New Biotechnology  Volume 27, Number 6  December 2010 REVIEW

Lignocellulose pretreatment
severity – relating pH to biomatrix
opening

Review
Mads Pedersen and Anne S. Meyer

Center for Bioprocess Engineering, Department of Chemical and Biochemical Engineering, Technical University of Denmark, Building 229, DK-2800 Kgs. Lyngby,
Denmark

In cellulose-to-ethanol processes a physico-chemical pretreatment of the lignocellulosic feedstock is a

crucial prerequisite for increasing the amenability of the cellulose to enzymatic attack. Currently
published pretreatment strategies span over a wide range of reaction conditions involving different pH
values, temperatures, types of catalysts and holding times. The consequences of the pretreatment on
lignocellulosic biomass are described with special emphasis on the chemical alterations of the biomass
during pretreatment, especially highlighting the significance of the pretreatment pH. We present a new
illustration of the pretreatment effects encompassing the differential responses to the pH and
temperature. A detailed evaluation of the use of severity factor calculations for pretreatment
comparisons signifies that the multiple effects of different pretreatment factors on the subsequent
monosaccharide yields after enzymatic hydrolysis cannot be reliably compared by a one-dimensional
severity factor, even within the same type of pretreatment strategy. However, a quantitative comparison
of published data for wheat straw pretreatment illustrates that there is some correlation between the
hydrolysis yields (glucose and xylose) and the pretreatment pH, but no correlation with the
pretreatment temperature (90–2008C). A better recognition and understanding of the factors affecting
biomatrix opening, and use of more standardized evaluation protocols, will allow for the identification
of new pretreatment strategies that improve biomass utilization and permit rational enzymatic
hydrolysis of the cellulose.

Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 740
Breaking down the diverse recalcitrant biomatrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 740
Chemical alteration of the biomatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 740
Acidic pretreatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 742
Alkaline pretreatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 742
Pretreatment at neutral pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 743
Comparison of the pretreatment conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 743
The severity factor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 744
Evaluation of the severity factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745
Comparing pretreatment strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745
Three strategies analyzed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745

Corresponding author: Meyer, A.S. (am@kt.dtu.dk)

1871-6784/$ - see front matter ß 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.nbt.2010.05.003 www.elsevier.com/locate/nbt 739
 REVIEW New Biotechnology  Volume 27, Number 6  December 2010

Wheat straw as a model substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745

Final remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748

Introduction
Plenty of focus is currently directed toward utilization of ligno- feedstocks, and differs between species and tissues in plants [6]. For
cellulosic biomass not only for liquid biofuels production, notably ‘second generation’ ethanol production special attention is cur-
ethanol [1] and butanol [2,3], but also for production of chemicals rently given to the utilization of four major lignocellulosic biomass
such as furans [4] and acetic acid [5]. Without pretreatment, the feedstocks: straw (wheat and barley), corn stover, hardwood and
lignocellulosic biomass is too resistant to enzymatic saccharifica- softwood. Their chemical compositions are shown in Table 1. In
Review

tion because of the tight bonding between the polymeric consti- softwood the amount of lignin is higher than in hardwood and also
tuents; cellulose, hemicellulose and lignin, and because of the higher than that found in agricultural residues such as straw and
crystalline nature of cellulose [6]. With pretreatment, the inten- corn stover (Table 1). This makes softwood more recalcitrant and
tion is to prepare the cellulose to become more accessible and resistant than the other feedstocks to enzymatic processing [18]. The
susceptible to enzymatic hydrolysis to provide a high monosac- cellulose levels in different hardwood and softwood feedstocks also
charide yield for the subsequent fermentation [7]. Various hydro- vary widely, for example the cellulose level in softwood might range
thermal methods have been examined as pretreatment strategies, from 34 to 50% of the total dry matter (Table 1), and make up
including steam explosion, ammonia fiber explosion and wet between 30 and 40% of the total dry matter in wheat and barley
oxidation [7–9]. The methods differ significantly with respect to straw (Table 1), whereas typical cellulose dry matter levels in corn
pH, temperature, holding time of treatment and types of additives stover vary from 39 to 42% (Table 1). Hence, the cellulose content is
used [7,10]. It is also known that different pretreatment strategies often higher in woody material than in straw and corn stover [19].
do induce differences in yields and enzyme treatment require- Regarding hemicellulose, both the amount and composition of
ments [11,12]. hemicellulose vary among different lignocellulosics (consult
Increased temperature has long been known to increase the Table 1). Furthermore, the hemicelluose, that mainly consists of
severity of the treatment and to improve the subsequent enzy- differentially substituted xylans, is more rich in galactogluco-man-
matic cellulose degradation [13]. Increased temperature might also nan in softwood than in hardwood and in straw, a difference which
result in losses, high-energy consumption and formation of inhi- also influences both the optimal pretreatment and the enzymatic
bitors for both the enzymes and the fermenting microorganisms digestibility of the biomass [20–22].
[14,15]. The purpose of this work is to systematically examine the The different pretreatment strategies currently available vary
influence of the pretreatment parameters on the severity and the with respect to pH, temperature and holding time (Table 2). These
hydrolytic biomass conversion. The advantages and disadvantages variations in conditions affect the severity of the pretreatment and
of different pretreatment strategies are compared and the effects the biomass composition differently during pretreatment. Differ-
and interactions of the major pretreatment parameters tempera- ent substrates therefore require different pretreatment procedures
ture, time and pH are illustrated quantitatively. The data highlight for degradation, and the pretreatment strategy must be designed
the complex, significant impact of the pretreatment pH on the for the particular substrate in mind. In addition to considering the
biomass hydrolysis yields. The severity factor, Ro or log(Ro), based substrate composition, the choice of pretreatment strategy is a
on holding time and temperature is widely used to compare compromise between positive and negative effects because of the
pretreatment yields. Together with time and temperature, the complex and sometimes opposing effects of various pretreatment
pretreatment pH is taken into consideration in the combined parameters and considerations (Tables 1–3).
severity factor R0o or log(R0o ) [16]. However, it is a major question
whether consideration of three factors, that is temperature, hold- Chemical alteration of the biomatrix
ing time and pH, suffice to compare and predict pretreatment The currently published hydrothermal pretreatment strategies can
severity across different pretreatment methods and feedstock sub- roughly be categorized as acidic, alkaline and neutral. In addition,
strates. To answer this question, the monosaccharide yields
obtained in response to different severity calculations are crucially
evaluated for different substrates and pretreatment strategies, and TABLE 1

