Novel aspects of the transamination reaction between Ti(NMe 2)4 and primary amines

Article abstract of DOI:10.1039/c3dt50840d

A detailed study of the transamination reaction between Ti(NMe ₂)₄ and primary amines RNH₂ is reported Alkylamines (1-adamantylamine, t-butylamine) and triphenylsilylamine yield dimers of the type [Ti(μ-NR)(NMe₂)₂]₂ (R = 1-adamantyl (1), ^tBu (2), SiPh₃ (3)). The experimental conditions and the nature of the amine are important issues that determine the scope and yield of the reaction. When reacted under the same conditions, aniline PhNH₂ does not yield the expected dimer, but instead the trimer [Ti₃(μ²-NPh)₃(μ³-NPh)(NMe ₂)₄(NHMe₂)] (5) and tetramer [Ti ₄(μ-NPh)₅(NMe₂)₆] (6) complexes were characterized. The reaction of Ti(NMe₂)₄ and 2,6-diisopropylaniline (Ar*NH₂) is essentially an equilibrium reaction of great complexity in which multiple species can coexist in solution. However, under certain circumstances, we were able to isolate and/or characterize several new complexes such as [Ti(μ-NAr*)(NMe ₂)₂]₂ (7), [(Me₂N) ₂Ti(μ-NAr*)₂Ti(NMe₂)(NHAr*)] (8), [Ti(NMe₂)₃]₂(μ-NAr*) (9), [(Ar*NH)₂Ti(μ-NAr*)₂Ti(NMe ₂)₂] (10), [(Me₂N)(Me₂NH)Ti(μ- NAr*)₂Ti(NMe₂)(NAr*)] (11), Ti(NHAr*)₄ (12), and [(Ar*NH)₂Ti(μ- NAr*)₂Ti(NMe₂)(NHAr*)] (13). Complexes 8 and 11 are tautomers. Alternatively, complex 7 was prepared by deprotonation of NHMe₂ ligands in Ti(NAr*)Cl₂(NHMe₂) ₂ with KN(SiMe₃)₂. Addition of pyridine (py) to the reaction of Ti(NMe₂)₄/Ar*NH₂ mixtures or to 7 or 8 allows characterization of transient [Ti(μ-NAr*) (NMe₂)₂]₂·py (14) and trapping of the intermediates [(Me₂N)(py)₂Ti(μ-NAr*) ₂Ti(NMe₂)(NAr*)] (15), Ti(NAr*)(NHAr*) (NMe₂)(py)₂ (16), and Ti(NAr*)(NHAr*) ₂(py)₂ (17). The crystal structures of 15 compounds were determined by X-ray diffraction studies.

Full text of DOI:10.1039/c3dt50840d

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Novel aspects of the transamination reaction between
Ti(NMe2)4 and primary amines†‡
Cite this: Dalton Trans., 2013, 42, 12203
Christian Lorber*a,b and Laure Vendiera,b
A detailed study of the transamination reaction between Ti(NMe2)4 and primary amines RNH2 is
reported Alkylamines (1-adamantylamine, t-butylamine) and triphenylsilylamine yield dimers of the type
[Ti(μ-NR)(NMe2)2]2 (R = 1-adamantyl (1), tBu (2), SiPh3 (3)). The experimental conditions and the nature of
the amine are important issues that determine the scope and yield of the reaction. When reacted under
the same conditions, aniline PhNH2 does not yield the expected dimer, but instead the trimer [Ti3(μ2-
NPh)3(μ3-NPh)(NMe2)4(NHMe2)] (5) and tetramer [Ti4(μ-NPh)5(NMe2)6] (6) complexes were characterized.
The reaction of Ti(NMe2)4 and 2,6-diisopropylaniline (Ar*NH2) is essentially an equilibrium reaction of
great complexity in which multiple species can coexist in solution. However, under certain circumstances,
we were able to isolate and/or characterize several new complexes such as [Ti(μ-NAr*)(NMe2)2]2 (7),
[(Me2N)2Ti(μ-NAr*)2Ti(NMe2)(NHAr*)] (8), [Ti(NMe2)3]2(μ-NAr*) (9), [(Ar*NH)2Ti(μ-NAr*)2Ti(NMe2)2] (10),
[(Me2N)(Me2NH)Ti(μ-NAr*)2Ti(NMe2)(vNAr*)] (11), Ti(NHAr*)4 (12), and [(Ar*NH)2Ti(μ-NAr*)2Ti(NMe2)-
(NHAr*)] (13). Complexes 8 and 11 are tautomers. Alternatively, complex 7 was prepared by deprotona-
tion of NHMe2 ligands in Ti(vNAr*)Cl2(NHMe2)2 with KN(SiMe3)2. Addition of pyridine (py) to the reac-
tion of Ti(NMe2)4/Ar*NH2 mixtures or to 7 or 8 allows characterization of transient [Ti(μ-NAr*)-
(NMe2)2]2·py (14) and trapping of the intermediates [(Me2N)(py)2Ti(μ-NAr*)2Ti(NMe2)(vNAr*)] (15),
Ti(vNAr*)(NHAr*)(NMe2)(py)2 (16), and Ti(vNAr*)(NHAr*)2(py)2 (17). The crystal structures of 15 com-
pounds were determined by X-ray diffraction studies.
Received 28th March 2013,
Accepted 3rd June 2013
DOI: 10.1039/c3dt50840d
www.rsc.org/dalton
In homogeneous catalytic applications2 such as in olefin meta-
thesis, olefin polymerization or oligomerization, the imido
1. Introduction
Over the last three decades, the now ubiquitous imido func- group has emerged as an ancillary ligand that is a viable
tionality has become an important ligand class, particularly in alternative to isolobal3 cyclopentadienyl ligands. Imido com-
high oxidation state early transition metal chemistry, that con- plexes have also been involved in a variety of stoichiometric or
tinues to be an attractive focus in various fields (organic, in- catalytic transformations in which the imido functionality is a
organic, industrial, and materials chemistry).1 Compounds direct participant in the reactivity, including such important
containing the dianionic π-donor terminal imido ligand [NR]2− industrial reactions as imines metathesis, C–H activation, reac-
(R = alkyl or aryl) have found practical use in multiple areas of tions with unsaturated C–C or C–X bonds, alkyne or alkene
research, driven by the ease of variation of the steric and elec- hydroamination, carboamination reaction, strained hetero-
tronic properties of the imido group, which allows, for cycles opening, propene ammoxidation, and hydrodenitro-
example, the fine-tuning of catalysts containing imido ligands. genation.4 In addition, in materials chemistry, imido
compounds are known precursors for metal nitride thin films
produced by MOCVD,5 and have also found applications in
aCNRS, LCC (Laboratoire de Chimie de Coordination), 205 route de Narbonne,
polyoxometalate chemistry.6
BP 44099, F-31077 Toulouse, France. E-mail: lorber@lcc-toulouse.fr
bUniversité de Toulouse, UPS, INPT, LCC, F-31077 Toulouse, France
†This paper is dedicated to Richard H. Holm, a brilliant chemist and mentor.
Various protocols have been developed for the assembly
of a terminal imido ligand onto early transition metals.1
‡Electronic supplementary information (ESI) available: Full experimental Considering titanium complexes, the most common
methods7 for accessing the TivNR motif are via dehydro-
halogenation (for example when treating a TiCl4 precursor with
a primary amine), deprotonation of an amine (with Ti–R or
Ti–NR2 bonds), α-hydrogen abstraction from a Ti–NHR bond,
and to a lesser extent transiminations (replacement of an
description of all compounds, CIF files, and tables of atomic coordinates, bond
distances and angles for the X-ray crystal structures of complexes 1, 3–5, 7–9, 10,
11–12, 15–17, [V(μ-N-1-adamantyl)(NMe2)2]2, and [V(μ-NArCl2)(NMe2)(NHArCl2)]2
(ArCl2 = o-Cl2–C6H3), as well as tables comparison of selected bond parameters
between compounds. CCDC 931095–931109. For ESI and crystallographic data in
CIF or other electronic format see DOI: 10.1039/c3dt50840d
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 12203–12219 | 12203

View Article Online
Paper
Dalton Transactions
Table 1 Designation and location in the article of titanium-imido or -amido
complexes (Arabic numbers) or non-isolated/transient species (letters)
Compound formula
Designation Section
[Ti(μ-N-1-adamantyl)(NMe2)2]2
[Ti(μ-NtBu)(NMe2)2]2
1
2
3
4
5
6
7
8
9
2.1
2.1
2.1
2.1
2.1
2.1
2.2
2.2, 2.3.2
2.3.1
2.3.2
2.3.2
2.3.2
2.3.2
2.4
2.4
2.4
2.4
2.2
[Ti(μ-NSiPh3)(NMe2)2]2
[(Me2N)2Ti(μ-NSiPh3)2Ti(NMe2)(NHSiPh3)]
[Ti3(μ2-NPh)3(μ3-NPh)(NMe2)4(NHMe2)]
[Ti4(μ-NPh)5(NMe2)6]
[Ti(μ-NAr*)(NMe2)2]2
[(Me2N)2Ti(μ-NAr*)2Ti(NMe2)(NHAr*)]
[Ti(NMe2)3]2(μ-NAr*)
[(Ar*NH)2Ti(μ-NAr*)2Ti(NMe2)2]
10
[(Me2N)(Me2NH)Ti(μ-NAr*)2Ti(NMe2)(vNAr*)] 11
Ti(NHAr*)4
12
13
14
15
16
17
A
[(Ar*NH)2Ti(μ-NAr*)2Ti(NMe2)(NHAr*)]
[Ti(μ-NAr*)(NMe2)2]2·py
[(Me2N)(py)2Ti(μ-NAr*)2Ti(NMe2)(vNAr*)]
Ti(vNAr*)(NHAr*)(NMe2)(py)2
Ti(vNAr*)(NHAr*)2(py)2
[Ti(NMe2)2(NHAr*)2]
[Ti(NMe2)3(NHAr*)]
B
C
D
E
F
G
H
2.2, 2.3.1
2.3.2
2.3.2
2.3.2
2.3.2
2.3.2
2.4
[(Ar*NH)(Me2N)Ti(μ-NAr*)2Ti(NMe2)(NHAr*)]
Scheme 1 Proposed intermediates involved in the mechanism of formation of
M(vNR)Cl2(NHMe2)2 (M = Ti, V) using the one-pot synthesis from M(NMe2)4/
RNH2/Me3SiCl mixtures.
[Ti(μ-NAr*)(NHAr*)2]2
Ti(vNAr*)(NHAr*)2(NHMe2)2
[Ti(vNAr*)(NHAr*)2]
[Ti(vNAr*)(NHAr*)2(NHAr*)]
Ti(vNAr*)(NMe2)2(py)2
existing imido by a different imido group) and trimethylsilyl
Ti(NMe2)4 and primary amines. The synthesis and reactivity
schemes are presented in Schemes 1–8. Isolated and/or charac-
terized complexes are denoted by Arabic numbers (1–17, see
Table 1), whereas non-isolated or transient species are denoted
by letters (A–I). The structures of 15 complexes are set out in
Fig. 1–8 or given as ESI;‡ selected metric data are collected in
Table 2.
chloride elimination reactions (in
complex).
a
[Ti(Cl)(NRSiMe3)]
In the early 2000s, we reported new routes to vanadium and
titanium terminal imido synthons of the general formula
[M(vN–R)Cl2(NHMe2)2] (M = Ti, V),2h,8–11 based on a very simple,
single-step, one-pot reaction between a commercially available
M(NMe2)4 precursor and primary amines in the presence of tri-
methylchlorosilane. For kinetic reasons,12 we propose the
mechanism for the one-pot formation of these complexes to
be the following (Scheme 1). The reaction starts with a trans-
amination between M(NMe2)4 and the amine RNH2, affording
an amido intermediate (either A or B) that undergoes deproto-
nation by an –NMe2 ligand (or a second –NHR) and generates
the imido motif. Exchange of the remaining –NMe2 ligand by
chlorides is done via silylation of the amido with trimethyl-
chlorosilane. The situation is further complicated by the pres-
ence of NHMe2, which can coordinate to the titanium center
in different intermediates. In order to address a few points in
this mechanism, we have studied, separately, the reactivity
between couples of these components (or reactants: M(NMe2)4,
RNH2, Me3SiCl). In this first paper, we report on our long-term
investigations on the reaction between Ti(NMe2)4 and 2,6-di-
isopropylaniline, and the synthesis of the new compounds that
were prepared.
