Electroreductive palladium-catalysed Ullmann reactions in ionic liquids: Scope and mechanism

Article abstract of DOI:10.1002/adsc.200606132

A room-temperature catalytic alternative to the Ullmann reaction is presented, based on electroreductive homocoupling of haloarenes catalysed by palladium naaoparticles. The particles are generated in situ in an electrochemical cell, and electrons are used to close the catalytic cycle and provide the motivating force for the reaction. This system gives good yields using iodo- and bromoarenes, and requires only electric current and water as reagents. Using an ionic liquid solvent combines the advantages of excellent conductivity and cluster stabilising. The solvent is reusable at least five times. Kinetic experiments at different electrode potentials show that the two-electron oxidation of water closes the catalytic cycle by regenerating the Pd(0). A mechanism involving radical anion formation is proposed. The advantages and limitations of this new system for carbon-carbon homocoupling and cross-coupling are discussed.

Full text of DOI:10.1002/adsc.200606132

FULL PAPERS
DOI: 10.1002/adsc.200606132
Electroreductive Palladium-Catalysed Ullmann Reactions in Ionic
Liquids: Scope and Mechanism
a
a
a,
Laura Durµn Pachón, Cornelis J. Elsevier, and Gadi Rothenberg *
a
van ꢀt Hoff Institute of Molecular Sciences, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam,
The Netherlands
Fax : (+31)-20-525-5604; e-mail: gadi@science.uva.nl
Received: March 23, 2006; Accepted: June 19, 2006
Abstract: A room-temperature catalytic alternative periments at different electrode potentials show that
to the Ullmann reaction is presented, based on elec- the two-electron oxidation ofwater closes the cata-
troreductive homocoupling ofhaloarenes catalysed
lytic cycle by regenerating the Pd(0). A mechanism
by palladium nanoparticles. The particles are gener- involving radical anion formation is proposed. The
ated in situ in an electrochemical cell, and electrons advantages and limitations ofthis new system of r
are used to close the catalytic cycle and provide the carbon-carbon homocoupling and cross-coupling are
motivating force for the reaction. This system gives discussed.
good yields using iodo- and bromoarenes, and re-
quires only electric current and water as reagents.
Using an ionic liquid solvent combines the advantag- Keywords: biaryls; CÀC coupling; electrochemistry;
es ofexcellent conductivity and cluster stabilising.
green chemistry; heterogeneous catalysis; homoge-
The solvent is reusable at least five times. Kinetic ex- neous catalysis; nanoparticles
Introduction
generated and reduced back to Pd(0) using an exter-
nal reducing agent, and oxidative coupling, which
Symmetrical biaryls are important intermediates for starts with Pd2 and needs an oxidising agent. Various
+
synthesising agrochemicals, pharmaceuticals and natu- reagents can be used for closing the reductive cou-
[
1]
À [4]
[5]
[6]
ral products. One ofthe simplest protocols to make
pling cycle, including HCO2
,
H gas, Zn/H O,
2 2
[2]
[7]
them is the Ullmann reaction, the thermal homo- and alcohols. The two pathways can even be joined,
coupling ofaryl iodides in the presence ofcopper giving a tandem system that converges on one prod-
metal. This reaction, although over a century old, is uct. All ofthese examples, however, require an ad-
still used today despite two main disadvantages: First, ditional chemical reagent.
[
8]
it uses stoichiometric copper, generating stoichiomet-
Using electrochemistry is an interesting alternative,
ric amounts ofCuI 2 waste [Eq. (1)]. Second, it only because in principle the reductive coupling requires
only two electrons for closing the catalytic cycle. Ull-
mann-type reactions using a combination ofelectro-
chemistry and Ni(0), Pd(0) or Co(0) complexes as
well as sacrificial Mg anodes have been reported.[
9,10]
In this paper, we present a different approach that
combines electrochemical reduction and palladium
nanoparticle catalysts in an ionic liquid solvent[
11]
[
Eq. (2)]. The catalytic cycle is closed by electrochem-
works with aryl iodides. This is a problem because ical regeneration ofthe Pd(0) species.
chemicals react by their molarity, but are quantified
by their mass, and aryl iodides are both expensive and
wasteful in this respect.[
3]
In the past five years, we have shown that heteroge-
neous Pd/C catalyses Ullmann-type reactions ofaryl
iodides, bromides, and chlorides. Two reaction path-
ways are possible: Reductive coupling, where Pd2+ is
Adv. Synth. Catal. 2006, 348, 1705 – 1710
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1705

Laura Durµn Pachón et al.
