b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 1 9 7 7 e1 9 8 2

Available at www.sciencedirect.com

http://www.elsevier.com/locate/biombioe

Optimal fermentation conditions for maximizing

the ethanol production by Kluyveromyces fragilis
from cheese whey powder

Giuliano Dragone a,*, Solange I. Mussatto a, João B. Almeida e Silva b, José A. Teixeira a
a
IBB e Institute for Biotechnology and Bioengineering, Centre of Biological Engineering, University of Minho, Campus de Gualtar,
4710-057 Braga, Portugal
b
Department of Biotechnology, Engineering College of Lorena, University of São Paulo, Estrada Municipal do Campinho s/n,
12602-810 Lorena/SP, Brazil

article info abstract

Article history: Cheese whey powder (CWP) is an attractive raw material for ethanol production since it is
Received 7 October 2009 a dried and concentrated form of CW and contains lactose in addition to nitrogen, phos-
Received in revised form phate and other essential nutrients. In the present work, deproteinized CWP was utilized
23 January 2011 as fermentation medium for ethanol production by Kluyveromyces fragilis. The individual
Accepted 25 January 2011 and combined effects of initial lactose concentration (50e150 kg m
3), temperature
Available online 17 February 2011 (25e35  C) and inoculum concentration (1e3 kg m
3) were investigated through a 23 full-
factorial central composite design, and the optimal conditions for maximizing the ethanol
Keywords: production were determined. According to the statistical analysis, in the studied range of
Biofuels values, only the initial lactose concentration had a significant effect on ethanol production,
Cheese whey resulting in higher product formation as the initial substrate concentration was increased.
Ethanol Assays with initial lactose concentration varying from 150 to 250 kg m
3 were thus per-
Fermentation formed and revealed that the use of 200 kg m
3 initial lactose concentration, inoculum
Kluyveromyces fragilis concentration of 1 kg m
3 and temperature of 35  C were the best conditions for maxi-
Lactose mizing the ethanol production from CWP solution. Under these conditions, 80.95 kg m
3 of
ethanol was obtained after 44 h of fermentation.
ª 2011 Elsevier Ltd. All rights reserved.

1. Introduction whey is an alternative of great interest for reuse of this

industrial by-product [4]. In addition, the development of
The dairy industry represents an important part of the food ethanol production methods is stimulated by the possibility of
processing industry and contributes significant liquid process using ethanol as a component in biofuels [5,6]. However, the
residues that can be used for the production of ethanol [1]. production of ethanol from non-concentrated CW is not
Cheese whey (CW), a by-product of the cheese manufacturing economically feasible because the levels of ethanol obtained
process whose major components are lactose (45e50 kg m
3), at the end of fermentation reach only about 20e30 dm3 m
3.
proteins (6e8 kg m
3), lipids (4e5 kg m
3), and mineral salts Distillation costs for ethanol separation from dilute fermen-
(8e10% of dried extract), constitutes an inexpensive and tation broths (20e30 dm3 m
3 EtOH) is a major cost item in
nutritionally rich raw material for the production of different ethanol fermentation of CW [2]. Ultrafiltration (UF) processes
compounds [2,3]. Ethanol production by bioconversion of have been used to concentrate lactose in CW before

* Corresponding author. Tel.: þ351 253 604 424; fax: þ351 253 604 429.
E-mail address: [email protected] (G. Dragone).
0961-9534/$ e see front matter ª 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biombioe.2011.01.045
1978 b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 1 9 7 7 e1 9 8 2

