• Open Access

Regional examination shows potential for native feedstock options for cellulosic biofuel production


Seth Debolt, e-mail: sdebo2@email.uky.edu


Kentucky, as with many regions around the globe, has a relatively long growing season with significant rainfall that could produce sizeable quantities of perennial herbaceous and woody biomass on land that does not compete with food crops. Additionally, there are limited options for renewable power production from low carbon sources such as solar-photovoltaic, wind and hydroelectric. Recent studies have shown that producing renewable energy from perennial cellulosic crops, as opposed to starch-based biofuel crops, will have a carbon-mitigating outcome. Currently, there is a lack of data regarding regionally suitable genotypes. Herein, we establish baseline values for multiple entry selections of three native C4 grass species, switchgrass (SW) (Panicum virgatum L.), eastern gamagrass (EG) (Trispicum dactyloides L.) and big bluestem (BB) (Andropogon gerardii Vitman). Yield potential examined over 7 years showed that environment, species and entries had a significant impact on yield, but EG had higher total yield over the duration of the study. Cellulosic biofuel potential was examined by measurement of saccharification efficiency, relative lignocellulosic energy density, cellulose content and lignin content during three growing seasons. EG had significantly higher digestibility rate than SW and BB. Underlying this was a negative correlation between lignification and saccharification efficiency. However, higher lignin content and higher cellulose content among SW entries resulted in higher energy density relative to EG and BB. These data reveal that locally bred EG varieties were most suited to cellulosic ethanol production under the growing conditions of central Kentucky, USA, compared with SW and BB and suggest the importance of regional examination.


national renewable energy laboratory


filter paper unit


degrees of freedom


eastern gamagrass




big bluestem

anova =

analysis of variance


Conversion of plant biomass into fermentable sugars and subsequent conversion to alcohol-based fuel offers renewable transportation fuel that can be employed in the majority of current vehicles. Plants also represent a combustible supplement to burning fossil fuels alone (coal) for electricity generation. During photosynthesis, solar energy is converted to chemical energy using atmospheric carbon to synthesize an array of carbon-based metabolites (Farquhar et al., 2001). One of these metabolites is cellulose, which is proportionally the largest contributor of plant biomass and is a semicrystalline organic polymer comprised of repeating units of sugar (β-1,4 linked glucose) (Brown, 1996; Mutwil et al., 2008). When biomass is combusted for fuel, essentially putting the carbon back into the atmosphere, the net atmospheric carbon gain is offset.

The proposed use of starch-based sugar to generate ethanol fuel has generated significant criticism (Searchinger et al., 2008) concerning placing competitive demands on land for food production, which is economically and environmentally not sustainable (Schmer et al., 2008). There also is much debate as to the plausibility of cellulosic biofuel due its recalcitrance to enzymatic hydrolysis, low energy density, additional agronomic infrastructure and possible environmental strains (Searchinger et al., 2008). It is evident that localized cropping solutions will be important due to the lower energy density of plant biomass making it uneconomical to transport at least in its raw form (herbaceous biomass has an energy density of 19.5 MJ kg−1) (Florine et al., 2006). Conversely, crude oil has a heating value on the order of 50 MJ kg−1 (EIA, 2007) and a much higher bulk density compared with biomass. Hence, short distance transportation would be necessary, especially considering that using enzymatic processes would only allow for 40% to 50% of the higher heating value to be converted to alcohol (Florine et al., 2006). These facts suggest that assessment of energy potential from biomass crops that proliferate within a given local biome will be necessary to drive policy level decisions as to the viability of plant-based renewable energy.

Warm-season C4 monocotyledons are excellent candidates as bioenergy crops, because they are some of the most efficient plants on Earth to utilize solar energy for photosynthesis to produce biomass (Heaton et al., 2008). Further, they are adapted over a wide geographic distribution (Cherney & Cherney, 2003) and have favorable yield potential (Redfearn & Nelson, 2003). Efficient carbon utilization is key to maximal feedstock biomass accumulation and C4 vs. C3 photosynthesis refers to an adaptation of the plants photosynthetic system that enhances carbon fixation efficiency (Farquhar et al., 2001). Of all the warm-season perennial grasses, switchgrass (SW) (Panicum virgatum L.) has been the most widely championed biomass feedstock for the central and southeastern United States (Cundiff & Marsh, 1996; Vogel et al., 2002; Teel & Barnhart, 2003; Parrish & Fike, 2005; Comis, 2006). In addition to SW, two of the top native grass candidates in Kentucky are eastern gamagrass (EG) (Trispicum dactyloides L.) and big bluestem (BB) (Andropogon gerardii Vitman) (Olson et al., 2007). The study performed herein occurred in Kentucky where there is a growing interest in the use of perennial native grasses for cellulosic biofuel and combustion-based energy. Although several native perennial C4 grasses can be grown (Olson et al., 2007), the focus has been solely on SW. Currently in USDA-NRCS programs, over 64 000 ha of native grasses have been planted in Kentucky (Brantley A, personal communication). We analyzed the yield potential for three native perennial grasses over a 7-year period. Within each species, we researched multiple cultivars, ecotypes and individuals derived from an experimental germplasm (referred to as entries). The goal of this study was to assess the potential of these entries for bioenergy agriculture in Kentucky.

