Renewable Energy 97 (2016) 252e265

Contents lists available at ScienceDirect

Renewable Energy
jou rnal h o me p a g e: ww w .e l s e v ie r . co m/ l o c a t e / r e n e n e

Pretreatment of cassava stems and peelings by thermohydrolysis to

enhance hydrolysis yield of cellulose in bioethanol production process
Paul Alain Kouteu Nanssou a, Yvette Jiokap Nono a, *, Ce´sar Kapseu b
a
University Institute of Technology (IUT) of the University of Ngaoundere, Department of Chemical Engineering and Environment, P.O. Box: 455,
Ngaoundere, Cameroon
b
National Advanced School of Agro-Industrial Sciences (ENSAI) of the University of Ngaoundere, Department of Process Engineering, P.O. Box: 455,
Ngaoundere, Cameroon

A R T I C L E I N F O
A B S T R A C T
Article history:
Received 5 November 2015 The potential of wastes obtained from the cultivation of Manihot esculenta Crantz as raw material for
Received in revised form bioethanol production was studied. The objective was to determine the optimal conditions of hemicel-
9 May 2016 lulose thermohydrolysis of cassava stems and peelings and evaluate their impact on the enzymatic hy-
Accepted 13 May 2016 drolysis yield of cellulose. An experimental design was conducted to model the influence of factors on
the pentose, reducing sugar and phenolic compound contents. Residues obtained from the optimal pre-
treatment conditions were hydrolysed with cellulase (filter paper activity 40 FPU/g). The hydrolysates
Keywords: from pretreatment and enzymatic hydrolysis were fermented respectively using Rhyzopus spp. and
Cassava stems Sacharomyces cerevisiae. The yield of enzymatic hydrolysis obtained under the optimal conditions were
Cassava peelings respectively 73.1% and 86.6% for stems and peelings resulting in an increase of 39.84% and 55.40%
Thermohydrolysis
respectively as compared to the non-treated substrates. The ethanol concentrations obtained after
Enzymatic hydrolysis
fermentation of enzymatic hydrolysates were 1.3 and 1.2 g/L respectively for the stem and peeling hy-
Rhizopus spp.
Bioethanol drolysates. The pentose and phenolic compound concentrations obtained from the multi-response
optimization were 10.2 g/L; 0.8 g/L and 10.1 g/L; 1.3 g/L respectively for stems and peelings. The hy-
drolysates of stems and peelings under these optimal conditions respectively gave ethanol concentra-
tions of 5.27 g/100 g for cassava stems and 2.6 g/100 g for cassava peelings.
© 2016 Published by Elsevier Ltd.

1. Introduction
In the recent years, the production of value-added compounds
from lignocellulosic wastes has been of great interest for economic
as well as environmental reasons [1]. These wastes constitute an
attractive and renewable raw material for bioethanol production
due to their abundance, availability and low cost [2,3]. In
Cameroon, cassava (Manihot esculenta Crantz), with an annual
production of
2.3 million tons, is amongst the most consumed agricultural
products. This production generates a great amount of residues
which include cassava peelings and stems (444 tons per hectare)
[4], often considered as waste. In some parts of Cameroon, the
stems are used as fuel or animal feeds while the peelings are used
as fertilizers. However, the use of these wastes for bioethanol
production would be of greater interest.
* Corresponding author.
E-mail address: jiokapnonoy@yahoo.fr (Y. Jiokap Nono).
Lignocellulosic biomass is principally composed of cellulose,
hemicellulose and lignin whose contents vary depending on the
species [5]. The bioethanol production process from lignocellulosic
feedstocks is very complex and involves pretreatment, hydrolysis
and fermentation [5,6]. The biomass pretreatment is necessary to
alter the physical and chemical properties, thereby enhancing
enzymatic hydrolysis [7,8]. Many factors related to lignocellulosic
substrate (lignin, hemicellulose, porosity …) or specifically to cel-
lulose characteristics (crystallinity index, degree of
polymerization, accessible specific surface …) are responsible for
low rate of enzyme hydrolysis of cellulose [5,7]. Amongst these
factors, the lignin-hemicellulose matrix surrounding the cellulosic
fraction of the biomass has been reported to be one of the
main factors
influencing the yield of hydrolysis. It acts as a physical barrier
preventing the access of cellulase on the cellulose surface [7,8].
Zhu et al. [9] reported that the hemicellulose hydrolysis increases
the pore size and substrate specific surface, thus facilitating the
access of cellulase on the cellulose structure. Some authors [10,11]
have shown that the removal of acetyl groups on hemicellulose
chains,

http://dx.doi.org/10.1016/j.renene.2016.05.050
0960-1481/© 2016 Published by Elsevier Ltd.
 P.A. Kouteu Nanssou et al. / Renewable Energy 97 (2016) 252e265
25
greatly improves enzymatic hydrolysis yields of the substrate.
Recent studies have shown that xylooligomers obtained from Cassava stems or
hemicellulose hydrolysis are cellulase inhibitors. Their release peelings
during enzymatic hydrolysis could lead to a low yield of
hydrolysis. This provides evidence of the need of hydrolyzing
hemicellulose during pretreatment [2,12].
Cutting
Amongst the pretreatments presented in literature, thermohy- h×L×l= 0,6 cm ×3 cm ×3 cm
drolysis which uses pressure to maintain water in the liquid state
at
◦
high temperatures (160e240 C) results in a good solubilisation of Drying
hemicellulose [7,13,14]. This pretreatment, also called autohy- (45°C, ventilated oven)
drolysis, hydrothermal treatment, hot compressed water or liquid
hot water [14], is very economical since there is no use of chemical Grinding
reagents such as sulfuric acid and ammonia [13] and avoids the
presence of inhibitors during fermentation of the hydrolysates
obtained during the pretreatment stage. Hinman et al. [15] re- Sieving
ported that the use of sugars from the hydrolysis of hemicellulose (Φ ≤ 1 mm)
offers an opportunity to reduce the cost of lignocellulosic ethanol
production in the order of 25%. However, during hydrolysis of Cassava stem or peeling powders
hemicellulose into monosaccharides, there is the simultaneous
cleavage of 1e4 alkyl-aryl bonds of lignin and lignin-hemicellulose
bonds resulting in the release of soluble phenolic compounds Characterization (Lignin, Cellulose, Hemicellulose,
[16,17]. These are inhibitors for the fermentation strains, and Proteins, Lipids)
therefore constitute a barrier to the use of hemicellulosic hydro-
lysates obtained during pretreatment.
In this work, thermohydrolysis is used to pretreat cassava stems
and peelings while determining the optimal pretreatment condi-
tions for the different fractions. From literature, very little work Thermohydrolysis
has been done on thermohydrolytic pretreatment of cassava stems (Modelling)
and peelings. Most of the optimizations done use classical methods
consisting of varying each independent variable with time while
maintaining the other variables fixed. This is costly in terms of
Optimization
time and does not permit a good appreciation of the possible in-
teractions between the factors. For this reason, the central com-
posite design response surface methodology was used to visualize
and determine the effects of time, temperature, solid
Maximizing pentose and Maximizing
concentration and possible interactions between them during
minimizing phenolic compounds pentose
thermohydrolysis of cassava peelings and stems. Thereafter,
optimal operating con- ditions for solid residues without
hemicellulose and secondly, hemicellulose hydrolysates with
maximum pentose content and minimum phenolic compound
content were determined. Enzy- matic hydrolysis of the residues
Enzymatic hydrolysis
ACCELLERASE® 1500
permitted the evaluation of its impact on the yield of cellulose
hydrolysis. Fermentation of hy- drolysates and products from
enzymatic hydrolysis by Rhizopus spp. and Saccharomyces Fermentation Fermentation
cerevisiae respectively were conducted. (Rhizopus spp.) (Saccharomyces cerevisiae

2. Materials and methods Fig. 1. Experimental approach.

Experiments were carried out as presented on Fig. 1.

