Syntheses, crystal structure and theoretical modelling of tetrahedral mono-β-diketonato titanocenyl complexes

Article abstract of DOI:10.1016/j.poly.2008.12.060

A comparative investigation on three novel bis(cyclopentadienyl) mono(β-diketonato) titanium(IV) complexes, [Cp₂Ti^IV(R¹COCHCOR²)] ⁺ClO₄^- (Cp = η⁵-C₅H_5<

Full text of DOI:10.1016/j.poly.2008.12.060

Polyhedron 28 (2009) 966–974
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Syntheses, crystal structure and theoretical modelling of tetrahedral
mono-b-diketonato titanocenyl complexes
*
Annemarie Kuhn, Alfred Muller, Jeanet Conradie
Department of Chemistry, University of the Orange Free State, Bloemfontein 9300, South Africa
a r t i c l e i n f o
a b s t r a c t
Article history:
A comparative investigation on three novel bis(cyclopentadienyl) mono(b-diketonato) titanium(IV) com-
ꢀ
plexes, [Cp2TiIV(R1COCHCOR2)]+ClO4 (Cp =
g
5-C5H5), i.e. [Cp2Ti(tfba)]+, [Cp2Ti(tfth)]+ and [Cp2Ti(tfba)]+
Received 22 October 2008
Accepted 26 December 2008
Available online 23 February 2009
where tfba = CF3COCHCOC6H5ꢀ, tfth = CF3COCHCOC4H3Sꢀ and tffu = CF3COCHCOC4H3Oꢀ, has been perꢀ-
formed based on structural data and DFT calculations. The preparation of [Cp2TiIV(b-diketonato)]+ClO4
involves the reaction of Cp2TiCl2 with AgClO4 and the respective b-diketones. The crystal structures show
Keywords:
Titanocene
b-Diketone
DFT
that the structures are isomorphous. All the complexes exhibit p-stacking between one Cp ring and the
aromatic R-group ring, i.e. the C6H5, C4H3S and C4H3O fragments, respectively. The DFT calculations show
that the formal 16-electron count of these d0 titanium(IV) complexes is increased via Ti O
p bonding.
The bonding mode in the [Cp2Ti(b-diketonato)]+ complexes is different from that in Cp2Ti(OR)2 and
Cp2Ti(dioxolene) complexes.
Crystal structure
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
showed that short Ti–O bonds and large Ti–O–R angles resulted
from in-plane Ti O
p bonding involving an orbital of al symmetry
Titanocene dichloride [Cp2TiIVCl2] (Cp =
g
5-C5H5) is a conve-
on the Ti atom [9]. Since the chemistry of the bent metallocene
fragment is mainly determined by the three low lying frontier orbi-
tals, 1a1, b2 and 2a1 of Cp2Ti2+, we found it worthwhile to investi-
gate both by experimental structure and computational chemical
calculations the way that the bidentate oxygen donor b-diketonato
ligand binds to the bent Cp2Ti2+ fragment.
nient starting material for the preparation of many other organo-
metallic compounds of titanium [1]. It has found application as a
catalyst or catalyst component for a wide variety of hydrometalla-
tion and carbometallation reactions, as well as for other transfor-
mations involving Grignard reagents. In addition, titanocene
dichloride [2] and budotitane, [TiIVꢀ(ba)2(OEt)2] [3] (ba = the b-dike-
In this study we report the synthesis and characterization by X-
ray crystallography of three new [Cp2TiIV(R1COCHCOR2)]+ com-
plexes with coordinated oxygen donor bidentate b-diketonato li-
gands. The need to further understand the titanium–oxygen bond
led to theoretical DFT studies on the structure of the present and
tonato ligand CH3COCHCOC6H5
) exhibit variable anti-tumor
activity for a variety of human tumors with reduced toxicity. Cat-
ionic [Cp2TiIV(b-diketonato)]+ complexes were first prepared by
Doyle and Tobias [4]. Although only one crystal structure of a tita-
nium(IV) compound containing the [Cp2Ti(b-diketonato)]+ cation,
related bis(g
5-cyclopentadienyl)titanium(IV) complexes.
[Cp2Ti(acac)][Ir3(l3-S)2(CO)6] [5], is known, structures of the anal-
ogous d1 titanium(III) compounds [Cp2TiIII(R1COCHCOR2)], with R1,
R2 = CH3 or C6H5 [6,7], have been reported. The coordination sphere
of these complexes does not change when switching the redox
state of the central coordinating metal from titanium(III) to tita-
nium(IV), but the change in electron density on the central coordi-
nating titanium should be reflected by changes in the bond lengths
of the bonds to titanium.
2. Results and discussion
2.1. Synthesis of complexes
The synthesis of the novel ꢀmono-b-diketonato titanium(IV)
salts [Cp2Ti(CF3COCHCOR)]+ClO4 with CF3COCHCORꢀ = CF3COCH-
ꢀ
COC6H5 (tfba) (1), CF3COCHCOC4H3Sꢀ (tfth) (2) and CF3COCH-
The electronic structure of the bent Cp2Ti2+ fragment had been
rationalized by Lauher and Hoffmann years ago [8]. Structure
COC4H3Oꢀ (tffu) (3) is based on an anion metathesis reaction
(which is driven by precipitation of one of the products), followed
by a b-diketone substitution reaction according to Scheme 1. An
important step is the solvation of the titanocene dichloride
(Cp2TiCl2) which hydrolyses in pure water with the displacement
of the two Clꢀ ligands, forming a wide range of cationic species
such as [Cp2Ti(H2O)(OH)]+, [Cp2Ti(H2O)(H2O)]2+ and [Cp2Ti(OH)2]
determinations
and
extended
Hückel
calculations
on
[Cp2TiIV(OR)2] complexes containing monodentate OR ligands
* Corresponding author. Tel.: +27 51 4012194; fax: +27 51 4446384.
E-mail address: conradj.sci@ufs.ac.za (J. Conradie).
0277-5387/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2008.12.060

A. Kuhn et al. / Polyhedron 28 (2009) 966–974
967
ClO4
2 Cl
2 ClO4
+
+
R
+
O
OH
O
O
OH2
OH
OH2
OH
Cl
Cl
R
CF3
2 AgClO4
H2O / THF
Ti
Ti
H
Ti
Ti
HClO4
+
-2AgCl
CF3
(and other aquated
species)
(and other aquated
species)
R = C6H5 (1), C4H3S (2), C4H3O (3)
Scheme 1. Synthesis of [Cp2Ti(CF3COCHCOR)]+ClO4 complexes.
