Aryl Grignard cross-coupling of aryl chlorides catalysed by new, highly active phosphine/imidazolium nickel(II) complexes

Article abstract of DOI:10.1016/j.molcata.2006.06.004

Two new zwitterionic nickel(II) complexes, bearing phosphine/imidazolium ligands, have been prepared. Their catalytic activity, and the activity of an analogous Ni(II) complex, previously described by us, has been evaluated with a range of aryl chlorides, an aryl bromide and arylmagnesium halides. The catalytic activity is related to the length of the tether between the carbene and phosphine moieties. Thus the catalysts possessing a six-membered metallacycle are more active than that having a seven-membered metallacycle. Changing from a mesityl (Mes) to a 2,6-diisopropylphenyl (DIP) on the imidazole ring does not strongly influence the activity. The three complexes show moderate to very good activities with most substrates, and enhanced selectivities compared to previously published Ni(II)/N-heterocyclic carbene systems. The low activity observed with 4-chlorobenzotrifluoride seems to be related to the presence of a phosphine on the ligands.

Full text of DOI:10.1016/j.molcata.2006.06.004

Journal of Molecular Catalysis A: Chemical 259 (2006) 205–212
Aryl Grignard cross-coupling of aryl chlorides catalysed by new, highly
active phosphine/imidazolium nickel(II) complexes
∗
∗∗
`
Joffrey Wolf, Agnes Labande , Michael Natella, Jean-Claude Daran, Rinaldo Poli
Laboratoire de Chimie de Coordination, UPR CNRS 8241 (lie´ par convention a` l’Universite´ Paul Sabatier et a` l’Institut
National Polytechnique de Toulouse), 205 Route de Narbonne, 31077 Toulouse Cedex 4, France
Received 27 February 2006; received in revised form 25 April 2006; accepted 7 June 2006
Available online 26 July 2006
Abstract
Two new zwitterionic nickel(II) complexes, bearing phosphine/imidazolium ligands, have been prepared. Their catalytic activity, and the
activity of an analogous Ni(II) complex, previously described by us, has been evaluated with a range of aryl chlorides, an aryl bromide and
arylmagnesium halides. The catalytic activity is related to the length of the tether between the carbene and phosphine moieties. Thus the catalysts
possessing a six-membered metallacycle are more active than that having a seven-membered metallacycle. Changing from a mesityl (Mes) to
a 2,6-diisopropylphenyl (DIP) on the imidazole ring does not strongly influence the activity. The three complexes show moderate to very good
activities with most substrates, and enhanced selectivities compared to previously published Ni(II)/N-heterocyclic carbene systems. The low activity
observed with 4-chlorobenzotrifluoride seems to be related to the presence of a phosphine on the ligands.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Nickel; Phosphine; N-Heterocyclic carbene; C–C coupling; Grignard
1. Introduction
ligand (1a and 1d) [19]. These complexes proved very active
as precatalysts for the coupling of phenylmagnesium chloride
and 4-chloroanisole, giving an essentially total conversion of
the starting aryl chloride after 1 h at room temperature, with
only 3 mol% of precatalyst.
Based on these results, we wanted to further explore the
field and study the scope and limitations of our systems in the
Kumada-Corriu reaction. In the present paper, we describe the
preparation and characterisation of two new members of this
family of nickel(II) phosphine–imidazolium complexes, and the
evaluation of their catalytic activity with a range of aryl chlo-
rides, one aryl bromide and arylmagnesium halides.
The development of very active, yet robust catalysts for
applications to organic synthesis is still a challenge. Com-
plexes bearing N-heterocyclic carbenes are actively studied
nowadays, because of their high thermal stability, and because
metal–NHC bonds have high dissociation energies [1–3]. On the
other hand, phosphine complexes of nickel(II) are very active in
the C–C coupling reaction between aryl halides and aryl Grig-
nard reagents (Kumada-Corriu reaction), but it is not rare to
observe some ligand dissociation [4–11]. Consequently, mixed
phosphine/N-heterocyclic carbene nickel complexes should be
goodcandidatesascatalystsforC–Ccouplingreactions[12–17].
To date, only two Ni(II)/N-heterocyclic carbene systems have
been successfully tested in the reaction of aryl chlorides with
aryl Grignard reagents [18].
2. Experimental
2.1. General comments
We have recently described the preparation of two zwitte-
rionic nickel(II) complexes, bearing a phosphine/imidazolium
All reactions were carried out under a dry argon atmosphere
using Schlenk glassware and vacuum line techniques. Solvents
for syntheses were dried and degassed by standard methods
before use. Elemental analyses were carried out by the ana-
lytical service of the Laboratoire de Chimie de Coordination in
Toulouse.
∗
∗∗
Corresponding author. Tel.: +33 561333173; fax: +33 561553003.
Corresponding author.
E-mail address: labande@lcc-toulouse.fr (A. Labande).
1381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.06.004

206
J. Wolf et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 205–212
1H data were recorded on Bruker AC200, AM-250 and AV-
1H, N–CH–N), 7.97 (t, J = 1.5 Hz, 1H, MesN–CH C), 7.16 (t,
J = 1.5 Hz, 1H, AlkN–CH C), 6.99 (s, 2H, CH (Mes)), 4.85
(t, J = 4.9 Hz, 2H, N–CH2), 4.03 (t, J = 4.8 Hz, 2H, CH2–O),
3.54 (br, 1H, OH), 2.33 (s, 3H, p-CH3), 2.21 (q, J = 5.7 Hz, 2H,
C–CH2–C), 2.07 (s, 6H, o-CH3). MS [DCI(NH3)] m/z, (%):
245.2, (100) [C15H21N2O+]; 79.2, (100) [Br−].
