One-pot synthesis of benzoic acid by electrocatalytic reduction of bromobenzene in the presence of CO2

Article abstract of DOI:10.1016/j.elecom.2011.05.009

The electrocatalytic properties of Ag for the reduction of organic halides have been exploited for one-pot electrosynthesis of benzoic acid. Among various investigated dipolar aprotic solvents, only dimethylformamide gives good results. The process has been investigated both under potentiostatic and galvanostatic control. In both cases, the reaction smoothly proceeds under mild conditions and lasts in a few hours, giving benzoic acid yields of the order of 80%.

Full text of DOI:10.1016/j.elecom.2011.05.009

Electrochemistry Communications 13 (2011) 810–813
Contents lists available at ScienceDirect
Electrochemistry Communications
journal homepage: www.elsevier.com/locate/elecom
One-pot synthesis of benzoic acid by electrocatalytic reduction of bromobenzene in
the presence of CO2
⁎
Abdirisak A. Isse, Christian Durante, Armando Gennaro
Department of Chemical Sciences, University of Padova, Via Marzolo, 1–35131 Padova, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The electrocatalytic properties of Ag for the reduction of organic halides have been exploited for one-pot
electrosynthesis of benzoic acid. Among various investigated dipolar aprotic solvents, only dimethylforma-
mide gives good results. The process has been investigated both under potentiostatic and galvanostatic
control. In both cases, the reaction smoothly proceeds under mild conditions and lasts in a few hours, giving
benzoic acid yields of the order of 80%.
Received 21 March 2011
Received in revised form 10 May 2011
Accepted 10 May 2011
Available online 15 May 2011
© 2011 Elsevier B.V. All rights reserved.
Keywords:
Electrocarboxylation
Benzoic acid
Electrocatalysis
1. Introduction
Herein, we report the electrocatalytic carboxylation of bromoben-
zene studied as a model system for the development of a simple method
Carbon dioxide is a cheap, nontoxic and abundant C1 source, which
represents an ideal C1 building block for organic synthesis [1]. Among
the possible products that may be prepared by C-fixation of CO2,
arylcarboxylic acids represent an important class of compounds in
organic synthesis [2]. However, practical use of CO2 in synthetic
processes leading to C―C bond formation is quite limited by the high
thermodynamic stability and low reactivity of the molecule. Therefore,
methods to catalytically activate CO2 for C―C bond forming reactions
have been extensively investigated. In particular, transition-metal-
catalyzed addition of carbon nucleophiles to CO2 has recently attracted
much attention as a viable method of synthesis of arylcarboxylic acids
[3–6]. These reactions use organozinc reagents [3,4] or organoboronic
esters [5,6], which are both prepared starting from haloarenes. Although
high yields have been reported in several cases, these catalytic systems
have some disadvantages. The process involves at least two distinct
stages of synthesis and requires highly expensive transition metal
catalysts. In addition, reactions are often very slow, requiring reaction
times as long as 12–36 h.
for the one-pot synthesis of functionalized benzoic acids under mild
conditions. The method fully exploits the powerful electrocatalytic
properties of Ag for the reduction of carbon–halogen bonds [8,9].
Although electrocarboxylation of organic halides has been extensively
applied for the synthesis of many carboxylic acids [10–14], especially
substituted acetic and propanoic acids, little attention has been devoted
to the electrosynthesis of aromatic acids.
2. Experimental
2.1. Chemicals
N,N-dimethylformamide (DMF) and acetonitrile were purified as
previously described [14]. THF was distilled in the presence of Na/
benzophenone under argon atmosphere. All other solvents (≥99.5%,
extra dry) were used as received. Et4NBF4 and n-Bu4NBF4 (Fluka, 98%)
were recrystallized from an ethanol/water mixture (2:1) and dried in
a vacuum oven at 70 °C.
Some progress has recently been made to eliminate the use of
expensive metal catalysts. Kobayashi and Kondo [4] have shown that
organozinc reagents can be carboxylated with CO2 in a transition-metal-
free process by appropriately choosing the reaction medium. A different
approach has been proposed by Correa and Martín [7], whoreportone-pot
synthesis of carboxylic acids by catalytic carboxylation of aryl bromides.
They use a Pd(0) complex as a catalyst and Et2Zn as a reducing agent.
