当前位置:首页 >> 能源/化工 >>

Structure and stability of colloidal liquid aphrons

Colloids and Surfaces A: Physicochemical and Engineering Aspects 131 ( 1998 ) 119 136 SURFACES


Structure and stability of colloidal liq

uid aphrons
G.J. Lye l, D.C. Stuckey *
Department of Chemical Engineering & Chemical Technology, Imperial College g/Science, Technology & Medici, c. London SW7 2BY, UK
Received 5 July 1996


The structure and stability of colloidal liquid aphrons (CLAs) have been investigated using a variety of experimental techniques, i.e. cryo-TEM, DSC, and light scattering. The findings support the structural model proposed by Sebba who suggested that polyaphron phases resemble a biliquid foam while the individual CLAs, when dispersed in a continuous aqueous phase, consist of spherical, micron-sized, oil droplets surrounded by a thin, aqueous "soapyshell". First-order half-lives of CLAs dispersed in a stirred vessel over a range of continuous phase ionic strengths, pH, and temperatures were also determined for the first time. These allowed quantitative comparison of CLA stability when dispersed under various conditions and of the influence of including various concentrations of lipase or erythromycin-A in the aphron formulation. Based on these results, a mechanism for the breakdown of dispersed CLA structure is proposed which involves destabilisation and loss of the "soapy-shell" followed by coalescence of the oil cores of the aphrons. However, direct evidence for the structure of the surfactant-laden interfaces responsible for the stabilisation of aphrons is still required if the structural model proposed by Sebba is to be fully confirmed. The similarities and differences between CLAs and high-internal-phase-ratio emulsions (HIPREst are also discussed. ~ 1998 Elsevier Science B.V.

Keywords: Aphron; Colloidal liquid aphron; Differential scanning calorimetry; High-internal-phase-ratio emulsion; Stability; Transmission electron microscopy

1. Introduction

Colloidal liquid aphrons (CLAs) are postulated to consist of a micron-sized solvent droplet encapsulated in a thin aqueous film ("soapy-shell"). This structure, as shown in Fig. 1, is stabilised by a mixture of non-ionic and ionic surfactants [1]. Since Sebba's original reports on biliquid foams [2] and subsequently "minute oil droplets encapsu* Corresponding author. ~Present address: The Advanced Centre for Biochemical Engineering, Department of Chemical and Biochemical Engineering, University College London, London WC1E 7JE, UK. 0927-7757 98/$19.00 <~',1998 Elsevier Science B.V. All rights reserved. P H S0927-77571 9 6 ) 0 3 9 3 1 - 3

lated in a water film" [3], these structures have been investigated for use in pre-dispersed solvent extraction processes. Because o f a favourable partition coefficient for non-polar solutes between the oil core of a CEA and a dilute aqueous solution, and their huge interfacial area per unit volume ( ~ 15 000 m 2 11"1 3), aphrons have been successfully applied to the extraction of antibiotics [4,5] and organic pollutants such as dichlorobenzene [6] and 3,4-dichloroaniline [7]. Similarly, colloidal gas aphrons (CGAs) have been investigated for use in flotation processes for the clarification of various biological media [6,8,9], the removal of organic pollutants and dyes [10 13] and also heavy metals [14] from aqueous solution.


G.J. Lye, D. C Stuck~:v / Colloids Surfaces A: Physicochem. Eng. Aspects 131 (1998) 119 136



," 9-

Ionic surfaotant



" "".,d'~l ~ ~ ~ . , ' ~ D i f f u s ? electrical " -~-~_6"..-~ ~ > - double layer Fig. 1. Proposed structure of a single colloidal liquid aphron (CLA) when dispersed in a continuous aqueous phase (redrawn from Sebba [ 1] ).

Although potential applications for CLAs have been investigated, little work has been carried out to either confirm or refute Sebba's proposed structures for polyaphron and CLA phases. Note the terminology that will be used in this paper. Upon aphron manufacture, the initial creamy-white phase consists of an aggregate of individual aphrons having a structure resembling that of a biliquid foam [1,2]: this is termed a "polyaphron" and data on the structure and stability of this phase are obtained when the polyaphron is not dispersed in an aqueous phase. Upon dispersion of a polyaphron in a continuous aqueous phase, however, the individual aphrons become separated to form the spherical droplets shown in Fig. 1 [1]; these will be termed "CLAs". Obviously the properties of the polyaphron will depend on the aphron formulation and method of manufacture [1,15], while the properties of the CLAs will, in addition, be influenced by the nature of the continuous phase in which they are dispersed. Sebba originally used fluorescent dyes to probe the structure of polyaphrons and CLAs [1], though transmission electron microscopy (TEM) might also be used to make size measurements of the various structural domains. This method is advantageous in that it does not require staining of the phases with compounds which may alter the surface properties of the system and whose location in the polyaphrons

may not be clear. Differential scanning calorimetry (DSC) is another technique which might fruitfully be applied to polyaphron phases since the various organic and aqueous domains of the polyaphron could be predicted to have different transition temperatures. DSC traces could thus be used to determine the "mixedness" of these two components. Finally, light scattering has previously been used to determine the size of dispersed CLAs and CGAs [ 15, 16 ], although it is not applicable to the study of polyaphron phases due to their opaque nature. Previous reports on the stability of aphrons have focused on the stability of isolated polyaphron phases (rather than dispersed CLAs or CGAs) by measuring the rate of creaming, i.e. the rate of separation of the dispersed solvent phase from the polyaphron. For gas polyaphrons, Matsushita et al. [15] studied the effect of surfactant type and concentration and stirring speed and duration on CGA half-life, as did Save and Pangarkar [17] who also looked at the effect of pH and additives such as enzymes, polymers, solvents and salts. In both cases the half-lives of these gas polyaphrons were of the order of 10 min or less. Matsushita et al. [15] and Sebba [1] also studied the influence of some of the above factors on the stability of liquid aphrons and found that stable polyaphron formulations can have half-lives of the order of months or even years. While such studies have provided useful information on factors influencing the formulation of stable polyaphrons, little is known about the stability of CGAs or CLAs when dispersed in various continuous phases. Obviously, this situation is more realistic in terms of adsorption/extraction processes using aphrons and will be vital for the design of suitable contacting equipment [ 1,8]. Recently, Chaphalkar et al. [ 16] have applied a light scattering technique to investigate the influence of surfactant type and concentration, and NaC1 concentration in the continuous phase on the stability of dispersed CGAs. Analysis of their data suggests dispersed CGA half-lives vary between 4 and 10 min. We have applied a similar technique to the study of dispersed CLAs [7] finding that, as in the case of liquid polyaphrons, the most stable CLAs are formed from relatively non-polar solvents (decanol > octanol >