FULL PAPERS
Results and Discussion
two electrodes, so that now the Pd electrode was the
cathode and the Pt electrode was the anode (Figure 1,
right). The rationale behind this experiment was that,
in theory, the catalysis could occur on the cathode sur-
Palladium Cluster-Catalysed Electrosynthesis of
Biphenyl
[
4]
face (Figure 2).
Direct electron transfer from a
The initial experiments were run with iodobenzene.
In
octylmethylimidazolium] [BF ] in a specially con-
4
a
typical reaction, PhI was stirred in
+
À
[
A
H
R
U
G
structed electrochemical cell (Figure 1, left) contain-
Figure 2. Testing the direct electron-transfer hypothesis,
where the PhI radical anion forms directly at the Pd cathode
Figure 1. Photo ofthe two-electrode cell ( left) and cartoons
showing the generation ofPd clusters using a Pd anode and
a Pt cathode (middle) and the reverse configuration (right).
(
this experiment was done using the reverse configuration,
cf. Figure 1, right).
Pd(0) atom on the cathode surface to PhI would give
À·
ing a Pd anode and a Pt cathode. Using an ionic a [PhI] radical anion, that would then dissociate to
[
12]
·
À [16]
liquid as solvent
has two main advantages: First, it Ph and I .
The constant supply ofelectrons to the
is a good conductor ofelectricity; second, it can stabi-
cathode would ensure the electron transfer (Figure 2).
lise metal nanoparticles via an ion bilayer mechanism, However, we did not observe any reaction in this
avoiding additional electrolytes or stabilisers. This case, and no Pd nanoparticles formed. Thus, we con-
type ofꢀligand- rf eeꢂ catalysis has gained considerable
importance, as it avoids using expensive ligands and sary for catalysing the homocoupling of aryl halides.
allows catalyst recycling.[13,14]
To examine the scope ofthis catalytic system, we
The solution was kept at 258C under a constant tested a variety of p-substituted haloarenes [Eq. (4)
current intensity of10 mA (1.6 V) and the reaction and Table 1]. Aryl iodides gave good yields (Table 1,
clude that, in our system, Pd nanoparticles are neces-
progress was monitored by gas chromatography (GC). entries 1–3) with the exception ofthe electron-donat-
After 20 min, the reaction mixture turned from a light ing iodoaniline. Aryl bromides were also active, but
yellow solution to a dark brown suspension. However, surprisingly less conversion was observed for the elec-
no conversion was observed by GC analysis at this tron-withdrawing bromobenzotrifluoride. The corre-
stage. The colour change reflects the anodic oxidation sponding chloride substrates showed little or no activ-
2
+
+
À
ofPd(0) to Pd
ions [Eq. (3)]. These are in turn re- ity. The ionic liquid [omim] [BF ] was recycled after
A
H
R
U
G
4
the reaction and re-used as solvent five times without
change in activity (see Experimental Section for de-
tails).
duced to adatoms at the Pt cathode, and finally form Mechanistic Studies
[
15]
stabilised Pd(0) nanoparticles (Figure 1, middle).
After 8 h, the PhI was totally consumed, giving 80% Palladium-catalysed systems have been described as
biphenyl and 20% benzene. Weighing the electrodes an alternative to the stoichiometric Ullmann protocol
before and after the reaction showed a difference of with copper. The reaction mechanism, however, is
~
2.5 mg in the Pd anode, equivalent to 0.1 mol% rel- still unclear. One proposal involves the formation of a
ative to the initial amount ofPhI. This corresponds to
a minimum TON of1000 (assuming that all the ꢀmiss-
ingꢂ Pd participates in the catalysis).
radical anion through a single electron transfer (SET)
À·
process that give a [PhX] radical anion that then
·
dissociates, forming a Ph radical and ultimately the
[
4,17]
To investigate the role ofpalladium nanoparticles
biaryl product.
