1-s2.0-0734975095000017-main

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 34

BiotechnelosyAdvaaom,Vol. 13, No. 2, pp.

175-208, 1995
CopyrishtO 1995ElsevierScienceLtd
Pergamon
0734-9750/95$29.00+ .00
0734-97S0(9S)00001-1

LIQUID EMULSION MEMBRANES: PRINCIPLES, PROBLEMS AND

APPLICATIONS IN FERMENTATION PROCESSES

P. IL PATNAIK
Institute of Microbial Technology
Sector 39-#., Chandigarh 160 014
India

ABSTRACT

Liquid emulsion membranes (LEMs) have developed into a versatile technique for a
variety of applications involving selective and controlled transport of biochemicals.
Biological applications cover the controlled delivery of drugs from capsules, detoxification
of the circulatory system, recovery of useful compounds from waste streams and selective
separation of products from fermentation broths. This review traces the development of
LEMs, discusses their key features, advantages and limitations, describes methods of
modelling LEM systems and highlights some applications with industrial potential.
Two kinds of LEM systems are considered. The first type are agitated emulsions,
which are relatively easy to prepare and use but may be limited in their selectivity and
long-term stability. Supported liquid membranes (SLMs) are a recent development; they
use porous solid supports and have excellent stability and selectivity. Their chemical engi-
neering aspects and applications in fermentation processes are considered.

KEY WORDS: Liquid emulsion membranes, bioseparation, extraction, fermentation.

INTRODUCTION
The efficient separation and concentration of fermentation products plays a key role in
the commercial success of a process. Even if the fermentation is successful, recovery of the
product can be a major bottleneck. Some processes, like conventional ion exchange, are
inherently batch operations that require pretreatment of the fermentation broth. Other
techniques, such as whole cell broth extraction and ultracentrifugation, can be used in a
continuous mode but require prederivatization of the solute, are limited in their range of
applications, and often have high capital costs. Chromatographic separations are very
efficient but they too require pretreatment of the broth, are capital and labour intensive
175
176 P.R. PATNAIK

and are not easily scaled [1]. To overcome these limitations, Li [2] developed the liquid
emulsion membrane (LEM) process. From their initial application for the fractionation
of hydrocarbons [3], LEMs have diversified into the recovery of heavy metal ions [4], the
removal of contaminants from wastewater [5,6], heterogeneous reactions replacing
conventional catalytic processes [7,8], treatment of disorders in the blood stream [9] and
the extraction (followed by reaction in some cases) of products of fermentations [10-13].
Applications of liquid membrane technology for the recovery of fermentation products
have grown rapidly because a substantial part of the technological and financial success
of bioprocesses depends on the post-fermentation steps [14,15]. In a broad sense, LEMs
may be considered to be an extension of liquid-liquid extraction but they have many
differences and advantages.
To understand the advantages of LEMs, consider their schematic representation
shown in Fig. 1. We see that an LEM contains three phases. The innermost phase is
usually aqueous, the intermediate phase is an organic solvent and the outer phase is
aqueous. There are thus two dispersed phases forming an emulsion with the continuous
phase. Solute is usually transported inward through the organic phase, called a liquid
membrane because it is thin and permits selective transport of molecules. Selectivity for
the desired species is often enhanced by incorporating suitable carriers in the membrane.
And since the interior aqueous phase has a much smaller volume than its external
counterpart, separation and concentration of the solute are achieved in one step. In
addition, by encapsulating appropriate reagents or enzymes or cells inside the aqueous
droplets, the solute can be concentrated against its concentration gradient [5,16] or reacted
further to generate other products such as secondary metabolites [17,18].

MODEL REAL
DROPLET LIQUID DROPLET
MEMBRANE

S p \ /
SUBSTRATE PRODUCT DISPERSED
PHASE

Figure I. Schematic representation of an L E M (from [10]).Reproduced with permission


of Elsevier Science.
LIQUIDEMULSIONMEMBRANES 177

LEM processes are readily scaled up from laboratory to commercial levels. An


LEM process for the recovery of zinc from wastewater has been successfully
eommercialised [19]. In spite of their complexity and stringent quality criteria, many
pharmaceutical products have been recovered by LEMs at the pilot scale with results that
show potential for industrial applications [10,20,21]. This scalability has been demonstrated
in different kinds of equipment [1,21-23] and in both free and solid-supported LEMs
[24,25]. Another advantage of LEMs is their robustness. Particulates and floes do not
affect the mass transfer rates [26], and LEMs can function efficiently even in the presence
of live bacterial cells [27] or enzymes [18]. Thus, LEMs can be used directly in
fermentation broths and to obtain both separation and subsequent reaction in one stage.
Some major limitations which have slowed the approach to commercial applications
of LEMs are: (a) swelling, (b) tendency to rupture, (c) long-term stability and (d) inac-
tivation of the encapsulated enzyme(s) during emulsion preparation. Recent studies have
focused on these problems and solutions such as impregnation in porous solid supports
[25], the use of appropriate carriers and emulsifiers [1,17], and optimisation of emulsion
composition and preparation conditions [13] have reduced the drawbacks. Later sections
discuss the problems, the solutions and some important applications.

PRINCIPLE OF LEMs
Most LEMs used in biochemical applications are of the 'water-oil-water' type described
above (see Fig. 1). However, sometimes the order of the phases may be reversed, resulting
in 'oil-in-water' emulsions dispersed in a continuous organic phase. These emulsions have
potential applications in enzymatic and cellular reactions carried out in nonaqueous media
[28,29]. However, most of today's applications are of the 'water-oil-water' type, and we will
therefore focus on this type of emulsions. The principle of transport is the same for both.
Since the aqueous and organic phases are immiscible, a stable emulsion is possible
only by decreasing the interracial tension between the liquids. This is achieved by using
surface active agents. Commercial surfactants which are commonly used are complex
compounds that have hydrophillie and lipophilic groups, the latter orienting the molecules
towards the organic membrane phase. The ratio of the two groups, called the
hydrophillic-lipophilic balance (HLB) determines the kind of emulsion obtained. If the
HLB is less than about 10 (as in Span 80), 'water-oil-water' emulsions are obtained;
178 P.R. PATNAIK

surfactants with larger HLB values (e.g. Tween 80, Triton X-100) favour 'oil-water-oil'
emulsions [10].
Sometimes, as in the controlled release of drugs from capsules implanted in the
body [30], the solute is transported from the inner droplets to the fluid outside. However,
in fermentation applications and most recovery processes the transport is inward. In either
case it is important that only the species of interest be transported from among many
others in solution, and this may be achieved simply by virtue of the preferential solubility
of the species [3,5] or by using carrier molecules that bind to the desired molecules and
ferry them across the membrane phase [12,14]. At the inner surface of the membrane, the
solute is released into the aqueous droplets and retained there by changing its ionic
character through reaction with an acid or an alkali (as in physical transport) or with a
counter-ion (as in carrier-mediated transport).
To be practically useful, an LEM should satisfy certain requirements. It should
provide rapid and selective transport of the desired species, be able to concentrate the
products, retain the activities of the encapsulated enzymes or cells close to their native
activities, be stable under variations in process conditions and allow easy and inexpensive
de-emulsification and recovery of the product [14,18,31]. Rarely is it possible to meet all
these requirements perfectly, so a judicious compromise is generally adopted.

PREPARATION OF LEMs
There are two stages in the preparation of an LEM (see Fig. 2). First a two-phase
emulsion of the aqueous phase in the organic phase is prepared by slowly adding the
aqueous solution to a fixed volume of the organic phase contained in a homogeniser
operated at high speeds of 8000 to 10,000 rpm [11] to generate adequate shear. This
emulsion is then redispersed in the outer aqueous phase; a lower speed of agitation (300
to 500 rpm) is used to get larger globules and to avoid rupture of the liquid membrane,
a problem which will be addressed separately.
The sizes of the aqueous and emulsion droplets depend on the fluid properties, the
composition of the membrane and the stirring speeds. Normal diameters are 20 to 40 ~m
for the aqueous droplets and 200 ~m to 2 mm for the two-phase globules [10,17]. A
proper choice of the diameters is important. If the aqueous droplets are too small, too
many of them are packed into each organic globule and consequently the liquid membrane
becomes too thin and ruptures easily. Large aqueous droplets, on the other hand, result
LIQUID EMULSIONMEMBRANES 179

in poor efficiency because of a low surface-to-volume ratio. Therefore considerable


attention has been devoted to the choice of membrane composition and emulsion
preparation conditions [1,17,18,32].