the problems of comparing monosaccharide yields based on a one- Cellulose, hemicellulose and lignin content in five major ligno-
dimensional model, encompassing only three pretreatment fac- cellulosic biomasses (% DM)
tors are highlighted. Substrate Cellulose Hemicellulose Lignin Reference
Wheat straw 31–39 22–24 16–25 [16,25,43,60]
Breaking down the diverse recalcitrant biomatrices Barley straw 33–40 20–35 8–17 [11,61–63]
The biomatrix that makes up lignocellulosic biomass is principally
Corn stover 39–42 22–28 18–22 [64–67]
built of three main polymers; cellulose (30–50% by weight),
Hardwood 40–55 19–40 18–25 [6,68]
hemicellulose (19–45% by weight) and lignin (15–35% by
weight) [6,17]. The level of each polymer varies in different plant Softwood 34–50 21–35 28–35 [6,69,70]

740 www.elsevier.com/locate/nbt
New Biotechnology  Volume 27, Number 6  December 2010 REVIEW

TABLE 2
Temperature, pH and holding time of selected pretreatment strategiesa
Temperature (8C) pH Time (min) Severity log(Ro) Reference
Lime 54–160 Alkaline 60–4800 2.1–3.9 [7,11,71–74]
AFEX 50–180 Alkaline 5–30 0.4–3.5 [7,9,10,18,44,75]
AWO 175–200 Alkaline 5–15 3.2–4.0 [53,76,77]
IBUS 180–200 Neutral 5–15 3.1–4.1 [35,43]
WO 175–195 Neutral 5–15 3.2–4.0 [10,76,78]
SE 160–230 Neutral 5–10 2.8–4.5 [7,79]
MIC 28–40 Neutral 14–72 (days) 2.4–2.8 [44–47]
2-step 140 Acid + alkaline 10 + 10 2.5 [28]
ASE 180–210 Acidic 2–12 3.0–4.3 [26,79]

Review
AH 30 Acidic 30 0.6 [10,27,80]
DAH 120–140 Acidic 15–60 1.9–2.7 [10,27,80]
The severity, Ro, is calculated from Ro = texp((T(t)  100)/14.74).
a
Lime: lime pretreatment, AFEX: ammonia fiber explosion, AWO: alkaline wet oxidation, IBUS: integrated biomass utilization system, WO: wet oxidation, SE: steam explosion, MIC: microbial,
2-step: 2-step pretreatment first with acid then alkaline catalyst, ASE: acidic steam explosion, AH: acid hydrolysis, DAH: diluted acid hydrolysis.

microbial pretreatment has been evaluated. Recently, some focus since ionic liquids do not have a pH, there is no interaction
has also been directed to the use of ionic liquids as a pretreatment between pH and temperature for pretreatment with ionic liquids.
procedure for lignocellulose [23], but the use of ionic liquids for Therefore, only hydrothermal and microbial pretreatment strate-
pretreatment is still at an exploratory stage [24]. Furthermore, gies are considered in the following.

TABLE 3
Advantages and disadvantages using the selected pretreatment methodsa.
Advantage Disadvantage Substrate Reference
Lime Mild temperatures, low inhibitor production, Slow action of lime causing long Corn stover, wheat straw, [7,11,71–74]
partial removal of lignin pretreatment time, salt production hardwood and other agricultural
residues
AFEX Reduces lignin and hemicellulose content, Expenses on ammonia, not effective Wheat straw, corn stover and [7,9,10,18,44,75]
swelling/decrystallization of cellulose, when high lignin content, other agricultural residues
low inhibitor formation, low temperature not suitable for softwood
AWO Increased digestibility of biomass, Costs on oxygen, temperature, Wheat straw, softwood and [53,76,77]
neutral after treatment alkaline catalyst agricultural residues
IBUS High DM, large-scale, no additives Equipment requirement Wheat straw, corn stover [35,43]
WO Easy to filter, minimal formation of inhibitors, Tank requirement, pressure, Wheat straw, softwood and other [10,76,78]
exothermic process temperature, cost of oxygen agricultural residues
SE High yield of glucose, hemicellulose Pressure, temperature, Wheat straw, wood and other [7,79]
degradation, lignin transformation formation of inhibitors agricultural residues
MIC Low energy requirement, no corrosive Time, microorganisms digest Wheat straw, wood and [44–47]
resistant reactor needed, degrades lignin some of the glucose agricultural wastes
and hemicellulose
2-step Most lignin removed, minor loss of Optimization needed toward Wheat straw [28]
fermentable glucose and xylose, continuous pretreatment
low formation of inhibitors,
low temperature and pressure
ASE High glucose release, removal of Uses SO2 or H2SO4, requires Wheat straw, hardwood and other [26,79]
hemicellulose high temperatures agricultural residues
AH High glucose yield, alters lignin structure, Tank requirements, corrosion, Agricultural residues [10,27,80]
ambient temperatures costs of acid and for acid recovery,
degradation products
DAH Removal of hemicellulose, Degradation products formed, Agricultural residues [10,27,80]
tends to remove some lignin corrosion
a
Lime: lime pretreatment, AFEX: ammonia fiber explosion, AWO: alkaline wet oxidation, IBUS: integrated biomass utilization system, WO: wet oxidation, SE: steam explosion, MIC: microbial,
2-step: 2-step pretreatment first with acid then alkaline catalyst, ASE: acidic steam explosion, AH: acid hydrolysis, DAH: diluted acid hydrolysis.

www.elsevier.com/locate/nbt 741
 REVIEW New Biotechnology  Volume 27, Number 6  December 2010

FIGURE 1
Acidic hydrolysis of polysaccharide backbone in biomass. R = H when hemicellulose, R = CH2OH when cellulose.
Review

FIGURE 2
Pathways for formation of degradation products from glucose and xylose.

