Synthesis and characterisation of [M(η5-NC4Me4)(CH2Ph) 3] (M=Ti, Zr, Hf) and [Ti(η5-NC4Me4)(Me)Cl2]. Structural determination and bonding of [Ti(η5</su

Article abstract of DOI:10.1016/S0022-328X(01)00824-5

The complex [Ti(η⁵-NC₄Me₄)(Me)Cl₂] (1) has been synthesised and characterised. Based upon the X-ray structure and ab initio theoretical calculations performed on the model complex [Ti(η⁵-NC₄

Full text of DOI:10.1016/S0022-328X(01)00824-5

Journal of Organometallic Chemistry 632 (2001) 157–163
www.elsevier.com/locate/jorganchem
5
Synthesis and characterisation of [M(h -NC Me )(CH Ph) ]
4
4
2
3
5
(
M=Ti, Zr, Hf) and [Ti(h -NC Me )(Me)Cl ]. Structural
4
5
4
2
determination and bonding of [Ti(h -NC Me )(Me)Cl ] depicting
4
4
2
an agostic interaction
Alberto R. Dias, Adelino M. Galv a˜ o *, Ana Castro Galv a˜ o
Centro de Qu ´ı mica Estrutural, Complexo I, Instituto Superior T e´ cnico, A6. Ro6isco Pais 1, 1049-001 Lisbon, Portugal
Received 26 February 2001; accepted 4 April 2001
On the occasion of Professor Alberto R. Dias 60th birthday.
Abstract
The complex [Ti(h -NC Me )(Me)Cl ] (1) has been synthesised and characterised. Based upon the X-ray structure and ab initio
5
4
4
2
5
theoretical calculations performed on the model complex [Ti(h -NC H )(Me)Cl ] an agostic interaction has been postulated. The
4
4
2
5
synthesis and characterisation of [M(h -NC Me )(CH Ph) ] (M=Ti, 2; Zr, 3; Hf, 4) complexes are also described as well as the
4
4
2
3
possibility of similar agostic interactions, which are discussed from the NMR data and the comparison with analogue carbocyclic
compounds. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Group IV transition metals; Pyrrolyl; Agostic interaction; X-ray diffraction; Ab initio calculations
5
5
5
1
. Introduction
Cl ], [Ti(h -C H )(h -NC Me )Cl ], [Ti(h -NC Me )-
2
5
5
4
4
2
4
4
(
SPh) ] (n=1, 3)} [8], we continued the reactivity stud-
n
5
5
Organometallic complexes with h -cyclopentadienyl
Cp) ligands constitute a very broad area within transi-
ies on [Ti(h -NC Me )Cl ] by testing the substitution of
4
4
3
(
the chlorides with alkyl ligands, in an attempt to repro-
5
tion metal chemistry [1], performing some important
roles, namely as homogeneous catalysts in polymerisa-
tion reactions [2] and anticarcinogenic agents [3,4].
However, the isoelectronic pyrrolyl ligands are consid-
duce the behaviour of the Cp and Cp* (h -C Me )
5
5
5
analogues [9,10]. In fact, the complex [Ti(h -
NC Me )(Me)Cl ] was obtained as indicated above, but
4
4
2
it was not possible to achieve selective substitution of
two or three chloride ligands by the same synthetic
route. Therefore, following the good results in the
preparation of Ti [11], Zr [12] and Hf [12] amide
complexes by adding 2,3,4,5-tetramethylpyrrole to an
appropriate amide starting material, we used an identi-
cal strategy to produce the compounds [M(h -
NC Me )(CH Ph) ] (M=Ti, Zr, Hf).
Bearing in mind the general tendency of Group IV d
5
erably less prone to undergo h -co-ordination to transi-
tion metals, the instability of the correspondent
compounds often being attributed to the lower ionisa-
tion potential of the non-bonding electron pair of the
nitrogen atom when compared to that of the p-elec-
trons [5–7].
5
5
In the sequence of the prior synthesis of [Ti(h -
5
4
4
2
3
NC Me )Cl ] and some derivatives {[Ti(h -NC Me ) -
4
4
3
4
4 2
0
metal atoms to establish agostic interactions to s-hy-
drocarbyl fragments [13–17] and the particular evi-
5
dences in the [Ti(h -C Me )(CH Ph) ] analogue [13],
5
5
2
3
the possible agostic interactions in complexes 1–4 were
*
Corresponding author. Tel./fax: +351-1-218417279.
E-mail address: adelino@alfa.ist.utl.pt (A.M. Galv a˜ o).
investigated.
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(01)00824-5

158
A.R. Dias et al. / Journal of Organometallic Chemistry 632 (2001) 157–163
2. Results and discussion
very similar to those of the pyrrolyl methyl groups in
compound 1 and the singlet at 1.27 ppm may certainly
result from three methyl ligands bonded to Ti.
