Bioresource Technology 98 (2007) 2824–2828

Fermentation of molasses by Zymomonas mobilis: Effects of

temperature and sugar concentration on ethanol production
a,*
M.L. Cazetta , M.A.P.C. Celligoi b, J.B. Buzato b, I.S. Scarmino c

a
Department of Food and Drugs, Londrina State University, Post Box 6001, CEP: 86051-990, Londrina-PR, Brazil
b
Department of Biochemistry and Biotechnology, Londrina State University, Post Box 6001, CEP: 86051-990, Londrina-PR, Brazil
c
Department of Chemistry, Londrina State University, Post Box 6001, CEP: 86051-990, Londrina-PR, Brazil

Received 12 July 2004; received in revised form 9 August 2006; accepted 10 August 2006
Available online 8 April 2007

Abstract

Fermentations utilizing strains of Zymomonas mobilis, in place of the traditional yeasts, have been proposed due their ethanol yields
being close to theoretical. Ethanol production from sugar cane molasses was analyzed under different culture conditions using Z. mobilis
in batch fermentation. The total reducing sugars (TRS) concentrations in the molasses, temperature, agitation and culture time effects
were studied simultaneously through factorial design. The best conditions for ethanol production were 200 g L1 of total reducing sugars
in the molasses, temperature of 30 C and static culture and time of fermentation of 48 h, achieving 55.8 g L1. The pH of the medium
was kept constant during the experiments, showing that molasses presents a buffering effect.
 2006 Elsevier Ltd. All rights reserved.

Keywords: Ethanol; Zymomonas mobilis; Sugar cane molasses; Factorial design

1. Introduction sources is important for the biofuel ethanol production

The depletion of fossil fuel reserves, the unstable pano-
rama of the petrol prices and more recently, increasing
environmental and political pressures (Davis et al., 2005)
has increased industrial focus toward alternative fuel
sources, and encouraged the search of products originated
from biomass, as renewable sources of energy.
In this context, fermentative processes stand out, where
microbial metabolism is used for the transformation of
simple raw materials in products with high aggregate value.
Among these, ethanol is one of the best examples of how
fermentation can match market needs satisfactorily. Even
though the fermentative process for ethanol production is
well known, the production costs are still the key impedi-
ment wide use of ethanol as fuel. Therefore, the develop-
ment of a fermentation process using economical carbon
*
Corresponding author. Tel.: +55 43 3371 4270; fax: +55 43 3371 4216.
E-mail addresses: macelligoi@uel.br (M.L. Cazetta), malulz@yahoo.
com.br (M.A.P.C. Celligoi).
on a commercial scale (Tanaka et al., 1999; Tao et al.,
2005). Many studies have been done that focus on produc-
tion improvement and decreasing its costs (Sreenath and
Jeffries, 2000; Davis et al., 2005; Ruanglek et al., 2006;
Mohagheghi et al., 2006).
Zymomonas mobilis, a Gram-negative bacterium, have
been attracting increasing attention for fuel ethanol. It is
an osmo- and ethanol-tolerant bacterium and it has shown
higher specific rates of glucose uptake and ethanol produc-
tion (Rogers et al., 1982, 1997) via the Entner-Doudoroff
pathway under anaerobic conditions. Z. mobilis may have
a greater potential for industrial ethanol production from
raw sugar, sugarcane juice and sugarcane syrup (Lee and
Huang, 2000).
Molasses is an agro-industrial by-product often used in
alcohol distilleries (Jiménez et al., 2004) due to the presence
of fermentative sugars, being an optimal carbon source for
the microorganism metabolism. Sugar cane molasses is an
abundant agro-industrial material produced in Brazil and
other tropical countries and its low cost is an important