Acidic pretreatment increase the digestibility of the biomass [28,38–40]. Alkaline

With acidic pretreatment strategies cellulose and hemicelluloses hydrothermal pretreatment (saponification) at mild conditions,
are degraded via autohydrolysis reactions (Fig. 1). This hydrolysis that is below 1408C, induces cleavage of ester linkages joining
of polysaccharides forms monosaccharides which are readily fer- phenolic acids: the nucleophilic acyl substitution of the ester
mentable [25–27]. Especially, the acid catalyzed pretreatment bonds normally takes place during reaction with an alkaline salt,
strategies show removal of hemicellulose from the biomass typically, sodium hydroxide, to form a carboxylic salt and an
[26,28], and this removal in turn displays the cellulose to enzy- alcohol at room temperature (Fig. 3) [39,40]. The ether bonds
matic attack [22,29]. The release of monosaccharides during acidic linking the lignols in the lignin matrix, however, will only be
pretreatment reduces the need for enzymes to accomplish the catalytically cleaved at high temperatures and the extent of clea-
saccharification process, which can take up a significant part of vage depends on the alkalinity and temperature [40]. In contrast to
the total costs in biofuel production [6,27]. However, acidic pre- the lignin redistribution that has recently been discovered to occur
treatment strategies are also known for the irreversible production with acidic pretreatment strategies [19,21,38,41], lignin droplets
of inhibitors, for example HMF (5-hydroxymethylfurfural) and 2- are not found on the surface of the biomass when highly alkaline
furfuralaldehyde, that are categorized as furan type inhibitors pretreatment methods are used. Rather, lignin appears to be
(Fig. 2) [28,30–34]. If these inhibitors are removed by washing removed from the solid biomass left from alkaline pretreatment
or filtration [35,36] there might be a loss of fermentable mono- [28,38]. Furthermore, with alkaline pretreatment methods, xylan
saccharides, which are not recovered. HMF and 2-furfuraldehyde is partially removed from the solid biomass fraction, resulting in
are also known to inhibit glycolytic enzymes. The furans can an increased hydrolysis yield of glucose as a result of increased
furthermore be hydrated to produce formic acid and levulinic enzyme catalyzed cellulose degradation [28]. Thus, when the
acid that constitutes a part of another group of potential inhibi- lignin is solubilized it is no longer acting as a shield, thus making
tors, that is the weak acids. This group of inhibitors also comprises
acetic acid that is formed by oxidation of acetyl groups released
directly from the substrate hemicellulose [37]. At high concentra-
tions, these inhibitors can affect the fermenting organism via
changing the pH of the cytosol of the microorganism [30].

Alkaline pretreatment
Pretreatment strategies accomplished at alkaline pH values are FIGURE 3
known for the ability to alter the lignin composition and thereby Saponification of ester bonds to yield a carboxylic salt and an alcohol.

742 www.elsevier.com/locate/nbt
New Biotechnology  Volume 27, Number 6  December 2010 REVIEW

both cellulose and xylan more susceptible to enzymatic attack.

Obviously, the extent of these events differs at various pH levels
and with various catalysts [28]. Unfortunately, monomeric lignin
compounds released during the lignin degradation comprise the
third group of potential inhibitors; that is phenolic compounds
such as aromatic acids, catechol, 4-hydroxybenzaldehyde and
vanillin. These also inhibit the hydrolytic enzymes used in the
glucose conversion by denaturation and precipitation [14,42].

Pretreatment at neutral pH
Pretreatment strategies accomplished without a catalyst changing
the pH have the advantage of the reduced need of addition of
chemicals [43]. Hydrothermal pretreatment using only (neutral)

Review
water or steam favors the enzymatic hydrolysis that takes place
afterwards (e.g. at pH 5 [28]). Even though run at neutral pH,
inhibitors are produced during pretreatment because of the high
temperatures [44]. If the pretreatment water is recycled acetic acid
might accumulate, and this will lower the pH and develop an
acidic pretreatment process [35]. FIGURE 4
Compared to hydrothermal strategies the temperature is kept Comparison of temperature and pH in main pretreatment methods. For
pretreatment abbreviations, please consult Table 2.
low in microbial pretreatment, thus requiring only low energy
input and no need for a corrosion resistant reactor [45]. However,
the chemical energy needed by the microorganisms is taken from utmost importance to consider at least the pH, temperature, as
the biomass carbohydrates. Even though some lignocellulolytic well as holding time, when attempting to design and understand
fungi are known to degrade lignin, they also break down and biomass pretreatment processes.
consume some cellulose and hemicellulose, and produce some The differential chemical alterations of the biomass in response
degradation products as well. Thus, microbial pretreatment does to the pretreatment conditions were considered when picturing
not release the maximal amount of fermentable monosaccharides the biomatrix opening in the model depicted in Fig. 6. As opposed
unless treated in very specific ways [46,47]. Moreover, the holding to previous sketches of the events occurring during lignocellulose
time needed for microbial pretreatment often exceeds two weeks biomass pretreatment (see e.g. Ref. [7]) this sketch illustrates the
(Table 2). This is very long as compared to the short hydrothermal very different effects on the biomass constituents of high and low
pretreatments, and microbial pretreatment therefore requires a
huge storage area if pretreatment shall take place continuously. To
overcome the time consuming procedure with microbes growing
on biomass for degradation and removal of lignin, it is more
feasible to identify and isolate the genes encoding the enzymes
responsible for making the biomass digestible [48]. Thus, to
develop novel microbial pretreatment strategies, screening and
characterization of new fungal species able to degrade lignocellu-
losic biomass remain crucially important [49,50].