ꢀ
[10]. Addition of silver perchlorate (AgClO4) removes the Clꢀ ions
by precipitating silver chloride (AgCl) while the hydrolysed per-
chlorate titanium salt, e.g. [Cp2Ti(H2O)(OH)]+ClO4ꢀ remains in solu-
tion. Finally, the b-diketone replaces the hydrolysed leaving
grouꢀps, forming an ionic species, [Cp2Ti(b-diketonato)]+ClO4ꢀ, with
ClO4 as the counter-ion. The problem of the insolubility of the b-
diketones in water was overcome by means of a mixed solvent sys-
tem, using a 1:2 ratio of H2O and THF. It was found that these salts
could also be synthesized in anhydrous THF, in which case, instead
of hydrolysis in water, a similar solvated species could be formed
by displacement of the two Clꢀ ligands in Cp2TiCl2 with OC4H4 in-
stead of OH2.
All the synthesized salts are insoluble in water, hexane and
ether, slightly soluble in organic solvents such as chloroform,
dichloromethane and are soluble in acetone, THF and ethanol.
However, dissolving the product in ethanol caused it to decom-
pose. The [Cp2Ti(b-diketonato)]+ cation is very sensitive to mois-
ture in solution, but has been found to be stable in the solid state.
0.014, 0.000 and 0.023 Å, respectively. This implies that the elec-
tron donating properties of the substituted non-CF3 groups Ph,
C4H3S, C4H3O (vR = 2.21 [12], 2.10 [13], unknown) of the b-diketo-
nato ligand do not show an observed effect on the Ti–O bond
length in the solid state compared to the stronger electron with-
drawing property of the CF3 group (vCF ¼ 3:01 [12]). The Ti–O
bond lengths are shorter in the TiIV com3plexes (average 1.996 Å)
compared to the TiIII analogues (average 2.076 Å), see Table 2,
and this indicates that the more electron deficient TiIV centre binds
oxygen more strongly than TiIII. The same trend is observed in the
Ti–Cp bonds: the Ti–(centroid Cp) bonds (average 2.036 Å) for the
TiIV complexes are ca. 0.03 Å shorter than those for the TiIII com-
plexes (average 2.062 Å).
The Cp rings of all the complexes tabulated in Table 2 are in a
staggered conformation. From an examination of the Cambridge
Structural Database (CSD) [14] it can be seen that the staggered
sandwich conformation in bent metallocene Cp2TiIV(O-ligand)2
and Cp2TiIV(O,O0-ligand) complexes is highly favored, although in
many cases it was found that the energy barrier to cyclopentadi-
enyl rotation is very small [15]. It has been shown that the choice
of conformation is attributed largely to crystal packing effects [16]
highlighted in the crystals of Cp2NbCl2 and Cp2MoCl2, where both
the staggered and eclipsed conformations are found in the same
crystal.
2.2. X-ray structure analysis
The molecular diagrams showing the atomic labelling, accom-
panied by diagrams showing the molecular packing of 1, 2 and 3,
respectively, are presented in Fig. 1. Crystal data and details of data
collection and refinement are summarized in Table 1. Selected
bond lengths and angles for the molecules are listed in Table 2.
Structural data for similar complexes are also included in Table 2.
Geometrical aspects: The crystal structures of complexes 1–3 re-
vealed a pseudo-tetrahedral structure with the centroids of the
cyclopentadienyl rings and the two oxygen atoms of the b-diketo-
nato ring in the ancillary positions of the central titanium(IV)
atom. It is evident, from examination of the bond angles (see Table
2), that the coordination tetrahedron is distorted. The (centroid
Cp1)–Ti–(centroid Cp2) angle is on average ca. 134° for complexes
1–3, which is a significant deviation from the ideal tetrahedral
bond angle of 109.5°. This large angle is attributed to steric hin-
drance between the Cp ligands. In contrast, the O1–Ti–O2 angle
is on average ca. 86°, which is considerably smaller than the stan-
dard 109.5° angle. This small angle is attributed to constraints in
accommodating the chelated b-diketonato ligand. The other angles
around the Ti atom, ranging from 103.4° to 108.4°, approach those
for an ideal tetrahedral geometry.
Typical C–C and C–O single and double bond lengths are: C–C
(1.54 Å), C@C (1.34 Å), C–O (1.43 Å), C@O (1.23 Å) [11]. The large
double bond character of the C–C and C–O bonds in the chelate
ring of the b-diketonato ligand (see Table 2) emphasize effective
electron delocalization in the pseudo-aromatic core of the b-diketo-
nato ligand. This electron delocalization enhances electronic com-
munication between the R-groups of the b-diketonato ligand and
the Ti centre. The b-diketonato ligand in [Cp2Ti(b-diketonato)]+
binds, within experimental error (taken as three times the bond
e.s.d.), symmetrically to the titanium core in 1 and 2 and asymmet-
rically in 3, with a difference between the two Ti–O bond lengths of
Crystallographic aspects: The structures 1–3 are isomorphous, i.e.
they crystallize in the same space group, P21/n, Z = 4, with unit cell
parameters varying slightly to accommodate the differences in the
substituted R-groups and are situated on fairly similar coordinates
within the unit cell. The refinements of [Cp2Ti(tfba)]+ (1) and
[Cp2Ti(tfth)]+ (2) showed large thermal vibrations on parts of the
periphery of the molecules, such as the Cp ring and the CF3 and
ꢀ
ClO4 groups. These large thermal vibrations may be attributed
to the fact that the collections were done at room temperature
and possibly also to loose packing environments in the periphery
of the molecules. The thermal vibrations were treated by disor-
dered refinement techniques, involving geometric and anisotropic
restraints to obtain more satisfactory refinement results. In the
data collection of 3, where sub-zero temperatures were utilized,
these disorders became less pronounced and were considered neg-
ligible in the final refinements. Site occupancies for both disor-
dered parts of 1 and 2 (indicated as A and B in Fig. 1) were
refined freely but restricted to add up to one. These refined to ra-
tios which are shown in Table 3.