INNER MEMBRANE OUTER


PHASE PHASE EMULSION PHASE

EMULSIFICATION DISPERSION

Figure 2. Steps in the preparation of a LEM (from [10]). Reproduced with permission of
Elsevier Science.

TYPES OF SEPARATIONS
Type h Physical Separation
The principle of physical separation is shown in Fig. 3. Transport across the membrane
is driven by the ability of the desired species to partition into the membrane phase and
its diffusion rate through the membrane. The driving force is simply the concentration
gradient. In order to sustain a large concentration gradient over a length of time, the
transported solute must be converted in the receiving aqueous droplets into a form which
cannot diffuse back.

HA¢ > HAc

A¢- Ae

Outer Phase ~embrene Phase I n t e r i o r Phase

Figure 3. Principle of physical separation. Reprinted from [14] by courtesy of Marcel


Dckker, Inc.
180 P.R. PATNAIK

An early application of this principle was the separation of acetic acid from
wastewater [5]. The nonionised form of acetic acid partitions into the organic phase and,
at the inner surface of the membrane it is released into the aqueous droplets, which
contain a concentrated alkali (usually NaOH) to ionise the acid back into CH3COO- , thus
preventing it from diffusing back. Other biological applications have been for the delivery
of drugs [30], the removal of cholesterol [33] and phenolic agents [6,9] from blood, and
emergency treatment of drug overdose [34,35].
While Type-I mechanism is simple and inexpensive, its selectivity and efficiency are
not always good. Moreover, as the separation proceeds, the neutralising or anchoring
compound in the inner aqueous phase gets depleted earlier in the droplets close to the
surface of a globule; hence the solute has to diffuse further and further inward before
being released and neutralised. Therefore as time progresses the process becomes
diffusion limited and slows down [36]. To overcome this limitation a carrier is used in the
membrane phase in order to increase speed and selectivity. This gives rise to the Type-II
mechanism.

Type II: Facilitated Diffusion


The principle of facilitated or carrier-mediated transport is depicted in Fig. 4.
Carrier-mediated transport may be of either of two types. In co-transport the carrier
molecule must be protonated before it can form a neutral complex with the anion to be
transported [37]. Secondary amines favour co-transport. In the reactive extraction of

s_) ~
penicillin G, Schuler and coworkers [12,38-40] found Amberlite LA-2
(N-lauryl-N-trialkylmethyl amine) and diisotridecyl amine to be the most suitable.

outer phase organic phase inner phase outer phase organic phase inner phase
Q~ q
S" H" --
-b S"

S ÷ -- -b H"

QH*S- ~ QH*S-
12- C"

counter t r a n s p o r t co-transport

Figure 4. Principle of facilitated transport (from [37]) Q = amine; C = counter-ion; S =


solute. Reprinted with permission of the author and VCH Publishers ©1989.
LIQUIDEMULSIONMEMBRANES 1$1

Quaternary amines are used for counter-transport. They have a long hydrophobic
tail section and a positive univalently charged hydrophillic head group; therefore, in order
to remain electrically neutral the carrier is always complexed with a counterion. This
counterion is exchanged for the charged solute at the outer surface of the liquid
membrane. After diffusing to the opposite interface, the transported species is released
into the inner aqueous phase in exchange for a new counterion. By this shuttle system the
desired solute accumulates in the dispersed aqueous droplets and the counterion is
liberated into the continuous phase. The extraction of penicillin G has also been reported
by this method [12]. Another useful application has been for the separation of
L-phenylalanine from a mixture of D- and L-isomers [1]. Both these are discussed later
in this review.
Sometimes a combination of Type-I and Type-II transport is used, as done by
Makryaleas and coworkers [17] for the enzymatic production of L-leucine. Leucine
dehydrogenase catalyses the reductive amination of tt-ketoisocaproate. Formate
dehydrogenase is necessary for the continuous regeneration of N A D H which becomes
oxidised during the amination. The substrates, formate and a-ketoisocaproate and the
product are transported by the carrier, Adogen 464, whereas ammonia, being soluble in
the organic phase, passes through by simple diffusion.

FACTORS AFFECTING LEM PERFORMANCE


Membrane Rupture
Since liquid membranes are thin, mobile and subject to shear, they tend to rupture during
preparation and use. The stability of an LEM depends on the shear produced by agitation
[41,42], internal droplet size [43,44] and composition of the membrane [1,42,43]. The
composition is crucial in determining robustness and, therefore, many workers have studied
methods of obtaining the 'best' formulations. One method of quantifying the rupture rate
is by encapsulating an enzyme in the inner aqueous droplets and measuring its activity in
the outer continuous phase as the enzyme leaks out. This method has been used with
•,-chymotrypsin [18], leueine dehydrogenase [17] and penicillin G acylase [18]. A
disadvantage of this method is that inactivation of the enzyme as it passes through the
membrane phase falsifies the measured leakage rates [10]. Therefore Meyer and associates
[32] encapsulated metal ions in place of enzymes. Potassium was used in their study of the
hydrolysis of 4-acetoxy-cyclopentanone by pig liver esterase, and that of Thien et al. [1] for
182 P.R. PATNAIK

L-phenylalanine transport across a paraffinic membrane. A leakage rate of less then 1%


per hour over the duration of the process is considered acceptable and both studies have
demonstrated that this is possible by selecting a proper composition for the membrane.
In some situations the metal ion may permeate through the intact membrane, thus
indicating leakage rates larger than the actual rates. Gadekar and coworkers [16] observed
that Ni and Cu individually permeated kerosene membranes but, in the presence of Cu,
nickel did not permeate; thus both metal ions had to be encapsulated to measure true
leakage rates. In their experiments on nitrophenol extraction, the rupture rate was less at
a low temperature but so was the extraction; this suggests that a judicious choice of
temperature is important.

Membrane Swelling
Swelling may be caused by shear or occlusion or any surface-active agent, including the
carrier. Data for nitrophenol [16] and L-phenylalanine [1] show that as the surfactant or
the carrier concentration is increased, the enhanced driving force initially overcomes the
negative effects of swelling. Eventually, however, swelling becomes significant and causes
dilution of the product in the interior phase (Fig. 5). Since most LEM separation systems
have large osmotic gradients [45], some transport of water along 'with the solute is
unavoidable and, hence, swelling may be minimised but not eliminated.

4O i I ~ I I
I00

~35 88


3o 76 --
0
t~
¢-
--64~
I1.

'7. 2O ÷ ,o,,,io, [Phe] 52


-A -A- % Swell
C
15 I I I I I 40
0 2 4 6
Corrier Concentrotion (% v/v)

Figure 5. LEM performance for L-phenylalanine extraction (from [1]). Reprinted with
permission of John Wiley & Sons, Inc.
LIQUID EMULSION MEMBRANES 183

The disadvantages of swelling are many: (a) the transferred water dilutes the solute
that was concentrated in the inner droplets, (b) the driving force is reduced, (c) since the
membrane occupies a larger fraction of the emulsion volume, productivity (per unit
volume) is reduced, and (d) the membrane becomes weaker [14].
Two mechanisms have been suggested for swelling. In the first [45], swelling is
considered to take place via a hydrated surfactant which traverses back and forth across
the membrane. The second mechanism is via reverse micelles. Unlike hydrated surfactants,
micelles can accommodate both solute and water molecules; this inference is based on
data [34,36] which indicate that charged solutes are transported across liquid membranes
without carders, especially when micelle-promoting suffaetants such as sorbitan monoole-
ate are present [47,48].
Studies by Kopp et al. [49] and Thien et al. [1] show that the percentage of swelling
generally increases linearly with the square root of time after an initial nonlinear period.
Their analyses suggest that transport during the nonlinear period is kinetieally controlled
and thereafter by diffusion. This also perhaps explains why in Thien et al.'s [1] experiments
the phenylalanine concentration in the inner water droplets went through a maximum as
the carrier concentration was increased.
Swelling may be controlled by adding certain hydrocarbons or anion exchangers.
Gadekar and coworkers [16] found that addition of 2.5 to 5% cyclohexanone reduced
swelling by more than half.

