JOURNAL OF APPLIED ELECTROCHEMISTRY 28 (1998) 359368

Macroscopic analysis of polarization characteristics

of gas-diusion electrodes in contact
with liquid electrolytes
Part I: First order reactions*
K. SUNDMACHERz , U. HOFFMANN
Institut fur Chemische Verfahrenstechnik, Technische Universitat Clausthal, Leibnizstrasse 17, D-38678
ClausthalZellerfeld, Germany
Received 29 April 1997; revised 25 June 1997

In this paper principles of gasliquid chemical reaction engineering are applied to analyse the currentpotential characteristics of gas-diusion electrodes (GDE) in contact with liquid electrolytes. A
macroscopic electrode model is formulated which accounts for mass transfer in the external diusion
lms, in the gas layer and in the ooded layer. The set of model equations accounts for material
balances, mass transport kinetics and ButlerVolmer polarization kinetics. Several dimensionless
parameter groups are introduced which allow a compact reformulation of the proposed model. For
rst order reactions its solution can be derived analytically. The introduced parameter groups allow a
classication of the dierent operating modes of a GDE, that is, slow reaction, fast reaction and
instantaneous reaction.
Keywords: gas diusion electrodes, currentpotential behaviour, liquid electrolyte, macroscopic model, rst order reactions

List of symbols
a
A1
Az2
2
BiG
m
BiLm
cL1
cLe
1
c1
d
d GL=FL
d
D
GL=FL

Deff;1

G=K
D
eff;1
G=L
D
1
E
F
H1
Ha
*

specic surface area (m1 )

gaseous reactant
ionic reactant with charge number z2
gas side Biot number
liquid side Biot number
molar liquid phase concentration of A1
(mol m3 )
solubility of A1 in liquid electrolyte (mol m3 )
dimensionless concentration of A1
thickness of electrode (m)
thickness of gas layer and ooded layer respectively (m)
dimensionless ooded layer thickness
ratio of mass transport resistances (gas layer/
ooded layer)
eective diusion coecient of A1 in gas
layer or ooded layer (m2 s1 )
eective gas phase diusion coecient of A1
for molecular (G) or Knudsen (K) transport
gas or liquid phase molecular diusion coefcient of A1
enhancement factor
Faraday's constant (C mol1 )
Henry's coecient (Pa m3 mol1 )
Hatta number

f
Ha
in
i0
in
ilim

G=L

k1
n

N1
P
p1GB
R
Rlayer
T
U
U0
z
z

modied Hatta number (dependent on overpotential)

transfer current per unit of inner pore surface
area (A m2 )
exchange current density (A m2 )
dimensionless transfer current density
supercial limiting current density (A m2 )
gas side or liquid side mass transfer coecient of A1 (m s1 )
number of electrons transferred in electrode
reaction
mass ux density of A1 (mol m2 s1 )
total operating pressure (Pa)
partial pressure of A1 in gas bulk (Pa)
universial gas constant (J mol1 K1 )
dimensionless mass transport resistance
(layer GF, GL, FL, LF)
temperature (K)
electrode potential (V)
equilibrium potential related to bulk phase
conditions (V)
spacial coordinate (m)
dimensionless spacial coordinate

Greek letters
aa ; ac anodic, cathodic transfer coecient
dimensionless total overpotential
g

Presented in part at the Fourth European Symposium on Electrochemical Engineering, Prague, 2830 August 1996.
Author for correspondence, e-mail: [email protected].

