G

B
growth rate on susceptor [cm/s] = exit of reactant section

h
Gr Grashof number (=gPD3e(th- tw)/v2) H = reaction zone
gravitational acceleration [cm/s2]

w
9 =wall

0
k mass transfer coefficient [cm/s] =inlet of reactor
I length of reaction zone [cm]
M molecular weight of SiCl4 (M= 170) H (Superscript)
m atomic weight of Si (m=28) [-] =average
Re Reynolds number ( = Dev/v) H
r radial distance [cm] Literature Cited
outer radius of inner tube [cm] 1) Bradshaw, S. E.: Int. J. Electron, 21, 205 (1966).
inner radius of outer tube [cm] 2) Fujii, E., H. Nakamura, K. Haruma and Y. Koga: /.
cross-sectional area of annulus Electrochem., 1106 (1972).
( = n(r2w-r2s)) [cm2]
Sh H
3) Hanzawa, T., K. Sakauchi, U. Ito, K. Kato and T. Tadaki: /.
average Sherwood number (=kDJ@) Chem. Eng. Japan, 10, 313 (1977).
local Sherwood number (=kDJ@) [-]
t [K]
4) Hanzawa, T., K. Sakauchi, K. Kato and T. Tadaki: /. Chem.
temperature Eng. Japan, 10, 319 (1977).
v velocity in z-direction [cm/s] 5) Hanzawa, T., K. Kato and T. Tadaki: Kagaku Kogaku
axial distance [cm] Ronbunshu, 9, 257 (1981).
axial distance from leading edge of 6) Hanzawa, T., A. Sako, H. Endo, M. Kagawa, T. Sunagaand
reaction surface [cm]
K. Kato: /. Chem. Eng. Japan, 19, 78 (1986).
7) Hanzawa, T., A. Sako and K. Kato: J. Chem. Eng. Japan, 19,
p
[K-1]
volumetric coefficient of expansion 96 (1986).
9
angle in cylindrical coordinates [rad] 8) Manke, C. W. and L. F. Donaghey: /. Electrochem. Soc,
\i viscosity [Pa-s] 124, 561 (1977).
v kinematic viscosity cm2/s]
9) Shepherd, W. H.: /. Electrochem. Soc, 112, 988 (1965).
P density
[g/cm3] 10) sherwin, K.: British Chem. Eng., 13, 1580 (1968).
(Subscripts) (Presented at World Congress III of Chemical Engineering, at
A = entrance of reactant section Tokyo, Sept. 1986.)

DEMULSIFICATION KINETICS OF W/O EMULSION IN AN

A.C. ELECTRIC FIELD

Tadashi HANO, Takaaki OHTAKEand Kazumi TAKAGI

Department of Environmental Chemistry and Engineering, Oita University, Oita 870-ll

Key Words : Extraction, Demulsification, W/O Emulsion, Kinetics, Viscosity, Electric Field
The demulsification kinetics of W/Oemulsion in a high A.C. electric field was investigated by using a batch
cylindrical demulsifier in which a glass-sealed electrode was placed above a grounded copper disk electrode. The
effects of emulsion preparation conditions (oil-phase viscosity, water drop size, water-phase holdup and surfactant
concentration) and demulsifying conditions (agitation speed and temperature) on the demulsification rates were
examined experimentally while keeping the applied voltage constant. The demulsification rates varied considerably
with slight change of these conditions. The rate equation that expressed the above contributions in suitable
functional forms gave satisfactory agreement with observed rates over a wide range of experimental conditions.
Based on comparison with thermal demulsification without electric field, the accelerating effect of temperature rise
was thought to be caused by the decrease of oil-phase viscosity and emulsion stability. Mixing of the demulsifier
contents during the operation was found to be effective in promoting demulsification together with temperature rise.

