Solution structures and dynamic properties of chelated d0 metal olefin complexes {η5 η1-C5R4SiMe2 NtBu}Ti(OCMe2CH2CH2 CH=CH2)+</s

Article abstract of DOI:10.1021/ja003209l

To model the Ti - olefin interaction in the putative {η⁵: η¹-C₅R₄SiMe₂N ^tBu}Ti(R′)(olefin)⁺ intermediates in "constrained geometry" Ti-catalyzed olefin polymerization, chelated a

Full text of DOI:10.1021/ja003209l

898
J. Am. Chem. Soc. 2001, 123, 898-909
Solution Structures and Dynamic Properties of Chelated d0 Metal
Olefin Complexes
{η5: η1-C5R4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)+ (R ) H, Me):
Models for the {η5: η1-C5R4SiMe2NtBu}Ti(R′)(olefin)+ Intermediates
in “Constrained Geometry” Catalysts
Jean-Franc¸ois Carpentier,† Vladimir P. Maryin,† Jeffrey Luci,† and Richard F. Jordan*,§
Contribution from the Department of Chemistry, The UniVersity of Iowa, Iowa City, Iowa 52242, and
Department of Chemistry, The UniVersity of Chicago, 5735 S. Ellis AVe., Chicago, IL 60637
ReceiVed August 29, 2000
Abstract: To model the Ti-olefin interaction in the putative {η5: η1-C5R4SiMe2NtBu}Ti(R′)(olefin)+ intermedi-
ates in “constrained geometry” Ti-catalyzed olefin polymerization, chelated alkoxide olefin complexes {η5:
η1-C5R4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)+ have been investigated. The reaction of {η5: η1-C5R4-
SiMe2NtBu}TiMe2 (1a,b; R ) H, Me) with HOCMe2CH2CH2CHdCH2 yields mixtures of {η5-C5R4Si-
Me2NHtBu}TiMe2(OCMe2CH2CH2CHdCH2) (2a,b) and {η5: η1-C5R4SiMe2NtBu}TiMe(OCMe2CH2CH2CHd
CH2) (3a,b). The reaction of 2a/3a and 2b/3b mixtures with B(C6F5)3 yields the chelated olefin complexes
[{η5: η1-C5R4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)][MeB(C6F5)3] (4a,b; 71 and 89% NMR yield). The
reaction of 2b/3b with [Ph3C][B(C6F5)4] yields [{η5: η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)]-
[B(C6F5)4] (5b, 88% NMR yield). NMR studies establish that 4a,b and 5b exist as mixtures of diastereomers
(isomer ratios: 4a/4a′, 62/38; 4b/4b′, 75/25; 5b/5b′, 75/25), which differ in the enantioface of the olefin that
is coordinated. NMR data for these d0 metal olefin complexes show that the olefin coordinates to Ti in an
unsymmetrical fashion primarily through Cterm such that the CdC π bond is polarized with positive charge
buildup on Cint. Dynamic NMR studies show that 4b/4b′ undergoes olefin face exchange by a dissociative
mechanism which is accompanied by fast inversion of configuration at Ti (“O-shift”) in the olefin-dissociated
intermediate. The activation parameters for the conversion of 4b to 4b′ (i.e., 4b/4b′ face exchange) are: ∆Hq
) 17.2(8) kcal/mol; ∆Sq ) 8(1) eu. 4a/4a′ also undergoes olefin face exchange but with a lower barrier (∆Hq
) 12.2(9) kcal/mol; ∆Sq ) -2(3) eu), for the conversion of 4a to 4a′.
Introduction
Scheme 1
Chain growth in insertion polymerization of olefins by early
transition metal catalysts involves generation of coordinatively
unsaturated, d0 LnMR alkyl species, coordination of monomer
to form transient LnM(R)(olefin) alkyl olefin adducts, and
migratory insertion.1 A principal chain transfer reaction, â-H
transfer to metal, generates transient LnM(H)(olefin) hydride
olefin adducts which undergo olefin dissociation or displace-
ment. These reactions are illustrated in simplified form in
Scheme 1. Understanding the properties of d0 metal olefin
complexes and in particular how the metal-olefin bonding is
influenced by the steric and electronic characteristics of the LnM
unit, may provide insights that are useful for the design and
application of new catalysts. However, d0 metal olefin com-
plexes are rare and little is known about their structures, bonding
or dynamic properties. Olefins coordinate weakly to d0 metal
centers due to the absence of conventional d-π* back-bonding.2
We have developed a simple strategy for the generation of
d0 metal olefin complexes that is based on the use of the flexible
alkoxide-olefin ligand -OCMe2CH2CH2CHdCH2.3 Alkyl
abstraction from LnM(R)(OCMe2CH2CH2CHdCH2) complexes
generates cationic LnM(OCMe2CH2CH2CHdCH2)+ species
which adopt chelated M-olefin-bonded structures if weakly
* Address correspondence to this author at The University of Chicago.
† The University of Iowa.
§ The University of Chicago.
(1) For reviews concerning olefin polymerization by well-defined early
metal catalysts see: (a) Jordan, R. F. AdV. Organomet. Chem. 1991, 32,
325. (b) Brintzinger, H. H.; Fischer, D.; Mu¨lhaupt, R.; Rieger, B.;
Waymouth, R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143. (c)
Bochmann, M. J. Chem. Soc., Dalton Trans. 1996, 255. (d) Kaminsky, W.;
Arndt, M. AdV. Polym. Sci. 1997, 127, 143. (e) Resconi, L.; Cavallo, L.;
Fait, A.; Piemontesi, F. Chem. ReV. 2000, 100, 1253. (f) Chen, E. Y.; Marks,
T. J. Chem. ReV. 2000, 100, 1391.
(2) For a review on metal-olefin bonding, see: Mingos, D. M. P. In
ComprehensiVe Organometallic Chemistry, 1st ed.; Wilkinson, G., Stone,
F. G. A., Abel, E. W., Eds.; Pergamon: New York, 1982; Vol. 3, p 1.
(3) (a) Carpentier, J.-F.; Wu, Z.; Lee, C. W.; Stro¨mberg, S.; Christopher,
J. N.; Jordan, R. F. J. Am. Chem. Soc. 2000, 122, 7750. (b) Wu, Z.; Jordan,
R. F.; Petersen, J. L. J. Am. Chem. Soc. 1995, 117, 5867.
10.1021/ja003209l CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/09/2001

Properties of d0 Metal Olefin Complexes
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 899
Chart 1
participation in the olefin dissociation, or steric inhibition of
chelate ring opening between the two cases. In contrast, cation
IV, which contains the ancillary N4-donor ligand octameth-
yldibenzotetraazaannulene (Me8taa2-), does not adopt a chelated
structure.4 The absence of olefin coordination in IV may reflect
the lower Lewis acidity and harder character of the (Me8taa)-
Zr(OR)+ cation as compared to (C5R5)2Zr(OR)+ cations.
Several other chelated olefin complexes of d0 bent metal-
locenes have been described.5 Royo generated V by benzyl
abstraction from Cp(η5-C5H4SiMe2CH2CHdCH2)Zr(CH2Ph)2,
and Casey prepared VI by reaction of the metallacyclobutane
complex Cp* Zr{CH2CH(CH2CHdCH2)CH2} (Cp* ) C5Me5)
2
with B(C6F5)3 (Chart 1).6,7 NMR data for V and VI are
consistent with unsymmetrical Zr-olefin bonding similar to that
in I and II, and dynamic NMR studies establish that the free
energy barriers to olefin face exchange for V (11.7 kcal/mol,
includes inversion of configuration at Zr) and VI (10.5 kcal/
mol) are similar to that observed for I. Neutral Cp* YCH2CH2-
2
CRR′CHdCH2 (R, R′ ) H, Me) species exhibit weak unsym-
metrical bonding of the pendant olefin.8 Simple, nonchelated
d0 metal olefin complexes are extremely rare, being limited to
the WVI cycloheptene adduct VII, which is stable below ca.
-25 °C, and the VV ethylene and propylene complexes VIII,
which have been characterized by NMR.9,10
An important current goal in this area is to determine if the
properties of nonmetallocene d0 olefin complexes are similar
to those of the metallocene species described above. Here we
report studies of {η5:η1-C5R4SiMe2NtBu}Ti(OCMe2CH2CH2-
CHdCH2)+ species (IX), which are models for the putative {η5:
η1-C5R4SiMe2NtBu}Ti(R′)(olefin)+ intermediates (X) in the
recently developed “constrained geometry” catalyst systems.11-13
The constrained geometry catalysts have gained technological
importance because of their utility in the synthesis of new
copolymers of ethylene with R-olefins, styrene, and even
isobutylene, and their stability at high polymerization temper-
atures. A 12-electron {C5R4SiMe2NR}TiX+ species is more
(4) Martin, A.; Uhrhammer, R.; Gardner, T. G.; Jordan, R. F.; Rogers,
R. D. Organometallics 1998, 17, 382
(5) (a) Horton, A. D.; Orpen, A. G. Organometallics 1992, 11, 8. (b)
Ahlers, W.; Erker, G.; Fro¨hlich, R. Eur. J. Inorg. Chem. 1998, 889. (c)
Karl, J.; Erker, G. J. Mol. Catal. A: Chem. 1998, 128, 858. (d) Temme,
B.; Karl, J.; Erker, G. Chem. Eur. J. 1996, 2, 919. (e) Ruwwe, J.; Erker,
G.; Fro¨hlich, R. Angew. Chem., Int. Ed. Engl. 1996, 35, 80. (f) Erker, G.;
Noe, R.; Kru¨ger, C.; Werner, R. Organometallics 1992, 11, 4174.
(6) Galakhov, M. V.; Heinz, G.; Royo, P. J. Chem. Soc., Chem. Commun.
1998, 17.
coordinating anions are used and the metal center is sufficiently
Lewis acidic to bind the olefin. In an initial study we
investigated the metallocene complexes Cp2Zr(OCMe2CH2CH2-
CHdCH2)+ (I, Cp ) C5H5) and rac-(EBI)Zr(OCMe2CH2CH2-
CHdCH2)+ (II, EBI ) ethylene-1,2-bis(1-indenyl)), which are
models for the corresponding (C5R5)2Zr(R)(R-olefin)+ active
species in metallocene catalysis (III, Chart 1). Complexes I and
II adopt chelated structures in the solid state and in chlorocarbon
solution. The Zr-olefin bonding in these species is unsym-
metrical and consists of a weak Zr-Cterm interaction and
minimal Zr-Cint interaction. The Zr-olefin bonding does not
perturb the structure of the coordinated olefin unit significantly
but does polarize the CdC bond such that positive charge is
delocalized from Zr to Cint. Similar unsymmetrical bonding and
polarization effects may contribute to the high insertion reactiv-
ity of nonchelated (C5R5)2Zr(R)(olefin)+ species. Dynamic NMR
studies show that I and II undergo olefin face exchange (i.e.,
exchange of the olefin enantioface that is coordinated to Zr) in
solution. The free energy barrier at 298 K for face exchange
(7) Casey, C. P.; Carpenetti, D. W.; Sakuri, H. J. Am. Chem. Soc. 1999,
121, 9483.
(8) (a) Casey, C. P.; Hallenbeck, S. L.; Pollock, D. W.; Landis, C. R. J.
Am. Chem. Soc. 1995, 117, 9770. (b) Casey, C. P.; Hallenbeck. S. L.;
Wright, J. M.; Landis, C. R. J. Am. Chem. Soc. 1997, 119, 9681. (c) Casey,
C. P.; Fagan, M. A.; Hallenbeck, S. L. Organometallics 1998, 17, 287.
(9) Kress, J.; Osborn, J. A. Angew. Chem., Int. Ed. Engl. 1992, 31, 1585.
(10) (a) Witte, P. T.; Meetsma, A.; Hessen, B. J. Am. Chem. Soc. 1997,
119, 10561. (b) Several cationic Nb(V) olefin complexes were recently
characterized by NMR. Humphries, M. J.; Douthwaite, R. E.; Green, M.
L. H J. Chem. Soc., Dalton Trans. 2000, 2952.
(11) (a) Stevens, J. C.; Timmers, F. J.; Rosen, G.; Knight, G. W.; Lai,
S. Y. (Dow Chemical Co.). Eur. Pat. Appl. EP 0416815 A2, 1991. (b)
Canich, J. A. (Exxon Chemical Co.) European Patent Application, EP
0420436 A1, 1991. (c) Stevens, J. C. In Studies in Surface Science and
Catalysis; Soga, K., Terano, M., Eds.; Elsevier: Amsterdam, 1994; Vol.
89, pp 277-284. (d) Stevens, J. C. In Studies in Surface Science and
Catalysis; Hightower, J. W., Delglass, W. N., Iglesia, E., Bell, A. T., Eds.;
Elsevier: Amsterdam, 1996; Vol. 101, pp 11-20. (e) Devore, D. D.;
Timmers, F. J.; Hasha, D. L.; Marks, T. J.; Deck, P. A.; Stern, C. L.
Organometallics 1995, 14, 3132.
for II (∆Gq ) 15.3(7) kcal/mol) is significantly greater than
FE
that for I (∆GqFE ) 11.1(8) kcal/mol). The face exchange of II
involves rate-limiting olefin dissociation and fast inversion of
configuration at Zr (“O-shift”) in the olefin-dissociated inter-
mediate. The difference in the face-exchange barriers of I and
II may reflect differences in Zr-olefin bond strengths, solvent
(12) For the initial design of this ligand system, see: (a) Shapiro, P. J.;
Bunel, E.; Schaefer, W. P.; Bercaw, J. E. Organometallics 1990, 9, 867.
(b) Piers, W. E.; Shapiro, P. J.; Bunel, E.; Bercaw, J. E. Synlett 1990, 2,
74. (c) Shapiro, P. J.; Cotter, W. D.; Schaefer, W. P.; Labinger, J. A.;
Bercaw, J. E. J. Am. Chem. Soc. 1994, 116, 4623.

