Communication to the Editor Crossflow Filtration of Baker's Yeast with Periodical Stopping of Permeation Flow and Bubbling

Takaaki Tanaka, Hiroshi Itoh, Kazutaka Itoh, and Kazuhiro Nakanishi" Department of Biotechnology, Faculty of Engineering, Oka yama University, Tsushima-Naka, Oka yama 700, Japan

Takaaki Kume and Ryuichi Matsuno Department of Food Science and Technology, Faculty of Agriculture, Kyoto University, Kitashirakawa-Oiwakecho, Sakyo-ku, Kyoto 606, Japan

Received May 2, 1994lAccepted March 17, 1995

The periodical stopping of permeation flow was applied to increase the permeation flux in crossflow filtration of commercially available baker's yeast cell suspension. The permeation flux after 3 h filtration in the crossflow filtration increased to 8 x m3/m2 s (290 Um2 h ) from 2 x m3/m2 s (72 Um2 h ) by applying the periodical stopping of permeation. Introduction of air bubbles dur- ing the stopping period of permeation fur ther increased the flux. 0 1995 John Wiley & Sons, Inc. Key words: crossflow filtration periodical stopping air bubbling baker's yeast

INTRODUCTION

In crossflow filtration of cell suspension, the permeation flux is reduced by deposition of cells on the membrane. Even in the crossflow filtration of baker's yeast cell sus- pension, the permeation flux is reduced by the formation of cell layer (cake) to 3 4 % compared to the water Furthermore, when the cell suspension contains fine parti- cles, the permeation flux decreases considerably even if its content is low. lo There have been reported several methods to improve the permeation flux during the crossflow filtra- tion of microbial cell suspension. The most popular method is backwashing during crossflow filtration."6 Rotating modules are also effective to improve the performance of crossflow f i l t r a t i ~ n . ~ ' ~ Matsumoto et al.7 showed that the intermittent operation of the feed pump was effective to increase the flux during crossflow filtration of model sus- pension. It was also reported that the periodical stopping of permeation flow was effective during the crossflow filtra- tion of microbial broth. '," This method is advantageous to the other methods described since additional equipment is not necessary. However, there have been few reports in which this method is studied in detail. Here, we applied the periodical stopping of the permeation flow to the crossflow filtration of suspension of commercially available baker's

* To whom all correspondence should be addressed.

yeast cell as a model case. Furthermore, the effect of in- troduction of air bubbling on the recovery of the permeation flux was examined.

MATERIALS AND METHODS

Yeast Cell Suspension

The yeast cell suspension was prepared by suspending the commercially available baker's yeast cells (Oriental Yeast Co., Ltd., Tokyo, Japan) in 0.9% saline after three-time washing with 4 volumes of saline.

Crossflow Filtration

The module for crossflow filtration was of thin-channel type with a filtration area of 7.32 X lor3 m2 (24 mm wide and 305 mm long), which was the same as used previously." The channel depth was 2.5 mm. As a microfiltration mem- brane, synthetic membranes with a nominal pore size of 0.45 pm were used: SE45 membrane (polysulfone), do- nated by Fuji Photo Film Co., Tokyo, and HV membrane [poly(vinylidene fluoride)], purchased from Millipore Corp. (Bradford, MA). The microfiltration membrane was replaced before each experiment. The cell suspension was circulated at 20°C at a circulation flow rate of 10-60 cm3/s with a rotary pump (RM10, Nakamura Metal Co., Osaka, Japan) equipped with a variable-speed drive (Ringcorn RXM-400, Shimpo Industry Co., Kyoto, Japan). The trans- membrane pressure was 49 kPa. The permeate was returned to the reservoir tank when the cell concentration was kept constant. The concentration of cells was either 70 or 350 kg/m3 each in wet weight. The volume of cell suspension was 1 dm3 in the experiments at constant cell concentration. In the concentration experiments, the initial volume of cell suspension was 2 dm3.

