Because small particles suspended in water have a higher interfacial energy than large particles, small particles have a tendency to form larger aggregates or flocs. Spontaneous flocculation of particles, however, is often prevented by electrostatic repulsion. This electrostatic repulsion occurs when particles carry a surface charge. Microalgae usually have a negative surface charge, the result of carboxylic, phos — phoryl, and amine/hydroxyl groups on the microalgal cell surface. Most important are the carboxylic groups, which are deprotonated at pH above 3-4, and thus have a negative charge. Phosphoryl groups become anionic above pH 7-8. Amine and hydroxyl groups lose a proton below pH * 10 and become, respectively, uncharged and negatively charged. The overall effect is net negative microalgal surface charge in most natural waters (Hadjoudja et al. 2010). Few studies, however, have inves­tigated the contribution of active groups on the microalgal cell surface to the surface charge, and more fundamental research is needed to better understand the linkage between microalgal surface charge and flocculation (Brady et al. 2014).

The electrical double layer consists of the negatively charged surface of the cell and the positively charged cloud of counterions close to the cell surface. The zeta potential, or Z potential, is the potential difference between the bulk solution and the slip plane in the electrical double layer (Fig. 12.2). The counterions that are between the surface and the slip plane remain associated with the cell when the cell is moving through the solution. The sign of the Z potential and the surface charge are the same. The Z potential is relatively easily quantified by measuring the mobility of the charged particles in an electric field (Ozkan and Berberoglu 2013) and is a useful measure of the stability of microalgal suspensions. A Z potential of -25 mV or less generally indicates a stable suspension, while a Z potential between -10 mV and 0 mV points to a low stability and will generally coincide with flocculation.

The negative surface charge of microalgal cells is largely controlled by depro­tonation of carboxylic functional groups. When pH is decreased in a microalgal suspension to levels below 3-4 where carboxylic groups are uncharged, the Z potential approaches 0 mV (Hadjoudja et al. 2010). Indeed, flocculation can be induced by reducing the pH to below 4 (e. g., Liu et al. 2013). Chemical modifi­cation of functional groups on the cell surface through oxidation by ozone, chlorine, potassium permanganate, or potassium ferrate addition may also facilitate floccu­lation (Sukenik and Shelef 1984; Henderson et al. 2008b).

The ionic strength of the medium has an important influence on the Z potential. At high ionic strength (e. g., seawater), the electrical double-layer surface is com­pressed and the Z potential becomes less negative, allowing particles to approach it other and to flocculate. For example, a suspension of clay particles will flocculate when the ionic strength of the medium is increased. High concentrations of divalent cations (e. g., Ca2+ or Mg2+) can also compress the double layer and, if present in sufficient concentrations, reverse the otherwise negative charge of surfaces. Despite the strong influence of ionic strength on the electrical double layer, there are no indications that suspensions of microalgae that live in seawater are less stable than

those of freshwater microalgae. This suggests that factors other than electrostatic repulsion also contribute to the stability of microalgal suspensions.

One factor that might play a role is steric stabilization by polymers that are associated with the microalgal cell surface (Fig. 12.3). Large polymers such as polysaccharides that are attached to the microalgal cell surface can extend into the

surrounding medium. These polymers may prevent cells from approaching each other and can therefore stabilize microalgal suspensions. More research is needed to better understand what factors contribute to the observed stability of microalgal suspensions, particularly in seawater.

A hydrodynamic fluidic separation system based on fluidic shear as the culture stream passes through a spiral pattern in the fluidic device has been developed at the Palo Alto Research Center. This system is a spiral channel through which the culture is passed; drag from the channel walls exerts forces that separate particles from the suspending solution such that separated, concentrated, and dilute streams come out at the end of the separator. The device can be run continuously, requires no filters, is generally large enough to prevent fouling, and requires no additional materials for the separation. They have demonstrated a benefit to the use of flocculants with small algae in this system. Such a system would also have minimal or no fouling due to the large size of the fluid channels. This device has been used with Scenedesmus dimorphus with chitosan as a flocculant and provided a 42-fold increase in concentration. With Arthrospira platensis and no flocculant, the device provided 5.8-fold concentration on one pass and 18.8-fold on two passes through the concentrator providing a 97.3 % harvesting efficiency (Hsieh et al. 2012; Volkel et al. 2011). There are currently no reports of use of this technology at large scale with algae.

