1686 Ind. Eng. Chem. Res.

1988, 27, 1686-1696

Energy-Saving Distillation Designs in Ethanol Production

Michael A. Collura*
Department of Chemistry and Chemical Engineering, University of New Haven, 300 Orange Avenue,
West Haven, Connecticut 06516

William L. Luyben
Department of Chemical Engineering, Lehigh University, Bethlehem, Pennsylvania 18015

A thorough study is made of energy-saving options for the distillation of a dilute, aqueous ethanol
solution to produce a distillate close to the azeotrope composition. Although presented in the context
of a fermentation ethanol plant, the system may be taken as a generic example of a dilute solution
with highly nonideal vapor-liquid equilibrium behavior. Conventional energy-saving techniques,
such as multiple heat-integrated columns, vapor recompression, and heat integration with the
evaporation section are analyzed and compared. The most economical alternatives and conditions
are identified. Sensitivity of the designs and costs to changes in design parameters and utility costs
is studied. For a 95 vol % ethanol distillate, the most economical designs included vapor recom-
pression or a two-column design with split feed for distillation and vapor recompression for the
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

evaporation section. For higher purity distillate, vapor recompression and a two-column design with
heavy split proved most attractive.

1. Introduction Table I. Distillation/Evaporation Design Configurations

In this work we examine energy-saving distillation op-
Downloaded via UNIV ULM on May 22, 2024 at 15:38:14 (UTC).

evapora-
tions for processing a dilute aqueous mixture which ex- case distillation tion comments
hibits highly nonideal equilibrium behavior. It is our intent 1 std column multieffect stand-alone
that this system be taken as a generic example of many 2 std column MVR0 stand-alone
such separations of commercial importance. The particular 3 std column multieffect energy-linked
4 MVR MVR stand-alone
system chosen for study is ethanol-water with a column MVR stand-alone
feed mixture as would be produced via grain fermentation
5 2-column/split feed
6 2-column/heavy split MVR stand-alone
and an alcohol product suitable for use as a fuel. Much 7 2-column/split feed/heavy split MVR stand-alone
interest has been expressed in the use of fermentation 8 2-column / prefractionator MVR stand-alone
ethanol as a premium fuel. Many in the industry (Eakin 9 complex heat integrated system integrated
et al. (1981) for example) advocate its use as a stand-alone 0
Mechanical vapor recompression evaporator.
fuel rather than as a gasoline additive (i.e., gasohol).
Strategic implications of the former option on petroleum However, for each design studied, the capital and operating
importation are obvious. The water content of a stand- costs are presented in sufficient detail that the reader may
alone alcohol fuel may be considerably higher than that draw his own conclusions using any economic criteria to
of an alcohol fuel additive. This permits a much simpler compare designs.
separation scheme since there is no need to employ com- For design purposes, only the binary system (ethanol-
plex techniques to break the ethanol-water azeotrope. water) was used in the distillation models. The effect of
Since separation is the most energy-intensive part of the trace components typically found in the feed stream (eg.,
production process (50% or more (Black, 1980)), a strong aldehydes, higher alcohols, etc.) was considered and found
economic incentive exists to study this part of the process. to have no significant impact on the design and operating
Considerable literature exists on the design of etha- conditions.
nol-water separation processes, and many authors directly After careful consideration of the possible design con-
address the energy-conservation issue. However, very little figurations, nine combinations of distillation and evapo-
information has been published concerning designs to ration section designs were chosen for further study. These
produce a hydrous alcohol product. Specifically lacking configurations are listed in Table I. The design and op-
in the literature is a comprehensive study which compares erating conditions for each case were optimized individu-
the various designs on an equivalent basis. The current ally. Comparisons were then made among the cases on an
study provides such a comparison for the hydrous fuel equivalent basis.
product case. In addition to its importance as a design
option for future commercial plants, the results of the 2. Survey of Published Designs
present study also provide a benchmark against which Only three relevant designs were found in the literature
alternative (e.g., nonthermal) separation techniques may for production of hydrous alcohol fuel. A design by Katzen
be compared. Associates (Brush, 1981; Ackley et al., 1980) uses a two-
The goal of this study was to define the optimum dis- column scheme in which the overhead vapor from a
tillation scheme to produce the desired product with regard high-pressure column reboils a second lower pressure
to both process conditions and the configuration of the column. The second published design (Mannfeld, 1981)
columns used. Heat integration with the evaporation is a small-scale, low-pressure column which uses a heat
section also considered. Designs and conditions were
was pump with an external refrigerant loop. Standiford and
chosen to minimize the annual operating cost, which
so as Weimer (1983) advocate a design in which the distillation
included both utility and fixed charges. Increases in capital and evaporation sections are heat-integrated. Use of vapor
costs were deemed acceptable only if they were accom- recompression in production of hydrous product is dis-
panied by an appropriate reduction in operating costs. A cussed by Eakin et al. (1981), although no design infor-
minimum value for the first year return on the incremental mation is given. It should be noted that there is consid-
investment of 15% was used to make this determination. erable discrepancy among the published energy-con-

0888-5885/88/2627-1686$01.50/0 © 1988 American Chemical Society

 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988 1687

Table II. Distillation Variables for Base Case and

Sensitivity Studies
variable base-case value sensitivity range
product rate 20 x 106 gal/year (4-60) x 106 gal/year
feed 2.9 mol % 0.5-9%
composition ethanol
top product 83 mol % 83-95%
composition ethanol
bottom product 0.01 mol % not varied
composition ethanol

sumption data. Thus, caution should be exercised in

comparing values from one author to another. Further
discussion of these designs may be found in Collura (1986)
along with a more extensive treatment of the relevant
literature on energy savings in distillation. Examples of
some relevant papers include a Department of Energy
Report (1980) and articles by Tyreus and Luyben (1975),
Flower and Jackson (1964), Stephenson and Anderson
(1980), and Petterson and Wells (1977).

