steel filter was introduced and its application for continu- ous ethanol fermentation was investigated. The filter performance was highly influenced by agitation ...
Cell Retention Culture with an Internal Filter Module: Continuous Ethanol Fermentation Ho Nam Chang,* Woo Gi Lee, and Beom So0 Kim Department of Chemical Engineering and Bioprocess Engineering Research Center, Korea Advanced Institute of Science and Technology, Taejon 305-701, South Korea Received May 21, 1992/Accepted October 20, 1992
A new internal filter feedback system with a stainless
steel filter was introduced and its application for continuous ethanol fermentation was investigated. The filter performance was highly influenced by agitation speed and yeast concentration. Retention coefficient with a filter of 2 p m pore size was found more than 97.5%, and the filter was suitable for yeast separation. Maximum yeast concentration was 157 g/L and the best operable cell concentration was between 90 and 150 g/L, which was similar to that obtained in the external membrane cell recycle culture. The cell concentration in the fermentor was maintained by manipulation of dilution rate and bleed ratio with the growth rate. The internal filter feedback system was successfully operated for more than 10 days. This study shows that the internal filter feedback system with a stainless steel filter can be used as successfully a s an external cell recycle system for high-density cell culture and ethanol fermentation. Furthermore, it can be scaled up more easily than the external Cell recycle system. 0 1993 John Wiley 84 Sons, Inc. Key words: internal filter feedback system high-density cell culture continuous ethanol fermentation
INTRODUCTION Achieving a high productivity in a bioreactor plays a crucial role in determining the economics of bulk biochemical products such as ethanol, amino acids, and single-cell proteins. The high-cell-density culture coupled with continuous operation should yield the desired theoretical productivity of the bioreactor, but in practice this has seldom been realized in bioreactor operation. In conventional continuous operation cells in the reactor will be washed out unless the dilution rate is kept lower than the growth rate of the cells. Maintaining a higher biomass concentration in a bioreactor can be accomplished by cell imm~bilization'~'~ and cell r e ~ y c l i n gHowever, .~ cell recycling is often preferred to cell immobilization since achieving desired oxygen transfer and stable reactor operation are more difficult for the latter. Cell recycling relies on one of three methods: sedimentation,2,6,s centrifugation,' and membrane f i l t r a t i ~ n . ~ ~ ~Membrane -'~.'~ cell recycle has been a very popular method in achieving a high bioreactor productivity. Recently, two-stage culture of recombinant deoxyribonucleic acid (DNA) cells was carried out using membrane cell recy~1e.I~ However, this process has several drawbacks that have delayed its industrial applications: (1) industrial substrates contain many particles *
To whom all correspondence should be addressed.
Biotechnology and Bioengineering, Vol. 41, Pp. 677-681 (1993) 0 1993 John Wiley & Sons, Inc.
which make pumping through external membrane device difficult; (2) oxygen supply and carbon dioxide removal may not be adequate while the broth is in the recycling loop; (3) sterilization of the external membrane device is difficult; and (4) recirculation of the broth requires pumps and additional energy for the operation. To overcome these problems, we have employed an internal filter feedback reactor, which allowed microbial separation to be carried out inside the fermentor. In this article we describe the design, operation, and performance of a high-cell-density culture system using the internal filter feedback reactor. Ethanol fermentation by Saccharomyces cerevisiae was studied as a model system.
MATERIALS AND METHODS Microorganism The yeast strain used in this study was Saccharomyces cerevisiae ATCC 24858.
Culture Medium The growth medium consisted of 5% glucose, 0.3% yeast extract, 0.3% malt extract, and 0.5% bacto peptone. The fermentation medium consisted of 100 g/L glucose, 8.5 g/L yeast extract, 1.3 g/L NH4C1, 0.12 g/L MgS04 7H20, and 0.06 g/L CaC12. Glucose used in the fermentation medium was of commercial grade (purity 99%, Miwon Co., Seoul, Korea), and other chemicals used were of reagent grade. The medium was sterilized for 2 0 4 0 min at 121°C. Glucose and the other medium components were sterilized separately to prevent caramelization.
-
Filter Module The filter module used in this study is shown in Figure la. The filter material was porous stainless steel with pore sizes of 2 or 10 p m (Cuno Co., USA). The filter module consisted of 13 vertical cylindrical filter rods with inner diameter, outer diameter, and height of 7.5, 9.0, and 120 mm, respectively, and an upper frame of stainless steel. The total surface area of the filter module was ca. 440 cm2. After the experiments the filter module was separated from the fermentor and cleaned with 1 N NaOH for several
CCC 0006-3592/93/060677-05
hours. I t was washed and backflushed with distilled water prior to reuse. After these treatments, the filter performance was checked to ensure that the distilled water flux returned to the original level.
