Transition metal imide/organic imine metathesis reactions: Unexpected observations

Article abstract of DOI:10.1039/a803719a

Mixtures of [Ti(NBu^t)Cl₂(py)₃] 1 and PhC(NAr)H (Ar = C₆H₃Me₂-2,6 or C₆H₄Me-4) gave quantitative conversion to [Ti(NAr)Cl₂(py)₃] and PhC(NBu^t)H, the products of Ti=N-Bu^t/C=NAr transition metal imide/organic imine metathesis; examination of the kinetics for Ar = C₆H₄Me-4 showed that the rate limiting step for this process is zero order in [1], demonstrating that these reactions do not involve metal imide particiption in the rate limiting step.

Full text of DOI:10.1039/a803719a

View Article Online / Journal Homepage / Table of Contents for this issue
Transition metal imide/organic imine metathesis reactions: unexpected
observations
Jacqueline M. McInnes and Philip Mountford*†‡
Department of Chemistry, University of Nottingham, Nottingham, UK NG7 2RD
t
But
Mixtures of [Ti(NBu )Cl
2
(py)
3
] 1 and PhC(NAr)H (Ar =
Tol
H
N
N
C
6
H
3
Me
2
-2,6 or C
(py) ] and PhC(NBu )H, the products of Ti§N-
Bu /C§NAr transition metal imide/organic imine meta-
thesis; examination of the kinetics for Ar = C Me-4
6
H
4
Me-4) gave quantitative conversion to
py)3Cl2Ti=NBut
(
+
(py)3Cl2Ti=NTol
+
(2)
t
[
Ti(NAr)Cl
2
3
1
Ph
6
Ph
H
t
5
4
6
H
4
showed that the rate limiting step for this process is zero
order in [1], demonstrating that these reactions do not
involve metal imide particiption in the rate limiting step.
fully reproducible in both CDCl
3 6 6
and C D . At first sight, eqns.
(1) and (2) are apparently straightforward examples of metal-
mediated imide/imine metathesis reactions. Furthermore, reac-
tion of 1 with certain N-aryl-a-diimines also leads to the
expected arylimido titanium products and the corresponding
N-tert-butyl-a-diimines.
One attraction of transition metal imido compounds is their
potential use in metal-mediated imido group (NR) transfer
reactions. The addition of unsaturated substrates to M§NR
linkages is now well documented, but examples of reactions
resulting in complete removal (functional group transfer) of NR
from metal to substrate are still somewhat more restricted,
especially by comparison with the rich functional group transfer
In a separate experiment [eqn. (3)] an equimolar mixture of
But
H
Tol
H
k1
N
N
t
Bu NH2
+
TolNH2
+
(3)
k–1
1
chemistry of metal oxo complexes. Very recently there have
Ph
Ph
been reports of transition metal imide/organic imine (L
n
M§NR/
5
4
2
R C§NRA) functional group metathesis reactions, and these
2
t
Bu NH
gave an equilibrium mixture (Keq ca. 1, i.e. k
the two starting compounds along with the products TolNH
2
with 5 in CDCl
3
(no titanium imido compound added)
have been proposed in some instances to arise via metal imide-
mediated processes, possibly involving cycloaddition of
C§NRA to M§NR. Such processes appear to be the aza-
1
≈ k21) containing
2
and imine 4 after 8 hours at 30 °C. Similarly, starting with an
equimolar mixture of pure TolNH2 and 4 also gave an
analogues of the widely used metal alkylidene (L
n
2
M§CR )/
alkene (RAC
2
§CRA
2
) metathesis reaction3 and are therefore
t
2
equilibrium mixture of TolNH , 4, Bu NH
2
and 5. Since there is
potentially hugely important transformations in stoichiometric
and catalytic synthesis. Identification of the possible role(s) of
the metal centre in such processes is clearly essential. As part of
an ongoing study of early transition metal imido chemistry, we
found that a number of titanium imido complexes undergo well
characterised cycloaddition and other coupling reactions of the
no thermodynamic driving force in eqn. (3) to account for the
net quantitative reactions in eqns. (1) and (2), we infer that it is
formation of the arylimide products 3 and 6 that causes these
two reactions to go to completion. This hypothesis is supported
by eqn. (4) which shows that addition of a stoichiometric
t
4
Ti§NR group (R = Bu or aryl) with organic substrates. In
k2
t
parallel studies we have observed reactions of [Ti(NBu )Cl
2
-
py)3Cl2Ti=NBut + ArNH2
t
(
(py)3Cl2Ti=NAr
+
Bu NH
2
(4)
4b
3
(py) ]
with certain mono- and a-di-imines that yield the
1
3 or 6
expected products of imide/imine functional group metathesis.
