Bioresource Technology 98 (2007) 28242828

Fermentation of molasses by Zymomonas mobilis: Eects of temperature and sugar concentration on ethanol production
M.L. Cazetta
b

a,*

, M.A.P.C. Celligoi b, J.B. Buzato b, I.S. Scarmino

a Department of Food and Drugs, Londrina State University, Post Box 6001, CEP: 86051-990, Londrina-PR, Brazil 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 dierent culture conditions using Z. mobilis in batch fermentation. The total reducing sugars (TRS) concentrations in the molasses, temperature, agitation and culture time eects 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 buering eect. 2006 Elsevier Ltd. All rights reserved.
Keywords: Ethanol; Zymomonas mobilis; Sugar cane molasses; Factorial design

1. Introduction The depletion of fossil fuel reserves, the unstable panorama 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 impediment wide use of ethanol as fuel. Therefore, the development 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). 0960-8524/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2006.08.026
*

sources is important for the biofuel ethanol production on a commercial scale (Tanaka et al., 1999; Tao et al., 2005). Many studies have been done that focus on production improvement and decreasing its costs (Sreenath and Jeries, 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 specic rates of glucose uptake and ethanol production (Rogers et al., 1982, 1997) via the Entner-Doudoro 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 (Jimenez 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

M.L. Cazetta et al. / Bioresource Technology 98 (2007) 28242828

2825

factor for the economical viability of substances production by fermentation. The traditional one-at-a-time optimization strategy is relatively simple, and the individual eects of medium factors can be graphically depicted without the need of the statistical analysis. Unfortunately, it frequently fails to locate the region of optimum response in such procedures. In this case, fractional and/or full factorial design provides an ecient approach to optimization. A combination of factors generating a certain optimum response can be identied though factorial design and the use of response surface methodology (RSM) (Box et al., 1978). The response-surface methodology is an empirical modeling system that assesses the relationship between a group of variables that can be controlled experimentally and the observed response (Sreekumar et al., 1999; Hamsaveni et al., 2001). Response surface methodology (RSM) is a useful model to study the eect of several factors inuencing the responses by varying them simultaneously and carrying out a limited number of experiments (Hamsaveni et al., 2001). The aim of this work was to study the inuence between four factors and their interaction to optimize the ethanol production by Z. mobilis ATCC 29191 in sugar cane molasses using factorial design and analysis by RSM. The selected factors were sugar concentration on molasses, temperature, agitation rate and culture time. The measured responses were ethanol and biomass. 2. Methods 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 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 ve 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 fermentations were carried out in duplicate in the sugar cane molasses culture medium, in the dierent culture conditions, 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 sample at 605 nm using a standard curve of absorbance against dry cell mass. The total reducing sugars (TRS) were quantied 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

Table 1 24 Factorial experimental design investigating the eect of TRS concentration in molasses, temperature, agitation rate and culture time to ethanol production by Z. mobilis ATCC 29191 Run Variables in coded levels X1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 + + + + + + + + X2 + + + + + + + + X3 + + + + + + + + X4 + + + + + + + + Measured responses Ethanol (g L1) 5.74 5.38 7.95 6.98 2.97 9.89 4.09 1.98 19.33 7.31 22.69 30.08 2.94 12.20 2.23 15.91 Biomass (g L1) 0.89 0.81 0.75 0.65 0.56 0.31 0.53 0.41 1.07 1.22 1.29 1.41 0.81 1.07 0.65 0.43 Real levels 1 X1 X2 X3 X4 Molasses (g L ) Temperature (C) Agitation rate (rpm) Time of cultivation (h)
1

Factors

+1 250 35 180 24

150 25 0 12

a ux of 40 ml min1 and isopropanol as an internal standard. 2.3. Experimental design The conditions to optimize Z. mobilis ethanol production 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 eects for each of the factors studied were dened by the Eq. (1): Efi  i  i y y 1

where Efi is the eect of the ith factor on the ethanol production, and  i and  i are the average ethanol proy y ductions values at the high (+) and low () levels of the factor. Interaction eects of two or more factors are also calculated using this equation. In these calculations, the ethanol production values attributed to the (+) and () levels 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 biomass production. This preliminary analysis facilitated

