FI catalysts: A molecular zeolite for olefin polymerization

Article abstract of DOI:10.1002/adsc.201000279

A bis(phenoxyimine) group 4 transition metal catalyst (now known as FI catalysts) can discern ethylene from a mixture of ethylene and propylene at more than 99% selectivity. Denisty function theory (DFT) calculations revealed a spatially confined reaction site in the transition states of the migratory insertion which is just the right size for an ethylene molecule but too small for a propylene one. The substituents adjacent to the phenoxy-oxygens are of crucial importance in developing the size/shape-selectivity.

Full text of DOI:10.1002/adsc.201000279

COMMUNICATIONS
DOI: 10.1002/adsc.201000279
FI Catalysts: A Molecular Zeolite for Olefin Polymerization
Haruyuki Makio,a,* Takashi Ochiai,a Hidetsugu Tanaka,a and Terunori Fujitaa,*
a
Catalysis Science Laboratory, Mitsui Chemicals, Inc., 580-32 Nagaura, Sodegaura, Chiba 299-0265, Japan
Fax : (+81)-438-64-2375; e-mail: Haruyuki.Makio@mitsui-chem.co.jp or Terunori.Fujita@mitsui-chem.co.jp
Received: April 9, 2010; Published online: June 30, 2010
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.201000279.
Abstract: A bis(phenoxyimine) group 4 transition
metal catalyst (now known as FI catalysts) can dis-
cern ethylene from a mixture of ethylene and pro-
pylene at more than 99% selectivity. Denisty func-
tion theory (DFT) calculations revealed a spatially
confined reaction site in the transition states of the
migratory insertion which is just the right size for
an ethylene molecule but too small for a propylene
one. The substituents adjacent to the phenoxy-oxy-
gens are of crucial importance in developing the
size/shape-selectivity.
However, in the field of olefin copolymerization
catalysis, the selectivity among the comonomers has
not been considered important, or rather the research
efforts have been directed towards widening the
scope of comonomer variety and increasing comono-
mer uptake (which of course have yielded an incredi-
bly rich variety of polyolefin products) and towards a
less selective direction. This trend has been accelerat-
ed by the advent of well-defined single-site catalysts,[1]
because they surpass, based on the rational under-
standing of polymerization mechanisms and molecu-
lar design, the limited performance attainable with
the classical heterogeneous Ziegler–Natta catalysts re-
garding comonomer incorporation. Thus, olefin co-
polymerizations have rarely been perceived as a selec-
tive catalysis and have not been directed towards the
other end of the spectrum, that is, a substrate-selec-
tive catalysis to enchain a particular monomer from
Keywords: ligand design; polymerization; size selec-
tivity; substituent effects; zeolites
Improvements in transition metal-catalyzed olefin among a mixture of more than two monomers. Such
polymerization performance can roughly be attributed selectivity among small molecules has been known to
to selectively increased or suppressed rates of specific exist in catalysis by zeolites, in which the molecules
elementary reaction(s), i.e., initiation, propagation, are discerned by uniform pore size and topologies.[2]
chain transfers, and irreversible terminations. Due to To the best of our knowledge, in transition metal-cat-
the inherent characteristics of polymerization reac- alyzed olefin polymerizations or any transition metal-
tions, where any errors tend to leave significant and catalyzed organic reactions, there are few examples
sometimes fatal consequences in the polymer prod- that exhibit a high selectivity based on the size or
ucts, any selectivity involved in these elementary reac- shape of the substrates in a similar way that zeolites
tions needs to be extremely high for the synthesis of or related catalysts do for simple alkanes or alkenes,[3]
macromolecules with desired structures, while such which have no directing functional groups to aid the
errors will simply reduce the yield for the synthesis of selective reactions through complementary non-cova-
small molecules. For example, ultra-high molecular lent interactions as seen in enzymatic catalysis.
