New Silica-Immobilized Nickel Catalysts for Cyclotrimerizations of Acetylenes

Article abstract of DOI:10.1002/adsc.200202206

New dicarbonylnickel catalysts with a variety of mono- and bidentate phosphine ligands, such as Ph₂P(CH₂) _xPPh(CH₂)_ySi(OEt)₃ (x = 2, 3; y = 2, 3), have been synthesized and fully characterized, three of them by X-ray analyses. All the catalysts have been immobilized on silica by reaction of the ethoxysilane or hydroxy groups of the phosphine linkers with the surface silanol groups. The clean immobilization was proved by ³¹P solid-state NMR. It has been demonstrated that the C-O-Si bridge, in contrast to the strong Si-O-Si bond, is easily cleaved by organic solvents and water. With respect to the cyclotrimerization of phenylacetylene the optimal reaction conditions for all molecular and immobilized catalysts have been determined. Cyclooctane was the best solvent, and the optimal reaction temperature was 100°C for reactions in homogeneous phase, and 155°C for surface-bound catalysts. The bite angles and spacer lengths of the chelating ligands did not play a major role for the performance of the homogeneous or immobilized catalysts. The latter could be recycled for 12 times, with final substrate conversions still typically between 30 and 40%, and turnover numbers (TON) around 1500. With an excess of substrate in one run a TON of 4810 was obtained for an immobilized carbonylnickel complex with a dppp-type chelate phosphine (x = 3, y = 2). Reduction of the surface coverage generally resulted in catalysts with longer lifetimes, for example, reduction to one fourth gave a record TON of 7616 for this catalyst. The selectivities of the catalysts all followed one trend. The main products were the cyclotrimers 1,2,4- and 1,3,5- triphenylbenzene, besides dimers, linear trimers and tetramers as side-products. The fraction of the latter decreased at higher temperatures with homogeneous catalysts, and vanished entirely after a few cycles with the immobilized catalysts. On recycling, the initial ratio of unsymmetrical to symmetrical cyclotrimers of typically 5:1 gradually changed to about 1:1 for all catalysts.

Full text of DOI:10.1002/adsc.200202206

FULL PAPERS
New Silica-Immobilized Nickel Catalysts for Cyclotrimerizations
of Acetylenes
Ï
S. Reinhard, P. Soba, F. Rominger, J. Blümel*
Department of Organic Chemistry, University of Heidelberg, Im Neuenheimer Feld 270, 69120 Heidelberg, Germany
Phone: / ‡ 49)-6221-54-8470, Fax: / ‡ 49)-6221-54-4904, e-mail: J.Bluemel@urz.uni-heidelberg.de
Received: December 16, 2002; Accepted: February6, 2003
Abstract: New dicarbonylnickel catalysts with a conversions still typically between 30 and 40%, and
varietyof mono- and bidentate phosphine ligands,
turnover numbers /TON) around 1500. With an
such as Ph2P/CH2)xPPh/CH2)ySi/OEt)3 /x ˆ 2, 3; y ˆ excess of substrate in one run a TON of 4810 was
2, 3), have been synthesized and fully characterized, obtained for an immobilized carbonylnickel complex
three of them by X-ray analyses. All the catalysts with a dppp-type chelate phosphine /x ˆ 3, y ˆ 2).
have been immobilized on silica byreaction of the
Reduction of the surface coverage generallyresulted
ethoxysilane or hydroxy groups of the phosphine in catalysts with longer lifetimes, for example, reduc-
linkers with the surface silanol groups. The clean tion to one fourth gave a record TON of 7616 for this
immobilization was proved by 31P solid-state NMR. It catalyst. The selectivities of the catalysts all followed
À À
has been demonstrated that the C O Si bridge, in one trend. The main products were the cyclotrimers
À À
contrast to the strong Si O Si bond, is easilycleaved
byorganic solvents and water. With respect to the
1,2,4- and 1,3,5-triphenylbenzene, besides dimers,
linear trimers and tetramers as side-products. The
cyclotrimerization of phenylacetylene the optimal fraction of the latter decreased at higher temper-
reaction conditions for all molecular and immobilized atures with homogeneous catalysts, and vanished
catalysts have been determined. Cyclooctane was the entirelyafter a few ccyles with the immobilized
best solvent, and the optimal reaction temperature catalysts. On recycling, the initial ratio of unsym-
was 1008C for reactions in homogeneous phase, and metrical to symmetrical cyclotrimers of typically 5:1
1558C for surface-bound catalysts. The bite angles graduallychanged to about 1:1 for all catalysts.
and spacer lengths of the chelating ligands did not
playa major role for the performance of the
Keywords: alkynes; cyclotrimerization; immobiliza-
homogeneous or immobilized catalysts. The latter tion; nickel catalysts; P ligands; silica; solid-state
could be recycled for 12 times, with final substrate NMR
Introduction
support allows the easyseparation of the catalyst by
decanting the supernatant reaction mixture. Thus,
Catalysis represents one of the most central aspects of ideally, the advantages of homogeneous and heteroge-
[1]
chemistry, and the quest for ªideal catalystsº is being neous catalysts should be combined. However, the
pursued in manylaboratories world-wide. Manytypes of
major drawback of immobilized catalysts is their
approaches are being undertaken. For example, homo- possible detachment from the support, the leaching. In
geneous catalysts are made more and more efficient by the course of our studies, we examined a varietyof
tailoring,[2] or two different catalysts are combined.[3] possible leaching mechanisms, and could successfully
Alternative recent strategies for improved removal of eliminate them bysimple means. For example, the
the catalysts from the reaction mixtures make use of support material best suited for the irreversible reaction
biphasic fluorous systems,[4] microencapsulation of cat- with ethoxysilane groups is silica, while oxides like MgO,
alysts,[5] or non-covalentlyfunctionalized dendrimers. [6] TiO2 or acidic Al2O3 lead to extensive and permanent
The keyfeature of all these different approaches is to
improve catalyst recovery and recycling.
linker leaching.[15] With silica, the different forms of
leaching can be prevented efficientlybyapplying the
For some time our group, like others,[7,8] has pursued right conditions concerning solvent[16] and tempera-
this idea byimmobilizing homogeneous catalysts on
ture[17] during the immobilization step, and using
oxidic inorganic supports.[9±14] Since the tethering is rigorouslydried silica.
As soon as there is one
[16,18]
achieved
bybifunctional
linkers
such
as
covalent Si-O-Si bond established, the linker is irrever-
Ph2P/CH2)3Si/OEt)3 /1),[9] the homogeneous character siblybound. [15] Detachment of the metal fragment from
of the catalysts should be retained, while the solid the phosphine moieties[11,12] can be prevented bychelat-
Adv. Synth. Catal. 2003, 345, 589 ± 602
DOI: 10.1002/adsc.200202206
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
589

S. Reinhard et al.
FULL PAPERS
ing phosphines with ethoxysilane groups.[10,12,14] As we gives two isomers, the unsymmetrical 1,2,4- and sym-
have demonstrated recently, catalyst deactivation by metrical 1,3,5-triphenylbenzene, besides minor
amount of linear dimers and trimers. The product
mixture can easilybe analyzed byGC, and in this way
a
dimerization can also be inhibited bydiluting the metal
centers on the surface.[12,14]
The surface-modified supports are amorphous mate- the activityand selectivityof the catalyst can be checked
rials, and thus cannot be characterized with conven- simultaneouslyin one catalytic run. The starting materi-
tional analytical methods other than IR. Nowadays, the al phenylacetylene is volatile and can easily be removed
method of choice to studythe immobilized species is
classical solid-state NMR, which we have optimized
especiallyfor the measurement of immobilized phos-
under vacuum from the reaction mixture if necessary.
Here, we present a systematic study of a variety of
carbonylnickel catalysts in the homogeneous phase, as
phines and molecular catalysts.[19] As a complementary well as immobilized on silica. We will examine the effect
method, suspension NMR spectroscopy[15,20] allows the of the reaction conditions, of the binding to the support,
rapid check of a great number of routine samples, and and factors like bite angle and spacer length of different
also helps to detect merelyadsorbed in contrast to
ligands. It will be shown that veryefficient recyclable
catalysts can be obtained by quite simple measures.
covalentlybound species. [15,18]
The cyclotrimerization of alkynes represents one of
the most useful methods for the selective synthesis of
various substituted aromatic compounds.[21] Since many Results and Discussion
transition metal complexes are active catalysts for
cyclotrimerizations, it is not surprising that the interest
in both mechanistic as well as synthetic aspects of
Synthesis of the Phosphine Ligands
cyclotrimerizations has even been growing recently.
Regarding the mechanism of cyclotrimerizations, it is All the phosphine ligands relevant for this work are
nowadays generally agreed that a metallacyclopenta- displayed in Scheme 1. The ligands 1 to 3 have been
diene is involved. This is corroborated bytheoretical
synthesized in high yields as described previously,[9,17] as
studies,[22,23] as well as byexperimental evidence. Using
well as the phosphines 4 ± 9 in yields of 89 ± 93%.[10] They
unreactive alkynes or bulky phosphine ligands, many have been fullycharacterized bythe usual analytical
trigonal planar bisphosphinemonoalkyne complexes methods, and the NMR signal assignments are based
[24]
could be isolated byPörschke,
for example, upon former results,[9,10,13,17] and 13C{31P}, H,H-, H,C-,
/Cy3P)2Ni/h2-C2H2),[24] and byothers, e.g., /Ph P)2Pd- H,P-COSY, and H,P-HOESY spectra. The still unpub-
3
/h2-MeO2CC ꢀ CCO2Me),[25]
[h2-i-Pr2PCH2CH2P-i- lished data of the compounds 7 ± 9 are given in the
Pr2]Ni/h2-PhC ꢀ CPh),[26] /h2-Cy2PCH2CH2PCy2)Ni/h2- Experimental Section. In solution all the phosphines are
PhC ꢀ CPPh2).[27] There are also recent preparative air-sensitive, and even the most robust triarylphosphine
mechanistic studies discussing metallacyclopentadienes, 2 should be handled under nitrogen or argon in Schlenk
e.g., with Fe as the metal center,[28] Ni,[29±31] Ti,[32,33] Zr,[34] flasks or in a glove-box carefully. However, as soon as
Rh,[35±37] Ru,[38,39] lanthanides,[40] and actinides.[41] Al- the phosphines are immobilized on silica, theycan be
though different types of model compounds have been fullyexposed to the air for hours, until measurable
synthesized,[42±44] the final step is still a matter of debate. amounts of phosphine oxide appear in the 31P NMR
Out of the manyrecent applications of cyclotrimeriza-
spectra.[19] The methods of choice for purifying the
tions for directed syntheses of aromatic sys- phosphines is Kugelrohr distillation or reversed phase
tems,[28±41,45±49] perhaps the most spectacular are the chromatography.
syntheses of helicenes,[50] and of dendrimers.[51]
Our goal has been now to further improve catalytic
cyclotrimerizations by applying immobilized catalysts,
and thus make these reactions more efficient. Out of the
manypossible metals we have chosen Ni/0) catalysts.
Carbonylnickel complexes in particular have the ad-
vantage that theyare cheap, easilyprepared, and good
for analytical purposes, because they are not para-
magnetic, which can, e.g., be demonstrated bytheir
61Ni NMR data.[52] The carbonylnickel complexes are
well-defined, and theyneed no cocatalsyt for the
cyclotrimerizations. Although they are known to oligo-
merize or polymerize a variety of different acetylenes,
depending on the reaction conditions,[53] for our detailed
and systematic study phenylacetylene as the substrate is
suited best. The cyclotrimerization of phenylacetylene linkers 1 ± 9.
Scheme 1. A varietyof monodentate and chelating phosphine
590
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 589 ± 602

