Biochemical Engineering Journal 28 (2006) 3643

Mathematical modeling to investigate temperature effect on kinetic parameters of ethanol fermentation

Muenduen Phisalaphong , Nuttapan Srirattana, Wiwut Tanthapanichakoon
Department of Chemical Engineering, Chulalongkorn University, Bangkok 10330, Thailand Received 21 June 2005; received in revised form 28 July 2005; accepted 27 August 2005

Abstract A mathematical model was developed to describe the effects of temperature on the kinetic parameters of ethanol fermentation by the occulating yeast, Saccharomyces cerevisiae M30, using cane molasses as the substrate. Three state variables, biomass, ethanol and the substrate and 12 kinetics parameters were used to describe the phenomenon. The kinetic parameters of the model were determined by using the least-square method. The inuence of temperature and initial sugar concentration on cell activities was investigated and quantied. Arrhenius relationships between operating temperature and the maximum specic growth rate, specic production rate, specic death rate were then established. The activation energy for growth, ethanol production and death rate were 3.461 104 , 3.496 104 and 1.777 105 kJ/kmol, respectively. Polynomial equations were established for the effects of temperature on the other kinetic parameters. A high temperature led to a decrease in the ethanol and cell yields but an increase in the inhibition effect of ethanol and sugar on cell growth and ethanol production. In addition, an inhibition effect of the initial sugar concentration on cell growth was clearly observed. The adopted mathematical model could describe very well the dynamics of ethanol fermentation from the beginning up to the stationary phase. 2005 Elsevier B.V. All rights reserved.
Keywords: Ethanol; Fermentation; Kinetics; Modeling; Temperature; Yeast

1. Introduction On account of limited global supply of oil, ethanol has reemerged as an alternative to, or extender for, petroleum-based liquid fuels. To effectively and efciently operate the fermentation process, the kinetic characteristics of cell growth and ethanol production are required. During fermentation of Saccharomyces cerevisiae, the activities of the microorganisms closely respond to changes in the environmental conditions, which are accompanied by variations in the mass transfer around and the metabolic behavior of the microorganisms. To gain insight into the morphology-associated time-variant process dynamics, various kinetic models associated with key parameters for ethanol fermentation have been proposed [111]. One of the most important variables is temperature. Temperature effects on fermentation performance of selected yeast strains for ethanol productivity were reported [1121]. In the work of Baranyl and Roberts [14], an arbitrary factor expressed as a function of the

Corresponding author. Tel.: +66 2 218 6875; fax: +66 2 218 6877. E-mail address: muenduen.p@chula.ac.th (M. Phisalaphong).

environmental conditions including the temperature was introduced to describe the physiological state of the cells and the growth kinetic equation was multiplied by this factor. The exponential increase of growth rate and productivity at high fermentation temperature were reported [15]. The deleterious effects of high temperature were considered to be due to the denaturation of ribosomes and enzymes and problems associated with the uidity of membranes [16]. In order to describe the effect of temperature, an empirical linear polynomial model was proposed to illustrate the effect of temperature and nutrient supplement on ethanol production rate [17]. The inuence of temperature on the kinetics parameters of ethanol and xylitol fermentation by Pachysolen tannophilus was investigated by S anchez et al. and the dependency of the maximum specic growth rate on the temperature was explained by the superposition of activation energies for cell growth and death [18]. The dependence of the kinetic coefcients on temperature using Arrhenius-type equations for autohydrolysis of brewerys spent grain was proposed [19]. On the contrary, according to Dalsenter et al. [20], the specic growth rate constant did not depend directly on temperature. The effect was suggested to be the delayed responses to

1369-703X/$ see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.bej.2005.08.039

M. Phisalaphong et al. / Biochemical Engineering Journal 28 (2006) 3643

37

Nomenclature KCM Kd KS KSP KSS KSSP Pm Pm CP CP0 CS CS0 CX CX0 E ED EP YP/S YX/S maintenance constant (h1 ) specic cell death rate (h1 ) saturation growth constant (g/L) saturation production constant (g/L) substrate growth inhibition term (g/L) substrate production inhibition term (g/L) ethanol inhibition term for growth (g/L) ethanol inhibition term for production (g/L) ethanol concentration (g/L) initial ethanol concentration (g/L) substrate concentration (g/L) initial substrate concentration (g/L) cell concentration (g/L) initial cell concentration (g/L) activation energy for cell growth (kJ/kmol) activation energy for cell death rate (kJ/kmol) activation energy for ethanol production (kJ/kmol) yield coefcient for product on substrate used for produce formation yield coefcient for cells on substrate used for cell formation