On the other hand, the reaction of the trichloride
2
.1. Chemical studies
5
Following the traditional chemical path used for the
complex [Ti(h -NC Me )Cl ] with three molar equiva-
4
4
3
synthesis of alkyl–Cp–Ti complexes [9,13,18], the com-
lents of MeLi yields an unidentified black oily solid,
proving that the chloride substitution becomes less and
less selective as we increase the molar amount of the
alkylating agent.
5
pound [Ti(h -NC Me )(Me)Cl ] (1) has been prepared
4
4
5
2
by reaction of [Ti(h -NC Me )Cl ] with one molar
4
4
3
equivalent of MeLi (Eq. (1)).
toluene
After verifying the difficulty in the selective prepara-
5
[Ti(h -NC Me )Cl ]+LiMe ꢀꢀꢀꢁ
4
4
3
−80°C
tion of the tri-methylated compound [Ti(NC Me )Me ]
4 4 3
5
by the synthetic route indicated above and considering
the successful path for the synthesis of Ti [11], Zr [12]
and Hf [12] amide complexes by using an appropriate
amide compound as starting material, we followed a
[
Ti(h -NC Me )(Me)Cl ]+LiCl (1)
4
4
2
The orange crystals of 1 are sensitive to oxygen and
moisture and present an elemental analysis compatible
with the formulation [Ti(NC Me )(Me)Cl ], corre-
sponding to the substitution of a chloride for a methyl
ligand.
The H-NMR spectrum reveals three singlets (l 1.45,
.75, 2.10) with relative areas 2:1:2 resulting from the
three types of methyl groups in the molecule; the minor
area of the peak at 1.75 ppm makes it readily assignable
to the s-hydrocarbyl ligand, while the other two signals
correspond to the two types of pyrrolyl methyl groups.
5
4
4
2
similar strategy to obtain the set of complexes [M(h -
NC Me )(CH Ph) ] (M=Ti, 2; Zr, 3; Hf, 4). In fact,
4
4
2
3
the reaction of the tetrabenzyl compounds of the three
group IV transition metals with an equimolar amount
of 2,3,4,5-tetramethylpyrrole led to the substitution of
one benzyl ligand for the heterocyclic ring, as depicted
in Eq. (2).
1
1
toluene
5
[
M(CH Ph )]+HNC Me ꢀꢀꢀꢁ [M(h -NC Me )
2
4
4
4
4
4
1
3
In the
C-NMR spectrum, the three quartets
JCH:130 Hz) arise from the three types of methyl
(
CH Ph) ]+CH Ph
2 3 2
1
(
2
(M=Ti); 3 (M=Zr); 4 (M=Hf)
(2)
groups indicated above: 11.02 and 16.24 ppm for the
pyrrolyl substituents and 84.24 ppm for the alkyl lig-
and. The quaternary carbons of the heterocyclic ring
originate the two singlets at 147.40 and 136.44 ppm,
considerably deshielded relative to those of the pro-lig-
and HNC Me (l 120.20 and 113.80). This significant
5
The [M(h -NC Me )(CH Ph) ] compounds are very
4
4
2
3
sensitive to oxygen and moisture, preventing reliable
elemental analyses or IR data. On the other hand, none
of the several attempts to obtain crystals suitable for
X-ray diffraction was successful and therefore the char-
acterisation of 2–4 was made exclusively from the
NMR data, as well as the investigation of the pre-
sumable agostic interactions.
4
4
deshielding of the ring carbons suggests an active dona-
tion of electron density to the metal, making reasonable
5
the postulation of an h -co-ordination of the pyrrolyl,
1
which was confirmed a posteriori by the X-ray struc-
ture. Additionally, the deshielding values are similar to
The H-NMR spectra of compounds 2–4 exhibit in
each case two singlets with equivalent areas and chemi-
cal shifts in the range 1.32–1.82 ppm, assigned to the
two types of pyrrolylic methyl protons; the remaining
5
those reported for the compounds [Ti(h -NC Me )Cl ]
4
4
3
5
t
2
[
8] and [Ti(h -NC H Bu -2,5)Cl ] [19] where the hetero-
4
2
3
5
cyclic ring is h -co-ordinated to titanium.