0960-8524/$ - see front matter  2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2006.08.026
 M.L. Cazetta et al. / Bioresource Technology 98 (2007) 2824–2828 2825

factor for the economical viability of substances produc- Table 1

tion by fermentation. 24 Factorial experimental design investigating the effect of TRS concen-
tration in molasses, temperature, agitation rate and culture time to ethanol
The traditional one-at-a-time optimization strategy is production by Z. mobilis ATCC 29191
relatively simple, and the individual effects of medium fac-
Run Variables in coded Measured responses
tors can be graphically depicted without the need of the levels
statistical analysis. Unfortunately, it frequently fails to
X1 X2 X3 X4 Ethanol (g L1) Biomass (g L1)
locate the region of optimum response in such procedures.
In this case, fractional and/or full factorial design provides 1     5.74 0.89
2 +    5.38 0.81
an efficient approach to optimization. A combination of 3  +   7.95 0.75
factors generating a certain optimum response can be iden- 4 + +   6.98 0.65
tified though factorial design and the use of response sur- 5   +  2.97 0.56
face methodology (RSM) (Box et al., 1978). 6 +  +  9.89 0.31
The response-surface methodology is an empirical mod- 7  + +  4.09 0.53
8 + + +  1.98 0.41
eling system that assesses the relationship between a group 9    + 19.33 1.07
of variables that can be controlled experimentally and the 10 +   + 7.31 1.22
observed response (Sreekumar et al., 1999; Hamsaveni 11  +  + 22.69 1.29
et al., 2001). Response surface methodology (RSM) is a 12 + +  + 30.08 1.41
useful model to study the effect of several factors influenc- 13   + + 2.94 0.81
14 +  + + 12.20 1.07
ing the responses by varying them simultaneously and car- 15  + + + 2.23 0.65
rying out a limited number of experiments (Hamsaveni 16 + + + + 15.91 0.43
et al., 2001). The aim of this work was to study the influ-
Factors Real levels
ence between four factors and their interaction to optimize
1 +1
the ethanol production by Z. mobilis ATCC 29191 in sugar
1
cane molasses using factorial design and analysis by RSM. X1 Molasses (g L ) 150 250
X2 Temperature (C) 25 35
The selected factors were sugar concentration on molasses,
X3 Agitation rate (rpm) 0 180
temperature, agitation rate and culture time. The measured X4 Time of cultivation (h) 12 24
responses were ethanol and biomass.

2. Methods a flux of 40 ml min1 and isopropanol as an internal

standard.
2.1. Microorganism and culture conditions

The strain used was Z. mobilis ATCC 29191. The strain
was maintained on agar plates containing (per liter): 200 g
glucose, 10 g yeast extract, 5 g peptone, 1 g (NH4)2SO4, 2 g
KH2PO4, 0.5 g MgSO4 Æ 7H2O and 0.5 g FeSO4 (Merck).
The culture medium was sterilized at 121 C for 15 min.
The cultures were maintained at 4 C and renewed every
five weeks.
The inoculum culture was grown composed with sucrose
at 200 g L1 and the components mentioned previously.
The cell concentration was standardized to 0.2 g L1,
determined by turbidimetry at k = 605 nm. The batch fer-
mentations were carried out in duplicate in the sugar cane
molasses culture medium, in the different culture condi-
tions, according to the experimental design (Table 1).
2.2. Analytical methods
After each fermentation, the culture was centrifuged
(10,0000 rpm for 15 min) and the biomass concentration
was determined by measuring the turbidity of diluted sam-
ple at 605 nm using a standard curve of absorbance against
dry cell mass. The total reducing sugars (TRS) were quan-
tified according to Somogy (1945) and Nelson (1944).
Ethanol was determined by Gas Chromatography (GC)
Shimadzu, using a DBWAX column (30.0 · 0.25 cm) with
2.3. Experimental design
The conditions to optimize Z. mobilis ethanol produc-
tion by controlling fermentation variables were performed
using a factorial design and analysis of the results by
response surface methodology (Box et al., 1978; Barros
et al., 1995). As a preliminary step for optimization, the
most important factors were screened by applying the full
24 factorial design. The main effects for each of the factors
studied were defined by the Eq. (1):
Efi ¼ ðy þ Þi  ðy  Þi ð1Þ
where Efi is the effect of the ith factor on the ethanol pro-
duction, and ðy þ Þi and ðy  Þi are the average ethanol pro-
ductions values at the high (+) and low () levels of the
factor. Interaction effects of two or more factors are also
calculated using this equation. In these calculations, the
ethanol production values attributed to the (+) and () lev-
els were determined by multiplying the sign in the columns
of design matrix for the factors involved in the interaction.
The following independent variables were included
X1 = total reducing sugars (TRS), X2 = temperature (C),
X3 = agitation (rpm) and X4 = culture time (h) shown in
Table 1. The dependent variables were ethanol and bio-
mass production. This preliminary analysis facilitated
2826 M.L. Cazetta et al. / Bioresource Technology 98 (2007) 2824–2828

