The extraordinary cocatalytic action of polymethylaluminoxane (MAO) in the polymerization of terminal olefins by metallocenes: Chemical change in the Group 4 metallocene dimethyl derivatives induced by MAO

Article abstract of DOI:10.1002/ejoc.200500371

In the polymerization of olefins with Group 4 metallocene dichlorides or dimethyl derivatives as procatalysts the use of polymethylaluminoxane (MAO) as the cocatalyst, especially in extreme excess (10²-10³ times the metallocene equivalent), has been shown to have an extraordinary accelerating effect on the rate of olefin polymerization, when compared with the cocatalytic action of alkylaluminum halides. In attempts at explaining the greatly superior catalytic activity of MAO in olefin polymerization (the MAO conundrum), hypotheses have generally paralleled the steps involved in the cocatalytic action of R_nAlCl_3-n, namely the alkylation of Cp₂M_tCl₂, ionization of Cp₂M _t(R)Cl into the metallocenium cation, [Cp₂M _t-R]⁺, and anion, [R_n-1AlCl_4-n] ^- and subsequent ion-pair separation. In order to understand any differences in catalytic action between such cocatalysts, we have studied the individual action of MAO (100 equiv.) and of MeAlCl₂ (1-2 equiv.) on each of the Group 4 metallocene derivatives, Cp₂TiCl₂, Cp₂ZrCl₂, Cp₂Ti(CH₃)₂ and Cp₂Zr(CH₃)₂. With MeAlCl₂ each of the metallocene derivatives appeared to form the cation, [Cp₂M _t-CH₃]⁺, with greater (Ti) or lesser (Zr) ease, because an alkyne such as diphenylacetylene was then found to insert into the M_t-CH₃ bond stereoselectively. In striking contrast, treatment of each metallocene with MAO gave two reactions very different from MeAlCl₂, namely a steady evolution of methane gas upon mixing and a finding upon hydrolytic workup that the diphenylacetylene present had undergone no insertion into the M_t-CH₃ bond but instead had been reductively dimerized completely to (E,E)-1,2,3,4-tetraphenyl-1,3-butadiene. To account for this astonishing difference in chemical behavior between MAO and MeAlCl₂ in their cocatalytic activation of Group 4 metallocenes to olefin polymerization, it is necessary to postulate a novel, unique sequence of reaction steps occurring between MAO and the metallocene. If one starts with the metallocene dichloride, then the free TMA present in the MAO would generate the Cp₂M_t(CH₃)₂. This metallocene dimethyl derivative, complexed with an oligomeric MAO unit, would undergo a transfer-epimetallation with added olefin or acetylene to form a metallacyclopropane or metallacyclopropene, respectively. With added diphenylacetylene the resulting 2,3-diphenylmetallacyclopropene would be expected rapidly to insert a second alkyne to form the 2,3,4,5-tetraphenyl-1- metallacyclopentadiene. Simple hydrolysis of the latter intermediate would generate (E,E)-1,2,3,4-tetraphenyl-1,3-butadiene while alternative workup with D₂O would give the 1,4-dideuterio derivative of this butadiene. Both such expectations were confirmed by experiment. In the case of added olefin, similar metallacyclopropane and metallacyclopentane intermediates should be produced until ring opening of the latter five-membered ring leads to an open-chain zwitterion, a process having ample precedent in the research of Gerhard Erker. The solution to the MAO conundrum then, namely the extraordinary cocatalytic activity of MAO in olefin polymerization by metallocenes, lies in the unique catalytic activation of the Group 4 metallocene dimethyl derivative, which occurs by transfer-epimetallation of the olefin monomer by the Cp ₂M_t(CH₃)₂-MAO complex. The most advantageous Lewis acidic sites in the MAO-oligomeric mixture for such metallocene-MAO complexation are suggested to be terminal Me₂Al-O- AlMe- segments of an open-chain oligomer. Wiley-VCH Verlag GmbH & Co. KGaA, 2005.

Full text of DOI:10.1002/ejoc.200500371

FULL PAPER
The Extraordinary Cocatalytic Action of Polymethylaluminoxane (MAO) in
the Polymerization of Terminal Olefins by Metallocenes: Chemical Change in
the Group 4 Metallocene Dimethyl Derivatives Induced by MAO[‡]
John J. Eisch,*[a] Peter O. Otieno,[a] John N. Gitua,[a] and Adetenu A. Adeosun[a]
Keywords: Polymethylaluminoxane (MAO) / Group 4 metallocene derivatives / Olefin polymerization / Cocatalysis /
Transfer-epimetallation
In the polymerization of olefins with Group 4 metallocene di-
chlorides or dimethyl derivatives as procatalysts the use of
polymethylaluminoxane (MAO) as the cocatalyst, especially
in extreme excess (102–103 times the metallocene equivalent),
has been shown to have an extraordinary accelerating effect
on the rate of olefin polymerization, when compared with the
cocatalytic action of alkylaluminum halides. In attempts at ex-
plaining the greatly superior catalytic activity of MAO in ole-
fin polymerization (the MAO conundrum), hypotheses have
generally paralleled the steps involved in the cocatalytic ac-
tion of RnAlCl3–n, namely the alkylation of Cp2MtCl2, ioniza-
necessary to postulate a novel, unique sequence of reaction
steps occurring between MAO and the metallocene. If one
starts with the metallocene dichloride, then the free TMA
present in the MAO would generate the Cp2Mt(CH3)2. This
metallocene dimethyl derivative, complexed with an oligo-
meric MAO unit, would undergo a transfer-epimetallation
with added olefin or acetylene to form a metallacyclopropane
or metallacyclopropene, respectively. With added diphenyla-
cetylene the resulting 2,3-diphenylmetallacyclopropene
would be expected rapidly to insert a second alkyne to form
the 2,3,4,5-tetraphenyl-1-metallacyclopentadiene. Simple hy-
tion of Cp2Mt(R)Cl into the metallocenium cation, [Cp2Mt–R]+, drolysis of the latter intermediate would generate (E,E)-
and anion, [Rn–1AlCl4–n]– and subsequent ion-pair separation.
