Structures and reactions of monomeric and dimeric lithium diazapentadienyl complexes with electrophiles: Synthesis of α-C,C′-dialkyl-β- diimines, and dissolution-reversible synthesis of an α-alkoxylithium- β-diimine

Article abstract of DOI:10.1039/b212079h

Direct acid-catalysed condensation of substituted anilines with acetylacetone was found to give convenient access to β-enamineimines [2-Pr¹-C₆H₄NCMeCHCMeNH-2-Pr¹-C ₆H₄] and [2-MeO-C₆

Full text of DOI:10.1039/b212079h

View Article Online / Journal Homepage / Table of Contents for this issue
Structures and reactions of monomeric and dimeric lithium
diazapentadienyl complexes with electrophiles: synthesis of
ꢀ-C,CЈ-dialkyl-ꢁ-diimines, and dissolution-reversible synthesis
of an ꢀ-alkoxylithium-ꢁ-diimine
David T. Carey, Elaine K. Cope-Eatough, Eva Vilaplana-Mafé, Francis S. Mair,*
Robin G. Pritchard, John E. Warren and Rebecca J. Woods*
Department of Chemistry, UMIST, P.O. Box 88, Manchester, UK M60 1QD.
E-mail: frank.mair@umist.ac.uk
Received 6th December 2002, Accepted 31st January 2003
First published as an Advance Article on the web 20th February 2003
Direct acid-catalysed condensation of substituted anilines with acetylacetone was found to give convenient access to
β-enamineimines [2-Pri-C6H4NCMeCHCMeNH-2-Pri-C6H4] and [2-MeO-C6H4NCMeCHCMeNH-2-MeO-C6H4],
whereas TiCl4-mediated condensation was required to produce [2,6-Pri2-C6H3NHC(CF3)CHC(CF3)N-2,6-Pri2-
C6H3], which was crystallographically characterized. All are conveniently metallated using BunLi. The structures
of monomer [2-Pri-C6H4NCMeCHCMeN-2-Pri-C6H4ؒLi(thf )2], and dimer [{ 2-MeO-C6H4NCMeCHCMeN-2-
MeO-C6H4ؒLi}2] are reported. The structure of the dimeric product of aldol addition of adamantan-2-one to
[2-Pri-C6H4NCMeCHCMeN-2-Pri-C6H4ؒLi.], the lithium scorpionate [{(C10H14OLi)CH(CMeN-2-Pri-C6H4)2}2],
is also reported. It undergoes retro-aldol dissociation upon dissolution in non-co-ordinating solvents. The
efficient synthesis of α-C,CЈ dialkylated ‘true’ β-diimines by repeat lithiation/alkylation of di- and mono-ortho-
isopropylanilino diketiminates is also reported. The differing reactivity of the monomers and dimer with
electrophiles, and its relation to the structures of the intermediates, are discussed.
to rationalise these on the basis of structural characterizations
of further examples of 2 while varying steric, electronic and
Introduction
Since the first diketiminate complexes appeared in 1968,1 there
functional characteristics. This we have achieved by employing
has been sporadic use of this ligand class, otherwise known as
trifluoromethyl substitution at position R1, thus introducing a
diazapentadienyl, vinamidine, β-iminatoaminate etc., in many
fluorinated version (1f ) of the most widely used variant 1a, and
different scenarios.2 However, the introduction of diaryl
by leaving R2 unsubstituted while employing isopropyl and
diketiminates possessing extreme ortho bulk in 1997,3 followed
methoxy substituents in position R3 to generate reactants 1c
rapidly by the first demonstrations of their use as N–N biden-
and 1g and their lithiated complexes 2c and 2g. We also extend
tate, monoanionic ligands of remarkable steric control in 1998,4
prompted a relative explosion of effort in the field: The ligand
discussion of some previously communicated results on use
of adamantanone as the electrophile.10 Finally, we report
has enabled significant milestones to be reached across the
the efficient α-C,CЈ-dialkylation of 2a and 2c to yield ‘true’ β-
Periodic Table.5 Different chemistry can be effected with subtle
diimines, devoid of α-C acidity or imino-azaenol tautomerism.
changes of steric demand in the R1, R2 and R3 positions,6 but
so far the ligands have been adopted overwhelmingly as inert
manifolds on which to hang unusual chemistry.2,5 Only recently
has effort been directed at manipulations at position R4.
Chlorination of 2a at R4 has recently been demonstrated7 to
give 1d, and a copper variant of the less bulky 2e has also been
Experimental
All manipulations requiring dry conditions were carried out
under a protective argon blanket, either in a double manifold
argon/vacuum line or argon-filled recirculating glovebox.
Argon was dried over phosphorus pentoxide supported on
vermiculite. Toluene, n-hexane and thf were used freshly
distilled under argon from sodium–benzophenone; acetonitrile
and dichloromethane from calcium hydride. 2,6-Diisopropyl-
aniline and 2-isopropylaniline were distilled from potassium
hydroxide prior to use. CDCl3 and hexamethylphosphoramide
(CAUTION: suspected carcinogen) were stored over 4 Å
molecular sieves. The hexane solution of BunLi was used as
received and standardised using N-benzylbenzamide.11 For
cryoscopy, spectrograde benzene was dried with freshly activ-
ated molecular sieves 3Å and standardised using benzil to an
experimental cryoscopic constant of 5.04. Meaurements were
made under argon in an air-jacketed Schlenk tube fitted with
a Beckman thermometer and placed in a cooling bath held
at 0 ЊC.
reported, though its synthesis was somewhat different.8
A
photochemical rearrangement of a PtIVMe3 complex of
deprotonated 1a led to methylation at R4.9 Continuing this
theme of investigating the reactions of species 2, rather than
their employment as spectator ligands, we report here our
studies of the reactions of 2 with electrophiles, and attempt
ϩ
Enamine-imine 1a was synthesised by a literature method.4
All other reagents were obtained from standard commercial
vendors and used as received.
Melting points were determined in sealed glass capillaries
under argon. Elemental analyses were performed by the micro-
analytical group in the Chemistry Department at UMIST.
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1083

View Article Online
1H NMR solution spectra were recorded on Bruker DPX 200
MHz, 300 MHz and 400 MHz NMR spectrometers. 13C{1H}
NMR solution spectra were recorded on a Bruker DPX 300
or 400 MHz spectrometer operating at 75 or 100 MHz
respectively. Chemical shifts are given in ppm and referenced
to residual H solvent shifts or 13C-NMR solvent shifts.
Assignments were made with the aid of DEPT and HMQC
experiments. Solid-state NMR spectra were acquired on a
Bruker AMX400WB spectrometer system operating at 100.6
MHz interfaced to Bruker temperature control and spinning
units. All experiments were performed using Bruker double
resonance MAS probes fitted with 4 mm spinning modules.
Samples were contained in zirconia MAS rotors fitted with
Kel-F caps and were spun at MAS frequencies of between 3 and
11 kHz at ambient temperature.
73.5; H, 7.1; N, 9.0. Found: C, 73.7; H, 7.2; N, 9.0%. IR: 1630
cmϪ1 (s, ν(C᎐N)), 1555 cmϪ1 (s, ν(C᎐C, aromatic)).
᎐
᎐
1,1,1,5,5,5-Hexafluoro-2-(2,6-diisopropyl)phenylamino-4-
(2,6-diisopropyl)phenyliminopent-2-ene (1f)
A solution of 2,6-diisopropylaniline (28.3 ml, 150 mmol) in
dry hexane (ca. 100 ml) was prepared under argon. Titanium
tetrachloride diluted in dry hexane (16 ml, 3.42 M, 54.7 mmol)
was added dropwise to the stirring solution of 2,6-diisopropyl-
aniline at 0 ЊC. A dense yellow-brown precipitate formed
immediately. The mixture was left to stir for 12 h under argon.
1,1,1,5,5,5-Hexafluoro-2,4-pentanedione (3.5 ml, 25 mmol) was
then added dropwise to the mixture which turned orange-
brown. After heating under reflux for 2 h, the mixture became
yellow-brown. Titanium dioxide was removed by filtration and
2,6-diisopropylaniline hydrochloride was removed by carrying
out a water/hexane extraction. The bright yellow-orange hexane
phase was reduced in volume and triturated with methanol to
yield a bright yellow crystalline solid, 1f, that was filtered off
and vacuum dried. Large yellow hexagonal-prism-shaped
crystals were grown from the filtrate at Ϫ25 ЊC; mp: 154–156 ЊC.
Yield: 3.42 g, 26%. 1H NMR (200 MHz; CDCl3): δ 1.15 (12H, d,
Infrared spectra were recorded on a Nicolet Nexus-FTIR/
Raman spectrometer using NaCl plates and Nujol mulls.
Raman spectra were recorded from powdered samples in
Lindemann capillaries using the same instrument.
2-(2-Isopropyl)phenylamino-4-(2-isopropyl)phenyliminopent-2-
ene (1c)
3
3JHH = 6.8 Hz, MeCHMe); 1.27 (12H, d, JHH = 6.8 Hz,
In a modification of a literature procedure,12 a solution of
2,4-pentanedione (9.48 ml, 92.03 mmol), 2-isopropylaniline
(25.50 ml, 184.07 mmol) and toluene (ca. 120 ml) was prepared.
A catalytic amount of para-toluenesulfonic acid was added and
the resultant mixture was heated under reflux for 7 h. The water
(≈ 3.3 ml) produced in the reaction was collected in a Dean–
Stark apparatus as a toluene azeotrope. The majority of the
toluene (≈ 100 ml) was then removed from the reaction by
distillation into the Dean–Stark arm and the remaining mixture
was triturated with methanol and filtered to yield a cream
crystalline solid, 1c. Cooling the toluene/methanol filtrate to
Ϫ25 ЊC yielded further cream crystals; mp: 105–108 ЊC. Com-
bined yield: 17.65 g, 57%. 1H NMR (200 MHz; CDCl3): δ 1.22
3
MeCHMe); 2.99 (4H, septet, JHH = 7.0 Hz, MeCHMe); 5.84
(1H, s, NC(CF3)CHC(CF3)N); 7.13–7.31 (6H, non-first-order
m, aromatic protons); 11.20 (1H, br s, NH ). 13C NMR (75
MHz; CDCl3): δ 23.1 (MeCHMe); 25.4 (MeCHMe); 28.9
1
(MeCHMe); 87.6 (NC(CF3)CHC(CF3)N); 118.2 (q, JCF
=
286 Hz, NC(CF3)CHC(CF3)N); 123.6 and 127.0 (aromatic
carbons); 138.0 (C–N, aromatic C); 142.0 (C(CH(Me)2)); 150.8
2
(q, JCF = 30.2 Hz, NC(CF3)CHC(CF3)N). 19F NMR (188
MHz, CDCl3): δ Ϫ65.9 (NC(CF3)CHC(CF3)N). Elemental
analysis, Calcd. for C29H36F6N2: C, 66.1; H, 6.9; N, 5.3; F,
21.6. Found: C, 66.2; H, 6.6; N, 5.4; F, 21.3%. IR: 1642 cmϪ1
(s, ν(C᎐N)).
