Light group 13 chloride diazadiene complexes: Consequences of varying substituent bulk

Article abstract of DOI:10.1039/b712911d

BCl₃ cyclizes diazadiene (2,6-Prⁱ₂C ₆H₃NCH)₂1 through a dichloroborated intermediate [(2,6-Prⁱ₂C₆H₃NCHCl) ₂BCl] 3 to give, in polar aprotic solvents, a spontaneously dehyrochlorinated C-chloro diazaborole 4. In contrast, reaction of AlCl ₃ with 1 forms only acyclic mono- or di-adducts 5a/b and 6. Alkali metal reductions of 3 gave mixtures of 4 and diazaborole [(2,6-Pr ⁱ₂C₆H₃NCH)₂BCl] 7. Pd(0) reduction cleanly gave diazaborole 7. Reduction of 6 gave a low yield of the closed shell C-C coupled dimer 8 of the putative diazadiene radical anion 1?AlCl₂ complex monomer. An alternative synthesis for diazadiene (2,6-Prⁱ₂C₆H₄NCPh)₂, 2, is reported. Reduction of 2/BCl₃, in which additional phenyl groups on the diazadiene C-2 and C-3 atoms hinder the radical coupling observed in 8, gave predominantly diazaborole 2?BCl, (9a) contaminated with 2?BCl₂, (9b) the first such stable radical diazadiene complex of boron. All compounds 2-9 were characterized by X-ray diffraction and NMR spectroscopy. Stable radical 9b was additionally characterized by EPR spectroscopy and density functional computation. The Royal Society of Chemistry 2008.

Full text of DOI:10.1039/b712911d

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Light group 13 chloride diazadiene complexes: consequences of varying
substituent bulk†‡§
Alan Hinchliffe,a Francis S. Mair,*a Eric J. L. McInnes,a Robin G. Pritcharda and John E. Warrenb
Received 24th August 2007, Accepted 1st October 2007
First published as an Advance Article on the web 23rd October 2007
DOI: 10.1039/b712911d
BCl3 cyclizes diazadiene (2,6-Pri2C6H3NCH)2 1 through a dichloroborated intermediate
[(2,6-Pri2C6H3NCHCl)2BCl] 3 to give, in polar aprotic solvents, a spontaneously dehyrochlorinated
C-chloro diazaborole 4. In contrast, reaction of AlCl3 with 1 forms only acyclic mono- or di-adducts
5a/b and 6. Alkali metal reductions of 3 gave mixtures of 4 and diazaborole [(2,6-Pri2C6H3NCH)2BCl]
7. Pd(0) reduction cleanly gave diazaborole 7. Reduction of 6 gave a low yield of the closed shell C–C
coupled dimer 8 of the putative diazadiene radical anion 1·AlCl2 complex monomer. An alternative
synthesis for diazadiene (2,6-Pri2C6H4NCPh)2, 2, is reported. Reduction of 2/BCl3, in which additional
phenyl groups on the diazadiene C-2 and C-3 atoms hinder the radical coupling observed in 8, gave
predominantly diazaborole 2·BCl, (9a) contaminated with 2·BCl2, (9b) the first such stable radical
diazadiene complex of boron. All compounds 2–9 were characterized by X-ray diffraction and NMR
spectroscopy. Stable radical 9b was additionally characterized by EPR spectroscopy and density
functional computation.
preliminary accountofthe unprecedentedcompound 3,7 a result of
imine chloroboration, possessed of remarkably long C–Cl bonds
Introduction
Complexes of ortho-bulky a-diimines (diazabutadienes, DAB)
such as 1 and 2 have enjoyed intense interest from main-group and
transition metal chemists alike since the discovery of Arduengo’s
carbenes.1 Using these ligands, anionic carbene analogues have
already been found for heavier group 13 elements.2,3
˚
(1.84 A). Here we report some reaction chemistry of 3, the product
of addition of BCl3 to 1, the markedly different behaviour of AlCl3
towards 1, the products of reaction of 2 with BCl3, and the products
of metal reduction of these compounds.
Experimental
Anilines and chlorinated solvents were distilled from calcium
hydride. Hexane was distilled from sodium benzophenenone ketyl
with 5% added tetraglyme. Other materials were purchased from
commercial vendors and used as received. Ligand 1 was prepared
by a literature route.8 All manipulations except the workup of
the ligand synthesis were performed under an argon atmosphere
using an argon/vacuum double manifold or argon-filled recir-
culating glovebox equipped with internally-mounted moisture-
and oxygen-scrubbing columns. Details of NMR spectroscopic
equipment and referencing procedures are as described elsewhere.9
X-Band EPR spectra were recorded at 9.5 GHz on a Bruker
continuous wave EPR spectrometer from fluid solution by the
EPSRC national service. Microanalyses were performed by the
UMIST microanalysis service.
We were originally motivated by the greater challenge of
accessing the lower valencies of the lighter group 13 elements
through alkali metal reduction of products of reaction of 1 and
2 with BCl3 or AlCl3. While anions [(DAB)Al]− and [(DAB)B]−
have been predicted by computational methods to be preparable,4
and in the case of boron have very recently proven preparable by
reduction of bromides,5 the chlorides proved remarkably robust
to total alkali metal reduction, in confirmation of findings of
related work by Weber and Schmid on less bulky variants.6
However, we found unexpectedly diverse behaviour in addition
and partial reduction chemistry. We have already communicated a
1,4-(2-isopropylphenyl)-2,3,-diphenyl-1,4-diazadiene 2
In a modification of the general method of Carlson,10 a solution
of TiCl4 (2.0 mL, 18 mmol) in n-hexane (20 mL) was added via
syringe to a rapidly stirred ice-cooled solution of 2-PriC6H4NH2
(15.6 mL, 109 mmol) in n-hexane (100 mL). After strirring for
30 min, the ice bath was removed, and stirring continued until
the mixture reached room-temperature. Powdered PhCOCOPh
(7.7 g, 37 mmol) was then added against a strong argon back-flush.
The temperature of the solution was raised to 50 ◦C, and stirring
was continued for 18 h. The resulting brown slurry was washed
aSchool of Chemistry, The University of Manchester, Manchester, UK M13
9PL. E-mail: mair@manchester.ac.uk
bDaresbury Laboratory, Warrington, Cheshire, UK
† In memory of Dr John H. Morris and his first PhD student, Professor
John Leach: two enthusiasts for boron chemistry and for life.
‡ CCDC reference numbers 658384–658391. For crystallographic data in
CIF or other electronic format see DOI: 10.1039/b712911d
§ Electronic supplementary information (ESI) available: Summary of key
results from Unrestricted-Density Functional computations on 9b and on
models. See DOI: 10.1039/b712911d
222 | Dalton Trans., 2008, 222–233
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
with deionized water and extracted into Et2O. The combined
extracts were dried over MgSO4, filtered, and concentrated by
rotary evaporation to a yellow oil, which solidified upon trituration
with a small volume of n-hexane. The resulting yellow solid was
re-crystallized from n-hexane, isolated by filtration and dried in
air to yield yellow crystals of 2 (10.0 g, 22.5 mmol, 60.8%), m.p.
