Rational synthesis of a family of neutral monomeric heteroleptic titanium complexes based on an octahedral TiO4N2 motif

Article abstract of DOI:10.1002/ejic.201200560

A rational approach leading to a novel family of monomeric heteroleptic titanium complexes formed with two biphenolate derivatives and six different bidentate nitrogenated chelates is proposed. Only one case of loss of structural predictability was observed in the case of a N,O-coordinating ligand and led to a structure containing two (μ-O)-bridged titanium centres. Diffusion-ordered NMR spectroscopy (DOSY) measurements have been systematically used to prove that the complexes characterised in the solid state remain identical in the liquid phase. All ligands have also been classified quantitatively according to their electron-withdrawal capabilities. A quantitative thermodynamic analysis using the PACHA software showed that the entropically favoured displacement of two monodentate-coordinated alcohol molecules needs to be assisted by solvation of the η²-chelate product to overcome the enthalpically unfavourable replacement of two highly polar Ti-O bonds by less polar Ti-N bonds. The occurrence of dichloromethane molecules in the crystalline state and the existence of more than one crystallographically nonequivalent complex within the unit cell has been associated with a low dipolar moment of the complexes. Other complexes with large dipolar moments (μ > 2 D) seem to have enough dipolar energy in the solid state to remove all the solvating dichloromethane molecules.

Full text of DOI:10.1002/ejic.201200560

FULL PAPER
DOI: 10.1002/ejic.201200560
Rational Synthesis of a Family of Neutral Monomeric Heteroleptic Titanium
Complexes Based on an Octahedral TiO4N2 Motif
David Michael Weekes,[a] Nathalie Baradel,[a] Nathalie Kyritsakas,[b] Pierre Mobian,*[a] and
Marc Henry*[a]
Keywords: Titanium / N ligands / Bond energy / Thermodynamics
A rational approach leading to a novel family of monomeric
heteroleptic titanium complexes formed with two biphenol-
ate derivatives and six different bidentate nitrogenated che-
lates is proposed. Only one case of loss of structural predict-
ability was observed in the case of a N,O-coordinating ligand
and led to a structure containing two (μ-O)-bridged titanium
centres. Diffusion-ordered NMR spectroscopy (DOSY) mea-
surements have been systematically used to prove that the
complexes characterised in the solid state remain identical in
the liquid phase. All ligands have also been classified quanti-
tatively according to their electron-withdrawal capabilities.
A quantitative thermodynamic analysis using the PACHA
software showed that the entropically favoured displacement
of two monodentate-coordinated alcohol molecules needs to
be assisted by solvation of the η2-chelate product to over-
come the enthalpically unfavourable replacement of two
highly polar Ti–O bonds by less polar Ti–N bonds. The occur-
rence of dichloromethane molecules in the crystalline state
and the existence of more than one crystallographically non-
equivalent complex within the unit cell has been associated
with a low dipolar moment of the complexes. Other com-
plexes with large dipolar moments (μ Ͼ 2 D) seem to have
enough dipolar energy in the solid state to remove all the
solvating dichloromethane molecules.
ular complex need the incorporation of solvent molecules,
whereas another one will lead to a crystal free of solvent
molecules? Why are some unit cells built from several crys-
tallographically nonequivalent copies of the same molecular
entity? Why are some structures found to be interweaved
whereas other are not? To provide answers, theoretical mod-
elling is mandatory to allow the observed chemical reac-
tions to be discussed by using reliable thermodynamic data.
This overlap between experiment and theory is rarely per-
formed in supramolecular chemistry, although it is very
common in physics. The idea here is to go beyond cor-
puscular quantum theory concepts that require the
Schrödinger equation to be solved and instead to consider
quantum field theory (QFT) concepts in which any kind of
interaction may be described as an exchange of virtual bos-
ons between fermions. In the case of the electromagnetic
interactions that form the basis of all chemical phenomena,
it is the exchange of virtual photons between protons of the
nucleus and electrons of the atom that should be modelled.
Within a QFT framework, a short bond should be strong
whereas a long bond should be weak. This point is easy
to understand by referring to the fundamental uncertainty
relationship of quantum physics involving energy and time.
Basically, for chemical bonding to occur, a virtual photon
should be emitted from a given fermion (proton or electron)
and travel along a certain distance Δx (bond length) to be
absorbed by another fermion after a time Δt = Δx/c, in
which c is the speed of light. If R stands for the reduced
Planck constant, the energy ΔE of this process should then
Introduction
Titanium complexes that incorporate a TiO4N2 frame-
work are an important class of compounds that are inten-
sively studied in the fields of asymmetric synthesis,[1] poly-
merisation,[2] cytotoxic activities,[3] materials science[4] and
supramolecular chemistry.[5] This large family of structur-
ally diverse compounds is also particularly attractive in the
context of basic coordination chemistry,[6] since in many
cases the rational synthesis of such complexes is particularly
difficult to perform. Very often, the behaviour in solution
of oxygenated ligand-based TiIV complexes is highly com-
plex due to their strong tendencies for oligomerisation by
means of the formation of oxo and/or alkoxy bridges be-
tween Ti atoms.[7] Thus, it is still an interesting task to de-
velop systematic approaches to synthesise monomeric TiIV
complexes incorporating a TiO4N2 framework. Moreover,
these complexes are simple enough to bring theoretical an-
swers to fundamental questions of coordination supra-
molecular chemistry that are often overlooked in the litera-
ture. For instance, why does crystallisation of a given molec-
[a] Laboratoire de Chimie Moléculaire de l’Etat Solide, UMR
7140, Université de Strasbourg,
67070 Strasbourg, France
E-mail: mobian@unistra.fr; henry@unistra.fr
Homepage: http://www-chimie.u-strasbg.fr/~lcmes/labo/
[b] Laboratoire de Chimie de Coordination Organique, UMR
7140, Université de Strasbourg,
67070 Strasbourg, France
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201200560.
Eur. J. Inorg. Chem. 2012, 5701–5713
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5701

D. M. Weekes, N. Baradel, N. Kyritsakas, P. Mobian, M. Henry
FULL PAPER
obey the inequality ΔE·Δt Յ R, as the virtual photon should ity of various nitrogenated bidentate ligands (Scheme 1, b)
not be seen (otherwise it becomes a real photon that will to be incorporated into the initial complex through the de-
be detected). The maximum energy that could be exchanged parture of two alcohol molecules has been screened. Ac-
between both fermions is thus ΔE·Δt ≈ R, that is, ΔE ≈ Rc/ cordingly, the cis-[Ti(dpbpol)2(HOiPr)2] complex presented
Δx. This clearly shows that at the QFT level there is no in Figure 1 (a) is known to retain two labile 2-propanol li-
fundamental difference between a covalent interaction gands in solution. Furthermore, the structure provided evi-
(short bond length Δx) and a van der Waals one (large dence that these two monodentate ligands are located in a
bond length Δx). The only difference lies in the fact that pocket created by two neighbouring phenyl rings belonging
the virtual photon should exist for a longer time in the case to the dpbpol ligands. Thus, to maximise the chance of dis-
of van der Waals interactions, thereby leading to a weaker placing the two monodentate ligands confined in such a
energy transfer than that in the covalent case. In summary, narrow environment, six markedly different flat aromatic
at the QFT level the observed molecular structure by fixing nitrogenated bidentate ligands have been chosen. 5,6-dmp
the distances that virtual photons should travel becomes and dphphen (Scheme 1, b) are 1,10-phenanthroline deriva-
completely equivalent to a set of interaction energies. This tives. They are thus highly rigid chelates. The dtbipy ligand
one-to-one correspondence is easily unveiled by considering is a flexible compound since the pyridine rings can rotate
that some virtual photons carry attractive interactions in relative to each another about the C2–C2Ј bond. The bpym
the case of a proton–electron couple, whereas theses inter- molecule is also flexible and provides a supplementary de-
actions are repulsive for electron–electron couples (proton–
proton electrostatic repulsion does not need to be consid-
ered here as the size of the nucleus is governed by the strong
interaction). Since electronegativity measures the aptitude
of a nucleus to attract electrons,[8a] this notion may be used
to characterise the ability of a nucleus to exchange attract-
ive virtual photons. Similarly, since atomic size[9] is gov-
erned by electron–electron repulsions, this concept may be
used to characterise the ability of a nucleus to exchange
repulsive virtual photons. As electronegativity is also acting
as a chemical potential for modelling the movements of
electrons,[8b] a full equalisation of all electronegativities be-
tween atoms of fixed size is the simplest way to implement
QFT concepts in chemistry. For a practical viewpoint, all
these ideas have been implemented in the PACHA software,
which is a graphical interface allowing any kind of struc-
tural data to be turned into a partial-charge distribution.[10]
This allows the interactions of electrons on average with
respect to nuclei, internal energies, solution entropies, sol-
vation energies, solid-state packing energies, bond energies
and molecular dipole moments to be readily derived.
To demonstrate the usefulness of this new approach of
thermodynamic modelling, a model system has been se-
lected based on a rational approach to control the coordi-
nation sphere of a titanium centre by means of the complex-
ation of two sterically hindered 2,2-biphenol-based ligands
to the metal, that is, 3,3Ј-diphenyl-2,2Ј-biphenol (named
dpbpolH2, see Scheme 1, a). The resulting titanium com-
plex formulated as cis-[Ti(dpbpol)2(HOiPr)2] (Scheme 1, a)
is a C2-symmetric mononuclear octahedral TiIV bis-biphe-
nolate complex with two remaining cis positions occupied
by two coordinated 2-propanol molecules.[11] These mono-
dentate ligands are labile molecules as attested by variable-
temperature NMR spectroscopy experiments. Thus, this be-
haviour offers a great opportunity to envisage a systematic
Scheme 1. a) Structure of the 3,3Ј-diphenyl-2,2Ј-biphenol
approach to form complexes bearing a TiO4N2 core
through the displacement of these labile alcohol molecules
by nitrogenated ligands. Consequently, we report herein the
synthesis and solid-state characterisation of a novel family
of TiO4N2 complexes, in which the oxygenated ligand sys-
(dpbpolH2) ligand. The crystal structure of cis-[Ti(dpbpol)2-
(HOiPr)2][11] is also represented. b) Various structures of ligands
employed in this study: 5,6-dmp = 5,6-dimethyl-1,10-phenan-
throline, dphphen = 4,7-diphenyl-1,10-phenanthroline, dtbipy =
4,4Ј-di-tert-butyl-2,2Ј-bipyridine, bpym = 2,2Ј-bipyrimidine, opda
= o-phenylene diamine, dafone = 4,5-diazafluoren-9-one, 8-Hq =
tem is two 3,3Ј-diphenyl-2,2Ј-biphenol molecules. The abil- 8-hydroxyquinoline.
5702
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 5701–5713

