C4 Plants as Biofuel Feedstocks: Optimising Biomass Production and Feedstock Quality from a Lignocellulosic Perspective


  • Caitlin S. Byrt,

    1. School of Environmental and Life Sciences, University of Newcastle, University Drive, Callaghan, NSW 2308, Australia
    Search for more papers by this author
  • Christopher P.L. Grof,

    1. School of Environmental and Life Sciences, University of Newcastle, University Drive, Callaghan, NSW 2308, Australia
    Search for more papers by this author
  • Robert T. Furbank

    Corresponding author
    1. CSIRO Plant Industry and High Resolution Plant Phenomics Centre, GPO Box 1600, Canberra, ACT 2601, Australia
      Corresponding author
      Tel: + 61 2 6246 5149; E-mail: Robert.Furbank@csiro.au
    Search for more papers by this author

Corresponding author
Tel: + 61 2 6246 5149; E-mail: Robert.Furbank@csiro.au


The main feedstocks for bioethanol are sugarcane (Saccharum officinarum) and maize (Zea mays), both of which are C4 grasses, highly efficient at converting solar energy into chemical energy, and both are food crops. As the systems for lignocellulosic bioethanol production become more efficient and cost effective, plant biomass from any source may be used as a feedstock for bioethanol production. Thus, a move away from using food plants to make fuel is possible, and sources of biomass such as wood from forestry and plant waste from cropping may be used. However, the bioethanol industry will need a continuous and reliable supply of biomass that can be produced at a low cost and with minimal use of water, fertilizer and arable land. As many C4 plants have high light, water and nitrogen use efficiency, as compared with C3 species, they are ideal as feedstock crops. We consider the productivity and resource use of a number of candidate plant species, and discuss biomass ‘quality’, that is, the composition of the plant cell wall.

  • image


Plants use energy from solar radiation to convert carbon dioxide into sugars and organic compounds which can be harvested and fermented to produce biofuels. In this way, photosynthesis may be exploited to generate liquid transportation fuels. For example, more than 50% of Brazil's transport fuel is comprised of bioethanol (Pohit et al. 2009) made from fermenting sugar extracted from sugar cane (Saccharum hybrid).

Plant based harvesting of solar energy for fuel production is not a recent innovation. The energy within fossil fuels also originally came from solar radiation harvested by plants into chemical energy, biomass. This biomass then decomposed, or was consumed by organisms that subsequently decomposed, over millions of years beneath layers of earth to form reserves of hydrocarbons. The calorific density of these hydrocarbon reserves is, of course, much higher than new plant biomass due to epochs of geological action. At present, there is a need for specialised cropping systems that will enable us to generate and harvest enough energy from crops to replace or at least partially substitute for the energy that we have been accustomed to harvesting from finite fossil resources; without the steps that takes millions of years.

There are limitations that will influence the types of energy-cropping systems that will be sustainable. The main limitation is the amount of land area and agronomic resources than can be made available for energy crops, without compromising food production. Thus, maximising the energy that can be generated from energy crops per unit land area, whilst minimizing the costly inputs, such as water and fertilizer and fuel, is a priority. This optimisation process for biofuel feedstocks can be divided into two components; biomass quantity and quality. Biomass quantity must be maximised, expressed in tonnes of dry plant material per unit land area and quality, in terms of litres of final fuel produced per unit biomass. In the final analysis, KJ of energy produced per KJ of energy input per unit land area is probably the most economically relevant yard stick for consideration (see Hammerslag et al. 2006).

There are a number of plant species that generate high yields of biomass with minimal inputs; many of these are C4 grasses. Grass species with C4 photosynthesis, such as Aleman grass (Echinochloa polystachya), Elephant grass (Pennisetum purpureum), fox tail millet (Setaria italica), miscanthus (Miscanthus giganteus), sweet sorghum (Sorghum bicolor), sugarcane and switchgrass (Panicum virgatum), are ideal energy crops because they possess the following traits: high conversion efficiency of light into biomass energy, high water use efficiency and high leaf level nitrogen use efficiency (Taylor et al. 2010), capacity to grow in marginal land areas, and a relatively high tolerance to soil constraints such as salinity and water-logging. Grasses with C4 photosynthesis possess an additional CO2 concentrating mechanism that enables them to outperform C3 species, particularly under high temperature and light conditions and are, therefore, capable of generating larger quantities biomass, even in resource limited environments. Plants with C3 photosynthesis such as poplar (Populus) and willow (Salix) also generate high yields of biomass, but take longer to grow and have higher contents of lignin, making the polysaccharides less accessible, thus the biomass quality is lower.

Plant species with C4 photosynthesis may have further advantage over C3 species in some of the types of climate scenarios that have been projected for the future as a result of global warming. In many areas a greater incidence of drought and higher temperatures are expected (Sheffield and Wood 2008). Grass species with C4 photosynthesis are well suited to those areas where the average temperature will increase and where the incidence of drought is expected to increase.

Discussion of what types of C4 plants may be used as bioenergy crops has been limited. Miscanthus (Heaton et al. 2004; Dohleman and Long 2009) and switchgrass (Schmer et al. 2008) have been the focus for considerable research and development activities in Europe and the USA. However, many promising energy crops appear to have been overlooked, in particular, variants of sugarcane and sweet sorghum. The genetic complexity and narrow genetic base of sugarcane presents considerable challenges (Nair et al. 1999). These issues are not pertinent to the diploid sorghum. Most of the literature about sorghum focuses on grain sorghum varieties, but the difference in potential energy yields between grain and sweet sorghum varieties is significant. There is enormous genetic diversity within Sorghum bicolor varieties, and some sweet sorghum varieties have been reported that produce similar yields to sugarcane (Ratnavathi et al. 2010). Sweet sorghum also competes well with miscanthus and switchgrass in regards to biomass yields (Propheter et al. 2010). Some sorghum varieties and species also ratoon and some are rhizomatous (Pritchard 1964). Extensive trials comparing the top varieties of candidate energy crops are needed to compare the energy yields and resources required.

