Tetraphosphines with tetra(biphenyl)silane and -stannane cores as rigid scaffold linkers for immobilized catalysts

Article abstract of DOI:10.1016/j.jorganchem.2017.03.034

The silane Si(p-C₆H₄-p-C₆H₄Br)₄ and the stannane Sn(p-C₆H₄-p-C₆H₄Br)₄ have been synthesized and characterized. For the silane a single crystal X-ray structure has been obtained. Tetralithiation of the silane with ⁿBuLi and quenching with the corresponding chlorophosphines gave the tetraphosphines Si(p-C₆H₄-p-C₆H₄PR₂)₄ with (R = Ph, Cy, ⁱPr, ^tBu). Tetralithiation of Sn(p-C₆H₄-p-C₆H₄Br)₄ led to cleavage of the Sn-C bonds. Therefore, Sn(p-C₆H₄-p-C₆H₄PPh₂)₄ was synthesized by Br/Li exchange of p-Br-C₆H₄-p-C₆H₄PPh₂ with ⁿBuLi and reaction with SnCl₄. The silanes have been immobilized on silica by generating three phosphonium groups per molecule that were bound to the surface via strong electrostatic interactions. The remaining phosphine group was subsequently coordinated to Wilkinson's catalyst via ligand exchange. All solids have been characterized by quantitative ³¹P solid-state NMR spectroscopy. The catalysts, immobilized via the tetraphosphine linker scaffolds with biphenyl spacers, showed high activities and selectivities with respect to the hydrogenation of 1-dodecene. They have been recycled 9 times in a batchwise manner. All immobilized catalysts eventually formed rhodium nanoparticles that retained their catalytic activity even after being exposed to air.

Full text of DOI:10.1016/j.jorganchem.2017.03.034

Journal of Organometallic Chemistry xxx (2017) 1e11
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Tetraphosphines with tetra(biphenyl)silane and -stannane cores as
rigid scaffold linkers for immobilized catalysts
Joseph H. Baker, Nattamai Bhuvanesh, Janet Blümel*
Department of Chemistry, Texas A&M University, P.O Box 30012, College Station, TX 77842-3012, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The silane Si(p-C6H4-p-C6H4Br)4 and the stannane Sn(p-C6H4-p-C6H4Br)4 have been synthesized and
characterized. For the silane a single crystal X-ray structure has been obtained. Tetralithiation of the
silane with nBuLi and quenching with the corresponding chlorophosphines gave the tetraphosphines
Si(p-C6H4-p-C6H4PR2)4 with (R ¼ Ph, Cy, iPr, tBu). Tetralithiation of Sn(p-C6H4-p-C6H4Br)4 led to cleavage
of the Sn-C bonds. Therefore, Sn(p-C6H4-p-C6H4PPh2)4 was synthesized by Br/Li exchange of p-Br-C6H4-p-
C6H4PPh2 with nBuLi and reaction with SnCl4. The silanes have been immobilized on silica by generating
three phosphonium groups per molecule that were bound to the surface via strong electrostatic in-
teractions. The remaining phosphine group was subsequently coordinated to Wilkinson's catalyst via
ligand exchange. All solids have been characterized by quantitative 31P solid-state NMR spectroscopy. The
catalysts, immobilized via the tetraphosphine linker scaffolds with biphenyl spacers, showed high ac-
tivities and selectivities with respect to the hydrogenation of 1-dodecene. They have been recycled 9
times in a batchwise manner. All immobilized catalysts eventually formed rhodium nanoparticles that
retained their catalytic activity even after being exposed to air.
Received 8 February 2017
Received in revised form
15 March 2017
Accepted 16 March 2017
Available online xxx
© 2017 Elsevier B.V. All rights reserved.
1. Introduction
continuous flow setting.
After experimenting with zirconium phosphate nanoplatelets
Catalysis represents one of the most central aspects of chemis-
try. Unfortunately, the “ideal catalyst” [1] with eternal lifetime,
superior activity and total selectivity has still not been discovered.
Therefore, different approaches have been investigated to improve
the removal and recyclability of originally homogeneous catalysts.
For example, one successful approach is based on biphasic catalysis
using two liquid phases, one fluorous and one conventional organic
phase, and controlling their miscibility via temperature under
different conditions [2]. In another approach a rhodium catalyst
could successfully be adsorbed on a teflon tape and reversibly
released into solution for performing as a homogeneous catalyst
[3].
Our group and others have successfully pursued the immobili-
zation of catalysts on solid supports via diverse linkers [4]. Opti-
mally, an immobilized catalyst retains all the qualities of the
homogeneous analog, while it allows for easy catalyst separation
from the reaction mixture, batchwise recycling, or application in a
[5] and various other oxides as support [6], silica [7] remains the
most favorable support for immobilized catalysts. Silica can easily
be modified with linkers incorporating ethoxysilane groups [8,9], it
is mechanically robust, stable at higher temperatures, and settles
within minutes after the reaction, so that the supernatant con-
taining the products can simply be decanted. After washing the
silica, the covalently bound immobilized catalyst can be used for
the next catalytic run.
Using bifunctional phosphine linkers incorporating ethox-
ysilane groups, immobilized nickel catalysts for cyclotrimerization
[10], palladium and copper catalyst systems for Sonogashira re-
actions [11], and rhodium catalysts for olefin hydrogenation [12]
have been generated. All these catalysts showed unprecedented
activities and lifetimes and could be recycled many times.
Nevertheless, some problems remain. For example, neighboring
catalyst molecules on the surface can dimerize or agglomerate.
Wilkinson's catalyst readily dimerizes already when stirred at room
temperature in benzene [13], and the dimer is catalytically not
active. Accordingly, it was found that diluting the immobilized
Wilkinson-type catalysts on the surface improved the catalytic
performance substantially [12b,c]. However, diluting surface-
Dedicated to Prof. Dr. John Gladysz on the occasion of his 65th birthday.
* Corresponding author.
E-mail address: bluemel@tamu.edu (J. Blümel).
http://dx.doi.org/10.1016/j.jorganchem.2017.03.034
0022-328X/© 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

2
J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
bound species entails a relative increase in the bulk of the support
material, which is an unpopular scenario in industrial settings.
Another, fatal problem arises when linkers decompose upon reac-
tion with silica [14]. Additionally, phosphine linkers incoporating
ethoxysilane groups can be transformed into the corresponding
ethylphosphonium salts at higher temperatures, which are no
longer able to coordinate to metal complexes [15]. The detachment
of the catalyst complex from monodentate linkers, a general
problem for immobilized catalysts which is typically classified as
“leaching”, leads to the gradual loss of metal complexes. We have
investigated this issue using palladium complexes by solid-state
NMR and found that this problem can only be ameliorated by us-
ing favorable chelate linkers [11b,c]. Finally, homogeneous catalysts
that are tethered to the support via linkers with flexible alkyl chains
can reach the support surface and decompose or form nano-
particles [5,12a].
tetraphenylmethane, tetraphenylsilane, and tetraphenylstannane
cores [16,17]. In the following the syntheses of the analogs 1e7
(Scheme 1) with tetra(biphenyl)silane and -stannane cores are
described. Different strategies for syntheses of rigid tetrahedral aryl
systems have been reported in the literature [27], making use of
multiple palladium coupling reactions and boronic acid derivatives,
as well as Grignard reagents. For the sake of simplicity and to avoid
multiple steps in the synthesis, our approach to the biphenyl sys-
tems was analogous to the successful previous syntheses of the
tetraphenylelement compounds [16,17].
Unfortunately, the syntheses of these tetra(biphenyl)element
compounds proved to be more difficult as compared to the previ-
ously synthesized molecules with tetraphenylelement cores. When
the syntheses of the tetrabromo compounds 1 and 6, the precursors
for 2e5 and 7, are pursued according to Scheme 2, the yields are not
optimal.
Recently, we have explored a rigid tetraphosphine linker system
with tetraphenylelement cores [16] that successfully prevented
rhodium catalysts from dimerizing and kept the metal centers at a
safe distance from the reactive silica surface [17]. Indeed,
Wilkinson-type rhodium catalysts with unprecedented activity and
selectivity could be obtained by using the rigid tetrahedral linker
Sn(p-C6H4PPh2)4 [17].
The reason for the lower yields is that 4,4’-dibromobiphenyl can
not be monolithiated selectively [28]. In comparison, the mono-
lithiation of 1,4-dibromobenzene is very selective and dilithiation
does not even take place when applying an excess of four equiva-
lents of n-butyllithium at room temperature (Scheme 3) [29]. The
monolithiation of 4,4’-dibromobiphenyl has been described in the
literature and often used as a benchmark for the selectivity of a
metalation reaction performed with an organolithium or Grignard
reagent. However, even with the use of micromixers and temper-
atures as low as ꢀ78 ꢁC absolute selectivity has not been achieved
[28b]. Scheme 3 presents a typical distribution of mono- versus
dilithiated dibromobiphenyl, as determined by quenching with
methanol [28b]. This distribution has been reproduced in our
hands, and the percentage of monolithiation of 4,4’-dibromobi-
phenyl is among the highest reported in the literature to the best of
our knowledge.
