The removal of low level organics via hydrogen peroxide formed in a reticulated vitreous carbon cathode cell, Part 1. The electrosynthesis of hydrogen peroxide in aqueous acidic solutions

Article abstract of DOI:10.1016/S0013-4686(98)00242-4

It is demonstrated that hydrogen peroxide can be produced with a current efficiency of 40-70% by the reduction of oxygen at a reticulated vitreous carbon cathode in a divided flow cell using catholytes consisting of aqueous chloride or sulfate media, pH ≈2. The influence of ferrous salts, potential and electrolyte concentration on the current efficiency and rate of H₂O₂ production is reported; ferrous ions can lead to the homogeneous decomposition of H₂O₂ away from the cathode surface but their effectiveness as a catalyst for this decomposition depends on their speciation in solution which changes during an electrolysis. The conclusions are supported by voltammetry at both a rotating vitreous carbon disc and the reticulated vitreous carbon electrodes.

Full text of DOI:10.1016/S0013-4686(98)00242-4

PERGAMON
Electrochimica Acta 44 (1998) 853±861
The removal of low level organics via hydrogen peroxide
formed in a reticulated vitreous carbon cathode cell,
Part 1. The electrosynthesis of hydrogen peroxide in
aqueous acidic solutions
Alberto Alvarez-Gallegos, Derek Pletcher *
The Department of Chemistry, The University, Southampton SO17 1BJ, U.K.
Received 8 April 1998; received in revised form 22 May 1998
Abstract
It is demonstrated that hydrogen peroxide can be produced with a current eciency of 40±70% by the reduction
of oxygen at a reticulated vitreous carbon cathode in a divided ¯ow cell using catholytes consisting of aqueous
chloride or sulfate media, pH12. The in¯uence of ferrous salts, potential and electrolyte concentration on the
current eciency and rate of H2O2 production is reported; ferrous ions can lead to the homogeneous decomposition
of H2O2 away from the cathode surface but their e€ectiveness as a catalyst for this decomposition depends on their
speciation in solution which changes during an electrolysis. The conclusions are supported by voltammetry at both a
rotating vitreous carbon disc and the reticulated vitreous carbon electrodes. # 1998 Elsevier Science Ltd. All rights
reserved.
1. Introduction
Hydrogen peroxide is, however, a much more attrac-
tive mediator since the reactant can be oxygen from
the air and any excess hydrogen peroxide eventually
decomposes only to oxygen and water, therefore leav-
ing no residual chemicals in the environment. It is also
well known that hydrogen peroxide is a much more
powerful oxidant in the presence of a trace of Fe(II)
and such additions of Fe(II) would also be acceptable
in most environments.
The past twenty years have seen a revival in interest
in the electrosynthesis of hydrogen peroxide by the re-
duction of oxygen and processes based on both three
dimensional cathodes and gas di€usion electrodes have
been described [1±5]. Universally, the cathode for this
2e reduction of oxygen is carbon. Electrochemical
technology for the manufacture of hydrogen peroxide
solutions has, however, been limited to high pH where
the HO2 anion has good stability. The same period
has seen an extensive e€ort to develop electrochemical
processes for the destruction of organic molecules in
e‚uents and waters [6±8]. One approach for the
removal of low concentrations of organics employs the
electrogeneration of a chemical oxidising agent to react
with the pollutants and systems using chlorine and/or
hypochlorite have been available for many years [1, 4].
The application of hydrogen peroxide electrogener-
ated directly in the waste water, however, requires the
reduction of oxygen to hydrogen peroxide under quite
di€erent conditions from those used in commercial
processes for the manufacture of hydrogen peroxide.
The medium will generally be close to neutral and may
also have a low ionic strength. Indeed, if Fe(II) is to
be employed as a catalyst, the medium needs to have
pH< 4 in order for the Fe(II) to remain in solution
and a small addition of acid to the wastewater may be
essential. It is, however, possible to envisage that both
the Fe(II) and H+ required could be generated in situ
at the anode of an undivided cell generating hydrogen
* Corresponding author. Tel.: +44-1703-593519; Fax: +44-
1703-676960; E-mail: DP1@soton.ac.uk
0013-4686/98/$ - see front matter # 1998 Elsevier Science Ltd. All rights reserved.
