Reaction of the diimine pyridine ligand with aluminum alkyls: An unexpectedly complex reaction

Article abstract of DOI:10.1021/om050936m

The diimine pyridine ligand 2,6-{2,6-ⁱPr₂C ₆H₃N=CMe}₂C₅H₃N (1) was reacted with a series of aluminum alkyls (Me₃Al, Et₃Al, ⁱBu₂Al, ⁱBU₂A1H, Et ₂AlCl). Depending on the choice of alkyl, addition to the imine carbon and the pyridine C2 and C4 positions was observed. Addition to C2 usually dominates but is reversible; the C4 alkylation product eventually dimerizes via double C-C coupling. Reaction of 1 with AlCl₃ gave the ionic complex [1·A1C1₂]⁺[AlCl₄]^-. DFT calculations were used to support NMR assignments of the various addition products and also to study alkyl transfer from Et₂AlCl to simplified model ligand 1′. Direct alkyl transfer from coordinated Et₂AlCl to the ligand C4 position is not possible. Introduction of a second molecule of Et₂AlCl results in formation of the ion pair [1′· AlEtCl]⁺[Et₃AlCl]^-, from which alkyl transfer to any position of the ligand is relatively easy. The dimerization of the C4-alkylated product is symmetry-forbidden and was calculated to follow a stepwise biradical path with an unusually low barrier.

Full text of DOI:10.1021/om050936m

1036
Organometallics 2006, 25, 1036-1046
Reaction of the Diimine Pyridine Ligand with Aluminum Alkyls:
An Unexpectedly Complex Reaction
Quinten Knijnenburg,† Jan M. M. Smits,† and Peter H. M. Budzelaar*,‡
Institute for Molecular Materials, Radboud UniVersity Nijmegen, ToernooiVeld 1,
6525 ED Nijmegen, The Netherlands, and Department of Chemistry, UniVersity of Manitoba,
Winnipeg, MB, R3T 2N2, Canada
ReceiVed October 31, 2005
The diimine pyridine ligand 2,6-{2,6-iPr2C6H3NdCMe}2C5H3N (1) was reacted with a series of
i
i
aluminum alkyls (Me3Al, Et3Al, Bu3Al, Bu2AlH, Et2AlCl). Depending on the choice of alkyl, addition
to the imine carbon and the pyridine C2 and C4 positions was observed. Addition to C2 usually dominates
but is reversible; the C4 alkylation product eventually dimerizes via double C-C coupling. Reaction of
1 with AlCl3 gave the ionic complex [1‚AlCl2]+[AlCl4]-. DFT calculations were used to support NMR
assignments of the various addition products and also to study alkyl transfer from Et2AlCl to simplified
model ligand 1′. Direct alkyl transfer from coordinated Et2AlCl to the ligand C4 position is not possible.
Introduction of a second molecule of Et2AlCl results in formation of the ion pair [1′‚AlEtCl]+[Et3AlCl]-,
from which alkyl transfer to any position of the ligand is relatively easy. The dimerization of the C4-
alkylated product is symmetry-forbidden and was calculated to follow a stepwise biradical path with an
unusually low barrier.
main-group metals Li,10 Al,11,12 Zn, and Mg.13 Alkylation can
occur at various positions of the ligand framework: so far,
examples of alkyl attack at the imine carbon,3,11,12 the pyridine
C23,7,8 and C4,9 and nitrogen atoms10,13 have been reported, as
well as double alkylation (at both C2 and C3 atoms).8 While
for this particular system ligand-centered reactivity has only
recently attracted attention, chemical noninnocence of the related
R-diimine and imine-pyridine systems has been known for a
long time.14 A recent report shows that also for these ligands
alkylation can be reversible.15
The reaction of aluminum alkyls with diimine pyridine ligands
is particularly important for two reasons. In the first place,
aluminum alkyls are the cocatalysts of choice for activating
transition metal complexes of diimine pyridine ligands in olefin
polymerization16 and hydrogenation.17 Secondly, the products
of the reaction of aluminum alkyls with this ligand have been
reported to generate transition-metal-free olefin polymerization
catalysts on activation with Lewis acids.11
Introduction
Since the parallel discovery by the groups of Brookhart and
Gibson of the performance of iron and cobalt complexes bearing
diimine pyridine ligands in olefin oligomerization and polym-
erization,1 numerous studies have been dedicated to this
particular ligand. In the case of cobalt, we and others showed
that MAO achieves reduction of the dichloride complex, prior
to alkylation in the activation pathway.2 For iron, dialkyl
complexes can be formed under the right circumstances,3,4 but
reduction of the metal center3,5 and (reversible) attack of alkyl
groups on the ligand framework3 have also been observed.
Furthermore, aluminum alkyls can displace the iron atom
partially or completely6 from the ligand. Ligand alkylation has
also been observed with the transition metals Ti,7 V,8 and Cr9
(in combination with various alkylating agents) and with the
* Corresponding author. E-mail: budzelaa@cc.umanitoba.ca.
† Radboud University.
‡ University of Manitoba.
(1) (a) Small, B. L.; Brookhart, M.; Bennett, A. A. J. Am. Chem. Soc.
1998, 120, 4049. (b) Britovsek, G. J. P.; Gibson, V. C.; Kimberley, B. S.;
Maddox, P. J.; McTavish, S. J.; Solan, G. A.; White, A. J. P.; Williams, D.
J. Chem. Commun. 1998, 849. (c) Small, B. L.; Brookhart, M. J. Am. Chem.
Soc. 1998, 120, 7143. (d) Britovsek, G. J. P.; Bruce, M.; Gibson, V. C.;
Kimberley, B. S.; Maddox, P. J.; Mastroianni, S.; McTavish, S. J.; Redshaw,
C.; Solan, G. A.; Stro¨mberg, S.; White, A. J. P.; Williams, D. J. J. Am.
Chem. Soc. 1999, 121, 8728.
(2) (a) Kooistra, T. M.; Knijnenburg, Q.; Smits, J. M. M.; Horton, A.
D.; Budzelaar, P. H. M.; Gal, A. W. Angew. Chem., Int. Ed. 2001, 40,
4719. (b) Gibson, V. C.; Humphries, M.; Tellmann, K.; Wass, D.; White,
A. J. P.; Williams, D. J. Chem. Commun. 2001, 2252.
(3) Scott, J.; Gambarotta, S.; Korobkov, I.; Budzelaar, P. H. M. J. Am.
Chem. Soc. 2005, 127, 13019.
(9) Sugiyama, H.; Aharonian, G.; Gambarotta, S.; Yap, G. P. A.;
Budzelaar, P. H. M. J. Am. Chem. Soc. 2002, 124, 12268.
(10) (a) Khorobkov, I.; Gambarotta, S.; Yap, G. P. A.; Budzelaar, P. H.
M. Organometallics 2002, 21, 3088. (b) Clenthsmith, G. K. B.; Gibson, V.
C.; Hitchcock, P. B.; Kimberley, B. S.; Rees, C. W. Chem. Commun. 2002,
14, 1498.
(11) Bruce, M.; Gibson, V. C.; Redshaw, C.; Solan, G. A.; White, A. J.
P.; Williams, D. J. Chem. Commun. 1998, 18, 2523.
(12) Milione, S.; Cavallo, C.; Tedesco, C.; Grassi, A. J. Chem. Soc.,
Dalton Trans. 2002, 8, 1839.
(13) Blackmore, I. J.; Gibson, V. C.; Hitchcock, P. B.; Rees, C. W.;
Williams, D. J.; White, A. J. P. J. Am. Chem. Soc. 2005, 127, 6012.
(14) (a) Klerks, J. M.; Stufkens, D. J.; Van Koten, G.; Vrieze, K. J.
Organomet. Chem. 1979, 181, 271. (b) Van Koten, G.; Jastrzebski, J. T. B.
H.; Vrieze, K. J. Organomet. Chem. 1983, 250, 49.
(4) Bouwkamp, M. W.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc.
2005, 127, 9660.
(5) Bouwkamp, M. W.; Bart, S. C.; Hawrelak, E. J.; Trovitch, R. J.;
Lobkovsky, E.; Chirik, P. J. Chem. Commun. 2005, 3406.
(6) Scott, J.; Gambarotta, S.; Korobkov, I.; Knijnenburg, Q.; De Bruin,
B.; Budzelaar, P. H. M. J. Am. Chem. Soc., submitted.