weight polyethylene having molecular weights in the
Our continuing research into bis(phenoxyimine)
millions can be prepared with high efficiency, which group 4 transition metal catalysts (now known as FI
means that the propagation reaction is about 99.999% catalysts) has revealed that the appropriate choice
selective relative to any possible side reactions that and design of phenoxyimine ligands, metal, cocata-
can interrupt its chain growth. In another example, lysts, and polymerization conditions make FI catalysts
stereospecific propagation of a-olefins achieves very selective in almost all aspects of polymeriza-
~99.9% enantioselectivity to give either isotactic or tion and polymer structures. FI catalysts are (i) extra-
syndiotactic polypropylene with extremely high melt- ordinarily active for ethylene polymerization
ing temperatures (Tm), which are readily degraded (TOF=64,900 sÀ1 atmÀ1),[4] and (ii) very selective
with only the slightest microstructural errors.
about chain propagation to yield UHMWPE,[4] ultra-
high molecular weight ethylene/a-olefin copolymers
Adv. Synth. Catal. 2010, 352, 1635 – 1640
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1635

Haruyuki Makio et al.
COMMUNICATIONS
(>10,000,000 Da)[5] and atactic polypropylenes high that the monomer conversions easily exceeded
(8,300,000 Da),[6] and also various living olefin poly- this limit. Therefore, the MRR for CGC could not be
mers, which represent ultimate selectivity in propaga- accurately evaluated under the conditions examined.
tion over any termination reactions that do not Nevertheless, the monomer compositions in feed and
exist.[7] FI catalysts can be (iii) highly iso- and syndio- in the copolymer are very close to each other
specific for both propylene[8] and styrene[9] polymeri- (Table 1), which suggests an even smaller r1 value
zations, and (iv) can be selective in chain end struc- than the r1 (=4) estimated previously for ethylene/1-
tures to produce exclusively vinyl-[10] or R2Al-termi- octene copolymerization using the same complex.[13]
nated[11] olefin polymers (R=alkyl).
The r1 of Cp2ZrCl2 was estimated at 13; meaning that
Here we would like to report on another unique se- the ethylene last-inserted species (Cp2Zr-CH2CH2-
lectivity pertaining to FI catalysts, namely, the sub- polymer) is 13 times more reactive toward ethylene
strate-selectivity in ethylene/propylene copolymeriza- relative to propylene.
tions, where only ethylene is enchained into polymers
FI catalysts 2–4 are substantially more ethylene-se-
out of ethylene/propylene mixtures at >99% selectiv- lective than Cp2ZrCl2 or CGC. In particular, FI cata-
ity with high efficiency. In addition, this remarkable lyst 4, that has cumyl groups adjacent to the phenoxy-
selectivity based on molecular size/shape has resulted oxygen (R2 substituents), exhibited a more than 150
in unique characteristics for the obtained polyethy- times higher reactivity towards ethylene than propyl-
lenes, or can be used for new olefin polymerization ene, i.e., a remarkable selectivity of >99%, despite
and other processes. We would also like to touch on the difference of only one carbon in the substrates.
these new and emerging applications based on this On the other hand, the fact that 1 (R2 =cyclooctyl)
unique substrate-selective catalysis.
showed the lowest selectivity demonstrates that the
The selectivity between ethylene and propylene in R2 plays a vital role in determining the selectivity of
copolymerizations can be quantitatively estimated by ethylene monomers from a mixture of ethylene and
monomer reactivity ratios (MRR). In this study, we propylene and that the steric bulk of R2 near the a-
assume the first-order Markov chain, where the reac- carbon (to the phenoxy ring) is more important
tivity of the propagating species depends on the last- rather than the bulk of the entire R2.
inserted monomer unit (terminal model).[12] On this
The r1 values of 2 and 3 (R2 =t-Bu), which are simi-
assumption, there are four propagating reactions as lar to each other and in between those of 1 and 4, in-
shown in Scheme 1, and the MRR (r1, r2) are defined dicate that the substituents on the imine-nitrogen (R1
as the ratios of the propagation rate constants, r1 = substituents) have a smaller effect on monomer selec-
k11/k12, r2 =k22/k21 (M1 =ethylene; M2 =propylene).
