Among major barriers to optimal H2-production yields in algal cultures are high sensitivity of algal [Fe—Fe]- hydrogenases to O2 inactivation, low light saturation levels of photosynthesis, competition for reductant from alternative metabolic pathways, state transition and establishment of cyclic electron flow around PSI, and the reversible nature of the hydrogenase— driven reaction. All these barriers have been exten­sively studied by different research groups in the last few years.

Clearly, the O2 sensitivity of [Fe—Fe]-hydrogenases is the major barrier preventing the application of green algal H2 photoproduction in commercial systems. Several approaches for solving the O2-sensitivity issue have been suggested: (1) identifying and implementing mutations, which narrow the channel(s) of the [Fe—Fe]-hydrogenase enzyme for blocking access of O2 molecules to the catalytic center (Cohen et al., 2005; Posewitz et al., 2009); (2) selecting for O2-tolerant en­zymes through random mutagenesis (Nagy et al., 2007; Stapleton and Swartz, 2010); (3) introducing enzymes from other organisms, which are more stable to O2 inac­tivation, into algal cells. None of these approaches have yet resulted in a mutant with improved O2 tolerance. Nevertheless, Stapleton and Swartz (2010) applying the directed evolution approach identified a version of C. reinhardtii HydA1 with a fourfold increase in catalytic activity as compared to the wild-type enzyme.

Low light-utilization efficiency in mass cultures is another important factor precluding the use of H2- producing green algae in practical applications (Torzillo et al., 2003). In algal suspensions, light intensity decreases with the depth of the culture. The light atten­uation is more pronounced in dense cultures, where shading limits the productivity of inner parts of the cul­ture. On the contrary, algae in the upper layers suffer from photoinhibition, which is more pronounced under high light intensities. The latter significantly limits application of high light intensities for improving the overall algae productivity. The problem can be addressed in part by immobilizing algae in thin layers or films. Immobilization fixes algal cells within a controllable volume and allows uniform light distribu­tion to the cells that makes light utilization per volume basis more efficient. Indeed, immobilization of sulfur — deprived C. reinhardtii cultures on glass fiber matrices demonstrated significant improvements both in the volumetric rate of H2 photoproduction and in the dura­tion of the process (Laurinavichene et al., 2006). This technique used the property of microalgae to form biofilm on the glass surface. The attachment of cells occurred through natural colonization that, if required, can be accelerated by activating glass fibers with 3-(2-aminoethyl-aminopropyl)-trimethoxysilane (Tsygankov et al., 1994). Later studies of immobilized algae with either a constant flow of medium containing micromolar sulfate concentrations or cycling of immobi­lized cells between minus and plus sulfate conditions improved the duration of H2 production up to at least 3 months (Laurinavichene at al., 2008). However, due to irregular colonization of glass fibers by the algal cells, the system showed significant physical and physiolog­ical heterogeneities in different parts of the matrix, resulting in irregular light and nutrient distributions, and decreasing the overall performance of H2 photopro­duction. In order to improve the light absorption proper­ties of immobilized microalgae, Kosourov and Seibert

(2009) entrapped cells within thin alginate films. This technique produced films with uniform distribution of algal cells within the matrix that had very high cell densities (up to 2000 mg total Chl per ml of the matrix). As a result, the light conversion efficiency in alginate films at ~ 60 mE/m2 s PAR (photosynthetic active radia­tion) achieved 1.5% for the period of the maximum H2- production rate and was close to 1% for the whole period of nutrient deprivation.

Another approach for improving light utilization effi­ciency in mass algal cultures is to find or generate algal mutants with a small chlorophyll antenna size. Strains with the truncated antennae allow greater transmittance of irradiance through the ultrahigh cell density culture without significant dissipation of light energy and, as a result, have a higher photosynthetic productivity in out­door conditions. Recently, C. reinhardtii mutants with truncated chlorophyll antennae were generated and characterized (Polle et al., 2000, 2003). These mutants have shown promise in increasing the light utilization efficiency and the overall productivity in mass cultures (Polle et al., 2002, 2003), but suspensions have not

established anaerobiosis and so have failed to produce H2 gas under sulfur-deprived conditions. Despite this, nutrient-deprived mutants with truncated chlorophyll antennae produced H2 after immobilization within thin alginate films (Table 21.1). These mutants showed higher efficiency of H2 photoproduction than the parental CC-425 strain under saturating light conditions (Kosourov et al., 2011).

H2 photoproduction in green algae competes with a number of different metabolic pathways for the reduc — tant originated in photosynthesis (Hemschemeier and Happe, 2011). Here, CO2 fixation is one of the most important. The affinity of Fd to FNR is very high and on the order of 0.6 gM (Kurisu et al., 2005), while the af­finity of Fd to hydrogenase enzyme is only about 10 gM (Roessler and Lien, 1984). It is clear that electrons in healthy algal cells will be preferably directed toward reduction of NADP+ and, hence, toward CO2 fixation. Sulfur-deprived algae, however, inactivate Rubisco, the key enzyme of CO2 fixation, by the time of the establish­ment of anaerobiosis in the photobioreactor (Zhang et al., 2002). Zhang and coauthors showed that only about 3% of this protein is present in cells during the H2 production stage. This finding suggests that the photosynthetically generated reductants in sulfur — deprived algae are preferably used for generation of H2, but not for CO2 fixation. It is important to note here, that according to Hemschemeier et al. (2008) the Rubisco-deficient C. reinhardtii, CC-2803 strain produces H2 gas even under sulfur-replete conditions (Table 21.1). H2 evolution in this strain is almost completely depen­dent on electron flow from PSII. This finding shows that flow of electrons in engineered green algae can be successfully redirected toward H2 photoproduction. The competition for the photosynthetically generated reductant from other metabolic pathways is less studied. There has been some evidence for the competition from nitrate reductase (Aparicio et al., 1985). However, the wild-type CC-124 and 137C strains of C. reinhardtii that are commonly used in sulfur-deprivation experiments (Table 21.1) carry the nitl and nit2 mutations and cannot grow on nitrate. Therefore, the question about possible competition for the reductant between hydrogenase and nitrate reductase should be studied in detail using the Sager’s line of C. reinhardtii wild-type strains (Praschold et al., 2005).

