Trichloro monophenoxide complexes of titanium(iv)

Article abstract of DOI:10.1039/a908435e

Thermalisation of TiCl₄ and phenol (1:1) in toluene gave [TiCl₃(OC₆H₅)] 1. The more soluble complex [TiCl₃(OC₆H₄CMe₃-4)] 2 is monomeric in benzene and reacts with 4, 4′-dimethyl-2, 2′-bipyridyl (dmbipy) to give mer-[TiCl₃(OC₆H₄CMe₃-4)(dmbipy)] 3 and the disproportionation product [TiCl₂(OC₆H₄CMe₃-4) ₂(dmbipy)]. The complex [TiCl₃(OC₆H₂Me₃-2, 4, 6)] 4 is monomeric in benzene whereas [TiCl₃(OC₆H₃Pr₂¹-2, 6)] 5 partially disproportionates in solution into [TiCl₂(OC₆H₃Pr₂¹-2, 6)₂] and reacts with dmbipy to give mer-[TiCl₃(OC₆H₃Pr₂¹-2, 6)(dmbipy)] 6 and [TiCl₂(OC₆H₃Pr₂¹-2, 6)₂(dmbipy)]. Thermalisation of 2, 6-di-/4 in toluene caused debutylation but [TiCl₃{OC₆H₂(CMe₃)₂-2, 6-Me-4}] 7 forms in light petroleum (bp range 40-60 °C). Complex 7 is monomeric in benzene and does not form adducts with dmbipy or other sigma donors. A crystal structure determination of 7 showed a monomer with distorted tetrahedral co-ordination, a Ti-O bond length of 1.750(2) A and Ti-Cl bonds longer than in TiCl₄ but shorter than in [TiCl₃(C₅H₅)] or [TiCl₃{C₅H₃(CMe₃)₂-l, 3}]· 2, 4, 6-Tri-rcrt-butylphenol debutylates when thermalised with TiCl₄ in toluene giving [TiCl₃{OC₆H₄(CMe₃)₂-2, 4}] 8. The complexes [TiCl₃{OC₆H₂(CMe₃)₂-2, 6-OMe-4}] 9, [TiCl₃(OC₆H₃CMe₃-2-Me-4)] 10, [TiCl₃(OC₆H₄Ph-2)] 11 and the 1-naphthoxide complex [TiCl₃(OC₁₀H₇)J 12 were also prepared. Density functional calculations performed on the models 4 and [TiCl₃(OMe)] showed both lone pairs on oxygen donate electron density to titanium but O(2p)-to-C=C (π*) donation weakens the Ti-O interaction in the phenoxide complex; CI(2p)-to-Ti(3d) donation is much reduced in the methoxide complex. The system [TiCl₃(OC₆H₄CMe₃-4)]/AlMe ₃ is 280 times more active than [TiCl₃Cp] (Cp = cyclopentadienyl)/AlMe₃ for low pressure (6 psi) ethylene polymerisation but 1/3 less active than TiCl₄/AlMe₃. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a908435e

View Article Online / Journal Homepage / Table of Contents for this issue
Trichloro monophenoxide complexes of titanium(IV)
Alastair J. Nielson,*a Peter Schwerdtfegerb and Joyce M. Watersa
a Chemistry, Institute of Fundamental Sciences, Massey University at Albany,
Private Bag 102 904, North Shore Mail Centre, Auckland, New Zealand
b Department of Chemistry, The University of Auckland, Private Bag 92 019, Auckland,
New Zealand
Received 22nd October 1999, Accepted 6th January 2000
Thermalisation of TiCl4 and phenol (1:1) in toluene gave [TiCl3(OC6H5)] 1. The more soluble complex
[TiCl3(OC6H4CMe3-4)] 2 is monomeric in benzene and reacts with 4,4Ј-dimethyl-2,2Ј-bipyridyl (dmbipy) to
give mer-[TiCl3(OC6H4CMe3-4)(dmbipy)] 3 and the disproportionation product [TiCl2(OC6H4CMe3-4)2(dmbipy)].
The complex [TiCl3(OC6H2Me3-2,4,6)] 4 is monomeric in benzene whereas [TiCl3(OC6H3Pri2-2,6)] 5 partially
disproportionates in solution into [TiCl2(OC6H3Pri2-2,6)2] and reacts with dmbipy to give mer-[TiCl3(OC6H3Pri2-
2,6)(dmbipy)] 6 and [TiCl2(OC6H3Pri2-2,6)2(dmbipy)]. Thermalisation of 2,6-di-tert-butyl-4-methylphenol and TiCl4
in toluene caused debutylation but [TiCl3{OC6H2(CMe3)2-2,6-Me-4}] 7 forms in light petroleum (bp range 40–60 ЊC).
Complex 7 is monomeric in benzene and does not form adducts with dmbipy or other sigma donors. A crystal
structure determination of 7 showed a monomer with distorted tetrahedral co-ordination, a Ti–O bond length of
1.750(2) Å and Ti–Cl bonds longer than in TiCl4 but shorter than in [TiCl3(C5H5)] or [TiCl3{C5H3(CMe3)2-1,3}].
2,4,6-Tri-tert-butylphenol debutylates when thermalised with TiCl4 in toluene giving [TiCl3{OC6H4(CMe3)2-2,4} ] 8.
The complexes [TiCl3{OC6H2(CMe3)2-2,6-OMe-4}] 9, [TiCl3(OC6H3CMe3-2-Me-4)] 10, [TiCl3(OC6H4Ph-2)] 11 and
the 1-naphthoxide complex [TiCl3(OC10H7)] 12 were also prepared. Density functional calculations performed on the
models 4 and [TiCl (OMe)] showed both lone pairs on oxygen donate electron density to titanium but O(2p)-to-C᎐C
᎐
3
(π*) donation weakens the Ti–O interaction in the phenoxide complex; Cl(2p)-to-Ti(3d) donation is much reduced
in the methoxide complex. The system [TiCl3(OC6H4CMe3-4)]/AlMe3 is 280 times more active than [TiCl3Cp]
1
–
(Cp = cyclopentadienyl)/AlMe3 for low pressure (6 psi) ethylene polymerisation but 3 less active than TiCl4/AlMe3.
Currently there is intense interest in finding replacements for
the cyclopentadienyl ligand in Group 4 transition metal chem-
istry due mainly to the impact metallocenes have had on olefin
polymerisations.1 A major challenge is finding replacements
that allow the metal to remain co-ordinatively and electronic-
ally unsaturated.2
Monocyclopentadienyl complexes [TiCl3Cp] (Cp = unsubsti-
tuted or substituted cyclopentadienide) and their derivatives are
known as active catalysts for polymerisations of ethylene and
propene,3 conjugated dienes,4 and the syndiospecific polymer-
isation of styrene5 but there have been few studies directed
towards replacing the 1σ, 2π donor ligand in this type of com-
plex. Both alkoxide (RO) and phenoxide ligands (ArO)6 are
capable of 1σ, 2π donation7 to transition metals with the latter
ligand being especially attractive since electronic and steric
properties can easily be assessed because of the commercial
availability of a wide range of substituted phenols.
We report here a comprehensive study of the monophen-
oxides [TiCl3(OAr)] (Ar = unsubstituted or substituted phenyl
group) prepared for use in catalytic applications. In particular
we have investigated the influence of steric factors on molecular
structure and on expansion of the co-ordination sphere. We
also present a fully optimised Density Functional Theory
(DFT) study of the bonding in the model complex [TiCl3-
(OC6H2Me3-2,4,6)] and compare it with the alkoxide model
[TiCl3(OMe)]. These studies represent the first detailed theor-
etical description of phenoxide and alkoxide bonding in an
early transition metal complex.
