ISIJ International, Vol, 33 (1993), No. 1, pp.

224-232

l
Physica Model of Slag Foaming

Yuji OGAWA.
Didier HUIN.1) Henri GAYE1)Naokl TOKUMITSU
Steelmaking Process. Process Technology Research Laboratories. Technical DevelopmentBureau. Nippon Steel Corporation,
Shintomi, Futtsu. Chiba-ken, 299-12 Japan. 1) Physical Chemistry Department. IRSiD, Voie Romaine. BP320.
Mazieres-les-Metz. 5721 4 France.

(Received on May27. l992, accepted in final form on October 16. 7992)

Physical model of slag foaming was developed by using results of cold and hot model experiments. The
bubble size at the slag/metal interface, the void fraction of foam and the film life of a bubble at the top
surface of slag were calculated. The governing factors of slag foaming have been clarified with this model.
It
was confirmed that the bubble size at the slag/metal interface is determined basically by the static
balance between the buoyancyforce and the adhesive force to the slag/metal interface. It
was resulted that
the slag/metal interfacial tension and the surface tension of metal also affect the foam height besides the
surface tension and the viscosity of slag because they change the bubble size.

KEYWORDS:
slag foaming; physical model; numerical calculation: foam; bubble; slag; metal; interface;
interfacial tension; contact angle; surface tension; viscosity.

following three principal stages, by X-ray fiuoroscopic

1. Introduction observations.4)
Slag foaming has been intensively investigated in 1) Formation ofCObubbles at the slag/metal interface
various steelmaking processes such as converter process, and their detachment,
bath smelting, hot metal treatment and electric 2) Rise of the bubbles in the slag layer and their
arc
furnace. accumulation under the free surface (formation of
But quite a few models of slag foaming have been the foam), and
proposed. Tatsukawa et al.1) proposed an empirical 3) Coalescence of bubbles in the foam and the rupture
equation which describes the change in foam height in of bubble films at the top surface of the slag which
BOFprocess. This model is simple and effective to use. tends to destroy the foam.
However, the knowledge of the average foam life is The physical model should correspond to these stages.
necessary to predict the foam height. It is difficult to Figure I schematically shows the concept of the •

knowthe average foam life in general. present model. The foam height is decided by the
Ito and Fruehan2,3) introduced balance of the gas evolution rate at the slag/metal
a foaming index as a
characteristic parameter of slag foaming. This param- interface and the gas escape rate at the top surface of
eter can be obtained from the slag properties. The slag the slag. The rate equation for the change in foam
height in the bath smelting and the electric arc furnace height is expressed as Eq. (1).
processes can be predicted with this model.3) The
coefficient of this model must be measured with each
dH
-
Qi*
-Q'~t ..........(1)

slag because it is an entirely empirical model, Iacking

dt A
physical background. For example, the effect of bubble where H is the foam height. Q. and
*" .~t are
the gas
Q
size
on the foaming index is ignored in this model. evolution rate and the gas escape rate respectively. A is
In this study, the authors have tried to make the cross section area of the vessel.
up a
slag foaming model with clear physical background. As for the estimation of the gas evolution rate,
Phenomenain three stages of slag foaming were data of previous experiments5,6) or somereaction rate
described based on the previous X-ray fluoroscopic models7) can be used. In this study, the works were
observations4) and two kinds of water model experi- concentrated to makeclear the governing factors of the
ments. With these physical models, the governing gas escape rate which determines the foam life.
factors of slag foaming were investigated. The gas escape rate is determined by the rupture rate
of bubble films at the top surface of slag, the size and
2. Model the numberof the bubbles. The rate of bubble rupture

2.1. Concept of Slag FoamingModel maybe governed by the bubble size, the film thickness
between bubbles and physical properties of the slag.
The slag foaming phenomenonis divided into the Moreover, the numberof bubbles and the film thickness

C 1993 ISIJ 224

 ISIJ International, Vol. 33 (1993), No, 1

foam hei9ht
~L
Qout

4~
~ rate ot film rupture
- film thickness

t between bubbles
If
distributiOn of blucll~tetbj,P"I_1_,ne)s
in fOam(void traetiOn)
e
+
9as evolutiOn
(bubble size) Fig. l.

