- Email: [email protected]
Enzymatic esterification of ethanol by an immobilised Rhizomucor miehei lipase in a perforated rotating disc bioreactor
Enzyme and Microbial Technology 26 (2000) 446 – 450
Enzymatic esterification of ethanol by an immobilised Rhizomucor miehei lipase in a perforated rotating disc bioreactor A.C. Oliveiraa,*, M.F. Rosaa, M.R. Aires–Barrosb, J.M.S. Cabralb a
INETI—Departamento de Energias Renova´veis, Estrada do Pac¸o do Lumiar, 1649-038 Lisbon, Portugal b IST—Centro de Engenharia Biolo´gica e Quı´mica, Av. Rovisco Pais, 1049-001 Lisbon, Portugal Received 22 September 1999; received in revised form 22 November 1999; accepted 4 December 1999
Abstract A perforated rotating disc bioreactor was developed to perform the esterification of ethanol with oleic acid, catalyzed by a lipase from Rhizomucor miehei immobilized by adsorption on to a hydrophobic support—Accurel EP700. The bioreactor with total recirculation operated at an optimum agitation rate of 400 rev./min. The experimental results, in this condition, were predict by a kinetic model using the constants obtained in the batch (Erlenmeyer flasks) assays: a catalytic constant, kcat ⫽ 5.78 mmol/h 䡠 mg protein; a Michaelis constant for ethanol, Km(Et) ⫽ 1.20 M; a Michaelis constant for oleic acid, Km(Ol) ⫽ 1.16 ⫻ 10⫺8 M, and a dissociation constant of the ethanol-lipase complex, K(Et) ⫽ 9.46 ⫻ 107 M. The efficiency of conversion gradually decreased during continuous operation of the reactor. The enzymatic activity decayed according to a first order deactivation model and the integrated equations of a continuous stirred tank reactor (CSTR) and a plug flow reactor (PFR). A half-life time of the lipase of about 10 days and a deactivation constant of 0.003 h⫺1 were obtained in the present system. © 2000 Elsevier Science Inc. All rights reserved. Keywords: Enzymatic esterification; Rhizomucor miehei lipase; Perforated rotating disc bioreactor
1. Introduction The enzymatic esterification of ethanol and oleic acid has been performed in a biphasic system [1– 4], with free and immobilized Rhizomucor miehei lipase as the esterification reaction catalyst. The use of immobilized enzyme has some industrial and economical advantages such as ease recovery and re-use, greater stability of the enzyme, and possibility of continuous operation. However, the esterification reaction referred above has been performed in a batch reactor that is not the most advantageous process if the objective is to work in large scale. In this case, the use of continuous processes will be preferable. With this purpose, some studies of the ethanol/oleic acid esterification reaction were carried out by Csa´nyi et al. [5] in a hollow fiber membrane bioreactor where the lipase was immobilized by adsorption onto the inner wall of the fibers. Perforated rotating disc contactor (PRDC) has been successfully applied as a continuous extraction equipment for
protein and enzyme, using reversed micellar systems [6,7]. With this apparatus, high-mass transfer rates between the phases are obtained, leading to high separation efficiencies [8]. This PRDC with free-flow areas has been industrially used as an alternative to the conventional extraction equipment [8,9]. The aim of this work was to use a type of reactor based on the advantages of PRDC to develop a novel process for the continuous enzymatic esterification of ethanol and oleic acid, using as catalyst a Rhizomucor miehei lipase immobilized by adsorption onto a hydrophobic support—Accurel EP700. The described system could take part in an integrated bioprocess consisting of a continuous fermentation followed by a continuous extraction of the inhibitory product (ethanol).
2. Materials and methods 2.1. Enzyme and chemicals
* Corresponding author. Tel.: ⫹351-1-712-72-11; fax: ⫹351-1-712-71 95. E-mail address: [email protected] (A.C. Oliveira)
A commercial lipase from R. miehei, Palatase M 1000 L, was a kind gift of Novo–Nordisk (Baegsraend, Denmark).
0141-0229/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S 0 1 4 1 - 0 2 2 9 ( 0 0 ) 0 0 1 8 5 - 4
A.C. Oliveira et al. / Enzyme and Microbial Technology 26 (2000) 446 – 450
447
The amount of protein immobilized onto the carrier was determined by a modified Folin assay and it is expressed per gram of support. 2.4 Experimental procedure The biphasic system used consisted of an organic phase, the oleic acid, and of an aqueous phase composed by phthalate-NaOH buffer (110 mM, pH 5.5) with ethanol. Initially, the column was filled with 55 ml of oleic acid (oleic acid concentration ⫽ 1.55 M) and 55 ml of aqueous phase (ethanol concentration ⫽ 0.64 M), at which the immobilized enzyme was added. Fig. 1. (A), perforated rotating disc bioreactor: 1, aqueous phase inlet; 2, organic phase outlet; 3, aqueous phase outlet; 4, organic phase inlet; 5, refrigerated water. (B), perforated disc.
