Ionic liquids as reaction media for palladium-catalysed cross-coupling of aryldiazonium tetrafluoroborates with potassium organotrifluoroborates

Article abstract of DOI:10.1002/ejic.200400637

The system comprising a palladium complex in a 1-butyl-3-methylimidazolium tetrafluoroborate/methanol mixture efficiently catalyses the cross-coupling reaction between p-tolyl-diazonium tetrafluoroborate and potassium phenyltrifluoroborate at room temperature. The presence of methanol (or water) in the reaction mixture is necessary in order to achieve quantitative conversions, due to its scavenging behaviour towards the BF₃ formed during the reaction. Yields higher than 90% were obtained using Pd ₂(dba)₃ or the azapalladacycle 10 as the palladium source. With the latter complex a turnover frequency of about 6000 h^-1 was attained in the coupling of aryldiazonium tetrafluoroborates with potassium vinyltrifluoroborate. Recycling of the catalytic solution could be performed provided that a slight excess of diazonium salt was used in the first run. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2005.

Full text of DOI:10.1002/ejic.200400637

FULL PAPER
Ionic Liquids as Reaction Media for Palladium-Catalysed Cross-Coupling of
Aryldiazonium Tetrafluoroborates with Potassium Organotrifluoroborates
[a]
[a]
[a]
[a]
Vito Gallo, Piero Mastrorilli,* Cosimo F. Nobile, Rossella Paolillo, and
Nicola Taccardi[
a]
Keywords: C-C coupling / Diazonium salts / Homogeneous catalysis / Ionic liquids / Organotrifluoroborates
The system comprising a palladium complex in a 1-butyl-3-
methylimidazolium tetrafluoroborate/methanol mixture effi-
ciently catalyses the cross-coupling reaction between p-tolyl-
10 as the palladium source. With the latter complex a turno-
ver frequency of about 6000 h was attained in the coupling
of aryldiazonium tetrafluoroborates with potassium vinyltri-
−
1
diazonium tetrafluoroborate and potassium phenyltrifluoro- fluoroborate. Recycling of the catalytic solution could be per-
borate at room temperature. The presence of methanol (or formed provided that a slight excess of diazonium salt was
water) in the reaction mixture is necessary in order to achieve used in the first run.
quantitative conversions, due to its scavenging behaviour to-
wards the BF
3
formed during the reaction. Yields higher than
(dba) or the azapalladacycle
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
90% were obtained using Pd
2
3
Introduction
der suitable conditions, the use of imidazolium-based IL/
H O mixtures leads to increased reactivity and selectivity
2
Palladium-catalysed
cross-coupling
utilizing
or- compared to conventional reaction media and permits the
ganoboron compounds (the Suzuki reaction) is a powerful recycling of the catalytic solutions. Other catalytic systems
tool for the formation of carbonϪcarbon bonds.[1,2] The employed for the same reaction include PdCl
in alkylam-
(dba) in a
the wide range of functionalized substrates that can be used ternary system made up of tetradecyltrihexylphosphonium
2
[
14]
practical advantages of this reaction are mainly related to monium tetrafluoroborate/H
O mixtures, Pd
2
2
3
[
15]
under mild conditions. Typical reaction protocols exploit chloride, water and toluene
aryl iodides, bromides and triflates, which are not as cheap Pd(OAc) or bisimidazolin-2-ylidene palladium complex in
and readily available as aryl chlorides, although the latter IL/MeOH mixtures under ultrasonic irradiation.
and phosphane-free
2
[
16]
are less reactive. It has been demonstrated[
3Ϫ8]
that aryldia-
In the framework of our studies on metal-catalysed
zonium tetrafluoroborates derived from inexpensive and carbonϪcarbon bond forming reactions in ILs
[
17Ϫ20]
we de-
easily accessible anilines can be efficiently used in the cross- cided to investigate the palladium-catalysed cross-coupling
coupling with organoboronic acids or esters. Moreover, the of aryldiazonium salts with organotrifluoroborates. Fig-
cross-coupling has also been performed at room tempera- ure 1 shows the ILs used in this work.
ture and without added base using potassium organotri-
fluoroborates, which are more stable and reactive than the
corresponding boronic acid derivatives.[
9]
Ionic liquids (ILs) have emerged as potentially useful re-
action media in catalysis because of their peculiar physical
and chemical properties.[10] The latest efforts of the scien-
tific community have been addressed towards finding new,
possibly recyclable catalytic systems that exhibit higher ac-
tivity than with classical solvents.
Figure 1. Ionic liquids employed: [bmim]BF
4 6
, [bmim]PF , [bmim]-
N(SO CF , [bmmim]BF , [bbim]BF
2
3
)
2
4
4
The palladium-catalyzed Suzuki reaction between bo-
ronic acids and aryl halides has been successfully carried
out in ILs. Welton et al.[
11Ϫ13]
Results and Discussion
have demonstrated that, un-
Optimisation of the Solvent
[
a]
Dipartimento di Ingegneria delle Acque e di Chimica del
Politecnico di Bari,
The cross-coupling reaction between p-tolyldiazonium
tetrafluoroborate (1a) and potassium phenyltrifluoroborate
(2) was chosen as the model system (Reaction 1).
