Various processes generating transport biofuels start from complex organic feed­stocks, including complete organisms or large chunks thereof, or from wastes such as sewage sludge, black liquor or household waste. The focus of such processes is often on lignocellulose. This is not surprising as the share of lignocellulose in all biomass has been estimated at about 50% (Claassen et al. 1999). Lignocellulose is a structural material of plants and a composite of lignin (a polymer composed of monolignols), cellulose (a glucose polymer) and hemicellulose (a polymer made up of diverse hexose and pentose sugars). The US Energy Law of 2008 stipulates that from 2016, transport bioethanol producers must switch to lignocellulosic feed­stocks.

There are a wide variety of lignocellulosic feedstocks. Wood, wood waste, har­vest residues, a variety of wastes and by-products originating in industries are rela­tively rich in lignocellulose (Prasad et al. 2007). It also has been suggested that nat­ural grasslands can be exploited as a source for lignocellulosic feedstock (Tilman et al. 2006; Zhou et al. 2008). It is furthermore possible to grow lignocellulosic crops on plantations. Examples of species that are considered for this purpose are woody perennials, such as eucalyptus, poplar, willow and black locust, and grasses and other non-woody perennials, such as switchgrass, elephant grass, reed canary grass, Miscanthus, cardoon, reeds and Bermuda grass. Of these species, Bermuda grass and reed canary grass are currently used as forage for livestock (Boateng et al. 2007; Pahkala et al. 2008).

The relative amount of lignin in lignocellulose is source dependent. In nutshells, the percentage of lignin may be 30-40% and in rice straw about 5.5% (Prasad et al.

2007) . Similarly, the composition of hemicellulose is source dependent. For in­stance, hemicellulose from agricultural residues or hardwood tends to be rich in pentose sugars, whereas such sugars are a minor component in hemicellulose from softwood (Hahn-Hagerdal et al. 2007).

Apart from lignocellulose, the lignocellulosic feedstocks also contain a variety of other compounds, both organic and inorganic in character. The latter, a for­tiori, holds for complex wastes such as sludges from wastewater treatment plants or household wastes, which have been proposed as sources of transport biofuel (Ptasin — ski et al. 2002). There is also a relatively limited supply of cellulosic wastes that may be used for biofuel production, such as sludges from (virgin) paper produc­tion and paper recycling (Mabee and Roy 2003; Prasad et al. 2007; Marques et al.

2008) .

There are several ways to generate substances that may serve as automotive bio­fuels from complex organic feedstocks. A first possibility, which applies both to biomass in general and to lignocellulose, is heating (‘thermochemical treatment’) to produce liquid biofuels. An option which has been exploited for centuries is the dry distillation or slow pyrolysis of wood. Apart from charcoal, methanol is an output (approximately 1-2% by weight) of slow pyrolysis of wood, which can in principle be used as a transport fuel (Reinharz 1985; Demirba§ 2001; Gullu and Demirba§ 2001; Huber et al. 2006). More recently, much attention has been given to fast and flash pyrolysis of biomass (Goyal et al. 2008). Fast and flash pyrolysis of biomass in principle produces charcoal or biochar, gas, organic fluids and water. The precise nature of the products and the relative shares of the different components can be varied, dependent on the character of the biomass, the presence of inorganic sub­stances (especially metals), reactor design, temperature, heating rate, catalysts and reaction time (Bridgwater et al. 1999; Yang et al. 2004; Demiral and §ensoz 2006; Huber et al. 2006; Boateng et al. 2007; Dobele et al. 2007; Lange 2007; Muller — Hagedorn and Bockhorn 2007; Demirba§ 2008; Di Blasi 2008; Fahmi et al. 2008; Ros et al. 2009). The fluid produced by fast and flash pyrolysis contains water and a variety of organic compounds, the latter collectively called ‘pyrolysis oil’. The pyrolysis oil tends to be unstable and to show polymerization reactions (Fahmi et al. 2008). It needs upgrading to serve as a basis for transportation fuel, for example, by hydrodeoxygenation, hydrogenation or treatment with zeolites (Huber et al. 2006; Esler 2007; Wang et al. 2008b). Such upgrading has proven difficult, and this has re­stricted the application of biomass pyrolysis technology (Wang et al. 2008b). It has also been proposed to view the pyrolysis oil as a basis for a biorefinery generating a number of chemicals besides transport fuel (Hayes 2008). As an alternative to fast and flash pyrolysis, a process has been proposed that combines pyrolysis with hy — drogenationby a formic acid-alcohol mixture (Kleinert and Barth 2008). Also under development is deoxy-liquefaction (Goyal et al. 2008; Wang et al. 2008), converting lignocellulosic biomass into a liquid that tends to be richer in hydrocarbons than the liquids commonly produced by fast pyrolysis.

A second possibility to deal with complex organic feedstocks is based on gasifi­cation of biomass or lignocellulosic materials resulting in the formation of synthesis gas (containing relatively high percentages of CO and H2). The formation of tar, and to a lesser extent char, and ash-related problems have emerged as problems in such gasification, necessitating major efforts in the field of optimizing gasification, tar reforming and syngas quality control and clean-up (Wang et al. 2008b). Using the water shift reaction, the amount of H2 in synthesis gas may be maximized, and H2 and the other main product of the water shift reaction (CO2) can be separated by processes such as pressure swing adsorption, membrane separation and cryogenic separation (Ferreira-Aparicio et al. 2005; Andersson and Harvey 2006; Haryanto et al. 2007; Barelli et al. 2008; Florin and Harris 2008; Wang et al. 2008b). It is also possible to subject syngas (after clean-up) to catalytic methanation, generating synthetic natural gas (Felder and Dones 2007).

