In order to use geophysical methods to monitor production from hydrate accumula­tions, a number of fundamental challenges must be considered. They can be sum­marized as follows:

Suitability of geophysical methods depends on geological setting and expected pro­duction behavior. The spatial and temporal evolution of physical properties in a hydrate accumulation differs dramatically depending on the type of the deposit and the dissociation method, thus requiring different monitoring methods. For example, the study of Moridis and Reagan [114] predicted that depressurization-induced pro­duction from the Tigershark deposit would effect changes in the HBL extending to a radius of 800 m from the production well, and including decreasing thickness of the HBL, increasing gas saturation, decreasing SH, and the formation of free gas lay­ers above and below the HBL. Kowalsky et al. [ 83] showed that such complex changes appear to be amenable to detection with time-lapse VSP measurements col­lected in a well 50 m away from the production well. Another study on production from a permafrost-associated hydrate accumulation in North Slope, Alaska [ 15 [ showed changes that were mostly limited to within 5 m from the production well after 2 years of production, making the system far less ideal for VSP monitoring than in the previous case. Cross-borehole measurements may be more promising in some cases, though the spatial coverage they provide is limited to the inter-borehole region, and there is the risk of non-detection if the changes in the HBL are not pro­nounced [210].

Rock physics models dependence on geological setting and time-varying hydrate configuration. A considerable amount of research has been performed to determine rock physics models (the relationships between sediment properties and geophysi­cal properties) for HBS based on theoretical considerations, laboratory experiments, and field data (e. g., [65, 100, 216]). Depending on the geological setting, GH can be distributed in a variety of ways (e. g., acting as cement between grains, acting as the matrix supporting the grains, or existing mainly in pore space), which dramatically affect the seismic and electrical properties of the sediment mixture [206] . For the purpose of geophysical monitoring at a given site, the rock physics model must be determined in advance using site-specific data such as well logging and core data. However, care must also be taken to account for the geophysical properties changes during production because the hydrate and gas saturations may move to ranges beyond those used to develop the models. It may be necessary to change models in the course of production as the hydrate conditions change, a subject that has not received much attention.

Simultaneously changing physical properties can lead to nonunique interpretations of time-lapse geophysical data. Because geophysical properties are a function of the saturations of all phases in the system, in addition to pore fluid pressures, it is difficult to uniquely attribute the change in a geophysical measurement exclusively to SH changes. For example, during depressurization-induced production from a GH deposit, the P-wave velocity can vary due to changes in both the effective pressure and in phase saturations. As the fluid pressure decreases, the stresses and the frame bulk moduli of the sediment increase [41]. At the same time, the increases in velocity are offset by the effects of the decreasing hydrate saturation and increased gas saturation [89].

In R. eutropha cells grown in the presence of O2 under heterotrophic conditions, no detectable Adh activity is seen (Sinskey laboratory, data not shown). Adh has been assayed in R. eutropha grown under anaerobic conditions [31]. To ensure IBT production under aerobic autotrophic conditions, the use of a constitutively expressed, broad substrate specificity Adh enzyme is required. A broad substrate specificity Adh has been characterized in R. eutropha, encoded by the adh gene (locus tag H16_A0757). This enzyme has been shown to exhibit activity on ethanol and 2,3-propanediol [29-31]. Constitutively expressed adh mutants have been iso­lated and characterized. These mutant strains contain alterations in the promoter region of adh, and in some cases have been shown to utilize short-chain alcohols for growth [29]. The Adh enzyme has recently been demonstrated to use isobutyralde- hyde as a substrate to produce IBT (Sinskey laboratory, data not shown). The R. eutropha strains constitutively expressing adh can be used as parental strains to produce the IBT production strain.

E. coli alcohol dehydrogenase YqhD catalyzes the reaction between many alco­hols and their corresponding aldehydes. YqhD belongs to the NADPH-dependent Adh superfamily and is subclassified as zinc-dependent long chain Adh [32] . The crystal structure of YqhD reveals Zn(II) and NADPH as cofactors for catalysis [76]. The active site contains a relative large substrate-binding pocket, which explains its ability to catalyze reactions involving both normal and branched-chain aliphatic and aromatic alcohols and their corresponding aldehydes and ketones. Using YqhD for IBT production could affect the cofactor balance. Because many native metabolic pathways require NADPH, an IBT biosynthetic pathway constructed with an NADPH-dependent enzyme could disturb the native metabolism by competing for this cofactor necessary for growth and maintenance, thus potentially resulting in slower growth or a decreased production yield. Clostridium acetobutylicum has two NADH-dependent Adhs, AdhEl, and AdhE2. AdhEl is expressed under anaerobic conditions and is active towards ethanol, acetaldehyde, butanol, and butyraldehyde [77]. AdhE2 has the same substrate specificity as AdhEl, but is only expressed in high NADH/NAD+ ratio alcohologenic cultures. AdhE2 has a conserved iron-bind­ing motif and is hypothesized to require Fe2+ as a cofactor for catalytic activity [78]. The activities of AdhEl and AdhE2 towards the reduction of isobutyraldehyde have not been explored. AdhEl and AdhE2 are potentially advantageous for IBT produc­tion because they are both NADH-dependent and can use NADH produced directly from the oxidation of hydrogen by the soluble hydrogenase.

