Photosynthetic microorganisms can be engineered to produce pharmaceuticals, chemical intermediates (1, 2), and clean energy (e.g., hydrogen) (3). They also fix atmospheric carbon dioxide (4)--an important consideration as increased levels of carbon dioxide are linked to global warming. It is expected that, in the future, photosynthetic microorganisms will play a larger role than higher plants in photosynthetic carbon dioxide fixation because they have higher photosynthetic rates per unit biomass and, if optimized, can be cultivated in a compact space (5).
The potential of microalgae as a food staple in the human diet has been investigated over many years in several countries (6). In Japan, algal biomass (e.g., Chlorella and Spirulina sp.) is produced commercially, primarily as health foods. Protein and chlorophyll contents are important parameters for assessing the quality of an algal biomass. Because no efficient large-scale photobioreactors are yet available, open cultivation ponds are used for almost all commercial algae production. However, it is difficult to obtain high productivity in open ponds because the temperature and light intensity vary throughout the day and year. In addition, open ponds require a large surface area, and problems with contamination arise. Because of the prohibitively expensive price of land in Japan, almost all algae is produced in Southeast Asian countries. Even so, the prices of microalgal biomass and alga-derived compounds are still very high. The development of large-scale tank-type photobioreactors is critical for the production of Chlorella and other algal products in Japan. To fully exploit the potentials of photosynthetic cells, efficient photobioreactors must be developed.
Many closed photobioreactors have been used or proposed for the cultivation of microalgae; the most common are vertical or horizontal tubular (7), helical (serpentine) (8), and inclined or horizontal thin-panel (9) photobioreactors. The critical design requirement in these photobioreactors is to supply light efficiently by maximizing the illumination surface-to-volume ratio of the reaction. As a result, tubes are often very narrow or the panels very thin. Some of the photobioreactors that work well in the laboratory may not work as well when scaled up because the surface-to-volume ratio decreases, causing poor light distribution inside the reactor. To produce alga-derived materials at competitive prices, efficient large-scale photobioreactors must be designed. We have designed and constructed a photobioreactor for the large-scale cultivation of photosynthetic cells of Chlorella sp. in which scale-up is a primary design criterion.
As a first step in the photobioreactor design, we investigated the relative significance of the exponential and the linear growth rates during light-limited batch cultivation of photosynthetic cells using various types and sizes of photobioreactors. The results indicated that there was not good correlation between the specific growth rates and the linear growth rates or between the specific growth rates and the final cell concentrations during the cultivation of Chlorella pyrenoidosa C-212 and Spirulina platensis M-135 cells (13). However, regardless of the type and size of the photobioreactor, we observed good correlation between the linear growth rates and the final cell concentrations for C. pyrenoidosa and S. platensis.
We also developed a mathematical model that could explain the existence of the various growth phases during the light-limited batch cultivation. The model predicts that the linear growth phase is longer than the exponential growth phase under various conditions (13). Thus, the linear growth rate is a better growth index than the specific growth rate and was used as the growth index in this study.
The concept of mean light intensity is an improvement over the incident light intensity but does not consider light distribution within the photobioreactor. We investigated the reliability of the incident and the average light intensities as indices of light conditions in cuboidal photobioreactors of various sizes (16). Although photobioreactors of the same size had similar relationships between the linear cell growth rates and the incident or average light intensities, we found no correlation between the linear growth rates and the above light parameters when data from photobioreactors of different sizes were considered. This finding implies that the apparently good correlation reported between the photosynthetic growth rates and the incident light intensities is probably the result of a single photobioreactor being used in each of those studies (14). The incident and the average light intensities are therefore not good indices of light supply efficiency of photobioreactors and cannot be used for meaningful evaluation of light conditions inside photobioreactors of different sizes.
According to Einstein's law of photochemical equivalence, the photosynthesis rate (and hence the cell growth rate) should be proportional to the rate of light energy absorbed by the cells. During the linear growth phase, the cell concentration in the photobioreactor is fairly high, so depending on the depth of the reactor, almost all the supplied light energy is absorbed by the growing cells. The total light energy supplied per unit volume of photobioreactor (Et/V) therefore would be a better measure of photobioreactor performance than the incident or the average light intensities. The relationships between the Et/V and the linear growth rates are shown in Figure 1. Although the results reveal a nearly linear relationship, the data are scattered near the curve. At a given Et/V, the linear growth rates decreased with an increase in depth of the photobioreactors, indicating that the light distribution inside the reactor must be considered in the rational design and scale-up of photosynthetic processes. Compared with the fairly homogenous distribution of light inside a very shallow photobioreactor, there is a distinct spatial heterogeneity of light intensities inside the deep photobioreactors.
