Synthesis of a series of new, highly electrophilic, monocyclopentadienyltitanium olefin polymerization initiators

Article abstract of DOI:10.1039/a703112b

The new compounds Ti(η-C₅Me₅)Me₂E (E = C₆F₅ or OC₆F₅) and Ti(η-C₅Me₅)Me(OC₆F₅)₂ have been synthesized and characterized by a variety of techniques, including ^47/49Ti NMR spectroscopy. All three compounds react with the borane B(C₆F₅)₃ to form the highly electrophilic but thermally unstable species Ti(η-C₅Me₅)Me(E)(η-Me)B-(C₆F ₅)₃ and [η-C₅Me₅)Ti(OC₆F₅) ₂][BMe(C₆F₅)₃], the solution structures and dynamics of which are investigated and compared with those of the known compound Ti(η-C₅Me₅)Me₂(η-Me)B(C ₆F₅)₃,. Interestingly, Ti(η-C₅Me₅)-Me(C₆F ₅)(η-Me)B(C₆F₅)₃ undergoes neither significant ion-pair dissociation to the solvent separated ions [(η-C₅Me₅)TiMe(C₆F₅)] ⁺ and [BMe(C₆F₅)₃]^- nor borane dissociation to its precursors (η-C₅Me₅)TiMe₂(C₆F₅) and B(C₆F₅)₃; indeed, both rotation about the Ti-C₆F₅ bond and inversion at the chiral metal are slow on the NMR time-scale. In contrast, Ti(η-C₅Me₅)Me(OC₆F ₅)(η-Me)B(C₆F₅)₃ is more labile and, like Ti(η-C₅Me₅)Me₂;(η-Me)-B(C ₆F₅)₃, undergoes ion-pair dissociation, while [(η-C₅Me₅)Ti(OC₆F₅) ₂][BMe(C₆F₅)₃] exists in solution as the solvent separated ion species [(η-C₅Me₅)Ti(OC₆F₅) ₂]^- and [BMe(C₆F₅)₃]^- in equilibrium with its precursors, (η-C₅Me₅)TiMe(OC₆F₅)₂ and B(C₆F₅)₃.

Full text of DOI:10.1039/a703112b

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis of a series of new, highly electrophilic,
monocyclopentadienyltitanium olefin polymerization initiators†
Mark J. Sarsfield, Sean W. Ewart, Tracey L. Tremblay, Aleksander W. Roszak and
Michael C. Baird*
Department of Chemistry, Queen’s University, Kingston, Ontario, Canada K7L3N6
The new compounds Ti(η-C5Me5)Me2E (E = C6F5 or OC6F5) and Ti(η-C5Me5)Me(OC6F5)2 have been synthesized
and characterized by a variety of techniques, including 47/49Ti NMR spectroscopy. All three compounds react with
the borane B(C6F5)3 to form the highly electrophilic but thermally unstable species Ti(η-C5Me5)Me(E)(µ-Me)B-
(C6F5)3 and [(η-C5Me5)Ti(OC6F5)2][BMe(C6F5)3], the solution structures and dynamics of which are investigated
and compared with those of the known compound Ti(η-C5Me5)Me2(µ-Me)B(C6F5)3. Interestingly, Ti(η-C5Me5)-
Me(C6F5)(µ-Me)B(C6F5)3 undergoes neither significant ion-pair dissociation to the solvent separated ions
[(η-C5Me5)TiMe(C6F5)]ϩ and [BMe(C6F5)3]Ϫ nor borane dissociation to its precursors (η-C5Me5)TiMe2(C6F5) and
B(C6F5)3; indeed, both rotation about the Ti᎐C6F5 bond and inversion at the chiral metal are slow on the NMR
time-scale. In contrast, Ti(η-C5Me5)Me(OC6F5)(µ-Me)B(C6F5)3 is more labile and, like Ti(η-C5Me5)Me2(µ-Me)-
B(C6F5)3, undergoes ion-pair dissociation, while [(η-C5Me5)Ti(OC6F5)2][BMe(C6F5)3] exists in solution as the
solvent separated ion species [(η-C5Me5)Ti(OC6F5)2]ϩ and [BMe(C6F5)3]Ϫ in equilibrium with its precursors,
(η-C5Me5)TiMe(OC6F5)2 and B(C6F5)3.
Since the early 1980s, there has been extensive research into
the use of Group 4 metallocene compounds as homogeneous
catalysts for olefin polymerization. The most successful cata-
lysts are the formally 14-electron, cationic complexes
[CpЈ2MR]ϩ (CpЈ = functionalized cyclopentadienyl; M = Ti, Zr
or Hf; R = alkyl or H),1–3 which contain a vacant co-
ordination site and an alkyl ligand and which can therefore
readily take part in co-ordination polymerization (Ziegler–
Natta) processes with a variety of olefins. However, these
complexes can also co-ordinate counter anions2 and Lewis
bases,2 and are sometimes found to react with solvents,4 all
processes which can reduce the activity of the cationic moiety,
and thus judicious choice of polymerization conditions is oft
imperative. Of great relevance here, crystallographically char-
acterized compounds of the type CpЈ2MMe(µ-Me)B(C6F5)3
have been studied intensively as precursors for the above-
mentioned cationic species [CpЈ2MR]ϩ, and have provided
valuable information concerning the optimal requirements of
non-co-ordinating counter anions.5–11
Another family of compounds which satisfy the above
requirements for Ziegler–Natta catalysis are 10-electron,
monocyclopentadienyl complexes of the type [CpЈMR2]ϩ.
These electronically less saturated and sterically less hindered
complexes might be expected to exhibit even higher reactivities
than their metallocene counterparts, and we12–20 have investi-
gated several such systems in detail. Thus reaction of Ti(η-
C5Me5)Me3 with the strong Lewis acid B(C6F5)3 results in the
formation of the methyl-bridged complex Ti(η-C5Me5)Me2-
(µ-Me)B(C6F5)3, which dissociates in solution to form [Ti-
(η-C5Me5)Me2]ϩ,17 an extremely active initiator for both carbo-
cationic [isobutylene,21 vinyl ethers,15 styrene (below 0 ЊC)16]
and Ziegler–Natta (ethylene,16 hex-1-ene,14 styrene16,22 and
propylene19) polymerization processes. Furthermore, while
metallocene complexes permit catalyst fine tuning only via
functionalizing of the cyclopentadienyl ring and changes in the
counter anion, monocyclopentadienyl complexes contain a
‘spectator’ anionic ligand {methyl in the case of [Ti-
(η-C5Me5)Me2]ϩ} and thus also offer the opportunity to alter
the steric and electronic properties of the initiator system by
changing the ‘spectator’ ligand.
The research described here began with the hypothesis that,
since cationic complexes generally form more active polymer-
ization catalysts than do neutral species,1–3 it seemed possible
that substitution of one of the methyl groups of [Ti(η-
C5Me5)Me2]ϩ with a more electron-withdrawing ligand could
result in the formation of more active initiators. This expect-
ation is contrary to current conventional wisdom for metal-
locene systems, since incorporation of electron-withdrawing
groups on the CpЈ rings of some metallocenes has been
reported to result in both decreased activity and reduced poly-
mer molecular weights.23 However, the relative rates of initi-
ation, propagation, termination and chain transfer are all
affected unpredictably by ligand substitution, and the reasons
for the observed deactivation were not assessed in detail.23 In
addition, as has been shown,5–11 enhanced Lewis acidity of the
metal can result in stronger binding of the counter anion, pos-
sibly resulting in both inhibition of monomer co-ordination
and more facile termination of polymer growth via olefin lig-
and substitution. Utilization of a counter anion which for
steric or electronic reasons could not co-ordinate would obviate
the latter modes of deactivation, and thus the effects of
increased metal ion electrophilicity on overall initiator activity
arguably remain moot.