Silica-Immobilized Ni Catalysts for Cyclotrimerizations of Acetylenes
FULL PAPERS
Synthesis of the Nickel Catalysts
conditions in organic solvents, and is therefore also
suitable for phosphines containing an ethoxysilane
The most relevant nickel catalysts investigated in this group that might lead to side reactions[17] or cross-
work are displayed in Scheme 2. Only the complexes linking[16] otherwise. For phosphines with OH instead of
10,[9] 12,[9] and 16[10] have been described previously. SiOEt groups, such as 3 and 4, route 2 can be applied.
Data for the new systems are given in the Experimental Here, for example in the case of 3, starting from NiCl2 ´
Section. Due to the alkyl chains and ethoxysilane 6 H2O the dichloronickel complex 21 can be formed
groups, all the compounds are viscous oils, which can first, and then reduced with Zn under addition of CO to
onlybe crystallized occasionally/see below).
the corresponding carbonylnickel complex 11.[55] Using
Nevertheless, preparing the molecular catalysts prior the sulfur compound NaSCH3 as an alternative to Zn[56]
to immobilization has two advantages. First, the com- is possible, but less pleasant and practicable.
plexes can be prepared with the correct stoichiometry
and purified. This leads to cleaner immobilization than
with catalysts prepared on silica. Furthermore their Characteristics of the Nickel Catalysts
analytical data, for example even d/61Ni) and 1J/61Ni-31P)
values can be obtained.[52] In order to avoid the toxic NMR Properties of the Nickel Complexes
Ni/CO)4 for the preparation of the carbonylnickel
compounds,[52] we applytwo routes /Scheme 3). In the
Interestingly, complex21 in C6D6 at ambient temperature
1
first route /COD)2Ni is treated with the phosphine, e.g., gave broad paramagnetic H NMR signals[57] at 27 ppm
7, and 20 forms, as one COD ligand is replaced. The /halfwidth about 1.5 kHz) and 9 ppm, besides broad
second coordinated COD is then replaced byCO to
form 17 bybubbling carbon monoxide into the reaction
mixture.[10,54] This route can be pursued under verymild
signals in the diamagnetic region. More detailed studies
and signal assignments will be a subject of future
investigations. However, this result is in accord, for
example, with the finding of Fyfeꢁs group[58] for the
analogous complex [/EtO)3SiCH2CH2PPh2]2NiCl2, which
was reported not to give a 31P NMR signal at room
temperature. Obviously, these complexes are tetrahedral,
although some other dichloronickel complexes with
diarylalkylphosphine ligands that we and others have
investigatedso far,[19,59] including15 /seebelow)are planar
and diamagnetic, and theygive 31P and 13C CP/MAS
signals with small halfwidths.[19,59] A more detailed over-
view of diamagnetic versus paramagnetic Ni/II) diphos-
phine complexes has been given byBooth. [60]
All the other diamagnetic nickel complexes have been
fullycharacterized bythe usual analytical methods, and
the NMR signal assignments are based upon former
results,[9,10,13,17] comparisons with values from the litera-
ture[61] and 13C{31P}, H,H-, H,C-, H,P-COSY, and H,P-
HOESY spectra. The obtained nickel complexes can also
nicelybe measured as polycrystalline materials by 31P-
and 13C CP/MAS NMR, and we even used them as
standard samples for recording the solid-state NMR
spectra for all sorts of similar immobilized species.[19] The
linewidths of these polycrystalline samples are rather
small, with about 100 to 150 Hz, depending on the
method of measurement. All polycrystalline dicarbonyl-
and dichloronickel complexes with chelating phosphines
show two well-resolved and narrow 31P NMR resonan-
ces[19] in the solid state, indicative of two magnetically
non-equivalent phosphorus nuclei in the elemental cell.
Scheme 2. Molecular nickel catalysts with monodentate and
chelating phosphine ligands.
Crystal Structures of the Nickel Complexes
The square planar arrangement of the ligands at the
metal center of 15 /Figure 1) can be demonstrated byan
Scheme 3. Two possible routes for the synthesis of carbonyl-
nickel phosphine complexes.
Adv. Synth. Catal. 2003, 345, 589 ± 602
asc.wiley-vch.de
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
591

S. Reinhard et al.
FULL PAPERS
Figure 2. Molecular structure of 13. Anisotropic replacement
parameters are shown at 50% probabilitylevel, hydrogen
atoms except OH have been omitted for clarity. Details of
measurement see ref.[64] Selected angles [8] and bond lengths
Figure 1. Molecular structure of 15. Anisotropic replacement
parameters are shown at 50% probabilitylevel, hydrogen
atoms except OH have been omitted for clarity. Details of
measurement see ref.[62] Selected angles [8] and bond lengths
À
À
À
À
À À
[Š]: P Ni P 101.35/2); C Ni C 114.46/10); P Ni C
À
107.20/7)/105.82/7)/114.24/7)/114.46/10); Ni C 1.768/2)/
1.774/2); Ni P 2.1888/6)/2.1953/6).
À
À
À
À
À
À À
[Š]: P Ni P 94.08/2); Cl Ni Cl 91.94/1); P Ni Clcis
À
À
À
86.47/2)/87.43/2); P Ni Cltrans 174.96/2)/178.26/3); Ni Cl
2.2191/6)/2.2089/6); Ni P 2.1652/6)/2.1800/6).
À
ingly, on going from the dichloronickel complex 15 to
the carbonylnickel complex 13 the two phenyl rings of
the lower half of the molecule, which in 15 lie
X-rayanalysis. [62] The arrangement of the molecules in approximatelyin one plane with the P atoms perpen-
À
À
the elemental cell corroborates the magnetic inequiva- dicular to the plane Cl Ni Cl, have to rotate about 908
lence of the two phosphorus nuclei found earlier in the in order to accommodate the steric needs of the CO
form of two 31P CP/MAS signals.[19] Due to the chelate ligand in 13. In contrast to the complexes 13 and 15 with
ring the phosphine ligands have to assume a cis position their bulkychelate ligands, intermolecular hydrogen
in contrast to manycomplexes with monodentate
bridges form in 11 /Figure 3, right display). Within the
phosphines, which prefer a trans arrangement and symmetrical assembly of the four OH groups, the
[60]
À
À
therefore displayno dipolar moment.
The P Ni P distances between the O atoms are 2.75 and 2.80 Š,
À
angle is 94.088, and thus in the usual range for this type of the two angles O H ´´´ O are 173 and 1788. Since every
complex; for example, the P Ni P angle in the com- complex has two phosphine ligands, in the solid state an
parable compound [/MeOCH2)MeC/CH2PPh2)2]NiCl2 extended network forms due to hydrogen bonding. This
À
À
is 95.498.[63]
might have implications on the immobilization of
In contrast to the dichloronickel compounds the phosphine 3 and its complexes on silica /see below).
dicarbonylnickel complexes are all diamagnetic and The crystal structure of 11 furthermore shows that there
thus tetrahedral. This can, for example, be seen in is no interaction of the oxygen atoms with the metal
Figure 2, which shows the crystal structure of 13.[64] The center, as it is claimed e. g. for Ru complexes with ether-
bite angle of the chelating phosphine P Ni P is widened containing phosphine ligands.[68]
to 101.348 now to accommodate the tetrahedral arrange-
À
À
ment of the substituents. However, a comparison with
the crystal structure of 11[65] /Figure 3) with a P Ni P Immobilization of the Linkers and Catalysts
angle of 108.78 for analogous alkyldiaryl, but mono-
À
À
dentate phosphine ligands, which comes close to the Immobilized species will be denoted by i in the follow-
theoretical tetrahedral angle, shows that the chelate ing. All the linkers and catalysts can be tethered to silica
phosphine still imposes its steric requirements to the /Merck silica, specific surface area 750 m2/g, average
[1±3,9±13]
complex. This might have an impact on the catalytic pore diameter 40 Š) as described previously.
activityof the nickel complexes /see below). Interest-
Special care has to be taken regarding the tempera-
592
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 589 ± 602

Silica-Immobilized Ni Catalysts for Cyclotrimerizations of Acetylenes
FULL PAPERS
Figure 3. Molecular structure /left display) and an arrangement of four molecules /right display) of 11. Anisotropic
replacement parameters are shown at 50% probabilitylevel, hydrogen atoms except O H have been omitted for clarity. Details
of measurement see ref.[65] P Ni P 108.67/2); C Ni C 113.50/12); P Ni C 104.88/8)/107.20/8)/107.99/8)/114.30/8); Ni C
1.772/3)/1.775/3); Ni P 2.2193/6)/2.2206/6).
À
À
À
À
À
À
À
À
ture[17] and the solvent[16] used for the immobilization. It ment studies, as we have previouslydone with ethoxy-
is important that the temperature of the immobilization silane species.[15] The surface-modified silica was stirred
[18]
be as low as possible,[17] the silica be as dry, and the with the solvents or mixtures as given in Table 1 for 24 h
solvent be as unpolar[16] as possible. In the case of at ambient temperature. Then, the amount of ligand
immobilization via Si/OEt)3 groups, as soon as one found in the supernatant solution was determined. In
covalent bond is formed, the linker cannot be detached case a TLC analysis did not indicate the presence of any
[11,15]
from the surface anymore.
The success of the
procedure can be checked with 31P CP/MAS NMR,[19] as
displayed in Figure 4 for the surface-bound ligand 6i and
complex 14i.
Due to the inhomogeneityof the surface, the line-
widths are increased byimmobilization, e.g., to 775 Hz
for 14i. In the case of 6i, the signals for both phosphorus
nuclei, which have different chemical shifts in solu-
tion,[10] now overlap. However, the spectra show that
ligands and complexes can be immobilized without any
side-products.[17] Importantly, in the case of the nickel
compound, no uncomplexed phosphine can be detected
on the surface.[12]
Recentlythe formation of EtOH laeyrs on silica
surfaces in cyclohexane has been investigated.[69] The
reaction of alcohols with silica has alreadybeen
described byIler [70] and one of us.[18] However, regarding
the robustness of covalent bonding to the support in the
case of ligands with OH groups, such as 3 and 4, no
systematic studies have been done yet to the best of our
Figure 4. 162.0 MHz 31P CP/MAS spectra of 6i and 14i on
silica. Rotational frequency13 kHz, other details of the
knowledge. Therefore, we undertook surface detach- measurement see ref.[19]
Adv. Synth. Catal. 2003, 345, 589 ± 602
asc.wiley-vch.de
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
593

