Synthesis, characterisation and a theoretical investigation of the formation of lithium dialkylaluminium amides

Article abstract of DOI:10.1016/S0022-328X(98)00954-1

Reaction of the pre-formed crystalline amides [(PhCH₂)₂NLi] and [Me₂AlN(CH₂Ph)₂] in the presence of pyridine results in the formation of the mixed metal complex [Me₂Al{(PhCH₂)₂N}₂Li·pyr] 1. Ab initio molecular orbital calculations indicate formation of the bimetallic product is energetically favourable. Using single crystal X-ray analysis for 1 and the pyridine solvate [{(PhCH₂)₂NLi·pyr}₂] 7, in combination with theoretical calculations, the possible driving forces for the reaction are discussed. A major contributing factor in the stabilisation of the bimetallic compound was found to be a reduction in steric crowding in the mixed metal base compared to the homometallic dialkylaluminium amide. In addition, complex 1 shows significant benzyl to lithium interactions which contribute to the overall bonding. Such interactions are unusual in that donor solvent is present as competing complexant.

Full text of DOI:10.1016/S0022-328X(98)00954-1

Journal of Organometallic Chemistry 572 (1999) 283–289
Synthesis, characterisation and a theoretical investigation of the
formation of lithium dialkylaluminium amides
William Clegg b, Stephen T. Liddle b, Kenneth W. Henderson a,*, Fiona E. Keenan a,
Alan R. Kennedy a, Arlene E. McKeown a, Robert E. Mulvey a
a Department of Pure and Applied Chemistry, Uni6ersity of Strathclyde, Glasgow G1 1XL, UK
b Department of Chemistry, Uni6ersity of Newcastle, Newcastle-upon-Tyne NE1 7RU, UK
Received 4 August 1998
Abstract
Reaction of the pre-formed crystalline amides [(PhCH2)2NLi] and [Me2AlN(CH2Ph)2] in the presence of pyridine results in the
formation of the mixed metal complex [Me2Al{(PhCH2)2N}2Li·pyr] 1. Ab initio molecular orbital calculations indicate formation
of the bimetallic product is energetically favourable. Using single crystal X-ray analysis for 1 and the pyridine solvate
[{(PhCH2)2NLi·pyr}2] 7, in combination with theoretical calculations, the possible driving forces for the reaction are discussed.
A major contributing factor in the stabilisation of the bimetallic compound was found to be a reduction in steric crowding in the
mixed metal base compared to the homometallic dialkylaluminium amide. In addition, complex 1 shows significant benzyl to
lithium interactions which contribute to the overall bonding. Such interactions are unusual in that donor solvent is present as
competing complexant. © 1999 Elsevier Science S.A. All rights reserved.
Keywords: Lithium; Aluminium; Alkyls; Amides; Molecular orbital calculations
1. Introduction
four equivalents of the lithiated organonitrogen com-
pound [4] (ii) reaction of LiAlH4 with four equivalents
of amine [5] (iii) lithiation of a dialkylaluminium pri-
mary amide (R2AlN(H)R%) to form the imido derivative
[6] and (iv) methane elimination from an amine sol-
vated complex of a lithium trialkylaluminium sec-
ondary amide 1–4 [7].
Heterobimetallic combinations of Group 1 metals
with other main group metallic elements represents an
area of ever increasing interest [1]. Attention has mainly
been directed towards mixed metal species involving
lithium due to a combination of the metal’s synthetic
utility and ease of use [2]. A number of mixed cation
organonitrogen compounds containing lithium and
Group 13 metals have been structurally characterised
[3–7]. To date, four synthetic methodologies have been
successfully applied for the preparation of mixed
lithium–aluminium organonitrogen complexes. These
are (i) transmetallation of an aluminium halide with
(1)
(2)
(3)
(4)
In this paper an alternative synthetic strategy involv-
ing the direct reaction of pre-formed amidolithium and
* Corresponding author. Fax: +44-141-552-0876; e-mail:
k.w.henderson@strath.ac.uk.
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(98)00954-1

284
W. Clegg et al. / Journal of Organometallic Chemistry 572 (1999) 283–289
Table 1
Crystallographic data
dialkylaluminium amide has been utilised to preparethe
mixed metal compound [Me2Al{(PhCH2)2N}2Li·pyr] 1.