from molasses at pH 5.0. The prepared medium was sterilized at 121 C for 20 min. 2.2. Batch fermentation Batch fermentation experiments were carried out in duplicate with 1722% (w/v) of the initial reducing sugar solution from cane molasses as the sole carbon source for S. cerevisiae. Experiments were performed in 500 mL Erlenmeyer asks, with 250 mL total liquid volume. Experiments were initiated by transferring 5% of the starter culture to the prepared medium. Fermentation asks were then shaken in the incubator at 150 rpm. The experiments were carried out for 3 days in isothermal conditions at 30, 33, 35, 38 and 42 C and monitored by removing 6 mL samples every 6 h for cell, sugar and ethanol analyses. 2.3. Analytical methods For cell dry weight determination, a 5 mL sample of the fermentation broth was centrifuged at 3000 rpm for 10 min. The cell pellet was resuspended in 0.1N HCl and washed twice with distilled water and then dried at 90 C for 48 h and then weighed. Concentrations of ethanol were determined by a gas chromatography system using a Shimadzu Model GC 7AG equipped with a ame ionization detector. A column (0.125 cm i.d., 2 m, SS) packed with Porapak Q 80100 mesh was used with N2 as carrier gas. The injector temperature was 280 C, and the detector temperature was 300 C. To measure the amount of sugar in the sample, a 0.2 mL of the sample solution was hydrolyzed in 33% HCl at 100 C for 10 min and neutralized with NaOH solution. Then the reducing sugar content in the sample solution was determined by Lane and Eynons method. 2.4. Mathematical modeling Mathematical model was developed to determine the quantitative linkage between an environmental parameter and cell kinetics. Yeast strains used in industrial processes normally have limited osmotolerance. The inhibition by the substrate and product was always observed, especially at high substrate concentration. To construct a mathematical model that could describe the dynamic process of ethanol fermentation by S. cerevisiae M30, a comprehensive kinetic model modied from the Monod kinetics responding to changes in the environmental conditions was proposed. The cell metabolism was strongly affected by the substrate and product concentrations, which could be classied into two types: limiting and inhibiting. Oxygen content in the medium could also play an important role for the cell growth and also varying with the operating temperature. However, rotary of the shaker should be effective enough to provide gentle mixing and surface aeration during the rst period of the growth phase. In view of the fact that ethanol fermentation was an anaerobic process, the dissolve oxygen was not concerned as a limiting substance of the system. The proposed kinetics was therefore, modied in both substrate and product terms and combined with death rate and cell maintenance. The rate of cell growth, ethanol production and substrate consumption were related to the cell

Greek symbols specic growth rate (h1 ) m maximum specic growth rate (h1 ) specic production rate (h1 ) m maximum specic production rate (h1 )

the shifts in temperature and depended on the level of an essential component within the biomass, with the rate of the synthesis and denaturation reactions of this component depending on temperature. In the present paper, insights into the inuence of temperature on ethanol fermentation can be consolidated through model validation. The mathematical model is proposed to quantify and describe the temperature effect on the kinetic parameters of ethanol fermentation. The results consequently provide a better understanding of temperature effects on the cell activities for further development of the process. 2. Materials and methods 2.1. Culture and culture media S. cerevisiae M30 kindly provided by the laboratory of Dr. Savithree Limthong (Department of Microbiology, Kasetsart University, Bangkok) was used in this study. Stock cultures were stored in a PDA agar slant. Starter cultures were prepared by transferring a loop of the stock culture to 100 mL of medium and incubated at 33 C for 20 h before each starter culture was transferred to the main culture. The medium for the starter culture contained 0.05% ammonium sulfate and 5% inverse sugar

38

M. Phisalaphong et al. / Biochemical Engineering Journal 28 (2006) 3643

concentration (CX ), ethanol concentration (CP ) and substrate concentration (CS ) as follows: cells : ethanol : dCX = CX Kd CX dt dCP = CX dt 1 dCS = dt YX/S + 1 YP/S dCX dt dCP dt + KCM CX (3) (1) (2)

substrate :

