3 with the bidentante NHGe ligand
5 in a ratio of 1:1 revealed a similar picture: in the 1H NMR
spectrum (C6D6), the hydride ligand resonates as a quintet signal
Figure 1. Hydroformylation with HRh(CO)(PPh3)3, HRh(CO)(1)3, HRh(CO)(2)3
and Rh(CO)2(acac) at 80 °C (left) and 50 °C (right). Reaction conditions: 40 g
toluene, nstyrene = 38 mmol, nRh = 0.01 mmol, n(1) = 4 molar eq., n(2) = 3 molar eq.,
p = 30 bar, rpm = 1200. The conversion X and l:b ratios were determined with
GC-FID.
The low activity of the in-situ generated NHSi rhodium complexes
HRh(CO)(1)3 and HRh(CO)(2)3 is potentially due to the increased
σ-donor strength of the NHSi ligands, compared to the TPP, which
reduces their overall activity compared to HRh(CO)(PPh3)3. On
the one hand the formation of the active 16 valence electrons
2
δ = -9.91 ppm (1J (H,Rh) = 10.4 Hz, J (H,P) = 5.0 Hz), whilst in
the 31P{1H} spectrum
a
new signal corresponding to
This article is protected by copyright. All rights reserved.

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
HRh(CO)(PPh3)(κ2-5) is seen at δ = 49.7 ppm (d, 1J (P,Rh) =
164.2 Hz) along with concomitantly liberated PPh3. The attempted
crystallisation of HRh(CO)(PPh3)(κ2-3b/4/5) for single crystal X-
ray diffraction analysis failed due to the inabaility to remove
triphenylphosphane despite several procedures.
Moreover, some results concerning the selectivity of the styrene
hydroformylation are worthy. Firstly, the linear-to-branched ratio
(l:b) increases by raising the temperature for all the catalysts
employed. According to the literature, the formation of the alkyl-
rhodium bond is crucial for the selectivity and this step becomes
reversible at higher temperatures, which leads to a higher l:b ratio
at high temperatures.[37,38] The bite angle of the formed complex
and the steric hindrance are the crucial parameters for the l:b
selectivity of bidentate ligands. Comparing the applied bidentate
ligands, the l:b selectivity increases in the following series:
3b > 3a > 4 > 5
Obviously, the benchmark XantPhos system has the best steric
conditions to increase the formation of the linear product.
Interestingly, although the ligands 3a, 4 and 5 are structurally
similar due to the ferrocendiyl bridge as backbone, the l:b
selectivity differs. Presumably, the sterically demanding tert-butyl
groups at the NHSi and NHGe ligands, compared to the phenyl
group in DPPF, affect also the l:b selectivity of the reaction.
Scheme 4. Coordination behaviour of the bis-NHSi (4) and -NHGe (5) ligands
with the rhodium precursor HRh(CO)(PPh3)3.
Table 1. Conversion X, l:b ratio and turn over frequency (TOF) of the
hydroformylation with HRh(CO)(PPh3)(L2) complexes containing bidentate
ligands L2 at different temperatures. Reaction conditions: 40 g toluene,
nstyrene = 38 mmol, nRh = 0.01 mmol, n(ligand) = 3 eq., p = 30 bar, rpm = 1200.
The conversion X, l:b ratio and the TOF is determined with GC-FID.