Via Orabona 4, 70125 Bari, Italy
Fax: ϩ 39-080-596-3611
E-mail: p.mastrorilli@poliba.it
582
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejic.200400637
Eur. J. Inorg. Chem. 2005, 582Ϫ588

Ionic Liquids as Reaction Media for Palladium-Catalysed Cross-Coupling
FULL PAPER
Table 1 summarises the results obtained when carrying
out reaction 1 in several ILs with 2.0 mol % Pd(OAc) as
2
catalyst, which is a palladium loading less than half of that
used in previous works.[
9]
Table 1. Solvent screening; reaction conditions: IL: 10 mmol;
Figure 2. Reaction course of Entry 1-1
2
Pd(OAc) : 0.02 mmol; 1a: 1.0 mmol; 2: 1.2 mmol
Entry
Solvent
Time
h]
Conv.[a]
(%)
Isolated yield
(%)
[
1-1
2-1
3-1
4-1
5-1
[bmim]BF
[bmim]N(SO
[bmim]PF
[bmmim]BF
[bbim]BF
4
6
6
14
12
12
51
60
50
53
45
49
38
41
45
34
This reaction could therefore be responsible for unsatis-
2
CF
3
)
2
factory conversions as the formed PhBF is inactive in Su-
2
6
4
zuki cross-coupling.
4
The detrimental effect of BF on the reaction course was
3
circumvented by using MeOH as a BF scavenger. In fact,
[
a]
3
Calculated at the time at which a plateau was reached in the
reactions performed in [bmim]BF /MeOH mixtures gave
consumption of the diazonium salt.
4
quantitative conversions. Figure 3 shows the effect of the
molar fraction (xMeOH) of added methanol on the catalysis
The results show that there is an effect of the IL used on
performed in [bmim]BF . Isolated yields of 3a and reaction
the course of the reaction. Using [bmim]BF (bmim ϭ 1-n-
4
4
times are dependent on the molar fraction of MeOH. At
xMeOH ϭ 0.25 the reaction reaches completion in 120 min
with a 68% yield of 3a. The discrepancy between yield and
conversion is due to decomposition of the diazonium salt,
since only traces of homo-coupling products (biphenyl and
butyl-3-methylimidazolium) gave 51% conversion after 6 h
with 96% selectivity towards 3a (Entry 1-1). Catalysis in
[
bmim]N(SO CF ) was more efficient (60% conv. after 6 h)
2 3 2
but significantly less selective (63%, Entry 2-1), whereas
when [bmim]PF was used as solvent a drop in both activity
6
4,4Ј-dimethylbiphenyl) were formed. At xMeOH ϭ 0.75 a
(
(
50% conv. after 14 h) and selectivity (82%) was observed
Entry 3-1 vs. Entry 1-1). The effect of the cation of the IL
67% yield of 3a was achieved along with the formation of
biphenyl (25%) and 4,4Ј-dimethylbiphenyl (6%) by-prod-
ucts. The reaction carried out in pure MeOH was very fast
and exothermic but poorly selective, giving only 47% yield
of 3a, 30% of biphenyl and 16% of 4,4Ј-dimethylbiphenyl.
Hence, MeOH has the side effect of accelerating the de-
composition of the diazonium salt over long reaction times
when used at low molar fractions; at high xMeOH the selec-
tivity of the reaction is strongly reduced, mainly owing to
formation of homocoupling by-products. Carrying out the
reaction with a xMeOH of 0.50 gave the best result (83%
yield of 3a after 30 min). In this case the formation of
is shown in Entries 1-1, 4-1 and 5-1, from which it is appar-
ent that comparable yields of 3a were obtained after signifi-
cantly different times. Among all the ILs tested [bmim]BF4
gave the best results in terms of both activity and selectivity
and was chosen as solvent for further investigations.
A common point which emerges from Table 1 is that,
after variable times, the reactions did not reach completion
but stopped when the conversion was about 50Ϫ60%. Fig-
ure 2 shows the reaction course in the case of Entry 1-1.
Such a behavior could, in theory, be ascribed to two
causes: a) catalyst deactivation, or b) formation of a prod-
uct that inhibits the reaction.
The first hypothesis was ruled out by the observation that
the conversion did not improve significantly when per-
forming the reaction with a higher amount of palladium:
carrying out the catalysis with 5 or 10 mol % Pd resulted in
65% or 70% conversion, respectively.
In search of the reaction product that could act as inhibi-
tor for the reaction we focused our attention on BF . It is
3
[
21,22]
known
that BF reacts with 2 in halogenated solvents
3
to afford potassium tetrafluoroborate and phenyldifluoro-
borane (Reaction 2).
Figure 3. Effect of the MeOH molar fraction
Eur. J. Inorg. Chem. 2005, 582Ϫ588
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
583

V. Gallo, P. Mastrorilli, C. F. Nobile, R. Paolillo, N. Taccardi
FULL PAPER
homo-coupling products was negligible and the decompo- 44% yield of 3a while the 1,2-bis(diphenylphosphanyl)-
sition of diazonium salt minimised.
ethane complex 5 was almost inactive (Entries 5-2 and 6-2).
Conversely, dinuclear phosphapalladacycles were found to
be active catalysts for reaction 1: complex 6 gave 64% con-
version together with 54% yield after 22 h (Entry 7-2), while
Palladium Source
Results obtained using several Pd sources in [bmim]BF4/
MeOH (xMeOH ϭ 0.50) as catalysts for reaction 1 are sum-
marized in Table 2. A comparison of Entry 1-2 with Entry
7
gave a quantitative conversion and 86% yield in a shorter
reaction time (Entry 8-2). The dinuclear azapalladacycles
Ϫ10 were found to be the most active catalysts. In all cases
8
2
-2 clearly shows the strong influence exerted by the anion
II
quantitative conversions were obtained in 30 min or less,
with selectivities towards the desired product 3a ranging
from 65% obtained with the 8-methylquinoline derivative 8
of the Pd salt on the catalysis: PdCl gave only 40% con-
2
version and 15% yield of 3a after 28 h. The reactions involv-
0
ing Pd complexes gave nearly quantitative conversions. In
(Entry 9-2) to 93% obtained with the azobenzene derivative
particular, Pd (dba) gave a higher yield (90%) than that
2
3
10 (Entry 11-2). A remarkably short reaction time was ob-
(
63%) obtained with Pd(PPh ) , but needed a longer reac-
3 4
served with the iminopalladacycle 9, which afforded an 84%
yield in less than 10 min (Entry 10-2). Complex 10 was also
tion time (Entries 3-2 and 4-2).