Selectivity
Selectivity is a key consideration in any separation process. Fermentation broths usually
contain many components and therefore the ability of the membrane to transport only the
species of choice becomes important. The mechanism of selection depends on the type of
LEM. In Type-I processes selectivity depends on the ability of the solute to dissolve in the
membrane phase and the driving force depends on the conversion efficiency of the solute
to a nontransportable form in the inner aqueous droplets. Together these factors imply
that solubility in the organic phase governs selectivity at short times while long-term
selectivity is determined by the driving force [14]. This feature is illustrated in Fig. 6 for
a phenol-acetic acid mixture [5]. Phenol, being more soluble than acetic acid in the
membrane phase, is initially taken up faster. Gradually, however, the stronger acetic acid
is preferentially taken up in the inner aqueous droplets.
184 P.R. PATNAIK

ao I I 1 1 1 1 1 1 1

~ 70
E
o
t~ 60
I-

~ 5o
o

"~3-0
c
0
~ ao
o Acetic Acid
& Phenol

i i I I i i I 1 I
o 4 8 12 16 20
Mixing Time (rain)

Figure 6. Uptake of phenol and acetic acid in a LEM system. Reprinted from [5] by
courtesy of Marcel Dekker, Inc.

Selectivity in Type-II processes is governed primarily by the carrier in the


membrane. In their studies of L-phenylalanine extraction, Thien and associates [1]
observed that the hydrophobicity of the solute plays a strong role in determining specificity
of the carrier; the initial flux of the solute increases with its hydrophobicity. In the same
work sodium sulfate was added to the continuous phase to simulate the use of inorganic
salts in fermentation media. Since the sulfate competes with the carder, the degree of
extraction decreased as sulfate concentration increased. Similar behaviour was observed
by Likidis and Schugerl [12] in the reactive extraction of penicillin G. With n-butyl acetate
as the membrane phase and diisotridecyl amine, a secondary amine, as the carrier, the
extraction efficiency decreased with increasing pH and with time at a given pH. The
development of commercially viable carriers with high selectivity still remains a fertile area
of work-

E~e Inactivation by the Membrane


The encapsulated enzyme may lose some of its activity during the process of making the
LEM. The choice of the membrane phase therefore depends both on its ability to
transport the solute selectively and its effect on the enzyme. Figures 7 and 8 illustrate this
LIQUID EMULSION h4EMBRANEs 185

for the coenzyme NADH used in the production of Lleucine [17]. In its pure form, thin
paraffin preserved NADH activity closest to its level in the native state. The addition of
emulsifiers reduced the activity in all cases. Thus emulsion stability is achieved at the cost
of enzyme activity.
-
100
-
-
92 95 -

85

dP

2
v
:
.E
s
f .E
.-F ! e :
Lm
5 k
Y x F
- - - - L

Figure 7. Influence of different Figure 8. Influence of differen It


membrane phases on the NADH surfactants on the NADH
concentration (from [lq). Reprinted concentration (from [17j). Reprinted
with permission of the author and with permission of the author and
VCH Publishers 81985. VCH Publishers 01985.

MATHEMATICAL MODELLING
The simplest approach to modelling LEMs is to consider the continuous phase to be
homogeneous and coalesce all the aqueous microdroplets within an emulsion globule into
one droplet (Fig. 9). In addition, the membrane is assumed to be of negligible thickness.
Such a model has been used for the sustained release of a drug out of a capsule implanted
in the body [30]. By this model,

(1)
186 P.R. PATNAIK

where M t is the amount of solute (drug) present in the droplet at any time t, M o is the
initial amount of solute, A is the surface area of the droplet, K is the partition coefficient
in the membrane phase, L is the thickness of the liquid membrane, VI is internal volume
of the emulsion phase and V2 is the volume of the continuous phase. Since V2 > > V1,
Eqn.(1) simplifies to

.2.

Figure 9. Lumped two-phase model of a LEM globule (from [30]). Reprinted with
permission of VCH Publishers ©1989.

This model suffers from a number of weaknesses. Coalescing the aqueous droplets
prevents analysis of the effects of droplet size and number of droplets per globule.
Assuming the membrane thickness to be negligible nullifies the role of diffusion through
the membrane. Moreover, the model does not take account of reactions promoted by
enzymes or cells encapsulated in the droplets, a feature often exploited in two-stage
reactions using LEMs [11,18,32]. Since these features are known to be significant
[1,14,17,50], the model expressed by Eqn.(1) is oversimplified.
These limitations have been overcome in the advancing front model (AFM) [35],
shown schematically in Fig. 10. Each emulsion globule is considered to contain a large
number of aqueous droplets, inside which there may be chemicals or enzymes or cells.
Local equilibrium is assumed between the aqueous and organic phases of a globule and,
therefore, a single average concentration is used inside the globule. This concentration,
LIQUIDEMULSIONMEMBRANES 187

however, varies with time and along the internal radius. It is assumed further that the
solute is consumed instantaneously upon reaching an aqueous droplet containing an active
reagent. As a result, there is a spherical reaction front (Fig. 10) which moves inward until
the globule is completely exhausted.
The governing equations of the AFM are [35]:
Globules:

Oc De//0[.20cl. Rf<r<R (3)


-~ " ~-7~-~ ' ~ I'

t-o; , - o (r<R) (4)

r-R; c-rc, (t>O) (S)

r- Rfq); ~- 0 (t > o) (e

Continuous phase:

dce 3 Oc.. (7)


-v,_dr - ~(v., + ~)D,r:-#,.R

t - 0; Ce - ¢eO (8)

Reaction front:

Vi dR, ~.
(9)

t- O; R f - R (10)

Here c is the solute concentration in the globule, c e is its concentration in the continuous
phase, R is the radius of the globule, R f is the radius of the reaction front, V/is the volume
of the internal aqueous phase, V m is the volume of the membrane phase, Ve is the volume
of the continuous phase, K is the partition coefficient and Def
t is the effective diffusivity
inside the globule.
188 P.R. PATNAIK

GLOBULE OF

~ /
/EMULSION

REACTION FRONT

/ ,' " • ~ _ I N~T E R N A L


~ DROPLET

/\ % • S O L U T E , I(,..)(,~ r ~ I N T E R N A L DROPLET
• -" - - CONTAINING
- REAGENT

Figure 10. The advancing front model (from [36]). Reproduced with permission of the
American Institute of Chemical Engineers. ©1982 AIChE. All rights reserved.

Lorbach and Hatton [50] extended the AFM to include axial dispersion, which is
significant in columnar vessels such as the ones used by Reschke and Schugerl [24] and
Likidis and Schugerl [22] for penicillin G. To Eqns.(3) to (10) is added the axial dispersion
equation

dce d2Ce ** 3
Ue "-'d-i"
-De -~z "- fO 7~ (Vm + Vi)Deff-~~'Rf(R)dr (II)

where u e is the superficial velocity of the continuous phase, D e is its effective diffusivity,
and z is the distance along the column. Equation (11) contains another extension of the
AFM by accounting for variation in the sizes of the globules through a distribution
function f(R); unlike well-mixed vessels, this variation may not be neglected in column
contactors. Ho et aL [36] and Lorbach and Hatton [50] applied their models to the
extraction of phenol from wastewater, while Chaudhuri and Pyle [51,52] applied the Ho
version to the recovery of lactic acid.
Although applicable to many LEM systems, the AFM (and its extended version)
has one major limitation. The assumption that the reaction rate of the transported species
is infinitely faster than its diffusion rate through the membrane is not always valid, as in
the removal of phenol from hepatic blood [9,53] and the reduction of nitrate to nitrite
using cell homogenate from Micrococcus denitrificans [46].
LIQUIDEMULSIONMEMBRANES 189

SURFACE OF
EMULSION GLOBULE

" ~ i : 2 ~i!2~: UNREACrEOCORE

Figure 11. The reaction zone model (from [55]).