180 C. H NMR (300 MHz, CDCl3) major isomer: d 7.94 (d,
4H, o-Ph), 7.43 (m, 6H, m-Ph + p-Ph), 6.95 (d, 2H), 6.85 (t, 2H),
6.71 (t, 2H), 6.55 (d, 2H), 2-PriC6H4, 2.47 (septet, 2H, Me2CH),
1.05, 0.65, (two d, 6H each, Me2CHA+B). Minor isomer (12%):
many peaks obscured by major isomer. Observed: 8.13 (broad d,
2H, o-Ph) 7.43 (m, 6H, m-Ph + p-Ph), 6.20 (d, 1H, 2-PriC6H4),
3.07 (septet, 1H, Me2CHA), 2.74 (septet, 1H, Me2CHB), 1.1 (d,
6H, Me2CHA), 0.90 (d, 6H, Me2CHB). 13C NMR (75.47 MHz,
A crystal of 4 was obtained by evaporation of CDCl3 from an
NMR sample of 3. The powder obtained as described matched the
single crystal by comparison of measured powder diffraction and
simulation from single-crystal data (Lazy Pulverix).11 1H NMR
(CDCl3, 300 MHz, 298 K) 7.31 (t, 7.6 Hz, 1H, p-aryl CHa), 7.26
(t, 1H, 7.8 Hz, 1H, p-aryl CHb), 7.17 (d, 7.8 Hz, 2H, m-aryl
◦
1
=
CHb), 7.14 (d, 7.6 Hz, 2H, m-aryl CHa), 6.15 (s, 1H, HC CCl),
2.89 (septet, 2H, 6.82 Hz, HCMe2) 2.79 (septet, 7.07 Hz, 2H,
HCMe2), 1.17 (two d, overlapped, 7.07 Hz, 12H, HCMe2), 1.11
(two d, overlapped, 6.82 Hz, 12H, HCMe2); 11B NMR (CD3CN,
64.2 MHz, 298 K): 21.7 (s); 13C NMR: (CDCl3, 75.47 MHz,
298 K): 147.0, 146.2 (2 × aryl C–N), 135.8, 132.8 (2 × aryl C–
i
=
Pr ), 128.6, 128.1, 123.5, 123.4 (4 × aryl CH), 115.8 (ClC CH),
=
115.0 (ClC CH), 28.75, 28.49 (2 × CHMe2), 24.06, 23.98, 23.87,
=
=
CDCl3): d 162.4 (C N), 146.2 (C–N), 142.7 (CC N), 138.7 (C–
Pri), 131.2, 129.1, 128.7, 126.2, 125.9, 125.6, 117.3 (7 × aryl CH),
28.1 (CHMe2), 23.3, 22.0 (Me2CHA + B). Found: C, 85.4; H, 7.9; N,
6.6%; C32H32N2 requires C, 86.4; H, 7.3, N, 6.3%.
23.43 (4 × CHMe2). Found: C, 68.6; H, 7.9; N, 6.1; Cl, 15.9%;
C26H35N2BCl2 requires C, 68.29; H, 7.71; N, 6.12; Cl, 15.50%.
[g1-N-{E,E-N,Nꢀ-Bis(2,6-diisopropylphenyl)-1,4-
diazadiene}trichloroaluminium(III)] 5a,
N,Nꢀ-Bis(2,6-diisopropylphenyl)-2,4,5-trichloro-1,3,
2-diazaborolidine 3
[g1-N-{E,Z-N,Nꢀ-bis(2,6-diisopropylphenyl)1,4-
diazadiene}trichloroaluminium(III)] 5b, and
[g2-N,Nꢀ-{E,E-N,Nꢀ-bis(2,6-diisopropylphenyl)-1,
4-diazadiene}bis{trichloroaluminium(III)}] 6
This compound has previously been reported in a preliminary
account.7 Some of the data was interpreted incorrectly in that
account; these errors are corrected here. A three-necked 500 mL
round bottomed flask was charged with dry n-hexane (150
mL) and cooled to −30 ◦C. Solutions of 1,4-diaza-1,4-(2,6-
diisopropylphenyl)-butadiene 1 (3.76 g, 10 mmol) in 20 mL n-
hexane and BCl3 (10 mL of a 1 M hexane solution, 10 mmol) were
concurrently added dropwise to the flask. The resulting mixture
was stirred for 1 h at −30 ◦C before being allowed to reach ambient
temperature. The yellow–orange powder obtained was filtered
under argon, washed with n-hexane and vacuum dried to give
3a (1.907 g, 38.6%), m.p. 109–110 ◦C (dec. with gas evolution);
Found: C, 63.05; H, 7.15; N, 5.40; Cl, 21.8%; C26H36N2BCl3
requires C, 63.25; H, 7.35; N, 5.68; Cl, 21.54%.
These three compounds were isolated as single crystals selected
from complex mixtures containing dark tars and orange–red
crystals. None were isolated as pure bulk materials. A typical
procedure is given:
To a vigorously stirred suspension of AlCl3 (0.35 g, 2.64 mmol)
in n-hexane (60 mL) held at −70 ◦C was added a n-hexane solution
of 1 (1.0 g, 2.66 mmol in 60 mL n-hexane), in 10 mL aliquots. An
orange colouration developed rapidly. The solution was allowed
to reach room temperature and stirred for 3 d, during which time
the orange colour deepened. Concentration in vacuo to 50% of
original volume, followed by overnight refrigeration (−20 ◦C),
afforded a crop of large orange crystals of 5/6, which were isolated
by filtration (0.27 g). Further concentration and refrigeration of
the filtrate produced a further 0.51◦g, combined yield 0.78 g, of a
mixture of 5 and 6, m.p. 120–125 C. Selection of crystals from
different batches of 5 for X-ray diffraction gave the structures of
the two isomers, E,E 5a and E,Z 5b, and of the doubly substituted
6. NMR spectra from all batches were identical, all containing
mixtures of 5 and 6 in solution, in approximate molar ratio 55% :
45%. Reported integrals are normalized. 1H NMR (CDCl3,
300 MHz, 298 K) peaks assigned to 5a/5b (isomers not observed at
The filtrate was refrigerated to produce colourless crystals of
3b (1.734 g, 35.1%). Microanalysis and NMR in CDCl3 of 3b
were identical to those of 3a. Single crystal diffraction data were
collected for a crystal of 3b. Powder X-ray diffraction of 3a
matched the Lazy Pulverix-computed pattern11 from single crystal
data for 3b, i.e. 3a = 3b = 3, total yield 73.7%.
Clean NMR spectra were not obtainable for 3, which instead
converted to 4 in all NMR solvents in which it was soluble, vide
infra.
=
=
room temperature): 9.08 (d, 1H, J = 8 Hz, N HC–CH N), 7.66
=
=
(d, 1H, J = 8 Hz, N HC–CH N), 3.01 (septet, 2H, J = 6.8 Hz,
Me2CHa) 2.77 (septet, 2H, J = 6.8 Hz, Me2CHb), 1.34 (d, 6H, J =
6.8 Hz, MeaMebCHa), 1.11 (d, 6H, J = 6.8 Hz, MeaMebCHa), 1.07
(d, 12H, J = 6.8 Hz, Me2CHb); peaks assigned to 6: 8.71 (s, 2H,
N,Nꢀ-Bis(2,6-diisopropylphenyl)-2,4-dichloro-1,3,2-diazaborole 4
A solution of 1 (2.28 g, 6 mmol), in dichloromethane (60 mL) was
chilled to −10 ◦C and stirred during dropwise addition of BCl3
(6.2 mL of a 1.0 M solution in n-hexane). A dark purple solid
precipitated instantaneously. The mixture was allowed to reach
room temperature and stirred for 2 d. Solvent was removed in
vacuo. The remaining purple solid was re-suspended in n-hexane
(60 mL), filtered in vacuo, washed with n-hexane (3 × 5 mL) and
dried in vacuo to yield a lilac powder (0.37 g). Concentration of
the filtrate produced a further 2.28 g of identical lilac powder,
combined yield 2.89 g, 88%) of 4, m.p. 106–108 ◦C. NMR spectra
were identical to those obtained by dissolution of 3 in CDCl3.