Hydroformylation reactions were again carried out with in-situ
generated HRh(CO)(PPh3)(κ2-4) and HRh(CO)(PPh3)(κ2-5)
respectively, and compared to the benchmark systems
HRh(CO)(PPh3)(κ2-3a/3b). Interestingly there is a clear difference
in the catalytic performance between the four bidentate ligands,
which is shown in Table 1. As expected, the bis-NHSi complex
HRh(CO)(PPh3)(κ2-4) is the most active catalyst at all
temperatures. Due to the enhanced σ donor ability of the bis-NHSi
ligand, the electron density at the metal center is increased,
facilitating the dissociation of PPh3 and stabilizing the active
rhodium complex. The enhanced σ donor strength of bidentate
NHSi ligands, compared to analogous bidentate phosphines, was
clearly proved in literature for iridium,[35] and most recently also
for nickel complexes.[27] However, not only the strong σ donor
ability causes the high activity of HRh(CO)(PPh3)(κ2-4) but also
the ability to act as π acceptor, which accelerates the hydride
migration step. The π acceptor properties of NHSi ligands of this
type has been documented in literature.[28,36] Hence
HRh(CO)(PPh3)(κ2-4) is more active and we achieve a higher
conversion and high turnover frequencies (TOF) compared to the
catalysts HRh(CO)(PPh3)(κ2-3a/3b) and the Ge analogue
HRh(CO)(PPh3)(κ2-5), respectively. Furthermore, due to the
structural similarity of the ligands 3a, 4 and 5 due to their
ferrocendiyl bridge, the electronic properties, particularly the -
donor/π-acceptor properties of the ligand, play the major role
concerning the different activity. Surprisingly, employing
Ligand L2
T [°C]
50
Conversion X[a] [%]
l:b ratio
-
TOF[b] [1/h]
3.0
3a
3b
4
traces
2.2
50
43:57
16:84
-
32
50
10
83
5
50
traces
3.9
1.1
3a
3b
4
80
34:66
47:53
12:88
35:65
49:51
25:75
14:86
91
80
29
650
80
99
2600
590
3a
3b
4
100
100
100
100
13
56
3000
9100
100
100
2.2
5
[a] determined at 50 °C after 240 min., at 80 °C after 120 min. and at 100 °C
after 60 min. [b] determined at 50 °C after 60 min., at 80 °C after 30 min. and
at 100 °C after 10 min.
HRh(CO)(PPh3)(κ2-5)
as
potential
precatalyst,
the
hydroformylation is hampered and almost no product formation is
observed. Similar results have been shown in the cobalt-mediated
[2+2+2] cycloaddition with the same ligands
Consequently, the activity of the respective HRh(CO)(PPh3)(L2)
complexes increases in the following series for the bidentate
ligands L2:
The general catalytic cycle of the rhodium-catalyzed
hydroformylation is well established and consistent with the
experimental and theoretical data.[39,40] Scheme 5 illustrates the
proposed catalytic cycle for the hydroformylation of alkenes with
diphosphine ligands. Starting from the trigonal-bipyramidal
hydride species 6, which is an 18-valence-electrons rhodium
4
and 5.[31]
4 > 3b > 3a > 5
This article is protected by copyright. All rights reserved.

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
complex generated in situ, the formation of the active complex 7
is initiated by dissociation of a PPh3 ligand. The π-coordination of
the alkene substrate to the unsaturated rhodium complex 7 results
in the formation of 8, which is converted through the migration of
the hydride to the corresponding rhodium alkyl complex 9. After
the coordination of an additional CO molecule to give 10, the
insertion of CO into the rhodium-alkyl bond affords the rhodium
acyl complex 11. Finally, the coordination of hydrogen, resulting
in complex 12, a reductive elimination step generates the
aldehyde along with the regeneration of the active 16-valence-
electrons hydrido rhodium complex 7.
revealed that the general rhodium hydroformylation mechanism
applies to both 3b and 4 supported Rh complexes using styrene
as substrate (Scheme 5). Moreover, it is clearly visible that the
styrene hydroformylation catalyzed by HRh(CO)(PPh3)(κ2-4)
proceeds via a notably lower energy pathway compared to
HRh(CO)(PPh3)(κ2-3b) (Figure 2).
The results from the DFT calculations further indicate that the
formation of the active Rh complex 7 is energetically more
favorable in the presence of NHSi 4 than that of XantPhos 3b.
The activation barrier leading to 7 is only 13.6 kcal mol-1 in the
case of HRh(CO)(PPh3)(2-4) while that is 16.9 kcal mol-1 for
HRh(CO)(PPh3)(2-3b). This 3.3 kcal mol-1 difference can be
attributed to the stronger σ-donor strength of the NHSi and can
result in higher concentration of the catalytically active Rh-silylene
species which might be one cause of the experimentally observed
higher TOF (see above). The calculations also suggest that the
hydride migration (8->9) is the rate determining step of the
catalytic cycle. The relative activation barrier of the hydride
migration step, which leads to the linear product, is 16.7 kcal mol-
1
in case of the phosphine ligand while 16.0 kcal mol-1 for the
silylene ligand, relative to the previous intermediate. The same
relative activation barrier of the hydride migration step which leads
to the branched product is 16.6 kcal mol-1 in case of the phosphine
ligand while 14.9 kcal mol-1 for the silylene ligand. These results
confirm that the formation of the branched product is definitely
more favorable than the linear as the barrier is 1.1 kcal mol-1
smaller for the branched version in case of the silylene ligand.
