Inuence of Substrate Particle Size and Wet Oxidation on Physical Surface

Structures and Enzymatic Hydrolysis of Wheat Straw

Mads Pedersen and Anne S. Meyer
Center for Bioprocess Engineering, Dept. of Chemical and Biochemical Engineering, Technical University of Denmark,
DK-2800 Kgs. Lyngby, Denmark

DOI 10.1021/bp.141
Published online February 26, 2009 in Wiley InterScience (www.interscience.wiley.com).

In the worldwide quest for producing biofuels from lignocellulosic biomass, the impor-
tance of the substrate pretreatment is becoming increasingly apparent. This work examined
the effects of reducing the substrate particle sizes of wheat straw by grinding prior to wet
oxidation and enzymatic hydrolysis. The yields of glucose and xylose were assessed after
treatments with a benchmark cellulase system consisting of Celluclast 1.5 L (Trichoderma
reesei) and Novozym 188 b-glucosidase (Aspergillus niger). Both wet oxidized and not wet
oxidized wheat straw particles gave increased glucose release with reduced particle size. Af-
ter wet oxidation, the glucose release from the smallest particles (53149 lm) reached 90%
of the theoretical maximum after 24 h of enzyme treatment. The corresponding glucose
release from the wet oxidized reference samples (24 cm) was
65% of the theoretical max-
imum. The xylose release only increased (by up to 39%) with particle size decrease for the
straw particles that had not been wet oxidized. Wet oxidation pretreatment increased the en-
zymatic xylose release by 5.4 times and the glucose release by 1.8 times across all particle
sizes. Comparison of scanning electron microscopy images of the straw particles revealed
edged, nonspherical, porous particles with variable surface structures as a result of the
grinding. Wet oxidation pretreatment tore up the surface structures of the particles to retain
vascular bundles of xylem and phloem. The enzymatic hydrolysis left behind a signicant
amount of solid, apparently porous structures within all particles size groups of both the not
wet oxidized and wet oxidized particles. VC 2009 American Institute of Chemical Engineers

Biotechnol. Prog., 25: 399408, 2009

Keywords: bioethanol, enzymatic hydrolysis, particle size, particle surface, pretreatment,
SEM, wet oxidation, wheat straw

Introduction cellulolytic enzymes and minimizing the enzyme addition

The conversion of cellulosic biomass into bioethanol is
currently studied worldwide to develop CO2-neutral biofuel
alternatives to fossil fuels. One of the rst prerequisites in
such ethanol production is the efcient generation of a fer-
mentable hydrolysate, rich in glucose, from the cellulose in
the biomass feedstock. Application of enzymes to accom-
plish the degradation of cellulose to glucose is now consid-
ered the most viable strategy to provide cost-efcient,
environmentally friendly processes that avoids generation of
byproducts that may inhibit the subsequent fermentation.1
However, exactly the efcient enzymatic conversion of the
(ligno)cellulose is currently a major bottleneck in lignocellu-
lose-to-ethanol processes, notably because of the signicant
robustness of lignocellulosic biomass to microbial and enzy-
matic deconstruction.2,3 To increase the liability of the cellu-
lose to enzymatic attack, lignocellulosic substrates are
subjected to a physicochemical/thermochemical pretreatment
prior to the enzymatic hydrolysis step.25 Signicant progress
has recently been made with respect to improving the
Correspondence concerning this article should be addressed to A. S.
Meyer at am@kt.dtu.dk.
levels for hydrolytic conversion of cellulose to glucose.69
This is why the pretreatment step is now considered as one
of the most expensive processing steps in cellulosic bio-
mass conversion to ethanol, thus requiring signicant
improvements.3,5
Wheat straw is in focus as a cellulosic feedstock for bioe-
thanol processes, notably in Europe.10 Wheat straw consists
of
3443% cellulose, 2635% hemicelluloses, and 1421%
lignin9,10the slight compositional variability seen in differ-
ent studies is due to varietal, geographical, and climatic
inuences on the makeup of the straw. A number of recent
reports have highlighted the inuence of different types of
pretreatments, the efcacy of alternative substrate loading
modes, and the molecular events occurring during enzymatic
hydrolysis of pretreated wheat straw.10,11 Thus, it is now
clear that signicant physical, notably viscosity lowering, as
well as chemical changes of the straw take place during pre-
treatment and enzyme catalysis.10,11 In spite of this, only
little is known about the changes of the substrate surface and
the supramolecular substrate structures taking place as a
result of pretreatment and enzymatic hydrolysis of wheat
straw. More detailed knowledge on the structural changes is
a prerequisite for obtaining an understanding of the impact