=
=
N HC–CH N), 3.04 (septet, 4H, J = 6.8 Hz, Me2CH), 1.27 (d,
12H, J = 6.8 Hz, Me2CH). Peaks not unambiguously assignable
to 5 or 6: 7.12–7.49 (unresolved m, aryl CH). 13C NMR (CDCl3,
=
=
75.47 MHz, 298 K): 5/6 mixture: 175.1 (AlN C), 153.9 (N C),
146.0, 141.1 (aryl C–N), 139.3, 138.2 (aryl C–Pri), 130.0, 128.8,
128.4, 124.9, 124.1, 123.9 (aryl CH), 29.0, 28.7, 28.1 (CH(CH3)2),
24.0, 23.8, 23.6, 23.5, (CH(CH3)2).
Similar reaction of 1 (1.0 g, 2.66 mmol) with two equivalents of
AlCl3 (0.71 g, 5.32 mmol) produced dark red powder, 1.46 g, m.p.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 222–233 | 223
©

View Article Online
158–160 ◦C, with NMR trace enriched in peaks assigned to 6, but
still containing some 5.
cold water-jacketed Schlenk flask containing lithium sand (0.09 g,
12.97 mmol) in hexanes (18 mL). The mixture was stirred for 3
weeks, filtered, the filtrate concentrated in vacuo to half-volume
then refrigerated for 6 d at −20 ◦C, during which time orange–
red crystals formed. These were isolated by filtration to yield 8,
(0.25 g, 10%). m.p. 191–193 C. H NMR (CDCl3, 300 MHz,
298 K): only weak signals attributable to impurities 5/6 were well-
resolved. Broad features centred at 7.23 (aryl C–H), 2.98 (HCMe2)
and 2.24 (Me2CH) were also apparent.
N,Nꢀ-Bis(2,6-diisopropylphenyl)-2-chloro-1,3,2-diazaborole 7
◦
1
Method 1. From 3 and lithium: To a stirred Schlenk flask
containing 1.0 g (2.0 mmol) of 3 and lithium sand (0.05 g,
7.2 mmol) at −10 ◦C, dry n-hexane (40 mL) was added, forming
a grey–tan suspension. After 3 d of stirring at room temperature,
a black precipitate was removed by filtration to yield a pale tan
solution. This was concentrated to 25% of its original volume in
vacuo and stored at −20 ◦C for 3 d to yield large prismatic crystals.
A small sample was collected for X-ray diffraction analysis, which
proved the formulation as 7, vide infra. The remaining solution
was concentrated in vacuo to yield impure crystals of 7 (0.82 g)
contaminated with some 4 and some silicone grease, as determined
by NMR and microanalysis. m.p. 164–168 ◦C. 1H NMR (CDCl3,
300 MHz, 298 K), minor peaks assignable to silicone grease and
4 (vide supra) were observed, in addition to: 7.25 (t, 7.6 Hz, 2H, p-
[N,Nꢀ-Bis(2,6-diisopropylphenyl)-2-dichloro-4,5-diphenyl-1,3,2-
diazaborolium]/[N,Nꢀ-bis(2,6-diisopropylphenyl)-2-chloro-4,5-
diphenyl-1,3,2-diazaborole] 9
To an ice-cooled solution of 2 (0.79 g, 1.78 mmol) in hexanes
(60 mL) and dichloromethane (2 mL) was added BCl3 (1.8 mL
of a 1 M solution in hexanes), producing an orange solution. An
orange suspension formed upon warming to room temperature.
The solvent volume was reduced by half in vacuo, and stirring at
room temperature continued for a further 2 d to yield an orange
powder which was isolated by filtration, washed with hexanes,
and dried to yield 0.77 g of an impure form of borolium salt.
This was used without further purification. Impure borolium salt
(0.3 g, approx. 0.5 mmol) and lithium sand were combined in
a Schlenk tube, then suspended in hexanes (20 mL). Stirring at
room temperature for 5 d gave an olive green–brown suspension.
Filtration gave a pale orange filtrate which was concentrated by
half in vacuo. Standing overnight at room temperature caused
deposition of a few yellow crystals of 2, which were manually
removed, and some fine orange solid. Further concentration
in vacuo, followed by heating to redissolve the orange solid,
followed by slow cooling, furnished a few orange crystals of
9 in low yield. 1H NMR (CDCl3, 300 MHz, 298 K): anti
conformer: 7.38 (d, 2H, NCCHCHCHCHCPri), 7.25 (AB dd, 4H,
NCCHCHCHCHCPri), 7.14 (d, 2H, NCCHCHCHCHCPri),
7.00 (d, 4H, o-Ph), 6.96 (t, 2H, p-Ph), 6.91 (t, 4H, m-Ph), 2.81
(sept. J = 6.8 and 7.1 Hz, 2H, CHMe2), 1.08 (d, J = 7.1 Hz, 6H,
Me2CH) and 0.58 (d, J = 6.8 Hz, 6H, Me2CH); syn conformer:
7.25 (AB dd, 4H, as above), 7.14 (d, 2H, as above, + overlapping d,
2H, NCCHCHCHCHCPri), 7.00, 6.96, 6.91 (Ph, as above), 3.09
(sept., J = 6.8 and 7.1 Hz, 2H Me2CH), 1.19 (d, J = 6.8 Hz, 6H,
Me2CH), 1.00 (d, J = 7.1 Hz, 6H, Me2CH). 13C NMR (298 K,
CDCl3, 75.47 MHz) (both conformers): 146.32, 145.97 137.24,
137.19, 131.83, 131.75, (quaternaryC,{C–N, C–Pri, ipso-C–Ph,} ×
2), 130.80, 129.93, 129.54, 127.85, 127.57, 127.40, 127.37, 126.46,
126.37, 126.02, 125.99, 125.73 (aryl C–H, 2 × PriC6H4, 2 × o-Ph,
m + p-Ph), 27.85, 27.71 (2 × CHMe2), 24.85, 24.24, 22.78, 21.60
(2 × CHMe2). EPR (293 K, CDCl3): g = 2.011, 14-line multiplet,
see discussion.
=
aryl CH), 7.16 (d, 4H, 7.6 Hz, m-aryl CH), 6.16 (s, 2H, HC CH),
2.91 (septet, 6.8 Hz, 4H, HCMe2), 1.12 (two d, overlapped, 6.8 Hz,
24H, HCMe2), 11B NMR (CD3CN, 64.2 MHz, 298 K): 21.5 (s); 13
C
NMR: (CDCl3, 75.47 MHz, 298 K): peaks for 4 (vide supra) plus
146.3 (aryl C–N), 136.7 (aryl C–Pri), 127.8, 123.45 (2 × aryl CH),
=
118.8 (HC CH), 28.46 (CHMe2), 24.07, 23.88 (2 × CHMe2).
Found: C, 64.75; H, 8.00; N, 5.15; Cl, 12.4%; C26H36N2BCl
requires C, 73.85; H, 8.58; N, 6.62; Cl, 8.38%. Compound 7 could
not be isolated in good yield free of 4 by this route.
Method 2. From
3 and Pd(Ph3P)4: A stirred Schlenk
flask containing 3 (0.20 g, 0.41 mmol) and tetrakis(triphenyl-
phosphine)palladium(0) (0.47 g, 0.41 mmol) in toluene (40 mL)
was heated under reflux for 3 h, allowed to cool, then filtered. The
filtrand was washed with cold toluene (3 × 5 mL) yielding a yellow
crystalline solid, [(Ph3P)2PdCl2], identified by crystallography,12
and a white powder. Evacuation of the combined filtrates to
dryness gave a yellow tar which was taken up in dry hexanes
(10 mL) and held at −20 ◦C for 6 d to yield crystals of 7 (0.11 g,
62%). NMR as above, but without contamination with 4. Found:
C, 73.11; H, 8.60; N, 6.37; Cl, 7.95%; C26H36N2BCl requires C,
73.85; H, 8.58; N, 6.62; Cl, 8.38%.