In the preliminary experiments studying the effect of

Biotechnology and Bioengineering, Vol. 47, Pp. 401-404 (1995) 0 1995 John Wiley & Sons, Inc. CCC 0006-3592/95/030401-04

stopping of permeation flow, the filtration was first allowed to continue for 30 min, and then the permeation flux was stopped. The recovery of the flux was followed by measur- ing the amount of permeate at appropriate time intervals. For the measurement of the flux, the outlet for permeation was opened for 5 s. Periodical stopping of the permeation flow was done by closing the valve at the outlet of permeate periodically. The permeation interval and the closing times were 2.5-10 min and 1-4 min, respectively. In the bubbling experiment, air bubbles were introduced directly into the filtration module during the stopping of the permeation flow. We introduced 30 cm3 of air into the module five times immediately after the closure of the valve for perme- ate.

RESULTS AND DISCUSSION

Effect of the Stopping of Permeation Flow on the Recovery of Permeation Flux

Figure 1 shows the effect of the stopping of permeation flow on the recovery of permeation flux in the crossflow filtra- tion of the suspension of commercially available baker's yeast. An HV membrane was used. The cell concentration was 70 kg/m3 in wet weight. The permeation flux decreased from 7 X to 5 X lo-' m3/m2 s (from 2500 to 18 Wm2 h) for 30 min of the filtration. After 30 rnin the permeation flow was stopped. The flux rapidly increased to reach nearly the initial flux. The increase of the flux was depen- dent on the circulation flow rate. At the circulation flow rate of 60 cm3/s, the recovery of the flux showed nearly the same tendency as at 30 cm3/s; at 10 cm3/s the recovery of the flux was lower, particularly at the beginning. A similar result was obtained with an SE45 membrane.

8x1 0 - 4 I I I I 1 I I I

Stopping of permeation flow

!i t Filtration

- ln N E 2 6x10-4 E Y

n -

1

1 I p I I I

" 0 10 20 30 40 50 60 70

Time [min]

Figure 1. Effect of circulation flow on the recovery of the permeation flux during stopping of permeation in crossflow filtration of yeast cell suspension of commercially available baker's yeast. An HV membrane was used. The crossflow filtration was camed out for 30 min. Then, the permeation flow was stopped and the recovery of the flux was followed (see text for details). Transmembrane pressure, 49 kPa; cell concentration, 70 kglm'. Circulation flow rate: (0) 10 cm3/s, (A) 30 cm3/s, (0) 60 cm3/s.

Periodical Stopping of the Permeation Flow

Figure 2 shows the permeation flux in the crossflow filtra- tion of yeast cell suspension with and without periodical stopping of permeation flow. The cell concentration was 70 kg/m3 in wet weight and an HV membrane was used. With- out periodical stopping of permeation flow, the flux dropped sharply to the order of m3/m2 s (360 L/m2 h) and then gradually decreased to 2 X lo-' m3/m2 s (72 L/m2 h). The steady-state permeation flux was on the order of

m3/m2 s in the crossflow filtration of baker's yeast cell suspension cultivated in the yeast extract-polypeptone- dextrose (YPD) medium." Thus, the low permeation flux in the crossflow filtration of commercially available baker's yeast cell suspension is due to the fine particles contained in the commercially available baker's yeast. lo The permeation flux increased by several times by applying periodical stop- ping of permeation flow, although the flux still decreased gradually, where the permeation flow was periodically stopped for 4 min after 10 min. The permeation flux after 3 h filtration was 4 times as high as that without the periodical stopping. To increase the net permeation flux, the period for stopping was reduced from 4 to 1 min. The permeation flux with 1 rnin stopping decreased more rapidly than the case with 4 rnin stopping. After 3 h filtration the module was opened and the membrane surface was observed. An appre- ciable amount of microbial cake was found, indicating that 1 min was not enough to sweep the cake from the membrane away when the permeation period was 10 min.

We introduced air bubbles into the module to enhance the sweeping effect by the stopping of permeation. The trans- membrane pressure decreased at the moment of introduction of the bubbles. However, the pressure recovered in a suf- ficiently short time before the valve was reopened. Further- more, there was no difficulty to circulate the suspension

15x1 0-5 I I I I I - ln N E . n L 10x10 X 3

S 0

(D 0)

0)

- +

5x10

E n

0 0 60 120 180

Filtration time [min]

Figure 2. Effect of periodical stopping of permeation flow on the flux in crossflow filtration of suspension of commercially available baker's yeast. An SEX5 membrane was used. Circulation flow rate, 30 cm3/s; transrnem- brane pressure, 49 kPa; cell concentration, 70 kg/m3. (0) Filtration pe- riod, 10 min; stopping period, 4 min; without bubbling. (A) Filtration period, 10 min; stopping period, 1 min; without bubbling. (0) Filtration period, 10 rnin; stopping period, 1 min; with bubbling. The broken line shows the permeation flux without periodical stopping.

