Probing the structure of methylalumoxane (MAO) by a combined chemical, spectroscopic, neutron scattering, and computational approach

Article abstract of DOI:10.1021/om4002878

The composition of methylalumoxane (MAO) and its interaction with trimethylaluminum (TMA) have been investigated by a combination of chemical, spectroscopic, neutron scattering, and computational methods. The interactions of MAO with donor molecules such as THF, pyridine, and PPh₃ as a means of quantifying the content of free and bound TMA have been evaluated, as well as the ability of MAO to produce [Me ₂AlL₂]⁺ cations, a measure of the electrophilic component likely to be involved in the activation of single-site catalysts. THF, pyridine, and diphenylphosphinopropane (dppp) give the corresponding TMA-donor ligand complexes accompanied by the formation of [Me ₂AlL₂]⁺ cations. The results suggest that MAO contains not only Lewis acid sites but also structures capable of acting as sources of [AlMe₂]⁺ cations. Another unique, but still unresolved, structural aspect of MAO is the nature of bound and free TMA. The addition of the donors OPPh₃, PMe ₃, and PCy₃ leads to the precipitation of polymeric MAO and shows that about one-fourth of the total TMA content is bound to the MAO polymers. This conclusion was independently confirmed by pulsed field gradient spin echo (PFG-SE) NMR measurements, which show fast and slow diffusion processes resulting from free and MAO-bound TMA, respectively. The hydrodynamic radius R_h of polymeric MAO in toluene solutions was found to be 12 ± 0.3 A, leading to an estimate for the average size of MAO polymers of about 50-60 Al atoms. Small-angle neutron scattering (SANS) resulted in the radius R_S = 12.0 ± 0.3 A for the MAO polymer, in excellent agreement with PFG-SE NMR experiments, a molecular weight of 1800 ± 100, and about 30 Al atoms per MAO polymer. The MAO structures capable of releasing [AlMe₂]⁺ on reaction with a base were studied by quantum chemical calculations on the MAO models (OAlMe)_n(TMA) _m for up to n = 8 and m = 5. Both -O-AlMe₂-O- and -O-AlMe₂-μ-Me- four-membered rings are about equally likely to lead to dissociation of [AlMe₂]⁺ cations. The resulting MAO anions rearrange, with structures containing separated Al₂O ₂ 4-rings being particularly favorable. The results support the notion that catalyst activation by MAO can occur by both Lewis acidic cluster sites and [AlMe₂]⁺ cation formation.

Full text of DOI:10.1021/om4002878

Article
pubs.acs.org/Organometallics
Probing the Structure of Methylalumoxane (MAO) by a Combined
Chemical, Spectroscopic, Neutron Scattering, and Computational
Approach
Fabio Ghiotto,† Chrysoula Pateraki,† Jukka Tanskanen,‡ John R. Severn,§ Nicole Luehmann,∥
Andre Kusmin,∥ Jorg Stellbrink,*,∥ Mikko Linnolahti,*,‡ and Manfred Bochmann*,†
́
̈
†Wolfson Materials and Catalysis Centre, School of Chemistry, University of East Anglia, Norwich, United Kingdom
‡Department of Chemistry, University of Eastern Finland, Joensuu Campus, FI-80101 Joensuu, Finland
§UHMwPE Chemistry & Catalysis, DSM, NL-6160MD Geleen, The Netherlands
∥Julich Centre for Neutron Science and Institute for Complex Systems, Forschungszentrum Julich, D-52425 Julich, Germany
̈
̈
̈
S
* Supporting Information
ABSTRACT: The composition of methylalumoxane (MAO) and its interaction with
trimethylaluminum (TMA) have been investigated by a combination of chemical,
spectroscopic, neutron scattering, and computational methods. The interactions of MAO
with donor molecules such as THF, pyridine, and PPh3 as a means of quantifying the content
of “free” and “bound” TMA have been evaluated, as well as the ability of MAO to produce
[Me2AlL2]+ cations, a measure of the electrophilic component likely to be involved in the
activation of single-site catalysts. THF, pyridine, and diphenylphosphinopropane (dppp) give
the corresponding TMA−donor ligand complexes accompanied by the formation of
[Me2AlL2]+ cations. The results suggest that MAO contains not only Lewis acid sites but also
structures capable of acting as sources of [AlMe2]+ cations. Another unique, but still
unresolved, structural aspect of MAO is the nature of “bound” and “free” TMA. The addition
of the donors OPPh3, PMe3, and PCy3 leads to the precipitation of polymeric MAO and shows that about one-fourth of the total
TMA content is bound to the MAO polymers. This conclusion was independently confirmed by pulsed field gradient spin echo
(PFG-SE) NMR measurements, which show fast and slow diffusion processes resulting from free and MAO-bound TMA,
respectively. The hydrodynamic radius Rh of polymeric MAO in toluene solutions was found to be 12 0.3 Å, leading to an
estimate for the average size of MAO polymers of about 50−60 Al atoms. Small-angle neutron scattering (SANS) resulted in the
radius RS = 12.0 0.3 Å for the MAO polymer, in excellent agreement with PFG-SE NMR experiments, a molecular weight of
1800 100, and about 30 Al atoms per MAO polymer. The MAO structures capable of releasing [AlMe2]+ on reaction with a
base were studied by quantum chemical calculations on the MAO models (OAlMe)n(TMA)m for up to n = 8 and m = 5. Both
−O−AlMe2−O− and −O−AlMe2−μ-Me− four-membered rings are about equally likely to lead to dissociation of [AlMe2]+
cations. The resulting MAO anions rearrange, with structures containing separated Al2O2 4-rings being particularly favorable. The
results support the notion that catalyst activation by MAO can occur by both Lewis acidic cluster sites and [AlMe2]+ cation
formation.