The development of the modied model was based on the conventional model structure introducing some modications to take into account in terms of death rate (Kd CX ) in Eq. (1) and the substrate consumption was separated into the terms of substrate used for the biomass production, the ethanol production and the cell maintenance (Eq. (3)). The specic growth rate , specic production rate and specic death rate Kd represented the rate constants for cell growth, ethanol production and cell death, respectively. The KCM represented the maintenance constant. The YX/S and YP/S represented the yield coefcient for cell on substrate used for cell formation and the yield coefcient for ethanol on substrate used for ethanol formation, respectively. By the modication, the inuence of temperature and initial sugar concentration on cell activities could be clearly investigated and quantied. In a previous publication [22], the similar modied model was applied to identify the kinetic parameters of A. awamori growing on whole wheat our. The and were controlled by substrate limiting effect and inhibition effects of the substrate and ethanol as follows: C S 1 CP = m (4) 2 CS Pm K +C +
S S

The initial YP/S was tentatively estimated from the mass balance of the substrate to ethanol. The initial YX/S was obtained from the literature [23]. The other parameters (Kd , KSS , Pm , KSSP , Pm , and KCM ) was tentatively estimated by manual adjustment until a good visual t with the experimental data. The initial guesstimates of kinetic parameters were then re-calculated by running the program iterations. The best-t values of the parameter were estimated using the least-squares method by minimization of the sum of squared errors between the predicted and experimental data. The differential equations from the developed model were solved using the routine ODE 15 s available in MATLAB. The Newton method was applied to improve the estimated parameters; the numerical solution was obtained using functions in MATLAB (version 6.1). 3. Results and discussion 3.1. Fermentation performance Batch fermentation in shake asks for ethanol production was carried out in duplicate for 72 h with initial reducing sugar concentrations of 17, 20 and 22% (w/v) and controlled at constant temperatures 30, 33, 35, 38 and 42 C. Fig. 1 demonstrated the experimental results, as well as the modeled values of the cell, substrate and ethanol concentrations at 22% (w/v) initial sugar concentration with different controlled temperatures. The model estimations gave an adequate t to the experimental data for all the growth, substrate and ethanol concentration from the beginning to the stationary phase. This suggested that the proposed kinetic model and associated kinetic parameters were valid for data interpretation. The above experimental results revealed the biomass increased exponentially at the beginning, entered a stationary phase after 18 h and the cellular concentration dropped after 60 h of fermentation for all operating temperatures. For general ethanol production by yeast, the maximum fermentation time in batch process was 72 h. The biomass and ethanol production rates were enhanced slightly by the increase of the isothermal control from 30 to 33 C but the rates slightly decreased at 35 C. Cell growth and ethanol production declined considerably at temperature exceeding 35 C. At the optimal temperature for cell growth and ethanol production, the residual sugar which the cells were unable to consume was less than 40 g/L. With cane molasses as the C-source, the partial residual sugars were unfermentable sugars including arabinose and xylose. High residual sugar concentration could suggest the occurrence of cell death or inactive conditions including nutritional limitations or high toxic metabolite accumulates. The maximum ethanol concentration was obtained after 35 h of the fermentation at 30 C, which was consistent with the report by Torija et al. [24]. 3.2. Inuence of temperature The inuences of temperature on the ethanol fermentation by S. cerevisiae M30 was studied with regard to the kinetic parameters related to biomass and ethanol production and sugar

= m CS KSP + CS +

KSS

1 CP Pm (5)

2 CS KSSP

where m and m were the maximum specic growth rate and the maximum specic production rate. KS , KSS , Pm were, respectively, the saturation constant, the substrate inhibition term and the ethanol inhibition term for cell growth. KSP , KSSP , Pm were, respectively, the saturation constant, the substrate inhibition term and the ethanol inhibition term for ethanol production. In order to investigate the effect of temperature, the following 12 kinetic parameters: m , m , Kd , KS , KSS , Pm , KSP , KSSP , Pm , YX/S , YP/S and KCM were allowed to vary as a function of the temperature in each experiment. Since the model was a non-linear model with multi-parameters, the optimization for parameter set signicantly depended on the initial guesses for the parameter values. Hereby, the initial parameter values, m , m , KS and KSP , were tentatively estimated from the experimental data during the growth phase with the initial rate method by using LineweaverBurk plot [11] based on Monod rate equation.