3. Analysis Procedures
A set of base specifications, listed in Table II, were used
in evaluating the various design configurations. The
product rate of 20 X 106 gal/year is based on the pro-
duction rate of alcohol on an anhydrous basis. Thus, for
an 83 mol % product the total product rate would be about
21 X 106 gal/year. This size plant corresponds to a small
to medium plant by industrial standards. A product
composition of 83 mol % corresponds to the standard 95
vol % product. Higher product concentrations were in-
vestigated by running the binary separation at subat- Figure 1. Pressure versus azeotrope composition.
mospheric pressures to raise the azeotrope composition.
Extractive and azeotropic distillation were not included erature data is excellent in the range of pressures used for
in this study. Fermentation with yeast typically results the study.
in a beer containing 6-8 wt % alcohol. Thus, 2.9 mol % In addition to the design variables which were optimized,
(7 wt %) was chosen for the base-case value. By use of there were several design parameters which did not
a 90% on-stream factor, the base-case feed rate to the strongly influence the choice of design configuration.
column was 13 200 mol/h. For the several promising de- These parameters, which were fixed for all cases, included
sign configurations, sensitivity studies were conducted by the following: tray design (valve trays with diameter and
varying the base specifications over typical commercial spacing dependent on internal flow rates), tray efficiency
ranges (also shown in Table II). (Murphree efficiency of 50% in enriching and 40% in
For each configuration, design and operating conditions stripping sections), and other design variables such as
were optimized with the objective of minimizing operating heat-transfer coefficients, compressor efficiencies, tem-
cost while requiring a minimum return of 15% on any perature approach in exchangers, and cooling water and
incremental investment. The design variables included steam temperatures. Values for heat-transfer coefficients
reflux ratio, number of trays, feed location, feed temper- were average values in the ranges presented by Peters and
ature, column pressure, and number of evaporator effects. Timmerhaus (1979) for the various types of exchangers
Capital cost estimates made for the optimization were (e.g., 200 BTU/(ft2*h*°F) for alcohol vapor condensing
based on fourth quarter 1982 costs. against boiling water). Exchanger designs and materials
Preheating of the feed to the distillation section with of construction for columns and exchangers were com-
an appropriate bottoms product was included in all designs patible with those chosen by Raphael Katzen Associates
except those which expressly forbid it. Heat integration (1978). For example, 304 stainless steel was chosen for
with equipment outside the boundaries of the distillation column shells and internals and for reboiler tubes. See
and evaporation sections was not considered. For example, Collura (1986) for further details.
use of excess heat from the distillation section to dry solids 3.2. Evaporator Model and Design Basis. The
to form distiller’s dried grains was not considered. evaporator model was developed and run on a Tandy 1200
3.1. Distillation Design Model and Design Basis. HD microcomputer. Details are available in Collura (1986).
The steady-state design models for the distillation section Wherever possible, information from industrial contacts
were developed and executed on a Hewlett-Packard 3000 was incorporated into the model. This includes such items
series computer. Details of the models are given in Collura as heat-transfer coefficients, boiling point elevations, feed
(1986). At the heart of the design package is a rigorous and product solids concentrations, and minimum heat-
model for vapor-liquid equilibria based on Van-Laar ac- transfer temperature differences. Four different evapo-
tivity coefficients and a vapor-phase imperfection correc- rator configurations were included in this study: (1)
tion. Figure 1 shows the azeotrope composition versus multieffect evaporator, steam heated; (2) multieffect
pressure as predicted by this model along with several evaporator, heated by condensing vapor from an ethanol
points from literature sources (Hala et al., 1968; Pemberton distillation column; (3) mechanical vapor recompression
and Mash, 1978; Barr-David and Dodge, 1959; Otsuki and (MVR) evaporator; (4) complex heat-integrated distilla-
Williams, 1953). Agreement between the model and lit- tion/evaporation.
1688 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988

Table III. Cost of Utilities solids, see Collura (1986). For sensitivity studies, the solids
cost per
concentration was varied from 0.5% to 9.3%. A final
utility unit unit ref concentrate of 50% was required for all designs.
3.3. Operating Cost Estimation. The annual costs
steam 1000 lb $5.00 Luyben, 1984
cooling 1000 gpm $0.22 Peters and Timmerhaus, 1979 incurred in operating the distillation section of a fuel
water ethanol plant can be broken into fixed costs and variable
electricity kW-h $0.07 Peters and Timmerhaus, 1979 costs. Fixed costs were estimated as 20% of the capital
cost. Thus, the comparisons make no allowance for dif-
For the multieffect designs, the number of effects was ferences in such things as labor costs or maintenance
varied to find the economic optimum. As with the reflux
among the designs, except to the extent that higher fixed
ratio in the distillation columns, the incremental return
charges were assessed on designs with higher capital costs.
on investment was used in this analysis. As a practical
Variable costs were assumed to consist entirely of utility
constraint, the minimum temperature difference for heat costs for this section of the plant. The rates at which these
transfer was set at 8 °F, as suggested by contacts with were charged are shown in Table III (Luyben, 1984).
evaporator manufacturers (Carlson, 1985). In most cases, All compressors were assumed to operate with a me-
this constraint proved to be a more severe limitation than chanical efficiency of 80% and to be driven by electric
that provided by the minimum incremental return on in- motors. The electricity requirement for refrigeration was
vestment of 15%. based on a coefficient of performance of 9.0 and a me-
All evaporators were a vertical tube design with natural chanical efficiency of 50%. Steam pressure was 100 psia
circulation. The multieffect designs used a reverse-flow for the calculation of latent heat of vaporization, but can
pattern for liquids in order to (1) reduce viscosity of be throttled to obtain lower condensing temperatures in
streams with high solids content and (2) provide a more reboilers. The plant was assumed to operate 8760 h/year.
even distribution of the sensible heat loads across the
effects. A steam temperature of 230 °F and a final con-
denser temperature of 120 °F were used. For MVR
4. Analysis of Chosen Design Configurations
evaporation and energy-linked distillation/evaporation, the The nine design configurations are listed in Table I. The
energy source temperatures differ from this steam tem- energy-linked scheme (case 3) is one in which the overhead
perature. The specific values used depend on the partic- vapors from the distillation column provide energy to the
ular configuration being analyzed. first effect of a multieffect evaporator system. This kind
In fermentation ethanol plants, it is common to recycle of heat integration is only possible with the standard
about half of the bottoms from the beer still back to the column design for distillation. In two-column designs, the
fermentation step. This liquid contains a considerable operating pressure of the first column is limited by the
amount of nutrients and thermal energy. Thus, the solids need to maintain a fairly low temperature in the reboiler
separation step must handle about half the total bottoms (i.e., 260 °F). Otherwise the solids entering with the feed
stream from the column: 115000 lb/h at the base-case will cause severe fouling of heat-exchange surfaces. The
conditions. In the study of the design sensitivity to second column, which is reboiled by the overhead vapor
changes in feed concentration, this rate varied from 32 000 from the first, must therefore be operated at subatmos-
to 685000 lb/h. pheric pressure. The resulting saturation temperature of
The liquid leaving the bottom of the beer still contains the overhead vapor from the second column is too low to
very little ethanol (<0.004 wt %). In the analysis of the drive the evaporator effectively. In the MVR distillation
evaporation step, the presence of ethanol in this stream scheme, most of the energy content of the overhead vapor
was neglected. Before entering the evaporator, the column is used to reboil the column itself, leaving little for use in
bottoms stream is used to preheat the distillation feed the evaporator.
stream. The latter stream leaves the fermenter at 90 °F. The complex heat-integrated design is based on the work
The cooled bottoms stream leaves the feed preheater at of Standiford and Weimer (1983). It differs from the
a temperature of 98-108 °F, depending on the particular energy-linked design in the degree of complexity of the
design. After half of this stream is split off for recycle, the heat integration scheme used.
remainder is preheated using evaporator vapor condensate 4.1. Comparison of Designs. Table IV is a summary
before being fed to the evaporator. The resulting evapo- of the required capital investment and annual expenses
rator feed temperature is 120 °F for all designs except the for combinations of distillation and evaporation sections.
complex heat integration scheme, which differs because The least expensive to run is the combination using MVR
of the complicated flow pattern used. in both distillation and evaporation (case 4). A two-column
A solids concentration of 3 wt % was set for the evap- distillation design using the split-feed configuration (case
orator feed stream as the base case. Typically, such a 5) is a close second. The heat-linked standard column/
stream would contain about 1 wt % suspended solids and multieffect evaporator (case 3) was only slightly higher in
2 wt % dissolved solids (Zurawski, 1984). For a more operating costs. The high capital cost of the MVR system
detailed discussion of the nature and removal of these is justified by the savings in yearly costs which yield an

