photobioreactors using led

8/12/2019 Photobioreactors Using LED

1/7

High-Density Algal PhotobioreactorsUsing Light-Emitting DiodesChoul-Gyun Lee and Bernhard0 PalssonDepartment o Chemical Engineering, University o Michigan,Ann Arbor, M ichigan 48109Received December 23 1993lAccepted June 17, 1994

Lack of high-density algal photobioreactors (PBR) hasbeen a limitation in exploiting th e biotechnological po-tential of alga e. Recent dev elo pm en ts of highly efficientlight-emitting diodes (LED using gallium aluminum arse -nide chips) have ma de the developm ent of a small LED-based PBR possible. We have calculated theoretical val-ues of g as m as s transfer requirements and light-intensityrequirements to support high-density algal cultures forthe 680 nm m onochroma tic red light from L E D a s a lightsource. A prototype PBR has been designed based onthe se calculations. A cell concentration of more than 2 xlo9cells/mL (m ore than 6.6% v/v), cell doubling times a slow as 1 2 h, and an oxygen production rate a s high a s 10mmol oxygen/L culture/h were achieved using on-lineultrafiltration to periodically provide fresh medium. 01994 John Wiley & Sons, Inc.Key words: Chlorella vulgaris light-emitting diodes( L E D ) oxygen production photobioreactor (PBR)

INTRODUCTIONAlgal biotechnology is drawing increasing interest due to itspotential as a source of valuable pharmaceutic als, pigments,carbohydrates, and other fine chemical^.^^^ * ^^^ Cur-rently, its applica tion is being ex tended to the areas of wastewater treatment6 0 4 and agricu1tu1-e.~~urther, the abilityto regenerate oxygen from carbon dioxide is key to thedesign of an efficiently controlled ecological life-supportsystem.21*3'However, lack of suitable photobioreactors(PBR) makes the co st of algally-derived compounds higherthan those derived by chem ical and thus hasprevented widespread use of algal cultures.Important factors in achieving high-density productivealgal culture are light, gas exchange, medium supply, andbiological limitations. 17734,35To effectively exploit thecommercial potential of algae, a cheap, durable, reliable,and highly efficient light source is needed. If the lightsource has a narrow spectral output that overlaps the pho-tosynthetic absorption spectrum, the emission of light atunusable frequencies would be eliminated and therefore theoverall energy conversion improved. One light source thatmeets these c riteria is a light-emitting diode (LED ).38 LEDare light and small enough to fit into virtually any PBR,they have a very long life-expectancy, and their electricalefficiency is so high that heat generation is minimized. LED

* To whom all correspondence should e addressed.

Biotechnology and Bioengineering, Vol. 44, p. 1161-1167 (1994)0 1994 John Wiley & Sons, Inc.

have a half-power band-width of 20-30 nm which can bematched with photosynthetic needs. H owever, the light in-tensity of conventional LED is too low to be used in PBR.Recently, double-power double-heterostructure (DDH) gal-lium aluminum arsenide (GaAlAs) chips which can emitlight at much higher intensity and efficiency have been de-~ e l o p e d . ~ 'hese DDH G aAlAs LED arrays can emit pho-ton fluxes as high as 900 pmol/m 2/sec,' which is sufficientto support photosyn thesis in many C, plants.46For the past decade, it has been shown that high-densityphotoautotrophic algal cultures can be established. A num-ber of PBR have been introduced to obtain productive high-density photosynthetic cultures. Most of them were thinpanel or tubular PBR,22*263333363423aut modified stirred-tank43,48or hollow-fiber PBR47 were also tested. New-concept PBR with internal illuminations through optical fi-bers5. 15,24,27,4 1 have also been employed. The key designchallenges were to maximize surface-to-volume ratio, toprovide a more efficient light source, to deliver the lightmore efficiently to the culture, and to provide a more ef-fective mass transfer of gases. Direct comparison of theperformance of these PBR is difficult, but the general per-formance characteristics can be enumerated. The maximumbiomass concentrations achieved were normally in the rangeof 5-6 g/L.22327,42,4374348ome PBR were able to obtainhigher cell concentrations than 10g/L,'5,41with the highestreported cell concentration being 20 g/L.,, Few studieshave reported the oxygen production rate. All the PBR re-ported used full-spectrum light source, such as sunlight ormetal halide lamps. Here we report that red LED can beused in PBR to obtain high cell densities and high oxygenproduction rates.PHOTOBIOREACTOR DESIGN CO NSIDERA TIONThe major factors that must be considered in the design ofhigh-density PBR are light delivery and distribution, gastransfer into and out of the reactor, mainte nance of mediumcomponents in balance, and prevention of accumulation ofpotentially detrimm tal secondary metabolites.GoalsThe goal is to achieve maximal biomass production ratesusing LED. The highest cell density previously obtained in

