The responsible production of genetically modified (GM) microalgae and its appropriate regulation in many ways parallels the previous emergence of GM crops utilised in terrestrial crop-based systems. GM crops have been in field testing for approximately three decades now and with their global scale now approaching almost 200 million hectares, their benefits have been demonstrated, although they have been beset by much controversy, and there are also some cautionary lessons learned. There are some important distinctions between the two forms of production (i. e. aquatic versus terrestrial), and microalgae systems are generally capable of much greater containment than conventional cropping systems. In order to preface this discussion, it is important to first examine the current issues with wild (non — GM) algae, both in the environment and in commercial production systems, and the current state of regulatory oversight.

Wild Algae in Aquatic Ecosystems ‘Toxic algae blooms’ are a regular headline in the mainstream media resulting in a public perception that algae are a menace. In water treatment industries, this fear of algal toxins is also relatively well established. In reality, the number of algae that produce any toxins is a tiny fraction of the existing biodiversity. Almost all of the known toxins attributed to algae are actually found in certain types of cyanobacteria and dinoflagellates, with a much smaller representation from some bacillariophytes (diatoms), haptophytes, pelagophytes and euglenoids. In some cases, there are groups who are cultivating specific species, e. g. dinoflagellates, to utilise toxic compounds for applications such as biomedical cytotoxins (generally under laboratory conditions), but this is the exception rather than the rule, and the overwhelming majority of the industry is focused upon avoiding toxic species. For example, the cultivation of the cyanobacteria Arthro — spira (Spirulina) for human food consumption must be free from the cyanobacteria Microcystis.

Environmental algal blooms, while an ongoing concern, are usually the result of anthropogenic nutrient outflows or natural processes of nutrient cycling. They are not generally the result of well-managed microalgae farming practices. Such blooms can occur during periods of elevated nutrient levels due to either natural processes (e. g. weather effected nutrient run-off from land or oceanic currents and upwellings) or from anthropogenic nutrients (e. g. municipal, agricultural or industrial waste waters), with the latter being more closely correlated with the increase in the frequency and the intensity of environmental algal blooms.

Algal blooms can be broadly divided into classes as (1) blooms that are transient and innocuous (2) both transient and persistent blooms that are generally considered to be harmful, and (3) blooms that are clearly detrimental and disruptive to eco­systems. As the algae themselves are by and large ubiquitously present, the primary underlying issue is the management of nutrients and eutrophication processes. While innocuous algae blooms are generally rapidly consumed by organisms higher up the food web (e. g. plankton and filter feeders) and are therefore transient, harmful algal blooms (HABs) (Anderson 2009; Anderson et al. 2002; Van Dolah et al. 2001) and ecosystem disruptive algal blooms (EDABs) (Sunda et al. 2006) can be comprised of algal species that are generally unpalatable to aquatic herbi­vores or that contain toxins. This is important because it disrupts the food web and the concordant transition of nutrition and chemical energy to higher trophic levels which can result in a loss of ecosystem biodiversity (with ecosystem biodiversity being closely correlated with ecosystem resilience). Historically, these problems are largely caused by agricultural nutrient outflows, and there has been significant analysis of how outflows of nutrients and chemicals from agricultural production can vary greatly in their ‘pollution footprint’, e. g. (Hill et al. 2006). The potential for reducing the pollution footprint is one of the strong benefits of microalgal production systems (Smith et al. 2010) in that they generally have no chemical outflows, and due to greater containment relative to fields of crops in soil, they can have much lower nutrient outflow, and in some cases a negative footprint where they utilise anthropogenic nutrients from other systems, e. g. wastewater integration and bioremediation systems. Nevertheless, forward thinking risk management strategies are needed to ensure that microalgae production systems at very large scale do not induce similar concerns to those experienced in traditional agriculture.