To understand the mechanism in
in this system, we switched the current between the our system, we must first clarify what closes the cata-
1706
www.asc.wiley-vch.de
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1705 – 1710

Electroreductive Palladium-Catalysed Ullmann Reactions in Ionic Liquids
FULL PAPERS
[
18–20]
ofalmost 3 h was observed.
This induction
period supports the involvement ofpalladium clusters
in the cycle, because ifthe Pd(0) atoms or adatoms
alone were responsible for the catalysis, there would
not be such a long induction period.
After 3 h, the reaction rate increased from k =
1
À1
2
À1
2
0
0
.0542 min
(R =0.979) to k =0.383 min
(R =
2
2
.999, both R values pertain to six observations in
duplicate experiments). We think this is due to the
optimal particle size” formation during the induction
period, as proposed by Finke and co-workers in the
“
[
20]
case ofhydrogenation.
The unusual kinetic profile
for such reactions, where an induction period was also
found, was interpreted by Schmidt and Smirnov using
[
20,21]
the concept ofa “cluster magic number”.
We
used transmission electron microscopy (TEM) to
characterise the particle size at the end ofthe induc-
tion period. Spherical well-dispersed particles were
observed (Figure 4, left) with a size distribution of
2
~
. 35 0Æ0 0a . t5 o nm ms .. This corresponds to a ꢀmagic numberꢂ of
lytic cycle and what role do the Pd clusters play in Figure 4. Transmission electron micrograph (left) and Pd
cluster size distribution (right, based on 600 particles count-
ed).
this reaction.
First, we monitored the kinetics ofthe model reac-
tion for PhI coupling (Figure 3). An induction period
À1
The TOF for these 2 h was 230 h , more than
double the average TOF during the entire reaction
À1
(
100 h ). This also fits well with the ꢀactive clusterꢂ
concept. Another interesting point is the reaction
order for this period. We fitted the kinetic data and
2
the best R value was obtained for a first-order reac-
À1
2
2
tion (k =0.383 min , R =0.999; cf. with R =0.977
2
for second-order). This agrees with a first-order rate-
À·
determining step, e.g., the dissociation ofthe [PhX]
radical anion.
An important mechanistic question is: “what closes
the catalytic cycle?” We think that it is oxidation of
water into dioxygen [Eq. (5)], recovering the Pd(0)
Figure 3. Time-resolved reaction profiles for the electrore-
ductive Pd-catalysed coupling ofPhI. Reaction conditions
are as in Table 1. The ꢀ*ꢂ and ꢀ*ꢂ symbols represent dupli- media. A similar reaction was observed by Raynal
catalyst and neutralising the halide anions in the
cate experiments.
et al. in their elegant Ni complex-catalysed electrosyn-
Adv. Synth. Catal. 2006, 348, 1705 – 1710
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1707

Laura Durµn Pachón et al.
FULL PAPERS
[a]
Table 1. Pd-catalysed electroreductive homocoupling ofvarious haloarenes.
[b]
[b]
[c]
Entry
1
Aryl halide
Conversion [%]
99
Yield [%]
TON
Time [h]
8
80 (55)
800
2
3
4
5
6
7
8
9
99
83
56
99
74
83
80
78
20
5
82
820
760
350
750
610
590
650
610
140
40
8
76
9
35
9
75
14
20
24
24
24
24
24
24
-
61
59 (38)
65
61
14
4
1
1
1
1
1
0
1
2
3
4
5
3
30
<1
-
-
<1
-
-
-
[
[
a]
+
À
4
Reaction conditions: 20 mmol aryl halide, 50 mL [omim]
A
H
R
U
G
b]
Based on GC analysis, corrected for the presence of an internal standard. The numbers in parentheses represent isolated
yields.