᎐
3
(12H, d, JHH = 7.0 Hz, MeCHMe); 1.95 (6H, s, NCMeCHC-
MeN); 3.25 (2H, septet, 3JHH = 7.0 Hz, MeCHMe); 4.96 (1H, s,
NCMeCHCMeN); 6.92–7.38 (8H, non-first-order m, aromatic
protons); 12.53 (1H, br s, NH ). 13C NMR (75 MHz; CDCl3):
δ 21.2 (NCMeCHCMeN); 23.6 (MeCHMe); 28.5 (MeCHMe);
96.6(NCMeCHCMeN); 124.3, 124.7, 125.9 and 126.2(aromatic
CH); 142.0 (C(CH(Me)2), aromatic C); 143.7 (C–NH, aromatic
Lithiation of 1c: 2cؒ2thf
To 1c (1.46g, 4.4 mmol) in thf (4 mL) and n-hexane (5 mL) was
added BunLi in hexanes (1.75 ml of a 2.5 M solution, 4.4 mmol)
with ice cooling. The solution was evacuated to ca. 4 ml volume
and rediluted with hexane (2 ml). Overnight refrigeration
yielded a crop of colourless rhomboids, which were isolated by
vacuum filtration, upon which solvent was lost from the crystal
lattice. Yield: 1.68 g. mp: 74–76 ЊC. 1H NMR (300 MHz; C6D6):
δ 1.35 (20H, overlapping m, MeCHMe, O(CH2CH2)2); 2.09
(6H, s, NCMeCHCMeN); 3.36 (8H, t, O(CH2CH2)2); 3.55 (2H,
C); 160.4 (C᎐N). Elemental analysis, Calcd. for C H N : C,
᎐
23 30
2
82.6; H, 9.0; N, 8.4. Found: C, 82.6; H, 9.1; N, 8.4%. IR: 1628
cmϪ1 (s, ν(C᎐N)), 1557 cmϪ1 (s, ν(C᎐C, aromatic)).
᎐
᎐
3
2-(2-Methoxy)phenylamino-4-(2-methoxy)phenylimino-pent-2-
ene (1g)
septet, JHH = 6.8 Hz, MeCHMe); 5.05 (1H, s, NCMeCHC-
3
4
MeN); 7.00 (2H, dd, JHH = 7.5 Hz, JHH = 1.5 Hz, aryl 2-H);
7.12 (2H, td, 3JHH = 7.5 Hz, 4JHH = 1.5 Hz, aryl 3-H); 7.21 (2H,
A solution of 2,4-pentanedione (25 mL, 0.24 mol), 2-methoxy-
aniline (60 mL, 0.53 mol) and toluene (150 ml) was prepared.
A catalytic amount of para-toluenesulfonic acid was added
and the resultant mixture was heated under reflux for 4 h with a
Dean–Stark trap, protected from the atmosphere by a drying
tube. The water (8 ml) produced in the reaction was continu-
ously removed as a toluene azeotrope. The majority of the
toluene (≈ 130 ml) was then removed from the reaction by
distillation into the Dean–Stark arm and the remaining brown
oil crystallized on cooling to room temparature. The semi-solid
mass was triturated with a small amount of methanol, filtered,
washed with cold methanol, and recrystallized from methanol/
hexane (4 : 1) to furnish 12.99 g (87.3%) of 1g as pale
td, 3JHH = 7.5 Hz, 4JHH = 1.5 Hz, aryl 4-H); 7.37 (2H, dd, 3JHH
=
7.5 Hz, JHH = 1.5 Hz, aryl 4-H). 13C NMR (75 MHz,
C6D6): δ 23.6, MeCHMe); 24.1 (NCMeCHCMeN); 25.8
(O(CH2CH2)2); 28.9 (MeCHMe); 68.2 (O(CH2CH2)2); 94.0
(NCMeCHCMeN); 122.8, 124.7, 125.1, 126.0 (aromatic
CH); 141.6 (C(CH(Me)2), aromatic C); 153.4 (aromatic C–N);
4
163.3 (C᎐N). Elemental analyses were variable (solvent
᎐
loss).
Lithiation of 1g: (2g)2
To a suspension of 1g (1.59 g, 5.12 mmol) in n-hexane (10 ml) at
0 ЊC was added BunLi (2.1 ml of a 2.5M solution in toluene)
with stirring. The mixture was heated to boiling, whereupon a
slightly turbid orange solution formed. Stirring was ceased, and
the solution deposited yellow crystals on slow cooling to room
temperature. These were isolated by filtration to yield 0.984 g
1
yellow plates; mp: 129–130 ЊC. H NMR (300 MHz; CDCl3):
δ 2.00 (6H, s, NCMeCHCMeN); 3.89 (6H, s, MeO); 4.94
(1H, s, NCMeCHCMeN); 6.88–7.28 (8H, non-first-order
m, aromatic protons); 12.75 (1H, br s, NH ). 13C NMR (75
MHz; CDCl3): δ 21.5 (NCMeCHCMeN); 56.1 (MeO); 98.3
(NCMeCHCMeN); 111.9, 120.9, 123.4 and 124.0 (aromatic
CH); 135.8 (COMe, aromatic C); 151.7 (C–NH, aromatic C);
1
(60.7%) of 2g; mp: 177–179 ЊC. H NMR (300 MHz, C6D6):
δ 2.10 (6H, s, NCMeCHCMeN); 3.11 (6H, s, OMe); 5.00 (1H, s,
NCMeCHCMeN); 6.06 (d), 6.85 (t), 6.95 (t), 7.05 (d), all 2H,
3JHH = 7.5 Hz, aromatic CH. 13C NMR (75 MHz, C6D6): δ 24.0
160.4 (C᎐N). Elemental analysis, Calcd. for C H N O : C,
᎐
19 22
2
2
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1084

View Article Online
(NCMeCHCMeN); 55.1 (OMe); 102.6 (NCMeCHCMeN);
110.6, 121.3, 121.7, 125.1 (aromatic CH); 143.4 (MeOC); 152.3
first-order m, aromatic protons). 13C NMR (75 MHz; CDCl3):
δ 8.6 (NCMeC(CH2Me)2CMeN); 17.7 (NCMeC(Et)2CMeN);
23.5 (MeCHMe); 23.8 (NCMeC(CH2Me)2CMeN); 28.3
(MeCHMe); 61.6 (NCMeC(Et)2CMeN); 118.6, 123.9, 126.0
and 126.6 (aromatic CH); 138.0 (C(CH(Me)2)); 149.4 (C–N);
(aromatic C–N); 164.2 (C᎐N). Elemental analysis, Calcd. for
᎐
C19H21N2O2Li: C, 72.1; H, 6.7; N, 8.8. Found: C, 76.2; H, 6.6;
N, 5.4%.
172.7 (C᎐N). Elemental analysis, Calcd. for C H N : C, 83.0;
᎐
27 38
2
H, 9.8; N, 7.2. Found: C, 83.3; H, 9.7; N, 7.2%. IR: 1645 cmϪ1
Cryoscopy. Addition of 2g (0.275 g) to 25 ml of benzene
(0.0348 M solution, expressed as monomer) depressed its
freezing point by 0.149 0.002 ЊC, corresponding to an average
molecular weight of 425.7 10, association state of 1.3.
(s, ν(C᎐N)), 1595 cmϪ1 (s, ν(C᎐C, aromatic)).
᎐
᎐
2,4-(2,6-Diisopropylphenylimino)-3,3-diethylpentane
(3ab).
Using 1a and EtBr. Crystals from hexane solution at Ϫ25 ЊC;
mp: 142–144 ЊC. Yield: 1.68 g, 74%. 1H NMR (200 MHz;
2,4-(2-Isopropylphenylimino)-3,3-dimethylpentane (3ca)
3
CDCl3): δ 1.2 (12H ϩ 12H, two d overlapped, JHH = 6.8 Hz,
BunLi (5.26 ml of a 1.42 M hexane solution, 7.49 mmol) was
added dropwise to a stirring solution of 1c (2.50 g, 7.47 mmol)
in hexane (40 ml) at 22 ЊC. The solution turned bright yellow.
Methyl iodide (0.49 ml, 7.47 mmol) diluted in hexane (6.71 ml)
was then added at 0 ЊC and the resultant mixture was allowed to
warm to 22 ЊC with stirring. After 1 h a dense white precipitate
had formed and the yellow colouration had disappeared. The
mixture was stirred for a further 4 h. A further 5.25 ml of 1.42
M BunLi in hexane was then added which turned the mixture
yellow. The mixture was again cooled to 0 ЊC and a further 7.20
ml of methyl iodide/hexane solution was added. The resultant
mixture was allowed to return to 22 ЊC and stirred overnight
under argon after which time the yellow colouration had
disappeared and a large amount of white precipitate had
formed. A water/hexane extraction was carried out to isolate
3ca in the hexane fraction. The hexane fraction was dried using
anhydrous MgSO4 and the hexane was evaporated, yielding
a pale brown oil. A hexane solution of the oil yielded large
colourless crystals of 3ca after ca. 7 days at Ϫ25 ЊC. However,
the yield was poor due to the high solubility of the β-diimine in
hexane. Recrystallisation from methanol gave a higher yield.
Yield: 1.95 g, 72%; mp: 64–66 ЊC. 1H NMR (200 MHz; CDCl3):
3
3JHHЈ = 6.8 Hz, MeCHMe); 1.01 (6H, t, JHH = 7.5 Hz,
NCMeC(CH2Me)2CMeN); 1.80 (6H, s, NCMeC(Et)2CMeN);
3
2.32 (4H, q, JHH = 7.5 Hz, NCMeC(CH2Me)2CMeN); 2.81
3
(4H, septet, JHH = 6.8 Hz, MeCHMe); 7.04–7.21 (6H, non-
first-order m, aromatic protons). 13C NMR (75 MHz; CDCl3):
δ 9.0 (NCMeC(CH2Me)2CMeN); 18.1 (NCMeC(Et)2CMeN);
23.5 and 23.9 (MeCHMe); 24.6 (NCMeC(CH2Me)2CMeN);
28.3 (MeCHMe); 62.3 (NCMeC(Et)2CMeN); 123.4 and 123.6
(aromatic CH); 136.0 (C(CH(Me)2), aromatic C); 147.0 (C–N,
aromatic C); 173.1 (C᎐N). Elemental analysis, Calcd. for
᎐
C33H50N2: C, 83.5; H, 10.6; N, 5.9. Found: C, 83.6; H, 10.4; N,
6.0%. IR: 1640 cmϪ1 (s, ν(C᎐N)), 1590 cmϪ1 (s, ν(C᎐C,
᎐
᎐
aromatic)).