C 2009 American Institute of Chemical Engineers

V 399
400
of the physicochemical substrate properties on the enzymatic
conversion to develop improved processing treatments.
Diminution of substrate particles has previously been
shown to be advantageous in enzymatic hydrolysis of differ-
ent lignocellulosic plant residues.1214 Presumably, when the
size of the substrate particles is reduced, the accessible sur-
face for enzymatic attack is increased, and the lengths of the
entry and exit paths for the enzymes and the hydrolysis
products, respectively, are reduced.15,16 These events pre-
sumably increase the enzymatic hydrolysis rate. However,
the available literature suggests that the relationship between
the biomass structure and enzymatic degradation is complex
and goes beyond the concept of accessible substrate surface
area and involves other features such as e.g. pore structures
and local concentration gradients.16,17
This work was undertaken to assess the effects of sub-
strate particle size diminution on (i) the structural appearance
as evaluated by scanning electron microscopy (SEM) images,
(ii) the quantitative geometrical properties, and (iii) the enzy-
matic hydrolysis of wheat straw before and after wet oxida-
tion pretreatment. The study was made on wheat straw
particles of three different particle sizes (53149 lm, 250
500 lm, and 7071,000 lm) and a reference particle size
(24 cm). Glucose and xylose releases were compared after
enzymatic hydrolysis treatments with a set blend (3:1, v/v)
of the commercial enzymes Celluclast 1.5L, a multicompo-
nent cellulolytic and hemicellulolytic enzyme preparation
from Trichoderma reesei, and Novozym 188, a b-glucosidase
derived from Aspergillus niger.
Materials and Methods
Substrate preparation
Wheat straw grown in Grumlse (southern Zealand, Den-
mark) in 2005 was obtained from The Danish Cooperative
Farm Supply, Barse, Denmark. The substrate was sorted
manually to contain only stems and ground repeatingly by
use of a cutting mill (Retsch SM 2000, Haan, Germany) to
reach particle sizes reaching from 4 cm to 53 lm. To sepa-
rate the different particle sizes, the ground straw samples
were passed through a series of stainless steel sieves (Ende-
cotts, London, UK). The sieving tower was shaken mechani-
cally for 510 min to obtain the samples of the different
particle sizes. Particles of three size ranges (sieve aperture
sizes) were selected for the study; 7071,000 lm, 250500
lm, and 53149 lm. The straw samples used for compari-
son, designated as reference substrate samples, were 24 cm
in length. To determine the substrate composition the stand-
ard procedure, including both the practical procedure and the
equations for calculating the composition, the standard pro-
cedure from the U.S. National Renewable Energy Laboratory
was used.18
Wet oxidation
The pretreatment of the wheat straw was carried out
according to the procedure described by Bjerre et al.19 in a
loop autoclave which was loaded with 60 g dry matter wheat
straw, 6.5 g Na2CO3, and 1 L water per batch (equivalent to
0.108 g Na2CO3 per kg substrate dry matter). Oxygen pres-
sure was 12 bar and the biomass was treated at 195 C for
10 min.19 After this wet oxidation treatment, the solid and
the liquid phase were kept together as a slurry and stored in
aliquots at 18 C until use.
Biotechnol. Prog., 2009, Vol. 25, No. 2
Enzymatic hydrolysis
Enzymatic hydrolysis reactions were in all cases carried
out by treatment with a set 3:1 (volume/volume) blend of
Celluclast 1.5L derived from Trichoderma reesei (Novo-
zymes A/S, Bagsvrd, Denmark) and Novozym 188 derived
from Aspergillus niger (Novozymes A/S, Bagsvrd, Den-
mark) in a sodium acetate buffer (0.1 M) with 0.02% sodium
azide at pH 5. Celluclast 1.5L had a lter paper unit (FPU)
activity of 67 FPU g1 using the NREL standardized lter
paper assay20 and a declared cellulase activity 700 EGU g1
(endoglucanase units). Novozym 188 had an activity of 246
CBU g1 (CBU, Cellobiase units). The CBU activity was
determined by measuring glucose production on cellobiose at
40 C, pH 5 (provided by Novozymes A/S). The hydrolysis
reactions took place in Eppendorf tubes with 2% dry matter
substrate (weight/weight), the straw in the reference substrate
samples were bent to be wetted during the reactions. The
dosage volume of the enzyme additions was adjusted to give
a total reaction volume of 1 mL and an enzyme activity of
0.6 FPU 0.7 CBU per reaction (equivalent to addition of
30 FPU 37 CBU per g substrate dry matter or addition of
600 g enzymes (Celluclast Novozym 188) per kg substrate
dry matter). Each enzyme hydrolysis treatment was accom-
plished during shaking at 750 rpm at 50 C using a Thermo-
mixer (Eppendorf, Hamburg, Germany).
The enzymatic hydrolyses were stopped by heating of the
samples at 100 C for 10 min. The samples were then cooled
to room temperature, centrifuged at 10,000 rpm for 10 min
and the levels of glucose and xylose liberated were deter-
mined by high-performance anion exchange chromatography
as described previously.21
Yields of enzymatic hydrolysis
The efciency of the enzymatic hydrolysis was evaluated
by calculating the release of glucose and xylose in the enzy-
matic hydrolysis as the percentage of the maximal theoretical
release (Eq. 1):
Cx;hydrolyzed;pretreated;grinded sample g=g
Yield x  100%
Cx;grinded sample g=g
(1)
where x denotes the type of monosaccharide (glucose or
xylose), the Cx in the numerator is the concentration of x
measured afterwards enzymatic hydrolysis and the Cx in the
denominator denotes the concentration of x found by compo-
sitional analysis. The yield was calculated for each particle
size after 0, 3, 6, 12, and 24 h.
SEM and energy dispersive spectroscopy
The inuence of the grinding, wet oxidation pretreatment,
and enzymatic hydrolysis on the structure of wheat straw
particles was evaluated by SEM. SEM was carried out in a