Results and discussion
Synthetic studies
The synthetic strategy employed in preparing the complexes
was dictated by the requirement to produce multigram quan-
tities of a pure product in high yield without recrystallisation,
via a simple and cheap method using commercially available
starting materials as supplied. This was best achieved by
thermalisation,10 eqn. (1). This method gave good yields of
The chemistry of the monophenoxides, [TiCl3(OAr)], is
poorly developed. For example, [TiCl3(OC6H5)] has been
known for many years8 but few of its properties have
been described.9 The complex [TiCl3(OC6H2Me3-2,4,6)] has
been shown to be monomeric in non-co-ordinating solvents and
the bis adducts [TiCl3(OC6H2Me3-2,4,6)(L)2] (L = pyridine,
2-, 3-, 4-methylpyridine, PhNH2, tetrahydrofuran, ¹ Ph2P-
reflux
(1)
TiCl4 ϩ ArOH
[TiCl3(OAr)] ϩ HCl
toluene
¯
²
CH2CH2PPh2 or ¹ 2,2Ј-bipyridyl) have been characterised
¯
²
by analytical data and IR spectroscopy.10 Uncharacterised
[TiCl3(OC6H4CMe3-2)] has been evaluated for regioselective
ortho-acylation11 and the 3,3Ј,5,5Ј tetrasubstituted biphen-
olate (tsb) complexes [{TiCl3(L)}2(tsb)] (L = diethyl ether
or tetrahydrofuran) prepared.12 The complex [TiCl3-
{OC6H3(CMe3)2-2,6}] has been tested in the presence of
methylaluminoxane [(MeAlO)n, MAO] co-catalyst for the
co-polymerisation of styrene and ethylene13 and its crystal
structure determined.14
solid complexes if the reactions were thermalised to completion
whereas incomplete thermalisation gave gums which proved
difficult to purify. Generally the complexes were obtained as
non-crystalline solids. Attempted recrystallisations mostly gave
gums of various composition and in some cases partial dis-
proportionation to the bis-phenoxide, [TiCl2(OAr)2], occurred
especially after extended periods of solvent contact. In general
the complexes were best isolated by filtering the reflux solution
and pumping off the solvent. In some instances this procedure
DOI: 10.1039/a908435e
J. Chem. Soc., Dalton Trans., 2000, 529–537
This journal is © The Royal Society of Chemistry 2000
529

View Article Online
Table 1 Yields and analytical data
Analysisa (%)
Crude
Complex
yield (%)
C
H
Cl
1
2
[TiCl3(OC6H5)]b,c
90
87
37.6
(37.7)
45.6
(45.5)
38.1
(38.1)
43.9
2.9
(3.0)
5.2
(4.9)
4.2
(3.9)
5.5
37.5
(37.1)
31.3
(31.0)
36.2
(36.3)
32.4
[TiCl3(OC6H4CMe3-4)]b–d
[TiCl3(OC6H2Me3-2,4,6)]c,e,f
[TiCl3(OC6H3Pri2-2,6)]
4
97
5
94
(43.5)
48.2
(48.5)
45.6
(46.7)
47.7
(47.7)
42.2
(41.6)
44.5
(44.6)
48.7
(5.2)
6.2
(5.7)
6.2
(5.9)
6.1
(6.0)
5.4
(4.8)
3.3
(2.8)
3.3
(32.1)
27.8
(28.6)
29.5
(29.6)
26.9
(26.4)
33.5
(33.5)
32.2
(32.9)
30.3
7
[TiCl3{OC6H2(CMe3)2-2,6-Me-4}]g
[TiCl3{OC6H3(CMe3)2-2,4}]
[TiCl3{OC6H2(CMe3)2-2,6-OMe-4}]
[TiCl3(OC6H3CMe3-2-Me-4)]
[TiCl3(OC6H4Ph-2)]
95
8
96
9
99
10
11
12
100
100
96
[TiCl3(OC10H7)]c,h
(48.0)
(3.5)
(30.4)
a Calculated values given in parentheses. b Calculated analytical data include 0.43 C7H8. c Solvent supported by NMR spectra. d M 317, calc. 303.5.
e Calculated analytical data include 0.041 C7H8. f M 284, calc. 289.3. g M 375, calc. 371.0. h Calculated data include 0.57 C7H8.
left toluene in the sample which was difficult to remove by
pumping. Washing introduced further solvent. Analytical data
are contained in Table 1 which indicates the solvent content.
The presence of solvent in the complexes was confirmed by
NMR spectroscopy.
Owing to the relevance of co-ordination expansion in
titanium chemistry and, in particular, the importance of
bidentate σ-donor ligands in stereospecific Ziegler–Natta cat-
alysis,16 a variety of such ligands was treated with complex 2
but were found to produce mixtures which were difficult to
distinguish by NMR spectroscopy. However, complex 2 reacts
with 4,4Ј-dimethyl-2,2Ј-bipyridyl (dmbipy) in CH2Cl2 to give
an orange solid which analysed as [TiCl3(OC6H4CMe3-4)-
(dmbipy)] when the solvent was removed from the reaction
mixture. The NMR spectra showed, however, that the CDCl3
solubles consisted of a mixture of mer-[TiCl3(OC6H4CMe3-
4)(dmbipy)] 3 (63%) and the bis-phenoxide [TiCl2(OC6H4-
CMe3-4)2(dmbipy)] (34%).† The 1H NMR spectrum shows
inequivalent dmbipy rings for 3 (see structure I) with widely
The reaction of TiCl4 and phenol in a 1:1 ratio produced
HCl gas for 12 h giving rise to [TiCl3(OC6H5)]8 1 for which
1
the H NMR spectrum (Table 2) showed the absence of the
phenolic proton and the 13C-{1H} NMR spectrum showed a
phenoxide ligand ipso-carbon at δ 169.65. The more soluble
complex [TiCl3(OC6H4CMe3-4)] 2 is monomeric in benzene
solution and also shows an ipso-carbon resonance in this region
(δ 169.6, CDCl3; 170.53, C6D6). Recrystallisations did not give
crystals suitable for X-ray analysis and contact with solvent
over time led to disproportionation, eqn. (2). Thus, if a solution
2[TiCl3(OC6H4CMe3-4)] →
[TiCl2(OC6H4CMe3-4)2] ϩ TiCl4 (2)
of 2 in light petroleum is allowed to stand to give slow
crystallisation over several days, crystals of the dichloro-
bis(phenoxide) are formed and the solution fumes
characteristically for TiCl4. However, after a series of such
crystallisations, if the solution is concentrated sufficiently then
complex 2 is deposited and fuming does not occur. We have
recently reported a much more rapid disproportionation for a
series of tungsten monophenoxide complexes.15 Based on the
ipso-carbon resonance position in the 13C-{1H} NMR spectra
of 1 and 2 and the monomeric nature of 2, tetrahedral co-
ordination geometry is predicted for these two complexes. This
allows the electron count to be maximised by π donation from
the chloro and phenoxide ligands. A theoretical model for this
type of bonding is discussed later.
separated doublets for the C1 and C10 hydrogens (δ 9.69 and
8.70; see dmbipy numbering scheme, Table 2), two singlets for
the C4 and C7 hydrogens (δ 8.02 and 7.96; meta coupling not
resolved), doublets for the C2 and C9 hydrogens (δ 7.47 and
7.35) and a broadened resonance containing the C3 and C8
methyl groups. Similarly, in the 13C-{1H} NMR spectrum there
are individual resonances for each of the carbons in the two
dmbipy rings and a broadened resonance for the C3 and C8
methyl groups. The phenoxide ipso-carbon resonance (δ 166.40)
lies to slightly higher field than for the four-co-ordinate parent
† The bis-phenoxide complexes mentioned in this work were independ-
ently prepared and will be reported elsewhere.