The chemical energy contained within biomass may be harvested by conversion to bioethanol or by conversion to alternate fuels, by direct combustion or by pyrolysis. In the context of bioethanol production, high quality biomass refers to a composition that can be easily and cheaply converted to liquid transport fuels. That is, the maximum accessible yield of firstly, monosaccharides and disaccharides, and secondly, easily extracted polysaccharides. Large quantities of polysaccharides, such as cellulose, contribute to biomass ‘quantity’ but biomass ‘quality’ is also important. Cellulose may be bound within lignin, and thus, inaccessible to processing. Lignin content, composition, and also the type of bonds between lignin, hemicellulose and cellulose are factors that influence biomass ‘quality’. Below we review the current knowledge of photosynthesis, sugar and biomass production, and resource use efficiency of a number of key existing and potential energy crops, in the context of bioethanol production.

Photosynthetic Efficiency, Radiation use Efficiency and Biomass Production in C4 Plants

The “yield equation” describing plant biomass production is well known to all crop physiologists; biomass production is a function of radiation use efficiency multiplied by light intercepted. Light interception is largely defined on a land area basis by the architecture of the canopy and the planting density. In most highly bred crop species, leaf area distribution and angle have undergone extensive genotypic selection and in crops such as maize, rice and wheat can be regarded as being close to optimal. For biofuel feedstock crops, such as miscanthus and switchgrass, this process has not occurred and improvements could be made through breeding for erect canopies and leaf area distribution in the canopy. However, in comparison to C3 species, C4 plants have an immediate and attractive advantage in terms of radiation use efficiency (RUE).

C4 plants utilise a photosynthetic mechanism whereby CO2 is pumped into specialised cells surrounding the vascular bundles, where Rubisco is exclusively localised, and CO2 can accumulate to levels in excess of ten fold atmospheric concentrations in these cells (see Furbank 1998). This biochemical pump requires both biochemical specialisation through cell specific gene regulation, and morphological specialisation in the form of “Kranz” anatomy (Figure 1A). Atmospheric CO2 from the intercellular spaces of the mesophyll cells is fixed by PEP carboxylase into C4 acids which move to the bundle sheath cells to be decarboxylated and release CO2 (Figure 1B). Three distinct mechanisms have evolved to achieve this biochemical CO2 pump (NAD-ME, NADP-ME and PCK types; von Caemmerer and Furbank 2003), and C4 photosynthesis is believed to have evolved more than 40 times in diverse genera of both grasses and dicots (Sage 2004).

Figure 1.

C4 photosynthesis mechanism.
A fluorescence micrograph of a transverse section of a maize leaf (above) after staining with acridine orange to visualise chloroplasts and a simplified scheme of C4 photosynthesis (below), illustrating the morphological and biochemical specialisation required for the C4 mechanism. CO2 is converted to HCO3 which is fixed by phosphoenolpyruvate (PEP) carboxylase to a C4 acid product, transported to the bundle sheath cells and decarboxylated to a C3 product, then recycled to the HCO3 acceptor PEP in the mesophyll. CO2 concentrations of up to 10 fold ambient atmospheric concentrations are built up in the bundle sheath cells by this biochemical pump, where Rubisco and the photosynthetic carbon reduction cycle (PCR) are located. Glycine decarboxylase, the step in photorespiration where CO2 is released under low CO2 conditions (indicated by a star), and the photosynthetic carbon oxidation cycle (PCO) are located exclusively in the bundle sheath compartment, maximising refixation in the event of photorespiratory conditions.
Adapted from Furbank et al. (2009).

The high RUE of C4 plants is directly related to this CO2 concentrating function as photorespiration is almost totally inhibited in C4 plants and Rubisco operates at close to its maximum capacity. The efficiency of conversion of total solar energy to biomass in a C4 plant is approximately 3.7% while in a C3 cereal, this value is only 2.4% (Zhu et al. 2010). This 60% increase in photosynthetic efficiency, if translated into fermentable biomass gives C4 plants a considerable advantage over C3 plants as biofuel feedstocks. The CO2 concentrating mechanism also results in large improvements in leaf level nitrogen use efficiency and transpiration efficiency. Leaf nitrogen use efficiency is high in C4 leaves because they contain less photosynthetic proteins, specifically; they contain only around 60% of the Rubisco of a C3 leaf (see Taylor et al. 2010). Transpiration efficiency is generally higher in C4 plants than C3 plants because they can maintain a lower intercellular CO2 concentration and, therefore, a greater diffusion gradient (Long and Ort 2010), due to the high affinity of PEP carboxylase for bicarbonate (see Furbank et al. 2009), thus C4 plants can fix CO2 with stomata open for a shorter time.