Alternatively, conversion is possible into liquids (biomass-to-liquids or BTL biofuels). One option is the use of synthesis gas to produce oxygenates such as methanol (Reed and Lerner 1973; Demirba§ 2001; Ptasinski et al. 2002) and dimethylether (Joelsson and Gustavsson 2008). Producing ethanol from syngas is also possible but is as yet not very efficient (Subramani and Gangwal 2008). Still another option is to use the Fischer-Tropsch reaction, after enrichment of syngas with hydrogen, to generate hydrocarbons (Dietenberger and Anderson 2007), or the methanol-to-synfuel synthesis to produce hydrocarbons (Takeshita and Yamaji 2008). The latter can be conveniently applied in diesel or Otto motors (Reinhardt et al. 2006) or in airplanes (Esler 2007). There are also bacteria that can convert synthesis gas into ethanol, and these are currently researched for use in biofuel production (Henstra et al. 2007; Tollefson 2008). Low conversion rates, product inhibition and problems in maintaining optimum conditions have for a substantial time prevented commercialization of this approach (Wang et al. 2008b), but such problems have now apparently been solved to the extent that a pilot plant has been announced (Ashley 2008).

Thirdly, cellulose and hemicellulose present in lignocellulose may be enzymat­ically converted into ethanol or butanol, to be applied in, for example, Otto mo­tors (Sanchez and Cardona 2008; Qureshi et al. 2008a, b). This requires separating hemicellulose from lignin, hydrolysis of cellulose and hemicellulose into sugars and fermentation of the sugars generated by hydrolysis (Lynd 1996; Lachke 2002; Palmarola-Adrados et al. 2005; Gray et al. 2006; Angenent 2007; Prasad et al. 2007; Gomez et al. 2008; Sanchez and Cardona 2008; Qureshi et al. 2008a, b).

Hydrolysis of cellulose generates glucose, which can be converted into ethanol. Important among the hydrolytic products of hemicellulose is often xylose, a 5-car­bon sugar (Fortman et al. 2008). Xylose can be converted into ethanol by fermenta­tion as follows:

3D-xylose (C5H10O5) ^ 5ethanol + 5CO2 .

Micro-organisms such as Pichia stipitis and genetically modified Escherichia coli are able to perform the fermentation of xylose (Rubin 2008). Minor sugars originat­ing in cellulose and hemicellulose are arabinose, rhamnose, glucose, galactose and mannose, which can be converted into ethanol, too (Numan and Bhosle 2006; Fort — man et al. 2008; Hayes 2008). It is also possible to ferment C6 and C5 sugars into a mixture of acetone, butanol and ethanol (Jones and Woods 1986; Qureshi et al. 2008c). Process design tends to be focused on a limited number of lignocellulosic feedstocks for which the process is optimized (Olofsson et al. 2008). In practice, the separation of hemicellulose from lignin currently causes most problems, which are in part linked to the heterogeneous structure of lignin polymers (Gomez et al. 2008; Wackett 2008). Building cell walls involves many enzymes (McCann and Carpita 2008), and it may well be that a combination of enzymes may be necessary for their deconstruction in a way that is optimal for the next step of biofuel production: sac­charification. However, most processes currently studied for near-term application rely on the use of rather brute physico-chemical force to separate the constituents of lignocellulose (which negatively impacts overall energy efficiency and the envi­ronmental burden). Examples are: the use of acid (whether or not combined with ionic liquid), steam explosion (sometimes combined with oxidation), high-pressure hot water treatment, treatment with alkaline peroxides and ammonia fibre explosion (Huber et al. 2006; Gomez et al. 2008; Li et al. 2008; S0rensen et al. 2008; Qureshi et al. 2008c). The difficulty of separation varies for different plant species (Bura­nov and Mazza 2008). Coniferyl lignin appears, for instance, more recalcitrant so far against physico-chemical methods of separation than syringyl lignin (Anderson and Akin 2008). And the presence of oxidatively coupled esterified or etherified fer- ulic acid residues has also been reported to inhibit separation (McCann and Carpita 2008).

Proposals to overcome the hurdles to separation of lignin and cellulose and hemicellulose include: application of lignin-degrading white rot fungi of micro­organisms derived from termite guts, of Clostridium phytofermentans, and pre­treatment with phenolic esterases (Warnick et al. 2002; Anderson and Akin 2008; Rotman 2008; Weng et al. 2008). Also, it has been suggested to use lignases and to convert degraded lignin into transport biofuels (Blanch et al. 2008). Furthermore, there have been proposals to downregulate lignin biosynthesis in plants by genetic modification to ease the release of cellulose and hemicellulose and ultimately sug­ars from plants (Chapple et al. 2007; Wackett 2008). Such downregulationhas led to plant characteristics that are unsuitable for biofuel crops, such as increased suscep­tibility to fungi, dwarfing and the collapse of vessels in xylem (Weng et al. 2008). Dwarfing has been linked to the simultaneous inhibition of flavonoid production (McCann and Carpita 2008). There have been new proposals for genetic modifica­tion, focusing on changes in lignin polymer structure and monolignol polymeriza­tion (Weng et al. 2008), but it is as yet not clear whether this approach will lead to suitable biofuel feedstocks.