[
c]
Minimum TON based on the difference in weight in the Pd anode before and after the reaction. This electrode weight
difference was independent from the reaction time.
thesis ofbiaryls in water. [22] Ionic liquids are notori- model reaction, the applied potential (1.60 V) was de-
ously hygroscopic, and a small water impurity would creased after a given time to below the potential nec-
suffice for closing the cycle.[ To test this hypothesis, essary for water oxidation (1.23 V). Again, we ob-
23]
we ran control experiments in the presence ofone
molar equivalent ofwater. The reaction was indeed
faster (complete conversion after 6 h, cf. with 8 h for 1 h. As expected, the same k value was found as
served an induction period of3 h. Subsequently, the
reaction rate increased, reaching ~30% conversion in
À1
2
the ꢀdryꢂ system). However, no difference in reactivity before at this point (k =0.383 min , R =0.999 for
3
was found when more equivalents of water were four observations in duplicate experiments). After
added.
4 h, the potential was decreased to 1.00 V for 3 h. GC
To obtain further evidence for the water oxidation, analysis showed that almost no reaction occurred.
we performed a simple experiment varying the poten- After 7 h, we increased the potential back to 1.60 V
tial during the reaction (Figure 5). For the same and the reaction re-started immediately, reaching the
1708
www.asc.wiley-vch.de
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1705 – 1710

Electroreductive Palladium-Catalysed Ullmann Reactions in Ionic Liquids
FULL PAPERS
Figure 5. Time-resolved reaction profiles observed for the
coupling ofPhI to biphenyl using palladium clusters catalyst.
Reaction conditions are as in Table 1. The ꢀ*ꢂ and ꢀ*ꢂ sym-
bols represent duplicate experiments.
Figure 6. Proposed catalytic cycle for the electroreductive
palladium-catalysed homocoupling ofaryl halides. The dark
grey spheres represent Pd ions. This cycle shows for clarity
min , R =0.920 for 12 observations in dupli- two single electron transfers from the same cluster, but in
+
maximum conversion after ~10 h (k =8.25”
4
À3
À1
2
10
cate experiments). This experiment shows that water principle these processes could also occur on different clus-
oxidation indeed closes the cycle. Furthermore, the ters.
reaction can be controlled by simple variation ofthe
potential.
From these studies we conclude that two processes The attractiveness ofthis system lies in the af ct that
take place in this system. First, the in situ formation only electrons and water are necessary for closing the
ofthe palladium nanoparticles. Second, the Pd-cata-
lysed electroreductive homocoupling ofbiaryls with a
catalytic cycle. However, any practical large-scale ap-
plication would have to include recovery ofthe ꢀlostꢂ
first-order rate-determining step. Based on this we palladium catalyst. Further investigations ofthis inter-
propose a catalytic cycle (Figure 6) wherein a SET esting system will be the subject of uf ture research in
occurs from a Pd cluster to an ArX molecule, forming our laboratory.
À·
+
a [ArX] radical anion and a positively charged Pd
cluster. In case ofa single Pd atom, this would be un-
stable, but the charge on the cluster would be stabi-
lised by the electron cloud ofthe neighbouring Pd(0)
Experimental Section
À·
·
atoms. First-order dissociation of[ArX]
to Ar and
-
Materials and Instrumentation
X followed by coupling of two aryl radicals would
give the biaryl product. Simultaneously, oxidation of
1H NMR spectra were recorded on a Varian Mercury vx300
+
water to dioxygen and 2H regenerates the Pd(0)
instrument at 258C. GC analysis was performed on an Inter-
science GC-8000 gas chromatograph with a 100% dimethyl-
polysiloxane capillary column (DB-1, 30 m ” 0.325 mm).
GC conditions (pentadecane internal standard): isotherm at
clusters and closes the catalytic cycle.
À1
Conclusions
1058C (2 min); ramp at 308C min to 2808C; isotherm at
2
808C (5 min). Electrochemical experiments were done
using a special home-made cell coupled to a dual current
supply with a maximum output of10 V/40 mA. The elec-
We show that palladium nanoparticles generated in
situ catalyse efficiently Ullmann-type reactions. To
the best ofour knowledge, this is the fi rst time this re-
action has been performed using electroreductive pal-
ladium catalysis. Room-temperature ionic liquids are
good solvents for preparing and stabilising palladium
nanoparticles. The kinetic data support a cycle involv-
2
trode surface areas were 1.5 cm for the Pd electrode (wire,
2
9
9.9%) and 3.0 cm for the Pt electrode (wire, 99.9%), re-
spectively. A detailed technical description ofthis system
[24]
+
has been published elsewhere.
