See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/243981090

The influence of bubble shape and the thickness of the wetting film on the
incremental electrical resistance caused by the presence of a single bubble in
Hall-Héroult cells

Article in Journal of Applied Electrochemistry · November 2006

DOI: 10.1007/s10800-006-9220-1

CITATIONS READS

4 153

3 authors, including:

Laszlo Istvan Kiss Sandor Poncsak

Université du Québec à Chicoutimi Université du Québec à Chicoutimi
169 PUBLICATIONS 1,398 CITATIONS 80 PUBLICATIONS 1,943 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Laszlo Istvan Kiss on 12 May 2014.

The user has requested enhancement of the downloaded file.

Journal of Applied Electrochemistry (2006)  Springer 2006
DOI 10.1007/s10800-006-9220-1

The influence of bubble shape and the thickness of the wetting film on the incremental
electrical resistance caused by the presence of a single bubble in Hall-Héroult cells

A.L. PERRON, L.I. KISS* and S. PONCSÁK

De´partement des sciences applique´es, Universite´ du Que´bec à Chicoutimi, Quebec, Canada G7H 2B1
(*author for correspondence, tel.:+1-418-545-5011, fax:+1-418-545-5012, e-mail: [email protected])

Received 21 February 2006; accepted in revised form 15 August 2006

Key words: alumina reduction cell, bubble shape, gas bubbles, interelectrode resistance

Abstract

Bubbles play an important role in the transport phenomena existing in an electrolysis cell. They increase the total
ohmic resistance of the electrolyte but their contribution is still not well quantified. During their movement under
the anode, the bubbles are separated from the solid by the so-called wetting film, that is by a thin liquid layer. In
order to develop a mathematical model to compute the increment of the electrical resistance of the electrolyte due to
the presence of several bubbles under the anode, the effects of the bubble shape and the thickness of the wetting film
for a single bubble must be quantified a priori. In this first paper, these effects are computed using the finite element
method (FEM). The results have shown that the influence of the bubble shape and that of the wetting film is small,
about 5% and 2%, respectively.

List of symbols Greek symbols

Aa anode area (m ) 2
j electrical conductivity (X)1 m)1)
Ap projected area (m2) H covering factor
A* cross-section of an electrical current tube (m2) q density (kg m3)
b bubble front width (m) r surface tension (N m)
d bubble diameter (m) t kinematic viscosity (m2 s)1)
deq equivalent diameter (m)
~
E electrical field (V/m) Subscripts
g gravitational acceleration (m s)2) 1 anode
h bubble height (m) 2 electrolyte
i local current density (A m)2) A upper limit of the computational domain
~
i current density vector (A m)2) (within the anode)
I current (A) B cathode surface (bath-molten aluminium
LC length of the elongated bubble (m) interface)
r geometrical ratio for large bubbles l liquid phase
R0 electrical resistance of the bubble-free electrolyte n index of an element
(X) n normal component
RT total electrical resistance of the electrolyte with N number of element
bubbles present (X) t tangential component
uT bubble terminal velocity (m s)1)
V voltage (V) Superscript
V* bubble volume (m3)
* electrical current tube