CCC 0006-3592/94/01001161-07


2/7

our laboratory was 1 X lo9cells/mL ( ~ 3 %/v)15 by usinga 1,OOO-W xenon lamp as the light source. The oxygenproduction rate of this PBR system was between 4-6mmoYL culture/h. The goal is to achieve equal or higherculture density and oxygen production rate using red LEDas compared to a full-spectrum light source.Estimationof the Specific Oxygen ProductionRate SOPR)Each m ol of chlorop hyll is reported to produce 50 400 molof oxygen per h,29 and the per-cell average chlorophyllcontent is in the range of 0.5-1.0 f m 0 1 . ~ ~he product ofthese numbers will yield a range of expected per-cell oxy-gen productivity; 25-400 fmol O,/cell/h.A simple experiment with a small, illuminated couvettewith a dissolved oxygen (DO) electrode was employed toverify this range. The measured SOPR fell within the esti-mated range. The SOPR was between 5-200 fmol/cell/h,decreasing with increasing cell concentration as expecteddue to mutual shading (Table I). S ince the measured SO PRwas inversely proportional to the cell concentration, thedesign challenge is to deliver and distribute the light effi-ciently to minim ize the effect of m utual shading. The SOPRwill have to be high at high-cell densities if the design of aPBR is to be optimized for oxygen produc tion. A target insitu SOPR is 100 fmol of oxygen per cell per h at a cellconcentration of 1 lo9 cells/mL.The Light RequirementEnough light to produce 100 fmol of ox yged celllh has to beprovided. According to the Z-scheme photosynthesis, 2 8mol of photons are required to generate 1 mol of oxygen.One m ol of 680 nm photons per h is equivalent to about 50W. As a result, each cell needs to experience 40 pW equiv-alent of photons to achieve the targeted SO PR of 100 fmollcelllh.The d esired cell density and the spec ific illuminated sur-face area are then used to compute the required light flux.For the lo9 cells/mL culture and 1.0 cm 2 illuminated area

per cm3 culture, one need s to provide 40 mW /cm2 or morebecause the maximum photosynthetic efficiency of Chlo-rella has been reported to be in the range of 32% to54y0.20,28.32 A closely packed LED array 90 ED on a 2X 4 printed circuit board, spaced about 0.76 cm apart)could provide a light intensity as high as 25 mW /cm2 oneach illuminating surface of the PBR.Light Delivery and DistributionIn addition to the light source, the light delivery and thelight distribution inside a PBR are key factors in PBR de-sign. Light should be delivered so that photon loss is min-imized, heat generation by the light source eliminated, andpotentially harmful wavelengths filtered. By using LED, allof these problems are circumvented. The best way to min-imize photon loss is to have internal illumination. The sm allvolume characteristics of LED makes this possible. Fu rther-more , their high efficiency and sharp spectru m eliminate theneed for a cooling system and filter. Therefore, LED rep-resent perhaps the optimal light source for PBR.Good light distribution will maximize the overall lightutilization by minimizing mutual sha ding as well as photo-inhibition. It should be noted that any light not abso rbed ornot used in photosynthesis will be converted into thermalenergy.Gas TransferGas transfer is an important issue in PBR design; carbondioxide needs to be supplied and oxygen generated re-moved. Carbon dioxide is the carbon source utilized forp h o t o s y n t h e ~ i s . ~ ~epending on the type of species, C O ,can be absorbed in dissociated (H C0 3- or C03-,) or un-dissociated form (CO, or H 2C0 3).2 In either case, carbondioxide needs to be dissolved into the culture broth. Oxygenis a product of photosynthesis as well as an inhibitor of algalgrowth. Thus, the produced oxygen needs to be removed toprevent oxygen from reaching inhibitory level.The target of SOPR is 100 fmol O,/celllh. For lo9 cells/mL, 100 mmol oxygen needs to be removed per L of culture

Table I.cell concentrationsSpecific oxygen production rate in units of fmol/celVh at various light intensities and

Cell concentration (cells/mL)1.3 X lo6 1.5 X lo7 2 .9 X lo7 8 .5 X lo7 2 .5 X 10 4.5 X 10'

0.17 0.7Approx. 0 .25 9 .2relative 0.4 2 209 15.7light 0.50 63 7.7 18.1 8. 0intensity 0.67 1310.83 188 127 8.3 29. 0 10.21.00 148 9.1 11.2

81Light intensities are expressed as relative intensities to the maximum (the intensity of 2.1 Vldevice1). A 22 W cool-white circline fluorescent light (General Electric Co.), or FL,was used as a

full-spectrum co ntrol.

