Work by SERI/NREL subcontractors in the early 1980s supported the idea that there is significant genetic variation within algal populations (i. e., Gallagher, Section ILB. Lc.). Therefore, one possible method for producing high lipid algal strains would be selection of natural genetic variants with desired traits, such as high lipid levels or increased tolerance to high salinity or temperature. The limiting factor to this approach has always been the difficulty of selecting for individuals exhibiting a desired trait among a large population of cells. The use of lipophilic dyes such as Nile Blue or Nile Red, coupled with flow cytometry, showed some potential for isolation of high lipid strains of microalgae (see work by Solomon and Cooksey, Sections II. B. 1.e and f). It was not clear, however, that the variations detected in subpopulations of cells were the result of genetic variations that would be passed on to progeny.

An alternative approach is to induce genetic variation in a population of cells using mutagenesis. Again, the ability to select for the desired trait is a limiting factor, but the production of large numbers of mutants by artificial means is a proven method for generating organisms with heritable traits, often the result of a mutation within a single gene. As a prelude to the initiation of mutagenesis and selection experiments with oleaginous microalgae, NREL researcher Ruth Galloway performed a series of experiments designed to understand the factors required to produce mutants in microalgae. These included media requirements for growth, the ability to form colonies on agar plates, sensitivity to herbicides and other growth inhibitors, and the sensitivity of algal strains to mutagens such as UV light or fluorodeoxyuridine (Galloway 1990). Nine algal strains from the SERI Culture Collection were tested, including organisms from three classes: the chlorophyceae (M. minutum MONOR1 and MONOR2), the eustigmatophyceae (Nannochloropsis (NANNP1 and NANNP2), and the bacilliarophyceae (C. cryptica T13L, C.

mulleri CHAET9, Amphora AMPHO17, Nitzschia pusilla NITSC12, and N. saprophila NAVIC1).

Growth of each strain was evaluated qualitatively after spotting the cultures onto media containing nitrogen or carbon sources, or after growing the cells under phototrophic, mixotrophic, or heterotrophic conditions. Cell growth was also evaluated in the presence of a large number of growth inhibitors including various antibiotics and herbicides. Although there was some variability between the algal strains, several generalizations could be made. Most strains could use either NO3- or NH4+ as a nitrogen source. Mixotrophic growth on various carbon sources was more variable, and only AMPHO17, MONOR2, and CYCLOT13L were able to grow heterotrophically, using glucose as a carbon source. The ability to grow heterotrophically would be important for the isolation of photosynthetic mutants.

Predictably, antibiotics that inhibit bacterial cell wall synthesis such as ampicillin and carbenicillin did not inhibit the growth of the algal strains. Antibiotics that inhibit bacterial protein synthesis by binding to the 30S ribosome showed variation in their effects on algal growth. For example, all strains tested grew well on kanamycin and neomycin and showed no growth on erythromycin; while the growth response differed for the strains on spectinomycin and streptomycin. Whether this result was due to differences in 30S (organellar) ribosomal structure between the algal strains, to differences in uptake of the antibiotics by the individual strains, or to other factors that affect sensitivity, is unclear. All strains showed sensitivity to photosynthesis inhibitors diuron, metronidazol, and atrazine, and to the herbicide glyphosate (“RoundUp”), which affects the shikimic acid pathway. However, sensitivity varied between the strains to compounds that affect the enzyme acetolactate synthase and to chemicals that inhibit microtubule synthesis. (The details of these growth experiments can be found in Tables 2, 3, 4, 5, and 7 of Galloway 1990). Many of the growth inhibitors used in this study affect specific proteins in the target organism, and many of these proteins have been well characterized in a number of systems. Isolating the corresponding gene from an algal mutant using heterologous gene probes to characterize the mutation and/or to use the mutant gene as a selectable marker for transformation studies should be relatively easy.

Attempts were also made to generate mutants in the algal strains by exposing the cells to UV light or to fluorodeoxyuridine, followed by plating the cultures on toxic levels of various growth inhibitors. Using UV mutagenesis, streptomycin-resistant mutants were obtained in MONOR2, as well as glyphosate-resistant mutants in both strains ofNannochloropsis, sulfometuron methyl — resistant mutants in NANNP1, and atrazine-diuron-resistant mutants in NAVIC1. In addition, tunicamycin-resistant mutants of NAVIC 1 were produced following treatment with fluorodeoxyuridine. Mutants were not obtained for the other diatoms, CY CLOT 13L, CHAET9, or NITZS12, whether this was due to poor colony formation by these strains, inefficacy of the mutagen, or inappropriately high levels of the selective agent is not known. One interesting point was that the green algal strains, Monoraphidium and Nannochloropsis, produced mutants with traits thought to be due to recessive nuclear gene mutations (i. e., glyphosate resistance or photosynthesis mutants). On the other hand, in Navicula, the only diatom in which mutants were generated, the types of mutations produced were indicative of dominant mutations, i. e.,

atrazine/diuron resistance (resulting from a chloroplast gene mutation) or resistance to tunicamycin, an inhibitor of n-glycosylation. These results indicate that Monoraphidium and Nannochloropsis are probably haploid; the diatoms are diploid. Design of strategies for generation of algal mutants will have to consider the ploidy of the target organism. For example, generating nitrate reductase-deficient mutants for use in a genetic transformation system using homologous selectable markers (described in detail later) should be relatively simple in haploid strains, but would be much more difficult in diploids. For the diatoms, a better approach would be to utilize a dominant gene as a selectable marker, such as a mutant form of the enzyme acetolactate synthase (discussed later), or a heterologous gene such as the neomycin phosphotransferase II (NPTII). The latter gene confers drug resistance by inactivating antibiotics such as kanamycin or geneticin (G418).