While complexes of the type [Ti(η-C5Me5)Me(OR)]ϩ have
previously been synthesized and shown to be less active than
[Ti(η-C5Me5)Me2]ϩ,16,24 the lower activities were probably a
result of decreased electrophilicity of the metal because of π-
electron donation of the alkoxy oxygen lone pairs to the empty
d orbitals of the metal and/or dimerisation via bridging alkoxy
ligands. Extending this work, we now describe the synthesis,
characterization and solution chemistry of the compounds
Ti(η-C5Me5)Me2E (E = C6F5 or OC6F5) and Ti(η-C5Me5)Me-
(OC6F5)2, which contain highly electronegative, poor π-electron
donors ligands. When reacted with B(C6F5)3, these form
respectively the anticipated species Ti(η-C5Me5)MeE(µ-
Me)B(C6F5)3 and [(η-C5Me5)Ti(OC6F5)2][BMe(C6F5)3], which
exhibit interesting solution chemistry. Successful efforts to
utilize these complexes as olefin polymerization catalysts will be
described elsewhere.
† Dedicated to the memory of Geoff Wilkinson, a generous mentor in
whose laboratory M. C. B. spent a very fruitful eighteen months.
J. Chem. Soc., Dalton Trans., 1997, Pages 3097–3104
3097

View Article Online
toluene (40 cm3) was stirred at 0 ЊC for 1 h before being
Experimental
warmed to room temperature and stirred for a further 5 h.
The orange suspension was filtered through Celite affording a
clear orange solution. The solvent was removed in vacuo and
crystals were grown from a toluene–hexanes solution at
Ϫ10 ЊC. Yield 66% (Found: C, 50.18; H, 3.11. Calc. for
All manipulations were carried out, under dry, prepurified
grade dinitrogen, using conventional Schlenk techniques, a
Vacuum Atmospheres glove box and dried, thoroughly
deoxygenated solvents. Proton, 13C-{1H} and 19F NMR spectra
were run using a Bruker AM 400 spectrometer operating at
1
C28H18F12O3Ti: C, 50.03; H, 2.90%). H NMR (C6D6): δ 1.85
1
400.14 MHz, 100.6 MHz and 376.5 MHz respectively; H and
(s, 15 H, C5Me5), 5.92 [tt, 4J(HF ) = 6.1, 1.9 Hz, 3 H, p-H].
13C-{1H} NMR (C6D6): δ 140.1 [d, J(CF ) = 247, o-CF], 139.4
[d, J(CF ) = 248 Hz, m-CF], 135.3 (p-CH), 128.3 (C5Me5),
14.0 (C5Me5), 14.0 (C5Me5). 19F NMR (C6D6): δ Ϫ141.5
[d, J(CF ) = 20.3, 2 F, o-F], Ϫ159.7 [d, J (CF ) = 20.3 Hz, 2 F,
m-F].
13C-{1H} NMR spectra are referenced with respect to internal
SiMe4 using residual proton resonances or carbon resonances,
respectively, of the solvents, while 19F spectra are referenced to
external CFCl3. Titanium-47,49 NMR spectra (Table 1) were
obtained (50,000–100,000 transients were acquired on 1  sam-
ples) on a Bruker AM400 spectrometer with a 10 mm diameter
broad-band probe at ambient temperature, and are referenced
to external TiCl4. A Hewlett-Packard 8452A diode array
spectrophotometer was used for measuring UV/VIS spectra.
Elemental analyses for carbon and hydrogen were carried
out by Canadian Microanalytical Services, Delta, British
Columbia. The compounds Ti(η-C5Me5)Cl3,25 Ti(η-C5Me5)-
{Ti(ç-C5Me5)(OC6HF4-p)2}2(ì-O) 5. While attempting to
grow crystals of Ti(η-C5Me5)(OC6HF4-p)3 in toluene, slow
hydrolysis by adventitious water gave orange crystals of 5. In
a separate experiment, a toluene solution of Ti(η-C5Me5)-
(OC6HF4-p)3 was left to stir in air overnight. The solvent was
1
removed in vacuo to give an orange powder of 5, H NMR
27
28a
MeCl2,26 Ti(η-C5Me5)Me2Cl,26 Ti(η-C5Me5)Me3 B(C6F5)3
(C6D6) δ 1.89 (15 H, C5Me5), 5.96 (m, 3 H, p-H), which was
characterized crystallographically (see below).
and LiC6F5 28b were prepared by literature methods.
The following compounds could not be isolated as pure
solids, and all work with them involved synthesis and spectro-
scopic characterization in situ on NMR scale samples prepared
in NMR tubes.
Syntheses
Ti(ç-C5Me5)Me2(C6F5) 1. A solution of Ti(η-C5Me5)Me2Cl
(4.01 g, 16.1 mmol) in hexanes (200 cm3) was added to a
solution of LiC6F5 (18.8 mmol) in diethyl ether (30 cm3)
cooled to Ϫ78 ЊC. After stirring at Ϫ78 ЊC for 2 h, the result-
ing suspension was allowed to warm to room temperature
before filtering through Celite. The deep red filtrate was con-
centrated in vacuo to a red oil before being recrystallised
twice from hexanes at Ϫ78 ЊC. Yield 58% (Found: C, 56.43;
Ti(ç-C5Me5)Me(C6F5)(ì-Me)B(C6F5)3 9.
A
solution of
B(C6F5)3 (31 mg, 0.06 mmol) in CD2Cl2 (0.3 cm3) was added to
a solution of Ti(η-C5Me5)Me2(C6F5) (23 mg, 0.06 mmol)
in CD2Cl2 (0.4 cm3) cooled to Ϫ78 ЊC. The red solution was
maintained at Ϫ78 ЊC for a few minutes, then placed in the
probe of an AM-400 NMR spectrometer at Ϫ50 ЊC where it
1
H, 5.99. Calc. for C18H21F5Ti: C, 56.86; H, 5.57%). H NMR
4
1
(CDCl3): δ 1.98 (s, 15 H, C5Me5), 1.41 [t, 6 H, J(HF ) = 2 Hz,
was examined by 1H, 13C-{1H} and 19F NMR spectroscopy. H
TiMe]. 13C-{1H} NMR (CDCl3): δ 127.0 (C5Me5), 80.0 [t,
3J(CF ) = 3 Hz, TiMe], 12.4 (C5Me5). 19F NMR (C6H6):
δ Ϫ121.4 (m, 2 F, o-F ), Ϫ155.6 (t, 1 F, p-F ), Ϫ163.0 (m, 2 F,
m-F ).
NMR (CD2Cl2): δ 2.61 [d, 4J(HF ) = 3 Hz, 3 H, TiMe], 2.10 (15
H, C5Me5), 1.36 (br s, 3 H, µ-Me). 13C-{1H} NMR (CD2Cl2):
δ 138.2 (C5Me5), 109.9 (TiMe), 13.6 (C5Me5). 19F NMR
(CD2Cl2): δ Ϫ118.6 (m, 1 F, o-F of TiC6F5), Ϫ124.3 (m, 1 F, o-F
of TiC6F5), Ϫ135.1 (m, 6 F, o-F of BC6F5), Ϫ150.1 (t, 1 F, p-F
of TiC6F5), Ϫ160.3 (m, 1 F, m-F of TiC6F5), Ϫ160.8 (t, 3 F,
p-F of BC6F5), Ϫ161.7 (m, 1 F, m-F of TiC6F5), Ϫ166.0 (m,
6 F, m-F of BC6F5).
Ti(ç-C5Me5)Me2(OC6F5) 2. A solution of Ti(η-C5Me5)Me2Cl
(4.0 g, 16.1 mmol) and LiOC6F5 (3.1 g, 16.1 mmol) in hexanes
(200 cm3) was stirred at 0 ЊC for 4 h and then warmed to room
temperature. The yellow suspension was filtered through Celite
and dried under vacuum. The bright yellow solid was isolated
pure after recrystallizing twice from hexanes at Ϫ78 ЊC. Yield
71% (Found: C, 54.34; H, 5.33. Calc. for C18H21F5OTi: C, 54.56;
Ti(ç-C5Me5)Me(OC6F5)(ì-Me)B(C6F5)3 10. In an NMR
experiment carried out as above, B(C6F5)3 (31 mg, 0.06
mmol) in CD2Cl2 (0.3 cm3) was added to a solution of
Ti(η-C5Me5)Me2(OC6F5) (24 mg, 0.06 mmol) in CD2Cl2 (0.4
cm3) cooled to Ϫ78 ЊC providing an orange solution. 1H
NMR (CD2Cl2): δ 2.04 (15 H, C5Me5), 1.89 (3 H, TiMe),
0.62 (br s, 3 H, µ-Me). 13C-{1H} NMR (CD2Cl2): δ 134.3,
(C5Me5), 82.2 (TiMe), 12.2 (C5Me5). 19F NMR (CD2Cl2):
Ϫ135.4 (m, 6 F, o-F of BC6F5), Ϫ159.7 (m, 2 F, o-F of
TiOC6F5), Ϫ160.9 (t, 3 F, p-F of BC6F5), Ϫ164.5 (m, 2 F, m-F
of TiOC6F5), Ϫ165.0 (t, 1 F, p-F of TiOC6F5), Ϫ166.0 (m, 6 F,
m-F of BC6F5).