Comparison of the pretreatment conditions

In Fig. 4 the strategies outlined in Table 2 are mapped with regard
to their temperature and pH requirements. This mapping signifies
the wide distribution of temperatures and pH values employed in
different pretreatment processes (Fig. 4). It is noteworthy that
there are no pretreatment methods in the center, that is at neutral
pH and medium temperatures, neither are there any methods at
low temperature and high pH values (Fig. 4). However, when
mapping the interactions of the factors, that is pH, temperature
and holding time, from the strategies presented in Table 2, it
becomes quantitatively evident that the level of one factor affects
that of another, for example an increase in temperature reduces
the pretreatment holding time requirement (Fig. 5). It is also clear
that at neutral pH values, the time required for the pretreatment
increases as compared to the requirements at the more extreme pH FIGURE 5
values (Fig. 5). Furthermore, the temperature requirement varies as Surface plot of factor interactions between temperature, pH and holding
a function of pH as well as of holding time. This is why it is of time.

www.elsevier.com/locate/nbt 743
 REVIEW New Biotechnology  Volume 27, Number 6  December 2010
Review

FIGURE 6
Sketch of pretreatment of lignocellulose as affected by temperature and final pH. Gray ‘veil’ indicates lignin sheath; orange and red tubes illustrate cellulosic fibrils
and microfibrils, respectively; black curved lines illustrate hemicellulose (xylan); the gray dots on the cellulose microfibrils in the low pH region illustrate
redeposited lignin.

pretreatment pH, and indicates the influence of increased tem- temperature, is generally used for comparison of various treat-
perature on the cellulose, hemicellulose and lignin compounds: at ments:
the alkaline conditions the lignin is degraded liberating phenolic Z b    
TðtÞ  100 TðtÞ  100
lignin monomers, whereas spherical lignin droplets (depicted as Ro ¼ exp dt ¼ t  exp (1)
a 14:75 14:75
gray bullets) are redeposited onto the solid fractions under acidic
pretreatment (Fig. 6). Figure 6 also shows the degradation of where t is the holding time of treatment in min, T(t) is the
hemicellulose and subsequently of cellulose to monosaccharides treatment temperature, 100 is the reference temperature. The
at strong acidic pretreatment – events that occur as a consequence fitted value (14.75) of the arbitrary constant v, is based on the
of the acid catalyzed hydrolysis shown in Fig. 1. Thus, in principle, activation energy when assuming pseudo first order kinetics [25].
with acid pretreatment, the biomass might be degraded directly to This constant has been evaluated and in some cases optimized;
monomers, an effect that is exploited in the Arkenol process however, 14.75 is often used without any concern [56]. The
[51,52]. severity of the pretreatment strategies studied in this work is
compared in Table 2 as the logarithmic value of Eqn (1).
The severity factor The severity factor was initially used to control the pulping
To compare the efficiencies of different pretreatment strategies in processes in the paper industry, but it was reintroduced for com-
preparing the cellulosic biomass for enzymatic conversion, the parison of steam explosion pretreatment severities on lignocellu-
outcomes of a benchmark cellulase treatment on the pretreated losic biomass. The equation then evolved into Eqn (2) to account
biomass are usually compared [12,16,53]. Alternatively, the fer- for the addition of catalysts [57]. Eqn (2) is used for correlation of
mentation broth can be assessed for its content of fermentable the xylan solubilization by sulfuric acid and the lignin reduction
monosaccharides and/or, if ethanol is the end product, the yields by sodium hydroxide where C is the concentration of the chemical
of ethanol produced on the hydrolysates might be assessed [10]. catalyst. The n-value is a new arbitrary constant fitted to be 0.849
However, to focus on the direct effect of the pretreatment itself, and 3.90 for sulfuric acid and sodium hydroxide, respectively [54].
the severity factor is often used to describe the lignin reduction  
TðtÞ  100
and xylan solubilization [54]. Overend and Chornet [13] used the P M o ¼ t  Cn  exp (2)
14:75
factor of Brasch and Free [55] to evaluate the severity of different
individual treatments. At this point the factor was referred to as the Eqn (1) does not take into account that all catalysts do not affect
reaction ordinate. Now, the factor Ro, incorporating an integration the biomass equally, and in particular that various acids or bases
of the time period used in the pretreatment done at a certain influence the biomass very differently (Fig. 6). Abatzoglou et al.

744 www.elsevier.com/locate/nbt
New Biotechnology  Volume 27, Number 6  December 2010 REVIEW

TABLE 4
Severity of different selected scenarios when varying temperature, holding time and pH
# Method Temperature (8C) Time (min) pH log(Ro) log(R0o ) log(R00o )
1 MIC 25 10,080 5 1.8 3.2 3.8
2 AH 30 60 1 0.3 1.3 5.7
3 Alkali 140 10 13 2.2 10.8 8.2
4 ASE 180 10 2 3.4 1.4 8.4
5 ASE 250 10 1 5.4 4.4 11.4
log(Ro) is derived from Eqn (1),log(R0o )
from Eqn (4) and log(R00o )
from Eqn (5).
MIC: microbial pretreatment, AH: acid hydrolysis, ASE: acidic steam explosion, Alkali: proposed highly alkaline pretreatment.