1. Introduction (Na3AlF6) at about 960
C [1]. In a typical cell, the

anode and the cathode are arranged horizontally. The
In the Hall-Héroult process, liquid aluminium is pro- primary cell reaction is
duced by the electrolytic reduction of alumina (Al2O3)
dissolved in an electrolyte of essentially cryolite 2=3Al2 O3 þ C ¼ 4=3Al þ CO2 : ð1Þ
 The current which passes through the cell is the in of accumulated gas. Aaberg et al. [6] investigated the
order of 100 kA. The consumption of the anode is about characteristics of the bubble layer under the anode in a
1.5 cm per day. The prebaked anodes have to be small scale electrolysis cell producing aluminum. The
replaced usually after 22–26 days. The liquid aluminium diameter of the carbon anode was less than 10 cm. By
is deposited at the cathode and bubbles, principally measuring simultaneously the volume of the accumu-
composed of CO2, are generated under the anode lated gas and the electrical resistance of the electrolyte,
surface. When the bubbles reach a critical size which they deduced the covering factor and the thickness of
depends principally on the anode local slope, the the bubble layer. The expressions to compute the latter
electrolyte velocity field as well as the retention forces, quantities were based on the work of Hyde and Welch
they start to move, escape from the anode sides. It is [5] which considered the bubbles as equivalent cylinders
believed that when a bubble reaches a certain velocity, with a diameter equal to the depth of the bubble layer.
there is a dynamical formation of a thin wetting film The volume of gas was obtained by measuring the rise of
separating the bubble and the anode surface. The the electrolyte level, as done by Solheim and Thonstad
formation of the film has been observed experimentally [7]. As pointed out by Quian et al. [8], this measurement
in an air–water–plexiglas system [2]. The direct obser- method may introduce an error since the bubbles
vation in a real industrial cell is impossible, owing to its outside the interpolar region are included in the total
small thickness and the opacity as well as the corrosivity accumulated gas. Nevertheless the typical covering
of the cryolite. The presence of the bubbles (stationary factor measured was about 45%. The thickness of the
and moving) under the anode contributes to increase the bubble layer varied from 4 to 6 mm. The average value
ohmic resistance of the electrolyte. According to Haupin of the accumulated gas under the anode was 0.445 cm3/
[3], the extra ohmic drop is in the range of 0.15–0.35 V. cm2. Haupin [3] measured the voltage gradient in the
Since the pioneering work of Dernedde and Cambridge interpolar space in an industrial reduction cell. He
[4] about three decades ago, a lot of research has been measured a thickness of 5 mm for the bubble layer. He
devoted to the role of bubbles in the Hall-Héroult found that occasionally, there were contacts between the
process. The contribution of the bubbles to the induced probe and gas bubbles as far as 2 cm from the anode
bath flow, to the cell thermal balance as well as to the surface.
total resistance of the electrolyte is still not well under- Zoric and Solheim [9] were the first to include the
stood. The knowledge of the contribution of the bubbles effect of the perturbation of the electrical field in the
to the overvoltage of the cell is an important parameter to entire ACD caused by the presence of bubbles in their
estimate the value of the anode–cathode distance ACD calculation of the electrical resistance. They studied the
under normal electrolysis conditions. In order to elabo- anodic and cathodic current distribution perturbed by
rate a mathematical model to compute the total resis- the presence of bubbles. The bubbles were represented
tance of the electrolyte, the effect of bubble shape and the by large discrete slabs 5 mm thick. They also modeled
thickness of the wetting film must be quantified a priori. large bubbles with a thicker (deeper) front (1–1.5 cm).
The term total resistance RT refers to the electrical The length of the bubble varied from 24 to 75 mm. In all
resistance of the electrolyte in the entire interelectrode cases, the anodic current density reached local maxima
spacing when bubbles are present. In this work, the close to the bubble. The local minima of the cathodic
increase of the electrical resistance caused by the presence current density were not significantly influenced by the
of the bubbles is expressed in a relative manner by bubble position because of the long distance between the
dividing the total electrical resistance RT by the resistance bubble and the cathode. They also presented some
R0 of the bubble-free electrolyte. correlations for the incremental voltage drop applicable
Hyde and Welch [5] studied the influence of accumu- in certain conditions.
lated gas under the anode, the ACD and the bubble The aim of this work is to compute the influence of the
shape on the bubble resistance. The bubbles were bubble shape and the thickness of the wetting film on the
simulated by ceramic objects of known volume and relative resistance caused by the presence of single
shape. The tested shapes were discs and spheres. The bubbles under the anode.
objects were inserted in a laboratory electrolysis cell
producing lead. The electrical resistance of the electro-
lyte with and without bubbles was obtained by measur- 2. The mathematical model
ing the voltage drop of the cell with a high sampling rate
when the current was suddenly stopped. This procedure To calculate the relative resistance RT/R0 induced by the
allowed the separation of the voltage components of the presence of the bubbles underneath the anode, the
laboratory cell. The results showed that the resistance Laplace equation
increase caused by the presence of the bubbles (solid r2 V ¼ 0 ð2Þ
objects) primarily depends on the bubble volume, and it
increases linearly with the accumulated gas (ceramic is solved within the domain with the cross-section shown
objects). The effect of bubble shape on the resistance in Figure 1. In order to analyze the influence of all the
increase has been found to be small for the same amount different bubble shapes used in the present paper, the
 @V
¼ 0: ð8Þ
@n
where n is the normal to the domain limits or to the
bubble interface. The boundary-value problem was
solved using the finite element method. To study the
effect of bubble shape, a mesh with about 60,000–
100,000 tetrahedral elements was generally used. The
characteristic length of the elements was less than 3 mm.
A mesh control was used to compute the effect of the
thin wetting film on the relative resistance. The thickness
of the elements in the y direction within the wetting film
was less than 0.1 mm.
Several calculations of the relative resistance were also
carried out with an equipotential electroactive area of
the anode, similarly as done by Vogt and Kleinschrodt
[10]. In other words, the distance between the upper
Fig. 1. Cross-section of the 3D model. equipotential plane and the cathode surface was reduced
to 5 cm. The relative resistances calculated within the
two different computational domains were very similar.
domain is three-dimensional. The computational This may be explained by the high ratio of resistivity of
domain is multiply connected: there is an external the electrolyte on the anode. In this situation, there were
boundary and an internal one around the bubble. The also current density maxima along the bubble contour.
conservation of the electric charge requires that the The total current provoked by a constant voltage
normal component of the current density vector is (boundary condition of first kind) through the domain
continuous at the anode–electrolyte interface: where a bubble is present is given by
Z XN
i1n ¼ i2n : ð3Þ
IT ¼ idA ¼ in dAn ð9Þ
n¼1
A
Using the differential form of the Ohm’s law ( ~i ¼ j~
E)
this becomes where A is the total cathode surface, N is the number of
elements on the cathode surface and, in and dAn are the
E1n j2 local current density and the area of the nth cathode
¼ : ð4Þ
E2n j1 element, respectively. The relative bubble resistance is
The other part of the coupling conditions comes from evidently independent of the order or type of the
the equality of the electric field strength on the two sides boundary conditions. Then