A novel approach for preventing competition for the reductant from other metabolic pathways is formation of a fused complex of Fd and hydrogenase. In vitro analysis of such a complex showed that replacing the hydroge — nase with the Fd/hydrogenase fusion switches the bias of electron transfer from FNR to hydrogenase and re­sults in an increased rate of H2 photoproduction (Yacoby et al., 2011). This experiment indicates that the idea of the formation of a fused Fd/hydrogenase complex is promising, but should be checked in the C. reinhardtii mutant in vivo.

Another barrier for the industrial H2 photoproduc­tion system involves the redirection of photosynthetic electron flow from linear to cyclic and production of ATP, which results in a nonproductive pathway and decreased H2 production under anaerobic conditions. In green algae, this process also involves phosphoryla­tion and dissociation of PSII-light-harvesting antenna and results in the so-called state 1 to state 2 transitions that lead to higher excitation of PSI over PSII. A prom­ising approach for prolongation of H2 production in algae has recently been proposed by Kruse et al.

(2005) . They generated the mutants affected in state tran­sition. These mutants are blocked in state 1 that inhibits cyclic electron flow around PSI. One of these mutants, stm6, accumulated larger starch reserves under sulfur deprivation and produced almost five times more H2 gas than the wild type (Table 21.1).

H2 photoproduction in green algae is driven by the bidirectional [Fe—Fe]-hydrogenase enzyme that cata­lyzes not only the forward (H2 photoproduction) but also the reverse (H2 uptake) reaction. Under high H2 partial pressure in the photobioreactor, the rate of reverse reaction is significant (Kosourov et al., 2012). The authors showed that the decrease in H2 partial pres­sure improves significantly the yields and rates of H2 photoproduction in algal cultures. They also suggested the existence in sulfur-deprived algae of H2-uptaking pathways, either photoreduction or oxy-hydrogen reac­tion. The possibility of photoreduction in nutrient — deprived algae is questionable because of the significant degradation of the Rubisco enzyme by the time H2 photoproduction begins (Zhang et al., 2002). Therefore, it is most likely that nutrient-deprived algae utilize H2 gas through the indirect oxy-hydrogen reaction involving the chlororespiration pathway from [Fe—Fe]- hydrogenase(s) to O2 through Fd, NADP+/NADPH and the PQ-pool. If H2-uptaking pathway(s) does exist, H2 photoproduction in green algae can be further improved by downregulating this pathway(s).

Acknowledgments

This work was financially supported by the Academy of Finland Center of Excellence project (118637) and by the Kone foundation (YA, SNK).

Bioethanol is usually produced out of organic-based matter with high contents of sugar fermentation by enzymes produced from yeast. The yeast converts
six-carbon sugars (mainly glucose) to ethanol, because starch is much easier than cellulose to convert to glucose (Nigam and Singh, 2011). Bioethanol is produced simi­larly to other alcohols such as spirits using natural prod­ucts like wheat, maize and sugar beet. Hence, the suitable raw materials required for bioethanol produc­tion could be any of those materials that contain consid­erable amounts of carbohydrates to provide fermentable sugars for bioconversion into bioethanol. Then an opti­mized microbial fermentation process can be used for the bioconversion of sugars released from carbohydrates into ethanol (Nigam and Singh, 2011).

Agricultural waste materials are inexpensively found outside the human food chain in large amounts and can be obtained throughout the year. These agricultural bio­masses are the potential feedstocks for bioethanol pro­duction, including the cellulosic biomass, as well as starchy waste agricultural materials, and they provide low-cost and uniquely sustainable resources, improve­ment on energy security, development of the economy, as well as cleaning the environment and atmosphere by the disposing of problematic solid wastes and getting wealth out of wastes. Synthetically, 7% ethanol can be made from petroleum resources and 93% ethanol through fermentation process using microorganisms to convert biomass materials into ethanol (Kahn et al., 2011).

Because of the overwhelming complexity of cell wall polymers in terms of their chemical compositions, linkages and structures, plant biomass formation and microbial degradation involve a surprisingly large num­ber of genes in plant and microbes, respectively. For example, presumably every different glycosidic bond in the polysaccharides will be formed using a different enzyme. It is estimated that ~ 10% of genes in Arabidopsis genome are involved in cell wall synthesis and modifi­cation (Yong et al., 2005), which account for ~2000 genes encoding enzymes for sugar and lignin precursor synthesis, polysaccharide and lignin synthesis and modification, lignin-polysaccharide cross-linking, tran­scription factors (TFs)and signaling proteins, etc.

