Production of Bio-Ethanol From Soybean M PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

Bioresource Technology 99 (2008) 8156–8163

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Production of bio-ethanol from soybean molasses by Saccharomyces cerevisiae


at laboratory, pilot and industrial scales
Paula F. Siqueira a, Susan G. Karp a, Júlio C. Carvalho a, Wilerson Sturm a, José A. Rodríguez-León a,
Jean-Luc Tholozan b, Reeta Rani Singhania c, Ashok Pandey c, Carlos R. Soccol a,*
a
Bioprocess Engineering and Biotechnology Division, Federal University of Paraná. P.O. Box 19 011, Curitiba, Brazil
b
Laboratory of Microbiology, IRD, University of Provence/University of the Mediterranean, Marseille, France
c
Biotechnology Division, National Institute for Interdisciplinary Science and Technology (formerly Regional Research Laboratory) CSIR, Trivandrum 695 019, India

a r t i c l e i n f o a b s t r a c t

Article history: The aim of this work was to develop an economical bioprocess to produce the bio-ethanol from soybean
Received 13 January 2008 molasses at laboratory, pilot and industrial scales. A strain of Saccharomyces cerevisiae (LPB-SC) was
Received in revised form 8 March 2008 selected and fermentation conditions were defined at the laboratory scale, which included the medium
Accepted 11 March 2008
with soluble solids concentration of 30% (w/v), without pH adjustment or supplementation with the min-
Available online 15 May 2008
eral sources. The kinetic parameters – ethanol productivity of 8.08 g/L h, YP/S 45.4%, YX/S 0.815%, m
0.27 h1 and lX 0.0189 h1 – were determined in a bench scale bioreactor. Ethanol production yields after
Keywords:
the scale-up were satisfactory, with small decreases from 169.8 L at the laboratory scale to 163.6 and
Soybean molasses
Bio-ethanol
162.7 L of absolute ethanol per ton of dry molasses, obtained at pilot and industrial scales, respectively.
Pilot scale Ó 2008 Elsevier Ltd. All rights reserved.
Industrial scale
Saccharomyces cerevisiae

1. Introduction stimulated the substitution of gasoline for sugarcane alcohol for


automobile use, and intensified the use of a mixture of ethanol
Natural energy resources such as petroleum and coal have been and gasoline as fuel for common cars (Soccol et al., 2005). Anhy-
consumed at high rates over the last decades. The heavy reliance of drous ethanol is added to gasoline at a 20–26% proportion in vol-
the modern economy on these fuels is bound to end, due to their ume (Cortez et al., 2003). Today, about 3 million automobiles run
environmental impact (and the corresponding pressure of society) on 100% alcohol, and about 60% of all new motor vehicles produced
and to the fact that they might eventually run out. Therefore, alter- in Brazil are ‘‘flex”, i.e. they can run on any mixture of alcohol/gas-
native resources as ethanol are becoming more important. Bio-eth- oline, as well as on 100% alcohol (Grad, 2006).
anol is one of the most important renewable fuels contributing to A worldwide interest in the utilization of bio-ethanol as energy
the reduction of negative environmental impacts generated by the source has stimulated studies on the cost and efficiency of indus-
worldwide utilization of the fossil fuels (Cardona and Sánchez, trial processes for ethanol production. Intense research has been
2007). Some biological processes have rendered possible routes carried out for obtaining efficient fermentative organisms, low-
for producing ethanol in large volumes using the cheap substrates cost fermentation substrates, and optimal environmental condi-
(Gunasekaran and Raj, 1999). tions for fermentation to occur (Cysewski and Wilke, 1978). Even
The worldwide production of bio-ethanol (all grades) reached though the fermentative process for ethanol production is well
around 51 billion liters in 2006, of which 17 billion were produced known, the production costs are still the key impediment for the
in Brazil from sugarcane. The United States produced around 18 wide use of ethanol as fuel. Therefore, the development of fermen-
billion liters from maize. Studies on bio-ethanol production from tation processes using economical carbon sources is important for
the cellulosic materials are being financed by the US Department the ethanol production in a commercial scale (Cazetta et al., 2007).
of Agriculture (RFA, 2008). The Brazilian production of soybean is estimated in 56.6 million
Ethanol can be used directly as a fuel, but most often it is tons (2006/2007), which represent around 30% of the global pro-
blended with gasoline to yield gasohol (Staniszewski et al., duction (IBGE, 2007). Soybean molasses is a co-product generated
2007). The Brazilian National Bio-Fuel Program, initiated in 1975, in the production of protein-concentrate soybean meal. The pro-
tein-concentrate is a soy bran with around 70% protein (in dry ba-
* Corresponding author. Tel.: +55 41 33613191; fax: +55 41 33613272. sis), obtained by the extraction of sugars from de-oiled soybean
E-mail address: soccol@ufpr.br (C.R. Soccol). meal using a mixture of water/ethanol as solvent (Fig. 1), its major

0960-8524/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2008.03.037
P.F. Siqueira et al. / Bioresource Technology 99 (2008) 8156–8163 8157

