Use of computational and synthetic chemistry in catalyst design: A new family of high-activity ethylene polymerization catalysts based on titanium tris(amino)phosphinimide complexes

Article abstract of DOI:10.1021/om049545i

DFT calculations of the mechanism of polymerization for the series of catalyst models derived from CpTiMe₂(NPR₃) (R = Me, NH₂, H, Cl, F) demonstrate the critical role of ion pairing in determining the overall barrier to polymerization and suggest that electron-donating substituents reduce this barrier. Based on these results, a family of precatalysts of general formula Cp′TiX₂(NP(NR ₂)₃) (X = Cl, Me) were developed. This approach using computational methods to guide the synthetic efforts has afforded a new, readily accessible, and easily varied family of highly active ethylene polymerization catalysts based on titanium tris(ammo)phosphinimide complexes.

Full text of DOI:10.1021/om049545i

5240
Organometallics 2004, 23, 5240-5251
Use of Com p u ta tion a l a n d Syn th etic Ch em istr y in
Ca ta lyst Design : A New F a m ily of High -Activity
Eth ylen e P olym er iza tion Ca ta lysts Ba sed on Tita n iu m
Tr is(a m in o)p h osp h in im id e Com p lexes
Chad Beddie, Emily Hollink, Pingrong Wei, J ames Gauld, and
Douglas W. Stephan*
Department of Chemistry and Biochemistry, University of Windsor,
Windsor, Ontario, Canada, N9B3P4
Received J une 22, 2004
DFT calculations of the mechanism of polymerization for the series of catalyst models
derived from CpTiMe2(NPR3) (R ) Me, NH2, H, Cl, F) demonstrate the critical role of ion
pairing in determining the overall barrier to polymerization and suggest that electron-
donating substituents reduce this barrier. Based on these results, a family of precatalysts
of general formula Cp′TiX2(NP(NR2)3) (X ) Cl, Me) were developed. This approach using
computational methods to guide the synthetic efforts has afforded a new, readily accessible,
and easily varied family of highly active ethylene polymerization catalysts based on titanium
tris(amino)phosphinimide complexes.
In tr od u ction
dienylborate,18 diketimine,19 tropidinyl,20 tridentate,21
and macrocyclic ligands.22 These and other efforts have
been recently reviewed by Gibson and co-workers.23,24
The “constrained geometry catalysts” (CGC) (B)25-33
derived from Cp-amido ligand complexes have been
Since the discovery that zirconocene derivatives can
act as single-site catalysts for olefin polymerization,
efforts have been ongoing both in academic and in
industrial labs to uncover alternative or modified cata-
lyst systems. While many groups have targeted the
control of polymer properties via judicious modification
of zirconocene precursors, others have probed the vi-
ability of early metal complexes with alternative ancil-
lary ligands as effective polymerization catalysts. In
such efforts, a number of systems have drawn consider-
able attention. For example, in 1996, McConville and
co-workers uncovered a nonmetallocene living polym-
erization catalyst (A), based on a bulky chelating
diamide ligand.1 A wide variety of other ligand modifi-
cations have been probed. Appreciable catalytic activi-
ties have been derived from titanium and zirconium
complexes containing amido,2 diamido,3-8 bidentate
aryloxide,9-11 borollide,12,13 boratabenzene,14,15 pendant
cyclopentadienyl borane,16 trimethylene,17 cyclopenta-
(11) Van Der Linden, A.; Schaverien, C. J .; Meijboom, N.; Ganter,
C.; Orpen, A. G. J . Am. Chem. Soc. 1995, 117, 3008-3021.
(12) Bazan, G. C.; Schaefer, W. P.; Bercaw, J . E. Organometallics
1993, 12, 2126.
(13) Bazan, G. C.; Donnelly, S. J .; Ridriguez, G. J . Am. Chem. Soc.
1995, 117, 2671-2672.
(14) Bazan, G. C.; Rodriguez, G.; Ashe, A. J ., III; Al-Ahmad, S.;
Muller, C. J . Am. Chem. Soc. 1996, 118, 2291-2292.
(15) Bazan, G. C.; Rodriguez, G. Organometallics 1997, 16, 2492-
2494.
(16) Spence, R. E. v. H.; Piers, W. E. Organometallics 1995, 14,
4617-4624.
(17) Rodriguez, G.; Bazan, G. C. J . Am. Chem. Soc. 1997, 119, 343-
352.
(18) Sun, Y.; Spence, R. E. v. H.; Piers, W. E.; Parvez, M.; Yap, G.
P. A. J . Am. Chem. Soc. 1997, 119, 5132-5143.
(19) Vollmerhaus, R.; Rahim, M.; Tomaszewski, R.; Xin, S.; Taylor,
N. J .; Collins, S. Organometallics 2000, 19, 2161-2169.
(20) Skoog, S. J .; Mateo, C.; Lavoie, G. G.; Hollander, F. J .; Bergman,
R. G. Organometallics 2000, 19, 1406-1421.
(21) Okuda, J .; Eberle, T.; Spaniol, T. P.; Piquet-Faure, V. J .
Organomet. Chem. 2000, 591, 127-137.
* To whom correspondence should be addressed. E-mail: stephan@
uwindsor.ca.
(1) Scollard, J . D.; McConville, D. H. J . Am. Chem. Soc. 1996, 118,
10008-10009.
(2) Sinnema, P. J .; Okuda, J . J . Organomet. Chem. 2000, 598, 179-
(22) Fokken, S.; Spaniol, T. P.; Kang, H.-C.; Massa, W.; Okuda, J .
Organometallics 1996, 15, 5069-5072.
(23) Gibson, V. C.; Spitzmesser, S. K. Chem. Rev. 2003, 103, 283-
315.
181.
(24) Britovsek, G. J . P.; Gibson, V. C.; Wass, D. F. Angew. Chem.,
Int. Ed. 1999, 38, 428-447.
(3) Lorber, C.; Donnadieu, B.; Choukroun, R. Organometallics 2000,
19, 1963-1966.
(25) Shapiro, P. J .; Bunel, E.; Schaefer, W. P.; Bercaw, J . E.
Organometallics 1990, 9, 867-869.
(4) Scollard, J . D.; McConville, D. H. J . Am. Chem. Soc. 1996, 118,
10008-10009.
(26) Chen, Y.-X.; Marks, T. J . Organometallics 1997, 16, 3649-3657.
(27) Tian, S.; Arredondo, V. M.; Stern, C. L.; Marks, T. J . Organo-
metallics 1999, 18, 2568-2570.
(5) Scollard, J . D.; McConville, D. H.; Vittal, J . J . Organometallics
1995, 14, 8.
(6) Warren, T. H.; Schrock, R. R.; Davis, W. M. Organometallics
1998, 17, 308-321.
(28) Deck, P. A.; Beswick, C. L.; Marks, T. J . J . Am. Chem. Soc.
1998, 120, 1772-1784.
(7) Cloke, F. G. N.; Geldach, T. J .; Hitchcock, P. B.; Love, J . B. J .
Organomet. Chem. 1996, 506, 343-345.
(8) Lee, C. H.; La, Y.-H.; Park, J . W. Organometallics 2000, 19, 344-
351.
(9) Nomura, K.; Naga, N.; Miki, M.; Yanagi, K.; Imai, A. Organo-
metallics 1998, 17, 2152-2154.
(10) Tsukahara, T.; Swenson, D. C.; J ordan, R. F. Organometallics
1997, 16, 3303-3313.
(29) Woo, T. K.; Margl, P. M.; Lohrenz, J . C. W.; Blochl, P. E.;
Ziegler, T. J . Am. Chem. Soc. 1996, 118, 13021-13030.
(30) Shapiro, P. J .; Cotter, W. D.; Schaefer, W. P.; Labinger, J . A.;
Bercaw, J . E. J . Am. Chem. Soc. 1994, 116, 4623-4640.
(31) Devore, D. D.; Timmers, F. J .; Hasha, D. L.; Rosen, R. K.;
Marks, T. J .; Deck, P. A.; Stern, C. L. Organometallics 1995, 14, 3132-
3134.
(32) Lindsay, K. F. Mod. Plast. 1993, 82.
10.1021/om049545i CCC: $27.50 © 2004 American Chemical Society
Publication on Web 09/25/2004

