Synthesis, structure, coordination expansion and theoretical modelling of dichlorobis(phenoxo) titanium(IV) complexes

Article abstract of DOI:10.1002/ejic.200400769

Thermalisation of TiCl₄ and two equivalents of a phenol in toluene is found to be the best preparative method for quantitative yields of a variety of dichlorobis(phenoxo) complexes. [TiCl₂(OC ₆H₄CMe₃-4)₂] (1) is monomeric in benzene, a phenoxo-bridged dimer in the solid state and undergoes coordination expansion with 4,4′-dimethyl-2,2′-bipyridine (dmbipy) to give [TiCl₂(OC₆H₄CMe₃-4) ₂(dmbipy)] (2). Also monomeric are [TiCl₂(OC ₆H₂Me₃-2,4,6)₂] (3) and [TiCl ₂(OC₆H₃iPr₂-2,6)₂] (5) which expand their coordination with dmbipy to give [TiCl₂(OC ₆H₂Me₃-2,4,6)₂(dmbipy)] (4) and [TiCl₂(OC₆H₃iPr₂-2,6) ₂(dmbipy)] (6). In contrast [TiCl₂(OC₆H ₂{CMe₃}₂-2,6-Me-4)₂] (7) is only partially formed by the thermolysis reaction and does not coordinatively expand with dmbipy. [TiCl₂(OC₆H₃Me₂-2,4) ₂] (8) is monomeric in benzene and reacts to form [TiCl ₂(OC₆H₃Me₂-2,4)₂(dmbipy)] (9). [TiCl₂(OC₆H₃{CMe₃} ₂-2,4)₂] (10) forms along with the tri-phenoxo complex [TiCl{OC₆H₃(CMe₃)₂-2,4} ₃]. [TiCl₂(OC₆H₃CMe ₃-2-Me-6)₂] (11) is monomeric in benzene and forms [TiCl₂(OC₆H₃CMe₃-2-Me-6) ₂(dmbipy)] (12). 2-Phenylphenol and 1-napthol form [TiCl ₂(OC₆H₄Ph-2)₂] (13) and [TiCl ₂(OC₁₀H₉)₂] (14) which are monomeric in benzene. DFT calculations give structural parameters for monomeric [TiCl₂(OC₆H₅)₂] (15) in good agreement with the X-ray data for [TiCl₂(OC₆H ₃Me₂-2,6)₂]. Each oxygen of the phenoxo ligand in 15 acts essentially as a 2π donor to titanium and there is substantial p_π(O)-p_π*(C=C) backbonding with the phenyl ring, which is absent in [TiCl₂(OCH₃)₂] (16). The global minimum for the dimer [(TiCl₂{OC₆H ₅}{μ-OC₆H₅})₂] is in almost perfect agreement with the crystal structure obtained for [(TiCl ₂{OC₆H₅}{μ-OC₆H ₅})₂] or [(TiCl₂{OC₆H ₄CMe₃-4}{μ-OC₆H₄CMe ₃-4})₂] (1). The dimerisation energy for 15 is -26.2 kJ·mol^-1.

Full text of DOI:10.1002/ejic.200400769

FULL PAPER
Synthesis, Structure, Coordination Expansion and Theoretical Modelling of
Dichlorobis(phenoxo) Titanium(IV) Complexes
Alastair J. Nielson,*[a] Chaohong Shen,[a] Peter Schwerdtfeger,[a] and Joyce M. Waters[a]
Keywords: Complexes / Phenoxo ligands / Structure elucidation / Theoretical structure / Titanium
Thermalisation of TiCl4 and two equivalents of a phenol in
toluene is found to be the best preparative method for quan-
titative yields of a variety of dichlorobis(phenoxo) complexes.
[TiCl2(OC6H4CMe3-4)2] (1) is monomeric in benzene, a phe-
[TiCl2(OC6H3CMe3-2-Me-6)2] (11) is monomeric in benzene
and forms [TiCl2(OC6H3CMe3-2-Me-6)2(dmbipy)] (12). 2-
Phenylphenol and 1-napthol form [TiCl2(OC6H4Ph-2)2] (13)
and [TiCl2(OC10H9)2] (14) which are monomeric in benzene.
noxo-bridged dimer in the solid state and undergoes coordi- DFT calculations give structural parameters for monomeric
nation
expansion
with
4,4Ј-dimethyl-2,2Ј-bipyridine
[TiCl2(OC6H5)2] (15) in good agreement with the X-ray data
for [TiCl2(OC6H3Me2-2,6)2]. Each oxygen of the phenoxo li-
gand in 15 acts essentially as a 2π donor to titanium and there
is substantial pπ(O)-pπ*(C=C) backbonding with the phenyl
ring, which is absent in [TiCl2(OCH3)2] (16). The global mini-
mum for the dimer [(TiCl2{OC6H5}{μ-OC6H5})2] is in almost
perfect agreement with the crystal structure obtained for
[(TiCl2{OC6H5}{μ-OC6H5})2] or [(TiCl2{OC6H4CMe3-4}{μ-
OC6H4CMe3-4})2] (1). The dimerisation energy for 15 is
(dmbipy) to give [TiCl2(OC6H4CMe3-4)2(dmbipy)] (2). Also
monomeric are [TiCl2(OC6H2Me3-2,4,6)2] (3) and [TiCl2(O-
C6H3iPr2-2,6)2] (5) which expand their coordination with
dmbipy to give [TiCl2(OC6H2Me3-2,4,6)2(dmbipy)] (4) and
[TiCl2(OC6H3iPr2-2,6)2(dmbipy)]
(6).
In
contrast
[TiCl2(OC6H2{CMe3}2-2,6-Me-4)2] (7) is only partially formed
by the thermolysis reaction and does not coordinatively ex-
pand with dmbipy. [TiCl2(OC6H3Me2-2,4)2] (8) is monomeric
in benzene and reacts to form [TiCl2(OC6H3Me2-2,4)2-
(dmbipy)] (9). [TiCl2(OC6H3{CMe3}2-2,4)2] (10) forms along
with the tri-phenoxo complex [TiCl{OC6H3(CMe3)2-2,4}3].
–26.2 kJ·mol–1
.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
ArNH2 and tetrahydrofuran) which have been characterised
only by analytical data and IR spectroscopy.[14] We have
reported[15] a comprehensive systematic study of the mono-
(phenoxo) complexes [TiCl3(OAr)] (Ar = unsubstituted or
substituted phenyl group) which included aspects of best
synthetic method, co-ordination expansion and theoretical
modelling. We report here a similar investigation for com-
plexes of the type [TiCl2(OAr)2] where changes are made
to the size of the ortho-substituents of the phenyl rings. In
particular we investigate the nature of the complexes in
non-coordinating solvents to ascertain if co-ordinative un-
saturation is maintained in solution. Some general compari-
sons of the bis(phenoxo) complexes with [TiCl2Cp2] (Cp =
cyclopentadienyl) are also made.
Introduction
The relationship between dichlorobis(phenoxo) com-
plexes [TiCl2(OAr)2] and titanium metallocences [TiCl2Cp2]
is now very apparent[1,2] with the phenoxo complexes
emerging as promising synthons and catalysts for organic
reactions[2,3,4] and catalysts for olefin polymerisation.[5]
Alkoxo complexes of titanium often form higher associates
than monomers in solution[6] but the solution behavior of
the bis(phenoxo) complexes and the coordination expan-
sion properties have not been studied.[7] From the synthetic
point of view, complexes containing tert-butyl or phenyl
groups in the 2,6- positions of the aryloxo ring have re-
ceived the most attention.[5,8–10] However, little is known
about bis(phenoxo) complexes containing ligands which are
less sterically demanding. For example, [TiCl2(OC6H5)2]
has a dimeric structure in the solid state[11] and there is
conflicting evidence for the structure in solution,[12,13]
[TiCl2(OC6H3Me2-2,6)2] has a monomeric solid-state struc-
ture,[9] and [TiCl2(OC6H2Me3-2,4,6)2] is monomeric in ben-
zene and expands its co-ordination sphere forming the bis-
Results and Discussion
Synthetic Studies
Bis(phenoxide) complexes have been prepared by ther-
malising TiCl4 and the phenol in a non-coordinating sol-
vent[14,16] or more commonly by reacting the lithium phe-
nolate with TiCl4 in diethyl ether.[2–10] We have found that
thermalisation in toluene of TiCl4 and the phenol in 1:2
ratio is the best procedure for most phenols [Equation (1)].