),
0.04 m (
), 0.06 m (
), 0.08 m (
),
and 0.16 m (
).
When a photobioreactor containing a high cell concentration is illuminated from the surface, light is absorbed rapidly by the cells at the surface, and light intensity decreases sharply into the center of the reactor. As a result, the photobioreactor can be divided into illuminated and nonilluminated (dark) volume fractions. The illuminated volume fraction is defined as the ratio of the photobioreactor volume that receives sufficient light for cell growth to the total volume of the photobioreactor. This illuminated volume fraction can be used as an indicator of the light distribution inside the photobioreactor. We therefore proposed the concept of a light distribution coefficient (Kiv) defined as the cell concentration at which 50% of the photobioreactor volume receives enough light for photosynthetic growth. The higher the Kiv, the more uniformly light is distributed within the photobioreactor.
The significance of this coefficient can be seen in the comparison of
the Kiv values of two 0.02-m-deep cuboidal photobioreactors
(Figure 2). Photobioreactor A is illuminated from one
surface with an incident light intensity of
325µmol/m2
s,
and B is
illuminated from two surfaces at incident light intensities of 162.5
µmol/m2s.
By assuming a critical light
intensity of 7.65 µmol/m2s and a light
extinction coefficient of 200 m2/kg (13), the
effect of cell concentration on the illuminated volume fraction was
calculated. Although the Et/V in the two
photobioreactors is the same, photobioreactor B is more
uniformly illuminated; consequently, the Kiv value for
B (3.1 kg/m3) is higher than that for
A (1.9 kg/m3).
s) is the same in the two photobioreactors, the calculated light
distribution coefficients were 1.9 kg/m3 () and 3.1 kg/m3 () for photobioreactors A and B, respectively.
We investigated the effects of Kiv on the linear growth rates and found that the linear growth rates increased with an increase in Kiv. However, as in the case of Et/V, the data were scattered, showing that Kiv alone is not a sufficient index of light supply efficiency of photobioreactors. At a constant Kiv, however, a linear relationship was observed between the linear growth rate and the Et/V. Similarly, when the Et/V was held constant, there was a good correlation between the Kiv and the linear growth rate.
On the basis of these results, we proposed the light supply
coefficient--defined as the product of the light energy supplied per
unit volume and the light distribution coefficient
(Et/V
Kiv)--as an index
of the light supply efficiency in photobioreactors. There was a linear
relationship between the light supply coefficient and the linear growth
rates of C. pyrenoidosa and S.
platensis in cuboidal photobioreactors of various sizes. When
other internally illuminated and externally illuminated cylindrical
photobioreactors were used, we found good correlation between the
linear growth rates of C. pyrenoidosa and the light supply
coefficient (16), indicating that the proposed light supply
coefficient can be used to quantitatively evaluate light condition
inside the photobioreactor, regardless of the cell type, reactor type,
or size.
Kiv; ) and the volumetric oxygen transfer coefficient (k
La; ) on the
photoautotrophic linear growth rate of C. pyrenoidosa.Because of the rapid light attenuation inside photobioreactors, spatial heterogeneity of light intensities occurs inside the photobioreactor. Thus, even when light energy is available in the entire photobioreactor, limitation of light energy is the most commonly encountered problem in large-scale cultures of photosynthetic cells. Conventionally, illumination surface area per unit volume is used as a photobioreactor design criterion. An efficient photobioreactor has a high surface area-to-volume ratio. In a cultivation pond, this is achieved by keeping the pond as shallow as possible. Tubular photobioreactors (7) and thin panels (9) are the most widely investigated closed systems for photosynthetic cell cultivation, because they have high surface area-to-volume ratios. However, many economic and technological problems in the scale-up of such photobioreactors result from the need to keep the surface area-to-volume ratio high, limiting tube diameters and panel depths.
The effects of Et/V
Kiv
on the linear growth rates of C. pyrenoidosa in various
types and sizes of photobioreactors are also shown in Figure 3. The
inflection observed at high
Et/V
Kiv values implies
a decrease in the photosynthetic efficiency in response to an excessive
supply of light energy. The relationship between the light supply
coefficient and the linear growth rate of S. platensis in
cuboidal photobioreactors of various sizes was also linear. Therefore,
this coefficient is a good index of light supply efficiency of
various types of photobioreactors. Because photobioreactor
performance is determined by the availability of light energy,
this light supply coefficient can be used not only to evaluate various
types of photobioreactors but also for rational design and scale-up of
photobioreactors. When a photobioreactor with the optimal light supply
coefficient for a target process is designed, it should be efficiently
scaled up by keeping this coefficient at its optimum.