C4 plants are among the most productive plants on the planet, both as agricultural crops and the worlds’ worst weeds (Hatch 1987). This superiority is illustrated in Figure 2, in a plot of final dry harvestable biomass for a number of C3 and C4 species under their optimal growing conditions. This plot is particularly informative as it clearly illustrates that the length of the growing season has a considerable influence on the advantage C4 grasses have over C3 crops. A similar comparison is shown in Figure 3; this figure is focusing on C4 grasses without consideration of seasonal duration but illustrates that even within C4 species there is considerable variation in biomass production. It is also apparent from these data and other reports for C4 grasses of standing biomass above 80 tonnes/ha (Somerville et al. 2010) that switchgrass and miscanthus are among the poorer performers on the biomass list.

Figure 2.

Standing dry weight at harvest for a variety of C3 and C4 crops plotted against length of growing season.
Adapted from Monteith (1978).

Figure 3.

A comparison of peak dry matter yields of field grown C3 and C4 potential biofuel species grown under optimal conditions in the field.
Adapted from El Bassam (1998).

While the advantages of the C4 mechanism on RUE are well accepted, the relationship between maximum photosynthetic rate and seasonal biomass accumulation has not been extensively explored. One of the few studies across a number of phylogenetically diverse C4 plants was carried out by Ghannoum et al. (2001). This study, while limited to pot grown material, indicates the large interspecific variation in biomass accumulation across C4 species and also indicates that there is little correlation between biomass accumulation and the two photosynthetic sub types examined (although NADP-ME types, the predominant C4 crop type, were superior overall relative to NAD-ME types; Figure 4). In fact, in this work, the correlation between above ground biomass at harvest and maximum photosynthetic rate was represented by an R2 value of only 0.3. This suggests that genetic variation in partitioning and the role of rhizomatous tissues in C4 grasses requires much more research before we understand how to select superior C4 biofuel feedstocks from photosynthetic parameters alone.

Figure 4.

Dry mass of C4 grasses of NAD-ME and NADP-ME types in winter or summer.
For winter (black bars) n= 3; for summer (hatched bars) n= 4. The numbers next to the means are the summer to winter ratios. Adapted from Ghannoum et al. (2001).
Species names:
NAD-ME types:
Astrebla lappacea (Lindl.) Domin; Astrebla pectinata (Lindl.) F. Muell. ex Benth.; Astrebla squarrosa C.E. Hubbard; Cynodon dactylon (L.) Pers.; Enteropogon acicularis (Lindl.) Lazarides; Eleusine coracana Gaertn; Eragrostis setifolia Nees; Eragrostis superba Peyr.; Leptochloa dubia (Kunth) Nees; Panicum coloratum L.; Panicum decompositum R.Br.
NADP-ME types:
Bothriochloa biloba S.T. Blake; Bothriochloa bladhii (Retz.) S.T. Blake; Bothriochloa pertusa (L.) A. Camus; Cenchrus ciliaris L.; Cymbopogon ambiguus A. Camus; Cymbopogon bombycinus (R. Br.) Domin; Dichanthium aristatum (Poir.) C.E. Hubb.; Dichanthium sericeum (R.Br.) A. Camus; Digitaria brownii (R&S) Hughes; Digitaria smutsi Stent; Echinochloa utilis Ohwi & Yabuno; Paspalum dilatatum Poir.; Panicum antidotale Retz.; Pennisetum alopecuroides (L.) Spreng.; Pennisetum clandestinum Hochst.ex Chiov.; Sehima nervosum (Rottler ex Roem. & Schult.) Stapf; Themeda triandra Forssk.

Saccharides for Fermenation

The sources of saccharides used for fermentation to produce bioethanol include soluble sugars, starch and structural sugars. The main soluble sugar harvested from plants is sucrose, and at present, the dominant crop grown for global sucrose production is sugarcane (approximately 75% of the worlds supply; Wu and Birch 2007). The biggest source of starch for bioethanol production is maize kernel and grain from other cereals. Structural polysaccharides, cellulose and hemicellulose, may be hydrolysed to produce monosaccharides for fermentation. Cellulose (D-glucose linked by β-1,4-glucosidic bonds) is the most abundant source of polysaccharides on Earth (Bayer et al. 1998). Crops considered as possible feedstocks for cellulosic ethanol production include grasses such as maize, switchgrass, miscanthus, sugarcane, sweet sorghum, and wood-chips from various C3 tree species such as poplar and willow.

At present, the cost of harvesting and processing sugar cane juice and maize kernel starch, by crushing stems to extract juice and milling grain followed by saccharification, respectively, is relatively low compared to the cost of harvesting and processing lignocellulosic biomass. Processing of lignocellulose is expensive due the energy and or enzymatic costs involved in separating cellulose from lignin, and the enzymatic costs of hydrolysing the cellulose. The source of the carbon atoms present in lignin polymers and cellulose polymers is the same, that is, carbon assimilated by photosynthesis; however, lignin is not useful as a carbon source for bioethanol production. This is because lignin is made from aromatic structures and the reactions to synthesize lignin are not reversible; whereas cellulose can be easily hydrolysed to release the six-carbon glucose monomers.

Processes to separate lignin from cellulose are costly, and include mechanical grinding; high-temperature treatments (such as pyrolysis >300°C); ammonia, carbon dioxide or steam explosion (autohydrolysis); acid or alkaline hydrolysis; oxidative delignification (with peroxide); lime treatment or treatment with fungi that produce lignin-degrading enzymes (such as peroxidases; Sun and Cheng 2002; Jorgensen et al. 2007). More recently use of ionic liquids to convert biomass has been investigated (Palkovits et al. 2010; Zhao et al. 2010).