adducts [TiCl2(OC6H2Me3-2,4,6)2(L)2] (L
= pyridine,
[a] Chemistry, Institute of Fundamental Sciences, Massey Univer-
sity at Albany,
Private Bag 102904, North Shore Mail Centre, Auckland, New
Zealand
Eur. J. Inorg. Chem. 2005, 1343–1352
DOI: 10.1002/ejic.200400769
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1343

A. J. Nielson, C. Shen, P. Schwerdtfeger, J. M. Waters
FULL PAPER
TiCl4 + 2 HOAr Ǟ [TiCl2(OAr)2] + 2 HCl
(1)
Ti–O(1)–C(1) bond angle [165.2(2) ] are similar to those
found for the mono(phenoxo) complex [TiCl3(OC6H2-
{CMe3}2-2,6-Me-4)] [1.750(2), 1.390(2) Å and 163.1°,
respectively[15]]. Theoretical calculations on the model com-
plex [TiCl3(OC6H2Me3-2,4,6)] show that the terminal phe-
noxo ligand oxygen atom acts essentially as a 1-σ, 2-π do-
nor to titanium but that there is also some O(2p) donation
to the phenyl ring C=C(π*) orbital.[15] The TiϪO(2) bond
length in 1 [1.912(2) Å], is significantly longer than the Ti–
O(1) bond length [1.749(2) Å] as a consequence of O(2)
forming a dative bond to the other titanium atom [Ti–O(2)Ј
bond length 2.142(2) Å] in the unsymmetrical bridge. The
O(2) lone pair donation to Ti occurs trans to the strongly
π-donating terminal phenoxo ligand so that the potentially
π-donating chloro ligands do not have to compete for the
same d orbitals. The increased length of the Ti–O(2) bond
indicates that O(2)-to-Ti π-donation is much less than for
the Ti–O(1) bond. Also the bridging phenoxo ligand O–C
bond [O(2)–C(11) bond length 1.406(3) Å] is significantly
longer than the terminal phenoxo ligand O–C bond [O(1)–
C(1) bond length [1.369(3) Å] indicating there is less O(2p)
donation to the phenyl ring C=C(π*) orbital for the bridg-
ing mode. This may be a consequence of the phenyl ring π-
orbitals being unable to overlap to any significant extent
with oxygen orbitals when the two oxygen lone pairs on
O(2) are otherwise involved with bonding to Ti [possible
O(2)p-to-Ti(1)d π-donation and O(2)p-to-Ti(2)d σ-do-
nation].
HOAr = HOC6H4CMe3-4, HOC6H2Me3-2,4,6, HOC6H3iPr2-2,6,
HOC6H3Me2-2,4, HOC6H3(CMe3)2-2,4, HOC6H3CMe3-2-Me-6,
HOC6H4Ph-2, HOC10H9 (naphthyl)
If the reaction is carried out to completion, then solid
complexes can be obtained in virtually quantitative yield
and in an analytically pure form. It is important to deter-
mine the end point of the reaction otherwise gummy mate-
rials often result which are difficult to purify. The last traces
of HCl gas produced in the reaction are easily detected by
a simple but very sensitive test which involves passing the
exhaust gases close to concentrated ammonia solution or a
liquid organic amine and observing any white hydrochlo-
ride cloud. On a 5–10 g scale in vigorously boiling toluene,
the reactions are usually complete in less than 14 h and on
a 50–100 g scale the reaction times are not much longer.
The crystal structure of [TiCl2(OC6H5)2] has been deter-
mined,[11] but we have found the complex is not suitably
soluble in benzene for an accurate cryoscopic molecular
weight determination. However, reaction of TiCl4 with 4-tert-
butylphenol gave highly soluble [TiCl2(OC6H4CMe3-4)2] (1)
and an X-ray crystal structure determination of the product
crystallised from benzene showed a dimeric complex con-
taining a phenoxo bridge in which each titanium atom
adopts a distorted pentagonal bipyrimidal coordination
geometry (Figure 1). The titanium atoms each have two cis-
coordinated chloro ligands and two cis-coordinated phen-
oxo ligands, one of which bridges to the adjacent titanium
atom where it lies trans to a terminal phenoxo ligand.
Table 1. Selected bond lengths [Å] and angles [°] for 1. Symmetry
transformations used to generate equivalent atoms: #1: – x + 1,
Ϫy + 1, –z + 1).
Ti–O(1)
Ti–O(2)
Ti–O(2)#1
Ti–Cl(1)
Ti–Cl(2)
Ti–Ti#1
O(1)–C(1)
O(2)–C(11)
1.749(2)
1.912(2)
2.142(2)
2.2186(11)
2.2256(11)
3.282(2)
1.369(3)
1.406(3)
2.142(2)
96.03(9)
167.96(9)
71.99(9)
96.52(8)
122.74(8)
89.25(6)
97.43(8)
119.23(7)
89.78(7)
114.06(5)
134.38(7)
38.36(6)
107.85(4)
106.38(4)
165.2(2)
123.5(2)
1.749(2)
108.01(9)
O(2)ϪTi#1
O(1)–Ti–O(2)
O(1)–Ti–O(2)#1
O(2)–Ti–O(2)#1
O(1)–Ti–Cl(1)
O(2)–Ti–Cl(1)
O(2)#1ϪTi–Cl(1)
O(1)–Ti–Cl(2)
O(2)–Ti–Cl(2)
O(2)#1ϪTi–Cl(2)
Cl(1)–Ti–Cl(2)
O(1)–Ti–Ti#1
O(2)–Ti–Ti#1
Cl(1)–Ti–Ti#1
Cl(2)–Ti–Ti#1
C(1)–O(1)–Ti
C(11)–O(2)–Ti
Ti–O(1)
Figure 1.Thermal ellipsoid diagram (at the 50% probability level)
for (1) showing the numbering system. Hydrogen atoms have been
omitted for clarity.
Ti–O(2)–Ti#1
The structure of 1 is essentially similar to that found for
dimeric [TiCl2(OC6H5)2],[11] but is described now in terms
of modern bonding concepts. Selected bond lengths and
The Ti–Cl(1) and Ti–Cl(2) bond lengths in complex 1
angles are given in Table 1. The Ti–O(1) and O(1)–C(1) [2.2186(11) and 2.2256(11) Å] are not very different from
bond lengths [1.749(2) and 1.369(3) Å, respectively] and the each other but are slightly longer than those found for
1344
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 1343–1352

Dichlorobis(phenoxo) Titanium(iv) Complexes
FULL PAPER
1
[TiCl3(OC6H2{CMe3}2-2,6-Me-4)] [Ti–Cl bond lengths showed the presence of only one isomer. Both the H and
2.1822(8), 2.1913(8) and 2.1945(9) Å] where π-donation 13C{1H} NMR spectra indicate a symmetrical molecule
from chloro ligands is needed to increase the overall elec- based on a single set of resonances for the relevant position
tron count. The theoretical calculations made on the model of the phenoxo and dmbipy ligands. Structure I (Scheme 2)
complex [TiCl3(OC6H2Me3-2,4,6)] show that significant is the preferred isomer since structure II contains trans-ori-
Cl(2p) to Ti(3d) π-donation occurs in these tetrahedral entated phenoxo ligands, which would lead to an unfavour-
molecules.[15] Assuming that O(2) in complex 1 makes a sin- able competition for d-metal orbitals and structure III is an
gle π-donor interaction with the titanium atom then the unsymmetrical molecule. A trans-chloro-cis-phenoxo con-
overall (formal) electron count for the complex is 16 so an figuration has been observed in the X-ray crystal structures
18-electron count can be attained if Cl(2p) to Ti(3d) do- [TiCl2(OC6H3Me2-2,6)2(THF)2][17] [i.e. consistant with
nation occurs.
structure I, Scheme 2], where the THF ligands coordinate
The X-ray crystal structure indicates complex 1 exists in cis to each other and trans to the strongly π-donating phe-
the solid state as a dimer but a molecular weight determi- noxo ligands. The NMR spectra for 2 show sharp reso-
nation in benzene, the solvent from which the dimer crystal- nances, which suggests that solution dynamics are not prev-
lised, indicates a monomer (Scheme 1). The 1H NMR spec- alent. The ipso-C resonance occurs at δ = 164.8 ppm, which
trum in C6D6 shows a single resonance for the tert-butyl lies slightly upfield to that observed for complex 1 (δ =
groups, a doublet for the meta-protons of the phenyl ring 166.3 ppm) and again this may represent a small change
but a broadened resonance for the ortho-protons. In the in the O(2p)–C=C(π*) bonding system associated with the
13C{1H} NMR spectrum the tert-butyl group methyl and smaller need for tight O(2)p–to–Ti(1)d orbital donation on
quaternary carbons appear as single sharp resonances but addition of the 4-electron donor bipy chelate. (see Table 2
the aromatic ring C resonances are somewhat broadened. for comparisons of the 13C{1H} spectra ipso-C resonance
Identical spectra are obtained for the complexes reported positions for complexes of the type [TiCl3(OAr)],
here whether the solvent is C6D6 or CDCl3 but the reso- [TiCl2(OAr)2] and [TiCl2(OAr)2(dmbipy)]).
nances are slightly sharper in CDCl3. Cooling CDCl3 or
[D8]toluene solutions of 1 produced little change in the
broadening observed leaving unclear the nature of any solu-
tion dynamics. The spectra are thus consistent with the mo-
nomeric structure shown by the molecular weight determi-
nation and not the dimer found in the solid state. The ipso-
C resonance occurs at δ = 166.3 ppm, which lies slightly
upfield to that observed for [TiCl3(OC6H4CMe3-4)] (δ =
169.6 ppm[15]) and this may represent a small change in the
O(2p)–C=C(π*) bonding system associated with the smaller
need for tight O(2)p-to-Ti(1)d orbital donation on addition
of the second phenoxide ligand which is a better π-bonder
than Cl.[15]
Scheme 1.