Degradation of hemicellulose may also be difficult. Hemicelluloses appear so far refractory against saccharification when esterified by ferulic or coumaric acids (Anderson and Akin 2008). And enzymatic hydrolysis of cellulose to fermentable sugars currently requires, per kg ethanol produced, 40-100 times more enzyme than the hydrolysis of starch (Eijsink et al. 2008). This has led to the proposal to include glycosyl hydrolases into plants by genetic modification (Taylor et al. 2008). Hy­drolysis and fermentation of cellulose and hemicellulose can be done in a two-step process (e. g. Zanichelli et al. 2007; Hayes 2008), with one hydrolytic and one fer­mentative step. When dilute acid is used for the pre-treatment of lignocellulosic biomass, there is often much hydrolysis of cellulose and hemicellulose. At higher temperatures, dilute acid treatment may also lead to much hydrolysis of cellulose (Hayes 2008). A variant of treatment with dilute acid may be used to generate sub­stantial amounts of the platform chemicals furfural and levulinic acid, in line with biorefinery concepts (Hayes 2008). Alternatively, hydrolytic enzymes produced by micro-organisms may be used (Lynd et al. 2002; Demain et al. 2005; Desvaux 2005).

The two steps in the conversion of cellulose and hemicellulose to ethanol may also be combined in a one-step process: simultaneous saccharification and fermen­tation (SSF). Simultaneous saccharification (hydrolysis of cellulose and hemicellu­lose giving rise to sugars) and fermentation by micro-organisms is often preferred as it is associated with shorter residence times and potentially higher yields and lower costs (Ballesteros et al. 2004; Demain etal. 2005; Huber etal. 2006; Angenent 2007; Marques et al. 2008). In simultaneous saccharification and fermentation, sacchari­fication is the rate-limiting step. Inhibition of fermentation by substances formed during pre-treatment and hydrolysis is a problem. Inhibitory compounds formed during pre-treatment and hydrolysis include salts, phenols, furfural, cinnamalde — hyde, ^-hydroxybenzaldehyde, lignin monomers and syringaldehyde (Zanichelli et al. 2007; Qureshi et al. 2008c; Sanchez and Cardona 2008; Royal Society 2008). The presence of inhibitors often necessitates the ‘detoxification’ by physical, chem­ical or biological methods. Another option is the use of fermenting organisms that are more tolerant to inhibitors (Hayes 2008; Olofsson et al. 2008).

Processes converting cellulose and hemicellulose into ethanol have as yet rel­atively low sugar-to-ethanol efficiencies, if compared with the well-established starch — or sucrose-to-ethanol conversion processes (Chang 2007; Hahn-Hagerdal et al. 2007; Olofsson et al. 2008). In view of the problems in converting ligno — cellulosic feedstocks into alcohol, there is a lively search for improvements, if not a ‘technological breakthrough’ or a ‘superbug’ that is able to perform the task of converting lignocellulose into ethanol with sufficient efficiency (Eijsink et al. 2008; Gomez et al. 2008; Rotman 2008).

In North America and Scandinavia, ethanol from woody lignocellulose has been, and still is, produced as a by-product of sulphite pulping for the paper industry (McElroy 2007). Wood hydrolysate is in this case converted into ethanol by yeast — based fermentation. Low nutrient concentrations, a large proportion of xylose and the presence of fermentation inhibitors have limited the efficiency thereof, and there are proposals for the optimization of sulphite liquor fermentation (Helle et al. 2008).

In Russia, there is a long-standing, large-scale, yeast-based production of ethanol from sugars obtained from wood chips hydrolyzed at elevated temperature by treat­ment with concentrated sulphuric acid (Bungay 2004; Zverlov et al. 2006). In Brazil, Europe and the USA, there are pilot plants producing ethanol from lignocellulose or components thereof such as cellulose and hemicellulose (Wheals et al. 1999; Bryner 2007a). Large-scale plants are under construction and consideration. In part, ethanol production from lignocellulose in such plants is combined with ethanol production based on sugar or starch.

Alternatively, a bacterial fermentation process for the production of the biofuel butanol from lignocellulose may be considered (Zverlov et al. 2006; Ezeji et al. 2007; Qureshi et al. 2008a, b). This process was used during the twentieth cen­tury in the Soviet Union for the fermentation of hydrolyzed lignocellulosic wastes (Zverlov et al. 2006). In this case, lignocellulose was hydrolyzed by treatment with high concentrations of sulphuric acid, and the hydrolysate was fermented in combi­nation with the fermentation of starch. H2 originated as a by-product (Zverlov et al.

2006) . Also, processes producing butanol from lignocellulose based on treatment with dilute acid followed by enzymatic treatment and fermentation (simultaneous saccharification and fermentation) of harvest residues have been proposed (Ezeji et al. 2007; Qureshi et al. 2008a, b), as have been processes to convert complex or­ganic feedstocks into mixtures of alcohols using mixtures of fermentative bacteria (Bagajewicz et al. 2007). Finally, there is research into the possibility of enzymati­cally converting lignocellulose into fatty acid ethyl esters (Royal Society 2008).