Review
[58] developed Eqn (3), where the effect of an acid catalyst is taken different pretreatment strategies seems to be too unreliable with
into account by adding the final proton concentration in the the severity factor calculations currently in use. This is shown in
pretreatment. Fig. 8 where enzymatically released glucose and xylose and lignin
R0o ¼ Ro  ½Hþ  (3) removal are shown in response to severity factor calculations from
Eqns (1), (4) and (5). No direct correlations can be discerned
The logarithm of Eqn (3) gives:
between the releases and the severity factors or the factors alone
logðR0o Þ ¼ logðRo  ½Hþ Þ ¼ logðRo Þ  pH (4) (Fig. 8 and Supplementary Figure 1). Hence, although it is clear
The severity factor logðR0o Þ has been used to compare the action that the pretreatment temperature, pH and holding time are the
of lignocellulose pretreatment on the digestibility of cellulose with major factors affecting pretreatment efficiency, many other para-
respect to xylan solubilization and lignin reduction [29]. However, meters apparently affect the biomatrix opening, making it unrea-
this severity factor ‘favors’ the effect of low pH, which is explained listic to explain or predict the yields obtained by one simple
by the origin of the equation, that is from acidic pretreatment of equation. Apart from the influence of the biomass composition
aspen [57]. When comparing cellulose digestibility as a result of (Table 1), another important factor, that has received too little
pretreatments at extremely different pH values, one must there- attention, is the enzymatic hydrolysis. Not all pretreatment stra-
fore reconsider the use of Eqn (4). When the pH is increased by tegies require subsequent enzymatic hydrolysis, but when enzy-
ammonia, sodium hydroxide or sodium carbonate in alkaline matic hydrolysis is used to evaluate the pretreatment, various
pretreatments the cellulose also becomes more susceptible to enzyme combinations have used, in various dosages and at various
enzymatic attack, an effect which is not considered by Eqn (4) reaction conditions [16,29,59]. It would be an important step
[9,28,29]. Thus, the influence of increasing pH should also show to forward if the pretreatment evaluations could be standardized
affect the severity of a certain pretreatment strategy. Eqn (5) is with respect to at least the enzymatic hydrolysis conditions, that
derived by having acidic and alkaline pretreatments affect the is a standardized enzyme blend, dosage and fixed enzyme reaction
biomass digestibility equally. conditions.

logðR00o Þ ¼ logðRo Þ þ jpH  7j (5)

Three strategies analyzed
By the use of Eqn (5), the severities of pretreatment strategies at A comparison of the monosaccharide yields (glucose and xylose)
various pH values are easily compared by the use of one equation. obtained in response to the different severity factor models for
A comparison of the use of Eqns (1), (4) and (5) is shown in Table 4 three principally different pretreatment strategies: acidic steam
for five different selected scenarios, and this comparison illustrates explosion (ASE), alkaline wet oxidation (AWO) and lime (Lime)
how the use of Eqn (5) gives a more fair comparison of the pretreatment, reveal the significant lack of correlation between the
pretreatment severities even at widely different pretreatment pH monosaccharide releases and the severity factor values (Fig. 9 and
values. Alkaline pretreatment is known to make biomass digesti- Supplementary Figure 2). In this case, it is important to note, that
ble, but the combined severity factor (R0o ) still rates the alkaline even though one single pretreatment strategy often has set limits
treatment the lowest. The very different effects of pH and tem- regarding pH and temperature, there might be differences in for
perature on the different severity calculation results, that is log(Ro), example substrate type, dry matter concentration and holding
log(R0o ) and log(R00o ), are shown in Fig. 7a–c. time. This comparison corroborates that several different factors
affect the severity of a certain pretreatment strategy, and that
Evaluation of the severity factors the influence of pretreatment on monosaccharide yields cannot
Comparing pretreatment strategies be reliably predicted with the currently available severity factor
Today most research on pretreatment is accompanied by the models.
calculation of the severity factor without any concern to the many
other factors affecting biomatrix opening than temperature, pH Wheat straw as a model substrate
and holding time. It has been shown previously that the pH and Another angle in the study of pretreatment severity is the design of
temperature do affect the pretreatment outcome drastically. Even a pretreatment strategy for a single substrate. In Europe the most
though producing the same final pH values, two distinct catalysts promising lignocellulosic biomass for second-generation ethanol
might affect the yields differently [28]. Therefore, comparison of production is wheat straw [16]. Hence, the enzymatic hydrolysis

www.elsevier.com/locate/nbt 745
 REVIEW New Biotechnology  Volume 27, Number 6  December 2010
Review

FIGURE 8
Enzymatic ( ) glucose release, ( ) xylose release and ( ) lignin removal as
responses to (a) log(Ro), (b) log(R0o ) and (c) log(R00o ). Data shown were
obtained from the references found in Tables 2 and 3.

yield data obtained in the most recent studies of wheat straw

pretreatment were plotted in response to the three different sever-
ity factor calculation methods (Fig. 10 and Supplementary Figure
3). The three types of severity factor equations did not produce any
correlations with the release of glucose and xylose. By contrast,
statistical tests of the data in direct response to pretreatment
temperature and pH, respectively, show some correlation (but
not a direct linear correlation) to the pretreatment pH, namely,
FIGURE 7 the higher the pH, the higher the release (Fig. 11). Hence, better
Influence of temperature and pH on the proposed severity factor calculations. correlations are seen when plotting the releases as responses to the
(a) log(Ro) is derived from Eqn (1), (b) log(R0o ) from Eqn (4) and (c) log(R00o ) main pretreatment factors, especially the pH, than with the use
from Eqn (5). Holding time is set to 10 min.
of the severity factor equations (Fig. 11). With respect to the
pretreatment temperature, it becomes apparent that to reach a

746 www.elsevier.com/locate/nbt
New Biotechnology  Volume 27, Number 6  December 2010 REVIEW

Review
FIGURE 9
Enzymatic ( ) glucose release and ( ) xylose release as response to (a) log(Ro), (b) log(R0o ) and (c) log(R00o ) for acidic steam explosion (ASE) (a)–(c), alkaline wet
oxidation (AWO) (d)–(f) and lime pretreatment (Lime) (g)–(i).