A Family of Neutral Monomeric Heteroleptic Titanium Complexes
gree of topicity compared to dtbipy. The 4,5-diazafluoren- ity as attested by the systematic solid-state analysis of the
9-one ligand is a less efficient chelate among the diimine resulting complexes. The effect of substituting an N,N- by
ligands proposed since it offers a subtantially larger bite an O,N-chelate was also investigated.
angle.[12] Opda is the only nitrogenated chelate that incor-
porates diamino functionalities. Furthermore, the opda li-
Methodology
gand is typically a noninnocent ligand for which the coordi-
nation chemistry has been intensively studied in the pres-
ence of late-transition metals.[13] Finally, the 8-Hq ligand is
a N,O-chelate.
From a synthetic view point, the starting cis-[Ti(dpbpol)2-
(HOiPr)2] complex was dissolved in dry dichloromethane
(CH2Cl2) with one equivalent of nitrogenated bidentate li-
gand. A counter solvent was added by diffusion. Crystalli-
sation occurred between one and three days, either at room
temperature or at +4 °C. The resulting crystals were iso-
lated from the solution and were analysed directly by X-ray
diffraction. To confirm that the structure characterised in
the solid state was unmodified in solution, diffusion-or-
dered NMR spectroscopy (DOSY) analysis was found to
be an ideal technique for studying this library of complexes.
The analysed solution resulted from the dissolution of crys-
tals in dry CD2Cl2. Thereby, the diffusion coefficients ob-
tained from the DOSY experiments allowed the hydrody-
namic radii of the species in solution to be calculated ac-
cording to the Stokes–Einstein equation. These values were
compared to the hydrodynamic radii evaluated from their
crystalline structure using a standard set of van der Waals
radii.[14] The chemical reaction to be modelled for N,N-che-
lates (N,N) may be written as shown in Equation (1).
cis-[Ti(dpbpol)2(HOiPr)2] + (N,N) Ǟ [Ti(dpbpol)2(N,N)] + 2 HOiPr
(1)
Consequently, all the crystalline structures have been char-
acterised with the PACHA software[10] to retrieve the
thermodynamic data needed to evaluate the free-energy
change associated with Equation (1). By applying the equi-
librium conditions ΔH = TϫΔS at T = 300 K, the minimal
number n of dichloromethane molecules that should be as-
sociated with the TiO4N2 complex to have a spontaneous
reaction should satisfy Equation (2).
Figure 1. a) Structure of [Ti(dpbpol)2(5,6-dmp)] as the (Δ,aS,aS)
enantiomer. Selected bond lengths [Å]: Ti1–N1 2.225, Ti1–N2
2.229, Ti1–O4 1.897, Ti1–O2 1.891, Ti1–O3 1.870, Ti1–O1 1.849.
Selected angles [°]: N1–Ti–N2 72.024, O1–Ti–O3 111.131, O4–Ti–
O2 172.850. b) Structure of [Ti(dpbpol)2(dphphen)] as the (Λ, aR,
aR) enantiomer. Selected bond lengths [Å]: Ti1–N1 2.239, Ti1–N2
2.232, Ti1–O4 1.873, Ti1–O2 1.908, Ti1–O3 1.865, Ti1–O1 1.843.
Selected angles [°]: N1–Ti–N2 72.866, O1–Ti–O3 110.649, O4–Ti–
O2 173.443. In both structures, hydrogen atoms are omitted for
clarity and the phenanthroline ligands are coloured differently
(light grey) from the two dpbpol ligands (black).
–nϫUpol(CH2Cl2) = 4345 + 0.3(S°L – S°C) – Upol(L) + Upol(C)(2)
in which Upol(CH2Cl2) is the average interaction energy (in
kJmol–1) per dichloromethane molecule; see the Supporting
Information and Table 1 for details.
Results
It is worth noting that among the 252 structures of com-
plexes containing a TiO4N2 backbone reported in the 5.32
release of the Cambridge Crystallographic Database, 27
complexes incorporate a bipy ligand and three a phenan-
throline ligand. For the dafone, bpym and opda ligands,
Complexation with Rigid Diimine Ligands: Solid- and
Liquid-State Characterisation of [Ti(dpbpol)2(5,6-dmp)],
[Ti(dpbpol)2(dphphen)] and [Ti(dpbpol)2(dafone)] Complexes
Our study began with the preparation of the [Ti(dpbpol)2-
there is no example reported until now. Fourteen structures (5,6-dmp)] and [Ti(dpbpol)2(dphphen)] complexes. Crystals
in which an 8-hydroxyquinolinato ligand is coordinated to of [Ti(dpbpol)2(5,6-dmp)]·CH2Cl2 and [Ti(dpbpol)2-
a titanium(IV) centre are described. Thus, our model sys- (dphphen)]·0.25CH2Cl2 were obtained from a CH2Cl2/n-
tem was also able to produce completely new TiO4N2 com- pentane solvent/counter-solvent system. The structures of
plexes. It was thus found that these flattened nitrogenated [Ti(dpbpol)2(5,6-dmp)] and [Ti(dpbpol)2(dphphen)] are de-
bidentate ligands produce exclusively neutral monomeric picted in parts a and b of Figure 1. These complexes crys-
complexes with a very high degree of structural predictabil- tallise in achiral space groups, that is, Pccn for [Ti(dpbpol)2-
Eur. J. Inorg. Chem. 2012, 5701–5713
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5703

D. M. Weekes, N. Baradel, N. Kyritsakas, P. Mobian, M. Henry
FULL PAPER
Table 1. Partial-charge distributions on Ti atoms (qTi), 3,3Ј-diphenyl-2,2Ј-biphenol ligands (qdpbpol), labile or bidentate ligand L (qL) and
coordinated solvent molecules (qS, with S = CH2Cl2 for dphphen, 5,6-dmp or opda; S = H2O or iPrOH for 8-q). For the 8-q ligand, the
charge on the oxo bridge was q(μ-oxo) = –0.80 in both cases. From these partial-charge distributions it was possible to estimate the polar
internal energy for ligands, Upol(L), and complexes, Upol(C), given in kJmol–1. Solution entropies in CH2Cl2 computed at T = 300 K are
given in JK–1 mol–1 for ligands (S°L) and complex (S°C). Values for free H2dpbpol molecules were found to be Upol = –251 kJmol–1 with
S° = 327 Jmol–1 K–1. For the water molecule, the assumption of a gas-phase structure characterised by d(O–H) = 0.957 Å and θ(H–O–
H) = 104.52° leads to Upol(H2O) = –179 kJmol–1 with S° = 191 Jmol–1 K–1. BE and PE columns give average Ti–O (dpbpol, 8-q) or Ti–
N (other ligands) bond energies without the covalent contribution and solid-state packing energies, respectively, in kJmol–1. Finally,
molecular dipole moments of all complexes are given in debyes (μ column).
L
Upol(L) S°L
qL
–0.21; –0.25 +2.88
–0.54 +2.77
–0.58; –0.55 +2.78;
+2.80
qTi
qdpbpol
qS
BE
μ
Upol(C) S°C PE
iPrOH
5,6-dmp
–382
–65
258
306
–1.21; –1.20
–1.05; –0.93
–1.13, –1.00; –1.07,
–1.03
–1.03; –1.05
–1.18; –1.15
–1.09, –1.04; –1.04,
–1.03
–
–770
2.07
1.70
0.60;
0.51
2.87
2.96
1.31;
1.11
1.12
0.85
–4810
–4173
353 –8.0
357 –2.4
–0.25 –589
–615;
–624
–544
–615
–602;
–592
–607
dphphen
–59
327
–0.43
–4230
362 –4.2
dafone
dtbipy
–77
–122
300
318
–0.69
–0.44
+2.78
+2.77
–
–
–4226
–4246
366 –5.3
359 –10.1
–0.70; –0.70 +2.80;
+2.79
m-bpym
–76
295
–
–4286
355 –9.1
o-bpym
opda
–76
–169
295
280
–0.68
–0.36
+2.80
+2.77
–1.08; –1.03
–0.93; –0.91
–1.19; –1.06
–
–4268
–4335
354 –10.3
353 –4.7
–0.19 –583
–0.92; –0.83 +2.53;
+2.55
8-q (H2O)
–145
291
–0.28 –609
1.11
–7527
361 –4.7
–0.93; –0.83 +2.56;
+2.54
–1.13; –1.11
8-q (iPrOH) –145
291
–0.30 –842
1.57
–7487
363 –7.0
(5,6-dmp)] and C2/c for [Ti(dpbpol)2(dphphen)]. These shows a good correlation with the expected calculated m/z
structures highlight the chirality of these molecules. For values {calculated for [Ti(dpbpol)2(dphphen) + H]+
=
each crystal system, the unit cells contain two enantiomers. 1053.317 and [Ti(dpbpol)2(5,6-dmp) + H]+ = 929.286}. It
Two sources of chirality could be identified from these is noteworthy that no ions related to species without the
architectures. A Δ or Λ configuration is adopted by these presence of the nitrogenated diimine ligands were detected,
complexes, since the two dpbpol ligands are wrapped which highlights a rather good stability of the complexes in
around the metal centres. The second source of chirality is the gas phase. To check that the structures characterised
an axial chirality that arises from the nonplanarity of the in the solid state are unmodified in solution, crystals were
two phenolate rings belonging to a dpbpol ligand. The two dissolved in dry CD2Cl2 and the resulting solutions were
oxygenated ligands show an identical configuration within analysed by NMR spectroscopy (1H and DOSY experi-
a complex and the axial chirality found for the dpbpol li- ments). The behaviour of these two complexes in solution
gands is directly correlated to the Δ or Λ configuration of is very similar. The two 1H NMR spectra indicate the pres-
the complex. As a consequence, these crystals are formed ence of one major product for both experiments in addition
by two enantiomers adopting a (Δ,aS,aS) and (Λ,aR,aR) to a very small amount of undetermined compounds. Ac-
configuration, respectively. Interestingly, these observations cording to the number of signals observed, the NMR spec-
arising from the solid-state analysis of [Ti(dpbpol)2(5,6- troscopy signatures of the complexes are in full accordance
dmp)] and [Ti(dpbpol)2(dphphen)] are very different in the with the C2 symmetry characterised in the solid state. Dif-
case of cis-[Ti(dpbpol)2(HOiPr)2]. In the crystal of cis- fusion measurements also confirm that the structures that
[Ti(dpbpol)2(HOiPr)2], two enantiomers (Δ or Λ) were were characterised by X-ray diffraction remain unaltered in
found with the two biphenyl derivatives adopting an oppo- solution. The hydrodynamic radii from the DOSY experi-
site configuration within a complex, thus only the (Δ,aR,aS) ments were calculated to be Rh
= (7.2Ϯ0.7) and
and (Λ,aS,aR) stereochemistry were observed.[8] In terms of (6.5Ϯ0.7) Å for [Ti(dpbpol)2(dphphen)] and [Ti(dpbpol)2-
the structural description of these two complexes, no (5,6-dmp)], respectively. These values are in good agreement
marked difference was noticed. As expected for these struc- with those evaluated from X-ray crystallographically deter-
tures, the metallic centres display a distorted octahedral ge- mined structures: Rh = (8.3Ϯ0.4) Å for [Ti(dpbpol)2-
ometry with two coordinated dpbpol ligands and one phen- (dphphen)] and Rh = (7.3Ϯ0.4) Å for [Ti(dpbpol)2(5,6-
anthroline ligand. The diimine ligands are positioned in a dmp)].
parallel manner in between two phenyl rings, which high-
Another noticeable difference between phenanthroline-
lights the intramolecular face-to-face π interactions. The based structures and 2-propanol-based ones is the occur-
measured closest centroid separations between pyridine rence of CH2Cl2 molecules in the lattice. Table 1 shows a
rings and phenyl rings range from 3.57 to 4.00 Å. The ob- flow of electronic density towards the CH2Cl2 molecule
tained crystals were dissolved in CH2Cl2 and analysed by leading to a less ionic bonding between the dpbpol ligand
mass spectrometry. The highest mass peaks for the analysis and the titanium atom and thus to a lower dipole moment,
of [Ti(dpbpol)2(dphphen)] and [Ti(dpbpol)2(5,6-dmp)] were particularly in the case of the dphphen ligand. Interestingly
found to be m/z = 1053.327 and 929.315, respectively, which enough, the charge on the dmp ligand is roughly the same
5704
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 5701–5713

