Influence of process parameters on the reaction kinetics of the chromiumcatalyzed irimerization of ethylene

Article abstract of DOI:10.1002/chem.201000533

In this paper we report the results of an extensive experimental kinetic study carried out on the novel ethylene trimerization catalyst system, comprising the chromium source [CrCl₃(thf)₃] (thf=tetrahydrofuran), a Ph₂P-N(ZPr)-P(Ph)-N(ZPr)H (PNPNH) ligand (Ph= phenyl, /Pr = isopropyl), and triethylaluminum (AlEt₃) as activator. It could be shown that the initial activity shows a first-order dependency on the ethylene concentration. Also, a first-order dependency was found for the catalyst concentration. The initial activity follows a typical Arrhenius behavior with an experimentally determined activation energy of 52.6 kJmol^-1. At elevated temperatures (ca. 80⁰C), a significant deactivation was observed, which can be tentatively traced back to a ligand rearrangement in the presence of AlEt₃. After a fast initial phase, a pronounced 'kink' in the ethylene-uptake curve is observed, followed by a slow, almost linear, further increase of the total ethylene consumption. The catalyst composition, in particular the ligand/chromium and the cocatalyst/chromium molar ratio, has a strong impact on the catalytic performance of the trimerization of ethylene. 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

Full text of DOI:10.1002/chem.201000533

FULL PAPER
DOI: 10.1002/chem.201000533
Influence of Process Parameters on the Reaction Kinetics of the Chromium-
Catalyzed Trimerization of ACTHNUTRGNEUEGN thylene
Anina Wçhl,[a] Wolfgang Mꢀller,*[a] Stephan Peitz,[b] Normen Peulecke,[b]
Bhaskar R. Aluri,[b] Bernd H. Mꢀller,[b] Detlef Heller,[b] Uwe Rosenthal,*[b]
Mohammed H. Al-Hazmi,[c] and Fuad M. Mosa[c]
Abstract: In this paper we report the
results of an extensive experimental ki-
netic study carried out on the novel
ethylene trimerization catalyst system,
comprising the chromium source
first-order dependency was found for
the catalyst concentration. The initial
activity follows a typical Arrhenius be-
havior with an experimentally deter-
was observed, which can be tentatively
traced back to a ligand rearrangement
in the presence of AlEt3. After a fast
initial phase, a pronounced ꢀkinkꢁ in
the ethylene-uptake curve is observed,
followed by a slow, almost linear, fur-
ther increase of the total ethylene con-
sumption. The catalyst composition, in
particular the ligand/chromium and the
cocatalyst/chromium molar ratio, has a
strong impact on the catalytic perfor-
mance of the trimerization of ethylene.
mined
activation
energy
of
[CrCl
U
52.6 kJmolꢀ1. At elevated temperatures
Ph2P-N
N
U
(ca. 808C), a significant deactivation
ligand (Ph=phenyl, iPr=isopropyl),
and triethylaluminum (AlEt3) as acti-
vator. It could be shown that the initial
activity shows a first-order dependency
on the ethylene concentration. Also, a
Keywords: chromium · homogene-
ous catalysis · kinetics · N,P ligands ·
oligomerization
Introduction
market segment strongly depends on the 1-olefin chain
length. Through the increasing demand for the high-value
comonomer fraction (C6 and C8) in recent years, many in-
dustrial and academic research groups focused their interest
on selective oligomerization, in particular, the tri- and tetra-
merization of ethylene.[2] Consequently, the number of pat-
ents and scientific publications for intentional routes to 1-
hexene and even 1-octene increased drastically in recent
years.
The most prominent and one of the earliest highly selec-
tive catalysts was developed by Chevron Phillips in the late
1980s,[3] yielding 1-hexene in an overall selectivity greater
than 90%. After various improvements, the Chevron Phil-
lips chromium-based catalyst became the basis of the first
commercialized trimerization technology, implemented in a
47000 taꢀ1 plant in Qatar that went successfully on stream
in 2003.[4]
The existing technologies for producing linear alpha-olefins
(LAOs) are mostly based on ethylene oligomerization
routes through an ethylene insertion/b-elimination mecha-
nism resulting in a mathematical product distribution like
Schulz–Flory or Poisson.[1] Thus, the conventional full-range
producers (e.g., Shell, Chevron Phillips, INEOS, UNITED)
have to meet a formidable challenge to match the market
demand. The reason is that the size and growth of each
[a] Dr. A. Wçhl, Dr. W. Mꢂller
Linde AG, Linde Engineering Division
Dr.-Carl-von-Linde-Strasse 6–14, 82049 Pullach (Germany)
Fax : (+49)89-744454981
E-mail: Wolfgang.mueller@linde-le.com
However, to reach the extraordinary selectivity for ob-
taining 1-hexene, the traditional insertion/b-elimination
mechanism has to be overcome. The selectivity towards 1-
[b] S. Peitz, Dr. N. Peulecke, Dr. B. R. Aluri, Dr. B. H. Mꢂller,
Dr. D. Heller, Prof. Dr. U. Rosenthal
Leibniz-Institut fꢂr Katalyse an der Universitꢃt Rostock e.V.
Albert-Einstein-Strasse 29A, 18059 Rostock (Germany)
Fax : (+49) 381-128151176
hexene is explained by
a
metallacycle mechanism
(Scheme 1), which was proposed initially by Manyik et al.[5]
and expanded later by Briggs.[6] This widely accepted mech-
anism could be supported by several studies for ethylene tri-
merization.[7]
E-mail: Uwe.rosenthal@catalysis.de
[c] Dr. M. H. Al-Hazmi, F. M. Mosa
Saudi Basic Industries Corporation
P.O. Box 42503, Riyadh 11551 (Saudi Arabia)
Chem. Eur. J. 2010, 16, 7833 – 7842
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7833

mental second-order rate dependency on ethylene concen-
tration for three different chromium-based trimerization cat-
alyst systems. These results led to the conclusion that the
rate-determining step is the formation of a metallacyclopen-
tane intermediate from two coordinated ethylene molecules.
Contrary to these results, modeling studies carried out by
van Rensburg et al.[18] for a “Cr–pyrrolyl” catalyst support a
first-order dependency with respect to ethylene concentra-
tion. Budzelaar et al.[19] as well as Tobisch and Ziegler[20]
theoretically investigated the titanium-catalyzed trimeriza-
tion mechanism and also identified the metallacycle growth
as the rate-determining step, consequently leading to a first-
order dependency. In contrast, a DFT study conducted by
Raybaud et al.[21] for a titanium-based system proposed that
the ring-opening reaction of the seven-membered metallacy-
cle is rate determining. Kuhlmann[15] found in his experi-
mental studies of the “Cr–PNP” system a first-order ethyl-
ene dependency for the trimerization reaction and a second-
order dependency for the tetramerization. This results in a
overall reaction order of 1.71, which is in good agreement
with the observed noninteger reaction order of 1.57 by
Walsh et al.,[14] suggesting a competition of two different
pathways for tri- and tetramerization.