The energy needed to produce cellulosic ethanol must be significantly lower than the energy yielded for the process to be viable. Yet, in terms of energy out verses energy in, cellulosic ethanol production far out-performs production of ethanol from corn starch. The energy out to energy in ratio of corn is 1.4–2.3, where as for switchgrass the energy out to energy in ratio is 10.8–11.3 (Vadas et al. 2008), and for miscanthus the ratio of energy output to input was 22, without fertilizer or irrigation, and 47 with fertilizer and irrigation (Ercoli et al. 1999). For traditional sugar crops; the energy balance of sugarcane, in terms of energy output verses fossil energy needed for production, was reported as 8.1–10, and sugar beet was reported as 2 (Boddey et al. 2008; Sanchez and Cardona 2008; Goldenberg and Guardabassi 2010). Previous studies report the energy out to energy in ratio for sugar cane as 21.3, and the ratio for sweet sorghum as 21.7 (Silva and Serra 1978). However, there are major inconsistencies within the literature where the energy balance data has been calculated, in many cases because energy inputs have been neglected or contributions to energy balance from recycling of lignin waste has not been included (Farrell et al. 2006; Hammerschlag 2006).

Soluble Sugar and Starch Production

As bacteria, yeast and fungi are prevalent in nature, it is no surprise that plants tissues, such as mature sweet fruits, sweet stems and cereal grains, where high concentrations of sugar or starch accumulate, readily ferment and generate ethanol. Humans have exploited this to make wine, beer and bread, among many other foods and drinks, possibly since 5000BC (Cavalieri et al. 2003). Ethanol has been used as a transport fuel, for internal combustion engines, since 1908 (DiPardo 2007). At present, global production of ethanol exceeds 66 billion litres per year (Balat and Balat 2009; Goldenberg and Guardabassi 2010), and the main feedstocks are sugar cane (60%), maize grain, sorghum grain, wheat grain and sugar beet (Balat and Balat 2009). It is possible that eight times this volume of ethanol could be produced if lignocellulosic material were used (Kim and Dale 2004). However, in the short term, the replacement of 10% of the gasoline used in the world by ethanol from sugarcane (Goldenberg and Guardabassi 2010), and or other crops such as sweet sorghum, is possible.

The main sugar yielding crops are sugar cane, sugar beet and sweet sorghum (Table 1). Of these, sweet sorghum may be under exploited relative to the sugar yield potential of this crop. Calculated ethanol yields from sweet sorghum stem juice, ∼10 000 L of ethanol per ha (Propheter et al. 2010), may exceed that of sugarcane and sugar beet, depending on the genotype and growing environment (Table 1). Sweet sorghum is a tropical crop that thrives in high-light and high-temperature environments; however, sweet sorghum sugar has also been shown to yield 5 235 L of ethanol per hectare in temperate areas (Smith 1993). Sweet sorghum generates high sugar yields over a wide range of geographical locations, from four tonnes/ha in cooler areas, up to 12 tonnes/ha of sugar in warmer climates (Smith et al. 1987). Sugar cane and sweet sorghum tend to yield best in warm climates where as sugar beet is well adapted to cooler climates.

Table 1.  Sugar contents and yields from sugar crops
CropGrowth periodWater requirement m3 ha−1 (Almodares and) (Hadi 2009)Brix (%)Sugar g per kg tissue (stem or beet)Sugar yield tonnes ha−1Calculated ethanol yield (L ha−1) from sugar only (not biomass)Reference
  1. aFour to six varieties were evaluated over four years and eight locations.

Sugar beet (Beta vulgaris)5–11 months18 0001548–80 (beets)7.2–124 000–7 000FAO (2006)
15–18% 6.2–9.1 Kaffka and Hills (1994)
   5 140Balat and Balat 2009
   5 500 (Europe)Goldenberg and Guardabassi 2010
  7.3–8.8 Renouf et al. 2008
   6 600Poitrat 1999
Sugar cane (Saccharum hybrid)12 months36 00013.7∼400 (stem)  Lingle et al. 2009
    10.4–17.4 Renouf et al. 2008
    15 Jackson 2005
     7 000Matsuoka et al 2009
     6 280Boddey et al. 2008
     6 000Lee and Bressan 2006
     6 470Goldenberg and Guardabassi
     5 345–9 381Sanchez and Cardona (2008)
    7–83 000–5 000Almodares and Hadi 2009
     6 641Balat and Balat 2009
Sweet sorghum (Sorghum bicolor)4–6 months12 00016–23%   Reddy et al. 2005
  16–18% 5.4–10.4 Wu et al. 2010
   50310.55 414Zhao et al. 2009
   (stem) (China) 
    9 Murray et al. 2008
    4–122 129–6 388Smith et al. 1987a
  11–23% 6–83 000Almodares and Hadi 2009
     10 000 (Taiwan –Liu and Lin 2009
     2 crops/yr) 
    138 000Hunter and Anderson (1997)
      sited by Bennett and Anex (2009)

Crop duration also impacts final potential yields. In warmer climates it is possible to harvest two crops per year from sweet sorghum; where as only a single sugar cane crop can be harvested in a year. The second sorghum crop need not require a second planting, as sorghum produces what is called a ratoon crop. After sorghum stems are chopped off, the base of the plant will grow a second set of shoots from which similar sugar yields may be harvested. Sweet sorghum may also prove to be cost effective relative to sugar production because sweet sorghum has a lower fertilizer requirement than sugarcane and pest and disease management is less complex for sweet sorghum (Almodares and Hadi 2009).