G..L Lye. D. C Stuckey / Colloids Surfaces A." Physicochum. Erie. Asl,ects 131 f1998) 119 136


butyl acetate) and non-ionic surfactants having high HLB numbers (Atlas G1300>Softanol 120 > Softanol 30). In this investigation we used the techniques of cryo-ultramicrotome T E M and DSC to analyse liquid polyaphrons in order to test Sebba's proposed structure for these phases. Results from these methods were also compared with those obtained by light scattering of dispersed CLAs in order to see if any correlation existed. In addition. we extended our initial investigation on dispersed CLA stability to study the influence of continuous phase properties on CLA half-lives, and see to what extent the data could be used to further elucidate the structure of CLAs. Some authors, noticeably Princen [8], refute Sebba's proposed structures for polyaphron phases and CLAs claiming that they are no different from those of highinternal-phase-ratio emulsions, Throughout this work, therefore, our results on the aphron systems under investigation are compared and contrasted with those published for HIPREs. Finally, these are discussed and suggestions are made which may clari~ the current confusion.

aqueous phase was typically 1.5 ml and the organic phase was added at an average flow rate of ~ 0 . 5 m l m i n -x until the desired phase volume ratio was reached (PVR= l';rg/l'~q). The organic phase initially disperses easily and can be added at a higher rate, but as the maximum PI"R (PgRma x) is approached [15], the mixture becomes extremely viscous and the rate of addition must be reduced. The polyaphron phases formed in this way had a creamy-white appearance and showed no signs of phase separation over a period of months. When used during aphron formulation, erythromycin was dissolved in the organic phase whilst lipase was dissolved in the aqueous surfactant solution at the desired concentrations. The influence of various solvents, surfactants and formulation procedures on P l/'Rmax and the stability of the polyaphrons have been published previously [15].

2.1. Crvo-ultramicrotomv and eh, ctron microsc?q~y of polyuphrons
Small droplets of the polyaphron phase ( ~ 2 ktl ) were rapidly frozen by plunging them into liquid propane at - 180C (Reichert Jung KF80 immersion cryofixation system). These were subsequently kept in liquid nitrogen before sections, approximately 0.2 jam thick, were cut at --156'C and collected on l\~rmvar-coated copper grids ( Reichert Jung FC40 ultracut system). These were then examined directly in a Philips 400 TEM with cryostage at - 1 6 5 ' C : surface ice. caused by grid transfer, was sublimed at - 1 0 0 C. Initially, there was little contrast between the aqueous and organic areas of the polyaphron. However. it was found that by sublimation at - 8 0 C the water present in the plateau borders of the polyaphron could be removed with no noticeable effect on the organic areas to allow easy differentiation of the lwo domains,

2. Materials and methods

The solvents used in this work, decan-l-ol and decane, were purchased from Aldrich and were >99% pure. The anionic surfactant used was sodium dodecyl sulphate (99%, Sigma) whilst the non-ionic surfactants, Softanol 30 and Softanol 120, were of the alcohol ethoxylate type (Honeywill & Stein) and were used as supplied. The non-ionics had EO mole numbers of 3 and 12, respectively. The chloride salts of various cations and glycerol were AnalR grade from Merck. Lipase from CandMa cvlindracea (Type VII) and erythromycin-A (98% pure) were from Sigma. Deionised water from a Purite RO50HP unit had a conductivity of < 0. I laS c m - 1 and was filtered through a 0.2 lam filter. Polyaphron phases were prepared by dropping the organic phase ( 1% (w/v) Softanol in the desired solvent) from a burette into a well-stirred ( ~ 800 rev min- ~), foaming aqueous solution containing 0.5% (w/v) SDS. The initial volume of the

2.2. D((/brential scanning calorhm, trv ~/ polyal)hrons
Freezing exotherms of polyaphron phases were obtained using a Perkin Elmer DSC-2 differential scanning calorimeter al a cooling rate of


G.J, Lye, D.C. Stuckev / Colloids Surfaces A: Physicochon. Eng Aspects 131 (1998) 119-136

5 K min -1, and a scanning range of 220-305 K. Measurements were made under a dry nitrogen atmosphere to prevent condensation of water vapour. Between 2 and 5 mg of the polyaphron phase was sealed in an aluminium pan, an empty sealed pan being used as the reference material. The error in the determination of the onset of the transition temperature was + 0.6 K, whilst that for the heat of phase transition (based upon the peak area) was _+7%.

When polyaphrons were dispersed under conditions which made them unstable, a clear solvent layer could be seen to develop on the surface of the bath. In all cases it was ensured that the speed of the impeller was sufficiently low to prevent re-entrainment of this solvent (or air) back into the dispersion, and hence the sample cell, as this could interfere with the size determination of the CLAs remaining in suspension.

2.3. Light scattering of dispersed CLAs
3. Results and discussion

Measurements were made using a Malvern 2600 (Model 3.0) particle size analyser. For a single determination of particle size, the instrument was fitted with a 15 ml sample chamber containing deionised water. This cell could be mixed by a small magnetic bar to ensure that the polyaphron phase was properly dispersed into CLAs and a measurement was taken within approximately 30 s of dispersion. Results for the CLA diameter, dov, are given here as the Sauter mean diameter which is a measure of the ratio of the total volume of particles to the total surface area. The error in the determination of dov was _+5%. Various routines were available in the software for analysing the scattering pattern obtained from the CLAs. The one which gave the best correlation between experimental and theoretical scattering patterns was that which assumed the sample consisted of spherical particles having a mono-modal size distribution. For determining the stability of the dispersed CLAs, the instrument was fitted with a flowthrough sample cell, and was programmed to take readings over a period of 40 rain. The sample cell was connected to an external stirred vessel in which the temperature (+0. I°C) and pH (by addition of 0.1 M NaOH or HC1) could be controlled, the dispersion being continually passed between the two at a flow rate of ~ 5 0 0 m l m i n - ' using a centrifugal pump. A schematic diagram of this equipment is shown in Fig. 2. In a typical experiment, 0.1 ml of a polyaphron phase was dispersed in 400 ml of the desired continuous phase, creaming of the dispersed CLAs to the surface being prevented by a Rushton-type impeller located near the base of the bath and operated at low speed.

