ISSN 00405795, Theoretical Foundations of Chemical Engineering, 2013, Vol. 47, No. 3, pp. 239253. Pleiades Publishing, Ltd.

., 2013.

Design and Economic Analysis of the Technological Scheme

for 1,3Propanediol Production from Raw Glycerol1
J. A. Posadaa, C. A. Cardonaa, J. C. Higuitaa, J. A. Tamayob, and Yu. A. Pisarenkoc
a Departamento de Ingeniera Qumica, Universidad Nacional de Colombia, Manizales, Colombia
b
Departamento de Ingeniera Industrial, Universidad Nacional de Colombia, Manizales, Colombia
c
Moscow State University of Fine Chemical Technology
email: ccardonaal@unal.edu.co; pisarenko_yu@mail.ru
Received May 22, 2012

AbstractA technological scheme for producing 1,3propanediol from raw glycerol was designed, simu
lated, and economically assessed. The production process was composed of three main stages, namely: glyc
erol purification, glycerol fermentation, and 1,3propanediol recovery and purification. First, a typical
stream of raw glycerol was purified up to 98 wt %, and then the fermentation took place in a two continuous
stages process by means of a Klebsiella pneumoniae strain. For the fermentation stage, a rigorous analysis was
carried out using a kinetic model considering both substrate and products inhibition. Thus, multiplicity of
steady states and hysteresis loops were studied for the first fermentation stage. Also, in order to optimize both
the outlet concentration of 1,3propanediol and its productivity, three different objective functions were ana
lyzed. As result, each objective function led to an optimal condition, such as: the highest global yield to 1,3
propanediol (0.599 mol/mol), the highest outlet concentration of 1,3propanediol (0.512 mol/L), and the
highest global productivity (1.157 102), respectively. Then, the downstream process for 1,3propanediol
recovery and purification was designed based on a reactiveextraction process and a reactivedistillation pro
cess. This downstream process was applied to each scenario analyzed on the fermentation stage. Finally, the
three scenarios were economically assessed and the lowest production cost was obtained for the third sce
nario. Simulation process and fermentation analysis were performed using Aspen Plus and MatLab respec
tively, while the economic assessment was carried out using the Aspen Icarus Process Evaluator.
DOI: 10.1134/S0040579513030093

1
INTRODUCTION
Glycerol is a byproduct of triglycerides transesterifi
cation with alcohol in the biodiesel production and it is
obtained in a weight ratio of 1 : 10 (glycerol/biodiesel)
[1]. Because of the high functionality of glycerol, it is a
particularly attractive raw material for chemical and
biochemical synthesis [2, 3]. Also, due to the increment
in the use of biodiesel, a growth in the glycerol produc
tion could cause an oversupply and consequently an
inevitable invasion of glycerol in the market [4, 5]. For
this reason glycerol could become a cheaper feedstock
for chemical synthesis obtained from biosustainable
sources [6]. In this sense, the biodiesel production
plants should work on glycerol transformation to prod
ucts with an added value and a high demand [7, 8].
Glycerol bioconversion to metabolites has been
widely studied in the last decade and many works have
been focused on 1,3propanediol production since it
exhibits a wide range of potential uses [9]. 1,3Pro
panediol can be used as an intermediate for the synthe
sis of heterocycles and as a monomer for polyesters syn
thesis (e.g., polypropyleneterephthalate (PPT) and
polyethyleneterephthalate (PET) [10]), where 1,3pro
panediol improves their chemical and mechanical
1 The article is published in the original.
properties in comparison with other conventional
monomers [11]. Also, biodegradability of plastics con
taining 1,3propanediol is higher than those totally syn
thetic polymers [12, 13]. 1,3Propanediol can also be
used as lubricant, solvent, and functional fluid (e.g.,
antifreeze, unfreezing, and transfer of heat) [14]. 1,3
Propanediol has other important uses as an additive in
foods, cosmetics, liquid detergents, wetter of tobacco,
paintings, flavoring and fragrances, and as a precursor
in the chemical and pharmaceutical industries [15,
16].Commercially, the oversupply of crude glycerol has
caused a price fall from US $1/lb in 1996 to US $0.07/lb
in 2010 [17]. Additionally, 1,3propanediol is produced
at large scale and it has a growing market with an annual
production up to one million pounds in the United
States and an annual market growth of 4%, with a sale
price around US $0.71/lb [18].
Fermentative production of 1,3propanediol
(13PD) under anaerobiosis takes place in two parallel
ways. In the first one, a fraction of glycerol is oxidized
by glyceroldehydrogenase (GlycDH) to dihydroxy
acetone (DHA), and then phosphorrylated by DHA
kinase to enter glycolysis. The remaining glycerol is
then dehydrated to 3hydroxypropionaldehyde
(3HPA) by glyceroldehydratase, where reduction con
tinues by propanedioldehydrogenase (PPDDH) and

239
240 POSADA et al.

by a dependent NAD oxidorreductase to 1,3pro
panediol [19, 20]. 1,3Propanediol production can be
performed biologically by several bacterial strains such
as Klebsiella pneumoniae (KPNEUMON), Citrobacter
freundii, Enterobacter agglomerans, Clostridium butyri
cum and Clostridium acetobutylicum [21, 22]. K. pneu
moniae and C. butyricum are the most commercially
promising bacterial strains because of their high yield,
productivity, and resistance to both substrate and prod
uct inhibition. Among these two bacteria, K. pneumo
niae DSM2026 has been presented as one of the most
appropriate bacterial strain for glycerol fermentation to
1,3propanediol [23].
On the other hand, recovering and purification of
1,3propanediol from the fermentation broth is a com
plex task because of both its high hydrophilic character
and its high boiling point. The purpose of this article is
to design and analyze the complete technological
scheme for the production of 1,3propanediol using raw
glycerol as feedstock.
process). The resulting product was neutralized with an
acid solution and formed salts and ashes were discarded
by centrifugation. The clarified mixture was then
washed with water and in a later evaporation process
more than 90% of the water content and the remaining
methanol were withdrawn. Finally, the required quality
of glycerol (98 wt %) was achieved throughout a distilla
tion process [17]. This quality of glycerol was the one
required for the fermentation process.
Glycerol Fermentation. The material balances for
continuous glycerol fermentation in one single stage
were solved using two independent variables namely the
glycerol concentration in the feed stream and the dilu
tion rate. Since, acetic acid and ethanol were also pro
duced during glycerol fermentation to 1,3propanediol
by K. pneumoniae, the material balances in a dynamic
state needed to be solved for biomass, substrate, and all
the obtained products as shown in equations (1) to (5):
Out
dX i
dt
(
= Di X i X i
In
)
Out
+ i X i ,
Out
(1)

dCGOut
METHODS
dt
,i
( )
= Di CG,i CG,i qG,i X i ,
In Out Out
(2)
In order to perform the design and economic analy
Out