Proper Management of Microalgal Production Systems Proper management of microalgal systems is an important aspect of any commercial operation. This will be increasingly important as systems are scaled for large-scale production and the varieties of engineered strains used increase. The establishment of production models aiming to exploit the benefits of GM microalgae contributes additional complexity to prudent regulatory frameworks. There is a duality to the responsible management of GM microalgae production systems in that (1) from a product perspective, farmed microalgae cultivations must be maintained at adequate purity and free from contaminants that can compromise product quality (e. g. in the Ar — throspira example given above), and (2) from an environmental perspective, the release of nutrients or microalgal biomass must be properly managed in order to mitigate any risk to local ecosystems. Given that for the production of biocrude, biomass will be subjected to thermochemical processing, it is the latter point which is central to this discussion.

Both the type of release (nutrients or biomass) and the scale of release are important parameters in a proper risk assessment. Gressel et al. (2013) have added to the discussion on mitigating spills and propose that spills from large-scale cul­tivations will be inevitable—however, there is an important consideration here regarding the terminology moving forward in this discussion. We expect that implementation of proper standards in prudent farm management should be able to mitigate the chances of large-scale spills into the environment; however, it is widely agreed that microalgae have a relatively high capacity for dispersion (e. g. micro­scopic size, and potential to form aerosols). Thus, if some aerosolised cells escape to the environment, it is certainly a release, but is this considered a spill? In terms of nutrients, the scale and/or persistence of release is generally the most important variable in terms of subsequent eutrophication potential and the corresponding risk assessment, but in terms of biomass, a single cell escaping as an aerosol particle has the potential to establish itself outside of the farm boundary even if there is no ‘spill’. Thus, in this respect, species release is indeed inevitable, and it is in this context that any discussion of GM strains must be conducted. Hence, if small-scale release cannot ultimately be avoided, then the discussion is inexorably dependent upon the biological character of what is released.

GM Microalgae and Their Regulation Considering the inevitability of release, risk assessments of GM microalgae must be conducted on a case-by-case basis, with specific attention to the nature of the modification and whether it actually conveys a competitive advantage of some kind to the strain in question when it is relocated within a natural ecosystem or whether the modification can result in disruption to ecosystems in some other way. Henley et al. (2013) have recently published an excellent examination of GM algae risk assessment which should serve well as a foundation study for this evolving discussion. They rightly stipulate that for a GM-specific environmental risk assessment, primary considerations of potential ecological impact include the following:

1. The potential of GM microalgae to be more highly competitive in natural ecosystems.

2. The potential of GM microalgae to result in altered communities of aquatic herbivores in terms of composition, dominance or biodiversity.

3. The potential of GM microalgae to be involved in horizontal gene transfer (HGT) to other micro-organisms.

Given that it is anticipated that, for the most part, new algae producers will not be cultivating microalgal species that are invasive or toxic—the primary consid­erations will be the genetic modifications themselves rather than the host strains (indeed popular host strains such as C. reinhardtii are quite easily outcompeted by many wild-type species); however, it has previously been seen that some potential production candidates have already been involved in bloom events that have resulted in their classification as EDABs (Sunda et al. 2006). Thus, we encourage a careful and iterative investigation of all aspects of microalgae production, but emphasise that in this discussion it is the specific genetic modifications relevant to high-density microalgae cultivations that is in need of far greater discussion in the literature. Consequently, we discuss here the implications of the engineering applications highlighted in section two, with respect to associated risk of species establishment, dominance and ecosystem disruption. Much can be gleaned from the parallels with GM crop species, especially pollen dispersal; however, there are distinct differences between terrestrial crops and communities of aquatic micro­organisms.

For microalgal strains engineered to have varied light-harvesting and photo­synthetic efficiency, the general desire is to increase net biomass productivity. As discussed above, this can be achieved through different methods. The down-regu­lation of LHC proteins or pigmentation can provide an overall net benefit to high- density cultures in high-light conditions (i. e. the artificial farming environment), e. g. (Oey et al. 2013); however, this generally makes individual cells less com­petitive in natural ecosystems where competing wild-type cells retain the capacity to modulate their antennae size and pigmentation levels. In theory though, genetic modifications that unilaterally increase total productivity (e. g. a higher efficiency rubisco enzyme or strains that can utilise a wider range of the spectrum) could potentially convey a competitive advantage irrespective of the growth environment.