In this contribution we extend the rigid linker theme and
explore whether using the longer biphenyl instead of phenyl units
will further improve the immobilized catalysts by preventing
dimerization and decomposition by contact with the surface, as the
distance between the substituents at the linker should increase
from about 10 to 15 Å. For this purpose, diverse new tetraphos-
phines with tetra(biphenyl)element cores have been synthesized
and fully characterized. They have been immobilized on silica and
Wilkinson-type rhodium catalysts have been coordinated.
Selected linkers have been tested for their activity and recycla-
bility in the important catalytic olefin hydrogenation [18] because
there is already a substantial body of data on rhodium-catalyzed
hydrogenation from our group [12,17]. A comparison with the hy-
drogenation results of other groups, using different catalytic sys-
tems, will be interesting as well [19]. Special interest in the work
described here is placed on a linker with a silicon atom at the core
instead of a tin atom [17] in order to exclude the participation of the
center atom of the linker in the catalytic hydrogenation.
The lack of selectivity of the monolithiation of 4,4'-dibromobi-
phenyl is problematic considering that the dilithiated biphenyl can
lead to oligomers when being reacted with SiCl4 or SnCl4 in the next
step. This further reduces the yields substantially.
After numerous experiments aimed at ameliorating the
outcome of the Br/Li exchange, it could be determined that the best
approach was to remove the dilithiated byproduct from the reac-
tion mixture prior to proceeding with the following synthetic steps.
Fortunately, the dilithiated biphenyl precipitates out of diethyl
ether while the monolithiated biphenyl is soluble in ether at room
temperature. Therefore, the separation of the ether solution from
the precipitate allows the isolation of the mono- and dilithium
salts. Starting from the clean lithium salts limits the amounts of
side products formed in subsequent reactions. In order to test the
Besides our application as linkers for immobilized catalysts,
tetrahedral scaffold-type molecules are also of general interest in
other fields. For example, they function as rigid structural elements
for metal organic frameworks (MOFs) [20], porous organic and
aromatic frameworks (POFs, PAFs) [21] and dendrimers [22].
Furthermore, they are of interest for studying photochemical ef-
fects [23], and for materials chemistry in general [24].
The most powerful method for characterizing amorphous solids
is solid-state NMR spectroscopy [25]. It has been an invaluable tool
for characterizing covalently bound linkers, catalysts, and supports,
and recently the dynamics of surface-adsorbed solid metal com-
plexes could be determined by solid-state NMR analysis [26]. Silica
as the support material is very advantageous for solid-state NMR
studies as it does not interfere with the measurements as other
supports do, for example, alumina containing quadrupolar nuclei
[6], or polymers producing 13C NMR background signals.
In this contribution, new tetraphosphines with tetra(biphenyl)
silane and -stannane cores will be synthesized, characterized,
immobilized, and tested as linkers for immobilized Wilkinson-type
rhodium catalysts for olefin hydrogenation.
2. Syntheses of tetra(biphenyl)element compounds
We have previously reported on tetraphosphines with
Scheme 1. Synthesized and characterized tetra(biphenyl)element compounds 1e7.
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
3
feature is reminiscent of the bending of the conjugated polyyne
chains described earlier in the context of “molecular wires” [32].
The bends in 1 create significant changes in the distances be-
tween the four bromine atoms in each molecule. The two that are
bent towards each other are 14.7 Å apart, whereas the two that are
bent away from each other are 19.0 Å apart. The bending even af-
fects the lengths of the biphenyl units, for example, it shortens the
length of one “arm” from 10.87 Å on the left to 10.84 Å on the right.
It should be noted that in the only other literature known crystal
structure of a tetra(biphenyl)silane [33] the bends are smaller. Due
to the strains in the crystal lattice, the deviations from the ideal
tetrahedral angle of 109.5ꢁ are substantial. For example, the right
molecule in Fig. 1 shows two deviating values of 107.0ꢁ and 110.7ꢁ.
This leads again to very different distances between the bromine
atoms within one molecule, 18.4 versus 16.5 Å. The tight packing of
the molecules within the crystal lattice results in a comparatively
high density of the crystal of 1.58 g/cm3. Finally, with respect to
using these tetrahedral scaffolds as linkers for immobilizing cata-
lysts, it should be noted that they will not clog the large pores of the
silica used in the presented studies, which are on average 40 Å in
diameter.
Scheme 2. Synthesis of 1 and 6.
Scheme 3. Br/Li exchange products with 1,4-dibromobenzene and 4,4’-dibromobi-
phenyl (5% remaining starting material).
Four new tetraphosphines with tetra(biphenyl)silane cores
(2e5) could be synthesized from 1 in high yields (Scheme 4). For
the tetralithiation of 1 with ether as the solvent, the number of
equivalents of nBuLi had to be increased to 20. This excess proved
necessary because even 8 equivalents, an amount that had suc-
cessfully been used previously with the tetraphenyl core linkers
[17], produced a mixture of starting bromides and phosphines after
quenching the reaction with the chlorophosphines at ꢀ78 ꢁC. The
incomplete tetralithiation was proven by 13C NMR after allowing
the reaction mixture to warm to room temperature and stirring for
3 h. Most diagnostic for this purpose was the 13C NMR signal of the
nucleus bound to Br at about 122 ppm. In THF the lithiation pro-
ceeds much more rapidly and a significantly smaller amount of
nBuLi is needed, which is more atom economic and beneficial for
the subsequent purification.
After quenching the tetralithium species with the correspond-
ing chlorophosphines, the products 2e5 (Scheme 4) could easily be
isolated in high purities and yields by removing the solvent, adding
ethanol to the residue, filtering the suspension through a frit and
thoroughly washing the solid product with ethanol. Fig. 2 displays
the solution 31P NMR spectra of the tetraphosphines 2e5 in CDCl3.
Noteworthy is the absence of phosphorus-containing sideproducts,
and that the phosphines are only moderately sensitive towards
oxidation and can tolerate chlorinated solvents. Furthermore, there
is only one signal for each tetraphosphine, confirming the
quality of the separation by the solubility in ether, the precipitate
was quenched with water. The resulting 1H and 13C NMR spectra of
the water-quenched precipitate show that it consisted of pure
dilithium salt. The signals in the spectra perfectly match the liter-
ature values for biphenyl [30]. The unoptimized yield of the isolated
compound after the lithiation was 19%.
Removal of the dilithiated biphenyl prior to subsequent syn-
thesis steps proved to be extremely beneficial for generating the
tetrabromo compound 1 (Scheme 1). The compound was obtained
with high purity in 88% yield. The product could be crystallized by
slow evaporation of acetone producing clear needle-like crystals
with high enough quality to subject them to single crystal X-ray
analysis (Fig. 1) [31]. There are two distinct molecules in the unit
cell which matches the 29Si solid-state NMR spectrum of poly-
crystalline 1, which shows two isotropic lines at ꢀ14.7
and ꢀ15.7 ppm, representing each molecule in the unit cell. This
corresponds nicely with our earlier results on tetra(p-bromo-
phenyl)silane [17].
It is interesting to note the very obvious high degree of bending
in the biphenyl spokes in the molecules that is most probably
enforced by the strain of packing in the crystal lattice (Fig. 1). In one
case, the deviation from linearity amounts to 0.7 Å (Fig. 1, left). This
Fig. 1. Single crystal X-ray structure of 1. Two independent molecules are displayed
[31].
Scheme 4. Synthesis of the tetraphosphines 2e5.
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

4
J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
3. Synthesis of tetra(biphenyl)stannane scaffolds
The synthesis of the tin compound 6 (Scheme 1) in analogy to 1
proceeded without difficulties. However, problems were encoun-
tered when trying to obtain the tetraphosphine 7 using the syn-
thesis that was successful for 2e5 (Scheme 4). The main problem
proved to be that a stoichiometric amount of nBuLi only led to
incomplete Br/Li exchange. A twofold excess of nBuLi, however,
resulted in the cleavage of the Sn-C bond and formation of dili-
thiated biphenyl. This is obvious, for example, when the tetra-
bromide 6 is treated with eight equivalents of nBuLi, the reaction is
quenched with chlorodiphenylphosphine, and 4,4'-bis(diphenyl-
phosphanyl)-1,1'-biphenyl is obtained as the sole phosphorus
containing product (Scheme 5). The product has been identified
unequivocally by 31P and 13C NMR spectroscopy and ESI MS. To
corroborate the attack of the lithiating reagent at the tin center,
tetraphenyltin has been subjected to the same treatment with nBuLi
and chlorodiphenylphosphine, and Ph3P was found as the sole
phosphorus containing product by 31P NMR. Therefore, tetrali-
thiation of 6 is not a viable method for generating 7, due to the Sn-C
bond cleavage under these reaction conditions.