2.1 Preliminary observations. Choice of model substrates
Pioneering studies by Bradley et al. suggested that the trans-
amination reaction (aminolysis) between numerous amido
compounds and amines is at equilibrium in solution.13–15
Given that the pKa of Me2NH/Me2N− is about 40,16 there is a
wide range of weakly acidic primary amines RNH2 which
should react with a transition metal amide M–NMe2 bond.
Such a reaction presupposes that the transamination reaction
(M–NMe2 + RNH2 → M–NHR + Me2NH) is simply an acid–base
reaction, and that the position of the equilibrium depends
only on the acidity of the species involved. However, it has
been demonstrated later that this reaction is sensitive to steric
effects.17
With the objective of studying the different issues and possi-
ble intermediates in the reaction of M(NMe2)4 and RNH2, and
the factors governing them, we initially chose a more suitable
model system for investigation (i.e., choice of the metal pre-
cursor and the amine). In general, monomeric imido–
vanadium(IV) d1-systems are paramagnetic8 and NMR silent (the
[V(μ-NR)(NMe2)2]2 dimers are, however, diamagnetic due to a
2. Results and discussion
This work constitutes the first systematic investigation of the V–V bond), which renders their mechanistic study more
synthetic and structural chemistry of titanium-imido and/or difficult. This complication induced us to study instead the
amido complexes generated by transamination reaction from reactivity of the titanium systems, starting from Ti(NMe2)4.18
12204 | Dalton Trans., 2013, 42, 12203–12219
This journal is © The Royal Society of Chemistry 2013

Table 2 Comparison of average interatomic distances (Å) and angles (°) in complexes 7–9, 10, 11, and 15
7
8
9
10
11
15
Ti–Nimido
2.0211(10) [Ti1–N1]
1.8818(11) [Ti1–N2]
2.0123(11) [Ti2–N2]
1.8694(10) [Ti2–N1]
1.9626(15) [Ti1–N1]
1.8485(15) [Ti1–N2]
2.0139(15) [Ti2–N2]
1.9173(16) [Ti2–N1]
1.9602(19) [Ti1–N1]
1.9522(18) [Ti2–N1]
1.8362(19) [Ti1–N3]
2.054(2) [Ti2–N3]
1.913(2) [Ti1–N4]
1.953(2) [Ti2–N4]
2.074(4) [Ti1–N1]
1.723(6) [Ti1–N2]
1.841(4) [Ti2–N1]
2.070(3) [Ti1–N1]
2.053(3) [Ti1–N2]
1.876(3) [Ti2–N2]
1.861(3) [Ti2–N1]
1.758(3) [Ti1–N5]
1.409(5) [C1–N1]
1.422(5) [C13–N2]
1.337(5) [C25–N3]
1.923(3) [Ti1–N6]
1.925(3) [Ti2–N7]
Nimido–C
1.4085(15) [N1]
1.4099(15) [N2]
1.418(2) [N1]
1.409(2) [N2]
1.426(3)
1.404(3) [N3]
1.427(3) [N4]
1.411(6) [C1–N1]
1.362(8) [C13–N2]
Ti–NMe2
1.8890(11) [Ti1–N3]
1.9197(11) [Ti1–N4]
1.8948(11) [Ti2–N5]
1.9224(12) [Ti2–N6]
1.8993(16) [Ti1–N4]
1.9115(17) [Ti2–N5]
1.8829(17) [Ti2–N6]
1.901(2) [Ti1–N2]
1.901(2) [Ti1–N3]
1.900(2) [Ti1–N4]
1.888(2) [Ti2–N5]
1.897(2) [Ti2–N6]
1.911(2) [Ti2–N7]
—
1.909(2) [Ti2–N5]
1.870(2) [Ti2–N6]
1.875(6) [Ti2–N3]
1.918(6) [Ti1–N5]
Ti–NHamido
—
—
1.9442(17)
1.9469(19) [Ti1–N1]
1.955(2) [Ti1–N2]
—
—
—
a
Ti–NL
—
—
2.185(8) [Ti2–N4]
2.359(3) [Ti2–N3]
2.240(3) [Ti2–N4]
2.8475(10)
Ti⋯Ti
2.8089(3)
2.7896(5)
3.621
2.8134(6)
2.8198(17)
Ti–Nimido–C
111.28(8) [C1–N1–Ti1]
152.99(8) [C1N1–Ti2]
152.28(9) [C13–N2–Ti1]
113.34(8) [C13–N2–Ti2]
130.17(12) [C1–N1–Ti1]
137.15(12) [C1–N1–Ti2]
158.65(14) [C13–N2–Ti1]
102.80(12) [C13–N2–Ti2]
111.97(14) [C1–N1–Ti1]
112.37(14) [C1–N1–Ti2]
170.07(18) [C25–N3–Ti1]
92.63(14) [C25–N3–Ti2]
134.43(17) [C37–N4–Ti1]
129.81(16) [C37–N4–Ti2]
123.4(3) [C1–N1–Ti1]
143.2(3) [C1–N1–Ti2]
168.2(5) [C13–N2–Ti1]
114.7(2) [C1–N1–Ti1]
151.0(3) [C1–N1–Ti2]
119.0(2) [C13–N2–Ti1]
147.7(3) [C13–N2–Ti2]
174.9(3) [C35–N5–Ti1]
92.67(13) [Ti1–N1–Ti2]
92.80(13) [Ti1–N2–Ti2]
92.67(13) [N1–Ti1–N2]
92.80(13) [N2–Ti2–N1]
123.79(14) [N5–Ti1–N1]
119.25(14) [N5–Ti1–N2]
—
Ti–Nimido–Ti
92.36(4) [Ti1–N1–Ti2]
92.27(4) [Ti1–N2–Ti2]
87.07(4) [N1–Ti1–N2]
87.67(4) [N2–Ti2–N1]
91.94(6) [Ti1–N1–Ti2]
92.38(6) [Ti1–N2–Ti2]
88.83(7) [N1–Ti1–N2]
85.46(6) [N2–Ti2–N1]
135.53(10)
92.48(9) [Ti1–N3–Ti2]
93.39(9) [Ti1–N4–Ti2]
82.97(8) [N3–Ti2–N4]
90.16(9) [N3–Ti1–N4]
91.97(16) [Ti1–N1–Ti2]
Nimido–Ti–Nimido
—
80.9(2) [N1–Ti1–N1′]
93.9(2) [N1–Ti2–N1′]
119.57(18) [N2–Ti1–N1]
Ti–NHamido–C
—
126.45(14)
105.01(7)
—
135.74(17) [C1–N1–Ti1]
136.48(16) [C13–N2–Ti1]
102.44(11) [N5–Ti2–N6]
—
—
NNMe2–Ti–NNMe2
103.09(5) [N3–Ti1–N4]
103.44(5) [N5–Ti2–N6]
103.87(9) [N3–Ti1–N4]
104.28(10) [N3–Ti1–N2]
105.19(9) [N4–Ti1–N2]
103.07(11) [N5–Ti2–N6]
104.40(10) [N5–Ti2–N7]
107.95(10) [N6–Ti2–N7]
—
—
NNMe2–Ti–Namido
—
—
107.35(7)
—
113.05(8) [N1–Ti1–N2]b
—
—
—
—
a
NL–Ti–NL
—
76.47(11) [N3–Ti2–N4]
a L = NHMe2 or Py. b
NNHAr*–Ti–NNHAr* angle.

View Article Online
Paper
Dalton Transactions
almost 50 years from the pioneering work of Bradley13 and sub-
sequent studies by Nugent who solved its structure,22 they are
currently less studied than terminal imido analogues. They
have often been obtained while seeking complexes with a term-
inal imido function.23 The synthesis of 2, as reported by
Bradley, is curiously poorly detailed but was obtained by react-
ing neat Ti(NMe2)4 with excess (no detail) boiling tBuNH2. A full
description of our procedure for the synthesis of 2 and its
characterization is available in the Experimental section.
Reaction of Ti(NMe2)4 with one equivalent of the silyl
amine Ph3SiNH2 proceeds in an analogous fashion as with
1-adamantylamine, and we isolated the expected dimer
[Ti(μ-NSiPh3)(NMe2)2]2 (3)24 as the major species, together with
a very minor amount of 4 (in ca. 5% by 1H NMR). Compound 4
had been selectively crystallized and identified by an X-ray
study to be the dimer [(Me2N)2Ti(μ-NSiPh3)2Ti(NMe2)-
(NHSiPh3)] with an additional –NHSiPh3 amido ligand repla-
cing one –NMe2 ligand in 3. The structure of 4 is given in the ESI‡
and will not be discussed here, as a related compound, 8, is
described later. The relatively straightforward preparation of 3
probably results from the stronger acidity of Ph3SiNH2 (vs.
tBuNH2) and the lower solubility of the product, which drives
the equilibrium forward, as with 1-adamantylamine.
Scheme 2 Formation of dimers [Ti(μ-NR)(NMe2)2]2 with alkyl- or silylamines.
In addition, in preliminary studies, we screened the reac-
tions between equimolar amounts of Ti(NMe2)4 and various
primary amines, in toluene or in C6D6 solutions. For example,
as depicted in Scheme 2, 1 equiv. of Ti(NMe2)4 was shown to
react differently depending on the alkylamines used. 1-Ada-
mantylamine reacts smoothly with Ti(NMe2)4 in toluene at RT
affording the expected unknown imido-bridged dimer [Ti(μ-N-
1-adamantyl)(NMe2)2]2 (1).19 A thermal ellipsoid plot of the
X-ray structure of 1 is given in Fig. 1. The coordination geometry
of titanium centers is distorted tetrahedral. The symmetric
Ti2N2 core is planar (torsion angle Ti1–N1–Ti1′–N1′ = 0°) and
is characterized by a short Ti⋯Ti distance [Ti1⋯Ti1′ = 2.8018(7)
Å]; such a short Ti⋯Ti distance is not unusual.5f,8,9,20 The
angle formed by the bridging nitrogen atoms with the two tita-
nium atoms [Ti1–N1–Ti1′ = 97.72(7)°] deviates significantly
from that expected for an sp2 nitrogen atom. The Ti–Nimido dis-
tances being comprised between 1.9121(15) and 1.9277(16) Å
are longer by ca. 0.2 Å than those in terminal imido complexes.
The (Me2N)2Ti moiety is normal with the expected Ti–
N bond distances and angles [Ti–Namido 1.9078(17) and
1.9087(18) Å].
The reactivity of different arylamines with Ti(NMe2)4 in a
1/1 ratio repeatedly produced crude mixtures of complexes.
Several previous studies have noted difficulties with the
Ti(NMe2)4/aniline systems that lead to black or dark, poorly
soluble and uncharacterized materials.4k,13,25 For example, in
our hands, aniline PhNH2 afforded mixtures of complexes (by
NMR), none of which is the expected dimer [Ti(NPh)-
t
By contrast, BuNH2 failed to react with Ti(NMe2)4 under
identical conditions (RT) or at 50–60 °C. It was only when it
was heated at 100 °C for 1.5 hours that we observed a color
change to red indicative of the formation of [Ti(μ-NtBu)-
(NMe2)2]2 (2). While the pKa value and the steric bulk of both
1
(NMe2)2]2.26 The unexpected complex H NMR spectrum (with
t
1-adamantylamine and BuNH2 are very similar,16 we explain
5 different –NMe2 signals) might be indicative of compounds
of higher nuclearity. Indeed we could crystallize out of the
mixture one of these species (5) and we determined its struc-
ture by X-ray crystallography (Fig. 2). The compound is a
t
this difference of reactivity (1-adamantylamine vs. BuNH2) to
be mainly due to the difference in the solubility of the product
formed: the 1-adamantyl substituent decreases the solubility,
which drives the equilibrium of the transamination reaction
toward the formation of less soluble 1. Although imido-
bridged group 4 metal complexes of this type (and in particular
t
this butyl-imido bridged derivative 2) have been known for
Fig. 1 Molecular structure of 1, showing 30% probability ellipsoids and partial
Fig. 2 Molecular structure of 5, showing 30% probability ellipsoids and partial
atom-labeling schemes. H atoms are omitted for clarity.
atom-labeling schemes. H atoms are omitted for clarity.