Scheme 1. Metallacycle mechanism for ethylene trimerization proposed
by Briggs.[6]
In general, selective oligomerization catalysts are complex
systems that consist of a transition metal, a ligand, and
mostly of an aluminum alkyl as cocatalyst. Besides the
usage of early transition metals (Ti,[8] Zr,[9] V,[10] Ta[7b,11]), in
the majority of all cases, homogeneous chromium-based eth-
ylene-trimerization catalysts are used. The most commonly
used ligands consist of aromatic or multidentate structures
with phosphine, amine, ether, or thioether donors.[2] To elu-
cidate the necessary characteristics of these ligands to cata-
lyze the selective oligomerization, a great number of studies
were conducted investigating the influences of different
ligand substituents, for example, for Sasolꢁs tri- and tetra-
As a result of the different studies and their inconsisten-
cies, one can speculate that the differences in catalytic sys-
tems lead to a different kinetic behavior as well. For a
better understanding of the metallacycle mechanism, further
detailed kinetic investigations are vitally important.
ACHTUNGTRENNUNG
In their extensive report on kinetic measurements for the
ethylene tri- and tetramerization system consisting of the
for selective tri- and tetramerization increased in recent
years, unfortunately only few publications[8a,14–16] deal with
kinetic investigations, although they are essential for the un-
derstanding of the catalytic mechanism and help to elucidate
the complex reaction network. With a good comprehension
of the influence of reaction conditions, the process can be
designed and even optimized, in particular, with regard to
activity, selectivity, and stability. Moreover, kinetic measure-
ments form the basis for the design of a pilot- and, eventual-
ly, a technical-scale plant.
precatalyst [Cr
ACHTUNGTERN(NUNG acac)3] (acac=acetylacetonate), Ph2PN-
AHCTUNGTRENNUNG
vator, Walsh et al.[14] found that the initial reaction rate fol-
lows a typical Arrhenius behavior with an apparent activa-
tion energy of 64.6 kJmolꢀ1 in the temperature range be-
tween 35 and 458C. However, above 458C the deactivation
rate is strongly influenced by temperature, which is consis-
tent with findings by Kuhlmann.[15] Although Sasolꢁs “Cr–
PNP” tetramerization system is highly sensitive to elevated
temperature, like other trimerization systems,[13] the
“Cr–MeOPNP” trimerization catalyst system shows, for exam-
ple, nearly no deactivation at 1108C.[13]
Table 1 gives an overview of reported kinetic measure-
ments. Some discrepancies regarding the influence of ethyl-
ene concentration are noteworthy. For instance, Manyik
et al.,[5] Yang et al.,[17] and Wass et al.[13e] reported an experi-
In this paper we report the results of an extensive experi-
mental kinetic study carried out
on the novel ethylene trimeriza-
tion catalyst system comprising
the chromium source [CrCl3-
Table 1. Overview of reported experimental kinetic measurements of selective ethylene tri- and tetrameriza-
tion processes.
Catalyst system[a]
Product Reaction order Reaction order
[catalyst] [C2H4]
Ea
[kJmolꢀ1
]
Ref.
AHCTUNGTRENNUNG
G
E
U
A
ACHTUNGTRENNUNG
(2-EH)3/PIBAO[b]
(2-EH)3/2,5-Me2-pyrrole/AlEt3/Cl compound
3A
(thf)3]/Ar2P(Me)NPAr2/MAO[c]
(acac)3]/Ph2P
(iPr)NPPh2/MAO[d]
(acac)3]/Ph2P
(THN)NPPh2/MMAO-3A[e]
C6
C6
C6
C6/C8
C6/C8
C6
–
–
–
1
0.6
1
2
2
2
1.57
1.71
1
“near 0”
[5]
[17]
[13e]
[14]
[15]
[8a]
(PNPNH) ligand (Ph=phenyl,
iPr=isopropyl), and triethylalu-
minum (AlEt3) as activator.[22,23]
The influence of temperature,
residence time, ethylene con-
centration, catalyst composi-
tion, and concentration on the
catalytic trimerization reaction
Cr
Cr
–
–
64.6
23.1
27.7
A
N
[Cr
[Cr
N
N
R
TiCl3/Me3SiC5H3Me2C6H3Me2/MAO
[a] Studies were conducted by using different parameters. [b] ChromiumACTHNURGTNE(UNG III)–2-ethylhexanoate/poly(isobutyl
aluminum oxide). [c] Ar: ortho-methoxyphenyl; MAO=methylaluminoxane. [d] acac=acetylacetonate.
[e] THN=1,2,3,4-tetrahydronaphthyl PNP; MMAO-3A=modified methylaluminoxane, type 3A.
7834
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7833 – 7842

Trimerization of Ethylene
FULL PAPER
was investigated in detail. The liquid-phase trimerization re-
action was performed in a semibatch stirred tank pressure
reactor (0.3 L), while isobaric conditions were provided by
feeding ethylene on demand. Through recording the ethyl-
ene uptake over time, statements about changes in the reac-
tion rate by, for example, catalyst deactivation could be de-
rived easily. Due to the relatively low conversion over time,
the reactor showed pseudodifferential behavior, thus provid-
ing the required sensitivity for the observation of changes in
the reaction rate as a function of process parameter varia-
tion. In a separate study, it could be shown that gas-to-liquid
phase-transfer limitation could be excluded and the intrinsic
kinetics of the ethylene trimerization reaction were ob-
served.
Figure 1. Ethylene-uptake curve for a typical ethylene trimerization reac-
tion catalyzed by the [CrCl3ACHTNUTRGNEG(UN thf)3]/PNPNH/AlEt3 catalyst system. (Run
conducted at 658C, 30 bar, [Cr]=1 mmolLꢀ1, t=4 h, [L]/[Cr]=1.5 mol/
mol, [Al]/[Cr]=70 mol/mol.)
Results and Discussion
In the following, the results of the evaluation of various pro-
cess parameters and kinetic studies are discussed. For this
purpose the ethylene-uptake curve over the oligomerization
run time is an integral part of the experimental evaluation.
Due to the fact that mass-transfer limitations can be exclud-
ed, the consumption of ethylene, monitored over time, can
be directly translated into the time characteristics of the cat-
alytic ethylene trimerization. For clarity, the filling process
of the reactorꢁs headspace and the rapid mass transfer (gas-
phase ethylene to dissolved ethylene) in the beginning of
each individual experiment is generally omitted from the il-
lustrations. However, the ethylene-uptake curve includes
both the ethylene that is converted to product and the ethyl-
ene that is required for saturation of the newly formed prod-
uct.