M. Phisalaphong et al. / Biochemical Engineering Journal 28 (2006) 3643

39

Fig. 1. Experimental results and simulation of cell, reducing sugar and ethanol concentrations at different operating temperatures; lines correspond to model simulation while dots correspond to experimental data of S. cerevisiae M30; growth on cane molasses: ( ) 30 C, ( ) 33 C, ( ) 35 C, ( ) 38 C, and ( ) 42 C.

consumption. Using the least-squares curve-tting method, values of the kinetic coefcients estimated from the experimental data in the range 3042 C at various initial sugar concentrations from 17 to 22% (w/v) were obtained. The effects of the temperature on the model parameters were clearly evident. The maximum specic growth rate (m ) and specic death rate (Kd ) of cells and the maximum specic production rate of cells (m ) increased exponentially as the temperature increased. Expressed by the Arrhenius relationship, the temperature dependency of the reaction rate was tted very well with the experimental results (Fig. 2). The estimated values of the activation energy for cell growth (E), ethanol production (EP ) and death rate (ED ) were 3.461 104 , 3.496 104 and 1.777 105 kJ/kmol, respectively. As was normal with most microorganisms, the value for ED was higher than those for E [15]. The apparent activation energies for cell growth (E = 5.690 104 kJ/kmol) and for cell death (ED = 1.389 105 kJ/kmol) on the fermentation of dxylose by P. tannophilus to produce ethanol has been reported previously [18]. As a result of the activation of both cell growth and death rate, the effect of temperature on the overall maximum growth

Fig. 2. Arrhenius plots illustrating the effect of temperature on (a) the maximum specic growth rate (m ), (b) the maximum specic production rate (m ) and (c) the specic death rate (Kd ) of S. cerevisiae M30; growth on cane molasses at initial sugar concentration: ( ) 17%, ( ) 20% and ( ) 22% (w/v).

rate could be estimated. From our testing, this relationship could satisfactorily be used to simulate the experimental data of various yeast strains for the effect of temperature on the overall maximum specic growth rate reported in the literature [12] (data not shown). The apparent (or overall) maximum specic growth rate could vary from 0.10 to 0.78 h1 depending on yeast strains and operating conditions [12]. Nevertheless, it is noteworthy that the maximum specic growth rate, m from this study was not the overall maximum specic growth rate. The estimated value of m hereby was obtained from the model in which the cell accumulation was separated into the terms of cell growth and cell death (Eq. (1)). Moreover, the equation for the specic growth rate was modied with the inhibition terms of ethanol and substrate concentrations (Eq. (4)). Therefore, the

40

M. Phisalaphong et al. / Biochemical Engineering Journal 28 (2006) 3643

Fig. 3. The effect of temperature on (a) the saturated growth constant (KS (g/L)) and (b) the substrate inhibition term (KSS (g/L)) of S. cerevisiae M30; growth on cane molasses at initial sugar concentration: ( ) 17%, ( ) 20% and ( ) 22% (w/v).

values of the estimated m in this work should be higher than the estimated values of the overall maximum specic growth rate obtained from the Monod formalism. In addition, polynomial expressions were derived for the dependence of the other estimated kinetic parameters on the

temperature. At high temperature, the saturation growth constant (KS ) was found to increase, whereas the saturation production constant (KSP ) remained relatively constant (Fig. 3). Repressive effect of the reducing sugar expressed as the inverse of the 1 substrate inhibition term on cell growth (KSS ) and on ethanol 1 production (KSSP ) increased as temperature increased. In addition, the inhibition effect of ethanol concentration expressed as 1 ) the inverse of the ethanol inhibition term on cell growth (Pm 1 and on ethanol production (P m ) signicantly increased with temperature as well. The effects of temperature on the substrate inhibition terms (KSS , KSSP ) and the ethanol inhibition terms (Pm , Pm ) were depicted in Figs. 4 and 5, respectively. It should be pointed out that the values of the activation energy of the specic growth rate together with the sugar and ethanol inhibition terms for cell growth (E, KSS , Pm ) and the values for ethanol production (EP , KSSP , Pm ) were of the same magnitude. The present investigation illustrated that ethanol was produced from the primary metabolite. As for the cause of the inhibition effect on cell growth, a high temperature could result in changing the transport activity or saturation level of soluble compounds and solvents in the cells, which might increase the accumulation of toxic concentration including ethanol inside cells. The indirect effect of high temperature might also be ascribed to the denaturation of ribosome and enzyme and problems with the uidity of membranes [16]. Although the yield coefcients (YX/S , YP/S ) for cell and ethanol on substrate declined as the temperature increased in Fig. 6, the ratio of cell yield to ethanol yield gradually increased with temperature until the optimum temperature was reached. According to Blanch and Clark [11], the cell yield coefcient was reduced at a sufciently high temperature due to the denaturation of enzymes within the cell and cell death. A similar effect of the temperature on the overall growth rate and cell yield coefcient was also observed in this work. Nevertheless, the maintenance constant (KCM ) was found to be relatively constant. The biomass (YX/S ) and ethanol yields (YP/S ) obtained in this study were 0.300.55 and 0.300.55, respectively. Again, it is noteworthy that the yields hereby obtained from the substrate balance equation (Eq. (3)) in which the substrate consumption term was

