journal of

MEMBRANE
SCIENCE
ELSEVIER Journal of Membrane Science 137 (1997) 231-239

Mass-transfer modeling of permeation of lactic acid across

amine-mediated supported liquid membranes
Ruey-Shin Juang*, Shwu-Hwa Lee, Ruey-Chang Shiau
Department o[ Chemical Engineering. Yuan Ze University Nei-Li. Taoyuan 320, Taiwan. ROC

Received 19 March 1997; received in revised form 12 August 1997: ~tccepted 12 August 1997

Abstract

The permeation of lactic acid from aqueous solutions through a supported liquid membrane containing tri-n-octylamine as
mobile carrier was studied experimentally and theoretically. The strip phase was either NaeCO3 solution or deionized water.
The membrane support was Durapore hydrophobic GVHP film. A permeation model was proposed considering aqueous film
diffusion, interracial chemical reaction, and membrane diffusion. The calculated rates were ffmnd to agree reasonably with the
measured ones (average standard error, 13%). Under the ranges investigated, it was shown that the present pernleation process
was governed by mixed interfacial reaction and membrane diffusion. V 1997 Elsevier Science B.V.

Kevwords: Mass-transfer modeling: Supported liquid membrane: Lactic acid; Tri-n-octylamine

1. I n t r o d u c t i o n studied [7,8], the contributions due to interracial

It has been shown that the amine-based extraction is
a prospective alternative to traditional calcium salt
precipitation process for recovering organic acids
from aqueous streams [1,2]. Certainly, the liquid
membrane technology is promising for this purpose
because it combines extraction and stripping processes
in one unit [3-6]. A supported liquid membrane
(SLM), which uses a porous-membrane support
impregnated with complexing carriers to separate feed
and strip phases, represents one of the feasible types of
liquid membranes [4-6].
Although the permeation of organic acids such
as lactic acid through amine-based SLMs has been
*Colrespnnding author. Fax: +886 3 4559373, e-mail:
cejuang (~,~ce.yzu.edu.tw.
reactions and aqueous film diffusion to the overall
process are usually neglected and only membrane
diffusion is regarded as a possible resistance. For
example, Ju and Verma [7] stated qualitatively
that the transport of lactic acid across an SLM
containing Alamine 336, a tertiary amine, is solely
controlled by diffusion of the a c i d - a m i n e complex
across the liquid membrane. Due to a considerable
effect of stirring on permeation rate, Lazarova
and Peeva [81 indicated that membrane diffusion
plays a significant role in the permeation of lactic
acid through bulk liquid membranes containing
Aliquat 336, a quaternary amine. Such simplified
types of mechanical analysis mainly result from the
lack of information on kinetic schemes of chemical
reaction occurring at or near the aqueous-organic
interface.

0376-7388/97/$17.00 .('. 1997 Elsevier Science B.V. All rights reserved.

P I I S()376-7388(97 )00206-8
232 R.-S. Juang et al./Journal of Membrane Science 137 (1997) 231 239

Basically, the extraction chemistry involved in SLM feed liquid strip

permeation is the same as that found in solvent phase membrane phase
extraction, but the overall process is controlled by
kinetic rather than equilibrium parameters [5]. It is bulk film film bulk
hence possible to model SLM permeation from a good
knowledge of extraction chemistry (mainly, kinetics) \HA-TOA J i
and the transport properties of the relevant geometry.
In this work, the permeation rates of lactic acid across HA .x
a TOA-SLM were measured and compared with the ~ HA
modeled results considering all possible resistances.
Rate-controlling mechanism for this permeation
process was also identified. /TOA

f,i s,i
2. Modeling of the permeation rates x:0 x=L
Fig. 1. Concentration profiles of the species across an SLM
2. ]. Chemistry o f solvent extraction containing TOA as mobile carrier. The asterisk indicates the strip
phase being Na2CO3.