fermentation [7]. UF improves the lactose concentration by The precipitates were removed by centrifugation at 5600 g and
a factor of 5e6, but is expensive (approx. 50 $ m
3 original 10  C for 15 min, and the supernatants were used as fermen-
dilute CW) [8]. tation medium.
Dry cheese whey powder (CWP) may be an attractive raw Batch fermentations were performed in 500 ml Erlenmeyer
material for ethanol production. CWP is a dried and concen- flasks containing 100 ml of medium. Flasks were maintained
trated form of CW and contains lactose in addition to nitrogen, in an orbital shaker at 2.5 Hz for 44 h. Different values of initial
phosphate and other essential nutrients [9]. Utilization of CWP lactose concentration, temperature and inoculum concen-
instead of CW for ethanol fermentations has significant advan- tration were used in the experiments (Table 1).
tages such as elimination of costly ultrafiltration processes to In the second step, the assays for determination of the best
concentrate lactose before fermentation, compact volume, long- initial lactose concentration (varying from 150 to 250 kg m
3)
term stability and high concentrations of lactose and other were also performed in Erlenmeyer flasks as above described.
nutrients yielding high ethanol concentrations by fermentation. However, in this case the temperature and initial inoculum
Moreover, the cost of CWP production from CW by spray or drum concentration were fixed at 35  C and 1 kg m
3, respectively.
drying varies between 0.2 and 0.4 $ per kg CWP (10e20 $ m
3 The fermentation runs were monitored through periodic
original dilute CW), which is much lower than distillation costs sampling in order to determine the cell growth, lactose
for pure ethanol production from dilute CW [8,10]. consumption and ethanol production. All the experiments were
It is known that the fermentation process performance is performed in duplicate and mean values are given. The kinetic
affected by operational conditions such as temperature, stir- parameters of fermentations were calculated at the end of the
ring rate, initial inoculum and substrate concentrations, dis- runs. The ethanol yield factor (YP/S, kg kg
1) was defined as the
solved oxygen, among others. A suitable control of these ratio between the ethanol concentration (kg m
3) and lactose
variables is of great importance for a good process perfor- consumed (kg m
3). The ethanol yield per cell (YP/x, kg kg
1) was
mance and obtainment of high-quality products. The present defined as the ratio between ethanol and total cell concentra-
study aimed to optimize the conditions for ethanol production tions (kg m
3). The ethanol productivity (QP, kg m
3 h
1) was
from CWP through RSM designed with central composite defined as the ratio between ethanol concentration (kg m
3) and
design. Three factors were selected as process (independent) fermentation time (h). The efficiency of ethanol production
variables: initial lactose concentration, temperature and (h, %) was defined as the ratio between the ethanol concentra-
inoculum concentration; while the ethanol concentration, tion (kg m
3) and the maximum theoretical ethanol concentra-
substrate consumption and fermentative parameters (ethanol tion (kg m
3) that could be achieved considering the theoretical
yield factor, YP/S; ethanol volumetric productivity, QP; ethanol value of 0.538 kg ethanol per kg consumed lactose [11].
yield per cell, YP/x; and bioconversion efficiency, h) were
selected as responses (dependent) variables. 2.3. Analytical methods

The fermented media samples were centrifuged at 2700 g for

2. Materials and methods 10 min and the supernatant was used for lactose and ethanol
quantification. The remaining solid was washed with distilled
2.1. Microorganism and inoculum preparation water, centrifuged and then, diluted with distilled water for
analysis of biomass. The cell concentration was determined in
Kluyveromyces fragilis (Kf1) from the culture collection of the a spectrophotometer at 600 nm, by means of a calibration
Centre of Biological Engineering, University of Minho (Portugal), curve (biomass dry weight vs. optical density (OD)) previously
was the yeast strain employed in the experiments. This strain obtained. Samples were diluted to give an absorbance in the
was supplied by University of Lavras (Department of Biology), range of 0.05e0.7.
Brazil, and was isolated from cocoa fermentation. Cells of this The lactose and ethanol concentrations in the supernatant
yeast were maintained at 4  C on YPD agar plates. The inoculum were determined by high-performance liquid chromatography,
was prepared by transferring a loopful of cells from a freshly in a Jasco chromatograph equipped with a refractive index (RI)
grown culture (incubated at 30  C for 30 h) to 500 ml Erlenmeyer detector (Jasco 830-RI) and a Chrompack (300  6.5 mm) column
flasks containing 100 ml sterile CWP solution (50 kg m
3 at 60  C, using 5 mM sulfuric acid as the eluent at a flow rate of
lactose). The flasks were incubated on a rotary shaker at 3.3 Hz, 0.5 ml min
1 and a sample volume of 20 ml.
30  C for 24 h. After this time, the cells were recovered by
centrifugation (4200 g, 15 min), washed with sterilized distilled
water, and directly resuspended in the fermentation medium. Table 1 e Experimental ranges and levels of the
independent process variables according to the 23
2.2. Medium and fermentation conditions full-factorial central composite design.
Independent variable Symbol Range and levels
Cheese whey powder (CWP) was kindly supplied by Lactogal