PII: S0013-4686(98)00242-4

854
A. Alvarez-Gallegos, D. Pletcher / Electrochimica Acta 44 (1998) 853±861
peroxide at the cathode. Two groups have reported
preliminary experiments in H cells with ¯at plate cath-
odes which demonstrate that destruction of
formaldehyde [9, 10] and phenol [11] via in situ gener-
ated hydrogen peroxide is possible and both papers
include some data on acidic solutions. A more recent
paper [12] describes the ecient destruction of formal-
dehyde in a ¯ow cell with a reticulated vitreous carbon
cathode; it was shown that formaldehyde was oxidised
only to formic in alkaline solutions but in slightly
acidic media containing Fe(II) oxidation to CO2 was
complete. A Spanish Group have developed the other
practical approach based of a gas di€usion cathode
and described the complete oxidation of aniline to
inorganic products using a slightly acidic solution con-
taining Fe(II) [13±15].
Voltammetric studies of the 2e reduction of oxygen
to hydrogen peroxide at carbon cathodes are mainly
restricted to high pH solutions where the peroxide is
completely stable although Hum€ray and Taylor [16±
18] have reported a study of oxygen reduction at vitr-
eous carbon disc electrodes using both alkaline and
acidic solutions.
In this paper, we report a systematic investigation,
using both a vitreous carbon rotating disc electrode
and a ¯ow cell with a reticulated vitreous carbon cath-
ode of the reduction of oxygen at in slightly acidic sol-
utions with low ionic strengths. Most experiments were
carried out with solutions, pH 2, sometimes also with
a low concentration of sodium chloride or sulfate. The
proton concentration was selected so that oxygen re-
duction did not produce a signi®cant pH change close
to the cathode surface and also so that the solution
had sucient conductivity even in the absence of
sodium salt for an acceptable potential/current distri-
bution within the three dimensional cathode.
Fig. 1. Sketches of (A) the ¯ow cell with reticulated vitreous
carbon cathode and (B) the ¯ow circuit.
2.2. Analysis
The hydrogen peroxide was determined by titration
with potassium permanganate using
procedure [19].
a
standard
2. Experimental
2.3. Electrochemical cells and instrumentation
2.1. Chemicals
Some voltammetric experiments were carried out in
a three electrode, three compartment glass cell with a 4
mm diameter vitreous carbon disc electrode (Pine
Instruments), a Pt gauze counter electrode and a lab-
oratory constructed saturated calomel electrode as the
reference electrode. All other voltammetry as well as
preparative experiments were carried out in a ¯ow cell,
see Fig. 1A, with a 50 mm  50 mm  12 mm reticu-
lated vitreous carbon (The Electrosynthesis Co., 60
pores per inch) cathode, separated from a Pt gauze
anode by a Na®on1 450 cation permeable membrane
(Aldrich). A plastic tube Luggin capillary entered the
three dimensional cathode via a hole drilled through
All solutions were prepared with deionised water
(resistivity>18 MO cm) from a MilliQ-MilliRho puri®-
cation system using hydrogen peroxide (Hogg, 20%),
sodium sulfate (BDH, Analar), sulfuric acid (BDH,
98%), sodium chloride (BDH, Analar), hydrochloric
acid (Fisons, speci®c gravity 1.18), ferrous sulfate
(Prolabo, 98%), ferrous chloride (Aldrich, 99%), ferric
chloride (Fisons, 97%), ferric sulfate (Aldrich, 97%)
and potassium permanganate (BDH Analar, 99.5%).
All solutions were saturated with either nitrogen or
oxygen (both BOC).

A. Alvarez-Gallegos, D. Pletcher / Electrochimica Acta 44 (1998) 853±861
855
3. Results and discussion
3.1. Voltammetry at a rotating vitreous carbon disc
electrode
Relatively little is known about the reduction of
oxygen at carbon in aqueous solutions of low acidity
and low ionic strength or the in¯uence of iron species
on the reaction. Hence a preliminary study was under-
taken using a rotating disc electrode.