Fourthly, methane may be produced from complex organic materials. Methane is also the molecule that makes natural gas a fuel, and natural gas supplies currently about 3% of primary energy for transport (de la Rue du Can and Price 2008). In 2004, there were about 3 million motorcars powered by natural gas, usually biva­lent vehicles able to drive on compressed natural gas and gasoline (Dondero and Goldemberg 2005; Janssen et al. 2006). Substantial use of vehicles powered by nat­ural gas is found in Argentina (world leader with about 800,000 of such vehicles by

2005) , India, Pakistan, Brazil, the USA and some countries in the European Union, such as Italy (Janssen et al. 2006). Large-scale application of methane in cars is dependent on a good refuelling infrastructure (Janssen et al. 2006). Natural gas is also used in ship and on-farm transport (Royal Society 2008; Boijesson and Mat — tiasson 2008). Alternatively, methane may be converted into liquid fuels using the Fischer-Tropsch reaction or via a process with ethylene as an intermediary (Hall

2005) . Currently, use of methane from natural gas in the Fischer-Tropsch synthesis of hydrocarbons is applied in diesel production, and this application is expected to increase in the future (Bagajewicz et al. 2007; Bryner 2007b; Takeshita and Yamaji 2008). Methane can furthermore be converted into methanol (Huber et al. 2006; Cantrell et al. 2008).

The use of methane in transport and the production of other transport fuels may be extended to biogenic methane (Murphy and McCarthy 2005; Boijesson and Mattiasson 2008; Lehtomaki et al. 2008). Above, the production of synthetic nat­ural gas from syngas has already been referred to. Methane can also be produced from a wide variety of biomass and biomass-derived materials, including complex wastes, using mixed cultures of micro-organisms in anaerobic reactors (Murphy and McCarthy 2005; Kleerebezem and van Loosdrecht 2007; Bocher et al. 2008; Ros et al. 2009). It has been proposed to use marginal lands for the large-scale growth of feedstocks and convert those into methane in decentralized biogas reac­tors (Schroder et al. 2008). A variant of this approach has been suggested that also allows for the bioconversion of CO2 to CH4 (Alimahmoodi and Mulligan 2008). Landfills can also be exploited for methane production. Before application in trans­port, CH4 production from biomass should be followed by upgrading. The extent of upgrading necessary varies, depending on the methane source. More upgrading is usually needed for methane from refuse in landfills and sewage sludge than for methane from manure (Rasi et al. 2007). Upgrading partly serves to remove com­pounds that may negatively affect engine performance or emissions, such as halo — genated compounds, siloxanes, H2S and NH3 (Ferreira-Aparicio et al. 2005; Rasi et al. 2007). Upgrading may also aim to increase methane content.

Hydrocarbons (‘Biocrude’) from Terrestrial Plants

In the 1930s, there were some efforts to cultivate Euphorbia, producing hydrocar­bons for biofuel production (Kalita 2008). Subsequently, there has been substantial research regarding plants producing latex which may be cracked to yield transport biofuels. The Euphorbia lathyris did relatively well in this respect and has been calculated to yield about 48 MJ ha-1 year-1 in biofuel: 26 MJ as hydrocarbons and 22 MJ as ethanol (Kalita 2008). As will be shown in Chap. 2, such energetic yields are relatively low compared with biofuels from current terrestrial crops, and there is no current commercial application of ‘biocrude’ from terrestrial plants.

4.1 Introduction

Here we will consider climate effects and non-greenhouse gas emissions associated with the life cycles of transport biofuels. Considering the whole seed-to-wheel life cycle of transport biofuels is important because at each stage, there may be signifi­cant environmental impacts. In cropping plants, there are usually inputs of fossil fu­els (needed to power tractors and generate N fertilizer) and emissions of substances such as N2O and nitrate (both derived from nitrogen fertilizers) and pesticides. The effects thereof may be substantial. For instance, Donner and Kucharik (2008) have shown that a US bioethanol production of 15 billion gallons would increase the an­nual average flux of dissolved inorganic fixed nitrogen (N) in the Mississippi and Atchafalaya rivers by 10-34%. And it has been argued that in Brazil, ethanol pro­duction from sugar cane will contribute to a rapidly changing tropical biogeochem­ical N cycle, because the N fertilizer use efficiency of production is low: about 30% of N fertilizer ends up in sugar cane tissue (Galloway et al. 2008).

Also the choice of plants grown as biofuel feedstocks may matter, as plants, for instance, differ in their production of isoprene, which may contribute oxidizing smog (Royal Society 2008). The stage of processing biomass to transport biofuels is associated with energy use and process-related emissions. For instance, dry mill fuel ethanol production leads to significant emissions of ethanol, acetaldehyde, acetic acid and ethylacetate (Brady and Pratt 2007). Leakage from storage facilities for bioethyl-tertiary-butylether (ETBE) may have a significant impact on groundwater quality (Rosell et al. 2007). And the stage of driving is important, because driving a car on biofuels generates emissions which may be different from those of fossil fuels.

This chapter starts with an overview of the different types of uncertainty related to life cycle studies (Sect. 4.2). After that, most attention will be given to life cycle emissions of biofuels based on terrestrial plants and wastes. These are considered in Sects. 4.3 to 4.5. Section 4.6 will deal with possibilities to reduce the life cycle emissions associated with the biofuels considered here.

