Continuous electrooxdiation of sulfuric acid on boron-doped diamond electrodes

Article abstract of DOI:10.1016/j.electacta.2014.09.133

This study reports on the electrochemical oxidation of highly concentrated sulfuric acid by Diachem boron-doped diamond electrodes. The scope of this work was to evaluate a proposed continuous electrooxidation process in order to prepare a resist removal

Full text of DOI:10.1016/j.electacta.2014.09.133

Electrochimica Acta 147 (2014) 589–595
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Continuous electrooxdiation of sulfuric acid on boron-doped diamond
electrodes
b
b
a
Felix Hippauf a, Susanne Dörfler a, , Ralf Zedlitz , Alfred Vater , Stefan Kaskel
*
a
Department of Inorganic Chemistry, Dresden University of Technology, Bergstrasse 66, 01069 Dresden, Germany
Infineon Technologies Dresden GmbH, Königsbrücker Strasse 180, 1099 Dresden, Germany
b
A R T I C L E I N F O
A B S T R A C T
This study reports on the electrochemical oxidation of highly concentrated sulfuric acid by Diachem1
boron-doped diamond electrodes. The scope of this work was to evaluate a proposed continuous
electrooxidation process in order to prepare a resist removal bath that contains equivalent amounts of
peroxosulfuric compounds as a common SPM (sulfuric acid and hydrogen peroxide mixture) bath. The
electrochemical processes at the electrode surface were investigated by cyclic voltammetry and by
titration of the reaction products under real working conditions (> 80 wt% H2SO4). Furthermore, an
alternative electrolysis test cell setup, exhibiting a thin and mechanically stable ion exchange membrane
will be used. For this process only one electrolyte circuit is performed. The initial electrolysis product
H2S2O8 immediately undergoes hydrolysis and forms H2SO5. The obtained concentration of H2SO5 was
similar to a reference SPM bath (> 0.14 mol lꢀ1 H2SO5 in 90 wt% H2SO4), even though the current efficiency
was lower than for more diluted solutions. This can be attributed to a complex interplay of side and
consumption reactions of peroxosulfuric compounds. This electrolysis bath has the potential to be a very
promising alternative to standard SPM baths, as it operates as a continuous and sustainable process.
ã 2014 Published by Elsevier Ltd.
Article history:
Received 6 June 2014
Received in revised form 5 September 2014
Accepted 5 September 2014
Available online 28 September 2014
Keywords:
Boron-doped diamond electrode
Peroxomonosulfuric acid
Electrooxdiation
SPM bath
Nafion
1. INTRODUCTION
peroxodisulfuric production depends strongly on the nature of the
anode material. As the discharge of water is a possible side
Peroxosulfuric compounds are widely used as bleaching and
cleaning agents due to their strong oxidizing properties. The SPM
(sulfuric acid and hydrogen peroxide mixture) process is used in
the semiconductor industry to remove patterned photoresist after
lithography. The mixing of hydrogen peroxide and sulfuric acid
forms Caro's acid (H2SO5, peroxomonosulfuric acid), which
oxidizes the resist [1]. However, a SPM bath becomes diluted by
adding more hydrogen peroxide leading to lesser oxidizing ability.
Thus, the SPM bath has to be removed, which causes problems in
sustainability and long down times. Therefore, it is highly desirable
to replace the SPM bath by a continuously working process and in
situ regeneration of the oxidant. The electrolysis of sulfuric acid on
platinum was often used to produce peroxodisulfuric acid
(H2S2O8), which was an important intermediate in the production
of hydrogen peroxide [2,3]. Peroxodisulfuric acid is known to form
peroxomonosulfuric acid at low pH-values. Sulfuric acid is
recovered after Caro's acid is consumed, hence the process can
in principle be run continuously [4–6]. The efficiency of the
reaction, anodes with a high overpotential for oxygen evolution are
required. Usually, carbon electrodes show high overpotentials, but
they lack the necessary electrochemical stability towards sulfuric
acid. In recent years, synthetic boron-doped diamond (BDD)
electrodes have been intensively investigated as a promising
electrode material due to their high electrochemical stability [7]
and high overpotential for both oxygen and hydrogen evolution
[8,9]. Outer sphere reactions proceed completely reversible [10],
whereas inner sphere reactions are strongly inhibited at diamond
surfaces and only take place in the potential region of water
discharge [11]. This is part of the special electrochemistry of BDD in
aqueous media, which can basically be attributed to the in-situ
formation of loosely adsorbed hydroxyl radicals at the electrode
surface [12,13]. The initial step is the decomposition of water to
adsorbed hydroxyl radicals (Eq. (1)):
H2O + BDD ! BDD(OHꢁ) + H+ + eꢀ
(1)
These radicals desorb and exist in a thin diffusion layer at the
electrode due to weak adsorption properties of diamond. They
can recombine to hydrogen peroxide or react with a suitable
reduction agent in a following reaction step [14]. This can lead to
the formation of e. g. exotic and strong oxidation agents such as
* Corresponding author. Tel.: +49 351 463 32029; fax: +49 351 463 37287.
E-mail address: Susanne.Doerfler@chemie.tu-dresden.de (S. Dörfler).
http://dx.doi.org/10.1016/j.electacta.2014.09.133
0013-4686/ã 2014 Published by Elsevier Ltd.