In summary, the research performed by Dr. Galloway demonstrated the potential to produce algal mutants with a wide variety of phenotypes, particularly in the green algae, using simple mutagenesis and selection techniques. It would be important to first optimize and understand the growth conditions for the target strains. The conditions to be used for selection (inhibitor specificities and concentrations) should be determined for each strain. However, the generation of mutants will probably be more useful as a tool in developing selectable marker systems, rather than as a method to directly produce high lipid algal strains, primarily because there is no simple way to screen for high-lipid phenotypes. The use of mutagenesis to develop of homologous selectable marker systems for algal transformation will be discussed in detail later.

Mutagenesis and selection was used successfully in another study at NREL to generate mutants in one aspect of lipid synthesis, fatty acid desaturation (Schneider et al. 1995, described in Section II. B.2.g.). In this experiment, UV mutagenized cells of Nannochloropsis were allowed to form colonies, then grown in small-scale liquid cultures. Lipids were extracted from each sample and analyzed by gas chromatography for any significant alteration in the proportion of fatty acids. This project resulted in the identification of a mutant lacking in 20:5 fatty acids, apparently due to a mutation in a 20:4 desaturase. In this case, a simple screen was used to look for changes in a quantitative trait. This result suggests that, with the right method to screen for mutants with the desired properties, mutagenesis could result in microalgae with altered lipid compositions. However, this project was very labor intensive, with hundreds of colonies screened to identify a single mutant.

Numerous microalgal strains were obtained from the SERI Culture Collection and tested in small-scale, 1.4-m2, ponds (Weissman and Goebel 1986; 1987). All strains could be grown quite successfully in these small units, although some, such as Amphora sp., did not survive more than 2 or 3 weeks before they were displaced by other algae. Cyclotella displaced Amphora under all conditions tested, even though Amphora was the most productive strain, producing 45 to 50 g/m2/d in short-term experiments. The green algae, e. g. Chlorella or Nannochloropsis, also could not be grown consistently. Their productivities were among the lowest, about 15 g/m2/d (similar to that in the prior year). Thus, one fundamental conclusion was that productivity is not necessarily correlated with dominance or persistence. However, these factors may be related to oxygen effects, as shown in later experiments.

Table III. B.2. summarizes the results of various experiments for the summer with small (3.5-m2) ponds and seven of the algal strains under different operating regimes, including controlling the oxygen tensions through degassing by air sparging.

A significant factor in pond operations was the oxygen level reached in the ponds, which influenced productivity and species survival. Ponds were operated with air sparging (and antifoam) to reduce DO levels, from typically 400% to 500% of saturation without air sparing, to 150% to 200% of saturation with sparging. Foaming, caused by air sparging, was still was a problem in some cases, as with the Cyclotella. However this alga exhibited approximately the same productivity with or without sparging despite the 20%-30% opaque foam cover, suggesting some positive effect of the lower pO2. For other algal species productivity differences of 10% to 20% were noted, and for some (e. g., C. gracilis), no specific effect of high versus low DO was noted.

These outdoor results were reproducible enough to detect differences of greater than about 10% between treatments. The major result of this project was that productivities were 50% to 100% higher than the previous year, with some species of diatoms producing 30 to 40 g/m2/d (AFDW, efficiency about 6% to 9% of PAR, or 3% to 4.5% total solar). The green algae were, as mentioned earlier, less productive than the diatoms.

A more detailed study of oxygen effects was also carried out in the laboratory, avoiding the confounding factors of CO2 supply, temperature, and light intensity. In general the diatoms were insensitive to high DO; most, but not all, of the green algal strains exhibited marked inhibition by high oxygen levels (Figure III. B. 8.). None of the oxygen-sensitive algae could be grown outdoors, suggesting this as a major factor in species dominance and productivity.

Laboratory studies were also carried out at both high light intensity and high DO, to determine the synergism between these factors. Both the apparent maximum growth rate and dense culture productivity were determined for comparisons. Higher levels of DO intensified the inhibitory effects of higher light observed in some cases. This was true in particular for productivity, with growth rates also affected. Of course, the actual density of the culture is a major factor

determining productivity, and dense cultures avoid most, if not all, the deleterious effects of high light intensity. High O2 and low CO2 are other factors influencing the response to high light, with O2 being more inhibitory at both low CO2 and high light levels. High oxygen also affects chlorophyll content, although this effect is most pronounced at low light intensities where chlorophyll levels are 50% higher compared to high light intensities.