402 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 47, NO. 3, AUGUST 5, 1995

including bubbles with a rotary pump. By introducing the bubbles during 1 min periodical stopping, the permeation flux was higher in a way similar to the case with 4 rnin periodical stopping. Thus, the net permeation flux increased by 20% by decreasing the period of stopping from 4 to 1 min. The effect of the bubbling was also found with an SE45 membrane.

Some researchers reported the effect of bubbling during crossflow filtration. Nakano et a1.8 showed the effect of bubbles in filtration of yeast cell suspension using a rotating ceramic module. The permeation flux increased by 50% by aeration. Imasaka et a1.2 also showed a turbulence- promoting effect of gas-liquid mixtures to suppress cake formation on the membrane module in a gas-liquid two- phase crossflow filtration system. However, in both the studies, bubbles were introduced during permeation. Dur- ing the crossflow filtration, the cell layer is pressed onto the membrane due to the permeation flow. The pressing force might be released when the permeation flow is stopped. Thus, it might be more effective to introduce bubbles during the stopping of permeation.

Figure 3 shows the effect of the periodical stopping of the permeation flow in the crossflow filtration of yeast cell suspension at the concentration of 350 kg/m3. The perme- ation flux decreased rapidly to 2 X 10K5 m3/m2 s (72 L/m2 h) in 15 rnin when the crossflow filtration was carried out without periodical stopping. By applying the periodical stopping of permeation flow with 4 min stopping and 10 rnin filtration, the flux became nearly constant at 4 X lop5 m3/m2 s (140 L/m2 h). The flux was also kept at 4 X m3/m2 s even by 1 min stopping when the air bubbles were introduced as described above. The net permeation flux with air bubbling was two times that obtained in the cross- flow filtration without periodical stopping. When the stop- ping time was 1 min without introduction of bubbles, the

8x1 0 5

E 2 6 ~ 1 0 - ~ E

- u) N

Y

E 0 m Q 2x

ta

.- CI

fi I I I

0 60 120 180

Filtration time [min]

Figure 3. Effect of periodical stopping of permeation flow on the flux in crossflow filtration of suspension of commercially available baker's yeast. An SFA5 membrane was used. Circulation flow rate, 30 cm3/s; transmem- brane pressure, 49 kPa; cell concentration, 350 kg/m3. (0) Filtration pe- riod, 10 min; stopping period, 4 min; without bubbling. (A) Filtration period, 10 min; stopping period, 1 min; without bubbling. (0) Filtration period, 10 min; stopping period, 1 min; with bubbling. The broken line shows the permeation flux without periodical stopping.

flux was in nearly the same level as that without periodical stopping. Thus, the periodical stopping of the permeation flow with air bubbling was found to be effective even at the high cell concentrations.

Based on these findings, we applied the periodical stop- ping method to the concentration of the yeast cell suspen- sion, with and without bubbling. In Table I the time needed for concentration of yeast cell suspension from 70 to 350 kg/m3 is summarized. The time was reduced to half com- pared to the continuous filtration by applying the periodical stopping (10 min permeation and 1 rnin stopping) of the permeation flow with air bubbling. By reducing the perme- ation period from 10 min to 5 and 2.5 min with 1 min stopping, the final flux as well as the initial flux during each permeation period increased. Accordingly, the time needed for concentration reduced. Membrane materials had little effect on the permeation flux (Table I). There was no ap- preciable difference in the time needed for concentration with or without the bubbling when the permeation period was 2.5 min in contrast to the result with the 10-min period of permeation (Table I). These findings imply that the de- gree of compaction of the cake of commercially available baker's yeast cell containing a trace amount of fine particles depends on the permeation period. The periodical stopping of permeation with such a short period as 2.5 rnin made the cake removal easy even without air bubbling. With a longer permeation period such as 10 min (Fig. 2), an introduction air bubbling was quite effective in the removal of the mi- crobial cake as described previously. The change of the properties of the cake such as the compactness and sticki- ness with the progress of the permeation time should be further investigated.

