Accepted Manuscript

Title: The electrocatalytic behaviour of Pt and Cu

nanoparticles supported on carbon nanotubes for the
nitrobenzene reduction in ethanol

Author: Benny Wouters<!–<query id="Q1">Please confirm

that given names and surnames have been identified
correctly.</query>–> Xia Sheng Andrea Boschin Tom
Breugelmans Elisabet Ahlberg Ivo F.J. Vankelecom Paolo P.
Pescarmona Annick Hubin

PII: S0013-4686(13)01495-3
DOI: http://dx.doi.org/doi:10.1016/j.electacta.2013.07.210
Reference: EA 21004

To appear in: Electrochimica Acta

Received date: 1-3-2013

Revised date: 27-7-2013
Accepted date: 28-7-2013

Please cite this article as: B. Wouters, X. Sheng, A. Boschin, T. Breugelmans,

E. Ahlberg, I.F.J. Vankelecom, P.P. Pescarmona, A. Hubin, The electrocatalytic
behaviour of Pt and Cu nanoparticles supported on carbon nanotubes
for the nitrobenzene reduction in ethanol, Electrochimica Acta (2013),
http://dx.doi.org/10.1016/j.electacta.2013.07.210.

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
*Revised Manuscript (including Abstract)

1
2
3
4
5
6
7
8
9 The electrocatalytic behaviour of Pt and Cu
10
11
nanoparticles supported on carbon nanotubes for the
12 nitrobenzene reduction in ethanol

t
13

ip
14
15 Benny Woutersa,1,∗, Xia Shengb,1 , Andrea Boschina,d , Tom Breugelmansa,c ,
16 Elisabet Ahlbergd , Ivo F. J. Vankelecomb , Paolo P. Pescarmonab , Annick

cr
17 Hubina
18 a Vrije Universiteit Brussel, Research Group Electrochemical and Surface Engineering,
19 Pleinlaan 2, 1050 Brussels, Belgium

us
20 b Centre for Surface Chemistry and Catalysis, University of Leuven (KU Leuven), Kasteelpark
21 Arenberg 23, 3001 Heverlee, Belgium
c Artesis University College of Antwerp, Applied Engineering and Technology - Chemistry,
22
23 Paardenmarkt 92, 2000 Antwerp, Belgium
d Department of Chemistry, University of Gothenburg, SE-41296 Gothenburg, Sweden
24

an
25
26
27
28
Abstract
M
29
30
31 In view of investigating the possibility of cogeneration of chemicals and electricity
32 in a fuel cell type reactor, a suitable electrocatalyst and optimal reaction condi-
33
d

34 tions have to be found. The nitrobenzene reduction was taken as the case study
35
te

reaction for this investigation. In literature no suitable electrocatalyst has been

36
37 found for this reaction. Pt and Cu nanoparticles, supported by multiple wall car-
38
bon nanotubes, were taken as electrocatalysts and ethanolic and acidic ethanolic
p

39
40 environments were studied as possible reaction media. The Cu nanoparticle
41
ce

42 catalyst displayed superior electrocatalytic behaviour since less overpotential

43 is needed for the nitrobenzene reduction in both ethanolic and acidic ethanolic
44
45 environment. The acidic ethanolic conditions were superior to the ethanolic
Ac

46
conditions as less overpotential was required for the nitrobenzene reduction
47
48 reaction and a higher conversion of nitrobenzene is obtained. In acidic ethanolic
49
50 environment, a four-electron reduction is taking place, which was assigned to
51 the reduction of nitrobenzene into phenylhydroxylamine. This compound is
52
53
54 ∗ Corresponding author
55 Email address: benny.wouters@vub.ac.be (Benny Wouters)
1 B. Wouters and X. Sheng have contributed equally to this paper.
56
57
58
59 Preprint submitted to Elsevier July 26, 2013
60
61
62
63
64 Page 1 of 19
65
 1
2
3
4
5
6
7
8
9 re-oxidised into nitrosobenzene when the solution comes into contact with oxygen,
10
and the formed nitrosobenzene selectively reacts with another phenylhydroxy-
11
12 lamine molecule, yielding azoxybenzene. The latter is an important compound

t
13

ip
14 in organic synthesis.
15 Keywords: nitrobenzene, electrocatalysis, carbon nanotubes, metal
16

cr
17 nanoparticles, porous rotating disc electrode
18
19

us
20 1. Introduction
21
22
23 In the chemical industry, many processes rely on oxidation and reduction
24

an
reactions for the production of various chemicals. As these processes are typically
25
26 heavily exothermic, this leads to a high amount of waste heat, which cannot
27 be fully recuperated. The energy efficiency of these processes is thus often low.
28
M
29 Moreover, the selectivity of the reactions that take place in these processes is
30
usually rather low. Apart from leading to loss of raw material, this implies that
31
32 extra purification steps, that introduce additional costs, are necessary.
33
d

34 The difficulties described above could be overcome by electrochemical produc-

35 tion of these chemicals in a fuel-cell type reactor. Such reactor would offer unique
te

36
37 opportunities both for energy saving (by cogeneration of electricity) and for selec-
38
tivity (by using the optimal electrocatalyst). In this work, the hydrogenation of
p

39
40 nitrobenzene is selected as a case study reaction, as its feasibility for cogeneration
41
ce

