Chairmen: Y. Kawase, L.N. Braginsky


Department of Mechanical Engineering,
Faculty of Engineering,
Himeji, Hyogo 671-22, JAPAN


Water-in-oil (W/O) or oil-in-water (O/W) emulsions are effectively utilized in many industries, including for example, the food, chemical, and pharmaceutical industries, while in the petroleum industry, demulsification is a major activity accounting for hundreds of millions of dollars each year. A number of different methods exist for breaking down petroleum emulsions of which the most common of these employ centrifuges, filters, chemical additives, and electrical treatments, etc. Electrical treatment has been successfully employed for crude oil refining by Cottrell precipitators. The technology offers in low operating and equipment costs and its adaptability to large throughputs.

As industrial application of a rich W/O emulsion, a novel extraction process using emulsion liquid membrane was developed in 1968. Since then, its application to the extraction and concentration of very dilute solutes has received a great deal of attention in waste water treatment, hydrometallurgy, and fermentation etc. An efficient operation for breaking the emulsion is a key step in the entire process. In this step, breakdown of water phase (droplets) containing the valuable substances enriched in the extraction step is necessary, as is full recovery of the oil phase containing emulsifiers and extractants without loss and deterioration, because they are used to remake the emulsion and the entire process is operated continuously. The present study then, is aimed at developing a novel membrane-based phase separator in order to efficiently break up rich W/O emulsions. Essentially, a flow-type column consisting of two perforated parallel electrodes inside the column was constructed to act as the electrostatic phase separator.

The effects of some parameters upon both the separation of W/O emulsions and the performance of the separator were experimentally examined; in particular, we considered various applied electrical voltages, differing diluting oil phases and other contributing operating conditions. The W/O emulsions were prepared by using aqueous solutions of NaCl or CuSO4 in combination with differing organic solvents namely, kerosene, heptane and a mixture of kerosene and carbon tetrachloride. Sobitan mono-oleate (Span 80) was used as the emulsifier. Varying the solvents in this way enabled a study to be made of the effect of correspondingly different dilution properties upon the separation process of the emulsions. Characteristically, the density of the aqueous solution is heavier than that of kerosene and heptane but, lighter than that of the mixed solvent. The water content of the emulsions were prepared to be in the range of 30 to 50 vol% with the Sauter mean diameter of water droplets approximately 2.5 m. The subsequent demulsification performance of the various emulsion combinations was quantified by measuring the rate at which the aqueous phase was resolved. Additionally, the effectiveness of the separation process was evaluated by measuring the residual content of the aqueous phase (in the demulsified oil phase) using a Karl-Fisher moisture- meter.

The demulsification performance of the present column showed good results largely irrespective of either the range of applied voltages or diluting oil phases. The production rate of the aqueous phase, however, did increase correspondingly with an increase in applied voltage but, was otherwise generally high for each of the dc, ac and pulsed dc voltage cases. With characteristics dc electric field strengths in the range of 80 to 200 V/cm, depending upon the physical properties of the emulsions, a stable interface regularly formed between the various emulsions and the demulsified oil phase. Compared with other published data in this area the field strengths quoted in this study represent some of the lowest. Regarding the emulsifiers contained in the demulsified oil phase, it was observed that they suffered little degradation and could be continuously reused without any appreciable deterioration in the efficacy of the oil phase-emulsifier combination. The rate of demulsification itself though was found to vary so that it diminished with decreasing water droplet size, increasing water content in the emulsion and, enhanced emulsifier concentration but, enhanced with the residence time of the emulsion in the electric field and the salinity of the water droplets.

Additional area of investigation, to the aforementioned demulsification properties, concerned the coalescent and dynamic behaviors of the droplets in the emulsion phase whilst being subjected to an electric field. To clarify these aspects, both the distributions of droplet sizes and the holdup fraction of the droplet phase in the emulsion which was demulsifying in the electric fields were measured; during demulsifying, the emulsions were sampled at given positions in the axial direction of the column.

A relation between the rate and electrical energy consumed in the demulsification was examined. A correlation for the dc and pulsed dc fields showed good agreement, but the correlation for the ac field was different from that of the dc fields.


Jason Stoyel *, Terje S°ntvedt **
* Imperial College of Science, Technology and Medicine,
** Norsk Hydro

ABSTRACT. A model has been developed which can be used to characterise the stability of water-in-oil emulsions formed by heavy oils. Its purpose is to transform data from small scale laboratory tests to field scale and may be used to select the size and type of separation equipment required.

The model characterises emulsions by means of four empirical coefficients which are used in a numerical approach to determine the appearance rate of both pure oil and pure water as well as the binary and the drop-homophase coalescence times. The two system parameters that have been investigated are the initial droplet size and the system water cut.

Variations in surfactant concentration at the drop interface, produced by changes in the drop diameter and/or the initial water cut are incorporated in the model. This effect can be considered in systems with or without added demulsifiers when the critical agglomeration concentration of demulsifier, and/or surfactant, is known previously. Two examples from the literature 1,2 and one heavy crude oil system are used to determine the proposed model. Although there are no artificially added demulsifiers the systems can not be considered surfactant free.

Expressions are described for the sedimentation velocity and rate of binary coalescence of drops in the sedimentation zone and for the rate of both interdroplet and drop- homophase coalescence in the dense packed zone.