2.2. Preculture and activation of microorganisms

Fermentations were conducted using Rhizopus spp. for the

2.1. Raw materials
pretreatment hydrolysates or Saccharomyces cerevisiae for the
enzymatic hydrolysis products. The Rhizopus genus was obtained
Sweet cassava stems and peelings were harvested from a
from the Microbiology Laboratory of the National Advanced School
plantation, in the month of November, in the Nom-Kandi locality,
of Agro-Industrial Sciences (ENSAI) of the University of
located 60 km from Ngaoundere in the Adamawa region
Ngaoundere in Cameroon. The Saccharomyces cerevisiae strain was
(Cameroon). The stems were sun dried till about 10% moisture
an instant yeast (“safe-instant”, packaged in white 11 g-composite
content and milled to a 1 mm particle size. The powders obtained
packaging, 59703-Marcq-France) bought from a local supermarket
were conditioned in sealed plastic bags and stored at ambient
◦ in Ngaoundere-Cameroon.
temperature (25 ± 2 C) till use. The moisture content was deter-
◦ The Rhizopus genus was identified following the identification
mined using a Heraeus oven at 105 C according to NREL key used by Chabasse et al. (2002) [32]. After 8 days of culture in
procedure [18]. Van Soest and Robertson [19] method was used to ◦
Sabouraud medium at 25 C, the Rhizopus had a cotton-like texture.
evaluate the cellulose, lignin and hemicellulose content. Ash At the beginning of the rapid and extensive growth, the colonies are
content was determined in a muffle furnace using LAP-NREL
procedure [20].
white and become gray and dark gray during aging. Microscopic
observations showed large non-septate filaments. Stolons, rhi-
zoides and sporocystophores generate from the same node. After
sporocyst rupture, the columelle falls on the sporocystophore giv-
Φ 35
ing an umbrella-shape which confirms ◦
the Rhizopus genus.
Rhizopus spp. was maintained at 30 ± 1 C in an inclined tube in a 1
medium containing, 40 g/l D-glucose; 10 g/l peptone and 20 g/l agar
2
at pH 5.5. After strain identification, culture and aging on Petri dish
◦
(Sabouraud medium at 27 C), the spores were collected and
introduced in a sterile tube. 10 mL of sterile distilled water was
added and the whole vigorously agitated using a vortex apparatus
to produce a spore suspension containing 7× 106 spores per
milliliter.
Saccharomyces cerevisiae was activated on Sabouraud medium
and a suspension at 6 ×107 colonies forming unit (CFU)/mL was
obtained.
The inoculum of the different strains was prepared using 2 ml
of each suspension in a 500 mL Erlenmeyer. The added medium
was composed of (g/L): 50 D-glucose; 5 yeast extracts; 7.5
ammonium sulfate ((NH4)SO4); 3.5 potassium
dihydrogenophopsphate (K2HPO4); 0.75 hydrated magnesium
sulfate (MgSO4 7H2O); 1 hy- drated calcium chloride (CaCl2, 2H20).
The medium was previously
◦
autoclaved at 121 C for 15 min and cooled before adding. The
◦
mixtures were incubated at 35 ± 0.5 C for 30 h. At the end of in-

250 mm
cubation, the content of each Erlenmeyer was aseptically centri-

210 mm
fuged and the collected sediment served as inoculum in the

185 mm
fermentation medium.

187 mm
2.3. Thermohydrolysis

153 mm
Pretreatment was conducted in a stainless steel tube reactor,

158 mm
187 mm long and 33 mm of inner diameter (Fig. 2). Polytetra-
fluoroethylene was used to seal the reactor as it is inert and resists 3
◦
to temperatures above 250 C [21].
The powder was introduced into the reactor at a solid concen-
tration given by the experimental design. After sealing, the reactor
was introduced in an oven (Heraeus) previously set at a desired
temperature. The treatment time was recorded from the moment
where the desired temperature was attained and the time taken to
reach the desired temperature was generally between 20 and
22 min. At the end of the treatment, the reactor was brought to
room temperature in a few minutes by immerging it in a water
bath. The liquid fraction was separated
◦
from the solid fraction using
vacuum filtration and stored at 4 C till analysis. 4

2.3.1. Experimental design

A central composite design with 19 experiments (6 axial and 5 Φ 33
central points) was elaborated to study the effect of independent
variables (temperature, time and solid concentration) on the
different responses (pentose, reducing sugar and phenolic com-
pound contents) and interaction of factors.
Pentose content was used as an indicator for hemicellulose
degradation as the hemicellulose of lignocellulosic biomass other
than wood is principally made of xylan [22].
The choices of factors as well as their levels were determined Fig. 2. Pretreatment reactor. 1, 3 and 4: Stainless steel; 2: Teflon.
according to literature review [23e25]. The domain of variation of
the factors and the experimental matrix of the different trials are
Table 1
presented on Table 1 and on Table 2 (“Experiment” columns) Experimental design: domain of variation of thermohydrolysis factors.

respectively. The independent variables were coded using Factors Factor levels
Equation
—a —1 0 1 þa
(1) in order to facilitate the comparisons between the model
◦
coefficients. Temperature ( C) 135.86 150 180 210 224.14
Time (min) 7.93 15 30 45 52.07
X—
— XXmin
Xn ¼ 2 ma
x
(1) Solid concentration (%) 1.35 3 6.5 10 11.65
X
Table 2
Central composite plan design: factors (real values) and responses of the stem and peeling thermohydrolysis.

Experiment Stems Peelings

Run Factors YThp (g/L) YThsr (g/L) YThcp (mg/L) YEhp (g/L) YEhsr (g/L) YEhcp (mg/L)

X1 X2 X3 Obs Aju Obs Aju Obs Aju Obs Aju Obs Aju Obs Aju

1 180.0 30.0 6.5 1.48 2.12 2.63 3.22 592.80 622.62 6.88 7.20 7.43 7.14 1086.38 1027.84
2 180.0 30.0 6.5 2.19 2.12 6.25 3.22 654.49 622.62 8.48 7.20 5.92 7.14 996.40 1027.84
3 180.0 30.0 6.5 2.36 2.12 1.73 3.22 652.19 622.62 6.96 7.20 7.21 7.14 1052.96 1027.84
4 180.0 30.0 6.5 1.43 2.12 1.86 3.22 595.37 622.62 6.89 7.20 8.39 7.14 1083.80 1027.84
5 180.0 30.0 6.5 2.69 2.12 3.84 3.22 664.78 622.62 7.31 7.20 6.27 7.14 932.13 1027.84
6 150.0 15.0 3.0 0.30 0.34 1.25 2.32 351.16 306.20 1.03 0.13 1.73 2.45 402.57 423.29
7 210.0 15.0 3.0 0.93 1.44 1.45 1.51 664.78 700.82 4.49 4.37 7.40 7.46 788.17 723.00
8 150.0 45.0 3.0 1.54 1.95 4.22 3.02 567.09 515.71 3.69 3.07 1.74 1.06 559.38 488.24
9 210.0 45.0 3.0 6.99 7.61 10.09 10.25 934.70 893.64 7.52 7.13 17.41 17.68 1407.71 1345.80
10 150.0 15.0 10.0 0.33 0.02 1.34 1.05 474.55 485.47 4.97 5.04 3.17 3.22 693.06 746.91
11 210.0 15.0 10.0 4.31 4.20 3.84 4.90 865.30 886.53 9.04 9.33 13.44 14.44 1358.87 1421.94
12 150.0 45.0 10.0 2.38 2.18 2.00 1.81 687.92 621.73 7.99 7.78 6.39 6.65 744.48 801.58
13 210.0 45.0 10.0 10.63 10.90 14.91 13.71 991.28 1006.1 11.31 11.89 29.90 29.49 2063.24 2034.46
14 135.9 30.0 6.5 1.48 1.65 3.71 4.07 531.11 621.03 2.97 3.96 3.59 3.49 657.07 612.41
15 224.1 30.0 6.5 9.61 8.88 12.35 12.23 1228.28 1194.06 10.49 10.11 24.46 23.98 1680.21 1739.77
16 180.0 7.93 6.5 0.79 0.83 1.61 0.25 402.57 373.65 4.15 4.46 4.22 3.11 783.03 730.26
17 180.0 52.0 6.5 7.55 6.95 5.64 7.24 531.11 615.73 8.21 8.50 12.63 13.15 1160.93 1228.60
18 180.0 30.0 1.35 0.21 —0.72 0.57 0.44 420.31 476.08 0.89 2.13 1.901 1.79 374.29 491.44
19 180.0 30.0 11.65 1.09 1.46 1.67 2.04 690.75 690.67 9.89 9.25 11.53 11.05 1338.30 1236.06

X1: Temperature; X2: Time; X3: Solid Concentration; Obs: Experimental Response; Aju: Theoretical Response.
YThp, YThsr and YThcp are respectively the pentose, the reducing sugar and the phenolic compound contents in the cassava stem hydrolysates.
YEhp, YEhsr and YEhcp are respectively the pentose, the reducing sugar and the phenolic compound contents in the cassava peeling hydrolysates.