Nitrogen source is very important in mixotrophic and heterotrophic cultures of mic­roalgae. Adequate concentration of nitrogen is required for cell growth, while nitrogen limitation is often used to enhance lipid accumulation. Inorganic nitrogen, organic nitrogen, and various waste products have been investigated for biodiesel oil pro­duction (Becker 1994). The use of ammonium as a nitrogen source for Ellipsoidion sp. resulted in higher growth rate and lipid content than when urea and nitrate were used (Xu et al. 2001). On the other hand, Neochloris oleoabundans grew faster and accumulated higher lipid with nitrate than with urea (Li et al. 2008a), but the cell grew poorly in medium with ammonium as the nitrogen source. Complex nitrogen sources are expected to be more effective than simple nitrogen sources in the heterotrophic culture of microalgae, since most of them contain amino acids, vitamins, and growth factors. However, the effectiveness of the nitrogen source depends on the species. For example, nitrate was the best, followed by urea, for the growth of Chlorella vulgaris, while peptone and beef extract did not improve cell growth. Furthermore, ammonium sulfate and ammonium nitrate were less effective than nitrate and urea (Kong et al.

2011) . The type of nitrogen source affects not only the cell growth, but also lipid accumulation. The lipid content of Chlorella vulgaris in mixotrophic culture was highest for peptone, followed by beef extract, but the lipid productivities were low because of low biomass concentration. Ammonium sulfate and ammonium nitrate gave the least lipid content. Potassium nitrate and urea gave intermediate lipid content, yet had the highest productivity as a result of the high biomass content (Kong et al. 2011).

Among the organic nitrogen sources, urea is a promising nitrogen source for large — scale production because it is relatively cheap (Becker 1994; Danesi et al. 2002; Matsudo et al. 2009). With urea as the nitrogen source, the lipid contents of Chlorella sp. decreased with the increase in urea concentration (Hsieh and Wu 2009). The optimal concentrations differ with the nitrogen source. The optimal sodium nitrate and yeast extract concentrations for growth and lipid production by Tetraselmis sp. in mixotrophic culture were 4.70 and 0.93 g/L, respectively (Iyovo et al. 2010). Industrial wastewater rich in nitrogen, such as monosodium glutamate waste, has been reported to be a good and cheap source of nitrogen for cultivation of Rhodotorula glutinis for the production of biodiesel (Xue et al. 2006; Becker 1994). The type of nitrogen source affects the pH of the culture broth. The pH was stable when urea or potassium nitrate were used, but dropped sharply when other nitrogen sources were used.

There are a number of technical bottlenecks that need to be addressed. Some of the

basic questions yet unanswered include the following:

1. In a full-scale field operation, what are the ideal strains of algae that will yield both a high-quality effluent and a high-quality biofuel?

2. What are the most efficient methods for separating and concentrating the algae? The transition from 300 mg/L in the pond to a slurry that is 20 % solids may take two or three steps.

3. What are the most efficient and cost-effective methods for breaking open algae cells for the production of a green crude or the separation of the lipid, aqueous, and solid fractions of the lysed cells?

4. What are the most efficient and cost-effective methods for converting the green crude or purified lipid into a commercially reliable biofuel? In the past few years, several new methods have been developed on a bench and demonstration scale, but no one process has been made the leap to be an industry standard for full-scale systems.

5. The companies that favor large photobioreactor systems have yet to show how their systems could be implemented on a grand scale. There is roughly an order of magnitude difference in capital and labor costs between photobioreactors and open pond systems. For algae-based biofuel system to be commercially viable, the output would need to be on the order of thousands to millions of liters of biofuel per day. While practitioners from the fields of biotechnology and bio­chemical engineering have very reliable data from their bench and demonstra­tion-scale bioreactors, the only two fields of engineering that have a long­standing history of working on systems that reliably process liquids on a grand scale are in the chemical/petroleum industry and wastewater treatment.