(7) Calderazzo, F.; Englert, U.; Pampaloni, G.; Santi, R.; Sommazzi, A.;
Zinna, M. J. Chem. Soc., Dalton Trans. 2005, 914.
(15) Riollet, V.; Cope´ret, C.; Basset, J.-M.; Rousset, L.; Bouchu, D.;
Grosvalet, L.; Perrin, M. Angew. Chem., Int. Ed. 2002, 41, 3025.
(16) Liu, J.; Li, Y.; Liu, J.; Li, Z. Macromolecules 2005, 38, 2559.
(17) (a) Horton, A. D.; Knijnenburg, Q.; Van der Heijden, H.; Budzelaar,
P. H. M.; Gal, A. W. WO 03042131 (2003); EP 2001-204376 (2001). (b)
Knijnenburg, Q.; Horton, A. D.; Van der Heijden, H.; Kooistra, T. M.; De
Bruin, B.; Hetterscheid, D. G. H.; Smits, J. M. M.; Budzelaar, P. H. M.;
Gal, A. W. J. Mol. Catal. A-Chem. 2005, 232, 151.
(8) Reardon, D.; Conan, F.; Gambarotta, S.; Yap, G.; Wang, Q. J. Am.
Chem. Soc. 1999, 121, 9318.
10.1021/om050936m CCC: $33.50 © 2006 American Chemical Society
Publication on Web 01/21/2006

Diimine Pyridine and Aluminum Alkyls
Organometallics, Vol. 25, No. 4, 2006 1037
Scheme 1. Ligand 1
Scheme 2. Alkylation Product Types Observed in This
Work
Figure 1. Thermal ellipsoid plot of 2a (50% probability). Hydrogen
atoms and isopropyl groups are omitted for clarity.
Figure 2. Thermal ellipsoid plot of 3b (50% probability). Hydrogen
atoms and isopropyl groups are omitted for clarity.
(8) has not been reported so far. We will be considering these
alternative products in the theoretical parts of this paper.
Scheme 3. Alternative Alkylation Product Types
Considered but Not Observed in This Work
Results and Discussion
We have investigated the stoichiometric reactions of alumi-
num alkyls with the free ligand 1. In all of these reactions, a
mixture of products is formed initially. Prolonged heating
changes the composition of the mixture, but never in such a
way that only a single product remains. Thus, pure compounds
could only be obtained in those cases where a single product
crystallized preferentially (other separation techniques failed).
Fortunately, we were able to isolate pure samples of a C2
alkylation product (3b) and two dimers of C4 alkylation products
(5d,18 5e) in this way. One imine alkylation product (2d) was
also obtained in crystalline form, but not in sufficient quantities
for NMR characterization;18 therefore, we prepared pure samples
of 2a and 2b following the route previously reported by
Gibson,11 using excess aluminum alkyl. Comparison of the
spectra of purified 2 and 3 with those of the reaction mixtures
allowed easy assignment of such alkylation products in all
mixtures. The C4 alkylation products could not be isolated pure
because they dimerize slowly as they are formed. They have,
however, very characteristic signals for the (former) pyridine
ring protons, which make assignment straightforward. As a
further support for the NMR assignments, we calculated 1H and
13C NMR shifts for the various products (potentially) formed
with Me3Al and Et3Al using the GIAO method. In the following
sections, we first discuss the pure compounds we managed to
isolate and the assignment of their spectra and then use these
to interpret the progress of the addition reactions.
For these reasons, we decided to study the reactions of ligand
1 (Scheme 1) with several aluminum alkyls in detail. Gibson
reacted both 1 and related bis(aldimine) pyridine ligands with
aluminum alkyls and from these reactions obtained the products
of addition to the imine CdN bond in pure form.11 Milione
studied the reaction of 1 with Et2AlCl12 and mentions com-
plexation and alkylation at the imine carbon as the main or sole
reactions occurring. As we will show below, the reaction is
much more complicated than indicated by these earlier studies
and features alkyl addition to at least three positions of the ligand
as well as dimerization of one of the alkylation products.
Depending on the choice of alkyl and reaction conditions, we
have observed alkylation at the imine position (2) and at pyridine
C2 (3) or C4 (4) as well as dimerization of 4 to its polycyclic
dimer 5 (see Scheme 2). Part of this work has been com-
municated.18
Additional alkylation products could in principle be formed,
as shown in Scheme 3. As mentioned above, products analogous
to 6 have been observed for Li,10 Mg, and Zn;13 alkylation at
C3 (without preceding C2 alkylation: 7) or at the imine nitrogen
Individual Alkylation Products.
Imine Alkylation: 2. Pure samples of complexes 2a and
2b, for use as reference materials in the analysis of mixture
spectra, were prepared by the method of Gibson11 (treating 1
with 2 equiv of Me3Al or Et3Al, 12 h of reflux in toluene,
removal of the solvent, crystallization from toluene). We also
(18) Knijnenburg, Q.; Smits, J. M. M.; Budzelaar, P. H. M. C. R. Chimie
2004, 7, 865.

1038 Organometallics, Vol. 25, No. 4, 2006
Knijnenburg et al.
Table 1. Calculated and Observed Chemical Shifts for 3b
1H
13C
position
Py 2
calc
obs
calc
obs
73.4
67.5
Py 3
Py 4
Py 5
Py 6
2-CH2Me
2-CH2Me
C(Me)dN
C(Me)dN
AlCH2Me
5.12
6.33
5.71
5.12
6.29
5.66
114.4
134.3
105.6
152.5
35.9
101.7
126.2
114.2
146.2
28.6a
6.2
183.1, 173.8
17.9, 16.8
1.4, -0.5
2.86, 0.97
1.08
2.4, 1.3
0.9
9.3
194.1, 181.7
23.2, 19.4
4.7, 1.7
2.05, 2.03
1.75, 1.55
0.25-0.0
-0.13, -0.14,
-0.20, -0.33
0.98, 0.31
AlCH2Me
Ar i,o
1.19, 0.74
13.6, 11.9
10.8, 8.7
150.3, 150.2, 147.5,
147.6, 146.7, 146.5
130.8, 130.5, 129.9,
129.5, 129.0, 128.6
33.5, 33.3, 32.2, 32.1
28.0, 27.8, 27.3, 26.7,
26.3, 25.5, 24.3, 24.0
143.2, 143.0, 141.1,
140.5, 140.0, 140.0
126.7, 126.5, 124.6,
124.5, 124.4, 123.8
28.3, 27.7, 27.7, 27.5a
26.0, 25.4, 25.0, 25.0,
24.9, 24.1, 23.7, 23.1
Ar m,p
7.14, 7.13, 7.12,
7.09, 7.09, 7.02
3.15, 2.87, 2.76, 2.60
1.45, 1.32, 1.28, 1.25,
1.07, 1.06, 0.97, 0.90
7.1
Ar CHMe2
Ar CHMe2
3.21, 3.09, 3.07, 3.01
1.41, 1.41, 1.39, 1.37,
1.11, 1.00, 0.99, 0.97
a Assignment tentative.
1
Table 2. Calculated and Observed Characteristic H
Resonances for Various Complexes
Table 3. Composition of Reaction Mixture of 1 with Al
Alkyls, Immediately after Mixing, in mol %
complex
position
calc
obs
Al alkyl
2
3
4
1
Py 3,5
Py 4
Ar CHMe2
Py 3
Py 4
Py 5
Py 3,5
Py 4
Py 3,5
Py 4
Py 3
Py 4
Py 5
Py 3
Py 4
Py 5
Ar CHMe2
Py 3
Py 4
Py 5
Py 3,5
Py 4
8.76
7.86
3.69, 2.84
5.05
6.29
5.77
5.10
3.92
6.56
4.88
3.54
5.49
6.38
6.33
4.94
5.64
3.88, 3.72, 2.50, 2.70
5.12
6.33
5.71
5.18
3.76
5.03
3.97
8.15
8.40
Me3Al (a)
Et3Al (b)
66
17
9
29
55
78
3a
28
13
6b
≈100
7.22
3.97, 2.94
5.16
6.32
5.74
5.03
3.52
a
a
a
a
a
2a
3a
Et2AlCl (d+d′)
iBu3Al (c)
iBu2AlH (e)
a Plus ca. 2 mol % 4f. b Plus ca. 1 mol % of 4e and ca. 3 mol % of a
second 4-iBu alkylation product; remainder appears to consist of coordination
complex(es) of 1.