Figure 2. Concentration–time profile of the formation of 1-hexene and
the most important by-products 1-butene and branched decenes with the
[CrCl3ACTHNUGTRNE(NUG thf)3]/PNPNH/AlEt3 catalyst system. (Runs conducted at 658C,
30 bar, [Cr]=1 mmolLꢀ1, [L]/[Cr]=1.5 mol/mol, [Al]/[Cr]=70 mol/mol.)
Influence of residence time on 1-hexene and by-product for-
mation: To resolve the kinetics of the ethylene trimerization
and the formation of the most important by-products, ex-
periments with termination after varying residence times
were conducted. As opposed to time-resolved sampling
during the experiment, this has the advantage that the reac-
tion is not disturbed by the removal of samples. Further-
more, an entire mass balance, including gas-phase product,
can be determined for each timescale.
The measured ethylene-uptake curves of the conducted
experiments are consistent with the example shown in
Figure 1. It illustrates clearly the complex time characteris-
tics of the ethylene-uptake curve with its steep initial slope,
followed by a ꢀkinkꢁ and, finally, a slow, almost linear, fur-
ther increase.
The kink kinetics observed in the ethylene-uptake curve
is also reflected by the concentration–time profile of 1-
hexene (Figure 2) and is, therefore, an intrinsic property of
the ethylene trimerization reaction. In conjunction with the
fact that more than 90 wt% of the product is 1-hexene, con-
clusions about the kinetics drawn from the ethylene-uptake
curve are in good accordance with the kinetics of the 1-
hexene formation.
Nevertheless, besides 1-hexene, also 1-butene and
branched decenes are formed as major by-products. Where-
as the trimerization of ethylene follows the kink kinetics,
the rate of the dimerization towards 1-butene is nearly con-
stant right from the beginning of the reaction. After approx-
imately one hour, the ratio of the formation rates of 1-
hexene and 1-butene decreases slightly, leading finally to a
decline in the C6 selectivity over time. The formation of 1-
butene might be the result of an early b-H elimination of
the C4 chain from the chromacyclopentane.[24] However,
due to the different kinetics of the formation of 1-butene in
contrast to 1-hexene, it is doubtful whether the same catalyt-
ically active species is responsible for both 1-olefins and
whether 1-butene is actually an inherent side product of the
catalytic cycle for 1-hexene. Accordingly, a different mecha-
nism from the metallacycle route, for example, insertion/b-
H elimination, is conceivable as well and was in fact pro-
posed for the nickel-catalyzed ethylene dimerization.[25] Un-
fortunately, only very little is known about the mechanism
of the formation of 1-butene as a typical by-product during
Chem. Eur. J. 2010, 16, 7833 – 7842
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7835

W. Mꢂller, U. Rosenthal et al.
ethylene trimerization because the gas-phase product was
not analyzed in most of the cases.
clusion that the ethylene concentration is not the determin-
ing factor for the strong loss in activity after an initial phase
of about 20 min (at 50 bar). Secondly, the induction period,
which indicates the time required for the formation of the
active catalytic species (time period between ethylene pres-
surization and start of trimerization reaction), follows the
same pattern of about 2.5 min for each pressure. Conse-
quently, one can conclude that the catalystꢁs activation is
not strongly affected by ethylene concentration.
In contrast, the concentration of branched decenes in-
creases in parallel with the concentration of 1-hexene, which
results in a constant mass relation of the C10 fraction to 1-
hexene over time (about 3 wt% of total product). Branched
decenes seem to be an inherent by-product of selective eth-
ylene oligomerization and were observed in several catalytic
systems.[7e,8,14,17] They are the result of secondary incorpora-
tion reactions, in which the primary product 1-hexene com-
petes with ethylene either in the coupling reaction resulting
in butyl-substituted chromacyclopentanes or in the subse-
quent insertion into the chromacyclopentane. Unfortunately,
it was not possible to conclude which pathway is applicable,
since the identification of the different C10 isomers was not
successful due to poor differentiation with GC–MS. Whereas
Hagen[8] found indications that the insertion of 1-hexene in
the metallacyclopentane is more likely, Deckers et al.[26]
postulated that 1-hexene is incorporated during the coupling
reaction. Regardless of what actual route the reaction takes
in detail, the effective reaction rate of cotrimerization with
1-hexene is presumably much slower than the rate of ethyl-
ene trimerization itself. This observation has significant con-
sequences with respect to the overall kinetics. Depending on
the actual ethylene/1-hexene ratio, a significant part of the
active catalytic species will be ꢀblockedꢁ by the slower cotri-
merization reaction. If this is true, one would also predict an
increasing reversible inhibition of the catalyst as more and
more product is being formed,[8] which results in decreasing
activity with increasing 1-hexene concentration.
Considering the initial activity, where interference with
secondary effects is not yet very likely, as a function of eth-
ylene pressure (Figure 4), the direct proportionality of the
Figure 4. Effect of pressure on the initial activity of the [CrCl3ACHTUNGTRENNUNG(thf)3]/
PNPNH/AlEt3 catalyst system. (Runs conducted at 658C, [Cr]=
1 mmolLꢀ1, [L]/[Cr]=1.5 mol/mol, [Al]/[Cr]=70 mol/mol.)
Effect of ethylene pressure: To elucidate the influence of
ethylene concentration on the performance of the novel
chromium-catalyzed ethylene trimerization system, the eth-
ylene pressure was varied at 658C in the range of 10 to
60 bar with 10 bar increments.
initial activity with respect to ethylene pressure becomes
evident in the measured pressure regime from 10 to 60 bar.
On the basis of Henryꢁs law and the fact that the phase
transfer of ethylene to toluene is fast relative to the reac-
tion, one can deduce a first-order dependency of the initial
*
From the ethylene-uptake curve of 10, 30, and 50 bar in
Figure 3, several effects can be observed. The first observa-
tion is that the rate of ethylene uptake increases with in-
creasing ethylene concentration, while the time characteris-
tics still show the typical kink kinetics. This leads to the con-
activity (ai) with respect to the ethylene concentration (cEt,L
in the liquid phase) [Eq. (1)]:
1
*
ai ꢁ ðcEt,LÞ
ð1Þ
Of course, this “first approximation” of the reaction order
by means of the pressure dependency of the initial activity
gives only an “integrated” order over the entire catalytic
process, inherently including all side reactions, like 1-butene
or branched-decenes formation. However, because of the
high selectivity towards 1-hexene, this observed first-order
dependency of the ethylene trimerization with respect to
ethylene concentration appears rather conclusive.
As discussed previously, some discrepancies do exist re-
garding the influence of ethylene concentration on the selec-
tive oligomerization processes. In contrast to the experimen-
tal observations of this study, Manyik et al.,[5] Yang et al.,[17]
and Wass et al.[13e] reported a second-order dependency for
different chromium-based catalyst systems. A first-order de-
pendency, as observed here, is supported by several theoreti-
cal modeling studies,[18–20] and was also observed experimen-
Figure 3. Effect of pressure on the total ethylene uptake of the [CrCl3-
ACHTUNGTRENNUNG
7836
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7833 – 7842

Trimerization of Ethylene
FULL PAPER
tally by Kuhlmann[15] for the trimerization reaction using
Sasolꢁs “Cr–PNP” tri- and tetramerization catalyst system.
and a further interaction of the catalytically active species
with ethylene is most likely. Naturally, all these individual
reaction steps are determined by their own temperature-de-
pendent kinetics.