Method 3. From 3 and Na–K alloy: To a Schlenk flask
containing 3 (0.50 g, 1.0 mmol) and Na–K alloy (0.11 g, 1.8 mmol,
3.2 mmol reducing equivalents) and a glass-coated magnetic
stir-bar were added hexanes (20 mL) with rapid stirring to
produce a grey–tan suspension. Stirring at room temperature
was continued for 5 d, after which filtration removed a black
precipitate. Concentration of the filtrate in vacuo to dryness yielded
a tan glass with NMR and microanalysis characteristics as for
Method 1. Recrystallization from cold hexanes gave crystals for
X-ray diffraction, identical to those from Method 2.
Crystallography‡
[g2:g2-N,NꢀNꢀꢀ,Nꢀꢀ-{Bis-1,6-N,Nꢀ-(2,6-diisopropylphenyl)-bis-3,
4-(2,6-diisopropylphenylamido)-1,6-diazahexa-1,5-diene}-
bis{dichloroaluminium(III)}] 8
Data collection for 3 previously communicated.7 Single crystals of
compounds 2 and 4–9 were coated in an inert perfluoropolyether
oil (1800 fomblin), mounted on a glass fibre then placed on
a Nonius Kappa CCD Diffractometer. Data were recorded at
−123 ◦C (Oxford Cryostream) using graphite monochromated
Mo-Karadiation (k = 0.71069) with the CCD detector placed at a
minimum distance of 33 mm from the sample. Data was collected
To a vigorously stirred solution of AlCl3 (0.70 g, 5.25 mmol) in
hexanes (100 mL) was added 1 (2.03 g, 5.39 mmol), under a strong
argon back-flush, instantly producing an orange solution. The
solution was stirred for 0.5 h then transferred by cannula to a
224 | Dalton Trans., 2008, 222–233
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
via a mixture of 1◦ u and x scans at different h and j settings using
the program COLLECT.13 The raw data were processed to produce
conventional data using the program DENZO-SMN.14 The data
sets were corrected for absorption using the program SORTAV.15
All structures were solved by using SIR9216 and were refined
by full matrix least squares refinement using SHELXL-97.17 All
non-hydrogen atoms were refined with anisotropic displacement
parameters. Table 1 lists key structural parameters; those for
5a, a geometric isomer of 5b, are available in the ESI§. Data
collection and refinement information for 2, 4, 5b, 6, 8 and 9a/b
is summarised in Table 2.
Computation
Energy minimisations using Unrestricted Density Functional
B3LYP methods on 9b using GAUSSIAN18 were performed
with 6-311G (d,p) basis sets, starting from the crystallographic
geometry. A full vibrational analysis was done in every case in
order to characterize the stationary points.
Results & discussion
Proligand synthesis
While ligand 1 is conveniently prepared by direct acid-catalysed
condensation,8 we find with benzil-derived diazadienes that the
TiCl4-mediated condensation is advantageous. In this way, good
yields of proligand 2 are obtained. This method has previously
been employed for a wide range of imine syntheses.9,10,19 Ligand 2
had been previously reported using a simple acid-condensation
route,20 as an intermediate to its cyclodehydrogenated dihy-
drophenanthrene form, but no data for 2 were reported in that
paper. We found it necessary to use TiCl4, because in our hands
the simple acid-catalysed route required forcing conditions under
which cyclization to 1,2-dihydro quinoxalines is a major side-
reaction.21 Compound 2 adopts the Z-s-cis-Z conformation in
the solid state (Fig. 1), and its 1H NMR spectrum indicates
that this conformation is retained in solution, in addition to
another minor conformer. As prochiral groups in an environment
lacking Cs symmetry in any conformation, the two methyls are
anisochronous.22
The two isopropyl groups are in an anti disposition in the solid
state. Adoption of a Cs transient conformer is impeded in the
Z,Z conformation, due to unfavourable interactions between aryl
hydrogens from the N-aryl and the a-phenyl (chart 1). In the E
conformation, this restriction is relaxed, as shown in Chart 1,
because analogous interaction between N-aryl and ipso-phenyl
are less-severe. Consequently, the minor component is assigned
an E/Z geometry. A free exchange of E and Z geometries in
imines during and prior to co-ordination is known.23 Observation
of cross-peaks between the sets of doublets in an 2-D EXSY NMR
experiment provided such evidence in this case. These results differ
from those found for the dihydrophenanthrene form, which was
reported to exist mainly as E/Z, with some E/E conformer in
solution.20
Reaction chemistry 1: addition
The normal path of reaction of a group 13 chloride with a
diazadiene is bidentate co-ordination of the diazadiene ligand with
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 222–233 | 225
©

View Article Online
Fig. 1 Crystal and molecular structure of proligand 2. The asymmet-
ric unit comprises one half-molecule; symmetry-equivalent atoms are
labelled with #. Hydrogen atoms are omitted. Thermal ellipsoids are drawn
at 50% probability level.
Chart 1 Steric interactions in possible Cs transient conformations for E
and Z imines.
Scheme 1
expulsion of a halide from the co-ordination sphere, Scheme 1,
path 1.6a–e
This results, in the case of boron, in a generation of diaz-
aborolium salts. Use of a further mole of Lewis acid AX3 can
226 | Dalton Trans., 2008, 222–233
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
−
promote isolation of clean, crystalline products such as AX4
observed spectrum. That is to say, in solution in solvents more
polar than the hexane from which the crystals were isolated,
elimination of HCl occurs. This strange hypothesis was proved
by isolation of crystals of 4 from an NMR tube CDCl3 sample
of 3. Furthermore, crystals of compound 4, the spontaneously
reduced diazaborole, were directly preparable by combination of 1
with BCl3 in the more polar solvent dichloromethane (Fig. 3). This
also explains the anomalous chemical shift of the ‘diazaborolidine’
ring hydrogen, since after dissolution for NMR analysis it is in
fact an alkenyl hydrogen from a diazaborole. We have no data to
inform mechanistic speculation on this curious reaction. We note,
however, that the C–Cl bonds in 3 are amongst the longest known
(Fig. 1), and that the transoid disposition of the chlorines makes
1,2-elimination of HCl a favourable process. While the compound
4 is among the first diazaboroles to possess halo substitution on
the ring,6i there is an interesting parallel to be drawn between
the reactivity of 1 with BCl3 and that of DAB ligands with PCl3,
which can also result in ring C-monochlorination.27 Further close
comparisons can be drawn with recently published work from the
group of Weber, which described the variety of products obtained
from reaction of BCl3 with mesityl and xylyl analogues of 1.6i His
findings have many parallels with our own, including isolation of
spontaneously reduced, mono-C-chlorinated diazaboroles, among
a variety of other products.6i
salts, as recently demonstrated crystallographically.6i,24 In contrast,
reaction of 1 with BCl3 at high dilution in hexane resulted
instead in highly diastereoselective addition of B–Cl across both
C N bonds to yield 3, Fig. 2. Such chloroboration of imines
7
=
(path 2) is without precedent, though chloroboration of other
unsaturated functions, such as isocyanates, isothiocyanates,25 and
alkynes26 is well-established. We originally hypothesised that this
reaction may have been caused by the reluctance of the boron
atom to entertain four-co-ordination in the presence of such
bulky aryl substituents, since it was the bulk on the aryls which
distinguished 3 from the borolium formulations of prior workers.6
However, the 2,6-diisopropylphenyl substituent has since been
found in a very similar environment in a dichlorodiazaborolium
salt isolated by Clyburne.24 In that case, a planar, p-delocalized
backbone on the diazabutadiene precluded the saturating attack of
chloride, hence more conventional borolium behaviour was found.
