Half-sandwich group 4 salicyloxazoline catalysts

Article abstract of DOI:10.1021/om060687h

A new class of zirconium and hafnium half-sandwich complexes bearing Cp* and salicyloxazolinato (L) ligands has been prepared by salt elimination and protonolysis routes. The analogous Cp and indenyl compounds are generally inaccessible, as are the titanium compounds. The molecular structures of four examples [Cp*MLX₂] (variously M = Zr, Hf; X = Cl, Me) reveal chiral-at-metal structures which persist in solution, according to variable-temperature NMR studies; ΔG?₂₉₈ for the racemization process was found to be ca. 75 kJ mol^-1. Treatment of these compounds with MAO, [Ph₃C][B(C₆F₅) ₄], or [PhNMe₂H]-[B(C₆F₅) ₄] leads to catalysts for alkene polymerization, the nature of which depends on the cocatalyst chosen. The anilinium salt smoothly produces a single chiral species, [Cp*ZrLMe][B(C₆F₅)₄], detected also by ¹H NMR spectroscopy, which is a highly active single-site catalyst for polymerization of ethene (and less active for copolymerization of ethene/hexene). The trityl activator produces the same catalyst and at least one other catalytically competent species, as evidenced by NMR spectroscopy and polymer modality. The use of MAO leads to a less well-defined catalyst system. The steric demand of the salicyloxazoline ligand affects the catalyst performance significantly, and computational studies show that access of ethene to either of two inequivalent coordination sites is restricted. This stability of the species [Cp*ZrLMe]⁺ with respect to addition appears to be the limiting factor for catalytic activity. Catalyst stability is addressed, and the steric and electronic factors affecting this are consistent with a mechanism of catalyst death by salicyloxazoline ligand loss.

Full text of DOI:10.1021/om060687h

Organometallics 2006, 25, 6019-6029
6019
Half-Sandwich Group 4 Salicyloxazoline Catalysts
Stuart R. Coles,† Guy J. Clarkson,† Andrew L. Gott,† Ian J. Munslow,†
Stefan K. Spitzmesser,‡ and Peter Scott*,†
Department of Chemistry, UniVersity of Warwick, CoVentry CV4 7AL, U.K., and
INEOS Technologies, Rue de Ransbeek 310, 1120 Brussels, Belgium
ReceiVed July 31, 2006
A new class of zirconium and hafnium half-sandwich complexes bearing Cp* and salicyloxazolinato
(L) ligands has been prepared by salt elimination and protonolysis routes. The analogous Cp and indenyl
compounds are generally inaccessible, as are the titanium compounds. The molecular structures of four
examples [Cp*MLX2] (variously M ) Zr, Hf; X ) Cl, Me) reveal chiral-at-metal structures which persist
in solution, according to variable-temperature NMR studies; ∆Gq for the racemization process was
298
found to be ca. 75 kJ mol-1. Treatment of these compounds with MAO, [Ph3C][B(C6F5)4], or [PhNMe2H]-
[B(C6F5)4] leads to catalysts for alkene polymerization, the nature of which depends on the cocatalyst
chosen. The anilinium salt smoothly produces a single chiral species, [Cp*ZrLMe][B(C6F5)4], detected
also by 1H NMR spectroscopy, which is a highly active single-site catalyst for polymerization of ethene
(and less active for copolymerization of ethene/hexene). The trityl activator produces the same catalyst
and at least one other catalytically competent species, as evidenced by NMR spectroscopy and polymer
modality. The use of MAO leads to a less well-defined catalyst system. The steric demand of the
salicyloxazoline ligand affects the catalyst performance significantly, and computational studies show
that access of ethene to either of two inequivalent coordination sites is restricted. This stability of the
species [Cp*ZrLMe]+ with respect to addition appears to be the limiting factor for catalytic activity.
Catalyst stability is addressed, and the steric and electronic factors affecting this are consistent with a
mechanism of catalyst death by salicyloxazoline ligand loss.
Introduction
The mechanism of alkene polymerization by group 4 met-
allocene catalysts of the type Cp2MX2 is now quite well
understood.1 In conjunction with cocatalysts such as methyl-
aluminoxane (MAO),2 the species [CpMR]+ are formed which
contain a vacant coordination site cis to the growing polymer
chain R, where the monomer alkene may bind as a precursor
to the 1,2-migratory insertion (propagation) process. Alternative
supporting ligand sets to the various cyclopentadienyls have
been studied intensively.3 Of particular relevance here are the
complexes [(PI)2MX2] (PI ) phenoxy-imine (I)) developed by
Fujita4 and Coates.5 A number of hybrid precatalysts containing
* To whom correspondence should be addressed. E-mail: peter.scott@
warwick.ac.uk. Fax: 024 7657 2710.
† University of Warwick.
one cyclopentadienyl ring and one monoanionic bidentate ligand
have also been developed, including the amidinates, e.g., II,6
and pyridyl-alkoxides, e.g., III.7 Recently, Bochmann explored
a range of phenoxy-imine, e.g., [CpM(PI)Cl2] (IV) and related
catalysts, noting that these inherently chiral systems may allow
stereoselective polymerization of R-olefins.8 Unfortunately,
however, multimodal polymers were produced, indicating the
presence of several catalytically active species, an issue which
‡ INEOS Technologies.
(1) (a) Jordan, R. F. AdV. Organomet. Chem. 1991, 32, 325. (b)
Brintzinger, H. H.; Fischer, D.; Mulhaupt, R.; Rieger, B.; Waymouth, R.
Angew. Chem., Int. Ed. 1995, 34, 1143. (c) Bochmann, M. J. Chem. Soc.,
Dalton Trans. 1996, 255. (d) Chen, E. Y. X.; Marks, T. J. Chem. ReV.
2000, 100, 1391. (e) Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. Chem.
ReV. 2000, 100, 1253. (f) Sillars, D. R.; Landis, C. R. J. Am. Chem. Soc.
2003, 125, 9894. (g) Liu, Z.; Somsook, E.; White, C. B.; Rosaaen, K. A.;
Landis, C. R. J. Am. Chem. Soc. 2001, 123, 11193.
(2) (a) Bryant, P. L.; Harwell, C. R.; Mrse, A. A.; Emery, E. F.; Gan,
Z.; Caldwell, T.; Reyes, A. P.; Kuhns, P.; Hoyt, D. W.; Simeral, L. S.;
Hall, R. W.; Butler, L. G. J. Am. Chem. Soc. 2001, 123, 12009. (b) Zurek,
E.; Ziegler, T. Prog. Polym. Sci. 2004, 29, 107. (c) Zakharov, V. A.; Talzi,
E. P.; Zakharov, I. I.; Babushkin, D. E.; Semikolenova, N. V. Kinet. Catal.
1999, 40, 836.
(4) (a) Furuyama, R.; Saito, J.; Ishii, S.; Makio, H.; Mitani, M.; Tanaka,
H.; Fujita, T. J. Organomet. Chem. 2005, 690, 4398. (b) Makio, H.; Fujita,
T. Bull. Chem. Soc. Jpn. 2005, 78, 52. (c) Mitani, M.; Saito, J.; Ishii, S.-I.;
Nakayama, Y.; Makio, H.; Matsukawa, N.; Matsui, S.; Mohri, J.-I.;
Furuyama, R.; Terao, H.; Bando, H.; Tanaka, H.; Fujita, T. Chem. Rec.
2004, 4, 137. (d) Suzuki, Y.; Terao, H.; Fujita, T. Bull. Chem. Soc. Jpn.
2003, 76, 1493. (e) Makio, H.; Kashiwa, N.; Fujita, T. AdV. Synth. Catal.
2002, 344, 477.
(3) (a) Gibson, V. C.; Spitzmesser, S. K. Chem. ReV. 2003, 103, 283.
(b) Britovsek, G. J. P.; Gibson, V. C.; Wass, D. F. Angew. Chem., Int. Ed.
1999, 38, 428. (c) Ittel, S. D.; Johnson, L. K.; Brookhart, M. Chem. ReV.
2000, 100, 1169.
10.1021/om060687h CCC: $33.50 © 2006 American Chemical Society
Publication on Web 11/16/2006

6020 Organometallics, Vol. 25, No. 26, 2006
Coles et al.
Scheme 1
is rather widespread in olefin polymerization by nonmetal-
locenes. We reasoned that the use of salicyloxazoline ligands
V9 might provide more stable catalysts, but found that the
complexes [CpM(V)X2] were generally inaccessible (vide infra).
We report here that, rather counterintuitively, the bulkier
complexes based on pentamethylcyclopentadienyl can be syn-
thesized readily. Under the appropriate conditions, these com-
plexes provide single-site polymerization catalysts. Stability and
mechanistic studies are also reported.
Results and Discussion
Synthesis of Proligands and Organometallic Complexes.
A range of salicyloxazolines HLn (n ) 1-4) was synthesized
from the corresponding salicylic acids and amino alcohols, as
previously described,9,10 or by similar methods (Scheme 1).
Treatment with excess NaH in THF produced the sodium salts
NaLn‚xTHF in quantitative yields. These were isolated by
filtration and evaporation or used in situ as ligand transfer
reagents.
Subsequent treatment of [Cp*MCl3] (M ) Zr,11 Hf12) with
NaLn‚xTHF produced the complexes [Cp*MLnCl2] (n ) 1-4),
in reasonable to good yields. Analogous reactions of the indenyl
complexes [(η5-C9H7)MCl3] (M ) Ti, Zr, Hf)13 and [CpZrCl3-
(DME)] failed to produce tractable complexes, despite the fact
that the latter is a precursor for the phenoxyimine (PI, vide
supra) complexes [CpZr(PI)Cl2].8 Treatment of [CpTiCl3]14 with
NaL1 gave the complex [CpTiL1Cl2] in good yield, although
the same reaction with the more bulky NaL2 was unsuccessful.
