World Journal of Microbiology & Biotechnology 19: 523526, 2003.

2003 Kluwer Academic Publishers. Printed in the Netherlands.

523

Optimization of fermentation conditions for the production of ethanol from sago

starch using response surface methodology
B.V.V. Ratnam*, M. Narasimha Rao, M. Damodar Rao
, S. Subba Rao and C. Ayyanna
Centre for Biotechnology, Department of Chemical Engineering, College of Engineering, Andhra University,
Visakhapatnam 530 003, India
*Author for correspondence: E-mails: bvvratnams@yahoo.com, ratnam72@redimail.com


Present address: Proteonomics lab, Centre for Disease Control and Prevention 4770, Buford Highway, f 25, Atlanta,
GA, USA
Received 11 October 2002; accepted 6 February 2003

Keywords: Central composite design, ethanol, glucoamylase, response surface methodology, sago starch,
simultaneous saccharication and fermentation (SSF), Zymomonas mobilis

Summary
The quantitative eects of temperature, pH and time of fermentation were investigated on simultaneous
saccharication and fermentation (SSF) of ethanol from sago starch with glucoamylase (AMG) and Zymomonas
mobilis ZM4 using a BoxWilson central composite design protocol. The SSF process was studied using free enzyme
and free cells and it was found that with sago starch, maximum ethanol concentration of 70.68 g/l was obtained
using a starch concentration of 140 g/l, which represents an ethanol yield of 97.08%. The optimum conditions for
the above yield were found to be a temperature of 36.74 C, pH of 5.02 and time of fermentation of 17 h. Thus by
using the central composite design, it is possible to determine the accurate values of the fermentation parameters
where maximum production of ethanol occurs.

Introduction
The production of fuel ethanol by fermentation requires
the ability to produce high ethanol concentrations
rapidly while maintaining good yields. Rapid fermentation and high alcohol levels are desirable to minimize
capital costs and energy required for distillation, while
good yields are necessary for process economics. The
substrate is the main cost component for industrial
ethanol production and it is essential that ethanol
production should be carried out with cheap substrates
such as starch or cellulose (Lee & Woodward 1983;
Elisson et al. 2001). Sago starch is an agricultural
material abundantly produced in India and other tropical
countries (Tan 1980). The starch is a product extracted
from the seeds or tubers, stems of palms and cycads such
as Metroxylon sagu and is a mixture of 27% (w/w)
amylose and 73% (w/w) amylopectin (Swinkels 1985).
Zymomonas mobilis could be grown in a medium
containing exogenously added glucoamylase (AMG) in
order to allow simultaneous starch saccharication and
ethanol fermentation (SSF) to occur. The SSF process
combines enzymatic hydrolysis of starch to glucose and
ethanol fermentation into a single operation. Consequently, this process oers a great potential of increased
rate of hydrolysis, reduction of fermentation time and
decreased capital cost (Kosaric et al. 1983).

The traditional one-factor at a time technique used

for optimizing a multivariable system is not only timeconsuming but also often easily misses the alternative
eects between components. Recently many statistical
experimental design methods have been employed in
bioprocess optimization. Among them, response surface methodology (RSM) is the one suitable for
identifying the eect of individual variables and for
seeking the optimum conditions for a multivariable
system eciently. This method has been successfully
applied to optimize alcoholic fermentation (Chen 1981;
Bowman & Geiger 1984; Zertuche & Zall 1985;
Ratnam 2001), optimize vegetable oil bioconversion
(Cheynier et al. 1983) and biomass production (Moresi
et al. 1980). Enzyme production was also eectively
promoted by optimizing the medium composition
using RSM (Maddox & Reichert 1977; Galas et al.
1981). A detailed account of this technique has been
outlined (Cochran & Cox 1968). Basically, this optimization process involves three major steps: performing the statistically designed experiments, estimating
the coecients in a mathematical model and predicting the response and checking the adequacy of the
model.
Hence, the authors report the application of the RSM
using the BoxWilson design (Box & Wilson 1951) of
experiments to develop a mathematical correlation

524

B.V.V. Ratnam et al.