The most important enzymes are clearly those involved in polysaccharide synthesis and lignin synthe­sis. To form polysaccharides, glycosyltransferases (GTs) take the activated sugar donors, nucleoside diphosphate sugars (NDP-sugars), as the substrates to build glyco — sidic bonds between two sugars. Except for celluloses, other cell wall polysaccharides are mostly synthesized in Golgi apparatus, where GTs, NDP-sugar biosynthetic enzymes and sugar transporters are located and work together. Glycoside hydrolases (GHs), on the other hand, are used to break glycosidic bonds through hydro­lysis reactions to release sugars from polysaccharides. In plants this is often used to modify existing polysaccha­rides, e. g. when plant cells are growing, while in mi­crobes GHs are the most critical enzymes degrading plant biomass. Clearly not all GH and GT enzymes are involved in cell wall polysaccharide metabolism, as many of them are involved in metabolism of storage polysaccharides, glycoproteins, glycolipids and other glycol-conjugates that are not relevant to plant cell walls.

All GT and GH enzymes are categorized by the CAZy (Carbohydrate Active enZyme) database (CAZyDB) (Cantarel et al., 2009), which provides a general classifi­cation scheme for all carbohydrate active enzymes (CAZymes) and is widely accepted by the carbohydrate research community. So far there are a limited number of enzymes biochemically or genetically characterized to be involved in plant cell wall synthesis or modification, many of which belong to some large GH and GT fam­ilies. For example, the GT2 family is known to include cellulose synthases and some hemicellulose backbone synthases (Lerouxel et al., 2006), such as mannan syn­thases (Dhugga et al., 2004; Liepman et al., 2005), puta­tive xyloglucan synthases (Cocuron et al., 2007), and mixed linkage glucan synthases (Burton et al., 2006). With respect to the synthesis of xylan, the most abun­dant hemicellulose, proteins of GT43, GT47 and GT8 are likely to be involved (Zhong et al., 2005; Brown et al., 2007; Lee et al., 2007; Pena et al., 2007; Persson et al., 2007; York and O’Neill, 2008; Brown et al., 2009; Wu et al., 2009).

Some of these known cell wall-related CAZyme fam­ilies are included in Purdue’s Cell Wall genomics data­base (Yong et al., 2005) and UC-Riverside’s (UCR) Cell Wall Navigator database (Girke et al., 2004), and more families are discussed in the literature or to be character­ized in terms of their roles in biomass-related polysac­charide formation and degradation. For example, a few recent papers (Scheller and Ulvskov, 2010; Driouich et al., 2012) and a book (Ulvskov, 2011) updated our knowledge about the GT family members involved in cell wall synthesis: GT2, 8, 31, 34, 37, 43, 47, 61, 64, 75, 77, while there must be more GT families not included and to be identified as cell wall related (CWR), e. g. GT92 (Liwanag et al., 2012).

Lignins are complex heterogeneous polymers with lots of aromatic rings. The monolignol synthesis pathway that starts from phenylalanine to synthesize G, S and H units has been relatively well known, with about 10 gene families characterized encoding most of the enzymes in the pathway (Humphreys and Chapple, 2002; Boerjan et al., 2003; Vanholme et al., 2008; Xu et al., 2009; Zhong and Ye, 2009; Li and Chapple, 2010; Weng and Chapple, 2010; Carpita, 2012). All these lignin synthesis-related enzymes have been extensively reviewed in the literature and are included in Purdue’s Cell Wall genomics database. Transporting the units to the outside of the cell and assembling them into lignin polymers are less understood but some candidate trans­porters and two major enzyme families, peroxidase and laccase, are suggested (McCaig et al., 2005; Liu et al., 2011; Zhang et al., 2011; Alejandro et al., 2012; Carpita, 2012; Handford et al., 2012; Sibout and Hofte, 2012).

As with all other metabolic pathways, biomass forma­tion and degradation are also under strict regulation. However, compared to enzymatic activities, regulatory mechanism is even more difficult to elucidate. In plants, only a handful of TFs are known to regulate cell wall synthesis. The most studied process is the regulation of lignin biosynthesis (Zhong and Ye, 2009; Zhong et al., 2010; Zhao and Dixon, 2011; Wang and Dixon,

2012) . TF families NAC, WRKY, and MYB among a few others have been shown to directly or indirectly con­trol the monolignol synthesis. Some of the TF family members are global regulators that regulate the entire secondary cell wall synthesis, including the synthesis of celluloses and xylans, suggesting that the different biopolymers in biomass are not synthesized indepen­dently but in a coordinated way. On the other hand, genetically modifying the regulation of cell wall biosyn­thesis represents a very promising way to improve the desired traits of bioenergy crops. For example, Wang et al. showed that a mutation found in a WRKY TF could rewire the regulatory network of secondary cell wall synthesis and improve 50% of the biomass production in Arabidopsis (Wang et al., 2010). Similarly, micro ribo­nucleic acids (miRNAs) are also excellent targets for controlling the regulation of cell wall synthesis (Fu et al., 2012), which is less discussed in the literature. Clearly looking for novel transcription regulators, either TFs and miRNAs, and further building the regulatory network of cell wall synthesis is the ultimate goal for the elucidation of the mechanism of biomass formation.

Recently a few plant journals published special issues on plant cell wall researches: Plant Physiology (McCann and Rose, 2010), Current Opinion in Plant Biology (Pauly and Keegstra, 2008b), Frontiers in Plant Science (Debolt and Estevez, 2012) and Molecular Plant (has a cell wall biology category). Particularly, a number of review articles published in these special issues and a few book chapters (Table 6.1) gave overviews of latest progress in a specific area of cell wall research and are very useful for pointing to the original research papers reporting the characterization of specific CWR genes.