When the aqueous pH is less than the pKa of lactic

acid (3.86), the extraction of lactic acid from aqueous
solutions with tertiary amines like tri-n-octylamine and 100 mol/m 3 Na2CO3 as a strip phase, respectively,
(TOA) is expressed by [9]: at 303 K [10].

pHA + qTOA e=~ (HA)p(TOA)q; Kpq (1)
where the overbar refers to the organic phase. The
equilibrium constant Kpq iS given by
Kpq = [(HA)r,(TOA)q]/{[HAJP[TO-AJq} (2)
In xylene, it was found that three types of the lactic
acid-TOA complexes (p,q) of (1,1), (1,2), and (3,1)
are formed [9]. The corresponding Kpq is Kll
=3.41 × 1 0 - 4 m 3 / m o l , Kl2 = 3.18x 10 s (m3/mol)2,
and K31 =5.48× 10 -H) (m3/mol) 3 at 303 K. Moreover,
the amount of lactic acid extracted by xylene alone is
negligible compared to that by TOA. Hence, the
contribution of unfacilitated transport in this SLM
process is not considered,
The kinetics of extraction and back-extraction of
lactic acid with TOA/xylene were separately investi-
gated using a membrane-based stirred cell [10]. Rate
equations for the formation and dissociation of the
acid-TOA complexes, R~ and R_l, are given by
Rj : kl [HA] [TOA] (3)
R_, = k I[T~OA]-2 [ ~ ] 2 (4)
where kt was 5.89×10 9m4/(mol s), and k ~ was
3.50x10 2 and 8.55×10 2mol/(m2s) with water
2.2. Permeation model in SLM
The concentration profiles of the species in the
present SLM system are shown in Fig. 1. It can be
seen that the following elementary steps are consid-
ered:
1. Diffusion of lactic acid in the feed phase toward
the membrane.
2. Reaction of lactic acid with the amine carrier to
form the complexes at the feed-membrane inter-
face (Eqs. (3) and (4)).
3. Diffusion of the complexes through the liquid
membrane.
4. Diffusion of the amine carrier through the liquid
membrane.
5. Reaction of the complexes with the strip agent at
the membrane-strip interface.
6. Diffusion of lactic acid in the strip phase outside
the membrane.
If the diffusion processes can be described by Fick's
equation, the permeation rates of these steps are given
as follows:
Jl = kHA,f([HA]f -- [HAJf.i) (5)
J2=k~[Ha]c~[UOA]fi-k ,[TOA]c~[ffN]2i (6)
 R.-S. Juang et al./Journal oj Membrane Science 137 (1997) 2.31 239
J3 kHA,m([HA]f.i- [HA]s,i )
J4 kToA,m([TOA]~.i - [TOA]f.i )
J5 = k 1[ T O A ] s ~ [ ~ ] ~ , i - k, [Ha]s.i[TOA]s.i
J~ = kuA.~([HA]~.i- [HA]s )
where HA represents the acid-TOA
(HA):,(TOA)q for simplicity,
Because the solubility of TOA in the aqueous phase
is negligible [ 11,12], the total amount of TOA in the
membrane phase is kept constant:
L prepared by dissolving lactic acid in water without pH
[TOA]o = (1/L) /
'0
([TOA] + [HA])dx
The stoichiometric ratio for HA of unity in Eq. (11 ) is
obtained from equilibrium results [9], meaning that,
on average, one TOA molecule reacts with one acid
per complex in the organic phase. If the concentration
profile of the species in the membrane phase is linear,