Finally, for comparison with the above, we attempted to
synthesize the hitherto unreported complexes [Cp*M(PI)Cl2] (M
) Zr, Hf) from Cp*MCl3 by various means, but this was
ultimately unsuccessful.
We achieved the syntheses of [Cp*MMe3] (M ) Zr, Hf) by
the methods of Bercaw11 and Marks.15 Direct reactions thereof
with HLn (n ) 1, 2) produced the four complexes [Cp*MLn-
Me2] (Scheme 1). The isolated yields were limited by the
difficulty in crystallizing these rather soluble complexes;
experiments in NMR tubes indicated very high conversions.
Crystallographic Analyses. The molecular structures of
several of the complexes have been determined by X-ray
crystallography. Table 1 gives crystallographic data, while Table
2 compares key structural parameters for all four structures.
Thermal ellipsoid plots are given in Figures 1-3. All of the
complexes have a four-legged piano-stool structure. Unsurpris-
ingly, [Cp*ZrL1Cl2] and [Cp*HfL1Cl2] are isomorphous, having
very similar unit cell and other parameters, and suffer from the
same lattice solvent disorder (vide infra).
We will compare first the three structures of dichloride
complexes (Figures 1 and 2). The M(1)-O(1) distances of ca.
2.02 Å and the C(1)-O(1)-M(1) angles of 137-146° are
consistent with sp2 hybridization at O,16 and thus the alkoxide
acts formally as a three-electron donor. In the structure of
[Cp*ZrL2Cl2] steric interactions between the oxazoline methyl
group at C(29) and the Cp* ligand cause a fold in the metal
chelate ring such that the angle θ between the least-squares
planes defined by N(1)-C(7)-C(8)-C(1)-O(1) and N(1)-
Zr(1)-O(1) is 21.17(50)°. In the complexes [Cp*ML1Cl2] (M
) Zr, Hf), where there is no significant steric pressure in this
region, the angles θ are 8.64(18) and 9.23(22)°, respectively.
The fold angle φ in the oxazoline ring (the angle between the
planes defined by N(1)-C(7)-O(2)-C(9) and N(1)-C(8)-
C(9)) is also the largest for [Cp*ZrL2Cl2]. As a further
consequence, C(1)-O(1)-M(1) for [Cp*ZrL2Cl2] is by far the
lowest of the three chloride complexes at 137.1(4)° and the
M(1)-N(1) distance is the largest. Interestingly, the M(1)-N(1)
distances for the L1 complexes of 2.295(3) and 2.277(4) Å are
rather shorter than those observed for analogous imino-phenolate
systems at ca. 2.35 Å,17 and this may be indicative of a stronger
bond resulting from the presence of the conjugated O atom in
(5) (a) Yoon, J.; Mathers, R. T.; Coates, G. W.; Thomas, E. L.
Macromolecules 2006, 39, 1913. (b) Cherian, A. E.; Lobkovsky, E. B.;
Coates, G. W. Macromolecules 2005, 38, 6259. (c) De Rosa, C.; Circelli,
T.; Auriemma, F.; Mathers, R. T.; Coates, G. W. Macromolecules 2004,
37, 9034. (d) Mason, A. F.; Coates, G. W. J. Am. Chem. Soc. 2004, 126,
10798. (e) Fujita, M.; Coates, G. W. Macromolecules 2002, 35, 9640. (f)
Hustad, P. D.; Coates, G. W. J. Am. Chem. Soc. 2002, 124, 11578. (g)
Hustad, P. D.; Tian, J.; Coates, G. W. J. Am. Chem. Soc. 2002, 124, 3614.
(h) Tian, J.; Coates, G. W. Angew. Chem., Int. Ed. 2000, 39, 3626.
(6) (a) Gomez, R.; Green, M. L. H.; Haggitt, J. L. J. Chem. Soc., Chem.
Commun. 1994, 22, 2607. (b) Chernega, A. N.; Gomez, R.; Green, M. L.
H. J. Chem. Soc., Chem. Commun. 1993, 18, 1415. (c) Gomez, R.;
Duchateau, R.; Chernega, A. N.; Teuben, J. H.; Edelmann, F. T.; Green,
M. L. H. J. Organomet. Chem. 1995, 491, 153. (d) Jayaratne, K. C.; Sita,
L. R. J. Am. Chem. Soc. 2000, 122, 958. (e) Keaton, R. J.; Jayaratne, K.
C.; Henningsen, D. A.; Koterwas, L. A.; Sita, L. R. J. Am. Chem. Soc.
2001, 123, 6197. (f) Kissounko, D. A.; Fettinger, J. C.; Sita, L. R. Inorg.
Chim. Acta 2003, 345, 121. (g) Zhang, Y. H.; Reeder, E. K.; Keaton, R. J.;
Sita, L. R. Organometallics 2004, 23, 3512. (h) Kissounko, D. A.; Zhang,
Y. H.; Harney, M. B.; Sita, L. R. AdV. Synth. Catal. 2005, 347, 426. (i)
Zhang, Y. H.; Sita, L. R. Chem. Commun. 2003, 18, 2358. (j) Koterwas, L.
A.; Fettinger, J. C.; Sita, L. R. Organometallics 1999, 18, 4183. (k) Sita,
L. R.; Babcock, J. R. Organometallics 1998, 17, 5228.
(7) Doherty, S.; Errington, R. J.; Jarvis, A. P.; Collins, S.; Clegg, W.;
Elsegood, M. R. J. Organometallics 1998, 17, 3408.
(8) Bott, R. K. J.; Hughes, D. L.; Schormann, M.; Bochmann, M.;
Lancaster, S. J. J. Organomet. Chem. 2003, 665, 135.
(9) Bott, R. K. J.; Hammond, M. L.; Horton, P. N.; Lancaster, S. J.;
Bochmann, M.; Scott, P. Dalton Trans. 2005, 22, 3611.
(10) (a) Westmoreland, I.; Munslow, I. J.; Clarke, A. J.; Clarkson, G.;
Deeth, R. J.; Scott, P. J. Organomet. Chem. 2006, 691, 2228. (b)
Vorbruggen, H.; Krolikiewicz, K. Tetrahedron 1993, 49, 9353. (c) Vor-
bruggen, H.; Krolikiewicz, K. Tetrahedron Lett. 1981, 22, 4471.
(11) Wolczanski, P. T.; Bercaw, J. E. Organometallics 1982, 1, 793.
(12) Roddick, D. M.; Fryzuk, M. D.; Seidler, P. F.; Hillhouse, G. L.;
Bercaw, J. E. Organometallics 1985, 4, 97.
(13) Morris, R. J.; Shaw, S. L.; Jefferis, J. M.; Storhoff, J. J.; Goedde,
D. M. Inorg. Synth. 1998, 32, 215.
(14) Mach, K.; Varga, V.; Antropiusova, H.; Polacek, J. J. Organomet.
Chem. 1987, 333, 205.
(15) Schock, L. E.; Marks, T. J. J. Am. Chem. Soc. 1988, 110, 7701.
(16) Bei, X. H.; Swenson, D. C.; Jordan, R. F. Organometallics 1997,
16, 3282.

Group 4 Salicyloxazoline Catalysts
Organometallics, Vol. 25, No. 26, 2006 6021
Table 1. Crystallographic Data, Collection Parameters, and Refinement Parameters for [Cp*ZrL1Me2], [Cp*HfL1Cl2],
[Cp*ZrL2Cl2], and [Cp*ZrL1Cl2]
[Cp*ZrL1Cl2]‚0.25Et2O
[Cp*ZrL1Me2]
[Cp*ZrL2Cl2]
[Cp*HfL1Cl2]‚0.25Et2O
empirical formula
fw
C28H41.5Cl2NO2.25Zr
599.76
0.18 × 0.15 × 0.14
monoclinic
P21/n
9.8486(3)
13.7408(4)
22.8455(6)
90
98.4590(10)
90
C29H45NO2Zr
530.88
C29H43Cl2NO2Zr
599.76
0.18 × 0.15 × 0.14
monoclinic
P21
9.4657(6)
16.7741(10)
9.6448(6)
90
102.54
90
1494.86(16)
1.332
0.571
C28H41.5Cl2HfNO2.25
677.51
0.25 × 0.23 × 0.08
monoclinic
P21/n
9.9224(2)
13.8597(3)
23.0046(5)
90
98.5900(10)
90
cryst size (mm3)
cryst syst
space group
a (Å)
0.32 × 0.10 × 0.06
triclinic
P1h
10.5992(15)
11.5384(17)
12.7577(19)
64.097(3)
82.870(3)
82.570(3)
1387.7(4)
1.271
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
3057.99(15)
1.282
0.558
3128.14(11)
1.439
3.529
D
calcd (Mg/m3)
µ (mm-1
)
0.420
564
F000
1234
628
1362
total no. of rflns
no. of indep rflns
Rint
19 166
7387
0.0912
17 565
6698
0.0506
9665
6630
0.0805
19 282
7376
0.0600
no. of data/restraints/ params
R1 (I > 2σ(I))
wR2
7387/2/322
0.0536
0.1314
6698/0/311
0.0431
0.0834
6630/1/330
0.0669
0.1398
7376/4/327
0.0389
0.0938
GOF on F2
0.914
0.962
0.873
0.968
Table 2. Selected Bond Lengths (Å) and Angles (deg) for [Cp*ZrL1Me2], [Cp*HfL1Cl2], [Cp*ZrL2Cl2], and [Cp*ZrL1Cl2]
[Cp*ZrL1Cl2]
[Cp*HfL1Cl2]
[Cp*ZrL2Cl2]
[Cp*ZrL1Me2]
M(1)-O(1)
2.004(3)
2.4577(11)
2.4641(11)
2.021(3)
2.4550(12)
2.4642(12)
2.013(5)
2.441(2)
2.4758(19)
2.0275(15)
M(1)-Cl(1)
M(1)-Cl(2)
Zr(1)-C(30)
Zr(1)-C(31)
M(1)-N(1)
2.282(3)
2.270(2)
2.3598(19)
2.231
2.295(3)
2.214
87.66(4)
143.21(8)
81.77(8)
2.277(4)
2.216
86.91(4)
142.25(11)
81.79(10)
2.358(5)
2.217
85.02(7)
144.20(15)
86.4(2)
M(1)-Cp*CEN
Cl(1)-M(1)-Cl(2)
N(1)-M(1)-Cl(2)
N(1)-M(1)-Cl(1)
C(30)-Zr(1)-C(31)
O(1)-M(1)-N(1)
C(1)-O(1)-M(1)
C(7)-N(1)-M(1)
θa
84.23(10)
74.18(6)
147.14(15)
127.77(16)
10.52
75.91(10)
145.6(2)
128.0(3)
8.64
76.50(13)
144.1(3)
128.1(3)
9.23
76.00(19)
137.1(4)
122.4(5)
21.17
φb
14.23
14.45
20.04
13.01
a The angle θ is defined as the angle between the planes defined by N(1)-C(7)-C(8)-C(1)-O(1) and N(1)-Zr(1)-O(1). b The angle φ is defined as the
angle between the planes defined by N(1)-C(7)-O(2)-C(9) and N(1)-C(8)-C(9).