1 0 þ1
(Porto, Portugal). CWP composition (sample of January, 2008)
included (w w
1): >73% lactose, 12% proteins, 1.5% lipids and Initial lactose concentration X1 50 100 150
<5% moisture. To be used as fermentation medium, CWP (kg m
3)
Temperature ( C) X2 25 30 35
solutions with different initial lactose concentrations were
Inoculum concentration X3 1 2 3
prepared, pH-adjusted to 5 by addition of 1 kmol m
3 citric
(kg m
3)
acid, and deproteinized by heat treatment at 115  C for 15 min.
 b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 1 9 7 7 e1 9 8 2 1979

2.4. Experimental design and optimization by response regression coefficients was tested by t-value. Results were
surface methodology analyzed by the Experimental Design Module of the Statistica
5.0 software (Statsoft, USA). The model permitted evaluation
A 23 full-factorial central composite design with three coded of the effects of linear, quadratic and interactive terms of the
levels, leading to 17 sets of experiments was made to establish independent variables on the chosen dependent variables.
the effects of the variables (initial lactose concentration,
temperature, and inoculum concentration) on ethanol prod-
uction from CWP solution. For statistical analysis, the inde- 3. Results and discussion
pendent variables were coded according to the Eq. (1), where
each independent variable is represented by xi (coded value), The yeast strain used in the present work was a K. fragilis
Xi (real value), X0 (real value at the center point), and DXi (step selected among 8 Kluyveromyces strains (unpublished results).
change value). The range and the levels of the variables are The experimental results obtained by cultivation of this yeast
given in Table 1. The ethanol concentration, substrate in deproteinized CWP solution, under different operational
consumption, ethanol yield factor, ethanol volumetric conditions according to a 23 central composite design, are
productivity, ethanol yield per cell, and bioconversion effi- shown in Table 2. It can be noted that K. fragilis was able to
ciency were taken as dependent variables or responses of the growth and produce ethanol under all the evaluated fermen-
experimental design. tation conditions, however, the production strongly varied
according to the levels employed for the independent vari-
xi ¼ ðXi
X0 Þ=DXi (1) ables. The highest ethanol concentration (55.9 kg m
3) was
The experimental results were fitted with a second-order obtained when using an initial lactose concentration of
polynomial equation by multiple regression analysis. The 150 kg m
3, 30  C, and 2 kg m
3 inoculum concentration
quadratic mode for predicting the optimal point was exp- (conditions of run 10). Under these same conditions, the
ressed according to eq. (2), where y b represents the response ethanol yield factor and volumetric productivity also achieved
i
variable, b0 is the interception coefficient, bi, bii and bij are the the highest values (YP/S ¼ 0.37 kg kg
1; QP ¼ 1.27 kg m
3 h
1).
regression coefficients, n is the number of studied variables, Due to the large difference observed in the ethanol
and Xi and Xj represent the independent variables. Where production, a statistical analysis was carried out to identify
possible, the model was simplified by elimination of statisti- the variables that had the greatest influence on this biocon-
cally insignificant terms. version process. Table 3 shows the Student’s t-test and
p-values used to determine the statistical significance of the
X
n X
n n
1 X
X n independent variables (initial lactose concentration, temper-
b ¼ b0 þ
y bi Xi þ bii X2i þ bij Xi Xj (2)
i ature, and inoculum concentration) on the response variables
i¼1 i¼1 i¼1 j¼iþ1
(ethanol concentration, YP/S, QP, h, YP/x, and substrate cons-
The quality of the fitted polynomial model was expressed by umption). According to this analysis the initial lactose
the coefficient of determination R2, and its statistical signifi- concentration was the variable that affected all the analyzed
cance was checked by the F-test. The significance of the responses. In addition, it was the unique variable with

