Synthesis and imido-group exchange reactions of tert-butylimidotitaniuni complexes

Article abstract of DOI:10.1039/a607735h

Treatment of TiCl₄ with Bu^tNH₂ (6 equivalents) followed by addition of bpy (2 equivalents, 4-/er/-butylpyridine) or pyridine (py) (2 or 4 equivalents) afforded the five- or six-co-ordinate /erf-butylimido complexes [ri(NBu

Full text of DOI:10.1039/a607735h

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis and imido-group exchange reactions of
tert-butylimidotitanium complexes†
Alexander J. Blake, Philip E. Collier, Simon C. Dunn, Wan-Sheung Li, Philip Mountford* and
Oleg V. Shishkin
Department of Chemistry, The University of Nottingham, Nottingham NG7 2RD, UK
Treatment of TiCl4 with ButNH2 (6 equivalents) followed by addition of bpy (2 equivalents, 4-tert-butylpyridine)
or pyridine (py) (2 or 4 equivalents) afforded the five- or six-co-ordinate tert-butylimido complexes
[Ti(NBut)Cl2(bpy)2] or [Ti(NBut)Cl2(py)n] (n = 2 or 3*) respectively in good yields. Reaction of [Ti(NBut)Cl2(py)3]
with RNH2 gave the corresponding arylimido derivatives [Ti(NR)Cl2(py)3] (R = Ph,* C6H4Me-4,* C6H4NO2-4,*
C6H3Me2-2,6 or C6H3Pri2-2,6). Addition of tmen or pmdien to [Ti(NBut)Cl2(py)n] (n = 2 or 3 respectively) gave
the corresponding bi- or tri-dentate amine complexes, [Ti(NBut)Cl2(tmen)] or [Ti(NBut)Cl2(pmdien)]*
(tmen = Me2NCH2CH2NMe2, pmdien = N,N,NЈ,NЉ,NЉ-pentamethyldiethylenetriamine). Complexes labelled *
have been crystallographically characterised.
The importance of imido complexes and their isolobal
analogues in industrial processes, organic synthesis and cata-
lysis is now very well established.1–4 Terminal imido titanium
complexes were only first fully characterised in 1990.5,6 A num-
ber of such complexes have since been isolated, most requiring
extractions, treatment with ca. 2 equivalents of bpy and stand-
ard work-up afforded orange-red [Ti(NBut)Cl2(bpy)2] 1 in 78%
yield on a 12 g scale of product. By similar procedures the
bis(pyridine) analogue [Ti(NBut)Cl2(py)2] 2 may be prepared in
64% yield on a 7 g scale. In contrast to the preparation of 1, the
isolation of spectroscopically pure 2 was occasionally hamp-
ered by co-ordination of residual ButNH2 to the Ti as indicated
by additional broad resonances at δ ca. 3.4 (NH2) and 1.5
(ButN) in the 1H NMR spectrum 15 (relative intensities indicated
that there was less than 1 equivalent of ButNH per Ti᎐NBut
a different synthetic strategy for introducing the Ti᎐NR moi-
᎐
ety.‡,1 We have recently developed a very simple synthetic
entry point to a range of imidotitanium complexes featuring a
variety of supporting ancillary ligands.7 The complexes are all
prepared by generally high-yielding substitution reactions of
[Ti(NBut)Cl2(bpy)2]7 (bpy = 4-tert-butylpyridine) or its bis- or
tris-(pyridine) homologue [Ti(NBut)Cl2(py)n] (n = 2 or 3, see
below) with the desired ancillary ligand(s) or their anion(s). The
derivatives accessible via this straightforward metathetical route
include tetraaza-macrocyclic, tris(pyrazolyl)borate, indenyl,
cyclopentadienyl and bis(cyclopentadienyl) imidotitanium
complexes.7–11 In this contribution we describe the multigram
syntheses of these versatile bis- and tris-(pyridine) tert-
butylimidotitanium synthons, together with their ligand-
exchange reactions with primary arylamines and bi- and tri-
dentate tertiary amines. The crystal structures of five of the new
complexes have been determined and allow a crystallographic
evaluation of the relative structural and trans influences of
the NBut, NPh, NC6H4Me-4 and NC6H4NO2-4 ligands in a
homologous series of d0, six-co-ordinate complexes. Part of this
work has been communicated.12
᎐
2
unit). However, NMR-tube experiments in CDCl3 established
that addition of 2–3 equivalents of pyridine to samples of 2
contaminated in this way readily effected displacement of
ButNH2.
Owing to the difficulties associated with reproducibly achiev-
ing good yields of pure complex 2, we modified the synthesis by
adding 4 equivalents of py (per Ti) to the first-formed orange
precipitate. This modification routinely allows near-
quantitative isolation of the spectroscopically pure, tris(pyrid-
ine) complex [Ti(NBut)Cl2(py)3] 3 on around a 35–40 g scale of
product. Analytically pure, single crystals of 3 suitable for X-
ray diffraction were obtained from a saturated pentane solution
and important bond lengths and angles are listed in Table 1.
Details of the structure determination and a thermal ellipsoid
plot were given in a preliminary communication.12
The structure of complex 3 confirms the trans arrangement
of the chloride and mer arrangement of the pyridine ligands
shown in Scheme 1. A comparison of the bonding parameters
with those of our other crystallographically characterised
tris(pyridine) complexes is presented later. Consistent with the
well known trans influence of multiply bonded ligands,2,16–19 the
pyridine ligand trans to the NBut group is less tightly bound
compared to the cis pyridine ligands [differences between Ti᎐N
(py) cis and trans bond lengths span 0.182(4) to 0.203(1) Å].
This trans labilisation is reflected in the solution NMR spectra
of 3 which are consistent with the solid-state structure and
show that the resonances for the trans pyridine occur at higher
field and are somewhat broader than those of the cis pyridine
ligands. The NMR line broadening is consistent with a dissoci-
ative equilibrium between 3 and (2 ϩ free pyridine). Indeed,
Results
The syntheses and proposed structures are summarised in
Scheme 1. All the new compounds are air- and moisture-
sensitive and have been fully characterised by spectroscopic
methods, combustion microanalysis (C, H and N) and X-ray
crystallography where possible. The characterising data are
listed in the Experimental section.
Treatment of a cold (Ϫ50 ЊC) solution of TiCl4 in dichloro-
methane solution with 6 equivalents of ButNH2 gave an orange
precipitate as has been described previously.13–15 Solvent
† Non-SI unit employed: D ≈ 3.33 × 10Ϫ30 C m.
‡ Although for ease of representation all titanium imido linkages are
1
ambient-temperature H NMR spectra of 3 with 2–3 equiv-
alents of additional pyridine show qualitatively that the trans
and ‘free’ pyridine are in fast exchange on the NMR time-scale
while the cis co-ordinated pyridine ligands do not show any
drawn ‘Ti᎐NR’, the formal titanium–imido nitrogen bond order in the
᎐
complexes [Ti(NR)Cl2Ln] described herein is probably best thought of
as three (pseudo-σ2 π4, triple bond) rather than as two.1
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558
1549

