Microstructured reactor for electroorganic synthesis

Article abstract of DOI:10.1016/j.electacta.2010.05.061

In the present work a brief overview of microstructured devices, the advantages and disadvantages as well as the principles of a multiscale design approach are presented. The advantages mainly comprise uniform current density distribution, local control of the process parameters, high single-pass conversion of the reactant and reduced concentration of the supporting electrolyte needed to perform the reaction. The main challenge with this type of cell is gas evolution, a typical reaction taking place at the counter-electrode. The phenomena discussed are documented using the example of anodic methoxylation of 4-methylanisole. An analysis was made of the two-phase flow hydrodynamics inside the microstructured cell. The range of operational conditions suitable for the desired reaction was identified. The results were used together with the information on the electrode reaction kinetics in the form of the Butler-Volmer-type equation for the design of a suitable bipolar microstructured cell. A bipolar cell optimized to perform the desired reaction is reported. The results are compared with the published data. An analysis was performed that proved that the performance of the existing technology is more demanding in terms of energy consumption for the separation of the final product from the reaction mixture. The process intensification was evaluated on the basis of the available data.

Full text of DOI:10.1016/j.electacta.2010.05.061

Electrochimica Acta 55 (2010) 8172–8181
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Microstructured reactor for electroorganic synthesis
Karel Bouzeka,∗,1, Vladimír Jirˇicˇny´ b, Roman Kody´ma, Jirˇí Krˇisˇt’álb, Tomásˇ Bystronˇ a
a Department of Inorganic Technology, Institute of Chemical Technology Prague, Technická 5, 166 28 Prague 6, Czech Republic
b Institute of Chemical Process Fundamentals, v.v.i., Academy of Sciences of the Czech Republic, Rozvojová 2, 165 02 Prague 6, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history:
In the present work a brief overview of microstructured devices, the advantages and disadvantages as
well as the principles of a multiscale design approach are presented. The advantages mainly comprise
uniform current density distribution, local control of the process parameters, high single-pass conversion
of the reactant and reduced concentration of the supporting electrolyte needed to perform the reaction.
The main challenge with this type of cell is gas evolution, a typical reaction taking place at the counter-
electrode. The phenomena discussed are documented using the example of anodic methoxylation of
4-methylanisole. An analysis was made of the two-phase flow hydrodynamics inside the microstructured
cell. The range of operational conditions suitable for the desired reaction was identified. The results were
used together with the information on the electrode reaction kinetics in the form of the Butler–Volmer-
type equation for the design of a suitable bipolar microstructured cell. A bipolar cell optimized to perform
the desired reaction is reported. The results are compared with the published data. An analysis was
performed that proved that the performance of the existing technology is more demanding in terms
of energy consumption for the separation of the final product from the reaction mixture. The process
intensification was evaluated on the basis of the available data.
Received 29 October 2009
Received in revised form 10 May 2010
Accepted 14 May 2010
Available online 24 May 2010
Keywords:
Microstructured reactors
Bipolar arrangement
Electroorganic synthesis
Gas evolution
Process intensification
© 2010 Elsevier Ltd. All rights reserved.
1. Introduction
ically viable, microstructured unit production technologies have
emerged and many of them have become routine. This is reflected
Microstructuring represents a novel approach to the design of
chemical apparatuses. For the last two decades this subject has been
intensively studied by numerous chemical engineering laborato-
ries around the world. In the case of electrochemical engineering
and technology, this endeavor has clearly lagged behind. This is
surprising since electrochemical engineering discovered some of
the advantages of microstructured design several decades ago
and implemented this type of unit in industrial practice. A typical
representative is the capillary-gap (CG) cell used in electroorganic
synthesis technology [1].
The discovery of the CG cell was largely intuitive due to the
basic characteristic of the systems it is used for, i.e. high ohmic
resistance. It is typically an important aspect in the case of elec-
troorganic synthesis because of the low ionic conductivity of the
electrolyte solutions used. At the time, the main obstacle to a
further reduction of the interelectrode distance was the limited
availability of suitable construction materials and the high produc-
tion price of this type of unit. Due to progress made in the fields
of material science and machining technologies during the last
decade, the situation has changed significantly. Recently econom-
by the fact that several companies now offer a number of standard-
ized microstructured units commercially [2]. The main deficit is an
inadequate fundamental understanding of the behavior of this type
of system. This makes it exceedingly difficult to reach a decision
on the suitability of this approach on an industrial scale and subse-
quently on the design and optimization of relevant electrochemical
cells.
This lack of information was already identified in the ꢀ90s. At the
time Belmont and Girault [3–8] published a series of papers dealing
with the utilization of interdigitated band electrodes for electroor-
ganic synthesis. The approach they used overcame the construction
problem of obtaining a cell with the electrodes fixed at a sub-
millimeter distance by depositing them at regular distances on a
suitable planar substrate. In the first part of the investigation, a the-
oretical analysis showed that a reduction of the cell voltage could be
obtained by this approach [3]. In the next step, a theoretical descrip-
tion was verified by furan methoxylation as a model reaction [4],
followed by propylene epoxidation as a second example of an elec-
troorganic reaction [5]. Both systems provided results comparable
to, or even better than, most efficient industrial cells. Subsequent
studies [6,7] focused on seawater electrolysis for hypochlorite pro-
duction. Scale-up of the original system was reported as well [8].
The last part of this series focused on a mathematical simulation of
paired reactions in an interdigitated electrode system. The results
obtained proved that this system geometry was advantageous for
∗
Corresponding author.
E-mail address: bouzekk@vscht.cz (K. Bouzek).
ISE member.
1
0013-4686/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2010.05.061

K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
8173
paired reaction syntheses primarily due to the overlap of the diffu-
sion layers.
[20,21] recommended operation at elevated pressure to reduce the
volume of the gas phase. It has, however, to be borne in mind that
this approach is limited by gas solubility in the electrolyte solu-
tion increasing with increasing pressure. Gas, in this particular case
hydrogen, dissolved in the electrolyte solution may be oxidized at
the working electrode, thus reducing process efficiency. The list
of the most important contributions to this field is rounded off by
the recent review of microstructured reactors in organic synthesis
published by Ziogas et al. [22]. The authors selected methoxylation
of 4-MA to demonstrate the advantages of this approach over the
classical capillary-gap cell.
From the above it is clear that two-phase flow hydrodynamics
in the interelectrode space, characterized by an extremely short
electrode distance, has a crucial impact on the overall process
characteristics. Flow pattern is one of the most important hydrody-
namic aspects of two-phase flow. A detailed understanding of these
aspects is necessary in order to optimize the reactor design for the
desired operating conditions [23]. Many design parameters, such as
heat transfer or overall reactor efficiency, are affected by the operat-
ing conditions as well as by the two-phase flow pattern. Up to now,
however, only little information is available on this problem, espe-
cially in the case of the in situ electrochemically evolved gas phase.
