11.2.1 Green biorefinery

A ‘green biorefinery’ represents the sustainable processing of green biomass into a spectrum of marketable products and energy (Cherubini et al., 2009). In the context of a green biorefinery, green biomass can include any naturally occurring wet biomass such as agricultural crops (e.g., grass, lucerne, clover and immature cereal) and agricultural residues (e.g., sugar beet leaves). These plants represent a natural chemical factory and can be rich in basic products such as carbohydrates, proteins, lignin and lipids, as well as various other substances such as vitamins, dyes and minerals (Kamm and Kamm, 2004). Green biorefineries are often described as multiple-product systems, with the potential range of products depending on the composition of the feedstock presented to the biorefinery. This multiple-product approach was born out of economic necessity, with single-product approaches often struggling to create sufficient revenue to cover the feedstock and subsequent processing costs (Mandl, 2010). As such, the viability of an industrial facility processing green biomass will depend on the range of suitable applications identified for the separated fractions (Kromus et al., 2004). Furthermore, this multiple-product approach can facilitate the efficient utilisation of the whole plant and any process residues.

11.2.2 Grassland composition

Poaceae (i.e., grass family) and Fabaceae (i.e., legume family) represent two of the major plant families used as forages. The Poaceae are a large and ubiquitous family of monocotyledonous flowering plants and include three of the most important crops in the world in wheat (Triticum spp.), maize (Zea mays) and rice (Oryza spp.), as well as many lawn and pasture grasses (e.g. Lolium spp.). The family Fabaceae are a large family of dicotyledonous flowering plants and also represent a large number of important agricultural and food plants including soybean (Glycine max), clover (Trifolium spp.) and lucerne (Medicago spp.).

Grasses and legumes can be further classified into two physiological groups, C₃ (i.e., cool season or temperate) and C₄ (i.e., warm season or tropical) plants, based on their photosynthetic pathway. There are distinct differences between the photosynthetic pathways of C₃ and C₄ species which affect light, water and nitrogen use efficiencies (Lattanzi, 2010). In warm regions, C₄ grasses can outyield C₃ grasses due to their more efficient photosynthetic pathway. However, the lower temperatures and shorter growing seasons in Northern Europe limit the growth of C₄ plants (Lewandowski et al., 2003). Furthermore, C₃ and C₄ plants may be classified as either annual (i.e., life cycle completed in one growth season) or perennial (i.e., life cycle completed in two or more years).

The term ‘grassland’ in this chapter refers to a plant community in which true grasses are usually the dominant species, with forbs (i.e., herbaceous dicotyledon species, including legumes) present in variable amounts (Hopkins, 2000). Grasslands are a major agricultural resource, supporting both intensive and extensive systems of ruminant livestock production. The main focus of this chapter is on C₃ or temperate grassland biomass that is currently or has recently been used as forage for animal production.

Intensively managed temperate grasslands are dominated by a small number of grass species including perennial ryegrass (Lolium perenne), Italian ryegrass (Lolium multiflorum), timothy (Phleum pratense), cocksfoot (Dactylis glomerata) and tall fescue (Festuca arundinacea) (Hopkins and Holz, 2006; Plantureux et al., 2005). These grasses may be present as indigenous species in permanent pastures (i.e., land used to grow grasses or other herbaceous forage naturally (self-seeded) or through cultivation (sown) and that is not included in the crop rotation of the holding for five years or longer; Commission Regulation E.U. No. 796/2004), be part of swards reseeded many years previously, or may be resown every few years as part of a crop rotation (Buxton and O’Kiely, 2003). The majority of reseeded temperate grassland in Europe is now dominated by perennial ryegrass due to its high digestibility when harvested at the appropriate growth stage, high yield in response to nitrogen fertiliser application, and ease of preservation as silage due to its relatively high water soluble carbohydrate content (Whitehead, 1995). However, in situations where ryegrasses are limited by winter survival, there is a continuing role for timothy, while in drier temperate regions there is a role for more drought-tolerant species such as cocksfoot and tall fescue (Hopkins and Wilkins, 2006). Legumes such as white clover (Trifolium repens), lucerne (Medicago sativa) and red clover (Trifolium pratense) can also represent a prominent component of temperate grasslands (Peeters et al., 2006). These legumes can be grown alone or in combination with compatible grasses and they can reduce the requirement for fertiliser N through fixation of atmospheric N, while also increasing herbage yield and quality (Peyraud et al., 2009).