500 spectrometers, operating at 200, 250 and 500 MHz, respec-
tively. 13C{H, P} and 31P{H} NMR data were recorded on the
Bruker AV-500 instrument, operating at 125.8 and 202.5 Hz,
respectively. The spectra were referenced internally using the
signal from the residual protiosolvent (1H) or the solvent sig-
nals (13C), and externally using 85% H3PO4 for 31P. Mass
spectra were obtained from acetonitrile or methanol solutions
on a TSQ7000 instrument from ThermoElectron (electrospray
ionisation or chemical ionisation), and from DMSO or DMF
solutions on a Nermag R10-10 instrument (FAB). GC chro-
matograms were recorded on a Fisons 8000 Series GC equipped
with a SPB-5 capillary column and the products were identi-
fied by comparison with authentic samples. Commercial chem-
icals were from Acros, Aldrich, Alfa Aesar, Avocado or Fluka
and used as received. N-(2,4,6-trimethylphenyl)imidazole (2a)
and N-(2,6diisopropylphenyl)imidazole (2b) were prepared as
described by Arduengo et al. [20] and Johnson [21]. Compound
1a was synthesised as previously described [19].
2.2.3. 1-(2-Bromoethyl)-3-(2,6-diisopropylphenyl)
imidazolium bromide (4b)
PBr3 (0.83 mL, 8.80 mmol) was slowly added to a cold
CH2Cl2 solution (70 mL, 0 ◦C) of 3b (2.59 g, 7.33 mmol). The
mixture was stirred for 20 h at room temperature, diluted with
CH2Cl2 (70 mL) and added to a cold, saturated aq. NaHCO3
solution (35 mL, 0 ◦C). The organic phase was extracted and
washed with a cold, saturated aq. NaHCO3 solution (1 × 10 mL,
0 ◦C). Subsequently, it was dried (MgSO4), filtered and the sol-
vent was removed in vacuo to give a white solid. Yield: 2.38 g
1
(78%). H NMR (δ, 500 MHz, CDCl3): 10.07 (t, J = 1.6 Hz,
1H, N–CH–N), 8.67 (t, J = 1.6 Hz, 1H, (DIP)N–CH C), 7.44
(t, J = 7.9 Hz, 1H, p-CH (DIP)), 7.20 (d, J = 7.9 Hz, 2H, m-
CH (DIP)), 7.13 (t, J = 1.6 Hz, 1H, AlkN–CH C), 5.15 (t,
J = 5.4 Hz, 2H, N–CH2), 3.96 (t, J = 5.4 Hz, 2H, CH2–Br),
2.23 (h, J = 6.8 Hz, 2H, CH–(CH3)2), 1.09 (d, J = 6.8 Hz,
2.2. Synthesis of ligands
2.2.1. 1-(2-Hydroxyethyl)-3-(2,6-diisopropylphenyl)-
imidazolium bromide (3b)
12H, CH–(CH3)2), 1.05 (d, J = 6.8 Hz, 12H, CH–(CH3)2); 13
C
NMR (δ, 125.8 MHz, CDCl3): 145.35 (o-C (DIP)), 138.21
(N–CH–N), 131.85 (p-CH (DIP)), 129.98 (N–C (DIP)), 124.57
(m-CH (DIP)), 124.39 ((DIP)N–CH ), 123.94 (AlkN–CH ),
50.73 (CH2–N), 32.00 (CH2–Br), 28.50 (CH–(CH3)2), 24.47
(CH–(CH3)2), 24.00 (CH–(CH3)2). m.p. 73–75 ◦C. Anal. calcd.
for C17H24Br2N2; C: 49.06, H: 5.81, N: 6.73%; found: C:
48.71, H: 6.11, N: 6.43%. MS (ESI) m/z, (%): 335, (100)
[C17H24BrN2+]; 79, (100) [Br−].
2-Bromoethanol (2.5 mL, 35 mmol) was added to a toluene
solution (150 mL) of 2b (6.2 g, 27.2 mmol). The mixture was
stirred at 120 ◦C for 20 h. During this time a yellow-orange
oil separated out in the flask. The mixture was cooled to
room temperature and the toluene phase was removed. The
oil was extracted with CH2Cl2 (20 mL), the resulting solution
was slowly precipitated in ether (400 mL), and the pale yel-
low solid which formed was filtered and dried in vacuo. Yield:
6.83 g (72%). 1H NMR (δ, 500 MHz, CDCl3): 9.50 (s, 1H,
N–CH–N), 8.19 (s, 1H, (DIP)N–CH C), 7.38 (t, J = 7.9 Hz, 1H,
p-CH (DIP)), 7.15 (d, J = 7.9 Hz, 2H, m-CH (DIP)), 7.09 (s,
1H, AlkN–CH C), 4.73 (t, J = 4.8 Hz, 2H, N–CH2), 3.96 (br,
1H, OH), 3.84 (t, J = 4.8 Hz, 2H, CH2–OH), 2.15 (h, J = 6.8 Hz,
2H, CH–(CH3)2), 1.01 (d of d, J = 6.8 Hz, 12H, CH–(CH3)2);
13C NMR (δ, 125.8 MHz, CDCl3): 145.35 (o-C (DIP)), 137.63
(N–CH–N), 131.75 (p-CH (DIP)), 130.00 (N–C (DIP)), 124.50
(m-CH (DIP)), 124.31 ((DIP)N–CH ), 123.73 (AlkN–CH ),
60.17 (CH2–OH), 52.01 (CH2–N), 28.43 (CH–(CH3)2), 24.33
(CH–(CH3)2), 24.00 (CH–(CH3)2). m.p. 169–170 ◦C. Anal.
calcd. for C17H25BrN2O; C: 57.79, H: 7.13, N: 7.93%; found:
C: 57.54, H: 6.90, N: 7.79%. MS (ESI) m/z, (%): 273.15, (100)
[C17H25N2O+]; 79, (100) [Br−].