3.1. Cryo-ultramicrotomy and electron microscopy of polyaphrons
A photomicrograph of a polyaphron phase prepared at PVR 4 is shown in Fig. 3. This is analogous to a cross-section through a normal gas-liquid foam. In the case of polyaphrons, however, their creamy-white appearance and the small size of the oil cores makes direct visual observation of this structure impossible. The dark areas of the photomicrograph represent the decanol cores of the polyaphron, whilst the white areas are those from which the water in the plateau borders was sublimed; white areas within the decanol cores are due to expansion of the underlying grid during sublimation causing the samples to "tear". Measurements taken from photomicrographs for samples of various PVR, and also those made by light scattering for polyaphrons dispersed in deionised water, are presented in Table 1. The agreement between the aphron diameters, dov, obtained from the two techniques is good, and the increase in dov with decreasing P VR is consistent with previous work [15]. These sizes are similar to those reported for the dispersed-phase oil droplets of HIPREs [19]. The advantage of the cryo-TEM technique is that it has allowed the thickness of the "soapy-shell" to be measured directly for the first time. For polyaphrons at P VR 10, the measured "soapy-shell" thickness of 0.03 gm is very similar to the minimum thickness predicted by Sebba of 10 n m [ 1]. The shape of the dispersed aphrons (CLAs) as determined by analysis of the scattering pattern

G.J. Lye, D.C. Stuckey / Colloids Surfaces A: Physicochem. Eng. Aspects 131 (1998) l 19 136


ooO oooOoa oo°o ooo2oo20
o owe

Or 0


O.oO.O z : : .o~.o .opo.%o o:o0:

0o°oO ~%o OoO . . . .




Unstable CLAs coalesce and rise to the surface of the stirred vessel. This results in an increase in droplet size and a decrease in the 'bulk' droplet volume concentration as measured by the Malvern detector.



/ /


/ " / / / / /




/ / .

/ J
. ?/ .. <



..... q


Fig. 2. Schematic diagram of the equipment used for determining the stability of dispersed CLAs and the coalescence mechanism responsible for the results obtained in this work.

detected by the Malvern was found to be spherical in all cases. Thermodynamically this would be expected since spheres have the minimum surface area to volume ratio, and hence give systems with the lowest free energy. The polyaphron photomicrographs show a progression from polyhedral structures at P V R 10 to ovoids as the P V R decreases, and ultimately to spheres as the aphrons are dispersed [ 1]. The elongation in the cryo-TEM polyaphron samples to give, for example, ovoids at P VR 4 could possibly also be one of the recognised artefacts to cryo-ultramicrotomy, namely elongation along the plane in which the sample is sectioned? The photomicrographs also

suggest a wider size distribution of the oil cores at lower PVR. This is again consistent with the light scattering data where, with decreasing PVR, it was found that the range of CLA sizes increased, and the size distribution became skewed to larger sizes as shown in Fig. 4. HIPREs are also known to become increasingly monodispersed at larger values of q< the dispersed-phase volume fraction [19]. Polyhedra will only form at high P VR where the volume fraction of the internal phase is greater than that which could be accommodated by closepacked spheres, i.e. ? =0.74. This onset of polyhedra formation has been suggested to occur, and


G.J. Lye, D.C. Stuckev / Colloids Surfaces A: Physicochem. Eng. Aspects 131 (1998) 119-136

Fig. 3. Photomicrographof a cryo-ultramicrotomesection of a polyaphron phase prepared to PVR 4. Aphrons formulated from 1% (w/v) Softanol 120 in decanol and 0.5% (w/v) SDS in deionised water. indeed does, between P VR 3 and 4 [1]. For the P V R l0 sample, where q~=0.91, the photomicrographs exhibit tetrakaidecahedron ( T K D H ) packing. Based on previous geometric calculations for HIPREs [20] it would be expected that rhomboidal dodecahedron ( R D H ) packing would be observed for ~b= 0.74-0.94 followed by T K D H packing for ~b>0.94. However, these calculations assumed a monodispersed system, and there may be small errors in the volumes of the two liquid phases from which the aphrons were formulated. These factors could account for the appearance of T K D H packing at q~<0.94 in polyaphron phases. The three-dimensional packing present in polyaphrons has also been observed using freeze-fracture electron microscopy [21].

G.J, Lye, D.C. Stuck~T / Colloids Surfaces A: Physicochem, Eng. AsTwCts 131 (1998) 119 136


Table 1 Comparison of the sizes and shapes of polyaphrons as determined by electron microscopy with those of dispersed ('LAs as determined by light scattering Initial P I'R Cryo-TEM (polyaphron) d~,,l I~.tml 10 4 3.5 4.6 1.5 5.5 Shell thickness (pm) 0.03 0.15 rib,. (r-tml 3.6 4.7 1.8 5.8 Shape Poh, hedral Ow,id Light scatlering (CLA) d~, t laml 4.0 7.11 Shupc Spherical Spherical

"Represents mean diameter of oil measured for major and minor axes. bRepresents overall diameter of oil core and soapy shell. Polyaphron phases were aged for at least 24 h prior to sectioning and were formulated as in Fig. 3.

3.2. D(~'erential scanning calorimeto, o[ polyaphrons
Freezing exotherms of polyaphron phases at various PVR are shown in Fig. 5 between 243 and 283 K; no other phase transitions were recorded over the range 220 305 K. Traces on the left represent polyaphron phases approximately 24 h after being formulated and, for polyaphrons stored at room temperature, were reproducible over a period of days. This suggests that no significant change in the microstructure of the polyaphron had occurred after the first 24 h. Prior to dispersion, the traces show distinct first-order crystallisation transitions for both water and decanol domains: when determined independently, water (+0,5% (w/v) SDS) froze at 258 K whilst decanol (+1.0% (w/v) Softanol 120) froze at 275 K, thus allowing identification of the individual peaks. The onset of the water freezing peak at 258 K (for PI'R 2) is similar to that found for the freezing of continuous-phase water in a water-oil water emulsion at 257 K [22]. Note that, due to the biliquid foam structure proposed for polyaphron phases [1], water is expected to remain as the continuous phase even at PVR as high as 20 (4,=0.95). Visual observation of the freezing of liquid polyaphron phases to 253 K in the laboratory freezer indicated that the structure of the polyaphron was maintained during the process. However, when left to defrost at room temperature considerable phase separation was observed. In the case of polyaphrons formulated from 1% (w/v) Softanol 30 in decane and 0.5% (w/v) SDS in

water, this phase separation went to completion, while for the Softanol 120-decanol polyaphrons used in the DSC experiments, only 10-20 vol.% of the polyaphron structure was preserved. Because of this extensive phase separation during defrosting, polyaphron phases were not subjected to repeated cooling and heating cycles in the calorimeter. The fact that at least two separate phase transitions are seen in the polyaphron samples indicates that the aqueous and organic domains are not mixed at a molecular level. The appearance of a second decanol freezing peak at PI"R 4 8 could be indicative of solvent molecules present in two distinct environments. These presumably are "bulk" decanol, located in the core of the aphrons, and "interfacial" decanol located at the inner oil core '~soapy-shell" interface of the aphrons. This interface is stabilised by the adsorption of the nonionic surfactant used for aphron formulation which would be expected to significantly alter the colligatire properties of the decanol in the immediate vicinity. For heterogeneous nucleation processes, such as those occurring at interfaces, transition temperatures will be shifted to lower temperatures [23,24], hence the appearance of this "'interfacial'" decanol peak at temperatures below the "'bulk'" decanol peak at 268 K in the P VR 2 sample. For the polyaphron sample at PI'R 10, only a single phase transition was found over the temperature range investigated. At this P I"R the cryo-TEM data showed that the thickness of the "'soapyshell" was very small compared to the diameter of the oil core. The absence of a freezing peak for water could thus be due to the polyaphron behav-