2,4-(2-Isopropylphenylimino)-3,3-dibenzylpentane
(3cc).
Using 1c and PhCH2Br, crystals from hexane; mp: 91–93 ЊC.
Yield: 2.96 g, 76%. 1H NMR (200 MHz; CDCl3): δ 1.14 (12H, d,
3JHH = 7.0 Hz, MeCHMe); 1.91 (6H, s, NCMeC(CH2Ph)2-
CMeN); 2.65 (2H, septet, 3JHH = 7.0 Hz, MeCHMe); 3.70 (4H,
s, NCMeC(CH2Ph)2CMeN); 6.20–6.30 and 7.01–7.49 (18H,
non-first-order m, aromatic protons). 13C NMR (75 MHz;
CDCl3): δ 18.9 (NCMeC(CH2Ph)2CMeN); 23.5 (MeCHMe);
28.1 (MeCHMe); 38.4 (NCMeC(CH2Ph)2CMeN); 63.0
(NCMeC(CH2Ph)2CMeN); 118.1, 124.1, 126.0, 126.6, 126.9,
128.6 and 130.8 (aromatic CH); 138.0 (C(CH(Me)2), aromatic
C); 138.9 (NCMeC(CH2Ph)2CMeN, substituted aromatic C);
3
δ 1.21 (12H, d, JHH = 7.0 Hz, MeCHMe); 1.60 (6H, s,
NCMeC(Me)2CMeN); 1.85 (6H, s, NCMeC(Me)2CMeN); 2.98
(2H, septet, 3JHH = 7.0 Hz, MeCHMe); 6.51–6.57 and 7.04–7.35
(8H, non-first-order m, aromatic protons). 13C NMR (75 MHz;
CDCl3): δ 17.2 (NCMeC(Me)2CMeN); 23.2 (MeCHMe); 24.3
(NCMeC(Me)2CMeN); 28.4 (MeCHMe); 54.8 (NCMeC-
(Me)2CMeN); 118.6, 123.9, 126.0 and 126.5 (aromatic CH);
138.1 (C(CH(Me)2), aromatic C); 149.2 (C–N, aromatic C);
148.8 (C–N, aromatic C); 171.4 (C᎐N). Elemental analysis,
᎐
Calcd. for C37H42N2: C, 86.3; H, 8.2; N, 5.4. Found: C, 86.4; H,
8.2; N, 5.5%. IR: 1636 cmϪ1 (s, ν(C᎐N)), 1595 cmϪ1 (s, ν(C᎐C,
᎐
᎐
aromatic)).
173.6 (C᎐N). Elemental analysis, Calcd. for C H N : C, 82.8;
᎐
25 34
2
H, 9.5; N, 7.7. Found: C, 82.9; H, 9.8; N, 7.9%. IR: 1645 cmϪ1 (s,
2,4-(2,6-Diisopropylphenylimino)-3,3-dibenzylpentane (3ac).
Using 1a and PhCH2Br, crystals from hexane; mp: 148–150 ЊC.
Yield: 2.96 g, 69%. 1H NMR (200 MHz; CDCl3): δ 1.0 (12H ϩ
ν(C᎐N)), 1593 cmϪ1 (s, ν(C᎐C, aromatic)).
᎐
᎐
Other β-diimines were prepared similarly.
3
12H, two d overlapped, JHH = 7.0 Hz, 3JHHЈ = 7.0 Hz,
2,4-(2,6-Diisopropylphenylimino)-3,3-dimethylpentane (3aa).
Using 1a and MeI, crystals from hexane solution at Ϫ25 ЊC;
mp: 104–106 ЊC. Yield: 2.16 g, 81%. 1H NMR (200 MHz;
MeCHMe); 1.86 (6H, s, NCMeC(CH2Ph)2CMeN); 2.22 (4H,
3
septet, JHH = 7.0 Hz, MeCHMe); 3.94 (4H, s, NCMeC-
(CH2Ph)2CMeN); 6.98–7.14, 7.31–7.43 and 7.54–7.63 (16H,
non-first-order m, aromatic protons). 13C NMR (75 MHz;
CDCl3): δ 19.2 (NCMeC(CH2Ph)2CMeN); 23.0 and 24.1
(MeCHMe); 28.1 (MeCHMe); 38.9 (NCMeC(CH2Ph)2-
CMeN); 64.4 (NCMeC(CH2Ph)2CMeN); 123.3, 123.6, 127.2,
128.9 and 131.3 (aromatic CH); 135.7 (C(CH(Me)2), aromatic
C); 139.0 (NCMeC(CH2Ph)2CMeN, substituted aromatic C);
CDCl3): δ 1.20 and 1.22 (12H ϩ 12H, two d overlapped, 3JHH
=
6.8 Hz, 3JHHЈ = 6.8 Hz, MeCHMe); 1.66 (6H, s, NCMeC(Me)2-
CMeN); 1.79 (6H, s, NCMeC(Me)2CMeN); 2.84 (4H, septet,
3JHH = 6.8 Hz, MeCHMe); 7.01–7.20 (6H, non-first-order
m, aromatic protons). 13C NMR (75 MHz; CDCl3): δ 17.9
(NCMeC(Me)2CMeN); 23.4 and 23.7 (MeCHMe); 24.8
(NCMeC(Me)2CMeN); 28.3 (MeCHMe); 55.2 (NCMeC(Me)2-
CMeN); 123.3 and 123.7 (aromatic CH); 136.5 (C(CH(Me)2);
146.5 (C–N, aromatic C); 171.5 (C᎐N). Elemental analysis,
᎐
Calcd. for C43H54N2: C, 86.2; H, 9.1; N, 4.7. Found: C, 86.2; H,
146.2 (C–N, aromatic C); 174.4 (C᎐N). Elemental analysis,
᎐
9.4; N, 4.8%. IR: 1620 cmϪ1 (s, ν(C᎐N)), 1590 cmϪ1 (s, ν(C᎐C,
᎐
᎐
Calcd. for C31H46N2: C, 83.3; H, 10.4; N, 6.3. Found: C, 83.2; H,
aromatic)).
10.7; N, 6.5%. IR: 1647 cmϪ1 (s, ν(C᎐N)), 1590 cmϪ1 (s, ν(C᎐C,
᎐
᎐
aromatic)).
Attempted preparation of 2,4-(2-methoxyphenylimino)-3,3-
dimethylpentane
2,4-(2-Isopropylphenylimino)-3,3-diethylpentane (3cb). Using
1c and EtBr, crystals from hexane; mp: 87–89 ЊC. Yield: 2.30 g,
To a suspension of 1g (0.10 g, 0.32 mmol) in 10 ml n-hexane
was added BunLi (0.22 ml of a 1.45 M solution in hexanes),
followed by hexamethylphosphoramide (0.12 ml, 0.64 mmol).
To the resultant orange solution was added MeI (0.20 ml, 0.32
mmol) at rt. After 18 h stirring, further equimolar aliquots of
BunLi and MeI were added. A cloudy solution resulted. This
1
3
76%. H NMR (200 MHz; CDCl3): δ 1.21 (12H, d, JHH
=
3
7.5 Hz, MeCHMe); 0.96 (6H, t, JHH = 7.0 Hz, NCMeC-
(CH2Me)2CMeN); 1.82 (6H, s, NCMeC(Et)2CMeN); 2.20 (4H,
3
q, JHH = 7.5 Hz, NCMeC(CH2Me)2CMeN); 3.00 (2H, septet,
3JHH = 7.0 Hz, MeCHMe); 6.49–6.55 and 7.04–7.36 (8H, non-
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1085

View Article Online
was taken up in hexane (5ml) and washed with water (3 × 5 ml),
dried over MgSO4 and filtered. Removal of solvent in vacuo
yielded an oil which gave a very complex NMR spectrum. TLC
on silica in 4 : 1 hexane/dichloromethane indicated components
of Rf 0.77, 0.34 and streaking from 0 to 0.11. Flash chromato-
graphy on silica in the same solvent system gave two fractions,
the first of which still contained numerous species. For example,
there appeared to be seven methoxy proton resonances in the
range 3.8–3.9 ppm. There was also a new peak at 2.8 ppm
(MeN?). Other runs in thf or hexane produced even greater
distributions of product. No further separation was attempted.
36.3, 38.4, 41.1 (adamantyl CH2, adamantyl OC(CH)2), 62.5,
63.4 (NCMeCH(Ad)CMeN); 85.6, 91.4 (CO) 121.5, 122.0,
123.2, 124.3, 124.7, (aromatic CH); 136.1, 137.3, 138.2, 138.8
(C(CH(Me)2), aromatic C); 148.6, 149.8 (aromatic C–N); 169.0,
170.6 (C᎐N); 180–230 ppm region totally silent.
᎐
Quenching experiments. (a) Solution phase. A crystalline
lump of 4 (0.26 g) was removed from the glovebox, taken up in
hexane (10 ml) and washed with distilled water (2 × 20 ml). The
organic phase was separated, dried in vacuo and subjected to
NMR analysis. Peaks corresponding to 1c and adamantanone
were observed. Using saturated aqueous NH4Cl in place of
pure water produced identical results.
Lithium {2-[2-(2-isopropyl)phenylimino-1-(1-(2-isopropyl)-
phenyliminoethyl)propyladamantan-2-olate} (4)
(b) Solid phase. (i) In a glovebox, a sample of 4 (0.32 g) was
ground in an agate mortar and pestle with solid anhydrous
NH4PF6 (0.55 g, excess) for 15 minutes. The mixture was
extracted into deuterobenzene and filtered through glass wool
into an NMR tube. NMR analysis indicated resonances
corresponding to 1c and adamantanone.
To a stirred suspension of 1c ( 2.44 g, 7.3 mmol) in hexane
(8 mL) in a Schlenk tube at 0 ЊC was added BunLi (4.86 ml of a
1.51 M solution in hexanes, 7.3 mmol). Heat and butane were
evolved. The resultant pale yellow solution was re-cooled to 0
ЊC with stirring, causing a pale cream, fine precipitate to form.
To this was added adamantan-2-one (1.095 g, 7.3 mmol), which
caused momentary dissolution of the precipitate, but solid re-
precipitated upon stirring for 5 min. The suspension was evacu-
ated to remove butanes and concentrated in vacuo to approxi-
mately 8 ml, then briefly heated to boiling to re-dissolve the
precipitate. A crop of pale yellow blocks was deposited after
standing overnight. These were isolated by filtration to yield
(ii) In a glovebox, a sample of 4 (0.27 g) and ButMe2SiCl
(0.35 g, excess) were ground as above for 15 minutes. The
mixture fully dissolved in deuterobenzene, and exhibited NMR
resonances corresponding to solutions of 4 plus ButMe2SiCl.
No reaction had occurred.