The UdhA reaction occurs in an energy (ATP)-independent manner. In a heterolo­gous E. coli system, the reversibility of UdhA was exploited via overexpression of the enzyme to increase NADPH availability and PHB production [82]. The increased intracellular NADPH concentration allowed a greater cofactor availability for the acetoacetyl-CoA dehydrogenase reaction [82]. A similar strategy was used in Pseudomonas fluorescens to increase production ofhydromorphone [83]. This strat­egy for altering redox cofactor pool size can be used in the IBT synthesizing strain of

R. eutropha to ensure efficient activities of NADPH-requiring enzymes in the biosyn­thetic pathway.

In most organisms, the redox cofactor NADPH is generated via the pentose — phosphate pathway by glucose-6-phosphate dehydrogenase and 6-phosphoglucon — ate dehydrogenase. Proteomic studies of R. eutropha revealed no active

6- phosphogluconate dehydrogenase under heterotrophic growth conditions [ 27], suggesting that R. eutropha synthesizes its cofactor NADPH from other pathway(s) during organoheterotrophic growth. One potential pathway involves the maeA or maeB genes, encoding malic enzyme. The malic enzyme catalyzes the conversion of malate to pyruvate and is part of a metabolic cycle that also includes pyruvate carboxylase and malate dehydrogenase (Fig. 5b). An increase in transcription of the malic enzyme gene also upregulates the transcription of both pyruvate carboxylase and malate dehydrogenase [84]. Another enzyme, nonphosphorylating glyceralde — hyde 3-phosphate dehydrogenase (GapN) bypasses 1,3-bisphospho-D-glycerate in glycolysis and generates an additional NADPH from NADH, at the expense of one ATP (Fig. 5c). In order to increase the intracellular NADPH pool, udhA and gapN can be heterologously expressed in R. eutropha or maeA could be overexpressed to increase IBT production.

Light impinges on the reactor surface, is absorbed by cells, scattered and reflected and thus light intensities necessarily decrease with increasing distance from the surface. Consequently, light cannot be provided with equal intensities for all cells

reaction

Fig. 2 Interdependency ofbiochemical reaction, light transfer, and fluid dynamics in a photobio­process [37]

within the reactor at the same time. Furthermore, the incident light intensity is sub­ject to daily and seasonal changes, as well as weather influences. Even increasing cell concentrations strongly alter light distribution in the time course of a single cultivation due to absorption, scattering, and mutual shading.

All photobioreactor concepts apply the same common design principle of a lim­ited light path length (Table 1). Light gradients in the reactor are inevitable. Nevertheless, a plate or tubular thickness that significantly exceeds the light path length, leads to an increased dark volume. This generally impairs the overall pro­ductivity because microalgae shift to respiratory metabolism when photosynthesis is stopped. The significance of respiratory losses can be deduced from respiratory maintenance metabolism during night hours. Respiration can cause a biomass loss of up to 25% of biomass produced during the day [11, 23].

Many attempts to simulate growth of algae cultures assume an exponential decline of light intensity with increasing distance from irradiated reactor surface [44]. High cell densities tremendously limit the light path length. Exemplary mea­surements show that at a cell concentration of 10 g/L (Arthrospira platensis), about 95% of incident light (I0 = 1,925 mE/m2/s) is absorbed or scattered along the first 2 mm of the light path [46].

However, high cell concentrations are desirable because downstream processing (DSP) usually requires high energy input, e. g., for centrifugation or spray drying. High cell concentrations increase efficiency of DSP by reducing energy input and cost per biomass yielded. Cell densities of about 5 g/L were reached with Chlorella in semicontinuous cultivation experiments in airlift-photobioreactors [12] and max­imal dry weight concentrations of Phaeodactylum tricornutum cultures under out­door conditions of around 7-8 g/L were attained in a Flat-Panel-Airlift reactor (Subitec, Germany) [35].