42 was proven by Yuan et al. [1] and it has a vast range of useful end-products,
43 including aniline, p-aminophenol, azoxybenzene and cyclohexylamine.
44
45 The reaction mechanism of the electrochemical reduction of nitrobenzene has
Ac

46 been studied extensively in the last decades [2–11]. It is generally understood that
47
48 the reduction of nitrobenzene occurs in two steps; in the first step, four electrons
49
are exchanged: nitrobenzene is reduced to phenylhydroxylamine through the
50
51 intermediate of nitrosobenzene (eq. 1).
52
53
54
55
56
57
58
59 2
60
61
62
63
64 Page 2 of 19
65
 1
2
3
4
5
6
7
8
9
10 NO2 NO NHOH
11
− −
12 + 2 e + 2 e

t
−−−−−→ −−−−−→ (1)
13

ip
14 nitrobenzene nitrosobenzene phenylhydroxylamine
15
16 Although the reaction mechanisms differ, the formation of phenylhydroxylamine

cr
17 can occur in acidic [6], neutral and alkaline conditions [8]. After the formation of
18
19 phenylhydroxylamine, in acidic environment, two pathways are possible, namely

us
20
21 the chemical rearrangement of phenylhydroxylamine to p-aminophenol, and the
22 electrochemical exchange of two more electrons, producing aniline. In neutral
23
24 and alkaline environment p-aminophenol is not formed.

an
25 NHOH
26
27 H+
−−→ HO NH2 (2)
28
M
29
p-aminophenol
30
31 NHOH NH2
32 + 2 e−
−−−−−→ (3)
33
d

34
aniline
35
te

36 The cathode material on which the reduction of nitrobenzene takes place is

37
expected to have a strong influence on the rate of the different reactions that can
38
p

39 occur. For instance, Kokkinidis et al. [6] found that Pb electrode materials and
40
41 Pb adsorbates on Ag completely suppress the formation of aniline. Cyr et al. [9]
ce

42 showed that Cu electrodes can promote the reduction of phenylhydroxylamine

43
44 to aniline, regardless the pH of the supporting electrolyte.
45 Besides the obvious effect of the acidity of the supporting electrolyte, the
Ac

46
47 electrolyte has another strong influence on the reactions taking place. This was
48
illustrated by Pintauro et al. [12]. They found that in tetraethylammonium
49
50 p-toluenesulphate, an aprotic solvent, nitrobenzene could be reduced as far as to
51
52 cyclohexylamine, whereas in protic solvents aniline could not be reduced further.
53 The stability of the intermediates has to be considered as well. Gros̃ková
54
55 et al. [13] determined that phenylhydroxylamine can easily disproportionate to
56
aniline and nitrosobenzene (eq. 4) in the presence of clay catalysts. Because
57
58
59 3
60
61
62
63
64 Page 3 of 19
65
 1
2
3
4
5
6
7
8
9 nitrosobenzene can react with an extra phenylhydroxylamine towards azoxy-
10
benzene (eq. 5), an equal amount of azoxybenzene and aniline is found in the
11
12 solution. Because of this distribution of end-products, the disproportionation of

t
13

ip
14 phenylhydroxylamine is not selective, making it undesirable in the fuel cell-type
15 reactor.
16

cr
17
18 NHOH NO NH2
19

us
20 2 −−→ + + H2 O (4)
21
22 NO NHOH
23 N
24 + −−→ N⊕ + H2 O (5)

an
25
O
26 azoxybenzene
27
28 All the described reaction mechanisms were derived from studies on bulk
M
29
electrodes. For a cogeneration fuel cell-type reactor, a suitable electrocatalyst is
30
31 needed, with high surface area and decreased costs compared to bulk electrodes.
32
33 No previous studies have been found on finding suitable metal nanoparticle
d

34 electrocatalysts for the nitrobenzene reduction. Therefore, the innovative goal of

35
te

36 this study is to find a suitable electrocatalyst and optimal reaction conditions for
37 the reduction to take place. The products that are formed during the reduction
38
p

39 are also investigated. Pt and Cu nanoparticles, supported by multiple wall

40
41 carbon nanotubes (MWCNT), are considered as electrocatalysts. Reaction
ce

42 conditions were set to favour aniline. Because aniline is only slightly soluble
43
44 in water but highly soluble in ethanol, ethanolic and acidic ethanolic solutions
45 where considered as supporting electrolyte. As the anodic reaction in a fuel cell
Ac

46
47 for cogeneration will consist of hydrogen oxidation, alkaline solutions are not
48
considered for the supporting electrolyte, even though they have been reported
49
50 as good electrolytes for the production of aniline in half-cell setups [7].
51
52
53 2. Experimental
54
55 The electrocatalysts used in this study are part of a group of catalysts specif-
56
ically developed for the nitrobenzene reduction. The two electrocatalysts present
57
58
59 4
60
61
62
63
64 Page 4 of 19
65
 1
2
3
4
5
6
7
8
9 in this study are Pt/MWCNT and Cu/MWCNT-H2 . They have been selected to
10
investigate the effect of reaction conditions. Pt/MWCNT was prepared following
11
12 a modification of a previously reported method [14, 15]. H2 PtCl6 · 6 H2 O was

t
13

ip
14 reduced by ethylene glycol and sodium acetate as stabiliser. The preparation
15 of the Cu/MWCNT-H2 catalyst consisted of the reduction of Cu(NO3 )2 by H2 .
16