we have
[TOA]o = ( 1/2)([TOA]t.. i + [TOA]s,i )
+ (1/2)([HA],,i + IHA]~.i )
Since counter-diffusion of TOA and its complexes
occurs in the membrane phase, the continuity of the
total flux of TOA is held. In addition, at steady state we
have:
Jl J2 =: ,13 J4 = J5 = J~, = JHA
If the mass-transfer coefficients and reaction-rate
constants in Eqs. (5)-(10) are known, it is possible to
calculate the permeation rate, JHA, by solving the
system of Eqs. (5)-(10) with the additional conditions
of Eqs. (12) and (13).
3. Experimental section
3.1. Apparatus, membrane, and solutions
The stirred cell used was the same as that employed
earlier [13]. Two chambers were stirred at the same
rate (300 rpm) but in opposite directions. This speed
was sufficient to minimize the thickness of stagnant
liquid fihns. The entire cell was immersed in a ther-
mostat. The hydrophobic PVDF membrane support
233
(7) used (Millipore, GVHP) had a mean thickness of
125 lain, a mean pore size of 0.22 btm, and a porosity
(g) of 75%.
(9) The water used in this work was deionized by a
Millipore Milli-Q system. TOA, lactic acid, xylene,
(10) and other inorganic chemicals were supplied by
Merck. The membrane phase was prepared by diluting
complex TOAwith xylene and pre-saturating with the acid-free
aqueous phase. The pores in the membrane support
were filled with TOA under vacuum as reported earlier
[14], and the resulting membranes were immersed in
the membrane phase before use. The teed phase was
(11) adjustment. The strip phase was 100 mol/m ~ Na2CO~
or water. All aqueous phases were pre-saturated with
xylene.
3.2. Procedure
In each run, the membrane soaked with TOA was
clamped and the apparatus was assembled. The feed
and strip phases (255 cm 3, each) were then introduced
into the lower and upper chambers, respectively, and
the mixing started. When a steady state was reached
(12) (about 30 min). samples (3 cm 3) were taken from the
strip phase at preset time intervals and the original
solution was immediately added to maintain the
volume unchanged. The concentration of lactic acid
was determined by HPLC (Jasco, Japan) with a
LiChrospher 100 RP-18 column (Merck, 5 btm) and
(13) an UV detector at 210 nm. The mobile phase was an
equimolar (5 mol/m 3) phosphate buffer of H3PO4 and
(NH4)H2PO4 and the flow rate was 0.5 cm~/min.
Corrections due to volume replacement were also
made.
4. Results and discussion
4.1. Determination o f transport parameters
For a microporous membrane separating the
organic and aqueous phases, the individual mass-
transfer coefficients in the aqueous, organic, and
membrane phases, k~,, k,,, and k .... respectively, can
be obtained from the flux measurements of one hydro-
philic (e.g. acetic acid) and one hydrophobic (e.g.
iodine) species [15]. Because lactic acid is essentially
234 R.-S. Juang et al./Journal of Membrane Science 137 (1997) 231-239