In order to understand any differences in catalytic action be-
tween such cocatalysts, we have studied the individual action
of MAO (100 equiv.) and of MeAlCl2 (1–2 equiv.) on each of
1,2,3,4-tetraphenyl-1,3-butadiene while alternative workup
with D2O would give the 1,4-dideuterio derivative of this buta-
diene. Both such expectations were confirmed by experiment.
In the case of added olefin, similar metallacyclopropane and
the Group 4 metallocene derivatives, Cp2TiCl2, Cp2ZrCl2, metallacyclopentane intermediates should be produced until
Cp2Ti(CH3)2 and Cp2Zr(CH3)2. With MeAlCl2 each of the me-
tallocene derivatives appeared to form the cation, [Cp2Mt–
CH3]+, with greater (Ti) or lesser (Zr) ease, because an alkyne
such as diphenylacetylene was then found to insert into the
Mt–CH3 bond stereoselectively. In striking contrast, treatment
of each metallocene with MAO gave two reactions very dif-
ferent from MeAlCl2, namely a steady evolution of methane
gas upon mixing and a finding upon hydrolytic workup that
ring opening of the latter five-membered ring leads to an
open-chain zwitterion, a process having ample precedent in
the research of Gerhard Erker. The solution to the MAO co-
nundrum then, namely the extraordinary cocatalytic activity
of MAO in olefin polymerization by metallocenes, lies in the
unique catalytic activation of the Group 4 metallocene di-
methyl derivative, which occurs by transfer-epimetallation of
the olefin monomer by the Cp2Mt(CH3)2–MAO complex. The
the diphenylacetylene present had undergone no insertion most advantageous Lewis acidic sites in the MAO–oligomeric
into the Mt–CH3 bond but instead had been reductively di-
merized completely to (E,E)-1,2,3,4-tetraphenyl-1,3-butadi-
ene. To account for this astonishing difference in chemical be-
havior between MAO and MeAlCl2 in their cocatalytic acti-
vation of Group 4 metallocenes to olefin polymerization, it is
mixture for such metallocene–MAO complexation are sug-
gested to be terminal Me2Al–O–AlMe– segments of an open-
chain oligomer.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
nized since the 1970s. Initially the clue to such cocatalysis
was the puzzling observation that small amounts of water
markedly increase the polymerization activity of titanocene
dichloride/alkylaluminum chloride systems.[2,3] Subsequent
research by Sinn and Kaminsky with halogen-free catalysts
typified by zirconocene dimethyl[4a] combined with trimeth-
ylaluminum (TMA) also identified water as a powerful po-
lymerization accelerant.[4b] As an explanation, Sinn and
Kaminsky then proposed that partial hydrolysis of TMA
causes the in situ formation of MAO, which consists of a
mixture of linear and cyclic oligomers of [–AlMe–O–] units
Introduction
The unexpected efficacy of polymethylaluminoxane
(MAO) as a cocatalyst in the polymerization of olefins by
Group 4 metallocene procatalysts has been widely recog-
[‡] Organic Chemistry of Subvalent Transition Metal Complexes,
36. Part 35: Ref.[1]
[a] Department of Chemistry, State University of New York at
Binghamton,
P. O. Box 6000, Binghamton, NY 13902-6000, U. S. A.
Fax: +1-607-777-4865
E-mail: jjeisch@binghamton.edu
4364
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200500371
Eur. J. Org. Chem. 2005, 4364–4371

Chemical Change in the Group 4 Metallocene Dimethyl Derivatives Induced by MAO
FULL PAPER
and functions as the superior cocatalyst because of the
Related studies concerned with generating the zir-
heightened Lewis acidity of an Al center with two O li- conocenium cation, [Cp2Zr–CH3]+ from Cp2Zr(CH3)Cl or
,
gands.[5] Thereupon, they found that MAO preformed from Cp2Zr(CH3)2 by abstraction of anionic Cl or CH3 by silver
the partial hydrolysis of TMA does in fact exhibit unusually ion have likewise viewed such zirconium cations as active
high polymerization activity when added to a Group 4 me- sites for olefin polymerization.[17,18] In such experiments,
tallocene procatalyst. Such catalyst systems were extraordi- however, the [Cp2Zr–CH3]+ generated was solvated by ace-
narily effective both in the polymerization of ethylene into tonitrile and was thus incapable of initiating olefin polyme-
high-density polyethylene and of propylene into atactic rization.