0021-891X

1998 Chapman & Hall

359

360

m1

K. SUNDMACHER AND U. HOFFMANN

stoichiometric coecient of A1

GL
I
LB
LS
LF
FL

Superscripts and subscripts

GB
gas bulk
GS
gas side external surface of GDE
GF
external gas lm

1. Introduction
Gas-diusion electrodes (GDE) are key devices in
fuel cells and batteries. Recently they have also been
increasingly applied to synthesis, sensors and separation processes. In Table 1 some gas consuming
reactions of technical relevance are given in which
GDEs were used successfully. It can be seen that
GDEs have been applied for the production of
electrical energy, as energy saving counter electrodes, for the production of substances and for gas
purication. Depending on the application, GDEs
are in contact with liquid or solid electrolytes. The
present contribution is dedicated to GDEs operating
in gas/liquid processes. This is motivated by the facts
that GDEs are in contact with liquid electrolytes
either in a number of technically relevant production
processes, for example, oxygen reduction for hydrogen peroxide generation, or as a rst step in a
development towards a later solid electrolyte application, for example, hydrogen oxidation for PEM
fuel cells.
Figure 1 shows a schematic sketch of a single pore
of a GDE. There are three phases: the gas phase,
where the reactant A1 is present, the electrically
conductive solid phase, which is often loaded or
coated with catalytically active particles, and the
liquid phase electrolyte solution.
The most important physicochemical processes
taking place are as follows:

porous gas layer of GDE

gasliquid interface
liquid bulk
liquid side external surface of GDE
external liquid lm
ooded layer of GDE

(i) The reactant A1 is transported from the gas

bulk (GB) to the external surface of the GDE
via lm diusion.
(ii) The reactant diuses through the porous gas
layer (GL) to the gasliquid interphase.
(iii) The reactant A1 is dissolved at the gasliquid
interphase. In some cases solvent evaporation
can occur.
(iv) The reactant A1 is transported through the
ooded layer (FL) via liquid phase diusion.
(v) The ionic reactant Az2
2 is transferred from the
liquid bulk (LB) to the external GDE surface
by diusion and migration. Due to electroneutrality condition the counterions Az3
3 are
.
dragged along with the ions Az2
2
z3
(vi) The ions Az2
2 and A3 are transported by diffusion and migration through the ooded part
of the pore.
(vii) At the wetted pore walls there takes place the
electrochemical reaction A1 and Az2
2 accompanied by electron charge transfer. In addition
sorption processes and chemical reactions can
be involved. The rate of these chemical reactions can be accelerated by catalysts.
(viii) The electrons are transported in the solid matrix by electrical conduction to or from the
current collector.
The present analysis is restricted to one single reaction of (pseudo)rst order with respect to A1 . The

Table 1. Some examples of gas consuming reactions that can be carried out at gas-diusion electrodes
Arrows indicate desired directions of reactions
Anodic reactants !Cathodic reactants

Electrolytes

Processes

References

H2 (g) ! 2 H (aq) 2 e

solid polymers
acid solutions

[13]

CH3 OH(g) H2 O(g) ! CO2 1 6 H (aq) 6 e

solid polymers,
acid solutions
solid polymers,
acid, alkaline
solutions;

fuel cell,
H2-consuming
counter anode,
gas purication
vapour phase
DMFC
fuel cell,
O2-consuming
counter cathode,
gas purication
production of
hydrogen
peroxide
production of
partial oxidized
hydrocarbons
fuel cells,
gas purication

H2 O1

O2 (g) 4 H (aq) 4 e

H2 O2 1 OH (aq)

O2 (g) 2 H2 O(1) 2 e

alkaline
solutions

HCOH1 H2 O1

CO2 (g) 4 H (aq) 4e

acid solutions

2 Cl (aq)

Cl2 (g) 2 e

acid solutions

[4, 5]
[68]

[9]
[10]
[11, 12]

CHARACTERISTICS OF GAS DIFFUSION ELECTRODESI

361

Fig. 1. One-dimensional, single pore model for the electrochemical reaction of a gaseous reactant A1 with a liquid reactant AZ2
2 in a GDE.

inuence of the ionic reactant Az2

2 is considered in
Part II of this contribution [12]. The components in
the liquid electrolyte phase are assumed not to be
evaporated. In addition we assume that the ohmic
drops in the electrode matrix and the electrolyte solution are negligible.