Table IV. Comparison of Total Costs for Distillation and Evaporation at Base-Case Conditions
capital costs, $106 annual costs, $106/year
case dist evap total dist evap total
std column dist, multieffect evap 1.1 2.4 3.5 2.3 1.7 4
std column dist, MVR evap 1.1 2.6 3.7 2.3 1.2 3.5
std dist integ, multieffect evap 1.3 2.4 3.7 2.4 0.8 3.2
MVR dist, MVR evap 2.6 2.6 5.2 1.5 1.2 2.7
split-feed dist, MVR evap 2.2 2.6 4.8 1.8 1.2 3
heavy split, MVR evap 1.6 2.6 4.2 2.4 1.2 3.6
HS/SF, MVR evap 1.9 2.6 4.5 1.8 1.2 3
prefractionator, MVR evap 1.9 2.6 4.5 2.4 1.2 3.6
complex integ (Standiford and Weimer, 1983) 4.4 3.5
 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988 1689

Figure 4. Two-column configurations.

Qc = Q. 1 5MM BTU/HR
173 F

Figure 3. Mechanical vapor recompression column: base case.

requires a much larger capital investment ($2.60 X 106),
incremental return on investment well in excess of the the reduction in operating costs (to $1.53 X 106/year) is
required 15%. enough to give a return of 53% on the additional invest-
In addition to the above designs, the use of inter-coolers ment.
and inter-heaters was studied briefly. This scheme was 4.2.3. Two-Column Designs. In the two-column de-
eliminated, however, as it appeared to provide no economic signs, the overhead vapor from one column (at higher
benefit in the ethanol-water separation. The nine designs pressure) condenses to reboil the second (lower pressure)
are discussed in more detail in the following sections. The column. Pressures are set to provide a 15 °F driving force.
standard column (case 2) and the three most economical This scheme theoretically has the potential to reduce the
cases (3, 4, and 5) are further analyzed for excursions be- energy requirement by one-half, although in practice this
yond the base conditions. cannot be achieved. Several different two-column flow
4.2. Stand-Alone Distillation. 4.2.1. Standard patterns (Figure 4) were studied for the desired separation.
Column. The final design and optimum operating pa- These differ in the manner in which products are taken
rameters for a standard distillation column at the base-case off. All two-column cases studied use forward integration,
conditions are shown in Figure 2 (capital cost, $1.10 X 106; i.e., column 1 is the high-pressure column. Reverse inte-
annual operating cost, $2.33 X 106/year). The energy gration is used in some separations but is undesirable for
consumption of 18.5 lb of steam/gal of ethanol (anhydrous the ethanol-water system because of the nature of the
basis) compares favorably with the range 15-18.5 reported vapor-liquid equilibrium. Such a pattern would require
by Maiorella (1983) for similar columns. The required that the distillate be taken from the high-pressure column
column diameter is 6 1/2 ft. The reflux ratio, number of and bottoms from the low-pressure column. However, the
trays, and column pressure shown represent the economic relative volatility for low concentrations is more favorable
optimum within the guidelines previously discussed. Al- at higher pressures, while that for high concentrations is
though vapor-liquid equilibrium behavior is more favor- more favorable at lower pressures. Similarly, other forward
able at lower pressures, operation under vacuum proved integration flow patterns are also possible, but were elim-
uneconomical due to increased capital investment (larger inated from consideration after brief examination.
diameter, thicker wall column) and operating cost (larger Split Feed. The simplest of the two-column configu-
latent heat value). rations is the split-feed case, shown in Figure 5. In this
4.2.2. Vapor Recompression Distillation. Addition design, the feed stream is divided and processed inde-
of a compressor to the standard column to pressurize the pendently in two columns. The only link between the two
overhead vapor is an expensive modification but results columns is via heat exchange. Both capital and operating
in substantial savings in operating costs. As shown in costs fall between that of the standard and the MVR de-
Figure 3, the column itself is almost identical with the signs.
standard column discussed previously. A slightly lower As pressure had little effect on the economics, the second
condenser pressure reduces by 2 the number of trays. column pressure was set by the available cooling water
Reflux ratios are essentially the same (3.3), although the temperature (120 °F on process side) to keep the bottoms
MVR column operates closer to its minimum reflux ratio. temperature of column 1 as low as possible. Pressure in
A temperature driving force of 15 °F was specified in the the first is high enough to provide a 15 °F heat-transfer
reboiler. Compressing the vapor to a saturation temper- driving force to reboil column 2. Bottoms temperatures
ature of 246 °F requires 2200 hp. While the MVR system much higher than the resulting 246 °F may lead to severe
1690 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988

Ft----------t* c-4f.;

CAP. COST: S1.63MM

ANN. COST: S2.42MM/YR

ENERCY: 34 4MM BTU/HR

* I , !

Figure 8. Base case: prefractionator.

Figure 6. Base case: heavy split.