Filter Performance Test The filter testing system consisted of a reservoir equipped with the filter module and two pumps with a filter chamber connected via a Tygon tube (Fig. lb). The pressure gauge WBS located between the reservoir and the pump. The fluid in the reservoir was removed through the filter module by the suction pump (Masterflex model no. 7535-10, ColeParmer) and then filtrate was recirculated back to the reservoir for homogeneity. Cultivated yeast cells in the fermentor were centrifuged and the aliquots were removed. The cells were remixed with a buffer saline solution which consisted of 8.5 g/L NaCI, 6 g/L NaHzPO4, and 3 g/L KHzP04. Experiments were performed at pH 7 and room temperature (25°C). Samples were removed at about 5-min intervals. A steady state was assumed when filtrate flux leveled off as evidenced by assaying several successive samples. The retention coefficient (R, in percentage) is defined as follows:
R=100X
(1 - - 3
Reactor System and Operation for Ethanol Fermentation The experimental setup for the internal filter feedback system is shown in Figure Ic. The fermentor used was a 1.5-L-capacity closed vessel of 1.0 L working volume (Bioflo model C30, New Brunswick Scientific Co, USA), in which the filter cartridge was located. Medium feed rate was adjusted by a peristaltic pump. The level of the fermentor was controlled at a fixed height so that the total inlet flow rate was exactly equal to the total outlet flow rate. Fresh medium was continuously supplied to the fermentor by a peristaltic pump, and at the same time the product was removed through the filter module by the suction pump. Part of the broth was continuously bled through the bleed port to control cell concentration in the fermentor." The inoculum was prepared by growing cells aerobically in a 500-mL flask containing 100 mL growth medium in a rotary shaker incubator for 24 h at 30°C before transfer to the fermentor. Inoculation was made with 10% (v/v) fermentation working volume and the culture was grown batchwise until substrate was limited. The fermentor and filter module were simultaneously sterilized for 40 min at 121°C. Fermentation was carried out at 30°C and the pH was controlled at 4.5 by automatic addition of 2 N NaOH and 2 N HCI. To maintain cell viability, filter sterilized air was supplied to the fermentor.
Assays
where C , and C , are concentrations of the cells in the filtrate and in the retentate, respectively. If the yeast cells are completely retained by the membrane, the retention coefficient R = 1.
Ethanol was measured by gas chromatography (Chromosorb W coated with Carbowax 20M) with a flame ionization detector (Varian 3300) using 2-butanol as an internal standard. Glucose was analyzed using the spectrophotometer (Milton Roy Co., 20D) at 525 nm by the glucose oxidase/peroxidase (GODIPOD) enzymatic method (glucose E-kit). Cell concentration was measured using the spectrophotomer at 520 nm. Cell dry weight was determined after centrifuging the
I i (a)
(b)
Figure 1. (a) Module configuration (stainless steel filter: 120 X 9 mm each, No. 13, surface area 440 cm'). (b) Schematic diagram of experimental apparatus f o r filter performance test: (1) fermentor equipped with filter module, (2) pressure gauge, (3) suction pump, (4) filtrate chamber, and
678
(C)
( 5 ) recirculation pump. (c) Schematic diagram of experimental apparatus for continuous culture: (1) medium reservoir, (2) fermentor equipped with filter module, (3) filtrate chamber, (4) blccd chamber, ( 5 ) feed pump, and (6) suction pump.
BIOTECHNOLOGY AND BIOENGINEERING, VOL. 41. NO. 6. MARCH 15 Iqq?
cell suspension twice, washing in distilled water, and drying at 105°C for 1 day.