A prototypical metallocene, Cp2ZrCl2, and a con- molecular weights are observed for these catalysts,[14]
strained geometry catalyst, Me2Si(Me4C5)(N-t- showing that polymer molecular weight and the sub-
tivity,[15] although considerable differences in polymer
A
ACHTUNGTRENNUNG
Bu)TiCl2 (CGC), were examined in ethylene/propyl- strate selectivity are determined independently as
ene copolymerization as a benchmark. Comonomer previously suggested.[4] These observations are quali-
uptake of Cp2ZrCl2 is modest or relatively low, while tatively consistent with an assumed octahedral cis-N,
CGC is known as one of the most efficient catalysts trans-O, cis-X (the reaction site) configuration of
for incorporating higher a-olefins in ethylene/a-olefin these FI catalysts, where the R2 is located above and
copolymerization.[13]
below the reaction site, while the R1 is positioned at
The MRR were evaluated using the Fineman–Ross the slightly skewed backside of the reaction site, and
method, and the conversions of propylene were kept thus does not directly interact with a coordinating
below 10% for almost all runs to minimize the change monomer.
from the initial monomer compositions.[14] However,
The DFT calculations further reinforce the argu-
the comonomer incorporation ability of CGC was so ment described above. Figure 1 compares the comput-
ed transition state structures for the migratory inser-
tions of ethylene and propylene to the alkyl com-
plexes of 1 and 4, where the alkyl chain is represented
by an n-propyl group, [hereafter, TSACTHNUGRTENNUG(1, E) and TSACHTUNGTRENNUNG(1,
P) denote the transition states of ethylene and propyl-
ene insertion to the cationic propyl derivative of 1, re-
spectively]. The bond lengths and geometry of the
four-membered metallacycles are very similar to each
other for these complexes.[14] However, the environ-
ments around these metallacycles are significantly dif-
ferent due to the R2 that is situated close to the reac-
tion site. The cyclooctyl groups of 1 can ease steric
tension by directing the smallest hydrogen on the a-
Scheme 1. Propagation reactions in the first order Markov
chain (L: ancillary ligand; i=1 or 2).
1636
asc.wiley-vch.de
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1635 – 1640

FI Catalysts: A Molecular Zeolite for Olefin Polymerization
Table 1. Monomer reactivity ratios in ethylene/propylene copolymerizations.[a]
Catalyst precursor
Propylene in feed [mol%][b]
Propylene conversion [%]
Propylene in polymer [mol%][c]
r1
Cp2ZrCl2
32.25
58.84
36.69
32.25
58.84
72.70
58.84
72.70
85.40
58.84
72.70
85.40
58.84
72.70
85.40
11.4
6.1
16.6
8.6
3.1
8.3
7.7
2.0
3.3
7.6
6.0
3.9
0.42
3.2
6.2
2.87
4.63
31.77
2.39
3.52
5.12
1.79
2.48
5.23
2.59
3.84
5.12
0.88
1.95
3.02
13
Cp2ZrCl2
CGC[d]
–
16
1
1
1
2
2
2
3
3
3
4
4
4
69
45
152
[a]
General conditions: 5.0 mL of hexane, Al
uous ethylene supply.
ACHTUNGTNER(NUNG dMAO)/Zr=300, 1008C, 10 min, 7.0 atm gauge pressure maintained by contin-
[b]
[c]
[d]
Calculated initial monomer composition.
Determined either by IR or 13C NMR.
5.0 mL of toluene, (i-Bu)3Al/Ph3CBACTHUNGTRENUNG(C6F5)4/CGC=300/2/1, 1008C, 5 min, 7.0 atm gauge pressure maintained by ethylene
supply.