590
F. Hippauf et al. / Electrochimica Acta 147 (2014) 589–595
Ag(II) (AgO) [15], Fe(VI) (FeO42ꢀ) [16], or peroxodisulfate [17]. The
assumed electrochemical mechanism on BDD surfaces is different
compared to the mechanism of commonly used platinum
electrodes (Diachem1) supplied by Condias. The boron-doped
diamond was coated on p-silicon substrates using hot-filament
chemical vapor deposition. The electrodes were inserted in PVDF
blocks and pressed against a PVDF frame. The resulting electrolysis
chamber had a geometric electrode surface of 20 cm2 and an inter-
electrode gap of 6 mm. The BDD electrodes were contacted on their
whole area of the electrode from the back side of the cell in order to
avoid inhomogeneous current distributions. A thin cation ex-
change membrane was placed between cathode and anode to
exclude the reaction products from the cathode chamber.
Therefore, a very thin membrane was manufactured by infiltrating
an ethanolic Nafion solution (DuPont) into a hydrophilic Gore PTFE
separator as a kind of framework with an average pore size of about
2ꢀ
ꢀ
electrodes. The sulfate anions SO4
and HSO4 are oxidized
directly on the platinum surface to sulfate radicals SO4ꢀ, which
recombine to H2S2O8 [18,19]. At the BDD electrode, hoꢀwever, the
hydroxyl radical reacts with H2SO4 (Eq. (2)) or HSO4 (Eq. (3))
to sulfate radicals:
ꢀꢁ
H2SO4 + OHꢁ ! SO4 + H3O+
(2)
ꢀꢁ
HSO4ꢀ + OHꢁ ! SO4 + H2O
(3)
5 mm. This procedure is reported elsewhere [25] and resulted in a
Serrano et al. proposed the above mentioned mechanism and
reportedontheelectrochemicalpreparationofperoxodisulfuricacid
using boron-doped diamond thin film electrodes and sulfuric acid
with concentrations below 5 mol lꢀ1, by observing current efficien-
cies of nearly 95% [17,20]. Recently, Kurita Water Industries
published a studyaboutthe resist removalpropertiesofelectrolyzed
sulfuric acid [21]. Both, Zwicker et al. [22] and Hayamizu et al. [23]
patented a process using boron-doped diamond electrodes and
sulfuric acid as a cleaning system, while Comninellis et al. patented
the electrochemical production of peroxodisulfates itself using BDD
electrodes [24]. However, up to now, the electrolysis in especially
highly concentrated sulfuric acid (ꢂ85 wt%) per se is not further
described in the literature.
The aim of this work is to prepare a solution of peroxodisulfuric
acid in sulfuric acid equivalentto a commonSPM resist removal bath
byelectrolysisonboron-dopeddiamondelectrodes.Therefore,cyclic
voltammetry was performed in highly concentrated H2SO4 of more
than 85 wt%, which should help to understand the formation and
consumption reactions of peroxodisulfuric acid at boron-doped
diamond electrodes. Moreover, the formation of H2S2O8 in an
electrochemicalflowcellundergalvanostaticconditionsisstudiedin
order to investigate the influence of consumption reactions on the
formation rate of peroxodisulfuric acid under real working
conditions. These experiments play an important role for the design
of a test setup to fulfill the special requirements of this process.
very thin, but mechanically stable cation exchange membrane,
which was used to separate the cathode chamber from the anode
chamber. The separated electrolyte cell was placed in an
unconventional setup as it will be described in the following
(Fig. 1). Instead of using two half circuits for the anolyte and
catholyte, only the anode chamber is flushed with fresh electrolyte.
The total electrolyte volume measured 1 liter for all experiments
with the flow cell. Electrolysis products are removed from the cell
with the electrolyte flow and stored in PP-bottles, in which they
were mixed with the residual electrolyte, thermally equilibrated by
a temperature control device, and pumped into the electrolyte cell
back again by PTFE membrane pumps as fast as possible assuming
homogeneous mixing of the electrolyte and equilibrium between
the electrolysis cell and electrolyte reservoir. The only convection
at the cathode surface is caused by gaseous electrolysis products
(Fig. 1).
The concentration of electrolysis products were determined
offline by titration. H2O2 was titrated using KMnO4. The sum of
peroxosulfuric species was back titrated with FeSO4 and KMnO4.
The electrolyte was titrated iodometrically with KI and NaS2O3 to
determine the peroxomonosulfuric acid concentration [26,27].
Cyclic voltammetry measurements were carried out using a
double walled three electrode glass cell (Radiometer Analytical). A
saturated calomel electrode was used as a reference electrode,
boron-doped diamond as the working electrode and a platinum
plate (Alfa Aesar, 99.9%) as the counter electrode. The data was
collected using a potentiostat from Iviumstat. The electrolyte was
degassed under argon flow for 30 min before the measurement
was started.
3. EXPERIMENTAL
Sulfuric acid was electrolyzed in a symmetric electrolytic flow
cell. Both, cathode and anode were boron-doped diamond
Fig.1. Schematic description of the modified electrolyte cell with two half cells separated by a manufactured Nafion membrane (left) and technical drawing of the setup used
for electrolysis of H2SO4 on BDD (right); C1 electrolytic flow cell, P1 PTFE membrane pump, HE1 PFA heat exchanger, V1 valve, T1 electrolyte reservoir.