530
J. Chem. Soc., Dalton Trans., 2000, 529–537

View Article Online
Table 2 Selected NMR spectral data (J/Hz)a
Complex
1Hb
13C-{1H}
1
2
7.16 (m, 3 H, m, p-H); 7.29 [d, 3J(HH) 7.1, 2 H, o-H]
118.77 (o-C); 126.24 ( p-C); 129.38 (m-C); 169.65 (ipso-C)
31.31 (Me); 34.80 (C); 118.14 (o-C); 126.15 (m-C); 150.07
( p-C); 169.60 (ipso-C)
21.77 (Me); 31.52 (CMe3); 34.67 (C); 119.53 [o-C (phenoxide)];
122.73 and 122.88 [C2 and C9 (dmbipy)]; 126.16 [m-C (phenox-
ide)]; 127.18 and 127.47 [C4 and C7 (dmbipy)]; 148.22 (C1 or
C10 (dmbipy)]; 148.36 [ p-C (phenoxide)]; 150.64 and 151.12 [C3
and C8 or C5 and C6 (dmbipy)]; 151.27 [C10 or C1 (dmbipy)];
152.52 and 153.28 [C6 and C5 or C8 and C3 (dmbipy)]; 166.40
[ipso-C (phenoxide)]
3
1.28 (s, 9 H, CMe3); 7.12 [d, J(HH) 8.4, 2 H, o-H]; 7.27 [d,
3J(HH) 8.4, 2 H, m-H]
3c
1.31 (s, 9 H, CMe3); 2.52 [s, 3 H, Me (dmbipy)]; 7.35 [d, 3J(HH)
3
5.6, 1 H, H2 or H9 (dmbipy)]; 7.38 [d, J(HH) 8.7, 2 H, o-H];
7.47 [d, 3J(HH) 5.6, 1 H, H9 or H2 (dmbipy)]; 7.54 [d, 3J(HH)
8.7, 2 H, m-H]; 7.96 [s, 1 H, H4 or H7 (dmbipy)]; 8.02 [s, 1 H,
H7 or H4 (dmbipy)]; 8.70 [d, 3J(HH) 5.6, H1 or H10 (dmbipy)];
9.69 [d, 3J(HH) 5.6, H10 or H1 (dmbipy)]
4
2.30 (s, 3 H, p-Me); 2.47 (bs, 6 H, o-Me); 6.83 (s, 2 H, m-H)
16.96 (o-Me); 20.95 ( p-Me); 128.70 (m-C); o-C, p-C and
ipso-C not observed
23.31 [CMe2]; 27.59 [CH]; 123.41 [m-C]; 126.93 [ p-C]; 139.32
[o-C]; 170.50 [ipso-C]
21.78 [Me(dmbipy)]; 24.54 (CMe2); 26.51 (CH); 122.81 [p-C
(phenoxide)]; 124.08 [m-C (phenoxide)]; 125.31 and 125.55 [C2
and C9 (dmbipy)]; 127.02 and 125.55 [C4 and C7 (dmbipy)];
142.91 [o-C (phenoxide)]; 148.22 [C1 or C10 (dmbipy)]; 150.80
and 150.94 [C3 and C8 or C5 and C6 (dmbipy)]; 151.01 [C10 or
C1 (dmbipy)]; 152.04 and 153.30 [C6 and C5 or C8 and C3
(dmbipy)]; 164.60 [ipso-C (phenoxide)]
21.69 (Me); 31.69 (CMe3); 45.06 (C); 125.51 (m-C); 135.67
(p-C); 140.40 (o-C); 174.87 (ipso-C)
30.47 (p-CMe3); 31.33 (o-CMe3); 35.00 (C); 35.26 (C); 123.26,
123.35 and 124.26 (o, m-CH); 136.05 (p-C); 150.02 (o-C);
169.90 (ipso-C)
5d
6c
1.38 [d, J(HH) 6.8, 6 H, CMe2]; 3.70 [sept, J(HH) 6.8, 1 H,
CH]; 7.18 [m, 3 H, m, p-H]
3
3
3
1.22 [d, J(HH) 6.6, 12 H, CMe2]; 2.56 and 2.59 [2s, 6 H, Me
(dmbipy)]; 4.50 [sept, 3J(HH) 6.6, 2 H, CH]; 7.13 [m, 1 H,
p-H]; 7.20 [m, 2 H, m-H]; 7.27 [d, J(HH) 5.2, 1 H, H2 or H9
3
(dmbipy)]; 7.47 [d, 3J(HH) 5.2, 1 H, H9 or H2 (dmbipy)]; 8.00
[bs, 2 H, H4 or H7 (dmbipy)]; 8.53 [d, J(HH) 5.2, 1 H, C1 or
3
C10 (dmbipy)]; 9.72 [d, 3J(HH) 5.2, 1 H, C10 or C1 (dmbipy)]
1.56 (s, 18 H, CMe3); 2.34 (s, 3 H, Me); 7.06 (s, 2 H, m-H)
1.35 (s, 9 H, p-CMe3); 1.55 (s, 9 H, o-CMe); 7.27 and 7.29 [dd,
7
8d
3J(HH) 8.5, J(HH) 2.3, 1 H, m-H]; 7.37 [d, J(HH) 2.3, 1H,
4
4
m-H]; 7.50 [d, 3J(HH) 8.5, 1 H, o-H]
9
1.57 (s, 18 H, CMe3); 3.82 (s, 3 H, OMe); 6.76 (s, 2 H, m-H)
31.51 (CMe3); 35.42 (C); 55.36 (OMe); 109.67 (m-C); 141.57
(o-C); 155.95 (p-C); 173.17 (ipso-C)
21.42 (Me); 30.45 (CMe3); 34.96 (C); 124.23 (o-CH); 127.01
(m-CH); 127.99 (m-CH); 136.53 (p-C); 136.92 (o-C); 170.21
(ipso-C)
10d
1.49 (s, 9 H, CMe3); 2.36 (s, 3 H, Me); 7.02 [d, 3J(HH) 8.2, 1 H,
m-H]; 7.10 (bs, 1 H, m-H); 7.41 [d, 3J(HH) 8.2, 1 H, o-H]
11
6.90–7.18 [m, 2 H, o-H(phenoxide), p-H(phenyl)]; 7.22 [td,
3J(HH) 7.6, 4J(HH) 1.6, 1 H, p-H (phenoxide)]; 7.30 [td,
3J(HH) 7.6, 4J(HH) 1.6, 2 H, m-H (phenoxide)]; 7.36 [bt,
120.36, 125.80, 127.90, 128.34, 128.62, 129.38, 130.25 (CH);
131.49 [C(phenyl)]; 136.49 [C(phenoxide)]; 166.63 (ipso-C)
3
4
3J(HH) 7.1, 2 H, m-H (phenyl)]; 7.44 [d, J(HH) 7.1, J(HH)
1.3, 2 H, o-H (phenyl)]
a Spectra obtained in dry CDCl3 solution. b bs = Broad singlet, bt = broad triplet, d = doublet, dd = doublet of doublets, m = multiplet, s = singlet,
sept = septet, td = triplet of doublets. c Spectra also contain resonances characteristic of [TiCl2(OAr)2(dmbipy)]. d Spectra also contain resonances
characteristic of [TiCl2(OAr)2].
complex 2 (δ 169.6), most likely as a result of decreased
π-electron donation from the phenoxide ligand in the six-
co-ordinate dmbipy adduct.
The appearance in the NMR spectra of the bis-phenoxide
complex (identified by comparison with an authentic
sample) suggests that addition of dmbipy to the monophen-
oxide complex 2 in CH2Cl2 solution involves a disproportion-
ation, eqn. (3). The presence of [TiCl4(dmbipy)] has not been
the solvent was driven off. The integrity of this material is
unknown but the process of dissolving in CDCl3 apparently
sets off solution dynamics. The material initially dissolves
but a precipitate rapidly forms and the 1H NMR spectrum
shows more of a variety of products than was observed for the
reaction in CH2Cl2 although the mono- and bis-phenoxide
products still dominate.
Reactions of TiCl4 were carried out with other phenols which
contain substituents in the 2,6 positions of the phenyl ring, in
order to assess the effect of increasing steric size. The complex
3[TiCl3(OC6H4CMe3-4)] ϩ 3dmbipy →
[TiCl3(OC6H4CMe3-4)(dmbipy)] ϩ
1
[TiCl3(OC6H2Me3-2,4,6)] 4 is monomeric in benzene; the H
NMR spectrum shows a broadened resonance for the 2,6-
dimethyl groups and in the 13C-{1H} NMR spectrum the
quaternary carbons of the aromatic ring were not observed due
to slow relaxation times. Octahedral adducts of this complex
have been described10 but NMR spectral characteristics have
not been reported. The complex [TiCl3(OC6H3Pri2-2,6)] 5 can
be obtained as a solid product from the thermalisation reaction
but the composition indicated by the NMR spectra differs from
that indicated by the analytical figures, and this suggests that
dynamic processes occur in solution. In CDCl3 disproportion-
ation apparently occurs giving a mixture of complex 5 (74%)
(ipso-carbon resonance δ 170.50) and [TiCl2(OC6H3Pri2-2,6)2]
(26%) whereas in C6D6–CDCl3 (1:1) the proportions are 60
and 40% respectively. The solution dynamics apparently occur
rapidly on dissolving the solid since the mixture proportions do
[TiCl2(OC6H4CMe3-4)2(dmbipy)] ϩ [TiCl4(dmbipy)] (3)
confirmed since it is extremely insoluble and remains con-
taminated when the mono- and bis-phenoxide complexes are
extracted. Whether the disproportionation involves an acceler-
ation of the solution dynamics observed for 2 is unclear but the
reaction does explain the overall analysis as [TiCl3(OC6H4-
CMe3-4)(dmbipy)] and the observed NMR spectra.
Further to understand this process and also to determine the
capacity of the analysed material to carry aromatic solvent,
dmbipy in light petroleum–benzene (80:20) was added to
complex 2 in light petroleum and the orange precipitate
collected. After repeated washings with light petroleum and
drying in vacuo, the complex analysed as [TiCl3(OC6H4CMe3-
4)(dmbipy)]ؒ0.33C6H6. Only with strong heating under vacuum
J. Chem. Soc., Dalton Trans., 2000, 529–537
531

View Article Online
Table 3 Selected bond lengths [Å] and angles [Њ] for [TiCl3{OC6H2-
(CMe3)2-2,6-Me-4}]
not change with time after the initial 1H NMR spectrum is run.
When 5 is dissolved on a larger scale in CHCl3 the solution
fumes characteristically for TiCl4. Complex 5 reacts with
dmbipy in CH2Cl2 solution to give an orange solid which anal-
yses as [TiCl3(OC6H3Pri2-2,6)(dmbipy)] but which NMR spectra
show to be a mixture of mer-trichloro [TiCl3(OC6H3Pri2-
2,6)(dmbipy)] 6 (55%) (ipso-carbon resonance δ 164.6) and the
bis-phenoxide [TiCl2(OC6H3Pri2-2,6)2(dmbipy)] (45%). When
the reaction was carried out in light petroleum–benzene (as
for complex 2) an orange solid was obtained analysing as
[TiCl3(OC6H3Pri2-2,6)(dmbipy)]ؒ0.83C6H6 and which gave up
benzene only on heating under vacuum. This material, which
initially dissolves in CDCl3 and then produces a precipitate,
shows more products in the NMR spectra than does the ori-
ginal CH2Cl2 reaction but the mono- and bis-phenoxide
dmbipy products still dominate.