S. Reinhard et al.
FULL PAPERS
Table 1. Detachment experiments for 3i and 4i on 1 g of SiO2
with 20 ml of solvent at ambient temperature.
hydrogen bonds. Since for the recycling of catalysts a
strong irreversible covalent bonding to the support is
indispensable, and the catalysts 11i and 13i indeed
showed too much leaching in repeated runs, in the
following we will focus on the catalysts immobilized via
siloxane bridges.
Solvent
Ligand Immobilized Detached ligand
on SiO2 substance
[mg]
[mg]
[%]
toluene[b]
toluene[b]
THF
3
4
3
4
3
4
3
4
3
4
3
63
72
78
82
103
75
70
45
86
78
±
±
±
±
< 6
< 10[b]
50
85
100
94
100
85
53
< 5
6[b]
52
64
70
42
86
63
136
Cyclotrimerization of Phenylacetylene
THF[b]
ethanol
ethanol
Improvement of Catalysis Conditions in Homogeneous
and Heterogeneous Phase
ethanol[a]
ethanol[a]
ethyl acetate[a]
ethyl acetate[a]
acetone[a]
All the experiments described in this manuscript were
done in a batch-wise manner, with a catalyst to substrate
ratio of 1:200, and a phenylacetylene /22) concentration
of 0.5 mol/L. The amount of catalyst applied here is
rather small with 0.5 mol %, as compared to the 2 ±
4 mol % used previously.[e.g.,46] As a first step we checked
the influence of the solvent on the performance of the
254
[a]
Oxygen-free solvents with residual water content.
608C.
[b]
ligand in solution, 12 more h of stirring at 608C followed. catalyst, and found that unpolar aliphatic hydrocarbons
As the examples in Table 1 show, with toluene 4 cannot function best.[12] But even within this group there are
be removed from the silica even at higher temperatures. marked differences. While, for example, catalyst 10
However, more polar solvents lead to nearlyquantita-
tive loss of 4 from the surface alreadyat room temper-
ature. Especiallythe solvents with residual water con-
gives quantitative conversion of the substrate after 3 h at
1008C in cyclooctane, under otherwise the same reac-
tion conditions in decane even after 3 days the con-
tent lead to major ligand leaching. This has to be taken version is lower than 10%. This effect cannot be due to
into account in case leaching of the catalyst and/or linker different polarities, because the dielectric constants of
is found later during recycling experiments. As com- cyclooctane /2.12) and decane /1.99) are very similar.[71]
pared to 4i, 3i is even more vulnerable to leaching. The Interestingly, there is even a noticeable difference
reason might be the tendencyto form intermolecular
between octane and cyclooctane /Figure 5). While the
hydrogen bridges, as they were found for 11 in the crystal conversion/time curve has the same character, in octane
structure /see above). This tendencyof 11 to arrange onlyabout 90% conversion can be obtained with 10.
itself in extended networks leaves less opportunities for This effect of the cyclic hydrocarbons, that we find for all
À À
bonding to the silica surface via C O Si bridges. Thus, the molecular and immobilized catalysts in this paper, is
larger aggregates of 3 might assemble in the pores, which known since a long time among chemists working in the
are later just flushed out bythe solvent. The higher
field of catalysis, but it has to the best of our knowledge
initial loading of silica with 3 as compared to 4 /Table 1) never been explained. For the rest of the work presented
could also be explained with this sort of cross-linking via here, cyclooctane was used as the solvent.
Figure 5. Representative dependence of the catalytic activity
on the solvent for catalyst 10 with phenylacetylene /22).
Figure 6. Temperature dependence of the catalytic activity of
10 in cyclooctane.
594
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 589 ± 602

Silica-Immobilized Ni Catalysts for Cyclotrimerizations of Acetylenes
FULL PAPERS
it as a compromise in order to compare the activities of
all the homogeneous catalysts 10 ± 19. The result is
displayed in Figure 7.
Under these conditions, the most active catalyst is 12,
while 13 displays the worst performance, which might be
due to its reduced solubilityin cyclooctane. Otherwise,
with the exception of 16, all the complexes display
similar activity. Complex 16 performs equallywell at
higher temperatures, though. Overall, one has to con-
clude that the bite angles or the spacer lengths of the
chelating ligands do not have a decisive influence on the
catalytic activity of 14 and 16 ± 19. Possiblythe ligands
are not rigid enough to show anyobvious dependence of
the catalytic performance on the bite angle.[72]
Figure 7. Comparison of the activities of 10, 12 ± 14, and
16 ± 19 in the homogeneous phase at 1008C.
The previous results also showed that the cyclotrime-
rization is not sensitive towards minor changes of the
linkers. Therefore, one could anticipate that the immo-
bilized versions of the catalysts, 10i, 12i, 14i, 16i, 17i, and
19i are equallysuited for cyclotrimerizing 22. Indeed, as
displayed in Figure 8, again all catalysts have similar
activityand achieve 100% conversion within one day.
The catalytic reactionis slower than in the homogeneous
phase, because now the substrate has to diffuse into the
pores of the support. This has been proved recentlywith
Rh catalysts by using supports with different pore
sizes.[14] However, this drawback is made up for bythe
easyseparation of the catalyst from the reaction mixture
bysimplydecanting the supernatant solution after the
Figure 8. Comparison of the catalytic activities of 10i, 12i, 14i,
16i, 18i, and 19i at 1558C. For 17i see ref.[12]
Another positive effect of cyclooctane versus octane
/bp 1268C) is the higher boiling point of 1528C, which
allows catalytic runs with a bath temperature of up to
1558C. With respect to the optimal temperature, all the
homogeneous catalysts 10 ± 19 have the same prefer-
ences as their immobilized analogues 10i ± 19i. For
catalysts with monodentate alkyldiarylphosphine and
triarylphosphine ligands, 1008C is a good choice. This is
displayed for example in Figure 6 for 10.
Triarylphosphine ligands lead to more active catalysts,
for example, quantitative conversion of 22 is reached
with 12 after about half an hour at 1008C already, while
about 3 hours are needed by 10 /Figure 6). All catalysts
with chelating ligands need 1558C for optimal activity.
For example, at 1558C with 19 quantitative conversion is
achieved after less than half an hour, while 6 hours are
needed at 1008C. As another example, for 17 the
relation is 3 h at 1558C versus 9 h at 1008C.
Comparison of Catalysts with Different Types of Ligands
Figure 9. Recycling of 18i in the cyclotrimerization of 22.
Top: Catalytic activities; bottom: conversion after maximally
Although 1008C is not the optimal temperature for
complexes with chelating ligands /see above), we chose 200 h.
Adv. Synth. Catal. 2003, 345, 589 ± 602
asc.wiley-vch.de
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
595

S. Reinhard et al.
FULL PAPERS
Table 2. Recycling of immobilized catalysts: maximal conversion of 22 [%], /ratio 23:24), and [ratio of dimer, linear trimers,
and linear tetramers] for everyrun with the amount of 24 set to 1 as the standard; TON gives the number of alkyne molecules
converted byeach metal center after 12 runs.
Cycle Catalyst
16i
17i
18i
19i
10i
12i
14i
1
2
100 /4.80)
93.1 /2.56)
100 /5.18)
100 /4.90)
100 /5.62)
100 /4.83)
100 /3.66)
[0.6, 0.5, 0.85] [0.06, ±, 0.05] [0.46, 0.66, 1.02] [0.4, 0.56, 0.95] [0.87, 0.51, 0.71] [0.77, 0.53, 1.01] [0.42, 0.2, 0.52]
95.9 /2.28)
94.6 /1.80)
98.8 /3.26)
100 /3.84)
95.9 /2.78)
95.1 /3.47)
91.3 /1.96)
[0.14, ±, 0.12] [±, ±, ±]
[0.26, ±, 0.36]
[0.34, 0.24, 0.48] [0.61, ±, 0.15]
[0.16, ±, 0.29]
[0.04, ±, 0.06]
3
89.1 /1.61)
[±, ±, 0.03]
77.9 /0.78)
[0.06, ±, ±]
93.1 /1.44)
[±, ±, 0.05]
90.5 /1.76)
[0.05, ±, 0.04]
87.1 /1.98)
[0.03, ±, ±]
83.4 /1.80)
[±, ±, ±]
71.4 /1.07)
[0.09, ±, ±]
4
84.3 /1.09)
[0.08,±,±]
81.7 /0.85)
[0.08,±,±]
88.6 /1.11)
[±, ±, ±,]
67.8 /1.04)
[±, ±, ±]
71.8 /1.53)
[0.10, ±, ±]
69.5 /1.34)
[±, ±, 0.07]
61.3 /0.86)
[0.11, ±, ±]
5
70.0 /0.69)
[0.12, ±, ±]
61.1 /0.94)
[±, ±, ±]
80.2 /0.99)
[±, ±, ±,]
61.8 /1.01)
[0.06, ±, ±]
63.9 /1.32)
[0.11, ±, ±]
57.4 /0.95)
[0.12, ±, 0.05]
65.5 /0.93)
[0.17, ±, ±]
6
60.6 /0.80)
[0.18, ±, ±]
38.3 /0.88)
[0.19, ±, ±]
69.0 /0.90)
[±, ±, ±,]
54.6 /0.92)
[0.11, ±, ±]
62.2 /1.0)
[0.10, ±, ±]
52.3 /0.87)
[0.13, ±, ±]
77.2 /1.11)
[0.10, ±, ±]
7
55.4 /0.88)
[0.16, ±, ±]
42.6 /0.82)
[0.19, ±, ±]
56.2 /0.75)
[0.10, ±, ±,]
47.8 /1.0)
[0.08, ±, ±]
40.3 /0.81)
[0.22, ±, ±]
48.8 /0.85)
[0.13, ±, ±]
58.4 /0.98)
[0.10, ±, ±]
8
41.7 /0.92)
[0.25, ±, ±]
68.8 /0.89)
[0.15, ±, ±]
52.4 /0.94)
[0.13, ±, ±]
36.0 /0.84)
[0.22, ±, ±]
63.9 /0.78)
[0.10, ±, ±]
37.4 /0.83)
[0.18, ±, ±]
59.1 /0.93)
[0.06, ±, ±]
9
34.0 /0.79)
[0.26, ±, ±]
43.4 /0.89)
[0.24, ±, ±]
48.6 /0.83)
[0.17, ±, ±]
32.6 /0.85)
[0.23, ±, ±]
58.1 /0.88)
[0.14, ±, ±]
34.0 /0.85)
[0.25, ±, ±]
68.6 /1.05)
[0.07, ±, ±]
10
11
12
31.9 /0.85)
[0.25, ±, ±]
61.7 /0.94)
[0.23, ±, ±]
45.3 /0.84)
[0.20, ±, ±]
33.2 /0.81)
[0.21, ±, ±]
52.9 /0.84)
[0.11, ±, 0.06]
27.0 /0.86)
[0.29, ±, ±]
49.8 /0.94)
[0.08, ±, ±]
36.1 /0.86)
[0.27, ±, 0.07] [0.24, ±, ±]
39.5 /0.91)
35.0 /0.80)
[0.28, ±, ±]
31.2 /0.86)
[0.24, ±, ±]
44.3 /0.91)
[0.17, ±, ±]
26.4 /0.81)
[0.33, ±, ±]
53.7 /1.04)
[0.10, ±, ±]
36.2 /0.84)
[0.28, ±, ±]
38.1 /0.83)
[0.17, ±, ±]
1481
29.1 /0.83)
[0.24, ±, ±]
1592
26.1 /0.84)
[0.23, ±, ±]
1363
35.9 /0.83)
[0.22, ±, 0.07]
1552
26.8 /0.86)
[0.30, ±, ±]
1317
36.1 /0.82)
[0.20, ±, ±]
1584
TON 1470
catalytic run. Furthermore, the supernatant solution the differences in the product yields are not decisive
does not have to be filtered prior to its application to the enough to make far-reaching assumptions about the best
GC column, as it has to be done for runs with ligand, those chelate ligands with C3 bridges seem to be
homogeneous catalysts. Thus, the work-up and analytics slightlysuperior to the ones with dppe-type bridges. No
are much easier for the immobilized catalysts.
correlation of the linker length or type with the
efficiencyof the catalysts can be found. Regarding the
monodentate phosphines, although the triarylphos-
phine ligand produces the most active catalyst in the
first run, in the course of 12 cycles it deteriorates more
Recycling of the Immobilized Catalysts
The main advantage of the immobilized catalysts, quickly/overall TON only1317) than the complexes
however, is their potential for easyrecycling. Here, the with alkyldiarylphosphine or aryldialkylphosphine li-
immobilized catalysts 10i, 12i, 14i, and 16i ± 19i show gands.
their value: theycan be subjected to 12 cycles with a final
As alreadyindicated in preliminarycommunications,
conversion of still between 30 and 40%. This is shown, the maximal yields of products can be increased by a
for example, for 18i in Figure 9. The maximal conversion continuous design of the reactions,[12] because the
of 22 after everyrun, and the ratio of 1,2,4-triphenyl-
benzene /23) to 1,3,5-triphenylbenzene /24) are given in cooling, opening and recharging the Schlenk flask. This
Table 2 for the immobilized catalysts.
is also proved bythe 31P CP/MAS spectra of the spent
catalytic systems are sensitive towards disturbances by
The maximal turnover numbers /TON) for all the catalysts after 12 runs: besides the signal of the original
catalysts after 12 runs /Table 2) are around 1500. The catalyst, peaks of oxidic species are always and in some
highest TON is achieved for 18i with 1592 in 12 cycles, a cases those of uncomplexed surface-bound phosphines
value much higher than obtained previously. Although are visible. The assumption that the system is disturbed
596
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 589 ± 602