Our previous study [24] represents one of the first approaches to
provide a deeper insight into this interesting problem.
Matlosz [9] chose an approach corresponding more closely to
a classical understanding of the microstructured electrochemical
cell. He published an analysis of a potential process enhancement
by segmenting a conventional porous electrode in order to control
its potential and thus the current load locally. An opportunity for a
significant increase in process efficiency was identified, the draw-
back being an enhanced ohmic drop in the electrolyte due to the
great distance to the counter-electrode. The results of this theo-
retical analysis were proven experimentally using the example of
the electrooxidation of sodium gluconate to D-arabinose [10]. This
study resulted in the identification of the optimum sectioning of a
porous anode for the given purpose. 500-␮m electrode slices were
identified as a realistic target.
At the end of the 1990s, Löwe and Ehrfeld published a survey of
the current situation in the field of micro-reactor technology which
also focused on electrochemical processes [11]. 4-Methylanisole
(4-MA) was reported to be one of the suitable model reactions to
be studied in a microstructured electrochemical cell. In a succeed-
ing study [12], the construction of a sophisticated microstructured
electrochemical reactor was described, allowing operation both in
monopolar as well as bipolar mode with divided as well as undi-
vided electrode chambers. The reactor was tested at overpressure
up to 3 MPa and was used for D-arabinose synthesis.
Gas–liquid flow patterns are complex in nature and are affected
by many factors. In closed channels with a diameter in the order
of a few millimeters, the flow patterns are dominated by the
effect of gravity. In microchannels, with diameters of 100 ␮m or
lower, the flow patterns are largely a function of interfacial ten-
sion, liquid phase viscosity and wall friction force. Various studies
of the gas–liquid patterns in microstructured devices have been
published [25–27]. The flow pattern determines the quality of
the dispersion of the two phases. It may be characterized by the
interfacial area between phases, void fraction, bubble size dis-
tribution, pressure drop and liquid film thickness, as well as by
the internal mixing quality in each phase. The physicochemical
properties (viscosity, density, interfacial tension) of the two phases,
the wettability of the microchannel walls and the cross-sectional
(circular, triangular or rectangular, with a wide range of aspect
ratios) and contacting geometries (T-junctions, co-axial capillaries)
of the microchannel must also be taken into account [26,28–35].
Another aspect is the presence of an electric field in microdevices.
In the engineering literature on hydrodynamics, the effect of elec-
tric forces is usually neglected, but in special cases (dimensions
reduced to the microscale) it plays an important role [36–39].
As follows from the above, microstructured devices have signif-
icant potential for utilization in industrial electroorganic synthesis.
The main challenge is the question of the accumulation of the
gaseous phase in the interelectrode space. Two main approaches to
solving this problem are possible. The first is operation at elevated
pressure, which has already been discussed [20,21]. An alternative
approach is periodic removal of the gas phase from the interelec-
trode space. It has, however, to be borne in mind that the residence
time of the electrolyte in the interelectrode space has to be suf-
ficient in order to achieve a high degree of conversion within a
single-pass. Thus the main aim of this study is:
The results of, for example, Belmont and Girault [5] stim-
ulated further intensive study of the potential utilization of
microstructured reactors for paired electrosynthesis, i.e. elec-
trosynthesis without added supporting electrolyte [13–17]. Such
an approach is advantageous in cases where added supporting
electrolyte is expensive, where it can cause significant prob-
lems of equipment corrosion and/or its separation and recycling
are difficult. The absence of added conducting salt can also
positively influence the reaction mechanism paths and allow
the synthesis of different products. On the other hand, elec-
trode processes should not be coupled if the desired reaction
product undergoes degradation at the counter-electrode. The
first four of these studies [13–16] provided proof of the prin-
ciple of utilizing synthesis of tetraethyl-ethanetetracarboxylate,
2,5-dimethoxy-2,5-dihydrofurane, 4,4-dinitrodibenzyl dimer, and
2,5-dimethoxy-2,5-dihydrofuran respectively. The last paper in the
series [17] comprises an extensive overview of potential paired
and coupled electrode processes. The authors concluded that a
number of novel electrosynthesis processes would be viable on con-
dition that the anode and cathode were brought together and that
they were designed to allow only selective transformation.
The basic idea of a segmented working electrode allowing locally
controlled electrode potential and current density was further
developed by Rode et al. [18] who carried out a theoretical study
of the influence of anode segmentation on electrochemical 4-MA
methoxylation in a rectangular channel with a short electrode
distance (100 ␮m). Model calculations have shown the superior
performance of the microstructured segmented cell over the indus-
trially used set-up. This conclusion initiated an extensive study
of this interesting problem. Attour et al. [19] reported an exper-
imental study of the methoxylation reaction in a thin-gap cell
with a segmented electrode and demonstrated the efficiency of
this type of cell. The main obstacle observed was the significant
heat generation. A second major problem was the intensive gas
evolution due to the reaction at the counter-electrode. Thus, in
the next step, the authors’ attention focused on electrolysis at
elevated pressure [20,21]. This study demonstrated that a pre-
requisite for operation with substrate concentration and current
density relevant to industrial application is to solve the problem
of ohmic resistance and current density irregularities enhanced
by the hydrogen phase generated at the electrode. The authors
•
to obtain more information on the influence of the operational
conditions of the microstructured electrochemical reactor on
the two-phase flow hydrodynamics of the electrolyte containing
hydrogen bubbles generated at the cathode;
•
to design and construct a corresponding microstructured electro-
chemical cell that is potentially suitable for industrial application;
•
to find experimentally optimal operational conditions of the
microstructured cell constructed in the previous point and to
determine its performance;

8174
K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
on the experimental arrangement and procedures are given in
[42].
2.3. Microstructured cell performance
The microstructured electrochemical cell designed and tested
in this study took the form of a bipolar cell of the filter-press type,
see Fig. 2. In the following it is denoted as FP BEMR. The cell body
was made of Hard Resin CR 39 (Great Lakes Chemical Co., USA)
which was also used for the construction of the flow visualiza-
tion cell. The glassy carbon electrodes (0.028 m long and 0.009 m
wide) were placed in the centre of the individual segments of a cell
stack. The interelectrode gap measured 1 × 10−4 m (alignment pre-
cision 1 × 10−5 m). The first and the last (terminal) electrodes were
connected to the power supply. The polished cathode surface was
treated with an emery paste of grade 800. The resulting average
surface roughness was characterized by atomic force microscopy
(AFM) as 0.298 ␮m. The aim was to introduce defects on the sur-
face causing a significant reduction in size of the hydrogen bubbles
formed. The experimental set-up was similar to that depicted in
Fig. 1. The electrolyte with 100 mol m−3 4-MA and 10 mol m−3 KF
was pumped to the FP BEMR by a linear syringe pump identical
to the flow visualization experiments. Product was collected at the
cell outlet for subsequent analysis.