fuller understanding of the chemistry involved has still not been
achieved.4−7,10−17
INTRODUCTION
■
Methylalumoxane (MAO), the product of the controlled
hydrolysis of trimethylaluminum (TMA), is a widely used
catalyst activator for olefin oligomerizations and polymer-
izations.1,2 However, despite its important industrial use, little
is known about the structure, due to a complex set of equilibria
and to the lack of isolable components that are amenable to
structural characterization.3−6 In solution-phase catalysts MAO
is typically employed at Al/transition-metal ratios on the order of
(103−104)/1,7 although Al/metal ratios of over 300000/1 have
been reported.8 MAO therefore contributes significantly to the
operating costs of polymerization processes.9 Over the last 30
years there have been many experimental and computational
attempts to shed light on the constitution of MAO, to find a
cheaper alternative or to improve its solubility and stability, but a
In the simplest analysis, solutions of MAO consist of two
different components: a polymeric MAO fraction, with a
chemical formula often written as [−Al(CH3)O−]n or as
[Al(CH3)1.4−1.5O0.75−0.80]n, and residual TMA.2,11,15−18 The
molecular weights assumed or proposed for MAO range typically
from about 700 (ca. 12 Al atoms)1a,15a to aggregates of 150−200
Al atoms (MW 9000−18000).6h Although many structures have
been proposed without reaching a consensus, cages with 4-
coordinated Al and 3-coordinated O centers are the most
favored,19 including nanotube-like structures.2,16 This was
Received: April 5, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
supported by DFT calculations13,15 and multinuclear NMR
studies.3,4 Barron synthesized a number of tert-butyl aluminox-
ane clusters as MAO models, such as [tBuAl(μ3-O)]6, which were
amenable to structural characterization.20 They were able to
activate metallocenes to give catalysts of modest activity. The
catalytic activity was found to be improved by adding TMA.21
TMA Determination. The TMA content has a strong
influence on the catalytic properties of MAO. High amounts of
TMA cause a decrease in catalyst activity7,22,23 and the molecular
weight of the polymer, in addition to having a strong influence on
copolymerizations.15a,24 TMA may be trapped by adding 2,6-di-
tert-butyl-4-methylphenol (BHT) to MAO solutions to mini-
mize chain transfer to aluminum and the leaching of the
supported catalysts promoted by TMA.25−27 For this reason, an
accurate determination of the residual TMA content is crucial,
especially considering that it is the only species clearly
recognizable spectroscopically in MAO solutions. Various
methods have been reported in the literature, but a systematic
comparison and a study on their reliability and reproducibility are
lacking. The TMA content can be determined by visual titration
with pyridine in the presence of phenazine as an indicator.28 This
method, which is still widely used, can give exaggerated TMA
values due to the reaction of pyridine not only with TMA but also
with MAO polymers.18 NMR techniques have proven to be more
reliable and faster to use. The direct quantification of TMA in the
1H NMR spectrum of MAO solutions is difficult due to
overlapping TMA and MAO signals, even if estimates can be
obtained.29 The TMA content can be better measured by adding
THF to give the adduct Me3Al·THF, the methyl protons of
which are shifted to higher field in comparison to those of TMA,
making signal integration easier.18 However, the influence of the
quantity of THF added on the TMA value obtained has not been
explored. Another method for TMA determination consists of
adding PPh3 in known excess.30 The 31P NMR chemical shift is
the weighted average of free and Me3Al-bound PPh3, and the
TMA content is determined using a calibration curve. However,
there are at least two different calibration curves in the
literature,4b,30 and a comparison of this method with that
based on THF was not established.
containing most of the MAO. This process is reversible: i.e., the
lower phase redissolves slowly on addition of TMA and
toluene.1a,32
Activation Process. The ability of MAO to activate catalysts
depends on the presence of Lewis acidic sites. EPR spectroscopic
studies with TEMPO (2,2,6,6-tetramethylpiperidine-N-oxyl) as a
spin probe have suggested at least two types of such sites,
attributed to −OAlMe2 and −O2AlMe structure fragments.12a
Since cages based on 4-coordinate Al contain no coordinatively
unsaturated centers, cleavage of an Al−O bond of a strained
Al2O2 4-membered ring during the catalyst activation step has
been suggested (“latent Lewis acidity”).20 More recently, a patent
proposed another catalyst activation mode by invoking the
formation of [AlMe2]+ cations during the interaction of
metallocene catalyst precursors with MAO,33,34 a suggestion
that follows from the earlier preparation by Klosin et al. of
[AlMe2(OEt2)2]+ salts and their use as catalyst activators.35 It
should be noted, however, that nonsolvated [AlMe2]+ is
extremely unstable and reactive.36
Further information about the structure of MAO and of the
mode of action of this industrially important activator is therefore
urgently needed. We report here a multidisciplinary approach to
determining the role of TMA in MAO formulations (“free”
versus “bound”) and the species involved in catalyst activation by
chemical, NMR spectroscopic, and neutron-scattering methods,
coupled with a computational interrogation of MAO models and
their propensity for ionization into [MAO]− and [AlMe2]+.
RESULTS AND DISCUSSION
■
Interaction of MAO with O and N Donors. The first step
in the characterization of MAO is the determination of the TMA
content. The addition of increasing quantities of tetrahydrofuran
(THF, pKb = 5)37 to MAO (30% solution in toluene) was studied
by 1H NMR spectroscopy (Figure 1).
Nature of TMA in MAO. The nature of the TMA−MAO
interaction remains controversial. TMA is present in MAO
solutions as the dimer Al2(μ-Me)2Me4. However, there are
different views of its interaction with MAO. Tritto et al. reported
a low-temperature NMR study on MAO samples showing that
TMA is mainly bound to MAO, and an analysis of the line
broadening of the “sharp” Al−methyl signal suggested that it
consisted of contributions from both free TMA and TMA
binding to and exchanging with MAO.4b In contrast, FT-IR
spectroscopic studies by Ystenes et al. suggested that there was
no reaction between TMA (or Al2Me6) and MAO.15a,31
Literature reports distinguish between two forms of TMA in
MAO: a fraction that can be removed under vacuum (“free”
TMA), and a form of TMA that can only be removed chemically
(“associated” or “bound” TMA).1,2 For the “bound” TMA the
following equilibrium is invoked:13,15b
Figure 1. 1H NMR spectra of MAO with increasing quantities of THF.
The Al/THF ratio is reported as total Al present in the sample. A single
asterisk (*) denotes the AlMe3·THF signal, and two asterisks (**)
denote the [Me2Al(THF)2]+ peak.
THF cleaves Al2Me6 to give AlMe3·THF, with a signal shift to
higher field (ca. δ −0.6), which facilitates integration and
quantification of the TMA content (Table 1). On THF addition
the MAO signal is shifted to higher field. For Al/THF ratios
between 1/1 and 1/10, the calculated TMA content of the
samples of MAO is consistent and reproducible at ca. 3.3 0.1 wt
%. This is in agreement with data by Imhoff et al., who reported
that the optimal ratio Al/THF to determine the TMA content is
1/4.18,38
(AlOMe)n + (m/2)(AlMe3)2 ⇌ (AlOMe)n(AlMe3)m
Moreover, the possibility was suggested that TMA molecules
might be occluded inside the MAO cages, in order to explain why
it is not possible to remove TMA quantitatively under vacuum.
The addition of diethyl ether to MAO solutions in toluene was
reported to lead to phase separation, with the upper phase
containing all the TMA and some MAO and the lower phase
B
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Table 1. TMA and [Me2Al(L)2]+ Content of the MAO Sample
as a Function of L and the TMA/L Ratio (L = THF, py)
L = THF
TMA
L = py
TMA
Al/L
TMA
[Me2AlL2]+
(mol/L)
TMA
(wt %) (mol/L)
[Me2AlL2]+
(mol/L)
ratio (wt %) (mol/L)
6/1
2.8
3.3
3.4
3.2
3.4
3.1
2.7
0.35
0.41
0.42
0.40
0.42
0.39
0.34
0.08
0.12
0.12
0.12
0.12
0.09
0.06
2.2
3.3
3.6
3.4
3.6
2.3
2.1
0.28
0.41
0.45
0.42
0.45
0.29
0.26
0.11
0.11
0.12
0.14
0.16
0.09
0.06
1/1
1/2
1/5
1/10
1/15
1/26
Figure 2. 1H NMR spectra of MAO with dppe (a) and dppp (b, c). The
Al/P ratio is reported as total Al present in the sample.