Catalyst Design
Organometallics, Vol. 23, No. 22, 2004 5241
Sch em e 1. Exa m p les of Non -m eta llocen e Ca ta lyst
P r ecu r sor s
of bulky electron-donating amides resulted in improved
catalyst performance.45 While computational studies of
polymerization catalysts have for the most part focused
on metallocene46-70 and constrained geometry-based
catalysts,48,49,51,57,60,62,71,72 group 4 phosphinimide cata-
lysts have only recently been investigated by Ziegler and
co-workers.58,59 While these experimental and theoreti-
cal investigations have illuminated various aspects of
successful olefin polymerization catalyst targets, in
general, the discovery of high-activity nonmetallocene
catalysts has, for the most part, been the product of
serendipity. In this article, we employ computational
methods to provide guidance for the synthetic chemistry
efforts of new phosphinimide catalysts. The energetics
of model reactions of phosphinimide catalysts with
ethylene, resulting in two consecutive insertions, are
computed as a function of the phosphinimide substitu-
ents using density functional theory. These results are
used together with judicious synthetic strategies to
uncover a new family of readily accessible, highly active
ethylene polymerization catalysts based on titanium
tris(amino)phosphinimide complexes.
commercialized by Dow and Exxon. More recently,
Fujita and co-workers at Mitsui Chemicals have devel-
oped the so-called “FI catalysts” (Fenokishi-Imin), which
are phenoxy-imine ligand complexes of zirconium and
titanium (C).34,35 In our own efforts to develop nonmet-
allocene catalyst systems, we have focused on Ti-
phosphinimide ligand complexes. While small substit-
uents on the phosphinimide ligands prompt unique
deactivation pathways, including formation of Ti(IV)
dications36 and multiple C-H activation processes,37-39
sterically demanding phosphinimide complexes such as
(C5H4t-Bu)TiCl2NPCy3 (D) or TiMe2(NPt-Bu3)2 (E) af-
ford highly active catalysts that sustain activity under
industrially relevant conditions.40-42 Further studies
involving a variety of sterically demanding phosphin-
imide ligands have revealed the complexity of catalyst
design issues for these ancillary ligands.43,44 The nature
of desirable electronic features is also unclear. Prelimi-
nary studies using a series of substituted triphenylphos-
phinimide Ti-precatalysts demonstrated increased ac-
tivity when electron-donating substituents were in-
corporated; however these catalysts operate in a rather
low activity regime.42 Nonetheless, a recent study of a
related CpTi-ketimide system showed that incorporation
Exp er im en ta l Section
Gen er a l Da ta . All preparations were done under an
atmosphere of dry, O2-free N2 employing both Schlenk line
techniques and an MBraun inert atmosphere glovebox. Sol-
(45) Kretschmer, W. P.; Dijkhuis, C.; Meetsma, A.; Hessen, B.;
Teuben, J . H. Chem. Commun. 2002, 608-609.
(46) Fusco, R.; Longo, L.; Masi, F.; Garbassi, F. Macromolecules
1997, 30, 7673-7685.
(47) Fusco, R.; Longo, L.; Proto, A.; Masi, F.; Garbassi, F. Macromol.
Rapid Commun. 1998, 19, 257-262.
(48) Lanza, G.; Fragala, I. L.; Marks, T. J . J . Am. Chem. Soc. 1998,
120, 8257-8258.
(49) Lanza, G.; Fragala, I. L. Top. Catal. 1999, 7, 45-60.
(50) Rappe, A. K.; Skiff, W. M.; Casewit, C. J . Chem. Rev. 2000,
100, 1435-1456.
(51) Lanza, G.; Fragala, I. L.; Marks, T. J . J . Am. Chem. Soc. 2000,
122, 12764-12777.
(52) Vanka, K.; Chan, M. S. W.; Pye, C. C.; Ziegler, T. Organo-
metallics 2000, 19, 1841-1849.
(53) Braga, D.; Grepioni, F.; Tedesco, E.; Calhorda, M. J . Z. Anorg.
Allg. Chem. 2000, 626, 462-470.
(54) Chan, M. S. W.; Ziegler, T. Organometallics 2000, 19, 5182-
5189.
(55) Nifant’ev, I. E.; Ustynyuk, L. Y.; Laikov, D. N. Organometallics
2001, 20, 5375-5393.
(56) Vanka, K.; Ziegler, T. Organometallics 2001, 20, 905-913.
(57) Lanza, G.; Fragala, I. L.; Marks, T. J . Organometallics 2002,
21, 5594-5612.
(58) Xu, Z.; Vanka, K.; Firman, T.; Michalak, A.; Zurek, E.; Zhu,
C.; Ziegler, T. Organometallics 2002, 21, 2444-2453.
(59) Vanka, K.; Xu, Z.; Ziegler, T. Isr. J . Chem. 2003, 42, 403-415.
(60) Woo, T. K.; Fan, L.; Ziegler, T. Organometallics 1994, 13, 2252-
2261.
(33) Stevens, J . C. In Catalyst Design for Taylor-made Polyolefins,
Studies in Surface Science and Catalysis; Soga, K., Terano, M., Eds.;
Elsevier: Amsterdam, 1994; Vol. 89, pp 277-284.
(34) Makio, H.; Kashiwa, N.; Fujita, T. Adv. Synth. Catal. 2002, 344,
477-493.
(35) Mitani, M.; Nakano, T.; Fujita, T. Chem. Eur. J . 2003, 9, 2396-
2403.
(36) Guerin, F.; Stephan, D. W. Angew. Chem., Int. Ed. 2000, 39,
1298-1300.
(61) Yoshida, T.; Koga, N.; Morokuma, K. Organometallics 1995,
14, 746-758.
(37) Kickham, J . E.; Guerin, F.; Stephan, D. W. J . Am. Chem. Soc.
2002, 124, 11486-11494.
(62) Woo, T. K.; Margl, P. M.; Ziegler, T.; Bloechl, P. E. Organome-
tallics 1997, 16, 3454-3468.
(38) Kickham, J . E.; Guerin, F.; Stewart, J . C.; Stephan, D. W.
Angew. Chem., Int. Ed. 2000, 39, 3263-3266.
(63) Margl, P.; Deng, L.; Ziegler, T. J . Am. Chem. Soc. 1999, 121,
154-162.
(39) Kickham, J . E.; Guerin, F.; Stewart, J . C.; Urbanska, E.;
Stephan, D. W. Organometallics 2001, 20, 1175-1182.
(40) Stephan, D. W.; Stewart, J . C.; Guerin, F.; Spence, R. E. v. H.;
Xu, W.; Harrison, D. G. Organometallics 1999, 18, 1116-1118.
(41) Stephan, D. W.; Guerin, 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. Organometallics 1999, 18, 2046-2048.
(64) Talarico, G.; Blok, A. N. J .; Woo, T. K.; Cavallo, L. Organome-
tallics 2002, 21, 4939-4949.
(65) Bierwagen, E. P.; Bercaw, J . E.; Goddard, W. A., III. J . Am.
Chem. Soc. 1994, 116, 1481-1489.
(66) Lohrenz, J . C. W.; Woo, T. K.; Fan, L.; Harrison, D.; Margl, P.;
Ziegler, T. Polym. Mater. Sci. Eng. 1996, 74, 391-392.
(67) Thorshaug, K.; Stovneng, J . A.; Rytter, E.; Ystenes, M. Mac-
romolecules 1998, 31, 7149-7165.
(42) Stephan, D. W.; Stewart, J . C.; Guerin, F.; Courtenay, S.;
Kickham, J .; Hollink, E.; Beddie, C.; Hoskin, A.; Graham, T.; Wei, P.;
Spence, R. E. v. H.; Xu, W.; Koch, L.; Gao, X.; Harrison, D. G.
Organometallics 2003, 22, 1937-1947.
(43) Carraz, C.-A.; Stephan, D. W. Organometallics 2000, 19, 3791-
3796.
(68) Margl, P. M.; Woo, T. K.; Ziegler, T. Organometallics 1998, 17,
4997-5002.
(69) Chan, M. S. W.; Vanka, K.; Pye, C. C.; Ziegler, T. Organo-
metallics 1999, 18, 4624-4636.
(44) Yue, N. L. S.; Stephan, D. W. Organometallics 2001, 20, 2303-
2308.
(70) Lanza, G.; Fragala, I. L.; Marks, T. J . Organometallics 2001,
20, 4006-4017.