2826

M.L. Cazetta et al. / Bioresource Technology 98 (2007) 28242828

Table 2 23 Central composite design for investigating the eects of TRS in molasses, temperature and culture time on the ethanol production by Z. mobilis ATCC 29191 Factors Molasses (g L1) Temperature (C) Time of growth (h) Real levels 1 X1 X2 X3 200 30 24 0 250 35 36 +1 300 40 48

where, i j, are linear and quadratic coecients, respectively, while b is the regression coecients, k the number of factors studied and optimized in the experiment and e is random error. The signicance of each coecient was determined using a students test. 3. Results and discussion The rst step of the statistical approach to the analysis optimization was to establish the criteria that will dene the experimental factors that have a signicant eect on the response variables. Therefore, to optimize the ethanol production it was rst performed as a 24 factorial design. Four relevant factors for the fermentative process were selected in a factorial design 24. The variables, studied simultaneously were: TRS concentration in sugar cane molasses (150 and 250 g L1), temperature (25 and 35 C), agitation (180 oscillations per minute and static culture) and culture time (12 and 24 h), as shown in Table 1. The main eects of the three factors concentration in sugar cane molasses (2.72), temperature (3.27), and (8.47) are all positives, and agitation rate is negative (6.66). The growth time (X4) main eect is the most signicant factorial design eect value for the production of ethanol. The inclusion of agitation rate reduces the average ethanol production. Therefore, the subsequent runs were performed in a static format. Higher ethanol productions, 22.69 g L1 and 30.08 g L1 were obtained in static culture (run 11 and 12 in Table 1). According to Lee and Huang (2000) Z. mobilis is able to obtain an ethanol production close to the theoretical one from glucose through EntnerDoudoro pathway under aerobic conditions. Based upon the results obtained in the 24 factorial design a 23 factorial design was developed using new variation levels in order to move sequentially in the direction of maximizing the ethanol production. To dene the best culture conditions it was necessary to test new sugar concentrations, temperature and culture time in a 23 factorial

selecting the statistically signicant factors, TRS concentration in sugar cane molasses (X1), temperature (X2) and growth time (X3), therefore, two new levels for each factor were chosen according to the experimental design, shown in Table 2. The results of this factorial design evidenced that TRS concentration in sugar cane molasses (X1) and temperature (X2) are signicant factors for ethanol production. In this case, a new full factorial design was employed to investigate the simultaneous eect of these two factors on the response. The experiments were carried out with a central point and star design, which consist in an identical planning, turning from 45 regarding to the original orientation, where the variables X1 and X2 were at a distance of p 2 (1.414) from the central point, adding up to 11 experiments, being 3 in the central point and 4 at the star design (Table 3). All the experiments were carried out in duplicate. The RMS used in the present study is a central composite involving two dierent factors. Once the experiments are performed, the coecients of linear and polynomial models are calculated using the Eqs. (2) and (3):
Y b0 Y b0
k X i1 k X i1

bi xi e linear model bi xi
k X i1

2
bij xij e quadratic model

bij x2 i

k k XX ii<j j

Table 3 22 Central composite design and star design investigating the eects of TRS in molasses and temperature on the ethanol production by Z. mobilis ATCC 29191 Run Coded levels X1 1 2 3 4 5 6 7 8 9 10 11
a b c

Real levels X2 + + 0 0 0 0 1414 0 1414 TRS (g L1) 150 250 150 250 200 200 200 80 200 270 200 Temperature (C) 25 35 25 35 30 30 30 30 37 30 18 Ethanol (g L1) 46.43 42.39 47.73 45.22 55.36 54.31 55.57 28.55 22.83 33.43 7.87

Measured responses Ypra (%) 64.26 73.98 77.01 74.56 62.13 58.37 63.03 73.04 44.66 53.20 46.62 Yp/sb (g g1) 0.35 0.40 0.42 0.40 0.34 0.32 0.34 0.40 0.24 0.29 0.24 Qpc (g L1 h1) 0.97 0.88 0.99 0.94 1.15 1.13 1.16 0.59 0.47 0.70 0.16 Biomass (g L1) 1.39 1.46 1.25 1.46 1.76 1.67 1.76 1.39 1.65 1.60 0.15

+ + 0 0 0 1414 0 1414 0

Ypr (%) = substrate conversion. Yp/s (g g10) = yield ethanol for substrate. Qp (g L1 h1) = ethanol productivity.

M.L. Cazetta et al. / Bioresource Technology 98 (2007) 28242828

2827

analysis, with a central point. The values of the new variables 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 production increased to more than 60% from 24 to 48 h. By comparison Bandaru et al. (2006) reported a maximum ethanol 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 production 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 sorbitol and levan. The molasses was an industrial sucrosecontaining 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. Numerous studies have shown that temperatures above 37 C are detrimental for ethanol production (Lee et al., 1981; Skotinicki 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 composite design, resulting in 11 experiments (Table 3). In this stage the time was xed in 48 h. With the central composite design it was possible to conrm 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 conrmed that the microor-

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 linear, quadratic and cross product terms for each factor. The signicance of the equation parameters on ethanol production was assessed by the F test. According to the RSM methodology, it was not possible to t 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 production can be observed clearly in these experiments (Tables 13). 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 converted into biomass. This occurs due to EntnerDoudoro 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 factor that strongly interferes in the fermentative processes. However, according to Diez and Yokoya (1996b) molasses exhibits a buering eect. 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 buering eects occur in the range of 6.0 and 7.0. Falcao 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 5.07.0. Buzato (1984), observed no substantial oscillations on the alco-

Fig. 1. Scheme showing results of 22 factorial analysis with star design to ethanol production by Z. mobilis ATCC 29191 in molasses.

