Journal of Biotechnology 92 (2001) 89 94 www.elsevier.

com/locate/jbiotec

A novel airlift photobioreactor with bafes for improved light utilization through the ashing light effect
Jorg Degen a,*, Andrea Uebele a, Axel Retze b, Ulrike Schmid-Staiger a, Walter Trosch a
a

Fraunhofer Institut fur Grenzachen- und Bio6erfahrenstechnik, Nobelstr. 12, D-70569 Stuttgart, Germany b Fraunhofer Institut fur Solare Energiesyteme, Oltmannstrasse 5, D-79100 Freiburg, Germany Received 13 June 2000; received in revised form 17 November 2000; accepted 27 November 2000

Abstract A newly developed at panel airlift photobioreactor with a dened circulation path was tested for microalgal culture. The bioreactor exposed the cells to intermittent light to improve the efciency of light utilization through the ashing-light effect. During batch cultures in the new photobioreactor, the biomass productivity of Chlorella 6ulgaris was 1.7 times greater than in a randomly mixed bubble column of identical dimension. A reduction in light path from 30 to 15 mm increased the biomass productivity by 2.5-fold. A maximum dry biomass productivity of 0.11 g l 1 h 1 was obtained at an articial illumination of 980 mE m 2 s 1. 2001 Elsevier Science B.V. All rights reserved.
Keywords: Microalgae; Flashing light effect; Airlift photobioreactor; Bafes; Productivity; Light path

1. Introduction Algae produce a variety of useful compounds. Algal biomass is used as human food and aquaculture feed. Potentially, algal culture provides a means of reducing the amount of carbon dioxide accumulation in the atmosphere. Despite many possible applications, only a few species of algae are cultured commercially because of poorly developed photobioreactor technology (Pulz and Scheibenbogen, 1998). Heterotrophic culture in which the cells use an organic carbon source for
* Corresponding author. Tel.: + 49-711-970-4037; fax: + 49-711-970-4200. E-mail address: deg@igb.fhg.de (J. Degen).

energy is signicantly less expensive than photosynthetic growth (Radmer and Parker, 1994; Running et al., 1994) but is not a sustainable production technique, and not all chemicals can be produced by heterotrophic growth. Biomass concentration and productivity attained in photobioreactors are comparatively low. One reason for low productivity is inefcient conversion of the available light into biomass (Tredici and Zitelli, 1998). Other factors contributing to low productivity are, an accumulation of the inhibitory oxygen in conventional photobioreactors; consumption of biomass by respiration in the dark zones of the reactor; insufcient mixing of CO2 and nutrients; and photoinhibition in the intensely illuminated outer zones.

0168-1656/01/$ - see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 1 6 5 6 ( 0 1 ) 0 0 3 5 0 - 9

90

J. Degen et al. / Journal of Biotechnology 92 (2001) 8994

When the illumination intensity is greater than the saturation value, the conversion of light energy to biomass can be enhanced if algal cells are made to repeatedly cycle between the well lit exterior and the dimly lit interior of the photobioreactor. This conversion enhancement by light dark cycling is known as the ashing-light effect (Kok, 1956; Terry, 1986). Random mixing does not enhance productivity as much as a regular light dark cycle of about 1 Hz frequency with a light versus dark residence time of 1:10. Although several designs of photobioreactors have attempted to use the ashing-light effect for productivity enhancement (Laws et al., 1983; Hu et al., 1996), a regular mixing pattern has not generally been implemented. This work reports on a new photobioreactor that enhances productivity by effectively using the ashing-light effect. The reactor is based on the airlift principle. It uses a at panel design to reduce the light path (Hu et al., 1996; Tredici et al., 1991) and bafes to induce a regular light cycling of microalgae.

The light compensation and saturation points for the alga were determined as follows; 40 ml of diluted cell suspension (optical density= 0.1) was placed in 250 ml tissue culture asks with darkended sides. The culture depth was about 5 mm. No self-shading occurred because the cell suspension was quite dilute. The asks were arranged on a rotary shaker (50 rpm, 22 C) and were illuminated from above with different light intensities from 0 to 400 mE m 2 s 1 (1400 W HQI-vapor lamp, Osram). The lamps used provided a light spectrum that was similar to that of natural sunlight. The inoculum for the photobioreactor was grown at a higher cell density and an illumination intensity of 80 mE m 2 s 1, all other conditions were as noted earlier.