Figure 1. Molecular structure of [Cp*ZrL1Cl2]. [Cp*HfL1Cl2] is
isomorphous.
Figure 2. Molecular structure of [Cp*ZrL2Cl2].
the oxazoline ring. The steric demand of the proligand has little
In [Cp*ZrL1Me2] (Figure 3) the Zr(1)-O(1) and Zr(1)-N(1)
distances at 2.0275(15) and 2.3598(19) Å, respectively, are
significantly longer than those in the analogous chloride complex
[Cp*ZrL1Cl2] at 2.004(3) and 2.295(3) Å, as expected on the
basis of electrostatics (Cl more electronegative than Me). In
other respects the last two structures are very alike, and other
structural parameters are similar to those observed in analogous
systems.6g,18
effect on the Cl(1)-M-Cl(2) angle (all 85-88°), although the
N(1)-Zr(1)-Cl(1) angle in [Cp*ZrL2Cl2] is ca. 4° larger than
in the complexes of L1.
(17) (a) Knight, P. D.; Clarkson, G.; Hammond, M. L.; Kimberley, B.
S.; Scott, P. J. Organomet. Chem. 2005, 690, 5125. (b) 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.

6022 Organometallics, Vol. 25, No. 26, 2006
Coles et al.
Figure 3. Molecular structure of [Cp*ZrL1Me2].
Figure 5. Eyring plot for [Cp*ZrL2Me2].
Table 3. Data from Eyring Plotsa for ∆Gq, ∆Hq, and ∆Sq
b
∆Hq/
∆Sq/
∆Gq
/
∆Gq
/
298
298
complex
kJ mol-1
J K-1 mol-1
kJ mol-1
kJ mol-1
[Cp*ZrL1Me2] 60.0 ( 1.9 -43.9 ( 5.6 73.1 ( 3.5 73.3 ( 1.5
[Cp*ZrL2Me2] 74.9 ( 3.4
-2.0 ( 10 75.7 ( 6.2 75.1 ( 1.1
a Data were obtained from line-shape analysis unless otherwise noted.
b Data were obtained from eq 1.
For each individual complex the two methods lead to the same
value for ∆Gq within standard errors.22,23 Notably, the values
of ∆Gq for the two complexes are effectively the same, but the
entropy of activation component is much more negative for the
complex of L1 than for that of L2. This may be related to the
rather more “ordered” structure of the complex [Cp*ZrL2Me2]
compared to that of [Cp*ZrL1Me2]; in the latter the salicylox-
azoline ligand is free to hinge about the N-O vector, whereas
in the former it is rather fixed in position. In any event, the
similarity of ∆Gq values between the two complexes is
consistent with an N-dissociative mechanism for racemization;
a rotation-in-place mechanism would give a comparatively high
value for [Cp*ZrL2Me2], as a result of steric interactions
between methyl groups on the oxazoline ring and the Cp* ligand.
We cannot, however, exclude the possibility that the two
complexes undergo exchange by different mechanisms; on the
basis that the more sterically encumbered system [Cp*ZrL2-
Me2] would be the most likely of the two to undergo N-
dissociative exchange, we would expect this compound to have
the most positive ∆Sq value, which is indeed the case.
Ethylene Polymerization Studies. The complexes
[Cp*MLnX2] (M ) Zr, Hf; n ) 1, 2; X ) Cl, Me) and
[CpTiL1Cl2] were tested for activity in ethylene polymerization
(Table 4). The zirconium complexes exhibited a range of pro-
ductivities, varying with the nature of the salicyloxazoline colig-
and, conditions, and cocatalyst, whereas the hafnium complexes
showed essentially no activity regardless of experimental
conditions.
Figure 4.
Variable-temperature 1H NMR spectra of
[Cp*ZrL2Me2].
VT NMR Studies. The chirality of all the complexes
described above is manifest in their 1H NMR spectra at ambient
temperature. For example, in [Cp*ZrL1Cl2] all four oxazoline
protons are inequivalent, giving rise to mutually coupled
complex multiplets around 3.50 and 3.77 ppm (1H each), and
two overlapping multiplets between 3.36 and 3.46 ppm (2H
total). In [Cp*ZrL2Cl2] a pair of doublets is observed at ca.
3.35 and 3.49 ppm and the two oxazoline methyl groups appear
as separate singlets at 1.04 and 1.49 ppm. In [Cp*ZrL1Me2]
the oxazoline CH2 groups appear as four distinct multiplets, and
the zirconium-bound methyl groups are also inequivalent (0.03
and 0.41 ppm, and 37.5 and 39.4 ppm in the 13C NMR
spectrum).
1
We recorded the temperature dependence of the H NMR
spectra for [Cp*ZrL1Me2] and [Cp*ZrL2Me2]. Figure 4 shows
the oxazoline H atom region for [Cp*ZrL2Me2]. At 298 K, these
are observed as two mutually coupled AB doublets. As the
temperature is increased, the peaks broaden and then coalesce
at ca. 373 K. These spectra are consistent with exchange between
enantiomers (racemization).19 A line shape analysis20 gave
exchange rate data for this process (Figure 5). The thermody-
namic parameters thus extracted are shown in Table 3 along
Treatment of [Cp*ZrLnX2] with 1000 equiv of MAO cocata-
lyst at 1.2 bar of ethylene pressure (runs 1-4) gave the most
with the values for ∆Gq calculated directly using eq 1.21
298
∆Gq ) aT[10.319 + log(T/k)]
a ) 0.01914 (1)
(21) Sandstro¨m, J. In Dynamic NMR Spectroscopy; Academic Press:
London, 1982; pp 93-123.
(22) Hamilton, L. C. In Regression With Graphics-A Second Course in
Applied Statistics; Brooks/Cole: Belmont, CA, 1991; pp 29-64. Blank, L.
In Statistical Procedures for Engineering, Management and Science;
McGraw-Hill: New York, 1980; pp 487-515.
(23) The standard errors involved in the Eyring plot were calculated using
Origin 7.0. The error involved in the single equation is based upon an error
in T of (1 K and an error in k of (40%. This error value for k comes from
the largest deviation from the linear fit in the Eyring plot.
(18) Zhang, Y. H.; Kissounko, D. A.; Fettinger, J. C.; Sita, L. R.
Organometallics 2003, 22, 21.
(19) Although we have not formally excluded the possibility of a
mechanism involving Zr-O bond cleavage, this is very unlikely under these
conditions.
(20) Line shape analysis was carried out using WIN-DYNA 32, version
1.01, release 981101, Bruker Analytik GmbH, 1994-1998.

Group 4 Salicyloxazoline Catalysts
Organometallics, Vol. 25, No. 26, 2006 6023
Table 4. Polymerization of Ethene and Hexene Using [Cp*ZrLnX2]
amt of cat./
µmol
amt of cocat/
equiv
cocat
concn/M
Al concn/
M
temp/
°C
av productivity/
1g (mol bar h)-1
run no.
precat.
cocat.