Table 2. Screening of palladium source; reaction conditions: de- used for comparison in pure methanol, giving only 55%
gassed solvent mixture IL/MeOH (xMeOH ϭ 0.50): 2.0 mL; Pd:
yield after 15 min (Entry 12-2).
0.02 mmol; 1a: 1.0 mmol; 2: 1.2 mmol
Pd(OAc) and complexes 6Ϫ10 were also used in an equi-
2
molar mixture of [bmim]BF and H O. In all cases quanti-
4
2
Entry
Catalyst
Pd(OAc)
Time
min]
Conv.
(%)
Isolated yield
(%)
tative conversions were achieved owing to the ability of
water to act as a BF scavenger. Pd(OAc) gave 65% yield
[
3
2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
-2
-2
-2
-2
-2
-2
-2
-2
2
30
100
40
95
100
80
23
64
98
100
100
100
100
96
83
15
90
63
44
traces
54
86
65
84
93
55
87
of 3a after 2 h whereas the yields obtained with complexes
and 7 were 85% and 87%, respectively, after 20 h. Com-
PdCl
Pd (dba)
2
1680
330
150
1800
1260
1320
900
30
10
30
15
120
720
6
2
3
·CHCl
3
plexes 8Ϫ10 gave 64%, 60% and 70% isolated yields, respec-
tively, in less than 30 min. Figure 5 shows a comparison of
Pd(PPh
3
)
4
4
5
the results obtained with Pd(OAc) and complexes 6Ϫ10 in
2
6
IL/MeOH with those in IL/H O. All the catalysts shown in
2
7
Figure 5, except phosphapalladacycles 6 and 7, were more
selective when the cosolvent was MeOH.
-2
8
0-2
1-2
2-2
3-2
4-2
9
10
[
[
[
a]
b]
c]
10
10
10
83
67
[
a]
[b]
Pure methanol (4 mL) as solvent.
Solvent mixture: 4 mL;
Ϫ3
[c]
Pd: 5.0 ϫ 10 mmol; 1a: 2.0 mmol; 2: 2.4 mmol.
1
Pd: 1.0 ϫ
Ϫ3
0
mmol.
Several palladium(ii) acetate derivatives (Figure 4) were
tested. Mononuclear complexes bearing chelating ligands
gave unsatisfactory results: 2,2Ј-bipyridyl complex 4 gave a
4
Figure 5. Performance of catalysts 4Ϫ10 in equimolar [bmim]BF /
cosolvent mixtures
Palladium Loading and Recycling
The effect of palladium loading was investigated using
complex 10 in an equimolar [bmim]BF /MeOH mixture. At
4
0.5% Pd loading (0.25 mol % of the dinuclear complex) the
catalyst was still active and selective, reaching nearly quan-
titative conversion with an 87% yield after 2 h (Entry 13-
2
8
); at 0.1% loading the reaction afforded 67% of 3a with an
3% conversion after 12 h (Entry 14-2).
Optimization tests aimed at recycling the catalytic solu-
tion, performed with complex 10 in IL, showed that an ex-
cess of the aryldiazonium 1a was necessary in order to ob-
tain a recyclable catalytic solution. When a 20% excess of
phenyltrifluoroborate 2 was employed the second cycle did
Figure 4. Palladium acetate derived catalysts; Naph ϭ naphthyl
584
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 582Ϫ588

Ionic Liquids as Reaction Media for Palladium-Catalysed Cross-Coupling
FULL PAPER
not reach completion (60% conv. after 3 d), giving only 37% Table 3. Reaction conditions: degassed solvent mixture LI/MeOH
of 3a; the catalyst decomposed into Pd black (formation of (xMeOH ϭ 0.50): 2.0 mL; Pd: 0.02 mmol; 1a: 1.0 mmol; 2: 1.2 mmol
Pd black during the reaction was observed in almost all the
Entry
Catalyst
Time
min]
Isolated yield
(%)[
reactions described so far). On the other hand, when a 20%
excess of 1a was used no Pd black formed after the first
run and the catalytic solution was recyclable (Figure 6). The
excess of 1a renders the catalytic system less efficient as, 2-3
with an excess of 2, a 93% yield was achieved after 30 min 3-3
a]
[
-3[b]
1
10
11
12
13
10
30
85
65
45
55
90
260
1600
480
40
4
5
-3[
-3
(Entry 11-2), whereas with 1a in excess a 72% yield was
c]
achieved after 2 h.
[
[
a]
Quantitative conversions. [b] Hg added; Hg/Pd ϭ 150 mol/mol.
Reaction performed in [bmmim]BF /MeOH.
4
c]
Among the possible candidates for the active species in
the present reaction we focused our attention on (car-
II
[10]
bene)Pd species. According to Wasserscheid, three sim-
ple experiments should be carried out to verify the interme-
diacy of carbene complexes as catalytically active species in
ILs, namely variation of the ligands, tests with indepen-
II
dently prepared, well-defined (carbene)Pd complexes, and
reactions in ILs that are not capable of carbene formation.