An ab initio molecular orbital study of the homometal-
lic precursor complexes and the mixed metal products
has been used to investigate the energetics of the reac-
tion and the driving forces for the stabilisation of the
bimetallic product.
Compound
1
7
Empirical formula
M
C
535.61
35H39AlLiN3
C38H38Li2N4
564.60
Crystal system
Space group
Unit cell dimensions
Orthorhombic
P212121
Triclinic
(
P1
˚
a (A)
15.661(4)
20.052(3)
9.840(2)
90
90
90
3090.2(11)
4
1.151
0.093
1144
9.2461(17)
9.4721(17)
10.593(2)
88.762(5)
76.565(4)
67.282(5)
829.9(3)
1
1.130
0.066
300
160
˚
2. Results and discussion
b (A)
˚
c (A)
h (°)
i (°)
2.1. X-ray crystallographic studies
k (°)
3
Pure crystalline samples of lithium dibenzylamide [8]
˚
V (A )
[(PhCH2)2NLi]
2
and
dimethyl(dibenzylamido)-
Z
Dcalc. (g cm−3
)
aluminium [Me2AlN(CH2Ph)2] 3 were prepared and
stored in an argon filled glove box. The solids were then
mixed in a 1:1 molar ratio and suspended in toluene.
One equivalent of pyridine was added to the reaction
mixture and the resultant deep red solution precipitated
complex 1 on cooling. An alternative synthesis involved
thermolysis of the amine solvated complex
[Me3Al{(PhCH2)2N}2Li·HN(CH2Ph)2] in a pyridine/
toluene mixture [7]. The identical composition of the
products from these reactions was confirmed by 1H-
NMR spectroscopy.
v (mm−1
F(000)
)
Temperature (K)
Crystal Size (mm)
qmax (°)
Maximum indices hkl
Reflections measured
Unique reflections
Rint
123
0.70×0.60×0.60 0.68×0.42×0.36
30.0
22, 28, 13
5484
5446
0.0379
363
0.1513
0.0487
1.050
28.6
12, 12, 13
6120
3701
0.0572
211
0.1678
0.0672
0.877
No. of refined parameters
wR2 (all data)a
R1 (observed data)a
Goodness-of-fit
An X-ray crystallographic study of 1 revealed a 1:1
complex between 2 and 3 with the lithium atom being
monosolvated by a single pyridine molecule (Fig. 1,
crystallographic data are given in Table 1). Bridging
−3
˚
Max/min el. density (e A
)
0.446, −0.663
0.359, −0.371
a wR2={S[w(F2o−Fc2)2]/S[w(F2o)2]}1/2, R1=SꢀFoꢁ−ꢁFcꢀ/SꢁFoꢁ.
amido anions link the aluminium and lithium centres,
derived from primary amines (RNH2). As expected,
comparison of the core dimensions between 1 and 4
finds only very small variances. The only notable differ-
ence is the substitution of the Li(1)–N(3) dative bond
¸¹º
resulting in a central four-membered LiNAlN ring.
Several complexes containing this bridging motif
have previously been characterised, including the
˚
closely related THF solvated analogue of
1
(2.027(5) A) in 1 for the Li(1)–O(1) dative bond
˚
[Me2Al{(PhCH2)2N}2Li·THF] 4 [7]. Another two com-
plexes related to 1 are [Mes2Al(NHtBu)2Li·THF] [6] 5
and [(tBu)2Al{Ph3C(H)N}2Li] [9] 6, which are both
(1.91(1) A) in 4, following the replacement of pyridine
for THF. All other dimensions, including the orienta-
tions of the benzyl groups, are similar in both struc-
tures (a detailed examination is given in the Section
2.2). Of interest is the highly pyramidal nature of the
lithium atom (sum of the angles 346.9 for 1 and 336.6°
for 4). Also, the four-membered cyclic cores of 1 and 4
are non-planar (mean deviation from the plane 0.159 in
˚
1 and 0.142 A in 4; sum of the endocyclic bond angles
354.7 for 1 and 355.8° in 4). Key bond lengths and
angles for 1 are listed in Table 2.