From Figure 5, it is clear that the maximum rates ob-
served in the beginning of each run and reflecting the intrin-
sic reaction rate of the catalytic trimerization reaction prior
to catalyst deactivation processes, increase strongly with in-
creasing temperature. By interpreting the temperature de-
pendency of the reaction rate/activity, one has, of course, to
take into account that the equilibrium concentration of eth-
Investigation of temperature dependency: The next set of
experiments consists of a temperature variation from 30 to
908C at otherwise constant parameters in which isothermal
conditions were ensured during all experiments. The selec-
tivity decreases dramatically above 608C (up to 84 wt% at
908C) due to the increased formation of the main by-prod-
uct 1-butene at elevated temperatures.
The ethylene-uptake curves at selected temperatures are
compared in Figure 5. It becomes evident that increased
temperature affects the kinetic behavior of the trimerization
*
ylene cEt,L in the liquid phase is also a strong function of
temperature (T) [Eq. (2) in which pEt is the ethylene pres-
sure and HEt is the Henry constant for ethylene in tolu-
ene].[27]
ꢀ
ꢁ
1228k
T
ꢀ6:12
ð2Þ
*
cEt,L ¼ pEtHEt ¼ pEte
The obtained direct proportionality between the initial ac-
tivity and the ethylene concentration helps to describe the
temperature dependency of the initial activity. Therefore, a
simplified approach to describe the relation between initial
activity and both temperature and ethylene concentration
was applied [Eq. (3), in which ki [sꢀ1] represents the temper-
ature-dependent reaction rate constant for the trimerization
reaction (4) and ki0 is a pre-exponential factor].
Figure 5. Effect of temperature on the total ethylene uptake of the
[CrCl3ACHTUNGTRENNUNG(thf)3]/PNPNH/AlEt3 catalyst system. (Runs conducted at 30 bar,
[Cr]=1 mmolLꢀ1, [L]/[Cr]=1.5 mol/mol, [Al]/[Cr]=70 mol/mol.)
1
*
ai ꢁ kiðcEt;LÞ
ð3Þ
ð4Þ
Ea
RT
ki ¼ ki0eꢀ
catalyst system in different ways. Firstly, the curvature of the
ethylene uptake changes completely from smoothly shaped
with nearly constant consumption over time at low tempera-
tures, to the already discussed kink kinetics at temperatures
above 608C. The strong kink is presumably the result of the
catalystꢁs deactivation or a change in the nature of the cata-
lystꢁs active species at high temperatures, leading to a less
active catalytic species. Naturally, this process is also driven
by temperature-dependent kinetics with its own activation
energy. Hints on a molecular level were already given by
the mechanistic investigation of the influence of elevated
temperatures on the PNPNH/AlEt3 system.[23] It could be
Figure 6 shows the Arrhenius diagram of the reaction rate
constant ki derived from the quotient of the initial activity ai
*
and the theoretical ethylene equilibrium concentration cEt,L
.
In the range from 30 to 708C, the reaction rate follows the
typical Arrhenius behavior with an experimentally derived
apparent activation energy (Ea) of 52.6 kJmolꢀ1. At 708C, a
relatively abrupt change of this empirical apparent activa-
tion energy is observed, giving approximately 12.7 kJmolꢀ1
proven that, at higher temperatures, a ligand reACTHNUTRGNEaUNG rrangement
takes place in the presence of AlEt3. Although this effect
was not verified until now for the chromium-containing cat-
alyst system, an analogous process might be a possible ex-
planation for the catalyst deactivation and the observed
kink in the ethylene-uptake curve.
The catalyst activation is also strongly affected by temper-
ature. For example, at 308C, the trimerization reaction did
not start until 30 min after pressurization with ethylene,
whereas the induction period at 658C lies around 2.5 min.
This leads to the aforementioned conclusion that the catalyt-
ically active species has still to be formed in the presence of
ethylene. During activation with AlEt3, the deprotonation of
the ligand, the consecutive coordination to the chromium,
Figure 6. Effect of temperature on the initial rate constant ki of the
[CrCl3ACTHNUGTRNE(NUG thf)3]/PNPNH/AlEt3 catalyst system. (Runs conducted at 30 bar,
[Cr]=1 mmolLꢀ1, [L]/[Cr]=1.5 mol/mol, [Al]/[Cr]=70 mol/mol.)
Chem. Eur. J. 2010, 16, 7833 – 7842
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7837

W. Mꢂller, U. Rosenthal et al.
for temperatures between 70 and 908C. The reason for this
behavior could be a very rapid onset of the catalyst deacti-
vation. It is also conceivable that elevated temperatures
might prevent the catalystꢁs complete activation with the
consequence that not the entire, theoretically possible,
amount of catalyst is involved in the trimerization reaction.
The pronounced reduction in the activation energy may also
be explainable by the strongly incipient 1-butene formation
at elevated temperatures. However, this general interpreta-
tion is slightly simplified, since the trimerization relies on a
complex reaction network.
After all, one has always to consider that the empirical ki-
netic figures are based on the interpretation of the initial
slope of the ethylene-uptake curves, reflecting the reaction
in its entirety, which means that all side reactions are includ-
ed. Due to the long activation period, depending on the
temperature, it is quite possible that the derived initial activ-
ity might slightly interfere with the formation of the active
catalytic species.
Figure 7. Effect of catalyst concentration on the total ethylene uptake of
the [CrCl3ACHTNUGTRNEUNG(thf)3]/PNPNH/AlEt3 catalyst system. (Runs conducted at
658C, 30 bar, [L]/[Cr]=1.5 mol/mol, [Al]/[Cr]=70 mol/mol.)
As mentioned earlier, only few publications[8a,14,15] deal
with kinetic investigations of selective ethylene oligomeriza-
tion catalyst systems. Currently, only one extensive report
on kinetic measurements is known for an ethylene tri- and
tetramerization system; it consists of the precatalyst [Cr-
ACHTUNGTRENNUNG(acac)3], Ph2PNACHTUNGTRENNUNG(iPr)PPh2 as ligand, and MAO as activa-
tor.[14] It is very interesting that the published results, with
respect to temperature effects, are in several points in good
agreement with the ones discussed here, although Sasolꢁs
“Cr–PNP” system is able to tetramerize ethylene as well.
Walsh et al. found a comparable temperature effect on the
ethylene-uptake behavior by regarding the increase in cur-
vature of the uptake curve profile with increasing tempera-
ture, which also resulted in a kind of kink kinetics at 608C.
