Construction and Building Materials 139 (2017) 551–561

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Virtual Special Issue

Bio Based Building Materials
Effect of hornification on the structure, tensile behavior and fiber matrix
bond of sisal, jute and curauá fiber cement based composite systems
Saulo Rocha Ferreira a, Flávio de Andrade Silva b,⇑, Paulo Roberto Lopes Lima c, Romildo Dias Toledo Filho a
a
Department of Civil Engineering, COPPE, Universidade Federal do Rio de Janeiro, P.O. Box 68506, 21941-972 Rio de Janeiro, RJ, Brazil
b
Department of Civil Engineering, Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio), Rua Marques de São Vicente 225, 22453-900 Rio de Janeiro, RJ, Brazil
c
Technology Department, Universidade Estadual de Feira de Santana, Av. Transnordestina, S/N, Novo Horizonte, 44.036-900 Feira de Santana, BA, Brazil

h i g h l i g h t s

Natural fiber tensile strength can be increased with the application of the hornification treatment.

New bonds between polymer chains in the microfibrills can be created with hornification.

Cellulose crystallinity can be altered with the used treatment.

Significant improvements in the fiber-matrix interface were verified through hornification.

a r t i c l e i n f o a b s t r a c t

Article history: Several fiber treatments are used to mitigate the high water absorption of vegetable fibers. Wetting and
Received 1 July 2016 drying cycles are usually performed in the industry of paper and cellulose to reduce the volume variation
Received in revised form 9 September 2016 of these fibers. This procedure stiffens the polymeric structure of the fiber-cells (this process is known as
Accepted 4 October 2016
hornification) resulting in a higher dimensional stability. The aim of this study is to determine the effect
Available online 13 October 2016
of the hornification on the chemical and mechanical behavior of natural fibers and how these properties
influence the fiber matrix bond. For this purpose, 5 and 10 cycles of wet and drying were applied to
Keywords:
curauá, jute and sisal fibers. Fiber pull-out tests were performed in the embedment length of 25 mm.
Hornification
Natural fibers
Direct tensile tests were performed in natural and hornified fibers. Furthermore, X-ray diffraction, ther-
Mechanical testing mogravimetry analysis, infra red spectroscopy and nuclear magnetic resonance were used to investigate
Chemical characterization the influence of the hornification on the chemical properties of the studied fibers. Modifications on the
fiber morphology were observed with a scanning electron microscope. The results indicate changes on
the tensile strength and strain capacity of the studied fibers, showing that morphology, physical aspects
and chemical composition play an important role on the efficiency rate of hornification. Significant
improvement in the fiber-matrix interface was observed through the pullout tests. It was concluded that
5 cycles promotes a better performance to curauá and sisal fibers. Only the sisal fibers show improvement
on its bond mechanisms after 10 cycles.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Vegetable fiber reinforced composites have the potentiality to
become the ultimate green material option, minimizing the use
of natural resources and overall lifetime impact. Due to its wide
availability, especially in tropical countries, these fibers have a
low cost and present the great advantage of being renewable.
Regarding the drawbacks in the use of those fibers, treatments
⇑ Corresponding author.
E-mail address: fsilva@puc-rio.br (F.A. Silva).
have been studied to reduce the impact of a low chemical bond
and the high water absorption capacity which causes a volumetric
instability. This volumetric instability results in a partial loss of the
fiber-matrix physical contact damaging the interfacial transition
zone.
Several fiber treatments can be used to reduce the natural fibers
volumetric instability and increase their affinity with a cementi-
tious matrix [1,2]. These procedures can clean the fiber surface,
modify its chemistry, lower the moisture uptake and increase the
surface roughness [3]. As natural fibers bear hydroxyl groups from
cellulose and lignin, they are subject to chemical modification. The