polypropylene.[6,7] Striking in this cocatalysis, however, is
Accordingly, most explanations for the activation of
metallocene dialkyl derivatives, such as
the necessity of employing an extreme excess of MAO (102– Group
4
103 times the metallocene equivalent) to attain such en- Cp2Zr(CH3)2 (14), by MAO (15), have begun by proposing
hanced activity. It is this aspect of its cocatalytic activity the generation of the zirconocenium cation 16 at an Al cen-
that may be termed the MAO conundrum.[8]
ter in 15 having an enhanced Lewis acidity, such as the [–
Initial attempts to solve the MAO conundrum have run O–AlMe–O–] unit mentioned above [Equation (6)].[8]
parallel to studies directed to studying the roles of the alkyl- Thereupon, complex ion-pairing equilibria would be cre-
aluminum halide cocatalyst in the activation of titanocene ated because of the diverse linear and cyclic oligomers pres-
dichloride (1), the titanocene methyl(chloro) derivative (2) ent in MAO and would give rise to a wide variety of anions
or titanocene dimethyl (3) for polymerization.[9–16] Starting 17 of different molecular mass and structure. Copious and
with titanocene dichloride (1) the first role of cocatalyst careful studies of the global structures present in MAO
MexAlCl3–x (4, x = 1, 2 or 3) is proposed to be the methyl- samples, as well as partial structures found in subunits,[19]
ation of 1 with the formation of 2 and/or 3 [Equation (1)]. have furnished much useful detail but no overall structural
The second proposed role exerted by the by-product Lewis elucidation useful in understanding MAO catalysis.[8] But
acid 5, either with 2 or 3, is the generation of titanocenium particularly pertinent to the action of MAO on Group 4
cation 6 as the active site for polymerization initiation metallocenes have been two meticulous NMR studies of
[Equation (2)] by the Lewis-acidic extraction of anionic E catalyst systems, the first involving Cp2Zr(CH3)2 (14) with
by 5.[10–12] Finally, in the third reaction, which is dependent MAO (15)[20] and the second with the procatalysts,
upon solvent polarity and concentration, the resulting ion Cp2TiCl2 (1) and the constrained-geometry system,
pair, 6 and 7, can be involved in equilibria leading to clus- (Me4C5)SiMe2NBuTiCl2, in their individual reactions with
tering (8) [Equation (3)] or providing tight (9) and solvent- MAO (15).[21] In the first study the types and sizes of the
separated (S, 10) ion pairs [Equation (4)]. Multinuclear ion pairs formed at various Zr/MAO ratios were examined
NMR and solvent-polarity evidence support the conclusion and the sizes of the MAO counteranions 17 were estimated
that solvent-separated ion pairs (10) are the most active ini- by pulsed field-gradient NMR spectroscopy. In the second
tiator of polymerization in the Cp2TiCl2/MeAlCl2 catalyst investigation focus was on identifying the most abundant
system.[13,14] The generation of the ion pair, [Cp2Ti–CH3]+ titanium cation present at different Ti/MAO ratios. Such
[AlCl4]– (11), in this system has been corroborated by its studies have indeed identified several cationic intermediates
chemical capture with trimethyl(phenylethynyl)silane (12) present, as well as having estimated the size of MAO anion
as adduct 13, whose 3-D structure has been verified by sin- 17 in the zirconocene system. But the pertinence of such
gle-crystal X-ray determination [Equation (5)].[9]
information to MAO catalysis remains unclear. Both stud-
ies make the unproven assumptions that: 1) The prevalent
ion pairs abundant at Mt/MAO = 1:100–300 are the most
active polymerization sites; and 2) the active site generated
in MAO cocatalysis of metallocenes exactly parallels the
active site generated in MexAlCl3–x cocatalysis, namely a
cation like 6. Thus, the explanation of MAO cocatalysis
arising from the foregoing NMR studies must be viewed as
plausible and consistent rather than convincing and decis-
ive.
In furthering our ongoing studies of Group 13 organome-
tallic Lewis acids,[15,22,23] we have undertaken a comparative
study of the activation of the Group 4 metallocene deriva-
tives, Cp2MtE2, where Mt = Ti or Zr and E = Cl or
Eur. J. Org. Chem. 2005, 4364–4371
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4365

J. J. Eisch, P. O. Otieno, J. N. Gitua, A. A. Adeosun
FULL PAPER
Me, by the individual reactions with MAO (15) and with
MeAlCl2 (18), respectively. A comparison of the alkylating
and Lewis-acidic reactions of 15 and of 18 on the metallo-
cenes, Cp2MtE2, seemed to be reasonable [cf. Equations (1)
and (2)], since both have as their principal structural sub-
unit an Al center bonded in turn to a methyl group and to
two identical electronegative atoms, 15a and 18.
Results and Discussion
Methane Gas Evolution from Solutions of Cp2MtE2
Derivatives 22–25 and MAO (15) in Toluene or
Dichloromethane
Individual solutions of metallocene derivatives 22–25 re-
acted promptly with methane evolution, as determined by
mass spectrometry, when treated at 25 °C with 100 equiv. of
MAO (15) in toluene. It is noteworthy that the dichlorides
22 and 23 gave a moderate but steady methane evolution
while the dimethyl compounds 24 and 25 exhibited an ini-
tially more vigorous and gradually subsiding gas evolution.
A likely explanation is that 22 and 23 are converted into 24
and 25, respectively, by the free Me3Al present in the MAO
and reaction of 24 and 25 with the MAO is the actual
source of the methane (Scheme 1).
But reaction mixtures of any one of the four Cp2MtE2
derivatives with 100 equiv. of MAO in toluene gave a clear
sign of an unexpected reaction, not immediately observable
with solutions of the Cp2MtE2 derivatives with an excess of
MeAlCl2 in CH2Cl2 or in toluene: A steady to vigorous evol-
ution of methane gas. Furthermore, when the resulting solu-
tions were then treated individually with diphenylacetylene
(19), in order to capture the methylmetallocenium ion [cf.
Equation (5)], a second arresting difference in chemical be-
havior between 15 and 18 became apparent. When a 1:2
mixture of Cp2TiCl2 and MeAlCl2 was allowed to react Scheme 1.
with 19 at 25 °C (in toluene or CH2Cl2), hydrolytic workup
A search of the literature then uncovered a pertinent
led to almost a quantitative yield of 1,2-diphenyl-1-propene
(20) as a 9:1 mixture of Z/E isomers [Equation (7)].[24] This
product is the hydrocarbon expected to result from the in-
sertion of 19 into the CH3–Ti bond of 6 and subsequent
hydrolysis. By sharp contrast in the present study, however,
we have now found that when a 1:1 mixture of diphenylace-
tylene and Cp2TiCl2 were allowed to react with 10–
100 equiv. of MAO in toluene at 25 °C, hydrolytic workup
showed that the diphenylacetylene had been completely
consumed but no 1,2-diphenyl-1-propene (20) whatsoever
had been formed. Instead, the diphenylacetylene had been
completely converted into (E,E)-1,2,3,4-tetraphenyl-1,3-bu-
tadiene (21) [Equation (8)]!