A Family of Neutral Monomeric Heteroleptic Titanium Complexes
as that found on the two 2-propanol ligands (–0.46) in the important than those determined in the phenanthroline
starting complex. Removing CH2Cl2 molecules allows a complexes, but the angle measured is only different by 3.6°
network solvation energy of Upol(CH2Cl2) = –26.3 kJmol–1 from the average N–Ti–N bite angle evaluated from the
for 5,6-dmp and of –24.3 kJmol–1 for dphphen, typical of [Ti(dpbpol)2(5,6-dmp)] and [Ti(dpbpol)2(dphphen)] struc-
strong dipole–dipole interactions, to be evaluated. The tures. A much more noticeable effect of the dafone ligand
packing energy is typical of van der Waals interactions with concerns the O4–Ti–O2 angle. A difference of 8.4 and 9°
a reduced ionic character of Ti–N bonds relative to Ti–O between the O4–Ti–O2 angle measured in [Ti(dpbpol)2-
ones. The number of solvent molecules per complex was (dafone)] and those measured in [Ti(dpbpol)2(5,6-dmp)]
evaluated as n = 8 in the case of the 5,6-dmp ligand and n and [Ti(dpbpol)2(dphphen)], respectively, was found. ESI-
= 7 for the dphphen one. As the complex appears to be MS analysis showed a good correlation between the experi-
solvated by four CH2Cl2 molecules in the 5,6-dmp crystal, mental peak found at m/z = 903.26 and a corresponding
this means that crystal growth implies the elimination of simulated peak for TiO5N2C59H38. As expected, the IR
about four solvent molecules per complex. However, the spectra showed a strong sharp band at 1731 cm–1 that can
dphphen complex appears to be solvated by two CH2Cl2 be attributed to the ketone C=O stretch. The DOSY mea-
molecules in the crystal, meaning that crystal growth im- surement found a diffusion at D = (830Ϯ83) μm2 s–1, which
plies the elimination of about five solvent molecules per corresponds to
a
hydrodynamic radius of Rh
=
complex. It is worth noting that the crystal structure of the (6.4Ϯ0.6) Å. This value, which is in the same range as the
dphphen complex consists of two similar interwoven net- theoretical one evaluated from the crystal structure [Rh =
works and that the interweaving energy of a Ti1-based net (7.3Ϯ0.4) Å], is indeed very similar to those obtained from
with a Ti2-based net was found to be cooperative at the phenanthroline-based complexes.
–12 kJmol–1.
From a polarity viewpoint, the dafone ligand appears to
Let’s now turn our attention to a much less efficient che- have a greater affinity for electrons than the dphphen li-
late: the dafone ligand. When a dafone ligand is coordi- gand, which leads to quite a large dipole moment. The
nated to a metal centre the nitrogen–metal overlap is less packing energy of the complex is again typical of van der
than that of a phenanthroline moiety owing to the carbonyl Waals interactions. The average Ti–N bond energy was
bridge between the pyridine rings. The same experimental found to be reduced, which reflects the less efficient chela-
procedure as the one used to isolate the complexes bearing tion. As the crystal is free of CH2Cl2 molecules, an average
the phenanthroline ligands was followed for the preparation value was used for Upol(CH2Cl2) = –25 kJmol–1. Conse-
of [Ti(dpbpol)2(dafone)]. The complex crystallises in a tri- quently, the number of solvent molecules per complex that
clinic crystal system with two molecules per unit cell. The should be eliminated during crystal growth was found to be
X-ray crystallographically determined structure of n = 7.
[Ti(dpbpol)2(dafone)] is shown in Figure 2.
Complexation with Flexible Diimine Ligands: Solid- and
Liquid-State Characterisation of [Ti(dpbpol)2(dtbipy)] and
[Ti(dpbpol)2(bpym)] Complexes
The study was continued by testing the dtbipy and bpym
diimine ligands that are much more flexible than the three
previously discussed ones. Crystals of [Ti(dpbpol)2(dtbipy)]
suitable for X-ray diffraction analysis grew within one day
from an equimolar CH2Cl2 mixture of the starting complex
and the bipyridine derivative layered with n-pentane. The
structure of the complex is depicted in Figure 3. The com-
plex crystallises in an achiral space group (P21/c) with four
molecules per unit cell. The crystal packing is generated by
columnar assemblies of homochiral complexes. Each of
these linear assemblies are surrounded by four columns in
Figure 2. Structure of [Ti(dpbpol)2(dafone)] as the (Δ,aS,aS) en-
antiomer. Selected bond lengths [Å]: Ti1–N1 2.296, Ti1–N2 2.232,
Ti1–O4 1.878, Ti1–O2 1.898, Ti1–O3 1.838, Ti1–O1 1.836. Selected which the complexes display an opposite chirality. The Ti–
angles [°]: N1–Ti–N2 76.206, O1–Ti–O3 108.626, O4–Ti–O2
N and Ti–O distances (see the figure caption) are unexcep-
164.443. Hydrogen atoms are omitted and the dafone ligand is col-
tional and the N1–Ti–N2 and O4–Ti1–O2 angles are sim-
oured differently (light grey) from the two dpbpol ligands (black)
ilar to those measured for the phenanthroline complexes.
for clarity.
The steric bulk provided by the tert-butyl groups at the
back of the diimine ligand has no influence on the metrical
In terms of chirality, the (Δ,aS,aS) and (Λ,aR,aR) parameters in comparison with the previously described
enantiomers were found in the crystal as observed for the complexes. However, it is noticed that the two pyridine
above-described phenanthroline complexes. As expected, rings do not present a perfect coplanar arrangement since
the N–Ti–N bite angle measured in the structure is more a torsion angle of 10.7° between these two aromatic rings
Eur. J. Inorg. Chem. 2012, 5701–5713
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5705

D. M. Weekes, N. Baradel, N. Kyritsakas, P. Mobian, M. Henry
FULL PAPER
was measured. The highest mass peak in the electrospray
mass spectrum of the complex appeared at 989.380 and was
assigned to [Ti(dpbpol)2(dtbipy) + H]+. In solution, a single
peak at δ = 1.38 ppm was observed for the resonance of the
tert-butyl protons, which indicates that the C2 symmetry of
the complex, observed in the solid state, is conserved in
solution. Finally, diffusion-ordered spectroscopy proved
that the species in solution is monomeric. The hydrody-
namic radius from the DOSY analysis was calculated to be
Rh = (7.2Ϯ0.7) Å, whereas a value of Rh = (7.8Ϯ0.4) Å
was evaluated from the crystal structure.
Figure 4. Structure of [Ti(dpbpol)2(bpym)] as the (Δ,aS,aS) enanti-
omer. Selected bond lengths [Å]: Ti1–N1 2.226, Ti1–N4 2.255, Ti1–
O4 1.840, Ti1–O2 1.855, Ti1–O3 1.879, Ti1–O1 1.888. Selected
angles [°]: N1–Ti–N4 70.649, O1–Ti–O3 177.646, O4–Ti–O2
113.524. Hydrogen atoms are omitted and the bpym ligand is col-
oured differently (light grey) from the two dpbpol ligands (black)
for clarity.
The structure provides evidence as to why dimerisation
was not observed despite the ditopic nature of the bpym
ligand since the two phenyl rings are located on both sides
of bpym thereby preventing the coordination of a second
Ti(dpbpol)2 moiety. The Ti–O and Ti–N distances are com-
parable to those measured in [Ti(dpbpol)2(dtbipy)] (the dis-
tances and the angles around the titanium centre are given
in the figure caption). The DOSY experiment unambigu-
ously proved the monomeric nature of the complex in a
CD2Cl2 solution since an experimental hydrodynamic ra-
dius Rh = (6.7Ϯ0.6) Å compatible with the hydrodynamic
radius evaluated from the structure Rh = (7.1Ϯ0.4) Å was
determined. An intense mass peak at 879.25 was found in
the electrospray mass spectrum that was attributed to the
monomeric species [Ti(dpbpol)2(bpym) + H]+ (calculated
mass 879.245).
The charge distribution obtained for both polymorphs
shows the quite high electronic affinity of the bpym ligand
associated with moderate dipole moments and a large pack-
ing energy typical of van der Waals interactions. The
average Ti–N bond energy appears to be very similar to
that found for the dtbipy ligands. The interweaving energy
of a Ti1-based net with a Ti2-based net was found to be
cooperative at –13 kJmol–1. As the crystal is free of CH2Cl2
molecules, a value of Upol(CH2Cl2) = –25 kJmol–1 was used
as before, which led to n = 5 for both monoclinic and ortho-
rhombic polymorphs that should all be eliminated during
crystal growth.
Figure 3. Structure of [Ti(dpbpol)2(dtbipy)] as the (Δ,aS,aS) enanti-
omer. Selected bond lengths [Å]: Ti1–N1 2.248, Ti1–N2 2.227, Ti1–
O4 1.886, Ti1–O2 1.914, Ti1–O3 1.859, Ti1–O1 1.846. Selected
angles [°]: N1–Ti–N2 72.015, O1–Ti–O3 110.215, O4–Ti–O2
175.041. Hydrogen atoms are omitted and the dtbipy ligand is col-
oured differently (light grey) from the two dpbpol ligands (black)
for clarity.
The dtbipy ligand appears to have less affinity for elec-
trons than the phenanthroline or dafone ligands. Despite
this, the complex retains quite a large dipole moment and
displays a packing energy about twice that of the dafone
complex. The average Ti–N bond energy was found to be
very similar to that evidenced with the dphphen ligand.
With the crystal being free of CH2Cl2 molecules, a value of
Upol(CH2Cl2) = –25 kJmol–1 was used as before leading to
n = 8 for the number of solvent molecules per complex that
should be eliminated during crystal growth.
The next complexation reaction was performed in the
presence of the bpym ligand in CH2Cl2. Two polymorphic
structures were obtained depending on the solvent used as
precipitant. A first polymorph crystallised if the CH2Cl2
mixture was overlaid with alkane solvents (n-pentane or n-
hexane). This polymorph crystallises in a monoclinic crystal
system with a P21/n space group. The unit cell contains
eight molecules. The second polymorph was obtained when
diethyl ether was allowed to diffuse into the CH2Cl2 mix-
ture. In this case, the complex crystallised in an orthorhom-
bic system with a Pbca space group. The unit cell also con-
tains eight molecules. However, the metrical description of
the molecular motif is identical in both cases. Figure 4 de-
picts the structure obtained from CH2Cl2/n-pentane.
Complexation with a Diamine Ligand: Solid- and Liquid-
State Characterisation of [Ti(dpbpol)2(opda)]
The last dinitrogenated ligand tested was the o-phenylene
diamine ligand. The reaction conditions used to synthesise
5706
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 5701–5713