Table 2 where, i j, are linear and quadratic coefficients, respectively,

23 Central composite design for investigating the effects of TRS in while b is the regression coefficients, k the number of
molasses, temperature and culture time on the ethanol production by
Z. mobilis ATCC 29191
factors studied and optimized in the experiment and e is
random error. The significance of each coefficient was
Factors Real levels
determined using a student’s test.
1 0 +1
X1 Molasses (g L1) 200 250 300 3. Results and discussion
X2 Temperature (C) 30 35 40
X3 Time of growth (h) 24 36 48
The first step of the statistical approach to the analysis
optimization was to establish the criteria that will define
the experimental factors that have a significant effect on
selecting the statistically significant factors, TRS concen- the response variables. Therefore, to optimize the ethanol
tration in sugar cane molasses (X1), temperature (X2) and production it was first performed as a 24 factorial design.
growth time (X3), therefore, two new levels for each factor Four relevant factors for the fermentative process were
were chosen according to the experimental design, shown selected in a factorial design 24. The variables, studied
in Table 2. The results of this factorial design evidenced simultaneously were: TRS concentration in sugar cane
that TRS concentration in sugar cane molasses (X1) and molasses (150 and 250 g L1), temperature (25 and
temperature (X2) are significant factors for ethanol produc- 35 C), agitation (180 oscillations per minute and static cul-
tion. In this case, a new full factorial design was employed ture) and culture time (12 and 24 h), as shown in Table 1.
to investigate the simultaneous effect of these two factors The main effects of the three factors concentration in
on the response. The experiments were carried out with a sugar cane molasses (2.72), temperature (3.27), and (8.47)
central point and star design, which consist in an identical are all positives, and agitation rate is negative (6.66).
planning, turning from 45 regarding to the original orien- The growth time (X4) main effect is the most significant fac-
tation,
pffiffiffi where the variables X1 and X2 were at a distance of torial design effect value for the production of ethanol. The
2 (1.414) from the central point, adding up to 11 experi- inclusion of agitation rate reduces the average ethanol pro-
ments, being 3 in the central point and 4 at the star design duction. Therefore, the subsequent runs were performed in
(Table 3). All the experiments were carried out in duplicate. a static format. Higher ethanol productions, 22.69 g L1
The RMS used in the present study is a central compos- and 30.08 g L1 were obtained in static culture (run 11
ite involving two different factors. Once the experiments and 12 in Table 1). According to Lee and Huang (2000)
are performed, the coefficients of linear and polynomial Z. mobilis is able to obtain an ethanol production close
models are calculated using the Eqs. (2) and (3): to the theoretical one from glucose through Entner–Dou-
X
k doroff pathway under aerobic conditions.
Y ¼ b0 þ bi xi þ e linear model ð2Þ Based upon the results obtained in the 24 factorial
i¼1 design a 23 factorial design was developed using new vari-
X
k X
k k X
X k ation levels in order to move sequentially in the direction of
Y ¼ b0 þ bi xi þ bij x2i þ bij xij þ e quadratic model maximizing the ethanol production. To define the best cul-
i¼1 i¼1 ii<j j
ture conditions it was necessary to test new sugar concen-
ð3Þ trations, temperature and culture time in a 23 factorial