11.2.3 Grassland in Europe – role and importance

Grasslands play a major role in global agriculture, accounting for approximately 70% of the world agricultural land area and 26% of total land area (FAO, 2012). In the EU-27, permanent and temporary grasslands represent more than 30 and 5% of the utilised agricultural area, respectively. However, this varies considerably between countries and regions with, for example, permanent grassland constituting 75% and 1.5% of the utilised agricultural area in Ireland and Finland, respectively (Eurostat, 2008). Furthermore, grasslands vary greatly in their degree and intensity of management, with grassland systems in Europe ranging from extreme Taiga vegetation in the far north to dry Mediterranean grassland in the south (Smit et al., 2008).

Grasslands are predominantly used for animal production, particularly as a principal source of feed for ruminants. However, animal production systems differ across Europe. In Northwest Europe, for example, grasslands can meet up to 100% of the nutrient requirements of the ruminant livestock population, while in other parts of Europe other forage crops and temporary grassland are more important. In many European countries grassland utilisation is decreasing due to a trend towards more controlled animal production systems or to decreasing livestock numbers. Grassland-based production is being increasingly challenged by the use of other forage and concentrate feeds (Wilkins et al., 2003). For example, the relatively low proportion of grassland in some specialised dairy farms can be explained through higher stocking rates and higher proportions of green maize and temporary grassland used as forage crops (Dillon, 2010). For grassland that is no longer needed for intensive dairy or beef production, intensive management is often unlikely to be continued (Stoate et al., 2009; Isselstein et al., 2005).

Grasslands are also becoming increasingly recognised for their contribution to environmental functions, and the protection of grassland area is integrated into the Common Agriculture Policy (CAP) cross-compliance system. This requirement for the protection of permanent grasslands is in recognition of the positive impacts of grassland compared with cropping/arable land on a range of ecosystem services (Dillon, 2010). Grasslands play an important multifunctional role in maintaining floral and faunal diversity (Isselstein et al., 2005), providing water catchment protection and soil erosion control (Cerdan et al., 2010), reducing the impact of global warming through carbon sequestration (Peeters and Hopkins, 2010) and providing landscape and amenity value (Gibon, 2005). For example, semi-natural grasslands are among the most species-rich habitats in Europe. Generally, natural and semi-natural grasslands are found in extensive, low stocking rate production systems in Europe (Dillon, 2010). The maintenance of semi-natural grassland habitats through traditional agricultural practices is vital for the protection of biodiversity. Periodic defoliation of extensive grassland, by mechanical cutting or grazing, is vital for controlling succession of plants (Rook et al., 2004) and is essential for maintenance of grassland habitats.

11.2.4 Grassland productivity and potential availability

Data on actual grassland productivity and its spatial distribution in Europe are scarce. Grassland productivity is affected by soil characteristics, botanical composition, climatic conditions (e.g., rainfall and temperature) and specific management practices (Peeters and Kopec, 1996). Smit et al. (2008) presented data on grassland productivity across Europe. The highest productivity, about 10 t DM ha^− 1, was reported to be achieved in temperate Atlantic zones (e.g., Netherlands, Britain and Ireland, North Germany), while regions with the lowest productivity (1.5 t DM ha^− 1) are located close to the Mediterranean. Central European countries (e.g., Germany, Austria and Switzerland) were reported to have intermediate yields of 6 t DM ha⁻¹ and higher. Aside from favourable climatic conditions, high fertiliser N inputs are also a major determinant of DM yields. In general, however, information on the biomass yields of different grassland species in Europe is difficult to compile since grass yields are rarely directly measured in farm practice. In an extensive study comparing the productivity of perennial ryegrass and timothy across 32 European sites, yields of perennial ryegrass varied from almost 2 t (Vila Real, Portugal) to 20 t DM ha⁻¹ (Kiel, Germany) and were higher than timothy in most cases (Peeters and Kopec, 1996).