Fig. 1. Experimental set-up for visualization of gas–liquid flow with electrochemi-
cally generated gas phase.
•
to evaluate the efficiency of the process with an integrated
microstructured reactor with respect to existing industrial tech-
nology.
2.4. Chemicals
During the experiments, the following chemicals were
used as received: 4-MA (Acros Organics, purissimum grade),
In order to accomplish these tasks, anodic methoxylation of
4-MA was chosen as a model reaction. It was selected because it
represents an industrially highly relevant reaction that has already
been studied in the literature. A basic understanding of the sys-
tem’s behavior and the optimum operational conditions in classical
electrochemical reactors is thus known [40].
4-methoxybenzaldehyde-dimethyl-acetal
and
4-methoxy-
benzaldehyde (Fluka, purissimum grade), methanol (Merck,
HPLC grade). Potassium fluoride (Acros Organics, analytical grade)
was dried at 125 ◦C for 8 h before use and was stored in a desiccator
over P4O10 to keep the amount of water in the electrolyte solution
as low as possible. The content of water in the electrolyte solution
was less than 0.5 vol.%.
2. Experimental
2.5. Analytical procedures
2.1. Two-phase flow hydrodynamics
Samples of the reaction mixture were analyzed using a Shi-
madzu GC17A gas chromatograph (Shimadzu Corporation, Japan).
Approximately 0.5 ml of inner standard (100 mol m−3 4-tert-
butyltoluene in methanol) was added to 1 ml sample of the reaction
mixture. The mass of the solutions was determined using an
analytical balance with a precision of 0.1 mg. A CP-SIL 8CB gas chro-
matography column (ChromPack International BV, Netherlands)
30 m long and 0.32 mm in diameter was used. The film thickness
was 0.25 ␮m. Helium was used as the carrier gas. The reproducibil-
ity of the analytical method was within 5%.
The flow visualization cell was made of glass and Hard Resin CR
39 (Great Lakes Chemical Co., USA). The glass channel was 0.01 m
wide and 0.10 m long. The channel gap was 1 × 10−4 m. Two pairs
of glassy carbon (Sigradur G) electrodes were embedded in the
glass channel walls, allowing electrochemical generation of the gas
phase. Electrolyte solution was pumped through the microchan-
nel by means of a linear syringe pump (KD Scientific Inc., USA).
The cell outlet was open to the atmosphere. Two-phase flow in
the microchannel was visualized by a Nikon D70 digital camera.
Digital images were analyzed by Matlab Image Processing Toolbox.
The flow pattern was determined by original images of two-phase
flow, and bubble size distribution was evaluated by image process-
ing data. The experimental set-up is shown in Fig. 1. More details
on the experimental set-up are provided in [41].
3. Results and discussion
3.1. Two-phase flow hydrodynamics
2.2. Electrode reaction kinetics
Flow visualization results showed that, in a thin-gap chan-
nel with a very low aspect ratio, four distinct flow patterns
exist: flow with no bubbles, dispersed bubbly flow, bubbly flow,
and churn flow. Examples of the flow patterns observed are shown
in Fig. 3. This wide range of flow patterns represents a significant
difference between two-phase flow hydrodynamics in a thin-gap
microchannel in comparison to circular or square microchannels,
where mostly a combination of slug (or Taylor) and annular flow
can be observed [26,29,32].
A standard three-electrode arrangement with a glassy car-
bon (GC) (Sigradur G, Germany) rotating disc electrode (RDE)
PINE MSRX rotator (PINE Research Instrumentation, USA) was
used. A Pt basket or sheet served as an auxiliary electrode. The
experimental electrochemical cell was made of glass. A saturated
calomel electrode (SCE) was used as the reference. All the elec-
trode potential values given in this work refer to this electrode.
The electrode potential was controlled by a HEKA PG 310 (HEKA
Elektronik GmbH, Germany) potentiostat/galvanostat. More details
For the desired electrochemical alkoxylation reaction, flow with
no bubbles or dispersed bubbly flow with bubble size below or

K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
8175
identified. The influence of electrolyte flow rate and current den-
sity was studied within this work. The third parameter, i.e. working
pressure, was not the subject of this particular study. An example
of the results obtained is depicted in Fig. 4. Generally, higher elec-
trolyte flow rates and lower current loads lead to the preferred flow
pattern. However, this is in direct contradiction to desirable condi-
tions from an industrial point of view. This further emphasizes the
issues of optimization of the cell geometry, overall system arrange-
ment and operational conditions in order to minimize the problems
induced by the existence of two-phase flow in the interelectrode
space.
3.2. Electrode reaction kinetics
The reaction kinetics for KF as the supporting electrolyte was
determined by means of quasi-stationary linear sweep voltamme-
try at the RDE and classical linear sweep voltammetry. A transfer
of 4e− is required to oxidize one 4-MA molecule to the desired
product, i.e. 4-methoxybenzaldehyde-dimethyl-acetal (4-MBA-
DMA). The reaction proceeds via 4-methoxybenzyl-methyl-ether
as a stable intermediate. The consequent oxidation of 4-MBA-DMA
to the corresponding orthoester (2e− transfer) or higher oxidation
products leads to an undesired degradation of the product, thus
reducing the efficiency of the overall process. A scheme of 4-MA
anodic methoxylation is shown in Fig. 5. The reaction mechanism
is influenced by the presence of KF. A detailed discussion of this
problem is provided in [42].
The kinetic parameters determined for anodic 4-MA oxidation
were evaluated in the form of Eq. (1).
)/b
jA = jEref,A 10(E−E
(1)
A
ref,A
The dependence of the first kinetic parameter, i.e. current
density at the formal reference potential Eref,A indicating a rise
in current, on the substrate concentration in the electrolyte is
described by Eq. (2). This approach was chosen because the reaction
under study is irreversible and classical exchange current density
cannot be evaluated [42].
jE
= nAFkAc4-MA
(2)
ref,A
In the present case number of exchanged electrons nA was
considered to be equal to 4, for Eref,A = 1.31 V kinetic constant kA
was determined experimentally as equal to 3.87 × 10−1 m s−1 and
c4-MA stands for the substrate concentration. The second kinetic
parameter bA, i.e. the anodic Tafel slope, was determined to be
0.107 V dec−1
.
The onset potential of undesired 4-MBA-DMA overoxidation is
only about 140 mV more positive than the onset potential of 4-MA
oxidation. Despite this rather small difference in onset potentials,
4-MBA-DMA overoxidation proceeded very slowly compared to
4-MA oxidation. This fact allows the process to be operated with
high selectivity.
Protons released during the anodic reaction, see Fig. 5, are
reduced at the counter-electrode under formation of gaseous
hydrogen. This reaction is known to proceed with high overpo-
tential at the GC electrode. This was confirmed experimentally.