Hydrogen-interactions, although relatively weak, also play a
part in site occupation ratios. There are H-interactions between
the atoms in the higher occupied site, e.g. the interactions C2–
H2ꢁꢁꢁF1A have higher occupation rates of 56.9% and 83.7% (position
A) for the disordered F atoms in [Cp2Ti(tfba)]+ (1) and [Cp2Ti(tfth)]+
(2), respectively. Similarly, H-interactions with perchlorate exist,
which lead to the higher occupation rates of 80.0% and 58.9% (po-
sition A) for the disordered O atoms in the perchlorate ion.
The fact that the Cp ring, marked C21 to C26 in the molecular dia-
gram of Fig. 1, is not disordered in all three cases, while the other

968
A. Kuhn et al. / Polyhedron 28 (2009) 966–974
Fig. 1. Molecular structure indicating the numbering scheme (with 30% probability displacement ellipsoids, left) and packing viewed along the a-axis (right) of (a)
[Cp2Ti(tfba)]+ClO4 and (b) [Cp2Ti(tfth)]+ClO4 at T = 20° and (c) [Cp2Ti(tffu)]+ClO4 at T = ꢀ123°. Monomeric cationic entities in the crystal are linked by
p–p stacking
ꢀ
ꢀ
ꢀ
interactions (indicated by double-headed arrows) forming dimeric units (perchlorate ions have been removed for the sake of clarity).
Cp ring (marked C11 to C16) is disordered in [Cp2Ti(tfba)]+ (1) and
[Cp2Ti(tfth)]+ (2) may be attributed to the fact that the non-disor-
planes, compared to 7.4° and 11.0° in the corresponding complexes
3 and 1, respectively. Geometrical parameters describing the
p–p
dered Cp rings are involved in the observed
below, forcing this ring (C21 to C26) into a more restricted packing
environment.
p
-stacking as described
stacking [17] are given in Fig. 2. In all three cases (tfba, tfth and
tffu), the rings are displaced from each other, in a slipped or off-
set alignment. The centroid–centroid separation, Rcen, of 3.724,
3.590 and 3.667 Å and the displacement (centre-normal) angle, h,
of 14.4°, 10.4° and 12.6° corresponds to a horizontal displacement
of the ring centroids, d, of 0.926, 0.648 and 0.800 Å, respectively.
The shortest C–H separations between the rings are 3.264, 3.486
and 3.217 Å. These parameters fall well within those required for
Packing features. The isomorphous complexes 1–3 have very
similar packing arrangements, with
p–p stacking between one
Cp ring and the R-group ring of the b-diketonato ligand, i.e. the
C6H5, C4H3S and C4H3O fragments of tfba, tfth and tffu, respec-
tively. The
plexes, forming dimeric units, see Fig. 1 (right).
The planes of the stacked rings are nearly parallel in
[Cp2Ti(tfth)]+ (2), with only a small deviation of 0.9° between the
p–p stacking interactions are between pairs of Ti com-
effective
pears to be unique to the TiIV bis(cyclopentadienyl) mono(b-dike-
tonato) complexes because the analogues, i.e.
TiIII
p–p stacking [18]. The packing pattern observed here ap-
p–p

A. Kuhn et al. / Polyhedron 28 (2009) 966–974
969
Table 1
Crystal data and structure refinement for [Cp2Ti(CF3COCHCOR)]+ClO4 with R = C6H5, C4H3S and C4H3O.
ꢀ
[Cp2Ti(tfba)]+ClO4
[Cp2Ti(tfth)]+ClO4
[Cp2Ti(tffu)]+ClO4
ꢀ
ꢀ
ꢀ
Formula
TiO6C20H16ClF3
492.68
red, cubic
monoclinic
P21/n
TiO6C18H14ClF3S
498.70
red, plate
monoclinic
P21/n
TiO7C18H14ClF3
482.64
red, plate
monoclinic
P21/n
Formula weight
Crystal colour/habit
Crystal system
Space group
Unit cell dimension
a (Å)
b (Å)
c (Å)
10.088(2)
16.613(3)
12.443(3)
90
96.86(3)
90
2060.4(7)
4
1.588
20
0.71073
0.607
9.872(2)
15.346(3)
13.134(3)
90
97.46(3)
90
1973.1(7)
4
1.679
20
10.0303(5)
16.0338(9)
11.6124(6)
90
97.046(2)
90
1853.45(17)
4
1.730
a
(°)
b (°)
(°)
c
Volume (Å3)
Z
Dcalc (Mg mꢀ3
)
Temperature (°C)
Wavelength (Å)
ꢀ123
0.71073
0.737
1008
0.71073
0.677
Absorption coefficient (mmꢀ1
F(000)
)
1000
976
Crystal size (mm3)
0.12 ꢂ 0.12 ꢂ 0.12
0.20 ꢂ 0.14 ꢂ 0.06
0.27 ꢂ 0.11 ꢂ 0.07
Theta range for data collection (°)
2.06–25.00
2.05–28.30
2.41–28.32
Index ranges
ꢀ11 6 h 6 11, ꢀ11 6 k 6 19,
ꢀ14 6 l 6 14
ꢀ13 6 h 6 9, ꢀ20 6 k 6 18,
ꢀ17 6 l 6 13
ꢀ13 6 h 6 13, ꢀ21 6 k 6 20,
ꢀ15 6 l 6 15
Reflections collected
11213
13654
17806
Independent reflections [Rint
Completeness to h = 28.35 (%)
Absorption correction
Max. and min. transmission
Refinement method
]
3622 [0.1037]
100.0
4871 [0.0727]
99.3
semi-empirical from equivalents
0.9571 and 0.8666
full-matrix least-squares on F2
4871/319/383
4606 [0.0316]
99.4
semi-empirical from equivalents
0.9542 and 0.8384
full-matrix least-squares on F2
4606/0/271
semi-empirical from equivalents
0.9307 and 0.9307
full-matrix least-squares on F2
3622/367/391
1.010
Data/restraints/parameters
Goodness-of-fit on F2
0.969
1.046
Final R indices [I > 2
R indices (all data)
r
(I)]
R1 = 0.0601, wR2 = 0.1294
R1 = 0.1852, wR2 = 0.1812
0.355 and ꢀ0.385
R1 = 0.0517, wR2 = 0.1166
R1 = 0.1628, wR2 = 0.1555
0.300 and ꢀ0.427
R1 = 0.0339, wR2 = 0.0787
R1 = 0.0465, wR2 = 0.0858
0.586 and ꢀ0.364
Largest difference in peak and hole
(e Åꢀ3
)
Table 2
Selected bond lengths (Å) and angles (°) for [Cp2TiIV(b-diketonato)]+ and Cp2TiIII(b-diketonato) complexes.