The mass ux densities in the nonreactive layers are

constant. In the ooded layer the mass ux varies
according to the transfer current density per electrode
ain . Depending on the reaction direction, the stoichiometric coecient m1 is positive (anodic reactant) or
negative (cathodic reactant).

2. Macroscopic model for gasliquid reactions in GDE

2.2. Mass transport

Many models of dierent complexity have been

proposed to describe the operating behaviour of
GDEs. Some authors favoured a quasihomogeneous
(macroscopic) approach [1315], others a heterogeneous model [1620]. The latter often require many
adjustable parameters. Due to the limited accuracy of
experimental data, some of the model parameters
have little statistical signicance. Therefore it is reasonable to follow Newman [21] and apply a macroscopic model in terms of volume-averaged quantities.

The mass ux densities of the gaseous reactant, N1 , in

the four layers are expressed by means of following
linear transport laws:
Gas lm:
G
pGB p1GS
D
1
5
with k1G G
N1GF k1G 1
RT
d
Gas layer:

with

2.1. Material balances

In the four layers considered (i.e., gas lm, gas layer,

ooded layer and liquid lm) the following mass
balances of component A1 are valid:
Gas lm:
dN1GF
0
dz

) N1GF const.

) N1GL const.

Gas layer:
dN1GL
0
dz
Flooded layer:
dN1FL
ain
m1
with
dz
nF

N1FL z 0 N1GL

3a,b,c

N1FL z d FL N1LF

Liquid lm:
dN1LF
0
dz

N1GL

) N1LF const.

DGL
eff;1

GS
I
DGL
eff;1 p1 p1
FL
d d
RT

6a

1 p1GB =p
1
K
G

Deff;1
Deff;1

6b

Flooded layer:
N1FL DFL
eff;1
Liquid lm:

dcL1
dz

L
D
LB
N1LF k1L cLS
8
with k1L L1
1 c1
d
In the stagnant gas lm, the ooded layer and the
stagnant liquid lm the component A1 is assumed to
be transported via molecular diusion. In the nonooded gas layer of the porous electrode, besides
molecular diusion there can be a contribution of
Knudsen diusion. Both transport mechanisms can
be lumped together by addition of the corresponding
mass transport resistances, that is, the reciprocal difK
G
fusion coecients D
eff;1 and Deff;1 . The gas and liquid
G
L
lm thicknesses d and d have to be determined
from secured mass transfer correlations for the used

362

K. SUNDMACHER AND U. HOFFMANN

electrochemical cell. For this purpose a cell with a

well-dened ow eld at both external surfaces of the
GDE should be used. This can be achieved best in a
horizontal cyclone ow cell [22].
2.3. Reaction kinetics
The local transfer current density related to the inner
pore surface in is expressed by a ButlerVolmer type
polarization equation which is of rst order with
respect to A1 :
 

n ac F
U U0
in i0 exp
RT



ac F
cL1
9
U U0 Le GB
exp
RT
c1 p1
Equation 9 is valid for electrochemical reduction of
A1 , the corresponding oxidation can be formulated in
analogous manner. The exchange current density, i0 ,
and the open circuit potential, U0 , are related to the
liquid phase molar equilibrium concentration of A1
under conditions of the gas bulk phase (i.e.,
GB
cLe
1 p1 ). This quantity is related to the partial
pressure in the gas bulk by Henry's law:
GB
p1GB H1 cLe
1 p1

10

2.4. Introduction of dimensionless parameter groups

The introduction of the dimensionless coordinate z
and the dimensionless concentration c1 , that is,
z
z  FL
11
d
cL
c1  Le 1GB
12
c1 p1

into the model equations yields a set of dimensionless

numbers given in Table 2.
The most important of these numbers are the
Hatta number, Ha, and the ooded layer thickness,
d  , which appear in the dimensionless mass balance
for the ooded layer:
d2 c1
m1 d 2 Ha2 in
dz2