singlet (1.90–2.94 ppm) arises from the CH groups of
2
5
Following the successful synthesis of [Ti(h -
the benzyl ligands and the equivalency of its area to
that of each pyrrolyl signal is consistent with the stoi-
chiometry [M(NC Me )(CH Ph) ]. Near the deuteriated
NC Me )(Me)Cl ] (1), the substitution reactions of two
4
4
2
and three chlorides for an equivalent number of methyl
groups were tested using the complex [Ti(h -
NC Me )Cl ] as starting material. However, a lack of
selectivity was verified when trying to reproduce the
reaction depicted in Eq. (1) with two and three molar
equivalents of the alkylating agent. In fact, the H-
NMR data from the dark green mixture obtained by
reaction of [Ti(h -NC Me )Cl ] and LiMe in the pro-
portion 1:2 strongly suggest that the substitution of all
the chloride ligands has occurred. Apart from 2,3,4,5-
tetramethylpyrrole signals and other impurities, we ob-
serve three singlets (l 1.27, 1.48, 2.07) with relative
areas 3:2:2 in the methyl protons range. The chemical
shifts of the peaks with the same area (l 1.48, 2.07) are
4
4
2
3
5
solvent broad band lie the peaks resulting from the
aromatic protons of the benzyl rings. In the Ti com-
pound 2 it was not possible to distinguish between the
several types of protons, which appear as multiplets in
the l range 6.99–7.17, with a 8 Hz coupling constant
which is typical of H–H interactions in six-membered
aromatic rings [20]. It is worth referring that the chem-
ical shifts exhibited by complex 2 (l 2.94 and 6.99–
7.17) are comparable to those reported for the parent
4
4
3
1
5
4
4
3
5
carbocyclic compound [Ti(h -C Me )(CH Ph) ] (l 2.73
5
5
2
3
and 6.84–7.36) [18].
For the Zr and Hf compounds, 3 and 4, it is possible
to differentiate the three types of aromatic protons: the

A.R. Dias et al. / Journal of Organometallic Chemistry 632 (2001) 157–163
159
5
doublet with J=7 Hz (l 6.68 for 3 and 6.85 for 4) with
the same area of the methylenic singlet results from the
resonance of the six ortho protons; the triplets at 6.99
although inferior to those of other h -pyrrolyl com-
pounds [8], are considerably superior to Zr and Hf
complexes with an average h -co-ordination of the
3
(
complex 3) and 6.92 ppm (compound 4) are assigned
pyrrolyl rings [12].
to the three para protons because of its area (half of the
Taking into consideration the well known tendency
of the d Group IV transition metals to establish in-
0
area of the CH peaks); finally the more deshielded
2
triplets (l 7.09 for 3 and 7.16 for 4), although partially
covered by the solvent band, show a similar area rela-
tive to the ortho protons signal, being compatible with
the presence of the six meta protons from the benzyl
rings.
tramolecular agostic interactions with alkyl ligands
(which was in fact elucidated for compound 1), we tried
to demonstrate this possibility in the set of complexes
5
[M(h -NC
Me
4
)(CH
Ph)
] (M=Ti, Zr, Hf).
4
2
3
As a possible way to investigate the existence of
agostic interactions, the H-NMR spectra of the Ti and
1
1
3
The C-NMR data of complexes 2–4 include (for
1
Zr complexes were recorded in d -toluene between
each one) two quartets with JCH=128–129 Hz and
8
room temperature and −80°C but no significant
changes were detected. This result does not exclude the
existence of agostic interactions, meaning that the
methylenic protons exhibit a fluxional behaviour in the
chemical shifts between 10 and 15 ppm, which are
assigned to the two types of pyrrolylic methyl sub-
stituents. On the other hand, in the range 70–100 ppm
1
3
lie the triplets of triplets ( J =124–125 Hz; JCH=4
CH
NMR experimental conditions. In fact, the same result
Hz) arising from the methylenic groups. Concerning the
tertiary carbons of the benzyl rings, it was possible to
make the individual assignments, mainly by analysing
the multiplicity of the signals and using the approxi-
mate area of the peaks as a clue. Therefore, the more
shielded doublets of triplets (:123 ppm) were at-
tributed to the para carbons, while the doublets of
doublets (128–129 ppm) correspond to the ortho car-
bons and the doublets of triplets between 127 and 129
ppm are assigned to the meta carbons. Finally, the
singlets in the range 130–150 ppm result from the
resonance of the three types of quaternary carbons
whose individual assignments were accomplished by
analogy with the similar Ti–Cp*–tri-benzyl complex
5
arises from the similar [Ti(h -C Me )(CH Ph) ] for
5
5
2
3
which a double agostic interaction has been proved
13]. Another consequence of the fluxionality are the
[
1
normal values for the J
coupling constants of the
CH
benzyl carbons (:124 Hz); this fact is also verified for
5
[
Ti(h -C Me )(CH Ph) ] [18], in contrast to the abnor-
5
5
2
3
5
mally low values of 105 and 113 Hz for [Ti(h -
C Me )(CH SiMe ) Cl] [18].
5
5
2
3 2
Despite the inconclusive results stated above, agostic
interactions in 2–4 can be postulated by the analysis of
the pyrrolyl ring carbons deshielding values. The pre-
sumed donation of at least one C–H benzylic bond to
the metal would make the heterocyclic ring donation
slightly weak, i.e. the deshielding would be lower than
for complexes without that type of interactions. This
hypothesis is confirmed by compound 1, showing
deshieldings of 27.2 and 22.6 ppm, considerably smaller
[
13,18] and using the area of the peaks as an indicative
factor.