12206 | Dalton Trans., 2013, 42, 12203–12219
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
trimer formulated as [Ti3(μ2-NPh)3(μ3-NPh)(NMe2)4(NHMe2)]. It offers steric protection, good solubility in hydrocarbons, and
is composed of three titanium centers bridged by three μ-NPh an easy 1H NMR probe for the imido linkage.
ligands [Ti–Nμ-imido ranging from ca. 1.87 to 2.07 Å]. The tita-
If conducted on a 20–100 mg metal precursor scale, the reac-
nium centers are further coordinated by a μ3-NPh [Ti–Nμ3-imido tion of one equiv. of Ti(NMe2)4 with one equiv. of Ar*NH2, in
1.8622(14), 2.4668(14), and 2.1024(15) Å, respectively for Ti1, toluene or benzene-d6 at room temperature, affords a yellow
Ti2, and Ti3], but are all non-equivalent, as the remaining co- solution that turns orange or red-orange very slowly. However,
ordinated ligands differ: one is a tetracoordinated Ti with by monitoring the reaction by 1H NMR spectroscopy over a
coordination to only one –NMe2 [Ti–NNMe2 1.9071(17) Å], the period of 15 minutes to 3 days, we noticed that: (i) the compo-
second Ti atom is pentacoordinated (two –NMe2, [Ti–NNMe2 sition of the mixture does not significantly evolve after about
1.9044(15) and 1.9030(16) Å]), and the third pentacoordinate 1 hour, and (ii) the ‘final’ mixture contains mainly unreacted
Ti atom is linked to one –NMe2 [Ti–NNMe2 1.8976(15) Å] and Ti(NMe2)4 and Ar*NH2, HNMe2, and multiple minor species.
one NHMe2 amino ligand [Ti–NNHMe2 2.2407(16) Å]. During At least three species with μ-NAr* and one with terminal
the course of our study, a related tetranuclear species formu- TivNAr* were identified by the characteristic δ region of the
lated as [Ti4(μ-NPh)5(NMe2)6] (6) was characterized by Paine27 CHiPr substituents (that we use as a good probe for dis-
while combining a 1 : 1 ratio of Ti(NMe2)4 and PhNH2 under tinguishing bridging vs. terminal imido). The presence of lib-
slightly different conditions. Comparing the 1H NMR signals erated HNMe2 proves that transamination has occurred, but
of our crude reaction products with those of Paine’s revealed the low integration29 and the fact that reagents are still the
that before crystallization we had both 5 and 6 (in a ratio of major species in solution tend to prove that the transformation
ca. 3 : 1) (Scheme 3).
is reversible and that the equilibrium is not in favor of the
As exemplified by these reactions with different primary transamination products. Heating the solution did not signifi-
amines, and in particular with aniline,28 the anticipated very cantly change the course of the reaction. Compounds like
simple reaction between equimolar amounts of the two com- neutral four-coordinate bis-anilido-Ti A or mono-anilido-Ti B
ponents (Ti(NMe2)4 and a primary amine) is obviously not a noted previously in Scheme 1 may have been formed, although
straightforward reaction: although in the case of alkyl- or silyl- we have no spectroscopic evidence for their formation.
amines the expected dimers are produced, the reaction can
We further noticed that removing the volatiles (in the pre-
yield complex mixtures of products, some of which are much vious solution) under vacuum afforded an oily residue that,
more complicated than initially expected, and were shown to upon dissolution in C6D6, did not present the same spectro-
be strongly dependent on the bulk, acidity but also solubility scopic features. The 1H NMR spectrum confirmed that most of
of the amine used.
the HNMe2 had been removed by vacuum pumping, but also
showed that the composition of the mixture had evolved (the
ratio between all species), with now only about 25% of the
Ti(NMe2)4 unreacted. Finally, the residue was treated by several
cycles (typically 3–5)30 of pumping under vacuum followed by
addition of pentane. Upon treatment the deep red residue
became less and less soluble in pentane, and (by monitoring
each cycle by NMR) we observed that all of the Ti(NMe2)4 was
progressively transformed. The final product contained two
products, which were identified as the dimers [Ti(μ-NAr*)-
(NMe2)2]2 (7) and [(Me2N)2Ti(μ-NAr*)2Ti(NMe2)(NHAr*)] (8), in
2.2 Reactivity of Ti(NMe2)4 and Ar*NH2 in a 1 : 1 ratio.
Towards the synthesis of [Ti(μ-NAr*)(NMe2)2]2
In a very recent article, Mindiola et al. pointed out that
“Ti(NMe2)4 and various anilines failed to produce any isolable
products”.23d These observations were quite puzzling, in par-
ticular in regard to the very selective and high yield way by
which complexes Ti(vNAr)Cl2(NHMe2)2 are obtained using
our related one-pot, three-component procedure that com-
bines Ti(NMe2)4, ArNH2, and Me3SiCl (including when using
PhNH2),9 or in apparent contradiction with our recent report
on the easy and selective formation of unique dimer complexes
with two distinct bridging μ-imido ligands, as well as hetero-
bimetallic species bridged by μ-imido ligands.20 Intrigued by
these observations, we decided to further investigate the trans-
amination reaction of Ti(NMe2)4 with anilines, and we chose
2,6-iPr-aniline (now abbreviated as Ar*NH2) as the model
amine. Ar*NH2 is very often used in imido chemistry, as it
1
respectively ca. 8 : 2 ratio (Scheme 4). In its H NMR spectrum,
7 exhibits a singlet at 2.99 ppm for the four equivalent –NMe2
ligands, whereas the three –NMe2 ligands in 8 are all inequiva-
lent, thus giving rise to three singlets (3.05, 2.91, 2.43 ppm).
In consequence, the reaction leading to 7 and/or 8 is a pre-
cipitation (or crystallization) driven equilibrium that is also
very much dependent on the initial elimination of NHMe2.
The difficulty in obtaining 7 free of 8 could originate from: (i)
kinetic reasons if A is formed more slowly than it reacts sub-
sequently with Ar*NH2 to form anilido 8 or (ii) thermodynamic
reasons if multiple species coexist in solution and recombine
to form the more stable dimer 8 (Scheme 4).31
A larger scale reaction in toluene (using 100 mg–2.00 g
Ti(NMe2)4) confirmed the previous observations, with the for-
mation of an oily residue that upon repeated similar treatment
(vacuum and pentane) produced 7 and 8 dimers similarly,
generally as a sticky dark red solid. Although selective
Scheme 3 Formation of two distinct complexes from Ti(NMe2)4 and PhNH2
used in a 1 : 1 molar ratio.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 12203–12219 | 12207

View Article Online
Paper
Dalton Transactions
Scheme 4 Syntheses of dimers 7 and 8, and proposed routes to explain the formation of 8. Molecular structure of 7 (30% probability ellipsoids).
recrystallization in a cold pentane solution could afford both 7 bridging two (Me2N)2Ti moieties. The coordination geometry
and 8 separately, we looked for different experimental con- of both titanium centers is distorted tetrahedral (although Ti1
ditions that would allow selective formation of 7 (or 8, see and Ti2 may be regarded as 5-coordinate, see below). The
below), and would also avoid the above treatment with Ti2N2 core deviates from planarity (torsion angle Ti1–N1–Ti2–
pentane and vacuum.
N2 = 5.85(4)°), and is characterized by a short Ti⋯Ti distance
Switching solvent from benzene-d6 or toluene to pentane,32 of 2.8089(3) Å common in imido dimers.5f,8,9,20,22b The angles
or conducting the reaction at different temperatures in toluene formed by the bridging nitrogen atoms with the two titanium
or in dilute solutions, did not really improve the reaction atoms [Ti1–N1–Ti2 = 92.36(4)°, Ti1–N2–Ti2 = 92.27(4)°] deviate
selectivity. Conducting the reaction in excess of Ti(NMe2)4 to significantly from that expected for an sp2 nitrogen atom. The
limit the formation of anilido 8 met with limited success (as it (Me2N)2Ti moiety is normal, with the expected Ti–N bond dis-
tends to form more of a new species 9 as discussed later). tances and angles [Ti–Namido ranging from 1.889 to 1.922 Å].
Nevertheless, we noticed that the selectivity in 7 was slightly The Ti2N2 core displays considerable asymmetry in its Ti–N–Ti
higher while conducting the reaction between Ti(NMe2)4 and bridges, both in the Ti–N distances [Ti1–N1 = 2.0211(10) Å,
Ar*NH2 in a Schlenk flask at 80 °C under solventless con- Ti2–N1 = 1.8694(10) Å, Ti1–N2 = 1.8818(11) Å, Ti2–N2 = 2.0123(11)
ditions and under a stream of argon or by pumping off the Å] and in the distortion which brings the aryl-imido Cipso
NHMe2 vapors that were being formed. Again, the argon to within η2-bonding distance of Ti atoms. The μ-[η1(N):
stream or vacuum pumping steps presumably shift the equili- η2(N,C)] bonding mode exhibited by the μ-arylimido ligands
brium of transamination towards the expected dimer by includes the following structural features: (i) Ti–C distance of
removing the NHMe2. During the reaction, the initially liquid 2.8521(11) Å (Ti1–C1) and 2.8783(12) Å (Ti2–C13) (this feature
mixture solidifies. Further extraction of this red-purple solid is however less pronounced than in our previously reported
with toluene affords 7 in 79% yield.
dimer with two distinct imido ligands {Ti2(μ-NAriPr2)( μ-NAda)-
Finally, we found a totally distinct synthesis to access 7 in a (NMe2)4} that has a Ti–Caryl distance of 2.6443(19) Å), and (ii)
pure form and without the need of a recrystallization step. The the shorter Ti–Nimido distance of 1.8694(10) Å is longer by ca.
synthesis, illustrated in Scheme 4, is based on the deprotona- 0.15 Å than in terminal imido complexes,9,33 (iii) the nearly
tion of NHMe2 ligands in Ti(vNAr*)Cl2(NHMe2)2. Treating a linear Ti2–N1–C1 and Ti1–N2–C3 bond angles of 152.99(8)°
solution of Ti(vNAr*)Cl2(NHMe2)2 with two equivalents of and 152.28(9)° lead to acute Ti1–N1–C1 and Ti2–N2–C13 of
KN(SiMe3)2 affords selectively 7 in 86% yield as a dark red 111.28(8)° and 113.34(8)°, and contrasts with the more sym-
microcrystalline material. This alternative synthesis might be metric M–Nimido–M bridge found in other related sys-
of significant importance for primary amines such as PhNH2 tems.4k,5f,8,20,22b Although unusual, this type of μ-[η1(N):
for which the desired [Ti(μ-NR)(NMe2)2]2 dimer cannot be η2(N,C)] bonding mode for an μ-arylimido ligand is not unprece-
obtained directly from Ti(NMe2)4.
dented,20,34 and it was proposed that compounds with such
The molecular structure of 7 (see the figure in Scheme 4) features would constitute a relevant structural model for
clearly demonstrates the compound to be the dimer formu- industrial hydrodenitrogenation catalysis (removal of nitrogen
lated as [Ti(μ-NAr*)(NMe2)2]2 with two μ-NAr* imido ligands compounds contaminating petroleum and coal): the
12208 | Dalton Trans., 2013, 42, 12203–12219
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
μ-[η1(N):η2(N,C)]-NAr ligand could represent a model sub-
strate–catalyst interaction on the way to C–N bond scission in
the hydrodenitrogenation of anilines.34
2.3 Reactivity of Ti(NMe2)4 and Ar*NH2 at higher
stoichiometry
In the Ti(NMe2)4/Ar*NH2 (1 : 1) preparative reaction described
above, we isolated two dimers: 7 and 8, and in solution other
species could be detected by 1H NMR, in particular at the early
stage of the reaction and when Ti(NMe2)4 was used in excess
(see above). We decided to study the reaction by varying the
stoichiometry of the reactants in order to: (i) determine if these
complexes (8 and 9) could be obtained in a pure form for full
characterization, (ii) study if other species were present in the
reaction medium by varying the stoichiometry of the reactants,
and (iii) try to correlate to the mechanism of the reaction.