Table 1 are found to be 141.9 and 98 cm3/mol, respectively, by

Valuesof parametersusedfor the calculationof permeationrates at the Tyn and Calus method [ 17]. The parachors for
303 K xylene, HA, TOA, and the complexes are estimated to
Parameter Lactic acid-YOA/xylene Reference be 285.1, 183.6, 1024, and 1713 cm 3 gl/4/St/2, respec-
Equilibrium constants: tively, by the method of additive group contributions
K~l 3.41 × 10 4 m3/mol [9] [17]. The parachor of lactic acid in xylene should be
K12 3.18× 10 8 (m3/mol)2 [9] doubled since it exists mainly as a dimer. It should be
K31 5.48×10 m (m3/mol)3 [9] noted that kHA,m and kTOA,m adopted here are on an
Reaction rate constants: average basis and assumed to be constant at whole
kl 5.89×10 9 m4/(mols) [10] amine concentrations. This effect will be discussed
k I (Na2CO3) 8.55× 10-2 mol/(m2 s) [10] below.
k 1 (water) 3.50× 10`-2 mol/(m2 s) [101
Mass-transfer coefficients: 4.2. Comparison of the calculated and measured
kHA,f, kHA,s 6.54× 10 5 m/s This work rates
knA.m 2.76× 10 - 6 m/s This work
kTOA,m 2.86x 10-° m/s This work
Fig. 2 illustrates the variation of lactic acid con-
centration in the strip phase and contact time. All data
points lie on respective straight lines passing through
more soluble in water than xylene [9], the value of ka the origin. The measured permeation rate JnA was
in the feed phase can be determined in this way from therefore obtained by the initial-rate technique accord-
diffusion experiments using the same stirred cell and ing to the following:
membrane support [15]. The result is listed in Table I, JHA = (V~/S)(d[HAk/dt),_o (14)
where kHA,s is assumed to equal knA.f.
Due to the ease of analysis, the mass-transfer Figs. 3-6 show the effects of the concentrations of
coefficient of the lactic acid complex, rather than lactic acid in the feed phase and of TOA in the
lactic acid itself, in the membrane phase km was
measured in this work. The fluxes were experimen- 10 , , ~ r ~ , , f
tally obtained using the same stirred cell, where the I I I ' t
complex was transported from a pre-loaded organic z5 [TOA]o
phase to a fresh organic phase through the same /- key](mol/rrO
membrane support. The concentration of the complex 8 // zx
El'I 200
50
was analyzed by HPLC after the sample was com- © s00
pletely stripped with 100 mol/m 3 Na2CO3. Due to the ,-"
lack of organic-aqueous interface, km can be directly ~ 6
obtained from the slope of the plot of concentration vs.
time which equals l/[(1/km)+(2/ko)]. The individual "-"
mass-transfer coefficient of the complex in the organic ~~
~ 4 a/ / ~
layer ko is estimated as below.
Here, the values ofko and km for TOA, and ko for the
complex are estimated from ka for HA and km for the 2
complex based on thewhererelationshipsis kmocDj and ka
the kinematicthat ~ ~ l "
(or ko) cx D2/3~ -1/6, ~ viscosity
of the medium [16]. The estimated results are also 0
listed in Table 1. In this correlation, the diffusivities of 0 5 10 is 20 2s
species in the media Dj are estimated by the Hayduk
and Minhas equation [ 17]. The value of 7 for xylene at t × 10-3 (s)
3 0 3 K is 7.67×10-7m2/s. The molar volumes of Fig. 2. Variationofthe strip-phaseconcentrationoflacticacidwith
xylene and lactic acid at their normal boiling points contact time.
 R.-S. Juang et al./Journal of Membrane Science 137 (19~J7) 231-239 235

10E-3 ~-r-r~-,
E
i,i~i111~,
r,,
,Iii;,I,]'"112
Ike,,l,i~e, IOA
1,,~ 3
1.0E-3!~
,II,III,J---
~--~--~-
F2 ' - _oo i S A A -

~" _, ~" ,El A

-'

1.0E_7 ~ I i I Ill[ll I I t IIIttl I ] I ,Itlll I 1 III 1,0E_7 ~_ I

1.0E+O 10E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+ 1 10E+2 10E+3

[HA]r (mol/m3) [TOA]o (mol/m3)

Fig. 3. Effect of lactic acid concentration in the feed phase on the Fig. 5. Efl;ect of initial TOA concentration in the membrane phase
permeation rate with water as a strip phase, on the permeation rate with 'a.ater as a strip p h a s e

1.0E-3 _--
i I 1111,,, 1 i 1111H, i I I,,I11, I I lllIT~
~
IOE-3 ~ i I I~-~-~-~rTT---r-
kev line [TOA]o 3
O" I m,50
re'a)L,2CR) ~ ..'~A
A ~/'k ~4 i ~{IH~'I' ru°Hn'~i ~ 3
A/s 8oo /.
lOE-4 j ,, ~ 2 _~ ,OE~ - L [] I ~ I ~'_!'_~

i ph~LsC Na ( ( )
Strip phase= Na2CO~
!