X is the coded value of temperature (X ), of time (X ) and of

P
n 1 2
solid concentration (X3). n jY obs —
Y prej

A second order polynomial model (Equation (2)) with interac- i ¼1

AAD ¼ Yobs
tion was used to adjust the experimental data. n (5)

y¼ a0 X X X ij
i i i i i j Where Yobs is the observed experimental response, Ypre the pre-
þ ax þ a x2 þ a xx þ ε (2) dicted response and n the total number of experiments.
i i i After validation of the models, analysis of variance (ANOVA) was
Where y is the pentose or reducing sugar or phenolic compound used to identify the factors or the interactions which significantly
concentrations; xi and xj are independent variables; a0 is the influence the responses. A factor is significant at 95% interval if its
equation constant; ai is the linear coefficient; aii is the quadratic p-value is less than 0.05.
coefficient; aij is the interaction coefficient and ε is the error. Contour plots were obtained with the aid of Sigmaplot Version
Several statistical parameters were used to validate the models 12.0 and STATGRAPHICS Centurion XVI software used for
as shown on Table 3. These include the determination coefficient response optimizations. The optimizations consisted of
(R2), the adjusted R2, the bias factor (Bf), the average absolute de- determining the combination of experimental factors which
viation (AAD) and the accuracy factor (A f). The R2 brings out in- simultaneously opti- mize many responses or a response. This is
formations on the variability of the responses. The adjusted R 2 is a done by maximizing a desirability function. The first optimization
measure of variation around the mean explained by the model consisted of finding out conditions leading to maximum pentose
[26]. Bf and Af indicate the precision of the model and the AAD the and minimum phenolic compounds solubilisation for the
de- viation of the model with respect to real values. They were hydrolysate intended for fermentation. The second was to
calcu- lated as follows: evaluate the enzymatic di- gestibility of the residue obtained
X under optimal pretreatment conditions that led to maximum
1 Ypre
Bf ¼ 10B; B¼ log (3) pentose solubilisation and there- fore maximum hemicellulose
removal.
n Yobs
2.3.2. Characterization of the liquid fractions
X For the different hydrolysates, the determination of reducing
A 1 Ypre
.
n
. sugar content was done using the 2,5-dinitrosalicyclic acid method
Af ¼ 10 ; A¼ log (4)
. obs .
Y [27], while the pentose and phenolic compound contents were
i¼1
determined as described by Hashimoto et al. [28] and Akpinar et
al.
[29] respectively.

Table 3
Standard values and acceptable range of model validation indicators.

Model validation indicators Standard values Acceptable range References

Adjusted R2 1 >80% Joglekar et May [39]

AAD 0 [0e0.3] Bas¸ and Boyac [41]
Bias factor 1 [0.75e1.25] Dalgaard and Jorgensen [40]
Accuracy factor 1 [0.75e1.25] Dalgaard and Jorgensen [40]
2.4. Enzymatic hydrolysis
Enzymatic hydrolysis was performed using commercial
solution of Trichoderma reesei cellulase (ACCELLERASE® 1500)
of activity
97.4 FPU/mL. The enzyme concentration used was 40 FPU/(g d-
◦ Proteins
b). The hydrolysis tests were done at 55 C for 72 h using 1% dry
matter
in 200 mL Erlenmeyer flasks. The preparation of the 1% dry
matter suspension was done by measuring moisture content of the
pre- treated residues. Then, the wet basis equivalent of 1 g dry
matter was weighed and the enzyme added. The mixture was then
completed to 200 mL with 50 mM citrate buffer (pH 4.5)
containing 1% sodium azide (2%) [30]. In parallel, three types of
controls were set: the first contained 1 g of non-pretreated matter
with citrate buffer and the enzyme. The second and third
respectively con- tained 1 g of pretreated matter and 1 g of non-
pretreated matter in citrate buffer without enzymes. The
experiments were done in triplicate.
The enzymatic hydrolysis was followed using 1 mL of the hy-
drolysate collected every 6 h interval. The sample was centrifuged
at 4000 rpm for 15 min to remove the unhydrolyzed residue and
the hydrolysate was used for reducing sugar analysis by 2,5 dini-
trosalicylic acid method [26]. The enzymatic hydrolysis yield was
expressed in terms of percentage of enzymatically converted cel-
lulose (CEC) into glucose [31]:
C$V$a
CEC $100 (6)
¼
m$cel$1:1
With C: reducing sugar concentration in glucose equivalent (g/
L); V: hydrolysate volume (L); a: dilution factor; m: biomass dry
matter (pretreated or not) before hydrolysis (g); cel: cellulose
percentage and 1.1: conversion factor of cellulose into glucose.
2.5. Fermentation
For the fermentation process, 1 g of Rhizopus spp. was used to
ferment 100 mL of non-detoxified hydrolysates obtained during
pretreatment. Also, 1 g of Saccharomyces cerevisiae was used to
ferment 100 mL of hydrolysates obtained from enzymatic
hydrolysis.
The ethanol produced during fermentation was measured by
colorimetry using potassium dichromate (K2Cr2O7). The ethanol
yield (Y) was calculated by dividing the total quantity of ethanol
(E) produced in the fermentation broth by the initial mass of the
reducing sugars (IRS) in the hydrolysates (Equation (7)). This can
equally be expressed in terms of a ratio of the ethanol effectively
produced (Eeff) to the theoretical ethanol value (Eth) (Equation (8)).
Yðg=gÞ¼
(7)
E IR
S [41] and corn cobs [42], cassava stems and peelings are less rich in
Eeff Eeff
Y (8)
¼ Eth ¼ 0:511$G
Where E and IRS are expressed in grams; 0.511 represents the
ethanol theoretical yield from glucose [33].
3. Results and discussion
3.1. Characterization
The chemical composition of cassava stems and peelings are
presented on Table 4. The proportions of carbohydrate (cellulose
and hemicellulose) and lignin Klason obtained were respectively
50.0% and 30.6% for the stems and 42.07% and 16.9% for the
Table 4
Chemical composition of cassava stems and peelings.
Component Peelings Stems
Dry matter (%) Dry matter (%)
Ash 11.38 ± 0.11 a
7.34 ± 0.27b
3.70 ± 0.15a 1.47 ± 0.4b
Lipids 1.70 ± 0.10a 0.70 ± 0.01b
Cellulose 9.71 ± 1.04a 28.86 ± 1.78b
Hemicellulose 32.36 ± 1.08a 21.12 ± 2.28b
Klason lignin 16.89 ± 0.76a 30.62 ± 0.95b
Others 24.26 9.89
On the same line. data with the same superscript letter are not different at the 5%
significance level.
peelings. Hemicellulose was found to be the main constituent of
the peelings against lignin for the stems. In additions to these,
cassava stems and peelings also contained ash, proteins, lipids and
other components, which were respectively: 7.3%; 1.47%; 0.7% and
9.89%,
and 11.38%; 3.7%; 1.7% and 24.26%. Literature also mentions the
presence of minor components such as uronic acid, starch or ex-
tractives [5] which were not evaluated in this work.
The hemicellulose and lignin content of the stems are in
accordance with those of Han et al. [34] who found respective
values of 30% and 24.3%. These differ from those of Chen et al.
[35] who found cellulose, hemicellulose, lignin and ash content of
35.86%; 19.2%, 26.10% and 4.97% respectively; and Pooja and Pad-
maja [36] who obtained respective values of 22.8%; 28.80%; 22.10%
and 3.68%.
For the cassava peelings, the cellulose content is similar to the
value presented by Lounglawan et al. [37] (11.58 ± 1.38%). However,
the lignin content (7.15 ± 0.24%) and the hemicellulose content
(52.24 ± 1.05%) are different. Besides, the values obtained differ
from those of Babayemi et al. [38]: lignin 11%; cellulose 14% and
hemicellulose 27%; and Pooja and Padmaja [36]: lignin 10%; cellu-
lose 14% and hemicellulose 24%. Differences obtained between our
results and the others probably result from the differences in cli-
matic conditions, soil fertility of the study area or the harvesting
period, soil types or variety of species used. Adler et al. [39] re-
ported an increase in polysaccharide content during winter,
inducing a higher yield in the harvested plant during spring than
that harvested during autumn. Moreover, Dien et al. [40] have
shown that the polysaccharides (cellulose and hemicellulose) and
lignin content were low for unmatured lignocellulosic samples
with respect to the more matured samples. The cassava used in
our work was harvested in the month of November while
Babayemi et al. [38] harvested theirs in the month of March and
August in Nigeria. Concerning Han et al. [34], the cassava stems in
their study
were collected during summer in Papouasie New Guinea.
Compared to other agricultural residues such as wheat straw
cellulose. The cellulose content is similar to that of Alfalfa stem
[43] and Coastal Bermuda grass [44]. The lignin content of
peelings and stems are close to that of sweet sorghum bagasse
[45] and euca- lyptus [46]. The hemicellulose content of stems and
peelings are close to Hybrid poplar [47] and corn cobs [42].
3.2. Effect of time, temperature and solid concentrations on the
different responses
At the end of each experiment, hydrolysates were collected and
analyzed and the results are as shown on Table 2. Exploitation of
these results has led to the establishment of six models.
3.2.1. Modelling and validation of the models
3.2.1.1. Thermohydrolysis models for the cassava stems.
YThp ¼ 2; 1245 þ 2; 4567 × X1 þ 2; 0783 × X2 þ 0; 7444 × X3
þ 1; 4513 × X21 þ 1; 1369 × X1 × X2 þ 0; 7676 × X1 ×
X3 þ 0; 8148 × X2 × X2 þ 0; 1349 × X2 × X3 — 0; 8109
× X3
YThsr ¼ 3; 2214 þ 2; 7722 × X1 þ 2; 3750 × X2 þ 0; 5435 × X3
þ 2; 2761 × X21 þ 2; 0104 × X1 × X2 þ 1; 1670 × X1 ×
X3
2 2
þ 0; 2404 × X2 þ 0; 0164 × X2 × X3 — 0; 9158 ×
X3
YThcp ¼ 622; 6180 þ 194; 7490 × X1 þ 82; 2718 × X
2
þ 72; 9329 × X 3 þ 131; 6430 × X1 — 4; 1739 × X1 ×
X2
þ 1; 6104 × X1 × X3 — 59; 1050 × X22 — 18; 3126 × X2
× X3 — 18; 1288 × X23
(11)
The models obtained were validated with the aid of validation
indicators. The values obtained from the validation model tools are
presented in Table 5. All these values are within the range recom-
mended by different works [48e50]. Joglekar and May [48]
consider that a model can be validated if the model explains at
least 80% of the response variability (adjusted R2). Dalgaard and
2
Jorgesen [49] consider a model to be valid if the truth and bias
factors are between 0.75 and 1.25. Bas and Boyac [50] judge a model
valid if the absolute average deviation (AAD) is between 0 and 0.3.
(9) From this analysis, the models explained 83.55e98.45% of the
response variations with a deviation percentage from the reality of
0.3e4.1%.
3.2.2. Effects of factors on the pentose concentration of the
(10) hydrolysates
The significance of the different effects was done by comparing
the quadratic mean of each effect with respect to experimental
error estimation. The effects with a probability less than 0.05 were
2
considered significant at the 95% confidence level (Table 6).
For all the pentose models obtained, all the direct effects
(temperature, time and solid concentration) are significant at 95%
confidence level and contribute to an increase of the extracted
pentose (Fig. 3). These contributions were respectively quantified