1162 BIOTECHNOLOGY AND BIOENGINEERING VOL. 44, NO. 10, NOVEMBER 20, 1994


3/7

per h. T o achieve this rate of oxygen transfer, the volumet-ric mass transfer coefficient k,a) needs to be larger than 80h- '. However, SOPR may be reduced as the cell concen-tration incre ases due to the light limitation in a high-densityculture.One way to remove 100 mmol of oxygen/L culture/h isby internal sparging. The overall mass transfer rates bysparging were calculated for two different bubble sizes (1mm and 3 mm in diameter).'' Both could give the requiredoxygen transfer k,a of about 200 h- ' and 100 h- ', re-spectively, at the aeration rate of 20 d h ) . The results fromthe above calculations, as well as those for typical bioreac-tors,' are summar ized in Table 11.MixingGood fluid mixing is im portant for achieving high ce ll con-centration in a PBR. Good mixing can keep the cells insuspension, eliminate thermal stratification, help nutrientdistribution, and improve gas exchange. Furthermore, mix-ing can reduce the degree of m utual shading and lower theprobability of photoinhibition. The last reason is perhapsthe most important one for requiring the right extent ofmixing to achieve a successful PBR. Proper mixing in thedirection perpendicular to flow (p arallel to light path) willsignificantly reduce the mutual shading and thus increasethe portion of the cells exposed to the light. This mixingwill also move the cells close to the illuminated surface intothe center of the PBR, which will give these photon-saturated cells the opportunity to use up all the absorbedlight energy for photosynthesis before they are exposed tothe light again. This mixing, in turn, will also provide thesame opportunity to the cells in the center (tha t already usedup the absorbed photons) to see the light. As a result, aproper mixing may significantly improve overall light-utilization efficiency. S parging will improve overall light-utilization efficiency in tw o ways: (1) existence of air bub-bles will increase light penetration depth, and (2) risingmotion of bubbles will induce some degree of mixing tan-gential to the flow d irection.On-Line Supply of Medium ComponentsYet another important parameter in PBR design is to keepall medium com ponents in balance with growth and photo-synthetic requirements. Discussion of this topic can beTable 11. Overall mass transfer coefficient under various conditions

System configurationStagnant water (specif ic surface area of 1 cm2/cm3)Prototype PBR ( 2 0 d h aeration, 3mm diam. bubble)Prototype PBR ( 2 0 d h aeration, 1 mm d im . bubble)Bubble fermentor (no agitation, 20 d h eration)Conical flasks (shaking at 200 rpm, o aerationyCSTR (500 rpm, 0 d h erationyCSTR (750 rpm, 0 d h erationyCSTR (1,680 rpm, 2 0 d h aerationy

~~ 7.561102001@3025-100325

1OOo2,650Data from ref. 8.

found elsewhere. 6 The on-line replenishment of medium(supply of m edium components as well as removal of anypotentially harmful secondary metabolites) was achievedeither by dialysis or ultrafiltration using C ell-Pharm modelBR170 (mo lecular weight cut-off of 70 kDa , U niSyn Tech-nologies, San Diego, CA ). The difference between the twomethods is the application of external force for ultrafiltra-tion; the dialysis mode supplies the medium componentspurely by diffusion.Ultrafiltration was preferred because it was faster andthus disturbed gas measurement less. By applying vacuum(about 400 mmHg) to the shell side of the ultrafiltrationcartridge, replenishment of 2 00 mL of medium (abo ut threetimes the volume of medium in the PBR system used) couldbe achieved in 10 min.

PHOTOBIOREACTOR DESIG NAND CONSTRUCTIONLED were chosen as a light source for all the reasons statedabove, such as high electrical energy conv ersion efficiency,sufficient light intensity, and sm all volume and w eight char-acteristics (which will make the most efficient internal il-lumination possible). For the required mass transfer rate,internal sparging was employed.Black polycarbonate plate (Ain Plastics, Inc., MountVerno n, NY) and black norprene tubing (Cole-Parmer In-strument Co., Niles, IL) were chosen as materials used forPBR construction and the tubing m aterial because of theirbiocompatibility , chemical and mechanical properties,opaqueness, oxygen impermeability, and autoclavability.From a series of simple preliminary experiments withvarious geome tries, a PBR w ith a vertical rectangular slab-shaped illumination chamber was found to be suitable. Thisshape meets all the stated criteria, such as good light dis-tribution with the smallest possible dark volume, propermixing without damaging cells by shear stress, and m ini-mization of wall growth.The volum es of the illumination chamber were either 80cm3 or 52 cm3, depending on the dep th of culture (1.55 cmand 1 OO cm , respectively). Both illumination chambe rs hadthe same illumination area of about 100 cm2 (8 in2, sum ofthe two sides). Each illumination chamber had two LEDunits on each side which consisted of either 24 LED forlow-intensity light delivery or 90 LED for high-intensitylight delivery. The two LED units we re mounted in such away that the LED on the differen t sides of the illumina tionchamber were alternated to have better light distributioninside the illumination chamber. The power to the LEDunits was increased as the cell concentration increased withtime.Mass transfer was achieved via internal sparging throughfour 3-mm diameter nozzles (a half-inch apart from eachother) located at the bottom of the illumination chamber. Ata gas flow rate of 100 mL/min with a liquid circulation rateof 100 mL/min, calculated superficial velocities were 0.42c d s e c and 0.67 c d s e c , depending on the thickness of theillumination cham ber.