JEOL microscope model JSM-5900. Prior to analysis the
samples were coated with gold to a thickness of
7 nm
using a Polaron SC7620 sputter coater. Microscopy was per-
formed at 10 and 12 kV acceleration voltage and with mag-
nication ranging from 50 to 1,000. Scale bars are given
at the bottom of each image. For analysis of the atomic com-
position of the plant material, energy dispersive spectroscopy
(EDS) was used together with SEM. The EDS system was
an INCA 400 from Oxford Instruments (Oxford, UK) using
Biotechnol. Prog., 2009, Vol. 25, No. 2
Table 1. Composition of Ground Wheat Straw Particles; Yield Values Are Given in % (w/w)*
Particle Size Glucose Xylose
Reference 24 cm 49.8ab  2.2 19.2a  0.9
7071000 lm 53.6a  2.1 20.8a  2.4
250500 lm 48.0ab  2.6 19.1a  0.9
53149 lm 43.4b  4.2 15.7a  3.6
* Levels in the same column followed by different roman letters a, b are signicantly different at P  0.05. ** Acid insoluble lignin.
a Si detector crystal with 133 eV resolution measured at
5.9 keV.
Results and Discussion
Compositional analysis
The monosaccharide compositions of the different par-
ticles of wheat straw were analyzed after acid hydrolysis to
monitor any variation in the biomass composition of the dif-
ferent particle size fractions. The compositional data indi-
cated, as expected,9,10 that the wheat straw particles were
mainly made up of glucan, xylans, and lignin which consti-
tuted on average
44%,
1620%, and 17% by weight,
respectively, of the dry matter of the milled straw particles
(calculated from the data in Table 1). The levels of glucose
and xylose appeared to decrease with decreasing particle size
within the samples analyzed, whereas the levels of arabinose,
galactose, lignin, and ash increased with decreasing particle
size (Table 1). These changes indicated a slight decrease in
the cellulose plus hemicellulose ratio to lignin in the smallest
particles as compared to the larger. EDS revealed that min-
eral particles were found to be particularly associated with to
the smallest substrate particles (Figure 1). The mineral par-
ticles were found in varying sizes and consisted mostly of
silicates, aluminum salts, and traces of salts of potassium,
iron, calcium, and magnesium. These ndings signied that
the relative increase in ash content in the smallest particles
might be due to collection of excess dust or sand from soil
crust associated with the wheat straw.22
Effect of substrate particle size reduction on
monosaccharide release from wheat straw
(not wet oxidized)
For the wheat straw particles which had not been wet oxi-
dized, the nely ground straw particles were more amenable
than the larger substrate particles to enzymatic hydrolysis as
assessed by monosaccharide release. Both the release of glu-
cose and the release of xylose thus increased when the parti-
cle size of the straw substrate was reduced (Figure 2, lower,
unlled labels). Compared to the reference substrate samples,
which had a length of 24 cm, the release from the smallest
straw particles (53149 lm) increased with 39% and 20% of
the theoretical maximal release of glucose and xylose,
respectively, after 24 h of hydrolysis (Figure 2).
It was also noticeable that the ground wheat straw had a
basal level of both glucose and xylose before enzymatic hy-
drolysis. This phenomenon might be a result of reactions cat-
alyzed by enzymes produced by microorganisms that
inhabited the wheat straw before and/or after harvest because
it may be speculated that the access of these enzymes to the
substrate was also increased on the smaller particles as com-
pared to the larger. We consider it unlikely that the grinding
itself would result in signicant release of monosaccharides.
401
Arabinose Galactose Lignin** Ash
2.8a  0.2 0.9ab  0.4 16.9b  0.7 1.9ab  0.9
2.8a  0.2 0.6b  0.2 16.2b  0.2 1.0b  0.1
3.0a  0.1 1.2ab  0.0 16.5b  0.3 1.4ab  2.0
3.2a  0.3 1.6a  0.0 19.3a  0.4 4.5a  0.1
Figure 1. SEM image of pretreated straw (53149 lm) show-
ing mineral-rich (dust and sand) particles of various
sizes.
Sizes of the particles can be measured by the bar given in the
bottom of each SEM image.
Effect of substrate particle size reduction on
monosaccharide release from wet oxidized wheat straw
After the pretreatment involving alkaline wet oxidation at
195 C for 10 min, the wheat straw was enzymatically hydro-
lyzed under the same conditions as the wheat straw samples
which had not been wet oxidized. In general, reduction of
the substrate particle size increased the glucose release when
wet oxidation was introduced prior to the enzymatic treat-
ment. The reference substrate samples and the particles of
7071,000 lm gave similar glucose releases during the enzy-
matic hydrolysis treatment, whereas the particles in the
ranges of 53149 and 250500 lm resulted in markedly
increased glucose releases as compared to the corresponding
reference substrate samples, especially the glucose release
from the smallest particles (53149 lm) was consistently
higher than the glucose release from the other wet oxidized
particles, releasing up to 90% of the theoretical maximal
release after 24 h of enzymatic hydrolysis (Figure 2A). Com-
pared to the wet oxidized reference substrate samples the
glucose release of the smaller particles increased by 28% of
the theoretical maximal release of glucose after 24 h of hy-
drolysis (Figure 2A). In contrast, the reduction of the particle
size (from 24 cm down to 53149 lm) only increased the
enzymatic xylose release on the wet oxidized straw particles
by maximum 8%, thus not showing a striking effect of parti-
cle size diminution prior to wet oxidation pretreatment and
24 h enzymatic hydrolysis (Figure 2B).
Comparison of the glucose and xylose release after the en-
zymatic hydrolysis of the wet oxidized vs. the not wet oxi-
dized wheat straw substrate samples thus clearly showed that
wet oxidation increased the monosaccharide release for all