X-Ray crystallography
1
2.42 g (4.93 mmol, 68%) of 4; mp: 76–78 ЊC. H NMR (400
The structures of 1f, 2c, 2g and 4 were determined by X-ray
crystallography. Crystals were selected from the mother-liquor
and mounted using the oildrop method (Fomblin 1800 oil).
Experimental parameters are summarized in Table 1.
CCDC reference numbers 199213 and 199312–199314.
See http://www.rsc.org/suppdata/dt/b2/b212079h/ for crystal-
lographic data in CIF or other electronic format.
MHz; C6D6, 300K): δ 1.2–1.5 (24H, overlapping m, MeCHMe
and adamantyl resonances); 1.85 (2H, apparent br s, adamantyl
OC(CH )2; 2.14 (6H, s, NCMeCH(Ad)CMeN); 3.68 (2H, sept,
3JHH = 7.0 Hz, MeCHMe); 5.20 (1H, s, NCMeCH(Ad)CMeN);
7.1–7.4 (8H, non-first-order m, aromatic protons). 1H NMR in
deuterohexane was also recorded. Though it was not possible to
lock on the signal, and some field drift was present, the peaks
were essentially the same. The alkenyl resonance (NCMeCH-
(Ad)CMeN) shifted slightly upfield to 4.85 ppm. 13C NMR (100
MHz; C6D6, 300 K): δ 23.8 (MeCHMe); 24.2 (NCMeCH-
(Ad)CMeN); 27.5 (MeCHMe); 28.3 (adamantyl CH); 36.1,
39.5 (adamantyl CH2) 47.3 (adamantyl OC(CH)2), 94.7
(NCMeCH(Ad)CMeN); 122.8, 125.0, 126.0 and 126.4 (aro-
matic CH); 141.8 (C(CH(Me)2), aromatic C); 153.3 (aromatic
Results and discussion
Enamineimine preparations
The direct acid-catalyzed condensation route from acetyl-
acetone and the appropriate aniline by Dean–Stark azeotropic
distillation which proceeded rapidly with modest though
acceptable yields in the previously discussed4 case of 1a worked
acceptably for 1c and exceptionally well for the methoxy-
substituted 1g, resulting in an 87% yield after 4 hours of
reaction. This provided a rapid and convenient route to the first
ortho-functional, potentially tetradentate diiminate ligands.
However, it did not proceed well for 1f, the fluorinated variant
of 1a; at the temperature required to remove the water of
condensation, the hexafluoroacetylacetone reactant was lost.
An alternative means of encouraging condensation was
required. Recently, it has been shown that a single CF3 group
could be introduced to diiminate ligands by a C–C coupling
methodology18 reminiscent of that used6 to prepare the exceed-
ingly bulky 1b, which also is not accessible by means of direct
condensation of 1,3-diketones with anilines. However, our
preparation of the bulky, symmetrically fluorinated 1f retained
the condensation route, but facilitated reaction at lower tem-
perature by reacting with a preformed titanium tetrachloride/
diisopropylaniline mixture. Such mixtures are not well charac-
terized in the case of diisopropylaniline, and resisted our
attempts to remedy this situation, but in cases concerning other
amines are known to contain amidotitanium and imido-
titanium species.19 These may be the true reacting species in the
condensation used to prepare 1f. The stoichiometry is explained
in Scheme 1. A similar method has been used to prepare N,NЈ-
cyclohexyl variants of 1.20
C–N); 163.7 (C᎐N); 227 (CO). Elemental analysis, Calcd. for
᎐
C33H43N2OLi: C, 80.8; H, 8.8; N, 5.7. Found: C, 79.6; H, 9.2; N,
5.4%. IR: 1650, 1626 cmϪ1 (s, ν(C᎐N), syn and anti), 1596 cmϪ1
᎐
(s, ν(C᎐C, aromatic)). Raman showed the same vibrations. On
᎐
exposure of the IR plates to moist air, 3566 (sharp, LiOH), 1720
(C᎐O), 1628, 1556 (free 1c) cmϪ1. Attempts to collect data by
᎐
dissolving 4 in freshly distilled hexane and injecting the solution
into a pre-dried and flushed IR solution cell were thwarted by
hydrolysis, as indicated by peaks in the 1720, 1628 and 1555
cmϪ1 regions.
Cryoscopy. Addition of 0.182 g of 4 to 25 ml of benzene
(0.0148 M solution, expressed as monomer) depressed its
freezing point by 0.092 2 ЊC, corresponding to an average
molecular weight of 456.3
10, association state of 0.93;
a 0.041 M solution gave an average molecular weight of
452.4 10.
Solid state CP-MAS NMR. A powdered sample of 4 was
packed into a rotor in a glovebox, and 13C{1H} spectra were
recorded in portions over a time period of 12 h to check for
decay. There was none, and so all data sets were combined
to give a low-noise spectrum. Resolution was good, with
symmetry inequivalent/chemically equivalent peaks often being
resolved, though in some cases the two (or four) resonances
overlapped with other groups of resonances, obscuring assign-
ments: δ 20.5, 21.1 (MeCHMe); 24.5 (NCMeCH(Ad)CMeN);
25.1, 26.2, 26.6, 27.3, 28.8 (MeCHMe, adamantyl CH); 34.7,
Though the yield was not high, usable quantities of the
fluorinated ligand are available by performing the reaction on a
large scale.
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1086

View Article Online
Table 1 X-Ray data collection and refinement detailsa
1f
2cؒ2thf
2g
4
Formula
M
Crystal system
a/Å
b/Å
c/Å
α/Њ
β/Њ
C29H36F6N2
526.6
C35H53LiN2O3
556.73
Trigonal
32.123(9)
32.123(9)
9.273(2)
(90)
(90)
(120)
R3m (160)
9
C38H42Li2N4O4
632.64
Triclinic
C66H86Li2N4O2
981.27
Orthorhombic
10.6269(3)
22.2662(3)
23.8287(4)
(90)
(90)
(90)
P212121 (19)
4
Monoclinic
11.5857(10)
9.3714(10)
26.8363(10)
(90)
92.389(10)
(90)
10.6917(4)
11.8407(5)
15.3648(9)
77.010(2)
85.387(2)
66.242(2)
γ/Њ
¯
Space group (no.)
Z
P21/n (15)
4
P1 (2)
2
T /K
293
0.097
203
0.062
150
0.078
150
0.068
µ/mmϪ1
Reflns. measd.
Reflns. obsdb (Rint
R1(observed)
wR2(all data)c
5173
4916 (0.0139)
0.0571
5339
21314
7179 (0.0748)
0.0592
0.1619
37520
10524 (0.0603)
0.0432
)
1817 (0.0682)
0.0605
0.1434
0.1482
0.0848
a Programs: Data collection: COLLECT,13 DENZO-SMN,14 SORTAV.15 Solution and refinement: SHELXS-9716 or SIR92,17 SHELXL-97.16
2
½
b I = I > 2σ(I ). c wR2 = {σ[w(Fo2 Ϫ Fc2)2]/σ[w(Fo )2]} .
Table 2 Selected bond lengths (Å) and angles (Њ) for 1f
F(1/5)–C(1/5)
N(6)–C(2)
N(6)–C(7)
N(19)–C(4)
N(19)–C(20)
1.306–1.341(3–4)
1.357(3)
1.442(3)
1.289(3)
1.434(3)
C(1)–C(2)
C(2)–C(3)
C(3)–C(4)
C(4)–C(5)
1.500(4)
1.363(4)
1.429(3)
1.531(4)
C(2)–N(6)–C(7)
C(4)–N(19)–C(20)
N(6)–C(2)–C(3)
N(6)–C(2)–C(1)
C(3)–C(2)–C(1)
129.6(2)
125.9(2)
124.2(2)
117.6(2)
118.1(2)
C(2)–C(3)–C(4)
N(19)–C(4)–C(3)
N(19)–C(4)–C(5)
C(3)–C(4)–C(5)
124.8(2)
121.3(2)
125.3(2)
113.3(2)
Scheme 1
A structure of 1f was obtained, and it is of interest to
examine what effect, if any, the fluorination of the backbone
has on the ligand. Examination of Fig. 1 (see also Table 2)
reveals that the iminoenamine form exists in the crystal, and in
this respect it is identical to its non-fluorinated analogue, 1a.21
Interestingly, the bulkier 1b is known to adopt a true β-diimine
form in the solid state.22 The ligand core bond lengths show
an alternating pattern consistent with the localized, though
strongly hydrogen bonded, proton on N(1). There appears to
be little significant structural effect caused by fluorination, the
C–C and C–N distances of 1f being closely comparable with
those of 1a, notably in the alternation of C–C and C–N lengths
within the aminoenimine ring. The intramolecular hydrogen
bond (freely refined) between the amine and imine nitrogens
appears to be marginally weaker, the amino N–H distance
being 0.88(4) and the imino N ؒ ؒ ؒ H distance being 1.97(4) Å,
cf. 0.97(4) and 1.86(4) Å respectively, for 1a;21 these differences
are, however, within error. The most significant difference is to
be found sterically: the fluorination of the backbone methyl
groups causes increased repulsion of the aryl groups, squeezing
the co-ordinating void of the ligand. This is manifested in
C–N(H)–C angles of 130Њ, and C᎐N–C angles of 126Њ, which
᎐
correspond to angles of 123Њ and 124Њ in the unfluorinated 1a.
In the yet bulkier 1b,22 the angles are more comparable with
those in 1f, but since the tautomer is different, the comparison
is less revealing.
While the structure of 1f showed little change with respect
to unfluorinated 1a, its reactivity was considerably different.
Successful alkylation experiments described below for 1a and 1c
failed for 1f. These, and further differences, were addressed by
probing the structures of the lithiated intermediates.
Preparation and structures of lithium diazapentadienyl
complexes
The preparation and structures of 2a have already been recently
reported by other workers.21 It is known to exist in two
unsolvated forms, one dimeric and one dodecameric, though it
Fig. 1 ORTEP46 diagram (40% probability) of 1f.