Fig. 2. Initial and nal pH values on the fermentation by Z. mobiliz ATCC 29191.

2828

M.L. Cazetta et al. / Bioresource Technology 98 (2007) 28242828 Lee, K.J., Skotnicki, M.L., Tribe, D.E., Rogers, P.L., 1981. The eect of temperature on the kinetics of ethanol production by strains of Zymomonas mobilis. Biotechnol. Lett. 3, 291296. 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, 217227. Lyness, E.W., Doelle, H.W., 1981. Fermentation pattern of sucrose to ethanol conversion by Zymomonas mobilis. Biotechnol. Bioeng. 27, 121128. Mohagheghi, A., Ruth, M., Schell, D.J., 2006. Conditioning hemicellulose hydrolysates for fermentation: eects of overliming pH on sugar and ethanol yields. Process Biochem. 41, 18061811. Nelson, N.A., 1944. Calorimetric adaptation of the Somogy method for determination of glucose. Biochemistry 153, 375380. Parker, C., Peekhaus, N., Zhang, X., Conway, T., 1997. Kinetics of sugar transport and phosphorylation inuence glucose and fructose cometabolism 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 lignocellulosics: potencial for Zymomonas-based process. Australasian Biotechnol. 7 (5), 304309. Ruanglek, V., Maneewatthana, D., Tripetchkul, S., 2006. Evaluation of Thai agro-industrial wastes for bio-ethanol production by Zymomonas mobilis. Process Biochem. 41, 14321437. Skotinicki, M.L., Tribe, D.E., Rogers, P.L., 1981. Comparison of ethanol production by dierent Zymomonas strains. Appl. Environ. Microbiol. 41, 889893. Sreekumar, O., Chang, N., Basappa, S., 1999. Optimization and interaction of media components in ethanol production using Zymomonas mobilis by response surface methodology. J. Biosci. Bioeng. 88, 334 338. Sreenath, H.K., Jeries, T.W., 2000. Production of ethanol from wood hydrolysate by yeasts. Biores. Technol. 72, 253260. Somogy, M.A., 1945. A new reagent for determination of sugar. J. Biol. Chem. 160, 6168. Swings, J., DeLey, J., 1977. The biology of Zymomonas mobilis. Bacteriol. Rev. 41, 146. 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, 642646. 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, 183187.

holic yield at a pH range of 5.06.0, showing that there is no major inuence of this factor when Z. mobilis is cultivated on molasses. Acknowledgements This project was nancially supported by the CAPESBrazil 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 amyloglucosidase and cells of Zymomonas mobilis using response surface methodology. Enz. Microbiol. Technol. 38, 209214. Barros, B., Scarmino, S.I., Bruns, E.R., 1995. Planejamento e Otimizacao de Experimentos. Unicamp, Campinas, pp. 299. Bekers, M., Vigants, A., Laukevics, J., Toma, M., Rapoports, Zikmanis, A.P., 2000. The eect of osmo-induced stress on product formation by Zymomonas mobilis on sucrose. Int. J. Food Microbiol. 36, 147150. 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 condicoes que afetam a fermentacao alcoolica 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 Zymomonas mobilis in sugar cane molasses. Process Biochem. 40, 747751. 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, 4959. 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. Fermentacao alcoolica de mosto de melaco de cana-de-acucar por processo descontnuo utilizando Zymomonas mobilis. Arq. Biol. Tecnol. 32, 419426. Diez, J.C., Yokoya, F., 1996b. Efeito da temperatura e pH na producao de etanol e levana durante a fermentacao de sacarose por Zymomonas mobilis. Arq. Biol. Tecnol. 39, 129137. Falcao de Moraes, J.O., Araujo, J.M., Rios, E.M., 1981. Fermentacao alcoolica de mosto de melaco em escala semi industrial por Zymomo nas mobilis. Rev. Instit. Antibiot. 20, 310. Hamsaveni, D.R., Prapulla, S.G., Divakar, S., 2001. Response surface methodology approach for the synthesis of isobutyl isobutyrate. Process Biochem. 36, 11031109. Jimenez, 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, 121132.

Further reading
Sprenger, G.A., 1996. Carbohydrate metabolism in Zymomonas mobilis: A catabolic highway with some scenic routes. FMS Microbiol. Lett. 145, 301307.