A newly emerging peak (labeled ** in Figure 1) appears on
THF addition. When [AlMe2(THF)2][MeB(C6F5)3] is added to
the solutions with Al/THF = 1/2 or 1/10, the intensity of this
peak increases while its chemical shift does not change; it is
therefore assigned to the cationic complex [Me2Al(THF)2]+.33
The assignments were further confirmed by calculated 1H NMR
chemical shifts (see the Supporting Information). This signal
originates from a species to high field of TMA, at δ −0.7, and
moves to δ −0.3 at higher THF concentrations;39 the nature of
this species was explored by modeling (vide infra).
Similar results were obtained on adding pyridine (py). Again
formation of a cation, [Me2Al(py)2]+, was observed. Table 1
shows the results of TMA and [Me2AlL2]+ quantification for
THF and pyridine. Within experimental error both methods give
comparable results (3.3 0.1 wt % TMA).
For Al/py = 6/1, the low TMA value can be tentatively
explained by insufficient pyridine to extract the TMA
quantitatively and to shift. For a large excess of pyridine, the
1H NMR spectra show new signals at ca. δ 0.15, −0.30, and
−0.55, indicating the formation of side products, which may
account for the lower TMA values calculated for high [py]. The
concentration of [Me2Al(py)2]+ increases with the [py] for Al/
py ratios between 1/1 and 1/10. These results explain why the
literature method for TMA determination by pyridine/
phenazine titration gives values higher than expected,28 as
pyridine does not simply form a 1/1 complex with TMA. The
higher relative concentrations of [Me2Al(py)2]+ found at high
Al/py ratios indicate the preferential binding of pyridine to the
more Lewis acidic cation-generating center.
The addition of dppp to MAO in the ratio Al/P = 19/1 shows the
expected shift of the TMA peak to higher field (δ −0.4),
accompanied by a new triplet at δ −0.25 (JHP = 4.4 Hz), which we
attribute to the cationic complex [Me2Al(dppp)]+. Two peaks at
δ −15.9 and −18.9 are seen in the 31P NMR spectrum (free
dppp: δ −15.7 in toluene). For comparison, the NMR spectra of
a toluene solution of TMA and dppp (Al/P = 1/1) show a proton
singlet at δ −0.26 and a 31P signal at δ −15.8. The peak at δ −15.9
is therefore attributed to a TMA−dppp complex, while that at δ
−18.9 can be related to the adduct between dppp and MAO.
Adding increasing quantities of dppp to MAO solutions sharpens
1
the H TMA signal, while the triplet turns into a broad singlet
and only one 31P resonance is seen, at δ −19.9, diagnostic of a
rapid exchange of dppp between the various available sites.
Indications for the presence of specific Lewis acidic sites in
MAO capable of preferential binding of donor ligands are not
only found with typical O, N, or P ligands. The addition of 1,2-
difluorobenzene (FF) to MAO also leads to the formation of a
well-defined triplet at δ −0.45 (JHF = 2.4 Hz), although in this
case the chemical shift does not change with increasing [FF] and
there is no formation of a defined dissociated cationic
dimethylaluminum complex in solution (see the Supporting
Information).40
MAO Fractionation. The addition of L = OPPh3, PMe3,
PCy3 leads to the formation of TMA(L) adducts. At higher [L]
these donors also induce the precipitation of the MAO polymer,
leaving only the soluble TMA(L) complex in the supernatant.
The oily MAO phase is soluble in solvents more polar than
toluene, such as 1,2-difluorobenzene (FF) and THF (Figure 3).
Interaction of MAO with Phosphines. Other bases, such
as phosphines, show variable degrees of interaction with MAO.
There is no formation of [Me2Al(L)2]+ cations when L = PPh3 is
added. We have established a new correlation for [TMA]
determinations with PPh3 in non-deuterated toluene (eq 1),
1
The H NMR spectrum in FF of the precipitate obtained with
OPPh3 shows a broad signal between δ 0.2 and −0.8 due to the
MAO component and a sharp peak at δ −0.72 attributable to
(1)
Al/P = −0.2957δ − 1.5378 R = 0.997
where Al indicates the moles of AlMe3(PPh3) and P the moles of
PPh3 added to the solution (see the Supporting Information).
The data are in good agreement with those obtained by the
pyridine and THF methods, showing that the [Me2AlL2]+
cations do not arise from TMA.
The diphosphines 1,1-diphenylphosphinomethane (dppm),
1,2-diphenylphosphinoethane (dppe), and 1,3-diphenylphosphi-
nopropane (dppp) show variable tendencies to form [Me2AlL2]+
adducts (Figure 2). Adding dppm leaves the shape of the MAO
1H NMR signal unaffected, and there is little detectable
interaction. Increasing quantities of dppe sharpen the TMA
signal at δ −0.3, but there is no formation of a cationic species.
Figure 3. 1H (a, b) and 31P (c, d) NMR spectra in THF (a, c) and 1,2-
difluorobenzene (FF; b, d) of the oily precipitate formed during the
addition of OPPh3 to MAO. The asterisk (*) denotes AlMe3·OPPh3.
C
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
AlMe3·OPPh3 (Figure 3b), as confirmed by the reaction of TMA
with OPPh3 (see the Experimental Section). The sharp peak at
ca. δ 39.0 in the 31P NMR spectrum is assigned to the compound
AlMe3·OPPh3 by comparison with Table 2, while the broad band
at ca. δ 40 is most likely due to the OPPh3 coordinated to MAO
(Figure 3d). An analogous situation is observed in THF solution
(Figure 3c).
reveals that for the two lowest concentrations under investigation
the obtained parameters are constant within an experimental
error of 4%. In particular, the constant Dtoluene, which was used as
the internal viscosity standard, ensures a dilute regime and
consequently a solution viscosity equal to that of pure toluene.
To increase the number of experimental data points and
therefore to improve fit stability and quality, we finally performed
a simultaneous fit of all measured concentrations. Here χ(Me) is
set as a global fit parameter (the same value for all
concentrations), whereas D(c) is allowed to vary with
concentration. As Table 2 shows, the simultaneous fit reproduces
the results of the THF addition quantitatively with respect to
χ(Me). This validates the application of PFG experiments to
quantify TMA concentrations within a complex MAO solution.
It is striking that the DTMA value in a dilute MAO solution is
strongly reduced in comparison to that for a TMA solution in
Independent of the ligands used, the TMA content of the
supernatant amounted to 2.6
0.1 wt % and that of the
precipitate 0.8 0.05 wt % of the original MAO. Precipitation
with these donor molecules therefore allows quantification of the
“free” and “bound” TMA fractions: about three-fourths of the
TMA is in toluene-soluble form, while the remaining one-fourth
is associated with or entrapped in the MAO. On the other hand,
precipitation of MAO with a stronger base, the phosphorus ylide
CH2PPh3, leads to the complete separation of TMA from the
MAO polymer fraction (see the Supporting Information); this
base is apparently able to break the MAO cages sufficiently to
allow quantitative removal of any bound TMA.