View Article Online
But
N
(i )
Ti
bpy
Cl
Cl
bpy
1
But
N
But
N
(ii )
Me2
N
(v )
TiCl4
Ti
Ti
Cl
Cl
py
Cl
Cl
py
N
Me2
2
11
But
N
But
N
Me2
N
(iii )
(vi )
py
Cl
Cl
Cl
Cl
py
Ti
Ti
N
Me
py
NMe2
3
12
Fig. 1 Thermal ellipsoid plot (30% probability) of [Ti(NPh)Cl2(py)3]
4. Hydrogen atoms and dichloromethane molecules of crystallisation
omitted for clarity
(iv )
R4
arylamine to a dichloromethane solution of complex 3 at am-
bient temperature affords generally good yields of the corre-
sponding arylimido derivatives [Ti(NR)Cl2(py)3] (R = Ph 4,
C6H4Me-4 5, C6H4NO2-4 6, C6H3Me2-2,6 7 or C6H3Pri2-2,6
8). Some of the products can be obtained in spectroscopically
pure form by simple removal of ButNH2 and dichloromethane.
Analytically pure samples may be obtained by recrystallisation.
The tris(pyridine) complexes 3–8 tend to lose the trans co-
ordinated pyridine ligand on repeated washing or handling in
solution. Indeed, prolonged exposure of 3 to a dynamic vac-
uum results in the near-quantitative formation of 2. The bis-
(pyridine) arylimido homologues [Ti(NC6H3RЈ2-2,6)Cl2(py)2]
(RЈ = Me 9 or Pri 10) may similarly be obtained.
Single crystals of [Ti(NPh)Cl2(py)3] 4, [Ti(NC6H4Me-4)Cl2-
(py)3] 5 and [Ti(NC6H4NO2-4)Cl2(py)3] 6 suitable for X-ray dif-
fraction studies were obtained from cold dichloromethane (4)
or by layering a dichloromethane solution with pentane or hex-
ane. A thermal ellipsoid plot of 4 is shown in Fig. 1. Important
bond lengths are listed in Table 1 and data collection and pro-
cessing parameters in Table 4. Details of the structure
determinations of 5 and 6 were given previously.12 The struc-
tures of 4–6 confirm the geometries shown in Scheme 1. A
comparison of their bonding parameters with those of [Ti-
(NBut)Cl2(py)3] 3 is presented later.
The solution NMR spectra of complexes 4–8 are consistent
with the solid-state geometries of 4–6 and suggest that the pyri-
dine ligand bound trans to the arylimido group is more weakly
bound than the cis co-ordinated pyridine ligands. This trans
labilisation is illustrated by the structures of 4–6 where the dif-
ference between cis- and trans-Ti᎐N (py) bond lengths spans
the range 0.181(4) to 0.201(4) Å. The 2,6-disubstituted aryl-
imido complexes 7 and 8 are stable in solution under Ar or N2
for at least several weeks. Solutions of the other arylimido
derivatives 4–6 give rise to unknown decomposition products
on standing at room temperature.
R2
R6
R2
R6
N
N
(vii )
py
Cl
Cl
py
Ti
py
Cl
Ti
Cl
py
py
9
R2 = R6 = Me
R2
H
R4
H
R6
H
10 R2 = R6 = Pri
4
5
6
7
8
H
Me
H
H
NO2
H
H
Me
Pri
Me
Pri
H
Scheme 1 Reagents and conditions: (i) ButNH2 (6 equivalents),
CH2Cl2, Ϫ50 ЊC for 15 min then room temperature (r.t.) for 5 h, bpy
(2.2 equivalents), yield 78%; (ii) ButNH2 (6.0 equivalents), CH2Cl2,
Ϫ50 ЊC, 15 min then r.t. for 2 h, py (2.3 equivalents), 64%; (iii) ButNH2
(6.3 equivalents), Ϫ50 ЊC for 15 min then r.t. for 2 h, py (4.1 equiva-
lents), 98%; (iv) aniline or substituted aniline (≈1 equivalent), CH2Cl2,
r.t., 1–48 h, 36–86%; (v) tmen (1.02 equivalents), CH2Cl2, r.t. 16 h,
>95%; (vi) pmdien (1 equivalent), Et2O, r.t. 3 h, 55%; (vii) 65 ЊC,
dynamic vacuum, 4–7 h, >95%
detectable broadening under these conditions. However, quali-
tative magnetisation-transfer (MT) experiments for the cis and
trans py ligands in pure 3 show that exchange between these
sites is significant on the MT time-scale at room temperature.
We also wished to prepare arylimido analogues of [Ti(NBut)-
Cl2(py)3] 3. However, attempts to produce clean samples of such
a complex by the addition of H2NC6H3Me2-2,6 to TiCl4 in
CH2Cl2 followed by treatment with pyridine (under conditions
similar to those used for the synthesis of 3) failed. The product
mixtures appeared to contain an arylimido product (see below)
but were always contaminated with free H2NC6H3Me2-2,6
which could neither be fully removed by recrystallisation nor by
washing with hexane. However, the ready availability of 3
prompted us to investigate this complex as a potential starting
material for the arylimido homologues.
The bis- and tris-(pyridine) tert-butylimido complexes 2 and
3 allow access to bi- and tri-dentate tertiary amine adducts.
Thus treatment of a dichloromethane solution of [Ti(NBut)-
Cl2(py)2]
2 with tmen (N,N,NЈ,NЈ-tetramethylethane-1,2-
diamine) gives essentially quantitative conversion to the previ-
ously described 15 yellow compound [Ti(NBut)Cl2(tmen)] 11
after removal of pyridine and dichloromethane. Addition of 1
equivalent of the tridentate amine pmdien (N,N,NЈ,NЉ,NЉ-
As shown in Scheme 1, addition of ca. 1 equivalent of
1550
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558

View Article Online
Table 1 Comparison of selected bond lengths (Å) and angles (Њ) for the complexes [Ti(NBut)Cl2(py)3] 3, [Ti(NPh)Cl2(py)3] 4, [Ti(NC6H4Me-
4)Cl2(py)3] 5 and [Ti(NC6H4NO2-4)Cl2(py)3] 6*
[Ti(NR)Cl2(py)3]; R =
But (3)
Ph (4)
C6H4Me-4 (5)
C6H4NO2-4(6)
Ti᎐N
᎐
1.705(3)
1.714(2)
1.705(4)
1.722(3)
1.706(3)
Ti᎐N (py cis)
Ti᎐N (py trans)
Ti᎐Cl
2.247(3), 2.249(3)
2.255(3), 2.258(3)
2.450(3)
2.225(3), 2.229(3)
2.410(3)
2.227(2)
2.428(4)
2.394(1)
1.378(6)
2.228(2)
2.428(3)
2.380(1)
1.359(4)
97.66(2)
180
2.440(3)
2.432(1), 2.436(1)
2.431(1), 2.452(1)
1.455(5)
2.413(1), 2.416(1)
1.382(4)
᎐N᎐C
᎐
1.446(5)
N᎐Ti᎐N (py cis)
᎐
95.32(12), 97.55(13)
94.39(13), 99.18(13)
176.23(14)
94.40(10), 92.68(10) 95.35(7)
N᎐Ti᎐N (py trans)
᎐
176.26(10)
180
175.96(13)
N᎐Ti᎐Cl
100.85(12), 94.84(12)
95.32(12), 98.32(12)
174.8(3)
96.38(9), 97.57(9)
177.5(2)
98.65(3)
180
95.33(5)
180
᎐
C᎐N᎐Ti
᎐
171.7(3)
* For compound 3 the two sets of figures (one below the other) correspond to the two crystallographically independent molecules; for 5 and 6 the
molecules lie on a crystallographic two-fold axis passing through the C᎐N᎐Ti᎐N (py trans) vector and so there is only one value of Ti᎐Cl, Ti᎐N (py),
᎐
etc. and N᎐Ti᎐N (py trans) must be 180Њ.
᎐
Table 2 Selected bond lengths (Å) and angles (Њ) for [Ti(NBut)Cl2(pm-
dien)] 12 with estimated standard deviations (e.s.d.s) in parentheses.
The two values for each parameter correspond to the two independent
molecules in the asymmetric unit
Ti(1)᎐N(1)
Ti(1)᎐N(2)
Ti(1)᎐N(3)
Ti(1)᎐N(4)
Ti(1)᎐Cl(1)
Ti(1)᎐Cl(2)
N(1)᎐C(1)
1.691(1)
2.595(3)
2.282(3)
2.373(3)
2.4261(11)
2.3608(12)
1.462(5)
1.696(3)
2.588(4)
2.274(3)
2.409(3)
2.3888(12)
2.3798(11)
1.455(5)
N(1)᎐Ti(1)᎐N(3)
N(1)᎐Ti(1)᎐N(4)
N(1)᎐Ti(1)᎐N(2)
N(1)᎐Ti(1)᎐Cl(1)
N(1)᎐Ti(1)᎐Cl(2)
N(3)᎐Ti(1)᎐N(4)
N(3)᎐Ti(1)᎐N(2)
N(3)᎐Ti(1)᎐Cl(1)
N(3)᎐Ti(1)᎐Cl(2)
N(4)᎐Ti(1)᎐Cl(1)
N(4)᎐Ti(1)᎐Cl(2)
N(4)᎐Ti(1)᎐N(2)
Cl(1)᎐Ti(1)᎐Cl(2)
Cl(1)᎐Ti(1)᎐N(2)
Cl(2)᎐Ti(1)᎐N(2)
C(1)᎐N(1)᎐Ti(1)
98.36(4)
92.05(4)
99.28(14)
93.26(14)
169.24(14)
91.84(12)
102.43(11)
76.92(12)
74.47(12)
93.92(9)
154.27(10)
170.12(9)
88.15(9)
93.79(12)
99.00(4)
168.99(14)
92.21(11)
103.47(12)
77.58(11)
74.22(12)
93.31(8)
155.03(10)
170.42(9)
89.56(9)
94.20(11)
97.78(4)
80.30(8)
85.65(9)
167.3(3)
Fig. 2 Thermal ellipsoid plot (30% probability) of one of the two
crystallographically independent molecules of [Ti(NBut)Cl2(pmdien)]
12. Hydrogen atoms omitted for clarity
80.03(9)
85.92(9)
164.8(3)
three types of methyl group carbon and two types of methylene
group carbon for the co-ordinated pmdien. Together with the
1H NMR spectrum, these solution data suggest: (i) that the
molecule adopts a static solution structure with, for example, a
mer- or a fac-pmdien ligand with the ‘central’ NMe nitrogen co-
ordinated trans to the tert-butylimido group; or (ii) that the
solid-state isomer is present in solution, but that rapid exchange
of the cis- and trans-co-ordinated (with respect to the NBut
ligand) NMe2 groups takes place on the NMR time-scale at
ambient temperature; or (iii) that a number of different solution-
phase structures is present and that these all rapidly inter-
convert on the NMR time-scale. To address these possibilities
pentamethyldiethylenetriamine) to a diethyl ether solution of
[Ti(NBut)Cl2(py)3] 3 followed by cooling at Ϫ25 ЊC for 24 h
afforded single crystals of orange [Ti(NBut)Cl2(pmdien)] 12 in
55% yield. A thermal ellipsoid plot of one of the two crystal-
lographically independent molecules of 12 in the asymmetric
unit is shown in Fig. 2, important bond lengths are listed in
Table 2 and data collection and processing parameters are given
in Table 4. The solid-state structure of 12 shows the pmdien lig-
and co-ordinated to the Ti(NBut)Cl2 fragment in a fac mode
with one of the terminal NMe2 nitrogen atoms [N(2) in Fig. 2]
bonded trans to the tert-butylimido group. Thus the structure
does not contain a molecular mirror plane and all five pmdien
methyl groups and all four methylene groups are chemically
inequivalent.
1
low-temperature H and 13C-{1H} NMR spectra of 12 were
recorded. On cooling a CD2Cl2 solution to 213 K the spectra
became very complicated and were consistent with the presence
of several different compounds. The 13C-{1H} DEPT (distor-
tionless enhancement by polarisation transfer) 135 NMR spec-
trum at 213 K showed nine different methylene group carbon
1
The ambient-temperature H and 13C-{1H} NMR solution
spectra of complex 12 are not consistent with the solid-state
structure. For example, the 13C-{1H} NMR spectrum shows only
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558
1551