Fig. 2 shows voltammograms recorded at 900 rpm
for an aqueous solution of 10 mM HCl + 0.1 M NaCl
saturated with nitrogen and oxygen respectively. In the
absence of oxygen, curve (a), little current is observed
positive to 1200 mV where H2 evolution
‡
Fig. 2. Voltammograms at a rotating vitreous carbon disc
electrode (area 0.13 cm2). (a) and (b) are for solutions of 10
mM HCl+ 100 mM NaCl saturated with N2 and O2 respect-
ively. (c) and (d) are for solutions of 10 mM HCl saturated
with N2 and O2 respectively. Rotation rate 900 rpm. Potential
scan rate 5 mV s 1. The inset shows the variation of the lim-
iting current for the reduction of O2 in 10 mM HCl+ 100
mM NaCl with the rotation rate of the disc.
2H ‡ 2e 4H2
can be seen to commence. In the presence of oxygen,
curve (b) the voltammogram shows a well formed but
rather drawn out reduction wave, E 1/2
and E1/4 E3/4 =220 mV. In contrast, when 10 mM
H2O2 was added to the deoxygenated solution, the vol-
tammogram was identical to that for the background
electrolyte alone; hydrogen peroxide does not reduce
on vitreous carbon in this solution at potentials posi-
tive to those required for H2 evolution. The expected
reduction of oxygen at carbon is [11, 16±18]
=
790 mV
‡
O2 ‡ 2H ‡ 2e 4H2O2
the cathode current contact. During electrolyte ¯ow
through the cell, solution passed through the Luggin
probe into the reference electrode compartment; hence,
this compartment was ®tted with a drain which re-
cycled the electrolyte to the reservoir. The reference
electrode was a laboratory constructed saturated calo-
mel electrode (SCE). The anolyte was 1 M H2SO4 and,
in most experiments, the catholyte volume was 2 l.
This cell and the ¯ow circuit, Fig. 1B, have been fully
described in an earlier paper [12].
and this will be con®rmed later as the major reaction.
The variation of the limiting current for oxygen re-
duction, measured at 1050 mV, is shown in the inset
to Fig. 2. It can be seen that at low rotation rates, the
limiting current is proportional to the square root of
the rotation rate. Moreover, the measured currents are
close to those estimated from the Levich equation
5
(assuming n = 2, c = 1.3 mM [20], D= 2.2Â 10
cm2
s
1). At rotation rates above 900 rpm, however,
the plot of IL vs o 1/2 is non-linear and the reduction
of oxygen is only partially mass transfer controlled in
these conditions.
All experiments were carried out at controlled poten-
tial using a HiTek potentiostat, model DT2101, and,
where appropriate, a HiTek function generator, model
PPR1, and a Gould Advance x±y recorder, model
60000. The rotation of the disc electrode was con-
trolled with an EG and G Parc unit, model 616 and
charges were measured on a laboratory constructed
digital integrator. Solution pH was determined with a
portable M90 pH-meter from Mettler Toledo. The
rotating disc voltammetry was carried out at 298 K
using a jacketed cell and water circulated from a ther-
mostat bath supplied by Grant Instruments, model
SE15. Flow cell experiments were carried out in ambi-
ent conditions but the temperature settled down to
300 2 K after a few minutes of electrolysis.
Fig. 2 also shows the responses when the voltammo-
grams were repeated in a solution containing only 10
mM HCl. In this low ionic strength solution, the re-
duction currents for oxygen reduction, curve (d), are
similar but the wave is less well de®ned. This is for
two reasons. Firstly, E 1/2 is shifted more negative by
070 mV. Secondly, and very surprisingly, H2 evolution
occurs signi®cantly more readily in this solution than
in 10 mM HCl + 0.1 M NaCl. This can also be seen
in the voltammogram for the deoxygenated solution,
curve (c), where the negative potential limit is shifted
positive by 0130 mV. The pH of both solutions were
measured and found to be 2.1 and 2.0 for the solution
with and without the added NaCl respectively. Hence,

856
A. Alvarez-Gallegos, D. Pletcher / Electrochimica Acta 44 (1998) 853±861
Fig. 3. Voltammograms recorded at a rotating vitreous car-
bon disc electrode (area 0.13 cm2) for solutions of O2 in 10
mM HCl+ 100 mM NaCl containing (a) no added Fe salt
(b) 1 mM Fe(II) and (c) 2 mM Fe(III). Rotation rate 900
1
Fig. 4. Voltammograms recorded at a 50 mm  50 mm 12
mm reticulated vitreous carbon electrode (60 ppi) in solutions
of (a) O2 saturated 10 mM HCl (b) N2 saturated 10 mM HCl
(c) 6 mM H2O2 in 10 mM HCl (d) O2 saturated 10 mM
HCl+ 500 mM NaCl (e) N2 saturated 10 mM HCl+ 500
rpm. Potential scan rate 5 mV s
.