6. The photobioreactors can provide the initial step of a high-quality starter culture into an open pond system, but when one takes into account the energy needs and capital cost of mega-scale bioreactors, the economic feasibility of hundreds of hectares of photobioreactors fades rapidly.

7. A number of companies are trying to outdo their competition based on the hopes of genetically altered strains of algae. Considering the fact that it takes the approval of several different local and national regulatory agencies just to restore a disturbed habitat with native plant species, the likelihood of a company being allowed to generate 50 tons/day of genetically altered algae in open ponds could face some very tough opposition that would include regulatory agencies and well-organized citizen groups. In addition, the fate of genetically introduced microbes into the environment can be precarious; for example, consider the failure of engineered Rhizobium strains introduced into soybean fields to inoc­ulate seedlings but were outcompeted by native strains (Kent and Triplett 2002)

8. What will make or break this industry is the ability to produce biofuels on a mega-scale basis with a high degree of reliability. A serious economic analysis is needed for each step (i. e, culturing, harvesting, dewatering, lysing, and bio­fuel processing) in the development of algae-based fuels.

9. Most large wastewater treatment plants that process 200-1000 ML/day are located adjacent to dense urban landscapes with little available room for large — scale algae ponds. The ideal candidates for the system proposed in this chapter would be rural wastewater treatment plants that process 2-40 ML/day. Most often, these plants are located at a good distance from populated areas, and there is ample land that could be developed into algae-culturing ponds. In parts of the world where there is a cold or monsoon season, the plant can revert back to its original treatment process and use the ponds for short-term storage of wastewater.

6.2Conclusion

The ability to culture and harvest algae has improved dramatically over the past five decades. There are numerous treatment options that can be used to make the transition from concept to demonstration to full-scale implementation of algae biofuel programs. This will require the ability to adapt preexisting technologies from several disciplines. Many of the answers are already out there but have yet to put in the proper sequence or combination. There is no one technical solution to make this process commercially viable. As demonstrated in this chapter, it will require contributions from several disciplines to go beyond their technical comfort zones. While this is an emerging field with great promise, it will be built on the fundamental principles of engineering and science.

Acknowledgment H. Ahmadzadeh thanks ATF Committee for the financial support.

Many methods can be used to validate the model’s proposed predictions; however, comparison of in silico predictions with in vivo experiments is a key method for validation. Measuring experimental growth phenotypes at specific conditions can validate the predicted growth under the same conditions. An alternative validation approach is to carry out in silico and in vivo gene deletion experiments (or to compare in silico results with available gene deletion literature) to check whether there is an agreement between model predictions and the actual deletion mutant phenotype. Moreover, omics data from transcriptomics, metabolomics, and pro- teomics experiments can be used to check the consistency of the model’s predicted results. Available simulation debugging tools may be used if the model has poor agreement with experimental data. Further refinement of the model can be done as described in the next section.

When using flocculation in wastewater treatment or fermentation, the cost of the flocculant is related to the volume of treated water or medium. But when floccu­lation is used for harvesting microalgae, the cost is calculated relative to the amount of biomass that is harvested, not relative to the volume of microalgal broth that is treated. A given dose of flocculant capable of removing 0.5 kg of particulates from an aqueous medium containing 0.5 g L 1 of particulates will yield 1000 L of treated water in the case of wastewater treatment, but only 0.5 kg of biomass in the case of microalgae harvesting. As a result, the cost of the flocculant is a critical factor when selecting a flocculation method for microalgae harvesting.