The coordination expansion properties of complex 1
were studied using 4,4Ј-dimethyl-2,2Ј-bipyridine (dmbipy),
Scheme 2.
as this ligand is useful in determining isomer structure by
NMR spectroscopy.[15] [TiCl2(OC6H4CMe3-4)2(dmbipy)]
Thermalisation of 2,4,6-trimethylphenol and TiCl4 gave
(2) was readily formed in CH2Cl2 and the NMR spectra [TiCl2(OC6H2Me3-2,4,6)2][14] (3) which is monomeric in
Eur. J. Inorg. Chem. 2005, 1343–1352
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1345

A. J. Nielson, C. Shen, P. Schwerdtfeger, J. M. Waters
FULL PAPER
Table 2. Comparison of 13C{1H} spectra ipso-C resonance positions for complexes of the type [TiCl3(OAr)], [TiCl2(OAr)2] and [TiCl2-
(OAr)2(dmbipy)].
Complex[a]
Ligand (OAr)
(OC6H4CMe3-4)
(OC6H2Me3-2,4,6)
(OC6H3iPr2-2,6)
(OC6H3Me2-2,4)
(OC6H3CMe3-2-Me-6)
(OC6H3{CMe3}2-2,6-Me-4)
[TiCl3(OAr)][b]
[TiCl2(OAr)2]
[TiCl2(OAr)2(dmbipy)]
169.6
166.3 (1)
166.4 (3)
165.7 (5)
165.2 (8)
168.3 (11)
170.0
164.8 (2)
164.5 (4)
162.6 (6)
164.3 (9)
[c]
170.5
169.9
172.8
174.9
166.4 (12)
[d]
[a] Spectra obtained in CDCl3. [b] Data taken from ref.[15]. [c] ipso-C resonance not observed due to long relaxation times.[15] [d] Dmbipy
complex does not form.
benzene and shows sharper NMR spectra than complex 1 ents represent the steric limit for coordination expansion
suggesting again that solution dynamics are not prevalent. about the titanium centre.
The complex reacts with dmbipy to give [TiCl2(OC6H2Me3-
Several unsymmetrically substituted phenols were also
2,4,6)2(dmbipy)] (4) for which the NMR spectra indicate thermalised with TiCl4 in toluene giving bis(phenoxo) com-
one isomer with a similar structure to 2. The ipso-C reso- plexes which have not been reported before.
nance in the 13C{1H} NMR spectrum of 4 occurs at δ = [TiCl2(OC6H3Me2-2,4)2] (8) (ipso-C resonance at δ = 165.2
164.5 ppm compared with δ = 166.4 ppm in the parent com- ppm, cf. δ = 169.9 ppm for [TiCl3(OC6H3Me2-2,4)],[15]
plex 3 (Table 2). Both complexes 3 and 4 did not crystallise, Table 2) is monomeric in benzene and reacts with dmbipy
so that comparisons of ipso-C resonance position with solid to give [TiCl2(OC6H3Me2-2,4)2(dmbipy)] (9) for which the
state bond length data cannot be made. However, complex NMR spectra show a single isomer similar to structure I,
3 is closely related to [TiCl2(OC6H3Me2-2,6)2] for which an Scheme 2 and an ipso-C resonance at δ = 164.3ppm. Reac-
X-ray crystal structure determination has shown a mono- tion of 2,4-di-tert-butylphenol and TiCl4 gave a mixture of
meric tetrahedral structure with Ti–O bond lengths of [TiCl2(OC6H3{CMe3}2-2,4)2] (10) (identified as a monomer
1.734(7) and 1.736(8) Å, Ti–Cl bond lengths of 2.192(4) and by NMR spectroscopy, one set of resonances in the NMR
2.211(4) Å[9] and an ipso-C resonance in the 13C NMR spectra, ipso-C resonance at δ = 165.7 ppm) and the tris-
spectrum at δ = 167.8 ppm.[8] In the octahedral adduct (phenoxo) complex [TiCl(OC6H3{CMe3}2-2,4)3], which
[TiCl2(OC6H3Me2-2,6)2(dpeda)] [dpeda = (1S,2S)-dipheny- could be crystallised from the solution and has a distinctly
lethylenediamine] the Ti–O bond lengths increase to different ipso-C resonance position (δ = 163.2 ppm). NMR
1.804(4) and 1.787(4) Å, the Ti–Cl bond lengths to spectroscopy showed that a reaction of two equivalents of
2.353(2) Å and the ipso-C resonance is at δ = 166.0 ppm.[7] LiOC6H3(CMe3)2-2,4 with TiCl4 in benzene also gave the
The standard deviations for the C–O bond lengths do not tris(phenoxo) complex as the major product. The reason for
allow reliable bond length comparisons to be made for the ease of formation of this tris-product in these reactions
these complexes.
is unclear but appears to be related to the presence of the
The reaction of 2,6-diisopropylphenol with TiCl4 gave an para-tert-butyl substituent. NMR spectral analysis of the
oil which failed to form a solid[9,18,19] but was identified as reaction of two equivalents of 2-tert-butyl-4-methylphenol
[TiCl2(OC6H3iPr2-2,6)2] (5) on the basis of its NMR spectra with TiCl4 shows the product is exclusively the bis-
(ipso-C resonance δ = 165.7, cf. δ = 170.5 ppm for [TiCl3(O- (phenoxo) complex (ipso-C resonance at δ = 165.8 ppm).
C6H3iPr2-2,6)],[15] Table 2) being identical to an authentic
Reaction of 2-tert-butyl-6-methylphenol and TiCl4 gave
sample.[19] Reaction of complex 5 with dmbipy gave [TiCl2(OC6H3CMe3-2-Me-6)2] (11) (ipso-C resonance at δ =
[TiCl2(OC6H3iPr2-2,6)2(dmbipy)] (6) for which the NMR 169.3 ppm, cf. δ = 172.8 ppm for [TiCl3(OC6H3CMe3-2-
spectra showed one isomer with a similar structure to 2 Me-6)],[15] Table 2) which is monomeric in benzene. The
[structure I, Scheme 2] and the ipso-C resonance at δ = complex is reported to be a liquid when prepared by ther-
162.6 ppm. Refluxing a solution of 2,6-di-tert-butyl-4-meth- molysis in CH2Cl2[18] and is formed in only 50% yield when
ylphenol with TiCl4 in toluene for 18 h, at which time HCl [Ti(NEt2)2(OC6H3CMe3-2-Me-6)2] is reacted with SiCl4.[4b]
production had essentially ceased, gave a gum which Using the present thermolysis method, production of HCl
NMR spectroscopy showed to be
a
mixture of appeared to have ceased after approximately 8 h of vigorous
[TiCl2(OC6H2- refluxing, but NMR spectroscopy indicated that the reac-
[TiCl3(OC6H2{CMe3}2-2,6-Me-4)],[15]
{CMe3}2-2,6-Me-4)2][8] (7) and unreacted phenol. This indi- tion was only about 80% complete. Starting the cooled
cated that the thermalisation reaction is not successful with solution refluxing again led to resumption of HCl gas pro-
this sterically demanding phenol.[8]
A
sample of duction and further NMR spectral analysis at intervals
[TiCl2(OC6H2{CMe3}2-2,6-Me-4)2] prepared by the reac- showed that another 10 h of reaction time was needed for
tion of TiCl4 and LiOC6H2(CMe3)2-2,6-Me-4[8] did not re- completion. On removing the solvent an oil remained but
act with dmbipy and is reported not to react with pyri- this solidified on standing. The reaction time can be com-
dine.[8] These results indicate that 2,6-di-tert-butyl substitu- pared with that when using 2,6-di-tert-butyl-4-methylphe-
1346
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 1343–1352

Dichlorobis(phenoxo) Titanium(iv) Complexes
FULL PAPER
nol where HCl gas production appeared to cease after
about 18 h but NMR spectral analysis showed the presence
of
[TiCl3(OC6H2{CMe3}2-2,6-Me-4)],
[TiCl2(OC6H2-
{CMe3}2-2,6-Me-4)2] (7) and unreacted phenol. This result
emphasises the need to correctly identify the end-point of
these reactions, especially in cases where the size of both
ortho-substituents becomes larger. The X-ray crystal struc-
ture of 11 shows a distorted tetrahedral structure in the
solid state with bond lengths and angles that are not signifi-
cantly different from those found for [TiCl2(OC6H3Me2-
2,6)2] except for the O–Ti–O bond angle which widens by
ca. 3.5° in 11.[4a] Complex 11 reacts readily with dmbipy to
form [TiCl2(OC6H3CMe3-2-Me-6)2(dmbipy)] (12) for which
the NMR spectra indicate a similar structure to complex
(2). The ipso-C resonance occurs at δ = 166.4 ppm.