Introduction has received a great deal of attention in, for example,

The application of (W/O)/W emulsion to the ex- the fields of hydrometallurgy, wastewater treatment,
and purification of fermentation products.7'944) In
traction and concentration of concerning
very dilute components
Received August 27, 1987. Correspondence this article should be ad- these processes, the breaking of W/O(water-in-oil)
dressed to T. Hano. emulsion after extraction is necessary in order to
VOL. 21 NO. 4 1988 345
recover the components concentrated in the inner conditions. The contributions of drop size, holdup,
water phase and to re-utilize the oil containing sur- surfactant concentration, agitation speed and operat-
factants and carriers. In the various demulsification ing temperature were studied to evaluate the de-
techniques proposed so far, the electrostatic method mulsification rates.
in which a high A.C. or D.C. voltage is applied to the
emulsion is considered to be the simplest and most 1. Experimental Procedure
effective.2'4'5) By observations under a microscope, it 1.1 Demulsification apparatus and operation
became clear that electrostatic forces caused the The batch demulsification apparatus shown in Fig.
coalescence of fine water drops and their growth to 1 is similar to that used by Fujinawa et al.1] The
larger drops which then readily fell by electrostatic demulsifier was made of acrylic-resin tube equipped
forces or gravity.2'5J2'13) The complex motion of water with an annularjacket to keep the emulsion tempera-
drops in the electric field makes it difficult to derive a ture constant during the run. The copper disk elec-
generalized expression for demulsification rates. trode, 41 mmin diameter, was set at the bottom and
Fromthe coalescence mechanismproposed so far, the grounded. The glass-sealed electrode was made of
following parameters are considered to affect the 8 mm-diameter glass tube of 1 mmthickness and was
demulsification kinetics:5} (operating conditions) vol- placed 80mmabove the disk electrode. The tube was
tage, frequency, temperature, degree of mixing, shape filled with saturated aqueous sodium chloride so-
and distance of separation of the two electrodes; lution into which a 2mm-diameter copper wire was
(emulsion properties) density, viscosity, inter facial immersed. A.C. 15kV of 60Hz was applied between
tension, water drop size, holdup, surfactant con- the two electrodes in all experiments. Unless other-
centration in oil phase, electrolyte concentration in wise noted, the experiments were carried out using
water phase; (electrical properties) conductivity, 120cm3 emulsion at 30°C without stirring the con-
dielectric constant. tents. After the start of demulsification, the variation
Because of the manyparameters to be examined, a of volume of both the water and oil layers was
complete rate equation which takes into account all measured at an appropriate time interval.
these factors has not yet been presented. A few papers 1.2 Preparation of emulsion
have reported the contribution of some of these The original W/Oemulsion was prepared by mix-
factors while keeping other conditions con- ing 40cm3kerosene solution containing 2wt%Span
stant.1'2'4'6 13) Their results, however, have not nec- 80 (Kao-Atlas Co.) and 80cm3 aqueous solution of
essarily agreed with each other, probably because of 20 mol/m3 copper sulphate with homogenizer (Nihon
the differences in demulsifier configuration employed. Seiki Co.) at 83s"1. The emulsion for the run was
In the parameters listed above, the coalescence and prepared by mixing 120cm3 original emulsion with
falling processes of water drops are greatly affected by 40cm3solution of kerosene and liquid paraffin con-
surroundings viscosity,3) but its effect has not been taining 2 wt% Span 80. The standard sample prepared
by this method contained water drops of 5.33/mi
analyzed quantitatively in past works. The viscosity is
also knownto be one of the major factors that control average diameter and had 50%water-phase holdup.
emulsion stability.5'8"11* Therefore, in the present The viscosity of the oil phase was altered by changing
work the demulsification kinetics was investigated at the composition of oils used to dilute the original
constant voltage by examining primarily the effects of emulsion. Samples of different drop sizes were pre-
viscosity for each emulsion prepared under various pared by changing the agitation speed of the homoge-