( )
sis of the 1,3propanediol production process from raw dCPD
= Di CPD,i CPD,i + qPD,i X i ,
,i In Out Out
glycerol, the complete technological scheme was first (3)
dt
stated using a sequential strategy based on knowledge. Out
This strategy allowed not only designing but also com
paring process alternatives, considering mainly techno
dCHAc
dt
,i
(
= Di CHAc In Out
)
,i CHAc,i + qHAc,i X i ,
Out
(4)
economic criteria and leading to a technological con Out
figuration with high performance [7, 8, 2426]. After
obtaining a high performance technological scheme for
dCEtOH
dt
,i
(
= Di CEtOH In Out
)
,i CEtOH,i + qEtOH,i X i ,
Out
(5)
1,3propanediol production, the process simulation where X, CG, CPD, CHAc, and CEtOH are the concentra
took place. Aspen Plus (Aspen Technology, Inc., USA) tions for biomass (g/L), glycerol (mol/L), 1,3pro
was the main tool used for defining, structuring, speci panediol (mol/L), acetic acid (mol/L), and ethanol
fying, and simulating the complete technological (mol/L), respectively, D is the dilution rate (i.e., ratio
scheme. Then as result of the simulation process, mass between volumetric flow rate and reactor volume)
and energy balances were obtained which gave the (h1), is the specific rate of cellular growth (h1), qG
requirements for raw materials, consumables, service is the specific rate of glycerol consumption, and qPD,
fluids, and energy requirements. Finally, these results qHAc, and qEtOH are the generation rates of each prod
were used to estimate the capital and operational costs
of 1,3propanediol production from raw glycerol using uct (h1). Subscript i indicates the fermentation stage
the software Aspen Icarus Process Evaluator (Aspen for a multistage system. In the ith fermentation stage,
Technologies, Inc., USA). In and Out superscripts indicate the in and the out
conditions, respectively.
For the production process of 1,3propanediol from
raw glycerol three main stages can be distinguished, The kinetic model of glycerol fermentation by
named: glycerol purification, glycerol fermentation, K. pneumoniae has been previously explained [2830].
and 1,3propanediol recovery and purification. Specific rates of cell growth, substrate consumption,
and products formation are given in equations (6)
Glycerol Purification. A typical composition of a raw to (11):
glycerol stream obtained in the biodiesel production
process is: methanol 32.59 wt %, glycerol 60.05 wt %, CG,i
NaOCH3 2.62 wt %, fats 1.94 wt %, and ash 2.8 wt % [7, i = max
CG,i + K S
8, 17]. The purification process of raw glycerol up to dif (6)
ferent commercial qualities (crude glycerol at 88 wt % CG,i CPD,i CHAc,i CEtOH,i
and pure glycerol at 98 wt %) was previously designed 1 1 1 1 ,

and analyzed by Posada et al. [7, 8, 17]. Initially the raw *
CG *
CPD *
CHAc CEtOH
*
glycerol was subjected to an evaporation process where CG,i
90% of methanol at 99 wt % was recovered and recycled qG,i = mG + mi + qGm , (7)
to the transesterification process (biodiesel production YG CG,i + K S*

JOURNAL OF CTHEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

 DESIGN AND ECONOMIC ANALYSIS OF THE TECHNOLOGICAL SCHEME
CG,i
qPD,i = mPD + i * YPD + qPD
m m
, (8)
CG,i + K PD
*
CG,i
qHAc,i = mHAc + i * YHAc + qHAc
m m
, (9)
CG,i + K HAc
*
qEtOH,i = qG,iY(EtOH
m
G),i ,
b1 b2
Y(EtOH G),i = +
m by substrate or products, fermentation systems could
. (11)
c1 + DiCG,i c2 + DiCG,i
Equation (6) describes the specific rate of cell growth
which represents a kinetic model with inhibition by
both substrate and products. The kinetic parameters
* and C *
CG*, C PD
* , C HAc EtOH are the critical concentra
tions (i.e., the concentration where biological activity
is stopped). The required parameters to solve the
kinetic model (valid at 37C and at neutral pH [24,
26]) are the maximum specific growth rate max=
0.67 h1 and the Monod saturation constant Ks = 2.8
104 gmol/L.
Constants for specific rate of substrate consump
tion and product formation are mG = 2.20 103,
mPD = 2.69 103, mHAc = 9.7 104, YGly m
= 8.2,
YPD = 6.769 10 , YHAc = 3.307 10 , qGly = 2.858
m 2 m 2 m method:
102, qPDm
= 2.659 102, qHAc
m
= 5.74 103,
* = 1.143 102, K PD
K Gly * = 1.55 102, K HAc
* = 8.571
2 objective function was the volumetric productivity in
10 . Also, b1, b2, c1, and c2 constants in equation (11)
are 2.5 105, 5.18 103, 6 105, and 5.045
102 mol/(L h), respectively.
The critical concentrations required in equation (6)
were taken from [22]. These concentrations have an
average global deviation of 8.6% for 29 stable states of
glycerol fermentation by K. pneumoniae and C. butyri
cum [29, 30]. The critical concentrations for glycerol,
1,3propanediol, acetic acid, and ethanol are 2.012,
0.8975, 0.7798, and 0.3975 mol/L, respectively.
On the other hand, glycerol fermentation to 1,3
propanediol by K. pneumoniae presents a metabolic
overflow of products causing dynamic phenomena of
nonlineal behavior such as multiplicity of steady states,
hysteresis, and oscillations [29]. Thus, it was required to
determine the operational conditions that take to mul
tiplicity of steady states.
In order to obtain the best performance in the first
fermentation stage, the volumetric productivity was
optimized using the LevenvergMarquardt method
[31]. Volumetric productivity is one of the most
important functions to be optimized from an operative
point of view, since it implies a high production in a
small reactive volume on a short period of time. The
volumetric productivity is shown in equation (12),
where Pv1 is the volumetric productivity, mol/(L h),
CPD1 is the outlet 1,3propanediol concentration,
mol/L, and D1 is the dilution rate, h1. The first fer
THEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47
241
mentation stage in this equation is indicated by the
subscript 1:
Pv1 = CPD1D1. (12)
In this case, when multiplicity occurs, the steady states
with the highest concentration of 1,3propanediol
were considered. This ensured that the steady states
(10) with the higher volumetric productivities were evalu
ated. Regardless of the inhibition phenomenon caused
be successfully performed in two continuous stages
simultaneously reaching a high productivity and prod
uct concentration in the first and second fermentation
stages respectively [28, 32].
To assess the glycerol fermentation process in two
continuous stages, three optimization models were
used. The first optimization model was a sequential
procedure in which the outlet stream from the first fer
mentation stage (operated at optimal conditions) was
directly fed on the second one. The volumetric produc
tivity on the second fermentation stage was calculated
as a function of both the dilution rate and the achieved
change on the 1,3propanediol concentration as shown
in equation (13). Also, the maximum outlet concentra
tion of 1,3propanediol on the second fermentation
stage was determined by the LevenvergMarquardt
Pv 2 = D1(CPD2 CPD1). (13)
Although in the second optimization model the
the second fermentation stage (see equation (13)), only
the optimal dilution rate in the first fermentation stage
was kept unchanged. Thus, the volumetric productivity
was calculated as a function of both the dilution rate on
the second stage and the feed glycerol concentration on
the first stage.
Finally, in the third optimization model the pro
ductivity of both fermentation stages was simulta
neously considered as a function of the dilution rate in
both stages and the feed concentration of glycerol in
the first fermentation stage. The objective function in
this model was the product of productivities between
both fermentation stages (Pv3):
Pv3 = Pv1Pv 2. (14)
1,3Propanediol Recovery and Purification.
Although different downstream process for 1,3pro
panediol recovery and purification from the fermenta
tion broth have been suggested (e.g., a series of filtration
stages followed by a train of a high vacuum distillation
columns [33] or a series of adsorption columns using
zeolites and coal [34]), one of the most promising
downstream was proposed by Posada et al. [35]. This
process was based on a reactiveextractive process
which changes the hydrophilic character of 1,3pro
panediol by mean of the reversible acetylation reaction
with isobutyl aldehyde (IBuAld) producing 2isopro
pyl1,3dioxane and water as shown in Fig. 1. 2Isopro
pyl1,3dioxane is a hydrophobic compound which can
No. 3 2013
242 POSADA et al.

H3C O H3C O CH2

H+
HO OH + HC CH2 + H2O
H3C H3C O CH2

Fig. 1. Acetylation reaction of 1,3propanediol with isobutyl aldehyde to 2isopropyl1,3dioxane.