The catalystꢁs activation is a function of temperature as
well. It was found that the initial reaction rate follows a typ-
ical Arrhenius behavior with an apparent activation energy
of 64.6 kJmolꢀ1 in the temperature range between 35 and
458C, lying in the same regime as for the “Cr–PNPNH”
system. In contrast to the system of this investigation, the
“Cr–PNP” system starts to deactivate at even lower temper-
atures of around 458C.[14,15]
Figure 8. Effect of catalyst concentration on the initial activity of the
[CrCl3ACTHNUGTRNE(NUG thf)3]/PNPNH/AlEt3 catalyst system. (Runs conducted at 658C,
30 bar, [L]/[Cr]=1.5 mol/mol, [Al]/[Cr]=70 mol/mol.)
the reaction kinetics, like the big difference in liquid-phase
composition and amount, presumably leads to the observed
discrepancy. The results also confirm that the experiments
are not inadvertently compromised by gas-to-liquid mass-
transfer limitations.
In general, the observed first-order dependency was ex-
pected and is in good accordance with the findings of Hagen
for the Ti-based trimerization system[8a] and with the results
of Walsh et al. for the Cr-based tri- and tetramerization
system.[14]
Influence of catalyst concentration: To investigate the influ-
ence of the catalyst concentration on the catalytic perfor-
mance of the “Cr–PNPNH” system, the concentration was
varied in the range of 0.25 and 1.5 mmolLꢀ1 at constant
ligand/chromium and Al/chromium molar ratio.
Effect of ligand/chromium ratio: Besides process conditions
like residence time, temperature, or ethylene pressure, the
catalyst composition has a major influence on the catalytic
performance as well. This pertains in particular to the molar
ratio of ligand to chromium and to the molar ratio of co-
The averaged apparent activity for all concentrations is in
the range of 5.0 kg
N
ACHTUNGTRENcNUNG atalyst to chromium. In the first step, the molar ratio of the
pendency in catalyst concentration, which is confirmed by
Figure 7, which shows the ethylene-uptake curves of experi-
ments with 50% more and 50% less catalyst concentration
than in the standard case.
When the initial activity is derived from the ethylene-
uptake curve (Figure 8) and normalized with the catalyst
amount, the first-order dependency in the catalyst concen-
tration is undeniable. Later on, secondary effects influencing
PNPNH ligand to the chromium precursor [CrCl
varied in the range from 0.5 to 3.0.
3ACHTUNGTERN(NUNG thf)3] was
From Figure 9 it becomes clear that the average activity is
a strong function of the molar ligand/chromium ratio, with a
distinctive maximum at [L]/[Cr]=1.75, which indicates that
at this ratio, the highest number of catalytically active spe-
cies for ethylene trimerization is formed. The corresponding
overall 1-C6 selectivity shows a similar dependency on the
7838
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7833 – 7842

Trimerization of Ethylene
FULL PAPER
or nitrogen-containing fragments (Scheme 2). 31P NMR
spectroscopy studies, stability investigations of the [CrCl3-
AHCTUNGRTEG(NNNU thf)3]/ligand slurry, and the isolation and characterization of
Figure 9. Effect of molar ligand to chromium ratio on the average activity
of the [CrCl3ACHTUNGTRENNUNG(thf)3]/PNPNH/AlEt3 catalyst system. (Runs conducted at
658C, 30 bar, 120 min, [Cr]=1.0 mmolLꢀ1, [Al]/[Cr]=70 mol/mol.)
Scheme 2. Possible reaction mechanism for the observed PNPNH ligand
decomposition.
ligand/Cr molar ratio, which also has a maximum between
1.5 and 1.75 (Figure 10), whereas the value at a ratio of 3.0
is an experimental uncertainty. The decrease in selectivity at
low and high ratios can be traced back to an increased for-
mation of 1-butene, while the decene fraction remains con-
stant for the entire experimental series.
an ammonium-ate complex of the PNPNH ligand coordinat-
ed to chromium tetrachloride, as a result of the ligand frag-
ments interacting with the chromium center, gave indication
that this decomposition occurs. Consequently, the ligand
also acts as scavenger for the liberated HCl during catalyst
activation.
However, on the other hand, a big surplus of ligand also
reduces the catalytic performance with regard to activity
and selectivity. Clearly, more than one ligand can coordinate
to chromium and hence shield the free coordination site,
which is necessary for the coordination of two ethylene mol-
ecules. Summing up, it becomes evident that a very fine bal-
ance between the amount of PNPNH ligand and chromium
precursor needs to be maintained, which is very susceptible
to small variations.
Effect of AlEt3/chromium ratio: As shown in several stud-
ies,[13e,17,28] the catalytic performance of selective oligomeri-
zation catalyst systems is also a function of the ratio of co-
Figure 10. Influence of the molar ligand to chromium ratio on the 1-C6
selectivity overall of the [CrCl3ACHTNUGTRENUNG(thf)3]/PNPNH/AlEt3 catalyst system.
AHCTUNGERTGcNNUN atalyst to catalyst. Therefore, the molar ratio of triethylalu-
(Runs conducted at 658C, 30 bar, 120 min, [Cr]=1.0 mmolLꢀ1, [Al]/
[Cr]=70 mol/mol.)
minum to chromium, in the following abbreviated as Al/Cr
ratio, was varied over a wide range between 5 and 170 at
otherwise constant process conditions for the “Cr–PNPNH”
trimerization catalyst system.
The Al/Cr ratio has a certain effect on the selectivity;
thus C6 selectivity and 1-hexene purity increase drastically
with decreasing Al/Cr ratio below 25, whereas at higher
ratios the influence is not as significant.
For the interpretation of the time characteristics of the tri-
merization reaction as a function of the Al/Cr ratio, selected
ethylene-uptake curves are shown in Figure 11. At ratios
above 70 the typical kink kinetics is observed, whereas with
an increasing excess of AlEt3 the kink in the ethylene-
uptake curve becomes sharper and occurs earlier. Interest-
ingly, both the initial and the final ethylene consumption
rate for molar ratios of 70 and 170 seem to be in the same
regime, leading to the conclusion that the amount of alumi-
num alkyl finally has no influence on the number of deacti-
vated catalyst species after the kink. With decreasing Al/Cr
ratio, the curvature of the ethylene-uptake curves decreases
At first glance, the observation that both activity and se-
lectivity show a maximum at a noninteger ligand/chromium
molar ratio seems very surprising. There is little doubt that
the active trimerization catalyst consists of one mononuclear
chromium center with one attached PNPNH ligand.[23] In
general, one can conclude that the formation of the catalyti-
cally active site is the result of a complex reaction equilibri-
um involving all three catalyst components, which leads to
the requirement of a large excess amount of ligand to reach
an optimum in activity.