5242 Organometallics, Vol. 23, No. 22, 2004
Beddie et al.
vents were purified employing a Grubbs type solvent purifica-
tion system manufactured by Innovative Technology. All
organic reagents were purified by conventional methods. H,
mL under vacuum, which resulted in the precipitation of a
white solid. In the glovebox the slurry was filtered and a white
solid collected in the filter frit. The solid was extracted with
∼250 mL of toluene. The solvent was removed from the filtrate
under vacuum to produce 6 as a white solid. Yield: 12.14 g,
31.0 mmol, 54%. 31P{1H} NMR: δ 110.4. 1H NMR: δ 7.17-
1
31P{1H}, and 13C{1H} NMR spectra were recorded on Bruker
Avance-300 and -500 spectrometers. All spectra were recorded
in C6D6 at 25 °C unless otherwise noted. Trace amounts of
protonated solvents were used as references, and chemical
shifts are reported relative to SiMe4. The 31P{1H} NMR spectra
were referenced to external 85% H3PO4. Combustion analyses
were done in-house employing a Perkin-Elmer CHN analyzer.
Me3SiN3, P(NMe2)3 (1), and P(NEt2)3 (2) were purchased from
Aldrich Chemical Co. Cp*TiCl3 was purchased from Strem
Chemical Co., and P(Nn-Pr2)3 (3) was purchased from Lan-
caster Chemicals. B(C6F5)3 and MAO were generously donated
by NOVA Chemicals Corporation. CpTiCl3,73 P(Nn-Bu2)3 (4),74
P(N-3-methylindolyl)3 (7),75 Me3SiNP(NMe2)3 (8),76 Me3SiNP-
(NEt2)3 (9),77 Me3SiNP(Nn-Bu2)3 (11),78 and CpTiCl2NP(NMe2)3
(15)79 were prepared via literature methods.
3
4
7.11 (m, 12H, NPh), 6.86 (tt, 3H, J H-H ) 7 Hz, J H-H ) 2 Hz,
3
3
NPh), 3.37 (dq, 6H, J P-H ) 3 Hz, J H-H ) 7 Hz, NCH2Me),
0.94 (t, 9H, J H-H ) 7 Hz, NCH2Me). 13C{1H} NMR: δ 147.4
3
(d, 2J P-C ) 22 Hz, NPh (ipso)), 129.6 (NPh), 122.1 (NPh), 121.9
(NPh), 40.9 (NCH2Me), 14.2 (NCH2Me). Anal. Found: C, 73.99;
H, 7.87; N, 10.74. Calcd: C, 73.63; H, 7.72; N, 10.73. X-ray
quality crystals were obtained by slow evaporation of toluene.
Syn t h eses of Me3SiNP (NMe2)3 (8),76 Me3SiNP (NE t2)3
(9),77 Me3SiNP (Nn -P r 2)3 (10), Me3SiNP (Nn -Bu 2)3 (11),78
Me3SiNP (Ni-P r Me)3 (12), a n d Me3SiNP (NEtP h )3 (13).78
These compounds were prepared in a similar manner, and thus
only one synthesis is detailed. To a solution of P(NMe2)3 (3.00
mL, 16.5 mmol) in toluene (30 mL) was added Me3SiN3 (3.29
mL, 24.8 mmol). The resulting solution was heated at refluxing
temperature for 12 h. The solvent and excess Me3SiN3 were
removed under vacuum, resulting in a crude oil that was
purified by vacuum distillation to give a clear, colorless oil.
Syn th esis of P (Ni-P r Me)3 (5).80 To a solution of PCl3 (2.07
mL, 23.7 mmol) and NEt3 (11.6 mL, 83.2 mmol) in ether (300
mL), cooled to -78 °C, was slowly added HNi-PrMe (8.65 mL,
83.1 mmoL) via syringe. The resulting white slurry was
allowed to warm to room temperature over a period of 1 h.
The slurry was filtered via a filter cannula to a produce a clear,
yellow solution. The white solid was washed with 3 × 75 mL
of ether, and the washings were transferred through the filter
cannula and added to the original filtrate. The solvent was
partially removed under vacuum to produce a white suspen-
sion. The white suspension was filtered through Celite, and
the residual white solid was washed with 10 mL of ether. The
solvent and excess HN(i-Pr)(Me) were removed from the
filtrate under vacuum to produce an oil, which was purified
by vacuum distillation, yielding a clear colorless liquid. 3: yield
1
8:76 Yield: 3.05 g, 12.2 mmol, 74%. 31P{1H} NMR: δ 14.9. H
3
NMR: δ 2.40 (d, 18H, J P-H ) 10 Hz, NMe2), 0.39 (s, 9H,
2
SiMe3). 13C{1H} NMR: δ 37.6 (d, J PC ) 3 Hz, NMe2), 5.1
(SiMe3). 9:77 Yield: 2.80 g, 8.38 mmol, 68%. 31P{1H} NMR: δ
1
3
3
9.7. H NMR: δ 2.94 (dq, 12H, J P-H ) 10 Hz, J H-H ) 7 Hz,
3
NCH2), 0.99 (t, 18H, J H-H ) 7 Hz, (CH2Me)), 0.39 (s, 9H,
SiMe3). 13C{1H} NMR: δ 40.0 (d, J PC ) 5 Hz, NCH2), 14.7 (d,
2
3J PC ) 3 Hz, CH2Me), 4.9 (SiMe3). 10: Yield: 2.25 g, 5.38 mmol,
89%. 31P{1H} NMR: δ 10.3. 1H NMR: δ 2.89 (m, 12H, N(CH2)),
3
1.53 (m, 12H, CH2CH2Me), 0.84 (t, 18H, J H-H ) 7 Hz, CH2-
1
3.17 g, 12.8 mmol, 54%. 31P{1H} NMR: δ 119.7. H NMR: δ
Me), 0.41 (s, 9H, SiMe3). 13C{1H} NMR: δ 49.2 (d, J P-C ) 4
2
3
3.58 (d of sept, 3H, J P-H ) 10 Hz,3J H-H ) 7 Hz CHMe2), 2.34
3
Hz, NCH2), 23.0 (d, J P-C ) 2 Hz, CH2CH2Me), 12.2 (s, CH2-
3
3
(d, 9H, J P-H ) 6 Hz, NMe), 1.09 (d, 18H, J H-H ) 7 Hz,
Me), 5.0 (s, SiMe3). Anal. Found: C, 59.33; H, 11.39; N, 14.65
Calcd: C, 60.24; H, 12.28; N, 13.38. 11:78 Yield: 2.19 g, 4.35
mmol, 91%. 31P{1H} NMR: δ 10.9 (s). 1H NMR: δ 3.02 (dt,
CHMe2). 13C{1H} NMR: δ 49.2 (d, 2J PC ) 36 Hz, CHMe2), 26.5
(d, J PC ) 5 Hz, NMe), 21.1 (d, J PC ) 4 Hz, CHMe2).
2
3
3
12H, J P-H ) 10 Hz,3J H-H ) 8 Hz NCH2), 1.59 (m, 12H, CH2-
Syn th esis of P (NEtP h )3 (6).81 To a solution of PCl3 (5.0
mL, 57.3 mmol) and NEt3 (28.0 mL, 201 mmol) in toluene (400
mL) cooled to 0 °C was slowly added HNEtPh (25.3 mL, 187
mmol) via syringe. The white suspension was heated at
refluxing temperature for 12 h. The mixture was cooled to room
temperature and then filtered through a cannula filter to
produce a yellow solution. The volatiles were removed under
vacuum to produce a yellow solid. 31P{1H} NMR revealed the
presence of two products; therefore 300 mL of toluene, NEt3
(16.0 mL, 115 mmol), and HNEtPh (7.2 mL, 57.2 mmol) were
added. The mixture was heated at refluxing temperature for
an additional 12 h. The mixture was cooled to room temper-
ature and then filtered through a cannula filter to produce a
yellow solution. The volume of the solution was reduced to ∼50
3
CH2CH2), 1.28 (pseudo sextet, 12H, J H-H ) 7 Hz, CH2CH2-
3
Me), 0.98 (t, 18H, J H-H ) 7 Hz, CH2Me), 0.43 (s, 9H, SiMe3).
2
3
13C{1H} NMR: δ 47.1 (d, J P-C ) 4 Hz, NCH2), 32.1 (d, J P-C
) 2 Hz, CH2CH2CH2), 21.5 (s, CH2CH2Me), 14.7 (s, CH2Me),
5.1 (s, SiMe3). 12: Yield: 2.52 g, 7.54 mmol, 68%. 31P{1H}
NMR: δ 10.6. 1H NMR: δ 4.00 (d(sept), 3H, J P-H ) 10
3
Hz,3J H-H ) 7 Hz, CHMe2), 2.25 (d, 9H, J P-H ) 10 Hz, NMe),
3
1.01 (d, 18H, J H-H ) 7 Hz, CHMe2), 0.36 (s, 9H, SiMe3). 13C-
3
{1H} NMR: δ 46.1 (d, J P-C ) 5 Hz, CHMe2), 26.7 (d, J P-C
)
2
2
4 Hz, NMe), 20.7 (CHMe2), 4.9 (SiMe3). Anal. Found: C, 53.21;
H, 11.30; N, 17.16. Calcd: C, 53.85; H, 11.75; N, 16.75. 13:
Yield: 3.56 g, 7.43 mmol, 90%. 31P{1H} NMR: δ -5.8. 1H
NMR: δ 7.21 (d, 6H, 3J H-H ) 8 Hz, NPh), 7.16 (t, 6H, 3J H-H
)
3
8 Hz, NPh), 7.00 (t, 3H, J H-H ) 8 Hz, NPh), 3.38 (dq, 6H,
3
3
3J P-H ) 7 Hz, J H-H ) 7 Hz, NCH2Me), 0.79 (t, 9H, J H-H ) 7
(71) Woo, T. K.; Margl, P. M.; Lohrenz, J . C. W.; Bloechl, P. E.;
Ziegler, T. J . Am. Chem. Soc. 1996, 118, 13021-13030.
(72) Xu, Z.; Vanka, K.; Ziegler, T. Organometallics 2004, 23, 104-
116.
(73) Gorsich, R. D. J . Am. Chem. Soc. 1960, 82, 4211-4214.
(74) Vogt, H.; Wulff-Molder, D.; Ritschl, F.; Mucke, M.; Skrabei, U.;
Meisel, M. Z. Anorg. Allg. Chem. 1999, 625, 1025-1027.
(75) Barnard, T. S.; Mason, M. R. Organometallics 2001, 20, 206-
214.
Hz, NCH2Me), 0.04 (s, 9H, SiMe3). 13C{1H} NMR: δ 144.9 (s,
NPh (ipso)), 129.9 (d, J P-C ) 3 Hz, NPh), 128.6 (NPh), 125.7
3
(NPh), 46.9 (NCH2Me), 14.5 (NCH2Me), 3.9 (SiMe3).
Syn th eses of Cp ′TiCl2(NP (NR2)3) (Cp ′ ) Cp , R ) Me
15,79 Et 16, P r 17, Bu 18, R2 ) i-P r Me 19, EtP h 20; Cp ′ )
Cp *, R ) Me 21, Et 22, P r 23, Bu 24, R2 ) i-P r Me 25, EtP h
26). These compounds were prepared in a similar manner, and
thus only one synthesis is detailed. To a yellow solution of
CpTiCl3 (0.400 g, 1.82 mmol) in toluene (30 mL) was added a
solution of Me3SiNP(NMe2)3 (8) (0.480 g, 1.92 mmol) in toluene
(20 mL). The resulting solution was stirred at room temper-
ature for 12 h. The volume of the solution was reduced under
vacuum to cause the formation of a yellow solid. The solid was
washed with hexanes (3 × 5 mL) and dried under vacuum to
give a yellow solid. 15: Yield: 0.630 g, 1.74 mmol, 96%. 31P-
(76) Schlak, O.; Stadelmann, W.; Stelzer, O.; Schmutzler, R. Z.
Anorg. Allg. Chem. 1976, 419, 275-282.
(77) Flindt, E. P.; Rose, H.; Marsmann, H. C. Z. Anorg. Allg. Chem.
1977, 430, 155-160.
(78) Marchenko, A. P.; Koidan, G. N.; Pinchuk, A. M.; Kirsanov, A.
V. Z. Obsh. Khim. 1984, 54, 1774-1782.
(79) Spence, R. E. V. H.; Koch, L.; J eremic, D.; Brown, S. J . (Nova
Chemicals (International) S.A., Switz.) PCT Int. Appl.; WO, 2000, 32
pp.
(80) Nagato, N.; Ogawa, M.; Naito, T. (Showa Denko K. K.) J pn.
Kokai Tokkyo Koho; J p, 1973; 6 pp.
(81) Marchenko, A. P.; Pinchuk, A. M.; Feshchenko, N. G. Z. Obsh.
Khim. 1973, 43, 1900-1903.
1
{1H} NMR: δ 5.6. H NMR: δ 6.44 (s, 5H, Cp), 2.24 (d, 18H,
3J P-H ) 10 Hz, NMe2). 13C{1H} NMR: δ 115.4 (Cp), 37.1 (d,