Fig. 1. Experimental apparatus

346 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN

nizer in the range of 50 to 183s *. The water-drop Such a sigmoidal pattern may be obtained due to
diameter in these emulsions showed a log-normal batch operation. The demulsiflcation rate dVw/dt was
distribution.8) The holdup of water phase was con- evaluated at the point of 50% demulsiflcation where
trolled by changing the volume ratio of dilution oil to the water layer was formed at nearly constant ve-
original emulsion. The viscosity of oil phase and locity. The electric current was maintained at ap-
emulsion was measured with a Type B rotational proximately 0.22mA until 80%demulsiflcation was
viscometer (Tokyo Keiki Co.). The inter facial tension attained and then it dropped gradually
between aqueous solution and oil phase which con-
stituted the emulsion was measured by a pendant 2. Experimental Results
drop method. Both measurements were conducted at The effects of operating and emulsification con-
the same temperature as in the demulsification run. ditions were studied by examining the viscosity de-
1.3 Evaluation of demulsification rates pendence of the demulsiflcation rates for each
Figure 2 shows an example of demulsification be- emulsion.
havior. The extent of demulsification, which was 2.1 Effect of water drop size
defined as the fraction of the separated water-phase The effect of water drop size was studied by
volume against the initial one, showed a gradual changing the diameter from 2.68 to 7.61 jam at con-
increase at an early stage followed by constant in- stant water-phase holdup and surfactant concen-
crease and subsequent convergence to completion. tration. The results are plotted against the viscosity of
emulsion and oil phase in Figs. 3 and 4, respectively.
Each emulsion had the same dependence on emul-
sion viscosity (-2.0power) and oil-phase viscosity
(-3.2power). In Fig. 3, the apparent viscosity at a
shear rate of 10s"1 was adopted as the emulsion
viscosity, since the emulsion was a non-Newtonian
fluid. The exponent of the viscosity nevertheless re-
mained constant at other shear rates because the
parallel linear relation between logarithmic viscosity
and logarithmic shear rates held for a series of the
emulsion tested.8) Ino et al. correlated demulsiflcation
Fig. 2. Time course of demulsification (/ioii=3.22mPa-s, rates linearly with emulsion viscosity, which varied as
(/>w=0.62, ^V-3.338-1)
a result of water-phase holdup change.2) In deriving

Fig. 3. Effect of emulsion viscosity on demulsification rate Fig. 4. Effect of oil-phase viscosity on demulsification rate
for emulsions of different drop size for emulsions of different drop size
VOL. 21 NO. 4 1988 347
 Fig. 7. Effect of water-phase holdup on demulsification
rate
Fig. 5. Effect of water drop size on demulsification rate

Fig. 8. Effect of surfactant concentration on demulsifl-

cation rate

close to that of Fujinawa et al. Whoreported 3.O.1*

2.2 Effect of water-phase holdup
Figure 6 shows the dependence on the oil-phase
viscosity of emulsions having 33 to 57% water-phase
holdup. The slopes of all straight lines were the same
as in Fig. 4. The contribution of holdup derived from
Fig. 6 could be expressed well by an exponential
function as shown in Fig. 7. The demulsification rate
was found to increase considerably with a slight
Fig. 6. Effect of oil-phase viscosity on demulsification rate decrease of water-phase volume. The rapid rate de-
for emulsions of different water-phase holdup crease with holdup may be attributed to the variation
the equation which correlates the demulsification
of electrical properties by the increase of water con-
tent in the emulsion.
kinetics, it is considered preferable to employ oil- 2.3 Effect of surfactant concentration
phase viscosity rather than emulsion viscosity because The increase of surfactant Span 80 concentration
the latter varies with drop size and water-phase retarded the demulsification exponentially as shown
holdup in addition to oil-phase viscosity. Therefore, in Fig. 8. Fujinawa et al. obtained a -0.5power
the oil-phase viscosity was used to analyze the exper- dependence concerning the same point.X) The de-
imental data in the following discussion. Figure 5 mulsification depression by surfactant has been ex-
shows the dependence on drop size at a certain oil- plained as an effect of an increase in stability of the
phase viscosity. The drop size in this figure is the interface surrounding each water drop.4 13) The range
Sauter mean diameter. The exponent 3.5 obtained is of surfactant concentration studied, however, is over
348 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN
1 wt%, which is rather in excess of the critical micelle phase viscosity under stirring was the same as in Fig.
concentration. Therefore, it is difficult to expect fur- 4 regardless of agitation speed.
ther stability increase by surfactant addition in the 2.5 Effect of operating temperature
present case. To elucidate the cause of rates decrease, Figure 10 shows Arrhenius-type plots of the de-
the experimental points shown in Fig. 8 were re- mulsification rates, whichdemonstrate a sharp rate
plotted in the same manner as in Fig. 4 while taking increase with temperature rise. Such an acceleration
into account the increase of oil-phase viscosity by effect of temperature has not been studied in detail so
surfactant addition. As a result, all the points agreed far, although temperature is one of the most funda-
with the correlating line for dw=5.33fim in Fig. 4. mental operating parameters. An apparent activation
This finding indicated that the rate decrease caused by energy of 80kJ/mol was calculated from this plot. As
the increase of surfactant concentration was mainly mentioned previously, the demulsification rate is pro-
because of viscosity rise. portional to the - 3.2 power of the oil-phase viscosity.
2.4 Effect of agitation speed in demulsifier The viscosity varied with temperature and its de-
There existed a linear relationship between the pendence was expressed satisfactorily by the well-
demulsification rates and the agaitation speeds of the known Andrade equation as shown in Fig. 10. To
demulsifier, as shown in Fig. 9. Although the exact evaluate the intrinsic temperature dependence of the
mechanismof the acceleration observed is not clear, demulsification rates where the viscosity effect was
stirring is supposed to increase the probability that separated, the product of the rates and the 3.2power
water drops are brought into the electric field formed of the oil-phase viscosity at each temperature were
by the two electrodes. The dependence on the oil- replotted in the same figure. The resulting tempera-
ture dependence also satisfied the Arrhenius-type
equation having an apparent activation energy of
22.5 kJ/mol. This may express the change of emulsion
stability with temperature rise.
3. Discussion
3.1 Derivation of the correlating equation
The correlating equation of the demulsification
kinetics was derived as follows by summarizing the
experimental results obtained in the preceding
section.
dVw/dt=4AS x \05 Qxp(-22500/RT)fi^-2d3w5
x (6.5+3.93A0exp(- 19.6(/v)
xexp(-0.111Q
Figure ll compares the observed and calculated
demulsification rates. There was relatively good
Fig. 9. Effect of agitation speed on demulsification rate agreement over a wide range of operating condi-