Where strains are photosynthetically superior irrespective of environment, they could potentially affect ecosystem dominance and diversity, and while microalgae composition might not be significantly changed, the increased availability of these microalgae could result in additional effects like changes in plankton composition. In contrast, LHC/pigment-reduced cells could lead to some immediate composi­tional changes when consumed, but this would be insubstantial at the community level, and as these strains are outcompeted by wild-type organisms, there would be no net change to dominance or biodiversity. Again, the real concern for HGT would be that microalgae with superior generic photosynthetic efficiency would be capable of transferring this trait to other phototrophic organisms enabling them to also have greater competitiveness in the natural ecosystem. The transfer of disabled antennae/ pigment modulation would not convey an advantage to other species.

GM strains that have a greater capacity for nutrient scavenging may have an increased competitiveness if released, but there is already a diverse range of strategies for nutrient uptake and usage among naturally occurring algae (Henley et al. 2013). Thus, while the predicted risk for these modifications is considered to be low, there has not yet been sufficient data from field trials to properly draw a conclusion.

Metabolic engineering is intended to alter the composition of microalgal bio­mass. While for biocrude producing systems, this will ideally result in strains that have higher overall carbon content, and it is not producing strains outside of the range of what occurs in nature. Nevertheless, if the available proportion of the population containing high carbon (i. e. abundance of GM microalgae relative to wild-type microalgae) is shifted, there is potential for an effect even if the conse­quences are low. If the nutritional value of the microalgae is altered, then this could also lead to changes in the nutritional value of plankton and filter feeders and subsequently lead up the food web to higher trophic levels. By the ecological risk assessment proposed by Henley et al. (2013), this risk is considered to be very low; however, this should be monitored in the longer term to obtain confirmation. In general, the accumulation of energy storage products in the form of reduced carbon molecules does not convey a competitive advantage to GM microalgae and it is likely that they will also be outcompeted by wild types within natural ecosystems.

GM traits that enhance the capacity of a microalga to remain dominant in the presence of predators, pathogens and competitors are varied in their approach and range from resistance to chemicals (e. g. herbicides and pesticides) to the use of allelopathy and toxins to maintain dominance. The use of chemicals is unlikely to become widespread for low-value commodity products such as biocrude due to the economic pressure it places upon business models; however, the engineering of endogenous chemicals into GM microalgae that prevent contamination is a potential risk that must be properly examined. Henley et al. (2013) propose that the risk of this approach is low to moderate depending upon whether the allelopathic chemical is naturally occurring or novel; however, we suggest that the range of potential risk assessment outcomes can be as variable as the potential allelopathic chemicals that can be engineered and that even for relatively low-level allelopathy, at the very large scales of production proposed for addressing fuel demand, even mild allelopathy could have ecosystem disruptive effects. Thus, we advise a strict examination of these strategies; though to the best of our knowledge, these strat­egies have not yet been employed. We do agree though that where traits are selected for from large populations and then elucidated and reproduced through engineering (rather than engineering of novel chemicals), the risk will be attenuated.

Other GM strategies to increase the harvestability and processability of micro­algae are unlikely to affect their dominance in natural systems, and the risk for these traits is considered very low. Similarly, where protein expression is used to create a primary revenue stream from a high-value product before HTL of residual biomass, these strains are unlikely to compete in natural systems due to diversion of much of their energy flow towards a product that is not useful to the microalga.

The theoretical risk assessment discussed here and that presented by Henley et al. (2013) can be quite informative, relying on an analysis of whether similar traits are already part of the ecosystem. However, a physical risk assessment strategy will be more convincing where laboratory-scale simulated ecosystems are developed from natural water bodies and the long-term survivability of GMO algae in mixed culture can be evaluated, e. g. by PCR.