Some EU countries are facing difficulties in ensuring regular utilisation of extensive grasslands by ruminants to meet conservation objectives (Wachendorf et al., 2009). In other regions, more controlled ruminant production systems or decreasing livestock numbers are reducing the requirement for intensively managed grassland. However, little or no information is available on the grassland biomass resource available in Europe for alternative applications. In Ireland, for example, grass is a biomass resource that is readily available (O’Keeffe et al., 2011). In a recent study, McEniry et al. (2013a) calculated the annual grassland resource available in Ireland as the difference between current estimated grass supply and the grass requirement of the national cattle herd and sheep flock. They reported that under current grassland management practices in Ireland, there is an estimated annual grassland resource of ca. 1.7 million t DM available in excess of livestock requirements. Furthermore, increasing N fertiliser input combined with increasing the grass utilisation rate by cattle has the potential to significantly increase this resource to 12.2 million t DM per annum. Thus, there is potential for grassland biomass to be a readily available resource in Europe. Consequently, numerous policy, research and commercial groups have considered alternative options for grassland biomass use, with the production of renewable energy, chemicals and materials from grassland biomass receiving considerable interest.

11.3.1 Green plant chemical composition

Temperate forages can be divided chemically into two main fractions: (a) cell walls composed mainly of structural carbohydrates and (b) cell contents which include the most readily and highly digestible components (Moore and Hatfield, 1994).

Cell walls

The structural carbohydrates cellulose and hemicellulose are the major components of the plant cell wall and provide structural integrity to the plant (Esau, 1977). Cellulose is the carbohydrate polymer of greatest abundance in plant cell walls and is composed of β-1,4-linked glucose units. Hemicellulose is composed of multiple polysaccharide polymers and consists of sugars including glucose, xylose, galactose, arabinose and mannose. The main hemicellulose in plants is xyloglucan; however, in most grasses the main hemicellulose component is arabinoxylan (Reiter, 1998). The remainder of the cell wall is composed of lignin, protein and pectin, with lipids and minerals in smaller amounts (Theander and Westerlund, 1993). Lignin is a phenolic polymer and can be formed from the polymerisation of three main monomeric units: P-coumaryl, coniferyl and sinapyl alcohols, with lignin in grass being formed from all three monomers (Evert, 2006).

The content of cell wall material is greater in the stem than in leaves of forages, with the difference between these parts being greater in legumes than grasses (Wilson, 1993). With advancing plant maturity, the proportion of cell wall increases in relation to cell contents (Fig. 11.1). This reflects the general decrease in leaf to stem ratio and the increasing cell wall content within the stems (Buxton, 1996). This process is accompanied by an increasing content of indigestible lignin within the cell wall fraction (Ugherughe, 1986), and it is the presence of lignin in the cell wall that limits polysaccharide breakdown (Buxton and Russell, 1988). For legumes, the cell wall concentration of leaves changes little with advancing maturity, whereas stems progressively undergo cambial growth which adds thick-walled xylem tissue during development increasing the diameter of the stem (Wilson, 1993).

11.3.2 Grassland management

The genome of each grass species imposes differences in adaptability, productivity and chemical composition. Although the majority of reseeded temperate grassland in Western Europe is dominated by perennial ryegrass, other grass species (e.g., cocksfoot, tall fescue and timothy) have different physical and chemical characteristics which may offer benefits for non-agricultural uses or be more suited to specific management (e.g., response to fertiliser) or environmental (e.g., temperature, water deficit) conditions. For example, in general, ryegrass and cocksfoot are reported to produce a higher DM yield response to high rates of N fertiliser input compared with other common grassland species (Reid, 1985). In addition, Davies and Morgan (1982) reported that perennial ryegrass had a higher crude protein concentration than other common grass species. Similarly, King et al. (2012a) reported that the high water soluble carbohydrate concentration and the lower buffering capacity observed for perennial and Italian ryegrasses, compared with other common grassland species, would make them more suitable for preservation as silage. Furthermore, legumes generally have a lower concentration of water soluble carbohydrate and a higher buffering capacity than grasses, making them more difficult to preserve as silage (Buxton and O’Kiely, 2003).