F. Hippauf et al. / Electrochimica Acta 147 (2014) 589–595
591
Small-angle X-ray scattering data were collected using a
NANOSTAR instrument from Bruker with Cu-K radiation. The
a
1
scattering was detected from 0 to 3 ꢃ2
and located within a copper sample holder for measurement.
Atomic force microscopy was performed on Dimension
u. Samples were activated
a
D3100 from Veeco Digital Instruments and the surface roughness
was determined with a RS-232 portable surface measurement
device from ATP-Messtechnik. X-ray photoelectron spectroscopy
measurements were carried out using a SPECS Phoibos 100 with
Mg-K radiation and an acceleration voltage of 12.5 kV.
a1
4. RESULTS AND DISCUSSION
The BDD electrode is characterized regarding to its morphology,
chemical composition and electrochemistry in order to get a better
understandingof the peroxodisulfuricacid formation and consump-
tionatboron-dopeddiamondelectrodes. Moreover, theformationof
H2S2O8inanelectrochemicalflowcellundergalvanostaticconditions
is studied under real working conditions. These experiments
influenced directly the specific tailoring of the experimental flow
cell equipment.
Fig. 3. Cyclic voltammogram of boron-doped diamond electrodes in 1 M H2SO4 at
different scan rates, T = 21 ꢃC, reference electrode SCE, counter electrode platinum,
lower magnification in the inset.
wide potential range between -1.0 V and 2.0 V vs. SCE in which
aqueous electrolytes remain stable (Fig. 3). Moreover, the BDD
electrode indicates a low background current in the double layer
domain. To investigate the electrochemical behavior in terms of
the generation of peroxodisulfuric acid, cyclic voltammograms
with different amounts of sulfuric acid and peroxodisulfuric acid
were measured. In pure 1 M H2SO4, the measured reversible redox
couple at 1.5 V vs. SCE is attributed to semi-hydroquinone/semi-
benzoquinone functional groups, which were proposed to be
present on the surface of the BDD electrode by Kapałka et al. (Fig. 3)
[14]. This observation is consistent with graphitic and amorphous
carbon regions at the electrode's surface (Fig. S1). Comninellis and
co-workers assigned the cathodic peak at -1.0 V (inset) vs. SCE to
the reduction of H2S2O8 (Eq. (4))[20], however, the reduction peak
at 0.1 V vs. SCE remains still unclear.
4.1. Electrode characterization
The morphology of the boron-doped diamond electrodes was
investigated by atomic force microscopy. The sample exhibits
randomly ordered diamond crystallites with an average size of
3
m
m and round aggregates indicating amorphous regions (Fig. 2)
beingrevealedbyRamanspectroscopy(seesupportinginformation).
The electrode surface shows a Ra and Rz roughness of 4 and 23.5 m,
m
respectively, measured by a portable surface measurement device
(see supporting information). X-ray photoelectron spectroscopy of
the electrode determined a boron concentration of 8700 ppm.
4.2. Electrochemical measurements
H2S2O8 + 2H+ + 2eꢀ ! 2H2SO4
(4)
Due to their high overpotential for oxygen- and hydrogen
generation, as well as their weak interactions with redox couples,
the electrochemical properties of BDD electrodes are unique and
favorable. The cyclic voltammogram of a BDD electrode shows a
Since it is known that BDD electrodes show no significant
adsorption peaks for hydrogen [28], the reduction peak at 0.1 V vs.
SCE may originate from another redox couple on the surface of the
diamond, as its intensity increases with higher scan rates and is
suppressed in highly concentrated H2SO4.
For the application as a resist removal bath in the semiconductor
Fig. 4. Cyclic voltammogram of boron-doped diamond electrodes in H2SO4 (90 wt
%), before and after electrolysis, and after addition of K2S2O8, scan rate = 50 mV sꢀ1
,
Fig. 2. Atomic force microscopy of the BDD electrode surface scanned in tapping
T = 21 ꢃC, reference electrode SCE, counter electrode platinum, lower magnification
mode with a resonance frequency of 150 kHz.
in the inset.

592
F. Hippauf et al. / Electrochimica Acta 147 (2014) 589–595
industry or for the production of peroxodisulfates, the electrolyte
concentration of 1 M H2SO4 is too low. Thus, we investigated the
electrochemical behavior of BDD in 90 wt% H2SO4 before and after
electrolysis. Under these conditions the semi-hydroquinone/semi-
benzoquinone redox couple is fully suppressed (Fig. 4). After
400 min of electrolysis, a peroxodisulfuric acid concentration of
0.16 mol lꢀ1 was determined by titration. In addition to the
cathodic reduction peak at 0 V vs. SCE, the cyclic voltamogramm
shows another peak of an anodic, irreversible reaction at 1.1 V vs.