cr
17 The support of both catalysts consists of commercially available multiwalled
18
carbon nanotubes (Nanocyl, 90%) with an average diameter of 9nm and lengths
19

us
20 of 20–50nm. Further information about the different catalysts can be found in
21
22 Sheng et al. [16]
23 All electrochemical measurements were carried out using an Autolab 302N
24

an
25 potentiostat or a BioLogic VMP3 multichannel potentiostat. Analysis of the
26 formed products after chronoamperometry was performed using HPLC (Shi-
27
28 madzu equipped with a Prevail C18 column, particle size 5µm, 250mm x4.6mm
M
29
I.D. and UV-vis detector SPD 10 Avp).
30
31 Porous rotating disc electrodes were used for all electrochemical measurements.
32
33 For the preparation of the electrodes, 8.0mg of electrocatalyst was dispersed
d

34 in a 1m% solution of polystyrene in toluene. 5µL of this dispersion was then

35
te

36 added in the form of a drop on a glassy carbon electrode, after which the
37
toluene was evaporated during 20min at 50◦ C. The glassy carbon electrode
38
p

39 consisted of a 6.0mm diameter glassy carbon rod encased in a 18.0mm diameter

40
41 PTFE sheet. The electrolyte used in the measurements consisted of either
ce

42 0.2M LiClO4 (Sigma-Aldrich, 98.0%) or 0.3M HClO4 (VWR, Analar Normapur,

43
44 70%) in absolute ethanol (VWR, Analar Normapur, 99.9%). Nitrogen gas was
45 bubbled through the electrolyte to remove oxygen. The temperature of the
Ac

46
47 electrolyte was set to 25◦ C using a thermostatic bath. A Pt grid was used
48
as counter electrode, while the reference electrode was a Ag/AgCl (sat. LiCl)
49
50 electrode (Autolab). The exact potential of the reference electrode was measured
51
52 by a cyclic voltammetry measurement of the ferrocene/ferrocenium (Fc/Fc+ )
53 0
reference couple (EFc/Fc + = −0.64V). All potentials in this paper are referred
54
55 to this redox couple.
56
The electrolyte resistance was calculated using the method described in [17]
57
58
59 5
60
61
62
63
64 Page 5 of 19
65
 1
2
3
4
5
6
7
8
9 using conductivity measurements, and the calculation was confirmed by the
10
impedance found in EIS measurements (ranging from 100mHz to 200kHz) at
11
12 the high frequency range. An electrolyte resistance of 190Ω and 96Ω was found

t
13

ip
14 for the 0.2M LiClO4 and the 0.3M HClO4 solutions in ethanol respectively.
15 The linear sweep voltammetry (LSV) measurements consisted of two steps;
16

cr
17 first the electrode ‘blank’ was measured three times in the solution without
18
nitrobenzene. Next, 5mM of nitrobenzene (Merck, 99%) was added, and three
19

us
20 measurements were performed in this solution. Moreover, similar tests were
21
22 carried out with nitrosobenzene (Sigma-Aldrich, 97%) as reagent instead of
23 nitrobenzene. The volume of the electrochemical cell was 50mL. The scan rate
24

an
25 was set at 5mV/s.
26 The chronoamperometry experiments were performed using an electrochemi-
27
28 cal cell that was split in two compartments, namely one for the working electrode
M
29
and one for the counter electrode. The two compartments were separated by
30
31 a Zirfonr membrane to prevent products which are formed on the working
32
33 electrode to be oxidised on the counter electrode. In this cell, the reference
d

34 electrode was placed in a compartment separated from the working electrode by

35
te

36 a fritted glass disc. Rotation speed of the working electrode was set at 500rpm.
37
In acidic solution, the potential of the working electrode was set at -0.8V vs
38
p

39 Fc/Fc+ , while in ethanolic solution the potential of the working electrode was
40
41 set at -1.35V vs Fc/Fc+ . These potentials were chosen to ensure an analysable
ce

42 conversion of nitrobenzene while avoiding the reduction of other compounds

43
44 present in the supporting electrolyte. The volume added to the working electrode
45 compartment was 170mL and the concentration of nitrobenzene was set at 15mM.
Ac

46
47 The electrodes for this experiment were prepared in the same manner as in the
48
linear sweep voltammetry measurements. Right before HPLC analysis, 1M
49
50 NaOH solution was added to the solution till pH 4–5. Afterwards this solution
51
52 was neutralised by phosphate buffer solution (pH 7) and the insoluble KClO4
53 was filtered out. Each HPLC analysis was done in triplicate and the results were
54
55 averaged.
56
The phenylhydroxylamine stability measurements were performed using
57
58
59 6
60
61
62
63
64 Page 6 of 19
65
 1
2
3
4
5
6
7
8
9 7.5mM of phenylhydroxylamine in ethanol, both under air and under nitrogen
10
atmosphere. After 52 of reaction, The formed products were identified using
11
12 HPLC analysis.

t
13

ip
14
15 3. Results and discussion
16

cr
17 3.1. Comparison of Pt and Cu catalysts
18 3.1.1. LSV measurements
19
Linear sweep voltammetry measurements were performed on both the Pt/MWCNT

us
20
21 and Cu/MWCNT-H2 nanoparticle electrocatalysts (figure 1). The voltammo-
22
23 grams for the electrolyte in absence and in presence of nitrobenzene are both
24

an
reported. Only the curves for the rotation rate of 1500rpm are given, although
25
26 the experiment was also performed with a rotation rate of 500 and 1000rpm.
27
The results from those experiments were in line with what is shown here.
28
M
29 In ethanolic solution (figure 1a), the onset potential for the nitrobenzene
30
31 reduction on the Pt electrocatalyst, calculated as the value of the potential
32 where the slope of the curve is higher than 0.1mA/V, has a value of -1.01V
33
d