Outdoor experiments were carried out to determine the effect of low CO2 (25 pM) and high (9-10) pH, which would be experienced in algal mass cultures, at least temporarily. Compared to the control cultures, one strain was not inhibited even at pH 10, two not at pH 9, and two were inhibited by about 33% at this pH, compared to the control at pH 8. Lowering pCO2 also resulted in similar levels of inhibition for the other strains. A role for bicarbonate in growth at high pH was established from the data, with metabolic costs estimated at about one-third of productivity, a major factor. This requires further investigation.

One strain, a Cyclotella species, exhibited an increase of lipid content of more than 40% of dry weight upon Si limitation. However, lipid productivity (9 g/m2/d), was not significantly different between Si-deficient and the Si-sufficient controls, because of the high productivity of the Si — sufficient culture. Optimizing for lipid productivity was considered possible, but requires more detailed study.

Perhaps most important, the data and simulations also suggest that maximizing productivity at an acceptable CO2/pH combination from the perspective of outgassing and CO2 loss from the ponds is possible, with operations above pH 8.0 required (for an alkalinity of 32 meq/L, higher for higher alkalinities) to avoid wasting of CO2.

Laboratory studies were also carried out during this project. These included a study of light conversion efficiencies that concluded that at low light intensities very high light conversion efficiencies can be achieved (near the theoretical maximum of about 10 photons/CO2 fixed). However, these and other laboratories studies carried out during this project would require a much longer review than possible here.

Finally, this project investigated harvesting of microalgae cultures with both polymers and FeCl3 (to enhance algal settling and sludge compaction) and cross-flow filtration. Organic flocculants at about 2 to 6 g/kg and FeCl3 at about 15 to 200 g/kg of algal biomass (AFDW) were required to remove 90% or more of the algal cells. Because of the high cost of the organic flocculants, costs were comparable for both flocculants tested. The organic polymers were also deemed to have significant potential for improvement and optimization. Cross-flow filtration, though effective, was estimated to be too expensive. A cost analysis of such a harvesting process was also presented.

In conclusion, this project significantly advanced the state-of-the art of low-cost microalgae biomass production, and provided the basis for the Outdoor Test Facility, discussed in Section III. B.5., following the review of the ASP-supported project in Israel.

Table III. B.2. Outdoor results summary for California pond operations.

Data from 3.5-m[9]ponds. (Source: Weissman and Goebel 1986.)

Kcal calculated fro* proximate composition, either 1) as aeasured or 2) deterained as SOX protein, lipid as aeasured, and CHO by difference.

Teaperature: Max 30-34*C, Hin 16-20 °С 1 Required re-inoculation 3 Required re-inoculation twice

3 Oxygen reaoval caused 20-302 coverage of surface with foam

In Japan, the Research for Innovative Technology of the Earth program (RITE) has carried out an extensive progam for microalgal CO2 utilization. The Ministry of International Trade and Industry (MITI) funds this program, through the New Energy Development Organization (NEDO). The program was established in 1990 as a 10-year effort, carried out by approximately two dozen private companies, with some supporting work at various national laboratories and academic institutions. The budget is generally stated to be approximately $80 million (10 billion Yen) for 10 years; however, this is only for direct costs provided by RITE to the companies. If all indirect costs and supporting R&D at various institutes were to be included, this would easily double this budget. Also, microalgae-CO2 capture related R&D has been going on at the Japanese electric utilities, projects that are not part of the MITI-NEDO-RITE program. Thus, perhaps it is no exaggeration to estimate that during the 1990s the Japanese government and private companies will have invested more than $200 million (and perhaps closer to $250 million) in this research.

Although the Japanese microalgae R&D program is very large, it can be summarized rather briefly. Contrary to the U. S. approach, the Japanese effort has focused on closed photobioreactors, and on higher-value products. The argument made for closed photobioreactors was that these would require less land area than open ponds, because of much higher productivities. The higher productivities were assumed to be possible by using optical fibers to diffuse light into the reactors, and by greater control over environmental conditions (such as the ability to supply high CO2 levels to the cultures). Lower land requirements were also assumed to be possible with the optical fiber devices, although the land required for the concentrating mirrors was apparently not considered. The Japanese RITE program has yet to carry out, or at least report on, any technical, engineering, or economic analysis on such processes. A Japanese report (by researchers not part of the RITE program) to the IEA Greenhouse Gas R&D

Programme (IEA 1994) on microalgae “direct biofixation” of CO2, was based on prior U. S. engineering and cost analyses (Ikuta 1994), though they dried the algal biomass to replace coal.

The Japanese RITE program has presented some results on their microalgae genetics program and photobioreactor development (Murakami and Ikenouchi 1997; Usui and Ikenouch 1975). But these add little detail to the development of this technology.

One major emphasis of the Japanese program has been on developing high-value coproducts, from animal feeds to antibiotics to specialty chemicals. Some are rather esoteric, such as algae-based paper and concrete additives. In brief, the Japanese RITE Biological CO2 Fixation Program, and other Japanese R&D activities, perhaps in part by concentrating on such higher — value products, have not significantly advanced the technology for biofuels production or CO2 utilization, despite large investments.