The most important factor influencing the chemical composition of a specific herbage is the growth stage at harvest (Buxton, 1996). Advancing plant maturity from the vegetative to the inflorescence growth stage is characterised by an increase in fibre components and a decrease in digestibility, buffering capacity, and crude protein and water soluble carbohydrate concentrations (Fig. 11.1; Buxton and O’Kiely, 2003). This decrease in buffering capacity, crude protein and water-soluble carbohydrate concentration can be attributed to changes in chemical composition resulting from the declining cell content to cell wall ratio (Buxton, 1996). A much faster rate of decline in herbage quality is observed during the primary reproductive spring growth than during subsequent vegetative regrowths (Balasko and Nelson, 2003).

As well as influencing herbage quality, harvest date also has a significant effect on herbage DM yield, with a delayed harvest generally increasing DM yield. Low DM yields may reduce the biomass availability within the catchment area of a green biorefinery (O’Keeffe et al., 2011). Furthermore, the unit cost of producing a grass feedstock for a biorefinery will decrease as the biomass yield increases, but there must be an economic equilibrium between yield and herbage quality. The optimal stage for harvesting will be determined by the particular grassland biomass use. For example, grass harvested at earlier growth stages may be more suitable for biogas production through anaerobic digestion (McEniry and O’Kiely, 2013a), while grass harvested at later growth stages may be more suitable for technical fibre or combustion (McEniry et al., 2012).

Fertilisation of grassland, in particular the application of N, is employed primarily to ensure that economically viable yields are available for harvesting at a time of adequate herbage quality (Keating and O’Kiely, 2000a). While the yield response is greatest for N application, P, K and regular applications of lime are also required to maintain grassland productivity (Balasko and Nelson, 2003). However, N fertiliser can also impact on herbage quality and ensilability. Increasing the rate of N fertiliser application generally increases herbage crude protein concentration (Keady and O’Kiely, 1998) and buffering capacity (O’Kiely et al., 1997), and reduces herbage DM (Whitehead, 1995) and water-soluble carbohydrate concentrations (Keating and O’Kiely, 2000b). This increase in herbage buffering capacity will make the grass more difficult to preserve as silage than indicated by the decrease in water-soluble carbohydrate concentration alone (Keating and O’Kiely, 2000b). Although inorganic N fertiliser can increase herbage yields, there are financial and environmental issues associated with very high rates of application. For example, the use of N fertiliser in agricultural systems is one of the biggest contributors to greenhouse gas emissions through fertiliser manufacture and N₂O emissions from soils (Dillon, 2010). The valuable nutrient content of animal manures and/or digestate (from biogas production), and the incorporation of N-fixing legumes into grassland, all have the potential to reduce the requirement for inorganic N fertiliser resulting in positive financial and environmental gains (McEniry et al., 2011; Peyraud et al., 2009; Gerin et al., 2008).

11.3.3 Preservation

Under normal practical conditions, grassland biomass can be harvested and utilised immediately in a biorefinery, or it can be conserved under anaerobic, acidic conditions (i.e., silage) or following drying (i.e., hay). Variable weather conditions can make efficient conservation of hay impractical. Therefore, in order to ensure a predictable quality and a constant year-round supply of feedstock to a biorefinery facility, grass usually needs to be harvested and stored as silage. Preservation is achieved by the combination of an anaerobic environment and the bacterial fermentation of sugar, the lactic acid produced from the latter process lowering the pH and preventing the proliferation of spoilage microorganisms (Muck, 1988). The main objective of ensilage is the efficient preservation of the crop at an optimum growth stage for later use during seasons when the fresh crop is unavailable (McDonald et al., 1991). Ideally crops for preservation as silage should have an adequate content of fermentable substrate in the form of water-soluble carbohydrates, a relatively low buffering capacity and a DM content above 200 g kg⁻¹ (Buxton and O’Kiely, 2003).

A reduction in water soluble carbohydrate concentration during ensiling is generally the most evident change in herbage chemical composition and indicates its substantial use as a substrate for fermentation (King et al., 2013a). Water-soluble carbohydrates are available for fermentation by a variety of microorganisms, of which lactic acid bacteria are the most important. Under good ensiling conditions, lactic acid bacteria become the dominant microbial population, producing mainly lactic acid as a fermentation product with a consequent decrease in pH (McEniry et al., 2008). In general, silage which has undergone a desirable fermentation is characterised by a low pH, high lactic acid content and low concentrations of butyric acid and ammonia-N (Haigh and Parker, 1985). Under sub-optimal ensiling conditions, a secondary clostridial fermentation may lead to considerable DM and energy losses due to extensive production of CO₂ and H₂ from the fermentation of lactate and hexose sugars, and also extensive degradation of proteinaceous compounds to ammonia-N (McDonald et al., 1991).