Waste
DC1
etanol Water 1 F1 F2
Waste RII1 Water
Water E2
E1 Water 2
R1 C1 Glycerol M1
Dec1
Raw (98 wt %)
Glycerol
Glycerol
(85 wt %)

Aqueous
glycerol Absorbate Diluted
Solids Organic
Phase Glycerol Fermentation
Dec3 Broth
Dec2
IBuAld

Waste Re1 Fresh

Water IBuAld
Distillated6 Distillated5
Waste Organic
Cen1
DC5 RDC Distillated4 Water DC2 Phase
Distillated3 DC3
DC4
Water Distillated2 M2
Aqueous
Phase Solids
M2
P3
13PD Bottoms5 Bottoms4 Bottoms3 Bottoms2

Fig. 2. Simplified flowsheet for 1,3propanediol production from raw glycerol: Eevaporator, Rreactor, Ccentrifuge,
Decdecanter, DCdistillation column, Mmixer, Ffermentator, REreactor extractor, RDCreactive distillation
column.

be easily extractable by an organic phase, allowing an
effective and low energy cost separation. Reactive
extraction of 1,3propanediol from an aqueous media
was first studied by Malinowski [36] and then continued
by Liu et al. [37, 38], who tested different solvents (i.e.,
oxylene, ethylbenzene, and toluene) and reactants
(i.e., propionaldehyde, butyraldehyde, and isobutyl
aldehyde). As result of their analysis, it was found that
isobutyl aldehyde was the best alternative to extract
1,3propanediol because of its low solubility in water
(11 g/100 g), its middle boiling temperature (64.3C),
and additionally its acetalization product is the less sol
uble in water [39]. Besides of the reactive extraction
process, Malinowski [36] and Liu et al. [37, 38] recom
mended that 1,3propanediol can be recovered by a
hydrolysis reaction of 2alkyl1,3dioxane in a reactive
distillation column recovering also the used aldehyde.
The downstream process here proposed for 1,3pro
panediol recovery from the fermentation broth con
sisted on a centrifugation stage where the cell mass was
discarded. The clarified product was mixed with both
recycled and fresh isobutyl aldehyde and then fed to the
reactive extractor where the acetylation reaction took
place and two phases were obtained. The aqueous phase
containing still isobutyl aldehyde and 2isopropyl1,3
dioxane was distilled in two times, the first distillated
stream was the azeotrope isobutyl aldehydewater,
while the second distillated stream was the azeotrope 2
isopropyl1,3dioxanewater. The organic phase was
mixed with the azeotrope 2isopropyl1,3dioxane
water. This mixture was distilled and the azeotrope
isobutyl aldehydewater was again obtained. The bot
toms stream containing mainly 2isopropyl1,3diox
ane and water were subjected to a reactive distillation
process which required additional water. Thus, a mix
ture of 1,3propanediol and water was obtained by bot
toms. This stream was finally purified up to 93 wt % by
a distillation process. Also, all distilled stream contain
ing mainly isobutyl aldehyde and water were decanted.
The simplified flowsheet for 1,3propanediol produc
tion is shown in Fig. 2.
Simulation and Economic Evaluation Procedure.
The process design and simulation of the production
process of 1,3propanediol from raw glycerol was per
formed in two main stages. First, the fermentation pro
cess was analyzed considering a configuration of two

JOURNAL OF CTHEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

 DESIGN AND ECONOMIC ANALYSIS OF THE TECHNOLOGICAL SCHEME
continuous stages. Because of the nonlinear behavior
of the glycerol fermentation the occurrence of multi
plicity of steady states in the first fermentation stage was
studied. Then, the productivity in the first fermentation
stage was optimized, and additionally three different
objective functions were considered in order to opti
mize the second fermentation stage.
On the other hand, Aspen Plus (Aspen Technology,
Inc., USA) was used for defining, structuring, specify
ing, and simulating the complete technological scheme
for glycerol conversion to 1,3propanediol. In this
sense, information required for simulating the most
basic technological scheme, such as: physical and
chemical properties, parameters of design, and opera
tion of processing units, were mainly obtained from
secondary sources (e.g., articles, technical reports,
databases, patents, among others). Then, the most
complex and detailed technological scheme was
obtained by mean of rigorous simulations which
involved sensitivity analysis and search of optimal oper
ation conditions.
Because of both the petrochemical character of
Aspen Plus and its modularsequential approach, there
are not available kinetic models describing fermentative
reactions. Therefore, it was required to work with the
available interface between Aspen Plus and Excel.
Additionally, the study of complex fermentation kinet
ics describing inhibition phenomena requires generat
ing more complex calculation routines using other kind
of software such as MatLab.
Compounds such as free fatty acids, alkyl esters,
proteins, salts, cell mass, and the complex molecule
produced by reactiveextractive process are not avail
able on the Aspen Plus Database. Thus, these com
pounds were created by means of group contribution
methods or specified as nonconventional or solids
(e.g., biomass). Also, the molecular formula used for
K. pneumoniae was CH1.75O0.46N0.23.
The analysis of conventional separation methods in
the distillation process was carried out with the help of
the corresponding modules of the process simulators.
For this, both shortcut methods and rigorous models
available in the simulation package were employed. For
simulation of the different technologies involving the
operation of distillation, the shortcut method
DSTWU incorporated in the package Aspen Plus was
applied. This method uses the equations and correla
tions of WinnUnderwoodGilliland in order to pro
vide an initial estimation of the minimum number of
theoretic stages, minimum reflux ratio, location of the
feed stage, and components distribution. The rigorous
calculation of the operating conditions in the distilla
tion columns was performed using the module RadFrac
based on the equilibrium method that employs the
MESH equations (mass balance equations, phase equi
librium equations, summation of the compositions, and
heat balance equations) using the insideout algorithm.
Residue curve maps were used for the conceptual
design of the distillation schemes applying the princi
THEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47
243
ples of topological thermodynamics (static analysis)
[4042]. Sensitivity analyses were performed in order
to study the effect of the main operating variables
(reflux ratio, temperature of the feed stream, ratio
between the distillate and the feed, etc.) on the biodiesel
purity and the energy consumption of this operation.
The final result was the determination of operating con
ditions that allowed developing energy efficient pro
cesses. The objective of this procedure was to generate
the mass and energy balances from which the require
ments for raw materials, consumables, service fluids
and energy needs were calculated. Additionally, the
design of the reactive distillation column was performed
applying the methodology named static analysis which
allows determining both the operational viability and
the best configuration in the reactivedistillation tower
(localization of the reaction zone). This methodology is
the main tool for the qualitative study of the reactive
distillation process which requires minimum initial
information. Also it is based on both the thermody
namic topological behavior of reactive system and the
selection of stable state limits of highest conversion.
This methodology was developed by Serafimov et al.
[4042] and has been sufficiently illustrated by
Pisarenko et al. [43] and validated in multiple reactive
systems [44, 45]. Some considerations should be made
to carry out this analysis: (i) the reaction takes place
under equilibrium conditions and (ii) the reactive distil
lation column operates to total both reflux and effi
ciency. In other words (/) conditions are consid
ered. The main operation parameters are: the flow rate
ratio of distillate to bottoms (P/W) and the volume and
localization of the reaction zone.
The process was designed and analyzed using a cal
culation base of 1000 kg/h of raw glycerol fed to the
glycerol purification process. As the simulation results,
mass and energy balances were obtained for the techno
logical scheme. Thus, was possible to obtain require
ments of additional raw material, solvents, utility fluids,
and energy.
The capital and operating costs were calculated
using the software Aspen Icarus Process Evaluator
(Aspen Technologies, Inc., USA). This software esti
mates the capital costs of process units as well as the
operating costs, among other valuable data, utilizing the
design information provided by Aspen Plus and the data
introduced by the user for specific conditions such as
project location among others. Also, analyses were
based on the strategy designed by Cardona et al. [24
27] for process assessment.
This analysis was estimated in US dollars for a
10year period at an annual interest rate of 16%, con
sidering the straight line depreciation method and a
33% income tax [45]. The cost of raw glycerol was
0.0554 US$/L. The labor cost used for operatives and
supervisors was US$2.14/h and US$4.29/h, respec
tively. Also, the prices used for electricity, water and low
pressure vapor were US$0.03044/kWh, US$1.252/m3
and US$8.18/ton, respectively [45]. These values were
No. 3 2013
244 POSADA et al.