When diffusion and reaction occur at comparable rates, the reaction is spread over
a finite interval of the radius of the globule, that is, instead of a reaction front there exists
an annular reaction zone (Fig. 11). For isotropic globules the zone remains of constant
thickness and moves inward until it collapses at the center. Janakiraman [54] proposed the
first model for this situation. Vyas and Patnaik [55] removed certain inconsistencies in the
model and compared it with the AFM. The equations are rather involved, so the original
references may be consulted.
The reaction zone model has been tested with data for the extraction of phenol
from wastewater. When only phenol is present, the A F M is applicable [36]. But if a
mixture of phenol and acetic acid is present, there is competitive diffusion and
counter-diffusion [1,5], which is described properly by a reaction zone model [55].

SELECTED APPLICATIONS
L-Phenylalanine Separation
Thien et al. [1] studied the reactive separation of L-phenylalanine from a racemic mixture.
L-phenylalanine is a zwitterion; it is positively charged at p H < 3 and negatively charged
at p H > 9. Aliquat 336, a tri-capryl quaternary ammonium salt, was used in its chloride
form as the carrier. To ensure complexation with the unipositive hydrophillie head of the
carrier molecules, the continuous phase aqueous solution of L-phenylalanine was
maintained at p H 11 so that the amino acid had a negative charge. L-phenylalanine dis-
places the C1- counterion and is thus transported across the membrane (Fig. 12).
190 P.R. PATNA/K

Two-third of the phenylalanine could be extracted in about 40 minutes, after which the
rate of removal slowed down. The concentrations of the carrier, the counterion and the
surfactant had a significant effect on the rate of transport and swelling of the membrane.
An interesting aspect of the study was that the optimal process conditions depended on
whether the solute was only to be separated or to be separated and concentrated. If only
separation is required, the C1- concentration should be as large as possible. However,
since this also promotes swelling of the membrane, a lower C1- concentration is better if
both separation and concentration are desired.

P Phe-

/
\
CI- I C'Phe- CI-
Outer Phase Membrane Phase Interior Phase

Figure 12. LEM separation of L-phenylalanine (from [1]). Reprinted with permission of
John Wiley & Sons, Inc.

Recently, Itoh and coworkers [56] used a di-(2- ethylhexyl)-phosphoric acid as the
carrier. Qualitatively similar findings have been reported. An encouraging new observation
was that 'impurities' such as sodium chloride, glucose and lactic acid, which are common
in fermentation broths, did not significantly affect the transport rates of L-phenylalanine.

Transport and Hydrolysis of D,L-Phenylalanine Methyl Ester


Ha and Hong [57] adopted a different approach to obtain L-phenylalanine. The
continuous aqueous phase had a mixture of D- and L-phenylalanine methyl esters. The
objective was to transport both isomers and hydrolyse them by using encapsulated
a-ehymotrypsin. Figure 13 is a schematic diagram of their method. As in Thien et al.'s
work [1], an ion exchange mechanism was used, with Adogen 464 in its hydroxyl form as
the quaternary ammonium salt, kerosene as the membrane phase and Span 80 as the
emulsifier. Both the inner and outer phases were buffered to pH 7. At this pH the ester
LIQUIDEMULSIONMEMBRANES 191

exists largely in its protonated form. The small amount of unprotonated ester permeated
the membrane by physical solubility. The inner aqueous droplets contained
a-chymotrypsin, which hydrolysed this neutral ester to produce L-phenylalanine and
D-ester. At a pH 7 the amino acid is zwitterionic, of which only a small fraction is in the
anionic state. This fraction was transported outward by the carrier, which also
counter-transported hydroxide ions into the inner aqueous phase (see Fig. 13), thus
maintaining neutrality.

~iii~iiiiiiiiiiiiiii!i
LE IIIIIII~!LE!IIIIIII b LE ,

L
f ::::::::::::::::::::::
H" H" i!iii!iiiiiiiiii!ii!iii
Enzyme, ]
::::::::::::::::::::::
MEOH MEOH4-
iii!i~::i!!ii!!!!!!ill
iii~i~iLiiiiiiiii!i
i~::!iiiiiiiiiiiiiiill
H~O -'~ ~ - LA*- ,I--]
i C a ~:ri~elriill
LA *" HzO
ib!embr~ Inner Phase
Outer Phase iiphas6iiiiiiiiii

Figure 13. LEM separation and hydrolysis of D,L-ester (from [57]). Reprinted with
permission of John Wiley & Sons, Inc.

Continuous extractions were run for up to seven hours. Ha and Hong [57] observed
that by optimising the composition of the membrane enzyme stability could be improved
and loss of activity upon encapsulation reduced. The LEM could be used three times and
encapsulation resulted in only 40% loss of activity compared to 70% in the case of
Scheper et al. [58] for the same system. An interesting feature was that, despite being
transported solely by physical diffusion, the transport rate of the ester through the
membrane was much faster than its rate of reaction. By contrast, in most applications
[1,17,18] mass transfer through the membrane is the rate determining step.

L-Leueine Biosynthesis
This complex system involves physical diffusion, two-way carrier-mediated transport and
enzymatic reaction. Consider the schematic representation shown in Fig. 14. The main
192 P.RP
.ATNAIK
substrates (in the outermost phase) are ,,-ketoisocaproate, ammonium formate, and
ammonia produced by the equilibrium reaction NH4 + + O H - "-, NH 3 + H20. In Makrya-
leas et aL's [17] experiments the membrane consisted of thin paraffin with 5% Span 80 as
the emulsifier. The carrier was Adogen 464 in its chloride form.

H ~ 0*-
substrates products
pit8
Figure 14. Reaction scheme for L-leucene (from [17]). Reprinted with permission of the
author and VCH Publishers ©1985.

The formate and ct-ketoisocaproate (in their ionic form at pH 8) are transported
inward by the carrier. Neutral ammonia, being soluble in the membrane, travels by simple
diffusion; it is protonated in the inner aqueous droplets, which contain two enzymes:
leucine dehydrogenase (LDH) and formate dehydrogenase (FDH). LDH catalyses the
reductive amination of a-ketoisocaproate to L-leucine. F D H is responsible for the
regeneration of NADH, which becomes o~ddised during the amination. The L-leucine (in
anionic form) is transported outwards through the membrane by the carrier, which thus
serves two purposes.
Batch experiments showed that mass transfer was the rate limiting step. In
continuous runs steady state was reached after four residence times, as also observed by
Scheper et al. [58]. "Conversions" of 70% or more were achieved at residence times of
more than 200 minutes, but smaller residence times are required for commercial viability.
Nevertheless, the study demonstrated the suitability of liquid membrane technology for
complex processes involving many steps of diffusion and reaction.
LIQUID E M U L S I O N M E M B R A N E S 193