4a
6a
7a
Table 4. Calculated Relative Stabilities (kcal/mol) for
Reaction Products of 1 with Me3Al and Et3Al
product
RIDFT-bp86/SV(P)
B3LYP/6-311G**a
8a
a
a
a
2a
3a
4a
6a
7a
8a
13.2
13.6
2.6
20.4
(0)
9.0
10.8
(0)
22.1
2.0
2b
3b
4.38, ...
5.12
6.29
5.66
5.03
3.7
4.94
3.53
7.92
8.63
a
27.0
29.0
4b
4f
2b
3b
4b
5b
16.5
14.5
15.5
15.0
(0)
Py 3,5
Py 4
Py 3,5
Py 4
(0)
-15.3
-4.8
+
(κ3-1)AlCl2
a At RIDFT-bp86/SV(P) geometries.
8.60
7.35
(κ1-im-1)AlMe3
Py 3
Py 4
Py 5
Py 3,5
Py 4
7.75
8.65
8.57
7.90
a
a
a
a
side of the complex. Apart from this, they show no resonances
that allow easy identification and quantification in mixtures.
From the reaction of 1 with Et2AlCl, we were able to isolate
a few orange crystals of complex 2d by “crystal picking”.
Although the quality of the crystals was poor, the connectivity
of the ligand framework could be confirmed by X-ray diffrac-
tion.18
(κ1-py-1)AlMe3
a No signals attributable to these compound were observed in any spectra.
determined the structure of 2a (Figure 1; for bond lengths in
this and other complexes, see Table 6); its geometry is fully
comparable to the analogous product of the reaction between
an aldimine ligand and AlMe3 as reported by Gibson.11 The
central Al is tetrahedral with a short Al-N3 bond (1.878(3) Å)
because N3 has now become anionic; the Al-N2 distance is
very long (2.591(5) Å) and probably does not represent a real
coordination bond.
Compared to complexes 3 and 4 and the free ligand 1,
complexes 2 show a characteristic low-field isopropyl methine
resonance, which (according to NOESY and calculations
mentioned below) is due to an isopropyl group on the “reacted”
Pyridine C2 Alkylation: 3. Pure samples of 3b could be
obtained by reacting 1 with Et3Al at room temperature in hexane,
cooling to -30 °C, and washing with cold hexane. According
1
to H NMR (see below and Table 3), 3b is initially formed in
about 55% yield, but the above partial crystallization and
washing procedure obviously results in significant losses, and
the isolated yield of 3b is typically about 10%.
The 1H NMR spectrum of 3b shows three characteristic
olefinic resonances for the (former) pyridine ring protons in the
region of 5.0-7.0 ppm, with clear couplings between them. The
aromatic, methine, and methyl resonances are unremarkable.

Diimine Pyridine and Aluminum Alkyls
Organometallics, Vol. 25, No. 4, 2006 1039
Table 5. Crystal Data and Structure Analysis Results
2a
3b
5e
9
cryst color
translucent brown-red
transparent dark
yellow-brown
regular thick platelet
0.32 × 0.23 × 0.10
C39H58AlN3
dark red
transparent light
yellow
rough fragment
0.19 × 0.13 × 0.10
C40H51Al2Cl6N3
840.50
cryst shape
cryst size (mm)
emp formula
fw
rough fragment
regular fragment
0.19 × 0.16 × 0.13
C82H124Al2N6
1247.83
0.31 × 0.31 × 0.22
C
46.50 H64AlN3
691.99
595.86
temp (K)
208(2)
Mo KR (graphite monochromated)/0.71073
radiation/wavelength (Å)
cryst syst space group
no. of reflns for cell
a (Å)
b (Å)
c (Å)
triclinic P1h
42
triclinic P1h
40
monoclinic P21/c
70
monoclinic P21/n
110 254
24.292(5)
13.808(3)
27.336(5)
90
102.06(3)
90
8966(3)
8, 1.245
0.453
9.6818(6)
10.8761(7)
21.6405(17)
100.973(6)
99.380(5)
104.004(5)
2117.7(3)
2, 1.085
12.8955(15)
16.811(3)
17.1893(19)
87.810(19)
79.367(11)
89.436(18)
3659.7(9)
4, 1.081
10.653(3)
16.9159(15)
21.622(3)
90
95.783(16)
90
3876.7(13)
2, 1.069
0.082
R (deg)
â (deg)
γ (deg)
volume (Å3)
Z, dcalc (Mg/m3)
abs coeff (mm-1
)
0.081
0.084
diffractometer/scan
F(000)
Nonius KappaCCD with area detector θ and ω scan
754
1304
1368
3520
θ range for data collection (deg)
2.22 to 25.00
-11 e h e 11
-12 e k e 12
-25 e l e 25
25 674/7437
[Rint ) 0.0474]
4772
3.07 to 24.00
-14 e h e 14
-19 e k e 19
-19 e l e 19
98 451/11 472
[Rint ) 0.1142]
7568
3.06 to 25.00
-12 e h e 12
-20 e k e 20
-25 e l e 25
65 449/6811
[Rint ) 0.1215]
500
4.51 to 22.00
-25 e h e 25
-14 e k e 14
-28 e l e 28
11 0254/10903
[Rint ) 0.1279]
7112
index ranges
reflns collected/unique
no. of obsd reflns ([Io>2σ(Io)])
abs corr
refinement
SADABS multiscan correction46
full-matrix least-squares on F2
computing
SHELXL-9747
no. of data/restraints/params
goodness-of-fit on F2
SHELXL-97 weight params
final R indices
[I > 2σ(I)]
R indices
7437/15/483
1.032
11 472/0/801
1.099
6811/0/420
1.068
10 903/0/941
1.042
0.0633, 2.4332
R1 ) 0.0701
wR2 ) 0.1544
R1 ) 0.1220
wR2 ) 0.1802
0.782 and -0.371
0.0347, 3.4821
R1 ) 0.0749
wR2 ) 0.1271
R1 ) 0.1251
wR2 ) 0.1431
0.223 and -0.242
0.0771, 3.5378
R1 ) 0.0728
wR2 ) 0.1737
R1 ) 0.1022
wR2 ) 0.1916
0.529 and -0.402
0.0912, 48.6697
R1 ) 0.0968
wR2 ) 0.2133
R1 ) 0.1474
wR2 ) 0.2422
1.757 and -1.457
(all data)
largest diff peak and hole (e/Å3)
Table 6. Selected Bond Distances (Å) and Angles (deg)
in the Supporting Information. Because of the addition, the
former pyridine ring is no longer planar, and the negative charge
is localized on its nitrogen atom. This produces a short Al-N1
bond (1.885(3) Å) and localized pyridine double bonds (1.364(5)
and 1.333(5) Å); the localization is somewhat less pronounced
than in the vanadium-containing C2 alkylation product reported
by Gambarotta.8 The geometry around the central Al atom is
distorted square pyramidal.
Pyridine C4 Alkylation Followed by Dimerization: 4 and
5. Mixtures of 1 and various aluminum alkyls frequently showed
a characteristic doublet at 4.9 ppm (coupled to a less distinctive
multiplet around 3.5 ppm) and/or two triplets at 4.8 and 3.5
ppm, each with a coupling of about 3.8 Hz. We assign these to
the products 4 of addition to the pyridine C4 position of an
alkyl group or a hydride, respectively. In most spectra, the
amounts of these products are small, but in the reaction of Et3Al
with 1 complex 4b can comprise up to 25 mol % of the product
mixture. We were unable to isolate any complexes of this type
in pure form. Our assignment is based on the following
arguments:
2a
3b
5e
9
Al-N(1)
2.036(2)
2.591(5)
1.878(3)
1.967(3)
1.979(3)
1.497(5)
1.283(4)
1.476(5)
1.355(4)
1.379(4)
1.380(5)
1.371(5)
1.389(4)
1.331(4)
1.513(4)
1.544(5)
1.463(4)
1.526(5)
1.885(3)
2.240(3)
2.221(3)
1.980(4)
1.981(4)
1.504(4)
1.284(4)
1.475(5)
1.366(4)
1.364(5)
1.433(5)
1.333(5)
1.513(5)
1.455(4)
1.516(5)
1.506(5)
1.278(4)
2.010(2)
2.147(2)
2.120(2)
2.026(2)
2.010(3)
1.493(4)
1.312(3)
1.436(4)
1.333(3)
1.503(3)
1.527(4)
1.535(4)
1.508(3)
1.347(3)
1.415(4)
1.502(4)
1.317(3)
1.958(6)
2.083(6)
2.076(6)
Al-N(2)
Al-N(3)
Al-C(41)
Al-C(51)
C(1)-C(2)
C(2)-N(2)
C(2)-C(3)
C(3)-N(1)
C(3)-C(4)
C(4)-C(5)
C(5)-C(6)
C(6)-C(7)
C(7)-N(1)
C(7)-C(8)
C(8)-C(9)
C(8)-N(3)
C(8)-C(61)
C(7)-(C61)
C(4)-C(6)#1
Al-Cl(1)
1.500(11)
1.289(9)
1.471(11)
1.322(9)
1.396(11)
1.370(12)
1.374(12)
1.391(11)
1.338(10)
1.475(11)
1.487(11)
1.286(9)
1.561(5)
1.557(4)
2.121(3)
2.112(3)
149.4(3)
Al-Cl(2)
• The shift and coupling are similar to those observed
previously for a C4 hydride addition product containing Rh19
and to shifts and couplings reported for more “normal”
4-hydropyridyl units.20
N(2)-Al-N(3)
N(1)-Al-C(41)
N(1)-Al-C(51)
152.34(11) 146.71(9)
138.63(14) 114.25(14) 142.35(12)
97.82(13) 133.02(14) 102.14(11)
C(41)-Al-C(51) 114.25(17) 112.70(16) 115.31(13)
N(1)-Al-Cl(1)
N(1)-Al-Cl(2)
Cl(1)-Al-Cl(2)
101.5(2)
146.2(2)
112.29(13)
• From the reaction of Et2AlCl with 1, we were able to isolate
and characterize dimer 5d of C4 alkylation product 4d.18
(19) Kooistra, T. M.; Hetterscheid, D. G. H.; Schwartz, E.; Knijnenburg,
Q.; Budzelaar, P. H. M.; Gal, A. W. Inorg. Chim. Acta 2004, 357, 2945.
