PCNCP ligands in the chromium-catalyzed oligomerization of ethylene: Triversus tetramerization

Article abstract of DOI:10.1002/chem.200900986

Chromium(III) complexes bearing R'N(CH₂PR₂) ₂ (PCNCP) ligands have been prepared. Upon activation with methylaluminoxane, these complexes proved to be effective in the selective tri- and tetramerization of ethylene. The fo

Full text of DOI:10.1002/chem.200900986

FULL PAPER
DOI: 10.1002/chem.200900986
PCNCP Ligands in the Chromium-Catalyzed Oligomerization of Ethylene:
Tri- versus Tetramerization**
Christian Klemps,[a, b] Elina Payet,[a] Lionel Magna,[b] Lucien Saussine,[b]
Xavier F. Le Goff,[a] and Pascal Le Floch*[a]
Abstract: ChromiumACHTNUTRGENN(UG III) complexes bearing R’NCAHTUNGTRNE(NGUN CH2PR2)2 (PCNCP) ligands
have been prepared. Upon activation with methylaluminoxane, these complexes
proved to be effective in the selective tri- and tetramerization of ethylene. The for-
mation of either 1-hexene or 1-octene was found to be highly dependent on the
steric bulk of the substituents R on the phosphine moieties. This observation was
rationalized by using density functional theory calculations on selected steps of the
metallacyclic mechanism of the ethylene oligomerization reaction.
Keywords:
density functional calculations
ethylene · homogeneous catalysis ·
oligomerization
chromium
·
·
Introduction
phorus, nitrogen, and sulfur donors. A chelating trisphos-
phine ligand, with the general formula R2P(CH2)2P(R’)-
AHCTUNGTRENNUNG
Selective oligomerization of ethylene has received growing
attention both in industrial and academic research due to
the paramount importance of short-chain linear a-olefins
such as 1-butene, 1-hexene, and 1-octene as comonomers in
the production of polyethylene (linear low-density polyethy-
lene, LLDPE).[1]
N
Due to their high activity and selectivity, chromium-based
catalytic systems take a predominant position notably in the
selective tri- and tetramerization of ethylene. Since the first
chromium-catalyzed selective ethylene transformation to 1-
hexene was reported by Manyik et al.,[2] a couple of highly
active and selective catalytic systems employing a chromium
source and a heteroatomic ligand have been disclosed.[3]
Besides the Phillips pyrrolide system,[4] which has found
large-scale industrial application, most other systems are
based on chelating bi- and tridentate ligands featuring phos-
Scheme 1. Chromium–bisphosphine-based catalyst precursors for ethyl-
ene tri- and tetramerization.
lective trimerization catalyst upon activation with various
aluminoxanes.[5] The asymmetric version with nꢀ3, resulting
in an increased chelate ring size was found to be superior in
terms of both activity and selectivity, which is in sharp con-
trast to later reports on the closely related Sasol P–N–P tri-
[a] Dipl.-Chem. C. Klemps, E. Payet, Dr. X. F. Le Goff,
Prof. Dr. P. Le Floch
merization system (B),[6] employing a (H)N
ACHTUNRGTNE[NUG (CH2)2PR2]2
symmetric ligand in its amine or amide form. The steric
bulk of the phosphine substituents R was found to have a
crucial role on the product selectivity of the catalyst. With
R=Cy, mainly polymerization was observed, whereas R=
Et led to overall 1-hexene selectivities of up to 97 wt%.
While conserving the overall ligand geometry, the Sasol S–
N–S trimerization system replaces the chelating phosphine
Laboratoire “Hꢀtꢀroꢀlꢀments et Coordination”
Ecole Polytechnique—CNRS
91128 Palaiseau cedex (France)
Fax : (+33)169-33-44-40
E-mail: lefloch@poly.polytechnique.fr
[b] Dipl.-Chem. C. Klemps, Dr. L. Magna, Dr. L. Saussine
IFP-Lyon, Rond-point de l’ꢀchangeur de Solaize
BP 3, 69360 Solaize (France)
donors by thioethers to yield (H)NACHTUNGTRNEUNG
[(CH2)2SR]2 ligands.[7] All
[**] PCNCP=bis(phosphinomethyl)amine (R’NACTHNUTRGNE(UNG CH2PR2)2).
three ligands feature a tridentate meridional coordination
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200900986.
mode towards CrIII metal centers.
Chem. Eur. J. 2009, 15, 8259 – 8268
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8259

This contrasts with the Sasol bisphosphinoamine (PNP)
catalytic system for ethylene tri- and tetramerization (C),[8]
and the BP system for ethylene trimerization (D),[9,10] which
equally employ a PNP bisphosphinoamine, bearing an addi-
tional hemilabile ether or thioether donor site, the coordina-
tion of which directs the selectivity towards trimerization.
Upon coordination of bisphosphinoamine ligands to CrIII
centers, a facial coordination mode of the ligand in the re-
sulting complexes is observed.[8,9]
new procedure adapted from a well-established route in the
literature. This involves the reaction of a secondary phos-
phine R2PH with formaldehyde followed by a Mannich-type
condensation of the resulting hydroxymethylphosphine
R2PCH2OH with
a primary amine R’NH2 (Scheme 2,
Table 1).[21,22] To assess the importance of substituents at
Carbon-bridged diphosphine ligands, like 1,2-(R2P)2-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
precursor (acac=acetylacetonate).[11] In addition to a poly-
mer fraction ranging from 21.3 to 65.6 wt%, an important
selectivity towards trimerization products was observed for
both the highly basic iPr-substituted ligand or the sterically
demanding diphenylphosphine derivative with R=o-Et-
(C6H4). On the contrary, the less bulky derivative with R=
Ph, exhibited a 1-octene selectivity of 57.1 wt% within the
liquid fraction of the catalytic run. The coordination com-
Scheme 2. General synthetic route to PCNCP ligands 1a–g.
Table 1. Substituents in 1a–g.
1
a
b
c
d
e
f
g
R
R’
Ph
iPr
Ph
tBu
Ph
Ph
Cy
iPr
Cy
tBu
Cy
Ph
nBu
iPr
plex of this ligand with [CrCl3ACHTNUTRGNENUG(thf)3] presents an octahedral
structure with facial coordination of the ligand.
A number of both experimental and theoretical studies
on the mechanism of the selective tri- and tetramerization
reaction have been reported recently. Whereas a metallacy-
clic pathway[12] has been established through deuterium la-
beling studies on both the above-mentioned BP[13–15] and the
Sasol PNP system,[16] and seems to have found general ac-
ceptance, the question of oxidation and spin state of the
active catalytic species upon activation with organoalumi-
num compounds still remains a subject of debate in some
cases.[17] Whereas a CrII–CrIV redox couple seems to be in-
volved in the Phillips pyrrolide system,[18] a recent study by
van Rensburg et al.,[19] which focuses on the importance of
the ion-pair model chosen for cationic chromium active in-
termediates in the Sasol PNP system, proposes a CrI–CrIII
redox couple. This same redox couple is postulated in a
computational investigation of the trimerization reaction
with cationic cyclopentadienyl–Cr complexes.[19]
phosphorus on the activity of the corresponding catalyst, re-
actions were carried out with bis(diphenylphosphine) (1a–
c), bis(cyclohexylphosphine) (1d–f), and bis(n-butylphos-
phine) (1g). Three different amines (isopropylamine, tert-
butylamine, and aniline) were employed in the case of
phenyl- and cyclohexyl-substituted ligands 1a–f. The n-
butyl-substituted ligand 1g, which proved to be much more
difficult to handle because of its air sensitivity, was only syn-
thesized with isopropylamine as substituent. All these li-
gands, which were obtained in fair to good yields, were char-
acterized by conventional NMR spectroscopy and MS tech-
niques.