CPD, mol/L
1.0
X, g/L 1 (c)
8 2
(a) 3 0.5
6 4
5 0 1.5 2.0
4 0.3 0.25 0.20 0.15 0.10 0.5 1.0
0.05 0
2 C HAc, mol/L
0.20
0 2.0 0.15 (d)
0.3 0.2 1.5
0.10
0.1 0.5 1.0
0 0.05
CG, mol/L 0 1.5 2.0
0.3 0.25 0.20 0.15 0.10 0.5 1.0
1.0 0.05 0
(b) CEtOH, mol/L
0.4
0.5 0.3 (e)
0.2
0 0.1
0.3 0.2 1.5 2.0 0 1.5 2.0
Di, h1 0.1 0.5 1.0 CGIn , mol/L 0.3 0.25 0.20 0.15 0.5 1.0
0 0.10 0.05 0

Fig. 3. Hysteresis loops and multiple steady states. Concentrations of (a) biomass, (b) residual glycerol, (c) 1,3propanediol, (d)
acetic acid, and (e) ethanol: D = 0.04 (1), 0.1 (2), 0.16 (3), 0.22 (4), and 0.28 (5) h1. Vertical lines indicate the limits of the mul
tiple steady states region. Dotted lines show the wash out conditions for each dilution rate.

based on Colombian conditions. The purification costs
of crude glycerol were obtained in previous work [45] as
follows: 0.098 US$/L and 0.1124 US$/L for glycerol at
88 and 98 wt %, respectively. The abovementioned
software estimates the capital costs of process units as
well as the operating costs, among other valuable data,
utilizing the design information provided by Aspen Plus
and the data introduced by the user for specific condi
tions such as project location among others. Briefly, the
1,3propanediol production cost was determined using
the aforementioned approach.
RESULTS AND DISCUSSION
Since the raw glycerol purification process was pre
viously reported, some papers are recommended to the
readers for detailed information about this process [7, 8,
17]. Then, here a complete analysis of the fermentation
process in one and two continuous stages was devel
oped.
Multiplicity of Steady States. Figure 3 shows the
multiple steady states, hysteresis loops, and the wash
out line for the continuous glycerol fermentation. Mul
tiplicity of steady states and hysteresis loops were stud
ied considering the dilution rate as a parameter from a
low glycerol concentration up to the wash out condi
tions. In the hysteresis loops when the feed glycerol
concentration increases, low 1,3propanediol yields
were obtained. Moreover, when the feed glycerol con
centration decreases, high 1,3propanediol yields were
acquired. The latter condition corresponds to the upper
curves in the hysteresis loops. The wash out line indi
cates the extreme operational conditions where the
dilution rate equals the cellular growth rate (i = Di).
Optimization of the First Fermentation Stage. Volu
metric productivity was calculated as a function of both
the feed glycerol concentration and the dilution rate,
using a polygon mesh (partition) from low feed glycerol
concentration up to the wash out line (see Fig. 4a). The
conditions that generate the higher 1,3propanediol
concentration were selected at the multiplicity of steady
states region. As a consequence, a discontinuity
between A and B was obtained and the higher volumet
ric productivity values were located at the right side of
these points.
For each dilution rate exists a feed glycerol concen
tration that generates a maximum volumetric produc
tivity as obtained by the polygon mesh distribution
shown in Fig. 4a. In order to obtain the global optimum
for volumetric productivity both independent variables
(i.e., feed glycerol concentrations and dilution rate)
must be simultaneously considered. Since the highest
volumetric productivity was close to the steady states
region, this area must be considered when selecting the
initial estimated to apply in the optimization method.
In order to find the conditions for feed glycerol con
centration and the dilution rate that generates the high

JOURNAL OF CTHEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

 DESIGN AND ECONOMIC ANALYSIS OF THE TECHNOLOGICAL SCHEME 245

D, h1 (a) (b)
PV, molL1h1
0.7 0.7
0.10 1
0.6 0.6 2
0.09
3
0.5 0.08 0.5 4
0.07
0.4 0.06 0.4
0.05
0.3 0.3
A 0.04
0.2 0.03 0.2
0.02
0.1 0.01 0.1
B

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
In
C G , mol L
Fig. 4. (a) 1,3Propanediol volumetric productivity (the column in the right side gives the scale); the dashed line represents the
wash out line. (b) Wash out line (1), multiplicity of steady states (2), region of optimal productivity for each dilution rate (3), and
global optimal productivity (4).

est volumetric productivity, the LevenvergMarquardt dilution rate of 1.9850 h1, with a productivity of
optimization method was employed [31]. Since volu 0.0625 mol/(L h).
metric productivity is a noncontinuous function, the
initial estimated for both the feed glycerol concentra In the second optimization model the dilution rate
tion and the dilution rate must be higher than the con for the first stage was kept from the first model and the
ditions obtained in the A point. productivity in the second fermentation stage was opti
The multiple steady states region for glycerol fer mized as a function of the feed glycerol concentration.
mentation was reported by Xiu et al. [30], but the used The reached productivity was 0.1128 mol/(L h), at a
critical concentration parameters have a smaller fitting dilution rate on the second stage of 0.79 h1, with a feed
than the ones used in this work. The optimal conditions glycerol concentration on the first stage of
for glycerol fermentation in one continuous stage were 0.8817 mol/L. The outlet concentration of 1,3pro
as follows: 0.2821 h1 for the dilution rate and panediol from the first and second stages were 0.3405
0.6882 mol/L for the feed glycerol concentration, with and 0.4833 mol/L, respectively. Since each dilution rate
a volumetric productivity of 0.1076 mol/(L h). has its own optimal productivity, it was necessary to cal
Additionally, the obtained outlet concentration of culate the global optimal productivity using the Leven
1,3propanediol is 0.3811 mol/L. This optimal volu
metric productivity was outside the multiple steady
states region as shown in Figure 4b. Since this volumet PV, mol/(L h) CPD, mol/L
ric productivity was very close to the multiple steady 0.12 0.425
states region, minimum requirements in the automatic 1
control of the equipments are recommended. Addition 0.10 0.420
ally, Figure 4b shows the wash out conditions and the
optimal volumetric productivity for each dilution rate. 0.08 0.415
Optimization of the Second Fermentation Stage. The
kinetic model given for the fermentation system by 0.06 0.410
equations (1) to (12) is equally applicable for simulation
of a fermentation process with two continuous stages. 0.04 2 0.405
Thus, three models to optimize the second fermenta
tion stage were used. 0.02 0.400
In the first model, optimal conditions for the first
fermentation stage were used to calculate the volumet 0.395
ric productivity on the second fermentation stage, but 0 1.0 2.0 3.0 4.0 5.0
this function increases proportionally to the dilution D, h1
rate, contrary to the behavior shown by the concentra
tion of 1,3propanediol (see Fig. 5). The optimal con Fig. 5. 1,3Propanediol productivity (1) and concentra
centration of 1,3propanediol was 0.4126 mol/L at a tion (2) in the second fermentation stage.

THEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

246
D2, h1
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.6 0.7 0.8
Fig. 6. Volumetric productivity in the second fermentation stage using the optimal dilution rate obtained by model 1 for the first
fermentation stage. The P point indicates the global optimal productivity in the second fermentation stage. The column in the
right side gives the scale.
vergMarquardt method (see Fig. 6). Optimal produc
tivities for each dilution rate in the second fermentation
stage are represented by the discontinuous curve. Also,
the P point indicates the global optimal productivity in
the second fermentation stage.
Finally, in the third optimization model the produc
tivities of both fermentation stages were simultaneously
optimized considering the two dilution rates and the
feed glycerol concentration as independent variables.
The obtained results using the optimal dilution rate in
the first fermentation stage are shown in Fig. 7. The
optimum product of productivities for each dilution
rate in the second fermentation stage is represented by
the discontinuous curve.
The Q point indicates the global optimum for the
product of productivities. This point corresponds to a
dilution rate of 0.2821 and 3.08 h1 in the first and sec
ond stages respectively, and a feed glycerol concentra
tion of 0.7362 mol/L in the first stage. The reached
product of productivities was 0.0116 (mol/(L h))2
where the outlet concentration of 1,3propanediol was
0.4124 mol/L and the global yield was 0.5602 (mol
13PD)/(mol glycerol).
JOURNAL OF CTHEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47
POSADA et al.
0.11
0.10
0.09
0.08
0.07
0.06
0.05
0.04
0.03
P
0.02
0.01
0.9 1.0 1.1
In
C G, I , mol L
Table 1 shows the results of the three different mod
els used to optimize the fermentation of glycerol in two
stages. The highest global yield and volumetric produc
tivity in the first fermentation stage were generated
using the sequential optimization model. On the other
hand, the volumetric productivity in the second fer
mentation stage was optimized using the combined
optimization model under the optimal dilution rate
in the first fermentation stage. But, when the dilution
rate in the first fermentation stage was decreased at
0.25 h1, the best final concentration of 1,3pro
panediol was obtained as shown in Table 1.
Also, when the dilution rate in the first fermentation
stage was increased to 0.30 h1, the highest volumetric
productivity in the second stage was obtained as shown
in Table 1. Then, the volumetric productivity in the sec
ond fermentation stage can be optimized only at a spe
cific dilution rate in the first fermentation stage.
Using the simultaneous optimization model a high
volumetric productivity in both fermentation stages and
the highest product of productivities were obtained.
Also, the obtained value for the optimal dilution rate in
the first stage was the same using the sequential optimi
zation model.
No. 3 2013
 DESIGN AND ECONOMIC ANALYSIS OF THE TECHNOLOGICAL SCHEME 247

D2, h1 103
3.5

Q 10
3.0

2.5 8

2.0 6

1.5
4

1.0
2

0.5

0
0.6 0.7 0.8 0.9 1.0 1.1
In
C G, I , mol L

Fig. 7. Product of productivities of both fermentation stages using the optimal dilution rate obtained by model 1 for the first fer
mentation stage. The Q point indicates the global optimum for the product of productivities. The column in the right side gives
the scale.

Thus, the use of a sequential optimization model
allowed obtaining the highest global yield for 1,3pro
panediol (0.599) and the maximum volumetric produc
tivity in the first fermentation stage (0.1075 mol/(L h)),
whereas the highest 1,3propanediol outlet concentra
tion (0.512 mol/L) was observed when the combined
optimization model was employed. Meanwhile, using
the simultaneous optimization model showed both:
high volumetric productivities in the two fermentation
stages and the highest product of productivities
(0.01157 (mol/(L h))2). In this way, for the fermenta
tion of glycerol in two continuous stages, three different
operational configurations were available depending on
the desired process objective namely global yield,
1,3propanediol outlet concentration, or high simulta
neous productivity.
Downstream Process. The purification scheme here
proposed was based on integrated reactionseparation
units to carry out the recovery of 1,3propanediol from
the fermentation media. The first integrated stage was a

Table 1. Results summary for each optimization model

Optimization CPD1, CPD2, Pv1, Pv2, Pv3,
CS0, mol/L D1, h1 D2, h1 YPD/S
Model mol/L mol/L mol/(L h) mol/(L h) (mol/(L h))2
Sequential 0.688 0.282 1.985 0.381 0.413 0.599 0.107 0.063 6.72 103
Combined 0.882 0.282 0.790 0.341 0.483 0.548 0.096 0.113 1.09 102
Combined 0.932 0.250 0.940 0.396 0.512 0.549 0.099 0.109 1.08 102
Combined 0.852 0.300 0.720 0.311 0.469 0.551 0.093 0.114 1.06 102
Simultaneous 0.736 0.282 3.080 0.377 0.412 0.560 0.106 0.109 1.15 102

THEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

248 POSADA et al.

Table 2. Singular points for acetylation system of 1,3propanediol with 2isopropyl1,3dioxane (2IP13DO) at 1 atm

Component Type T, C X1 X2

Azeotrope: H2OIBuAld Heterogeneous 61.35 0.21 0.79

IBuAld Homogeneous 64.10 1
Azeotrope: H2O2IP13DO Heterogeneous 92.87 0.74 0.25
H2O Homogeneous 100.00 1
2IP13DO Homogeneous 138.12 1
1,3Propanediol Homogeneous 214.40 1

reactionextraction process where the hydrophilic hyde/1,3propanediol). The reactiveextraction pro

nature of 1,3propanediol was changed by the acetyla
tion reaction with isobutyl aldehyde to produce 2iso
propyl1,3dioxane. Then, due to 2isopropyl1,3
dioxane has a hydrophobic character, it was dragged
into the organic phase containing mainly isobutyl alde
hyde. This aldehyde acted as both reactant and solvent
for the reactive extraction process [3638]. Subse
quently, 1,3propanediol was recovered by reactive dis
tillation of 2isopropyl1,3dioxane with water
throughout the reverse reaction of acetylation. Thus,
1,3propanediol at high purity was obtained by bottom
while isobutyl aldehyde was recovered as distillated and
it was able to be reused in the downstream process.
Then, aldehyde was not consumed in the purification
process.
Otherwise, the three objective functions previously
analyzed for the glycerol fermentation process were
now considered as three different scenarios in order to
perform both the analysis of the downstream process for
each case and the economic comparison of each sce
nario.
In order to perform the reactive extractive process,
the clarified fermentation broth should be mixed with
isobutyl aldehyde in a weight ratio of 5/1 (isobutyl alde
Water
Feed
Pseudo initil Separatriz
mixture Material balance
Bottoms lines
Residue Distillate
curve
2isopropyl
1.3dioxane
Isobutyl
1.3propanediol aldehyde
Chemical interaction
surface
Fig. 8. Direct separation with a feeding ratio of 2isopro
pyl1,3dioxane to water of 0.38/0.62.
cess took place at 10C since better distribution coeffi
cients were obtained at lower temperatures. During this
downstream process, isobutyl aldehyde recovered by
decantation was obtained at 97.3 wt %.
On the other hand, to carry out the static analysis of
the hydrolysis system of 2isopropyl1,3dioxane with
water to produce 1,3propanediol, the singular points
of this mixture were characterized as shown in Table 2.
Thus, only one distillation region was obtained with a
bunch of residue curves starting from the azeotrope of
minimum boiling point (isobutyl aldehydewater) and
ending in the 1,3propanediol pure. By direct separa
tion (formulated distilled) three distillation subregions
were obtained while in the case of indirect separation
(bottoms formulated) two distillation subregiones were
founded. Thus, ten different possibilities for the feeding
ratio of 2isopropyl1,3dioxane to water at the reac
tive distillation column were analyzed (results are not
shown). Then, the feeding ratio that generates the high
est products distribution (P/W) at a total conversion of
2isopropyl1,3dioxane (see Fig. 8) corresponded to
the 2isopropyl1,3dioxane/water molar mixture of
0.3776/0.6224, as it is shown in Fig. 9. Also, based on
both the chemical equilibrium and the residue curve
maps, it was determined that the reactive zone must be
localized in the stripping section of the reactive distilla
tion column. Also, the product obtained in the distil
lated stream is the heterogeneous azeotrope isobutyl
aldehydewater.
Then, in order to verify the predicted behavior, it
was possible to obtain the actual conditions and the tra
jectory predicted by the static analysis method. Thus,
simulations at both / (stages/reflux) and finite con
dition were carried out. Also, it was found that a self
extractive phenomenon (which can be understood as a
nonlineal variation in the relative volatility of a system
with the change of concentrations in a multicomponent
mixture (see Fig. 10) affects strongly the reactive distil
lation column performance. The analysis of the isovol
atility curves allowed determining that a redistribution
of the feeding streams lead to a high conversion keeping
the configuration obtained by the static analysis
method.