The local current density, in , can be expressed in

terms of the total overpotential, g , as follows:
in expfaa g g expfac g gc1

Denition

Dimensionless total overpotential

g 

Hatta number

14

2.5. Boundary conditions

For the formulation of the two boundary conditions
needed for the dierential Equation 13, the continuity of mass uxes between the layer interfaces, as
expressed in Equation 3(b) and (c), must be accounted for. After rearrangement the following
boundary conditions are obtained:

dc 
1 c z 0
 1 D 1d 
15a
1 
D
dz z 0

G 
d
Bim d


dc1 

BiLm d  c1 z 1 cLB
15b
1
dz z 1
The parameter group D in Equation 15(a) describes
the ratio of pore diusion velocities in the gas layer
and in the ooded layer (see Table 2). The resistances
in the gas and the liquid lms at the outer electrode
surface are considered by the Biot mass numbers BiG
m
and BiLm , respectively.

Table 2. Denition and interpretation of used dimensionless numbers

Number

13

U U0 p1GB ; cLB
2
RT =F
s
ai0 =nF
Ha  d
Le GB
DFL
eff;1 c1 p1
d FL
d

Interpretation

overpotential related to bulk conditions

25:7 mV

reaction rate under LB conditions

pore diffusion of A1 in FL

thickness of flooded layer

thickness of electrode

pore diffusion of A1 in FL
pore diffusion of A1 in GL

Dimensionless ooded layer thickness

d 

Ratio of mass transport resistances (GL/FL)

D 

Biot number for gas side

BiG
m 

k1GF d
DGL
eff;1

mass transfer of A1 in GF
pore diffusion of A1 in GL

Biot number for liquid side

BiLm 

k1LF d
DFL
eff;1

mass transfer of A1 in LF
pore diffusion of A1 in FL

Dimensionless liquid bulk concentration of A1

cLB

1

cLB
1
GB
cLe
1 p1

LB concentration of A1
equilibrium concentration of A1 at p1GB

Dimensionless local current density

in 

in
i0 p1GB ; cLB
2

local current density on pore surface

exchange current density

Dimensionless total current density

i 

m1 id GL
G
nF Deff;1 p1GB =RT

total current density

maximum limiting current density

Le GB
DFL
eff;1 c1 p1

GB
DGL
eff;1 p1 =RT

CHARACTERISTICS OF GAS DIFFUSION ELECTRODESI

3. Polarization characteristics

GB
DGL
eff;1 p1

16

d GL RT

It is reasonable to use this quantity for the denition

of a dimensionless current density, i , that has values
in the range 0 < i < 1:

i
L LB
i g ; Ha; n; ac ; d  ; D ; BiG
i 
m ; Bim ; c1
ilim;max
17
From Equation 17 it can be seen that the shape of the
current potential characteristics is determined by the
set of dimensionless numbers presented in Table 2
and, in addition, by the transfer coecient, ac , and
the number of electrons transferred, n.
In Table 3 typical ranges for the values of these
inuence parameters are given. From the values of
the Hatta number, Ha, it can be seen that slow, as
well as fast, reactions are carried out with GDEs. The
degree of ooding of a GDE can take values from
nearly d  0 to d  1 depending on the electrode
wettability. As expected, the values for D , that is, the
ratio of mass transfer resistances in the gas layer and
in the ooded layer, are very low according to the
eective diusion coecients in the two layers.
Moreover the solubility of gases in electrolyte solutions are low and of magnitude cLe 1 mol m3 . The
Biot numbers on the gas side, BiG
m , typically are one
decade below the liquid side numbers, BiLm , because
the gas lm is signicantly thicker than the corresponding lm on the liquid side.
The total current density at a GDE can be calculated from the dierence of the absorption rate of A1
from the gas bulk and the leakage rate of A1 to the
liquid bulk. Therefore, the dimensionless total current density, i , is related to the dierence in dimensionless concentration gradients at z 0 and at
z 1:
Table 3. Typical estimated values for the inuence parameters of the
polarization characteristics of a GDE for rst order reaction
Number
Dimensionless total
overpotential
Hatta number
Anodic and cathodic
transfer coecient
Dimensionless ooded
layer thickness
Ratio of mass transport
resistances (GL/FL)
Biot number for gas side
Biot number for liquid side
Dimensionless liquid bulk
concentration of A1