View Article Online
resonances and twelve methyl group carbon resonances for
pmdien ligands. We conclude that several different isomers of
12 are present in solution and rapidly interconvert at ambient
temperature; the solid-state structure shown in Fig. 2 thus
corresponds to only one of the solution isomers.
derivatives of the type [Ti(NR)X2L2].1 We note that a 1963
report suggested that [Ti(NSiMe3)Cl2(py)2], obtained by treat-
ing [Ti{N(SiMe3)2}Cl3] with pyridine, has a proposed structure
analogous to those suggested here for 1 and 2.28
The reaction of TiCl4 with H2NC6H3Me2-2,6 followed by
addition of pyridine is not suitable for producing pure samples
of complex 7. We expect that attempts at preparing 4–6 and 8
using this route would also give disappointing outcomes. This
failure is apparently due to difficulties in removal of the excess
of aniline either by dynamic vacuum or recrystallisation.
Indeed, Winter and co-workers15 have previously shown that
H2NC6H3Pri2-2,6 can react with TiCl4 to give [Ti(NC6H3Pri2-
2,6)Cl2(H2NC6H3Pri2-2,6)2]. Similar difficulties in removing
excess of disubstituted anilines have been encountered for
related zirconium and hafnium arylimido systens.29
However, addition of 1 equivalent of RNH2 to complex 3
does give generally good yields of [Ti(NR)Cl2(py)3] (R = Ph 4,
C6H4Me-4 5, C6H4NO2-4 6, C6H3Me2-2,6 7 or C6H3Pri2-2,6 8).
Although this arylamine/tert-butylimido exchange strategy has
also recently found success in the synthesis of chromium,30
molybdenum,31,32 osmium 33 and iridium 34 arylimido deriv-
atives, our examples are the first from Group 4. Mechanisms for
the exchange reactions of the osmium and chromium systems
have been proposed by Gibson 30 and Bergman 33 and co-
workers. Factors contributing to the overall driving force for
these exchange reactions presumably include the greater bas-
icity of the imido nitrogen in the ButN ligand compared to that
in RN (R = aryl).
The complexes [Ti(NR)Cl2(py)3] 4–8 are the first examples of
six-co-ordinate titanium arylimido derivatives although the six-
co-ordinate zirconium arylimido congener [Zr(NC6H3Pri2-
2,6)Cl2(py)3] has recently been described.29 The 2,6-disubsti-
tuted arylimido derivatives 7 and 8 are the most stable in solu-
tion and are also more soluble than the un- or 4-substituted
arylimido homologues 4–6. Moreover, since 7 and 8 are readily
obtained in spectroscopically pure form and in good yield, they
are potentially highly valuable synthons for the development of
arylimidotitanium chemistry.
Substitution of the pyridine ligands in complexes 2 or 3 by
the bi- or tri-dentate tertiary alkylamines tmen or pmdien to
form 11 or 12 is straightforward, presumably with entropy
factors helping to drive the reactions. The tmen derivative,
[Ti(NBut)Cl2(tmen)] 11, probably first prepared by Nielson 14
but formulated as a binuclear complex with µ-NBut ligands, has
recently been structurally characterised as a mononuclear
species.15 The solution NMR behaviour of the pmdien deriv-
ative [Ti(NBut)Cl2(pmdien)] 12 is in contrast to that of the
tris(pyridine) complexes [Ti(NR)Cl2(py)3] (R = But or aryl). Its
fluxional nature may reflect weaker binding of the pmdien ali-
phatic amine nitrogen atoms to the metal centre (see below for
discussion of X-ray data). Indeed the tris(tert-butylamine)
complex [Ti(NBut)Cl2(ButNH2)3] also gives rise to complicated
1H NMR spectra possibly indicative of a mixture of geometric
isomers.15
Discussion
Syntheses and spectroscopic data
The reaction of TiCl4 with ButNH2 followed by treatment with
bpy or py is an efficient and highly convenient route to multi-
gram quantities of the tert-butylimidotitanium synthons [Ti-
(NBut)Cl2Ln] (n = 2, L = bpy 1 or py 2; n = 3, L = py 3). The
reactions of TiCl4 with primary amines were first described by
Cowdell and Fowles13 in 1960 but the identities of the various
products obtained from the reaction with ButNH2 were only
recently elucidated by Winter and co-workers.15 Depending on
the amount of ButNH2 available in the system three types of
orange/orange-red titanium tert-butylimido species were
obtained, namely mononuclear [Ti(NBut)Cl2(ButNH2)3] (mix-
ture of isomers), trinuclear [{Ti(µ-NBut)Cl2(ButNH2)2}3] or
hexanuclear [{Ti(µ3-NBut)Cl2}6]. Interconversion between the
three complexes is achieved by addition or removal of ButNH2.
We have not attempted to identify which (if any) of these
species is formed under our reaction conditions and it is likely
that any of them could react with bpy or py to give 1–3. For
example, Winter and co-workers15 have shown that treatment
of isolated trinuclear [{Ti(µ-NBut)Cl2(ButNH2)2}3] with O-
donor monodentate or N-donor bidentate compounds affords
the five-co-ordinate complexes [Ti(NBut)Cl2L2] (L = OPPh3 or
L2 = tmen or N,NЈ-diisopropylethane-1,2-diamine).
The similarity of the IR and NMR spectra of the three
tert-butylimido complexes 1–3 is consistent with monomeric,
terminal tert-butylimido structures for 1 and 2 (as opposed to
alternative, six-co-ordinate bi- or tri-nuclear formulations
[{TiCl2L2(µ-NBut)}x] where L = bpy or py and x = 2 or 3). We
and others15 have found that terminal tert-butylimidotitanium
complexes show strong-to-medium intensity absorptions in
the ca. 1200–1260 cmϪ1 region of their IR spectra. The iden-
tity of the vibrational mode(s) corresponding to such absorp-
tions for terminal transition-metal imido complexes is the
subject of ongoing debate.20–22 Nonetheless, the presence of
strong absorptions in this region for 1–3 is consistent with
them all having terminal NBut groups. Additional strong
bands in the 520–550 cmϪ1 region also appear to be associ-
ated with the Ti᎐NBut functionality. Pyridine and bpy do not
᎐
give rise to strong absorptions in the 1200–1260 and 520–550
1
cmϪ1 regions.23 The H NMR spectra of 1–3 in CDCl3 show
resonances for the NBut ligand at δ 0.99, 0.85 and 0.92
respectively, again consistent with this ligand having terminal
co-ordination modes in all three complexes. In related studies
we have found that bridging NBut groups appear at some-
what lower field compared to their terminal NBut counter-
parts in the 1H NMR spectra of otherwise closely related
complexes.11,24 The 13C NMR data for the NBut ligand in the
three complexes are also comparable and consistent with
terminal imido groups.
Crystal structures
Thermal ellipsoid plots of complexes 4 and 12 are shown in
Figs. 1 and 2 respectively; selected bond lengths and angles are
presented in Tables 1 (3–6) and 2. In all five complexes the
Mutually trans arrangements of the pyridine and chloride
ligands in complexes 1 and 2 are assumed by comparison with
the structure of [Ti(NBut)Cl2(OPPh3)2],25 and the crystal-
lographically characterised tris(pyridine) complexes described
herein. The square-pyramidal (as opposed to trigonal-
bipyramidal) limiting geometry depicted is expected by com-
parison with other structurally characterised, five-co-ordinate
imidotitanium dichloride adducts of the type [Ti(NR)Cl2-
L2].15,25,26 The tris(pyridine) complex [Ti(NBut)Cl2(py)3] 3 is
analogous to [Ti{NP(S)R2}Cl2(py)3] (R = Ph or Pri),6,27 but
bis(pyridine) complexes [Ti(NBut)Cl2L2] (L = bpy 1 or py 2) are
examples of the more usual five-co-ordinate imidotitanium
linear or near-linear Ti᎐N᎐R linkage implies that the NR lig-
᎐
and acts as a four-electron donor to titanium.1–3
A preliminary examination of the bond lengths and angles
for these complexes immediately reveals that (i) the nitrogen
atom bonded trans to the imido group is less tightly bound than
those bonded cis and that (ii) the angles subtended at titanium
between the imido nitrogen and the donor atoms of the ligands
cis to it are all greater than 90Њ. These features are consequences
of the trans influence frequently exerted by a multiply bonded
ligand in a transition-metal complex. The likely origins (steric
1552
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558