Replacing the bridging acetates with bridging bromides
Figure 6. Recyclability of the catalytic solution (initial 1a/2 molar
ratio ϭ 1.2)
(
(
Figure 7) in complex 10 resulted in a drop of activity
Entry 11-2 vs 2-3); this might be rationalised by invoking
the inability of bromide to generate carbene species by ab-
straction of the acidic proton in the 2-position of the imida-
zolium ring. However, reactions performed with the pre-
formed carbene complexes 12 and 13 (Entries 3-3 and 4-3)
were much slower than that performed with complex 10
Possible Active Species: Preliminary Investigations
Mechanistic investigations on Suzuki coupling involving
aryldiazonium salts have not been performed so far. How-
ever, different active species have been suggested for ‘‘classi-
cal’’ Suzuki reactions in ILs. Zou et al. have proposed that
the catalytic activity observed in alkylammonium tetra-
(Entry 11-2). This seems to indicate that, even if they form
in the reaction mixture, carbene species are not mainly
responsible for the catalytic activity observed with aza-
palladacycles. Accordingly, replacing [bmim]BF4 with
[
bmmim]BF , an imidazolium-based IL that is not capable
4
fluoroborate/H O mixtures is due to palladium metal stabil-
2
of carbene formation due to the presence of the methyl
group in the 2-position, gave essentially equal activity and
selectivity (see Entries 11-2 and 5-3).
ised by the reaction medium.[14] On the other hand, Srini-
vasan’s group has ruled out colloidal Pd nanoparticles as
the active species in the Suzuki reaction performed in imid-
azolium-based IL/MeOH mixtures.[ Welton et al. have
16]
[12]
suggested the intermediacy of mixed (carbene)(phos-
II
phane)Pd complexes formed by reaction of Pd(PPh3)4
with [bmim]BF in the presence of aryl halide and a halide
4
salt under basic conditions.
Although palladium nanocolloids are known to catalyze
the strictly related Heck reaction in ILs,[ in our case the
progressive formation of palladium black observed nearly
invariably during catalysis is more probably the conse-
quence of catalyst decomposition, since no Pd black formed
23]
Figure 7. (Azobenzene)PdII complexes
The above-mentioned detrimental effect of bromides on
in the catalytic solutions that could be successfully recycled. catalysis could, therefore, be due to the strong coordinating
That palladium metal is not involved in the catalytic act ability of halide species towards palladium. This strong
is further demonstrated by poisoning experiments with Hg tendency to bind palladium could also be held responsible
[
24]
metal. When performing reaction 1 catalyzed by complex for the difference in activity between bromo complex 12 and
0 in the presence of added Hg (Hg/Pd ϭ 150 mol/mol) acetonitrile complex 13, and can also rationalize the slug-
no inhibition was observed (Entry 1-3, Table 3). This result gishness of the reaction catalyzed by PdCl (Entry 2-2).
1
2
suggests that palladium nanocolloids are not involved in At this stage, the following points seem to emerge: (i)
the catalytic cycle and that a homogeneous mechanism catalysis is molecular and influenced by the ligands on the
should be operating.
palladium atom; (ii) the presence of easily dissociable li-
Eur. J. Inorg. Chem. 2005, 582Ϫ588
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
585

V. Gallo, P. Mastrorilli, C. F. Nobile, R. Paolillo, N. Taccardi
FULL PAPER
gands in palladium(ii) precatalysts has a beneficial effect
on catalysis (compare activities of acetate complex 10 vs.
bromide complex 11 or of acetonitrile complex 13 vs. bro-
mide complex 12). Although no definitive conclusion can
be drawn about the active species that triggers the catalysis
starting from azapalladacycle 10, these observations lead us
0
II
to suggest, on the basis of the commonly accepted Pd /Pd
0
cycle, the anionic Pd complex depicted in Figure 8 as the
possible active species. Such a species, which could be gen-
erated by reduction of 10 by 2, might start the catalytic
cycle by forming an arylpalladium complex by reaction
with the diazonium salt.
Figure 8. Possible active species in reactions catalysed by complex
10
Anionic Pd0 complexes have been suggested as active
species in related cross-coupling reactions. Amatore et
Conclusion
This study demonstrates that ILs are suitable solvents for
the title reaction. Used neat, ILs give high selectivities for
the cross-coupling product. High activity and selectivity
were shown by catalytic systems comprised of azapallada-
al.,[
25]
for example, have demonstrated that [Pd(PPh ) -
3
2
Ϫ
(
OAc)] , formed from Pd(OAc) and PPh , is an active
2
3
species for the catalytic cycle of Heck reactions between
aryl halides and olefins, and Hermann et al.[ have pro-
posed the palladacycle [(di-o-tolylphosphanyl)benzyl]-
palladate(0) as the active species in Heck couplings cata-
lyzed by complex 6.
26]
cycles in [bmim]BF /MeOH mixtures. Using a slight excess
4
of diazonium salt with respect to aryltrifluoroborate al-
lowed the recycling of the catalytic solution. Preliminary
mechanistic studies carried out with the precatalyst 10 indi-
cate that a homogeneous process takes place and that car-
bene species are not the main active species. Further studies
on this subject are in progress.
Screening of Substrates
The reactivity of several substrates was explored
(Table 4). Substituted aryldiazonium salts (Reaction 3) gave
fair to good yields: 4-methoxybiphenyl was obtained with
a 78% yield after 3 h (Entry 1-4) while aryldiazonium salt
Experimental Section
1c, containing an electron-withdrawing group (Entry 2-4),
reacted faster to give the relevant biphenyl in 82% yield. General: Unless otherwise specified all the manipulations were con-
ducted under an inert gas (nitrogen) using standard Schlenk tech-
Halogen-substituted substrates 1d and 1e were less reactive
niques. Flash-chromatography was performed on SiO
2
Kieselgel
and required 4% palladium to give quantitative conversions
and 79% and 95% yields, respectively (Entries 3-4 and 4-4).