2.2. Flat panel airlift (FPA) photobioreactor

The rectangular channel airlift photobioreactor (Trosch et al., 2000, German patent 199 16 597.1 41) was made of Plexiglas, and had an illuminated front area of 0.084 m2. A small downcomer zone was located on one side of the reactor (Fig. 1). The riser was subdivided into interconnected chambers by means of horizontal bafes. The bafes were attached alternately to the front and the back of the larger at faces of the reactor. The riser zone was injected with compressed air. The uid in the chambers circulated so that the cells regularly moved between the narrow illuminated zone and the deeper dark zone (Fig. 1). Two different culture depths (15 and 30 mm) were tested to establish the effects of changes in the ratio of the light and dark zones. The reactor volumes were 1.5 and 3.0 l, respectively, of which the bafes supplant about 7%. The temperature was controlled by circulating cooling water through a transparent jacket located at the front of the reactor. The temperature and pH sensors were placed in the upper region of the reactor next to the gas outlet.

2. Material and methods

2.1. Microorganism and growth conditions

Chlorella 6ulgaris SAG 211 12 was obtained from the Gottingen Algae Culture Collection (Schlosser, 1994). The DS-medium of Pohl et al. (1987) was used (after supplementation with phosphate and nitrate) to grow the alga. The concentration of the additional salts were, 5 g l 1 KNO3, and 0.75 g l 1 K2HPO4. The initial pH was 6.8. The optical density was determined with triple samples at 625 nm in a spectrophotometer (Hitachi). Cell dry weight was determined previously with 10-ml aliquots. Each sample was ltered through an acetate lter (Millipore), washed twice with distilled water, dried overnight at 100 C until constant weight. The correlation between optical density and cell dry weight is cell dry weight (g l 1) =0.15 OD625 (r 2 = 0.98)
nm

2.3. Growth in the FPA photobioreactor

For obtaining a dense culture for inoculation of the reactor, Chlorella was grown in a 800-ml tubular loop reactor circulated with a peristaltic

J. Degen et al. / Journal of Biotechnology 92 (2001) 8994

91

pump (Watson Marlow, 100 rpm). The culture was illuminated from both sides of the tubes with four Osram Fluora lamps (120 mE m 2 s 1). The pH was controlled at 6.89 0.1 by injecting carbon dioxide by demand. The cells were grown to an optical density of 15 20 before being withdrawn for inoculation of the FPA. Growth experiments were done in the FPA (Fig. 1) and in a 3.0-l bubble column with identical volume and shape (except the bafes) for comparison. The reactors were sterilized overnight with hydrogen peroxide (3% vol.) and washed (3) with sterile water prior to use. The alga was grown in the reactor from an initial optical density of 3.5 to a nal optical density of 11 9 1 and at an average illumination of 300 mE m 2 s 1 (2 400W HQI-vapor lamp, Osram) to preadapt to the reactor conditions. The algal suspension was then diluted to an optical density of

3.5 before each batch experiment. Photoinhibition of the diluted culture was avoided by gradual adjustment of the light intensity to match the culture biomass concentration. Full average light intensity of 9809 80 mE m 2 s 1 was applied when the optical density was 11. Illumination at the surface was relatively uniform, except at the periphery of the reactor where the illumination intensity was 80 mE m 2 s 1 less than elsewhere on the reactor. Luminous intensities where measured with a lux meter (TES 1332; A.T.P. Messtechnik GmbH, Kappel) at several points on the reactors surface and converted to PAR light intensity by correlation previously done with a light meter (Zeiss MMS1/VIS-enh. 300 1100 nm, Jena). Average light intensity on the surface was calculated by integration. The temperature (29 C) and pH (6.890.1) were controlled as previously noted. The cultures were bubbled with

Fig. 1. Flat panel airlift (FPA) photobioreactor.