Catalyst Design
Organometallics, Vol. 23, No. 22, 2004 5243
2J PC ) 4 Hz, NMe2). Anal. Found: C, 36.31; H, 6.77; N, 15.39.
Calcd: C, 36.59; H, 6.42; N, 15.52. 16: Yield: 1.28 g, 2.88
mmol, 95%. 31P{1H} δ NMR: 3.4. 1H NMR: δ 6.48 (s, 5H, Cp),
Anal. Found: C, 51.47; H, 9.14; N, 10.85. Calcd: C, 51.27; H,
8.80; N, 10.87. 26: Yield 0.786 g, 1.19 mmol, 85%. 31P{1H}
1
NMR: δ -5.9. H NMR: δ 7.16 (d, 6H, partially obscured by
3
3
3
2.82 (dq, 12H, J P-H ) 11 Hz, J H-H ) 7 Hz, CH2Me), 0.90 (t,
18H, 3J H-H ) 7 Hz, CH2Me). 13C{1H} NMR: δ 115.3 (Cp), 39.9
(d, 2J P-C ) 5 Hz, CH2Me), 14.2 (d, 3J P-C ) 3 Hz, CH2Me). Anal.
Found: C, 45.39; H, 7.99; N, 12.59. Calcd: C, 45.86; H, 7.92;
N, 12.58. 17: Yield: 0.234 g, 0.442 mmol, 96%. 31P{1H}
NMR: 3.8. 1H NMR: 6.50 (s, 5H, Cp), 2.85 (m, 12H, CH2CH2-
C6D6, NPh (ortho)), 7.09 (pseudo t, 6H, J H-H ) 8 Hz, NPh
3
(meta)), 6.96 (t, 3H, J H-H ) 7 Hz, NPh (para)), 3.62 (dq, 6H,
3
3J P-H ) 7 Hz, J H-H ) 7 Hz, NCH2Me), 2.15 (s, 15H, Cp*),
3
0.57 (t, 9H, J H-H ) 7 Hz, NCH2Me). 13C{1H} NMR: δ 142.5
(NPh (ipso)), 131.0 (d, 3J P-C ) 3 Hz, NPh (ortho)), 129.4 (NPh
2
(meta)), 126.9 (s, Cp*), 126.7 (NPh (para)), 47.0 (d, J P-C ) 5
3
3
Hz, NCH2Me), 13.8 (d, J P-C ) 5 Hz, NCH2Me) 13.5 (Cp*).
Me), 1.48 (m, 12H, CH2CH2Me), 0.82 (t, 18H, J H-H ) 7 Hz,
CH2Me). 13C{1H} NMR: 115.3 (Cp), 48.5 (d, 2J P-C ) 4 Hz, CH2-
Anal. Found: C, 61.65; H, 7.16; N, 8.74. Calcd: C, 61.92; H,
6.88; N, 8.50.
3
CH2Me), 22.5 (d, J P-C ) 3 Hz, CH2CH2Me), 12.0 (s, CH2Me).
Anal. Found: C, 51.42; H, 8.96; N, 10.64. Calcd: C, 52.18; H,
Syn th eses of Cp ′TiMe2(NP (NR2)3) (Cp ′ ) Cp , R ) Me
27,79 Et 28, P r 29, Bu 30, R2 ) i-P r Me 31, EtP h 32; Cp ′ )
Cp *, R ) Me 33, Et 34, P r 35, Bu 36, R2 ) i-P r Me 37, EtP h
38). These compounds were prepared in a similar manner, and
thus only one synthesis is detailed. To a yellow slurry of
CpTiCl2(NP(NMe2)3) (0.201 g, 0.557 mmol) in ether (10 mL)
was added 1.4 M MeLi in ether (0.83 mL, 1.162 mmol). The
solvent was removed immediately under vacuum to produce
a green residue. The green residue was extracted with hexanes
(10 mL), and the LiCl was removed by filtration. The solvent
was removed under vacuum to produce a green residue. 27:
Yield 0.152 g, 0.475 mmol, 85%. 31P{1H} NMR: δ 1.5. 1H
8.95; N, 10.58. 18: Yield: 0.277 g, 0.451 mmol, 97%. 31P{1H}
1
NMR: δ 4.5. H NMR: δ 6.53 (s, 5H, Cp), 3.00 (m, 12H, CH2-
CH2CH2), 1.57 (m, 12H, CH2CH2CH2Me), 1.28 (pseudo sextet,
3
3
12H, J H-H ) 7 Hz, CH2CH2CH2Me), 0.94 (t, 18H, J H-H ) 7
Hz, CH2Me). 13C{1H} NMR: δ 115.3 (Cp), 46.6 (d, J P-C ) 4
Hz, NCH2), 31.5 (d, J P-C ) 3 Hz, CH2CH2CH2), 21.2 (s, CH2-
2
3
CH2Me), 14.6 (s, CH2Me). Anal. Found: C, 55.83; H, 9.23; N,
8.23. Calcd: C, 56.77; H, 9.69; N, 9.13. 19: Yield: 1.15 g, 2.59
mmol, 86%. 31P{1H} NMR: δ 3.4. 1H NMR: δ 6.47 (s, 5H, Cp),
3
4.04 (d(sept), 3H, J P-H ) 10 Hz,3J H-H ) 7 Hz CHMe2), 2.07
3
3
(d, 9H, J P-H ) 10 Hz, NMe), 0.96 (d, 18H, J H-H ) 7 Hz,
CHMe2). 13C{1H} NMR: δ 115.2 (Cp), 46.9 (d, J P-C ) 5 Hz,
2
3
NMR: δ 6.22 (s, 5H, Cp), 2.40 (d, 18H, J P-H ) 10 Hz, NMe),
2
3
0.77 (s, 6H, TiMe). 13C{1H} NMR: δ 111.2 (Cp), 41.2 (TiMe),
CHMe2), 26.6 (d, J P-C ) 4 Hz, NMe), 20.7 (d, J P-C ) 3 Hz,
CHMe2). Anal. Found: C, 45.56; H, 8.14; N, 12.45. Calcd: C,
45.86; H, 7.92; N, 12.58. X-ray quality crystals were obtained
from slow evaporation of a toluene/hexanes solution. 20:
Yield: 0.946 g, 1.61 mmol, 88%. 31P{1H} NMR: δ -11.1. 1H
NMR: δ 7.21 (d, 6H, 3J H-H ) 8 Hz, NPh (ortho)), 7.13 (pseudo
2
37.4 (d, J P-C ) 3 Hz, NMe). Anal. Found: C, 47.54; H, 9.09;
N, 17.85. Calcd: C, 48.76; H, 9.13; N, 17.50. 28: Yield 0.198
1
g, 0.490 mmol, 99%. 31P{1H} NMR: δ -1.0. H NMR: δ 6.22
(s, 5H, Cp), 2.98 (dq, 12H, 3J P-H ) 10 Hz, 3J H-H ) 7 Hz, NCH2-
3
Me), 0.97 (t, 18H, J H-H ) 7 Hz, CH2Me), 0.65 (s, 6H, TiMe).
3
3
2
13C{1H} NMR: δ 111.1 (Cp), 40.9 (TiMe), 39.9 (d, J P-C ) 4
t, 6H, J H-H ) 8 Hz, NPh (meta)), 6.97 (t, 3H, J H-H ) 8 Hz,
NPh (para)), 6.03 (s, 5H, Cp), 3.38 (dq, 6H, 3J P-H ) 7 Hz, 3J H-H
Hz, NCH2), 14.5 (s, CH2Me). Anal. Found: C, 55.69; H, 10.03;
) 7 Hz, NCH2Me), 0.66 (t, 9H, J H-H ) 7 Hz, NCH2Me). 13C-
3
N, 13.53. Calcd: C, 56.43; H, 10.22; N, 13.85. 29: Yield: 0.130
{1H} NMR: δ 141.9 (d, J P-C ) 4 Hz, NPh (ipso)), 130.9 (d,
2
1
g, 0.266 mmol, 87%. 31P{1H} NMR: δ -0.6. H NMR: δ 6.27
3J P-C ) 3 Hz, NPh (ortho)), 129.6 (s, NPh (meta)), 127.2 (s,
(s, 5H, Cp), 2.97 (m, 12H, NCH2CH2Me), 1.53 (m, 12H, NCH2-
2
3
NPh (para)), 116.2 (s. Cp), 46.8 (d, J P-C ) 4 Hz, NCH2Me),
CH2Me), 0.82 (t, 18H, J H-H ) 7 Hz, CH2Me), 0.71 (s, 6H,
3
2
TiMe). 13C{1H} NMR: δ 111.2 (Cp), 48.9 (d, J P-C ) 4 Hz,
14.0 (d, J P-C ) 4 Hz, NCH2Me). Anal. Found: C, 58.72; H,
3
6.08; N, 9.61. Calcd: C, 59.10; H, 5.99; N, 9.51. 21: Yield:
NCH2CH2), 41.1 (s, TiMe), 22.7 (d, J P-C ) 3 Hz, NCH2CH2-
0.505 g, 1.17 mmol, 68%. 31P{1H} NMR: 6.1 (s). 1H NMR: 2.35
Me), 12.1 (s, CH2Me). Anal. Found: C, 60.58; H, 11.27; N 11.28.
Calcd: C, 61.46; H, 10.93; N, 11.47. 30: Yield: 0.198 g, 0.346
mmol, 76%. 31P{1H} NMR: δ 0.0. 1H NMR: δ 6.30 (s, 5H, Cp),
3.12 (m, 12H, NCH2CH2), 1.61 (m, 12H, NCH2CH2CH2), 1.27
(pseudo sextet, 12H, 3J H-H ) 7 Hz, CH2Me), 0.94 (t, 18H, 3J H-H
) 7 Hz, CH2Me), 0.72 (s, 6H, TiMe). 13C{1H} NMR: δ 111.2
(d, 18H, J P-H ) 10 Hz, NMe2), 2.21 (s, 15H, Cp*). 13C{1H}
3
2
NMR: 126.1 (Cp*), 37.5 (d, J P-C ) 4 Hz, NMe2), 13.4 (Cp*).
Anal. Found: C, 44.33; H, 7.57; N, 13.05. Calcd: C, 44.57; H,
7.71; N, 12.99. X-ray quality crystals were obtained from slow
evaporation of a toluene/hexanes solution. 22: Yield: 0.586 g,
1.14 mmol, 57%. 31P{1H} NMR: δ 5.6. 1H NMR: δ 2.97 (dq,
2
(Cp), 46.9 (d, J P-C ) 4 Hz, NCH2CH2), 41.2 (TiMe), 31.8 (d,
3
3
3J P-C ) 3 Hz, NCH2CH2CH2), 21.4 (s, CH2Me), 14.7 CH2Me).
Anal. Found: C, 63.98; H, 10.83; N, 10.72. Calcd: C, 65.01;
H, 11.44; N, 9.78. 31: Yield: 0.175 g, 0.433 mmol, 93%. 31P-
{1H} NMR: δ -1.0. 1H NMR: δ 6.21 (s, 5H, Cp), 4.27 (d(sept),
12H, J P-H ) 11 Hz, J H-H ) 7 Hz, NCH2Me), 2.22 (s, 15H,
Cp*), 0.95 (t, 18H, J H-H ) 7 Hz, CH2Me). 13C{1H} NMR: δ
3
2
3
125.7 (Cp*), 39.8 (d, J P-C ) 5 Hz, CH2Me), 14.1 (d, J P-C ) 3
Hz, CH2Me), 13.4 (Cp*). Anal. Found: C, 50.81; H, 8.60; N,
11.06. Calcd: C, 51.27; H, 8.80; N, 10.87. X-ray quality crystals
were obtained from slow evaporation of a toluene/hexanes
solution. 23: Yield: 0.574 g, 0.957 mmol, 96%. 31P{1H} NMR:
δ 5.0.1H NMR: δ 2.99 (m, 12H, NCH2), 2.24 (s, 15H, Cp*),
3H, 3J P-H ) 10 Hz,3J H-H ) 7 Hz CHMe2), 2.21 (d, 9H, 3J P-H
)
3
10 Hz, NMe), 1.01 (d, 18H, J H-H ) 7 Hz, CHMe2), 0.64 (s,
6H, TiMe). 13C{1H} NMR: δ 111.1 (Cp), 46.5 (d, 2J P-C ) 5 Hz,
2
CHMe2), 40.8 (TiMe), 26.6 (d, J P-C ) 3 Hz, NMe), 20.7 (d,
3
3J P-C ) 3 Hz, CHMe2). Anal. Found: C, 55.48; H, 10.34; N,
14.21. Calcd: C, 56.43; H, 10.22; N, 13.85. 32: Yield: 0.196 g,
0.357 mmol, 99%. 31P{1H} NMR: δ -16.1. 1H NMR: δ 7.24
1.52 (m, 12H, CH2CH2Me), 0.85 (t, 18H, J H-H ) 7 Hz, CH2-
Me).13C{1H} NMR: δ 125.8 (Cp*), 48.4 (NCH2), 22.1 (CH2CH2-
Me), 13.5 (Cp*), 12.0 (CH2Me). Anal. Found: C, 55.15; H, 9.41;
N, 9.27. Calcd: C, 56.09; H, 9.58; N, 9.35. 24: Yield: 0.670 g,
3
3
(d, 6H, J H-H ) 8 Hz, NPh (ortho)), 7.14 (pseudo t, 6H, J H-H
0.980 mmol, 91%. 31P{1H} NMR: δ 5.4. H NMR: δ 3.00 (m,
1
3
) 8 Hz, NPh (meta)), 6.98 (t, 3H, J H-H ) 7 Hz, NPh (para)),
3
3
12H, NCH2CH2), 2.25 (s, 15H, Cp*), 1.61 (m, 12H, CH2CH2-
5.85 (s, 5H, Cp), 3.47 (dq, 6H, J P-H ) 7 Hz, J H-H ) 7 Hz,
3
3
CH2), 1.32 (pseudo sextet, 12H, J H-H ) 7 Hz, CH2CH2Me),
NCH2Me), 0.77 (t, 9H, J H-H ) 7 Hz, NCH2Me), 0.66 (s, 6H,
3
0.96 (t, 18H, J H-H ) 7 Hz, CH2Me). 13C{1H} NMR: δ 125.8
TiMe). 13C{1H} NMR: δ 143.8 (d, J P-C ) 4 Hz, NPh (ipso)),
2
2
3
131.2 (d, 3J P-C ) 3 Hz, NPh (ortho)), 129.3 (NPh (meta)), 126.3
(NPh (para)), 111.7 (Cp), 46.5 (d, 2J P-C ) 4 Hz, NCH2Me), 43.7
(TiMe), 14.4 (CH2Me). Anal. Found: C, 68.03, H, 7.42; N, 9.94.
Calcd: C, 67.88; H, 7.53; N, 10.21. 33: Yield: 0.171 g, 0.438
mmol, 90%. 31P{1H} NMR: δ 0.4. 1H NMR: δ 2.50 (d, 18H,
3J P-H ) 10 Hz, NMe2), 2.08 (s, 15H, Cp*), 0.49 (s, 6H, TiMe2).
(Cp*), 46.4 (d, J P-C ) 4 Hz, NCH2), 31.1 (d, J P-C ) 3 Hz,
CH2CH2CH2), 21.3 (CH2CH2Me), 14.7 (CH2Me), 13.5 (Cp*).
Anal. Found: C, 59.50; H, 10.51; N, 8.18. Calcd: C, 59.73; H,
10.17; N, 8.19. 25: Yield: 0.874 g, 1.70 mmol, 70%. 31P{1H}
NMR: δ 4.4. 1H NMR: δ 3.99 (d(sept), 3H, 3J P-H ) 10 Hz,3J H-H
3
) 7 Hz CHMe2), 2.22 (d, 9H, J P-H ) 11 Hz, NMe), 2.21 (s,
3
2
15H, Cp*), 1.01 (d, 18H, J H-H ) 6.7 Hz, CHMe2). 13C{1H}
13C{1H} NMR: δ 119.0 (Cp*), 43.1 (TiMe), 37.7 (d, J P-C ) 4
2
NMR: δ 125.7 (Cp*), 46.7 (d, J P-C ) 5 Hz, CHMe2), 27.0 (d,
Hz, NMe2), 12.6 (Cp*). Anal. Found: C, 55.17; H, 10.16; N,
2J P-C ) 4 Hz, NMe), 20.8 (d, 3J P-C ) 3 Hz, CHMe2), 13.4 (Cp*).
14.59. Calcd: C, 55.38; H, 10.07; N, 14.35. 34: Yield: 0.180 g,