carbon (to the phenoxy ring) toward the metallacycles
Kretschmer, Kempe and their co-workers reported
radius of hydrogen is 1.20 ꢁ, TSACUTHGTNRENNU(NG 1, E) and TSACHTUNGTRENNUNG
(1, P) conium complexes bearing aminopyridinato ligands.[3a]
(Figure 1 a,b). Considering that the van der Waals a similar ethylene selective polymerization using zir-
have no significant constraint between R2 substituents Under the given conditions in their report, the selec-
and the reaction site. On the other hand, one of the tivity seems to roughly correspond to the level of 2
methyl groups in the cumyl group of 4 needs to pro- (propylene mole fraction in the reaction media:
trude into the reaction sphere, and besides, the con- 43.5%; propylene content in polymer: <1 mol%) at
formation of the cumyl group is much more rigid than 808C (vs. 1008C for our experiments), which implies
that of the cyclooctyl group because of the high rota- that the ethylene selectivity of 4 will exceed that of
À
tional barrier around the two C(quaternary) C(ipso their catalysts under comparable conditions. In addi-
carbons in phenyl and phenoxy groups) bonds. This tion, despite the significant steric difference in their li-
geometry and rigidity of the cumyl group considera- gands, the ethylene selectivity did not vary for their
bly confine the space around the reaction site as can two complexes, from which they concluded that this
be clearly seen in the shortest distances between R2 selectivity is electronic in nature. This is in sharp con-
and the reaction site (Figure 1 c,d). The structure of trast to the FI catalysts that exhibit a clear structure-
TS
cule seems to be close-fitting in the confined space tivity is primarily steric in nature and thus tunable.
formed by the cumyl groups. With regard to TS(4, P), This remarkably high ethylene-selectivity of the FI
ACHTUNGTRENNUNG(4, E) demonstrates that even the ethylene mole- selectivity relationship where the origin of the selec-
AHCTUNGTRENNUNG
the spatial constraints between the cumyl groups and catalysts has resulted in applications that have not
the transition state metallacycle seem to be critically been possible using catalysts with ordinary selectivity.
À
large, apparently verified by the very short H H dis-
One of the consequences of the unique ethylene-se-
tances (Figure 1 d). From the calculated rate con- lectivity is that FI catalysts can produce a completely
stants (k11, k12), the r1 values were estimated at 125.5 linear polyethylene free from any short or long chain
and 57.5 for 4 and 1, respectively, which are in appre- branches. Commercially available high density poly-
ciably good agreement with the experiments.
ethylenes (HDPE) have methyl branches (0.2–0.6 Me/
Adv. Synth. Catal. 2010, 352, 1635 – 1640
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1637

Haruyuki Makio et al.
COMMUNICATIONS
Figure 1. Calculated transition states of olefin insertion into Zr-propyl species of FI catalysts: a) TSACTHNUTRGENNGU(4, E); b) TSCAHTUNGTERN(NUGN 4, P); c) TS-
ACHTUNGTRENNUNG(1, E); d) TSACHTUNGTRENNUNG(1, P). The important distances are given in ꢁ. The atoms highlighted in red and blue are oxygens and nitro-
gens, and in light blue and pink represent inserting monomers and propyl chains, respectively. Some hydrogen atoms are
omitted for clarity.
1000 C) or longer branches that are not derived from branches (under the detection limit, <0.1 branch/
a-olefin comonomers,[16] but presumably from chain- 1000 C).[16] Since the highly ethylene-selective FI cata-
end isomerization (Scheme 2 a) or reinsertion of in- lysts possess such a confined reaction site, the isomer-
situ generated vinyl macromers (Scheme 2 b). For ex- ization, which will generate a sterically demanding
ample, a heterogeneous Ziegler catalyst yielded poly- secondary a-carbon or the reinsertion of large macro-
ethylene with 0.3 and 0.6 methyl branches per 1000 mers, is virtually impossible. The defect-free, com-
carbons in ethylene homo-polymerizations, while pletely linear PE is expected to show higher crystal-
polyethylenes prepared by complex 2 under compara- linity, stronger intermolecular forces, and thus greater
ble conditions are truly linear and have virtually no tensile strength. These features are most suitable for
Scheme 2. Probable pathways for the formation of a) methyl, and b) long chain branches.