http://dx.doi.org/10.1016/j.conbuildmat.2016.10.004
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
552 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561
hydroxyl groups may be involved in the hydrogen bonding within
the cellulose molecules thereby reducing the activity towards the
matrix. Chemical modifications may activate these groups and
can introduce new moieties that can effectively lead to a chemical
interlock with the matrix [4,5]. Mercerization, isocyanate treat-
ment, acrylation, permanganate treatment, acetylation, methane
cold plasma, silane treatment and peroxide treatment with various
coupling agents and other pretreatments of natural fibers have
achieved various levels of success for improving fiber strength,
fiber stiffness and fiber-matrix bond [2,6–8]. Furthermore, thermal
treatments can also be used to improve the natural fiber properties
[9,10]. Usually a combination of chemical and thermal treatments
are applied. Acid and alkaline treatments were proposed by Abou-
Yousef et al. [9] as a solution for sugar cane delignification. At
200 °C, the loss of material due to pyrolysis is still low, and only
the extractives and hydrogen bonds are destroyed [11]. Depending
on the type of lignocellulosic fiber, partial lignin may be removed
in this process [12]. The application of the mentioned treatments
have shown to be effective in promoting changes in the natural
fiber morphologies as well as in their physical and mechanical
behavior [1]. A particular procedure to reduce the water absorption
capacity of natural fibers and to improve the fiber-matrix bond
have been pursued by applying cycles of wetting and drying in
the fibers [1,2,13]. These treatments promoted a reduction in vol-
umetric changes of pulps and fibers of natural origin as well as
alterations in their mechanical properties. This is a result of
repeated wetting and drying cycles which increase the degree of
crosslink within the fiber microstructure. Fibers treated by
methane cold plasma during 10 min presented a higher hydropho-
bicity. The treatment promoted a better adhesion with the cement
matrix which was confirmed by pullout tests [8]. The use of car-
boxylated styrene butadiene polymer latex not only promotes a
fiber surface that is hydrophobic but the polymer latex acts as a
bridge between fiber and the cementitious matrices, strengthening
the interfacial bonding between them [2,4]. The fiber hornification
treatment have been successfully used by Claramunt and co-
workers in vegetable fiber reinforced cement composites [14].
The previous treatment of the fibers improved the durability and
also the mechanical behavior of the studied cement based compos-
ites. However, there still exists a lack of information on how the
use of such treatment modifies the chemical properties and mor-
phology of the natural fibers affecting the fiber-matrix bond with
a cement based matrix. The objective of the present research is
to investigate the effect of cycles of wet and drying on the fiber
properties (chemical and mechanical) and fiber-matrix interface
behavior of natural sisal, curauá and jute fibers. Tensile tests were
carried out to study the effect of the treatments on the fiber. After
the treatments, the fibers were analyzed through SEM, thermo-
gravimetry and X-ray diffraction. Pullout tests were carried out
to investigate the potential benefits caused by the treatments in
the interface bond.
2. Experimental program
2.1. Materials and processing
The sisal fibers used in the present study were obtained from
the sisal plant (Agave sisalana) cultivated in farms located in the
Bahia state, Brazil. They were extracted from the sisal plant leaves
in the form of long fiber bundles. Sisal fibers are extracted from the
leaf by a process called decorticaion. In this process, the leaves are
crushed by a rotating wheel with blunt knives, where only the
fibers prevail [15]. Decorticated fibers were dried under the sun
light. The curauá fibers were provided by the company Pematec,
located in the city of Santarém, in the state of Pará, Brazil. The
curauá fibers are produced in the Amazon region, and extracted
from the leaf of the plant Ananas erectifolius by a process similar
to the process used for sisal fiber. The jute fiber came from the
Amazon region. It is extracted from the stem of the plant Corchorus
capsularis by a combination of processes which comprises the fol-
lowing steps: cutting, retting, shredding, drying, packing and clas-
sification [15].
Regarding to the natural fiber microstructure it is formed by
numerous individual fibers (fiber-cells) (see Fig. 1). The individual
fiber-cells are linked together by means of the middle lamella.
More information on the studied fibers microstructure can be
found in the authors previous works [1,16]. Even though all natural
fibers present a similar morphology the cell wall size, thickness,
porosity and cross-section area may vary. This variation between
fibers can result in a different mechanical behavior. According to
Fidelis et al. [15] curauá, jute and sisal fibers can present an aver-
age tensile strength of 540, 250 and 480 MPa, respectively. These
natural fibers can be classified as high performance fibers. Not only
the morphology but also the chemical composition rules the fiber
mechanical behavior. These fibers present a different amount of
cellulose, each one presenting a different crystallinity degree.
Another important parameter correlated to the tensile strength is
the microfibrillar angle. According to Satyanarayana and Mukher-
jee [17,18], curauá, jute and sisal has a microfibrillar angle of
18.8°, 17.1° and 20°, respectively. Fidelis et al. [15] correlate the
mechanical strength of natural fibers not only to cellulose content,
but also due to a high microfibrillar angle. Coir fibers, with a
microfibrillar angle of 8°, present a tensile strength of 177 MPa
