A fully integrated high-throughput screening methodology for the discovery of new polyolefin catalysts: Discovery of a new class of high temperature single-site group (IV) copolymerization catalysts

Article abstract of DOI:10.1021/ja020868k

For the first time, new catalysts for olefin polymerization have been discovered through the application of fully integrated high-throughput primary and secondary screening techniques supported by rapid polymer characterization methods. Microscale 1-octene primary screening polymerization experiments combining arrays of ligands with reactive metal complexes M(CH₂Ph)₄ (M = Zr, Hf) and multiple activation conditions represent a new high-throughput technique for discovering novel group (IV) polymerization catalysts. The primary screening methods described here have been validated using a commercially relevant polyolefin catalyst, and implemented rapidly to discover the new amide-ether based hafnium catalyst [η²- (N,O)-(2-MeO-C₆H₄) (2,4,6-Me₃C₆H₂)N]Hf (CH₂Ph)₃ (1), which is capable of polymerizing 1-octene to high conversion. The molecular structure of 1 has been determined by X-ray diffraction. Larger scale secondary screening experiments performed on a focused 96-member amine-ether library demonstrated the versatile high temperature ethylene-1-octene copolymerization capabilities of this catalyst class, and led to significant performance improvements over the initial primary screening discovery. Conventional one gallon batch reactor copolymerizations performed using selected amide-ether hafnium compounds confirmed the performance features of this new catalyst class, serving to fully validate the experimental results from the high-throughput approaches described herein.

Full text of DOI:10.1021/ja020868k

Published on Web 03/18/2003
A Fully Integrated High-Throughput Screening Methodology
for the Discovery of New Polyolefin Catalysts: Discovery of a
New Class of High Temperature Single-Site Group (IV)
Copolymerization Catalysts
Thomas R. Boussie, Gary M. Diamond, Christopher Goh, Keith A. Hall,
Anne M. LaPointe, Margarete Leclerc, Cheryl Lund, Vince Murphy,*
James A. W. Shoemaker, Ursula Tracht,‡ Howard Turner, Jessica Zhang,
Tetsuo Uno,§ Robert K. Rosen,† and James C. Stevens†
Contribution from Symyx Technologies Inc., 3100 Central Expressway,
Santa Clara, California 95051, and The Dow Chemical Company, 2301 North Brazosport BlVd.,
B-3827, Freeport, Texas 77541
Received June 21, 2002; E-mail: vmurphy@symyx.com
Abstract: For the first time, new catalysts for olefin polymerization have been discovered through the
application of fully integrated high-throughput primary and secondary screening techniques supported by
rapid polymer characterization methods. Microscale 1-octene primary screening polymerization experiments
combining arrays of ligands with reactive metal complexes M(CH2Ph)4 (M ) Zr, Hf) and multiple activation
conditions represent a new high-throughput technique for discovering novel group (IV) polymerization
catalysts. The primary screening methods described here have been validated using a commercially relevant
polyolefin catalyst, and implemented rapidly to discover the new amide-ether based hafnium catalyst [η2-
(N,O)-(2-MeO-C6H4)(2,4,6-Me3C6H2)N]Hf(CH2Ph)3 (1), which is capable of polymerizing 1-octene to high
conversion. The molecular structure of 1 has been determined by X-ray diffraction. Larger scale secondary
screening experiments performed on a focused 96-member amine-ether library demonstrated the versatile
high temperature ethylene-1-octene copolymerization capabilities of this catalyst class, and led to significant
performance improvements over the initial primary screening discovery. Conventional one gallon batch
reactor copolymerizations performed using selected amide-ether hafnium compounds confirmed the
performance features of this new catalyst class, serving to fully validate the experimental results from the
high-throughput approaches described herein.
Introduction
cesses employing metallocene complexes for the preparation
of commodity polyolefins such as linear low-density polyeth-
Within the polyolefins industry, new single-site catalysts with
improved performance features represent highly desirable
research goals. Improvements in catalyst performance features
such as temperature stability, productivity, comonomer incor-
poration capability, high molecular weight capability, comono-
mer sequence distribution control, and the ability to copoly-
merize new monomer combinations, are currently being sought
through the design and modification of catalyst structures.
Indeed, modifications of the ligand frameworks of single-site
catalysts have led to impressive advances in catalyst technologies
over the last twenty years. During this period, research has
centered mostly on metallocene complexes comprising high-
valent group (IV) metal centers supported by cyclopentadienyl-
based ligands.1 Structural modifications of the cyclopentadienyl
ligand framework have led to substantial improvements in
catalyst performance, and the emergence of commercial pro-
ylene (LLDPE).1-3 More recent developments include single-
site catalyst classes based upon new noncyclopentadienyl
ligands,4 and mid-late transition metals.5,6 Further discoveries
of new classes of single-site catalysts will undoubtedly lead to
improved and more versatile performance features, in addition
to unique polymer product families with advanced properties.
An assessment of the development of polyolefin catalyst
technologies over the last twenty years leads to the undeniable
(1) For selected reviews on metallocenes, see: (a) Jordan, R. F. AdV.
Organomet. Chem. 1991, 32, 325-387. (b) Kashiwa, N.; Imuta, J.,-i. Catal.
SurV. Jpn. 1997, 1, 125-142. (c) McKnight, A. L.; Waymouth, R. M. Chem.
ReV. 1998, 98, 2587-2598. (d) Brintzinger, H. H.; Fischer, D.; Mu¨lhaupt,
R.; Rieger, B. Waymouth, R. M. Angew. Chem., Int. Ed. Engl. 1995, 34,
1143-1170. (e) Bochmann, M. J. Chem. Soc., Dalton Trans. 1996, 255-
270. (f) Alt, H. G.; Ko¨ppl, A. Chem. ReV. 2000, 100, 1205-1221.
(2) For a review of the polyolefin field, refer to articles within: Chem. ReV.
2000, Volume 100.
(3) (a) Turner, H. W. U. S. Patent 4,752,597, 1988. (b) Turner, H. W.; Hlatky,
G. G.; Eckman, R. R. U. S. Patent 5,470,927, 1995. (c) Stevens, J. C.;
Timmers, F. J.; Wilson, D. R.; Schmidt, G. F.; Nickias, P. N.; Rosen, R.
K.; Knight, G. W.; Lai, S-y, Eur. Patent Appl. 416-815-A2, 1991. (d)
Stevens, J. C.; Neithamer, D. R. Eur. Patent Appl. 418-044-A2, 1991. (e)
Stevens, J. C.; Neithamer, D. R. U. S. Patent 5,064,802, 1991. (f) Canich,
J. A. M.; Licciardi, G. F. U. S. Patent 5,057,475, 1991.
† Dow Chemical Company.
‡ Current Address: Bayer AG, Corporate Technology, 51368 Leverkusen,
Germany.
§ Current Address: Genomics Institute of the Novartis Research Founda-
tion, 10675 John Jay Hopkins Drive San Diego, CA 92121.
9
4306
J. AM. CHEM. SOC. 2003, 125, 4306-4317
10.1021/ja020868k CCC: $25.00 © 2003 American Chemical Society