LEE AN D PAL SSON: LED-BASED ALGAL PHOTOBIOREACTORS 1163


4/7

The resulting PBR had a gadliquid phase separationmodule allowing for gas exit (Fig. 1). Introducing spargingadded some complexity to the system because the gas ana-lyzer used had to have a hermetically sealed gas phase tomeasure the gas consumptiodproduction rate accurately(see Materials and Methods section).MATERIALS AND METHODSCell Line and Culture MediumC. vulgaris (UTEX 398, recently renamed as C . kessleri)was obtained from The Culture Collection of Algae atUTEX (Austin, TX) on proteose agar. The N-8 culture me-d ium was used wi thout m ~ d if ic a t io n . ~ ~LED, Light Measurements, and Power SuppliesRed DDH GaAlAs LED were obtained from Quantum De-vices Inc. (Barneveld, WI). These LED have narrow spec-tral outputs that peak at wavelengths of 680 nm. The LEDwere powered by simple DC power supplies (GP-105,GoldS tar Precision, C erritos, CA) w ith constant voltage be-tween 1.7 V and 2.1 V per device, except one point where2.4 V were applied.The light intensity of the LED units was measured by asilicon photo cell (M odel 0560.0500, Testoterm GmbH &Co., Germany). The light intensities of a high-intensityLED unit were about 3 and 12 mW/cm2 when powered at1.7 and 2.1 V , respectively.Culture ConditionsTwo different types of LED units were constructed: a low-intensity unit with 24 L ED in a 2 X 4 printed circuit boardand a high-intensity one (a closely mounted LED arraywhich had 90 LED in the same PCB, about 0.76 cm apart).The maximum intensity achieved by two high-intensityunits (one on each side of the PBR) was 50 mW/cm2 byproviding a power of 2.4 V per L ED. However, the PBRwere normally powered at 2.1 V/device to give a light in-

icro

L iq u id G a s Separator

Figure 1internal sparging and on-line ultrafiltration.

Schematic diagramof the LED-based prototypePBR hat uses

tensity inside the PBR of 23 m W/cm 2. All the experime ntswere performed at a constant temperature of 25C and initialpH of 6.2 , and were done batch-wise with N-8 medium. Aperistaltic pump was used to circulate the culture mediumand to keep the cells in suspension. Premixed gas with thecomposition of 15% O,, 5% CO,, and balanced with N,was introduced through the gas analyzer and refreshedwhenever either 0 or CO, concentration had changed morethan 0.5% from the premixed concentration.The seed culture was prepared by 500-mL flasks with150-m L working volume using an L ED unit w ith 65 of 680nm LED. The flask was doubly wrapped with extra-heavy-duty aluminum foil to block the ambient lights.

Cell AnalysisSamples were prepared by diluting the cell culture brothwith the appropr iate amount of iso tonic diluent (Fisher Sci-entific, Pittsburgh, PA). The cell concentration was deter-mined by a m icroprocess or-controlled electronic particlecounter, Coulter Counter model ZM with a Channelizermodel C256 (both from Coulter Electronics, Inc., Hialeah,FL) as well as by occasional confirming counts using ahemacytometer. Sinc e the average cell size was highly vari-able (about 100 fL/cell or pm3/cell while the cells are ac-tively growing and about 30 fL/cell at stationary phase), th etotal cell volume [ cell concentration (cells/mL) X aver-age cell volume (fL/cell)] was used to calculate the volu-metric specific growth rate and thus the doubling time.Gas Composition Measu rementThe gas concentrations were measured by a M icro-Oxymaxgas analyzer (Columbus Instruments, Columbus, OH ). Thissystem was chosen due to its high sensitivity to detect thesmall gas composition change by 100 mL of algal culture.To obtain the desirable sensitivity, a slightly pressuredclosed system in which the air in the gas phase of a testchamber is pumped through the gas sensors and then re-turned to the test chamb er, was necessar y. The 0 and CO,gas conce ntrations in the air in the test chamber were m ea-sured periodically, and the changes in the conce ntrations aswell as the volumes of the algal culture and gaseous phasewere used to com pute gas consum ption or produ ction rates.Whenev er the concentration of either gas in the closed gasloop deviated more than -C0.5 , the gas phase was re-freshed from a gas cylinder with the premixed gas. A ref-erence chamber containing a reference supply of air wasused to recalibrate the sensors prior to each measurement.The oxygen sensor operated as an oxygen battery (fuel cell)and the carbon dioxide sensor was a single-beam nondis-persive infrared device.RESULTS AND DISCUSSIONPerformance of the Prototype PBRA schematic diagram of prototyp e PBR w ith internal sparg-ing with gad liquid s eparator is shown in Figu re 1. A typical

