Binary neodymium alkoxide/dialkylmagnesuim polymerization systems: Studies on the nature of the reaction intermediates and active species

Article abstract of DOI:10.1002/ejic.200400036

Attempts to identify the intermediates and/or active species generated from binary combinations of a lanthanide alk(aryl)-oxide with a dialkylmagnesium reagent, which behave as efficient olefin polymerization systems, are reported. The well-defined trinuclear complex [Nd₃(μ₃-OtBu) ₂(μ₂-OtBu)₃-(μ-OtBu)₄(THF) ₂] (1) and the monomeric precursor [Nd(OC₆H ₂tBu₂-2,6-Me-4)₃(THF)] (2) were used in association with IMg(CH₂SiMe₃)2(Et₂O)] (3). The new heterodimetallic complex [(THF)Nd(μ₃-OtBu) ₂(μ₂-OtBu)₂(OtBu)Mg₂-(CH ₂TMS)₂] (4) and the alkyllanthanide complex [Nd(OC ₆H₂tBu₂-2,6-Me-4)₂(CH ₂SiMe₃)(THF)₂] (5) have been isolated and characterized in the solid state and in solution. Complex 4 is proposed to be a reaction intermediate in the active species formation, while complex 5 is the first alkyllanthanide species isolated from an Ln(OR)₃ / MgR₂ mixture, consistent with the observed behavior of these combinations in olefin polymerization. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004.

Full text of DOI:10.1002/ejic.200400036

FULL PAPER
Binary Neodymium Alkoxide/Dialkylmagnesium Polymerization Systems:
Studies on the Nature of the Reaction Intermediates and Active Species
[a]
[a]
[b]
[c]
´ ˆ
´
´ ´
Jerome Gromada, Andre Mortreux, Guy Nowogrocki, Frederic Leising,
Thomas Mathivet,[c] and Jean-Francois Carpentier*[d]
¸
Keywords: Alkoxides / Alkyl complexes / Homogeneous catalysis / Magnesium / Neodymium / Polymerization
Attempts to identify the intermediates and/or active species
generated from binary combinations of a lanthanide alk(aryl)-
oxide with a dialkylmagnesium reagent, which behave as
efficient olefin polymerization systems, are reported. The
well-defined trinuclear complex [Nd3(µ3-OtBu)2(µ2-OtBu)3-
(µ-OtBu)4(THF)2] (1) and the monomeric precursor
[Nd(OC6H2tBu2-2,6-Me-4)3(THF)] (2) were used in associ-
ation with [Mg(CH2SiMe3)2(Et2O)] (3). The new heterodime-
tallic complex [(THF)Nd(µ3-OtBu)2(µ2-OtBu)2(OtBu)Mg2-
(CH2TMS)2] (4) and the alkyllanthanide complex
[Nd(OC6H2tBu2-2,6-Me-4)2(CH2SiMe3)(THF)2] (5) have been
isolated and characterized in the solid state and in solution.
Complex 4 is proposed to be a reaction intermediate in the
active species formation, while complex 5 is the first alkyllan-
thanide species isolated from an ‘‘Ln(OR)3’’/MgR2 mixture,
consistent with the observed behavior of these combinations
in olefin polymerization.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
species is likely to be a homonuclear alkyllanthanide, or a
related mixed alkyl-bridged lanthanide-magnesium spec-
ies.[1b,1c] Variable-temperature 1H NMR monitoring experi-
ments of mixtures of a dialkylmagnesium, for example
Mg(n-Hex)2 or nBuMgEt, with 1 or 2 showed that fast reac-
tions take place in the absence of monomer.[1b] Complete
conversion of the initial alk(aryl)oxides to unidentified
species was observed at Ϫ70 °C, followed upon raising of
the temperature by the appearance of the corresponding 1-
alkene (1-hexene, 1-butene; ethylene was immediately poly-
merized in situ), presumably arising from β-hydride elimin-
ation of a putative alkylneodymium species. The release of
1-alkene was observed at temperatures as low as Ϫ60 °C
with the tert-butoxide precursor 1 while a higher tempera-
ture, typically 0Ϫ20 °C, was required for the aryloxide pre-
cursor 2. As a consequence of this ready decomposition, a
reaction product could only be isolated and fully charac-
terized in the case of the 2/Mg(n-Hex)2 combination — the
We have recently described new versatile polymerization
systems based on combinations of a neodymium alk(aryl)-
oxide and a dialkylmagnesium reagent that enable the syn-
thesis of original materials such as high molecular weight
poly(ethylene-b-methyl methacrylate)[1] and poly(butadiene-
b-glycidyl methacrylate)[2] diblock copolymers. For this pur-
pose, the well-defined trinuclear complex [Nd3(µ3-
OtBu)2(µ2-OtBu)3(µ-OtBu)4(THF)2] (1) and the monomeric
precursor [Nd(OC6H2tBu2-2,6-Me-4)3(THF)] (2) were em-
ployed.
The active species formed in situ from these combi-
nations act either as initiators with lanthanidocene-like
properties, for example for ethylene[1] and methyl methacry-
late (MMA)[3] controlled homopolymerizations, or as cata-
lysts for 1,3-diene polymerization[2] along with a fast, re-
versible transfer reaction to the magnesium centers. From
this polymerization behavior, we anticipated that the active
alkylmagnesium
aryloxide
compound
[(n-Hex)Mg-
(OC6H2tBu2-2,6-Me-4)]2, consistent with the envisioned
transmetallation process.[1b] Herein we report further
studies of these binary polymerization systems, using
[Mg(CH2SiMe3)2(Et2O)] (3) as a co-reagent, and the iso-
lation and structural characterization of two intermediate
alkyllanthanide alk(aryl)oxide complexes.