Concept of present foaming model of siag.

C+(O)--CO

in the foam are considered to be determined by the gas

~volution rate, the bubble size and the slag properties.
Therefore, if the bubble size, the distribution of bubbles Ys
in the foam and the rupture rate of bubble films at the
top surface of the slag are known, the foam height can eG eS
be predicted.
Mathematical modeling in each stage of slag foaming
was constructed as follows: 7lvl
C YSM
l)
a model to calculate the size of bubbles evolved at @M
the slag/metal interface,
Flg' 2. Schematic drawing of the joining point between three
2) a model to estimate the distribution of bubbles in phases.
the foam, and
3) a model to calculate the rate of film rupture at the At the joining point of the three phases, if the
top surface of the slag. interfacial tensions are considered as force vector (Fig.
2), the angles formed by the three phases are expressed
2.2. Size of COBubbles at the Slag/Metal Interface
by the relationship:
2.2.1. Theoretical Basis
= (- y~ - + y~M)/2ysyM ............,,
Recently, Terashima et al.8) calculated the configura- cos((9G) y~ (4)

tion of CO bubbles at the slag metal interface. The cos((9M)

= (- y~ + - y~M)/2ySMyM...........
y~ (5)
theoreticai basis of the calculation in this study is

= (y~ - -
Also in cos((9s) y~ y~M)/2ysMys (6)
almost the same as that of their calculation. """"""" . . .

this study, the size of CObubbles at the slag/metal where yM, ys and ysM are the interfacial tensions of
interface was assumedto be determined by the static gas/slag, gas/metal and slag/metal respectively. The
balance of the three interfacial tensions and the static angle Os is commonlyreferred to as the contact angle
pressure. between slag and metal.
The calculations of the configuration of a bubble at
the slag/metal interface are based on the resolution 2.2.2. Equations for the Three Interfaces
of Laplace's equation of capillarity. This equation Gas/Slag Interface
expresses the difference of pressure across the interface The gas bubble assumedto be axisymetric as Fig.
is
between two phases I and 2. 3(a). At the apex O of the bubble, the two radii of
curvature equal to the samevalue called b.
AP=yl2(1/R +1/R ) ..........(2)

A point M(x, z) maybe associated with an angle ep

where yl2 is the interfacial tension between the two between the normal at the interface and the vertical
phases and R1 and R2 are the two principal radii of axis. If the curvilinear abscissa on the interface is noted
curvature of the interface. s, the two principal radii of curvature of the interface
At equilibrium this pressure difference is balanced by are such that:
the difference of static pressure, which is a function of
the position on the interface, the difference between the l/R1=ec/as and 1/R2=sinc/x ........,.(7)

densities of the two phases and the gravity. If the

Thus, the gas/slag interface is described by the system:
descending vertical axis is called Oz, this balance gives
at any point of the interface: ac/as
= -sin c/x+ ((pG- Ps)9/ys)z+ 2/b .....
(8a)

ex/as cos c az/as sin c ..............(8b)

AP=yl2(1/RI + l/R2) = =
p2)9(z-zo) + yl2(11RiO + 1/R20) -"""(3) with the conditions at the apex O:
= (pl - initial
x=z=s=
where, pl and p2 are the densities of the two phases, g c= O.

is the gravity acceleration and Ois a point of reference Gas/Metal Interface

where the principal radii of curvature are known. It is possible to show that the curvature of this

225 C 1993 ISIJ

 ISIJ International, Vol. 33 (1993), No. 1

o x

s~~ggl
oJr~ x
a
c*1
cc
9G_s
sls_ag
M Os
-----C s"
-- - ---- -

meral a
z' z
z

(a) (b) (c)

Fig. 3. Schematic drawing of the interface between two phases.