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1087

View Article Online
Table 3 Selected bond lengths (Å) and angles (Њ) for 2c
diazapentadienyl complex which has been established to be
4-co-ordinate, since high degrees of steric bulk tend to be
employed. The first was a dimeric hexamethylphosphoramide-
bridged structure, with unsubstituted phenyl groups.24 In
common with all structures derived from 1a or 1b, the aryls
in 2c lie almost perpendicular to the diazapentadienyl plane
(the aryls are twisted only 7.9Њ from the perpendicular to the
NCCCN plane). It is the syn arrangement of the isopropyl
groups which generates the space for the second molecule of thf
to bind. In fact, the lithium lies close (only 0.49 Å above) to the
plane formed by O(2) and the two nitrogen atoms such that
these atoms together almost form a basal plane, with the other
thf occupying an ‘axial’ position. This plane is shared by the
diazapentadienyl fragment and the equatorial thf in a manner
which makes the mapping with the related diisopropylaryl
structure 2a very close. This is mirrored in the similarity of
behaviour with respect to electrophiles, vide infra. While 1a and
1b have proved, by virtue of the shape of the ligating cavity
of the anions, to be highly efficient in stabilizing unusual
examples of trigonal planar co-ordination or bent di-co-ordin-
ation, the less bulky 1c may be expected to offer frequent
examples of heavily distorted tetrahedral co-ordination. Alter-
natively, the metal may distort out of the diazapentadienyl
plane to give envelope conformations, which was the outcome
in the homoleptic zinc complex which is the only other
structurally characterized complex from 1c.12
C(1)–N(1)
C(1)–C(2)
C(1)–C(3)
1.3121(17)
1.3979(13)
1.5110(16)
N(1)–Li(1)
O(1)–Li(1)
O(2)–Li(1)
1.955(2)
1.994(3)
1.947(3)
N(1)–C(1)–C(2)
N(1)–C(1)–C(3)
C(2)–C(1)–C(3)
C(1)–C(2)–C(1)#1
C(1)–N(1)–C(4)
C(1)–N(1)–Li(1)
123.88(10)
119.86(9)
116.26(12)
128.80(16)
121.93(9)
121.42(9)
C(4)–N(1)–Li(1)
O(2)–Li(1)–N(1)
N(1)#1–Li(1)–N(1)
O(2)–Li(1)–O(1)
N(1)–Li(1)–O(1)
115.88(11)
122.61(10)
95.84(14)
102.98(15)
105.49(12)
Symmetry transformation used to generate equivalent atoms: #1 Ϫy ϩ
1, Ϫx ϩ 1, z.
is assumed that in solution monomers are present. From diethyl
ether or thf, monomeric monosolvates with trigonal planar
lithium environments are obtained,21 as is the case with 2b.6
Complex 2c, being less bulky, rapidly precipitates from hexane
solution upon metallation with BunLi, but crystals were
obtained from a thf/hexane solution (Fig. 2, Table 3). The
molecule crystallizes in the space group R3m; the trigonal
packing arrangement generates channels which are occupied by
disordered solvent. A further thf solvent is trapped along
a crystallographic plane. These are predominantly lost upon
isolation by vacuum filtration, but served to make elemental
analysis figures irreproducible. It proved difficult to satis-
factorily model all the solvent in the crystal, since the three-fold
symmetry of the channel was difficult to reconcile with the
five-membered ring of thf.23 The NMR spectra confirm that
both co-ordinated thf molecules survive the isolation pro-
cedure. The spectra indicate that, with regard to the isopropyl
configurations, rapid equilibration of syn and anti conformers
is occurring at room temperature. Low temperature studies
were not attempted.
The reaction of 1f with BunLi in hexane produced a crop
of small crystals of poor quality. A preliminary X-ray deter-
mination showed 2f to be co-crystallized with 1f, but the data
are not of publishable standard. It did show that the lithium in
2f was 2-co-ordinate monomeric, with none of the unusual
oligomerization motifs of 2a.
Lithiation of 1g in hexane gave a good crop of crystalline 2g,
which has the dimeric solid-state structure depicted in Fig. 3. In
this case, the single isopropyl substitution of 1c has been
replaced with the sterically similar though functionally different
methoxy group, intended to act as an internal chelation site. As
such, 2g represents only the second structurally characterized
tetradentate acyclic diketiminate ligand complex, the first being
the recently reported [{N(CH2CH2NEt2)C(Me)}2CHؒScCl2].25
There is a conflict between optimum co-ordination geometry
at lithium and ideal bond lengths and angles within the ligand.
This conflict is substantially resolved by dimerization (Fig. 3).
One diazapentadienyl unit of the dimer (binding through N(1)
and N(2)) has an anti configuration of 2-methoxy substituents,
Fig. 2 ORTEP diagram (50% probability) of the monomer 2cؒ2thf.
Hydrogens, one contributor to thf disorder, and a further disordered
non-co-ordinated solvent, are omitted for clarity. A crystallographic
plane cuts through C(2), Li(1), O(1) and O(2).
A crystallographic plane bisects the molecule down the C(2)–
Li(1)–O(2)–O(1) plane. It dictates the syn orientation of both
isopropyl groups, and (indirectly) the perfect planarity of the
diazapentadienyl ring and the thf containing O(2), to which it is
perpendicular.
The structure of monomer 2cؒ2thf appears to be similar
in most respects to the previously published monomer 2aؒthf
isolated from the same solvent,21 save for the obvious difference
that two molecules of thf are co-ordinated, rather than one,
thus giving the lithium its favoured tetrahedral environment,
though heavily distorted. This is in fact only the second lithium
Fig. 3 ORTEP diagram (30% probability) of the cisoid dimer 2g.
Hydrogens are omitted for clarity.
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1088

View Article Online
Table 4 Selected bond lengths (Å) and angles (Њ) for 2g
plane being 37.7Њ. It remains somewhat surprising that the
diazapentadienyl nitrogens were pressed into service as bridging
atoms when a plausible monomeric structure seemed able
to furnish the lithium with its favoured 4-co-ordination,
especially when the structure of 2g is compared to the structure
of [{PhN(CH)3NPhLiؒOP(Me2)3}2], in which dimerization
occurred via the bridging phosphoric amide oxygens in prefer-
ence to the diazapentadienyl nitrogens.24
O(1)–Li(2)
N(1)–C(2)
N(1)–Li(1)
N(1)–Li(2)
N(2)–C(4)
N(2)–Li(1)
O(3)–Li(1)
N(3)–Li(1)
Li(1)–Li(2)
C(1)–C(2)
C(2)–C(3)
C(3)–C(4)
2.000(4)
1.356(3)
2.009(4)
2.151(4)
1.317(3)
1.898(4)
1.997(4)
2.085(4)
2.432(5)
1.520(3)
1.376(3)
1.423(3)
C(4)–C(5)
1.502(3)
2.168(3)
1.343(2)
1.430(3)
2.148(3)
1.325(2)
1.995(4)
1.512(3)
1.389(3)
1.415(3)
1.505(3)
O(4)–Li(2)
N(3)–C(21)
N(3)–C(25)
N(3)–Li(2)
N(4)–C(23)
N(4)–Li(2)
C(20)–C(21)
C(21)–C(22)
C(22)–C(23)
C(23)–C(24)
In solution in C6D6, the simplicity and sharpness of the
NMR data are consistent only with rapidly equilibrating
isomers. Scheme 2 is intended to represent just two of the many
possible isomers showing exchange of cisoid and transoid
forms through a putative dissociated monomeric intermediate.
This hypothesis is supported by the fact that cryoscopy in
benzene at a similar concentration to that used for the NMR
measurements gives an average association state of 1.3. It is
tempting to speculate that the isolation of the cisoid arrange-
ment is due to some geometric or steric conflict in more
symmetrically bound dimers. There is some intramolecular
(intermonomer) offset π-stacking interaction between the aryls
carrying O(1) and O(3), e.g. C(7)–C(25) = 3.26 Å. However, the
reason may simply be because the higher dipole moment of a
non-centrosymmetric molecule makes it less well-solvated by
hexane; hence the cisoid isomer is deposited. Furthermore, the
relatively high melting point suggests efficient intermolecular
crystal packing in this conformation.
N(2)–Li(1)–N(3)
O(3)–Li(1)–N(3)
N(1)–Li(1)–N(3)
N(1)–C(2)–C(3)
N(1)–C(2)–C(1)
C(3)–C(2)–C(1)
C(2)–C(3)–C(4)
N(2)–C(4)–C(3)
N(2)–C(4)–C(5)
C(3)–C(4)–C(5)
C(21)–N(3)–C(25)
C(21)–N(3)–Li(1)
C(2)–N(1)–C(6)
C(2)–N(1)–Li(1)
C(6)–N(1)–Li(1)
C(2)–N(1)–Li(2)
C(6)–N(1)–Li(2)
Li(1)–N(1)–Li(2)
C(4)–N(2)–C(13)
C(4)–N(2)–Li(1)
C(13)–N(2)–Li(1)
N(2)–Li(1)–O(3)
N(2)–Li(1)–N(1)
O(3)–Li(1)–N(1)
143.2(2)
C(25)–N(3)–Li(1)
C(21)–N(3)–Li(2)
C(25)–N(3)–Li(2)
Li(1)–N(3)–Li(2)
C(23)–N(4)–C(32)
C(23)–N(4)–Li(2)
C(32)–N(4)–Li(2)
N(4)–Li(2)–O(1)
N(4)–Li(2)–N(3)
O(1)–Li(2)–N(3)
N(4)–Li(2)–N(1)
O(1)–Li(2)–N(1)
N(3)–Li(2)–N(1)
N(4)–Li(2)–O(4)
O(1)–Li(2)–O(4)
N(3)–Li(2)–O(4)
N(1)–Li(2)–O(4)
N(3)–C(21)–C(22)
N(3)–C(21)–C(20)
C(22)–C(21)–C(20)
C(21)–C(22)–C(23)
N(4)–C(23)–C(22)
N(4)–C(23)–C(24)
C(22)–C(23)–C(24)
104.03(16)
121.69(15)
115.71(14)
70.11(15)
121.62(17)
127.90(17)
109.49(15)
165.0(2)
88.84(14)
96.35(15)
111.08(19)
82.29(14)
97.56(15)
77.99(13)
92.27(14)
157.0(2)
104.72(15)
123.52(17)
120.04(19)
116.43(18)
130.27(18)
121.60(18)
123.16(18)
115.18(18)
82.01(15)
104.36(18)
124.03(17)
119.52(19)
116.44(19)
130.49(19)
123.27(19)
120.75(19)
115.96(19)
118.56(16)
113.99(17)
118.19(16)
117.51(17)
117.43(15)
117.07(15)
105.27(17)
71.45(15)
120.20(18)
121.42(17)
118.36(17)
116.9(2)
99.98(17)
106.57(17)
in contrast with the 2-isopropyl variant 2c. It shares with most
other N–NЈ diaryl diazapentadienyl complexes a near per-
pendicular relationship between the aryl planes and the diaza-
pentadienyl NCCCN plane (deviations of ϩ11.9 and Ϫ8.1Њ).
However, the other anion, containing ligating sites N(3) and
N(4), has one ring near perpendicular (ϩ13.4Њ), but another
ring much closer to coplanarity, at a rotation angle of 43.9Њ
from perpendicular. It is this aryl which is the only one to
engage in ‘intramonomer’ chelation, that is to say, O(4) of this
aryl binds to the lithium which forms the six-membered ring
through N(3) and N(4), part of the same anion. In contrast,
two other methoxy arms serve to bind the dimer together by
forming inter-monomer links. The final methoxy arm lies
unbound. There are no significant intermolecular interactions.