Vinyltrifluoroborates were far more reactive than the aro-
matic ones (Reaction 4): quantitative conversions and yields
2
30Ϫ400 mesh. The NMR spectra were recorded with a Bruker
1
AX400 spectrometer (400 MHz for H) at 295.0 K; chemical shifts
are reported in ppm referenced to SiMe for H and C, and CFCl
4 3
1
13
19
for F. LC-MS analyses of complexes 11, 12 and 13 were per-
formed with an Agilent HPLC system equipped with DAD, auto-
sampler and MS systems (Agilent 1100 LC-MS SL series). All
samples were dissolved in the same HPLC-grade solvent that was
used as eluent. The used interfaces were APCI for neutral com-
plexes 11 and 12, and ESI for cationic complex 13. APCI con-
ditions: positive-ion mode, flow rate 0.5 mL/min, nitrogen as nebul-
izing and drying gas, nebulizer pressure 4 atm, vaporizer tempera-
ture 350 °C, corona current 4.0 µA, drying-gas flow 5 L/min, dry-
ing-gas temperature 350 °C, capillary voltage 4000 V. ESI
conditions: positive-ion mode, flow rate 0.3 mL/min, nitrogen as
nebulizing and drying gas, nebulizer pressure 2 atm, capillary volt-
age 4000 V. Mass spectrometry data were acquired in the scan
mode (mass range m/z ϭ 50Ϫ3000). C, H, N elemental analyses
were carried out with a Eurovector CHNS-O Elemental Analyser.
Pd analyses were performed with a PerkinϪElmer SIMAA6000
Atomic Absorption spectrometer. The IR spectra were recorded
with a Bruker Vector 22 FT-IR spectrometer. GC analyses were
of up to 83% were obtained after 10 min using as little as
Ϫ1
0.1% palladium (turnover frequency ഠ 6000 h ; Entries
5-4 and 6-4).
Table 4. Screening of substrates; reaction conditions: degassed sol-
vent mixture LI/MeOH (xMeOH 0.50): 2 mL; ArN BF
.0 mmol; KRBF : 1.2 mmol; quantitative yields were obtained
ϭ
2
4
:
1
3
Entry ArN
2
BF
4
3
K(RBF ) Pd loading Time Product Isolated yield
[mmol]
[min]
(%)
1
2
3
4
5
6
-4 1b
-4 1c
-4 1d
-4 1e
-4 1a
-4 1f
2
0.02
0.02
0.04
0.04
0.001
0.001
180 3b
78
82
79
95
83
80
2
5
3c
2
2
14
14
180 3d
3
10
10
3e
15a
15f
586
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 582Ϫ588

Ionic Liquids as Reaction Media for Palladium-Catalysed Cross-Coupling
FULL PAPER
performed with an HP5890 instrument equipped with a Supelco
SPB-1 capillary column (30 m ϫ 320 µm ϫ 0.25 µm). GC-MS
analyses were performed with an HP6890-HP5973MSD instrument
202Ϫ205°. IR (KBr): ν˜ max ϭ 3156 (w), 3124 (m), 3042 (w), 2954
(m), 2929 (m), 2870 (m), 1574 (m), 1551 (m), 1460 (m), 1445 (m),
1388 (m), 1254 (m), 1233 (m), 1204 (w), 1107 (w), 774 (s), 731 (m),
Ϫ1
1
equipped with an HP-5MS capillary column (30 m ϫ 250 µm ϫ 716 (m), 691 (s), 584 (w), 547 (w), 522 (w) cm
.
H NMR
), 1.25 (m, 2 H,
), 3.78 (s, 3 H, NCH ),
), 6.12 (d, J ϭ 7.92 Hz, 1 H), 7.00 (m, 3 H),
0.25 µm). The halide content of the ionic liquids was determined
(CD
CH
4.18 (m, 2 H, NCH
2
Cl
CH
2
): δ ϭ 0.80 (t, J ϭ 7.31 Hz, 3 H, CH
2 3
CH
in aqueous solution by a potentiometric automatic titration using
a Metrohm 716 DMS Titrino (300 mg sample, ca. 30 mL deionized
2
2
CH ), 1.77 (m, 2 H, CH CH CH
3
2
2
2
3
2
water, three runs) while the water content was determined by the 7.16 (t, J ϭ 7.31 Hz, 1 H), 7.41 (m, 3 H), 7.94 (d, J ϭ 6.7 Hz, 1
1
3
KarlϪFischer method using
a
Metrohm 716 DMS Titrino
H), 7.99 (m, 2 H, NCH) ppm. C NMR (CD
(CH ), 19.78 (CH CH CH ), 32.20 (CH CH
(NCH CH ), 51.12 (NCH ), 121.44, 122.54, 124.66, 125.14, 127.86,
130.35 (NCH), 130.48 (NCH), 132.57, 135.45, 151.93, 155.82,
CN/CH OH 99:1): exact
Pd: 504.00; found 505 [M ϩ H] , 465
CN] , 424 [M Ϫ Br]. C20 23BrN Pd: calcd. C
2
Cl
2
): δ ϭ 13.40
equipped with a 703 titration stand using sodium tartrate-2-hydrate
as standard and dry methanol as solvent (1 g sample, ca. 30 mL
MeOH, three runs). All commercial reagents were used as received
3
2
2
3
2
2
CH ), 38.40
2
2
2
3
without further purification. Solvents were dried and distilled 164.54, 169.35 (NCN) ppm. LC-MS (CH
3
3
ϩ
under nitrogen according to standard procedures; water (resistivity
mass calcd. for C20
[M Ϫ Br ϩ CH
H23BrN
4
Ϫ1
ϩ
Ͼ 17.5·MΩ cm ) was deionized with a Millipore Simplicity unit.