2.2.4. 1-(3-Bromo-n-propyl)-3-(2,4,6-trimethylphenyl)
imidazolium bromide (4c)
PBr3 (1.1 mL, 11.5 mmol)wasslowlyaddedtoacoldCH2Cl2
solution (80 mL, 0 ◦C) of 3c (3.26 g, 10 mmol). The mixture
was stirred for 15 h at room temperature, diluted with CH2Cl2
(80 mL) and added to a cold, saturated aq. NaHCO3 solution
(25 mL, 0 ◦C). The organic phase was extracted and washed
with cold, saturated aq. NaHCO3 (1 × 10 mL, 0 ◦C). Subse-
quently, it was dried (MgSO4), filtered and the solvent was
removed in vacuo to give a pale yellow solid. Yield: 3.37 g
(87%). 1H NMR (δ, 500 MHz, CDCl3): 10.23 (s, 1H, N–CH–N),
8.23 (s, 1H, MesN–CH C), 7.23 (s, 1H, AlkN–CH C), 6.95
(s, 2H, CH (Mes)), 4.86 (t, J = 6.8 Hz, 2H, N–CH2), 3.48 (t,
J = 6.3 Hz, 2H, CH2–Br), 2.63 (q, J = 5.2 Hz, 2H, C–CH2–C),
2.29 (s, 3H, p-CH3), 2.02 (s, 6H, o-CH3); 13C NMR (δ,
125.8 MHz, CDCl3): 141.28 (p-C (Mes)), 137.78 (N–CH–N),
134.10 (o-C (Mes)), 130.62 (N–C (Mes)), 129.85 (CH (Mes)),
124.04 (AlkN–CH ), 123.39 (MesN–CH ), 48.79 (N–CH2),
33.01 (C–CH2–C), 29.13 (CH2–Br), 21.11 (p-CH3), 17.72 (o-
CH3). m.p. 148–149 ◦C. Anal. calcd. for C15H20Br2N2; C:
46.42, H: 5.19, N: 7.22%; found: C: 46.30, H: 4.92, N: 7.11%.
MS (ESI) m/z, (%): 399, (100) [C15H20BrN2+]; 79–81, (100)
[Br−].
2.2.2. 1-(3-Hydroxy-n-propyl)-3-(2,4,6-trimethylphenyl)
imidazolium bromide (3c)
3-Bromo-1-propanol (2.2 mL, 24.0 mmol) was added to a
toluene solution (60 mL) of 2a (3.72 g, 20.0 mmol). The mix-
ture was stirred at 120 ◦C for 15 h. During this time a cream
solid precipitated in the flask. The suspension was cooled to
room temperature, the solid was decanted, washed with toluene
(2 × 15 mL) and dried in vacuo. A white solid was obtained.
1
Yield: 4.47 g (69%). H NMR (δ, 250 MHz, CDCl3): 9.64 (s,

J. Wolf et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 205–212
207
2.2.5. 1-(2-Diphenylphosphinoethyl)-3-(2,6-
diisopropylphenyl)imidazolium bromide (5b)
(br s), 6.69 (br s), 6.13 (br s), 3.86 (br s), 2.94 (br s), 2.62 (br s),
1.30 (br s), 0.23 (br s), 0.07 (br s), −2.22 (br s); m.p. 157–159 ◦C.
Anal. calcd. for C29H34Br3NiN2P; C: 47.07, H: 4.63, N: 3.79%;
found: C: 46.72, H: 4.85, N: 3.59%. MS (ESI) m/z, (%): 441.6,
(100) [C29H34N2P+]; 79, (100) [Br−].
KPPh2, freshly made from t-BuOK (337 mg, 3.00 mmol)
and HPPh2 (0.54 mL, 3.15 mmol) in DMSO (8 mL), was added
to a DMSO solution (8 mL) of 4b (1.19 g, 2.86 mmol). The
solution was allowed to stir for 2 h at room temperature. The
solvent was then removed under vacuum. Methanol (5 mL)
was added to quench the excess KPPh2, then removed under
vacuum. Dichloromethane (10 mL) was added, the resulting
mixture was filtered and diethyl ether (100 mL) was added.
The resulting precipitate was separated by filtration and dried
in vacuo. The product was obtained as a white, air-sensitive
solid. Yield: 1.37 g (92%). 1H NMR (δ, 200 MHz, CDCl3):
10.20 (s, 1H, N–CH–N), 7.85 (s, 1H, MesN–CH C), 7.46 (t,
J = 7.8 Hz, 1H, p-CH (DIP)), 7.39 (m, 4H, o-CH (C6H5)), 7.30
(m, 6H, m/p-CH (C6H5)), 7.23 (d, J = 8 Hz, 2H, m-CH (DIP)),
7.07 (s, 1H, AlkN–CH C), 4.84 (d of t, J = 12.4–6.9 Hz, 2H,
N–CH2), 2.84 (t, J = 6.8 Hz, 2H, CH2–P), 2.25 (h, J = 6.8 Hz,
2H, CH–(CH3)2), 1.17 (d, J = 6.6 Hz, 6H, CH–(CH3)2), 1.07 (d,
J = 6.9 Hz, 6H, CH–(CH3)2); 31P NMR (δ, 81 MHz, CDCl3):
−20.77 (CH2–PPh2) (data consistent with those found in the
literature [13]).