The ‘intramonomerЈ chelation through O(4) has the longest
Li–O bond. (Table 4). The two monomeric units are also bound
by bridging nitrogens, N(1) and N(3), in a manner reminiscent
of the first structurally characterized lithium diketiminate,
[{Me3SiNC(Ph)CHC(Ph)CNSiMe3Li}2].26 The co-ordination
of two lithium ions to each of N(3) and N(1) serves to localise
more negative charge on those nitrogens, so that the alternation
of bond lengths is almost as clear in both diazapentadienyl
rings of 2g as in the case of 1f, in stark contrast to 2c, where the
completeness of delocalization is evidenced by the crystallo-
graphic symmetry plane bisecting the diazapentadienyl ring.
This incomplete delocalization is also seen in [{Me3SiNC(Ph)-
CHC(Ph)CNSiMe3Li}2]. In that case, each lithium was
merely 3-co-ordinate,26 whereas in 2g there is one 4-co-ordinate
(pseudo-tetrahedral) and one 5-co-ordinate (pseudo-square-
based-pyramidal) environment. The other major difference
between these two dimeric lithium diketiminates is that the
3-co-ordinate example was transoid, whereas 2g has an unusual
cisoid arrangement of diazapentadienyl planes.27 There is a
significant butterfly puckering of the central Li2N2 ring, the
angle between the Li(1)N(1)Li(2) plane and the Li(1)N(3)Li(2)
Scheme 2
The fact that one 5- and one 4-co-ordinate lithium is found
in 2g suggests that there is a fine balance between the extra
energy to be gained from the fifth co-ordination event and the
increased steric and angle-strain necessary to make this happen.
Reactions of diazapentadienyllithiums with electrophiles
While compounds 1 are often referred to as β-diketimines, their
observation in this form is rare; ‘enamineimine’ is a more
accurate term for the normal tautomer in the solid state and in
solution.21 The very bulky 1b adopts the β-diketimine tauto-
mer,22 and 1a has been observed once in this neutral diimine
form, co-ordinated to nickel.3 This first report of the now
omnipresent ligand 1a also included a palladium complex, but
it was one where the palladium atom was co-ordinated to the
α-carbon. The complex 1aؒNiBr2 was activated using methyl-
aluminoxane to give an ethylene polymerization catalyst of very
modest efficacy.3 Almost certainly, the α-CH would have been
deprotonated under these conditions.28 Noting the recent inter-
est in α-diimines,29 we reasoned that it should be possible to
access ‘true’ β-diketimines by a double alkylation protocol,
removing the problematic C–H acidity by α,αЈ dialkylation.
Such rational double C-alkylation of eneamineimines has, to
our knowledge, never been previously attempted, though in one
of only two prior reports, concerning alkylation of a variant of
1 lacking ortho substitution, a dialkylated species was reported
in low yield among the numerous products of an intended
single alkylation.30 The second, also with only a single alkyl-
ation, was more recent, and involved reaction with CH2Br2
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1089

View Article Online
in order to link two diketiminate ligands.31 This report is the
first of which we are aware specifically directed at synthesis of
simple β-diimines lacking α-carbon reactivity. There have been
some other more recent examples of mono-substitution at the
α-carbon.7–9
attack is not hindered for the α-carbon position. More puzzling
is the case of 1g. Repetition of the double lithiation/methyl-
ation procedure on 1g in thf, where a single methoxy group
took the place of the single isopropyl group of 1c, produced a
forest of products which resisted chromatographic separation.
The mixture was so complex as to indicate that the route, which
was clean for 1a and 1c, would not be preparatively useful
for 1g. However, there was evidence of N-alkylation, and in
some of the fractions also there was evidence of addition of
BunLi across imine bonds; both modes of reactivity have been
previously observed in the case where no ortho substituents
were present.30 Despite the fact that sterically the methoxy
group of 1g almost matches the isopropyl group of 1c, func-
tionally it differs. Since the cryosopic result implies significant
amounts of monomer in benzene solution, we can assume a
predominance of monomer in thf. In 2g, ortho-chelation of the
methoxy groups is likely to occur, giving rise to an aryl–diaza-
pentadienyl angle more akin to the unsubstituted case (as
shown in the structure of 2g by the sole intramolecularly
co-ordinated methoxyaryl) and therefore again failing to
prevent approach of methyl iodide to the diazapentadienyl
nitrogen atoms. The reaction was also attempted in hexane,
with and without added hexamethylphosphoric triamide
(hmpa). The least cluttered data, where N-alkylation was
clearly contributing, were obtained from the hmpa run, where
the strong donor was employed in order to free the methoxy
group from its chelating role, but clean C,C-dialkyl product
from 1g was not accessible by any means attempted.
By a cycle of deprotonation/alkylation (Scheme 3), it proved
possible to dimethylate, diethylate, or dibenzylate 1a and 1c
cleanly. Isolated yields were limited only by the high solubility
of the products in most organic solvents. Most notable in
these syntheses was the absence of N-alkylation, or any other
identifiable side-reaction, by NMR analysis of crude product.
The double alkylation provides convenient access to diaryl
β-diimines with significant ortho bulk. The previous work on
alkylation of diaryl diazapentadienyls lacked this feature, and
was plagued by multiple by-products.30 Other conceivable
routes involving direct condensation of anilines with 3,3Ј-di-
methylpentan-2,4-dione are likely to be difficult in cases of such
bulk, and in any case would not give control of imine geometry,
which was exclusively E/E, as appropriate for metal chelation,
possibly as a result of freezing of the structure of the chelated
lithium intermediate. While so-called β-diimines have been
known since the early work of Holm,1 these are the first which
exist exclusively in the tautomeric form implied by that name,
rather than the much more abundant eneamine-imine form,21
save for 1b. In the case of 1b, two isomers were present in the
solid state, one E/Z and one E/E with respect to imine conform-
ation, both differerent from the E/E conformation of 3.†
Notwithstanding these limitations, dialkylation appears to be
an attractive method for provision of gram quantities of bulky
diaryl β-diimines lacking acidic protons, the co-ordination
chemistry of which will be compared with their α-diimine
counterparts in a forthcoming paper.32
The fluorination in 1f appears to calm the reactivity of the
central carbon atom, perhaps by absorbing excess electron
density. No alkylation was found; only unreacted 1f was
recovered.
In order to determine if mixed alkylations could be per-
formed, a single methylation was attempted on 2a, but only a
reduced amount of 3aa, and unreacted 1a, were isolated. This
implied that a fast equilibrium was set up as the rather slow
alkylation proceeded. As the slow alkylation began, a neutral
monoalkylated molecule which could exchange protons with
unalkylated 2a was produced (see Scheme 4).
The greater electron-richness of the monomethylated 2a
α-carbon made it the preferred site of attack for the remainder
of electrophile. Interestingly, the older paper on the matter,
where monoalkylation was the target, also reported significant
amounts of dialkylated product.30 More recent, successful
monofunctionalizations at position R4 have utilised more
reactive electrophiles (CF3SO2Cl), where the greater rate of
reaction and the opposite inductive effect of the substituent
eliminates the problem of equilibration.7
Scheme 3
The complexity experienced by the earlier workers, on di-
ketiminates bearing no ortho substitution,30 suggests that the
structural features of the diazapentadienyllithium inter-
mediates are critical in determining the efficacy of the
alkylation. The structure of 2a is found to be a weakly-bound
hexamer or polymer depending on crystallization conditions,
but in all cases a solution-state monomer, even in non-co-ordi-
nating solution.21 In thf a single solvent molecule is bound to
the lithium in 2a; in 2c, two thf molecules can be incorporated.
These are most likely the identity of the reacting species in thf.‡
In both cases the nitrogen atoms are well shielded from
attacking nucleophiles by the isopropyl and solvent molecules;
the aryl rings lie close to perpendicular with the diazapenta-
dienyl ring. The same is not true of diaryldiazapentadienyl-
lithium complexes lacking ortho substitution; in this case, the
aryls lie between coplanarity and orthogonality,24 thereby
allowing attack from the electrophile from above or below the
diazapentadienyl plane on the nitrogen atoms. In all cases, such
The most interesting results on electrophilic attack of 2 were
obtained by use of adamantanone as the electrophile. These
results have been previously communicated.10 Addition of
adamantanone to a hexane suspension of 2c gave a solution
from which the addition product 4 was isolated. A new C–C
bond had formed by addition of the carbanion form of
the diazapentadienyl unit across the C᎐O double bond. Where
᎐
the diazapentadienyl anion is seen as a resonance-stabilized
aza version of an enolate, this reaction is a type of aldol
addition. The dimeric structure of the resultant lithium
β-diimine-alkoxide is depicted in Fig. 4. The new ligand has a
‘scorpionate’ architecture, giving an O,N,N alternative33 to the
N,N,N, donor set of the tris(pyrazolyl)borates.34
The bulk identity of the crystallographically characterised
4 was confirmed by IR, Raman, elemental analysis and solid-
state NMR, as well as a reproduction of the single-crystal
structure determination from independently prepared batches.
† The differing substitution patterns change nomenclature priorities
about the C᎐N bonds such that Z/Z 4 would be topologically equiv-
᎐
alent to E/E 2,2,6,6-tetramethyl-3,5-bis(2,6diisopropylphenylimino)-
heptane. The conformations were estabished by crystallography. The
structures of the free ligands will be reported along with co-ordinated
examples in a forthcoming paper.
‡ While the details in the Experimental section report our alkylations
performed in hexane, we have found equally clean products and high
yields using thf as solvent.
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1090

View Article Online
Scheme 4
Table 5 Selected bond lengths (Å) and angles (Њ) for 4 (values for one
crystallographically independent monomer only; the other is chemically
comparable). Hydrogen atoms have been removed for clarity
C(1)–C(2)
C(2)–N(1)
C(2)–C(3)
C(3)–C(4)
C(3)–C(24)
C(4)–N(2)
C(4)–C(5)
C(24)–O(1)
1.507(3)
1.284(3)
1.520(3)
1.514(3)
1.642(3)
1.278(3)
1.502(3)
1.353(2)
C(24)–C(25)
C(24)–C(29)
N(1)–Li(1)
N(2)–Li(1)
O(1)–Li(2)
O(1)–Li(1)
O(2)–Li(1)
O(2)–Li(2)
1.559(3)
1.562(3)
2.090(4)
2.079(4)
1.812(4)
1.874(4)
1.795(4)
1.875(4)
N(2)–C(4)–C(5)
N(2)–C(4)–C(3)
C(5)–C(4)–C(3)
O(1)–C(24)–C(25)
O(1)–C(24)–C(29)
C(25)–C(24)–C(29)
O(1)–C(24)–C(3)
C(25)–C(24)–C(3)
N(1)–C(2)–C(1)
N(1)–C(2)–C(3)
C(1)–C(2)–C(3)
C(4)–C(3)–C(2)
C(4)–C(3)–C(24)
C(2)–C(3)–C(24)
O(1)–Li(2)–O(2)
O(1)–Li(2)–N(42)
O(2)–Li(2)–N(42)
O(1)–Li(2)–N(41)
124.09(19)
119.71(18)
116.19(19)
111.58(17)
111.23(15)
106.29(17)
108.48(16)
109.78(15)
124.4(2)
120.33(17)
115.30(19)
115.02(17)
109.87(16)
109.25(16)
97.61(17)
129.36(19)
93.92(16)
139.3(2)
C(29)–C(24)–C(3)
C(2)–N(1)–C(6)
C(2)–N(1)–Li(1)
C(6)–N(1)–Li(1)
C(4)–N(2)–C(15)
C(4)–N(2)–Li(1)
C(15)–N(2)–Li(1)
C(24)–O(1)–Li(2)
C(24)–O(1)–Li(1)
Li(2)–O(1)–Li(1)
O(2)–Li(1)–O(1)
O(2)–Li(1)–N(2)
O(1)–Li(1)–N(2)
O(2)–Li(1)–N(1)
O(1)–Li(1)–N(1)
N(2)–Li(1)–N(1)
N(42)–Li(2)–N(41)
O(2)–Li(2)–N(41)
109.45(16)
120.35(16)
110.02(17)
125.02(16)
120.21(17)
111.09(16)
126.92(16)
154.79(17)
121.48(16)
81.80(15)
98.24(17)
128.86(19)
94.91(16)
137.6(2)
Fig. 4 An ORTEP diagram (50% probability) of 4 showing the
pseudo-centrosymmetric dimerization via bridging oxygens. Hydrogens
are omitted for clarity. Bonds linking the putative monomers are
greyed.