Table 3
22 Central composite design and star design investigating the effects of TRS in molasses and temperature on the ethanol production by Z. mobilis ATCC
29191
Run Coded levels Real levels Measured responses
X1 X2 TRS (g L1) Temperature (C) Ethanol (g L1) Ypra (%) Yp/sb (g g1) Qpc (g L1 h1) Biomass (g L1)
1   150 25 46.43 64.26 0.35 0.97 1.39
2 +  250 35 42.39 73.98 0.40 0.88 1.46
3  + 150 25 47.73 77.01 0.42 0.99 1.25
4 + + 250 35 45.22 74.56 0.40 0.94 1.46
5 0 0 200 30 55.36 62.13 0.34 1.15 1.76
6 0 0 200 30 54.31 58.37 0.32 1.13 1.67
7 0 0 200 30 55.57 63.03 0.34 1.16 1.76
8 1414 0 80 30 28.55 73.04 0.40 0.59 1.39
9 0 1414 200 37 22.83 44.66 0.24 0.47 1.65
10 1414 0 270 30 33.43 53.20 0.29 0.70 1.60
11 0 1414 200 18 7.87 46.62 0.24 0.16 0.15
a
Ypr (%) = substrate conversion.
b
Yp/s (g g10) = yield ethanol for substrate.
c
Qp (g L1 h1) = ethanol productivity.
 M.L. Cazetta et al. / Bioresource Technology 98 (2007) 2824–2828
analysis, with a central point. The values of the new vari-
ables are listed in Table 2.
The results of the 23 design showed that the condition of
200 g L1 of TRS and temperature 30 C was the most
favorable, achieving 54.83 g L1 after a 48-hour-culture
time. The time was a decisive factor, once the ethanol pro-
duction increased to more than 60% from 24 to 48 h. By
comparison Bandaru et al. (2006) reported a maximum eth-
anol concentration (55.3 g L1) at 32.4 C, pH of 4.93 after
17.24 h from sago starch using Z. mobilis MTCC 92. Davis
et al. (article in press) reported similar values (54 g L1) for
Z. mobilis ZM4 from hydrolysed waste starch stream.
In the central point, 250 g L1 and 35 C, ethanol pro-
duction was an average of 31.15 g L1. The decrease in
ethanol production at high sugar concentration occurred
due to an increase in the osmotic pressure that is one of
the essential factors for by-products synthesis such as sor-
bitol and levan. The molasses was an industrial sucrose-
containing substrates that has been reported to contain
substantial salt content (Bekers et al., 2000). At 35 C
and 300 g L1 sugars concentration on molasses Cazetta
et al. (2005) obtained maximum sorbitol production by
Z. mobilis ATCC 29191.
The temperature of 40 C was negative for fermentative
process, resulting in lower productions, 4.6 g L1. Numer-
ous studies have shown that temperatures above 37 C are
detrimental for ethanol production (Lee et al., 1981; Sko-
tinicki et al., 1981; Lyness and Doelle, 1981; Diez and
Yokoya, 1996a). Based on the results of 23 factorial design,
it was performed as a 22 factorial design, with central com-
posite design, resulting in 11 experiments (Table 3). In this
stage the time was fixed in 48 h.
With the central composite design it was possible to con-
firm that maximum ethanol concentration occurred at the
central point, 55.8 g L1 on average (Fig. 1 and Table 3).
These values are similar to the ones described for ethanol
production from sucrose (Skotinicki et al., 1981; Lyness
and Doelle, 1981; Sreekumar et al., 1999) and sago starch
(Bandaru et al., 2006), which confirmed that the microor-
Fig. 1. Scheme showing results of 22 factorial analysis with star design to
ethanol production by Z. mobilis ATCC 29191 in molasses.
2827
ganism showed an optimal adaptation to the non-treated
molasses. The ethanol productivity was a mean of
1.1 g L1 h1.
A multiple regression analysis of the data was used to
describe the variables under study taking into account lin-
ear, quadratic and cross product terms for each factor. The
significance of the equation parameters on ethanol produc-
tion was assessed by the F test.
According to the RSM methodology, it was not possible