Light intensity on reactor surface l0 / (|jEm*2S‘1)

Fig-3 Effect of light intensities on growth kinetics of P purpureum in turbidostat cultivation mode under homogenous light conditions (filled circle growth rates resulting from continuous illumina­tion; filled square growth rates when cells are exposed to light/dark cycles)

Provided that the light path length exceeds the plate thickness in flat plate reac­tors, exponential growth of the culture can generally be achieved. Transmitted light is not necessarily “lost” but can be captured by other compartments of the facility, e. g., when parallel reactors are arranged in fence-like structures (Fig. 4). Otherwise, all photons will be absorbed thus leading to linear growth on condition that no sub­strates become limiting.

The tremendous implication on scale-up is that the rule of geometric similarity on different scales cannot be applied to photobioreactors. Instead, one dimension is more or less fixed. Scale-up is limited to the remaining dimensions.

The assumption that exposure of microalgae cultures to high irradiances neces­sarily increases productivity would be misleading.

A look at Fig. 3 reveals that growth rates show a linear increase with light inten­sities only in a very narrow range. As shown here for the model organism, Porphyridium purpureum cultures are light limited when light intensities impinging on the reactor surface range up to ca. 100 mE/m2/s. Higher intensities have almost no advantageous effect on growth kinetics since dark reactions in the CO2 fixing Calvin-Benson cycle become kinetically limiting and therefore the availability of NADP+ and ADP for any further conversion of the H+ gradient across the thylakoid membrane to NADPH and ATP is restricted. Excess light that is harvested by the algae is dissipated as heat or fluorescence by pigments [51]. Cells are said to be light saturated. Growth rates do not increase linearly when light intensities are raised but rather stay constant over a wide range. Such inefficient utilization of light will necessarily result in low PCE values.

Further increase in light intensities can even damage proteins involved in the photosynthetic apparatus and inhibit cell growth. This phenomenon is called

light inhibition. It is characterized by reduced growth rates when light intensities are further increased [2].

The reaction of microalgae to various light intensities is affected by adaptation processes over the day cycle and also dependent on the specific strain. The latter should already be taken account of in screening programs for isolation of new strains.

An optimized operating point for photobioreactors lies in a range where light is limiting and saturation is avoided (as indicated by the vertical, dashed lines in Fig. 3). These conditions fundamentally determine photobioreactor geometry.

A high surface to aperture area (or ground area) is attributable to the fact that high midday summer light intensities need to be avoided and thus light is spread over a larger surface area. Otherwise, cultures would be exposed to the high light intensities that lead to saturation or even inhibition. This generally applied concept is referred to as “light dilution.” Furthermore, limited light penetration depth confines reactor geometry in one dimension. Consequently, reactors are flat or con­sist of tubes with small diameters. Finally, a high surface to volume ratio is attribut­able to the other two demands on geometry.

The Green Wall Panel reactor is one example for light dilution in practice [45]. Vertical reactor compartments of flat panels in fence like arrangements collect light rays at large angles and therewith achieve a dilution effect (Fig. 4). In particular, high areal productivities and efficient light utilization can be obtained since the modules can be placed in relatively short distances (e. g., 1 m high modules, 0.9 m distance [36]) and even light-averted surfaces collect reflected and diffuse light [46].

Table 1 gives an overview of different reactor concepts and their corresponding surface to aperture area ratios.

Although individual reactor and scale-up concepts significantly vary, it becomes clear that all reactors provide high surface areas to collect the incident light (surface:volume ratios of 30-86 are shown) and small depths (3-7 cm) account for limited light path length. Moreover, all reactors dilute light that is collected from a certain area at least by a factor 3.6 (surface:aperture area ratios: 3.61-5.45) in order to avoid excess light intensities.

Midday light intensities of about 1,000 pE/m2/s are not uncommon in Europe in the summer season (2,000 pE/m2/s in equatorial regions) [ 10, 30]. Therewith, a surface to aperture area ratio of 10 or even more would be reasonable if the specific algal strain reacted similarly to high photon flux densities like P. purpureum, as depicted in Fig. 2.

Adjustments for every individual facility to the requirements and characteristics of the specific algae as well as the location, latitude, and climatic conditions need to be considered.

The biodiesel in this study is produced by conversion of dry microalgae biomass from the VIP using a one-step thermochemolysis reaction into an open pyrolysis prototype reactor. The biomass may be used partially wet, but Hatcher and Liu rec­ognized that drying increases the yield of FAMEs [15] . During this reaction, the transesterification occurs at a temperature sufficient to hydrolyze and alkylate trig­lycerides in the biomass [15] . This process is possible because the association of TMAH and methanol provides a single step high-temperature saponification and methylation of the ester functional groups of the triglycerides [11]. It is also likely that a transesterification occurs at lower temperatures (approximately 100°C) to produce the FAMEs, much like what is observed with alkali like sodium hydroxide [12] as long as there is residual methanol to act as a transesterification reagent. Using temperatures above 100°C and evaporating the methanol before insertion of the reaction mixture into the reactor created conditions that minimized the transesterification product at low temperatures. The produced FAMEs become vola­tile at reaction temperatures above 250°C and can be condensed at the exit of the reactor. To establish the optimum reaction temperature, multiple assays at tempera­tures ranging from 250 to 550°C were used. The volatile products trapped in an ice bath were collected and analyzed by GC-MS in order to identify and quantify the individual FAMEs and other products for each assayed reactor temperature.