of the interface. Thus the tangential components of ~ E RT VT

I0


¼ ¼ ð10Þ
are equal on the two sides: R0 V0
I¼const IT
V¼const
E1t ¼ E2t : ð5Þ where RT and R0 are the electrical resistance of the total
The anode–cathode distance ACD is assumed to be electrolyte and the bubble-free electrolyte, respectively.
5 cm. The distance between the upper equipotential It is clear that this model neglects some phenomena. The
plane (within the anode) and the cathode surface is fixed model computes only the primary current distribution.
at 10 cm. The resistivity of the electrolyte and the anode The addition of the anodic activation overvoltage would
were taken to be 0.5 Xcm and 0.005 Xcm, respectively tend to homogenize the current distribution close to the
[9]. The Laplace equation is an elliptic PDE type bubble. It assumes a non-deformable cathode surface
therefore the boundary conditions on the complete and the bubble shapes are also well defined. It neglects
contour enclosing the geometry must be specified to the effect of convection on the transport of sodium
obtain a solution. The boundary conditions are simply charge carriers. In other words, in the solution domain
those of the first kind the conductivity of the bubble-free electrolyte is assumed
to be homogeneous and isotropic. In this work, the
Vðy ¼ 0Þ ¼ VB ; ð6Þ results are presented in a relative form which emphasizes
the role of the bubble on the total electrical resistance.