SCE. To elucidate the origin of the latter peak, an equal amount of
K2S2O8 was added to 90 wt% H2SO4. The resulting cyclic
voltammogram shows the same cathodic reduction peak at 1.1 V
vs. SCE and is therefore attributed to an oxidation reaction of either
peroxomonosulfuric acid or peroxodisulfuric acid. It is known that
peroxodisulfate undergoes an acid catalyzed hydrolysis forming
sulfuric acid and peroxomonosulfuric acid (Eq. (5))[6]:
During the electrolysis the concentration of H2O2 and H2S2O8
(representative for the sum of H2SO5 and H2S2O8, determined by
titration against Fe2+, even though the predominant peroxosulfate
species in solution is H2SO5) was constantly monitored to verify
how ongoing reactions interact. In the early stage of the
electrolysis, the concentration of the peroxodisulfuric acid
increases rapidly and the current efficiency reaches 40% (see
supporting information). A steady state is achieved at 21 ꢃC after
1000 min and the current efficiency decreases to 0% until the
peroxodisulfuric acid reaches a plateau of 0.13 mol lꢀ1 (Fig. 5). The
current is completely extinguished by side reactions, due to
complex anodic and cathodic consumption reactions (vide supra)
(Fig. 6). One byproduct of the electrolysis is hydrogen peroxide,
which was detected by titration. A plateau concentration of 0.01 M
was observed after 600 min at 21 ꢃC and was formed either by the
recombination of two hydroxyl radicals (Eq. (7)) or by the
hydrolysis of Caro's acid (Eq. (8)):
H2S2O8 + H2O ! H2SO5 + H2SO4
(5)
2OHꢁ ! H2O2
(7)
To gain further information, we performed an Iodometric
titration, which confirms the hydrolysis of peroxodisulfate to
peroxomonosulfate in 90 wt% H2SO4 during the electrolysis, even
at ambient temperature (Table 1). Additionally, peroxodisulfate
remains stable in 1 M H2SO4, due to the absence of this peak at 1.1 V
vs. SCE in diluted H2SO4 (Fig. 3). Therefore, we conclude that this
behavior can be attributed to the oxidation of peroxomonosulfate
(Eq. (6)):
H2SO5 + H2O ! H2SO4 + H2O2
(8)
Hydrogenperoxidedecomposesthermallyandelectrochemically
[31]. Another side-reaction is the generation of oxygen, which was
detected by hydrogen-oxygen reaction, and the formation of ozone
[29]. These competitive reactions of the hydroxyl radical limit the
currentefficiency inthe earlystage of the electrolysis, evenwhenthe
product concentration is still low. Iodometric titration revealed that
the majorperoxosulfuricspeciesisCaro's acid evenat10 ꢃC (Table 1).
This shows that H2S2O8, being the primary reaction product, forms
immediately H2SO5 and is consumed at both cathode and anode,
following reactions (4) and (6). The consumption rate increases with
theconcentrationofH2SO5andlimitsthecurrentefficiencyinthelater
stageoftheelectrolysiswhentheplateauconcentrationisreached.In
previous works, the side reactions have not been taken into account,
becausetheelectrolyteconcentrationshavebeenmuchlowerandthe
final product concentration only reached levels that do not cause
significantelectrochemicaldecomposition[17,20,32].Recently,Davis
etal.reportedtheformationofaplateauconcentrationinacirculated
electrolyte bath at even lower sulfuric acid concentrations, which
stands in good agreement with our results [33].
H2SO5 + H2O ! H2SO4 + O2 + 2H+ + 2eꢀ
(6)
This oxidative consumption was only described for platinum
electrodes [28] and implies a distinctive interaction between the
electrolyte and the electrode [29]. The observed oxidation
potential of 1.1 V vs. SCE is by 0.711 V higher than the standard
potential J. Balej gave (0.389 V vs. SCE). This over voltage is
attributed to a high electrolyte concentration (90 wt%) and weak
interactions between boron doped diamond and the electrolyte
[30]. The anodic and cathodic consumption reactions of perox-
odisulfuric acid, as well as the side reactions of the hydroxyl
radicals affect the process described below and considerably
reduce the current efficiency.
4.4. Electrochemical generation of Caro's acid by oxidation of sulphuric
acid
4.4.1. Influence of the working temperature on the process
No significant differences between 10 ꢃC, 21 ꢃC, and 55 ꢃC
working temperature are observed at the early stage of the
Due to the reduced current efficiency the described consump-
tion reactions (4), (5), and (6) will limit the yield of the product.
Sulfuric acid was oxidized in an electrochemical flow cell and
different parameters were adjusted to elucidate the influence of
the side reactions on the current efficiency under real working
conditions. For this purpose, the temperature of the electrolyte, the
current density, and the electrolyte concentration have been
systematically altered and the results were investigated by
titration of the reaction products. The electrolyte circuit and the
flow cell were used as described in chapter 2, except that Valve
V1 was opened during the electrolysis and the membrane was
removed. Hence, anodic and cathodic reactions can take place.
Table 1
H2SO5 and H2S2O8 sum and H2SO5 single concentration in 90 wt% H2SO4 at
different working temperatures and different time determined by iodometric
titration and back titration with FeSO4 and KMnO4 respectively.
T/ꢃC
cH2SO5+H2S2O8/mol lꢀ1
cH2SO5/mol lꢀ1
10
21
55
0.055
0.135
0.064
0.056
0.144
0.064
Fig. 5. Concentrations of H2S2O8 and H2O2 vs. time in sulfuric acid (90 wt%) using
three different working temperatures, current density = 0.15 A cmꢀ2
.