One reason for the need of a large excess amount of
ligand relative to chromium might be that part of the ligand
decomposes during the activation procedure. During the co-
ordination of the PNPNH ligand to [CrCl3ACTHUNTRGNE(UGN thf)3] the NH
function of the ligand is deprotonated under potential liber-
ation of HCl. In a further step, the free HCl is able to split a
noncoordinated ligand molecule into several phosphorous-
Chem. Eur. J. 2010, 16, 7833 – 7842
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7839

W. Mꢂller, U. Rosenthal et al.
is not as distinctive, supposedly for the reason that not
enough ꢀfreeꢁ AlEt3 can compete with chromium for the
PNPNH ligand and initiate a rearrangement of the ligand
leading to a less active catalyst species. At higher Al/Cr
molar ratios, the equilibrium is shifted towards AlEt3,
making a deactivation more likely, which again becomes evi-
dent in the ethylene-uptake curve. However, even through
the simple interaction of the ligand with AlEt3, like the for-
mation of Al–amide bonds, the number of available ligand
molecules for the chromium-based trimerization catalyst is
reduced.
Figure 11. Effect of the AlEt3/Cr ratio on the total ethylene uptake of the
[CrCl3ACHTUNGTRENNUNG(thf)3]/PNPNH/AlEt3 catalyst system. (Runs conducted at 658C,
30 bar, [Cr]=1 mmolLꢀ1, [L]/[Cr]=1.5 mol/mol).
Conclusion
The reported experimental series were carried out to inves-
tigate the influence of single parameters on the novel “Cr–
PNPNH” trimerization catalyst system, while all other pro-
cess parameters were held constant. The results are required
for a better understanding of the catalytic system and they
deliver data for an upscaling of the process. The catalyst
system was not only defined by its chemical composition
and structure, but also by its individual window of operabili-
ty in terms of process conditions and catalyst composition.
The high number of possible process parameters comprises
a multidimensional parameter (hyper)space, in which the
catalyst can act. Consequently, the establishment of the best
set of parameters to reach an optimum in catalytic perfor-
mance is a great challenge.
as well, eventually leading to a smoothly bent ethylene-
uptake curve for a Al/Cr molar ratio of 15. However, the in-
itial slopes of all ethylene-uptake curves of experiments
with Al/Cr molar ratios above 15 are identical, while at
ratios below 15 a decline is observed, which indicates that
not the entire number of possible catalytic species is activat-
ed. The induction period for all Al/Cr molar ratios is
2.5 min, a typical value for experiments conducted at 658C.
Consequently, the kinetics of catalyst activation do not seem
to be a function of the amount of AlEt3 relative to chromi-
um.
In general, it becomes clear that no large excess of the
aluminum alkyl is necessary to activate the catalyst and,
moreover, that such an excess even has detrimental effects
on the catalytic performance, as was already shown for the
“Cr–PNP/SNS” trimerization system.[13c,28b] The optimum
lies around 15 mol cocatalyst per 1 mol chromium, which is
presumably also dependent on the amount of impurities
scavenged by the aluminum alkyl. This is quite low in com-
parison with the “Cr–PNP” tri- and tetramerization catalyst
system, for which at least 190 equiv MAO per 1 equiv chro-
mium are required for a complete activation, while a surplus
of MAO does not seem to have any impact on the catalytic
performance.[14]
Due to the fact that, besides the chromium precursor, the
PNPNH ligand also interacts intensely with the cocatalyst,
the essentially required amount of AlEt3 for catalyst activa-
tion is also a function of the molar ratio of PNPNH ligand
to chromium. Thus, a deprotonation of the ligandꢁs amine
function occurs presumably during activation, while a coor-
dination of the PNPNH ligand to the aluminum atom is also
possible.[23] Both, ligand deprotonation[13c,d,28b,29] and further
interaction with/stabilization of the active site by a chelating
framework[28b,30] is conceivable and was verified for several
systems.
However, some basic concepts for the novel trimerization
catalyst can be derived from previous results.
First of all, in the beginning the applied catalyst consists
of its single components. These are the chromium precursor
[CrCl3ACTHNUGTRNE(UNG thf)3] and the PNPNH ligand, which finally form the
active catalytic species in situ, by activation with AlEt3 in
the presence of ethylene. It is most likely that the PNPNH
ligand is just a precursor and undergoes deprotonation upon
addition of AlEt3, a reaction that is also highly dependent
on the reaction temperature.
It could be shown that the ethylene concentration has a
significant influence on the catalytic performance. So the in-
itial activity shows a first-order dependency on the ethylene
concentration. Also, a first-order dependency was found for
the catalyst concentration. The initial activity follows a typi-
cal Arrhenius behavior with an experimentally determined
activation energy of 52.6 kJmolꢀ1. Therefore, a simplified
expression for the initial activity could be derived [Eq. (5)]:
52:6 kJ molꢀ1
RT
1
1
ai ꢁ eꢀ
ðcEt,LÞ ðccatalyst
Þ
ð5Þ
*
Assuming that the ligand rearrangement at elevated tem-
peratures, as discussed in ref. [23], is responsible for the de-
activation, eventually leading to the observed kink in the
ethylene-uptake curve, one can conclude that this process of
deactivation is also a function of the excess amount of alu-
minum alkyl. For low Al/Cr molar ratios (e.g., 15) the kink
This empirically derived expression describes the depen-
AHCTUNGTRENNUNG
7840
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7833 – 7842

Trimerization of Ethylene
FULL PAPER
means that the pressure and
temperature
dependencies
might vary with different cata-
lyst compositions.
At elevated temperatures
(ca. 808C) a significant deacti-
vation was observed, which can
be tentatively traced back to a
ligand rearrangement in the
presence of AlEt3. After a fast
initial phase, a pronounced kink
in the ethylene-uptake curve
was observed, followed by a
slow, almost linear, further in-
crease of the total ethylene con-
sumption. Although it is not
Figure 12. Schematic of the 0.3 L Parr kinetic test rig.
clear so far whether the ligand
rearrangement is really respon-
autoclave could be depressurized and evacuated by means of a rotary
vane vacuum pump (10ꢀ1 mbar). A low-pressure argon supply (1.5 bar
absolute) was used to obtain an inert atmosphere. A schematic of the ex-
perimental test rig is given in Figure 12. Before conducting an experi-
ment, the reactor was heated to 1008C at reduced pressure for several
hours and purged with argon several times to eliminate traces of water
and oxygen. The reactor was then allowed to cool down to reaction tem-
perature and the prepared catalyst solution was transferred to the evacu-
ated reactor by means of the pressure differential to the argon inert at-
mosphere. The speed of the gas entrainment stirrer was set to 1000 rpm
to exclude gas-to-liquid mass-transfer limitations. In the next step, the re-
actor was pressurized with ethylene to start the experiment. All reactions
were performed at a constant reaction temperature. After the reaction
time the ethylene supply was closed and the reaction solution was dis-
charged by transferring the liquid by means of the headspace pressure
via the reactorꢁs dip tube into a round-bottomed flask with HCl-acidified
water for rapid catalyst quenching. The volume of the gaseous product
phase was measured by a gas meter and collected completely in a gas
bag. The mass balance of the experiment was determined by quantifica-
tion and GC/flame ionization detection (FID) analysis of the gaseous
and liquid product separately, followed by comparison with the ethylene-
uptake data.
sible for the deactivation, one can still assume that, at high
temperatures, the catalytically active species is converted to
a less active one. Moreover, some evidence for competitive
reversible product inhibition was found, which also had a
significant effect on the kinetics. The formation of branched
decenes is an indication of this blocking of the active sites
by the product, since this side reaction reflects the “irrever-
sible part” of this blocking process.