(a) gaslslag interface (b) gas/metal interface (c) slag/metal interface

interface has an opposite sign compared to that of 2.2.3. ComputingProcedure

gas/slag interface as shownin Fig. 3(b). Let O'(O, h) be The calculations were performed as follows:
the position (unknown) of the bottom of the bubble (1) the value of the radius
First of curvature at b
and b' the value of the two radii of curvature at this the apex of the gas/.slag interface is fixed, and the
point. The generic equation is written in the reference theoretical gas/slag contour satisfying Eq. (8) is com-
axis (O'x. O'z) where z' =z-h. puted.
The shape of the gas/metal interface is described by (2) For each one of these points C, there is only one
the following system: configuration of the gas/metal interface that satisfies the
contact angle condition expressed in Eq. (4). The proper
ac'/es'
=- sin ep'/x
+ ((pG- pM)g/yM)z' + 2/b' ,...,..(9a)
value of b' is obtained by successive approximations.
ax/as cos ep' az'las' sin c' (9b) (3) Starting from each point C, the slag/metal
, =- ... . . ...
interface is computed according to Eq. (lO). The con-
with the initial conditions at the reference point O':
figuration of the slag/metal interface depends on the
= = c' =
z' s' O.
x= position of C. The proper position with which the
Slag/Metal Interface slag/metal interface far from the bubble becomesflat is
For this interface the joining point C is the reference numerically determined by successive approximations.
point as shownin Fig. 3(c). The shape of the slag/metal b
When is increased, there appears a value beyond
interface can be computedby resolution of the system: which no position of C
Ieading to a flat slag/metal
interface can be found: no bubble with such a large
ac"/as"
= - smc"/x+((pM Ps)9/ysM)z
value of the top radius of curvature can be in a stable
+ (2ys/b-2yM/b' + (pG-pM)gh)/vsM ""(10a) position at the flat slag/metal interface.
If the properties of slag and metal are given, the
ex/as" =cos c" ,
az/as"
= sin ep"
.....
(10b)
maximumvolume of a bubble which exists stably at
with the initial conditions at the point C: the flat slag/metal interface can be obtained. It was
assumedthat this volume is that of a bubble just before
x=xc; z zc, s" Oc" cc+6, Ir
detachment from the slag/metal interface. The size of a
Equation
curvature
(ofIO) the
takes into account the actual equilibrium
slag/metal interface at the point C,
bubble evolved at the slag/metal interface was obtained
from this maximum volume, assuming that the bubble
resulting from the pressure difference across this inter- after detachment is spherical.
face .

For each one of the three interfaces, the positions of 2.3. Bubble Distribution in the Slag Layer
successive points spaced at fixed distance ~s apart, are
calculated step by step. In order to compute the
The X-ray observations of slag foaming showed the
displacement (dx, dz) corresponding to the relative emergenceof two superposed layers in the slag. .4)
(1) in lower layer, the bubbles seemed to
position of two successive points, the following third move
freely and the void fraction is low,
order Taylor'series expansion is made:
(2) in upper layer, they formed
a dense packed bed
(a2x/ds2)5s2/2
dx= (ax/es)8s+ + (a3x/as3)8s3/6 ...(1 Ia) (foam) where the void fraction is muchhigher.
(a2z/as2)~s2/2
This behavior is typical two-phase flows. Thus it seems
dz
= (azlas)~s + + (a3z/as3)8s3/6 ....
(1 Ib)
possible to analyze the foaming phenomenon with the
the various partial derivatives being obtained by dif- help of two-phase flow correlations.
ferentiation of Eq. (8), (9) or (lO). Wallis9) proposed
a theoretical model to describe