to fit the data obtained to either the linear or quadratic
mathematical model, however, there was evidence of a
slight curvature in the response surface. Since the average
response at the center point was larger than the average
response at the vertices, the surface was slightly convex.
The uncoupling between the biomass and ethanol pro-
duction can be observed clearly in these experiments
(Tables 1–3). Low biomass production is normally
observed in Z. mobilis, and cell growth and fermentation
are not linked (Parker et al., 1997). According to Rogers
et al. (1982) approximately 2% of the carbon source is con-
verted into biomass. This occurs due to Entner–Doudoroff
pathway used by this microorganism. This pathway yields
only a single mole of ATP per mole of sugar fermented,
giving Zymomonas the lowest molar growth yield reported
for a bacterium (Swings and DeLey, 1977).
The pH of the medium remained constant during the
experiments, varying from 6.0 at the beginning to 5.6, on
average n (Fig. 2). The pH has also been described as a fac-
tor that strongly interferes in the fermentative processes.
However, according to Diez and Yokoya (1996b) molasses
exhibits a buffering effect. This regulatory action depends
of molasses chemical composition. The main stabilizer
compounds of the pH are weak acids and amino acids that
act in the acid range, mainly between pH 3.0 and 5.0, or
phosphates, whose buffering effects occur in the range of
6.0 and 7.0. Falcão de Moraes et al. (1981) noted that
Z. mobilis possesses hugh tolerance at pH variations from
3.5 to 7.5, and its optimum at a range of 5.0–7.0. Buzato
(1984), observed no substantial oscillations on the alco-
Fig. 2. Initial and final pH values on the fermentation by Z. mobiliz
ATCC 29191.
2828 M.L. Cazetta et al. / Bioresource Technology 98 (2007) 2824–2828
holic yield at a pH range of 5.0–6.0, showing that there is
no major influence of this factor when Z. mobilis is culti-
vated on molasses.
Acknowledgements
This project was financially supported by the CAPES-
Brazil and CNPq-Brazil.
References
Bandaru, V.V.R., Somalanka, S.R., Mendu, D.R., Madicherla, N.R.,
Chityala, A., 2006. Optimization of fermentation conditions for the
production of ethanol from sago starch by co-immobilized amyloglu-
cosidase and cells of Zymomonas mobilis using response surface
methodology. Enz. Microbiol. Technol. 38, 209–214.
Barros, B., Scarmino, S.I., Bruns, E.R., 1995. Planejamento e Otimização
de Experimentos. Unicamp, Campinas, pp. 299.
Bekers, M., Vigants, A., Laukevics, J., Toma, M., Rapoports, Zikmanis,
A.P., 2000. The effect of osmo-induced stress on product formation by
Zymomonas mobilis on sucrose. Int. J. Food Microbiol. 36, 147–150.
Box, G.E.P., Hunter, S.W., Hunter, J.S., 1978. Statistic for Experiments
an Introduction to Design, Data Analysis and Model Building. John
Wiley, Inc., New York, pp. 653.
Buzato, J.B., 1984. Estudo de alguns fatores e condições que afetam a
fermentação alcoólica por Zymomonas mobilis CP4. Thesis (Master of
Science) – Rio Claro, SP, pp. 96.
Cazetta, M.L., Celligoi, M.A.P.C., Buzato, J.B., Scarmino, I.S., Silva,
R.S.F., 2005. Optimization study for sorbitol production by Zymo-
monas mobilis in sugar cane molasses. Process Biochem. 40, 747–751.
Davis, L., Jeon, Y., Svenson, C., Rogers, P., Pearce, J., Peiris, P., 2005.
Evaluation of wheat stillage for ethanol production by recombinant
Zymomonas mobilis. Biomass Bioenerg. 29, 49–59.
Davis, L., Rogers, P., Pearce, J., Peiris, P. Evaluation of Zymomonas
mobilis-based ethanol production from a hydrolysed waste starch
stream. Process Biochemistry. Article in press, www.elsevier.com/
locate/biombioe.