The chromium
1a–g were easily prepared in almost quantitative yield by
addition of [CrCl3A(thf)3] to a solution of 1a–g in thf at room
ACHTUNGTRENN(UNG III) complexes 2a–g of the PCNCP ligands
CTHUNGTRENNUNG
Herein we report on the synthesis of bis(phosphinome-
thyl)amine (PCNCP) ligands, their coordination to [CrCl3-
ACHTUNGTRENNUNG(thf)3], and the application of the resulting complexes in the
temperature, the coordination being indicated by the imme-
diate color change from pink to dark blue. Due to the
strong paramagnetic nature of the octahedral d3 CrIII center,
complexes 2a–g are 31P NMR silent and their 1H and 13C
spectra only exhibited broad signals that could not be ex-
ploited to establish their structure. Therefore 2a–g were
only characterized by elemental data. Fortunately, crystals
suitable for X-ray structure determination could be grown
in the case of 2d by slow evaporation of a concentrated so-
lution of 2d in thf at room temperature. A view of one mol-
ecule of 2d is presented in Figure 1 and the most significant
metric parameters are listed in the corresponding legend. As
can be seen, complex 2d adopts a monomeric structure and
the overall geometry around chromium is octahedral.
chromium-catalyzed selective tri- and tetramerization of
ethylene. The influence of the steric bulk of the phosphine
moieties on the selectivity towards either 1-hexene or 1-
octene is discussed and thoroughly rationalized through
DFT calculations that compare the formation of 1-hexene
from chromacycloheptane and growth of this chromacyclo-
heptane complex with the formation of a chromacyclono-
nane complex through ethylene insertion.
Results and Discussion
Interestingly, the PCNCP ligand behaves as a bidentate
ligand and the central nitrogen donor atom does not partici-
Synthesis of ligands 1a–g and chromium complexes 2a–g:
Symmetric bis(phosphinomethyl)amines of general formula
(R2PCH2)2NR’ (PCNCP) 1a–g were prepared following a
À
pate in the coordination, as indicated by the very long Cr N
distance of 4.017(1) ꢂ. As a consequence, the coordination
8260
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 8259 – 8268

Chromium-Catalyzed Oligomerization of Ethylene
FULL PAPER
tectable polymeric material is formed in each experiment.
Equally, no cyclic C6 byproducts, as observed in the Sasol
PNP tetramerization system,[8] have been detected. Ligands
with phenyl substitution on phosphorous systematically ex-
hibit a lower activity than those with alkyl substitution (Cy
and nBu; entry 1–3 vs. 4–7). Concomitantly, their selectivity
towards trimerization is generally lower and oligomer distri-
butions are observed. On the other hand, cyclohexyl-substi-
tuted derivatives show higher activities alongside high selec-
tivities towards the formation of 1-hexene. Even though less
pronounced, the great steric bulk of the tBu group on the
central amine moiety seems to have a negative influence on
the activity (entry 5). The most important observation con-
cerns the effect of steric bulk at phosphorus atoms. Indeed,
it becomes obvious that decreasing the steric bulk by replac-
ing the cyclohexyl groups (2d–f) by the smaller nBu group
(2g) results in a shift of the selectivity towards tetrameriza-
tion, while a selectivity is maintained. Finally, it clearly ap-
pears that the most satisfactory results are obtained with
complex 2 f, which shows an excellent selectivity towards tri-
merization (99%). Therefore, further experiments were car-
ried out to optimize its performance. It was found that a
temperature increase from 45 to 708C was beneficial for cat-
alyst productivity (entry 6 vs. 8). However an increase in
ethylene pressure from 30 to 50 bar did not improve the
productivity. Finally we found that improvement of the cata-
lyst activity was achieved upon doubling of the MAO quan-
tity and carrying out the oligomerization at 708C (entry 10).
Figure 1. Molecular structure of one molecule of 2d in the crystal. Ther-
mal ellipsoids are drawn at the 30% probability level (hydrogen atoms
À
are omitted for clarity). Selected bond lengths [ꢂ] and angles [8]: P1 Cr1
À
À
À
À
2.518(2), P1 C1 1.836(5), C1 N1 1.457(6), N1 C2 1.465(7), C2 P2
À
À
À
À
1.863(5), P2 Cr1 2.466(2), Cr1 Cl1 2.320(2), Cr1 Cl2 2.314(2), Cr1 Cl3
À
2.311(2), Cr1 O1 2.104(3); P1-C1-N1 116.3(4), C1-N1-C2 112.8(4), N1-
C2-P2 116.1(4), C2-P2-Cr1 113.6(2), P2-Cr1-P1 85.90(5), P2-Cr1-Cl1
86.94(6), P2-Cr1-Cl3 90.16(6), P2-Cr1-Cl2 92.82(6), P2-Cr1-O1 177.4(1),
P1-Cr1-Cl1 85.53(6), P1-Cr1-Cl2 89.72(6), Cl3-Cr1-Cl1 91.44(6).
sphere of chromium is completed by coordination of a thf
molecule that is located trans to one phosphorus atom (P2).
Note that this complex is not the first example of a biden-
tate phosphino-[CrCl3ACTHUNRGTNE(NUG thf)] complex. Recently, Overett
et al. reported on the structure of a similar complex featur-
ing 1,2-bis(diphenylphosphino)benzene as ligand.[11] Metric
parameters of complex 2d are very close to those of this
complex except for the P-Cr-P angle, which is larger in the
case of 2d (85.90(5) vs. 77.82(2)8). Apart from this, the
structure of complex 2d does not deserve further comment
and all bond lengths and angles fall in the usual range.
Theoretical calculations: Theoretical calculations were car-
ried out to shed light on the influence of steric effects. We
deliberately focused our study on the formation of chroma-
cyclononane versus 1-hexene formation. Indeed, according
to the postulated metallacycle mechanism, formation of 1-
hexene and 1-octene results from the transient formation of
a chromacycloheptane ring that can either undergo a 3,7-hy-
drogen transfer or a b-hydrogen transfer followed by a re-
ductive elimination (to afford 1-hexene), or insert an addi-
tional molecule of ethylene to form a chromacyclononane.[23]
This large metallacycle, in turn, would yield 1-octene
through the same reductive processes. These different cata-
lytic steps are presented in Scheme 3.
Ethylene oligomerization: The catalytic activity of 2a–g in
ethylene oligomerization was then evaluated. Experiments
were carried out at 45 or 708C in the presence of methylalu-
minoxane (MAO) with a MAO/Cr ratio of 300:1 or 600:1
and toluene was used as solvent. Catalytic performances are
summarized in Table 2. Importantly, we noted that no de-
All our calculations were realized at the quantum level
using the Gaussian 03[24] set of
programs in combination with
Table 2. Ethylene oligomerization with complexes 2a–g.[a]
the PBEPBE functional[25] (see
[b]
Catalyst
T [8C]
%C
N
%C
G
%C
(1-C8)
%C10+
Productivity [gg(Cr)À1 hÀ1
]
the Experimental Section for
further details regarding opti-
mization of geometries and re-
search of transition states).