[a]
Laboratoire de Catalyse de Lille, UPRESA 8010 CNRS Ϫ
´
Universite de Lille 1, ENSCL,
B. P. 108, 59652 Villeneuve d’Ascq Cedex, France
[b]
Laboratoire de Cristallochimie et Physicochimie du Solide,
´
UPRESA 8012 CNRS Ϫ Universite de Lille 1, ENSCL,
B. P. 108, 59652 Villeneuve d’Ascq Cedex, France
Rhodia Recherches,
[c]
93308 Aubervilliers Cedex, France
[d]
´
Laboratoire Organometalliques et Catalyse, UMR 6509 CNRS
Results and Discussion
´
Ϫ Universite de Rennes 1,
35042 Rennes Cedex, France
To obtain a better understanding of these binary catalyst
systems, we decided to concentrate our efforts on the iso-
Fax: (internat.) ϩ 33-2-23236939
E-mail: jean-francois.carpentier@univ-rennes1.fr
Eur. J. Inorg. Chem. 2004, 3247Ϫ3253
DOI: 10.1002/ejic.200400036
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3247

J.-F. Carpentier et al.
FULL PAPER
lation of the reaction intermediates and active species. For
this purpose, the typical co-reagent Mg(n-Hex)2 used for
polymerization and initial mechanistic studies was replaced
by [Mg(CH2SiMe3)2(Et2O)] (3). The latter reagent, in com-
bination with 1 or 2, has been shown to also generate active
polymerization systems towards ethylene, MMA and 1,3-
dienes, with properties similar to those of combinations
based on Mg(n-Hex)2, although somewhat less active.[1b,2]
The dialkylmagnesium 3 was selected to stabilize the corre-
sponding alkyllanthanide species, in particular by avoiding
β-H elimination reactions.[4] Another potential advantage
of 3 over Mg(n-Hex)2 and higher long-chain dialkylmag-
nesium reagents is the better propensity of the CH2SiMe3
moiety for crystallization.
Combinations of [Nd3(OtBu)9(THF)2] and
[Mg(CH2SiMe3)2(Et2O)]
Toluene solutions of 1 and 3 in various ratios were pre-
pared and stored at low temperature (Ϫ30 °C). When the
reaction was performed in a 1:3 molar ratio (i.e. Nd/Mg ϭ
1), as optimized for polymerization purposes, in the absence
of additive, it was not possible to isolate any species from
the resulting dark-brown solution. On the other hand, in
the presence of THF (10 equiv. vs. Nd), the solution re-
1
mained clear and blue. The H NMR spectrum of the blue
crystals grown from this solution is identical to the starting
trinuclear alkoxide 1, indicating that an excess of THF pre-
vents the alkylation reaction from taking place. This fact is
in direct line with the observed inactivity of the 1/Mg(n-
Hex)2 system in the presence of minute amounts of THF
for ethylene polymerization.[1b] Evans et al. have observed
a similar phenomenon with their [Y(OtBu)7Cl2]/AlMe3 sys-
tems and assigned this lack of reactivity to the formation
of the unreactive [AlMe3(THF)].[5]
Given the successful isolation by Yasuda et al. of an in-
termediate in lanthanidocene-based MMA polymeriz-
ation[6] and the efficient controlled syndiotactic polymeriz-
ation of this monomer achieved with 1/Mg(n-Hex)2,[3]
MMA (0.03 equiv. vs. Nd) was added to some crystalliza-
tion experiments. Under those conditions, colorless crystals
of [(Me3SiCH2)Mg(OtBu)(THF)] (5) formed rapidly
(15Ϫ20% isolated yield vs. Mg) in a reproducible manner,
followed after a few weeks by pale-blue crystals of a neo-
dymium complex (4; 30Ϫ40% yield vs. 1) [Equation (1)]. As
established by NMR spectroscopy and an X-ray diffraction
analysis (Figure 1, Table 1), the product 4 corresponds to
the new heterodimetallic complex [(THF)Nd(µ3-OtBu)2(µ2-
OtBu)2(OtBu)Mg2(CH2SiMe3)2], which obviously, however,
does not contain any MMA. The neodymium center in 4 is
hexacoordinate, as are most tert-butoxyneodymium deriva-
tives,[7] with six oxygen atoms in its environment, one ter-
minal tert-butoxy ligand and a THF molecule. Doubly and
triply bridging tBuO groups link the neodymium atom to
the two tetracoordinate magnesium atoms, each of which
bears an alkyl chain. Such segregation of the alkoxide and
alkyl ligands in complex 4 can be correlated with the higher
oxophilicity of Nd vs. Mg (bond disruption energies:
NdϪO, 703 Ϯ 13 kJ·molϪ1; MgϪO, 363 Ϯ 13 kJ·molϪ1).[8]
Figure 1. Two views of the molecular structure of [Nd(THF)(µ3-
OtBu)2(µ2-OtBu)2(OtBu)Mg2(CH2SiMe3)2] (4); ellipsoids are
drawn at the 40% probability level and hydrogen atoms omitted
for clarity
Several well-characterized aluminium-lanthanide alkyl-alk-
(aryl)oxide complexes have been described previously, most
of which were prepared from the interaction of lanthanide
alk(aryl)oxides with AlR3.[9] Examples of heterodimetallic
organometallic Ln-Mg complexes are much rarer and none
of them contain alk(aryl)oxide ligands as in 4.[10]