(20) See e.g.: Kita, Y.; Maekawa, H.; Yamasaki, Y.; Nishiguchi, I.
Tetrahedron 2001, 57, 2095.
The crystal structure of 3b contains two independent molecules.
Figure 2 shows one of these, and Table 6 gives geometrical
parameters for it; full details for both molecules can be found

1040 Organometallics, Vol. 25, No. 4, 2006
Knijnenburg et al.
cannot be used, and we resorted to calculations,22 using the
GIAO method23 (as implemented in Gaussian9824) at the
B3LYP25/6-311G**26 level. Perfect agreement is not expected
for several reasons.27 Nevertheless, we hoped that calculated
data might be good enough to aid spectrum interpretation and
assignments. Table 1 compares all calculated and observed shifts
for complex 3b. Table 2 lists calculated and observed shifts of
complexes 2a-4a, 6a-8a, and 2b-4b and related species in
the “characteristic” regions of 3.5-6.5 and 7.5-10.0 ppm,
which are not obscured by aliphatic or aromatic ligand signals.
The agreement is seen to be quite good. Moreover, the
calculations confirm that no other characteristic resonances
should be observed for any of the products discussed here.
Formation of pyridine N-alkylation products 6, as reported
recently for Li,10 Mg, and Zn,13 can be ruled out here both on
the basis of their calculated NMR shifts and by comparison with
the NMR data for the Mg and Zn analogues. The C3 (7) and
imine-N (8) alkylation products are also calculated to have very
characteristic shifts, allowing us to rule out their formation as
well. In our opinion, calculation of NMR shifts can be a valuable
tool in the assignment of spectra of mixtures.
Following the Reactions of 1 with Al Alkyls. The free ligand
and aluminum alkyl (Me3Al, Et3Al, Et2AlCl, iBu3Al, or iBu2AlH)
were reacted in toluene (CH2Cl2 for iBu2AlH). The solvent was
then removed in vacuo, the residue was dissolved in toluene-
d8, and a spectrum was recorded. For Me3Al and Et2AlCl, the
mixture was then sealed in an NMR tube, and spectra were
recorded after keeping the sample for hours or days at elevated
temperature. For Et3Al, the main product (3b) was first isolated
and sealed in an NMR tube in toluene-d8, and its further
rearrangements were followed by NMR. Table 3 lists the
composition of the initial reaction mixture for all compounds,
and Figure 4 summarizes the variations in concentration on
heating.
Figure 3. Thermal ellipsoid plot of 5e (50% probability). Hydrogen
atoms and isopropyl groups are omitted for clarity.
• A spectrum recorded immediately after mixing 1 and
iBu2AlH appears to consist of mainly 4e, with no other
complexes recognizable in solution. A crystal obtained from
this solution was of very poor quality, but X-ray structure
determination confirmed its connectivity (see Supporting In-
formation). From the mixture, we were able to isolate dimer 5e
after heating and workup.
• The shifts of the H3,5 protons and (where assignable) the
H4 protons show satisfactory agreement with the calculated
shifts.
In parallel experiments, EPR spectra were recorded of the
reaction mixtures. These measurements indicated the presence
The X-ray structures of 5d18 and 5e (Figure 3) show a tricyclic
ligand skeleton, similar to that reported by Gambarotta for the
product of the reaction of 1‚CrCl3 with BzMgCl.9 In 5d, as in
the Cr analogue, the alkyl groups added to the ligand C4 position
of each pyridyl unit avoid the second monomer unit. The Al
atoms are five-coordinate with the usual square pyramidal
geometry.
(22) See e.g.: Sefzik, T. H.; Turco, D.; Iuliocci, R. J.; Facelli, J. C. J.
Phys. Chem. A 2005, 109, 1180, and references therein.
(23) Wolinski, K.; Hilton, J. F.; Pulay, P. J. Am. Chem. Soc. 1990, 112,
8251; Dodds, J. L.; McWeeny, R.; Sadlej, A. J. Mol. Phys. 1980, 41, 1419;
Ditchfield, R. Mol. Phys. 1974, 27, 789; McWeeny, R. Phys. ReV. 1962,
126, 1028; London, F. J. Phys. Radium (Paris) 1937, 8, 397.
(24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Rega,
N.; Salvador, P.; Dannenberg, J. J.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A.
G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.;
Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.;
Replogle, E. S.; Pople, J. A. Gaussian 98, Revision A.11.4; Gaussian, Inc.:
Pittsburgh, PA, 2002.
1
Neither 5d nor 5e gave an acceptable H NMR spectrum in
benzene, toluene, or any other solvent we tested. This might be
simply due to their surprisingly low solubility. However, the
intense purple color of the complexes in the solid state and as
dilute toluene solutions suggests that easy formation of para-
magnetic species might also play a role.
Calculation of NMR Shifts. Given the impossibility of
separating isomeric complexes obtained in most reactions of
aluminum alkyls with 1, we decided to obtain more support for
the NMR assignments of the various isomeric structures through
prediction of the expected 1H and 13C NMR shifts. For
“ordinary” organic compounds, shift prediction based on refer-
ence molecules, group contributions, and empirical corrections21
is probably the most reliable method available at present. For
the present cases, involving organometallic complexes with
unusual and highly delocalized anionic ligands, such a procedure
(25) Becke, A. D. J. Chem. Phys. 1996, 104, 1040.
(26) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639.
Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980,
72, 650.
(27) (a) Solvent effects are not included. Electronic solvent effects are
expected to be small, but the preferred orientation of highly anisotropic
solvent molecules such as benzene or toluene can still have large effects in
particular on 1H NMR shifts. (b) The molecules were optimized at a
relatively low level (RI-bp86/SV(P), see computational details); optimization
at a higher level was not feasible for these large species. (c) These molecules
all have considerable conformational freedom. Exploring the full confor-
mational space was not feasible, so we studied only a single conformation
for each complex. This may be important in particular for ethyl complexes
(where at least three conformations of each Et group should be considered)
and for 2 (where the five-membered NAlNCC ring can pucker in different
ways).
(21) See e.g.: (a) Satoh, H.; Koshino, H.; Uno, T.; Koichi, S.; Iwata,
S.; Nakata, T. Tetrahedron 2005, 61, 7431. (b) Abraham, R. J.; Byrne, J.
J.; Griffiths, L.; Koniotou, R. Magn. Reson. Chem. 2005, 43, 611, and
references in these papers.

Diimine Pyridine and Aluminum Alkyls
Organometallics, Vol. 25, No. 4, 2006 1041
Figure 4. Changes in composition (mol %) of 1 + aluminum alkyl mixtures. Note the ×10 scale used for 4a, 4f, and 4b.