The ionic liquid [omim]
À
A
H
R
U
G
4
[25]
dried prior to use.
All other chemicals were purchased
ing the formation of a phenyl radical anion that coor- from commercial sources (>98% pure). The ionic liquid
dinates on the surface of the palladium nanoparticles. solvent was recycled following a published procedure.[26]
Adv. Synth. Catal. 2006, 348, 1705 – 1710
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1709

Laura Durµn Pachón et al.
FULL PAPERS
Procedure for Pd-Catalysed Electroreductive
Ullmann Coupling
[10] D. Kweon, Y. Jang, H. Kim, B. Kor. Chem. Soc., 2003,
2
4, 1049–1050.
[
11] For recent references on coupling reactions with palla-
dium nanoparticles in ionic liquids see: a) D. B. Zhao,
Z. F. Fei, T. J. Geldbach, R. Scopelliti, P. J. Dyson, J.
Am. Chem. Soc. 2004, 126, 15876–15882; b) V. Calò,
A. Nacci, A. Monopoli, A. Fornaro, L. Sabbatini, N.
Cioffi, N. Ditaranto, Organometallics 2004, 23, 5154–
5158; c) C. C. Cassol, A. P. Umpierre, G. Machado, S. I.
Wolke, J. Dupont, J. Am. Chem. Soc. 2005, 127, 3298–
3299.
Example 1 – Biphenyl (2a) from iodobenzene (1a): The
electrochemical cell was charged with PhI (4.09 g,
+
À
[BF ] . After stirring for
4
20.0 mmol) and 50 mL [omim]
A
H
R
U
G
5
min, a constant current (10 mA, 1.6 V) was applied and
the mixture was further stirred for 8 h at 258C. Reaction
progress was monitored by GC. After 8 h, the product was
extracted with ether (3”50 mL). The ether phases were
combined and evaporated under vacuum to give the product
as colourless crystalline platelets; yield: 0.81 g (55 mol%
[12] For reviews on ionic liquids as solvents see: a) J. S.
Wilkes, J. Mol. Catal. A 2004, 214, 11–17; b) J. Dupont,
R. F. de Souza, P. A. Z. Suarez, Chem. Rev. 2002, 102,
3667–3692; c) P. Wasserscheid, W. Keim, Angew.
Chem. Int. Ed. 2000, 39, 3772–3789; d) M. J. Earle,
K. R. Seddon, Pure Appl. Chem. 2000, 72, 1391–1398;
e) T. Welton, Chem. Rev. 1999, 99, 2071–2083.
[
27]
based on PhI); mp 71–738C (lit.,
69–728C). The solvent
was recycled by washing with aqueous NaBF4. 1H NMR
(
Me Si): d=7.36–7.42 (m, 2H), 7.45–7.51 (m, 4H), 7.62–7.68
4
(
m, 4H). Good agreement was found with the literature
[14]
values.
Example 2 – 4,4’-Dimethylbiphenyl (2c) from 1-bromo-
4
-methylbenzene (1g): The electrochemical cell was charged
[
13] J. Dupont, G. S. Fonseca, A. P. Umpierre, P. F. P. Ficht-
ner, S. R. Teixeira, J. Am. Chem. Soc. 2002, 124, 4228–
4229.
+
with CH C H Br (3.50 gr, 20.0 mmol) and 50 mL [omim]
3
6
4
À
A
C
H
T
R
E
U
N
G
[BF ] . The reaction was performed as above to give the
4
product as light yellow crystalline needles; yield: 0.71 g (38
[
14] V. Calò, A. Nacci, A. Monopoli, F. Montingelli, J. Org.
Chem. 2005, 70, 6040–6044.
[28]
mol% based on CH C H Br); mp 118–1218C (lit.,
119–
3
6
4
1
218C). The solvent was recycled by washing with aqueous
[
15] The electrochemical synthesis ofmetal nanoparticles
was developed by M. T. Reetz. For a mechanistic de-
scription, see: M. T. Reetz, W. Helbig, J. Am. Chem.