Ti–O(1)
Ti–Cl(3)
Ti–Cl(2)
1.750(2)
2.1822(8)
2.1913(8)
Ti–Cl(1)
O(1)–C(1)
2.1945(9)
1.390(2)
O(1)–Ti–Cl(3)
O(1)–Ti–Cl(2)
Cl(3)–Ti–Cl(2)
O(1)–Ti–Cl(1)
112.66(6)
109.18(6)
108.57(3)
112.84(5)
Cl(3)–Ti–Cl(1)
Cl(2)–Ti–Cl(1)
C(1)–O(1)–Ti
105.78(4)
107.58(3)
163.08(14)
When 2,6-di-tert-butyl-4-methylphenol was refluxed with
TiCl4 in a 1:1 ratio in toluene NMR spectroscopy showed that
a significant amount of debutylation occurred. However
[TiCl3{OC6H2(CMe3)2-2,6-Me-4}] 7 was prepared in almost
quantitative yield by refluxing TiCl4 and the phenol in light
petroleum (bp 40–60 ЊC) until the production of HCl
ceased. The complex [TiCl3{OC6H3(CMe3)2-2,6}] has been
synthesized13,14 by treating LiOC6H3(CMe3)2-2,6 with TiCl4 in
benzene but the reflux method represents a cheaper and more
convenient synthesis for this type of complex. A 50 g prepar-
ation of complex 7 is easily obtainable using this method.
A molecular weight determination showed that complex 7
is monomeric in benzene. The NMR spectrum shows a single
resonance for each of the appropriate protons and carbons with
the phenoxide ligand ipso carbon positioned in the 13C-{1H}
NMR spectrum at δ 174.87. This is further downfield than that
observed for all the other complexes and may represent an
increase in π donation from the phenoxide ligand with the elec-
tron donating tert-butyl groups present. The NMR spectra
do not show any evidence for disproportionation to the bis-
phenoxide suggesting that, although the tert-butyl groups are
more electron donating than the isopropyl groups in complex 5,
steric influences may now be important. Preliminary studies
show that even after extended refluxing of TiCl4 and 2 equiv-
alents of 2,6-di-tert-butyl-4-methylphenol in toluene there is
more of the monophenoxide present than the bis-phenoxide. In
comparison with complexes 2 and 5 the di-tert-butyl complex 7
does not react with dmbipy to expand its co-ordination sphere
nor does it react with a variety of other σ-donor ligands such as
tetrahydrofuran, pyridine or PMe3.
Reaction of 2,6-diphenylphenol with TiCl4 gave a mixture
of two complexes which the NMR data suggest are [TiCl3-
(OC6H3Ph2-2,6)] and the bis-phenoxide [TiCl2(OC6H3Ph2-
2,6)2].17 Owing to the expense of this phenol, in comparison
with the others used here, we have not yet persevered with the
monophenoxide preparation although a comparison of its
steric properties with those of the 2,6-di-tert-butyl substituted
phenoxide in complex 7 are of interest.
When 2,4,6-tri-tert-butylphenol was refluxed in toluene with
TiCl4 debutylation of one of the tert-butyl groups occurred
giving [TiCl3{OC6H3(OCMe3)2-2,4}] 8. In a preliminary study
on the electronic effect of substituents on the 2,6-di-tert-
butylphenoxide system 2,6-di-tert-butyl-4-methoxyphenol gave
[TiCl3{OC6H2(CMe3)2-2,6-OMe-4}] 9 (ipso-carbon resonance,
δ 173.17), whereas the reaction with 2,6-di-tert-butyl-4-nitro-
phenol generated HCl only very slowly and did not produce a
characterisable product.
Fig. 1 Molecular structure of complex 7; atoms are depicted as 50%
probability surfaces. Hydrogen atoms have been omitted for clarity.
structural features of 7 with other related complexes which
carry out this function. Selected bond lengths and angles for 7
18
are given in Table 3; comparable data for TiCl4 and the
monocyclopentadienyl complexes [TiCl3Cp] [Cp = C5H5,19
C5H3(SiMe3)2 20 or C5H3(CMe3)2 21] are given in Table 4.
The Ti–O(1) bond length in complex 7 [1.750(2) Å] is longer
than that found for the tetrahedral bis-phenoxide complex
[TiCl2(OC6H3Ph2-2,6)2] [1.726(2) Å]17 and a range of neutral
octahedral monoisopropoxide complexes of titanium [range
1.702(4)–1.726(4) Å]22 all of which require strong π donation
from oxygen to maximise the electron count on Ti. However, it
is shorter than the Ti–O distances found in the tris-phenoxide
complex [TiCl{OC6H3(CMe3)2-2,6}3] [1.810(9), 1.802(7) and
1.782(8) Å]23 where three phenoxide oxygens are able to
π-donate to the metal. The Ti–O–C bond angle in 7 [163.1(1)Њ]
is smaller than that in [TiCl2(OC6H3Ph2-2,6)2] (168.5(2)Њ)17 but
larger than those found for the octahedral isopropoxide com-
plexes [range 153.7(4)–157.5(6)Њ]22 suggesting that the M–O–C
bond angle is not greatly affected by the π nature of the ligand
to metal bonding.
In complex 7 the phenyl ring is oriented so that the tert-butyl
groups are positioned over Cl(1) and Cl(3) leaving Cl(2)
exposed. The 3 Ti–Cl bond lengths are all similar (range
2.1822(8)–2.1945(9) Å but are longer than those found in TiCl4
[2.170(2) Å]18 where each chloride must π-donate 2 electrons to
attain an electron count of 16 for the metal. However, on aver-
age they are shorter than in the [TiCl3Cp] complexes (Table
4) where the π-donor ability of the Cp ligands appears to
reduce the amount of π donation needed from chlorine. This is
especially so with the 1,3-di-tert-butylcyclopentadienyl ligand
where the Ti–Cl bond lengths increase to ca. 2.240(1) Å when
the electron donating tert-butyl substituents are present. In 7
the two chloro ligands lying closer to the tert-butyl groups
[Cl(1) and Cl(3)] are directed away from the phenoxide
ligand more than is the more isolated chloro ligand [O(1)–
Ti–Cl(1), 112.84(5); O(1)–Ti–Cl(2), 109.18(6); O(1)–Ti–Cl(3),
112.66(6)Њ]. Overall, however, the ligand is less sterically
demanding than its Cp counterparts of Table 4 where the
Cp(centroid)–Ti–Cl angles widen to 117Њ. As a result the Cl–Ti–
Cl angles of 7 are all larger than those for the Cp complexes of
Table 4.
A single-crystal structure determination confirmed the solid
state monomeric nature of [TiCl3{OC6H2(CMe3)2-2,6-Me-4}] 7.
The co-ordination geometry of the Ti is that of a distorted
tetrahedron with three chloro ligands and a phenoxide oxygen
as donor atoms (Fig. 1) and is similar to that reported recently
for [TiCl3{OC6H3(CMe3)2-2,6}].14 Owing to the importance of
this type of complex in olefin polymerisations13 we compare the
532
J. Chem. Soc., Dalton Trans., 2000, 529–537

View Article Online
Table 4 Comparative bond length (Å) and bond angle (Њ) data
[TiCl3{OC6H2(CMe3)2-
2,6-Me-4}]
[TiCl3{C5H3(SiMe3)2-
1,3}]c
[TiCl3{C5H3(CMe3)2-
[TiCl3(C5H5)]b
1,3}]d
a
TiCl4
Ti–Cl(1)
Ti–Cl(2)
Ti–Cl(3)
Ti–O
2.1945(9)
2.1913(8)
2.1822(8)
1.750(2)
2.170(2)
2.170(2)
2.170(2)
2.170(2)e
2.201(3)
2.248(5)
2.221(2)
[2.01] f
2.232(3)
2.229(3)
2.229(3)
[2.007(8)] f
2.240(1)
2.243(1)
2.243(1)
[2.022] f
Cl(1)–Ti–Cl(2)
Cl(2)–Ti–Cl(3)
Cl(1)–Ti–Cl(3)
105.78(4)
108.57(3)
107.58(3)
109
109
109
102.0
104.1(2)
102.3(3)
102.0(1)
102.6(1)
102.0(1)
100.4(1)
101.9(1)
100.4(1)
O–Ti–Cl(1)
O–Ti–Cl(3)
O–Ti–Cl(2)
112.84(5)
112.66(6)
108.57(3)
109g
109g
109g
117.2h
114.3h
115.0h
116.4(2)h
116.4(2)h
115.3(3)h
—
—
—
a Gas diffraction data taken from ref. 18. b X-Ray data taken from ref. 19. c X-Ray data taken from ref. 20. d X-Ray data taken from ref. 21. e Ti–Cl
bond length. f Ti–Cp centroid distance. g Cl–Ti–Cl bond angle. h Cp centroid–Ti–Cl bond angle.