Fig. 10. Temperature dependence of demulsification rate and oil-phase viscosity

VOL. 21 NO. 4 1988 349
 erties other than viscosity could be considered neg-
ligibly small in the present case.
As pointed out in the introduction, a number of
factors affect the demulsification kinetics. It should be
noted that the present equation was obtained for
emulsions whose water and oil phases were respec-
tively a copper sulfate solution and a mixture of
kerosene, liquid paraffin and Span 80. The applica-
bility of the proposed rate equation to emulsions
prepared with other organic solvents, surfactants and
carrier molecules will be discussed in a subsequent
paper. The contribution of demulsifier configuration,
which becomes important in scale-up, will also be
treated.
3.2 Comparison with thermal demulsification
Examination of the temperature effect suggested a
decrease of emulsion stability at high temperature. In
Fig. ll. Comparison of calculated and observed demulsifi- general, the emulsifying ability of surfactants de-
cation rates creases with temperature rise.n) To compare elec-
trostatic demulsification behavior with its thermal
counterpart where demulsification occurs due only to
tions, although some data showed over 50%devia- stability decrease, demulsification without electric
tion. Fujinawaet al. presented a correlation that con- field was carried out at high temperature. Emulsions
sidered the effects of water-drop size, surfactant con- having large water drops were utilized to accelerate
centration, holdup, electrode distance, and applied the demulsification, and the contents were agitated to
voltage.l) Their results concerning the dependence on prevent the creaming of emulsion during long oper-
water drop size and surfactant concentration do not ation time. The thermal demulsification in the range
differ much from those in this study. As for the of 50 to 75°C yielded an apparent activation energy of
holdup effect, however, the exponential decrease about 90kJ/mol, which was close to the 80kJ/mol
shown in Fig. 7 was quite different from their results observed under an electric field. This result means
where the rates decreased moderately. According to that the temperature effect observed under an electric
observations under a microscope, demulsification field indicates a decrease of emulsion stability.
proceeds by way of coalescence and subsequent sed- 3.3 Favorable conditions to accelerate demulsification
imentation of water drops.2'5-12'13* It is not clear
The increase of oil-phase viscosity, which is desir-
which is the rate-controlling process of the demulsifi- able to suppress emulsion breakup in extraction pro-
cation. Therefore, the above-mentioned rate equation cesses,8"^ is unfavorable for demulsification. Two
should be regarded as expressing the overall effect of ways to accelerate demulsification without adding any
various parameters.
organics to decrease the oil-phase viscosity may be
In the present study, the oil-phase composition was suggested by the present study. 1) Operation at high
altered to increase the viscosity. Therefore, it is
temperature is effective as described in the preceding
necessary to confirm that the variation of oil-phase section, although it requires warming. 2) Stirring of
properties other than viscosity is negligible during the the emulsion during operation is another way which is
composition change. The density change was slight simple compared to raising the temperature. Even
since it differed by only about 10% for kerosene and mild stirring results in a marked increase in the rate.
liquid paraffin. The inter facial tension between oil These suggestions maybe useful for continuous ex-
phase and water phase was also found to remain
traction processes with W/Oemulsion.
almost constant. The change of electrical parameters Conclusion
was studied for the dielectric constant since the
electrical force on a water drop is proportional to the The kinetics of electrostatic demulsification was
dielectric constant of the continuous phase.3) The investigated by examining primarily the effects of
dielectric constant of kerosene containing Span 80 is viscosity for each emulsion prepared under various
reported as 2.1-2.2F/m,1) and this is close to that of
conditions, and the following results were obtained.
alkyl naphthenes, which are a major component of 1) Demulsification behavior was affected by both
liquid paraffin. Therefore, the dielectric constant of emulsifying and demulsifying conditions. The de-
oil phase is thought to remain constant. Con-
sequently, the contribution of the oil-phase prop- mulsification rate increased considerably with increas-
ing water-drop size and with decreasing oil-phase
350
JOURNAL OF CHEMICAL ENGINEERING OF JAPAN
viscosity, water-phase holdup and surfactant con- /iOii = oil-phase viscosity [mPaà"s]
(j)w = water-phase holdup [-]
centration. These effects were expressed by power or
exponential functions. Temperature had a pronounc- Literature Cited
ed acceleration effect which could be described by an
Arrhenius-type equation. Stirring of the emulsion 1) Fujinawa, K., M. Morishita, M. Hozawa, N. Imaishi and H.
during the operation was found to promote de- Ino: J. Chem. Eng. Japan, 17, 632 (1984).
2) Ino, H., N. Imaishi, M. Hozawa and K. Fujinawa: Kagaku
mulsification and its contribution was expressed by a Kogaku Ronbunshu, 9, 263 (1983).
linear function. 3) Joos, F. M., R. W. L. Snaddon and N. A. Johnson: IEEE
2) A rate equation composed of suitable functions 4)
Trans. Ind. AppL, IA-21,
Kataoka,
1366
T. and T. Nishiki:
(1985).
/. Chem. Eng. Japan, 12, 16
for each factor showed satisfactory agreement with
(1986).
observed rates over a wide range of experimental 5) Kriechbaumer, A. and R. Marr: "Macro- and Micro-
conditions. emulsions (ACS Symp. Ser. 272, ed. by D.O. Shah)," p. 381,
3) The temperature dependence of thermal de- Am. Chem. Soc, U.S.A. (1985).
mulsification was close to that of electrostatic de- 6) Mori, Y. and W. Eguchi: Proc. World Congr. Ill Chem. Eng.,
8J-157, Tokyo, Japan (1986).
mulsification. Therefore, the acceleration by tempera-
ture rise under an electric field was thought to be 7) Mori, Y., H. Uemae, S. Hibino and W. Eguchi: Kagaku
caused by a decrease of emulsion stability and oil- Kogaku Ronbunshu, 13, 412 (1987).
8) Ohtake, T., T. Hano, K. Takagi and F. Nakashio: J. Chem.
phase viscosity. Eng. Japan, 20, 443 (1987).