F. Hippauf et al. / Electrochimica Acta 147 (2014) 589–595
593
Fig. 6. Possible reaction paths of hydroxyl radicals after desorption, including anodic and cathodic consumption reactions.
electrolysis (<150 min), whereas the plateau concentꢀra1tion
reaches only 0.07 mol lꢀ1 for 55 ꢃC instead of 0.14 mol l
for
limited over a wide range of parameters [33]. As a result of higher
peroxodisulfuric acid concentrations, the hydrogen peroxide
concentration increases as well. This supports the formation of
H2O2 via Reaction (5) and (8) rather than via reaction (7).
10 ꢃC (Fig. 5). Thus, the anodic and cathodic consumption reactions
are influenced much more by increased temperatures than the
competitive side reactions. The thermal decay of H2SO5 can be
neglected at this temperature values, because the product
concentrations remained stable for 2 h at 55 ꢃC. The setup had
to be cooled down to 21 ꢃC for further experiments, as the process
generates heat and the electrolyte's temperature rises without a
heat exchanger cooling down the electrolyte.
4.4.3. Influence of electrolyte concentration on the process
This parameter is limited to concentration restrictions of a
resist removal bath and was varied from 95 wt% to 90 wt% and
85 wt%, respectively (Fig. 8). Serrano et al. reported an increasing
formation rate of peroxodisulfuric acid with rising electrolyte
concentration up to 40 wt% H2SO4 since hydroxyl radicals only
react with molecular sulfuric acid or hydrogen sulfate and their
amount rises constantly with higher mass fractions of H2SO4
[20,34]. In contrast, the formation rate of H2S2O8 and the plateau
concentration decrease with increasing concentrations from 85 wt
% to 90 wt%. An explanation for this observation might be that in a
solution of more than 76 wt% of pure H2SO4 in H2O practically
contains no free water, which is not solvated to a proton and the
equilibrium (Eq. (9)) is strongly shifted to the right side.
4.4.2. Influence of current density on the process
The current density was examined in the range from 0.05 to
0.5 A cmꢀ2 (Fig. 7). An increasing current density leads to a higher
concentration of hydroxyl radicals causing a faster increase of the
peroxodisulfuric acid concentration. The formation rate of
peroxodisulfuric acid increases linearly with the current density
and the plateau concentration is already reached at 600 min for
0.5 A cmꢀ2, (note that other settings still generate peroxosulfate
after 600 min). Hence, the formation and the consumption
reactions of peroxodisulfuric acid are accelerated to the same
degree by increasing current densities and no diffusion processes
or slow reaction steps seem to limit the process in the observed
current density range. The influence of the current density is less
pronounced (Fig. S4). Davis et al. reported a similar behavior for
current densities from 0.1 to 0.3 A cmꢀ2 and more diluted
electrolyte concentrations showing that this process runs current
5H2SO4 + H2O $ HSO4ꢀ + H3 O+
(9)
Water is essential for the first step of the peroxodisulfate
formation comprising the water discharge and the hydroxyl radical
formation. Consequently, the hydrogen peroxide and peroxodi-
sulfate concentration decreases and electrochemical consumption
reactions must increase to consume the electrical charge at the
electrodes under galvanostatic conditions. Lower overvoltages and
Fig. 7. Concentrations of H2S2O8 species and H2O2 vs. time in sulfuric acid (90 wt%),
Fig. 8. Concentrations of H2S2O8 and H2O2 vs. time at constant temperature (21 ꢃC),
constant temperature (21 ꢃC) using six different current densities.
constant current (0.15 A cmꢀ2) using three different H2SO4 concentrations.

594
F. Hippauf et al. / Electrochimica Acta 147 (2014) 589–595
current densities minimize this effect. It is not clear which
consumption reaction is more crucial regarding to the current
efficiency. At these concentrations, no free water is present in
solution and H2SO4 molecules are not fully hydrated and may even
form polysulfuric acids. Hence, there are many details involved in
this electrolysis, which are not fully understood at this point.
peroxodisulfate depleted condition (after it was used in a resist
removal bath), less peroxodisulfuric acid will face the cathode and
the cathodic reduction of H2S2O8 (Eq. (4)) will be reduced. If the
electrolyte enters the cathode chamber after recirculation con-
taining considerable amounts of H2S2O8, reaction (4) will take
place.
In our work, the electrolyte is pumped exclusively through the
anode chamber. The cathode chamber is filled with electrolyte
and closed by a valve prior to the electrolysis start. Only one
electrolyte loop is necessary, which is highly desirable in an
industrial process. The proton gradient, which is formed during
electrolysis, can only be destroyed by convection of the rising
hydrogen bubbles and diffusion (Fig. 1). In contrast to the setup of
Kobayashi et al. using a diaphragm and a series connection of
anode and cathode chamber, peroxodisulfuric acid is excluded
from the cathode chamber by the membrane in order to avoid the
reductive consumption (Eq. (4)). After the electrolysis was
started, no electrolyte leaching was observed from the cathode
chamber demonstrating that the membrane works successfully as
a pneumatic barrier. A maximum current density of 0.25 A cmꢀ2
was reached and no loss of membrane material was detected after
an electrolysis period of 3.9 Ah cmꢀ2. No significant benefits were
noticed until the peroxodisulfuric acid concentration reaches
0.08 mol lꢀ1 (Fig. 10). After this point, the new Nafion/PTFE hybrid
membrane setup produces more peroxodisulfuric acid compared
to the setup without membrane. The membrane is expected to
eliminate only the cathodic consumption reaction. Hence, the
competitive reactions, which are dominating the early stage of
the electrolysis, are not suppressed and anodic consumption is
still present leading into a plateau of 0.16 mol lꢀ1 H2S2O8.