901
RESULTS AND DISCUSSION Filter Performance The effects of cell concentration, pressure difference, and agitation speed in the fermentor on filtrate flux are shown in Figures 2a, b. The decline in the flux with time is likely to be due to fouling or concentration p~larization.'~ As shown in Figure 2, as long as the yeast concentration at the membrane surface is less than a gel formation concentration, the filtrate flux increases with the pressure difference until the threshold. However, any further increase of pressure difference entails a compaction of the gel layer and increase of its thickness. Thus, there is a compromise between the benefits of increased pressure difference and the negative effects of gel layer compaction. The latter dominates in the process used in this study, and as a result, a small additional flux increase is obtained. For this reason there is an optimal pressure difference for the filtrate flux. This phenomenon is very similar to that obtained with any other external membrane devices. The most significant improvement of filtrate flux could be obtained by keeping the membrane as free of deposits as possible by employing a sufficiently high shear force (Fig. 2b). The decline in the flux with operation time at a higher agitation speed was relatively low because the shear force generated in the reactor was large enough to remove cells deposited on the membrane surface. The agitation speed in the fermentation must be optimized since too high an agitation speed brings about high energy consumption. Thus agitation speed requirement may vary depending on the applications. Generally, the retention coefficient of solutes is determined by the properties of the membrane-like pore size distribution and adsorption characteristics and by operation conditions such as concentration, pH, pressure, and shear force. Figure 3 presents the retention coefficients for two different pore sizes as a function of cell concentration at 700 rpm and a pressure difference of 0.13 bar. The gel layer deposited on the surface of the membrane actually constitutes a dynamic membrane and retains the cells. The retention coefficient generally increases with gel layer thickness. The retention coefficients of the 2-pm-pore-size filter for yeast cells are higher than those of 10 pm. But the membrane pore size has little effect on cell retention as the gel layer becomes thicker. The retention coefficient using the filter of 2 p m pore size was more than 97.5% and the filter was suitable for yeast cell separation.
Total Cell Retention Culture A typical yeast culture profile with an internal filter feedback system using stainless steel filter of 2 p m pore size
Pressure Difference (bar)
(a)
Pressure Difference (bar)
( b)
Figure 2. (a) Flux changes with pressure difference for the different cell concentration at agitation speed of 700 rpm: (m) X = 2 g yeast/l; (A) X = 8.65 g yeas& (0)X = 16.5 g yeast/l; (X) X = 26 g yeast/L. (b) Flux changes withe pressure difference for the different agitation speed, X = 16.5 g yeast/L (B) 700 rpm; (A) 400 rpm; (0) 100 rpm.
is shown in Figure 4. The cells grown in a flask were used as inoculum for the fermentor. Total cell retention culture was started with an initial cell concentration of 4.8 g/L. In the beginning of the fermentation the glucose concentration increased rapidly to 70 g/L due to the low concentration of yeast cells. The glucose concentration gradually decreased and became completely depleted after 12 h. After 12 h, the cell concentration began to increase linearly with glucose consumption having the growth yield coefficient of 0.085 g dry weight/g glucose. The maximum cell concentration obtained in this study was 157 g/L, and operation was successful at cell concentrations between 90 and 150 g/L. This is ca. 9-14 times higher than what can be obtained by chemo~tat.~ There has been a report that in ethanol fermentation with S. cerevisiue ATCC 4126, the cell concentrations were 50 g/L for settling recycle process and 124 g/L for the vacuum fermentation with cell recycle.6 In the membrane cell recycle system, however, reactor operation is possible with a cell concentration above 100
COMMUNICATIONS TO THE EDITOR
679
0
°
7
10
‘0
40
80
120 160 200 240
2
Time (h)
0
10
20
30
40
50 60 Cell Concentration (g/L)
Figure 5. Results of a long-term continuous ethanol fermentation at steady state, D = 0.41 h-’, B = 0.15: (W) cell; (A) ethanol; (13) glucose.
70
Figure 3. Retention coefficient changes with cell concentration for the different pore size of filter at agitation speed of 700 rpm, A P = 0.13 bar: (W) 2 pm; (A) 10 pm.
g/L.4,11 The cell concentration obtained in this study was similar to that in an external cell recycle system, but the internal filter feedback system has significant advantages over the traditional external cell recycle systems, due to the reduced maintenance and operation requirements. The average ethanol concentration of filtrate during substrate limitation was 47.4 g/L and the actual average ethanol yield coefficient (YPls= 0.47) was 92.7% of the theoretical yield coefficient (YPls= 0.51). This value generally ranges between 90 and 95% of the theoretical yield since a portion of the substrate is utilized for the formation of biomass and secondary products. We carried out the total cell retention culture successfully using molasses as the substrate. A cell concentration of 100 g/L was obtained with a reducing sugar concentration of 110 g/L at a dilution rate of 0.26 h-’.