particle sizes, and thus signied that the enzymatic glucose
release was more affected than the xylose release by sub-
strate particle size reduction prior to wet oxidation.
402
Figure 2. Comparison of glucose (A) and xylose (B) release from hydrolysis of pretreated wheat straw (n, 53149 lm; ^, 250500
lm; ~, 7071,000 lm) and hydrolysis of wheat straw that had not been wet oxidized (&, 53149 lm; ^, 250500 lm; ~,
7071,000 lm).
Reference particles are given by punctured lines.
However, when comparing the monosaccharide releases
obtained from the wet oxidized straw particles with those
from the not wet oxidized particles, the calculated ratios
showed that the wet oxidation generally affected the xylose
yields more than the glucose yields because the relative
increase in xylose release with wet oxidation was higher
than the relative glucose increase across all particle sizes
(Figure 3). The ratios of the enzymatic hydrolysis data
between 3 and 12 h revealed a particularly large difference
between the relative increases in glucose and xylose yields
with wet oxidation, and moreover showed a signicantly
larger effect of the wet oxidation pretreatment on the larger
particles (7071,000 lm) than on the smaller particles (53
149 lm) (Figure 3). Similar results were recently reported
by Zeng et al. in 200722 on hot water pretreated corn stover
samples. They found that the increases in conversion of cel-
lulose to glucose, xylan to xylose, and galactan to galactose,
obtained by hot water pretreatment of corn stover, were
more pronounced for larger substrate particles than for small
substrate particles.22 We hypothesize that a possible explana-
tion for this phenomenoni.e. that the larger substrate par-
ticles were affected relatively more by the wet oxidation
pretreatment than the smaller particlesmay be a result of a
larger impact of topological changes on the substrate par-
ticles imposed by the wet oxidation treatment on the larger
particles relative to on the smaller particles. In the present
work, the largest difference between the impact of the parti-
cle sizes on the effect of the wet oxidation was found after
12 h of hydrolysis for both glucose and xylose (Figure 3).
After 24 h the difference among the particle sizes had dimin-
ished. Comparison of the yields of monosaccharide per kg
dry matter obtained after 24 h of enzymatic hydrolysis
revealed a clear increase in monosaccharide yields, i.e. glu-
cose and xylose, with substrate particle size reduction for the
ground wheat straw which had not been wet oxidized (Table
2). However, for the wet oxidized straw samples, only the
glucose yields increased in response to particle size reduc-
tion, whereas the xylose yields in fact decreased when the
substrate particles size was reduced, although they were still
much higher than those obtained for the corresponding wheat
straw samples that had not been wet oxidized (Table 2).
With respect to the total yield of monosaccharides, i.e. glu-
cose xylose, the yields increased with particle size reduc-
tion when the wheat straw had not wet oxidized, but
remained almost constant among the differently sized
Biotechnol. Prog., 2009, Vol. 25, No. 2
Figure 3. Plot of the increase in monosaccharide release when
comparing the wet oxidized wheat traw with wheat
straw that had not been wet oxidized.
Ratios plotted are labeled by n, 53149 lm; ^, 250500 lm;
~, 7071,000 lm for glucose ratios and &, 53149 lm; ^,
250500 lm; ~, 7071,000 lm for xylose ratios. Ratios were
calculated by comparing release of monosaccharide from hy-
drolysis of pretreated wheat straw particles with monosaccha-
rides from hydrolysis of wheat straw particles that had not
been wet oxidized.
particles for the wet oxidized samples (Table 2). This nding
showed that grinding affected the yield of monosaccharides,
however, the decrease in xylose yields with particles size
reduction due to the wet oxidation erased the effect of grind-
ing on the total yield of monosaccharides.
On the basis of previously obtained data,13,14 we had
expected that the enzyme catalyzed monosaccharide release
would increase when the substrate particles were diminished
prior to the wet oxidation treatment. However, the increase
in glucose and xylose release obtained after wet oxidation as
a result of the particle size reduction was lower than
expected. Although this study was not designed to examine
the inuence of wet oxidation treatment on the subsequent
enzyme catalyzed monosaccharide release, the data indicate
that the wet oxidation was equally harsh on all particles sizes
thus in effect overriding any potential inuences of the struc-
tural differences, e.g. surface to volume ratios, of differently
sized straw particles. These results thus indicated that an
effective pretreatment method can reduce the need for reduc-
ing the substrate particle size. This conclusion is in accord-
ance with the ndings published by Mosier et al.5 However,
it is worth noting, that our results did signify an increased
glucose release when the straw substrate particle sizes were
reduced from 7071,000 lm to 250500 lm. In turn, this
Biotechnol. Prog., 2009, Vol. 25, No. 2
Table 2. Yield of Glucose, Xylose, and Glucose 1 Xylose Due to Particle Size and Treatment
53149 lm
Particle size Ground Wet Oxidized
1
Yield glucose (g kg DM ) 253 452
Yield xylose (g kg DM1) 40 160
Yield glucose xylose (g kg DM1) 293 612
Yield values are given in g monosaccharide per kg dry matter of unprocessed wheat straw.
suggests that grinding of straw to particles smaller than
1,000 lm in diameter enhance the release of monosaccha-
rides during the subsequent enzymatic hydrolysis.
Another noteworthy difference between the results
obtained on the wet oxidized wheat straw as compared to
the not wet oxidized wheat straw samples was the basal lev-
els of glucose and xylose. The basal levels of glucose and
xylose on the not wet oxidized wheat straw corresponded to