Choice of the PBE functional
was justified by the fact that it
usually gives the most satisfac-
tory results in not overestimat-
ing the stability of the low-spin
forms. Indeed, it must be kept
in mind that CrIII complexes
1
2
3
4
5
6
7
8
2a
2b
2c
2d
2e
2 f
2g
2 f
2 f
2 f
45
45
45
45
45
45
45
70
45
70
6(51)
4(59)
3(57)
1(35)
1(20)
–
–
–
–
–
88(98)
67(91)
90(98)
97(95)
90(92)
99(97)
78(99)
99(98)
98(98)
99(98)
4(81)
20(85)
4(87)
1(99)
2(99)
1(99)
21(98)
1(99)
1(99)
1(99)
2
9
3
1
7
–
1
–
1
–
2106
527
2335
4579
4003
5300
5127
8535
4892
9783
9[c]
10[d]
[a] Conditions: 30 bar ethylene pressure, 30 min, MAO (300 equiv). [b] No polymeric material was discovered.
[c] Conditions: 50 bar ethylene pressure, 30 min. [d] 600 equiv of MAO used.
Chem. Eur. J. 2009, 15, 8259 – 8268
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8261

P. Le Floch et al.
other conformations of the metallacycle, not featuring this
b-H agostic interaction, were computed but were found to
be slightly higher in energy on the potential energy surface
(PES). A view of the chromacyclopentane IaEt is presented
in Figure 2 and the most significant bond lengths and angles
are listed in the corresponding legend.
Scheme 3. Possible evolution pathways of a chromacycloheptane com-
plex.
(s1d2) can adopt two electronic configurations, quartet (S=
3
= ) or doublet (S=1= ). Throughout this study CrIII com-
2
2
plexes were also found to be more stable in the quartet elec-
tronic configuration. Note that experimental data reported
in the literature (measure of magnetic susceptibility)[14,26]
confirmed that a quartet spin state is often encountered as
the ground state for CrIII complexes.[18,27] As will be seen
later, CrI complexes (d5 complexes), which can adopt either
3
5
low-spin or high-spin configurations (S=1= or = or = ),
2
2
2
were also found to be more stable in the quartet spin state.
Four approximations were done to reduce computing time
due to the size of the considered systems. First, as we were
exclusively interested in assessing steric effects at phospho-
rus, the nitrogen substituent was systematically replaced by
a methyl group. Second, the n-butyl groups were replaced
by ethyl groups assuming that the ending ethyl moiety of
this n-butyl group on phosphorous is situated far away from
the metal during the catalytic process. Third, substituents at
phosphorus (either ethyl or cyclohexyl) were described
using the 3-21G* basis set (6-31G* for all other nonmetal
atoms) to minimize the number of basis functions. Fourth,
the role of MAO was not considered.[28] This approximation
is corroborated by recent calculations on real systems de-
rived from the Sasol PNP system, which have shown that
dissociated ion-pair complexes are probably the favored spe-
cies in the reaction medium when additional ethylene mole-
cules are introduced as ligands on cationic tetrasubstituted
chromacyclic derivatives.[19] These calculations also revealed
the presence of dissociated ion pairs to be a prerequisite for
the reaction to proceed.
Figure 2. Optimized geometry of theoretical complex IaEt. Most signifi-
À
À
À
cant bond lengths [ꢂ] and angles [8]: P1 Cr 2.505, P2 Cr 2.411, Cr C1
À
À
À
À
2.027, Cr C6 2.010, C6 C5 1.516, C5 H5 1.171, H5 Cr 1.955; P1-Cr-P2
90.630, Cr-C6-C5 83.523, C6-C5-H5 111.390, C5-H5-Cr 95.676.
Coordination of one additional molecule of ethylene on
IaR was then studied. Two possibilities were considered.
Indeed, one may consider that the ethylene molecule can
attach to chromium on the remaining apical position to
form a distorted octahedral complex, or by coordination on
the equatorial site occupied by the b-agostic interaction to
form a distorted square-pyramidal complex. Both structures
IbEt and IbEt’ were found as minima in the case of the ethyl
derivative, but optimizations of the cyclohexyl complex only
led to the distorted square-pyramidal complex IbCy. At-
tempts to minimize a structure analogous to IbEt’ in the case
of the cyclohexyl derivative only yielded a structure featur-
ing no interaction between IaCy and the incoming molecule
of ethylene (Scheme 4). As two components are involved in
this transformation, Gibbs free energy (DG) values were
preferred over internal energy (DE) values, though it is well
known that entropy is always slightly overestimated in gas-
phase calculations.
The energy required to coordinate one molecule of ethyl-
ene to IaCy to yield IbCy (DG= +18.9 kcalmolÀ1) was found
to be much more important than the one needed to form
IbEt and IbEt’ (DG= +9.9 and +9.7 kcalmolÀ1, respectively).
This observation very likely reflects the stronger steric hin-
drance provided by the cyclohexyl groups in IbCy. Notably,
in all complexes, bonding of the ethylene molecule to the
cationic center is probably very weak as evidenced by the
Several chromacycloheptane rings adopting different geo-
metries and conformations were computed in the case of
both ethyl and cyclohexyl substituents. Indeed, it is impor-
tant to keep in mind that, at least two geometries should be
accessible in the case of the chromacycloheptane complex.
A first one, in which the bidentate ligand and the ring meth-
ylene substituents lie in the same plane (yielding a distorted
square-planar geometry), and a second one, in which the
complex adopts a distorted tetrahedral geometry. Despite
many attempts to optimize a square-planar structure, all our
calculations converged towards the distorted tetrahedral
structures IaR (R=Et, Cy), in which one of the metallacycle
a-carbon atoms occupies the apical position. Notably, both
À
structures feature a b-H agostic interaction (dACTHNUGTRNEUNG(Cr H)=
À
1.955 ꢂ in IaEt and d
H atom occupies the remaining equatorial position. Several
ACHTUNGTRENN(NUG Cr H)=1.971 ꢂ in IaCy) in which the
8262
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 8259 – 8268

Chromium-Catalyzed Oligomerization of Ethylene
FULL PAPER
quadruplet spin state (the S=2 state lying 16.5 and 16.6 kcal
molÀ1 above for IcEt and IcCy, respectively). A view of the
transition state IaEt-TS-IcEt is presented in Figure 3 and the
Scheme 4. Bonding of ethylene to structures IaEt and IaCy. All values are
given in kcalmolÀ1
.
À
long Cr CCcentroid bond lengths (2.396 ꢂ in IbEt; 2.567 ꢂ in
IbEt’; 2.395 ꢂ in IbCy) and to the relatively short C=C bond
lengths (1.360 ꢂ in IbEt; 1.355 ꢂ in IbEt’; 1.359 ꢂ in IbCy).
These data are indicative of weak p back bonding from the
chromium center to the p* system of the ethylene ligand.