Furthermore, recent results from Weber’s group suggest that the
behaviour is not restricted to cases of extreme steric bulk, since
C-chlorinated compounds were obtained from reaction of BCl3
with N-mesityl analogues of 1.6i Rather, the N-aryl substituent is
the distinguishing factor, since such behaviour is not observed with
N-alkyl-substituted diazadienes. All synthetic transformations are
summarised in Scheme 2.
Fig. 2 Crystal and molecular structure of 3. Previously communicated.7
Bond lengths (A): Cl(3)–B(1) 1.743(3); N(1)–B(1) 1.403(3); N(2)–B(1)
Fig. 3 Crystal and molecular structure of 4. A crystallographic plane
lies along B1–Cl1. Symmetry-equivalent atoms are denoted by #. There is
substitutional disorder on C2: the C2-substituent is 50% Cl, 50% H. The
Cl2 only is shown on C2, the H2# only on C2#. Other hydrogen atoms are
omitted. Thermal ellipsoids are drawn at 50% probability level.
˚
1.416(3); N(1)–C(1) 1.413(3); N(2)–C(2) 1.417(3); C(1)–C(2)
1.528(3); C(1)–Cl(2) 1.840(3); C(2)–Cl(1) 1.841(3). Bond Angles (◦):
B(1)–N(1)–C(1) 109.10(19); B(1)–N(1)–C(15) 128.9(2); C(1)–N(1)–C(15)
121.98(19); B(1)–N(2)–C(2) 108.82(19); B(1)–N(2)–C(3) 129.4(2);
C(3)–N(2)–C(2) 121.78(19); N(1)–B(1)–N(2) 109.4(2); N(1)–B(1)–Cl(3)
125.5(2); N(2)–B(1)–Cl(3) 125.1(2); N(1)–C(1)–C(2) 105.45(19);
N(1)–C(1)–Cl(2) 112.93(17); C(2)–C(1)–Cl(2) 107.26(16); N(2)–C(2)–C(1)
105.05(19); N(2)–C(2)–Cl(1) 112.85(17); N(2)–C(2)–Cl(1) 112.85(17);
C(1)–C(2)–Cl(1) 107.51(17). Diazaborinane ring hydrogen atoms are
shown, others are omitted.
Chloro substitution has a minimal effect on metrical parameters
(see Table 1), but does offer a path for further synthetic elaboration
=
of the ligand. Indeed, in Weber’s case, some C C coupled
bis(diazadienyl) complexes were isolated as byproducts.6i We find
similar, but not identical, behaviour for aluminium, vide infra.
AlCl3 behaves quite differently from BCl3 towards 1, mirroring
more closely the behaviour of BF3 with the less bulky mesityl
analogues of 1 as reported by Weber.6d Among the mixture of
products from 1 : 1 reactions of AlCl3 with 1, simple adducts 5
were found, i.e., path 3. Crystals of both E/E (5a) and E/Z (5b)
isomers were isolated. The structure of the adduct 5a can be found
in the ESI; that of 5b is presented in Fig. 4.
In the original communication,7 we erroneously interpreted the
multiple isopropyl methine resonances in the 1H NMR spectrum
of 3 as being due to restricted aryl rotation, whereby one isopropyl
group on the same side of the diazaborolidine ring as the chlorine
substituent gave rise to one signal, while the isopropyl from the
other side of the same phenyl ring gave rise to the other signal.
It is now clear that two inequivalent phenyl groups, both with
chemically equivalent 2- and 6-isopropyl groups, produce the
Adduct 5a bears the closest resemblance to that of the free ligand
1, which also crystallizes in an E/E conformation.28 It is the less
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 222–233 | 227
©

View Article Online
Scheme 2
bond lengths being closely similar to those found for the free ligand
1. Bond lengths to the co-ordination nitrogen N1 increase slightly,
as expected for an atom with increased co-ordination number. The
geometry around Al1 is that of a compressed tetrahedron, the Cl–
Al–Cl angles being systematically several degrees larger than the
Cl–Al–N ones (Table 1). This is common to related adducts of
AlCl3 with single monodentate nitrogen ligands.29
When a further molar equivalent of AlCl3 co-ordinated, gener-
ating circumstances where abstraction of chloride by one molecule
of AlCl3 might have caused cyclization as found for related boron
systems,6d,24 instead the 2 : 1 adduct 6 was produced (Fig. 5), again
mirroring Weber’s results with BF3. This is in stark contrast to the
situation for BCl3, and, less surprisingly, in contrast to the findings
of Jones and co-workers for AlI3.30 For both BF3 and AlCl3, the
bonds resisting cleavage are between atoms of the same period. It
remains possible that other non-crystalline products which contain
diazadiene metallacycles exist in the complex mixtures produced
by these reactions. Indeed, Weber’s recent results on xylyl and
mesityl analogues of 1 with BCl3 display a wide array of reactivity,
depending on temperature and reaction solvent, including cases
where adducts analogous to 5 were postulated as low-temperature
intermediates.6i
Fig. 4 Crystal and molecular structure of E/Z isomer 5b. Diazabutadiene
hydrogens H2 and H3 are shown, others are omitted.
densely packed of the two isomers of 5. Adduct 5b has isomerized
at the co-ordinated nitrogen (E/Z), but otherwise is similar: Both
have a non-crystallographic plane of symmetry running through
Al1 and Cl2 as well as N1, C2, C3, N4, C10 and C40. The aryls
are almost orthogonal to this diazabutadiene conjugated plane
(aryl–diazabutadiene plane angles of 88.3 and 89.5◦ in 5a, for
example). Co-ordination of AlCl3 to the diazadiene 1 in either
conformation had a negligible effect on conjugation, the ligand
The configuration at both imine bonds in 6 is Z. It seems
probable that this minimises steric conflict between the AlCl3
and diisopropylphenyl moieties. It does not favour formation
228 | Dalton Trans., 2008, 222–233
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
in preparation, it was reported by Nozaki that use of BBr3 in
place of BCl3 allows full reduction to B(I) with Li naphthalide,
even for prototype diazadiene 1.5 Some early attempts by us
to use 4,4ꢀ-bis-But-biphenyl as a catalyst of lithium reduction
were plagued with problems of contamination with the lithium
biphenyl, while Nozaki was able to remove the less-soluble excess
lithium naphthalenide by filtration.5
The structure of 7 is shown in Fig. 6. It has metrical parameters
in accordance with those of other N,Nꢀ-diaryl-diazaboroles,6d,g,i
the shortening of all the bonds within the diazaborole ring upon
reduction of 3 to form 7 being evident. The planarity of the
diazaborole ring is not shared by N,Nꢀ-dialkyl analogues.6c
Fig. 5 Crystal and molecular structure of 6. The asymmetric unit
comprises one half-molecule; symmetry-equivalent atoms are labelled with
#. Diazabutadiene hydrogens are shown, others are omitted. Thermal
ellipsoids are drawn at 50% probability level.
of metallacycles. However, rapid E/Z equilibration has been
previously noted in aryl imines, and is not seriously inhibited by
complexation with group 13 Lewis acids.23
NMR spectra of the crystalline products from a 1 : 1 reaction
of 1 with AlCl3 showed 5 and 6 in an approximate 1 : 0.8 ratio.
Two isomers of 5 were not detectable. It was clear from the pair
of doublets for the aldimine hydrogens at 9.1 and 7.6 ppm that
the major component was 5. There was also doubling of all other
peaks. The minor component exhibited a single aldimine hydrogen
resonance at 8.7 ppm, shifted from the analogous peak in free
ligand 1 at 8.0 ppm. This, and corresponding single groups of
resonances from the isopropyl groups, suggested the presence of 6
in samples of 5. The peaks assigned to 6 increased when a further
molar equivalent of AlCl3 was used in the synthesis, strengthening
confidence in assignments, but pure bulk samples of neither 5 nor
6 proved isolable.