A Family of Neutral Monomeric Heteroleptic Titanium Complexes
the [Ti(dpbpol)2(opda)] complex are similar to those em- troscopy. Crystals were dissolved in dry CD2Cl2 and a series
ployed for isolating the previous complexes. The single crys- of experiments were performed. First of all, DOSY mea-
tal analysed by X-ray diffraction revealed that the complex surements confirmed that the analysed species in solution
¯
crystallises in the P1 space group with two molecules per is analogous to that characterised in the solid state [experi-
unit cell. Interestingly, the stereochemical feature of the mental Rh = (6.5Ϯ0.6) Å; calculated Rh = (7.1Ϯ0.4) Å
[Ti(dpbpol)2(opda)] complex characterised in the solid state evaluated from the structure]. The room-temperature 1H
is unique compared with the compounds built from the di- NMR spectrum (CD2Cl2) shows first the presence of a not
imine ligands. The crystal structure highlights the presence negligible amount of free opda and dpbpolH2 ligands. The
of the (Δ,aS,aR) and (Λ,aR,aS) enantiomers per unit cell. most intense signals appear as large resonances that were
In contrast, the complexes mentioned previously present a attributed to the NMR spectroscopic signature of the com-
common stereochemical feature since only two types of plex. This observation is in good accordance with a rapid
stereomers were characterised [(Δ,aS,aS) and (Λ,aR,aR)] in ΔiΛ inversion of configuration at the metal centre for
the solid state. All of the five X-ray crystallographically de- [Ti(dpbpol)2(opda)] as it is well documented that six-coor-
termined structures of the diimine-ligand-based complexes dinate bis(chelate)–titanium complexes usually display such
revealed that the dpbpol moieties display an identical con- a fluxional behaviour.[17] For instance, the NH protons ap-
figuration within a molecule. This observation suggests that pear as two broad signals located at δ = 3.504 and
the nature of the bidentate nitrogenated ligand governs the 2.975 ppm. Indeed, lowering the temperature induces dras-
overall stereochemical features of the resulting complex in tic changes in these resonances. At 194 K, a set of four dou-
the solid state. Thus,
a careful observation of the blets (J = 11.5 Hz) located between δ = 3.89 and 2.64 ppm
[Ti(dpbpol)2(opda)] structure shown in Figure 5 is particu- were observed as expected for the resonances of the NH
larly informative. The structure shows that at least two diastereotopic protons (Figure 6).
amino hydrogen atoms are oriented near the mean plane of
the phenyl rings.
Figure 6. 1H NMR spectra (CD2Cl2) of [Ti(dpbpol)2(opda)] at 298
(bottom) and 194 K (top). The region shown here concerns the
resonances of the NH protons belonging to the opda coordinated
ligand.
Next, a 2D rotating-frame NOE (ROESY) analysis was
performed on the [Ti(dpbpol)2(opda)] complex to detect
close spatial proximity between the NH protons and the
aromatic protons of dpbpol. Unfortunately, owing to the
complexity of the aromatic region, we were not able to es-
tablish a clear correlation between the NH protons and
neighbouring dpbpol protons.
The [Ti(dpbpol)2(opda)] complex was also analysed by
ES-MS. To our surprise, the two highest mass peaks that
appeared at m/z = 827.24 and 933.29 are different from
Figure 5. Structure of [Ti(dpbpol)2(opda)] as the (Δ,aS,aR) enanti-
omer. Selected bond lengths [Å]: Ti1–N1 2.273, Ti1–N2 2.227, Ti1–
O4 1.838, Ti1–O2 1.889, Ti1–O3 1.909, Ti1–O1 1.838. Selected
angles [°]: N1–Ti–N2 74.577, O4–Ti–O1 106.448, O2–Ti–O3
162.685. Hydrogen atoms are omitted (with the exception of the
NH hydrogen atoms) and the opda ligand is coloured differently
(light grey) from the two dpbpol ligands (black) for clarity.
The four centroid distances measured between the amino those expected for the [Ti(dpbpol)2(opda)] complex (exact
hydrogen atoms and the closest phenyl rings are found to mass: 828.247). As the ES source could behave like an elec-
be 2.61 (H1B), 2.86 (H2B), 2.915 (H2A) and 3.30 Å (H1A), trolysis cell[18] and the redox behaviour of coordinated opda
respectively. These short distances are consistent with pos- to form the o-benzoquinone diimine ligand (bqdi) is well
sible attractive NH–π interactions between the aromatic documented, we assumed that the analysed species is a
rings and the amino hydrogen atoms. Although NH–π in- complex in which the opda ligand has been oxidised into
teractions in proteins have been known for almost three the diimine analogue. Thus, the isotopic profile correspond-
decades,[15] in transition-metal complexes only a few exam- ing to [Ti(dpbpol)2(bqdi) + H]+ was simulated and a perfect
ples show this type of low interaction energy particularly if match was found with the experimental profile as attested
NH2 groups coordinated to transition metals are consid- by Figure 7.
ered.[16] Thus, to provide evidence of the presence of such
weak interactions in solution, we turned to NMR spec- only in the presence of CH2Cl2 solvent molecules, which
The crystallisation of the opda-based complex is possible
Eur. J. Inorg. Chem. 2012, 5701–5713
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5707

D. M. Weekes, N. Baradel, N. Kyritsakas, P. Mobian, M. Henry
FULL PAPER
Figure 7. Experimental (top) and simulated (bottom) profile for [Ti(dpbpol)2(bqdi) + H]+. A second peak found at a higher mass (m/z =
933.29) could be attributed to [Ti(dpbpol)2(bqdi)2 + H]+.
leads to an average network solvation energy of structures of [Ti2(μ-O)(dpbpol)2(8-q)2(HOiPr)] and [Ti2(μ-
Upol(CH2Cl2) = –21.4 kJmol–1 typical of strong dipole–di- O)(dpbpol)2(8-q)2(H2O)], respectively. [Ti2(μ-O)(dpbpol)2-
pole interactions. The electronic affinity of this ligand is (8-q)2(HOiPr)] and [Ti2(μ-O)(dpbpol)2(8-q)2(H2O)] crystal-
quite low, which leads to small dipolar moment and pack- lise in an orthorhombic [space group Pna2(1)] and a mono-
ing energy with an average Ti–N bond energy. The number clinic (space group C2) crystal system, respectively. These
n of CH2Cl2 molecules needed for the reaction was found two structures display the same central core; in one case
to be 7. As each complex is interacting with three CH2Cl2 a 2-propanol molecule is coordinated to a titanium atom,
molecules in the crystal this means that crystal growth im- whereas for the second structure it is a water ligand that is
plies elimination of about four solvent molecules per com- bound to the metallic centre as attested by the rather long
plex.
The six structures obtained from the six different biden- highly comparable and were found to be 2.134 and 2.222 Å
tate N,N-chelate ligands attest a posteriori the judicious for [Ti2(μ-O)(dpbpol)2(8-q)2(HOiPr)] and [Ti2(μ-
Ti–O distances. The Ti–O(monodentate ligand) distances are
choice of the chelating ligands to form in a predictable way O)(dpbpol)2(8-q)2(H2O)], respectively. These values are in a
heteroleptic mononuclear complexes bearing a TiO4N2 core straight line with average Ti–O distances of 2.172 Å deter-
starting from cis-[Ti(dpbpol)2(HOiPr)2]. However, to fur- mined from the 5.29 release of the Cambridge Crystallo-
ther develop the project, we sought to investigate the influ- graphic Database. As noted before, these two bimetallic
ence of an alternative type of ligand.
complexes possess an identical backbone. Both structures
highlight a dissymmetrical framework with two sixfold-co-
ordinated titanium(IV) atoms bridged by two types of oxy-
gen atoms. Infrared spectroscopy investigations of both
crystals confirmed the nature of these two monodentate li-
gands found in the structures.
Complexation with a N,O-Chelating Ligand: Solid- and
Liquid-State Characterisation of [Ti(μ-O)(dpbpol)2(8-q)2-
(HOiPr)] and [Ti(μ-O)(dpbpol)2(8-q)2(H2O)]
Although nitrogenated ligands generate only tris-chelate
Infrared spectra in the region between 3400 and
mononuclear complexes, we were intrigued to examine the 3600 cm–1 display for [Ti2(μ-O)(dpbpol)2(8-q)2(HOiPr)] a
influence of a N,O-chelate on the structure of the final com- single broad band at ν = 3546 cm–1 and for [Ti2(μ-
˜
plex. The 8-hydroxyquinoline was selected as it still con- O)(dpbpol) (8-q) (H O)] a sharp double band at ν = 3497
˜
2
2
2
tains a nitrogen-donor atom in an aromatic environment, and 3479 cm–1. The single signal for the analysis of [Ti(μ-
and the bite angle between binding sites is not dissimilar to O)(dpbpol)2(8-q)2(HOiPr)] was attributed to the O–H
those previously tested. Indeed, in this particular case, the stretch of the alcohol ligand bonded to the metal. The
formation of a complex incorporating a TiO4N2 core is not double band observed for the aquo complex is in good ac-
expected. Herein, we wanted to inspect whether the mono- cordance with the symmetric and antisymmetric O–H stret-
meric nature of the compounds described previously would chings. In general, coordinated water in transition-metal
be affected by employing a N,O-chelating ligand. The same complexes exhibits no observable bands in the high-fre-
reaction conditions were maintained as before. Thus, cis- quency region,[19] since water ligands in the crystalline lat-
[Ti(dpbpol)2(HOiPr)2] and 8-hydroxyquinoline (1 equiv.) tice are involved in hydrogen bonding. In the present case,
were dissolved in CH2Cl2. n-Pentane was allowed to slowly as attested by the molecular structure, the water ligand is
diffuse into the resulting solution and dark red crystals suit- trapped in a sterically hindered aromatic environment,
able for X-ray analysis appeared in the medium in a few which prohibits the possible formation of hydrogen bonds
days. The experiment was repeated several times and two with neighbouring complexes. Next, crystals of [Ti2(μ-
very similar structures were characterised from this series O)(dpbpol)2(8-q)2(HOiPr)] were dissolved in CD2Cl2 and
of crystallisation studies. Figures 8 (a and b) represent the the resulting solution was analysed by 1H NMR spec-
5708
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 5701–5713