2.2. Instrumentation and procedures
Electrochemical experiments were carried out at 25 °C in a three-
electrode cell by using a computer-controlled EG&G PARC Model 273A
potentiostat/galvanostat. The working electrodes used for cyclic voltam-
metry were glassy carbon and silver, fabricated and activated as reported
previously [8], whereas the counter-electrode and the reference electrode
were a Pt wire and Ag|AgI|0.1 M n-Bu4NI in DMF, respectively. The latter
was always calibrated against the ferrocenium/ferrocene couple for the
conversion of the potentials to the SCE scale.
⁎
Corresponding author. Tel.: +39 049 8275132; fax: +39 049 8275829.
E-mail address: armando.gennaro@unipd.it (A. Gennaro).
1388-2481/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.elecom.2011.05.009

A.A. Isse et al. / Electrochemistry Communications 13 (2011) 810–813
811
Table 2
Preparative scale electrolyses were performed at an Ag plate of
Electrocarboxylation of PhBr in different CO2-saturated solvents.a
3.6 cm2 in an undivided cell with a sacrificial Al anode. Analyses of the
products were performed on untreated electrolyzed solutions by
HPLC (JASCO 2075, equipped with a UV detector (λ=256 nm) and a
15 cm, 4.6-mm Prevail organic acid column). The eluent was a
mixture of MeCN and phosphate buffer at pH=2.5. Identification and
quantitative determination of the products were based on comparison
with authentic compounds and the use of calibration curves. In a few
cases the target product was isolated according to the following
procedure. At the end of the electrolysis, the solvent was evaporated
off under reduced pressure and the residue was treated with 20 mL of
1 M HCl and then extracted with Et2O (5×15 mL). The ether was
evaporated and the residue was re-dissolved in 20 mL of 1 M NaOH
and extracted with Et2O (5×10 mL). Finally, the aqueous fractions
were acidified with 6 M HCl (5 mL) and extracted with Et2O
(5×15 mL). The ether was again evaporated and the residue was
dried under vacuum for several hours to give a pale yellow solid,
which was analyzed by HPLC and NMR. 1H-NMR (250 MHz, DMSO)
(ppm): 7.46–7.53 (m, 2 H), 7.59–7.66 (m, 1 H), 7.93–7.96 (m, 2 H),
10.73 (br, 1 H). 13C-NMR (250 MHz, DMSO) (ppm): 128.5, 129.2,
130.7, 132.8, 167.3.
b
Entry
Solvent
Eapp
jb
Conversion
(%)
PhH
(%)c
PhCO2H
(V vs SCE)
(mA/cm2)
(%)c
Controlled-potential electrolysis
1
2
MeCN
DMSO
PC
NMP
NMP
THF
DMF
DMF
DMF
DMF
DMF
−1.85
−1.85
−1.85
−1.85
−2.25
−2.40
−1.85
−1.85
−1.85
−2.00
−2.20
93
67
75
46
84
72
67
80
70
73
70
94
89
8
6
6
18
36
31
59
3
4d
5
64
67
41
16
20
21
17
14
6e
7f
8
83 (74)g
79 (71)g
78 (68)g
83
9h
10
11
85
Galvanostatic electrolysis
12
13
14
DMF
DMF
DMF
11
17
22
77
70
65
15
13
14
83
86
86
a
Unless otherwise stated, other conditions were: 0.1 M Et4NBF4 background
electrolyte, CPhBr =95 mM; T=25 °C, Ag cathode, Al sacrificial anode.
b
Applied potential (Eapp) or current density (j).
Yield calculated with respect to converted PhBr after a charge consumption of 2.1
c
e−/molecule.
3. Results and discussion
d
Electrolysis interrupted after charge consumption of 1.18 e–/molecule.
0.25 M Bu4NBF4 was used as base electrolyte.
CPhBr =53 mM.
Isolated yield.
e
The success of electrocarboxylation relies on the selective
reduction of RX to generate R−, which immediately captures CO2.
Therefore, we started the study by examining the redox properties of
bromobenzene and CO2 in different solvents at Ag and GC electrodes
to check whether reduction of CO2 overlaps that of the substrate. In
cyclic voltammetry each of the two compounds shows a single,
irreversible cathodic peak. The peak potentials measured at
v=0.2 V s−1 in all solvents are collected in Table 1. These data
show that Ag has extraordinary electrocatalytic properties for the
reduction of bromobenzene. The peak potentials of PhBr are more
than 0.6 V more positive at Ag than at GC, except in THF and DMSO.