The structures of the homometallic reagents are of
interest when considering the possible mechanism of
formation of 1. Reaction of pre-formed crystalline 2
with one equivalent of pyridine in toluene solution
yielded a red solution. Standing this solution at room
temperature (r.t.) deposited crystals of the solvated
complex [{(PhCH2)2NLi·pyr}2] 7 which was character-
1
ised by H-NMR spectroscopy and single crystal X-ray
Fig. 1. Molecular structure of 1 with hydrogen atoms omitted for
clarity.
analysis. In the solid state 7 is found to be dimeric,

W. Clegg et al. / Journal of Organometallic Chemistry 572 (1999) 283–289
285
Table 2
Key distances (A) and angles (°) for 1
Table 3
˚
˚
Key distances (A) and angles (°) for 7
˚
˚
Bond distance (A)
Bond distance (A)
Al1–N1
Al1–C1
Al1···Li1
N2–Li1
1.904(2)
1.980(3)
2.652(5)
2.063(6)
Al1–N2
Al1–C2
N1–Li1
N3–Li1
1.934(2)
1.982(3)
2.078(5)
2.027(5)
Li1–N1
Li1–N2
1.983(4)
2.012(4)
Li1–N1a
Li1···Li1a
1.988(4)
2.410(7)
Bond angle (°)
N1–Li1–N1a
N1a–Li1–N2
C7–N1–Li1a
C8–N1–Li1
105.29(16)
123.43(18)
116.43(16)
119.00(16)
74.71(16)
N1–Li1–N2
C7–N1–Li1
C7–N1–C8
C8–N1–Li1a
N1–C8–C9
129.53(19)
116.59(18)
110.25(15)
116.06(17)
114.45(17)
Bond angle (°)
N1–Al1–N2
N1–Al1–C2
N2–Al1–C1
Al1–N1–C3
C3–N1–C10
C10–N1–Li1
Al1–N2–C24
C17–N2–C24
C24–N2–Li1
C31–N3–Li1
N1–C3–C4
98.76(9)
113.03(11)
111.47(11)
121.14(16)
110.4(2)
107.0(2)
112.14(15)
111.3(2)
110.6(2)
120.4(3)
116.8(2)
114.4(2)
139.8(2)
128.3(3)
N1–Al1–C1
C1–Al1–C2
N2–Al1–C2
Al1–N1–C10
C3–N1–Li1
Al1–N2–C17
Al1–N2–Li1
C17–N2–Li1
C31–N3–C35
C35–N3–Li1
N1–C10–C11
N2–C24–C25
N1–Li1–N2
N1–Li1–N3
106.13(12)
116.86(14)
113.03(11)
120.74(17)
108.5(2)
118.14(16)
83.07(16)
118.8(2)
116.9(3)
122.1(2)
111.53(19)
117.3(2)
89.4(2)
Li1–N1–Li1a
N1–C7–C1
110.47(15)
˚
˚
N(2) 2.012(4) A) for THF (Li(1)–O(1) 1.915(3) A). The
effect of solvent substitution is also seen in a shortening
of the Li–N bonds of the central four-membered ring
of 7 compared to 8, Li(1)–N(1) 1.983(4) and Li(1)–
N(1a) 1.988(4) A in 7, but Li(1)–N(1) 2.028(3) and
Li(1)–N(1a) 2.058(3) A in 8. Nevertheless, overall the
N2–C17–C18
Al1–Li1–N3
N2–Li1–N3
˚
˚
129.2(3)
structures adopt very similar gross geometrical features.
Key bond lengths and angles for 7 are given in Table 3.
bridging through the amido anions and with each
lithium terminally solvated by one pyridine molecule
(Fig. 2).
2.2. Theoretical calculations
Ab initio molecular orbital calculations run at the
HF/6-31G* level were used to probe the nature of the
reaction between the homometallic bases. Initially
dimethylamido units were used to model the dibenzy-
lamido anions to simplify the calculations. Dimers were
taken to be the most likely structures for the ho-
mometallic bases. This is the most common structural
arrangement for dialkylaluminium amides [11] and sol-
vated amidolithiums. The homometallic starting materi-
als [{Me2AlNMe2}2] I and [{Me2NLi·pyr}2] II and the
heterobimetallic complex [Me2Al(Me2N)2Li·pyr] III
were geometry optimised with no imposed constraints.