Fig. 6 Crystal and molecular structure of diazaborole 7. Diazaborole ring
hydrogens are shown, others are omitted.
Lithium reduction of 6 proceeded slowly, but after three weeks
of stirring, refrigeration produced a small yield of complex 8,
Fig. 7.
For the bulkier ligand 2, reaction with BCl3 is presumed to
proceed along path 1, given that the increased bulk on the
diazadiene carbon atoms would be expected to prevent the attack
of chloride. Only insoluble powders were obtained, which provided
neither satisfactory analysis figures nor readily interpretable
spectra, so samples of the presumed borolium salt were used
directly in attempted reductions without full characterization.
Reaction chemistry 2: reduction
If 3 is reacted with [Pd(PPh3)4], Na–K alloy or Li sand, C-
dechlorination to produce the diazaborole 7 occurs, as for anal-
ogous, less bulky systems.6 Thus, a C-chlorinated or C-unchlori-
nated diazaborole is available by either spontaneous HCl ejection
or chemical reaction with a variety of reductants, respectively. The
cleanest product was obtained using the Pd reductant. Interest-
ingly, Pd/C was used in the earliest preparations of diazaboroles
by dehydrogenation of diazaborolines.31 Where alkali reductants
were employed, some contamination with 4 was evident. In
confirmation of the results of Weber, in no case was further
reduction to B(II) or B(I) species achieved via chlorides.6a–e Weber
has accessed B(II) compounds via reduction of chalcogenides,6f
and Morris has reduced chlorides of analogous tetrazaboroles to
diborane(4) derivatives with Na–Hg.32 While this manuscript was
Fig. 7 Crystal and molecular structure of 8. Amido-imine ring hydrogens
are shown, others are omitted. The asymmetric unit comprises one
half-molecule; symmetry-equivalent atoms are labelled with #. Thermal
ellipsoids are drawn at 50% probability level.
While reduction succeeds in closing the diazaaluminolium ring,
the open-shell AlCl2 complex of a hemi-reduced diazabutadiene
appears to be unstable. This is surprising in view of the fact that an
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 222–233 | 229
©

View Article Online
AlI2 complex of the same diazabutadiene radical anion has been
spectroscopically and crystallographically characterized.30 In the
case of AlCl2, the open-shell species undergoes C–C coupling to
produce the linked, closed-shell complex 8. The coupling happens
in such a way as to generate meso diamides. The crystallographic
from ligand 2. When it was reduced with lithium sand, the
reaction proceeded as far as diazaborole 9a, but crystals showed
shadow positions above and below the single remaining chlorine
substituent whose 14-line regular multiplet solution phase X-band
EPR spectrum indicates approximately equal couplings to two 14N,
one 11B and two 35Cl atoms. This indicated that approximately 11%
of lattice points in the crystal were occupied not with the closed
shell diazaborole 9a, but with the less-reduced open-shell 9b, a
neutral complex of a diazadiene radical anion with (BCl2) cation.
Both 9a and 9b are shown as a composite structure in Fig. 8. The
compound possesses a crystallographic C2 axis along the B(1)–
Cl(1a) bond vector. This dictates an anti arrangement of isopropyl
groups. The imine-aryls lie at an angle of 72.5◦ with respect to the
central diazachloroborole plane indicating limited delocalization,
but the C-aryls lie at an angle of 44.4◦ to the central plane,
suggesting some stabilization of spin density through partial p-
delocalization from the diazadiene. While the open shell 9b might
be expected to exhibit longer C–C and shorter C–N bonds than
the more fully reduced 9a, there is no evidence of split sites, or even
of distorted ellipsoids, for C(2). The changes are so slight as to be
undetectable. The bond lengths found are comparable with those
from other known diazaboroles.6 Fortunately, the presence of some
paramagnetism in the sample, while allowing EPR to be observed
for 9b, did not hamper the observation of NMR data for 9a.
This revealed, when recorded immediately after dissolution, major
peaks assigned to the anti conformation exhibited in the crystal,
and some additional small peaks assigned to a syn conformation of
isopropyl groups. On standing for 100 min, the populations of both
conformations became essentially equal, and remained constant
thereafter, indicating a negligible energy difference between them
in the solution state, and a relatively slow fluxion, with no exchange
broadening.
#
¯
centre of the P1 space group lies at the midpoint of the C3–C3
bond. This coupling serves to localise the p-system, such that
the ligand is now best described as a bis(a-amino-imine). The two
nitrogen–aluminium interactions, which might be termed covalent
and dative, are distinguished by differing bond lengths, 1.82 and
˚
1.95 A, respectively. A notable feature is the pyramidalization of
˚
the amido-nitrogen. N1 lies 0.197 A above the plane defined by
its three substituents. Most amidoaluminum complexes have the
amido groups in bridging positions; we recently published data
which allows a close comparison, a b-amidoalane with a terminal
amide where the organic substituent on the amide was identical
(2,6 diisopropylphenyl), and the aluminium was co-ordinated also
by a neutral imine donor and two hydride or chloride ligands. In
that case, the pyramidalization was greatly reduced (distance of
33
˚
N above plane defined by its three substituents: 0.013 A). The
pyramidalization in 8 may therefore be the consequence of steric
repulsion between the two halves of the dimeric species, since the
distortion observed maximised this separation. Its NMR spectra
were very broad, suggesting the presence of paramagnetism in the
sample. No NMR data have been reported for the similar AlI2
complex.30
This coupling of diazabutadienes has been seen before in zinc
chemistry,34 and recently in a pyridyl imine on gallium,35 but
this is the first observation of the behaviour involving the very
bulky ligand 1. The bulk would have been expected to help to
stabilize the radical, as it did for 1·AlI2.30 The differing sterics of
Cl vs. I would appear to be too remote from the site of coupling
to have a bearing on the relative stability of open and closed-
shell species. Quantum-chemical computations were employed to
probe the electronic aspects of this difference. These indicated
that substitution of iodide for chloride had a negligible effect
on spin density on the carbon atoms involved in the coupling
(see ESI§). The electronic structure calculations confirmed the
established view of such compounds as AlX2+ complexes of ligand-
based radical anions.36 The remaining possibilities are (i) that
there is a negligible energy difference between C–C coupled closed
shell and radical complexes, and better packing in the solid state
favours the formation of the closed shell dimeric species 8 while
in solution it is in equilibrium with open-shell monomer, or (ii),
that differing synthetic routes (the iodide case was a product
of in situ reduction with Al powder)30 caused the chloride and
iodide reactions to follow different courses. We favour the former
explanation, since we recently showed in a related adduct of
a similarly bulky diketiminate-ketone complex that a solution-
promoted bond scission equilibrium was operating on a C–C bond
9
#
˚
of length 1.64 A. The C3–C3 length is shorter than this, but
Fig. 8 Crystal and molecular structure of 9a/b. Thermal ellipsoids are
˚
drawn at 50% probability level. Hydrogen atoms are omitted.
at 1.58 A, is still a long C–C bond. Furthermore, such solution
equilibria have been well-established in the case of zinc-diazadiene
radicals.33
Returning to 9b, similar diazadiene radical anion complexes
are known for Al alkyls37 and iodides,30,36 Ga iodides and
bromides,30,35,36,38 and In chlorides,35,36,39 but we find AlCl3 be-
haviour with diazadiene 1 to be different from all the above,
and from BCl3 behaviour. In addition, stable radical (DAD)2M
complexes have long been known for M = Al and Ga,40 and
While thepresumed intermediate radicalcomplexproved elusive
for aluminium due to C–C coupling, and our attempts to charac-
terise it spectroscopically were frustrated by sample decomposition
and low yields, phenyl substituents on the diazabutadiene carbons
prevented any such path for the diazaborolium salt derived
230 | Dalton Trans., 2008, 222–233
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Fig. 9 (a) 1st derivative plot of the EPR spectrum of the fluid solution of 9b in CDCl3 at 293 K. (b) SIMFONIA simulation using coupling constants
a11-B = 4.0, a10-B = 1.34, a14-N = 5.5; a35-Cl = 4.5; a37-Cl = 4.0 G, Gaussian–Lorentzian parameter 0.7, and linewidth 2.4 G.
for M = Li, Mg and Zn,41 but this is the first case of which we
are aware of a stable diazadiene radical anion complex of boron.