A Family of Neutral Monomeric Heteroleptic Titanium Complexes
tre, contrary to what is observed in the case of the cis-
[Ti(dpbpol)2(HOiPr)2] precursor. Both complexes were also
submitted to mass spectrometry analysis (ESI). However,
we were able to detect only in the case of [Ti2(μ-O)(dpbpol)2-
(8-q)2(H2O)] a series of very low intensity molecular peaks
matching the structure characterised in the solid state {m/z
calcd. for [Ti2(μ-O)(dpbpol)2(8-q)2 + H]+ = 1073.219; found
1073.224; m/z calcd. for [Ti2(μ-O)(dpbpol)2(8-q)2 + Na]+ =
1095.201; found 1095.206}. The signals with more intense
peaks correspond to m/z values much higher than those ex-
pected for [Ti2(μ-O)(dpbpol)2(8-q)2(H2O)] and [Ti2(μ-
O)(dpbpol)2(8-q)2(HOiPr)].
The charge distributions for both the water and 2-propa-
nol complexes show the rather high electronic affinity of
the 8-q ligand with larger dipolar moment and packing en-
ergy for the 2-propanol complex. The Ti–O(H)iPr bond en-
ergy for this structure was found to be more stabilising than
in the reference complex, whereas that involving the aquo
ligand is rather similar to that of Ti–N bonds. For the
thermodynamic modelling we have to take into account the
fact that an oxolation reaction has occurred. The most
plausible way to explain this phenomenon is to assume that
in a first step, the bidentate 8-Hq ligand substitutes, as be-
fore, the two labile 2-propanol molecules. With L represent-
ing the dpbpol ligand, we may thus write:
[TiL2(HOiPr)2] + (8-Hq) Ǟ [TiL2(8-Hq)] + 2 iPrOH
Now coordination of the OH moiety to a very acidic TiIV
centre would favour an intramolecular proton transfer
towards one dpbpol ligand:
[TiL2(8-Hq)] Ǟ [TiL(LH)(8-q)]
Figure 8. a) Structure of [Ti2(μ-O)(dpbpol)2(8-q)2(HOiPr)]. Two
positions have been found for the 2-propanol ligand; here only one
is shown. Selected bond lengths [Å]: Ti1–N1 2.217, Ti2–N2 2.212,
Ti1–O8 2.134, Ti1–O1 1.887, Ti1–O7 1.777, Ti2–O7 1.903, Ti2–O4
1.887, Ti2–O6 1.947, Ti1–O5 2.051, Ti2–O5 2.086. Selected angles
This protonation clearly weakens the Ti–LH bond that
becomes sensitive to the unavoidable presence of some
moisture in the solvent:
[°]: N1–Ti1–O5 74.588, N2–Ti2–O6 76.252, O8–Ti1–O1 84.196, [TiL(LH)(8-q)] + H2O Ǟ [TiL(8-q)(OH)] + LH2
O7–Ti1–O5 79.022, O7–Ti2–O5 75.439. b) Structure of [Ti(μ-
Now we have an unstable fivefold coordinated Ti centre
that is prone to dimerisation for retrieving a more stable
sixfold coordination:
O)(dpbpol)2(8-q)2(H2O)]. Selected bond length [Å]: Ti1–O7 2.222.
Hydrogen atoms are omitted (except for the water molecule in b)
and the 8-q ligand as well the monodentate ligands are coloured
differently (light grey) from the two dpbpol ligands (black) for clar-
ity.
2 [TiL(8-q)(OH)] Ǟ [Ti2(μ-O)L2(8-q)2(H2O)]
From a thermodynamic viewpoint, we should discard all
the unstable intermediates leading to an overall scheme:
1
troscopy. The resulting room-temperature H NMR spec-
trum is highly complex owing to the asymmetric nature of
the complex and to the presence of very large signals attrib-
uted to intramolecular slow dynamic processes at the NMR
2cis-[Ti(dpbpol)2(HOiPr)2] + 2(8-Hq) + 2H2O Ǟ [Ti2(μ-
O)(dpbpol)2(8-q)2(H2O)]
(CH2Cl2)3}
+ 2H2dpbpol + 2{(iPrOH)2·
1
spectroscopic timescale. The H NMR spectroscopy room-
temperature DOSY experiment indicates the presence of a
From Table 1, we may evaluate for this reaction ΔH =
nϫUpol(CH2Cl2) kJmol–1 and ΔS
single complex in solution that diffuses at
D
=
1441
+
=
(760Ϯ80) μm2 s–1 corresponding to an experimental hydro- 377 Jmol–1 K–1. From the equilibrium condition ΔH =
dynamic value of Rh = (7.0Ϯ0.7) Å, whereas the theoretical TϫΔS and assuming Upol(CH2Cl2) = –25 kJmol–1 as be-
value extracted from the crystal structure is Rh
=
fore, this then leads to n = 53. This rather large value may
(7.4Ϯ0.4) Å. This result indicates that the dimeric structure explain why the crystal structure of this compound displays
obtained in the solid state is conserved in solution. In ad- large cavities holding a large amount of solvent molecules
dition, this result proved that at room temperature the 2- that are strongly disordered (the structure was solved using
propanol ligand is indeed coordinated to the titanium cen- the SQUEEZE command owing to this large disorder). Ac-
Eur. J. Inorg. Chem. 2012, 5701–5713
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5709

D. M. Weekes, N. Baradel, N. Kyritsakas, P. Mobian, M. Henry
FULL PAPER
cordingly, if these solvent molecules were not present in the (covalent contributions have been added relative to
crystal, crystallisation will not be thermodynamically al- Table 1): iPrOH (–1015 and –943 kJmol–1) Ͻ dphpen
lowed.
(–810 kJmol–1)
Ͻ
Ͻ
Ͻ
dtpby (–805 kJmol–1)
H2O (–782 kJmol–1)
opda (–773 kJmol–1)
Ͻ
bpym
For the 2-propanol solvate, the same overall scheme may (–787 kJmol–1)
be assumed, but as water is much less abundant than 2- (–779 kJmol–1)
Ͻ dmp
Ͻ
dafone
propanol, the mass action law may explain the substitution (–734 kJmol–1). The Ti–O bond energy (average at
of the aquo ligand by an 2-propanol molecule:
–913 kJmol–1) is significantly more negative than the Ti–N
bond energy (average at –781 kJmol–1); the positive entropy
variation associated with the liberation of the two 2-propa-
nol ligands was, in all the investigated cases, not large
enough to overcome the large endothermic contribution as-
sociated with the replacement of two highly polar Ti–O
bonds by two less polar Ti–N bonds, when considering un-
solvated complexes. Consequently, a necessary condition
for these substitutions to occur spontaneously (ΔG Յ 0) is
a better solvation of the η2-chelate final complex relative to
the bis(η1-solvate) starting complex. In terms of solvation
number, this translates to about Δn = 5–8 CH2Cl2 molecules
(except for the 8-Hq ligand for which this number is much
larger: Δn = 53–56) in a specific interaction (about
–25 kJmol–1) with the η2-chelate final complex relative to a
nonsolvated bis(η1-solvate) starting complex. This manda-
tory solvation of the η2-chelate final complex is in accord-
ance with the experimental fact that upon crystallisation
some CH2Cl2 molecules are not expelled during crystal
growth. This is particularly well illustrated in the case of
the 8-Hq ligand in which cavities are maintained within the
[Ti2(μ-O)L2(8-q)2(H2O)] + iPrOH Ǟ [Ti2(μ-O)L2(8-q)2-
(iPrOH)] + H2O
Consequently, the following scheme may be assumed:
2cis-[Ti(dpbpol)2(HOiPr)2] + 2(8-Hq) + H2O Ǟ [Ti2(μ-
O)(dpbpol)2(8-q)2(iPrOH)] + 2H2dpbpol + {(iPrOH)2·
(CH2Cl2)3} + {iPrOH·(CH2Cl2)2}
which leads to ΔH = 1500 + nϫUpol(CH2Cl2) kJmol–1 and
ΔS = 312 Jmol–1 K–1. From the equilibrium condition ΔH
= TϫΔS and assuming Upol(CH2Cl2) = –25 kJmol–1 as be-
fore, this then leads to n = 56 CH2Cl2 molecules. This is of
the same order of magnitude as for the water-based dimer
and could explains the large amount of solvent molecules
that are strongly disordered (the structure was solved using
the SQUEEZE command owing to this large disorder).
Discussion
From the above results, one may deduce some interesting crystal to hold a large number of disordered CH2Cl2 mole-
general rules concerning monomeric octahedral Ti-based cules. To explain why some complexes crystallise without
complexes. First it could be observed that the charge on the any associated CH2Cl2 molecules, it is worth considering
Ti atom is roughly constant (+2.77 Յ q Յ +2.80) among the molecular dipole moment of each complex. Basically, if
the TiO4N2 species and significantly lower than that in the this dipolar moment is large, dipole–dipole interactions
TiO6 starting complex (q = +2.88). This clearly reflects the upon stacking of these complexes in the solid state could
greater covalent character of the Ti–N bond relative to the compensate for the energy loss coming from the desol-
Ti–O ones. The presence of an oxo bridge in the dimeric vation, whereas if the dipolar moment is low, solvent encap-
species explains the much lower partial charge on Ti atoms sulation becomes mandatory to minimise lattice energy. Ac-
(ϽqϾ = +2.54) owing to the strong electron-donating capa- cordingly, we may order our complexes according to their
bilities of both O2– and 8-q– ligands relative to neutral li- decreasing molecular dipolar moment indicating in each
gands such as iPrOH, dmp, dphphen, dafone, bpym, dtbipy case the number n of CH2Cl2 molecules retained within the
and opda. This also reflects the hard character of Ti atoms crystal: dtbpy (μ = 2.97 D, n = 0) Ͼ dafone (μ = 2.87 D, n
that allows electronic density to flow almost freely from the = 0) Ͼ iPrOH (μ = 2.07 D, n = 0) Ͼ dmp (μ = 1.70 D, n =
electron-rich dpbpol ligands towards the neutral electron- 4) Ͼ 8-q (μ = 1.34 D, n Ͼ 0) bpym (μ = 1.21 D, n = 0) Ͼ
acceptor moieties. For the charge on the dpbpol ligands the opda (μ = 0.85 D, n = 3) Ͼ dphphen (μ = 0.55 D, n = 2).
following order was found: (iPrOH)2 (–1.20) Ͻ opda (–1.14) CH2Cl2 molecules seems thus to be trapped within the crys-
Ͻ 8-q (–1.12) Ͻ dtbipy (–1.11) Ͻ dmp (–1.08) Ͻ dphphen tal when μ Ͻ 2.0 D with the exception of bpym, which has
(–1.07) Ͻ bpym (–1.03) Ͻ dafone (–1.02). For the negative a rather low dipolar moment and forms a crystal without
charge gained by each ligand, the following order was any CH2Cl2 molecules. However in this last case, two crys-
found: opda (–0.48) Ͻ (iPrOH)2 (–0.52) Ͻ dtbipy (–0.55) tallographically nonequivalent complexes are found within
Ͻ dmp (–0.62) Ͻ dphphen (–0.67) Ͻ bpym (–0.73) Ͻ the unit cell that create a new, more polar dimeric entity (μ
dafone (–0.74) Ͻ 8-q (–0.88). These two series quantify the = 1.67 D). As the bpym complex is the one displaying the
relative electron-withdrawing ability of these ligands and lowest solvation number (Δn = 5), the squeezing of the
are in accordance with what may be expected based on pu- CH2Cl2 molecules from the crystal could be the conse-
rely qualitative electronegativity considerations.
quence of such a dimerisation process. This observation
Nucleophilic substitution of two solvent molecules in the also applies to the dphphen complex that also has two crys-
cis position by a bidentate ligand within an octahedral com- tallographically nonequivalent complexes found within the
plex is a simple process that can be under enthalpic and/or unit cell. However, the dipolar moment in this case is so
entropic control. It was thus found that total iono-covalent low that dimerisation alone is not enough to remove the
Ti–X bond energies (X = O or N) increase in the order CH2Cl2 molecules.
5710
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 5701–5713