Harvesting of a forage crop is followed by rapid and extensive proteolysis (Muck, 1988). Degradation of plant protein during wilting and ensiling is inevitable and results in changes in the N-constituents of the ensiled herbage. Two stages of protein degradation can be considered. Firstly, peptide bond hydrolysis takes place releasing free amino acids and peptides, and these amino acids can subsequently be degraded to a variety of end products including NH₃, organic acids and amines. Ammonia and amines are largely the end products of microbial activity rather than plant enzyme activity (Rooke and Hatfield, 2003). During wilting and in well-fermented silages dominated by lactic acid bacteria, proteolysis is largely a result of plant enzymes (Woolford, 1972). Proteolysis can be minimised with a rapid wilt under dry conditions (Carpintero et al., 1979). In general, the extent of proteolysis during ensiling depends largely on the herbage DM concentration and the rapidity with which acid conditions are established (McDonald et al., 1991).

Furthermore, small increases in the relative proportions of neutral detergent fibre, acid detergent fibre, ash and nitrogen can also occur during ensiling as a result of a loss in organic matter during fermentation and through effluent production (McEniry et al., 2013b). For some silages, a decrease in the hemicellulose fraction is also observed, suggesting the hydrolysis of hemicellulose by plant enzymes at the early stages of ensiling or acid hydrolysis during longer term storage, and this can provide additional substrate for lactic acid fermentation (McDonald et al., 1991).

11.3.4 Wet fractionation

The majority of the technological options for a grass-based biorefinery involve the essential primary process of fractionating green plant biomass into a fibre-rich press-cake and a nutrient-rich press-juice. The press-cake fraction represents mainly the cell wall fraction of the herbage, which is rich in cellulose, hemicellulose and lignin. The press-juice fraction represents mainly the cell contents and therefore contains proteins, water-soluble carbohydrates, organic acids, minerals and other substances (Kamm and Kamm, 2004).

Mandl et al. (2005, 2006) stated that the overall goal of fractionation is to transfer as much of the valuable components (e.g., lactic acid, protein) as possible into the press-juice fraction, and reported recovery rates of 85–95% for lactic acid and 55–65% for crude protein from clover-grass and lucerne silages. The process of fractionation involves the mechanical dehydration or pressing of plant biomass to remove water from the plant structural framework, with or without an additional hydrothermal conditioning step to enhance the removal of plant solubles (King et al., 2012b). Prior to pressing, the plant material is generally chopped to reduce particle size, to increase flow rates through the press and reduce blockages, and to aid in the release of plant cell contents. Screw presses, which provide a high degree of maceration of cell walls as a result of axial movement and abrasion of the tissue under high pressure, have been used with some success. The use of less effective piston and roller presses to dehydrate plant material has also been described (Carlsson, 1982).

Wachendorf et al. (2009) and Richter et al. (2009) described the process of hydrothermal conditioning of silages, in which water and silage were mixed (4 : 1 ratio) and heated to 60°C under continuous stirring for 15 minutes. This treatment aimed to macerate cell walls and produce a ‘mash’ which could be mechanically pressed to allow the transfer of soluble minerals (e.g., K, Mg, P and Cl) and organic compounds (e.g., carbohydrates, proteins and lipids) into the press-juice fraction. Consequently, there was a reduction in water and mineral content in the press-cake fraction, together with an overall enrichment of the fibrous constituents. McEniry et al. (2012) reported that fractionation resulted in an average reduction in Cl (95%) and K (98%) and greater than 55% of the N and ash in the press-cake fraction relative to the parent material. Mahmoud et al. (2011) and Arlabosse et al. (2011) have also described the use of thermally assisted mechanical dewatering processes to separate alfalfa and spinach into solid and liquid fractions without the addition of water.