Vðy ¼ LÞ ¼ VA : ð7Þ
3. Results and discussion
Furthermore, the vertical limits on the left and right
side of the domain and the bubble interface were In this section, the results concerning the difference of
considered insulated the relative bubble resistance caused by four different
bubble shapes of the same volume are presented first. In
the second part, the effect of the thickness of the wetting
film on the electrical resistance is discussed. The liquid
film is located between the anode and the moving
bubble. It separates the anode surface and the moving
bubbles.

3.1. The influence of bubble shape

The dynamic shape of a bubble depends of the forces

which act on it. In a gravity driven flow, Kiss et al. [2]
and Perron et al. [11] have shown that there may exist
Fig. 2. Bubble shapes used in this work: (a) lateral view, (b) top
two main classes of bubbles moving under a solid view.
surface: the ‘‘creeping bubble’’ and the ‘‘bubble on the
wetting film’’. In the former, the terminal velocity is
controlled by the surface tension, the buoyancy and The four basic bubble shapes are presented in
the viscous forces. However, in an industrial cell, the Figure 2. The volume of the bubble A is given by
buoyancy acting on the bubble is not the only driving
pd2A
force, i.e. there is a drag force generated by the flow V
A ¼ hA ð12Þ
around the bubble. The distance traveled by the 4
creeping bubble in a real cell is very small. Conse- where hA is the depth of the bubble and its value is
quently in this work, the bubble shape observed in the 0.5 cm. The bubble shape A is not an observed one. It is
creeping regime is not studied. For the bubble on the used only as a simplification to ease computational
wetting film, the inertia may play the dominant role in work, since its geometry is simple. Henceforth, the
controlling the movement. However, a recent study bubble shape A is called circular disc. Although the
[12] showed that the viscous and surface tension forces shape B has been observed experimentally in water at
play also a non-negligible role at both low bubble both low bubble volumes and low inclination angles, the
volumes and low inclination angles of the solid principal reason to study this shape is the fact that the
surface. bubble simulator developed by Kiss et al. [15–17] uses
The bubble shapes studied in this paper are based on this bubble shape. The volume of the bubble B is
shapes observed in the above mentioned study [12], on  2 
d hB dB h2B
other visual observations and on the well-known work V
B ¼ ph B  þ : ð13Þ
of Fortin et al. [13]. All the bubble shapes reported 4 4 6
here including the Fortin’s shape are observed in an
air–water two-phase flow system. A recent study [14] The depth of this bubble is also 0.5 cm. From now,
has shown that the bubble motion under an inclined the bubble shape B is called rounded disc. For larger
surface may be described by three independent dimen- bubble volumes and/or for higher bubble velocities, the
sionless parameters such as the Bond and Morton bubble becomes more deformable and the depth of the
numbers, the inclination angle and by one dependent bubble may be greater than 0.5 cm. This deformation is
parameter, the Reynolds number. If the effects of the caused by a perturbation moving with the same velocity
physical properties of the dispersed phase are ne- as the bubble and it is called a hydraulic jump. The
glected, the Morton, Bond and Reynolds numbers are
expressed as
gt4l q3l ql d2 g uT d
Mo ¼ Bo ¼ Re ¼ ð11Þ
r3 r tl
where ql is the liquid density, tl is the liquid kinematic
viscosity, g is the gravitational acceleration, r is the
surface tension, uT is the bubble terminal velocity and
d is the bubble equivalent diameter defined by
d ¼ ð6V
=pÞ1=3 , where V
is the bubble volume. Thus,
the Morton number depends strictly on the physical
properties of the working liquid. The values of the
Morton number for the water at 20
C and cryolite
under normal operation conditions are 2.59  10)11
and 9.46  10)11, respectively. Therefore, one assumes
that the observed bubble shapes in water may exist in
the cryolite under normal operation conditions. Fig. 3. Definition of h used to describe the bubble shape C.
Fig. 4. Bubble shape D (Fortin shape): (a) lateral view, (b) bottom view.

phenomenon has been first observed (indirectly) by 2

3

RT
RT

Haupin [3]. The shapes C and D represent this situation.
 R0
SB 7
6 R0
The former has been observed by Perron et al. [12] for difference ¼ 4 CD