93.34(15)
90.18(15)
88.52(14)
93.32(15)
This unusual care was deemed necessary because of the
unexpected solution behaviour of 4: in deuterobenzene the
resonances were more in keeping with complex 2c co-ordinated
by adamantanone, which indicated that the aldol addition
reaction responsible for the formation of 4 had reversed upon
dissolution. Notable were the diagnostic solution chemical
shifts of H{C(3)/C(43)}, C(3/43), C(2/4/42/44) C(24/64) which
were almost identical to the corresponding shifts of 2cؒ2thf.
Most crucially, the C(24/64) resonance, which was, as expected
for an alkoxide carbon, at 91/85 ppm in the 13C CP MAS NMR
spectrum, had in deuterobenzene solution shifted to 227 ppm,
firmly in the carbonyl region. This represented a downfield
co-ordinative shift from free adamantanone at 215 ppm. In
hexane there was only a very slight upfield shift in H{C(3)/
C(43)}, which at 4.8 ppm was still in the alkenyl range expected
for 2c in solution, indicating that even in the solvent from which
4 crystallized, it existed predominantly as an uncoupled diaza-
pentadienyllithium–adamantanone complex.
It is known that lithium-mediated aldol additions are
disfavoured by polar, co-ordinating solvents and high temper-
atures,35 but these results indicate that it is possible to harvest
aldol addition product in the solid state at room temperature,
even where dissolution in the least polar of solvents totally
reverses the process. To shed further light on the phenomenon,
cryoscopy in benzene, the solvent used in most of the NMR
studies, was carried out. This indicated that in the con-
centration range used for spectroscopic measurement, the
species had dissociated to monomer 2cؒAd (Scheme 5).
bonds to correspond to the true β-diimine tautomer, and most
bond lengths and angles are within the expected ranges for such
a situation. However, at the two carbons which have undergone
the sp2/sp3 transition, C(3/43) and C(24/64), there are some
deviations. The angle C(2/42)–C(3/43)–C(4/44) lies, at 115Њ,
almost precisely between the sp2 and the sp3 angle. More signifi-
cantly, the newly formed bond length C(3/43)–C(24/64) was
1.644(2) Å averaged over the two monomers. While this is sig-
nificantly longer than the standard C–C distance of 1.52 Å, it is
not among the longest of such bonds.36 Furthermore, the C–O
bond was short (1.35 and 1.36 Å) in comparison to those from
the analogous 2-cumyl adamantanol and a tantalum-ligated
adamantan-alkoxide (both 1.44 Å).37 Given the long C–C and
short C–O bonds, it is tempting to view the structure of 4 as a
transition state analogue (TS, Scheme 5) for the aldol addition.
However, computations on lithium aldols, and aza-versions38
more similar to the situation in 4, all indicate a very early transi-
tion state, with C–C distances of 2.2–2.5 Å. The only previously
characterizaed lithium aldolate displayed normal C–C and
C–O bond lengths in the tetrameric aggregated structure it
assumed, such that it represented the terminus of the aldol add-
ition reaction.39 Clearly, 4 is not a transition state analogue, but
a ground state, albeit one in which the steric bulk of each
reagent and the electronic stability of the 6π-planar delocalized
reagent 1,5-diazapentadienyllithium has compressed the reac-
tion co-ordinate to the extent that the terminus of the reaction
Inspection of the detailed structural parameters (Table 5)
reveals why this may be occurring. Addition of the electrophile
to the diazapentadienyl unit resulted in localization of the
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1091

View Article Online
Scheme 5
is separated from the reagents by an extraordinarily small
barrier.
Aside from the only other crystallized lithium aldolate,39
anhydrous ammonium salts under argon, also resulted in the
identification only of uncoupled starting materials. This may
be because protonation, even in the solid state, destabilizes a
structure held together only by the co-ordinative demands of
the lithium cations, or it may mean that the retro-aldol reaction
on the alcohol occurred upon dissolution for NMR analysis.
Silylation did not proceed in the solid state or in solution.
The main factor in the equilibrium is clearly the transition
from the solid to the solution state: the solid state favours
higher aggregation states, and dimerization favours C–C
coupling in this case; the greater entropy in solution reverses the
process.
It should be recognized that the analysis of the structure
of intermediates in lithium-mediated syntheses is infrequent.
Had this not been done, neither the solution spectroscopic
evidence nor the quenched product analysis would have
indicated that any reaction had occurred. This leaves open the
question of whether such phenomena may be common in
reactions deemed to have ‘failed’. In cases of aldol reactions
known to proceed, but with poor yields, carrying out quenches
in the solid state may offer a means of improved product
recovery. It has not escaped our attention that aldol additions
are one of the most widely studied and employed methods of
C–C bond formation,35 and so this first demonstration of
the importance of physical phase in the aldol addition merits
further study.
there also exists structural data for a sodium enolate complexed
by unenolized ketone.40 Complex 4 represents a data point some
way along the reaction co-ordinate for carbanion-addition to
ketone, between these two extremes. It clearly lies in a shallow
potential well, since dissolution in non-co-ordinating, non-
polar solvents reverses its formation, an event likely to be under
entropic control. We postulate that the transfer from solution
monomer to dimer upon crystallization is intimately connected
to the aldol addition reaction in this case. The monomer species
2cؒAd, with a 3-co-ordinate planar lithium, has ample prece-
dent in the thf and diethyl ether complexes of 2a and 2b.21,6
Inspection of Fig. 4 reveals that the shortest Li–O bonds lie
almost in the Li–N–Li plane; the lithium atom Li(1) lies only
0.18 Å above the plane of N(41), N(42) and O(1) (0.20 Å in the
other monomer) such that little more than re-hybridization of
C(2/42) and C(24/64) to sp2 and rupture of the longest/weakest
Li–O and C–C bonds, would generate a plausible solution-
phase monomer with regained delocalization-derived stability
and increased entropy. In both directions of this equilibrium,
the putative intermediate would be a dimer (2cؒAd)2 with ter-
minal diazapentadienyls and bridging carbonyls (see Scheme 5).
A similar structure is known where hexamethylphosphoramide
acts as the bridge.24 Furthermore, a computational study con-
firmed the capacity of carbonyls to bridge in lithium dimers.41
In fact, a primary motivation for this experiment was the
isolation of just such a model for the precursor of ketone
enolization by amidolithium bases, in support of the hypothesis
that such reactions proceed via bridging ketones.41 While the
neutral bridging ketone motif has yet to be seen in lithium
chemistry, it has precedent for sodium.42 As to the reason why
aldol addition occurs at all for this resonance-stabilized carb-
anion, upon co-ordination of two lithium ions to the carbonyl
oxygens, the greater polarization of the carbonyl41 facilitates
carbanion attack. Dimerization would also push the two react-
ants together in the correct fashion, whereas in the monomer,
close approach is unlikely. Given that the only other character-
ized lithium aldolate is a tetramer, with more normal C–C and
C–O bond lengths and three lithiums attached to each alkoxide
oxygen,39 the aggregation phenomenon appears to play a major
role in the positon of the aldol/retro-aldol equilibrium.43
Conclusion
Direct acid catalysed, or TiCl4-mediated, condensation was
employed to extend the range of diketiminate (diazapenta-
dienyl) ligands available to co-ordination chemists with cases
of variable bulk, electron donicity and functionality. Double
electrophilic alkylation proceeded smoothly only for diaryl-
diketiminates with significant ortho bulk, which serves to
protect the nitrogen atoms from side reactions. This was estab-
lished by structural characterization of the lithiated inter-
mediates. Simple bulky diaryl β-diimines protected from α-C
reactivity result, the co-ordination chemistry of which will be
discussed in comparison with otherwise identical α-diimines
in a future paper. Where adamantanone was used as the
electrophile on the lithium diazapentadienyl complex, the
reaction proceeded in an aldol-like manner, but did not go to
completion, resulting in long C–C and short C–O bonds to the
adamantanone. The product undergoes retro-aldol dissociation
upon dissolution in even the least polar of solvents.
Lithium alkoxides, and their close relatives, lithium enolates,
are known to exist in the solid state and in solution frequently
as tetramers and hexamers.40 Dimers are also not unusual, but
are normally accompanied by extreme bulk and/or ancillary
co-ordinating solvent ligands.44 Complex 4 is unusual in
employing intramolecular imine nitrogen co-ordination in
completing the lithium’s co-ordination sphere, though a recent
result almost duplicates this feature: The samarium-mediated
(irreversible) reductive coupling of an α-diimine with benzo-
phenone also gives a tridentate N,N,O ligand, though one of the
imine nitrogens is reduced and carries an acidic proton.45
Acknowledgements
Thanks are due to Dr D. A. Middleton for practical assistance
in collection of solid state NMR data, and to Mr I. Crossley for
collection of Raman scattering data. EPSRC (R. J. W., J. E. W.)
and the European Social Fund (E. K. C.-E.) are thanked
for studentship funding. Creators Ltd is thanked for Total
Technology funding (J. E. W.). EPSRC are also thanked for
equipment grants for NMR, crystallography and vibrational
spectroscopy.
Aqueous quenches on solutions of 4 produced recovered,
uncoupled starting materials 1 and adamantanone, a result
entirely consistent with the solution-state spectroscopic results.
However, our attempts at solid-state quench, by grinding with
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1092

View Article Online
24 F. S. Mair, D. Scully, A. J. Edwards, P. R. Raithby and R. Snaith,
References and notes
Polyhedron, 1995, 14, 2397.