Fig. 2. 31P NMR spectra of tetraphosphines 2e5. Numbers next to the signals corre-
spond to the chemical shift values (ppm).
Since the tetralithiation of 6 with the, according to experience,
required excess of nBuLi [16,17] did not work, other approaches
were tested. Unfortunately, addition of tetramethylethylene
diamine (TMEDA) to activate the lithiating reagent [35] did not lead
to the desired outcome even after varying the stoichiometries and
reaction conditions. For example, using 4 equivalents of nBuLi only
about 50% lithiation was observed. With 6 equivalents about 33% of
arylbromides remained, besides decomposed material. Diagnosti-
cally most valuable in these attempts proved to be 13C NMR that
allowed to identify residual bromoaryl moieties by the C-Br signal
at 121.8 ppm.
A more successful attempt to synthesize 6 is outlined in
Scheme 6. In a first step, the phosphine 8 was prepared in high
yields using the protocol for obtaining the clean monolithiated
biphenyl described above. The lithiation of 8 with nBuLi had to be
optimized next. Since triarylphosphines can undergo a P-C bond
cleavage with an excess of lithiating reagents, only one equivalent
of nBuLi was used with 8. The optimal conditions were identified as
the solvent THF, a reaction temperature of ꢀ78 ꢁC, and 1.25
equivalents of TMEDA with respect to nBuLi. After lithiation of 8 the
quench of the reaction mixture with SnCl4 leads to the tetraphos-
phine 7 (Scheme 6) in a respectable crude yield of 52%.
symmetry of the compounds in solution. None of the signals shows
any “fine structure” or additional signals, which would indicate the
presence of triphosphines or diphosphines stemming from
incomplete Br/Li exchange in the previous step of the synthesis
(Scheme 4). It should also be noted that the tetraphosphines 2e5
are rather soluble in organic solvents, excluding the formation of
oligomers.
The signal assignments in the 1H and 13C NMR spectra are
straightforward. They are based on the assignments of the tetra-
phosphines with tetraphenylelement cores reported earlier [16,17],
and on various two-dimensional techniques, such as 1H,1H COSY,
and 13C,1H COSY NMR spectra. Fig. 3 shows, for example, the COSY
spectrum detailing the assignments of the cyclohexyl ring signals,
which are additionally aided by the NMR study of PhPCy2 reported
in the literature [34]. Furthermore, especially with respect to dis-
tinguishing the signal sets of the two aryl rings in the biphenyl
units, the 1ꢀ3J(31P-13C) values are very helpful. The splittings of the
carbon NMR signals due to 1ꢀ4J(31P-13C) couplings can clearly be
seen in the 13C NMR trace in Fig. 3 (left).
4. Immobilization of the tetraphosphine linkers
It has been described previously that phosphines form ethyl-
phosphonium salts with siloxide counteranions in the presence of
ethoxysilanes and silica [15b,16,17]. Alkylation of arylphosphines
with chloroalkanes does not take place, for example, Ph3P does not
react with the silane used here, 3-chloropropyltriethoxysilane, in
Fig. 3. 13C,1H COSY NMR spectrum of 3 used for signal assignments in the cyclohexyl
rings.
Scheme 5. Generating 4,4'-bis(diphenylphosphanyl)-1,1'-biphenyl by attempted tet-
ralithiation of 6.
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
5
Scheme 6. Syntheses of the tetraphosphine 7 via phosphine 8.
solution even if the components are stirred in toluene at 90 ꢁC for
several days [29]. Furthermore, ethoxysilanes do not react with
arylphosphines in solution even under these harsh conditions. The
silica surface is needed to activate the ethoxysilanes in order for an
ethylgroup to be taken up by the phosphine [15a]. Any di-, tri-, or
tetraethoxysilane can form ethylphosphonium salts from phos-
phines in the presence of silica [15b]. As a precursor in the synthesis
of other phosphine linkers, chloropropylethoxysilane is readily
available in our lab and was used for the immobilization described
here. While the transferred ethyl group clearly stems from the
ethoxysilyl group, the siloxide counteranion could be associated
with the silica surface, or the ethoxysilyl group. The latter is cova-
lently linked to the surface via the residual ethoxy groups, so the
linkage of the phosphine might not only be electrostatic in nature,
but it could be a combination of electrostatic cation-anion in-
teractions and covalent bonding via the surface-bound ethoxysilyl
group. The latter is most likely, as the phosphonium salts are
strongly bound to the support surface, and they cannot be washed
off the support even by polar organic solvents. It has been
demonstrated with quantitative 31P MAS NMR spectroscopy that in
the case of tetrahedral tetraphosphine linker scaffolds with tetra-
phenylelement cores three phosphine groups are transformed into
phosphonium groups that are bound to the surface under the right
conditions [16,17]. Due to the tetrahedral geometry, one phosphine
group cannot react with surface-activated ethoxysilyl groups and
remains unchanged, ready to coordinate a metal complex [16,17].
Using this principle, together with the optimized reaction condi-
tions, the tetraphosphines 2e4 have been immobilized successfully
to give 2i-4i as depicted in Scheme 7.
Fig. 4. 31P MAS NMR spectra of the immobilized linkers 2i-4i. The chemical shifts are
given in ppm.
spectra in Fig. 4 have been recorded only with high-power 1H
decoupling and MAS without magnetization transfer, combined
with a long relaxation delay of 10 s [16]. The chemical shifts of the
downfield phosphonium and upfield phosphine resonances lie
within the expected ranges [15e17]. The signal intensity ratios are
about 3:1 for the phosphonium and the phosphine signals. The
exact values are given in Table 1. Therefore, it can be concluded that
the tetraphosphines with tetrabiphenylsilane cores match the
immobilization behavior of the previously studied tetraphosphines
with tetraphenylsilane and -stannane cores [16,17]. The configu-
ration with three “feet” down and one up, as displayed in Scheme 7
is confirmed.
The surface coverages of 2i-4i have been obtained by reacting
the silica with known specific surface area with an amount of tet-
raphosphine that creates a submonolayer of molecules on the
surface, as estimated by the footprint of the tetrahedral molecules,
bound by three “feet”. After the reaction the supernatants have
been checked carefully for residual phosphines by 31P NMR. The
surface coverages are given in Table 1. They were chosen to lie
within the typical range for tetraphosphines with tetraphenylele-
ment cores [16].
The number of binding phosphonium groups versus phosphine
groups in 2i-4i has been determined by quantitative 31P MAS
(Fig. 4). For this purpose, cross-polarization (CP) [25] cannot be
applied, because it would give the phosphonium signals an “unfair”
boost in signal intensity due to the many alkyl protons in the ethyl
groups that provide optimal magnetization transfer. Therefore, the
5. Generating immobilized wilkinson-type catalysts
The immobilized Rh complexes 2iRh, 3iRh and 4iRh have been
generated from the immobilized linker phosphines 2i, 3i, and 4i,
respectively. This has been achieved via ligand exchange by stirring
the modified silica with a slight excess of Wilkinson's catalyst,
ClRh(PPh3)3, at room temperature (Scheme 8). Within a few mi-
nutes the originally white silica support turns orange while the
color of the supernatant is fading. After decanting the supernatant
and washing the silica with toluene to remove PPh3 and excess
Wilkinson's catalyst, the solid is dried in vacuo and subjected to 31
P
solid-state NMR analysis.
The 31P MAS NMR spectra of 2iRh-4iRh show that no surface-
Table 1
31P NMR chemical shifts of resonances from phosphonium (Pþ) and phosphine (P)
groups, their integral ratios, and surface coverages for 2i-4i.
Immobilized Species
d
(
31Pþ)
d(
31P) Ratio Pþ:P Molecules per 100 nm2
2i
3i
4i
23.5
31.3
37.5
ꢀ5.5
1.0
10.4
3:1.3
3:1.5
3:1.1
2.9
4.7
4.0
Scheme 7. Immobilization of 2e4 on silica to generate 2i-4i.
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

6
J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
quantitative within a time window of 60 h [17].
Regarding, in comparison, the immobilized catalysts 2iRh-4iRh,
there are three important aspects. (a) The primary question is
whether the catalytic performance and especially the longevity
improves when the longer biphenyl units are used in the linkers,
since the distances to the support and to neighboring catalysts are
increased. (b) Furthermore, it is of interest to check whether the
record hydrogenation activity reported for the rigid stannane linker
reported previously [17] is due in part to the presence of the tin in
the center that could function as an activator for hydrogen.