The onset potential Eref,C of approx. −1.6 V and Tafel slope of
−0.200 V dec−1 indicate very slow reaction kinetics. The kinetic
equation takes the form of Eq. (3).
Fig. 2. (A) Photograph of the filter-press bipolar electrochemical micro-reactor
designed and tested within the framework of this study; (B) Details of one
bipolar electrode plate, plate thickness 0.005 m, electrolyte inlet/outlet 0.002 m
I.D., electrode channel (H × W × D) 0.05 m × 0.009 m × 1 × 10−4 m, active electrode
area 0.028 m × 0.009 m, non-conductive part of electrode channel (H × W × D)
0.011 m × 0.009 m × 1 × 10−4 m, 1—electrolyte outlet, 2—active electrode area,
3—electrolyte inlet, 4—non-conductive part of electrode channel, 5—electrode chan-
nel; (C) Schematic sketch of the electrolyte flow and electric current connections of
the cell.
jC = jEref,C 10(E−E
(3)
)/b
C
ref,C
In this particular case, the value of jE
is equal to −1.0 A m−2
.
ref,C
equal to the channel depth is the desired flow pattern [24]. How-
ever, in an industrially relevant case flow with no bubbles is not
realistic because of the relatively high substrate concentration,
which necessitates a sufficiently high current load. Thus opera-
tional conditions yielding a dispersed bubbly flow pattern had to be
This is given by the fact that the species reacting at the cathode is
methanol, i.e. a solvent. Its concentration in the electrolyte can be
considered to be constant. The value follows from the methodology
used to determine the value of Eref,C described above.

8176
K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
Fig. 3. Typical flow patterns observed in the visualization cell for the 4-methylanisole methoxylation reaction. The following experimental conditions correspond to the
individual cases: (A) current density 400 A m−2, flow velocity 0.17 m−1, (B) current density 1500 A m−2, flow velocity 0.17 m s−1, (C) current density 1500 A m−2, flow velocity
0.017 m s−1 and (D) current density 5 500 A m−2, flow velocity 0.017 m s−1. Width of the visualization cell: 0.01 m.
design which would solve the main issues relating to the utilization
of this type of equipment on an industrial scale.
One of the common problems connected with increasing the
production capacity of a technology based on microstructured reac-
tors by increasing the number of elementary cells is the extremely
complicated hydraulic and electric circuitry. This problem can
be partly solved by means of a bipolar cell arrangement. As is
known, however, a significant risk of this approach is a decrease
in process efficiency. This is caused by the existence of by-pass or
shunt currents [43] which are influenced by the system geometry,
number of bipolar electrodes, electrolyte conductivity, electrode
reaction kinetics and the conductivity of the electrode mate-
rial. Careful optimization of the cell parameters was, therefore,
ensured by means of mathematical modeling of the local poten-
tial and current density distribution. The cell design proposed was
already shown in Fig. 2. The filter-press arrangement allows flexible
modification of the number of cells in a single stack. The inter-
Fig. 4. Hydrogen bubble size distribution in the thin-gap channel at atmospheric
electrode channels have a conical shape at the electrolyte inlet
and on the top around the outlet collector. The reason is to pro-
vide homogeneous flow distribution at the electrolyte inlet into
the interelectrode space as well as efficient removal of the mix-
ture of hydrogen and electrolyte at the cell outlet. Each electrode
gap comprises an electrochemically non-active part below and
above the electrode, increasing ohmic resistance to the current
flowing through the hydraulic circuitry for electrolyte distribution
pressure and current density of 200 A m−2. Electrolyte flow rate at which indi-
vidual distributions were obtained: (ꢁ) 8.3 × 10−3 m s−1, (ꢂ) 16.7 × 10−3 m s−1 and
(ꢀ)41.7 × 10−3 m s−1
.
3.3. Microstructured cell design
Information obtained from the previous parts of the study was
used to propose a novel microstructured electrochemical reactor
Fig. 5. Mechanism of anodic 4-methylanisole methoxylation; (A) 4-methylanisole, (B): 4-methoxybenzyl-methyl-ether, (C) 4-methoxybenzalhyde dimethylacetal,
(D) 4-methoxybenzoic acid orthoester.

K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
8177
As is apparent, at this stage the model used neglects the effect
of the gas phase present in the interelectrode space on the current
density distribution and parasitic current values. This is due to the
fact that the main purpose of the model was to quantify the effect of
the cell geometry and current load on the parasitic current value. In
order to solve the problem of the influence of the gas phase on the
current density distribution along the electrode, additional infor-
mation is needed on the size distribution of the hydrogen bubbles
in dependence on the position in the cell. Otherwise, simplification
assumptions have to be adopted. The main impact of the presence
of the gas phase in the interelectrode space on the current density
distribution consists in locally enhanced electrolyte ohmic resis-
tivity. This effect may be neglected at the cell inlet and it increases
with increasing distance from that point. At the cell outlet, however,
the hydrogen produced accumulates. The outlet channels, there-
fore, may be considered as to a great extent non-conductive. This
is also the case of the proposed model which considers the elec-
trolyte collecting channel as filled by gas and, thus, non-conductive.
This approach represents an approximation of the real situation. Its
accuracy is sufficient for the desired purpose, i.e. to predict the effi-
ciency «P of the proposed bipolar cell geometry with respect to
the parasitic currents. This parameter («P) is defined as the ratio of
the electric charge utilized for the desired electrochemical process
to the entire charge supplied to the individual cells [43]. It can be
expressed in the form of Eq. (7), where N denotes the number of
the bipolar electrochemical cells, (N − 1) gives the number of the
bipolar electrodes in the stack, IT corresponds to the current load
on the bipolar micro-reactor and finally Ibe,i is the current flowing
through the bipolar electrode i.
Fig. 6. Local Galvani potential distribution in the electrode and electrolyte phase
of the FP BEMR of the optimized geometry; (a) terminal anode, (b) bipolar elec-
trodes, (c) terminal cathode, (d) electrolyte distribution channel, (e) electrolyte inlet;
arrows indicate direction of the electrolyte flow in the domain, CC–current collector,
CF–current feeder. The color bar indicates the Galvani potential value in volts. Input
parameters: current load at terminal electrodes 790 A m−2, electrolyte conductivity
0.051 S m−1 and electrode phase conductivity 2.2 × 104 S m−1. For the calculation,
electrode reaction kinetics expressed by Eq. (6) was used. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of
the article.).
ꢀ
and collection. The intention is to decrease the value of the by-pass
current.
(N−1)
IT
+
I
be,i
i=1
P
=
· 100%
(7)
NIT
2-D and 3-D mathematical models of the filter-press type
reactor were proposed and implemented in the COMSOL Multi-
physics environment. Both of them are based on the calculation
of the secondary local current density (j) and Galvani potential
(ϕ) distribution. The electric field in the individual model domains
(electrolyte, electrodes) is described by Laplace’s equation, Eq. (4),
and local current densities are evaluated by means of Eq. (5). In both
domains constant conductivity (ꢀ) is considered. The subscripts E
and M stand for electrolyte and metal, respectively.