900 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
Carpentier et al.
Scheme 2
t
sterically open, is more electronically unsaturated, and has harder
character than a 14-electron (C5R5)2ZrX+ species. The objectives
of this work were to generate cationic species of type IX,
determine if they adopt chelated structures, and compare the
metal-olefin bonding and dynamic properties to those of
zirconocene systems I and II.
those of 1a,b. In contrast, the Bu NMR resonances for 3a,b
are similar to those of 1a,b. The H and 13C NMR spectra of
3a contain two Si-Me and four C5H4 resonances, and the
spectra of 3b contain two Si-Me and four C5Me4 resonances,
consistent with C1-symmetric structures.
The 2a/3a and 2b/3b product ratios are not affected by the
presence of excess 2-methyl-5-hexene-2-ol. Mixtures of 2a/3a
and 2b/3b were obtained as yellow-brown oils after evaporation
of the solvent. These compounds are highly soluble in hexane
(even at -78 °C), benzene, and chlorinated solvents and could
not be separated by crystallization. The mixtures were used
directly for generation of cationic species.
1
Results
Neutral Ti(OCMe2CH2CH2CHdCH2) Complexes. As shown
in Scheme 2, the reaction of {η5: η1-C5R4SiMe2NtBu}TiMe2
(1a,b; R ) H, Me) with 1 equiv of 2-methyl-5-hexene-2-ol in
benzene (23 °C, 5 min) proceeds by competitive Ti-N and
Ti-C bond alcoholysis to yield the monocyclopentadienyl
alkoxide complexes 2a,b as the major products (94 and 89%
NMR yield, respectively, versus an internal standard) and the
ansa-cyclopentadienyl-amido alkoxide complexes 3a,b as the
only other NMR-observable organometallic products (6 and 11%
{η5: η1-C5R4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)+
Species. Direct addition of solid B(C6F5)3 to neat 2a/3a and
2b/3b mixtures, followed by addition of C6D6 by vacuum
transfer, affords chelated olefin complexes 4a/4a′ and 4b/4b′
in 71 and 89% NMR yield, respectively, vs. an internal standard
(Scheme 2).15 The reaction of the 2b/3b mixture with [Ph3C]-
[B(C6F5)4] under similar conditions generates the corresponding
B(C6F5)4- salt 5b/5b′ in 88% NMR yield along with Ph3CCH3.15
The formation of the chelated olefin complexes from 2a,b
proceeds by methyl abstraction by B(C6F5)3 or Ph3C+ and
subsequent Ti-Me protonolysis by the NH group, which
eliminates CH4 (detected by 1H NMR, δ 0.2). Compounds 4a/
4a′, 4b/4b′ and 5b/5b′ separate as oils from aromatic and
aliphatic hydrocarbon solvents. Washing the oils with hexane
and drying under high vacuum yields 4a/4a′, 4b/4b′ and 5b/
5b′ in 70-90% purity as red-orange gummy solids. The isolated
materials are readily soluble in dichloromethane and 1,2-
dichloroethane. These solutions are quite stable and can be
stored for several weeks at -30 °C without noticeable reaction.
However, attempts to further purify 4a/4a′, 4b/4b′, or 5b/5b′
were unsuccessful.
1
vs. internal standard, respectively).13l,m The H NMR spectra
of 2a,b contain NH resonances (δ 0.76, 0.57; CD2Cl2) that are
close to those for the corresponding free amines (C5R4H)SiMe2-
NHtBu.14 Additionally, the tBu resonances for 2a,b (δ 1.09, 1.09;
C6D6) are shifted upfield compared to the corresponding
resonances for 1a,b (δ 1.52, 1.57) and are close to the free amine
resonances. A similar trend is observed for the -CMe3 13C NMR
resonances of 2a,b which are shifted ∼10 ppm upfield from
(13) (a) Chen, Y.-X.; Marks, T. J. Organometallics 1997, 16, 3649. (b)
Soga, K.; Uozomi, T.; Nakamura, S.; Toneri, T.; Teranishi, T.; Sano, T.;
Arai, T. Macromol. Chem. Phys. 1996, 197, 4237. (c) Sernetz, F. G.;
Mu¨lhaupt, R.; Waymouth, R. M. Macromol. Chem. Phys. 1996, 197, 1071.
(d) Shiomura, T.; Asanuma, T.; Inoue, N. Macromol. Rapid Commun. 1996,
17, 9. (e) Eberle, T.; Spaniol, T. P.; Okuda, J. Eur. J. Inorg. Chem. 1998,
237. (f) Sernetz, F. G.; Mu¨lhaupt, R.; Amor, F.; Eberle, T.; Okuda, J. J.
Polym. Sci., Part A: Polym. Chem. 1997, 35, 1571. (g) Sinnema, P.-J.;
Liekelema, K.; Staal, O. K. B.; Hessen, B.; Teuben, J. H. J. Mol. Catal. A:
Chem. 1998, 128, 143. (h) McKnight, A. L.; Masood, M. A.; Waymouth,
R. M. Organometallics 1997, 16, 2879. (i) McKnight, A. L.; Waymouth,
R. M. Chem. ReV. 1998, 98, 2578. (j) Shaffer, T. D.; Canich, J. A. M.;
Squire, K. R. Macromolecules 1998, 31, 5145. (k) Hultzsch, K. C.; Voth,
P.; Beckerle, K.; Spaniol, T. P.; Okuda, J. Organometallics 2000, 19, 228.
(l) For other reactions in which the {C5R4SiMe2NtBu}M chelate ring in
constrained geometry complexes is cleaved, see: Carpenetti, D. W.;
Kloppenburg, L.; Kupec, J. T.; Petersen, J. L. Organometallics 1996, 15,
1572 and (m) Kloppenburg, L.; Petersen, J. L. Organometallics 1996, 15,
7.
The structures of the cations of 4a/4a′, 4b/4b′, and 5b/5b′ in
CD2Cl2 solution were established by NMR studies. The low-
temperature (-60 °C) 1H and 13C NMR spectra of these species
each contain two complete sets of resonances and establish that
(15) Extensive NMR experiments utilizing an internal standard showed
that 4b/4b′ and 5b/5b′ are formed in 89% yield along with 11% of a ∼1/1
mixture of two diamagnetic cationic Ti compounds with free vinyl groups.
These species could not be unambiguously identified by NMR or attempted
derivatization/trapping experiments. However variable temperature 1D- and
(14) (C5H5)SiMe2NHtBu (5-isomer): 1H NMR (C6D6) δ 6.58 (br, 2H,
C5H4), 6.52 (br, 2H, C5H4), 3.50 (br s, 1H, C5H4H), 1.20 (s, 9H, tBu), 0.62
1
2D-EXSY H NMR experiments established that the unidentified species
1
(br s, H, NH), -0.03 (s, 6H, SiCH3). (C5Me4H)SiMe2NHtBu (5-isomer):
do not influence the dynamic properties of 4b/4b′ and 5b/5b′ and that the
free vinyl groups of the unidentified species do not exchange with those of
4b/4b′ and 5b/5b′.
1H NMR (C6D6) δ 2.75 (br s, 1H, C5Me4H), 2.00 (s, 6H, C5CH3), 1.84 (s,
t
6H, C5CH3), 1.10 (s, 9H, Bu), 0.39 (br s, 1H, NH), 0.10 (s, 6H, SiCH3).