JOURNAL OF CTHEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

 DESIGN AND ECONOMIC ANALYSIS OF THE TECHNOLOGICAL SCHEME 249

In this case, the water required as reactive to carry P/W

out the hydrolysis reaction of 2isopropyl1,3dioxane
was fed in five different stages of the reactive distillation 1.6
column. For instance, in scenario 1 the reactive distilla
tion column had 45 stages and the water stream was fed 1.2
in the stages: 9, 15, 22, 29, 36, and 43. And the respec
tive mass flow rates were 7.56, 11.46, 15.36, 19.27, and
23.17 kg/h. 0.8
In this way high conversion levels of 2isopropyl
1,3dioxane were obtained. But, because of the 0.4
1,3propanediol purity ranges 83.3 and 93.9 wt %, a
final distillation process was required in order to achieve
a higher purity of 1,3propanediol. A summary of the 0 0.2 0.4 0.6 0.8 1.0
main simulation results for each scenario is given in
Table 3.
Fig. 9. P/W ratio, direct separation with a feeding ratio of
The final production of 1,3propanediol was mainly 2isopropyl1,3dioxane to water of 0.38/0.62.
related to the fermentation yield of both fermentation
stages. Thus, while the decreasing order for the final
concentration of 1,3propanediol after the second fer Isobutyl aldehyde
mentation stage was scenario 2 > scenario 1 > scenario
3 and the decreasing order for the fermentative yield of
glycerol to 1,3propanediol was scenario 1 > scenario
3 > scenario 2, the decreasing order for the actual pro
duction of 1,3propanediol was scenario 1 > scenario
3 > scenario 2 (see Table 3).
Otherwise, high recovery percentages were achieved
for isobutyl aldehyde, indicating that low requirements
of fresh reactive were required. Higher differences were
noticed when the whole technological scheme was ana
lyzed. The maximum global molar yield from glycerol 2isopropyl Water
to 1,3propanediol was obtained for scenario 1, while 1.3dioxane
the minimum was obtained for scenario 2. The relative
difference between these two scenarios was 5.21%, Fig. 10. Isovolatility curve (waterisobutyl aldehyde2
isopropyl1,3dioxane).
which was close to the relative difference for the fer
mentation yield obtained for the same both scenarios,
11.14%. Thus, it can be stated that the technological In order to obtain more detailed information about
performance of 1,3propanediol production from raw the 1,3propanediol production from raw glycerol, the
glycerol depended mostly on the global conversion of total production cost was divided in two processing sec
substrate to the main product during the fermentation
stage (see Table 4). tions: (i) glycerol purification plus glycerol fermenta
tion and (ii) 1,3propanediol recovery and purification,
ECONOMIC ASSESSMENT as shown in Table 5. The first section (i.e., glycerol puri
Although, for most industrial processes the cost of fication and fermentation) represented in all cases
raw material represents near to 50% of the total produc between 29.4 to 32.4%, while the second section (i.e.,
tion costs, if raw glycerol is used for the 1,3propanediol 1,3propanediol purification) was between 67.6 to
production by means of K. pneumoniae strains this value 70.6%. The 1,3propanediol purification cost was esti
is lower i.e., between 8.2 to 9.2% of the total production mated between 0.652 to 0.758 US$/kg of 1,3pro
cost. The values here obtained did not consider the panediol. Also, the commercial sale price for 1,3pro
transportation costs, since the 1,3propanediol produc panediol was compared to the production cost and it
tion process was assumed to be a biorefinery adjacent to
was found that this ratio (i.e., sale price/production
the biodiesel production process. Both utilities and cap
ital costs represented the highest production cost which cost) was higher than the unity for all scenarios, and for
combined vary between 64.46% for scenario 3 to scenario 3 this ratio was 1.83. These results indicate that
66.99% for scenario 1. Also, it was noticed that capital 1,3propanediol production from raw glycerol by
costs increased when final concentration of 1,3pro means of K. pneumoniae could be a profitable alterna
panediol increased. tive for glycerol usage.

THEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

250 POSADA et al.

Table 3. Summary of the main simulation results

Parameters Dilgly 13PD2 Organic2 Aqueous1 IP13DO3 13PD3 13PD4

Scenario1
T, K 310 310 283.2 283.1 404.2 457.9 486.8
M, kg/h 9115.8 8900.1 920.0 9270.3 453. 8 272.6 252.7
Water 8536.2 8536.2 14.28 8582.9 0 0.752 0
Glycerol 578.9 6.45 0.273 6.18 0.273 0.273 0.273
KPNEUMON 0 27.16 0 0 0 0 0
1,3P01 0 273.67 1.633 10.95 1.63 251.1 250.83
Acetic acid 0 62.95 0.387 59.07 1.08 1.08 0.980
Ethanol 0 20.15 1.335 17.48 3.74 0.003 0
IBuAld 0 0 491.32 557.88 0.449 0.183 0
2IP13DO 0 0 410.77 35.83 446.60 19.24 0.656
Scenario2
T, K 310 310 283.2 283.1 409.1 444.4 487.7
M, kg/h 6766.1 6594.4 925.9 6847.5 423.5 247.9 245.0
Water 6186.5 6186.5 15.51 6227.5 0 2.74 0.007
Glycerol 578.9 89.0 5.95 83.04 5.95 5.95 5.95
KPNEUMON 0 20.26 0 0 0 0 0
1,3P01 0 250.30 1.70 7.54 1.703 238.4 238.4
Acetic acid 0 57.58 0.473 52.84 1.021 0.747 0.686
Ethanol 0 10.37 0.560 8.50 1.001 0 0
IBuAld 0 0 515.1 442.4 1.515 1 x 103 0
2IP13DO 0 0 386.6 25.73 412.3 0.017 0.002
Scenario3
T, K 310 310 283.2 283.1 411.4 424.7 487
M, kg/h 8531.1 8338.9 872.37 8687.6 427.6 253.6 248.4
Water 7951.5 7951.5 13.77 7995.5 0 5.15 0.027
Glycerol 578.9 53.03 2.277 50.7 2.28 2.28 2.28
KPNEUMON 0 23.73 0 0 0 0 0
1,3P01 0 259.07 1.535 10.20 1.53 245.6 245.6
Acetic acid 0 60.30 0.374 56.55 0.7 0.59 0.507
Ethanol 0 14.28 0.573 12.37 0 0 0
IBuAld 0 0 465.07 527.9 0 0.001 0
2IP13DO 0 0 388.77 34.32 423.1 0.011 0

JOURNAL OF CTHEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

 DESIGN AND ECONOMIC ANALYSIS OF THE TECHNOLOGICAL SCHEME 251

Table 4. Data representing the behavior of the downstream process

Data Scenario 1 Scenario 2 Scenario 3
Reactive extraction
Extraction efficiency, % 96.00 96.99 96.06
Distribution coefficient 204.09 203.39 204.91
Loading (Z) 0.177 0.182 0.177
Downstream process
Global 1,3propanediol recovery, % 91.66 95.26 94.80
Isobutyl aldehyde recovery, % 98.66 99.48 99.54
Global process
Global process yield from glycerol to lactic acid, % 0.5244 0.4984 0.5134