b-Diketonato ligand
[Cp2TiIVb]+
Cp2TiIII
b
tfba
tfth
tffu
acac5a
acac6a
ba7a
dbm7a
Ti–O1
Ti–O1
Ti–O2
O1–C1
O2–C3
C1–C2
C2–C3
1.979(4)
1.979(4)
1.993(5)
1.281(7)
1.283(8)
1.389((9)
1.350(9)
2.007(13)
2.059(15)
2.023(3)
2.705
85.91(19)
136.3
120.7(6)
124.7(6)
127.1(6)
46.6
1.996(3)
1.996(3)
1.996(3)
1.277(4)
1.283(5)
1.408(6)
1.359(6)
2.000(2)
2.033(2)
2.037(2)
2.720
85.98(12)
133.4
121.8(4)
123.0(4)
127.8(4)
48.4
1.9914(12)
1.9914(12)
2.0144(13)
1.283(2)
1.281(2)
1.411(3)
1.969
1.969
1.981
1.295
1.279
1.344
1.404
2.033
2.068
2.068
2.068
1.287
1.260
1.360
1.390
2.066
2.085
2.085
2.078
1.261
1.268
1.388
1.385
2.067
2.067
2.067
2.087
1.261
1.286
1.401
1.388
2.060
1.369(3)
2.0333(9)
Ct1–Tib,c
Ct1–Tib,c
Ct2–Tib
O1ꢁꢁꢁO2
O1–Ti–O2
Ct1–Ti–Ct2b
O1–C1–C2
C1–C2–C3
O2–C3–C2
ud
2.0333(8)
2.729
85.90(5)
133.3
123.06(16)
121.79(16)
128.15(17)
47.4
2.036
2.710
86.6
133.8
123.4
126.1
123.4
47.5
2.052
2.775
84.3
134.4
126.8
123.4
125.7
45.9
2.070
2.758
83.0
135.2
123.8
124.9
125
2.064
2.740
82.5
134.3
124.8
122.8
125.4
44.9
46.6
1.0
xe
7.9
19.4
19.9
16.8
4.6
13.3
a
ꢀ
acac, CH3COCHCOCH3ꢀ; ba, CH3COCHCOC6H5ꢀ; dbm, C6H5COCHCOC6H5
.
b
c
Ct1, centroid of plane defined by C11–C15; Ct2, centroid of plane defined by C21–C25.
Cp1 disordered, two positions.
d
e
u, angle between the two Cp planes.
x
, dihedral angle between the b-diketonato ligand plane (through atoms O1–C1–C2–C3–O2) and the O1–Ti–O2 plane.
[Cp2TiIII(CH3COCHCOC6H5)] [6] and [Cp2TiIII(C6H5COCHCOC6H5)]
2.3. Theoretical studies
[7], display no
p–p
stacking. These Cp2TiIII(b) complexes do not
ꢀ
have a counter-ion such as ClO4 in the case of TiIV which might
Density functional theory (DFT) calculations were carried out
influence the observed packing.
on complexes 1–3 and [Cp2TiIV(CH3COCHCOCH3)]+ (4) to study

970
A. Kuhn et al. / Polyhedron 28 (2009) 966–974
Table 3
calculated for non-hydrogen atoms for the best three-dimensional
superposition of calculated structures on experimental structures
give a qualitative measurement of the accuracy of the ground state
geometry of the calculated structures. Excellent agreement be-
tween experimental and theoretical structures is obtained, as re-
flected by the RMSD values of 0.20, 0.08, 0.07 and 0.04 Å for 1–4,
respectively. For 1 and 2, containing disordered parts, the parts
with the highest site occupancy were used in the RMSD calcula-
tion. The largest deviation from the experimental structure was
found for 1, containing a large % of disordered parts. Fig. 3 displays
the optimized structures of 1–4 with selected calculated geometri-
cal parameters. All the bonds in the Ti-b-diketonato ring structure
of 1–4 were reproduced by DFT calculations to within 0.01–0.03 Å
of the experimental values. Since comparisons of experimental me-
tal–ligand bond lengths with calculated bond lengths below a
threshold of 0.02 Å are considered as meaningless [19], the compu-
tational method used thus gives a good account of the experimen-
tal bond lengths. The O1–Ti–O2 angles were calculated accurately
to within 1°. The DFT optimized structures containing titanocene
with the Cp rings in the staggered or in the eclipsed conformation
are approximately equi-energetic, with no preference for either
conformation. This result is in agreement with the experimental
observation that the choice of conformation is attributed to crystal
packing effects and is counter-ion dependant [16].
Site occupancies (%) for disordered parts (marked A and B in Fig. 1) of the Cp1-ring and
the CF3 and ClO4 groups in [Cp2Ti(tfba)]+ClO4 (1) and [Cp2Ti(tfth)]+ClO4 (2).
Crystal data collected at T = 20°.