1
H, 5.34%). H NMR (CD2Cl2): δ 1.86 (s, 15 H, C5Me5), 0.55
(s, 6 H, TiMe). 13C-{1H} NMR (CD2Cl2): δ 141.1 [d, J(CF ) =
231, o-CF], 138.7 [d, J(CF ) = 241, m-CF], 135.5 [d, J(CF ) = 241
Hz, p-CF], 124.2 (C5Me5), 59.0 (TiMe), 12.1 (C5Me5). 19F NMR
(CD2Cl2): δ Ϫ160.1 (m, 2 F, o-F ), Ϫ167.0 (m, 2 F, m-F ), Ϫ171.3
(t, 1 F, p-F ).
Ti(ç-C5Me5)Me(OC6F5)2 3. A solution of Ti(η-C5Me5)MeCl2
(0.50 g, 1.85 mmol) and LiOC6F5 (0.70 g, 3.72 mmol) in
hexanes (50 cm3) was stirred at 0 ЊC for 4 h and then warmed to
room temperature. The yellow suspension was filtered through
Celite and dried under vacuum. The bright yellow solid proved
difficult to purify and could only be isolated with small
amounts of impurities of Ti(η-C5Me5)(OC6F5)3. Yield 40%
(Found: C, 48.00; H, 3.49. Calc. for C23H18F10O2Ti: C, 48.86; H,
3.22%). 1H NMR (CD2Cl2): δ 1.89 (s, 15 H, C5Me5), 1.17 (3 H,
TiMe). 13C-{1H} NMR (CD2Cl2): δ 139.9 [d, J(CF ) = 251,
o-CF], 138.1 [d, J(CF ) = 241, m-CF], 135.2 [d, J(CF ) = 251 Hz,
p-CF], 127.0 (C5Me5), 61.5 (TiMe), 10.9 (C5Me5). 19F NMR
(CD2Cl2): δ Ϫ162.0 (m, 2 F, o-F ), Ϫ167.2 (m, 2 F, m-F ), Ϫ171.7
[t, 2J(FF ) = 21 Hz, 1 F, p-F].
[{Ti(ç-C5Me5)Me(OC6F5)}2(ì-Me)][BMe(C6F5)3] 11. In an
NMR experiment carried out as above, B(C6F5)3 (15 mg,
0.03 mmol) in CD2Cl2 (0.3 cm3) was added to a sol-
ution of Ti(η-C5Me5)Me2(OC6F5) (24 mg, 0.06 mmol)
in CD2Cl2 (0.4 cm3) cooled to Ϫ78 ЊC providing an orange
solution. 1H NMR (CD2Cl2):
δ 2.01 (30 H, C5Me5),
1.50 (3 H, TiMe), 1.47 (3 H, TiMe), 0.34 (br s, 3 H,
BMe), Ϫ0.31 (3 H, µ-Me). 13C-{1H} NMR (CD2Cl2):
δ 132.0 and 131.9 (C5Me5), 79.7 and 79.3 (TiMe), 12.3 and 12.2
(C5Me5). 19F NMR (CD2Cl2): Ϫ134.3 (m, 6 F, o-F of BC6F5),
Ϫ159.4 and Ϫ159.6 (m, 4 F, o-F of TiOC6F5), Ϫ163.6 and
Ϫ164.1 (m, 4 F, m-F of TiOC6F5), Ϫ165.0 (m, 2 F, p-F of
TiOC6F5), Ϫ165.0 (m, 3 F, p-F of BC6F5), Ϫ167.7 (m, 6 F, m-F
of BC6F5).
Ti(ç-C5Me5)(OC6HF4-p)3 4. A solution of Ti(η-C5Me5)Cl3
(1.00 g, 3.45 mmol) and LiOC6HF4-p (1.79 g, 10.40 mmol) in
3098
J. Chem. Soc., Dalton Trans., 1997, Pages 3097–3104

View Article Online
Table 1 Crystallographic data for compounds 4 and 5
Compound
4
5
Empirical formula
M
C28H18F12O3Ti
678.32
C44H34F16O5Ti2
1042.51
Crystal symmetry
Space group
Triclinic
P1 (no. 2)
Triclinic
P1 (no. 2)
¯
¯
a/Å
b/Å
c/Å
α/Њ
8.196(2)
10.583(2)
15.997(2)
91.939(12)
95.066(12)
87.123(12)
1379.9(4)
2
1.633
0.422
0.35 × 0.25 × 0.10
Orange
0.710 69
Ϫ8 р h р 8, 0 р k р 11,
Ϫ17 р l р 17
1.2–23.0
298(2)
11.3152(17)
12.6242(20)
17.7777(19)
106.297(9)
90.171(11)
113.476(13)
2216.6(5)
2
β/Њ
γ/Њ
U/Å3
Z
Dc/g cmϪ3
1.562
0.472
µ(Mo-Kα)/mmϪ1
Crystal dimensions/mm
Crystal color
λ(Mo-Kα)/Å
Data collected
0.40 × 0.25 × 0.15
Yellow-brown
0.710 69
0 р h р 12, Ϫ14 р k р 13,
Ϫ20 р l р 20
1.2–24.0
θ Scan range/Њ
T/K
298(2)
No. of reflections
4432
7327
No. of independent reflections
Final R
3818
0.0477
0.051, 0.27
0.0876, 0.1460
1.148
0.003
0.288
Ϫ0.252
6924
0.0514
0.0204, 0.938
0.1034, 0.1280
1.327
0.004
0.505
Ϫ0.406
a, b in weighting schemea
R,b wR2,c all independent reflections
Goodness of fit indicator d
Maximum shift/error, final cycle
Maximum resid. electron dens./e ÅϪ3
Minimum resid. electron dens./e ÅϪ3
¹
a
2
2
2
2
w = 1/[σ2(Fo ) ϩ aP2 ϩ bP] where P = (Fo2 ϩ 2Fc )/3. b R = Σ Fo| Ϫ |Fc /Σ|Fo|. c wR2 = {[Σw(Fo2 Ϫ Fc )2]/[Σw(Fo )2]}². d Goodness-of-fit on F 2,
¹
S = {[Σw(Fo2 Ϫ Fc )2]/(No Ϫ Nv)}², all independent data.
2
Table 2 Selected bond lengths (Å) and angles (Њ) for complex 4*
{Ti(ç-C5Me5)(OC6F5)2}{BMe(C6F5)3} 12. In an NMR experi-
ment carried out as above, B(C6F5)3 (31 mg, 0.06 mmol) in
CD2Cl2 (0.3 cm3) was added to a solution of Ti(η-C5Me5)-
Me(OC6F5)2 (34 mg, 0.06 mmol) in CD2Cl2 (0.4 cm3) cooled to
Ϫ78 ЊC to give a red solution. 1H NMR (CD2Cl2): δ 2.17 (15 H,
C5Me5), 0.37 (br s, 3 H, BMe). 13C-{1H} NMR (CD2Cl2):
δ 139.3 (C5Me5), 12.3 (C5Me5). 19F NMR (CD2Cl2): Ϫ134.7
(m, 6 F, o-F of BC6F5), Ϫ159.7 (m, 4 F, o-F of TiOC6F5),
Ϫ165.1 (t, 2 F, p-F of TiOC6F5), Ϫ164.2 (m, 4 F, m-F of
TiOC6F5), Ϫ166.4 (m, 3 F, p-F, of BC6F5), Ϫ169.0 (m, 6 F, m-F
of BC6F5).