(1)
3248
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 3247Ϫ3253

Binary Neodymium Alkoxide/Dialkylmagnesium Polymerization Systems
FULL PAPER
˚
Mg(3)ϪO(6)ϪNd(2): 95.6(1)°] compare well with the corre-
sponding values observed in [(CMeNC6H3iPr2-2,6)Mg-
Table 1. Selected bond lengths [A] and angles [°] for 4
˚
Nd(2)ϪO(8)
Nd(2)ϪO(7)
Nd(2)ϪO(6)
Nd(2)ϪO(9)
Nd(2)ϪO(5)
Nd(2)ϪO(4)
Nd(2)ϪMg(2)
Nd(2)ϪMg(3)
Mg(2)ϪO(7)
Mg(2)ϪO(4)
Mg(2)ϪO(5)
Mg(2)ϪMg(3)
Mg(3)ϪO(6)
Mg(3)ϪO(4)
Mg(3)ϪO(5)
Mg(2)ϪC(37)
Mg(3)ϪC(33)
2.241(6) O(8)ϪNd(2)ϪO(7)
2.304(3) O(8)ϪNd(2)ϪO(6)
2.307(4) O(7)ϪNd(2)ϪO(6)
2.350(5) O(8)ϪNd(2)ϪO(9)
2.600(4) O(7)ϪNd(2)ϪO(9)
2.762(4) O(8)ϪNd(2)ϪO(5)
3.178(2) O(7)ϪNd(2)ϪO(5)
3.178(2) O(8)ϪNd(2)ϪO(4)
1.984(4) O(5)ϪNd(2)ϪO(4)
2.017(4) O(7)ϪMg(2)ϪO(4)
2.046(4) O(4)ϪMg(2)ϪO(5)
3.118(2) O(6)ϪMg(3)ϪC(33) 130.9(2)
1.974(4) O(4)ϪMg(3)ϪC(33) 127.9(2)
2.033(4) O(5)ϪMg(3)ϪC(33) 121.9(2)
2.040(4) Mg(2)ϪO(4)ϪMg(3) 100.7(2)
2.118(5) Mg(2)ϪO(4)ϪNd(2)
2.117(5) Mg(2)ϪO(5)ϪMg(3) 99.5(2) tion unambiguously. Minimal overlapping of the reson-
109.5(2)
105.2(2)
135.0(1)
93.3(2)
(µ2-OtBu)(THF)] [MgϪO: 1.997(2) and 1.994(2) A;
MgϪOϪMg: 98.56(6)°].[11]
The solution structure of complex 4 was briefly studied
1
by H NMR spectroscopy in [D8]toluene. As shown in Fig-
99.6(1)
107.3(2) ure 2, interpretation of the variable-temperature NMR
71.4(1)
162.4(1)
55.3(1)
93.2(2)
spectra is complicated by the high fluxionality of the mol-
ecule and, above all, the paramagnetic Nd center, which
broadens signals and affects intensity measurements. Thus,
the correlation between chemical shift and coordination
mode of the alkoxide groups previously noted in diamag-
netic yttrium and lanthanum complexes of this type — in-
75.6(2)
creased bridging is associated with lower-field shifts[12]
—
cannot be applied to determine the geometry of 4 in solu-
81.8(1)
O(7)ϪMg(2)ϪC(37) 129.5(2)
O(4)ϪMg(2)ϪC(37) 129.8(2)
O(5)ϪMg(2)ϪC(37) 121.1(2)
Mg(2)ϪO(5)ϪNd(2)
Mg(3)ϪO(5)ϪNd(2)
Mg(2)ϪO(7)ϪNd(2)
85.4(1)
85.6(2)
95.4(1)
ances is observed in the NMR spectrum at Ϫ80 °C (Fig-
ure 2), which features eight major singlet-like resonances
AϪH with relative integrations 9:9:18:2:18:4:9:4 (from high
field to low field). This pattern is consistent with the struc-
ture observed in the solid state, considering that both 18 H
singlets C and D correspond to the µ3-OtBu and µ2-OtBu
groups, which are magnetically equivalent on the NMR ti-
In the solid state, complex 4 has basically the same struc- mescale at this low temperature,[12] and that the CH2SiMe3
tural framework as the starting trinuclear alkoxide 1, except groups are nonequivalent, one of the methylene resonance
that the [Nd(OtBu)(THF)] and [Nd(OtBu)2] moieties have being obscured by residual solvent resonances (e.g. I) or a
been replaced by two [Mg(OtBu)(CH2SiMe3)] units. It is stronger intensity signal.
therefore interesting to compare the geometric features of
Complex 4 was exposed to ethylene (1 bar) in toluene
both structures. The NdϪO distance of the terminal tert- solution. At room temperature, even after a long time per-
˚
butoxide ligand in 4 [Nd(2)ϪO(8): 2.241(6) A] is somewhat iod (48 h), the formation of polyethylene was not observed.
[1b]
˚
longer than those found in 1 [2.147(4)Ϫ2.163(3) A],
This was not unexpected considering that 4 does not con-
while the NdϪO(THF) distance is significantly shorter tain any NdϪC bond susceptible to insert ethylene. Ad-
˚
˚
[2.350(5) A vs. 2.661(4) A]. It is worthwhile mentioning that dition of a slight excess of 3 to 4 under ethylene led to an
these THF and terminal OtBu ligands are statistically dis- effective polymerization. We can therefore hypothesize that
ordered (see Exp. Sect.), which may lead to averaged values. 4 is an intermediate in the formation of the real active spec-
The doubly bridging NdϪO distances are, as expected, ies in 1/MgR2 systems for ethylene polymerization. On the
longer than those for the terminal alkoxide ligand other hand, the 1:1 Nd/Mg stoichiometry imposed for the
˚
[Nd(2)ϪO(7): 2.304(3); Nd(2)ϪO(6): 2.307(4) A], and reaction between 1 and 3 was not recovered in the final
slightly
shorter
than
those
measured
in
1
complex 4, and we cannot discard the potential catalytic
˚
[2.333(3)Ϫ2.458(3) A]. Also, the nonequivalent triply bridg- activity of other products remaining in the mother solution
ing NdϪO bonds [Nd(2)ϪO(4): 2.762(4); Nd(2)ϪO(5): and so far unidentified, such as species of the type
˚
2.600(4) A] are the longest ones observed in the structure [(tBuO)2NdϪR]n.