Diez, J.C., Yokoya, F., 1996a. Fermentação alcoólica de mosto de melaço
de cana-de-açúcar por processo descontı́nuo utilizando Zymomonas
mobilis. Arq. Biol. Tecnol. 32, 419–426.
Diez, J.C., Yokoya, F., 1996b. Efeito da temperatura e pH na produção de
etanol e levana durante a fermentação de sacarose por Zymomonas
mobilis. Arq. Biol. Tecnol. 39, 129–137.
Falcão de Moraes, J.O., Araújo, J.M., Rios, E.M., 1981. Fermentação
alcoólica de mosto de melaço em escala semi industrial por Zymomo-
nas mobilis. Rev. Instit. Antibiot. 20, 3–10.
Hamsaveni, D.R., Prapulla, S.G., Divakar, S., 2001. Response surface
methodology approach for the synthesis of isobutyl isobutyrate.
Process Biochem. 36, 1103–1109.
Jiménez, A.M., Borja, R., Martin, A., 2004. A comparative kinetic
evaluation of the anaerobic digestion of untreated molasses and
molasses previously fermented with Penicillium decumbens in batch
reactors. Biochem. Eng. J. 18, 121–132.
Lee, K.J., Skotnicki, M.L., Tribe, D.E., Rogers, P.L., 1981. The effect of
temperature on the kinetics of ethanol production by strains of
Zymomonas mobilis. Biotechnol. Lett. 3, 291–296.
Lee, E.C., Huang, C.T., 2000. Modeling of ethanol fermentation using
Zymomonas mobilis ATCC 10988 grown on the media containing
glucose and fructose. Biochem. Eng. J. 4, 217–227.
Lyness, E.W., Doelle, H.W., 1981. Fermentation pattern of sucrose to
ethanol conversion by Zymomonas mobilis. Biotechnol. Bioeng. 27,
121–128.
Mohagheghi, A., Ruth, M., Schell, D.J., 2006. Conditioning hemicellulose
hydrolysates for fermentation: effects of overliming pH on sugar and
ethanol yields. Process Biochem. 41, 1806–1811.
Nelson, N.A., 1944. Calorimetric adaptation of the Somogy method for
determination of glucose. Biochemistry 153, 375–380.
Parker, C., Peekhaus, N., Zhang, X., Conway, T., 1997. Kinetics of sugar
transport and phosphorylation influence glucose and fructose come-
tabolism by Zymomonas mobilis. Appl. Environ. Microbiol. 63, 3519–
3525.
Rogers, P.L., Lee, K.J., Skotinich, M.L., Tribe, D.E., 1982. Ethanol
production by Zymomonas mobilis. Adv. Biochem. Eng. 23, 27–
84.
Rogers, P., Joachimsthal, E., Haggett, K., 1997. Ethanol from lignocell-
ulosics: potencial for Zymomonas-based process. Australasian Bio-
technol. 7 (5), 304–309.
Ruanglek, V., Maneewatthana, D., Tripetchkul, S., 2006. Evaluation of
Thai agro-industrial wastes for bio-ethanol production by Zymomonas
mobilis. Process Biochem. 41, 1432–1437.
Skotinicki, M.L., Tribe, D.E., Rogers, P.L., 1981. Comparison of ethanol
production by different Zymomonas strains. Appl. Environ. Microbiol.
41, 889–893.
Sreekumar, O., Chang, N., Basappa, S., 1999. Optimization and interac-
tion of media components in ethanol production using Zymomonas
mobilis by response surface methodology. J. Biosci. Bioeng. 88, 334–
338.
Sreenath, H.K., Jeffries, T.W., 2000. Production of ethanol from wood
hydrolysate by yeasts. Biores. Technol. 72, 253–260.
Somogy, M.A., 1945. A new reagent for determination of sugar. J. Biol.
Chem. 160, 61–68.
Swings, J., DeLey, J., 1977. The biology of Zymomonas mobilis. Bacteriol.
Rev. 41, 1–46.
Tanaka, K., Hilary, Z.D., Ishizaki, A., 1999. Investigation of the utility of
pineapple juice and pineapple waste material as low-cost substrate for
ethanol fermentation by Zymomonas mobilis. J. Biosci. Bioeng. 87,
642–646.
Tao, F., Miao, J.Y., Shi, G.Y., Zhang, K.C., 2005. Ethanol fermentation
by an acid-tolerant Zymomonas mobilis under non-sterilized condition.
Proc. Biochem. 40, 183–187.
Further reading
Sprenger, G.A., 1996. Carbohydrate metabolism in Zymomonas mobilis:
A catabolic highway with some scenic routes. FMS Microbiol. Lett.
145, 301–307.