FIGURE 10
Enzymatic ( ) glucose release and ( ) xylose release as response to (a) log(Ro), (b) log(R0o ) and (c) log(R00o ) for pretreated wheat straw.

www.elsevier.com/locate/nbt 747
 REVIEW New Biotechnology  Volume 27, Number 6  December 2010
Review

FIGURE 11
Enzymatic ( ) glucose release and ( ) xylose release as response to temperature (T, 8C) and pH for pretreated wheat straw. (Open blue square is an outlier not
included in the curve fitting.)

glucose release above 80% of the biomass glucose potential, it is studies encompassing different substrates and pretreatment strate-
necessary to increase the temperature to above 1708C (Fig. 11). gies – revealed that the monosaccharide yields obtained after the
However, at temperatures close to 2008C, both high and low yields subsequent enzymatic hydrolysis cannot be reliably predicted by
might be found. This contradictory effect might be a result of use of a one-dimensional severity factor calculation, even when this
various enzymatic hydrolysis strategies, or simply because of that factor encompasses the pretreatment pH. It seems clear that the
the increased temperature might yield high glucose releases but at general use of the same fitted value (14.75) for the derived activation
the same time induce degradation of parts of the fermentable energy constant on different substrates is an inherent limitation. We
monosaccharides released. By avoiding too acidic pretreatment also suggest the need for employing a standardized, uniform enzy-
conditions, the formation of degradation products might be matic hydrolysis treatment using a defined cellulase blend at fixed
reduced, and this might explain the positive effects on the yields temperature, pH, enzyme dosages, reaction time and substrate
obtained at higher pH (Fig. 11). Hence, when comparing the concentration. Our understanding of the effects of pretreatment
influence of pretreatment factors from various studies on the factors on different types of lignocellulosic biomass can only be
enzymatic degradation of a single type of biomass feedstock, for improved if more specific methodologies and benchmarks are used.
example wheat straw, the pretreatment pH appears to be the major In view of the wide divergence of results currently available there is a
factor influencing the yields. strong need for more valid models describing the quantitative
influences of the multiple factors affecting the impact of pretreat-
Final remarks ment on the enzymatic digestibility of lignocellulosic biomass.
A detailed understanding of the chemical reactions that occur
during pretreatment is crucial for designing better pretreatment Acknowledgements
strategies. However, the multiple factors affecting the biomatrix This project was supported by Novozymes A/S via the Novozymes
opening make the modeling and prediction of the events very Bioprocess Academy at The Technical University of Denmark. We
complex. The pretreatment pH and temperature need to be parti- furthermore thank Henrik Kolind for the assistance with the
cularly considered when visualizing the effect of pretreatment of graphical illustrations.
biomass. Together with treatment duration, both temperature and
pH might be included when comparing the severities of different Appendix A. Supplementary data
pretreatment strategies, but a comparison of the yield responses Supplementary data associated with this article can be found, in
to different pretreatment severity factor models – across multiple the online version, at doi:10.1016/j.nbt.2010.05.003.

References
1 Lange, J. (2007) Lignocellulose conversion: an introduction to chemistry, process 5 Hsieh, Y. (2009) Alkaline pre-treatment of rice hulls for hydrothermal production
and economics. Biofuel. Bioprod. Bior. 1, 39–48 of acetic acid. Chem. Eng. Res. Des. 87, 13–18
2 Liu, S. and Qureshi, N. (2009) How microbes tolerate ethanol and butanol. New 6 Howard, R. (2003) Lignocellulose biotechnology: issues of bioconversion and
Biotechnol. 26, 117–121 enzyme production. Afr. J. Biotechnol. 2, 602–619
3 Atsumi, S. et al. (2008) Metabolic engineering of Escherichia coli for 1-Butanol 7 Mosier, N. et al. (2005) Features of promising technologies for pretreatment of
production. Metab. Eng. 10, 305–311 lignocellulosic biomass. Bioresour. Technol. 96, 673–686
4 Binder, B. and Raines, R.T. (2009) Simple chemical transformation of 8 Bjerre, A.B. et al. (1996) Pretreatment of wheat straw using combined wet
lignocellulosic biomass into furans for fuels and chemicals. J. Am. Chem. Soc. 131, oxidation and alkaline hydrolysis resulting in convertible cellulose and
1979–1985 hemicellulose. Biotechnol. Bioeng. 49, 568–577

748 www.elsevier.com/locate/nbt
New Biotechnology  Volume 27, Number 6  December 2010 REVIEW

9 Holtzapple, M. et al. (1991) The ammonia freeze explosion (AFEX) process. Appl. 37 Thomsen, M.H. et al. (2009) Identification and characterization of fermentation
Biochem. Biotechnol. 28, 59–74 inhibitors formed during hydrothermal treatment and following SSF of wheat
10 Olsson, L. et al. (2005) Bioethanol production from lignocellulosic material. In straw. Appl. Microbiol. Biotechnol. 83, 447–455
Polysaccharides: Structural Diversity and Functional Versatility (2nd edn) (Dumitriu, 38 Kristensen, J.B. et al. (2008) Cell-wall structural changes in wheat straw pretreated
S., ed.), pp. 957–993 for bioethanol production. Biotechnol. Biofuels 1, 5
11 Saha, B.C. and Cotta, M.A. (2009) Comparison of pretreatment strategies for 39 Durot, N. et al. (2003) The unmasking of lignin structures in wheat straw by alkali.
enzymatic saccharification of barley straw to ethanol. New Biotechnol. 10.1016/ Phytochemistry 63, 617–623
j.nbt.2009.10.005 40 Iiyama, K. et al. (1990) Phenolic acid bridges between polysaccharides and lignin in
12 Rosgaard, L. et al. (2007) Comparison of different pretreatment strategies for wheat internodes. Phytochemistry 29, 733–737
enzymatic hydrolysis of wheat and barley straw. Appl. Biochem. Biotechnol 143, 41 Donohoe, B.S. et al. (2008) Visualizing lignin coalescence and migration through
284–296 maize cell walls following thermochemical pretreatment. Biotechnol. Bioeng. 101,
13 Overend, R.P. and Chornet, E. (1987) Fractionation of lignocellulosics by steam- 913–925
aqueous pretreatments. Philos. Trans. R. Soc. Lond. A 321, 523–536 42 Panagiotou, G. and Olsson, L. (2006) Effect of compounds released during
14 Garcı́a-Aparicio, M.P. et al. (2006) Effect of inhibitors released during steam- pretreatment of wheat straw on microbial growth and enzymatic hydrolysis rates.
explosion pretreatment of barley straw on enzymatic hydrolysis. Appl. Biochem. Biotechnol. Bioeng. 96, 250–258
Biotechnol. 129, 278–288 43 Petersen, M.Ø. et al. (2009) Optimization of hydrothermal pretreatment of wheat
15 Hector, R.W. et al. (2009) The Saccharomyces cerevisiae YMR315W gene encodes an straw for production of bioethanol at low water consumption without addition of