YThp, YThsr, YThcp, X1, X2, X3 are respectively the pentose, reducing as 23.63%, 19.99% and 7.16% for direct effect of the temperature,
sugar, phenolic compound contents in the cassava stem hydroly- time and solid concentration for the stems models. From the
sates, temperature, time and solid concentration (coded values). contribution of the effects in the peelings model, direct effects of
the temperature, time and solid concentration were respectively
29.41%, 19.35% and 34.06%. These results are in accordance with
3.2.1.2. Thermohydrolysis models for the cassava peelings.
those of Laser et al. [51] who observed a positive effect of time and
YEhp ¼ 7; 2031 þ 2; 0886 × X1 þ 1; 3741 × X2 þ 2; 4184 × X3 temperature on the solubilisation of hemicellulosic fraction of
sugarcane bagasse (34e88%) for an increase of the process time
◦
— 0; 0781 × X21 — 0; 0456 × X1 × X2 þ 0; 0124 × X1 × from 2 to 10 min at 200 C and for an increase of the temperature
◦
X3 from 200 to 220 C at 2 min. Nitsos et al. [52] also observed an
2 2
— 0; 3331 × X2 — 0; 0498 × X2 × X3 — 0; 6993 × increase in hemicellulose hydrolysis and xylose content of the hy-
◦
X3 (12) drolysates when the temperature was raised from 150 to 220 C
during the pretreatment of beech wood.
YEhsr ¼ 7; 1422 þ 6; 9607 × X1 þ 3; 4124 × X2 þ 3; 1461 × X3 To understand these different direct effects, it should be noted
that the chemical reactions can only take place when the reactive
þ 3; 0444 × X21 þ 2; 9042 × X1 × X2 þ 1; 5547 × X1 ×
molecules get in contact, evaluated by the collision frequency be-
X3 þ 0; 4550 ×2 X þ 1; 2072 × X2 × X3 — 0; 3337 ×2 X tween particles which has an impact on the reaction speed con-
2 3
stant. As the collision frequency increases, the reaction kinetic
(13) considered also increases. Increase in temperature on one hand
increases the kinetic energy of molecules and on the other hand,
YEhcp ¼ 1027; 8400 þ 383; 1470 × X1 þ 169; 3660 × the modification of physico-chemical properties of water such as a
decrease in density and viscosity. This facilitates the diffusion of
X2
2
þ 253; 0690 × X3 þ 68; 4925 × X1 þ 139; 4600 × X1 AAD 0.041 0.023 0.003 0.006 0.004 0.007
Bias Factor 1.056 1.034 1.002 0.998 0.993 0.988
× X2 þ 93; 8303 × X1 × X3 — 22; 3678 × X22 — 2; 5707 Accuracy factor 1.056 1.034 1.002 1.009 1.007 1.011

× X2 × X3 — 75; 8151 × X3 2

(14)
YEhp, YEhsr, YEhcp, X1, X2, X3 are respectively the pentose, reducing
sugar, phenolic compound contents in the cassava peeling hydro-
lysates, temperature, time and solid concentration (coded values).

Table 5
Model validation indicators of the different thermohydrolysis pretreatment models
for cassava stems and peelings.

Validation Indicators Stems Peelings

YThp YThsr YThcp YEhp YEhsr YEhcp

R2 (%) 97.72 91.77 95.51 95.68 99.23 97.65

Adjusted R2 (%) 95.44 83.55 91.01 91.35 98.45 95.30
water into the material and increases the ionic product of water
leading to an auto-ionization of water, hence liberating
hydronium ions in solution. The biomass then undergoes
hydrolytic reactions in the presence of hydronium ions
generated by auto-ionization of water which plays the role of a
catalyst. The ether heterocyclic bonds of hemicelluloses are the
most susceptible bonds in this type of reaction resulting
simultaneously to the generation of soluble oligosaccharides and
to the liberation of acetic acid, by saponifi- cation reaction of the
acetyl groups present in the hemicellulosic fraction [22,25].
The direct effect of solid concentration can be explained by
the establishment of the concentration gradient which facilitates
the diffusion in the reverse direction of the concentration
gradient. This direct effect cannot be easily dissociated from the
temperature- solid concentration (X1X3) interaction because the
movement of molecules, amplified by the temperature rise, is
done in the reverse direction of the concentration gradient
imposed by the direct effect of the solid concentrations (Fig. 3).
This interaction is significant (p- value of 0.0115) in the stems
pentose model as well as for the quadratic effects of temperature,
time and solid concentration, and
Table 6
Cassava stem and peeling thermohydrolysis models: significance of the effects. X1:Temperature; X2:Time; X3: Solid Concentration.

Effects Stems Peelings

YThp YThsr YThcp YEhp YEhsr YEhcp

Coefficient P-value Coefficient P-value Coefficient P-value Coefficient P-value Coefficient P-value Coefficient P-value
X1 2.4567 0.0000 2.7722 0.0002 194.7490 0.0000 2.0886 0.0000 6.9607 0.0000 383.1470 0.0000
X2 2.0783 0.0000 2.3750 0.0006 82.2718 0.0017 1.3741 0.0004 3.4124 0.0000 169.3660 0.0001
X3 0.7444 0.0042 0.5435 0.2729 72.9329 0.0035 2.4184 0.0000 3.1461 0.0000 253.0690 0.0000
X1 × X1 1.4513 0.0001 2.2761 0.0021 131.6430 0.0002 —0.0781 0.7924 3.0444 0.0000 68.4925 0.0520
X1 × X2 1.1369 0.0011 2.0104 0.0069 —4.1739 0.8607 —0.0456 0.8868 2.9042 0.0000 139.4600 0.0023
X1 × X3 0.7676 0.0115 1.1670 0.0742 1.6104 0.9460 0.0124 0.9690 1.5545 0.0013 93.8303 0.0196
X2 × X2 0.8148 0.0055 0.2404 0.6631 —59.1050 0.0219 —0.3331 0.2774 0.4550 0.1784 —22.3678 0.4834
X2 × X3 0.1349 0.5922 0.0164 0.9780 —18.3126 0.4487 —0.0498 0.8767 1.2072 0.0059 —2.5707 0.9398
X3 × X3 —0.8109 0.0056 —0.9158 0.1204 —18.1288 0.4181 —0.6993 0.0382 —0.3337 0.3123 —75.8151 0.0351

The coefficients having P-value more than 0.05 are considered non-significant.
experiments 2, 5, 8, 10 and 12 of the experimental design of peel- ings
and 11, 12 and 17 of the experimental design of stems. Taking
time-temperature interaction whose contributions are
respectively 13.96%; 10.93%; 7.38%; 7.83% and 7.80% (Fig. 3).
All these different effects can be explained by the generation of
hydronium ions by acetic acid auto-ionization at more advanced
reaction stages, which catalyze the degradation of oligomers into
monomers. Liu et al. [53] proposed a model to explain the hydra-
tion of acetyl groups leading to the acidification of the medium,
inducing the formation of hydronium ion in the medium during
thermohydrolysis. Heitz et al. [54] showed that hydronium ion
formation from acetic acid is more important than that obtained
from water; although the role of auto-ionization of water is limited
to the initial reaction step, the difference in hemicellulose acety-
lation rate of stems and peelings could explain their different hy-
drothermal behaviors. The quadratic effect of time being positive
implies that an important increase in time contributes to an in-
crease in pentoses in solution. This result is supported by
Mehlberg and Tsao [55] who mentioned the existence of two
hemicellulose fractions in the biomass, one being hydrolysed
easily whereas the other requiring a very long time.