Enzymatic Conversion of Penicillin G to 6-APA


The conversion of penicillin G to 6-APA (6-amino penicillanic acid) is catalysed by
penicillin acylase. Thus, in the work of Scheper et al. [18] the inner aqueous droplets had
the acylase. The membrane phase consisted of kerosene, Span 80 and Amberlite LA-2 (as
the carrier). Figure 15 shows that penicillin G forms a carrier-penicillin-proton complex
(CHP) which traverses the membrane and decomposes at the inner interface. Penicillin
anion (P-) and a proton are released, and P- is enzynaatically converted to 6-APA and
phenylacetic acid (PhA). Ionised PhA, the free hydrogen and the free carrier form another
complex, which travels outward across the membrane and splits at its outer interface,
releasing PhA- and H + into the continuous phase. The carrier molecules are thus engaged
in two-way traffic, moving penicillin inward and phenylacetic acid outward. 6-APA accumu-
lates inside the aqueous droplets and is recovered later after splitting the emulsion.
In a Kuhni extraction column, 30% of the penicillin G was extracted and converted
in four hours and an enrichment factor of 7 to 8 was achieved for 6-APA [18]. These
results were improved later [12,22,59] and extraction efficiencies between 80 and 98%
could be achieved.
While four important applications have been described here, many others are
being developed. Table 1 presents their salient features. It may be seen that although
there are numerous and different applications, certain surfactants, carriers and membranes
are common to them. This has contributed to standardising LEM systems so as to make
them commercially feasible process units.

outer phase

membrane

PhA"

Figure 15. Penicillin G extraction and synthesis of 6-APA (from [18]). Reprinted with
permission of Butterworth Heinemann.
194 P.R. PATNAIK

¢
°,
e~
Q

÷
f
¢
O

~8
LIQUID EMULSION MEMBRANES 195

~ ~ ~
o o~ o~ ~

0 m

~J
0

0
_o

O0

b5 O0

O0
O0

o
o

O3
196 P.R. PATNAIK

b"
• i~ ~
t.,

o
@
=., ,.9.o ~d r 3 ~ o= @

il

e~ ee) O
t~

ii =l

[--., ca., ~
O
Z

O
t..t

II
,.a

"6 +
"6 ~o
o~ o
t'tl o
.o
,a

t~

E
°~
I:
II o
¢~ I=l
.e.-

g.i t"..
[ .t-
Z
O
LIQUID EMULSIONMEMBRANES 197

SUPPORTED LIQUID MEMBRANES


Despite their many advantages, certain drawbacks have inhibited the industrial use of
LEMs. The main drawbacks are: swelling, rupture and poor long-term stability. The last
problem has been attributed to the loss of the membrane by solubility, osmotic flow of
water across the membrane (which causes swelling) and the effects of shear [23,25]. The
use of supported liquid membranes (SLMs) has overcome or reduced these problems. The
membrane phase is immobilised inside the pores of a solid matrix or between two solid
supports, and the donor (or feed) phase and receiving (or strip) phase flow on either side.
There is no emulsification and hence no surfactant is needed. There are two common
configurations for SLMs: flat sheets and hollow fibres.

Flat Porous Supports


The membrane phase is immobilised inside the pores of a fiat solid sheet. If the
membrane phase is an organic solvent, the support material should be hydrophobic; with
an aqueous membrane, a hydrophillic support is used.

Hollow Fibre Supports


The support matrix comprises a bundle of hollow porous fibres contained inside an
impervious shell. Alternate fibres or alternate rows of fibres carry the feed solution and
the strip solution respectively. The inter-fibre spaces are filled by the membrane liquid
(Fig. 16). Depending on the hydrophobic/hydrophillic nature of the fibre, the membrane
phase and the feed and strip solutions, there are four basic possibilities for the interfaces
at which solute transfer takes place. The concentration profiles vary accordingly, as shown
in Fig. 17.
Since the membrane phase is contained inside the pores of a solid support, rupture
and swelling are much less than in unsupported LEMs. Nevertheless, some rupture may
occur due to the supports being defective. When this happens in flat sheet supports, the
feed and strip solutions mix directly and the integrity of separation is lost. On the other
hand, in hollow fibre systems only the feed or strip solution and/or the membrane phase
is affected. Small leaks of the membrane phase into some of the fibres do not seriously
affect the performance and if the membrane liquid is contaminated by an aqueous solution
it can easily be replenished [61]. However, fiat porous supports can be fabricated easily
in the laboratory and they permit direct in situ measurements, both of which are difficult
198 P.R. PATNAIK

in hollow fibre modules [23,62]. Moreover, the pressure drop in fibre bundles is larger
than in fiat sheets and this aspect may become important in large-scale applications,
especially if there are viscous liquids or there is evolution of gas.

f'

EATORSHELL

IIYDROPHILIC HOLLOW
FIBERS CARRYING THE. k
STRIPSOLUTION ~ ~ ~
HYDROPHILIC HOLLO~I ~:~::~:Z.'('~J
FIBERS CARRYING THE \ ~ ~ /
FEEDSOLUTI~

OR~IC LIQUID ;4U4BRANE --


CONTAINEDIN BE~EENTHE
~ 0 SETSOF HOLL~FIBERS TWO SETSOF MICROPOROUS
CARRYING AQUEOUSSOLUTIONS HOLLOW FIBERS CLOSELY
PACKED INSIDE THE SHELL
BUT SEPARATEDAT THE ENDS

Figure 16. Structure of a hollow fibre liquid membrane module/Reprinted from [61] by
courtesy of Marcel Dekker, Inc.

B, HYI)ROPHILIC SUBSTRATE
A..~op.o.,c SU.',aATE
I IORGANIC l I I IORGANIC I I

I ILo,, STRIP
I
I
I
I
I
I
I AOUEOUS
I STem

I IAO°~OUSI I I IAO.~OUSI 1
ORGANICF E E D ~ E M B R A N E ~ 4 I"~"RA'~l I
ORGANIC F E E O ~

I I---'-'-"-'1 I ORGANIC
II II sTR'P I
I
I
I
I
I
] ORGANIC
I STRIP

Figure 17. Concentration profiles for extraction by hollow fibre SLMs. Reprinted from [61]
by courtesy of Marcel Dekker, Inc.
LIQUIDEMULSIONMEMBRANES 199

Applications of SIMs
Since SLMs are of recent origin, there are fewer studies than for weU-mixed emulsions.
However, a number of commercially important processes have been studied (Table 1) and
some of them can be compared with their LEM counterparts.
Citric acid recovery has attracted the attention of many workers. Conventional
methods use either precipitation as calcium citrate, followed by filtration, or liquid-liquid
extraction followed by precipitation or back-extraction. Friesen et aL [63] used fiat sheet
SLMs to study citric recovery. The membrane phase consisted of Shell Sol 71, an aliphatic
hydrocarbon solvent, or Alkane 56, an aromatic hydrocarbon solvent. Trilaurylamine, a
tertiary amine, was the carrier, and higher alcohols were added to increase the partition
coefficient for citric acid. The support material was Celgard 2400, a microporous
hydrophobic polypropylene film. Citric acid could be recovered from fermentation beer
directly or as monosodium citrate. In principle, almost complete recovery was possible but
this required very low citric acid flux and low concentrations in the strip solution; therefore
Friesen and coworkers have suggested optimum conditions to balance extraction efficiency
with the concentration and flux.
Basu and Sirkar used hollow fibres for two applications. Polyprelene fibres (Celgard
X-10) were employed for both. In the first application, citric acid extraction [2.5], the
membrane phase had tri-n-octyl amine, also a tertiary amine, dissolved in xylene, heptane
or methyl isobutyl ketone. Plain water as well as sodium hydroxide solution were used as
stripping agents, the former yielding citric acid and the latter monosodium citrate. The
mass transfer rate remained almost constant over 1500 hours of continuous operation,
indicating good stability. About 99% extraction was possible and the degree of extraction
varied linearly with the membrane area, suggesting easy scale-up.
Their second work [64] was the reactive recovery of diltiazem, a cardiac drug with
calcium blocking activity. Excellent recovery (80-100%) was achieved with decanol as the
membrane phase and no carrier. To prevent reverse transfer of the drug, the strip solution
contained L-malic acid to convert dfltiazem to its malate. High concentrations of dfltiazem
malate could be obtained even from dilute feeds if the strip solution was recirculated.
They report that the extraction efficiency remained above 70% for 900 hours and that the
economics are competitive with liquid-liquid extraction.
Both fiat sheets and hollow fibre modules were used by Bryjak and colleagues [13]
to study the extraction of phenylalanine hydrochloride (PAH) and other amino acid
200 P.R. PATNAIK