Methods and materials

Plant material and sampling

Plant samples were from experimental research plots on the University of Kentucky campus and the University of Kentucky Agricultural Experimental Station research farm, Lexington, KY 38.05 °N 84.06 °W (Elev. 981 ft.). A description of the breeding history for grass entries tested in this study is presented in Table 1.

Table 1.   Description of the organisms being used for bioenergy agriculture in this study including origin, ecological and breeding background
  1. Data is provided as a resource for each of the native grass entries used for yield analysis within this study (cultivars and germplasm).

Big Bluestem (BB):
 KYAG9601: ecotype collection from western KY/near land between the lakes area with minimal selection for disease resistance.
 Wapiti: ecotype collection from Central Kentucky (southwestern Hart County) by Randy and John Seymour/Roundstone Native Seeds.
 Pawnee: released cultivar from Nebraska.
 Kaw: released cultivar from Kansas.
 Roundtree: released cultivar from Missouri (original plant collection from Iowa).
Eastern gamagrass (EG):
 Meade County: ecotype collection from Meade County, Kentucky by Henry Burkwat with minimal selection for disease resistance.
 Hart: ecotype collection from Central Kentucky (Hart County) by Randy and John Seymour/Roundstone Native Seeds.
 Highlander: released cultivar from Mississippi.
 Pete: released cultivar (original plant material from Kansas and Oklahoma).
 Iuka: released cultivar
 Jackson: released cultivar
Switchgrass (SW):
 Alamo: released cultivar from Texas (lowland type).
 Cave in Rock: released cultivar from Illinois (upland type).
 KYPV9504: original plants collected from West Virginia to form KY-1625 then selected for uniform leaf color and width (10 plant synthetic) (upland type).
 KYPV9505: original plants collected from West Virginia to form KY-1625 then plants selected for improved agronomic characteristics (upland type).
 KYPV9506: original plants collected from West Virginia to form KY-1625 then plants selected for improved agronomic characteristics (upland type).
 Trailblazer: released cultivar from Nebraska with selection for improved digestibility (25 plant synthetic) (upland type).

Agronomic cultivation and environmental conditions

Field experiments were established in the summer of 2000 into a well-drained Maury silt loam soil (fine, mixed, semiactive, mesic Typic Paleudalfs). Plots were established in June 2000, by transplanting small plants raised in greenhouse float trays from seed or sprigs to ensure uniform stands. These experimental plots were allowed to become established during the remainder of 2000. Transplants were set 1 foot apart using four rows per plot. The research plots were arranged in a complete randomized block design with four replications and 1.5 m × 4.5 m plot dimensions. Total biomass was harvested twice per year at physiological maturity for each species and weighed for each plot during 2002 through 2008. Representative samples were obtained from each entry at each harvest for dry matter yield (kg ha−1) and to conduct subsequent laboratory analyses. Precipitation and temperature measurements were taken daily during 2001–2008 and were summarized as monthly averages (Table S1). Plots were harvested to a 15 cm stubble height using an agricultural forage plot harvester (Hege 212, Wintersteiger AG, Ried, Austria). Plots were fertilized with 67 kg ha−1 nitrogen each spring when plant height reached 10–20 cm. Phosphorus and potassium were applied annually based on University of Kentucky soil test recommendations (P were maintained above 67 kg ha−1 and K were maintained above 336 kg ha−1).