1164 BIOTECHNO LOGY AND BIOENGINEERING, VOL. 44, NO. 10, NOVEMBER 20, 1994


5/7

growth curve in this PBR with low light-intensity units (a24-LED unit on each side, 1.55 cm apart each other) isshown in Figure 2 (closed symbols). Ultrafiltration (UF)was used to semicontinuously feed the nutrients withoutremoving cells from the PBR system. The first ultrafiltra-tion was performed at 92.3 h and the average cell sizeimmediately increased. This means that cells actively di-vided. UF was performed every other day starting at 115 .5h, w here the average cell size started to increase again. Thesuccessive UF did not show the same visible effect as thefirst two. How ever, it was necessa ry to keep the cells alive.The volume of fresh medium used by each UF was 2- to4-fold of the total PBR volume, depend ing on cell concen-tration. The specific growth rate in the logarithmic phasewas 0.035 h- ' (a doubling time of 20 h), resulting in amaximum cell density at the end of the operation of morethan 2% (v/v). However, it took over 350 h to increase thecell concentration from 1 * lo8 (at 1 40 h) to 8 lo8 (closedsymbols). This growth was extremely slow because a 100-fold increase in cell num ber could be achieved in about 100h following the initial lag phase.Monitoring the oxygen production rate was challenging.The gas analyzer required a hermetically sealed loop tomeasure the gas composition, and on-line ultrafiltration wasrequired to support cell concentration greater than 1 lo8cells/mL. Even though the setup was carefully designed tomeasure gas concentration, the gas reading (oxygen produc-tion rate per L culture, OPW L) fluctuated whenever the gasloop was disturbed by cell sampling, UF, or to refresh thegases in the hermetically sealed gas loop. The highest ox-ygen production rate achieved in this configuration (thePBR with low-intensity LED units and 1.55-cm light path)was about 2 mmol/L culture/h at 150 h (dotted line in Fig.3). The second peak in OPWL at 330 h was a result ofincreasing the power to the LE D units to 2.4 V . By applyingthis voltage, the light intensity could be increased about

Time (days)0 5 10 15 20 25. . . ~. . . 1 . . . ; . . . I . . , . f 1 0 0

~-9080

-- 70--

-- 6-~ 50-- 4 0-- 30

0 100 200 300 400 500 600Time (hr)

Figure 2. The growth curves 0nd 0 ellslml) and correspondingaverage cell sizes 0 nd W Wcell) in the two different light intensities.Closed symbols for the low-intensity 48 LEDlchamber or 24 LED/illuminating surface) and open markers for high-density LED units 180LEDlchamber 90 LED each on both sides).

Time (days)0 5 10 15 20 25

Time (hr)Figure 3.and high-intensity (-) PBR rom the same conditions as in Figure 2.Oxygen production comparison between low-intensity (----)

100% (more than 40 mW /cm2). However, this voltage wasnot within the normal operating range of the LED used.Therefore at 2.4 V , the LED generated a significant amountof heat and finally some of the individu al LED were burnedout. All the later experiments used 2.1 V as the maximumpower.These results implied that this prototype PBR was oper-ating in a light-limiting environment. High efficiency inlight absorption of algae resulted in uneven accessibility tothe incident light or mutual shading. P erfect mixing wouldovercome this limitation. A high degree of m ixing generallyintroduces a high degree of shear stress, to which algae aresensitive. 1 Another way to address the light-limitation wasto deliver more photons to overcome mutual shading. High-intensity LED assemblies were constructed by increasingthe number of LED per unit (to 90 LEDhnit) rather thanincreasing the pow er supplied to each unit.Modifications to Improve the Performance ofthe PBRThe original low-intensity LED units w ere replaced by twohigh-intensity units 1.55 cm apart from each other. Theexpected change in cell growth was obtained (open symbolsin Fig. 2). The higher number of LED cou ld support a cellconcentration of 1.2 X lo9cells/mL ( ~ 5 %/v) and a spe-cific growth rate during the expon ential phase of0.046 h- '(a doubling time of 1 5 h). T he 2.5-fold differe nce in thetotal biomass ( v/v) between a low-intensity PBR and ahigh-intensity one was consistently observed in several in-dependent experiments. The difference in the total biomasswas sma ller than the difference in the light intensity, whichwas 3.75-fold. Ag ain, once the culture entered the station-ry phase, the growth rate dropped significantly and theculture was able to grow to very high cell concentrationsonly if a series of UF were performed.The pe k oxygen production rate at the en d of the loga-rithmic phase (170 h) was abo ut 4.5 rnmoVL culture/h (solid