L. Reijnders, M. A.J. Huibregts, Biofuls for Road Transport © Springer 2009

2.1 Introduction

The adjective ‘sustainable’ is frequently used regarding biofuels (e. g. Abrahamson et al. 1998; Krotscheck et al. 2000; Buckley and Schwarz 2003; Bhattacharya et al. 2003; Goldemberg and Teixeira Coelho 2004; Butterworth 2006; Demirba§ 2007; Robert et al. 2007; Goldemberg et al. 2008; Karp and Shield 2008; Royal Society 2008). Also, biofuels are a regular subject in scientific journals dealing with re­newable or sustainable energy. The apparent rationale of using ‘sustainable’ and ‘renewable’ in the context of biofuels is the following: biomass may be argued to temporarily store solar energy, based on photosynthesis (see Chap. 1). In doing so, carbon is sequestered, and on burning transport biofuel, it is de-sequestered. In the meantime, photosynthesis proceeds, generating new feedstocks for biofuels. As solar irradiation and photosynthesis are expected to last for many millions of year, doing so would seem sustainable and transport biofuels renewable. However, this is not the ‘whole story’. Energy inputs in the world economy are currently, as pointed out in Chap. 1, overwhelmingly fossil fuels, and the use of fossil fuels extends to the production and distribution of transport biofuels. This is at variance with renew — ability and sustainability, as fossil fuels are non-renewables, and their use cannot be sustained indefinitely at the present level. The cumulative life cycle fossil fuel demand of biofuels will be discussed in Sect. 2.3.

For converting solar irradiation into transport kilometres, there are a variety of technologies available with widely varying efficiencies. Such efficiencies matter: they are major determinants of spatial requirements of energy supply. These spatial requirements, in turn, are important determinants of competition of energy supply with food production and habitats for living nature. Because this competition is an important matter in the current transport biofuel debate and will return later in this book, this chapter will deal with the solar conversion efficiency of transport biofuels (Sects. 2.4 and 2.5). Other methods for solar energy conversion do not involve organisms but rely on physical conversion technologies. Photovoltaic cells generating electricity are examples thereof, for which the solar conversion efficiency will be discussed in Sect. 2.6.

L. Reijnders, M. A.J. Huibregts, Biofuls for Road Transport © Springer 2009

Biofuels and the output from photovoltaic cells can be used to perform work or to deliver energy services. Work (a thermodynamic concept) or energy services (an economic concept) include, for instance, car kilometres. The performance of work often includes the use of intermediaries (e. g. power plants, batteries or motors). The energy efficiencies of such intermediaries will be discussed in Sect. 2.6. In Sect. 2.6, we will also consider the overall efficiency of a variety of methods to convert solar energy to car kilometres, giving a ‘seed-to-wheel’ perspective.

As pointed out in Chap. 1, the production of biofuels is often accompanied by by-products or co-products. For instance, in making biodiesel from rapeseed, both glycerol and an ingredient of animal feed (rapeseed cake) are produced. Before we go into the calculations of cumulative (fossil) energy demand, it should be decided how much of that demand is allotted to biodiesel and how much to glycerol and rapeseed cake. This is called ‘allocation’ and will be discussed in Sect. 2.2.

Lignocellulosic biomass produced in a way that does not significantly change C sequestration in ecosystems will do better than fossil fuels in electricity produc­tion (see Fig. 4.2). Biodiesel made from palm oil may well be inferior to biodiesel when forests are cleared to grow oil palms (Danielsen et al. 2008; Fargione et al. 2008; Reijnders and Huijbregts 2008a). The same will hold for coconut biodiesel (contrasting the conclusion of Tan et al. 2004). But the opposite will hold when palms are established on abandoned agricultural land where there is currently little sequestration of C, provided that cultivation does not lead to lowering of soil carbon stocks (e. g. Germer and Sauerborn 2007). When, under market conditions, perennial grasses, short rotation woody perennials and herbaceous species such as Miscanthus are used in industrialized countries as lignocellulosic biomass for the production of Fischer-Tropsch diesel or ethanol, it would seem likely that there will be a large indirect effect on land use involving clearance of natural vegetation (Searchinger et al. 2008). In this case, it is doubtful whether lignocellulosic biofuels will much outperform fossil fuels regarding life cycle greenhouse gas emissions (see also the last row of Table 4.2).

The use of a variety of algae has been considered because of their direct and indirect biocatalytic effect on the splitting of water in H2 and O2 (Melis and Happe 2001; Hallenbeck and Benemann 2002; Nath and Das 2004; Savage et al. 2008). In spite of a nearly 70-year history of research, actual production of H2 by algal systems is still very low, about 2 g of H2 per square metre of culture area per day (Melis and Happe 2001), and H2 has to be withdrawn continually as the overall conversion of glucose into H2 is energetically only slightly favourable to H2 (Savage et al. 2008). At realistic solar irradiation, solar conversion efficiencies in optimized systems for direct and indirect biophotolysis seem to be in the order of 1% or lower, when pure cultures can be maintained (Hallenbeck and Benemann 2002; Yoon et al. 2006; Rupprecht et al. 2006; Burgess and Fernandez-Velasco 2007). And as pointed out before, the cumulative energy demand for algal H2 production is probably of the same order of magnitude as the energetic output, when the solar energy conversion efficiency does not exceed 1% (Burgess and Fernando-Velasco 2007).

A trend in the production of ethanol is increased interest in the combined conversion of starch, cellulose and hemicellulose into ethanol. In this way, the traditional pro-

L. Reijnders, M. A.J. Huibregts, Biofuls for Road Transport © Springer 2009

duction of two outputs (ethanol and dried distillers grains with or without solubles — to be used in, for example, animal feed) is replaced by one output: ethanol (Linde et al. 2008). Additionally, efforts are under way to eliminate a second co-product from ethanol production: glycerol (Bideaux et al. 2006). Similarly, in butanol pro­duction, there is currently much effort focused on getting rid of the by-products acetone and ethanol (Antoni et al. 2007; Durre 2008). Also, anaerobic conversion by mixtures of micro-organisms converts a wide variety of organic substances to one fuel: methane.