[TiCl2(OC2H4Ph-2)2] (13) and the 1-napthoxo complex
[TiCl2(OC10H9)2] (14) were also prepared and found to be
monomeric in benzene.
Theoretical Studies
Density functional theory (DFT) calculations were car-
ried out on model complexes to study the electronic proper-
ties of the phenoxo and chloro ligand interactions with the
titanium centre in the monomeric and dimeric complexes
shown by the solution and solid-state structures. Unsubsti-
tuted phenoxo ligands were used in the models to simplify
the calculations. The optimised geometry for the monomers
Figure 2. Optimized B3LYP structures of [TiCl2(OC6H5)2] (15) and
[TiCl2(OCH3)2] (16).
[TiCl2(OC6H5)2] (15) and [TiCl2(OCH3)2] (16) (used for those found in the X-ray structures [Ti–O bond lengths
comparisons) are shown in Figure 2. Bond lengths and 1.734(7) and 1.753(4) Å, respectively]. The model struc-
angles for these two structures are given in Table 3 where tures, however, indicate that the Ti–O bond lengths are
they are compared with data obtained from the X-ray struc- longer in phenoxo model 15 (1.753 Å) than in the alkoxo
ture of the monomeric complexes [TiCl2(OC6H3Me2- model 16 (1.740 Å). Theoretical calculations have shown
2,6)2][9] and [TiCl2(OC6H3CMe3-2-Me-6)2].[3]
previously that a short Ti–O bond length of 1.750(2) Å in
As expected, the ligands in 15 are tetrahedrally coordi- [TiCl3(OC6H2Me3-2,4,6)] corresponds to a phenoxo ligand
nated to titanium with the O–Ti–Cl, Cl–Ti–Cl and O–Ti– acting formally as a 2π-donor to Ti[15] but the Ti–O bond-
O bond angles (108.7 , 111.1 , and 111.0°, respectively) be- ing is clearly stronger in 16, where there is no pπ(O)-
ing very close to those found in the two X-ray structures. pπ*(C=C) backbonding with a phenyl ring (the C–O bond
The Ti–O bond lengths in 15 [1.753 Å] are also similar to lengths in 15 and [TiCl2(OMe)2] (16) are 1.384 and 1.436 Å,
Table 3. Comparison of the bond lengths [Å] and angles [°] for the X-ray structure of [TiCl2(OC6H3Me2-2,6)2] and [TiCl2(OC6H3CMe3-
2-Me-6)2] with the calculated structures [TiCl2(OC6H5)2] (15) and [TiCl2(OCH3)2] (16).
[TiCl2(OC6H3Me2-2.6)2]
(X-ray)[9]
[TiCl2(OC6H3CMe3-2-Me-6)2]
(X-ray)[5]
[TiCl2(OC6H5)2] (15)
[TiCl2(OCH3)2] (16)
(calcd.)
(calcd.)
Ti–O(1)
Ti–O(2)
Ti–Cl(1)
Ti–Cl(2)
C–O(1)
C–O(2)
O(1)–Ti–O(2)
O(1)–Ti–Cl(1)
O(1)–Ti–Cl(2)
O(2)–Ti–Cl(1)
O(2)–Ti–Cl(2)
Cl(1)–Ti–Cl(2)
Ti–O(1)–C(11)
Ti–O(2)–C(21)
1.734(7)
1.736(8)
2.192(4)
2.211(4)
1.353(12)
1.365(15)
109.1(4)
109.4(3)
108.1(3)
109.2(3)
110.1(3)
111.0(2)
167.3(7)
168.9(8)
1.753(4)
1.752(5)
2.217(3)
2.207(3)
1.403(7)
1.401(9)
112.6(2)
109.2(2)
107.8(2)
107.0(2)
110.0(2)
110.3(1)
162.2(4)
163.4(4)
1.753
1.753
2.249
2.249
1.384
1.384
111.1
108.7
108.7
108.7
108.7
111.1
175.8
175.8
1.740
1.740
2.257
2.257
1.436
1.436
112.2
108.6
108.2
108.6
108.3
111.1
178.3
178.3
Eur. J. Inorg. Chem. 2005, 1343–1352
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1347

A. J. Nielson, C. Shen, P. Schwerdtfeger, J. M. Waters
FULL PAPER
respectively). A similar situation was found for the models which allows assignment of energetic contributions to indi-
[TiCl3(OC6H2Me3-2,4,6)] and [TiCl3(OMe)].[15]
vidual donor acceptor pairs, shows there is substantial do-
The Ti–Cl bond lengths in 15 [2.249 Å] are somewhat nation from the two oxygen 2p lone pairs on the phenoxo
longer than in [TiCl2(OC6H3Me2-2,6)2] [2.192(4) and ligands to both the phenyl C=C (π*) orbital (E2
=
2.211(4) Å] and [TiCl2(OC6H3CMe3-2-Me-6)2] [2.217(3) 104.6 kJmol–1) and more importantly to the unoccupied ti-
and 2.207(3) Å], but this is to be expected from the level of tanium 3d orbitals (E2 = 376.6 kJmol–1). Both the pπ(O)-
theory applied, and they compare with 2.257 Å in pπ*(C=C) and the pπ(O)-Ti interactions are stronger than
[TiCl2(OMe)2] where the strong Ti–O bonding allows the that in the mono(phenoxo) model [TiCl3(OC6H2Me3-2,4,6)]
Ti–Cl bond to relax its π-bonding component somewhat (E2 = 54.4 kJmol–1 and 230 kJmol–1 per lone pair, respec-
when the electron count is maximised.
tively). Even so these values in 15 are clearly overestimated
The Ti–O–C bond angle in model 15 [175.8°] is close to by the NBO procedure, but nevertheless it clearly points
being linear which is also the case in the calculated structure towards Ti–O multiple bonding. The O2p-phenyl C=C (π*)
of [TiCl3(OC6H2Me3-2,4,6)] [177.7° ][15] but the angles are back-bonding is also sufficiently strong to prevent rotation,
smaller in the X-ray structures [e.g. 167.3(7)° in so that at ambient temperatures the configuration of the
[TiCl2(OC6H3Me2-2,6)2]
and
163.2(4)°
in two phenoxo ligands is essentially locked. A decrease in the
[TiCl2(OC6H3CMe3-2-Me-6)2]. Ti–O–C bond angle linear- Ti–O–C bond angle diminishes the backbonding into the Ti
ity is also favoured by pπ(O)-pπ*(C=C) back bonding with 3d-orbitals. This bonding situation would leave no orbitals
the phenyl ring. However, the Ti–O–C bond angle in the available for π-donation from the two chloro ligands, which
model [TiCl2(OMe)2] (16) in which there is no pπ(O)- could build the electron count formally to 18. However, the
pπ*(C=C) back bonding with a phenyl ring is 178.3°. In perturbation analysis reveals that this picture is too simple
this molecule the bond angles about the titanium atom are as there is some orbital mixing of the Cl(3p) lone-pair on
similar to those in 15, but more importantly the Ti–O bond both chlorines via back donation into unoccupied titanium
lengths are slightly shorter in [TiCl2(OMe)2] (1.740 Å, 3d orbitals. In comparison, for the complex
1.753 Å in 15), where the oxygen atom can bind to the me- [TiCl3(OC6H2Me3-2,4,6)] in which there is only one phe-
tal without the influence of π-bonding to an aromatic ring. noxo ligand and thus a greater need for Clp–Ti π-donation,
An angle scan reveals, however, that for the Ti–O–C bond calculations show that two of the lone pairs donate more
angle in [TiCl3(OC6H2Me3-2,4,6)] the energy difference be- strongly (E2 = 108.8 kcalmol–1.[15]) In [TiCl2(OMe)2] (16)
tween the calculated angle (177.7°) and the experimental the calculations show that the contribution from the chloro
value in the crystal [163.1(1)°] is only 3 kJ·mol–1.[15] The ligands is much less than in 15 as without the pπ(O)-
reduced bond angle in the crystal structure is thus due to pπ*(C=C) interaction the strong Ti–O bonding allows the
intermolecular effects similar to those apparent in the re- Ti–Cl bond to relax its π-bonding component and this is
lated organoimido complexes, where M–N–C bond angles reflected in the longer Ti–Cl bond length in 16.