A Family of Neutral Monomeric Heteroleptic Titanium Complexes
1H/X z-gradient BBI probe and by applying a PFGSTE pulse se-
quence using bipolar gradients. DOSY spectra were generated with
the DOSY module of NMRNotebookTM software, through maxi-
mum entropy and inverse Laplace transform calculations. 1H NMR
DOSY measurements were performed at 500.13 MHz with a 5 mm
1H/X z-gradient BBI probe and by applying a PFGSTE pulse se-
quence using bipolar gradients. Mass spectrometry was performed
at the Service Commun d’Analyses, Université de Strasbourg
(France). The electrospray analyses were performed on a Micro-
TOF (Bruker) apparatus equipped with an electrospray (ES)
source. The analysed mass peaks refer to the most intense peaks
(unless mentioned otherwise in the Discussion). IR spectra were
recorded on a FTIR-8400S Shimadzu apparatus. Ligands were
used as received from Alfa–Aesar, TCI, or Sigma–Aldrich. The
dafone ligand was prepared by literature methods.[22] All solvents
were dried before use with activated 3 Å molecular sieves.
Conclusion
Herein, we have described a rational approach leading to
a novel family of monomeric heteroleptic titanium com-
plexes formed with two biphenolate derivatives and a biden-
tate nitrogenated chelate that have been classified quantita-
tively according to their electron-withdrawing capabilities.
The proposed strategy involved the entropically favoured
displacement of two monodentate-coordinated alcohol mo-
lecules belonging to the starting cis-[Ti(dpbpol)2(HOiPr)2]
complex assisted by solvation of the η2-chelate product to
overcome the enthalpically defavourable replacement of two
highly polar Ti–O bonds by less polar ones. The systematic
crystal-structure determination of the resulting architec-
tures provided important information concerning the influ-
ence of a donor ligand on the overall stereochemistry of the
resulting complexes. In particular, we found that diimine
chelates lead exclusively to complexes in which (Δ,aS,aS) or
(Λ,aR,aR) stereochemistry is adopted, whereas in the spe-
cific case of the opda ligand only the (Δ,aS,aR) and
(Λ,aR,aS) enantiomers were characterised in the crystal.
Such a difference may be explained by the presence of rarely
reported NH–π interactions in the case of the [Ti(dpbpol)2-
(opda)] derivative. Over the course of this study, we also
intensively used DOSY measurements to prove that the
complexes characterised in the solid state remain identical
in the liquid phase. The occurrence of CH2Cl2 molecules in
the crystalline state and the existence of two crystallograph-
ically nonequivalent complexes within the unit cell have
been associated with a low dipolar moment of the com-
plexes. Other complexes with large dipolar moments (μ Ͼ
2 D) have enough dipolar energy in the solid state to re-
move all the solvating CH2Cl2 molecules. Finally, we have
also tested the influence of a N,O-coordinating ligand. The
structure obtained contains a doubly bridged titanium cen-
tre that indicates a loss of structural predictability. All these
findings have been interpreted by using a QFT-based for-
malism for deducing thermodynamic data directly from a
crystalline structure. Even if the modelled reaction was
quite trivial, it should be understood that the proposed
methodology is quite general and can be applied as soon
as a crystal structure becomes available. In particular, no
empirical input is needed for computing internal energies
and solution entropies. This is a great advantage that allows
an easy overlap between theory and experiments, thereby
bringing supramolecular chemistry to the same level as
physics. Current studies are underway to exploit the struc-
tural information extracted from this study to elaborate in
a predictable way novel helical metallo-supramolecular
architectures constructed around heteroleptic TiO4N2 mo-
tifs in which the oxygenated ligand system used incorpo-
rates two 2,2Ј-biphenol units.[20,21]
General Procedure: Under nitrogen (glovebox), a counter solvent
was allowed to slowly diffuse into a solution containing the ligand
(1 equiv.) and the cis-[Ti(dpbpol)2(HOiPr)2] complex (5 mg) in a
minimum amount of dry CH2Cl2 in a small vial.
[Ti(dpbpol)2(dtbipy)]: Crystals were obtained from CH2Cl2/n-pent-
ane. TiO4N2C66H56; T = 173(2) K; monoclinic; space group P2(1);
a = 19.3783(6) Å, b = 13.5855(5) Å, c = 20.3398(6) Å; α = 90°, β =
104.459(2)°,
γ = 90°; V = Z = 4; Dcalcd. =
5185.1(3) Å3;
1.267 Mgm–3; reflections collected: 49457, Rint = 0.0370; refine-
ment factors R = 0.0409, RW = 0.1048; GOF on F2 = 1.012. H
1
3
NMR ([D8]toluene, 400 MHz): δ = 8.22 (d, J = 5.9 Hz, 2 H), 7.73
(dd, 3J = 8.5 Hz, 4J = 1.2 Hz, 3 H), 7.42 (dd, 3J = 7.8 Hz, 4J =
3
4
3
1.9 Hz, 2 H), 7.33 (dd, J = 7.5 Hz, J = 1.9 Hz, 2 H), 7.26 (dd, J
= 7.7 Hz, 4J = 1.9 Hz, 2 H), 7.13 (t, 3J = 8 Hz, 4 H), 7.06–6.98
(complex due to the superimposition with the signals of toluene),
6.75 (d, 4J = 1.7 Hz, 2 H), 6.69 (dd, 3J = 7.5 Hz, 4J = 2.1 Hz, 2
3
3
H), 6.59 (d, J = 7.5 Hz, 2 H), 6.43–6.30 (m, 5 H), 6.14 (dd, J =
5.9 Hz, 4J = 1.9 Hz, 2 H), 1.031 (s, 18 H) ppm. 13C NMR (CD2Cl2,
125 MHz): δ = 163.20 (C), 161.55 (C), 160.54 (C), 151.26 (C),
149.56 (CH), 139.50 (C), 139.16 (C), 132.06 (CH), 131.83 (CH),
130.46 (C), 130.42 (C), 130.27 (CH), 129.89 (CH), 129.60 (CH),
129.17 (C), 128.83 (CH), 127.95 (C), 127.77 (CH), 126.55 (CH),
126.24 (CH), 126.13 (CH), 121.99 (CH), 121.94 (CH), 119.51 (CH),
117.80 (CH), 35.40 (C), 30.42 (CH3) ppm. 1H NMR DOSY
(CD Cl , 500 MHz): D = (740Ϯ10%) μm2 s–1. IR: ν = 1397, 1411,
˜
2
2
1618, 2336, 2359, 2996, 3053 cm–1. MS (ESI): m/z calcd. for
[Ti(dpbpol)2(dtbipy) + H]+ 989.380; found 989.380.
[Ti(dpbpol)2(dphphen)]: Crystals were obtained from CH2Cl2/n-
pentane. 4[TiO4N2C72H48]·CH2Cl2; T = 173(2) K; monoclinic;
space group C2; a = 38.5715(18) Å, b = 13.8581(6) Å, c =
42.327(2) Å; α = 90°, β = 101.909(2)°, γ = 90°; V = 22138.2(18) Å3;
Z = 4; Dcalcd. = 1.289 Mgm–3; reflections collected: 50661, Rint
=
0.0533; refinement factors R = 0.1026, RW = 0.2452; GOF on F2
1
3
= 1.066. H NMR (CD2Cl2, 500 MHz): δ = 8.69 (d, J = 5.3 Hz, 2
H), 7.45–7.66 (m, 18 H), 7.29–7.33 (m, 4 H), 7.17 (dd, 3J = 7.6 Hz,
4J = 1.8 Hz, 2 H), 7.01–7.10 (m, 9 H), 6.61 (t, J = 7.4 Hz, 2 H),
3
3
4
6.53 (dd, J = 7.6 Hz, J = 1.7 Hz, 2 H), 6.35–6.47 (m, 9 H) ppm.
13C NMR (CD2Cl2, 125 MHz): δ = 161.42 (C), 160.76 (C), 150.89
(C), 149.56 (CH), 143.35 (C), 139.50 (C), 138.86 (C), 136.82 (C),
132.07 (CH), 131.91 (CH), 130.46 (C), 130.34 (C), 130.05 (CH),
129.87 (CH), 129.76 (CH), 129.40 (CH), 129.23 (C), 128.71 (CH),
128.08 (C), 127.89 (CH), 127.86 (CH), 126.50 (CH), 126.26 (CH),
124.93 (CH), 124.06 (CH), 122.13 (CH), 119.70 (CH) ppm. 1H
Experimental Section
NMR Spectroscopy: Bruker Avance-300, Avance-400 and Avance-
500 were used for solution NMR spectroscopic analysis. H NMR
DOSY measurements were performed at 500.13 MHz with a 5 mm
NMR DOSY (CD Cl , 500 MHz): D = (740Ϯ10%) μm2 s–1. IR: ν
˜
2
2
= 1397, 1416, 1603, 2956, 3044 cm–1. MS (ESI): m/z calcd. for
1
[Ti(dpbpol)2(dphphen) + H]+ 1053.317; found 1053.327.
Eur. J. Inorg. Chem. 2012, 5701–5713
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5711