34 vs Fc/Fc+ (I). On this catalyst, the nitrobenzene reduction remains the only
35 reaction that occurs until -1.33V vs Fc/Fc+ , the onset potential for the hydrogen
te

36
37 evolution (II). In this solution the Cu catalyst outperforms the Pt catalyst as
38
its onset potential is much more positive (-0.74V vs Fc/Fc+ , III). On the other
p

39
40 hand, the ethanol reduction appears at a slightly more negative potential (-1.44V
41
ce

42 vs Fc/Fc+ , IV) than on the Pt catalyst, leaving a larger potential window where
43 the nitrobenzene reduction is the only reaction taking place.
44
45 In acidic ethanolic solution (figure 1b), the curve of the Pt/MWCNT catalyst
Ac

46
shows that the hydrogen evolution starts around the same potential as the
47
48 nitrobenzene reduction, since the onset potential for the measurement in absence
49
50 of nitrobenzene is only slightly less negative than the measurement in presence
51 of nitrobenzene. On Cu/MWCNT-H2 the onset potential of nitrobenzene and
52
53 hydrogen evolution are more distinct, and the onset potential for the nitrobenzene
54 reduction is 0.30V more positive than the hydrogen evolution on Pt.
55
56
57
58
59 7
60
61
62
63
64 Page 7 of 19
65
 1
2
3
4
5
6
7
8
9 3.1.2. Discussion
10
11 In both acidic and nonacidic medium, the Cu electrocatalysts outperforms
12 the Pt electrocatalyst for the nitrobenzene reduction. The distinction is most

t
13

ip
14 visible in acidic ethanolic environment, as the nitrobenzene reduction and the
15 hydrogen evolution occur around the same potential on the Pt catalyst under
16

cr
17 these conditions. This is a surprising result, considering that Yuan et al. [1]
18
have successfully performed cogeneration of cyclohexylamine and electricity on
19

us
20 a Pt catalyst in ethanolic environment in a PEM fuel cell. The current in the
21
22 LSV measurement is however slightly higher than in absence of nitrobenzene,
23 indicating that some nitrobenzene is indeed being reduced. As it can thus
24

an
25 be concluded that the Cu catalyst is much more suited for the nitrobenzene
26
reduction than the Pt one in either environment, the electrocatalytic behaviour
27
28 of this catalyst was studied in more detail.
M
29
30
FIGURE 1
31
32
33 3.2. LSV investigation of the Cu/MWCNT-H2 catalyst
d

34 3.2.1. Reduction of nitrobenzene and intermediates

35
te

36 The electrocatalytic behaviour of Cu/MWCNT-H2 was further investigated

37 to find out which reactions are taking place. To that purpose, the number of
38
electrons exchanged in the reaction was determined by making use of Koutécky-
p

39
40 Levich plots. The linear sweep voltammetry curves as a function of the rotation
41
ce

42 speed of the electrode are shown in figure 2a for the ethanolic solution. Figure
43
2b shows the Koutécky-Levich plots of the curves shown in figure 2a at four
44
45 potentials prior to the potential at which the hydrogen evolution starts to have
Ac

46
47 a significant influence, ranging from -1.15V till -1.6V vs Fc/Fc+ . In these plots
48 the inverse of the current density j is plotted against the inverse of the square
49
50 root of the rotation speed ω. The slope of these plots is constant in this potential
51
window. Figure 3a shows the nitrobenzene reduction in acidic ethanolic solution.
52
53 The Koutécky-Levich plots in this solution have also been calculated. They are
54
55 shown in figure 3b. It can be seen that the slope of the Koutécky-Levich plots
56 remains unchanged over a potential range of -0.6V to -0.85V vs Fc/Fc+ .
57
58
59 8
60
61
62
63
64 Page 8 of 19
65
 1
2
3
4
5
6
7
8
9 FIGURE 2
10
FIGURE 3
11
12

t
13 To investigate the nitrobenzene reduction on Cu/MWCNT-H2 further, it

ip
14
15 was tried to perform linear sweep voltammetry tests using known intermediates
16 of the reduction (eq. 1) as reagents. Unfortunately, it was not possible to start

cr
17
18 the reduction from phenylhydroxylamine since this product proved too unstable
19 in the different supporting electrolytes that were tested: it converted into other

us
20
21 compounds before the LSV scan was completed. On the other hand, using
22
nitrosobenzene this was not the case. Both in ethanolic (a) and acidic ethanolic
23
24 (b) solution, the reduction of nitrosobenzene occurs at a much more positive

an
25
26 potential than the reduction of nitrobenzene (figure 4).
27
28 FIGURE 4
M
29
30 3.2.2. Discussion
31
32 The constant slope of the Koutécky-Levich plots, seen both in figures 2b and
33
d

3b, can be interpreted as a constant number of number of electrons transferred

34
35 (see the Levich equation, eq. 6). This interpretation is valid as all other
te