can be as low as 150° without greatly effecting the electronic
Calculations were also carried out on models of the di-
meric complex [(TiCl2{OC6H4CMe3-4}{μ-OC6H4-CMe3-
nature of the multiple bond.[20]
Interestingly, the phenyl rings in 15 face each other with 4})2]. For each titanium centre a trigonal bipyramidal ar-
the faces of the phenyl rings inclined to each other at an rangement for the ligands with the bridging oxygen atoms
angle of 92.6°. This fact can be easily explained from elec- in the equatorial position was chosen in agreement with the
tronic arguments. As there is strong π-overlap with both the crystal structure. With the 4-tert-butyl substituent replaced
Ti centre and the phenyl ring carbon atom (cummulene by H for computational convenience, a number of different
type arrangement), one oxygen atom overlaps with the dxz structures according to all possible permutations was ob-
orbital whilst the other overlaps with the dyz orbital re- tained. However, all the possible starting geometries op-
sulting in phenyl rings perpendicular to each other (Fig- timised to only three different structures and these are
ure 3). The pπ(O)–pπ*(C=C) interaction extends the π sys- shown in Figure 4. The global minimum (Figure 4A) is in
tem in the complexes which accounts for the deep red color- almost perfect agreement with the crystal structures
ation observed. In comparison, bis(alkoxo) complexes [(TiCl2{OC6H4CMe3-4}{μ-OC6H4-CMe3-4})2] (complex 1)
which do not contain this feature are usually colourless.[6]
and [(TiCl2{OC6H5}{μ-OC6H5})2][11] (see Table 4). In par-
ticular the terminal Ti–O bond lengths are 1.747, 1.749 and
1.744 Å, respectively, and the Ti–O bridging bond lengths
are 1.908, 1.912 and 1.910 Å, respectively. The C–Ot–Ti
bond angles are also in reasonable agreement with experi-
mental data (169.9, 165.2 and 165.9°, respectively) as are
the C–Ob–Ti bond angles (126.6, 123.5 and 128.3°, respec-
tively). It is clear that solid state effects and limitations in
the computational procedure accounts for the small differ-
ences in the geometries. The other two structures (Figure 4
B and C) are 13.2 and 23.3 kJ·mol–1 above the global mini-
Figure 3. Opz–Tidxz and Opx–Cpx orbital overlaps in the Ti–O–Ar
system.
A natural bond orbital (NBO) analysis[21] of 15 using mum (Figure 4, A). The dimerization energy for the reac-
second order perturbation analysis of the Fock matrix, tion
1348
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 1343–1352

Dichlorobis(phenoxo) Titanium(iv) Complexes
FULL PAPER
Figure 4. Optimized B3LYP structures of the dimeric titanium compound, [TiCl2(OC6H5)2]2. (A) global minimum; (B); [ΔE =
13.2 kJ·mol–1 compared to (A)]; (C) [ΔE = 23.3 kJ·mol–1 compared to (A)].
Table 4. Comparison of the bond lengths [Å] and angles [°] for the calculated structure [(TiCl2{OC6H5}{μ-OC6H5})2] with the X-ray
structures [(TiCl2{OC6H5}{μ-OC6H5})2][11]and [(TiCl2{OC6H4CMe3-4}{μ-OC6H4CMe3-4})2] (1).
Bond/angle[a]
[(TiCl2{OC6H5}{μ-OC6H5})2]
[(TiCl2{OC6H5}{μ-OC6H5})2]
[(TiCl2{OC6H4CMe3-4}{μ-OC6H4CMe3-4})2] (1)
(calcd.)
(X-ray[11]
)
(X-ray)
Ti–Ob
Ti–Ot
Ti–Ob
Ti–Cl
Ti–Cl
C–Ot
C–Ob
C–Ot–Ti
C–Ob–Ti
2.151
1.747
1.908
2.270
2.276
1.382
1.423
169.9
126.6
2.122(9)
2.142(2)
1.749(2)
1.912(2)
2.219(1)
2.226(1)
1.369(3)
1.406(3)
165.2(2)
123.5(2)
1.744(10)
1.910(9)
2.209(6)
2.219(6)
1.359(19)
1.422(14)
165.9(6)
128.3(4)
[a] b = bridging. t = terminal.
there are no substituents in both ortho-positions of the phe-
noxo ligand phenyl rings, a dimeric phenoxo-bridged struc-
ture is possible in the solid state. Coordination expansion
occurs where a ligand such as dmbipy is used but this ceases
when both ortho-positions of the phenoxo ligand phenyl
ring contain tert-butyl substituents. This compares with the
dicyclopentadienyl complex [Cp2TiCl2] which is not known
to undergo coordination expansion easily and can be pre-
pared in the presence of coordinating solvents.[22] The
chloro ligands in the six-coordinate bis(phenoxo) adducts
are expected to lie trans to each other which is not the ori-
entation required for many catalytic reactions. It is thus im-
portant to keep potentially coordinating ligands away from
the complexes if a cis-dichloro geometry is to be main-
tained. The ipso-C resonance position in the 13C NMR
spectra of the phenoxo ligand phenyl rings indicate that
there is a trend towards less O(2p)-C=C(π*) bonding com-
ponent through the series [TiCl3(OAr)], [TiCl2(OAr)2] and
[TiCl2(OAr)2(dmbipy)]) and this would appear to be a sen-
sitive probe for determining Oπ-orbital-Tid orbital bonding
tightness. The comparable feature in cyclopentadienyl com-
plexes is ring slippage which is uncommon and is only
found in circumstances where a strong π-donating ligand is
2 [TiCl2(OC6H5)2] Ǟ [(TiCl2{OC6H5}{μ-OC6H5})2]
[i.e. for dimer (A)] is only –26.2 kJmol–1. This supports the
results obtained here indicating complex (1) is monomeric
in solution but a dimer in the solid state and also gives
insight as to why molecular weight studies could indicate
either monomeric or dimeric structures in solution.[12,13] In
the gas phase only the monomeric compound was found
from mass spectroscopic studies.[13]
Conclusions
The results of this work show that a range of dichloro-
bis(phenoxo)titanium(iv) complexes can be prepared in es-
sentially pure form and in quantitative yield if the end-point
of thermalising TiCl4 and two equivalents of a phenol in
toluene is determined. This represents a major advance in
the ease of preparing these complexes when larger quanti-
ties are required. In a non coordinating solvent such as ben-
zene the complexes are monomeric and thus have a four-
coordinate geometry typical of [Cp2TiCl2] complexes. When
Eur. J. Inorg. Chem. 2005, 1343–1352
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1349

A. J. Nielson, C. Shen, P. Schwerdtfeger, J. M. Waters
FULL PAPER
CH2Cl2 (30 cm3) and the mixture was stirred for 2 h. The solution
was filtered, the solvent removed and the residue allowed to stand
under petroleum ether (20 cm3) overnight giving the complex as a
noncrystalline orange solid. Yield: 0.75 g, 97%. C32H38Cl2N2O2Ti:
able to compete for available metal orbitals.[23] DFT calcu-
lations on the model [TiCl2(OC6H5)2] show that the struc-
tural parameters compare well with the X-ray structures of
[TiCl2(OC6H3Me2-2,6)2] and [TiCl2(OC6H3CMe3-2-Me-6)2]
and as already discussed, an NBO analysis indicates sub-
stantial donation from the two oxygen 2p lone pairs on the
phenoxo ligands to the unoccupied titanium 3d orbitals.
This is consistent with each ligand acting formally as a 5-
electron donor and gives credence to the analogy with the
1
calcd. C 63.6, H 6.3, N 4.6; found C 63.5, H 6.2, N 4.9. H NMR:
δ = 1.30 (s, 18 H, CMe3), 2.39 (s, 6 H, Me-dmbipy), 7.23 and 7.27
[2d, (AB quartet),3JH,H = 8.8 Hz, 8 H, o,-m-H], 7.27 (d, obsc., H2-
dmbipy), 7.93 (br. s, 2 H, H4-dmbipy), 8.96 (d,3JH,H 5.6 Hz, 2 H,
H1-dmbipy) ppm. 13C NMR: δ = 21.5 (Me-dmbipy), 31.4 (CMe3),
34.3 (C), 118.5 (o-C), 122.5 (C2-dmbipy), 125.8 (m-C), 126.9 (C4-
5-electron donor Cp ligand.[1] There is also sufficient dmbipy), 145.5 (p-C), 148.7 (C1-dmbipy), 150.4 (C3 or C5-dmbipy),
152.4 (C5 or C3-dmbipy), 164.8 (ipso-C) ppm.