Review
NADP(H)-specific oxidoreductase regulated by the transcription factor Stb5p in chemicals. Biomass Bioenerg. 33, 340–834
response to NADPH limitation. New Biotechnol. 26, 171–180 44 Sun, Y. and Cheng, J. (2002) Hydrolysis of lignocellulosic materials for ethanol
16 Kabel, M.A. (2007) Effect of pretreatment severity on xylan solubility and production: a review. Bioresour. Technol. 83, 1.11
enzymatic breakdown of the remaining cellulose from wheat straw. Bioresour. 45 Keller, F.A. et al. (2003) Microbial pretreatment of biomass. Appl. Biochem.
Technol. 98, 2034–2042 Biotechnol. 105, 27–41
17 Fang, T.J. et al. (2009) Enhanced production of xylanases by Aspergillus carneus M34 46 Shi, J. et al. (2009) Effect of microbial pretreatment on enzymatic hydrolysis and
in solid-state fermentation with agricultural waste using statistical approach. New fermentation of cotton stalks for ethanol production. Biomass Bioenerg. 33, 88–96
Biotechnol. 10.1016/j.nbt.2009.09.008 47 Taniguchi, M. et al. (2005) Evaluation of pretreatment with Pleurotus ostreatus for
18 McMillan, J.D. (1994) Pretreatment of lignocellulosic biomass. In Enzymatic enzymatic hydrolysis of rice straw. J. Biosci. Bioeng. 100, 637–643
Conversion of Biomass for Fuel Production, ACS Symposium Series 566 (Himmel, M.E. et 48 Wang, H. et al. (2004) Heterologous expression of lignin peroxidase of
al. eds), pp. 292–324, American Chemical Society Phanerochaete chrysosporium in Pichia methanolica. Biotechnol. Lett. 26, 1569–1573
19 Kumar, P. et al. (2009) Methods for pretreatment of lignocellulosic biomass for 49 Pedersen, M. et al. (2009) Screening for cellulose and hemicellulose degrading
efficient hydrolysis and biofuel production. Ind. Eng. Chem. Res. 48, 3713–3729 enzymes fromthefungal genus Ulocladium. Int. Biodeterior. Biodegradation 63, 484–489
20 Singh, S. et al. (2009) Production of high level of cellulase-poor xylanases by wild 50 Sohail, M. et al. (2009) Cellulase production from Aspergillus niger MS82: effect of
strains of white-rot fungus Coprinellus disseminatus in solid-state fermentation. temperature and pH. New Biotechnol. 25, 437–441
New Biotechnol. 26, 165–170 51 Cuzens, J.C. and Miller, J.R. (1997) Acid hydrolysis of bagasse for ethanol
21 Selig, M.J. et al. (2007) Deposition of lignin droplets produced during dilute acid production. Renew. Energ. 10, 285–290
pretreatment of maize stems retards enzymatic hydrolysis of cellulose. Biotechnol. 52 Yamada, T. et al. (2002) Performance of immobilized Zymomonas mobilis 31821
Prog. 23, 1333–1339 (pZB5) on actual hydrolysates produced by Arkenol technology. Appl. Biochem.
22 Kabel, M.A. et al. (2007) Structural differences of xylans affect their interaction Biotechnol. 98–100, 899–907
with cellulose. Carbohydr. Polym. 69, 94–105 53 Pedersen, M. and Meyer, A.S. (2009) Influence of substrate particle size and wet
23 Singh, S. et al. (2009) Visualization of biomass solubilization and cellulose oxidation on physical surface structures and enzymatic hydrolysis of wheat straw.
regeneration during ionic liquid pretreatment of switchgrass. Biotechnol. Bioeng. Biotechnol. Prog. 25, 399–408
104, 75–86 54 Silverstein, R.A. et al. (2007) A comparison of chemical pretreatment methods for
24 Zhu, S. (2008) Use of ionic liquids for the efficient utilization of lignocellulosic improving saccharification of cotton stalks. Bioresour. Technol. 98, 3000–3011
materials. J. Chem. Technol. Biotechnol. 83, 777–779 55 Brasch, D.J. and Free, K.W. (1965) Prehydrolysis-Kraft pulping of Pinus radiata
25 Carvalheiro, F. et al. (2009) Wheat straw autohydrolysis: process optimization and grown in New Zealand. TAPPI 48, 245–248
products characterization. Appl. Biochem. Biotechnol. 153, 84–93 56 Chen, S.-F. et al. (2007) Pseudo reaction kinetics of organic degradation products in
26 Ramos, L.P. (2003) The chemistry involved in the steam treatment of dilute-acid-catalyzed corn stover pretreatment hydrolysates. Biotechnol. Bioeng. 98,
lignocellulosic materials. Quim. Nova 26, 863–871 1135–1145
27 Sluiter, A. et al. (2006) Determination of Structural Carbohydrates and Lignin in 57 Chum, H.L. et al. (1990) Pretreatment-catalyst effects and the combined severity
Biomass. National Renewable Energy Laboratory, LAP-002, Golden, CO parameter. Appl. Biochem. Biotechnol. 24–25, 1–14
28 Pedersen, M. et al. Monosaccharide yields and lignin removal from wheat straw in 58 Abatzoglou, N. et al. (1992) Phenomenological kinetics of complex systems: the
response to pH and catalyst type during mild thermal pretreatment. Process development of a generalized severity parameter and its application to
Biochem (in press) DOI:10.1016/j.procbio.2010.03.020. lignocellulosics fractionation. Chem. Eng. Sci. 47, 1109–1122
29 Yang, B. and Wyman, C.E. (2004) Effect of xylan and lignin removal by batch and 59 Dien, B. et al. (2006) Enzymatic saccharification of hot-water pretreated corn fiber
flow through pretreatment on the enzymatic digestibility of corn stover cellulose. for production of monosaccharides. Enzyme Microb. Tech. 39, 1137–1144
Biotechnol. Bioeng. 86, 89–95 60 Thomsen, M.H. et al. (2008) Hydrothermal treatment of wheat straw at pilot plant
30 Almeida, J.R.M. et al. (2007) Increased tolerance and conversion of inhibitors in scale using a three-step reactor system aiming at high hemicellulose recovery, high
lignocellulosic hydrolysates by Saccharomyces cerevisiae. J. Chem. Technol. Biot. 82, cellulose digestibility and low lignin hydrolysis. Bioresour. Technol. 99, 4221–4228
340–349 61 Jin, A.X. et al. (2009) Comparative characterization of degraded and non-
31 Suryawati, L. et al. (2009) Effect of hydrothermolysis process conditions on degradative hemicelluloses from barley straw and maize stems: comparison,
pretreated switchgrass composition and ethanol yield by SSF with Kluyveromyces structure, and thermal properties. Carbohydr. Polym. 78, 609–619
marxianus IMB4. Process Biochem. 44, 540–545 62 Adapa, P. et al. (2009) Compaction characteristics of barley, canola, oat and wheat
32 Liu, Z.L. et al. (2004) Adaptive response of yeasts to furfural and 5- straw. Biosyst. Eng. 104, 335–344
hydroxymethylfurfural and new chemical evidence for HMF conversion to 2,5-bis- 63 Persson, T. et al. (2009) Fractionation of wheat and barley straw to access high-
hydroxymethylfuran. J. Ind. Microbiol. Biotechnol. 31, 345–352 molecular-mass hemicelluloses prior to ethanol production. Bioresour. Technol.
33 Saha, B.C. et al. (2005) Dilute acid pretreatment, enzymatic saccharification and 100, 3906–3913
fermentation of wheat straw to ethanol. Process Biochem. 40, 3693–3700 64 Kim, T.H. et al. (2003) Pretreatment of corn stover by aqueous ammonia. Bioresour.
34 Lohmeier-Vogel, E.M. et al. (1998) Intracellular acidification as a mechanism for Technol. 90, 39–47
the inhibition by acid hydrolysis-derived inhibitors of xylose fermentation by 65 Zhang, Y.H.P. et al. (2007) Fractionating recalcitrant lignocellulose at modest
yeasts. J. Ind. Microbiol. Biotechnol. 20, 75–81 reaction conditions. Biotechnol. Bioeng. 97, 214–223
35 Larsen, J. et al. (2008) The IBUS process – lignocellulosic bioethanol close to a 66 Kaar, W.E. and Holtzapple, M.T. (2000) Using lime pretreatment to facilitate the
commercial reality. Chem. Eng. Technol. 31, 765–772 enzymatic hydrolysis of corn stover. Biomass Bioenerg. 18, 189–199
36 Weng, Y.-H. et al. (2009) Separation of acetic acid from xylose by nanofiltration. 67 Zhu, Z. et al. (2009) Comparative study of corn stover pretreated by dilute acid and
Sep. Purif. Technol. 67, 95–102 cellulose solvent-based lignocellulose fractionation: enzymatic hydrolysis,