Nevertheless, the plateau concentration could be increased by
0.02 mol lꢀ1 due to the implementation of the developed
membrane. This is the first time, a mechanically stable Nafion
membrane was implemented in electrolysis process for the
generation of peroxodisulfuric acid in highly concentrated
sulfuric acid. Furthermore, a membrane is essential for industrial
use, because gaseous byproducts, namely oxygen and hydrogen,
can form explosive mixtures.
5.1. Development of an adjusted flow cell design
The previous experiments revealed that both competitive
reactions and consumption reactions limit the current efficiency.
The current density does not show any significant influence on these
mentionedreactions. A promisingapproach concerning the efficien-
cyenhancementistoremovetheanodicreactionproductsfrequently,
thus to prevent cathode contact as well as critical concentrations.
Usually, an electrolyte cell is separated by an ion exchange
membrane to isolate the cathode from the anode chamber.
Commercially available membranes such as Nafion failed under
these conditions. Their ion conductivity is too low to ensure the
required current densities. As a result, the electrical current
collapsed after a few seconds. Verbrugge et al. showed that the
ion conductivity of perfluorosulfonic acid membranes decreases
with increasing sulfuric acid concentrations due to the decreasing
water content of the membrane [35]. Small angle X-ray scattering
showed that the ionomer peak still exists after treatment with 90 wt
% H2SO4, but it is shifted to higher q-values (Fig. 9). A higher q-value
indicatesasmalleraveragedistanceLbetweenthesolventclustersin
the membrane (Eq. (9)), which is attributed to the partial
dehydration of the membrane and therefore, decreasing ion
conductivity [36–38]:
L = 2
p
/q
(10)
A thinner membrane gives better ion conductivities, but
thinner membranes usually suffer from the mechanical stability
being required in a flow cell. Therefore, a commercial, hydrophilic,
and porous PTFE separator was impregnated with an ethanolic
Nafion solution according to Nouel and Fedkiw [25]. This thin, but
mechanically stable cation exchange membrane was implemented
in a special setup that was designed to meet the requirements of
this process. The membrane divides the electrolyte cell in two
This modified setup was used to prepare 0.1 molperoxodisulfuric
acid in 90 wt% sulfuric acid (5.9 V, 3 A), which is equivalent to a
reference SPM-bath (50:1H2SO4:H2O2, 100 ꢃC), with an average
current efficiency of 18.8% and a specific energy consumption of
8.66 kWh per kg H2S2O8. This concentration stays constant by
further electrolysis without changing or diluting the bath.
compartments. Usually, such
a cell requires two different
electrolyte loops. Recently, Kobayashi et al. patented a setup in
which the electrolyte circulates successively in the cathotic
chamber before it enters the anodic chamber. Anodic and cathodic
compartments are separated only by a diaphragm [39]. If the
electrolyte enters the anodic chamber in a consumed and
Fig. 10. Concentrations of H2S2O8 and H2O2 vs. time in sulfuric acid (90 wt%),
current density = 0.15 A cmꢀ2 for electrolysis cell with and without Nafion
membrane.
Fig. 9. Small angle X-ray scattering curves of Nafion NE 450 membranes after
different activation procedures at room temperature.

F. Hippauf et al. / Electrochimica Acta 147 (2014) 589–595
595
6. CONCLUSIONS
[12] T.A. Enache, A.-M. Chiorcea-Paquim, O. Fatibello-Filho, A.M. Oliveira-Brett,
Hydroxyl radicals electrochemically generated in situ on a boron-doped
diamond electrode, Electrochem. Commun. 11 (2009) 1342–1345, doi:http://
dx.doi.org/10.1016/j.elecom.2009.04.017.
[13] N. Vatistas, Electrocatalytic Properties of BDD Anodes: Its Loosely Adsorbed
Hydroxyl Radicals, Int. J. Electrochem. 2012 (2012) 1–7, doi:http://dx.doi.org/
10.1155/2012/507516.
[14] A. Kapałka, G. Fóti, C. Comninellis, Investigations of electrochemical oxygen
transfer reaction on boron-doped diamond electrodes, Electrochim. Acta.
53 (2007) 1954–1961, doi:http://dx.doi.org/10.1016/j.electacta.2007.08.066.
[15] M. Panizza, I. Duo, P.A. Michaud, G. Cerisola, C. Comnellis, Electrochemical
Generation of Silver (II) at Boron-Doped Diamond Electrodes, Electrochem.
Solid-State Lett. 3 (2000) 550–551, doi:http://dx.doi.org/10.1149/1.1391205.