View Article Online
Table 3 Comparison of metric data for pairs of homologous tert-butyl- and aryl-imido complexes
a
Homologous pair
[Ti(NBut)Cl2(tmen)]
M᎐N/Å
∆
Ref.
᎐
M ᎐N
᎐
1.662(4)
1.702(4)
1.724(4)
1.720(4)
1.712(7)
1.749(7)
1.712(7)
1.729(7)
1.704(5)
1.712(18)
1.704(5)
1.712(7)
1.752(3) and
1.744(3)c
1.761(6)
15
26
7
[Ti(NPh)Cl2(tmen)]
0.040(6)
Ϫ0.004(6)
0.037(10)
0.017(10)
0.008(18)
0.008(9)
[Ti(omtaa)(NBut)]b
[Ti(tmtaa)(NC6H3Me2-2,6)]b
[Ir(η-C5Me5)(NBut)]
8
34
34
34
34
44
45
44
46
47
[Ir(η-C5Me5)(NC6H3Pri2-2,6)]
[Ir(η-C5Me5NBut)]
[Ir(η-C5Me5)(NC6H3Me2-2,6)]
[W2(NBut)2Cl6(µ-Cl)2]
[W2(NC6H4Me-4)2Cl6(µ-Cl)2]
[W2(NBut)2Cl6(µ-Cl)2]
[W2(NC6H3Me2-2,6)2Cl6(µ-Cl)2]
[Nb(η-C5H5)(NBut)Cl2]
[Nb(η-C5H5)(NC6H3Pri2-2,6)Cl2]
3 [Ti(NBut)Cl2(py)3]
0.017(7) and
0.009(7)
47
1.705(3) and
1.706(3)c
1.714(2)
This work
This work
This work
4 [Ti(NPh)Cl2(py)3]
0.009(3) and
0.008(3)
0.000(5) and
Ϫ0.001(5)
5 [Ti(NC6H4Me-4)Cl2(py)3]
1.705(4)
a
Given by (M᎐N bond length for arylimido homologue) Ϫ (M᎐N bond length for tert-butylimido homologue). The e.s.d. of the difference of two
᎐
᎐
values having individual e.s.d.s x and y is given by √(x2 ϩ y2).48 Arylimido complexes with highly electron-withdrawing substituents on the ring (e.g.
NO2 or F) are excluded. b H2tmtaa and H2omtaa = tetra- and octa-methyldibenzotetraaza[14]annulene. c Two molecules in the asymmetric unit.
and/or electronic) of this phenomenon have been discussed in
detail elsewhere.2,16–19 More recently, density functional theory
(DFT) calculations on the complex [OsNCl5]2Ϫ showed that a
stabilisation of the metal to (multiply bonded) ligand π inter-
actions is the major force behind the trans influence, although
the effects of steric repulsion between the multiply bonded
atom and the ligands cis to it are also important.35 Such calcul-
ations have also been used to study the bonding in six-co-
ordinate molybdenum imido complexes.36 For the purposes of
discussion here we shall use the term ‘trans influence’ to
describe the lengthening of the bond trans to the imide ligand
as compared to the equivalent bond cis to the imide ligand in
the same complex.2 We shall use the term ‘structural influence’
to describe the more general effects of the imide ligand on other
groups co-ordinated.
It has been suggested that the overall order of trans influence
between different multiply bonded ligands is nitrido > oxo >
imido,19 although exceptions to this general hierarchy exist.37,38
The homologous series of tris(pyridine) d0 derivatives 3–6
allows us the best crystallographic opportunity to date of com-
paring the structural and trans influences of an imide ligand
within a simple six-co-ordinate complex as a function of the
imide ligand N-substituent. Moreover, the absence of the usual
aryl ring 2,6 substituents in 4–6 means that steric effects are
minimised as far as possible. Not surprisingly, homologous
series of imido complexes have been structurally characterised
before. The series of d2 18-electron complexes [Re(NR)Cl3-
(PEt2Ph)2] [R = Me, C6H4OMe-4 or C6H4C(O)Me-4]39,40 is
the closest previous study of the possible different structural
influence of different imido N-substituents in six-co-ordinate
complexes. However, the X-ray data showed no apparent trends
and both Nugent and Haymore3 and Maatta and co-workers41
have noted that the trans influence in these and certain related
six-co-ordinate complexes can be negligible for d2 18-electron
complexes (and greatest for d0 16-electron species). Lyne and
Mingos35,36 have very recently described the electronic origins
of these empirical observations using DFT. The structures of
the d0 16-electron complexes [W2(NR)2Cl6(µ-Cl)2] (R = Et, Pri,
But, C6H4Me-4 or C6H3Pri2-2,6)42–46 have been reported, but
the binuclear nature of these complexes probably places non-
perturb the intrinsic trans influence. The η-cyclopentadienyl
and -pentamethylcyclopentadienyl derivatives [Nb(η-C5H5)-
(NR)Cl2] (R = Me, But or C6H3Pri2-2,6) and [Ir(η-C5Me5)(NR)]
(R = But, SiMe2But, C6H3Me2-2,6 or C6H3Pri2-2,6) have recently
been described by Gibson 47 and Bergman and co-workers.34
These last two series of homologous half-sandwich complexes
allow some comparisons of the relative structural influences of
tert-butyl- and aryl-imide ligands but do not allow an estima-
tion of their formal trans influences.
Let us now examine the structural data in detail. The two
crystallographically independent Ti᎐N bond lengths in [Ti-
᎐
(NBut)Cl2(py)3] 3 are significantly shorter than the correspond-
ing bonds in Roesky’s [Ti{NP(S)R }Cl (py) ] [R = Ph, Ti᎐N
᎐
2
2
3
1.720(2); R = Pri, Ti᎐N 1.723(2) Å].6,27 Furthermore, the Ti᎐Cl
᎐
and Ti᎐N (py cis and trans) distances in 3 are generally sig-
nificantly longer than in either of the [Ti{NP(S)R2}Cl2(py)3]
analogues, indicating that the NBut ligand has a greater labilis-
ing structural influence than that of N(S)PR2.
The Ti᎐N bond lengths in complex 3 and in the phenyl- and
᎐
tolyl-imido homologues [Ti(NC6H4R-4)Cl2(py)3] (R = H 4 or
Me 5) are equivalent within experimental error. A comparison
of crystallographic data for a collection of homologous pairs of
arylimido and tert-butylimido complexes (Table 3) shows that
the transition metal to nitrogen bond lengths are generally
longer for the arylimido complex. Indeed, one would expect
that imido nitrogen electron density could be delocalised onto
the phenyl ring of the arylimido complexes thus reducing the
metal to nitrogen bond order. However, for individual hom-
ologous pairs of tert-butylimido and arylimido complexes
(Table 3) the difference (∆M ᎐N) between the M᎐N bond lengths is
᎐
significant only in a minori᎐ty of cases.
Despite the experimentally identical Ti᎐N bond lengths in
᎐
complexes 3–5, the NBut and NC6H4R-4 (R = H or Me) ligands
have significantly different structural influences on the Ti-
Cl2(py)3 fragment with the Ti᎐N (py) and Ti᎐Cl bond lengths
for 3 being nearly all significantly longer than the correspond-
ing ones in 4 and 5. This ability of the NBut ligand to labilise
other ligands present (as compared to arylimido homologues)
generally appears also to exist in the literature compounds iden-
tified above, but the effects are not as straightforwardly traced
as in 3–5.
quantifiable constraints on W᎐Cl bond lengths and RN᎐
᎐
W᎐Cl angles cis and trans to the W᎐NR linkage which may
However, in the 4-nitrophenyl analogue [Ti(NC6H4NO2-4)-
᎐
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558
1553