2.3.2. 1-(3-Diphenylphosphinopropyl)-3-(2,4,6-
trimethylphenyl)imidazolium nickel tribromide (1c)
NiBr2(MeCN)2 (312 mg, 1.01 mmol) was added to a suspen-
sion of 5c (435 mg, 0.88 mmol) in THF (10 mL). The mixture
was stirred at room temperature for 1 h. The precipitate was
filtered, the green solid/oil obtained was washed with THF
(2 × 5 mL), CHCl3 (2 × 10 mL) and dried under vacuum to give
agreensolid. Yield:612 mg(97%). Greencrystalswereobtained
1
by diffusion of diethyl ether into an acetonitrile solution. H
NMR (δ, 250 MHz, d6-(Me)2CO): 20.72 (br s), 8.81 (br s), 8.63
(br s), 7.92 (s), 7.23 (s), 6.13 (s), 6.02 (br s), 4.53 (br s), 3.94 (br
s), 3.70 (br s), 2.39 (br s), 2.23 (br s), 1.86 (s), −0.34 (br s), −2.43
(br s); m.p. 226–227 ◦C. Anal. calcd. for C27H30Br3NiN2P; C:
45.55, H: 4.25, N: 3.93%; found: C: 45.12, H: 3.95, N: 3.79%.
MS (FAB, MNBA matrix) m/z, (%): 413 (100) [C27H30N2P+];
631 (0.5) [C27H30Br2NiN2P+].
2.2.6. 1-(3-Diphenylphosphinopropyl)-3-(2,4,6-trimethyl-
phenyl)imidazolium bromide (5c)
2.4. Standard procedure for Kumada-Corriu reactions
KPPh2, freshly made from t-BuOK (152 mg, 1.35 mmol) and
HPPh2 (0.25 mL, 1.42 mmol) in DMSO (3 mL), was added to a
DMSO solution (2 mL) of 4c (500 mg, 1.29 mmol). The solu-
tion was allowed to stir for 15 h at room temperature. The
solvent was then removed under vacuum. Methanol (5 mL) was
added to quench the excess KPPh2, then removed under vacuum.
Dichloromethane (10 mL) was added and the resulting mixture
filtered. The filtrate was concentrated (ca. 1.5 mL) and diethyl
ether (20 mL) was added. The resulting precipitate was sepa-
rated by filtration and dried in vacuo. The product was obtained
The aryl halide (1 mmol) and diethyleneglycol-di-n-
butylether (30 ␮L as internal standard) were mixed with 1a,
1b or 1c (3 mol% based on aryl halide) in THF (1 mL). The
arylmagnesium halide (1.5 mmol) was slowly added to the
mixture. Reactions were kept at room temperature (between
23 and 25 ◦C) for 18 h and quenched with a minimum of
ethanol. In order to remove the catalyst and the magne-
sium salt, mixtures were filtered on silica with ether as sol-
vent. The resulting mixtures were analysed by gas chromato-
graphy.
1
as a white, air-sensitive solid. Yield: 500 mg (79%). H NMR
(δ, 200 MHz, CDCl3): 10.35 (t, J = 1.7 Hz, 1H, N–CH–N), 7.75
(t, J = 1.8 Hz, 1H, MesN–CH C), 7.36 (t, J = 7.8 Hz, 1H, o-
CH (C6H5)), 7.28 (m, 6H, m/p-CH (C6H5)), 7.16 (t, J = 1.7 Hz,
1H, AlkN–CH C), 6.92 (s, 2H, CH (Mes)), 4.76 (d of t,
J = 25.6–6.5 Hz, 2H, N–CH2), 2.97 (d of t, J = 23–6.5 Hz, 2H,
CH2–P), 2.27 (s, 3H, p-CH3), 2.63 (m, 2H, C–CH2–C), 1.98
(s, 6H, o-CH3); 31P NMR (δ, 81 MHz, CDCl3): −14.1 (s,
CH2–PPh2) (data consistent with those found in the literature
[13]).
2.5. X-ray analyses
A single crystal was mounted under inert perfluoropolyether
at the tip of a glass fibre and cooled in the cryostream
of an Oxford-Diffraction XCALIBUR CCD diffractometer.
Data were collected using the monochromatic Mo K␣ radi-
ation (λ = 0.71073). The final unit cell parameters were
obtained by the least-squares refinement of a large num-
ber of selected reflections. Only statistical fluctuations were
observed in the intensity monitors over the course of the data
collection.
2.3. Synthesis of nickel(II) complexes
The structure was solved by direct methods (SIR97 [22]) and
refined by least-squares procedures on F2 with the SHELXL-97
program [23] using the integrated system WINGX(1.63) [24].
Hydrogen atoms attached to carbon atoms were introduced at
calculated positions and treated as riding on their parent atoms
2.3.1. 1-(2-Diphenylphosphinoethyl)-3-(2,6-
diisopropylphenyl)imidazolium nickel tribromide (1b)
NiBr2(DME) (410 mg, 1.33 mmol) was added to a solution
of 5b (700 mg, 1.34 mmol) in THF (20 mL). The mixture was
stirred at room temperature for 30 min to give a green solution.