Nomenclature

°Brix percentage (w/v) of soluble solids rS sugar consumption rate, dS/dt (g/L h)
S0 initial sugar concentration (g/L) rX biomass formation rate, dX/dt (g/L h)
Sf final sugar concentration (g/L) rP product formation rate, dP/dt (g/L h)
X0 initial biomass concentration (g/L) lS specific sugar consumption rate, rS  X1 (h1)
Xf final biomass concentration (g/L) lX specific biomass production rate, rX  X1 (h1)
P0 initial ethanol concentration (g/L) lP specific ethanol production rate, rP  X1 (h1)
Pf final ethanol concentration (g/L) YX/S biomass yield from sugar, rX/rS (g/g)
m maintenance coefficient (h1) YP/S ethanol yield from sugar, rP/rS (g/g)

application field being the animal feed industry. An important Bra- material is stable, and was stored at room temperature, being di-
zilian soybean-processing company produces 600 tons per day of luted with distilled water to concentrations of 15–30% soluble sol-
protein-concentrate, generating 220 tons per day of molasses ids prior to fermentation tests. The percentage (w/v) of soluble
(Siqueira, 2006). solids (°Brix) was determined with a portable refractometer for su-
Yeasts are the most commonly used microorganisms for etha- gar – Instrutherm, model RT-30 ATC (Singh et al., 1996).
nol fermentation. Anaerobic cultivation of Saccharomyces cerevisiae The carbohydrate composition of soybean molasses was deter-
generates, besides ethanol, carbon dioxide, glycerol and cell mined by HPLC (High Performance Liquid Chromatography; see
biomass as the most significant byproducts. Carbon dioxide is an Section 2.2). Protein concentration was determined by the Kjeldahl
inevitable fermentation product, but the off-gas can be sold as a method. Lipids concentration was determined by gravimetric anal-
high-quality raw material and is, therefore, more of a logistic ysis after solvent extraction with hexane (Soxhlet method). Ashes
problem. Glycerol can be produced as a compatible solute during were quantified by gravimetric analysis after burning samples at
osmotic stress (Brandberg et al., 2007). 550 °C for 5 h. Fibers concentration was calculated by difference.
The main objective of this work was to develop an economical Moisture content was determined by gravimetric analysis after
bioprocess to produce bio-ethanol from soybean molasses by a se- drying at 105 °C to constant weight.
lected strain of the yeast S. cerevisiae, through experiments at lab-
oratory, pilot and industrial scales. Fermentative assays evaluated 2.2. Determination of biomass, sugars and ethanol concentrations
the effects of initial soluble solids concentration, supplementation
with mineral sources, pH, operational mode and addition of anti- Biomass concentration in cells/mL was quantified with a Neu-
foam and dispersant agents. bauer counting chamber; biomass concentration in g/L was deter-
mined by gravimetric analysis after drying to constant weight. The
2. Methods viability of yeast cells was determined by methylene-blue staining
(Alfenore et al., 2002).
2.1. Characterization of the soybean molasses Individual sugars and ethanol were quantified by HPLC (Varian
Liquid Chromatography: solvent delivery module 240; column
The soybean molasses was received from a soybean-processing valve module 500; RI Detector 350; Workstation software 5.0) using
company in the concentrated form (75–80% soluble solids); this a Shodex KS-801 column, that separates sugars by molecular size, at

DE-OI LED
SOYBEAN MEAL
1 ton

BOLTER
ETHANOL
TANK EXTRACTOR PRESS

MISCELLA DESOLVENTIZER-
TANK TOASTER

EVAPORATOR FILTER DRYER

DEALCOHOLIZING COOLER
COLUMN
PROTEIN -
CONCENTRATOR CONCENTRATE
768.1 kg

MOLASSES
231.9 k g

Fig. 1. Production process of the soybean protein-concentrate.