However, all concepts developed so far for a reaction
mechanism are based on one single catalyst system chosen
from a large family of catalysts consisting of a chromium
precursor and a ligand with the basic chemical structure
PNPNH. Through variation of substituents of the ligand,
through the variation of the chromium precursor, or even
through the choice of the solvent, the kinetics can be influ-
enced fundamentally.
Experimental Section
Catalyst preparation: The catalyst was always activated ex situ under
argon. Hence, the required amount of the chromium precursor ([CrCl3-
Materials: For the selective oligomerization experiments, the following
ACHTUNGRTEN(NUGN thf)3]) and the PNPNH ligand were weighed under an inert atmosphere.
chemicals were used: [CrCl
(thf)3] (97% purity) and AlEt3 solution
Then the catalyst components were transferred into a Schlenk tube and
dissolved under stirring in anhydrous toluene (100 mL). After suspending
the Cr compound and the ligand, the required amount of AlEt3 was
added, yielding the activated catalyst solution. The solution was immedi-
ately transferred to the suitably tempered reactor and the reaction was
started by opening the ethylene supply.
(1.9 molLꢀ1 in toluene) was obtained from Sigma–Aldrich and used with-
out further purification; toluene (>99.9% purity) was obtained from
Merck Chemicals and was dried over sodium with benzophenone and
was then distilled in an inert gas atmosphere (Ar); argon 5.0 and ethyl-
ene 3.0 were purchased from Linde Gas and used as received; the
Ph2PNACHTUNGTRENNUNG(iPr)P(Ph)NACHTUNGTRENNUNG(iPr)H (PNPNH) ligand was synthesized according to
published procedures.[22,23] All air- and moisture- sensitive compounds
were handled under an argon atmosphere using standard Schlenk tech-
niques or in a glove box.
Acknowledgements
Experimental setup and procedural methods for ethylene trimerization:
The syntheses were performed in a 0.3 L Parr autoclave equipped with a
hollow-shaft gas-entrainment stirrer. The ethylene supply was maintained
by an aluminum pressurized gas cylinder on a balance to monitor the eth-
ylene consumption over time by means of a computerized data-acquisi-
tion system. The ethylene-supply pressure was controlled by a pressure
reducer, which held the pressure constant over the entire run time. The
temperature of the autoclave was measured by a thermocouple inside the
reactorꢁs liquid phase and controlled electronically by an electrical heat-
ing jacket and a water cooling coil, submerged in the liquid phase. The
The authors would like to express their gratitude to Linde Engineering
and SABIC for the permission to publish this work. Furthermore, we
wish to thank Linde Engineeringꢁs R&D Chemistry Department for their
valuable contributions as well as the chemical staff at LIKAT, in particu-
lar Petra Bartels for her assistance with the ligand synthesis.
[1] a) E. Burridge, Chemical Profile: Alpha olefins, ICIS Chemical
Business, February 2008; b) G. R. Lappin, J. D. Sauer (Eds.), Alpha
Chem. Eur. J. 2010, 16, 7833 – 7842
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7841

W. Mꢂller, U. Rosenthal et al.
Olefins Applications Handbook, Marcel Dekker, New York/Basel,
1989.
[14] R. Walsh, D. H. Morgan, A. Bollmann, J. T. Dixon, Appl. Catal. A
2006, 306, 184–191.
[15] a) S. Kuhlmann, Dissertation, Technische Fakultꢃt an der Friedrich-
Alexander-Universitꢃt Erlangen-Nꢂrnberg, 2006; b) S. Kuhlmann,
J. T. Dixon, M. Haumann, D. H. Morgan, J. Ofili, O. Spuhl, N. Tac-
cardi, P. Wasserscheid, Adv. Synth. Catal. 2006, 348, 1200–1206.
[16] T. Jiang, Y. Ning, B. Zhang, J. Li, G. Wang, J. Yi, Q. Huang, J. Mol.
Catal. A 2006, 259, 161–165.
[2] a) J. T. Dixon, M. J. Green, F. M. Hess, D. H. Morgen, J. Organomet.
Chem. 2004, 689, 3641–3668; b) A. Forestiꢅre, H. Olivier-Bourbi-
gou, L. Saussine, Oil Gas Sci. Technol. 2009, 64, 649–667.
[3] W. K. Reagan (Phillips Petroleum Co.), EP 0417477, 1991.
[4] Chevron Phillips Chemical Company, Press release 10/31/2005 via
web, http://www.cpchem.com.
[17] Y. Yang, H. Kim, J. Lee, H. Paik, H. G. Jang, Appl. Catal. A 2000,
193, 29–38.
[5] R. M. Manyik, W. E. Walker, T. P. Wilson, J. Catal. 1977, 47, 197–
209.
[18] W. Janse van Rensburg, C. Grovꢆ, J. P. Steynberg, K. B. Stark, J. J.
Huyser, P. J. Steynberg, Organometallics 2004, 23, 1207–1222.
[19] A. N. J. Blok, P. H. M. Budzelaar, A. W. Gal, Organometallics 2003,
22, 2564–2570.
[20] S. Tobisch, T. Ziegler, Organometallics 2003, 22, 5392–5405.
[21] T. J. M. de Bruin, L. Magna, P. Raybaud, H. Toulhoat, Organometal-
lics 2003, 22, 3404–3413.
[22] P. M. Fritz, H. Bçlt, A. Wçhl, W. Mꢂller, F. Winkler, A. Wellenhof-
er, U. Rosenthal, M. Hapke, N. Peulecke, B. H. Mꢂller, M. H. Al-
Hazmi, V. O. Aliyev, F. M. Mosa (Linde AG/SABIC), WO 2009/
006979A2, 2009.
[23] S. Peitz, N. Peulecke, B. R. Aluri, S. Hansen, B. H. Mꢂller, A. Span-
nenberg, U. Rosenthal, M. H. Al-Hazmi, F. M. Mosa, A. Wçhl, W.
Mꢂller, Eur. J. Inorg. Chem. 2010, 1167–1171.
[6] J. R. Briggs, J. Chem. Soc. Chem. Commun. 1989, 674–675.
[7] a) R. Arteaga-Mꢂller, H. Tsurugi, T. Saito, M. Yanagawa, S. Oda, K.