G.J, Lye, D.C. Stuckey / Colloids SurJaces A: Physicochem. Eng. Aspects 131 (1998) 119-136

30 25
"o t(o .Q .c E o >

PVR I I 10 2

20 15 10 5
0 t____._.__
<1.2 1.2 1.$ 1.9 2.4 3 3.9 5 6.4 8.2 10.5 13.6 17.7 23.7 33.7 54.9<118.4

Size band (microns)
40 35 30
oC. 25

Time (Mins)

11 17

40 2O


E o >


<1.2 1.2 1.5

1.9 2.4 3 3.9 S 6.4 8.2 t0.8

13.6 17.7 23.7 33.7 54.9<113.4

Size band (microns)

Fig. 4. Variation of CLA sizedistribution with (top) P VR after dispersion in water and (bottom) time after dispersion in 0.3 M NaC1, pH 7. Aphron formulation as in Fig. 3. ing as a single phase or because the water in the sample did not freeze over the temperature range investigated. This latter point could again be due to a large alteration of the colligative properties of the water when in such close proximity to the number of charged interfaces postulated to be present in polyaphrons and CLAs [ 1]. For polyaphron phases reformed by creaming after dispersion in deionised water (50 wt.%) the freezing exotherms, which are shown on the right in Fig. 5, clearly indicate a change in the structure of the polyaphrons compared to the initial traces. Qualitatively, the transitions for the reformed polyaphrons resemble those for polyaphrons with an initial P V R of 2-4. Assuming that the heat of crystallisation (AHc) of the "interfacial" and

G.J. Lye, D. C Stuckey / Colloids Surfaces A: Physicochem. Eng. Aspects 131 (1998) 119 136



Before dispersion

Initial PVR


After dispersion






| -30

i -20

I -lO

l 0

I I0

I -30

I -20

I -I0

I 0

I lO

Temperature (°C)



Fig. 5. Freezing exotherms of polyaphron phases at various PVR before and after dispersion ( 50 wt.%) in deionised water. Aphrons formulated from 1% (w/v) Softanol 120 in decanol and 0.5% (w/v) SDS in deionised water.

"bulk" decanol peaks are similar, then it is possible to correlate the initial polyaphron P V R with the total decanol peak areas. These gave values of A H c between - 72 and - 204 J g 1 with increasing PVR, and are higher than values reported for the freezing of oil droplets in a 20wt.%

n-hexadecane-octadecane-Tween 20-water emulsion of - 3 1 to - 4 7 J g 1 [25]. Using this correlation of AHc with the initial P VR, Fig. 6 shows the estimated equilibrium P V R for the polyaphron phases reformed after dispersion. Allowing for experimental error, the dispersed polyaphrons


G.J. Lye D. C. Stuckey / Colloids Surfaces A: Physicochem. Eng. Aspects 131 (1998) 119-136



8 6


0 0


4 6 8 Initial polyaphron PVR



Fig. 6. Estimated equilibrium polyaphron PVR after dispersion in deionised water (50 wt.%); solid line represents initial PVR before dispersion. Aphron formulation as in Fig. 5.

Sebba proposed to be responsible for the stability of CLAs. The only evidence for these interfaces comes from the appearance of various decanol freezing peaks in the DSC scans. With the light scattering technique, it was also found that when CLAs were dispersed in 0.3 M NaC1 rather than deionised water (Fig. 4) the mean size of the aphrons was destroyed. A similar effect has been reported for dispersed gas aphrons formulated from both cationic (hexadecyltrimethylammonium bromide) and anionic (sodium dodecyl benzene sulphate) surfactants [13]. This suggests that by studying the effect of the continuous phase properties on the stability of dispersed CLAs, further information may be inferred regarding the properties of the surfactant-laden interfaces present in CLAs.

appear to reform with an equilibrium PVR of 2.8_+0.7 regardless of their initial PVR. This corresponds to ~b~0.74, but the coincidence of this value of ~bwith that found for close-packed spheres is purely coincidental since photomicrographs showed the oil cores of polyaphron phases to have a wide size distribution at this P VR. The reformed polyaphron phases thus contain more water than they did prior to dispersion which would tend to increase the average film thickness between the oil cores of the polyaphron. The formation of HIPREs by creaming as a function of NaC1 concentration in the continuous phase has previously been reported [26]. Although working with different solvent/surfactant systems, the measured values of q~ in the absence of added electrolyte are very similar. It would thus be interesting to determine, and then compare, the relationship between NaC1 concentration and ~bfor aphron formulations with that found for HIPREs. The application of cryo-TEM and DSC to polyaphron phases, and of light scattering to dispersed CLAs, appears to confirm the macroscopic structures of these phases as first proposed by Sebba [1]. The photomicrographs shown in Fig. 3, and the DSC traces shown in Fig. 5, clearly identify the existence of an oil core droplet surrounded by an aqueous film. Unfortunately, with the TEM technique it was not possible to "stain" the surfactants present and hence determine anything about the structure of the various surfactant layers that

3.3. Stability of disTersed CLAs
The effect of NaC1 concentration in the continuous phase on the mean CLA diameter, dov, with time is shown in Fig. 7 together with data for CLAs dispersed in deionised water. As was suggested by the changing size distributions shown in Fig. 4, when NaC1 is present the mean size of the CLAs increases with time, this process being faster at higher salt concentrations. This is in contrast to the case of CLAs dispersed in deionised water where dov does not change significantly over the course of the experiment. Also shown in Fig. 7 is the simultaneous variation in the normalised CLA volume concentration, V?, with time. The parameter Vc represents the number of scattering particles in the laser beam, and can be used as a surrogate measure of the droplet (CLA) concentration or "sample loading" [16]. Similarly to when this technique was used with CGAs [16], it was found that the initial measurement of V~ increased linearly with the amount of CLAs added to the circulating continuous phase, the upper limit being set by saturation of the Malvern detector. In the experiments reported here V~ typically had an initial value of 0.03% after dispersion of the polyaphron phases. Given the data presented in Fig. 7, it is suggested that with increasing salt concentration the CLAs become unstable leading to coalescence of the oil

G.J. Lye. D.C. Stuckey / Colloids Surfaces A: Physicochem. Eng. A,s7~ects 131 (1998) 119 13(~


8 .o l I 6 l l ?


speed sufficiently low to maintain stable CLAs in dispersion whilst not re-entraining the observed solvent layer (or air) back into the dispersion. The kinetics of this proposed coalescence process, as indicated in Fig. 7 by the decrease in the volume concentration of CLAs, Vc, with time, have been found to follow pseudo-first-order reaction kinetics --=exp(-kt)
I~(0) l~(t)

q 2
0 5










Time (Mins)

In 2
ill 2 =

0.8 l?

e 1 ?