3.2.3. Effects of pretreatment parameters on hydrolysates reducing

sugar concentration
The direct effects of temperature and time, temperature-time
interactions and the quadratic effect of temperature significantly
and positively influence the extraction of reducing sugars from the
stems at 95% confidence level (Table 6); the contributions are
respectively 22.5%, 19.28%, 18.47% and 16.32%. Here also like in the
case of pentose, the water auto-ionization and acetic acid reactions
that occur are responsible for the reducing sugar liberation.
Eight effects are significantly and positively implicated in the
reducing sugars extraction from the cassava peelings (Fig. 4).
These different effects respectively contribute to 30.23%; 14.82%;
13.66%;
13.22%; 12.61%; 6.75% and 5.24% for direct effects of temperature,
time and solid concentration, quadratic effect of temperature,
time- temperature interaction and temperature-solid
concentration in-
teractions. These results are in accordance with those of Pin
´ kowska
et al. [56] who observed an increase in reducing sugar content
from 14.1% to 39.6% for an increase of temperature from 180 ◦C to
220 ◦C during beech wood thermohydrolysis. Whatever the
substrate studied, the effects of temperature, time and their
interactions (Fig. 4) appears to be significant. This observation is
similar to that reported
by Ruiz et al. [14], where the relation between the temperature
and time significantly influences the physico-chemical
characteristics of lignocellulosic biomass during thermohydrolysis
treatment.
Results show that most of pentose concentrations obtained are
slightly lower than those of reducing sugars with an exception in
into account that pentose is a reducing sugar, this slight
difference predicts on one hand, that thermohydrolysis leads
to a selective solubilisation of hemicelluloses, and on the other
hand that pentose is present in solution in the form of non-
reducing oligomers. These results are confirmed by Garrote et
al. [57] who showed that hy- drolysis of hemicelluloses of corn
cobs into monomers passes through an intermediate stage
where oligomers are formed and then are hydrolyzed into
monomer units such as glucose, xylose and arabinose.

3.2.4. Effects of pretreatment parameters on the

hydrolysates phenolic compounds concentration
From the analysis of variance of the stems model, it is
observed that the direct effects of temperature, time, solid
concentration and quadratic effect of temperature have a
significant and positive in- fluence on the extraction of phenolic
compounds (Table 6) from the cassava stems in the order
33.47%; 14.11%; 12.5% and 22.58% respectively. However,
the quadratic effect of time had a significant and negative
influence on the extraction in the order of 10.14%.
For the peelings model, direct effects of temperature, time
and substrate concentration, time/temperature and
temperature/solid concentration interactions have a
significant and positive influence on the extraction of phenolic
compounds with respective values of 31.71%; 14.02%; 20.95%
and 11.54% (Fig. 5). These results are in accordance with the
works of Akpinar et al. [29], who observed a positive influence
of time, temperature and quadratic effect of temperature on
the extraction of phenolic compounds from wheat straw. From
these analyses, it could be deduced that for any of the
substrates, temperature is the main factor that causes the
liberation of phenolic compounds in the medium (Fig. 5).
Lignin is a polymer made up of aromatic phenol units, inter-
linked by ether and C-C bonds [58]. To understand
time/tempera- ture and temperature/solid concentration
interactions many authors have revealed that ether bonds are
not stable in hydro- thermal medium [59,60] and that a rise in
temperature leads to a decrease in lignin aliphatic hydroxyl
groups. This is probably due to two simultaneous reactions:
lateral chains acid catalyzed elimina- tion reaction and
formation of hydroxyl phenolic compounds from the cleavage
of aryl ether bonds. Furthermore, Lora and Wayman [61],
showed that b-aryl ether bonds are degraded in acid medium
from acetic acid released by hemicellulose, which plays the role
of a catalyst. For a long time process, this release from
hemicellulose leads to lignin depolymerisation and/or
simultaneous lignin repo- lymerization by C-C new bonds
formation. This explains the time quadratic, time/temperature
and temperature/solid concentration interactions effects. For a
high quantity of substrate, the acidifica- tion medium will be
high, hence the increase of catalytic hydrolysis of b-aryl bonds
and phenolic compounds liberation.
Fig. 3. Hydrolysates of cassava stem (left) and peelings (right) from thermohydrolysis pretreatment: pentose response surface curves as a function of (a) Time and Temperature with
◦
solid concentration set at 6.5%. (b) Time and Solid concentration with temperature set at 180 C. (c) Temperature and Solid concentration with the time set at 30 min.
Fig. 4. Hydrolysates of cassava stem (left) and peelings (right) from thermohydrolysis pretreatment: reducing sugar response surface curves as a function of (a) Time and Tem-
◦
perature with solid concentration set at 6.5% (b) Time and Solid concentration with temperature set at 180 C. (c) Temperature and Solid concentration with the time set at 30 min.
Fig. 5. Hydrolysates of cassava stem (left) and peelings (right) from thermohydrolysis pretreatment: phenolic compound response surface curves as a function of (a) Time and Tem-
◦
perature with solid concentration set at 6.5%. (b) Temperature and Solid concentration with the time set at 30 min. (c) Time and Solid concentration with temperature set at 180 C.

3.3. Enzymatique hydrolysis

from the optimization of pretreatment conditions leading to
maximum pentose concentration in the liquid phase. These
3.3.1. Characterization of residues from optimal conditions of
optimal conditions are presented on Table 7. The theoretical
pentoses extraction
optimum ob-
The different residues used for enzymatic hydrolysis resulted ◦
tained for the peelings model is 224.1 C, 50.24 min and 11.48% for
X1, X2 and X3 respectively, for the corresponding optimal pentose in
Table 7
Operating conditions for maximum pentose liberation in solution.
◦
Substrate Temperature ( C) Time (min) Solid concentration (%) Pentoses (g/L) Reducing sugar (g/L)

Predicted Experimental Predicted Experimental

Peelings 224.1 50.3 11.48 13.22 14.01 31.29 32.37

Stems 223.9 52.1 10.7 17.40 18.20 22.82 23.76

solution of 13.22 mg/mL. For the stems model, the optimal point
6
◦
YThp (X1, X2, X3) is 223.9 C for X1, 52.1 min for X 2 and 10.7% for X3.
The optimal pentose value was 17.40 mg/mL. A verification of
5
optimal values led to respective pentose concentrations of
14.01 ± 0.80 mg/ml and 18.20 ± 0.67 mg/ml for peelings and stems.

Reducing sugar (g/L)