1 J. E. Parks and R. H. Holm, Inorg. Chem., 1968, 7, 1408.
2 A rececent review: L. Bourget-Merle, Michael F. Lappert and
J. R. Severn, Chem. Rev., 2002, 102, 3031.
3 J. Feldman, S. J. McLain, A. Parthasarathy, W. J. Marshall,
J. C. Calabrese and S. D. Arthur, Organometallics, 1997, 16,
1514.
25 A. M. Neculai, H. W. Roesky, D. Neculai and J. Magull,
Organometallics, 2001, 20, 5501.
26 P. B. Hitchcock, M. F. Lappert and D.-S. Liu, J. Chem. Soc., Chem.
Commun., 1994, 1699.
27 While in non-D2h lithium dimers the transoid form is generally
favoured for steric reasons, there are some cisoid examples, e.g.:
W. Clegg, L. Horsburgh, S. T. Liddle, F. M. Mackenzie and
R. E. Mulvey, J. Chem. Soc., Dalton Trans., 2000, 1225.
28 C. E. Radzewich, I. A. Guzei and R. F. Jordan, J. Am. Chem. Soc.,
1999, 121, 8673.
29 D. J. Tempel, L. K. Johnson, R. L. Huff, P. S. White and
M. Brookhart, J. Am. Chem. Soc., 2000, 122, 6686 and references
therein.
4 W. Clegg, E. K. Cope, A. J. Edwards and F. S. Mair, Inorg. Chem.,
1998, 37, 2317.
5 At the time of writing there were 111 entries in the Cambridge
Crystallographic Database containing ligand 1a alone. Some leading
references: B. Qian, D. L. Ward and M. R. Smith, Organometallics,
1998, 17, 3070; V. C. Gibson, P. J. Maddox, C. Newton, C. Redshaw,
G. A. Solan, A. J. P. White and D. J. Williams, Chem. Commun.,
1998, 651; P. H. M. Budzelaar, R. de Gelder and A. W. Gal,
Organometallics, 1998, 17, 4121; W.-K. Kim, M. J. Fevola, L. M.
Liable-Sands, A. L. Rheingold and K. H. Theopold, Organo-
metallics, 1998, 17, 4541; M. Chen, E. B. Lobkovsky and G. W.
Coates, J. Am. Chem. Soc., 1998, 120, 11018; L. W. M. Lee, W. E.
Piers, M. R. J. Elsegood, W. Clegg and M. Parvez, Organometallics,
1999, 18, 2947; B. Qian, W. J. Scanlon, M. R. Smith III and D. H.
Motry, Organometallics, 1999, 18, 1693; B. Qian, S. W. Bach and
M. R. Smith, Polyhedron, 1999, 18, 2405; P. L. Holland and
W. B. Tolman, J. Am. Chem. Soc., 1999, 121, 7270; P. L. Holland
and W. B. Tolman, J. Am. Chem. Soc., 2000, 122, 6331; P. J. Bailey,
C. M. E. Dick, S. Fabre and S. Parsons, J. Chem. Soc., Dalton Trans.,
2000, 1655; C. Cui, H. W. Roesky, H. Hao, H.-G. Schmidt and
M. Noltemeyer, Angew. Chem., Int. Ed., 2000, 39, 1815; D. J. E.
Spencer, N. W. Aboelella, A. M. Reynolds, P. L. Holland and
W. B. Tolman, J. Am. Chem. Soc., 2002, 124, 2108.
30 R. Todoriki, M. Ono and S. Tamura, Chem. Pharm. Bull., 1983, 31,
4277.
31 P. B. Hitchcock, M. F. Lappert and D.-S. Liu, J. Chem. Soc., Chem.
Commun., 1994, 2637.
32 E. K. Cope-Eatough, F. S. Mair, R. G. Pritchard, J. E. Warren and
R. J. Woods, Polyhedron, 2003, in press.
33 K. Isele, V. Broughton, C. J. Matthews, A. F. Williams,
G. Bernardinelli, P. Franz and S. Decurtins, J. Chem. Soc., Dalton
Trans., 2002, 3899.
34 S. Trofimenko, Scorpionates – The co-ordination Chemistry of the
Polypyrazolylborate Ligands, Imperial College Press, London, 1999.
35 H. O. House, D. S. Crumrine, A. Y. Teranishi and H. D. Olmstead,
J. Am. Chem. Soc., 1973, 95, 3310; W. A. Kleschick, C. T. Buse and
C. H. Heathcock, J. Am. Chem. Soc., 1977, 99, 247; C. H.
Heathcock, C. T. Buse, W. A. Kleschick, M. C. Pirrung, J. E. Sohn
and J. Lampe, J. Org. Chem., 1980, 45, 1066; C. H. Heathcock,
Science, 1981, 214, 395; J. L. G. Ruano, D. Barros, M. C. Maestro,
A. M. Z. Slawin and P. C. B. Page, J. Org. Chem., 2000, 65, 6027;
R. W. Hoffmann, G. Mas and T. Brandl, Eur. J. Org. Chem., 2002,
3455; I. Paterson, S. B. Blakey and J. C. Cowden, Tetrahedron Lett.,
2002, 43, 6005; C. Palomo, M. Oiarbide and J. M. Garcia, Chem.
Eur. J., 2002, 8, 37.
6 P. H. M. Budzelaar, A. B. van Oort and A. G. Orpen, Eur. J. Inorg.
Chem., 1998, 1485.
7 B. A. Jazdzewski, P. L. Holland, M. Pink, V. G. Young, Jr., D. J. E.
Spencer and W. B. Tolman, Inorg. Chem., 2001, 40, 6097.
8 S. Yokota, Y. Tachi, N. Nishiwaki, M. Ariga and S. Itoh, Inorg.
Chem., 2001, 40, 5316.
9 U. Fekl, W. Kaminski and K. I. Goldberg, J. Am. Chem. Soc., 2001,
123, 6423.
10 L.-J. Ball, A. P. Dickie, F. S. Mair, D. A. Middleton and R. G.
Pritchard, Chem. Commun., 2003, DOI: 10.1039/b211745b.
11 A. F. Burchatt, J. M. Chong and N. Nielsen, J. Organomet. Chem.,
1997, 542, 281.
12 M. Cheng, D. R. Moore, J. J. Reczek, B. M. Chamberlain,
E. B. Lobkovsky and G. W. Coates, J. Am. Chem. Soc., 2001, 123,
8738.
13 COLLECT, data collection software, Bruker-Nonius B.V., Delft,
The Netherlands, 1999.
36 Bonds of 1.7–1.8 Å can be found in some polycyclic compounds:
J. D. Dunitz, X-ray Analysis and the Structure of Organic Molecules,
Cornell University Press, Ithaca, 1979, p. 341. In a cumyl
adamantanol of similar steric proportions, the equivalent length is
1.61 Å, see ref. 37.
37 H. Choi, A. Pinkerton and J. L. Fry, J. Chem. Soc., Chem. Commun.,
1987, 225; H. C. Strauch, R. Roesmann, S. Pacheiner, G. Erker,
R. Fröhlich and B. Wibbeling, J. Organomet. Chem., 1999, 584, 318.
38 R. Glaser, C. M. Hadad, K. B. Wiberg and A. Streitwieser, J. Org.
Chem., 1991, 56, 6625; M. Ruiz, V. Ojea and J. M. Quintela,
Tetrahedron: Asymmetry, 2002, 13, 1863.
39 P. G. Williard and J. M. Salvino, Tetrahedron Lett., 1985, 26, 3931.
40 P. G. Williard and G. B. Carpenter, J. Am. Chem. Soc., 1986, 108,
462.
41 J. M. Hayes, J. C. Greer and F. S. Mair, New J. Chem., 2001, 25, 262.
42 H. Krautscheid, J.-F. Lekieffre and J. Besinger, Z. Anorg. Allg.
Chem., 1996, 622, 1781.
43 There is a variety of other circumstancial evidence pointing to the
importance of aggregated forms in lithium aldol chemistry, e.g.:
J. H. Horner, M. Vera and J. B. Grutzner, J. Org. Chem., 1986, 51,
4212; E. Juaristi, A. K. Beck, J. Hansen, T. Matt, T. Mukhpadhyay,
M. Simson and D. Seebach, Synthesis, 1993, 1271.
44 G. Beck, P. B. Hitchcock, M. F. Lappert and I. A. MacKinnon,
J. Chem. Soc., Chem. Commun., 1989, 1312; W. Clegg, E. Lamb,
S. T. Liddle, R. Snaith and A. E. H. Wheatley, J. Organomet. Chem.,
1999, 573, 305; T. J Boyle, D. M. Pedrotty, T. M. Alam, S. C. Vick
and M. A. Rodriguez, Inorg. Chem., 2000, 39, 5133.
45 J. Scholz, H. Görlz, H. Schumann and R. Weimann, Organo-
metallics, 2001, 20, 4394.
46 M. N. Burnett and C. K. Johnson, ORTEP-III: Oak Ridge Thermal
Ellipsoid Plot Program for Crystal Structure Illustrations, Report
ORNL-6895, Oak Ridge National Laboratory, Oak Ridge, TN,
USA, 1996; L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.
14 Z. Otwinowski and W. Minor, Methods Enzymol., 1996, 276,
307.
15 R. H. Blessing, Acta Crystallogr., Sect. A, 1995, 51, 33.
16 G. M. Sheldrick, SHELX97, Programs for Crystal Structure
Analysis (Release 97–2), University of Göttingen, Germany, 1998.
17 A. Altomare, G. Cascarano, C. Giacovazzo and A. Guagliardi,
J. Appl. Crystallogr., 1993, 26, 343.
18 S. Fustero, M. G. de la Torre, B. Pina and A. S. Fuentes, J. Org.
Chem., 1999, 64, 5551.
19 A. J. Nielson, Inorg. Chim. Acta, 1988, 154, 177; C. H. Zambarro,
R. D. Profilet, J. E. Hill, P. E. Fenwick and I. P. Rothwell,
Polyhedron, 1993, 12, 689.
20 W. Clegg, S. J. Coles, E. K. Cope and F. S. Mair, Angew. Chem., Int.
Ed., 1998, 37, 796.
21 M. Stender, R. J. Wright, B. E. Eichler, J. Prust, M. M. Olmstead,
H. W. Roesky and P. P. Power, J. Chem. Soc., Dalton Trans., 2001,
3465.
22 P. J. Bailey, S. T. Liddle and S. Parsons, Acta Crystallogr., Sect. E,
2001, 57, 863.
23 ‘Squeeze’ was used to account for the electron density of the
disordered solvent in the channel. A. L. Spek, PLATON, University
of Utrecht, The Netherlands, 2001.
D a l t o n T r a n s . , 2 0 0 3 , 1 0 8 3 – 1 0 9 3
1093