We then focused our study on the two main transforma-
tions that either lead to 1-hexene or to the chromacyclono-
nane complexes. Two different processes were envisaged for
the formation of 1-hexene: the intramolecular 3,7-hydrogen
transfer and the transient formation of an hydrido-hexenyl
complex through a b-H transfer from carbon to chromium.
No such hydrido complexes could be successfully optimized,
each attempt yielding back the starting precursors IaEt and
IaCy. In each case, the 3,7-hydrogen transfer process proved
to be the only pathway to form the chromium(I) (h2-1-
hexene) complexes IcR (R=Et, Cy). These findings are con-
sistent with a recent experimental study by Bercaw et al.[15]
Thus the formation of IcEt was found to be only slightly en-
dothermic (DG= +0.2 kcalmolÀ1), involving an energetic
barrier of DG° = +13.3 kcalmolÀ1 (Scheme 5). Note that
both complexes IcEt and IcCy were found as minima in the
Figure 3. Optimized geometry of the transition structures IaEt-TS-IcEt
(upper) and IbEt-TS-IdEt (lower). Most significant distances [ꢂ] and
À
À
À
angles [8]: For (IaEt-TS-IcEt): P1 Cr 2.486, P2 Cr 2.416, Cr C1 2.219,
À
À
À
À
À
Cr C6 2.087, C6 C5 1.452, C5 H5 1.443, H5 Cr 1.654, H5 C1 1.482;
P1-Cr-P2 82.468, Cr-C6-C5 74.511, C6-C5-H5 114.532, C5-H5-Cr 90.320,
À
À
À
C1-H5-Cr 89.844. For (IaEt-TS-IcEt): P1 Cr 2.496, P2 Cr 2.527, Cr C1
À
À
À
À
2.039, Cr C6 2.195, Cr C8 2.121, C6 C7 2.168, C7 C8 1.450; P1-Cr-P2
82.988, P1-Cr-C8 82.321, Cr-C8-C7 78.094, C8-C7-C6 120.786, C6-C6-Cr
63.504.
most significant distances and bond angles are listed in the
corresponding legend. As can be seen, if one does not take
À
into account the C1 Cr bond, which is significantly elongat-
ed, this transition complex adopts a distorted square-planar
pyramidal structure, with only P1, P2, H5, and C6 being co-
À
ordinated. As expected, the C5 H (1.443 ꢂ) bond is elon-
À
gated and close to the C1 H5 (1.482 ꢂ) bond that is being
formed. Concomitantly, the C5–C6 (future double bond of
1-hexene) is shortened to 1.452 ꢂ. Most importantly, it ap-
pears that this hydrogen transfer is strongly assisted by chro-
Scheme 5. Overall energetic pathway showing the transformation of IaEt
into the 1-hexene complex IcEt and the chromacyclononane IdEt. All
À
mium as attested by the relatively short Cr H5 distance of
1.482 ꢂ.
values are given in kcalmolÀ1
.
Chem. Eur. J. 2009, 15, 8259 – 8268
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8263

P. Le Floch et al.
The ring expansion that leads to chromacyclononane IdEt
was also computed and the general energetic pathway is de-
picted in Scheme 5. This transformation was found to be
slightly exothermic (DG=À7.4 kcalmolÀ1 versus IaEt and
À17.3 kcalmolÀ1 versus IbEt) and involves an activation bar-
rier of DG= +17.2 kcalmolÀ1 from IaEt and +7.3 kcalmolÀ1
from IbEt, respectively. A view of the transition state con-
necting IbEt and IdEt is presented in Figure 3 as well as the
most significant metric parameters.
This ring expansion simply consists of a 1,2-insertion pro-
cess by means of the C6 carbon atom of the ring attacking
the C7 carbon atom of the coordinated ethylene moiety. Not
surprisingly, if we take into account the high exothermicity
of the process, the overall geometry around chromium is
closer to that of the starting precursor IbEt than to that of
important information is given by the comparison of the two
processes in the case of the cyclohexyl substituent. Contrary
to what is observed with the ethyl derivative, formation of
1-hexene is favored (DG° = ++10.9 kcalmolÀ1) over the for-
mation of complex IbCy, which is the key intermediate to
allow the growth of the metallacycle (DG° = +25.5 kcal
molÀ1 with IaCy as reference). Examination of Scheme 6 led
us to conclude that formation of the chromacyclononane is
strongly disfavored when cyclohexyl substituents are pres-
ent. Note that these conclusions are in good agreement with
experimental results that show that only 1-hexene is formed.
To complete this study, we then envisaged the existence
of a second type of pentacoordinated ethylene complex IIbR
(R=Et, Cy), which could be formed by coordinating one
molecule of ethylene on the apical position of the unsaturat-
ed square-pyramidal derivative IIaR, though this latter unsa-
turated complex could not be optimized in the case of our
PCNCP ligands (see above). Optimizations of the structures
IIbR were successful and both complexes were found to be
energetically close to isomers IbR (+13.5 kcalmolÀ1 for IIbEt
versus +9.9 and +9.7 kcalmolÀ1 for IbEt and IbEt’, respec-
tively, and +17.0 kcalmolÀ1 for IIbCy versus +18.9 kcalmolÀ1
for IbCy; Scheme 7). A view of one molecule of complex
IIbCy is shown in Figure 4. Note that an analogue of this un-
À
IdEt. The most significant information are given by the C6
À
C7 (2.168 ꢂ) and Cr C8 (2.121 ꢂ) bonds, which are still
À
rather long and the C7 C8 (1.419 ꢂ) bond, which is slightly
lengthened. Note that to complete this first part of the
study, additional calculations were done postulating the
hexaACHTUNGTRENNUNGcoordinated IbEt’ complex as starting precursor for the
synthesis of a chromacyclononane, but all attempts proved
unsuccessful. We conclude that both pathways starting from
IaEt, which lead to IcEt and IdEt, require an activation energy
of 13.3 and 17.2 kcalmolÀ1, respectively, and therefore are
both kinetically accessible.
The same approach was followed in the case of the cyclo-
hexyl derivatives IaCy and IbCy and the complete energetic
pathways leading to the 1-hexene IcCy and chromocyclono-
nane IdCy complexes were computed. In each case, transition
states IaCy-TS-IcCy and IbCy-TS-IdCy were localized and were
found to be geometrically close to that computed in the case
of the ethyl derivatives. These two energetic pathways are
presented in Scheme 6. Two conclusions can be drawn by
comparing Schemes 5 and 6. It appears that the kinetic for-
mation of 1-hexene is slightly favored in the case of the cy-
clohexyl substituent compared with the ethyl derivative
(DG° = +10.9 vs. +13.3 kcalmolÀ1), whereas the formation
of the chromacyclononane is comparatively slightly exother-
mic (DG=À1.0 versus À7.4 kcalmolÀ1). However, the most
Scheme 7. Relative energies (in kcalmolÀ1) of IIbEt and IIbCy
.
Figure 4. Optimized geometry of complex IIbCy. Cyclohexyl substituents
are drawn in wireframe for the sake of clarity. Most significant bond
À
À
À
À
Scheme 6. Overall energetic pathway showing the transformation of IaCy
into the 1-hexene complex IcCy and the chromacyclononane IdCy. All
lengths [ꢂ] and angles [8]: P1 Cr 2.764, P2 Cr 2.593, Cr C1 2.089, Cr
À
À
À
À
C6 2.0675, C5 H5 1.164, H5 Cr 1.946, Cr C7 2.284, Cr C8 2.261; P1-Cr-
P2 87.766, Cr-C6-C5 82.393, C6-C5-H5 112.496, C5-H5-Cr, 97.789.
values are given in kcalmolÀ1
.