LEE AND PALSSON: LED-BASED ALGAL PHOTOBIOREACTORS 1165


6/7

line in Fig. 3). A higher rate of up to 6 mmoVL culture/hcould be obtained by the higher voltage (2.4 V; see previoussection) at 430 h. This was a 3-fold increase over the per-formance obtained by using the original lower-intensityLED units in the PBR. The ratio of improvement wassmaller than the ratio of light intensity, suggesting that evenat the higher light intensity the PBR eventually becamelight-limited.A final design modification was to reduce the cross-section of the P BR to allow the light to penetrate the entireculture volume inside the PBR. The depth of the culture (orthe distance between the two illuminating surfaces) was1.55 cm in all the PBR mentioned before. A PBR with ashorter light path of 1.00 cm and two high -intensity LEDunits was constructed. Figure 4 hows the growth data froman experime nt with the new PBR (open sym bols) and a PBRwith original thickness with high density (closed symb ols).A specific volumetric growth rate of 0.058 h- ' (a doublingtime of 12 h) in exp onential phase and a final cell concen-tration greater than 2 X lo9cells/mL (more than 6.6% v/v)were obtained in the PBR with the reduced depth. By re-ducing the light path by 35 %, there was a 2-fold increase inthe maximal cell density, and the exponential growth ratewas increased by 20%. Thus, a com parable cell concentra-tion to the high est algal cell con centration reported33 wasachieved by monochromatic red lights from LED.A similar improvement in OP R/L was obtained (solid linein Fig. 5 , whose peak production rate was 10 mmol/Lculture/h at 105 h. The S OPR at this point was less than 20fmol/cell/h at a cell density of about 5 lo8cells/mL. Thisresult re-emphasizes the need to distribute the light moreevenly to ind ividual cells. Better light distribution could beobtained by either (1) achieving a higher degree of fluidmixing without damag ing the cells, o r (2) reducing the cul-ture depth.

Time (days)0 2 4 6 8 10 12

--120

-- 100

--80

--60

200 50 100 150 200 250 300Time (hr)

Figure 4. The cell concentration 0nd 0 ells/mL) and the averagecell size 0 and W fllcell) of the original PBR (closed symbols, thicknessof the illumination chamber: 1.55 cm) and PBR with thinner illuminationchamber (open symbols, thickness: 1.0 cm).

Time (days)0 2 4 6 8 10 12

10 --

0 100 150 200 250 30 0Time (hr)

Figure 5. The oxygen production rates in a thicker PBR (----) and athinner PBR (-), both with high light intensity 180 LED/PBR). Condi-tions the same as in Figure 4.

CONCLUSIONLED-based photobioreactor systems were designed, con-structed, and tested to achieve high-density photoauto-trophic algal cultures. These PBR used LED e mitting at 680nm as the sole light source. The LED units were locatedinternally for direct illumination and could em it the photo-synthetic photon flux of up to 50 mW /cm2. By introducing(1) internal sparging (which improves overall light utiliza-tion), (2) on-line ultrafiltration (supplying medium co mpo -nents semicontinuously while removing any secondar y me-tabolites), (3) higher light intensity (increasing the nu mbe rof LED per illumination area), and (4) a shorter light path,both the highest cell concentration (more than 2 X lo9cells/mL) and the highest total cell volume fraction (morethan 6.6% v/v) were achieved. By modifying these fourfactors, the oxygen production rate in the final PBR wasincreased almost 5-fold from the initial PBR, giving themaximal oxygen production rate of 10mmol oxyge dL cul-ture/h. Further challenges are to improve light distributionand continuously perfuse fresh medium of the prope r com-position to reach the stated goal of a S OPR of 100 fmol/cell/h at a cell concentration of 1 * lo9 cells/mL.

This project was funded in part by Grant 10-56943 from NASA.

References1. Barta, D. J. Tibbitts, T. W., Bula, R. J. Morrow, R. C. 1990. Ap-plication of light emitting diodes for plant irradiation in space bases.COSPAR Meeting in The Hague, The Netherlands, June 28, 1990.2. Beardall, J. 1985. Occurrence and importance of HC0,- utilizationin microscopic algae, pp. 83-96. In: W . J. Lucas and J. A. Berry(eds.), Inorganic carbon uptake by aquatic photosynthetic organisms.American Society of Plant Physiologists, Rockville, MD.3. Benemann, J. R., Koopman, B. L., Weissman, J. C., Eisenberg,D. E., Goebel, R. P. 1980. Development of microalgae waste watertreatment and harvesting technologies in California, pp. 457-496. In:G. Shelef and C. J. Soeder (eds.), Algae biomass: Production and

use. Elsevier/North-HollandBiomedical Press, Amsterdam, Holland.