This is in very good agreement with the experimental results
which suggests 12:88 ratio in favor of the branched product. The
1.1 kcal mol-1 barrier difference can be translated into 4.8 times
faster reaction rate in case of the branched product (see details
in the Supporting Information) which means 17:83 product ratio in
favor of the branched product at 80 °C. For the phosphine ligand,
the relative activation barrier difference for the hydride migration
step is only 0.1 kcal mol-1 which still suggests minor favor of the
branched product. Using similar reaction rate considerations, as
we just did for the silylene ligand, we can calculate 46:54 ratio in
favor of the branched product. This is also in very good agreement
with the experimentally observed 47:53 ratio at 80 °C. We can
also compare the relative activation energy barrier of the
phosphine and silylene ligands which leads to the same product.
For the branched product, which is the major product, this
activation energy barrier difference is 1.7 kcal mol-1. Using this,
the relative reaction rate can be estimated (see details in
Supporting Information) which gives 11.2 times higher rate for the
silylene ligand. This is higher than the experimentally observed
only 4 times larger TOF at 80 °C, however, considering the error
of DFT calculations this is still a good qualitative agreement which
provide another good explanation for the higher reaction rate.
Scheme 5. Proposed catalytic cycle for the hydroformylation with bidentate
ligands bearing two donor atoms X (X = P, Si) leading to the major branched
product. The same numbering scheme of the intermediates is employed in
Figure 2.
To gain further insight into the catalytic process facilitated by the
bis-NHSi ligand 4 supported rhodium complexes, we performed
Density Functional Theory (DFT) calculations at ωB97X-D/cc-
pVTZ(-PP)(SMD=toluene)//B97-D/6-31G(d)[Rh:cc-pVTZ-PP]
level which have been successfully applied to explore the
reactivity of transition-metal silylene complexes before.[27,41,42] We
wanted to prove that employing 4 is energetically more favorable
than the benchmark ligand XantPhos 3b and it is easier to form
the branched product than the linear, as we observed
experimentally (see above). The calculated reaction profiles
This article is protected by copyright. All rights reserved.

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
Figure 2. Reaction profile of the full catalytic process with the bidentate ligands
3b and 4, respectively. The red line is for HRh(CO)(PPh3)(κ2-3b) and blue line
is for HRh(CO)(PPh3)(κ2-4), the latter clearly following a slightly lower energy
pathway. Selected intermediates (7, 9, 10, 12) along the catalyst pathway for
the HRh(CO)(PPh3)(κ2-4) precursor are shown.
128.0; [D8]THF 67.6 (left signal) ppm) or with an external standard (SiMe4
for 29Si NMR). Concentrated solutions of samples were sealed off in a
Young-type NMR tube for measurements. The NHSi ligands were
prepared according to literature procedure and unless otherwise stated,
general Schlenk and glove box techniques were used thoroughout.[21,29,31]
The rhodium precursor (HRh(CO)(PPh3)3) and the bidentate ligands
4,5-Bis(diphenylphosphino)-9,9-dimethylxanthene
(XantPhos,
purity
Conclusions
97 %) and 1,1´-Bis(diphenylphosphino)ferrocene (dppf, purity 97 %) were
obtained from Sigma Aldrich and used without further purification. The
synthesis gas (1:1 mixture by volume of CO and H2) was achieved from
AIR Liquide. These chemicals are used without further purification.
Styrene was purified by passage over activated alumina, and degassed
three times via freeze-pump-thaw cycles and stored over molecular sieves
for use. Toluene was freshly distilled over sodium and collected prior to
use. The analytical standards 2-Phenylpropionaldehyde (Sigma-Aldrich,
purity 98 %) and 3-Phenylpropionaldehyde (Sigma-Aldrich, purity 95 %)
were obtained from Sigma Aldrich to calibrate the gas chromatograph.
The results presented here reveal the suitability of bis-NHSi
ligands for the catalytic hydroformylation of alkenes. Probably,
due to their enhanced -donor/π-acceptor properties, the activity
is significantly higher compared with the diphosphines DPPF and
XantPhos, but the l:b ratio in selectivity is lower due to the
structural difference of the NHSi ligand. In contrast to the
bidentate silylenes the monodentate ones hamper the
hydroformylation of styrene, probably due to the formation of a
fully substituted rhodium complex. We are currently investigating
this process further, in particular using other bis-NHSi ligands,
and other substrates and will report on these endeavours in due
course.
