(583d) Feedstock Mixture Effects On Sugar Monomer Recovery Following Dilute Acid Pretreatment and Enzymatic Hydrolysis

Feedstock Mixture
Effects on Sugar Monomer Recovery Following Dilute Acid Pretreatment and
Enzymatic Hydrolysis

Michael
Brodeur-Campbell¹, Jordan Klinger¹, and David
Shonnard^1,2

[1] Department of Chemical Engineering,
Michigan Technological University

[2] Director Sustainable Futures Institute,
Michigan Technological University

Increased
use of lignocellulosic biomass for bioenergy
production holds promise for improved domestic energy security, reduced
greenhouse gas (GHG) emissions, rural economic development, and developing a
more sustainable energy infrastructure [Chum and Overend
2001; Perlack, Wright, et al.  2005; U.S. DOE 2006].  Transportation fuels are especially difficult
to replace as their replacement has to have a high energy density, be compatible
with existing fuel distribution infrastructure, and be compatible with existing
automobile technology.  In 2009, 97% of
transportation fuels were derived from fossil sources [EIA 2008], and
transportation accounted for one-third of carbon dioxide emissions in 2008 [EIA
2009].  Liquid fuels derived from biomass
are particularly attractive as a replacement for fossil-derived transportation
fuels.  Cellulose is the most abundant
biopolymer on Earth, and lignocellulosic feedstocks represent one of the few
natural resources available in sufficient quantity to replace a significant
fraction of liquid transportation fuel.

In a
commercial biorefinery feedstocks are likely to come from a variety of sources
and include many different materials, such as softwoods, hardwoods, agricultural
residues, and herbaceous or woody energy crops. 
Much work has been done to understand and optimize the pretreatment and
enzymatic hydrolysis conditions for pure feedstocks, but little is known
about mixture effects on enzymatic hydrolysis. 
A primary concern in processing woody biomass to produce a range of
biofuels is to maximize the production of desired intermediates, in this case,
determining the quantity of monomer sugars that are recoverable.  The biochemical platform as described by the
National Renewable Energy Laboratory (NREL) was the process studied.  Briefly, lignocellulosic biomass ? which is
composed primarily of hemicellulose, cellulose, and lignin ? is treated with
dilute sulfuric acid to hydrolyze the hemicellulose into soluble sugars.  The solid residue remaining after dilute acid
pretreatment is then treated with cellulose-degrading enzymes to hydrolyze the
cellulose into soluble sugars.  The
remaining lignin may be utilized as a high energy co-product stream to provide
heat and power for the cellulosic ethanol conversion processes.

This study
seeks to clarify the effects of processing mixtures of feedstocks on overall sugar
recovery.  We wish to test the hypothesis
that sugar yields from mixed feedstocks can be predicted by a simple linear
interpolation model based on the behavior of the pure species.  Experiments were performed on pure species
(aspen, balsam, and switchgrass), and on 50/50 blends of aspen/balsam and
aspen/switchgrass.  Aspen was chosen as a
model species due to its high suitability for processing in a commercial
biorefinery.  Balsam was chosen to mix
with aspen to study the effects of high lignin content on mixed feedstock
processing, and switchgrass was chosen to mix with aspen to investigate the
effects of a high ash content energy crop.

One-half
gram (0.5 g) of dry sample was weighed and placed into one of a set of 9
stainless steel Swagelok cylindrical reactor tubes, and 4.5 ml of dilute
sulfuric acid (0.5% w/v) was added to achieve 10% (wt) solids loading in the
reactor.  Reactors were rapidly heated to
160°C in a silicon oil bath (Dow Corning 550 fluid), and three tubes were
removed from the oil bath at previously determined time intervals.  From previous research involving pure samples
of aspen, balsam, and switchgrass (Jensen et al. 2010), the time to achieve
maximum yields of monomer and oligomer sugars from combined dilute acid
hydrolysis and enzymatic hydrolysis was determined; the ?optimum? time. Three
reactors were removed at one-half the optimal time, three reactors were removed
at the optimal time point, and the last three were removed at two times the
optimal time.  Reactor tubes removed from
the oil bath were immediately submerged in an ice-water bath to quench the
reaction.  One milliliter of liquid from
each reactor was removed, filtered, neutralized, and analyzed on an Agilent
1200 series HPLC for glucose and xylose concentration.  One milliliter of liquid from each reactor
was filtered and then further acidified to 4% (w/v) sulfuric acid, and treated
at 121°C for one hour to
hydrolyze any soluble oligomers into monomer form.  After neutralization these samples were analyzed
by HPLC as above.  Results from these two
liquid samples were combined into total soluble (monomer plus oligomer) sugar
recovery from dilute acid pretreatment. 
Solids were washed with excess distilled water on a glass-fiber filter
(VWR) and carried on to enzymatic hydrolysis.