In terms of degradation, cell wall polysaccharides are degraded by microbial GHs and other CAZymes that are defined and categorized in CAZyDB. Lignins are mostly degraded by microbes too particularly by certain fungi (Dashtban et al., 2009). Enzymes involved in the degradation include fungal laccases and peroxidases, which are categorized in the FOLy (fungal oxidative enzymes) database (Levasseur et al., 2008). Note that these two families are not restricted to fungi. Instead they both belong to large protein families having many homologs in various organisms such as plants, animals and bacteria, bearing slightly different biochemical activities (Welinder, 1992). As mentioned above, these enzyme families are also used for lignin polymerization

in plants. There is also increasing evidence to show that such enzymes are also used for lignin degradation in bacteria (Claus, 2003; Li et al., 2009; Bugg et al., 2011a, 2011b).

Notably many cell wall biosynthesis-related gene families are also rooted in bacteria (Royo et al., 2000; Nobles and Brown, 2004; Emiliani et al., 2009; Yin et al., 2009; Weng and Chapple, 2010; Yin et al., 2010, 2011; Popper et al., 2011). In other words, although car­bohydrate and lignin-rich plant cell walls are almost unique to plants, the biosynthetic machinery has evolved from ancient gene families that were already present in early prokaryotes. On the degradation side, microbes are responsible for breaking down biomass, while plants also contain homologs of many microbial degrading enzymes such GHs and peroxidases. Obvi­ously plants also inherited these enzymes for different purposes: modify existing polysaccharide or complete the lignification process.

Similarly, less is known about the regulation of enzymes for the polysaccharide and lignin degradation in microbes than the enzymes themselves. As opposed to plants, microbes involved in biomass degradation are more taxonomically distributed spanning from eukaryotic fungi (e. g. Neurospora crassa) to prokaryotic bacteria (Clostridium thermocellum). As a result, the regu­lation systems in these divergent organisms are often not very conserved, e. g. many of the TFs found in fungi are not present in bacteria and vice versa. Furthermore, there are numerous model microbes used for bioenergy research and the transcription regulators regulating cellulases, hemicellulases or ligninases are highly dispersed in the literature, e. g. (Aro et al., 2005; Portnoy et al., 2011; Coradetti et al., 2012; Sun et al., 2012). All these make the curation and annotation of the regulators and targeting cis elements to be very difficult. Recently, global gene expression data (e. g. microarray) and other omics data have been generated to help study the regu­lation of biomass degradation (Nataf et al., 2010; Raman et al., 2011; Riederer et al., 2011; Yang et al., 2012), which represents the future trend of understanding biofuel production at the systems biology level. Similar to regu­lators for cell wall synthesis, there is a lack of web-based bioinformatics databases to include the regulatory genes for bioenergy-related degradation enzymes.

INTRODUCTION

First-generation, or conventional, biofuels are derived from sugars, starches, or vegetable oils from traditional agricultural crops and waste oils. Given first — generation biofuels’ impact on agricultural crop demand and prices, alternative feedstocks have been sought out. Microalgae have since been identified as a viable second-generation biofuels feedstock
(Figure 10.1). The advantages of using microalgae for biofuel production in comparison with other available feedstocks have been extensively reported.

There are an estimated 100,000 microalgae species, each with specific properties that allow them to exist in nearly every environment on Earth, including arid cli­mates that do not sustain most agricultural crops. There­fore, microalgal production systems need not displace other traditional land-based crops intended for human

Bioenergy Research: Advances and Applications http://dx. doi. org/10.1016/B978-0-444-59561-4.00010-3

FIGURE 10.1 The progression from first — to second-generation biofuels. (For color version of this figure, the reader is referred to the online version of this book.)

or livestock consumption, which in turn greatly reduces the impact to the food distribution chain. Further, micro­algae may be harvested multiple times a year, which greatly increases yearly production yields. The cultiva­tion of microalgae for biofuels production can also be coupled with other beneficial production schemes to improve net income and positively address environ­mental concerns. Some possibilities currently being investigated include the following:

• Reclamation of nutrients such as NH^, NO^, POfi, and others from wastewater, which reduces costs associated with cultivating the algae and treating wastewater (Zhu, 2013; Batten, 2013).

• Utilization of waste CO2 from industrial flue gases, which reduces greenhouse gas emissions while producing biofuel (Gonzalez-Lopez, 2012).

• Cultivation and extraction of value-added metabolites within microalgae intended for biofuel production. In this scenario, the value-added metabolite is extracted prior to, or during, the biofuel production stream. Commercially relevant products include a large range of fine chemicals and bulk products, such as polyunsaturated omega fatty acids, antioxidants, high-value bioactive compounds, natural dyes, sugars, and proteins (Mimouni, 2012; Skjanes et al., 2013).

• After oil and target metabolite extraction, the processed algal biomass can be used as a nutrient-rich livestock feed, or used as sustainable organic fertilizer due to its high N:P ratio (Mulbry, 2005; Stamey, 2012).

Because of this variety of value-added biological de­rivatives, coupled with environmental sustaining

strategies, microalgae intended for biofuel production can potentially revolutionize a large number of biotech­nology areas concurrently, including pharmaceuticals, cosmetics, nutrition and food additives, aquaculture, and pollution prevention.