1166 BIOTECHNOLOGYAND BIOENGINEERING, VOL. 44, NO. 10, NOVEMBER 20, 1994


7/7

4 . Borowitzka, M. A ,, Borowitzka, L. J . 1988. Micro-algal biotechnol-ogy. Cambridge University Press, Cambridge, U.K.5. Burgess, J. G., Iwamoto, K., Miura, Y., Takano, H., Matsunaga, T.1993. An optical fiber photobioreactor for enhanced production of the

marine unicellular alga Isochrysis aff. albana T-Iso (UTEX LB2307) rich in docosahexaenoic acid. Appl. M icrobiol. B iotechnol. 39:4 5 6 4 5 9 .6 . Damall, D. W., Greene, B., Henzi, M. T., Hosea, J . M., McPher-son, R. A., Snedson, J., Alexander, M. D. 1986. Selective recoveryof gold and other metal ions from an algal biomass. Environ. Sci.Technol. 20: 06-208.

7 . Faulkner, D. J. 1986. Marine natural products. Nat. Prod Rep. 3:8. Finn, R. K. 1954. Agitation-aeration in the laboratory and in industry.

Bacteriol. Rev. 18: 254-274.9 . Glombitza, K. W., Koch, M. 1989. Secondary metabolites of phar-

maceutical potential, pp. 161-238. In: R. C. C resswell, T . A. V.Rees, and H. Shah (eds.), Algal and cyanobacterial biotechnology.Longman Scientific & Technical, Harlow, UK.10. Greene, B., Henzi , M. T. , Hosea, J. M., Dam all, D. W . 1986. Elim-ination of bicarbonate interference in the binding of U(V1) in mill-waters to freeze-dried Chlorella vulgaris. Biotechnol. Bioeng. 28:764-767.

1 1 . Gudin, C., Chaumont, D. 1991. Cell fragility-the key problem ofmicroalgae mass production in closed photobioreactor. BioresourceTechnol. 38 2-3): 145-151.12. Hill, R., Bendall, F. 1960. Function of the two cytochrome compo-nents in chloroplasts: a working hypothesis. Nature 186: 13613 7 .

13. Hoppe, H. A,, Levring, T., Tanaka, Y. 1979. Marine algae in phar-maceutical science. Walter de Gruyter, Berlin, Germany.

14. Hosea, M., Greene, B., McPherson, R., Henzel, M., Alexander,M. D ., Damall , D. W. 1986. Accumulation of elemental gold on thealga Chlorella vulg aris. Inorg. Chim. Acta 123: 161-165.15. Javanmardian, M ., Palsson, B. 0 . 1991. High-density photoau-totrophic algal cultures: design, co nstruction, and operation of a novelphotobioreactor system. Biotechnol. Bioeng. 38: 1182-1 189.16. Javanmardian, M., Palsson, B. 0. 1992. Continuous photoau-totrophic cultures of the eukaryotic alga Chlorella vulgaris can exhibitstable oscillatory dynamics. Biotechnol. Bioeng. 39: 4 8 7 4 9 7 .17. Javanmardian, M. , Palsson, B. 0. 1992. Design and operation of analgal photobioreactor system. Adv. Space Res. 12:231-235.18. Kargi, F., Moo-Young, M. 1985. Transport phenomena in biopro-cesses, pp. 5-56. In: M. Mm-Young (ed.), Comprehensive biotech-nology, vol. 2. Pergamon P ress, Oxford, U.K.

19. Kessler, E. , Huss, V. A. R. 1992. Comparative physiology and bio-chemistry and taxonomic assignment of the Chlorella (Chloro-phyceae) strains of the culture collection of the University of Texas atAustin. J . Phycol. 28: 50-553.

20. Kok, B. 1960. Efficiency of photosynthesis, pp. 566-633. In: W.Ruhland (ed .), Encyclopedia of plant physiology. S pringer, Berlin,Germany.21. Langhans, R. W ., Dreesen, D. R. 1988. Challenges to plant growingin space. HortScience 23: 86-293.

22. Lee, E. T.-Y., Bazin, M. J. 1990. A laboratory scale air-lift helicalphotobioreactor to increase biomass output rate of photosynthetic al-gal cultures. New Phytol. 116: 331-335.23. Metting, B. 1988. Micro-algae in agriculture, pp. 288-304. In: M. A.Borowitzka and L. J. Borowitzka (ed s.), Micro-algal biotechnology.Cambridge University Press, Cambridge, U.K.

24. Mignot, L., Junter, G. A,, Labbe, M. 1989. A new type of immo-bilized-cell photobioreactor with internal illumination by optical fi-bers. Biotechnol. Techniques 3:299-304.

25. Miller, R. L. 1969. Design and preliminary evaluation of a m an-ratedphotosynthesis exchanger. USAF School of Aerospace Medicine,SAM-TDR-69-64, Brooks AFB, T X.26. Miyamoto, K., Wable, O., Benemann, J. R. 1988. Vertical tubularreactor for microalgae cultivation. Biotechn ol. Lett. 10: 03-708.

1-31.