8158 P.F. Siqueira et al. / Bioresource Technology 99 (2008) 8156–8163

a flow rate of 1.0 mL/min, mobile phase ultra pure H2O, temperature Fermentations were conducted until there was no liberation of
80 °C. Samples were diluted (10-fold) with ultra pure water and fil- CO2 bubbles. Enough agitation was provided by the liberation of
tered with hydrophilic PVDF membranes (0.22 lm pore size, 13 mm CO2 during fermentation and by circulation of the broth through
diameter, Millipore). Standards were Stachyose tetrahydrate (Acros the plate heat exchangers. Temperature was maintained at 30 °C.
Organics), D(+)-Raffinose pentahydrate (Acros Organics), Sucrose New inoculum (30 g/L of pressed yeast LPB-SC) was added at the
(Synth), D(+)-Glucose anhydrous (Acros Organics), D(+)-Fructose beginning of each fermentation. Antifoam and dispersant agents
anhydrous (Vetec), D(+)-Galactose (Acros Organics) and absolute tested were AE-2002 and AD-1009, respectively. These products
ethanol (Merck, 99%+). All reagents were of analytical grade. Con- consist of a mixture of silicon oil, solvents, emulsifying waxes
centrations were calculated by means of standard curves relating and tensioactive agents. Dispersant agent was added at the begin-
individual concentration to peak area. Total sugars concentration ning of feeding to avoid foam formation and antifoam was added to
was calculated by the sum of individual sugars’ concentrations. break the formed foam after feeding. Best concentrations were
determined considering effect on foam reduction and cost. The fer-
2.3. Selection of the yeast mented broth was discarded after fermentation and treated as a
common effluent.
Ten different S. cerevisiae strains (LPB 1–6, LPB-SC, LPB-MA, LPB-
JP and LPB-FR), from the Culture Collection of the Bioprocess Engi- 2.6. Fermentation tests at industrial scale
neering and Biotechnology Division/Federal University of Paraná
(DEBB/UFPR), were tested for ethanol production from soybean After the results obtained at laboratory scale were reproduced
molasses. The six strains LPB 1–6, stored in PDA agar slants at at the pilot plant and the necessary parameters were defined, the
4 °C, were reactivated in sterile YM broth, incubated in shaker at process was scaled-up again, being transferred to an industrial
30 °C, 100 rpm, for 24 h. This broth was transferred, at the propor- plant with a production capacity of 10 m3 ethanol per day. This
tion of 10% (v/v), to the inoculum medium (soybean molasses with plant consists of eight 20 m3 fermentation tanks, a system for bio-
15% soluble solids). The other four strains, available in the form of mass recovery that includes one biomass centrifuge and an agi-
pressed yeast and stored at 4 °C, were reactivated directly in the tated tank for biomass acid treatment and a distillation device
molasses medium. Inocula for fermentation assays were incubated composed of four columns (depurator, exhauster, concentrator
in shaker at 100 rpm, 30 °C, for 24 h. Volumes transferred to the and rectifier) that work at atmospheric pressure.
fermentation media were calculated so that initial biomass con- Molasses was collected directly with 20–30 °Brix. The concen-
centration was 1  108 viable cells/mL. tration was adjusted to 30 °Brix, when necessary, by the addition
Selection of the yeast strain was performed in molasses med- of concentrated molasses (80 °Brix), in an agitated tank of 5 m3. Fer-
ium with 30% soluble solids, non-sterilized, without nutrients sup- mentations were conducted as fed-batches. Biomass was diluted in
plementation or pH adjustment, in Erlenmeyer flasks incubated in water in an agitated tank and transferred to one pair of fermenta-
shaker at 100 rpm, 30 °C, for 24 h. tion tanks. Air was injected at 1 VVM to allow biomass multiplica-
tion and molasses was fed at the flow rate of 2 m3/h in each tank for
2.4. Fermentation tests at laboratory scale 4 h, until tanks were half-filled. Then, half of the content of each
tank was transferred to a pair of subsequent tanks and the proce-
Fermentation assays using the selected strain were conducted dure for biomass multiplication was repeated. The first pair of tanks
in Erlenmeyer flasks, incubated in shaker (Innova, model 4080) at was fed again with molasses at 4 m3/h, until tanks were filled to 80%
30 °C, 100 rpm. Inoculum (yeast biomass 10 g/L, with 57% moisture of their capacities, and fermentation was conducted, with no aera-
and a minimum of 95% viable cells) was previously cultivated in di- tion, to the end of CO2 liberation. This procedure was repeated for
luted molasses (15% soluble solids, not sterilized) at 30 °C, the other pairs of tanks. Dispersant agent was added at the begin-
100 rpm, for 12 h. Media were then concentrated by addition of ning of feeding and antifoam was used to break the formed foam
molasses 75 °Brix to the desired final concentration. Biomass con- after feeding. Temperature control (at 30 °C) and agitation were as-
centration at the beginning of fermentation was adjusted to sured by circulation of the broth through plate heat exchangers.
1  108 viable cells/mL. Fermentations were conducted until CO2 When fermentation was finished, biomass was separated by
liberation was over (visual analysis). The effects of initial °Brix, continuous centrifugation, the yeast cream containing around
addition of inorganic salts as sources of magnesium (MgSO4.7H2O, 60% solids was treated with acid (H2SO4 to pH 2.2 for 2 h), in order
Reagen) and nitrogen (NH4NO3, Nuclear) (Machado, 1999) and ini- to avoid flocculation and inactivate the weak cells, in a 5.5 m3 agi-
tial pH were tested. tated tank, and then transferred to the fermentation tanks to be
Fermentations to determine kinetic parameters were conducted used as inoculum. Biomass concentration at the beginning of fer-
in bioreactor (8 L, MDL B.E. Marubishi), filled with 6 L of non-ster- mentation was adjusted to at least 3  108 viable cells per mL,
ilized medium. The bioreactor had agitation and temperature the inoculum volume being around 30% of the total fermentation
controlled. volume. The wine was distilled for ethanol recovery.

2.5. Fermentation tests at pilot scale 2.7. Determination of fermentation yields and kinetic parameters

The process developed at laboratory scale was transferred to a The maximum theoretical ethanol yield from sugar was calcu-
pilot scale plant. Fermentations were carried out in two bioreactors lated according to the stoichiometric relation represented by Eq.
with a total capacity of 1 m3 each. Molasses was collected directly (1), i.e., 100 g of hexose produce 51.1 g of ethanol and 48.9 g of
with 20–30 °Brix, before the concentration step to 80 °Brix. One of CO2. Ethanol yields over total initial sugars (Y1) and consumed sug-
the tanks was used for storing the must and the other for inoculum ars (Y2) were calculated according to Eqs. (2) and (3).
preparation and fermentation. The process was carried out as batch
C6 H12 O6 ! 2CH3 CH2 OH þ 2CO2 ð1Þ
and fed-batch. For the batch process, biomass (pressed yeast) was
½p  p0 ; g=L  100
diluted with molasses and, after homogenization, the tank was Y 1 ð%Þ ¼ f ð2Þ
filled with molasses to 80% of its capacity. For the fed-batch pro- ½S0 ; g=L  0:511
cess, biomass was diluted in water and molasses was fed at the ½pf  p0 ; g=L  100
Y 2 ð%Þ ¼ ð3Þ
flow rate of 200 L/h (feeding time 4 h). ½S0  Sf ; g=L  0:511
P.F. Siqueira et al. / Bioresource Technology 99 (2008) 8156–8163 8159