3
H
4
[
27]
[27]
[28]
[29]
[bmim]Cl,
[bmmim]Cl,
[bbim]Br
and [bmim]Br
were 47.50, H 4.58, N 11.08, Pd 21.04; found C 47.78, H 4.75, N 10.86,
synthesised from 1-chlorobutane or 1-bromobutane and the rel-
evant imidazole salt according to literature procedures. [bmim]Cl,
Pd 20.96.
Acetonitrile(1-butyl-3-methylimidazolin-2-ylidene)[o-(phenylazo)-
phenyl]palladium(II
[bmmim]Cl and [bmim]Br were purified by crystallisation from
)
Tetrafluoroborate (13): AgBF
4
(0.110 g,
acetonitrile/ethyl acetate (1:4) until a colourless mother liquor was
obtained, and then filtered and dried under vacuum; all com-
pounds were obtained as white solids. [bbim]Br was obtained as a
0.57 mmol) dissolved in dry acetonitrile (5 mL) was added to a
THF (30 mL) solution of complex 12 (0.286 g, 0.57 mmol). After
stirring for 45 min, the solvent was evaporated in vacuo. The pale-
orange residue was extracted with dry dichloromethane (30 mL)
and the resulting suspension filtered through Celite. Crystallization
from THF/hexane gave 13 as a pale-orange powder (0.276 g, 88%)
m.p. 168° (dec.). IR (nujol mull): ν˜ max ϭ 3167 (w), 3136 (w), 2318
viscous brown liquid, washed with CH
2 2
Cl /diethyl ether (1:4) and
dried under vacuum. All compounds were stored under nitrogen.
[
[
bmim]BF
bbim]BF
4
,
[bmim]PF
6
,
[bmim]N(SO
2
CF
3
)
2
,
4
[bmmim]BF ,
4
were synthesised by metathesis from the corresponding
alkali metal salts in dichloromethane; the resulting suspension was
filtered through Celite and washed with the minimal amount of
(w) [ν(CϵN)], 2290 (w) [ν(CϵN)], 1575 (w), 1555 (w), 1463 (s),
1401 (w), 1377 (m), 1257 (w), 1236 (m), 1056 (vs) [ν(BϪF)], 773
water until the AgNO
solvent was then evaporated and the resulting IL was dried at
0Ϫ50 °C under vacuum for 48 h and stored under nitrogen. The
3
test on the aqueous layer was negative. The
Ϫ1
1
(
(
m), 713 (m), 696 (m), 588 (w), 548 (w), 523 (w) cm . H NMR
CD Cl ): δ ϭ 0.81 (br. t, J ϭ 7.31 Hz, 3 H, CH CH ), 1.25 (br.
CH CH ), 1.76 (br. quint, J ϭ 7.31 Hz, 2
), 2.13 (s, 3 H, NCCH ), 3.83 (s, 3 H, NCH ), 4.15
br. t, J ϭ 7.31 Hz, 2 H, NCH ), 6.03 (d, J ϭ 7.31 Hz, 1 H), 7.00
2
2
2
3
4
sext, J ϭ 7.31, 2 H, CH
H, CH CH CH
(
2
2
3
ionic liquids were obtained as pale-yellow or colourless liquids and
fully characterised by multinuclear NMR spectroscopy. Their water
content was less than 460 ppm whereas their halide content was
2
2
2
3
3
2
[
30]
[31]
[26]
(t, J ϭ 7.31 Hz, 1 H), 7.14 (s, 1 H, NCH), 7.18 (s, 1 H, NCH), 7.22
(t, J ϭ 7.31 Hz, 1 H), 7.51 (br. s, 3 H), 7.72 (m, 2 H), 7.97 (d, J ϭ
below the detection limit of 0.38 mg/L. Complexes 4, 5, 6,
[
32]
[33] [34]
[35]
7,
8,
9
and 10
were synthesised according to literature
1
3
7
.31 Hz, 1 H) ppm. C NMR (CD
CH CH ), 19.62 (CH CH CH ), 32.64 (CH
NCH CH ), 51.18 (NCH ), 122.48, 122.75, 123.98, 129.03, 131.51
NCH), 131.62 (NCH), 133.25, 136.17, 151.22, 151.56, 164.92,
65.04 ppm. LC-MS (CH CN): exact mass calcd. for the cation
2
Cl
2
): δ ϭ 2.78 (NCCH
3
), 13.35
procedures.
(
(
(
2
3
2
2
3
2
CH CH ), 38.45
2
2
Di-µ-bromo-bis[o-(phenylazo)phenyl]dipalladium(II) (11): Complex
(0.173 g, 0.25 mmol) and tetrabutylammonium bromide
0.167 g, 0.52 mmol) were dissolved in dry dichloromethane (10
2
2
3
10
1
3
(
ϩ
ϩ
C
C
22
H
H
26
N
5
Pd: 466.12; found 466 [M] , 425 [M Ϫ CH
3
CN] .
mL). After stirring for 3 h, the volume was reduced in vacuo to a
half and dry methanol (50 mL) was added. The resulting red solid
was filtered off, washed with methanol (3 ϫ 5 mL) and hexane (15
mL) and dried in vacuo to afford 11 as a red powder (0.151 g,
22
4 5
26BF N
Pd: calcd. C 47.72, H 4.73, N 12.65, Pd 19.22; found
C 47.61, H 4.85, N 12.74, Pd 19.12.