One exception is the work carried out by Mitsubishi Heavy Industries (MHI) and several electric utilities, in particular Tohoku Electric Co., near the northern city of Sendai, in the early 1990s. There, a small pond (approximately 3-m2) project was carried out on the mass culture of diatoms and green algae. These studies initially used algal strains obtained from the NREL culture collection, and then with strains that spontaneously appeared and dominated the cultures at this site (Negoro et al. 1992, 1993). Productivity data were obtained and were generally in accord with the work at Roswell, New Mexico. (Two authors of this report, Benemann and Weissman, were consultants to this project.)

Another interesting project was carried out by MHI and Tokyo Electric Power Co. (TEPCO), which demonstrated actual increased productivity in optical fiber bioreactors. However, the complication and costs of these devices resulted in this project shifting to more conventional, air-lift tubular reactors. Recently TEPCO-MHI released a publicity announcement of a major breakthrough in the production of ethanol from microalgae biomass. However, little specific technical information is available on this work.

I Publications:

Ikuta, Y. (1994) “Design of a biological system for CO2 fixation, in carbon dioxide utilization- direct biofixation.” Report to the Int. Energy Agency Greenhouse Gas R&D Programme.

Murakami, M.; Ikenouchi, M. (1997) “The biological CO2 fixation and utilization project by RITE. 2. Screening and breeding of microalgae with high capability of fixing CO2.” Energy Conver. Mgmt. 38: Suppl. 493-498.

Negoro, M.; Shioji, K.; Ikuta, Y.; Makita, T.; Utiumi, M. (1992) “Growth characteristics of microalgae in high-concentrations of CO2 gas: Effects of culture medium, trace components and impurities thereon.” Biochem Biotech. 34/35:681-692.

Negoro, M.; Hamasaki, K. A.; Ikuta, Y.; Makita, T.; Hirayama K.; Suzuki, S. (1993) “CO2 fixation by microalgae photosynthesis using actual flue gas discharged from a boiler.” Biochem Biotech. 39/40:643-653.

Usui, N.; Ikenouchi, M. (1997) “The biological CO2 fixation and utilization project by RITE. 1. Highly effective photobioreactor system.” Energy Conver. Mgmt. 38: Suppl. 487-492.

Previous to the research performed by researchers in the ASP at NREL, very little work had been done in the area of microalgal strain improvement, particularly with a goal of developing a commercial organism. Although much remains to be done, significant progress was made in the understanding of environmental and genetic factors that affect lipid accumulation in microalgae, and in the ability to manipulate these factors to produce strains with desired traits.

The evidence for a specific lipid trigger is not overwhelming. Interpreting exactly what is happening in the nutrient-deprived cells is difficult, particularly when cells are starved for N, as the lack of an important nutrient is likely to produce multiple and complex reactions in a cell. However, lipid accumulation in some algal species can be induced by nutrient limitation. Cell division slows or stops, and the cells begin to accumulate lipid as cytoplasmic droplets, formed primarily of neutral TAGs. The trigger hypothesis is supported by microscopic and flow cytometric evidence that showed that the lipid droplets do not form gradually within all cells in a population; rather, individual cells seem to sense the trigger and lipid accumulation occurs rapidly within the individual cells. However, lipid accumulation is always correlated with the cessation of cell division. Other factors that inhibit cell division, such as a pH shift, can also induce lipid accumulation in some strains. The evidence suggests that the rate of synthesis of all cell components, including lipids, proteins, and carbohydrates, is decreased in nutrient-stressed cells. However, the rate of lipid synthesis remains, at least for some strains of diatoms, higher

than the rate of protein or carbohydrate synthesis, resulting in a net accumulation of lipid in nutrient-starved cells. Another hypothesis is that cessation of cell division in nutrient-limited cells leads to decreased utilization of storage lipid while new synthesis of lipid continues, causing a net accumulation of lipid in the cells.

One of the most important findings from the studies on lipid accumulation in the microalgae is that, although nutrient stress causes lipid to increase in many strains as a percentage of the total biomass, this increase is generally accompanied by a decrease in total cell and lipid productivity. As discussed elsewhere in this report, economically viable production of algal lipids for fuel production will require optimization of productivity as well as a clear understanding of the kinetics of lipid accumulation, in order to harvest the cells at a time when lipid production is maximal.

In addition to the effect on total lipid production, nutrient deprivation seems to have other effects on the lipid biosynthetic pathways. Several laboratories reported that nutrient limitation also resulted in a change in the types of lipids seen in the algal cells, specifically, an increase in the ratio of neutral lipids (storage TAGs that are important in the production of biodiesel) as compared to the polar membrane lipids. It will be important to characterize this phenomenon further in any algal strain targeted for biomass production to maximize the desired lipid product.

The progress made by the ASP in the understanding of the biochemistry and molecular biology of lipid biosynthesis in algae, and the success in the area of algal genetic engineering are more clear cut. An enzyme that appears to play a key regulatory role in the synthesis of lipids in plants, and likely in algae, ACCase, was purified from the diatom C. cryptica. The gene that codes for this enzyme was then cloned; this was the first report of cloning a full-length ACCase sequence from any photosynthetic organism. Obtaining this gene was advantageous for two reasons:

1. The regulatory sequences from this gene were used to develop a genetic transformation system for diatoms, and

2. Having this gene (in combination with the transformation system) allowed researchers to test for the effects of overexpression of this enzyme on lipid accumulation.

