Documente Academic
Documente Profesional
Documente Cultură
pubs.acs.org/EF
Laboratory of Optimization, Design and Advanced Control (LOPCA), School of Chemical Engineering,
University of Campinas (UNICAMP), Campinas, SP, Brazil
ABSTRACT: Acetonebutanolethanol (ABE) facilities have traditionally presented unattractive economics because of the large
energy consumption during recovery of the products from a dilute fermentation broth (13 g/L butanol). This problem results
from the high toxicity of butanol to microorganisms that catalyze its production. Flash fermentation is a continuous fermentation
system with integrated product recovery. The bioreactor is operated at atmospheric pressure and the broth is circulated in a closed
loop to a vacuum chamber where ABE is continuously boiled o at 37 C and condensed afterward. With this technology the beer
achieved a concentration of butanol as high as 3037 g/L. This paper studies the energy requirements for butanol recovery using the
ash fermentation technology and its eect on the energy consumption by the downstream distillation system. Compressors are
used to remove the vapors from the ash tank, thus maintaining the desired vacuum. The heat recovery technique of vapor
recompression is used to reduce energy requirements. With this technique the heat generated by the compression and partial
condensation of the vapors provides the energy for boil up (heat of vaporization) in the ash tank. Thus the energy requirement for
the ash fermentation is essentially the electrical power demanded by compressors. Energy for recirculation pumps accounts for
approximately 0.5% of the total energy consumption. Small increments in butanol concentration in the beer can have important
positive impacts on the energy consumption of the distillation unit. Nonetheless, the energy use of the recovery technology must be
included in the energy balance. For a fermentation with a wild-type strain, the total energy requirement for butanol recovery (ash
fermentation distillation) was 17.0 MJ/kg butanol, with 36% of this value demanded by the ash fermentation. This represents a
reduction of 39% in the energy for butanol recovery in relation to the conventional batch process. In the case of a fermentation with a
hyper-butanol producing mutant strain, the use of the ash fermentation could reduce the energy consumption for butanol recovery
by 16.8% in relation to a batch fermentation with the same mutant strain.
1. INTRODUCTION
Butanol is highly toxic to microorganisms that catalyze its
production, and for this reason less than 13 g/L of butanol is
produced during batch fermentation. For the sake of comparison,
in the ethanol fermentation the yeast cells tolerance to ethanol is
approximately ten times greater. Therefore, typical acetone
butanolethanol (ABE) fermentation has been plagued by the
use of dilute sugar solutions as substrates, large process volumes,
high downstream process costs due to intensive energy requirements for recovery of low concentrations of ABE in the beer, and
large quantities of wastewater.
A solution to these problems can be addressed by using genetic engineering techniques to develop strains that could
tolerate higher concentrations of butanol and sugar.1,2 Another
option is the use of technologies designed to remove butanol
continuously from the fermentation broth (integrated recovery technologies). The product recovery reduces the eect
of product inhibition and allows an increase in the substrate
concentration which results in a reduction in process streams,
and higher productivity. With either approach (biological and
r 2011 American Chemical Society
ARTICLE
2. METHODOLOGY
The principles of the ash fermentation process are presented
in the scheme shown in Figure 1. The fermentation broth of a
continuous fermentor circulates through a vacuum ash tank.
The partial vaporization of water and fermentation products that
takes place in the ash tank generates a vapor phase rich in
solvents (ABE) and an ABE-depleted liquid stream. The vapor is
subsequently condensed and combined with the beer stream and
then sent to distillation. The ABE-depleted liquid stream returns
to the fermentor, which is operated at atmospheric pressure.
Temperature in the fermentor and in the ash tank is set to
37 C. Vacuum is regulated in order to keep constant the amount
of broth vaporized.
ARTICLE
Figure 2. Vapor recompression heating system. Heat integration between the recompressed ABE-enriched vapor stream from the ash tank and the
reboiler of the acetone column, and fermentation beer stream (stream conditions refers to scenario 2, Table 1).
ARTICLE
Figure 3. Downstream distillation unit (stream conditions refers to scenario 2, Table 1).
dS
F PU
Fr
F0
rS
S Sr S0
dt
V
V
V
dPi
F PU
Fr
r Pi
Pi Pr;i
dt
V
V
ARTICLE
III
IV
VI
S0 (g/L)
100
150
150
170
170
150
500
500
750
500
750
750
strain
a
II
substrate and produced solvents and biomass previously determined from eqs 13 were obtained in the Aspen Plus simulations. This procedure enabled the simplied reactor model
used in the Aspen simulator (eqs 49) to reproduce the steady
state concentrations values generated by a mathematical model
that incorporates a sophisticated experimental kinetic model
(eqs 13), including a nonlinear product inhibition model.