(2 )


oo °' 1
0.5 ~ 0.2 0


ii A l&A

0 o




? ?

(M) 0
0.05 0.1 0.3

? ? ?

? ?











T i m e (Mins)

Fig. 7. Variation of (top) mean diameter and (bottoml normalised volume concentration with time for CLAs dispersed in continuous phases of various ionic strengths at pH 7. Aphrons formulated from 1% (w/v) Softanol 120 in decanol and 0.5% (w/v) SDS in deionised water to an initial P V R of 9.

where ["~ is the volume concentration of CLAs, k is a first-order rate constant, t is the time elapsed since the polyaphron phase was initially dispersed, and tl/e is the mean CLA half-life. The 1~ vs. time curves do represent a first-order process since, as shown in Fig. 8, a plot of ln[f~(t)/l'c(0)] against time yields a straight line. In addition, the calculated values of tl/2 were found to be independent of the initial droplet concentration, I~(0), over the range of concentrations permitted by the instrumentation. If this process results from a pseudo-chemical reaction between surfactant molecules and salt ions, and it is assumed to be firstorder with respect to each, then an overall pseudofirst-order process would indeed be expected given

cores of the droplets upon collision (the reasons for this will be discussed in the following section). This would account for the increase in dov with time. In addition, the larger oil droplets would rise to the surface of the external sample bath more rapidly, thereby reducing the number of droplets being circulated through the sample cell, and hence reducing the measured volume concentration of the dispersed phase, l~c. During experiments at high NaC1 concentrations, it was observed that a clear solvent layer developed on the surface of the external bath, and that the turbidity of the dispersion being circulated round the system visibly decreased. Note that the decrease in V~ with time does not simply represent the rise of stable CLAs to the surface of the external vessel, i.e. creaming, as this was fitted with an impellar operated at a












Time (Mins)

Fig. 8. Plot of first-order reaction kinetics for the coalescence of CLAs with time: solid lines fitted by linear regression. Data and aphron formulation as in Fig. 7.


G.J. Lye, D.C. Stuckey / Colloids Surfaces A: Physicochem. Eng. Aspects 131 (1998) 119-136

the great excess of salt present under the experimental conditions. Replicate experiments for polyaphrons dispersed under various conditions indicated that the maximum error in the determination of tl/2 values was less than 15%. Using experimentally determined values of tl/2 it is now possible to determine quantitatively the influence of the continuous phase properties on the stability of dispersed CLAs.

3.4. Effect of ionic strength and cation type
The effect of NaC1 concentration in the continuous phase on the half-lives of dispersed CLAs at 25°C is shown in Fig. 9. Compared to CLAs dispersed in deionised water, which have a halflife of approximately 60 min, it is seen that values of tl/2 decrease with increasing ionic strength reaching a constant value of around 15 min at ionic strengths greater than 0.3 M. The data presented by Chaphalkar et al. [16] for CGAs dispersed in 3.4 mM NaC1 also indicate a decrease in tl/2 for CGAs formulated from cationic or anionic surfactants, but no significant change for those formulated from a non-ionic surfactant. Their tl/z values are estimated to be around 4 min which indicates that dispersed CGAs are considerably less stable than the CLAs investigated here. Addition of salts to gas or liquid polyaphrons, especially those of polyvalent ions, has previously

70 60

? ?

t0 0 0 0.'2 0.'4 0.'6 0.'8 Ionic strength (M) 1'

Fig. 9. Effect of continuous phase ionic strength (NaC1) on the half-life of CLAs dispersed at pH 7. Aphron formulation as in Fig. 7.

been shown to reduce the stability of non-dispersed polyaphron phases, and even break them [1,17]. According to Sebba's proposed structure for CLAs [1] as shown in Fig. 1, the outer interface of the "soapy-shell" is stabilised by a bilayer of ionic surfactant molecules. These create an electrical double-layer, due to the charges on the surfactant head groups, which together with the Gibbs elasticity of this thin liquid film is responsible for the greater stability of dispersed CLAs compared to normal emulsion droplets. The electrostatic repulsion of the highly charged droplets [12], and the mechanical stability of the liquid film, prevent CLA coalescence. Upon the addition of salt to the dispersion these electrical double-layers are compressed, hence the CLAs formed are initially smaller than those dispersed in deionised water (see initial dov data in Fig. 7 for CLAs, and those of Ref. [16] for CGAs). However, with increasing salt concentration the electrical double-layer is increasingly destabilised and we suggest that this causes the "soapy-shell" around the oil core to be stripped offthe CLAs. This would therefore reduce the energy barrier to droplet coalescence, allowing the decanol cores of the aphrons to fuse upon collision and thus leading to the type of process indicated by the data in Fig. 7. For "soapy-shells" stabilised by anionic or cationic surfactants, this mechanism would be expected to show a strong dependence on continuous phase ionic strength. Whilst the outer interface of the "soapy-shell" is comprised of ionic surfactants, that located at the inner decanol-"soapy-shell" interface is primarily stabilised by the oil soluble, non-ionic surfactant used for polyaphron formulation (Softanol 120 in this case). Increasing ionic strength would be predicted to have less influence on this interface than on the outer charged interface. This may be the reason for the relatively constant value of tu2 found at ionic strengths greater than 0.3 M, where the outer surfactant interface and "soapy-shell" would have been stripped off the CLAs very rapidly. This would also explain the insignificant effect of 3.4 mM NaCI on the half-lives of CGAs formulated from non-ionic surfactants [16]. Given the strong influence of ionic strength on q/2 as described above, it would be expected that the nature of the cation would also be important

G.J. Lye. D. C Stuckev / Colloids Surfaces A: Physieochem. Eng. Aspects 131 (1998) 119 136


due to their selective interactions with the electrical double-layers. The influence of various cation types in the continuous phase on the half-lives of CLAs dispersed at 25'~'C is shown in Table 2. Except for phases comprised of sodium ions (which happen to be the surfactant counter-ions in this case) the CLAs appear to be more stable in continuous phases of divalent cations rather than monovalent ones. This behaviour is difficult to interpret: divalent ions, due to their smaller hydrated radius and higher charge density, may be expected to exchange with the surfactant counter-ions leading to greater compression of the electrical double-layers and hence smaller CLAs. This, however, does not appear to be the case. The comparatively high stability of CLAs dispersed in continuous phases comprised of the surfactant counter-ion could be due to the formation of a relatively stable interface between the two phases across which there are no exchange or mass transfer processes occurring. Such processes are known to create local variations of surfactant concentration, and hence interfacial tension, which can destabilise the interface due to spontaneous interfacial convection [27].