C 1993 ISIJ 226

 ISIJ International. Vol. 33 (1993), No. 1
two-phase flows in a vertical pipe. He assumed fiat obtained from F values of neighboring cells. The
velocity profiles in the cross section of the pipe. Second
pressure and velocity field are calculated with use of the
assumption is that the rise velocity W*
in an infinite surface tension of liquid and this curvature.
liquid mediumfollows the modified Stokes law: The fiow of slag around a bubble at the top surface
of the slag was calculated with the VOFmethod. The
W*= (pL - PG)9R213,4L .....
..........(12)
bubble was fixed to be spherical and the size was
where pL and pG are the densities of liquid and gas, /lL ranged from 2to 6cmin diameter, because of difficulty
isthe viscosity of liquid and R is the radius of the in convergence for smaller bubbles. The initial position
bubbles. of slag surface was taken as I .5 cmabove the top of the
From the relationship between the actual and the bubble. The average decreasing rate in film thickness
superficial velocities of gas phase, two values of void from 5 to O at the apex of the bubble was calcu-
fraction (oci and oc2) which correspond to two layers are
mm
lated. Although the effect of the bubble deformation is
theoretically obtained. ignored and the bubble size is large, this calculation will
be available to evaluate the effects of slag properties
o(1=(1+fr4J/W*)/2 and oc2=1-0(1 ......(13)
and the bubble size on the rate of film rupture.
where J is the velocity of gas phase (ratio gas
suD*erficial
flow rate to cross section area). Wallis proposed a 3. Water Model Experiment
diagram A simple reading of
as shown in Fig. 4. this
Two kinds of water model experiments with gas
diagram two values of void fraction.
gives directly the
bubbling as shown in Fig. 5 were performed. Experi-
Thus, for given properties of gas and slag and a
fixed bubble size this model allows to estimate the void
ment I is to verify the ability of the Wallis model to
describe the foaming phenomenon.Experiment 11 is to
fraction in the two parts of the flow, that is, the
know the difference of film life of a bubble at the top
dispersed gas layer (10wer part) and the foam layer
of foam from that at the top surface of the slag without
(upper part).In the present study, the ability of Wallis
foam.
model to describe the void fraction in foam was verified.
In experiment I, an initial height Ho of liquid
2.4. Rupture of Bubble Films at the Top Surface of Slag (water + saccharose) was put in a vertical cylindrical
tube (about I in height, inner diameter of 37mm).
As mentioned before, if the film life of single bubble m
Nitrogen gas was injected at the bottom through a glass
at the top surface of the foam is estimated, the gas
filter. In order to reduce the bubble coalescence at the
escape rate from the foam can be evaluated as function
of the numberand the size of the bubbles. However, it outset of the glass filter, 2"/o ethanol was added in
water;1 1) the size of the bubbles
is difficult
to calculate directly the rate of film rupture was muchlarger (about
of a bubble at the top of the foam. Therefore, in the 5 mm in diameter) without ethanol. Increasing the gas
flow rate, two superposed layers appeared. The void
present model, it was assumedthat the film life of a
fractions in the two layers were evaluated by the height
bubble at the top of the foam is the sameas that at the
of the layers H1 and H2'
top surface of slag without foam.
In experiment II, argon gas was injected into aqueous
VOF (Volume Of Fluid) methodlo) was adopted to
solution containing I o/o gelatin. The diameter of the
the calculation of film rupture of the bubble. In this
cylinder was 60 and the four kinds of glass fiters
method, the free surface is treated by introducing a mm
which have different size of pores were used. The
function F. The value of F
in a cell equals to the
fractional volume of the cell occupied by fluid. By
calculating F, the new free surface orientation in each
-=~ flow meter
cell is determined. In the cell with value between zero F
and unity, the curvature of the free surfaces are

t
H1 fOamlayer
0.15
OOO
O OO
J/ W= O._;2 o
O15~~
OOO
ooo
O.1Q oe( I -a)2 Ooo
OOO
Ooo
". OO
OOo H2 dispersed
O OO
(1-a)J/W* o OO
OOa gas layer
0.00 O OO
O OO
o OQ
oo
glass fi,ter

O a I O.2 0.4

void traction
0.6

Oe
0.8 a2 1.o
t
gas In]ection
Fig. 4. Wallrs diagram 8) Fig. 5. Schematic representation of the water model.