The immobilization support, a micropouros polyamide powder (Accurel EP700; particle size 300 –1000 m, void volume 75%, pore size 5 to 30 nm) from Akzo (Obernburg, Germany) was a kind gift of S.C.I.E. Irma˜os Planas Almasque Lda (Lisbon, Portugal). All of the reagents used were of analytical grade from Merck (Damstadt, Germany). 2.2. Experimental apparatus The esterification reactions were carried out in a perforated rotating disc (PRD) bioreactor (Fig. 1) made of Perspex tube 32 mm internal diameter (i.d.) and 160 mm high. Four perforated discs equally separated were mounted on a central shaft, which was rotated at different velocities (rev./ min). The perforated discs had a diameter of 30 mm and were drilled with six holes of 7.5-mm diameter and six holes of 1.5-mm diameter. The reactor was maintained at 30°C using water in a refrigerated jacket. To prevent the immobilized lipase from leaving the PRD bioreactor, a tight porous filter net was provided at the reactor organic phase outlet.
2.4.1. Batch operation The effect of agitation rate (100 –500 rev./min) on the ethanol/oleic acid esterification reaction, using the immobilized enzyme, was carried out in the PRD bioreactor with total recirculation. Samples of the organic phase were taken at similar residence times and the ester concentration was analyzed by CGC. The initial reaction rates were estimated from the slope of plots of ethyl esters produced versus residence time and expressed as mmol/h 䡠 mg of immobilized protein. 2.4.2. Continuous operation The start-up of the reactor included, initially, the operation with total recirculation during 24 h, at a constant agitation rate, following the continuous operation after this period of time. The aqueous and the organic phases get into the reactor in counter-current flow as shown in Fig. 1. The flow rates of the two phases were maintained constant by using peristaltic pumps with a flow rate of 0.035 or 0.07 ml/min, depending on the assay. The samples were collected in the organic phase outlet, at regular time intervals, and the amount of ester produced, measured by CGC, was used to calculated the ethanol conversion degree obtained. 2.5. Ethyl ester determination
2.3. Preparation of the immobilized enzyme The lipase was immobilized by adsorption on Accurel EP700. For the bioreactor experiments, 3 ml of the lipase solution (10.2 mg of protein/ml of enzyme solution) was added to 900 mg of the immobilization support. After being vortex-mixed for 1 min, the enzyme-support contact time was 1 h [3], at room temperature. At this time, 10 ml of buffer (110 mM phthalate-NaOH buffer, pH 5.5) was added to the preparation and the support was separated by filtration (Millipore system) and washed twice with buffer.
Ethyl esters produced during the bioreactor operation were quantified by capillary gas chromatography (CGC) by using a Varian 3300 chromatograph equipped with a FID and a Supelcowax 10 column (i.d. ⫽ 0.32 mm; 30 m; film thickness ⫽ 0.25 m). Helium was used as the carrier gas and the injector and detector temperatures were kept at 250°C and 260°C, respectively. The oven temperature was kept at 200°C for 11.5 min and then heated to 225°C. The organic phase samples (20 mg) were analyzed directly, after the addition of methyl heptadecanoate as internal standard (0.53 mg).
448
A.C. Oliveira et al. / Enzyme and Microbial Technology 26 (2000) 446 – 450
Fig. 2. Effect of agitation rate on ester production when using the perforated rotating disc bioreactor with total recirculation. Agitation rate (rev./ min): 〫, 100; F, 200; 䡺, 300; E, 400; ‚, 500. Reaction conditions: organic/aqueous phases ratio of 1; initial ethanol and oleic acid concentration, respectively, 0.64 and 1.55 M; immobilized enzyme ⫽ 3.09 mg/g support.
3. Results and discussion 3.1. Batch operation The influence of the agitation rate on the enzymatic esterification of ethanol and oleic acid performed in a perforated rotating disc bioreactor with total recirculation was evaluated. Ethanol and oleic acid were initially at 0.64 and 1.55 M respectively, and 3.09 mg of immobilized protein (R. miehei lipase)/g support was used as initial reaction conditions. Different performances were obtained when using different agitation rates, namely for lower agitation rates (100 and 200 rev./min) where the equilibrium conversion was only achieved for higher normalized residence times, ( ⫽ ET 䡠 t/V, where ET is the amount of immobilized protein, t is the time and V is the liquid volume) (Fig. 2). The increase of agitation rate considerably enhanced the initial esterification rate up to 300 rev./min (Fig. 3). Above this value, these differences were less pronounced, which showed that at higher agitation rates (300 –500 rev./min) the