57% and 13% of the theoretical maximal release of glu-
cose and xylose, respectively. This level of available mono-
saccharides disappeared after wet oxidation pretreatment.
This monosaccharide loss might be a result of the high tem-
perature employed during the wet oxidation which can lead
to degradation of free monosaccharides. The wheat straw
particles were wet oxidized at 195 C and the melting point
of b-D-glucose is approximately 150 C.23 This slight loss of
fermentable monosaccharides during the wet oxidation pre-
treatment thus reduced the maximal yield obtainable from
the raw material. In this work the whole slurry, including
both dissolved and insoluble dry matter was used and all cal-
culations were calculated to the dry matter base of the origi-
nal straw biomass (not wet oxidized). Thus, all other things
being equal, the results of this work indicate that an optimal
balance exists between the size of the lignocellulose sub-
strate particles and the severity of the pretreatment with
respect to obtainment of maximal yields of glucose and
xylose from the biomass raw material.
SEM of wheat straw
SEM showed that the surface of wheat straw particles
changed signicantly with grinding, wet oxidation, and enzy-
matic hydrolysis. Images of the ground, not wet oxidized,
wheat straw particles displayed the smooth, palisadic surfaces
of the epidermal layers of wheat straw and revealed the elon-
gated nature of the cells in the inner parts of the straw, which
are most likely the parenchyma cells (Figure 4A). The outer
layers of the straw, i.e. the epidermis, were found to be made
up of a double-layer comprising an outer, smooth layer and
an inner layer having a porous, creased structure in between
(Figures 4AC). The grinding appeared to induce partial sepa-
ration of these epidermal cell layersespecially at the cut
ends of the particlesresulting in the display of the more po-
rous internal structures, interpreted as the parenchyma cells,
of the straw particles (Figures 4B,C). The SEM also revealed
that further grinding of the particles to sizes between 707 and
1,000 lm resulted in roughly cut ends of particles and further
display of the xylem and phloem structures of the straw,
appearing as internal porous, creased tube-like structures (Fig-
ure 4D). Moreover, the SEM images showed that the particles
varied in size and shape (Figures 4D,E). Particles that ranged
between 250500 lm and 53149 lm displayed the palisadic
surface of the epidermis, which appeared resistant to enzy-
matic attack (Figure 4E). At the same time, the epidermal sur-
403
250500 lm 7071,000 lm
Ground Wet Oxidized Ground Wet Oxidized
153 430 123 395
28 208 23 215
182 638 147 610
face layer appeared more torn and ripped with decreased
particle size (Figures 4E,F). In addition, highly torn, rough
edges were seen for the smallest particles (53149 lm) (Fig-
ure 4F). The images of the particles between 53 and 149 lm
also revealed that the particles were mainly cubic and edged,
i.e. nonspherical (Figure 4F)this nding is in contrast to the
assumptions made by Zeng et al.22 regarding the spherical
shape of corn stover particles.
Images of the wet oxidized reference substrate samples
showed that the wet oxidation pretreatment resulted in an
altered surface structure and rupture of the parenchyma cells
(Figure 5A). The cell structures seemed more loosely con-
nected and the plane surfaces seemed rounded after the wet
oxidation pretreatment. Moreover, spherical depositions and
various crumbled sheet like or tube like structures, probably
disrupted phloem cells, were found to be attached to the
solid structures (Figure 5B). Spherical depositions were also
observed by Selig et al.24 in 2007 when evaluating dilute
acid pretreatment of maize stems. The images of the straw
particles of 7071,000 lm revealed similar effects of the
pretreatment, although the consequences of the wet oxidation
pretreatment seemed more pronounced as compared to the
reference substrate samples; in addition, a number of sheet-
and soft tube-like structures, most likely disintegrated
phloem and xylem cells, could be discerned from the SEM
images (Figure 5C), such sheet like structures were not seen
on the corresponding particles that had not been subjected to
wet oxidation pretreatment (Figure 4D). By comparison of
the size bars on the SEM pictures the wet oxidized particles
in the range 250500 lm (Figure 5C) and 53149 lm (Fig-
ure 5D) appeared to be somewhat swollen as compared to
the corresponding particles prior to wet oxidation. This appa-
rent increase in particle size might be due increased water
binding because of the rupture and display of the hemicellu-
lose fraction, which has a high water binding capacity. The
particles in the range 250500 lm (Figure 5C) and 53149
lm (Figure 5D) only had little intact epidermal structures
left, although areas of smooth, palisadic surfaces were still
visible on even the smallest particles ranging from 53 to 149
lm (Figure 5E). In general, the smaller particles, of 250500
lm and 53149 lm, displayed only few signs of rupture of
the internal parencyma, phloem and xylem structures, as the
tubings making up the porous elongated particle structures
were relatively intact (Figures 5C,D), but crevices and holes
appeared after wet oxidation pretreatment (Figure 5F).
Although some of these holes might be physiological struc-
tures of the straw, i.e. simple pits, bordered pits, or plasmo-
desmata, it cannot be excluded that some of them were a
direct result of the wet oxidation treatment. The relatively
insignicant effect of the extent of particle size reduction
with respect to yields obtained from the wet oxidized wheat
straw particles could be explained by the increased surface
area of the larger particles and notably to be a result of the
removal of the palisadic epidermal surface layer as a
404 Biotechnol. Prog., 2009, Vol. 25, No. 2