11.4.1 Press-juice fraction

The composition of the press-juice fraction will vary considerably between fresh and ensiled herbage. Silage has a different composition to fresh herbage with its N fraction dominated by amino acids instead of proteins, and can contain lactic acid and a range of other products (e.g., acetic acid, propionic acid, butyric acid and ethanol) from the fermentation of sugar during ensiling.

Green juice

The recovery of plant protein is the main focus of green biorefinery concepts based on the processing of fresh biomass (Fig. 11.2; Kamm et al., 2010). Green forage plants such as grass and legumes can have a high crude protein content of up to 250 g kg⁻¹ DM (Lyttelton, 1973). Green crop fractionation has been investigated for the last two centuries as a means to mechanically extract protein from green forage crops in a form that could be used efficiently by non-ruminants. The extracted crude protein can be used for feeding pigs and poultry, thus reducing the dependence on imported protein rich feeds. The remaining press-cake fraction can be fed to ruminants on the premise that green crops contain a much higher concentration of protein than ruminants actually require (Jones, 1977). Thus, the main objective of this partial extraction was often the provision of a more nutritionally balanced crop containing the correct amount of protein for ruminants, with the primary product of economic value being the pressed-crop.

Furthermore, during times of crisis (e.g., Second World War) the high crude protein content of green forage crops stimulated interest in the use of leaf protein concentrate for human nutrition (Carlsson, 1994; McDougall, 1980; Pirie, 1971). The removal of surplus protein prior to feeding green forage crops to ruminants was proposed as a mechanism to alleviate food-supply shortages in Europe. In this case the primary product of economic value was the green juice fraction. More recently, novel protein foods from plants have been proposed as a protein-rich replacement for meat in human diets so as to reduce the strain that intensive animal husbandry practices pose to the environment (Dijkstra et al., 2003; Linnemann and Dijkstra, 2002).

The first industrial process for leaf protein extraction, the Rothamsted process, was developed by Pirie. The procedure was based on heat coagulation (70°C) of protein in green juice (Pirie, 1966). Procedures based on two-step heating of green juice subsequently enabled extraction of protein fractions of different composition (Edwards et al., 1975; Defremery et al., 1973). Fraction II or green fraction leaf protein concentrate consists of a mixture of proteins originating from the chloroplasts and cytoplasm and can be separated from the green press-juice fraction by thermal coagulation (60–70°C) and centrifugation. After removal of the green fraction leaf protein concentrate, the Fraction I or water-soluble white fraction leaf protein concentrate can be separated by a number of methods including thermal precipitation (∼ 85°C), acid precipitation, solvent extraction or membrane filtration (Dijkstra et al., 2003; Koschuh et al., 2004; Edwards et al., 1975; Pirie, 1971).

Kamm et al. (2010) proposed two protein product streams in the Havelland (Germany) green biorefinery demonstration plant (Table 11.1; Fig. 11.3). Firstly, green fraction leaf protein concentrate is produced by thermal coagulation for use as an animal feed. A second protein product stream focuses on the extraction of white fraction leaf protein concentrate, via thermal coagulation and ultrafiltration, with potential for innovative food processing and cosmetic applications. Similarly, the goal of the green biorefinery research project GRASSA in the Netherlands (www.grassnederland.nl) is the recovery of a protein-rich feed product from fresh grass, in addition to the utilisation of the press-cake fraction in the pulp and paper industries.