www.elsevier.com/locate/nbt 749
 REVIEW New Biotechnology  Volume 27, Number 6  December 2010

supramolecular structure, and substrate accessibility. Biotechnol. Bioeng. 103, 74 O’Dwyer, J.P. et al. (2007) Enzymatic hydrolysis of lime-pretreated corn stover and
715–724 investigation of the HCH-1 model: inhibition pattern, degree of inhibition,
68 Kim, S.B. et al. (2000) Step-change variation of acid concentration in a percolation validity of simplified HCH-1 model. Bioresour. Technol. 98, 2969–2977
reactor for hydrolysis of hardwood hemicellulose. Bioresour. Technol. 72, 289–294 75 Eggeman, T. and Elander, R.T. (2005) Process and economic analysis of
69 Miller, S. and Hester, R. (2007) Concentrated acid conversion of pine softwood to pretreatment technologies. Bioresour. Technol. 96, 2019–2025
sugars. Part 1: use of a twin-screw reactor for hydrolysis pretreatment. Chem. Eng. 76 Schmidt, A.S. and Thomsen, A.B. (1998) Optimization of wet oxidation
Commun. 194, 85–102 pretreatment of wheat straw. Bioresour. Technol. 64, 139–152
70 Mabee, W.E. et al. (2006) Updates on softwood-to-ethanol process development. 77 Martin, C. et al. (2007) Wet oxidation as a pretreatment method for enhancing the
Appl. Biochem. Biotechnol. 129–132, 55–70 enzymatic convertibility of sugarcane bagasse. Enzyme Microb. Tech. 40, 426–432
71 Sierra, R. et al. (2009) Short-term lime pretreatment of poplar wood. Biotechnol. 78 Schmidt, A.S. et al. (2002) Comparison of the chemical properties of wheat straw
Prog. 25, 323–332 and beech fibers following alkaline wet oxidation and laccase treatments. J. Wood
72 Rabelo, S.C. et al. (2009) Lime pretreatment of sugarcane bagasse for bioethanol Chem. Technol. 22, 39–53
production. Appl. Biochem. Biotechnol. 153, 139–150 79 Bura, R. et al. (2002) SO2-catalyzed steam explosion of corn fiber for ethanol
73 Maas, R.H.W. et al. (2008) Lactic acid production from lime-treated wheat straw by production. Appl. Biochem. Biotechnol. 98–100, 59–72
Bacillus coagulans: neutralization of acid by fed-batch addition of alkaline 80 Esteghlalian, A. et al. (1997) Modeling and optimization of the dilute-sulfuric-acid
substrate. Appl. Microbiol. Biot. 78, 751–758 pretreatment of corn stover, poplar and switchgrass. Bioresour. Technol. 59, 129–136
Review

750 www.elsevier.com/locate/nbt