Properties of d0 Metal Olefin Complexes
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 901
the diastereomer ratios 4a/4a′ and 4b/4b′ are 62/38 and 75/25,
-
respectively. The diastereomer ratio of the B(C6F5)4 salt 5b/
5b′ is the same as that for the corresponding MeB(C6F5)3- salt
4b/4b′ as expected.
The NMR data for 4a/4a′, 4b/4b′, and 5b/5b′ show that in
each case the pendant olefin is coordinated to Ti. The vinyl 1H
resonances for each species were assigned by analysis of the
JH-H coupling constants and confirmed by low-temperature
COSY experiments. The vinyl 1H resonances are shifted
substantially from the free olefin positions; in particular, the
Hint resonance appears at low field (δ 4a: 7.02; 4a′: 7.24; 4b:
6.52; 4b′: 7.30) compared to the corresponding free olefin
resonance in 2a,b (δ 5.90).16 The terminal vinyl 13C resonances
are shifted significantly upfield (δ Cterm 4a: 104.3; 4a′: 98.0;
4b: 106.0; 4b′: 101.6) while the internal vinyl 13C resonances
are shifted significantly downfield (δ Cint 4a: 159.4; 4a′: 163.1;
4b: 162.8; 4b′: 167.8) from the corresponding free olefin
resonances in 2a,b (δ Cterm 114.5; Cint 139.2). However, the
vinyl JC-H values for 4a/4a′ and 4b/4b′ are nearly identical to
those for the free olefin, 2a,b and 3a,b (JC-H ) 154 ( 4 Hz)
which shows that coordination of the olefin to Ti does not
significantly perturb the hybridization of the vinyl carbons. The
Figure 1. Side views (left) and front views (right) of R,S-4b and R,R-
4b′. The curved lines in the side views show observed NOE correlations.
The SiMe2NtBu unit has been omitted for clarity in the front views.
As illustrated in Figure 1, Hcis and Hint of the major diastereomer
4b show NOE correlations to C5Me4 hydrogens (δ 2.18 and
2.22), while only Hcis (but not Hint) of the minor diastereomer
4b′ correlates with a C5Me4 group (δ 2.24). Analysis of the
models shows that these correlations are consistent with the R,S
(ent ) S,R) configuration for 4b and the S,S (ent ) R,R)
configuration for 4b′.19
1
-60 °C H NMR spectrum of 4b/4b′ contains two Si-Me
signals and four C5Me4 signals for each diastereomer in the
expected intensity ratios, consistent with C1-symmetric struc-
tures. Similarly, the -60 °C 1H NMR spectrum of 4a/4a′
contains two Si-Me and four C5H4 resonances for each
diastereomer.
Dynamic Properties of 4b/4b′. Compound 4b/4b′ was
selected for detailed dynamic NMR studies because this species
is obtained in higher purity than 4a/4a′ and the vinyl and Si-
The 1H, 13C, and 19F NMR parameters for the MeB(C6F5)3
-
anion in 4a/4a′ and 4b/4b′ are characteristic of the free anion.17
1
-
The H and 13C NMR spectra of the B(C6F5)4 salt 5b/5b′ are
identical to the spectra of 4b/4b′ except for the anion resonances.
The 19F NMR spectrum of 5b/5b′ confirms the presence of a
1
Me regions of the H NMR spectra are relatively simple. To
simplify the nomenclature in the following discussion, the
diastereomers 4b and 4b′ will be denoted as A and B
respectively. As shown in Figure 2, the pairs of Hint, Htrans, and
Hcis 1H NMR resonances each collapse to a single resonance as
the temperature is raised from -60 to 60 °C. The A/B (i.e.,
4b/4b′) ratio is constant from -60 to 20 °C (75/25 ( 2%), and
the coalesced vinyl resonances at 60 °C appear at the weighted
averages of the individual low-temperature (-60 °C) chemical
shifts.20 These observations establish that the A/B ratio does
not change significantly between -60 and 60 °C. Additionally,
as shown in Figure 3, the four Si-Me resonances (two for each
diastereomer) broaden, coalesce, and sharpen to one singlet
between -60 and 60 °C. The eight C5Me4 resonances (four R,
four â; vs. SiMe2NtBu bridge) also broaden and coalesce to
two singlets (R, â) over this temperature range. These line-shape
changes are all reversible upon lowering the temperature.
These observations imply that 4b/4b′ undergo two exchange
processes which are illustrated in Scheme 3: (a) interconversion
of 4b/4b′, that is, “olefin face exchange”, which permutes the
corresponding vinyl hydrogens, Si-Me and C5Me4 groups of
the two diastereomers (horizontal process in Scheme 3), and
(b) inversion of configuration at Ti, which permutes the two
Si-Me groups, the two R-C5Me4 groups and the two â-C5Me4
groups of each diastereomer. This second process presumably
involves migration of the η1-alkoxide ligand between lateral sites
on Ti (i.e., “O-shift”, vertical process in Scheme 3) in the olefin-
-
free B(C6F5)4 anion.18
Addition of THF to CD2Cl2 solutions of 4a/4a′ and 5b/5b′
results in the formation of the THF complexes [{η5: η1-
C5H4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)(THF)][MeB-
(C6F5)3] (6a) and [{η5: η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2-
CH2CHdCH2)(THF)][B(C6F5)4] (6b). In both cases the vinyl
1H and 13C NMR resonances shift back to the free olefin
positions, but the anion resonances are unchanged. These results
confirm that the pendant olefin and not the counterion is
coordinated to Ti in 4a/4a′ and 5b/5b′.
The diastereomeric cations of 4a/4a′, 4b/4b′, and 5b/5b′ differ
in the enantioface of the olefin that is coordinated and therefore
in the relative configurations of the Ti center and the internal
vinyl carbon (Figure 1). The structures of 4b and 4b′ were
1
assigned from a low-temperature (-40 °C) H 2D NOESY
spectrum. For this purpose, models of 4b and 4b′ were generated
by molecular modeling (see Experimental Section). The stere-
ochemistry of these species is denoted by the descriptors R,S
(ent ) S,R) and S,S (ent ) R,R), in which first entry denotes
the configuration at Ti and the second entry denotes that at Cint.
(16) Htrans and Hcis are trans and cis to Hint, respectively.
-
(17) The MeB(C6F5)3 NMR resonances of 4a/4a′, 4b/4b′ and 6a are
nearly identical to those of [NBu3CH2Ph][MeB(C6F5)3] (see ref 3a): 1H
NMR (CD2Cl2, -60 °C) δ 0.33; (-20 °C) δ 0.41; (20 °C) δ 0.47 (br s,
3H, BCH3). 13C NMR (CD2Cl2, -60 °C) δ 147.3 (dm, JC-F ∼ 238), 136.8
(dm, JC-F ≈ 240, Cpara), 135.7 (dm, JC-F ≈ 243), 127.2 (br s, Cipso), 9.1
(br q, JC-H ) 119, CH3B). 19F NMR (CD2Cl2) δ -131.7 (d, 3JF-F ) 19.9,
(19) H-H contacts estimated by molecular modeling are as follows.
4b: Hint-C5Me4 ) 2.7 Å; Hcis-C5Me4 ) 2.4 Å; 4b′: Hint-C5Me4 > 4.8
Å; Hcis-C5Me4 ) 2.4 Å. NOE correlations between Htrans and the C5Me4
groups were not observed for either diastereomer, despite the close contact
(Htrans-C5Me4 ) 2.5 Å) predicted for the R,R isomer. Close Htrans-C5Me4
contacts are not present in the S,R isomer (Htrans-C5Me4 ) 3.9 Å).
(20) The chemical shifts (δobserved, δcalculated) for the vinyl hydrogen
resonances at coalescence are as follows: Hint: 6.75, 6.76; Htrans: 4.90,
4.92; Hcis: 4.39, 4.37.
3
3
2F, o-F), -163.5 (t, JF-F ) 19.8, 1F, p-F), -166.1 (t, JF-F ) 19, 2F,
m-F). 11B NMR (CD2Cl2) δ -13.3.
(18) B(C6F5)4 resonances of 5b/5b′: 13C NMR (CD2Cl2, -60 °C) δ
-
146.0 (dm, JC-F ≈ 238), 137.1 (dm, JC-F ≈ 240, Cpara), 135.4 (dm, JC-F
≈
243), 122.9 (br s, Cipso). 19F NMR (CD2Cl2) δ -132.8 (d, 3JF-F ) 21, 8F,
o-F), -163.5 (t, 3JF-F ) 20, 4F, p-F), -167.4 (br t, 3JF-F ) 19, 8F, m-F).
11B NMR (CD2Cl2) δ -13.5.

902 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
Carpentier et al.
1
Figure 2. Experimental (left) and simulated (right) H NMR spectra (vinyl region) of [{η5: η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)]-
[MeB(C6F5)3] (4b/4b′ ) A/B). Best-fit first-order rate constants (kvinylAB) are shown with the simulated spectra. The spectra at -20 °C and -60 °C
(not shown) are nearly identical. Assignments (4b/4b′): Hint: 6.52/7.30; Htrans: 4.99/4.56; Hcis: 4.45/4.01. * ) impurities (see ref 15). S ) solvent
(-20 to 20 °C: CD2Cl2; 40 and 60 °C: CD2ClCD2Cl).
Scheme 3
AB
dissociated intermediate. The O-shift cannot occur if the olefin
remains coordinated.21
diastereomer (A) to the minor diastereomer (B) (∆Hq
)
vinyl
AB
17.2(8) kcal/mol; ∆Sq
) 8(1) eu) were obtained from a
least-squares Eyring analysis (Figure 4) according to eq 1, where
kvinyl is the corresponding first-order rate constant and kB is
vinyl
Olefin Face Exchange of 4b/4b′. The kinetics of the olefin
face-exchange process (i.e., A/B interconversion) were probed
by line-shape analysis of the vinyl region of the 1H NMR spectra
(-20 to 60 °C, Figure 2). Spectra were simulated for an
AB
the Boltzmann constant.
ln(kvinylAB/T) ) -∆Hq AB/(RT) + (∆Sq AB/R) +
ensemble of three two-site exchange systems, each with a 75/
vinyl
vinyl
A
25 relative population ratio, corresponding to HintA/HintB, Htrans
/
ln(kB/h) (1)
HtransB, and HcisA/HcisB. Satisfactory agreement between the
observed and simulated spectra was only obtained when the
exchange rate for each two-site system was equal, which
confirms that the vinyl group undergoes face exchange as a unit.
The activation parameters for the conversion of the major
The face-exchange process was also investigated in the slow-
exchange region (-40 to 0 °C) by two-dimensional exchange
spectroscopy. Representative phase-sensitive H 2D NOESY-
1
EXSY spectra (vinyl region) of A/B (4b/4b′) in CD2Cl2 are
shown in Figure 5. At -40 °C a nearly pure NOESY spectrum
is obtained, as cross-peaks arise almost exclusively between
signals of the same diastereomer. Additionally however, low-
(21) Four TiOCMe2- signals are observed at -60 °C for 4b/4b′ (two
for each diastereomer). These signals coalesce to one singlet at 60 °C, which
is consistent with the existence of the two exchange processes noted in the
text.