the shift is not associated with a pH di€erence; nor
does it result from uncompensated IR drop. It would
appear that the presence of NaCl inhibits the kinetics
of H2 evolution.
mM NaCl. Catholyte ¯ow rate 0.13 m s 1. Potential scan
1
rate 5 mV s
.
The in¯uence of Fe(II) and Fe(III) on the reduction
of oxygen in 10 mM HCl + 100 mM NaCl is reported
in Fig. 3 (and voltammograms in sulfate solution are
very similar). It can be seen that the addition of either
2 mM Fe(III) or 1 mM Fe(II) has little in¯uence on
the reduction wave for oxygen although the presence
in solution of an iron species does lead to the catalysis
of H2 evolution and this results from the deposition of
some iron metal on the vitreous carbon surface at po-
tentials more negative than 900 mV. The presence of
Fe(III) results, however, in a reduction wave for the
process Fe3+4Fe2+ at E1/2 = + 150 mV. Although
the concentrations of Fe(III) and O2 are similar (2 and
1.3 mM [20], respectively), the mass transfer controlled
limiting currents are quite di€erent because of di€erent
wide potential range negative to 200 mV, the current
is much higher than for the deoxygenated solution,
curve (b). Indeed, the current increases smoothly with
potential and reaches a value 01.1 A (37 mA cm 3) at
900 mV. This compares with the expected mass
3
transfer limited current of 115 mA cm
for a 1.3 mM
solution of O2 calculated using n = 2 and a value of
kmAe = 0.44 s 1, estimated using a value from [21]
and allowing for the large di€usion coecient for O2.
The degree to which the response is distorted by IR
drop is unknown but it is certainly clear that the use
of the three dimensional electrode permits the re-
duction of oxygen at a high rate. It should also be
noted that the cathodic current for oxygen reduction is
observed over the same potential range as at the rotat-
ing vitreous carbon disc electrode.
5
5
di€usion coecients (0.5 10
and 2.2 Â 10
cm2
s
1, respectively). The experiments also showed that
the reaction of oxygen with Fe(II) is very slow in these
solutions ±the voltammogram for the solution contain-
ing Fe(II) + O2 did not change over several hours.
When the experiments were repeated in sulfate
media, the voltammograms were very similar to those
reported above for chloride media.
Fig. 4 reports two further results. Curve (c) for 6
mM H2O2 in 10 mM HCl shows that hydrogen per-
oxide does not reduce on reticulated vitreous carbon,
at least positive to 1000 mV where H2 evolution com-
mences. Curves (d) and (e) illustrate the e€ect of
recording voltammograms for solutions with a large
excess of sodium chloride. It can be seen that the ad-
dition of 0.5 M NaCl (which changes the pH from 2.1
to 1.7) decreases substantially the current for H2 evol-
ution. This change in the voltammograms is the same
as observed with the rotating vitreous carbon disc elec-
trode but it is contrary to that predicted from con-
siderations of pH or uncompensated IR drop or
structure of the double layer. The voltammogram for
the oxygen saturated solution also shows less current
(e.g. 0.7 A cf 1.1 A at 900 mV) after the addition of
the 0.5 M NaCl. This results partly from a decrease in
3.2. Voltammetry at a reticulated vitreous carbon
cathode in a ¯ow cell
The voltammetry was repeated in the cell with a 60
ppi reticulated vitreous carbon cathode (50 mm  50
mm  12 mm) and a mean linear catholyte ¯ow rate
of 0.13 m s 1. Typical voltammograms in 10 mM HCl
are reported in Fig. 4. The oxygen saturated solution
gives a rather featureless response without a clear re-
duction wave, curve (a). On the other hand, over a

A. Alvarez-Gallegos, D. Pletcher / Electrochimica Acta 44 (1998) 853±861
857
Fig. 5. Plots of hydrogen peroxide formation in the catholyte versus charge passed during electrolyses of 10 mM HCl+ 50 mM
NaCl continuously saturated with oxygen. The electrolyses were carried out at constant potentials of (a) 500 mV, (b) 600 and
1
700 mV vs SCE. Catholyte ¯ow rate 0.13 m s
.