227 C 1993 ISIJ

 ISIJ International, Vol. 33 (1993). No. 1
foaming phenomenonwas recorded by using a high increasing the iron oxide content in slag. Therefore, it is
speed camera(400 frames/sec). The foam height and the concluded that the decrease in the contact angle by
bubble size was measuredwith these images. After the these surface-active elements causes the decrease in the
gelatin solution was foamed, gas supply was stopped bubble size.
and the change in the height of solution was measured. Figure 7 shows the effect of surface tension of the
slag on the bubble size whenthe other parameters were
4. Results and Discussion fixed. The bubble size decreased with decreasing the

4.1. Bubble Size at the Slag/Metal Interface

surface tension of slag. Figure shows the effect of 8
surface tension of metal on the bubble size. The bubble
From the calculation of the bubble at the slag/metal size decreased with the increase in the surface tension
interface, it
was found that the critical size of bubbles
was dependent on all of three interfacial tensions.
Figure 6 shows the effect of the contact angle on 7~ = 1.6 (N/m)
the bubble size by changing the slag-metal interfacial E 8
E 7 T*~= 1.3 (N/m)
tension. The critical diameter of bubbles increased 2.5 (glcnlP)
a'
~ p.=
almost linearly with the increase in the contact angle. ,D 6
The relationship between the critical bubble diameter E
a 5
~i
measured in the X-ray fiuoroscopic observations and
o 4
the contact angle is also shownin
the samefigure in the da
n
case of two sulphur concentration. The contact angle
data used were those obtained by Crambet al,12) based
B 3
~lo
2
on the measurementsof Gaye et a!.13) The measured ,~-

1
bubble size agreed well with the calculated one with this
model. o 0.4 0.5 0.6
Cramb et al. showed that the increase in oxygen surface tension of slag (N/m)
content and the decrease in sulphur content lower the
Fig. 7. The effect of the surface tension of slag on the critical
contact angle between slag and metal. Mukai et al.14) size of adhesive bubble to the slag/metal interface.
indicated that the contact angle decreases sharply with

E 12
E Is = 0.5(N/m) E 8
E 7
~*O 10 TM= 1.6(N/m)

~
O
E8
cc
ps = 2.5(glem3)
~
O
,g
11
6
5
1,
6 e'F- 4
,D exp. 1
!: (S:0.40/0) 3
Q
=
4 Present work
:
J:I
7s = 0.5 (Nlm)
=L \ex p.2 -8 2 lr.*= 1.3 (N/m)
cc 2 :E
o (S:0.050/0)
*O 1 p.= 2.5 (glcnf)
:S:

o Oo 20 40 60 80 1oo i 20
o 1.O 1.2 1.4 I .6
COntact angle (deg.) surface tension of metal (Nlm)
Fig. 6. The effect of the contact angle between slag and Fig. 8. The effect of the surface tension of slag on the critical
metal on the critical bubble size. size of adhesive bubble to the slag/metal interface.

mm(a)
2
mm(b)
8

6
1 Fig. 9.
Successive calculated contours of a bubble
4 growing at the slag/metal interface for two
values of the contact angle ((a) es=20
degrees, (b) Os=100 degrees).
l
, 2 Dotted lines and continuous lines show the
intermediate contours and the maximum
bubble size.

O 1 mm O 2 4 6 8 mm

C 1993 ISIJ 228

 ISIJ International, Vol. 33 (1993), No. 1

~LA

15
RbubbLc
=0.18
\ e water + ethanol
-
uL= l0-3 kg/m/S
A A
mm
'r water + ethanol uL = 2. 10 3kg/m/s
A
JL
e , A + saccharose
eA
O, 16
~,L mm A

~
~a 10 JL
A A

E e

E e
0.18 mm

~
1 A
,
t t O, 16 mJ11 e

5 JL

t J';

JL

O 0.5 1.O
VOid traetiOn oc (-)

Fig. lO. The relationship between the void fraction and the superficial velocity of gas obtained in the water model
experiment, for two values of viscosity.

of metal beyond 1
200 dyn/cm. It is considered that Table l. The relation between the pore size of filters and
these effects are due to the change in contact angle by the observed bubble size.

changing the surface tension of slag or metal.

Bubble contours during their growth are lllustrated
Filter No. 1 2 3 4
in Fig. 9 for two extreme values of contact angle. In Pore size (pm) 5l O 2(~30 40-50 lOC~120
the case of low contact angle, the bubble was almost Bubble diameter (mm) O. l0.3 0.3-0.5 0.6-1 .O 0.7-1 .5

spherical and the gas/metal contact area was small.