Microscale enzymatic saccharification

Saccharification, as referred to in this paper, is the process of breaking a complex carbohydrate (cellulose) into its monosaccharide components and efficiency is the percentage of cellulosic glucans converted to the monosaccharide form [National Renewable Energy Laboratory Laboratory Analytical Procedure (NREL, LAP-009) 1996]. Each biomass sample was homogenized by grinding to a 1 mm particle diameter. These saccharification efficiency assays for all years were repeated in batches of four for each independent replicate and compared on the basis of the amount of sugar recovered after 72 h enzyme reaction with the cellulase cocktail added in excess [defined as 60 filter paper units (FPU) per gram of cellulose].

Micro-scale saccharification experiments were performed as described in Harris et al. (2009), with the difference that two different sources of enzyme cocktails were used to establish the saccharification efficiency of cellulosic biomass. One enzyme cocktail was obtained by mixing equal volumes of Celluclast 1.5 L (cellulase from Trichoderma reeseii) Sigma-Aldrich (St Louis, MO, USA) and Novozyme 188 (cellobiase from Aspergillus niger) Sigma-Aldrich the two enzymes contained an enzymatic activity of 486-endoglucanase units (EGU) mL−1 for cellulase which equals 45.6 FPU as standardized in NREL, LAP-009 (1996) and 134 cellobiase unit (CBU) mL−1 for cellobiase. Additional experiments were conducted using an enzyme cocktail from a commercial supplier (Alltech Inc., Nicholasville, KY, USA) that contained a mixture of cellulases, xylanases and cellobiases derived from T. reeseii. The commercial enzymes had a cellulase activity of 10 000 CMCU g−1 (carboxymethyl cellulose units per gram measured at a pH of 4.8 and a temperature of 50 °C) and a xylanase activity of 150 000 XU g−1 (xylanase units g−1 measured at a pH of 5.3 and a temperature of 50 °C). The cellulase activity in filter paper units was measured as 184 FPU g−1 (Adney & Baker, 1996). Enzymatic saccharification of lignocellulosic material was done according to the laboratory analytical procedure of the NREL, LAP-009 (1996). Modification for the microscale experiment was made by using 0.04 g dry biomass per sample ground through a 1 mm screen and mixed with 2 mL of buffer with a pH of 4.8. The cellulase loading was 60 FPU g−1 of dry biomass. The samples were incubated in an incubator/shaker (Innova 4300, New Brunswick Scientific Edison, NJ, USA) at 50 °C and shook in a horizontal position at 100 rpm. The progress of the reaction was measured by taking individual aliquots after 24 h and determining the glucose concentration using a glucose analyzer standardized for glucose determination using buffer and membranes purchased from Yellow Springs Instruments (YSI, Yellow Springs, OH, USA). Enzyme blanks and Whatman #1 filter paper (Whatman International Ltd., Maidstone, England) were digested alongside the samples.

The experimental ethanol per hectare unit was calculated by multiplying yield data (kg ha−1) with the cellulose content (% of dry biomass), the conversion efficiency (saccharification) and a factor of 1.11 to account for the weight gain during hydrolysis by the addition of a water molecule. The resulting kilogram glucose per hectare data was multiplied by 0.5114 to account for the weight loss of two carbon dioxide molecules during the glucose to ethanol fermentation and 1.2764 to convert ethanol weight to volume (kilogram to liter). Theoretical conversion efficiency was calculated based on 100% conversion of cellulose to glucose.

Cellulose content measurement

Crude cell walls were prepared as described in Reiter et al. (1993). In brief, samples were homogenized using a grinder (Arthur H Thomas Co Scientific, Philadelphia, PA, USA) equipped with a 1 mm sieve. Twenty-five milligram plant material were incubated in 1 mL 70% ethanol overnight at 65 °C, washed twice with 1 mL 70% ethanol for 1 h and once with 1 mL acetone for 5 min. After removing the wash solutions, the samples were dried under vacuum. Cellulose content was determined by weighing out 5 mg of dry biomass extract in triplicate and boiled in acetic-nitric reagent (acetic acid : nitric acid : water 8 : 1 : 2) for 30 min to remove lignin and hemicellulose (Updegraff, 1969). The samples were allowed to cool to room temperature and the reagents were carefully removed. The plant cell wall material was washed twice with 8 mL MQ-water and 4 mL acetone and dried under vacuum. The cellulose samples were then hydrolyzed in 67% sulfuric acid for 1 h. The glucose content of the samples was determined by the anthrone method (Updegraff, 1969). Twenty-five microliters of the sulfuric acid hydrolyzed samples were mixed with 475 μL water and 1 mL 0.3% anthrone in concentrated sulfuric acid on ice. The samples were boiled for 5 min then placed immediately back on ice. The absorbance of the samples was measured using a Bio-Mate Thermo Scientific spectrophotometer (Thermo Fischer, Waltham, MA, USA) set at OD 620 nm and compared with a standard curve obtained from known (10–50 mm) concentrations of glucose (the standard curve was set each time for each reaction). The cellulose content was calculated by multiplying the measured glucose concentration of each sample by the total volume of the assay and then by the hydration correction factor of 0.9 to correct for the water molecule added during hydrolysis of the cellulose polymer.