Ethanol yields from C4 sugar crops, such as sugarcane and sweet sorghum (around 8 000 L/ha), exceed the ethanol yields from starch from grain crops (Table 1). Ethanol yields from starch from corn starch is in the order of 1 140 L/ha (Lee and Bressan 2006) to 4 180 L/ha (Balat and Balat 2009; Goldenberg and Guardabassi 2010). Ethanol yields from wheat starch are in the order of 1 075–1 730 L/ha (Balat and Balat 2009). Furthermore, food grain must be prioritized for human consumption rather than fuel production.

Sweet sorghum has such high levels of sugar in the stems that ground, fresh or dry stem may be fermented directly; referred to as solid state fermentation. This process has been shown to yield 7.9 g of ethanol per 100 g of fresh stalks (100 L ethanol per tonne of stalks, where 0.46 g ethanol was generated per g sugar), which is 90% of the theoretical yield (Yu et al. 2008). The maximum theoretical yield from fermentation is 0.51 g of ethanol per gram of glucose. Solid state ethanol production can also overcome problems with the short shelf life of high-sugar biomass, as dry stem tissue can be stored for up to eight months (Shen and Liu 2009). Sweet sorghum biomass has also been used to generate hydrogen and methane; Antonopoulou et al. (2008) suggest sweet sorghum as an ideal substrate as they demonstrated anaerobic digestion of sorghum residues may generate 10.4 L H2/Kg sweet sorghum, and yields of 78 L of methane/Kg of sweet sorghum residue.

Sugar yields of sweet sorghum and sugar cane may be increased through transgenic approaches. One promising approach induces the transgenic plant to synthesise sugars that cannot be metabolised by plants. In sugar cane, sugar yields were doubled by introduction of a bacterial isomerise enzyme that converts sucrose into isomaltose (Wu and Birch 2007). The high sugar yielding lines also exhibited increased photosynthesis, sucrose transport and sink strength (Wu and Birch 2007).

Calculated Ethanol Yields

Biomass production (tonnes/ha/yr) may be the single greatest factor limiting global lignocellulosic ethanol yields. However, as the performance of plant species is influenced by the soil and climate the highest yielding plants may not be suitable for all locations and tailoring biofuel feedstock crops to microclimates may be important. Some environments will support the production of biomass in forestry systems and others may support grassland systems. Therefore, it is likely that many different crops may be adopted in different areas to supply biomass for lignocellulosic bioethanol production.

Corn stover, miscanthus, switchgrass and willow are currently being used as biomass feedstocks in trial lignocellulosic bioethanol studies (Li et al. 2010; Van Hulle et al. 2010). However, many promising plants with greater biomass production, such as Echinochloa polystachya (Table 2) are yet to be tested as lignocellulsic feedstock crops. Echinochloa polystachya may produce ten times the quantity of biomass per year as willow (Table 2).

Table 2.  Biomass yields
CropApproximate growth periodFresh weight tonnes ha−1Dry weight tonnes ha−1Calculated Ethanol yield from biomass (L ha−1)Reference
  1. aFour to six varieties were evaluated over four years and eight locations.

Aleman grass (Echinochloa polystachya)12 months 99 (South America) Piedade et al. 1991
  100 Somerville et al. 2010
  (per year)  
Cassava (Manihot esculenta)9–24 months 204 500Lee and Bressan 2006
 6.1–21.4  Cock 1982
Corn (Zea mays spp.)6s months10.6–23.5 (Iowa-Texas, USA)  Rooney et al. 2007
 6.59–11.06 1 500–2 518Kim and Dale (2004)
  20 Dohleman and Long (2009)
 10 3 800Somerville et al. 2010
Elephant grass (Pennisetum purpureum) 50  Kubota et al. 1994
 88  Somerville et al. 2010
Miscanthus (Miscanthus x giganteus)3–4 months 27–44 (Europe) Heaton et al. 2004
  30 Dohleman and Long 2009
  14–26 Amougou et al. 2010
  14–404 600–12 400Somerville et al. 2010
 3.3–12.8 1 035–3 963Propheter et al. 2010
   (Manhattan, U.S.) 
Poplar (Populus trichocarpa)  5 (per year, UK) Cannell et al. 1988
  6–17 Labrecque and Teodorescu (2005)
  (per year, Canada)  
  5–111 500–3 400Somerville et al. 2010
Sugar beet5–11 months47.8–57.3 (UK)  Renouf et al. 2008; Tzilivakis et al. 2005
 63.425 Koga 2008
Sugar cane12 months79–112 (Queensland, Australia)  Renouf et al. 2008
 76.6  Boddey et al. 2008
 75  Matsuoka et al 2009
   5 345–9 381Sanchez and Cardona (2008)
  19 Hattori and Morita (2010)
 80 9 950Somerville et al. 2010
Sweet sorghum4–6 months 32.513 032 (Beijing, China)Zhao et al. 2009
  24.77 013Bennett and Anex (2009)
   (Texas, USA) 
 17.5–31.6  Rooney et al. 2007
 43–15024–25 Smith et al. 1987a
 28.2–32.6 9 656–10 184Propheter et al. 2010
   (Manhattan, U.S.) 
Switchgrass (Panicum virgatum)2–7 months10.3–10.6  Rooney et al. 2007
 1.28–8.92 (Pennsylvania, U.S.) 555–3 871Adler et al. 2006
 4.1–9.2 1 288–2 851Propheter et al. 2010
   (Manhattan, U.S.) 
Willow (Salix viminalis)  10 (per year, UK) Cannell et al. 1988
  17–18 (per year, Canada) Labrecque and Teodorescu (2005)

Lignocellulosic Feedstock Quality

Biomass recalcitrance, the natural resistance of plant cell walls to microbial and enzymatic deconstruction, increases the cost of conversion of cellulose to glucose. Lignin significantly contributes to biomass recalcitrance. Not surprisingly, tree species such as poplar and willow have higher lignin contents than grasses (Table 3).