9) Ohtake, T., T. Hano, K. Takagi and F. Nakashio: J. Chem.

Acknowledgment Eng. Japan, 21, 272 (1988).

The authors wish to express their thanks to Prof. F. Nakashio of 10) Thien, M. P., T. A. Hatton and D. I. C. Wang: "Separation,
Kyushu University for his helpful suggestions. Thanks are also due Recovery and Purification in Biotechnology (ACS Symp. Ser.
to Mr. Y. Atago, Mr. H. Nakajima, Mrs. Y. Nakayama and Mr. Y. 314,ed. byJ. A. AsenjoandJ. Hong),"p. 67,Am. Chem. Soc,
U.S.A. (1986).
Takeuchi for their experimental assistance.
l) Tsuji, S.: "Nyuuka Kayouka no Gijutsu," Kogaku Tosho
Nomenclature Co., Japan (1976).
12) Yamaguchi, M., A. Kobayashi and T. Katayama: Kagaku
Q
=surfactant concentration [wt%]
dw [/mi] Kogaku Ronbunshu, ll, 599 (1985).
= water drop diameter in W/Oemulsion
N = agitation speed of demulsifier 13) Yamaguchi, M.,
[s"1] A. Kobayashi, K. Ohbori and T. Katayama:
R =gas constant [J/mol - K] Kagaku Kogaku Ronbunshu, ll, 729 (1985).
[K] 14) Yoshizuka, K., K. Kondo and F. Nakashio: /. Chem. Eng.
T =temperature
[s] Japan, 19, 396 (1986).
t =time
Vw volume of water phase separated [cm3]

VOL 21 NO. 4 1988 351