The results herein show that examples of apparently metal-
mediated imide/imine metathesis reactions must be subjected to
mechanistic investigations before they can be properly au-
Ar = C6H3Me2-2,6 or Tol
t
amount of arylamine to [Ti(NBu )Cl
2
(py)
3
] 1 gives quantitative
thenticated.
Reaction of [Ti(NBu )Cl
t
4b
formation of Bu NH
2
and the corresponding arylimide 3 or 6.
t
2
(py)
3
] 1 with one equivalent of the
This process is (qualitatively) very much faster than any of the
other three [i.e. eqns. (1)–(3)]. Further evidence for the role of
N-2,6-dimethylphenylbenzaldehyde imine 2 at 60 °C for 96
hours in CDCl gave (by NMR) quantitative and stoichiometric
conversion to PhC(NBu )H 4 and the arylimido compound
3
1
as an effective trap for ArNH
in which 1 was added to a pre-equilibrated mixture of Bu NH
, TolNH and 4 [eqn. (3)]: 1 and TolNH were consumed and
there was quantitative conversion of 5 to 4.
2
came from NMR experiments
t
t
2
,
[Ti(NC
H
6 3
Me
2
-2,6)Cl
2
(py)
3
] 3 [eqn. (1)].§ Repeating the
5
2
2
But
N
Ar
N
py)3Cl2Ti=NBut
Examination of the kinetics of eqn. (3) under pseudo-first
order conditions established that this reaction [from left to right
as shown in eqn. (3)] is first order in both Bu NH
(
+
(py)3Cl2Ti=NAr
+
(1)
1
Ph
H
3
Ph
H
t
2
and imine 5
2
4
and also allowed the direct deduction of k
1
. However, reaction
t
of [Ti(NBu )Cl (py) ] 1 with PhC(NTol)H 5 (10, 16.5 and 20
2 3
Ar = C6H3Me2-2,6
equivs.) under otherwise identical pseudo-first order conditions
showed (Fig.1) that the rate for eqn. (2) is zero order with
respect to 1 and has no simple dependence on [5]. These
observations are clearly not consistent with the expected
titanium imide-mediated metathesis mechanism of the type
illustrated in Scheme 1, in which a diazametallacycle 7 is
formed in the rate limiting step. For such a process a rate
dependency on [1] is required.¶ Indeed, to our knowledge, all
reaction on a preparative scale confirmed the exchange of tert-
butylimide for the 2,6-dimethylphenylimide ligand. Similarly,
reaction of 1 with one equivalent of the less bulky N-arylimine
PhC(NTol)H 5 (Tol = C
6
H
4
Me-4) gave a quantitative yield of
, along with the tolylimido
4
after only 4 hours at 60 °C in CDCl
3
product 6 [eqn. (2)]. Products 3, 4 and 6 were identified by
comparison with authentic samples4b and the reactions were
Chem. Commun., 1998
1669

View Article Online
reached. Further investigations of these and other related
reactions are currently under way.
We thank the EPSRC and Leverhulme Trust for support,
Professors M. Bochmann, R. H. Grubbs and J. J. Turner for
3
helpful comments and Dr P. Scott for a gift of [Sc(OTf) ].