Scheme 4. Diastereomeric Products that Could Be Formed
of LAlR2 radicals in amounts varying from 1% (Me3Al) to ca.
from 1 and Et2AlCl
0.01% (iBu3Al). These radicals have been discussed in a separate
paper.6 Gibson also noted formation of paramagnetic species
in reactions of diimine pyridine ligands with alkylmagnesium
and alkylzinc compounds.13
Reaction of Me3Al with 1. The initial mixture showed
formation of two major products, resulting from imine (2a) and
C2 (3a) alkylation, formed in a ca. 1:2 ratio. In addition, small
quantities of 4a and 4f28 (ca. 2 mol % each) could be observed.
On heating, 3a slowly converted to 2a, but the reaction appeared
to stop at a 2a:3a ratio of about 3:1; even prolonged heating
did not cause further conversion. Complex 4f already disap-
peared early in the reaction, while 4a remained visible a little
longer (both probably end up as dimers, as described below for
other alkyls).
From these observations, we draw the following conclusions:
• C2 alkylation product 3a is one of the initial reaction
products, and it can rearrange to 2a, most likely via reversion
close to 1:1. Instead, it resulted in slow decrease of all NMR
signals attributable to derivatives of 1.
From these observations for Et3Al we conclude the following:
of the alkyl addition.
• Complex 2b can be formed via 3b, like 2a from 3a. The
• The formation of 2a from 3a is too slow to explain the
surprise is that we now see hardly any direct formation of 2b.
initial formation of ca. 30% of 2a. This means that either 2a is
Complex 4b can be formed via 3b; whether it can also be
in part also an initial product (competing with 3a) or conditions
formed directly cannot be established with certainty.
later on in the reaction are less favorable for converting 3a to
• The most plausible explanation for the “stopping” of the
2a (possibly because the conversion is catalyzed by free Me3Al,
conversion of 3b to 2b is the attainment of chemical equilibrium
at a ca. 1:1 ratio; similarly, the above constant 2a:3a ratio then
most likely represents the equilibrium composition for the
methyl case. It seems reasonable that steric hindrance would
which at that time has been consumed).
• The final 3:1 2a:3a ratio might represent an equilibrium
composition (at 120 °C), or it might be that the reaction stopped
at this point because any factor that assisted the reaction at an
be more severe for alkylation at the imine carbon than at
earlier stage has disappeared at this point.
pyridine C2, and higher for ethyl than for methyl groups, which
is consistent with the observed ratios (2a/3a > 2b/3b).
It seems clear that 3a is a kinetic product, which for the largest
part slowly converts to thermodynamically more stable 2a, the
The cause of the eventual decrease of the signals is not clear
product reported previously by Gibson.
at present. Some solid material is formed, but not enough to
Reaction of Et3Al with 1. The initial reaction mixture here
explain the dramatic decrease in signal intensity. It could be
contained little of the imine alkylation product. Instead, the
that the eventual product is dimer 5b of 4b (we have never
major products were 3b and 4b, in a ratio of about 2:1.30 On
observed NMR signals for such dimers).
Reaction of Et2AlCl with 1. The initial reaction mixture
showed mainly pyridine ring alkylation to 3d and 4d (ca. 6:1);
heating, the concentration of 3b slowly decreased and 2b was
formed as a major product. We monitored the behavior of
purified 3b on heating. On gentle heating, it partially converted
traces of a second isomer 3d′ of 3d were also visible, as was
to 2b and a small quantity of 4b. Prolonged heating at higher
about 9 mol % of 2d. On heating, larger amounts of 2d/2d′
temperature did not change the ratio 2b:3b, which remained
were formed, and the intensities of 3d and 3d′ became similar.
Conversion of 3d/3d′ to 2d/2d′ seemed to stop at a 2:3 ratio of
about 3:1.
(28) The formation of 4f, a product of hydride addition, is somewhat
surprising here, since Me3Al does not contain significant quantities of
hydride impurities. One possible route is via intermoleculer transfer of the
pyridine H4 from a molecule of 4a to unreduced 1, similar to the transfer
we proposed to explain 1,4-dihydropyridyl formation from 1‚RhCl and
Et2Zn.19 This type of hydrogen transfer is similar to the well-known addition
of “a proton and two electrons” converting coenzyme NAD+ into NADH.29
(29) Stryer, L. Biochemistry, 4th ed.; W. H. Freeman: New York, 1995;
pp 449-450.
(30) The amount of 4b observed in the initial spectrum is somewhat
variable. It may depend on speed of mixing, time required to measure the
first spectrum, and other factors. However, the initial order of concentrations
is always [3b] > [4b] > [2b].
The formation of two isomers of 3d and 2d can reasonably
be explained by partial exchange of the Et and Cl groups at Al
to give diastereomers 3d′ and 2d′ as illustrated in Scheme 4.31
Even though preparative-scale experiments produced crystals
of 5d,18 no resonances clearly attributable to this complex are
observed in any of the spectra.
(31) Disproportionation to AlCl2 and AlEt2 complexes would be expected
to give even more complex spectra.

1042 Organometallics, Vol. 25, No. 4, 2006
Knijnenburg et al.
crystals deposited on cooling. One type looked like the free
ligand; the other (complex 9) was shown by X-ray diffraction
-
(see below) to contain [1‚AlCl2]+ cations and AlCl4 anions.
The NMR spectrum of a solution of these crystals shows clear
but broadened low-field resonances for the pyridine H4 (8.77
ppm) and H3,5 (8.06) protons, which agree well with shifts
calculated for the cation (Table 2).
Crystals of 9 contain two independent molecules as well as
two molecules of toluene. Figure 5 and Table 6 refer to one of
the cations; full details for all molecules in the structure can be
found in the Supporting Information.
Computational Study of Alkylation and Dimerization. The
observations described above demonstrate that the reaction of
1 with alkylaluminum compounds is quite complex. Addition
can occur at different positions of the ligand (at least C2, C4,
and imine C), and the products can undergo further reactions.
Cation-anion pairs can also form, and this may be relevant to
further reactivity. Therefore, we turned to theoretical methods
(DFT) for a better understanding of this complex system. In
particular, we wanted to understand the preferences for addition
at various positions of the ligand and the nature of the unusual
dimerization reaction. As a model system we used the reaction
between simplified ligand 1′ (bearing only hydrogens at the
imine carbon and nitrogen atoms) and Et2AlCl.
Figure 5. Thermal ellipsoid plot of the cation of 9 (50%
probability). Hydrogen atoms and isopropyl groups are omitted for
clarity.
Reaction of iBu3Al and iBu2AlH with 1. Spectra of the initial
reaction mixture of iBu3Al and 1 showed formation of only very
small amounts of alkylation products (ca. 6 mol % 4c and 1
mol % 4e). In addition, shifted ligand resonances were observed,
which might indicate formation of one or more coordination
complexes. On heating for longer periods, a precipitate was
formed, which X-ray diffraction studies indicated to be the dimer
5e of 4e. The same dimer 5e was prepared more conveniently
i
from Bu2AlH and 1. The spectrum from the latter reaction,
directly after mixing the reagents in CH2Cl2 and evaporation
of the solvent, showed mainly 4e (judging from the characteristic
triplets at 4.8 and 3.5 ppm) and no other recognizable
complexes; workup after warming produced pure 5e.
Reaction of AlCl3 with 1. Heating 1 and AlCl3 (1:1) in
toluene produced an orange solution, from which two types of
Alkyl Addition to 1′. Preliminary studies revealed that after
complexation of Et2AlCl direct alkyl transfer to the pyridine
C4 position has a very high barrier. This is hardly surprising,
given the physical distance between the Al atom and the pyridine
C4 position. Thus, more complicated mechanisms need to be
considered for the formation of 4′. One possibility might be
Scheme 5. Calculated Free Energies (kcal/mol; 273 K, 1 bar) of Minima and [Transition States], relative to 1′ + 2 Et2AlCl

Diimine Pyridine and Aluminum Alkyls
Organometallics, Vol. 25, No. 4, 2006 1043
Figure 7. Energy profile for formation of the second C-C bond
of 5′: relative energy (2) and S2 value (9).
Figure 6. Bond distances in the transition state for the first C-C
bond formation from two units of 4′. Distances are symmetry
averages over the two approaching monomers.
the second C-C bond. The energy profile for this rotation is
extremely flat. The final stages of the approach were followed
by constrained geometry optimizations; Figure 7 shows the
energy profile and the decrease in biradical character on giving
final product 5′. This is 30.5 kcal/mol below 9′ and 7.0 kcal/
mol below 4′ and is probably the most stable product accessible
from 9′.