Properties of d0 Metal Olefin Complexes
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 903
1
Figure 3. Experimental (left) and simulated (right) H NMR spectra
(Si-Me region) of [{η5: η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2CH2CHd
CH2)][MeB(C6F5)3] (4b/4b′ ) A/B). Best-fit first-order rate constants
(kMeAxBy) are shown with the simulated spectra. Assignments: 4b: δ
0.87, 0.79; 4b′: δ 0.91, 0.71.
Figure 5. 360-MHz 1H 2D-NOESY/EXSY spectra (vinyl region)
of [{η5: η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)][MeB-
(C6F5)3] (4b/4b′ ) A/B) in CD2Cl2 solution. The spectrum at -40 °C
(top) shows almost exclusively NOESY correlations within 4b and 4b′
(correlations drawn), while the spectrum at -10 °C (bottom) shows
exchange between 4b and 4b′ (exchange networks drawn).
spectra display high-intensity cross-peaks for each pair of vinyl
hydrogens, HtransA/HtransB, HcisA/HcisB, and HintA/HintB, as expected
for the face-exchange process.
The 2D EXSY data were treated quantitatively using the
methods summarized by Perrin and Dwyer.22 Each of the three
sets of diagonal and cross-peak intensities were processed,
assuming a two-site exchange system according to eqs 2 and 3,
in which k′ is the sum of forward and reverse exchange rate
constants in s-1, τm is the mixing time, X denotes the mole
AB
Figure 4. Eyring plot of “olefin face-exchange” rate constants (kvinyl
)
for [{η5: η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)][MeB-
(C6F5)3] (4b/4b′ ) A/B) using line-shape (0) and 2D-EXSY (×) data.
The line corresponds to the overall least-squares fit.
A
B
intensity cross-peaks are detected between Hcis and Hcis
,
(22) (a) Perrin, C. L.; Dwyer, T. J. Chem. ReV. 1990, 90, 935. See also:
(b) Orrell, K. G.; Sik, V. Dynamic NMR Spectroscopy in Inorganic and
Organometallic Chemistry. In Annual Reports in NMR Spectroscopy; Webb,
G. A., Ed.; Academic: New York, 1993; Vol. 27, pp 103-172.
indicating that chemical exchange is not totally frozen at this
temperature. As the temperature is increased, the intensities of
the exchange cross-peaks increase. Between -35 and 0 °C, the

904 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
Carpentier et al.
Scheme 4
Table 1. Olefin Face-Exchange Rate Constants Derived from the
1H EXSY Spectra of 4b/4b′a
10 to 30 °C (intermediate exchange region), but the polarity
and donor properties of these solvents are very similar. 4b/4b′
is not stable in C6D5Cl. The positive ∆Sq value (8(1) eu) is
consistent with simple olefin dissociation with minimal solvent
involvement, although this value should be interpreted with
caution.25
T, °C
τm, s
k′ (Htrans), s-1
k′ (Hcis), s-1
kvinylAB, s-1 b
-40
-35
-30
-20
-10
0
1.0
0.5
0.4
0.3
0.2
0.05
0.02(2)
0.08(2)
0.29(4)
1.23(7)
4.5(3)
0.05(2)
0.11(2)
0.32(3)
1.35(8)
4.3(3)
0.009(5)
0.024(5)
0.076(9)
0.32(2)
1.10(8)
2.6(2)
An alternative possible face-exchange mechanism is a non-
dissociative mechanism involving transit of the Ti center through
the alkene π nodal plane via a C-H σ-complex without
complete olefin dissociation, as illustrated in Scheme 4. Mech-
anisms of this type have been proposed previously for the
interconversion of diastereomeric Re-olefin complexes.26 Analy-
10(1)
10(1)
a See Experimental Section. b Rate constant calculated from k′average
) [k′(Htrans) + k′(Hcis)]/2 according to eq 5.
fractions of the two sites (XA ) 0.75; XB ) 0.25), IAA and IBB
are the diagonal peak intensities, and IAB and IBA are the cross-
peak intensities.
1
sis of the vinyl region of the H NMR spectra of 4b/4b′ does
not allow the dissociative and σ-complex mechanisms to be
distinguished. However, this issue can be addressed by analysis
of the Si-Me exchange, as described in the next section.
Si-Me Exchange. The Si-Me exchange system comprises
four sites which are labeled A1, A2, B1, and B2 in Scheme 3.
A1 and A2 correspond to the diastereotopic Si-Me groups of
the major diastereomer A (4b), and B1 and B2 correspond to
the diastereotopic Si-Me groups of the minor diastereomer B
(4b′). Interconversion of R,S-A and S,R-A permutes A1 and A2,
while interconversion of R,R-B and S,S-B permutes B1 and B2.
Because the diastereomer ratio A/B ) 3/1, the relative popula-
tions (mole fractions) of sites A1, A2, B1, and B2 are 3/8, 3/8,
1/8, and 1/8, respectively.
Assuming that (i) olefin face exchange (i.e., A/B intercon-
version) occurs by olefin dissociation/recoordination and (ii)
the O-shift is much faster than olefin recoordination in the olefin-
dissociated intermediate, that is, kOS . kcoord in Scheme 3, then
exchange of the Si-Me groups should proceed statistically,
according to the relative populations of the different sites. The
existence of four Si-Me sites implies six possible site-to-site
exchanges (A1-A2, A1-B1, A1-B2, A2-B1, A2-B2, B1-
B2). The rate of each site-to-site exchange Rij is given by eq 6
k′ ) (1/τm) × ln[(r + 1)/(r - 1)]
(2)
r ) {4XAXB(IAA + IBB)/(IAB + IBA)} - (XA - XB)2 (3)
Considering that the forward and reverse rates (RAB, RBA) are
AB
equal (eq 4), the rate constant for conversion of A to B (kvinyl
)
is given by eq 5:
R
AB ) XAkvinylAB ) XBkvinylBA ) RBA
kvinylAB ) k′/(1 + XA/XB)
(4)
(5)
The k′ values determined by analysis of the Htrans and Hcis signals
between -40 and 0 °C are listed in Table 1 and agree within
experimental error, as expected for an exchange process in which
the whole vinyl unit is involved.23 For each temperature, the
k′(Htrans) and k′(Hcis) values were averaged and used to determine
kvinylAB. Activation parameters for the face-exchange process
AB
AB
(∆Hq
) 17.5(8) kcal/mol; ∆Sq
) 8(2) eu) were
vinyl
vinyl
AB
obtained by an Eyring analysis of these kvinyl values (Figure
4). These values agree well with the values determined by the
line-shape analysis described above.
Rij ) pij × Pi × k
(6)
Possible Olefin Face-Exchange Mechanisms. The most
likely mechanism for the olefin face exchange is a simple olefin
dissociation/recoordination process (Scheme 3). Associative
mechanisms involving displacement of the olefin by the anion
can be ruled out because the 1H NMR spectra of the MeB(C6F5)3-
salt 4b/4b′ and B(C6F5)4- salt 5b/5b′ are identical in the range
-60-70 °C (except for the anion resonances).24 It is more
difficult to probe the extent to which olefin dissociation is
assisted by the solvent, due to the solubility and reactivity
properties of these compounds. The 1H NMR spectra of 4b/4b′
in CD2Cl2 are identical to those in ClCD2CD2Cl over the range
where pij is the probability of exchanging from site i to j, Pi is
the population of site i, and k is defined in terms of the mean
lifetime τ of all sites by k ) 1/τ.22,27 Also, equilibrium mass
(25) Purely associative mechanisms are generally characterized by ∆Sq
values below -10 eu. However, activation entropies are difficult to
determine precisely and must be interpreted carefully because of possible
contributions from solvent reorganization, especially for polar solvents and
charged metal complexes; see: (a) Atwood, J. D. Inorganic and Organo-
metallic Reaction Mechanisms; Brooks/Cole: Monterey, CA, 1985; p 17.
(b) Jordan, R. B. Reaction Mechanisms of Inorganic and Organometallic
Systems; Oxford University: New York, 1991; pp 56-57.
(26) (a) Peng, T.-S.; Gladysz, J. A. J. Am. Chem. Soc. 1992, 114, 4174.
See also: (b) Kegley, S. E.; Walter, K. A.; Bergstrom, D. T.; MacFarland,
D. K.; Young, B. G.; Rheingold, A. L. Organometallics 1993, 12, 2339.
(c) Quiro´s-Me´ndez, N.; Mayne, C. L.; Gladysz, J. A. Angew. Chem., Int.
Ed. Engl. 1990, 29, 1475.
(27) For a discussion of the relationship between rate constants measured
by dynamic NMR methods and those of the chemical process(es) giving
rise to the exchange, see: Green, M. L. H.; Wong, L. L.; Sella, A.
Organometallics 1992, 11, 2660.
(23) Accurate kinetic data could not be extracted from the Hint signals
due to the high multiplicity and the significant broadening of these signals
with increasing temperature, which precluded determination of accurate
cross-peak volume intensities.
(24) Variable temperature 1H NMR spectra were recorded for 4b/4b′
over a 2-fold concentration range, and no concentration dependence was
noted.