Fig. 3 Molecular structure of 9, showing 30% probability ellipsoids and partial
atom-labeling schemes. H atoms are omitted for clarity.
2.3.1 Reactivity of Ti(NMe2)4 and Ar*NH2 in a 2 : 1 ratio.
Treatment of two equivalents of Ti(NMe2)4 with only one equi-
valent of Ar*NH2 in toluene afforded a red-orange solution
As illustrated in Scheme 5, we propose the intermediate for-
(Scheme 5), in contrast to the darker solution obtained when mation of a transient mono-amido (–NHAr*) species B, which
an equimolar ratio of reactants is used, and from which after would rapidly react with a second molecule of Ti(NMe2)4 to
work-up a new complex 9 could be obtained as an orange solid give 9. Importantly, dimer 9 slowly evolves in solution to
1
in 80% yield. The H NMR spectrum of 9 indicated: (i) a rela- reform Ti(NMe2)4 and 7 (Scheme 5). At equilibrium (about 10
tive ratio between Ar*N and –NMe2 ligands of 1 : 6 in accord- days at room temperature), the composition follows: 9 : Ti-
ance with the stoichiometry of the reactants (if transamination (NMe2)4 : 7 = 2 : 2 : 1 (4 : 6 : 3 after 24 h). The reaction is revers-
i
has occurred). In addition, signals of the Pr substituents were ible, giving back 9 upon low temperature crystallization in
found in the region of the bridging imido NAr* group pentane or by slow evaporation of the solvent. A reexamination
1
(3.76 ppm). X-ray quality crystals of 9 were obtained. A thermal of the reaction leading to 9 in solution (monitored by H NMR
ellipsoid plot is presented in Fig. 3. Metric parameters are in a C6D6 solution) confirmed these findings and showed 7 to
summarized in Table 2 to facilitate a comparison between appear after about 1–2 hours of reaction and to increase until
members of the same family of complexes.
representing only 10% after 14 hours. After several days (13
The X-ray crystallographic study revealed 9 to be a new dimer days), at equilibrium, unreacted Ti(NMe2)4 was still present at
formulated as [Ti(NMe2)3]2(μ-NAr*). It is composed of two about 30%, with 10 and 30% of 7 and 9, respectively, and an
{(Me2N)3Ti} units held together by an extremely rare (in tita- appreciable amount of 8 (30%). (The thermal reaction leading
nium and its group 4 congeners)35,36 single bridging NAr* to 7 was already shown to form more 8; see above.) Unfortu-
motif. Both Ti atoms are tetra-coordinated and separated by a nately, but not surprisingly, the formation of 9 is (like 7) influ-
short Ti⋯Ti distance [Ti1⋯Ti2 = 3.621 Å] and two Ti–Nimido enced by the elimination of NHMe2, and can further react with
bond distances of 1.9522(18) Å and 1.9602(19) Å. These Ti– Ar*NH2 to give 7 and 8. Therefore a clean preparation of 9 is
Nimido distances are only slightly longer by ca. 0.02 Å than the only possible under crystallization driven equilibrium con-
mean value observed for bridging imido in 7. The (Me2N)3Ti ditions. Alternatively, as the reaction is reversible, we verified
moiety is normal and displays the expected Ti–N bond distances that 9 can be obtained from Ti(NMe2)4 and 7. However, in that
and angles [Ti–Namido ranging from 1.888(2) to 1.911(2) Å].
case even under crystallization-driven equilibrium conditions
Scheme 5 Formation of dimer 9 from Ti(NMe2)4 and Ar*NH2 in a 2 : 1 ratio, and suggested intermediates.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 12203–12219 | 12209

View Article Online
Paper
Dalton Transactions
the equilibrium seems to favor 7 over 9 (we obtained a mixture
of approximate composition 7 : Ti(NMe2)4 : 9 = 1 : 2 : 1).
2.3.2 Reactivity of Ti(NMe2)4 and Ar*NH2 in a 1 : y ratio
(y > 1). Using now a Ti(NMe2)4/Ar*NH2 ratio of 2 : 3 (see
Scheme 6), i.e. an appropriate stoichiometry for getting anilido
dimer 8, in toluene at 100 °C indeed cleanly afforded com-
pound 8 as red-purple crystals in 91% yield. The molecular
structure of 8 resembles that of dimer 7, with an amido
–NHAr* ligand in place of an –NMe2 group (see Fig. 4). A
second difference comes from the μ-NAr* bridges: one μ-NAr*
bridge (Ti1–N2–Ti2) displays a considerable asymmetry, both
in the Ti–N distances [Ti1–N2 = 1.8485(15) Å, Ti2–N2 = 2.0139(15)
Å] and in the distortion which brings the aryl-imido Cipso
to within η2-bonding distance of Ti2 (as already observed in
both bridges of 7). The μ-[η1(N):η2(N,C)] bonding mode exhibi-
ted by the μ-arylimido ligand includes the following structural
features: a short Ti2–C13 distance of 2.7014(19) Å, (ii) Ti–
Nimido distances of 1.8485(15) Å and 2.0139(15) Å, and (iii) the
nearly linear Ti1–N2–C13 bond angle of 158.65(14)° leads to an
acute Ti2–N2–C13 of 102.80(12)°. By contrast, the second μ-NAr*
bridge (Ti1–N1–Ti2) with the more symmetric Ti–NAr*–Ti bridge
exhibits Ti–N–C angles of 137.15(12) and 130.17(13)° associated
Fig. 4 Molecular structure of 8, showing 30% probability ellipsoids and partial
atom-labeling schemes. Hydrogen atoms are omitted for clarity (except for
NHAr*).
Next we examined the possibility to prepare dimer [Ti-
with Ti–NAr* bond distances of 1.9173(16) and 1.9626(15), (μ-NAr*)(NMe2)(NHAr*)]2 (C) with two-symmetry related
respectively, for Ti1 and Ti2. The Ti–NHAr* bond is characterized –NHAr* anilido ligands (or its less likely isomers) by increas-
by a distance of 1.9442(17) Å and a Ti–N–CAr* angle of 126.4°.
ing the stoichiometry of Ar*NH2 to 2 equiv. (vs. 1 equiv.
Scheme 6 Reactivity of Ti(NMe2)4 and Ar*NH2 at higher stoichiometry, and tautomeric equilibrium between 8 and 11.
12210 | Dalton Trans., 2013, 42, 12203–12219
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Ti(NMe2)4) (Scheme 6). Instead of giving the expected C, the –NHAr* in 8 jumping to the N atom of one –NMe2 ligand
room temperature reaction led only to 8 and unreacted (prototropy) in 11, thus generating the terminal imido and the di-
Ar*NH2. After heating at 100 °C, we noted the slow but incom- methylamine ligands (Scheme 6).37 NMR studies showed both
plete transformation into a new species (11, see later). The molecules to be in equilibrium, and dissolving crystals of 11 in
reasons for the difficulty to form C are unclear, in particular C6D6 showed the complex to slowly (2 days) transform almost
since we have been able to obtain related vanadium dimers (such completely in favor of 8. At present, the mechanism of the
as [V(μ-NArCl2)(NMe2)(NHArCl2)]2 (ArCl2 = o-Cl2–C6H3), the struc- tautomerization is unclear.38 Furthermore, on dissolution
ture of which is given in ESI‡). Later in the text we will however in CD2Cl2 (but not in C6D6), 8 affords an equilibrium mixture
see that one isomer of C, namely 10, was indeed detected and of both tautomers 8 and 11 in a ca. 3 : 1 ratio after 14 h (6 : 1
isolated in reactions with a higher Ar*NH2/Ti(NMe2)4 ratio.
Increasing the Ar*NH2 : Ti(NMe2)4 ratio to 3 : 1, in order to
after 1 h).
11 was later observed in reactions of Ti(NMe2)4/Ar*NH2 in
force the formation of C or eventually to substitute more various ratios (from 1 : 1 to 1 : 5, see also below) and various
–NMe2 ligands and get new dimers such as the fully substi- experimental conditions (RT to 120 °C), but the common
tuted [Ti(μ-NAr*)(NHAr*)2]2 (D) (or its bis-dimethylamine point of all these syntheses was that the reactions should be
adduct monomeric counterpart E), produces mixtures of com- performed in a closed vial. The closed vial reaction retains
plexes at room temperature and at 100 °C. Among these NHMe2 in solution, therefore avoiding further Ti–NMe2 substi-
species, we could selectively crystallize and identify two new tution by Ar*NH2. Moreover, we suspect that NHMe2 is likely
complexes, namely 11 and 12 (Scheme 6).
involved in the prototropic rearrangement of 8 to 11. Finally,
The first complex, 11, is easily separated from the reaction 11 was also obtained by addition of 4 equiv. of Ar*NH2 to 7.
medium owing to its strong insolubility in pentane. Com- Notably, to the best of our knowledge, 11 is the first example
pound 11 exhibits in its 1H NMR spectrum signals corres- of a dinuclear titanium complex with both bridging and termi-
ponding to both terminal and bridging imido functionalities. nal imido ligands.39
An X-ray structure determination revealed 11 to be the dimer of
The second compound (12) identified in the previous reac-
formula [(Me2N)(Me2NH)Ti(μ-NAr*)2Ti(NMe2)(vNAr*)] (Fig. 5). tion (Scheme 6) was also unambiguously characterized by an
It is composed of two different fragments bridged by two X-ray study to be the homoleptic tetraanilido complex Ti-
μ-imido ligands: one with an extremely long Ti–N bond dis- (NHAr*)4. This result was rather unexpected due to the extreme
tance [Ti1–N1 2.074(4) Å] and the other in the normal range bulkiness of the four –NHAr* groups, which we thought
[Ti2–N1 1.841(4) Å]. In the first fragment the Ti center (Ti1) is would, more likely, have eliminated one Ar*NH2 to form either
coordinated to one terminal imido ligand [Ti1–N2 1.723(6) Å, the dimer D or its monomer [Ti(NAr*)(NHAr*)2] (F). An ORTEP
Ti1–N2–C13 168.2(5)°] and one –NMe2 ligand [Ti1–N5 1.918(6) drawing is presented in Fig. 6 together with a space filling view
Å], whereas in the second fragment the metal center is that gives a better idea of the steric hindrance around the
bounded to one –NMe2 ligand [Ti2–N3 1.875(6) Å] and one di- metal center. The tetrahedral Ti center is surrounded by four
methylamine molecule [Ti2–N4 2.185(6) Å]. It appears that closely imbricated Ar*NH– ligands, with Ti–N bond distances
complex 11 is a tautomer of 8, with the H atom of the anilido of 1.9028(15) Å (Ti1–N1) and 1.8962(15) Å (Ti1–N2) and associ-
ated Ti–N–C angles of respectively 137.49(12)° and 136.41(13)°.
Large angles are typically observed with bulky amido substitu-
ents like tris-tbutylsilylamido.40 There is also a distortion away
from tetrahedral geometry with N–Ti–N angles varying from
105.71(10)° to 112.67(7)°. The parent homoleptic Zr(NHAr*)4
had been reported two decades ago by treating Zr(CH2Ph)4
with 4 equivalents of Ar*NH2.41
Considering the size of the substituent on the imido nitro-
gen, and the much smaller Ti radius, isolation of the titanium
analogue 12 is remarkable and prompted us to determine
whether 12 could be prepared selectively. The attempted syn-
thesis of 12 by treating Ti(NMe2)4 with 4 or 5 equivalents of
Ar*NH2 (RT and 100 °C, in toluene or under solventless con-
ditions) unfortunately led to mixtures of complexes: namely
12, 13, and depending on the conditions an isomer of C (10)
was also detected and isolated as a minor compound. Among
these complexes could be identified the new complex 13,
which according to NMR spectroscopic studies is the dimer
formulated as [(Me2N)(Ar*HN)Ti(μ-NAr*)2Ti(NHAr*)2] (Scheme 6).