Therefore, linkers with silicon in the center that does not influence
catalysis are chosen here. (c) Finally, it will be important to see
whether the increased distance between the metal centers pre-
vents nanoparticle formation or at least delays its onset. Nano-
particle formation at an early stage of catalysis, due to reduction of
Rh(I) to Rh(0) by hydrogen, has been found primarily when using
linker systems with flexible alkyl chains [5,12a].
In order to allow for optimal comparability with the previous
studies, the same substrate, 1-dodecene, reaction conditions, and
hydrogenation apparatus were applied for the catalytic runs
[5,12a]. The immobilized catalyst 2iRh was studied first because it
provides a silicon atom in the center and the biphenyl “distance
holders”, but leaves the coordination of the rhodium center by
triarylphosphine groups in place for a direct comparison with the
results in the literature [17].
Fig. 6 shows the results of 1-dodecene hydrogenation with 2iRh.
The data points were collected during each hydrogenation cycle by
monitoring the hydrogen gas consumption and recording the vol-
ume of H2 every hour. After each catalytic run, the immobilized
catalyst was allowed to settle and the supernatant was removed.
The catalyst was subsequently washed with toluene before the next
catalytic run was started for checking the batchwise recyclability.
Fig. 6 shows that the catalytic activity decreases gradually with
every run. However, even in the eighth run the catalytic reaction is
completed within 40 h. This performance of 2iRh is comparable to
the results obtained previously [17], but does not exceed expecta-
tions. Interestingly, the differences between the completion times
are not as pronounced as the ones reported in earlier work using
the tetraphenylstannane linker [17].
Overall, at this point regarding question (a) one can safely
conclude that the activity and recyclability of the catalyst is similar
when using linkers with phenyl versus biphenyl spacers. Perhaps
the advantage that the larger distances between the metal centers
and to the surface provides is partly diminished by the linkers
filling a large portion of the pore space and thus impeding the
diffusion of the substrate. This would explain why the reaction is
Scheme 8. Immobilized catalysts 2iRh-4iRh, generated from 2i-4i by ligand exchange
with Wilkinson's catalyst.
adsorbed PPh3 [29] is left in the silica. This can, for example, be seen
in Fig. 5. Furthermore, the signals of the uncoordinated phosphines
vanish when the complexes are bound. This can clearly be seen for
3iRh versus 3i (Fig. 5), where the signal of the PCy2 group at
1.0 ppm vanishes upon coordination. Unfortunately, the 31P NMR
signals of the bound Wilkinson-type catalyst at about 30 ppm [12d]
are overlapping with the phosphonium resonance and are there-
fore not resolved. However, the presence of unwanted side-
products, like phosphine oxides, can be excluded, as those
resonances [15a,36] would have larger chemical shift anisotropies
[25a,b] and their rotational sidebands would show prominently in
the spectrum of Fig. 5.
6. Catalytic hydrogenation and recycling
Rigid scaffold-type linkers are favorable for immobilized cata-
lysts because they hold the catalyst at a safe distance from the
reactive support surface. In this way, the premature decomposition
of the catalyst by contact with the silica surface is prevented.
Furthermore, due to the rigid nature of the linkers, the coordinated
metal complexes cannot form dimers [13] with neighboring cata-
lysts that would no longer be catalytically active. It has been
demonstrated with Wilkinson-type catalysts immobilized via rigid
linkers with tetraphenylstannane cores previously that this is a
winning concept [17]. These catalysts could be recycled in a
batchwise manner for 30 times, and the olefin conversion remained
Fig. 6. Percent consumption of H2 over time for nine batchwise catalytic runs with
2iRh. The last catalytic run was performed without inert atmosphere.
Fig. 5. 31P MAS spectra of 3i and 3iRh.
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
7
slower already in the first run as compared to the characteristics
found in earlier work [17]. Experiments with silica with larger pores
will be necessary to answer this question.
Furthermore, one can tentatively answer question (b), the cat-
alytic activity and recyclability are not influenced in a major way by
the tin [17] versus silicon center of the linker.
One of the factors indicating the presence of rhodium nano-
particles is that catalytic activity persists even when oxygen is
allowed into the system. This has been described previously for
rhodium nanoparticles formed when Wilkinson-type catalysts
were immobilized via linkers with long alkyl chains [12a]. When
2iRh was exposed to oxygen for several hours after the eighth run
and used for the catalytic cycle #9 without inert gas atmosphere,
the catalytic activity deviate only minimally from the one observed
in run 8 (Fig. 6). Therefore, one has to conclude that question (c) can
also be answered now. The more extended biphenyl containing
linker scaffold does not prevent nanoparticle formation. Taking a
close look at the characteristics displayed in Fig. 6 one might
speculate that the onset of nanoparticle formation occurs in the
fourth run, when the activity transitions into a slower, then
persistent mode. This would mean that the onset of nanoparticle
formation is delayed for 2iRh, as nanoparticles already form within
the first hours of the initial run for the catalyst tethered with long
alkyl chains [12a].
In order to test the influence of alkyl versus aryl substituents at
the phosphine groups of the biphenyl linker system, the charac-
teristics of the catalyst 4iRh with PiPr2 groups at the ends of the
linker scaffold were investigated. Most importantly, this catalyst
needs a reaction temperature of 80 ꢁC before it becomes active.
When catalyst 4iRh is then recycled in a batchwise manner, the
activity stays about the same in the first three cycles (Fig. 7).
However, in the 4th cycle, the hydrogenation slowed down sub-
stantially without obvious reason.
Fig. 8. Images of the catalyst 4iRh before (left) and after (right) the first hydrogenation
cycle. The white spheres are the stirring bar.
synthesized and characterized. Three tetraphosphines were
immobilized onto silica (2i-4i) and investigated by solid-state NMR.
Surface-bound Wilkinson-type rhodium catalysts have been
generated by ligand exchange with the immobilized linkers (2iRh-
4iRh). Catalytic hydrogenation was performed using the catalysts
immobilized with 2iRh and 4iRh. The former could be recycled at
least 9 times without significant loss of activity and the reactions
were completed within 40 h. Based on the appearance change of
the catalyst and previous results with similar linkers with tetra-
phenylelement cores, the formation of nanoparticles was found.
Exposing the used catalyst to air and running another catalytic
cycle confirmed the presence of catalytically active rhodium
nanoparticles that were inert against oxygen. In future work
different pore sizes for the support will be explored in combination
with 2iRh to check whether substrate diffusion slows the catalytic
runs. Furthermore, a chelate phosphine group will be attached to
the scaffolds which will reduce the detachment of the metal
complex as compared to the present monodentate phosphine
groups at the linker scaffolds.
It is noteworthy with respect to point (c) that the material 4iRh
darkened gradually and turned from originally orange to black at
the end of the first run (Fig. 8). This indicates the eventual forma-
tion of nanoparticles which proceeded faster than for 2iRh. This
might be due to the higher temperature applied, or the nature of
the phosphine groups. Again, in order to prove the nanoparticle
formation during catalysis, the catalyst was exposed to air after the
fourth run. The material remained catalytically as active as in run 4,
meaning that no molecular rhodium catalyst that would get
deactivated under these conditions remained.
8. Experimental section
All reactions were performed by using standard Schlenk tech-
niques or in a glove-box in an oxygen-free argon atmosphere. The
solvents were dried by boiling over Na, distilled and kept under
nitrogen. Alternatively, they were obtained from a solvent purifi-
cation system. All the immobilization experiments were carried out
with Merck silica 40 (specific surface area 750 m2/g; average pore
size 40 Å; particle size 0.063e0.2 mm) that was dried at 300 ꢁC in
vacuo (10ꢀ2 Pa) for at least three days in order to remove adsorbed
water and condense surface silanol groups [8].
The 1H, 13C, and 31P NMR spectra of molecular compounds were
recorded at 499.70, 125.66, and 202.28 MHz on a 500 MHz Varian
Inova spectrometer. The 13C and 31P NMR spectra were measured
with 1H decoupling if not stated otherwise. Neat Ph2PCl
7. Conclusions
Two tetrabromo compounds (1,6) and five tetraphosphines
(2e5,7) with tetra(biphenyl)silane and -stannane cores were
(
d
(
31P) ¼ þ81.92 ppm) in a capillary centered in the 5 mm NMR
tubes was used for referencing the 31P chemical shifts of the
compounds. For referencing the 1H and 13C chemical shifts the
residual proton signals of the solvent CDCl3 and the carbon signal
have been used (
d
(1H) ¼ 7.26 ppm,
d(
13C) ¼ 77.00 ppm). The signal
assignments have been obtained by two-dimensional 1H,1H COSY,
13C,1H HSQC, and 13C,1H HMBC NMR measurements, and by com-
parisons with analogous tetraphosphines with tetraphenylelement
cores [16,17]. All 31P solid-state NMR spectra were recorded on a
Bruker Avance 400 spectrometer, equipped with a 2.5 mm broad-
band MAS probehead and ZrO2 rotors. The modified silica was
loosely filled into the rotors under argon in a glove-box. The
relaxation delays were 10 s for all surface-immobilized compounds,
and the rotational frequency 4 kHz if not mentioned otherwise.