The symmetry-forbidden character of the direct, synchronous
[3+3] dimerization path explains why we find a nonsynchronous
biradical path for this reaction. It does not, however, explain
the surprisingly low barrier we calculate for this nonsynchronous
path. One possible explanation is the more efficient delocal-
ization of negative charge over the two imine nitrogens in 5′
(in 4′, it is formally localized on the pyridine nitrogen). In
addition, 4′ is cross-conjugated, which is in general less
favorable than the linear conjugation over the NCCNCCN path
seen in 5′. However, we do not believe that the calculated barrier
of 13.4 kcal/mol for the biradical path is very accurate, since it
is notoriously difficult to calculate accurate relative energies
for open-shell and closed-shell species. Presumably, the most
one can say on the basis of our calculations is that C-C coupling
is “relatively easy”.34
Relative Stabilities of the Various Alkylation Products.
The above calculations on model compound 1′, though useful
in a qualitative sense, cannot be expected to accurately reproduce
the relative stabilities of the various isomers in the crowded
alkylation products derived from real ligand 1. For more
quantitative information, we optimized the complete structures
of all possible products derived from Me3Al and 1 (2a-4a and
6a-8a) as well as the most relevant reaction products of Et3Al
(2b-4b and dimer 5b). The resulting energies (Table 4) do not
include thermal corrections, which would be too expensive for
such large systems. Steric bulk is seen to have a major impact
on the relative stabilities. It makes the imine and C2 alkylation
products (2 and 3) comparable in energy and significantly higher
than those of alkylation at C4 and C3 (4 and 7), which are also
close to each other. The two products of N alkylation (6 and 8)
are both much higher in energy. The change from Me3Al to
Et3Al favors 3 over 2, as expected. These calculated relative
energies are consistent with the interpretation given earlier that
the observed final ratios of 2 and 3 correspond to the equilibrium
position. They are also compatible with eventual “disappear-
ance” on heating of all of 2b/3b via isomerization to the more
homolytic cleavage of an Al-C bond of 1‚AlR3, followed by
intra- or intermolecular transfer of the essentially free alkyl
radical, comparable to the mechanism proposed for alkyl
addition to R-diimine and Pyca systems.14 We did indeed
observe some formation of 1‚AlR2 radicals, but it is not clear
at present what their relation is to the observed diamagnetic
products; they might also represent a side reaction. In any case,
our observation of an ionic product 9 from 1 and AlCl3 prompted
us to consider the possibility of similar ionic species in our
alkylations, and we decided to introduce a second molecule of
Et2AlCl in the calculations (all energies are therefore calculated
relative to one ligand 1′ and 2 equiv of Et2AlCl). This led to
the reactions summarized in Scheme 5.
The loosely bound ion pair 9′ is calculated to be lower in
energy than the separated molecules 1′ and [Et2AlCl]2 even in
the gas phase.32 Ethyl migration from the Et3AlCl- anion to
various positions of the cation can now proceed smoothly:
calculated free energy barriers are 15.0 (C2), 12.9 (C3), 12.1
(C4), and 8.6 (imine) kcal/mol. Transfer to C2 is mildly
exothermic (12.0 kcal/mol); the other reactions are highly
exothermic (24.8, 23.5, and 22.6 kcal/mol, respectively) and
lead to products of comparable stability. However, steric
hindrance present in the real ligand might make transfer to the
imine position less favorable, as discussed later on.33 The
calculated barriers shown in the scheme are low enough to be
compatible with initial alkylation via an ionic mechanism. Of
course, this does not prove that the reaction of real ligand 1
with aluminum alkyls actually follows such an ionic path, but
it seems that the ionic path can at least be a realistic alternative
to the direct-transfer and free-radical paths.
Formation of Dimer 5′. At first sight, it seems reasonable
to expect dimerization of 4′ to follow a more or less synchronous
path for forming the two new C-C bonds. However, despite
extensive searches such a symmetric path could not be located.
Inspection of relevant orbitals shows that the symmetric
approach is symmetry-forbidden, making a nonsynchronous
biradical path likely. Indeed, we found that approach of two
monomers of 4′ easily leads to single C-C coupling with a
modest barrier of 13.4 kcal/mol (see Figure 6).
The resulting intermediate is essentially a biradical ( S2
)
1.01). After it has been formed, the two halves of the molecule
rotate with respect to each other around the newly formed C-C
bond to let the radical centers approach each other and so form
(34) 1,4-Dihydropyridines bearing electron-withdrawing substituents
easily undergo photochemical dimerization, even in the solid state. However,
the structures of the [2+2] dimers thus obtained are very different from
the [3+3] ones reported in the present work and refs 9 and 18. See e.g.:
Hilgeroth, A.; Baumeister, U.; Heinemann, F. W. Eur. J. Org. Chem. 1998,
1213, 3.
(32) Surprisingly, 9′ prefers a broken-symmetry electronic structure ( S2
) 0.28).
(33) The steric hindrance in 1 was earlier shown to have a large effect
on the relative stabilities of iron dialkyl ligand alkylation products.3

1044 Organometallics, Vol. 25, No. 4, 2006
Knijnenburg et al.
(s, 3H each, NdCCH3), 1.39, 1.35, 1.34, 1.32, 1.00, 0.94, 0.85,
0.85 (d, 3H each, 3JHH ) 6.8 Hz, CH(CH3)2), -0.43, -0.66 (s, 3H
each, AlCH3).
stable 4b and its subsequent dimerization to 5b. However,
neither the barriers shown in Scheme 5 nor the product stabilities
listed in Table 4 explain why we have never observed any 7,
which is calculated to be as low in energy as 4 and equally
easy to form. It might be that alkylation follows a path
completely different from the ionic one studied here and that
this alternative path has a large kinetic bias against attack at
C3. Alternatively, it could be that 7 is even more reactive than
4 and rapidly dimerizes or polymerizes to nonobservable
products.
1
3
4a. H NMR (400 MHz, C7D8): δ 5.03 (d, 2H, JHH ) 3.8 Hz,
Py H3,5) 3.52 (m, 1H, Py H4).
1
3
4f. H NMR (400 MHz, C7D8): δ 4.94 (t, 2H, JHH ) 3.8 Hz,
3
Py H3,5) 3.53 (t, 2H, JHH ) 3.8 Hz, Py H4).
Synthesis of 3b. To 0.296 g of 1 (0.615 mmol) in 3 mL of hexane
was added 0.74 mL of 1 M Et3Al in hexanes, which resulted in a
color change from yellow to green. The suspension was stirred
overnight at room temperature, after which it was cooled to -30
°C for one week. It was filtered, and the orange-yellow residue
was washed with 5 mL of ice-cold hexane and dried in vacuo. 1H
NMR showed that this was pure 3b, of which all peaks could be
assigned; yield 0.044 g (12%).
Conclusions
The reaction of 1 with aluminum alkyls is surprisingly
complex and features alkylation at the C2, C4, and imine
carbons. These additions are in part reversible. The C4 alkylation
product can dimerize to form the tricyclic ligand skeleton
observed earlier for a chromium derivative.9 Calculations
indicate that for Al complexes 4 the dimerization follows a
nonsynchronous biradical pathway; there seems to be no reason
the chromium system would not follow the same route. The
newly formed six-membered ring has all substituents in a well-
defined stereoselective orientation, which suggests this reaction
might have some potential in organic synthesis.