5244 Organometallics, Vol. 23, No. 22, 2004
Beddie et al.
Ta ble 1. Ca lcu la ted Solu tion P h a se Rela tive En er giesa (k ca l/m ol) for th e Mech a n ism Dep icted in F igu r e 1
R
1′
2′
3′
4′
TS1
5′
6′
7′
8′
TS2
9′
Me
NH2
Hb
0
0
0
0
0
0
-23.7
-22.9
-19.1
-14.3
-5.9
17.3
17.7
25.4
74.8
42.5
40.5
0
0
6.5
53.4
22.1
19.8
6.4
6.6
12.5
60.2
27.2
24.5
-9.1
-9.0
-2.9
44.9
12.2
9.6
-12.8
-12.8
-5.8
41.8
9.9
-42.2
-44.9
-37.2
-34.4
-27.4
-27.2
-21.3
-22.5
-15.9
28.0
0.0
-2.6
-18.5
-19.2
-13.2
31.2
0.4
-2.3
-35.5
-35.8
-30.0
15.1
-16.1
-18.8
Cl
F
-5.6
7.4
a
b
B3LYP/BS2//B3LYP/LANL2DZ COSMO. Solvent toluene (ꢀ ) 2.379). Italicized numbers refer to B3LYP/BS2//B3LYP/LANL2DZ
gas phase relative energies.
0.379 mmol, 93%. 31P{1H} NMR: δ -0.9. 1H NMR: δ 3.11 (dq,
X-r a y Da ta Collection a n d Red u ction . Crystals were
manipulated and mounted in capillaries in a glovebox, thus
maintaining a dry, O2-free environment for each crystal.
Diffraction experiments were performed on a Siemens SMART
System CCD diffractometer. The data were collected in a
hemisphere of data in 1329 frames with 10 s exposure times.
The observed extinctions were consistent with the space groups
in each case. The data sets were collected (4.5° < 2θ < 45-
50.0°). A measure of decay was obtained by re-collecting the
first 50 frames of each data set. The intensities of reflections
within these frames showed no statistically significant change
over the duration of the data collections. The data were
processed using the SAINT and XPREP processing packages.
An empirical absorption correction based on redundant data
was applied to each data set. Subsequent solution and refine-
ment was performed using the SHELXTL solution package
operating on a Pentium computer.
3
3
12H, J P-H ) 10 Hz, J H-H ) 7 Hz, NCH2Me), 2.08 (s, 15H,
3
Cp*), 1.00 (t, 18H, J H-H ) 7 Hz, CH2Me), 0.43 (s, 6H, TiMe).
2
13C{1H} NMR: δ 118.7 (Cp*), 43.2 (TiMe), 39.7 (d, J P-C ) 4
Hz, NCH2Me), 14.4 (CH2Me), 12.5 (Cp*). Anal. Found: C,
60.00; H, 10.89; N, 11.55. Calcd: C, 60.75; H, 10.83; N, 11.81.
35: Yield: 0.163 g, 0.272 mmol, 83%. 31P{1H} NMR: δ -1.0.1H
NMR: δ 3.10 (m, 12H, NCH2CH2), 2.10 (s, 15H, Cp*), 1.55
(m, 12H, NCH2CH2), 0.85 (t, 18H, 3J H-H ) 7 Hz, CH2Me), 0.45
(s, 6H, TiMe). 13C{1H} NMR: δ 118.4 (Cp*), 48.2 NCH2), 42.9
(TiMe), 21.8 (NCH2CH2Me), 12.2 (Cp*), 11.8 (CH2Me). Anal.
Found: C, 64.74; H, 11.53; N, 9.83. Calcd: C, 64.49; H, 11.37;
N, 10.03. 36: Yield: 0.118 g, 0.173 mmol, 90%. 31P{1H}
NMR: δ -0.7. 1H NMR: δ 3.23 (m, 12H, NCH2), 2.12 (s, 15H,
Cp*), 1.63 (m, 12H, NCH2CH2CH2), 1.30 (m, 12H, CH2Me),
3
0.95 (t, 18H, J H-H ) 7 Hz, CH2Me), 0.45 (s, 6H, TiMe). 13C-
{1H} NMR: δ 118.4 (Cp*), 46.2 (NCH2), 43.0 (TiMe), 31.4 (d,
3J P-C ) 3 Hz, NCH2CH2CH2), 21.1 CH2Me), 14.4 (s, CH2Me),
12.2 (Cp*). Anal. Found: C, 66.58; H, 11.60; N, 8.34. Calcd:
C, 67.26; H, 11.76; N, 8.72. 37: Yield: 0.171 g, 0.360 mmol,
93%. 31P{1H} NMR: δ -1.7. 1H NMR: δ 4.31 (d(sept), 3H,
St r u ct u r e Solu t ion a n d R efin em en t . Non-hydrogen
atomic scattering factors were taken from the literature
tabulations.82 The heavy atom positions were determined using
direct methods employing the SHELXTL direct methods
routine. The remaining non-hydrogen atoms were located from
successive difference Fourier map calculations. The refine-
ments were carried out by using full-matrix least-squares
techniques on F, minimizing the function w(|Fo| - |Fc|)2 where
the weight w is defined as 4Fo2/2σ(Fo2) and Fo and Fc are the
observed and calculated structure factor amplitudes. In the
final cycles of each refinement, all non-hydrogen atoms were
assigned anisotropic temperature factors in the absence of
disorder or insufficient data. In the latter cases atoms were
treated isotropically. C-H atom positions were calculated and
allowed to ride on the carbon to which they are bonded
assuming a C-H bond length of 0.95 Å. H atom temperature
factors were fixed at 1.10 times the isotropic temperature
factor of the C atom to which they are bonded. The H atom
contributions were calculated, but not refined. The locations
of the largest peaks in the final difference Fourier map
calculation as well as the magnitude of the residual electron
densities in each case were of no chemical significance.
Addition details are provided Table 4 and in the Supporting
Information.
3
3J P-H ) 10 Hz,3J H-H ) 7 Hz CHMe2), 2.32 (d, 9H, J P-H ) 10
3
Hz, NMe), 2.07 (s, 15H, Cp*), 1.05 (d, 18H, J H-H ) 7 Hz,
CHMe2), 0.41 (s, 6H, TiMe). 13C{1H} NMR: δ 118.8 (Cp*), 46.2
2
2
(d, J P-C ) 6 Hz, CHMe2), 43.0 (TiMe), 27.0 (d, J P-C ) 3 Hz,
3
NMe), 20.9 (d, J P-C ) 3 Hz, CHMe2), 12.5 (Cp*). Anal.
Found: C, 60.70; H, 11.12; N, 11.59. Calcd: C, 60.75; H, 10.83;
N, 11.81. 38: Yield: 0.163 g, 0.263 mmol, 88%. 31P{1H}
NMR: δ -13.4. 1H NMR: δ 7.21 (d, 6H, J H-H ) 7 Hz NPh
3
3
(ortho)), 7.13 (pseudo t, 6H, J H-H ) 7, NPh (meta)), 6.97 (t,
3
3
3H, J H-H ) 7 Hz, NPh (para)), 3.65 (dq, 6H, J P-H ) 7 Hz,
3J H-H ) 7 Hz, NCH2Me), 1.97 (s, 15H, Cp*), 0.71 (t, 9H, 3J H-H
) 7 Hz, NCH2Me), 0.55 (s, 6H, TiMe). 13C{1H} NMR: δ 144.1
(NPh (ipso)), 130.0 (NPh), 129.2 (NPh), 125.9 (NPh), 119.6
(Cp*), 46.5 (NCH2Me), 45.9 (TiMe), 14.2 (NCH2Me), 12.5 (Cp*).
Anal. Found: C, 69.45, H, 8.48; N, 9.08. Calcd: C, 69.89; H,
8.31; N, 9.06.
P olym er iza tion P r otocol. Purification of reagents used
in the polymerizations was similar to that described above.
Polymerization experiments were performed using a Bu¨chi
polymerization reactor. In all polymerizations, 600 mL of
toluene was transferred into the reactor, heated to 30 ( 2 °C,
and presaturated with ethylene prior to injection of precatalyst
and cocatalyst. The solution was stirred at 1000 rpm for the
duration of the polymerization experiment. At the end of the
experiments, the contents of the reactor were emptied into a
4 L beaker containing approximately 200 mL of 10% HCl (v/
v) in MeOH. The polymer that precipitated was collected by
filtration, washed with water and acetone, and dried to a
constant weight. Polymerization experiments using MAO as
the cocatalyst were conducted for 30 min. In these polymeriza-
tions, 500 equiv of MAO was injected into the reactor and the
solution was stirred for 5 min prior to injecting a toluene
solution of the titanium dichloride precatalyst. Polymerization
experiments using B(C6F5)3 as the cocatalyst were conducted
for 10 min. In these experiments, 20 equiv of triisobutylalu-
minum hydride (TiBAl) was injected into the reactor and the
solution was stirred for 5 min prior to injecting toluene
solutions of the titanium dimethyl precatalyst and the B(C6F5)3
cocatalyst.
Com p u ta tion a l Ch em istr y. Molecular mechanics calcula-
tions were conducted using the MM3 force field as imple-
mented in CAChe 6.0 for windows developed by CAChe Group,
Fujitsu. All DFT calculations were performed using the
Gaussian 98 suite of programs.83 Optimized gas phase geom-
(82) Cromer, D. T.; Mann, J . B. Acta Crystallogr. A 1968, A24, 321-
324.
(83) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.;
Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.; Daniels,
A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .; Barone, V.;
Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford,
S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma,
K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J . B.;
Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Keith,
T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.;
Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen, W.; Wong, M.
W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.; Pople, J . A.
Gaussian 98 (Revision A.7); Gaussian, Inc.: Pittsburgh, PA, 1998.