MICROALGAL BIOFUEL HISTORY

Beginning in the 1950s, Golueke et al. (1957) conduct­ed early work on the anaerobic digestion of microalgal biomass for the production of methane fuel. The energy crisis in 1973 prompted the formation of The National Renewable Energy Laboratory (NREL) under the Jimmy Carter Administration. From 1978 to 1996, NREL con­ducted the most authoritative study to date on the devel­opment of biofuels from algae (Sheehan et al., 1998). The study concluded that under controlled conditions, algae are capable of producing 40 times the amount of oil for biodiesel per unit area of land when compared to terres­trial oilseed crops such as soy and canola, and that the use of wastewater as a nutrient source for algae propa­gation was the most practical approach for near-term production of algal biodiesel (Sheehan et al., 1998; Oswald, 2003). Despite the promise of cost-effective fuel production from microalgae, interest in renewable energy quickly waned as the energy crisis subsided and fuel prices fell. The recent world-wide escalation in oil prices has renewed interest in microalgae as a bio­fuels feedstock.

Since the original NREL study, other groups also have conducted analyses of full-scale algae-to-biodiesel pro­duction (Benemann et al., 1982; Weissman and Goebel,

1987; Beal, 2012a). Although these and other studies have indicated a great potential for profitable biofuel from microalgae, they also highlighted the need for system im­provements, in both cultivation management and pro­cessing schemes to improve yields and reduce costs in order to be competitive with fossil fuels. For example, even when robust algae growth was achieved, inefficient processing techniques such as biomass centrifugation and drying followed by solvent extraction made recovery of biofuels cost-prohibitive. To overcome this barrier, changes to the system have been introduced, including processing techniques that eliminate the need for expen­sive dewatering regimens such as centrifugation and dry­ing of the harvested biomass prior to oil extraction with solvents. One suggested path forward is a solventless wet stream process whereby microalgae are concentrated using pH-driven flocculation using inexpensive lime, fol­lowed by rupturing of the cells by pulsed electric field, and ultimate recovery of released lipids by cross-flow filtration. When coupled with waste streams for CO2 and nutrients, this process has a positive return on invest­ment (Beal, 2012b). Another suggested path forward to­ward practical biofuel extraction from microalgae is the use of hydrothermal liquefaction (HTL) processing. This method eliminates the need for solvents to break open algae cells, instead relying on heat and pressure to remove the water from the biomass. An ancillary benefit of the HTL method is that in addition to lipids, other organic metabolites such as carbohydrates, proteins, and nucleic acids can likewise be converted to biocrude during the HTL process. Thus, a cultivation strategy needs only to focus on the production of biomass rather than inducing the accumulation of lipids at the expense of cellular proliferation. Ultimately, cultivation and pro­cessing strategies should be firmly supported by real­time analysis of fuel precursors such as lipids that can be converted to biodiesel, carbohydrates that can be con­verted to bioethanol, and the organic biomass that can be converted to biocrude. Detailed analytical feedback is necessary to optimize growth conditions to maximize specific biofuel precursors.

Following extraction from biomass, biolipids can be used as pure oil (generally plant) or can be converted to biodiesel by a process known as transesterification, described later. However, the use of PPO as a fuel re­quires the modification of diesel engines unlike bio­diesel, which, particularly when blended with petroleum diesel, can be used in unmodified diesel engines. These engine modifications are needed as PPO is more than 10 times as viscous as biodiesel. As a result, it has a tendency to gum up in cold weather, which can be somewhat overcome by blending with traditional fossil diesel. Nevertheless, it has some advantages: with a flash point of over 300 °C, storage and transport are simplified. According to the VwVwS (Verwaltungsvorschrift wassergefahrdende Stoffe), which is the national German regulation on water hazard classification, PPO is not designated as even a hazard to water given that it is biodegradable. In an unmodified engine, poor atomization of the fuel will lead to coking of the injectors and accumulation of soot deposits. Modification is designed to preheat fuel or involves installation of a two-tank system. In the latter, the engine is started with diesel and only changes to PPO when the operating temperature has been reached. It must switch back to diesel before being turned off, to flush out the remainder of the PPO in order to ready the engine for the next operation. Other options exist, such as the specialist engine developed by Ludwig Elsbett in the 1970s. The fuel emissions of PPO are also much lower in sulfur emissions when compared to the fossil equivalent. For a detailed over­view see (Russo et al., 2012). After extraction, if the biolipid is not to be used as PPO, or other pure oil, it needs to be further processed into a more useable bio­fuel, usually biodiesel. Here the biolipid goes through a series of processing steps beginning with degumming.

Just a small number of papers have been devoted so far to the use of acid-activated clay minerals in the cata­lytic esterification of fatty acids (Vijayakumar et al., 2005;

Konwar et al., 2008; Nascimento et al., 2011; Neji et al., 2011; Zatta et al., 2012,2013; Rezende et al., 2012; Olowo — kere et al., 2012), as well as in the transesterification of oils and fats (Bokade and Yadav, 2009). Some specific cases will be evaluated in sequence.