Fig. 4. The effect of temperature on the substrate (sugar) inhibition term on (a) cell growth (KSS (g/L)) and (b) on ethanol production rate (KSSP (g/L)) of S. cerevisiae M30; growth on cane molasses at initial sugar concentration: ( ) 17%, ( ) 20% and ( ) 22% (w/v).

M. Phisalaphong et al. / Biochemical Engineering Journal 28 (2006) 3643

41

Fig. 5. The effect of temperature on the ethanol inhibition term on (a) cell growth (Pm (g/L)) and (b) ethanol production rate (Pm (g/L)) of S. cerevisiae M30; growth on cane molasses at initial sugar concentration: ( ) 17%, ( ) 20% and ( ) 22% (w/v).

separated into the terms of substrate used for biomass production, ethanol production and cell maintenance. The biomass yield (YX/S ) was dened as the ratio of biomass production to substrate consumption for biomass in particular. Similarly, the ethanol yield (YP/S ) was the ratio of ethanol production to substrate consumption for ethanol in particular. From these denitions, the values of YX/S and YP/S should be higher than those of the yields (YX/S and YP/S ) based on the overall substrate consumption. In this study, the biomass (YX/S ) and ethanol yields (YP/S ) derived from the overall substrate consumption were 0.0110.059 and 0.220.51 depending on the operating temperature and initial sugar concentration (data not shown) which were comparable to those reported elsewhere [2,7,8,11]. 3.3. Inuence of initial reducing sugar Signicant effects of the initial reducing sugar on the metabolic activation and repression of cell metabolism have been reported [1,25]. In order to examine possible interaction between temperature in the range from 30 to 42 C and initial sugar concentration in the range from 17 to 22% (w/v) and their effect on cell activities, the relevant kinetic parameters were estimated as described above. The range of experimental conditions was based upon the operating conditions in fuelethanol production plants. Within the controlled range, a signicant inhi-

bition by the initial sugar concentration on the maximum specic growth rate was clearly observed as shown in Fig. 2(a). However, no signicant inuence of the initial sugar concentration on the maximum specic ethanol production rate or specic death rate was observed (Fig. 2(b) and (c), respectively). A negative relationship between the initial sugar concentration and the saturated constants for cell growth (KS ) and for ethanol production (KSP ) was determined as shown in Fig. 3. Furthermore, interactions between temperature and initial sugar concentration on substrate inhibition term (KSS , KSSP ) and ethanol inhibition term (Pm , Pm ) were observed as illustrated in Figs. 4 and 5, respectively. Nevertheless, a signicant effect of the initial substrate on the reduction of yield coefcients (YX/S , YP/S ) and the increase of maintenance constant (KCM ) was observed only at high concentration (22%, w/v) (Figs. 6 and 7, respectively). Obviously, the inuences of temperature on the kinetic parameters at various sugar concentrations all showed a common trend. The effects of the initial sugar concentration on cell activities were consistent with published reports [9,21]. An excessively high initial sugar concentration could result in growth inhibition by osmotic stress that occurred when the concentration of a certain solute became so high that a large osmotic pressure gradient was established across the cell membrane. In this case, osmotic pressure was increased linearly with temperature. The loss of sugar transport activity could limit the growth rate and lead to more cell death.

Fig. 6. The effect of temperature on the yield coefcient for (a) cell on substrate used for cell formation (YX/S ) and (b) ethanol on substrate used for ethanol formation (YP/S ) of S. cerevisiae M30; growth on cane molasses at initial sugar concentration: ( ) 17%, ( ) 20% and ( ) 22% (w/v).