D. M. Weekes, N. Baradel, N. Kyritsakas, P. Mobian, M. Henry
FULL PAPER
[Ti(dpbpol)2(5,6-dmp)]: Crystals were obtained from CH2Cl2/n-
pentane. TiO4N2C62H44·CH2Cl2; T = 173(2) K; orthorhombic;
β = 93.4830(10)°, γ = 90°; V = 8513.4(3) Å3; Z = 8; Dcalcd.
=
1.371 Mgm–3; reflections collected: 132781, Rint = 0.0486; refine-
ment factors R = 0.0502, RW = 0.1067; GOF on F2 = 1.020.
space group Pccn;
a = 39.769(7) Å, b = 12.598(3) Å, c =
19.318(4) Å; α = 90°, β = 90°, γ = 90°; V = 9678(3) Å3; Z = 8;
Dcalcd. = 1.392 Mgm–3; reflections collected: 53714, Rint = 0.0976;
refinement factors R = 0.0836, RW = 0.1448; GOF on F2 = 1.207.
1H NMR (CD2Cl2, 500 MHz): δ = 8.58 (dd, 3J = 4.9 Hz, 4J =
Crystals were obtained from CH2Cl2/n-pentane. TiO4N4C56H38; T
= 173(2) K; orthorhombic; space group Pbca; a = 11.9476(4) Å, b
= 21.8675(8) Å, c = 33.3302(12) Å; α = 90, β = 90°, γ = 90°; V =
8708.0(5) Å3; Z = 8; Dcalcd. = 1.341 Mgm–3; reflections collected:
77102, Rint = 0.0514; refinement factors R = 0.0437, RW = 0.1058;
3
4
1.2 Hz, 2 H), 8.13 (dd, J = 8.3 Hz, J = 1.1 Hz, 2 H), 7.47–7.50
(m, 4 H), 7.26–7.29 (m, 4 H), 7.12–7.17 (m, 4 H), 6.99–7.07 (m, 8
1
GOF on F2 = 1.017. H NMR (CD2Cl2, 600 MHz): δ = 8.69 (dd,
3
3
4
H), 6.56 (t, J = 7.5 Hz, 2 H), 6.44 (dd, J = 7.3 Hz, J = 1.9 Hz,
2 H), 6.22–6.34 (m, 10 H), 2.64 (s, 6 H) ppm. 13C NMR (CD2Cl2,
125 MHz): δ = 161.27 (C), 160.59 (C), 148.50 (CH), 141.88 (C),
139.48 (C), 138.89 (C), 135.15 (CH), 132.08 (CH), 131.84 (CH),
130.78 (C), 130.44 (C), 130.34 (C), 130.27 (CH), 129.96 (CH),
129.78 (CH), 129.30 (C), 129.13 (C), 129.11 (CH), 128.97 (CH),
128.15 (CH), 127.42 (CH), 126.21 (CH), 126.18 (CH), 124.62 (CH),
123.51 (CH), 122.06 (CH), 119.49 (CH), 14.75 (CH3) ppm. 1H
3
4
3
4J = 2.2 Hz, J = 4.8 Hz, 1 H), 8.53 (dd, J = 2.3 Hz, J = 5.3 Hz,
1 H), 7.45 (dd, 4J = 1.9 Hz, 3J = 8.4 Hz, 2 H), 7.29 (d, 3J = 7.7 Hz,
2 H), 7.15 (dd, 4J = 1.9 Hz, J = 7.5 Hz, 1 H), 7.10 (t, 3J = 7.8 Hz,
3
1 H), 7.08–7.02 (complex, 3 H), 6.92 (t, 3J = 4.9 Hz, 1 H), 6.83
(dd, 4J = 0.9 Hz, 3J = 8.6 Hz, 2 H), 6.74 (t, J = 7.7 Hz, 2 H), 6.72
3
(dd, 4J = 1.9 Hz, 4J = 7.7 Hz, 1 H), 6.67 (t, 3J = 7.6 Hz, 2 H) ppm.
13C NMR (CD2Cl2, 125 MHz): δ = 160.93, 160.37, 160.34, 158.02,
156.21, 139.11, 138.72, 132.15, 131.83, 130.19, 130.10, 129.92,
129.66, 129.48, 129.41, 129.18, 129.13, 128.47, 127.87, 127.52,
126.47, 126.40, 122.63, 121.87, 120.18 ppm. 1H NMR DOSY
NMR DOSY (CD Cl , 500 MHz): D = (820Ϯ10%) μm2 s–1. IR: ν
˜
2
2
= 1397, 1416, 1604, 3041, 3064 cm–1. MS (ESI): m/z calcd. for
[Ti(dpbpol)2(5,6-dmp) + H]+ 929.286; found 929.315.
(CD Cl , 500 MHz): D = (860Ϯ10%) μm2 s–1. IR: ν = 837, 864,
˜
2
2
1075, 1082, 1221, 1235, 1247, 1400, 1412, 1573 cm–1. MS (ESI):
[Ti(dpbpol)2(dafone)]: Crystals were obtained from CH2Cl2/n-pent-
m/z calcd. for [Ti(dpbpol)2(bpym) + H]+ = 879.245; found 879.25.
¯
ane. TiO5N2C59H38; T = 173(2) K; triclinic; space group P1; a =
11.5181(17) Å, b = 14.227(2) Å, c = 15.466(4) Å; α = 107.753(8)°,
[Ti(μ-O)(dpbpol)2(8-q)2(HOiPr)]: Crystals were obtained from
CH2Cl2/n-pentane. Ti2O8N2C69H51; T = 173(2) K; orthorhombic;
space group Pna2(1); a = 19.8443(5) Å, b = 27.0583(8) Å, c =
11.0010(3) Å; α = 90°, β = 90°, γ = 90°; V = 5907.0(3) Å3; Z = 4;
Dcalcd. = 1.273 Mgm–3; reflections collected: 37614, Rint = 0.0292;
refinement factors R = 0.0586, RW = 0.1486; GOF on F2 = 1.054.
1H NMR (CD2Cl2, 600 MHz): the NMR spectrum is highly com-
plex to analyse owing to the presence of very large signals, however,
partial identification of several signals could be performed: δ = 7.93
(large doublet, 3J = 7.4 Hz, 2 H), 7.49 (large doublet, 3J = 7.5 Hz),
7.38 (large triplet, 3J = 7.7 Hz), 7.14 (large triplet, 3J = 7.5 Hz),
β = 96.340(7)°, γ = 110.550(5)°; V = 2191.9(8) Å3; Z = 2; Dcalcd.
=
1.368 Mgm–3; reflections collected: 21400, Rint = 0.0550; refine-
ment factors R = 0.1011, RW = 0.2709; GOF on F2 = 1.131. H
1
3
4
NMR (CD2Cl2, 500 MHz): δ = 7.77 (dd, J = 5.3, J = 1.1 Hz, 1
3
4
3
4
H), 7.62 (dd, J = 7.7, J = 1.3 Hz, 1 H), 7.42 (dd, J = 8.4 Hz, J
3
4
= 1.7 Hz, 2 H), 7.28 (m, 2 H), 7.19 (dd, J = 7.7 Hz, J = 1.9 Hz,
3
3
1 H), 6.96–7.10 (m, 4 H), 6.97 (dd, J = 5.3 Hz, J = 7.6 Hz, 1 H),
3
4
3
3
6.83 (dd, J = 8.2 Hz, J = 1.8 Hz, 2 H), 6.77 (dd, J = 1.9 Hz, J
= 7.7 Hz, 1 H), 6.71 (t, 3J = 7.5 Hz, 1 H), 6.60–6.64 (m, 3 H) ppm.
13C NMR (CD2Cl2, 125 MHz): δ = 185.81 (C), 161.27 (C), 160.59
(C), 159.13 (C), 151.23 (CH), 139.52 (C), 139.08 (C), 134.05 (CH),
132.77 (CH), 131.95 (CH), 131.46 (CH), 131.08 (CH), 130.09 (CH),
130.02 (CH), 129.78 (CH), 129.63 (CH), 129.54 (CH), 129.13 (CH),
129.01 (C), 128.50 (C), 128.25 (C), 128.02 (C), 127.92 (CH), 127.32
(CH), 126.84 (CH), 126.45 (CH), 126.39 (CH), 122.75 (CH), 121.74
3
3
7.07 (large doublet, J = 8.2 Hz), 6.75 (large doublet, J = 6.6 Hz)
ppm. 13C NMR (CD2Cl2, 125 MHz): δ = 161.13, 160.29, 150.26,
146.69, 141.87, 138.91, 137.97, 131.47, 131.08, 130.56, 130.14,
130.04, 129.94, 129.79, 129.56, 129.48, 129.14, 128.07, 128.03,
126.70, 125.91, 125.42, 121.97, 121.78, 121.74, 116.80, 111.03, 64.7,
1
(CH), 120.21 (CH) ppm. H NMR DOSY (CD2Cl2, 500 MHz): D
25.51 ppm. 1H NMR DOSY (CD2Cl2, 600 MHz):
D =
= (830Ϯ10%) μm2 s–1. IR: ν = 1397, 1417, 1587, 1731, 3053 cm–1.
˜
(760Ϯ10%) μm2 s–1. IR: ν = 698, 748, 873, 983, 1110, 1238, 1320,
˜
MS (ESI): m/z calcd. for [Ti(dpbpol)2(dafone) + H]+ = 903.234;
1379, 1406, 1420, 1468, 1498, 1578, 3546 cm–1.
found 903.261.
[Ti(μ-O)(dpbpol)2(8-q)2(H2O)]: Crystals were obtained from
CH2Cl2/n-pentane. Ti2O8N2C46H66; T = 173(2) K; monoclinic;
space group C2; a = 29.1950(17) Å, b = 18.3101(10) Å, c =
27.118(2) Å; α = 90°, β = 122.178(2)°, γ = 90°; V = 12269.4(14) Å3;
[Ti(dpbpol)2(opda)]: Crystals were obtained from CH2Cl2/n-pent-
ane. TiO4N2C54H40; T = 173(2) K; triclinic; space group P1; a =
¯
10.68695(3) Å, b = 11.7051(4) Å, c = 20.8110(8) Å; α = 82.2920
(10)°, β = 84.6260 (10)°, γ = 74.9230 (10)°; V = 2486.25(15) Å3; Z
Z = 8; Dcalcd. = 1.181 Mgm–3; reflections collected: 115897, Rint
=
= 2; Dcalcd. = 1.334 Mgm–3; reflections collected: 28700, Rint
=
0.0475; refinement factors R = 0.0589, RW = 0.1519; GOF on F2
0.0458; refinement factors R = 0.0879, RW = 0.2442; GOF on F2
= 1.028. 1H NMR (CD2Cl2, 600 MHz): δ = 7.60 (large signals, 4
H), 7.38 (large, 6 H), 7.25–6.9 (large and complex, 20 H), 6.8–6.7
(large signal, 4 H), 6.43 (dd, J = 4.2 Hz, J = 5.6 Hz, 2 H), 3.55
(large, 2 H, NH), 3.02 (large, 2 H) ppm. 13C NMR (CD2Cl2,
125 MHz): δ = 160.56, 159.94, 150.37, 140.80, 139.30, 138.08,
136.83, 132.44, 132.24, 131.58, 131.20, 130.42, 130.18, 130.08,
129.90, 129.58, 129.26, 128.19, 127.91, 127.02, 126.98, 126.66,
122.29, 121.87, 120.82 ppm. 1H NMR DOSY (CD2Cl2, 500 MHz):
= 1.070. IR: ν = 703, 754, 759, 799, 828, 1071, 1166, 1215, 1228,
˜
1246, 1326, 1420, 1434, 1448, 3480, 3497 cm–1. MS (ESI): m/z
calcd. for [Ti(μ-O)(dpbpol)2(8-q)2 + H]+ 1073.219; found 1073.224;
m/z calcd. for [Ti(μ-O)(dpbpol)2(8-q)2 + Na]+ 1095.201; found
1095.206.
CCDC-881439 {for [Ti(dpbpol)2(dtbipy)]}, -881440 {for
[Ti(dpbpol)2(dafone)]}, -881441 {for [Ti(dpbpol)2(5,6-dmp)]},
-881442 {for [Ti(dpbpol)2(dphphen)]}, -881443 {for [Ti(dpbpol)2-
(opda)]}, -881444 {for [Ti(μ-O)(dpbpol)2(8-q)2(H2O)]}, -881445
{for [Ti(μ-O)(dpbpol)2(8-q)2(HOiPr)]}, -881446 {for [Ti(dpbpol)2-
(bpym)] from CH2Cl2/diethyl ether} and -881447 {for [Ti(dpbpol)2-
(bpym)] from CH2Cl2/n-pentane} contain the supplementary crys-
tallographic data for this paper. These data can be obtained free
of charge from The Cambridge Crystallographic Data Centre via
D = (810Ϯ10%) μm2 s–1. IR: ν = 872, 1227, 1248, 1398, 1419 cm–1.
˜
ES (ESI): calcd. for [Ti(dpbpol)2(opda) – 2H + H]+ = 827.24; found
827.24.
[Ti(dpbpol)2(bpym)]: Crystals were obtained from CH2Cl2/diethyl
ether. TiO4N4C56H38; T = 173(2) K; monoclinic; space group P21/
n; a = 12.3274(2) Å, b = 16.5431(3) Å, c = 41.8230(7) Å; α = 90°, www.ccdc.cam.ac.uk/data_request/cif.
5712 Eur. J. Inorg. Chem. 2012, 5701–5713
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