Discovery of New Polyolefin Catalysts
A R T I C L E S
Scheme 1. High-Throughput Discovery Workflow
conclusion that a limited understanding of the complex relation-
ships between catalyst structure and performance has impeded
progress. For example, it is currently not possible to predict
which new metal-ligand combinations will lead to active
catalyst classes, and which metal-ligand combinations will have
low or no activity. Furthermore, upon discovering a novel
catalytically active metal-ligand combination, catalyst optimi-
zation is always desirable, and yet for any new catalyst class it
is extremely difficult to predict the structural features necessary
to impart the desired improvements in catalyst performance.
Thus, research directed toward the discovery and optimization
of new catalyst classes for commercial applications is an
expensive, time-consuming “trial and error” process with an
uncertain outcome. It is therefore not surprising that high-
throughput synthesis and screening strategies are beginning to
emerge within the field of polyolefin catalysis. In fact, such
strategies are now being adopted for a broad range of catalytic
transformations.7-11 High-throughput approaches offer signifi-
cant advantages over more conventional methods of catalyst
discovery. First, a high-throughput screen can (i) rapidly identify
catalytically active systems, and importantly, identify and reject
inactive systems, and (ii) allow a much broader range of catalysts
candidates and conditions to be evaluated. As a consequence,
time and resources may be directed toward maximizing the
potential of the most promising catalyst candidates rather than
optimizing catalyst structures around less important local
“maxima” in catalyst performance. Additionally, the ability to
generate sufficient quantities of meaningful catalyst performance
data can provide the beneficiary with the opportunity to establish
comprehensive and ultimately predictive structure-property
relationships. Thus, to overcome the limitations of more
conventional methods of polyolefin catalyst discovery, and to
extend the capabilities of our own high-throughput program,
we have designed and implemented a fully integrated, high-
throughput discovery and optimization infrastructure targeted
toward next-generation catalysts for olefin polymerization.
Our approach utilizes a microscale high-throughput primary
screen in which arrays of new metal-ligand combinations are
rapidly surveyed for catalytic activity. The purpose of our
primary screen is to assess the potential of each metal-ligand
combination as a polymerization catalyst for a specifically
targeted application. Catalytically active metal-ligand combina-
tions identified using this primary screening format are then
subjected to a secondary screen in which the catalyst perfor-
mance properties for the targeted application are assessed using
larger scales and more commercially relevant conditions.
Catalysts that meet predetermined performance criteria in these
secondary screens are then subjected to rapid optimization
through the structural elaboration of the newly discovered ligand
class (Scheme 1). Secondary screens are typically intended to
improve upon the catalyst performance, and importantly, to
establish relationships between catalyst structure and catalyst
performance. Rapid polymer characterization techniques have
been developed to accommodate the synthetic throughput and
the nature of the polymer products from our primary and
secondary screens.
(4) See ref 6 and: (a) Baumann, R.; Davis, W. M.; Schrock, R. R. J. Am.
Chem. Soc. 1997, 119, 3830-3831. (b) Liang, L.-C.; Schrock, R. R.; Davis,
W. M.; McConville, D. H. J. Am. Chem. Soc. 1999, 121, 5797-5798. (c)
Scollard, J. D. McConville, D. H. J. Am. Chem. Soc. 1996, 118, 10 008-
10 009. (d) Tshuva, E. Y.; Goldberg, I.; Kol, M.; Goldschmidt, Z. J. Chem.
Soc. Chem. Commun. 2001, 2120-2121. (e) Tshuva, E. Y.; Goldberg, I.;
Kol, M. Goldschmidt, Z. Organometallics, 2001, 20, 3017-3028. (f)
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, 6847-6856. (g) Stephan, D.
W.; Gue´rin, F.; Spence, R. E. v. H.; Koch, L.; Gao, X.; Brown, S. J.;
Swabey, J. W.; Wang, Q.; Xu, W.; Zoricak, P.; Harrison, D. G. Organo-
metallics, 1999, 18, 2046-2048.
(5) (a) Ittel, S. D.; Johnson, L. K.; Brookhart, M. Chem. ReV. 2000, 100, 1169-
1203. (b) Johnson, L. K.; Killian, C. M.; Brookhart, M. J. Am. Chem. Soc.
1995, 117, 6414-6415. (c) Small, B. L.; Brookhart, M.; Bennett, A. M.
A. J. Am. Chem. Soc. 1998, 120, 4049-4050. (d) Britovsek, G. J. P.;
Gibson, V. C.; Kimberley, B. S.; Maddox, P. J.; McTavish, S. J.; Solan,
G. A.; White, A. J. P.; Williams, D. J. J. Chem. Soc. Chem. Commun.
1998, 849-850.
(6) For reviews on nonmetallocene olefin polymerization catalysts, see: (a)
Britovsek, G. J. P.; Gibson, V. C.; Wass, D. F. Angew. Chem., Int. Ed.
Engl. 1999, 38, 428-447. (b) Gibson, V. C.; Spitzmesser, S. K. Chem.
ReV. 2002, ASAP Web Release Date Dec 17.
(7) For recent reviews, see: (a) Jandeleit, B.; Schaefer, D. J.; Powers, T. S.;
Turner, H. W.; Weinberg, W. H. Angew. Chem., Int. Ed. Engl. 1999, 38,
2494-2532. (b) Wennemers, H. Combinatorial Chemistry & High Through-
put Screening 2001, 4, 273-285. (c) Gennari, F.; Seneci, P.; Miertus, S.
Catal. ReV.-Sci. Eng. 2000, 42(3), 385-402. (d) Gennari, C.; Nestler, H.
P.; Piarulli, U.; Salom, B. Liebigs Ann./Recueil 1997, 637-647. (e) Shimizu,
K. D.; Snapper, M. L.; Hoveyda, A. H. Chem. Eur. J. 1998, 4(10), 1885-
1889. (f) Crabtree, R. H. J. Chem. Soc. Chem. Commun. 1999, 1611-
1616.
(8) (a) Weinberg, W. H.; McFarland, E.; Goldwasser, I.; Boussie, T.; Turner,
H.; van Beek, J. A. M.; Murphy, V.; Powers, T. U.S. Patents 6,030,917,
2000; and 6,248,540, 2001. (b) Weinberg, W. H.; McFarland, E.; Gold-
wasser, I.; Boussie, T.; Turner, H. W.; van Beek, J. A. M.; Murphy, V.;
Powers, T. Eur. Patent Appl. 978-499-A2, 2000.
(9) High-throughput strategies for the discovery of polyolefin catalysts are
limited to a few examples; see: (a) Tian, J.; Coates, G. W. Angew. Chem.,
Int. Ed. Engl. 2000, 39, 3626-3629. (b) Stork, M.; Herrmann, A.; Nemnich,
T.; Klapper, M.; Mu¨llen, K. Angew. Chem., Int. Ed. Engl. 2000, 39, 4367-
4369. (c) Boussie, T. R.; Coutard, C.; Turner, H.; Murphy, V.; Powers, T.
S. Angew. Chem., Int. Ed. Engl. 1998, 37, 3272-3275. (d) Boussie, T. R.;
Murphy, V.; Hall, K. A.; Dales, C.; Petro, M.; Carlson, E.; Turner, H. W.;
Powers, T. S. Tetrahedron 1999, 55, 11699-11710. (e) Hinderling, C.;
Chen. P. Angew. Chem., Int. Ed. Engl. 1999, 38, 2253-2256. (f) Jones, D.
J.; Gibson, V. C.; Green, S. M.; Maddox, P. J. J. Chem. Soc. Chem.
Commun. 2002, 1038-1039.
(10) For high-throughput screening techniques, see: (a) Connolly, A. R.;
Sutherland, J. D. Angew. Chem., Int. Ed. 2000, 39, 4268-4271. (b)
Holzwarth, A.; Schmidt, H.-W.; Maier, W. F. Angew. Chem., Int. Ed. Engl.
1998, 37, 2644-2647. (c) Taylor, S. J.; Morken, J. P. Science 1998, 280,
267-270. (d) Stambuli, J. P.; Stauffer, S. R.; Shaughnessy, K. H.; Hartwig,
J. F. J. Am. Chem. Soc. 2001, 123, 2677-2678.
Herein, we report our first results in this area, including the
design of a 1-octene primary screen, its validation through the
use of a commercially relevant olefin polymerization catalyst,
and the subsequent discovery of a new 1-octene polymerization
catalyst using the primary screening methodology. Using our
secondary screen, we then performed high-temperature ethylene-
(11) For high-throughput catalyst discovery, see ref 7 and: (a) Francis, M. B.;
Jacobsen, E. N. Angew. Chem., Int. Ed. Engl. 1999, 38, 937-941. (b)
Sigman, M. S.; Jacobsen, E. N. J. Am. Chem. Soc. 1998, 120, 4901-4902.
(c) Porte, A. M.; Reibenspies, J.; Burgess, K. J. Am. Chem. Soc. 1998,
120, 9180-9187. (d) Gilbertson, S. R.; Collibee, S. E.; Agarkov, A. J.
Am. Chem. Soc. 2000, 122, 6522-6523. (e) Loch, J. A.; Crabtree, R. H.
Pure Appl. Chem. 2001, 73, 119-128. (f) Dahmen, S.; Bra¨se, S. Synthesis
2001, 10, 1431-1449.
9
J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003 4307