The presence of a compound like 13, with incomplete
Fig. 5 Molecular structure of 11, showing 30% probability ellipsoids and
–NMe2 substitution and bridging imido, demonstrates both
the extreme difficulty of substituting all –NMe2 due to the
partial atom-labeling schemes. Hydrogen atoms and solvate molecules (two
C6H6) are omitted for clarity (except for NHMe2 ligand).
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 12203–12219 | 12211

View Article Online
Paper
Dalton Transactions
Fig. 7 Molecular structure of 10, showing 30% probability ellipsoids and
partial atom-labeling schemes. Hydrogen atoms are omitted for clarity (except
for NHAr* groups).
–NMe2 ligands on the second Ti center (Fig. 7). Interestingly,
the regiochemistry of the formation of 10 might suggest that it
results from the transamination reaction between two –NHAr*
anilido ligands in tetra-anilido monomeric 12 and Ti(NMe2)4
(instead of the reaction of 8 or 11 with Ar*NH2, see also ref.
42). The overall molecular structure and bond parameters of
10 are very similar to that of parent mono-anilido dimer 8 (see
selected bond distances and angles in Table 2).
Fig. 6 Molecular structure of 12 (top: Ortep drawing, down: spacefilling view,
with nitrogen atoms in blue and titanium in magenta) showing 30% probability
ellipsoids and partial atom-labeling schemes. Hydrogen atoms are omitted for
clarity (except for NHAr* groups).
2.4 Reaction of Ti(NMe2)4 with Ar*NH2 in the presence of
pyridine
steric constraints imposed by the bulky ligands, and the high As detailed in the previous sections, substitution of –NMe2
instability of Ti–NHAr* containing species, which are easily ligands in Ti(NMe2)4 by treating with Ar*NH2 in ratios varying
prone to α-H abstraction and generation of μ-imido complexes. from 1 : 0.5 to 1 : 5 afforded generally equilibrium mixtures of
Indeed, when placed in solution, crystals of 12 are often con- imido and amido complexes. In certain cases, using crystalliza-
taminated by Ar*NH2, which confirms the α-H abstraction tion or precipitation techniques, we have been able to shift the
with concomitant formation of elusive [Ti(vNAr*)(NHAr*)2] equilibrium and favor the synthesis of single compounds. In
(F) (or its dimer D or its aniline adduct [Ti(vNAr*)(NHAr*)2- other studies, however, it appeared extremely difficult to
(NHAr*)] (G)). The latter, however, could not be clearly identified control the reaction towards selective formation of single mole-
due to possible exchange between all species in solution and cules. In an effort to shift the equilibrium we used donor sol-
also with free Ar*NH2, as in elusive G (Scheme 6).
As noted above, the reaction also produced small amounts donor atom coordination to the metal center, and hence help
new compound that was initially spectroscopically in trapping key intermediates. For that purpose, we studied
vents or molecules that can favor both NHMe2 elimination and
of
a
assigned as being the trans-dianilido compound C. Moreover, the effect of the addition of selected donors such as THF,
the NMR studies suggested that only one isomer (of the 3 possi- NHMe2, bis-tBu-bipyridine and pyridine on the reaction of Ti-
ble isomers of C) was present in solution, even at 193 K. Selec- (NMe2)4/Ar*NH2 or on isolated dimers (7, 8). Among the
tive crystallization from cold pentane solutions afforded low donors tested, only pyridine gave clear observations (or iso-
yields40 of crystals suitable for an X-ray diffraction study. Unex- lated products) that are reported here.
pectedly, the study revealed unequivocally the compound to be
Unfortunately, no intermediate could be trapped by
the new dimer formulated as [(Ar*HN)2Ti(μ-NAr*)2Ti(NMe2)2] addition of pyridine (6 equiv.) to the reaction of Ti(NMe2)4
(10) (Scheme 6), i.e. a less likely isomer of the expected C due with 1 equivalent of Ar*NH2. However, while monitoring the
to steric constraints. Instead of having the two –NHAr* anilido addition of pyridine (1–8 equiv.) to 7 by 1H NMR spectroscopy,
ligands located in a trans-position on the two Ti centers, the we observed at room temperature the presence of 7 and broad
anilido ligands in 10 are on the same metal center, with two signals attributed to a new pyridine adduct [(Me2N)2Ti-
12212 | Dalton Trans., 2013, 42, 12203–12219
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Scheme 7 Reaction of 7 in the presence of pyridine.
(μ-NAr*)2(NMe2)2(py)] (14) and/or its monomer [Ti(vNAr*)-
(NMe2)2(py)2] (H) (Scheme 7). Both H and 14 might be present
in equilibrium with excess pyridine. Attempts to isolate the
adducts H or 14 by precipitation or under vacuum failed, with
reformation of 7. However, a VT NMR study in toluene-d8
demonstrated that at low temperature (193 K) in the presence
Fig. 8 Molecular structure of 15, showing 30% probability ellipsoids and
partial atom-labeling schemes. H atoms are omitted for clarity.
of excess pyridine dimer, 7 is transformed reversibly into a parameters (Table 2). Compound 15, like 11, is composed of
mono-pyridine adduct 14 (see the Experimental section). two different fragments bridged by two asymmetric μ-imido
Under thermal conditions, a slow transformation into new ligands [Ti1–N1 2.070(3) Å, Ti1–N2 2.053(3) Å, Ti2–N1 1.861(3)
compounds was observed, including Ti(NMe2)4 and a new Å, Ti2–N2 1.876(3) Å], the two μ-NAr* substituents being bent
complex 15 that crystallized out of the C6D6 or toluene solu- toward Ti1 [Ti1–C1 2.951 Å, Ti1–C13 3.011 Å]. In the first frag-
tion, and that, upon being subjected to vacuum, regenerated ment the Ti center (Ti1) is coordinated to one terminal imido
starting dimer 7. Single crystals of only one compound (15) ligand [Ti1–N5 1.758(3) Å ] and one –NMe2 ligand [Ti1–N6
were obtained from the solution, and were suitable for struc- 1.923(3) Å], whereas in the second fragment the metal center
tural characterization by X-ray crystallography.
is bound to one –NMe2 ligand [Ti2–N7 1.925(3) Å] and two pyri-
The structure is presented in Fig. 8 and revealed 15 to be dine molecules [Ti–N 2.359(3) and 2.240(3) Å].
the dimer complex [(Me2N)(py)2Ti(μ-NAr*)2Ti(NMe2)(vNAr*)].
As illustrated in Scheme 8, addition of pyridine (2–8 equiv.)
Compound 15 is closely related to 11, but with two pyridine to a solution of 8 in C6D6 (or toluene) was found to lead to the
ligands instead of one dimethylamine. We suggest 15 to be rapid (<1 hour) formation of a new species that crystallized
generated through the intermediacy of H (Scheme 7) and spontaneously from the solution. This compound was shown
1
further recombination with a fragment of 7 (or its pyridine by H NMR spectroscopy to be 15. The equilibrium is reversi-
adduct 14) with elimination of Ti(NMe2)4.
ble under vacuum (elimination of pyridine) with reformation
The molecular structure of 15 (Fig. 8) shows strong simi- of 8. As a difference from the previous observation of 15 from
larities to that of 11, with the exception that one titanium 7, here the pyridines should probably shift the tautomeric 8/11
atom is now pentacoordinate (Ti2) due to the coordination of equilibrium in favor of 11, which upon NHMe2 substitution by
two pyridines, which has an impact on the geometric py affords 15.
Scheme 8 Reactions of Ti(NMe2)4 and excess Ar*NH2 in the presence of pyridine, synthesis of 15–17.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 12203–12219 | 12213

View Article Online
Paper
Dalton Transactions
By contrast, working in the presence of py (4 and 8 equiv.) triphenylsilylamine. With Ar*NH2, if NHMe2 is not released
while conducting the reaction of 2 equivalents of Ti(NMe2)4 from the solution, the equilibrium is not favorable for the pro-
with 3 equiv. of Ar*NH2 (i.e., the right stoichiometry to form duction of a particular or single compound, and multiple
anilido-complex 8) led to less clean reaction. However, 15 was species were shown to coexist in solution (Ti(NMe2)4, 7–13).
still observed (and liberated NHMe2) (Scheme 8). When left for Nevertheless, using an appropriate and controlled stoichi-
longer periods (4 days) at room temperature, the reaction ometry of the reactants and experimental conditions, and by
evolved and we were able to crystallize a new species, namely applying a suitable physical or chemical treatment, we have
the monomeric complex Ti(vNAr*)(NHAr*)(NMe2)(py)2 (16). been able to synthesize, isolate and characterize a variety of
As shown by its solid-state structure (see ESI‡), 16 results from novel imido or amido complexes.
the cleavage of the dimer 8 by pyridine ligands. The other part
of the cleaved dimer (H), however, could not be isolated, and
A few other conclusions can be drawn from our studies:
(1) Dimers [Ti(μ-NR)(NMe2)2]2 are generally obtained with
we know from the reaction of 7 + pyridine that H is probably bulky alkyl- or silylamines such as tBuNH2, 1-adamantyl-
not stable under vacuum giving back 7.
amines, and Ph3SiNH2.
Attempts to synthesize 16 more rationally (i.e. with the stoi-
(2) Substitution of –NMe2 ligands in Ti(NMe2)4 by treating
chiometry of reactants Ti(NMe2)4 : Ar*NH2 = 1 : 2) in the pres- with Ar*NH2 in ratios varying from 2 : 1, 1 : 2 to 2 : 3 produced
ence of an excess of pyridine (8 equiv.) and at room dimers 7, 8 and 9. Compound 9 is the first example of a
temperature43 in toluene afforded an orange crystalline solid. Ti-dimer complex bridged by only one μ-imido ligand.
On the basis of elemental analysis, 1H NMR data, and X-ray
(3) Substitution of –NMe2 ligands in Ti(NMe2)4 by treating
data (the same cell parameters), we conclude that this com- with Ar*NH2 in ratios varying from 1 : 2 to 1 : 5 afforded gener-
pound is 16. ally mixtures of imido and amido complexes that appeared
When heated at 80 °C, the previous reaction (Ti(NMe2)4 : extremely difficult to control for selective formation of single
Ar*NH2 : py = 2 : 3 : 4 or 2 : 3 : 8) afforded less soluble species, molecules, even with the addition of a donor molecule (pyri-
among which the new complex Ti(vNAr*)(NHAr*)2(py)2 (17) dine). Nevertheless, several novel complexes were isolated and
could be identified by 1H NMR and isolated in 35% yield. characterized, including complexes of rare or unprecedented
Compound 17 was better prepared in 76% yield by treating structures (single bridged imido compounds and compounds
Ti(NMe2)4 with 3–5 equivalents of Ar*NH2 (Scheme 8), and its with both bridging and terminal imido linkages).
structure was unequivocally confirmed by an X-ray crystallo-
(4) The study highlighted an unprecedented prototropic
graphic study (see ESI‡) as well as by NMR spectroscopy. Com- tautomerism equilibrium between imido and amido functions
pound 17 is probably formed via pyridine coordination in (8/11).
elusive intermediates D or F, but we cannot exclude the inter-
(5) Dimer 7 was also cleanly synthesized by deprotonation
mediacy of 12. Indeed, an NMR tube reaction showed 12 to be of a coordinated dimethylamine ligand with KN(SiMe3)2 in the
transformed into 17 in the presence of added pyridine. During known Ti(vNAr*)Cl2(NHMe2)2.
the course of our studies compound 17 was independently pre-
This study helps us to better understand the mechanism by
pared by Nielson and Mountford,44 who used it as a precursor to which complexes Ti(vNAr)Cl2(NHMe2)2 are selectively obtained
bis- and tris-imido systems, and therefore 17 will not be described from our one-pot, three-component procedure that combines
here. The zirconium parent complex was prepared by Rothwell Ti(NMe2)4, ArNH2, and Me3SiCl.2h,k,8–10 In fact, in these systems,
from Zr(NHAr*)4 and Wigley from ZrCl4(THF)2/LiNHAr*.34,41a
Me3SiCl has a dual role that drives the equilibrium by substi-
tution of the –NMe2 ligands which generates a formally three-
coordinate titanium center that can capture liberated NHMe2
to form pentacoordinate Ti(vNR)Cl2(NHMe2)2.45
3. Summary and conclusion
In a similar way, our work also helps us to shed some light
A systematic study of the reaction between Ti(NMe2)4 and on our previous report, at first sight contrary to preliminary
Ar*NH2, in various ratios and under different experimental observations made in this manuscript, concerning an easy and
conditions, revealed the great complexity of these systems in selective formation of unique dimer complexes with two dis-
which multiple species coexist in solution, which presents tinct bridging μ-imido ligands, as well as heterobimetallic
difficulty for the isolation of single or well-defined compounds. species bridged by μ-imido ligands.20 As was demonstrated in
But it also demonstrates the possibility of great diversity and this article, these reactions are governed by precipitation or
richness of this chemistry if one can circumvent some of the crystallization of these less soluble dimers.
difficulties by specific physical or chemical treatments.