Case 1 (Zatta et at., 2012)

Standard Texas Montmorillonite STx-1 with the chemical formula ((Ca0.27Na0.04K0.01)[Al2.41Fe(III)0.09

Mg0.71Ti0.03]Si8.00O20(OH)4), supplied by the Clay Min­eral Society repository, was activated using phosphoric, nitric and sulfuric acids under different conditions of temperature, time and acid concentrations and the resulting materials were characterized by X-ray diffrac­tion (XRD), nitrogen adsorption isotherms and Fourier transform infrared spectroscopy. Also, the presence of Lewis and Bronsted acid sites in the structure of the cata­lyst was characterized by pyridine adsorption. After­ward, the materials were evaluated as catalysts in the methyl esterification of lauric acid. Blank reactions car­ried out in the absence of any added catalyst presented conversions of 32.64, 69.79 and 79.23% for alcohol:lauric acid molar ratios of 60:1,12:1 and 6:1, respectively. In the presence of the untreated clay and using molar ratios of 12:1 and 6:1 with 12 wt% of catalyst, conversions of 70.92 and 82.30% were obtained, respectively. For some key samples obtained by the acid activation, conversions up to 93.08% of lauric acid to methyl laurate were ob­tained, much higher than those observed for the thermal conversion (TC) or using raw montmorillonite. Relative good correlations were observed between the catalytic activity and the development of acid sites and structural and textural properties of the acid-leached materials.

Case 2 (Zatta et at., 2013)

The same sample of montmorillonite STx-1 described above was submitted to acid activation using aqueous solutions of phosphoric acid. The acid treatment was carried out under vigorous stirring at 100 °C in a flat- bottomed flask connected to a reflux condenser and a heating mantle. The mineral clay and the acid solution were mixed in a 1:4 ratio (mass per volume) using acid concentrations of 0.5, 1, 2 and 4 mol/l.

After the acid activation process, the samples were repeatedly washed with distilled water until pH close to 7, dried at 110 °C for 24 h and then heated in an oven at 250 °C for 2 h. To check the influence of other acids in the activation of montmorillonite STx-1, this clay mineral was subjected to activation with hydrochlo­ric (37% proof), nitric (65% proof) and sulfuric (98% proof) acids. The resulting acid-activated clay materials were characterized and subsequently used in the cata­lytic conversion of lauric, oleic and stearic acids, as well as of a complex mixture of fatty acids (tall oil) to their corresponding fatty acid methyl esters (FAMEs).

The results obtained for the best phosphoric acid — activated sample (PA) were compared to those of the TC and from a standard commercial Lewis acid catalyst (K10) and raw montmorillonite (STX). In all experi­ments, conversion of all samples were carried out for 2 h at a methanol:fatty acid molar ratio of 12:1 and 160 °C with 12 wt% of the catalyst in relation to the oil mass.

In general, the PAs and the standard Lewis acid cata­lyst (K10, Sigma—Aldrich), which is produced by HCl activation of mineral clays at boiling temperatures, had similar catalytic activities. However, in some cases, the catalytic performance of PA was even better than that of K10 (Figure 16.4). These data were a strong indication that PA and K10 have similar chemical and physical characteristics, even though part of the layered structure is still retained in PA after acid activation.

Tests of reuse of the best PA were performed (Figure 16.5) and no significant losses of activity were observed during the first four consecutive reaction cy­cles (see dotted horizontal line in Figure 16.5). This is a very important observation since, from any practical use in industrial processes, the catalysts must last for long time before deactivation.

Depolymerization of lignin in sub — and supercritical water (pc > 22.1 MPa; Tc > 374 °C) lead to extensive lower molar mass fragments, dealkylation and deme — thoxylation, but a part of these fragments tend to cross-link in larger fragments. The economic viability of this process is severely controlled by the extent to which the heat is recovered from the effluents. The yield of monomers is positively correlated with base concen­tration added with maximum yield of one-third of the initial lignin. Low molecular weight fraction yields increased with longer reaction times in supercritical water without catalysts at 350—400 °C and 25—40 MPa. The water-soluble fraction consists of catechol (28%), phenol (7.5%), and cresols (11%), suggesting the cleav­age of both ether and carbon—carbon (Wahyudiono et al., 2008). Addition of phenolics (e. g. phenol and p-cresol Okuda et al., 2004a, b, 2008; Fang et al., 2008) gives a complete depolymerization of lignin into dimers without char formation. Phenol and p-cresol depressed

cross-linking reactions due to entrapment of reactive fragments, like formaldehyde, and capping of active sites like Ca in the lignin structure.

In addition to specific plants, staple crops also pro­vide phytochemicals at large scale, as coproducts of well-developed, comprehensive food production pro­cesses, which may be considered as the first generation of biorefineries.

Soybean processing involves multitier steps yielding multiple streams. The major soybean products include oil, feedstuff, and fermented soy food. Minor products include full-fat soy flours, soy concentrate, soy protein iso­lates, and lecithin. Phytochemicals that can be produced as coproducts from soybean processing include caroten­oids, isoflavones and saponin; protease inhibitors from protein fractions; as well as lecithin, phytosterols, and tocopherols from oil. Soybean processing in general consists of preparatory steps (cleaning, drying, mechanic disruption or grinding, or conditioning) and oil-extraction steps (mechanical pressing or solvent extraction, refining, bleaching, and hydrogenation). Coproducts are prepared by extractive distillation, adsorption, membrane filtration, and super — or subcritical fluid extraction (Kannan et al., 2012; Zijlstra et al., 2012) (Figure 20.5).