1638
asc.wiley-vch.de
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1635 – 1640

FI Catalysts: A Molecular Zeolite for Olefin Polymerization
UHMWPE having >106 molecular weights, whose ap-
plications range from lithium ion battery separators
to bulletproof vests. In fact, complex 4 produces such
UHMWPE under industrially relevant conditions.[14]
Scientists at Dow Chemical Company have recently
revealed that the high ethylene-selectivity of FI cata-
lysts can also be applied to make an olefin block co-
polymer in a continuous process.[17] The process in-
cludes two catalysts, one of which is a highly ethyl-
ene-selective FI catalyst (closely related to complex 2
in this work) that provides the block copolymers with
crystalline segments with low comonomer contents
and high melting temperatures (Tm). The hard crystal-
line segments produced with the FI catalyst form
physical cross-linking points connecting the very elas-
tic soft segments, which were simultaneously pro-
duced with the other catalyst, generating an infinite
network and serving as a thermoplastic elastomer.
The strength of the cross-linking points is directly re-
lated to the crystallinity of the hard segments, which
in turn is attributed to the excellent ethylene-selectivi-
ty of the FI catalysts over 1-octene.
toluene. Propylene was introduced until the reactor pressure
reached the predetermined value and the reactors were
closed. The reactor temperature was raised to 1008C and
ethylene gas was applied so that the total gauge pressure of
the reactors was maintained at 7 bar all the time. Through
the reagent addition ports, a prescribed amount of dMAO
in toluene, and then the catalyst solution were added to ini-
tiate the polymerization. After 10 min, the polymerization
was quenched by adding isobutyl alcohol. After removal of
volatiles, the obtained polymer was dried under vacuum at
1008C for 12 h.
Elucidation of MRR
The initial monomer compositions in solution under the
given conditions were calculated by using Aspen Plusꢂ
(Aspen Technology, Inc.) and the copolymer compositions
were measured either by IR or 13C NMR (Table 1). Accord-
ing to the Fineman-Ross equation [Eq. (1)], where F is the
initial molar ratio of ethylene/propylene in reaction media
and f is copolymer compositions (ethylene/propylene) in the
obtained polymers, FACTHNUTRGNEUNG
(f - 1)/f was plotted against F2/f and
linear approximation of the plots (mean square correlation
coefficient, R2 =0.96–0.99) gave the r1 as slope of a line and
the r2 as a Y-intercept.
In conclusion, we have developed highly ethylene-
selective FI catalysts based on the size/shape recogni-
tion in the substrates. An FI catalyst can polymerize
only ethylene monomers from mixtures of ethylene
and propylene at >99% selectivity, the highest selec-
tivity ever reported. Zeolites are known to be highly
selective in size and shape of substrates or products in
the reactions of small molecules based on the size and
topology of uniform pores. Similarly, the molecular
size/shape selectivity of FI catalysts stems from the
confined reaction site consisting of the metal center
and the appropriate substituents on the phenoxyimine
ligands, which is just right for ethylene but too small
for propylene, allowing FI catalysts therefore to func-
tion as a molecular zeolite. As such, we envisage that
the FI ligand framework, which is relatively rigid as
an organometallic compound but significantly more
flexible than inorganic zeolites, might be applicable
for other reactions which require high selectivity at
the molecular level.
References
[1] T. Fujita, H. Makio, in: Comprehensive Organometallic
Chemistry III, Vol. 11, (Eds.: R. H. Crabtree, D. M. P.
Mingos; Volume Ed.: T. Hiyama), Elsevier, Amster-
dam, 2007, pp 691–734.
[2] For a textbook, see: N. Y. Chen, W. E. Garwood, F. G.
Dwyer, Shape Selective Catalysis in Industrial Applica-
tions, Marcel Dekker Inc., New York, 1989.
[3] a) W. P. Kretschmer, B. Hessen, A. Noor, N. M. Scott,
R. Kempe J. Organomet. Chem. 2007, 692, 4569–4579.
b) The kinetic resolution of racemic a-olefins was in-
vestigated, where the selectivity stems from chirality:
J. A. Byers, J. E. Bercaw Proc. Natl. Acad. Sci. U S A.
2006, 103, 15303–15308, and references cited therein.