The values of the individual constants used for the calcula-
tion are summarized in the previous Section 3.2. The electrode
length has to be considered since it is one of the important factors.
Increased electrode length reduces the importance of the parasitic
current with respect to the total cell current load. On the other hand,
it causes excessive accumulation of the gas phase in the interelec-
trode space. This phenomenon can be compensated in two different
ways: (i) by reducing the average current density used; (ii) by
increasing the electrolyte flow rate. The first approach substantially
reduces the positive effect on the significance of parasitic cur-
rents gained by lengthening the electrodes. The second approach
then imposes increasing demands on the pump’s performance and
energy consumption (bearing in mind the small interelectrode dis-
tance). Thus, the definition of a suitable electrode length clearly
needs to be optimized. In the present case an electrode length of
0.028 m was considered. This was an optimal length from the point
of view of the laboratory experiment (preciseness of the cell pro-
duction, pump’s performance, chemicals consumption, amount of
product needed for the analysis, etc.). For pilot plant scale or indus-
trial application a detailed analysis has to be provided. The potential
and current fields presented here, i.e. in the domain of the elec-
trolyte distribution and collection channels, are influenced by the
electrode extension, providing the remaining aspects of the cell
geometry remain unchanged.
2
∇ ϕE(M) = 0
(4)
(5)
−ꢀE(M)∇ϕ = j
The Galvani potential discontinuity at the electrode–electrolyte
interface is evaluated by Butler–Volmer-type electrode reaction
kinetics in the form of Eq. (6) determined experimentally, see
Section 3.2. The continuity of the charge flux at the electrode-
electrolyte interface is considered.
/b
)/b
C
j = jEref,A 10(E−E
) + jEref,C 10(E−E
(6)
A
ref,A
ref,C
In accordance with the previous notation, E in Eq. (6) denotes
electrode potential at the electrode-electrolyte interface. It is
defined as E = ϕM − ϕE. Eq. (6) represents the boundary condition
on the boundaries of the electrodes in contact with the electrolyte.
The boundary conditions used for the boundary of the terminal
electrodes connected to the power supply are denoted in Fig. 6 as
“current collector” and “current feeder”. Their boundary connected
to the external power supply is characterized by zero Galvani
potential (ϕM = 0) in the case of the cathode (current collector)
and by constant current density (j = I/S) in the case of the anode
(current feeder). S denotes electrode surface area. The remaining
boundaries within the model represent electric insulation charac-
terized by the zero gradient of the Galvani potential in the direction
normal to the boundary (∂ϕE(M)/∂nb = 0).
The Galvani potential field distribution in the cell with an opti-
mized geometry (except the interelectrode distance) is shown in
Fig. 6. In this particular case an interelectrode distance of 0.001 m
was used for graphical purposes to render the potential distribu-
tion clearly visible. As can be seen, the main potential modifications
relating to the parasitic current flow occur in the flow distribu-
tion channel at each individual cell. The reason is that its active
cross-section is substantially smaller compared to the electrolyte
supply channel. This indication is further confirmed by the depen-

8178
K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
Fig. 7. Dependence of cell efficiency on the current load for various interelectrode
distances (i.e. cross-sections of the flow distribution channels). Input parameters:
current load at terminal electrodes 0–16 A m−2, electrolyte conductivity 0.051 S m−1
and electrode phase conductivity 2.2 × 104 S m−1. For the calculation, electrode reac-
tion kinetics expressed by Eq. (6) was used.
Fig. 8. Dependence of cell efficiency on current load for various numbers of bipolar
electrodes. Input parameters: current load at terminal electrodes 0–16 A m−2, elec-
trolyte conductivity 0.051 S m−1 and electrode phase conductivity 2.2 × 104 S m−1
.
For the calculation, electrode reaction kinetics expressed by Eq. (6) was used.
In the present study four bipolar electrodes were used. This
arrangement multiplies the capacity of the single-channel. This
set-up is characterized by sufficient capacity and at the same
time uniform flow distribution between the individual channels.
The mathematical modeling results obtained indicate that the uti-
lization of a high electric current load can be attained using the
proposed system geometry. Moreover, the proposed cell design
allows the gas phase to be separated from the liquid phase during
the transition between two electrochemical cells. The electrolyte
solution is then brought into the second cell connected in series
in order to increase the degree of substrate conversion without
the negative impact of the gas phase on process selectivity. The
number of cells connected in series determines the entire degree
of conversion achieved.
dence of cell efficiency on the current load calculated for various
interelectrode distances, i.e. the varying thickness of the flow distri-
bution channel, shown in Fig. 7. As is apparent from the calculated
dependencies, the efficiency of current utilization increases with
the increasing current load of the cell. This is logical behavior,
reflecting the influence of the ohmic drop necessary to compensate
the equilibrium voltage of the single-cell reaction and the kinetics
of the electrode reactions. As the current load increases, the relative
portion of the current flowing through the electrolyte distribution
channels necessary to fulfil this requirement decreases. This is due
to the more or less constant equilibrium potential of the electrodes
and the exponential shape of the polarization curves [43]. At the
same time it is obvious that the cell efficiency decreases by increas-
ing the active cross-section of the flow distribution channels. Once
again, this behavior reflects the basic laws governing the utilization
of a bipolar electrode. Quantification of this phenomenon by the
mathematical model designed shows that a reduction of the inter-
electrode distance, and thus the thickness of the flow distribution
channels, to below 5 × 10−4 m allows a significant enhancement
of the cell’s efficiency, especially in the low current load region.
Therefore microstructuring is also advantageous from this point of
view.
Since in this particular study an increase in the production
capacity of the microstructured cell is proposed in the form of an
increase in the number of bipolar cells, the dependence of cell effi-
ciency on this parameter is of great interest. The results of the model
calculations are summarized in Fig. 8. The efficiency clearly signif-
icantly decreases with the increasing number of bipolar cells up to
approximately 50. With a higher number of cells the influence of
this parameter becomes negligible. This result indicates that the
number of cells has to be kept as low as possible to maintain the
advantage of high cell efficiency. This is, however, in contradiction
to the required increase in cell capacity. A further approach is to
increase the number of cells as much as possible in order to take
advantage of the high capacity, by which the increasing number of
cells no longer has an additional negative impact on cell efficiency.
Here, however, the implications have to be considered of a signif-
icantly more complex design to ensure uniform flow distribution
between the individual cells as well as heat management.