Table 3.  Biomass composition of a selection of lignocellulosic feedstock plants
Corn stover35%28%10.4%Karp and Shield (2008)
Miscanthus57.6%15.9%10.5%Karp and Shield (2008)
Poplar40%14%20%Karp and Shield (2008)
Sugarcane24%8%7%Karp and Shield (2008)
Sweet sorghum26.3%20%7.1%Rooney et al. 2007
Switchgrass31.6%36%6.1%Karp and Shield (2008)
Willow55.9%14%19%Karp and Shield (2008)

Greater energy (temperature and pressure), and or quantity of enzymes (cellulases), are needed to hydrolyse cellulose that is embedded within lignin. In addition to blocking the liberation of sugars from cellulose and adhering to hydrolytic enzymes, lignin may also release aromatic compounds that inhibit fermentation. Modifying lignin content, composition, hydrophobicity and cross-linking can improve the enzymatic hydrolysis of cell walls.

Lignin content and composition may vary due to natural mutation in the genes involved in the lignin biosynthesis pathway, such as observed for brown-midrib (bmr) mutant plants. Bmr mutants produce a lignin which differs to the lignin of normal plants, this results in a red-brown colour which is seen in the mid vein of the leaves, mutant plants have significantly less lignin than normal plants (Oliver et al. 2005). Lignin may also be altered by selective plant breeding and transgenic approaches. Approaches to reducing lignin have targeted each of the steps in phenylpropanoid metabolism (Vanhholme et al. 2010a). In general, reduction of the mRNA transcript levels of the genes involved in lignin biosynthesis reduces lignin content (see Table 3 of Baucher et al. 2003 for a summary), with the exception of ferulate 5-hydroxylase, which may increase lignin content when down-regulated (Reddy et al. 2005). Increasing mRNA transcript levels of many lignin biosynthesis genes may also increase lignin content (Baucher et al. 2003). In many cases, altered regulation of lignin biosynthesis genes influences the ratios of p-hydroxyphenyl (H), guaiacyl (G) and syringyl (S) lignin monomers, and it may be possible to improve the digestibility of lignin by modifying the monomeric composition without the need to reduce the total lignin content (see Simmons et al. 2010). For example, transgenic poplar plants with higher S lignin due to overexpression of the ferulate 5-hydroxylase gene digest more easily than wild type plants (Stewart et al. 2009). However, some plant species may have mechanisms to compensate for changes in lignin content and composition, such as increasing the amount of cross-linking, which may increase recalcitrance (Grabber et al. 2005).

Increasing the biosynthesis of ferulic acid, and its export to the cell wall, has been suggested as a strategy to increase the value of bio-energy crops (Vanholme et al. 2010b). For example, transgenic poplar plants with defective Cinnamoyl-CoA reductase had elevated ferulic acid levels which were associated with improved saccharification potential relative to wild type poplar (Leplé et al. 2007). Incorporation of ferulic acid into the lignin monomer may lead to acetal bonds that are easily cleaved in pre-treatment processes to degrade lignin (Vanholme et al. 2010b). Sweet sorghum naturally contains a significant quantity of ferulic acid; 33 and 44 μmol/g in pith and bark, respectively (Billa et al. 1997); thus, it would be of interest to investigate the relationship between ferulic acid content and cell wall digestibility in a range of sorghum lines differing in ferulic acid content.

Use of Resources

The most sustainable systems for growing energy are those that maximise the production of photoassimilate and minimise the requirement for valuable inputs such as nitrogen, phosphorus and water. Furthermore, an ideal energy-cropping system would be one which increases soil organic carbon over time, by sequestering carbon and minimising the net green house gas emissions. Boosting soil organic carbon over time would have the additional benefit of improving the soil structure and nutrient availability.

Nitrogen and Phosphorus Use

A low nitrogen requirement, nitrogen use efficiency and a high C:N ratio in biomass are ideal traits for biofuel feedstock crops. A low nitrogen requirement and high nitrogen use efficiency is beneficial as the costs of supplying fertilizer can be minimized, and a low biomass N concentration reduces dry matter loss during storage (Heaton et al. 2009), and contributes less to greenhouse gas emissions (Karp and Shield 2008). Less than 0.6% N may be optimal for biomass crops (Kauter et al. 2003). C4 species tend to have lower shoot N concentration (higher N-use efficiency) values than C3 species, which is related to a lower content of photosynthetic proteins, as C4 leaves have 60% of the Rubisco of C3 leaves (Gastal and Lemaire 2002; Levang-Brilz and Biondini 2003; Taylor et al. 2010).

Perennial rhizomatous grasses can translocate nutrients from aboveground organs to belowground organs. This trait may be valuable in a biofuel feedstock system as it decreases the loss of nutrients from the system when aboveground parts of the plant are harvested. Use of remobilized leaf N for growth in the subsequent season minimizes the need for additional N from fertilizers. Furthermore, fuel costs may be reduced in systems where a second or third crop can be harvested from a single planting of seed.