and modulus of 3.4 GPa while Jute fibers have a tensile strength
of 353 MPa and a modulus of elasticity of 26 GPa. It is still not well
understood how the combination of all the aforementioned param-
eters influence the mechanical performance of the natural fibers.
The chemical composition of the used natural fibers are shown in
Table 1. More information on the natural fibers microstructure
and mechanical properties can be found in the authors previous
work [15,19,20].
The used matrix presented a mix design of 1:0.5:0.4 (binder:
sand: water/binder ratio) by weight following the author’s previ-
ous works [21]. The binder was composed by 30% of Portland
cement CP II – F32 defined by the Brazilian standard [22] as com-
posed with filler (in mass: 85% < clinker < 91%; 3% < gypsum < 5%;
6% < limestone filler < 10%) with 32 MPa of compressive strength
at 28 days. Following the recommendations of previous studies,
in order to increase the durability of the composites, the Portland
cement was replaced by 30% of metakaolin (MK) and 40% of fly
ash. This ratio of metakaolin and fly ash aims to guarantee the
durability of the fiber once a matrix free of calcium hydroxide is
obtained, as shown in the previous research [16]. The fly ash also
ensures greater workability to the matrix that, within the context
of high-performance composites, is a desirable property as it
ensures a better homogenization of the vegetable fibers [1].
The sand was processed to obtain a maximum diameter of
840 lm. The used superplasticizer was the Glenium 51 (type PA)
with solids content of 31%. A viscosity modifier Rheomac UW
410, (manufactured by BASF), at a dosage of 0.8 kg/m3 was also
used in order to avoid segregation and bleeding during molding.
The matrix used in this study showed a flow table spread value
of 450 mm according to ASTM C1437 [23] and compressive
strength at 28 days of 31 MPa, according to ASTM C 39 [24].
The mixtures were produced in a room with controlled temper-
ature (21 ± 1 °C) using a mixer with capacity of 5 L. The mixing
procedure is described as follow. All dry components were homog-
enized in the mixer. The water and superplasticizer were added
and mixed for 2 min at a speed of 125 RPM. After that, the process
was stopped during 30 s to remove the material retained in the
 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561
Fig. 1. Natural fibers cross-section morphology: (a) curauá, (b) jute and (c) sisal.
Table 1
Summary of average values of the chemical composition and morphological characterization of sisal, curauá and jute fibers.
Fibers Cellulose (%) Hemicellulose (%) Lignin (%)
Curauá 58.8 23.8 14.7
Jute 74.4 15 8.4
Sisal 60.5 25.7 12.1
mixer. Then, the mixing procedure continued for 2 min at 220 RPM
and, finally for a further 5 min at 450 RPM.
To mold the pullout specimens a special mold was used. After
filling the mold with the matrix, the top cap was fixed and the fiber
stretched slightly for alignment. The mortar was placed in plastic
bags before being placed in the mold as to facilitate the casting
process. The embedment length of 25 mm was used based on the
author’s previous works [2]. After 24 h, the specimens were
demolded and placed in a fog room (HR% P 95%) to moist curing
for 7 days for the pullout test [2].
2.2. The fiber hornification treatment
The detailed description of the used treatment is described as
follow. The natural fibers were placed in a container with water
(T = 22 °C) during 3 h. The drying process was carried out in a fur-
nace at a temperature of 80 °C. The used furnace is equipped with
an electronic temperature control and connected to a scale, with a
precision of 0.01 g to record the mass loss during the drying
process.
The furnace was programmed to reach 80 °C at a heating rate of
1 °C/min and to maintain this temperature for 16 h. After 16 h of
drying, the furnace was cooled down to the temperature of 22 °C
in order to avoid possible thermal shock to the fibers. This proce-
dure was repeated 5 and 10 times. More details can be obtained
elsewhere [1].
2.3. Fiber morphology characterization
The fiber’s morphology was investigated using an SEM Hitachi
TM3000. The microscope was operated under an accelerating volt-
age of 15 kV.
A pre-coating with a thin layer of approximately 20 nm of gold
was done to make the fiber conductive and suitable for analysis. In
order to measure the fiber’s cross-sectional area, for each single
fiber used in the pullout and tensile test, an adjacent piece of the
fiber (immediately next to the one tested) was kept for future mea-
surement and morphology characterization using the SEM. The
obtained images were post-processed using ImageJ, a Java-based
image processing program.
553
Number of fiber-cells Lumens area (lm2) Diameter of lumens (lm) Total area (mm2)
300 358.93 0.8 0.008
26 1014.1 6.7 0.004
144 5796.9 8.2 0.023
2.4. Fiber water absorption capacity
For the water absorption test a bundle of fibers weighting 200 g
in the dry state were used. The specimens were dried in a furnace
at 80 °C until a constant mass value was reached. The fiber bundles
were then immersed in distilled water until its maximum water
absorption capacity was reached. At the immersion times of 10,
30, 60, 120, 180, 240, 300, 360, 420, 480 and 1440 min the fibers
were removed, patted dry with lint free cloth, and weighed. For
the drying process a furnace at a constant temperature of 80 °C
was used. The used system is equipped with an electronic temper-
ature control and connected to a scale, with a precision of 0.01 g.
The furnace was programmed to reach 80 °C at a heating rate of
1 °C/min and to maintain this temperature for 16 h. Water absorp-
tion is expressed as increase in weight percent. Fig. 2 shows an
example of the used process performed for a sisal fiber.
2.5. Thermogravimetrical analysis (TGA)
The tests were performed in a SDT Q600 simultaneous TGA/