11.4 Conclusion

The commercially profitable production of algal biocrude, at scale, will represent the culmination of a long and parallel development of algal agronomy, biology, GM, bioreactor engineering, harvesting and chemical conversion processes and the development of suitable sensors and control systems, along with their associated modelling and control software. No one innovation will suffice to overcome the formidable challenges faced by this nascent industry, and no actor will have ownership of all the important intellectual property. Since the most significant competitive challenges are between algal technologies and other fuel systems and secondary markets, the field of algal biotechnology stands to benefit greatly from relative openness of sharing data, technology and experience. This suggests that the modified algal strains used for biocrude production in the future will be heavily modified fuel factories equipped with streamlined metabolism, externally control­lable cellular programs, and both sensors and reporting systems for monitoring the state of the system. Biocrude production appears, at this stage, to offer one of the most promising production pathways for algal biofuel production, and genetic manipulation offers a powerful tool for fine-tuning microalgal biofuel production all the way along the development pathway.

Acknowledgments The authors would like to acknowledge support from Australian federal research grants: NHMRC Project Grant APP1074296, ARC Project Grant DP130100346 and the Queensland State Government NIRAP Grant High Efficiency Microalgal Biofuel Systems.

Centrifugation is a standard and efficient process for collection of algal biomass from a dilute solution. It is flexible and can be run continuously to handle huge volumes at scale. The major obstacles to the use of centrifugation in algal biofuel production are high OpEx and CapEx as well as being prone to mechanical problems due to freely moving parts (Bosma et al. 2003). Additionally, high lipid — containing algae are harder to centrifuge and therefore require additional energy to recover by centrifugation.

Generally, it is believed that centrifugation for the primary and secondary dewatering would only be feasible for high-value applications (Molina Grima et al. 2003). However, others feel that a continuous centrifuge would be economical at large-scale Briggs (2004). It is most likely that the use of a centrifuge for secondary concentration with improved and dependable centrifuges could be useful for biofuel production in a continuous and large-scale process.

Hydrocyclones are low-cost continuous centrifuges that have been used for clearing algae and suspended solids from ballast waters. Their efficiency increases with decrease in size of the hydrocyclone and higher flow rates, which could be an issue for industrial scale-up. An example of this issue is a study of solid separation efficiency of 5- and 18-cm-diameter hydrocyclones that delivered 34 and 29 % removal, respectively (Martinez et al. 2007). Recent analyses suggest that the use of hydrocyclones for algal biofuels could be economical (Packer 2009).

Evaluating the available techno-economic studies for the production of biofuels from microalgae leads to the inventible conclusion that using existing technologies, the biofuel production is at least 2 but more likely 4-5 times more expensive than current fossil fuels and biofuels (final costs are dependent on the expected yields, growth system design and harvesting system employed). The major roadblocks to economic fuel production include:

1. High yields: Regardless of the system, it seems a minimum annual average productivity of 30 g/m2/day is required with at least 30 % lipid content. Although highly possible, this has yet to be demonstrated and will require continued strain development either through traditional methods or genetic modification.

2. Low Capital costs: At low productivities (<25 g/m2/day), growth system capital costs dominate and in the case of open ponds, liners are cost prohibitive. With this limitation, it is difficult to see how PBR’s could ever be cost effective, as they typically do not increase aerial productivity rates (Benemann et al. 2012) relative to open ponds. Furthermore, capital and operating costs for dewatering must be lowered through the use of innovative concentrating steps (self — or electro-flocculation followed by settling or DAF). Using centrifuges for com­plete harvesting is not viable (Klein-Marcuschamer et al. 2013).

3. Co-products: In production systems that produce a protein for animal feed, selling prices will typically need to be in excess of $350/tonne to compete with on-site anaerobic digestion of wastes (ANL et al. 2012).

In addition to these challenges, the climate in most locations causes significant seasonal variation in microalgae productivity (5-10 times in the American context), resulting in underutilised capital for significant periods of the year. Furthermore, assumptions associated with low-cost water availability and CO2 availability are very optimistic as very few sites around the world have these resources in com­bination with the correct climate.

With this in mind, two observations can be made as follows:

1. Biofuels from microalgae will be viable as long as significant (disruptive) improvements are made in the growth/harvesting stages.

2. There are a handful of sites around the world where climate, land availability, CO2 and water availability align to transform this potential viability into com­mercial reality.

Considering the current pace of development and intense interest in this space, the former is likely to occur in the next 10-20 years. In the meantime, microalgae will continue to be exploited commercially for high-value products.