Large-scale processes. Many photobioreactors that have been proposed work well in the laboratory but are extremely difficult to implement. Commercial, scale-up potentials should be a primary design criterion for photobioreactors. A successful scale-up will be achieved only if the data obtained with a small-scale reactor can be reproduced with large-scale reactors. To achieve this, the factors that affect cell growth and productivity must be maintained within the optimal range as the reactor is scaled up from the laboratory to the industrial scale.
Like other microorganisms, the growth and productivity of photosynthetic cells are affected by many factors, including media components, temperature, mass transfer characteristics, pH, and concentrations of O2 and CO2 in the reactor. However, as shown in Figure 3, the light supply is more important than mass transfer rate during autotrophic cultivation of photosynthetic cells. Based on these results, we used the light supply coefficient as the design and scale-up parameter.
To maintain the constant light condition inside the photobioreactor (the light supply coefficient), we proposed a new concept in design and scale-up. We conceptualized the photobioreactor as being comprised of units. One unit is a reactor volume (space) with a single light-distributing object. If the intensity of the light-distributing object is constant, then the light supply coefficient of a unit decreases with an increase in the size of the unit. At a constant light intensity, therefore, there is an optimal unit size for a given cell and process. An optimal unit size for a process is determined first, and the photobioreactor is scaled up by increasing the number of this unit in three dimensions. In this way, the optimal light supply coefficient of the reactor is maintained during the scale-up.
One reactor for different processes. Usually, one bioreactor is used to cultivate various types of cells and produce various metabolites. These processes can be done by using the appropriate substrate and controlling the temperature, aeration, pH, and other factors as desired. Because the difference between an ordinary bioreactor and a photobioreactor is the presence or absence of light, it is reasonable to consider light as a part of the photobioreactor. The light requirements of cells and processes vary greatly; consequently, the optimal light supply coefficient (optimal unit size) depends on the cell type and the process. Thus, each cell and process requires a different photobioreactor.
For economic reasons, it is desirable to use the same photobioreactor for several processes; therefore, it should be possible to change the light supply coefficient of the photobioreactor to suit the process. The light supply coefficient of a photobioreactor is a function of the size of each unit (the distance between two light-distributing objects) and the light intensity. It is technically easier to change the light intensity than the distance between the light-distributing objects. A photobioreactor can be used for various processes by changing the intensity of the light-distributing objects, either by using a light source with controllable intensity or by changing the light source altogether. In this way, the light supply coefficient of the photobioreactor can be changed, even during cell cultivation. This design allows for starting with low light intensity at the initial stages of growth when the cell concentration is still low and then increasing the light intensity as the cultivation progresses.
Hydrodynamic stress. Mixing is very important in photobioreactors. It helps to keep the cells in suspension, distribute the nutrients and the generated heat within the reactor, improve CO2 transfer into the reactor, degas the photosynthetically produced O2, improve mass transfer between the cells and the liquid broth, and facilitate the movement of cells in and out of the illuminated part of the photobioreactor. However, because the growth rates of most photosynthetic cells are very low, only a very low degree of mixing is required to achieve most of these objectives (17, 18). Furthermore, at high light distribution coefficients, the variation in light intensity within the reactor is minimal, so there is little advantage in moving the cells toward and away from the light source (18). Many photosynthetic cells have no cell wall, and some are mobile or filamentous, making them fragile and sensitive to shear stress. Therefore, it is desirable to keep the hydrodynamic stress as low as possible.
Cultivation under a sterile condition. The risk of contamination by heterotrophic microorganisms is low when there is no organic carbon source in the medium. However, at facilities where many other photoautotrophic cells are cultivated, contamination by other photoautotrophs can be a serious problem. Thus, the new photobioreactor should be able to withstand sterilization procedures.
On the basis of these objectives, we focused on constructing a sterilizable photobioreactor with good mixing properties but low hydrodynamic stress. We also wanted to ensure that the reactor could be efficiently scaled up for large-scale processes and could have a controllable light supply coefficient so that the same photobioreactor could be used to cultivate various cells with different optimal light supply coefficients.
), yield coefficient
(), and light supply coefficient
() of C. pyrenoidosa.
Each unit was internally illuminated by a 4-W fluorescent lamp producing
a light intensity of 163 µmol/m2s.
The light supply coefficient and thus the linear growth rate decreased with increasing unit diameter. The highest productivity (in this case, the linear growth rate) is obtained with very narrow units corresponding to high light supply coefficients. However, under such conditions, more light is supplied than the cells can use efficiently because of photoinhibition or energy loss in the form of heat. This process leads to low yield coefficients. The yield coefficient increases with increasing unit diameter, because at relatively low light supply coefficients, most of the supplied light energy is absorbed and efficiently used for cell growth and product formation. However, with large unit sizes, the light supply coefficient decreases and the illuminated volume fraction of the unit is low (i.e., a large portion of the unit is dark). This increases maintenance energy and decreases productivity and the yield coefficient (18, 19).