[16] J. Lee, D.A. Tryk, A. Fujishima, S.-M. Park, Electrochemical generation of ferrate
in acidic media at boron-doped diamond electrodes, Chem. Commun. 48
(2002) 486–487, doi:http://dx.doi.org/10.1039/B111317H.
In this work, the electrolysis of highly concentrated sulfuric acid
to peroxodisulfuric acid by BDD electrodes was investigated. The
morphology of the BDD electrode was characterized and the
electrochemical behavior of BDD in 90 wt% was analyzed by cyclic
voltammetry revealing that both anode and cathode consume the
electrolysis product. Electric current density only has a marginal
influence on the current efficiency in the electrolytic flow cell,
whereas the high electrolyte concentration initiates hydrolysis and
consumption reactions that are responsible for the lower current
efficiency [20]. SAXS measurements confirmed that conventional
membranes exhibit too low ion conductivities and are thus not
suitable for our requirements. Therefore, a porous PTFE-separator
was impregnated with a Nafion solution resulting in a thinner but
mechanically stableion exchange membrane that was implemented
in a specifically designed setup for the electrooxidation of sulfuric
acid. This approach successfully hindered the peroxosulfuric acid
reduction at the cathode and improved the performance of the
process. Finally, a peroxodisulfate concentration was achieved as a
steady state that is equivalent to a common SPM resist removal bath
with a current efficiency of 18.8% without dilution.
[17] P. Michaud, E. Mahé, W. Haenni, A. Perret, C. Comninellis, Preparation of
Peroxodisulfuric Acid Using Boron-Doped Diamond Thin Film Electrodes,
Electrochem. Solid-State Lett.
1.1390963.
3 (2000) 77, doi:http://dx.doi.org/10.1149/
[18] M.W. Breiter, Anodische Bildung von Perverbindungen am Platin, Chem.
Ing. Tech. 35 (1963) 376–378, doi:http://dx.doi.org/10.1002/cite.330350512.
[19] K. Nagel, H. Dietz, Spannungs-aktivitäts-diagramm der wichtigsten elektro-
denreaktionen im System {Platin/Platinoxid/Lösung.}, Electrochim. Acta
(1961) 141–154, doi:http://dx.doi.org/10.1016/0013-4686(61) 80013-3.
[20] K. Serrano, P.A. Michaud, C. Comninellis, A. Savall, Electrochemical preparation
of peroxodisulfuric acid using boron doped diamond thin film electrodes,
Electrochim. Acta 48 (2002) 436–688, doi:http://dx.doi.org/10.1016/S0013-
4686(02)-6.
[21] T. Nagai, H. Yamakawa, M. Uchida, T. Ohtsu, N. Ikemiya, H. Morita, Study on
Resist Removal Using Electrolysed Sulfuric Acid Solution in Comparison with
SPM, Solid State Phenom. 187 (2012) 109–112.
ACKNOWLEDGEMENTS
[22] G. Zwicker, M. Fryda, T. Matthee, Method for the oxidative treatment of
components comprised of or containing elementary silicon and/or substan-
tially inorganic silicon compounds, US 2005/0139487 A1, 2005
[23] N. Hayamizu, Y. Shibata, M. Kato, H. Fukui, Cleaning System and Cleaning
Method, US 2008/0110766 A1, 2008
[24] C. Comninellis, P. Michaud, W. Hänni, A. Perret, M. Fryda, Electrochemical
Production of peroxopyrosulphuric acid using diamond coated electrodes, US
6855,242 B1, 2005
The author wants to thank Condias GmbH, especially Dr.
Thorsten Mathhée, for supplying the boron-doped diamond
Diachem electrodes and Infineon Technologies Dresden GmbH
for technical and financial support. Further thanks are due to Tim
Biemelt for SAXS measurements, Andreas Maier for AFM images,
Dr. rer. nat. Steffen Oswald for XPS measurements, and
Kai Schwedtmann for extensive proof reading
[25] K.M. Nouel, P.S. Fedkiw, Nafion1-based composite polymer electrolyte
membranes, Electrochim. Acta. 43 (1998) 2381–2387, doi:http://dx.doi.org/
10.1016/S0013-4686(97)10151-7.
[26] Y.K. Gupta, Iodometric Determination Of Persulphate In Sulphuric Acid
Solution, Fresenius J. Anal. Chem. (1961) BF00463208 180, 260-265, 10 .1007/.
[27] A. Baeyer, V. Villiger, Ueber die Sulfomonopersäure (Caro'sche Säure), Eur. J.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.elec-
tacta.2014.09.133.
Inorg.
Chem.
34
(1901)
853–862,
doi:http://dx.doi.org/10.1002/
cber.190103401148.
[28] A. Kraft, G. Gmbh, K. Str, Doped Diamond: A Compact Review on a New
Versatile, Electrode Material 2 (2007) 355–385.
[29] A. Schmidt, Angewandte Elektrochemie: Grundlagen d. elektrolyt Produk-
tionsverfahren, Verlag Chemie, 1976.
[30] J. Balej, Thermodynamics Of Reactions During The Electrolysis Of Peroxodi-
sulphates, Electrochim. Acta. 29 (1984) 1239–1242.
References
[1] H. Caro, Hauptversammlung des Vereins deutscher Chemiker am 1. bis 4. Juni
1898 zu Darmstadt, Angew. Chem. 11 (1898) 845–846, doi:http://dx.doi.org/
10.1002/ange.18980113602.