[4] a) H. Makio, N. Kashiwa, T. Fujita, Adv. Synth. Catal.
2002, 344, 477–493; b) M. Mitani, J. Saito, S. Ishii, Y.
Nakayama, H. Makio, N. Matsukawa, S. Matsui, J.
Mohri, R. Furuyama, H. Terao, H. Bando, H. Tanaka,
T. Fujita, Chem. Rec. 2004, 4, 137–158.
Experimental Section
[5] S. Ishii, J. Saito, S. Matsuura, Y. Suzuki, R. Furuyama,
M. Mitani, T. Nakano, N. Kashiwa, T. Fujita, Macro-
mol. Rapid Commun. 2002, 23, 693–697.
Ethylene/Propylene Copolymerization
[6] J. Saito, M. Onda, S. Matsui, M. Mitani, R. Furuyama,
H. Tanaka, T. Fujita, Macromol. Rapid Commun. 2002,
23, 1118–1123.
[7] a) M. Mitani, T. Nakano, T. Fujita, Chem. Eur. J. 2003,
9, 2396–2403; b) G. J. Domski, J. M. Rose, G. W.
Coates, A. D. Bolig, M. Brookhart, Prog. Polym. Sci.
2007, 32, 30–92; c) M. C. W. Chan, Macromol. Chem.
Phys. 2007, 208, 1845–1852.
Copolymerizations were carried out in a parallel, multi-reac-
tor (Endeavor Catalyst Screening Systemꢂ, Biotage AB)
equipped with the reagent addition ports and automated
system of gas delivery, heating and mixing control, and the
entire system was placed in a nitrogen-filled glove box. Into
thoroughly dried reactors, hexane (5 mL) was added via a
gas-tight syringe through the reagent addition ports and sa-
turated with propylene gas at 258C under vigorous stirring.
In a separate vessel, a catalyst solution was prepared by
[8] a) A. V. Prasad, H. Makio, J. Saito, M. Onda, T. Fujita,
Chem. Lett. 2004, 33, 250–251; b) M. Mitani, R. Fur-
mixing an FI ligand and ZrACHTNUGTRNEUNG(CH2Ph)4 in 2:1 molar ratio in
Adv. Synth. Catal. 2010, 352, 1635 – 1640
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1639

Haruyuki Makio et al.
COMMUNICATIONS
uyama, J. Mohri, J. Saito, S. Ishii, H. Terao, T. Nakano,
H. Tanaka, T. Fujita, J. Am. Chem. Soc. 2003, 125,
4293–4305.
[13] A. L. McKnight, R. M. Waymouth, Chem. Rev. 1998,
98, 2587–2598.
[14] See the Supporting Information.
[9] K. Michiue, M. Onda, H. Tanaka, H. Makio, M. Mitani,
T. Fujita, Macromolecules 2008, 41, 6289–6291.
[10] H. Terao, S. Ishii, J. Saito, S. Matsuura, M. Mitani, N.
Nagai, H. Tanaka, T. Fujita, Macromolecules 2006, 39,
8584–8593.
[11] J. Saito, Y. Tohi, N. Matsukawa, M. Mitani, T. Fujita,
Macromolecules 2005, 38, 4955–4957.
[15] It is possible that the bulkier 2-methylcyclohexyl group
makes 2 slightly more ethylene-selective than 3.
[16] M. Takahashi, T. Kasai, S. Otsuzuki, T. Matsumoto, T.
Tsutsui, J. Saito, N. Matsukawa, K. Tsuru, M. Mitani, T.
Fujita, (Mitsui Chemicals, Inc.), Patent WO00/078828,
2000.
[17] D. J. Arriola, E. M. Carnahan, P. D. Hustad, R. L. Kuhl-
[12] G. Odian, in: Principles of Polymerization, 4th edn.,
man, T. T. Wenzel, Science 2006, 312, 714–719.
John Wiley & Sons, Inc., Hoboken, 2004, pp 464–480.
1640
asc.wiley-vch.de
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1635 – 1640