Catalyst Design
Organometallics, Vol. 23, No. 22, 2004 5245
Ta ble 2. Ca lcu la ted En er gies for [Cp TiR′(NP R3)]+
(R′ ) Me, P r )a
b
c
d
e
f
g
h
R
Eipf
Eips EC-Me
Eib-Me Eips-Pr EC-Pr
Eib-Pr
Me
-23.7 41.1 -17.4
6.4
6.6
5.9
6.8
5.1
4.8
29.4
32.1
31.4
76.2
37.3
34.6
-8.6
2.8
3.4
2.8
3.2
0.4
0.3
NH2 -22.9 40.6 -17.7
-9.8
-10.1
-11.0
-9.9
Hi
-19.1 44.6 -18.9
-14.3 89.1 -21.4
-5.9 48.4 -20.4
-5.6 46.1 -20.8
Cl
F
-10.0
a
B3LYP/BS2//B3LYP/LANL2DZ COSMO. Solvent toluene (ꢀ )
b
2.379). Ion-pair formation energy, corresponding to 1′ + BCl3 f
2′ + Eipf
.
c Me-cation/anion separation energy, corresponding to
2′ f 3′+ [MeBCl3]- +Eips-Me
.
Me-cation-C2H4 binding, corre-
d
sponding to 3′ + C2H4 f 4′ + EC-Me.
e Insertion barrier (Me-
cation), corresponding to 4′ f TS1 + Eib-Me.
f Pr-cation/anion
separation energy, corresponding to 7′ f 6′ + [MeBCl3]- + Eips-Pr
.
g
Pr-cation-C2H4 binding energy, corresponding to 6′ + C2H4
f
h
8′ + EC-Pr
TS2 + Eib-Pr
LANL2DZ gas phase values.
.
Insertion barrier (Pr-cation), corresponding to 8′ f
Italicized numbers refer to B3LYP/BS2//B3LYP/
i
.
Ta ble 3. Ca lcu la ted Atom ic Ch a r ges on Tia
R
CpTiMe2NPR3
[CpTiMeNPR3]+
Me
NH2
H
Cl
F
0.78
0.79
0.84
0.94
0.91
0.90
0.91
0.96
1.05
1.05
a
Mulliken Charges from B3LYP/BS2//B3LYP/LANL2DZ. COS-
MO solvent toluene (ꢀ ) 2.379) calculations.
etries were obtained using the Becke3 exchange functional,84
as implemented in Gaussian 98,85 in combination with the Lee,
Yang, and Parr correlation functional,86 i.e., the B3LYP
method. The LANL2DZ basis set was used for geometry
optimizations and energy calculations. Relative energies were
obtained by performing single-point calculations at the B3LYP/
BS2 level of theory, based on the above geometries. The BS2
basis set is a larger basis set consisting of the 6-31G(d,p) basis
set87-90 for H, B, C, N, and F atoms, and the LANL2DZ basis
sets for Ti, Ni, P, and Cl. As previously recommended by
Torrent, Sola`, and Frenking,91 for Ti, the LANL2DZ basis set
was supplemented with a set of (n)p functions for transition
metals developed by Couty and Hall.92 In addition, for P, a d
function with an exponent of 0.34 was added.93 This basis set
is expected to provide more reliable relative energies than the
smaller LANL2DZ basis set. Solvent effects were approximated
with single-point calculations using B3LYP/BS2 on gas phase
B3LYP/LANL2DZ geometries using the COSMO method94
with the dielectric constant of ꢀ ) 2.379 for toluene. Toluene
was chosen since all polymerization experiments reported
herein used toluene as the solvent. The nature of transition
structures was confirmed by calculating the harmonic vibra-
tional frequencies. Relaxed potential energy scans along the
C-C bond distances that served as the reaction coordinates,
followed by full geometry optimizations, were used to confirm
that all transition states were connected backward to the
(84) Becke, A. D. Phys. Rev. A 1988, 38, 3098-3100.
(85) Stephens, P. J .; Devlin, F. J .; Chabalowski, C. F.; Frisch, M. J .
J . Phys. Chem. 1994, 98, 11623-11627.
(86) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condensed Matter
Mater. Phys. 1988, 37, 785-789.
(87) Ditchfield, R.; Hehre, W. J .; Pople, J . A. J . Chem. Phys. 1971,
54, 724-728.
(88) Hehre, W. J .; Ditchfield, R.; Pople, J . A. J . Chem. Phys. 1972,
56, 2257-2261.
(89) Hariharan, P. C.; Pople, J . A. Theor. Chim. Acta 1973, 28, 213-
222.
(90) Hariharan, P. C.; Pople, J . A. Mol. Phys. 1974, 27, 209-214.
(91) Torrent, M.; Sola, M.; Frenking, G. Chem. Rev. 2000, 100, 439-
493.
(92) Couty, M.; Hall, M. B. J . Comput. Chem. 1996, 17, 1359-1370.
(93) Feldgus, S.; Landis, C. R. Organometallics 2001, 20, 2374-2386.
(94) Barone, V.; Cossi, M. J . Phys. Chem. A 1998, 102, 1995-2001.

5246 Organometallics, Vol. 23, No. 22, 2004
Beddie et al.
F igu r e 1. Generalized mechanism and energy profile for two consecutive solution phase (bold) and gas phase (thin line)
insertion reactions.
a barrier of just 5.9 kcal mol-1. Rearrangement to the
reactants and forward to the products. Atomic charges on
titanium reported herein are Mulliken charges from the
â-agostic propyl cation is thermodynamically favored.
Prior studies of the polymerization mechanism for
B3LYP/BS2 COSMO single-point calculations (Table 3). All
relative energies are reported in kcal mol-1, bond lengths in
constrained geometry and metallocene catalysts sug-
angstroms (Å), and angles in degrees (°), unless otherwise
gested such conversions from γ- to â-agostic conforma-
noted.
tions are facile.61,71,96 Re-coordination of the [MeBCl3]-
counterion to form the thermodynamically favored ion
pair [CpTiPr(NPH3)][MeBCl3], 7′ (31.4 kcal mol-1 lower
Resu lts a n d Discu ssion
in energy), represented both the conclusion of the first
P olym er iza tion Mech a n ism . Our strategy to cata-
insertion of ethylene and the starting point for the
lyst design began with a computational examination of
subsequent ethylene insertion step. Consideration of the
the mechanism of the first two insertions of ethylene
solvent effects48,49,51,56-58,69 for toluene using the COS-
polymerization using the model precursor CpTiMe2-
MO94,97 method revealed that the slightly polar solvent
NPH3 and the Lewis acid BCl3. Ziegler and co-workers
generally stabilizes all ions and ionic complexes, as
expected. This solvent correction significantly reduces
the magnitude of the ion pair separation and ethylene
binding energies and increases the magnitude of the ion
pair formation energy for the phosphinimide-complex
reaction sequence (Figure 1, Table 1). Similar trends
have been previously observed.57,59
have previously described similar computations for the
first insertion of ethylene for CpTiMe2(NPR3), where R
) H, Me, t-Bu, employing B(C6F5)3 as a Lewis acid.58,59
The mechanism begins with the formation of an ion pair
[CpTiMe(NPH3)][MeBCl3], 2′, lying 19.1 kcal mol-1
lower in energy. The transition state structure for this
reaction was not located due to the rather soft potential
surface; nonetheless, experimental data has confirmed
the formation and identity of the zwitterionic species
[CpTiMe(NPt-Bu3)][MeB(C6F5)l3].41 In addition, the ex-
perimental barrier for formation of the related ion pair
[(1,2-Me2Cp)2ZrMe][MeB(C6F5)3] was estimated by Marks
Calculations for a second ethylene insertion followed
a similar reaction coordinate with lower solvent phase
and gas phase energy profiles (Figure 1). These results
are also consistent with computational studies of eth-
ylene insertion barrier into [Cp2ZrR]+,67 [[H2Si(C5H4)-
(t-BuN)]TiR]+ ,57 and [(CpSiH2NH)TiR]+ 71 (R ) Me, Pr).
As observed for the first insertion, the largest barrier
to the second insertion occurs for the ion pair dissocia-
tion. Subsequent reaction of the separated â-agostic
propyl cation [CpTiPr(NPH3)]+, 6′, with C2H4 proceeds
to give the ethylene π-complex, which is 10.1 kcal mol-1
lower in energy. In the final step, ethylene insertion into
the Ti-Pr bond requires only 2.7 kcal mol-1 of energy
to give a γ-agostic pentyl cation, which is an additional
14.1 kcal mol-1 lower in energy. Consistent with these
computations, Landis et al. have previously shown that
the first olefin insertion can be ca. 400 times slower than
subsequent insertions for metallocene catalysts.98
and co-workers to be approximately 3 kcal mol-1 95
.
Separation of the ion pair [CpTiMe(NPH3)][MeBCl3], 2′,
to discrete ions was calculated to require 44.6 kcal mol-1
of energy in the solution phase. Subsequent steps
considered separated cations, although studies of coun-
terion effects by Marks et al.57 have inferred that this
approach overestimates energy barriers. The initial
interaction of the cation with ethylene results in the
formation of an energetically favorable ethylene π-com-
plex, 18.9 kcal mol-1 lower in energy, which agrees well
with the complexation energy previously reported by
Ziegler and co-workers (-20.1 kcal mol-1, gas phase).59
Insertion of ethylene into the Ti-Me bond proceeds via
a metallacycle transition state and results in the forma-
tion of a propyl cation with a γ-agostic interaction with
(96) Lohrenz, J . C.; Woo, T. K.; T., Z. J . Am. Chem. Soc. 1995, 117,
12793-12800.
(95) Deck, P. A.; Beswick, C. L.; Marks, T. J . J . Am. Chem. Soc.
1998, 120, 1772-1784.
(97) Andzelm, J .; Kolmel, C.; Klamt, A. J . Chem. Phys. 1995, 103,
9312-9320.