Zymomonas mobilis ZM4 obtained from National

Chemical Laboratory, Pune, India, was used throughout
this study.

per unit volume of enzyme (ml) at 60 C. Ethanol was

estimated by GLC in which a ame ionization detector
and stainless steel column (2.0 m length, 3.0 mm i.d.)
packed with Porapak-Q (5080 mesh, manufactured by
Nucon Engineers, India) were used. The column oven
was operated isothermally at 150 C and the detector
and injection ports were kept at 170 C. Nitrogen was
used as carrier gas at a ow rate of 30 cm3/min and the
combustion gas was a mixture of hydrogen and air.
Ethanol was also estimated by using a chemical oxidation method (Caputi & Wright 1969) for comparative
purposes. Quantication of ethanol by both GC analysis
and the chemical method were found to be in good
agreement, indicating the accuracy of the present results.
All experiments were conducted in duplicate and average values were reported.

Enzymes

Experimental design and optimization

a-amylase from Bacillus licheniformis and AMG from

Aspergillus niger were obtained from Sigma Chemical
Co. (St Louis, Mo). The activities of the two enzymes
were 60 KNU/g and 200 AGU/ml, respectively.

Central composite experimental design (CCD, Box &

Wilson 1951) was used in the optimization of ethanol
production. Temperature (X1, C), pH (X2) and time of
fermentation (X3, h) were chosen for the independent
variables and were shown in Tables 1 and 2. Ethanol
concentration (Yi, g/l) was used as the dependent output
variable. For statistical calculations the variables Xi
were coded as xi according to Equation (1)

between the temperature, pH and time of fermentation

and yield of ethanol.

Materials and methods

Substrate
Sago starch was collected from cultivators, East Godavari District, Andhra Pradesh, India.
Organism

Growth medium and growth conditions

The culture was maintained on agar slants having
composition (g/l): glucose-100, yeast extract-10,
KH2PO4-1, (NH4)2SO4-1 and MgSO47H2O-0.5.
Fermentation conditions
Fermentation media were composed of 100250 g sago
starch and 10 g yeast extract in 1 l tap water. The pH of
the solution was adjusted to 6.5 by the addition of the
required volume of 1 M NaOH. Liquefaction was
carried out by adding 0.2% (v/w) a-amylase to the
slurry at pH 6.5 and heating at 95 C for 1 h. No
problem was faced during the solubilization of starch
because of the reduction in viscosity of the fermentation
mashes by enzymes.
SSF using free enzyme was carried out in a Biostat M
fermentor supplied by B. Braun Co., Germany with all
necessary controls. The reactor was of 2 l capacity and
the working volume was 1 l. The operating conditions
were maintained at a temperature of 30 C, pH 5.0 and
stirrer speed 200 rev/min. The inoculum size and the
AMG dosage were 5% (v/v) and 0.5 (v/w) respectively.
The reactor was maintained under anaerobic conditions
with no aeration.
Analytical techniques
Reducing sugars were estimated by the dinitrosalicylic
acid (DNS) method (Miller 1959). AMG activity was
determined by allowing the enzyme preparation to react
with 10% partially hydrolysed starch in 0.1 M acetate
buer, pH 4.4 at 60 C. One unit of AMG activity was
dened as the production of 1 g of free sugar per hour

Table 1. Independent variables in the experimental plan.

Variables

Coded levels

Temperature, X1 (C)
pH, X2
Time, X3 (h)

)1.682

)1

1.682

26.59
3.318
7.9

30
4
12

35
5
18

40
6
24

43.41
6.682
28.092

Table 2. The CCD matrix employed for three independent variables

(actual values given in Table 1).
Run no.

X1

X2

X3

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

)1
1
)1
1
0
0
)1
1
)1
1
0
0
)1.682
1.682
0
0
0
0
0
0

)1
)1
1
1
0
0
)1
)1
1
1
0
0
0
0
)1.682
1.682
0
0
0
0

)1
1
1
)1
0
0
1
)1
)1
1
0
0
0
0
0
0
)1.682
1.682
0
0

525
xi X i

xi =Dxj i 1; 2; 3; . . . ; k;

Table 3. Experimental and the predicted yields for ethanol.