Figure 4. SEM images of wheat straw samples (not wet oxidized).

Numbers in parentheses correspond to magnication. (A) Reference substrate sample (50); (B) Reference substrate sample (250), (C) Reference sub-
strate sample (250), (D) 7071,000 lm (100), (E) 250500 lm (100), (F) 53149 lm (500). Sizes of the particles can be measured by the bars given
in the bottom of each SEM image.

consequence of the general tearing up of the substrate sur-
face structures induced by the wet oxidation. Nevertheless,
the conclusion still was that relatively more glucose and
xylose was released from the smaller particles (Figure 2).
This conclusion could be a result of the presence of an
increased number of edges and more accessible cellulose
chains or cellulose (micro)brils allowing improved sub-
strateenzyme interactions (22).
The SEM images of the particles that had been wet oxi-
dized and hydrolyzed enzymatically for 24 h revealed partial
presence of the initial structures of the wheat straw (Figure
6A). The images showed that the structures remaining after
the enzymatic hydrolysis were similar for the differently
sized particles. Edges of rough cuts and signicant presence
of aligned, tube-like structures, probably remnants from the
xylem or phloem, were visible (Figures 6B,C). For the
smallest, hydrolyzed particles, 53149 lm, mainly granular
structures remained, but ordered, physiological structures,
presumably rudiments of the epidermis and xylem, could
also be seen (Figure 6D). These SEM images thus clearly
Biotechnol. Prog., 2009, Vol. 25, No. 2 405

Figure 5. SEM images of wet oxidized wheat straw samples.