prior study by Kaminsky and Steiger in which zirconocene
derivatives 23 and 25 were found consistently to give meth-
ane gas evolution when combined with MAO in toluene.[25]
Indeed, in a comprehensive doctoral investigation Steiger
has investigated the activation of 23 or 25 by MAO as an
olefin polymerization catalyst system and has correlated
that polymerization activity with methane evolution. Meth-
ane evolution was observed to progress with the ageing or
loss of catalyst activity.[26] The source of the methyl group
in such methane evolution has been shown to be zir-
conocene dimethyl (25), by means of a study of the reaction
of 26 with MAO (15), which latter reagent had been pro-
duced from Me3Al and H2O [Equation (9)].[27]
The methane evolved proved to be Ͼ99% CD3H. In the
case of individual mixtures of Cp2TiCl2 (22) or Cp2ZrCl2
(23) with MeAlCl2 in CH2Cl2 or toluene, on the other hand,
no significant gas evolution could be detected at 25 °C over
a period of 24 h. In fact, 1 equiv. of an alkyne, such as 12
or 19, could be carbaluminated in over 90% yield (13 and
20) by 1 equiv. of CpTiCl2 with 2 equiv. of MeAlCl2. This
These two surprising differences in chemical behavior be- finding shows that the Me group of MeAlCl2 is not signifi-
tween mixtures of Cp2TiCl2 and MAO and mixtures of cantly diminished through loss of methane. Many similar
Cp2TiCl2 and MeAlCl2 clearly shows that different chemi- alkylations of alkynylsilanes by combinations of CpTiCl2
cal reactions may be involved in their individual cocatalytic with Me2AlCl, EtAlCl2 or Et2AlCl likewise show no notice-
action in olefin polymerization. The present study has able loss of alkane.[28] As discussed in the next section, how-
sought to elucidate these reaction differences between MAO ever, the reaction of alkynes with [Cp2ZrCl2]/CH3AlCl2
and MeAlCl2 as cocatalysts.
mixtures was surprisingly slow.
4366
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4364–4371

Chemical Change in the Group 4 Metallocene Dimethyl Derivatives Induced by MAO
FULL PAPER
With Me3Al or Et3Al itself, however, the loss of alkane attempted under similar conditions (18 h, 25 °C, CH2Cl2)
can be significant over time. The well-known Tebbe reagent but no addition whatsoever was observed. Only with pro-
(27) results from the reaction of Cp2TiCl2 (22) with an ex- tracted reaction times or with use of (CH3)2AlCl as cocata-
cess of Me3Al in toluene at room temperature over 60 h lyst could low yields of 20 (Ͻ10%) be attained [Equa-
[Equation (10)].[29] Similarly, the reaction of Cp2ZrCl2 (23) tion (12)].
with Et3Al can lead to the 1,2-dizirconio compound 28 with
ethane evolution [Equation (11)].[30]
From these observations one can conclude that ion pair
29b is not as readily generated from 23 by CH3AlCl2 (18)
as is ion pair 29a from 22 and CH3AlCl2 (18). This observa-
tion may explain why 18 and other alkylaluminum chlorides
are relatively poor cocatalysts in olefin polymerizations
with zirconocene procatalysts.[8]
But what has been most disconcerting is the failure of
diphenylacetylene (19) to undergo any such alkyne insertion
whatsoever when 19 is added individually to systems sup-
posedly containing 30 or 31, respectively. Such an observa-
tion could indicate either that such ion pairs are not present
or that they react in an entirely different manner in the pres-
ence of an extreme excess of MAO.
In light of the foregoing observations made both in our
group and in that of Kaminsky, it is uncomprehensible that
no mention is made of methane gas evolution or NMR
spectral instability in either of the foregoing NMR spectro-
scopic studies of the catalyst systems, Cp2Zr(CH3)2 with
MAO[20] or Cp2TiCl2 with MAO.[21] For methane gas evol-
ution not to have been a significant factor in altering such
systems, NMR sample preparation and spectral measure-
ments would had to have been carried out quickly and at
lower temperatures. But the limited experimental descrip-
tions provided[20,21] make further speculation on this point
futile.
Synergetic Interaction of Group 4 Metallocene Dialkyl
Derivatives and MAO with Alkynes: Catalyzed Transfer-
Epimetallation
That the latter possibility obtains is clear from Equa-
tion (8), which shows that diphenylacetylene undergoes ex-
clusively reductive dimerization, rather than insertion into
an Mt–CH3 bond. Further insight has come from the four
individual reactions of Cp2TiCl2, Cp2ZrCl2, Cp2Ti(CH3)2
or Cp2Zr(CH3)2 with one equiv. of 19 and 100 equiv. of
MAO, followed by workup with D2O. The organic product
was in each case exclusively (E,E)-1,4-dideuterio-1,2,3,4-tet-
raphenyl-1,3-butadiene (21a). The latter dideuteriated de-
rivative presupposes the presence of the titanocycle 34, in
the case of Cp2TiE2, which could have arisen from the gen-
eration of titanocene (32), the rapid addition of 32 to alkyne
19 to yield 33 and the further insertion of 19 into 33
(Scheme 2).[31] However, from what else is known about the
stability of titanocene (32), such a divalent carbene-like in-
Attempted Capture of Postulated Methylmetallocenium
Cations in the MeAlCl2/Cp2MtE2 and the MAO/Cp2MtE2
Catalyst Systems
From the interaction of a 1:2 mixture of Cp2TiCl2 (22)
or Cp2ZrCl2 (23) with MeAlCl2 (18), the polymerization-
initiating ion pair 29 has been postulated to be generated.
Likewise, from the two foregoing cited NMR studies of
Cp2Zr(CH3)2 or Cp2TiCl2, each with 200 equiv. of MAO,
the principal ion pairs concluded to be present are 30 and
31.
In order to capture such ions, an alkyne such as alkynyl-
silane 12 or diphenylacetylene (19) has been introduced into
each of the four individual catalyst systems. The success of
both alkyne insertions with 1:2 mixtures of Cp2TiCl2 and
CH3AlCl2, as depicted in Equations (5) and (7), therefore
represents strong corroboration for the presence (or avail-
ability) of ion pair 29a. Contrary to expectations, however,
the reaction of 1:2 mixtures of Cp2ZrCl2 and CH3AlCl2 was Scheme 2.
Eur. J. Org. Chem. 2005, 4364–4371
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4367

J. J. Eisch, P. O. Otieno, J. N. Gitua, A. A. Adeosun
FULL PAPER
termediate would not be expected to persist at room tem- phenylacetylene (19) with complexed ion pair 31 via transi-
perature. For example, the reduction of Cp2TiCl2 by so- tion state 37 [Equation (14)].
dium[32] or by sodium naphthalenide[33] or the reduction of
Cp2Ti(CH3)2 by dihydrogen[34] at 20 °C, each of which reac-
tions was intended to generate 32, led instead to the iso-
meric titanocene(iii) hydride dimer 35 [Equation (13)].[35]
The methyl radicals thereby liberated would attack MAO
or the toluene solvent and thus lead to the evolution of
methane. A similar transfer-epizirconation involving 30 and
19 would lead to the zirconacyclopropene analogue of 33
and, similar to the sequence, 33 Ǟ 34 Ǟ 21 in Scheme 2,
would explain the isolation of 21 from the reaction of 25
and 19.