Notes and References
†
Email: Philip.Mountford@Nottingham.ac.uk,
www: http://www.nottingham.ac.uk/ ~ pczwww/Inorganic/PMount.html
Philip Mountford is the Royal Society of Chemistry Sir Edward Frankland
Fellow.
3
For all the rate determinations the disappearance of [Ti(NBu )Cl ] 1
‡
t
§
2
(py)
2
was monitored over at least three half-lives under pseudo-first-
t
or Bu NH
order conditions with p-dimethoxybenzene as the internal standard. Both
before and after each experiment the temperature in the NMR probe was
calibrated using ethylene glycol (80%) in DMSO. The temperature did not
vary by more than 1 °C. The samples were prepared in a dry-box using
Teflon valve (Young’s) 5 mm tubes, which had been rigorously dried.
Fig. 1 Linear plots of [1]
of PhC(NTol)H 5 in CDCl
rate on [1]. Legend: 5 10 equivalents of 5; :, 16.5 equivalents of 5; -, 20
equivalents of 5.
t
/[1]
0
versus time for the reactions with an excess
3
showing clearly the zero order dependence of
CDCl
3 6 6
and C D were dried over calcium hydride and potassium
1
respectively, at rt, then distilled under reduced pressure. H NMR spectra
were recorded on a Bruker DPX 300 spectrometer at 30 °C.
But
¶ For the process shown in Scheme 1, rate = kobs[metal imide][organic
imine], where kobs = k k /(k23 + k ), assuming the k24 step is not significant
3 4 4
at low concentrations of product imine and a steady state approximation for
[7]. In the reactions in eqns. (1) and (2), pyridine dissociation may be
required before formation of hypothetical 7. In this instance the rate should
still be dependent on [metal imide], but kobs would show a dependence on
Tol
H
N
Ph
H
k3
N
[Ti]
Ti] NBut
[
+
N
Ph
k-3
Tol
7
[
pyridine]. However, we have found that the five-coordinate bis(pyridine)
5
t
2 2
complex [Ti(NBu )Cl (py) ] also reacts with organic imines at a rate not
But
significantly different from that of 1 under the same conditions and so the
presence of the third pyridine molecule is not important.
But
N
k4
N
Ph
∑
Combining eqns. (3) and (4) and assuming a steady state approximation of
[
Ti]
[Ti] NTol
+
t
H
[TolNH
pseudo-first order conditions used here, [5] >> [4] and hence the k21 back-
reaction is negligible, especially since 1 is an effective trap for TolNH
2 1 2
] leads to the expression: rate = k [Bu NH ][5] since, under the
N
k-4
Ph
H
Tol
4
2
.
7
** From kinetic analysis of eqn. (3) we know independently the value of k
1
2
3
21 21
(
1.35 3 10
indeed rate = k
to 0.43
maximum possible value of [Bu NH
M
s
) under analogous reaction conditions. Therefore, if
Scheme 1
t
t
1
[Bu NH
2
][5] then [Bu NH
2
] has to be approximately 0.10
M
for the experiments shown in Fig. 1. From the NMR spectra the
t
kinetically characterised examples of alkene metathesis by
homogeneous metal alkylidene (carbene) complexes (which
proceed via metallacyclobutane intermediates) show a first
order dependence on metal concentration. Furthermore, studies
by Lee and Bergman of cycloaddition reactions of the imido
zirconium complex [Cp
L = THF, OPPh
be first order in imido complex. Previously reported zirco-
nium-mediated imine metathesis reactions also follow a first
order relationship between reaction rate and metal complex,
although in this instance the resting state of the complex is a
diazametallacycle.2a
2
] in these experiments is < 0.005
M
.