On the other hand, there is also a trend to widening the variety of outputs of production processes generating transport biofuels. At the factory level, the anal­ogy to petrochemical refineries has given rise to the concept of a ‘biorefinery’ that produces a variety of products from biomass feedstocks (Kamm and Kamm 2004; Arifeen 2007; Hayes 2008). In producing more than one product in the context of fermentative ethanol production, a variety of technologies may be used, including several biotechnologies and chemical synthesis technologies, the latter starting from ‘platform chemicals’ such as levulinic acid (Hayes 2008; Huang et al. 2008). When synthesis gas or hydrocarbons are produced from biomass, one might envisage the development of refineries with synthetic technologies which are now commonly ap­plied in the petrochemical industry (e. g. Chew and Bhatia 2008; Rowlands et al. 2008). Biorefinery concepts including the production of monomers for current bulk chemicals such as eth(yl)ene and caprolactam have been proposed, too (Kamm and Kamm 2004). Also, biorefineries based on hydrocarbons containing oxygen have been suggested starting from pyrolysis oil (Hayes 2008).

In line with the biorefinery concepts focusing on the use of biotechnology and separation technologies, there is, for instance, an operational factory that con­verts corn into the biofuel ethanol and also produces citric acid, lactic acid, amino acids and enzymes. And a wheat biorefinery has been proposed generating, besides ethanol, ferulic acid, arabinoxylan, amino acids and gluten (Arifeen et al. 2007).

There is probably a place for both the biorefinery and single output approaches. And both may have significant implications elsewhere in the economy. Eliminating current by-products of biofuel production which are used as animal feed, such as dried distillers grains, will probably have an upward effect on other types of animal feed production (Searchinger et al. 2008).

As pointed out in Sect. 1.7, if transport biofuels are going to replace current fossil fuels on a large scale by multi-output types of production, markets for by-products may be easily flooded, leading, among other things, to major price reductions for such by-products. This has already happened in the case of glycerol, a by-product of biodiesel production (Yazdani and Gonzalez 2007). In the summer of 2007, a glut of dried distillers grains with solubles, a co-product of bioethanol production, in the USA led to relatively low prices paid for its use as an ingredient in animal feed (Tyner 2008). And flooding markets by biorefineries may also happen in marketing co-products such as xylitol, xylo-oligosaccharides and lignin (Kadam et al. 2008).

A variety of algae are currently cultivated commercially, especially for applications in food and feed production, but also for other applications such as fertilizer and the production of materials. Also, there is limited harvesting of uncultivated algae (Critchley et al. 2006). There have been proposals to exploit aquatic biomass for the production of biofuels.

Marine Phytobiomass

Most of the surface of the Earth consists of seas, mainly oceans. A variety of pro­posals exist to exploit the seas for the production of biofuels. Macroalgae, macro­phytes and microalgae have been considered in this context. Microalgae include both prokaryote and eukaryote photosynthetic micro-organisms. In the context of exploiting macrophytes, floating man-made structures to cultivate the Macrocystis pyrifera (giant kelp) have been proposed (Wilcox 1982; Bungay 2004). Varieties of the brown seaweed Laminaria, which is currently harvested for food (Chopin et al. 2001), have been suggested as a convenient source of carbohydrates to be con­verted into ethanol (Hornet al. 2000). The highly salt-tolerant microalgaDunaliella, which, for instance, occurs in the Dead Sea, has also been proposed as a source of transport biofuel (Ben-Amotz et al. 1982).

However, there is a major snag regarding the proposal to use the sea for the production of algae, which may serve the supply of transport biofuels. Actual phy­tobiomass in the seas is in the order of 1-2% of total global plant carbon. Photosyn­thesis in the seas is much higher (in the order of 40-50% of total photosynthesis; Rosing et al. 2006), but most of the photosynthetic yield (approximately 80-88%) is quickly consumed. In the case of microalgae, consumption is mainly by zooplankton (‘grazers’), while 2-10% is subject to viral lysis (Wilhelm and Suttle 1999). Thus, substantial direct appropriation of the products of photosynthesis by humans in the seas in general would necessitate a major overhaul of the marine food web. For the successful growth of desired microalgae, probably dramatic changes in ocean com­position, such as a switch to much higher salinity, may be required (Sawayama et al. 1999; Joint et al. 2002; Ugwu et al. 2008). Large-scale exploitation of macroalgae is cumbersome. Proposals to exploit the giant kelp Macrocystis require pumping of deep seawater to the ocean surface, massive man-made structures to support kelp growth and regular replanting (Bungay 2004).

Estimates of life cycle impacts are subject to uncertainty. In life cycle assessments, there is uncertainty in input data (parameter uncertainty), in normative choices (sce­nario uncertainty) and mathematical relationships (model uncertainty) (Huijbregts et al. 2003; Lloyd and Ries 2007). As the focus in this chapter is largely on com­paring fuels, model uncertainty tends to be rather similar for all fuels, which is favourable to the comparative value of life cycle assessments.