The phenoxide ligand in complex 7 bends towards Cl(2) [Ti–
O(1)–C(1) is 163.1(1)Њ] which allows the tert-butyl group
methyls to maximise their distances from Cl(1) and Cl(3). How-
ever, it is unlikely that the Ti–O(1)–C(1) bond angle is substan-
tially influenced by such steric contacts since small rotations of
the bulky groups can minimise the interactions. The angles sub-
tended at C(8) and C(12) by the tert-butyl carbon atoms facing
Cl(1) and Cl(3) widen [C(9)–C(8)–C(11), 111.9(2); C(14)–
C(12)–C(15), 112.1(2)Њ] compressing the other C–C–C bond
angles of the tert-butyl group [range 105.9(2)–106.4(2)Њ] and, as
well, the angles C(1)–C(2)–C(8) and C(1)–C(6)–C(12) [123.7(2)
and 123.4(2)Њ respectively] are increased slightly above the ideal
angle of 120Њ to relieve the steric pressures.
The unsymmetrically substituted phenols 2-tert-butyl-4-
methylphenol and 2,4-di-tert-butylphenol on refluxing with
TiCl4 gave [TiCl3(OC6H3CMe3-2-Me-4)] 10 and [TiCl3{OC6H3-
(CMe3)2-2,4}] 8 the latter complex having previously been
prepared by debutylation of 2,4,6-tri-tert-butylphenol. NMR
natural charges (right) for [TiCl3(OC6H2Me3-2,4,6)] 4. All bond
Fig. 2 DFT (B3LYP) optimised structure (left) and DFT (B3LYP)
spectra show disproportionation occurs in CDCl3 solution for
these complexes. The monophenoxides (composition 70% for
10, 61% for 8) show ipso-carbon resonances at δ 170.21 for 10
and 169.90 for 8 and there are NMR resonances characteristic
of the bis-phenoxides (composition 30 and 39% respectively).
Refluxing 2-tert-butyl-6-methylphenol with TiCl4 in toluene
distances (in Å) and angles (in Њ) are indicated in the drawing. The
important optimised torsion angles τ are as follows: τ (CCOTi) = 90.0;
τ (COTiCl1) = 119.3; τ (COTiCl2) = 2.6Њ.
the B3LYP level), which in the crystal structure is 163.1Њ. How-
ever, the calculation does not take into account the effect of the
bulky CMe3 substituents in the 2,6 positions of the phenyl ring
or any crystal packing forces. In addition, approximations in
the basis sets, pseudopotentials, etc. may overestimate the oxy-
gen lone-pair back bonding into the unoccupied Ti(3d) or
gave
a mixture of monophenoxide [TiCl3(OC6H3CMe3-
2-Me-6)] [tentatively identified in the 13C-{1H} NMR spectrum
from its ipso-carbon position (δ 172.61)], and the bis-phenoxide
[TiCl2(OC6H3CMe3-2-Me-6)2]. The toluene reflux reaction
using 2-phenylphenol gave [TiCl3(OC6H4Ph-2)] 11 (ipso-carbon
resonance δ 166.63) and with 1-naphthol [TiCl3(OC10H7)] 12
(ipso-carbon resonance δ 166.62). Complexes 11 and 12 are
relatively insoluble and have not been characterised further.
phenyl C᎐C(π*) orbitals (see discussion below). Nevertheless,
᎐
we confirm an approximately linear arrangement of the Ti–O–
C unit as an angle scan demonstrates (Fig. 3). This figure shows
that the minimum at α = 177.7Њ is shallow, and the difference
between both geometries (the B3LYP minimum at α = 177.7Њ
Theoretical studies
and the crystal structure at α = 177.7Њ) is only 3 kJ molϪ1
.
Density-functional calculations (DFT) were carried out on the
model complex [TiCl3(OC6H2Me3-2,4,6)] 4 to obtain an under-
standing of the chloro and phenoxide ligand bonding to the
titanium centre. In the absence of π donation from any of the
ligands the complex has an overall electron count of 8. This
increases to 12 if the phenoxide oxygen makes 2 π-donor inter-
actions and to 16 if chloro ligands add 2 extra π-donor
interactions.
Fig. 2 shows the fully optimised B3LYP structure of complex
4. From a comparison with the crystal structure of 7 we see that
the geometry of the model compound is generally in good
agreement with the crystallographic interatom distances and
angles. For example, the calculated Ti–O and the Ti–Cl bond
lengths of 1.73 and 2.23 Å, respectively, are close to those
obtained from the crystal structure (1.75 and 2.19 Å). The only
large deviation observed is the Ti–O–C bond angle (177.7Њ at
There appear to be no crystal structure determinations of the
alkoxide complexes [TiCl3(OR)] (R = alkyl group) but molecu-
lar weight determinations indicate monomeric structures in
solution.24 Therefore a DFT calculation was carried out on the
model [TiCl3(OMe)] (Fig. 4) for comparison with the model of
4. Whereas the Ti–Cl bond distances are almost identical with
those of compound 4, the Ti–O bond distance is shortened
significantly and the C–O bond distance increases. The
latter effect is understandable since bonding to an sp2 hybrid-
ised carbon should lead to a smaller bond distance compared
with an sp3 hybridised carbon. In addition, the π bonding
between the oxygen and the phenyl π system is eliminated in
[TiCl3(OMe)]. This should also lead to an increase in the
oxygen 2p lone-pair back donation into the empty titanium
d orbitals. Indeed, the Ti–O bond distance in [TiCl3(OMe)]
(l.718 Å) is considerably shorter than that in the model of 4
J. Chem. Soc., Dalton Trans., 2000, 529–537
533

View Article Online
Fig. 5 Occupied MOs for [TiCl3(OC6H2Me3-2,4,6)] 4 showing (A) the
overlap between the O(pπ) and CPh(pπ), (B) the overlap between the
O(pπ) and Ti(dπ), (C) the interaction between the Cl(p) lone pair with
O(p) and Ti(d) orbitals.
Fig. 3 Ti–O–C angle scan from the linear arrangement (α = 180Њ) to a
Ti–O–C angle of 100Њ. The calculations were carried out at the B3LYP
level of theory.
Fig. 6 Occupied MOs for [TiCl3(OMe)] showing (A) one of the Cl(p)
lone pairs, (B) the overlap between the O(pπ) and Ti(dπ).
2p oxygen lone-pair back donation to titanium for this com-
pound. Indeed, the NBO analysis for [TiCl3(OMe)] does not
show oxygen lone pairs but rather identifies two oxygen–
titanium π bonds with 88% oxygen and 12% titanium character
as shown in Fig. 6B. This explains the linear Ti–O–C
arrangement.
Fig. 4 DFT (B3LYP) optimised structure (left) and DFT (B3LYP)
natural charges (right) for [TiCl3(OMe)]. All bond distances (in Å) and
angles (in Њ) are indicated in the drawing.
The NBO atomic charges for both compounds are shown in
Figs. 2 and 4. Both the Mulliken (qTi = ϩ0.77) and the NBO
analysis (qTi = ϩ0.87) for compound 4 assign a much smaller
atomic charge compared with the titanium in TiCl3Me
(qTi = ϩ1.27).25 The difference is large between the two com-
pounds and can mostly be attributed to differences in the basis
sets used. Indeed, the structure of TiCl3Me at the B3LYP level
was optimised using the same basis sets and pseudopotentials
as in complex 4. For the optimised C3v structure (Ti–Cl, 2.210;
Ti–C, 2.08; C–H, 1.098 Å; C–Ti–Cl, 106.0; C–Ti–Cl, 108.7Њ),
the NBO charges are similar, i.e. qTi = ϩ0.73, qCl = Ϫ0.19. It is
expected that the more electronegative oxygen atom would
result in a much higher atomic charge for titanium but this is
not the case and again indicates strong back donation of the
oxygen lone-pair 2p orbitals into the unoccupied titanium 3d
orbitals. Interestingly, the NBO analysis for TiCl3Me shows
only a small lone-pair Cl(3p) back donation into the un-
occupied titanium 3d orbitals in accordance with that found for
[TiCl3(OMe)].
(1.732 Å). Interestingly, the Ti–O–C bond is now linear which
also indicates a substantially increased oxygen 2p lone-pair
donation to titanium.
Second order perturbation theory analysis of the Fock
matrix within the natural bond order (NBO) basis assigns
energetic contributions E2 to individual donor–acceptor
bonding pairs. The analysis for complex 4 reveals substantial
donation from the oxygen 2p lone pairs into both the phenyl
C᎐C (π*) orbital (E = 13 kcal molϪ1 per lone pair) as shown in
᎐
2
Fig. 5A, and the unoccupied titanium 3d orbitals (E2 = 55 kcal
molϪ1 per lone pair) as shown in Fig. 5B. This explains the
slightly increased bond length of 1.414 Å of the first C᎐C unit
᎐
in the phenyl ring system compared with the other C᎐C bonds.
᎐
The two oxygen lone-pair donations to titanium are approxi-
mately equal in size. The perturbation analysis further reveals
very strong lone-pair Cl(3p) back donation into unoccupied
titanium 3d orbitals which vary between the different lone pairs
(one of the three lone pairs shows only weak donation, the
other two donate more strongly with a maximum E2 = 26 kcal
molϪ1). One of these lone pairs is shown in Fig. 5C. In contrast,
[TiCl3(OMe)] shows only little Cl(3p) back donation to
titanium (maximum E2 = 7 kcal molϪ1).