A Family of Neutral Monomeric Heteroleptic Titanium Complexes
[7] a) K. Mikami, Y. Matsumoto, L. Xu, Inorg. Chim. Acta 2006,
359, 4159–4167; b) K. M. Waltz, P. J. Carroll, P. J. Walsh, Orga-
nometallics 2004, 23, 127–134; c) J. P. Corden, W. Errington, P.
Moore, M. G. Partridge, M. G. H. Wallbridge, Dalton Trans.
2004, 1846–1851; d) T. J. Davis, J. Balsells, P. J. Carroll, P. J.
Walsh, Org. Lett. 2001, 3, 699–702; e) K. Gigant, A. Rammal,
M. Henry, J. Am. Chem. Soc. 2001, 123, 11632–11637; f) S.
Pandiaraju, G. Chen, A. Lough, A. K. Yudin, J. Am. Chem.
Soc. 2001, 123, 3850–3851; g) M. Terada, Y. Matsumoto, Y.
Nakamura, K. Mikami, Inorg. Chim. Acta 1999, 296, 267–272;
h) T. J. Boyle, D. L. Barnes, J. A. Heppert, L. Morales, F. Taku-
sagawa, J. C. Connolly, Organometallics 1992, 11, 1112–1126.
[8] a) L. Pauling, J. Am. Chem. Soc. 1932, 54, 3570–3582; b) R. G.
Parr, R. A. Donnelly, M. Levy, W. E. Palke, J. Chem. Phys.
1978, 68, 3801–3807; c) J. B. Mann, T. L. Meek, L. C. Allen, J.
Am. Chem. Soc. 2000, 122, 2780–2783; d) J. B. Mann, T. L.
Meek, E. T. Knight, J. F. Capitani, L. C. Allen, J. Am. Chem.
Soc. 2000, 122, 5132–5137.
[9] J. T. Waber, D. T. Cromer, J. Chem. Phys. 1965, 42, 4116–4123.
[10] a) M. Henry, ChemPhysChem 2002, 3, 561–569; b) M. Henry,
ChemPhysChem 2002, 3, 607–616; c) M. Henry, M. W. Hosse-
ini, New J. Chem. 2004, 28, 897–906; d) C. Carpanese, S. Ferlay,
N. Kyritsakas, M. Henry, M. W. Hosseini, ChemComm. 2009,
1, 6786–6788; e) M. Henry, in: Advances in Quantum Chemical
Bonding Structures (Ed.: M. V. Putz), Transworld Research
Network, Kerala, India, 2009, pp. 153–211; f) H. Senouci, B.
Millet, C. Volkringer, C. Huguenard, F. Taulelle, M. Henry, C.
R. Chim. 2010, 13, 69–96; g) M. Henry, in: Encyclopedia of
Nanoscience and Nanotechnology (Ed.: H. S. Nalwa), American
Scientific Publishers, vol. 14, 2011, pp. 1–43.
Supporting Information (see footnote on the first page of this arti-
1
cle): H NMR spectra, 13C NMR spectra and ES-MS analyses.
Acknowledgment
This work was performed at Université de Strasbourg with public
funds allocated by the Centre National de la Recherche (CNRS)
and the French government. The authors warmly thank Dr. S. Bau-
dron and Dr. J.-P. Collin for the gift of some ligands.
[1]
a) S. Kondo, K. Saruhashi, K. Seki, K. Matsubara, K. Miyaji,
T. Kubo, K. Matsumoto, T. Katsuki, Angew. Chem. 2008, 120,
10349; Angew. Chem. Int. Ed. 2008, 47, 10195–10198; b) A.
Yeori, S. Groysman, I. Goldberg, M. Kol, Inorg. Chem. 2005,
44, 4466–4468; c) L. Z. Flores-Lopéz, M. Parra-Hake, R. Som-
anathan, P. J. Walsh, Organometallics 2000, 19, 2153–2160.
[2]
a) A. Stopper, I. Goldberg, M. Kol, Inorg. Chem. Commun.
2011, 14, 715–718; b) A. L. Zelikoff, J. Kopilov, I. Goldberg,
G. W. Coates, M. Kol, Chem. Commun. 2009, 6804–6806; c)
G.-J. M. Meppelder, H.-T. Fan, T. P. Spaniol, J. Okuda, Orga-
nometallics 2009, 28, 5159–5165; d) P. J. Aragón, F. Carrillo-
Hermosilla, E. Villaseñor, A. Otero, A. Antiñolo, A. M.
Rodríguez, Eur. J. Inorg. Chem. 2006, 965–971; e) T. Gueta-
Neyroud, B. Tumanskii, M. Botoshansky, M. S. Eisen, J. Or-
ganomet. Chem. 2007, 692, 927–939; f) A. J. Chmura, M. G.
Davidson, M. D. Jones, M. D. Lunn, M. F. Mahon, Dalton
Trans. 2006, 887–889; g) A. J. Chmura, M. G. Davidson, M. D.
Jones, M. D. Lunn, M. F. Mahon, A. F. Johnson, P. Khunkam-
choo, S. L. Roberts, S. S. F. Wong, Macromolecules 2006, 39,
7250–7257.
[11] C. Diebold, P. Mobian, C. Huguenard, L. Allouche, M. Henry,
Dalton Trans. 2009, 10178–10180.
[12] a) H. Jiang, E. Stepowska, D. Song, Dalton Trans. 2008, 5879–
5881; b) A. K. Biju, M. V. Rajasekharan, Polyhedron 2008, 27,
2065–2068; c) S. Menom, M. V. Rajasekharan, Inorg. Chem.
1997, 36, 4983–4987; d) L. J. Henderson, F. R. Fronczek, W. R.
Cherry, J. Am. Chem. Soc. 1984, 106, 5876–5879.
[3]
a) A. Tzubery, E. Y. Tshuva, Inorg. Chem. 2012, 51, 1796–1804;
b) D. Peri, C. M. Manna, M. Shavit, E. Y. Tshuva, Eur. J. Inorg.
Chem. 2011, 4896–4900; c) D. Peri, S. Meker, M. Shavit, E. Y.
Tshuva, Chem. Eur. J. 2009, 15, 2403–2415; d) M. Guo, H.
Sun, S. Bihari, J. A. Parkinson, R. O. Gould, S. Parsons, P. J.
Sadler, Inorg. Chem. 1999, 38, 206–215.
[13] For a general review, see: K. G. Caulton, Eur. J. Inorg. Chem.
2012, 435–443; K. P. Butin, E. K. Beloglazkina, N. V. Zyk,
Russ. Chem. Rev. 2005, 76, 531–553; a) L. F. Warren, Inorg.
Chem. 1977, 16, 2814–2819, 47; b) R. C. Elder, D. Koran, H. B.
Mark, Inorg. Chem. 1974, 13, 1644–1649; c) G. G. Christoph,
V. L. Goedken, J. Am. Chem. Soc. 1973, 95, 3869–3875.
[14] Average between the maximum value corresponding to the
longest distance from the barycentre of a molecule Rmax and a
minimum value Rmin = (3V/4π)1/3 derived from the molecular
volume; estimated according to: A. Gavezotti, J. Am. Chem.
Soc. 1983, 105, 5220–5225. Van der Waals radii used, H: 1.2 Å,
C: 1.7 Å, O: 1.52 Å, N: 1.55 Å; see: A. Bondi, J. Phys. Chem.
1964, 68, 441–451; Ti: 1.45 Å (assumed).
[4] a) N. Hollingsworth, M. Kanna, G. Kociok-Köhn, K. C. Mol-
loy, S. Wongnawa, Dalton Trans. 2008, 631–641; b) M. Braun,
A. Hahn, M. Engelmann, R. Fleischer, W. Frank, C. Kryschi,
S. Haremza, K. Kürschner, R. Parker, Chem. Eur. J. 2005, 11,
3405–3412; c) T. J. Boyle, M. A. Rodriguez, T. M. Alam, Dal-
ton Trans. 2003, 4598–4603; d) J. M. Tanski, T. P. Vaid, E. B.
Lobkovsky, P. T. Wolczanski, Inorg. Chem. 2000, 39, 4756.
[5] M. L. N. Rao, H. Houjou, K. Hiratani, Chem. Commun. 2002,
420–421.
[6] a) T. Hermon, E. Y. Tshuva, J. Org. Chem. 2008, 73, 5953–
5958; b) T. Okamatsu, R. Irie, T. Katsuki, J. Organomet. Chem.
2007, 692, 645–653; c) K. V. Zaitsev, S. S. Karlov, A. A. Selina,
Y. F. Oprunenko, A. V. Churakov, B. Neumüller, J. A. K. How-
ard, G. S. Zaitseva, Eur. J. Inorg. Chem. 2006, 1987–1999; d)
K. C. Fortner, J. P. Bigi, S. N. Brown, Inorg. Chem. 2005, 44,
2803–2814; e) A. G. Maestri, S. N. Brown, Inorg. Chem. 2004,
43, 6995–7004; f) H. Chen, P. S. White, M. R. Gagné, Organo-
metallics 1998, 17, 5358–5366; g) S. M. Harben, P. D. Smith,
R. L. Beddoes, D. Collison, C. D. Garner, Angew. Chem. 1997,
109, 1985; Angew. Chem. Int. Ed. Engl. 1997, 36, 1897–1898;
h) F. Franceschi, E. Gallo, E. Solari, C. Floriani, A. Chiesi-
Villa, C. Rizzoli, N. Re, A. Sgamellotti, Chem. Eur. J. 1996, 2,
1466–1476; i) K. M. Carroll, J. Schwartz, D. M. Ho, Inorg.
Chem. 1994, 33, 2707–2708; j) R. Minhas, R. Duchateau, S.
Gambarotta, C. Bensimon, Inorg. Chem. 1992, 31, 4933–4938;
k) F. J. Feher, S. L. Gonzales, J. W. Ziller, Inorg. Chem. 1988,
27, 3440–3442; l) P. H. Bird, A. R. Fraser, C. F. Lau, Inorg.
Chem. 1973, 12, 1322–1328.
[15] T. Steiner, G. Koellner, J. Mol. Biol. 2001, 305, 535–557.
ˇ
[16] a) D. J. Miodragovic´, Z. J. Vitnik, S. M. Milosavljevic´, M. J.
Malinar, I. O. Juranic´, Eur. J. Inorg. Chem. 2005, 3172–3178;
b) H. Kumita, T. Kato, K. Jitsukawa, H. Einaga, H. Masuda,
Inorg. Chem. 2001, 40, 3936–3942.
[17] R. C. Fay, Coord. Chem. Rev. 1996, 154, 99–124.
[18] D. Felder, H. Nierengarten, J.-P. Gisselbrecht, C. Boudon, E.
Leize, J.-F. Nicoud, M. Gross, A. Van Dorsselaer, J.-F. Nieren-
garten, New J. Chem. 2000, 24, 687–695.
[19] I. Nakagawa, T. Shimanouchi, Spectrochim. Acta 1964, 20,
429–439.
[20] C. Diebold, P. Mobian, C. Huguenard, L. Allouche, M. Henry,
Inorg. Chem. 2010, 49, 6369–6371.
[21] C. Diebold, D. M. Weekes, M. T. Navarrete, P. Mobian, N. Ky-
ritsakas, M. Henry, Eur. J. Org. Chem. 2010, 6949–6956.
[22] M. J. Plater, S. Kemp, E. Lattmann, J. Chem. Soc. Perkin Trans.
1 2000, 971–979.
Received: May 24, 2012
Published Online: November 2, 2012
Eur. J. Inorg. Chem. 2012, 5701–5713
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5713