ꢀ
ꢀ
ꢀ
Complex
Disordered group
ꢀ
CF3
FA:FB
ClO4
Cp1
OA:OB
CA:CB
[Cp2Ti(tfba)]+ (1)
[Cp2Ti(tfth)]+ (2)
56.9:43.1
83.7:16.3
80.0:20.0
58.9:41.1
42.0:58.0
80.0:20.0
Surface
normal
Rcen
Ring centroid
θ
Rcd
ϕ
d
d = displacement of ring centroids
Rcen = centroid-centroid separation
R
cd = closest contact distance between rings
θ = centre-normal angle
ϕ =angle between π-stacked planes
Fig. 2. Definition of the
p–p stacking geometrical parameters, Rcen (centroid–
An interesting observation for experimental structure 2 is that
the sulfur on the aromatic thienyl group is orientated towards
the methine hydrogen of the b-diketonato ligand; this is in contrast
to the orientation in the uncoordinated Htfth ligand [20]. The same
result was found for the orientation of the O atom of the furyl
group in the crystal structure of 3 [21]. Theoretical calculations,
however, did not give any preference for the orientation of the thi-
enyl or furyl group in the optimized ground state structure of 2 or
3, respectively. The energy needed for rotation between the orien-
tations is 11.3 kcal/mol for 2, showing that this rotation may easily
occur at room temperature [22], see Fig. 4.
centroid separation), Rcd (closest contact distance) and h (centre-normal angle)
between the two aromatic rings involved in slipped or offset non-parallel alignment
of
p–p stacking. Note: The two benzene rings displayed in the figure can be any
aromatic ring, e.g. C6H5, C4H3O or C4H3O.
the ground state geometrical and electronic properties of the b-
diketonato interactions with the titanium centre in monomeric
complexes of the type [Cp2TiIV(R1COCHCOR2)]+. Since these density
functional methods are applied to b-diketonato titanocenyl com-
plexes for the first time and reported here, some measure on the
reliability of the approach was obtained by comparing the calcu-
lated data with the known single crystal X-ray diffraction struc-
tural data of 1–4. The root-mean-square distances (RMSD)
Electronic structure: A bent Cp2Ti2+ fragment has C2v symmetry if
the Cp ligands have an eclipsed geometry, and Cs symmetry if the
Fig. 3. The PW91/TZP calculated minimum energy geometries of 1–4. Angles (°) and bond lengths (Å) are as indicated.

A. Kuhn et al. / Polyhedron 28 (2009) 966–974
971
-5642
-5644
-5646
-5648
-5650
-5652
-5654
-5656
0
50
100
150
200
dihedral angle
Fig. 4. Rotation of 2 from the isomer with the S orientated towards the methine H
to the isomer with S orientated in the same direction as the O, as a function of the
O1–C1–C31–S dihedral angle where all other internal coordinates have been
optimized at each point.
rings are staggered. The Cs symmetry is preserved when a symmet-
rical b-diketonato ligand is complexed to Cp2Ti2+. The electronic
structure of [Cp2Ti(acac)]+ with Cs symmetry is thus presented here
as a representative example of the [Cp2Ti(b-diketonato)]+ com-
plexes 1–4. To understand the way in which the acacꢀ ligand binds,
the interaction between the frontier orbitals of a bent Cp2Ti2+ frag-
ment [8] and the HOMOs of the acacꢀ ligand is schematically illus-
trated in Fig. 5. The usual in-plane
r-type Ti–O interaction is
observed where the unoccupied b2 LUMO+1 of Cp2Ti2+ interacts
with the occupied b2 HOMO of acacꢀ (forming HOMOꢀ5 of
[Cp2Ti(acac)]+) and the unoccupied 2a1 LUMO+2 of Cp2Ti2+ inter-
acts with the occupied a1 HOMOꢀ2 of acacꢀ (forming HOMOꢀ6
of [Cp2Ti(acac)]+). The [Cp2Ti(acac)]+ complex is formally a 16-elec-
Fig. 5. A schematic molecular orbital (MO) diagram for the interaction between
Cp2Ti2+ and the acacꢀ unit in 4. The y-axis denotes the relative energy in eV for the
molecular orbitals of the [Cp2Ti(acac)]+ complex 4. The MO energy levels indicated
in blue do not have any Ti–O interaction.
tron d0 complex if only Ti–O
count is increased by Ti O
r
-bonding is considered. The electron
donation. A moderate to weak
p
p-
type interaction is observed in the HOMOꢀ2 and HOMOꢀ7 of
the [Cp2Ti(acac)]+ complex, see Fig. 5 for a schematic presentation
of the interaction and Fig. 6 for presentation of the orbitals in-
volved in Ti–O bonding. The above described Ti–O interactions re-
sult in Ti–O bonds of an average 1.99 Å for 1–4, slightly longer than
that for determined for Cp2TiIV(OR)2 complexes containing two
monodentate OR ligands, i.e. 1.96 Å for Cp2Ti(cyanoacetato)2 [23]
and 1.93 Å for Cp2Ti(benzoato)2 [9] and comparable to the Ti–O
bond length of 1.97 Å in [g
2-OC(Ph)@C(Ph)O]TiCp2, containing a
dioxolate bidentate ligand [24].
One interesting feature of the [Cp2Ti(b-diketonato)]+ structures
is the degree of out-of-plane folding of the b-diketonato ligand, as
schematically shown in Fig. 7. The folding of [Cp2Ti(b-diketonato)]+
along the OꢁꢁꢁO axis is
x = 7.9°, 19.4°, 19.9° and 16.8° for 1–4,
respectively. Similar folding is observed in the series of d0 penta
metallacyclic structures, i.e. Cp2Ti(dioxolene) [24], Cp2Ti(dithio-
lene) [25] and Cp2Ti(diselenolene) [26] complexes, with an
increasing degree of folding along the OꢁꢁꢁO, SꢁꢁꢁS and SeꢁꢁꢁSe axes
of 38°, 42–50° and 48–52°, respectively. It was rationalized by Lau-
her and Hoffmann [8] using extended Hückel (EH) calculations that
the folding in the Cp2Ti(dithiolene) complexes was due to a stabi-
lizing interaction between the empty acceptor LUMO 1a1 orbital of
the Cp2Ti2+ fragment and the HOMO b1 orbital of the dithiolene li-
gand, which was only possible by a symmetry lowering of the
Cp2Ti(dithiolene) complex to the C2 conformation and associated
folding of the TiS2C2 metallacycle. In this study we confirmed this
Ti L (L = O, S, Se)
p interaction (between the empty acceptor 1a1
LUMO of the Cp2Ti2+ and the b1 HOMO of the ligand) with DFT cal-
culations for the above pentacyclic series, Cp2(TiO2C2R2),
Fig. 6. Selected occupied MO’s of 4, illustrating the Ti–O
r and Ti O p interaction.