The development of the transformation system for an oleaginous microalgal strain was a major goal of the ASP, and a significant effort was put into this project during the early 1990s. This was the first successful transformation of any non-green alga. The method was simple and reproducible, and should work for a variety of diatom strains, as long as the cells can form colonies on solid medium and are sensitive to one of the known selectable agents, such as G418 or kanamycin. Although little work is currently being done on the development of genetically engineered algal strains for commercial applications, the ability to transform these algae should have positive ramifications for the algal biotechnology community.

Preliminary experiments were also performed within the ASP to use this genetic transformation system to introduce genes into the algal cells, with the goal of manipulating lipid biosynthesis. Additional copies of the ACCase gene were introduced into cells of C. cryptica and N. saprophila. Although ACCase activity was increased in these cells, there was no detectable increase in lipid accumulation. The project was terminated before these experiments, and similar experiments designed to down-regulate genes involved in carbohydrate synthesis could be pursued further. This could be an interesting and possibly rewarding path for future research, if only to help in understanding the biochemical and molecular biological factors that affect lipid accumulation in these cells.

III. Outdoor Studies and Systems Analysis

This report (Vignon et al. 1982), comissioned by the ASP, was the first comprehensive discussion of the resource requirements for microalgae production. It covered the criteria that should and could be used to identify available water, land and other resources, to estimate their relative importance, and to evaluate various legal, institutional, and other resource constraints. These issues were discussed at some length, focusing on the southwestern United States, although most of the discussion was of a rather general nature. For example, land and water rights issues are addressed, which are certainly important, but are difficult to extrapolate over large areas. Similarly, permits for such facilities will be very important, but will also involve site specific considerations.

The authors calculated various costs and energy inputs for water (seawater) pumping and transportation to arrive at permissible lift and distance criteria for water resources. Water lifting of some 75 m and pipeline distances of 6 km, with an approximately 1-m diameter pipeline, were estimated to cost about $31 million for a 400-ha system. This estimate for water supply is as high as later estimates for the total cost of building and operating an entire microalgae production system (see Section III. B.5.), which puts some perspective on the limits of lift and distance for water supplies. This report did not arrive at a prediction for the resource base, but was an important early introduction to the complexities of such resource assessments.

I Publications:

Vigon, B. W.; Arthur, M. F.; Taft, L. G.; Wagner, C. K.; Lipinsky, E. S.; Litchfield, J. H.; McCandlish, C. D.; Clark, R. (1982) “Resource assessment for microalgal/emergent aquatic biomass in the arid southwest.” Battelle Columbus Laboratory Report, Solar Energy Research Institute, Golden, Colorado.

The second year of this project emphasized the use of “flashing light to enhance algal mass culture production” (Laws, 1982; see also Laws et al. 1983). The basic idea was that a “foil array” in the pond culture would generate a vortex that would create organized mixing in the ponds, expected to result in exposure of the cells to regular dark-light cycles (Figure III. B.2.). Based on data in the literature, this effect would be predicted to increase overall productivity. These a priori arguments were not supported by the algal physiological literature (the flashing light productivity enhancements are observed at much shorter time constants), and neither were the hydraulic arguments plausible (organized mixing would be seen only in a small fraction of the pond volume). However, the key issue here is not the theory but the actual experimental results.

From November 1981 to January 1982, an average productivity of only about 3.3 g/m2/d was recorded for the 50-m2-flume reactor, a very low value for Hawaii, even in winter. After installation of the foils, productivities, from February to March 1982, increased to about 11 g/m2/d. This increase was attributed to the effect of the foils, though lack of a control did not allow isolation of this variable from other effects. Five-day running mean average photosynthetic efficiencies (PAR) are shown in Figure III. B.3. The author stated that productivity could be doubled with semi-continuous operations. One observation was infestation of the culture by algal predators, which could have been one reason for the rather large variability in productivities observed during this operation (Figure III. B.3.). However, day-to-day variability in productivities is a fact of outdoor pond microalgae cultivation, even in the best of cases.

Heat exchanger

Drain box

FV—«

Lilt box

Recirculation and

airlift

LOW PRESSURE OVER

FLOW»

HIGH PRESSURE

UNDER

CROSSOVER

VORTEX

Figure III. B.2. Hawaii ARPs with mixing foils

a. (Top). Schematic of the 48-m2 flume, showing heat exchangers, lift box, drain box and airlift mixer.

b. (Bottom) Schematic of mixing resulting from foils inserted in the shallow flumes.

(Source: Laws 1982.)

This report (Weissman and Goebel 1987) originated from a competition held by the ASP for the development of a pilot plant (“test facility”) for microalgae production. As mentioned in Section III. A.4., two companies were selected to develop competing processes: Aquasearch, Inc., of San Diego, California, and Microbial Products, Inc., which had carried out the ASP pond project in California. The objectives were to arrive at cost projections for such a test facility, and also for scale-up costs of a future full-scale facility based on the selected process. Aquasearch, Inc., developed a concept for a closed system microalgae production process, using large plastic bag tubular reactors contained in a greenhouse. This system was not selected for further development, and no final report is available. Aquasearch, Inc. recently established in Hawaii a closed photobioreactor process for cultivating Haematococcus pluvialis, an alga high in astaxanthin.