Figs. 3–5 show the energy minima geometries with
selected bond lengths, bond angles and charges. A
comparison of the geometry of I with its known crystal
structure shows excellent agreement, with the average
difference between the bond lengths and between the
The structural chemistry of the lithium dibenzy-
lamide system has been extensively studied over the
past two decades as a model for bulky amido anions
[10]. As found in 7 the most common structural motif
for solvates of lithium dibenzylamide is that of a disol-
vated dimeric ring. A closely related analogue of 7 is
the THF solvate [{(PhCH2)2NLi·THF}2] 8 ([10]b).
Comparison of the bond lengths and angles between 7
and 8 show only small variances. As expected, the
major difference is the replacement of pyridine (Li(1)–
˚
angles being (B0.05 A) and (B1°), respectively ([11]a).
Table 4 lists the energies of the calculated structures.
The energy of the reaction between I and II to form
two equivalents of III is calculated to be exothermic by
−18.40 kcal mol−1. An inspection of the bond lengths
and angles within the structures of I–III gives some
indication of the source of this stabilisation. The core
ring in II is distinctly rhomboidal with large N–Li–N
angles of 106.8° and acute Li–N–Li angles of 73.2°. In
contrast the core of I is much nearer to a square with
N–Al–N angles of 88.1 and Al–N–Al angles of 90.9°.
Ring angles in III are intermediate between these two
extremes with N–Li–N at 93.2, N–Al–N at 96.9 and
Li–N–Al at 83.6°. The dimeric rings in I and III are
Fig. 2. Molecular structure of 7 with hydrogen atoms omitted for
clarity.

286
W. Clegg et al. / Journal of Organometallic Chemistry 572 (1999) 283–289
Fig. 5. Energy minimum calculated structure of III, showing bond
˚
angles (°) as normal text, bond lengths (A) in italics and atomic
charges in brackets.
III, 16.9, 8.7, 18.0 and 20.5°, respectively. Hence, the
cause of this distortion is the reduction of repulsions
between adjacent methyl groups on the periphery of the
dimeric ring. Another important effect on moving from
I to III is the change in the N···N distance (N···N 2.78
Fig. 3. Energy minimum calculated structure of I, showing bond
angles (°) as normal text, bond lengths (A) in italics and atomic
charges in brackets.
˚
puckered with the sum of the endocyclic bond angles
being 358.0 for I and 357.6° for III. Puckering of the
ring aids in relieving torsional strain induced by adja-
cent methyl groups. If all the methyl groups in I are
replaced by hydrogens to give [{H2AlNH2}2] IV, the
ring reverts to planarity. Adding methyl groups to the
aluminium, i.e. [{Me2AlNH2}2] V, or the amido nitro-
gen in isolation, i.e. [{H2AlNMe2}2] VI, also has no
effect on the planarity of the ring. Placing methyl
groups on both the aluminiums and the amido nitro-
gens induces puckering. This is reflected in the torsional
angles R–Al–N–R (for syn substituents) lying between
5.1 and 7.8° for the planar rings in IV–VI, whereas
substantially more staggered angles are found in I and
˚
in I, 3.20 in II and 2.92 A in III) which occurs with a
concomitant change in metal–metal separations
˚
(Al···Al 2.84 in I, Li···Li 2.38 in II and Li···Al 2.65 A in
III). These effects combine to make formation of the
bimetallic compound thermodynamically favourable.
A comparison of theoretical and experimental bond
lengths and angles shows good agreement. One notice-
able difference between the calculated model III and the
experimental structure of 1 is the positioning of the
pyridine solvent molecule. In III the Al(1)–Li(1)–N(3)
angle is 173.0° whereas a much more acute angle of
139.8° is found in 1. This is due to the pyridine
molecule in the crystal 1 sitting out the ‘plane’ of the
¸¹¹º
buckled LiNAlN ring. To investigate if this is a conse-
quence of steric influences a calculation was run on a
more realistic model, where the methyl groups on the
amide were replaced by benzyl units to give the actual
Table 4
Calculated energies (Hartrees) of the complexes
Compound
Total energy (Hartrees)
I
II
−909.747528082
−775.726802458
−842.751823916
−597.490159408
−753.668756250
−753.574249292
−1760.93313543
−1693.92484358
III
IV
V
VI
VII
VIII
Fig. 4. Energy minimum calculated structure of II, showing bond
˚
angles (°) as normal text, bond lengths (A) in italics and atomic
charges in brackets.