8264
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 8259 – 8268

Chromium-Catalyzed Oligomerization of Ethylene
FULL PAPER
saturated square-pyramidal complex was found to be a mini-
mum in the study on strained PNP ligands reported by van
Rensburg et al.[19]
Though we have no experimental evidence about the for-
mation of these ethylene complexes, two hypotheses may ac-
count for their formation. First, one may propose that one
MAO counteranion interacts with this unsaturated complex
and forces it to adopt a distorted square-pyramidal geome-
try, with an incoming ethylene molecule then substituting
this MAO anion in a second step (Scheme 8). A second hy-
Interestingly, the creation of this vacant site only requires
a small amount of energy (ca. 10.6 kcalmolÀ1). Finally the
last step of this mechanism involves the recoordination of
the pendant phosphine moiety to chromium trans to the
apical alkyl substituent to yield IIbR. However, these path-
ways involve important free energies higher than
22 kcalmolÀ1 and are probably disfavored, in both cases,
compared with the formation of 1-hexene and chromacyclo-
nonanes complexes (Schemes 5 and 6).
Nevertheless, to be exhaustive, we considered complexes
IIbR (R=Et, Cy) as possible intermediates in the ring-ex-
pansion process to chromacyclononanes. The two corre-
sponding structures IIdEt and IIdCy were successfully opti-
mized as well as the two corresponding transition states
IIbEt-TS-IIdEt and IIbCy-TS-IIdCy that connect them to the
two ethylene complexes IIbR. The energetic pathways of
these two transformations are depicted in Scheme 10. In the
Scheme 8. Possible precursors for the formation of IIbEt and IIbCy
.
pothesis would be that complexes IaR may rearrange to
yield complexes IIbR. These complexes are obviously ener-
getically close to the IbR isomers and could be relevant in-
termediates in the mechanism. It was therefore important to
verify whether IIbR could indeed be formed from the penta-
coordinated chromacycloheptane IaR.
From pursuing our investigations we found a plausible
mechanism that may possibly account for transformation of
complexes IaR into complexes IIbR (R=Et, Cy). This mech-
anism, which is depicted in the Scheme 9, involves, as a first
Scheme 10. Overall energetic pathways explaining the formation of chro-
macyclononanes IIdR from the ethylene complexes IIbR (R=Et, Cy). All
values are given in kcalmolÀ1
.
case of the ethyl substituent, formation of the chromacyclo-
nonane is competitive with the elimination leading to 1-
hexene (DG° between IaEt and IIbEt-TS-IIdEt =+16.0 kcal
molÀ1 versus DG° between IaEt and IaEt-TS-IcEt =+13.3 kcal
molÀ1) and compares well with the energetic barrier needed
to form chromacyclononane IdEt from IaEt (DG° between
IaEt and IbEt-TS-IdEt =+17.2 kcalmolÀ1). However, when cy-
clohexyl substituents are present at phosphorus, the forma-
tion of the chromacyclononane is disfavored with respect to
the formation of the 1-hexene complex. Indeed, +24.7 kcal
molÀ1 are needed to reach the transition state IIbCy-TS-IIdCy,
whereas formation of 1-hexene requires an activation
energy of only 10.9 kcalmolÀ1. Here we also note that the
formation of chromacyclononanes IdCy and IIdCy roughly re-
quire the same activation barriers (+24.7 kcalmolÀ1 for
isomer II versus +25.5 kcalmolÀ1 for isomer I).
Scheme 9. Overall energetic pathway showing the possible formation of
IIbR (R=Et, Cy) ethylene complexes from the pentacoordinated IaR
chromacycloheptanes. All values are given in kcalmolÀ1
.
step, the decoordination of the ligandꢃs phosphine moiety lo-
cated trans to the b-agostic hydrogen and, in a subsequent
second step, the coordination of the incoming ethylene mol-
ecule on the vacant site created.
These last series of results led us to conclude that isomers
IIbR, if they are formed, can lead to 1-hexene complexes
through a 3,7-elimination and the corresponding chromacy-
clononanes through the insertion of the ethylene molecule.
Chem. Eur. J. 2009, 15, 8259 – 8268
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8265

P. Le Floch et al.
However, as stated in the case of isomers IbR, formation of
chromacyclononanes is strongly disfavored when cyclohexyl
groups are present. Even though no additional detailed stud-
ies were undertaken in the case of the two isomers consid-
ered, one may reasonably propose that this difference does
not stem from electronic effects since both substituents are
alkyl groups, but rather results from steric effects that limit
the access of incoming ethylene molecules to the cationic
chromium center.
Experimental Section
Computational details: Calculations were performed with the Gaussi-
an 03 set of programs.[24] Geometries were optimized using the Perdew–
Burke–Ernzerhof (PBE) functional.[25] All the atoms were computed at
the quantum level of theory. The standard 6-31G* basis set was used for
all nonmetallic atoms (H, C, N, and P) belonging to the cyclic part of the
ligand and chromacycles, whereas the 3-21G* basis set was used for de-
scribing ethyl and cyclohexyl substituents at phosphorus atoms. The basis
set used for chromium is the Hay and Wadt[29] small-core quasi-relativis-
tic effective core potential with the double-z valence basis set (441s/
2111p/41d) augmented with an f-polarization function (exponent=
1.941).[30] All optimized geometries were characterized by complete fre-
quency calculations at the same level of calculations (no imaginary fre-
quencies), and single-point calculations were performed on all optimized
geometries at the same level of theory. Transition states were determined
by using the synchronous transit-guided quasi-Newton (STQN) method
and characterized by one imaginary frequency.
Conclusion
In this contribution, new chromiumACHTUNTRGNEUNG(III) complexes bearing
bidentate bis(phosphinomethyl)amine (PCNCP) ligands
have been prepared. Monomeric octahedral complexes were
obtained, in which the bisphosphine ligand adopts a facial
coordination mode. Those complexes were evaluated for
their activity in the selective ethylene oligomerization reac-
tion upon activation with excess MAO. Activities of up to
9783 g(oligomers)g(Cr)À1 hÀ1 were observed, with concomi-
tant selectivities of up to 99% towards 1-hexene. The selec-
tivity towards the ethylene trimerization product was found
to be strongly dependent on the steric bulk on the phos-
phine moieties, and, to a lesser extent, on the basicity of the
phosphinyl groups, and the substitution on the central nitro-
gen of the ligands. Bulky dicyclohexylphosphinyl groups led
to high selectivities towards 1-hexene, whereas less bulky di-
n-butylphosphinyl substituents led to the appearance of
21% of 1-octene during catalysis.