This study further developed the concept of the HRP system for large-scale microalgae production, providing considerable additional detail, and performing extensive sensitivity analysis of various design options. A site was identified near Brawley, California, in the Imperial Valley, for locating a pilot plant and a full-scale system. Ample groundwater resources were available at this site. There was also significant water supply available from the Salton Sea, which is as saline as ocean water, although of different ionic composition. As discussed in Benemann et al. (1978), the Salton Sea is a potential source of saline water and land for more than 10,000 ha of algal ponds, as such ponds could help manage the salinity of this inland sea, a major problem. The report proposed building two 0.4-ha ponds to validate the process, as well as several smaller ponds for process development and inoculum production. A covered anaerobic lagoon was to be included to test the digestion of the algae and the recycling of nutrients to the algal ponds.

The design of these experimental and pilot plant-scale ponds was provided in great detail. A larger-scale, 400-ha, pond system was also designed and costed. This report presents the most detailed, comprehensive, and realistic cost estimates currently available for large-scale, low-cost microalgae biomass and fuels production.

Several significant advances were made in this design and analysis. Perhaps most importantly, the report presented a fundamental analysis of CO2 supply and in-pond transfer issues, in combination with water chemistry and transit times between carbonation issues. The analysis concluded that CO2 utilization efficiencies can overall be very high, more than 95%, within parameters that would allow high microalgae productivity. Another innovation was the use of small amounts of high molecular weight polymers to improve the flocculation, settling characteristics and harvesting efficiency of the basic bioflocculation process. The polymers can be used in very small amounts, without contributing a major cost to the overall process.

The base case (30 g/m2/d) capital costs were estimated at almost $72,000/ha, without working capital, or almost twice as high as the prior effort (Benemann et al. 1982). This was due to higher costs for many components, such as earthworks, which were several-fold higher. Among other

things, higher costs were assumed for rough and fine (laser) grading, which depends on the type of site assumed to be available. Also the 1987study estimated about $5,000/ha to provide a 3-5 cm crushed rock layer, specified to reduce the suspension of silt from the pond bottom. There is, however, little evidence for a need for such erosion prevention, except perhaps for some areas around the paddle wheel and perhaps the turns. Further, the Weissman and Goebel (1987) study selected slipform poured concrete walls and dividers (baffles) as the design choice. For the curved portions of the walls and berms, the authors specified corrugated walls, with an average cost of about $25/m. This resulted in a cost of over $8,000/ha for the walls (perimeter central, etc.). Clearly, such design options and engineering specifications can result in very large differences in capital costs. For another example, in the present design, a power generation system was specified to produce electricity from the methane generated from the algal residues (at about 10% of total costs). This had not been included in the earlier study. Despite these higher costs, and perhaps because of them, this engineering design and cost analysis effort may be considered the most detailed and realistic one available.

Table III. D.5. summarizes two design cases (from more than a dozen presented in the report), with 30 and 60 g/m2/d average productivities. By using an annual capital charge of 25% (depreciation, return on investment, insurance, taxes), biomass costs of some $273/mt and $185/mt were estimated for the two productivity cases (Table III. D.5.). These costs were more than twice the cost derived from the “conservative” case in the earlier study (Benemann et al.1982), which used only the lower productivity. Note that this even included a significant credit for power production from the methane produced (most of which was used internally). Although this report is the most detailed and complete analysis of microalgae biomass production for fuels available, it can also be criticized for not attempting to examine cost reduction possibilities in the various design options, which would be required to make microalgae fuels production viable. Possible strategies for cost reduction were the objective of the study discussed in the following section.

I Publications:

Weissman, J. C.; Goebel, R. P. (1987) “Design and analysis of pond systems for the purpose of producing fuels.” Report, Solar Energy Research Institute, Golden, Colorado, SERI/STR-231-2840.

Table III. D.5. Capital and Operating Costs for an Open Pond System*

(Source: Weissman and Goebel, 1987.)

Table III. D.5. Capital and Operating Costs for an Open Pond System*

(Source: Weissman and Goebel, 1987.)

Introduction:

During the past 2 decades, manipulation of organisms via genetic engineering has become routine in a number of animal, bacterial, fungal, and plant systems. However, before the research was done at NREL, very little work in this area had been done with microalgae. In fact, the only species for which there was a reproducible transformation system was the single-celled, flagellated green alga C. reinhardtii, which is studied extensively in laboratories as a model photosynthetic cell. The focus of the research in the ASP during the early 1990s was to develop

genetic transformation methods for microalgae with potential for biodiesel production. Based on the collection and screening efforts of the 1980s, this approach was considered to have the highest potential to produce organisms with high constitutive lipid levels, and to use genetic manipulation to understand the molecular regulation of lipid synthesis in the oleaginous algae. Studies on the biochemistry and molecular biology of lipid production in C. cryptica had identified acetyl-CoA carboxylase as a key regulatory enzyme in lipid synthesis (Section II. B.2.e.). One initial goal was to introduce additional copies of this gene into C. cryptica with the hope of increasing the activity of the enzyme and the flux of fixed carbon into lipid.