What is more important is that at Ag, Ep of PhBr is always considerably
more positive than that of CO2. The data show that selective reduction
of PhBr at Ag can be carried out in the presence of CO2 in all
investigated solvents.
A series of controlled-potential electrolyses (CPEs) was performed
in CO2-saturated solvents containing 95 mM PhBr. As a rule, the
applied potential (Eapp) was chosen to be slightly more negative than
Ep of PhBr and the electrolysis was interrupted after a charge
consumption of 2.1 e−/molecule of PhBr. At the end of the electrolysis
the solution was analyzed by HPLC for the quantitative determination
of the products, benzene and benzoic acid, and the results are
summarized in Table 2. Neither fouling of the Ag electrode during
electrolysis in most solvents nor any weight loss at the end of the
experiment has been observed. Both the distribution of the products
and the overall conversion of PhBr are strongly affected by the nature
of the solvent. In particular, the process in MeCN and DMSO gives
f
g
h
CPhBr =210 mM.
almost exclusively benzene. A slightly better result is obtained in PC
and NMP with an acid yield of 18% and 31–36%, respectively. When
the electrolysis was performed at −1.85 V in NMP (Table 2, entry 4),
good conversion could not be achieved because of passivation of the
Ag electrode. This problem of passivation could be avoided by shifting
Eapp to more negative values (Table 2, entry 5).
The best results are obtained in THF and DMF. It is worth noting,
however, that THF is characterized by a high resistivity, which
represents a severe drawback to its use in preparative electrochem-
istry. In fact, the electrolysis in THF had to be carried out using a high
concentration of background electrolyte and Eapp =−2.40 V instead
of −1.85 V to achieve a satisfactory current. The effective value of Eapp
is not precisely known, but is certainly much more positive than
−2.40 V because of the contribution of the uncompensated ohmic
drop. DMF, which like all other polar solvents does not present these
problems, shows good results both in terms of conversion and acid
yield (Table 2, entries 7–11).
Fig. 1A shows variations in the quantity (n) of PhBr and its
reduction products during the course of a controlled-potential
electrolysis in DMF. As the consumed charge increases, both
conversion of PhBr and formation of PhH and PhCO2H smoothly
increase, so that the yield of the acid remains constant (ca 80%).
Almost complete conversion of bromobenzene is achieved with ca
30% excess charge with respect to the theoretical 2 e−/molecule.
Some control experiments performed in DMF and MeCN in the
absence of CO2 show complete conversion of PhBr into PhH with the
consumption of only 2 e−/molecule. The excess charge consumed in
the electrocarboxylation experiments is therefore to be attributed to a
partial involvement of CO2 reduction in the process. Isolation of
benzoic acid after electrolysis results in a decrease of its yield from ca
80% to ca 71% (Table 2, entries 7–9).
The effects of Eapp and CPhBr on the performance of the process
have been investigated in DMF. As shown in entries 7–11 of Table 2,
neither of these parameters significantly affects the yield of the target
acid. However, decreasing Eapp to more negative values lowers the
conversion from 80% to 70%, possibly because of a greater involve-
ment of CO2 reduction at very negative potentials.
Table 1
Voltammetric data for the reduction of PhBr and CO2 in different solvents at 25 °C.a
Solvent
Ep(GC) (V)
Ep(Ag) (V)
ΔEp (V)b
PhBr
CO2
PhBr
CO2
MeCN
THFc
PC
DMF
DMSO
NMP
−2.58
−2.05
−2.32
−2.71
−2.27
−2.53
−2.48
b−2.8
−2.45
b−3
−2.47
−2.49
−1.78
−1.80
−1.69
−1.79
−1.81
−1.63
−2.37
−2.68
−2.26
−2.35
−2.33
−2.27
0.80
0.25
0.63
0.92
0.46
0.90
Background electrolyte: 0.1 M Et4NBF4; Ep (vs SCE) measured at v=0.2 Vs−1
.
a
b
ΔEp is the difference between the reduction peak potentials of PhBr at Ag and GC
cathodes (ΔEp =Ep(Ag) −Ep(GC)).
c
0.1 M Bu4NBF4 was used as base electrolyte.

812
A.A. Isse et al. / Electrochemistry Communications 13 (2011) 810–813
Increasing the charge consumption to 2.7 e–/molecule slightly increases
the conversion of PhBr to 79, 79 and 77% for j=11, 17 and 22 mA cm−2
,
respectively. It is important to note that neither the current density nor
the amount of consumed charge affects the yield of the target acid.