Mn
Mw
1a
[Cp*ZrL1Cl2]
[Cp*ZrL1Me2]
[Cp*ZrL2Cl2]
[Cp*ZrL2Me2]
[CpTiL1Cl2]
2.1
11
12
18
9.8
11
11
11
11
11
1.8
10
10
13
1.8
11
11
11
10
13
11
MAO
MAO
MAO
MAO
1000
1000
1000
1000
1000
1.1
1.2 × 10-2
6.5 × 10-2
7.1 × 10-2
0.10
30
30
30
30
30
30
30
30
30
30
70
30
70
30
70
30
70
30
30
30
30
1.5 × 105
2 340
1 640
5 750
460 000
266 000
126 000
2a
8.5 × 104
3a
8.8 × 104
4a
3.4 × 103
5a
MAO
5.6 × 10-2
3.74 × 10-3
2.43 × 10-3
3.74 × 10-3
2.43 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
6.05 × 10-3
1.7 × 103
6a
[Cp*ZrL1Cl2]
[Cp*ZrL1Cl2]
[Cp*ZrL1Cl2]
[Cp*ZrL1Cl2]
[Cp*ZrL1Cl2]
[Cp*ZrL1Cl2]
[Cp*ZrL2Cl2]
[Cp*ZrL2Cl2]
[Cp*ZrL3Cl2]
[Cp*ZrL3Cl2]
[Cp*ZrL4Cl2]
[Cp*ZrL4Cl2]
[Cp*ZrL1Cl2]
[Cp*ZrL2Cl2]
[Cp*ZrL3Cl2]
[Cp*ZrL4Cl2]
1
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
7.48 × 10-5
5.3 × 10-5
7.48 × 10-5
5.3 × 10-5
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
1.1 × 10-4
2.3 × 105
32 200
28 600
18 900
4 660
12 000
4 680
644 000
59 600
47 000
445 000
39 300
24 000
63 000
37 300
44 000
67 900
51 100
109 000
49 900
28 700
52 800
41 800
7a
1.1
1.1
1.1
1.1
6.6
1.1
1.1
1.1
6.6
1.1
1.1
1.1
1.1
1.1
1.1
1.9 × 105
8b
5.6 × 104
9b
8.7 × 104
10c
11c
12c
13c
14c
15c
16c
17c
18d
19d
20d
21d
2.7 × 105
1.0 × 106
5.3 × 104
8 660
5 560
9.0 × 104
7.0 × 104
14 400
4 650
13 100
12 500
18 800
5 650
6.1 × 105
2.7 × 104
1.3 × 105
8.5 × 104 (5.2)e
3.3 × 104 (trace)e
1.4 × 105(2.7)e
2.7 × 104 (0.4)e
22 800
10 500
a Solvent toluene (180 mL). b Solvent bromobenzene (100 mL). c Solvent toluene (100 mL). d Solvent 20 wt % 1-hexene in toluene (100 mL). e 1-Hexene
contents (mol %) in parentheses.
Figure 7. GPC data for PE from [Cp*ZrL1Cl2].
Figure 6. GPC data for PE from [Cp*ZrL1Cl2] and various
activators.
is the case (runs 8 and 9). The molecular weights of the polymers
promising results for [Cp*ZrL1Cl2], the bulkier ligand L2 giving
consistently lower productivity. Surprisingly, the complexes X
) Me led to relatively poor catalysts. Also, [CpTiL1Cl2] (run
5) has low productivity; the analogous imino-phenolate com-
plexes gave ca. 1.0 × 104-2.0 × 105 g of PE (mol bar h)-1
under very similar conditions.8
The stoichiometric activators [Ph3C][B(C6F5)4] (1) and
[PhNMe2H][B(C6F5)4] (2), when used in conjunction with
AlBui3 as an alkylating agent and scavenger, were significantly
more productive (runs 6 and 7). The variation of cocatalyst also
leads to interesting changes in the physical properties of the
polymers produced (Figure 6). Polyethylene produced when
either MAO or 1/AlBui3 is used as the cocatalyst has a bimodal
molecular weight distribution, consistent with the formation of
more than one active species (vide infra). The use of 2 gives a
unimodal distribution with a narrow PDi (2.1), characteristic
of single-site catalysis. In contrast, precatalysts [CpM(PI)Cl2]
gave multimodal molecular weight distributions, although only
activation with MAO was reported.8
produced in chlorobenzene are also slightly lower than those
in toluene.
We have been concerned with the thermal stability and
deactivation mechanisms of PI and related catalyst systems9,17a
and thus tested the most active of the current catalysts using a
gas delivery measurement system in order to profile the catalyst
activity over time, at a constant temperature. The average
productivities for these runs (runs 10-17) calculated from the
isolated yield of polyethylene in a 1 h run correlated well with
the measured gas flow and with earlier runs.25 The average
productivity of each catalyst is higher at 70 °C than at 30 °C,
particularly in the case of [Cp*ZrL1Cl2], which at 1.0 × 106 g
of PE (mol bar h)-1 is highly active.3b Molecular weight
distributions from GPC of the polymers thus made at 30 °C
(run 10) and 70 °C (run 11) are shown in Figure 7. At the lower
temperature, the polymer (unimodal, PDI ) 3.3) is similar to
that from run 7. At 70 °C there is a reduction in molecular
(24) (a) Jayaratne, K. C.; Keaton, R. J.; Henningsen, D. A.; Sita, L. R.
J. Am. Chem. Soc. 2000, 122, 10490. (b) Jayaratne, K. C.; Sita, L. R. J.
Am. Chem. Soc. 2000, 122, 958.
(25) In the context of catalyst stability studies, a comparison with [Cp*Zr-
(PI)Cl2] would be instructive, but unfortunately we have thus far been unable
to synthesise these compounds (vide supra).
Sita has shown that chlorobenzene is a more successful
solvent than toluene for the polymerization and copolymerization
of alkenes with related half-sandwich group 4 catalysts.24 We
found that for [Cp*ZrL1Cl2] activated with 1 or 2 the opposite

6024 Organometallics, Vol. 25, No. 26, 2006
Coles et al.
mined by 13C NMR spectroscopy using the method of Hsieh
and Randall.27 The catalysts incorporating the bulkier salicy-
loxazolines, i.e., [Cp*ZrL2Cl2] and [Cp*ZrL4Cl2], gave very
low or negligible incorporation compared with [Cp*ZrL1Cl2]
or [Cp*ZrL3Cl2]. The rate of ethylene uptake and productivity
with [Cp*ZrL3Cl2] is higher than with [Cp*ZrL1Cl2].
NMR Studies of the Catalyst Activation Process. We
studied the reactions of [Cp*ZrL1Me2] in d5-bromobenzene with
the cocatalysts used in the above polymerization processes.
Treatment with ca. 20 equiv of dried MAO led to broad features
1
in the H NMR spectrum, and identification of a product or
products was not feasible. In particular the very broad resonance
for MAO around 0 ppm obscures the region where metal-bound
methyl signals are expected. The reaction with a slight excess
of trityl (1) and anilinium (2) led to complete consumption28 of
the starting materials and production of a major new organo-
1
Figure 8. Ethylene polymerization activity profile of
[Cp*ZrL1Cl2].
metallic species, common to both experiments. The H NMR
spectrum is consistent with the formulation [Cp*ZrL1Me]-
[B(C6F5)4] (Figure 9). The oxazoline region reveals that, like
the starting material, this compound is chiral (four mutually
coupled oxazoline H atoms), but as can be seen most clearly in
Figure 9a, it has only one methyl group (δ 0.24 ppm). The
expected coproducts Ph3CMe and PhNMe2 were also formed
in appropriate quantities. A small peak at the correct chemical
shift for methane in this solvent (0.142 ppm) was also detected
in the reaction with 2.
The product from 1 (Figure 9b) has a significant impurity of
unidentified oxazoline (or similar) species which display
resonances in the metal alkyl region (0.02-0.12 ppm) and thus
represent potentially catalytic impurities. We speculate that the
acid-sensitive oxazoline unit may have taken part in side
reactions with the trityl ion. The product from 2 (Figure 9a)
included a small amount of a different oxazoline-containing
impurity, but there was no corresponding Zr-Me resonance.
Our proposal that this arises from the dicationic species
[Cp*ZrL1(NMe2Ph)]2+ is encouraged by the observation of two
distinct resonances at 2.61 and 2.45 ppm (1:1 ratio) with an
appropriate integral for the diastereotopic NMe2 group. Notably,
the major species [Cp*ZrL1Me][B(C6F5)4] does not form
an adduct with PhNMe2, even at this relatively high concentra-
tion.29
The properties of the polymers produced under catalytic
conditions by [Cp*ZrL1X2] systems are consistent with the
information available from the above catalyst activation
studies. (i) The [Cp*ZrL1Me2]/2 system gives a single catalytic
species (Figure 9b) and a unimodal polymer (Figure 6). (ii)
[Cp*ZrL1Me2]/1 gives the same catalyst as 2 along with at least
one other potential catalyst; this is also reflected in the GPC
data. In addition, we might also speculate that the catalyst
species responsible for the higher weight fractions present in
the [Cp*ZrL1Me2]/1 and [Cp*ZrL1Me2]/MAO polymers arises
from reactions of these strong Lewis acids with the oxazoline
ligand.
weight of the main peak and a small amount of higher molecular
weight polymer was also formed, presumably as a result of
catalyst decomposition to another active species at this higher
temperature.
The activity profiles of [Cp*ZrL1Cl-2] at 30 and 70 °C (runs
10 and 11) are shown in Figure 8. After an initial induction
period the activity of [Cp*ZrL1Cl2] increases steadily across a
1 h run at 30 °C. This may be due to a relatively low rate of
activation of the precatalyst with AlBui3 and [PhNMe2H]-
[B(C6F5)4] (vide infra); unfortunately, preactivation of the
catalyst for longer periods generally led to poor activity on
admission of ethylene gas. At 70 °C uptake of ethylene is
observed instantaneously and the initial rate of polymerization
is ca. 2.4 × 106 g of PE (mol bar h)-1. This decreases fairly
steadily over 1 h, with t1/2 ) 20 min, although after 1 h the
activity is still a moderate, 2.8 × 105 g of PE (mol bar h)-1
.
Under the same conditions, [Cp*ZrL2Cl2] showed an initial
activity an order of magnitude lower at 2.9 × 105 g of PE (mol
bar h)-1 and decomposed much more rapidly (t1/2 ca. 5 min).
[Cp*ZrL2Cl2] also decomposed at 30 °C, with t1/2 ) 11 min
(runs 12 and 13). Since the mechanism of catalyst death is most
probably transfer of the salicyloxazoline ligand to aluminum
(vide infra), the relative instability of the catalyst derived from
[Cp*ZrL2Cl2] may be an effect of the larger angle θ for the L2
complex (Table 2), leading to greater accessibility of the
phenolate oxygen atom to attack by AliBu3.