On the other hand, when the contact angle is large,
the bubble is compressed near the bottom and the good in spite of the important infiuence of the bubble
gas/metal contact area is relatively large. The shape of size. Somecoalescence occurred during the rise of
bubble agreed with that observed in the experirnent bubbles in the column and their size depended on the
with X-ray fiuoroscopy.4) The buoyancy force calcu- gas flow rate; this explains the scattered results for high
lated by the volume of the bubble and the density of gas fiow rate.
slag and metal agreed well with the sumof the vertical Even if it seemsrelatively simplistic, the Wallis model
componentof slag/metal interfacial tensions around the predicts the emergenceof two flow configurations in the
joining boundary of the three interfaces. Therefore, the samecolumn: dispersed gas layer and foam layer. It was
decrease in the gas/metal contact area will cause the confirmed that Wallis model is applicable to evaluate
decrease in the force which should be balanced with the the void fraction of the foam as functions of the gas
buoyancy forces. evolution rate, the bubble size and the slag properties.

4.2. Void Fraction in Foam 4.3. Film Life of a Bubble at the Top Surface of Slag
In the water model experiment, it was verified at first 4.3.1. Water Model Experiment
that the value of W*
compatible with the values of ocl In the experiment with gelatin solution, the average
and c(2 remained reasonably constant, which showsthat bubble size increased with the increase in the pore size
the bubble size is independent of the gas flow rate. as shown in Table l. The average bubble size was
Therefore, the bubble size was estimated by assuming almost constant with the change in the gas flow rate
that Wallis model is true. from 100 to I OOOcm3/min.
Measurementswere madewith two kinds of liquid Figure 11 shows the effect of the average bubble
which have different viscosities. The results are pre- size
on the gas escape rate. The gas escape rate was
sented in the Wallis diagram (Fig. lO). For comparison calculated by the foam decline rate after gas supply
with Wallis theory, the curves corresponding to fixed was stopped. The gas escape rate increased with the
values of bubble radius (0.16 and O.18 mm)have been increase in the bubble size. In other words, the foam life
drawn in the samefigure. The correlation was relatively increases with decreasing the bubble size. The result

229 C 1993 ISIJ

 ISIJ International, Vol. 33 (1993), No. 1
40o e,
\
=
)(
~
~
10.5

~3oo
E o
~o ~
EI 0.4
~ 20Q
co*
~
o -o
a o P*= 2.5 (g/cnf)
,5
o ~
~ = IO (P)
11*

~ Ioo
o
,D
c IO 3 d = 5 (cm)
co
2
~
cao'

0.4 0.5 0.6

o surface tension of slag (Nlm)

o o.5 1
Fig. 13. The effectof the surface tension of slag on the
average bubble diameter (mm) decreasing rate of thickness of bubble film.
Fig, Il, The effect of the average bubble size
on the gas
escape rate from the foam.
co
.\
E P.= 2.5 (g/cnf)
o
*J n.= 10 (P)
20
T* = 0.5 (N/m)
J'
,S
5
N:
E
*O 15

~
(i)
d=2 (cm) (ii)
d=3 (cm)