The relation between steric bulk on phosphorous and the
oligomer product distribution was studied through DFT cal-
culations concentrating on the key steps of the metallacyclic
mechanism, which are decisive for the selectivity towards
either 1-hexene or 1-octene. Results of these theoretical in-
vestigations reveal first that both CrI and CrIII complexes
were found to adopt the quartet spin state as ground states,
while other spin configurations proved to be energetically
less stable. The second piece of important information is re-
lated to the formation 1-hexene. Indeed, the results ob-
tained clearly show that in the cases of these PCNCP-based
systems, formation of CrI–(1-hexene) complexes results
from an intramolecular 3,7-hydrogen transfer and not from
a b-hydrogen transfer followed by a reductive elimination,
with hydrido–CrIII–(1-hexenyl) complexes being found un-
stable. Finally, the third important conclusion concerns the
selectivities observed. Our calculations clearly demonstrate
that, though the two substituents considered (ethyl and cy-
clohexyl) are probably not very different from an electronic
point of view, formation of chromacyclononane complexes is
strongly disfavored when cyclohexyl groups are present at
phosphorus, relative to the formation of CrI–(1-hexene)
complexes. This result, which is in very good agreement
with experimental data, reveals that steric bulk at phospho-
rus is also an important factor that can probably be further
exploited to finely tune selectivities in this important cata-
lytic transformation.
Synthesis: All reactions were routinely performed under an inert atmos-
phere of argon or nitrogen using Schlenk and glove-box techniques and
dry deoxygenated solvents. Nuclear magnetic resonance spectra were re-
corded on a Bruker AC-300 SY spectrometer operating at 300.0 MHz for
1H, 75.5 MHz for 13C, and 121.5 MHz for 31P. Solvent peaks were used as
the internal reference relative to Me4Si for 1H and 13C chemical shifts
(ppm); 31P chemical shifts are relative to a 85% H3PO4 external refer-
ence. In 13C NMR spectroscopy, only coupling to 31P nuclei are reported.
The following abbreviations are used: s, singlet; d, doublet; t, triplet; m,
multiplet. Mass spectra were obtained with a HP 5989B spectrometer
using ammonia as reagent gas. Elemental analyses were performed by
the “Service d’analyse du CNRS”, at Gif sur Yvette, France. Compounds
3ACTHNUTRGNEUNG
1b,[31] 1c,[21] nBu2PH,[32] and [CrCl (thf)3][33] were prepared as described in
the literature; all other reagents were obtained from commercial sources.
General procedure for the preparation of 1a–g: Solid paraformaldehyde
(74 mg, 2.47 mmol) was added to an equimolar quantity of secondary
phosphine R2PH, and the resulting mixture was heated to 1108C over
1 h.[22] Subsequently, toluene (5 mL) was added to the mixture, followed
by the corresponding primary amine R’NH2 (1.24 mmol). This reaction
mixture was heated to 608C over 2 h. The solvent was removed in vacuo,
typically resulting in a colorless viscous oil, which was triturated with a
mixture of methanol and petroleum ether, upon which white powdery
solids were obtained.
Compound 1a: Yield: 463 mg (1.02 mmol, 82%); 31P{1H} NMR
([D6]benzene, 208C): d=À26.5 ppm (s; P); 1H NMR ([D6]benzene,
3
2
208C): d=0.80 (d, J
(H,H)=6.6 Hz, 6H; NCH
N
ACHTUNGTRENNUNG(H,P)=
3
3.8 Hz, 4H; NCH2P), 3.78 (sept, J
ACHUTGTNRENN(GU H,H)=6.6 Hz, 1H; NCHACTHUNGTRENNNUG
(m, 12H; m,p-CH (PPh2)), 7.54 ppm (dd, 3J
U
ACHTUNGTRENNUNG
A
A
U
ACHTUNGTRENNUNG
5.7 Hz; NCH2P), 128.7 (s; p-CH (PPh2)), 128.9 (d, 3J
ACHTUNGTRENNUNG
m-CH (PPh2)), 132.8 (d, 2J
G
G
1
(d, J
(C,P)=87.5 Hz; ipso-C
R
Compound 1d: Yield: 381 mg (0.794 mmol, 64%); 31P{1H} NMR
([D6]benzene, 208C): d=À14.9 ppm (s; P); 1H NMR ([D6]benzene,
208C): d=0.99 (d, 3J
ACTHNUGTRENN(UNG H,H)=6.7 Hz, 6H; NCHACTHUNGTRENNUNG
44H; CH2 (Cy)), 3.89 (s, 4H; NCH2P), 4.09 ppm (m, 1H; NCH
13C NMR ([D6]benzene, 208C): d=16.8 (s; NCH
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
Compound 1e: Yield: 429 mg (0.868 mmol, 70%); 31P{1H} NMR
([D6]benzene, 208C): d=À14.7 ppm (s; P); 1H NMR ([D6]benzene,
208C): d=0.82 (s, 9H; NCACTHNUTRGNEUN(G CH3)3), 1.09–1.70 (m, 44H; CH2 (Cy)),
3.88 ppm (s, 4H; NCH2P); 13C NMR ([D6]benzene, 208C): d=14.2 (s;
NC (P,C)=15.8 Hz,
(CH3)3), 27.0, 27.6, 30.2, 33.2 (Cy), 47.2 (dd, 1J
3J(P,C)=11.7 Hz; NCH2P), 47.5 ppm (t, 3J
(P,C)=10.1 Hz; NC(CH3)3);
MS (CI): m/z: 494 [M+H]+.
T
ACHTUNGTRENNUNG
G
T
ACHTUNGTRENNUNG
Compound 1 f: Yield: 433 mg (0.843 mmol, 68%); 31P{1H} NMR
([D6]benzene, 208C): d=À15.0 ppm (s; P); 1H NMR ([D6]benzene,
8266
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 8259 – 8268

Chromium-Catalyzed Oligomerization of Ethylene
FULL PAPER
208C): d=1.14–1.83 (m, 44H; CH2 (Cy)), 3.93 (s, 4H; NCH2P), 6.77 (t,
Acknowledgements
3
3J
A
E
(NPh)), 7.24 ppm (pseudo-t; m-CH (NPh)); 13C NMR ([D6]benzene,
The authors thank the CNRS (Centre National de la Recherche Scientifi-
que), the Ecole Polytechnique, and the IFP-Lyon for financial support of
this work and IDRIS for the allowance of computer time (project
no. 060616).
208C): d=26.9, 27.7, 30.3, 33.3 (Cy), 47.9 (dd, JACTHUNTGRENNUG(P,C)=16.1 Hz, JACHTUNGTRNE(NUGN P,C)=
11.8 Hz; PCH2N), 118.2 (s; o-CH (NPh)), 118.8 (s; p-CH (NPh),
129.2 ppm (s; m-CH (NPh)); MS (CI): m/z: 514 [M+H]+.
Compound 1g: Yield: 247 mg (0.657 mmol, 53%); 31P{1H} NMR
([D6]benzene, 208C): d=À48.4 ppm (s; P); 1H NMR ([D6]benzene,
208C): d=0.90 (t, 3J
ACHTUNGTRENNUNG(H,H)=6.9 Hz, 12H; PACTHUNGTRENNNUG
[1] D. Vogt in Applied Homogeneous Catalysis with Organometallic
Compounds, Vol. 1 (Eds.: B. Cornils, W. A. Herrmann), Wiley-VCH,
New York, 2002, p. 240.
[2] R. M. Manyik, W. E. Walker, T. P. Wilson, J. Catal. 1977, 47, 197–
209.
[3] a) J. T. Dixon, M. J. Green, F. M. Hess, D. H. Morgan, J. Organomet.