Catalyst Design
Organometallics, Vol. 23, No. 22, 2004 5247
Liga n d Effects. Analogous calculations employing
the model systems based on CpTiMe2(NPR3), 1′, (R )
Me, NH2, Cl, F), revealed that the atomic charges for
the precursors and the cations [CpTiMe(NPR3)]+, 3′ (R
) Me, NH2, H, Cl, F), were consistent with the order of
the electron-donating ability of the phosphinimides: Me
= NH2 > H > Cl = F. It is clear that, relative to H,
electron-donating substituents (Me, NH2) on the phos-
phinimide ligand produce lower relative energy profiles
for the first and second ethylene insertions, while
electron-withdrawing substituents afford higher energy
profiles (Table 1). While the electronic properties of the
phosphinimide ligand influence the ethylene complex-
ation energies (Ec-Me, Ec-Pr) and the ethylene insertion
barriers (Eib-Me, Eib-Pr) to a limited extent, the effects
are small, with differences of only approximately 3 kcal
mol-1 at most (Table 2).
In contrast, the magnitude of the ion pair formation
energies follows the trend Me > NH2 > H > Cl > F,
with approximately an 18 kcal mol-1 difference between
the most electron-rich titanium complex, CpTiMe2-
(NPMe3), and the most electron-poor titanium complex,
CpTiMe2(NPF3) (Table 2). In related experimental work,
Marks and co-workers showed that in reactions of the
Lewis acids B(C6F5)2Ar (Ar ) C6F5, 3,5-F2C6H3, Ph, 3,5-
Me2C6H3) with (1,2-Me2Cp)2ZrMe2 the ion pair forma-
tion energies and, thus, the extent to which the ion pair
formed in equlibria, were qualitatively related to the
polymerization activities.95 In the present systems, if
entropy contributions are assumed to be similar to those
described by Marks et al., then electron-withdrawing
substituents on the phosphinimide ligand would be
expected to reduce the ability of cocatalysts to effect
methyl abstraction, and thus reduce polymerization
activity.
The ion pair separation energies for the methyl
cations are also strongly affected by the electronic
properties of the phosphinimide ligand, as electron-
donating substituents result in reduced barriers to ion
pair separation (Table 2). Such ligands should result in
faster ethylene insertion. This view is in general agree-
ment with the previous suggestions for other catalyst
systems in which incorporation of electron-donating
groups resulted in enhanced polymerization activi-
ties.42,99-102
F igu r e 2. J IMP depictions109 of propyl cation ion pairs
for (a) R ) NH2 and (b) R ) Me (H ) white, B ) pink, C )
gray, N ) blue, P ) purple, Cl ) green, Ti ) yellow).
[CpTiPr(NPF3)][MeBCl3] was constrained to 179.99°.
Although this was also unsuccessful for the determina-
tion of a fully optimized structure, it did provide
geometries in which the ion pair was bridged through
the methyl group. Thus, for the subsequent B3LYP/BS2
COSMO calculations, the lowest energy geometry for
the constrained model was employed. For the Cl-
substituted model, [CpTiPr(NPCl3)][MeBCl3], a separate
problem was encountered, as a Cl atom from the ligand
and the anion eliminated to form free Cl2. Constraint
of the B-Cl and P-Cl bond lengths resolved this
problem and afforded a fully optimized geometry. In
both cases, the use of constraints resulted in an under-
estimation of the ion pair separation. Nonetheless, it is
clear that the electronic properties of the phosphinimide
ligands result in a significant influence on the ion pair
separation energies for the propyl cations (Table 2). It
is interesting to note that for the propyl cation with R
The geometries of the propyl ion pairs 7′ (R ) Me,
NH2, H) were minimized by introducing the [MeBCl3]-
counterion to the â-agostic propyl cation; however, for
the F analogue it was not possible to ascertain an
optimized geometry using this method. Instead, a Cl
atom transfer from [MeBCl3]- to titanium resulted in
the formation of CpTiPrCl(NPF3) and MeBCl2. A similar
problem was reported by Lanza and co-workers during
their investigations on constrained geometry catalysts.49
To prevent the chloride atoms of [MeBCl3]- from ap-
proaching the titanium, the Ti-C-B bond angle in
) NH2 the ion pair separation energy is 32.1 kcal mol-1
,
which is slightly more than the 29.4 kcal mol-1 for the
R ) Me analogue. This stands in contrast to that
observed for the ion pair separation for the methyl
cations, where the reverse is the case (R ) NH2: 40.6
kcal mol-1; R ) Me: 41.1 kcal mol-1). This reversal may
be due to the stronger binding of the anion in the propyl
cation for R ) NH2 compared to the R ) Me analogue,
although N-H‚‚‚Cl (2.39 Å) hydrogen bonding between
the tris(amino)phosphinimide ligand and the anion may
contribute to the stabilization energy of this ion pair
(Figure 2).
(98) Landis, C. R.; Rosaaen, K. A.; Sillars, D. R. J . Am. Chem. Soc.
2003, 125, 1710-1711.
The electronic properties of the phosphinimide ligand
have little effect on the ethylene complexation energy
for the propyl cations. On the other hand, complexes
with electron-withdrawing groups Cl and F have inser-
tion barriers that are 2-3 kcal mol-1 lower than for the
corresponding electron-donating groups, Me (2.8 kcal
mol-1) and NH2 (3.4 kcal mol-1). This may be the result
(99) Piccolrovazzi, N.; Pino, P.; Consiglio, G.; Sironi, A.; Moret, M.
Organometallics 1990, 9, 3098-3105.
(100) Lee, I.-M.; Gauthier, W. J .; Ball, J . M.; Iyengar, B.; Collins,
S. Organometallics 1992, 11, 2115-2122.
(101) Mohring, P. C.; Coville, N. J . J . Organomet. Chem. 1994, 479,
1-29.
(102) Klosin, J .; Kruper, W. J ., J r.; Nickias, P. N.; Roof, G. R.; De
Waele, P.; Abboud, K. A. Organometallics 2001, 20, 2663-2665.

5248 Organometallics, Vol. 23, No. 22, 2004
Beddie et al.
F igu r e 4. ORTEP drawings of (a) 6 and (b) 7; 30% thermal
ellipsoids are shown. Hydrogen atoms are omitted for
clarity. 6: P-Nav 1.725(4) Å, N-P-Nav 102.2(2)°. 7: P-N
1.709(3) Å, N-P-N′ 100.60(13)°.
F igu r e 3. J IMP depictions109 of propyl cation ethylene
π-complexes for (a) R ) H and (b) R ) F (H ) white, C )
gray, N ) blue, P ) purple, F ) green, Ti ) yellow).
wide range of phosphinimide ligands are readily acces-
sible from inexpensive starting materials. Thus, the tris-
(amino)phosphines P(NMe2)3, 1, P(NEt2)3, 2, and P(N(n-
Pr)2)3, 3, were obtained from commercial sources, while
several other derivatives including P(N(n-Bu)2)3, 4,74
P(Ni-PrMe)3, 5,80 and P(NEtPh)3, 6,81 were readily
synthesized in moderate to good yields. In addition, P(N-
3-methylindolyl)3, 7,75 was prepared by generation of the
lithium salt of 3-methylindole, followed by subsequent
reaction with PCl3. These species were spectroscopically
characterized, and in the case of 6 and 7, X-ray crystal-
lographic studies confirmed the formulations. While the
metric parameters were unexceptional, the solid state
structures did reveal that the phenyl substituents on
the nitrogen atoms of 6 and the indolyl arene rings of 7
were oriented in the same direction, i.e., canted up
toward the “lone pair” on phosphorus (Figure 4).
Subsequent oxidation of these phosphines proceeds,
in general, via standard methods.103-105 Reaction of 1-6
with Me3SiN3 in refluxing toluene yielded the deriva-
tives Me3SiNP(NR2)3 (R ) Me 8, Et 9, Pr 10, Bu 11; R2
) i-PrMe 12, EtPh 13) as colorless oils in yields ranging
from 68 to 95% (Scheme 2). These species were identi-
fied by multinuclear NMR spectroscopy. In contrast,
attempts to synthesize Me3SiNP(N-3-methylindolyl)3,
14, via treatment of 7 with Me3SiN3 while heating at
refluxing temperature for an extended period of time
(2 weeks), failed to effect oxidization of this phosphine,
suggesting that steric congestion and the rigidity of the
substituents precluded reaction. Experimental and theo-
retical data have previously supported the view that
steric factors can alter the course of such Staudinger
reactions.106,107 Furthermore, P(N-3-methylindolyl)3, 7,
previously displayed slow reactivity in other oxidation
of â-agostic interactions that are present in the ethylene
π-complexes for R ) Cl or F, but absent for R ) Me,
NH2, or H. Such â-agostic interactions may destabilize
the ethylene π-complexes, thus decreasing the gap in
energy between the π-complex and the transition state,
and consequently lower the insertion barrier. In addi-
tion, the ethylene is bound in a conformation parallel
to the Ti-N bond in the complexes with R ) Me, NH2,
or H, while it is perpendicular to the Ti-N bond for R
) Cl or F. The latter conformation is similar to the
geometry of the transition state for ethylene insertion,
thus facilitating insertion (Figure 3).
The present computational data suggest that increas-
ing the electron-donating capability of the phosphin-
imide ligand reduces interactions between the titanium
cation and anion, thus facilitiating ethylene coordination
and increasing the rate of polymer progation. At the
same time, electron-donating substituents increase the
barrier from the ethylene π-complexes to the insertion
transition state. However, the latter barrier is signifi-
cantly smaller than the ion pair separation energy, and
thus, in these catalyst systems, electron-donating groups
enhance catalyst activity.
Syn th etic Ch em istr y. The above computations sug-
gest that amino and alkyl substituents on the phos-
phinimide ligand should facilitate ion separation and,
thus, polymerization catalysis to a similar extent. Our
previous experimental work has shown that while high-
activity olefin polymerization catalysts can be derived
from trialkylphosphinimide ligand complexes, the range
of ligands that offer high activity is limited to those that
are sterically demanding. From a synthetic perspective,
the tris(amino)phosphine ligand precursors are much
easier to prepare in high yield than sterically similar
trialkylphosphines. In addition, since tris(amino)phos-
phines are derived from secondary amines and PCl3, a
(103) Staudinger, H.; Meyer, J . Helv. Chim. Acta 1919, 2, 635-646.
(104) Gololobov, Y. G.; Zhmurova, I. N.; Kasukhin, L. F. Tetrahedron
1981, 37, 437-472.
(105) Gololobov, Y. G.; Kasukhin, L. Tetrahedron 1992, 48, 1353-
1406.

Catalyst Design
Organometallics, Vol. 23, No. 22, 2004 5249
Ta ble 5. Selected Metr ic P a r a m eter s fr om X-r a y Str u ctu r es
Ti-Cl (Å) Ti-N (Å) Ti-P (Å)
2.298(2), 2.323(2) 1.759(3)
compd
N
Ti-N-P (deg)
15
19
21
22
24
25
1.594(3)
164.3(2)
174.3(2) 173.1(3)
163.8(2)
175.7(2)
175.5(2)
2.311(2), 2.309(2) 2.290(2), 2.298(2)
2.323(2), 2.310(1)
1.761(3), 1.769(4)
1.788(3)
1.593(3) 1.592(4)
1.589(3)
2.317(3), 2.337(3)
1.786(4)
1.586(4)
2.303(1), 2.313(2)
1.782(3)
1.592(3)
2.311(1), 2.317(1)
1.778(2)
1.591(2)
170.3(2)
Sch em e 2. Gen er a lized Syn th etic Rou te to
P r eca ta lysts
reactions, which was attributed to the poor electron-
donating ability of phosphine resulting from the aro-
maticity N-3-methylindolyl substituents.75
The series of titanium complexes Cp′TiCl2(NP(NR2)3)
(Cp′ ) Cp, R ) Me 15, Et 16, Pr 17, Bu 18, R2 ) i-PrMe
19, EtPh 20; Cp′ ) Cp*, R ) Me 21, Et 22, Pr 23, Bu
24, R2 ) i-PrMe 25, EtPh 26) were prepared by reaction
of Cp′TiCl3 and the corresponding phosphinimine Me3-
SiNP(NR2)3. Generally the yields were good, ranging
from 68 to 95%. Subsequent alkylation by MeLi or
MeMgX (X ) Cl, Br) afforded the corresponding Ti-
dimethyl derivatives 27-38, also in high yields of ca.
90% (Scheme 2). All complexes were characterized by
elemental analysis and multinuclear NMR spectroscopy.
In addition, X-ray crystallographic studies of 19, 21, 22,
24, and 25 are reported herein. It is worth mentioning
that the structure of 15 has been previously reported.108
Representative ORTEP depictions of these molecules
are presented in Figure 5. The features of these com-
pounds are typical of Ti-phosphinimide complexes (Table
5), with Ti-N distances of 1.759(3)-1.769(4) Å for the
Cp derivatives and slightly longer (1.778(3)-1.788(3) Å)
for the Cp* analogues. These variations are consistent
with the greater electron-donating ability of the Cp*
ligand. The Ti-N-P angles approach linearity, ranging
from 163.8(2)° to 175.7(2)°.
F igu r e 5. ORTEP drawings of (a) 19 and (b) 25; 30%
thermal ellipsoids are shown. Hydrogen atoms are omitted
for clarity.
Ta ble 6. P olym er iza tion Testin g of P r eca ta lysts
Usin g MAO a s a Coca ta lysta
[cat]
activity
(µmol PE yield (g mmol-1
precatalyst L-1
)
(g)
h-1 atm-1
)
Mn
Mw
PDI
15
16
19
20
21
22
25
26
100
100
50
0.15
1.57
0.40
4.62
1.26
2.33
1.66
6.08
2.4
26
13
150
21
39
257 100 533 600 2.08
129 900 464 000 3.57
39 100 348 400 8.91
135 500 624 800 4.61
82 600 141 800 1.72
90 100 149 100 1.65
127 800 352 200 2.76
126 100 507 200 4.02
50
100
100
50
56
200
50
a
Polymerization conditions: C2H4 P ) 2 atm, T ) 30 °C, solvent
) 600 mL of toluene, stir rate ) 1000 rpm, 500 equiv of MAO,
reaction time ) 30 min. Results of at least duplicate runs.
Eth ylen e P olym er iza tion . The complexes 15, 16,
19-22, 25, and 26 were tested for ethylene polymeri-
zation activity using 500 equiv of MAO as a cocatalyst.
The polymerization experiments were conducted for 30
min, under an ethylene pressure of 2 atm, using catalyst
concentrations of either 50 or 100 µmol L-1. Toluene
(600 mL) was used as the solvent, with a stir-rate of
1000 rpm. While 15, 16, 19, 21, 22, and 25 resulted in
low catalytic activity upon activation with MAO, the
more sterically demanding phosphinimide ligand com-
plexes afforded significantly higher activities (Table 6).
For example, 20 and 26 exhibited higher catalyst
activities of 150 and 200 g mmol-1 h-1 atm-1, respec-
tively. These data are in agreement with the observa-
tions of related titanium trialkylphosphinimide com-
plexes and suggest that enhanced steric bulk of the
ligand has a positive effect on catalyst activity, possibly
by slowing the rate of catalyst deactivation via attack
at the phosphinimide-N atom.37-39 The Mn values of the
polyethylene derived under these conditions range from
39 to 257 000 (Table 6). The polydisperities range from
1.6 to 8.9 for these systems, suggesting that some side
(106) Widauer, C.; Grutzmacher, H.; Shevchenko, I.; Gramlich, V.
Eur. J . Inorg. Chem. 1999, 1659-1664.
(107) Alajarin, M.; Conesa, C.; Rzepa, H. S. J . Chem. Soc., Perkin
Trans. 1999, 1811-1814.
(108) Von Haken Spence, R. E.; Stephan, D. W.; Brown, S. J .;
J eremic, D.; Wang, Q. (Nova Chemicals (International) S.A., Switz.)
PCT Int. Appl.; WO, 2000, 38 pp.