Silica-Immobilized Ni Catalysts for Cyclotrimerizations of Acetylenes
FULL PAPERS
Table 3. Product ratio of 1,2,4-triphenylbenzene 23 to the
dimer, linear trimers, and linear tetramers with respect to 24
/set to 1 as standard) at 1008C, [1558C], and /708C).
byrecharging the reaction vessel is further corroborated
bythe fact that one single addition of an excess of 22 to
one batch of 18i gives the high TON of 4810.
Another option that immobilized catalysts offer as
compared to homogeneous ones is that the metal
centers can be diluted on the support surface. In this
Product
Ratio Catalyst
23
Linear
Trimers
Linear
Tetramers
10
12
13
14
16
17
18
10.1 /14.1)
8.1 /10.4)
5.3 [5.4]
7.0 [5.6]
7.6 [2.9]
7.5 [2.4]
7.1 [5.0]
6.9 [4.6]
2.0 /2.7)
1.7 /1.8)
1.3 [1.0]
1.7 [1.2]
2.3 [0.7]
2.0 [0.6]
1.5 [1.2]
1.7 [1.0]
1.4 /1.7)
1.0 /1.0)
1.0 [1.1]
1.4 [1.2]
1.3 [0.7]
1.7 [0.7]
1.4 [1.3]
1.3 [1.1]
[73]
way, deactivation by dimerization has, for example,
successfullybeen curbed back in the case of Wilkinson-
type Rh hydrogenation catalysts.[14] Although in the case
of carbonylnickel catalysts the main decomposition
mechanism is the loss of the metal fragment from the
phosphine linker, the surface dilution also seems to play
a role.[12] For example, if 18i is diluted to one fourth 19
/about 4 metal centers per 100 nm2) of the maximal
surface density/16), a record TON of 7616 in one cycle
with an excess of 22 can be obtained. With the usual
recycling procedure /Table 2) 18i with one fourth of the
maximal surface coverage shows still a conversion of
88% in the 7th run. In contrast to this, one half of the
maximal coverage with 18i makes no difference as
compared to 18i with maximal densityon the surface,
the conversion in the 7th run is about 56%. This result
shows that the metal centers need a minimum distance
to each other, so that theyare kept from forming dimers
or oligomers. So, yet another deactivation mechanism
for carbonylnickel phosphine catalysts seems to be an
aggregation of the activated nickel centers. In the
homogeneous phase this is obvious bythe formation of
a shinylayer of metallic nickel on the glass surface of the
flask. The effect of surface dilution on the catalytic
activityof the catalysts also corroborates the assumption
that the molecules are evenlydistributed over the surface
afterimmobilization, and nopatchesareformedbycross-
linking of the ethoxysilane groups or ªself-assemblyº of
the linkers. It is also possible to improve the efficiencyof
Figure 10. Product ratio of 1,2,4- to 1,3,5-triphenylbenzene
23:24 for the catalysts 10i, 12i, 14i, and 16i in the course of 12
cycles.
immobilized catalysts with monodentate ligands. For trimers and tetramers is greater at lower temperatures.
example, in the case of 10i half /5 molecules per 100 nm2) For 17, their ratio with respect to 24 /set to 1) is 0.6:0.7 at
of the maximal surface coverage alreadygives a decisive
improvement: In the seventh run the conversion of 22 is temperature effect in order to drive the catalysis in the
still about 89%, while 10i with maximal densityis already preferred direction. Another interesting feature is that
1558C, and 2.0:1.7 at 1008C. Thus, one can use the
down to about 40%. So, the effect of dilution of the the catalysts with monodentate ligands are more
catalyst on the surface is of a general nature, and selective with respect to the formation of 23 than those
independent of the ligands or metal centers.
with chelating ligands. For example, for 10 at 1008C a
ratio of 23 to 24 of 10.1:1 is obtained, and at 708C even
of 14.1:1. The highest ratio obtained with the chelate
complex 16 at 1008C is only7.6:1. While at 100 8C there
is no difference in selectivityamong the chelate
Selectivity of the Catalysts
In homogeneous catalysis the main products with complexes, at 1558C those with the larger dppp-type
phenylacetylene /22) as the substrate are the unsym- bridge are more selective towards the formation of 23
metric and symmetric cyclotrimer 23 and 24, linear than those with the dppe-type phosphine ligand. So,
trimers, and linear tetramers. Onlytraces of dimers are
overall the homogeneous catalysts become more selec-
found. As shown in Table 3, the catalysts are overall tive, the better theycan assume tetrahedral ligand
more selective towards the formation of the unsymmet- arrangement.
ric cyclotrimer 23, when the temperature is lower. For
As can be seen in Table 2, in the case of the
example, in the case of 17 the ratio of 23:24 increases immobilized catalysts the amount of linear products is
from 2.4:1 to 7.5:1 when the temperature is lowered reduced substantiallyin all runs. Generall,y after the
from 1558C to 1008C. However, the percentage of linear fifth run, there are no more trimers or tetramers to be
Adv. Synth. Catal. 2003, 345, 589 ± 602
asc.wiley-vch.de
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
597