Fujita noted that the introduction of a methoxy group para
to the phenolate O atom in certain [Zr(PI)2Cl2] complexes led
to higher productivities.26 While it is not apparent whether this
arises from an improvement in intrinsic activity or reduction in
the rate of catalyst decomposition, we decided to investigate
the effect on the present salicyloxazoline system (ligands L3
and L4). In runs 14 and 15, [Cp*ZrL3Cl2] showed a slightly
lower average productivity than [Cp*ZrL1Cl2] but at 70 °C was
rather longer lived; with t1/2 ) 27.5 min [Cp*ZrL4Cl2] (runs
16 and 17) behaved similarly (t1/2 ) 30 min) at this higher
temperature. The thermal stability of the catalyst does thus
appear to improve with the introduction of a p-methoxy group,
without substantially altering the productivity.
Computational Investigations of the Catalytic Reac-
tion. DFT calculations of the mechanism of the insertion of
one ethylene unit into the metal-bound methyl group of
[Cp*ZrL1Me]+ were carried out.30 Two general directions of
Ethylene/1-Hexene Copolymerizations. While propene and
hexene were not homopolymerized to a significant degree by
the present catalysts, ethylene/hexene copolymers were produced
(runs 18-21). The hexene content of the polymer was deter-
(27) Hsieh, E. T.; Randall, J. C. Macromolecules 1982, 15, 1402.
(28) In the case of 2 the reaction took place overnight, although the
product alkyl cation was remarkably stable, existing at ambient temperatures
for several days in solution before noticeable decomposition.
(29) Bochmann, M.; Lancaster, S. J. J. Organomet. Chem. 1992, 434,
C1.
(26) Matsukawa, N.; Matsui, S.; Mitani, M.; Saito, J.; Tsuru, K.; Kashiwa,
N.; Fujita, T. J. Mol. Catal. A: Chem. 2001, 169, 99.
(30) These studies were carried out without consideration of counteranion
and solvent effects.

Group 4 Salicyloxazoline Catalysts
Organometallics, Vol. 25, No. 26, 2006 6025
Figure 9. 1H NMR spectra of the reactions of [Cp*ZrL1Me2] with (a) [PhNHMe2][B(C6F5)4] and (b) [Ph3C][B(C6F5)4]. The asterisk indicates
an impurity of silicone vacuum grease, and the plus sign indicates methane.
(pathway A) and 3.7 kcal mol-1 (pathway B) lower in energy
than the corresponding Me cation plus free ethylene. A
subsequent transition state for insertion of ethylene into the
Zr-Me bond was found 3.4 and 10.0 kcal higher in energy,
respectively. These barriers are within the range typically
calculated for metallocenes and other highly active single-site
catalysts.31 Both transition states have an R-agostic interaction,
with the Cγ-H bond of the Zr-bound methyl group lengthened
from 1.09 to 1.11 Å and a Zr-H distance of 2.36 Å. The
Zr-propyl reaction products are similar in energy for both
pathways and are 22.0 and 22.8 kcal mol-1 below the sum of
energies of the reactants. In neither structure was a γ-agostic
interaction observed. The complex adopts a distorted three-
legged piano-stool configuration with the propyl chain bent
Figure 10. Projections of the computed isomeric structures of
[Cp*ZrL1Me(C2H4)]+ corresponding to pathways A and B, viewed
along the Zr-Cp* centroids.
toward the former ethylene coordination site. Chain straightening
leads in both cases to a nearly symmetrical piano-stool config-
uration, with the perfectly staggered propyl chain directed away
from the Cp* ring and an additional energy release of 2.4 kcal
mol-1 (pathway A) and 1.1 kcal mol-1 (pathway B). A â-agostic
confomer was only found for pathway B, with the agostic
interaction trans to the oxygen donor and an additional energy
release of 3.2 kcal mol-1. In this structure the Câ-H bond of
the agostic hydrogen is lengthened to 1.15 Å and the Zr-H
distance is 2.22 Å. No attempts were made to determine the
barriers for these propyl chain isomerizations.
For [Cp2ZrPr]+, all confomers contain agostic ineractions.
In addition to the strong â-agostic ineraction with one â-H in
the “beta” isomer, two weaker interactions with both â-H in
the “straight” isomer and an interaction with one γ-H in the
“bent” isomer can be observed. Structure files for all calculated
species are available as Supporting Information.
For [Cp*ZrL1Me]+ the ethylene coordination energy (1.2 or
3.7 kcal mol-1) is very low compared to the values for
[Cp2ZrMe]+ obtained in this work (20.8 kcal mol-1) and by
Figure 11. Optimized energy levels, relative to starting materials
Ziegler and co-workers.32 Steric hindrance encountered in the
ethylene adduct [Cp*ZrL1Me(C2H4)]+, either with the hydrogen
atoms in the 4-position on the oxazoline (pathway A) or the
o-tBu substituent on the salicyl ring (pathway B) is depicted in
Figure 10. The steric constraints imposed by the ligand
framework are also evidenced by the absence of any γ-agostic
interaction in the [Cp*ZrL1Pr]+ insertion products. It is also
notable also in this context that in the three-legged piano-stool
species [Cp*ZrL1Me]+ the methyl group has to move aside in
order to accommodate the incoming ethylene ligand. Our
optimized structure of [Cp2ZrMe]+, very similar to that produced
by Ziegler,32 is pyramidal and is thus preorganized for coordina-
tion of alkene.
for the insertion of C2H4 into the Zr-Me bond of [Cp2ZrMe]+ and
[Cp*ZrL1Me]+ along pathways A and B.
approach of the ethylene monomer toward the metal center were
considered (Figure 10). The computed energies and selected
bond distances are presented in Figure 11 and Table 5, together
with the values obtained for [Cp2ZrMe]+ for comparison,
calculated at the same level of theory.
The crystallographic paramaters of Cp*ZrL1Me2 were used
as a starting point for the calculations. Removing one methyl
ligand leads to two unoptimized [Cp*ZrL1Me]+ complexes with
a vacant coordination site trans to either the O or N donor of
L1. A subsequent geometry optimization leads in both cases to
a three-legged piano-stool configuration with nearly equivalent
geometry energy (within 0.1 kcal mol-1). Addition of ethylene
leads to an intermediate ethylene adduct at 1.2 kcal mol-1
(31) Rappe´, A. K.; Skiff, W. M.; Casewitt, C. J. Chem. ReV. 2000, 100,
1435.
(32) Woo, T. K.; Fan, L.; Ziegler, T. Organometallics 1994, 13, 2252.

6026 Organometallics, Vol. 25, No. 26, 2006
Coles et al.
Table 5. Calculated Relative Energies (kcal mol-1) and Selected Bond Distances (Å) for Optimized Geometries for the Ethylene
Insertion into [Cp*ZrL1Me]+ and [Cp2ZrMe]+
bond dist
compd
path
rel energy
Zr-CR
Zr-Câ
Zr-Cγ
CR-Câ
Câ-Câ
CR-Ha
Câ-H a
Cγ-H a
[Cp*ZrL1Me]+ + C2H4
[Cp*ZrL1Me(C2H4)]+
TS for C2H4 insertion
A
A
A
A
A
A
B
B
B
B
B
B
0.0
-1.2
2.2
-22.0
-24.4
2.24
2.29
2.38
2.26
2.42
2.23
2.25
2.34
2.72
1.40
1.40
1.53
1.54
not obsd
2.22
1.53
1.54
1.11
[Cp*ZrL1Me(Pr)]+ (bent)
[Cp*ZrL1Me(Pr)]+ (straight)
[Cp*ZrL1Me(Pr)]+ (beta)
[Cp*ZrL1Me]+ + C2H4
[Cp*ZrL1Me(C2H4)]+
TS for C2H4 insertion
-0.1
-3.7
6.3
2.23
2.97
2.67
2.29
2.44
2.24
2.24
2.25
2.95
2.34
1.37
1.41
1.53
1.54
1.51
213
1.53
1.54
1.11
1.15
1.14
[Cp*ZrL1Me(Pr)]+ (bent)
[Cp*ZrL1Me(Pr)]+ (straight)
[Cp*ZrL1Me(Pr)]+ (beta)
[Cp2ZrMe]+ + C2H4
-22.8
-23.9
-27.1
0.0
-20.8
-16.6
-29.2
-31.6
-35.3
2.26
2.27
2.32
[Cp2ZrMe(C2H4)]+
2.84
2.41
2.24
2.26
2.28
2.82
TS for C2H4 insertion
1.41
1.56
1.53
1.50
2.19
1.57
1.55
1.53
[Cp2ZrMe(Pr)]+ (bent)
[Cp2ZrMe(Pr)]+ (straight)
[Cp2ZrMe(Pr)]+ (beta)
1.12
1.11
1.17
a C-H bond distances for agostic hydrogens.
The optimized structures for the [Cp*ZrL1Me(C2H4)]+ in-
termediates in both pathways show a considerable difference
in their Zr-ethylene bond distances: 2.26 and 2.34 Å (pathway
A) versus 2.95 and 2.97 Å (pathway B). A cut through the total
electron density map along the O-Zr-N plane for both
intermediates (Figure 12) reveals a weaker electron density along
the Zr-N bond compared to that along the Zr-O bond. This
weaker electron-donating capability of the nitrogen donor should
render the coordination site trans to N more electrophilic and
hence strengthen the Zr-ethylene bond. However, the Zr-C
bond distances at this site are longer than those in the trans-O
site. This implies that the ethylene addition at the site trans to
nitrogen is disfavored for purely steric reasons. This is in
agreement with the absence of a â-agostic interaction at the
trans-N site.