Catalytic activity studies
Preliminary results are reported for catalytic activity. The com-
plex [TiCl3(OC6H4CMe3-4)] 2 was tested as a catalyst in a low-
pressure (6 psi) polymerisation of ethylene. The various runs
(Table 5) were conducted in toluene solution, each with similar
mole quantities of catalyst, similar pressure of ethylene, a
Fig. 6A clearly demonstrates this showing almost pure
Cl(p) lone-pair character. This is probably due to the increased
534
J. Chem. Soc., Dalton Trans., 2000, 529–537

View Article Online
Table 5 Ethene polymerisationa
106 concentration/
Run
Complex
mol dmϪ3 × 10Ϫ6
Solvent
Pressureb
Yield/g
Activityc
1
2
3
4
5
6
7
2
7.4
6.3
7.3
6.3
7.7
5.9
6.5
Toluene
Toluene
Toluene
Light petroleum
Light petroleum
Toluene
6
6
6
0.768
1.900
0.0027
0.227
1.128
0.550
0.353
5.1797
15.0
0.0185
1.8130
7.2840
4.6368
2.7328
TiCl4
[TiCl3Cp]
2
2
2
9
6
20
20
6
Toluene
a 21 ЊC; 0.12 to 0.16 mmol of catalyst; 50 cm3 total solvent volume; approximately 7 mol equivalents of AlMe3; 1 h reaction period. b psi (lbf
inϪ2 ≈ 6895 Pa). c kg polyethylene (mol cat)Ϫ1 hϪ1
.
7-fold excess of AlMe3, and reaction times of 1 h so that
catalyst activity could be directly compared by the production
of polyethylene.
were distilled from sodium wire and dichloromethane from
freshly ground CaH2. Proton and 13C-{1H} NMR spectra were
recorded at 400 and 100 MHz respectively in CDCl3 solution on
a Bruker AM400 spectrometer; CDCl3 was dried over, and dis-
tilled from, freshly ground CaH2. Molecular weights were
determined cryoscopically in benzene with a Knauer molecular
weight determination apparatus under N2 gas conditions using
concentrations in the vicinity of 0.065 mol dmϪ3. The C, H
and N analyses were determined by Dr A Cunninghame and
associates, University of Otago, New Zealand. Chlorine was
gravimetrically determined.
Based on the yield of polyethylene, complex 2 is approxi-
mately 3 times less active than TiCl4/AlMe3 (Ziegler–Natta
catalysis) but 280 times more active than [TiCl3Cp]/AlMe3
where minimal polyethylene was formed. When the polymeris-
ation was carried out in light petroleum (bp 40–60 ЊC) (run 4)
polyethylene production dropped to 30% of the toluene reac-
tion (run 1). At 20 psi the production in light petroleum (run 5)
was 1.4 times that of run 1 whereas in toluene (run 6) produc-
tion dropped to about 90% of run 1. The 2,6-di-tert-butyl-
phenoxide complex 9 had approximately only ¹ the activity at
Syntheses
¯
²
6 psi (run 7) found for complex 2 in run 1.
[TiCl3(OC6H5)] 1. Phenol (1.5 g, 15.9 mmol) in toluene (30
cm3) was added via a cannula to TiCl4 (3.1 g, 16.3 mmol) in
toluene (40 cm3) and the solution refluxed until the exhaust
gases no longer produced a white cloud when passed over
N,N,NЈ,NЈ-tetramethylethylenediamine (8 h). The solution
was filtered, the solvent removed and the residue washed with
light petroleum (5 × 20 cm3) and dried under vacuum for
several hours. This procedure leaves traces of toluene in the
dark red sample (see analytical data, Table 1). A solvent-free
product has been obtained by preparing the complex in a mix-
ture of CHCl3 and light petroleum and boiling off the CHCl3.8
Conclusion
The results of these studies show that whereas monophenoxide
complexes, [TiCl3(OAr)], can be prepared by a simple thermal-
isation reaction, dynamic processes complicate the solution
chemistry, especially where the phenyl substituents are 2,6-
diisopropyl or 2,4-di-tert-butyl substituents, but more import-
antly when a bidentate donor such as dmbipy is added. The
complex [TiCl3{OC6H2(CMe3)2-2,6-Me-4}] 7, in which the
phenoxide ligand has structural similarities longitudinally
to the 1,3-bis-tert-butylcyclopentadienyl ligand, remains
unchanged in solution and is unaffected by dmbipy. Theoretical
studies show that O(2p) lone pair donation to the phenyl ring
[TiCl3(OC6H4CMe3-4)] 2. p-tert-Butylphenol (4.75 g, 31.6
mmol) in toluene (50 cm3) was added via a cannula to TiCl4 (6.0
g, 31.6 mmol) in toluene (60 cm3) and the mixture refluxed until
HCl gas was no longer produced (12–18 h). The solution was
filtered and the solvent removed to give the complex as a deep
red solid which was dried under vacuum for 3 h. This procedure
leaves traces of toluene in the sample (see analytical data, Table
1). When smaller quantities (ca. 1–2 g) were allowed to stand in
light petroleum for crystallisation during long periods (e.g.
several days) in more dilute solution red crystals of [TiCl2-
(OC6H4CMe3-4)2] were slowly formed and the remaining solu-
tion fumed vigorously in moist air as TiCl4 hydrolysed. Found:
C, 58.0; H, 6.5. C20H26Cl2O2Ti requires C, 57.6; H, 6.3%. After
several such crystallisations the solution was then concentrated
and [TiCl3(OC6H4CMe3-4)] formed within several hours.
Found: C, 40.2; H, 4.7. C10H13Cl3OTi requires C, 39.7; H, 4.3%.
C᎐C (π*) orbital reduces electron donation to the metal and
᎐
structural comparisons of complex 7 and [TiCl3Cp] complexes
reflect this in the longer Ti–Cl bonds in the Cp complexes.
Whereas a phenoxide ligand can be regarded as a lσ, 2π donor,
it is likely to create a more electron deficient titanium centre
than does a Cp ligand when chloro ligands are replaced by
methyl ligands. This may be reflected by the higher catalytic
activity of complex 2 towards ethylene polymerisation than
[TiCl3Cp] when AlMe3 co-catalyst is used.
Preliminary studies indicate that the [TiCl3(OAr)] complexes
are more easily reduced than [TiCl3Cp] and this, coupled with
the solution dynamics, suggests the co-ordination chemistry of
the monophenoxides will be complicated. However, the poten-
tial of the less highly substituted phenoxide complexes, and in
particular 7, to act as catalysts in a variety of situations should
not be underestimated. Studies are at present underway to
establish the wider applicability of these systems.
1
3
NMR (C6D6): H, δ 1.19 (s, 9 H, CMe3); 6.94 [d, J(HH) 8.2,
2 H, o-H] and 7.00 [d, J(HH) 8.2 Hz, 2 H, m-H]; 13C-{1H}
3
δ 31.69 (CMe3); 35.05 (C); 118.82 (o-C); 126.75 (m-C); 150.35
( p-C) and 170.53 (ipso-C).
Reaction of complex 2 with dmbipy. Procedure A. 4,4Ј-
Dimethyl-2,2Ј-bipyridine (0.61 g, 3.3 mmol) in CH2Cl2 (30 cm3)
was added to a rapidly stirred solution of complex 2 (1.0 g, 3.3
mmol) in CH2Cl2 (20 cm3) via a cannula and the stirring con-
tinued for 2 h. The solution was filtered and the solvent
removed to give an orange solid which was then allowed to
stand under light petroleum overnight to give an orange
powder [Found: C, 57.1; H, 5.5; N, 5.3. C22H25Cl3NOTi
Experimental
All preparations and manipulations were carried out under
dry oxygen-free nitrogen using standard bench-top techniques
for air sensitive substances. Titanium tetrachloride and the
phenols were used as received from commercial sources. 4,4Ј-
Dimethyl-2,2Ј-bipyridine and phenanthroline were dried under
vacuum before use. Light petroleum (bp 40–60 ЊC) and toluene
J. Chem. Soc., Dalton Trans., 2000, 529–537
535

View Article Online
1
requires C, 57.0; H, 5.4; N, 5.3%]. The H NMR spectrum in
to TiCl4 (3.0 g, 15.8 mmol) in toluene (30 cm3) and the mixture
CDCl3 shows the presence of mer-[TiCl3(OC6H4CMe3-4)-
(dmbipy)] 3 and [TiCl2(OC6H4CMe3-4)2(dmbipy)] in a ratio of
63:34.
refluxed until the production of HCl ceased (14 h). The solution
was filtered and the solvent removed to give a deep red solid
which, on dissolving in light petroleum (100 cm3) and reducing
the volume, was washed with light petroleum (20 × 30 cm3) and
dried under vacuum.