View Article Online
Cl2(py)3] 6 there is a substantial effect of the imido N-
substituent on the Ti᎐N bond length compared to
ible. The closest Group 4 analogues of 12 are the macrocycle-
3
supported, oxotitanium complexes [Ti(O)X2(Me3[9]aneN3)]
(X = Cl or NCS) and [Ti(O)X2(Pri3[9]aneN3)] (X = NCO or
NCS) which also have the two X ligands cis to the multiply
bonded moiety (R3[9]aneN3 = 1,4,7-tri-R-1,4,7-triazacyclo-
nonane, R = Me or Pri).52,53 The two Me3[9]aneN3 derivatives
have been crystallographically characterised but no solution
NMR data were reported. The trans influence of the oxide
ligand in [Ti(O)X2(Me3[9]aneN3)] [0.172(8) and 0.182(7) Å for
X = NCS; 0.157(6) and 0.149(5) Å for X = Cl] is somewhat
smaller compared to that of the NBut ligand in 12. This could
as much imply a greater flexibility of the open-chain pmdien
ligand compared to the cyclic Me3[9]aneN3 as it might a larger
trans influence of the NBut ligand compared to oxo.19 However,
the average Ti᎐Cl bond length in 12 [2.389(1) Å] is slightly
longer than that in [Ti(O)Cl2(Me3[9]aneN3)] [2.364(2) Å] and
may imply that the NBut ligand has the better donor (or at
least, cis-labilising) ability.2
᎐
[∆M ᎐M = 0.016(4) or 0.017(4)], 4 [0.008(4)] and 5 [0.017(5) Å].
Although some of these individual differences are not highly
significant, the consistency of the direction of the difference is
convincing. The imido nitrogen to ipso-carbon bond length in 6
is shorter by 0.023(6) and 0.019(7) Å than that in 4 and 5
respectively. This presumably indicates greater delocalisation of
the imido nitrogen electron density onto the aromatic ring in 6.
Surprisingly, therefore, we find the Ti᎐N (py cis and trans) bond
lengths in 4–6 are equivalent within error despite these implied
differences in Ti᎐N bonding. In contrast the Ti᎐Cl bond length
᎐
in 6 is significantly shorter than those in 4 and 5. This might
reflect enhanced π donation from the Cl ligands in 6 since any σ
effects would presumably be reflected in the Ti᎐N (py) bond
lengths.
Although there are no apparent structural effects of different
electron-withdrawing or -donating groups in the imido aryl ring
in the d2 eighteen-electron complexes [Re(NC6H4R-4)Cl3-
(PEt2Ph)2] [R = OMe or C(O)Me],40 solution dipole moments
Conclusion
for
a homologous series of these complexes have been
reported.49 The molecular dipole moments µ decrease in the
order R = OMe (7.2) > Me (6.5) > H (5.9) > Br (5.2) > Cl
(5.0) > F (4.6) > C(O)Me (4.5 D). These data are consistent
with a through-ring, electron-withdrawing effect of the 4-nitro
ring substituent on the imido nitrogen atom in 6. We note that
in the recently described [Mo(NC6F5)(NC10H15)Cl2(dme)] (C10-
We have described simple routes to synthetically useful quan-
tities of pyridine-supported, tert-butylimidotitanium dichlor-
ide complexes. Straightforward arylamine/tert-butylimido ex-
change reactions give corresponding arylimidotitanium deriv-
atives. The 2,6-disubstituted arylimido complexes are the most
stable and soluble, and are readily obtained in a synthetically
useful way. The bis- and tris-(pyridine) tert-butylimido com-
plexes allow easy access to corresponding tmen and pmdien
adducts and the solid-state structure and solution behaviour of
the latter have been examined. The crystal structures of 3–6
suggest that there is no clear experimental relationship between
H15 = adamant-1-yl, dme = 1,2-dimethoxyethane)31 the Mo᎐N
᎐
(adamantyl) bond lengths [1.715(3) and 1.716(3) Å, two mole-
cules in the asymmetric unit] are substantially shorter than the
Mo᎐NC F bond lengths [1.775(3) and 1.759(3) Å]. The Mo᎐O
᎐
6
5
bonds trans to the adamantylimide ligands are longer [Mo᎐O
(average) 2.370(2) Å] than those trans to NC6F5 [Mo᎐O (aver-
age) 2.346(2) Å].
Ti᎐NR bond length and the ability of the NR group to distort
᎐
other Ti᎐py or ᎐Cl bonds present. The formal trans influences
of the NBut, NPh, NC6H4Me-4 and NC6H4NO2-4 ligands are
experimentally similar to each other and to those of NP(S)Ph2
and NP(S)Pri2 in the d0, six-co-ordinate series [Ti(NR)Cl2(py)3].
Consistent with most previous studies, the NBut ligand has the
greatest labilising effect on other ligands present in the metal
co-ordination sphere.
Despite the different structural features of complexes 3–6
there is no difference in formal trans influence between them.
For 4–6 the trans influence spans the range 0.181(4) to 0.201(4)
Å while for 3 it lies between 0.182(4) and 0.203(1) Å. Interest-
ingly, the trans influences of the imide ligands in [Ti{NP(S)-
R2}Cl2(py)3] [0.210(3) and 0.215(4) Å for R = Ph;6 0.178(4)
and 0.192(4) Å for R = Pri 27] span the values found for 3–6.
These results suggest that the general structural influence of the
various imide ligands is not manifested as a specific experi-
mentally significant perturbation of the trans ligand relative to
the cis ligands. Finally we note that there are no apparent trends
Experimental
General methods and instrumentation
in the N᎐Ti᎐Cl or N᎐Ti᎐N (py, cis) angles for the complexes
All manipulations were carried out under an atmosphere of di-
nitrogen or argon using either standard Schlenk-line or dry-box
techniques. Solvents were predried over activated 4A molecular
sieves and refluxed over sodium (toluene), sodium–potassium
alloy (1:3 w/w) (pentane, diethyl ether), potassium (hexane), or
calcium hydride (dichloromethane) under dinitrogen and col-
lected by distillation; CDCl3 and CD2Cl2 were dried over freshly
ground calcium hydride, vacuum distilled and stored under
dinitrogen in Teflon-valve ampoules. The NMR samples were
prepared in a dry-box in Wilmad 505-PS tubes fitted with a
J. Young NMR/5 valve. 4-Nitroaniline, 4-methylaniline and
titanium tetrachloride (Aldrich) were used as received; 4-tert-
butylpyridine was dried over activated 4A molecular sieves and
used without further purification. Other reagents were dried
over finely ground CaH2 and either distilled at atmospheric
pressure (ButNH2, pyridine) or under reduced pressure (aniline,
H2NC6H3Me2-2,6 and H2NC6H3Pri2-2,6).
᎐
᎐
3–6.
Consider now the structure of [Ti(NBut)Cl2(pmdien)] 12.
The Ti᎐N (pmdien) bond lengths (Table 2) are all substantially
longer than the corresponding Ti᎐N (py) bond lengths in 3,
consistent with the solution NMR fluxional behaviour of 12
and the good σ-donor ability of pyridine. The average Ti᎐NBut
᎐
bond length for the two crystallographically independent mol-
ecules of 12 is slightly shorter than that in 3. This may reflect
the different donor abilities of the ligand trans to the imido
group. The average Ti᎐Cl bond length in 12 [2.389(1) Å] is sub-
stantially shorter than for 3 [2.438(1) Å]. It is also possible that
this may reflect a greater trans influence of pyridine as com-
pared to pmdien. Insufficient comparative structural data are
available for early transition-metal d0 complexes for a quanti-
tative evaluation to be made at this time.
Although a search of the Cambridge Structural Database50,51
showed that a number of complexes of the pmdien ligand have
been structurally characterised for the later transition metals,
[Ti(NBut)Cl2(pmdien)] 12 is the first example of an early
transition-metal derivative. The fac co-ordination mode is
unusual for pmdien, but this comparison is based on the appar-
ent preferences of later transition-metal complexes which lack a
multiply bonded ligand. Furthermore, the NMR data imply
that different solution co-ordination modes are readily access-
Proton and 13C-{1H} NMR spectra were recorded on either a
Bruker WM 250, DPX 300 or AM 400 spectrometer. Where
necessary 13C assignments were confirmed by DEPT-135 and
-90 13C-{1H} experiments. The spectra were referenced intern-
ally to the residual protiosolvent (1H) or solvent (13C) reson-
ances and are reported relative to tetramethylsilane (δ 0) at
ambient probe temperature. Infrared spectra were recorded
on a Nicolet 205 FTIR spectrometer as Nujol mulls between
1554
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558