Determination of the TMA Content by NMR Diffusion
Coefficient and Small-Angle Neutron Scattering (SANS)
Measurements. A series of pulsed field gradient spin echo
(PFG-SE) NMR experiments of MAO solutions in toluene
(Figure 4) allows the quantitative determination of diffusion
toluene (D′TMA = (10.4 0.38) × 10−6 cm2 s−1 and DTMA
=
(12.86 0.28) × 10−6 cm2 s−1 at w(Al) = 1% w/w, respectively).
For dilute solutions it is reasonable to assume that DTMA is
identical with the diffusion coefficient of TMA in toluene. With
regard to the concept of free and associated TMA, this slowing
down of TMA can be explained by a fast exchange of these two
species. The measured value of DTMA is a composition of a
fraction of fast-diffusing TMA and a fraction of TMA with the
same diffusion coefficient as MAO (eq 3). Consequently, a 75%
D′TMA = χ(TMAfree)DTMA + (1 − χ(TMAfree))DMAO
(3)
fraction of TMA diffuses quickly, whereas a 25% fraction of TMA
is bound to MAO. Several experiments with different diffusion
times Δ (50−200 ms) have been performed to estimate the
exchange rate of this process.43 Even at Δ = 50 ms there are no
effects observable on the diffusion behavior of TMA. Therefore,
the exchange process is too fast to be seen directly with PFG
experiments.
Therefore, PFG-NMR experiments of MAO are a means not
only of determining the total TMA content but also of
distinguishing quantitatively between associated and free TMA.
These results are in excellent agreement with those obtained
independently from the phosphine addition studies; both
techniques demonstrate that about 25% of TMA is “bound” or
associated to MAO.41
Figure 4. Stejskal−Tanner plot of a concentration series of MAO in
toluene with a biexponential decay of the signal intensity, clearly
indicating the presence of fast- and slow-diffusing species.
coefficients and TMA and MAO concentrations.41,42 Stejskal−
Tanner plots of a concentration series of MAO (0.08% w/w ≤
w(AlMAO) ≤10.7% w/w; Figure 4) clearly indicate the presence
of a fast and a slow diffusing species for all concentrations. The
peak integral between δ −0.8 and 0.6 of the 1H NMR spectrum
was used as the peak intensity. This area contains MAO signals as
well as the overlapping TMA signal; therefore, they can be
attributed to the slow and the fast diffusing species, respectively.
To analyze the data, a biexponential fit according to the
Stejskal−Tanner plot of a concentration series is necessary (eq
2), where I/I0 is the normalized peak intensity, χ(Me) is the mole
The interpretation of diffusion coefficients D in terms of a
hydrodynamic radius Rh via the Stokes−Einstein equation (Rh ∝
D−1) assuming a spherical structure raises several problems.44
Different calibration methods (see the Supporting Information)
give a hydrodynamic radius for MAO in toluene: Rh(MAO) ≈ 12
0.3 Å. If one assumes a dense spherical shape of MAO
polymers, a comparison of volumes based on this experimental
hydrodynamic radius with the volumes of calculated structures
for methyl-rich MAO [AlMe1.5O0.75]n (i.e., a composition close
to the experimentally determined Al/Me ratio of typical MAO)
suggests that an MAO polymer molecule contains about 80−100
Al atoms (see the Supporting Information). This estimate is
somewhat smaller than that by Babushkin and Brintzinger for
zirconocene cations paired with [Me-MAO]− anions (150−200
Al atoms).6h However, such estimates depend strongly on the
method used to calculate the volume: for example, volumes
circumscribed by a limiting electron density (where the
molecular volume is defined as a volume inside a contour of
0.001 e/bohr3) or volumes taking steric hindrance and geometric
constraints into account (e.g., the cube method, which takes into
TMA(Δ−δ/3)(γδg)2
I
I0
χ(Me)TMA e−D
+ χ(Me)MAO
MAO(Δ−δ/3)(γδg)2
e−D
(2)
fraction of methyl groups with χ(Me)TMA = 1 − χ(Me)MAO, D is
the diffusion coefficient, γ is the gyromagnetic ratio, Δ is the
diffusion length, δ is the gradient pulse length and g is the
gradient.
To take concentration effects into account, in a first approach
each concentration was fitted separately according to eq 2, which
D
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Table 2. Mole Fractions of Methyl Groups and the Diffusion Coefficients of MAO and TMA in a Dilute MAO Sample (w(AlMAO) =
0.8% w/w)
method
THF addition
PFG NMR simultaneous fit
χ(Me)TMA
DTMA, 10−6 cm2 s−1
χ(Me)MAO
DMAO, 10−6 cm2 s−1
0.31
0.69
0.31 0.01
10.40 0.38
0.71 0.01
3.06 0.04
account all solvent-inaccessible cavities within a given
structure).45,46 In fact, a detailed comparison for a well-
characterized compound, the polyhedral oligomeric silsesquiox-
ane isooctyl-POSS, reveals a factor of about 2 between measured
(3053 Å3 by PFG-NMR) and calculated volumes (2304 Å3 by the
cube method, 1662 Å3 by electron density) (see the Supporting
Information). Therefore, if the number of Al atoms per MAO
polymer is deduced from dividing an experimental volume by the
calculated volume per Al structural unit, this uncertainty has to be
taken into account. The number of Al atoms per MAO should
therefore be corrected by the empirically determined factor of
1.8, which results in an average of about 50−60 Al atoms per
MAO polymer.
Determination of the MAO Molecular Weights and
TMA Content by Small-Angle Neutron Scattering (SANS)
Measurements. The hydrodynamic radius obtained by PFG-
SE NMR methods could be confirmed independently by small-
angle neutron scattering (SANS) experiments on a dilute MAO
solution in toluene. Analysis of the SANS data give the radius of
gyration of the MAO-TMA adduct as Rg = 9.3 0.2 Å, which can
be converted into the sphere radius RS = √(5/3)Rg = 12.0 0.3
Å, in excellent agreement with PFG-NMR experiments. The
good agreement between fitted model and experimental SANS
data is shown in Figure 5; details of the analysis can be found in
the Supporting Information.
bound fraction = NTMA(mMAO/MMAO)/(mTMA /MTMA
)
where MMAO and MTMA are molecular masses of MAO and TMA
and mMAO and mTMA are their respective mass fractions in
solution. The resulting value is NTMA ≈ 0.78; a similar number,
NTMA ≈ 0.64, also results from analysis of the slow component in
the PFG NMR data. One can speculate whether this number, i.e.
approximately one TMA per MAO polymer, might indicate that
there is only one special site per MAO polymer which interacts
with TMA and which may be the site responsible for catalyst
activation.
Computational Modeling of the MAO−THF Interac-
tion. As the results discussed above have shown, MAO contains
some structural elements that are capable of reacting with donor
molecules with liberation of [AlMe2L2]+ cations. We have
explored the potential origin of these intriguing species by
modeling the formation of THF−MAO adducts and the
concomitant release of [AlMe2]+ cations.
The reaction of THF with MAO was explored by first focusing
on small MAO polymers, likely produced during the initial steps
of TMA hydrolysis. The MAOs are described by the structural
formula (AlMeO)n(AlMe3)m, and we chose species with (n,m) =
(1,1), (1,2), (1,3), (2,2) as a starting point of the investigation.