hydrochlorides. Flat sheet supports were prepared from polyacrylonitrile by a procedure


described by the authors. They were soaked in the membrane phase (methanol and
1-decanol) and the carrier solution. A number of carriers were used. Hollow fibre modules
were of polysulfone or polyacrylonitrile or polyethylene. Membrane phase and carriers
were the same. The studies revealed some interesting findings: (i) hollow fibres produced
lower fluxes of P A H than fiat sheets but had superior reusability, (ii) P A H is transported
partially by solubility and partially by the carrier, but in only five out of 24 combinations
of carrier and activation solvent was the contribution of carrier-mediated transport more
than 50%, (iii) the fluxes did not depend on the hydrophobicity of the amino acid, contrary
to observations with LEMs [1], or on the acidity of the source medium, and (iv) the
recovery was sensitivity to pH, just as for lactic acid [65] and for L-valine [66].
The observed fluxes for several amino acid hydrochlorides were two to three orders
of magnitude higher than those reported previously for similar systems. An important
observation was that the SLM had to be activated in order to enable it to function.
Without activation there was no transport of solute. Activation was done with plain water
or with aqueous solutions of ethanol and propanol.
In a recent study, Lee and coworkers [23] have built on the work of Schugerl and
associates for the reactive extraction of penicillin G. Like Likidis and Schugerl [12], they
used Amberlite LA-2 as the carrier, which is a little surprising because Likidis and
Schugerl found diisotridecylamine (DITDA) to be equally good and recommended it since
it is cheaper. The liquid membrane consisted of 1-decanol and the carrier. Decanol was
selected because in this medium the equilibrium constant for complexation of penicillin
G with LA-2 is one to three orders of magnitude larger than for many other solvents [67].
A fiat hydrophobic polypropylene membrane was used for impregnating the organic phase
and the flux of penicillin G was measured in a Minitan-S (Millipore) system. The feed was
an aqueous penicillin G solution and the aqueous strip solution was buffered at different
p H values.
More than 95% of the penicillin G could be extracted if the p H was between 5.0
and 5.5 and the carrier concentration was greater than 20 mM. This appears to be as good
as the results of Likidis and Schugerl [12] for an LEM. But Lee et al.'s [23] results on mass
transfer offer useful information. Film resistance disappeared at flow rates exceeding 360
mL/min. At carrier concentrations less than 20 mM the mass transfer was controlled by
LIQUIDEMULSIONMEMBRANES 201

the liquid membrane resistance, beyond 200 mM aqueous film resistance was dominant
and both resistances were significant in the intermediate range.
The permeability of the membrane depended strongly on the flow rate, carrier
concentration and pH of the feed and strip solutions. It increased with the first two
variables and saturated at 360 mL/min and 20raM. The permeability also increased with
the pH of the strip solution but varied inversely with the pH of the penicillin G solution.
Although a low pH of penicillin G favours extraction, it also promotes undesirable
decomposition. The authors recommended pH 6-6.5 for the feed and pH 7 for the strip
solution.
Unfortunately Lee et aL [23] have provided no data on long-term performance and
reusability of the SLM. Nor is there a comparison with previous work based on LEMs
[12,22,38-40]. Since the extraction efficiencies are only as good as for LEMs and the
carrier concentrations are larger than for DITDA [12], it is difficult to recommend SLMs
using Amberlite LA-2 for penicillin G recovery until they have been shown to be better
for actual fermentation broth.
Some authors have combined an SLM module with a fermenter to achieve
continuous extraction of the product while fermentation is going on. Deblay et al. [66] used
such a configuration to separate L-valine from fermentation broth containing
Corynebacterium gultamicum. The SLM was a flat polytetrafluoroethylene (PTFE) sheet
impregnated with 1-decanol and Aliquat 336 (as the carrier). Stable operation was possible
for 18 days. A useful observation from a later study [68] by them is that permeability of
the membrane increases with the distribution coefficient of valine and with decreasing
viscosity of the organic phase and tortuosity of the support material. However, low
viscosity and low tortuosity reduced the stability of the liquid membrane, as did a high
concentration of the carrier. Thus a judicious choice of these variables is required.
Tsikas and coworkers [69] used hollow fibre modules as separators-turn-reactors.
The feed solution was penicillin G from a fermentation broth, maintained at pH 6 in 0.1M
phosphate buffer. The strip fibres had penicillin G arnidase in 0.1M phosphate buffer, pH
8, containing 100raM potassium iodide. The membrane phase in the inter-fibre spaces was
paraffin and decanol containing Adogen 464. Penicillin G was selectively transported into
the strip phase and converted into 6-APA according to the scheme shown in Fig, 15.
Other recent applications of SLMs have been for continuous extractive
fermentation of ethanol [62], recovery of carboxylic acids from spent medium [70],
202 P.R. PATNAIK

selective transport of L-pbenylalanine from a racemic mixture [71] and the isolation of
histamine produced by histidine deoxycarboxylase from histidine [72].

CONCLUDING REMARKS
Liquid emulsion membranes offer interesting possibilities for immobilising enzymes and
cells, for extracting fermentation products, for combining separation, concentration and
subsequent reaction in an integrated process and for numerous medical and biomedical
applications. Because liquid membranes can be tailored for specific applications and the
separation is based on the chemical nature of the solute, much better selectivity is possible
than with solid membranes [14]. This advantage also extends to immobilised enzymes and
cells because mass transfer resistances are smaller for LEMs than for solutes encapsulated
inside solid membranes.
Rupture and long-term stability were two major limitations in the commercial
exploitation of LEMs. However, the use of porous solid supports for the membrane phase
has reduced these problems considerably and now SLMs perform efficiently for long
durations [25,64]. Mass transfer across the liquid film (membrane) is often the rate
controlling step, as in many two-phase processes. Reducing this resistance enhances the
rate of extraction considerably but there have been only a few systematic studies [23,25,61]
to improve performance by optimising the operating strategy. Laboratory experiments
using commercially available solid membrane modules have demonstrated the viability of
SLMs under simulated industrial conditions for important products such as citric acid [25],
amino acids [13], penicillin G [69] and lactic acid [65].
To gain acceptability as a large-scale bioseparation technique, LEMs and SLMs
have to compete with other downstream methods. As mentioned in the Introduction, there
are distinct advantages over ion exchange, chromatographic separation, ultrafiltration and
ultracentrifugation, but the strongest competition may be from liquid-liquid extraction
(LLE), with which LEMs have some similarity. However, their three-phase nature,
carrier-mediated selectivity and encapsulation ability makes LEMs superior to LLE in a
number of situations: (i) the separation of biochemical zwitterions by manipulating the pH
in the two aqueous phases [1,56], (ii) the recovery of stereospecific isomers from racemic
mixtures [1,57], (iii) extzaction of pure compounds from liquids containing particulate
matter such as cells [20,62], (iv) separation of products with minimal loss and high purity
LIQUIDEMULSIONMEMBRANES 203

from dilute solutions [33,66], and (v) integrated processes where product recovery is
combined with further reaction catalysed by immobilised enzymes or cells [17,69].
Three classes of applications have potential for large-scale use of LEMs and SLMs
in fermentation processes. The first are processes where separation and reaction can be
combined through an SLM system, as in the extraction of penicillin G from fermentation
broth and its conversion to 6-APA [69]. The second are bioreactions in which the product,
which is often based on a genetically engineered strain, is expensive, is produced in low
titres and has to be recovered almost stoichiometrically in high purity. Examples are
streptokinase, L-valine [66], L-lysine [33], L-phenylalanine [1,72] and growth hormones.
No studies have yet been reported for the first and last examples, but they are good candi-
dates for a third area which has been suggested for LEMs. This involves the use of reverse
micelles as carriers in the membrane phase [14] for separation of proteins. The protein
in the bulk solution is trapped within the micelle, transported across the membrane and
released into the interior phase. The method thus combines the benefits of high selectivity
through carrier-mediated transport and the retention of the activity of the protein by
shielding it inside the micellar cage.
Given its numerous advantages, wide range of applications, versatility and ease of
scale-up, liquid membrane technology is on the threshold of being established as an
important bioseparation technique.