Table 2 e Experimental matrix and results of ethanol concentration (Et), ethanol per biomass yield factor (YP/x), and ethanol
yield factor (YP/S), with coded levels of the variables according to a 23 full-factorial central composite design.
Runs Independent variablesa Responses

X1 X2 X3 Cell S cons. Et YP/x YP/S QP h

(kg m
3) (%) (kg m
3) (kg kg
1) (kg kg
1) (kg m
3 h
1) (%)

1
1
1
1 4.9 100 12.7 2.59 0.25 0.29 45.67

2
1
1 þ1 6.8 100 13.4 1.97 0.26 0.30 48.83
3
1 þ1
1 5.5 100 12.0 2.18 0.27 0.27 54.86
4
1 þ1 þ1 7.2 100 13.4 1.86 0.35 0.30 66.07
5 þ1
1
1 7.4 94.6 48.2 6.51 0.34 1.10 60.13
6 þ1
1 þ1 9.1 95.3 47.0 5.16 0.36 1.07 64.29
7 þ1 þ1
1 7.4 100 48.5 6.55 0.35 1.10 64.76
8 þ1 þ1 þ1 7.5 100 41.5 5.53 0.28 0.94 52.44
9
1 0 0 6.6 100 10.3 1.56 0.22 0.23 40.65
10 þ1 0 0 8.7 100 55.9 6.43 0.37 1.27 69.04
11 0
1 0 7.8 100 28.8 3.69 0.33 0.66 62.03
12 0 þ1 0 7.3 100 24.5 3.36 0.28 0.56 51.81
13 0 0
1 7.5 100 32.4 4.32 0.33 0.74 60.58
14 0 0 þ1 9.0 100 28.8 3.20 0.29 0.66 53.26
15 0 0 0 8.4 100 26.1 3.11 0.30 0.59 55.64
16 0 0 0 8.3 100 25.6 3.08 0.30 0.58 55.00
17 0 0 0 8.3 100 23.8 2.87 0.29 0.54 53.88

a X1 ¼ coded values of lactose; X2 ¼ coded values of temperature; X3 ¼ coded values of inoculum concentration.
1980 b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 1 9 7 7 e1 9 8 2

Table 3 e Effect estimates, standard errors and ethanol concentration (Et), substrate consumption (Sc), bioconversion
efficiency (h), ethanol yield per cell (YP/x), ethanol yield factor (YP/S), and ethanol volumetric productivity (QP) during the
bioconversion of cheese whey power solution by Kluyveromyces fragilis, according to the 23 full-factorial central composite
design.
Estimated Standard tvalue Estimated Standard tvalue Estimated Standard tvalue
effects errors effects errors effects errors

Variables and interactions Et Sc h

X1 35.860 2.498 14.357a
2.016 0.570
3.538a 10.916 4.724 2.310b
X21 7.641 4.826 1.583
0.994 1.101
0.903
1.468 9.127
0.161
X2
2.040 2.498
0.817 2.016 0.570 3.538a 1.798 4.724 0.381
X22
5.259 4.826
1.090
0.994 1.101
0.903 2.682 9.127 0.294
X3
1.940 2.498
0.777 0.132 0.570 0.232
0.222 4.724
0.047
X23 2.641 4.826 0.547
0.994 1.101
0.903 2.682 9.127 0.294
X 1X 2
1.125 2.793
0.403 2.520 0.637 3.956a
8.413 5.282
1.593
X1X3
2.575 2.793
0.922 0.165 0.637 0.259
5.633 5.282
1.066
X 2X 3
1.275 2.793
0.457
0.165 0.637
0.259
2.108 5.282
0.399