O(2p)–C=C(π*) interaction to prevent rotation of the
phenyl rings about the C–O bond at room temperature. The
[TiCl2(OC6H2Me3-2,4,6)2] (3): 2,4,6-Trimethylphenol (4.31 g,
calculations also show that there is only a small dimeris-
ation energy for [TiCl2(OC6H5)2] and the global minimum
structure is the one observed in the solid state. Overall the
bis(phenoxo) complexes show a range of differences com-
31.6 mmol) in toluene (50 cm3) was added to TiCl4 (3.0 g,
15.8 mmol) in toluene (50 cm3) and the mixture was refluxed vigor-
ously until the production of HCl gas ceased (10 h). The solution
was cooled, filtered and the solvent removed to give the complex as
pared with [Cp2TiCl2] and in the absence of cyclometall- a noncrystalline deep red solid. Yield: 5.97 g, 97%. C18H22Cl2O2Ti:
calcd. C 55.6, H 6.1, Cl 18.2; found C 55.6, H 6.1, Cl 18.8. Cryo-
scopic mol. mass: found 395.4; required: 389.1. 1H NMR: δ = 2.22
(s, 6 H, p-Me), 2.29 (s, 12 H, o-Me), 6.75 (s, 4 H, m-H) ppm. 13C
NMR: δ = 16.8 (o-Me), 20.8 (p-Me), 127.3 (o-C), 128.6 (m-C),
134.1 (p-C),166.4 (ipso-C) ppm.
ation involving the ortho-substituents on the aromatic
rings[24] these differences may be exploited in tuning stoi-
chiometric and catalytic reactions in organic synthesis. Such
metal-mediated reactions are now emerging.[3]
[TiCl2(OC6H2Me3-2,4,6)2(dmbipy)]
(4):
Dmbipy
(0.33 g,
1.79 mmol) in CH2Cl2 (30 cm3) was added to complex 3 (0.69 g,
1.78 mmol) in CH2Cl2 (30 cm3) and the mixture was stirred for 3
h. The solution was filtered, the solvent removed and the residue
allowed to stand under petroleum ether (20 cm3) overnight giving
the complex as a noncrystalline orange solid. Yield: 1.02 g, 100%.
C30H34Cl2N2O2Ti: calcd. C 62.8, H 6.4, N 4.9; found C 62.4, H
6.4, N 4.8. 1H NMR: δ = 2.23 (s, 6 H, p-Me), 2.38 (s, 6 H, Me-
Experimental Section
Syntheses: All preparations and manipulations were carried out un-
der dry, oxygen-free nitrogen using standard bench-top techniques
for air-sensitive substances. TiCl4 and the phenols were used as
received from commercial sources. 4,4Ј-Dimethyl-2,2Ј-bipyridine
(dmbipy) was dried under vacuum before use. Light petroleum (b.p.
40–60 °C) and toluene were distilled from sodium wire and dichlo-
romethane from freshly ground CaH2. 1H and 13C{1H} NMR spec-
tra were recorded at 400 and 100 MHz respectively with a Bruker
AM400 spectrometer. CDCl3 and CD2Cl2 were dried with, and dis-
tilled from, freshly ground CaH2 and [D6]benzene from sodium
wire. Molecular mass values were determined cryoscopically in ben-
zene with a Knauer molecular weight determination apparatus un-
der N2 gas conditions using concentrations in the vicinity of
5.3×10–8 mol·L–1. C, H and N analyses were determined by Dr A.
Cunninghame and associates, University of Otago, New Zealand.
Chlorine was determined by gravimetric measurements. The pro-
duction of HCl gas in the thermalisation reactions was monitored
by passing the exhaust gases from the nitrogen bubbler over
N,N,NЈ,NЈ-teramethylethylenediamine and observing the white
cloud produced.
3
dmbipy), 2.51 (s, 12 H, o-Me), 6.75 (s, 4 H, m-H), 7.22 (d, JH,H
5.5 Hz, 2 H, H2- dmbipy), 7.90 (br. s, 2 H, H4-dmbipy), 8.99 (d,
3JH,H 5.5 Hz, 2 H, H1-dmbipy) ppm. 13C NMR: δ = 18.1 (o-Me),
20.6 (p-Me), 21.4 (Me-dmbipy), 122.6 (C2-dmbipy), 126.5 (C4-
dmbipy), 128.9 (m-C), 129.3 (o-C), 131.6 (p-C), 148.6 (C1-dmbipy),
150.4 (C3 or C5-dmbipy),151.9 (C5 or C3-dmbipy), 164.5 (ipso-C)
ppm.
[TiCl2(OC6H3iPr2-2,6)2] (5): 2,6-Diisopropylphenol (3.76 g,
21.1 mmol) in toluene (50 cm3) was added to TiCl4 (2.0 g,
10.5 mmol) in toluene (50 cm3) and the mixture was refluxed vigor-
ously until the production of HCl gas ceased (13 h). The solution
was cooled, filtered and the solvent removed to give an oil which
failed to solidify over time. Yield: 4.96 g. The product gave an iden-
tical NMR spectra to a literature complex.[19]
[TiCl2(OC6H4CMe3-4)2] (1): 4-tert-Butylphenol (4.75 g, 31 6 mmol)
in toluene (50 cm3) was added to TiCl4 (3.0 g, 15.8 mmol) in tolu-
ene (50 cm3) and the mixture was refluxed vigorously until the pro-
duction of HCl gas ceased (9 h). The solution was cooled, filtered
and the solvent removed to give a deep red microcrystalline solid
which was pumped on for several hours. Yield: 6.57 g, 100%.
C20H26Cl2O2Ti: calcd. C 57.6, H 6.3, Cl 17.0; found C 57.5, H 6.3,
Cl 17.3. Cryoscopic mol. mass: found 420.6; required: 417.2. 1H
NMR: δ = 1.25 (s, 18 H, CMe3), 6.92 (b, 4 H, o-H), 7.18 (d,
3J(H,H) = 7.7 Hz, 4 H, m-H) ppm. 13C NMR: δ = 31.4 (CMe3),
34.4 (C), 118.8 (o-C), 125.8 (m-C), 148.2 (p-C), 166.3 (ipso-C) ppm.
Recrystallisation of a portion of the solid from benzene gave well
formed crystals. [C20H26Cl2O2Ti: calcd. C 57.6, H 6.3; found C
57.5, H 6.5]. One of these crystals was used in the X-ray analysis.
[TiCl2(OC6H3iPr2-2,6)2(dmbipy)] (6): Dmbipy (0.31 g, 1.79 mmol)
in CH2Cl2 (30 cm3) was added to complex 5 (0.8 g, 1.79 mmol) in
CH2Cl2 (30 cm3) and the mixture was stirred for 3 h. The solution
was filtered, the solvent removed and the residue allowed to stand
under diethyl ether (50 cm3) overnight giving the complex as a
noncrystalline
deep
red
solid.
Yield:
1.03 g,
88%.
C36H46Cl2N2O2Ti: calcd. C 65.8, H 7.1, N 4.3; found C 65.9, H
1
3
7.0, N 4.8. H NMR: δ = 1.10 (d, JH,H = 6.7 Hz, 24 H, CMe2),
3
2.43 (s, 6 H, Me-dmbipy), 4.20 (sept, JH,H = 6.7 Hz, 4 H, CH),
7.07 (m, 6 H, m,p-H), 7.25 (d, JH,H = 5.4 Hz, 2 H, H2-dmbipy),
3
7.92 (br. s, 2 H, H4-dmbipy), 8.89 (d, JH,H = 5.4 Hz, 2 H, H1-
3
dmbipy) ppm. 13C NMR: δ = 21.6 (Me-dmbipy), 24.5 (CMe2), 26.1
(CH), 122.7 (C2-dmbipy), 122.9 (p-C), 123.6 (m-C), 126.4 (C4-
dmbipy), 140.2 (o-C), 148.5 (C1-dmbipy), 150.4 (C3 or C5-dmbipy),
151.8 (C5 or C3-dmbipy), 162.6 (ipso-C) ppm. The complex is
slightly soluble in diethyl ether.
[TiCl2(OC6H4CMe3-4)2(dmbipy)] (2): Dmbipy (0.24 g, 1.3 mmol) in
CH2Cl2 (25 cm3) was added to complex 1 (0.54 g, 1.3 mmol) in
1350
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 1343–1352

Dichlorobis(phenoxo) Titanium(iv) Complexes
[TiCl2(OC6H3Me2-2,4)2] (8): 2,4-Dimethylphenol
FULL PAPER
(5.15 g,
was cooled, filtered from a small amount of solid and the solvent
removed to give the complex as an oil which turned into a noncrys-
talline orange-red solid on standing. Yield: 6.80 g, 96%.