Several projects will be discussed in the following section of this report that were directed towards the development and use of genetic transformation systems in oleaginous microalgae. The initial approach was to use use available promoters and marker genes that were reported to function in other eukaryotic systems. Various methods were also tried to get DNA into the cell, initially focusing on enzymatically removing the cell wall or perturbating the cell membrane using electroporation. Unsuccessful experiments represented a “Catch 22” scenario, as negative results could mean either the DNA was not getting into the cells, or the DNA entered but could not be expressed at detectable levels. Subsequent experiments were designed to increase the understanding of the processes involved in DNA uptake and expression and to increase the probability of obtaining transformants by developing methods for detecting rare transformation events within a population of cells.

The projects that will be discussed here include a basic study on the DNA composition of microalgal strains, with implications for the choice of reporter or marker genes used to monitor gene expression in transgenic algae. Other aspects of the research that will be discussed include:

• the use of the luciferase gene to monitor DNA uptake and expression in Chlorella protoplasts,

• attempts to develop heterologous and homologous genetic markers for algal transformation,

• the development of methods to introduce DNA into algal cells through the cell wall, and

• the successful development of a stable genetic transformation system for diatoms.

Once the methods were available to obtain genetic transformants, efforts were made to use the transformation system to manipulate lipid content in the algae by overexpressing or downregulating key genes. In addition, the transformation system was used to introduce a reporter gene under the control of various regulatory sequences, to better understand the regulation of gene expression in microalgae.

In the mid-1980s an algal mass culture project for biodiesel production was supported by the ASP in Israel (Arad 1984, 1985, 1986), as a cooperative project among the following groups:

1. Israel Oceanographic and Limnological Research Institute, with Dr. Ben-Amotz, who investigated lipid production at the laboratory and micropond scale;

2. Ben-Gurion University of the Negev, with Professor A. Richmond as Principal Investigator, investigating algal mass cultures with outdoor ponds, essentially of the HRP design.

3. Technion University, with Professor Gedaliah Shelef in charge of developing suitable microalgae harvesting technology.

During the first 2 years of the project Ben-Amotz (1984, 1985) screened laboratory cultures of unicellular algae isolated in Israel and elsewhere. Of a score of strains tested, Nanochloropsis salina and B. braunii were the highest lipid producers, with lipid content as high as 50% in semi­continuous nitrate-limited cultures. Other strains had lipid contents <20%. Lipid composition and chemical characteristics (e. g., hydrocarbon contents) were also determined for many cultures. Nannochloropsis sp., P. tricornutum and C. gracilis were studied in more detail in 0.5- Liter, pH-controlled chemostats for effects of temperature, light intensities, nutrients (Fe and nitrate), salinity and other parameters. The author concluded that “nitrogen limitation does not induce the production and accumulation of lipids,” but the “algae attain a low protein-carbohydrate ratio.” Previous reports in the literature describing lipid accumulation in algae induced by N limitation were attributed to trace element limitations. Actually, the data is typical of chemostat results, in which growth rate imposed by culture dilution do not allow lipid induction as is observed in batch or semi-continous cultures.

During the final year of this project, Ben-Amotz (1986), optimized the growth of two cultures, C. gracilis and Nannochloris atomus in laboratory chemostats and in 0.35-m2 outdoor “microponds.” The ponds were mixed by air sparging, which would reduce pO2 levels. Maximal productivities of 40 g/m2/d were obtained with C. gracilis during June-August, and highest photosynthetic efficiency (9.5% PAR) was achieved in the fall (when productivity was 27.3 g/m2/d, AFDW). During the winter, productivity decreased by about half, but lipid contents in the N-sufficient algal cells increased almost as much, reproducing the low-temperature effect on lipid content seen in the laboratory cultures. Attempts were also made to increase lipid production by Si limitation, but this was unsuccessful due to rapid contamination with green algae.

The work by Professor Richmond and colleagues (1984; Boussiba et al. 1985, 1986), started with laboratory growth and lipid content experiments with more than a dozen algal strains. Outdoor cultivation was carried out for 2 years with small (2.5-m2, 12-cm deep) paddle wheel-mixed high rate ponds. Among other factors, the effects of culture density on productivity and lipid content

were studied, with the expected result that maximal productivity depended on the culture density (actually, on the areal concentration), but that this does not have major effects on lipid content. At the optimal density of 350 mg/L (45 g/m2), productivity in the summer was 24.5 g/m2/d and lipid content about 16% for N. salina, and somewhat higher (28.1 g/m2/d and 22 %) for Isochrysis galbana (both SERI Culture Collection strains). However, experiments with varying pond depths but constant areal biomass densities resulted in productivity differences of up to twofold, contrary to theory and expectations. Other factors (pO2, etc.) likely accounted for this. However, mixing speed had no significant effect on productivity. The authors stated: “These data reflect the complexity of the process of optimizing outdoor biomass production….”