5250 Organometallics, Vol. 23, No. 22, 2004
Ta ble 7. P olym er iza tion Testin g Usin g B(C6F 5)3 a s a Coca ta lysta
[cat] ) 10 [cat] ) 4
PE yield (g) PE yield (g)
Beddie et al.
precatalyst
activity
activity
Mn
Mw
PDI
27
28
29
30
31
32
33
34
35
36
37
38
350
1800
0.70
3.53
2200
3500
5500
3600
3600
4200
4200
4700
10000
6100
4900
4200
5200
5600
16000
1.73
2.81
4.37
2.88
2.85
3.37
3.37
3.79
8.33
4.88
3.94
3.39
4.17
4.50
12.61
315 000
394 000
646 200
752 900
2.05
1.91
1900
2600
1200
2000
3.79
5.30
2.44
4.06
388 600
432 500
140 800
insoluble
717 800
829 100
692 200
1.85
1.92
4.92
2100
2300
1600
2900
3500
4.25
4.54
3.24
5.80
6.93
288 100
324 600
493 400
437 800
175 000
617 900
660 100
1 012 000
786 300
331 100
2.14
2.03
2.05
1.80
1.89
Cp*TiMe2 NPi-Pr3
CpTiMe2 NPt-Bu3
Cp2ZrMe2
a
Polymerization conditions: ethylene pressure ) 2 atm, temperature ) 30 °C, Solvent ) 600 mL of toluene, stir rate ) 1000 rpm, 2
equiv of B(C6F5)3; 20 equiv of TiBAL, reaction time ) 10 min. Results of at least duplicate runs. Activities are reported in the units g
mmol-1 h-1 atm-1. [Cat] units: µmol/L.
reactions may provide access to variation in the active
species. Similar observations have been previously
reported for MAO activation of phosphinimide-based
catalysts.40-44
The analogous methylated precatalysts 27, 28, 31-
34, 37, and 38, in concentrations of 10 µmol/L, were
tested for ethylene polymerization activity using 2 equiv
of B(C6F5)3 as a cocatalyst and 20 equiv of TiBAl as the
solvent purifier. These experiments were conducted for
10 min, under an ethylene pressure of 2 atm, using 600
mL of toluene as the solvent, and with a stir-rate of 1000
rpm. In contrast to the MAO activation strategy de-
scribed above, all of these systems resulted in very high
ethylene polymerization activity (1200-2600 g mmol-1
h-1 atm-1), although 27 exhibited polymerization activ-
ity that was somewhat lower (350 g mmol-1 h-1 atm-1
)
F igu r e 6. Plot of phosphinimide cone angle (deg) versus
polymerization activity (g mmol-1 h-1 atm-1) for the
catalysts derived from CpTi Me2(NPR3) (b) and Cp*TiMe2-
(NPR3) (9).
(Table 7). Among these catalysts, there appears to be a
good correlation between the steric bulk of the ligands
and the activity. For example, except for the analogues
31 and 38, replacement of Cp with Cp* results in higher
activity. Similarly, increased steric demands on the
amino groups of the tris(amino)phosphinimide ligands
also result in higher activity. The observation that
increased electron-donating ligands (Cp* vs Cp) and
increased steric bulk of the phosphinimide ligands yield
higher catalyst activities is consistent with the domi-
nant role for ion-pair separation energies in the polym-
erization energy profile.58 Reduction of the catalyst
concentrations to 4 µmol/L with conditions otherwise
unchanged resulted in a further dramatic increase in
the polymerization activities to the range 2200-10 000
g mmol-1 h-1 atm-1 (Table 7). It should be noted that
further reduction of the concentration of 38 to 2 µmol/L
did not result in a further increase in polymerization
activity, indicating that at 4 µmol/L diffusion-limited
kinetics are no longer an issue. It is particularly
noteworthy that among this family of highly active
catalysts the precursor 36 exhibits activity that is
almost twice as high as that achieved for the previously
reported precatalysts derived from the trialkylphos-
phinimide complexes, CpTiMe2NP(t-Bu)3 and Cp*TiMe2-
NP(t-Bu)3.40,42 The Mn of the resulting polymers range
from 140 000 to 433 000 (Table 7), while the polydis-
persities, in general less than 2.0, are consistent with
single-site polymerization catalyst behavior.
Ster ic Effects. The steric bulk of the phosphinimide
ligands can be expressed in terms of a “cone angle”.110
These angles, calculated by doubling the average of the
three half-angles defined at titanium by the phosphorus
atom and the outermost atom of the three substituents,
were obtained both from crystallographic data and
optimized models derived from molecular mechanics
calculations (MM3) (Table 8). In the latter determina-
tions, the Ti-N and P-N bond distances were fixed at
values derived from X-ray data. In general, the cone
angles computed in either fashion showed generally
good agreement, with the exception of the values
computed for NP(Nn-Bu2)3 (X-ray: 117.5°; MM3: 146.3°).
This discrepancy was attributed to the flexibility of the
long butyl chains and indicates that cone angles should
be used only as a rough estimate of the steric size.
Correlations of these cone angles with the polymeriza-
tion activities showed qualitative relationships for both
the Cp and Cp* series of tris(dialkylamino)phosphin-
imide-based catalysts (Figure 6). In the case of the tris-
(ethylphenylamino)phosphinimide ligand complexes,
(109) Manson, J .; Webster, C. E.; Hall, M. B. J IMP Development,
Version 0.1 ed.; Texas A&M University: College Station, TX 77842.
(110) Tolman, C. A. Chem. Rev. 1977, 77, 313.

Catalyst Design
Organometallics, Vol. 23, No. 22, 2004 5251
Ta ble 8. P h osp h in im id e Con e An gles for
Cp *TiCl2(NP R3) Com p lexes
although the cone angle is calculated to be 112.5°, the
activity of the derived catalyst is lower than those
derived from Cp*TiMe2NP(NEt2)3 and Cp*TiMe2NP-
(Ni-PrMe)3. This is attributed to the diminished electron-
donating ability as a result of the presence of the aryl
ring on nitrogen. In addition, this may be a reflection
of the inherent error in the description of the steric bulk
via an estimated cone angle.
NPR3
R
cone angle (deg)
X-raya
MM3
R ) NMe2
R ) NEt2
77.5
94.7
115.3
117.5
93.5
112.5
87.2
75.7
77.2
100.9
129.8b
146.3
101.3
R ) Nn-Pr2
R ) Nn-Bu2
R ) Ni-PrMe
R ) NEtPh
R ) t-Bu
Con clu sion s
84.2
81.3
R ) i-Pr
The DFT calculations of the mechanism of polymer-
ization for the series of catalyst models derived from
CpTiMe2NPR3 (R ) Me, NH2, H, Cl, F) demonstrated
the critical role of ion pairing in the determination of
the overall barrier to polymerization, in agreement with
other theoretical and experimental observations. The
present study further suggests that ligand modifications
that incorporate electron-donating substituents reduce
this barrier and thus enhance polymerization activity.
Based on these calculations and synthetic limitations,
strategies to the readily accessible and easily varied
family of compounds of general formula Cp′TiX2NP-
(NR2)3 were developed. Activation of the dichloride
derivatives with MAO revealed widely variable polym-
erization activities, an observation previously seen for
phosphinimide-based catalysts. However, activation of
the dimethyl analogues using B-based activators re-
sulted in very high ethylene polymerization activities.
In addition, for electronically similar tris(amino)phos-
phinimide ligands, polymerization activity was seen to
increase with increasing steric bulk. Thus, in general,
optimization of steric bulk and electronic characteristics
to facilitate ion-pair separation and prolonged catalyst
lifetime has been achieved, affording a readily accessible
and easily varied family of highly active ethylene
a
These values were derived from the X-ray reported herein.
Estimated from the X-ray data for 24.
b
polymerization catalysts based on the tris(amino)phos-
phinimide ligand. Furthermore, this study effectively
demonstrates the synergy of theoretical and experimen-
tal methods for the catalyst design. Further develop-
ments of this new class of single-site catalysts are the
subject of ongoing efforts.
Ack n ow led gm en t. Financial support from NSERC
of Canada, NOVA Chemicals Corporation, and the
ORDCF is gratefully acknowledged. C.B. is grateful for
the award of NSERC-IPGS and CCMR research schol-
arships. E.H. is grateful for the award of an OGS
scholarship.
Su p p or tin g In for m a tion Ava ila ble: Tables of calculated
gas phase relative energies (kcal/mol) for the mechanism; gas
phase ion-pair formation, ion-pair separation, and ethylene
binding energies and insertion barriers for the methyl and
propyl cations derived from CpTiMe2NPR3; and crystallo-
graphic information files (CIF). This material is available free
of charge via the Internet at http://pubs.acs.org.
OM049545I