27. Mori, K. 1985. Photoautotrophic bioreactor using solar rays con-densed by fresnel lenses. Biotechnol. Bioeng. Symp. 15: 331-345.28. Myers, J. 1980. On the algae: thoughts about physiology and mea-surements of efficiency, pp. 1-16. In: P. G. Falkowski (ed.), Primary

productivity in the sea. Plenum Press, New York.29. Myers, J . Graham, J . 1971. The photosynthetic unit in Chlorellu

measured by repetitive short flashes. Plant Physio l. 48: 282-286.30. Nielsen, E. S . 1966. The uptake of free CO, and HC 0,- during

photosynthesis of plankton algae w ith special reference to th e cocco-lithophorid Coccolirhus hu ley i. Physiol. Plant. 19: 32-240.

31. Olson, R. L., Oleson, M. W., Slavin, T. J. 1988. CELSS for ad-vanced manned mission. HortScience 23: 75-286.32. Pirt, S. J., Lee, Y.-K., Richmond, A., Watts-Pirt, M. 1980. Thephotosynthetic efficiency of Chlorella biomass growth w ith referenceto solar energy utilization. J. Chem. Tech. Biotechnol. 30: 5-34.

33. Pirt, S. J . Lee, Y .-K., W alach, M. R., Pir t, M . W. , Balyuzi ,H. H. M., Bazin, M. J . 1983. A tubular photobioreactor for photo-synthetic production of biomass from CO,: design and performanc e.J. Chem. Tech. Biotechnol. 33B: 5-58.

34. Pratt, R. 1944. Studies on Chlorella vulgaris. IX. Influence on growthof Chlorellu of continuous removal of chlorellin from the culturesolution. Am. J. Bot. 31: 18-421.

35. Pratt, R., Daniels, T. C., Eiler, J . B., Stain, H. H. 1944. Chlorellin,an antibacterial substance from Chlorellu. Science 99: 351-352.36. Ratchford, I. A. J., Fallowfield, H. J . 1992. Performance of a flatplate, air-lift reactor for the growth of high biomass algal cultures. J .37. Richmond, A. 1986. Microalgae of econom ic potential, pp. 199-243.

In: A. Richmond (ed.), Handbook of microalgal mass culture. CRCPress, Boca Raton, FL.

38. Schreiber, U. 1983. Chlorophyll fluorescence yield ch anges as a toolin plant physiology. I. The measuring system. Photosynth. Res. 4:39. Shinho, K. 1986. Antitumor glycoproteins from Chlorella and other

species. Jpn. Kokai Tokk yo Koho Japanese patent JP 61-069728,105:030034t.40. Steranka, F. M., DeFevere, D. C., Ca ma s, M. D., Tu , C.-W.,

McElfresh, D. K. , Rudaz, S. L., Cook, L. W., Snyder, W. L. 1988.Red A lGaAs light-emitting diodes. H ewlett-Packard J. 39 8 4 8 8 .

41. Takano, H. , Takeyama, H., Nakamura, N . , Sode, K., Burgess,J. G., Manabe, E., Hirano, M., Matsunaga, T. 1992. CO, removalby high-density culture of a marine cyanobacterium Synechococcussp. using an improved photobioreactor employing light diffusing o p-tical fibers. Appl. Biochem. Biotechnol. 34/35: 4 9 4 5 8 .

42. Torzillo, G . , Carlozzi, P., Pushparaj, B., M ontaini, E., Materassi, R.1993. A two-plane tubular photobioreactor for outdoor culture of Spir-ulina. Biotechnol. Bioeng. 42: 91-898.

43. Treat, W . J . Castillon, J. , Soltes, E. J. 1990. Photobioreactor cultureof photosynthetic soybean cells: growth and biomass characteristics.Appl. Biochem. Biotechnol. 24/25: 97-510.44. Tredici, M. R., Materassi, R. 1992. From open ponds to verticalalveolar panels: the Italian experience in the development of reactorsfor the mass cultivation of phototrophic microorganisms. J. Appl.

45. Vonshak, A. 1986. Laboratory techniques for the cultivation of mi-croalgae , pp. 117-145. In: A. Richmond (ed .), Handbook of microal-gal mass culture. CRC Press, Boca Raton, FL.46. Walker, D. A. 1989. Automated m easurement of leaf photosynthetic0 evolution as a function of photon flux de nsity. Phil. Trans. R. SocLand. B323: 313-326.

47. Wang, S. C., Jin, M. R., Hall, D. 0. 1991. Immobilization of An-abaena azollne in hollow fibre photobioreactors for ammonia produc-tion. Bioresource Technol. 38 2-3): 85-90.

48. Yongmanitchai, W., Ward, 0. P. 1992. Grow th and eicosapentaenoicacid production by Phaeodactylum tricornutum in batch and contin-uous culture systems. JAOCS 69: 584-590.

Appl. Phycol. 4: 1-9.

361-373.

Phycol. 4: 21-231.

LEE AND PALSSON: LED-BASED ALGAL PHOTOBIOREACTORS 1167

photobioreactors using led

Documents