For a light source of constant intensity, the optimal unit size depends on the cell and the process economy. If light represents a significant percentage of the total production cost, then greater importance should be attached to the efficiency of light use, and the unit size giving the highest yield coefficient should be selected. However, if the cost of light is relatively cheap (such as the case of solar energy), the design criterion should be to obtain the highest productivity. A high light supply coefficient is desirable, provided that the light intensity is not high enough to cause photoinhibition. In most cases, a compromise is made between productivity and the yield coefficient.
For mixing, we used an impeller modified in shape so that it did not
touch the glass housing unit during rotation. This impeller has low
shear stress and a high mixing capacity. With this impeller, a
kLa value of 100 h-1 was achieved
at an aeration rate (with air) of
0.5 V/Vmin and an agitation
speed of 250 rpm. This value is enough to prevent cell sedimentation,
achieve a sufficient rate of CO2 transfer to the culture,
and prevent O2 build-up within the reactor (20).
Aeration was done through a 5.5-cm-diameter ring sparger with four
1-mm-diameter holes. The glass housing units serve as baffle plates in
breaking the gas bubbles, thus increasing the
kLa. When a higher gas transfer rate is
required, a sparger with numerous smaller diameter holes should be
used. One benefit of this design is that only the reactor is heat
sterilized. The lamps are inserted after cooling, thus making it
possible to cultivate under sterile conditions (21).
We investigated production of C. pyrenoidosa biomass in our
photobioreactor using 4-W fluorescent lamps, which give a light supply
coefficient of
0.374 kJkg/m6
s.
The linear growth rate (0.164 kg/m3day)
we obtained was consistent with the light supply coefficient, and the
final cell concentration was 1.37 g/L. This concentration is
relatively low and would result in high cost of downstream
processing. When the light supply coefficient was increased to
0.692 kJkg/m6
s,
the linear growth rate and the final cell concentration increased
relative to the light supply coefficient; however, as expected, the
efficiency of the cell's use of the supplied light energy decreased.
Chlorella and some other algae can metabolize organic carbon sources, so heterotrophic cultivation can be used to achieve high cell concentrations. Under such conditions, however, the photosynthetically derived products are not accumulated and the main advantage of photosynthetic cell cultures is lost. Experiments in our laboratory have shown that the protein and chlorophyll contents of Chlorella biomass produced from heterotrophic cultures are much lower than those of the photoautotrophic cultures. In some other species (e.g., Euglena), chlorophyll synthesis is almost completely inhibited under heterotrophic conditions.
We investigated a sequential heterotrophic-photoautotrophic
cultivation system to produce a high concentration of C.
pyrenoidosa biomass with high cellular protein and chlorophyll
contents. This method involves passing a high concentration of
monoalgal biomass from a heterotrophic culture through a
photobioreactor for accumulation of protein and chlorophyll. Our system
was composed of the conventional mini-jar fermentor for the
heterotrophic phase and the new photobioreactor for the
photoautotrophic phase. The exhaust gas from the heterotrophic phase
was used for aeration of the photoautotrophic phase to reduce
CO2 emission (Figure 6). The mini-jar fermentor and the
photobioreactor were each filled with the media and inoculated with the
preculture of C. pyrenoidosa. The lamps of the
photobioreactor were turned off, and both reactors were wrapped with
aluminum foil to shut out the light. The cultures were then grown
heterotrophically using glucose as the carbon source. When the glucose
was completely consumed, the photobioreactor lamps were turned on;
continuous feeding of fresh medium into the mini-jar fermentor was
started; and the effluent was continuously passed into the
photobioreactor for protein and chlorophyll accumulation. Changes in
the cell concentration as well as chlorophyll and protein
contents of the cells in the photobioreactor are shown in Figure 7. It
was possible to produce high C. pyrenoidosa biomass
concentration (14 g/L) containing 63.5% protein and 2.5% chlorophyll
continuously for >600 h. During the steady state, the
CO2 concentration in the exhaust gas was reduced by
15% and the cell productivity was 4 g/Lday. This
productivity is much higher than the values reported to date in
photoautotrophic cultures.
), chlorophyll concentration (), and protein concentration () of C. pyrenoidosa cells in the photobioreactor during the
continuous sequential heterotrophic-photoautotrophic cultivation.
ChemCenter
Pubs Div. Home Page