The total ion chromatograms (TIC) for VIP algae assayed at 250 and 450°C are shown and peak identifications and their concentrations are reported (Fig. 2 and Table 1). Mostly linear saturated and unsaturated FAMEs are detected in the chro­matogram. The dominant peaks are fatty acids, typical of biodiesel, containing C140, C160, C161, and C181 corresponding to peaks 1, 6, 5, and 9a, b in the table [8].

It is also worth noticing that both chromatograms of biodiesel produced at 250 and 450°C are very similar showing that the conversion of triglycerides to biodiesel using TMAH/MeOH reagent does not change its selectivity for FAMEs with respect to temperature. A similar pattern was observed at all assay temperatures. Minor amounts of other compounds are observed in the products.

All of the peaks correspond to FAMEs except peaks 4, 8, and 11 which are respectively ascribed to a C20 isoprenoid hydrocarbon, a lactone, and dodecanamide. The isoprenoid hydrocarbon probably is produced from chlorophyll and the dode- canamide from reaction of ammonia with triglycerides in the reaction chamber.

The yields of FAMEs are summed for each assay temperature along with an estimate of the standard deviation of the analysis for multiple assays at each tem­perature (Fig. 3). A yield of approximately 3% (±0.3%) is achieved at 250°C for the VIP algae. The highest yield of FAMEs, approximately 4% (±0.8 and 0.6%) of the dry biomass, is achieved at 350 and 450°C. At higher temperature, a slight decrease in yield of FAMEs is observed with a yield below 3% (±0.2%) at 550°C. Total lipid content obtained by Bligh-Dyer extraction was found to be on average 12% (±3.6%) from Scenedesmus spp., which is the dominant species from the VIP algae, and similar to reported values [3, 6, 31]. The lipid extract is determined by gravimetric

yield and includes triglyceride lipids as well as other extractable organic and inorganic materials. The TMAH procedure only converts the triglycerides and free fatty acids to methyl esters, and we can determine that approximately 30-40% of the total extract is converted to FAMEs. Samples run at each temperature range show standard

deviation overlap of percent yield. Though a slightly higher yield is observed for algae run at 350 and 450°C, they are not significantly different from those run at the lower temperature of 250 and higher temperature of 550°C. The effect of tempera­ture difference, however, is apparent in the algae residue after biodiesel collection.

The algal residue obtained following the thermochemolysis may potentially be useful as a fertilizer and even as an animal feed. Our initial goal was to examine its chemical composition for possible future evaluation as these end uses. Raw algae in coastal regions, such as seaweeds, have a long history of use as soil conditioners and fertilizers [14]. Seaweed composting, however, presents some problems such as high salinity and excessive sand content that can limit plant growth and develop­ment [10 ] . Using algal residue produced in our thermochemolysis reactor from freshwater algae would potentially eliminate these problems. The carbon and nitro­gen content of algae and algal residues are collected at different temperatures (Table 2). The algal residue collected at 250°C had an average carbon of 46.67% and nitrogen of 5.83%. Both carbon and nitrogen values decreased slightly with increasing temperatures. The C/N ratios, however, were highest at a temperature of 450°C with a value of 9.41. Further testing is being conducted to verify the use of algae and algae residue as an organic fertilizer.

The solid-state CPMAS 13C NMR spectra for whole algae and residue collected at various temperatures are displayed (Fig. 4). Whole algae samples collected both

200 50

Fig. 4 13C solids NMR (a) whole algae VIP, (b) whole algae Scenedesmus/Desmodesmus spp. from algae farm Spring Grove, VA, (c) VIP algae residue from reactor at 250°C, (d) VIP algae resi­due from reactor at 450°C from the VIP and the algal farm are shown. Dominant species in the VIP samples were microscopically identified as pennate diatoms and in the algae farm samples as Scenedesmus/Desmodesmus spp. The NMR spectra of the two algal collections are quite similar, being dominated by strong signals in the 0-60 ppm region, repre­senting lipid-like aliphatic carbons and proteins. Carbohydrates, characterized by signals between 60 and 105 ppm, are subordinate components of the spectra indicat­ing that they constitute a small fraction of the algal biomass. Peaks in the region for aromatic/olefinic carbons (105-160 ppm) are also subordinate, reflecting aromatic amino acids comprising proteinaceous components of the algae and olefinic struc­ture contained in the lipids. The large peak at 175 ppm is assigned to amide and carboxyl groups, structural components of both lipids and proteins.