S. Reinhard et al.
FULL PAPERS
found, and onlytraces of the dimer. Interestingly, in the
course of the recycling experiments, and also during one
single run, the product ratio of 23 to 24 gradually
becomes smaller, and finallyreaches a plateau value of
about 1:1 for everycatalsyt. This can be seen in
Figure 10 for a varietyof immobilized catalysts with
mono- and bidentate phosphine linkers.
The same result is found in the case of 10i, operated in
a pseudo-continuous manner.[12] Therefore, we deduce
that the selectivityof immobilized catalsyts is not
governed bythe bite angle or spacer length of the linker
phosphine. The change also does not correlate with the
loss of catalytic activity, or the decomposition of the
catalyst, as seen by solid-state NMR. The catalysts must
change in a common, general way. The elucidation of
this process will be the subject of future investigations
with still more elaborate analytical tools.
Synthesis of Phosphines 6 to 9
Compounds
6 to 9 were synthesized by irradiating
Ph2P/CH2)2PPhH[74] or Ph2P/CH2)3PPhH[74] in the presence
of 1.2 equivalents of commercially available vinyl- or allyl-
triethoxysilane with UV light in a standard quartz apparatus
with stirring and water cooling. The surplus of starting silane
was then removed byKugelrohr distillation at 120 8C under
reduced pressure /0.02 mbar). Typical yields of the viscous oils
were 89% /6), 92% /7), 93% /8), and 91% /9). Compound 6 has
[10]
been described previously.
Phosphine 7: 1H NMR /300.1 MHz, C6D6): dˆ7.45 ±7.38 /m,
Ho, PPh, 2H), 7.38±7.29 /m, Ho, PPh2, 4H), 7.10±7.05 /m, Hm, p
,
ˆ
PPh, 3H), 7.05 ±6.95 /m, Hm, p, PPh2, 6H), 3.73 /q, OCH2, 3JH,H
7.0 Hz, 6H), 2.25 ±2.04 /m, Ph2PCH2, 2H), 1.90±1.76 /m,
CH2CH2Si, 2H), 1.73 ±1.60 /m, PhPCH2, 4H), 1.11 /t, CH3,
3JH,H ˆ7.0 Hz, 9H), 0.85 ±0.76 /m, CH2Si, 2H); 13C NMR
1
/75.5 MHz, C6D6): dˆ139.4, 139.2 /d, Ci, PPh2, JP, C ˆ15.2 Hz),
139.1 /d, Ci, PPh, 1JP, C ˆ17.3 Hz), 133.2, 133.0 /d, Co, PPh2, 2JP, C
ˆ
ˆ
18.0 Hz), 132.9 /d, Co, PPh, 2JP, C ˆ19.4 Hz), 128.7 /d, Cm, JP, C
3
5.5 Hz), 128.9, 128.6 /s, Cp, PPh2), 128.6 /d, Cp, PPh, 4JP, C ˆ1.4 Hz),
1
58.4 /s, OCH2), 32.0 /d, CH2CH2CH2Si, JP, C ˆ13.1 Hz), 24.7
1
/pseudo-t, Ph2PCH2, JP,C ˆ2JP, C ˆ15.3 Hz), 24.6 /pseudo-t,
Conclusion
1
2
PhPCH2, JP, C ˆ2JP, C ˆ14.5 Hz), 20.1 /d, CH2CH2Si, JP, C
ˆ
15.9 Hz), 18.6 /s, CH3), 12.9 /d, CH2Si, 3JP, C ˆ11.1 Hz); 31P NMR
/121.5 MHz, THF): dˆÀ12.7 /d, PPh2, 3JP, P ˆ26.8 Hz), À21.6 /d,
In this work we have synthesized and characterized a
variety of new carbonylnickel catalysts, and we have
immobilized them cleanlyon silica bymono- or
bidentate phosphine linkers. The success of the immo-
bilization has been shown by 31P solid-state NMR, as
well as the fate of the catalysts during catalysis. The
temperature and solvent playmajor roles for the
cyclotrimerizations. The immobilized catalysts can be
recycled many times, and record turnover numbers are
achieved. The selectivities can be tuned with the choice
of temperature and phosphine bite angle in the case of
homogeneous catalysts, and with reaction time for
immobilized species. The activities of the immobilized
catalysts can be further enhanced by continuous catal-
ysis and diluting the metal centers on the surface.
3
PPh, JP, P ˆ26.8 Hz); anal. calcd. for C29H40O3P2Si /526.67): C
66.14, H 7.66, P 11.76; found: C 66.33, H 7.40, P 11.25.
Phosphine 8: 1H NMR /300.1 MHz, C6D6): dˆ7.52±7.44 /m,
Ho, PPh, 2H), 7.43±7.33 /m, Ho, PPh2, 4H), 7.13±7.00 /m, Hm, p
,
9H), 3.73/q, OCH2, 3JH,H ˆ7.0 Hz, 6H),2.34±2.28/m, Ph2PCH2,
2H), 2.05±1.92 /m, CH2CH2Si, 2H), 1.90±1.83 /m, PhPCH2,
3
2H), 1.73±1.68 /m, Ph2PCH2CH2, 2H), 1.12 /t, CH3, JH,H
ˆ
7.0 Hz, 9H), 0.90±0.69 /m, CH2Si, 2H); 13C NMR /75.5 MHz,
C6D6): dˆ139.8 /d, Ci, PPh2, 1JP, C ˆ15.2 Hz), 139.7 /d, Ci, PPh2,
1JP, C ˆ14.5 Hz), 139.4 /d, Ci, PPh, 1JP, C ˆ18.0 Hz), 133.2, 133.1,
2
3
132.9 /d, Co, JP, C ˆ18.7 Hz), 128.6 /d, Cm, PPh2, JP, C ˆ6.9 Hz)
128.5 /d, Cm, PPh, 3JP, C ˆ6.9 Hz), 128.8, 128.6, 128.6 /s, Cp), 58.5
1
3
/s, OCH2), 30.1 /dd, Ph2PCH2, JP, C ˆ13.1 Hz, JP, C ˆ11.1 Hz),
1
3
29.8 /dd, PhPCH2, JP, C ˆ14.5 Hz, JP, C ˆ12.5 Hz), 22.9 /dd,
Ph2PCH2CH2, 2JP, C ˆ15.2 Hz, 2JP, C ˆ17.3 Hz), 21.4 /d,
CH2CH2Si, 1JP, C ˆ14.5 Hz), 18.5 /s, CH3), 6.7 /d, CH2Si, 2JP, C
ˆ
10.4 Hz); 31P NMR /121.5 MHz, C6D6): dˆÀ17.2 /s, PPh2), À
18.8 /s, PPh); anal. calcd. for C29H40O3P2Si /526.67): C 66.14, H
7.66, P 11.76, found: C 66.20, H 7.41, P 11.21.
Experimental Section
1
Phosphine 9: H NMR /300.1 MHz, C6D6): d ˆ 7.45 ± 7.39
/m, Ho, PPh, 2H), 7.38 ± 7.28 /m, Ho, PPh2, 4H), 7.08 ± 6.92 /m,
General Remarks
3
Hm, p, 9H), 3.71 /q, OCH2, JH,H ˆ 7.0 Hz, 6H), 2.06 ± 1.92 /m,
Ph2PCH2, 2H), 1.75 ± 1.50 /m, PhPCH2, CH2CH2CH2Si,
All reactions were performed bystandard Schlenk techniques
or in a glove-box in an oxygen-free nitrogen atmosphere. The
solvents were dried byboiling over Na /cyclooctane over
CaH2) and distilled and kept under nitrogen. The NMR spectra
were recorded on Bruker instruments at the indicated mag-
netic fields. The solid-state NMR measurements were done
with an Avance 400 wide-bore machine equipped with a 4 mm
MAS probehead. For details of the solid-state NMR param-
eters see ref.[19]
3
CH2CH2CH2Si, Ph2PCH2CH2, 8H), 1.08 /t, CH3, JH,H
ˆ
7.0 Hz, 9H), 0.86 ± 0.71 /m, CH2Si, 2H); 13C NMR /75.5 MHz,
C6D6): d ˆ 140.1, 139.9 /d, PPh2, Ci, 1JP, C ˆ 13.1 Hz), 140.0 /d, Ci,
PPh, 1JP, C ˆ 15.2 Hz), 133.4 /d, Co, PPh, 2JP, C ˆ 18.7 Hz), 133.4.
2
3
133.3, /d, Co, PPh2, JP, C ˆ 18.8 Hz), 129.1 /d, Cm, PPh, JP,C
ˆ
5.7 Hz), 129.1, 129.0 /d, Cm, PPh2, 3JP, C ˆ 6.6 Hz), 129.2, 128.8,
1
128.8 /s, Cp), 58.9 /s, OCH2), 32.7 /d, CH2CH2CH2Si, JP,C
ˆ
13.2 Hz), 30.8 /pseudo-t, Ph2PCH2, 1JP, C ˆ 3JP, C ˆ 13.2 Hz), 30.6
1
3
/dd, PhPCH2, JP, C ˆ 13.2 Hz, JP, C ˆ 11.3 Hz), 23.4 /pseudo-t,
Ph2PCH2CH2, JP, C ˆ 2JP, C ˆ 16.5 Hz), 20.7 /d, CH2CH2Si,
2
2JP, C ˆ 16.0 Hz), 19.1 /s, CH3), 13.4 /d, CH2Si, JP, C ˆ 11.3 Hz);
3
Syntheses of Phosphines 1 to 5
The compounds 1 ± 3,[9,17] and 4, 5[10] have been synthesized and
31P NMR /121.5 MHz, C6D6): d ˆÀ 17.5 /s, PPh2), À 27.7 /s,
PPh); anal. calcd. for C30H42O3P2Si /540.69): C 66.64, H 7.83, P
11.46; found: C 67.04, H 7.86, P 11.33.
characterized as described previously.
598
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 589 ± 602