--->3-*6C

! CAP. COST: S1.87MM

ANN. COST: S1.82MM/YR
j
ENERGY: 23.7MM BTU/HR
‘Tfl
'
-1
20.9km S"U/h»
> BTu/hR ! / '-'

Figure 7. Base case: heavy split/split feed. Figure 9. Steam-heated multieffect evaporator.

fouling of heat-transfer surfaces by solids in the reboiler. 1 .0 otrp 1.4 atm

Both columns were designed with reflux ratios of about

1.1 times the minimum, the economic optimum. Pro-
cessing 56% of the feed in the first column leads to bal-
anced heat duties for heat integration. A small trim re-
boiler on column 2 and a trim condenser on column 1 are
recommended for start-up and control purposes.
The energy requirement for the split-feed case is 22.0
X 106 BTU/h, about 60% of that required in the standard
column design. The goal of a 50% reduction in reboiler
duty cannot be achieved because of the low pressure in
column 2 and a substantial quantity of sensible heat
leaving in the bottoms streams. Further discussion of this Figure 10. Mechanical vapor recompression evaporator.
phenomena can be found in Collura (1986).
Heavy Split. The main disadvantage of the split-feed temperature of the column 1 bottoms (232 °F) is lower
design is the rather high temperature of the bottoms. A than that of the split-feed design; energy requirements and
design in which the first column operates as a stripper, operating costs are slightly higher. This design may be of
Figure 6, results in a lower bottoms temperature (227 °F). interest in cases where higher purity distillate is of interest
The feed is processed in series fashion, rather than parallel, since all distillate comes from the second, low-pressure
as in the split-feed design. All feed enters the first column.
(high-pressure) column. Vapor leaves this column at a Prefractionator. The last of the two-column designs
concentration of 22 mol % ethanol, the equilibrium vapor to be considered for the base-case conditions is the pre-
concentration for the feed, and is further processed in the fractionator. In this design, the first (higher pressure)
second column to provide the required distillate concen- column receives all of the feed. Intermediate products are
tration. produced at both ends of the column and fed to the second,
Energy consumption in this design is only slightly lower lower pressure column. The design specifications for the
than in the standard column. Such performance is due prefractionator are shown in Figure 8. Energy con-
to the relatively dilute feed stream and consequent high- sumption and cost figures indicate no advantage for using
energy requirement for the stripping action. Note that the this design in the ethanol-water separation.
condenser duty of the first column is much higher than Interested readers will find more detailed discussion of
the reboiler duty of the second, thus limiting the energy these designs in Collura (1986), including details of the
reduction potential of this design configuration. optimization procedure and explanation of the observed
Heavy Split/Split Feed. The energy loads of the two results.
columns in the heavy-split configuration can be balanced 4.3. Stand-Alone Evaporator. Two stand-alone
by diverting some of the feed directly to the second col- evaporator designs were studied: a steam-heated mul-
umn, thus combining the heavy-split and the split-feed tieffect design and a mechanical vapor recompression
designs. As shown in Figure 7, the first column is a stripper (MVR) design. The final designs are shown in Figures 9
which processes 69% of the feed. The second column has and 10.
two feeds: the remaining 31% of the feed as well as the Energy was provided to the steam-heated evaporator by
condensed overhead vapor from the first column. The saturated steam at 1.4 atm (230 °F). The optimum of six
 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988 1691

Figure 11.

Table V. Costs of Complex Heat-Integrated Scheme

capital, $106 (allocated part of
fixed capital investment) annual. $106/vear
beer stripper 0.314 steam 1.530
condensate stripper 0.260 electricity 1.124
rectifier 0.677 fixed costs 0.880
effect I 0.392
small compressor 0.427
effect II 0.500
effect III 0.547
compressor 1.280
total $4,400 $3,350

the one-fourth of the condenser duty not used by the

evaporator for the heat-integrated design discussed above.
However, brief examination of this scheme indicated no
Figure 12. Complex heat integration scheme for distillation/evap- reduction in costs.
oration. The more complex heat-integration scheme (Figure 12)
is based on an article by Standiford and Weimer (1983).
effects (3920-ft2 area each) had a capital cost of $2.4 X 106 In this scheme the vapor leaving the beer stripper is sent
and an annual cost of $1.7 X 106/year, primarily for the to a heat exchanger in the first evaporator effect where it
required 21000 lb/h of steam. is partially condensed to provide heat to that effect. The
The other stand-alone evaporator design used mechan- uncondensed vapor is further processed in a rectifier.
ical vapor recompression. Three effects (6960 ft2 each) Condensed liquid from this partial condensation along with
were optimal, with a 1530-hp compressor, yielding a capital bottoms from the rectifier are stripped in a condensate
cost of $2.6 X 106 and an operating cost of $1.2 X 106/year. stripper. Vapor leaving the stripper is combined with that
If a stand-alone evaporator was used, the MVR design is from the beer stripper. Steam to run both strippers and
preferred since it saves half a million dollars per year in the second evaporator effect is provided by compressing
operating expenses. The difference in capital investment the water vapor leaving the three evaporator effects. For
between the two is not significant. the strippers, this steam is injected directly into the base
4.4. Heat-Integrated Designs. Two schemes with heat of the column. Energy in the third evaporator effect is
integration of the evaporator and the distillation sections provided by fresh steam.
were studied and are shown in Figures 11 and 12. In the Detailed information was not provided in the reference
simpler design, a single distillation column is operated with to calculate energy requirements and costs. Consequently,
a condenser pressure of 2.0 atm (206 °F). Three-quarters some design parameters had to be set in order to perform
of its condenser duty is transferred to the first evaporator mass and energy balances and estimate costs. Values for
effect. The total energy requirement of this system is the pressures in columns and in the evaporator, the fraction
37 X 106 BTU/h reboiler duty of the column. A capital vapor condensed in the first effect exchanger, and tem-
investment of $3.7 X 106 includes five 5300-ft2 evaporator perature differences for heat transfer were all chosen to
effects, a 70-in.-diameter column with 75 trays, and various be consistent with the other designs studied.
smaller equipment items. Annual operating cost is $3.1 With the previous paragraph serving as somewhat of a
X 106, somewhat higher than case 4, but a viable design
disclaimer, the capital cost estimate of $4.4 X 106 and
choice. A major concern with this design is the high tem- annual operating expenses at $3.5 X 106/year place this
perature at the bottom of the column, 262 °F. This may design among the less economical of the design cases. The
lead to severe fouling of reboiler heat-transfer surfaces. cost breakdown is shown in Table V.
It is not possible to heat integrate the evaporator with
any of the other distillation designs, since the available 5. Sensitivity Analysis
energy is at too low a temperature for use in the evapo- In the discussion that follows, we examine the sensi-
rator. The only obvious combination is to use MVR on tivities of energy consumption and costs to changes in the
1692 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988