Professor Shelef (1984a, b; Shelef et al. 1985) carried out experimental and engineering studies of algal harvesting. The major effort was on the use of chemical flocculants for affecting algal sedimentation. Much of the work focused on I. galbana, grown, as above, on seawater of various concentrations. As expected, the higher the ionic strength (salinity), the greater amounts of chemical flocculants (alum, ferric chloride, chitosan) were required to induce algal flocculation. Autoflocculation, achieved by interrupting the CO2 supply, was also very effective. Other processes investigated were sand bed filtration, microstrainers (a 21 pm polyester weave allowed some 80%-90 % harvest efficiency), dissolved air flotation (after chemical flocculation, the method of choice for most commercial installations), and again, chemical “enforced” flocculation (recycling some of the precipitate to reduce flocculant needs). An economic analysis suggested various “allowable” flocculant costs for assumed biomass values. Overall, however, chemical flocculants are too expensive for biodiesel production.

During the final year of this project, a 100-m2 pond was operated with I. galbana for 1 month (Arad 1986). In batch culture it took 12 days for the culture to enter stationary phase, and a productivity of 23.6 g/m2/d was measured for about 2 weeks after starting dilutions. The culture was harvested with FeCl3 and alum using a dissolved air flotation unit from Technion. The flocculated algae had rather low lipid contents, compared to centrifuged algae. In conclusion, the Israeli project provided another dimension to the ASP effort, generally supporting the conclusions and results obtained by the U. S. work.

There are several arguments for and against a U. S. microalgae biodiesel R&D program. One of the more important, and perhaps contentious, issues, is the potential impact of such technologies on U. S. energy supplies, specifically liquid transportation fuels. The review in Section III. C. of the NREL resource analyses for microalgae biodiesel concluded that there is a potential for production of several quads (1015 Btu) of biodiesel fuels in the southwestern United States alone. However, as stated earlier, it will be difficult to find many locations where all the resources required for microalgae cultivation, flatland, brackish or waste waters, and low-cost CO2 supplies, are all available in juxtaposition. And, as also pointed out, the southwestern United States is not the ideal climatic location for such systems. For both these reasons, the resource potential estimated by these resource studies must be significantly discounted.

In the case of utilization of power plant CO2, diurnal and seasonal factors would restrict direct CO2 (e. g., flue gas) utilization to about one-third of the power plant CO2. Even with CO2 capture and transportation (which greatly increases costs), only about half of the CO2 would be useable. With most coal-fired power plants located in the north, or in otherwise unfavorable climates, only a rather small fraction of power plant CO2 resources would likely be captured with microalgae systems in the United States.

A conservative estimate is that microalgae systems would be able to mitigate, directly or indirectly, perhaps only about 1% of current power plant CO2 emissions, supplying an approximately equivalent amount of current transportation fuels. Herzog (1995) argued that such a potential, in fact, anything less than 10%, is not sufficient to justify a R&D effort, and that scarce resources should be devoted only to potentially high-impact technologies, such as the disposal of CO2 in the oceans or geological formations. However, Benemann (1995) countered that such a resource-only argument is too limited, as it ignores the issue of economics and technological risks. For example, the technical feasibility of ocean disposal is far from established, and the costs of such a process are not currently constrained by credible engineering and economic analyses. A balanced R&D portfolio would need to account for such factors. Also, it is inherently more attractive to use and recycle CO2, thus increasing economic activity, rather than to bury it or dump it into oceans. In addition, microalgae CO2 utilization could spin-off other technologies, as in the case of the ASP. Thus, although a decisive role for microalgae fuel production and greenhouse gas mitigation cannot be extrapolated, a modest R&D effort in this area is appropriate in the context of developing many such alternative technologies.

IV. B. I.e. Summary of Major Conclusions from the ASP Microalgal Mass Culture Work

This report cannot do justice to the extensive and long-term R&D effort in applied microalgae mass culture carried out by DOE and the ASP over a 20-year period. Here only a very brief summary of the major conclusions is provided to put into context the recommendations for future R&D, which follow.

Two major conclusions can be derived from the outdoor cultivation projects and engineering/economic analyses under the ASP, and can be briefly summarized:

1. There appear to be no fundamental engineering and economic issues that would limit the technical feasibility of microalgae culture, either in terms of net energy inputs, nutrient (e. g., CO2) utilization, water requirements, harvesting technologies, or general system designs.

2. Productivities, in terms of total biomass and algal lipids (oils) currently achieved during the ASP are substantially higher than those reported and even projected before the ASP, but still well below the theoretical potential, and the requirements for economical viability.

The first conclusion should not imply that all these issues and problems have been solved. It does, however, suggest that the immediate R&D needs are not for engineering designs or cost analysis, or even in the operation of large, outdoor algal mass culture systems. Rather, from the second conclusion, the emphasis of any R&D effort must be on more fundamental and early-stage applied research issues faced in developing very high productivity algal strains. Ideal strains would dominate the pond cultures, achieve near-maximal productivities, efficiently biosynthesize large amounts of lipids, and be easy to harvest.

Another conclusion from the DOE-ASP program is that the only plausible near — to mid-term application of microalgae biofuels production is integrated with wastewater treatment. In such cases the economic and resource constraints are relaxed, allowing for such processes to be considered with well below maximal productivities.