Ti᎐Cg
Ti᎐O(2)
O(1)᎐C(11)
O(3)᎐C(31)
2.023(3)
1.826(3)
1.339(5)
1.317(5)
Ti᎐O(1)
Ti᎐O(3)
O(2)᎐C(21)
1.867(3)
1.830(3)
1.330(5)
Ti᎐O(1)᎐C(11)
Ti᎐O(3)᎐C(31)
O(2)᎐Ti᎐Cg
O(1)᎐Ti᎐O(2)
O(2)᎐Ti᎐O(3)
137.1(3)
156.6(3)
116.49(13)
106.2(2)
102.2(2)
Ti᎐O(2)᎐C(21)
O(1)᎐Ti᎐Cg
O(3)᎐Ti᎐Cg
O(1)᎐Ti᎐O(3)
165.0(3)
111.14(12)
118.39(13)
100.59(14)
* Cg is the center of gravity of the (η-C5Me5) ring.
Crystallography: structure solution and refinement
Table 3 Selected bond lengths (Å) and angles (Њ) for complex 5*
The structure of complex 4 was solved by direct methods using
the program SHELXS 86,29 and that of complex 5 by Patterson
and Fourier methods using the program SHELXS 86.29 The
pertinent crystallographic data are given in Table 1. For both
structures full-matrix least-squares refinement on F 2 data with
the anisotropic displacement parameters for all non-H atoms
was performed using the program SHELXL 9330 (the neutral-
atom scattering factors and anomalous dispersion corrections
used are taken from ref. 31).
In structure 4 the aromatic hydrogen atoms were placed in
calculated positions with C᎐H distances of 0.93 Å, and those of
(η-C5Me5) methyl groups with C᎐H 0.96 Å and orientations
based on features of the Fourier-difference map. In the refine-
ment, all H atoms were riding on the carbon atoms to which
they are attached with their isotropic thermal displacement
parameters kept as Uiso(H) = 1.2Ueq(C). The refinement for
structure 4 converged at the final R (conventional, based on F )
of 0.0477. There was no residual solvent accessible area in this
structure. Selected bond lengths and bond angles are listed in
Table 2.
Ti(1)᎐Cg(1)
Ti(1)᎐O(1)
Ti(1)᎐O(2)
Ti(1)᎐O(3)
2.042(3)
1.814(3)
1.838(3)
1.883(3)
Ti(2)᎐Cg(2)
Ti(2)᎐O(1)
Ti(2)᎐O(4)
Ti(2)᎐O(5)
2.038(3)
1.823(3)
1.844(3)
1.875(3)
Ti(1)᎐O(1)᎐Ti(2)
O(1)᎐Ti(1)᎐Cg(1)
O(2)᎐Ti(1)᎐Cg(1)
O(3)᎐Ti(1)᎐Cg(1)
O(1)᎐Ti(1)᎐O(3)
O(1)᎐Ti(1)᎐O(2)
160.2(2)
O(1)᎐Ti(2)᎐Cg(2)
O(4)᎐Ti(2)᎐Cg(2)
O(5)᎐Ti(2)᎐Cg(2)
O(1)᎐Ti(2)᎐O(4)
O(1)᎐Ti(2)᎐O(5)
116.82(15)
116.24(16)
110.31(15)
103.45(13)
102.63(14)
118.23(15)
114.48(14)
110.29(14)
104.58(13)
102.51(13)
* Cg is the center of gravity of the (η-C5Me5) ring.
were placed in calculated positions with C᎐H distances of 0.93
Å, and those of (η-C5Me5) methyl groups with C᎐H 0.96 Å and
orientations based on features of the Fourier-difference map. In
the refinement, all H atoms were riding on the carbon atoms to
which they are attached with their isotropic thermal displace-
ment parameters kept as Uiso(H) = 1.2Ueq(C). The refinement
for structure 5 converged at the final R (conventional, based on
F ) of 0.0514. There was no residual solvent accessible area in
this structure. Selected bond lengths, and bond angles are listed
In structure 5 atoms C(54), C(55), F(55), C(56) and F(56) of
one of the tetrafluorophenyl rings show large anisotropy con-
sistent with some slight disorder. The aromatic hydrogen atoms
J. Chem. Soc., Dalton Trans., 1997, Pages 3097–3104
3099

View Article Online
unselective, giving inseparable mixtures of mono-, di- and
tri-substituted compounds. Often the final product can only be
recovered pure when all three Ti᎐Me bonds are transformed,
and facile formation of Ti(η-C5Me5)Me2(OR) via protonolysis
has been limited to alcohols ROH with bulky R groups (SiPh3,34
CPh3,24 But 24 or Pri 24).
As an alternative, more selective methodology Royo and co-
workers34,35 have described the isolation of partially methylated
titanium derivatives Ti(η-C5Me5)MenCl3Ϫn (n = 1–3) by treating
Ti(η-C5Me5)Cl3 with methyllithium or MeMgX (X = Cl or Br)
[equation (4)]. We have used this latter route to synthesize the
Ti(η-C5Me5)Cl3 ϩ nMeLi →
Ti(η-C5Me5)MenCl3Ϫn ϩ nLiCl (4)
desired substituted complexes Ti(η-C5Me5)Me3ϪnEn. Treatment
of Ti(η-C5Me5)Cl3 with 1 equivalent of MeLi in hexanes pro-
vides the yellow complex Ti(η-C5Me5)MeCl2, while addition of
2 equivalents in tetrahydrofuran (thf) gives the orange com-
pound Ti(η-C5Me5)Me2Cl.34,35 We also find that reacting
Ti(η-C5Me5)Cl3 with Ti(η-C5Me5)Me3 in a molar ratio of 1:2 at
room temperature in hexanes provides a clean, high yield route
to Ti(η-C5Me5)Me2Cl via comproportionation. Subsequent
replacement of the chloride ligands in Ti(η-C5Me5)Me3ϪnCln
is then easily achieved by addition of an appropriate amount
of the lithium salt of the desired substituent ELi in hexanes,
resulting in LiCl precipitation and the formation of Ti(η-
C5Me5)Me3ϪnEn (n = 1, E = C6F5 1 or OC6F5 2; n = 2, E = OC6F5
3; n = 3, E = OC6HF4-p 4) [equation (5)].
Fig. 1 An ORTEP plot of complex 4
Ti(η-C5Me5)Me3ϪnCln ϩ nELi →
Ti(η-C5Me5)Me3Ϫn(E)n ϩ nLiCl (5)
Compounds 1–4 have been fully characterized by elemental
analyses and spectroscopic methods. Of some interest is the
observation of long range spin–spin coupling (2 Hz) between
the Ti᎐Me resonances and the ortho fluorine atoms of 1. While
the chemical shifts of the C5Me5 resonances of this series of
compounds vary little (<0.1 ppm), the TiMe resonances do.
Indeed, substitution of one of the TiMe methyl groups of Ti(η-
C5Me5)Me3 by the much more electronegative ligand C6F5
results in a downfield shift of 0.52 ppm. In contrast, substit-
ution by the OC6F5, OPri and OBut groups results in upfield
shifts of 0.11, 0.70 and 0.75 ppm, respectively, presumably
because of π donation from the oxygen lone pairs to the vacant
d orbitals of the titanium centre. Previous studies show that
alkoxy36 and aryloxy34 ligands readily donate electrons to the
titanium center with their lone pair orbitals conveniently sit-
uated to overlap with empty metal d orbitals, forming partial
double bonds. For the more electronegative OC6F5 ligand,
the donating ability of the oxygen is apparently suppressed,
resulting in a more electrophilic metal centre with respect to the
non-fluorinated alkoxy compounds.
Fig. 2 An ORTEP plot of complex 5
in Table 3. The geometrical analysis of the structures was done
with programs SHELXL 9330 and PLATON.32
The ORTEP 33 drawings of molecules 4 and 5, with the atom
numbering schemes, are given in Figs. 1 and 2 respectively.
CCDC reference number 186/641.