Similarly to the characterisation of 1, the deshieldings
of the pyrrolyl quaternary carbons allowed us to postu-
5
to those reported for [Ti(h -NC Me )Cl ] [8]. Following
5
4
4
3
late an h -co-ordination of the heterocyclic ring in
this trend, the values of 24.2 and 16.5 ppm for com-
pound 2 strongly suggest the presence of at least one
agostic interaction. On the other hand, the parent Cp*
compounds 2–4, reproducing the structural tendency of
the Cp and Cp* analogues. In fact, the deshielding of
the ring carbons relative to the free pro-ligand (l 24.2
and 16.5 for 2; 17.6 and 15.5 for 3; 19.1 and 16.0 for 4),
5
compound [Ti(h -C Me )(CH Ph) ] behaviour [13,18],
5
5
2
3
as well as the reduction of pyrrolyl deshielding when
compared to 1 may suggest a double C–H ···Ti interac-
2
tion, helped by the lower donating power of the
pyrrolyl relative to Cp*.
For compounds 3 and 4 the postulated double agos-
tic interaction would be even more favourable due to
the better overlap of the C–H bonds to the expanded
Zr and Hf orbitals.
2.2. Structural studies
5
Confirming the h -co-ordination of the heterocyclic
13
ligand pointed out by the C-NMR data, compound 1
reveals a piano-stool configuration as we can observe in
the X-ray structure presented in Fig. 1, depicting a ring
slippage of the heteroatom towards the metal centre.
5
Fig. 1. X-ray structure of [Ti(h -NC H )(Me)Cl ] (1) with 40% ellip-
4
4
2
soids.

160
A.R. Dias et al. / Journal of Organometallic Chemistry 632 (2001) 157–163
Table 1
LANL2MB and B3LYP basis set were carried out with
Selected bond lengths (A), bond angles (°) and torsion angles (°) for
,
complex 1
GAUSSIAN-94/DFT [22]. Both methods predict the ex-
perimental structures with good accuracy (the experi-
mental value lying between the values given by each
method), namely the N–centroid–Ti slippage angle
,
Bond lengths
Ti–N
2.1882(2)
Ti–Cl (average)
C–Me (average) 1.494
Ti–centroid
Ti–H(1A)
Ti–H(1B)
Ti–H(1C)
2.239
Ti–C(1)
Ti–C(2)
Ti–C(3)
Ti–C(4)
Ti–C(5)
2.0771(1)
2.3105(1)
2.4375(1)
2.4362(1)
2.2778(2)
(
80.3 and 81.3°), average Ti–Cl (2.319 and 2.195 A),
2.008
2.28
2.61
2.56
Ti–Me (2.053 and 2.096 A
). The Ti–H bonds predicted by both ab initio meth-
,
) and Ti–N (2.177 and 2.192
A
,
ods are slightly longer then the experimental values
with the HF method predicting the closest results to the
X-ray values. The hydrogen trans opposite to the cen-
troid shows a shorter distance to the metal centre (2.47
Bond angles
Cl(1)–Ti–Cl(2)
Centroid–Ti–Me
Cl(1)–Ti–centroid
Cl(2)–Ti–centroid
105.57(6)
110
113
Cl(1)–Ti–Me
Cl(2)–Ti–Me
N–centroid–Ti
100.2(2)
100.8(2)
82
against 2.70 A of the other two hydrogens) and a
,
124
bending of the Ti–C–H (99° HF, 93° X-ray) when
compared to the near tetrahedral values of the other
two hydrogens (115° average in both HF and X-ray).
An Extended H u¨ ckel calculation performed in the opti-
mised geometry has shown a positive overlap popula-
tion between the 1s hydrogen orbital and the dxz (xz
being the centroid–Ti–C–H plane) on the metal centre,
confirming the agostic interaction.
Torsion angles
N–centroid–Ti–Cl(1)
N–centroid–Ti–Cl(2)
N–centroid–Ti–Me
195
325
84
Ti–C(1)–H(1A)
Ti–C(1)–H(1B)
Ti–C(1)–H(1C)
93
117
113
As reported previously [8] for similar complexes, this
structural trend can be quantified by the N–centroid–
Ti angle of 82°, Table 1, which is comparable to those
5
obtained for the similar compounds [Ti(h -NC Me )L ]
4
4
3
3
. Experimental
(
L=Cl, SPh) [8] (82–80°, respectively).
The projection of the four ligands in a plane perpen-
3
.1. Synthesis
dicular to the Ti–centroid axis shows the five-mem-
bered ring in an approximately staggered conformation
relative to the triangle formed by the two chlorides and
the methyl, characterised by the torsion angles N–cen-
troid–Ti–X (X=Cl, Me) of 84, 195 and 325°.