High-power decoupling, but no cross-polarization (CP) [25a,c] was
applied. All spectra were recorded at room temperature (298 K).
The 31P MAS NMR spectra were referenced with respect to 85%
Fig. 7. Hydrogenation of 1-dodecene using the catalyst 4iRh.
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

8
J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
H3PO4 (aq) by setting the 31P NMR peak of solid (NH4)H2PO4 as the
external standard to þ0.81 ppm. For the 29Si NMR spectra the
external chemical shift standard Si(SiMe3)4 was used.
Synthesis of Si(p-C6H4-p-C6H4Br)4 (1):
The synthesis of 3 was performed in analogy to the synthesis of
2 described above. Compound 3 was obtained as a white powder in
a yield of 93%.
In a 500 ml Schlenk flask, 4,4’-dibromobenzene (2.718 g,
8.71 mmol) dissolved in 300 ml of ether. The solution is cooled
to ꢀ78 ꢁC and 4.00 ml of 2.5 M nBuLi (10.0 mmol) in hexanes is
added dropwise. After allowing the mixture to stir at room tem-
perature for 90 min, the solution is separated from the precipitate
via cannula and placed in a different 500 ml Schlenk flask. The flask
is cooled to ꢀ78 ꢁC and 0.20 ml of SiCl4 (0.297 g, 1.75 mmol) is
added dropwise. The reaction mixture is warmed to room tem-
perature and stirred overnight. The solvent is then removed in
vacuo and the crude product is redissolved in 80 ml of chloroform.
Flash chromatography is then performed with chloroform as the
eluent. The solvent is removed in vacuo and the powder is washed 3
times with 20 ml of hexanes. Residual solvent is removed in vacuo
and the product is obtained as white powder in a yield of 88%
(1.470 g, 1.54 mmol). Crystals suitable for single crystal X-ray
analysis are grown by taking 10 ml of a saturated solution of 1 in
acetone, diluting with another 10 ml, and slowly allowing the sol-
vent to evaporate.
1H NMR (CDCl3, 500.1 MHz):
d
¼ 7.75 (H2, d, 3J(H-H) ¼ 8.1 Hz),
7.70 (H3, d 3J(H-H) ¼ 8.3 Hz), 7.64 (H7, d, 3J(H-H) ¼ 7.8 Hz), 7.56 (H6,
dd, 3J(H-H) ¼ 7.3 Hz, 3J(P-H) ¼ 7.3 Hz), 1.98e1.66 (9, 10e, 11e, 13e,
12e, 14e, m), 1.43e0.91 (11a, 13a, 10a, 12a, 14a, m) ppm (Fig. 3); 13
C
NMR (CDCl3, 125.8 MHz): d 141.70 (C4, s), 140.91 (C5, s), 136.91 (C2,
s), 135.18 (C7, d, 2J(P-C) ¼ 19.1 Hz), 134.02 (C8, d, 1J(P-C) ¼ 17.7 Hz),
133.05 (C1, s), 126.54 (C3, s), 126.40 (C6, d, 3J(P-C) ¼ 7.4 Hz), 32.49
(C9, d, 1J(P-C) ¼ 11.2 Hz), 30.01 (C10, d, 2J(P-C) ¼ 15.8 Hz), 28.82
(C14, d, 2J(P-C) ¼ 7.0 Hz), 27.24 (C11, d, 3J(P-C) ¼ 12.6 Hz), 27.00
(C13, d, 3J(P-C) ¼ 7.4 Hz), 26.39 (C12, s) ppm; 31P NMR (CDCl3,
162.0 MHz) 1.92 ppm. mp 117 ꢁC. ESI-MSþ: [Mþ1] 1425.78 (96%)
and [Mþ2] 1426.76 (100%) plus decomposition products, calculated
1425.81 (92%), 1426.81 (100%).
Synthesis of Si(p-C6H4-p-C6H4PiPr2)4 (4):
1H NMR (CDCl3, 500.1 MHz):
d
¼ 7.73 (H2, d, 3J(H-H) ¼ 7.5 Hz),
7.62 (H3, d, 3J(H-H) ¼ 7.8 Hz), 7.58 (H7, d, 3J(H-H) ¼ 8.2 Hz), 7.50
(H6, d, 3J(H-H) ¼ 8.5 Hz) ppm; 13C NMR (CDCl3, 125.8 MHz):
The synthesis of 4 was performed in analogy to the synthesis of
2 described above. Compound 4 was obtained as a white powder in
a yield of 64%.
d
¼ 141.28 (C4, s), 139.67 (C5, s), 136.94 (C2, s), 133.07 (C1, s), 131.98
(C6, s), 128.72 (C7, s), 126.50 (C3, s) 121.95 (C8, s) ppm; 29Si NMR
(CDCl3, 79.5 MHz):
d
d
¼
ꢀ14.36 ppm; 29Si CP/MAS:
1H NMR (CDCl3, 500.1 MHz):
d
¼ 7.75 (H2, d, 3J(H-H) ¼ 8.3 Hz),
¼ ꢀ14.7, ꢀ15.7 ppm (nrot ¼ 10 kHz). mp 149 ꢁC.
7.70 (H3, d, 3J(H-H) ¼ 8.3 Hz), 7.64 (H6, d, 3J(H-H) ¼ 7.8 Hz), 7.57
(H7, dd, 3J(H-H) ¼ 8.0 Hz, 3J(P-H) ¼ 6.6 Hz), 2.16 (H9, m), 1.13 (H10,
dd (3J(H-H) ¼ 7.1 Hz, 3J(P-H) ¼ 15.1 Hz), 0.98 (H11, dd (3J(H-
H) ¼ 6.9 Hz, 3J(P-H) ¼ 11.6 Hz) ppm; 13C NMR (CDCl3, 125.8 MHz):
Synthesis of Si(p-C6H4-p-C6H4PPh2)4 (2):
d
141.86 (C4, s), 141.12 (C5, s), 136.93 (C2, s), 135.05 (C7, d, 2J(P-
C) ¼ 18.5 Hz), 134.06 (C8, d, 1J(P-C) ¼ 16.0 Hz), 133.10 (C1, s), 126.58
(C3, s), 126.46 (C6, d, 3J(P-C) ¼ 7.6 Hz), 22.79 (C9, d, 1J(P-
C) ¼ 10.9 Hz), 19.86 (C10, d, 2J(P-C) ¼ 18.5 Hz), 18.79 (C11, d, 2J(P-
C) ¼ 8.4 Hz) ppm; 31P NMR (CDCl3,162.0 MHz) 10.54 ppm. 29Si NMR
In a 250 ml Schlenk flask, 1 (0.600 g, 0.63 mmol) is dissolved is
(CDCl3, 79.5 MHz):
d
¼ ꢀ15.3 ppm. mp 205 ꢁC. ESI-MSþ: [Mþ1]
75 ml of THF. The solution is cooled to ꢀ78 ꢁC and 1.51 ml of 2.5 M
nBuLi (3.8 mmol) in hexanes is added dropwise and the solution is
stirred for 90 min. Then 0.84 g of ClPPh2 (3.8 mmol) is added
dropwise, the reaction mixture is allowed to warm to room tem-
perature and stirred overnight. The solvent is removed in vacuo
which results in an oil. The product is precipitated with ethanol and
filtered onto a frit under N2. The powder is washed three more
times with 25 ml of ethanol and dried in vacuo. The product is
obtained as 0.695 g of a white powder (0.50 mmol, yield 80%).
1105.51 (100%) and [Mþ2] 1106.49 (78%) plus decomposition
products, calculated 1105.56 (100%), 1106.56 (83%).
Synthesis of Si(p-C6H4-p-C6H4PtBu2)4 (5):
1H NMR (CDCl3, 500.1 MHz):
d
¼ 7.72 (H2, d, 3J(H-H) ¼ 7.8 Hz),
The synthesis of 5 was performed in analogy to the synthesis of
2 described above. Compound 5 was obtained as a white powder in
a yield of 75%.