A R T I C L E S
Boussie et al.
1-octene copolymerizations using a 96-member focus library
comprising metal-ligand combinations structurally related to
the primary screening discovery. The results illustrate that for
this catalyst class, there is a strong dependence of copolymer
yield, molecular weight and 1-octene content of the copolymer
upon ligand structure. Additionally, many of the catalysts display
substantial performance improvements relative to the perfor-
mance of the initial primary screening discovery. Thus, our
secondary screen has identified and developed a versatile new
catalyst class capable of high-temperature ethylene-1-octene
copolymerizations. Finally, conventional large scale batch
reactor experiments performed on selected isolated catalysts
confirm the performance features of this new class of catalyst,
and further validate the high-throughput approaches adopted
herein.
to homopolymerize R-olefins would serve as a meaningful
primary screen for new catalyst discovery.15 Primary screening
discoveries could then be screened for their ability to copoly-
merize ethylene with 1-octene under more controlled and
commercially relevant conditions in our secondary screen (vide
infra). We are aware that ethylene-1-octene copolymerization
catalysts that are incapable of successive R-olefin insertions
will not be detected using this primary screen. However, we
adopted this approach in the belief that a rapid primary screen
in which multiple arrays of metal-ligand combinations were
assessed for their ability to polymerize 1-octene would provide
sufficient numbers of new catalyst discoveries to offset such a
limitation.
For validation purposes, we chose to assess the feasibility of
performing small-scale 1-octene polymerizations using the well-
studied catalyst [(η5-C5Me4)SiMe2(η1-NBut)]TiMe2, a highly
active olefin polymerization catalyst known to be effective for
the commercial production of ethylene-1-octene copolymers.3c-f
Additionally, we wished to determine relevant conditions under
which to perform the primary screening experiments using new
metal-ligand combinations. The validation experiment was
performed using an [8 × 12] array of 1 mL glass vials fitted
with miniature stir-bars. Polymerizations were conducted for 1
h in 220 µL total reaction volume containing 120 µL of 1-octene.
Reagents were added as toluene solutions using a commercially
available liquid handling robot driven by Symyx software
packages.16 Figure 1 illustrates the experiment design, which
encompasses three catalyst concentrations, three activation
methods (plus one control with no activator), and importantly,
eight replicates for each polymerization condition. Figure 2
displays the resultant 1-octene conversion data and clearly
demonstrates the excellent reproducibility for the eight replicates.
Moreover, some of the known performance features of [(η5-
C5Me4)SiMe2(η1-NBut)]TiMe2 were captured in these experi-
ments (e.g., single-site behavior and dependence of Mw on
1-octene conversion). A detailed description of the experiment
and the results can be found in the Supporting Information.
Importantly, reliable data from this well-studied, highly active
polymerization catalyst can be used as a figure of merit for new
metal-ligand combinations studied using our primary screen.
On the basis of these results, we conclude that meaningful
polymerization experiments can be conducted at microliter
scales.
Results and Discussion
Strategy. High-throughput approaches to catalyst discovery
typically require (i) an adaptation of catalyst synthesis and
screening to smaller scales, which is necessary to accommodate
arrays of reaction vessels,12 (ii) large numbers of ligands that
express suitably diverse electronic and steric properties, (iii)
an efficient method of attaching the ligand to the metal to form
the catalyst candidate, and (iv) a suitable rapid screening
technique. Requirements (i), (ii), and (iii) will be addressed in
the following sections. With regard to requirement (iv), the
sample analysis times of traditional polymer characterization
techniques such as GPC are too slow to accommodate arrays
of polymer products produced in a high-throughput fashion. We
have therefore utilized a Rapid]GPC technique that can provide
polymer concentration and molecular weight information at a
rate of ninety seconds per sample. Additionally, for the products
from our secondary screen, we used a Rapid]FT-IR technique
for the determination of weight % 1-octene incorporation within
ethylene-1-octene copolymers. The details of these techniques
have been reported elsewhere.13
Our previous work encompassing the requirements outlined
in (i)-(iv) above utilized simple “mix and screen” methods for
the discovery of late transition metal olefin polymerization
catalysts.9c-d Because commercial single-site catalysts are based
upon metallocene complexes comprising group (IV) metal
centers, we wished to extend our high-throughput strategies to
include new group (IV) metal-ligand combinations in our
primary screen. There is also a growing recognition that
nonmetallocene complexes hold great promise as olefin polym-
erization catalysts, and consequently we have turned our
attention to high-throughput syntheses and screening of non-
metallocene group (IV) complexes.
Primary Screening Discovery Experiments: Ligand Selec-
tion. Our starting point for discovery screening was a small
ligand array (arranged in a [8 × 3] format) containing an
assortment of simple bidentate and tridentate ligands with
(14) Linear low-density polyethylene is an ethylene-R-olefin copolymer in
which the R-olefin, typically 1-hexene or 1-octene, is statistically
distributed throughout the polymer chain. Commercial solution-phase
polymerizations are conducted at temperatures in excess of 100 °C. New
high-temperature catalysts that are capable of producing high molecular
weight, high R-olefin content copolymers with high productivities are
highly desirable, see: Klosin, J.; Kruper, W. J., Jr.; Nickias, P. N.; Roof,
G. R.; De Waele, P.; Abboud, K. A. Organometallics 2001, 20, 2663-
2665.
(15) Additionally, the use of a liquid R-olefin, such as 1-octene, in a primary
screening format is convenient, since this allows liquid handling robots to
deliver reagents and monomer rapidly to an array of reaction vessels.
Furthermore, poly-1-octene is a soluble material that can be conveniently
and rapidly manipulated using automated liquid handling equipment. See:
Diamond, G. M.; Goh, C.; Leclerc, M. K.; Murphy, V.; Turner, H. W.
WO Patent Application 01/98371 2001.
Design and Validation of the Primary Screen. The specific
objective of this primary screen is to discover new catalysts
with potential for the production of LLDPE.14 Insertion barriers
for R-olefins are higher than for ethylene, and catalysts capable
of polymerizing R-olefins would be predicted to copolymerize
ethylene with the R-olefin. We therefore rationalized that an
assessment of activated metal-ligand complexes for their ability
(12) Pooled approaches can also be used, see refs 7 and 9.
(13) (a) Nielson, R. B.; Kuebler, S. C.; Bennett, J.; Safir, A.; Petro, M. U.S.
Patent 6,175,409, 2001; Boussie, T. R.; Devenney, M. Eur. Patent Appl.
1-160-262-A1, 2001. (b) Komon, Z. J. A.; Diamond, G. M.; Leclerc, M.
K.; Murphy, V.; Okazaki, M.; Bazan, G. C.; J. Am. Chem. Soc. 2002, 124,
15 280-15 285.
(16) Lacy, S. D.; McFarland, E. W.; Safir, A. L.; Turner, S. J.; Van Erden L.;
Wang, P. WO Patent Application 00/23921, 2000; Rust, W. C.; Nielson,
R. B. WO Patent Application 00/67086, 2000.
9
4308 J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003