and longer than the corresponding values in 1 [2.624(3) and
2.409(3) A],[1b] showing that the Nd atom is highly distorted
from an ideal octahedral geometry. The angles around the
neodymium center are similar to those observed in 1 except
for the O(µ3)ϪNdϪO(µ3) angle, which is smaller in the het-
˚
Combinations of [Nd(OC6H2tBu2-2,6-Me-4)3(THF)] and
[Mg(CH2SiMe3)2(Et2O)]
We focused next on [Nd(OC6H2tBu2-2,6-Me-4)3(THF)]
erodimetallic complex [55.3(1)° in 4 vs. 66.0(1)° in 1]. This (2). This precursor, thanks to its bulky and readily dis-
difference may be ascribed to the asymmetry in the complex placeable ligand commonly used in the chemistry of group
caused by the shorter MgϪO bonds compared to the 1Ϫ4 metals, might lead to more stable and/or more easily
NdϪO ones. The equivalent magnesium atoms are four-co- crystallizable complexes, as evidenced by the isolation of
ordinate, showing a distorted tetrahedral geometry. This [Mg(OC6H2tBu2-2,6-Me-4)(n-Hex)]2 from a 2/Mg(n-Hex)2
geometry has been observed previously in some mixed al- (1:1) mixture in toluene.[1b] The quantitative formation of
koxy-amido complexes such as [(CMeNC6H3iPr2- the latter compound is in agreement with the assumed in
2,6)Mg(µ2-OtBu)(THF)] and [(CMeNC6H3iPr2-2,6)Mg(µ2- situ formation of a bis(aryloxy)alkylneodymium complex.
OC6H9)].[11] The doubly bridging MgϪO distances However, the latter could not be isolated either from Mg(n-
˚
[Mg(2)ϪO(7): 1.984(4); Mg(3)ϪO(6): 1.974(4) A] and Hex)2 or upon using the dialkylmagnesium compound 3 in
MgϪOϪNd
angles
[Mg(2)ϪO(7)ϪNd(2):
95.4(1); pure toluene, despite repeated efforts. On the other hand,
Eur. J. Inorg. Chem. 2004, 3247Ϫ3253
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3249

J.-F. Carpentier et al.
FULL PAPER
1
Figure 2. Representative variable-temperature H NMR spectra of [Nd(µ3-OtBu)2(µ2-OtBu)2(OtBu)Mg2(CH2SiMe3)2(THF)] (4) in [D8]-
toluene; top: Ϫ80 °C; bottom: 0 °C (I denotes residual toluene resonance and/or a signal from 4)
the addition of THF (10 equiv. vs. Nd) to a mixture of 2 (THF)2] (6) along with [(4-Me-2,6-tBu2C6H2O)Mg(CH2-
and 3 (1:1) enabled the ready crystallization of the pentaco- SiMe3)]2 (7), isolated in 34% and 39% yields, respectively
ordinate complex [(4-Me-2,6-tBu2C6H2O)2Nd(CH2SiMe3)- [Equation (2)].
3250
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 3247Ϫ3253

Binary Neodymium Alkoxide/Dialkylmagnesium Polymerization Systems
FULL PAPER
also been claimed in the case of [Nd(µ2-OAr)(OAr)2-
(py)2]2.[9h]
˚
Table 2. Selected bond lengths [A] and angles [°] for 6
Nd(1)ϪO(1A)
2.235(3) O(1A)ϪNd(1)ϪO(1B) 146.21(10)
2.227(3) O(1A)ϪNd(1)ϪO(3B) 83.70(10)
Nd(1)ϪO(1B)
Nd(1)ϪO(3A)
Nd(1)ϪO(3B)
Nd(1)ϪC(16)
O(1A)ϪC(1A)
O(1B)ϪC(1B)
C(16)ϪSi
(2)
2.500(3) O(1B)ϪNd(1)ϪO(3B)
91.69(10)
2.487(3) O(1A)ϪNd(1)ϪC(16) 113.02(12)
2.482(4) O(1B)ϪNd(1)ϪC(16) 100.26(12)
1.352(5) O(3B)ϪNd(1)ϪC(16)
88.54(12)
1.355(5) O(1A)ϪNd(1)ϪO(3A) 84.40(9)
1.843(4) O(1B)ϪNd(1)ϪO(3A) 93.56(10)
The solid-state structure of 6 features the metal center
coordinated in a slightly distorted trigonal-bipyramidal
mode by one equatorial methyltrimethylsilyl and two ary-
loxy groups (sum of the angles around Nd ϭ 359.49°), and
two axial THF ligands (Figure 3, Table 2). A plane of sym-
metry (noncrystallographic) embraces the Nd, both
O(THF) and the C(CH2SiMe3) atoms. Similar trigonal-bi-
pyramidal coordination has been observed previously in
Ln(OR)3L2 complexes such as [Nd(OCtBu3)3(CH3CN)2],[13]
as well as in mixed halogen-aryloxide complexes, such as
C(16)ϪNd(1)ϪO(3A) 103.82(12) O(3B)ϪNd(1)ϪO(3A) 165.46(9)
The NdϪO(Ar) distances are nearly identical for both
˚
ligand units [2.227(3), 2.235(3) A] and are somewhat longer
than those in 2, but similar to those observed in related
complexes.[16] Also, the NdϪO(THF) distances [2.487(3),
˚
2.500(3) A] are within the range of values reported for
NdϪTHF complexes.[17] The NdϪC distance of 2.482(4) A
˚
[SmI(OAr)2(THF)2]
and
[(4-Me-2,6-tBu2C6H2-
can be compared to the terminal SmϪC distance of 2.45(1)
O)2LnCl(THF)2] (Ln ϭ Er, Yb),[14] and in the mixed aryl-
halogen complex [(2,6-dimesitylphenyl)YbCl2(THF)2],[15] in
which the chlorine atoms occupy the axial positions. The
O(1A)ϪNd(1)ϪO(1B) angle [146.2(1)°] in 6 is much larger
than 120°, reflecting the larger steric requirement of the ar-
yloxide ligands compared to the CH2TMS group. The
bulkiness of the equatorial ligands affects the geometry of
the axial ligands, with a lower (THF)OϪNdϪO(THF) an-
gle [165.46(9)°] than the theoretical value (180°). The in-
crease in coordination number from the four-coordinate,
distorted tetrahedral tris(aryloxide) precursor 2 to pentaco-