42

M. Phisalaphong et al. / Biochemical Engineering Journal 28 (2006) 3643

Fig. 7. The effect of temperature on the maintenance constant (KCM (h1 )) of S. cerevisiae M30; growth on cane molasses at initial sugar concentration: ( ) 17%, ( ) 20% and ( ) 22% (w/v).

growth kinetics from the initial condition until the stationary phase. The model parameters were tentatively estimated by the initial rate method and then adjusted using a non-linear regression technique assisted by a computer program to minimize the sum-of-squares deviation between the model predictions and experimental data. The estimated values were well tted with the experimental data. The maximum specic growth rate, the maximum specic production rate and the specic death rate of cells increased with temperature as Arrhenius functions. The estimated activation energies for cell growth, ethanol production and death rate were 3.461 104 , 3.496 104 and 1.777 105 kJ/kmol, respectively. High temperature also accelerated the inhibition effect of the substrate and ethanol on the cell activities, thereby lowering both cell and ethanol yields. Furthermore, positive relationships of the initial sugar concentration and the substrate inhibition on cell growth and on ethanol production were determined. Besides, the interaction between temperature and initial sugar concentration to accelerate the inhibition effect was observed. From the scale-up test, the developed model from the small scale process reproduced adequately the kinetic behavior of the large-scale experimental data. However, for the application in production scale, some kinetic parameters might have to be adjusted according to the variation in the microorganism strain and the aeration and shear rate caused by differences in the geometry and mixing intensity between the two systems. In this study, we have shown how the effects of the environmental parameters, temperature and initial sugar concentration, on the cell activities could be elucidated and quantied by the utilization of mathematical models. These will contribute to a better understanding of the environmental effect on cell activities and could be applied as a tool for further process development. Acknowledgments

Fig. 8. Experimental results and simulation of cell ( ), reducing sugar ( ) and ethanol ( ) concentrations in a 10 l fermentor at 33 C and 300 rpm; lines correspond to model simulation while dots correspond to experimental data of S. cerevisiae M30; growth on cane molasses.

In order to investigate possible scale-up effects on the application to fermentation process, batch ethanol fermentation experiments using cane-molasses as the main substrate were carried out in duplicated with 22% (w/v) of the initial reducing sugar in a 10 l fermentor at 33 C and 300 rpm. From Fig. 8, it can be said that the mathematical model and estimated kinetic parameters developed from the small scale could adequately predict the dynamics of ethanol fermentation in 10 l fermentor. The simulation results were close to the experimental data throughout fermentation; only some minor deviations of ethanol concentration prole at the beginning and substrate concentration prole at the end of the fermentation were observed. 4. Conclusions In this paper, we have cultivated S. cerevisiae M30 in batch, at different temperature and different initial molasses concentration. A mathematical model was developed in order to follow

We thank the Thailand Research Fund (TRF), Project TRG 4580047, for nancial support, the Thailand-Japan Technology Transfer Project (TJTTP-OECF) for the support of analytical equipments and S. Limthong for provision of the cell cultures. WT also received support from Research Team Award, TRF. References
[1] G. Birol, P. Doruker, B. Kirdar, Z.I. Onsan, K. Ulgen, Mathematical description of ethanol fermentation by immobilized Saccharomyces cerevisiae, Process Biochem. 33 (1998) 763771. [2] A.E. Ghaly, A.A. El-Taweel, Kinetic modeling of continuous production of ethanol from cheese whey, Biomass and Bioenergy 12 (1997) 461472. [3] R.D. Tyagi, T.K. Ghose, Batch and multistage continuous ethanol fermentation of cellulose hydrolysate and optimum design of fermentor by graphical analysis, Biotechnol. Bioeng. 22 (1980) 19071928. [4] J. Sainz, F. Pizarro, J.R. P erez-Correa, E.A. Agosin, Modeling of yeast metabolism and process dynamics in batch fermentation, Biotechnol. Bioeng. 81 (2003) 818828. [5] B. Andr es-Toro, J.M.A. Giron-Sierra, Kinetic model for beer production under industrial operational conditions, Math. Comput. Simulat. 48 (1998) 6574.