A backwashing method is usually used in crossflow fil- tration to increase the permeation However, spe-

Table I. Concentration of suspension of commercially available baker's yeast cells during crossflow filtration with periodical stopping of perme- ation flow with air bubbling.

Permeation period Initial permeation Final permeation Time need for (min) flux (m3/m2 s) flux (m3/m2 s) concentration (min)

10 1.1 x 1 0 - ~ 5 . 1 x 10-5 56 10" 1.0 x 1 0 - ~ 4.0 x 10-5 59 5 1.4 x 1 0 - ~ 9.5 x 1 0 - ~ 41 2.5 1.8 x 1 0 - ~ 1 . 1 x 10-4 39 2 S b 1 .7 x 1 0 - ~ 1.1 x 1 0 - ~ 41

Continuous filtrationb 1.4 x 1.9 x 1 0 - 5 d 1 1 1

Note: The stopping time was 1 min. A 2-dm3 suspension with the concentration of 70 kg/m3 was concentrated to 0 .4 dm3 of 350 kg/m3. Circulation flow rate, 30 cm31s; transmembrane pressure, 49 Wa. An SEM membrane was used unless otherwise noted. The mean permeation fluxes during the initial and final permeation period were measured in the cross- flow filtration with periodical stopping.

"An HV membrane was used. bAir bubbles were not introduced. 'Mean permeation flux for 3 rnin from the start of crossflow filtration. dMean permeation flux for 3 min at the end of crossflow filtration.

COMMUNICATION TO THE EDITOR 403

cia1 equipment is necessary for conducting backwashing and sometimes organic membranes made of cellulose or synthetic polymers is not usable.

The method developed in this study is one used to in- crease the flux without any additional equipment.

We thank Fuji Photo Film Co. for their kind gifts of membrane.

References

1 . Hanish, W. 1986. Cell harvesting. pp. 61-88 In. W. C. McGregor (ed.), Membrane separations on biotechnology. Marcel Dekker, New York.

2. lmasaka, T . , Kanekuni, E., So, H., Yoshino, S. 1989. Gas-liquid two-phase cross-flow filtration by ceramic membrane modules. Kagaku Kogaku Ronbunshu 15: 638-644.

3. Kroner, K . H., Nissinen, V . , Ziegler, H. 1987. Improved dynamic filtration of microbial suspensions. Bio/technology 5: 921-926.

4. Kroner, K . H., Schiitte, H., Hustedt, H., Kula, M.-R. 1984. Cross-

filtration in the downstream processing of enzyme. Proc. Biochem., April, 67-74.

5. Matsumoto, K . , Katsuyama, S. , Ohya, H. 1987. Separation of yeast by cross-flow filtration with backwashing. J. Ferment. Technol. 65: 77-83.

6. Matsumoto, K., Kawahara, M., Ohya, H. 1988. Cross-flow filtration of yeast by microporous ceramic membrane with backwashing. J. Ferment. Technol. 66: 199-205.

7. Matsumoto, Y., Totsuka, Y., Miwa, T., Nakao, S.-I., Kimura, S. 1992. Kagaku Kogaku Ronbunshu 18: 677-683.

8. Nakano, K., Matsummura, M., Kataoka, H. 1993. Application of a rotating ceramic membrane to dense cell culture. J. Ferment. Bioeng. 73: 49-54.

9. Riesmeier, B., Kroner, K. H., Kula, M.-R. 1989. Tangential filtra- tion of microbial suspensions: filtration resistances and model devel- opment. I. Biotechnol. 12: 153-171.

10. Tanaka, T . , Kamimura, R. , Itoh, K., Nakanishi, K . , Matsuno, R. 1993. Factors affecting the performance of crossflow filtration of yeast cell suspension. Biotechnol. Bioeng. 41: 617-624.

11. Yamasaki, H., Lee, M.-S., Tanaka, T., Nakanishi, K. 1993. Im- provement of performance for cross-flow membrane filtration of pul- Man broth. Appl. Microbiol. Biotechnol. 39: 21-25.

404 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 47, NO. 3, AUGUST 5, 1995