1.0E+0 1.0E+I 1.0E+2 1.0E+3 1.0E+4 1.0E+I 1.0E+2 1 0E+3

[HA]r (mol/m3) [TOA]o (mol/m3)

Fig. 4. Effect of lactic acid concentration in the feed phase on Fig. 6. Effect of initial TOA concentration m the membrane phase
permeation rate with 100 mol/m ~ NaeCO~ as a strip phase, on the permeation rate with 100 mol/m ~ Na,C()~ as a slrip phase.
236 R.-S. Juang et al./Journal of Membrane Science 137 (1997) 231-239
membrane phase on the permeation rate with water
and 100 mol/m 3 Na~CO3 as a strip phase, respectively,
It is found that the transport rates increase with both
concentrations of lactic acid and TOA carrier. The
solid curves in these figures are calculated by the
present permeation model. Close agreement between
the measured and calculated rates is obtained under
the conditions studied (average standard error, 13%).
However, relatively large deviations are observed at
high concentrations of lactic acid and TOA. The
overestimated rates are mainly due to the oversimpli-
fication of the model that kHA,m and kTOA,m remain
unchanged for the described concentration range of
TOA. In fact, an increase in the viscosity of the
membrane phase at high concentrations of TOA and
lactic acid, due to possible aggregation of the acid-
TOA complexes, would result in a decrease in the
values of kin.
As clearly shown in Figs. 3-6, the TOA-SLM gives
a larger rate while using 100 mol/m 3 Na2CO3 as a strip
phase, compared to the use of water. This is because
in the former case the following reaction occurs,
N a + + H A ~ , N a + A + H + when the lactic acid-TOA
complexes diffuse and reach the membrane-strip
interface. The equilibrium constant is reported to be
1.29 × 10 7 (mol/m3)2 at zero ionic strength and 298 K
[9]. The amount of the undissociated acid at the
interface [HA]~,i and in the bulk phase [HA]s will
be small because this strip phase has pH 11.36 at
298 K.
4.3. Rate-controlling mechanisms
To elucidate the permeation mechanism, the rates
under extreme conditions of aqueous film diffusion,
interracial chemical reaction, and membrane diffusion
control are calculated. They are given as follows:
(1) aqueous film diffusion control, JHA,a,
JHA,a = kHA,f[HA]f (15)
(2) chemical reaction control, JHA,~,
J,A,c ------k, [HA]/[T-OA]f,i - k-1 [YOA]f,i
-2 [HA]f,i
-- 2
= k_, [T-OA]~,i2[~]~,i - kl [HA]~[T-O-A]~,i (16)
(3) membrane diffusion control, JHA,m,
JHA,m = kHA,m[HA]f.i (17)
where [HA]f,i in Eq. (17) can be obtained from the
mass balance of lactic acid in the membrane phase
based on equilibrium relationships [9].
Conceptually, the fractional resistance of each step
to the overall process, Aa, A~, and Am, can be
obtained from Eqs. (18)-(20).
~Xa=(I/JHA'a)/{(I/JHA'a) + (I/JHA,c) + (1/JHA,m)}
(18)
~c=(I/JHA,c)/{(]/JHA,a) ~-(I/JHA.c)-~ (l/JHA,m)}
(19)
/km=(l/JHA.m)/{(]/JHA,a) + (I/JHA,c) ~- (1/JHA,m)}
(20)
By comparing the above A values, we can quantita-
tively identify the rate-controlling steps for this SLM
co-transport process. Table 2 lists A n, A c, and Am
calculated at different conditions. It is found that the
contribution of diffusional resistance of lactic acid in
the aqueous stagnant film is negligibly small under
the ranges studied. That is, the permeation process is
predominantly governed by mixed interfacial chemi-
cal reaction and membrane diffusion.
Although the contribution of aqueous film diflhsion
is negligible, its role is rather significant at low [HAIl
and high [TOA]0. The rate-controlling nature of aqu-
eous film diffusion at extremely low solute concentra-
tion and high carrier concentration has also been
reported earlier in the SLM counter-transport of
metal ions containing chelating extractant as mobile
carriers [13,18-20] and in the co-transport of citric-
acid-containing amine carriers [14]. In practice, the
aqueous-phase mass-transfer coefficients of solutes
in SLM systems are often determined under such
conditions.
On the other hand, the role of membrane diffusion
becomes more important at low [HA]f and high
[TOA]0; furthermore, Ac is relatively large under
the conditions of high [HAIl and low ITOA]0. These
findings are different from those observed earlier in
the co-transport of citric acid across a TOA-SLM [141
and in the SLM counter-transport of metal ions con-
taining chelating extractant [I 3,18,20]. These discre-
pancies are probably due to different types of rate
equations (for example, reaction orders).
It should be noted that the present mixed mechan-
ism is different from the literature one [7,8]. Lazarova
and Peeva [8] investigated the recovery of lactic acid
 R.-S. duang et al. ~Journal of Membrane Science 137 (1977~ 231-239 237