Lignin content determination

Acid hydrolysis

Acid-soluble lignin, acid-insoluble lignin and ash were measured using the laboratory analytical protocols NREL, LAP-004 (1996). Three-gram samples (±0.1 g) of biomass were placed into 30 mL test tubes with three replicates followed by an addition of 3 mL of sulfuric acid (72% v/v) to each test tube and stirred thoroughly with a glass rod. Samples were incubated in a water bath at 30 °C for 2 h and stirred every 15 min. After hydrolysis, samples were diluted with 84 mL of deionized water (4% concentration) into a 250 mL Erlenmeyer flask that was autoclaved at 121 °C at 15 psi for 40 min.

Acid soluble lignin measurement

Acid soluble lignin was measured using NREL, LAP-004 (1996). Briefly, the autoclaved solution was allowed to cool to room temperature before withdrawing 1 mL of the hydrolyzate without disturbing any solid particles in the solution. A 4% solution of sulfuric acid was used as a reference blank. To accommodate the absorbance range (0.2–0.7) of the spectrophotometer at 205 nm, both the hydrolyzate and reference 4% sulfuric acid solution were diluted 1 : 30 with deionized water. The absorbance of the hydrolyzate at 205 nm using a 1 cm light path cuvette was used to calculate the acid soluble lignin.

Acid insoluble lignin measurement

Porcelain filter crucibles containing a glass microfiber filter (Whatman 934-AH) were placed in a furnace at 575 °C for 4 h. The crucibles were removed from the furnace and placed into a desiccator to cool for a minimum of 1 h before being weighed. The autoclaved samples were vacuum filtered through the crucibles. The crucibles and their contents were dried in an oven at 105 °C overnight before transferring into a desiccator to cool. The weight of the crucibles were recorded and placed in a furnace at 575 °C for 6 h before reducing the temperature to 105 °C overnight. The crucibles were removed and placed into a desiccator to cool and weighed.

Data analysis

Statistical analyses were performed using sas statistical software (Version 9.1, SAS Institute, Cary, NC, USA). Yield data for the three highest yielding entries for each species were analyzed using a two-way analysis of variance (anova) with year as a blocking factor and entry as a factor nested within species using the general linear model procedure (proc glm) to test for significant differences between species (SW: Alamo, Cave in Rock and KYPV 9504; EG: Meade County, Hart and Highlander; BB: KYAG 9601, Pawnee and Wapiti). In order to analyze across species, we followed the assumption that species were randomized within each replication. In reality, we were not able to randomize species due to harvest timing, space and equipment limitations. Therefore, species' means and standard errors of the entries within species were calculated to illustrate annual yield trends. Yield data for each species were subjected to a one-way anova with year as a blocking factor to determine entry effect. Mean separation was conducted by pairwise difference comparison using the mixed procedure (proc mixed). Data were considered significant at P≤0.05.

Laboratory analysis of cellulose, saccharification efficiency and glucose were conducted on single composite samples for each year from 2006 to 2008. As such, cellulose, saccharification efficiency and glucose data were analyzed using an incomplete block design with year as the blocking factor and entry as the treatment factor using the proc glm of sas. Data were considered significant at P≤0.05.