The input data for the ash calculation in Aspen Plus were
temperature (37 C) and vapor fraction. The latter was obtained
from the Fortran calculations.
Properties for biomass (CH1.8O0.5N0.2) were obtained from
the NREL database.17 The energy demanded by the equipment
(MJ/h) was divided by the mass ow rate (kg/h) of butanol at
the bottom of the butanol column to get the specic energy
requirement (MJ/kg butanol) for butanol purication.
Design specications and operating conditions used in the
simulation of the distillation unit were based on the optimum
congurations determined by Luyben18 and van der Merwe.19
The nal setup used here aimed the conciliation between capital
(minimum number of trays) and energy (minimum reux ratio)
costs. For this, design specications were fed to the simulator in
terms of recovery or purity for a specic compound in either the
top or bottom sections. The design spec/vary feature of Aspen
2351
ARTICLE
ARTICLE
Figure 6. Eects of pressure in the ash tank (Pash) on fermentation parameters determined in a simulated fermentation with the following operating
conditions: F0 = 50 m3/h, S0 = 150 g/L, V = 500 m3, Fc = 300 m3/h.
Table 2. Performance Parameters of the Flash Fermentation Considering the Scenarios Presented in Table 1
scenarios
I
II
III
IV
VI
6.49
6.47
6.53
6.47
6.50
6.45
8.20
18.5
13.9
24.1
20.9
26.7
1.8
0.18
3.0
0.20
2.1
0.21
3.6
0.22
2.5
0.22
3.2
0.33
0.29
0.34
0.35
0.36
0.37
0.43
5.6
6.9
7.8
7.9
8.3
10.3
8.20
10.5
11.9
12.4
13.0
12.8
81.8
69.4
91.6
67.0
77.3
82.1
75.3
84.9
81.2
88.0
86.2
90.0
17.9
30.0
31.0
36.3
37.1
48.6
6.60
4.1
12.6
5.5
10.6
4.1
16.6
5.5
14.4
4.6
14.9
4.5
4.8
6.0
4.4
6.5
5.5
5.4
17.5
10.9
11.0
9.5
9.3
8.0
22.3
17.0
15.4
16.0
14.8
13.4
54.3
32.3
31.3
26.7
26.2
21.0
ARTICLE
4. CONCLUSIONS
With the ash fermentation technology, high conversion of
concentrated sugar solution into ABE can be obtained, resulting
in high productivity and a more concentrated fermentation beer.
For dierent fermentation conditions (substrate concentration,
dilution rate, and microbial strain), two operating variables of the
ash tank, inlet ow rate and pressure, must be regulated in order
to ensure the desired sugar conversion. These two variables regulate the amount of fermentation broth vaporized, minimization of which is crucial to enhance the energy eciency of the
ash fermentation.
Small increments in butanol concentration in the beer can
have important positive impacts on the energy consumption of
the distillation unit. Nonetheless, the energy use of the recovery
technology must be included in the energy balance. For a fermentation with a wild-type strain, the total energy requirement for butanol recovery (ash fermentation distillation) was
17.0 MJ/kg butanol, with 36% of this value demanded by the ash
fermentation. This represents a reduction of 39% in the energy
for butanol recovery in relation to the conventional batch
process. In the case of a fermentation with a hyper-butanol
2354
AUTHOR INFORMATION
Corresponding Author
ACKNOWLEDGMENT
We thank Fundac-~ao de Amparo a Pesquisa do Estado de S~ao
Paulo (FAPESP) (Contract grant 2007/00341-1) for the nancial support.