Table 2
Fractional resistances calculated at different conditions at 303 K

IHAh rTOA]~l Case 1 '~ Case 2 ''

(mol/m;, (mol/m ~)

I0 50 nil 0.454 0.546 nil 0.4i)3 0.5ql7

50 nil 0.589 0.411 nil 0.553 0.447
20(I nil 0.727 0.273 nil 0.705 0.295
500 nil 0.842 0.158 nil ().82t,b (I. 171
I0 200 0.002 0.374 0.624 0.002 0.316 (1.682
50 0.002 0.484 0.514 0.002 (i.43~ ().562
2(10 (I.O01 0.627 (I.372 (t.O(ll 0.59(I I).41)9
500 0.001 0.768 0.231 1).001 (I.744 0.255
10 800 0.008 0.331 0.661 l).ll09 0.268 0.723
50 0.007 0.415 0.578 0.008 I).355 0,637
200 0.006 0.546 0.448 0.()07 0.496 11.497
5()(1 0.005 0.701 0.294 0.006 0.663 0.331
21) 20 nil 0.583 0.417 nil 0.547 I1.453
50 nil 0.508 (/.492 nil (1.463 (I.537
1O0 0.001 0.457 0.542 0.001 11.407 (t.592
200 tl.002 (I.415 0.583 0.002 (L361 I).(>~7
500 0.005 0.375 I).620 0.0(t5 I). 314 l l.(~8 I
800 0.0(18 0.361 11.631 11.009 0.29 ~) 0(~92
100 20 nil 0.729 0.271 nil 0.70~ 0.291
5(1 nil 0.656 0.344 nil 0.626 0374
100 nil 0.600 0.400 0.001 0.564 0 435
200 0.001 0.548 (1.451 0.002 0.505 0493
500 0.004 0.492 (I.504 0.0(t4 0.440 0.556
800 0.007 0.471 11.522 0.007 0.41 li.578

~'Strip phase is deionized water (Case I) or 100 mol/m~ Na2CO~(Case 2).

by bulk liquid membrane containing Aliquat 336.
Owing to a large effect of stirring on the permeation
rate, they suggested the contribution of diffusion
processes across the two feed/organic and organic/
strip interfaces to the overall resistance of mass trans-
fer to be significant. In addition, Ju and Verma [7]
examined the transport of lactic acid in an SLM
containing Alamine 336 and suggested that diffusion
of complex across the liquid membrane alone is rate-
limiting. On the other hand, Reisinger and Man" 121l
proposed a model to obtain mass-transfer resistances
in lactic acid permeation by liquid surfactant mem-
branes using Amberlite LA-2 as mobile carriers. They
indicated that it is controlled by the rate of chemical
reaction under optimal conditions. These discrepan-
cies may partly result from different structural proper-
ties of the membrane support such as the thickness,
porosity, pore size, and tortuosity. To our knowledge,
this work first identifies the significance of chemical
reaction on a quantitative basis. It is also noted from
Table 2 that this SLM permeation Is more reaction-
controlled with water as a strip phase compared to the
use of Na2CO3 solution. This is because the latter
gives a fast dissociation of the complexes at the strip-
membrane interface.
In fact, the present mechanistic results are also
inconsistent with those obtained in other organic
acid/carrier systems [12,22-24]. For example, the
permeation of citric acid from fermentation beer
across SLMs containing trilaurylamine [22] or TOA
[23] could be qualitatively described by assuming
membrane diffusion control [22]. In a slightly differ-
ent approach. Basu and Sirkar [ 121 analyzed the
separation of citric acid from aqueous streams with
TOA in methyl isobutyl ketone in microporous hydro-
phobic hollow-fiber extractors. They assumed that the
reaction between citric acid and TOA occurring at the
interface on the organic side is very rapid, compared to
aqueous- and organic-phase film diffusion processes.
Similarly, Nuchnoi et al. [24] proposed that diffusion
238 R.-S. Juang et al./Journal of Membrane Science 137 (1997) 231-239