DTA/DSC from TA Instruments. Samples weighing 10 mg were sub-
mitted to a heating rate of 10 °C/min until reaching 1000 °C in a
platinum crucible using 100 mL/min of nitrogen as the purge gas.
2.6. X-ray diffraction (XRD)
The XRD measurements were performed on a Bruker D8 Focus
(Cuka radiation = 1.54056 A). The diffracted intensity of radiation
was measured in a 2h range between 10° and 40° with a step size
of 0.05° at 3°/min. For sample preparation the fibers were scissor
cut to particle sizes of less than 0.5 mm (digitally measured from
optical microscopy by imageJ software). The test was performed
using a ring specimen holder made of PMMA.
The crystallinity index (Crl) was calculated using the intensity
values corresponding to the diffraction of the crystalline structure
and the amorphous fraction, according to the method proposed by
Müller et al. [25]:
Acryst
Cri ð%Þ ¼ ð1Þ
Atotal
554 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561
(a)
Fig. 2. Example of the wetting and drying cycle performed for the sisal fiber: (a) water absorption and desorption curve, (b-1) furnace used in the drying cycles and (b-2) the
scale inside the furnace.
where Cri is the crystalline index; Acryst is the sum of crystalline
band areas; and Atotal is the total area under the diffractograms.
The crystalline area correspond to the area under peaks located at
2h = 16° and 22.9°.
2.7. Nuclear magnetic resonance (NMR): relaxometry measurements
The proton spin–lattice relaxation time (T11H and T12H) mea-
surements were carried out in a Maran Ultra 0.54 T device (Reso-
nance – Oxford, UK) at 30 °C (±1), using a receiver gain of 4%, 4
scans, recycle delay of 3 s, spin-lock of 23.4 kHz and fiber samples
in the solid state.
2.8. Fourier transform infrared (FTIR) spectroscopy
FT-IR analyses were performed in a Nicolet iN10 Spectrometer
(Thermo Scientific, U.S.A), using a MCT-B detector and a diffuse
reflectance probe, with resolution of 4 cm1 in the range [4000–
675] cm1. Stored results were averages of 24 scans.
The energy of the hydrogen bonds (Eh) for OH stretching bands
has been calculated according to Eq. (2) proposed by [26]:
1 v0v
Eh ðkJÞ ¼
k v0
where v 0 is the standard frequency corresponding to the vibration
of free OH groups (3650 cm1); v is the frequency of the bonded OH
groups; and k is a constant (k = 1.6  102 kcal1).
2.9. Direct tensile tests
The tensile tests were performed in an electromechanical test-
ing machine Shimadzu AG-X with a load cell of 1 kN. The tests
were performed in 15 specimens using a displacement rate of
0.1 mm/min. The fibers with a gage length of 50 mm were glued
to a paper template for better alignment in the machine and for
a better griping with the upper and lower jaws in accordance with
ASTM C1557 [27]. The fiber strain was calculated based on the dis-
placement measured by the cross head movement.
2.10. Single fiber pullout tests
The pullout tests were performed in an electromechanical test-
ing machine Shimadzu AG-X with a load cell of 1 kN. The tests
were carried out using a displacement rate of 0.1 mm/min. The
samples were fixed in the machine grips through a system with
(b-1)
(b-2)
hinged-fixed boundary conditions as mentioned by [2]. Ten repeti-
tions were performed for each condition (natural and hornified
fibers) to understand the influence of the treatment. The embed-
ment length of 25 mm was chosen based on a previous study [2].
All tests were carried out in laboratory condition with samples
with approximately 65% of internal moisture. The samples were
cured for 6 days in a fog room and left for one day in laboratory
controlled condition (1 °C and 65 ± 2% RH) before testing.
3. Results and discussion
3.1. Fiber morphology and water absorption capacity
Several modifications were observed in the fiber morphology
after the hornification treatment. It was observed by SEM a reduc-
tion of the fiber cross-section area after 5 and 10 cycles of wetting
and drying. The diameter of the lumens was also reduced after the
cycles. The average reduction was 18, 15 and 17% for curauá, jute
and sisal, respectively. Curauá and jute fibers presented a fiber-
cell desconfiguration after the cycles. This fiber structure collapse
results in a reduction of the lumen size but also leads to a delam-
ination between fibercell and the middle lamella. Rupture at the
ð2Þ cell wall is observed in Fig. 3. It is known already that hornification
is promoted by excessive fatigue due to wetting and drying in the
cell walls, causing shrinkage of the natural fibers due to the forma-
tion of hydrogen bonds [28]. This modification promotes an inten-
sely bonded structure [29]. This process results in a higher stiffness
of the fibers leading to a brittle behavior [30]. It is possible to
assume that due to the increase of this higher interaction based
in h-bonds, the stress promoted due to the swelling process may
cause not only delamination but also rupture in the cell wall, start-
ing after 5 cycles. This same rupture and delamination process is
only observed in sisal after 10 cycles.
According to Murata and Masuda [31] after absorption of mois-
ture, the lumen shrinks and the cell wall swells mainly in the thick-
ness direction. It was found that the more distant it is from the
middle lamella, greater is the expansion in the thickness direction.
The middle lamella, composed mainly of lignin (hydrophobic),
does not present a significant swell. In this way, the higher amount
of cellulose presented by curauá and jute may promote a rapid
stiffening of the cell wall layer, shrinking the fibercells and pro-
moting the delamination and layer rupture (see Fig. 3). According
to Santos [32] and Lima et al. [33] sisal fibers present a similar fail-
ure after 23 cycles of wetting and drying. Apart to this phe-
nomenon, all fibers present a continuous reduction of cross-
 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561 555