X2

X1

where xi is dimensionless value of an independent

variable, Xi is real value of an independent variable, xi
is real value of the independent variable at the centre
point and Dxj is step change.
p
A 23-factorial CCD, with six axial points (a 3)
and six replications at the centre points (n0 6) leading
to a total number of 20 experiments was employed
(Table 2) for the optimization of the conditions of
fermentation. The second degree polynomials (Equation
(2)) were calculated with the statistical package (StatElse Inc, Minneapolis, MN, USA) to estimate the
response of the dependent variable.
Y i b0 b1 X 1 b2 X 2 b3 X 3 b11 X 21 b22 X 22
b33 X 23 b12 X 1 X 2 b23 X 2 X 3 b13 X 1 X 3 ;

where Yi is predicted response, X1, X2, X3 are independent variables, b0 is oset term, b1, b2, b3 are linear
eects, b11, b22, b33 are squared eects and b12, b23, b13
are interaction terms.

Results and discussion

The most important physical factors which aect the
fermentative production of ethanol are the temperature,
initial pH and time of fermentation. The suitable levels
for these parameters were also determined using statistical CCD. The experimental design matrix is given in
Tables 1 and 2. Twenty experiments were performed
using dierent combinations of the variables as per the
CCD. Using the results of the experiments the following
second order polynomial equation giving the ethanol
concentration as a function of temperature (X1, C), pH
(X2) and time of fermentation (X3, h) was obtained.

30
40
30
40
35
35
30
40
30
40
35
35
26.59
43.41
35
35
35
35
35
35

4
4
6
6
5
5
4
4
6
6
5
5
5
5
3.318
6.682
5
5
5
5

Ethanol yield (g/l)

X3

12
24
24
18
18
24
12
12
24
18
18
18
18
18
18
18
7.9008
28.092
18
18

Experimental

Predicted

9.25
41.51
22.73
44.81
66.34
65.97
10.46
50.42
27.29
48.73
66.25
66.73
4.82
44.61
60.11
63.33
43.24
65.49
66.49
66.82

9.855
50.23
30.34
55.49
65.96
64.07
9.855
48.48
30.34
55.49
65.96
65.96
2.685
34.41
60.28
50.82
46.26
56.97
65.96
65.96

also indicates that only 1% of the variation is not

explained by the model. The value of R is 0.96. The
corresponding analysis of variance (ANOVA) is presented in Table 4. v2 test shows that the model is a good t
since v2cal < v2tab , where v2cal is 16.11 and v2tab is 30.14. The
predicted optimum levels of temperature, initial pH and
time of fermentation were obtained by applying the

Yi
1278:79 53:39893X1 43:32943X2

30:13357X3
0:67033X12
3:67826X22

0:14073X32 0:669338X1 X2
1:82141X2 X3

0:44089X1 X3
3
The predicted production of ethanol using the above
equation is given in Table 3 along with the experimental
value. The coecient of determination, R2 is 0.92, implies
that the sample variation of 92.35% for ethanol production is attributed to the independent variables, viz.,
temperature, pH and fermentation time. The R2 value

Figure 1. Response surface and contour plot of temperature vs. pH on

ethanol production (time was kept constant at 18 h).

Table 4. ANOVA for full quadratic model.

Source of variation

Sum of squares (SS)

Degrees of
freedom (DF)

Mean squares (MS)

F-value

Probe > F

Regression
Residual
Total

7972.435
659.732
8632.166

9
10
19

885.83
65.973

13.427

526

B.V.V. Ratnam et al.

and fermentation with AMG and Zymomonas mobilis
ZM4.

References

Figure 2. Response surface and contour plot of temperature vs. time

on ethanol production (pH was kept constant at 5).