Conclusions
A new series of group 4 half-sandwich salicyloxazoline
complexes bearing Cp* ligands has provided new precatalysts
for alkene polymerization. Of the catalyst activators studied,
[PhNMe2H][B(C6F5)4] is the best matched, smoothly producing
the single, chiral, catalytically competent species [Cp*ZrL1Me]-
[B(C6F5)4] and subsequently promoting the highly active single-
site polymerization of ethene (and less active copolymerization
of ethene/hexene). The Lewis acidic activator [Ph3C][B(C6F5)4]
produces the same catalyst and at least one other, perhaps
through a reaction of the basic oxazoline unit. Activation with
MAO leads to a qualitatively different and less well-defined
system.
This type of catalyst is sensitive to steric effects, and light is
cast on the origins of this by the structural and computational
work. Approach of ethene to either coordination site is hampered
by the o-tBu substituents on the salicyl ring or the oxazoline
4-substituents (R2), even in the least sterically demanding cases
(L1 and L3, where R2 ) H). This reluctance to coordinate
ethylene is at the heart of the modest (by metallocene standards)
productivity of these catalysts in ethylene polymerization. The
stability of the three-legged piano-stool structure [Cp*ZrL1Me]+
(observed in the NMR experiments; e.g., Figure 9) is also
reflected in the fact that it does not coordinate dimethylaniline.
The catalysts based on L2 and L4 (R2 ) Me) have intrinsic
activities at least an order of magnitude lower. They also
Figure 12. Total electron density in the N-Zr-O plane for
[Cp*ZrL1Me(C2H4)]+ from pathways A (upper) and B (lower).
Cp* and the remaining salicyloxazoline ligand are removed for
clarity.
decompose more rapidly, probably because a structural change
caused by the greater steric demand makes the phenoxide unit
susceptible to ligand transfer to aluminum in the cocatalyst
mixture.
Interestingly, the introduction of a p-methoxy substituent has
a more profound effect on stability than it does on activity,

Group 4 Salicyloxazoline Catalysts
Organometallics, Vol. 25, No. 26, 2006 6027
dried thoroughly in vacuo. Desolvation occurred during this
latter process, and only traces of solvent were detected by
NMR spectroscopy. Yield: 178 mg, 56%. Anal. Calcd for
C27H39Cl2NO2Zr: C, 56.72; H, 6.88; N, 2.45. Found: C, 55.88; H,
6.90; N, 2.59. 1H NMR (400 MHz, 298 K, C6D6): δ 7.83 (d, 4JHH
commensurate with the growing body of evidence for a
mechanism of catalyst death in phenoxyimine and related
systems by ligand loss.9,33
Experimental Details
4
) 3 Hz, 1H, Ar H), 7.72 (d, JHH ) 3 Hz, 1H, Ar H), 3.70-3.85
(m, 1H, NCH2), 3.48-3.55 (m, 1H, OCH2), 3.36-3.46 (m, 2H,
NCH2CH2O), 1.90 (s, 15H, Cp*), 1.66 (s, 9H, C(CH3)3), 1.26 (s,
9H, C(CH3)3) ppm. 13C{1H} NMR (100 MHz, 298 K, C6D6): δ
169.2 (NdCO), 162.1, 141.3, 140.2, 130.7, 125.2, 123.6 (Ar C),
112.1 (C5Me5), 67.1 (OCH2), 57.0 (NCH2), 35.8, 34.5 (C(CH3),
31.5, 30.3 (C(CH3), 12.1 (C5Me5). MS (EI): m/z 571 [M+].
General Comments. Air-sensitive procedures were carried out
under an inert atmosphere of argon using a dual-manifold vacuum/
argon line and standard Schlenk techniques or in an MBraun dry
box (<5 ppm of O2/H2O). Solvents were dried by refluxing for 3
days under dinitrogen over the appropriate drying agents (sodium
for toluene; potassium for THF; sodium-potassium alloy for diethyl
ether and pentane; calcium hydride for dichloromethane, acetonitrile,
and 1-hexene) and were degassed before use. Solvents were stored
in glass ampules under argon. All glassware and cannulas were
stored in an oven >373 K. Deuterated solvents were freeze-thaw
degassed and dried by refluxing over potassium (for benzene,
toluene, and THF) or calcium hydride (for dichloromethane,
pyridine, and bromobenzene) for 3 days before trap-to-trap distil-
lation and storage in the drybox. Deuterated chloroform was dried
in the bottle over molecular sieves (4 Å).
Cp*ZrCl3,11 Cp*HfCl3,11 Cp*ZrMe3,11 Cp*HfMe3,15 and
CpTiCl314 were synthesized using literature methods. Unless stated
otherwise, purchased chemical reagents were used as received.
Triethylamine and carbon tetrachloride were dried by refluxing
under dinitrogen over calcium hydride for 3 days. Triphenylphos-
phine was recrystallized from hot hexanes and dried under vacuum.
A sodium hydride dispersion in mineral oil was placed in a Schlenk
vessel under an inert atmosphere and washed three times with
diethyl ether to remove the oil. The solid was subsequently dried
thoroughly and stored in the glovebox. Flash chromatography was
performed with a FlashMaster Personal chromatography system and
a selection of prepacked disposable columns. Thin-layer chroma-
tography was performed using Merck 0.25 mm silica layer foil-
backed plates.
NMR spectra were recorded on Bruker DPX-300, DPX-400, and
AV-400 spectrometers. 1H and 13C spectra were referenced
internally using residual protio solvent resonances relative to
tetramethylsilane (δ 0 ppm). Proton and carbon NMR assignments
were routinely confirmed by 1H-1H (COSY) or 1H-13C (HMQC)
experiments. Infrared spectra were obtained directly using an Avatar
320 FTIR instrument. EI mass spectra were obtained on a VG
Autospec mass spectrometer. Elemental analyses were performed
by Warwick Analytical Services. Low percentage C values were
obtained in the elemental analysis of a few of the organometallic
complexes, despite their apparent purity as judged by 1H NMR (see
the Supporting Information). This is attributed to the partial
formation of metal carbides in the combustion process, as frequently
observed with systems of this type.34
[Cp*ZrL1Cl2]. HL1 (153 mg, 0.56 mmol) and NaH (54 mg, 2.22
mmol) were cooled to -78 °C in a Schlenk vessel and dissolved
in THF (10 cm3). The yellow mixture was warmed to room
temperature and stirred for 15 h. The mixture was filtered onto a
solution of Cp*ZrCl3 (185 mg, 0.56 mmol) in THF (10 cm3) at
-78 °C. The resulting yellow solution was stirred for 24 h, affording
a fine white precipitate. The solvent was removed under vacuum,
and the crude solid was dissolved in toluene (10 cm3) and this
solution filtered. The solvent was removed under vacuum, leaving
an off-white solid, which was crystallized from diethyl ether and
[Cp*ZrL2Cl2]. The procedure was as for Cp*ZrL1Cl2, using
HL2. Yield: 321 mg, 58%. Anal. Calcd for C29H43Cl2NO2Zr: C,
1
58.07; H, 7.23; N, 2.34. Found: C, 57.97; H, 7.24; N, 1.87. H
4
NMR (400 MHz, 298 K, C6D6): δ 8.00 (1H, d, Ar H, JHH ) 3
Hz), 7.71 (1H, d, Ar H, 4JHH ) 3 Hz), 3.49 (1H, d, OCH2, 2JHH
)
2
8 Hz), 3.35 (1H, d, OCH2, JHH ) 8 Hz), 1.90 (15H, s, C5Me5),
1.64 (9H, s, tBu), 1.49 (3H, s, NCMe2), 1.28 (9H, s, tBu), 1.04 (3H,
s, NCMe2) ppm. 13C{1H} NMR (100 MHz, 298 K, C6D6): δ 169.2
(NdCO), 162.0, 141.8, 140.3, 130.6, 126.2, 123.7 (Ar C), 113.3
(C5Me5), 78.7 (OCH2), 71.5 (NCMe2), 35.7, 34.6 (C(CH3)3), 31.5,
30.1 (C(CH3)3), 27.7, 24.9 (NCMe2), 12.6 (C5Me5) ppm. MS (EI):
m/z 598 [M+].
[Cp*ZrL3Cl2]. HL3 (225 mg, 0.90 mmol) and NaH (87 mg, 3.63
mmol) were cooled to -78 °C in a Schlenk vessel and dissolved
in THF (10 cm3). The yellow mixture was warmed to room
temperature and stirred for 15 h. The mixture was filtered onto a
solution of Cp*ZrCl3 (300 mg, 0.90 mmol) in THF (10 cm3) at
-78 °C. The resulting yellow solution was stirred for 24 h, affording
a fine white precipitate. The solvent was removed under vacuum,
and the crude solid was dissolved in dichloromethane (10 cm3) and
the solution filtered. The solvent was removed under vacuum,
leaving an off-white solid, which was washed with diethyl ether,
yielding [Cp*ZrL3Cl2] as a white solid. Yield: 309 mg, 63%. Anal.
Calcd for C24H33Cl2NO3Zr: C, 52.83; H, 6.10; N, 2.57. Found: C,
1
50.85; H, 6.01; N, 2.35. H NMR (400 MHz, 298 K, CD2Cl2): δ
7.12 (2H, s, Ar H), 4.58-4.64 (1H, m, OCH2), 4.22-4.44 (2H, m
OCH2CH2N), 3.96-4.04 (1H, m, CH2N), 3.73 (3H, s, OMe), 1.95
(15H, s, C5Me5), 1.37 (9H, s, tBu) ppm. 13C{1H} NMR (100 MHz,
298 K, CD2Cl2): δ 169.3 (NdCO), 158.6, 152.2, 141.9, 125.8,
122.8 (Ar C), 111.7 (C5Me5), 107.8 (Ar C), 68.2 (OCH2), 57.2
(NCH2), 55.7 (OMe), 35.3 (C(CH3)3), 29.7 (C(CH3)3), 11.9 (C5Me5)
ppm. MS (EI): m/z 545 [M+].