W. Clegg et al. / Journal of Organometallic Chemistry 572 (1999) 283–289
287
for 7. Therefore changing methyl for benzyl only has a
relatively small effect on the solvation mode and cer-
tainly it is a very minor change in comparison to the
distinct pyramidalisation of the lithium in the hetero-
bimetallic systems.
A close examination of the crystal structure of 1 and
the model structure of VII gives some insight into this
pyramidalisation at lithium. One of the benzyl units is
significantly twisted towards the lithium centre leading
to relatively short Li···C(H) interactions. In 1 these
contact distances are Li(1)···C(10)benzyl 2.89,
˚
Li(1)···C(11)ipso 2.93 and Li(1)···C(16)ortho 2.63 A.
Similarly, from the calculated structure of VII the
contact distances are Li···C(benzyl) 2.88, Li···C(ipso)
3.05 and Li···C(ortho) 3.02 A. The significance of these
Fig. 6. Energy minimum calculated structure of VII, showing bond
lengths (A) in italics.
˚
˚
benzyl to lithium interactions is highlighted when com-
pared to the previously calculated structure of unsol-
vated monomeric lithium dibenzylamide [(PhCH2)2NLi]
IX [12]. The geometry optimised structure of IX has
both phenyl rings directed towards the lithium. The
shortest Li–C contacts are Li···C(benzyl) 2.85,
composition of 1 i.e. [Me2Al{N(CH2Ph)2}Li·pyr] VII
(Fig. 6).
The model complex VII optimised to a structure very
similar in geometry to that of 1. Indeed the Al–Li–
N(pyr) angle of 140.5° is remarkably close to the exper-
imental value of 139.8°. In the THF derivative 8, an
even more acute Al–Li–O(THF) angle of 130.8° was
found. In this case the larger size of THF compared to
the flat pyridine forces the solvent even further out of
the dimeric plane. In contrast, the homometallic lithium
amide complex II has planar solvation of the dimeric
ring by pyridine (Li–Li–N(pyr) 180.0°). As before, to
investigate if increasing the steric bulk on nitrogen
affects the solvation mode, a calculation was run by
replacing the methyl groups for benzyl units to give a
˚
Li···C(ipso) 2.90 and Li···C(ortho) 3.36 A. These inter-
actions, although not particularly short, account for a
stabilisation of −14.3 kcal mol−1 compared with the
geometry where the phenyl rings are directed away
from lithium. Hence, since the Li···C(H) contact dis-
tances calculated for VII are similar to those of IX, they
may make a contribution to the stability of the
heterodimer.
That this interaction occurs in the heterodimer and
not the homodimer is probably due to substituting a
trigonal planar lithium by a pseudo-tetrahedral alu-
minium. The projection of the methyl groups on alu-
minium above and below the plane of the dimeric ring
causes the proximate benzyl groups to move towards
the lithium. This is seen in the distortion of the angles
around one of the amido nitrogens (N(1)), where there
is significant tilting towards lithium (Li(1)–N(1)–
C(benzyl) mean angle in 1 is 107.8° and the Al(1)–
N(1)–C(benzyl) mean angle is 120.9°, compared with a
Li–N–C(benzyl) mean angle of 117.0° in 7). The re-
maining amido anion (N(2)) has a much smaller distor-
tion, with a mean Li(1)–N(2)–C(benzyl) at 114.6 and
Al(1)–N(2)–C(benzyl) at 115.1°. As expected it is the
amido anion with the greater distortion which is in-
volved in the Li(1)···C(16) interaction. This distortion
facilitates the movement of the phenyl ring towards the
lithium.
complex with the same composition as
7
i.e.
[{(PhCH2)2NLi·pyr}2] VIII (Fig. 7).