As corroborating experiments for the model reaction
pathway proposed in Scheme 2 for titanocene dimethyl (24),
the following reactions were carried out: 1) The reaction
between 24 and 1 equiv. of 19 was attempted in refluxing
toluene but neither 21 nor any other product was detected;
2) stirring a toluene solution of 1 equiv. of 19 with
100 equiv. of MAO at 25 °C led to the complete recovery
of 19; and 3) as expected by the model, reaction of 24 with
19 in a 1:2 molar ratio, in the presence of 100 equiv. of
MAO, led to the complete consumption of 19 and the
quantitative formation of diene 21.
By reason of the instability of Cp2Ti at 25 °C, therefore,
we conclude that the formation of the titanacyclopropene
33 is highly unlikely to arise by the direct epititanation of
diphenylacetylene (19) by 32,[36] as shown in Scheme 2. The
instability of 32 can be circumvented, however, if the re-
cently observed process of transfer-epititanation is instead
involved, as exemplified in Scheme 3.[37] In this process
TiCl2 is not liberated from nBu2TiCl2 by prior elimination
of the two butyl radicals (path a) but rather the loss of
butyl groups occurs by way of the octahedral complex 36,
in which the TiCl2 group is simultaneously transferred to
the alkyne. In the latter process of transfer-epititanation, no
free TiCl2 need ever be formed.
Nature of the Interaction of Group 4 Metallocene Dimethyl
Derivatives with MAO
Introduction of diphenylacetylene (19) into a mixture of
Cp2M(CH3)2 (M = Ti, 24; M = Zr, 25) and MAO (15) in a
1:100 equiv. ratio gives no detectable evidence for the gener-
ation of 30 or 31 nor of their capture by 19 to give the
hydrocarbon 20 on hydrolysis. Accordingly, some chemical
reaction between Cp2M(CH3)2 and MAO must be sought
other than a simple M–CH3 heterolysis. If this interaction
of MAO would also depend on the Lewis acidity of specific
MAO subunits, we would propose as the specifically strong
Lewis-acidic site, the bidentate grouping at the ends of
open-chain aluminoxane subunits, namely (CH3)2Al–O–
Al(CH3)–O– (38). Both on steric and electronic grounds, 38
should best be able to form a Lewis complex (39) with both
CH3 groups of 24 and 25 simultaneously. By drawing elec-
tron density from the transition metal toward the methyl
groups, compound 38 should lower the bond-homolysis en-
ergy required to attain transition state 37 and its zirconium
analogue [Equation (15)].
Scheme 3.
Our first attempt to test for the occurrence of such trans-
fer-epimetallation was disappointing: Neither Cp2Ti(CH3)2
In his kinetic analysis of propene polymerization with
nor Cp2Zr(CH3)2 reacted with diphenylacetylene even in re- the Cp2ZrE2/MAO system (E = Cl, CH3) it is interesting to
fluxing toluene. But these attempted transfer-epimetallation note that Steiger also assigns a special importance to the
reactions omit the obvious catalytic role of MAO in the proportion of end units like 38 in MAO. Their importance
reactions of Cp2Ti(CH3)2 (24) or of Cp2Zr(CH3)2 (25), as in his view is their superior methylating action on Cp2ZrCl2
well as the demonstrated fact that 24 and 25 exist princi- and Cp2Zr(CH3)Cl precursors. However, both Steiger and
pally as zwitterion-like intermediates, 30 and 31, respec- our group assign the same reason for the necessity of em-
tively, in the presence of more than 100-fold excess of ploying extreme excesses of MAO for the extraordinary
MAO.[20,21] We therefore propose that the source of titana- catalytic action: Only such an excess will furnish a sufficient
cyclopropene 33 is the transfer-epititanation reaction of di- proportion of the truly catalytically active end groups 38.
4368
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4364–4371

Chemical Change in the Group 4 Metallocene Dimethyl Derivatives Induced by MAO
FULL PAPER
Relevance of Transfer–Epimetallation to the Cocatalytic
Action of MAO in Olefin Polymerization by Titanocene
and Zirconocene Derivatives
from 42 to 43 there is excellent precedent in the work of
Erker’s research group. For example, a mixture of (s-cis-η4-
butadiene)zirconocene (shown as 44) and (s-trans-η4-buta-
diene)zirconocene (not shown) react with the Lewis acid,
(C6F5)3B, to form zwitterionic 45 [Equation (17)], whose
structure has been established by X-ray crystal structure
analysis.[38] In keeping with our suggestion above for 43,
compound 45 immediately initiates polymerization when
exposed to ethylene or propylene gas.
The present study has demonstrated that distinctly dif-
ferent, irreversible reactions take place between titanocene
or zirconocene procatalysts 22–25 and the respective cocat-
alysts, MeAlCl2 (18) or MAO (15). While MeAlCl2 activates
these metallocenes (at least Cp2TiCl2) by the straightfor-
ward reactions of alkyl transfer from Al to Mt, followed by
catalyst ion-pair formation [cf. Equations (2)–(5)], acti-
vation of these same metallocenes by MAO is accompanied
by considerable evolution of methane and the formation of
ion pairs 30 and 31 that do not insert alkynes in the ex-
pected polar manner shown in Equations (3) and (7). In-
stead, these intermediates apparently undergo transfer-epi-
metallation with alkyne 19, for only via this route can the
isolation of the reductive dimer of 19, namely 21, be ex-
plained.
Because of differences shown in the activation of
metallocenes for polymerization by MeAlCl2 or by MAO,
we now suggest that the two cocatalysts follow different ini-
tiating steps. With MeAlCl2 the initiation would involve the
well-established electrophilic attack of the methylmetall-
ocenium cation of ion pair 40 on ethylene followed by many
repetitive insertions [Equation (16)].