Chem. 2004, 689, 3641–3668, and references therein; b) D. F. Wass,
Dalton Trans. 2007, 816–819.
[4] a) W. K. Reagan (Phillips Petroleum Company), EP 0417477, 1991 ;
b) W. K. Reagan, T. M. Pettijohn, J. W. Freeman (Phillips Petroleum
Company), US Patent 5523507, 1996; c) J. W. Freeman, J. L. Buster,
R. D. Knudsen (Phillips Petroleum Company), US Patent 5856257,
1999.
[5] F. J. Wu (Amoco Corporation), US Patent 5811618, 1998.
[6] a) J. T. Dixon, J. J. C. Groove, P. Wasserscheid, D. S. McGuinness,
F. M. Hess, H. Maumela, D. H. Morgan, A. Bollmann (Sasol Tech-
nology Pty.), WO03053891, 2001; b) D. S. McGuinness, P. Wassersc-
heid, W. Keim, J. T. Dixon, J. J. C. Groove, C. Hu, U. Englert, Chem.
Commun. 2003, 334–335.
3J(H,H)=6.4 Hz, 6H; NCH
3.72 (s, 4H; PCH2N), 4.18 ppm (m, 1H; NCH
([D6]benzene, 208C): d=15.4 (s; NCH
ACHTUGTNRENGUN(CH3)2), 1.24–1.69 (m, 24H; PCAHTUNGTRENNNUG
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
PCH2CH2CH2), 25.9 (d, 2J
ACHTNUGTRENUNG(P,C)=3.8 Hz; PCH2CH2), 33.4 ppm (d,
3J(P,C)=17.0 Hz; PCH2); MS (CI): m/z: 410 [M+H]+.
General procedure for the preparation of 2a–g: To a stirred suspension
of [CrCl3ACHTUNGTRENNUNG(thf)3] (225 mg, 0.601 mmol) in thf (5 mL), an equimolar quanti-
ty of ligand 1 was added, which resulted in an immediate color change of
the reaction mixture to dark blue. After stirring for 2 h, the solvent was
removed in vacuo and the resulting blue powder was washed with petro-
leum ether (2ꢄ5 mL), followed by further drying in vacuo.
Compound 2a: Yield: 391 mg (0.570 mmol, 95%) of a dark blue powder;
elemental analysis calcd (%) for C33H39Cl3CrNOP2: C 57.78, H 5.73, N
2.04; found: C 57.98, H 5.21, N 1.95.
Compound 2b: Yield: 408 mg (0.583 mmol, 97%) of a dark blue powder;
elemental analysis calcd (%) for C34H41Cl3CrNOP2: C 58.34, H 5.90, N
2.00; found: C 57.91, H 5.49, N 2.03.
[7] a) D. S. McGuinness, P. Wasserscheid, W. Keim, D. H. Morgan, J. T.
Dixon, A. Bollmann, H. Maumela, F. M. Hess, U. Englert, J. Am.
Chem. Soc. 2003, 125, 5272–5273; b) D. S. McGuinness, P. Wasser-
Compound 2c: Yield: 402 mg (0.559 mmol, 93%) of a dark blue powder;
elemental analysis calcd (%) for C36H37Cl3CrNOP2: C 60.05, H 5.18, N
1.95; found: C 59.89, H 5.02, N 2.05.
AHCTUNGTREGsNNUN cheid, D. H. Morgan, J. T. Nixon, Organometallics 2005, 24, 552–
556.
Compound 2d: Yield: 423 mg (0.594 mmol, 99%) of a blue powder; ele-
mental analysis calcd (%) for C33H63Cl3CrNOP2: C 55.81, H 8.94, N 1.97;
found: C 55.73, H 8.90, N 1.89.
[8] a) K. Blann, A. Bollmann, J. T. Dixon, A. Neveling, D. H. Morgan,
H. Maumela, E. Killian, F. M. Hess, S. Otto, L. Pepler, H. A. Ma-
homed, M. J. Overett (Sasol Pty.), WO2004056479, 2004; b) A. Boll-
mann, K. Blann, J. T. Dixon, F. M. Hess, E. Killian, H. Maumela,
D. S. McGuinness, D. H. Morgan, A. Neveling, S. Otto, M. J. Over-
ett, A. M. Z. Slavin, P. Wasserscheid, S. Kuhlmann, J. Am. Chem.
Soc. 2004, 126, 14712–14714; c) K. Blann, A. Bollmann, J. T. Dixon,
F. Hess, E. Killian, H. Maumela, D. H. Morgan, A. Neveling, S.
Otto, Chem. Commun. 2005, 620–621.
[9] a) A. Carter, S. A. Cohen, N. A. Cooley, A. Murphy, J. Scutt, D. F.
Wass, Chem. Commun. 2002, 858–859; b) D. F. Wass (BP Chemicals
Ltd.), WO0204119, 2002; c) M. J. Overett, K. Blann, A. Bollmann,
J. T. Dixon, F. Hess, E. Killian, H. Maumela, D. H. Morgan, A. Nev-
eling, S. Otto, Chem. Commun. 2005, 622–624.
[10] T. Agapie, M. W. Day, L. M. Henling, J. A. Labinger, J. E. Bercaw,
Organometallics 2006, 25, 2733–2742.
[11] M. J. Overett, K. Blann, A. Bollmann, R. de Villiers, J. T. Dixon, E.
Killian, M. C. Maumela, H. Maumela, D. S. McGuinness, D. H.
Morgan, A. Rucklidge, A. M. Z. Slavin, J. Mol. Catal. A 2008, 283,
114–119.
[12] a) J. R. Briggs, J. Chem. Soc. Chem. Commun. 1989, 674–675; b) R.
Emrich, O. Heinemann, P. W. Jolly, C. Krꢅger, G. P. J. Verhovnik,
Organometallics 1997, 16, 1511–1513.
[13] T. Agapie, S. J. Schofer, J. A. Labinger, J. E. Bercaw, J. Am. Chem.
Soc. 2004, 126, 1304–1305.
[14] S. J. Schofer, M. W. Day, L. M. Henling, J. A. Labinger, J. E. Bercaw,
Organometallics 2006, 25, 2743–2749.
[15] T. Agapie, J. A. Labinger, J. E. Bercaw, J. Am. Chem. Soc. 2007, 129,
14281–14295.
Compound 2e: Yield: 396 mg (0.547 mmol, 91%) of a blue powder; ele-
mental analysis calcd (%) for C34H65Cl3CrNOP2: C 56.39, H 9.05, N 1.93;
found: C 56.09, H 8.93, N 1.87.
Compound 2 f: Yield: 438 mg (0.589 mmol, 98%) of a blue powder; ele-
mental analysis calcd (%) for C36H61Cl3CrNOP2: C 58.10, H 8.26, N 1.88;
found: C 57.98, H 8.12, N 1.75.
Compound 2g: Yield: 339 mg (0.559 mmol, 93%) of a blue-grey powder;
elemental analysis calcd (%) for C25H55Cl3CrNOP2: C 49.55, H 9.15, N
2.31; found: C 49.28, H 9.07, N 2.22.