972
A. Kuhn et al. / Polyhedron 28 (2009) 966–974
R
R1
ϕ O
R
R
O
O
L
L
θ
θ
θ
Ti
Ti
Ti
Ti
ω
L
ϕO
R2
R
L = O, S, Se
Cp2TiIV(L,L−BID)
[Cp2TiIV(β)]+
e.g. Cp2TiIV(dioxolene)
Cp2TiIV(OR)2
Fig. 7. Folding round the LꢁꢁꢁL axis in the Cp2Ti(LL-BID) complexes (LL-BID = bidentate ligand with donor atoms L, L = O, S or Se) is described by the dihedral angle,
x, between
the cyclic ligand plane and the L–Ti–L plane. In the Cp2Ti(OR)2 complexes containing monodentate ligands OR, the large h and u angles describe the degree of bending along
Ti–O–R.
Fig. 8. The 1a1 type MO of the [Cp2Ti(b-diketonato)]+, Cp2Ti(dioxolene), Cp2Ti(dithiolene), Cp2Ti(diselenolene) and Cp2Ti(OR)2 complexes illustrating the 1a1–
p interaction.
Cp2(TiS2C2R2) and Cp2Ti(Se2C2R2) (R = CH3), see Fig. 8. However, no
interaction was observed for the hexacyclic
plexes exhibit p–p stacking between one Cp ring and the aromatic
such Ti O
p
R-group ring, i.e. the C6H5, C4H3S and C4H3O fragments, respec-
tively. DFT calculations show that [Cp2TiIV(b-diketonato)]+ com-
plexes behave differently to Cp2TiIV(OR)2 and Cp2TiIV(dioxolene)
complexes in that no significant bonding interaction between the
1a1 LUMO of the Cp2Ti2+ fragment and the coordinating ligand is
observed.
[Cp2Ti(b-diketonato)]+ complexes 1–4 where the 1a1 LUMO of
the Cp2Ti2+ fragment remains non-bonding in [Cp2Ti(acac)]+, see
Fig. 7. Thus we conclude that the folding in the [Cp2Ti(b-diketo-
nato)]+ complexes is not due to the stabilization of the same elec-
tronic interaction as in the Cp2Ti(dioxolene) [24], Cp2Ti(dithiolene)
[25] and Cp2Ti(diselenolene) [26] series.
Bending along Ti–O–R, stabilized by Ti O
p interactions, is ob-
4. Experimental
served in Cp2TiIV(OR)2 complexes. Extended Hückel (EH) calcula-
tions [9] revealed that the in-plane Ti O
p interaction
4.1. Materials and apparatus
involving the 1a1 LUMO of Cp2Ti2+ and the b1 HOMO of the ligand
was made possible by the large O–Ti–O (h = 89–91°) and Ti–O–R
Solid reagents (Merck, Aldrich and Sigma) were used without
further purification. Liquid reactants and solvents were distilled
prior to use using standard purification techniques [28] and water
was double distilled. Melting points were determined with a Reic-
hert Thermopan microscope equipped with a Kofler hot-stage and
are uncorrected. NMR measurements were recorded on a Bruker
Advance DPX 300 NMR spectrometer. Chemical shifts are reported
as d values relative to SiMe4 (0 ppm). IR spectra were recorded
from neat samples on a Digilab FTS 2000 Fourier transform spec-
trometer utilizing a He–Ne laser at 632.6 nm.
(u = 143–148°) angles [9,23, 27]. The in-plane Ti O
p interaction
was reproduced with DFT in the present study for Cp2TiIV(OCH3)2
and is illustrated in Fig. 8. The strain of the b-diketonato ring struc-
ture in [Cp2Ti(b-diketonato)]+ complexes with a O–Ti–O angle of
ca. 86°, however, makes large O–Ti–O and Ti–O–C angles impossi-
ble. Thus stabilization through in-plane Ti O
p interaction is thus
geometrically not possible for [Cp2TiIV(b-diketonato)]+ complexes.
The potential energy as a function of the folding along the OꢁꢁꢁO,
SꢁꢁꢁS and SeꢁꢁꢁSe axes, with all other internal coordinates being opti-
mized at each point, was calculated for the model complexes
[Cp2Ti(b-diketonato)]+, Cp2Ti(dioxolene), Cp2Ti(dithiolene) and
Cp2Ti(diselenolene). A plot of potential energy as a function of fold-
ing indicates that it takes less than 1 kcal/mol to fold the [Cp2Ti(a-
cac)]+ complex from the minimum energy conformation, with a
folding angle of 17°, to planar. This energy is substantially in-
creased in the pentacyclic series, where the energy needed is 7.4,
4.2. Synthesis
Safety note: Perchlorate complexes are potentially explosive,
should be treated with caution and handled in small quantities.
–
4.2.1. Preparation of [Cp2Ti(tfba)]+ClO4 (1)
16.7
and
16.0 kcal/mol
for
the
Cp2Ti(O2C2(CH3)2),
Titanocene dichloride, Cp2TiCl2 (249.0 mg, 1.0 mmol) was dis-
solved in H2O/THF, 1:2 mixture (9 ml) and stirred under nitrogen
for ½ h. The solution was cooled down on an ice-bath before silver
perchlorate, AgClO4 (406 mg, 1.96 mmol), dissolved in water (2 ml)
was added. The cooled mixture was stirred for a further ½ h. Ag-
ClO4 is light sensitive and the reaction must be shielded from light.
Silver chloride, a white precipitate, was filtered off and washed
with H2O/THF (2 ml). Trifluorobenzoylacetone, Htfba (432 mg,
Cp2Ti(S2C2(CH3)2) and Cp2Ti(Se2C2(CH3)2) complexes, respectively.
3. Conclusions
New titanocenyl complexes of the type [Cp2Ti(CF3COCH-
ꢀ
COR)]+ClO4 with R = C6H5, C4H3S and C4H3O were synthesized
and characterized. Crystallographic studies showed that all com-

A. Kuhn et al. / Polyhedron 28 (2009) 966–974
973
2.0 mmol), dissolved in cold THF (2 ml, ca. 4 °C) was added drop-
wise to the orange filtrate while stirring, causing a slight colour
change. After ½ h of stirring the solution was removed and left
for 7 days to precipitate (turned dark brown/black). The dark sticky
precipitate was collected and washed with water (2 ꢂ 100 ml) and
diethyl ether (4 ꢂ 50 ml). Recrystallisation from DCM/hexane
afforded the pure product. Spectroscopically pure crystals suitable
for single X-ray analysis were obtained by recrystallisation from
DCM/hexane. Yield 51% (246 mg). M.p. = 200 °C. Colour: metallic
ions, CF3 moieties and those on the Cp rings were treated with var-
ious geometrical and vibrational restraints.