Heretofore, only computational data has been available.42 These
heavier analogues have been well-studied by EPR and ENDOR
spectroscopies,36 and instructive comparisons can be drawn with
the data presented in Fig. 9. Computation of accurate a-values
is notoriously difficult,43 and this proved to be the case for 9b:
Unrestricted density functional-computed a-values for B, N, and
Cl at b3lyp using 6-311G(d,p) basis were 6, 3 and 4 G, respectively.
In particular, the value of 3 G for aN in what is essentially a complex
of a diazadiene radical seems small,36,37 in common with more
recent computational work on similar systems.43 The generally
regular features of the spectrum, where each peak is separated by
around 5 G, suggest that all coupling constants converge close to
this value. Indeed, in Ga and Al complexes of similar diazadiene
radicals, values of 5 to 6 G have been observed for aN.36 The
slightly sinusoidal envelope of the spectrum, and the deviation
of intensities from the theoretical 1 : 5 : 15 : 33 : 57 : 81 : 96 :
96 : 81 : 57 : 33 : 15 : 5 : 1 ratio, suggests that some of the a-
values deviate slightly from 5 G, but the breadth of the signals,
partly due to minor isotopomers from 10B (I = 3) and 37Cl (I =
3/2) and numerous small, unresolved aH couplings (all aH < 0.7
G by computation), means that reasonable fits can be obtained
for a range of values, and the only conclusion that can drawn
with certainty is that two N, two Cl and one B nuclei couple to
the unpaired spin density, all with values around 5 1 G for the
major isotopes. Such modest agreement with gas-phase ab initio
computation is not unexpected, given the presence of different
conformers in solution, and the room-temperature solution phase
of the spectral data collection. Inclusion of the small couplings
from the aryl hydrogens failed to improve the fit. Of interest is the
fact that chlorine couplings could be observed, in contrast to the
case with [1·InCl2·thf] and [(1·InCl)2].39 This is presumably due to
the closer proximity of the chlorine nuclei to the concentration
of unpaired spin density on the diazadiene fragment in 9b, and
the greater Cl–N and Cl–B overlap than for the indium case.
Notwithstanding the modest agreement with experiment, a plot
of the total Unrestricted Density functional spin density is shown
in Fig. 10a. A plot of the SOMO is shown in Fig. 10b. From this
it is clear that the unpaired electron density is concentrated on
the diazadiene fragment, in common with complexes of 1, though
significant delocalization from the diazadiene into the C-phenyls is
evident, while negligible amounts reside on the near-orthogonal N-
phenyls. The large a-spin density on nitrogen polarizes the boron
spin density to be b; there is minimal contribution to the SOMO
from boron. However, the significant a-spin on the chlorines can
be attributed to the contribution of their p-orbitals to the p-
SOMO via hyperconjugative donation of B–Cl r-bonding elecron
density, principally from a Cl 3p orbital, mixing with the partially
occupied diazadiene p-system. Thus, spin-density on the chlorines
appears to reside in a p-type orbital directed along the B–Cl bond
vector.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 222–233 | 231
©

View Article Online
References
1 A. J. Arduengo, III, R. L. Harlow and M. Kline, J. Am. Chem. Soc.,
1991, 113, 361.
2 E. S. Schmidt, A. Jockisch and H. Schmidbaur, J. Am. Chem. Soc.,
1999, 121, 9758; E. S. Schmidt, A. Schier and H. Schmidbaur, J. Chem.
Soc., Dalton Trans., 2001, 505.
3 R. J. Baker, R. D. Farley, C. Jones, M. Kloth and D. Murphy, J. Chem.
Soc., Dalton Trans., 2002, 3844.
4 A. Sunderman, M. Reiher and W. W. Schoeller, Eur. J. Inorg. Chem.,
1998, 305; N. Metzler-Nolte, New J. Chem., 1998, 793.
5 Y. Segawa, M. Yamashita and K. Nozaki, Science, 2006, 314, 113.
6 (a) L. Weber and G. Schmid, Angew. Chem., Int. Ed. Engl., 1974,
13, 1; (b) G. Schmid and J. Schulze, Chem. Ber., 1977, 110, 2744;
(c) G. Schmid, M. Polk and R. Boese, Inorg. Chem., 1990, 29, 442;
(d) L. Weber, E. Dobbert, R. Boese, M. T. Kirchner and D. Bla¨ser,
Eur. J. Inorg. Chem., 1998, 1145; (e) L. Weber, M. Schneider, R. Boese
and D. Bla¨ser, J. Chem. Soc., Dalton Trans., 2001, 378; (f) L. Weber,
M. Schnieder and P. Lo¨nneke, J. Chem. Soc., Dalton Trans., 2001,
3459; (g) L. Weber, A. Rausch, H. B. Wartig, H.-G. Stammler and B.
Neumann, Eur. J. Inorg. Chem., 2002, 2438; (h) L. Weber, I. Domke, A.
Rausch, A. Chrostowska and A. Dargelos, Dalton Trans., 2004, 2188;
(i) L. Weber, J. Fo¨rster, H.-G. Stammler and B. Neumann, Eur. J. Inorg.
Chem., 2006, 5048.
7 F. S. Mair, R. Manning, R. G. Pritchard and J. E. Warren, Chem.
Commun., 2001, 1136.
8 H. tom Dieck, M. Svoboda and T. Greiser, Z. Naturforsch., B: Anorg.
Chem. Org. Chem., 1981, 36, 823.
9 D. T. Carey, E. K. Cope-Eatough, E. Vilaplana-Mafe´, F. S. Mair,
R. G. Pritchard, J. E. Warren and R. J. Woods, Dalton Trans., 2003,
1083.
Fig. 10 (a) Plot of total spin density from an Unrestricted B3LYP/6-311G
(d,p) geometry optimised density functional computation on 9b; a-spin
denoted as dark blue, b-spin as dark green. Atom colours: chlorine: light
green; carbon: grey; hydrogen: white; boron: pink; nitrogen: light blue.
(b) Plot of SOMO, atom colours as above, a-orbital red–brown, b-orbital
green.
10 R. Carlson, U. Larsson and L. Hansson, Acta Chim. Scand., 1992, 46,
1211.
11 K. YvonW. Jeitschko and E. Par,LAZY PULVERIX, Software for the
calculation of theoretical powder diffraction patterns from single crystal
data.
12 G. Ferguson, R. McCrindle, A. J. McAlees and M. Parvez, Acta
Crystallogr., Sect. B., 1982, B38, 2679.