Silica-Immobilized Ni Catalysts for Cyclotrimerizations of Acetylenes
FULL PAPERS
2
1
d ˆ 201.7 /t, CO, JP, C ˆ 1.2 Hz), 139.2 /pseudo-t, Ci, JP,C
ˆ
Synthesis of the Nickel Complexes 10 ± 19
3JP, C ˆ 22.3 Hz), 137.6 /dd, Ciꢀ, JP, C ˆ 15.0 Hz, JP, C ˆ 13.3 Hz),
1
3
The syntheses and characterization of the compounds 10,[9]
12,[9] and 16[10] have been described in the indicated references.
Compound 11: To 1.46 g /2 mmol) [HO/CH2)4PPh2]2NiCl2
/synthesis analogous to the one described below for 15) in
10 mLTHF was added 0.19 g /3 mmol, 1.5 equiv.) of Zn powder
in a stream of CO. The mixture turned pale yellow during 1 h of
stirring at ambient temperature. After removal of the excess
Zn byfiltration and removal of the solvent under vacuum, the
residue was recrystallized at 808C from EtOH/H2O /1:1).
After further recrystallization from EtOH/hexane a slightly
yellow crystalline product was obtained; yield: 0.29 g /24%,
0.46 mmol). Data from the crystal structure see Figure 3 and
ref.[65] IR /KBr): n ˆ 3371 /OH) 1991 /CO), 1935 cmÀ1 /CO);
132.6 /d, Co, 2JP, C ˆ 7.3 Hz), 132.0 /d, Coꢀ, 2JP, C ˆ 7.9 Hz), 129.4 /s,
3
Cp), 129.3 /s, Cpꢀ), 128.7 /d, Cm, JP, C ˆ 4.8 Hz), 128.5 /d, Cmꢀ,
3JP, C ˆ 4.8 Hz), 68.1 /s, CHOSi), 59.4 /s, OCH2), 39.8 /pseudo-t,
CH2, 1JP,C ˆ 3JP, C ˆ 9.9 Hz), 18.3 /s, CH3); 31P NMR /121.5 MHz,
THF): d ˆ 18.4 /s); HRMS /FAB): calcd. for C33H40NiO4P2Si
[M ± 2 CO]: 648.1536, 650.1516; found: 648.1525 /100),
650.1479 /51.1); anal. calcd. for C35H40NiO6P2Si /705.42): C
59.59, H 5.72, P 8.78; found: C 59.55, H 6.06, P 8.44.
Compound 15: To a solution of 0.86 g /3.6 mmol) of NiCl2 ´
6 H2O in EtOH a solution of 1.56 g /3.6 mmol) of 4 in THF was
added. After 1 h of stirring, the dark red precipitate was
filtered off and washed with EtOH wherebydark red crystals
were obtained which were insoluble in the usual organic
solvents; yield: 1.14 g /2.0 mmol, 56%). Therefore it was
characterized byan X-rayanalysis /Figure 1 and ref. [62]) and
bysolid-state NMR; 13C CP/MAS NMR /TPPM decoupled,[19]
100.6 MHz, nrot ˆ 15 kHz): d ˆ 146.4 ± 121.4 /m, Caryl), 65.2 /s,
1H NMR /300.1 MHz, acetone-d6): d ˆ 7.42 ± 7.33 /m. Haryl
,
3
2OH), 3.35 /t, CH2OH, JH,H ˆ 6.7 Hz, 4H), 1.83 /dt, PCH2,
2JP, H ˆ 9.8 Hz, 3JH,H ˆ 6.7 Hz, 4H), 1.33 /q, CH2CH2OH, 3JH,H
ˆ
6.7 Hz, 4H), 1.21 /m, PCH2CH2, 4H); 13C NMR /75.5 MHz,
THF-d8): d ˆ 201.2 /t, CO, 2JP, C ˆ 2.2 Hz), 139.7 /dd, Ci, 1JP,C
ˆ
CH), 39.7, 34.8 /s, CH2); 31P CP/MAS NMR /162.0 MHz, nrot
ˆ
3
2
18.2 Hz. JP, C ˆ 14.7 Hz), 132.9 /t, Co, JP,C ˆ 4JP, C ˆ 6.6 Hz),
13 kHz): d ˆ 13.1, 6.9; MS /FAB): m/z ˆ 521.1 /M ± Cl); anal.
calcd. for C27H26Cl2NiOP2 /558.05): C 58.11, H 4.70, P 11.10;
found: C 57.78, H 4.96, P 10.97.
129.3 /s, Cp), 128.6 /t, Cm, JP, C ˆ 5JP, C ˆ 4.4 Hz), 61.9 /s,
3
CH2OH), 35.3 /t, PCH2, JP, C ˆ 3JP, C ˆ 7.2 Hz), 30.7 /dd,
1
PCH2CH2,
2JP, C ˆ 12.2 Hz,
4JP, C ˆ 10.0 Hz),
21.5
/t,
Compound 17: A solution of 239 mg /0.45 mmol) of 7 in
5 mL of toluene was added in a dropwise manner to a solution
of 125 mg /0.45 mmol) of /COD)2Ni in 30 mL of toluene,
precooled to À 308C. The suspension turned dark red during
30 minutes of stirring at À 308C. During 2 h of bubbling in CO
at À 308C, the color became lighter. Then the mixture was
warmed to ambienttemperature under prolonged admission of
CO. After removal of the solvent and flash chromatography
with reversed-phase silica, 17 was obtained as an orange oil;
yield: 266 mg /0.41 mmol, 91.2%); IR /KBr): n ˆ 1995 /CO),
1932 cmÀ1 /CO); 1H NMR /300.1 MHz, C6D6): d ˆ 7.72 /d, Ho,
PPh, 3JH,H ˆ 8.4 Hz, 2H), 7.67 /d, Ho, PPh2, 3JH,H ˆ 9.0 Hz, 2H),
7.52 /d, Ho, PPh2, 3JH,H ˆ 8.7 Hz, 2H), 7.13 ± 6.89 /m, Hm, p, 9H),
CH2CH2OH, 3JP, C ˆ 5JP, C ˆ 3.9 Hz); 31P NMR /121.5 MHz, ace-
tone-d6): d ˆ 25.2; anal. calcd. for C34H38NiO4P2 /631.304 g/
mol): C 64.69, H 6.07, P 9.81; found: C 64.19, H 6.12, P 9.66.
Compound 13: To a suspension of 2.69 g /4.82 mmol) of 15 in
40 mL EtOH 0.47 g of Zn powder were added in a stream of
CO. After 1.5 h of stirring at ambient temperature the bubbling
in of CO was stopped. After filtration and removal of the
solvent under vacuum a light brown oil remained. Recrystal-
lization with EtOH/H2O gave 13 as slightly yellow crystals;
yield: 1.56 g /2.85 mmol, 59.5%). Data from the crystal
structure, see Figure 2 and ref.[64] IR /KBr): n ˆ 2000 /CO),
1937 cmÀ1 /CO); 1H NMR /300.1 MHz, acetone-d6): d ˆ 7.23 ±
7.77 /m, Haryl, 20H), 4.43 /d, OH, 3JH,H ˆ 5 Hz), 3.61 /m, CH),
3
3.76 /q, OCH2, JH,H ˆ 6.9 Hz, 6H), 2.03 ± 1.53 /m, Ph2PCH2,
2
3
3.13 /dd, CH2, JH,H ˆ 13.2 Hz, JH,H ˆ 11.7 Hz, 2H), 2.15 /dd,
CH2, 2JH,H ˆ 13.2 Hz, 3JH,H ˆ 12.8 Hz, 2H); 13C NMR
CH2CH2Si, PhPCH2, CH2CH2CH2Si, 8H), 1.14 /t, CH3, 3JH,H
ˆ
6.9 Hz, 9H), 0.81 ± 0.66 /m, CH2Si, 2H); 13C NMR /75.5 MHz,
2
/125.8 MHz, acetone-d6): d ˆ 199.0 /t, CO, JP, C ˆ 3.4 Hz),
2
2
C6D6): d ˆ 202.9 /d, CO, JP,C ˆ 3.5 Hz), 202.7 /d, CO, JP,C
ˆ
1
3
139.0 /dd, Ci, JP,C ˆ 23.3 Hz, JP, C ˆ 20.8 Hz), 136.6 /dd, Ciꢁ,
1
3
4.8 Hz), 138.6 /dd, Ci, PPh2, JP, C ˆ 26.0 Hz, JP, C ˆ 5.2 Hz),
138.4 /dd, Ci, PPh2, 1JP, C ˆ 25.6 Hz, 3JP, C ˆ 6.2 Hz), 137.7 /dd, Ci,
1JP, C ˆ 15.7 Hz, JP, C ˆ 12.3 Hz), 132.1 /pseudo-t, Co, JP, C
ˆ
3
2
4JP, C ˆ 7.2 Hz), 131.3 /pseudo-t, Coꢀ, JP, C ˆ 4JP, C ˆ 8.1 Hz),
2
1
3
PPh, JP, C ˆ 22.1 Hz, JP, C ˆ 5.5 Hz), 133.2, 133.0, 132.8 /d, Co,
129.4 /s, Cp), 129.2 /s, Cpꢀ), 128.4 /pseudo-t, Cm, JP, C ˆ 5JP,C
ˆ
3
2JP, C ˆ 14.5 Hz), 130.3 /d, Cp, PPh2, 4JP, C ˆ 1.4 Hz), 130.1 /s, Cp,
4.7 Hz), 128.2 /pseudo-t, Cmꢀ, JP, C ˆ 5JP, C ˆ 4.7 Hz), 65.0 /t,
3
4
PPh), 129.8 /d, Cp, PPh2, JP, C ˆ 2.1 Hz), 129.4, 129.3 /d, Cm,
CHOH, 2JP, C ˆ 16.5 Hz), 38.6 /dd, CHꢁ2, 1JP, C ˆ 11.4 Hz, 3JP,C
ˆ
PPh2, 3JP, C ˆ 9.0 Hz), 129.3 /d, Cm, PPh, 3JP, C ˆ 9.7 Hz), 59.0 /s,
8.3 Hz); 31P NMR /121.5 MHz, THF): d ˆ 18.4 /s); HRMS
/FAB): calcd. for C27H26NiO3P2 [M ± 2CO]: 486.0821, 488.0777;
found: 486.0812 /100), 488.0778 /42.9).
1
3
OCH2), 35.7 /dd, CH2CH2CH2Si, JP, C ˆ 15.6 Hz, JP,C
ˆ
1
2
4.5 Hz), 30.3 /dd, Ph2PCH2, JP, C ˆ 24.9 Hz, JP, C ˆ 21.5 Hz),
1
2
28.9 /dd, PhPCH2, JP, C ˆ 23.9 Hz, JP, C ˆ 22.5 Hz), 20.0 /d,
Compound 14: A solution of 259 mg /0.94 mmol) of
/COD)2Ni in 30 mL of toluene was cooled to À 308C. Then,
a solution of 554 mg /0.94 mmol) of 5[10] in 5 mL of toluene was
added dropwise. The yellow reaction mixture turned dark red
immediately. The mixture was stirred for 0.5 h at À 308C. Then
CO was bubbled through the sample for 2 h at À 308C. The
mixture was warmed to room temperature and became yellow.
Removal of the solvent and flash chromatographyon reversed
phase silica gave 14 as a slightly yellow oil; yield: 530 mg
/0.75 mmol, 80%); IR /KBr): n ˆ 1998 /CO), 1933 cmÀ1 /CO);
1H NMR /300.1 MHz, C6D6): d ˆ 7.83 ± 7.59 /m, Ho, 8H), 7.12 ±
CH2CH2Si, 2JP, C ˆ 6.2 Hz), 19.0 /s, CH3), 13.4 /d, CH2Si, 3JP,C
13.1 Hz); 31P NMR /121.5 MHz, C6D6): d ˆ 46.6 /d, PPh, 3JP, P
ˆ
ˆ
35.9 Hz), 42.2 /d, PPh2, 3JP, P ˆ 35.9 Hz); HRMS /FAB): calcd.
for C29H40NiO3P2Si [M ± 2 CO]: 584.1599, 586.1553; found:
584.1576 /100), 586.1530 /95.2); anal. calcd. for C31H40NiO5P2Si
/641.37): C 58.05, H 6.29, P 9.66; found: C 58.66, H 6.44, P 9.82.
Compound 18: The nickel complex 18 was synthesized
according to the procedure given for 17. IR /KBr): n ˆ 1992
/CO), 1930 cmÀ1 /CO); 1H NMR /300.1 MHz, C6D6): d ˆ
7.66 ± 7.62 /m, Ho, PPh2, 4H), 7.54 ± 7.51 /m, Ho, PPh, 2H),
3
7.40 ± 7.20 /m, Hm, p, 9H), 3.71 /q, OCH2, JH,H ˆ 6.7 Hz, 6H),
6.84 /m, Hm, p, 12H), 4.01 ± 3.84 /m, CH), 3.70 /q, OCH2, 3JH,H
ˆ
2.34 ± 2.28 /m, Ph2PCH2, 2H), 2.05 ± 1.92 /m, CH2CH2Si, 2H),
1.90 ± 1.83 /m, PhPCH2, 2H), 1.73 ± 1.68 /m, Ph2PCH2CH2, 2H),
7.0 Hz, 6H), 3.57 ± 3.39 /m, CH2, 2H), 2.26 ± 2.06 /m, CH2, 2H),
1.06 /t, CH3, 3JH,H ˆ 7.0 Hz, 9H); 13C NMR /125.8 MHz, C6D6):
Adv. Synth. Catal. 2003, 345, 589 ± 602
asc.wiley-vch.de
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
599