Soc. 1994, 116, 7401–7402.
16] The dissociation [PhX] into [Ph] and X is well docu-
mented, see: T. T. Tsou, J. K. Kochi, J. Am. Chem. Soc.
1979, 101, 6319–6332.
1
NaBF4. H NMR (Me Si): d=2,51 (m, 6H), 7.41–7.49 (m,
4
2
H), 7.53–7.57 (m, 4H), 7.60–7.66 (m, 4H). Good agree-
[
28]
ment was found with the literature values.
À·
·
-
[
[
Acknowledgements
17] A. J. Fry, J. M. Porter, P. F. Fry, J. Org. Chem. 1996, 61,
3191–3194.
We thank B. van Groen and P. F. Collignon for excellent
technical assistance and Dr. F. Hartl for discussions.
[18] L. Durµn Pachón, J. H. van Maarseveen, G. Rothen-
berg, Adv. Synth. Catal. 2005, 347, 811–815.
[
[
19] C. Besson, E. E. Finney, R. G. Finke, J. Am. Chem.
References
Soc. 2005, 127, 8179–8184.
20] C. Besson, E. E. Finney, R. G. Finke, Chem. Mater.
2005, 17, 4925–4938.
[
[
[
1] G. Bringmann, R. Walter, R. Weirich, Angew. Chem.
Int. Ed. Engl. 1990, 29, 977–991.
2] F. Ullmann, Ber. Dtsch. Chem. Ges. 1903, 36, 2359–
[21] A. F. Schmidt, V. V. Smirnov, Top. Catal. 2005, 32, 71–
75.
[22] F. Raynal, R. Barhdadi, J. Perichon, A. Savall, M. Trou-
pel, Adv. Synth. Catal. 2002, 344, 45–49.
[23] U. Schrçder, J. D. Wadhawan, R. G. Compton, F.
Marken, P. A. Z. Suarez, C. S. Consorti, R. F. de Souza,
J. Dupont, New. J. Chem. 2000, 24, 1009–1015.
2
362.
3] S. Mukhopadhyay, G. Rothenberg, G. Lando, K. Agba-
ria, M. Kazanci, Y. Sasson, Adv. Synth. Catal. 2001,
3
43, 455–459.
[4] S. Mukhopadhyay, G. Rothenberg, D. Gitis, H. Wiener,
Y. Sasson, J. Chem. Soc., Perkin Trans. 2 1999, 2481–
[
[
[
[
[
24] L. Durµn Pachón, M. B. Thathagar, F. Hartl, G. Roth-
enberg, Phys. Chem. Chem. Phys. 2006, 8, 151–157.
25] J. D. Holbrey, K. R. Seddon, J. Chem. Soc., Dalton
Trans. 1999, 2133–2139.
2
484.
[
[
[
[
[
5] S. Mukhopadhyay, G. Rothenberg, H. Wiener, Y.
Sasson, Tetrahedron 1999, 55, 14763–14768.
6] S. Mukhopadhyay, G. Rothenberg, D. Gitis, Y. Sasson,
Org. Lett. 2000, 2, 211–214.
26] R. Barhdadi, C. Courtinard, J. Y. Nedelec, M. Troupel,
Chem. Commun. 2003, 1434–1435.
7] D. Gitis, S. Mukhopadhyay, G. Rothenberg, Y. Sasson,
27] Y. Tamura, M. W. Chun, K. Inoue, J. Minamikawa, Syn-
thesis 1978, 11, 822–823.
28] V. Courtois, R. Barhdadi, M. Troupel, J. Perichon, Tet-
Org. Process. Res. Dev. 2003, 7, 109–114.
8] S. Mukhopadhyay, G. Rothenberg, D. Gitis, Y. Sasson,
J. Org. Chem. 2000, 65, 3107–3110.
rahedron 1997, 53, 11569–11576.
9] J. Y. NØdØlec, J. Perichon, M. Troupel, Top. Curr.
Chem. 1997, 185, 141–173.
1710
www.asc.wiley-vch.de
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1705 – 1710