Atomic scattering factors were taken from the International Ta-
bles for Crystallography Volume C [37]. The molecular plot was
drawn using the DIAMOND program [38] with a 30% thermal envelope
probability for non-hydrogen atoms. Hydrogen atoms were drawn
as arbitrary sized spheres with a radius of 0.135 Å.
4.4. Quantum computational methods
brown.
m
CO = 1554 cmꢀ1. NMR: dH (300 MHz, acetone–d6)/ppm:
7.16 (s, 10H, 2 ꢂ C5H5), 7.55 (s, 1H, CH), 7.68 (t, 2H, C6H5), 7.86
(t, 1H, C6H5), 8.35 (d, 2H, C6H5). Anal. Calc. for TiC20H16O6ClF3: C,
48.75; H, 3.27. Found: C, 48.13; H, 3.22%.
Pure Density Functional Theory (DFT) calculations were carried
out using the Amsterdam Density Functional 2007 (ADF) program
system [39] with the PW91 (Perdew–Wang, 1991) exchange and
correlation functional [40]. The TZP (Triple f polarized) basis set,
a fine mesh for numerical integration (5.2 for geometry optimiza-
tions), a spin-restricted formalism and full geometry optimization
(gas-phase) with tight convergence criteria, as implemented in the
ADF 2007 program, were used. The accuracy of the computational
method was evaluated by comparing the root-mean-square devia-
tions (RMSD’s) between the optimized molecular structure and the
crystal structure, using the non-hydrogen atoms in the molecule.
RMSD values were calculated using the ‘‘RMS Compare Structures”
utility in CHEMCRAFT Version 1.5 [41]. Whether artificially generated
atomic coordinates or coordinates obtained from X-ray crystal data
were used in the input files, optimizations for each compound re-
sulted in the same optimized geometry. The optimized structures
were verified as a minimum through frequency calculations. Un-
less indicated, no symmetry limitations were imposed in the
calculations.
4.2.2. Preparation of [Cp2Ti(tfth)]+ClO4 (2)
–
Preparation as for 1, but replacing the trifluorobenzoylacetone
with trifluorothenoylacetone, Htfth (461 mg, 2.0 mmol). Yield
48% (235 mg). M.p. = 191 °C. Colour: dark brown. m .
CO = 1566 cmꢀ1
NMR: dH (300 MHz, acetone–d6)/ppm 7.13 (s, 10H, 2 ꢂ C5H5), 7.38
(s, 1H, CH), 7.50 (t, 1H, C4H3S), 8.42 (d, 1H, C4H3S), 8.56 (d, 1H,
C4H3S). Anal. Calc. for TiC18H14SO6ClF3: C, 43.35; H, 2.83. Found:
C, 43.77; H, 2.87%.
–
4.2.3. Preparation of [Cp2Ti(tffu)]+ClO4 (3)
Preparation as for 1, but replacing the trifluorobenzoylacetone
with trifluorofuroylacetone, Htffu (412 mg/2.0 mmol). Yield 43%
(203 mg). M.p. = 195 °C. Colour: brown.
m
CO = 1568 cmꢀ1. NMR:
dH (300 MHz, acetone–d6)/ppm: 7.00 (t, 1H, C4H3O), 7.13 (s, 10H,
2 ꢂ C5H5), 7.18 (s, 1H, CH), 8.07 (d, 1H, C4H3O), 8.28 (d, 1H,
C4H3O). Anal. Calc. for TiC18H14O7ClF3: C, 44.79; H, 2.92. Found:
C, 44.71; H, 2.87%.
Acknowledgements
Financial assistance from the South African National Research
Foundation under Grant No. 61093 and the Central Research Fund
of the University of the Free State is gratefully acknowledged.
4.3. X-ray crystal structure determination
The X-ray intensity data for 1 and 2 were measured on a Bruker
SMART 1K CCD area detector and that for 3 on a Bruker X8 Apex II
4K CCD area detector. Both instruments were equipped with a
Appendix A. Supplementary data
graphite monochromator and a Mo K
a fine-focus sealed tube
CCDC 693411, 693410 and 6934 contains the supplementary
(k = 0.71073 Å) operated at 1.5 kW power (50 kV, 30 mA). The
detector was placed at a distance of 4.00 cm from the crystal. Data
collection temperature for 1 and 2 was room temperature (293 K),
whereas the temperature during the collection of 3 was 150(2) K,
using an Oxford 700 series cryostream cooler.
The initial unit cell and data collection of 3 were achieved by
the APEX2 software [29] utilizing COSMO [30] for optimum collection
of more than a hemisphere of reciprocal space. A total of 948
ꢀ
crystallographic data for [(C5H5)2Ti(CF3COCHCOR)]+ClO4 with
R = C6H5 (1), C4H3S (2) and C4H3O (3), respectively. These data
can be obtained free of charge via http://www.ccdc.cam.ac.uk/con-
ts/retrieving.html, or from the Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: (+44) 1223-
336-033; or e-mail: deposit@ccdc.cam.ac.uk. Supplementary data
associated with this article can be found, in the online version, at
doi:10.1016/j.poly.2008.12.060.
frames were collected with a scan width of 0.5° in u and
an exposure time of 20 s per frame. The initial unit cell and data
collection for 1 and 2 were achieved by the SMART NT software
x with
References
-
[31], set up for data collection of more that a hemisphere of reci-
procal space. A total of 1350 frames were collected with a scan
width of 0.3° in u and exposure times of 10 s per frame. The frames
were integrated using a narrow frame integration algorithm and
reduced with the Bruker SAINTPLUS and XPREP software [32] packages,
respectively. Analysis of all of the data collections showed no sig-
nificant decay during the data collection. Data were corrected for
absorption effects using the multi-scan technique SADABS [33].
All the structures were solved by the direct methods package
SIR97 [34] and refined using the WINGX software package [35] incor-
porating SHELXL [36]. The largest peaks on the final difference elec-
tron densities and the deepest holes were all within 1 Å from non-
hydrogen atoms and presented no physical meaning in the final
refinements. Aromatic protons were placed in geometrically ideal-
ized positions (C–H = 0.93–0.98 Å) and constrained to ride on their
parent atoms with Uiso(H) = 1.2Ueq(C). Disorders on the perchlorate
[1] N.J. Long, Metallocenes: An Introduction to Sandwich Complexes, Blackwell
Science, London, 1998. p. 23.