Discovery of New Polyolefin Catalysts
A R T I C L E S
Figure 1. [8 × 12] Array design for the primary screening validation experiment.
Figure 2. Conversion data from the primary screening validation experiment.
considerable diversity (Figure 3). Because this primary screening
experiment was directed toward the discovery of new early metal
catalysts, the ligands chosen mostly contain “hard” oxygen or
nitrogen based donor atoms with one or two acidic protons.
Oxygen donors include ethers, aliphatic and aromatic alcohols,
whereas nitrogen donors include imines, aliphatic and aromatic
amines and pyridines. Through these donor sets, ligands with
considerable differences in steric demands, electronic properties
and potential bite angles are represented. It is important to note
that the control cell [IIC], which contains no ligand, provides a
valuable measure of the catalytic activity of the M(CH2Ph)4-
activator combinations alone, allowing us to determine the effect
of the presence of each ligand upon the polymerization. Finally,
we are aware that some of the ligands present in this initial
array have been investigated previously as supporting ligands
for new group (IV) polyolefin catalysts.17
Primary Screening Discovery Experiments: Metal-
Ligand Complexation and Activation. As previously dis-
9
J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003 4309

A R T I C L E S
Boussie et al.
Figure 3. [8 × 3] Discovery ligand array.
cussed, high-throughput approaches to catalyst discovery require
an efficient method of attaching the ligand to the metal to form
the catalyst candidate. Our strategy for metal-ligand complex-
Scheme 2. Metal-Ligand Complexation Strategy Employed for
the Discovery Screen
(17) (a) Ligand [IH] see: van der Linden, A.; Schaverien, C. J.; Meijboom, N.;
Ganter, C.; Orpen, A. G. J. Am. Chem. Soc. 1995, 117, 3008-3021;
Katayama, H.; Nabika, M.; Imai, A.; Kawamura, N.; Hanaoka, H. Eur.
Patent Appl. 0-761-694, 1996. (b) Ligand [IIH] see: Scollard, J. D.;
McConville, D. H.; Vittal, J. J. Organometallics 1997, 16, 4415-4420;
Shaffer, T. D.; Squire, K. R. WO Patent Application 98/37109, 1998; Imuta,
J.; Saito, J.; Sugimura, K.; Fujita, T. WO Patent Application 98/34961,
1998. (c) Ligand [IIIA] and [IIIB], see: Fuhrmann, H.; Brenner, S.; Arndt,
P.; Kempe, R. Inorg. Chem. 1996, 35, 6742-6745. For related ligands,
see: Polamo, M.; Leskela¨, M.; Hakala, K.; Lo¨fgren, B. WO Patent
Application 97/45434, 1997. (d) Ligand [IIIC] see: Nitabaru, M.; Yoshida,
Y.; Mitani, M.; Terunori, F. Eur. Patent Appl. 1-008-595-A3, 2000. (e)
Ligand [IIID] see: Takahashi, F.; Naito, Y. JP 11060624, 1999. (f) Ligand
[IIIE], see: Theopold, K. H.; Kim, W.-K.; Power, J. M.; Mora, J. M.;
Masino, A. P. WO Patent Application 01/12637, 2001. (g) For ligands
related to [IG] see: Fujita, T.; Tohi, Y.; Mitani, M.; Matsui, S.; Saito, J.;
Nitabaru, M.; Sugi, K.; Makio, H.; Tsutsui, T. Eur. Patent Appl. 0-874-
005-A1 1998.
ation in our primary screen involves contacting a solution of
one equivalent of each ligand with a solution of M(CH2Ph)4
(M ) Zr, Hf) for sufficient time to afford the products of a
toluene elimination reaction. Scheme 2 illustrates one such
reaction between a bidentate ligand containing one acidic proton
(the ligand is represented generically as L-XH in Scheme 2)
and M(CH2Ph)4 (M ) Zr, Hf). The targeted reaction products
from this approach would contain the associated ligands and,
importantly, metal-carbon bonds to support olefin insertion
9
4310 J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003

Discovery of New Polyolefin Catalysts
A R T I C L E S
events in the presence of a suitable activator. Additionally,
because the removal of undesirable byproducts is not necessary
using this approach, activators can be added directly to the
reaction mixture in the presence of 1-octene to initiate the
primary screen. To assess the feasibility of this approach, we
1
conducted H NMR experiments to investigate the reactivity
of M(CH2Ph)4 (M ) Zr, Hf) with representative ligands from
the [8 × 3] ligand array. These studies revealed that reactions
between the less sterically encumbered ligands, such as ligand
[IIIC] (Figure 3), and M(CH2Ph)4 (M ) Zr, Hf) proceeded
rapidly through toluene elimination to generate metal-ligand
complexes containing metal-carbon bonds.18 Unfortunately,
reactions between more sterically encumbered ligands (such as
[IIH]) and M(CH2Ph)4 (M ) Zr, Hf) were either very slow, or
did not proceed at all. However, we recognized that additional
opportunities for the ligand to react with the group (IV) metal
center could arise through the use of carefully chosen activation
conditions. For example, alkyl aluminum reagents (e.g., R3Al),
are known to undergo facile alkyl group exchange reactions
with group (IV) metals, potentially generating a group (IV) metal
alkyl or metal hydride bond that may react more readily with
the acidic proton(s) on the ligand.19 Additionally, the alkyl
aluminum reagents may react directly with the acidic proton(s)
on the ligands, producing an aluminum-ligand species which
can subsequently react with M(CH2Ph)4 (M ) Zr, Hf) to form
the targeted group (IV) metal-ligand compound(s). Further-
more, the use of ionizing activators such as [Me2PhNH]-
[B(C6F5)4] may generate an open coordination site available for
coordination of the ligand. In our view, the use of carefully
chosen activation conditions increases the likelihood of ligand
association to the metal, and decreases the likelihood of missing
a new catalyst.20 Thus, in an attempt to overcome limitations
in complexation, and to minimize the probability of missing a
new catalyst, we chose to use eight different activation condi-
tions for each ligand/M(CH2Ph)4 combination. Specifically, we
chose to combine the ionizing activators [Me2PhNH][B(C6F5)4],
[Ph3C][B(C6F5)4] and B(C6F5)3, with two aluminum alkyl
reagents, Et3Al and Bui3Al (Figure 4). In fact, the combination
of [Me2PhNH][B(C6F5)4] with Bui3Al was shown to be par-
ticularly effective for the activation of [(η5-C5Me4)SiMe2-
(η1-NBut)]TiMe2 in our primary screening validation experi-
ment. In summary, we adopted this approach believing that high-
throughput screening of multiple arrays of metal-ligand
combinations under multiple activation conditions would provide
sufficient numbers of new catalyst discoveries to offset any
limitations in the complexation chemistry.
Figure 4. Activation conditions and plate designs for the primary screening
discovery experiment.
Scheme 3. Sequence of Reagent Additions and Polymer
Work-Up Protocol for the Primary Screening Discovery Experiment
[8 × 3] ligand array was repeated in four zones conveniently
allowing four different activation methods to be employed per
plate (1 plate ) 1 [8 × 12] array). To accommodate eight
activation conditions, two plates were used for each of the metal
compounds Zr(CH2Ph)4 and Hf(CH2Ph)4, totaling four plates
comprising 384 polymerization experiments. Scheme 3 displays
the sequence of reagent additions to initiate the screen. It is
noteworthy that using the high-throughput methodologies
described herein, these 384 polymerization experiments were
completed in just a few hours.21 The results indicate that of the
eight activation conditions chosen, the four conditions employed
in plate 1 proved to be the most productive for both Zr(CH2-
Ph)4 and Hf(CH2Ph)4. Figures 5 and 6 illustrate the resultant
1-octene conversion data for Zr(CH2Ph)4 and Hf(CH2Ph)4
respectively, for the plate 1 activation conditions. It is important
to note that the control cells containing M(CH2Ph)4 (M ) Zr,
Hf) in the absence of ligands, yield poly-1-octene with very
low 1-octene conversions for all eight activation conditions
(e1-8%, for plate positions 2C, 5C, 8C, and 11C in Figures 5
and 6). Figure 5 indicates that for Zr(CH2Ph)4 in combination
with this ligand set, 1-octene conversions range from < 1% to
60% using the conditions described, with five ligand/Zr(CH2-
Ph)4 combinations yielding poly-1-octene with conversions >
10%. Figure 6 indicates that for Hf(CH2Ph)4 in combination
with this ligand set, 1-octene conversions range from < 1% to
100% using the conditions described, with five ligand/Hf(CH2-
Ph)4 combinations yielding poly-1-octene with conversions >
10%. Interestingly, the same five ligand structures form produc-
tive combinations with both Zr(CH2Ph)4 and Hf(CH2Ph)4
(Ligands [IC], [IE], [IID], [IIIA], and [IIIC] from Figure 3).
Strikingly, using eight activations conditions, 1-octene conver-
sions can vary dramatically for a given metal-ligand combina-
tion. For example, 1-octene conversions for the combination
Primary Screening Discovery Experiments: 1-Octene
Primary Screening Results. Based upon results from the
primary screening validation experiments described above, the
M(CH2Ph)4/ligand concentration was set at 1.6 mM, and
polymerizations were conducted for 1 h in 250 µL total reaction
volume containing 120 µL 1-octene. As described previously,
reagents were added as toluene solutions using a liquid handling
robot. Figure 4 illustrates the experiment design in which the
(18) The products are dependent upon the ligand. For example, with [IIIC] the
reaction proceeds cleanly to one product of the type [ligand]Hf(CH2Ph)3;
with [IIIA], product mixtures containing species of the types [ligand]M-
(CH2Ph)3 and [ligand]2M(CH2Ph)2, (along with unreacted M(CH2Ph)4) were
observed. The notation [ligand] used here refers to the deprotonated ligand.
(19) Siedle, A. R.; Newmark, R. A.; Schroepfer, J. N.; Lyon, P. A. Organo-
metallics 1991, 10, 0, 400-404.
(20) We have seen evidence that activation conditions can assist the reaction
between bulky ligands and M(CH2Ph)4 (M ) Zr, Hf).
(21) Polymer yields were determined using an automated weighing assay.
9
J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003 4311