92

J. Degen et al. / Journal of Biotechnology 92 (2001) 8994

reactor, and an energy content (enthalpy) of Chlorella biomass of 25 kJ g 1 dry weight (Javanmardian and Palsson, 1991), the data suggest a photosynthetic efciency of 4.7% in the FPA and only 2.9% in the bubble column. Because the reactors and the operating conditions were identical except for the bafes, the results demonstrate that productivity is improved by installing the bafes. The use of bafes can contribute to productivity enhancement in several possible ways. (1) The

Fig. 2. Light intensity growth rate relationship of C. 6ulgaris SAG 211 12 at 22 C.

0.45 vvm compressed air. The optical density was determined several times per day.

3. Results and discussion

3.1. Light response cur6e of Chlorella 6ulgaris

The specic growth rate versus irradiance relationship is shown in Fig. 2. The light compensation point (i.e. the irradiance below which there is no net photosynthesis) was between 5 and 10 mE m 2 s 1 and the light saturation intensity was about 250 mE m 2 s 1. The maximum specic growth rate was 0.08 h 1. These results suggest that photoinhibition can be avoided in dilute Chlorella cultures (e.g. at the beginning of the batch) by keeping the irradiance at about 250 mE m 2 s 1.

3.2. Culti6ation and producti6ity in the FPA photobioreactor in comparison to a bubble column
The growth and productivity data for the two reactors with 3-cm light path are shown in Fig. 3a and b. Volumetric biomass productivity was about 1.7-fold higher in the FPA than in the comparable bubble column. The maximum productivity attained was 0.045 and 0.027 g l 1 h 1, respectively. Considering an average light intensity of 980 mE m 2 s 1 at the surface of the

Fig. 3. (a) Comparison of batch growth of C. 6ulgaris in a at panel airlift photobioreactor (30-mm path length) and a similar bubble column. The illumination intensity was 980 mE m 2 s 1 in both cases. (b) Comparison of productivity of C. 6ulgaris in a at panel airlift photobioreactor (30-mm path length) and a similar bubble column. The illumination intensity was 980 mE m 2 s 1 in both cases.

J. Degen et al. / Journal of Biotechnology 92 (2001) 8994

93

bafes increase the residence time of gas bubbles in the reactor and this can affect the mass transfer rates of carbon dioxide and oxygen. However, because oxygen did not accumulate and CO2 was supplied on demand so that it was never a limiting factor, the mass transfer effects do not explain the better performance of the bafed reactor. (2) The bafes affect mixing; however, both reactors were always mixed sufciently to supply the cells with dissolved nutrients. Indeed, when biomass concentration was low (i.e. homogenous illumination), the two reactors attained similar productivities (Fig. 3b). Therefore, any effects of bafes on the supply of dissolved nutrients does not explain the difference in performance of the reactors. Nutrient limitations could be further excluded because nitrate and phosphate were added in excess, and in earlier experiments Chlorella could attain a concentration of 7.5 g DW per l in the same medium. The FPA bioreactor performed dramatically better than the bubble column only when biomass concentration had become substantial, i.e. when the volume of the reactors had become demarcated into light and dark zones. The organized mixing in the FPA now greatly enhances performance. An optical density of 12 in Chlorella cultures corresponds to 1.8 g DW per l and with 980 mE m 2 s 1 illumination, less than 1% of the light remained at a culture depth of 3 mm. At this depth the light intensity fell below the light compensation point of about 7 mE m 2 s 1; thus, the deeper region of the reactor was effectively a dark zone. Repeatedly moving the dark culture volume to the light zone improved productivity. In the bubble column, the light dark movement was chaotic and did not have a well-dened frequency. The dened mixing pattern in the FPA effectively simulated a regular ashing-light.

Fig. 4. (a) Comparison of batch growth of C. 6ulgaris in a at panel airlift photobioreactor with light path lengths of 15 and 30 mm. The illumination intensity was 980 mE m 2 s 1 in both cases. (b) Comparison of productivity of C. 6ulgaris in a at panel airlift photobioreactor with light path lengths of 15 and 30 mm. The illumination intensity was 980 mE m 2 s 1 in both cases.