A mathematical model (polynomial equation of third order) was Table 2


adjusted to experimental data, representing concentrations as time Ethanol production yields over total initial sugars (Y1) after 24 h of fermentation for
different Saccharomyces cerevisiae strains
functions. The maintenance coefficient was determined by solving
the mass balance for substrate consumption represented by Eq. (4). Strain Total sugars (g/L) Ethanol (g/L) Y1 (%) Standard deviation
By-products formation was not considered. CO2 formation was cal- N.f.m.a 222.1 – – –
culated in relation to produced ethanol, according to Eq. (1). Con- LPB 1 135.2 38.9 34.3 1.57
version yields, specific rates and the maintenance coefficient LPB 2 125.0 35.9 31.6 0.0567
LPB 3 131.5 37.0 32.6 1.29
were determined either using experimental data or the values pre- LPB 4 135.6 42.9 37.8 2.04
dicted by the model. LPB 5 134.9 37.5 33.0 1.82
LPB 6 118.3 33.2 29.3 1.68
ds dX dP dCO2
¼ þ þ þ mX ð4Þ LPB-SC 126.9 47.1 41.5 2.07
dt dt dt dt LPB-MA 145.0 30.6 27.0 0.551
LPB-JP 121.7 30.5 26.9 0.905
LPB-FR 127.9 30.3 26.7 2.13
3. Results and discussion
Initial concentrations were 30 °Brix for the soybean molasses (non-supplemented,
natural pH of 5.5) and 1  108 cells/mL for biomass. Average of two experiments.
3.1. Characterization of soybean molasses a
N.f.m. – non-fermented medium.

The composition of soybean molasses is shown in Table 1. Car-


bohydrates, representing 57.3% of molasses’ dry mass, include tion of yeasts in industrial ethanol production as the result of high
mainly sucrose (28.4%), which is a fermentable disaccharide for osmotic pressure (Takeshige and Oushi, 1995). On the other side,
yeasts, and also stachyose (18.6%) and raffinose (9.68%), complex low dilution rates represent an economy of equipment and process
tetra- and trisaccharides, respectively, that are not fermentable costs (e.g., distillation costs). In this way, for industrial applica-
by many microorganisms due to the presence of a-1,6 bonds (Lan tions, it was established that the concentration of 30 °Brix would
et al., 2007). The molasses contains also significant amounts of pro- be more appropriate.
teins (9.44%) and lipids (21.2%).
3.3.2. Effect of nutrients supplementation to the soybean molasses
3.2. Selection of the yeast medium in bio-ethanol production
Results presented in Table 4 show that the addition of nutrients
Table 2 shows the results of the microorganism screening test. to the medium did not improve ethanol production and process
Analysis of variance at the level of 5% demonstrated that at least yield. The highest yield (48.86%) was obtained when the medium
one average yield is statistically different. According to the Tukey was not supplemented with salts. The soybean molasses is rich
test, the ethanol yield presented by the strain LPB-SC (41.5% of in mineral salts (6.36% ashes) and contains a considerable amount
the theoretical from total initial sugars) was significantly higher of proteins that can be used as nitrogen source by the yeast. Effect
than the other averages, except for strains LPB 1, 4 and 5. Since of nutrients addition was tested in a medium with 20% soluble sol-
the strain LPB-SC was available in the form of pressed yeast (thus ids to avoid inhibition by osmotic pressure. For other substrates
facilitating inoculum preparation – see Section 2.3), it was chosen such as beet molasses, addition of nitrogen source (ammonium
for the subsequent fermentation tests with soybean molasses. sulfate 0.2 g/L) can improve ethanol production in about 10%
(Nahvi et al., 2002). Because of the digestible nitrogen deficiency
3.3. Definition of fermentative conditions at laboratory scale of sugar beet molasses, ammonium phosphate, ammonium dihy-
drogen phosphate and ammonium sulfate are usually added to
3.3.1. Effect of initial concentration of soybean molasses in bio-ethanol the fermentation medium for better productivity (Ergun and Mut-
production lu, 2000).
Results presented in Table 3 show that the only significant drop
in fermentation yield occurs for the initial concentration of 35 °Brix 3.3.3. Effect of initial pH in bio-ethanol production
(38.53%), the highest value being 42.57%, obtained at 25 °Brix. As it There was no significant difference among the initial pH tested
was expected, fermentation time increases in direct proportion to (4.3, 5.0 and 5.5) regarding ethanol production, comparing with
initial soluble solids concentration. However, decrease in produc- non-supplemented soybean molasses diluted to 30 °Brix. The
tivity was observed only at 35 °Brix. Several studies have reported medium fermented at the natural pH of soybean molasses (5.5)
that high substrate concentrations inhibit growth and fermenta- presented an average yield over total initial sugars (Y1) of 50.39%
(2.92 standard deviation) after 25 h of fermentation. Yields
obtained in this experiment were high in comparison to the ones
Table 1 presented in Table 3 (average 42.08% for the same conditions). This
Composition of soybean molasses difference may be due to changes in the molasses’ sugar
Component % in dry basis Standard deviation
composition.

Glucose 0.243 0.06860


3.3.4. Kinetics of bio-ethanol production at laboratory scale
Fructose 0.127 0.06576
Galactose 0.254 0.03250 The fermentation development was adjusted according to the
Sucrose 28.4 2.069 desired condition at industrial scale, a 6 h fermentation step. For
Lactose – – that, 30 g/L of pressed yeast had to be added at the beginning of
Raffinose 9.68 1.287
fermentation, providing an initial concentration of 3  108 viable
Stachyose 18.6 2.828
Total carbohydrates 57.3 1.381
cells/mL (6 g cells/L in dry basis). Table 5 shows the kinetics of
Proteins 9.44 1.160 bio-ethanol production from soybean molasses under optimized
Lipids 21.2 2.680 conditions in bioreactor. The maximum ethanol productivity
Fibers 5.7 2.675 (10.8 g/L h) occurred after 3 h of fermentation. A product yield
Ash 6.36 1.387
(YP/S) of 45.4% from consumed substrate represents 88.8% of the
Average of three different samples. theoretical maximum. Considering only the initial substrate
8160 P.F. Siqueira et al. / Bioresource Technology 99 (2008) 8156–8163