Fig. 12. Calculated configuration of the top surface of slag
around a bubble just before the film rupture.
agreed with the result in the case of molten slag.4)
Gas escape rate for a single bubble (Imm in
diameter) at the top surface of the liquid without foam
was estimated in a following method.
(1) The time from beginning of the surface de-
formation by the effect of the bubble rise to the
momentof the rupturc of the bubble film was measured
by using the high speed TVcamera.
(2) From the result of the foaming experiment, the
void fraction was calculated by the foam height in the
case that the average bubble diameter was I mm.
(3) The numberofbubbles at the top surface of the
foam were evaluated by the void fraction.
(4) The gas escape rate was estimated by the time of
film rupture of a single bubble and the number of
bubbles at the top surface of the foam calculated in (3).
The X mark In Fig. II represents the result of estima-
tion. The estimated value agreed with that obtained by
the foam decline rate. Fromthis result, it is considered
that the gas escape rate from the foam can be predicted
from the rate of film rupture of a single bubble at the
top surface of liquid without foam.
4.3.2. Numerical Calculation
Figure 12 shows the configuration when the filrn
thickness becomeszero. In the case of cmin diameter
of the bubble, the level of slag surface was about the
half height of the bubble. On the other hand, the level
of the slag surface was near the bottom of the bubble in
~
lo
2 3 4 5 6
bubble diameter (cm)
Fig. 14. The effect of the bubble diameter on the decreasing
rate of the thickness of bubble film.
the case of 3
cmof the bubble. The curvature of the slag
surface at the same level of the surface increases with
the decrease in the diameter of the bubble. Therefore, it
is considered that the decrease in film thickness is
accelerated by the pressure difference between at the
apex and near the bottom of the slag surface due to this
curvature.
The effects of surface tension on the rate of the
decrease in film thickness is shown in Fig. 13. The
increase in the surface tension of slag enlarged the
decreasing rate of film thickness. The smaller the
surface tension of slag becomes, the smaller is the
pressure difference mentioned above. This may results
in the suppression of the flow in the film.
Figure 14 presents the effect of the bubble diameter.
The decreasing rate of film thickness increased with the
decrease in the bubble size below 5
cmin diameter. This
may be because of the increase in pressure difference
due to the increase in curvature of the slag surface.
However, there is a ininimum point in this rate between
5 and 6cmin diameter of the bubble. The decreasing
2 rate of film thickness may be accelerated also by the
hydrostatic pressure. Therefore, it is regarded that the
interaction between the pressure due to the curvature
of the surface and the hydrostatic pressure causes the

C 1993 ISIJ 230

 [SIJ International, Vol. 33 (1993), No. 1

'O
s]agand metal on the foam height could be made.
*~
S
O
Table 2showsthe effects of the properties of slag and
~
'O
'O
metal on the foam height. The effect of each property is
O 10 considered as follows:
:
~O Viscosity of Slag
E According to Wallis diagram, the void fraction of the
E
~: foam decreases with the increase in the slag viscosity,
.h
O 9 then the film between two bubbles in the foam
'D
may
~S become thick. The downward flow in the film of
p. = 2.5 (glcnf) the bubble at the top surface of the foam
O
'O
T. = 0.45 (N/m)
may be
,e suppressed by the increase in the slag viscosity. These
O d = 5 (cm)
88
U
two effects will cause the decrease in the rate of film
rupture of the bubbels at the top of the foam and then
1 5 10 50 100
(P)
the increase in the foam height.
slag viscosity
Fig. 15. The effect of of slag on he decreasing
the viscosity Surf~ce Tension of Slag
rate of the thickness of bubble film. According to Fig. 6, the bubble size increases with
increasing the surface tension of slag. The film between
Table 2. The effects ofthe properties ofslag and metal on bubbles becomesthin, because the void fraction in the
the bubble size at the slag/metal interface, the foam increases with the increase in the bubble size. The
void fraction of foam, the rupture rate of bubble downwardflow in the film of the bubble at the top of
film and the foam height. the foam will be promoted because of the increase in
ure ra e the pressure difference due to the curvature of the slag
Foamheight up ot fllm Bubble s~ze j Void
, fraction
bubble of foam
surface around the bubble. Therefore, the rate of film
Slag viscos~y A increase v v rupture of the top bubbles will increases with the surface
Surface tension tension of slag and the foam height will decrease.
of slag A
l
decrease A A A
Slag/Metal Interfacial Tension
Slag-metal
interfacia] tension A decrease
,
A A + According to Fig. the bubble size increases with
8,
Surface tension
of metal A
l
increase
v A v the slag/metal interfacial tension. It will also cause the
increase in the rate of film rupture of the top bubbles
and then the decrease in the foam height.
minimumpoint. The maximum
point in bubble lifetime
Surface Tension of Metal
was recognized also in the measurementwith p]astisols According to Fig. 7, the increase in the surface
by Harrell et al.Is) According to the measurementwith
tension of metal decreases the bubble size. Therefore,
aqueous solution of soap,16) the lifetime increased with the foam height increases with the surface tension of
decreasing the bubble size. In this case, because the
metal.
bubble is quite stable, the lifetime may be governed
mainly by the stability of the liquid film, not the
mechanism above. In the case of the actual slag The effects of viscosity and surface tension of slag on
foaming caused by slag/metal reaction, the bubble size
the foam height obtained in this study agree with those
of most of the previous works. It has been clarified that
is less than
2mm.4) Therefore, the rate of film rupture the slag/metal interfacial tension and the surface tension
of bubbles in the slag foam will increase with the
decrease in the bubble size. But the increasing rate is of metal also affect the foam height because they
not so much, while the volume of a bubble is propor- may change the bubble size evolved at the slag/metal
interface.
tional to the third power of the diameter. Thus, total
gas escape rate will increase with the increase of
the bubble size as observed in the water and X-ray 7. Conclusion
experiments.
Physical model of slag foaming was derived by
Figure 15 shows the effect of the slag viscosity. The
using results of cold and hot model experiments. The
rate of the decrease in film thickness decreased with
increasing the slag viscosity.
governing factors of slag foaming have been clarified
more in with this model. The effects of the
detail
physical properties of slag and metal on the foam
6. The Effects of Properties of Slag and Meta] on the
height have been made clear. It was confirmed that
FoamHeight the bubble size evolved at the slag/metal interface is
,