7.66 (H6, dd, 3J(H-H) 7.5 Hz), 4J(P-H) ¼ 1.6 Hz), 7.62 (H3, d, 3J(H-
H) ¼ 7.4 Hz), 7.39 (H7, t, 3J(H-H) ¼ 7.6 Hz, 3J(P-H) ¼ 7.6 Hz), 7.37
(H10-H12, m) ppm; 13C NMR (CDCl3, 125.8 MHz):
d
141.70 (C4, s),
1H NMR (CDCl3, 500.1 MHz):
d
¼ 7.78 (H7, dd, 3J(H-H) ¼ 7.6 Hz,
141.01 (C5, s), 136.92 (C2, s), 134.18 (C7, d, 2J(P-C) ¼ 19.5 Hz), 133.77
(C10, d, 2J(P-C) ¼ 19.5 Hz), 133.69 (C1, s), 132.69 (C8, d, 1J(P-
C) ¼ 18.6 Hz), 130.31 (C9, d, 1J(P-C) ¼ 21.4 Hz), 128.79 (C12, s),
128.54 (C11, d, 3J(P-C) ¼ 7.0 Hz), 127.15 (C6, d, 3J(P-C) ¼ 6.5 Hz),
126.59 (C3, s) ppm; 31P NMR (CDCl3, 162.0 MHz) ꢀ6.11 ppm. mp
125 ꢁC. ESI-MSþ: [Mþ1] 1377.37 (96%) and [Mþ2] 1378.34 (100%)
plus decomposition products, calculated 1377.44 (96%), 1378.44
(100%).
3J(P-H) ¼ 7.6 Hz), 7.75 (H2, d, 3J(H-H) ¼ 7.3 Hz, 7.70 (H3, d, 3J(H-
H) ¼ 7.3 Hz), 7.62 (H6, d, 3J(H-H) ¼ 7.6 Hz), 1.24 (H10, d, 3J(P-
H) ¼ 11.7 Hz) ppm; 13C NMR (CDCl3, 125.8 MHz):
d 141.80 (C4, s),
141.15 (C5, s), 136.93 (C2, s), 136.90 (C7, d, 2J(P-C) ¼ 15.2 Hz), 136.15
(C8, d, 1J(P-C) ¼ 22.0 Hz), 133.10 (C1, s), 126.59 (C3, s), 126.13 (C6, d,
3J(P-C) ¼ 8.4 Hz), 32.02 (C9, d, 1J(P-C) ¼ 20.2 Hz), 30.48 (C10, d, 2J(P-
31
C) ¼ 14.3 Hz); P NMR (CDCl3, 162.0 MHz) 38.18 ppm. mp 240 ꢁC
(decomp.). ESI-MSþ: [Mþ1] 1217.66 (100%) and [Mþ2] 1218.66
Synthesis of Si(p-C6H4-p-C6H4PCy2)4 (3):
(100%) plus decomposition products, calculated 1217.69 (100.0%),
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
9
1218.69 (92.8%).
to ꢀ78 ꢁC and 4.30 mL of 1.6 M nBuLi (6.88 mmol) in hexanes is
added dropwise. After allowing the mixture to stir at room tem-
perature for 90 min, the solution is separated from the precipitate
via cannula and placed in a different 500 ml Schlenk flask. The flask
is cooled to ꢀ78 ꢁC and 1.12 ml of ClPPh2 (6.88 mmol) is added. The
reaction mixture is warmed to room temperature and stirred
overnight. The solvent is then removed in vacuo and the crude
product is filtered through a frit and washed 3 times with 20 ml of
ethanol. The residual ethanol is removed in vacuo and the product
is obtained as a white powder in a yield of 48% (1.230 g, 2.95 mmol).
Synthesis of Sn(p-C6H4-p-C6H4Br)4 (6):
In a 500 ml Schlenk flask, 4,4’-dibromobenzene (2.718 g,
8.71 mmol) is dissolved in 300 ml of ether. The solution is cooled
to ꢀ78 ꢁC and 6.25 ml of 1.6 M nBuLi (10.0 mmol) in hexanes is
added dropwise. After stirring the mixture at room temperature for
90 min, the supernatant is separated from the precipitate via can-
nula and placed in a different 500 ml Schlenk flask. The flask is
cooled to ꢀ78 ꢁC and 0.20 mL of SnCl4 (0.442 g, 1.70 mmol) is added
dropwise. The reaction mixture is warmed to room temperature
and stirred overnight. The solvent is then removed in vacuo and the
crude product is redissolved in 80 ml of dichloromethane. Flash
chromatography is then performed with dichloromethane as the
eluent. The solvent is removed in vacuo and the powder is washed 3
times with 20 ml of hexanes. Residual solvent is removed in vacuo
and the product is obtained as white powder in a yield of 29%
(0.660 g, 0.49 mmol).
1H NMR (CDCl3, 500.1 MHz):
d
¼ 7.57 (H2, d, 3J(H-H) ¼ 8.3 Hz),
7.53 (H6, d, 3J(H-H) ¼ 7.8 Hz), 7.47 (H3, d, 3J(H-H) ¼ 8.3 Hz), 7.39
(H7,dd, 3J(H-H) ¼ 7.8 Hz, 3J(P-H) ¼ 8.3 Hz), 7.37 (H10-12) ppm; 13
C
NMR (CDCl3, 125.8 MHz):
d
¼ 140.17 (C5, s), 139.41 (C4, s), 136.96
(C9, d, 1J(P-C) ¼ 10.2 Hz), 136.75 (C8, d, 1J(P-C) ¼ 10.2 Hz), 134.43
(C7, d, 2J(P-C) ¼ 19.5 Hz), 133.75 (C10, d, 2J(P-C) ¼ 19.5 Hz), 131.93
(C2, s), 128.83 (C3, s), 128.63 (C12, s), 128.56 (C11, d, 3J(P-
C) ¼ 7.0 Hz,) 126.92 (C6, d, 3J(P-C) ¼ 7.0 Hz), 121.83 (C1, s) ppm; 31
NMR (CDCl3, 162.0 MHz) ꢀ6.17 ppm.
P
Immobilization of Si(p-C6H4-p-C6H4PPh2)4 to give 2i:
Immobilization of 2 on SiO2 via three phosphonium groups:
2.480 g of rigorously dried SiO2 is suspended in 10 ml of toluene,
and a solution of 0.120 g (0.08 mmol) of 2 in 20 ml of toluene,
together with 1.70 g (7.0 mmol) of Cl(CH2)3Si(OEt)3 is added. The
mixture is heated to 90 ꢁC and stirred for 6 d in a gas storage vessel.
After cooling to ambient temperature and allowing the silica to
settle, the supernatant is decanted. Then the silica is washed with
three 25 ml aliquots of toluene and dried in vacuo. Since no traces of
phosphorus containing substances are found in the supernatant,
the surface coverage can be determined to be about 2.9 linker
molecules per 100 nm2 of silica surface. 31P MAS (quantitative):
diso ¼ 23.5 (PPh2Etþ), ꢀ5.5 ppm (PPh2), intensity ratio 3.0:1.3.
Immobilization of Si(p-C6H4-p-C6H4PCy2)4 to give 3i:
1H NMR (CDCl3, 500.1 MHz):
d
¼ 7.74 (H2, d, 3J(H-H) ¼ 8.3 Hz),
7.63 (H3, d, 3J(H-H) ¼ 8.0 Hz), 7.57 (H7, d, 3J(H-H) ¼ 8.8 Hz), 7.47
(H6, d, 3J(H-H) ¼ 8.8 Hz) ppm; 13C NMR (CDCl3, 125.8 MHz):
d
¼ 140.95 (C4, s), 139.75 (C5, s), 137.70 (C2, s), 136.80 (C1, s), 131.98
(C6, s), 128.71 (C7, s), 127.19 (C3, s) 121.88 (C8, s) ppm; 119Sn NMR
(CDCl3, 149.2 MHz):
d
¼ ꢀ124.3 ppm.
Synthesis of Sn(p-C6H4-p-C6H4PPh2)4 (7):
The immobilization procedures for 3i was analogous to the one
described above for 2i. Surface coverage: 4.7 particles/100 nm2, 31
MAS: 31.3 ppm (PCy2Etþ), 1.0 ppm (PCy2), intensity ratio 3.0:1.5.
Immobilization of Si(p-C6H4-p-C6H4PiPr2)4 to give 4i:
The immobilization procedure for 4i was analogous to the one
described above for 2i. Surface coverage of 4.0 particles/100 nm2,
31P MAS: 37.4 ppm (PiPr2Etþ), 10.4 ppm (PiPr2), intensity ratio
3.0:1.1.
P
In a 50 ml Schlenk flask, 8 (0.320 g, 0.77 mmol) is dissolved is
10 ml of THF. The solution is cooled to ꢀ78 ꢁC and 5 ml of THF,
together with nBuLi (0.77 mmol) and TMEDA (0.97 mmol), is added
dropwise while stirring the solution for 5 min. Then SnCl4 (0.050 g,
0.19 mmol) is added dropwise, the reaction mixture is allowed to
warm to room temperature and stirred overnight. The solution is
then passed through SiO2 and the solvent is removed in vacuo. The
residue is washed with ethanol and the suspension is filtered
through a frit under N2. The retained powder is washed three more
times with 15 ml aliquots of ethanol and dried in vacuo. The product
is obtained as 0.148 g of a light yellow powder (0.10 mmol, crude
yield 52%).