One of the two major corn processings is wet milling. Wet milling yields major products ranging from starch, starch-fermented ethanol (first-generation bioethanol), Wet or Dried Distillers Grains (residues from ethanol fermentation) or Dried Distillers Grains and Solubles or Stillage (WDDG or DDG, DDGS, which are widely used as feed), and steep liquor. Phytochemicals that might be generated as coproducts include flavonoids, phytosterols, carotenoids, polyamine-hydroxycinnamic acid amide conjugates (Rausch, 2012; Moreau et al., 2009) from steep liquor, steeped corn, oil-extracted resi­dues, stillage, or unfermented residues, although their production has not been widely integrated in current corn wet milling factories. Corn wet milling process include mechanic disruption, liquid extraction (steep­ing, acidic, basic or SO2 impregnation), screening, oil pressing, evaporation, centrifugation, fermentation and distillation. Corn dry milling is another major corn pro­cessing, mainly geared for bioethanol production. The process does not have steeping and germ-processing (oil extraction) as the wet milling does, resulting in sep­aration and enrichment of most corn phytochemicals in unfermented residues and stillage (Rausch, 2012; Rausch and Belyea, 2006) (Figure 20.6).

Vegetable oil production involves multitier, multi­phase steps (Figure 20.7). Major plant sources for vege­table oils include palm, soybean, rapeseed, sunflower seed, peanut, cotton seed, coconut, and olive, and (to less extent) corn, hazelnut, grape seed, sesame, flax seed, safflower, rice bran, etc. Besides oil and cake (or meal, oil-extracted residues), other coproducts come from mechanically or chemically separated substances prior to oil extraction as well as refining by-products. Degumming and deodorizing of crude vegetable oils result in the production of lecithin and tocopherols or phytosterols, respectively. General processes of vegetable oil production include feedstock disruption by mechanical, chemical or enzymatic means, mechani­cal pressing, phase separation, solvent extraction, and refining (degumming, neutralizing, bleaching and deo­dorizing) (Panpipat et al., 2012; Febrianto and Yang, 2011; Muth et al., 1998; Dunford, 2012).

Postharvest processing of wheat, rice, oat, or other cereals generates not only flour or milled rice, bran, and germ as primary products, but also gluten, fiber, bran oil, or other phytochemicals as secondary products. The processing mainly comprises different levels of mill­ing and fractionation (air classification, sieving, etc.), sometimes also with extraction (Kraus, 2006).

As with ethylene, isoprene is a medium-value biochemical that is produced through steam cracking of oil. It is actually an important by-product of ethylene production and is almost entirely used for production of a synthetic substitute for natural rubber. It is also natu­rally produced by many plants as a heat stress response, where it was shown to increase the stability of photosyn­thetic membranes at high temperatures (Sharkey et al.,

2001) . It can represent as much as 2% of all carbon fixed by oak leaves at a temperature of 30 °C (Sharkey, 1996), showing the physiological importance of this com­pound. The enzyme isoprene synthase (ispS) was shown to produce isoprene in plants, converting one of the products of the methylerythritol phosphate (MEP) pathway, dimethylallyl-diphosphate (DMADP), into isoprene (Silver and Fall, 1991; Silver and Fall, 1995). Prokaryotes were suggested to be able to produce isoprene after reports of the detection of this compound in the headspace of culture broth on many species (Kuzma et al., 1995), with emphasis on Bacillus subtilis. Not surprisingly, sequence analysis of bacterial genome could not identify any gene homologous to the ispS. found in plants (Withers et al., 2007). So far, functional genomics has also failed to identify the pathway for isoprene pro­duction in prokaryotes. Sequence-independent methods showed that 19,000 E. coli clones transformed with DNA fragments from B. subtilis in an environment where DMADP and IPP (isopentenyl pyrophosphate) levels were selectively toxic, showed that no single enzyme was sufficient to convert DMADP to isoprene, where the few clones that managed to survive, preferably converted it to a prenyl alcohol (Withers et al., 2007). As all isopre — noids are thought to be solely produced from DMADP and IPP (Xue and Ahring, 2011), the conversion of the metabolites involved in MEP or mevalonate pathway

Coa-dependent pathway

FIGURE 22.4 Pathway alternatives for n-butanol bioproduction. The alcohol n-butanol is naturally produced in different microorganisms in small quantities, where it can be synthesized either through the CoA-dependent pathway or the keto acids pathway. (For color version of this figure, the reader is referred to the online version of this book.)

to isoprene in bacteria could be a phenotype derived from convergent evolution using a multistep reaction diverged from those pathways (Izumikawa et al., 2010; Withers et al., 2007; Xue and Ahring, 2011).

Bioproduction of isoprene is feasible and has already been demonstrated in E. coli expressing heterologous ispS (Miller et al., 2001; Zhao et al., 2011). Of course pro­ductivity is an issue and different strategies were tried to increase isoprene production. Simultaneous expres­sion of heterologous enzymes involved in MEP or meval — onate pathways was shown to be effective in both cases (Yang et al., 2012; Zhao et al., 2011). Julsing et al. also showed that the individual expression of the genes encoding enzymes involved in the MEP pathway did not affect isoprene production with the exception of the dxs gene, encoding the enzyme that catalyzes the first re­action of the MEP pathway, which significantly improved isoprene production (Julsing et al., 2007). Cya­nobacteria produce DMAPP through the MEP pathway for secondary metabolites and, albeit with no natural pro­duction of isoprene being reported yet, transformation and expression of heterologous ispS were shown to be sufficient for production of isoprene. Lindberg et al. re­ported isoprene production using Synechocystis sp. PCC 6803 as a model organism harboring ispS from Pueraria montana (Kudzu) (Lindberg et al., 2010). The transgene was inserted at the psbA2 locus and mutants did not
show any disturbance in growth when compared to the wild type. This was a well-achieved proof of concept, and the low productivity reported, 50 mg per gram of dry cell weight per day, can be much improved through metabolic engineering. However, the use of cyanobacte­ria to produce isoprene has issues different from meta­bolic yield: to develop a production system of a molecule with a half-life of only a couple hours in the presence of light is particularly challenging in a photo­synthetic organism. To overcome this issue, the develop­ment of special photobioreactors is made in parallel to the molecular research, where the properties of isoprene as a volatile hydrophobic compound, easily separated from a culture broth and concentrating in the headspace, are exploited (Lindblad et al., 2012). The production of a gas in microorganisms is an interesting strategy because one does not need to harvest the cells, the product is concentrated in the gaseous phase of the culture. Howev­er, the cultivation techniques and the purification of this gas from a complex mixture represents an important step in the production chain and, as shown in this case, should develop together.