These obtained values are not significantly different from the
theoretical values obtained with the software, at 95% confidence
level. These optima were used to produce residues for enzymatic
hydrolysis.
The characterization of residues (cellulose, hemicellulose and
lignin contents) was carried out prior to enzymatic hydrolysis and
the results are presented on Table 8. These results show that more
than half of the different residues are composed of cellulose;
meanwhile, the poor hemicellulose residues content confirms the
efficiency of the hemicellulose hydrolysis. Nitsos et al. [52] also
obtained residues without hemicellulose from wood thermohy-
◦ 0
drolysis treatment at 220 C for 15 min. After a thermohydrolysis
◦ Time (h)
pretreatment at 220 C for 50 min, Li et al. [64] obtained a bamboo
residue made of 61.76% cellulose, 3.43% hemicellulose and 34.76%
lignin contents meanwhile non-pretreated raw material had
43.44% cellulose, 30.10% hemicellulose and 27.52% lignin contents,
and concluded that thermohydrolytic pretreatment can be used
for hemicellulose hydrolysis in bamboo. The lignin content of the
different residues showed an increase with respect to that of the
raw materials by 13.61% for peelings and 19.48% for stems. This
increase in lignin content of pretreated residues can be due to a
repolymerization of polysaccharides or products from the degra-
dation of polysaccharides on lignin chains as reported by
Jakobsons et al. [62] during steam explosion treatment of birch
wood. Other authors [63], concluded that some organic and
inorganic matters present in biomass can be condensed to lignin
during 72% sulfuric acid treatment (used for lignin content
determination) resulting in an overestimation of the lignin
content.
Low ash content was observed. Protein contents were not
quantified; however, taking into account its low content in the
initial matter, its absence in the residues can be envisaged.
3.3.2. Enzymatic hydrolysis kinetics of pretreated substrates
Fig. 6 presents the results of 72 h kinetic. Control residues
without addition of cellulase confirmed the complete removal of
reducing sugars, through washing, from the residues after
pretreatment.
From the curve, no significant increase of reducing sugar con-
tent was observed after 42 h for the pretreated residues while for
the non pretreated residues, no significant increase was observed
from the 60th hour. However, the yields of hydrolysis obtained
4
3
2
1
0
10 20 30 40 50 60 70
Pretreated stems Pretreated peelings Non pretreated stems Non pretreated peelings
Fig. 6. Enzymatic hydrolysis kinetics of the residues obtained with the optimal pre-
treatment conditions compared to those obtained with the non-pretreated raw
material.
didn’t reach 100%; the values are 86.65% and 71.46% for peelings
and stems respectively. This can be attributed to either the inacti-
vation and/or enzyme inhibition by cellobiose or glucose or com-
plete hydrolysis of cellulose. Many studies [65e67] report the
inhibition of cellulase by glucose and cellobiose. According to
Gruno et al. [68] Trichoderma reessei cellulase inhibition is strictly
dependent on the nature of the substrate; this explains the
behavioral difference observed with the two substrates. This inhi-
bition could be competitive, non-competitive or mix type
according to the authors [68e70]. On the contrary, hydrolysis yield
of non- pretreated peelings and stems are 46.81% and 16.06%,
which gives an increase of 39.84% and 55.40% respectively.
Nitsos et al.
[52] obtained similar results during wood thermohydrolysis, with
an increase of cellulose digestibility of 63%. Hydrolysis yield of
pretreated stems is close to those of Han et al. [34] who obtained a
hydrolysis yield of 70%. This increase of cellulose hydrolysis in stem
and peeling residues can be attributed to the increase of specific
surface and porosity due to the hydrolysis of hemicellulose as re-
ported by Nitsos et al. [52].
3.4. Fermentation of hydrolysates
The hydrolysates obtained from enzymatic hydrolysis and from
the pretreatment were autoclaved then fermented using

Table 8
Cellulose, hemicellulose, lignin and ash contents of the residues obtained from the optimal pretreatment conditions, compared to those of non-pretreated substrates.

◦
Constituents (% ) Non pretreated stems Pretreated stems Non pretreated peelings Pretreated peelings
Cellulose 28.86 ± 1.78a 45.9 ± 0.4b 9.71 ± 1.64c 57.7 ± 1.9d
Hemicellulose 21.12 ± 2.28a 1.9 ± 0.2b 32.36 ± 1.08c 1.3 ± 0.1d
Lignin 30.62 ± 0.95a 50.1 ± 1.9b 16.89 ± 0.76c 30.5 ± 1.2d
Ash 7.34 ± 0.27a 2.1 ± 0.4b 11.38 ± 0.11c 10.5 ± 0.9c

On the same line, data with the same superscript letter are not different at the 5% significance level.
Table 9
Thermohydrolysis multi-response optimization results: combination of factors and optimal values of pentose, phenolic compound and reducing sugar contents.

◦
Component Temperature ( C) Time (min) Solid concentration (%) Y Optimal value

Predicted Experimenal

Stems
Pentose 198.57 52.07 6.97 10.24 g/L 10.35 g/L
Phenolic compounds 789.01 mg/L 776.25 mg/L
Reducing sugar 11.81 g/L 12.03 g/L
Peelings
Pentose 175.59 52.07 11.46 10.14 g/L 10.50 g/L
Phenolic compounds 1325.9 mg/L 1299.8 mg/L
Reducing sugar 17.74 g/L 18.09 g/L

Table 10
Ethanol concentrations and yields of the different hydrolysates fermented by Rhyzopus spp. or Saccharomyces cerevisiae.

Treatment Substrate Ethanol concentration (g/L) Ethanol yield Gay Lussac (%) Ethanol yield (g/g reducing sugar)

Enzymatic hydrolysis Stems 1.24 ± 0.19 67.40 0.34

Peelings 1.30 ± 0.03 49.25 0.24
Thermohydrolysis Stems 6.15 ± 0.07 66.7 0.34
Peelings 5.03 ± 0.12 76.84 0.40

Saccharomyces cerevisiae and Rhizopus ssp. respectively.