interfacial area and, consequently, the esterification reaction were independent of the agitation rate. At low agitation rates, diffusional (mass transfer) and hydrodynamic (interfacial area) effects can be responsible for the lower initial esterification rate values achieved. In fact, at 100 and 200 rev./min, the bioreactor observation showed a clear separation between the aqueous and the organic phases. Based on the results referred above, an agitation rate of 400 rev./min was selected for the subsequent studies in the PRD bioreactor. The observed behavior at this agitation rate was compared to the one obtained when the esterification reaction, using the same initial substrates and immobilized enzyme concentration, was carried out in 100-ml Erlenmeyer flasks (batch reactor), at a temperature of 30°C and a shaking rate of 150 rev./min (not shown here). In both cases, an equilibrium conversion of 57% was achieved, at the same normalized residence time. Similar initial esterification rates were also observed being the values of 0.915 and 1.06 mmol/mg protein 䡠 h obtained, respectively, for the PRD with total recirculation and the batch bioreactor. The comparison of results attained in both bioreactors, PRD and batch, for the initial rate and the conversion degree, allowed the use of the apparent kinetic constants previously determined for the immobilized enzyme in the batch reactor [1]. In this way, the apparent kinetic constants values used in the integrated Michaelis–Menten equation (equation (1)) to model the experimental results obtained when the PRD bioreactor was operated with total recirculation were: a catalytic constant, kcat ⫽ 5.78 mmol/h 䡠 mg protein; a Michaelis constant for ethanol, Km(Et) ⫽ 1.20 M; a Michaelis constant for oleic acid, Km(Ol)) ⫽ 1.16 ⫻ 10⫺8 M, and a dissociation constant of the ethanol-lipase complex, K(Et) ⫽ 9.46 ⫻ 107 M [1]. The integrated equation, considering that the reaction follows a ternary complex mechanism [1], is given by: k cat ⫽ [Et]0X⫺K m(Et) ln(1 ⫺ X) ⫺ K m(O1) 䡠
冉
冊
[Et]0 K (Et)K m(O1) 1 ln(1 ⫺ X) ⫹ ⫺1 [O1]0 [O1]0 1⫺X
(1)
where is the normalized residence time (ET 䡠 t/V), X is the conversion degree, [Et]0 and [Ol]0 are, respectively, the initial ethanol and oleic acid concentration. The model satisfactorily predicts the experimental results up to the equilibrium (Fig. 4) but not above this point, as the reverse reaction was not considered. 3.2. Continuous operation
Fig. 3. Effect of agitation rate on the initial esterification rate obtained in the PRD bioreactor with total recirculation. Reaction conditions as in Fig. 2.
The perforated rotating disc bioreactor was operated continuously by constant addition of aqueous (0.64 M ethanol) and organic (1.55 M oleic acid) phase, after the equilibrium conversion degree (⬃57%) having been achieved. For this purpose, the esterification reaction was initially performed with total recirculation conditions, during 24 h. An agitation
A.C. Oliveira et al. / Enzyme and Microbial Technology 26 (2000) 446 – 450
449
grated equation of a plug flow reactor [PFR; Eq. (4)]. In both cases, it was accounting that the reaction follows a ternary complex mechanism and that the apparent kinetic constants are the ones previously referred: kcat ⫽ 5.78 mmol/h 䡠 mg protein, Km(Et) ⫽ 1.20 M, Km(Ol) ⫽ 1.16 ⫻ 10⫺8 M, and K(Et) ⫽ 9.46 ⫻ 107 M [1]. CSTR:
K m(Et)X K m(O1)[Et]0X k catET ⫽ [Et]0X ⫹ ⫹ Q (1 ⫺ X) [O1]0(1 ⫺ X) ⫹
Fig. 4. Progress of ethanol conversion degree in the perforated rotating disc bioreactor with total recirculation. Symbols represent the experimental data points and the line is the kinetic model prediction. Reaction conditions as in Fig. 2, with 400 rev./min for agitation rate.
PFR:
K (Et)K m(O1)X [O1]0(1 ⫺ X)2
k catET ⫽ [Et]0X ⫺ K m(Et) ln(1 ⫺ X) Q ⫺ K m(O1)
rate of 400 rev./min was the selected value, as previously referred. The conversion degree was determined in a long-term continuous operation to assess the enzyme stability (Fig. 5). The experimental data obtained at a flow rate of 0.035 and 0.07 ml/min showed that, in both cases, a conversion degree decrease in time was observed due, probably, to enzyme thermal deactivation. To model the observed behavior, the first order deactivation model was considered: ETt ⫽ ET0e⫺kdt
⫹
[Et]0 ln(1 ⫺ X) [O1]0
冉
K (Et)K m(O1) 1 ⫺1 [O1]0 1⫺X
冊
CSTR:
ln
冤
K m(Et) X0 K m(O1) [Et]0X0 K (Et) K m(O1)X0 ⫹ ⫹ (1 ⫺ X0) [O1]0 (1 ⫺ X0) [O1]0 (1 ⫺ X0)2 K m(Et) Xt K m(O1) [Et]0Xt K (Et) K m(O1)Xt [Et]0 Xt ⫹ ⫹ ⫹ (1 ⫺ Xt) [O1]0(1 ⫺ Xt) [O1]0 (1 ⫺ Xt)2
[Et]0 X0 ⫹
⫽ k d(t t ⫺ t 0)
冥
(5)
PFR:
冤
[Et]0 X0 ⫺ K m(Et) ln(1 ⫺ X0) ⫺ K m(O1) ln(1 ⫺ X0) ⫹
冉
冊
⫹
K(Et) Km(O1) 1 ⫺1 [O1]0 1 ⫺ Xt
冉
冊
K (Et)K m(O1) 1 ⫺1 [O1]0 1 ⫺ X0 ln [Et]0 Xt ⫺ Km(Et) ln(1 ⫺ Xt) ⫺ Km(O1) ln(1 ⫺ Xt)
⫽ k d(t t ⫺t 0)
Fig. 5. Modeling of the experimental results obtained in the continuous perforated rotating disc bioreactor with a flow rate of 0.035 ml/min (䡺) and 0.07 ml/min (f). The symbols represent the experimental data points and the line is the model prediction. Reaction conditions as in Fig. 2, with 400 rev./min for agitation rate.