The calculated structure VIII was found to have an
Li–Li–N(pyr) angle of 170.9° which compares fa-
vourably with the experimental value of 168.1° found
Such benzyl to lithium interactions have previously
been noted in several structures, including unsolvated
lithium dibenzylamide which is a ring trimer [8]. Short
phenyl to lithium interactions were also found in the
bimetallic complex [(tBu)2Al{Ph3C(H)N}2Li] [9]. Note-
worthy is the fact that the phenyl to lithium interac-
tions in 1 persist even though donor solvent, in the
form of pyridine, is present. Generally, short contacts
Fig. 7. Energy minimum calculated structure of VIII, showing bond
angles (°) as normal text and bond lengths (A) in italics.
˚

288
W. Clegg et al. / Journal of Organometallic Chemistry 572 (1999) 283–289
of this type are thought to be relatively weak, and occur
only when the metal is forced into a low coordination
environment. It is therefore commonplace for these
interactions to disappear when Lewis base is present as
complexant [13].
dardised by titration with diphenylacetic acid before
use [16]. Trimethylaluminium was purchased from
Aldrich as a 1 M solution in hexane and used as
received. Lithium dibenzylamide was prepared by liter-
ature methods [8]. 1H-NMR spectra were recorded on a
Bruker AMX 400 spectrometer at r.t.
Another possible stabilising factor of the het-
erodimers is the relocation of electron density within
the molecule i.e. a ‘push–pull’ stabilisation [14]. This
was investigated by natural charge population analysis
of the homometallic and heterometallic complexes I–
III. Figs. 3–5 show the charges located on each atom.
On changing from the homodimers I and II to bimetal-
lic III there is a modest movement of electron density
away from lithium towards aluminium (difference in
charge is −0.06 for Al and +0.05 for Li). Apart from
the amido nitrogen atoms, which equilibrate to a value
of −1.11, the charges on the remaining atoms do not
alter to any significant degree. From these calculations
it appears that charge relocation makes only a minor
contribution in these systems. Charges on the calculated
structure of VII are consistent with the lithium to
phenyl interaction at the ortho position. The charge at
the ortho carbon in contact with lithium is −0.28
which is larger than that found for the average value of
−0.23 at the remaining ortho positions of VII (Mul-
liken charges quoted).
In conclusion, we have shown that heterobimetallic
combinations of dialkylaluminium amides and lithium
amides can be prepared by simply mixing the pre-
formed homometallic compounds in the presence of a
donor solvent. Ab initio calculations indicate that the
formation of the mixed metal species is energetically
favourable. A major factor contributing to this energy
gain is the alleviation of steric strain imposed in the
dialkylaluminium amide dimer starting material.
Changing from a system with four interconnected tetra-
coordinate atoms (i.e. two aluminiums and two nitro-
gens) to one in which there is one tricoordinate and
three tetracoordinate atoms allows an increased degree
of steric freedom. In addition, the bimetallic complex 1
gains significant stability from the formation of benzyl
to lithium interactions which exist in spite of the pres-
ence of donating solvent.
3.1.1. Synthesis of [Me2AlN(CH2Ph)2]
A 10 ml solution of 1 M trimethylaluminium (10
mmol) in hexanes was placed in a Schlenk tube and
cooled to 0°C. Dibenzylamine (10 mmol) was added
dropwise over a period of 10 min to the clear colourless
solution. The solution was allowed to warm to r.t. and
was then refluxed for 3 h. After cooling to r.t., the
solution was stored at −28°C for 2 days. Small crystals
of [Me2AlN(CH2Ph)2] were precipitated from solution
and isolated into an argon-filled glove box. Non-opti-
1
mised yield 56%. Melting point 128–130°C. H-NMR
(400 MHz, benzene-d6): l 0.07 (s, 6H, AlMe2), l 4.16
(s, 4H, PhCH2), l 7.03 (m, 10H, o, m, p-Ph)
3.1.2. Synthesis of [Me2Al(PhCH2)2NLi ·pyr] 1
Equimolar amounts (3 mmol) of lithium dibenzy-
lamide and [Me2AlN(CH2Ph)2] were added to a Schlenk
tube inside an argon-filled glove box. The Schlenk tube
was removed from the glove box and 10 ml of toluene
added to suspend the solids. Pyridine (3 mmol) was
added dropwise by syringe resulting in the dissolution
of the solids and the production of a deep red solution.