Yield comparison between SW, EG and BB

To determine if species yields were significantly different, the three highest yielding varieties over the 7-year trial period were compared (SW: Alamo, Cave in Rock and KYPV 9504; EG: Meade County, Hart and Highlander; BB: KYAG 9601, Pawnee and Wapiti)(Table 2). Species was significantly different (F=110.12, df=2, 42, P<0.0001) and the species and year interaction was also significantly different (F=40.74, df=12, 42, P<0.0001). Therefore, the highest yielding species varied depending on year. Entry and entry-year interactions were both significantly different (F=7.00, df=6, P<0.0001; F=6.54, df=36, P<0.0001). Due to the assumptions of this analysis and the significant factor interactions, comparison between species cannot be rigorously tested and entry data have to be analyzed within species and year. The average annual yield and standard errors among entries within species were provided to demonstrate annual yield trends (Table 2). EG had the highest overall average total yield at 78 888 kg ha−1 compared with SW with 73 784 kg ha−1 and BB with 59 531 kg ha−1. However, EG entries were not consistently greater than SW or BB in all years (Table 2). The highest yield potential for an EG entry was Meade County, which averaged 14 455 kg ha−1, compared with the highest yielding SW entry Cave in Rock 12 356 kg ha−1 and 10 192 kg ha−1 for BB- KYAG9601 (Table 2). EG Meade County also displayed the greatest individual yield potential of 27 573 kg ha−1 observed in the 2003 growing season.

Table 2.   Average total dry matter yield (kg ha−1) over a 7-year trial comparing big bluestem (BB), eastern gamagrass (EG) and switchgrass (SW) entries
BB Entry20022003200420052006200720087-year total
  • *

    Average total yields within year followed by the same letter are not significantly different (P≤0.05). Pairwise difference comparison was conducted using proc mixed.

  • †Standard Error was calculated to illustrate the data spread among the varieties within species.

KYAG 960110 200 a*7748 b16 171 a7062 a9203 a10 243 a10 715 a71 343
Wapiti8473 b10 116 a14 910 ab5905 a7452 b8879 b9614 ab65 349
Pawnee7562 b8570 ab14 234 b5883 ab6896 b7906 bc8924 bc59 975
Kaw7605 b8940 ab10 801 c5816 ab5350 c7805 bc7699 cd54 017
Roundtree6208 c4013 c11 630 c4537 b6193 bc7245 c7143 d46 969
Mean8010787713 549584070198416881959 531
Standard Errorb657103810114006505276444253
EG Entry
Meade County17 944 a27 573 a18 796 a10 404 a9428 a7979 ab9064 a101 188
Hart16 303 a24 864 b17 160 a9249 ab9455 a8466 a8351 ab93 848
Highlander16 060 a24 079 b16 868 a7571 bc8776 ab8448 a8578 ab90 379
PMK 2413 000 b19 241 c11 992 b6871 bc7777 ab7294 abc7854 ab74 030
Iuka7267 c15 915 d10 429 b5234 cd6471 bc5537 c6705 ab57 557
Jackson8484 c17 222 cd10 831 b2938 d4682 c5689 bc6480 b56 325
Mean13 17621 48214 346704577657235783878 888
Standard Error18061910149811047715424267833
SW Entry
Cave in Rock12 654 b16 674 b13 995 a9335 a10 427 a14 625 a8846 a86 556
Alamo17 927 a25 971a10 874 b4555 c6811 bc9299 c5833 bc81 271
KYPV 950410 365 bc16 423 b14 120 a7161 ab8895 ab11 645 b7968 ab76 578
KYPV 950510 781 bc16 618 b14 204 a6935 b8625 ab9491 bc6579 abc73 234
KYPV 950611 372 bc15 787 b13 248 a7361 ab8090 b9062 c6798 abc71 719
Trailblazer9602 c12 889 c10 693 b5329 bc5075 c5041 d4715 c53 344
Mean12 11717 39412 856677979879861679073 784
Standard Error1235181267068575412946024652

Yield comparison between SW, EG and BB entries

Native C4 grass varieties were selected from three different genera for yield determination in this study (Table 1). To determine if entry within species is a significant factor influencing the yield potential of bioenergy feedstocks grown in Kentucky, we measured and compared total dry matter yield (kg ha−1) over a 7-year period for each species and entry for years 2002–2008. Entry was shown to be a significant factor for all three species (SW: F=26.88, df=5, 105, P<0.0001; EG: F=71.27, df=5, 105, P<0.0001; BB: F=55.81, df=4, 84, P<0.0001)(Table 2). Entry and year interaction was also significant for all three species (SW: F=7.59, df=30, 105, P<0.0001; EG: F=3.63, df=30, 105, P<0.0001; BB: F=4.40, df=24, 84, P<0.0001). Therefore, the entries within species' yield ranked differently depending on year. As a result, the entries were analyzed within year (Table 2).