of the complex inside the membrane is rate-control- S membrane cross-sectional area (m 2)

ling in the extraction of volatile fatty acids like formic t time (s)
and acetic acids by a TOPO (tri-n-octylphosphine TOA tri-n-octylamine
oxide)-SLM. However, Basu and Sirkar [11] modeled V volume (m 3)
the separation of citric acid with TOA by diffusion x distance from the feed-membrane inter-
processes only inherent in the system in a hollow fiber face (m)
contained liquid membrane, and found that there [] molar concentration of species in the
exists an indirect evidence for interracial resistance brackets (mol/m 3
in the back-extraction step.
6.1. Greek letters

5. Conclusions A a, Ac, A m fractional resistances of aqueous film

diffusion, interfacial reaction, and
Rates and mechanism of lactic acid permeation membrane diffusion to the overall
from aqueous solutions across an SLM containing process, respectively
TOA as mobile carriers have been examined. A
permeation model is presented considering all 6.2. Subscripts
possible resistances based on a good knowledge
of extraction chemistry and hydrodynamics of f, m, s teed, membrane, and strip phases, respec-
the stirred cell. The calculated rates agree well tively
with the measured ones (average standard deviation, i aqueous-membrane interface
13%). The rate-controlling steps of the present process 0 initial (total)
are quantitatively identified by comparing the
values of A a, Ac, and A m listed in Table 2. Under 6.3. Superscript
the conditions studied ([HA]t=10-2000mol/m -~,
[TOAJ0 -- 20-1200 tool/m3), this co-transport SLM overbar species in membrane or organic phase
process is controlled by combined interfacial chemical
reaction and membrane diffusion (At+A,,>0.99);
also, the role of membrane diffusion becomes domi- Acknowledgements
nant at low [HA]f and high [TOA]0, and vice versa.
This work was supported by the ROC National
Science Council under Grant No. NSC86-2214-
6. List of symbols E155-002, which is greatly appreciated.