Many strains of microalgae are known to grow either heterotrophically or mixo — trophically. They include Haematococcus pluvialis, Chlamydomonas reinhardtii, Chlamydomonas globosa, Scenedesmus acutus, Selenastrum capricornutum, Scenedesmus bijuja, Ankistrodesmus sp., and many strains of Chlorella (Salim 2013; Ogbonna et al. 1998; Chojnacka and Marquez-Rocha 2004; Chojinacka and Noworyta 2004; Chen 1996). Growth of cyanobacteria can also be enhanced by mixotrophic culture, depending on the carbon source used (Lodi et al. 2005).

However, only a few of these strains that can utilize organic carbon sources accumulate more than 20 % oil under normal growth conditions. The oleaginous strains (those that accumulate more than 20 % oil) include Chlamydomonas, Dunaliella, Botryococcus, Chlorella, Phaeodactylum, Thalassiosira, Nannochlor- opsis, and Isochrysis. Out of these, many are known to be capable of growing heterotrophically. The choice of the strain to be used for heterotrophic or mixo — trophic oil production depends on the oil productivity (a product of the cell growth rate and oil contents of the cells), and the quality of the oils.

Many reports have shown that photoautotrophic and heterotrophic culture conditions result in different biomass and lipid yields for the same microalga strain (Xu et al. 2006; Cheng et al. 2009). Chlorella emersonii and C. protothecoides gave the highest average lipid and biomass yield among many strains of microalgae tested (Suali and Sarbatly 2012). Liang et al. (2009) also reported very high oil productivities with Chlorella vulgaris, while Mandal and Mallick (2009) reported that Scenedesmus obliquus has very high potential for oil production. Heterotro — phically cultivated C. protothecoides was reported to be composed of 40-60 % lipid, 10-28 % protein, 11-15 % carbohydrate, and 6 % ash (Xu et al. 2006; Miao and Wu 2004; Zhang et al. 2008). In view of their lipid content, C. vulgaris, C. protothecoides, and C. zofingiensis were reported as candidates for biodiesel production under photoautotrophic or heterotrophic culture conditions (Liu et al. 2008; Miao and Wu 2006; Hsieh and Wu 2009). Park et al. (2012) found that under mixotrophic conditions, oleic acid is comprised of 41-62 % fatty acid in many strains of microalgae, but in some Chlamydomonas isolates, oleic acid comprised only 9-16 % fatty acid, while palmitic and linoleic acid constituted 47-49 % of the total fatty acid content. Although C. vulgaris and C. minutissima are capable of producing high lipid contents, the triglyceride content is low, making them unsuitable for biodiesel production (Stephenson et al. 2010). The biodiesels pro­duced from some Chlorella species were acid methyl esters, linoleic acid methyl esters, and oleic acid methyl esters (Gao et al. 2009). Unsaturated fatty acid methyl esters comprised over 82 % of the total biodiesel content of Chlorella species (Xu et al. 2006; Cheng et al. 2009). Therefore, the properties of the biodiesel produced from Chlorella comply with ASTM 6751, the US Standard for biodiesel (Li et al. 2007). From various reports on the potentials of several strains of mic­roalgae for oil production, no strain can be selected as the best for biodiesel oil production, since both their oil contents and the composition of the oils vary with culture conditions. The choice, therefore, depends on the culture condition and the composition of the culture medium.

Heterokont algae are a monophyletic group with chloroplasts containing chloro­phyll a and c and the accessory pigment fucoxanthin which gives the group a golden-brown color. Marine, freshwater, and terrestrial heterokonts are known and range in the form of giant kelp (brown seaweeds), diatoms, Eustigmatophytes, and Chrysophytes. The later three groups have species that are high-lipid producers. Nannochloropsis (Eustigmatophytes) are primarily known from marine environ­ments but also occur in fresh and brackish water (Fawley and Fawley 2007). All of the species are small (diameter of about 2-3 pm) non-motile spheres with no distinct morphological features, and many are mixotrophic (Das et al. 2011). Nannochloropsis strains contain 30 % lipids under nutrient-replete growth condi­tions and over 60 % lipid content after nitrogen deprivation (Rodolfi et al. 2009; Huerlimann and de Nys 2010). The Chrysophyceae and Xanthophyceae are predominately freshwater organisms, although a substantial number of xantho — phytes are terrestrial.