When the algal samples are treated with TMAH at different temperatures, some significant changes are observed in what remains as a residue. The spectrum shown has many of the same signals as the original algae, but contains significantly more olefinic/aromatic carbon (Fig. 4c). This is most likely an indicator that the 250°C heating has transformed the algae such that increased aromatization occurs. The transformation, however, is not overly drastic, suggesting that this residue probably has sufficiently preserved structures to be used as a fertilizer. It is well known that
heating materials containing carbohydrates and proteins (algae) together produces furanosic materials often referred to as melanoidins [2, 20].

The diminution of carbohydrates (60-105 ppm) and portions of the proteina­ceous signals (NCH at 50 ppm) observed in this residue is consistent with the involvement of carbohydrates and proteins to form these aromatic substances. The presence of a peak at 160 ppm (assigned to the O-bearing aromatic carbons), along with a broad signal in the range between 105 and 150 ppm, is suggestive of furans, but also could be derived from heterocyclic N rings such as indoles or pyrroles. Usually, a signal at 160 ppm is assigned to phenolic substances that are commonly found in lignin. We know that lignin is not present within algae, so the emergence of this peak at 160 ppm can only be rationalized as that derived from furans or heterocyclic N. Heating to a temperature of 450°C causes drastic changes in the structure of the residue, which shows signals mainly like charred material. The high aromatic content and minor aliphatic content is suggestive of extensive polymeriza­tion into condensed aromatic moieties. We can speculate that the furans forming from carbohydrates might undergo cyclization and aromatization and eventual for­mation of dibenzofurans. The protein-derived heterocyclic N might probably undergo increased aromatization to structures similar to carbazoles. Interestingly, the C/N ratio for this material (9.41) coupled to the high aromatic content are suggestive of the fact that a significant amount of N-containing structures are embodied within aromatic units. Use of this material as a fertilizer where the N can be released via decomposition is unknown at this time.

A basis of 90% total lipid recovery is assumed for design calculations. Using the proximal composition in Fig. 3 , the annual yield of oil from P. tricornutum with 90% recovery is ~9,000 tonnes. However, this amount could reduce when the extract is purified and all unsaponifiable components are removed. The design is focussed on the use of solvent extraction on wet biomass with no cell lysis unit operation. Lee et al. [17] demonstrated that cell lysis before solvent extraction does not significantly affect lipid yield. This extraction method is employed due to its high lipid yields and minimal extraction steps. It also involves the use of low-solvent quantities. The total amount of saponifiable lipids extracted from P tricornutum is approximately 6.4% of biomass (dry weight) [27]. Therefore, the daily yield of purified lipid from the extraction is approximately 7.78 tonnes. Figures 6 and 7 show the product yields from the crude lipid extraction and purification processes.

Based on economic and design considerations, the two-stage dewatering process of chitosan 2 and disc-stack centrifugation is recommended. The recommendation is based on an overall study of the process steps: cultivation, dewatering, extraction and biodiesel production. The net emission results for the complete process are

shown in Table 7. The results from the Scope 2 emission audit indicate that the over­all emissions from the HTR and ELR options are by a factor of 10 greater than that for the RP option. As seen in Fig. 15, the majority of the emissions for the HTR and ELR are due to emissions from the cultivation stage (98 and 97% of total emissions, respectively), whilst the emissions from the cultivation stage for the RP only amounts to 33% of total emissions. If the cultivation emissions were ignored, as seen in Fig. 16, the data indicate that to dewater, extract and produce biodiesel more emis­sions are produced for the RP, than for the ELR or the HTR. This is due to the emis­sion rating to dewater large volumes of less concentrated algae culture from the RP.

The fundamental importance of this project is CO2 biosequestration. By captur­ing CO2 , the process reduces the overall emissions which would otherwise be released into the atmosphere. This would reduce the number of permits the facility or an industry is required to obtain. The financial savings analysis shows that the HTR and ELR options save $2.18 million (87,000 permits) whilst the RP saves $2.04 million (82,000) per year. The higher permit saving for the HTR and ELR

150000

100000

50000

0

-50000

-100000

■ Net NGER □ Net CPRS

Fig. 17 The net emissions according to the CPRS and NGER requirements options is due to the lower Scope 1 emissions resulting from their higher CO2 cap­turing efficiencies. In terms of an overall outcome, the best option is RP cultivation followed by the two-stage dewatering process. As seen in Fig. 17, the HTR and ELR actually produce more emissions than the RP due to the high cultivation emission rating, whilst the RP has an overall negative emissions rating. As the use of HTR or ELR produces significantly more emissions than the RP, it has a greater negative environmental impact.