The objective of steam pretreatment, steam explo­sion or liquid hot water, is to solubilize the hemicellu — lose to make the cellulose better accessible for enzymatic hydrolysis and to avoid the formation of inhibitors (Hendriks and Zeeman, 2009). During steam pretreatment parts of the hemicellulose hydrolyze and form acids, which could catalyze the further hydrolysis of the hemicellulose. To avoid the formation of inhibi­tors, the pH should be kept between 4 and 7 during the pretreatment (Hendriks and Zeeman, 2009). The aqueous fractionation of native lignocellulosic mate­rials with hot, compressed water (also known as hydro­thermal processing or autohydrolysis) has been proposed as a fractionation method for biorefineries, as it enables the simultaneous removal of water — soluble extractives and the solubilization of hemicellu — loses, yielding a solid phase enriched in lignin and cellulose (Gullon et al., 2012). Liquid hot water has the major advantage that solubilized hemicellulose and lignin products are present in lower concentra­tions, when compared to steam pretreatment, due to higher water input. These lower concentrations reduce the risk on degradation products like furfural and the condensation and precipitation of lignin compounds (Hendriks and Zeeman, 2009).

Wet oxidation is another oxidative pretreatment method, which uses oxygen as oxidation agent. The sol­uble sugars produced during wet-oxidation pretreat­ment are mainly polymers opposite to the monomers produced during steaming or acid hydrolysis as pre­treatment. Phenolic monomers are no end products during wet oxidation but are further degraded to car­boxylic acids (Hendriks and Zeeman, 2009; Martin and Thomsen, 2007).

Carbon dioxide pretreatment is conducted with high — pressure carbon dioxide at high temperatures of up to 200 °C with duration of several minutes. Explosive steam pretreatment with high-pressure carbon dioxide causes the liquid to be acidic and this acid hydrolyses especially the hemicellulose. Carbon dioxide is also applied as supercritical carbon dioxide (35 °C, 73 bars) for depolymerization of the sugars present in biomass, increasing the glucose yield probably caused by increase in pore size (Hendriks and Zeeman, 2009). This method is considered as a "green" pretreatment because it does not require neutralization or pH adjustment prior to enzymatic hydrolysis (King et al., 2012).

Ammonia fiber explosion (AFEX), ammonia recycled percolation (ARP) and soaking aqueous ammonia (SAA) are alkaline pretreatment methods that use liquid ammonia to pretreat biomass. The difference between AFEX and ARP processes is that the first is carried out in liquid ammonia and the second one in an aqueous ammonia solution.

AFEX is a physicochemical pretreatment process in which lignocellulosic biomass is exposed to liquid ammonia at high temperature and pressure for a period of time, and then the pressure is suddenly reduced (Kumar et al., 2009). The AFEX pretreatment simulta­neously reduces lignin content and removes some hemi — cellulose while decrystallizing cellulose and it has the advantage of ammonia being recyclable due to its high volatility (Yang and Wyman, 2008). AFEX has been shown to complete conversion of cellulose to ferment­able sugars but removes or loses little lignin or hemicel — lulose. In a typical AFEX process, the dosage of liquid ammonia is 1—2 kg of ammonia/kg of dry biomass, the temperature is 90 °C, and the residence time is 30 min (Kumar et al., 2009). However, AFEX pretreat­ment at 40 ° C and longer residence times, up to 8 h, has also been proposed with comparable yields of sugar and ethanol (Bals et al., 2012).

AFEX treatment is a batch process while continuous processing in an extruder is an approach called FIBEX (fiber extrusion) that significantly reduces both the time required for treatment and the ammonia levels required with similar hydrolysis results to those for AFEX (Yang and Wyman, 2008).

ARP is another process based on ammonia, which recycles aqueous ammonia solution (5—15 wt%) through a reactor packed with biomass at elevated temperatures (80—180 °C). Ammonia in aqueous solution and at high temperature breaks down lignin via the ammoniolysis reaction but has virtually no effect on carbohydrates (Geddes et al., 2011). A major challenge for ARP is to reduce liquid loadings to keep energy costs low (Yang and Wyman, 2008). SAA is a modified version of AFEX but it uses moderate temperatures (25—60 °C) to reduce the liquid amount during pretreatment. At ambient temperatures the duration could be up to

10— 60 days while at higher temperatures (150—190 °C) the duration of pretreatment is reduced to minutes (Agbor et al., 2011). The cost of ammonia, and especially of ammonia recovery, drives the cost of the ammonia — related pretreatments (Kumar et al., 2009).