[Cp*ZrL4Cl2]. The procedure was as for Cp*ZrL1Cl2, using
HL4. Yield: 332 mg, 68%. Anal. Calcd for C26H37Cl2NO3Zr: C,
1
54.43; H, 6.53; N, 2.44. Found: C, 54.70; H, 6.59; N, 2.81. H
4
NMR (400 MHz, 298 K, C6D6): δ 7.32 (1H, d, Ar H, JHH ) 8
Hz), 7.31 (1H, d, Ar H, 4JHH ) 8 Hz), 3.48 (1H, d, OCH2, 2JHH
)
8 Hz), 3.40 (3H, s, OMe), 3.34 (1H, d, OCH2, 2JHH ) 8 Hz), 1.91
(15H, s, C5Me5), 1.54 (9H, s, tBu), 1.50 (3H, s, NCMe2), 1.04 (3H,
s, NCMe2) ppm. 13C{1H} NMR (100 MHz, 298 K, CD2Cl2) δ 169.3
(NdCO), 159.2, 152.6, 142.4, 126.2, 122.9 (Ar C), 113.3 (C5Me5),
108.2 (Ar C), 78.7 (OCH2), 71.6 (NCMe2), 55.2 (OMe), 35.6
(C(CH3)3), 29.9 (C(CH3)3), 27.6, 24.9 (NCMe2), 12.6 (C5Me5) ppm.
MS (EI): m/z 571 [M+].
[Cp*HfL1Cl2]. The procedure was as for [Cp*ZrL1Cl2],
using Cp*HfCl3. Yield: 298 mg, 53%. Anal. Calcd for
C27H39Cl2NO2Hf: C, 49.21; H, 5.97; N, 2.13. Found: C, 48.77;
H, 5.93; N, 2.08. 1H NMR (400 MHz, 298 K, d8-toluene): δ 7.74
(33) (a) Pennington, D. A.; Clegg, W.; Coles, S. J.; Harrington, R. W.;
Hursthouse, M. B.; Hughes, D. L.; Light, M. E.; Schormann, M.; Bochmann,
M.; Lancaster, S. J. Dalton Trans. 2005, 561. (b) Pennington, D. A.; Hughes,
D. L.; Bochmann, M.; Lancaster, S. J. Dalton Trans. 2003, 3480. (c) Makio,
H.; Fujita, T. Macromol. Symp. 2004, 213, 221. (d) Bryliakov, K. P.;
Kravtsov, E. A.; Pennington, D. A.; Lancaster, S. J.; Bochmann, M.;
Brintzinger, H. H.; Talsi, E. P. Organometallics 2005, 24, 5660.
(34) Carpenetti, D. W.; Kloppenburg, L.; Kupec, J. T.; Petersen, J. L.
Organometallics 1996, 15, 1572.
4
4
(1H, d, JHH ) 3 Hz, Ar H), 7.66 (1H, d, JHH ) 3 Hz, Ar H),
3.60-3.70 (1H, m, CH2), 3.28-3.50 (3H, m, CH2CH2), 1.88 (15H,
s, Cp*), 1.57 (9H, s, C(CH3)3), 1.22 (9H, s, C(CH3)3). 13C{1H}
NMR (100 MHz, 298 K, d8-toluene): δ 169.3 (NdCO), 162.3,
141.1, 140.8, 131.0, 123.4, 123.0, (Ar C), 111.8 (C5Me5), 67.3
(OCH2), 56.8 (NCH2), 35.7, 34.5 (C(CH3), 31.5, 30.3 (C(CH3)),
11.8 (C5Me5). MS (EI): m/z 659 [M+].

6028 Organometallics, Vol. 25, No. 26, 2006
Coles et al.
[Cp*HfL2Cl2]. The procedure was as for [Cp*ZrL1Cl2], using
Cp*HfCl3 and HL2. Yield: 305 mg, 58%. Anal. Calcd for
C29H43Cl2NO2Hf‚0.1Et2O: C, 50.85; H, 6.39; N, 2.02. Found: C,
(9H, s, C(Me)3), 0.25 (3H, s, HfMe), -0.21 (3H, s, HfMe).
13C{1H} NMR (100 MHz, 298 K, d8-toluene): δ 168.4 (NdCO),
163.4, 140.2, 139.2, 130.0, 123.4, 118.1 (Ar C), 111.3 (C5Me5),
66.7 (OCH2), 55.1 (NCH2), 45.2, 43.8 (HfMe), 35.7, 34.5 (C(Me)3),
31.6, 30.2 (C(Me)3), 11.1 (C5Me5). MS (EI): m/z 604 [M+].
1
49.87; H, 6.09; N, 1.92. H NMR (400 MHz, 298 K, CD2Cl2): δ
7.66 (1H, d, Ar H, 4JHH ) 3 Hz), 7.54 (1H, d, Ar H, 4JHH ) 3 Hz),
2
2
1
4.27 (1H, d, OCH2, JHH ) 8 Hz), 3.99 (1H, d, OCH2, JHH ) 8
Hz), 1.92 (15H, s, C5Me5), 1.75 (3H, s, NCMe2), 1.38 (9H, s, tBu),
[Cp*HfL2
Me2]. The procedure was as for [Cp*ZrL Me2], using
HL2 and Cp*HfMe3. Yield: 222 mg, 59%. Anal. Calcd for
1.35 (3H, s, NCMe2), 1.26 (9H, s, Bu) ppm. 13C{1H} NMR (100
C31H49NO2Hf: C, 57.62; H, 7.64; N, 2.17. Found: C, 56.75; H,
t
1
MHz, 298 K, CD2Cl2): δ 169.2 (NdCO), 142.0, 140.0, 132.4,
131.1, 124.4, 123.9 (Ar C), 112.9 (C5Me5), 79.6 (OCH2), 72.0
(NCMe2), 35.4, 34.7 (C(CH3)3), 31.5, 29.9 (C(CH3)3), 28.1, 25.1
(NCMe2), 12.3 (C5Me5) ppm. MS (EI): m/z 687 [M+].
7.56; N, 2.30. H NMR (400 MHz, 298 K, d8-toluene): δ 8.02
(1H, d, 4JHH ) 3 Hz, Ar H), 7.70 (1H, d, 4JHH ) 3 Hz, Ar H), 3.52
2
2
(2H, d, JHH ) 8 Hz, OCH2), 3.36 (2H, d, JHH ) 8 Hz, OCH2),
1.88 (15H, s, Cp*), 1.57 (9H, s, C(Me)3), 1.32 (9H, s, C(Me)3),
1.24 (3H, s, NCMe), 0.96 (3H, s, NCMe), 0.43 (3H, s, HfMe), 0.14
(3H, s, HfMe). 13C{1H} NMR (100 MHz, 298 K, d8-toluene): δ
162.6 (NdCO), 139.9, 139.6, 129.1, 123.8, 119.0, 112.6 (Ar C),
78.6 (OCH2), 70.2 (NCMe2), 48.8, 46.3 (HfMe), 35.6, 34.5
(C(Me)3), 31.6, 30.1 (C(Me)3), 27.8, 25.1 (NCMe), 11.7 (Cp*). MS
(EI): m/z 632 [M+ - Me].
[CpTiL1Cl2]. HL1 (300 mg, 1.09 mmol) and NaH (105 mg, 4.38
mmol) were cooled to -78 °C in a Schlenk vessel, and dissolved
in THF (10 cm3). The yellow mixture was warmed to room
temperature, and left stirring for 15 h. The mixture was filtered
onto a solution of CpTiCl3 (239 mg, 1.09 mmol) in THF (10 cm3)
at -78 °C. The resulting red solution was stirred for 24 h, affording
a fine white precipitate. The solvent was removed under vacuum,
and the crude solid was dissolved in toluene (10 cm3) and filtered.
The solvent was removed under vacuum, leaving a dark red solid,
which was purified by recrystallization from dichloromethane.
Yield: 217 mg, 43%. Anal. Calcd for C22H29Cl2NO2Ti: C, 57.66;
i
Typical Polymerization. A solution of Bu3Al in toluene (0.1
M, 6 cm3) was added to toluene (100 cm3) under an atmosphere of
argon. The vessel was evacuated and heated to the desired
temperature. A constant pressure supply of ethylene (1.2 bar) was
introduced, and the system was allowed to equilibrate. During this
time, a catalyst solution was prepared by dissolving the precatalyst
(1-6 mg) and [PhNMe2H]+[B(C6F5)4]- (9 mg) in toluene (5 cm3).
The catalyst solution was injected, and the gas flow and temperature
((0.5 °C) were monitored. After 1 h, the ethylene atmosphere was
then replaced with argon and 5% HCl in methanol (2 × 10 cm3)
was then added to the slurry. The slurry was then poured into
methanol (500 cm3) with stirring, and the polymer was allowed to
precipitate for approximately 1 h. The polymer was isolated by
filtration and washed with 5% HCl in methanol (50 cm3) and
methanol (2 × 50 cm3). The polymer was then dried in a vacuum
oven at 60 °C until constant weight was achieved.