A R T I C L E S
Boussie et al.
Figure 5. Poly-1-octene conversion data for Zr(CH2Ph)4 with discovery ligand set.
Scheme 4. Reaction Scheme for the Preparation of 1
of Zr(CH2Ph)4 and ligand [IIIA] varied from 1 to 60%! Thus,
the right choice of activation condition can be essential for
detecting new catalytically active metal-ligand combinations,
and since this is not possible to predict, the use of multiple
activation conditions is warranted.
Overall, from the 384 primary screening experiments de-
scribed here, 10 ligand/M(CH2Ph)4 combinations (M ) Zr, Hf)
were identified as 1-octene polymerization catalysts when
combined with appropriate activation conditions.22,23 Thus, using
our high-throughput strategies, these 10 metal-ligand combina-
tions would be selected for secondary screening. Of the ligands
that have been reported as supporting ligands for group (IV)
polyolefin catalysts in the literature, [IIIA] and [IIIC] were
identified as poly-1-octene catalysts when combined with both
Zr(CH2Ph)4 and Hf(CH2Ph)4 under the conditions of our primary
screen.24 The most exciting observation is that from these small-
scale primary screening experiments, many catalytically active
metal-ligand combinations were identified, thereby demonstrat-
ing the power of this high-throughput approach.
Of particular interest to us was the combination of Hf(CH2-
Ph)4 with ligand [IID] (plate positions 2D, 5D, 8D, and 11D in
Figure 6). When activated with [Me2PhNH][B(C6F5)4] and 5
eq. Et3Al, this ligand/Hf(CH2Ph)4 combination generates a
catalyst capable of polymerizing 1-octene with 100% conversion
under the conditions of our primary screen. We felt that this
observation warranted an investigation into the reaction between
Hf(CH2Ph)4 and ligand [IID]. The reaction between Hf(CH2-
Ph)4 and [IID] leads quantitatively, and rapidly via toluene
elimination, to the formation of the new amide-ether hafnium
tribenzyl complex [η2-(N,O)-(2-MeO-C6H4)(2,4,6-Me3C6H2)N]-
Hf(CH2Ph)3 (1) (Scheme 4). The molecular structure of 1 has
been determined by X-ray diffraction (Figure 7) and indicates
a distorted trigonal bipyramid structure with η1-benzyl groups.
Selected bond lengths and angles for 1 are presented in Table
1, crystal, intensity collection and refinement data for 1 are
presented in the Supporting Information.
(22) The catalytically active metal-ligand combinations produced poly-1-octenes
with a wide range of molecular weights. Data for all 384 experiments can
be found in the Supporting Information.
(23) A repeat experiment in which the ligand-to-metal ratio was adjusted to 2:1
significantly altered the 1-octene conversions. However, no new catalytically
active metal-ligand combinations were identified.
(24) With respect to zirconium and hafnium complexes of the other ligands
listed in ref 17, R-olefin polymerization activity has only been reported
for the combination of ligand [IIH] with zirconium. The [IIH]/zirconium
combination, when activated with [Ph3C][B(C6F5)4], has been reported as
a low activity R-olefin oligomerization catalyst.17b In our primary screen,
the combination of [IIH]/Zr(CH2Ph)4/[Ph3C][B(C6F5)4]/5 Bui3Al produced
small quantities of a material that was not detected by our Rapid]GPC
technique. The Rapid]GPC protocol chosen has a low molecular weight
detection limit of around 1000.
Interestingly, there is no evidence for the formation of the
2:1 complex upon addition of excess [IID] to 1. The results
from the NMR and X-ray studies are consistent with the notion
that compound 1 was formed in our primary screening experi-
ment, and under suitable activation conditions is capable of
9
4312 J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003

Discovery of New Polyolefin Catalysts
A R T I C L E S
Figure 6. Poly-1-octene conversion data for Hf(CH2Ph)4 with discovery ligand set.
new ligand/Hf(CH2Ph)4 combinations through the preparation
of a 96-member focus library of structurally related ligands.
Secondary Screening: The Parallel Polymerization Reac-
tor. The next step in our high-throughput discovery research is
an assessment of primary screening catalyst discoveries using
larger scales under more controlled and more commercially
relevant copolymerization conditions. As previously discussed,
the rationale for our primary screen is that catalysts capable of
polymerizing R-olefins would be predicted to copolymerize
ethylene with the R-olefin to produce a LLDPE copolymer.
However, because commercial solution processes for the
production of LLDPE operate at temperatures in excess of 100
°C, high-temperature stability is an additional requirement for
new catalysts. Therefore, in our secondary screen we chose to
perform ethylene-1-octene copolymerization experiments at 130
°C. To meet the throughput required to perform secondary
screening on primary screening discoveries within our program,
we have designed and built our own 48-cell parallel polymer-
ization reactor. Although this parallel reactor has been described
elsewhere, the important features of this instrument are worth
mentioning.13b,25 The instrument consists of 48 individually
controlled 15 mL high-pressure batch reactors with magnetically
coupled mechanical overhead stirring. Each reactor cell pos-
sesses its own pressure and temperature control, and gaseous
monomer feed line. Reagent stations with liquid dispensing
robots are situated alongside the reactors, and are used to mix
reagents where necessary, and add reagents to each cell of the
reactor. Because reagents can be added to the reactor cells at
Figure 7. Molecular structure of 1.
Table 1. Selected Bond Lengths (Å) and Angles (deg) for 1a
Hf-N
2.094 (3) Hf-O
2.311 (2) Hf(1)-C(1) 2.226 (4)
Hf-C(2) 2.254 (4) Hf-C(3) 2.236 (3)
N-Hf-C(1)
N-Hf-C(3)
O-Hf-C(2)
C(1)-Hf-C(2)
C(2)-Hf-C(3)
114.0 (1)
91.0 (1)
86.5 (1)
119.0 (2)
98.9 (2)
N-Hf-C(2)
O-Hf-C(1)
O-Hf-C(3)
C(1)-Hf-C(3)
O-Hf-N
122.8 (1)
92.9 (1)
160.8 (1)
100.3 (2)
70.8(1)
a average of two crystallographically independent molecules.
polymerizing 1-octene to high conversions. In the following
sections, we describe preliminary secondary screening results
for this novel ligand/Hf(CH2Ph)4 [IID] combination, along with
(25) Turner, H. W.; Dales, G. C.; Van Erden, L.; van Beek, J. A. M. U.S. Patent
6,306,658, 2000.
9
J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003 4313