the H2 evolution current and partly from the decreased
solubility of O2 in the higher ionic strength medium.
The addition of 1 mM Fe(II) to either electrolyte
medium did not change the voltammograms for the re-
duction of oxygen. Moreover, the responses in sulfate
media were essentially identical to those shown above
for chloride solutions.
within the catholyte for three electrolyses carried out
in 10 mM HCl + 50 mM NaCl at di€erent potentials.
It can be seen that the concentration of hydrogen per-
oxide increases linearly with time and the slopes of the
plots are not a strong function of potential. In all
three electrolyses, the current eciency is 60 3% for
the formation of hydrogen peroxide. Moreover in
longer timescale electrolyses, the concentration of
hydrogen peroxide could be taken to >20 mM.
3.3. Electrolyses in the absence of Fe(II)
Table 1 and 2 report data from more extensive set
of experiments, carried out as a function of potential
and ionic strength in both chloride and sulfate sol-
utions of pH close to 2. Over an extended range of
conditions, the current eciency fell within the range
50±70% and electrolyte ¯ow rate also has no signi®-
cant e€ect. At the most negative potentials, the current
eciency dropped due to competing H2 evolution. As
expected from the voltammetry, this was most marked
Electrolyses were carried out in the ¯ow cell with a
reticulated vitreous carbon electrode and a cation per-
meable membrane. In all experiments, the catholyte
was continuously saturated with oxygen by bubbling
the gas through the reservoir and the hydrogen per-
oxide concentration was monitored periodically by
titrating samples with potassium permanganate. Fig. 5
reports the build up of hydrogen peroxide with time
Table 1. Cell currents, I, current eciencies, f and rates of H2O2 generation, R, in chloride solutions as a function of electrode po-
tential and ionic strength
E/V vs SCE 10 mM HCl
10 mM HCl+
50 mM NaCl
10 mM HCl+
100 mM NaCl
10 mM HCl+
500 mM NaCl
1
1
1
)
1
)
I/A f (%) R (mmoles s
)
I/A f (%) R (mmoles s
)
I/A f (%) R (mmoles s
I/A f (%) R (mmoles s
400
500
600
700
800
900
0.17 57
0.32 56
0.44 48
0.61 41
0.82 16
0.54
0.90
1.06
1.25
0.66
0.15 61
0.25 61
0.39 62
0.55 57
0.64 49
0.85 39
0.46
0.76
1.21
1.57
1.57
1.66
0.12 43
0.26 62
0.35 63
0.44 63
0.57 56
0.73 48
0.26
0.81
1.10
1.39
1.62
1.75
0.12 39
0.27 54
0.33 51
0.42 58
0.57 58
0.60 58
0.24
0.86
0.84
1.22
1.65
1.74
Flow cell with 60 ppi reticulated vitreous carbon cathode. Mean linear catholyte ¯ow rate 0.13 m s 1. Temperature 300 K.

858
A. Alvarez-Gallegos, D. Pletcher / Electrochimica Acta 44 (1998) 853±861
Table 2. Cell currents, I, current eciencies, f and rates of H2O2 generation, R, in sulfate solutions (columns 1±4) for 10 mM
H2SO4 +50 mM Na2SO4 as a function of electrode potential and (columns 5±8) at 500 mV vs SCE as a function of sodium sul-
fate concentration added to 10 mM H2SO4
1
)
1
)
E/V vs SCE
I/A
f (%)
R (mmoles s
[Na2SO4] (mM)
I/A
f (%)
R (mmoles s
400
0.13
62
68
61
63
52
36
0.40
0.88
0.94
1.35
1.43
1.39
0
50
0.30
50
68
62
69
0.75
0.88
0.81
0.79
500
600
700
800
900
0.26
0.29
0.43
0.55
0.77
0.26
0.26
0.23
100
500
Flow cell with 60 ppi reticulated vitreous carbon cathode. Mean linear catholyte ¯ow rate 0.13 m s 1. Temperature 300 K.
in solutions without the addition of a sodium salt to
increase the ionic strength. Clearly, the current ecien-
cies are always signi®cantly below 100% and this is in
contrast to electrolyses in NaOH solutions where the
current eciencies were generally >90% [12, 21].