Results and Discussion
Preparations and general characterization
There are several synthetic routes which one can take to syn-
thesize derivatives of Ti(η-C5Me5)Me3. For instance, addition
of aldehydes or ketones to Ti(η-C5Me5)Me3 results in insertion
Titanium NMR spectroscopy
Titanium NMR spectroscopy is becoming increasingly useful
as an investigative tool for assessing electronic effects of ligand
substitution in titanium compounds,6,37–46 including systems of
interest in catalytic processes.6,44–46 Although spectral acquisi-
tion of the NMR active 47Ti and 49Ti nuclei is generally hin-
dered by low natural abundances (7.75%, 5.51%, respectively)
of a C᎐O group into a Ti᎐Me bond to produce secondary and
᎐
tertiary alkoxy compounds, respectively [equations (1) and
(2)].24 Protonolysis of Ti(η-C5Me5)Me3 by alcohols gives similar
Ti(η-C5Me5)Me3 ϩ MeCHO →
Ti(η-C5Me5)Me2(OCHMe2) (1)
and low sensitivities (both ≈10Ϫ3 that of H), as well as by quad-
rupolar broadening of resonances of these two nuclei (I = ,
1
5
7
– –
2
2
Ti(η-C5Me5)Me3 ϩ Me2CO →
Ti(η-C5Me5)Me2(OCMe3) (2)
respectively), useful spectra of titanium() complexes contain-
ing a wide variety of ligands have been reported.6,37–46 Since the
47Ti and 49Ti nuclei have almost identical magnetogyric ratios,
resonances of both are readily apparent in the spectra of most
compounds, with separations ≈266 ppm. In practice the 49Ti
resonances are easier to observe because of greater quadru-
polar broadening of the 47Ti resonances, although the chemical
products [equation (3)], but all three processes are quite
Ti(η-C5Me5)Me3 ϩ nROH →
Ti(η-C5Me5)Men(OR)3Ϫn ϩ nCH4 (3)
3100
J. Chem. Soc., Dalton Trans., 1997, Pages 3097–3104

View Article Online
shifts of both are of course identical when referenced to the
corresponding isotopomer of TiCl4, the usual reference.
Chemical shifts are usually described by the Ramsey equa-
tion, which describes the overall shielding σ as in equation (6);
σ = σd ϩ σp ϩ σo
(6)
here σd is the diamagnetic shielding, σp is the paramagnetic
shielding and σo is a contribution from other atoms.47 However,
σp seems generally to dominate the chemical shifts of
transition-metal nuclei, and equation (6) has been simplified to
the form shown in equation (7). Here A and B are constants
σ = A Ϫ (B/∆)
(7)
for complexes with similar ligands and ∆ is the average energy
difference between the ground state and low-lying electronic
excited states.47 To date, most 47/49Ti NMR chemical shift data
have been interpreted in terms of equation (7),6,37–46 with linear
relationships between the chemical shifts and the longest wave-
length transition in the electronic spectra being observed for
several series of compounds.6,37,38,46
Fig.
3
Plot of λmax against δ49Ti: A = Ti(η-C5Me5)Cl3, B = Ti-
(η-C5Me5)MeCl2, C = Ti(η-C5Me5)Me2Cl, D = Ti(η-C5Me5)Me2(C6F5),
E = Ti(η-C5Me5)Me3, F = Ti(η-C5Me5)Me2(OC6F5), G = Ti(η-C5Me5)-
Me(OC6F5)2
Table 4 Titanium-47/49 NMR data *
Indeed, reasonably linear correlations between 47/49Ti chem-
ical shifts and the longest wavelength electronic transitions have
been found for each of the series of compounds TiX4 (X = Cl,
Br, I, OPri or NEt2),37,38 CpЈ2TiX2 (CpЈ = C5H5 or substituted
C5H5; X = F, Cl, Br, I, N3 or NCS)38,44,46 and (indЈ)TiCl3
(indЈ = substituted indenyl),6 although the three series of com-
pounds do not exhibit the same relationship. Differences in the
nature of the ligands apparently result in significant variations
in A and/or B of equation (7).45,46 However, for the individual
series TiX4 and CpЈ2TiX2, equation (7) does generally apply and
results in chemical shifts becoming less shielded in the order
X = F > Cl > Br > I.47 These various types of TiIV complexes
thus apparently exhibit a direct correlation between 47/49Ti
shielding and ligand electronegativities, a relationship which
may possibly be fortuitous but which has now been utilized
extensively. Indeed, several correlations of titanium chemical
shifts with electron-donating properties of CpЈ ring substit-
uents have recently been reported.6,40,45,46
Compound
δ49Ti ∆ν¹49Ti
δ47Ti ∆ν¹47Ti λmax
¯
²
¯
²
6 Ti(η-C5Me5)Me3
551
484
343
143
69
317
1664
567
143
620
11
288
nd
Ϫ121
nd
nd
Ϫ354
nd
902
nd
1056
nd
nd
322
330
348
368
338
440
340
1 Ti(η-C5Me5)Me2(C6F5)
7 Ti(η-C5Me5)Me2Cl
8 Ti(η-C5Me5)MeCl2
2 Ti(η-C5Me5)Me2(OC6F5)
Ti(η-C5Me5)Cl3
Ϫ88
27
nd
3 Ti(η-C5Me5)Me(OC6F5)2 Ϫ380
460
* nd = Not determined.
linear decrease in the chemical shift with an increase in λmax as
methyl ligands are substituted by chloride, a ligand of lower
ligand field strength. The opposite trend is found for the series
of compounds TiX4,36,37 CpЈ2TiX2
and (indЈ)TiCl3, 6 as
38,44,46
well as for many other types of transition-metal systems where
strong field ligands generally result in shifts to higher field.47
While the reasons for the anomalies are not known, we note
that 195Pt chemical shifts also often exhibit a poor correlation
which is attributed to a number of influences, mainly the vari-
ation of B with ligand substitution.48–51 The term B relates to
the angular momentum of valence orbitals, and is very sensitive
to changes in the distance of the d electrons/orbitals from the
metal nucleus (B ∝ 〈rϪ3〉d).48
With a view to possibly assessing futher the relative electro-
philicities of the compounds Ti(η-C5Me5)Me2E and related
species, we have obtained 47/49Ti NMR data for several of the
compounds reported here. As yet, few Ti NMR data have been
available for monocyclopentadienyl titanium compounds of the
6,40,44,45
type CpЈTiX3
and, in general, chemical shifts of substi-
As a group, the methyl-substituted complexes 6, 7 and 8 are
the most deshielded of any titanium compounds yet known,
more so even than the sulfur-bound compounds TiClx-
(S2CNR2)4Ϫx (x = 2–0).42 Substituting methyl ligands with the
oxygen-donor ligand OC6F5 causes an upfield shift 6 (551) > 2
(69) > 3 (Ϫ380 ppm). Such an upfield shift upon increasing the
number of oxygen-donor ligands is also apparent for the com-
pounds CpTiCl3 (δ Ϫ390) and CpTi(2-PrO)2Cl (δ Ϫ917),44 and
for the series of compounds TiClx(OEt)4Ϫx [x = 4 (δ 0), 3
(Ϫ360), 2 (Ϫ560), 1 (Ϫ750) or 0 (Ϫ825)].42
tuted cyclopentadienyl compounds CpЈTiCl3 are found in the
region δ Ϫ270 to Ϫ390.44,45 However, chemical shifts of similar
indenyl compounds are found in the range δ Ϫ114 to Ϫ269,6
that of Ti(η-C5Me5)Cl3 at δ Ϫ85.44,45 Only a single report for
alkyltitanium compounds has appeared, the chemical shifts of
the compounds TiMeCl3 and TiMeBr3 being reported at δ 618
and δ 825, respectively,43 intermediate between those of TiBr4
and TiI4.38 We have also, therefore, in the course of this work
obtained further information on the effects of alkyl ligands on
47/49Ti chemical shifts.