Enzymatic saccharification studies in native grass crops

To further determine the suitability of each species and entry for biofuel production, we measured the efficiency of enzymatic conversion to fermentable sugar (defined as saccharification). Plant material from selected harvests during 2006, 2007 and 2008 were used for the laboratory analyses reported in this paper (Table S2). One-way anova was used to compare individual entries with year as a blocking factor. These analyses were performed on representative batch samples for each year and for each species and entry. These data revealed that average saccharification efficiency for EG as a species was 12.93% of total cellulose converted to glucose (fermentable sugar) by a cellulase cocktail, whereas average saccharification efficiency for SW was significantly lower 8.42% (F=10.12, df=2, P<0.0003) (One-way anova and also confirmed by bimodal histogram plot via the Wilcoxon Signed-Rank Test P<0.001, data not shown) (Table 3). EG Highlander and Meade County had the highest saccharification efficiency (18.12% and 16.76% respectively), whereas SW KYPV9404 and KYPV9405 were the least favorable entries (5.3% and 5.93% respectively) based on saccharification efficiency of all entries analyzed in this study (Table 4, Table S2) This illustrates up to a threefold variation in raw biomass digestibility. We have presented all saccharification data as a reference resource for future studies using these entries (Table S2).

Table 3.   Statistical comparison of conversion efficiency between switchgrass, eastern gamagrass and big bluestem over a 3-year period (six varieties were analyzed from each species)
SpeciesAverage saccharification
Efficiency (%)Standard error
  1. Average saccharification efficiency refers to the amount of cellulose converted to fermentable sugar as a percentage to total cellulose:

  2. *Average saccharification efficiency followed by the same letter are not significantly different (P≤0.05) comparison was conducted using proc glm.

Table 4.   Saccharification efficiency (percentage of total cellulose converted to fermentable sugar), cellulose content, apparent and potential ethanol yield (liters per hectare) data are calculated for BB, EG and SW varieties examined in this study
VarietyAverage conversionStandard
Average celluloseStandard
Apparent ethanolPotential ethanol
Efficiency (%)Content (%)Yield l ha−1Yield l ha−1
KYAG 960113.092.7439.709.953812914
Meade County15.741.5232.282.425293361
PMK 2412.031.9239.375.873612998
Cave in Rock10.841.8541.808.614033723
KYPV 95047.671.7343.105.452603396

The theoretical ethanol yield per hectare (L ha−1) for each entry was determined (Table 4). This relative measure of cellulosic biofuel potential showed that EG Meade County (529 L ha−1) and Highlander (502 L ha−1) had the highest yield and saccharification efficiency (Tables 2 and 4). However, if measuring theoretical ethanol yield based on 100% cellulose conversion (which was not achieved), it was found that SW entries Cave in Rock and Alamo had the highest biorefinery yield potential up to 3852 L ha−1 compared with EG Meade County, at 3361 L ha−1.

Correlation between cellulose, lignin and saccharification efficiency in native perennial grasses

Cellulose content was necessary to determine the saccharification efficiency. Hence, these two measurements are interlinked. Quantification of lignin levels, as acid soluble, acid insoluble and total lignin, were performed to establish the correlation between these three central factors contributing to overall recalcitrance of cellulosic biomass to enzymatic hydrolysis. Cellulose contents were not significantly different between EG, BB and SW (df=2, F=0.18, P>0.05) (Table 4). Sample cellulose content was found to range between 27% and 59% of dry biomass (Table S2). Total lignin content was analyzed for a subset of samples that corresponded to the greatest variation in saccharification values. Entries analyzed were correlated with cellulose and saccharification efficiency by pairwise analysis using proc glm. Lignin content for the samples analyzed ranged from 15% to 29% of the dry biomass (Table S2). When lignin was plotted against the saccharification efficiencies in a pairwise plot and a linear regression analysis performed a clear trend was observed (Fig. 1). As lignin increases, so did the recalcitrance of cellulose. We previously attempted to determine if any structural change in cellulose occurred within a large range of grasses using X-ray scattering (Harris & DeBolt, 2008) and concluded that no detectable change in the relative crystallinity occurred among any of the species analyzed. A negative correlation between lignification and digestibility was determined (y=−0.4587x+22.769, R2=0.2583, Fig. 1a). The SW entry Trailblazer had one of the highest digestibility amongst the SW entries (6/19/2007 harvest 14.27% efficiency) and also displayed the lowest lignin content (19%). On the contrary, in the comparison between SW, BB and EG it appeared that as lignin increased, the total amount of cellulose also increased (y=1.3356x+10.001, R2=0.4082) (Fig. 1b). This was due to a trend of greater cellulose content in SW entries when compared with EG (Table 4, Table S2).