Numbers in parentheses correspond to magnication. (A) Reference substrate sample (50), (B) Reference substrate sample (500), (C) 7071,000 lm
(500), (D) 250500 lm (250), (E) 53149 lm (500), (F) 53149 lm (1,000). Sizes of the particles can be measured by the bars given in the bottom
of each SEM image.

indicated that a signicant part of the straw structures was
left almost intact by wet oxidation and the enzymatic
actions, and that this was the case irrespective of the initial
particle size. The presence of apparently dense, compact
structuresin addition to the more porous structures with
clear tube-like structuresin the images (Figure 6C) might
be a result of the centrifugation that was required to prepare
the particles for the SEM analysis after the enzymatic
hydrolysis.
In general, the yields of glucose and xylose were above
78% of the theoretical maximal release, except for the glu-
cose release from the particles between 707 and 1,000 lm
and from the reference substrate samples. The reference sub-
strate samples (Figure 6A) thus still contained up to 38% by
weight of glucose and 22% by weight of xylose after wet ox-
idation pretreatment and enzymatic hydrolysis. However, the
particles between 53 and 149 lm (Figure 6D) only contained
up to 5% (w/w) glucose and 12% (w/w) xylose. The remain-
ing part of the plant cell wall material visible on the SEM
images of the hydrolyzed wheat straw particles must then be
made up of other structural polysaccharides as well as
lignin.
406
Figure 6. SEM images of wet oxidized wheat straw after enzymatic hydrolysis for 24 h.
Numbers in parentheses correspond to magnication. (A) Reference substrate sample (250), (B) 7071,000 lm (250), (C) 250500 lm (250),
(D) 53149 lm (250). Sizes of the particles can be measured by the bars given in the bottom of each SEM image.
Together with the analyses of monosaccharide release, the
microscopic images of the particles revealed that the struc-
ture of the wheat straw particles changed as a result of the
grinding, with wet-oxidation, and nally with enzymatic hy-
drolysis. The main alterations included physical changes of
the epidermal, palisadic surface structures, enhanced display
of the inner, porous structuresincluding the parenchyma,
xylem, and phloem cellsas well as tear and rupture of the
sides and edges of the particles to display more edges and
surface area. A comparison of the SEM images indicated
that the different physical changes together contributed to
the increased release of monosaccharides by enzymatic hy-
drolysis. The increased monosaccharide release thus seemed
to be a result of improved enzyme to substrate interaction
because of the increase in accessible surface area including
the increase in edges as well as enhanced access of enzymes
to the inner surfaces of the straw particles. It is tempting to
infer that the apparently shorter entry and exit path lengths
of the smaller particles also improved the mass transfer and
in turn enhanced the enzymatic degradation. These ndings
invite to further experimental substantiation and modeling of
the kinetics of the enzymatic reactions.
Effects of particle geometry on accessibility of
degradable polysaccharides
The experimental data showed that a decrease in particle
size increased the release of monosaccharides as accom-
Biotechnol. Prog., 2009, Vol. 25, No. 2
plished by enzymatic hydrolysis (Table 2). This increase in
monosaccharide release and yield was expected and may be
assumed to be a result of the increase in accessible surface
area of the particles in turn increasing the accessible area for
enzymatic attack. The ground wheat straw particles analyzed
in this work were found to form cubic, edged particles with
porous ends upon grinding rather than spherical, porous par-
ticles (Figure 4D). The cubic particles from the grinding
resulted in enlarged outer surface areas. However, by grind-
ing of the particles also the entry to the porous inner surface
appeared to be enhanced, which could give rise to increased
enzymesubstrate interactions. Reducing the particle size
from 1,000 lm in length to 53 lm resulted in a decrease of
the outer surface area of one particle from 6.0  106 lm2 to
1.7  104 lm2 (Table 3). However, if assuming similar den-
sities of the smallest particles (53 lm) and the largest (1,000
lm) then the total surface area of the smallest particles will
be 18.9 times larger per unit weight (Table 3). However, the
particles of the ground wheat straw were cubic and porous
and thus appeared to have both an outer and an inner surface
which increases the surface area. In addition, to increase the
hydrolysis yield also the binding of enzyme (or in this case
rather the binding of the cellulose binding modules of the
cellulolytic enzymes) to the substrate needs to be taken into
consideration. The cellulose binding modules (CBMs) of the
two dominant Trichoderma reesei cellobiohydrolases, T. ree-
sei Cel7A [CBHI] CBM, and T. reesei Cel6A [CBHII], have
been shown to have dened binding sites on corners of
Biotechnol. Prog., 2009, Vol. 25, No. 2
Table 3. Properties of the Wheat Straw Particles Analyzed
Sieve Size Outer Surface Area Per Surface Ratio Increase**
(lm) Particle* (lm2) (Index 1,000 lm Particles)
7071,000 3.00  1066.00  106 1.41.0
250500 3.75  1051.50  106 4.02.0
53149 1.69  1041.33  105 18.96.7
Surface area and edge length are shown for each particle size. Ratios show the increases in surface area and edge length when comparing one particle
size with a particle of 1,000 lm in size.
* Assuming one cubic particle and only outer surface. ** Per unit weight ratios between total outer surface area of one particles size and a particle
with size of 1,000 lm (assuming similar density for large and small particles). Assuming cubic particles and only outer surface. Per unit weight ratios
between edge length of one particle size and the edge length of particles with size of 1,000 lm (assuming similar density for large and small particles).
cellulose crystals.25 From this it can be inferred that an
increased display of the particular face that provide binding
sites for the CBMs would increase the rate of reaction of the
cellulolytic enzymes, notably the cellobiohydrolases, on in-
soluble cellulose substrates. Although an increased surface as
a result of a simple cut through a straw substrate may not
directly display the binding sites for the CBMs of the cello-
biohydrolases, the worn corners of bundles of crystalline vided a new insight into the changes in surface properties of
cellulose will increase the available binding sites for CBMs,
as shown for Valonia cellulose crystals.25 In relation to this
work, it can thus be hypothesized that the more ends and
edges the substrate particle has, the more substrate sites may
be available for binding of CBMs and in turn for enzymatic
attack.25,26
By reducing the particle size from 1,000 lm to 53 lm the
length of edges decreased from 1.2  104 to 6.4  102 (Ta-
ble 3). Comparison of the total edge length achieved by par-
ticles occupying the volume of one 1,000 lm particle (i.e. a
volume of 1  109 lm3), the reduced particle size increased
the total surface area by up to 356 times (Table 3). When
including the inner edges, the increase in accessible substrate
surface most likely increased even more. The increase in
edges and ends could in addition reduce the nonproductive
adsorption of cellulases to the lignin, which is found to
occur unspecically in the plant material during hydrolysis.27
The increase in the enzymatic monosaccharide release due to
particle size reduction has previously been attributed to be a
result of a higher surface area.22 However, although the total
surface area and the total length of edges increased 18.9 and
356 times, respectively, when the particle size was reduced
from 1,000 lm to 53 lm (Table 3), the release of monosac-
charides obtained in this work (Figures 2A,B and 3) did not