3.5. Effect of pH
Since the stability of CLAs displays a strong dependence on ionic strength, due to electrostatic interactions associated with the surfactant headgroups, pH should also have an influence on CLA
Table 2 Effect of continuous phase cation type on the half-life of dispersed C L A s Continuot, s phase cation type Monovalent Li Na K Divalent Mg Ca 17.8 26.7 17.8 26.7 24.8 3.8 3.8 3.8 4.0 4.1

half-lives. The effect of continuous phase pH on the stability of CLAs dispersed in deionised water at 25~C is shown in Fig. 10. Above a pH of 6-7, tl/2 values are essentially constant, but begin to decline as the continuous phase becomes more acidic. At low pH values, the excess hydrogen ion concentration leads to protonation of the sulphonate head-groups of the SDS molecules located at the outer "soapy-shell" interface. This would have two effects on aphron stability. Firstly, it would reduce the surface charge on the aphrons and hence the energy barrier to droplet coalescence. We have previously found that the zeta potential of a CLA suspension falls from - 4 5 mV at pH 8.4 to - 3 6 m V at p H 4 [21]. Secondly, protonation of the head-groups reduces the polarity of the surfactant monomers making it energetically less favourable for the hydrophobic tails of the surfactant molecules to remain in an aqueous environment. Experimentally, it is easy to show that the solubility of SDS in 20 mM buffer solutions rapidly decreases below pH 7 and this may also be a factor responsible for decreasing CLA stability at low pH. No other literature data exist on the stability of dispersed CLAs at various pH, though for nondispersed polyaphrons of CGAs formed from anionic surfactants, some authors report t~,2 values to be independent of pH over the range 2-7 [17] and 3 12 [8], whilst others indicate a strong dependence on pH [28].


? ?

t~.2 (min)

dov ( t = 0 ) (~tm)



0 2
4 6 s 1o 12

Aphrons formulated as in Fig. 7 were dispersed in an aqueous solution of the chloride salts of the various cations at pH 5.9 and a total ionic strength of 0.2 M.


Fig. 10. Effect of continuous phase p H on the half-life of CLAs dispersed in deionised water. Aphron formulation as in Fig. 7.


G.Z Lye, D. C Stuckey / Colloids SurJaces A: Physicochem. Eng. Aspects 131 (1998) 119-136

3.6. Effect of temperature
If the collision-coalescence mechanism proposed to explain the data presented in Fig. 7 is correct, then the stability of CLAs should depend upon the temperature at which this process occurs. The increased kinetic energy of the droplets at higher temperatures should result in more frequent collisions with an increased probability of fusion occurring. This indeed was found experimentally, with t m values falling from 96 to 4 min for a corresponding increase in temperature of the continuous phase from 10 to 60°C. Note that the highest temperature used in the experiments reported here is far below the boiling point of decanol (231°C) which precludes evaporation of the solvent as a possible mechanism of aphron destabilisation. For non-dispersed polyaphrons of both CGAs and CLAs, Sebba has shown that they can be broken by slightly superheated steam and by freezing to temperatures between - 1 0 and -50°C [11. The temperature dependence of the measured pseudo-first-order rate constants, k (min-~), can be described by the well-known Arrhenius equation: k = A e x p ( - Ea/RT)







1/1" (K-l)

Fig. 11. Arrhenius plot for the effect of temperature on CLAs dispersed in deionised water at pH 5.8, solid line fitted by linear regression. Aphron formulation as in Fig. 7.

3.7. Effect of additives to the CLA formulation
For "basic" polyaphron formulations comprised of an organic solvent, water and a mixture of surfactants, it has been shown that their stability depends upon the conditions under which they are dispersed (see Figs. 9-11 ), and also upon the properties of the phase components [1,7,15]. One approach to improving the stability of dispersed CGAs or CLAs is to add species during polyaphron formulation which either alter the physical properties of the system, or more specifically the surface properties. Addition of solvents such as ethanol, methanol, tributyl phosphate [ 1] or ether [17] have been shown to break polyaphrons due to reductions in surface tension and viscosity, and also dehydration effects. However, Sebba showed that kerosene polyaphrons formulated with a "soapy-shell" comprised of 60% methanol and 40% water were able to withstand far lower temperatures than normal, water-shell, polyaphrons [ 1]. Following this approach, aphrons were formulated here in which the water of the "soapy-shell" was completely replaced by glycerol (using 1% (w/v) Softanol 120 in decanol and 0.5% (w/v) SDS in glycerol). In contrast to the results presented in Fig. 9, these "glycerol-shell" aphrons actually show an increase in tl/2 values with increasing ionic strength, being three times more stable than the "water-shell" CLAs when dispersed in 0.2 M NaC1 at pH 6.2 and 30°C (tl/2=63 min). This is


where A (min -1) is the pre-exponential factor which is synonymous with the frequency of droplet collisions, E a (kJmo1-1) is the activation energy required for a droplet collision to result in fusion, R (J K - 1 mol - 1) is the gas constant, and T is the absolute temperature. An Arrhenius plot showing the influence of temperature on the experimental first-order rate constant is shown in Fig. 11. Allowing for experimental error, a linear relationship was found between In k and 1/T which justifies the application of Eq. (3) to the data. Performing linear regression yields a single value of the activation energy, E a, for the collision process of 50 kJ mol-~. A value for E a of this magnitude would suggest that, over the temperature range investigated, the collision fusion process of the CLAs is controlled by both diffusion and chemical reaction.

G.J. Lye, D. C Stuckey / Colloids SurJiwes A: Physicochem. Eng As7~ects 131 (1998) 119 136


attributed, in part, to the increased viscosity of the glycerol shell which would improve the mechanical properties of the aphrons. Unfortunately, the high viscosity of the glycerol meant that it was only possible to formulate polyaphrons up to a P VRm,~ of 2, and this may also have a detrimental effect on mass transfer rates of solutes across the interface. For non-dispersed gas polyaphrons, Save and Pangarkar [17] reported improved stability upon the addition of a viscosity enhancing polymer, polyethylene glycol, and the addition of a polar polymer, polyacrylamide. The rheological characteristics of both polyaphrons [21,29] and H I P R E s [29 31] have also been reported. Recent work on CLAs in this laboratory has looked at the extraction of erythromycin from dilute solutions [4,5], and the immobilisation of the enzyme lipase in the "soapy-shell" of the CLAs for use in a membrane bioreactor [32]. Consequently, it is interesting to determine the influence of these compounds on the stability of the dispersed CLAs as shown in Table 3. The oilcore concentrations of erythromycin were chosen to represent those typically found at equilibrium for extraction of 0.5 gl ~ erythromycin at CLA/feed solution ratios between 1:50 and 1:100 [4], whilst the lipase concentrations corresponded to the range over which it was possible to formulate stable polyaphrons [32]. The results in Table 3 show that 0-40 g l ~ of erythromycin has relatively little effect on CLA half-life, although the stability of lipase-containing CLAs increases with increasing concentrations of the enzyme. The