Table 5. Economic results for raw glycerol conversion to CONCLUSIONS

1,3propanediol: cost (US$/kg) and share (%)
In the past, important efforts have been made in
Characteristics Scenario 1 Scenario 2 Scenario 3 order to introduce the biotechnological production of
Raw materials: 1,3propanediol from glycerol to the industry. How
cost 0.0889 0.0892 0.0887 ever, research tendencies were focused on microorgan
share 8.61 8.19 9.20
ism development and on the analysis of some specific
process conditions. On the other hand, the drastic
Utilities:
increment in the use of biodiesel has caused an oversup
cost 0.2151 0.2245 0.1458 ply of glycerol in the market. Thus, in order to suitably
share 20.83 20.61 15.12 use the glycerol obtained from the biodiesel industry,
Operating labor: the mass production of 1,3propanediol from glycerol
cost 0.0698 0.0746 0.0714 needs additional analysis. Here, a complete technolog
share 6.76 6.85 7.40 ical scheme for its production has been not only pro
Maintenance: posed but also assessed. Even more, the fermentation
cost 0.0479 0.0503 0.0486 stage was optimized by three different ways and as
share 4.64 4.62 5.04 result it was possible to assess economically three
Operating charges: scenarios where the third one had the lowest produc
cost 0.0172 0.0187 0.0178 tion cost. The relative differences respect to the first
share 1.66 1.71 1.85 and the second scenarios were 7.2 and 13.0%. The
Plant overhead: obtained results provide enough information to
understand the different possibilities for process
cost 0.0577 0.0625 0.0600
intensification using this technology and also to
share 5.59 5.73 6.22 compare it with other new industrial alternatives for
Other costs: the utilization of glycerol as a raw material.
cost 0.0650 0.0712 0.0615
share 6.29 6.54 6.38
Depreciation of capital: ACKNOWLEDGMENTS
cost 0.4767 0.5042 0.4757 The authors express their acknowledgments to the
share 46.16 46.28 49.34 National University of Colombia at Manizales and the
Coproducts sales: Russian Foundation for Basic Research (project
cost 0.0055 0.0057 0.0054 no. 13030222a) for funding this research.
share 0.53 0.52 0.56
Glycerol purification and NOTATION
fermentation:
cost 0.304 0.331 0.313 b1, b2parameters used in equation (11);
share 29.40 30.38 32.42 CS0feed glycerol concentration in the first stage,
Lactic acid purification: mol/L;
cost 0.729 0.758 0.652 CPD11,3propanediol concentration in the first
share 70.60 69.62 67.58
fermentation stage, mol/L;
Total cost 1.033 1.089 0.964 CPD21,3propanediol concentration in the sec
Sale price/production 1.71 1.62 1.83
ond fermentation stage, mol/L;
In
cost CEtOH inlet ethanol concentration, mol/L;

THEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47 No. 3 2013

252 POSADA et al.
Out
CEtOH outlet ethanol concentration, mol/L;
* critical ethanol concentration, mol/L;
CEtOH
CGIn inlet glycerol concentration, mol/L;
C GOut outlet glycerol concentration, mol/L;
C *critical glycerol concentration, mol/L;
G
In
inlet acetic acid concentration, mol/L;
C HAc
Out
C HAcoutlet acetic acid concentration, mol/L;
* critical acetic acid concentration, mol/L;
C HAc
In
CPD inlet 1,3propanediol concentration,
mol/L;
Out
C PD outlet 1,3propanediol concentration,
mol/L;
* critical 1,3propanediol concentration,
CPD
mol/L;
c1, c2parameters used in equation (11);
Ddilution rate, h1;
D1dilution rate in the first fermentation stage,
h1;
D2 dilution rate in the second fermentation stage,
h1;
KSMonod saturation constant, mol/L;
* parameter used in equation (9);
K HAc
* parameter used in equation (8);
K PD
K S* parameter used in equation (7);
Mmass flow rate, kg/h;
mGparameter used in equation (7);
mHAc parameter used in equation (9);
mPDparameter used in equation (8);
Pdistillate flow rate, mol/s;
Pv volumetric productivity, mol/(L h);
Pv1volumetric productivity in the first fermenta
tion stage, mol/(L h);
Pv2volumetric productivity in the second fer
mentation stage, mol/(L h);
Pv3product of volumetric productivities in the
two fermentation stages, (mol/(L h))2;
qEtOHspecific rate of ethanol generation, h1;
qGspecific rate of glycerol consumption, h1;
qHAcspecific rate of acetic acid generation, h1;
qPDspecific rate of 1,3propanediol generation,
h1;
qGmparameter used in equation (7);
qHAc
m
parameter used in equation (9);
qPDparameter used in equation (8);
m
Ttemperature, C;
Wbottoms flow rate, mol/s;
X In nlet biomass concentration, g/L;
X Out outlet biomass concentration, g/L;
JOURNAL OF CTHEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47
YPD/Sglobal fermentation yield;
m
Y(EtOH G)parameter used in equation (10);
YGm parameter used in equation (7);
m
YHAc parameter used in equation (9);
m
YPD parameter used in equation (8);
conversion;
specific rate of cellular growth, h1;
maxmaximum specific growth rate, h1.
REFERENCES
1. Posada, J.A., Orrego, C.E., and Cardona, C.A., Biodie
sel production: biotechnological approach, Int. Rev.
Chem. Eng., 2009, vol. 1, p. 571.
2. Zheng, Y., Chen, X., and Shen, Y., Commodity chemi
cals derived from glycerol, an important biorefinery
feedstock, Chem. Rev., 2008, vol. 108, p. 5253.
3. Silva, G.P., Mack, M., and Contiero, J., Glycerol: a
promising and abundant carbon source for industrial
microbiology, Biotechnol. Adv., 2009, vol. 27, p. 30.
4. Pagliaro, M. and Rossi, M., The future of Glycerol: New
Usages for a Versatile Raw Material, Cambridge: RSC,
2008.
5. Morita, T., Konishi, M., Fukuoka, T., Imura, T., and
Kitamoto, D., Microbial conversion of glycerol into gly
colipid biosurfactants, mannosylerythritol lipids, by a
basidiomycete yeast, Pseudozyma Antarctica JCM
10317T, J. Biosci. Bioeng., 2007, vol. 104, p. 78.
6. Posada, J.A., Cardona, C.A., Cetina, D.M., and Or
rego, C.E., Bioglycerol as Raw Material to Obtain Added
Value Products, Researching Advances for Biofuels Pro
duction, Cardona, C.A., Ed., Manizales: Universidad
Nacional de Colombia, 2009, p. 103.
7. Posada, J.A. and Cardona, C.A., Design and analysis of
fuel ethanol production from raw glycerol, Energy, 2010,
vol. 35, p. 5286.
8. Posada, J.A., Naranjo, J.M., Lpez, J.A., Higuita, J.C.,
and Cardona, C.A., Design and analysis of poly3
hydroxybytyrate production processes from crude glyc
erol, Process Biochem., 2000, vol. 46, p. 310.
9. Saxena, R.K., Anand, P., Saran, S., and Isar, J., Micro
bial production of 1,3propanediol: recent developments
and emerging opportunities, Biotechnol. Adv., 2009,
vol. 27, p. 895.
10. Ma, C., Zhang, L., Dai, J., and Xiu, Z., Relaxing the
coenzyme specificity of 1,3 propanediol oxidoreductase
from Klebsiella pneumonia by rational design, J. Biotech
nol., 2010, vol. 146, p. 173.
11. Menzel, K., Zeng, A.P., and Deckwer, W.D., Enzy
matic evidence for an involvement of pyruvate dehydro
genase in the anaerobic glycerol metabolism of Klebsiella
pneumonia, J. Biotechnol., 1997, vol. 56, p. 135.
12. Barbirato, F., Himmi, E.H., Conte, T., and Bories, A.,
1,3Propanediol production by fermentation: an inter
esting way to valorize glycerin from the ester and ethanol
industries, Ind. Crops Prod., 1998, vol. 7, p. 281.
13. Zheng, P., Sun, J., van den Heuvel, J., and Zeng A.P.,
Discovery and investigation of a new, second triose
No. 3 2013
 DESIGN AND ECONOMIC ANALYSIS OF THE TECHNOLOGICAL SCHEME
phosphate isomerase in Klebsiella pneumonia, J. Biotech
nol., 2006, vol. 125, p. 462.
14. Deckwer, W.D., Microbial conversion of glycerol to
1,3propanediol, FEMS Microbiol. Rev., 1995, vol. 16,
p. 143.
15. Menzel, K., Zeng, A.P., Biebl, H., and Deckwer, W.D.,
Kinetic, dynamic and pathway studies of glycerol metab
olism by Klebsiella pneumonia in anaerobic continuous
culture. Part I: The phenomena and characterization of
oscillation and hysteresis, Biotechnol. Bioeng., 1996, vol.
52, p. 549.
16. Zeng, A.P., Menzel, K., and Deckwer, W.D., Kinetic,
dynamic and pathway studies of glycerol metabolism by
Klebsiella pneumonia in anaerobic continuous culture.
Part II: Analysis of metabolic rates and pathways under
oscillation and steadystate conditions, Biotechnol.
Bioeng., 1996, vol. 52, p. 561.
17. Posada, J.A. and Cardona, C.A., Validation of glycerin
refining obtained as a byproduct of biodiesel produc
tion, Ing. Univ., 2010, vol. 14, p. 2.
18. Dasari, M.A., Kiatsimkul, P.P., Sutterlin, W.R., and
Suppes, G.J., Lowpressure hydrogenolysis of glycerol
to propylene glycol, Appl. Catal., A., 2005, vol. 281,
p. 225.
19. Barbirato, F., Soucaille, P., and Bories, A., Physiologic
mechanisms involved in accumulation of 3hydroxy
propionaldehyde during fermentation of glycerol by
Enterobacter agglomerans, Appl. Environ. Microbiol.,
1996, vol. 62, p. 4405.
20. Papanikolaou, S., RuizSanchez, P., Pariset, B., Blan
chard, F., and Fick, M., High production of 1,3pro
panediol from industrial glycerol by a newly isolated
Clostridium butyricum strain, J. Biotechnol., 2000, vol. 77,
p. 191.
21. GonzalezPajuelo, M., MeynialSalles, I., Mendes, F.,
Andrade, J.C., Vasconcelos, I., and Soucaille, P., Meta
bolic engineering of Clostridium acetobutylicum for the
industrial production of 1,3propanediol from glycerol,
Metab. Eng., 2005, vol. 7, p. 329.
22. Ito, T., Nakashimada, Y., Senba, K., Matsui, T., and
Nishio, N., Hydrogen and ethanol production from
glycerolcontaining wastes discharged after biodiesel
manufacturing process, J. Biosci. Bioeng., 2005, vol. 100,
p. 260.
23. Zhao, Y.N., Chen, G., and Yao, S.J., Microbial produc
tion of 1,3propanediol from glycerol by encapsulated
Klebsiella pneumonia, Biochem. Eng. J., 2006, vol. 32,
p. 93.
24. Cardona, C.A. and Sanchez, O.J., Energy consumption
analysis of integrated flowsheets for production of fuel
ethanol from lignocellulosic biomass, Energy, 2006, vol.
31, p. 2447.
25. Quintero, J.A., Montoya, M.I., Sanchez, O.J., Giral
do, O.H., and Cardona, C.A., Fuel ethanol production
from sugarcane and corn: comparative analysis for a
Colombian case, Energy, 2008, vol. 33, p. 385.
26. Cardona, C.A., Quintero, J.A., and Paz, I.C., Produc
tion of bioethanol from sugarcane bagasse: status and
perspectives, Bioresour. Technol., 2009, vol. 101, p. 4754.
27. Gutierrez, L.F., Sanchez, O.J., and Cardona, C.A., Pro
cess integration possibilities for biodiesel production
from palm oil using ethanol obtained from lignocellu
losic residues of oil palm industry, Bioresour. Technol.,
2009, vol. 100, p. 1227.
THEORETICAL FOUNDATIONS OF CHEMICAL ENGINEERING Vol. 47
253
28. Menzel, K., Zeng, A.P., and Deckwer, W.D., High
concentration and productivity of 1,3propanediol from
continuous fermentation of glycerol by Klebsiella pneu
moniae, Enzyme Microb. Technol., 1997, vol. 20, p. 82.
29. Xiu, Z.L., Song, B.H., Sun, L.H., and Zeng, P., Theo
retical analysis of effects of metabolic overflow and time
delay on the performance and dynamic behavior of a
twostage fermentation process, Biochem. Eng. J., 2002,
vol. 11, p. 101.
30. Xiu, Z.L., Song, B.H., Wang, Z.T., Sun, L.H.,
Feng, E.M., and Zeng, A.P., Optimization of dissimila
tion of glycerol to 1,3propanediol by Klebsiella pneumo
niae in one and twostage anaerobic cultures, Biochem.
Eng. J., 2004, vol. 19, p. 189.
31. Edgar, T.F., Himmelblau, D.M., and Lasdon, L.S.,
Optimization of Chemical Processes, New York:
McGrawHill, 2001, 2nd ed.
32. Raval, K.N., Hellwig, S., Prakash, G., RamosPlasen
cia, A., Srivastava, A., and Biichs, J., Necessity of a two
stage process for the production of azadirachtinrelated
limonoids in suspension cultures of Azadirachta indica,
J. Biosci. Bioeng., 2003, vol. 96, p. 16.
33. Adkesson et al., US Patent 200510069997 A1, 2005.
34. Corbin, D.R. and Norton, T., US Patent 603 048, 2003.
35. Cardona, C.A., Posada, J.A., and Montoya, M.I., Use of
glycerol from biodiesel production: conversion to added
value products, Proc. 6th Eur. Congr. on Chemical Engi
neering, Copenhagen, 2007, p. 35.
36. Malinowski, J.J., Reactive extraction for downstream
separation of 1,3propanediol, Biotechnol. Prog., 2000,
vol. 16, p. 76.
37. Liu, D., Hao, J., and Liu, H., Novel route of reactive
extraction to recover 1,3propanediol from a dilute
aqueous solution, Ind. Eng. Chem. Res., 2005, vol. 44,
p. 4380.
38. Liu, D., Hao, J., Xu, F., and Liu, H., Downstream pro
cessing of 1,3propanediol Fermentation broth, J. Chem.
Technol. Biotechnol., 2006, vol. 81, p. 102.
39. Samant, K. and Ng, K.M., Synthesis of extractive reac
tion processes, AIChE J., 1998, vol. 44, p. 1363.
40. Serafimov, L.A., Zharov, V.T., and Timofeyev, V.S.,
Rectification of multicomponent mixtures. I. Topologi
cal analysis of liquid vapor phase equilibrium diagrams,
Acta Chim. Acad. Sci. Hung., 1971, vol. 69, no. 4, p. 383.
41. Pisarenko, Yu.A., Serafimov, L.A., and Cardona, K.A.,
Optimization of reactive distillation processes, Theor.
Found. Chem. Eng., 1999, vol. 33, p. 455.
42. Serafimov, L.A., Pisarenko, Yu.A., and Kulov, N.N.,
Coupling chemical reaction with distillation: thermody
namic analysis and practical applications, Chem. Eng.
Sci., 1999, vol. 54, p. 1383.
43. Pisarenko, Yu.A., Serafimov, L.A., Cardona, C.A.,
Efremov, D.L., and Shuwalov, A.S., Reactive distilla
tion design: analysis of the process statics, Rev. Chem.
Eng., 2001, vol. 17, no. 4, p. 253.
44. Giessler, S., Danilov, R.Y., Pisarenko, R.Y., Serafi
mov, L.A., Hasebe, S., and Hashimoto, I., Feasible sep
aration modes for various reactive distillation systems,
Ind. Eng. Chem. Res., 1999, vol. 38, p. 4060.
45. Posada, J.A., Cardona, C.A., and Rinc, L.E., Sustain
able biodiesel production from palm using in situ pro
duced glycerol and biomass for raw bioethanol, Proc.
32nd Symp. on Biotechnology for Fuels and Chemicals,
Clearwater Beach, Fla., 2010, p. 172.
No. 3 2013