ACKNOWLEDGEMENT

Ritu Saihj Pal is thanked for preparation of the figures.

REFERENCES

Thien, M.T., Hatton, T.A. and Wang, D.I.C., BiotechnoL Bioeng., 32, 604-615
(1988). Separation and concentration of aminoaeids using liquid emulsion mem-
branes.

2. Li, N.N., U.S. Patent 3,410,794 (1988). Separating hydrocarbons with liquid mem-
branes.

3. Li, N.N., Ind. Eng. Chem. Proc. Des. Develop., 10, 215- 221 (1971). Separation of
hydrocarbons by liquid membrane permeation.

4. Melzner, D., Tilkowski, J., Mohrmann, A., Poppe, W., Halwachs, W. and Schugefl,
IC, Hydrometallu~, 13, 105-123 (1984). Selective extraction of metals by the fiquid
membrane technique.
204 P. R, PATNAIK

5. Terry, R.E., Li, N.N. and Ho, W.S., J. Membr. Sci., 10, 305-323 (1982). Extraction
of phenolic compounds and organic acids by liquid membranes.

6. Halwachs, W., Vokel, W. and Schugerl, IC, Int. Solv. Extr. Confr., Liege, Belgium,
1980, pp.80-88.

7. Ollis, D.F., Thompson, J.B. and Wolynic, E.T., AIChE J., 18, 457-458 (1972).
Catalytic liquid membrane reactor I: Concept and preliminary experiments in
acetaldehyde synthesis.

8. Cussler, E.L and Evans, D.F., J. Membr. Sci~, 6, 113 (1980). Liquid membranes for
separations and reactions.

9. Volkel, W., Bosse, J., Poppe, W., Halwachs, W. and Schugerl, IC, Chem. Eng.
Commun., 30, 55-66 (1984). Development and design of a liquid membrane reactor
for the detoxification of blood.

10. Scheper, T., Adv. Drug De~. Revs., 4, 209-231 (1990). Enzyme immobilization in
liquid surfactant membrane emulsions.

11. Schugerl, IC and Scheper, T., 4th Int. Confr. Chemical Biotechnol. Biological~ Active
NaturalProds., Budapest, Hungary (1987) pp.133-152. Biotransformation in enzyme
liquid membrane reactors.

12. Likidis, Z. and Schugerl, K., J. Biotechnol., 5, 293-303 (1987). Reactive extraction
and reextraction of penicillin with different carriers.

13. Bryjak, M., Wieczorek, P., Kafarski, P. and Lejczak, B., Z Membr. Sci~, 56, 167-180
(1991). Transport of amino acids and their phosphoric acid analogues through
supported liquid membranes containing macrocyclic carriers. Experimental
parameters.

14. Thien, M.P. and Hatton, T.A., Sep. Sc~ Technol., 23, 819-853 (1988). Liquid
emulsion membranes and their applications in biochemical processing.

15. Mukerjea, R.N. (ed.), Downstream Processingin Biotechnology, Tata McGraw-Hill,


New Delhi (1992).

16. Gadekar, P.T., Mukkolath, A.V. and Tiwari, K.IC, Sep. Sci. Technol., 27, 427-445
(1992). Recovery of nitrophenols from aqueous solutions by a liquid emulsion
membrane system.

17. Makryaleas, K., Scheper, T., Schugerl, K. and Kula, M.-R., Ger. Chem. Eng., 8,
345-350 (1985). Enzymatic production of L-amino acid with continuous coenzyme
regeneration by liquid membrane technique.

18. Scheper, T., Likidis, Z., Makryaleas, IC, Nowottny, Ch. and Schugerl, IC, Enzyme
Microb. TechnoL, 9, 625-631 (1987). Three different examples of enzymatic
bioconversion in liquid membrane reactors.
LIQUIDEMULSIONMEMBRANES 205

19. Protsch, M. and Marr, R., Proc. Int. Solvent Extr. Confr., Denver, Colorado (1983)
p.66.

20. Likidis, Z. and Schugerl, IC, BiotechnoL Bioeng., 30, 1032-1040 (1987). Recovery of
penicillin by reactive extraction in centrifugal extractors.

21. Likidis, Z., Schlichting, E., Bischoff, L and Schugerl, K., BiotechnoL Bioeng., 33,
1385-1392 (1989). Reactive extraction of penicillin G from mycel- containing broth
in a counter-current extraction decanter.

22. Lilddis, Z. and Schugerl, IC, Chem. Eng. Sc£, 43, 27-32 (1988). Continuous reactive
extraction of penicillin G and its reextraction in three different column types - a
comparison.

23. Lee, C.-J., Yeh, H.-J., Yang, W.-J. and Kan, C.-R., BiotechnoL Bioeng., 42, 527-534
(1993). Extractive separation of penicillin G by facilitated transport via carrier
supported liquid membranes.

24. Reschke, M. and Schugerl, IC, Chem. Eng. Sc~, 31, B19-B26 (1985). Continuous
reactive extraction of penicillin G in a Karr column.

25. Basu, R. and Sirkar, ICIC, AIChE J., 37, 383-393 (1991). Hollow fibre contained
liquid membrane separation of citric acid.

26. Hayworth, H.C., Ho, W.S. and Bums, W.A., Sep. ScL Technol., 15, 493 (1983).

27. Mohan, R.R. and Li, N.N., Biotechnol. Bioeng., 17, 1137-1156 (1975). Nitrate and
nitrite reduction by liquid membrane encapsulated whole cells.

28. Arnold, F.H., Trends BiotechnoL, 8, 244-249 (1990). Engineering enzymes for
nonaqueous solvents.

29. Hailing, P.J., Biotechnol. Adv., $, 47-84 (1987). Biocatalysis in multiphase reaction
mixtures containing organic liquids.

30. Florence, A.T., Omotosho, J. and Whateley, T.L, in Rosoff, M. (ed.), Controlled
Release of Drugs, VCH Publishers, N.Y. (1989), pp.163-184. Multiple w/o/w
emulsions as drug vehicles.

31. Frankelfeld, J.W. and Li, N.N., in Rousseau, R.W. (ed.), Handbook of Separation
Process Technology, John Wiley, N.Y. (1987), pp.840-861. Recent advances in liquid
membrane technology.

32. Meyer, E.R., Scheper, T., Hitzmann, B. and Schugerl, IC, BiotechnoL Techniques,
2, 127-132 (1988). Immobilization of enzymes in liquid membranes for enanti-
oselective hydrolysis.

33. Yagodin, G., Lopukhin, Y., Yurtov, E., Guseva, T. and Sergienko, V., Proc. lnt
Solvent Extr. Confr., Denver, Colorado (1983), p.385.
206 P.R. PATNAIK

34. Chilamkurti, R.N. and Rhodes, C.T., J. Appl. BiocherrL, 2, 17-24 (1980). Transport
across liquid membranes: Effect of molecular structure.

35. Clailamkurti, R.N. and Rhodes, C.T., J. Appl. Biochem., 2, 7-16 (1980). Transport
across liquid membranes: Effect of formulation variables.

36. Ho, W.S., Hatton, T.A, Lightfoot, E.N. and Li, N.N., AIChE J., 28, 662-670 (1982).
Batch extraction with liquid surfactant membranes: a diffusion controlled model.