(4)
By using the exponential model (equation (2)) and the integrated equations described above, and after some mathematical manipulation, the following equations describing the reactor performance with enzyme deactivation can be achieved:
(2)
where ETt and ET0 are the total enzyme concentration, respectively, at t ⫽ t and t ⫽ 0; kd is the deactivation constant and t is the operation time. The half-life time (time after which half activity was lost), defined as t1/2 ⫽ ln 2/k d, was used as a stability indicator. It was also analyzed if the PRD bioreactor performance was better described by the integrated equation of a continuous stirred tank reactor [CSTR; Eq. (3)] or by the inte-
(3)
冥 (6)
where X0 and Xt are, respectively, the conversion degree obtained at t ⫽ 0 and t ⫽ t. These equations were then applied to model the experimental data allowing the determination of the deactivation constant values reported in Table 1. As it can be seen, there is no significant differences between the CSTR and the PFR in the decay of immobilized enzyme activity, which indicates that the ethanol/oleic acid esterification reaction has a zero-order reaction kinetics [10] under the substrate concentrations used.
450
A.C. Oliveira et al. / Enzyme and Microbial Technology 26 (2000) 446 – 450
Table 1 Deactivation constant (kd) and half-life time (t1/2) values considering a continuous stirred tank reactor (CSTR) or a plug flow reactor (PFR).
CSTR PFR
Flow rate (ml/min)
kd (h⫺1)
t1/2 (h)
r2
0.035 0.070 0.035 0.070
0.0034 0.0031 0.0028 0.0028
203.9 223.6 247.6 247.6
0.9859 0.9098 0.9884 0.9214
vation model and the integrated equations of a CSTR and a PFR were used to predict the observed behavior. Both equations lead to similar results, which indicates a zero order reaction. A deactivation constant of about 0.003 h⫺1 and an half-life time of about 10 days were obtained in the present system.
References
r2 represents the square of correlation coefficient
A good correlation between the experimental data and the model was found when using the lower flow rate (0.035 ml/min) (Table 1, Fig. 5). However, at a flow rate of 0.07 ml/min (Fig. 5) the model did not describe the experimental results up to 48 h of continuous operation as well as in the previous case. This could be probably due to some initial loss of immobilized enzyme by the organic phase outlet because the small dimension support particles may have been washed out, at this flow rate. Concerning the lipase operational stability, the half-life time values obtained (Table 1) showed that this parameter is independent of the flow rate, being almost the same in all cases. In general, the lipase lost 50% of its initial activity in about 10 days of operation. The mechanical agitation could probably be one of the factors that lead to the deactivation profile achieved.
4. Conclusions The performance of the PRD bioreactor operating with total recirculation allowed to select an agitation rate of 400 rev./min to be used in later studies. The experimental results obtained with the continuous PRD bioreactor showed an enzymatic activity decay, which was independent of the flow rate. The exponential deacti-
[1] Oliveira AC, Rosa MF, Aires–Barros MR, Cabral JMS. Enzymatic esterification of ethanol and oleic acid—a kinetic study. 1999 (submitted). [2] Oliveira AC, Cabral JMS. Kinetic studies of Mucor miehei lipase in phosphatidylcholine microemulsions. J Chem Tech Biotechnol 1993; 56:247–52. [3] Oliveira AC, Rosa MF, Cabral JMS, Aires–Barros MR. Immobilization of Saccharomyces cerevisiae cells and Rhizomucor miehei lipase for the production and extractive biocatalysis of ethanol. Bioprocess Engineering 1997;16:349 –53. [4] Oliveira AC, Rosa MF, Cabral JMS, Aires–Barros MR. Improvement of alcoholic fermentations by simultaneous extraction and enzymatic esterification of ethanol. J Molec Catalysis B: Enzymatic 1998;5:29 – 33. [5] Csa´nyi E, Be´lafi–Bako´ K, Harasek M, Sisak CS, Gubicza L. Ethanol removal from fermentation broth by enzymatic esterification. Med Fac Landbouww Univ Gent (Proceedings of Twelfth Forum for Applied Biotechnology) 1998;63/4b:1549 –52. [6] Carneiro– da–Cunha MG, Aires–Barros MR, Tambourgi EB, Cabral JMS. Recovery of a recombinant cutinase with reversed micelles in a continuous perforated rotating disc contactor. Biotechnol Techn 1994;8:413– 8. [7] Carneiro– da–Cunha MG, Aires–Barros MR, Tambourgi EB, Cabral JMS. Continuous extraction of a recombinant cutinase from Escherichia coli disrupted cells with reversed micelles using a perforated rotating disc contactor. Bioprocess Engineering 1996;15:253– 6. [8] Tambourgi EB, Pereira JAFR, Perazolli LA. The efficiency of a rotating disc contactor. Latin Am Appl Res 1993;23:257– 60. [9] Reman GH. A new efficiency apparatus: the rotating disc contactor. Proc 3rd World Petroleum Congress 1953;121. [10] Veith WR, Venkatasubramanian K, Constantinides A, Davidson B. Design and analysis of immobilized-enzyme flow reactors. Appl Biochem Bioeng 1976;1:221–327.