When evaluating the cost of flocculation-based harvesting, the cost of the chemical flocculant in relation to the dosage needed is obviously important. Some flocculants have been used since many years in various industries and are low-cost commodities that are available in bulk. An example is alum, which is commonly used in wastewater treatment. Although alum is cheap, the high dosage that is required to flocculate microalgae makes its use relatively expensive (about 300 $ ton-1 harvested biomass) (Molina Grima et al. 2003; Uduman et al. 2010). Chitosan requires lower dosages, but it is a more expensive chemical and therefore has a similar overall cost (about 500 $ ton-1 harvested biomass). Autoflocculation by magnesium hydroxide is often proposed as low-cost method for flocculation. As most waters contain sufficient magnesium, the main cost for flocculation is the cost of the base required to increase the pH. This cost is very low when calcium hydroxide or slaked lime is used as a base (about 30 $ ton 1 harvested biomass) (Schlesinger et al. 2012). When evaluating the cost of ECF, not only the electricity cost for electrolysis should be taken into account, but also the dissolution of the sacrificial anode (Lee et al. 2013a). Aluminum anodes generally have a higher flocculation efficiency than iron anodes, yet iron electrodes may nevertheless be preferred due to a lower anode cost and lower energy con­sumption (Dassey and Theegala 2014). Some novel flocculants such as for modified iron oxide nanoparticles are not yet available on the market, which makes it difficult to estimate their cost. No chemicals are required to harvest microalgae using bio­flocculation. When flocculation of microalgae is induced by addition of cultured bacteria, fungi or other microalgae, the cost to cultivate these flocculating microor­ganisms should be taken into account. It is important to realize that the dosage of flocculant needed and thus the cost of the flocculant depend strongly on the species of microalgae that is harvested and on the culture conditions. The amount of algal organic matter excreted in the culture medium can result in a strong increase in the flocculant dosage.

When estimating the cost of a flocculation method, not only the cost of the flocculant should be taken into account, but also the energy demand for mixing and pumping (Elmaleh and Jabbouri 1991; Danquah et al. 2009a). Intensive mixing of the flocculant solution with the microalgal culture broth is essential to achieve good flocculation efficiency (e. g., Lee et al. 2013a, b). Other indirect costs are costs related to the sedimentation or flotation system (construction costs, operational cost, land cost) (Richardson and Johnson 2014). In life cycle analysis or LCA studies, the environmental burden associated with the production of the flocculant should be taken into account as this cost can differ substantially between different flocculants (Brentner et al. 2011). Examples of flocculants with a potentially high environ­mental burden might include those that are Al-based.

The primary aim of using flocculation for harvesting microalgae is to pre­concentrate the microalgal biomass and to reduce the volume of broth that needs to be processed by a mechanical dewatering method such as centrifugation or filtra­tion. The higher the degree of pre-concentration that can be achieved during the pre­concentration stage, the higher the energy savings in the mechanical dewatering stage (Milledge and Heaven 2012). A flocculation method that can produce a small volume of algal sludge will require less energy for mechanical dewatering than a method that produces a large sludge volume. Pre-concentration of the biomass can be achieved by combining flocculation with sedimentation or with flotation. Flo­tation has higher investment costs and requires more energy than sedimentation, but it can achieve a higher degree of biomass concentration (Besson and Guiraud

2013) .

The use of flocculation for harvesting may interfere with other stages of mic­roalgae cultivation and processing. For instance, the flocculant may prevent recy­cling of the culture medium, which can result in high costs for water to prepare fresh culture medium or a high cost for treatment of the spent medium before discharge into the environment. The flocculant can interfere with downstream processing of the biomass, e. g. lipid extraction. Some flocculants form toxic resi­dues in the harvested biomass and can limit the use of the protein fraction as animal feed, and thus limit the income that can be generated from the microalgal biomass. These indirect costs of using a certain flocculation method should also be taken into account. More information on the economics of harvesting and downstream pro­cessing is given in Chap. 14.

12.2 Conclusions

It is clear that the energy requirements for harvesting microalgae could be reduced by at least an order of magnitude if the biomass could be pre-concentrated using floc­culation. There are several technologies available to flocculate microalgae, including metal salt coagulants, electro-coagulation-flocculation, polymer flocculants, or the use of clays or iron oxide nanoparticles. Flocculation can also be induced by a pH increase (autoflocculation; result of precipitation of Ca or Mg salts) or it can occur spontaneously (bioflocculation). The main disadvantage of using flocculation for harvesting microalgae is that, in most cases, the biomass is contaminated with a foreign substance. This can limit the use of the biomass or can interfere with downstream processing. Therefore, the flocculant should ideally be non-toxic or, better still, it should be possible to remove the flocculant from the biomass after harvesting. Because cost reduction is an important issue in production of microalgal biofuels, the cost of a flocculation technology is an important criterion. Microalgae excrete organic matter in the culture medium that may interfere strongly with floc­culation. This interference of microalgal organic threatens the applicability of floc­culation for harvesting microalgae. So far, no ideal universal flocculation method has yet emerged. It is most likely that the optimal flocculation method will be species specific and depends on the final application of the biomass.