In summary, the explanation for the extraordinary co-
catalytic activity of MAO in olefin polymerization with
metallocenes, known as the MAO conundrum, lies in realiz-
ing that MAO activates metallocenes not for the usual in-
sertions of olefins into polar transition metal–carbon bonds
[such as that in Equations (5) and (7)] but rather for homo-
polar transfer-epimetallations of olefins, starting from 30 or
31 and leading successively to the metallacyclopropanes 41,
the metallacyclopentanes 42 and ultimately to the open-
chain zwitteronic initiator 43 (Scheme 4).
Experimental Section
Instrumentation, Analysis and Starting Reagents: All reactions were
carried out under a positive pressure of anhydrous, oxygen-free ar-
gon. All solvents employed with organometallic compounds were
dried and distilled from a sodium metal/benzophenone ketyl mix-
ture prior to use.[39] The IR spectra were recorded with a Perkin–
Elmer instrument (model 457) and samples were measured either
as mineral oil mulls or as KBr films. The NMR spectra (1H and
13C) were recorded with a Bruker spectrometer (model EM-360)
and tetramethylsilane (Me4Si) was used as the internal standard.
The chemical shifts reported are expressed on the δ scale in ppm
from the Me4Si reference signal. The GC/MS measurements and
analyses were performed with a Hewlett–Packard GC 5890/Hew-
lett–Packard 5970 mass-selective-detector instrument. The gas
chromatographic analyses were carried out with a Hewlett–Packard
instrument (model 5880) provided with a 2-m OV-101 packed col-
umn or with a Hewlett–Packard instrument (model 4890) having a
30-m SE-30 capillary column. Melting points were determined with
a Thomas-Hoover Unimelt capillary melting point apparatus and
are uncorrected. In all of the foregoing spectral or physical mea-
surements any air- or moisture-sensitive samples were handled in
glass tubes filled and sealed under dry argon.
With MAO, on the other hand, zwitterionic-like complex
39 would undergo a transfer-epimetallation with ethylene
(accompanied by methane evolution) to form zwitterionic
metallacyclopropane 41. Ring strain should favor rapid in-
sertion of a second ethylene monomer to form zwitterionic
metallacyclopentane 42 (Scheme 4).
Reactions of Group 4 Metallocene Dichlorides/Methylaluminum Di-
chloride Mixtures with Diphenylacetylene (19). (a) Titanocene Di-
chloride (1): According to a previously published general pro-
cedure,[24] which is here modified and specified, a solution of
996 mg (4.0 mmol) of titanocene dichloride (1) and 712 mg
(4.0 mmol) of diphenylacetylene in 100 mL of anhydrous dichloro-
methane was cooled in an ice bath and then treated with 8 mmol
Scheme 4.
At the point the active site for further polymer growth
will likely arise from the ring-opening isomerization of 42,
into open-chain zwitterion 43. For the feasibility of the step of methylaluminum dichloride (18) in hexane. After 18 h of stirring
Eur. J. Org. Chem. 2005, 4364–4371
www.eurjoc.org © 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 4369

J. J. Eisch, P. O. Otieno, J. N. Gitua, A. A. Adeosun
FULL PAPER
at room temp., the reaction mixture was quenched with 1.0 mL of
water. Basic alumina was added to the CH2Cl2 solution, volatiles
were removed under vacuum and the solid residue packed on top
of an alumina column prepared with hexane. Elution with hexane
gave 500 mg of a mixture of 81% of (Z)-1,2-diphenyl-1-propene
(20a) and 19% of (E)-1,2-diphenyl-1-propene (20b) with no reco-
vered 19. When the foregoing procedure was repeated and worked
up with D2O, 20a and 20b were largely (Ͼ90%) deuteriated at their
respective vinylic positions. It should be noted that the precise con-
ditions of hydrolysis or deuterolysis can influence both the ratio of
20a/20b and the extent of vinylic deuteriation. Addition of a mix-
ture of D2O and triethylamine for such a hydrolytic workup gives
a 42:58 mixture of 20a/20b with only a modest incorporation of
deuterium (20a: 28%). These results indicate that hydrolysis of the
C–Ti bond involves a competition between heterolytic and homo-
lytic cleavages. This important finding will be addressed in a subse-
quent publication. When the foregoing procedure was carried out
as described, but using only 4.0 mmol of MeAlCl2 (18), then 50%
except with 200 mg (1.54 mmol) of 19, again upon hydrolytic
workup only diene 21 was quantitatively formed with Ͻ1% of re-
maining 19. (b) Temperature: When any of the foregoing reactions
were conducted at –78 °C by mixing and stirring the components
in a reaction flask cooled by an acetone/solid CO2 mixture, in a
Dewar bath protolysis at –78 °C after 24 h with methanol led to
the complete recovery of 19 with no sign of 21, 20a or 20b. (c)
Deuteriation: When either 22 or 23 was allowed to react with
100 equiv. of MAO in the presence of 1 equiv. of 19 in toluene
at 5 °C for 24 h, according to section (a), workup first with O-
deuterioacetic acid (Ͼ99%) and then with ordinary 6 n aqueous
HCl, the (E,E)-1,2,3,4-tetraphenyl-1,3-butadiene was completely
1,4-dideuteriated (21a), as shown by 1H and 2H NMR spec-
troscopy. (d) Control Experiment: In order to substantiate the nec-
essary presence of all three reactants, 22, 23, 24 or 25 together with
MAO (15) and diphenylacetylene (19), one further control experi-
ment was carried out. Stirring of a toluene solution of 1 equiv. of
19 with 100 equiv. of MAO (15) at room temperature for 24 h led
of 19 remained and 50% of a 24:1 mixture of 20a/20b was formed. to the complete recovery of unchanged 19. (e) Methane Evolution:
(b) Zirconocene Dichloride (23): In an experiment identical with the
titanocene dichloride procedure, 4.0 mmol each of 23 and 19 were
allowed to react with 8 mmol of 18 in CH2Cl2 solution at room
temp. for 18 h. Usual workup revealed that no 1,2-diphenyl-1-pro-
penes (20a and 20b) whatsoever were formed and 19 was recovered
unchanged. Repetition of the reaction attempt with (CH3)2AlCl
and prolongation of the reaction time (Ͼ24 h) did eventually lead
to small amounts of 20a (ca. 10%).