ordinate complex 6 also likely stems from the decreasing
steric requirement of CH2SiMe3 compared to the very
bulky aryloxide. We assume that this may be the reason why
it was not possible to isolate a putative alkyl complex, such
as [(ArO)2Nd(CH2SiMe3)], in the absence of THF. The
necessary addition of a Lewis base for crystallization has
˚
A
in [Sm(OAr)3(CH2SiMe3)2][Li(THF)]2 (Ar ϭ 2,6-
iPr2C6H3), reflecting the minimal 2.6% difference in the
ionic radii.[18] It is, however, much lower than that of 2.68(1)
˚
A in [{1,1Ј-(3,5-tBu2-2-C6H2O)2}La{CH(SiMe3)2}(THF)3]
due to the larger ionic radius of lanthanum,[18] its hexa-
coordination and the bulkiness of both aryloxide and
alkyl groups.[19]
1
The characterization of 6 in solution by H NMR spec-
troscopy was unambiguous and indicated a symmetric
structure on the NMR time scale, as observed in the solid
state. Complex 6 exhibits the same pattern as 2 for the
equivalent aryloxide ligands — narrow signals for the aro-
matic and Me protons and a broad singlet for the tert-butyl
groups at higher field. The signals assigned to the alkyl
group appear at δ ϭ Ϫ1.25 and Ϫ7.15 ppm for the methyl-
ene and trimethylsilyl groups, respectively, while the two
equivalent coordinated THF molecules give broad reson-
ances at δ ϭ Ϫ6.90 and Ϫ9.66 ppm.
As expected, despite the presence of an NdϪC bond,
complex 6 does not initiate ethylene polymerization under
the conditions used for catalytic combinations of 2 and 3
(1 atm, 0Ϫ80 °C). This is, of course, due to the presence of
the two coordinated THF molecules that ‘‘block’’ the active
site for the weaker Lewis base ethylene.
In conclusion, some reaction products of the combi-
nations of neodymium alk(aryl)oxides 1 and 2 with
[Mg(CH2SiMe3)2(Et2O)] (3) have been isolated and fully
characterized. The new heterodimetallic complex 4, ob-
tained from the association of 1 with 3, appears to be a
plausible intermediate towards the species active in ethylene
and methyl methacrylate (co)polymerizations. Indeed, al-
though inert towards ethylene, complex 4 can be further
activated upon MgR2 addition, eventually yielding species
that are active in polymerization. The alkylneodymium
complex 5 is also inert towards olefins despite its reactive
NdϪC bond. The isolation of this complex is, however,
Figure 3. Molecular structure of [(4-Me-2,6-tBu2C6H2O)3Nd(CH2-
SiMe3)(THF)2] (6); ellipsoids are drawn at the 50% probability level
and hydrogen atoms omitted for clarity
Eur. J. Inorg. Chem. 2004, 3247Ϫ3253
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3251

J.-F. Carpentier et al.
FULL PAPER
0.10 mmol) was added at room temperature to the resulting purple
solution, and the mixture was immediately cooled to Ϫ30 °C. After
a few days, colorless crystals of [(4-Me-2,6-tBu2C6H2O)Mg(CH2S-
iMe3)]2 (7) were removed from the solution (0.013 g, 39%) and blue
crystals of 6 suitable for X-ray analysis were obtained from the blue
solution after a month (0.025 g, 34%).
noteworthy in this area since it is the first example of an
alkyl complex prepared from a neodymium alk(aryl)oxide/
dialkylmagnesium mixture. Formation of 5 is the strongest
evidence to date corroborating that an alkyllanthanide
could be the active species in olefin polymerizations in-
itiated by these new binary catalytic systems.
6: 1H NMR (200 MHz, [D6]benzene, 23 °C): δ ϭ 21.36 (s, 2 H,
Ar), 10.47 (s, 3 H, Me), 7.99 (br. s, 18 H, tBu), Ϫ1.25 (s, 2 H, CH2),
Ϫ6.90 (br. s, THF), Ϫ7.15 (s, SiMe3), Ϫ9.66 (br. s, THF) ppm.
C42H73NdO4Si (814.35): calcd. C 61.94, H 9.04; found C 60.45,
H 8.96.
Experimental Section
7: 1H NMR (200 MHz, [D6]benzene, 23 °C): δ ϭ 7.19 (s, 2 H, Ar),
2.29 (s, 3 H, Me), 1.59 (s, 18 H, tBu), 0.06 (s, 9 H, SiMe3), Ϫ1.16
(s, 2 H, CH2) ppm. 13C{1H} NMR (50 MHz, C6D6, 23 °C): δ ϭ
159.4 (CAr), 139.8 (CAr), 128.9 (CAr), 127.6 (CAr), 35.7 (Cq, tBu),
33.9 (CH3, tBu), 21.9 (CH3), 4.1 [Si(CH3)], Ϫ6.2 (CH2-Mg) ppm.
C19H34MgOSi (330.86): calcd. C 68.97, H 10.36; found C 68.4, H
9.82.
General: All operations were performed in a glove box under nitro-
gen. Solvents and deuterated solvents were freshly distilled from
sodium-potassium amalgam under argon and degassed prior to
use. NMR spectra were recorded in Teflon-valved NMR tubes with
a Bruker AC-200 spectrometer. 1H (200 MHz) and 13C{1H}
(50 MHz) chemical shifts are expressed in ppm vs. SiMe4 and are
referenced to the residual solvent peaks. The complexes [Nd3-
(OtBu)9(THF)2],[1b] [Nd(OC6H2tBu2-2,6-Me-4)3(THF)],[20] and
[Mg(CH2SiMe3)2(OEt2)][21] were prepared according to the re-
ported procedures. Methyl methacrylate (99%, Aldrich) was vac-
uum-distilled from CaH2 and stored at Ϫ20 °C under argon.