After classification of AlMe2 centers in the MAOs according to
their chemical environment by taking into account the
neighboring oxygen atoms, we investigated the reaction of
THF with the centers producing THF−MAO species. In the
classification, C refers to terminal methyl and Cb to bridging
methyl, while the number in parentheses after O gives its
coordination number. The ring suffix refers to an Al species in a
4-membered Al2O2 ring. The MAO models and optimized
molecular structures of the THF−MAOs are illustrated in Figure
6, whereas the calculated reaction energetics and the identified Al
centers are given in Table 3.
The first interaction of Al2Me6 with THF leads to Me3Al(μ-
Me)AlMe2(THF) (Figure 6, (0,2)). This is a local minimum; the
further reaction to 1/2Al2Me6 + AlMe3(THF) is exergonic, with
ΔG = −54 kJ mol−1. THF coordinates to the MAOs via the
formation of an Al−O bond. The adduct formation is
spontaneous for the MAOs except for the (1,3) species and
AlC2O(3)2-ring site in (2,2). The unfavorable reaction with the
AlC2CbO(4) site of the (1,3) MAO originates from the already
preferable 4-coordination of the Al atom in the MAO. Reaction
of an analogous Al center in (1,2) is favorable due to structural
distortion, which produces structurally analogous species for the
THF−(1,2) adducts. The Al−O distance from the Al center to O
in MAO (shown by rAl−O in Table 3) provides a measure of the
feasibility of the formation of [Me2Al(THF)2]+[MAO]− cation−
anion separation. The largest separation is obtained for (1,3),
because the oxygen atom neighboring the Al center is 4-
coordinated and the elongation of this Al−O bond to form a 3-
coordinate O is energetically favorable.16
Figure 5. SANS intensity I(Q) in absolute units (cm−1) vs scattering
vector Q for a dilute MAO solution in toluene.
Moreover, in combination with chemical analysis and NMR
results, SANS results suggest an average molecular weight of
MAO (without “bound” TMA) of 1800
100. Taking the
stoichiometric formula of MAO to be Al(CH3)1.5O0.75, the
number of Al atoms in a MAO polymer can be calculated to be
1800/60 ≈ 30. This number is about 50% less than the number
estimated from our NMR results; however, considering the
approximations involved in obtaining this estimate, the agree-
ment is good.
Anionization potentials, i.e. energies of the reactions MAO →
[AlMe2]+ + [MAO]−, provide a computationally practical way to
probe the feasibility of anion formation. A detailed analysis of
MAOs produced by hydrolysis of TMA is in progress, which will
In addition, we can calculate the average number of TMA
molecules bound to one MAO polymer, NTMA, from
E
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
The rearrangement of the anions following [AlMe2]+
dissociation is an important contributing factor to their stability.
Both −O−AlMe2−O− and −O−AlMe2−μ-Me− structures are
capable of generating [AlMe2]+, with no total energy difference
due to structural rearrangement. Of these, the structure with
(n,m) = (8,2) gives a particularly stable anion; this MAO model
has an ideal environment for [AlMe2]+ abstraction from the
AlC2CbO(4) site. Also, cleavage of [AlMe2]+ removes the only
methyl bridge, leaving a cage consisting solely of 4-coordinate Al,
3-coordinate O, and terminal methyl groups. The resulting
structure has three Al2O2 4-membered rings, all separated by 6-
rings, whereas energetically less favored structures contain edge-
or corner-sharing 4-rings.
CONCLUSIONS
■
One of the most important characteristics of methylalumoxane
that determines its reactivity as a catalyst activator, as well as its
shelf life prior to catalyst preparation, is its trimethylaluminum
content. MAO is a complex substance, and a combined chemical,
spectroscopic, and computational approach has been employed
to try and identify some key characteristics. A comparison
confirmed that the addition of donor molecules such as THF,
pyridine, and triphenylphosphine provides a consistent method
of TMA quantification, while the addition of trialkylphosphines
led to the precipitation of polymeric MAO and the partitioning of
TMA into “free” and “bound” fractions. These results could be
confirmed independently by pulsed field gradient NMR
methods. Measurements of the diffusion coefficients of TMA
and MAO can be used to determine the average hydrodynamic
radius of MAO, data which are in close agreement with those
obtained from small-angle neutron scattering. Conversion of
these radii into molecular volumes for molecules of ill-defined
geometry such as MAO is less straightforward; our estimates
suggest that an MAO molecule contains about 30−60 Al atoms.
It could be shown that MAO contains small amounts of
structures which, on reaction with suitable bases, can lead to the
dissociation of [AlMe2L2]+ cations. Computational modeling has
shown that these structures are likely to be associated with Al2(μ-
Me)(μ-O) and Al2(μ-O)2 4-rings. Similar processes are highly
likely involved in catalyst activation. The computations do
indeed indicate that for larger MAO cages this ionization is
becoming increasingly energetically feasible. The ability of the
resulting anion to relax into a more stable cage structure
contributes crucially to the overall energetics of this process, and
cages where the 4-membered rings are isolated from one another
by 6-rings are particularly favored. These studies lay the
foundation for more detailed consideration of these processes
on larger MAO models.
Figure 6. Models for (AlMeO)n(AlMe3)m MAOs labeled according to
(n,m) and structures of the MAO + THF → MAO−THF reaction
products. Structural isomers are distinguished by a and b. H, C, O, and
Al atoms are illustrated in white, gray, red, and pink, respectively.
Table 3. RI-MP2/TZVP-Calculated Standard Gibbs Energies
of the Reaction MAO + THF → MAO−THF (T = 298 K),
(AlMeO)n(AlMe3)m MAOs, Chemical Environments of the Al
Centers, and Al Center−O(MAO) Distances (rAl−O
)
a
(n,m)
Al center
ΔG (kJ/mol)
rAl−O (Å)
EXPERIMENTAL SECTION
■
b
(0,2) TMA dimer
(1,1)
AlC2C2
−3.3
−74.0
−92.7
−87.7
1.0
Materials and Techniques. All manipulations were conducted
under an inert atmosphere of dry argon using standard Schlenk
techniques. All reagents were used as purchased without further
purification unless otherwise stated. A solution of MAO (30 wt % MAO
+ TMA in toluene) was provided by Chemtura Organometallics GmbH.
It was stored at −20 °C, and it was 10 months old at the time of analysis.
Solvents were dried over Na/benzophenone (tetrahydrofuran) or
sodium (toluene) before use and purged with argon. 1,2-Difluor-
obenzene (FF, Apollo Scientific), extra-dry pyridine (Py, Acros), and
PMe3 (1 M in toluene, Sigma Aldrich) were degassed with a stream of
argon and stored over activated 4 Å molecular sieves. Deuterated NMR
solvents (CD2Cl2 and toluene-d8) were degassed by several freeze−thaw
cycles and dried over activated 4 Å molecular sieves. NMR spectra were
recorded using a Bruker Avance DPX-300 spectrometer. 1H (300 MHz)
AlC2O(2)
1.737
1.793
1.796
1.903
1.791
1.856
(1,2)-a
AlC2O(3)
(1,2)-b
AlC2CbO(4)
AlC2CbO(4)
AlC2O(3)
(1,3)
(2,2)-a
−88.0
5.6
(2,2)-b
AlC2O(3)2-ring
a
Where Al is bound to two oxygen atoms, an average of the two Al−O
bond distances is reported.
provide potential MAO sources for [AlMe2]+ abstraction.47
Figure 7 summarizes key preliminary findings regarding the
anionization potentials of (AlOMe)n(AlMe3)m, where n ≤ 8.