5
100
RT
bubble volumes ranging from about 5 to 15 cm3. The R0
CD
length of the bubble C in the direction parallel to the  
RT jCD RT jSB
movement, lC, was nearly constant at 1.5 cm, indepen- ¼
100 ð16Þ
RT jCD
dent of the bubble volume in this bubble volume range.
The dimension of the bubble C perpendicular to the
movement increases with the volume and its depth is
considered constant at 1 cm. The volume of the bubble
C is given by
 

l2C LC ðh  sin hÞ
VC ¼ p ð14Þ
2 2
where lC = 1.5 cm and h = 2.46 rad is defined in
Figure 3 below. Henceforth, the bubble shape C is called
truncated cylinder. As the bubble volume increases, the
length of the bubble in the direction of the movement
also increases and the well-known shape of Fortin [13] is
obtained. The depth of the front part reaches 1 cm while
the depth of the rear part stays at 0.5 cm. The
geometrical ratio r = b/dD is defined in Figure 4. Three
different values of the ratio r have been used in the Fig. 5. Difference in the electrical resistance for the rounded disc
(shape B) and the truncated cylinder (shape C) bubbles as function of
calculations such as 1/4, 1/3 and 1/2. The volume of the the bubble volume for the intermediate bubble sizes. The base of
bubble with shape D is given by comparison is the circular disc shape (A).
 