Properties of d0 Metal Olefin Complexes
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 905
Table 2. Comparison of kVinyl and kMe
Valuesa
2kMeAxBy, s-1
AB
AxBy
balance requires that
T, °C
kvinylAB, s-1
Rij ) Rji ) pji × Pj × k
(7)
-5
0
5
2.3
4.4
8.2
2.7
4.8
8
Thus, the rate of exchange of A1 and A2, which corresponds
to interconversion of the two enantiomers of the major diaste-
reomer A, is expected to be 9 times faster (RA1A2 ) 3/8 × 3/8
× k) than the rate of exchange of B1 and B2, which corresponds
to the interconversion of the two enantiomers of the minor
diastereomer B (RB1B2 ) 1/8 × 1/8 × k). The four “cross-
exchanges”, that is, A1-B1, A1-B2, A2-B1, and A2-B2
(Ax-By; x,y ) 1 or 2), which correspond to interconversion of
the major and minor diastereomers, occur at an intermediate
rate (RAxBy ) 3/8 × 1/8 × k). Thus, if assumptions (i) and (ii)
are correct, it is expected that the six possible site-to-site
exchanges A1-A2, A1-B1, A1-B2, A2-B1, A2-B2, B1-
B2 should occur at relative rates of 9/3/3/3/3/1.
10
20
30
40
60
16
46
133
358
2183
16
32
96
360
1600
a Values of kvinyl were calculated using activation parameters
derived from the least-squares Eyring analysis of the combined line-
shape and EXSY data (see Figure 4).
AB
On the other hand, if the O-shift is not fast relative to olefin
recoordination (i.e., kOS e kcoord in Scheme 2), then the four
Si-Me exchanges that require olefin dissociation and O-shift,
that is, A1-A2, A1-B2, A2-B1, and B1-B2, would occur at
slower rates than the A1-B1 and A2-B2 exchanges. This result
would also be observed if a nondissociative pathway such as
the “σ-complex” mechanism (Scheme 4) contributes signifi-
cantly to the face-exchange process.
The Si-Me region was simulated assuming that the six
possible exchanges of the four Si-Me sites (A1-A2, A1-B1,
A1-B2, A2-B1, A2-B2, B1-B2) occur at relative rates of 9/3/
3/3/3/1. The site-to-site exchange rates were optimized to fit
the simulated and experimental spectra of 4b/4b′ over the
temperature range -5 to 60 °C, which includes the intermediate
and fast exchange regions. As shown in Figure 3, a good fit
was obtained using this procedure. Exhaustive simulations
showed that for the spectrum at 10 °C (intermediate exchange
region), which proved to be very sensitive to the simulation
parameters, a satisfactory match of observed and simulated
spectra requires that the relative A1-A2, A1-B1, A1-B2, A2-
B1, A2-B2, and B1-B2 exchange rates be close to 9/3/3/3/
3/1.28 On the basis of these results, we conclude that σ-complex
intermediates do not play a significant role in the face exchange
Figure 6. Eyring plot of “olefin face-exchange” rate constants
(2kMeAxBy) for [{η5: η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)]-
[MeB(C6F5)3] (4b/4b′) determined from simulation of the Si-Me
region.
and (c) the olefin dissociation (kdiss). As is evident from Scheme
3, if assumptions (i) and (ii) above are correct, then because
the isomer ratio A/B ) 3/1, the exchange of a given Si-Me of
A with a given Si-Me of B occurs in one out of eight
dissociation events, that is,
AxBy
of 4b/4b′. The rate constants for the “cross-exchanges” (kMe
A1B1
A1B2
A2B1
) kMe
) kMe
) kMe
) kMeA2B2) obtained by this
kMeAxBy ) kdiss/8
(8)
procedure are given in Figure 3.29
At this stage it is useful to relate the rate constants for (a)
the Si-Me site-to-site exchanges (kMeAxBy), (b) the olefin face
exchange determined from analysis of the vinyl region (kvinylAB),
Similarly, the exchange of each of the three vinyl hydrogens of
A with the corresponding vinyl hydrogens of the B occurs in
one out of four dissociation events, that is,
(28) However, it should be pointed out that at other temperatures, the
spectra are less sensitive to the simulation parameters. For example, for 0
°C, a satisfactory fit between experimental and simulated spectra can also
be obtained by using relative rates for the six Si-Me permutations varying
by up to ∼(40% from the statistical ratio 9/3/3/3/3/1.
kvinylAB ) kdiss/4
Consequently it is expected that
(9)
(29) The results of the simulation were confirmed by line-broadening
analysis in the slow exchange region (-5-5 °C). Under conditions of
dissociative olefin face exchange and fast O-shift (Scheme 3), the rate
constant for exchange of a given Si-Me group of A to any of the other
three Si-Me sites in A or B (kMeA) is related to the site-to-site rate constants
AB
2kMeAxBy ) kvinyl
(10)
AxBy
AB
As shown in Table 2, the values of kMe
and kvinyl are
A
A1A2
A1B1
A1A2
AxBy
+ kMeA1B2. From eq 6, kMe
) 3kMe
,
consistent with eq 10. Accordingly, the values of 2kMeAxBy were
used in an Eyring analysis (Figure 6) to calculate the activation
parameters for the olefin face exchange determined by the Si-
by kMe ) kMe
+ kMe
so that kMeA ) 5kMeAxBy. Similarly, the rate constant for exchange of a given
Si-Me group of B to any of the other three Si-Me sites (kMeB) is given by
B
B1B2
B1A1
B1B2
AxBy
kMe ) kMe
+ kMe
+ kMeB1A2. From eq 6, kMe
) kMe
and
BxAy
B
A
B
) 3kMeAxBy, so that kMe ) 7kMeAxBy. Values for kMe and kMe
AB
AB
Me simulation (∆Hq
) 17.0(7) kcal/mol; ∆Sq
) 7(2)
kMe
Me
Me
were obtained from the excess line widths of the Si-Me resonances in the
slow exchange region (-5-5 °C) using the relation k ) π(∆W), where
∆W ) W - W0, W is the line width at half-height, and W0 is the line
width in the absence of exchange (W0 ) 1.5 Hz, measured at -40 and
-60 °C). The rate constants determined by this analysis agree well with
eu). These values agree well with those determined from analysis
of the vinyl region.
In summary, the following key results emerge from our study
of the dynamic properties of 4b/4b′: (i) the vinyl group
undergoes face exchange as a unit, (ii) the dynamic properties
are not influenced by the counterion, which implies that the
A
the values determined by simulation. kMe (line broadening, simulation;
B
s-1): -5 °C: 6.0, 6.7; 0 °C: 11.2, 12.0; 5 °C: 20.4, 20.0. kMe (line
broadening, simulation): -5 °C: 8.5, 9.3; 0 °C: 16.7, 16.8).

906 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
Carpentier et al.
face exchange is not assisted by the anion, and (iii) the Si-Me
exchange is statistical, and the activation parameters for the vinyl
exchanges and Si-Me exchange are identical within experi-
mental uncertainty. These results are best accommodated by
Scheme 3, in which the rate-limiting step is olefin dissociation
and inversion of configuration at Ti in the olefin-dissociated
intermediate is fast relative to olefin recoordination. Nondis-
sociative olefin face exchange via σ-complex intermediates is
not important in this system. It is possible that the solvent assists
the olefin dissociation and stabilizes the olefin-dissociated
intermediate, but the positive ∆Sq value argues against kineti-
cally significant solvent participation.
Chart 2
Table 3. Activation Parameters for Olefin Face Exchange of
LnM(OCMe2CH2CH2CHdCH2)+ Complexes from Dynamic H
1
NMR Studies
∆Gq at 298 K
kcal/mol
LnM+
∆Hq kcal/mol ∆Sq eu
Cp2Zr+ (I)
9.6(5)a
16.2(4)b
12.2(9)c
-5(2)a
11.1(8)a
15.3(7)b
13(1)c
Dynamic Properties of 4a/4a′. The dynamic properties of
4a/4a′ are similar to those of 4b/4b′ and are discussed briefly
rac-(EBI)Zr+ (II)
{η5: η1-C5H4SiMe2NtBu}Ti+
(4a/4a′)
3(2)b
-2(3)c
1
in this section. The H NMR spectrum of 4a/4a′ (CD2Cl2) at
{η5: η1-C5Me4SiMe2NtBu}Ti+
(4b/4b′)
17.2(8)c
8(2)c
15(1)c
23 °C contains a single set of sharp resonances for the Si-Me,
tBu, TiOCMe2, R-C5H4, and â-C5H4 hydrogens, indicating that
interconversion of 4a and 4a′ (olefin face exchange) and
inversion of configuration at Ti (O-shift) are fast on the chemical
shift time scale at this temperature. The kinetics of the 4a/4a′
exchange were probed by analysis of the vinyl region of the 1H
NMR spectrum. As the temperature is raised from -60 to 23
°C, the pairs of Hint, Htrans, and Hcis 1H NMR resonances each
collapse to a single resonance. The 4a/4a′ ratio is constant from
-60 to 23 °C (62/38 ( 2%), and the coalesced vinyl resonances
at 23 °C appear close to the weighted average of the individual
low-temperature (-60 °C) chemical shifts.30 Activation param-
a Determined by analysis of ZrOCMe2 region; ref 3a. b Determined
by analysis of Mesyn region; ref 3a. c Values for conversion of major
to minor diastereomer determined by analysis of vinyl region; this work.
R-olefin complexes in both metallocene and constrained ge-
ometry systems.
The activation parameters for olefin face exchange of the
LnM(OCMe2CH2CH2CHdCH2)+ complexes that we have in-
vestigated to date are summarized in Table 3. For 4b/4b′ and
II the dynamic NMR results establish that the face exchange
involves rate-limiting decomplexation of the olefin and is
accompanied by fast inversion at the metal center (i.e., fast
O-shift) in the olefin-dissociated intermediate, and that the anion
is not involved. We presume that the face-exchange processes
of 4a/4a′ and I are similar. Interestingly, in both the constrained
geometry Ti series and the zirconocene series, the face-exchange
barrier is higher for the more crowded examples. Several factors
may influence the magnitude of the face-exchange barrier,
including the metal-olefin bond strength, solvent involvement
in the decomplexation of the olefin, and steric inhibition of
chelate ring-opening. As a {η5: η1-(C5Me4)SiMe2NtBu}Ti(OR)+
cation is expected to be less Lewis acidic than a {η5: η1-
(C5H4)SiMe2NtBu}Ti(OR)+ cation, the Ti-olefin bond strength
should be lower for 4b/4b′ than for 4a/4a′. The fact that the
face-exchange barriers follow the reverse trend suggests that
other factors besides the Ti-olefin bond strength contribute to
the face-exchange barriers. In both the constrained geometry
Ti series and the zirconocene series, the activation entropies
for the less crowded cases are ∼8-12 eu more negative than
for the more crowded cases. While the differences are small,
this trend suggests that solvent participation in the olefin
decomplexation step may play a more important role in the less
crowded cases. We are investigating the synthesis of nonchelated
LnM(OR)(olefin)+ species to investigate these issues and to
obtain more direct estimates of d0 metal-olefin bond strengths.
eters for the face-exchange process (∆Hq 4a/4a′ ) 12.2(9) kcal/
vinyl
4a/4a′
mol; ∆Sq
) -2(3) eu) were determined by line-shape
analysis of the Hcis resonances over the temperature range -40-
vinyl
23 °C.31
Discussion
Methyl abstraction from 2a/3a and 2b/3b mixtures yields the
cationic d0 metal olefin complexes {η5: η1-C5R4SiMe2NtBu}-
Ti(OCMe2CH2CH2CHdCH2)+ (R ) Me in 4b/4b′; R ) H in
4a/4a′). The NMR data for these cations are very similar to the
data for I and II, for which unsymmetrical Zr-olefin bonding
was established by X-ray crystallography. In particular, the Hint
1H resonances and the Cint 13C resonances for 4b/4b′ and 4a/
4a′ are shifted significantly downfield, and the Cterm 13C
resonances are shifted significantly upfield, from the free olefin
positions, as observed for I and II. The olefin JC-H values in
4b/4b′ and 4a/4a′ are unchanged from the free olefin values.
These results are consistent with unsymmetrical Ti-olefin
bonding in 4b/4b′ and 4a/4a′; that is, the Ti-olefin bond in
these systems comprises a weak Ti-Cterm and minimal Ti-Cint
interaction which results in polarization of the CdC π bond
with positive charge buildup at Cint. As proposed for I and II,
a σ-bonded resonance form XII thus contributes to the structures
4b/4b′ and 4a/4a′ (Chart 2). The NMR data for the coordinated
olefin units in the chelated alkyl olefin complexes V, VI, and
Cp*2YCH2CH2CRR′CHdCH2, and the nonchelated propene
complex VIII are also consistent with unsymmetrical M-olefin
bonding. Thus, unsymmetrical metal-olefin bonding appears
to be a general feature of the ground-state structures of d0 metal
Experimental Section
General Procedures. All experiments were carried out under
purified nitrogen using a Vacuum Atmospheres glovebox or standard
Schlenk techniques. Hydrocarbon solvents, diethyl ether, and tetrahy-
drofuran were distilled from Na/benzophenone. Chlorinated solvents
were distilled from calcium hydride. All solvents were stored under
N2 prior to use. 2-Methyl-5-hexene-2-ol was prepared as described
previously.3a B(C6F5)3 (Boulder Scientific) was purified by double
sublimation. [Ph3C][B(C6F5)4] was obtained from Asahi Glass Co. and
used as received. The Ti complexes {η5: η1-C5R4SiMe2NtBu}TiMe2
(1a,b; R ) H, Me) were prepared by literature procedures.11a,b
NMR spectra were recorded on Bruker AMX-600 (1H), AMX-360
(1H, 13C, 11B) or AC-300 (19F) spectrometers, in Teflon-valved tubes
(30) The chemical shifts (δobserved, δcalculated) for the vinyl hydrogen
resonances at 20 °C are as follows: Hint: 7.08, 7.10; Htrans: 5.11, 5.05;
H
cis: 4.64, 4.61.
(31) The free energy of activation at the coalescence temperature
determined using these parameters (Tcoal ) -10 °C; ∆Gqvinyl4a/4a′ ) 12.7-
(3) kcal/mol) agrees well with the value (12.5(2) kcal/mol) estimated from
the frequency difference between the unequal intensity Hcis doublets in the
absence of exchange using the method of Shanan-Atidi, H.; Bar-Eli, K.
H. J. Chem. Phys. 1970, 74, 961.