The solvent was removed under vacuum and CHCl3 (20 mL)
was added. The mixture was filtered and the solvent removed to
give a green solid. Yield: 953 mg (97%). 1H NMR (δ, 200 MHz,
CDCl3): 1H NMR (δ, 250 MHz, d6-(Me)2CO): 20.79 (br s), 7.05
˚
[d(CH) = 0.96–0.98 A] with a displacement parameter equal to
1.2 (C6H5, CH2, OH) or 1.5 (CH3) times that of the parent
atom. Owing to the poor quality of the crystal, only reflections
with theta below 23.3◦ were used in the refinement procedure.
The molecular view was realised with the help of ORTEP-

208
J. Wolf et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 205–212
Table 1
Crystal data and structure refinement for 1c
Fig. 1. Proposed structures of Ni(II) carbenic complexes from 1a: monomer (cis
coordination) and dimer (trans coordination).
Empirical formula
Formula weight
Temperature
C27H30Br3N2Ni P
711.94
293(2) K
˚
Wavelength
Crystal system
Space group
0.71073 A
3. Results and discussion
Orthorhombic
Pbca
˚
˚
Following our recent discovery that the nickel(II) com-
plexes 1a and 1d are excellent precatalysts for the coupling
of phenylmagnesium chloride and 4-chloroanisole [19], it was
of interest to see whether small structural changes in the
phosphine–imidazolium ligand could affect the properties or
catalytic activity. For this purpose, we considered two modifica-
tions: (i) replacement of the ortho substituents on the aryl group
of the imidazolium moiety by bulkier isopropyl substituents; (ii)
an increased length of the tether from 2 to 3 carbons, thus form-
ing a seven-membered metallacycle instead of a six-membered
one in the putative phosphino–carbene complex (Fig. 1).
Ligands 5b and 5c were prepared by a natural exten-
sion of the procedure previously described for ligand 5a
[19] (Scheme 1). Thus, 2-bromoethanol reacted with N-(2,6-
diisopropylphenyl)imidazole to give imidazolium salt 3b in
good yield. Analogously, the reaction of 3-bromopropanol with
N-mesitylimidazole gave imidazolium salt 3c in good yield (in
Unit cell dimensions
a = 13.2103(8) A,
b = 14.6501(9) A,
◦
˚
c = 30.0552(17) A, α = 90 ,
β = 90◦, γ = 90◦
3
˚
Volume
Z
Density (calculated)
Absorption coefficient
F(0 0 0)
Crystal size
Theta range for data collection
Index ranges
5816.6(6) A
8
1.626 Mg/m3
4.862 mm−1
2832
0.23 mm × 0.13 mm × 0.11 mm
2.86–23.25◦
−14 ≤ h ≤ 14, −16 ≤ k ≤ 16,
−29 ≤ l ≤ 33
Reflections collected
32586
Independent reflections
Completeness to theta = 23.25◦
Absorption correction
Maximum and minimum transmission
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I > 2σ(I)]
R indices (all data)
4176 [R(int) = 0.1041]
99.8%
Semi-empirical from equivalents
0.5426 and 0.4162
Full-matrix least-squares on F2
4176/0/311
1.236
¨
this case, the procedure described by Furstner [26] was slightly
R1 = 0.1019, wR2 = 0.1407
R1 = 0.1290, wR2 = 0.1524
0.716 and −0.468 Einstein A
modified and allowed us to improve the yield from 28% to 69%).
Bromination of 3b and 3c with PBr3 gave compounds 4b and 4c.
Finally, nucleophilic substitution with KPPh2 gave the expected
imidazolium salts 5b and 5c, with Br− as the anion, in good to
excellent yields.
Zwitterionic nickel(II) complexes 1b and 1c were obtained,
like the previously reported 1a [19], in excellent yields by stir-
ring a mixture of the desired ligand and NiBr2L2 for 1 h in
THF at room temperature (Scheme 2). Like 1a, the two new
complexes are green, paramagnetic compounds, with a dis-
torted tetrahedral geometry. They have been characterized by
NMR, elemental analysis and mass spectrometry, and by an
X-ray diffraction study for 1c (Fig. 2). The paramagnetism
−3
˚
Largest diffraction peak and hole
III [25]. Crystal data and refinement parameters are shown in
Table 1.
Crystallographic data (excluding structure factors) have been
deposited with the Cambridge Crystallographic Data Centre as
supplementary publication no. CCDC 299074. Copies of the
data can be obtained free of charge on application to the Director,
CCDC, 12 Union Road, Cambridge CB2 1EZ, UK (Fax: +44
1223 336 033; e-mail: deposit@ccdc.cam.ac.uk).
Scheme 1. Reagents and conditions: (i) 2-bromoethanol, toluene, 120 ◦C, 20 h; (ii) 3-bromo-1-propanol, toluene, 120 ◦C, 15 h; (iii) PBr3, CH2Cl2, 0 ◦C, 15–20 h;
(iv) HPPh2, t-BuOK, DMSO, rt, 2 h (5a, 5b), 15 h (5c).

J. Wolf et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 205–212
209
Table 2
◦
˚
Hydrogen bonds for 1c (A and )
C–H· · ·Br(1)
d(C–H)
d(H· · ·Br)
d(C· · ·Br)
(CHBr)
C(1)–H(1)· · ·Br(1)a
C(11)–H(11A)· · ·Br(1)a
C(3)–H(3)· · ·Br(1)b
0.93
0.97
0.93
2.76
2.90
2.92
3.566(11)
3.666(13)
3.707(13)
146.0
137.1
143.5
a
Symmetry transformation used to generate equivalent atom: x − 1/2, y,
−z + 3/2.
b
Symmetry transformation used to generate equivalent atom: −x + 2, y + 1/2,
−z + 3/2.