1
3b. H NMR (400 MHz, C6D6): δ 7.11-7.07 (m, 6H, Ar H),
3
3
6.29 (dd, 1H, JHH ) 5.8 and 8.8 Hz, Py H4), 5.66 (dd, 1H, JHH
4
3
) 5.8 Hz, JHH ) 0.7 Hz, Py H5), 5.12 (dd, 1H, JHH ) 8.8 Hz,
4JHH ) 0.7 Hz, Py H3), 3.21, 3.09, 3.07, 3.01 (sept, 1H each, 3JHH
) 6.8 Hz, CH(CH3)2), 2.45-2.36, 1.31-1.23 (m, 1H each, Py-
C2-CH2CH3), 1.74, 1.55 (s, 3H each, NdCCH3), 1.41 (2×), 1.39,
1.37, 1.11, 1.00, 0.99, 0.97 (d, 3H each, 3JHH ) 6.8 Hz, CH(CH3)2),
3
1.19, 0.73 (t, 3H each, JHH ) 8.0 Hz, AlCH2CH3), 0.90 (t, 3H,
3JHH ) 8.0 Hz, Py-C2-CH2CH3), 0.23-0.09, 0.10-0.02 (m, 2H
each, AlCH2CH3). 13C NMR (50 MHz, C6D6): δ 183.1, 173.8 (Nd
CCH3) 146.2, 143.2, 143.0, 141.1, 140.5, 140.0 (2×) (Py C6, Ar
Ci, Ar Co), 126.7, 126.5, 124.6, 124.5, 124.4, 123.8 (Ar Cm, Ar
Cp), 126.2 (PyC4), 114.2 (Py C5), 101.7 (Py C3), 67.5 (Py C2),
28.6, 28.3, 27.7 (2×), 27.5 (CH(CH3)2), Py-C2-CH2CH3), 26.0,
25.4, 25.0 (2×), 24.9, 24.1, 23.7, 23.1 (CH(CH3)2), 17.9, 16.8 (Nd
CCH3), 10.8, 8.7 (AlCH2CH3), 6.2 (Py-C2-CH2CH3), 1.4, -0.5
(AlCH2CH3). Anal. Calcd (%) for C39H58N3Al: C 78.61, H 9.81,
N 7.05. Found: C 78.74, H 9.05, N 7.87.
The mechanism(s) of ligand alkylation remain unclear. Direct
alkyl transfer may play a role for imine and C2 alkylation but
is unlikely for C4 alkylation. Transfer of free alkyl radicals is
•
another possibility, although for the extremely reactive CH3
and H• radicals one would expect many more side reactions.
The calculations reported here show that an ionic mechanism
is also a realistic possibility. Clearly, more work, both experi-
mental and theoretical, is required before any possibility can
be ruled out.
Heating of 3b. Spectra were recorded after heating a sealed C7D8
solution of pure 3b at 60 °C for 22 h, at 85 °C for another 24 h, at
110 °C for another 24 h, at 110 °C for another 72 h, at 145 °C for
another 48 h, and at 145 °C for a final 48 h. Apart from 3b (vide
supra) the following signals could be assigned:
Experimental Section
General Procedures. All manipulations were carried out under
an atmosphere of argon using standard Schlenk techniques or in a
conventional nitrogen-filled glovebox. Solvents were refluxed over
an appropriate drying agent and distilled prior to use. NMR spectra
were recorded on Varian and Bruker spectrometers at ambient
temperature. For NMR spectra of mixtures, only clearly separated
signals are reported here.
1
3
2b. H NMR (200 MHz, C6D6): δ 4.38 (sept, 1H, JHH ) 6.8
Hz, CH(CH3)2) (assignments verified by comparison with an
authentic sample of pure 2b11).
4b. 1H NMR (400 MHz, C6D6): δ 5.03 (d, 2H, 3JHH ) 3.8 Hz,
Py H3,5), 3.70-3.65 (m, 1H, Py H4), 1.70 (s, 6H, NdCCH3).
Crystallization of 3b. In a separate experiment 0.26 g of 1 (0.55
mmol) was suspended in 10 mL of toluene at -50 °C, and 0.6 mL
of 1 M Et3Al in hexanes (0.60 mmol; 1.1 equiv) was added, which
resulted in a color change from yellow to light orange. Warming
up to room temperature resulted in a darkening of this color. On
slow evaporation crystals were grown, which were suitable for
X-ray diffraction.
1 + Me3Al. Light yellow 1 (218 mg, 453 µmol) was suspended
in 5 mL of toluene. Me3Al was added (65.3 µL, 679 µmol; 1.5
equiv), which resulted in the formation of a yellow-brown solution.
The solution was evaporated to dryness, and a 1H NMR spectrum
1
was recorded in C7D8, after which the NMR tube was sealed. H
NMR spectra at room temperature were recorded again after heating
at 60 °C for 22 h, after heating at 85 °C for 72 h, after heating at
120 °C for 72 h, and after heating at 120 °C for a final 96 h. The
1 + Et2AlCl. To 0.140 g of 1 (291 µmol) in 5 mL of toluene
was added 0.34 mL of a ca. 1 M solution of Et2AlCl in hexanes
(336 µmol, 1.16 equiv), which resulted in a color change from
yellow to yellow-green. The solution was evaporated to dryness,
redissolved in C7D8, and sealed in an NMR tube. A 1H NMR
spectrum was recorded immediately, then after heating 22 h at 60
°C (solution has become yellow-brown) and after 24 h at 85 °C
(solution has become brown). Spectra recorded after further heating
were broad and featureless. The following signals could be assigned:
1
following signals could be assigned (part of the H NMR data is
from a separate experiment and recorded in C6D6, where COSY
and NOESY data allowed a more complete assignment; assignments
for 2a were verified by comparison with an authentic sample of
pure 2a11):
1
2a. H NMR (400 MHz, C6D6): δ 3.97, 2.94 (sept, 2H each,
3JHH ) 6.8 Hz, CH(CH3)2), 1.67 (s, 3H, NdCCH3), 1.46 (s, 6H,
3
NC(CH3)2), 1.39, 1.38, 1.32, 0.97 (d, 6H each, JHH ) 6.8 Hz,
1
3
CH(CH3)2), -0.52 (s, 6H, Al(CH3)2). 13C NMR (50 MHz, C6D6):
δ 174.0 (NdCCH3) 163.1 (Py C6), 155.6 (Py C2) 62.0 (NC(CH3)2),
-4.0 (Al(CH3)2).
2d. H NMR (300 MHz, C7D8): δ 4.52 (sept, 1H, JHH ) 6.8
Hz, CH(CH3)2).
3d. 1H NMR (300 MHz, C7D8): δ 6.12 (dd, 1H, 3JHH ) 5.8 and
8.8 Hz, Py H4), 5.56 (d, 1H, 3JHH ) 6.0 Hz, Py H5), 5.16 (d, 1H,
3JHH ) 9.0, Py H3), 1.72, 1.56 (s, 3H each, NdCCH3) 0.74 (t, 3H,
3JHH ) 8.0 Hz, AlCH2CH3), 0.10 (quart, 2H, 3JHH ) 8.0 Hz, AlCH2-
CH3).
3a. 1H NMR (400 MHz, C6D6): δ 6.32 (dd, 1H, 3JHH ) 6.0 and
3
4
8.8 Hz, Py H4), 5.74 (dd, 1H, JHH ) 6.0 Hz, JHH ) 0.6 Hz, Py
H5), 5.16 (dd, 1H, 3JHH ) 8.8, 4JHH ) 0.6 Hz, Py H3), 3.16, 3.07,
3.04, 2.89 (sept, 1H each, JHH ) 6.8 Hz, CH(CH3)2), 1.77, 1.52
3

Diimine Pyridine and Aluminum Alkyls
Organometallics, Vol. 25, No. 4, 2006 1045
4d. 1H NMR (300 MHz, C7D8): δ 5.10 (d, 2H, 3JHH ) 3.8 Hz,
Py H3,5).
did not locate or calculate hydrogen positions for this solvent
molecule. As a consequence, the thermal displacement parameters
of the toluene molecules are rather large, and no conclusions may
be derived based on these molecules. Nevertheless, we are
convinced that the connectivity and geometry of the Al complex
are correct and accurate enough.
3b. The unit cell contains two independent molecules with only
minor geometrical differences.
5e. The asymmetric unit consists of only half a molecule, since
the dimer is positioned around an inversion center.
1 + iBu3Al. To 0.214 g of 1 (444 mmol) in 5 mL of toluene
i
was added 0.367 g of 1 M Bu3Al (1.19 equiv), which resulted in
a color change from yellow to brown. The solution was evaporated
to dryness, redissolved in C7D8, and sealed in an NMR tube. A 1H
NMR spectrum was recorded immediately. The solution changed
color on heating (first to red and then to purple-red), but spectra
recorded after heating showed mostly broad lines and were not
readily interpretable; purple crystals formed from the solution after
heating proved to be 5e (X-ray). In the initial spectrum, the
following signals could be assigned:
9. This structure, like that of 3b, contains two crystallographically
independent molecules. In addition, one molecule of toluene per 9
was found to be present in the lattice. The crystal structure
determination of 9 was hampered by the poor crystal quality. Some
of the chlorine atoms of the AlCl4 moieties show large molecular
displacement parameters, even though the data were collected at
-65 °C. It proved to be impossible to describe this disorder; all
attempts to do so led to unstable refinements. The same holds to a
lesser extent for one of the isopropyl CH3 carbons and one of the
ipso carbons of a diisopropylphenyl group. The four highest peaks
in the difference Fourier map (1.76 to 1.14 e/Å3) are all close to
the above-mentioned chlorine atoms; the fifth peak is only 0.40
e/Å3.