View Article Online
caesium bromide windows. Elemental analyses were carried out
by the analysis department of this laboratory.
urated pentane solution at r.t. NMR (CDCl3): 1H (250 MHz), δ
9.23 (4 H, d, J = 5.4, o-H of cis-NC5H5), 8.58 (2 H, br d, coup-
ling not resolved, o-H of trans-NC5H5), 7.79 (2 H, t, J = 7.3, p-H
of cis-NC5H5), 7.56 (1 H, br t, J = 7.3, p-H of trans-NC5H5),
7.35 (4 H, apparent t, apparent J = 7.0, m-H of cis-NC5H5),
7.10 (2 H, br, apparent t, apparent J = 7.0 Hz, m-H of trans-
NC5H5) and 0.92 (9 H, s, NBut); 13C-{1H} (62.9 MHz), δ 151.9
(o-C of cis-NC5H5), 150.4 (o-C of trans-NC5H5), 139.3 (p-C of
cis-NC5H5), 136.5 (p-C of trans-NC5H5), 123.7 (m-C of cis-
NC5H5), 123.3 (m-C of trans-NC5H5), 71.5 (NCMe3) and 30.0
(NCMe3). IR: 1603s, 1484m, 1353m, 1247s, 1216s, 1071m,
1041m, 1012w, 756m, 701s, 636m, 541s and 429m cmϪ1 [Found
(Calc. for C19H24Cl2N4Ti): C, 52.6 (53.4); H, 5.9 (5.7); N, 12.9
(13.1)%].
Syntheses
[Ti(NBut)Cl2(bpy)2] 1. To a stirred solution of TiCl4 (3.50
cm3, 31.6 mmol) in dichloromethane (80 cm3) at Ϫ50 ЊC was
added ButNH2 (20.0 cm3, 190 mol, 6.0 equivalents) over 15 min.
An orange precipitate formed, the mixture was stirred at r.t. for
5 h and the volatiles were removed under reduced pressure. The
residue was extracted into toluene (75 cm3), filtered away from
ButNH3Cl and the volatiles were removed under reduced pres-
sure. These residues were re-extracted into dichloromethane
(100cm3), treated with 4-tert-butylpyridine(10.1cm3, 68.0mmol,
2.2 equivalents) and the resultant dark orange solution was
stirred for 16 h. Volatiles were removed under reduced pressure
and the residue extracted into toluene and filtered away from
further residual ButNH3Cl. Subsequent evaporation of the
volatiles under reduced pressure, washing with hexane (2 × 30
cm3) and drying in vacuo gave complex 1 as an orange-red
solid. Yield: 12.2 g (78%). It can be recrystallised from hexane
if required. NMR (CDCl3): 1H (250 MHz), δ 9.17 (4 H,
d, J = 6.7, o-H of NC5H4But), 7.46 (4 H, d, J = 6.7 Hz, m-H
of NC5H4But), 1.36 (18 H, s, NC5H4But) and 0.99 (9 H,
s, NBut); 13C-{1H} (62.9 MHz), δ 163.2 (p-C of C5H4But),
151.1 (o-C of C5H4But), 121.1 (m-C of C5H4But), 73.6
(NCMe3), 35.1 (C5H4CMe3), 30.5 (NCMe3) and 30.2 (C5H4-
CMe3). IR: 1614s, 1500m, 1419m, 1352w, 1275m, 1254m, 1235s,
1209m, 1071m, 1021m, 843m (sh), 832s, 571s and 547m cmϪ1
[Found (Calc. for C22H35Cl2N3Ti): C, 56.0 (57.4); H, 8.0 (7.7);
N, 8.9 (9.1)%].
[Ti(NPh)Cl2(py)3] 4. A solution of [Ti(NBut)Cl2(py)3] 3 (773
mg, 1.80 mmol) in dichloromethane (30 cm3) at r.t. was treated
with aniline (0.17 cm3, 1.80 mmol). After 1.5 h the volatiles were
removed under reduced pressure to give a brown solid which
was washed with hexane (2 × 10 cm3) and dried in vacuo. Yield
800 mg (96%). This compound may be recrystallised at Ϫ25 ЊC
from dichloromethane. Recrystallised yield 288 mg (36%).
1
NMR (CDCl3): H (250 MHz), δ 9.11 (4 H, d, J = 4.9, o-H
of cis-NC5H5), 8.71 (2 H, d, J = 4.3, o-H of trans-NC5H5),
7.81 (2 H, t, J = 7.6, p-H of cis-NC5H5), 7.63 (1 H, t, J = 7.6,
p-H of trans-NC5H5), 7.36 (4 H, apparent t, apparent J = 6.9,
m-H of cis-NC5H5), 7.25 (2 H, apparent t, apparent J = 6.6,
m-H of trans-NC5H5), 7.03 (2 H, apparent t, J = 7.6, m-H of
Ph), 6.94 (2 H, d, J = 7.04, o-H of Ph) and 6.76 (1 H, t,
J = 7.2 Hz, p-H of Ph); 13C-{1H} (62.9 MHz), δ 159.7 (ipso-C
of Ph), 151.6 (o-C of cis-NC5H5), 150.3 (o-C of trans-
NC5H5), 138.6 (p-C of cis-NC5H5), 136.3 (p-C of trans-
NC5H5), 128.1 (o-C of Ph), 124.1 (m-C of cis-NC5H5), 123.7
(m-C of trans-NC5H5), 122.2 (p-C of Ph), m-C of Ph
obscured by py resonances [Found (Calc. for C21H20Cl2N4Ti):
C, 57.5 (56.4); H, 4.7 (4.5); N, 13.0 (12.5)%].
[Ti(NBut)Cl2(py)2] 2. To a stirred solution of TiCl4 (3.50 cm3,
31.6 mmol) in dichloromethane (80 cm3) at Ϫ50 ЊC was added
ButNH2 (20.0 cm3, 190 mol, 6.0 equivalents) over 15 min. An
orange precipitate formed and the mixture was allowed to
warm to r.t. and stirred for 2 h. It was filtered to give a clear
orange solution to which was added pyridine (6.00 cm3, 74.2
mmol, 2.3 equivalents). After 2 h the volatiles were removed
from the orange-red mixture and the residue was extracted into
toluene–dichloromethane (10:1, 50 cm3). This solution was
again filtered (to remove residual ButNH3Cl), evaporated to
dryness under reduced pressure and the orange-red residue
washed with hexane (2 × 20 cm3) to give complex 2 after drying
in vacuo. Yield: 7.0 g (64%). Samples thus prepared are suf-
[Ti(NC6H4Me-4)Cl2(py)3] 5. A solution of complex 3 (483 mg,
1.13 mmol) in dichloromethane (30 cm3) was added to a solu-
tion of 4-methylaniline (139 mg, 1.30 mmol) in dichloro-
methane (20 cm3) at r.t. After 2 h volatiles were removed under
reduced pressure to afford a brown solid which was dissolved in
dichloromethane (20 cm3). Careful layering of this solution
with pentane (15 cm3) at r.t. gave brown crystals of 5 suitable
for X-ray diffraction after 16 h. These were washed with pen-
tane (2 × 5 cm3) and dried in vacuo. Yield: 224 mg (42%).
1
ficiently pure (according to H NMR spectroscopy) to use in
1
further reactivity studies and an analytically pure sample was
obtained by slow cooling of a saturated hot toluene solution.
NMR (CDCl3): H (250 MHz), δ 9.12 (4 H, d, J = 4.6, o-H
of cis-NC5H5), 8.73 (2 H, br s, o-H of trans-NC5H5), 7.83 (2
H, t, J = 7.6, p-H of cis-NC5H5), 7.68 (1 H, t, J = 7.1, p-H of
trans-NC5H5), 7.39 (4 H, apparent t, apparent J = 6.6 Hz, m-
H of cis-NC5H5), 7.26 (2 H, br s, m-H of trans-NC5H5), 6.86
(4 H, apparent s, C6H4Me) and 2.23 (3 H, s, C6H4Me); 13C-
{1H} (75.5 MHz), δ 151.5 (o-C of cis-NC5H5), 150.7 (o-C of
trans-NC5H5), 138.7 (p-C of cis-NC5H5), 136.9 (p-C of trans-
NC5H5), 128.6 (o-C of C6H4Me), 124.2, 123.8, 123.7 (m-C
of cis- and trans-NC5H5 and of C6H4Me), 21.1 (C6H4Me),
ipso-C for C6H4Me not observed [Found (Calc. for C22H22-
Cl2N4Ti): C, 57.1 (57.3); H, 5.0 (4.8); N, 12.2 (12.2)%].
1
NMR (CDCl3): H (250 MHz), δ 9.20 (4 H, d, J = 5.3, o-H of
NC5H5), 7.78 (2 H, J = 7.5, p-H of NC5H5), 7.34 (4 H, apparent
t, apparent J = 7.0 Hz, m-H of NC5H5) and 0.85 (9 H, s, NBut);
13C-{1H} (100.6 MHz), δ 151.7 (o-C of NC5H5), 138.5 (p-C of
NC5H5), 123.9 (m-C of NC5H5), 73.1 (NCMe3) and 30.3
(NCMe3). IR: 1605s, 1483w, 1443s, 1361w, 1245s, 1215m,
1206w (sh), 1066m, 1043m, 1014w, 754m, 695s, 639m, 539m,
524m and 431m cmϪ1 [Found (Calc. for C14H19Cl2N3Ti): C, 48.2
(48.3); H, 5.4 (5.5); N, 12.5 (12.1)%].
[Ti(NBut)Cl2(py)3] 3. To a stirred solution of TiCl4 (10.0 cm3,
0.091 mol) in dichloromethane (150 cm3) at Ϫ50 ЊC was added
ButNH2 (60 cm3, 0.47 mol, 6.3 equivalents) over 15 min to give
an orange-yellow precipitate. The mixture was stirred at r.t. for 2
h and pyridine (ca. 30 cm3, 0.37 mol, 4.1 equivalents) was added
to give a dark orange solution which was stirred at r.t. for 16 h.
After filtration, the volatiles were removed under reduced pres-
sure and the orange solid was washed with hexane (2 × 40 cm3)
to give complex 3 after drying in vacuo. Yield: 38.6 g (98%).
Samples thus prepared are sufficiently pure (1H NMR spec-
troscopy) to use in further reactivity studies. An analytically
pure, X-ray-quality crystalline sample was obtained from a sat-
[Ti(NC6H4NO2-4)Cl2(py)3] 6. A solution of complex 3 (443
mg, 1.04 mmol) in dichloromethane (15 cm3) was added to a
solution of 4-nitroaniline (143 mg, 1.04 mmol) in dichloro-
methane (30 cm3) at r.t. After 48 h the volatiles were removed
under reduced pressure and the red-brown solid was extracted
into dichloromethane (15 cm3). Careful layering of this solution
with hexane (10 cm3) at r.t. gave brown crystals of 6 suitable
for X-ray diffraction after 16 h. These were washed with cold
(Ϫ20 ЊC) hexane–dichloromethane (1:1, 2 × 15 cm3) and dried
in vacuo. Yield: 300 mg (59%). NMR (CDCl3): 1H (250 MHz), δ
9.07 (4 H, d, J = 4.9, o-H of cis-NC5H5), 8.76 (2 H, d, J = 4.2,
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558
1555