d2 hD 1  
V
D ¼ D p þ cos1 ð1  2rÞ  sin 2 cos1 ð1  2rÞ
8 2
ð15Þ
where hD = 1 cm.
In the following, the influence of the bubble shape
on the electrical resistance is presented for all the
shapes introduced in the Figure 2. The bubble volumes
are divided in two groups: the intermediate bubbles
with a volume less than 20 cm3 and the larger ones
called macro-bubbles with a volume ranging from 20
to 80 cm3. The former class includes the rounded disc
and the truncated cylinder shapes while the latter class
includes only the Fortin shape. All the results are
presented in a relative form using the circular disc
(shape A) as reference case. For instance, the differ- Fig. 6. Difference in the electrical resistance for the Fortin shape
ence presented on the y-axis of the Figures 5, 6 is bubble (shape D) as function of the bubble volume for the macro-
defined as bubble sizes. The base of comparison is the circular disc shape (A).
where CD is the Circular Disc bubble and SB is the given bubble volume, the resistance difference value
Studied Bubble shape such as the rounded disc, the between the circular disc and Fortin shapes is positive
truncated cylinder or the Fortin shape. and it increases as the ratio increases. These two
Figure 5 shows the influence of the bubble shape on remarks show the importance of the screening effect
the electrical resistance for the intermediate bubble sizes. compared to the current line perturbation since as the
The discrepancy between the circular disc bubble and geometrical ratio increases, the covering factor de-
the rounded disc (lower curve) was calculated at a low creases. In other words, the total electrical resistance
value of the covering factor to ensure that there is no RT or the relative resistance RT/R0 due to the presence
interaction between the bubble and the limits of the of a Fortin shape bubble decreases as the geometrical
calculation domain. The covering factor is defined as ratio increases. Furthermore, for the three curves
presented in Figure 6, the difference increases slightly
Ap with the bubble volume. It is also a very interesting that
H¼ ð17Þ at the same value of the covering factor the large
Aa
bubbles provoke a higher electrical resistance than the
where Ap is the projected bubble area on the anode smaller ones. This remark reflects directly the contribu-
surface Aa. The difference between the circular disc tion of the current line deformation to the electrical
bubble and the truncated cylinder bubble (upper curve) resistance caused by the presence of the bubble under
was computed on an identical surface area of the anode. In other words, one can say simply that a
40  40 cm2. The difference increases with the bubble large bubble deforms the electrical field more than a
volume for the truncated cylinder bubble while it smaller one even at identical values of the covering
decreases for the rounded disc bubble. For the former factor. The consequence of the latter conclusion may be
(truncated cylinder vs circular disc), the increase in very important in terms of the influence of the morpho-
difference may not be attributed to the effect of the logical structure of the two-phase layer on the total
covering factor since in both cases, it increases propor- electrical resistance due to the presence of a large
tionally with the bubble volume. The increase of the number of bubbles under the anode.
difference may be explained by the fact that the circular To summarize, the preceding results show that the use
disc provokes a stronger, three-dimensional perturba- of the circular disc or the rounded disc bubble instead of
tion of the electrical field during the increase of its the Fortin bubble in a mathematical model overesti-
volume than the truncated cylinder. For the latter, the mates the total electrical resistance by less than 2% for
perturbation pattern of the equipotential and current the macro-bubble sizes at low covering factors.
lines stays the same as the increase of the volume is The last paragraphs were useful to understand the role
realized by a linear elongation. The difference in the of the two contributions to the increment of the
resistance between the rounded disc and the circular electrical resistance when a bubble is present under
disc(lower curve) is caused mainly by the shape of the the anode: the screening effect and the deformation of
bubble extremities. The importance of the edge contri- the current lines under the bubble. But one question still
bution decreases as the bubble volume increases. Then, remains: what is the best geometrical configuration to
the difference decreases as the bubble volume increases. compute the effect of bubble shape on the electrical
These first results show that for intermediate bubble resistance: the situation where the volume of gas is
sizes, the error caused by the use of the circular disc conserved or when the value of the covering factor is
shape (A) to simulate the bubbles in a mathematical constant ? We must keep in mind that for a bubble-free
model that compute the total electrical resistance of the electrolyte, the current lines are vertical (with the
electrolyte is less than 0.7% at low covering factors. assumptions made earlier). The presence of a bubble
Figure 6 shows the influence of the bubble shape on deforms the electrical field and a horizontal component
the electrical resistance for the macro-bubbles, i.e. of the electrical field is created close to the bubble. Far
bubble volumes ranging from 20 to 80 cm3. Once again, away from the bubble, the electrical field is homoge-
the numerical calculations were carried out with a neous and the current lines are all vertical. In order to
covering factor of 10% based on the circular disc bubble characterize the size of the domain influenced by a
to avoid the edge effect. The difference presented on this bubble, a perturbed region (electrical current tube) of
graph is also defined by Equation 16 but, in this case, cross-section A*, can be associated which each bubble.
the studied bubble is the Fortin shape. First, the values The shape of the cross-section of this tube is circular in
of the resistance difference are one order of magnitude the case of the bubble shapes A, B, and D and
greater than those presented in Figure 5 concerning the rectangular for the shape C. In this study, the limits of
intermediate bubble sizes. Second, the influence of the the perturbed zone were defined where the horizontal
geometrical ratio r on the resistance difference is clear component of the current density equals 2% of its
and very interesting. It allows us to compare the nominal value, using the results of the numerical
contribution of the two phenomena responsible for the solutions. For the evaluation of the effect of the bubble
increase of the electrical resistance of the electrolyte shape, the increment of the electrical resistance caused
when a bubble is present: the screening effect and the by a gas volume with a specific shape (A–D) inside its
deformation of the current lines under the bubble. At a current tube is proposed. The cross-section of the
electrical current tube A* depends on the volume as well
as the shape of the gas pocket. Figure 7 shows the
variation of the covering factor Q* for the current tubes
with cross-section A* as function of the bubble volume
for the different shapes studied in this work. For the
Fortin shape, only the value of r = 0.25 is used since it
is the most realistic. Indeed, the two other values of r
would imply that there is a variation of the pressure
within the bubble. It can be seen from Figure 7, that the
value of the covering factor within an electrical current
tube reaches values higher than 50%. The difference in
the relative electric resistances (Equation 16) for the
different bubble shapes compared to the circular disc (as
reference) as function of the bubble volume is presented
in Figure 8. The tendency of the curves is the same than Fig. 8. Difference in the relative resistance as function of the bubble
the previous ones (computed at low covering factor) volume for the shapes studied in the present work.
with values of the relative difference slightly higher. If
the results concerning the Fortin shape are extrapolated,
the difference may reach 5% at very high bubble surface anymore. There is a liquid film between the
volumes (V  200 cm3). According to the interpretation bubble and the solid surface. The determination of the
of Richards et al. [18] of direct measurements on thickness of the wetting film is complicated because of
industrial cells, the latter values of bubble volumes are the proximity of two interfaces. Presently, the thickness
only reached under new anodes. With this approach of of the wetting film has been estimated to be less than
the characterization of the perturbed zone, the previous 1.5 mm. Several calculations have been carried out with
conclusions concerning the effect of bubble shape on the different bubble shapes and volumes. Figure 9 shows
electrical resistance are still valid. results obtained for the circular disc shape for three
different volumes. Calculations were done with a cov-
3.2. The influence of the wetting film ering factor of 30%. Here, the difference presented on
the y-axis of the graph is defined the same way as earlier
In this section, the effect of the thickness of the wetting by the Equation 16. However in the present case, the
film on the relative electrical resistance is calculated. It reference or the base of comparison is taken for the case
was mentioned above that there exist different regimes when the thickness of the wetting film is zero. This
of movement for a bubble initially attached under a situation corresponds to a bubble in direct contact with
solid surface. When the equilibrium between the con- the solid surface. The graph shows that the effect of the
trolling forces is broken for a creeping bubble, the wetting film on the relative electrical resistance is almost
formation of a wetting film begins. At the end of this independent of the bubble volume and the difference
transition process, the velocity of the bubble is one order increases linearly with the thickness of the wetting film.
of magnitude higher than that in the preceding regime, The results presented in Figure 9 are representative of
about 15 cm/s in an air–water gravity driven system. In all the calculations performed. Therefore it can be
this regime, the gas is not in contact with the solid concluded that the influence of the wetting film is less
than 2% on the total electrical resistance.