S. Reinhard et al.
FULL PAPERS
3
1.10 /t, CH3, JH,H ˆ 6.7 Hz, 9H), 0.85 ± 0.76 /m, CH2Si, 2H);
Homogeneous Catalysis
13C NMR /75.5 MHz, CD2Cl2): d ˆ 201.6 /s, CO), 138.7 /dd, Ci,
For all experiments with homogeneous catalysts, this standard
procedure was applied, unless otherwise noted: X mg /usually
10 to 25 mg) of catalyst /0.5 mol % with respect to the
substrate) in Y mL /usually5 to 12 mL) of freshlydistilled
cyclooctane /0.5 mol/L substrate concentration) was heated to
the reaction temperature in a 25-mL Schlenk flask. Then the
catalysis was started by adding Z mL /usually 0.3 to 0.5 mL) of
freshlydistilled 22 via a syringe. Samples of 100 mL were taken
at regular intervals from the reaction mixture, filtered through
a small column of silica, and applied to the GC. The peak
assignments in the chromatograms were made with standard
samples of the pure compounds 23 and 24, as well as with GC-
MS for the linear oligomers. Everycatalytic run was done
twice, and the average value was used for the Figures and
Tables presented here.
PPh2, 1JP, C ˆ 27.3 Hz, 3JP, C ˆ 5.2 Hz), 138.6 /dd, Ci, PPh2, 1JP,C
ˆ
3
1
27.0 Hz, JP,C ˆ 6.2 Hz), 136.5 /dd, Ci, PPh, JP,C ˆ 24.2 Hz,
3JP, C ˆ 5.5 Hz), 132.5, 132.4, 132.0 /d, Co, 2JP, C ˆ 13.8 Hz), 129.8,
4
129.6, 129.3 /d, Cp, JP,C ˆ 1.4 Hz), 128.7, 128.5, 128.4 /d, Cm,
3JP, C ˆ 9.0 Hz), 58.6 /s, OCH2), 29.7 /dd, Ph2PCH2, JP,C
ˆ
1
3
1
21.1 Hz, JP, C ˆ 2.4 Hz), 29.2 /dd, PhPCH2, JP,C ˆ 19.0 Hz,
3JP, C ˆ 2.4 Hz), 25.7 /dd, Ph2PCH2CH2, JP, C ˆ 15.9 Hz, JP,C
ˆ
2
2
6.9 Hz), 19.8 /pseudo-t, CH2CH2Si, JP,C ˆ 3JP, C ˆ 3.8 Hz), 18.4
/s, CH3), 4.4 /s, CH2Si); 31P NMR /121.5 MHz, C6D6): d ˆ 17.1
/s, PPh), 15.1 /s, PPh2); HRMS /FAB): calcd. for C29H40NiO3P2-
575 /100), 586.1530 /49.1); anal. calcd. for C31H40NiO5P2Si
/641.37): C 58.05, H 6.29; found: C 57.72, H 6.73.
1
Compound 19: The nickel complex 19 was synthesized
according to the procedure given for 17. IR /KBr): n ˆ 1992
/CO), 1929 cmÀ1 /CO); 1H NMR /500.1 MHz, C6D6): d ˆ 7.75
3
3
/d, Ho, PPh, JH,H ˆ 7.0 Hz, 2H), 7.67 /d, Ho, PPh2, JH,H
ˆ
3
6.7 Hz, 2H), 7.56 /d, Ho, PPh2, JH,H ˆ 7.0 Hz, 2H), 7.14 ± 6.89
3
Catalysis with Immobilized Catalysts
/m, Hm, p, 9H), 3.79 /q, OCH2, JH,H ˆ 7.1 Hz, 6H), 2.04 ± 1.94
/m, Ph2PCH2, 2H), 1.92 ± 1.84 /m, PhPCH2, 2H), 1.79 ± 1.39 /m,
6H, CH2CH2CH2Si, CH2CH2Si, Ph2PCH2CH2), 1.17 /t, CH3,
3JH,H ˆ 7.1 Hz, 9H), 0.80 ± 0.68 /m, CH2Si, 2H); 13C NMR
/125.8 MHz, C6D6): d ˆ 201.4 /d, CO, 2JP, C ˆ 1.9 Hz), 201.2 /d,
X g of modified silica /depending on the surface coverage 80 to
160 mg, corresponding to about 7 to 20 mg of surface-bound
catalyst) was placed into a 25-mL Schlenk flask. The silica was
then suspended in Y mL /5 to 15 mL) of cyclooctane and
heated to the reaction temperature with stirring. Then, the
catalysis was started by addition of Z mL /0.25 to 0.6 mL) of
freshlydistilled 22. Samples of 100 mL of the supernatant
solution were taken at regular intervals from the reaction
mixture, filtered through a small column of silica, and applied
to the GC. For peak assignments see above. The maximal
reaction time for one cycle was 8 days. After catalysis the
reaction mixture was cooled, and the supernatant solution
decanted. The solid catalyst was washed by stirring with 5 mL
of heptane, and dried under vacuum for about 30 min after
decanting all liquids. Then, the catalyst was either subjected to
solid-state NMR analysis, or used for the next cycle. Every
catalytic run was done twice, and the medium value was used
for the Figures and Tables presented here.
2
1
3
CO, JP, C ˆ 3.8 Hz), 138.7 /dd, Ci, PPh2, JP,C ˆ 26.9 Hz, JP,C
ˆ
6.1 Hz)#, 138.4 /dd, Ci, PPh, JP, C ˆ 26.8 Hz, JP, C ˆ 8.0 Hz)#,
1
3
2
132.3, 132.1 /d, Co, PPh2, JP,C ˆ 15.1 H), 131.6 /d, Co, PPh,
2JP, C ˆ 14.1 Hz), 129.1 /s, Cp, PPh2), 128.6 /s, Cp, PPh), 128.2 /d,
Cm, PPh2, 3JP, C ˆ 9.4 Hz), 128.1 /d, Cm, PPh, 3JP, C ˆ 9.4 Hz), 58.1
1
3
/s, OCH2), 35.8 /dd, CH2CH2CH2Si, JP, C ˆ 17.0 Hz, JP, C
ˆ
1
3
6.6 Hz), 29.3 /dd, Ph2PCH2, JP, C ˆ 20.3 Hz, JP, C ˆ 8.9 Hz),
1
3
29.2 /dd, PhPCH2, JP, C ˆ 20.1 Hz, JP,C ˆ 8.5 Hz), 22.7 /br. s,
Ph2PCH2CH2), 19.4 /pseudo-t, CH2CH2Si, JP, C ˆ 4JP,C
ˆ
2
3
4.2 Hz), 18.2 /s, CH3), 12.6 /d, CH2Si, JP, C ˆ 13.2 Hz);
31P NMR /121.5 MHz, C6D6): d ˆ 15.6 /s, PPh), 10.0 /s, PPh2);
HRMS /FAB): calcd. for C30H42NiO3P2Si [M ± 2 CO]:
598.1705, 600.1625; found: 598.1732 /100), 600.1687 /49.7).
#The signals of the ipso-C nuclei have onlylow intensity.
Acknowledgements
General Immobilization Procedure
We thank the Deutsche Forschungsgemeinschaft /SFB 623), and
the Fonds der Chemischen Industrie /FCI) for financial support.
We are also grateful for the help of Mrs. D. Rambow for
preparing countless samples and supervising the catalytic runs.
All the molecular phosphines and nickel complexes were
immobilized on Merck silica /specific surface area 750 m2/g,
average pore diameter 40 Š, particle size 70 to 230 mesh,
predried at 6008C under vacuum) bystirring an excess of the
molecular compounds with a suspension of the support
material in toluene at 608C for about 5 hours.[16±18] After
decanting the supernatant solution, the silica was washed twice
with toluene and twice with THF, and the surface coverages
were determined byweighing back the molecular compounds
after removing the solvents from the combined supernatant
and washing solutions.[16±18] The surface coverages of all
immobilized species were usuallyin the range of 12 to 19
molecules per 100 nm2 of surface area. For suspension NMR[15]
the THF slurries were used directly, for solid-state NMR the
material was dried under vacuum at ambient temperature for
at least 4 h. The solid-state NMR chemical shifts of the
immobilized species were comparable to the ones of the
molecules in solution /see Figure 4).
References and Notes
[1] J. A. Gladysz, Pure Appl. Chem. 2001, 73, 1319.
[2] B. Cornils, W. A. Herrmann, Applied Homogeneous
Catalysis with Organometallic Compounds, VCH, New
York, 1996.
[3] H. Gao, R. J. Angelici, Organometallics 1999, 18, 989.
[4] M. Wende, R. Meier, J. A. Gladysz, J. Am. Chem. Soc.
2001, 123, 11490, and literature cited therein.
[5] R. Akiyama, S. Kobayashi, Angew. Chem. 2001, 113,
3577.
600
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 589 ± 602

Silica-Immobilized Ni Catalysts for Cyclotrimerizations of Acetylenes
FULL PAPERS
[6] D. de Groot, B. F. M. de Waal, J. N. H. Reek, A. P. H. J.
Schenning, P. C. J. Kamer, E. W. Meijer, P. W. N. M.
van Leeuwen, J. Am. Chem. Soc. 2001, 123, 8453.
[7] F. R. Hartley, Supported Metal Complexes, D. Reidel
Publishing Co, Dordrecht, Holland, 1985.
[8] J. H. Clark, Supported Reagents in Organic Reactions,
VCH, Weinheim, 1994.
[9] K. D. Behringer, J. Blümel, Inorg. Chem. 1996, 35, 1814.
[10] G. Tsiavaliaris, S. Haubrich, C. Merckle, J. Blümel,
Synlett 2001, 391.
[11] K. D. Behringer, J. Blümel, Chem. Commun. 1996, 653.
[12] S. Reinhard, K. D. Behringer, J. Blümel, New J. Chem., in
press.
[13] C. Merckle, S. Haubrich, J. Blümel, J. Organomet. Chem.
2001, 627, 44.
[33] D. Suzuki, H. Urabe, F. Sato, J. Am. Chem. Soc. 2001,
123, 7925; R. Tanaka, Y. Nakano, D. Suzuki, H. Urabe, F.
Sato, J. Am. Chem. Soc. 2002, 124, 9682; D. Suzuki, R.
Tanaka, H. Urabe, F. Sato, J. Am. Chem. Soc. 2002, 124,
3518.
[34] T. Takahashi, F.-Y. Tsai, Y. Li, K. Nakajima, M. Kotora,
J. Am. Chem. Soc. 1999, 121, 11093; T. Takahashi, Z. Xi,
A. Yamazaki, Y. Liu, K. Nakajima, M. Kotora, J. Am.
Chem. Soc. 1998, 120, 1672.
[35] Y. Kishimoto, P. Eckerle, T. Miyatake, M. Kainosho, A.
Ono, T. Ikariya, R. Noyori, J. Am. Chem. Soc. 1999, 121,
12035.
[36] H. Yan, A. M. Beatty, T. P. Fehlner, Organometallics
2002, 21, 5029; H. Yan, A. M. Beatty, T. P. Fehlner,
Angew. Chem. 2001, 113, 4630.
[14] C. Merckle, J. Blümel, Adv. Synth. Catal. 2003, 345, 584.
[15] C. Merckle, J. Blümel, Chem. Mater. 2001, 13, 3617.
[16] K. D. Behringer, J. Blümel, J. Liquid Chromatogr. 1996,
19, 2753.
[37] I. Ojima, A. T. Vu, J. V. McCullagh, A. Kinoshita, J. Am.
Chem. Soc. 1999, 121, 3230.
Â
Â
[38] B. Witulski, T. Stengel, J. M. Fernandez-Hernandez,
Chem. Commun. 2000, 1965.
[17] J. Blümel, Inorg. Chem. 1994, 33, 5050.
[18] J. Blümel, J. Am. Chem. Soc. 1995, 117, 2112.
[19] S. Reinhard, J. Blümel, Magn. Reson. Chem. 2003, 41, in
press.
[39] Y. Yamamoto, R. Ogawa, K. Itoh, Chem. Commun. 2000,
549.
[40] R. Ruan, J. Zhang, X. Zhou, R. Cai, Chem. Commun.
2002, 538.
[20] K. D. Behringer, J. Blümel, Z. Naturforsch. 1995, 50b,
1723.
[41] A. Haskel, T. Straub, A. K. Dash, M. S. Eisen, J. Am.
Chem. Soc. 1999, 121, 3014.
[21] Reviews on cyclotrimerization: S. Saito, Y. Yamamoto,
Chem. Rev. 2000, 100, 2901; H.-W. Frühauf, Chem. Rev.
1997, 97, 523; M. Lautens, W. Klute, W. Tam, Chem. Rev.
1996, 96, 49; G. G. Melikyan, K. M. Nicholas, in Modern
Acetylene Chemistry, /Eds.: P. J. Stang, F. Diederich),
VCH, Weinheim, 1995, pp. 99 ± 138; D. B. Grotjahn, in
Comprehensive Organometallic Chemistry II, /Eds.:
E. W. Abel, F. G. A. Stone, G. Wilkinson), /Vol. Ed.:
L. S. Hegedus), Pergamon, Oxford, 1995, Vol. 12,
pp. 741 ± 770.
[42] R. Diercks, B. E. Eaton, S. Gürtzgen, S. Jalisatgi, A. J.
Matzger, R. H. Radde, K. P. C. Vollhardt, J. Am. Chem.
Soc. 1998, 120, 8247.
[43] L. Guo, J. D. Bradshaw, D. B. McConville, C. A. Tessier,
W. J. Youngs, Organometallics 1997, 16, 1685.
[44] T. Nickel, R. Goddard, C. Krüger, K.-R. Pörschke,
Angew. Chem. 1994, 106, 908.
[45] S. Saito, T. Kawasaki, N. Tsuboya, Y. Yamamoto, J. Org.
Chem. 2001, 66, 796.
[46] N. Mori, S. Ikeda, K. Odashima, Chem. Commun. 2001,
181.
[47] J. Li, H. Jiang, M. Chen, J. Org. Chem. 2001, 66, 3627.
[48] T. Sugihara, A. Wakabayashi, Y. Nagai, H. Takao, H.
Imagawa, M. Nishizawa, Chem. Commun. 2002, 576.
[49] N. Mori, S. Ikeda, Y. Sato, J. Am. Chem. Soc. 1999, 121,
2722.
[50] S. Han, D. R. Anderson, A. D. Bond, H. V. Chu, R. L.
Disch, D. Holmes, J. M. Schulman, S. J. Teat, K. P. C.
Vollhardt, G. D. Whitener, Angew. Chem. 2002, 114,
3361; S. Han, A. D. Bond, R. L. Disch, D. Holmes, J. M.
Schulman, S. J. Teat, K. P. C. Vollhardt, G. D. Whitener,
Angew. Chem. 2002, 114, 3357.
[22] S. A. Macgregor, E. Wenger, Organometallics 2002, 21,
1278.
[23] R. Herges, A. Papafilippopoulos, Angew. Chem. 2001,
113, 4809.
[24] K.-R. Pörschke, J. Am. Chem. Soc. 1989, 111, 5691.
Â
Â
Ä
[25] D. Pena, D. Perez, E. Guitian, L. Castedo, J. Org. Chem.
2000, 65, 6944 and literature cited therein.
[26] C. Müller, R. J. Lachicotte, W. D. Jones, Organometallics
2002, 21, 1118 and 1975.
[27] M. A. Bennett, C. J. Cobley, A. D. Rae, E. Wenger, A. C.
Willis, Organometallics 2000, 19, 1522.
Â
[28] K. Ferre, L. Toupet, V. Guerchais, Organometallics 2002,
Â
21, 2578.
[51] S. Hecht, J. M. J. Frechet, J. Am. Chem. Soc. 1999, 121,
[29] H. Hoberg, W. Richter, J. Organomet. Chem. 1980, 195,
355.
4084.
[52] K. D. Behringer, J. Blümel, Magn. Reson. Chem. 1995,
33, 729.
[53] L. S. Meriwether, M. F. Leto, E. C. Colthup, G. W.
Kennerly, J. Org. Chem. 1962, 27, 3930.
[54] K.-R. Pörschke, C. Pluta, B. Proft, F. Lutz, C. Krüger, Z.
Naturforsch. 1993, 48b, 608.
[30] M. Shanmugasundaram, M.-S. Wu, C.-H. Cheng, Org.
Lett. 2001, 3, 4233; A. Jeevanandam, R. P. Korivi, I.
Huang, C.-H. Cheng, Org. Lett. 2002, 4, 807.
[31] J. Montgomery, Acc. Chem. Res. 2000, 33, 467, and
literature cited therein.
[32] O. V. Ozerov, F. T. Lapido, B. O. Patrick, J. Am. Chem.
Soc. 1999, 121, 7941; O. V. Ozerov, B. O. Patrick, F. T.
Lapido, J. Am. Chem. Soc. 2000, 122, 6423.
[55] P. Giannoccaro, A. Sacco, G. Vasapollo, Inorg. Chim.
Acta 1979, 37, L 455.
[56] K. Tanaka, Y. Kawata, T. Tanaka, Chem. Lett. 1974, 831.
Adv. Synth. Catal. 2003, 345, 589 ± 602
asc.wiley-vch.de
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
601