Enzymatic esterification of ethanol by an immobilised Rhizomucor miehei lipase in a perforated rotating disc bioreactor A.C. Oliveiraa,*, M.F. Rosaa, M.R. Aires–Barrosb, J.M.S. Cabralb a
INETI—Departamento de Energias Renova´veis, Estrada do Pac¸o do Lumiar, 1649-038 Lisbon, Portugal b IST—Centro de Engenharia Biolo´gica e Quı´mica, Av. Rovisco Pais, 1049-001 Lisbon, Portugal Received 22 September 1999; received in revised form 22 November 1999; accepted 4 December 1999
Abstract A perforated rotating disc bioreactor was developed to perform the esterification of ethanol with oleic acid, catalyzed by a lipase from Rhizomucor miehei immobilized by adsorption on to a hydrophobic support—Accurel EP700. The bioreactor with total recirculation operated at an optimum agitation rate of 400 rev./min. The experimental results, in this condition, were predict by a kinetic model using the constants obtained in the batch (Erlenmeyer flasks) assays: a catalytic constant, kcat ⫽ 5.78 mmol/h 䡠 mg protein; a Michaelis constant for ethanol, Km(Et) ⫽ 1.20 M; a Michaelis constant for oleic acid, Km(Ol) ⫽ 1.16 ⫻ 10⫺8 M, and a dissociation constant of the ethanol-lipase complex, K(Et) ⫽ 9.46 ⫻ 107 M. The efficiency of conversion gradually decreased during continuous operation of the reactor. The enzymatic activity decayed according to a first order deactivation model and the integrated equations of a continuous stirred tank reactor (CSTR) and a plug flow reactor (PFR). A half-life time of the lipase of about 10 days and a deactivation constant of 0.003 h⫺1 were obtained in the present system. © 2000 Elsevier Science Inc. All rights reserved. Keywords: Enzymatic esterification; Rhizomucor miehei lipase; Perforated rotating disc bioreactor
1. Introduction The enzymatic esterification of ethanol and oleic acid has been performed in a biphasic system [1– 4], with free and immobilized Rhizomucor miehei lipase as the esterification reaction catalyst. The use of immobilized enzyme has some industrial and economical advantages such as ease recovery and re-use, greater stability of the enzyme, and possibility of continuous operation. However, the esterification reaction referred above has been performed in a batch reactor that is not the most advantageous process if the objective is to work in large scale. In this case, the use of continuous processes will be preferable. With this purpose, some studies of the ethanol/oleic acid esterification reaction were carried out by Csa´nyi et al. [5] in a hollow fiber membrane bioreactor where the lipase was immobilized by adsorption onto the inner wall of the fibers. Perforated rotating disc contactor (PRDC) has been successfully applied as a continuous extraction equipment for
protein and enzyme, using reversed micellar systems [6,7]. With this apparatus, high-mass transfer rates between the phases are obtained, leading to high separation efficiencies [8]. This PRDC with free-flow areas has been industrially used as an alternative to the conventional extraction equipment [8,9]. The aim of this work was to use a type of reactor based on the advantages of PRDC to develop a novel process for the continuous enzymatic esterification of ethanol and oleic acid, using as catalyst a Rhizomucor miehei lipase immobilized by adsorption onto a hydrophobic support—Accurel EP700. The described system could take part in an integrated bioprocess consisting of a continuous fermentation followed by a continuous extraction of the inhibitory product (ethanol).
2. Materials and methods 2.1. Enzyme and chemicals
* Corresponding author. Tel.: ⫹351-1-712-72-11; fax: ⫹351-1-712-71 95. E-mail address: [email protected] (A.C. Oliveira)
A commercial lipase from R. miehei, Palatase M 1000 L, was a kind gift of Novo–Nordisk (Baegsraend, Denmark).
0141-0229/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S 0 1 4 1 - 0 2 2 9 ( 0 0 ) 0 0 1 8 5 - 4
A.C. Oliveira et al. / Enzyme and Microbial Technology 26 (2000) 446 – 450
447
The amount of protein immobilized onto the carrier was determined by a modified Folin assay and it is expressed per gram of support. 2.4 Experimental procedure The biphasic system used consisted of an organic phase, the oleic acid, and of an aqueous phase composed by phthalate-NaOH buffer (110 mM, pH 5.5) with ethanol. Initially, the column was filled with 55 ml of oleic acid (oleic acid concentration ⫽ 1.55 M) and 55 ml of aqueous phase (ethanol concentration ⫽ 0.64 M), at which the immobilized enzyme was added. Fig. 1. (A), perforated rotating disc bioreactor: 1, aqueous phase inlet; 2, organic phase outlet; 3, aqueous phase outlet; 4, organic phase inlet; 5, refrigerated water. (B), perforated disc.