Properties of d0 Metal Olefin Complexes
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 907
(q, JC-H ) 126, OCMe2), 29.7 (t, JC-H ) 134, CH2), 15.0 (q, JC-H
at ambient probe temperature (23 °C) unless otherwise indicated. NMR
probe temperatures were controlled with a Bruker B-VT-1000E
accessory coupled with a Eurotherm 818 controller calibrated versus a
)
126, C5CH3), 11.9 (q, JC-H ) 126, C5CH3), 5.2 (q, JC-H ) 120, SiCH3).
3b: 1H NMR (C6D6) δ 2.19 (s, 3H, C5CH3), 2.01 (s, 3H, C5CH3), 1.93
1
t
MeOH thermometer. Quoted temperatures are accurate to (1 °C. H
(s, 3H, C5CH3), 1.78 (s, 3H, C5CH3), 1.39 (s, 9H, Bu), 1.23 (s, 6H,
and 13C NMR chemical shifts were determined versus the solvent
resonances and are reported relative to tetramethylsilane. 19F and 11B
NMR chemical shifts are referenced to external CFCl3 and external
OCMe2), 0.54 (s, 3H, SiCH3), 0.52 (s, 3H, SiCH3), 0.35 (s, 3H, TiCH3);
the other signals overlap with those of 2b. {1H}-13C NMR (C6D6) δ
114.3 (CH2d), 83.4 (OC), 49.6 (CMe3), 44.5 (CH2), 34.6 (CMe3), 30.7
(OCCH3), 29.3 (CH2), 14.2 (C5CH3), 12.7 (C5CH3), 11.8 (C5CH3), 11.4
(C5CH3), 6.8 (SiCH3), 6.5 (SiCH3); the other signals overlap with those
of 2b.
.
BF3 Et2O, respectively. Coupling constants are reported in Hz. The
NMR spectra of cationic complexes contain resonances for the free
anions.17,18 1H-1H COSY and 13C-detected 13C-1H phase-sensitive
HETCOR spectra were acquired on the AMX-360 spectrometer and
processed using standard Bruker programs. Detailed information on
[{η5:η1-C5H4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)][MeB-
(C6F5)3] (4a/4a′).32 A C6D6 solution of 2a/3a (95/5) was prepared in
an NMR tube as described above and concentrated under vacuum to
give a dark, yellow-brown oil. The oil was cooled to -30 °C, and
cold, solid B(C6F5)3 (1 equiv vs 1a) was added. The color of the oily
mixture immediately turned brown-orange, and gas (CH4) evolution
was observed. The tube was rapidly cooled to -78 °C and C6D6 (∼2
mL) was added by vacuum transfer. The tube was slowly warmed to
room temperature. The resulting biphasic mixture, consisting of a dark
orange lower layer (ionic complexes) and a bright yellow upper layer
(C6D6), was shaken vigorously and maintained at room temperature
until gas evolution ceased (∼15-30 min). The upper benzene layer
was removed by pipet, and the remaining oil was dried under vacuum.
The gummy residue was dissolved in CD2Cl2 to give a clear, bright
orange solution. NMR analysis established the formation of 4a (44%
yield vs internal standard) and 4a′ (27% yield). 4a: 1H NMR (CD2-
1
the H 2D NOESY-EXSY experiments is provided below.
1
{η5:η1-C5H4SiMe2NtBu}TiMe2 (1a): H NMR (C6D6) δ 6.74 (t, J
t
) 2.3, 2H, C5H4), 5.82 (t, J ) 2.3, 2H, C5H4), 1.52 (s, 9H, Bu), 0.65
(s, 6H, TiCH3), 0.25 (s, 6H, SiCH3). 13C NMR (C6D6) δ 122.1 (d, JC-H
) 172, C5H4), 121.4 (d, JC-H ) 172, C5H4), 103.8 (s, CSiMe2), 58.7
(s, CMe3), 51.0 (q, JC-H ) 121, TiCH3), 34.4 (q, JC-H ) 121, CMe3),
0.9 (q, JC-H ) 120, SiCH3).
1
{η5:η1-C5Me4SiMe2NtBu}TiMe2 (1b): H NMR (C6D6) δ 1.96 (s,
6H, C5CH3), 1.85 (s, 6H, C5CH3), 1.57 (s, 9H, tBu), 0.51 (s, 6H, TiCH3),
0.43 (s, 6H, SiCH3). 13C NMR (C6D6) δ 133.8 (s, C5CH3), 129.6 (s,
C5CH3), 97.8 (s, CSiMe2), 57.8 (s, CMe3), 51.1 (q, JC-H ) 119, TiCH3),
34.7 (q, JC-H ) 125, CMe3), 15.1 (q, JC-H ) 125, C5CH3), 12.0 (q,
JC-H ) 125, C5CH3), 6.3 (q, JC-H ) 119, SiCH3).
{η5-C5H4SiMe2NHtBu}TiMe2(OCMe2CH2CH2CHdCH2) (2a) and
{η5:η1-C5H4SiMe2NtBu}TiMe(OCMe2CH2CH2CHdCH2) (3a). Neat
2-methyl-5-hexene-2-ol (27 µL, 0.198 mmol) was added to a solution
of 1a (53.8 mg, 0.198 mmol) in C6D6 (2 mL) in a Teflon-valved NMR
tube. The tube was shaken vigorously and maintained at 23 °C for 30
d
Cl2, -60 °C) δ 7.02 (m, 1H, Hint), 6.87 (m, J ) 2.3, 1H, C5H4 ), 6.78
c
b
(m, J ) 1.2, 1H, C5H4 ), 6.57 (m, J ) 1.6, 1H, C5H4 ), 6.34 (m, J )
2.0, 1H, C5H4 ), 5.06 (d, 3J ) 18.4, 1H, Htrans), 4.78 (d, 3J ) 10.8, 1H,
a
Hcis), 2.51 (m, 2H, CH2), 2.20 (m, 2H, CH2), 1.32 (s br, 6H, OC(CH3)2),
1.24 (s, 9H, tBu), 0.73 (s, 3H, SiCH3 ), 0.58 (s, 3H, SiCH3 ). 13C NMR
b
a
min. The color of the solution was bright yellow. H and 13C NMR
1
(CD2Cl2, -60 °C) δ 159.4 (d, JC-H ) 155, dCH), 124.8 (d, JC-H
)
spectra were obtained and revealed clean conversion of 1a to a mixture
of 2a (95% NMR yield, determined using 1,4-(SiMe3)2-C6H4 as an
internal standard) and 3a (5% NMR yield). 2a: 1H NMR (C6D6) δ
6.37 (t, J ) 2.4, 2H, C5H4), 6.09 (t, J ) 2.4, 2H, C5H4), 5.87 (m, 1H,
Hint), 5.11 (dq, J ) 17.1, J ) 1.8, 1H, Htrans), 5.01 (dm, J ) 10.0, 1H,
Hcis), 2.32 (m, 2H, CH2), 1.75 (m, 2H, CH2), 1.31 (s, 6H, OCMe2),
1.09 (s, 9H, tBu), 0.65 (s, 6H, TiCH3), 0.34 (s, 6H, SiCH3), NH signal
a
b
172, C5H4 ), 123.9 (d, JC-H ) 172, C5H4 ), 122.6 (d, JC-H ) 172,
d
c