Figure 1.

 Pairwise linear regression analysis of cellulose/lignin and lignin/saccharification. (a) Pairwise comparison between lignin content and saccharification, (b) Pairwise comparison between cellulose and lignin levels.


In Kentucky, along with many states in the region, (TN, IN, OH, WV and IL) a substantial amount of energy is currently derived from fossil sources of coal. Moreover, there are limited options for renewable power production from low carbon sources such as solar, wind and hydro. Therefore, regions like Kentucky that have a relatively long growing season, significant rainfall and abundant marginal land could produce sizeable quantities of perennial herbaceous and woody biomass, which are well suited to bioenergy agriculture. Several recent studies and reviews have researched the global challenge of using plants as a source of carbon neutral biofuel for alcohol based transportation fuels (Campbell et al., 2008; Schmer et al., 2008; Piñeiro et al., 2009). Others have commented on the problems associated with using starch-based plant components as biofuels, since they compete with food demand and are largely inefficient (Searchinger et al., 2008; Piñeiro et al., 2009). In general, these reports have looked favorably on the use of cellulose as the next generation of biofuel (Tilman & Downing, 1994; Hill et al., 2009). Herein, we study the feedstock characteristics of several native grasses to address the current lack of available data regarding regionally suitable genotypes that are suitable for cellulosic bioenergy agriculture. EG, a less researched C4 grass, displayed significantly greater yield and digestibility traits than SW or BB.

Providing renewable energy alternatives at the global level is one of the central challenges facing the current generation of scientists. A common phrase, ‘think globally, act locally’ is particularly true for renewable energy given the distinct geographical locations and climatic criteria that dictate where specific energy plants will grow. We examined several energy grass candidates grown in the experimental station in Lexington, KY for yield and glucose conversion efficiency to identify the entries best suited for local energy plantation. Specifically, baseline values for biofuel production for locally endemic C4 perennial grasses EG, SW and BB were established. Both the environmental and meteorological conditions of the growing year and the different energy grass entries planted had a significant impact on the biomass yield in Kentucky. Over a 7-year trial period, we determined EG to be the highest yielding species relative to SW and BB. It is worth noting that all three entries were high yielding sources of biomass for cellulosic biofuel, consistent with C4 grass capacity to efficiently fix atmospheric carbon. Finding that growing year had a significant effect on crop yields was consistent with Kentucky's proximity to the US continental jet stream (Casler et al., 2007) and experiences substantial year-to-year variability in weather patterns (Table 3). A plausible explanation for environmental effect (year variation) is that meteorological circumstances in certain years favor one species over another. High yielding entries that are incapable of tolerating stress may result in less overall yield than lower yielding entries that have higher tolerance level. For example, Alamo is a lowland cultivar from Texas that can suffer from winter damage in Kentucky (Alderson & Sharp, 1994). It may not have persisted as well as other SW entries due to the combination of cold winters, summer temperatures and soil conditions during the study. Others have found latitude of origin to have a significant effect on performance of SW (Hultquist et al., 1997) and recommend populations not be moved more than one USDA plant hardiness zone from their area of origin. Most lowland SW cultivars are tetraploid with 2n=36, while upland cultivars are usually hexaploid, with 2n=54 (Hultquist et al., 1997). Lowland cultivars tend to have larger stem diameters and plant heights, so yields can be higher than upland cultivars especially using one cut per year harvest systems (Lemus et al., 2002). Both EG and BB also have polyploidy series in native germplasm. Many of the entries in this study are ecotype cultivars, and as such, may not have been subjected to intensive selection pressure for higher biomass production. Adaptation to local conditions includes pest resistance and stress tolerance, so yield measurements alone may not be the most relevant character to measure. Persistence and stand density may be as important as yield, particularly if crops are to be grown on marginal lands (Tilman & Downing, 1994; Campbell et al., 2008). The main findings of the study were that EG entries Meade County, Hart and Highlander yielded greater amounts of biomass over a 7-year period than other EG, SW or BB entries. Hence, the data lead us to postulate that regional assessment of energy crops may lead to significant improvements in yield and entry performance was in part dependent on environment.