Apart from the bond and torsion angles, Table 1 also
presents the bond lengths for complex 1 and we can
All reactions and manipulations were carried out
under an atmosphere of argon, using standard Schlenk-
tube techniques. The NMR samples were prepared in a
Mbraun glove-box as well as the mounting of the
crystals, carried out in Lindemann glass capillaries. All
the solvents were dried with sodium and distilled over
Na/benzophenone, under nitrogen. Deuteriated ben-
zene and toluene were dried with molecular sieves and
conclude that the average Ti–Cl distance of 2.239 A
,
compares well with the indenyl analogue [21a] also
depicting a slightly longer Ti–Me bond length of 2.115
deoxygenated by several freeze–pump–thaw cycles.
[
21a] against 2.077 A
derived from X-ray diffraction data are not entirely
,
. Although the hydrogen positions
5
[
Ti(h -NC Me )Cl ] [8], [Ti(CH Ph) ] [23], HNC Me
4
4
3
2
4
4
4
[24] [Zr(CH Ph) ] [25] and [Hf(CH Ph) ] [25] were pre-
2 4 2 4
reliable, an agostic interaction can be postulated by the
pared as described. Methyllithium was obtained from
Aldrich and used without further purification.
shorter distance Ti–H1A of 2.28 A
.56 A of the other two hydrogens. The geometry of the
,
against 2.61 and
2
,
methyl group can also be compared with the original
work of Green et al. [21b] where neutron diffraction
data confirmed a similar agostic Ti–methyl bond with
3
.2. Analytical procedures
Microanalysis was performed by Laborat o´ rio de
d(Ti–C)=2.122 A
,
, d(Ti–Hagostic)=2.45 A
,
, d(Ti–H)=
An a´ lises of Instituto Superior T e´ cnico on a Fisons
Instruments 1108 spectrometer. Proton and carbon
NMR spectra were recorded in deuteriated benzene or
toluene-d8 on a Varian 300 MHz spectrometer and
referenced internally to the residual solvent resonance.
2
1
.80 and 2.74 A
18.4 and 112.9°.
Theoretical calculations were also used in the present
,
, Ti–C–Hagostic=93.5°, Ti–C–H=
work to investigate further the postulated agostic
interaction.
5
3.2.1. Preparation of [Ti(p -NC Me )(Me)Cl ] (1)
4
4
2
−
3
2.3. Calculations
A 1.51 mol dm
solution of methyllithium (0.77
3
cm , 1.16 mmol) in diethyl ether was added to a stirred
5
Using as a starting model the X-ray crystal structure
with pyrrolylic methyls replaced by hydrogens ab initio
calculations at HF and DFT levels using, respectively,
and cooled solution (−80°C) of [Ti(h -NC Me )Cl ]
(0.32 g, 1.17 mmol) in toluene (ca. 50 cm ). The mixture
was left stirring overnight and it was noticeable the
4
4
3
3

A.R. Dias et al. / Journal of Organometallic Chemistry 632 (2001) 157–163
161
1
changing of the initial orange colour to dark yellow.
After filtering, the solution was evaporated to dryness,
to give a yellowish–brown oil which was extracted with
n-hexane. The resulting yellow–orange solution was
cooled to −80°C and orange crystals (0.4×0.3×0.3
H-NMR spectrum (benzene-d , r.t.), l (ppm): 1.50
6
(s, 6H, pyrrolyl CH ); 1.63 (s, 6H, pyrrolyl CH ); 1.97
(s, 6H, benzyl CH ); 6.68 (d, 6H, [ J( H– H)=7],
benzyl ortho-H); 6.99 (t, 3H, [ J( H– H)=8], benzyl
para-H); 7.09 (t, 6H, [ J( H– H)=7], benzyl meta-H).
C-NMR spectrum (benzene-d , r.t.), l (ppm): 10.63
(q, [ J( C– H)=129], pyrrolyl CH3); 14.72 (q,
[ J( C– H)=129], pyrrolyl CH ); 70.15 (t, [ J( C–
3
3
1
1
1
2
1
1
1
1
1
1
3
13
mm ) were obtained after filtration and drying under
vacuum (yield 0.12 g, 42%). Anal. Found: C, 40.97; H,
6
1
13
1
1
13
1
1
13
5
.78; N, 5.25. Calc. for C H Cl NTi: C, 42.22; H, 5.91;
9
15
2
3
1
2
1
1
13
1
N, 5.47%.
H)=125], benzyl CH ); 123.75 (dt, [ J( C– H)=160;
2
1
13
1
2
13
H-NMR spectrum (benzene-d , r.t.), l (ppm): 1.45
J( C– H)=8], benzyl para-C); 128.76 (dt, [ J( C–
H)=6], benzyl meta-C); 129.30 (s, pyrrolyl ring C);
6
(
2
s, 6H, pyrrolyl CH ); 1.75 (s, 3H, CH bonded to Ti);
3
3
13
1
13
1
2
13
1
.10 (s, 6H, pyrrolyl CH3). C-NMR spectrum (ben-
129.68 (dd, [ J( C– H)=158; J( C– H)=7], benzyl
ortho-C); 137.80 (s, pyrrolyl ring C); 143.09 (s, benzyl
quaternary C). There were also performed H-NMR
1
13
1
zene-d , r.t.), l (ppm): 11.02 (q, [ J( C– H)=129],
pyrrolyl CH ); 16.24 (q, [ J( C– H)=129], pyrrolyl
CH ); 84.24 (q, [ J( C– H)=130], CH bonded to Ti);
6
1
13
1
1
3
1
13
1
spectra between r.t. and −80°C (in toluene-d ) and a
3
3
8
1
3
36.44 (s, pyrrolyl ring C); 147.40 (s, pyrrolyl ring C).
slight broadening of one of the pyrrolyl signals was
observed.