A R T I C L E S
Boussie et al.
Scheme 5. Reaction Scheme for the Preparation of the
Amine-Ether Focus Library
Figure 8. 96-Member focus library containing 3 classes of amine-ether
ligands.
high pressures and high temperatures, metal-ligand combina-
tions can be prepared, activated and quickly injected into each
cell at the equilibrated temperature and pressure of the copo-
lymerization. Moreover, upon catalyst injection, real-time tem-
perature, pressure and ethylene uptake measurements can be
made on each of the 48 cells simultaneously. Thus, 48 individual
polymerization events can be monitored in real-time under
conditions that provide meaningful information about the
performance capabilities of each catalyst at high temperatures.
Typically, 100-200 mg of copolymer product is targeted from
our secondary screening studies. Such a quantity allows a
comprehensive analysis of the polymer properties, but will not
foul the reactor or prevent the diffusion of the monomer to the
catalyst sites during the polymerization experiment.
Secondary Screening: High-Temperature Ethylene-1-
Octene Copolymerizations. Catalysts were prepared in situ by
mixing solutions of the ligands (1 eq.) with Hf(CH2Ph)4, in a
manner similar to that used in the primary screen. A high-
temperature reactor study comparing the performances of 1 and
the in situ prepared combination of Hf(CH2Ph)4 and ligand [IID]
(Figure 3) revealed little difference in the yield of polymer
obtained or the nature of the polymer produced.28 Only one
activation condition was chosen for this catalyst optimization
experiment, the combination of [Me2PhNH][B(C6F5)4] with Bui3-
Al, which was selected from a combination of primary screening
data and preliminary high-temperature secondary screening runs
with compound 1.29 Thus, 96 high-temperature ethylene-1-
octene copolymerization experiments were performed in toluene
solutions (6.1 mL total reaction volume containing 0.25 mL
1-octene, and Bui3Al as an impurity scavenger). Upon addition
of toluene, 1-octene and Bui3Al to the reactor cells, the stirring
rate was set to 600 rpm, the reactor temperature was set to 130
°C, and the pressure was set to 100 psi with ethylene. Using
the reagent station situated alongside the reactor cells, solutions
of the ligands were treated with solutions of Hf(CH2Ph)4, and
allowed to mix for 1 h at room temperature, whereupon the
resultant mixtures were treated with Bui3Al. At this point,
solutions of [Me2PhNH][B(C6F5)4] were injected into each
reactor cell, followed immediately by an aliquot30 from each
solution containing the ligand/Hf(CH2Ph)4 combinations to
initiate the copolymerizations. The copolymerizations were then
allowed to continue for 30 min, whereupon the reactors were
vented and cooled. The polymer products were isolated,
weighed, and characterized using our Rapid]FT-IR and
Rapid]GPC techniques. Figure 9 displays the results from this
experiment, plotted as polymer yield against molecular weight
for each ligand/Hf(CH2Ph)4 combination. Selected ligand struc-
tures are also displayed in Figure 9.
Secondary Screening: Focus Library Preparation. The
objective of our secondary screen is the optimization of catalyst
performance, and the development of relationships between
catalyst structure and catalyst performance. Preliminary reactor
experiments performed at 130 °C revealed that 1 is capable of
producing high molecular weight ethylene-1-octene copolymers
with high levels of 1-octene incorporation.26 To determine the
structural features necessary to further improve the performance
of 1, we prepared a focused 96-member amine-ether ligand
library containing significant structural variations on ligand
[IID]. As depicted in Figure 8, three structurally similar classes
of ligands were synthesized for this library, all of which were
based upon amine-ether binding groups, but contain three
different linkers situated between the amine and ether moieties,
namely 1,2-phenyl (both with and without additional substituents
on the linker; Class I), 1,2-ethyl (Class II), and 1,3-propyl linkers
(Class III). Within each of the three structural classes, additional
steric and electronic diversity is expressed by substituents on
the amine and ether moieties, denoted generically as R1 and R2
respectively in Figure 8. Because the principal focus of this
library was the structural elaboration of the ligand framework
discovered in the primary screening experiment, ligand Class I
is represented by 64 of the 96 ligands. Additionally, to examine
the influence of the linker group situated between the amine
and ether moieties, we prepared a total of 32 examples from
ligand Classes II and III. All ligand structures were prepared
using a previously developed aryl-amination procedure (Scheme
5).27 For a list of the 96 structures prepared, see the Supporting
Information.
Several important conclusions can be drawn from this
experiment: (i) many of these catalysts are clearly capable of
(27) See the Supporting Information, and: (a) Wolfe, J. P.; Wagaw, S.; Marcoux,
J.-F.; Buchwald, S. L. Acc. Chem. Res. 1998, 31, 805-818. (b) Hartwig,
J. F. Angew. Chem., Int. Ed. Engl. 1998, 37, 2046-2067.
(28) It should be noted that for bulkier ligands, particularly those containing
phenyl ether moieties, we typically observe identical products, but lower
polymer yields using the in situ preparation method. Thus, catalyst
performance is convolved with complexation efficiency, and this limitation
needs to be considered when analyzing the results of this study.
(29) It should be noted that this activation condition is unlikely to be optimal
for every ligand/Hf(CH2Ph)4 combination.
(30) The aliquot contained 1 µmol of the ligand/Hf(CH2Ph)4 mixture. Final
reactor concentrations: ligand/Hf(CH2Ph)4 mixture: 0.16 mM.; [PhMe2-
NH][B(C6F5)4]: 0.16 mM.; Bui3Al: 1.97 mM.
(26) An earlier version of our parallel polymerization reactor was used for these
experiments.
9
4314 J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003