Scheme 2. Reagents and conditions: (i) NiBr2(DME), THF, rt, 30 min to 1 h
(1a, 1b); (ii) NiBr2(MeCN)2, THF, rt, 1 h (1c).
acteristic 1H NMR signal at ca. 10.5 ppm corresponding to the
acidic proton of the imidazolium moiety. The 31P NMR spec-
trum of the deprotonation product of 1c shows a similar pattern,
with two close singlets at −5.2 and −5.5 ppm, and a very small
signal at −15.0 ppm.
is indicated by the large paramagnetic shift of certain reso-
1
nances in the H NMR spectrum. Unfortunately, we did not
manage to isolate the corresponding carbenic Ni(II) complexes,
obtained by deprotonation of the imidazolium moiety in the
presence of LiN(SiMe3)2. However, the deprotonation of 1a–c
induced a colour change to yellow-brown, consistent with the
formation of diamagnetic, square–planar complexes. An NMR
monitoring of the deprotonation of 1a confirmed the forma-
tion of nickel phosphine–carbene complexes. Indeed, the 31P
NMR spectrum shows two sharp, very close signals at 3.28 ppm
(major) and 2.99 ppm (minor), in addition to a sharp signal at
−18.3 ppm which could correspond to the free phosphine. In the
13C NMR spectrum, two small doublets were observed in the
typical region of coordinated NHCs at, respectively, 168.3 ppm
(JPC = 108.2 Hz) and 159.85 ppm (JPC = 35.2 Hz). A selective
31P decoupling experiment showed that these two doublets
are coupled to the phosphorus nuclei resonating at 3.28 and/or
2.99 ppm, but not to the one at −18.3 ppm. The presence of two
doublets, with very different JPC coupling constants, could be
explained by the formation of two carbenic complexes (Fig. 1):
a monomer, bearing the chelating carbene-phosphine ligand (cis
coupling), and a dimer, with the carbene and the phosphine being
in trans on the nickel. Moreover, we did not observe the char-
A view of compound 1c is illustrated in Fig. 2. The
Ni atom is tetrahedrally coordinated by the phosphorus
atom of the phosphine group and by three bromine atoms.
The Ni–P distance is comparable with the value found
in [NiBr3(PPh3)]−(C24H20As)+ (2.323(1) A)] [27]. The Br
˚
atoms are rotationally staggered with respect to the phenyl
and the imidazolium groups. The three Br–Ni–Br angles
show a significant deviation from their mean, 112.8(1)◦,
with two values greater than the ideal tetrahedral angle.
The mesityl group is found roughly orthogonal to the
imidazolium ring with a dihedral angle of 88.2(4)◦, as
previously observed in [1-(2,4,6-Me3C6H2)imidazolium-3-
{CH2C(t-Bu) N(i-Pr)}]bromide[28]. Theinternalbondlengths
and angles of the imidazolium ring are unexceptional and lie
within the range expected [29]. It is interesting to note that
therearethreeweakintermolecularC–H· · ·Br(1)hydrogenbond
interactions (Table 2), the shortest one being with the acidic imi-
dazolium proton.
The new complexes, together with the previously reported
complex 1a, were tested as precatalysts for the coupling reac-
tion between a variety of aryl chlorides and Grignard reagents,
◦
˚
Fig. 2. Molecular view of compound 1c with atom labelling scheme. H atoms have been omitted for clarity. Selected bond lengths (A) and bond angles ( ): Ni(1)–P(1)
2.330(3); Ni(1)–Br(1) 2.3723(18); Ni(1)–Br(2) 2.3774(19); Ni(1)–Br(3) 2.3518(19); N(1)–C(11) 1.476(13); N(1)–C(1) 1.327(12); N(1) C(3) 1.365(14); N(2)–C(1)
1.313(12); N(2)–C(2) 1.362(14); N(2)–C(211) 1.448(13); C(2)–C(3) 1.326(16); P(1)–Ni(1)–Br(1) 100.52(10); P(1)–Ni(1)–Br(2) 105.26(9); P(1)–Ni(1)–Br(3)
110.73(10); Br(1)–Ni(1)–Br(2) 106.07(8); Br(1)–Ni(1)–Br(3) 112.77(8); Br(2)–Ni(1)–Br(3) 119.57(8); N(1)–C(1)–N(2) 108.7(10); C(1)–N(1)–C(3) 107.7(10);
C(1)–N(2)–C(2) 108.7(10); N(1)–C(3)–C(2) 107.8(11); N(2)–C(2)–C(3) 107.1(11).

210
J. Wolf et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 205–212
Table 3
Cross-coupling reactions of aryl chlorides with arylmagnesium chloride or bromidea,b
d
Entry
R1
E
R2
X
Cat. 1ac
Cat. 1bc
Cat. 1cc
IPr/Ni(acac)2
1
2
3
4
5
6
7
8
9
H
4-CF3
N
C
C
C
C
C
N
C
C
C
C
H
H
H
H
H
H
Cl
Cl
Cl
Cl
Cl
Cl
Br
Br
Br
Br
Br
>99
37
86
95
78
4
>99
36
99
92
48
>99
53
96
92
82
3
>99
33
98
87
87
>99
29
63
80
23
7
>99
42
76
89
12
>99
96
81
71
73
4-CH3
4-OCH3
2-CH3
Bromomesitylene
H
4-CF3
H
4-CH3
2-CH3
5
4-OCH3
4-OCH3
4-OCH3
4-OCH3
4-OCH3
>99
>99
93
88
77
10
11
GC yield (%) of heterocoupling product 8.
a
For the meaning of R1, R2, E and X and for the nature of the possible reaction products, see Scheme 3.
b
c
Conditions: 1.0 equiv. aryl halide, 1.5 equiv. aryl Grignard, 3 mol% 1a–1c, THF, 25 ◦C, t = 18 h.