Parameter uncertainty may be limited by using relatively good quality inven­tories of emission and resource use data, such as the JLCA-LCA inventory from Japan (Suguiyama et al. 2005) and the Ecoinvent database (cf. Zah et al. 2007), as well as recent peer-reviewed research into emissions and resource use. In this way, uncertainties in contributions of industrial and transport activities to impacts of bio­fuels, especially in industrialized countries, can be limited. However, uncertainties about industrial activities in some developing countries may remain relatively large because of major uncertainty about fuels, energy efficiency and environmental tech­nology (e. g. Reijnders and Huijbregts 2008a; de Vries 2008). Uncertainties linked to the fate of C and N in cropping and harvesting feedstocks are relatively large. In the case of N2O emissions associated with intensive cropping, uncertainty in net greenhouse gas emissions may well be ±20% (Reijnders and Huijbregts 2008b), and uncertainties about changes in C stocks of ecosystems may also be quite substantial.

Normative choices are an important source of uncertainty. One of these choices relates to the time that land will be used to produce biofuel. Choices regarding this time are important in calculating net greenhouse gas emissions due to land use change (e. g. Reijnders and Huijbregts 2008a; Wicke et al. 2009). Allocation in the case of multi-output production is another normative choice that is important. As explained in Chap. 2, there are three major ways to allocate. The first is based on prices, the second on physical categories, such as energy or weight, and the third on subtracting avoided processes (also called substitution). Apart from choosing the basis for allocation, there may also be other matters to consider. Take, for example, the conversion of lignocellulose in dried distillers grains with or without solubles [DDG(S)] to ethanol. Lignocellulosic outputs of ethanol production such as dried distillers grains are currently an ingredient of animal feed (Taylor-Pickard 2008). If one goes back in history, there have been advocates for replacing ingredients that had high starch contents by lignocellulosic ingredients, such as DDG(S), in ani­mal feed (e. g. Sarkanen 1976). Now, when these lignocellulosic components are diverted to the production of transport fuel, will this give rise to an increase in the starch content of animal feed? And should the effects thereof on the net emission of greenhouse gases be allocated to lignocellulosic ethanol, and if so to what ex­tent? Also, an objection has been raised against considering dried distillers grains as a product output of ethanol production, to which non-product outputs should be allocated (Patzek 2004). According to Patzek (2006) dried distillers grains should become an input in cropping and spread on the fields to diminish the need for nitro­gen fertilizer, decrease soil erosion and improve the energy efficiency of cropping. Patzek (2006) has also argued that if there is any crediting at all of dried distillers grains, the energy credit should be somewhat negative.

The complications of dealing with co-products such as DDG(S) and alloca­tion may well seem so problematic that no ‘iron-clad’ estimate of net greenhouse gas emissions associated with biofuels from multi-output processes seems feasible. Only limited study has been made as to the differences in estimated environmental impacts of transport biofuels following from the different approaches to allocation. Eickhout et al. (2008) found that the substitution approach and allocation on the basis of energy-generated outcomes for ethanol and biodiesel were in the range of ±15%. Curran (2007) looked at the impact of different ways of allocation (based on price, weight, volume and energy) on the relative environmental ranking of con­ventional gasoline and bioethanol and found that this ranking was the same in all instances. On the other hand, Reijnders and Huijbregts (2005) found that alloca­tion based on either price or energy may lead to a difference in the environmental ranking of fossil-fuel- and manure-based electricity. And Malga and Freire (2006) did show that in the case of wheat ethanol, different ways to allocate have a major influence on results.

In the following, we will indicate what type of time frame and allocation has been used in arriving at specific results.

There are three major ways to allocate. The first is based on prices, the second on physical categories, such as weight or energy, and the third on subtracting avoided processes (also called substitution). We will look at these in turn. The first way to allocate is on the basis of price (market values). The idea behind this type of allocation is that prices drive production (Weidema 1993). This method is, however, not without problems. Firstly, market prices are not constants. So, if, for example, ethanol prices go up, whereas the prices of other outputs do not, the emissions and cumulative fossil energy demand allocated to this transport fuel increase. The same happens when by-products go down in price, but the transport biofuel price remains constant (or increases). A good example of the latter is the tenfold price decrease of glycerol between 2004 and 2006 (Yazdani and Gonzalez 2007).

A second problem is that currently, much transport biofuel production is not driven by market value but by market value plus subsidy. This leads to the question of whether, for instance, in the case of ethanol production from cornstarch, alloca­tion should be on the basis of the market value of cornstarch or on the basis of the subsidized value. Another problem arises when wastes are considered. These may well have negative prices (being a cost to the producer). For instance, the producer of the waste may have to pay a price for the incineration or treatment of his waste. If so, allocation on the basis of price may mean that the waste, because of its negative price, is apparently associated with a negative cumulative energy demand (Reijnders and Huijbregts 2005). Usually, this has been felt unsatisfactory by proponents of al­location based on prices, and this often leads to the decision the give a zero price to wastes. However, this seems inconsistent. An implication of a zero price is that the life cycle leading to the generation of wastes has no impact on the environmental evaluation of such biofuels. The problem may also arise as to whether something is a waste or a by-product. An example thereof is sawdust. This may be used for firing industrial installations or power plants, and then may be categorized as by-product (with a positive monetary value), but sawdust may also be left in the woods and may then be categorized as a waste (with zero monetary value). Decisions regarding such categorizations may be far from easy and may have a substantial impact on the greenhouse gas emissions calculated.