Acknowledgements D. Vandamme is a Postdoctoral Researcher funded by the Research Foundation—Flanders (FWO).

While both the crystalline silicon solar cells and amorphous silicon solar cell examples shown in Fig. 15.3 are optimized to have a broad spectral response, it is known that photovoltaic devices would work with a higher efficiency if they only had to absorb monochromatic light (Sark et al. 2008) or light from a very narrow spectral range. One way to provide a limited portion of the spectrum to a solar cell, to make the most of its peak efficiencies, is to use a luminescent solar concentrator (LSC). A LSC is a flat-plate solar concentrator made from a thin transparent polymer (such as acrylic) containing a luminescent material (Sark et al. 2008). They work by accepting light from the AM1.5 solar spectrum and directing a portion of the light toward the edges of the flat polymer sheet. Solar cells are located at the edge of this flat sheet for converting the light into electricity. Photons incident upon the polymer sheet with enough energy will excite the luminescent materials. These materials will then re-emit a photon with a longer wavelength. As the photons are emitted in random directions, a portion of these will be captured by total internal reflection in the flat sheet and transmitted to the edge where it can be collected by a solar cell (Sark et al. 2008). With the photons being re-emitted in random direc­tions, there will be some that are transmitted out of the concentrator, and however, a portion is captured and directed toward the edge for collection. Photons without insufficient energy to excite the luminescent material will be transmitted through the concentrator with very low loss. The beauty of these flat-panel concentrators is that increasing the size of the polymer sheet will directly increase the number of photons captured via total internal reflection and able to be converted to electricity. That is, the concentration factor increases with the area of the concentrator. This increase in output from the concentrator can be achieved without increasing the size of the solar cell itself.

Although the efficiency of LSC has historically been fairly low (Sark et al.

2008) , there have been devices reported with efficiencies of up to 7.1 % (Sark et al. 2008) which bodes well for the technology. This style of concentrator relies on the use of only a small portion of the solar spectrum. For example, if a luminescent material that emitted green photons was used, all incident light with longer wavelengths would be transmitted through the concentrator. The re-emitted green photons would be directed to a solar cell with a high efficiency in this part of the spectrum. This would potentially be a good match for a system that would convert solar energy into electricity and grow microalgae.

Hydrocarbons are able to be ‘milked’ (without cell death) from Botryococcus braunii using a solvent added to the growth medium (Moheimani et al. 2013a). Advances in this research further demonstrated that hydrocarbons could be repeatedly extracted (milked) from B. braunii using the solvent every 5 days for a total of 70 days with no addition of fertilisers (N and P) to the culture (Moheimani et al. 2013b). In this experiment, the cells were not dividing and therefore, nutrients were not required for the production of proteins and other cell elements. Instead, the majority of the light energy was used to convert CO2 to hydrocarbons to replace those previously milked.

Microalgae play an important role in many wastewater treatment facilities around the world. In developing economies in tropical and subtropical countries, waste­water is often treated using facultative ponds or oxidation ponds (Duncan 2004; Rahman et al. 2012). These consist of relatively deep and non-mixed ponds that are spontaneously colonized by microalgae. In these ponds, microalgae serve mainly to supply oxygen for the aerobic oxidation of organic matter present in the wastewater. Because these ponds are relatively deep and are poorly mixed, microalgal pro­ductivity is relatively low; only about 10 ton of dry biomass ha-1 year-1. The microalgal biomass is not harvested at the end of the wastewater treatment process, and either settles to the bottom of the pond or is washed out of the ponds. Because the microalgal biomass is not harvested, removal of nutrients from the wastewater by the microalgae is inefficient.