1 d
dc1 
dc1 
i D


d
dz z 0 dz z 1


The maximum limiting current ilim;max is achieved if

only the gas layer resistance dominates the electrochemical conversion. This is calculated from
ilim;max nF

363

Symbol


Typical range
(min, max)
2

)10 , 10

Ha
aa ; ac

100, 104
0.1, 2.0

d

10)4, 100

D

10)7, 10)3

BiG
m
BiLm
cLB
1

100, 102
101, 103
0.0, 1.0

18

3.1. General analytical solution

The combination of the mass balance, Equation 13,
with the polarization Equation 14 yields a linear
dierential equation for the concentration prole
c1 z . Its solution [23] is:


f z
yz 1 sinh Ha
 
yz
f
sinh Ha


f z
yz 0 sinh Ha1
 
19

f
sinh Ha
with the transformed concentration
yz c1 z 1 expng

20

The shape of the concentration prole is inuenced

by the modied Hatta number


f  d  Ha exp ac g
Ha
21
2
This important parameter group is a combination of
the ooded layer thickness, d  , the Hatta number, Ha,
and the applied total overpotential, g .
From Equation 19 the absorption rate of A1 (i.e.,
the concentration gradient at z 0) and the leakage
rate, (i.e., the concentration gradient at z 1) are:



dc1 

E c1 z 0 expng
dz z 0

c1 z 1 expng
22a

f
cosh Ha

  
dc 
c z 0 expng
1 
E 1
f
dz z 1
cosh Ha



 
22b
c1 z 1 expng
In both equations the enhancement factor E appears.
This describes the acceleration of gas absorption by
electrochemical reaction. For a rst order reaction
this is given by
f
Ha
E
23
f
tanh Ha
The enhancement factor, E, is inuenced by the
overpotential, g , which enters the modied Hatta
number in Equation 21. Therefore, the operator can
adjust the operating regime of a GDE by changing the
overpotential. With Equations 22(a) and (b) the total
current density, i , (Equation 18) can be determined:

h
1 d
1
E
1

i D
c1 z 0
f
d
cosh Ha
i
c1 z 1 2 expng
24
The unknown concentrations c1 z 0 and
c1 z 0 are derived from the boundary conditions,

364

Equations 15(a) and (b). These are given by the following expressions:
c1 z 0 expng
1 expng

 

1
RFL RL
1 RFL
e
cosh Ha
25

L
G
GL
RL RFL RG RGL RFL R R R 2
cosh Hea
c1 z 1 expng

cLB
expng
1




 1
RFL RG RGL
1 RFL
cosh Hea

26
L
G
GL
RL RFL RG RGL RFL R R R 2
cosh Hea
Four dimensionless resistances, which dominate the
performance of a GDE, are contained in Equations 25 and 26:
D 1
27
Resistance of gas film:
RGF   G
d Bim
D
Resistance of gas layer:
RGL   1 d  28
d
1
Resistance of flooded layer: RFL 
29
E
1 1
30
Resistance of liquid film:
RLF   L
d Bim
In the following we discuss three operating modes:
slow reaction, fast reaction and instantaneous reaction. In each cases, we assume no reactant A1 to be
present in the liquid bulk (i.e., cLB
0). This cor1
responds to the desired operation under technical
conditions.
3.2. Operating modes
3.2.1 Slow reaction. For modied Hatta numbers
f < 0:3 we have a slow electrochemical reaction
Ha
compared to the mass transport in the ooded layer.
For this case Fig. 2 illustrates the concentration
prole of A1 in a GDE.
For slow reaction the enhancement factor, E, can
be approximated as follows:

K. SUNDMACHER AND U. HOFFMANN

f
f2
1
Ha
Ha
E
31
) RFL 
1
f 2 =2
tanh Ha
2
1 Ha
Obviously the enhancement factor, E, has values
slightly above unity. From Equation 29 it can be seen
that the resistance of the ooded layer, RFL , also has
values near unity. If the gas sided transport resistances RG and RGL with RFL are compared, for slow
reactions we can conclude:
RG RGL  RFL O1


D 1


1
) 
d BiG
m

32

For most GDEs this condition is met, because D has

quite low values (see Table 3). With the help of
Equation 32 the general solution, Equations 24, 25
and 26, can be simplied to formulate an approximated solution for slow reactions:
f2
1 d  Ha
i D
d
2


L
f
R 4 Ha2 2 11 expng 2 expng

f 2 2
RL 2 Ha
33
 a 
c
f  d  Ha exp g < 0:3
for Ha
2
To analyse the dependence of current, i , on overpotential, g , we consider two limiting cases concerning RL :
As RL ! 0 :
i  D
As RL ! 1 :

f2
1 d  Ha
1 2 expng 34a
d
2

f2
f 2 4 Ha
1 d  Ha
1 expng
f2
d
2 2 Ha
1 d f 2
34b
Ha 1 expng
 D
d

i  D

In both cases the current density is proportional to

the square of the modied Hatta number. The terms
in brackets in Equations and 34(a) and (b) account

f < 0:3.
Fig. 2. Concentration prole of reactant A1 in GDE for a slow electrochemical reaction Ha

CHARACTERISTICS OF GAS DIFFUSION ELECTRODESI

for the backward reaction. These terms are of importance close to the electrochemical equilibrium
(i.e., for g ! 0). Neglecting this, in the limiting case
RL ! 0, the current density takes half the value observed in the limiting case RL ! 1. This is due to the
leakage eect of the reactant to the liquid bulk.
From Equations 34 we can express the current
density, i (in dimensions), in terms of the electrode
potential, U:


F
35
U U0
i  d FL aio exp ac
RT
It is evident that, for slow reactions, the current depends on the electrochemically active surface area,
that is the product d FL a and the exchange current
density i0 . The dependence on electrode potential
corresponds to the polarization kinetics, Equation 9.
3.2.2. Fast reaction. By increasing the overpotential
jg j we achieve a transition from a slow to a fast
reaction. The concentration prole in Fig. 3 illustrates the fast reaction. The liquid lm resistance RL
has no relevance, because the reactant A1 is completely consumed inside the ooded layer and no
leakage occurs. A fast reaction is characterized by
f > 3 and in this case the enhanceHatta numbers Ha
ment factor can be approximated as follows:
E

f
Ha
f ) RFL  1
 Ha
f
f
Ha
tanh Ha

36

Now, the E values are signicantly above unity, because the gas absorption is signicantly enhanced by
the electrochemical reaction. Therefore, the resistance
of the ooded RFL is dramatically reduced by the
reaction. Starting from the general solution, Equations 24, 25 and 26, the total current density of fast
reactions is approximated by
i D

1
BiG
m

1 d
1

G
GL
f
d
R R 1=Ha

1 d
1 d  D Ha exp1 ac g
2

37

365

for



f  d  Ha exp ac g > 3
Ha
2

and

expng ! 0

For moderately fast reactions, the resistances on

the gas side RG and RGL are negligible in comparison
to the ooded layer resistance RFL . For these reactions the dimensionless current density is given by
 a 
1 d f
c