reach the maximal theoretical release. It was found to
increase by maximum 2.5 times (Figures 2A,B). Hence,
other mechanisms may still be in play to retard the enzy-
matic cellulose and xylan hydrolysis on lignocellulosic
biomass.
A prevalent theory is that the inuence of particle size on
rates of enzymatic hydrolysis may be more related to mass
transfer conditions within the particles that made up the sub-
strates rather than correlated to the available surface area
available for enzymatic attack.15,16 Our data supported the
idea that other factors or events than the apparently accessi-
ble surface area may govern (or respectively promote) the
enzymatic hydrolysis. Such factors presumably include the
nonproductive adsorption of enzymes to the solid substrate
structure, different impact of product inhibition as a result of
local elevated concentration gradients, which may be pro-
nounced inside the porous structures if mass transfer is low.
Another limiting factor may be that the enzymes are simply
heat inactivated during the prolonged hydrolysis reaction at
50 C.11 Also other mechanisms, such as the cellulases
407
Edge Length Edge Length Ratios
(lm) (Index 1,000 lm Particles)
8.48  1031.20  104 2.01.0
3.00  1036.00  103 16.04.0
6.36  1021.79  103 356.045.0
getting stuck within the cellulose may slow down the en-
zymatic hydrolysis during cellulolytic hydrolysis reactions.28
Nevertheless, the results obtained in this work showed that
the reduction of the particle size had the potential to enhance
the enzymatic monosaccharide release. The inclusion of mi-
croscopic images showing the topological structures of the
substrate and their changes during different treatments pro-
particles during wet oxidation and enzymatic hydrolysis. The
kinetics of the action of multiple enzymes on insoluble,
porous lignocellulosic substrates is highly complex and
requires further study.
Acknowledgments
We acknowledge Novozymes A/S for partial funding of this
work through The Novozymes BioProcess Academy. We also
thank Anne Belinda Thomsen and Tomas Fernqvist from Ris-
DTU for assistance with wet oxidation and Leila Leth from
IPL-DTU for assistance with SEM. At last we also thank Jrn
Erik Pedersen (The Danish Cooperative Farm SupplyDLG)
for supplying the wheat straw.
Literature Cited
1. Sheehan J, Himmel M. Enzymes, energy, and the environment:
a strategic perspective on the U.S. Department of Energys
research and development activities for bioethanol. Biotechnol
Prog. 1999;15:817827.
2. Himmel ME, Ding S-Y, Johnson DK, Adney WS, Nimlos MR,
Brady JW, Foust TD. Biomass recalcitrance: engineering plants
and enzymes for biofuels production. Science. 2007;315:804
807.
3. Chandra RP, Bura R, Mabee WE, Berlin A, Pan X, Saddler JN.
Substrate pretreatment: the key to effective enzymatic hydroly-
sis of lignocellulosics? Adv Biochem Eng Biotechnol. 2007;
108:6793.
4. Mabee WE, Gregg DJ, Arato C, Berlin A, Bura R, Gilkes N,
Mirichnik O, Pan X, Pye EK, Saddler JN. Updates on soft-
wood-to-ethanol process development. Appl Biochem Biotech-
nol. 2006;129132:5570.
5. Mosier N, Wyman C, Dale B, Elander R, Lee YY, Holtzapple
M, Ladisch M. Features of promising technologies for pretreat-
ment of lignocellulosic biomass. Bioresource Technol. 2005;
96:673686.
6. Zhang Y-HP, Himmel ME, Mielenz JR. Outlook for cellulase
improvement: screening and selection strategies. Biotechnol
Adv. 2006;24:452481.
7. Merino ST, Cherry J. Progress and challenges in enzyme devel-
opment for biomass utilization. Adv Biochem Eng Biotechnol.
2007;108:95120.
8. Rosgaard L, Pedersen S, Cherry JR, Harris P, Meyer AS. Ef-
ciency of new fungal cellulase systems in boosting enzymatic
degradation of barley straw lignocellulose. Biotechnol Prog.
2006;22:493498.
9. Rosgaard L, Pedersen S, Langston J, Akerhielm D, Cherry JR,
Meyer AS. Evaluation of minimal Trichoderma reesei cellulase
408
mixtures on differently pretreated barley straw substrate. Bio-
technol Prog. 2007;23:12701276.
10. Kabel MA, Bos G, Zeevalking J, Voragen AGJ, Schols HA.
Effect of pretreatment severity on xylan solubility and enzy-
matic breakdown of the remaining cellulose from wheat straw.
Bioresource Technol. 2007;98:20342042.
11. Rosgaard L, Pedersen S, Meyer AS. Comparison of different
pretreatment strategies for enzymatic hydrolysis of wheat and
barley straw. Appl Biochem Biotechnol. 2007;143:284296.
12. Dusterhoft EM, Engels FM, Voragen AGJ. Parameters affecting
the enzymic hydrolysis of oil-seed meals, lignocellulosic by-prod-
ucts of the food industry. Bioresource Technol. 1993;44:3946.
13. Meyer AS, Jepsen SM, Srensen NS. Enzymatic release of anti-
oxidants for human low-density lipoprotein from grape pomace.
J Agric Food Chem. 1998;46:24392446.
14. Pinelo M, Tress AG, Pedersen M, Arnous A, Meyer AS. Effect
of cellulases, solvent type and particle size distribution on the
extraction of chlorogenic acid and other phenols from spent
coffee grounds. Am J Food Technol. 2007;2:641651.
15. Manseld SD, Mooney C, Saddler JN. Substrate and enzyme
characteristics that limit cellulose hydrolysis. Biotechnol Prog.
1999;15:804816.
16. Sangseethong K, Meunier-Goddik L, Tantasucharit U, Liaw ET,
Penner MH. Rationale for particle size effect on rates of enzy-
matic saccharication of microcrystalline cellulose. J Food Bio-
chem. 1998;22:321330.
17. Converse AO, Grethlein HE. On the use of an adsorption model
to represent the effect of steam explosion pretreatment on the
enzymatic hydrolyses of lignocellulosic substances. Enzyme
Microbiol Technol. 1987;9:7982.
18. Sluiter A. Laboratory analytical procedure 002. Determination
of structural carbohydrate and lignin in biomass. Retrieved from
http://www.nrel.gov/biomass/analytical_procedures.html. 2006.
19. Bjerre AB, Olesen AB, Fernqvist T, Ploger A, Schmidt AS. Pre-
treatment of wheat straw using combined wet oxidation and
alkaline hydrolysis resulting in convertible cellulose and hemi-
cellulose. Biotechnol Bioeng. 1996;49:568577.
Biotechnol. Prog., 2009, Vol. 25, No. 2
20. Adney B, Baker J. Laboratory Analytical Procedure 006. Mea-
surement of cellulose activities. Retrieved from http://
www.nrel.gov/biomass/analytical_procedures.html. 1996.
21. Srensen HR, Meyer AS, Pedersen S. Enzymatic hydrolysis of
water-soluble wheat arabinoxylan. I. Synergy between a-L-arabi-
nofuranosidases, endo-1,4-b-xylanases, and b-xylosidase activ-
ities. Biotechnol Bioeng. 2003;81:726731.
22. Zeng M, Mosier NS, Huang C-P, Sherman DM, Ladisch MR.
Microscopic examination of changes of plant cell structure in
corn stover due to hot water pretreatment and enzymatic hydro-
lysis. Biotechnol Bioeng. 2007;97:265278.
23. ONeil MJ, Heckelman PE, Koch CB, Roman KJ, Kenny CM,
editors. The Merck Index: An Encyclopedia of Chemicals,
Drugs, and Biologicals, 14th ed. NJ, Whitehouse Station: Merck
& Co.; 2006:769.
24. Selig MJ, Viamajala S, Decker SR, Tucker MP, Himmel ME,
Vinzant TB. Deposition of lignin droplets produced during
dilute acid pretreatment of maize stems retards enzymatic hy-
drolysis of cellulose. Biotechnol Prog. 2007;23:13331339.
25. Lehtio J, Sugiyama J, Gustavsson M, Fransson L, Linder M,
Teeri TT. The binding specicity and afnity determinants of
family 1 and family 3 cellulose binding modules. PNAS
2003;100:484489.
26. Divne C, Stahlberg J, Teeri TT, Jones TA. High-resolution crys-
tal structures reveal how a cellulose chain is bound in the 50 A
long tunnel of cellobiohydrolase I from Trichoderma reesei.
J Mol Biol. 1998;275:309325.
27. Palonen H, Tjerneld F, Zacchi G, Tenkanen M. Adsorption of
Trichoderma reesei CBH I and EG II and their catalytic
domains on steam pretreated softwood and isolated lignin.
J Biotechnol. 2004;107:6572.
28. Yang B, Willies DM, Wyman CE. Changes in enzymatic hydro-
lysis rate of Avicel cellulose with conversion. Biotechnol Bio-
eng. 2006;94:11221128.
Manuscript received July 28, 2008, and revision received Aug. 7, 2008.