Finally, some of the observations made in this work might
The transamination reaction between Ti(NMe2)4 and a be useful for the understanding of the first step (transami-
primary amine appears to be very much dependent on impor- nation) in the mechanism of hydroamination of alkynes or
tant factors such as the nature of the amine (its basicity and alkenes with Ti(NMe2)4.4k,46 In these systems, the Ti-catalyst is
its steric demand), but more importantly, the equilibrium used in about 5 mol%. This means that at the beginning of
reaction is governed by the elimination of liberated NHMe2 the catalytic reaction (at a high amine/pre-catalyst ratio),
and/or crystallization (or precipitation) of the products with Ti(NHR)4 species might be the major species present in solution,
less soluble alkylamines (like 1-adamantylamine) or and in equilibrium with other likely imido species of the type
12214 | Dalton Trans., 2013, 42, 12203–12219
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Ti(vNR)(NHR)2Ln. However, with the amine substrate being solid. Yield: 89%. 1H NMR (300 MHz, C6D6): δ 3.40 (s, 24H,
consumed, the number and the nature of the species present NMe2), 1.20 (s, 18H, MetBu). 13C{1H} NMR (75.5 MHz, C6D6): δ
in the catalytic medium would certainly evolve during the 69.2 (CMe3), 45.5 (NMe2), 34.2 (CMe3). Anal. Calcd for
course of the reaction which may impact the selectivity of the C16H42N6Ti2 (414.28): C, 46.39; H, 10.22; N, 20.29. Found: C,
reaction (i.e. the Markovnikov/anti-Markovnikov ratio), a point 46.34; H, 10.20; N, 20.21.
that has, however, not been reported in these systems and that
will be investigated in our group.
[Ti(μ-NAr*)(NMe2)2]2 (7). From Ti(NMe2)4 under solventless
reaction. Schlenk tube with 500 mg of Ti(NMe2)4
A
(2.230 mmol) and 395 mg Ar*NH2 (2.228 mmol) was heated at
100 °C under a very slow stream of argon for 1 hour (via a
needle connected to a bubbler). The neat liquid solidifies
within about 20–30 minutes. The solid was vacuum dried at
100 °C for 30 minutes, and extracted with toluene (3 × 1 mL).
The toluene solution was filtered through celite, and the
solvent was removed under vacuum, to afford 565 mg of 7
(yield 81%). The solid generally contained less than 1–2% 8
(that can be removed by recrystallization in pentane or
toluene–pentane solutions at −20 °C).
4. Experimental section
This section contains only a few representative procedures,
and a complete description of all complexes (syntheses and
full characterization) and reactivity can be found in ESI.‡
General methods and instrumentation
All manipulations were carried out using standard Schlenk
From Ti(vNAr*)Cl2(NHMe2)2). To a toluene solution (2 mL)
line or dry box techniques under an atmosphere of argon. Sol- of 100 mg of Ti(vNAr*)Cl2(NHMe2)2 (0.2603 mmol) were
vents were refluxed and dried over appropriate drying agents added by portions equiv. of KN(SiMe3)2 (103.8 mg,
2
under an atmosphere of argon and collected by distillation. 0.5204 mmol) at room temperature. The resulting red solution
NMR spectra were recorded on Bruker DPX300 and Avance500 was stirred for 2 hours at RT. The volatiles were pumped off,
spectrometers. Chemical shifts are expressed in δ (ppm). and the dark red-purple residue was extracted with pentane (3
Elemental analyses were performed at the Laboratoire de × 2 mL). The pentane solution was dried under vacuum to give
1
Chimie de Coordination/CNRS (Toulouse, France) (C, H, N).
70 mg of crystals of 7 (86%). H NMR (300 MHz, C6D6): δ 7.21
The Ti(NMe2)4 used in this study was prepared by modifi- (d, 3J = 7.8 Hz, 4H, C6H3Pri2), 7.07 (t, 3J = 7.6 Hz, 2H, C6H3Pri2),
cation of a literature procedure,47 or was purchased from com- 4.05 (sept, J = 6.8 Hz, 4H, CHMe2), 2.99 (s, 24H, NMe2), 1.32
3
mercial sources (Aldrich). V(NMe2)4 was prepared by (d, 3J = 6.7 Hz, 24H, CHMe2). 13C{1H} NMR (75.5 MHz, C6D6): δ
modification of a literature procedure.14 Trimethylchlorosilane 152.3 (ipso-C6H3), 140.1 (o-C6H3), 123.7 (p-C6H3), 123.4 (m-
was distilled and stored over 4 Å molecular sieves under argon C6H3), 45.1 (NMe2), 28.1 (CHMe2), 24.8 (CHMe2). Anal. Calcd
before use. Ar*NH2 was stirred over KOH, distilled, dried over for C32H58N6Ti2 (622.58): C, 61.73; H, 9.39; N, 13.50. Found: C,
CaH2, redistilled, and stored over 4 Å molecular sieves under 61.67; H, 9.30; N, 13.40.
argon before use in the glove box. [Ti(vNAr*)Cl2(NHMe2)2]
[(Me2N)2Ti(μ-NAr*)2Ti(NMe2)(NHAr*)] (8). A Schlenk tube
with 200 mg of Ti(NMe2)4 (0.892 mmol), 237 mg Ar*NH2
was prepared according to our published procedure.9
[Ti(μ-N-1-adamantyl)(NMe2)2]2 (1). To a toluene solution (1.337 mmol), and 4 mL of toluene was heated for 1 hour at
(3 mL) of 500 mg of Ti(NMe2)4 (2.230 mmol) was added 1-ada- 100 °C under argon. The volatiles were removed under
mantyl amine (337 mg, 2.230 mmol) at room temperature. The vacuum, first at RT until dry and then for 5 minutes at 100 °C.
resulting red solution was left without stirring and red crystals Extraction of the residue with toluene (6 × 1 mL) and filtration
started to form after 1–2 hours at RT. Pentane (2 mL) of the resulting solution to remove small amounts of pale
was slowly added to complete crystallization overnight. The purple insoluble materials afforded a red-purple solution. The
crystals were separated by decantation, washed with pentane solvent was removed under vacuum to give 306 mg of dark red
1
(3 × 3 mL), and dried under vacuum. Yield: 420 mg (66%). H crystals of 8 (yield 91%). The product can further be recrystal-
NMR (300 MHz, CD2Cl2): δ 3.33 (s, 24H, NMe2), 1.88 (s, 6H, lized from cold pentane solutions. 1H NMR (300 MHz, C6D6): δ
3
CH), 1.54 (s, 12H, CH2), 1.53 (s, 12H, CH2). 13C{1H} NMR 7.41 (s, 1H, NH), 7.28 (d, J = 7.8 Hz, 4H, m-C6H3Pri2 μ-NAr*),
(125.81 MHz, CD2Cl2): δ 47.7 (CH2), 45.6 (NMe2), 36.3 (CH2), 7.08–6.94 (5H, C6H3Pri2 μ-NAr* + NHAr*), 4.14 (sept, 3J = 6.6
30.6 (CH) (CQ was not observed due to very poor solubility). IR: Hz, 4H, CHMe2 μ-NAr*), 3.30 (s, 6H, NMe2), 3.05 (s, 6H, NMe2),
3
2907s, 2848s, 2763s, 1447m, 1418w, 1345w, 1297w, 1149m, 2.91 (sept, J = 6.6 Hz, 2H, CHMe2 NHAr*), 2.43 (s, 6H, NMe2),
3
3
1117m, 1091m, 947vs, 813w, 655s, 631s, 595m, 541w. Anal. 1.46 (d, J = 6.7 Hz, 24H, CHMe2 μ-NAr*), 1.04 (d, J = 6.7 Hz,
Calcd for C28H54N6Ti2 (570.50): C, 58.95; H, 9.54; N, 14.73. 12H, CHMe2 NHAr*). 13C{1H} NMR (75.47 MHz, C6D6): δ 152.1
Found: C, 58.83; H, 9.66; N, 14.61.
(ipso-C6H3 μ-NAr*), 147.9 (ipso-C6H3 NHAr*), 140.4 (o-C6H3
[Ti(μ-NtBu)(NMe2)2]2 (2). In a closed vial, a toluene solution NHAr*), 139.1 (o-C6H3 μ-NAr*), 124.2 (p-C6H3 μ-NAr*), 123.3
(2 mL) of 400 mg of Ti(NMe2)4 (1.784 mmol) and 142 mg of (p-C6H3 NHAr*), 122.5 (m-C6H3 μ-NAr*), 121.8 (m-C6H3
tBuNH2 (1.941 mmol) was heated at 100 °C for 1.5 hours. The NHAr*), 47.5 (NMe2), 45.4 (NMe2), 42.0 (NMe2), 28.9 (CHMe2
resulting red solution crystallized on cooling to room tempera- μ-NAr* + NHAr*), 24.7 (CHMe2 NHAr*), 24.3 (CHMe2 μ-NAr*).
ture, and was dried under vacuum. The sticky residue was Anal. Calcd for C42H70N6Ti2 (754.78): C, 66.83; H, 9.35; N,
washed with cold pentane (3 × 2 mL) to afford 330 mg of a red 11.13. Found: C, 66.82; H, 9.42; N, 11.16.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 12203–12219 | 12215

View Article Online
Paper
Dalton Transactions
[Ti(NMe2)3]2(μ-NAr*) (9). A toluene solution (2 mL) of NOESY and ROESY NMR experiments. Prof. James P. Donahue
Ti(NMe2)4 (200 mg, 0.892 mmol) and 79 mg Ar*NH2 (Tulane University) is thanked for helpful discussions and
(0.446 mmol) was stirred at RT for 4 hours. The yellow solution insightful comments.
turned orange-red. The volatiles were removed under vacuum
to afford a red oil. This oil was treated with 2 mL of pentane
and evaporated to dryness (3 times). The red sticky solid was
solubilized in the minimum of pentane (ca. 1 mL), filtered
Notes and references
through celite, and left overnight in the glove box with an
opened cap for slow evaporation that led to complete crystalli-
zation. The solid was further dried under vacuum. Yield:
190 mg (80%) (alternatively, 160 mg of crystals were also
1 For selected reviews on imido complexes, see:
(a) D. E. Wigley, in Progress in Inorganic Chemistry, ed.
K. D. Karlin, Interscience, 1994, vol. 42, pp. 239–482;
(b) W. A. Nugent and J. M. Mayer, Metal–Ligand Multiple
Bonds, Wiley-Interscience, New York, 1998; (c) P. D. Bolton
and P. Mountford, Adv. Synth. Catal., 2005, 347, 355–366;
(d) N. Hazari and P. Mountford, Acc. Chem. Res., 2005, 38,
839–849.