Solid
residues containing cellulose and lignin were subjected to enzymatic hydrolysis
using Accelerase 1500 (Genencore).  Each of the three reactors from the same
pretreatment time point were loaded with a different enzyme concentration ?
0.1, 0.25, and 0.5 ml Accelerase 1500 per gram of
carbohydrate in the feedstock ? in order to study the interaction between
pretreatment severity and enzyme loading on sugar yields.  Reaction conditions were 50mM citrate buffer,
pH 4.8, 50°C,
250 RPM orbital shaking.  Liquid samples
were removed at 0, 1, 2, and 3 days, filtered, and analyzed for sugar concentrations
by HPLC as above.

Maximum
yields of xylose from pretreatment were seen to vary with both time and
feedstock composition in the expected manner.  Maximum glucose yields following enzymatic
hydrolysis were seen to be strongly affected by both pretreatment severity and
enzyme loading.  Results are summarized
in Table 1 below.

Table 1:  Experimental sugar recoveries for Dilute Acid
Hydrolysis (DAH), Enzymatic Hydrolysis (EH), and experimental and predicted
recoveries for Total Sugar (TS) yield as percent of theoretical.

[1]
Predicted value based on Aspen (2) pretreatment results.

Xylose
yields from pretreatment varied from 69-93% of the theoretical maximum.  Pure switchgrass gave the lowest yield and
pure aspen the highest.  Each pure
species and mixture showed a peak hydrolysis yield near the expected optimal
time.  Earlier time points showed
incomplete hydrolysis with greater quantities of soluble oligomeric
sugars in the hydrolyzate.  Later time
points yielded sugar decomposition products such as furfural and hydroxymethylfurfural (HMF) in the hydrolyzate.  Longer pretreatment reaction times always
resulted in higher glucose yields, but due to the degradation of xylose at
later time points total sugar yields were generally lower for extended
pretreatment times. 

Glucose
yields following enzymatic hydrolysis varied from 2-61% with pure balsam giving
the lowest yield and pure aspen with supplemental β-glucosidase giving the
highest.  The first aspen experiment was
repeated due to incomplete pretreatment (indicated by lower than expected sugar
recoveries), and subsequently treated with the highest Accelerase
concentration plus supplemental β-glucosidase to study the effect of
additional enzyme loading on yield.

Using a
simple linear interpolation model to predict sugar recoveries for mixed
feedstock streams based on the known yields obtainable for pure species gives a
result that is accurate to within about 2% of the experimental data for both
glucose and xylose recovery following dilute acid pretreatment, and to within
3% for glucose recovery following enzymatic hydrolysis.  Xylose recovery following enzymatic
hydrolysis for blends containing balsam could not be predicted accurately from
these data, but since that is only a very small amount of the total sugar yield
the discrepancy did not affect the model results for total sugar yield.  Total sugar yield was predicted to within 5%
of experimental results.

References:

Chum, Helena L. and Ralph P Overend (2001).  Biomass and renewable fuels, Fuel
Processing Technology 71, 187-195.

Perlack,
R. D., L. L. Wright, et al. (2005). Biomass as Feedstock for a Bioenergy and Bioproducts Industry: The Technical Feasibility of a
Billion-Ton Annual Supply. U. S. D. o. A. (USDA) and U. S. D. o. E. (DOE).

U.S. DOE (2006).  Breaking
the Biological Barriers to Cellulosic Ethanol: A Joint Research Agenda,
DOE/SC-0095, U.S. Department of Energy Office of Science and Office of Energy
Efficiency and Renewable Energy (www.doegenomestolife.org/biofuels/).

EIA (2008).  Emissions
of Greenhouse Gases in the United States.  U.S. Energy Information Administration Office of Integrated Analysis and
Forecasting (http://www.eia.doe.gov/oiaf/1605/ggrpt/carbon.html).

EIA (2009).  Annual Energy Review 2009.  U.S.
Energy Information Administration, (http://www.eia.doe.gov/aer/pecss_diagram.html).

Jensen, Jill R., Juan E. Morinelly, Kelsey R. Gossen,
Michael J. Brodeur-Campbell, and David R. Shonnard (2010).  Effects
of dilute acid pretreatment conditions on enzymatic hydrolysis monomer and
oligomer sugar yields for aspen, balsam, and switchgrass.  Bioresource
Technology 101, 2317-2325.