Alternatively, one may allocate on the basis of physical categories such as ‘en­ergy content’ (heating value) or weight. For instance, the European Union in its 2008 draft Renewables Directive has proposed to allocate on the basis of energy (Eickhout et al. 2008). This type of allocation has the advantage of stable outcomes, unaffected by movements of prices. However, there are curious consequences, too. For instance, in this allocation system, there is an obvious way to improve the en­vironmental performance of a transport biofuel, and that is to produce more waste. To evade this problem, there is a tendency to restrict allocation to product outputs. Matters related to quality may also emerge. If one, for instance, allocates to the out­puts of electricity and low temperature heat on the basis of ‘energy content’, one may be criticized for neglecting the quality of these outputs and be advised to use exergy instead of energy. Thus, allocation on the basis of physical categories may encounter criticism if the physical property chosen is at variance with the perceived value of the co-products.

Another way to deal with a multi-output process is to ‘correct the system’. In the case of biofuels, one may consider biofuel to be the only output and correct for the other outputs by subtracting ‘avoided processes’ which such outputs can substitute (Ekvall and Finnveden 2001). This approach has also been called substi­tution. For instance, in the case of ethanol production from corn or wheat, it has been argued that by-products such as dried distillers grains (DDG) or dried distillers grains with solubles (DDGS) may be a substitute of soybean meal in cattle feed (Kim and Dale 2005). Thus, producing DDG(S) may be valued on the basis of the avoided process of producing soybean meal. However, soybean meal and DDG(S) are not identical. This then raises the question of the basis for conversion: should it be on the basis of price, or protein content, or metabolizable joules (energy)? Moreover, DDG(S) is not a straightforward substitute of soybean meal, as its com­position is relatively variable, and its consumption by animals may be linked to increased mycotoxicosis risk and increased intakes of mycotoxins (Taylor-Pickard 2008). This has led to a more limited recommended use of DDG(S) in animal feed than in the case of soybean meal (Taylor-Pickard 2008). Then there is the mat­ter of applications other than animal feed. For instance, soybean meal may also be used to generate vegetarian alternatives to meat, and DDG(S) may be used to produce protease and peptones (Romero et al. 2007), methane (Murphy and Power 2008) or ethanol. Such alternative applications may have environmental impacts that are very different from the use as an ingredient of animal feed. Suppose, fi­nally, that DDG(S) is indeed valued on the basis of avoiding soybean meal; the problem is that soybean meal is a co-product, just as DDG(S) is. This may be argued to imply that substitution in this case means plugging one hole with an­other.

So, each way to allocate has its weak points, and there is no agreement on the best way to allocate. In this book, we will not make a choice in favour of a specific way to allocate but rather will explicitly indicate what type of allocation has been used in arriving at specific results.

There is substantial combustion of biowastes for the generation of electric power, which in principle may be used for traction. There is also substantial anaerobic con­version of biomass wastes into methane, which in turn can be used for automotive purposes. Biodiesel made from waste fats and oils (yellow grease) is currently the main biowaste-derived road transport biofuel. As pointed out in Chap. 1, there are a wide variety of processes that have been proposed to convert wastes into liquid biofuels, such as ethanol, methanol and Fischer-Tropsch liquid transport fuels, and there is also the possibility for conversion into H2.

How does one evaluate the environmental impact of using biowastes to power transport? This is a tricky question. We will illustrate this with two examples, drawn from major applications of such wastes: burning to generate electricity and conver­sion into methane. Firstly, the allocation of emissions has a large impact on those evaluations. In this context, one may consider the example of electricity production from chicken manure in the European Union (Reijnders and Huijbregts 2005). At the time of this study, such manure had a negative price. Reijnders and Huijbregts (2005) calculated the emissions of greenhouse gases associated with electricity pro­duction using allocation on the basis of actual prices, allocation on the basis of prices but with an assumed zero price for manure and allocation on the basis of energetic outputs of chicken production. The results thereof are shown in Table 4.2.

As also pointed out in Chap. 2, so far it has been usual in life cycle assessments of biofuels from wastes to allocate on the basis of a zero price, implying that the life cycle leading to the generation of wastes has no impact on the environmental eval­uation of such biofuels. A first problem with this approach is that when the current

waste indeed turns out to be viable as a source of biofuel production, it will turn into a ‘secondary resource’ which may well have a positive price. A second complication centres around the reference to be used and the ‘normal fate’ of the waste used. This again may have a very large impact. For instance, a study about Dutch projects to convert manure into methane (Zwart et al. 2006) concluded that the energetic out­put (methane) was roughly the same as the energetic input (fossil fuels). However, Zwart et al. (2007) calculated a very low net greenhouse gas emissions for fermen­tation of manure, because they did not compare greenhouse gas emissions linked to methane from manure with the life cycle emissions of natural gas, but rather they compared fermentation with other ways of handling manure. Also, for the energetic application of manure, they assumed a major reduction in the emission of methane and N2O due to a much-reduced storage time for manure. So estimates about the environmental impacts of biowastes used for fuelling transport are dependent on subjective assumptions.

Zah et al. (2007) have studied emissions associated with methane production from a variety of wastes, using allocation on the basis of prices and a zero value of the waste itself. Thus, the calculation of emissions linked to methane production from wastes was restricted to the waste-to-wheel stages of the life cycle. Compar­ison was with natural gas. The wastes considered were: sewage sludge, biowaste, manure and manure plus co-substrates. The emission of greenhouse gases for the production of methane from these wastes was in the order of 50-80% of the fossil reference. When allocation would have been seed-to-wheel on the basis of energy or mass, the emission of greenhouse gases linked to waste-based methane production would have been much higher (cf. Table 4.2).