5
.2.2. Preparation of [Ti(p -NC Me )(CH Ph) ] (2)
4
4
2
3
5
A solution of 2,3,4,5-tetramethylpyrrole (0.15 g, 1.24
3.2.4. Preparation of [Hf(p -NC Me )(CH Ph) ] (4)
4
4
2
3
3
mmol) in toluene (ca. 30 cm ) was added to a stirred
and cooled solution (−70°C) of tetrabenzyltitanium
A solution of 2,3,4,5-tetramethylpyrrole (0.14 g, 1.10
3
mmol) in toluene (ca. 20 cm ) was added to a stirred
3
(
0.51 g, 1.24 mmol) in the same solvent (ca. 30 cm ).
and cooled solution (−60°C) of tetrabenzylhafnium
3
The mixture was left stirring for 3.5 days and was then
evaporated to dryness, to give a dark red oil which was
extracted with n-hexane. The resulting red solution was
slightly concentrated and then cooled to −80°C. Dark
red crystals unsuitable for X-ray diffraction analysis
were obtained after filtration and drying under vacuum
(0.59 g, 1.09 mmol) in the same solvent (ca. 15 cm ).
The mixture was left stirring overnight and was then
evaporated to dryness, to give a yellow–orange oil
which was extracted with n-hexane. The resulting or-
ange solution was cooled to −80°C but it was not
possible to obtain any solid.
1
(
2
yield 0.15 g, 27%). Anal. Found: C, 69.08; H, 6.02; N,
.89. Calc. for C H NTi: C, 78.54; H, 7.50; N, 3.16%.
H-NMR spectrum (benzene-d , r.t.), l (ppm): 1.42
6
(s, 6H, pyrrolyl CH ); 1.76 (s, 6H, pyrrolyl CH ); 1.90
2
9
33
3
3
1
1
1
1
H-NMR spectrum (benzene-d , r.t.), l (ppm): 1.32
(s, 6H, benzyl CH ); 6.85 (d, 6H, [ J( H– H)=7],
6
2
1
1
1
(
s, 6H, pyrrolyl CH ); 1.82 (s, 6H, pyrrolyl CH ); 2.94
benzyl ortho-H); 6.92 (t, 3H, [ J( H– H)=7], benzyl
para-H); 7.16 (t, 6H, [ J( H– H)=8], benzyl meta-H).
C-NMR spectrum (benzene-d , r.t.), l (ppm): 10.05
(q, [ J( C– H)=128], pyrrolyl CH3); 14.02 (q,
3
3
1
1
1
1
1
1
(
s, 6H, benzyl CH ); 6.99–7.17 (m, 15H, [ J( H– H)=
2
13
13
8
], benzyl aromatic H). C-NMR spectrum (benzene-
6
1
13
1
1
13
1
d , r.t.), l (ppm): 10.35 (q, [ J( C– H)=128], pyrrolyl
6
1
13
1
1
13
1
1
13
CH ); 14.79 (q, [ J( C– H)=128], pyrrolyl CH );
[ J( C– H)=128], pyrrolyl CH ); 83.38 (tt, [ J( C–
3
3
3
1
13
1
3
13
1
1
3
13
1
97.48 (tt, [ J( C– H)=124; J( C– H)=4], benzyl
H)=124; J( C– H)=4], benzyl CH ); 123.18 (dt,
2
1
13
1
2
13
1
1
13
1
2
13
1
CH ); 123.16 (dt, [ J( C– H)=162; J( C– H)=7],
[ J( C– H)=163; J( C– H)=8], benzyl para-C);
2
1
13
1
2
13
2
13
1
benzyl para-C); 127.60 (dt, [ J( C– H)=157; J( C–
128.44 (dm, [ J( C– H)=9], benzyl meta-C); 128.82
1
2
13
1
2
13
1
H)=6], benzyl meta-C); 128.51 (dd, [ J( C– H)=7],
(dd, J( C– H)=7], benzyl ortho-C); 129.81 (s,
pyrrolyl ring C); 139.32 (s, pyrrolyl ring C); 144.38 (s,
benzyl quaternary C).
benzyl ortho-C); 130.29 (s, pyrrolyl ring C); 144.45 (s,
pyrrolyl ring C); 148.96 (s, benzyl quaternary C). There
were also performed H-NMR spectra between r.t. and
1
−
80°C (in toluene-d ) but there were no noticeable
3.3. Crystallography
8
changes in the compound signals.