C22H30Cl2O2Ti: calcd. C 59.3, H 6.8, Cl 15.9; found C 59.3, H 6.7,
Cl 15.8. Cryoscopic mol. mass: found 454.4; required: 445.3. 1H
42.2 mmol) in toluene (50 cm3) was added to TiCl4 (4.0 g,
21.1 mmol) in toluene (50 cm3) and the mixture was refluxed vigor-
ously until the production of HCl gas ceased (10 h). The solution
was cooled, filtered and the solvent removed to give the complex as
a noncrystalline deep red solid. Yield: 7.28 g, 96%. C16H18Cl2O2Ti:
calcd. C 54.6, H 5.2, Cl 19.0; found C 54.6, H 5.4, Cl 19.4. Cryo-
scopic mol. mass: found 376.5; required: 360.0. 1H NMR: δ = 2.23
3
NMR: δ = 1.52 (s, 18 H, CMe3), 2.29 (s, 6 H, Me), 6.93 (t, JH,H
3
4
= 7.6 Hz, 2 H, p-H), 7.00 (d, JH,H = 7.6, JH,H = 1.2 Hz, 2 H, m-
H), 7.17 (d, JH,H = 7.6, JH,H = 1.2 Hz, 2 H, m-H) ppm. 13C
3
4
3
(s, 6 H, Me), 2.25 (s, 6 H, Me), 6.82 (m, 4 H, m-H), 6.98 (d, JH,H
NMR: δ = 18.5 (Me), 30.3 (CMe3), 35.0 (C), 124.6 (m-C), 124.7
= 7.2 Hz, 2 H o-H) ppm. 13C NMR: δ = 16.5 (Me), 20.8 (Me), (m-C), 129.1 (p-C), 131.1 [o-C(methyl)], 137.7 [o-C(tert-butyl)],
119.3 (o-CH), 126.4 (p-C), 127.0 and 131.0 (m-CH), 134.2 (o-C), 169.3 (ipso-C) ppm.
165.2 (ipso-C) ppm.
[TiCl2(OC6H3CMe3-2-Me-6)2(dmbipy)] (12): Dmbipy (0.32 g,
1.75 mmol) in CH2Cl2 (25 cm3) was added to complex 11 (0.78 g,
1.75 mmol) in CH2Cl2 (30 cm3) and the mixture was stirred for 3
h. The solution was filtered, the solvent removed and the residue
allowed to stand under petroleum ether (20 cm3) overnight giving
the complex as a noncrystalline orange solid. Yield: 0.60 g, 55%.
C34H42Cl2N2O2Ti: calcd. C 64.9, H 6.7, N 4.5; found C 64.5, H
[TiCl2(OC6H3Me-2,4)2(dmbipy)] (9): Dmbipy (0.25 g, 1.36 mmol)
in CH2Cl2 (20 cm3) was added to complex 7 (0.49 g, 1.78 mmol) in
CH2Cl2 (20 cm3) and the mixture was stirred for 2 h. The solution
was filtered, the solvent removed and the residue allowed to stand
under petroleum ether (20 cm3) overnight giving the complex as a
noncrystalline orange solid. Yield: 0.69 g, 93%. C28H30Cl2N2O2Ti:
1
1
calcd. C 61.8, H 5.6, N 5.1; found C 61.4, H 5.9, N 4.9. H NMR:
7.0, N 4.5. H NMR: δ = 1.48 (s, 18 H, CMe3), 2.46 (s, 6 H, Me-
3
δ = 2.27 (s, 6 H, Me), 2.42 (s, 6 H, Me), 2.43 (s, 6 H, Me-dmbipy),
dmbipy), 2.76 (s, 6 H, Me-phenoxo), 6.84 (t, JH,H = 7.4 Hz, 2 H,
3
4
3
6.74 (d, JH,H = 8.1 Hz, JH,H = 1.2, 2 H, m-H), 6.85 (br. s, 2 H,
p-H), 6.99 (bd, 2 H, m-H), 7.17 (bd, 2 H, m-H), 7.22 (d, JH,H
m-H), 7.23 (d, JH,H = 5.5 Hz, 2 H, H2-dmbipy), 7.46 (d, JH,H
=
=
3
3
5.5 Hz, 2 H, H2-dmbipy), 7.93 (br. s, 2 H, H4-dmbipy), 8.84 (d,
3JH,H 5.5 Hz, 2 H, H1-dmbipy) ppm. 13C NMR: δ = 21.5 (Me-
dmbipy), 21.8 (Me), 31.7 (CMe3), 35.6 (C), 122.1(CH), 124.5 (CH),
129.8 (CH), 133.0 [o-C(methyl)], 140.8 [o-C(tert-butyl)], 148.6 (C1-
dmbipy), 150.6 (C3 or C5-dmbipy), 151.5 (C5 or C3-dmbipy), 166.4
(ipso-C) ppm. The complex is slightly soluble in petroleum ether.
8.1 Hz, 2 H, o-H), 7.92 (br. s, 2 H, H4-dmbipy), 8.88 (d, JH,H
3
5.5 Hz, 2 H, H1-dmbipy) ppm. 13C NMR: δ = 17.0 (Me), 20.8 (Me),
21.5 (Me-dmbipy), 119.7 (o-CH), 122.5 (C2-dmbipy), 126.9 (C4-
dmbipy), 127.0 (m-C), 127.2 (p-C), 130.8 (m-C), 131.9 (o-C), 148.5
(C1-dmbipy), 150.4 (C3 or C5-dmbipy),152.2 (C5 or C3-dmbipy),
164.3 (ipso-C) ppm.
[TiCl2(OC6H4Ph-2)2] (13): 2-Phenylphenol (5.38 g, 31.6 mmol) in
CH2Cl2 (60 cm3) was added to TiCl4 (3.0 g, 15.8 mmol) in CH2Cl2
(40 cm3) and the mixture was refluxed vigorously until the pro-
duction of HCl gas ceased (12 h). The solution was cooled, filtered
and the solvent removed to give the complex as an oil which turned
into a noncrystalline deep red solid on standing for several days.
Yield: 7.21 g, 100%. C24H18Cl2O2Ti: calcd. C 63.1, H 4.0, Cl 15.5;
found C 63.4, H 4.1, Cl 15.5. Cryoscopic mol. mass: found 471.3;
[TiCl2(OC6H3{CMe3}2-2,4)2] (10) and [TiCl(OC6H3{CMe3}2-2,4)3].
Procedure A: 2,4-Di-tert-butylphenol (6.53 g, 31.6 mmol) in toluene
(60 cm3) was added to TiCl4 (3.0 g, 15.8 mmol) in toluene (50 cm3)
and the mixture was refluxed vigorously until the production of
HCl gas ceased (12 h). The solution was cooled, filtered and the
solvent removed to give an oil which solidified on standing over-
night. NMR spectroscopy showed the product consisted of com-
plex 10 [1H NMR: δ = 1.33 (s, 18 H, CMe3), 1.52 (s, 18 H, CMe3),
1
required: 457.2. H NMR: δ = 6.91–7.49 (m, 18 H, aromatic-H’s)
4
7.14–7.25 (m, 4 H, o,m-H), 7.34 (d, JH,H 2.1 Hz, 2 H, m-H) ppm.
ppm. 13C NMR: δ = 115.9, 120.3, 124.3, 127.5, 128.4, 128.5, 129.0,
129.3, 130.1, 137.0 (o-C), 161.9 (ipso-C) ppm.
13C NMR: δ = 30.3 (CMe3), 31.4 (CMe3), 34.7 (C), 35.1 (C), 123.3
(o-CH), 123.4 (m-CH), 124.2 (m-CH), 135.8 (p-C), 148.0 (o-C),
165.7 (ipso-C) ppm], and [TiCl(OC6H3{CMe3}3-2,4)3], which was
crystallised from the mixture by dissolving the reaction product in
petroleum ether (100 cm3), reducing the volume to ca. 50 cm3 while
keeping the solution hot, then allowing it to stand and cool.
[C42H63ClO3Ti: calcd. C 72.1, H 9.1; found C 72.7, H 9.4].
[TiCl2(OC10H9)2] (14): 1-Napthol (4.56 g, 31.6 mmol) in toluene
(60 cm3) was added to TiCl4 (3.0 g, 15.8 mmol) in toluene (30 cm3)
and the mixture was refluxed vigorously until the production of
HCl gas ceased (12 h). The solution was cooled, filtered and the
solvent removed to give the complex as a noncrystalline solid.
Yield: 6.8 g, 100%. C20H14Cl2O2Ti: calcd. C 61.3, H 3.9, Cl 16.4;
found C 61.4, H 4.1, Cl 15.9. Cryoscopic mol. mass: found 426.9;
required: 405.0. The 1H NMR spectrum shows the product con-
tains a small amount of toluene which is difficult to remove by
pumping.
Procedure B: A solution of n-butyllithium (13.2 cm3, 1.6 mol·L–1)
in hexane was added dropwise to a solution of 2,4-di-tert-butylphe-
nol (4.35 g, 21.1 mmol) in benzene (50 cm3) chilled with ice-water.