Procedure B. 2,4,6-Tri-tert-butylphenol (4.2 g, 16.0 mmol) in
toluene (40 cm3) was added to TiCl4 (3.0 g, 15.8 mmol) in tolu-
ene (50 cm3) and the mixture refluxed until the production of
HCl ceased (ca. 16 h). The solution was filtered and the solvent
removed to give a dark red solid which, on dissolving in light
petroleum (100 cm3) and reducing the volume, gave the complex
as dark red microcrystals (yield 3.4 g, 60%) [Found: C, 47.2;
H, 6.1. C14H21Cl3OTi requires C, 46.7; H, 5.9%]. The product
showed identical 1H and 13C-{1H} NMR spectra to the sample
prepared under A.
Procedure B. Complex 2 (0.45 g, 1.5 mmol) was dissolved in
light petroleum (70 cm3) and the solution filtered from a small
amount of solid. To this rapidly stirred solution was added
dmbipy (0.28 g, 1.52 mmol) in light petroleum (10 cm3) and
benzene (20 cm3) to give an immediate orange precipitate. After
stirring for 1 h the solution was filtered and the solid washed
with light petroleum (5 × 20 cm3) and held under vacuum for
1 h [Found: C, 55.7; H, 5.4; N, 5.8. C22H25Cl3NOTiؒ0.33C6H6
requires C, 56.1; H, 5.4; N, 5.5%]. The solid dissolves in CDCl3
1
and then produces a precipitate. A H NMR spectrum shows
mer-[TiCl3(OC6H4CMe3-4)(dmbipy)]
3 and [TiCl2(OC6H4-
CMe3-4)2(dmbipy)] in a ratio of 20:80. Benzene solvent (δ 7.28)
decreased after heating the solid under vacuum.
[TiCl3{OC6H2(CMe3)2-2,6-OMe-4}] 9. 2,6-Di-tert-butyl-4-
methoxyphenol (3.72 g, 15.7 mmol) in toluene (50 cm3) was
added to TiCl4 (3.0 g, 15.8 mmol) in toluene (50 cm3) and the
mixture refluxed until the production of HCl ceased (11 h). The
solution was filtered and the solvent removed to give a gum
which gave the complex as a purple crystalline mass after
extended drying under vacuum (5 h).
[TiCl3(OC6H2Me3-2,4,6)] 4. 2,4,6-Trimethylphenol (2.1 g,
15.4 mmol) in toluene (50 cm3) was added to TiCl4 (3.0 g, 15.8
mmol) in toluene and the mixture refluxed until the production
of HCl ceased (10 h). The solution was filtered, the solvent
removed and the dark red solid washed with light petroleum
(5 × 20 cm3). The residue was then dried under vacuum for
several hours. This procedure leaves a trace of toluene in the
sample (see analytical data, Table 1).
[TiCl3(OC6H3CMe3-2-Me-4)] 10. 2-tert-Butyl-4-methyl-
phenol (2.6 g, 15.8 mmol) in toluene (50 cm3) was added to
TiCl4 (3.0 g, 15.8 mmol) in toluene (30 cm3) and the mixture
refluxed until production of HCl gas ceased (13 h). The solu-
tion was filtered and the solvent removed to give a deep red
solid which was dried under vacuum for 4 h [Found: C, 42.6;
H, 5.4. C11H15Cl3OTiؒ0.0625C7H8 requires C, 42.6; H, 4.9%].
The solid was dissolved in boiling light petroleum (120 cm3),
the solution filtered while hot and the volume reduced while
keeping the solution hot to give an analytically pure sample
as a deep red solid.
[TiCl3(OC6H3Pri2-2,6)] 5. 2,6-Diisopropylphenol (4.7 g, 26.4
mmol) in toluene (50 cm3) was added to TiCl4 (5.0 g, 26.4
mmol) in toluene (50 cm3) and the mixture refluxed until the
production of HCl ceased (12 h). The solution was filtered and
the solvent removed to give a deep red gum which solidified on
gentle heating (water bath 60–70 ЊC) under vacuum for several
hours.
Reaction of complex 5 with dmbipy. Procedure A. The com-
pound dmbipy (0.47 g, 2.6 mmol) in CH2Cl2 (30 cm3) was added
to a rapidly stirred solution of complex 2 (0.85 g, 2.6 mmol) via
a cannula and stirring continued for 2 h. The orange solution
was filtered, the solvent removed and the residue allowed to
stand under light petroleum (50 cm3) for 2 d. Filtration gave an
orange powder [Found: C, 56.4; H, 5.8; N, 5.3. C24H29Cl3NOTi
[TiCl3(OC6H4Ph-2)] 11. 2-Phenylphenol (1.8 g, 10.6 mmol)
in CH2Cl2 (25 cm3) was added to TiCl4 (2.0 g, 10.6 mmol) in
CH2Cl2 (25 cm3) and the mixture refluxed until the production
of HCl gas ceased (12.5 h). The solution was filtered, the
solvent removed and the residue held under vacuum for 4 h
giving the complex as a deep red solid. The complex is partially
soluble in CHCl3, less so in benzene or toluene and only slightly
soluble in light petroleum.
1
requires C, 55.9; H, 5.7; N, 5.4%]. The H NMR spectrum
in CDCl3 shows the presence of mer-[TiCl3(OC6H3Pri2-2,6)-
(dmbipy)] 6 and [TiCl2(OC6H3Pri2-2,6)2(dmbipy)] in a ratio of
55:45.
Procedure B. The compound dmbipy (0.35 g, 1.9 mmol) in
benzene (25 cm3) was added rapidly to complex 5 (0.63 g, 1.9
mmol) in light petroleum (70 cm3) and the mixture stirred
for 1 h. The orange solid was filtered off, washed with light
petroleum (2 × 30 cm3) and held under vacuum for 2 h [Found:
C, 60.5; H, 6.2; N, 5.4. C24H29Cl3NOTiؒ0.83C6H6 requires C,
60.7; H, 6.0; N, 4.7%]. The solid dissolves in CDCl3 and then
produces a precipitate. A 1H NMR spectrum shows mer-
[TiCl3(OC6H3Pri2-2,6)(dmbipy)] 6 and [TiCl2(OC6H3Pri2-2,6)2-
(dmbipy)] in a ratio of 55:45. Benzene solvent (δ 7.28)
decreases after heating the solid in vacuum.
[TiCl3(OC10H7)] 12. 1-Naphthol (2.25 g, 15.6 mmol) in tolu-
ene (50 cm3) was added to TiCl4 (3.0 g, 15.8 mmol) in toluene
(20 cm3) and the mixture refluxed for 11.5 h. The solution
was filtered, the solvent removed and the product held under
vacuum for 4 h. This procedure leaves toluene in the sample (see
analytical data, Table 1). The solid is only slightly soluble in
CDCl3 but the 1H NMR shows the presence of extracted
toluene.
Polymerisations
Polymerisations were performed in a 400 cm3 flame-dried pres-
sure bottle equipped with a head containing inlet and outlet
taps and a pressure gauge. Toluene (20 cm3) and AlMe3 in tolu-
ene [1 cm3 of a 0.072 g cmϪ3 solution (approximately 7 mol
equivalents)] were added via a syringe to the vessel which con-
tained a Teflon stirring bar and the mixture was saturated with
ethylene until the head pressure remained at 6 psi. Complex 2
(0.045 g, 1.5 mmol) in toluene (29 cm3) was added via a cannula
under nitrogen keeping a stream of ethylene flowing from the
pressure bottle during the addition. The mixture was stirred
vigorously for 1 h at 21 ЊC while maintaining the head pressure
at 6 psi. The polymerisation was terminated by degassing the
solution (bubbling N2 through the solution) and adding meth-
[TiCl3{OC6H2(CMe3)2-2,6-Me-4}] 7. 2,6-Di-tert-butyl-4-
methylphenol (2.3 g, 10.5 mmol) in light petroleum (50 cm3)
was added to TiCl4 (2.0 g, 10.5 mmol) in light petroleum (30
cm3) and the mixture refluxed until the production of HCl
ceased (12–15 h). The solution was filtered and the solvent
removed to give a solid which is essentially pure (NMR spec-
troscopy). An analytically pure crystalline sample was obtained
by dissolving a portion of the bulk sample in hot light petrol-
eum and allowing the sample to stand.
[TiCl3{OC6H3(CMe3)2-2,4}] 8. Procedure A. 2,4-Di-tert-
butylphenol (3.25 g, 15.8 mmol) in toluene (50 cm3) was added
536
J. Chem. Soc., Dalton Trans., 2000, 529–537

View Article Online
D. Jeremic, Q. Wang, R. Quyoum and M. C. Baird, J. Organomet.
anol (100 cm3) containing 5% HCl solution. The polyethylene
was filtered off, broken into small pieces, washed extensively
with methanol to remove impurities and dried under vacuum to
constant weight.