Substrates: The aryldiazonium tetrafluoroborates 1aϪ1f were syn-
8
(
1
2%); m.p. Ͼ 250 °C. IR (KBr): ν˜ max ϭ 1575 (m), 1553 (w), 1482
w), 1458 (w), 1444 (w), 1393 (m), 1318 (w), 1306 (w), 1263 (w),
242 (w), 762 (m), 753 (m), 706 (w), 691 (m), 595 (w), 550 (w), 525
[36]
1
thesised as reported in the literature and characterised by H and
1
3
3
C NMR (CD CN) and IR (nujol mull) spectroscopy. Potassium
aryltrifluoroborate 2 and vinyltrifluoroborate 14 were also syn-
Ϫ1
1
(w) cm . H NMR ([D]
6
DMSO): δ ϭ 7.07 (t, J ϭ 7.31 Hz, 1 H),
thesised according to a literature procedure.[
9]
7
.17 (t, J ϭ 7.31 Hz, 1 H), 7.44 (m, 3 H), 7.65 (m, 2 H), 7.86 (d,
13
J ϭ 7.31 Hz, 1 H), 7.89 (d, J ϭ 7.92 Hz, 1 H) ppm. C NMR
[D] DMSO): δ ϭ 124.36, 126.06, 128.84, 130.24, 131.38, 132.22,
39.12, 150.88, 153.88, 164.13 ppm. LC-MS (CH CN/CH Cl , 9:1):
exact mass calcd. for C24 Pd : 731.80; found 732 [M] ,
Pd : calcd. C 39.21, H 2.47, N
.62, Pd 28.95; found C 39.65, H 2.70, N 7.41, Pd 28.79.
General Procedure for the Catalytic Reactions: All catalytic runs
were repeated at least three times and yields and conversions were
reproducible within Ϯ3%. The course of the reactions was moni-
(
1
6
3
2
2
ϩ
H
18Br
2
N
4
2
2
tored by measuring the evolution of N vs. time by means of a
ϩ
7
7
73 [M ϩ CH
3
CN] . C24
H18Br
2
N
4
2
mercury gas-burette. Reactions were stopped when no further gas
evolution was observed. In neat ILs, the decomposition of di-
azonium salt was negligible (4Ϫ5% per day) as demonstrated by
blank tests.
Bromo(1-butyl-3-methylimidazolin-2-ylidene)[o-(phenylazo)phenyl]-
palladium(II) (12): Complex 10 (0.39 g, 0.56 mmol) and [bmim]Br
(0.25 g, 1.13 mmol) were dissolved in dry THF (50 mL) in a round-
Typical Reaction in Neat IL: Compounds 2 (221 mg, 1.2 mmol) and
bottomed flask. After refluxing under nitrogen for 3 h and cooling 1a (206 mg, 1.0 mmol) were added, with vigorous stirring, to a sus-
to ambient temperature, the solvent was evaporated in vacuo to
give an orange solid. Crystallization from acetone/hexane afforded
2 4
pension of Pd(OAc) (4.48 mg, 0.02 mmol) in [bmim]BF (2.260 g,
10 mmol) in a 10-mL Schlenk tube. At the end of reaction, the
mixture was diluted with water (100 mL) and the products were
12 as a dark-orange, microcrystalline solid (0.470 g, 82%); m.p.
Eur. J. Inorg. Chem. 2005, 582Ϫ588
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
587

V. Gallo, P. Mastrorilli, C. F. Nobile, R. Paolillo, N. Taccardi
FULL PAPER
[
[
[
[
13]
14]
15]
16]
extracted with diethyl ether (3 ϫ 25 mL); the organic phases were
collected and dried with Na SO . The products were purified by
2 4
flash chromatography and characterised by GC-MS and the selec-
tivity assessed by GC using naphthalene as internal standard.
C. J. Mathews, P. J. Smith, T. Welton, J. Mol. Catal. A 2004,
214, 27Ϫ32.
G. Zou, Z. Wang, J. Zhu, J. Tang, M. Y. He, J. Mol. Catal. A
003, 206, 193Ϫ198.
2
J. McNulty, A. Capretta, J. Wilson, J. Dyck, G. Adjabeng, A.
Robertson, Chem. Commun. 2002, 1986Ϫ1987.
Reactions in [bmim]BF
mixture was poured into a 10-mL Schlenk tube containing Pd cata-
lyst (0.02 mmol of Pd), KRBF (1.2 mmol) and ArN BF
1.0 mmol) and stirred vigorously.
4
/Cosolvent: 2 mL of the degassed solvent
R. Rajagopal, D. V. Jarikote, K. V. Srinivasan, Chem. Commun.
2002, 616Ϫ617.
3
2
4
[17]
M. M. Dell’Anna, V. Gallo, P. Mastrorilli, C. F. Nobile, G.
Romanazzi, G. P. Suranna, Chem. Commun. 2002, 434Ϫ435.
P. Mastrorilli, C. F. Nobile, V. Gallo, G. P. Suranna, G. Farin-
ola, J. Mol. Catal. A 2002, 184, 73Ϫ78.
[19]
(
[
18]
Recycle: Compounds 2 (184 mg, 1.0 mmol) and 1a (247 mg,
1.2 mmol) were added, with vigorous stirring, to a suspension of
V. Gallo, P. Mastrorilli, C. F. Nobile, G. Romanazzi, G. P. Sur-
anna, J. Chem. Soc., Dalton Trans. 2002, 4339Ϫ4342.