Table 11.1

Overview of some green biorefinery activities in Europe

Location	Company/partners	Feedstock	Biorefinery products	Reference(s)
Schaffhausen, Switzerland	2B AG	Fresh grass	• Blow-in insulation (2B Gratec®) • Biogas	Grass (2004)
Obre, Switzerland	Biomass Process Solution	Grass silage	• Insulation boards (Gramitherm®) • Biogas	www.bpsag.ch www.gramitech.ch
Utzenaich, Austria	Okoenergie Utzenaich GmbH	Grass silage	• Lactic acid • Amino acids • Biogas	Mandl (2010)
Groningen, The Netherlands	Avebe	Fresh grass	• Protein-rich animal feedstuff • Technical fibre	Hulst, 2002 www.grassanederland.nl
Esbjerg, Denmark	Agroferm	Fresh lucerne	• Lysine additive for pigs and poultry	Thomsen et al. (2004) www.vitalys.dk
Potsdam, Germany	Leibniz Institute for Agricultural Engineering Postdam-Bornim (ATB)	Grass, cereals (e.g. rye wholemeal)	• Lactic acid	Venus and Richter (2007)
Brensbach, Germany	Biowert	Grass silage	• Protein-rich feedstuff for pigs and poultry • Blow-in insulation • Bio-composite materials • Biogas	www.biowert.de
Havelland, Germany	Research Institute Biopos e.V., Agro-Farm GmbH, FMS-Futtermittel GmbH, biorefinery.de GmbH, LINDE KCA, GmbH and Markischer Hof GmbH.	Fresh lucerne and grass	• Green pellets for animal feeding • Protein-rich animal feedstuff • White protein concentrates for R&D • Biogas	Kamm et al. (2010)

Deproteinated brown juice is an inevitable co-product of leaf protein concentrate extraction and will contain protein, water-soluble carbohydrates and minerals. This brown juice has been proposed as a fertiliser, a ruminant feedstuff, a feedstock for biogas production and as a fermentation medium (Worgan and Wilkins, 1977). Anderson and Kiel (2000) and Thomsen et al. (2004) described the use of deproteinated brown juice from the green crop drying industry as a fermentation medium for the production of amino acids and organic acids. For example, the production of lysine would be attractive as it is an essential amino acid used as a feed additive for pig and poultry feed mixtures. This technology is currently being implemented by the Danish company VitaLys (www.vitalys.dk; Agroferm).

Silage juice

Preservation of grassland biomass as silage ensures a more predictable quality and a constant supply of feedstock to a biorefinery facility. The microbial fermentation of water-soluble carbohydrates during ensiling results in the production of lactic acid among a range of other fermentation products. In addition, much plant protein is hydrolysed to peptides and free amino acids during ensiling. Lactic acid and amino acids are currently seen as the key compounds of green biorefinery concepts processing grass silage (Fig. 11.4; Mandl, 2010). Lactic acid is the major fermentation product produced during ensilage, with values ranging from 20 to 40 g kg⁻¹ in well-preserved, extensively fermented grass silage. Lactic acid is widely used in food, pharmaceutical and cosmetic applications, and as a building block for biodegradable plastic (Datta and Henry, 2006). Free amino acid values can range from 2 to 24 g kg^− 1 in well-preserved grass silage. Amino acids are considered as valuable building blocks and can potentially be used for a wide variety of applications in industry including the synthesis of drugs, cosmetics and food additives (Scott et al., 2007).

The ‘Austrian’ green biorefinery concept is based on a decentralised system with small-scale production of grass silage by local farmers and the subsequent separation of lactic acid and various amino acids from the grass silage press-juice fraction (Kromus et al., 2004). Results from Utzenaich highlight the importance of well-preserved grass silage as a starting material with lactic acid as the main fermentation product (Ecker and Harasek, 2010). Downstream processing of the press-juice fraction involves a series of separation technologies to purify high-grade amino acids and food-grade lactic acid including ultrafiltration, nanofiltration, electrodialysis, reverse osmosis and ion exchange chromatography (Fig. 11.4; Ecker et al., 2012; Mandl, 2010; Novalin and Zweckmair, 2009; Thang and Novalin, 2008; Koschuh et al., 2005; Thang et al., 2005; Thang et al., 2004).

11.4.2 Press-cake fraction

The press-cake fraction is rich in cellulose, hemicellulose and lignin and can potentially be used as a raw material for a wide range of applications (Table 11.2). Furthermore, the press-cake fraction is a potential feedstock for thermal combustion (McEniry et al., 2012; Richter et al., 2011; Wachendorf et al., 2009), biogas (Murphy et al., 2011) and bioethanol (Sieker et al., 2011; Neureiter et al., 2004; Koegel et al., 1999) production. In addition, the separated press-cake fraction is a potential feedstock for a lignocellulosic biorefinery and if this lignocellulosic matrix can be successfully depolymerised to obtain smaller molecules (e.g., sugars), a wide range of potential products (e.g., platform chemicals, biofuels) could be realised.

11.4.3 Energy products