GC yield using diethyleneglycol-di-n-butylether as the internal standard. The relative yields in percent are normalized to the reagent in defect (6):
6 + 8 + 9 + 10 = 100.
d
Ref. [18].
Table 4
Detailed results of cross-coupling reactions with phenylmagnesium chloridea
Entry
R
E
Cat. 1a
Cat. 1b
Cat. 1c
IPr/Ni(acac)2
1
2
3
4
5
6
H
4-CF3
4-CH3
4-OCH3
2-CH3
N
C
C
C
C
C
100/0/0/0/6
37/63/0/0/30
86/2/5/7/17
95/4/0/1/19
78/10/12/0/24
4/64/32/0/27
100/0/0/0/17
53/47/0/0/30
96/2/1/1/10
92/6/1/1/22
82/11/7/0/20
3/71/26/0/23
100/0/0/0/21
29/71/0/0/27
63/32/4/1/15
80/18/1/1/19
23/67/10/0/22
7/53/40/0/26
100/0/0/0/4
96/4/0/0/4
81/7/3/9/11
71/19/1/9/16
73/9/10/8/15
5/80/15/0/12
Bromomesitylene
a
The product distribution is given as 8/6/9/10/11; the amount of 10 was approximated by comparison of the peak integrations.
Table 5
Detailed results of cross-coupling reactions with 4-anisylmagnesium bromidea
Entry
R
E
Cat. 1a
Cat. 1b
Cat. 1c
IPr/Ni(acac)2
7
8
9
10
11
H
4-CF3
H
4-CH3
2-CH3
N
C
C
C
C
100/0/0/0/7
36/63/1/0/17
99/0/0/1/4
92/7/0/1/5
48/45/6/1/6
100/0/0/0/6
33/66/1/0/15
98/1/0/1/6
87/11/1/0/3
87/9/3/1/5
100/0/0/0/10
42/57/1/0/16
76/23/0/1/4
89/10/1/0/5
12/83/5/0/5
100/0/0/0/1
100/0/0/0/8
93/3/3/1/16
88/5/4/3/15
77/20/0/3/10
a
The product distribution is given as 8/6/9/10/11; the amount of 10 was approximated by comparison of the peak integrations.
extending the previous study of the coupling of 4-chloroanisole
with phenylmagnesium chloride catalyzed by 1a [19]. The cor-
responding trichloride complex 1d was not further investigated,
because our previous study showed that the activity does not
depend on the nature of the halide ligand. As we have previously
shown [19], the catalytic results are independent on whether the
Grignard reagent or an external base (e.g. LiN(SiMe3)2) is used
to deprotonate the precatalysts. Therefore, we consider that a
chelating phosphine–carbene coordination mode is present in
the catalyst as indicated by the NMR results described above.
The results are given in Table 3. A more detailed analysis of the
product distribution yields the results shown in Tables 4 and 5
(Scheme 3).
tionsofthecouplingsinvolvingtheCF3-substitutedarylchloride
substrate (runs 2 and 8), also show an equal or slightly bet-
ter activity than the Ni(acac)2/IPr system [18], as can be seen
from the kinetic studies (Figs. 3 and 4). Indeed, in the reac-
tion of 4-chloroanisole with phenylmagnesium chloride, the 4-
chloroanisole conversion reached 95% with 1a (Fig. 5) and 82%
Some general trends emerge from this work. Complex 1a,
bearing a mesityl group on the imidazolium moiety, is a bit
more active than 1b. Both complexes, with the notable excep-
Scheme 3. Reagents and products for the catalytic Kumada-Corriu reaction (for
the results, see Tables 3–5).

J. Wolf et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 205–212
211
the formation of a seven-membered cycle slows the reaction
down. For several substrates, complex 1c also shows a reduced
selectivity, even relative to the Ni(acac)2/IPr system.
One noticeable exception was observed with 4-
chlorobenzotrifluoride, where conversions were low, with
up to 71% starting material left for entry 2 (Table 4) and 66%
for entry 8 (Table 5). A similar behaviour was reported for
a monophosphine-based system, namely Ni(acac)2/P(t-Bu)3
(up to 53% starting material left), whereas the use of a simple
imidazolium salt gave excellent conversions under the same
conditions (Table 4, entry 2) [18]. A GC monitoring of this
reaction provided insight as to the possible cause of this reduced
activity.
The rate of conversion is initially high but dramatically drops
after ca. 2 h. This result clearly indicates that the catalyst is rather
active at the early stages of the reaction but dies as the reaction
progresses. This stability problem appears closely related to the
simultaneous presence of CF3 groups and a phosphine in the
ligand system.
Fig. 3. Conversion of 4-chloroanisole vs. time with 1a–c (Table 3, entry 4).
Finally, the presence of ortho substituents relative to the chlo-
rineatomhadadramaticeffectonconversionwith1c, butamuch
less pronounced effect with 1b. A curious effect, for which we
cannot offer an obvious rationalization, is the significantly lower
activity of 1a relative to 1b for entry 11, whereas about the same
activity of the two catalysts is observed for entry 5.