Variables and interactions YP/x YP/S QP

X1 4.004 0.246 16.288a 0.070 0.024 2.968b 0.818 0.058 14.182a
X21 1.006 0.475 2.119
0.011 0.046
0.232 0.164 0.111 1.470
X2
0.088 0.246
0.358
0.002 0.024
0.085
0.050 0.058
0.867
X22 0.066 0.475 0.140 0.009 0.046 0.207
0.116 0.111
1.043
X3
0.886 0.246
3.604a 0.000 0.024 0.000
0.046 0.058
0.798
X23 0.536 0.475 1.130 0.019 0.046 0.427 0.064 0.111 0.573
X 1X 2 0.233 0.275 0.846
0.045 0.026
1.707
0.028 0.064
0.426
X 1X 3
0.358 0.275
1.301
0.035 0.026
1.327
0.058 0.064
0.892
X 2X 3 0.158 0.275 0.573
0.005 0.026
0.190
0.028 0.064
0.426

a p < 0.01.
b p < 0.05; X1 ¼ coded values of initial lactose concentration; X2 ¼ coded values of temperature; X3 ¼ coded values of inoculum concentration.

significant influence on ethanol production, YP/S, QP, and h 1 kg m
3, an average increase of 3.604 kg kg
1 was observed in
values. For all of these responses, only the linear effect of the YP/x.
initial lactose concentration was significant at 95% confidence Regarding the substrate consumption, although the lactose
level. Such effect had a positive signal, indicating that the was completely consumed in almost all the fermentations, the
ethanol concentration, YP/S, QP, and h values increased by statistical analysis pointed out that this response was influ-
increasing the initial lactose concentration. Temperature and enced by the initial lactose concentration and temperature
inoculum concentration did not present main significant (Table 3). The initial lactose concentration had a main and
effect for these responses, suggesting that temperature negative effect, indicating that the substrate consumption
between 25 and 35  C and inoculum concentration varying increased as the initial substrate concentration used in the
from 1 to 3 kg m
3 did not affect the YP/S, QP, h, and ethanol experiments decreased. On the other hand, the temperature
production by K. fragilis from CWP solution. Interaction effects had a main and positive effect, revealing that the temperature
among the studied variables were also not significant at 95% increase favored the substrate consumption by the microor-
confidence level. ganism. Increasing the fermentation temperature from 23  C to
The ethanol yield per cell (YP/x) was also mainly affected by 42  C also enhanced the lactose utilization by Lactobacillus hel-
the initial lactose concentration, resulting in higher ethanol veticus [12]. According to the authors, the rate of reaction for
production by cell as the initial substrate concentration was microorganisms really increases with increasing temperature,
increased (positive effect) (Table 3). Moreover, the inoculum until a limiting maximum value is reached.
concentration had also a significant effect at 95% confidence After identification of the main variables affecting the ethanol
level, which had a negative signal, indicating that the lower production, a multiple regression analysis was performed to fit
the inoculum concentration, the higher the ethanol amount the experimental data to polynomial equations, obtaining the
produced by cell. In addition, the statistical analysis carried coefficients given in Table 4. The models were simplified by
out for substrate consumption revealed that the inoculum elimination of statistically insignificant terms. The quality of the
variation from 1 to 3 kg m
3 did not have influence in the fitted polynomial models was expressed by the coefficient of
lactose consumption by yeast. Inoculum variation from 1 to determination R2. As can be observed, models explaining more
3 kg m
3 was also not significant for ethanol production. By than 90% of the variations observed in the responses (R2 > 0.9)
considering these facts it can be concluded that when using could be adjusted for the responses ethanol concentration, QP,
1 kg m
3 inoculum, the cells were able to consume the same and YP/x. The high R2 values mean that the models accurately
lactose amount and produce the same final ethanol concen- represent the data in the experimental region studied, explaining
tration that when using 3 kg m
3 inoculum. Consequently, the more than 90% of the variability in the responses.
product formation by cell was higher. As can be seen in Table The relation between variables and ethanol concentration
3, when the inoculum concentration was decreased from 3 to can be best visualized by examining the surface plots given in
 b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 1 9 7 7 e1 9 8 2 1981