Figure 3. Response surface and contour plot of pH vs. time on ethanol

production (temperature was kept constant at 35 C).

regression analysis to the Equation (3). The predicted and

experimental ethanol production at the optimum levels of
fermentation conditions were also determined. Figures
13 represent the isoresponse contour and surface plots
for the optimization of fermentation conditions of
ethanol production. The maximum ethanol concentration of 70.68 g/l appeared at temperature, pH and time of
fermentation of 36.74 C, 5.02 and 17.01 h, respectively.
Thus the present study using the technique of central
composite design enables us to nd the accurate values
of the fermentation parameters for the maximum
product of ethanol using simultaneous saccharication

Bowman, L. & Geiger, E. 1984 Optimization of fermentation

conditions for alcohol production. Biotechnology and Bioengineering 26, 14921497.
Box, G.E.P. & Wilson, K.B. 1951 On the experimental attainment of
optimum conditions. Journal of the Royal Statistical Society
(Series B) 13, 145.
Caputi, A. & Wright, D. 1969 Collaborative study of the determination of ethanol in wine by chemical oxidation. Journal of AOAC
International 52, 8588.
Chen, S.L. 1981 Optimization of batch alcohol fermentation of glucose
syrup substrate. Biotechnology and Bioengineering 23, 18271836.
Cheynier, V., Feinberg, M., Chararas, C. & Ducauze, C. 1983
Application of response surface methodology to evaluation of
bioconversion experimental conditions. Applied and Environmental
Microbiology 45, 634639.
Cochran, N.G. & Cox, G.M. 1968 Experimental Designs. John Wiley
and Sons Inc., p. 611.
Elisson, A., Hofmeyr, J.H.S., Pedler, S. & Hahn-Hagerdal, B. 2001
The xylose reductase/xylitol dehydrogenase/xylulokinase ratio
aects product formation in recombinant xylose-utilizing Saccharomyces cerevisiae. Enzyme and Microbial Technology 29, 288297.
Galas, E., Bieleeki, S., Antezak, T., Weiczorek, A. & Blaszezyk, R.
1981 Optimization of cultivation medium conditions for lytic
enzyme biosynthesis. In Advances in Biotechnology. Proceedings of
the 6th International Fermentation Symposium, Vol. 3. eds. MooYoung, M., Vezina, C., Singh, K. pp. 301306. Canada: Pergamon
Press. ISBN 0-08025385-7.
Kosaric, N., Wieczorek, A., Cosentino, G. & Magee, R. 1983 In
Biotechnology, vol. 3. ed. Dellweg, H. p. 315. (Weinheim, Verlag
Chemie) ISBN 3-52725765-9.
Lee, J.M. & Woodward, J. 1983 Properties and application of
immobilized b-D -glucoside coentrapped with Zymomonas mobilis
in calcium alginate. Biotechnology and Bioengineering 25, 2441.
Maddox, I.S. & Reichert, S.H. 1977 Use of response surface
methodology for the rapid optimization of microbiological media.
Journal of Applied Bacteriology 43, 197204.
Miller, G.L. 1959 Use of DNS reagent for determination of reducing
sugars. Analytical Chemistry 31, 426428.
Moresi, M., Colicchio, A. & Sansovini, F. 1980 Optimization of whey
fermentation in a jar fermenter. European Journal of Applied
Microbiology and Biotechnology 9, 173183.
Ratnam, B.V.V. 2001 Studies on physico-chemical and nutritional
parameters for the production of ethanol from palmyra jaggery by
submerged fermentation using Saccharomyces cerevisiae. PhD
Thesis, Andhra University, Visakhapatnam, AP, India.
Tan, K. 1980 Production of sago starch. In Proceedings of the Second
International Sago Symposium. eds. Stanton, W.R. & Flach, M.
p. 56. (The Hague: Martinus Nijho) ISBN 90-2472470-8.
Swinkels, J.J.M. 1985 In Starch Conversion Technology, eds. van
Beynum, G.M.A., Roels, J.A. p. 23. New York: Marcel Dekker.
ISBN 0-82477194-X.
Zertuche, L. & Zall, R.R. 1985 Optimizing alcohol production from
whey using computer technology. Biotechnology and Bioengineering 27, 547554.