X-ray Crystal-Structure Determination of 4: A pale-blue crystal of
approximate dimensions 0.30 ϫ 0.25 ϫ 0.10 mm was mounted on
a glass fiber. X-ray data were collected with a Bruker SMART
CCD area-detector diffractometer with Mo-Kα radiation (λ ϭ
Reaction of [Nd3(µ3-OtBu)2(µ2-OtBu)3(µ-OtBu)4(THF)2] (1) with
[Mg(CH2SiMe3)2(Et2O) (3)]. Synthesis of [Nd(µ3-OtBu)2(µ2-
OtBu)2(µ-OtBu)Mg2(CH2SiMe3)2(THF)] (4) and [(Me3SiCH2)-
Mg(OtBu)(THF)] (5): [Nd3(OtBu)9(THF)2] (1; 0.124 g, 0.10 mmol)
was dissolved in a minimal amount of toluene (ca. 0.25 ml). Methyl
methacrylate (10 µL, 0.010 mmol) and [Mg(CH2SiMe3)2(OEt2)] (3;
0.084 g, 0.33 mmol) were then added at room temperature to the
resulting deep-blue solution. The solution turned green after ad-
dition of the magnesium reagent and was immediately cooled to
Ϫ30 °C. After 2 d, colorless crystals of [(Me3SiCH2)Mg(Ot-
Bu)(THF)] (5) were removed from the solution (0.015 g, 17%) and
after typically about 10 d, pale-blue crystals of 4 had grown from
the dark-green solution (0.027 g, 37%), which proved suitable for
X-ray diffraction.
˚
0.71073 A) at T ϭ 100 K. Crystal data: monoclinic, Cc; a ϭ
˚
26.312(5), b ϭ 11.595(2), c ϭ 22.888(5) A; β ϭ 109.127(3)°; V ϭ
3
˚
6597(2) A ; formula unit: C32H75O6Mg2NdSi2 with Z ϭ 6; molecu-
lar weight: 804.96; calculated density ϭ 1.216 g·cmϪ3; F(000) ϭ
2562; µ(Mo-Kα) ϭ 1.297 mmϪ1. 22487 reflections collected (2.27°
Ͼ Θ Ͼ 25.00°). The structure was solved by direct methods.[22] Of
the six neodymium atoms in the cell, two are on the twofold axis
and the four others are in general positions. The resolution was
complicated by the disorder between the monodentate OtBu ions
and THF molecules, which had to be refined as independent parts.
Hydrogen-atom positions were calculated on a geometrical basis.
Full-matrix least-squares refinement on F2 based on 11197 inde-
pendent reflections converged with 620 variable parameters and
three restraints. R1 ϭ 0.0360 [for 10265 data with I Ͼ 2σ(I)];
wR2 ϭ 0.0864; GoF (F2) ϭ 1.057. ∆ρmax ϭ 1.776 and Ϫ0.851
4: 1H NMR (200 MHz, [D8]toluene, 23 °C): δ ϭ 19.56 (s, 1 H),
12.82 (br. s, 0.6 H), 10.40 (s, 2 H), 1.37 (s, 2 H), Ϫ15.87 (s, 0.4 H),
Ϫ30.11 (br. s, 0.7 H), Ϫ33.43 (br. s, 0.4 H) ppm. 1H NMR
(200 MHz, [D8]toluene, Ϫ35 °C): δ ϭ 27.54 (s, 1 H), 19.09 (s, 1 H),
15.16 (s, 2 H), 2.25 (s, 0.2 H), 2.10 (s, 0.1 H), 1.69 (s, 1.5 H), Ϫ24.53
(s, 0.4 H), Ϫ44.86 (s, 1 H), Ϫ51.62 (s, 0.4 H) ppm. 1H NMR
(200 MHz, [D8]toluene, Ϫ80 °C): δ ϭ 39.38 (s, 9 H), 28.60 (s, 9 H),
22.06 (s, 18 H), 3.48 (s, 2 H), 2.37 (s, 2 H), 2.15 (s, 18 H), Ϫ37.21
(s, 4 H), Ϫ66.03 (s, 9 H), Ϫ78.01 (s, 4 H) ppm. C32H75Mg2NdO6Si2
(804.96): calcd. C 47.75, H 9.39; found C 47.1, H 9.11.
Ϫ3
˚
e·A
.
X-ray Crystal Structure Determination of 6: A pale-blue crystal of
approximate dimensions 0.45 ϫ 0.25 ϫ 0.12 mm was mounted on
a glass fiber. X-ray data were collected with a Bruker SMART
CCD area-detector diffractometer with Mo-Kα radiation (λ ϭ
˚
0.71073 A) at T ϭ 100 K. Crystal data: orthorhombic, Pca21; a ϭ
3
˚
˚
23.732(5), b ϭ 9.799(2), c ϭ 18.902(4) A; V ϭ 4395.7(16) A ; for-
mula unit: C42H73O4SiNd with Z ϭ 4; molecular weight: 814.33;
calculated density ϭ 1.119 g·cmϪ3; F(000) ϭ 1724; µ(Mo-Kα) ϭ
1.244 mmϪ1. 30722 reflections collected (2.15° Ͼ Θ Ͼ 28.04°). The
structure was solved by direct methods.[22] Hydrogen-atom coordi-
nates were generated by geometrical calculations. Full-matrix least-
squares refinement on F2 based on 9437 independent reflections
converged with 450 variable parameters and one restraint. R1 ϭ
0.0339 [for 7321 data with I Ͼ 2σ(I)]; wR2 ϭ 0.0637; GoF (F2) ϭ
5: 1H NMR (200 MHz, [D6]benzene, 23 °C): δ ϭ 3.68 (m, 4 H,
THF), 1.29 (s, 9 H, tBu), 1.24 (m, 4 H, THF), 0.41 (s, 9 H, SiMe3),
Ϫ1.33 (s, 2 H, CH2) ppm. 13C{1H} NMR (50 MHz, [D6]benzene,
23 °C): δ ϭ 69.7 (CH2O, THF), 67.8 (Cq, tBu), 35.3 (CH3, tBu),
25.5 (CH2, THF), 5.6 [Si(CH3)], Ϫ6.0 (CH2Mg) ppm.