To avoid the strain inhibition by phenolic compounds present
in the hydrolysate, a multi-response optimization was used to look
for the combination factors that permit maximum pentoses and
reducing sugar but minimum phenolic compounds in solution. The
different combination factors as shown in Table 9 produced non-
toxic fermentable hydrolysate.
Table 10 presents the different ethanol concentrations, the
ethanol yields with respect to Gay Lussac theoretical values and
the ethanol yields (expressed in ethanol g/g reducing sugar) of
hydrolysates.
From the results, it appears that Rhyzopus spp. can grow on
non- detoxified hydrolysates from pretreatment. It shows that
Rhizopus spp can be used for fermentation of hydrolysates from
different treatments. The results are similar to those of Millati et
al. [71] who observed that certain strains of Rhizopus are capable
of producing ethanol from hydrolysates obtained from the
pretreatment of forest residues using dilute sulfuric acid.
Fermentation yield of peeling hydrolysates is greater than that of
stems. This can be attributed to the higher ash content of cassava
peelings and hence higher min- eral content which can be
favorable for mould growth. Castro and Roberto [72] reported that
mineral supplementation plays an important role in rice straw
hydrolysates fermentation. The yield of
0.34 ethanol g/g reducing sugar obtained from stem hydrolysates
are similar to those of Millati et al. [71] who obtained 0.35 ethanol
g/g reducing sugar yield by fermenting, with the aid of Rhizopus
oryzae B., hydrolysates from sulfuric acid pretreatment.
As concerns enzymatic hydrolysates, it was noticed that peeling
hydrolysates with high reducing sugar content presented a lower
ethanol yield. This may be due to the fact that the medium might
have contained cellobiose, a reducing glucose dimer and an enzy-
matic hydrolysis intermediate, which cannot be assimilated by
Saccharomyces cerevisiae. Many works report that without genetic
modification, Saccharomyces cerevisiae cannot assimilate cello-
biose [73,74] due to the absence of cellobiose transporter and b
glucosidase necessary for cellobiose hydrolysis into glucose.
4. Conclusion
This work shows that thermohydrolysis can be used to hydro-
lyse hemicellulose in cassava stems and peelings. This process led
to an increase in hydrolysis yield of cellulose in cassava peelings
and stems of 39.84% and 55.4% respectively. From the results, the
hydrolysis of hemicellulose leads to an increase in the enzymatic
hydrolysis of cellulose. The hydrolysates from pretreatment can be
fermented without prior detoxification with the aid of Rhizopus
spp. Respective ethanol yields of 76.84% and 66.70% for peelings
and stems are obtained. As perspective, other works can be carried
out on hydrolysates fermentation optimization.
Acknowledgment
The authors thank the Microbiology Laboratory of the National
Advanced School of Agro-Industrial Sciences (ENSAI) of the Uni-
versity of Ngaoundere-Cameroon for providing the microbial
strains used in this work and the Ministry of Higher Education of
Cameroon for its financial support through the Special Fund Ac-
count for the modernization of research in state Universities
◦
(Presidential Decree N 2009/121 of April 8, 2009).
References
[1] U.S. Orlando, A.U. Base, W. Nishijima, M. Okada, Preparation of agricultural
residue anion exchangers and its nitrate maximum adsorption capacity,
Chemosphere 48 (10) (2002) 1041e1046.
[2] H. Palonen, Role of Lignin in the Enzymatic Hydrolysis of Lignocellulose, 520,
VTT publication, Finland, 2004, pp. 1e80. VTT Technical Research Centre,
Finland, JULKAISIJA-UTGIVARE-PUBLISHER.
[3] B. Karki, D. Maurer, S. Jung, Efficiency of pretreatments for optimal enzymatic
saccharification of soybean fiber, Bioresour. Technol. 102 (11) (2011)
6522e6528.
[4] M. Takam, Fore^t et Valorisation E´ nerge´tique de la Biomasse
pour le
De´veloppement d’une Economie Verte en Zone COMIFAC, 1er congre`s
des fore^ts communales d’Afrique centrale, Yaounde´-Cameroun, 2011.
[5] M. Balat, Production of bioethanol from lignocellulosic materials via the
biochemical pathway: a review, Energy Convers. Manag. 52 (2) (2011)
858e875.
[6] A. Singh, S. Punita, S. Alok Kumar, S. Namita, R.B. Narsi, Comparative study on
ethanol production from pretreated sugarcane bagasse using immobilized
Saccharomyces cerevisiae on various matrices, Renew. energy 50 (C) (2013)
488e493.
[7] P. Alvira, E. Toma´s-Pejo´, M. Ballesteros, M.J. Negro, Pretreatment
technologies
for an efficient bioethanol production process based on enzymatic hydrolysis:
a review, Bioresour. Technol. 101 (13) (2010) 4851e4861.
[8] Teng-Chieh Hsu, Gia-Luen Guo, Wen-Hua Chen, Wen-Song Hwang, Effect of
dilute acid pretreatment of rice straw on structural properties and enzymatic
hydrolysis, Bioresour. Technol. 101 (13) (2010) 4907e4913.
[9] L. Zhu, J.P. O’Dwyer, V.S. Chang, C.B. Granda, M.T. Holtzapple, Structural fea-
tures affecting biomass enzymatic digestibility, Bioresour. Technol. 99 (9)
(2008) 3817e3828.
[10] L.C. Teixeira, J.C. Linden, H.A. Schroeder, Simultaneous saccharification and
 cofermentation of peracetic acid-pretreated biomass, Appl. Biochem. Bio-
technol. 84e86 (1e9) (2000) 111e127.
[11] S. Kim, M.T. Holtzapple, Lime pretreatment and enzymatic hydrolysis of corn
stover, Bioresour. Technol. 96 (18) (2005) 1994e2006.
[12] Q. Qing, B. Yang, C.E. Wyman, Xylooligomers are strong inhibitors of cellulose
hydrolysis by enzymes, Bioresour. Technol. 101 (24) (2010) 9624e9630.
[13] Yu Qiang, Xinshu Zhuang, Shuang Liang, Min Chao He, Yu Zhang,
Zhenhong Yuan, Wei Qi, Qiong Wang, Wen Wang, Xuesong Tan, Liquid hot
water pretreatment of sugarcane bagasse and its comparison with chemical
pretreatment methods for the sugar recovery and structural changes, Bio-
resour. Technol. 129 (2013) 592e598.
[14] H.A. Ruiz, R.M. Rodríguez-Jasso, B.D. Fernandes, A.A. Vicente, J.A. Teixeira,
Hydrothermal processing, as an alternative for upgrading agriculture residues
and marine biomass according to the biorefinery concept: a review, Renew.
Sustain. Energy Rev. 21 (2013) 35e51.
[15] N.D. Hinman, J.D. Wright, W. Hoagland, C.E. Wyman, Xylose fermentation, an
economic analysis, Appl. Biochem. Biotechnol. 20/21 (1) (1989) 391e401.
[16] G. Garrote, J.M. Cruz, A. Moure, H. Dominguez, J.C. Parajo´, Antioxidant
activity
of byproducts from the hydrolytic processing of selected lignocellulosic ma-
terials, Trends Food Sci. Technol. 15 (3e4) (2004) 191e200.
[17] D. Nabarlatz, A. Ebringerova´, D. Montane´, Autohydrolysis of agricultural
by-
products for production of xylooligosaccharides, Carbohydr. Polym. 69 (1)
(2007) 20e28.
[18] A. Sluiter, B. Hames, D. Hyman, C. Payne, R. Ruiz, C. Scarlata, J. Sluiter,
D. Templeton, J. Wolfe, Determination of Total Solids in Biomass and Total
Dissolved Solids in Liquid Process Samples, National Renewable Energy Lab-
oratory (NREL), 2008. Golden, CO p.6. NREL/TP-510e42621.
[19] P.J. Van Soest, J.B. Robertson, B.A. Lewis, Methods for dietary fiber, neutral
detergent fiber, and nonstarch polysaccharides in relation to animal nutrition,
J. Dairy Sci. 74 (10) (1991) 3583e3597.
[20] A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton, Determination
of Ash in Biomass, National Renewable Energy Laboratory (NREL), 2005.
Golden, CO p.5. NREL/TP-510e42622.
[21] R.K. Sinnott, Chemical engineering design, in: Butterworth-Heinemann (Ed.),
Chemical Engineering, fourth ed. 6, 2005, p. 1056.
[22] G. Garrote, H. Domınguez, J.C. Parajo´, Kinetic modelling of corncob
autohy-
drolysis, Process Biochem. 36 (6) (2001) 571e578.
[23] O. Bobleter, Hydrothermal degradation of polymers derived from plants, Prog.
Polym. Sci. 19 (5) (1994) 797e841.
[24] G. Garrote, H. Dominguez, J.C. Parajo´, Mild autohydrolysis: an
environmen-
tally friendly technology for xyloolisaccharide production for wood, J. Chem.
Technol. Biotechnol. 74 (11) (1999) 1101e1109.
[25] H. Boussarsar, Application de traitements thermique et enzymatique de sol-
ubilisation et saccharification de la fraction he´micellulosique en vue de
la valorisation de la bagasse de canne `a sucre. Reims, 2008.
http://www.theses. fr/2008REIMS019.
[26] Behzad Satari Baboukani, Manouchehr Vossoughi, Iran Alemzadeh, Optimi-
sation of dilute-acid pretreatment conditions for enhancement sugar recovery
and enzymatic hydrolysis of wheat straw, Biosyst. Eng. 111 (2) (2012)
166e174.
[27] G.L. Miller, Use of dinitrosalicylic acid reagent for determination of reducing
sugar, Anal. Chem. 31 (1959) 426e428.
[28] S. Hashimoto, M.D. Shogren, Y. Pomeranz, Cereal pentosans: their estimation
and significance. I. Pentosans in wheat and milled wheat products, Cereal
Chem. 64 (1) (1987) 30e34.
[29] Akpinar Ozlem, Serdal Sabanci, Okan Levent, Abdulvahit Sayaslan, Evaluation
of antioxidant activity of dilute acid hydrolysate of wheat straw during xylose
production, Industrial Crops Prod. 40 (2012) 39e44.
[30] M. Selig, N. Weiss, Y. Ji, Enzymatic Saccharification of Lignocellulosic Biomass:
Laboratory Analytical Procedure (LAP), National Renewable Energy Labora-