Table VI. Sensitivity of Operation Cost to Cost of Steam

and Electricity0
change in oper.
cost per 10%
configuration base cost, change in
distillation / evaporation $106/year steam electr.
1. std col/multieffect 4.0 0.29 0.00
2. std col/MVR 3.5 0.19 0.07
3. std col/energy-linked 3.2 0.18 0.00
4. MVR/MVR 2.7 0.00 0.17
5. split feed/MVR 3.0 0.11 0.07
6. heavy split/MVR 3.6 0.17 0.07
7. HS-SF/MVR 3.0 0.12 0.07
8. prefractionator/ MVR 3.6 0.16 0.07
9. complex heat integration 3.5 0.15 0.11
0 1-1-1-1-1—i—I—I—l-1--
0 2 4 6 8 10
°Steam cost =
$5.00 X 106 BTU; electricity cost =
$0.07/(kW-h). MOLE % ETHANOL IN FEED MOLE % ETHANOL IN FEED

Figure 13. Costs versus feed concentration.

base-case parameters. The results should prove helpful
75
in discerning trends in the effects of design parameters as 1
1
1
1
1
1

70
well as in applying the results obtained here to cases which w
>-
_

. ,

differ from the base-case conditions. 65 -

· · —

5.1. Utility Costs. Throughout this study, the costs

used were $5.00/1000 lb of steam and $0.07/(kW-h) of z 55
electricity. Clearly, these costs will vary significantly from 50 1_1_1 ...1_L_
one plant location to another (e.g., use of waste steam from 80 120 160 200 240

another process), and thus operating costs will also vary.

3.6
Table VI presents the expected variation in operating cost
1
r~1 r 1

as the cost of steam or electicity is varied.

Use of cost sensitivities lends an interesting perspective
O

§
3 2
/ -1

to the differences in estimated operating costs among the

2.8 -

/ :
designs. For example, a 10% increase in the cost of & 2 4 /
electricity combined with a 20% decrease in steam cost i _J_1
2 L-1--1-1-1-1-1-1- 1 1 1_

would cause cases 3 and 5 to be less expensive to operate 80 120 160 200 240

than case 4. If the criterion of 15% return on incremental Figure 14. Feed temperature of standard column.
investment is considered, a steam cost below $4.30 (elec-
ticity cost unchanged at $0.07/(kW-h)) makes case 3 more at the very dilute concentrations.
economical than case 4. At this steam cost, the savings 5.3. Feed Temperature. Shown in Figure 14 are the
in operating costs for the mechanical vapor recompression changes in the conventional column design needed to ac-
design is not sufficient to justify the extra capital cost of commodate a reduction in the feed temperature. The
the compressor. Thus, in the use of the results of this range of temperatures covered is from 90 to 218 °F. This
study, care should be taken to note the bases used in range covers the case of no heat exchange with the bottoms
making cost estimations and to adjust these as required (feed at fermentation temperature) through the case in
by the situation of particular interest to the reader. which the feed is 20 °F below the bottoms temperature.
5.2. Mole Fraction of Ethanol in Feed. The changes This latter case was used throughout the study for design
in capital and operating costs with changes in feed con- comparisons.
centration of ethanol are illustrated in Figure 13 for several The primary effect of reduced feed temperature is an
designs. These plots include both the distillation and the increase in the energy required in the reboiler. The re-
evaporation sections. In all cases, the feed rate has been boiler duty varies almost linearly with feed temperature.
adjusted to give a product rate of 20 X 106 gal/year (an- This results from the need to heat the large liquid mass
hydrous basis). The solids concentration in the feed to the from the feed temperature to that of the bottoms. Closer
evaporator was assumed to vary linearly with the ethanol examination of Figure 14 reveals a definite change in slope
concentration. Thus, for dilute concentrations, more water as temperature decreases. At lower temperatures, up to
must be evaporated in concentrating the solids. about 160 °F, the slope is equal to the flow rate times the
Some observations concerning general trends can be heat capacity. In the higher temperature range, 160-218
made based on the figures. There is considerable economic °F, the slope is much lower. This phenomenon is an ar-
incentive to increase the ethanol concentration in the tifact of the dilute feed stream. Some other results seen
fermentation broth. An increase from current levels of 2.9 earlier in the discussions of the designs can also be at-
mol % (7 wt %) to 4.5 mol % (11 wt %) will result in tributed to this condition. See Collura (1986) for further
savings of from $4 X 105/year to $1 X 106/year. An in- explanation of this phenomenon.
crease in concentration beyond 4.5 mol %, however, does The effect of feed temperature on capital cost of the
not yield as much of a reduction in costs. Both capital and column is minor, as the number of trays and column di-
operating cost curves tend to flatten out as concentration ameter do not change appreciably.
increases further. The figure also shows that schemes 5.4. Product Rate. Different sized plants have dif-
which produce a beer with a concentration lower than 2.9% ferent optimal designs. The trade-off between capital and
are very costly. Separation of such a feed could easily operating costs changes as the economy-of-scale principle
require an additional million dollars per year. works to reduce the unit cost of larger equipment. To test
There is essentially no change in cost ranking of the four the sensitivity of this parameter, the standard column and
configurations as feed concentration is changed. At higher the MVR column were designed for 4 and 60 X 106 gal/
levels, the differences are less pronounced than they are year of product in addition to the base-case designs for 20
 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988 1693

Table VII. Effect of Plant Capacity on Distillation Design and Cost

standard column MVR column
capacity, 106
gal/year 4 20 60 4 20 60
capital cost, $/gal 0.14 0.056 0.040 0.21 0.13 0.11
annual cost, $/gal 0.14 0.12 0.12 0.10 0.077 0.072
energy consumption0 16.9 16.5 16.5 0.767 0.717 0.717
reflux ratio 3.46 3.29 3.13 3.81 3.31 3.31
no. of trays 60 65 73 53 63 63

Units for standard column BTU/gal; for MVR column

0 =
103 =
(kW-h)/gal.

Table VIII. Costs for Distillation Designs for 90%

Distillate
energy,
capital annual cost, 106
design cost, $10e $106/year BTU/h
1. std col 4.34 3.87 49.8
2. MVR 6.20 2.98 3630 hp
3. heavy split 4.00 3.47 42.8
4. prefraction 4.20 3.54 45.0

X 106 gal/year. Results are shown in Table VII.