increased half-life of CLAs containing 6 mg ml 1 lipase corresponds to a situation where enzyme immobilisation was found to be most efficient, and where there is approximately monolayer coverage of the oil core of the aphrons with enzyme molecules ( in polyaphrons formulated with > 8 mg ml ~ lipase, evidence of phase separation was observed within 24 h of manufacture) [32]. Similarly, for non-dispersed gas polyaphrons, Save and Pangarkar [17] reported improved stability upon the addition of 0.1 and 0 . 2 m g m l [3-galactosidase to the aqueous surfactant solution from which the gas polyaphrons were generated. This was attributed to decreased diffusivity of material across the interface, and an increase in mechanical strength which might also be the reasons for the enhanced stability of the lipasecontaining CLAs reported in the present investigation.

4. Summary and conclusions
Since Sebba's original publications on the t\3rmulation and manufacture of aphrons [2, 3], subsequent work has focused on the influence of the various system components on the stability of both gas and liquid polyaphrons [1,15-17, 28]. For nondispersed liquid polyaphrons, the current heuristic rules suggest that optimum stability is achieved using relatively apolar solvents with non-ionic surfactants having a high HLB number. Additionally, the size and structure of the aphrons

Table 3 Effect of lipase and erythromycin addition on the stability and sizes of CLAs dispersed in 0.05 M NaCI at 20 C and pH 6.3 and pH 6.8, respectively [Lipase] (gl ~)a 0 1 2 5 6 lu2 (min) 6.5 6.5 8.1 9.9 11.3 dov (t=0) ( p m ) 24.7 24.9 18.4 17.4 15.4 [Erythromycin] (gl ~)b 0 10 20 30 40 t~,= (min) 27.2 28.7 26.3 32.4 22.4 d,,,. (t-0) ({tm) 4.3 4.7 4.4 4.3 4.3

aConcentration of lipase dissolved in aqueous SDS solution prior to polyaphron formation. bConcentration of erythromycin dissolved in organic phase prior to polyaphron formation. Lipase-containing aphrons formulated from 1% (w/v) Softanol 30 in decane and 0.5% (w/v) SDS in deionised water to PVR 4 [32]. Erythromycin-containing aphrons formulation as in Fig. 3 to PVR 7.


G.J. Lye, D.C. Stuckev / Colloids Surfaces A: Physicochem. Eng. Aspects 131 (1998) 119 136

is influenced by the type and concentration of both the non-ionic and ionic surfactants, the physical properties of the components, inclusion of various additives to the formulation and also by the conditions under which the polyaphron is manufactured such as oil addition rate, mixing time and speed, P VR and vessel geometry [1,3,15,17]. In contrast to these earlier investigations, in this work we have developed a method by which the stability of dispersed CLAs can be quantitatively compared by determination of their pseudo-first-order half-lives. For the polyaphron formulation used here, which consisted of the anionic surfactant SDS, optimum CLA stability was found when they were dispersed in aqueous media of relatively low ionic strength, low temperature and high pH. Likewise, for aphrons formulated from cationic surfactants, optimum stability would be expected under conditions of low ionic strength, low temperature, and low pH. It was also found that half-lives could be increased by using various additives to alter the physical and surface properties of the CLAs such as divalent cations, glycerol and other surface active polymers such as proteins. One problem, however, is how to relate the measured values of tl/z obtained at the low volume concentrations permitted by the Malvern instrument (i.e. 0.03%) to the more realistic situation of a concentrated CLA dispersion of between 1 and 50%. In this latter case, a polyaphron phase can be dispersed, allowed to reform by creaming, and then redispersed a number of times over a period of days without any visible sign of deterioration. Also, in the case of a 1 vol.% CLA-immobilised lipase preparation retained within a continuous membrane reactor, no change in the CLA concentration (by absorption at 640 nm) or size (by light scattering) could be detected for over 6 h [32]. The half-lives of dispersed CLAs in more concentrated dispersions are thus considerably longer than those determined by the light scattering technique used here of between 5 and 100 min. It has been confirmed experimentally, however, that the stability of concentrated CLA dispersions also follows the same trends predicted in Figs. 9 and 10, decreasing with increasing ionic strength and decreasing pH. The light scattering technique reported here is thus a useful method to quantitatively examine the

stability of various aphron formulations and the influence of the continuous phase properties on them. Concerning the structure of aphrons, the model originally proposed by Sebba [1] and redrawn in Fig. 1 of a spherical oil-core droplet surrounded by a thin aqueous film is essentially correct. Sebba's direct microscopic examination of liquid polyaphrons [1-3], the cryo-ultramicrotome sections shown in Fig. 3 and the DSC data shown in Fig. 5 clearly indicate an oil core surrounded by an aqueous film. However, evidence for the structure and location of the three surfactant layers is less obvious. If the outermost interface consists of the water-soluble, ionic surfactant then the most logical conformation for the molecules to adopt is that of a bilayer [1]. In this case the stability of this interface, and hence the CLAs themselves, will depend strongly upon electrostatic interactions with the phase in which they are dispersed. Data relating to the ionic strength (Fig. 9) and pH (Fig. 10) of the continuous phase does indeed show this is the case and, therefore, provides some indirect evidence to support at least a monolayer of ionic surfactant. Other than the fact that the non-ionic surfactant is chosen because of its solubility in the organic phase, and that its concentration strongly influences the size of the CLAs [ 1,15], little evidence exists to prove that the inner oil core-"soapy-shell" interface is stabilised by a monolayer of non-ionic surfactant.