With the metallocene dichlorides 22 and 23, admixture with MAO
and alkyne 19 gave an initially slow but steady and prolonged
methane evolution. On the other hand, with the metallocene di-
methyls 24 and 25, admixture with MAO and alkyne 19 produced
initially a very vigorous gas evolution that gradually became slow
over the first hour. The identification of the gas evolved as Ͼ99%
methane, and not ethane or dihydrogen, was achieved by with-
drawing samples of the evolved gas by a gastight syringe through
a septum covering one neck of the reaction vessel and by injecting
the sample into the aforementioned Hewlett–Packard GC/MS in-
strument.
Attempted Thermal Reactions of Group 4 Metallocene Dimethyl De-
rivatives with Diphenylacetylene (19): Published procedures were ap-
plied for the preparation of titanocene dimethyl (24)[34] and zir-
conocene dimethyl (25),[40,41] respectively. Then, under argon
840 mg of 24 or 1.00 g of 25 (4.0 mmol) in 40 mL of toluene was
added to a solution of 1.43 g (8.0 mmol) of 19 in 20 mL of toluene
at 25 °C. The resulting solution was heated at reflux for 8 h and
then hydrolyzed with 6 n aqueous HCl. Usual ether extraction and
workup led to the quantitative recovery of 19 and no sign of 20a.
Even when the reaction vessel was covered with Al foil to exclude
light, there was no sign of reaction.
Acknowledgments
The senior author is pleased to thank various industrial chemical
sponsors for their financial support of and ongoing confidence in
our research group’s efforts in Ziegler–Natta olefin polymerization.
Initial support was provided some years ago by the Schering Cor-
poration of Bergkamen, Germany (later the Witco and then the
Crompton Corporation). Continuing support during the 1990s
stemmed from the Solvay Corporation, Brussels, Belgium. Cur-
rently we have enjoyed the generous support of the Boulder Scien-
tific Company and a stimulating scientific exchange with its presi-
dent, Dr. John M. Birmingham, concerning the generation of titan-
ocene (32).
Reactions of Group 4 Metallocene Derivatives/MAO Mixtures with
Diphenylacetylene (19). (a) General Procedure: A solution of titano-
cene dichloride (22), titanocene dimethyl (24), zirconocene dichlo-
ride (23) or zirconocene dimethyl (25) (0.57 mmol) in 40 mL of
toluene was treated under argon at 25 5 °C with 57 mmol of
MAO (15) in 12.2 mL of toluene (source of MAO: Sigma–Aldrich,
used as freshly purchased and containing 5 1% of free Me3Al as
determined by 1H NMR spectroscopy). The vigor and extent of
gas evolution (CH4 in each case) was noted. After 5 min, a solution
of 19 (100 mg, 0.57 mmol) in 20 mL of toluene was added to the
reaction mixture. After 24 h of stirring at room temperature, usual
hydrolytic workup with 6 n aqueous HCl, the reaction mixtures of
19 individually with 22, 24, 23 or 25 led in each case to the quanti-
tative isolation of (E,E)-1,2,3,4-tetraphenyl-1,3-butadiene (21) with
no trace of 19 or of the (Z)- and (E)-1,2-diphenyl-1-propenes (20a,
20b). The (E,E)-1,2,3,4-tetraphenyl-1,3-butadiene (21) was further
purified by recrystallization from glacial acetic acid to give white
crystals, m.p. 183–184 °C. The melting point of a mixture with au-
thentic 21 was undepressed and the NMR spectroscopic data of
the sample isolated from each of the four foregoing reactions were
identical with those of authentic 21. 1H NMR (CDCl3): δ = 42–
7.30 (m, 10 H), 7.03–6.99 (m, 6 H), 6.75–6.72 (m, 4 H), 6.30 (s, 2
H) ppm. 13C NMR (CDCl3): δ = 145.60, 139.76, 137.25, 131.64,
130.38, 129.47, 128.77, 127.76, 127.31, 126.58 ppm. When the fore-
going reaction between 19 and 24 was repeated as described above,
[1] J. J. Eisch, S. Dutta, Organometallics 2005, 24, 3355–3358.
[2] K. H. Reichert, K. R. Meyer, Makromol. Chem. 1973, 169,
163–176.
[3] W. P. Long, D. S. Breslow, Justus Liebigs Ann. Chem. 1975,
463–469.
[4] a) In this article Cp2Ti(CH3)2 and Cp2Zr(CH3)2 will be referred
to as “titanocene dimethyl” and “zirconocene dimethyl”,
respectively, in order to stress their structural kinship with their
metallocene dichloride derivatives, Cp2MtCl2. To refer to such
methyl derivatives as “dimethyltitanocene” or “dimethylzir-
conocene” runs the risk of implying that the methyl groups
might be substituents on the metallocene rings. As classes such
derivatives will be referred to as “metallocene dimethyl deriva-
tives” for Cp2Mt(CH3)2 and as “metallocene methyl(chloro) de-
rivatives” for Cp2Mt(CH3)Cl; b) A. F. Andresen, H. G. Cordes,
J. Herwig, W. Kaminsky, A. Merck, R. Mottweiler, J. Pein, H.
Sinn, H. J. Vollmer, Angew. Chem. 1976, 88, 689–690.
[5] H. Sinn, W. Kaminsky, H. J. Vollmer, R. Woldt, Angew. Chem.
1980, 92, 396–402.
4370
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4364–4371

Chemical Change in the Group 4 Metallocene Dimethyl Derivatives Induced by MAO
FULL PAPER
[6]
halten der Produktkomplexe gegenüber Propen, Doctoral Dis-
sertation, University of Hamburg, 1990.
H. Sinn, W. Kaminsky, Adv. Organomet. Chem. 1980, 18, 99–
149.