As expected, the cost per gallon of product decreases as
the plant capacity increases. The change in capital cost
is quite striking in going from 4 X 106 to 20 X 106 gal/year.
Operating expenses do not vary quite as much. Also, the
changes are generally more pronounced for the standard
column than for the MVR column. It is interesting to note Figure 15. Distillate concentration sensitivity of standard column.
the change in optimum reflux ratio with plant size. Again,
owing to the lower capital cost per unit of capacity at high
plant capacity, the reflux ratio is generally higher for the
smaller plants. The change in number of trays must, of
course, follow an opposite trend.

6. Higher Purity Distillate

In some circumstances, it might be desirable to produce
an ethanol product with concentration higher than 95 vol
% (83 mol %). For example, if the product is used directly
as a fuel, there is a trade-off between the energy content
per unit volume (related to the alcohol concentration) and
the cost of separation. Similarly, if the distillate is to be
further processed to produce anhydrous alcohol, the op-
timum concentration for feeding to the drying step may Figure 16. Standard column design for 90% distillate.
be higher than 83 mol %. In either case, it is necessary
to know the cost of increasing the ethanol concentration atm as the distillate concentration changes from 83 to 90
beyond that of the base case. mol % ethanol. This is the lowest pressure at which
Higher distillate purity can be produced in the standard cooling water can be used in the condenser. Operating
column by operating with a subatmospheric condenser costs increase rather sharply as 90% is approached due
pressure. Figure 1 shows the shift in azeotrope composition primarily to the increased energy requirement at higher
with pressure. Theoretically it is possible to obtain an- Xd values.
hydrous ethanol by distillation at a pressure below 0.05 Optimal design conditions for a 90 mol % distillate in
atm. However, the difference in volatility between ethanol the standard column are shown in Figure 16. The annual
and water above 95 mol % is so small that it makes it cost of separation is $0.19/gal (contained ethanol basis)
impractical to obtain distillate concentrations much higher versus $0.12/gal for the 83 mol % product: a 66% increase
than 90 mol % with vacuum distillation. in costs.
In the study of designs for higher purity distillate, only 6.2. Mechanical Vapor Recompression. Costs and
the distillation section was considered. The evaporation design parameters for mechanical vapor recompression
section remains essentially the same as for the base case columns are shown in Figure 17 versus distillate concen-
and thus need not be considered again. Energy con- tration. The trends are similar to those of the standard
sumptions and costs to produce a 90 mol % distillate are column except that the reflux ratios tend to be somewhat
summarized in Table VIII for four of the design configu- higher and the number of trays lower. This difference is
rations. due to the economic trade-off between capital and oper-
6.1. Standard Column. The changes in design needed ating costs and is related to the lower cost penalty asso-
to produce higher distillate concentrations in a standard, ciated with the use of a compressor in place of steam.
steam-heated column are presented in Figure 15. In- Changes in capital cost are more pronounced for the
creases in the number of trays and the column diameter MVR case than for the standard column because of the
account for the rise in capital cost. The diameter is in- added cost of the compressor. Horsepower requirements
fluenced both by the reduction in pressure and by the increase substantially as the distillate concentration is
increase in reflux with increasing Xd. The optimum raised because of two effects: (1) the increase in reflux
column operating pressure drops from atmospheric to 0.275 ratio results in a higher molar flow rate to the compressor;
1694 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988

90
Table X. Ratio of Auxiliary Condenser Duty to Compressor
' 1 1
1 1

i
/ : 5 80
Horsepower

Xd
compressor, aux. cond.,
ratio
-

CAP COST
70 _
10s BTU/h 102
***6

BTU/h
o
Z 0.83 5.59 0.15 0.03
0.90 9.24 3.25 0.35
*____ ^' . <pOST
-

‘
60 _3_i_i__i_
80 85 90 80 86 90
MOLE % ETHANOL IN DISTILLATE MOLE % ETHANOL IN DISTILLATE

80 85 90 80 86 90

Figure 17. Distillate concentration sensitivity of MVR column.

0olx = 3,3MM B’U/^R

Figure 19. Heavy-split design for 90% distillate.

the higher concentration product. This results from the

larger amount of work done per mole of vapor compressed
at the higher Xd values (for the reasons discussed above).
Compressor inefficiency (i.e., friction losses) contributes
to the energy content of the compressed vapor. As the
compressor power per unit mole of vapor increases, this
inefficiency contribution becomes more pronounced. This
is a very undesirable phenomeon since it is equivalent to
Figure 18. Mechanical vapor recompression column for 90% dis-
tillate. providing some of the reboiler heat by electric resistance
heating rather than from the latent heat available in the
Table IX. MVR Column Cost Advantage over Standard overhead vapor. The coefficient of performance is lost in
Column this portion of the duty, thus reducing the advantage of
% reduction in ROI for MVR
the MVR design with increasing Xd.
Xd annual costs vs std col
6.3. Two-Column Designs. Three different two-col-
umn configurations were studied: a heavy-split design, a
0.83 38 57
0.85 38 80 combined heavy-split/split-feed design, and a prefrac-
0.875 28 53 tionator design. A design which produces specification-
0.90 23 48 grade product in both columns (split-feed design) is not
suitable since the 90% distillate cannot be produced in the
(2) the increased number of trays results in a higher tem- first, higher pressure column. This column must operate
perature difference across the column, thus requiring a above atmospheric pressure in order to provide heat to the
higher compression ratio. second column and thus is limited to a distillate concen-
A sharp increase in all costs occurs as the 90% level is tration below the atmospheric azeotrope.
approached, as with the standard column. For the MVR Unlike the 83% case, the more difficult separation oc-
design, this trend begins at lower values of Xd due to the curs in the enriching section for the higher concentration
complex interaction between Xd and the compressor distillate. To reduce the energy consumption, it is nec-
horsepower. It is interesting to compare the savings in essary to split the enriching effort between the two col-
operating costs at various levels of Xd. The economic umns. However, since the high-pressure column is limited
advantage of the MVR system drops as the distillate to a distillate concentration of about 86 mol %, it is im-
concentration increases, as shown in Table IX. Even at possible to split the separation work in a way which will
the higher values of Xd, however, the higher capital cost significantly reduce the total energy requirement.
of the MVR system (compared to the standard column) A heavy-split configuration can be designed with a feed
can easily be justified on the basis of the large return on to the second column ranging in composition from 0.22 to
investment which is realized. 0.86 mole fraction of ethanol. A composition of 0.85 results
Details of the MVR column design for a 90 mol % in a balance in the energy requirements of the two col-
distillate are shown in Figure 18. The optimum pressure umns, as shown in Figure 19. The reboiler duty is about
is below that of the standard column because there is no 14% less than that of the standard column, and the capital
quantum jump in costs for the MVR design when the and annual costs are somewhat lower. Since the standard
pressure drops below 0.275 atm. By contrast, the standard column is designed for partial vacuum service, the diameter
column requires refrigeration to operate below this pres- is rather larger (10.5 ft), resulting in a very expensive
sure, thus making it an unattractive operating condition. column ($4 X 106). The heavy-split columns have fewer
Note that an auxiliary condenser is needed for the MVR trays and smaller diameters, leading to lower capital cost
design. The ratio of energy in the auxiliary condenser to ($3.4 X 106 for both). The cost of extra heat exchangers
energy supplied by the compressor is about 10 times needed for the two-column case is easily offset by the
greater for the 90% product than for the 83% product as savings in column cost.
shown in Table X. The change in the ratio above indicates With the heavy-split case, it is not possible to reduce
that the inefficient use of energy is more pronounced for the reboiler duty much below the value above because of
 Ind. Eng. Chem. Res., Vol. 27, No. 9, 1988 1695