Curauá fiber

(a) (b) (c)

50 µm
Jute fiber

(a) (b) (c)

50 µm
Sisal fiber
(a) (b) (c)

50 µm

Fig. 3. Natural fibers cross-section morphology after sequential cycles of wetting and drying: (a) natural, (b) 5 and (c) 10 cycles. In the sisal fiber cross section it was observed
an evolution process of fiber cell detachment from the middle lamella after the 10th cycle.

section followed by the closure of lumens until the cycle number
10.
Fig. 4 shows the results of water absorption after the wetting
and drying cycles for all fibers. Sisal and jute present a similar
behavior. After each cycle the fibers reduce its water retention
value. The same behavior was reported by Ferreira et al. [2]. The
curauá fiber behaves in an opposite way. After each cycle, the
curauá fiber absorbs more water, reaching its maximum sorption
capacity. This higher absorption capacity may be correlated to
the micro crack formations due to the cycles, increasing the fiber
porosity.
3.2. Thermogravimetry analysis (TGA)
Fig. 5 shows the thermogravimetry analysis (TGA) performed in
the natural fibers. It is possible to observe three intervals of decom-
position for all studied fibers, where the first stage is correlated to
water loss (around 100 °C). The second step occurs between 250
and 350 °C. It can be traced back to the simultaneous decomposi-
tion of cellulose, hemicelluloses, lignin and pectins [34].
The third peak occurs in intervals from 350 and 400 °C and it is
characteristic of cellulose degradation. At the end of the third peak
there is a ‘‘sholder” relative to the slow degradation of lignin,
556 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561

Fig. 4. Water absorption and desorption curves before and after the wetting and drying curves for (a) curauá, (b) jute and (c) sisal.