[2] C.H. Hamann, W. Vielstich, Elektrochemie, WILEY-VCH, 2005.
[3] Oloman, A.P. Watkinson, The electroreduction of oxygen to hydrogen peroxide
on fixed bed cathodes, Can. J. Chem. Eng. 54 (1976) 312–318, doi:http://dx.doi.
org/10.1002/cjce.5450540412.
[4] J.H. Boyer, C. Huang, Oxidation of nitrobenzofuroxans, J. Chem. Soc. Chem.
Commun. 36 (1981) 365–366, doi:http://dx.doi.org/10.1039/c39810000365.
[5] H. Palme, Studien über die Zersetzung der Überschwefelsäure, Z. Anorg. Allg.
Chem. 112 (1920) 97–130, doi:http://dx.doi.org/10.1002/zaac.19201120105.
[6] I.M. Kolthoff, I.K. Miller, The Chemistry Of Persulfate 1. The Kinetics And
Mechanism Of The Decomposition Of The Persulfate Ion In Aqueous Medium, J.
Am. Chem. Soc. 73 (1951) 3055–3059, doi:http://dx.doi.org/10.1021/
ja01151a024.
[7] N. Katsuki, Water Electrolysis Using Diamond Thin-Film Electrodes, J.
Electrochem. Soc. 145 (1998) 2358, doi:http://dx.doi.org/10.1149/1.1838643.
[8] W. Haenni, H. Baumann, C. Comninellis, D. Gandini, P. Niedermann, A. Perret,
et al., Diamond-sensing microdevices for environmental control and analytical
applications, Diam. Relat. Mater. 7 (1998) 569–574, doi:http://dx.doi.org/
10.1016/S0925-9635(97).
[9] X. Chen, G. Chen, Proper Hot Filament CVD Conditions for Fabrication of Ti-
Boron Doped Diamond Electrodes, J. Electrochem. Soc. 151 (2004) B214, doi:
http://dx.doi.org/10.1149/1.1651529.
[10] S. Alehashem, F. Chambers, J.W. Strojek, G.M. Swain, R. Ramesham, Cyclic
Voltammetric Studies of Charge Transfer Reactions at Highly Boron-Doped
Polycrystalline Diamond Thin-Film Electrodes, Anal. Chem. 67 (1995)
2812–2821, doi:http://dx.doi.org/10.1021/ac00113a014.
[11] M. Panizza, G. Cerisola, Application of diamond electrodes to electrochemical
processes, Electrochim. Acta. 51 (2005) 191–199, doi:http://dx.doi.org/
10.1016/j.electacta.2005.04.023.
[31] B. Marselli, J. Garcia-Gomez, P.-A. Michaud, M.A. Rodrigo, C. Comninellis,
Electrogeneration of Hydroxyl Radicals on Boron-Doped Diamond Electrodes, J.
Electrochem.Soc.150(2003)D79–D83,doi:http://dx.doi.org/10.1149/1.1553790.
[32] D. Khamis, E. Mahé, F. Dardoize, D. Devilliers, Peroxodisulfate generation on
boron-doped diamond microelectrodes array and detection by scanning
electrochemical microscopy, J. Appl. Electrochem. 40 (2010) 1829–1838, doi:
http://dx.doi.org/10.1007/s10800-010-0114-x.
[33] J.R. Davis, J.C. Baygents, J. Farrell, Effect of current density and sulfuric acid
concentration on persulfuric acid generation by boron-doped diamond film
anodes, J. Appl. Electrochem. 44 (2014) 841–848, doi:http://dx.doi.org/
10.1007/s10800-014-0692-0.
[34] N.B. Librovich, V.D. Maiorov, Ionic-molecular composition of aqueous sulfuric
acid solutions at 25, B. Acad. Sci. USSR Ch. 26 (1977) 621–623, doi:http://dx.
doi.org/10.1007/BF01179485.
[35] M.W. Verbrugge, R.F. Hill, Experimental and theoretical investigation of
perfluorosulfonic acid membranes equilibrated with aqueous sulfuric acid
solutions, J. Phys. Chem. 92 (1988) 6778–6783, doi:http://dx.doi.org/10.1021/
j100334a056.
[36] W.Y. Hsu, T.D. Gierke, Ion transport and clustering in nafion perfluorinated
membranes, J. Membr. Sci. 13 (1983) 307–326, doi:http://dx.doi.org/10.1016/
S0376-7388(00)81563-X.
[37] C. Heitner-Wirguin, Recent advances in perfluorinated ionomer membranes:
structure, properties and applications, J. Membr. Sci. 120 (1996) 1–33, doi:
http://dx.doi.org/10.1016/0376-7388(96)155-X.
[38] G. Gebel, J. Lambard, Small-Angle Scattering Study of Water-Swollen
Perfluorinated Ionomer Membranes, Macromolecules 30 (1997) 7914–7920,
doi:http://dx.doi.org/10.1021/ma970801v.
[39] N. Kobayashi, Y. Shibata, N. Hayamizu, M. Kato, Apparatus for Electrolyzing
Sulfuric Acid, Method of performing electrolysis, and apparatus for Processing
a Substrate, US 8236,161 B2, 2012