P. Mastrorilli, C. F. Nobile, R. Paolillo, G. P. Suranna, J. Mol.
Catal. A 2004, 214, 103Ϫ106.
1
0 (6.93 mg, 0.01 mmol) in [bmim]BF (2.260 g, 10 mmol) and
4
[
[
[
20]
21]
22]
MeOH (405 µL 10 mmol) in a 10-mL Schlenk tube. At the end of
reaction, the MeOH was evaporated off in vacuo and the products
were extracted with dry n-hexane (6 ϫ 3 mL); after evaporation of
the solvent, fresh MeOH (405 µL, 10 mmol), 2 (184 mg, 1.0 mmol)
and 1a (206 mg, 1.0 mmol) were added to the resulting catalytic
solution.
H. J. Frohn, F. Bailly, V. V. Bardin, Z. Anorg. Allg. Chem. 2002,
6
28, 723Ϫ724.
H. J. Frohn, H. Franke, P. Fritzen, V. V. Bardin, J. Organomet.
Chem. 2000, 598, 127Ϫ135.
N. A. Hamill, C. Hardacre, S. E. J. McMath, Green Chem.
[23]
2
002, 4, 139Ϫ142.
[24]
0
The suppression of catalysis by Hg is often used distinguish
0
heterogeneous from homogeneous catalysis since Hg is known
Acknowledgments
to poison metal(0) catalysts by formation of amalgams: Y. Na,
S. P. S. B. Han, H. Han, S. Ko, S. Chang, J. Am. Chem. Soc.
2004, 126, 250Ϫ258 and references cited therein.
Dr. G. Ciccarella is gratefully acknowledged for performing the
LC/MS experiments. We also thank Mr. G. Di Gregorio for exper-
imental assistance and Engelhard for a generous loan of palladium
acetate. The Polytechnic of Bari is acknowledged for financial sup-
port (F.R.A. ex 60%).
[
[
[
25]
26]
27]
C. Amatore, A. Jutand, J. Organomet. Chem. 1999, 576,
2
54Ϫ278.
V. P. W. Böhm, W. A. Herrmann, Chem. Eur. J. 2001, 7,
4
191Ϫ4197.
L. Cammarata, S. G. Kazarian, P. A. Salter, T. Welton, Phys.
Chem. Chem. Phys. 2001, 3, 5192Ϫ5200.
R. R. Deshmukh, R. Rajagopal, K. V. Srinivasan, Chem. Com-
mun. 2001, 1544Ϫ1545.
P. Bonh oˆ te, A. P. Dias, N. Papageorgiou, K. Kalyanasunda-
ram, M. Grätzel, Inorg. Chem. 1996, 35, 1168Ϫ1178.
T. A. Stephenson, S. M. Morehouse, A. R. Powell, J. P. Heffer,
G. Wilkinson, J. Chem. Soc. 1965, 3632Ϫ3640.
C. Bianchini, H. M. Lee, A. Meli, W. Oberhauser, M. Peruz-
zini, F. Vizzi, Organometallics 2002, 21, 16Ϫ33.
B. L. Shaw, S. D. Perera, E. A. Staley, Chem. Commun. 1998,
1361Ϫ1362.
[
[
[
1]
2]
3]
A. Suzuki, N. Miyaura, Chem. Rev. 1995, 95, 2457Ϫ2483.
A. Suzuki, J. Organomet. Chem. 1999, 576, 147Ϫ168.
S. Darses, T. Jeffery, J. L. Brayer, J. P. Demoute, J. P. Genet,
Bull. Soc. Chim. Fr. 1996, 133, 1095Ϫ1102.
S. Darses, T. Jeffery, J. P. Genet, J. L. Brayer, J. P. Demoute,
Tetrahedron Lett. 1996, 37, 3857Ϫ3860.
S. Sengupta, S. Bhattacharyya, J. Org. Chem. 1997, 62,
[28]
[
[
[
[
29]
30]
31]
32]
[
[
[
4]
5]
6]
3
405Ϫ3406.
D. M. Willis, R. M. Strongin, Tetrahedron Lett. 2000, 41,
6271Ϫ6274.
[
[
7]
8]
M. B. Andrus, C. Song, Org. Lett. 2001, 3, 3761Ϫ3764.
K. Selvakumar, A. Zapf, A. Spannenberg, M. Beller, Chem.
Eur. J. 2002, 8, 3901Ϫ3906.
[33]
[34]
[35]
[36]
A. J. Deeming, I. P. Rothwell, J. Organomet. Chem. 1981,
205, 117Ϫ131.
[
9]
S. Darses, G. Michaud, J. P. Genet, Eur. J. Org. Chem. 1999,
J. Granell, D. Sainz, J. Sales, X. Solans, M. Font-Altaba, J.
Chem. Soc., Dalton Trans. 1986, 1785Ϫ1790.
I. Aiello, A. Crispini, M. Ghedini, M. LaDeda, F. Barigelletti,
Inorg. Chim. Acta 2000, 308, 121Ϫ128.
1
875Ϫ1883.
[
[
[
10]
11]
12]
Ionic Liquids in Synthesis (Eds.: P. Wasserscheid, T. Welton),
Wiley-VCH, Weinheim, 2003.
C. J. Mathews, P. J. Smith, T. Welton, Chem. Commun. 2000,
1
A. Roe, Organic Reactions, Coll. Vol. 1949, 5, 193Ϫ228.
249Ϫ1250.
Received July 20, 2004
Early View Article
Published Online December 6, 2004
F. McLachlan, C. J. Mathews, P. J. Smith, T. Welton, Or-
ganometallics 2003, 22, 5350Ϫ5357.
588
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 582Ϫ588