Fig. 7. Covering factor in terms of A*, the cross-section of the per- Fig. 9. Difference in the electrical resistance as function of the wet-
turbed zone (electrical current tube), as function of the volume for ting film thickness for three different bubble volumes. The base of
the different bubble shapes studied in the present work. comparison is when the thickness of the wetting film is zero.
Fig. 10. Increase of the relative electrical resistance due to the pres- Fig. 12. Relative electrical resistance for the configurations shown in
ence of the wetting film. Figure 11.

An interesting result presented in Figure 10 is that the studied. Three of these shapes have been reported or
relative resistance increases slightly with the thickness of observed in laboratory. The bubble volumes varied from
the wetting film. The results for three different bubble 1 to 150 cm3. The results show that the influence of the
volumes are presented such as 2, 8 and 15 cm3. The bubble shape for the same volume of gas on the total
bubble shape is the circular disc and the covering factor electrical resistance computed within its perturbed zone
is fixed at 35%. To understand this phenomenon, (electrical current tube) is less than 5%. The present
several tests were carried out with cylindrical bubbles paper confirms the conclusion of Hyde and Welch [5]
of four different diameters. The geometry is presented in that the effect of bubble shape is weak. Furthermore the
Figure 11. The bubble a is attached under the surface use of different computational domain sizes allowed us
with a thickness of 5 mm. In the second situation b, the to increase the understanding about the two contribu-
same bubble glides on a wetting film of a thickness of tions, such as the screening effect and the deformation of
1 mm. The third bubble c is also attached under the the current lines under the bubble, which are responsible
surface but its thickness is 6 mm. The anode–cathode of the resistance increase due to the presence of a bubble
distance ACD is kept constant at 5 cm in the three under the anode. The effect of the wetting film on the
situations. The relative electrical resistance computed electrical resistance has also been investigated. The
for these three situations is presented in Figure 12 thickness of the wetting film varied from 0 to 1.5 mm.
below. The results show that the additional resistance The influence of the wetting film has been found
caused by the presence of the bubbles b and c is almost negligible with a maximal value of only 2%.
the same. Therefore the increase of the electrical
resistance as the wetting film is increased is principally
due to a deeper penetration of the deformed zone in the Acknowledgements
current distribution. Thus the contribution of the small
increment of the conducting surface caused by the The first author gratefully acknowledges the support of
presence of the wetting film (situation b) is negligible the Fonds québécois de recherches sur la nature et les
compared to the deflection of the electrical field under- technologies (FQRNT) and that of the Conseil de Re-
neath the bubble. cherches en Sciences Naturelles et en Génie du Canada
(CRSNG) in the form of post-graduate scholarships.