Chem., 1995, 497, 143; S. W. Ewart, M. J. Sarsfield, D. Jeremic,
T. L. Tremblay and M. C. Baird, Organometallics, 1998, 17, 1502.
4 G. Ricci and L. Porri, Macromolecules, 1994, 27, 868.
5 N. Ishichara, T. Seimiya, M. Kuramoto and M. Uoi,
Macromolecules, 1986, 19, 2464; N. Ishihara, M. Kuramoto and
M. Uoi, Macromolecules, 1988, 21, 3356; A. Zambelli, L. Oliva and
C. Pellecchia, Macromolecules, 1989, 22, 2129; J. C. W. Chien,
Z. Salajka and S. Dong, Macromolecules, 1992, 25, 3199; T. E.
Ready, R. O. Day, J. C. W. Chien and M. D. Rausch,
Macromolecules, 1993, 26, 5822; D. J. Duncalf, H. J. Wade,
C. Waterson, P. J. Derrick, D. M. Haddleton and A. McCamley,
Macromolecules, 1996, 29, 6399; W. Kaminsky, S. Lenk, V. Scholz,
H. W. Roesky and A. Herzog, Macromolecules, 1997, 30, 7647.
6 D. C. Bradley, R. C. Mehrotra and D. P. Gaur, Metal Alkoxides,
Academic Press, New York, 1978; K. C. Malhotra and R. L. Martin,
J. Organomet. Chem., 1982, 239, 159.
7 M. H. Chisholm, Chemtracts, 1992, 4, 273.
8 G. P. Lutschinsky and E. S. Al’tman, Z. Anorg. Allg. Chem., 1935,
225, 321.
9 R. W. Crowe and A. N. Caughlan, J. Am. Chem. Soc., 1950, 72,
1694.
10 S. Shah, A. Singh and R. C. Mehrotra, Polyhedron, 1986, 3, 1285.
11 G. Sartori, G. Casnati, F. Bigi and G. Predieri, J. Org. Chem, 1990,
55, 4371.
X-Ray crystallography
Crystals of complex 7 were grown from a light petroleum solu-
tion. Data were collected on a Siemens SMART diffractometer.
The collection covered a nominal sphere of reciprocal space, by
a combination of four sets of exposures. Each set had a differ-
ent φ angle for the crystal and each exposure covered 0.3Њ in α.
Coverage of the unique data set is at least 98% complete to 56Њ
in 2θ. Crystal decay was monitored by repeating the initial
frames at the end of data collection and analysing the duplicate
reflections. Unit cell parameters were obtained by a least
squares fit of all data with I > 10σ(I). Data were corrected for
Lorentz-polarisation and absorption effects. The structure was
solved by direct methods and refined by the full-matrix least-
squares technique. All non-hydrogen atoms were allowed to
assume anisotropic thermal motion. Hydrogen atoms were in
calculated positions (C–H, 0.96 Å) and refined with a riding
model with Uiso = 0.05. Programs used were SHELXS26 for
structure solution and SHELXL27 for refinement. Diagrams
were prepared with ORTEP 3.28
12 D. L. Barnes, N. W. Eilerts and J. A. Heppert, Polyhedron, 1994, 13,
743.
13 P. Aaltonen, J. Seppala, L. Matilainen and M. Leskela, Macro-
molecules, 1994, 27, 3136.
14 L. Matilainen, M. Klinga and M. Leskela, Polyhedron, 1996, 15,
153.
15 M. W. Glenny, A. J. Nielson and C. E. F. Richard, Polyhedron, 1998,
17, 851.
16 P. Sobota and S. Szafert, Inorg. Chem., 1996, 35, 1778; G. Morini,
E. Albizzati, G. Balbontin, I. Mingozzi, M. C. Sacchi, F. Forlini and
I. Tritto, Macromolecules, 1996, 29, 5770; P. Sobota and S. Szafert,
Inorg. Chem., 1996, 35, 1778; J. Xu, L. Feng, S. Yang, Y. Yang and
X. Kong, Macromolecules, 1997, 30, 7655.
17 J. R. Dilworth, J. Hanich and M. Krestel, J. Organomet. Chem., 1986,
315, C9.
Crystal data. C15H23Cl3OTi, M = 373.58, monoclinic, space
group P21/c, a = 17.294(3), b = 6.1220(10), c = 17.833(4) Å, β =
108.36(3)Њ, U = 1791.9(6) Å3, T = 203 K, Z = 4, µ(Mo-Kα) =
0.918 mmϪ1, 3416 observed reflections, final wR(F 2) for all
4032 data 0.1024, R1 = 0.0466.
CCDC reference number 186/1795.
See http://www.rsc.org/suppdata/dt/a9/a908435e/ for crystal-
lographic files in .cif format.
18 Y. Morino and H. Uehara, J. Chem. Phys., 1966, 45, 4543.
19 L. M. Engelhardt, R. I. Papasergio, C. L. Raston and A. H. White,
Organometallics, 1984, 3, 18.
20 C. H. Winter, X.-X. Zhou, D. A. Dobbs and J. Heeg, Organo-
metallics, 1991, 10, 210.
21 I. Jibril, S. Abu-Orabi, S. A. Klaib, W. Imhof and G. Huttner,
J. Organomet. Chem., 1992, 433, 253.
22 H.-M. Gau, C.-S. Lee, C.-C. Lin, M.-K. Jiang, Y.-C. Ho
and C.-N. Kuo, J. Am. Chem. Soc., 1996, 118, 2936.
23 S. L. Latesky, J. Keddington, A. K. McMullen, I. P. Rothwell and
J. C. Huffman, Inorg. Chem., 1985, 24, 995.
Theoretical
Density functional calculations29 were carried out on the com-
plexes [TiCl3(OC6H2Me3-2,4,6)] 4, [TiCl3(OMe)] and TiCl3Me.
The hybrid Becke-3 parameter functional (B3)30 together with
the Lee–Yang–Parr correlation functional (LYP)31 has been
used in all calculations. Owing to the large size of molecule 4
the basis set was restricted to a Dunning/Huzinaga valence
double-zeta set for H, C and O32 using Hay–Wadt pseudo-
potentials with valence double-zeta basis sets for Cl and Ti.33
This resulted in 378 basis functions contracted to 158 and the
geometry optimisation required several days on a 16-processor
R10000 SGI supercomputer. At the optimised geometry a
subsequent natural bond orbital analysis was carried out.
24 R. L. Martin and G. Winter, J. Chem. Soc., 1961, 2947.
25 C. N. Field, J. C. Green, N. Kaltsoyannis, G. S. McGrady,
A. N. Moody, M. Siggel and M. De Simone, J. Chem. Soc., Dalton
Trans., 1997, 213.
26 G. M. Sheldrick, SHELXS 86, Acta Crystallogr., Sect. A, 1990, 46,
467.
27 G. M. Sheldrick, SHELXL 93, Program for the refinement of crystal
structures, University of Göttingen, 1993.
28 ORTEP 3 for WINDOWS, Based on ORTEP III, C. K. Johnson and
M. N. Burnette, Report ORNL-6895, modified by L. J. Farrugia,
University of Glasgow, 1996.
29 M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G.
Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A. Petersson,
J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V. G.
Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B.
Stefanov, A. Nanayakkara, M. Challacombe, C. Y. Peng, P. Y.
Ayala, W. Chen, M. W. Wong, J. L. Andres, E. S. Replogle,
R. Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley, D. J. DeFrees,
J. Baker, J. J. P. Stewart, M. Head-Gordon, C. Gonzalez and
J. A. Pople, GAUSSIAN 94, Revision D.1, Gaussian, Inc.,
Pittsburgh, PA, 1996.
Acknowledgements
P. S. thanks the High Performance Computer Committee
(HPCC), the Auckland University Research Committee
(AURC) and the Marsden Fund Wellington (contact number
96-UOA-PSE-0081) for financial support.
References
1 H. H. Brintzinger, D. Fischer, R. Mulhaupt, B. Rieger and
R. Waymouth, Angew. Chem., Int. Ed. Engl., 1995, 34, 1143;
M. Bochmann, J. Chem. Soc., Dalton Trans., 1996, 255; W.
Kaminsky, J. Chem. Soc., Dalton Trans., 1998, 1413; G. J. P.
Britovsek, V. C. Gibson and D. F. Wass, Angew. Chem., Int. Ed.,
1999, 38, 429.
2 T. V. Lubben, P. T. Wolczanski and G. D. Van Duyne,
Organometallics, 1984, 3, 977.
3 K. Soga, J. R. Park and Y. Shiono, Polym. Commun., 1989, 22, 2126;
J. R. Park, T. Shiono and K. Soga, Macromolecules, 1992, 25, 521;
C. Pellecchia, A. Immirzi, A. Grassi and A. Zambelli, Organo-
metallics, 1993, 12, 4473; R. Quyoum, Q. Wang, M. J. Tudoret,
M. C. Baird and D. J. Gillis, J. Am. Chem. Soc., 1994, 116, 6435;
30 A. D. Becke, Phys. Rev. A, 1988, 38, 3098.
31 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785.
32 T. H. Dunning and P. J. Hay, in Modern Theoretical Chemistry,
ed. H. F. Schaefer III, Plenum, New York, 1976, pp. 1–28.
33 W. R. Wadt and P. J. Hay, J. Chem. Phys., 1985, 82, 284.
Paper a908435e
J. Chem. Soc., Dalton Trans., 2000, 529–537
537