Continuous Culture with Bleed The continuous culture with bleed is beneficial to maintain a cell concentration suitable for operation, since the increase of the cell concentration continues until the operation becomes difficult without bleeding. The steady state cell concentration at constant dilution rate and bleed ratio could be obtained when the rate of cell formation in the fermentor was equal to the bleeding rate of the cells. Figure 5 shows a continuous ethanol kinetics at steady state. A dilution rate of 0.41 h-I and a bleed ratio of 0.15 were used in this experiment. Bleed ratio was taken as the fraction of total effluent stream which contains concentrated cell. Flux was nearly constant. There exists a steady state condition in which the substrate, cell, and ethanol concentrations remain relatively constant.16 A productivity of about 20 g/L h in given conditions was obtained and 5 times that of a classical continuous system: Long-term operation of more than 10 days employing the internal filter feedback system was found to be possible using glucose medium. No significant loss of the separation performance of the filter module was observed. Since there were little or no difficulties in the reactor operation, the internal filter feedback system proved to be efficient and stable over the course of its operation. This suggests that clogging of the membrane pore by cells or other particulates is not a major problem when we use the internal stainless steel filter module in continuous ethanol fermentation. This work was supported by a grant-in-aid from the Ministry of Energy and Resources of Korea and the Korea Science and Engineering Foundation.
0 .-c
L
c
8
References
D
Time (h)
Figure 4. Total cell retention culture of S. cerevisiue, (G1ucose)o 100 g/L, D = 0.55 h-l: (W) cell; (A) ethanol; (0)glucose.
680
=
1. Black, G. M., Webb, C., Mattews, T. M., Atkinson, B. 1984. Practical reactor systems for yeast cell immobilization using biomass support particles. Biotechnol. Bioeng. 26:134- 141. 2. Bu’lock, J. D., Comberbach, D. M., Ghommidh, C. 1984. A study of continuous ethanol production using highly flocculent yeast in the gas lift tower fermenter. Chem. Eng. J. 29B9-B24. 3. Chang, H.N., Furusaki, S. 1991. Membrane bioreactor: Present and prospects. pp. 27-64 In: A. Fiechter (ed.), Advances in biochemical engineeringbiotechnology. Springer-Verlag, Berlin.
BIOTECHNOLOGY AND BIOENGINEERING, VOL. 41, NO. 6, MARCH 15, 1993
4. Cheryan, M., Mehaia, M.A. 1984. Ethanol production in a membrane recycle reactor, conversion of glucose using Saccharomyces cerevisiae. Proc. Biochem. 19204-208. 5 . Cysewski, G. R., Wilke, C. R. 1976. Utilization of cellulosic materials through enzymatic hydrolysis. 1. Fermentation of hydrolysis to ethanol and single cell protein. Biotechnol. Bioeng. 18:1297- 1313. 6. Cysewski, G. R., Wilke, C. R. 1977. Rapid ethanol fermentations using vacuum and cell recycle. Biotechnol. Bioeng. 19:1125- 1143. 7. Cysewski, G.R., Wilke, C.R. 1978. Process design and economic studies of alternative fermentation methods for the production of ethanol. Biotechnol. Bioeng. 20:1421- 1444. 8. Ghose T. K., Tyagi, R. D. 1979. Rapid ethanol fermentation of cellulose hydrolyte. I. Batch versus continuous systems. Biotechnol. Bioeng. 21:1387-1400. 9. Jassens, J.H., Bernard, A., Bailey, R.B. 1984. Ethanol from whey: Continuous fermentation with cell recycle. Biotechnol. Bioeng. 26:1-5. 10. Kroner, K. H., Schutte, H., Hustedt, H., Kula, M. R. 1984. Cross-flow filtration in the downstream processing of enzymes. Proc. Biochem. 19:67-74.
11. Lee, C. W., Chang, H.N. 1987. Kinetics of ethanol fermentations in membrane cell recycle fermentors. Biotechnol. Bioeng. 29:1105- 1112. 12. Margaritis, A., Wilke, C. R. 1978. The rotofermentor 11. Application to ethanol fermentation. Biotechnol. Bioeng. 20727-754. 13. Nagashima, M., Azuma, M., Noguchi, S., Inuzuka, K., Samejima, H. 1984. Continuous ethanol fermentation using immobilized yeast cells. Biotechnol. Bioeng. 26:992-997. 14. Nagata, N., Herouvis, K.J., Dziewulski, P.M., Belfort, G. 1989. Cross-flow membrane microfiltration of a bacterial fermentation broth. Biotechnol. Bioeng. 34:447-466. 15. Park, T.H., Hong, J., Lim, H.C. 1991. Theoretical analysis of the effect of cell recycling on recombinant cell fermentation process. Biotechnol. Prog. 7:77-84. 16. Pirt, S. J., Kurowski, W. M. 1970. An extention of the theory of the chemostat with feed back of organisms. Its experimental realization with a yeast culture. J. Gen. Microbiol. 63:357-366.
COMMUNICATIONS TO THE EDITOR
68 1