At the end of the 1950s, high-rate algal ponds (HRAPs) were proposed as an alternative to facultative ponds (Oswald and Golueke 1960). HRAPs are raceway — type ponds in which the water is mixed by a paddle wheel. Compared to facultative ponds, HRAPs are much shallower and better mixed and, as a result, have higher microalgal productivity, about 30 ton dry biomass ha-1 year-1. The productivity of HRAP’s can be further increased with CO2 addition (Craggs et al. 2013). Because microalgal productivity is higher, a larger volume of wastewater can be treated on the same land area when compared to facultative ponds. Akin to facultative ponds, microalgae in HRAPs supply oxygen for the aerobic oxidation of organic matter, and if the microalgal biomass is harvested at the end of the treatment process, the microalgae remove nutrients from the wastewater. Harvesting of the microalgal biomass, however, is costly and many HRAPs today do not harvest the biomass. Although HRAPs are used in wastewater treatment plants around the world, the technology is much less widespread than oxidation ponds or conventional elec­tromechanical wastewater treatment systems (Craggs et al. 2013).

Computational tools that help to improve synthetic biology have been developed and are currently being expanded. Improvement of algae production to increase biofuel yields through synthetic biology involves many distinct processes that can be aided by different computational tools. Genome-scale metabolic network reconstructions and models are available for a number of algal species. These can be useful for identifying and selecting gene targets for knockout and strain engi­neering. Some of the available tools and algorithms that are able to perform such tasks include (but are not restricted to) Optknock (Burgard et al. 2003) and Optstrain (Pharkya et al. 2004).

The standard approach for computing metabolic fluxes is flux balance analysis (FBA) through using toolboxes such as COBRA or Pathway Tools (these are discussed in the accompanying Chap. 10, Towards applications of Genomics and Metabolic Modeling to Improve Biomass Productivity). FBA allows prediction of optimal flux distribution throughout the network for a given cellular phenotypic state. Through the use of computational tools associated with FBA, consequences of changes introduced in a metabolic network can be predicted. For example, quantitative mapping of intracellular fluxes in relation to single or multiple gene deletions can easily be carried out by FBA. There are many strategies available to predict alterations that result in increased production of a desired metabolite. For instance, one method entails identifying key and relevant pathways that are impacted using simulated gene knockouts (Reed et al. 2010). The accuracy of this method can be enhanced by integrating experimental data; such as metabolite concentration, gene expression data, and uptake and secretion rates.

‘Pathway Tools’ (Karp et al. 2010) is an integrated reconstruction, analysis, and visualization software created by the Bioinformatics Research Group at SRI Inter­national (http://bioinformatics. ai. sri. com/ptools/). Pathway tools can automatically

generate organism-specific metabolic network databases and provides details of genes/proteins, reactions, and compound associations, as well as create pathway databases (called Pathway Genome Database, or PGDBs). To create a PGDB, one of the components of the Pathway Tools called PathoLogic is used. This tool allows users to create PGDBs using the genome annotation of an organism of interest directly from the organism’s GenBank annotation file. Users can manually adjust, edit, or add (new) content as needed. There are already many well developed and intensively curated pathway databases or PGDBs including BioCyc, EcoCyc and MetaCyc (Caspi et al. 2014), which aid metabolic analysis and network recon­structions. BioCyc alone has a collection of about 3530 PGDBs, which users can query, visualize, manage and analyze. Among these, algal PGDBs include Tha — lassiosira pseudonana, Nannochloropsis gaditana, Acaryochloris marina, Ana — baena cylindrica, Anabaena variabilis, Synechococcus elongatus and

Chlamydomonas reinhardtii. Other offered functionalities in Pathway Tools include tools that can be used in downstream analyses to identify the shortest path between two metabolites, identify dead-end metabolites, fill pathway gaps, identify choke — points (potential drug targets), and infer operons and transport reactions. Many new metabolic reactions have been added to EcoCyc using the dead-end metabolite analysis approach (Mackie et al. 2013). Pathway Tools can aid synthetic biology experiment designs by identifying potential pathways, which may be targets for alterations.