[2] P. Köpf-Maier, Eur. J. Clin. Pharmacol. 47 (1994) 1.
[3] B.K. Keppler, C. Friesen, H.G. Moritz, H. Vongerichten, E. Vogel, Struct. Bonding
78 (1991) 97.
[4] (a) G. Doyle, R.S. Tobias, Inorg. Chem. 6 (1967) 1111;
(b) G. Doyle, R.S. Tobias, Inorg. Chem. 7 (1968) 2484.
[5] M. Casado, J.J. Perez-Torrente, M.A. Ciriano, A.J. Edwards, F.J. Lahoz, L.A. Oro,
Organometallics 18 (1999) 5299 (CSD Ref.: LIHDUB).
}
[6] A.M. Bond, R. Colton, U. Englert, H. Hugel, F. Marken, Inorg. Chim. Acta 235
(1995) 117 (CSD Ref.: HIDPOZ).
[7] S.S. Yun, I. Suh, E.H. Kim, S. Lee, J. Coord. Chem. 51 (2000) 219 (CSD Ref.:
MOYJEP, MOYJIT).
[8] J.W. Lauher, R. Hoffmann, J. Am. Chem. Soc. 98 (1976) 1729.
[9] D.M. Hoffman, N.D. Chester, R.C. Fay, Organometallics 2 (1983) 48.
[10] J.H. Toney, T.J. Marks, J. Am. Chem. Soc. 107 (1985) 947.
[11] R.C. Weast, Handbook of Chemistry and Physics, 63rd Ed., The Chemical
Rubber Co., Ohio, p. F180.
[12] W.C. Du Plessis, J.C. Erasmus, G.J. Lamprecht, J. Conradie, T.S. Cameron, M.A.S.
Aquino, J.C. Swarts, Can. J. Chem. 77 (1999) 378.

974
A. Kuhn et al. / Polyhedron 28 (2009) 966–974
[13] M.M. Conradie, A.J. Muller, J. Conradie, S. Afr. J. Chem. 61 (2008) 13.
[14] Cambridge Structural Database (CSD), Version 5.27, August 2007 update.
[15] (a) B.E. Mann, C.M. Spencer, B.F. Taylor, J. Chem. Soc., Dalton Trans. (1984)
2027;
(b) L.N. Mulay, A. Attalla, J. Am. Chem. Soc. 85 (1963) 702.
[16] C.K. Prout, G.B. Allison, L.T.J. Delbaere, E. Gore, Acta Crystallogr. B28 (1972)
3043.
[28] B.S. Furniss, A.J. Hannaford, P.W.G. Smith, A.R. Tatchell, Vogel’s Textbook of
Practical Organic Chemistry, 5th Ed., Longman Scientific and Technical, John
Wiley and Sons, New York, p. 395 (Chapter 4).
[29] APEX2, Version 1.0–27, Bruker AXS Inc., Madison, Wisconsin, USA, 2005.
[30] COSMO, Version 1.48, Bruker AXS Inc., Madison, Wisconsin, USA, 2003.
[31] Bruker SMART
1998.
-NT, Version 5.050, Bruker AXS Inc., Madison, Wisconsin, USA,
[17] G.B. McGaughey, M. Gagne, A.K. Rappe, J. Biol. Chem. 273 (1998)
15458.
[32] SAINTPLUS, Version 7.12 including XPREP, Bruker AXS Inc., Madison, Wisconsin,
USA, 2004.
[18] C. Janiak, J. Chem. Soc., Dalton Trans. (2000) 3885.
[33] SADABS, Version 2004/1, Bruker AXS Inc., Madison, Wisconsin, USA, 1998.
[34] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, C. Giacovazzo, A.
Guagliardi, A.G.G. Moliterni, G. Polidori, R. Spagna, J. Appl. Crystallogr. 32
(1999) 115.
[35] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837 (WINGX 1.70.01).
[36] G.M. Sheldrick, SHELXL97, Program for Crystal Structure Refinement, University
of Göttingen, Germany, 1997.
[19] W.J. Hehre,
A Guide to Molecular Mechanisms and Quantum Chemical
Calculations, Wavefunction Inc., 2003. p. 153.
[20] R.D.G. Jones, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 32
(1976) 1224 (CSD Ref.: THTFBD).
[21] R. Boese, M. Yu Antipin, D. Blaser, K.A. Lyssenko, J. Phys. Chem. 102 (1998)
8654 (CSD Ref.: LIWPIQ).
[22] J.S. Lomas, J. Phys. Org. Chem. 20 (2007) 221.
[23] D.A. Edwards, M.F. Mahon, T.J. Paget, Polyhedron 16 (1997) 25.
[24] L.C. Song, P.C. Liu, C. Han, Q.M. Hu, J. Org. Chem. 648 (2002) 119.
[25] F. Guyon, M. Fourmigue, P. Audebert, J. Amaudrut, Inorg. Chim. Acta 239
(1995) 117.
[26] O. Jeannin, M. Nomura, M. Fourmigué, J. Organomet. Chem. 692 (2007)
4113.
[27] M.A.A.de C.T. Carrondo, M.J. Calhorda, M.B. Hursthouse, Acta Crystallogr. C43
(1987) 880.
[37] International Tables for Crystallography, Kluwer Academic Publishers,
Dordrecht, The Netherlands, vol. C, 2002.
[38] K. Brandenburg, H. Putz, DIAMOND Release 3.1a, Crystal Impact GbR, Bonn,
Germany, 2005.
[39] G.T. Velde, F.M. Bickelhaupt, E.J. Baerends, C.F. Guerra, S.J.A. van Gisbergen, J.G.
Snijders, T.J. Ziegler, J. Comp. Chem. 22 (2001) 931.
[40] J.P. Perdew, J.A. Chevary, S.H. Vosko, K.A. Jackson, M.R. Perderson, D.J. Singh, C.
Fioihais, Phys. Rev. B46 (1992) 6671.
[41] G.A. Zhurko, D.A. Zhurko, CHEMCRAFT, Version 1.5 (build 282), 2007.