2 Conclusion

In terms of the design outcomes, the HTR and ELR appeared to be attractive options due to their ability to achieve high biomass concentrations during cultivation. However, the fixed capital cost involved with the HTR and ELR are up to 493% greater than that for the RP. The tubular reactors are difficult to scale-up due to issues of dark zones and dissolved oxygen build-up. Therefore, the number of units required proved to be significant. The results of the carbon audit indicated that the overall emissions from the HTR and the ELR were greater than that for the RP by a factor of 10. This was largely due to energy consumptions associated with the use of airlift pumps.

In dewatering, a two-stage process involving flocculation preceding centrifuga­tion heavily reduced energy consumption with high reduction in emissions, com­pared with a single-stage process. The carbon study also indicated that dewatering using a two-stage process was more attractive as culture volume increased, even with a low-efficiency flocculant. In the extraction and transesterification stages only one design alternative was investigated, and on the basis of 50,000 tonnes biomass processing, the results were identical for the different cultivation system.

The overall findings from this study indicate that a RP cultivation stage, followed by a two-stage dewatering process is the optimum alternative. However, the eco­nomic study showed that this option is not feasible presently due to an excessively high cost of production of $74/L of biodiesel, which leads to an annual operating loss of $190 million. The carbon audit, however, indicates that the process is carbon neutral, capturing ~49,000 net tonnes of CO2-e. In terms of environmental impact, the project is attractive. However, it would not be financially viable, as the value of carbon permits is only $2.0 million.

Microalgae contain both chlorophyll a and chlorophyll b. Intracellular chlorophyll a content can vary from 0.0041 g/g dried microalgae (Synechococcus sp.) to 0.0185 g/g dried microalgae (Nannochloropsis gaditana) [53]. In green microalgae, the ratio of chlorophyll a/chlorophyll b ranges broadly from 0.64 to 5, in contrast to higher plants which have a narrower range from 1 to 1.4. The chlorophyll content and profile of a microalgal species continuously change depending on its life cycle and cultivation conditions (medium composition, nutrient availability, temperature, illumination intensity, ratio of light and dark cycle, aeration rate). Certain green microalgae, such as Chlamydomonas, Chlorella, and Scenedesmus sp., have mutants that can synthesize chlorophyll in the dark during heterotrophic growth [53] .

Figure 2 shows the downstream processing steps required to produce chlorophyll from microalgae, while Table 1 provides a list of different technologies currently

available for each step. After the microalgal culture is harvested from its cultivation system, it is concentrated in the dewatering step to yield a wet paste. Afterwards, the microalgal pellet undergoes a pre-treatment step for preparation towards chloro­phyll extraction. The chlorophylls are then extracted from cellular materials before being purified in a fractionation step [24].

The content and composition of algal lipids vary with species, geographical location, season, temperature, salinity, light intensity, or combination of these factors. In gen­eral, algae contain up to 1-3% of dry weight of lipids, being glycolipids the major lipid class in all algae, followed by neutral and phospholipids.

The major polar lipids that can be found in microalgae are monogalactosyl dia — cylglycerols (MGDGs), digalactosyl diacylglycerols (DGDGs), and phosphatidylg — lycerol (PG) [2]. Although these compounds, primarily MGDGs and DGDGs, have been known for more than 40 years, their importance has been recently raised by the description of their different, mainly anti-inflammatory, functional activities [15]. For example, glycol analogs of ceramides and of PG with antithrombotic and anti­inflammatory activities have been reported in cyanobacteria [2] . MGDGs and DGDGs contain a galactose linked to the sn-3 position of the glycerol backbone. These polar lipids are found in the thylakoid membrane of the cells. For instance, several polar lipids have been identified in Spirulina platensis, such as, four MGDGs, three PGs and two sulfoquinovosyl diacylglycerol [57], in Croococcidiopsis sp. [2], in Sargassum thunbergii [81], and Phormidium tenue [124] among others.

On the other hand, most of the alga’s lipid content is made of polyunsaturated fatty acids (PUFAs) which accumulation also relies on environmental factors. For example, it is known that algae accumulate PUFAs when there is decrease in the environmental temperature [80]. In this sense, it has been described that tropical species contain less lipid (<1%) than cold water species (1.6%) [125].

PUFAs are essential nutrients for humans, and must be obtained from food. w-3 and w-6 long chain PUFAs are structural and functional components of cell mem­branes. The w-3 to w-6 ratio is closely matched, a factor that has been found to be important in balanced diet [176]. Likewise, these fatty acids are precursors of eico — sanoids, which exert hormonal and immunological activity. This means w-3 and w-6 should be consumed in a balanced proportion, with the ideal ratio w-6: w-3 ranging from 3:1 to 5:1 [184].