tory, 2008. Golden, CO. NREL/TP-510e42629.
[31] E. Varga, K. Re´czey, G. Zacchi, Optimization of steam pretreatment of
corn
stover to enhance enzymatic digestibility, Appl. Biochem. Biotechnol. 114
(1e3) (2004) 509e523.
[32] D. Chabasse, J.P. Bouchara, L. De gentile, S. Bran, B. Cimon, P. Penn, Cahier de
formation biologie me´dicale, les moisissures d’inte´re^t me´dical.
Laboratoire de parasitologie- mycologie du CHU de d’Angers-4, rue Larrey
49033 Angers cedex, Bioforma (2002) 159.
[33] K.C. Thomas, S.H. Hynes, W.M. Ingledew, Practical and theoretical consider-
ations in the production of high concentrations of alcohol by fermentation,
Process Biochem. 31 (4) (1996) 321e333.
[34] Han Minhee, Yule Kim, Youngran Kim, Bongwoo Chung, Gi-Wook Choi, Bio-
ethanol production from optimized pretreatment of cassava stem, Korean J.
Chem. Eng. 28 (1) (2011) 119e125.
[35] Chen Cong Jin, Yue Li, Qiu Ming Su, Li Juan Qu, Dong Chen, Study on chemical
compositions of Manihot esculenta Crantz (M. utilissima Pohl) stalks, Adv.
Mater. Res. 236e238 (2007) 394e398.
[36] N.S. Pooja, G. Padmaja, Pretreatment techniques to enhance the enzymatic
degradability of agricultural and processing residues of cassava, J. Microbiol.
Biotechnol. Res. 4 (1) (2014) 57e67.
[37] P. Lounglawan, M. Khungaew, W. Suksombat, Silage production from cassava
peel and cassava pulp as energy source in cattle diet, J. Animal Veterinary Adv.
10 (8) (2011) 1007e1011.
[38] O.J. Babayemi, O.J. Ifut, U.A. Inyang, L.J. Isaac, Quality and chemical composi-
tion of cassava wastes ensiled with Albizia saman pods, Agric. J. 5 (3) (2010)
225e228.
[39] P.R. Adler, M.A. Sanderson, A.A. Boateng, P.J. Weimer, H.-J.G. Jung, Biomass
yield and biofuel quality of switchgrass harvested in fall or spring, Agron. J. 98
(6) (2006) 1518e1525.
[40] B.S. Dien, G. Jung H-J, K.P. Vogel, M.D. Casler, J.F.S. Lamb, L. Iten, R.B. Mitchell,
G. Sarath, Chemical composition and response to dilute-acid pretreatment
and enzymatic saccharification of alfalfa, reed canarygrass, and switchgrass,
Biomass Bioenergy 30 (10) (2006) 880e891.
[41] T. Ingram, K. Wo€rmeyer, J.C.I. Lima, V. Bockemü hl, G. Antranikian, G. Brunner,
I. Smirnova, Comparison of different pretreatment methods for lignocellulosic
materials. Part I: conversion of rye straw to valuable products, Bioresour.
Technol. 102 (2011) 5221e5228.
[42] K. Liu, X. Lin, J. Yue, X. Li, X. Fang, M. Zhu, J. Lin, Y. Qu, L. Xiao, High con-
centration ethanol production from corncob residues by fed-batch strategy,
Special Issue on Lignocellulosic Bioethanol: Current Status and Perspectives,
Bioresour. Technol. 101 (2010) 4952e4958.
[43] J. Ai, U. Tschirner, Fiber length and pulping characteristics of switchgrass,
alfalfa stems, hybrid poplar and willow biomasses, Bioresour. Technol. 101
(2010) 215e221.
[44] A. Goshadrou, K. Karimi, M.J. Taherzadeh, Bioethanol production from sweet
sorghum bagasse by Mucor hiemalis, Ind. Crops Prod. 34 (2011) 1219e1225.
[45] Z. Wang, D.R. Keshwani, A.P. Redding, J.J. Cheng, Sodium hydroxide pre-
treatment and enzymatic hydrolysis of coastal Bermuda grass, Bioresour.
Technol. 101 (2010) 3583e3585.
[46] R. Gonzalez, T. Treasure, R. Phillips, H. Jameel, D. Saloni, R. Abt, J. Wright,
Converting Eucalyptus biomass into ethanol: financial and sensitivity analysis
in a co-current dilute acid process, Part ii. Biomass Bioenergy 35 (2011)
767e772.
[47] X. Pan, N. Gilkes, J. Kadla, K. Pye, S. Saka, D. Gregg, K. Ehara, D. Xie, D. Lam,
J. Saddler, Bioconversion of hybrid poplar to ethanol and co-products using an
organosolv fractionation process: optimization of process yields, Biotechnol.
Bioeng. 94 (2006) 851e861.
[48] A.M. Joglekar, A.T. May, Product excellence through design of experiments.
Cereal Food, World 32 (1987) 857e868.
[49] P. Dalgaard, L.V. Jorgensen, Predicted and observed growth of Listeria mon-
ocytogenes in seafood challenge tests and in naturally contaminated cold
smoked salmon, Int. J. Food Microbiol. 40 (1e2) (1998) 105e115.
[50] D. Bas¸ , I.H. Boyac, Modeling and optimization I: usability of response
surface methodology, J. Food Eng. 78 (3) (2007) 836e845.
[51] M. Laser, D. Schulman, S.G. Allen, J. Lichwa, M.J. Antal Jr., L.R. Lynd,
A comparison of liquid hot water and steam pretreatments of sugar cane
bagasse for bioconversion to ethanol, Bioressource Technol. 81 (1) (2002)
33e44.
[52] C.K. Nitsos, K.A. Matis, K.S. Triantafyllidis, Optimization of hydrothermal
pretreatment of lignocellulosic biomass in the bioethanol production process,
ChemSusChem 6 (1) (2013) 110e122.
[53] S. Liu, Woody biomass: niche position as a source of sustainable renewable
chemicals and energy and kinetics of hot-water extraction/hydrolysis, The
biorefinery: sustainably renewable route to commodity chemicals, energy,
and materials Selected papers from the Second International Biorefinery
Conference, Biotechnol. Adv. 28 (2010) 563e582.
[54] M. Heitz, E. Capek-Me´nard, P.G. Koeberle, G. Gagne´, E. Chornet, R.P.
Overend,
J.D. Taylor, E. Yu, Fractionation of Populus tremuloides at the pilot plant scale:
optimization of steam pre-treatment conditions using the STAKE II technol-
ogy, Bioresour. Technol. 35 (1) (1986) 23e32.
[55] R. Mehlberg, G.T. Tsao, Low liquid hemicellulose hydrolysis of hydrochloric
acid, Washington, in: 178th ACS National Meeting, 1979, pp. 8e12.
[56] H. Pin´kowska, W. Pawe1, Z. Adrianna, Hydrothermal decomposition of xylan
as a model substance for plant biomass waste hydrothermolysis in subcritical
water, Biomass Bioenergy 35 (9) (2011) 3902e3912.
[57] G. Garrote, Herminia Domınguez, Juan C. Parajo´, Kinetic modelling of corncob
autohydrolysis, Process Biochem. 36 (6) (2001) 571e578.
[58] B. Saake, R. Lehnen, Lignin, in Ullmann’s Encyclopedia of Industrial Chemistry,
Wiley-VCH Verlag GmbH & Co. KGaA, 2007.
[59] J.M.L. Penninger, R.J.A. Kersten, H.C.L. Baur, Reactions of diphenylether in
supercritical water - mechanism and kinetics, J. Supercrit. Fluids 16 (2) (1999)
119e132.
[60] G. Gonza´lez, J. Salvado´, D. Montane´, Reactions of vanillic acid in sub- and
supercritical water, J. Supercrit. Fluids 31 (1) (2004) 57e66.
[61] J.h. Lora, M. Wayman, Delignification of Hardwoods by Autohydrolysis and
Extraction, Tappi, 1978 [Technical Association of the Pulp and Paper Industry],
http://agris.fao.org/agris-search/search.do?recordID¼US19780336126.
[62] J. Jakobson, B. Hortling, P. Erins, J. Sundquist, Characterization of alkali soluble
fraction of steam exploded birch wood, Holzforschung - Int. J. Biol. Chem.
Phys. Technol. Wood 49 (1) (1995) 51e59.
[63] F.A. Agblevor, S. Batz, J. Trumbo, Composition and ethanol production po-
tential of cotton gin Residues, Appl. Biochem. Biotechnol. 105 (1e3) (2003)
219e230.
[64] M.-F. Li, C.-Z. Chen, R.-C. Sun, Effect of pretreatment severity on the enzymatic
hydrolysis of bamboo in hydrothermal deconstruction, Cellulose 21 (6) (2014)
4105e4117.
[65] Y. Zhao, B. Wu, B. Yan, P. Gao, Mechanism of cellobiose inhibition in cellulose
hydrolysis by cellobiohydrolase, Sci. China C Life Sci. 47 (2004) 18e24.
[66] P. Andri´c, A.S. Meyer, P.A. Jensen, K. Dam-Johansen, Reactor design for mini-
mizing product inhibition during enzymatic lignocellulose hydrolysis: I.
 Significance and mechanism of cellobiose and glucose inhibition on cellulo-
lytic enzymes, Biotechnol. Adv. 28 (2010) 308e324.
[67] L. Murphy, C. Bohlin, M.J. Baumann, S.N. Olsen, T.H. Sørensen, L. Anderson,
K. Borch, P. Westh, Product inhibition of five Hypocrea jecorina cellulases,
Enzyme Microb. Technol. 52 (2013) 163e169.
[68] M. Gruno, P. Va€ljam€ae, G. Pettersson, G. Johansson, Inhibition of the Tricho-
derma reesei cellulases by cellobiose is strongly dependent on the nature of
the substrate, Biotechnol. Bioeng. 86 (2004) 503e511.
[69] Y.-H. Lee, L.T. Fan, Kinetic studies of enzymatic hydrolysis of insoluble cellu-
lose: analysis of the initial rates, Biotechnol. Bioeng. 24 (1982) 2383e2406.
[70] K.-K. Oh, S.-W. Kim, Y.-S. Jeong, S.-I. Hong, Bioconversion of cellulose into
ethanol by nonisothermal simultaneous saccharification and fermentation,
Appl. Biochem. Biotechnol. 89 (2000) 15e30.
[71] R. Millati, L. Edebo, M.J. Taherzadeh, Performance of Rhizopus, Rhizomucor,
and Mucor in ethanol production from glucose, Xylose and wood hydroly-
sates, Enzyme Microb. Technol. 36 (2e3) (2005) 294e300.
[72] R.C. Castro, A. de, I.C. Roberto, Effect of nutrient supplementation on ethanol
production in different strategies of saccharification and fermentation from
acid pretreated rice straw, Biomass Bioenergy 78 (2015) 156e163.
[73] S.-J. Ha, J.M. Galazka, S.R. Kim, J.-H. Choi, X. Yang, J.-H. Seo, N.L. Glass,
J.H.D. Cate, Y.-S. Jin, Engineered Saccharomyces cerevisiae capable of simul-
taneous cellobiose and xylose fermentation, Proc. Natl. Acad. Sci. 108 (2011)
504e509.
[74] Y.-H. Bae, K.-H. Kang, Y.-S. Jin, J.-H. Seo, Molecular cloning and expression of
fungal cellobiose transporters and b-glucosidases conferring efficient cello-
biose fermentation in Saccharomyces cerevisiae, J. Biotechnol. 169 (2014)
34e41.