4.1. Aphrons or high-internal-phase-ratio emulsions?
Although there is now a substantial body of literature concerning gas and liquid aphrons, the structure of these colloidal systems is disputed by a number of investigators; see, e.g., Ref. [18]. It is suggested that the structures of liquid polyaphrons and HIPREs are, in fact, the same and that the term aphron should be dropped to prevent confusion. This may well be the case since it has been shown throughout this work that polyaphrons and HIPREs display similar characteristics with respect to structure, viscosity and creaming behaviour. There are, however, a number of significant differ-

G.Z Lye. D. C Stuckey / Colloids Surfaces ,4: Physicochem. En,e. ,4s7~ect.v131 f 1998) 119 I3~


ences in the formulation and manufacture of the two systems which should be pointed out. (a) Formulation. HIPREs are generally formulated using a pure solvent phase and an aqueous phase containing 10-50% of a water-soluble, non-ionic surfactant [ 19, 20, 30, 31 ]. Liquid polyaphrons, on the other hand, are formulated using a solvent phase containing 1-2% of an oil-soluble, non-ionic surfactant and an aqueous phase containing 0.1-0.5% of a water-soluble, ionic surfactant [1,15]. We have never been able to manufacture liquid polyaphrons using only one or other of the two types of surfactant. (b) Mam{facmre. Polyaphrons are formed by the spreading of an oil film on the surface of an aqueous surfactant solution in a region of intense mixing (hence the requirement of the oil-soluble surfactant to ensure that the spreading pressure of the oil is sufficiently high). HIPREs, on the other hand, are formed by either manually shaking the required volumes of the two phases [19] or by using a laboratory mixer [20]. It has also been reported that HIPREs can be formed by the creaming of a dilute emulsion prepared by homogenisation

our current investigation of solute mass transfer kinetics, should enable the successful design and operation of a CLA (or HIPRE!) extraction process.

The authors would like to thank the SERC ['or financial support of this work (grant number GR/H/84727) and Susan Wilson and Chris Gilpin of the EM Unit, University of Manchester, for performing the cryosectioning and help with the TEM.

[1] F. Sebba. Foams and Biliquid Foams. Aphrons, Wiley, New York, 1987. [2] F. Sebba, J. Colloid Interface Sci. 40 11972) 468. [3] F. Sebba, Colloid Polym. Sci. 257 (19791 392. [4] G.J. Lye, D.C. Stuckey, in: D.L. Pylc I Ed.), Separations for Biotechnology llI, SC1 Publishing, 1994, pp. 280 286. [5] G.J. Lye, D C. Stuckey, in: D.('. Shallcross e! al. (Eds.), Proceedings 1SEC'96, Vol. 1I. University or" Melbourne, 1996, pp. 1399 1405. [6] D.A. Wallis, D.L. Michelsen, F. Sebba, J.K. Carpenter, D. Houle, Biotechnol. Bioeng. Symp. 15 (1985) 399. [7] G.J. Lye, UV. Poutiainen, D.C. Stuckey. in: Biotechnology "94, Vol. 11, 2nd Int. Syrup. on Environmental Biotechnology, Institution ot ('heroical Engineers. 1994, pp. 25 27. [8] M.B. Subramaniam, N. Blakebrough, M.A. Hashim. J. Chem. Technol. Biotechnol. 48 (1990) 41. [9] S.V. Save, V.G. Pangarkar, J. Chem. Technol. Biotechnol. 61 11994) 367. [10] P.G. Chaphalkar, K.T. Valsaraj, D. Roy, Sep. Sci. Technol. 29 (1994) 907. [11] M. Caballero, R. Cela, J.A. Pcrez-Bustamcmc, Sep. Sci. Technol. 24 11989) 629. [12] D.L. Michclsen, K.W. Ruettimann, KR. Hunter, I-. Sebba, Chem. Eng. Commun. 148 (1986) 155. [13] D. Roy. K.T. Valsaraj, S.A. Kottai, Sep. Sci. Technol. 27 11992) 573. [14] S. Ciriello, S.M. Barnett, F.J. Delusc. Sep. Sci. Tcchnol. 17 { 1982) 521. [15] K. Matsushita, A.H. Mollah. D.C. Stuckcy, C. del Cerro, A.I. Bailey, Colloids Surfaces A 69 11992) 65. [16] P.G. Chaphalkar, K.T. Valsaraj, I). Ro 3. Sep. Sci. Technol. 28 (1993) 1287. [17] S.V. Save, V.G. Pangarkar, Chem. Eng. ('ommun. 127 (19941 35.

It would thus appear that it is "time to perform a definitive structural study" on HIPREs and liquid po[yaphrons [33]. The crux of such an investigation would be to determine the location and orientation of the surfactant molecules present in the phases. This is seen as the key to deciding whether these two colloidal systems are structurally identical or not since an elaborate series of surfactant-laden interfaces have been proposed to account for the properties of polyaphrons [1]. In conclusion, we have shown that the macroscopic structure of CLAs as proposed by Sebba [l] is essentially correct and is somewhat similar to that found for HIPREs. Unfortunately, there is still little direct evidence for the microstructure of the surfactant interfaces present in either CLAs or HIPREs. A definitive structural study is needed before we can rationalise our understanding and nomenclature in this area. From an engineering point of view, however, we now have quantitative data on the stability of CLAs which, together with


G.J. Lye, D.C. Stuckev / Colloids Surfaces A." Physicochem. Eng. Aspects 131 (1998) 119-136
[26] H.M. Princen, M.P. Aronson, J.C. Moser, J. Colloid Interface Sci. 75 (1980) 246. [27] E.S. Perez de Ortiz, in: J.D. Thornton (Ed.), Science and Practice of Liquid Liquid Extraction, Vol. 1, Clarendon Press, 1992, pp. 157-209. [28] M.C. Amiri, E.T. Woodburn, Trans. IChemE, Part A 68 (1990) 154. [29] J.P. Heller, M.S. Kuntamukkula, Ind. Eng. Chem. Res. 26 (1987) 318. [30] R.J. Mannheimer, J. Colloid Interface Sci. 40 (1972) 370. [31] H.M. Princen, J. Colloid Interface Sci. 105 (1985) 150. [32] G.J. Lye, O.P. Pavlou, M. Rosjidi, D.C. Stuckey, Biotechnol. Bioeng. 51 (1996) 69 78. [33] S.E. Taylor, Personal communication, 1996.

[18] H.M. Princen, Langmuir 4 (1988) 486. [19]K.J. Lissant, B.W. Peace, S.H. Wu, K.G. Mayhan, J. Colloid Interface Sci. 47 (1974) 416. [20] K.J. Lissant, J. Colloid Interface Sci. 22 (1966) 462. [21] D.C. Stuckey, K. Matsushita, A.H. Mollah, A.I. Bailey, in: 3rd Int. Conf. on Effective Membrane Processes: New Perspectives, BHR Group Publications, 1993, pp. 3 19. [22] S. Raynal, I. Pezron, L. Potier, D. Clausse, J.L. Grossiord, M. Seiller, Colloids Surfaces A 91 (1994) 191. [23] D. Clause, J.P. Dumas, P.H.E. Meijer, F. Broto, J. Dispers. Sci. Technol. 8 (1987) 1. [24] S. Ganguly, K.S. Adiseshaiah, Colloids Surfaces 66 (1992) 105. [25] D.J. McClements, S.R. Dungan, J.B. German, J.E. Kinsella, Food Hydrocolloids 6 (1992) 415.