1

d
Ha
exp
g
Ha

D
i D
d
2
38
for
G

R R

GL

R

FL



D 1
1

) 
1 d 
f
d BiG
Ha
m

Based on Equation 38, the following dependence of

the current density on the system parameters ensues:


q
ac F
FL
39
U U0
i  Deff;1 ai0 exp
2 RT
According to Equation 39 for moderately fast reactions the current is inuenced by the geometric average of two rate constants: that is, the exchange
current density, i0 , and the eective diusion coecient in the ooded layer, DFL
eff;1 . The dependence on
the electrode potential is less distinct than for slow
reactions. This is due to the halving of the transfer
coecient ac in Equation 39. The ooded layer
thickness, aFL , has no inuence according to the fact
that even for moderately fast reactions, the component A1 is completely depleted inside the ooded
layer.
3.2.3. Instantaneous reaction. For very high modied
f there is a transition from a fast
Hatta numbers Ha
reaction to an instantaneous reaction. In this situation the reactant is totally consumed at the gasliquid
interface and its concentration there drops to zero.
This is illustrated in Fig. 4.
The ooded layer resistance RFL has no importance, because the gas component A1 no longer enters
the ooded layer. The resulting total current density,

f > 3.
Fig. 3. Concentration prole of reactant A1 in GDE for a fast electrochemical reaction Ha

366

K. SUNDMACHER AND U. HOFFMANN

f ! 1.
Fig. 4. Concentration prole of reactant A1 in GDE for an electrochemical instantaneous reaction Ha

i achieves its maximum value, the limiting current

density, ilim
1 d
1
1 d
i ilim D

40
d  RG RGL Bi1G 1 d 
m

for
G

R R

GL

R

FL



D 1
1

) 
1 d 
G
f
d Bim
Ha

The maximum limiting current density is characterized by negligible gas lm resistance, RG , in comparison to the resistance that is exerted by the gas
layer, RGL :
i ilim; max D
for
RG  RGL ;

1 d 1
1
d  RGL

RGL  RFL ) BiG

m 
D
1
1 d  
f
d
Ha

1
;
1 d

41

4. Discussion and conclusions

Figure 5 shows a simulated dimensionless polarization curve of a GDE for a typical set of parameters.
Increasing the overpotential, the reaction rate is more
and more accelerated and we have a transition from a
slow reaction to fast reaction and nally to an instantaneous reaction.
The kinetic parameters of the electrochemical reaction can be determined from a plot of the logarithm
of the ratio i =1 d  against the applied overpotential. This term does not depend on the unknown
ooded layer thickness, d FL , and therefore can be
determined directly from experimental data. From
Fig. 6 a linear dependence in the region of moderately
fast reaction is evident. This is in agreement with
Equation 38 which can be expressed in logarithmic
form:
ln

i
ac
lnD Ha g
1 d
2

Fig. 5. Simulated polarization curve of a GDE for an electrochemical reduction of rst order: v1 A1 (g) ne . . . ! products.

42

CHARACTERISTICS OF GAS DIFFUSION ELECTRODESI

367

Fig. 6. Determination of cathodic transferr coecient, ac and Hatta number, Ha, from the currentpotential characteristics of a GDE for a
rst order reaction: m1 A1 (g) ne . . . ! products.

Thus, the transfer coecient, ac , can be determined

from the slope of the curve and the intersection of the
dashed approximation line with the ordinate yields the
group D Ha. The parameter D can be estimated from
component properties and therefore we obtain the
Hatta number Ha. From the Hatta number the volume
based exchange current density aio can be extracted.
To determine the essential mass transport parameter, the gas side pore diusion coecient, DGL
eff;1 ,
measurements in the mode of instantaneous reaction
must be carried out. At high overpotentials we obtain
the limiting current ilim . According to Equation 40
this depends on the gas Biot mass number, BiG
m , and
on the ooded layer thickness, d  . If d  is known, it is
possible to calculate the Biot number and from this

the pore diusion coecient, DGL

eff;1 , of the reactant
A1 . This requires an experimental cell where a well
dened gas lm mass transfer coecient, k1GF can be
adjusted. The unknown parameter d  can be found
from measurements in the region of slow reaction,
that is, at low overpotentials. At higher overpotentials the modied current density, i =1 d  , is independent on d  , and in the region of instantaneous
reaction this parameter group shows only a weak
dependence on d  . In the case that the degree of
ooding cannot be determined exactly, it is advised to
start with d  ! 0. This is reasonable because GDEs
are very often weakly wetted by the electrolyte solution. In many cases this is a precondition to run
successfully GDE-based electrochemical reactors in

Fig. 7. Simulated concentrations at the gas side z 0 and at the liquid side z 1 of the ooded layer for a rst order reaction:
m1 A1 (g) ne . . . ! products.