C12H28MgO2Si (256.74): calcd. C 56.14, H 10.99; found C 55.86,
H 10.53.
1.010. ∆ρmax ϭ 2.015 and Ϫ0.808 e·AϪ3. The largest residuals are
˚
close to one of the THF molecule and are probably due to some
orientational disorder.
Reaction
of
[Nd(OC6H2tBu2-2,6-Me-4)3(THF)]
(2)
with
[Mg(CH2SiMe3)2(Et2O)] (3). Synthesis of [Nd(OC6H2tBu2-2,6-Me- CCDC-228005 (4) and -228006 (6) contain the supplementary crys-
4)2(CH2SiMe3)(THF)2] (6) and [(4-Me-2,6-tBu2C6H2O)Mg(CH2S- tallographic data for this paper. These data can be obtained free
iMe3)]2 (7): [Nd(OC6H2tBu2-2,6-Me-4)3(THF)] (2) (0.087 g, of charge at www.ccdc.cam.ac.uk/conts/retrieving.html [or from the
0.10 mmol) was dissolved in a mixture of toluene (0.140 g,
0.16 mL) and THF (0.050 g, 0.70 mmol) THF was weighted to in-
troduce an exact amount. [Mg(CH2SiMe3)2(Et2O)] (3) (0.027 g;
Cambridge Crystallographic Data Centre, 12 Union Road, Cam-
bridge, CB2 1EZ, UK; Fax: (internat.) ϩ 44-1223-336033; E-mail:
deposit@ccdc.cam.ac.uk].
3252
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 3247Ϫ3253

Binary Neodymium Alkoxide/Dialkylmagnesium Polymerization Systems
FULL PAPER
[(THF)2Sm(OtBu)2(µ2-OtBu)Al(iBu)2]: G. R. Giesbrecht, J. C.
Gordon, D. L. Clark, B. L. Scott, J. G. Watkin, K. J. Young,
Inorg. Chem. 2002, 41, 6372Ϫ6379.
Acknowledgments
We are grateful to Rhodia Electronics and Catalysis for financial
support of this research (PhD grant for J. G.).
[10] [10a]
W. L. Wu, M. W. Chen, P. Zhou, Organometallics 1991,
[10b]
10, 98Ϫ104.
J. Lin, Z. Wang, J. Organomet. Chem. 1999,
[1]
[1a] J. Gromada, T. Chenal, A. Mortreux, J. W. Ziller, F. Leising,
589, 127Ϫ132.
[11]
J.-F. Carpentier, Chem. Commun. 2000, 2183Ϫ2184. [1b] J. Gro-
mada, A. Mortreux, T. Chenal, J. W. Ziller, F. Leising, J.-F.
Carpentier, Chem. Eur. J. 2002, 8, 3773Ϫ3788. [1c] J. Gromada,
T. Chenal, A. Mortreux, F. Leising, J.-F. Carpentier, J. Mol.
Catal. 2002, 182/183, 537Ϫ543.
M. H. Chisholm, J. Gallucci, K. Phomphrai, Inorg. Chem.
2002, 41, 2785Ϫ2794.
[12] [12a]
W. J. Evans, M. S. Sollberger, T. P. Hanusa, J. Am. Chem.
[12b]
Soc. 1988, 110, 1841Ϫ1850.
W. J. Evans, M. S. Sollberger,
Inorg. Chem. 1988, 27, 4417Ϫ4423. [12c] W. J. Evans, J. M. Olof-
[12d]
[2]
son, J. W. Ziller, Inorg. Chem. 1989, 28, 4309Ϫ4311.
W. J.
J. Gromada, L. Le Pichon, F. Leising, A. Mortreux, J.-F. Carp-
Evans, J. M. Olofson, J. W. Ziller, J. Am. Chem. Soc. 1990, 112,
entier, J. Organomet. Chem. 2003, 683, 44Ϫ55.
J. Gromada, C. Fouga, T. Chenal, A. Mortreux, J.-F. Carpent-
[12e]
[3]
2308Ϫ2314.
W. J. Evans, R. E. Golden, J. W. Ziller, Inorg.
Chem. 1991, 30, 4963Ϫ4968.
W. A. Herrmann, R. Anwander, M. Kleine, W. Scherer, Chem.
Ber. 1992, 125, 1971Ϫ1979.
Z. M. Hou, A. Fujita, T. Yoshimura, A. Jesorka, Y. G. Zhang,
H. Yamazaki, Y. Wakatsuki, Inorg. Chem. 1996, 35,
7190Ϫ7195.
ier, Macromol. Chem. Phys. 2002, 203, 550Ϫ555.
It should be pointed out, as commented by a referee, that nu-
[13]
[14]
[4]
merous n-alkyllanthanide complexes, including active polymer-
ization catalysts, are known, which indicates that β-H elimin-
ation is not necessarily a problem. Also, β-H elimination is not
the only possible decomposition process in the lanthanide tert-
butoxide/MgR2 systems described in this paper. When the R
alkyl group, transferred from the magnesium to the neodym-
ium atom, does not undergo β-H elimination, a second path-
way can take place that involves decomposition of the tert-
butoxy ligand into isobutene to give the corresponding alkane
(RH) and lanthanide oxide. See ref.[1b]
[15]
G. W. Rabe, C. D. Berube, G. P. A. Yap, K.-C. Lam, T. E.
Concolino, A. L. Rheingold, Inorg. Chem. 2002, 41,
1446Ϫ1453.