Discovery of New Polyolefin Catalysts
A R T I C L E S
Figure 9. Results from the 96 Ligand/Hf(CH2C6H5)4 ethylene-1-octene copolymerizations performed at 130 °C.
Scheme 6. Reaction Scheme for the Preparation of 2
ethylene-1-octene copolymerizations at 130 °C which, we
believe, represents a significant technological advance for
nonmetallocene olefin polymerization catalysts; (ii) the data
displayed in Figure 9 illustrates that for this class of catalyst
there is a strong dependence of polymer yield, molecular weight
and 1-octene incorporation on the ligand structure, with many
of the catalysts demonstrating significant performance improve-
ments over the initial primary screening discovery. For instance,
Mw varied between 2k and 381k, 1-octene incorporation values
varied between 0 and 20 wt. %, and yields varied between 15
and 482 mg; and (iii) many of the product polymers possess
low polydispersities, demonstrating that under these conditions,
single-site catalysts are being generated.31 Thus, the results
suggest that we have discovered a new and versatile class of
high temperature, single-site LLDPE catalyst that can provide
access to many different types of polymer products through
structural changes to a very simple ligand framework.32
performance properties, and thus, a proper assessment of the
performances of the lower activity catalysts would therefore
require an increase in catalyst concentrations. The conditions
employed for these preliminary experiments necessarily repre-
sent a compromise, and clearly, the results should be considered
with this in mind.34 Additionally, we are aware that the results
from this experiment are convolved with the complexation
efficiency of the ligands with Hf(CH2Ph)4.28 In view of these
limitations, further experiments employing isolated complexes
are therefore necessary to establish quantitative structure-
property relationships.
Scaled-Up Batch Reactor Copolymerizations. To rigorously
test the quality of the results from our parallel polymerization
reactor, we prepared two amide-ether hafnium tribenzyl com-
plexes for larger scale batch reactor experiments. For this
comparison, 1 was chosen for along with a second catalyst, the
combination of Hf(CH2Ph)4 and ligand [7F], identified from
the secondary screening experiments as possessing superior
performance features in comparison to 1 (Figure 9). The reaction
between Hf(CH2Ph)4 and ligand [7F] proceeds rapidly and
quantitatively, via toluene elimination, to the formation of the
new amide-ether hafnium tribenzyl complex 2 (Scheme 6).35
Batch reactor evaluations of compounds 1 and 2 were
performed in a 1 gallon autoclave under commercially relevant
It is also important to discuss the limitations of screening
this many catalysts under a single set of experimental conditions
(refer to Figure 9). Using the conditions employed in this
experiment, we are aware that the highest activity catalysts
perform at an activity level that exceeds the ethylene diffusion
limitations of the reactor.33 Under these conditions, the results
for these catalysts, including polymer yields and polymer
properties will not represent the capabilities of the catalysts.
To properly assess the performances of the higher activity
catalysts would require a reduction in the catalyst concentrations.
Additionally, for the lowest activity catalysts, the yields are too
low to fully assess the polymer properties, and catalyst
(31) Of the 24 copolymers containing > 10 wt % 1-octene, 20 had polydisper-
sities below 3.5, see the Supporting Information.
(32) Goh, C.; Diamond, G. M.; Murphy, V.; Leclerc, M. K.; Hall, K.; Lapointe,
A. M.; Boussie, T. R.; Lund, C.; Uno, T. WO Patent Application 01/74910,
2001.
(33) At a stirring rate of 600 rpm. The real-time ethylene uptake monitoring
capabilities of these reactors enable us to readily identify catalysts that are
running under ethylene diffusion limitations.
(34) Under identical conditions, the results from this secondary screening
experiment are reproducible.
(35) The analogous zirconium compound has been reported, see: Murray, R.
E. U.S. Patent 6,103,657, 2000.
9
J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003 4315

A R T I C L E S
Boussie et al.
Table 2. One Gallon Batch Reactor Copolymerization Data
was oven-dried. Polymerizations were carried out in a parallel pressure
reactor, which is fully described in ref 25. High-temperature Size
Exclusion Chromatography was performed using an automated
Rapid]GPC system as described, for example in ref 13, and U.S.
Patents 6,294,388; 6,260,407; 6,296,771; and 6,265,226. All of the
molecular weight data were measured relative to linear polystyrene
standards. Rapid]FT-IR was performed on a Bruker Equinox 55 +
IR Scope II in reflection mode to determine weight % (wt %) of
1-octene incorporated in the copolymer. Samples were prepared in a
thin film format by evaporative deposition techniques. Wt.% 1-octene
was obtained from ratio of peak heights at 1378 and 4335 cm-1. This
method was calibrated using a set of ethylene-1-octene copolymers with
a range of known wt. % 1-octene content.13 NMR spectra were recorded
using a Bruker ADVANCE DPX 300 MHz spectrometer. GC-MS
analyses were recorded using a Hewlett-Packard 6898 gas chromato-
graph coupled to a Hewlett-Packard 5973 Mass Selective Detector. For
detailed descriptions of (i) the primary screening validation and
discovery experiments, and (ii) the secondary screening high-temper-
ature ethylene-1-octene copolymerizations, see the Supporting Informa-
tion.
Preparation of 96-Member Amine-Ether Ligand Library. Reac-
tions were performed on a 1 mmol scale in 20 mL glass vials. In a
typical procedure, Pd(dba)2, NaOBut, and L1 or L2 (Scheme 5) were
first combined in the glass vial under an atmosphere of argon. A
degassed toluene solution (ca. 5 mL) containing the amine and the aryl
bromide were then added to the reaction mixture by syringe transfer.
The mixture was then heated to 100 °C overnight. The crude reaction
product was then filtered, concentrated and purified using column
chromatography (silica gel; 5% ethyl acetate in hexane). L1 was initially
used for the preparation of this library, but was found to be less effective
for aryl bromides with bulky substituents at the 2 and 6 positions. For
these transformations L2 was used with an extended reaction time of
48 h.36 The aryl bromides employed for this library are commercially
available. The anilines were either obtained commercially or prepared
using simple procedures. All ligands were assayed for purity using
GCMS, and 1H and 13C NMR were recorded for representative members
of the 96 ligands (see Supporting Information).
complexa
CGC
1
2
activityb
Mw
Mw/Mn
52.30
59 500
2.0
7.56
97 000
3.0
13.76
181 000
2.8
polymer density (g/mL)
0.912
0.910
0.917
a Conditions: 1440 g Isopar E; 126 g 1-Octene, 130 °C, 450 psi Ethylene,
10 mmol H2. b grams of polymer/µmol metal.
high-temperature solution polymerization conditions. The runs
were performed at 130 °C and 450 psig of ethylene pressure,
using an isoparaffinic solvent, 1-octene as comonomer, and
hydrogen for molecular weight control. A comparative run was
performed using the titanium based catalyst [(η5-C5Me4)SiMe2-
(η1-NBut)]Ti(η4-C5H8) (CGC), which has proven performance
as a commercially viable catalyst in the Dow solution polyeth-
ylene process.
Key results from these runs are summarized in Table 2. It is
clear from these results that the new hafnium complexes 1 and
2 perform at a level comparable to a commercially relevant
catalyst system, particularly in the areas of polymer molecular
weight and comonomer incorporation (reflected by the polymer
density), where the new complexes equal or exceed the
performance of the CGC complex. Polymerization activity is
somewhat lower than the CGC complex, although still reason-
able considering the early stage of development of these catalyst
systems. It is particularly significant that the molecular weight
and activity trends from the commercial screening results for 1
and 2 mirror the results from the high-throughput screens. The
correspondence between the large-scale catalyst performance
and the secondary screening results clearly serves to validate
the new high-throughput screening methodologies for the
discovery and optimization of polyolefin catalysts.
Conclusions
Preparation of [η2-(N,O)-(2-MeO-C6H4)(2,4,6-Me3C6H2)N]Hf-
(CH2Ph)3 (1). A solution of (2-MeO-C6H4)(2,4,6-Me3C6H2)NH (0.076
g, 0.31 mmol) in 5 mL toluene was added to a solution of Hf(CH2Ph)4
(0.18 g, 0.33 mmol) in 5 mL toluene. After stirring the resultant mixture
for 12 h the toluene was removed under a stream of argon and the
product was then dried in vacuo. The product was then washed with
cold hexane and isolated as a pale yellow solid. (0.15 g, 70%). Crystals
for an X-ray structure determination were grown from a hexane/diethyl
ether solution at -35 °C. 1H NMR data (C6D6): δ 7.10, t, 6.82, t, 6.75
d [15H, Hf(CH2Ph)3], 6.9, [2H, s, [η2-(N,O)-(2-MeO-C6H4)(2,4,6-
Me3C6H2)N]Hf(CH2Ph)3], 6.75, t 6.50, t, 6.35, d 6.14, d [4H, [η2-
(N,O)-(2-MeO-C6H4)(2,4,6-Me3C6H2)N]Hf(CH2Ph)3], 2.98 [3H, s,
[η2-(N,O)-(2-MeO-C6H4)(2,4,6-Me3C6H2)N]Hf(CH2Ph)3], 2.18 [9H,
s, [η2-(N,O)-(2-MeO-C6H4)(2,4,6-Me3C6H2)N]Hf(CH2Ph)3], 2.00 [6H,
s, Hf(CH2Ph)3]; 13C{1H} (C6D6): δ 148.7, 144.6, 143.1, 141.1, 135.9,
135.6, 130.4, 128.9, 124.6, 122.7, 117.2, 112.9, 109.3 [η2-(N,O)-(2-
MeO-C6H4)(2,4,6-Me3C6H2)N]Hf(CH2Ph)3], 87.1 [Hf(CH2Ph)3], 58.6
[η2-(N,O)-(2-MeO-C6H4)(2,4,6-Me3C6H2)N]Hf(CH2Ph)3] 21.0, 18.3
[η2-(N,O)-(2-MeO-C6H4)(2,4,6-Me3C6H2)N]Hf(CH2Ph)3].
For the first time, new polyolefin catalysts have been
discovered and developed through the application of a fully
integrated high-throughput primary and secondary screening
approach supported by rapid polymer characterization tech-
niques. Using this approach, 384 microscale primary screening
polymerization experiments were performed in just a few hours
and resulted in the identification of 10 interesting 1-octene
polymerization catalysts. Larger scale secondary screening
experiments performed on a focused 96-member library led to
the development of a new and versatile high temperature LLDPE
catalyst class based upon a nonmetallocene ligand. Moreover,
we have subsequently used our high-throughput infrastructure
to screen many thousands of catalyst candidates for a variety
of targeted applications. Numerous catalyst discoveries have
emerged that were entirely unpredictable, and that most likely
would not have been discovered through more conventional
methods of catalyst research. The results described herein
illustrate the power of well-designed high-throughput experi-
ments, and highlight the tremendous advantages of the high-
throughput approach over more conventional methods of catalyst
discovery. Polyolefin catalysts can now be discovered and
optimized at unprecedented rates.
Preparation of [η2-(N,O)-(C4H7O-2-CH2)(2,6-Pri2-C6H3)N]Hf-
(CH2Ph)3 (2). A solution of (C4H7O-2-CH2)(2,6-Pri2-C6H3)NH (0.084
g, 0.32 mmol) in 5 mL toluene was added to a solution of Hf(CH2Ph)4
(0.20 g, 0.37 mmol) in 5 mL toluene. After stirring the resultant mixture
(36) (a) Bei, X.; Crevier, T.; Guram, A. S.; Jandeleit, B.; Powers, T. S.; Turner,
H. W.; Uno, T.; Weinberg, W. H. Tetrahedron Lett. 1999, 40 (20), 3855-
3858. (b) Bei, X.; Guram, A. S.; Turner, H. W.; Weinberg, W. H.
Tetrahedron Lett. 1999, 40 (7), 1237-1240. (c) Bei, X.; Turner, H. W.;
Weinberg, W. H.; Guram, A. S.; Petersen, J. L. J. Org. Chem. 1999, 64,
6797-6803. (d) Bei, X.; Uno, T.; Norris, J.; Turner, H. W.; Weinberg, W.
H.; Guram, A. S. Organometallics 1999, 18, 1840-1853.
Experimental Section
General. All reactions were performed under a purified argon or
nitrogen atmosphere in a glovebox. All solvents used were dried, de-
oxygenated and purified according to known techniques. All glassware
9
4316 J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003