3.4. Microstructured cell performance
The proposed bipolar microstructured cell was constructed and
tested using the model reaction selected. In the first instance, the
influence of the current load of the cell and electrolyte flow rate
was tested in a single-channel cell in order to eliminate the effect
of the by-pass current on the results. In order to permit a direct
comparison of the results attained, a dimensionless current I* was
introduced. It is defined by Eq. (8),
1
I∗
=
(8)
n FQLCA
0
where I stands for the absolute current load, n for the num-
ber of electrons exchanged during the desired reaction, F for
Faraday charge, QL for the volumetric flow rate of the electrolyte
and CA0 for the substrate concentration at the cell inlet. A dimen-
sionless current indicates the fraction of current load required for
the total conversion of the substrate brought to the cell in the
desired product. The results are shown in Fig. 9. It is evident that,
for the two electrolyte flow rates studied, a degree of conversion of
90% or higher can be obtained when the dimensionless current I*
approaches unity, i.e., from the point of view of substrate conver-
sion, the current efficiency exceeds 90%. The electrolyte flow rate
does not have a significant effect on this value.

K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
8179
Fig. 10. Dependence of the degree of conversion of 4-methylanisole (ꢀ) and reac-
tion selectivity (ꢂ) on the number of electrolytic cells in series, between individual
cells the gas phase was separated from the electrolyte; dimensionless current used
in the individual cells is indicated on the right-hand side y axis as (ꢁ); electrolyte
flow velocity 1.34 × 10−3 m s−1; cell configuration with four gaps (three bipolar elec-
trodes); substrate concentration 100 mol m−3, supporting electrolyte concentration
10 mol m−3, pressure 0.1 MPa, room temperature.
Fig. 9. Effect of electrolyte flow rate on the degree of conversion of 4-methylanisole
and reaction selectivity dependence on the dimensionless current load. Electrolyte
flow velocityꢁ 7.66 × 10−4 m s−1 and ꢂ 1.92 × 10−3 m s−1;filledsymbols correspond
to the degree of conversion of the substrate, empty symbols to process selec-
tivity; substrate concentration 100 mol m−3, supporting electrolyte concentration
10 mol m−3, room temperature.
The situation is completely different in the case of pro-
cess selectivity with respect to the desired oxidation product,
i.e. 4-MBA-DMA. Here the lower linear flow velocity of the elec-
trolyte of 7.66 × 10−4 m s−1 makes it possible to attain selectivity of
70% at a degree of conversion of 82%. These values are in good agree-
ment with the optimum electrochemical conditions predicted by
the mathematical model [18]. In the case of electrolyte flow velocity
of 1.92 × 10−3 m s−1 at the same dimensionless current value, a con-
version of 93% and selectivity of 40% were achieved. This is clearly
connected to the higher absolute current load used in the case of
the higher electrolyte flow rate. If a higher amount of substrate is
brought into the cell within the time unit, a higher current load
is required to ensure the conversion of the substrate. Although a
higher amount of hydrogen generated at the counter-electrode is
counterbalanced by a higher electrolyte flow rate, different bub-
ble sizes can be expected in the interelectrode space, which causes
more extensive disturbance in the local potential and current den-
sity distribution, thus diminishing process selectivity.
substrate concentration and electrolyte flow rate. Thus, the over-
all system performance increases. It should be pointed out that,
in this particular case, the cell was working in a bipolar arrange-
ment with three bipolar electrodes, i.e. four electrolytic cells. Since
the substrate conversion of nearly 100% was obtained at dimen-
sionless current I* equal to unity, it is possible to conclude that a
parasitic current is not a major issue in the given geometry. This is
in agreement with the results of the mathematical model.
Although this represents a significant improvement in process
efficiency, the substrate concentration of 100 mol m−3 is still unre-
alistically low for industrial application. Electrolysis with increased
4-MA concentration in the electrolyte of 500 mol m−3 was, there-
fore, tested in the next step. In order to minimize the negative
effect of the hydrogen phase, an elevated operational pressure of
0.3 MPa was used. The results obtained are summarized in Fig. 11.
Elevated pressure allowed utilization of similar dimensionless cur-
rent values even though the concentration was five times higher
than in the previous case. A higher degree of conversion per single-
pass was attained. In this case, the by-pass current is an even less
In the next step, the efficiency of the proposed method of elim-
inating the negative influence of the gas phase was tested, see
Fig. 10. Four cells in a configuration with three bipolar electrodes
(four interelectrode gaps) were connected in series. Between each
two of them the gas phase was separated from the electrolyte solu-
tion. The larger total area of the electrode allowed the electrolyte
flow rate to be increased, while maintaining correspondingly low
current density. The dimensionless current used for each individual
reactor is shown in the same figure. The values were set empirically
to obtain the highest process selectivity. In the first reactor, with
the highest substrate concentration, a current load corresponding
to a dimensionless current of I* = 0.53 was used. This corresponds
to a current density of 100 A m−2 due to the highest substrate
concentration at the cell inlet coupled with a relatively short work-
ing electrode. As the concentration of substrate decreases, the
dimensionless current load increases in order to ensure complete
substrate conversion. In the last cell it reached I* = 2.09, i.e. a current
density of 36 A m−2
.
The gas separation between the individual cells has proven
to be highly efficient. As shown in Fig. 10, with the single-pass
arrangement this set-up permitted 4-MA conversion of 99%, while
maintaining process selectivity at 79%. This is an extremely pos-
itive result in contrast to the results obtained for the single-cell
arrangement shown in Fig. 9. This documents the fact that increas-
ing the number of cells allows a further increase in the initial
Fig. 11. Dependence of the degree of conversion of 4-methylanisole (ꢀ) and reac-
tion selectivity (ꢂ) on the number of electrolytic cells in series, between individual
cells the gas phase was separated from the electrolyte; dimensionless current used
in the individual cells is indicated on the right-hand side y axis as (ꢁ); electrolyte
flow velocity 6.64 × 10−4 m s−1; cell configuration with four gaps (three bipolar elec-
trodes); substrate concentration 500 mol m−3, supporting electrolyte concentration
10 mol m−3, pressure 0.3 MPa, room temperature.

8180
K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
Fig. 12. Scheme of the simulated process, COL1 and COL2 stand for the distillation columns, DIST1 for separated methanol, DIST2 for separation of unreacted substrate and
LIQ2 for the product stream.
important issue due to the increased absolute value of the current
load, see Section 3.3. As documented in Fig. 11, this process inten-
sification was not obtained at the expense of the overall process
selectivity.
A further important issue is the increase in process selectivity. It
is still not clear whether selectivity can be enhanced by improving
the cell design, which would ensure an even higher uniformity of
electrolyte flow under conditions of intensive hydrogen evolution.
It is important to bear in mind that, under the conditions used,
the flow pattern inside the electrochemical micro-reactor coincides
with bubbly flow, i.e. this is not an optimum situation from the point
of view of current density distribution.
separation process of the current technology is shown in Fig. 12. The
first distillation column (COL1) was mainly designed for methanol
separation from the electrolyte, the second column for the separa-
tion of the unreacted 4-MA from the product (4-MBA-DMA). The
purity of 4-MBA-DMA in a LIQ2 stream was set at 99.5% (mol). The
concentration of 4-MBA-DMA in the distilled output of the second
column (DIST2) was held at around 2% to avoid undesired recircu-
lation of the product. The results of the simulations for individual
cases are presented in Table 2.