obtained for the QP response (not shown). A comparable

Table 4 e Model equations for the response surfaces fitted
to the experimental data points, and the respective R2. behavior was also verified for Kluyveromyces marxianus DSMZ-
7239 yeast using initial lactose concentrations up to 75 kg m
3
Response Model equations R2
[13]. However, maxima ethanol amounts produced in the
Ethanol concentration Et ¼ 28.99 þ 17.93X1 0.94 present study were higher than those obtained by direct
(Et, in kg m
3) fermentation of crude (non-concentrated) cheese whey [14] or
Ethanol volumetric productivity QP ¼ 0.659 þ 0.409X1 0.94
cheese whey powder [13].
(QP, in kg m
3 h
1)
Based on the statistical analysis results, assays were per-
Ethanol yield by cell YP/x ¼ 3.376 þ 2.002X1 0.96
(YP/x, in kg kg
1) þ 0.658X21
0.443X3 formed in a following step to evaluate the possibility of
increasing the ethanol concentration by increasing the initial
X1 ¼ initial lactose concentration; X3 ¼ inoculum concentration.
lactose concentration to values above 150 kg m
3 (up to
250 kg m
3). The values of temperature and inoculum used in
these experiments were fixed at 35  C and 1 kg m
3, respectively.
Fig. 1, which were plotted as a function of two variables at
This temperature was chosen because the solubility of lactose
a time and holding the other variable at a fixed level. Fig. 1
solutions increase with the temperature increase and this is
clearly shows that increasing initial lactose concentration
important as higher the concentration of the cheese whey
resulted in higher ethanol production, with maxima values
solution used. On the other hand, the inoculum concentration
(41.5 kg m
3) being achieved under the maximum tested
was fixed at 1 kg m
3 due to economical and practical reasons.
concentration (150 kg m
3). Similar plot surfaces were
Fig. 2 shows the experimental results obtained in these
experiments. As can be seen, initial lactose concentrations up to
A 200 kg m
3 favored the bioconversion to ethanol, but higher
lactose concentration values drastically affected all the
65
fermentative parameters. Cell growth was slightly increased by
55 increasing the lactose concentration between 150 and
200 kg m
3 (Fig. 2A). In this same range of values, the substrate
45
ETHANOL (kg m )

35
A
25

15
-3

1.0

0.5
TE

0.0
M

1.0
PE

0.5
RA

-0.5
0.0
TU

-0.5 E
TO S
RE

-1.0 -1.0 LAC

B
65
B
55

45
ETHANOL (kg m )

35

25

15
-3

1.0

0.5
INO

0.0
1.0
CU

0.5
LU

-0.5 0.0
M

-0.5 E
-1.0 -1.0 TO S
LAC
Fig. 2 e Effect of the initial lactose concentration on (A) cell
Fig. 1 e Response surface of ethanol production from growth, lactose consumption, ethanol production, and (B)
cheese whey powder by K. fragilis as a function of: (A) in the fermentative parameters for ethanol production by
initial lactose concentration and temperature, (B) initial K. fragilis in cheese whey powder solution. Fermentation
lactose and inoculum concentrations. temperature: 35  C; inoculum concentration: 1 kg mL3.
1982 b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 1 9 7 7 e1 9 8 2