Complex 1 crystallises in the tetragonal system, space
5
3
.2.3. Preparation of [Zr(p -NC Me )(CH Ph) ] (3)
group I4 /a, with a=b=26.910(2) and c=6.7597(5)
4
4
2
3
1
3
−3
A solution of 2,3,4,5-tetramethylpyrrole (0.18 g, 1.46
mmol) in toluene (ca. 20 cm ) was added to a stirred
A
,
, V=4895.0(6) A
,
, Z=16, Dcalc=1.39 g cm
and
3
−1
v(Mo–K )=10.2 cm . 8571 reflections 1.5BqB25°
a
and cooled solution (−40°C) of tetrabenzylzirconium
were collected by the ꢀ–2q scan mode, in an Enraf-No-
nius MACH3 diffractometer using graphite monochro-
mated radiation. Two standard reflections were
monitored during data collection with no instrumental
instability detected. Data were corrected for linear de-
cay (average correction 1.07) and using the CAD4
software for Lorentz and polarisation effects and em-
3
(
0.67 g, 1.47 mmol) in the same solvent (ca. 20 cm ).
The mixture was left stirring overnight and was then
evaporated to dryness, to give a yellow–orange oily
solid, which was extracted with n-hexane. The resulting
orange solution was cooled to −80°C but it was not
possible to obtain any crystals.

162
A.R. Dias et al. / Journal of Organometallic Chemistry 632 (2001) 157–163
pirically for absorption (minimum transmission factor
2.3%, average transmission factor 96.5%). 2154 unique
In fact, agostic interactions between a Group IV unsat-
urated transition metal and a C–H bond are well
known as a means to minimise the metal centre unsatu-
ration acting simultaneously as powerful tool for C–H
activation.
9
2
reflections with F ]0 (R =0.04) were used in struc-
int
ture solution and refinement of 130 parameters. The
position of the Ti atom was obtained by a tridimen-
sional Patterson synthesis and all the other non-hydro-
gen atoms were located in subsequent difference
Fourier maps and refined with anisotropic thermal
motion parameters. With exception of the co-ordinated
methyl group the hydrogen atoms were inserted in
calculated positions and refined isotropically with fixed
distances to the parent carbon atom. The co-ordinated
methyl hydrogens were allowed to refine isotropically
Acknowledgements
A.C.G. acknowledges FCT for funding.
References
without restrains. Final refinement converged at R =
1
0
.04 and wR =0.11. The largest peak in the final
[1] C. Elschenbroich, A. Salzer, Organometallics, a Concise Intro-
2
−
3
duction, 2nd edn., VCH, New York, 1992, pp. 315–342.
difference Fourier synthesis was 0.35 e A
,
. The
[
[
2] M. Bochmann, J. Chem. Soc. Dalton Trans. (1996) 255.
3] T.M. Klap o¨ tke, H. K o¨ pf, I.C. Tornieporth-Oetting, P.S. White,
Organometallics 13 (1994) 3628.
molecular structure of complex 1 is shown in Fig. 1 and
lists of observed and calculated structure factors, tables
of anisotropic thermal parameters, atomic co-ordinates,
bond lengths and angles and inter and intra molecular
contact distances are available as supplementary mate-
rial. The structure solution and refinement were per-
formed with SHELX-86 [27] and SHELXL-93 [28],
respectively, and the illustrations were drawn with OR-
TEP-III [26]. The atomic scattering factors and anoma-
lous scattering terms were taken from International
Tables [29].
[4] P. K o¨ pf-Maier, in: B.K. Keppler (Ed.), Complexes in Cancer
Chemotherapy, VCH, New York, 1993, pp. 259–296.
[
5] K.H. Pannell, B.L. Kalsotra, C. Parkanyi, J. Heterocycl. Chem.
5 (1978) 1057.
1
[
[
6] D. L: Kershner, F. Basolo, Coord. Chem. Rev. 79 (1987) 279.
7] K. Y u¨ nl u¨ , F. Basolo, A.L. Rheingold, J. Organomet. Chem. 330
(1987) 221.
[8] A.R. Dias, A.M. Galv a˜ o, A.C. Galv a˜ o, M.S. Salema, J. Chem.
Soc. Dalton Trans. (1997) 1055.
[
9] R. Poli, Chem. Rev. 91 (1991) 509 and references therein.
[
[
10] G.J. Erskine, G.J.B. Hurst, E.L. Weinberg, B.K. Hunter, J.D.
McCowan, J. Organomet. Chem. 267 (1984) 265.
11] A.R. Dias, A.M. Galv a˜ o, A.C. Galv a˜ o, Collect. Czech. Chem.
Commun. 63 (1998) 182.
3.4. Calculations
[
[
12] A.R. Dias, A.M. Galv a˜ o, A.C. Galv a˜ o, Unpublished results.