F
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 7. Anionization potentials of selected MAOs, (AlOMe)n(AlMe3)m, relative to the TMA dimer: Cb, bridging methyl; O(n), n-coordinate O;
purple Al, site of AlMe2+ dissociation.
and 31P NMR (121.5 MHz) chemical shifts were referenced to the
residual solvent peaks of two separate external standards of CD2Cl2 and
of H3PO4 (85% in H2O), each contained in a capillary. 1,3,5-Tri-tert-
butylbenzene was used as an internal standard. In order to compensate
Computational Details. The MAO−THF calculations were
performed by the RI-MP2 method48−51 in combination with the def-
TZVP basis set,52 as implemented in TURBOMOLE version 6.3.53 The
structures were fully optimized and verified as true local minima by
determining their vibrational frequencies, also required for calculation of
Gibbs energies. No scaling factors were employed. NMR calculations
were performed at the MP2/TZVP level of theory using the gauge-
independent atomic orbital (GIAO) method as implemented in
Gaussian09.54 Fast rotation of the methyl groups was assumed in the
NMR peak visualization. Anionization potentials, i.e. relative energies
for the reaction MAO → [AlMe2]+ + [MAO]−, were calculated by the
density functional M062X/TZVP method.55 Molecular volumes were
defined as a volume inside a contour of 0.001 e/bohr3 density and were
calculated at the HF/SVP level56 of theory for MAO models up to 148
Al atoms. The volumes and Cartesian coordinates of the MAOs are
given in the Supporting Information.
1
for sloping baselines and peak overlaps, the H NMR signals were
integrated by printing the spectra on A4 sheets of paper (80 g/m2),
cutting the peaks, and weighing on an analytical balance. Pulse delays
(D1) of 25 and 16 scans were used for 1H NMR spectra. The D1 value
was chosen in line with the longer proton NMR relaxation times
reported by Imhoff et al. for MAO solutions.3
Diffusion experiments were performed on a Varian Inova 400
spectrometer. All measurements were carried out at 293 1 K; during
an experiment the temperature was stable within 0.1 K. The
Dbppste_CC (DOSY bipolar pulse pair stimulated echo with
convection compensation) pulse program was used.4 The diffusion
time Δ was adjusted (typically 20 and 200 ms); the gradient pulse length
δ was 2 ms. The number of transients was adjusted to the concentration.
Small-angle neutron scattering (SANS) experiments were performed
with a KWS-2 instrument (JCNS Outstation, FRM II, Munich,
Germany). We measured an MAO solution in a toluene/d8-toluene
mixture with the following concentrations (w/w): MAO, 1.39%; TMA,
0.17%; toluene, 3.8%; toluene-d8 94.57%.
Reactions and Syntheses. Determination of the TMA Content in
30% MAO by THF and Py. MAO (2 mL, 9.72 mmol Al) and THF (1.6−
250 mmol) were stirred for 30 min at room temperature. An aliquot (0.5
mL) of the resulting solution was transferred to an NMR tube
containing a CD2Cl2 capillary and 1,3,5-tri-tert-butylbenzene (10 − 40
μL, 0.5 M in toluene), and 1H NMR spectra were recorded.
G
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Determination of TMA in MAO by PPh3. MAO (2 mL, 9.72 mmol
Al) and PPh3 (1.18 g in 5 mL of toluene, 4.5 mmol) were stirred at room
temperature for 1 h. An aliquot (50 μL) was transferred to an NMR tube
containing toluene (0.5 mL) and capillaries with CD2Cl2 and H3PO4
(85%) to record 1H and 31P NMR spectra. The calibration curve (Figure
S2, Supporting Information) was obtained by plotting the 31P NMR
chemical shifts of standard TMA/PPh3 mixtures with Al/P molar ratios
between 1/10 and 1/1. The following linear relationship was found:
(4) (a) Tritto, I.; Sacchi, M. C.; Locatelli, P.; Li, S. X. Macromol. Chem.
Phys. 1996, 197, 1537. (b) Tritto, I.; Mealares, C.; Sacchi, M. C.;
Locatelli, P. Macromol. Chem. Phys. 1997, 198, 3963.
(5) Smith, G. M.; Malpass, D. B.; Bernstein, S. N.; Jones, P. D.;
Monfiston, D. J.; Rogers, J.; Su, B. M.; Palmaka, S. W.; Tirendi, C. F.;
Crapo, C. C.; Deavenport, D. L.; Hodges, J. T.; Ngo, R. H.; Post, C. W.;
Rabbitt, C. S. Proc. Polyolefins X - 10th Internat. Conf. Polyolefins 1997, 49.
(6) For equilibria in MAO-activated metallocene catalysts see for
example: (a) Tritto, I.; Li, S. X.; Sacchi, M. C.; Locatelli, P.; Zannoni, G.
Macromolecules 1995, 28, 5358. (b) Tritto, I.; Donetti, R.; Sacchi, M. C.;
Locatelli, P.; Zannoni, G. Macromolecules 1997, 30, 1247. (c) Babushkin,
D. E.; Semikolenova, N. V.; Zakharov, V. A.; Talsi, E. P. Macromol. Chem.
Phys. 2000, 201, 558. (d) Bryliakov, K. P.; Semikolenova, N. V.; Yudaev,
D. V.; Ystenes, M.; Rytter, E.; Zakharov, V. A.; Talsi, E. P. Macromol.
Chem. Phys. 2003, 204, 1110. (e) Bryliakov, K. P.; Semikolenova, N. V.;
Yudaev, D. V.; Zakharov, V. A.; Brintzinger, H. H.; Ystenes, M.; Rytter,
E.; Talsi, E. P. J. Organomet. Chem. 2003, 683, 92. (f) Wieser, U.;
Schaper, F.; Brintzinger, H. H. Macromol. Symp. 2006, 236, 63.
(g) Babushkin, D. E.; Naundorf, C.; Brintzinger, H. H. Dalton Trans.
2006, 4539. (h) Babushkin, D. E.; Brintzinger, H. H. J. Am. Chem. Soc.
2002, 124, 12869. (i) Bryliakov, K. P.; Semikolenova, N. V.; Panchenko,
V. N.; Zakharov, V. A.; Brintzinger, H. H.; Talsi, E. P. Macromol. Chem.
Phys. 2006, 207, 327. (j) Babushkin, D. E.; Brintzinger, H. H. Chem. Eur.
J. 2007, 13, 5294. (k) Bryliakov, K. P.; Talsi, E. P.; Voskoboynikov, A. Z.;
Lancaster, S. J.; Bochmann, M. Organometallics 2008, 27, 6333.