The solution was stirred for 2 h at r.t. and then placed
in a −28°C freezer. After 24 h a pale yellow solid was
precipitated from solution and isolated by filtration.
Non-optimised yield 39%. Melting point 119–121°C.
1H-NMR (400 MHz, benzene-d6): l 0.10 (s, 6H,
AlMe2), l 4.23 (s, 4H, PhCH2), l 6.49 (t, 2H, i-pyr),
l 6.85 (t, 1H, k-pyr), l 7.06 (t, 2H, p-Ph), l 7.16 (t, 4H,
o-Ph), l 7.23 (d, 4H, o-Ph), l 7.59 (d, 2H, h-pyr).
3.1.3. Synthesis of [{(PhCH2)2NLi ·pyr}2] 7
A Schlenk tube was charged with dibenzylamine (10
mmol) and 10 ml of toluene. BunLi (10 mmol as a 1.6
M solution in hexanes) was added dropwise to the 0°C
cooled solution. The solution was allowed to warm to
r.t. and stirred for 1 h. Pyridine (10 mmol) was added
to the solution resulting in the formation of a deep red
colour. Upon standing the solution at r.t. for 24 h,
crystals of 7 were deposited. Non-optimised yield 82%.
Melting point 170–172°C. 1H-NMR (400 MHz, d6-
DMSO): l 3.66 (s, 4H, PhCH2), l 7.22 (t, 2H, i-pyr),
l 7.37 (m, 10H, o, m, p-Ph), l 7.78 (t, 1H, k-pyr), l
8.58 (d, 2H, h-pyr).
3. Experimental
3.1. Syntheses
All solvents were distilled over sodium/benzophenone
until blue and stored over 4A molecular sieve before
use. Amines were distilled over CaH2 prior to use.
Standard Schlenk techniques [15] were employed for
the preparation and manipulation of the highly air and
moisture sensitive materials. All reactions were carried
out under a prepurified argon blanket. BunLi was stan-
3.2. X-ray crystallography
Crystallographic data is presented in Table 1. Data
for 1 were collected on a Rigaku AFC7S diffractometer

W. Clegg et al. / Journal of Organometallic Chemistry 572 (1999) 283–289
289
˚
[3] For lithium/gallium complexes see: (a) D.A. Atwood, A.H.
Cowley, R.A. Jones, Organometallics 12 (1993) 236. (b) G. Linti,
H. No¨th, K. Polborn, C. Robl, M. Schmidt, Chem. Ber. 128
(1995) 487. (c) M. Niemeyer, P.P. Power, Organometallics 14
(1995) 5488. (d) S.T. Barry, D.S. Richeson, Chem. Mater. 6
(1994) 2220. For an example of a lithium/indium complex see: (e)
M. Veith, M. Zimmer, S. Muller-Becker, Angew. Chem. Int. Ed.
Engl. 32 (1993) 1731.
fitted with a graphite monochromator (u=0.71069 A).
Data for 7 were collected on a Bruker AXS CCD
diffractometer with graphite-monochromated Mo–Kh
˚
radiation (u=0.71073 A). Absorption corrections were
unnecessary for both structures.
The structures were solved by automatic direct meth-
ods and refined on F2 values for all unique data, with
anisotropic displacement parameters and with a riding
model for isotropic H atoms. Programs were standard
diffractometer control software, members of the
SHELX family (G.M. Sheldrick, University of Go¨ttin-
gen, Germany), and local programs.
Complete crystallographic results have been de-
posited at the Cambridge Crystallographic Data
Centre.
[4] H.M.M. Shearer, R. Snaith, K. Wade, J. Chem. Soc. Chem.
Commun. (1971) 1275.
[5] (a) St. Bo¨ck, H. No¨th, P. Rahm, Z. Naturforsch. Teil B 43
(1988) 53. (b) M.M. Andrianarison, M.C. Ellerby, I.B. Gorrell,
P.B. Hitchcock, J.D. Smith, D.R. Stanley, J. Chem. Soc. Dalton
Trans. (1996) 211. (c) A. Heine, D. Stalke, Angew. Chem. Int.
Ed. Engl. 31 (1992) 854. (d) M.L. Montero, H. Wessel, H.W.