Generating Immobilized Catalysts 2iRh, 3iRh and 4iRh:
The linker-modified silica 2i-4i are suspended in 20 ml of
toluene and combined with a solution of ClRh(PPh3)3 (slight excess,
1.1 mmol per 1 mmol of linker molecule) in 10 ml of toluene. After
stirring for 5 h at ambient temperature, the silica is allowed to settle
and the supernatant is decanted. The silica is washed with three
15 ml aliquots of toluene to remove excess ClRh(PPh3)3 and PPh3,
and dried in vacuo. The surface coverages are calculated based on
1H NMR (CDCl3, 500.1 MHz):
d
¼ 7.75 (H2, d, 3J(H-H) ¼ 7.8 Hz),
7.66 (H6, dd, 3J(H-H) 7.5 Hz), 4J(P-H) ¼ 1.5 Hz), 7.63 (H3, d, 3J(H-
H) ¼ 7.8 Hz), 7.40 (H7, dd, 3J(H-H) ¼ 7.5 Hz, 3J(P-H) ¼ 8.3 Hz), 7.37
i
the fact that no signals for uncoordinated PR2 (R ¼ Ph, Cy, Pr)
groups are visible in the 31P CP/MAS spectra, and the knowledge of
the linker surface coverages. For 2iRh 2.9 Rh complexes, 3iRh 4.7
Rh complexes, and for 4iRh 4.0 Rh complexes are bound on
100 nm2 of silica surface.
(H10-H12, m) ppm; 13C NMR (CDCl3, 125.8 MHz):
d 140.98 (C4, s),
139.77 (C5, s), 136.92 (C2, s), 134.20 (C7, d, 2J(P-C) ¼ 19.6 Hz), 133.79
(C10, d, 2J(P-C) ¼ 19.4 Hz), 136.82 (C1, s), 137.19 (C8, d, 1J(P-
C) ¼ 17.4 Hz), 136.92 (C9, d, 1J(P-C) ¼ 25.5 Hz), 128.81 (C12, s),
128.57 (C11, d, 3J(P-C) ¼ 7.0 Hz), 127.22 (C3, s), 127.16 (C6, d, 3J(P-
C) ¼ 7.0 Hz) ppm; 31P NMR (CDCl3, 162.0 MHz) ꢀ6.12 ppm.
Synthesis of p-BrC6H4-p-C6H4PPh2 (8):
9. General hydrogenation procedure
Immobilized catalyst 2iRh (240 mg, containing 10 mg of Wil-
kinson's catalyst, corresponding to 0.010 mmol Rh) is suspended in
5 ml of toluene in a Schlenk flask. The mixture appears opaque and
orange/pink in color. The flask is then attached to the hydrogena-
tion apparatus described earlier [12d] and 1 mmol of 1-dodecene,
dissolved in toluene (5 ml), is added to the suspension of 2iRh with
a syringe through the stopcock. Subsequently the suspension is
stirred vigorously and the hydrogen consumption is monitored.
In a 500 ml Schlenk flask, 4,4’-dibromobenzene (1.952 g,
6.26 mmol) is dissolved in 350 ml of ether. The solution is cooled
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

10
J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
[18] (a) , in: first ed., in: J.G. de Vries, C.J. Elsevier (Eds.), The Handbook of Ho-
mogeneous Hydrogenation, vols. 1e3, Wiley-VCH, Weinheim, 2007;
After complete substrate conversion the catalyst is allowed to
settle, the supernatant is removed via syringe and the material is
washed three times with 5 ml of toluene. To start the second and
following cycles, fresh toluene is added and the described pro-
cedure is repeated.
€
(b) G. Ertl, H. Knozinger, F. Schüth, J. Weitkamp (Eds.), Handbook of Hetero-
geneous Catalysis, second ed., Wiley-VCH, Weinheim, 2008;
(c) M. Wende, R. Meier, J.A. Gladysz, J. Am, Chem. Soc. 123 (2001) 11490 ( and
lit. cited therein);
(d) Adv. Synth. Catal. 345 (2003) vols. 1þ2, Special Issue on Hydrogenation,
including, e.g.: W.S. Knowles, Adv. Synth. Catal. 345 (2003) 3; R. Noyori, Adv.
Synth. Catal. 345 (2003) 15; T. Imamoto, Adv. Synth. Catal. 2003, 345, 79;
Competing financial interest
ꢀ
(e) D.M. Heinekey, A. Lledos, J.M. Lluch, Chem. Soc. Rev. 33 (2004) 175e182.
ꢀ
[19] (a) D. Rutherford, J.J.J. Juliette, C. Rocaboy, I.T. Horvath, J.A. Gladysz, Catal.
The authors declare no competing financial interest.
Acknowledgment
Today 42 (1998) 381e388;
(b) K. Kromm, B.D. Zwick, O. Meyer, F. Hampel, J.A. Gladysz, Chemistry 7
(2001) 2015e2027;
(c) K. Kromm, P.L. Osburn, J.A. Gladysz, Organometallics 21 (2002)
4275e4280;
ꢀ
This material is based on work supported by the National Sci-
ence Foundation (CHE-1300208). Furthermore, we thank Kevin
Dong and Scott Leyendecker for contributing some experiments.
(d) T. Soos, B.L. Bennett, D. Rutherford, L.P. Barthel-Rosa, J.A. Gladysz, Or-
ganometallics 20 (2001) 3079e3086.
[20] (a) M. Zhang, Y.-P. Chen, M. Bosch, T. Gentle, K. Wang, D. Feng, Z.U. Wang, H.-
C. Zhou, Angew. Chem. Int. Ed. 53 (2014) 815e818;
(b) D. Feng, K. Wang, Z. Wei, Y.-P. Chen, C.M. Simon, R.K. Arvapally,
R.L. Martin, M. Bosch, T.-F. Liu, S. Fordham, D. Yuan, M.A. Omary,
M. Haranczyk, B. Smit, H.-C. Zhou, Nat. Commun. (2014) 5723e5732;
(c) L. Wen, P. Cheng, W. Lin, Chem. Sci. 3 (2012) 2288e2292;
(d) L. Ma, A. Jin, Z. Xie, W. Lin, Angew. Chem. Int. Ed. 48 (2009) 9905e9908.
[21] (a) A. Fraccarollo, L. Canti, L. Marchese, M. Cossi, Langmuir 30 (2014)
4147e4156;
Supplementary material
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jorganchem.2017.03.034
(b) M. Cossi, G. Gatti, L. Canti, L. Tei, M. Errahali, L. Marchese, Langmuir 28
(2012) 14405e14414;
References
(c) H. Yu, C. Shen, M. Tian, J. Qu, Z. Wang, Macromolecules 45 (2012)
5140e5150;
[1] J.A. Gladysz, Pure Appl. Chem. 73 (2001) 1319e1324.
[2] Selected references: (a) J.A. Gladysz, D.P. Curran, I.T. Horvath, Handbook of
Fluorous Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, 2004;
€
(d) W. Lu, D. Yuan, D. Zhao, C.I. Schilling, O. Plietzsch, T. Muller, S. Brase,
J. Guenther, J. Blümel, R. Krishna, Z. Li, H.-C. Zhou, Chem. Mater 22 (2010)
5964e5972.
ꢀ
(b) J.J.J. Juliette, D. Rutherford, I.T. Horvath, J.A. Gladysz, J. Am. Chem. Soc. 121
[22] (a) J. You, G. Li, Z. Wang, RSC Adv. 2 (2012) 9488e9494;
(b) S.-O. Kim, Q. Zhao, K. Thangaraju, J.J. Kim, Y.-H. Kim, S.-K. Kwon, Dyes
Pigments 90 (2011) 139e145.
[23] (a) L. Wang, D. Wang, H. Wang, S. Feng, J. Organomet. Chem. 767 (2014)
40e45;
(1999) 2696e2704;
(c) M.-A. Guillevic, C. Rocaboy, A.M. Arif, I.T. Horvath, J.A. Gladysz, Organo-
metallics 17 (1998) 707e717;
(d) J.J.J. Juliette, J.A. Gladysz, I.T. Horvath, Angew. Chem. Int. Ed. 36 (1997)
ꢀ
ꢀ
1610e1612.
(b) I. Eryazici, O.K. Farha, O.C. Compton, C. Stern, J.T. Hupp, S.T. Nguyen, Dalton
Trans. 40 (2011) 9189e9193;
(c) P.K. Iyer, S. Wang, Tetrahedron Lett. 47 (2006) 437e439;
(d) S. Wang, W.J. Oldham Jr., R.A. Hudack Jr., G.C. Bazan, J. Am. Chem. Soc. 122
(2000) 5695e5709;
[3] L.V. Dinh, J.A. Gladysz, Angew. Chem. Int. Ed. 44 (2005) 4095e4097.