4. Conclusion
Zwitterionic complexes of Ni(II), bearing phos-
phine/imidazolium ligands, have been prepared and show
high activities and selectivities for the coupling of aryl
chlorides with aryl Grignard reagents at room temperature.
We have observed that small structural changes, such as
increasing the length of the ligand tether from 2 to 3 carbons,
could have a dramatic effect on the catalytic activity. Very
high conversions have been achieved, except in the case of
4-chlorobenzotrifluoride, where the presence of a phosphine
moiety in the catalyst seems to hamper the reaction. We are now
studying the coordination chemistry of these ligands on other
metals, and the catalytic properties of the resulting complexes
will be tested in due course.
Fig. 4. Conversion of 4-chlorotoluene vs. time with 1a–c (Table 3, entry 10) and
comparison with the Ni(acac)2/IPr system.
with 1b after 1 h (Fig. 3). On the other hand, complex 1b was a bit
more selective than 1a and gave less by-products (Tables 4 and 5)
[30]. Both are more selective than the Ni(acac)2/IPr system.
In most cases, complex 1c showed a poorer activity than 1a
or 1b, as can be seen from Table 3 and Fig. 3. It seems that
Acknowledgement
We are grateful to the CNRS for support of this work and to
´
the “Fonds Social Europeen” for a PhD fellowship to JW.
References
[1] W.A. Herrmann, Angew. Chem. Int. Ed. Engl. 41 (2002) 1290.
[2] D. Bourissou, O. Guerret, F.P. Gabbai, G. Bertrand, Chem. Rev. 100 (2000)
39.
[3] K.J. Cavell, D.S. McGuinness, Coord. Chem. Rev. 248 (2004) 671.
[4] K. Tamao, K. Sumitani, M. Kumada, J. Am. Chem. Soc. 94 (1972) 4374.
[5] M. Kumada, Pure Appl. Chem. 52 (1980) 669.
Fig. 5. Conversion of 4-chlorobenzotrifluoride vs. time with 1a (Table 3,
[6] K. Tamao, J. Organomet. Chem. 653 (2002) 23.
entry 2).
[7] T. Banno, Y. Hayakawa, M. Umeno, J. Organomet. Chem. 653 (2002) 288.

212
J. Wolf et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 205–212
[8] E. Wenkert, E.L. Michelotti, C.S. Swindell, M. Tingoli, J. Org. Chem. 49
(1984) 4894.
[9] L.N. Pridgen, J. Org. Chem. 47 (1982) 4319.
[10] A. Minato, K. Tamao, K. Suzuki, M. Kumada, Tetrahedron Lett. 21 (1980)
4017.
[11] J.W. Dankwardt, Angew. Chem. Int. Ed. Engl. 43 (2004) 2428.
[12] J. Huang, S.P. Nolan, J. Am. Chem. Soc. 121 (1999) 9889.
[13] N. Tsoureas, A.A. Danopoulos, A.A.D. Tulloch, M.E. Light,
Organometallics 22 (2003) 4750.
[14] N. Stylianides, A.A. Danopoulos, N. Tsoureas, J. Organomet. Chem. 690
(2005) 5948.
[21] A.L. Johnson, US Patent 3,637,731 (1972).
[22] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, C. Giacovazzo, A.
Guagliardi, A.G.G. Moliterni, G. Polidori, R. Spagna, J. Appl. Crystallogr.
32 (1999) 115.
[23] G.M. Sheldrick, SHELX97, Programs for Crystal Structure Analysis
(Release 97-2), Institu¨t fu¨r Anorganische Chemie der Universita¨t, Tam-
manstrasse 4, D-3400 Go¨ttingen, Germany, 1998.
[24] L.J. Farrugia, J. Cryst. Appl. 32 (1999) 837.
[25] (a) M.N. Burnett, C.K. Johnson, ORTEP-III—Report ORNL-6895, Oak
Ridge National Laboratory, Oak Ridge, Tennessee, 1996;
(b) L.J. Farrugia, J. Appl. Crystallogr. 30 (1997) 565.
[15] T. Focken, G. Raabe, C. Bolm, Tetrahedron: Asymmetry 15 (2004)
1693.
[26] S. Pru¨hs, C.W. Lehmann, A. Fu¨rstner, Organometallics 23 (2004)
280.
[16] E. Bappert, G. Helmchen, Synletters (2004) 1789.
[17] S. Gischig, A. Togni, Organometallics 23 (2004) 2479.
[18] V.P.W. Bo¨hm, T. Weskamp, C.W.K. Gsto¨ttmayr, W.A. Herrmann, Angew.
Chem. Int. Ed. Engl. 39 (2000) 1602.
[19] J. Wolf, A. Labande, J.-C. Daran, R. Poli, J. Organomet. Chem. 691 (2006)
433.
[27] L.R. Hanton, P.R. Raithby, Acta Cryst. B 36 (1980) 2417.
[28] K.S. Coleman, H.T. Chamberlayne, S. Turberville, M.L.H. Green, A.R.
Cowley, Dalton Trans. (2003) 2917.
[29] A.J. Arduengo III, S.F. Gamper, M. Tamm, J.C. Calabrese, F. Davidson,
H.A. Craig, J. Am. Chem. Soc. 117 (1995) 572.
[30] The reaction of a Grignard reagent with 1-chloro-2,6-dimethylbenzene was
carried out but only gave 5–10% of the expected heterocoupling product,
which is comparable to the results found in the literature.
[20] A.J. Arduengo III, F.P. Gentry Jr., P.K. Taverkere, H.E. Simmons Jr., US
Patent 6,177,575 (2001).