Table 3
Effect of media initial soluble solids concentration (°Brix) on ethanol productivity and yield over total initial sugars (Y1)

°Brix Fermentation time (h) Total sugars (g/L) Ethanol (g/L) Average productivity (g/L h) Y1 (%) Average yield (%) Standard deviation
20 0 169.5 0 –
20 87.4 38.0 1.820 43.87 42.08 1.266
20a 93.4 34.9 40.29
25 0 207.3 0 –
25 108.2 46.6 1.804 43.99 42.57 1.001
25a 107.8 43.6 41.16
30 0 267.5 0 –
30 138.3 52.1 1.867 38.11 40.97 2.019
30a 134.3 59.9 43.82
35 0 311.6 0 –
40 152.3 63.5 1.530 39.88 38.53 0.9546
40a 151.3 59.2 37.18

Media fermented without nutrients supplementation and pH adjustment. Initial biomass concentration was 1  108 cells/mL.
a
Represents the duplicate experiment.

concentration (182 g/L), ethanol yield was 45.8% of the theoretical Values representing yield in biomass (YX/S, 0.00815) and growth
maximum. This is equivalent to 134.1 kg or 169.8 L of absolute eth- rate (lX, 0.0189) are very small in comparison to product yields
anol per ton of dry molasses, since one ton of dry molasses con- (Table 6). These small values were expected since there is no oxy-
tains 573 kg of sugar (Table 1). gen supply and the soluble solids concentration is high (30%) and,
Experimental data (presented in Fig. 2) were used to derive consequently, dissolved oxygen concentration is low. Echegaray
third order polynomials, representing concentrations (S, X and P) et al. (2000) reported values as high as 0.116 for YX/S and ranging
as functions of fermentation time (t). These are Eqs. (5)–(7) pre- from 0.24 to 0.019 for lX for the anaerobic fermentation of sugar-
sented below. Kinetic parameters based on experimental data cane molasses by S. cerevisiae containing around 170 g/L of sugars
and on the values predicted by the model are presented in Table 6. (for this substrate, equivalent to 17% soluble solids). The decrease
in lX along time was possibly due to product inhibition.
SðtÞ ¼ 0:536t3  5:74t 2  0:631t þ 182; R2 ¼ 0:999 ð5Þ
The calculated maintenance coefficient of 0.27 h1 (Table 6)
3 2 2
XðtÞ ¼ 0:0036t þ 0:0367t þ 0:0296t þ 6:00; R ¼ 0:998 ð6Þ was similar to the value reported by Pirt (1975) for anaerobic fer-
PðtÞ ¼ 0:244t3 þ 2:57t 2 þ 0:468t þ 0:338; R2 ¼ 0:998 ð7Þ mentation by S. cerevisiae cultivated in glucose medium with 1 M
NaCl (0.36 h1), which has a osmotic pressure comparable to that
on the 30% molasses medium.
Table 4
Ethanol yields over total initial sugars (Y1) after 20 h of fermentation 3.3.5. Consumption of individual sugars
Structural analysis of the main sugars present in soybean
Sample Total sugars (g/ Ethanol Y1 Average yield Standard
molasses (stachyose, raffinose and sucrose) shows that all of them
L) (g/L) (%) (%) deviation
have a b-1,2 bond, which is cleaved by the enzyme invertase (b-D-
M 156.1 0 – – –
fructofuranoside fructohydrolase, E.C. 3.2.1.26). It is known that
Ma 151.9 0 –
A 73.3 39.1 49.69 48.86 0.5870 the yeast Saccharomyces cerevisiae produces intra and extracellular
Aa 65.2 37.8 48.03 invertase. The extracellular invertase resides in the periplasmic
B 70.1 36.7 46.64 43.15 2.471 space and is responsible for cleaving sucrose into glucose and fruc-
Ba 70.2 31.2 39.65 tose, monomers that are assimilated and converted into ethanol
C 76.3 33.0 41.93 42.95 0.7112
Ca 75.2 34.6 43.97
(Zech and Görisch, 1995). The other bonds, except for sucrose
D 72.5 34.7 44.09 44.28 0.1343 which is a disaccharide, are of the a-1,6 type, cleaved by the en-
Da 75.8 35.0 44.47 zyme a-galactosidase (E.C. 3.2.1.22). This enzyme is probably not
produced by the yeast LPB-SC, since there is such a high concentra-
A, no addition of nutrients; B, addition of magnesium source (MgSO4, 0.1 g/L); C,
addition of nitrogen source (NH4NO3, 3.5 g/L); D, addition of magnesium and tion of residual sugars after fermentation (Table 5).
nitrogen sources; M, non-fermented medium. Initial concentrations were 20 °Brix The main sugars that remain after fermentation are tri- and
for soybean molasses and 1  108 cells/mL for biomass. disaccharides, whereas 48% of the soybean molasses’ sugars are
a
Represents the duplicate experiment.
not fermented by the strain LPB-SC. Trisaccharides present in

Table 5
Kinetic study of bio-ethanol production from soybean molasses in bench scale bioreactor under optimized conditions: medium with 30% soluble solids, without pH adjustment or
nutrients supplementation, initial biomass concentration 3  108 cells/mL

Time (h) S (g/L) X (g/L) P (g/L) rS (g/L h) rX (g/L h) rP (g/L h) YX/S (%) YP/S (%)
0 182.0 6.00 0.400 – – – – –
1 176.1 6.05 3.20 5.9 0.05 2.8 0.847 47.5
2 163.6 6.18 8.80 12.5 0.13 5.6 1.04 44.8
3 140.5 6.31 19.6 23.1 0.13 10.8 0.563 46.7
4 123.9 6.49 27.0 16.6 0.18 7.4 1.08 44.6
5 102.1 6.59 36.4 21.8 0.10 9.4 0.459 43.1
6 87.6 6.72 43.0 14.5 0.13 6.6 0.897 45.5
Average 15.7 0.12 8.08 0.815 45.4