Hydrogen peroxide was found to be stable over many
hours in these pH 2 solutions even when ¯owing
through the cell on open circuit and hence the lower
current eciencies cannot be attributed to homo-
geneous chemical decomposition. Moreover, H2O2
does not reduce at the vitreous carbon cathode, see the
voltammogram shown as Fig. 4(c). Therefore, it must
be concluded that at these vitreous carbon surfaces,
oxygen reduction occurs by a mechanism where there
is competition between a 2e pathway and a 4e
pathway without H2O2 as a discrete intermediate; simi-
lar conclusions have been drawn by other
authors [11, 16±18]. The ratio of the two pathways is
independent of the potential and it is likely that they
correspond to reduction at two di€erent types of site
on the carbon surface. The tables also reports the aver-
age cell currents during the electrolyses, which, of
course increase as the potential is made more negative
that the current eciency for the production of hydro-
gen peroxide was negligible over an extended period of
time. It was also found that the addition of 1 mM
Fe(II) to solutions of hydrogen peroxide in either
chloride or sulfate media, pH12, led to the homo-
geneous decomposition over a period of a few minutes.
On the other hand, it was reported in Sections 3.1
and 3.2, that the addition of Fe(II) to the solutions
had little in¯uence on the voltammetry. If the Fe (II)
catalysed disproportionation of hydrogen peroxide
were rapid, one would expect to a change from 2e
waves to 4e reduction waves, re¯ecting the regener-
ation of oxygen with the boundary layer at the elec-
trode surface. Moreover, conclusive evidence for the
formation of hydrogen peroxide throughout the elec-
trolyses in the presence of Fe(II) results from studies
of the in situ oxidation of organic pollutants during
the reduction of oxygen [22]; a number of organic pol-
lutants are oxidised with good eciency. Hence, it
must be concluded that hydrogen peroxide is, indeed,
formed at the reticulated vitreous carbon cathode even
in the presence of Fe(II) but it decomposes in solution
in a reaction which is slow enough that it occurs away
from the cathode surface.
as well as the rates of production of H2O2 in mmoles
1
s
since in e‚uent treatment applications this may be
It was also found during extended electrolyses that
eventually the hydrogen peroxide concentration in the
catholyte began to rise. Fig. 6A shows a plot of hydro-
gen peroxide concentration versus time for an electro-
lysis carried out at 700 mV vs SCE and a catholyte
consisting of 50 mM Na2SO4 + 1 mM Fe(II), adjusted
to pH 1.6 with H2SO4, with data from a similar exper-
iment without Fe(II) for comparison. The cell current
remained at 0.4 A throughout both electrolyses. In the
presence of Fe(II), the current eciency for hydrogen
peroxide is close to 0% during the passage the ®rst
4000 C of charge (0100 min). Thereafter, the current
eciency begins to improve and by 12 000 C, it is
030%. The ®gure also shows that the addition of a
further aliquot of Fe(II) equivalent to 1 mM after 500
min leads to the rapid destruction of all the hydrogen
peroxide in the catholyte as well as preventing the ac-
cumulation of further hydrogen peroxide for another
more important than the current eciency when values
are so similar over a range of conditions. With the
exception of one solution, the highest rate of H2O2
production always occurs at 900 mV; the increase in
current is more important than the loss in current e-
ciency.
It should also be stressed that no bulk pH change
was observed during the electrolysis; this is to be
expected since 1H + /1e passes through the mem-
brane and 1H + /1e is consumed at the cathode. The
local increase in pH at the cathode surface will also be
‡
small since cH > cO
.