Diatoms, widely studied as a feedstock for biodiesel production, are important members of planktonic and attached biofilm communities in both freshwater and marine environments (Round et al. 1990). Over 100,000 species are known which are estimated to contribute up to 45 % of total primary productivity in open oceans (Yool and Tyrrell 2003). Diatoms are distinguished by a unique silica cell wall composed of two separate valves and yellowish brown chloroplasts, surrounded by four membranes and containing the carotenoid pigment fucoxanthin as a photo­synthetic accessory pigment. Other xanthophylls are present as well as P-carotene and chlorophylls a and b. The main storage compounds are lipids (TAGs) and a P (1 ^ 3-linked carbohydrate chyrsolaminarin (Horner 2002). Several genera include species known for high lipid content including Nitzschia, Navicula, Amphiprora, Amphora, and Phaedodactylum (Griffiths and Harrison 2009). Diatoms lack flagella (except in sperm of some species), and their dense cell walls cause them to sink in the water column. Planktonic forms rely on turbulence to keep them in the photic zone, and many species regulate their buoyancy using intercellular lipids. Most diatoms are phototrophic, but a few groups are either obligate heterotrophs or are diurnally heterotrophic in the dark when supplied with a carbon source.

Silicon metabolism has relevance in the culture of diatoms as a feed source for biodiesel production. The silicon-laden cell wall is synthesized intercellularly by polymerizing silicic acid monomers (taken up by transporters from the media) in a specialized membranous compartment (Pickett-Heaps et al. 1990). Because silicate is a relatively expensive and an essential nutrient for diatoms, production costs can be raised significantly. However, silicon limitation prevents cell division and trig­gers rapid lipid biosynthesis which may allow for methods to control oil production in a two-stage production process. Nutrient limitation, including N and P that promote lipid hyper-accumulation in a variety of microalgae, has also been shown to promote lipid accumulation in diatoms (McGinnis et al. 1997). However, several studies suggest that Si deficiency stimulates lipid biosynthesis more rapidly and can result in up to 70 % (dry weight) lipid content (Adams et al. 2013).

Metabolic models build a characteristic description of the cell’s phenotypic state and give insights into systems’ emergent properties with respect to metabolic functions, adaptability, robustness, and optimality. Moreover, a metabolic model serves as a basis to investigate questions of major biotechnological importance, such as the effects of directed modifications of enzymatic activities to improve a desired property of cellular systems (Alper et al. 2005). Reconstruction of genome — scale metabolic models has led to a better systems level understanding of microbial metabolism, bridging the genotype-phenotype gap. The steady increase in the number of new genome-scale metabolic models over the past decade is clear evi­dence of their utility in investigating biological systems for many applications, including those with applicability for pharmaceutical, chemical, and environmental industries (Feist and Palsson 2008).

Soon after the release of Chlamydomonas’s genome sequence in 2007, a number of groups began the reconstruction of metabolic network models for this alga, resulting in the reconstruction of its central metabolic network in 2009 (Boyle and Morgan 2009; Manichaikul et al. 2009). Two years later, genome-scale recon­structed networks of Chlamydomonas were released by two groups independently. Chang et al. (2011) published a genome-scale metabolic network model for C. reinhardtii, iRC1080, describing and accounting for *2000 reactions, *1000 metabolites and over 1000 associated gene products. Dal’Molin et al. (2011) described a slightly smaller genome-scale reconstruction (AlgaGEM), which encompassed *1700 reactions, *900 genes, and *1900 metabolites. Both are constraint-based models that can predict genome-scale reaction fluxes under steady state growth conditions, as well as a wide range of other metabolic outcomes (see Sects. 10.4 and 10.5 for more details on steady state models).