3.3. Effect of the light path length on producti6ity in the FPA photobioreactor
Maximum utilization of available light energy for maximum production in a given reactor depends on optimal selection of the population densities in relation to the light path length. In a light limited system, the shorter the light path, the

higher the optimal cell density and volumetric productivity (Hu et al., 1996). Theoretically the best light conditions in the FPA with 30 mm were at an optical density of 12, where the proportion of light/dark volumes was 1:10. If the reactors depth is only 15 mm, the same light/dark proportion was attained at a higher optical density and, therefore, a greater productivity. This was experimentally veried by using a modied FPA with 15-mm culture depth. Fig. 4a and b

94

J. Degen et al. / Journal of Biotechnology 92 (2001) 8994

compares the results of batch experiments in 15mm FPA with data obtained in the 30-mm FPA. Peak productivity in the thinner FPA was 0.11 g l 1 h 1 dry weight at an optical density of about 13, corresponding to 1.95 g l 1 biomass dry weight. This was about 2.5-times greater than the values attained in the 30-mm FPA. These results concur with those of Hu et al. (1996); an increase in volumetric productivity with decreasing light path. Because a reduction in light path length from 30 to 15 mm reduces the culture volume by half, a greater than twofold increase in volumetric productivity implied an even bigger increase in areal productivity. No serious photoinhibition related damage was detected in Chlorella cells. Microscopic examinations showed no damage to internal cell organelles, nor the built up of clearly visible carotenoids. Carotenoid content was not measured. A change in color towards a yellowish were detected in both FPA-bioreactors near the end of the batch growth, when the conditions necessary for attaining the ashing-light effect were no longer satised. A visible build up of carotenoids was detected much earlier in the bubble column.

References
Hu, Q., Gutermann, H., Richmond, A., 1996. A at inclined modular photobioreactor for the outdoor mass cultivation of photoautotrophs. Biotechnol. Bioeng. 51, 51 60. Javanmardian, M., Palsson, O.P., 1991. High density photoautotrophic algal cultures: design, constructions, and operation of a novel photobioreactor system. Biotechnol. Bioeng. 38 (10), 1182 1189. Kok, B., 1956. Photosynthesis in ashing light. Biochim. Biophys. Acta 21, 245 258. Laws, E.A., Terry, K.L., Wickman, J., Chalup, M.S., 1983. A simple algal production system designed to utilize the ashing light effect. Biotechnol. Bioeng. 25, 2319 2335. Pohl, P., Kohlhase, M., Krautwurst, S., Baasch, K.-H., 1987. An inexpensive inorganic medium for the mass cultivation of freshwater microalgae. Phytochemistry 26 (6), 1657 1659. Pulz, O., Scheibenbogen, K., 1998. Photobioreactors: design and performance with respect to light energy input. Adv. Biochem. Eng. Biotechnol. 59, 123 152. Radmer, J.R., Parker, B.C., 1994. Commercial applications of algae: opportunities and constraints. J. Appl. Phycol. 6, 93 98. Running, J.A., Huss, R.J., Olson, P.T., 1994. Heterotrophic production of ascorbic acid by microalgae. J. Appl. Phycol. 6, 99 104. Schlosser, U.G., 1994. SAG-Sammlung von Algenkulturen at the University of Gottingen, Catalogue of Strains 1994. Bot. Acta 107, 113 186. Terry, K.L., 1986. Photosynthesis in modulated light: quantitative dependence of photosynthetic enhancement on ashing rate. Biotechnol. Bioeng. 28, 988 995. Tredici, M.R., Zitelli, G.C., 1998. Efciency of sunlight utilization: tubular versus at photobioreactors. Biotechnol. Bioeng. 587 (2), 187 197. Tredici, M.R., Carlozzi, P., Zitelli, G.C., Materassi, R., 1991. A vertical alveolar panel (VAP) for outdoor mass cultivation of microalgae and cyanobacteria. Bioresour. Technol. 39, 153 159. Trosch, W., Schmid-Staiger, U., Zastrow, A., Retze, A., Brucker, F., 2000. Photobioreactor with improved supply of light by surface enlargement, wavelength shifter bars or light transport. Patent announced (199 16 597.1-41, PCT/EP00/03089).

4. Concluding remarks In view of the observations discussed a highly dened mixing pattern that produces light dark cycling at a given frequency is required for enhancing productivity through the ashing-light effect. By contrast, chaotic mixing is not as effective in enhancing productivity as is organized mixing.