NOMENCLATURE
A = Heat-transfer area, m2
Aij = Parameter for GE (molar Gibbs energy) for the UNIQUAC
activity coecient model (cal/mol)
AA = Acetic acid
ABE = Acetonebutanolethanol
BA = Butyric acid
F0 = Fresh broth ow rate (continuous feed), m3/h
Fc = Inlet ow rate of the ash tank, m3/h
Fpu = Fermentor beer ow rate (beer), m3/h
Fr = Flash tank liquid outlet ow rate (liquid stream depleted in
ABE), m3/h
NREL = National Renewable Energy Laboratory (USA)
Pash = Flash tank pressure, kPa
Pi = Fermentor product concentration, g/L
Pr = Product concentration in the ash tank liquid outlet ow, g/L
q = Molecular area parameter for the UNIQUAC activity
coecient model, Q = Heat, MW
r = Molecular volume parameter for the UNIQUAC activity
coecient model, S = Fermentor substrate concentration, g/L
Sr = Substrate concentration in the ash tank liquid outlet ow, g/L
S0 = Substrate concentration in fresh broth, g/L
X = Fermentor biomass concentration, g/L
Xr = Biomass concentration in the ash tank liquid outlet ow, g/L
U = Overall heat-transfer coecient, W/m2 K
V = Fermentation volume, m3
ARTICLE
(10) Mariano, A. P.; Angelis, D. F.; Atala, D. I. P.; Maugeri Filho, F.;
Wolf Maciel, M. R.; Maciel Filho, R. Chem. Prod. Process Model. 2008, 3
(A), 34.
(11) Mariano, A. P.; Costa, C. B. B.; Angelis, D. F.; Atala, D. I. P.;
Maugeri Filho, F.; Wolf Maciel, M. R.; Maciel Filho, R. Chem. Eng. Res.
Des. 2010, 88, 562571.
(12) Maiorella, B.; Wilke, C. R. Biotechnol. Bioeng. 1980, 22, 17491751.
(13) Roer, S.; Blanch, H. W.; Wilke, C. R. Biotechnol. Prog. 1987,
3, 131140.
(14) Mulchandani, A.; Volesky, B. Can. J. Chem. Eng. 1986, 64, 625.
(15) Sandler, S. I. Chemical & Engineering Thermodynamics; John
Wiley & Sons: New York, 1999.
(16) Qureshi, N.; Blaschek, H. P. Appl. Biochem. Biotechnol. 2000,
84, 225235.
(17) Wooley, R.; Putsche, V. Development of an ASPEN PLUS
Physical Property Database for Biofuels Components; Report MP-42520685; NREL: Golden, CO, 1996.
(18) Luyben, L. W. Energy Fuels 2008, 22, 42494258.
(19) van der Merwe, A. B. Evaluation of Dierent Process Designs
for Biobutanol Production from Sugarcane Molasses. Master degree
thesis. University of Stellenbosch, South Africa, 2010.
(20) Fisher, K.; Gmehling, J. J. Chem. Eng. Data 1994, 39, 309315.
(21) Votruba, J.; Volesky, B.; Yerushalmi, L. Biotechnol. Bioeng. 1985,
26, 247255.
(22) Jones, D. T.; Woods, D. R. Microbiol. Rev. 1986, 50, 484524.
(23) Ezeji, T. C.; Qureshi, N.; Blaschek, H. P. In Handbook on
Clostridia; Durre, P., Ed.; Taylor and Francis: New York, 2005; Ch. 36,
pp 797812.
(24) Godin, C.; Engasser, J. M. Biotechnol. Lett. 1988, 6, 389392.
(25) Matsumura, H. K.; Sueki, M.; Araki, K. Bioprocess Eng. 1988,
3, 93100.
REFERENCES
(1) Ezeji, T. C.; Qureshi, N.; Blaschek, H. P. Curr. Opin. Biotechnol.
2007, 18, 220227.
(2) Durre, P. Biotechnol. J. 2007, 2, 15251534.
(3) Groot, W. J.; van der Lans, R. G. J. M.; Luyben, Ch. A. M. Proc.
Biochem. 1992, 27, 6175.
(4) Durre, P. Appl. Microbiol. Biotechnol. 1998, 49, 639648.
(5) Qureshi, N.; Hughes, S.; Maddox, I. S.; Cotta, M. A. Bioprocess.
Biosyst. Eng. 2005, 27, 215222.
(6) Vane, L. M. Biofuels Bioprod. Bioref. 2008, 2, 553588.
(7) Oudshoorn, A.; Van der Wielen, L. A. M.; Straathof, A. J. J. Ind.
Eng. Chem. Res. 2009, 48, 73257336.
(8) Ezeji, T. C.; Milne, C.; Price, N. D. Appl. Microbiol. Biotechnol.
2010, 85, 16971712.
(9) Mansur, M. C.; ODonnell, M. K.; Rehmann, M. S.; Zohaib, M.
ABE fermentation of sugar in Brazil. Senior Design Report, University of
Pennsylvania, 2010.
2355