36
37 parameters in the Levich equation, the Faraday number F , the geometrical
38 surface A, the diffusion coefficient D, the rotation speed ω, the viscosity ν and
p

39
40 the bulk concentration C, remain constant for the duration of the experiment.
41
ce

42
43 2 1 1
44 Il = 0.620nF AD 3 ω 2 ν − 6 C (6)
45
Ac

46 In the original plot of the LSVs in acidic reaction media (figure 3a) the
47
48 diffusion limited current is not apparent, which means mixed kinetics are in play,
49
in which both electron transfer and mass transport limit the reaction in the
50
51 entire potential region where only nitrobenzene reduction occurs. The number
52
53 of electrons that is exchanged cannot be determined from the Koutécky-Levich
54 plots yet as the diffusion coefficient of nitrobenzene in the supporting electrolytes
55
56 are unknown.
57
58
59 9
60
61
62
63
64 Page 9 of 19
65
 1
2
3
4
5
6
7
8
9 From the lower onset potential of the nitrobenzene reduciton, it can be
10
concluded that in the investigated reaction conditions on this electrocatalyst the
11
12 reduction of nitrobenzene is always slower than the reduction of nitrosobenzene,

t
13

ip
14 which is in line with the results reported in most solutions in literature [2–11].
15 Looking at equation 1, it can thus be concluded that during the nitrobenzene
16

cr
17 reduction on Cu/MWCNT-H2 at least four electrons are exchanged under these
18
conditions. Comparing the slopes of the Koutécky-Levich of nitrosobenzene and
19

us
20 nitrobenzene under both acidic ethanolic an ethanolic conditions, a four-electron
21
22 reduction is suggested.
23
24 3.3. Chronoamperometric tests

an
25 3.3.1. HPLC analysis
26
27 To investigate which products are formed under the different conditions,
28 chronoamperometry combined with HPLC analysis was performed. The po-
M
29
30 tential of these tests was selected based on two criteria from the linear sweep
31
voltammograms; the current had to be sufficient to obtain a measurable conver-
32
33 sion in a time frame of around 50h, and nitrobenzene reduction had to be the
d

34
35 only reaction taking place at the chosen potential.
te

36 The reaction conditions, the conversion of nitrobenzene and the yields of

37
38 the different products are shown in table 1. It is clear that the reduction of
p

39 nitrobenzene is much faster in the acidic ethanolic environment than in the

40
41 ethanolic environment, as the conversion is 3.5 times higher with a comparable
ce

42
43 reaction time despite a much less negative applied potential. In both ethanolic
44 and acidic ethanolic experiments, azoxybenzene is the main product of the
45
Ac

46 chronoamperometric test.
47
48 3.3.2. Discussion
49
Using Faraday’s law, the conversion in the acidic ethanolic environment
50
51 compared to the total charge that passed through the system suggests a four-
52
53 electron reduction. This result is in agreement with the comparison of the
54 Koutécky-Levich slopes of nitrosobenzene in this environment. When the same
55
56 calculation is done for the ethanolic environment, the results are less clear. At
57
58
59 10
60
61
62
63
64 Page 10 of 19
65
 1
2
3
4
5
6
7
8
9
Table 1: Results of the HPLC analysis after chronoamperometry performed in a two-
10 compartment cell starting from a solution containing 15mM of nitrobenzene. The working
11 electrode was a porous rotating disc with a surface area of 28mm2 .
12

t
13 Electrolyte 0.2 M LiClO4 in EtOH 0.3 M HClO4 in EtOH

ip
14 Potential -1.15V vs Fc/Fc+ -0.62V vs Fc/Fc+
15 Reaction time 46h 52h
16
Average current -1.5mA -2.3mA

cr
17
Conversion 12.5% 44.0%
18
19 Yields:

us
20 • azoxybenzene 9.8% 36.0%
21 • aniline 2.3% 1.0%
22 • nitrosobenzene - 1.3%
23 • p-ethoxyaniline - 4.3%
24

an
• azobenzene - 0.4%
25
26
27
28 this current, with a four-electron reduction, the conversion of nitrobenzene should
M
29
30 have been 24%. This strongly suggests that some side reactions are taking place
31
that do not show in the electrochemical measurement. It is therefore impossible
32
33 to confirm the number of electrons exchanged in this environment based on the
d

34
35 chronoamperometry.
te

36 As four electrons are exchanged in the acidic ethanolic supporting electrolyte,

37
38 it can be assumed that, looking at equation 1, phenylhydroxylamine is the main
p

39 product under these circumstances. However, this product is not found in the
40
41 HPLC analysis after the chronoamperometry. This is not surprising, because
ce

42
43 phenylhydroxylamine is known to be a very unstable compound that tends to
44 undergo disproportionation and oxidation with air.
45
Ac