An example of the trends of PhBr disappearance and product
formation during a galvanostatic electrolysis is shown in Fig. 1. The
figure shows also the dependence of PhCO2H yield and current
efficiency (CE) on the consumed charge. CE is defined as the ratio
between the charge consumed for the production of benzoic acid and
the total charge (QPhCOOH /Qtotal). The trends are similar to those
observed for the CPEs; during electrolysis formation of PhCO2H and
PhH smoothly increases with increasing PhBr conversion. As shown in
Fig. 1C CE decreases as the reaction proceeds and PhBr conversion
increases, while the yield remains high and constant (ca 85%).
4. Mechanistic considerations
Based on the known electrochemistry of aromatic bromides both
at inert [17] and catalytic electrodes [9], we propose the general
mechanism shown in Scheme 1 for the reduction of bromobenzene at
Ag in the presence of CO2. The initial step is adsorption of PhBr
followed by dissociative electron transfer, according to either a
concerted or a stepwise mechanism, leading to adsorbed products. In
a previous study on the electrocatalytic activation of bromobenzenes
at Ag we have shown that the process involves surface bound species
and the concerted mechanism is more likely for PhBr [9].
Different reaction pathways can be distinguished in Scheme 1.
Ph•ads may either leave the electrode surface and abstract a H atom
from the solvent or get reduced to a carbanion. The reaction route
involving dissolved Ph• is very unlikely because reduction of the
radical (Eo =0.05 V vs SCE [18]) is much easier than that of PhBr
(Ep =−1.79 V vs SCE in DMF at v=0.2 V/s). This means Ph•ads will be
immediately reduced to Ph−ads, which could be less prone to
adsorption because of an unfavorable electrostatic interaction with
the negatively charged electrode surface. However, very recently it
has been shown that, in the case of the reduction of benzyl chloride at
Ag, the interaction between the carbanion and silver is very similar to
the formation of an organometallic intermediate PhCH2-Ag−[19].
Fig. 1. Electrocarboxylation of 95 mM PhBr in CO2-saturated DMF+0.1 M Et4NBF4 at
Ag. Dependence of the amount of unreacted PhBr and its reduction products on
consumed charge during (A) controlled-potential electrolysis at −1.85 V or (B)
galvanostatic electrolysis with j=17 mA cm−2: PhBr ( ), PhH ( ), PhCO2H ( ).
(C) Variation of PhCO2H yield ( ) and current efficiency ( ) with consumed charge
during galvanostatic electrolysis.
Thus, Ph−
in Scheme 1 could stand for a phenyl-silver anionic
ads
adduct that can react with electrolphiles such as CO2 or H2O. Two
reaction channels can be envisaged for Ph− and Ph− after
ads
The reaction is found to be sensitive to the nature of the solvent,
hydrodehalogenation of PhBr to PhH being strongly favored over
electrocarboxylation in most investigated solvents, especially MeCN
and DMSO, whereas formation of benzoic acid is favored in DMF. A
similar trend has previously been reported for the electrocarboxylation
of 2-amino-5-bromopyridine [14] and, more recently, for the reaction of
phenylzinc iodide with CO2[4]. The role of the solvent on the
electrocarboxylation process cannot be easily rationalized. For example,
the observed trend does not correlate with the solubility of CO2, which
increases in the following order: DMSObNMPbDMFbTHFbMeCN
[15,16]. Perhaps several parameters including CO2 solubility, acidity of
the residual water in the solvent, H-atom donor and proton donor
abilities of the solvent, etc., should be taken into consideration.
desorption, both involving nucleophilic attack on CO2 and any proton
Galvanostatic electrolysis is often more simple and more preferable
than the potentiostatic one, especially when large scale applications are
considered. We therefore checked the efficiency of the electrosynthesis
under galvanostatic conditions. Some electrolyses were performed with
different currentdensities underotherwisethe sameconditionsused for
the potentiostatic experiments. The course of the reaction was
monitored during electrolysis, which was interrupted, as usual, after a
charge consumption of 2.1 e−/molecule. The results are reported in
Table 2 (entries 12–14). These electrolyses give slightly better acid
yields (ca 85%) than the experiments performed under potential
control. The overall conversion is smaller than that of the controlled-
potential electrolysis and is significantly affected by the current density.