heat integration scheme did not appear to be economical.

There is some financial incentive for the development
of fermentation processes which yield higher concentration
ethanol solutions. An increase to about 10 wt % seems
optimal. Beyond that level, there is little cost savings.
Plant capacity can have a significant effect on the choice
of operating and design parameters but does not seem to
alter the ranking of design configurations. The economic
optimum reflux ratio for large designs is closer to the
minimum reflux ratio than it is for smaller plants.
To produce a distillate concentration above 95 vol %
(83 mol %), the column must be operated at subatmos-
20. Prefractionator design for 90% distillate.
pheric pressure. This rules out the use of certain design
Figure
options which prove desirable for a lower concentration
the nature of the vapor-liquid equilibrium in the high- product, such as the split-feed design and heat integration
concentration range. The shape of the equilibrium curve with the evaporator. The vapor recompression design
in the region close to the azeotrope is such that the min- again provides significant energy savings. A two-column
imum reflux ratio is set by a pinch point occurring at about design with heavy-split pattern is also attractive.
85 mol % when the feed is above 22%. Thus, an increase The specific utility costs play a critical role in deter-
in the concentration of the feed to the second column does mining the best design for a given situation. Sensitivity
not reduce the reboiler duty, since the reflux ratio and the analysis shows that a moderate departure (e.g., 10-20%)
vapor rate do not change. from the costs used in this study can reverse the order of
A prefractionator design was also investigated for the design preferences. This must be kept in mind when ap-
90% product, as shown in Figure 20. This configuration plying the results presented here.
shows no apparent advantage over the heavy-split design
discussed above. Nomenclature
As seen in Table VIII, the standard column and pre- B =
bottoms flow rate from distillation column, mol/h
fractionator are higher in both capital and annual costs Bl, B2 = B for first or second column in a two-column design
than the heavy-split design., The extra $2.2 X 106 capital D = distillate flow rate from distillation column, mol/h
required for the vapor recompression design yields a 22% DI, D2 = D for first or second column in a two-column design
return in annual savings when compared to the heavy-split F = feed flow rate, mol/h for distillation column (solids-free
design. While this does exceed the criteria of 15% return basis); lb/h for evaporator
on investment, it is rather marginal, especially in light of HS = heavy split
the mechanical complexity involved in adding a com- L = flow rate of evaporator concentrate stream, lb/h
pressor. Choice between these two designs must ultimately MVR = mechanical vapor recompression
depend on the details of the specific application. Qr = reboiler duty for distillation column, BTU/h
Qc = condenser duty, BTU/h
7. Conclusions and Design Recommendations Q = heat-exchanger duty, BTU/h
can be stated that there is potential for
In general, it ROI = return on investment
saving energy and reducing operating costs in fuel ethanol RR = reflux ratio
RRmin = minimum reflux ratio
production based on the standard column commonly in
use. The unique features of the fermentation ethanol SF split feed
=

stream to be distilled are the dilute alcohol concentration Tf temperature of feed to column, °F
=

Wi = mass fraction solids in stream i

and the highly nonideal nature of the vapor-liquid equi-
Wf mass fraction solids in feed stream to evaporator
=
librium. These features distinguish this system from the
Xd, Xdl, Xb, Xbl mole fraction ethanol in respective
=

hydrocarbon separation schemes which have received distillate or bottoms streams

much attention in the literature. Within the constraints Zf mole fraction ethanol in distillation column feed stream
=

imposed by these features, several designs are possible

which reduce both the energy consumption and the op- Registry No. Ethanol, 64-17-5.
erating costs while returning at least 15% profit on the Literature Cited
extra capital needed compared to the standard column.
An obvious and important feature of any of the designs Ackley, W. R.; Moon, G. D.; Messick, J. R.; Brush, B. F.; Kaupiso,
is the provision for feed preheating with the bottoms K. F. (Katzen Associates), Prepr.—Am. Chem. Soc., Dw. Pet.
Chem. 1980, 25(4), 309.
stream. Savings in energy are very large and capital in-
Barr-David, F.; Dodge, B. F. J. Chem. Eng. Data 1959, 4(2), 107.
vestment very small for this step. It is the dilute feed Black, C. Chem. Eng. Prog. 1980, 76(9), 78-85.
concentration which makes this feature so important. Brush, B. F. (Katzen Associates) U.S. Patent 4306942, 1981.
For the assumed cost figures, mechanical vapor recom- Carlson, A. P. V. Equipment, Inc., Tonawanda, NY, personal com-
pression appears to be the least expensive to operate both munication, 1985.
for distillation and for evaporation. The relatively high Collura, . A. “Energy-Saving Distillation in Fuel Ethanol Produc-
tion: Steady-State and Dynamic Analysis”. Ph.D. Dissertation,
capital cost is justified by the reduction in annual costs Lehigh University, 1986. Available through University Microfilms
compared to other alternatives. International, Ann Arbor, MI.
Designs using two heat-linked columns can offer at- Department of Energy “Energy Conservation: A Route to Improved
tractive alternatives to MVR distillation, particularly if Distillation Profitability”. DOE 4431T1 and T2, 1980.
the cost of electricity is high or that of steam is low. A Eakin, D. E.; Donovan, J. M.; Cysewski, G. R.; Petty, S. E.; Maxham,
J. V. “Preliminary Evaluation of Alternative Ethanol/Water
split-feed design is least costly. Heavy split/split feed is
a good choice if the temperature of the column bottoms Separation Processes”. US DOE PNL-3823, 1981.
Flower, J. R.; Jackson, R. Trans. Inst. Chem. Eng. 1964, 42, T249.
is a serious concern. Hala, E.; Pick, J.; Fried, V.; Vilim, O. Vapor-Liquid Equilibrium
Simple heat integration of the distillation and evapo- Data at Normal Pressures; Pergamon: London, 1968.
ration sections shows some promise, while a more complex Luyben, W. L. Lehigh University, personal communication, 1984.
1696 Ind. Eng. Chem. Res. 1988, 27, 1696-1701