1
obtained by placing the pentane solution at −20 °C). H NMR
(300 MHz, C6D6): δ 7.20 (d, 3J = 7.7 Hz, 2H, C6H3Pri2), 7.00 (t, 3J
= 7.6 Hz, 2H, C6H3Pri2), 3.76 (sept, 3J = 6.7 Hz, 2H, CHMe2),
3
3.01 (s, 36H, NMe2), 1.27 (d, J = 6.7 Hz, 12H, CHMe2). 13C{1H}
NMR (75.47 MHz, C6D6): δ 150.0 (ipso-C6H3), 140.8 (o-C6H3),
123.7 (p-C6H3), 122.7 (m-C6H3), 44.8 (NMe2), 27.7 (CHMe2),
24.4 (CHMe2). Anal. Calcd for C24H53N7Ti2 (535.46): C, 53.83;
H, 9.98; N, 18.31. Found: C, 53.66; H, 9.82; N, 17.85.
2 For selected examples of complexes with an imido function
acting as
a spectator ligand, see: olefin metathesis
(a) R. R. Schrock, Acc. Chem. Res., 1990, 23, 158–165;
(b) R. R. Schrock and A. H. Hoveyda, Angew. Chem., Int. Ed.,
[(Me2N)(Me2NH)Ti(μ-NAr*)2Ti(NMe2)(vNAr*)] (11). In
a
2003,
42,
4592–4633.
Olefin
polymerization;
closed vial, a toluene solution (2 mL) of Ti(NMe2)4 (100 mg,
0.4460 mmol) and Ar*NH2 (158 mg, 0.8912 mmol) was stirred
at 100 °C for 2 hours. Red crystals of 11 separated on cooling
to RT. The crystals were collected, washed with portions of
pentane (2 × 3 mL) and dried under vacuum (yield: 100 mg,
59%). 1H NMR (500 MHz, CD2Cl2, 193 K): δ 7.08 (d, 3J = 7.8 Hz,
(c) P. D. Bolton, N. Adams, E. Clot, A. R. Cowley,
P. J. Wilson, M. Schroder and P. Mountford, Organometal-
lics, 2006, 25, 5549–5565; (d) H. R. Bigmore,
M. A. Zuideveld, R. M. Kowalczyk, A. R. Cowley,
M. Kranenburg, E. J. L. McInnes and P. Mountford, Inorg.
Chem., 2006, 45, 6411–6423; (e) N. Adams, H. J. Arts,
P. D. Bolton, D. Cowell, S. R. Dubberley, N. Friederichs,
C. M. Grant, M. Kranenburg, A. J. Sealey, B. Wang,
P. J. Wilson, M. Zuideveld, A. J. Blake, M. Schroder and
P. Mountford, Organometallics, 2006, 25, 3888–3903;
(f) D. A. Pennington, M. Bochmann, S. J. Lancaster,
P. N. Horton and M. B. Hursthouse, Polyhedron, 2005, 24,
151–156; (g) A. J. Nielson, M. W. Glenny and
C. E. F. Rickard, J. Chem. Soc., Dalton Trans., 2001, 232–239;
(h) C. Lorber, B. Donnadieu and R. Choukroun, J. Chem.
Soc., Dalton Trans., 2000, 4497–4498; (i) K. Nomura and
W. Zhang, Chem. Sci., 2010, 1, 161–173; ( j) A. Arbaoui,
D. Homden, C. Redshaw, J. A. Wright, S. H. Dale and
M. R. J. Elsegood, Dalton Trans., 2009, 8911–8922;
(k) V.-H. Nguyen, L. Vendier, M. Etienne, E. Despagnet-
Ayoub, P.-A. R. Breuil, L. Magna, D. Proriol, H. Olivier-Bour-
bigou and C. Lorber, Eur. J. Inorg. Chem., 2012, 97–111.
3 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, 739–751.
4 For selected examples of reactivity of the imido function,
see: imine metathesis (a) K. E. Meyer, P. J. Walsh and
R. G. Bergman, J. Am. Chem. Soc., 1995, 117, 974–982. CH
activation: (b) J. L. Bennett and P. T. Wolczanski, J. Am.
Chem. Soc., 1997, 119, 10696–10709; (c) H. M. Hoyt,
F. E. Michael and R. G. Bergman, J. Am. Chem. Soc., 2004,
126, 1018–1019. Reaction with unsaturated C–X bonds:
(d) F. E. Michael, A. P. Duncan, Z. K. Sweeney and
R. G. Bergman, J. Am. Chem. Soc., 2005, 127, 1752–1764;
(e) T. E. Hanna, I. Keresztes, E. Lobkovsky,
W. H. Bernskoetter and P. J. Chirik, Organometallics, 2004,
3
2H, ma-C6H3Pri2 μ-NAr*), 6.99 (d, J = 7.8 Hz, 2H, mb-C6H3Pri2
μ-NAr*), 6.85 (t, 3J = 7.6 Hz, 2H, p-C6H3Pri2 μ-NAr*), 6.56 (d,
3
3J = 7.6 Hz, 2H, m-C6H3Pri2 TivNAr*), 6.37 (d, J = 7.7 Hz, 1H,
p-C6H3Pri2 TivNAr*), 3.87 (m, 2H, CHaMe2 μ-NAr*), 3.52 (s, 3H,
NMeaMeb), 3.39 (s, 3H, NMeaMeb), 3.29 (br m, 1H, NHMe2),
3.19 (m, 2H, CHbMe2 μ-NAr*), 3.06 (s, 6H, NMe2), 2.58 (d, J =
3
6.5 Hz, 6H, NHMe2), 2.08 (sept, 3J = 6.6 Hz, 2H, CHMe2
TivNAr*), 1.33 (d, 3J = 6.6 Hz, 6H, CHMeaMeb μ-NAr*), 1.19 (d,
3
3J = 6.6 Hz, 6H, CHMeaMeb μ-NAr*), 1.14 (d, J = 6.6 Hz, 6H,
CHMecMed μ-NAr*), 1.07 (d, 3J = 6.6 Hz, 6H, CHMecMed
μ-NAr*), 0.54 (br d, 12H, CHMe2 TivNAr*). 13C{1H} NMR
(125.8 MHz, CD2Cl2, 193 K): δ 152.1 (ipso-C6H3 TivNAr*),
154.8 (ipso-C6H3 μ-NAr*), 141.4 (o-C6H3 TivNAr*), 138.1 and
134.1 (o-C6H3 μ-NAr*), 128.4 and 124.0 (m-C6H3 μ-NAr*), 121.6
(p-C6H3 μ-NAr*), 121.5 (m-C6H3 TivNAr*), 117.7 (p-C6H3
TivNAr*), 53.3 and 46.3 (NMeaMeb), 43.0 (NMe2), 39.6
(NHMe2), 30.0 (CHaMe2 μ-NAr*), 27.45 (CHbMe2 μ-NAr*), 27.38
(CHMe2 TivNAr*), 25.7 (CHMeaMeb μ-NAr*), 25.3 (CHMeaMeb
μ-NAr*), 25.0 (CHMe2 TivNAr*), 24.6 (CHMecMed μ-NAr*), 24.1
(CHMecMed μ-NAr*). Anal. Calcd for C42H70N6Ti2 (754.78): C,
66.83; H, 9.35; N, 11.13. Found: C, 67.25; H, 9.44; N, 10.42. (A
perfect analysis could not be obtained, probably due to the pres-
ence of residual toluene molecules even after extensive drying
(half a molecule), as seen by NMR spectroscopy and X-ray study.)
Acknowledgements
We are grateful to the CNRS for financial support. Dr Christian
Bijani is thanked for experimental assistance with VT NMR,
12216 | Dalton Trans., 2013, 42, 12203–12219
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
23, 3448–3458; (f) P. D. Bolton, M. Feliz, A. R. Cowley, 13 D. C. Bradley and E. G. Torrible, Can. J. Chem., 1963, 41,
E. Clot and P. Mountford, Organometallics, 2008, 27, 6096– 134–138.
6110; (g) N. Vujkovic, B. D. Ward, A. Maisse-François, 14 D. C. Bradley and M. H. Gitlitz, J. Chem. Soc. A, 1969, 980–
H. Wadepohl, P. Mountford and L. H. Gade, Organometal- 984.
lics, 2007, 26, 5522–5534; (h) A. E. Guiducci, C. L. Boyd, 15 For other transamination studies of interest to our studies,
E. Clot and P. Mountford, Dalton Trans., 2009, 5960–5979.
Hydroamination and carboaminations: (i) T.-G. Ong,
G. P. A. Yap and D. S. Richeson, Organometallics, 2002, 21,
see ref. 32 and A. K. Hugues, S. M. B. Marsh,
J. A. K. Howard and P. S. Ford, J. Organomet. Chem., 1997,
528, 195–198.
2839–2841; ( j) R. K. Thomson, J. A. Bexrud and 16 For the amines used in this study: the pKa value for proto-
+
L. L. Schafer, Organometallics, 2006, 25, 4069–4071;
(k) C. Lorber, R. Choukroun and L. Vendier, Organometal-
lics, 2004, 23, 1845–1850; (l) F. Basuli, H. Aneetha,
J. C. Huffman and D. J. Mindiola, J. Am. Chem. Soc., 2005,
127, 17992–17993. Heterocycles opening: (m) P. J. Walsh,
F. J. Hollander and R. G. Bergman, J. Am. Chem. Soc., 1993,
nated amines (RNH3 /RNH2) tBuNH2 10.7, 1-adamantyl-
amine 10.5, Ph3SiNH2 6.3, PhNH2 4.2, Ar*NH2 3.2, Me2NH
10.7. Available pKa value for deprotonated amines (RNH2/
RNH−): Me2NH 42.0, PhNH2 30.6 (from http://www.chem.
wisc.edu/areas/reich/pkatable/,
and http://www.chemicalize.org).
https://ilab.acdlabs.com
12, 3705–3723; (n) S. A. Blum, V. A. Rivera, R. T. Ruck, 17 W. A. Herrmann, M. J. A. Morawietz and T. Priermeier,
F. E. Michael and R. G. Bergman, Organometallics, 2005, 24, J. Organomet. Chem., 1996, 506, 351–355.
1647–1659. Propene ammoxidation: (o) T. E. Hanna, 18 Other group 4 analogues Zr(NMe2)4 and Hf(NMe2)4 were
Coord. Chem. Rev., 2004, 248, 429–440. Hydrodenitrogena-
tion: (p) K. J. Weller, P. A. Fox, S. D. Gray and D. E. Wigley,
Polyhedron, 1997, 16, 3139–3163.
also considered; however, their larger atomic radii make
them more prone to further coordinate NHMe2 that often
lead to more complex adducts (see ref. 5f, 20, 21 and
unpublished results).
5 For selected examples: (a) C. J. Carmalt, A. Newport,
I. P. Parkin, P. A. Mountford, J. Sealey and S. R. Dubberley, 19 Reaction with V(NMe2)4 proceeds similarly with 1-adaman-
J. Mater. Chem., 2003, 13, 84–87; (b) C. J. Carmalt,
tylamine to give the V congener (see ESI‡ for experimental
A. C. Newport, I. P. Parkin, A. J. P. White and
details and X-ray structure).
D. J. Williams, Dalton Trans., 2002, 4055–4059; 20 C. Lorber and L. Vendier, Inorg. Chem., 2011, 50, 9927–9929.
(c) C. H. Winter, P. H. Sheridan, T. S. Lewkebandara, 21 S. R. Dubberley, S. Evans, C. L. Boyd and P. Mountford,
M. J. Heeg and J. W. Proscia, J. Am. Chem. Soc., 1992, 114,
Dalton Trans., 2005, 1448–1458.
1095–1097; (d) O. J. Bchir, K. M. Green, M. S. Hlad, 22 (a) D. L. Thorn, W. A. Nugent and R. L. Harlow, J. Am.
T. J. Anderson, B. C. Brooks, C. B. Wilder, D. H. Powell and
L. McElwee-White, J. Organomet. Chem., 2003, 684, 338–
Chem. Soc., 1981, 103, 357–363; (b) W. A. Nugent and
R. L Harlow, Inorg. Chem., 1979, 18, 2030–2032.
350; (e) O. J. Bchir, K. M. Green, H. M. Ajmera, E. A. Zapp, 23 For recent examples of group 4 and 5 imido-bridged
T. J. Anderson, B. C. Brooks, L. L. Reitfort, D. H. Powell,
K. A. Abboud and L. McElwee-White, J. Am. Chem. Soc., 2005,
127, 7825–7833; (f) S. D. Cosham, A. L. Johnson, K. C. Molloy
and A. J. Kingsley, Inorg. Chem., 2011, 50, 12053–12063.