NMR experiments of HRh(CO)(PPh3)3 + Ligand (1):
In the glove box HRh(CO)(PPh3)3 and Ligand 1 were added in a 1:3 ratio
and [D6]benzene added (0.5 mL) affording a yellow solution. The NMR
spectra indicate selective formation of the complex HRh(CO)(1)3. After the
NMR spectra were recorded an ESI-MS was also recorded showing
selective formation of the trisubstituted product. 1H-NMR (200 MHz,
Experimental Section
1
[D6]benzene, 298 K, ppm): δ = -10.45 (d, 1H, J (H,Rh) = 9.2 Hz) (no P
coupling visible), 1.14 (s, 18 H, 6 x H- Me), 1.43 (s, 54 H, 6 x H-tBu-N),
6.94-6.97 (m, 15 H, 3 x H-Ph), 7.03-7.05 (m, 45 H, H-PPh3). 31P-NMR
(81 MHz, [D6]benzene, 298 K, ppm): δ = -5.3 (s, P-PPh3) (exclusively Free
PPh3, no other resonance signals visible). 29Si-NMR (79 MHz, [D6]benzene,
298 K, ppm): δ = 62.8 (d, 1J (Si,Rh) = 80.4 Hz) (no P coupling visible).
ESI- MS(Toluol): 1042.5480 (expt.) 1042.5503 (calcd.) M●+ , 738.3333
(expt.) 738.3338 (calcd.) [M ─ SiNMe2]●+ , 304.2201 (expt.) 304.2159
(calcd.) [free Ligand + H ]+.
General
All experiments were carried out under dry oxygen-free nitrogen using
standard Schlenk techniques. Solvents were dried by standard methods
and freshly distilled and degassed prior to use. The NMR spectra were
recorded on Bruker spectrometers (AV400 or AV200) referenced to
residual solvent signals as internal standards (1H NMR: [D6]benzene, 7.15;
[D8]THF 3.58 (left signal) ppm and 13C{H} NMR: [D6]benzene,
Repeated experiments of HRh(CO)(PPh3)3 with Ligand 1 in a 1:1 ratio
showed the formation of the trisubstituted product HRh(CO)(1)3 along with
unreacted HRh(CO)(PPh3)3 and free PPh3 in the NMR spectra.
NMR experiments of HRh(CO)(PPh3)3 + Ligand (2):
This article is protected by copyright. All rights reserved.

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
In the glove box HRh(CO)(PPh3)3 and Ligand 2 were added in a 1:3 ratio
and [D6]benzene added (0.5 mL) affording a light orange solution. The
NMR spectra indicate selective formation of the trisubstituted complex. 1H-
NMR (200 MHz, [D6]benzene, 298 K, ppm): δ = -9.68 (d, 1H, 1J (H,Rh) =
5.2 Hz) (no P coupling visible), 1.58 (s, 54 H, 6 x H-tBu), 6.57 (s, 6 H, 3 x
CH2=CH2), 7.03-7.05 (m, 45 H, H-PPh3). 31P-NMR (81 MHz, [D6]benzene,
298 K, ppm): δ = -5.3 (s, P-PPh3) (exclusively free PPh3, no other
resonance signals visible). 29Si-NMR (79.5 MHz, C6D6, 298 K, ppm): δ =
106.0 (d, 1J (Si,Rh) = 72.0 Hz) (no P coupling visible).
autoclave was heated to 110 °C for 1 h and after that the reactor was
evaporated and flushed with nitrogen three times. Then the autoclave was
cooled down to reaction temperature and the reaction mixture was added
during a nitrogen counter flow. The mixture was pressurized to 30 bar
during stirring at 200 rpm. The consumed gas was readjust through a mass
flow controller to realise isobar conditions. The reaction was started at a
stirrer speed of 1200 rpm. Samples were taken, diluted with acetone and
analysed via gas chromatography. Gas chromatography analysis were
operated using a Shimadzu GC-2010 gas chromatograph equipped with
flame ionisation detector (FID) and a Supelcowax 10 fused silica column
(30 m x 0.53 mm x 1 µm film thickness. The temperature programme for
the column was: 140 °C (5 min.), Δ 15 °C min-1 to 200 °C (5 min.). The
temperature of the injector was 220 °C, that of the detector 270 °C. The
split ratio was 30:1. Nitrogen was used as carrier gas with a flow rate of
3 mL min-1. For further details consult the SI.