M. Phisalaphong et al. / Biochemical Engineering Journal 28 (2006) 3643 [6] H.J. Rehm, G. Reed, A. P uhler, P. Stadler, Biotechnology, McGraw-Hill, New York, 2001. [7] S. Aiba, M. Shoda, M. Nagatani, Kinetics of product inhibition in alcohol fermentation, Biotechnol. Bioeng. 67 (2000) 671690. [8] S.C. Oliveira, H.F. De Castro, A.E.S. Visconti, R. Giudici, Continuous ethanol fermentation in a tower reactor with occulating yeast recycle: scale-up effects on process performance, kinetic parameters and model predictions, Bioprocess Eng. 20 (1999) 525530. [9] N.K. Sree, M. Sridhar, K. Suresh, I.M. Banat, L.V. Rao, High alcohol production by repeated batch fermentation using an immobilized osmotolerant Saccharomyces cerevisiae, J. Ind. Microbiol. Biotechnol. 24 (2000) 222226. [10] A. Nishiwaki, I.J. Dunn, Analysis of the performance of a two-stage fermentor with cell recycle for continuous ethanol production using different kinetic models, Biochem. Eng. J. 4 (1999) 3744. [11] H.W. Blanch, D.S. Clark, Biochemical Engineering, Marcel Dekker, New York, 1997. [12] K. Evans, A. Mohagheghi, J. Hamilton, M. Zhang, Effect of corn stover hydrolysate and temperature on fermentation performance of selected yeast strains, in: Proceedings of the 24th Biotechnology for Fuels and Chemicals Symposium, Gatlinburg, TN, USA, 2002, pp. 213. [13] W. Messens, J. Verluyten, F. Leroy, L.D. Vuyst, Modelling growth and bacteriocin production by Lactobacillus curvatus LTH 1174 in response to temperature and pH values used for European sausage fermentation processes, Int. J. Food Microbiol. 81 (2003) 4152. [14] J. Baranyl, T.A. Roberts, A dynamic approach to predicting bacterial growth in food, Int. J. Food Microbiol. 23 (1994) 277294. [15] J.R. Hettenhaus, Ethanol fermentation strains: present and future requirements for biomass to ethanol commercialization, Report to United States Department of Energy and National Renewable Energy Laboratory, 1998.

43

[16] T.A. McMeckin, J. Olley, D.A. Ratkwsky, T. Ross, Predictive microbiology: towards the interface and beyond, Int. J. Food Microbiol. 73 (2002) 395407. [17] G. Dragone, D.P. Silva, J.B.A. Silva, Factors inuencing ethanol rates at high-gravity brewing, Lebensm. Wiss. Technol. 37 (2004) 797802. [18] S. S anchez, V. Bravo, A.J. Moya, E. Castro, F. Camacho, Inuence of temperature on the fermentation of d-xylose by Pachysolen tannophilus to produce ethanol and xylitol, Process Biochem. 39 (2004) 673679. [19] F. Carvalheiro, G. Garrote, J.C. Paraj o, H. Pereira, F.M. G rio, Kinetic modeling of brewerys spent grain autohydrolysis, Biotechnol. Prog. 21 (2005) 233243. [20] F.D.H. Dalsenter, G. Viccini, M.C. Barga, D.A. Mitchell, N.A. Krieger, A mathematical model describing the effect of temperature variations on the kinetics of microbial growth in solid-state culture, Process Biochem. 40 (2005) 801807. [21] J.M. Salmon, J.C. Maurico, Relationship between sugar uptake kinetics and total sugar consumption in different industrial Saccharomyces cerevisiae strains during alcohol fermentation, Biotechnol. Lett. 16 (1994) 8994. [22] A.A. Koutinas, R. Wang, I.K. Kookos, C. Webb, Kinetic parameters of Aspergillus awamori in submerged cultivations on whole wheat our under oxygen limiting condition, Biochem. Eng. J. 16 (2003) 2334. [23] M.L. Shuler, F. Kargi, Bioprocess Engineering, Prentice-Hall, New Jersey, 2001. [24] M.J. Torija, N. Roz` es, M. Poblet, J.M. Guillam on, A. Mas, Effects of fermentation temperature on the strain population of Saccharomyces cerevisiae, Int. J. Food Microbiol. 80 (2003) 4753. [25] S. Helle, D. Cameron, J. Lam, B. Whit, S. Duff, Effect of inhibitory compounds found in biomass hydrolysates on growth and xylose fermentation by a genetically engineered strain of Saccharomyces cerevisiae, Enzyme Microb. Technol. 33 (2003) 786792.