Roesky, M. Teichert, I. Uso´n, Angew. Chem. Int. Ed. Engl. 36
(1997) 629.
[6] D. Rutherford, D.A. Atwood, J. Am. Chem. Soc. 118 (1996)
11535.
[7] D.A. Armstrong, F.J. Craig, A.R. Kennedy, R.E. Mulvey, J.
Organomet. Chem. 550 (1998) 355.
3.3. Theoretical calculations
[8] D. Barr, W. Clegg, R.E. Mulvey, R. Snaith, J. Chem. Soc.
Chem. Commun. (1984) 285.
[9] M.A. Petrie, K. Ruhlandt-Senge, P.P. Power, Inorg. Chem. 32
(1993) 1135.
[10] (a) D.R. Baker, R.E. Mulvey, W. Clegg, P.A. O’Neil, J. Am.
Chem. Soc. 115 (1993) 6472. (b) P.C. Andrews, D.R. Armstrong,
D.R. Baker, et al., Organometallics 14 (1995) 427. (c) D. Barr,
W. Clegg, R.E. Mulvey, R. Snaith, J. Chem. Soc. Chem. Com-
mun. (1984) 285. (d) D.R. Armstrong, R.E. Mulvey, G.T.
Walker, et al., J. Chem. Soc. Dalton Trans. (1988) 617. (e) D.R.
Armstrong, M.G. Davidson, D. Moncrieff, Angew. Chem. Int.
Ed. Engl. 34 (1995) 478.
The Gaussian 94 program, revision E.2, was used for
the calculations [17]. For the larger molecules VII and
VIII geometry optimisation was carried out using the
HF/6-31G basis set as an initial guess and then reopti-
mised at the higher 6-31G* level. No constraints were
used in any of the optimisations.
Acknowledgements
[11] For a review see: J.J. Eisch in: G. Wilkinson, F.G.A. Stone,
E.W. Abel (Eds.), Comprehensive Organometallic Chemistry,
vol. 1, Pergamon Press, Oxford, 1982, ch. 6. For some examples
see: (a) H. Hess, A. Hinderer, S. Steinhauser, Z. Anorg. Allg.
Chem. 377 (1970) 1. (b) T. Belgardt, J. Storre, A. Klemp, et al.,
J. Chem. Soc. Dalton Trans. (1995) 3747. (c) T.Y. Her, C.C.
Chang, J.O. Tsai, et al., Polyhedron 12 (1993) 731.
[12] D.R. Armstrong, D.B. Baker, F.J. Craig, R.E. Mulvey, W.
Clegg, L. Horsburgh, Polyhedron 15 (1996) 3533.
We would like to thank the Royal Society for grant-
ing K.W. Henderson a University Research Fellowship
and the EPSRC for an equipment research grant (W.
Clegg). We also wish to thank BP Chemicals for partial
funding (A.E. McKeown).
References
[13] W. Clegg, K.W. Henderson, R.E. Mulvey, P.A. O’Neil, J. Chem.
Soc. Chem. Commun. (1994) 769.
[14] K.W. Henderson, A.E. Dorigo, Q-Y. Liu, P.G. Williard, P.v.R.
Schleyer, P.R. Bernstein, J. Am. Chem. Soc. 118 (1996) 1339.
[15] D.F. Shriver, M.A. Drezdzon, The Manipulation of Air-Senti-
tive Compounds, Wiley, New York, 1986.
[16] W.G. Kofron, L.M. Baclawski, J. Org. Chem. 4 (1976) 1879.
[17] M.J. Frisch, G.W. Trucks, H.B. Schlegel, et al., Gaussian Inc.,
Pittsburgh PA, 1995.
[1] (a) M.A. Beswick, D.S. Wright in: E.W. Abel, F.G.A. Stone, G.
Wilkinson (Eds.), Comprehensive Organometallic Chemistry,
vol. 1, 1995, ch. 1. (b) M.A. Beswick, N. Choi, C.N. Harmer, et
al., Inorg. Chem. 37 (1998) 2177. (c) R.E. Mulvey, Chem. Soc.
Rev. 27 (1998) 339.
[2] For uses of lithium reagents see: B.J. Wakefield, Organolithium
Methods, Academic Press, London, 1988.
.