The iRC1080 model allows for quantitative growth prediction for a given light source. This was accomplished by setting up new reactions that treat light as metabolites. More precisely, reactions for the absorption of light by photosystem I and II (as well as other light driven reactions such as vitamin D3 synthesis and photoisomerase) were defined with their wavelength specificities and stoichiome­tries. The introduced light reactions can accept different values corresponding to different light intensities the cell is exposed to. In summary, the absorption of photons drives photosynthesis and other reactions according to specified absorption coefficient, stoichiometry of the absorbed photon, and wavelengths. The described “light reactions” of the model were experimentally validated by photobioreactor growth studies under different light sources and intensities, i. e., photon fluxes, demonstrating the general agreement of actual biomass and oxygen yields with those predicted by the model (Chang et al. 2011).

The aforementioned network models (i. e., iRC1080 and AlgaGEM) can greatly facilitate future developments of network reconstructions for other species of green algae by providing a framework that can be modified according to the alga’s species-specific metabolic properties. We note that nongreen algal groups, such as diatoms which are evolutionarily distant to green algae, are likely to have distinct metabolic processes relative to green algae, differing in metabolic wiring and presence or absence of various subsystems in the network. The reconstruction of metabolic networks for these organisms is likely to require a significant adjustment of the existing green algae models, if these were to be used as the framework.

Microalgae are a highly heterogeneous group of microorganisms belonging to many different evolutionary lineages. Different species of unrelated microalgae can have very different surface properties (Eldridge et al. 2012). When the functional groups on the cell surface differ between species of microalgae, this will cause differences in the flocculation behavior (Henderson et al. 2010). Cell surface properties may even differ between strains of the same species (Cheng et al. 2011). For example, mutant Chlamydomonas strains that lack a cell wall were found to flocculate more easily than wild type strains (Scholz et al. 2011). In addition, smaller species have a higher surface to volume ratio and require a higher flocculant dose per unit of biomass (Vandamme et al. 2010).

Flocculation of microalgae differs from flocculation of inorganic colloids and particles in that microalgae are living organisms that can modify their surface properties and interact with their environment through uptake or excretion of substances (Pieterse and Cloot 1997). The flocculation behavior of microalgae can depend on the culture conditions. Zhang et al. (2012) showed that the cell surface of stationary-phase Chlorella zofingiensis cells had lower concentrations of carboxylic groups than during exponential phase and required a lower dosage of alum to induce flocculation. Because the microalgal Z potential changes from exponential to stationary growth, the flocculation efficiency is likely to vary with the growth phase (Danquah et al. 2009b).

Microalgae excrete a substantial fraction of the photosynthesis products as extracellular organic matter. Hulatt & Thomas (2010) showed that up to 17.3 % of the organic matter produced during photosynthesis is excreted in the culture medium. Organic matter concentrations in microalgal culture medium are often 10-100 mg C L-1, depending on species and culture conditions. This organic matter consists mainly of polysaccharides and proteins (Henderson et al. 2008a). The secreted material has a strong inhibitory effect on flocculation, as is evident from much higher flocculant dosages in media with microalgal organic matter compared to media without the organic matter (Bernhardt et al. 1989; Vandamme et al. 2012b). Most of the chemical demand required for flocculating microalgae may be due to the microalgal organic matter present in the media rather than the microalgal biomass itself, as well as impurities present in the medium, such as humic matter (de Godos et al. 2011; Beuckels et al. 2013). Chen et al. (2008) suggested that microalgal organic matter may cause complexation of free metal ions and as such prevent flocculation by metal hydroxides. It is also possible that negative charges on the extracellular organic matter compete with the microalgal cell surface for the positive charges of the flocculants (Garzon-Sanabria et al. 2013).

The semiconductor bandgap determines the optoelectronic properties of the semi­conductor material. A semiconductor’s bandgap has a significant influence on the properties (including absorption in the case of a solar cell) of devices produced from them. The bandgap defines the minimum amount of energy needed for an electron to jump from the valence band to the conduction band.

The value of the bandgap (Eg) is characteristic of each semiconductor. This value affects the properties of the solar cells produced from each semiconductor (McEvoy et al. 2003). For example, semiconductors are effectively transparent to photons of energy less than the bandgap energy as these photons have insufficient energy to excite an electron from the valence to the conduction band and hence are not absorbed.