368

K. SUNDMACHER AND U. HOFFMANN

order to avoid ooding of the gas compartment. The

liquid side Biot number, BiLm , is only of importance
for slow reactions. However, in practical applications
mostly fast or instantaneous reactions are used, so
that there is nearly no signicant inuence of BiLm .
Surprisingly at low overpotentials there can be
c1 z 0 > 1 (see Fig. 7). For the interpretation of
this phenomenon one has to keep in mind that the
cathodic reaction depends on the overpotential, g ,
and the concentration, c1 , whereas the anodic reaction only depends on the overpotential. Due to the
fact that no A1 is assumed to be present in the liquid
bulk, the anodic backward reaction becomes dominant at low overpotentials and forms A1 . This material leaves the GDE preferably at the gas side.
References
[1]
[2]
[3]
[4]
[5]

R. M. Q. Mello and E. A. Ticianelli, Electrochim. Acta 42

(1997) 1031.
J. J. T. T. Vermeijlen and L. J. J. Janssen, J. Appl. Electrochem. 23 (1993) 1237.
K. Petrov, I. Nikolov and T. Vitanov, Int. J. Hydrogen
Energy 9 (1984) 901.
A. S. Arico, P. Creti, H. Kim, R. Mantegna, N. Giordano
and V. Antonucci, J. Electrochem. Soc. 143 (1996) 3950.
K. Scott, J. Cruickshank and P. C. Christensen, 4th European Symposium on Electrochemical Engineering,
Symposium Proc. (Ed.: I. Rous ar) UTAX, Prague,
(1996), p. 39.

[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]

A. S. Arico, A. K. Shukla, V. Antonucci and N. Giordano,

J. Power Sources 50 (1994) 177.
E. Brillas, A. Maestro, M. Moratalla and J. Casado, J. Appl.
Electrochem. 27 (1997) 83.
G .R. Dieckmann and S. H. Langer, ibid. 27 (1997) 1.
E. Hillrichs and K. Lohrberg, Dechema-Monographien 125,
VCH, Weinheim (1992), p. 221.
A. Bandi, C. U. Maier, J. Schwarz, M. Specht and W.H.
Bloss, Dechema-Monographien 128, VCH, Weinheim
(1993), p. 563.
E. B. Anderson, E. J. Taylor, G. Wilemski and A. Gelb, J.
Power Sources 47 (1994) 321.
K. Sundmacher and U. Homann, J. Appl. Electrochem.
(1998) in preparation.
D. M. Bernardi and M. W. Verbrugge, AIChE. J. 37 (1991)
1151.
Idem, J. Electrochem. Soc. 139 (1992) 2477.
T. E. Springer, M. S. Wilson and S. Gottesfeld, ibid. 140
(1992) 3513.
S. J. Ridge, R. E. White, Y. Tsou, R. N. Beaver and G. A.
Eisman, ibid. 136 (1989) 1902.
P. Bjornbom, Electrochim. Acta 32 (1987) 115.
R. P. Iszkowski and M. B. Ncutlip, J. Electrochem. Soc. 127
(1980) 1433.
K. Broka and P. Ekdunge, J. Appl. Electrochem. 27 (1997)
281.
M. A. Al-Saleh, S. Gultekin, S. Rahman and A. Al-Zakri,
J. Power Sources 55 (1995) 33.
J. S. Newman, `Electrochemical Systems', 2nd edn, Prentice
Hall, Englewood Clis, NJ (1991).
K. Sundmacher and U. Homann, J. Appl. Electrochem.
(1998) in preparation.
K. R. Westerterp, W. P. M. van Swaaij and A. A. C. M.
Beenackers, `Chemical Reactor Design and Operation',
Wiley, Chichester, UK (1990).