Biomass yield is an important measurement of an energy plantation, however a second factor is the type of energy required. Biomass can be burned to generate electricity or fermented to generate liquid transportation fuel for internal combustion engines. The major hurdle to overcome in fermenting cellulosic material is the recalcitrance of cellulose to enzymatic hydrolysis. Several factors alter digestibility of cellulose such as lignification (Chapple et al., 2007) and crystallinity (Harris & DeBolt, 2008; Harris et al., 2009). Analysis of saccharification efficiency showed that EG entries Meade County, Hart and Highlander also had significantly higher values when compared with other EG entries, SW and BB. Multiplying yield by saccharification efficiency and cellulose content showed that EG, relative to SW and BB, had significantly greater potential to serve as an efficient cellulosic ethanol biofuel feedstock in Kentucky. Although SW produced the highest cellulose yield per hectare the lower digestibility of its cellulose makes it difficult and\or more expensive to produce bioethanol from it.

If our focus was placed on energy density per se, then SW would be a superior feedstock to EG despite having saccharification efficiency values that were on average only half that of EG. Underlying the higher energy density of SW was a trend towards higher cellulose as well as lignin in SW entries and lower lignin in EG entries was evident. However, energy density would be more sought after for thermochemical energy production (such as supplementing electricity generation by burning coal plus biomass) and not necessarily for the production of alcohol by enzymatic hydrolysis. Based on the calculation that cellulose yields approximately 17 500 kJ kg−1 and lignin yields 25 600 kJ kg−1 (NREL), we conclude that growing SW for thermochemical processes would have approximately 10% greater apparent energy dense than EG (based on the average of values for cellulose and lignin for seven cultivars for both SW and EG multiplied by the relative energy density of the products).

Pretreating lignocellulosic biomass is commonly used to loosen the interaction of lignin with cellulose (Schmer et al., 2008). Pretreatment of the biomass with weak acid or basic solutions were avoided in this study since we aimed at assessing the native biomass saccharification efficiency in experimental micro scale assays. To further explore the underlying cause for the range of saccharification efficiencies among entries we analyzed lignin composition since other studies suggest it has an influence on digestibility rates (Chen & Dixon, 2007). Linear regression analysis showed that low lignin levels were correlated with high saccharification values, particularly in the EG entries. The saccharification efficiency of all the EG and BB entries were significantly higher than the SW entries showing that the more digestible cellulose is not just the result of breeding of the entries, but plausibly the result of the different lignification level of the species. Further supporting this hypothesis, analysis of SW Trailblazer showed that it had greater saccharification efficiency than other entries among SW, but also had the lowest lignin level among the SW entries. These data are consistent with the breeding of Trailblazer for digestibility as a forage crop (Alderson & Sharp, 1994). In agreement with our data for EG, Ritchie et al. (2006) also found that EG entries generally had low lignin/fiber content and high digestibility. Hence, we propose that EG is a potentially useful alternative native feedstock for alcohol based fuel production in central Kentucky.

A recent study of EG grown on depleted soil and under environmental stress has demonstrated that it has comparable composition and digestibility traits to many forages currently used in agroecosystems (Ritchie et al., 2006). Models of global bioenergy cropping systems suggest that in order to avoid competing with food crops or cutting of pristine forested lands, energy crops will be grown on reclaimed or depleted agricultural and mining lands (Tilman & Downing, 1994; Campbell et al., 2008). This is particularly true in Kentucky, where tens of thousands of hectares of abandoned agricultural and mine lands need to be evaluated for energy crop potential. Findings of this study suggest that EG is a native, well-adapted feedstock to meet yield, digestibility and sustainability requirements. Moreover, given the natural diversity of native grasses it will be important to screen endemic populations for hardiness traits as well as identify high cellulose, low lignin variants among natural populations for efficient cellulosic alcohol production and lignin rich variants for energy density. A recent study has shown that some of these traditionally grown forage crops have high yield and fermentation characteristics consistent with cellulosic biofuel needs (Weimer & Springer, 2007). These authors found that BB was overall more efficiently converted to fermentable sugars than EG, but they noted that higher yields and a modest trend toward increased fermentability was seen with EG samples grown in the more eastern locations. Examination of native biofuel plants for yield and digestibility is an important step in optimizing bioenergy production systems, but further work is needed to establish pilot scale trials for energy crops on abandoned and depleted lands and use of life cycle assessment to determine their value for communities and energy synthesis.


The authors want to thank Gene Olson for managing the experimental plots, Angela Schoergendorfer for her valuable help with the statistical analyses and three anonymous reviewers for helpful comments. This article is published with approval of the Director of the Kentucky Agricultural Experiment Station as article number 08-11-130 Funding was provided as start up funds to SD.