37. Scheper. T., Barenschee, E.R., Barenschee, T., Hasler, A., Makryaleas, IC and
Schugerl, K., Ber. Bunsenges. Phys. Chem., 93, 1034-1038 (1989). A combination of
selective mass transport and enzymatic reaction: enzyme immobilization in liquid
surfactant membranes.

38. Reschke, M. and Schugerl, IC, Chem. Eng. J., 28, B1-B9 (1984). Reactive extraction
of penicillin G I:Stability of penicillin G in the presence of carriers and
relationships for distribution coefficients and degrees of extraction.

39. Reschke, M. and Schugerl, K., Chem. Eng. J., 28, Bll-B20 (1984). Reactive
extraction of penicillin G II:Distribution coefficients and degrees of extraction.

40. Reschke, M. and Schugerl, IC, Chem. Eng. J., 29, B25-B29 (1984). Reactive
extraction of penicillin G III: Kinetics.

41. Volkel, W., Halwachs, W. and Schugerl, IC, J. Membr. ScL, 6, 1 (1980). Copper
extraction by means of a liquid surfactant membrane process.

42. Stroeve, P. and Varansi, P., J. Colloid Interface Sci., 99, 360-373 (1984). An
experimental study on double emulsion drop breakup in uniform shear flow.

43. Calm, R.P., Frankenfeld, J.W., Li, N.N., Naden, D. and Subramanian, K.N., in Li,
N.N. (ed.), Recent Developments in Separation Technology, Chemical Rubber Co.,
Cleveland, OH (1982), 6, p.51. Extraction of metal ions by liquid membranes.

44. Frankenfeld, J.W., Cahn, R.P. and Li, N.N., Sep. Sci. Technol., 16, 385-402 (1981).
Extraction of copper by liquid membranes.

45. Colinart, P., Delepine, S., Trouve, G. and Renon, H., Z Membr. ScL, 20, 168-187
(1984). Water transfer in emulsified liquid membrane processes.

46. Mohan, R.R. and Li, N.N., BiotechnoL Bioeng., 16, 513-523 (1974). Reduction and
separation of nitrate and nitrite by liquid membrane encapsulated enzymes.

47. Matsumoto, S., Inoue, T., Kohda, M. and Ikura, K., J. Colloid Interface ScL, 77,
555-563 (1980). Water permeability in oil layers in w/o/w emulsions under osmotic
pressure gradients.

48. Magdasi, S., Frenkel, M., Garti, N. and Kassan, R., J. Colloid Interface ScL, 97,
374-379 (1984). Multiple emulsions II: HLB shift caused by emulsifier migration to
external interface.
LIQUIDEMULSIONMEMBRANES 207

49. Kopp, A.G., Marr, R. and Moser, EE., Inst. Chem. Engrs. @mp. Ser., 54, 279-290
(1978). A new concept for mass transfer in liquid suffactant membranes without
carriers and with carriers that pump.

50. Lorbaeh, D.M. and Hatton, T.A., Chem. Eng. ScL, 43, 405-418 (1988).
Polydispersity and baekmixing effects in diffusion controlled mass transfer with
irreversible chemical reaction: an analysis of liquid emulsion membrane processes.

51. Chaudhuri, J.B. and Pyle, D.L, Chem. Eng. ScL, 47, 41-48 (1992). Emulsion liquid
membrane extraction of organic liquids - I.A theoretical model for lactic acid
extraction with emulsion swelling.

52. Chaudhuri, J.B. and Pyle, D.L, Chem. Eng. ScL, 47, 49-56 (1992). Emulsion liquid
membrane extraction of organic liquids - II.Experimental.

53. Halwachs, W., Flaschel, E. and Schugerl, K., J. Membr. Sc£, 6, 33 (1980). Liquid
membrane transport - highly selective separation process for organic solutes.

54. Janakiraman, B., Sep. ScL Technol., 20, 423-443 (1985). Liquid membranes:
advancing reaction zone model for finite reactions.

55. Vyas, V.V. and Patnaik, P.R., 4th Int. Symp. Ana~t. Meth. @stems Strategies
BiotechnoL, Noordwijkerhout, Netherlands (1992). A modified reaction zone
approach to analyse reactive extraction with liquid membrane emulsions.

56. Itoh, H., Thien, M.P., Hatton, T.A. and Wang, D.I.C., BiotechnoL Bioeng., 35,
853-860 (1990). A liquid emulsion membrane process for the separation of amino
acids.

57. Ha, H.Y. and Hong, S.-A., BiotechnoL Bioeng., 39, 125-131 (1992). A study on
enzymatic reaction using a liquid emulsion membrane technique.

58. Scheper, T., Halwachs, W. and Schugerl, K., Chem. Eng. J., 29, B31-B37 (1984).
Production of L-amino acid by continuous enzymatic hydrolysis of D,L- amino acid
methyl ester by the liquid membrane technique.

59. Muller, B., Schlichting, E., Bischoff, L and Schugerl, K., Appl. MicrobioL Bio-
technoL, 26, 206-210 (1987). Reactive extraction of penicillin G in a pilot plant Karr
column II:Fermentation broths.

60. Boey, S.C., Del Corro, M.C.G. and Pyle, D.L, Chem. Eng. Res. Design, 65, 218-223
(1987). Extraction of citric acid by liquid membrane extraction.

61. Sengupta, A., Basu, R., Prasad, R. and Sirkar, K.K., Sep. Sci. Technol., 7,3,
1735-1751 (1988). Separation of liquid solutions by contained liquid membranes.

62. Christen, P., Minier, M. and Renon, H., Biotechnol. Bioeng., 36, 116-123 (1990).
Ethanol extraction by supported liquid membrane during fermentation.
208 P.R. PATNAIK

63. Friesen, D.T., Babcock, W.C., Brose, D.J. and Chambers, A.R., Z Membr. Sci., 56,
127-141 (1991). Recovery of citric from fermentation beer using supported liquid
membranes.

64. Basu, R. and Sirkar, tCK., Z Membr. Sci., 75, 131-149 (1992). Pharmaceutical
product recovery using a hollow fibre contained liquid membrane: A case study.

65. Scholler, (2., Chaudhuri, J.B. and Pyle, D.L, Biotechnol. Bioeng., 42, 50-58 (1993).
Emulsion liquid membrane extraction of lactic acid from aqueous solutions and
fermentation broth.

66. Deblay, P., Minier, M. and Renon, H., Biotechnol. Bioeng., 35, 123-131 (1990).
Separation of L- valine from fermentation broths using a supported liquid
membrane.
67. Schugerl, IC and Degener, W., Int. Chem. Eng., 33, 2%40 (1992). Recovery of low
molecular weight compounds from complex aqueous mixtures by extraction.

68. Deblay, P., Delepine, S., Minier, M. and Renon, H., Sep. Sci. Technol., 26, 97-116
(1991). Selection of organic phases for optimal stability and efficiency of fiat sheet
supported liquid membranes.

69. Tsikas, D., Kaltsidou-Schottelius, E. and Brunner, G., Chem. lng. Technik, 64,
545-548 (1992). Hollow fibre supported liquid membranes for the extraction of
penicillins and synthesis of 6-aminopenicillanic acid.

0. Yano, T., Nuchnoi, P., Nishio, N. and Nagai, S., Bioprod. Bioprocesses 2nd Confr.,
Japan/U.S. Jr. Pro]. Coop. Biotechnol. (1986), pp.281-293. Extraction of volatile fatty
acids from spent medium with a supported liquid membrane.

71. Molinari, R., De bartolo, L. and Diroli, E., J. Membr. Sci~, 73, 203-215 (1992).
Coupled transport of amino acids through a supported liquid membrane
I:Experimental optimization.

72. Takeshima, S. and Hiroso, S., Sep. Sci~ TechnoL, 26, 1195-1205 (1991). A organic
liquid membrane system for the separation and production of histamine.

73. Man', R. and Kopp, A., Chem. Ing. Technik, 52, 399-410 (1980).
Flussingmembran-technik- ubersicht uber phanomene, transportmechanismen und
modellbildunge.

You might also like