Concentrations represent the average of two experiments. Average standard deviations were 4.77% for sugar, 2.10% for biomass and 5.24% for ethanol concentrations.
P.F. Siqueira et al. / Bioresource Technology 99 (2008) 8156–8163 8161

200 6.8 3.4. Bio-ethanol production at pilot scale


180 6.7
The main objective of scaling-up a process is to identify prob-
ConcentrationS, P (g/L)

160

Concentration X (g/L)
6.6 lems that were not significant at laboratory scale, and also check
140 S 6.5 if the fermentation yield is maintained. For the first batch fermen-
120 tation at pilot scale, a major problem identified was the high foam
P 6.4
100 formation. The soybean molasses has a significant concentration of
X 6.3
80 proteins and lipids that may contribute to increase viscosity and
6.2 surface tension. Although molasses’ viscosity should diminish
60
6.1 while ethanol concentration increases, the intense production of
40
CO2 bubbles is a factor that strongly favors foam formation (Togrul
20 6
and Arslan, 2004).
0 5.9 An alternative to control foam formation is the fed-batch pro-
0 1 2 3 4 5 6
cess, which is considered one of the most useful systems for
Time (h)
economical ethanol fermentation (Roukas, 1996). Controlled
Fig. 2. Evolution of substrate (S), ethanol (P) and biomass (X) concentrations during substrate feeding, besides avoiding intense CO2 production, pre-
bench scale fermentation of soybean molasses under optimized conditions, and the vents inhibition and catabolite repression, improving productivity
corresponding third order polynomial tendency lines. of fermentation by maintaining a low substrate concentration (Pra-
sad et al., 2007). Addition of antifoam and dispersant agents was
also necessary to prevent foam formation.
Table 6
Fermentation kinetic parameters determined by using experimental data and by the
Table 7 shows the results of 11 fermentation cycles conducted
resolution of the mathematical model (average values) at the pilot scale plant. Effects of antifoam and dispersant agents
addition are shown in Table 8. An average product yield of
YX/S YP/S m lS lX lP
44.13% (standard deviation 1.323) over the total initial sugars (Ta-
Experimental data 0.00815 0.454 0.270 2.47 0.0189 1.12 ble 7), represents an ethanol yield of 129.2 kg or 163.6 L of absolute
Mathematical model 0.00837 0.456 0.274 2.48 0.0189 1.12
ethanol per ton of dry molasses. The average productivity for the
fed-batch process was 7.882 g/L h (standard deviation 0.8560).
Roukas (1996) reported a maximum productivity value of 3.8 g/
soybean molasses are composed of raffinose and partially hydro- L h for the fed-batch fermentation of beet molasses with initially
lyzed stachyose, which is a tetrasaccharide. In the same way, disac- 250 g/L sugars and 3.7  108 cells/mL. These conditions are similar
charides include sucrose and partially hydrolyzed raffinose. In this to those of the eleventh fermentation cycle (Table 7), when a pro-
way, once it is known that sucrose is utilized by the yeast, it is sup- ductivity of 6.916 g/L h was achieved. However, it is important to
posed that the residual carbohydrates are oligomers linked by remark that the sugarbeet molasses is composed mainly of sucrose
a-1,6 bonds. From stachyose and raffinose, only the terminal fruc- (Vicik et al., 1990), while soybean molasses contains almost 50% of
tose and glucose units would be consumed, respectively. non-fermentable sugars, consequently, ethanol concentration is
Based on these considerations, it is possible to calculate the lower and product inhibition may be less expressive.
maximum theoretical yield in ethanol from the soybean molasses, The dispersant agent added at 5 mL/m3 showed good efficiency
counting only the fermentable sugars. Since glucose, fructose and in preventing the foam formation during the feeding. A minimum
galactose have the same molecular mass, the fourth and the third concentration of 35 mL/m3 of antifoam agent had to be added after
part of stachyose’s and raffinose’s mass concentrations, respec- feeding to dissolve the formed foam and prevent its formation dur-
tively, can be considered fermentable sugars. In this way, according ing the last 2 h of ‘‘batch” fermentation (Table 8). After the results
to Table 1, the fermentable sugars’ concentration in soybean obtained at lab scale were reproduced at the pilot plant and the
molasses is 36.90% (dry basis), or 64.40% of the total amount of necessary parameters were defined, the process was transferred
sugars. Therefore, the maximum theoretical yield in ethanol from to an industrial plant for ethanol production, with a production
soybean molasses’ dry mass would be 18.86% (w/w) or 23.70% capacity of 10 m3 ethanol per day.
(v/w).