Perennial grasses remove less N, P and K than high yielding annual crops (Propheter and Staggenborg 2010). N loss and erosion estimates for switchgrass systems were very favorable over corn systems. Switchgrass production systems lost less N (6–8 kg N/ha) to leaching than corn systems (9–16 kg N/ha), had less denitirification (9–14 kg N/ha for switchgrass verses 24–39 kg N/ha for corn), and switchgrass systems would cause a small fraction of the erosion (78 kg/ha) as compared to corn systems (2 307 kg/ha) (Vadas et al. 2008). Commercial sweet sorghum cultivars tend to remove similar quantities of N and P as corn crops; however, sweet sorghum produces 25 to 50% greater total biomass for the same nutrient removal (Propheter and Staggenborg 2010). Overall, miscanthus and switchgrass remove much less N, P and K than corn and sorghum (Propheter and Staggenborg 2010).

Recent studies have investigated the optimal harvest time for miscanthus and Panicum virgatum to minimize the N removed with harvesting of above ground biomass (Heaton et al. 2009; Amougou et al. 2010). In general nitrogen content aboveground is highest mid-season and as plants senesce N is translocated to the crown, rhizomes and roots (Karp and Shield 2008). The C:N ration of M. x giganteus was 142.6 and P. virgatum was 95.9, and overall P. virgatum had a higher [N] (Heaton et al. 2009). Miscanthus is particularly efficient at reusing N; three seasons after application of 15N labeled fertilizer 14% of the labeled N was still found in the rhizome (Christian et al. 2006). However, fertilizer application is required for miscanthus if soils are very nutrient poor, and after two years of continuous planting; applications are typically 40–100 kg/ha N, 10–20 kg/ha P and 40–100 kg/ha K (Heaton et al. 2004).

Perennial sorghum has been developed by crossing Sorghum bicolor with Sorghum halepense, a species of Sorghum which is rhizomatous (DeHaan et al. 2005). Sorghum almum and S. halepense are also rhizomatous (Pritchard 1964). Continuous cultivation of sorghum increases soil organic carbon, thus, Sorghum sp. may be useful for carbon sequestration purposes (Wang et al. 2010). Sorghum bicolor had a C:N ratio of 63 (Wang et al. 2010).

Photosynthetic phosphorus (P) use efficiency in C4 species increases as P becomes more limiting, to a greater extent than in C3 species (Ghannoum and Conroy 2007). Thus, in P limiting environments C4 species may be better able to sustain photosynthesis, and thus, production of photoassimilate, than C3 species.

Water Use

Evapotranspiration causes water loss; water loss can be reduced by lowering stomatal conductance, but this compromises CO2 capture. As mentioned previously, C4 plants tend to have higher water use efficiency because of the ability of C4 plants to fix CO2 with less open stomata by virtue of the high affinity of PEP carboxylase for bicarbonate (see Furbank et al 2009; Long and Ort 2010); thus, enabling C4 plants to maintain a lower intercellular CO2 concentration and, therefore, a greater diffusion gradient (Long and Ort 2010). The water use efficiency of C3 plants is 2–3 g dry mass/kg H2O (Kramer and Boyer 1995), which is one third to one half of the efficiency of C4 plants, many of which may achieve 4–6 g dry mass/kg H2O (Figure 5).

Figure 5.

Water use efficiency of C4 grasses of NAD-ME and NADP-ME types in winter or summer.
For winter (black bars) n= 3; for summer (hatched bars) n= 4. The numbers next to the means are the summer to winter ratios Adapted from Ghannoum et al. (2001).
Species names:
NAD-ME types:
Astrebla lappacea (Lindl.) Domin; Astrebla pectinata (Lindl.) F. Muell. ex Benth.; Astrebla squarrosa C.E. Hubbard; Cynodon dactylon (L.) Pers.; Enteropogon acicularis (Lindl.) Lazarides; Eleusine coracana Gaertn; Eragrostis setifolia Nees; Eragrostis superba Peyr.; Leptochloa dubia (Kunth) Nees; Panicum coloratum L.; Panicum decompositum R.Br.
NADP-ME types:
Bothriochloa biloba S.T. Blake; Bothriochloa bladhii (Retz.) S.T. Blake; Bothriochloa pertusa (L.) A. Camus; Cenchrus ciliaris L.; Cymbopogon ambiguus A. Camus; Cymbopogon bombycinus (R. Br.) Domin; Dichanthium aristatum (Poir.) C.E. Hubb.; Dichanthium sericeum (R.Br.) A. Camus; Digitaria brownii (R&S) Hughes; Digitaria smutsi Stent; Echinochloa utilis Ohwi & Yabuno; Paspalum dilatatum Poir.; Panicum antidotale Retz.; Pennisetum alopecuroides (L.) Spreng.; Pennisetum clandestinum Hochst. ex Chiov; Sehima nervosum (Rottler ex Roem. & Schult.) Stapf; Themeda triandra Forssk.

Future climate scenarios predict that higher atmospheric carbon dioxide levels will lead to more days with higher temperatures and more erratic rainfall leading to an increase in the incidence of drought. Generally, C4 grasses have the capacity to generate more biomass than C3 grasses in conditions where the maximum daily temperature is higher (Rubio et al. 2010). Conditions of elevated temperature and CO2 improve yields of sugar and biomass in sugar cane and sorghum, respectively (Prasad et al. 2009; Vu and Allen 2009). The water use efficiency of sweet sorghum increases, exceeding that of maize and grain sorghum, under water stress conditions (Steduto et al. 1997; Conley et al. 2001). The normalized leaf level transpiration efficiency of sorghum was 10.5 μmol/mmol kPa, where as C3 crops averaged around 4.8 μmol/mmol kPa (Steduto and Albrizio 2005).