2
3.4. Electrolyses in the presence of Fe(II)
When the electrolyses were repeated with 1 mM
Fe(II) added to the catholytes studied, it was found

A. Alvarez-Gallegos, D. Pletcher / Electrochimica Acta 44 (1998) 853±861
859
Fig. 6. Electrolyses at 700 mV vs SCE of 50 mM Na2SO4 adjusted to pH 1.6 continuously saturated with O2 at a reticulated vitr-
eous carbon cathode in a ¯ow cell. (A) Plots of H2O2 formed versus charge passed for solutions with Q 0 mM added Fe(II) . 1
mM added Fe(II) R further 1 mM Fe(II) added. (B) concentrations of Q Fe(II) . Fe(III) R added Fe no longer present as soluble,
electroactive species as a function of charge passed.
extended period. During the electrolysis, aliquots of
the catholyte were taken and voltammogams were
recorded for each at a rotating vitreous carbon disc
electrode between the potential limits, +1100 and
400 mV vs SCE. The oxidation and reduction waves
observed at E1/2 = + 780 and +15 mV were used to
estimate the concentrations of Fe(II) and Fe(III) re-
spectively in the catholyte as a function of time during
the electrolysis. The results are shown in the inset to
Fig. 6B. It can be seen that the concentration of Fe(II)
decays steadily and has dropped to zero after about
4000 C. In contrast, the concentration of Fe(III)
increases over this period of electrolysis and then
passes through a broad maximum until it is almost
zero after 13 000 C. It is clear, however, that the total
concentration of electroactive Fe species in the catho-
lyte drops throughout the electrolysis and these data
are also shown in Fig. 6B. The addition of a further 1
mM Fe(II) to the catholyte after 13 000 C leads to the
instant conversion of most of this Fe(II) to Fe(III).
The total loss of Fe(II) from the catholyte coincides
with the formation of some stable hydrogen peroxide
and it would appear that the Fe(II) is a much more
ecient catalyst than the Fe(III) for the decomposition
of hydrogen peroxide.

860
A. Alvarez-Gallegos, D. Pletcher / Electrochimica Acta 44 (1998) 853±861
Fig. 7. Electrolyses at 700 mV vs SCE of 10 mM HCl+ 10 mM NaCl continuously saturated with O2 at a reticulated vitreous
carbon cathode in a ¯ow cell. (A) Plots of H2O2 formed versus charge passed for solutions with Q 0 mM added Fe(II) . 1 mM
added Fe(II) R further 1 mM Fe(II) added. (B) concentrations of Q Fe(II) . Fe(III) R added Fe no longer present as soluble, elec-
troactive species as a function of charge passed.
Fig. 7 reports the results from an identical exper-
iment using catholyte consisting of 10 mM
oxygen and/or hydrogen peroxide and this results in a
loss of catalytic activity for the decomposition of
hydrogen peroxide. Thereafter, the Fe(III) is also con-
verted to an electroinactive form. Most likely the
Fe(III) precipitates as a ferric oxide/hydroxide solid or
colloid or adsorbs on a surface within the ¯ow system
although a precipitate cannot be seen in the catholyte.
a
NaCl+ 10 mM HCl both with and without the ad-
dition of 1 mM Fe(II). Although the current eciency
for hydrogen peroxide production is always lower than
in the sulfate medium, the variation of hydrogen per-
oxide concentration with charge passed is similar to
that in sulfate medium. Fig. 7B shows the charge evol-
ution of Fe(II) and Fe(III) concentrations in the chlor-
ide medium and this also follows the trends noted
above for the sulfate solution.
4. Conclusions
Therefore, in both media it is apparent that the
Fe(II) is converted to Fe(III) by reaction either with
The voltammetry and preparative electrolyses both
con®rm that hydrogen peroxide may be formed by

A. Alvarez-Gallegos, D. Pletcher / Electrochimica Acta 44 (1998) 853±861
861
[3] F.C. Foller, R.T. Bombard, J. Appl. Electrochem. 25
(1995) 613.
cathodic reduction of oxygen at vitreous carbon in
slightly acidic, aqueous solutions. The hydrogen per-
oxide may be prepared in 10 mM acid without other
electrolytes and although the addition of electrolyte
does improve the current eciency to some extent, it
does not increase the rate of production. The current
eciency at pH12 is, however, always signi®cantly
below 100% and the current eciency is not a strong
function of either potential or the catholyte compo-
sition; this is consistent with 2e and 4e reductions
of oxygen occurring in parallel at di€erent sites on the
carbon surface.