[27] P. G. Gassman, Ziegler–Natta Symposium, Regional Meeting
of the American Chemical Society, Corpus Christi, Texas, held
during November 30 to December 2, 1988. Both Professor Ka-
minsky and one of the present authors (J. J. E.) attended Pro-
fessor Gassman’s lecture and noted these findings. However,
Professor Gassman has since passed away, so no documenta-
tion of these results has been obtainable.
[28] J. J. Eisch, R. J. Manfre, D. A. Komar, J. Organomet. Chem.
1978, 159, C13–19.
[29] F. N. Tebbe, G. W. Parshall, G. S. Reddy, J. Am. Chem. Soc.
1978, 100, 3611.
[7]
[8]
P. Pino, R. Mülhaupt, Angew. Chem. 1980, 92, 869–887.
H. H. Brintzinger, D. Fischer, R. Mülhaupt, B. Rieger, R. M.
Waymouth, Angew. Chem. Int. Ed. Engl. 1995, 34, 1143–1170.
J. J. Eisch, A. M. Piotrowski, S. K. Brownstein, E. J. Gabe,
F. L. Lee, J. Am. Chem. Soc. 1985, 107, 7219–7221.
J. J. Eisch, M. P. Boleslawski, A. M. Piotrowski, in Transition
Metals and Organometallics as Catalysts for Olefin Polymeriza-
tion (Eds.: W. Kaminsky, H. Sinn), Springer-Verlag, Berlin,
Heidelberg, 1988, pp. 371–378.
J. J. Eisch, K. R. Caldwell, S. Werner, C. Krüger, Organometal-
lics 1991, 10, 3417–3419.
J. J. Eisch, K. R. Caldwell, in Advances in Chemistry Series
(Eds.: W. R. Moser, D. W. Slocum), American Chemical Soci-
ety, Washington, D.C., 1992, no. 230, pp. 575–590.
J. J. Eisch, S. I. Pombrik, G. X. Zhang, Organometallics 1993,
12, 3856–3863.
[9]
[10]
[11]
[12]
[30] J. Kopf, H. J. Vollmer, W. Kaminsky, Cryst. Struct. Commun.
1980, 9, 271–276.
[31] In a similar experimental finding, the irradiation of (η5-allyl)-
bis(η5-cyclopentadienyl)titanium(iii) by UV light of 254 nm
wavelength in the presence of 2 equiv. of diphenylacetylene led
to the isolation of 38% of dark green crystals of 34. Appar-
ently, an allyl radical was lost photolytically from the starting
titanocene(iii) and the resulting titanocene 32 captured in-
stantly by the alkyne to produce successively 33 and 35: J. J.
Eisch, S. I. Pombrik, X. Shi, S. C. Wu, Macromol. Symp. 1995,
89, 221–229.
[13]
[14]
[15]
J. J. Eisch, S. I. Pombrik, X. Shi, S. C. Wu, Macromol. Symp.
1995, 89, 221–229.
J. J. Eisch, F. A. Owuor, P. O. Otieno, J. N. Gitua, X. Shi, A. A.
Adeosun, in Advances in Chemistry Series (Eds.: P. J. Shapiro,
D.A. Atwood), American Chemical Society, Washington, D.C.,
2002, no. 822, pp. 88–103.
[32] G. W. Watt, L. J. Baye, F. O. Drummond, Jr., J. Am. Chem.
Soc. 1966, 88, 1138–1140.
[33] J. J. Salzmann, P. Mosimann, Helv. Chim. Acta 1967, 50, 1831–
1836.
[16] J. J. Eisch, P. O. Otieno, Eur. J. Org. Chem. 2004, 3269–3276.
[17] R. F. Jordan, W. E. Dasher, S. F. Echols, J. Am. Chem. Soc.
1986, 108, 1718–1719.
[18] D. J. Crowther, S. L. Borkowsky, D. Swenson, T. Y. Meyer,
R. F. Jordan, Organometallics 1993, 12, 2897–2903. A leading
reference to research of the Jordan group since 1986.
[19] S. Pasynkiewicz, Polyhedron 1990, 9, 429–453.
[34] K. Clauss, H. Bestian, Justus Liebigs Ann. Chem. 1962, 654, 8–
19.
[35] H. H. Brintzinger, J. E. Bercaw, J. Am. Chem. Soc. 1970, 92,
6182–6185.
[20] D. E. Babushkin, H. H. Brintzinger, J. Am. Chem. Soc. 2002, [36] J. J. Eisch, J. Organomet. Chem. 2001, 617–618, 148–157.
124, 12869–12873.
[37] J. J. Eisch, J. N. Gitua, Organometallics 2003, 22, 24–26.
[38] G. Erker, Acc. Chem. Res. 2001, 34, 309–317.
[39] A detailed description for conducting organometallic reactions
in a safe and reproducible manner is given in: J. J. Eisch, Orga-
nometallic Syntheses, Academic Press, New York, 1981, vol. 2,
pp. 1–84.
[21] K. P. Bryliakov, E. P. Talsi, M. Bochmann, Organometallics
2004, 23, 149–152.
[22] J. J. Eisch, B. W. Kotowicz, Eur. J. Inorg. Chem. 1998, 761–769.
[23] J. J. Eisch, K. Mackenzie, H. Windisch, C. Krüger, Eur. J. Inorg.
Chem. 1999, 153–162.
[24] J. J. Eisch, J. E. Galle, A. Piotrowski, in: Transition Metal Cata-
lyzed Polymerizations – Alkenes and Dienes (Ed.: R. P. Quirk),
Harwood Academic Publishers, New York, pp. 799–823.
[25] W. Kaminsky, R. Steiger, Polyhedron 1988, 7, 2375–2381.
[26] R. Steiger, Reaktionen von Bis(cyclopentadienyl)zirconium(IV)-
Verbindungen mit Methylalumoxan und Polymerisationsver-
[40] E. Samuel, M. D. Rausch, J. Am. Chem. Soc. 1973, 95, 6263–
6267.
[41] P. C. Wailes, H. Weigold, A. P. Bell, J. Organomet. Chem. 1972,
34, 155–164.
Received: May 24, 2005
Published Online: August 29, 2005
Eur. J. Org. Chem. 2005, 4364–4371
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4371