A recent study investigated the sugar and ethanol yields of sweet sorghum grown under soil salinity and reduced irrigation stresses (Vasilakoglou et al. 2010). Sweet sorghum plants grown in soil salinity of 3.2 dS/m and limited (210 mm) irrigation water yielded 27.1–33.5 t/ha biomass, 2.6–3.86 T/ha total sugar and 4 926–7 620 L/ha theoretical ethanol yields (Vasilakoglou et al. 2010). Thus, this crop may yield sufficient sugar for ethanol production in areas that are not ideal for food production such as arid environments where soil and water may be moderately saline.

Major Areas for Improving Feedstocks

We do not yet have a fully domesticated crop dedicated for energy production (Sang 2011). However, investigation of the natural variation within species of C4 grass for biomass production, cell wall composition and nitrogen and water use efficiency may reveal species with useful traits. For example, there are approximately 4 000 sweet sorghum cultivars distributed throughout the world (Grassi et al. 2004), this number does not include other Sorghum species such as wild rhizomatous grasses that are yet to be exploited, these provide a diverse genetic background from which to develop regionally specific, highly productive cultivars (Bennett and Anex 2009). Sugar cane was originally developed from only two crosses; greater genetic variability may be introduced into sugar cane by re-making sugar cane with alternative Saccharum species or even using wide crossing into miscanthus or other related C4 grasses with desirable attributes (Nair et al. 1999).

Generation of feedstock crops that are optimized for biofuel production may be achieved by selective breeding for natural differences, or by genetic modification. While the search for biofuel relevant traits and genes has been accelerated by the sequencing of the maize and sorghum genomes, verification of gene function by reverse and forward genetics is hampered in these species by their long generation times and large stature. A new C4 grass model species, Setaria viridis, is now being proposed to facilitate the testing of strategies to optimize biomass quantity and quality by way of plant transformation (Brutnell et al. 2010). This species has a small genome, compact stature and most importantly a short lifecycle making it an ideal C4 biofuel feedstock model and it is closely related to the C4 crop species Setaria italica, already a potential candidate for biofuel production (Brutnell et al. 2010).

Future crop productivity may be increased by improving photosynthetic efficiency. Towards this goal Zhu et al. (2010) suggest improving leaf display to avoid light saturation, and engineering carboxylases that are better adapted to forthcoming CO2 concentrations. Using a transgenic approach, the major limiting steps in C4 photosynthesis have been quantitatively elucidated in Flaveria bidentis, a model C4 dicot (see Furbank et al. 1997). From these studies, control of photosynthetic flux in current air levels of CO2 appears to be shared between PEP carboxylase, Rubisco and an enzyme involved in regenerating PEP for PEP carboxylase, pyruvate Pi-dikinase. Increasing amounts of or improving kinetic properties of these enzymes to increase catalytic rate are the most attractive targets (see Matsuoka et al. 2001).

Greater lignocellulosic ethanol yields may be achieved by modifying cell wall composition, for example, altering lignin content and composition. To retain fitness in future low-lignin crops, it may be necessary for lignin to be reduced in only specific tissues, or cell types (Shadle et al. 2007), rather than in the whole plant. For example, it may be beneficial to retain normal lignin synthesis and deposition in tissues such as stomata, roots, and sheath or rind tissue in grasses, but reduce lignin content in areas where maximum cellulose and hemicellulose yields may be harvested, such as in the storage parenchyma tissue within the stems of large-stemmed grasses, such as sweet sorghum. Specifically, altering lignin in different cell types, such as retaining lignin in xylem cell walls but lowering lignin content in storage parenchyma cell walls may improve digestibility without compromising xylem function. By modifying cell wall composition in a tissue specific manor lignin in the sheath may still provide structural support and possibly defence against microbial attack, lignin in stomata may ensure efficient stomatal function and prevent unnecessary water loss, and lignin in xylem may contribute to normal xylem function, but stem parenchyma cell wall cellulose would be more accessible to degradation during processing. An alternative approach may be co-polymerizing alternative lignin monomers into the lignin polymers so as to improve degradation (Vanholme et al. 2010a).

Data directly comparing the performance of potential biofuel feedstock crops under different climate and field conditions is lacking, there is a significant gap in research in this area. Furthermore, these studies need to consider use of water and fertiliser and be undertaken in various climates to assess which crops may yield the most bioethanol in different environments.

Concluding Comments

The data reviewed here, for sugar yields, biomass yields and composition, calculated ethanol yields and water use of a number of crops of interest, was collected from a range of studies and does not provide a direct comparison of the yield potential of these crops in the environments that they may be grown in for biofuel production. Data comparing the top varieties of candidate energy crops in regards to energy yields and resource use in target environments is lacking.

While biofuels are unlikely to replace fossil fuels in our global energy requirements, they will undoubtedly form part of a distributed solution to meeting our energy needs. C4 plants are among the most efficient convertors of sunlight into biomass on the planet and present great opportunities for exploitation of genetic diversity in growth and nutrient use efficiency both within species and between species. So far plant biologists have barely scratched the surface of the rich genetic resources available for improving C4 grasses and plant breeders have only recently begun focussing their activities on improving accepted biofuel crops such as miscanthus and switchgrass, and adapting sorghum and sugarcane to these non-food uses. Considering that crops such as maize have been subjected to thousands of years of agricultural selection during cultivation, C4 biofuel feedstock improvement is only in its infancy.

(Co-Editor: Martin A. J. Parry)


CSB and CPLG are supported by the Australian Research Council (ARC) though ARC-linkage project LP0883808.