The immobilization support, a micropouros polyamide powder (Accurel EP700; particle size 300 –1000 m, void volume 75%, pore size 5 to 30 nm) from Akzo (Obernburg, Germany) was a kind gift of S.C.I.E. Irma˜os Planas Almasque Lda (Lisbon, Portugal). All of the reagents used were of analytical grade from Merck (Damstadt, Germany). 2.2. Experimental apparatus The esterification reactions were carried out in a perforated rotating disc (PRD) bioreactor (Fig. 1) made of Perspex tube 32 mm internal diameter (i.d.) and 160 mm high. Four perforated discs equally separated were mounted on a central shaft, which was rotated at different velocities (rev./ min). The perforated discs had a diameter of 30 mm and were drilled with six holes of 7.5-mm diameter and six holes of 1.5-mm diameter. The reactor was maintained at 30°C using water in a refrigerated jacket. To prevent the immobilized lipase from leaving the PRD bioreactor, a tight porous filter net was provided at the reactor organic phase outlet.
2.4.1. Batch operation The effect of agitation rate (100 –500 rev./min) on the ethanol/oleic acid esterification reaction, using the immobilized enzyme, was carried out in the PRD bioreactor with total recirculation. Samples of the organic phase were taken at similar residence times and the ester concentration was analyzed by CGC. The initial reaction rates were estimated from the slope of plots of ethyl esters produced versus residence time and expressed as mmol/h 䡠 mg of immobilized protein. 2.4.2. Continuous operation The start-up of the reactor included, initially, the operation with total recirculation during 24 h, at a constant agitation rate, following the continuous operation after this period of time. The aqueous and the organic phases get into the reactor in counter-current flow as shown in Fig. 1. The flow rates of the two phases were maintained constant by using peristaltic pumps with a flow rate of 0.035 or 0.07 ml/min, depending on the assay. The samples were collected in the organic phase outlet, at regular time intervals, and the amount of ester produced, measured by CGC, was used to calculated the ethanol conversion degree obtained. 2.5. Ethyl ester determination
2.3. Preparation of the immobilized enzyme The lipase was immobilized by adsorption on Accurel EP700. For the bioreactor experiments, 3 ml of the lipase solution (10.2 mg of protein/ml of enzyme solution) was added to 900 mg of the immobilization support. After being vortex-mixed for 1 min, the enzyme-support contact time was 1 h [3], at room temperature. At this time, 10 ml of buffer (110 mM phthalate-NaOH buffer, pH 5.5) was added to the preparation and the support was separated by filtration (Millipore system) and washed twice with buffer.
Ethyl esters produced during the bioreactor operation were quantified by capillary gas chromatography (CGC) by using a Varian 3300 chromatograph equipped with a FID and a Supelcowax 10 column (i.d. ⫽ 0.32 mm; 30 m; film thickness ⫽ 0.25 m). Helium was used as the carrier gas and the injector and detector temperatures were kept at 250°C and 260°C, respectively. The oven temperature was kept at 200°C for 11.5 min and then heated to 225°C. The organic phase samples (20 mg) were analyzed directly, after the addition of methyl heptadecanoate as internal standard (0.53 mg).
448
A.C. Oliveira et al. / Enzyme and Microbial Technology 26 (2000) 446 – 450
Fig. 2. Effect of agitation rate on ester production when using the perforated rotating disc bioreactor with total recirculation. Agitation rate (rev./ min): 〫, 100; F, 200; 䡺, 300; E, 400; ‚, 500. Reaction conditions: organic/aqueous phases ratio of 1; initial ethanol and oleic acid concentration, respectively, 0.64 and 1.55 M; immobilized enzyme ⫽ 3.09 mg/g support.
3. Results and discussion 3.1. Batch operation The influence of the agitation rate on the enzymatic esterification of ethanol and oleic acid performed in a perforated rotating disc bioreactor with total recirculation was evaluated. Ethanol and oleic acid were initially at 0.64 and 1.55 M respectively, and 3.09 mg of immobilized protein (R. miehei lipase)/g support was used as initial reaction conditions. Different performances were obtained when using different agitation rates, namely for lower agitation rates (100 and 200 rev./min) where the equilibrium conversion was only achieved for higher normalized residence times, ( ⫽ ET 䡠 t/V, where ET is the amount of immobilized protein, t is the time and V is the liquid volume) (Fig. 2). The increase of agitation rate considerably enhanced the initial esterification rate up to 300 rev./min (Fig. 3). Above this value, these differences were less pronounced, which showed that at higher agitation rates (300 –500 rev./min) the