S. Reinhard et al.
FULL PAPERS
[57] J. Blümel, M. Herker, W. Hiller, F. H. Köhler, Organo-
metallics 1996, 15, 3474; M. Schnellbach, F. H. Köhler, J.
Blümel, J. Organomet. Chem. 1996, 520, 227.
[58] L. Bemi, H. C. Clark, J. A. Davies, C. A. Fyfe, R. E.
Wasylishen, J. Am. Chem. Soc. 1982, 104, 438.
[59] P. S. Jarrett, P. J. Sadler, Inorg. Chem. 1991, 30, 2098.
[60] G. Booth, Adv. Inorg. Chem. Radiochem. 1964, 6, 1.
[61] J. Mason, Multinuclear NMR, Plenum Press, New York,
1987.
residual values R1/F) ˆ 0.028, wR/F2) ˆ 0.064 for ob-
served reflections, residual electron density±0.19 to 0.27
eŠ±3. Further crystallographic data for this structure
have been deposited with the Cambridge Crystallo-
graphic Data Center as supplementarypublication no.
CCDC 199726. These data can be obtained free of
charge via www.ccdc.cam.ac.uk/conts/retrieving.html /or
from the CCDC, 12 Union Road, Cambridge CB2 1EZ,
UK; fax: ‡ 44 1223 336033; e-mail: deposit@ccdc.cam.
ac.uk).
[62] Red crystal /polyhedron), dimensions 0.28 Â 0.23 Â
0.19 mm3, crystal system monoclinic, space group P21/c,
Z ˆ 4, a ˆ 8.6869/4) Š, b ˆ 24.5197/11) Š, c ˆ 12.6866/6)
[65] Colorless crystal /polyhedron), dimensions 0.34 Â 0.16 Â
0.06 mm3, crystal system monoclinic, space group P21/c,
Z ˆ 4, a ˆ 19.4224/3) Š, b ˆ 10.0773/1) Š, c ˆ 17.6950/2)
Š, a ˆ 908, b ˆ 109.101/1)8, g ˆ 908, Vˆ 2553.5/2) Š3,
Š, a ˆ 908, b ˆ 112.4390/10)8, g ˆ 908, Vˆ 3201.14/7) Š3,
3
ˆ
ˆ
r 1.452 g/cm , Tˆ 200/2) K, 2Vmax 27.488, radiation
Mo-Ka, l ˆ 0.71073 Š, 0.38 w-scans with CCD area
detector, covering a whole sphere in reciprocal space,
26130 reflections measured, 5855 unique /R/int) ˆ
0.0331), 4821 observed /I > 2s/I)), intensities were
corrected for Lorentz and polarization effects, an
empirical absorption correction was applied using SA-
DABS[66] based on the Laue symmetry of the reciprocal
3
ˆ
ˆ
r 1.310 g/cm , Tˆ 200/2) K, 2Vmax
25.578, radiation
Mo-Ka, l ˆ 0.71073 Š, 0.38 w-scans with CCD area
detector, covering a whole sphere in reciprocal space,
23209 reflections measured, 5573 unique /R/int) ˆ
0.0375), 4247 observed /I > 2s/I)), intensities were
corrected for Lorentz and polarization effects, an
empirical absorption correction was applied using SA-
DABS[66] based on the Laue symmetry of the reciprocal
space, m ˆ 1.11 mm±1, Tmin 0.76, Tmax 0.83, structure
ˆ
ˆ
space, m ˆ 0.74 mm±1, Tmin 0.79, Tmax 0.96, structure
solved byPatterson-method and refined against F 2 with a
full-matrix least-squares algorithm using the SHELXTL-
PLUS /5.10) software package,[67] 378 parameters re-
fined, hydrogen atoms were treated using appropriate
riding models except for H33 and H63 at the hydroxy
groups, which were refined isotropically, goodness of fit
1.04 for observed reflections, final residual values
R1/F) ˆ 0.031, wR/F2) ˆ 0.064 for observed reflections,
residual electron density±0.35 to 0.28 eŠ ±3. Further
crystallographic data for this structure have been depos-
ited with the Cambridge Crystallographic Data Center as
supplementarypublication no. CCDC 199728. These
data can be obtained free of charge via www.ccdc.ca-
m.ac.uk/conts/retrieving.html /or from the CCDC, 12
Union Road, Cambridge CB2 1EZ, UK; fax: ‡ 44 1223
336033; e-mail: deposit@ccdc.cam.ac.uk).
ˆ
ˆ
2
solved bydirect methods and refined against F with a
full-matrix least-squares algorithm using the SHELXTL-
PLUS /5.10) software package,[67] 302 parameters re-
fined, hydrogen atoms were treated using appropriate
riding models except for H1, which was refined isotropi-
cally, goodness of fit 1.06 for observed reflections, final
residual values R1/F) ˆ 0.033, wR/F2) ˆ 0.077 for ob-
served reflections, residual electron density±0.35 to 0.50
eŠ±3. Further crystallographic data for this structure
have been deposited with the Cambridge Crystallo-
graphic Data Center as supplementarypublication no.
CCDC 199727. These data can be obtained free of
charge via www.ccdc.cam.ac.uk/conts/retrieving.html /or
from the CCDC, 12 Union Road, Cambridge CB2 1EZ,
UK; fax: ‡ 44 1223 336033; e-mail: deposit@ccdc.cam.
ac.uk).
[63] S.-T. Liu, G.-J. Liu, C.-H. Yieh, M.-C. Cheng, S.-M. Peng,
J. Organomet. Chem. 1990, 387, 83.
[66] Program SADABS V2.03 for absorption correction:
G. M. Sheldrick, Bruker Analytical X-ray Division,
Madison, Wisconsin, 2001.
[67] Software package SHELXTL V5.10 for structure solu-
tion and refinement: G. M. Sheldrick, Bruker Analytical
X-rayDivision, Madison, Wisconsin, 1997.
[68] C. Nachtigal, S. Al-Gharabli, K. Eichele, E. Lindner,
H. A. Mayer, Organometallics 2002, 21, 105.
[69] M. Mizukami, M. Moteki, K. Kurihara, J. Am. Chem.
Soc. 2002, 124, 12889.
[70] R. K. Iler, The Chemistry of Silica, John Wiley, New
York, 1979.
[71] D. R. Lide, in C. R. C. Handbook of Chemistry and
Physics, 76th ed., C. R. C. Press, New York, 1995.
[72] P. Dierkes, P. W. N. M. van Leeuwen, J. Chem. Soc.
Dalton Trans. 1999, 1519.
[73] U. Schubert, K. Rose, Transition Met. Chem. 1989, 14,
291.
[74] T.-S. Chou, C.-H. Tsao, S. C. Hung, J. Organomet. Chem.
1986, 312, 53.
[64] Colorless crystal /polyhedron), dimensions 0.36 Â 0.21 Â
0.15 mm3, crystal system monoclinic, space group P21/c,
Z ˆ 4, a ˆ 8.3643/1) Š, b ˆ 31.4479/3) Š, c ˆ 10.4164/1)
Š, a ˆ 908, b ˆ 104.4210/10)8, g ˆ 908, Vˆ 2653.60/5) Š3,
3
ˆ
ˆ
r 1.360 g/cm , Tˆ 200/2) K, 2Vmax 25.648, radiation
Mo-Ka, l ˆ 0.71073 Š, 0.38 w-scans with CCD area
detector, covering a whole sphere in reciprocal space,
19711 reflections measured, 4561 unique /R/int) ˆ
0.0268), 3788 observed /I > 2s/I)), intensities were
corrected for Lorentz and polarization effects, an
empirical absorption correction was applied using SA-
DABS[66] based on the Laue symmetry of the reciprocal
space, m ˆ 0.88 mm±1, Tmin 0.77, Tmax 0.90, structure
ˆ
ˆ
2
solved bydirect methods and refined against F with a
full-matrix least-squares algorithm using the SHELXTL-
PLUS /5.10) software package,[67] 320 parameters re-
fined, hydrogen atoms were treated using appropriate
riding models except for H1, which was refined isotropi-
cally, goodness of fit 1.03 for observed reflections, final
602
ꢀ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2003, 345, 589 ± 602