Table 7
Results of pilot scale fermentations for bio-ethanol production from soybean molasses at various °Brix using an initial biomass concentration of 3  108 cells/mL

Cycle nr./op. modea °Brix Initial sugar (g/L) Final sugar (g/L) Ethanol (g/L) Time (h) Productivity (g/L h) Yieldb (%)
c
1/B 20.0 116.6 54.00 26.10 6 4.350 43.80
2/B 30.0 251.8 118.2 58.60 6 9.767 45.54
3/FBd 21.0 166.9 59.1 39.28 7 5.611 46.06
4/FB 22.0 168.9 79.2 41.92 5 8.384 48.57
5/FB 23.0 185.0 89.9 43.72 5 8.744 46.25
6/FB 21.0 182.1 83.3 40.01 4 10.00 43.00
7/FB 21.0 189.5 76.9 40.50 6 6.750 41.82
8/FB 21.5 190.7 85.0 40.81 6 6.802 41.88
9/FB 22.0 180.8 83.2 41.14 5 8.228 44.53
10/FB 23.0 173.2 80.0 38.00 4 9.500 42.93
11/FB 30.0 230.7 74.37 48.41 7 6.916 41.06
Average productivity/yield 7.882e 44.13
a
Operational mode.
b
Yield over total initial sugars.
c
B, batch.
d
FB, fed batch.
e
For the fed-batches only.
8162 P.F. Siqueira et al. / Bioresource Technology 99 (2008) 8156–8163

Table 8 acid treatment. The pre-treatment may also preserve the equip-
Effect of antifoam (R$ 4.40/kg) and dispersant (R$ 11.50/kg) agents addition on foam ments (tanks, pipelines, distillation columns).
control and production cost for fed-batch fermentations at pilot scale
As presented in Fig. 3, the mass balance based on industrial data
Cycle Anti-foam Dispersant Foam Ethanol Additional showing all products obtained from soybean processing, the fer-
number (mL/m3) (mL/m3) reduction (L/m3) cost (R$/L) mentation yield (18.39 kg of absolute ethanol per ton of soybean),
4 100 100 + 49.72 0.0309 representing a yield of 43.89% of the theoretical over the total ini-
5 30 30 + 53.06 0.0092 tial sugars, or 162.7 L of absolute ethanol per ton of dry soybean
6 30 30 + 52.34 0.0091
7 40 20 + 55.34 0.0068
molasses, was maintained after the second scale-up.
8 20 20 + 50.65 0.0063 Because of environmental problems, the treatment of the resi-
9 40 20 + 51.27 0.0081 due generated by the distillation process, called vinasse, is one of
10 40 10 + 54.30 0.0053 the most significant and challenging issues in the industrial pro-
11 35 5 + 51.66 0.0041
duction of ethanol. A typical distillery with a daily ethanol produc-
+ Represents sufficient foam reduction. tion of 100 m3 has a vinasse discharge of 1300 m3 having a high
pollution load, with BOD values ranging from 30 to 60 gO2/L (Nav-
arro et al., 2000). The BOD of the vinasse produced by the fermen-
3.5. Implementation of the process at industrial scale tation of soybean molasses was determined (LCK 555 kit, Dr.
Lange) and presented a value of 77.2 gO2/L.
For the industrial scale fermentations, the molasses concentra- Siqueira (2006) reported that enzymatic and acid hydrolyses
tion was adjusted to 30% soluble solids, in order to provide an eth- are alternatives to increase ethanol yield from soybean molasses’
anol concentration in the fermented broth of 40–50 g/L. This is the sugars and, consequently, reduce the BOD of the vinasse. Enzy-
minimum concentration required for an economically feasible dis- matic hydrolysis of the molasses by the enzyme a-1,6-galactosi-
tillation process (Siqueira, 2006). Foam formation was easily con- dase provided an increase of 20% in ethanol yield from total
trolled by the techniques developed at pilot scale. sugars. However, the yield improvement did not compensate for
A major problem identified after starting-up the plant was the the additional cost. The acid hydrolysis of the vinasse to release
contamination with bacteria. Microscopy analysis indicated the fermentable sugars was also tested and provided an increase of
presence of gram-positive bacilli. This problem was expected since 17% in ethanol yield, but due to the severe conditions demanded
the biomass is recycled along the process and may lose activity, (HCl at 10% v/v, pH 0.1, 100 °C, 60 min), the process demonstrated
allowing the development of opportunistic contaminants. The to be economically unfeasible (data not shown).
problem was solved by the addition of appropriate antibiotic – Currently, the vinasse produced at the industrial plant is being
commonly used in industrial ethanol fermentations, does not affect concentrated and incinerated to produce energy, since it cannot be
yeast metabolism – together with the previous treatment of soy- treated as a common wastewater due to its high organic charge.
bean molasses (removal of solids), in order to facilitate the separa- The utilization of the vinasse as raw material for other fermenta-
tion of yeast biomass at the end of fermentation and the inoculum tive processes is being studied.

SOYBEAN 1 ton
35% protein
13% moisture

LECITHIN OIL DE-OILED SO YBEAN HULLS


10 kg 195 kg MEAL 716 kg 50 kg
48% protein
13.8% moisture

MOLASSES 190.8 kg PROTEIN-CONC ENTRATE


57.3% sugars 522.0 kg
25% moisture 72% protein
9.18% moisture

ETHANOL 18.39 kg

VINASSE 533.6 kg
35.0% sugars
80.5% moisture

Fig. 3. Mass balance: from soybean to ethanol. Protein and sugar percentages in dry basis.
P.F. Siqueira et al. / Bioresource Technology 99 (2008) 8156–8163 8163