The properties of the long-chain w-3 fatty acids eicosapentaenoic acid (EPA) (w-3 C20 5) and docosahexaenoic acid (DHA) (w-3 C22 6) have been followed with considerable interest in the last few years. In particular, the vascular protective effects of long-chain w-3 fatty acids are well documented [17, 170, 207]. Green algae show interesting levels of alpha linolenic acid (w-3 C18:3). The red and brown algae are particularly rich in fatty acids with 20 carbon atoms: EPA and arachidonic acid (w-6 C20:4).

S. platensis is a microalga belonging to the group of cyanobacteria (or blue-green algae) and is a natural source of DHA, which can account for up to 9.1% of the total fatty acids content [199].

Table 1 [133, 161] presents the typical composition of different fatty acids in algae. As can be seen, in all algae studied except Undaria pinnatifida and Ulva lac — tuca the single most abundant fatty acid was palmitic acid (which in Phorphyra sp. accounted for 63.19% of all fatty acids) while in U. pinnatifida the palmitic acid

content (16.51%) was only exceeded by that of octadecatetraenoic acid (w-3 C18.4) (22.6%), and in U. lactuca the C16.0 content (14.0%) was only exceeded by that of oleic acid (w-9 C181) (27.43%). However, all the seaweeds also contained the essen­tial fatty acids linoleic acid (w-6 C18.2) and linolenic acid and the icosanoid precur­sors, arachidonic acid and EPA. Furthermore, the w-6:w-3 ratio, which the WHO currently recommends should be no higher than 10 in the diet as a whole, was at most 1.49 so that these algae may be used for reduction of w-6:w-3 ratio. Saturated fatty acid contents were higher in the red algae (Palmaria sp. and Porphyra sp.) than in the brown and green algae, and vice versa for relative total unsaturated fatty acid contents. Whereas in the red algae, C20 PUFAs were as a class 8-12 times more abundant than C18 PUFAs, in green algae the opposite occur while in brown algae these two classes of fatty acids were more or less equally abundant. Relative essen­tial fatty acid contents were higher in brown and green algae than in red algae.

Several researchers have reported the fatty acid composition of total lipids of different species of Sargassum. Heiba et al. [47] studied the fatty acids present in four different Sargassum species in the Phaeophyta class that contained heptade — canoic acid (C17.0), eicosanoic acid (C20.0), eicosatrienoic acid (w-3 C20.3), and DHA. On the other hand, Khotimchenko [ 80 ] . working with seven Sargassum species from different parts of the world, determined similar fatty acid compositions in all of them. The site of collection only seemed to affect palmitic acid (C16.0) and C.0 PUFA contents and was connected mainly with water temperature.

Aquatic plants possess conjugated fatty acids (CFA) with carbon chain length varying from 16 to 22, as natural constituents in their lipids; both trienes and tet — raenes occur in aquatic plant lipids. There is not much information available on the literature, only a few reports on the occurrence of these conjugated polyenes in Tydemania expeditionis, Hydrolithon reinboldii [69], Ptilota [205], Acanthophora [8], and Anadyomene stellata [6] have been published. Various enzymes in aquatic plants are thought to be responsible for the formation of conjugated trienes/tet — raenes endogenously. The enzymes responsible for the formation of CFA can be grouped into three main categories of conjugases, oxidases, and isomerases. Hideki and Yuto [58] studied the selective cytotoxicity of eight species of marine algae extracts to several human leukemic cell lines. It has been reported recently that conjugated PUFA, such as conjugated EPA, conjugated AA, and conjugated DHA, prepared by alkali isomerization had profound cytotoxic effects against human can­cer cell lines [102].

Besides fatty acids, unsaponifiable fraction of algae contain carotenoids (see Sect. 4.1), tocopherols (see Sect. 4.5), and sterols. The distribution of major sterol composition in macroalgae has been used for chemotaxonomic classification. Recent biological studies have demonstrated that sterols and sterol derivatives pos­sess biological activities. Currently, phytosterols (C28 and C29 sterols) are playing a key role in nutraceutic and pharmaceutical industries because they are precursors of some bioactive molecules (e. g., ergosterol is a precursor of vitamin D2, also used for the production of cortisone and hormone flavone and has some therapeutic applica­tions to treat hypercholesterolemia). Phytosterols have also been shown to lower total and LDL cholesterol levels in human by inhibiting cholesterol absorption from the intestine [37]. High serum concentrations of total or LDL cholesterol are major risk factors for coronary heart disease, a major cause for morbidity and mortality in developed countries. In addition to their cholesterol lowering properties, phytoster­ols possess anti-inflammatory and anti-atherogenicity activity and may possess anticancer and antioxidative activities [37] .