Mashima, J. Am. Chem. Soc. 2009, 131, 5370–5371; b) A. Epshteyn,
E. F. Trunkely, D. A. Kissounko, J. C. Fettinger, L. R. Sita, Organo-
metallics 2009, 28, 2520–2526; c) T. Agapie, J. A. Labinger, J. E.
Bercaw, J. Am. Chem. Soc. 2007, 129, 14281–14295; d) T. Agapie,
S. J. Schofer, J. A. Labinger, J. E. Bercaw, J. Am. Chem. Soc. 2004,
126, 1304–1305; e) R. Emrich, O. Heinemann, P. W. Jolly, C.
Krꢂger, G. P. J. Verhovnik, Organometallics 1997, 16, 1511–1513.
[8] a) H. Hagen, Ind. Eng. Chem. Res. 2006, 45, 3544–3551; b) H.
Hagen, W. P. Kretschmer, F. R. van Buren, B. Hessen, D. A. van
Oeffelen, J. Mol. Catal. A 2006, 248, 237–247.
[9] M. Wang, Y. Shen, M. Qian, R. Li, R. He, J. Organomet. Chem.
2000, 599, 143–146.
[10] R. Santi, A. M. Romano, M. Grande, A. Sommazzi, F. Masi, A.
Proto (Enichem S.P.A.), WO 2001/68572, 2001.
[11] K. Iwakura, M. Yanagawa, S. Oda (Sumitomo Chemical Company),
WO 2005/044761, 2005.
[24] T. Agapie, M. W. Day, L. M. Henling, J. A. Labinger, J. E. Bercaw,
Organometallics 2006, 25, 2733–2742.
[25] D. Steinborn, Grundlagen der metallorganischen Komplexkatalyse,
Teubner/GWV Fachverlage, Wiesbaden, 2007, and references there-
in.
[12] a) T. Jiang, S. Zhang, X. Jiang, C. Yang, B. Niu, Y. Ning, J. Mol.
Catal. A 2008, 279, 90–93; b) K. Blann, A. Bollmann, H. de Bod,
J. T. Dixon, E. Killian, P. Nongodlwana, M. C. Maumela, H. Maume-
la, A. E. McConnell, D. H. Morgan, M. J. Overett, M. Prꢆtorius, S.
Kuhlmann, P. Wasserscheid, J. Catal. 2007, 249, 244–249; c) S. Kuhl-
mann, K. Blann, A. Bollmann, J. T. Dixon, E. Killian, M. C. Maume-
la, H. Maumela, D. H. Morgan, M. Prꢆtorius, N. Taccardi, P. Was-
serscheid, J. Catal. 2007, 245, 279–284; d) E. Killian, K. Blann, A.
Bollmann, J. T. Dixon, S. Kuhlmann, M. C. Maumela, H. Maumela,
D. H. Morgan, P. Nongodlwana, M. J. Overett, M. Pretorius, K. Hçf-
ener, P. Wasserscheid, J. Mol. Catal. A 2007, 270, 214–218; e) M. J.
Overett, K. Blann, A. Bollmann, J. T. Dixon, F. Hess, E. Killian, H.
Maumela, D. H. Morgan, A. Neveling, S. Otto, Chem. Commun.
2005, 622–624; f) A. Bollmann, K. Blann, J. T. Dixon, F. M. Hess, E.
Killian, H. Maumela, D. S. McGuinness, D. H. Morgan, A. Neveling,
S. Otto, M. Overett, A. M. Z. Slawin, P. Wasserscheid, S. Kuhlmann,
J. Am. Chem. Soc. 2004, 126, 14712–14713.
[13] a) D. S. McGuinness, D. B. Brown, R. P. Tooze, F. M. Hess, J. T.
Dixon, A. M. Z. Slawin, Organometallics 2006, 25, 3605–3610; b) K.
Blann, A. Bollmann, J. T. Dixon, F. M. Hess, E. Killian, H. Maume-
la, D. H. Morgan, A. Neveling, S. Otto, M. J. Overett, Chem.
Commun. 2005, 620–621; c) D. S. McGuinness, P. Wasserscheid,
D. H. Morgan, J. T. Dixon, Organometallics 2005, 24, 552–556;
d) D. S. McGuinness, P. Wasserscheid, W. Keim, C. Hu, U. Englert,
J. T. Dixon, C. Grove, Chem. Commun. 2003, 334–335; e) A. Carter,
S. A. Cohen, N. A. Cooley, A. Murphy, J. Scutt, D. F. Wass, Chem.
Commun. 2002, 858–859.
[26] P. J. W. Deckers, B. Hessen, J. H. Teuben, Organometallics 2002, 21,
5122–5135.
[27] H. Wesenfeld, K. H. Reichert, Chem. Ing. Tech. 1993, 65, 577–579.
[28] a) D. S. McGuinness, A. J. Rucklidge, R. P. Tooze, A. M. Z. Slawin,
Organometallics 2007, 26, 4696–4704; b) A. J. Rucklidge, D. S.
McGuinness, R. P. Tooze, A. M. Z. Slawin, J. D. A. Pelletier, M. J.
Hanton, P. B. Webb, Organometallics 2007, 26, 2782–2787; c) D. S.
McGuinness, M. Overett, R. P. Tooze, K. Blann, J. T. Dixon,
A. M. Z. Slawin, Organometallics 2007, 26, 1108–1111; d) T. Jiang,
X. Liu, Y. Ning, H. Chen, M. Luo, L. Wang, Z. Huang, Catal.
Commun. 2007, 8, 1145–1148; e) A. Jabri, C. Temple, P. Crewdson,
S. Gambarotta, I. Korobkov, R. Duchateau, J. Am. Chem. Soc. 2006,
128, 9238–9247.
[29] J. Zhang, A. Li, T. S. A. Hor, Organometallics 2009, 28, 2935–2937.
[30] a) A. Jabri, C. B. Mason, Y. Sim, S. Gambarotta, T. J. Burchell, R.
Duchateau, Angew. Chem. 2008, 120, 9863–9867; Angew. Chem. Int.
Ed. 2008, 47, 9717–9721; b) K. Albahily, D. Al-Baldawi, S. Gambar-
otta, R. Duchateau, E. KoÅ, T. J. Burchell, Organometallics 2008, 27,
5708–5711; c) K. Albahily, E. KoÅ, D. Al-Baldawi, D. Savard, S.
Gambarotta, T. J. Burchell, R. Duchateau, Angew. Chem. 2008, 120,
5900–5903; Angew. Chem. Int. Ed. 2008, 47, 5816–5819; d) K. Al-
bahily, D. Al-Baldawi, S. Gambarotta, E. KoÅ, R. Duchateau, Orga-
nometallics 2008, 27, 5943–5947; e) A. Jabri, P. Crewdson, S. Gam-
barotta, I. Korobkov, R. Duchateau, Organometallics 2006, 25, 715–
718.
Received: March 1, 2010
Published online: May 28, 2010
7842
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7833 – 7842