consumption by the microorganism was not affected, and the references

ethanol production increased with the initial lactose concen-
tration increase, achieving a maximum value of 80.95 kg m
3
(77.4% of the theoretical value) when a deproteinized CWP [1] Ghaly AE, El-Taweel AA. Kinetic modelling of continuous
solution containing 200 kg m
3 lactose was fermented. Lactose production of ethanol from cheese whey. Biomass Bioenerg
concentration values higher than 200 kg m
3 affected the 1997;6:461e72.
ethanol production by the yeast, as can be seen in the profiles [2] González Siso MI. The biotechnological utilization of cheese
whey: a review. Bioresour Technol 1996;57:1e11.
given in Fig. 2A and B. The observed ethanol concentration
[3] Panesar PS, Kennedy JF, Gandhi DN, Bunko K. Bioutilisation
reduction by elevated initial sugar concentration is probably of whey for lactic acid production. Food Chem 2007;105:1e14.
related with a substrate inhibition, which might have inacti- [4] Dragone G, Mussatto SI, Oliveira JM, Teixeira JA.
vated the cells due to high osmotic pressure encountered at high Characterisation of volatile compounds in an alcoholic
sugar content, causing high maintenance requirements [8]. beverage produced by whey fermentation. Food Chem 2009;
The results here attained can be favorably compared with 112:929e35.
[5] Kim S, Dale BE. Global potential bioethanol production from
others reported in the literature. For example, fermentation of
wasted crops and crop residues. Biomass Bioenerg 2004;26:
whey by K. marxianus MTCC 1288 under different initial lactose
361e75.
concentrations yielded maximum ethanol production [6] Staniszewski M, Kujawski W, Lewandowska M. Semi-
(3.98 kg m
3) when using 50 kg m
3 lactose. Higher lactose continuous ethanol production in bioreactor from whey with
concentrations led to a drastic decrease in product formation co-immobilized enzyme and yeast cells followed by
and substrate utilization [15]. Lactose concentrations higher pervaporative recovery of product e Kinetic model
than 100 kg m
3 had an inhibitory effect on the specific growth predictions considering glucose repression. J Food Eng 2009;
91:240e9.
rate, lactose utilization rate, and ethanol production rate by
[7] Guadix A, Sorenson E, Papageorgiou LG, Guadix EM. Optimal
Candida pseudotropicalis [16]. Ethanol bioconversion by a rec- design and operation of ultrafiltration plants. J Membr Sci
ombinant Saccharomyces cerevisiae was also affected by initial 2004;235:131e8.
lactose concentrations higher than 100 kg m
3. The highest [8] Ozmihci S, Kargi F. Fermentation of cheese whey powder
value (59%) decreased to 53% when using initial lactose solution to ethanol in a packed-column bioreactor: effects of
concentrations higher than 100 kg m
3 [17]. Ethanol production feed sugar concentration. J Chem Technol Biotechnol 2009;
by K. fragilis (present work) was only affected by lactose 84:106e11.
[9] Kargi F, Ozmihci S. Utilization of cheese whey powder (CWP)
concentrations higher than 200 kg m
3. This finding represents
for ethanol fermentations: effects of operating parameters.
an advantage because the costs of the process can be signifi- Enzym Microb Technol 2006;38:711e8.
cantly reduced with the lactose concentration increase [2]. [10] Ozmihci S, Kargi F. Continuous ethanol fermentation of
cheese whey powder solution: effects of hydraulic residence
time. Bioproc Biosyst Eng 2007;30:79e86.
4. Conclusions [11] Mawson AJ. Bioconversions for whey utilization and waste
abatement. Bioresour Technol 1994;47:195e203.
[12] Tango MSA, Ghaly AE. Effect of temperature on lactic acid
The initial lactose concentration in deproteinized cheese
production from cheese whey using Lactobacillus helveticus
whey powder solution exerted great influence on ethanol under batch conditions. Biomass Bioenerg 1999;16:61e78.
production by K. fragilis, being the maximum product forma- [13] Ozmihci S, Kargi F. Kinetics of batch ethanol fermentation of
tion (80.95 kg m
3) obtained when using an initial lactose cheese-whey powder (CWP) solution as function of substrate
concentration of 200 kg m
3. This value is about 4 times higher and yeast concentrations. Bioresour Technol 2007;98:
than that achieved during the fermentation of non-concen- 2978e84.
[14] Zafar S, Owais M. Ethanol production from crude whey by
trated cheese whey, and represents thus an interesting
Kluyveromyces marxianus. Biochem Eng J 2006;27:295e8.
alternative to decrease the distillation costs for ethanol
[15] Zafar S, Owais M, Saleemuddin M, Husain S. Batch kinetics
production from cheese whey. and modelling of ethanolic fermentation of whey. Int J Food
Sci Tech 2005;40:597e604.
[16] Ghaly AE, El-Taweel AA. Effect of lactose concentration on
batch production of ethanol from cheese whey using Candida
Acknowledgments
pseudotropicalis. Trans ASAE 1995;38:1113e20.
[17] Guimarães P, Klein J, Domingues L, Teixeira JA. Fermentation
The authors acknowledge the financial support from "CAPES/ performance of a recombinant lactose-consuming
Grices (BEX2150/07-7) and Lactogal for supplying cheese whey flocculating Saccharomyces cerevisiae strain. Braz J Food
powder. Technol 2005;3:34e9.