The presence of Fe(II) in the catholyte does not pre-
vent the formation of hydrogen peroxide at the cath-
ode but can catalyse the homogeneous decomposition
and in the presence of 1 mM Fe(II) the half life of the
hydrogen peroxide is minutes. The ability of the iron
species in solution to catalyse hydrogen peroxide de-
composition is, however, very dependent on its form
which changes during the electrolysis; while freshly
added Fe(II) is an ecient catalyst for hydrogen per-
oxide decomposition, its e€ectiveness decreases with
electrolysis time.
[4] P. Tatapudi, J.N. Fenton, 1994. In: Sequeira, C.A.C.
(Ed.), Environmental Orientated Electrochemistry.
Elsevier, Amsterdam.
[5] C. Oloman, 1996. Electrochemical Processing for the
Pulp and Paper Industry. The Electrochemical
Consultancy, Romsey, UK.
[6] J.D.
Genders,
Electrochemistry for
N.L.
Weinberg,
(Eds.),
1992.
Cleaner Environment. The
a
Electrosynthesis, Lancaster, New York.
[7] K. Rajeshwar, J. Ibanez, 1997. Environmental
Electrochemistry: Fundamentals and Applications in
Pollution Abatement. Academic Press, San Diego.
[8] C. Comninellis, 1994. In: Sequeira, C.A.C. (Ed.),
Environmental Orientated Electrochemistry. Elsevier,
Amsterdam.
[9] J.-S. Do, C.-P. Chen, J. Electrochem. Soc. 140 (1993)
1632.
[10] J.-S. Do, C.-P. Chen, J. Appl. Electrochem. 24 (1994)
936.
[11] M. Sudoh, T. Kodera, K. Sakai, J.Q. Zhang, K. Koide,
J. Chem. Eng. Jpn. 19 (1986) 513.
[12] C. Ponce de Leon, D. Pletcher, J. Appl. Electrochem. 25
(1995) 307.
[13] E. Brillas, R.M. Bastida, E. Llosa, J. Casado, J.
Electrochem. Soc. 142 (1995) 1733.
Certainly, the use of a three dimensional electrode
fabricated from reticulated vitreous carbon allows
hydrogen peroxide to be produced at pH12 at a rate
which makes possible, for example, the use of this
approach for e‚uent treatment [22].
[14] E. Brillas, E. Mur, J. Casado, J. Electrochem. Soc. 143
(1996) L49.
[15] E. Brillas, R. Sauleda, J. Casado, J. Electrochem. Soc.
144 (1997) 2374.
[16] R.J. Taylor, A.A. Hum€ray, J. Electroanal. Chem. 64
(1975) 63.
[17] R.J. Taylor, A.A. Hum€ray, J. Electroanal. Chem. 64
(1975) 85.
Acknowledgements
[18] R.J. Taylor, A.A. Hum€ray, J. Electroanal. Chem. 64
(1975) 95.
The authors would like to thank the Consejo
Nacional de Ciencia y Tecnologõa de Mexico for sup-
port of the work through a grant to A. A.-G.
[19] I.M. Koltho€, R. Belcher, 1957. Volumetric Analysis,
vol. 3. Titration Methods: Oxidation Reduction
Reactions. Interscience, New York.
[20] E.W. Washburn, (Ed.), 1928. International Critical
Tables of Numerical Data in Physics, Chemistry and
Technology, vol. III. McGraw-Hill, p. 272.
[21] D. Pletcher, I. Whyte, F.C. Walsh, J.P. Millington, J.
Appl. Electrochem. 21 (1991) 659.
References
[1] D.
Pletcher,
F.C.
Walsh,
1990.
Industrial
Electrochemistry. Chapman and Hall, London.
[2] J.A. McIntyre, Interface 4 (1995) 29.
[22] D. Pletcher, A. Alvarez-Gallegos, in preparation.