1
H, 6.38; N, 3.06. Found: C, 57.33; H, 6.32; N, 3.06. H NMR
4
(400 MHz, 298 K, CD2Cl2): δ 7.69 (1H, d, Ar H, JHH ) 3 Hz),
4
7.61 (1H, d, Ar H, JHH ) 3 Hz), 6.50 (5H, s, C5H5), 4.54-4.60
(3H, m, OCH2CH2N), 4.10-4.15 (1H, m, CH2N), 1.39 (9H, s, tBu),
1.28 (9H, s, tBu) ppm. 13C{1H} NMR (100 MHz, 298 K, CD2Cl2):
δ 167.2 (NdCO), 165.7, 144.4, 138.0, 130.9, 123.1, 122.1 (Ar C),
112.4 (C5H5), 68.5 (OCH2), 59.2 (NCMe2), 35.5, 35.0 (C(CH3)3),
31.5, 29.9 (C(CH3)3) ppm. MS (EI): m/z 458 [M+].
[Cp*ZrL1Me2]. HL1 (280 mg, 1.02 mmol) and Cp*ZrMe3 (276
mg, 1.02 mmol) were charged into a Schlenk vessel and dissolved
in toluene (10 cm3). Gas was observed almost instantaneously, and
the mixture was stirred for 15 h. The solvent was removed, and
the crude off-white solid was purified by recrystallization from
pentane. Yield: 221 mg, 41%. Anal. Calcd. for C29H45NO2Zr : C
65.61; H 8.54; N 2.64. Found: C 64.61; H 8.37; N 2.70. 1H NMR
Crystallography. Crystals of [Cp*ZrL1Cl2], [Cp*ZrL2Cl2], and
[Cp*HfL1Cl2] were obtained as colorless blocks by cooling a
concentrated solution in diethyl ether to -30 °C. Crystals of
[Cp*ZrL1Me2] were similarly obtained from a concentrated solution
in pentane. Crystals were coated in an inert oil prior to transfer to
a cold nitrogen gas stream on a Bruker-AXS SMART three-circle
CCD area detector diffractometer system equipped with Mo KR
radiation (λ ) 0.710 73 Å). Data were collected using narrow (0.3°
in ω) frame exposures. Intensities were corrected semiempirically
for absorption, on the basis of symmetry-equivalent and repeated
reflections (SADABS). The structures were solved by direct
methods (SHELXS) with additional light atoms found by Fourier
methods. All non-hydrogen atoms were refined anisotropically. All
H atoms were constrained with a riding model; U(H) was set at
1.2 (1.5 for methyl hydrogen atoms as applicable) times Ueq for
the parent atom. Programs used were Bruker AXS SMART
(control), SAINT (integration), and SHELXTL for structure solu-
tion, refinement, and molecular graphics.
The crystal structures of [Cp*ZrL1Cl2] and [Cp*HfL1Cl2] are
isomorphous and contain diffuse electron density modeled as a
molecule of diethyl ether from the crystallization. This was found
to reside in a position of higher symmetry than is possible for diethyl
ether; therefore, it was modeled as two molecules at 25% occupancy
related by inversion symmetry at the oxygen atom (O103). No
hydrogens were added for the disordered diethyl ether molecules
but were included in the formula for calculation of the density.
The crystal chosen of [Cp*ZrL2Cl2] was chiral with a small twin
component BASF refined to 0.079(72).
4
(400 MHz, 298 K, d8-toluene): δ 7.89 (1H, d, JHH ) 3 Hz, Ar
4
H), 7.66 (1H, d, JHH ) 3 Hz, Ar H), 3.51-3.57 (1H, m, OCH2),
3.36-3.44 (1H, m, OCH2), 3.24-3.32 (1H, m, NCH2), 3.07-3.14
(1H, m, NCH2), 1.77 (15H, s, Cp*), 1.54 (9H, s, C(Me)3), 1.27
(9H, s, C(Me)3), 0.41 (3H, s, ZrMe), 0.03 (3H, s, ZrMe). 13C{1H}
NMR (100 MHz, 298 K, d8-toluene): δ 167.4 (NdCO), 162.0,
138.4, 136.1, 128.5, 123.7, 117.7 (Ar C), 110.2 (C5Me5), 65.1
(OCH2), 53.9 (NCH2), 39.4, 37.5 (ZrMe), 34.3, 33.1 (C(Me)3), 30.3,
28.8 (C(Me)3), 9.9 (C5Me5). MS (EI): m/z 514 [M+ - Me].
[Cp*ZrL2Me2]. The procedure was as for [Cp*ZrL1Me2], using
HL2. Yield: 145 mg, 48%. Anal. Calcd for C31H49NO2Zr: C, 66.61;
1
H, 8.84; N, 2.51. Found: C, 66.52; H, 8.99; N, 2.69. H NMR
4
(400 MHz, 298 K, d8-toluene): δ 7.98 (1H, d, JHH ) 3 Hz, Ar
4
2
H), 7.63 (1H, d, JHH ) 3 Hz, Ar H), 3.47 (2H, d, JHH ) 8 Hz,
2
OCH2), 3.28 (2H, d, JHH ) 8 Hz, OCH2), 1.79 (15H, s, Cp*),
1.50 (9H, s, C(Me)3), 1.26 (9H, s, C(Me)3), 1.15 (3H, s, NCMe),
0.92 (3H, s, NCMe), 0.59 (3H, s, ZrMe), 0.25 (3H, s, ZrMe).
13C{1H} NMR (100 MHz, 298 K, d8-toluene): δ 168.5 (NdCO),
162.7, 139.4, 128.8, 124.0, 120.0 (Ar C), 112.6 (C5Me5), 78.3
(OCH2), 70.1 (NCMe2), 45.1, 40.0 (ZrMe), 35.6, 34.5 (C(Me)3),
31.6, 30.1 (C(Me)3), 27.8, 25.2 (NCMe), 11.8 (C5Me5). MS (EI):
m/z 542 [M+ - Me].
[Cp*HfL1Me2]. The procedure was as for [Cp*ZrL1Me2], using
Cp*HfMe3. Yield: 353 mg, 66%. Anal. Calcd for C29H45NO2Hf:
C, 56.35; H, 7.34; N, 2.27. Found: C, 55.99; H, 7.22; N, 1.95. 1H
NMR (400 MHz, 298 K, d8-toluene): δ 7.87 (1H, d, 4JHH ) 3 Hz,
Ar H), 7.67 (1H, d, 4JHH ) 3 Hz, Ar H), 3.50-3.55 (1H, m, OCH2),
3.37-3.44 (1H, m, OCH2), 3.21-3.30 (1H, m, NCH2), 3.10-3.17
(1H, m, NCH2), 1.80 (15H, s, Cp*), 1.55 (9H, s, C(Me)3), 1.26
Computations. All geometry optimizations were performed with
the DMol3 density functional theory (DFT) code35 as implemented
in the MaterialsStudio (versions 3.2 and 4.0) program package of
(35) (a) Delley, B. J. Chem. Phys. 1990, 92, 508. (b) Delley, B. J. Phys.
Chem. 1996, 100, 6107. (c) Delley, B. J. Chem. Phys. 2000, 113, 7756.

Group 4 Salicyloxazoline Catalysts
Organometallics, Vol. 25, No. 26, 2006 6029
Accelrys Inc. DMol3 utilizes a basis set of numeric atomic functions,
which are exact solutions to the Kohn-Sham equations for the
atoms.36 These basis sets are generally more complete than a
comparable set of linearly independent Gaussian functions and have
been demonstrated to have small basis set superposition errors.36
In the present study a polarized split valence basis set, termed the
double numeric polarized (DNP) basis set, has been used, containing
a p-polarization function for H and d-polarization functions for other
atoms. The DNP basis set was chosen because it is equivalent in
quality and size to the Gaussian 6-31G** split-valence double-ú
plus polarization basis set, generally accepted as the standard basis
set in quantum chemistry. All geometry optimizations employed
highly efficient delocalized internal coordinates.37 The use of
delocalized coordinates significantly reduces the number of geom-
etry optimization iterations needed to optimize larger molecules
compared to the use of traditional Cartesian coordinates.
The convergence criteria for these optimizations consisted of
threshold values of 2 × 10-5 hartree, 0.004 hartree Å-1, and 0.005
Å for energy, gradient, and displacement convergence, respectively,
while a self-consistent field (SCF) density convergence threshold
value of 1 × 10-5 was specified.
Preliminary transition state geometries were obtained by the
integrated linear synchronous transit/quadratic synchronous transit
(LST/QST) algorithm39 available in MaterialsStudio. These pre-
liminary structures were then subjected to full TS optimizations
using an eigenvector following algorithm. All transition structure
geometries exhibited only one imaginary frequency in the reaction
coordinate. All calculated reaction energies were derived from the
total electronic energies of the geometries after optimization.
Acknowledgment. PS thanks EPSRC and BP Chemicals/
INEOS for support.
The generalized gradient approximation (GGA) functional by
Becke and Perdew (BP)38 was used for all geometry optimizations.
Supporting Information Available: CIF files for the molecular
structure determinations and structure files for the computed species
in Table 5. This material is available free of charge via the Internet
at http://pubs.acs.org.
(36) (a) Delley, B. In Density Functional Theory: A Tool for Chemistry;
(b) Seminario, J. M., Politzer, P., Eds.; Elsevier: Amsterdam, The
Netherlands, 1995.
(37) Andzelm, J.; King-Smith, R. D.; Fitzgerald, G. Chem. Phys. Lett.
2001, 335, 321.
OM060687H
(38) (a) Becke, A. D. J. Chem. Phys. 1988, 88, 2547; (b) Perdew, J. P.;
Wang, Y. Phys. ReV. B 1992, 45, 13244.
(39) Govind, N.; Petersen, M.; Fitzgerald, G.; King-Smith, D.; Andzelm,
J. Comput. Mater. Sci. 2003, 28, 250.