The minimum room temperature bandgap energy values for some common semiconductors range from 0.67 eV for germanium (Lide 2005) to 1.35 eV for gallium arsenide (Lide 2005). The semiconductors used for solar cells should ideally have a bandgap energy close to the peak of the energy range of light in the AM1.5 spectrum (1-3 eV). Not all semiconductors are appropriate for the use in solar cells. The most suitable semiconductors will have a bandgap of about 1-1.6 eV(Wenham et al. 1994). Silicon, with abandgap of 1.12 eV (Lide 2005), is a good candidate material use in solar cells. Ideally, a solar cell should have a flat response to irradiance of different wavelengths. However, this is not usually the case as each will respond differently to different parts of the spectrum. A mea­surement known as the spectral response can characterize the quantum efficiency of the solar cell to different wavelengths of light.

The spectral response of a solar cell is defined as the short-circuit current (output current under short-circuit conditions) per unit power of incident monochromatic light, as a function of the wavelength of the incident light (Cuevas et al. 2002). The spectral response measurement shows how the solar cell will perform under dif­ferent spectral conditions and can have implications on which technology is deployed in the field. For example, Ruther et al. (2002) have shown that crystalline cells are more suitable for “red” spectra and that amorphous silicon solar cells are more suitable for “blue” spectra (Ruther et al. 2002). This can contribute to the better performance of a-Si:H cells during summer months and the better perfor­mance of c-Si cells during winter months, due to the seasonal variations in the spectra of light received by the solar cell (Ruther et al. 2002). Comparison and analysis of the spectral response measurements for different solar cell technologies enables the most appropriate solar cell to be deployed given the spectra of light they are likely to encounter.

Post-dewatering, the microalgae biomass can be used directly as a source of animal feed or human food. The cultural and economic development of society has resulted in changes in human lifestyles with developed countries’ diets highly caloric, rich in

saturated fats and sugars, with lower consumption of complex carbohydrates and dietary fibre. This has brought about a greater interest in new foods that can contribute to improve nutritional health and well-being (Plaza et al. 2008). Microalgae are certainly candidates for producing high protein (Spirulina), high carbohydrate (Chlorella) and high essential oil similar to fish oil (Diatoms). Fur­thermore, microalgae biomass can be converted to renewable fuels. The three different pathways that can be used to extract and convert microalgae wet biomass (20 % solid) into bioenergy are summarised in Fig. 1.2. To date, hydrothermal liquefaction seems to be the most energetically positive method for biofuel pro­duction from microalgae (de Boer et al. 2012). However, extensive research and development is still required to determine the most energetically favourable and economically feasible process for extracting and converting the algal biomass for renewable bioenergy.

As previously discussed, mixotrophic cultures have very high potentials for oil production by many strains of photosynthetic organisms. However, optimization of mixotrophic cultures is difficult as the conditions that favor photoautotrophic metabolic activities may not favor heterotrophic metabolic activities. The relative contribution of these two metabolic activities depends on such factors as light intensity, the nature, and concentration of the organic carbon source, and carbon dioxide concentration (Ogbonna et al. 2002a, b). For instance, it has been observed that the assimilation of glucose in a certain strain of Chlorella is suppressed by light, even at low light intensities (Haass and Tanner 1974; Kamiya and Kovallik 1987); therefore, only photoautotrophic metabolic activities are observed under mixotrophic conditions. Sequential heterotrophic-photoautotrophic culture system can be used to overcome such limitations. In this system, the cells are first culti­vated to high densities in heterotrophic cultures, and the condition is then changed to photoautotrophic. This system ensures that the advantages of heterotrophic cultures (such as high cell densities) and those of the photoautotrophic cultures (such as light-induced synthesis of metabolites) are realized. The effectiveness of this culture system has been demonstrated for Chlorella (Ogbonna et al. 1997), Euglena (Ogbonna et al. 1999), and Haematococcus (Hata et al. 2001). Sequential heterotrophic-mixotrophic cultivation, in which light illumination is started before the organic carbon source is completely utilized, has also been investigated for biodiesel oil production (Mitra et al. 2012).