The 47/49Ti NMR data obtained for the series of compounds
Ti(η-C5Me5)Me2E (E = Me 6, Cl 7, OC6F5 2 or C6F5 1), Ti-
(η-C5Me5)MeE2 (E = Cl 8, OC6F5 3) and Ti(η-C5Me5)Cl3 are
shown in Table 4. The individual 47Ti and 49Ti resonances were
observed in most cases, the former being relatively broad, and
there is good agreement between the observed 47/49Ti chemical
shift and linewidths (∆ν¹) of Ti(η-C5Me5)Cl3 and previously
Crystal structures of Ti(ç-C5Me5)(OC6HF4-p)3 4 and
{Ti(ç-C5Me5)(OC6HF4-p)2}2(ì-O) 5
Compound 4 was characterized crystallographically because we
wished to assess the possibility of alkoxy ligand π donation in
compounds of this type but could not obtain satisfactory crys-
tals of Ti(η-C5Me5)(OC6F5)3. The structure of 5 was obtained
because our initial attempt to grow crystals of 4 produced
instead 5, the product of hydrolysis by adventitious water; the
structure is reported because it was thus available for purposes
of comparison and complements the structure of 4. Fig. 1
shows a perspective of the molecular structure of 4, Table 2
most important bond distances and angles. Compound 4 is
monomeric, and assumes a classical three-legged piano stool
structure. The (η-C5Me5 centroid)᎐Ti᎐O angles vary from
¯
²
published values.45 The λmax values for the lowest energy elec-
tronic transitions of Ti(η-C5Me5)Cl3 and the compounds 1–3
and 6–8 are also listed in Table 4, and a plot of chemical shifts
vs. λmax is shown in Fig. 3. As can be seen, the expected relation-
ship between δ(49Ti) and λmax is not observed for these com-
pounds, suggesting that the relationship shown in equation (7)
does not apply well. In fact, the plot in Fig. 3 illustrates a large
downfield shift for the trimethyl compound 6 (δ 551) and a
J. Chem. Soc., Dalton Trans., 1997, Pages 3097–3104
3101

View Article Online
especially characteristic of the proposed zwitterionic structures
because of their linewidths, a consequence of quadrupolar
relaxation by the 10,11B nuclei.60 Shielding of the terminal
methyl protons of the compounds Ti(η-C5Me5)MeE(µ-Me)-
B(C6F5)3 clearly decreases in the order E = Me (δ 1.53) > OC6F5
(1.89) > C6F5 (2.61), consistent with the relative electro-
negativities of the groups E and strongly implying a high degree
of electrophilicity of the metal in compounds 9 and 10.
However, while shielding of the MeB protons decreases in a
different order, E = OC6F5 (δ 0.62) Me > (1.19) > C6F5 (1.35),
solution and crystallographic studies of the similar methyl-
zirconocene compounds CpЈ2ZrMe(µ-Me)B(C6F5)3 indicate
that the MeB resonances become increasingly deshielded as the
strength of the borate–zirconium interaction, affected by both
electronic and steric factors, increases. Thus it seems likely that
the metal–borate interaction in 10 may be weakened by steric
interactions.
Spin saturation experiments have previously been carried out
on a solution of Ti(η-C5Me5)Me2(µ-Me)B(C6F5)3, the Ti᎐Me
and µ-Me resonances being irradiated and difference spectra
being inspected for transfer of magnetization.17 Interestingly,
while no transfer between these two resonances was observed,
implying that the B(C6F5)3 does not ‘hop’ from one methyl
group to another, irradiation of the µ-Me resonance resulted in
enhancement of a vanishingly weak resonance at δ ≈0.38, the
chemical shift of free borate anion. It follows that the borate
anion, [BMe(C6F5)3]Ϫ, of Ti(η-C5Me5)Me2(µ-Me)B(C6F5)3 dis-
sociates to some extent in solution, and that a low, steady-state
concentration of the (possibly solvated) cation [Ti(η-C5Me5)-
Me2]ϩ exists in solution. Similar NMR experiments performed
on 9 and 10 over the temperature range Ϫ50 to Ϫ10 ЊC pro-
vided no evidence for borate dissociation in the former but did
result in magnetization transfer, and hence evidence for free
borate ion, in the spectrum of 10. This result indicates an equi-
librium between the methyl-bridged species and the solvent
separated ions, [Ti(η-C5Me5)Me(OC6F5)]ϩ and [BMe(C6F5)3]Ϫ,
in the case of 10. Thus the strength of the titanium–borate
interaction in the compounds Ti(η-C5Me5)MeE(µ-Me)B(C6F5)3
decreases in the order E = C6F5 > OC6F5 = Me, correlating
reasonably well with the apparent electrophilicities of the com-
plexes [Ti(η-C5Me5)MeE]ϩ and providing an indication of the
relative amounts present in solution of these cations, believed to
be the active species in olefin polymerizations.1,2,9
Ti+
Me
H
H
C6F5
E
C
B
C6F5
C6F5
H
Fig. 4 Structure proposed for Ti(η-C5Me5)MeE(µ-Me)B(C6F5)3 (E =
Me, C6F5 or OC6F5)
111.14 for O(1) to 118.4Њ for O(3), similar to those found for
Ti(η-C5Me5)(NMe2)3, where the (η-C5Me5 centroid)᎐Ti᎐N
angles range from 112 to 119Њ.52 The O᎐Ti᎐O angles lie between
100.59(14) and 106.2(2)Њ, while the Ti᎐C (C5Me5) distances are
in the normal range. The angles between titanium, oxygen and
the ipso carbon of the aryl rings vary significantly, i.e.
Ti᎐O(3)᎐C(31) = 156.6(3), Ti᎐O(2)᎐C(21) = 165.0(3), Ti᎐O(1)᎐
C(11) = 137.1(3)Њ.
Large Ti᎐O᎐C (aryl) angles and short Ti᎐O (aryl) distances
would be indicative of double bond character for the Ti᎐O
bond due to π-electron donation from the alkoxy ligands.53
Indeed, the Ti᎐O bond in Ti(η-C5Me5)Cl2(OC6H3Me2-2,6)
[1.785(2) Å, 162.3Њ]34 is believed to have significant multiple
character in contrast to that of Ti(η-C5Me5)2{OC6H2But-2,6-
Me-4} [1.892(2) Å, 142.3Њ].53 However, the Ti᎐O᎐aryl bond
angles of 4 vary considerably, and it appears distortions are
very facile, possibly because of the steric effects, and it would
seem that the effects of π-electron donation in 4 must be min-
imal. All other bond lengths and angles are normal.54
Fig. 2 shows a perspective of the molecular structure, Table 3
the most important bond distances and angles of the oxygen
bridged compound {Ti(η-C5Me5)(OC6HF4-p)2}2(µ-O) 5. The
molecular structure of 5 consists of two identical Ti(η-
C5Me5)(OC6HF4-p)2 units bridged by an oxygen atom situated
such that there is a trans arrangement of the (η-C5Me5) rings.
As with 4, the OR ligands in 5 exhibit quite different Ti᎐O᎐R
bond angles [Ti(1)᎐O(2)᎐C(21) = Ti(1)᎐O(4)᎐C(41) ≈161,
Ti(2)᎐O(3)᎐C(31) = Ti(2)᎐O(5)᎐C(51) ≈144Њ]. Significant
π
bonding involving the µ-O ligand might also be anticipated, and
would be implied by shortened Ti᎐O bonds and a linear or near
linear arrangement of the Ti᎐O᎐Ti bonds.55 However the Ti᎐O
bond lengths and Ti᎐O᎐Ti bond angles in {Ti(η-C5H5)2-
(CN)}2(µ-O) [1.840(2) Å, 174.3(1)Њ],56 {Ti(η-C5H5)2(Et)}2(µ-O)
[1.840(3) Å, 173.7(2)Њ],57 {Ti(η-C5H5)2}2(µ-O) [1.838(1) Å,
170.9(4)Њ]58 and {Ti(η-C5H5)2(Cl)2}2(µ-O) [1.809(9) Å,
167.5(6)Њ]59 do not correlate as expected, and thus the deviation
of the Ti(1)᎐O(1)᎐Ti(2) bond angle (160.2Њ) from a linear
arrangement and the titanium–oxygen distances of 1.814 Å for
Ti(1)᎐O(1) and 1.823 Å for Ti(2)᎐O(1) are in the range expected
for titanium–oxo bridged dimers (1.820 Å),54 and little can be
said about the Ti᎐O bonding. All other bond lengths and angles
are normal.54
The 19F NMR spectrum of complex 9 at Ϫ50 ЊC exhibits five
resonances of equal intensity and assignable to five magnetic-
ally inequivalent fluorine environments in the C6F5 ligands. The
ortho and meta resonances broaden as the temperature is raised
but the compound decomposes before coalescence is reached,
preventing acquisition of accurate site exchange data. However,
using the dynamic NMR simulation program DNMR 5,61 the
spectra were fitted for a number of temperatures and the
data did provide activation data for the dynamic process(es)
involved,62 and a least-squares fitting of the data resulted in
determination of activation enthalpies and entropies: ∆H ‡
=
Zwitterionic complexes
4.5 ± 0.1 kcal molϪ1 (cal = 4.184 J) and ∆S ‡ = Ϫ25.8 ±
0.8 cal KϪ1 molϪ1 for the ortho F resonances; ∆H ‡ = 3.2 ± 0.1
kcal molϪ1 and ∆S ‡ = Ϫ32.6 ± 0.8 cal KϪ1 molϪ1 for the meta
F resonances. The coalescence temperatures for the exchange
of the ortho and meta fluorines are 11.7 ± 1 ЊC and 1.0 ±
1 ЊC, respectively and, using equation (8) for an equally
Treatment of solutions of Ti(η-C5Me5)Me2(C6F5) 1 and Ti-
(η-C5Me5)Me2(OC6F5) 2 with 1 equivalent each of the borane
B(C6F5)3 in CD2Cl2 at Ϫ78 ЊC results in immediate colour
changes from yellow to red and the formation of the zwit-
terionic complexes Ti(η-C5Me5)MeE(µ-Me)B(C6F5)3 (E = C6F5
9, OC6F5 10) (Fig. 4). These complexes decompose above 10 ЊC
and could not be isolated, but were characterized by H, 13C
∆Gc‡ = aT [9.972 ϩ log(T/δν)]
(8)
1
and 19F NMR spectroscopic studies (CD2Cl2, Ϫ50 to 10 ЊC);
the structure proposed, with two bridging hydrogen atoms,
is based on metallocene precedents.2,9 Unfortunately, 47/49Ti
resonances could not be detected, presumably because of severe
quadrupolar broadening.60 Compounds 9 and 10 were also
readily identified on the basis of comparison with the pre-
viously characterized Ti(η-C5Me5)Me2(µ-Me)B(C6F5)3.17 The
broadened µ-MeB resonances of all three compounds are
populated two-site system, good agreement was obtained
for the two calculated free energy of activation values:
∆Gc‡ = 11.8 ± 0.4 kcal molϪ1 and 12.2 ± 0.2 kcal molϪ1
respectively.