View Article Online
o-H of trans-NC5H5), 7.96 (2 H, m, m-H of C6H4NO2), 7.87 (2
H, t, J = 7.7, p-H of cis-NC5H5), 7.71 (1 H, t, J = 7.5, p-H of
trans-NC5H5), 7.41 (4 H, apparent t, apparent J = 7.0, m-H of
cis-NC5H5), 7.26 (2 H, apparent t, apparent J = 6.8 Hz, m-H of
trans-NC5H5) and 6.86 (2 H, m, o-H of C6H4NO2); 13C-{1H}
(100.5 MHz), δ 151.3 (o-C of cis-NC5H5), 150.3 (o-C of trans-
NC5H5), 138.9 (p-C of cis-NC5H5), 136.7 (p-C of trans-NC5H5),
124.6 (m-C of C6H4NO2), 124.1 (m-C of cis-NC5H5), 123.6
(m-C of trans-NC5H5), 123.5 (o-C of C6H4NO2), ipso-C for
C6H4NO2 not observed [Found (Calc. for C21H19Cl2N5O2Ti): C,
50.7 (51.2); H, 3.9 (3.9); N, 14.3 (14.2)%].
t, J = 7.5 Hz, p-H of C6H3Me2) and 2.56 (6 H, s, C6H3Me2);
13C-{1H} (75.5 MHz), δ 151.0 (o-C of NC5H5), 139.0 (p-C of
NC5H5), 134.5 (o-C of C6H3Me2), 126.9 (m-C of C6H3Me2),
124.6 (m-C of NC5H5), 122.5 (p-C of C6H3Me2), 19.0
(C6H3Me2), ipso-carbon resonance not observed [Found (Calc.
for C18H19Cl2N3Ti): C, 54.3 (54.6); H, 5.2 (4.8); N, 10.8
(10.6)%].
[Ti(NC6H3Pri2-2,6)Cl2(py)2] 10. The complex [Ti(NC6H3Pri2-
2,6)Cl2(py)3] (555 mg, 1.04 mmol) was heated at 65 ЊC under a
dynamic vacuum for 4 h to give 10 as a green powder. Yield:
>95%. NMR (CDCl3): 1H (300 MHz), δ 9.18 (4 H, d, J = 4.9, o-
H of NC5H5), 7.94 (2 H, t, J = 7.7, p-H of NC5H5), 7.53 (4 H,
apparent t, apparent J = 7.0, m-H of NC5H5), 6.96 (2 H, d,
J = 7.8, m-H of C6H3Pri2), 6.86 (1 H, t, J = 7.6, p-H of
C6H3Pri2), 4.34 (2 H, spt, J = 6.9, CHMe2 of Pri) and 1.18 (12 H,
d, J = 6.9 Hz, CHMe2 of Pri); 13C-{1H} (75.5 MHz), δ 150.9 (o-
C of NC5H5), 145.1 (o-C of C6H3Pri2), 139.2 (p-C of NC5H5),
124.8 (m-C of NC5H5), 123.5 (p-C of C6H3Pri2), 122.2 (m-C
of C6H3Pri2), 27.7 (CHMe2 of Pri), 23.9 (CHMe2 of Pri), ipso-
carbon resonance not observed [Found (Calc. for C22H27-
Cl2N3Ti): C, 57.9 (58.4); H, 6.0 (6.0); N, 9.3 (9.3)%].
[Ti(NC6H3Me2-2,6)Cl2(py)3] 7. A solution of complex 3 (485
mg, 1.13 mmol) in dichloromethane (30 cm3) at r.t. was treated
with 2,6-dimethylaniline (0.14 cm3, 1.14 mmol). The initially
orange solution became dark red over 2 h. Volatiles were
removed under reduced pressure to afford 7 as a red-brown
powder which was washed twice with pentane (2 × 15 cm3) and
dried in vacuo. Yield: 465 mg (86%). Samples thus prepared are
normally >95% pure (1H NMR spectroscopy). An analytically
pure sample was obtained by careful layering of a dichloro-
methane solution of 7 (containing a small quantity of added
1
pyridine) with hexane at r.t. NMR (CDCl3): H (250 MHz), δ
9.15 (4 H, d, J = 6.4, o-H of cis-NC5H5), 8.63 (2 H, br d,
J = 4.3, o-H of trans-NC5H5), 7.87 (2 H, t, J = 7.7, p-H of cis-
NC5H5), 7.66 (1 H, br t, J = 7.6, p-H of trans-NC5H5), 7.43 (4
H, d of d, J = 6.4 and 7.7, m-H of cis-NC5H5), 7.23 (2 H, br
apparent t, apparent J = 7.4, m-H of trans-NC5H5), 6.77 (2
H, d, J = 7.4, m-H of C6H3Me2), 6.58 (1 H, t, J = 7.4 Hz, p-H
of C6H3Me2) and 2.40 (6 H, s, C6H3Me2); 13C-{1H} (62.9
MHz), δ 151.5 (o-C of cis-NC5H5), 150.4 (o-C of trans-NC5H5),
138.7 (p-C of cis-NC5H5), 136.3 (p-C of trans-NC5H5), 134.9
(o-C of C6H3Me2), 126.8 (m-C of C6H3Me2), 124.2 (m-C of
cis-NC5H5), 123.5 (m-C of trans-NC5H5), 121.7 (p-C of
C6H3Me2), 18.9 (C6H3Me2), ipso-C for C6H3Me2 not observed
[Found (Calc. for C23H24Cl2N4Ti): C, 58.5 (58.1); H, 5.2 (5.1);
N, 12.1 (11.8)%].
[Ti(NBut)Cl2(tmen)] 11. To a solution of [Ti(NBut)Cl2(py)2] 1
(429 mg, 1.23 mmol) in dichloromethane (30 cm3) was added
tmen (0.19 cm3, 1.26 mmol) giving a change from orange-red
to yellow. After 16 h the volatiles were removed under reduced
pressure to yield 11 as a spectroscopically pure yellow powder
1
which was characterised by comparison of its H NMR data
with the literature values.15 Yield: >95%.
[Ti(NBut)Cl2(pmdien)] 12. To a stirred suspension of [Ti(N-
But)Cl2(py)3] 3 (510 mg, 1.19 mmol) in diethyl ether (50 cm3)
was added pmdien (0.20 cm3, 1.19 mmol). After 3 h the yellow
solution was placed in a freezer at Ϫ25 ЊC to afford 12 as light
orange, X-ray-quality crystals after 24 h. These were washed
with the minimum volume of cold (Ϫ25 ЊC) diethyl ether and
1
dried in vacuo. Yield: 238 mg (55%). NMR (CDCl3): H (250
[Ti(NC6H3Pri2-2,6)Cl2(py)3] 8. A solution of complex 3 (474
mg, 1.11 mmol) in dichloromethane (30 cm3) at r.t. was treated
with 2,6-diisopropylaniline (0.21 cm3, 1.11 mmol). The orange
solution became dark red over 3 h. Volatiles were removed
under reduced pressure to afford 8 as a brown powder which
was washed twice with hexane (2 × 15 cm3) and dried in vacuo.
Yield of the analytically pure product: 475 mg (84%). NMR
MHz), δ 3.93 (2 H, apparent t, J = 12.7, CH2), 3.01–2.90 (16 H,
overlapping s and m, NMe2 and CH2), 2.32 [3 H, s, CH2-
N(Me)CH2], 1.87 (2 H, d, J = 12.7 Hz, CH2) and 1.03 (NBut);
13C-{1H} (62.9 MHz), δ 61.8 (CH2), 56.4 and 55.2 (NMe2), 55.0
(CH2), 44.5 [CH2N(Me)CH2], 30.8 (NCMe3), CMe3 quaternary
carbon not observed [Found (Calc. for C13H32Cl2N4Ti): C,
43.2 (43.0); H, 9.2 (8.9); N, 15.5 (15.4)%].
1
(CDCl3): H (250 MHz), δ 9.17 (4 H, d, J = 4.9, o-H of cis-
NC5H5), 8.66 (2 H, br s, o-H of trans-NC5H5), 7.90 (2 H, t,
J = 7.7, p-H of cis-NC5H5), 7.68 (1 H, br s, p-H of trans-
NC5H5), 7.48 (4 H, apparent t, apparent J = 7.0, m-H of cis-
NC5H5), 7.25 (2 H, br s, m-H of trans-NC5H5), 6.92 (2 H, d,
J = 7.6, m-H of C6H3Pri2), 6.81 (1 H, t, J = 7.6, p-H of
C6H3Pri2), 4.33 (2 H, spt, J = 6.9, CHMe2 of Pri) and 1.11 (12 H,
d, J = 6.9 Hz, CHMe2 of Pri); 13C-{1H} (62.9 MHz), δ 156.9
(ipso-C of C6H3Pri2), 151.1 (o-C of cis-NC5H5), 150.3 (o-C of
trans-NC5H5), 145.3 (o-C of C6H3Pri2), 139.0 (p-C of cis-
NC5H5), 136.3 (p-C of trans-NC5H5), 124.5 (m-C of cis-
NC5H5), 123.6 (p-C of C6H3Pri2), 123.0 (m-C of trans-NC5H5),
122.1 (m-C of C6H3Pri2), 27.5 (CHMe2 of Pri) and 23.9 (CHMe2
of Pri) [Found (Calc. for C27H32Cl2N4Ti): C, 60.6 (61.0); H, 6.2
(6.1); N, 10.6 (10.5)%].
Crystallography
For complex 4 a crystal was taken from under the dichloro-
methane mother-liquor, mounted on a glass fibre with RS3000
oil and transferred on a goniometer head to a Stoë Stadi-4
four-circle diffractometer equipped with an Oxford Cryo-
systems low-temperature device.54 For 12 a crystal was sealed in
a Lindemann glass capilliary under N2 and transferred to a
Siemens SMART CCD diffractometer. Data collection and
processing parameters are listed in Table 4. The data were cor-
rected for Lorentz-polarization and absorption effects, equiva-
lent reflections were merged (for 12) and systematically absent
reflections rejected. The structures were solved by direct methods
(4) or from a Patterson synthesis (12). Subsequent Fourier-
difference syntheses revealed the positions of all other non-
hydrogen atoms. The crystals of 4 contained 1.5 CH2Cl2 mol-
ecules of crystallisation per asymmetric unit. Non-hydrogen
[Ti(NC6H3Me2-2,6)Cl2(py)2] 9. The complex [Ti(NC6H3Me2-
2,6)Cl2(py)3] (457 mg, 1.01 mmol) was heated at 65 ЊC under a
dynamic vacuum for 7 h. The resultant green powder was
extracted into dichloromethane (20 cm3) and filtered. Cooling
to Ϫ25 ЊC and washing with cold hexane yielded 9 as a green
powder. Yield: >95%. NMR (CDCl3): 1H (300 MHz), δ 9.16 (4
H, d, J = 4.8, o-H of NC5H5), 7.94 (2 H, t, J = 7.7, p-H of
NC5H5), 7.53 (4 H, apparent t, apparent J = 7.0, m-H of
NC5H5), 6.87 (2 H, d, J = 7.9, m-H of C6H3Me2), 6.68 (1 H,
2
atoms were refined against Fo with anisotropic displacement
parameters by full-matrix least-squares procedures. Hydrogen
atoms were located from Fourier-difference syntheses (4) or
placed in estimated positions (12). Standard weighting schemes
were applied. For 12 the data were corrected for isotropic extinc-
tion (not necessary for 4). All crystallographic calculations were
performed using SHELXS 8655 and SHELXL 96.56
1556
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558