interfacial area and, consequently, the esterification reaction were independent of the agitation rate. At low agitation rates, diffusional (mass transfer) and hydrodynamic (interfacial area) effects can be responsible for the lower initial esterification rate values achieved. In fact, at 100 and 200 rev./min, the bioreactor observation showed a clear separation between the aqueous and the organic phases. Based on the results referred above, an agitation rate of 400 rev./min was selected for the subsequent studies in the PRD bioreactor. The observed behavior at this agitation rate was compared to the one obtained when the esterification reaction, using the same initial substrates and immobilized enzyme concentration, was carried out in 100-ml Erlenmeyer flasks (batch reactor), at a temperature of 30°C and a shaking rate of 150 rev./min (not shown here). In both cases, an equilibrium conversion of 57% was achieved, at the same normalized residence time. Similar initial esterification rates were also observed being the values of 0.915 and 1.06 mmol/mg protein 䡠 h obtained, respectively, for the PRD with total recirculation and the batch bioreactor. The comparison of results attained in both bioreactors, PRD and batch, for the initial rate and the conversion degree, allowed the use of the apparent kinetic constants previously determined for the immobilized enzyme in the batch reactor [1]. In this way, the apparent kinetic constants values used in the integrated Michaelis–Menten equation (equation (1)) to model the experimental results obtained when the PRD bioreactor was operated with total recirculation were: a catalytic constant, kcat ⫽ 5.78 mmol/h 䡠 mg protein; a Michaelis constant for ethanol, Km(Et) ⫽ 1.20 M; a Michaelis constant for oleic acid, Km(Ol)) ⫽ 1.16 ⫻ 10⫺8 M, and a dissociation constant of the ethanol-lipase complex, K(Et) ⫽ 9.46 ⫻ 107 M [1]. The integrated equation, considering that the reaction follows a ternary complex mechanism [1], is given by: k cat ⫽ [Et]0X⫺K m(Et) ln(1 ⫺ X) ⫺ K m(O1) 䡠
冉
冊
[Et]0 K (Et)K m(O1) 1 ln(1 ⫺ X) ⫹ ⫺1 [O1]0 [O1]0 1⫺X
(1)
where is the normalized residence time (ET 䡠 t/V), X is the conversion degree, [Et]0 and [Ol]0 are, respectively, the initial ethanol and oleic acid concentration. The model satisfactorily predicts the experimental results up to the equilibrium (Fig. 4) but not above this point, as the reverse reaction was not considered. 3.2. Continuous operation
Fig. 3. Effect of agitation rate on the initial esterification rate obtained in the PRD bioreactor with total recirculation. Reaction conditions as in Fig. 2.
The perforated rotating disc bioreactor was operated continuously by constant addition of aqueous (0.64 M ethanol) and organic (1.55 M oleic acid) phase, after the equilibrium conversion degree (⬃57%) having been achieved. For this purpose, the esterification reaction was initially performed with total recirculation conditions, during 24 h. An agitation
A.C. Oliveira et al. / Enzyme and Microbial Technology 26 (2000) 446 – 450
449
grated equation of a plug flow reactor [PFR; Eq. (4)]. In both cases, it was accounting that the reaction follows a ternary complex mechanism and that the apparent kinetic constants are the ones previously referred: kcat ⫽ 5.78 mmol/h 䡠 mg protein, Km(Et) ⫽ 1.20 M, Km(Ol) ⫽ 1.16 ⫻ 10⫺8 M, and K(Et) ⫽ 9.46 ⫻ 107 M [1]. CSTR:
K m(Et)X K m(O1)[Et]0X k catET ⫽ [Et]0X ⫹ ⫹ Q (1 ⫺ X) [O1]0(1 ⫺ X) ⫹
Fig. 4. Progress of ethanol conversion degree in the perforated rotating disc bioreactor with total recirculation. Symbols represent the experimental data points and the line is the kinetic model prediction. Reaction conditions as in Fig. 2, with 400 rev./min for agitation rate.
PFR:
K (Et)K m(O1)X [O1]0(1 ⫺ X)2
k catET ⫽ [Et]0X ⫺ K m(Et) ln(1 ⫺ X) Q ⫺ K m(O1)
rate of 400 rev./min was the selected value, as previously referred. The conversion degree was determined in a long-term continuous operation to assess the enzyme stability (Fig. 5). The experimental data obtained at a flow rate of 0.035 and 0.07 ml/min showed that, in both cases, a conversion degree decrease in time was observed due, probably, to enzyme thermal deactivation. To model the observed behavior, the first order deactivation model was considered: ETt ⫽ ET0e⫺kdt
⫹
[Et]0 ln(1 ⫺ X) [O1]0
冉
K (Et)K m(O1) 1 ⫺1 [O1]0 1⫺X
冊
CSTR:
ln
冤
K m(Et) X0 K m(O1) [Et]0X0 K (Et) K m(O1)X0 ⫹ ⫹ (1 ⫺ X0) [O1]0 (1 ⫺ X0) [O1]0 (1 ⫺ X0)2 K m(Et) Xt K m(O1) [Et]0Xt K (Et) K m(O1)Xt [Et]0 Xt ⫹ ⫹ ⫹ (1 ⫺ Xt) [O1]0(1 ⫺ Xt) [O1]0 (1 ⫺ Xt)2
[Et]0 X0 ⫹
⫽ k d(t t ⫺ t 0)
冥
(5)
PFR:
冤
[Et]0 X0 ⫺ K m(Et) ln(1 ⫺ X0) ⫺ K m(O1) ln(1 ⫺ X0) ⫹
冉
冊
⫹
K(Et) Km(O1) 1 ⫺1 [O1]0 1 ⫺ Xt
冉
冊
K (Et)K m(O1) 1 ⫺1 [O1]0 1 ⫺ X0 ln [Et]0 Xt ⫺ Km(Et) ln(1 ⫺ Xt) ⫺ Km(O1) ln(1 ⫺ Xt)
⫽ k d(t t ⫺t 0)
Fig. 5. Modeling of the experimental results obtained in the continuous perforated rotating disc bioreactor with a flow rate of 0.035 ml/min (䡺) and 0.07 ml/min (f). The symbols represent the experimental data points and the line is the model prediction. Reaction conditions as in Fig. 2, with 400 rev./min for agitation rate.