Conclusion
BCl3 cyclizes 1 through a chloroborated intermediate 3, isolable
from hexane, to give in more polar solvents a spontaneously
dehyrochlorinated diazaborole 4. In contrast, reaction of AlCl3
with 1 forms only acyclic mono- or di-adducts 5a/b and 6. Such
preformed M(III) chloride complexes of diazadiene ligands do not
lead to M(I) or M(II) species in the cases investigated. Reduction
of boron diazadienes proceeded further than those of aluminium
analogues, under similar conditions: Alkali metal reductions of
3 give mixtures of 4 and diazaborole 7. Pd(0) reduction cleanly
gave diazaborole 7. Reduction of 6 gave a low yield of closed
shell C–C coupled dimer 8 of the putative diazadiene radical
anion-AlCl2 complex monomer. Reduction of 2/BCl3, in which
additional phenyl groups on the diazadiene carbon atoms hinder
such radical coupling, gave predominantly diazaborole 2·BCl
9a, contaminated with 2·BCl2 9b, the first such stable radical
diazadiene complex of boron. Thus, high degrees of ortho-aryl
bulk on the N-substituents of the diazadiene promote varying
and unusual chemistry, including the first documented case of
chloroboration of an imine bond, the first example of C–C
coupling of a diazadiene radical bearing 2,6-diisopropylphenyl
substituents, and the first spectroscopic and structural data on
a stable diazabutadiene radical complex of boron, such species
having first been observed in heavier group 13 element compounds.
13 COLLECT, Data collection software, Bruker-Nonius B.V., Delft, The
Netherlands, 1999.
14 Z. Otwinowski and W. Minor, Methods Enzymol., 1996, 276, 307.
15 R. H. Blessing, Acta Crystallogr., Sect. A, 1995, A51, 33.
16 A. Altomare, G. Cascarano, C. Giacovazzo and A. Guagliardi, J. Appl.
Crystallogr., 1993, 26, 343.
17 SHELXL-97, Software for crystal structure refinement, G. Sheldrick,
University of Go¨ttingen, Germany, 1997.
18 GAUSSIAN 98, Revision A.7, Software for quantum chemical computa-
tion, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E.
Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels,
K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R.
Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski,
G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V.
Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-
Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill,
B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-
Gordon, E. S. Replogle and J. A. Pople, Gaussian, Inc., Pittsburgh PA,
1998.
19 W. Clegg, S. J. Coles, E. K. Cope and F. S. Mair, Angew. Chem., Int. Ed.,
1998, 37, 796–798; H. Hamaki, N. Takeda, T. Yamasaki, T. Sasamori
and N. Tokito, J. Organomet. Chem., 2007, 692, 44.
20 R. van Belzen, R. A. Klein, W. J. J. Smeets, A. L. Spek, R. Benedix and
C. J. Elsevier, Recl. Trav. Chim. Pays-Bas, 1996, 115, 275.
21 F. Sannicolo`, J. Org. Chem., 1983, 48, 2924.
22 W. B. Jennings, Chem. Rev., 1975, 75, 307.
Acknowledgements
23 J. E. Johnson, N. M. Morales, A. M. Gorczyca, D. D. Doliver and
M. A. McAllister, J. Org. Chem., 2001, 66, 7979; J. M. Blackwell,
W. E. Piers, M. Parvez and R. McDonald, Organometallics, 2002, 21,
1400.
We gratefully acknowledge Mr John Friend of the EPSRC
National Service for EPR spectroscopy for data collection and
processing, Dr Robert Manning and Mr Martyn Peers for
synthetic assistance, and the Nuffield Foundation, the EPSRC
and Creators Ltd for financial support (JEW).
24 H. A. Jenkins, C. L. Dumaresque, D. Vidovic and J. A. C. Clyburne,
Can. J. Chem., 2002, 80, 1398.
25 M. F. Lappert and B. Prokai, J. Chem. Soc., 1963, 4223.
26 A. Suzuki, Rev. Heteroat. Chem., 1997, 17, 271, and refs therein.
232 | Dalton Trans., 2008, 222–233
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
27 A. M. Kibardin, I. A. Litvinov, V. A. Naumov, Yu. T. Struchkov, T. V.
Gryaznova, Yu. B. Mikhailov and A. N. Pudovik, Dokl. Akad. Nauk
SSSR, 1988, 298, 369.
28 T. V. Laine, M. Klinga, A. Maaninen, E. Aitola and M. Leskela, Acta
Chem. Scand., 1999, 53, 968.
29 U. Klingebiel, M. Noltemeyer, H. G. Schmidt and D. Schmidt-Base,
Chem. Ber., 1997, 130, 753.
37 H. Schumann, M. Hummert, A. N. Lukoyanov and I. L. Fedushkin,
Organometallics, 2005, 24, 3891.
38 T. Pott, P. Jutzi, W. Kaim, W. W. Schoeller, B. Neumann, A.
Stammler, H.-G. Stammler and M. Wanner, Organometallics, 2002, 21,
3169.
39 R. J. Baker, R. D. Farley, C. Jones, M. Kloth and D. M. Murphy, Chem.
Commun., 2002, 1196.
30 R. J. Baker, R. D. Farley, C. Jones, M. Kloth and D. M. Murphy,
J. Chem. Soc., Dalton Trans., 2002, 1196.
31 J. S. Merriam and K. Niedenzu, J. Organomet. Chem., 1973, 51, C1.
32 B. Hessett, J. B. Leach, J. H. Morris and P. G. Perkins, J. Chem. Soc. A,
1972, 131.
33 D. T. Carey, F. S. Mair, R. G. Pritchard, J. E. Warren and R. J. Woods,
Dalton Trans., 2003, 3792.
34 M. Kaupp, H. Stoll, H. Preuss, W. Kaim, T. Stahl, G. van Koten, E.
Wissing, W. J. J. Smeets and A. L. Spek, J. Am. Chem. Soc., 1991, 113,
5606; G. van Koten, J. T. B. H. Jastrzebski and K. Vrieze, J. Organomet.
Chem., 1983, 250, 49; E. Wissing, E. Rijnberg, P. A. van der Schaaf,
K. van Gorp, J. Boersma and G. van Koten, Organometallics, 1994, 13,
2609; E. Wissing, S. von der Linden, E. Rijnberg, J. Boersma, W. J. J.
Smeets, A. L. Spek and G. van Koten, Organometallics, 1994, 13, 2602.
35 R. J. Baker, C. Jones, M. Kloth and D. P. Mills, New J. Chem., 2004,
28, 207.
40 F. G. N. Cloke, C. I. Dalby, M. J. Henderson, P. B. Hitchcock, C. H. L.
Kenna, R. N. Lamb and C. L. Raston, J. Chem. Soc., Chem. Commun.,
1990, 1394; F. G. N. Cloke, C. I. Dalby, P. J. Daff and J. C. Green,
J. Chem. Soc., Dalton Trans., 1991, 181; F. G. N. Cloke, G. R. Hanson,
M. J. Henderson, P. B. Hitchcock and C. L. Raston, J. Chem. Soc.,
Chem. Commun., 1989, 1002; M. J. Henderson, C. H. L. Kennard,
C. L. Raston and G. Smith, J. Chem. Soc., Chem. Commun., 1990,
1203.
41 M. G. Gardiner, G. R. Hanson, M. J. Henderson, F. C. Lee and C. L.
Raston, Inorg. Chem., 1994, 33, 2456; P. J. Bailey, C. M. Dick, S. Fabre,
S. Parsons and L. J. Yellowlees, Dalton Trans., 2006, 1602.
42 W. W. Schoeller and S. Grigoleit, J. Chem. Soc., Dalton Trans., 2002,
405.
43 T. Chivers, D. J. Eisler, C. Fedorchuk, G. Schatte, H. M. Tuononen and
R. T. Boere´, Chem. Commun., 2005, 3930; H. M. Tuononen and A. F.
Armstrong, Dalton Trans., 2006, 1885; H. M. Tuononen, T. Chivers,
A. F. Armstrong, C. Fedorchuk and R. T. Boere´, J. Organomet. Chem.,
2007, 692, 2705.
36 R. J. Baker, R. D. Farley, C. Jones, D. P. Mills, M. Kloth and D. M.
Murphy, Chem.–Eur. J., 2005, 11, 2972.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 222–233 | 233
©