C5H4 ), 119.4 (d, JC-H ) 172, C5H4 ), 113.4 (s, CSiMe2), 104.3 (t, JC-H
) 159, CH2d), 90.3 (s, OC), 66.5 (s, CMe3), 49.3 (t, JC-H ) 130,
CH2), 33.2 (q, JC-H ) 128, CMe3), 31.6 (t, JC-H ) 128, CH2), 30.2 (q,
b
a
JC-H ) 129, OCCH3 ), 26.5 (q, JC-H ) 129, OCCH3 ), -0.6 (q, JC-H
a
b
1
) 120, SiCH3 ), -0.85 (q, JC-H ) 120, SiCH3 ). 4a′: H NMR (CD2-
d
1
Cl2, -60 °C) δ 7.24 (m, 1H, Hint), 6.73 (m, J ) 1.6, 1H, C5H4 ), 6.66
obscured. H NMR (CD2Cl2) δ 6.46 (t, J ) 2.4, 2H, C5H4), 6.28 (t, J
(2m, J ) 2.0, 2H, C5H4b,c), 6.55 (m, J ) 2.3, 1H, C5H4 ), 5.04 (d, J
a
3
) 2.4, 2H, C5H4), 5.90 (m, 1H, Hint), 5.05 (dq, J ) 17.1, J ) 1.8, 1H,
3
) 18.4, 1H, Htrans), 4.32 (d, J ) 10.8, 1H, Hcis), 2.51 (m, 2H, CH2),
H
trans), 4.95 (dm, J ) 10.0, 1H, Hcis), 2.33 (m, 2H, CH2), 1.80 (m, 2H,
b
a
t
2.20 (m, 2H, CH2), 1.35 (s, 3H, OCCH3 ), 1.30 (s, 3H, OCCH3 ), 1.18
CH2), 1.42 (s, 6H, OCMe2), 1.12 (s, 9H, Bu), 0.76 (br s, 1H, NH),
0.30 (s, 6H, SiCH3), 0.27 (s, 6H, TiCH3). 13C NMR (C6D6) δ 139.2 (d,
JC-H ) 150, CHd), 126.1 (s, CSiMe2), 118.6 (d, JC-H ) 172, C5H4),
115.9 (d, JC-H ) 172, C5H4), 114.5 (t, JC-H ) 155, CH2d), 85.0 (s,
b
a
(s, 9H, tBu), 0.72 (s, 3H, SiCH3 ), 0.64 (s, 3H, SiCH3 ). 13C NMR (CD2-
Cl2, -60 °C) δ 163.1 (d, JC-H ) 151, CHd), 125.9 (d, JC-H ) 172,
a
b
d
C5H4 ), 123.0 (d, JC-H ) 172, C5H4 ), 121.7 (d, JC-H ) 172, C5H4 ),
c
119.9 (d, JC-H ) 172, C5H4 ), 114.1 (s, CSiMe2), 98.0 (t, JC-H ) 159,
OC), 49.7 (s, CMe3), 47.8 (q, JC-H ) 122, TiCH3), 44.1 (t, JC-H
)
CH2d), 90.5 (s, OC), 63.1 (s, CMe3), 48.8 (t, JC-H ) 130, CH2), 33.0
126, CH2), 33.8 (q, JC-H ) 125, CMe3), 30.3 (q, JC-H ) 126, OCCH3),
29.5 (t, JC-H ) 134, CH2), 2.5 (q, JC-H ) 118, SiCH3). 3a: 1H NMR
(C6D6) δ 6.46 (m, 1H, C5H4), 6.32 (m, 1H, C5H4), 6.24 (m, 1H, C5H4),
a
(q, JC-H ) 128, CMe3), 32.3 (t, JC-H ) 129, CH2), 30.8 (OCCH3 ),
b
b
28.7 (q, JC-H ≈ 131, OCCH3 ), -0.4 (q, JC-H ) 120, SiCH3 ), -0.86
a
t
(q, JC-H ) 120, SiCH3 ).
1.34 (s, 9H, Bu), 1.07 (s, 6H, OCMe2), 0.61 (s, 3H, TiCH3), 0.40 (s,
[{η5:η1-C5Me4SiMe2NtBu}Ti{OCMe2CH2CH2CHdCH2}][MeB-
(C6F5)3] (4b/4b′).32 This compound was generated from a 2b/3b mixture
(89/11) and 1 equiv of B(C6F5)3 using the procedure described for 4a/
4a′. NMR analysis revealed that 2b/3b underwent conversion to 4b
(67% yield vs internal standard) and 4b′ (22% yield). 4b: 1H NMR
(CD2Cl2, -20 °C) δ 6.52 (m, 1H, Hint), 4.99 (d, 3J ) 18.4, 1H, Htrans),
3H, SiCH3), 0.37 (s, 3H, SiCH3); the other signals overlap with those
of 2a.
{η5-C5Me4SiMe2NHtBu}TiMe2(OCMe2CH2CH2CHdCH2) (2b)
and {η5:η1-C5Me4SiMe2NtBu}TiMe(OCMe2CH2CH2CHdCH2) (3b).
A 2b/3b mixture was generated from 1b (43.7 mg, 0.137 mmol) in
C6D6 (∼2 mL) and 2-methyl-5-hexene-2-ol (18.3 µL, 0.137 mmol)
using the procedure described for 2a/3a. The final solution was bright
3
2
b
4.45 (dd, J ) 8.6, J ) 2.5, 1H, Hcis), 2.53 (m, 2H, CH2 ), 2.25 (m,
2H, CH2 ), 2.22 (s, 3H, C5CH3 ), 2.18 (s, 3H, C5CH3 ), 2.13 (s, 3H,
C5CH3 ), 2.03 (s, 3H, C5CH3 ), 1.43 (s, 3H, OCCH3 ), 1.41 (s, 3H,
a
d
c
1
yellow. H and 13C NMR spectra were recorded and revealed clean
b
a
b
conversion to 2b (89% NMR yield vs internal standard) and 3b (11%
NMR yield). 2b: 1H NMR (C6D6) δ 5.90 (m, 1H, Hint), 5.15 (dq, J )
17.1, J ) 1.8, 1H, Htrans), 5.02 (dq, J ) 8.0, J ) 1.8, 1H, Hcis), 2.34
(m, 2H, CH2), 2.12 (s, 6H, C5CH3), 1.94 (m, 2H, CH2), 1.88 (s, 6H,
OCCH3 ), 1.31 (s, 9H, tBu), 0.87 (s, 3H, SiCH3 ), 0.79 (s, 3H, SiCH3 ).
13C NMR (CD2Cl2, -60 °C) δ 162.8 (d, JC-H ) 152, CHd), 138.5 (s,
C5CH3), 136.5 (s, C5CH3), 135.8 (s, C5CH3), 133.2 (s, C5CH3), 109.3
(s, CSiMe2), 106.0 (t, JC-H ) 157, CH2d), 90.2 (s, OC), 64.7 (s, CMe3),
a
b
a
t
C5CH3), 1.47 (s, 6H, OCMe2), 1.09 (s, 9H, Bu), 0.47 (s, 6H, TiCH3),
a
49.4 (t, JC-H ) 122, CH2 ), 33.5 (q, JC-H ) 127, CMe3), 31.5 (t, JC-H
) 134, CH2 ), 31.0 (q, JC-H ) 124, OCCH3 ), 27.8 (q, JC-H ) 128,
0.41 (s, 6H, SiCH3). 1H NMR (CD2Cl2) δ 5.89 (m, 1H, Hint), 5.05 (dm,
J ) 17.0, 1H, Htrans), 4.94 (dm, J ) 10.0, 1H, Hcis), 2.32 (m, 2H, CH2),
2.08 (s, 6H, C5CH3), 1.98 (s, 6H, C5CH3), 1.91 (m, 2H, CH2), 1.51 (s,
6H, OCMe2), 1.11 (s, 9H, tBu), 0.57 (br s, 1H, NH), 0.28 (s, 6H, SiCH3),
0.05 (s, 6H, TiCH3). 13C NMR (C6D6) δ 139.2 (d, JC-H ) 158, CHd),
129.2 (s, C5CH3), 127.2 (s, C5CH3), 119.9 (s, CSiMe2), 114.4 (t, JC-H
) 155, CH2d), 85.0 (s, OC), 49.9 (q, JC-H ) 121, TiCH3), 49.6 (s,
CMe3), 44.6 (t, JC-H ) 127, CH2), 33.7 (q, JC-H ) 125, CMe3), 30.5
b
b
a
a
d
OCCH3 ), 15.0 (q, JC-H ) 128, C5CH3 ), 14.8 (q, JC-H ) 128, C5CH3 ),
b
c
13.1 (q, JC-H ) 128, C5CH3 ), 12.2 (q, JC-H ) 128, C5CH3 ), 5.1 (q,
a
b
JC-H ) 121, SiCH3 ), 4.2 (q, JC-H ) 121, SiCH3 ). 4b′: 1H NMR (CD2-
(32) The superscript labels (a, b, c, d) following the 1H and 13C chemical
shift assignments of 4a/4a′ and 4b/4b′ denote resonances that are correlated
in 13C-1H HETCOR spectra.