4. Conclusions IBGE – Instituto Brasileiro de Geografia e Estatística (online). <http://


www.ibge.gov.br/home/presidencia/noticias/noticia_visualiza.php?id_noticia=
832&id_pagina=1> (accessed 25.01.07).
From the results, it could be concluded that soybean molasses is Lan, Y., Williams, B.A., Verstegen, M.W.A., Patterson, R., Tamminga, S., 2007. Soy
an attractive raw material for the production of bio-ethanol. Since oligosaccharides in vitro fermentation characteristics and its effect on caecal
microorganisms of young broiler chickens. Animal Feed Sci. Techol. 133, 286–
it provided the necessary nitrogen and magnesium and the appro-
297.
priate hydrogen balance for the fermentation, there was no need Machado, R.P., 1999. Producßão de etanol a partir de melacßo de soja. Master
for supplementing these additionally or making any pH adjust- Dissertation, Federal University of Rio Grande do Sul, Porto Alegre, Brazil.
Nahvi, I., Emtiazi, G., Alkabi, L., 2002. Isolation of a flocculating Saccharomyces
ment. The laboratory process was scaled-up with promising re-
cerevisiae and investigation of its performance in the fermentation of beet
sults. The residue of bio-ethanol production, called soybean molasses to ethanol. Biomass Bioenergy 23, 481–486.
vinasse, did not present any environmental problem since it was Navarro, A.R., Sepúlveda, M.C., Rubio, M.C., 2000. Bio-concentration of vinasse from
employed as raw material for energy co-generation. the alcoholic fermentation of sugar cane molasses. Waste Manage. 20, 581–585.
Pirt, S.J., 1975. Principles of Microbe and Cell Cultivation, first ed. Blackwell
Scientific, Oxford, UK.
References Prasad, S., Singh, A., Joshi, H.C., 2007. Ethanol as an alternative fuel from
agricultural, industrial and urban residues. Res. Conserv. Recycl. 50, 1–39.
Alfenore, S., Molina-Jouve, C., Guillouet, S.E., Uribelarrea, J.L., Goma, G., Benbadis, L., RFA – Renewable Fuels Association (online). <http://www.ethanolrfa.org/>
2002. Improving ethanol production and viability of Saccharomyces cerevisiae by (accessed 29.02.08).
a vitamin feeding strategy during fed-batch process. Appl. Microbiol. Roukas, T., 1996. Ethanol production from non-sterilized beet molasses by free and
Biotechnol. 60 (1–2), 67–72. immobilized Saccharomyces cerevisiae cells using fed-batch culture. J. Food Eng.
Brandberg, T., Gustafsson, A., Franzén, C.J., 2007. The impact of severe nitrogen 27, 87–96.
limitation and microaerobic conditions on extended continuous cultivations of Singh, P.C., Bhamidipati, S., Singh, R.K., Smith, R.S., Nelson, P.E., 1996. Evaluation of
Saccharomyces cerevisiae with cell recirculation. Enzyme Microb. Technol. 40, in-line sensors for prediction of soluble and total solids/moisture in continuous
585–593. processing of fruit juices. Food Control 7 (3), 141–148.
Cardona, C.A., Sánchez, O.J., 2007. Fuel ethanol production: process design trends Siqueira, P.F., 2006. Production of bio-ethanol from soybean molasses by
and integration opportunities. Biores. Technol. 98, 2415–2457. Saccharomyces cerevisiae. Master Dissertation, Federal University of Paraná/
Cazetta, M.L., Celligoi, M.A., Buzato, J.B., Scarmino, I.S., 2007. Fermentation of Universities of Provence and of the Mediterranean, Curitiba, Brazil.
molasses by Zymomonas mobilis: effects of temperature and sugar Soccol, C.R., Vandenberghe, L.P.S., Costa, B., Woiciechowski, A.L., Carvalho, J.C.,
concentration on ethanol production. Biores. Technol. 98, 2824–2828. Medeiros, A.B.P., Francisco, A.M., Bonomi, L.J., 2005. Brazilian biofuel program:
Cortez, L.A.B., Griffin, M.W., Scaramucci, J.A., Scandiffio, M.I.G., Braunbeck, O.A., an overview. J. Sci. Ind. Res. 64, 897–904.
2003. Considerations on the worldwide use of bio-ethanol as a contribution for Staniszewski, M., Kujawski, W., Lewandowska, M., 2007. Ethanol production from
sustainability. Manage. Environ. Qual. 14, 508–519. whey in bioreactor with co-immobilized enzyme and yeast cells followed by
Cysewski, G.R., Wilke, C.R., 1978. Process design and economic studies of alternative pervaporative recovery of product – kinetic model predictions. J. Food Eng. 82,
fermentation methods for the production of ethanol. Biotechnol. Bioeng. 20, 618–625.
1421–1427. Takeshige, K., Oushi, K., 1995. Effect of yeast invertase on ethanol production in
Echegaray, O.F., Carvalho, J.C.M., Fernandes, A.N.R., Sato, S., Aquarone, E., Vitolo, M., molasses. J. Ferm. Bioeng. 79 (5), 513–515.
2000. Fed-batch culture of Saccharomyces cerevisiae in sugar-cane blackstrap Togrul, H., Arslan, N., 2004. Mathematical model for prediction of apparent viscosity
molasses: invertase activity of intact cells in ethanol fermentation. Biomass of molasses. J. Food Eng. 62, 281–289.
Bioenergy 19, 39–50. Vicik, S.M., Fedor, A.J., Swartz, R.W., 1990. Defining an optimal carbon source/
Ergun, M., Mutlu, S.F., 2000. Application of a statistical technique to the production methionine feed strategy for growth and cephalosporin C formation by
of ethanol from sugar beet molasses by Saccharomyces cerevisiae. Biores. Cephalosporium acremonium. Biotechnol. Prog. 6 (5), 333–340.
Technol. 73, 251–255. Zech, M., Görisch, H., 1995. Invertase from Saccharomyces cerevisiae: reversible
Grad, P., 2006. Biofuelling Brazil: an overview of the bioethanol success story in inactivation by components of industrial molasses media. Enzyme Microb.
Brazil. Refocus 7 (3), 56–59. Technol. 17, 41–46.
Gunasekaran, P., Raj, K.C., 1999. Ethanol fermentation technology – Zymomonas
mobilis. Curr. Sci. 77, 56–68.

You might also like