Repeated experiments of HRh(CO)(PPh3)3 and Ligand 2 in a 1: 1 ratio
showed the formation of the trisubstituted product HRh(CO)(2)3 along with
unreacted HRh(CO)(PPh3)3 and free PPh3 in the NMR spectra.
NMR experiment of HRh(CO)(PPh3)3 + Ligand (4):
DFT calculations
The rhodium precursor HRh(CO)(PPh3)3 and Ligand 4 was dissolved in
0.5 mL [D6]benzene in a 1:1 ratio and transferred to an NMR tube. A light
orange colour is noticed immediately. The NMR spectra show selective
formation of the product HRh(CO)(PPh3)(κ2-4). 1H-NMR (200 MHz,
[D8]THF, 298 K, ppm): δ = -9.43 (ps t: 1H, 1J (H,Rh) and 2J (H,P) = 11.4
Hz) (P coupling visible), 0.87 (s, 18 H, 2 x H-tBu-N), 1.29 (s, 18 H, 2 x H-
tBu-N), 4.15 (m, 8 H, 4 x H-Ferrocene), 7.03-7.70 (m, 55 H, H-PPh3 + m,
10 H, H-Ph). 31P-NMR (81 MHz, [D8]THF, 298 K, ppm): δ = -5.4 (s, P-PPh3)
(Free Ligand), 44.7 (d, 1J (P,Rh) = 98.7 Hz).
DFT calculations were performed at the ωB97X-D/cc-pVTZ(-
PP)(SMD=toluene)//B97-D/6-31G(d)[Rh:cc-pVTZ-PP] level of theory.[43–48]
Stationary points on the potential energy surface (PES) were
characterized by harmonic vibrational frequency calculations.
Temperature was set to the experimentally applied 353 K. Calculations
were carried out using GAUSSIAN 09 program. Further details are
described in the Supporting Information.
Repeated NMR experiments of HRh(CO)(PPh3)3 with excess 4 revealed
the formation of HRh(CO)(PPh3)(κ2-4) with unreacted 4 and no other
products.
Acknowledgements
Financial support by the Deutsche Forschungsgemeinschaft
(Cluster of Excellence UniCat, ExIn 314-2 and Collaborative
Research Center/Transregio 63 "Integrated Chemical Processes
in Liquid Multiphase Systems") is gratefully acknowledged.
NMR experiment of HRh(CO)(PPh3)3 + Ligand (5):
The rhodium precursor HRh(CO)(PPh3)3 and Ligand 5 was dissolved in
0.5 mL [D6]benzene in a 1:1 ratio and transferred to an NMR tube. A light
orange colour is noticed immediately. After the NMR spectra were
recorded, a ESI-MS spectrum of the mixture was recorded. The NMR
spectra show selective formation of the product HRh(CO)(PPh3)(κ2-5). 1H-
NMR (200 MHz, [D6]benzene, 298 K, ppm): δ = -9.91 (ps q, 1H, 1J (H,Rh)
= 10.4 Hz, 2J (H,P) = 5.0 Hz) (P coupling visible), 0.99 (s, 18 H, 2 x H-tBu-
N), 1.38 (s, 18 H, 2 x H-tBu-N), 4.47 (br s, 4 H, 4 x H-Ferrocene), 4.66 (br
s, 4 H, 4 x H-Ferrocene), 7.03-7.05 (m, 30 H, H-PPh3), 7.34-7.42 (m, 10 H,
H-Ph). 31P-NMR (81 MHz, [D6]benzene, 298 K, ppm): δ = -5.4 (s, P-PPh3)
(Free Ligand), 49.7 (d, 1J (P,Rh) = 164.2 Hz). ESI-MS(Toluene):
1186.2065 (expt.) 1186.2119 (calcd.) M●+, 1158.2233 (expt.) 1158.2170
(calcd.) [M ─ CO + H]+ , 898.1246 (expt.) 898.1249 (calcd.) [M ─ CO ─
PPh3 ─ H]+ , 792.5467 (expt.), 792.2125 (calcd.) [free ligand + H]+.
Keywords: Metal Catalysis• Homogeneous Catalysis • Silicon •
Ferrocene Ligands • Germylenes
[1]
B. Cornils, W. A. Herrmann, M. Rasch, Angew. Chem. Int. Ed. 1994,
33, 2144–2163.