However, it is possible that under the reaction conditions of eqns. (1) or (2),
eqn. (3) is accelerated by Lewis and/or Brønsted acids. Indeed, tests of this
hypothesis for eqn. (3) with added [Sc(OTf) ] demonstrated a rate
3
5
acceleration suggesting that Lewis acids (e.g. 1 and/or its trace decomposi-
tion products) may make a tandem combination of eqns. (3) and (4) a
possibility. Toluene-p-sulfonic acid had no measurable effect on the rate of
t
2
Zr(NR)(L)] (R = Bu or C
6 3
H Me
2
-2,6,
3
or 4-tert-butylpyridine) have shown them to
eqn. (3). We cannot directly determine the influence of additional Bu NH
t
2
6
t
on the rate of reaction of 1 with 5 [eqn. (2)]: any Bu NH
2
added would be
rapidly scavenged by 5 [eqn. (3)] since t1/2 for the consumption of 5 in eqn.
(3) is ca. 25 times shorter than for 5 in eqn. (2) under otherwise identical
conditions.
The absence of a rate dependence on [1] implies that one or
more alternative mechanisms (to that shown in Scheme 1) must
be operating in eqns. (1) and (2). At first sight, one possibility
1
D. E. Wigley, Prog. Inorg. Chem., 1994, 42, 239; W. A. Nugent and J. M.
Mayer, Metal-Ligand Multiple Bonds, Wiley-Interscience, New York,
1
988; R. H. Holm, Chem. Rev., 1987, 87, 1401; E. W. Harlan and R. H.
appears to be a tandem combination of eqns. (3) and (4) where
Holm, J. Am. Chem. Soc., 1990, 112, 186; L. K. Woo, Chem. Rev., 1993,
93, 1125.
t
traces of Bu NH
2
(in principle impossible to exclude for these
very moisture sensitive systems, although we have no NMR
evidence for its presence in mixtures of 1 and 5) give rise to
2 (a) K. E. Meyer, P. J. Walsh and R. G. Bergman, J. Am. Chem. Soc., 1995,
117, 974; (b) G. K. Cantrell and T. Y. Meyer, Organometallics, 1997, 16,
5
381; (c) G. K. Cantrell and T. Y. Meyer, Chem. Commun., 1997, 1551;
formation of imine 4, the liberated TolNH
forming the imido product 6. However, the kinetic results do
2
being scavenged by
(d) G. K. Cantrell, T. Pontz and T. Y. Meyer, Abstracts of Papers of the
1
American Chemical Society, 1997, 214, 312-INOR.
not support this mechanism since the rate law in this case∑
3
4
For reviews see: V. C. Gibson, Adv. Mater., 1994, 6, 37; R. H. Grubbs and
S. Chang, Tetrahedron, 1998, 54, 4413, and references therein.
(a) P. Mountford, Chem. Commun., 1997, 2127 (Feature Article);
requires a first-order dependence on [5]. Furthermore there is no
t
evidence for significant enough concentrations of Bu NH
2
in
these NMR mixtures to account for the observed rate of
formation of 4 and 6 in these reactions.**
(b) A. J. Blake, P. E. Collier, S. C. Dunn, W.-S. Li, P. Mountford and
O. V. Shishkin., J. Chem. Soc., Dalton Trans., 1997, 1549; (c) A. J.
Blake, P. E. Collier, L. H. Gade, P. Mountford, M. Schubart and I. J.
Scowen, Chem. Commun., 1997, 1555.
In summary we have demonstrated that apparently straight-
forward transition metal imide/organic imine metathesis reac-
tions do not necessarily involve metal imide participation in the
rate determining step. This is in sharp contrast to the well known
metal alkylidene/alkene metathesis reactions. Cases of metal
imide/imine metathesis must clearly be subjected to mecha-
nistic scrutiny before conclusions concerning either the role of
the metal centre or any structure–activity relationships can be
5
6
E. L. Dias, S. T. Nguyen and R. H. Grubbs, J. Am. Chem. Soc., 1997, 119,
3
887 and references therein.
S. Y. Lee and R. G. Bergman, Tetrahedron, 1995, 51, 4255 and
references therein.
Received in Cambridge, UK, 18th May 1998; 8/03719A
1670
Chem. Commun., 1998