HA lactic acid
J permeation rate (mol/m 2 s) References
k mass-transfer coefficient (m/s)
kl rate constant for complex formation de- [11 J. Hartl. R. Marr. Extraction processes for bioproduct
fined in Eq. (3) (mY/tool 2 s) separation, Sep. Sci. Technol. 28 (1993) 805.
k_] rate c o n s t a n t for complex dissociation I2] C.J. King, Amine-based systems for carboxylic acid recovery,
Cbemtech 22(5) (1992) 285.
d e f i n e d in Eq. (4) ( m o l 2 / m 5 s) [31 C. del Cerro, D. Boey, Liquid membrane extraction, Chem.
Kpq extraction equilibrium constant defined in Ind. 7(1 I) (1988) 681.
Eq. (2) (tool/m3) ~l-p q) 14l A.E. Eyal, E. Bressler, Industrial separation of carboxylic and
L membrane thickness (m) amino acids by liquid membranes: Applicability, process
Rj reaction rate of complex formation per considerations, and potential advantages, Biotechnol. Bioeng.
41 (1993) 287.
unit area (mol/m 2 s) 151 R.D. Noble, C.A. Koval, J.J. Pellegrino, Facilitated trans-
R_t r e a c t i o n rate of complex dissociation per port membrane system, Chem. Eng. Prog. 85(3) (1989)
unit area (mol/m 2 s) 58.
 R.-S. Juang et al./Journal of Membrane Science 137 (1997) 231-239
[61 J.J. Pellegrino, R.D. Noble, Enhanced transport and liquid
membranes in bioseparations, Trends Biotechnol. 8 (1990)
21~5.
[71 L.K. Ju, A. Verma, Characteristics of lactic acid transport in
supported liquid membranes, Sep. Sci. Technol. 29 (1994)
2299.
181 Z. Lazarova, L. Peeva, Facilitated transportoflacticacidin a
stirred cell, Biotechnol. Bioeng. 43 (1994) 907.
191 R.S..Iuang, R.H. Huang, Equilibriurn studies on reactive
exlraction of lactic acid with an amine extractant, Chem. Eng.
J.(~5 (1997147.
[10[ R.S. Juang, R.H. Huang, Kinetic studies on lactic acid
extraction with amine using a microporous membrane-based
stirred cell, J. Membr. Sci. 129 (1997) 185.
111] U. Basu, K.K. Sirkar, Hollow fiber contained liquid
membrane separation of citric acid, AIChE J. 37 (1991) 383.
[I 21 R. Basu, K.K. Sirkar, Citric acid extraction with microporous
hollow fibers, Solvent Extr. lon Exch. 10 (1992) 119.
[131 RS. Juang, R.H. Lo, Kinetics of the coupled transport of
vanadium(W) from sultktte solutions through supported liquid
membranes, Ind. Eng. Chem. Res. 33 (1994) 11)1 I.
[141 RS. Juang, L.J. Chen, Analysis of the transport rates of citric
acid Ihrough a supported liquid membrane containing trim-
octylamine, Ind. Eng. Chem. Res. 35 (1996) 1673.
[151 R.S. Juang, R,H. Lo. Mass-transfer characteristics of a
membrane permeation cell and its application to the kinetic
sludies of solvent extraction, Ind. Eng. Chem. Res. 33 (19941
1OOl
?,39
[16] V.G. Levich. Physicochemical Hydrodynamics, 2nd edn.,
Prentice-Hall, Englewood Cliffs, NJ, 1962, pp. 6(J 72.
I17] R.C. Reid, J.M. Prausnitz, B.E. Poling, The Properties of
Gases and Liquids. 4th edn. McGraw-Hill, New York, 1987,
pp. 52-55 and 598-606.
[ 18] A.B. de Haan, RV. Barrels, J. de Graauw, Extraction of metals
from waste water. Modeling of the mass transfer in a
supported liquid membrane process, J. Membr. Sci. 45 (1989)
281.
119[ A.K. Guha, C.H. Yun. R. Basu, K.K. Sirkar, Heavy metal
reruoval and recovery by contained liquid membrane
permeator, AIChEJ. 40(19941 1223.
12(/I R Plucinski, W. Nitsch, Calculation of permeation rates
through supported liquid membranes based on the kinetics of
liquid-liquid extraction, J. Membr. Sci. 39 (1988) 43.
[211 H. Reisinger, R, Mart, Comparison of the separation of lactic
acid and L-leucine by liquid elnulsion membranes, J. Mcmbr.
Sci. 80 (199~) 85.
[221 D.T. Friesen, W.C. Babcock. l),J. Brose, A.R. Chambers,
Recovery of citric acid from fermentation beer using
supported liquid membranes. J. Membr. Sci. 56 i l991~ 127.
[231 G. Rajalo, N Livantsova, M. Arro, Transpor! of organic acids
through a supported liquid membrane. Proc Est~mma Acad
Sci. Chem. 40 (1991) 193.
124] R Nuchnoi, T. Yano, N. Nishio, S. Nagai, Extraction ol
volatile fatty acids from diluted aqueous solution using a
supported liquid membrane, J. Fermem. Technol. 65 11087)
301.