It can be seen from these results that, under the desired
conditions, the energy consumption for methanol and 4-MA
recycling/separation is, to great extent, dependent on two factors:
the yield of 4-MBA-DMA and the concentration of the organics in
the methanol. The values presented of boiler duties clearly show
that the bulk of the energy is used in the methanol separation
column (COL1). Energy consumption in COL1 as well as in COL2
decreases substantially with increasing 4-MA conversion. This
demonstrates that a considerable amount of energy can be saved
in a high-conversion (integral) reactor compared to a low con-
version (differential) one. On the other hand, it is evident that
the use of a higher 4-MA concentration in a micro-reactor than
that currently used (100–500 mol m−3) is of great importance. If
it can be operated with a 4-MA content of 1000 mol m−3 as in the
BASF process, the advantage gained will be enormous. The second
aspect is the size of the distillation columns, which corresponds
to the flux of the separated mixture. From this point of view, the
high-conversion micro-reactor also seems to be the best choice.
3.5. Process intensification with respect to an industrial process
Four-electron anodic methoxylation of 4-MA is currently per-
formed by BASF on an industrial scale with a production rate
of 3500 tons of anisaldehyde per year. The electrolysis step is
performed in a capillary-gap cell operated in continuous mode
and placed in a recycle loop. A process diagram, including the sep-
aration steps, is given in [40]. The aim of the present section is to
evaluate potential process intensification related to a change in the
reactor design. The calculations are based on performances deter-
mined in the microstructured reactors tested within the framework
of this work. This approach is based on one of the advantages of
using microstructured devices, namely process scale-up by increas-
ing the number of elementary cells. The results obtained using
a laboratory cell can thus be directly utilized to predict the per-
formance of an industrial system. For purposes of comparison,
the parameters of the two processes considered are shown in
Table 1. The currently installed production capacity of 3500 tons
of anisaldehyde per year is considered in all calculations in this
section.
Table 1
Comparison of the BASF process as depicted by Pütter in [40] and measurements
performed in a bipolar cell with four gaps.
Process parameters
BASF industrial process
(Pütter [40])
Microstructured bipolar
cell
Process intensification related to the utilization of a microstruc-
tured reactor can be identified in more steps of the entire
production technology. The two most obvious ones are the elec-
trolysis itself and the product separation. Since the content of
the supporting electrolyte is intentionally kept as low as possible
and due to the more pronounced effect of the hydrogen accumula-
tion in the interelectrode space in the case of microstructured cell
(distance between the electrodes comparable with the diameter of
the bubbles), in both cases (the conventional and the microstruc-
tured cell) the cell voltage will retain a similar value. In this case,
therefore, the electrical energy consumption remains very similar
to the classical technology. The separation part of the process offers
more possibilities for saving energy and, thus, costs. A scheme of the
Overall process flow
scheme
Electrode material
Interelectrode gap
4-methoxytoluene inlet 10–25% mass
concentration
Continuous with recycle
Single-pass
high-conversion
Glassy carbon
Bipolar graphite rings
5–10 × 10−4
m
1 × 10−4
m
1.4% mass (100 mol m−3
)
(650–1650 mol m−3
)
Supporting electrolyte
concentration
0.3–3% mass potassium
fluoride, sodium sulfonate fluoride (0.01 M)
Atmospheric Atmospheric
0.07% mass potassium
Operating pressure
Operating temperature 40–50 ◦C
45 ◦
C
Current density
Voltage per gap
Overall conversion
Selectivity
300–500 A m−2
4–6 V
90–99%
85%
10–400 A m−2
4–6V
99%
79%
78%
Material yield
>80%

K. Bouzek et al. / Electrochimica Acta 55 (2010) 8172–8181
8181
Table 2
Comparison of the necessary stages and the energy consumption of the distillation columns for the separation of the product from the electrolyte related to a production
capacity of 3500 tons of anisaldehyde per year. For each variant tested, two columns are considered (see Fig. 12). Conversion typically attained by the BASF process is not
known, therefore a parametric study is provided in an interval considered as the most realistic one.
BASF (a) BASF (a) BASF (b) BASF (b) BASF (c) BASF (c) Micro-reactor
100 mol m−3
Micro-reactor
100 mol m−3
Micro-reactor
500 mol m−3
Micro-reactor
500 mol m−3
Conversion
Selectivity
Column
Stages
0.2
0.2
0.3
0.3
0.4
0.4
0.99
0.79
COL1
3
0.99
0.79
COL2
9
0.95
0.83
COL1
3
0.95
0.83
COL2
12
0.85
COL1
2
0.85
COL2
11
0.85
COL1
3
0.85
COL2
11
0.85
COL1
3
0.85
COL2
12
Molar reflux ratio
Flux to column [kg/hr]
Reboler duty [kW]
4-MBA DMA purity [%
(mass)]
0.1
16720
5339
2.5
0.05
10962
3230
2.6
0.05
8060
2390
2.6
0.05
29440
9708
0.11
666
96
0.05
6227
1916
1.2
606
91
1762
830
99.7
1605
636
99.8
1172
337
99.5
99.6
99.6
References
4. Conclusions
[1] D. Pletcher, F.C. Walsh, Industrial Electrochemistry, second ed., Chapman and
Hall, London-New York, 1990.
[2] W. Ehrfeld, V. Hessel, H. Löwe, Microreactors-New Technology for Modern
Chemistry, Wiley-VCH Verlag, Weinheim, 2000.
[3] C. Belmont, H.H. Girault, J. Appl. Electrochem. 24 (1995) 475.
[4] C. Belmont, H.H. Girault, J. Appl. Electrochem. 24 (1995) 719.
[5] C. Belmont, H.H. Girault, Electrochim. Acta 40 (1995) 2505.
[6] C. Belmont, R. Ferrigno, O. Leclerc, H.H. Girault, Electrochim. Acta 44 (1998)
597.
[7] R. Ferrigno, J. Josserand, P.F. Brevet, H.H. Girault, Electrochim. Acta 44 (1998)
587.
[8] R. Ferrigno, C. Comninellis, V. Reid, C. Modes, R. Scannell, H.H. Girault, Elec-
trochim. Acta 44 (1998) 2871.
In the present study a bipolar microstructured reactor, consist-
ing of four electrochemical cells, was designed and constructed.