(7) (a) Kleinschmidt, R.; van der Leek, Y.; Reffke, M.; Fink, G. J. Mol.
Al/P = −0.2957δ − 1.5378 R = 0.997
where Al indicates the moles of AlMe3(PPh3) and P the moles of PPh3
added to the solution.
Reaction of MAO with OPPh3, PCy3, and PMe3. OPPh3, PCy3, or
PMe3 (0.5−5 mL, 1 M in toluene) and MAO (30% in toluene; 2 mL,
9.72 mmol of Al) were stirred at room temperature for 30 min. A clear
colorless oil precipitated, and the suspension was decanted. An aliquot
(0.5 mL) of the supernatant solution was studied by 1H and 31P NMR
using CD2Cl2 and H3PO4 (85%) capillaries and 1,3,5-tri-tert-
butylbenzene as internal standard (10−40 μL, 0.5 M in toluene). The
oil was washed with toluene (2 × 5 mL) and dissolved in THF or 1,2-
difluorobenzene (2 mL). An aliquot of the solution was transferred to an
NMR tube, again containing CD2Cl2 and H3PO4 (85%) capillaries and
1
1,3,5-tri-tert-butylbenzene (20 μL, 0.5 M in toluene). The H and 31P
NMR spectra were recorded.
1H NMR of TMA(L): L = PMe3, δ −0.41, C6D6;8 L = PCy3, δ −0.19,
C6D6;9 L = OPPh3, δ− 0.3, toluene. In the case of PMe3, a broad 31P
NMR signal at δ −8.0 is observed which is not a weighted average of
PMe3 and AlMe3(PMe3). This signal may be due to equilibria involving
OPMe3 (δ 36.2, C6D6), from the partial oxidation of PMe3. Adding
OPPh3 (δ(31P) 24.3) to MAO results in a slightly broad 31P signal at δ
27.0, due to a fast exchange between OPPh3 and AlMe3(OPPh3).
Catal. A: Chem. 1999, 148, 29. (b) Jungling, S.; Mulhaupt, R. J.
̈
̈
Organomet. Chem. 1995, 497, 27. (c) Reddy, S. S.; Radhakrishnan, K.;
Sivaram, S. Polym. Bull. (Berlin) 1996, 36, 165.
(8) Matsui, S.; Mitani, M.; Saito, J.; Tohi, Y.; Makio, H.; Matsukawa,
N.; Takagi, Y.; Tsuru, K.; Nitabaru, M.; Nakano, T.; Tanaka, H.;
Kashiwa, N.; Fujita, T. J. Am. Chem. Soc. 2001, 123, 6847.
(9) See for example: Quevedo-Sanchez, B.; Nimmons, J. F.; Coughlin,
E. B.; Henson, M. A. Macromolecules 2006, 39, 4306.
ASSOCIATED CONTENT
* Supporting Information
■
S
(10) Pasynkiewicz, S. Polyhedron 1990, 9, 429.
Text, figures, and tables giving experimental and spectroscopic
details, details for SANS and PFG NMR measurements,
optimized structures for calculated molecular volumes, and
coordinates for calculated structures. This material is available
free of charge via the Internet at http://pubs.acs.org.
(11) (a) Panchenko, V. N.; Zakharov, V. A.; Danilova, I. G.; Paukshtis,
E. A.; Zakharov, I. I.; Goncharov, V. G.; Suknev, A. P. J. Mol. Catal. A:
Chem. 2001, 174, 107. (b) Zakharov, V. A.; Zakharov, I. I.; Panchenko,
V. N. In Organometallic Catalysts and Olefin Polymerization; Springer:
Berlin, 2001; pp 63−71.
(12) (a) Talsi, E. P.; Semikolenova, N. V.; Panchenko, V. N.; Sobolev,
A. P.; Babushkin, D. E.; Shubin, A. A.; Zakharov, V. A. J. Mol. Catal. A:
Chem. 1999, 139, 131. (b) Pedeutour, J. N.; Radhakrishnan, K.; Cramail,
H.; Deffieux, A. Macromol. Rapid Commun. 2001, 22, 1095.
(13) (a) Zurek, E.; Ziegler, T. Prog. Polym. Sci. 2004, 29, 107.
(b) Zurek, E.; Ziegler, T. Inorg. Chem. 2001, 40, 3279.
(14) (a) Glaser, R.; Sun, X. J. Am. Chem. Soc. 2011, 133, 13323. (b) van
Rensburg, W. J.; van den Berg, J. A.; Steynberg, P. J. Organometallics
2007, 26, 1000.
AUTHOR INFORMATION
Corresponding Author
*M.B.: tel, +44 1603 592044; e-mail, m.bochmann@uea.ac.uk.
M.L.: e-mail, mikko.linnolahti@uef.fi. J.S.: e-mail, j.stellbrink@
fz-juelich.de.
■
Notes
The authors declare no competing financial interest.
(15) (a) Ystenes, M.; Eilertsen, J. L.; Liu, J. K.; Ott, M.; Rytter, E.;
Stovneng, J. A. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 3106.
(b) Zakharov, I. I.; Zakharov, V. A.; Potapov, A. G.; Zhidomirov, G. M.
Macromol. Theory Simul. 1999, 8, 272.
(16) Linnolahti, M.; Severn, J. R.; Pakkanen, T. A. Angew. Chem., Int.
Ed. 2008, 47, 9279.
ACKNOWLEDGMENTS
■
This work was supported by the European Commission (Grant
NMP4-SL-2010-246274). We thank Chemtura Organometallics
GmbH for samples of methylaluminoxane. The computations
were made possible by use of the Finnish Grid infrastructure
resources.
(17) Hansen, E. W.; Blom, R.; Kvernberg, P. O. Macromol. Chem. Phys.
2001, 202, 2880.
(18) Imhoff, D. W.; Simeral, L. S.; Sangokoya, S. A.; Peel, J. H.
Organometallics 1998, 17, 1941.
REFERENCES
■
(19) For an early example of this structural type see: Atwood, J. L.;
Hrncir, D. C.; Priester, R. D.; Rogers, R. D. Organometallics 1983, 2, 985.
(20) (a) Harlan, C. J.; Bott, S. G.; Barron, A. R. J. Am. Chem. Soc. 1995,
117, 6465. (b) Mason, M. R.; Smith, J. M.; Bott, S. G.; Barron, A. R. J.
Am. Chem. Soc. 1993, 115, 4971. (c) Harlan, C. J.; Mason, M. R.; Barron,
A. R. Organometallics 1994, 13, 2957.
(1) (a) Sinn, H. Macromol. Symp. 1995, 97, 27. (b) Kaminsky, W.
Macromol. Symp. 1995, 97, 79. (c) Sinn, H.; Kaminsky, W.; Vollmer, H.
J.; Woldt, R. Angew. Chem., Int. Ed. Engl. 1980, 19, 390. (d) Sinn, H.;
Kaminsky, W. Adv. Organomet. Chem. 1980, 18, 99.
(2) Severn, J. R. In Tailor-Made Polymers. Via Immobilization of Alpha-
Olefin Polymerization Catalysts; Wiley-VCH: Weinheim, Germany,
2008, p 95.