46 The main product of the chronoamperometry being azoxybenzene is in con-

47 trast with literature, as there seems to be no clear reaction pathway that suggests
48
49 the formation of only azoxybenzene. Azoxybenzene can be formed through the
50
condensation reaction between nitrosobenzene and phenylhydroxylamine [18].
51
52 Since nitrosobenzene is rapidly converted to phenylhydroxylamine electrochem-
53
54 ically, only trace amounts of nitrosobenzene are expected to be available for
55 this reaction. The disproportionation of phenylhydroxylamine cannot account
56
57
58
59 11
60
61
62
63
64 Page 11 of 19
65
 1
2
3
4
5
6
7
8
9 for the formed products, since it would lead to an equal amount of aniline and
10
azoxybenzene, as opposed to the observed selectivity towards azoxybenzene.
11
12

t
13 3.4. Phenylhydroxylamine stability measurements

ip
14 3.4.1. HPLC analysis
15
16 To find out how a solution of phenylhydroxylamine can generate the observed

cr
17 product distribution (table 1), stability measurements have been performed,
18
19 where phenylhydroxylamine was placed in ethanol and left to react for 52h (table

us
20 2). In an oxygen-rich environment (sample 1), the conversion of phenylhydroxy-
21
22 lamine to azoxybenzene is almost complete, indicating that phenylhydroxylamine
23
24 is oxidised into nitrosobenzene by oxygen, after which the nitrosobenzene reacts

an
25 with the phenylhydroxylamine to form azoxybenzene. Under nitrogen atmo-
26
27 sphere (sample 2), the oxidation of phenylhydroxylamine is much slower, but
28 still occurring, possibly due to residual oxygen in the sample.
M
29
30
31 Table 2: Stability test of phenylhydroxylamine (PHA (initial concentration: 7.5mM)) in
32 ethanol. XPHA : conversion of PHA. Y: yield of azoxybenzene (AOB), nitrosobenzene (NSB)
and nitrobenzene (NB).
33
d

34
35 Sample Solution Atmosphere XPHA YAOB YNSB YNB
te

36 1 EtOH air 99% 95% 4% 1%

37 2 EtOH N2 7% 3% 4% 0%
38
p

39
40
41
ce

3.4.2. Discussion
42
43 Based on these results, it can be concluded that, during the chronoampero-
44
45 metric experiment, phenylhydroxylamine is selectively formed via a four-electron
Ac

46 reduction of nitrobenzene by the Cu/MWCNT-H2 electrocatalyst in a solution

47
48 of perchloric acid in ethanol. After the experiment, when the solution comes
49
into contact with oxygen from the atmosphere, the formed phenylhydroxylamine
50
51 is entirely oxidised and condensed into azoxybenzene. As a consequence, the
52
53 overall process results in the reduction of nitrobenzene into azoxybenzene with a
54 selectivity of 82%.
55
56
57
58
59 12
60
61
62
63
64 Page 12 of 19
65
 1
2
3
4
5
6
7
8
9 4. Conclusions
10
11 The electrocatalytic behaviour of Pt and Cu nanoparticles, supported on
12

t
13 multiple wall carbon nanotubes, towards the nitrobenzene reduction in ethanolic

ip
14
and acidic ethanolic environments has been investigated. The Cu electrocatalyst
15
16 (Cu/MWCNT-H2 ) proved to be the superior electrocatalyst in both environments,

cr
17
18 since the overpotential for the nitrobenzene reduction is lower while the hydrogen
19 evolution occurs at a more negative potential, thus minimising competition with

us
20
21 the nitrobenzene reduction.
22 During the electrochemical reduction of nitrobenzene on the Cu/MWCNT-
23
24 H2 catalyst, at least four electrons are being exchanged in both ethanolic and

an
25
26 acidic ethanolic environments. In the latter environment, it can be seen that in
27 the potential region from -0.6V to -0.85V vs Fc/Fc+ , four electrons are being
28
M
29 exchanged during the nitrobenzene reduction. At more negative potentials, both
30 nitrobenzene reduction and hydrogen evolution occur. In the former environment,
31
32 in the potential range of -1.15V to -1.6V vs Fc/Fc+ , a four-electron transfer is
33
d

also suggested via Koutécky-Levich analysis, however HPLC results were unable
34
35 to confirm this finding.
te

36
37 Because of the very high electrolyte resistance, the low conversion and the high
38 potential necessary for the nitrobenzene reduction, the ethanolic environment is
p

39
40 not suited for use in a fuel-cell design reactor for cogeneration of chemicals and
41
ce

electricity. In the acidic ethanolic electrolyte, phenylhydroxylamine is formed

42
43 electrochemically. Stability measurements of phenylhydroxylamine showed that
44
45 when the solution comes into contact with oxygen from the atmosphere, azoxy-
Ac

46 benzene is formed selectively from phenylhydroxylamine. A straightforward

47
48 manner to produce azoxybenzene from nitrobenzene has thus been found, and
49 this supporting electrolyte is thus very suited for the selective cogeneration of
50
51 azoxybenzene and electricity in a fuel-cell design reactor. In the chronoampero-
52
metric measurement, a selectivity of 82% towards azoxybenzene was obtained.
53
54 This means that both a suitable electrocatalyst and a suitable environment for
55
56 the nitrobenzene reduction have been found during this study. This work shows
57
58
59 13
60
61
62
63
64 Page 13 of 19
65
 1
2
3
4
5
6
7
8
9 that the electrocatalytic reduction of nitrobenzene into azoxybenzene is possible
10
with high selectivity and forms the basis for a more detailed and quantitative
11
12 investigation of the underlying mechanism in the future.

t
13

ip
14
15 Acknowledgements
16

cr
17 The authors would like to thank the Flemish Institute for support of Scientific-
18
19 Technological Research in Industry (IWT SBO, project acronym: OCPEC,

us
20 contract number: 090027).
21
22
23 References
24

an
25 [1] X.-Z. Yuan, Z.-F. Ma, Q.-Z. Jiang, W.-S. Wu, Cogeneration of cyclohexy-
26
27 lamine and electrical power using PEM fuel cell reactor, Electrochem.
28
Commun. 3 (2001) 599.
M
29
30
31 [2] J. Marquez, D. Pletcher, A study of the electrochemical reduction of
32 nitrobenzene to p-aminophenol, J. Appl. Electrochem. 10 (1980) 567.
33
d

34
35 [3] J. Marquez, D. Pletcher, The cathodic reduction of o-halonitrobenzenes in
te

36 acidic propanol/water, Electrochim. Acta 26 (1981) 1751.