13] M. Mena, M.A. Pellinghelli, P. Royo, R. Serrano, A. Tiripic-
chio, J. Chem. Soc. Chem. Commun. (1986) 118.
14] Z. Dawoodi, M.L.H. Green, V.S.B. Mtetwa, K. Prout, J. Chem.
Soc. Chem. Commun. (1982) 802.
GAUSSIAN-94/DFT [22] was used in all ab initio HF
and DFT calculations using respectively LANL2MB and
5
B3LYP basis set. An h -NC H co-ordinated ring was
4
4
[
used as a model to our more complex tetramethyl–
pyrrolyl ligand. The starting geometry was built from
X-ray structural data and fully optimised without any
symmetry restraints. All calculations were performed
on a 3400 Alfa Station running under OpenVMS.
[15] Z. Dawoodi, M.L.H. Green, V.S.B. Mtetwa, K. Prout, J. Chem.
Soc. Chem. Commun. (1982) 1410.
[
[
[
[
[
16] S. Obara, N. Koga, K. Morokuma, J. Organomet. Chem. 270
1984) C33.
17] N. Koga, S. Obara, K. Morokuma, J. Am. Chem. Soc. 106
1984) 4625.
(
(
18] M. Mena, P. Royo, R. Serrano, M.A. Pellinghelli, A. Tiripic-
chio, Organometallics 8 (1989) 476.
19] N. Kuhn, S. Stubenrauch, R. Boese, D. Bl a¨ ser, J. Organomet.
Chem. 440 (1992) 289.
20] W. Kemp, NMR in Chemistry, a Multinuclear Introduction, 3rd
edn., MacMillan Press, London, 1992, p. 215.
4
. Conclusions
The structural features of the complexes studied are
5
similar to those observed in other pyrrolyl h -deriva-
tives, namely the slippage of the heterocyclic ring to-
wards the metal centre. This particular behaviour
increases the reactivity of this type of compounds rela-
tive to Cp and Cp* analogues, making it comparable to
that of indenyl complexes.
[21] (a) S.L. Shaw, J.J. Storhoff, S. Cullison, C.E. Davis, G. Hol-
loway, R.J. Morris, J.C. Huffman, J.C. Bollinger, Inorg. Chim.
Acta 292 (1999) 220;
(
b) Z. Dawoodi, M.L.H. Green, V.S.B. Mtetwa, K. Prout, A.J.
Schultz, J.M. Williams, T.F. Koetzle, J. Chem. Soc. Dalton
Trans. (1986) 1629.
The X-ray structure of 1 has shown an agostic inter-
action between an a C–H bond and the titanium atom,
corroborated by ab initio calculations. The comparison
of the carbon atoms deshielding relative to the free
pro-ligand in complexes 2–4 with previously reported
data in carbocyclic analogues strongly indicates the
establishment of at least one of that type of interaction.
[
22] M.J. Frisch, G.W. Trucks, H.B. Schlegel, P.M.W. Gill, B.G.
Johnson, M.A. Robb, J.R. Cheeseman, T. Keith, G.A. Pe-
tersson, J.A. Montgomery, K. Raghavachari, M.A. Al-Laham,
V.G. Zakrzewski, J.V. Ortiz, J.B. Foresman, J. Cioslowski, B.B.
Stefanov, A. Nanayakkara, M. Challacombe, C.Y. Peng, P.Y.
Ayala, W. Chen, M.W. Wong, J.L. Andres, E.S. Replogle, R.
Gomperts, R.L. Martin, D.J. Fox, J.S. Binkley, D.J. Defrees, J.
Baker, J.P. Stewart, M. Head-Gordon, C. Gonzalez, and J.A.

A.R. Dias et al. / Journal of Organometallic Chemistry 632 (2001) 157–163
163
Pople, GAUSSIAN-94, Revision C.3, Gaussian Inc., Pittsburgh
PA, 1995.
23] U. Zucchini, E. Albizzati, U. Giannini, J. Organomet. Chem. 26
thermal ellipsoid plot program for crystal structure illustrations,
Oak Ridge National Laboratory Report ORNL-6895, 1996;
(b) P. McArdle, J. Appl. Crystallogr. 28 (1995) 65.
[27] G.M. Sheldrick, Acta Crystallogr. Sect. A 46 (1990) 467.
[28] G.M. Sheldrick, SHELX-93, crystallographic calculation program,
University of G o¨ ttingen, 1993.
[
(1971) 357.
[
[
24] A.W. Johnson, R. Price, Org. Synth. Coll. V (1973) 1022.
25] J.J. Felten, W.P. Anderson, J. Organomet. Chem. 36 (1972)8
7
.
[29] International Tables for X-ray Crystallography, vol. IV, Kynoch
Press, Birmingham, UK, 1974.
[
26] (a) M.N. Burnett and C.K. Johnson, ORTEP III: Oak Ridge
.