Possible reasons for the non-equivalences are restricted rot-
ation of the C6F5 ring about the Ti᎐C6F5 bond and/or borate
ligand dissociation/reorganisation; the latter is observed to
3102
J. Chem. Soc., Dalton Trans., 1997, Pages 3097–3104

View Article Online
+
for which rather similar exchange is implied by variable-
temperature NMR studies although not the type of shift in
equilibrium found here. As pointed out above, [(η-C5Me5)-
Ti(OC6F5)2]ϩ is expected to be a relatively strong Lewis acid
and its apparently unusual disinclination to bind the [BMe-
(C6F5)3]Ϫ anion is to be attributed to steric hindrance by the
three bulky ligands on the titanium hindering co-ordination
of the latter. Indeed, it appears that the strong but sterically
hindered Lewis acid [(η-C5Me5)Ti(OC6F5)2]ϩ effectively com-
petes with B(C6F5)3 for possession of the methyl group, pre-
sumably via a transient methyl-bridged species although none
was detected in the spin saturation experiments.
OC6F5
Me
Ti
Ti+
[BMe(C6F5)3]–
Me
Me
F5C6O
Fig. 5 Structure proposed for [{Ti(η-C5Me5)Me(OC6F5)}2(µ-Me)]-
[BMe(C6F5)3]
The results reported here are in accord with the findings of
Marks and co-workers5,7–10 that bulky substituents at a metal
center can result in reduced [metal cation]–anion interactions in
spite of the intrinsic electrophilicity of the metal center. Com-
pound 12 presents an extreme case where the steric effects com-
pletely overcome any strong electrostatic interactions.
occur at higher temperatures for the metallocenes CpЈ2MMe(µ-
Me)B(C6F5)3 (CpЈ = substituted cyclopentadienyl, M = Zr or
Hf), with free energies of activation of 13.5 to >19 kcal
molϪ1.7,10,11 Since, as shown above, spin saturation transfer
experiments on 9 show no sign of TiMe–MeB methyl exchange,
it seems that the site exchange phenomenon observed is a result
of restricted rotation of the C6F5 ring rather than an intra/
intermolecular methyl exchange between the borane and the
metal.
Addition of 0.5 equivalent of borane to 2 or of 1 equivalent
of 2 to 10 results in formation of an unstable species which, on
the basis of its NMR spectrum, is the dititanium complex
[{Ti(η-C5Me5)Me(OC6F5)}2(µ-Me)][BMe(C6F5)3] 11 (Fig. 5);
this contains two chiral titanium centres and thus exists as
diastereomers. The bridged methyl resonance of 11 is shifted
upfield by 2 ppm relative to the terminal TiMe resonance of
10, a characteristic of Group 4 metal methyl-bridged com-
pounds of the type CpЈ2ZrMe(µ-Me)ZrMeCpЈ2.8,17,63,64 The
diastereomeric nature of 11 is attested to by the 1H NMR TiMe
resonances at δ 1.50 and 1.47, and by pairs of resonances in the
13C-{1H} NMR spectrum: C5Me5 (δ 132.0 and 131.9), TiMe
(δ 79.7 and 79.3), C5Me5 (δ 12.3 and 12.2). The 19F NMR spec-
trum also shows pairing of resonances for both the o-fluorines
(δ Ϫ159.4 and Ϫ159.6) and the m-fluorines of TiOC6F5
(δ Ϫ163.6 and Ϫ164.1) although the p-fluorine resonance is
obscured by that of MeB(C6F5)3. The pairs of diastereomers
occur in about a 1:1 ratio, as all of the above-mentioned pairs
of resonances are of equal intensity. Irradiation experiments
(Ϫ50 ЊC, CD2Cl2) show the bridged and terminal methyl groups
to be in exchange, a result found for the similar system [{Ti(η-
C5Me5)Me2}2(µ-Me)][BMe(C6F5)3].17
Treatment of Ti(η-C5Me5)Me(OC6F5)2 3 with 1 equivalent of
B(C6F5)3 (Ϫ50 ЊC, CD2Cl2) results in an immediate colour
change from yellow to deep red. The 1H NMR spectrum shows
the formation of the complex [Ti(η-C5Me5)(OC6F5)2]-
[BMe(C6F5)3] 12, with a downfield shift of the (η-C5Me5) res-
onance of 3 (δ 1.89) to δ 2.17 and a broad methyl resonance at
δ 0.37, attributable to free borate, [BMe(C6F5)3]Ϫ. Thus, in con-
trast to the methyl-bridged complexes 9 and 10, complex 12
shows no co-ordination of the methyl group to the metal, sur-
prising given the apparent eletrophilic nature of the titanium
centre. Thus the inability of [Ti(η-C5Me5)(OC6F5)2]ϩ to bind
the borate anion must be attributed to steric hindrance by the
three bulky ligands on the titanium, which would arguably have
greater steric requirements than the ligands of 9 and 10. Inter-
estingly, warming the solution results in the reappearance of the
resonances of 3, and at Ϫ10 ЊC there are substantial amounts
of both species present. Spin saturation transfer and variable-
temperature experiments show them to be in equilibrium
[equation (9)], with ∆H = Ϫ1.25 ± 0.1 kJ molϪ1 and ∆S =
Acknowledgements
We thank the Natural Sciences and Engineering Research
Council of Canada (Research and Strategic Grants to M. C. B.
and Graduate Scholarship to S. W. E.) and Alcan (Graduate
Scholarship to S. W. E.) for financial support.
References
1 H. H. Brintzinger, D. Fischer, R. Mülhaupt, B. Rieger and
R. M. Waymouth, Angew. Chem., Int. Ed. Engl., 1995, 34, 1143.
2 M. Bochmann, J. Chem. Soc., Dalton Trans., 1996, 255.
3 P. C. Möhring and N. J. Coville, J. Organomet. Chem., 1994, 479, 1.
4 S. L. Borkowsky, N. C. Brenzinger and R. F. Jordan, Organometal-
lics, 1993, 12, 486.
5 L. Jia, X. Yang, C. L. Stern and T. J. Marks, Organometallics, 1997,
16, 842.
6 Y. Kim, B. H. Koo and Y. Do, J. Organomet. Chem., 1997, 527,
155.
7 C. Sishta, R. M. Hathorn and T. J. Marks, J. Am. Chem. Soc., 1992,
114, 1112.
8 M. A. Giardello, M. S. Eisen, C. L. Stern and T. J. Marks, J. Am.
Chem. Soc., 1995, 117, 12 114.