[2]
[3]
R. Franke, D. Selent, A. Börner, Chem. Rev. 2012, 112, 5675–5732.
M. Beller, B. Cornils, C. D. Frohning, C. W. Kohlpaintner, J. Mol.
Catal. A Chem. 1995, 104, 17–85.
[4]
A. Gual, C. Godard, S. Castillón, C. Claver, Tetrahedron Asymmetry
2010, 21, 1135–1146.
[5]
[6]
J. T. Carlock, Tetrahedron 1984, 40, 185–187.
Repeated NMR experiments of HRh(CO)(PPh3)3 with excess 5 revealed
the formation of HRh(CO)(PPh3)(κ2-5) with unreacted 5 and no other
products.
V. K. Srivastava, R. S. Shukla, H. C. Bajaj, R. V. Jasra, Appl. Catal.
A Gen. 2005, 282, 31–38.
[7]
[8]
A. M. Trzeciak, J. J. Ziółkowski, J. Mol. Catal. 1986, 34, 213–219.
P. W. N. M. van Leeuwen, C. F. Roobeek, J. Organomet. Chem.
1983, 258, 343–350.
Preparation of the catalyst reaction mixture
The rhodium precursor (0.01 mmol, 9.188 mg,
1 eq.) and the
[9]
M. L. Clarke, D. Ellis, K. L. Mason, A. G. Orpen, P. G. Pringle, R. L.
Wingad, D. A. Zaher, R. T. Baker, Dalt. Trans. 2005, 3, 1294.
A. Van Rooy, J. N. H. de Bruijn, K. F. Roobeek, P. C. J. Kamer, P.
W. N. M. van Leeuwen, J. Organomet. Chem. 1996, 507, 69–73.
R. L. Pruett, J. A. Smith, J. Org. Chem. 1969, 34, 327–330.
C. A. Tolman, J. Am. Chem. Soc. 1969, 1, 2953–2956.
M. Sparta, K. J. Børve, V. R. Jensen, J. Am. Chem. Soc. 2007, 129,
8487–8499.
corresponding ligand was weighed in a glovebox into a 100 mL Schlenk
tube. Distilled styrene (0.038 mol, 4.0 g, 3800 eq.) and distilled toluene
(0.434 mol, 40.0 g) was added.
[10]
Experimental procedure for the hydroformylation
[11]
[12]
[13]
A 100 mL stainless steel autoclave from PREMEX was used to perform
the hydroformylation reactions. Before adding the reaction mixture the
This article is protected by copyright. All rights reserved.

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
[14]
[15]
[16]
[17]
P. C. J. Kamer, P. W. N. M. Van Leeuwen, J. N. H. Reek, Acc.
Chem. Res. 2001, 34, 895–904.
124, 6271–6275; Angew. Chem. Int. Ed. 2012, 51, 6167–6171.
J. Feng, M. Garland, Organometallics 1999, 18, 417–427.
C. Bergounhou, D. Neibecker, R. Mathieu, J. Mol. Catal. A Chem.
2004, 220, 167–182.
[32]
[33]
P. W. N. M. van Leeuwen, P. C. J. Kamer, J. N. H. Reek, P.
Dierkes, Chem. Rev. 2000, 100, 2741–2769.
M. Kranenburg, Y. E. M. Van Der Burgt, P. C. J. Kamer, P. W. N. M.
Van Leeuwen, Organometallics 1995, 14, 3081–3089.
G. K. Ramollo, M. J. López-Gómez, D. C. Liles, L. C. Matsinha, G.
S. Smith, D. I. Bezuidenhout, Organometallics 2015, 34, 5745–
5753.
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
L. Damoense, M. Datt, M. Green, C. Steenkamp, Coord. Chem.
Rev. 2004, 248, 2393–2407.
A. Brück, D. Gallego, W. Wang, E. Irran, M. Driess, J. F. Hartwig,
Angew. Chem. Int. Ed. 2012, 51, 11478–11482.
B. Blom, S. Enthaler, S. Inoue, E. Irran, M. Driess, J. Am. Chem.
Soc. 2013, 135, 6703–6713.
[18]
[19]
W. Gil, A. M. Trzeciak, Coord. Chem. Rev. 2011, 255, 473–483.
A. C. Chen, L. Ren, A. Decken, C. M. Crudden, Organometallics
2000, 19, 3459–3461.
E. Boymans, M. Janssen, C. Müller, M. Lutz, D. Vogt, Dalt. Trans.
2013, 42, 137–142.