[16] [16a]
˚
[Nd(Odpp)3(DME)] (Odpp ϭ O-2,6-Ph2C6H3): 2.191 A:
G. B. Deacon, T. Feng, P. C. Junk, B. W. Skelton, A. H. White,
[16b]
Chem.
Ber./Recueil
1997,
130,
851Ϫ857.
˚
˚
[5]
[Nd(Odpp)3(THF)2]: 2.190 A; [Nd(Odpp)3(THF)]: 2.193 A: G.
B. Deacon, T. Feng, B. W. Skelton, A. H. White, Aust. J. Chem.
1995, 48, 741Ϫ756.
W. J. Evans, T. J. Boyle, J. W. Ziller, J. Am. Chem. Soc. 1993,
115, 5084Ϫ5092.
H. Yasuda, H. Yamamoto, K. Yokota, S. Miyake, A. Naka-
[6]
[17] [17a]
[NdCl3(THF)4]: C. Wenqi, J. Zhonsheng, X. Yan, F. Yu-
mura, J. Am. Chem. Soc. 1992, 114, 4908Ϫ4910.
[17b]
[7]
guo, Y. Guangdi, Inorg. Chim. Acta 1987, 130, 125Ϫ129.
R. Anwander, Top. Curr. Chem. 1996, 179, 151Ϫ245, and refer-
[Nd2(OCHiPr2)6(THF)2]: D. M. Barnhart, D. L. Clark, J. C.
Gordon, J. C. Huffman, J. G. Watkin, B. D. Zwick, J. Am.
Chem. Soc. 1993, 115, 8461Ϫ8462.
ences cited therein.
J. A. Kerr, in CRC Handbook of Chemistry and Physics (Ed.:
[8]
D. R. Lide), 81st ed., CRP Press, Boca Raton, Florida, USA,
Effective ionic radii for six-coordinate metal centers: La3ϩ
1.172 A; Nd : 0.983 A; Sm : 0.958: A. R. D. Shannon, Acta
Crystallogr., Sect. A 1976, A32, 751.
:
[18]
[19]
2000.
3ϩ
3ϩ
˚
˚
˚
[9] [9a]
[(tBuO)(THF)Y{(µ2-OtBu)(µ2-Me)AlMe2}2], [(tBuO)(Cl)-
(THF)2Y(µ2-OtBu)2AlMe2]: see ref.[5] [9b][Ln(µ2-OtBu)3(µ2-Me)3-
(AlMe2)3] (Ln ϭ Y, Pr, Nd): P. Biagini, G. Lugli, L. Abis, R.
C. J. Schaverien, N. Meijboom, A. G. Orpen, J. Chem. Soc.,
Chem. Commun. 1992, 124Ϫ126.
[9c]
Millini, J. Organomet. Chem. 1994, 474, C16ϪC18.
[(ArO)Sm{(µ2-OAr)(µ2-R)AlR2}2] (Ar ϭ 2,6-iPr2C6H3, R ϭ
Me, Et): J. C. Gordon, G. R. Giesbrecht, J. T. Brady, D. L.
[20] [20a]
P. B. Hitchcock, M. F. Lappert, A. Singh, J. Chem. Soc.,
[20b]
Chem. Commun. 1983, 1499Ϫ1501.
M. F. Lappert, A.
Singh, R. G. Smith, Inorg. Synth. 1990, 27, 164Ϫ168.
R. A. Andersen, G. Wilkinson, J. Chem. Soc., Dalton Trans.
1977, 809Ϫ811.
Clark, D. W. Keogh, B. L. Scott, J. G. Watkin, Organometallics
[21]
[9d]
2002, 21, 127Ϫ131.
[(ArO)La{(µ-OAr)AlMe2}2]: G. R.
Giesbrecht, J. C. Gordon, J. T. Brady, D. L. Clark, D. W.
[22] [22a]
SMART Software Users Guide, version 5.0, Bruker Ana-
Keogh, R. Michalczyk, B. L. Scott, J. G. Watkin, Eur. J. Inorg.
[22b]
[9e]
lytical X-ray Systems, Inc., Madison, WI, USA, 1999.
SAINT Software Users Guide, version 6.0, Bruker Analytical
X-ray Systems, Inc., Madison, WI, USA, 1999. [22c] G. M. Shel-
drick, SADABS, Bruker Analytical X-ray Systems, Inc., Madi-
son, WI, USA, 1999. [22d] G. M. Sheldrick, SHELXTL, version
5.10, Bruker Analytical X-ray Systems, Inc., Madison, WI,
USA, 1999.
Chem. 2002, 723Ϫ731.
[(ArO)Y{(µ2-OAr)(µ2-Me)AlMe2}2]
(Ar ϭ 2,6-iPr2C6H3), [(ArO)2Ln{(µ2-Me)2AlMe2}] (Ln ϭ Y,
Lu; Ar ϭ 2,6-tBu2C6H3): A. Fischbach, E. Herdtweck, R. An-
wander, G. Eickerling, W. Scherer, Organometallics 2003, 22,
[9f]
499Ϫ509.
[(ArO)2(THF)2Ln{(µ2-OAr)2AlMe2}] (Ln ϭ Nd,
Yb, Ar ϭ 2,6-Me2C6H3, R ϭ Me, Et): W. J. Evans, M. A.
[9g]
Ansari, J. W. Ziller, Inorg. Chem. 1995, 34, 3079Ϫ3082.
[Nd{(µ2-OAr)2AlMe2}3] (Ar ϭ 4-MeC6H4): W. J. Evans, M. A.
Ansari, J. W. Ziller, Polyhedron 1997, 16, 3429Ϫ3434.
Received January 15, 2004
Early View Article
Published Online June 17, 2004
[9h]
[{(iPrO)(iBu)Al(µ2-OiPr)2Sm(OiPr)(HOiPr)}(µ2-OiPr)]2,
Eur. J. Inorg. Chem. 2004, 3247Ϫ3253
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3253