4. Conclusions References

In this paper, the influence of the bubble shape on the 1. K. Grjotheim and H. Kvande (eds.), ÔIntroduction to Aluminium
electrical resistance due to the presence of a single ElectrolysisÕ, 2nd edn. (Aluminium-Verlag, 1993).
bubble under the anode has been computed by math- 2. L.I. Kiss, S. Poncsák, D. Toulouse, A.L. Perron, A. Liedtke and
ematical modeling. Four different shapes have been V. Mackowiak, in L. Nastac and B.Q. Li (eds.), Light Metals
Multiphase Phenomena and CFD Modeling and Simulation in
Materials Processes (TMS, Warrendale, PA, 2004), pp. 159–167.
3. W.E. Haupin, J. Metals 23 (1971) 46.
4. E. Dernedde and E.L. Cambridge. in R. Rentsch (ed.), Light
Metals, (TMS, Warrendale, PA, 1975), pp. 111–122.
5. T.M. Hyde and B.J. Welch. in R. Huglen (ed.), Light Metals,
(TMS, Warrendale, PA, 1997), pp. 333–340.
6. R.J. Aaberg, V. Ranum, K. Williamson and B.J. Welch, in
R. Huglen. (ed.), Light Metals (TMS, Warrendale, PA, 1997),
Fig. 11. Different bubble geometries. pp. 341–346.
 7. A. Solheim and J. Thonstad. in R.B. Miller and W.S. Peterson 14. A.L. Perron, L.I. Kiss, S. Poncsák, Int. J. Multiphase Flow (2006)
(eds), Light Metals, (TMS, Warrendale, PA, 1986), pp. 397–403. (in press).
8. K. Quian, D. Chen and J.J.J. Chen, J. Appl. Electrochem. 28 15. L.I. Kiss and S. Poncsák. in W. Schneider (ed.), Light Metals,
(1998) 1141. (TMS, Warrendale, PA, 2002), pp. 217–223.
9. J. Zoric and A. Solheim, J. Appl. Electrochem. 30 (2000) 787. 16. L.I. Kiss, S. Poncsák and J. Antille. in H. Kvande (Ed.), Light
10. H. Vogt and H.D. Kleinschrodt, J. Appl. Electrochem. 33 (2003) Metal (TMS, Warrendale, PA, 2005), pp. 559–564.
563. 17. S. Poncsák, L.I. Kiss, D. Toulouse, A.L. Perron and S. Perron. in
11. A.L. Perron, L.I. Kiss and S. Poncsák. in H. Kvande (ed.), Light T.J. Galloway (ed.), Light Metals, (TMS, Warrendale, PA, 2006),
Metals, (TMS, Warrendale, PA, 2005), pp. 565–570. pp. 457–462.
12. A.L. Perron, L.I. Kiss and S. Poncsák, Int. J. Multiphase Flow 32 18. N. Richards, H. Gudbrandsen, S. Rolseth and J. Thonstad. in
(2006) 606. P.N. Crepeau (ed.), Light Metals, (TMS, Warrendale, PA,
13. S. Fortin, M. Gerhardt and A.J. Gesing. in J.P. McGeer (ed.), 2003), pp. 315–322.
Light Metals, (TMS, Warrendale, PA, 1984), pp. 721–741.

View publication stats