(4)
By using the exponential model (equation (2)) and the integrated equations described above, and after some mathematical manipulation, the following equations describing the reactor performance with enzyme deactivation can be achieved:
(2)
where ETt and ET0 are the total enzyme concentration, respectively, at t ⫽ t and t ⫽ 0; kd is the deactivation constant and t is the operation time. The half-life time (time after which half activity was lost), defined as t1/2 ⫽ ln 2/k d, was used as a stability indicator. It was also analyzed if the PRD bioreactor performance was better described by the integrated equation of a continuous stirred tank reactor [CSTR; Eq. (3)] or by the inte-
(3)
冥 (6)
where X0 and Xt are, respectively, the conversion degree obtained at t ⫽ 0 and t ⫽ t. These equations were then applied to model the experimental data allowing the determination of the deactivation constant values reported in Table 1. As it can be seen, there is no significant differences between the CSTR and the PFR in the decay of immobilized enzyme activity, which indicates that the ethanol/oleic acid esterification reaction has a zero-order reaction kinetics [10] under the substrate concentrations used.
450
A.C. Oliveira et al. / Enzyme and Microbial Technology 26 (2000) 446 – 450
Table 1 Deactivation constant (kd) and half-life time (t1/2) values considering a continuous stirred tank reactor (CSTR) or a plug flow reactor (PFR).
CSTR PFR
Flow rate (ml/min)
kd (h⫺1)
t1/2 (h)
r2
0.035 0.070 0.035 0.070
0.0034 0.0031 0.0028 0.0028
203.9 223.6 247.6 247.6
0.9859 0.9098 0.9884 0.9214
vation model and the integrated equations of a CSTR and a PFR were used to predict the observed behavior. Both equations lead to similar results, which indicates a zero order reaction. A deactivation constant of about 0.003 h⫺1 and an half-life time of about 10 days were obtained in the present system.
References
r2 represents the square of correlation coefficient
A good correlation between the experimental data and the model was found when using the lower flow rate (0.035 ml/min) (Table 1, Fig. 5). However, at a flow rate of 0.07 ml/min (Fig. 5) the model did not describe the experimental results up to 48 h of continuous operation as well as in the previous case. This could be probably due to some initial loss of immobilized enzyme by the organic phase outlet because the small dimension support particles may have been washed out, at this flow rate. Concerning the lipase operational stability, the half-life time values obtained (Table 1) showed that this parameter is independent of the flow rate, being almost the same in all cases. In general, the lipase lost 50% of its initial activity in about 10 days of operation. The mechanical agitation could probably be one of the factors that lead to the deactivation profile achieved.
4. Conclusions The performance of the PRD bioreactor operating with total recirculation allowed to select an agitation rate of 400 rev./min to be used in later studies. The experimental results obtained with the continuous PRD bioreactor showed an enzymatic activity decay, which was independent of the flow rate. The exponential deacti-
[1] Oliveira AC, Rosa MF, Aires–Barros MR, Cabral JMS. Enzymatic esterification of ethanol and oleic acid—a kinetic study. 1999 (submitted). [2] Oliveira AC, Cabral JMS. Kinetic studies of Mucor miehei lipase in phosphatidylcholine microemulsions. J Chem Tech Biotechnol 1993; 56:247–52. [3] Oliveira AC, Rosa MF, Cabral JMS, Aires–Barros MR. Immobilization of Saccharomyces cerevisiae cells and Rhizomucor miehei lipase for the production and extractive biocatalysis of ethanol. Bioprocess Engineering 1997;16:349 –53. [4] Oliveira AC, Rosa MF, Cabral JMS, Aires–Barros MR. Improvement of alcoholic fermentations by simultaneous extraction and enzymatic esterification of ethanol. J Molec Catalysis B: Enzymatic 1998;5:29 – 33. [5] Csa´nyi E, Be´lafi–Bako´ K, Harasek M, Sisak CS, Gubicza L. Ethanol removal from fermentation broth by enzymatic esterification. Med Fac Landbouww Univ Gent (Proceedings of Twelfth Forum for Applied Biotechnology) 1998;63/4b:1549 –52. [6] Carneiro– da–Cunha MG, Aires–Barros MR, Tambourgi EB, Cabral JMS. Recovery of a recombinant cutinase with reversed micelles in a continuous perforated rotating disc contactor. Biotechnol Techn 1994;8:413– 8. [7] Carneiro– da–Cunha MG, Aires–Barros MR, Tambourgi EB, Cabral JMS. Continuous extraction of a recombinant cutinase from Escherichia coli disrupted cells with reversed micelles using a perforated rotating disc contactor. Bioprocess Engineering 1996;15:253– 6. [8] Tambourgi EB, Pereira JAFR, Perazolli LA. The efficiency of a rotating disc contactor. Latin Am Appl Res 1993;23:257– 60. [9] Reman GH. A new efficiency apparatus: the rotating disc contactor. Proc 3rd World Petroleum Congress 1953;121. [10] Veith WR, Venkatasubramanian K, Constantinides A, Davidson B. Design and analysis of immobilized-enzyme flow reactors. Appl Biochem Bioeng 1976;1:221–327.