General oligomerization procedure: All catalytic reactions were carried
out in a magnetically stirred stainless steel autoclave (300 mL), equipped
with a pressure gauge and needle valves for injections, and heated in an
oil bath. The interior of the autoclave was protected from corrosion by a
Teflon/protective coating. A typical reaction was performed by introduc-
ing a suspension of the complex (32 mmol) in toluene (80 mL) into the re-
actor under a nitrogen atmosphere. After injection of the MAO solution
(10 wt% in toluene, Aldrich), the reactor was immediately brought to
the desired working pressure, and continuously fed by ethylene using a
reserve bottle. The reaction was stopped by closing the ethylene supply
and cooling down the system to À708C. After release of residual pres-
sure, the reaction was quenched by adding acidified methanol (5 mL). n-
Heptane, used as internal standard, was also introduced and the mixture
was analyzed by quantitative GC (PERICHROM 2100 gas chromato-
graph equipped with a HP PONA column (50 mꢄ0.2 mmꢄ0.5 mm)), first
calibrated with authentic samples.
X-ray crystallographic study: Data were collected at 150.0(1) K on a
Nonius Kappa CCD diffractometer using a MoKa (l=0.71070 ꢂ) X-ray
source and a graphite monochromator. The crystal structure of 2d was
solved using SIR97.[34] ORTEP drawings were made using ORTEP III.[35]
CCDC-725970 (2d) contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from the Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[16] M. J. Overett, K. Blann, A. Bollmann, J. T. Dixon, D. Haasbroek, E.
Killian, H. Maumela, D. S. McGuinness, D. H. Morgan, J. Am.
Chem. Soc. 2005, 127, 10723–10730.
[17] a) C. N. Temple, A. Jabri, P. Crewson, S. Gambarotta, I. Korobkov,
R. Duchateau, Angew. Chem. 2006, 118, 7208–7211; Angew. Chem.
Int. Ed. 2006, 45, 7050–7053; b) A. Jabri, C. Temple, P. Crewdson, S.
Gambarotta, I. Korobkov, R. Duchateau, J. Am. Chem. Soc. 2006,
128, 9238–9247; c) C. N. Temple, S. Gambarotta, I. Korobkov, R.
Chem. Eur. J. 2009, 15, 8259 – 8268
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8267

P. Le Floch et al.
Duchateau, Organometallics 2007, 26, 4598–4603; d) 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; e) R. D. Kçhn, Angew. Chem. 2007, 119, 249–253; Angew.
Chem. Int. Ed. 2007, 46, 245–247.
[26] a) M. J. Carney, N. J. Robertson, J. A. Halfen, L. N. Zakharov, A. L.
Rheingold, Organometallics 2004, 23, 6184–6190; b) M. Ganesan,
F. P. Gabbaꢆ, Organometallics 2004, 23, 4608–4613.
[27] a) R. Cohen, E. Weitz, J. M. L. Martin, M. A. Ratner, Organometal-
lics 2004, 23, 2315–2325; b) E. Collange, D. Duret, R. Poli, Dalton
Trans. 1999, 875–879; c) I. Cacelli, D. W. Keogh, R. Poli, A. Rizzo,
New J. Chem. 1997, 21, 133–135; d) J. S. Hess, S. Leelasubcharoen,
A. L. Rheingold, D. J. Doren, K. H. Theopold, J. Am. Chem. Soc.
2002, 124, 2454–2455; e) C. W. Hoganson, D. J. Doren, K. H. Theo-
pold, Macromolecules 2004, 37, 566–572.
[28] In their important study on the role of MAO in chromium-mediated
tri- and tetramerization processes, van Rensburg et al.[19] concluded
that abstraction of the methyl group from Cr–MAO complexes was
found to proceed with relatively low energies in the gas-phase DFT
analysis (DEZPE =12–13 kcalmolÀ1; ZPE=zero-point energy). They
also concluded that in reality this methyl abstraction is likely to be
even more favorable when a solvent is present. Hence it was demon-
strated that introduction of ethylene results in the immediate elon-
gation of Cr–Me bond length making conceivable the formation of
[18] W. J. van Rensburg, C. Grovꢀ, J. P. Steynberg, K. B. Stark, J. J.
Huyser, P. J. Steynberg, Organometallics 2004, 23, 1207–1222.
[19] W. J. van Rensburg, J. A. van den Berg, P. J. Steynberg, Organome-
tallics 2007, 26, 1000–1013.
[20] B. Blom, G. Klatt, J. C. Q. Fletcher, J. R. Moss, Inorg. Chim. Acta
2007, 360, 2890–2896.
[21] A. L. Balch, M. M. Olmstead, S. P. Rowley, Inorg. Chim. Acta 1990,
168, 255–264.
[22] L. Maier, Helv. Chim. Acta 1965, 48, 1034–1039.
[23] It should be noted that, whereas the formation of metallacyclohep-
tanes is generally accepted, formation of metallacyclononanes has
been calculated to be unlikely in the case of both Ti (bearing an
ansa-(arene)(cyclopentadienyl) ligand) and TaCl3. See, Z.-X. Yu,
K. N. Houk, Angew. Chem. 2003, 115, 832–825; Angew. Chem. Int.
Ed. 2003, 42, 808–811, and A. N. J. Blok, P. H. M. Budzelaar, A. W.
Gal, Organometallics 2003, 22, 2564–2570.
[24] Gaussian 03, Revision B.05, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomer-
y, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyen-
gar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y.
Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg,
V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O.
Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Fores-
man, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Na-
nayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen,
M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Pittsburgh
PA, 2003.
À
“dissociated” ion-pair structures (Cr Me bond lengths about 5 ꢂ).
In this study we exclusively focused on the energetic difference be-
tween free energies of the unsaturated and saturated complex as-
suming that interactions of unsaturated complexes with MAO would
be of similar amplitude.
[29] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[30] A. W. Ehlers, M. Bohme, S. Dapprich, A. Gobbi, A. Hollwarth, V.
Jonas, K. F. Kohler, R. Stegmann, A. Veldkamp, G. Frenking, Chem.
Phys. Lett. 1993, 208, 111–114.
[31] M. Keles, Z. Aydin, O. Serindag, J. Organomet. Chem. 2007, 692,
1951–1955.
[32] C. A. Busacca, J. C. Lorenz, N. Grinberg, N. Haddad, M. Hrapchak,
B. Latli, H. Lee, P. Sabila, A. Saha, M. Sarvestani, S. Shen, R. Varso-
lona, X. Wei, C. H. Senanayake, Org. Lett. 2005, 7, 4277–4280.
[33] P. Boudjouk, J. H. So, Inorg. Synth. 1992, 29, 108–111.
[34] SIR97, An integrated package of computer programs for the solu-
tion and refinement of crystal structures using single-crystal data, A.
Altomare, M. C. Burla, M. Camalli, G. Cascarano, G. Giacovazzo,
A. Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna.
[35] ORTEP III, L. J. Farrugia, University of Glasgow.
[25] a) J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77,
3865–3868; b) A. D. J. Becke, J. Phys. Chem. 1993, 97, 5648–5662;
c) J. P. Perdew, Y. Wang, Phys. Rev. B 1992, 45, 13244–13249;
d) A. K. Rappe, C. J. Kasewit, K. S. Colwell, W. A. Goddard III,
W. M. Skiff, J. Am. Chem. Soc. 1992, 114, 10024–10035.
Received: April 14, 2009
Published online: July 16, 2009
8268
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 8259 – 8268