From a chemotaxonomic point of view, literature data show that major sterols in red algae are C27 compounds and cholesterol occur in substantial amount. It is gen­erally the primary sterol. Desmosterol and 22E-dehydrocholesterol are present in high concentrations and may even be the major sterols in any red algae.

Sterol content in green algae is similar to higher plants, and also contains large amounts of cholesterol. But in green algae, the dominant sterol seems to vary within the order and within the family.

In brown algae, the dominant sterol is fucosterol and cholesterol is present only in small amounts.

Fucosterol content in H. elongata and U. pinnatifida was 1,706 mg/g of dry weight and 1,136 mg/g of dry weight, respectively, as demonstrated by Sanchez — Machado et al. [ 162] . Mean desmosterol content in the red algae ranged from 187 mg/g for Palmaria sp. to 337 mg/g for Porphyra sp. Cholesterol, in general, was present at very low quantities, except in Porphyra sp. that can contain up to 8.6% of the total content of sterols as cholesterol [162].

Sterol content determined in red alga Chondrus crispus showed that the main sterol was cholesterol (>94%), containing smaller amounts of 7-dehydrocholes­terol and stigmasterol and minimum amounts of campesterol, sitosterol, and 22-dehydrocholesterol [188].

According to the investigation carried out by Kapetanovic et al. [77], the sterol fractions of the green alga Codium dichotomum and the brown alga Fucus vir — soides contained practically one sterol each, comprising more than 90% of the total sterols (cholesterol in the former and fucosterol in the latter). The main sterols in the green alga U. lactuca were cholesterol and isofucosterol, while in the brown algae Cystoseira adriatica, the principal sterols were cholesterol and stigmast-

5- en-3 beta-ol, while the characteristic sterol of the brown algae, fucosterol, was found only in low concentration [77]. However, fucosterol was the major sterol present in Cystoseira abies-marina (96.9%), containing low concentration of 24-methylenecholesterol (1.1%), brassicasterol (1.2%), and cholesterol (0.7%) [120].

A summary of the methane yield and volumetric production rate during algal diges­tion in continuous reactors appears in Fig. 11. The curves drawn through the data points on each figure represent general trends for easy comparison of the nonuni­form data.

Based on these data, algae as a substrate for ADP can be classified into three groups. The first group consists of algae with the methane yield larger than 0.3 L/gVS and includes brown macroalga M. pyrifera with high mannitol content, cyanobac­terium Arthrospira. green microalgae Chlorella and Scenedesmus. The second group has methane yield about 0.2 L/g VS and includes brown macroalga Laminaria and green macroalgae Ulva, Cladophora. and Chaetomorpha. Lastly, the third group has methane yield lower than 0.15 L/g VS and includes brown macroalga Sargassum, red macroalga Gracilaria, and green macroalga Enteromorpha.

Another important conclusion is that AD of M. pyrifera is stable at values of OLR up to 10 gVS/L-day. High values of the methane volumetric production rate are achieved 2.7 L(CH.)/L(digester)-day, reducing the required volume and the capital costs of the digester.

Finally, several methods such as application of advanced digestion reactors, co­digestion of algae with other substrates, algal hydrolysis and extraction of cellular liquids, and digestion of alginate extraction residues, significantly enhance methane yield, production rate, and the overall process efficiency.

The ability of green algae to produce hydrogen was discovered over 70 years ago [438, 439]. Hydrogen is recognized in the US and EU as a promising future fuel [440]. While the technology of generating hydrogen from algae is far from being at the industrial scale, several laboratories are working on improving the hydrogen production efficiency by using transcriptomics, proteomics, and metabolomic data [441-443].

The ADP can be used to process surplus algal biomass into methane or hydro­gen. A combined biorefinery concept has recently been proposed [157]. During the first step, C. reinhardtii produces hydrogen via a sulfur deprivation method. During the second step, it is digested anaerobically for methane production. The authors reported approximately 123% more biogas production from algae after the hydro­gen production cycle compared to fresh algae due to the accumulation of lipids and carbohydrates during the sulfur deprivation step [443, 444].

Phototrophic Microbial Fuel Cell

Light energy can be converted into electrical energy via a PMFC [445-448]. The combination of an ADP and a PMFC in a closed loop system can produce methane and electricity [237]. In this system, algal biomass serves as a substrate for AD and provides oxygen (final electron acceptor) for the fuel cell. The liquid phase of digested biomass contains compounds and nutrients to supply electrons in the fuel cell and for subsequent use as media for algal growth. While the system failed to operate in continuous-flow mode, in batch experiments it successfully produced methane (0.32 L/gVS) and power (1.33 mW/m2 illuminated footprint area). Possible reasons for the failure of a continuous system are low biomass concentration, high nitrate concentration, or high water circulation flow rate [237] .