Maiorella, B. Ph.D. Thesis, University of California, Berkeley, 1983. Standiford, F. C.; Weimer, L. D. Chem. Eng. Prog. 1983, 79(1), 1.
Mannfeld, R. L. U.S. Patent 4 308106, 1981. Stephenson, R. M.; Anderson, T. F. Chem. Eng. Prog. 1980, 76(8),
Otsuki, H.; Williams, F. C. Chem. Eng. Prog. Symp. Ser. 1953, 1.
49(55), 55-68. Tyreus, B. D.; Luyben, W. L. Hydrocarbon Processing, 1975, 93.
Pemberton, R. C.; Mash, C. J. J. Chem. Thermodyn. 1978,10, 867. Zurawski, D A.P.V. Equipment Inc., Tonawanda, NY, personal
Peters, M.; Timmerhaus, K. Plant Design and Economics For communication, 1984.
Chemical Engineers, 3rd ed.; McGraw-Hill: New York, 1979.
Petterson, W. C.; Wells, T. A. Chem. Eng. 1977, 84(19), 1. Received for review August 6, 1986
Raphael Katzen Associates “Grain Motor Fuel Alcohol Technical Revised manuscript received April 18, 1988
and Economic Assessment”. US DOE HCP/J6639-01, June 1978. Accepted May 5, 1988

A Systematic Method for the Study of the Rate-Controlling Mechanisms

in Liquid Membrane Permeation Processes. Extraction of Zinc by
Bis(2-ethylhexyl)phosphoric Acid
Inmaculada Ortiz Uribe,f Supriya Wongswan, and E. Susana Pérez de Ortiz*
Department of Chemical Engineering, Imperial College of Science and Technology, London SW7 2BY,
England

A new systematic method based on a mathematical model of metal ion extraction with interfacial
reaction is described and applied to the study of the rate-controlling mechanisms in liquid membrane
permeation. Four different controlling regimes are predicted by the model depending on the range
of concentrations of the species involved and the hydrodynamic conditions of the contactor. Ex-
perimental results on the extraction of zinc by bis(2-ethylhexyl)phosphoric acid obtained in a spray
column and a stirred tank under a wide range of concentrations are analyzed using the proposed
method and are found to cover three of these regimes.

Following the pioneering work of Li in 1971, there has phase into the emulsion droplets (or internal phase). There
been an increasing interest in the study of the kinetics and are two types of facilitation methods used to increase the
mechanisms of liquid membrane permeation. The process mass-transfer rate across the membrane. In one case a
seems particularly attractive when very dilute solutions reactant is added into the internal phase so that the solute
are involved since the volume ratio between the stripping concentration at the membrane-internal phase interface
phase and the feed can be reduced drastically. The process is effectively zero, thus maximizing the concentration
is also capable of giving a higher degree of concentration gradients through the membrane. Chan and Lee (1984)
of solute in the extract in fewer stages while maintaining present a review of the various models which are commonly
the high selectivity of conventional solvent extraction. used to describe this type of facilitated mass-transfer
Substantial savings can also be made in the organic solvent phenomena.
inventory. In the other type of facilitated transport, a reactant is
Two forms of membrane geometry are commonly used, added to the membrane which reacts with the solute at
the liquid surfactant membranes or emulsion-type liquid the external phase-membrane interface, as schematically
membranes (Biehl et al., 1982; Bock and Valint, 1982; shown in Figure 1 for the transfer of zinc ion. The complex
Boyadzhiev and Kyvchoukov, 1980; Cahn et al., 1981; formed diffuses across the membrane, and on reaching the
Casamatta et al., 1978; Frankenfeld and Li, 1977; Frank- other side of the membrane, the reverse reaction takes
enfeld et al., 1981; Hochhauser and Cussler, 1975; Kitagawa place, regenerating the extractant and liberating the solute
et al., 1977; Kondo et al., 1979,1981; Kremesec, 1981; Lee into the internal phase. This mechanism of transfer is
et al., 1978; Marr et al., 1981; Martin and Davies, 1976/ usually called carrier-mediated transport.
1977; Melling, 1979; Nakashio and Kondo, 1980; Reddy In both types of facilitated transport, the reactions
and Doraiswamy, 1971; Reusch and Cussler, 1973; Schiffer taking place at the internal side of the membrane can be
et al., 1974; Schlosser and Kossaczky, 1980; Strzelbicki, very fast, so they are not expected to be rate controlling.
1978; Strzelbicki and Charewicz, 1978; Strzelbicki and Then the possible rate-controlling steps are diffusion in
Charewicz, 1980; Volkel et al., 1980) and the supported the continuous and membrane phases and chemical re-
liquid membranes (Barker et al., 1977; Carraciolo et al., action at the external interface for the carrier-mediated
1975; Chiarizia et al., 1983; Choy et al., 1974; Cussler, 1971; type of transport.
Danesi et al., 1981,1983; Imato et al., 1981; Komasawa et Most theoretical treatments of membrane kinetic be-
al., 1983). havior developed during the last decade, however, have
The surfactant liquid membrane is the continuous phase assumed extreme conditions in which negligible contri-
of an emulsion dispersed into a third phase. Usually, butions due to chemical reaction and aqueous-phase dif-
phases separated by a membrane are completely miscible. fusional process are assumed, as Komasawa et al. (1983)
The emulsion is stabilized by surfactants. In general, the and Danesi et al. (1981) point out in their respective works.
solute transfers through the membrane from the external Thus, only the diffusional process in the membrane has
been considered as a possible resistance to the permeation
+
Permanent address: Departamento de Ingeniería Química, process. The reason for this is most likely due to the lack
Facultad de Ciencias, Universidad del País Vasco, Apdo. 644, of available information on the kinetic behavior of the
Bilbao 48080, Spain. chemical reaction occurring at the aqueous-organic in-

0888-5885/ 88/ 2627-1696$01.50/0 &copy; 1988 American Chemical Society