9 X. Yang, C. L. Stern and T. J. Marks, J. Am. Chem. Soc., 1994, 116,
10 015.
10 P. A. Deck and T. J. Marks, J. Am. Chem. Soc., 1995, 117, 6128.
11 A. R. Siedleand R. A. Newmark, J. Organomet. Chem., 1995, 497, 119.
12 D. J. Gillis, M.-J. Tudoret and M. C. Baird, J. Am. Chem. Soc., 1993,
115, 2543.
13 D. J. Gillis, R. Quyoum, M.-J. Tudoret, Q. Wang, D. Jeremic,
A. W. Roszak and M. C. Baird, Organometallics, 1996, 15, 3600.
14 M. C. Murray and M. C. Baird, J. Mol. Catal., in the press.
15 Q. Wang and M. C. Baird, Macromolecules, 1995, 28, 8021.
16 Q. Wang, R. Quyoum, D. J. Gillis, M.-J. Tudoret, D. Jeremic,
B. K. Hunter and M. C. Baird, Organometallics, 1996, 15, 693.
17 Q. Wang, R. Quyoum, D. J. Gillis, D. Jeremic and M. C. Baird,
J. Organomet. Chem., 1997, 527, 7.
18 R. Quyoum, Q. Wang, M.-J. Tudoret, M. C. Baird and D. J. Gillis,
J. Am. Chem. Soc., 1994, 116, 6435.
19 S. W. Ewart, M. J. Sarsfield, T. L. Tremblay, A. W. Roszak and
M. C. Baird, unpublished work.
20 T. L. Tremblay, S. W. Ewart, M. J. Sarsfield and M. C. Baird, Chem.
Commun., 1997, 831.
21 F. Barsan and M. C. Baird, J. Chem. Soc., Chem. Commun., 1995,
1065.
22 C. Pellecchia, D. Pappalardo, L. Oliva and A. Zambelli, J. Am.
Chem. Soc., 1995, 117, 6593.
23 I. M. Lee, W. J. Gauthier, J. M. Ball, B. Iyengar and S. Collins,
Organometallics, 1992, 11, 2115.
24 D. J. Gillis, Ph.D. Thesis, Queens University, Kingston, 1993.
25 G. Hidalgo-Llinás, M. Mena, F. Palacios, P. Royo and R. Serrano,
J. Organomet. Chem., 1988, 340, 37.
26 A. Martín, M. Mena, M. A. Pellinghelli, P. Royo, R. Serrano and
A. Tiripicchio, J. Chem. Soc., Dalton Trans., 1993, 2117.
27 M. Mena, P. Royo, R. Serrano, M. A. Pellinghelli and A.
Tiripicchio, Organometallics, 1989, 8, 476.
3 ϩ B(C6F5)3
[Ti(η-C5Me5)(OC6F5)2][BMe(C6F5)3] (9)
Ϫ46 ± J KϪ1 molϪ1.
The ability of 12 to engage in Ti᎐Me/B᎐Me exchange is
unique among the mono(η-C5Me5) compounds studied here,
and also stands in contrast to some zirconocene systems,5,7–10
28 (a) A. G. Massey and A. J. Park, J. Organomet. Chem., 1964, 2, 245;
(b) R. Uson and A. Laguna, Inorg. Synth., 1982, 21, 72.
J. Chem. Soc., Dalton Trans., 1997, Pages 3097–3104
3103

View Article Online
29 G. M. Sheldrick, SHELXS 86, Program for the Solution of Crystal
Structures, University of Göttingen, 1985.
30 G. M. Sheldrick, SHELXS 92, Program for the Refinement of
Crystal Structures, University of Göttingen, 1992.
31 A. J. C. Wilson, in International Tables for Crystallography, ed.
A. J. C. Wilson, Kluwer, Dordrecht, 1992.
32 A. L. Speck, Acta Crystallogr., Sect. A, 1990, 46 (Suppl.), C34.
33 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National
Laboratory, Oak Ridge, TN, 1976.
34 P. M. Gómez-Sal, M. Mena, P. Royo and R. Serrano, J. Organomet.
Chem., 1988, 358, 147.
48 R. G. Kidd, Annu. Rep. NMR Spectrosc., 1980, 10, 1.
49 R. R. Dean and J. C. Green, J. Chem. Soc. A, 1968, 3047.
50 P. L. Goggin, R. J. Goodfellow, S. R. Haddock, B. F. Taylor and
R. H. Marshall, J. Chem. Soc., Dalton Trans., 1976, 459.
51 A. Pidcock and R. E. Richards, J. Chem. Soc. A, 1968, 1970.
52 A. Martin, M. Mena, C. Yelamos, R. Serrano and P. R. Raithby,
J. Organomet. Chem., 1994, 467, 79.
53 J. W. Lauher and R. Hoffmann, J. Am. Chem. Soc., 1976, 98, 1729.
54 A. G. Orpen, L. Brammer, F. H. Allen, O. Kennard, D. G. Watson
and R. Taylor, J. Chem. Soc., Dalton Trans., 1989, S1.
55 B. Cetinkaya, P. B. Hitchcock, M. L. Lappert, S. Torroni,
J. L. Atwood, W. E. Hunter and M. J. Zaworotko, J. Organomet.
Chem., 1980, 188, C31.
35 P. M. Gómez-Sal, A. Martín, M. Mena, P. Royo and R. Serrano,
J. Organomet. Chem., 1991, 419, 77.
36 J. C. Huffman, K. G. Maloy, J. A. Marsella and K. G. Caulton,
J. Am. Chem. Soc., 1980, 102, 3009.
37 R. G. Kidd, R. W. Mathews and H. G. Spinney, J. Am. Chem. Soc.,
1972, 94, 6686.
38 N. Hao, B. G. Sayer, G. Denes, D. F. Bickley and M. J. McGlinchey,
J. Magn. Reson., 1982, 50, 50.
39 P. G. Gassmann, W. H. Campbell and D. W. Macomber, Organo-
metallics, 1984, 3, 385.
40 A. Dormond, M. Fauconet, J. C. Leblanc, and C. Moise,
Polyhedron, 1984, 3, 897.
41 K. M. Chi, S. R. Frerichs, S. B. Philson and J. E. Ellis, J. Am. Chem.
Soc., 1988, 110, 303.
42 P. R. Trail and C. G. Young, J. Magn. Reson., 1990, 90, 551.
43 S. Berger, W. Bock, C. F. Marth, B. Raguse and M. T. Reetz, Magn.
Reson. Chem., 1990, 28, 559.
56 U. Thewalt and W. Nuding, J. Organomet. Chem., 1996, 512, 127.
57 H. G. Alt, K.-H. Schwind, M. D. Rausch and U. Thewalt,
J. Organomet. Chem., 1988, 349, C7.
58 B. Honold, U. Thewalt, M. Herberhold, H. G. Halt, L. B. Kool and
M. D. Rausch, J. Organomet. Chem., 1986, 314, 105.
59 P. Gowik, T. Klapötke and J. Prickardt, J. Organomet. Chem., 1990,
393, 343.
60 R. K. Harris, in Nuclear Magnetic Resonance Spectroscopy, Pitman,
London, 1983.
61 D. S. Stephenson and G. Binsch, Quantum Chemistry Program
Exchange, 1978, 11, 365.
62 J. Sandström, Dynamic NMR Spectroscopy, Academic Press,
New York, London, 1982, pp. 93–100.
63 M. Bochmann and S. J. Lancaster, Angew. Chem., Int. Ed. Engl.,
1994, 33, 1634.
44 A. Hafner, R. O. Duthaler, R. Marti, G. Rihis, P. Rothe-Streit and
F. Schwarzenbach, J. Am. Chem. Soc., 1992, 114, 2321.
45 A. Hafner and J. Okuda, Organometallics, 1993, 12, 949.
46 W. C. Finch, E. W. Anslyn and R. H. Grubbs, J. Am. Chem. Soc.,
1988, 110, 2406.
64 M. Bochmann and S. J. Lancaster, J. Organomet. Chem., 1992, 434,
C15.
Received 6th May 1997; Paper 7/03112B
47 J. J. Dechter, Prog. Inorg. Chem., 1985, 33, 393.
3104
J. Chem. Soc., Dalton Trans., 1997, Pages 3097–3104