6 (a) Z. Peng, Angew. Chem., Int. Ed., 2004, 43, 930–935;
(b) A. R. Moore, H. Kwen, A. M. Beatty and E. A. Maatta,
Chem. Commun., 2000, 1793–1794; (c) J. B. Strong,
G. P. A. Yap, R. Ostrander, L. M. Liable-Sands,
A. L. Rheingold, R. Thouvenot, P. Gouzerh and
E. A. Maatta, J. Am. Chem. Soc., 2000, 122, 639–649.
7 For a recent review, see: A. R. Fout, U. J. Kilgore and
D. J. Mindiola, Chem.–Eur. J., 2007, 13, 9428–9440.
8 C. Lorber, R. Choukroun and B. Donnadieu, Inorg. Chem.,
2002, 41, 4217–4226.
9 C. Lorber, R. Choukroun and L. Vendier, Eur. J. Inorg.
Chem., 2006, 4503–4518.
10 C. Lorber, R. Choukroun and L. Vendier, Inorg. Chem.,
2007, 46, 3192–3202.
11 C. Lorber and L. Vendier, Organometallics, 2010, 29, 1127–
1136.
12 The formation of the imido bond via reaction of RNH2 to
Ti(NMe2)4 seems fast (in the presence of Me3SiCl) com-
pared to the silylation of the amine or the chlorination of
Ti(NMe2)4 by Me3SiCl, unpublished results.
dimers, see ref. 4k, 5f, 8, 10, 20 and 21: (a) X. Yu,
S.-J. Chen, X. Wang, X.-T. Chen and Z.-L. Xue, Organometal-
lics, 2009, 28, 4269–4275; (b) R. F. Munha, L. F. Veiros,
M. T. Duarte, M. D. Fryzuck and A. M. Martins, Dalton
Trans., 2009, 7494–7508; (c) E. Smolensky, M. Kapon and
M. S. Eisen, Organometallics, 2005, 23, 5495–5498;
(d) F. Basuli, B. Wicker, J. C. Huffman and D. J. Mindiola, J.
Organomet. Chem., 2011, 696, 235–243; (e) G. Szigethy and
A. F. Heyduk, Inorg. Chem., 2011, 50, 125–135;
(f) A. I. Nguyen, R. A. Zarkesh, D. C. Lacy, M. K. Thorson
and A. F. Heyduk, Chem. Sci., 2011, 2, 166–169;
(g) K. Kaleta, P. Arndt, A. Spannenberg and U. Rosenthal,
Inorg. Chim. Acta, 2011, 370, 187–190; (h) U. Radius,
A. Schorm, D. Kairies, S. Schmidt, F. Möller, H. Pritzkow
and J. Sundermeyer, J. Organomet. Chem., 2002, 655, 96–
104; (i) S.-J. Chen and Z.-L. Xue, Organometallics, 2010, 29,
5575–5584; ( j) F. Cheng, S. M. Kelly, S. Clark, N. A. Young,
S. J. Archibald, J. S. Bradley and F. Lefebvre, Chem. Mater.,
2005, 17, 5594–5602; (k) K. Kaleta, P. Arndt, T. Beweries,
A. Spannenberg, O. Theilmann and U. Rosenthal, Organo-
metallics, 2010, 29, 2604–2609; (l) Y. Li, S. Banerjee and
A. L. Odom, Organometallics, 2005, 24, 3272–3278;
(m) R. M. Porter, S. Winston, A. A. Danopoulos and
M. B. Hursthouse, J. Chem. Soc., Dalton Trans., 2002,
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 12203–12219 | 12217

View Article Online
Paper
Dalton Transactions
3290–3299; (n) G. I. Nikonov, A. J. Blake and P. Mountford, 36 For other examples of mono-bridged imido with additional
Inorg. Chem., 1997, 36, 1107–1112.
bridging ligands (oxo, azo), see: (a) S. Gambarotta,
C. Floriani, C. Villa and C. Guastini, J. Am. Chem. Soc.,
1983, 105, 7275–7301; (b) H. Fric and U. Schubert, Z. Natur-
forsch., B: Chem. Sci., 2007, 62, 487–490; (c) P. Gomez-Sal,
A. Martin, M. Mena, del C. Morales and C. Santamaria,
Chem. Commun., 1999, 1839–1840.
24 The X-ray structure of 3 is given in ESI,‡ and a related syn-
thesis of 3 was recently reported during the course of our
investigations by Johnson et al. who used a thermal reac-
tion [see ref. 5f] to produce the same compound that crys-
tallized with different cell parameters.
25 E. Smolensky, M. Kapon and M. S. Eisen, Organometallics, 37 For an example of the related tautomeric eneamido-amido/
2007, 26, 4510–4527.
26 The absence of [Ti(NPh)(NMe2)2]2 was unequivocally
assessed by comparison with the 1H NMR data of a sample
amidinate-imido tantalum system, see: B. L. Yonke,
A. J. Keane, P. Y. Zavalij and L. R. Sita, Organometallics,
2012, 31, 345–355.
prepared by deprotonation of NHMe2 ligands in Ti(vNPh)- 38 Variable-temperature NMR experiments (starting from low
Cl2(NHMe2)2 precursor, see the text for a related synthesis of 7.
27 M. Fan, E. N. Duesler, J. F. Janik and R. T. Paine, J. Inorg.
Organomet. Polym. Mater., 2007, 17, 423–437.
28 We tested several other substituted anilines (ArNH2), and
most often we observed side-products along with the
expected dimer [Ti(NAr)(NMe2)2]2.
temperature CD2Cl2 solutions of 11) suggest the equili-
brium to be irreversible. Curiously, heating solutions of 8
affords some 11 (before decomposition occurs). The
prototropy might be base sensitive or even induced by
bases present in the reaction medium (excess of Ar*NH2 or
liberated NHMe2), and in the presence of pyridine the
reaction affords 15 which is an analogue of 11, see later in
the text.
29 The composition is the following: the ratio HNMe2/Ti–
NMe2 is 25/75 after 19 hours (20/80 after 1 hour). This rep-
resents about one Ti–NMe2 bond that was transaminated 39 There is one example of a tetranuclear complex with a
(over the four possible bonds in Ti(NMe2)4, but in theory
only two bonds due to the initial stoichiometry used). On
the remaining Ti–NMe2 bonds, 50% is from Ti(NMe2)4
(70% after 1 hour).
Ti4(μ-NH)3(μ3-N) cube-like motif, see: A. Abarca, A. Martin,
M. Mena and C. Yelamos, Angew. Chem., Int. Ed., 2000, 39,
3460–3463.
40 (a) C. C. Cummins, G. D. Van Duyne, C. P. Schaller and
P. T. Wolczanski, Organometallics, 1991, 10, 164–170;
(b) C. C. Cummins, C. P. Schaller, P. T. Wolczanski,
A. W. E. Chan and R. Hoffmann, J. Am. Chem. Soc., 1991,
113, 2985–2994.
30 In some cases, when too many cycles are done, we noticed
the progressive transformation into a different grey-brown
poorly soluble complex that was later characterized to be 11.
31 Noteworthy, the reaction conducted under reflux con-
ditions afforded more 8 than 7 (ca. 60 : 40). We explain the 41 (a) R. D. Profilet, C. H. Zambrano, P. E. Fanwick, J. J. Nash
formation of 8 from 7, by ligand redistribution in the pres-
ence of HNMe2. Later in the text we will also see that under
heating, 8 is transformed into isomer 11, and that upon
dissolution 11 returns to 8.
and I. P. Rothwell, Inorg. Chem., 1990, 29, 4362–4364;
(b) C. H. Zambrano, R. D. Profilet, J. E. Hill, P. E. Fanwick
and I. P. Rothwell, Polyhedron, 1993, 12, 689–708.
42 Knowing that compound 10 had been isolated as a minor
species in the reaction Ti(NMe2)4 with 4 or 5 equivalents of
Ar*NH2 incited us to attempt again its selective synthesis
with the right stoichiometry (from Ti(NMe2)4 and 2 equiv.
of Ar*NH2) and by applying experimental conditions used
successfully for the preparation of 13 (2 cycles of heating/
vacuum, and under solventless conditions). These
attempts, however, failed to give 10 in higher yields, and 11
was still the major isolated component (in a manner
similar to what had been previously described in a toluene
solution). However, under these experimental conditions,
we have been able to detect new signals (characteristic
signals in 1H NMR (C6D6): 7.98 (Ar*NH), 4.55 (iPr), 3.25
(–NMe2)) that could correspond to C that was formed in
about the same low amount as its isomer 10. Although at
first sight curious, this observation could mean that at
high Ar*NH2 concentrations the tetraanilido complex 12 is
the kinetic species that react with Ti(NMe2)4 to form
isomer 10, whereas at lower Ar*NH2 concentrations isomer
C is formed by substitution of one –NMe2 in mono-anilido
8 or imido species 11.
32 (a) The reaction in pentane seems, however, slower which
is ideal for the observation of intermediate 9: (b) Donor sol-
vents have also been tested as they can change the course
of the reaction by displacement of NHMe2 and/or trapping
transient species. The example of pyridine is described
later in the text: (c) Dichloromethane as a solvent gave rise
to the formation of chlorinated products.
33 For comparison with related TivNR complexes, see, for
example, ref. 2f, 5c, 9 and (a) N. Adams, H. R. Bigmore,
T. L. Blundell, C. L. Boyd, S. R. Dubberley, A. J. Sealey,
A. R. Cowley, M. E. G. Skinner and P. Mountford, Inorg.
Chem., 2005, 44, 2882–2894; (b) C. Lorber and L. Vendier,
Organometallics, 2008, 27, 2774–2783; (c) C. Lorber and
L. Vendier, Dalton Trans., 2009, 6972–6984.
34 D. J. Arney, M. A. Bruck, S. R. Huber and D. E. Wigley,
Inorg. Chem., 1992, 31, 3749–3755.
35 There is only one other example of a single bridged
Ti(μ-NAr)Ti complex (see: S. H. Lee, C. J. Wu, U. G. Joung,
B. Y. Lee and J. Park, Dalton Trans., 2007, 4608–4614);
however, in this example the bridging mode of the aryl-
imido ligand is likely to be due to further additional 43 At 80 °C the same reaction was shown to lead to complex
coordination of two pyrrol ortho-substituents.
17 (yield: 35%) and unidentified species.
12218 | Dalton Trans., 2013, 42, 12203–12219
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
44 A. D. Schwarz, A. J. Nielson, N. Kaltsoyannis and 46 Moreover, even if unwanted complexes such as 8 are
P. Mountford, Chem. Sci., 2012, 3, 819–824.
present in solution, we verified that in the presence of an
excess of Me3SiCl, the Ti–NHAr* bond (as well as Ti–NMe2
bonds) is chlorinated to form a Ti–Cl bond (i.e., 8 +
45 (a) Y. Shi, J. T. Ciszewski and A. L. Odom, Organometallics,
2001, 20, 3967–3969; (b) C. Cao, J. T. Ciszewski and
A. L. Odom, Organometallics, 2001, 20, 5011; (c) K. Takaki,
S. Koizumi, Y. Yamamoto and K. Komeyama, Tetrahedron
Lett., 2006, 47, 7335–7337; (d) J. A. Bexrud, J. D. Beard,
D. C. Leitch and L. L. Schafer, Org. Lett., 2005, 7, 1959–
4Me3SiCl
+
4NHMe2
→
2Ti(vNAr*)Cl2(NHMe2)2
+
3Me3SiNMe2 + Me3SiNHAr*). In certain cases oligomeric
complexes are obtained, see the recent study: C. Lorber
and L. Vendier, Inorg. Chem., 2013, 52, 4756–4758.
1962; (e) L. Ackermann and R. G. Bergman, Org. Lett., 2002, 47 D. C. Bradley and I. M. Thomas, J. Chem. Soc., 1960, 3857–
4, 1475–1477.
3861.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 12203–12219 | 12219