View Article Online
Table 4 Crystal data collection and processing parameters* for [Ti(NPh)Cl2(py)3]ؒ1.5CH2Cl2 4ؒ1.5CH2Cl2 and [Ti(NBut)Cl2(pmdien)] 12
4ؒ1.5CH2Cl2
12
Molecular formula
M
C21H20Cl2N4Tiؒ1.5CH2Cl2
574.60
C13H32Cl2N4Ti
363.23
a/Å
b/Å
c/Å
30.493(8)
9.117(4)
18.980(4)
95.14(2)
5255(3)
22.321(2)
19.535(2)
17.601(2)
96.681(4)
7622.7(13)
16
β/Њ
U/Å3
Z
8
Dc/Mg mϪ3
1.452
0.853
1.266
0.727
µ/mmϪ1
F(000)
2344
3104
Crystal size/mm
θ Range for data collection/Њ
Reflections collected
Independent reflections
Rint
0.90 × 0.87 × 0.58
2.54–25.05
4633
0.40 × 0.20 × 0.20
1.74–26.05
15 775
6537
0.0509
4633
No reflections to be merged
3949
ψ Scans
0.682, 0.600
9.1
4633, 0, 386
0.0388, 0.0839
0.0512, 0.0949
1.161
Reflections with [I > 2σ(I)]
Absorption correction
Maximum, minimum transmission
Decay correction (%)
Data, restraints, parameters
Final R1, wR2 [I > 2σ(I)]
(all data)
5502
Analytical
0.775, 0.706
None
6532, 0, 553
0.0581, 0.1198
0.0728, 0.1302
1.196
Goodness of fit on F2
Final (∆/σ)max
0.037
0.57, Ϫ0.66
0.003
0.44, Ϫ0.44
Largest residual peaks/e ÅϪ3
* Details in common: 150(2) K; Mo-Kα radiation (λ 0.710 73 Å); monoclinic, space group C2/c; R1 = Σ Fo| Ϫ |Fc /Σ|Fo|; wR2 = [Σw(Fo2 Ϫ Fc )2/
2
¹
¹

Σw(Fo ) ]²; goodness of fit = [Σw(Fo2 Ϫ Fc )2/(Nobs Ϫ Nparam)]² (based on all data).
2
2
2

15 T. S. Lewkebandara, P. H. Sheridan, M. J. Heeg, A. L. Rheingold and
Atomic coordinates, thermal parameters, and bond lengths
and angles have been deposited at the Cambridge Crystallo-
graphic Data Centre (CCDC). See Instructions for Authors,
J. Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the
CCDC for this material should quote the full literature citation
and the reference number 186/446.
C. H. Winter, Inorg. Chem., 1994, 33, 5879.
16 R. Mason and A. D. C. Towl, J. Chem. Soc. A, 1970, 1601.
17 J. K. Burdett and T. A. Albright, Inorg. Chem., 1979, 18, 2112.
18 D. Bright and J. A. Ibers, Inorg. Chem., 1969, 8, 709.
19 E. M. Shustorovich, M. A. Porai-Koshits and Y. A. Buslaev, Coord.
Chem. Rev., 1975, 17, 1.
20 K. Dehnicke and J. Strähle, Angew. Chem., Int. Ed. Engl., 1981, 20,
413.
21 W. P. Griffith, A. J. Nielson and M. J. Taylor, J. Chem. Soc., Dalton
Trans., 1988, 647.
22 J. H. Osborne, A. L. Rheingold and W. C. Trogler, J. Am. Chem.
Soc., 1985, 107, 7945.
23 K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds, Wiley, New York, 4th edn., 1986.
24 P. E. Collier, A. J. Blake and P. Mountford, unpublished work.
25 C. H. Winter, P. H. Sheridan, T. S. Lewkebandara, M. J. Heeg and
J. W. Proscia, J. Am. Chem. Soc., 1992, 114, 1095.
26 R. Duchateau, A. J. Williams, S. Gambarotta and M. Y. Chiang,
Inorg. Chem., 1991, 30, 4863.
Acknowledgements
We thank the University of Nottingham for a demonstrator-
ship (to P. E. C.), the EPSRC for a studentship (to S. C. D.) and
use of the X-Ray Crystallography Service (Cardiff), the Royal
Society for an equipment grant and Professor J. A. K. Howard
(Durham) for the use of X-ray crystallographic facilities. We
also acknowledge the use of the EPSRC’s Chemical Database
Service at Daresbury.
27 H. W. Roesky, T. Raubold, M. Witt, R. Bohra and M. Noltemeyer,
Chem. Ber., 1991, 124, 1521.
References
1 D. E. Wigley, Prog. Inorg. Chem., 1994, 42, 239.
2 W. A. Nugent and J. M. Mayer, Metal–Ligand Multiple Bonds,
Wiley-Interscience, New York, 1988.
28 H. Burger and U. Wannagat, Monatsh. Chem., 1963, 94, 761.
29 D. J. Arney, M. A. Bruck, S. R. Huber and D. E. Wigley, Inorg.
Chem., 1992, 31, 3749.
3 W. A. Nugent and B. L. Haymore, Coord. Chem. Rev., 1980, 31, 123.
4 V. C. Gibson, Adv. Mater., 1994, 6, 37.
30 M. P. Coles, C. I. Dalby, V. C. Gibson, W. Clegg and M. R. J.
Elsegood, Polyhedron, 1995, 14, 2455.
5 J. E. Hill, R. D. Profilet, P. E. Fanwick and I. P. Rothwell, Angew.
Chem., Int. Ed. Engl., 1990, 29, 664.
6 H. W. Roesky, H. Voelker, M. Witt and M. Noltemeyer, Angew.
Chem., Int. Eng. Engl., 1990, 29, 669.
7 S. C. Dunn, A. S. Batsanov and P. Mountford, J. Chem. Soc., Chem.
Commun., 1994, 2007.
31 A. Bell, W. Clegg, P. W. Dyer, M. R. J. Elsegood, V. C. Gibson and
E. L. Marshall, J. Chem. Soc., Chem. Commun., 1994, 2247.
32 A. Bell, W. Clegg, P. W. Dyer, M. R. J. Elsegood, V. C. Gibson and
E. L. Marshall, J. Chem. Soc., Chem. Commun., 1994, 2547.
33 R. I. Michelman, R. G. Bergman and R. A. Andersen, Organo-
metallics, 1993, 12, 2741.
8 P. Mountford and D. Swallow, J. Chem. Soc., Chem. Commun.,
1995, 2357.
34 D. S. Glueck, J. Wu, F. J. Hollander and R. G. Bergman, J. Am.
Chem. Soc., 1991, 113, 2041.
9 K. A. Butakoff, D. A. Lemenovskii, P. Mountford, L. G. Kuz’mina
and A. V. Churakov, Polyhedron, 1996, 15, 489.
10 S. C. Dunn, P. Mountford and O. V. Shishkin, Inorg. Chem., 1996,
35, 1006.
35 P. D. Lyne and D. M. P. Mingos, J. Chem. Soc., Dalton Trans., 1995,
1635.
36 P. D. Lyne and D. M. P. Mingos, J. Organomet. Chem., 1994, 478,
141.
11 S. C. Dunn, P. Mountford and D. A. Robson, J. Chem. Soc., Dalton
Trans., 1997, 293.
12 P. E. Collier, S. C. Dunn, P. Mountford, O. V. Shishkin and
D. Swallow, J. Chem. Soc., Dalton Trans., 1995, 3743.
13 R. T. Cowdell and G. W. A. Fowles, J. Chem. Soc., 1960, 2522.
14 A. J. Nielson, Inorg. Chim. Acta, 1988, 154, 177.
37 R. C. B. Copley, P. W. Dyer, V. C. Gibson, J. A. K. Howard,
E. L. Marshall, W. Wang and B. Whittle, Polyhedron, 1996, 15, 3001.
38 W. Clegg, R. J. Errington, D. C. R. Hockless and C. Redshaw,
J. Chem. Soc., Dalton Trans., 1993, 1965.
39 D. Bright and J. A. Ibers, Inorg. Chem., 1969, 8, 703.
40 D. Bright and J. A. Ibers, Inorg. Chem., 1968, 7, 1099.
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558
1557

View Article Online
41 C. Y. Chou, J. C. Huffman and E. A. Maatta, Inorg. Chem., 1986, 25,
822.
49 J. Chatt, J. D. Garforth, N. P. Johnson and G. A. Rowe, J. Chem.
Soc., 1964, 1012.
42 D. C. Bradley, R. J. Errington, M. B. Hursthouse and R. L. Short,
J. Chem. Soc., Dalton Trans., 1990, 1043.
43 B. R. Ashcroft, G. R. Clark, A. J. Nielson and C. E. F. Rickard,
Polyhedron, 1986, 5, 2081.
44 W. Clegg and R. J. Errington, Acta Crystallogr., Sect. C, 1987, 43,
2223.
45 K. Stahl, F. Weller, K. Dehnicke and P. Paetzold, Z. Anorg. Allg.
Chem., 1986, 534, 93.
46 A. Bell, personal communication cited in ref. 1.
47 D. N. Williams, J. P. Mitchell, A. D. Poole, U. Siemeling, W. Clegg,
D. C. R. Hockless, P. A. O’Neil and V. C. Gibson, J. Chem. Soc.,
Dalton Trans., 1992, 739.
50 F. H. Allen and O. Kennard, Chem. Design Autom. News, 1993, 8, 1.
51 D. A. Fletcher, R. F. McMeeking and D. Parkin, J. Chem. Inf.
Comput. Sci., 1996, 36, 746.
52 A. Bodner, P. Jeske, T. Weyhermüller, K. Wieghardt, E. Dubler,
H. Schmalle and B. Nuber, Inorg. Chem., 1992, 31, 3737.
53 P. Jeske, G. Haselhorst, T. Weyhermüller, K. Wieghardt and
B. Nuber, Inorg. Chem., 1994, 33, 2462.
54 J. Cosier and A. M. Glazer, J. Appl. Crystallogr., 1986, 19, 105.
55 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467.
56 G. M. Sheldrick, SHELXL 96, Institut für Anorganische Chemie der
Universität Göttingen, 1996.
48 G. H. Stout and L. H. Jensen, X-Ray Structure Determination,
Wiley, Toronto, 2nd edn., 1989.
Received 14th November 1996; Paper 6/07735H
1558
J. Chem. Soc., Dalton Trans., 1997, Pages 1549–1558