37
38
[4] T. R. Nolen, Quantitative Kinetic Expressions for the Reduction of Ni-
p

39
40 trobenzene, J. Electrochem. Soc. 135 (1988) C29.
41
ce

42 [5] T. R. Nolen, P. S. Fedkiw, Kinetic study of the electroreduction of nitroben-

43
44 zene, J. Appl. Electrochem. 20 (1990) 370.
45
Ac

46 [6] G. Kokkinidis, K. Jüttner, The Electrocatalytic influence of underpoten-

47
48 tiallead adsorbates on the reduction of nitrobenzene and nitrosobenzene on
49 silver single crystal surfaces in methanolic solutions, Electrochim. Acta 26
50
51 (1981) 971.
52
53 [7] A. Cyr, P. Huot, G. Belot, J. Lessard, The efficient electrochemical re-
54
55 duction of nitrobenzene and azoxybenzene to aniline in neutral and basic
56
aqueous methanolic solutions at devarda copper and raney nickel electrodes:
57
58
59 14
60
61
62
63
64 Page 14 of 19
65
 1
2
3
4
5
6
7
8
9 electrocatalytic hydrogenolysis of N-O and N-N bonds, Electrochim. Acta
10
35 (1990) 147.
11
12

t
13 [8] A. Cyr, P. Huot, J.-F. Marcoux, G. Belot, E. Laviron, J. Lessard, The

ip
14 electrochemical reduction of nitrobenzene and azoxybenzene in neutral and
15
16 basic aqueous methanolic solutions at polycrystalline copper and nickel

cr
17
18 electrodes, Electrochim. Acta 34 (1989) 439.
19
[9] A. Cyr, E. Laviron, J. Lessard, Electrochemical behaviour of nitrobenzene

us
20
21
and phenylhydroxylamine on copper rotating disk electrodes, J. Electroanal.
22
23 Chem. 263 (1989) 69.
24

an
25 [10] H. C. Rance, J. M. Coulson, The electrolytic preparation of p-aminophenol,
26
27 Electrochim. Acta 14 (1969) 283.
28
M
29 [11] J. Jiang, R. Zhai, X. Bao, Electrocatalytic properties of Cu-Zr amorphous
30
31 alloy towards the electrochemical hydrogenation of nitrobenzene, J. Alloys
32 Compd. 354 (2003) 248.
33
d

34
35 [12] P. N. Pintauro, J. R. Bontha, The role of supporting electrolyte during the
te

36 electrocatalytic hydrogenation of aromatic compounds, J. Appl. Electrochem.

37
38 21 (1991) 799.
p

39
40 [13] D. Gros̃ková, M. S̃tolcová, M. Hronec, Reaction of N-phenylhydroxylamine
41
ce

42 in the presence of clay catalysts, Catal. Letters 69 (2000) 113.

43
44 [14] Z. H. Zhou, S. L. Wang, W. J. Zhou, G. X. Wang, L. H. Jiang, W. Z. Li,
45
Ac

S. Q. Song, J. G. Liu, G. Q. Sun, Q. Xin, Novel synthesis of highly active

46
47 Pt/C cathode electrocatalyst for direct methanol fuel cell, Chem. Commun.
48
49 (2003) 394.
50
51 [15] G. Yu, W. Chen, J. Zhao, Q. Nie, Synthesis of highly dispersed Pt/C
52
53 electrocatalysts in ethylene glycol using acetate stabilizer for methanol
54 electrooxidation, J. Appl. Electrochem. 36 (2006) 1021.
55
56
57
58
59 15
60
61
62
63
64 Page 15 of 19
65
 1
2
3
4
5
6
7
8
9 [16] X. Sheng, B. Wouters, H. Van Parys, T. Breugelmans, A. Hubin, I. F. J.
10
Vankelecom, P. P. Pescarmona, J. Catal. (submitted).
11
12

t
13 [17] W. Oelßner, F. Berthold, U. Guth, The iR drop – well-known but of-

ip
14 ten underestimated in the electrochemical polarization measurements and
15
16 corrosion testing, Mater. Corros. 57 (2006) 445.

cr
17
18 [18] S. Oae, T. Fukumoto, M. Yamagami, The Mechanism of Azoxybenzene
19
Formation, Bull. Chem. Soc. Jpn. 36 (1963) 728.