It was tested using 4-MA anodic methoxylation as a model reac-
tion. The results obtained confirmed that this reactor provides
electrochemical characteristics (process selectivity and substrate
conversion) very close to the industrial capillary-gap cell oper-
ated by BASF. At the same time it permits substantial investment
and operational costs savings in the subsequent product separation
step of the technology. This is due to the reduction of the elec-
trolyte volume to be treated. The exact value strongly depends on
the concentration of the substrate in the electrolyte at the inlet
and degree of conversion reached. Savings of approximately 50% of
the heat energy were calculated for the mean values of the parame-
ters used. However, savings can substantially increase by increasing
the substrate concentration in the electrolyte solution, provided a
high degree of conversion is maintained in a single-pass of the elec-
trolyte through the series of synthesis cells. The reduction in the
investment costs is related to the reduction in size of the distillation
column.
[9] M. Matlosz, J. Electrochem. Soc. 142 (1995) 1915.
[10] C. Vallieres, M. Matlosz, J. Electrochem. Soc. 146 (1999) 2933.
[11] H. Löwe, W. Ehrfeld, Electrochim. Acta 44 (1999) 3679.
[12] M. Küpper, V. Hessel, H. Löwe, W. Stark, J. Kinkel, M. Michel, H. Schmidt-Traub,
Electrochim. Acta 48 (2003) 2889.
[13] C.A. Paddon, G.J. Pritchard, T. Thiemann, F. Marken, Electrochem. Commun. 4
(2002) 825.
[14] D. Horii, M. Atobe, T. Fuchigami, F. Marken, Electrochem. Commun. 7 (2005)
35.
[15] P. He, P. Watts, F. Marken, S.J. Haswell, Electrochem. Commun. 7 (2005) 918.
[16] D. Horii, M. Atobe, T. Fuchigami, F. Marken, J. Electrochem. Soc. 153 (2006)
D143.
[17] C.A. Paddon, M. Atobe, T. Fuchigami, P. He, P. Watts, S.J. Haswell, G.J. Pritchard,
S.D. Bull, F. Marken, J. Appl. Electrochem. 36 (2006) 617.
[18] S. Rode, S. Altmeyer, M. Matlosz, J. Appl. Electrochem. 34 (2004) 671.
[19] A. Attour, S. Rode, A. Ziogas, M. Matlosz, F. Lapicque, J. Appl. Electrochem. 38
(2008) 339.
[20] S. Rode, A. Attour, F. Lapicque, M. Matlosz, J. Electrochem. Soc. 155 (2008) E193.
[21] A. Attour, S. Rode, F. Lapicque, A. Ziogas, M. Matlosz, J. Electrochem. Soc. 155
(2008) E201.
As anticipated, the main challenge is the gas evolution at the
counter-electrode. The presence of gas in the interelectrode space
disturbs the homogeneity of the local potential and current density
field as well as the flow hydrodynamics, thus causing a deteriora-
tion of the process selectivity. This effect is of special interest in the
case of increased substrate concentration in the electrolyte solu-
tion. In this particular case, the negative effect of this aspect was
[22] A. Ziogas, G. Kolb, M. O’Connell, A. Attour, F. Lapicque, M. Matlosz, S. Rode, J.
Appl. Electrochem. 39 (2009) 2297.
[23] K.A. Triplett, B.N. McCord, S.M. Ghiaasiaan, S.I. Abdel-Khalik, A. LeMouel, Int. J.
already well visible for a substrate concentration of 100 mol m−3
,
where process selectivity substantially deteriorated with increas-
ing degree of conversion. The method proposed in this work, i.e. the
separation of hydrogen from the reaction mixture has proven to be
a promising step towards solving this problem. It has also allowed
an operating cell with sufficient selectivity at a substrate concentra-
tion of 500 mol m−3. Given an additional increase in the number of
cells connected in series, a further concentration increase is possi-
ble. If the number of hydrogen separation steps is too high and the
process set-up thereby too complex, a combination of gas phase
separation with elevated operational pressure will be necessary
to minimize the negative effect of the gas phase. Further work is
needed to ensure the homogeneity of the electrolyte flow in the
cell.
Multiphas. Flow 25 (1999) 395.
[24] J. Krˇisˇt’ál, R. Kody´m, K. Bouzek, V. Jirˇicˇny´, Electrochem. Commun. 10 (2008) 204.
[25] T. Bayraktar, S.B. Pidugu, Int. J. Heat Mass Trans. 49 (2006) 815.
[26] S. Waelchli, P.R. von Rohr, Int. J. Multiphas. Flow 32 (2006) 791.
[27] V. Hessel, S. Hardt, H. Löwe, Chemical Micro Process Engineering: Fundamen-
tals, Modelling and Reaction, Wiley-VCH, Weinheim, 2004.
[28] T.S. Zhao, Q.C. Bi, Int. J. Heat Mass Trans. 44 (2001) 2523.
[29] T.S. Zhao, Q.C. Bi, Int. J. Multiphas. Flow 27 (2001) 765.
[30] K.A. Triplett, S.M. Ghiaasiaan, S.I. Abdel-Khalik, D.L. Sadowski, Int. J. Multiphas.
Flow 25 (1999) 377.
[31] B.R. Fu, C. Pan, Int. J. Heat Mass Trans. 48 (2005) 4397.
[32] P.M.Y. Chung, M. Kawaji, Int. J. Multiphas. Flow 30 (2004) 735.
[33] L. Chen, Y.S. Tian, T.G. Karayiannis, Int. J. Heat Mass Trans. Transfer 49 (2006)
4220.
[34] J.W. Coleman, S. Garimella, Int. J. Heat Mass Trans. 42 (1999) 2869.
[35] K. Mishima, T. Hibiki, H. Nishihara, Int. J. Multiphas. Flow 19 (1993) 115.
[36] T. Postler, Z. Slouka, M. Svoboda, M. Pribyl, D. Snita, J. Colloid Interface Sci. 320
(2008) 321.
[37] W. Schrott, M. Pribyl, J. Stepanek, D. Snita, Microelectron. Eng. 85 (2008) 1100.
[38] M. Pribyl, D. Snita, P. Hasal, M. Marek, Chem. Eng. J. 101 (2004) 303.
[39] M. Pribyl, D. Snita, M. Marek, Chem. Eng. J. 105 (2005) 99.
[40] H. Pütter, in: H. Lund, O. Hammerich (Eds.), Organic Electrochemistry, fourth
ed., Marcel Dekker, New York, 2001 (chapter 1).
Acknowledgements
Financial support of this study by the European Commission
under Project No. 0111816-2 IMPULSE and by the Ministry of Edu-
cation, Youth and Sports of the Czech Republic under Project No.:
CEZ: MSM6046137301 is gratefully acknowledged.
[41] J. Krˇisˇt’ál, PhD Thesis: Study of gas-liquid flow in the thin channel, ICT Prague
and ICPF CAS, Prague 2008.
[42] T. Bystron, K. Bouzek, Z. Hasnik, J. Electrochem. Soc. 156 (2009) E179.
ˇ
[43] R. Kody´m, K. Bouzek, D. Snita, J. Thonstad, J. Appl. Electrochem. 37 (2007) 1303.