The crystal data and a summary of the data collection and
structure refinement for all four structures are given in Table 5.
The structure and part of the atomic numbering is shown in Figures
1 (2a), 2 (3b), 3 (5e), and 5 (9).37 A selection of bond distances
and angles is given in Table 6. Complete crystallographic data (CIF)
are provided as Supporting Information. A drawing of 4e and details
of the low-quality structure determination for this compound are
given in the Supporting Information.
Calculations. All geometry optimizations were carried out with
the Turbomole program38,39 coupled to the PQS Baker optimizer.40,41
Geometries were fully optimized as minima or transition states at
the unrestricted (for 1′) or restricted (for 1) bp8642,43/RIDFT44 level
using the Turbomole SV(P) basis set45 on all atoms. For model
system 1′ bearing only hydrogens at the imine carbon and nitrogen,
all stationary points were characterized by vibrational analyses
(numerical frequencies); ZPE and thermal (enthalpy and entropy)
corrections (1 bar, 273 K) from these analyses are included. For
these systems, all energies mentioned in text and tables are free
energies. For the full systems derived from 1, Gaussian9824 was
used to calculate chemical shifts using the GIAO formalism,23 the
B3LYP functional,25 and the 6-311G** basis.26 Thermal corrections
were not feasible for these larger systems.
1
3
4c. H NMR (400 MHz, C6D6): δ 5.09 (d, 2H, JHH ) 4.1 Hz,
Py H3,5), 3.81-3.73 (m, 1H, Py H4).
1
3
4e. H NMR (400 MHz, C6D6): δ 4.87 (t, 2H, JHH ) 3.8 Hz,
3
PyH3,5), 3.54 (t, 2H, JHH ) 3.8 Hz, Py H4).
1 + iBu2AlH. To 0.505 g of 1 (1.05 mmol) in 10 mL of hexane
i
was added 1.15 mL of 1 M Bu2AlH (1.10 equiv), which resulted
in a color change from yellow to green. Crystals formed from this
solution on cooling proved to be green 4e (X-ray; for details see
Supporting Information). The solution was heated at 110 °C for 2
h, which resulted in a color change to purple. After cooling
overnight, 70 mg (10.7%) of crystals suitable for X-ray diffraction
had formed; they were identified as 5e, identical to the crystals
formed from iBu3Al, by comparison of the observed and calculated
powder diffraction patterns. Unfortunately, the poor solubility of
the product did not allow recording of NMR spectra. The clearest
1
i
initial H NMR spectrum was obtained by mixing 1 and Bu2AlH
in CH2Cl2, removing the solvent and then recording the spectrum
in toluene-d8; data in Table 3 are based on such a spectrum. Anal.
Calcd (%) for C82H124N6Al2 (5e): C 78.93, H 10.02, N 6.73, Al
4.32. Found: C 78.64, H 9.90, N 6.84, Al 4.22.
1 + AlCl3. To 224 mg of 1 (yellow, 0.47 mmol) were added 62
mg of AlCl3 (yellow; 0.465 mmol; 1.0 equiv) and 15 mL of toluene,
which resulted in a light orange suspension. This was heated to 87
°C and then slowly cooled, upon which two types of crystals
formed: small yellow needles that looked liked free ligand, as well
as larger shiny yellow-orange cubes that were structurally character-
ized as 9.
9. 1H NMR (200 MHz, C6D6): δ 8.63 (t, 1H, 3JHH ) 7.7 Hz, Py
H4), 7.92 (d, 2H, 3JHH ) 7.7 Hz, Py H3,5), 7.17-7.02 (m, 6H, Ar
3
H), 2.88 (sept, 4H, JHH ) 6.6 Hz, CH(CH3)2), 2.03 (s, 6H, Nd
CCH3), 1.33, 1.03 (d, 12H each, 3JHH ) 6.6 Hz, CH(CH3)2). Anal.
Calcd (%) for C33H43N3Cl6Al2‚C7H8: C 57.16, H 6.12, N 5.00, Cl
25.31. Found: C 56.96, H 6.13, N 5.09, Cl 25.53.
Structure Determination of 2a, 3b, 5e, and 9. The molecular
structures of 2a, 3b, 5e, and 9 were determined by single-crystal
X-ray diffraction. For measurement of 3b, the crystal was mounted
in a glass capillary. The structures were solved by the PATTY
option35 of the DIRDIF program system.36 Notes on individual
structures follow.
2a. The structure contains one cocrystallized molecule of toluene
in a general position and one molecule of toluene located over an
inversion center. Although the geometry of the Al complex is quite
acceptable, the geometries of the toluene solvent molecules certainly
are not. The toluene molecule in a general position had to be
constrained rather strictly in order to keep the geometry acceptable.
The other toluene molecule has a symmetry-imposed disorder; we
Supporting Information Available: Complete crystallographic
data (CIF format) for complexes 2a, 3b, 5e, 9, and 4e; H NMR
1
(37) Spek, A. L. PLATON. A Multipurpose Crystallographic Tool;
Utrecht University: Utrecht, The Netherlands, 2003.
(38) Ahlrichs, R.; Ba¨r, M.; Baron, H.-P.; Bauernschmitt, R.; Bo¨cker, S.;
Ehrig, M.; Eichkorn, K.; Elliott, S.; Furche, F.; Haase, F.; Ha¨ser, M.; Ha¨ttig,
C.; Horn, H.; Huber, C.; Huniar, U.; Kattannek, M.; Ko¨hn, A.; Ko¨lmel, C.;
Kollwitz, M.; May, K.; Ochsenfeld, C.; O¨ hm, H.; Scha¨fer, A.; Schneider,
U.; Treutler, O.; Tsereteli, K.; Unterreiner, B.; Von Arnim, M.; Weigend,
F.; Weis, P.; Weiss, H. Turbomole Version 5; Theoretical Chemistry Group,
University of Karlsruhe: Germany, January 2002.
(39) Treutler, O.; Ahlrichs, R. J. Chem. Phys. 1995, 102, 346.
(40) PQS version 2.4; Parallel Quantum Solutions: Fayetteville, AR,
2001 (the Baker optimizer is available separately from PQS upon request).
(41) Baker, J. J. Comput. Chem. 1986, 7, 385.
(35) Beurskens, P. T.; Beurskens, G.; Strumpel, M.; Nordman, C. E. In
Patterson and Pattersons; Glusker, J. P., Patterson, B. K., Rossi, M., Eds.;
Clarendon Press: Oxford, 1987; p 356.
(36) Beurskens, P. T.; Beurskens, G.; Bosman, W. P.; de Gelder, R.;
Garcia-Granda, S.; Gould, R. O.; Israel, R.; Smits, J. M. M. DIRDIF-96. A
computer program system for crystal structure determination by Patterson
methods and direct methods applied to difference structure factors;
Crystallography Laboratory, University of Nijmegen: The Netherlands,
1996.
(42) Becke, A. D. Phys. ReV. A 1988, 38, 3089.
(43) Perdew, J. P. Phys. ReV. B 1986, 33, 8822.
(44) Eichkorn, K.; Weigend, F.; Treutler, O.; Ahlrichs, R. Theor. Chem.
Acc. 1997, 97, 119.
(45) Schaefer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571.
(46) Sheldrick, G. M. SADABS, Program for Empirical Absorption
Correction; University of Go¨ttingen: Germany, 1996.
(47) Sheldrick, G. M. SHELXL-97, Program for the refinement of crystal
structures; University of Go¨ttingen: Germany, 1997.

1046 Organometallics, Vol. 25, No. 4, 2006
Knijnenburg et al.
spectra for the reaction between 1 and AlMe3 (Figure S1), Et2AlCl
(Figure S3), and the heating of 3b (Figure S2); molecular structure
(Figure S4) and crystallographic details (Table S1) for the low-
quality X-ray structure determination of 4e; total energies and
thermal corrections for geometries of minima and transition states
derived from 1′ as calculated by DFT (Table S2); selected bond
distances for these species (Figures S5-S7); calculated total
energies for 2a-4a, 6a-8a, and 2b-4b at RIDFT-bp86/SV(P) and
B3LYP/6-311G** levels (Table S3), and calculated NMR chemical
shifts for various alkylation products and coordination complexes
of 1. This material is available free of charge via the Internet at
http://pubs.acs.org.
OM050936M