us
20
21
22
23
24

an
25
26
27
28
M
29
30
31
32
33
d

34
35
te

36
37
38
p

39
40
41
ce

42
43
44
45
Ac

46
47
48
49
50
51
52
53
54
55
56
57
58
59 16
60
61
62
63
64 Page 16 of 19
65
 1
2
3
4
5
6
7
8
9 Figures
10
11
12

t
13 0 0

ip
a IV II I III b
14 −2 −2
15
−4 −4
16

cr
j / mA cm−2

j / mA cm−2
17 −6 −6

18 −8 −8

19 −10 −10

us
20 Cu/MWCNT−H2 (0.2M LiClO4) Cu/MWCNT−H2 (0.3M LiClO4)
−12 Cu/MWCNT−H2 (0.2M LiClO4 + NB) −12 Cu/MWCNT−H2 (0.3M LiClO4 + NB)
21 Pt/MWCNT (0.2M LiClO4) Pt/MWCNT (0.3M LiClO4)
Pt/MWCNT (0.2M LiClO4 + NB) Pt/MWCNT (0.3M LiClO4 + NB)
22 −14 −14
−1.6 −1.4 −1.2 −1 −0.8 −1 −0.9 −0.8 −0.7 −0.6 −0.5 −0.4
23 + +
E vs (Fc/Fc ) / V E vs (Fc/Fc ) / V
24

an
25
Figure 1: Comparison between Pt and Cu nanoparticles supported on MWCNT for the
26 nitrobenzene (NB) reduction in ethanolic (a) and acidic ethanolic (b) environments. Onset
27 potentials in ethanolic solution: (I) Nitrobenzene reduction on Pt catalyst, (II) H2 evolution
28 on Pt catalyst, (III) Nitrobenzene reduction on Cu catalyst, (IV) H2 evolution on Cu catalyst.
M
29 Rotation speed: 1500rpm.
30
31
32
33
d

34
35
te

36
37
38
p

39
40
41
ce

42
43
44
45
Ac

46
47
48
49
50
51
52
53
54
55
56
57
58
59 17
60
61
62
63
64 Page 17 of 19
65
 1
2
3
4
5
6
7
8
9
10
11
12

t
13

ip
0 0
14 a b
15 −2
−0.05
16 −4

cr
17

j−1 / mA−1cm2
j / mA cm−2

−0.1
−6
18
−8
19 −0.15

us
20 −10

21 −12
0.2M LiClO4 (500rpm)
0.2M LiClO4 + 5mM NB (500rpm)
−0.2
−1.15V
−1.25V
22 0.2M LiClO4 + 5mM NB (1000rpm)
0.2M LiClO4 + 5mM NB (1500rpm)
−1.4V
−1.6V
−14
23 −1.8 −1.6 −1.4 −1.2 −1 −0.8
−0.25
0 0.05 0.1 0.15 0.2
24 +
ω−1/2 / rad−1/2

an
E vs (Fc/Fc ) / V
25
26 Figure 2: Nitrobenzene reduction in ethanolic environment. The supporting electrolyte consists
27 of 0.2M LiClO4 . After three scans in the supporting electrolyte (only one is shown), 5mM
28 nitrobenzene was added. Figure a shows the linear sweep voltammetries at 500, 1000 and
M
1500rpm. Figure b shows the Koutécky-Levich plots of the nitrobenzene reduction at 4 selected
29
potentials, prior to the hydrogen evolution.
30
31
32
33
d

34
35
te

36
37
38
p

39 0 0
40 a b
−2
41 −0.1
ce

42 −4
−0.2
j−1 / mA−1cm2

43
j / mA cm−2

−6
44 −0.3
−8
45
Ac

−0.4
46 −10
0.3M HClO4 (500rpm) −0.6V
47 −12 0.3M HClO4 + 5mM NB (500rpm) −0.5 −0.7V
0.3M HClO4 + 5mM NB (1000rpm) −0.8V
48 −14
0.3M HClO4 + 5mM NB (1500rpm) −0.85V

49 −1 −0.9 −0.8 −0.7 −0.6 −0.5 −0.4 0 0.05 0.1 0.15 0.2

50 E vs (Fc/Fc+) / V ω−1/2 / rad−1/2

51
52 Figure 3: Nitrobenzene reduction in acidic ethanolic environment, with the supporting elec-
53 trolyte containing 0.3M HClO4 . The same procedure as in figure 2 was used.
54
55
56
57
58
59 18
60
61
62
63
64 Page 18 of 19
65
 1
2
3
4
5
6
7
8
9
10
11
12

t
13

ip
14
15
16

cr
17
18
19

us
20
21
22
23
24

an
25
26 0 0
a b
27 −2 −2
28
M
−4 −4
29
j / mA cm−2

j / mA cm−2

30 −6 −6

31 −8 −8
32
−10 −10
33
d

34 −12
0.2M LiClO4 (500rpm)
0.2M LiClO4 + 5mM NB (1500rpm)
−12 0.3M HClO4 (500rpm)
0.3M HClO4 + 5mM NB (1500rpm)

35 0.2M LiClO4 + 5mM NSB (1500rpm)

−14
0.3M HClO4 + 5mM NSB (1500rpm)
te

36 −1.6 −1.4 −1.2

+
−1 −0.8 −1 −0.8 −0.6 −0.4
+
−0.2 0
E vs (Fc/Fc ) / V E vs (Fc/Fc ) / V
37
38
Figure 4: Linear sweep scans of nitrobenzene (NB, green curves) and nitrosobenzene (NSB,
p

39
red curves) on Cu/MWCNT-H2 in ethanolic (a) and acidic ethanolic (b) environments. Scans
40 without active compound are also shown. Rotation speed is set at 1500rpm.
41
ce

42
43
44
45
Ac

46
47
48
49
50
51
52
53
54
55
56
57
58
59 19
60
61
62
63
64 Page 19 of 19
65