Discovery of New Polyolefin Catalysts
A R T I C L E S
for 12 h, the toluene was removed under a stream of argon and dried
in vacuo. The product was then washed with cold hexane and isolated
as a white solid. 0.16 g (70%). H NMR data (C6D6): δ 7.22-7.15
(0.005 M in toluene) were prepared in the drybox, and the catalyst
was activated using a soluble anilinium borate activator, [(C18H37)2-
PhNH][B(C6F5)4] (1.2 equivs). An alumoxane (MMAO Type 3A from
Akzo Nobel Polymer Chemicals) was used as an impurity scavenger
(5 eq). In the comparative run, the titanium complex [(η5-C5Me4)SiMe2-
(η1-NBut)]Ti(η4-C5H8) was activated using B(C6F5)3 (3 eq) and MMAO-
3A (10 eq). The mixture of the metal complex, borane or borate, and
alumoxane was injected to the reactor over several minutes using a
flow of high pressure solvent, and the polymerization was allowed to
proceed for 10 min, with ethylene provided on demand to maintain
the pressure at 450 psig. The amount of ethylene consumed during the
reaction was monitored using a mass flow meter. The polymer solution
was dumped from the reactor into a nitrogen-purged glass kettle
containing 10-20 mL of 2-propanol. An aliquot of an additive solution
(prepared by dissolving 6.66 g of Irgaphos 168 and 3.33 g of Irganox
1010 in 500 mL of toluene) was added to this kettle and the solution
stirred thoroughly. The polymer solution was dumped into a tray, air-
dried overnight, and then thoroughly dried in a vacuum oven for 2 d.
The weights of the polymers were recorded and the efficiency calculated
as grams of polymer per micromole of transition metal.
1
[3H, m, (C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 7.15 t, 6.88 t,
6.81 d [15H, Hf(CH2Ph)3], 4.11 [1H, m, (C4H7O-2-CH2)(2,6-{(CH3)2-
CH}2-C6H3)N], 3.99 [1H, sept, (C4H7O-2-CH2)(2,6-{(CH3)2CH}2-
C6H3)N], 3.56 [1H, m, (C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N],
3.35 [1H, sept, (C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 3.24 [1H,
m, (C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 2.93 [1H, m, (C4H7O-
2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 2.78 [1H, m, (C4H7O-2-CH2)(2,6-
{(CH3)2CH}2-C6H3)N], 2.08 [6H, m, Hf(CH2Ph)3], 1.37 [3H, d,
(C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 1.29 [3H, d, (C4H7O-
2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 1.22 [3H, d, (C4H7O-2-CH2)(2,6-
{(CH3)2CH}2-C6H3)N], 1.20 [3H, m, (C4H7O-2-CH2)(2,6-{(CH3)2-
CH}2-C6H3)N], 1.20-1.02 [3H, m, (C4H7O-2-CH2)(2,6-{(CH3)2CH}2-
C6H3)N], 0.82 [1H, m, (C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N];
13C{1H} (C6D6): δ 149.6, 145.8 (br), 145.4, 144.9, 128.8, 127.3, 126.0,
124.6, 124.4, 122.0 [(C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N]Hf-
(CH2Ph)3, 86.5 [(C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 83.0 [br,
Hf(CH2Ph)3,], 72.0 [(C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 63.9
[(C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N], 28.5, 28.4, 27.8, 27.1,
26.8, 26.2, 24.22, 24.17 [(C4H7O-2-CH2)(2,6-{(CH3)2CH}2-C6H3)N].
X-ray Structure Determination for Compound 1. X-ray diffraction
data were collected on a Bruker P4 diffractometer equipped with a
SMART CCD detector. The structure was solved using direct methods
and standard difference map techniques, and were refined by full-matrix
least-squares procedures on F2 with SHELXTL (Version 5.03).37
Hydrogen atoms on carbon were included in calculated positions.
One Gallon Batch Reactor Copolymerization Procedure. Polym-
erization experiments were performed using a 1 gallon stirred Autoclave
Engineers reactor. The reactor was charged with 1440 g of Isopar E
(mixture of isoparaffins, available from ExxonMobil Chemical), 126
g of 1-octene, and 10 mmol of hydrogen, then heated to 130 °C and
saturated with ethylene to 450 psig. Solutions of compounds 1 and 2
Acknowledgment. The authors wish to express gratitude to
the dedicated Symyx scientists who have contributed to various
aspects of this work: Isabel Galdo, Jeffrey Harris, Rod Howden,
Vivian Luo, Eric Low, Fedri Marrugo, Jeff Norris, and Jackie
Regan.
Supporting Information Available: Tables of crystal structure
data for compound 1. Experimental designs and results from
the primary screening validation and discovery experiments.
Experimental designs and results from the secondary screening
ethylene-1-octene copolymerization experiments. GCMS data
1
for the 96-member amine-ether ligand library, and H and 13C
NMR data for representative examples. This material is available
free of charge via the Internet at http://pubs.acs.org.
(37) Sheldrick, G. M. SHELXTL, An Integrated System for Solving, Refining
and Displaying Crystal Structures from Diffraction Data; University of
Go¨ttingen, Go¨ttingen, Germany, 1981.
JA020868K
9
J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003 4317