[20]
[21]
[22]
M. Bortenschlager, J. Schütz, D. von Preysing, O. Nuyken, W. A.
Herrmann, R. Weberskirch, J. Organomet. Chem. 2005, 690, 6233–
6237.
R. Lazzaroni, A. Raffaelli, R. Settambolo, S. Bertozzi, G. Vitulli, J.
Mol. Catal. 1989, 50, 1–9.
I. Jacobs, B. de Bruin, J. N. H. Reek, ChemCatChem 2015, 7,
1708–1718.
M. Denk, R. Lennon, R. Hayashi, R. West, A. V. Belyakov, H. P.
Verne, A. Haaland, M. Wagner, N. Metzler, J. Am. Chem. Soc.
1994, 116, 2691–2692.
I. del Rıo, O. Pàmies, P. W. N. M. van Leeuwen, C. Claver, J.
́
Organomet. Chem. 2000, 608, 115–121.
M. Haaf, T. a. Schmedake, R. West, Acc. Chem. Res. 2000, 33,
704–714.
N. C. Breit, T. Szilvási, T. Suzuki, D. Gallego, S. Inoue, J. Am.
Chem. Soc. 2013, 135, 17958–17968.
[23]
[24]
[25]
[26]
N. J. Hill, R. West, J. Organomet. Chem. 2004, 689, 4165–4183.
B. Blom, D. Gallego, M. Driess, Inorg. Chem. Front. 2014, 1, 134.
B. Blom, M. Stoelzel, M. Driess, Chem. - A Eur. J. 2013, 19, 40–62.
S. Raoufmoghaddam, Y.-P. Zhou, Y. Wang, M. Driess, J.
Organomet. Chem. 2016, in press,
T. T. Metsänen, D. Gallego, T. Szilvási, M. Driess, M. Oestreich,
Chem. Sci. 2015, 6, 7143–7149.
[43]
[44]
T. H. Dunning Jr, J. Chem. Phys. 1989, 90, 1007.
M. J. Frisch, J. a. Pople, J. S. Binkley, J. Chem. Phys. 1984, 80,
3265.
10.1016/j.jorganchem.2016.07.014.
[45]
[46]
[47]
[48]
A. D. Becke, J. Chem. Phys. 1997, 107, 8554.
S. Grimme, J. Comput. Chem. 2006, 27, 1787–1799.
J.-D. Chai, M. Head-Gordon, Chem. Phys. 2008, 10, 6615–6620.
K. L. Schuchardt, B. T. Didier, T. Elsethagen, L. Sun, V.
Gurumoorthi, J. Chase, J. Li, T. L. Windus, J. Chem. Inf. Model.
2007, 47, 1045–1052.
[27]
[28]
[29]
Y.-P. Zhou, S. Raoufmoghaddam, T. Szilvási, M. Driess, Angew.
Chem. Int. Ed. 2016, 55, 12868–12872.
T. Schmedake, M. Haaf, B. J. Paradise, A. J. Millevolte, D. R.
Powell, R. West, J. Organomet. Chem. 2001, 636, 17–25.
C. W. So, H. W. Roesky, P. M. Gurubasavaraj, R. B. Oswald, M. T.
Gamer, P. G. Jones, S. Blaurock, J. Am. Chem. Soc. 2007, 129,
12049–12054.
[30]
[31]
Z. Benedek, T. Szilvási, RSC Adv. 2015, 5, 5077–5086.
W. Wang, S. Inoue, S. Enthaler, M. Driess, Angew. Chem. 2012,
This article is protected by copyright. All rights reserved.

10.1002/ejic.201700148
European Journal of Inorganic Chemistry
FULL PAPER
FULL PAPER
Successful silicon bite: Due to the
unique electronic properties of N-
heterocyclic silylene (NHSi) ligands
toward transition-metals, the rhodium
mediated hydroformylation of styrene
in the presence of a ferrocendiyl-
based bis-NHSi ligand outperforms
that of comparable diphosphine
complexes. Its higher activity was
investigated experimentally and
theoretically.
Marcel Schmidt, Burgert Blom, Tibor
Szilvási, Reinhard Schomäcker, Matthias
Driess
[Rh]
Page No. – Page No.
Improving the Catalytic Activity in
Rhodium-Mediated Hydroformylation
of Styrene by a Bis(N-heterocyclic
silylene) Ligand
X = Si
Superior
Hydroformylation
Catalyst
This article is protected by copyright. All rights reserved.