The governing of the bubble size evolved at

factors
the slag/metal interface, the void fraction in the foam
and the rate of fiim rupture of the bubbles at the top
surface of the foam was madeclear with the three mod-
els of each foaming process. Therefore, the qualitative
evaluation of the effects of the physical properties of
determined basically by the static balance between the
buoyancy force and the adhesive force to the slag/metal
interface,and the slag/metal interfacial tension and the
surface tension of metal also affect the foam height
besides the surface tension and the viscosity of slag
through the change in the bubble size.

231 C 1993 ISIJ

 ISIJ International, Vol. 33 (1993), No. 1
9) G. B. Wallis: International Developments in Heat Transfer,
REFERENCES ASME,(1963), 319.
1) M. Tatsukawa, M. Shimada, M. Ishibashi and T. Shiraishi:
1O) M. D. Torrey, L. D. Cloutman, R. C. Mjolsness and C. W.
Tetsu-to-Hagan~, 60 (1974). A19. A
Hirt: NASA-VOF2D: ComputerProgram for Incompressible
2) K. Ito and R. J. Fruehan: Metall. Trans. B, 20B (1989), 509. Flows with Free with Free Surface, LA-l0612-MS, Los Alamos
3) K. Ito and R. J. Fruehan: Metall. Trans. B, 20B (1989), 515. Scientific NewMexico, (1985).
Laboratory, Los Alamos,
4) Y. Ogawaand N. Tokumitsu: Proc. of 6th Int. Iron and Steel 11) B. J. Ackers: Forms. AcademicPress, London, (1976), 147.
Cong. Vol. l, ISIJ. Tokyo, (1990), 147. l 2) A. W. Cramband J. Jimbo: Stee/. Res., 60 (1989), 157.
5) W. Pan, M. Hiramasa, M. Sanoand K. Mori: Tetsu-to-Hagan~, l 3) H. Gaye, L. D, Lucas, M. Olette and P. V. Riboud: Can.
73 (1988), 61. Metall. Q., 23 (,1984), 179.
6) A. Shinotake, N. Tokumitsu and Y. Hayashi: Tetsu-to-Hagan~, 14) K. Mukai, H. Furukawaand T. Tsuchikawa: Tetsu-to-Haganb,
73 (1987), S121. 63 (1977), 1484.
7) For example. K. Shibata, T. Kitamura and N. Tokumitsu: ISIJ l 5) E. R. Harrell, J. S. Kim and G. A. Gall: J. Vinyl Technol., 5
Int., to be published. (1983), 143.
8) H. Terashima, T. Nakamura,K. Mukai and D. Izu: Trans. Jpn. 16) T. Sasaki ei a/.: Kaimen Gensh5-no-Kiso, Asakura-shoten,
Inst. Met., 56 (1992), 422. Tokyo, (1973), 171.

C 1993 ISIJ 232