(21) Watanabi, M.; McMahon, C. N.; Harlan, C. J.; Barron, A. R.
(3) Babushkin, D. E.; Semikolenova, N. V.; Panchenko, V. N.; Sobolev,
A. P.; Zakharov, V. A.; Talsi, E. P. Macromol. Chem. Phys. 1997, 198,
3845.
Organometallics 2001, 20, 460.
(22) Ghiotto, F.; Pateraki, C.; Severn, J. R.; Friederichs, N.; Bochmann,
M. Dalton Trans. 2013, DOI: 10.1039/C3DT00107E.
H
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(23) (a) Bochmann, M.; Lancaster, S. J. Angew. Chem., Int. Ed. Engl.
1994, 33, 1634. (b) Bochmann, M.; Lancaster, S. J. J. Organomet. Chem.
1995, 497, 55.
(24) (a) Wu, Q.; Ye, Z.; Gao, Q. H.; Lin, S. G. Macromol. Chem. Phys.
1998, 199, 1715. (b) Reddy, S. S.; Sivaram, S. Prog. Polym. Sci. 1995, 20,
309. (c) Reddy, S. S.; Shashidhar, G.; Sivaram, S. Macromolecules 1993,
26, 1180. (d) Chien, J. C. W.; Wang, B.-P. J. Polym. Sci., Part A: Polym.
Chem. 1988, 26, 3089.
(25) Tynys, A.; Eilertsen, J. L.; Seppala, J. V.; Rytter, E. J. Polym. Sci.,
Part A: Polym. Chem. 2007, 45, 1364.
(26) Busico, V.; Cipullo, R.; Cutillo, F.; Friederichs, N.; Ronca, S.;
Wang, B. J. Am. Chem. Soc. 2003, 125, 12402.
(27) Turunen, J. P. J.; Pakkanen, T. T. J. Appl. Polym. Sci. 2006, 100,
(49) Weigend, F.; Haser, M.; Patzelt, H.; Ahlrichs, R. Chem. Phys. Lett.
1998, 294, 143.
(50) Hattig, C.; Hellweg, A.; Kohn, A. Phys. Chem. Chem. Phys. 2006, 8,
1159.
(51) Hattig, C.; Weigend, F. J. Chem. Phys. 2000, 113, 5154.
(52) Schafer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829.
̈
(53) TURBOMOLE V6.3 2011, a Development of University of
Karlsruhe and Forschungszentrum Karlsruhe GmbH, 1989−2007,
TURBOMOLE GmbH, since 2007; available from http://www.
Turbomole.Com.
(54) Frisch, M. J., et al. Gaussian 09, Revision A.02; Gaussian, Inc.,
Wallingford, CT, 2009.
(55) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215.
(56) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571.
̈
4632.
(28) Jordon, D. E. Anal. Chem. 1968, 40, 2150.
(29) Resconi, L.; Bossi, S.; Abis, L. Macromolecules 1990, 23, 4489.
(30) Barron, A. R. Organometallics 1995, 14, 3581.
(31) Eilertsen, J. L.; Rytter, E.; Ystenes, M. Vib. Spectrosc. 2000, 24, 257.
(32) Bliemeister, J.; Hagendorf, W.; Harder, A.; Heitmann, B.;
Schimmel, I.; Schmedt, E.; Schnuchel, W.; Sinn, H.; Tikwe, L.;
Vonthienen, N.; Urlass, K.; Winter, H.; Zarncke, O. In Ziegler Catalysts:
Recent Scientific Innovations and Technological Improvements; Springer:
Heidelberg, Germany, 1995; pp 57−82.
(33) Luo, L.; Sangokoya, S. A.; Wu, X.; Diefenbach, S. P.; Kneale, B.
U.S. Pat. Appl. 2009/0062492 A1 (2009, Albemarle Corp.).
(34) Another recent patent refers to an “unknown species present in
MAO” observed after THF addition, i.e. [AlMe2(THF)2]+, and
correlates the concentration of this species and of TMA with the
composition and molecular weights of the ethylene copolymers:
Crowther, D. J.; Fiscus, D. M. PCT Int. Appl. WO2012/096698
(2012, Exxon Mobil).
(35) (a) Klosin, J.; Roof, G. R.; Chen, E. Y.-X. Organometallics 2000,
19, 4684. (b) Klosin, J. PCT Int. Appl. WO2000/11006 (2000, Dow
Chemical Co.). (c) Chen, E. Y.-X.; Kruper, W. J.; PCT Int. Appl. WO
2001/5642 (2001, Dow Chemical Co.).
(36) Bochmann, M.; Sarsfield, M. J. Organometallics 1998, 17, 5908.
(37) Yamashita, Y.; Kasahara, H.; Suyama, K.; Okada, M. Makromol.
Chem. 1968, 117, 242.
(38) For Al/THF = 6/1, the value of TMA is slightly lower, probably
due to the fact that there is not enough THF to extract all the TMA
(“free” and “bound”) or due to the partial overlapping of the
AlMe3·THF signal with an unidentified species (see the shoulder of
the * signal in Figure 1; Al/THF = 6/1), which affects the integration
value.
1
(39) Similar solvent-induced H NMR shifts of the Al-Me signal are
observed for [AlMe2(THF)2][MeB(C6F5)3]: δ −0.94 ppm in toluene,
−0.25 ppm in THF, and −0.45 ppm in toluene/THF (1/2 v/v).
(40) Attempts to generate [AlMe2(1,2-difluorobenzene)]+ cations
from AlMe3 and B(C6F5)3 in difluorobenzene/toluene produced no
reaction at room temperature and only decomposition products at 40
°C.
(41) The measured diffusion coefficient of MAO, DMAO = (3.06
0.04) 10−6 cm2 s−1, at high dilution is in excellent agreement with the
results of Hansen et al. (DMAO = (3.1 0.1) 10−6 cm2 s−1); however,
these authors decreased the TMA content to χ(MeTMA) = 0.09 0.01
and completely neglected the influence of the residual TMA.17
(42) Rocchigiani, L.; Busico, V.; Pastore, A.; Macchioni, A. Dalton
Trans. 2013, DOI: 10.1039/C3DT00041A.
(43) Karger, J.; Pfeifer, H.; Heink, W. In Advances in Magnetic
̈
Resonance; Waugh, J., Ed.; Academic Press: New York, 1988; Vol. 12, pp
1−89.
(44) Macchioni, A.; Ciancaleoni, G.; Zuccaccia, C.; Zuccaccia, D.
Chem. Soc. Rev. 2008, 37, 479.
(45) Muller, J. J. J. Appl. Crystallogr. 1983, 16, 74.
̈
(46) Kusmin, A.; Lechner, R. E.; Kammel, M.; Saenger, W. J. Phys.
Chem. B 2008, 112, 12888.
(47) Linnolahti, M.; Laine, A.; Pakkanen, T. A. Chem. Eur. J. 2013, 19,
7133.
(48) Hattig, C. J. Chem. Phys. 2003, 118, 7751.
I
dx.doi.org/10.1021/om4002878 | Organometallics XXXX, XXX, XXX−XXX