  1. Barnea E., Mizrahi J., Separation Mechanism of Liquid-Liquid Dispersions in a Deep-Layer Gravity Settler :Part I-IV, Trans. Instn. Chem. Engrs, Vol.54, 1975
  2. Jeelani S.A.K., Pandit A., Hartland S., Factors Affecting the Decay of Batch Liquid-Liquid Dispersions, Canadian Journal of Chemical Engineering, Vol. 68, December, 1990


Dreher, T.M., and Stevens, G.W.
Department of Chemical Engineering, The University of Melbourne
Parkville, Victoria, 3052, Australia


Understanding the factors that influence the coalescence of single drops at an interface in liquid- liquid systems is an important precursor to understanding emulsion stablity and the behaviour of droplet swarms such as occurs in many liquid extraction pocesses. Three techniques have been used to examine the coalescence process of single drops. Atomic force microscopy enables direct measurement of the fundamental forces at a liquid-liquid interface, scanning ellipsometry/interferometry is used to measure the film thickness profile during the drainage process, and single drop coalescence time measurement is used for quantitative comparative studies of effects of various emulsifiers.


Our work has centred on an understanding of the processes that control the drainage of thin films and the role of interfacial forces during the drainage process. We have been applying atomic force microscopy to liquid interfaces (4). This technique, in principle, enables the determination of the fundamental forces between the two surfaces. Typical results for a silica sphere approaching a water decane interface in the presence of 1 m M SDS is shown in Fig. l. The data has been fitted to a constant surface charge model and shown to fit; however, the interpretation of the data is based on immobile changes and no surface deformation. Further work is required in the development of the interpretation of the data before quantitative information about the double layer structure can be obtained ˝om this equipment.

Figure 1. (a) Raw displacement-distance curve for silica-water-n-decane in the presence of I.0 mM SDS and 10 mM NaCl. The oscillations are still present but the approach is repulsive at all separations, and there is no adhesion during retraction in the presence of surfactant. (b) Fit to a force-separation curve at low scan rates using A = 6x10-21 J, (sphere) = -60 mV, (oil) = -30 mV, k-1 = 100 ┼. Interaction under conditions of constant surface charge. The pH was 5.6.

Figure 2. Schematic of Coalescence Apparatus

The addition of long chain polymers to the continuous organic phase affects the rate of coalescence by modifying the rheology of the continuous phase. This in turn affects the performance of extraction equipment, it can also be used in some cases to stabilise an emulsion.


Single droplet coalescence experiments were carried out in the apparatus illustrated in Figure 2.


The Newtonian fluids consisted of n-decane (>99% pure, Sigma-Aldrich) and Hyvis 3 polybutene fluid (BP Chemicals) mixed in various proportions to give a viscosity range between 2x10-3 and 0.70Pas. Small quantities of PIB (Exxon Vistanex 140-MML polyisobutylene, MW=2,100,000) were added to some of the Newtonian fluids to make them elastic. Water distilled in an all glass apparatus () was used as the aqueous phase.

Median Coalescence Time

The median drop coalescence time increased linearly with the continuous phase viscosity. The addition of polymer to the continuous phase had little effect, within experimental error, on the coalescence time of the drops with a diameter of approximately 2.5mm. Linear regressions through the data resulted in .

Coalescence time for the experiments conducted with drops of diameter approximately 0.3mm are given in Table 1. It was obvious for small drops, that elasticity significantly increased the coalescence time. This was due to the small drops experiencing high enough shear rates, in the process of coalescence, that the non-Newtonian fluid exhibited some elastic characteristics.

Rheological Properties and Coalescence Time for 90% of 100 Drops (d=0.3mm) to Coalesce
Fluid (mPas) (ms)

time (s)
20oC 30oC 20oC 30oC
N 294.3 174.9 -- -- 851.2 30.78 685
O 304.8 169.1 33.10 4.11 847.1 32.07 1940


The results indicated that increased emulsion stability may be achieved by the addition of a suitable polymer to the continuous phase. The extent of stabilisation was dependent on the rheological properties of the polymer solution, and the dispersed phase drop size. A method has been proposed which enables an estimation of when elastic effects become significant. This relies on a knowledge of the physical and rheological properties of the continuous and dispersed phases, in particular the Trouton ratio, of the continuous phase.


This work was funded by City West Water Ltd., the Advanced Mineral Products Special Research Centre, and the Australian Research Council. The polybutene fluids were supplied by BP Chemicals Ltd.


d drop diameter m
m, m' power law parameters Nm-2s-n, n'
N number of drops -
n, n' power law slopes -
u velocity ms-1
interfacial tension Nm-1
relaxation time s
continuous and dispersed phase viscosity Pas
density kgm-3
time taken for half the drops to coalesce, half life s


  1. Garg, G., and Pratt, H.R.C., AIChE J. 30, 432 (1984).
  2. Hamilton, J.A., and Pratt, H.R.C., AIChEJ. 30, 442 (1984).
  3. Jeffreys, G.V., Davies, G.A., and Pitt, K., AIChEJ.16, 823 (1970).
  4. Mulvaney, P., Perera, J.M., Biggs, S., Grieser, F. and Stevens, G.W., J.Coll. and Int. Sci., 183, 614,1996.

Future Meetings | Past Meetings | Proceedings on Sale | Related Links

ICHMT World Wide Web Administrator