Scheme 1. Possible mechanism of electrocarboxylation of PhBr at Ag.

A.A. Isse et al. / Electrochemistry Communications 13 (2011) 810–813
813
donor HA available in the reaction medium. Since CO2 is the best
electrophile present in solution, carboxylation of desorbed Ph− is
expected to be much faster than its protonation. Indeed, this is what is
observed for the electrocarboxylation of a large number of organic
halides in DMF and MeCN [10–13]. In these cases high yields of
carboxylated products are obtained. The other reaction route involves
Ph−ads, which is engaged in the same competition as before. In this
case, however, the nucleophilic center suffers steric hindrance from its
interaction with the electrode surface and this may drastically affect the
competition in favor of the protonation. We may therefore speculate
that the solvent plays a crucial role in the fate of Ph−ads in terms of both
its reactivity with the electrophiles and its adsorption/desorption
behavior.
particular attention to the tolerance to reducible functional groups
such as ketones, nitriles, halogen atoms, etc.
Acknowledgment
This work was financially supported by the University of Padova.
References
[1] T. Sakakura, J.C. Choi, H. Yasuda, Chem. Rev. 107 (2007) 2365.
[2] C. Wang, I. Piel, F. Glorius, J. Am. Chem. Soc. 131 (2009) 4194.
[3] C.S. Yeung, V.M. Dong, J. Am. Chem. Soc. 130 (2008) 7826.
[4] K. Kobayashi, Y. Kondo, Org. Lett. 11 (2009) 2035.
[5] K. Ukai, M. Aoki, J. Takaya, N. Iwasawa, J. Am. Chem. Soc. 128 (2006) 8706.
[6] T. Ohishi, M. Nishiura, Z. Hou, Angew. Chem. Int. Ed. Engl. 47 (2008) 5792.
[7] A. Correa, R. Martin, J. Am. Chem. Soc. 131 (2009) 15974.
[8] A.A. Isse, S. Gottardello, C. Durante, A. Gennaro, Phys. Chem. Chem. Phys. 10
(2008) 2409.
5. Conclusions
[9] A.A. Isse, P.R. Mussini, A. Gennaro, J. Phys. Chem. C 113 (2009) 14983.
[10] G. Silvestri, S. Gambino, G. Filardo, A. Gulotta, Angew. Chem. Int. Ed. Engl. 23
(1984) 979.
[11] O. Sock, M. Troupel, J. Perichon, Tetrahedron Lett. 26 (1985) 1509.
[12] A.A. Isse, A. Gennaro, Chem. Commun. (2002) 2798.
[13] C.L. Scortichini, S.J. Babinec, J. Electroanal. Chem. 379 (1994) 111.
[14] A. Gennaro, C.M. Sánchez-Sánchez, A.A. Isse, V. Montiel, Electrochem. Commun. 6
(2004) 627.
[15] A. Gennaro, A.A. Isse, E. Vianello, J. Electroanal. Chem. 289 (1990) 203.
[16] M.G. Florentino, A.T. Rodriguez, J. Chem. Eng. Data 29 (1984) 456.
[17] C.P. Andrieux, C. Blocman, J.-M. Dumas-Bouchiat, J.-M. Savéant, J. Am. Chem. Soc.
101 (1979) 3431.
[18] C.P. Andrieux, J. Pinson, J. Am. Chem. Soc. 125 (2003) 14801.
[19] Y.-F. Huang, D.-Y. Wu, A. Wang, B. Ren, S. Rondinini, Z.-Q. Tian, C. Amatore, J. Am.
Chem. Soc. 132 (2010) 17199.
In conclusion, we report an efficient method of synthesis of
benzoic acid, which offers some advantages over methods based on
transition-metal catalysis. Among these, short reaction time (ca 3 h),
no need for expensive metal catalysts and/or preparation of reactive
organometallic intermediates, and simplicity of the electrolysis
procedure may be mentioned. Both galvanostatic and potentiosatic
electrolyses in DMF give good acid yields, whereas higher PhBr
conversions approaching 100% can be achieved in the case of CPE
without a significant decrease of current efficiency. Further in-
vestigations are underway to understand the role of the solvent and
also to define better the scope and limitations of the method, with