908 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
Carpentier et al.
3
2
Cl2, -20 °C) δ 7.30 (m, 1H, Hint), 4.56 (dt, J ) 17.6, J ) 2.5, 1H,
diastereomer. The angles between Ti-C-C and CH2dCH- planes
for the lowest-energy structures are 78° for the R,R-diastereomer and
88° for the R,S-diastereomer.
H
trans), 4.01 (dd, 3J ) 9.4, 2J ) 2.9, 1H, Hcis), 2.60 (m, 2H, CH2), 2.25
d c
(m, 2H, CH2), 2.24 (s, 3H, C5CH3 ), 2.14 (s, 3H, C5CH3 ), 2.11 (s, 3H,
b
a
b
C5CH3 ), 2.07 (s, 3H, C5CH3 ), 1.41 (s, 3H, OCCH3 ), 1.33 (s, 3H,
NMR Simulations for 4b/4b′. Simulations were performed using
the software package gNMR (Cherwell Scientific, version 3.6.5). The
vinyl region was simulated as a set of three two-site exchange systems.
The chemical shifts and coupling constants for the three vinyl hydrogens
of each diastereomer in the slow limit exchange (below -40 °C) were
used to set up the spin systems of the two species, and the relative
population ratio was fixed at 75/25. The natural line width in the absence
OCCH3 ), 1.28 (s, 9H, tBu), 0.91 (s, 3H, SiCH3 ), 0.71 (s, 3H, SiCH3 ).
13C NMR (CD2Cl2, -60 °C) δ 167.8 (d, JC-H ) 155, CHd), 108.6 (s,
CSiMe2), 101.6 (t, JC-H ) 157, CH2d), 90.1 (s, OC), 63.7 (s, CMe3),
49.3 (t, JC-H ) 122, CH2), 33.3 (q, JC-H ) 127, CMe3), 29.7 (2 s,
a
b
a
c
OC(CH3)2), 15.5 (q, JC-H ) 128, C5CH3 ), 15.0 (overlaps with a
d
resonance of the major diastereomer, C5CH3 ), 13.4 (q, JC-H ) 128,
a
b
b
A
C5CH3 ), 12.1 (q, JC-H ) 129, C5CH3 ), 5.2 (q, JC-H ) 120, SiCH3 ),
of exchange (W0 ) 1.4 Hz) was measured at -40 °C for Hcis and
a
HcisB and confirmed by observation of the same line width at -60 °C.
The accuracy of the NMR parameters and population ratio was assessed
by comparison of the simulated spectrum with the experimental
4.4 (q, JC-H ) 119, SiCH3 ); the other Cp signals overlap with anion
resonances.
[{η5:η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)][B-
(C6F5)4] (5b/5b′). This compound was generated by the procedure
described for 4b/4b′ using a 89:11 mixture of 2b/3b (0.082 mmol)
spectrum at -40 °C. Then, spectra were calculated assuming that
A
exchange of each pair of corresponding hydrogens, HintA/HintB, Htrans
/
and 1 equiv of [Ph3C][B(C6F5)4]. The variable-temperature 1H and 13
C
HtransB, and HcisA/HcisB, occurs at the same rate, which was optimized
to match the simulated and experimental spectra. The uncertainties in
the exchange rate constants determined in this manner are ∼( 20%.
Activation parameters were determined by a standard Eyring analysis
(Figure 4). The standard deviations for the least-squares fit were used
spectra revealed that 5b/5b′ are formed in the same yield and ratio as
4b/4b′ (5b: 67% yield vs internal standard; 5b′: 22%) and that the
dynamic behavior of 5b/5b′ is identical to that of 4b/4b′. The spectra
of 5b/5b′ are identical to those of 4b/4b′ except for the anion and Ph3-
CCH3 resonances.
AB
AB
to estimate the uncertainties in ∆Hq
and ∆Sq
.
vinyl
vinyl
[{η5:η1-C5H4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)(THF)]
[MeB(C6F5)3] (6a). Excess THF (50 µL, 0.60 mmol) was added to a
solution of 4a/4a′ (.∼0.1 mmol) in CD2Cl2 (∼2 mL). The color of the
solution changed rapidly from bright orange to yellow-orange. The
reaction mixture was stirred for 10 min, the volatiles were removed
under vacuum, and CD2Cl2 (∼2 mL) was added by vacuum transfer.
This procedure was repeated a second time. NMR spectra were recorded
and revealed that 4a/4a′ was completely converted to 6a. 6a: 1H NMR
δ 7.19 (m, 1H, C5H4), 6.70 (m, 2H, C5H4), 6.26 (m, 1H, C5H4), 5.82
(m, 1H, Hint), 5.05 (dd, J ) 17.3, J ) 2.2, 1H, Htrans), 4.98 (dd, J )
10.0, J ) 2.2, 1H, Hcis), 4.05 (m, 2H, THF), 3.93 (m, 2H, THF), 2.2-
2.0 (m, 6H, THF and CH2-CHd), 1.43 (s, 3H, OCCH3), 1.42 (s, 3H,
The Si-Me region was simulated as a four-site system (A1, A2,
B1, B2) in 3/8:3/8:1/8:1/8 relative population ratio. The chemical shifts
observed in the slow-limit exchange (below -5 °C) were used to set
up the spin system using a natural line width W0 ) 1.4 Hz. The chemical
shifts of the Si-Me resonances vary slightly with temperature between
-5 and 10 °C (δ SiMeA1 0.788 to 0.792; SiMeA2 0.875 to 0.890; SiMeB1
0.712 to 0.720; SiMeB2 0.910 to 0.915); chemical shifts in the absence
of exchange at higher temperatures were estimated by linear extrapola-
tion of the -5 to 10 °C values. Spectra were simulated as described in
the text, assuming that the six possible site-to-site exchanges (A1-
A2, A1-B1, A1-B2, A2-B1, A2-B2, B1-B2) occur at relative rates
of 9/3/3/3/3/1. Exchange rates were obtained by comparison of
experimental spectra with simulated spectra for eight temperatures in
the range -5-60 °C. As shown in Figure 3, a good fit was obtained
using this procedure. The uncertainty in the Si-Me site-to-site exchange
rates was probed by extensive simulations using different values for
t
OCCH3), 1.31 (s, 9H, Bu), 0.72 (s, 3H, SiCH3), 0.68 (s, 3H, SiCH3).
13C NMR: δ 137.6 (d, JC-H ) 154, CHd), 125.8 (d, JC-H ) 175,
C5H4), 123.7 (d, JC-H ) 175, C5H4), 123.0 (d, JC-H ) 175, C5H4),
122.1 (d, JC-H ) 175, C5H4), 115.4 (t, JC-H ) 156, CH2d), 112.9 (s,
CSiMe2), 92.5 (s, OC), 79.3 (t, JC-H ) 153, OCH2 THF), 64.1 (s,
CMe3), 44.3 (t, JC-H ) 130, CH2-CH), 30.4 (q, JC-H ) 127, OCCH3),
29.9 (t, JC-H ) 128, CH2), 29.7 (t, JC-H ) 132, CH2 THF), 0.6 (q,
JC-H ) 121, SiCH3), -0.1 (q, JC-H ) 121, SiCH3).
R
A1A2, RAxBy, and RB1B2.
Variable-Temperature 1H 2D-EXSY NMR Spectra of 4b/4b′.
1
Two-dimensional H spectra were obtained using the Bruker AMX-
360 spectrometer with the program “noesysh”. The pulse sequence was
D1-90°-D0-90°-D9-90°-FID. The initial relaxation delay time D1 was
typically 3 s (>4 × T1, vide infra), the initial D0 was 3 µs, and the
mixing time D9 (τm) was varied according to the experimental
temperature (vide infra). The pulse sequence was repeated for 128
values of D0, that is, the F1 dimension contained 128 points, which
was then zero-filled to 512 points. The F2 dimension contained 2048
points. The number of scans per experiment was 16, giving a total
experiment time of ∼135 min. The optimal mixing time τm,opt was
chosen on the basis of eq 11 to minimize the relative error in the rate
constant.34,35 The activation parameters determined from the Eyring
analysis of the line-shape data (Figure 4) were used to estimate rate
constants kAB and kBA. Spin-lattice relaxation times T1 were measured
by using the standard inversion-recovery method. T1 values for the
vinyl hydrogens of 4b/4b′ (A/B) measured at 0 and -20 °C are listed
in Table 4.
[{η5:η1-C5Me4SiMe2NtBu}Ti(OCMe2CH2CH2CHdCH2)(THF)]-
[B(C6F5)4] (6b). Compound 6b was generated from 5b/5b′ using the
procedure described for 6a. 6b: 1H NMR (CD2Cl2) δ 5.84 (m, 1H,
Hint), 5.06 (dd, J ) 17.3, J ) 2.2, 1H, Htrans), 5.00 (dd, J ) 10.0, J )
2.2, 1H, Hcis), 4.05 (m, 2H, THF), 3.95 (m, 2H, THF), 2.33 (s, 3H,
C5CH3), 2.24 (s, 3H, C5CH3), 2.09 (s, 3H, C5CH3), 2.05 (s, 3H, C5CH3),
2.2-2.0 (m, 4H, THF), 1.46 (s, 3H, OCCH3), 1.45 (s, 3H, OCCH3),
1.34 (s, 9H, tBu), 0.81 (s, 3H, SiCH3), 0.79 (s, 3H, SiCH3); resonances
of Ph3CCH3 were also observed.
Molecular Modeling. Molecular mechanics calculations were
performed using the software package Cerius2 (version 2.1, Molecular
Simulations, Inc.). The universal force field (UFF) developed by Rappe´
was used.33 The parameters used to generate the UFF include a set of
hybridization-dependent atomic bond radii, a set of hybridization angles,
van der Waals parameters, torsional and inversion barriers, and a set
of effective nuclear charges. The {η5: η1-C5Me4SiMe2NtBu}Ti unit was
constructed and constrained to atomic coordinates taken from crystal-
lographic data for {η5: η1-C5Me4SiMe2NtBu}TiCl2 (centroid-Ti-N
angle ) 108°).11c Constraints were implemented by the “Open Force
Field Setup-Energy Terms-Restraints” package. The Ti-O-C angle
was constrained to 160°, and the Ti-C distances were constrained to
Ti-Cint ) 2.70 Å, Ti-Cter ) 2.55 Å, based on crystallographic data
for I and II. No other constraints were applied. The structure was
minimized by conjugate gradient methods to a 0.1 kcal mol-1 Å-1 root-
mean-square force. Different conformations of the alkoxide-olefin
chelate ring were examined to find a global energy minimum for each
τm,opt ∼ 1/(T1-1 + kAB + kBA
)
(11)
Control experiments established that EXSY-determined rate constants
are not significantly affected by a (50% variation in τm,opt. Uncertainties
in rate constant values were estimated assuming (1% uncertainty in
the finite S/N ratios of the computed spectra, and (2% uncertainty in
(34) The mixing time must be of sufficient magnitude to give relatively
strong cross-peaks but not so long that the signal intensities become
insensitive to exchange rates, or all signals lose appreciable intensity due
to spin-lattice relaxation.
(35) Abel, E. W.; Coston, T. P. J.; Orrell, K. G.; Sik, V.; Stephenson,
D. J. Magn. Res. 1986, 70, 34.
(33) (a) Rappe´, A. K.; Casewitt, C. J.; Colwell, K. S.; Goddard, W. A.,
III; Skiff, W. M. J. Am. Chem. Soc. 1992, 114, 10024. (b) Rappe´, A. K.;
Colwell, K. S.; Casewitt, C. J. Inorg. Chem. 1993, 32, 3438.

Properties of d0 Metal Olefin Complexes
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 909
Table 4. T1 Values for the Vinyl Hydrogens of 4b/4b′ at 0 and
-20 °C
Acknowledgment. This work was supported by NSF grant
CHE-9413022 (R.F.J.). J.-F.C. thanks the French “Centre
National de la Recherche Scientifique” and Elf Atochem for
support.
A
B
A
B
A
B
T, °C
Hint
Hint
Htrans
Htrans
Hcis
Hcis
0
-20
0.85
0.60
0.85
0.55
0.60
0.40
0.50
0.35
0.50
0.35
0.50
0.35
Supporting Information Available: NMR spectra of 4b/
4b′ and 4a/4a′ (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
2D signal integrations. Two separate experiments performed at -30
°C gave results consistent with the uncertainty estimated using these
assumptions.
JA003209L