The cooling bath was removed and the mixture stirred for 1 h. The
solution and precipitated solid was added slowly to TiCl4 (2.0 g,
10.5 mmol) in benzene (50 cm3) via a cannula and the mixture
stirred overnight. The mixture was allowed to stand for 24 hours
to settle the precipitated LiCl, the solution was filtered and the
solvent removed to give a red solid which was shown to be a mix-
ture of complex 10 and [TiCl(OC6H3{CMe3}3-2,4)3] by NMR spec-
troscopy. Crystallisation of the mixture as for procedure A, gave
[TiCl(OC6H3{CMe3}3-2,4)3] [C42H63ClO3Ti: calcd. C 72.1, H 9.1;
found C 71.1, H 9.9] which had identical NMR spectra with the
sample prepared under procedure A.
Theoretical: Density functional (DFT) calculations were carried
out on the model compound [TiCl2(OC6H5)2] and its possible di-
meric structures using a hybrid Becke three parameter function to-
gether with the Lee–Yang–Parr correlation function (B3LYP).[25,26]
For H, C and O a Dunning/Huzinaga valence double-zeta set was
used.[27] For Ti and Cl we applied the pseudopotential approxi-
mation using the Hay–Wadt parametrization together with valence
double-zeta basis.[28] This resulted in 934 basis functions contracted
to 368 for the dimeric species. The geometries were fully optimized
until the gradients were below 10–5 a.u. which took several months
of supercomputer time on a 16-processor SGI. The structures are
shown in Figure 2 and Figure 4. All calculations were performed
with a parallel version of Gaussian98[29] The bonding situation has
[TiCl2(OC6H3CMe3-2-Me-6)2] (11): 2-tert-Butyl-6-methylphenol
(5.2 g, 31.7 mmol) in toluene (40 cm3) was added to TiCl4 (3.0 g,
15.8 mmol) in toluene (30 cm3) and the mixture was refluxed vigor-
ously until the production of HCl gas ceased (18 h). The solution
Eur. J. Inorg. Chem. 2005, 1343–1352
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1351

A. J. Nielson, C. Shen, P. Schwerdtfeger, J. M. Waters
FULL PAPER
24, 995; c) R. A. Jones, J. G. Hefner, T. C. Wright, Polyhedron
1984, 3, 1121.
[11] K. Waterpaugh, C. N. Caughlan, Inorg. Chem. 1966, 5, 1782.
[12] M. J. Frazer, Z. Goffer, J. Inorg. Nucl. Chem. 1966, 28, 2410.
[13] A. Flamini, D. Cole-Hamilton, G. Wilkinson, J. Chem. Soc.
Dalton Trans. 1978, 454.
[14] A. Shah, A. Singh, R. C. Mehrotra, Polyhedron 1986, 5, 1285.
[15] A. J. Nielson, P. Schwerdtfeger, J. M. Waters, J. Chem. Soc.
Dalton Trans. 2000, 529.
[16] British Titan Products Co. Ltd., French Patent Number:
1377130, 1964.
[17] a) H. Yasuda, Y. Nakayama, K. Takei, A. Nakamura, Y. Kai,
N. Kanehisa, J. Organomet. Chem. 1994, 473, 105; b) N.
Kanehisa, Y. Kai, N. Kasai, H. Yasuda, Y. Nakayama, K.
Takei, A. Nakamura, Chem. Lett. 1990, 2167.
[18] M. Ouchi, M. Kamigaito, M. Sawamoto, Macromolecules
1999, 32, 6407.
[19] H. Noeth, M. Schmidt, Organometallics 1995, 14, 4601.
[20] a) A. Bel, W. Clegg, P. W. Dyer, M. R. Elsegood, V. C. Gibson,
E. L. Marshall,. Chem. Soc. Chem. Commun. 1994, 224; b) P.
Barrie, T. A. Goffey, G. D. Forster, G. Hogarth, J. Chem. Soc.
Dalton Trans. 1999, 4519.
[21] NBO version 3.1, E. D. Glendening, A. E. Reed, J. E. Carpen-
ter, F. Weinhold.
been examined using the natural bond orbital (NBO) partitioning
of Weinhold and Reed.[21]
X-ray Crystallography: Data were collected with a Siemens Bruker
CCD instrument with graphite-monochromated Mo-Kα radiation
(λ = 0.71073 Å) at T = 203(2) K. Crystal decay was monitored by
repeating the initial frames at the end of the data collection and
analysing the duplicate reflections. This was negligible. Unit cell
parameters were obtained by a least-squares fit of all data with I
Ͼ 10 σ(I). Data were corrected for Lorentz, polarisation and ab-
sorption effects.[30] The structures were solved by direct methods
and refined on all F2 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, 0.94 Å
for phenyl ring H and 0.97 Å for tert-butyl H) and refined with a
riding model with Uiso = 0.08. Programs used were SHELXS[31] for
structure solution and SHELXL[32] for refinement. Diagrams were
prepared with ORTEP-3 for windows®.[33]
Crystal Data: [C20H26Cl2O2Ti], M = 417.21, monoclinic, space
group P 21/c, a = 11.265(2), b = 18.603(3), c = 11.491(2) Å, β =
116.59(3)°, V = 2153.4(6) Å3, Z = 4, Dcalcd. = 1.287 mg m3, μ (Mo-
Kα) = 0.655 mm–1, crystal size 0.23×0.22×0.14 mm, ω scans, θmax.
= 28.19°, F(000) = 872, 12833 reflections measured, 4783 unique
(Rint. = 0.0381), Imin./Imax. = 0.8639/0.9139, R1 = 0.0547 [I Ͼ 2σ(I)],
wR2 = 0.1210 (all data).
[22] N. J. Long, Metallocenes – An Introduction to Sandwich com-
plexes, Blackwell Science, Oxford, 1998.
[23] J. C. Green, M. L. H. Green, J. T. Thomas, P. C. Konidaris,
G. H. Maunder, P. Mountford, J. Chem. Soc. Chem. Commun.
1992, 1361.
[24] I. P. Rothwell, Acc. Chem. Res. 1988, 21, 153.
[25] A. D. Becke, Phys. Rev. A. 1988, 38, 3098.
CCDC-249718 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
[26] C. Lee, W.Yang, R. G. Parr, Phys. Rev. B. 1988, 37, 785.
[27] T. H. Dunning, P. J. Hay, in: Modern Theoretical Chemistry
(Ed.: H. F.Schaefer III), Plenum, New York, 1976, p. 1–28.
[28] W. R. Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 284.
[29] Gaussian 98 (Revision A.1), M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G.
Zakrzewski, J. A. Montgomery, R. E. Stratman, J. C. Burant,
S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C.
Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi,
B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski,
G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Ma-
lick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cios-
lowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A. Liashenko, P. Pis-
korz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T.
Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gon-
zalez, M. Challacombe, P. M. W. Gill, B. G. Johnson, W. Chen,
M. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle,
J. A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.
[1] V. C. Gibson, Angew. Chem. Int. Ed. Engl. 1994, 33, 1565.
[2] M. G. Thorn, J. E. Hill, S. A. Waratuke, E. S. Johnson. P. E.
Fanwick, I. P. Rothwell, J. Am. Chem. Soc. 1997, 119, 8630.
[3] L. G. Quan, J. K. Cha, Tetrahedron Lett. 2001, 42, 8567.
[4] a) B. P. Santora, P. S. White, M. R. Gagne, Organometallics
1999, 18, 2557; b) B. P. Santora, A. O. Larsen, M. R. Gagne,
Organometallics 1998, 17, 3138; c) O. V. Ozerov, F. T. Ladipo,
B. O. Patrick, J. Am. Chem. Soc. 1999, 121, 7941.
[5] S. Fokken, T. P. Spanniol, J. Okuda, F. G. Sernetz, R. Mul-
haupt, Organometallics 1997, 16, 4240.
[6] D. C. Bradley, R. C. Mehotra, Metal Alkoxides, Academic
Press, New York, 1978.
[7] A. O. Larsen, P. S. White, M. R. Gagne, Inorg. Chem. 1999, 38,
4824.
[8] A. W. Duff, R. A. Kamarudin, M. F. Lappert, R. J. Norton, J. [30] R. H. Blessing, Acta Crystallogr. Sect. A 1995, 51, 33.
Chem. Soc. Dalton Trans. 1986, 489.
[9] S. A. Waratuke, M. G. Thorn, P. L. Fanwick, A. P. Rothwell,
I. P. Rothwell, J. Am. Chem. Soc. 1999, 121, 9111.
[10] a) J. Okuda, S. Fokken, H.-C.Kang, W. Massa, Chem. Ber.
1995, 128, 221; b) S. L. Latesky, J. Keddington, A. K.
McMullen, I. P. Rothwell, J. C. Huffman, Inorg. Chem. 1985,
[31] G. M. Sheldrick, Acta Crystallogr. Sect. A 1990, 46, 467.
[32] G. M. Sheldrick, Program for the refinement of Crystal Struc-
tures, University of Göttingen, Germany, 1997.
[33] L. J. Farrugia, ORTEP 3 for Windows, Version 1.06, University
of Glasgow, Scotland, U.K.
Received: September 16, 2004
1352
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 1343–1352