[4] Selected books and reviews: (a) P. Barbaro, F. Liguori (Eds.), Heterogenized
Homogeneous Catalysts for Fine Chemicals Production, Springer, Heidelberg,
2010;
(b) J. Blümel, Coord. Chem. Rev. 252 (2008) 2410e2423;
(c) F.R. Hartley, Supported Metal Complexes, D. Reidel, Publ. Co., Dordrecht,
The Netherlands, 1985;
(d) D.E. DeVos, I.F.J. Vankelecom, P.A. Jacobs (Eds.), Chiral Catalyst Immobili-
zation and Recycling, Wiley-VCH, Weinheim, 2000;
(e) G. Rothenberg, Catalysis: Concepts and Green Applications, Wiley-VCH,
Weinheim, 2008.
(e) D.-R. Bai, S. Wang, Organometallics 23 (2004) 5958e5966;
(f) Q. Shen, S. Ye, G. Yu, P. Lu, Y. Liu, Synth. Met. 158 (2008) 1054e1058.
[24] (a) C.Y. Yang, S. Wang, M.R. Robinson, G.C. Bazan, A.J. Heeger, Chem. Mater 13
(2001) 2342e2348;
(b) A. Pegenau, T. Hegmann, C. Tschierske, S. Diele, Chem. - Eur. J. 5 (1999)
1643e1660;
(c) M. Lequan, C. Branger, J. Simon, T. Thami, E. Chauchard, A. Persoons, Adv.
Mat. (Weinheim, Ger.) 6 (1994) 851e853;
[5] R. Silbernagel, A. Díaz, E. Steffensmeier, A. Clearfield, J. Blümel, J. Mol, Catal. A
Chem. 394 (2014) 217e223.
(d) S. Murphy, X. Yang, G.B. Schuster, J. Org. Chem. 60 (1995) 2411e2422;
(e) E.S. Blake, W.C. Hammann, J.W. Edwards, T.E. Reichard, M.R. Ort, J. Chem.
Eng. Data 6 (1961) 87e98.
[6] C. Merckle, J. Blümel, Chem. Mater 13 (2001) 3617e3623.
[7] Merck silica 40, Average Particle Size 0.063e0.2 mm, Specific Surface Area 750
m2/g, Average Pore Diameter 40 Å.
[25] (a) C.A. Fyfe, Solid-state NMR for Chemists, C.F.C. Press, Guelph, Canada, 1983;
(b) T.M. Duncan, A Compilation of Chemical Shift Anisotropies, Farragut Press,
Chicago, IL, 1990;
(c) S. Reinhard, J. Blümel, Magn. Reson. Chem. 41 (2003) 406.
[26] (a) K.J. Cluff, J. Blümel, Organometallics 35 (2016) 3939e3948;
(b) K.J. Cluff, J. Blümel, Chem. Eur. J. 22 (2016) 16562e16575;
(c) K.J. Cluff, J. Blümel, Organometallics 33 (2014) 2671e2680;
(d) K.J. Cluff, M. Schnellbach, C.R. Hilliard, J. Blümel, J. Organomet. Chem. 744
(2013) 119e124.
[8] (a) J. Blümel, J. Am. Chem. Soc. 117 (1995) 2112e2113;
(b) K.D. Behringer, J. Blümel, J. Liq. Chromatogr. 19 (1996) 2753e2765.
[9] (a) E.F. Vansant, P. VanDer Voort, K.C. Vrancken, Characterization and
Chemical Modification of the Silica Surface, Elsevier, Amsterdam, 1995;
(b) R.P.W. Scott, Silica Gel and Bonded Phases, John Wiley and Sons, New York,
1993.
[10] (a) K.D. Behringer, J. Blümel, Inorg. Chem. 35 (1996) 1814e1819;
(b) S. Reinhard, P. Soba, F. Rominger, J. Blümel, Adv. Synth. Catal. 345 (2003)
589e602;
[27] (a) S. Kotha, V.R. Shah, Amino Acids 35 (2008) 83e88;
(b) B. Liu, T.T.T. Dan, G.C. Bazan, Adv. Funct. Mater 17 (2007) 2432e2438;
(c) Y. Takeuchi, M. Nishikawa, H. Hachiya, H. Yamamoto, Magn. Reson. Chem.
43 (2005) 662e664;
(c) F. Piestert, R. Fetouaki, M. Bogza, T. Oeser, J. Blümel, Chem. Commun.
(2005) 1481e1483.
[11] (a) T. Posset, J. Blümel, J. Am. Chem. Soc. 128 (2006) 8394e8395;
(b) T. Posset, J. Guenther, J. Pope, T. Oeser, J. Blümel, Chem. Commun. 47
(2011) 2059e2061;
(c) J.C. Pope, T. Posset, N. Bhuvanesh, J. Blümel, Organometallics 33 (2014)
6750e6753.
[12] (a) J. Guenther, J. Reibenspies, J. Blümel, Adv. Synth. Catal. 353 (2011)
443e460;
(b) C. Merckle, J. Blümel, Top. Catal. 34 (2005) 5e15;
(c) C. Merckle, J. Blümel, Adv. Synth. Catal. 345 (2003) 584e588;
(d) C. Merckle, S. Haubrich, J. Blümel, J. Organomet. Chem. 627 (2001) 44e54.
[13] J.A. Osborn, F.H. Jardine, J.F. Young, G. Wilkinson, J. Chem. Soc. A Inorg. Phys.
Theor. (1966) 1711e1732.
[14] T. Posset, F. Rominger, J. Blümel, Chem. Mater 17 (2005) 586e595.
[15] (a) J. Blümel, Inorg. Chem. 33 (1994) 5050e5056;
(b) J. Sommer, Y. Yang, D. Rambow, J. Blümel, Inorg. Chem. 43 (2004)
7561e7563.
(d) T. Alaviuhkola, J. Bobacka, M. Nissinen, K. Rissanen, A. Ivaska, J. Pursiainen,
Chem. Eur. J. 11 (2005) 2071e2080;
(e) X.-M. Liu, C. He, X.-T. Hao, L.-W. Tan, Y. Li, K.S. Ong, Macromolecules 37
(2004) 5965e5970;
(f) B.N. Biddle, J.S. Gray, Appl. Organomet. Chem. 3 (1989) 537e543;
(g) W.C. Schumb, J. Ackerman Jr., C.M. Saffer Jr., J. Am. Chem. Soc. 60 (1938)
2486e2488.
[28] (a) A. Nagaki, N. Takabayashi, Y. Tomida, J.-I. Yoshida, Beilstein, J. Org. Chem. 5
(2009) 16e27;
(b) A. Nagaki, N. Takabayashi, Y. Tomida, J.-I. Yoshida, Org. Lett. 10 (2008)
3937e3940;
(c) S.J. Dolman, F. Gosselin, P.D. O'Shea, I.W. Davies, Tetrahedron 62 (2006)
5092e5098.
[29] Y. Yang, Doctoral thesis, University of Heidelberg, 2007.
€
[30] C.F. Nising, U.K. Schmid, M. Nieger, S. Brase, J. Org. Chem. 69 (2004)
[16] Y. Yang, B. Beele, J. Blümel, J. Am, Chem. Soc. 130 (2008) 3771e3773.
[17] B. Beele, J. Guenther, M. Perera, M. Stach, T. Oeser, J. Blümel, New J. Chem. 34
(2010) 2729e2731.
6830e6833.
[31] CCDC 1527673 contains the supplementary crystallographic data for this
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034

J.H. Baker et al. / Journal of Organometallic Chemistry xxx (2017) 1e11
11
compound. This data can be obtained free of charge from The Cambridge
[33] R.J. Hinch, J. Krc, Anal. Chem. 29 (1957) 1550e1551.
ꢁ
Crystallographic Data Centre. The unit cell parameters are a ¼ 16.7257(9), b ¼
[34] J. Schraml, M. Capka, V. Blechta, Magn. Reson. Chem. 30 (1992) 544e547.
€
16.7257(9), c ¼ 7.1979(4). The space group is P-4.
[35] F.H. Kohler, N. Hertkorn, J. Blümel, Chem. Ber. 120 (1987) 2081e2082.
[32] J. Stahl, J.C. Bohling, E.B. Bauer, T.B. Peters, W. Mohr, J.M. Martin-Alvarez,
F. Hampel, J.A. Gladysz, Angew. Chem. Int. Ed. 41 (2002) 1871e1876.
[36] C.R. Hilliard, S. Kharel, K.J. Cluff, N. Bhuvanesh, J.A. Gladysz, J. Blümel, Chem.
Eur. J. 20 (2014) 17292e17295.
Please cite this article in press as: J.H. Baker, et al., Journal of Organometallic Chemistry (2017), http://dx.doi.org/10.1016/
j.jorganchem.2017.03.034