Fig. 5. Thermal analysis of the curauá, jute and sisal fibers before and after the wetting and drying cycles: TG/DTG and Differential Scanning Calorimetry.

which occurs between 250 and 600 °C. Each macromolecule pre-
sents a different thermal resistance. Cellulose presents a higher
resistance to temperature. It is possible to observe the decomposi-
tion phase of lignin, which begins around 250–300 °C, same value
according to literature [35,36]. According to the DSC curve, the
small peak at 250–300 °C is due to the endothermic reaction from
the hemicellulose decomposition.
Fig. 5 also shows the influence of the treatments in the compo-
sition of the fibers. Cycles of wetting and drying promote an
increase in the enthalpy of decomposition of cellulose. The DSC
curves of hornified fibers present a higher enthalpy when com-
pared to natural untreated fibers. This higher energy necessary to
the depolimerization of the macromolecules may be correlated to
a higher number of bonding (probably hydrogen bonds) meaning
a higher crystalline macromolecule. That means the water present
in the curauá is clearly strongly bonded with the fibers macro-
molecules. This enthalpy is also directly correlated to the com-
bined water, which means, more hydrogen bonds, which is
directly correlated to the crystalline level of the fibers.
The hornification treatment modifies the DSC curves for all
fibers, increasing the cellulose peak intensity. Another similar
change is the intensity of the water peak, indicating a higher
energy to evaporate the water, probably due to the higher bounded
water.
3.3. X-ray diffraction (XRD)
X-ray diffractograms are presented in Fig. 6. The degree of cel-
lulose crystallinity is one of the most important crystalline struc-
ture parameters. The stiffness of cellulose fibers increases and
their flexibility decreases with increasing the ratio of crystalline
to amorphous regions [37]. Curauá present a higher crystallinity
index followed by jute. Sisal presents the lowest values, as also
observed by Poletto et al. [38]. All fibers presented an increase in
the crystallinity index after 5 cycles. The curauá fiber has the most
expressive increase of 23% (see Table 2). When cycled for 10 times
the crystallinity started to reduce.
 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561 557

Fig. 6. X-ray diffractograms for treated and untreated fibers: (a) Curauá, (b) Jute and (c) Sisal.

Table 2
Energy of the hydrogen bonds and summary of crystallinity values (Crl) of curauá, jute and sisal fibers before and after the cycles of wetting and drying.

FTIR NMR XRD

Lignin Crystalline cellulose 1b T11H T12H Crystallinity
3567 cm1 3278 cm1
Fiber Eh (kJ) Eh (kJ) (ms) (ms) (%)
Curauá 1.81 7.23 1.1 54.1 40.88
Curauá 5 1.73 17.68 1.42 39.1 53.20
Curauá 10 2.43 19.27 1.2 72.1 48.49
Jute 1.50 9.10 7.7 87.2 33.81
Jute 5 9.92 153.4 11 126.3 37.56
Jute 10 1.88 14.76 8.4 132.6 25.87
Sisal 3.03 7.79 5.19 98.6 33.33
Sisal 5 1.99 14.87 3.3 65.5 38.52
Sisal 10 5.37 18.06 2.26 43.5 32.26

3.4. Nuclear magnetic resonance (NMR): relaxometry measurements with Methoxyl groups (AOCH3), CAOAC and C@C aromatic ring
groups [41,42].
The NMR T11H and T12H values are presented in Table 2. The
domain with lowest relaxation time T11H refers to the population
of most mobile hydrogen, while T12H is the time of longitudinal
relaxation of the more rigid region. The percentage of each field
indicates the domain size, or percentage of hydrogen atoms com-
posing the region. The mobile hydrogen can be described as free
water in the fiber and T12H as the hydrogen present in the con-
stituents bonding as hydrogen-bonds.
In a general view, the increase of relaxation time to both hydro-
gen (T11H and T12H) indicate stiffness in the h-bonding. This
increase was observed for curauá and jute fibers. This increase is
directly correlated to a densification in the hydrogen bond popula-
tion and its energy. The increase in T11H and T12H can be directly
correlated to the increase in the cellulose degradation enthalpy.
Sisal fiber presents a gradual reduction in T12H values. This
decrease can be correlated to a cellulose structure plasticizing, pro-
moting a reduction in the Young’s modulus after 10 cycles.

3.5. Fourier transform infrared (FTIR) spectroscopy

The FTIR spectra of sisal fibers are shown in Fig. 7. The spectra of
the natural and treated fibers show the vibration modes in agree-
ment with the literature [39] and [40], characteristics of lignocellu-
losic fibers. Natural fibers have cellulose, hemicellulose and lignin
as the most important components. These three important ele-
ments have peaks in the largest region between 800 and Fig. 7. Infrared spectroscopy of natural fibers after 5 and 10 cycles of wetting and
2000 cm1. Mostly of the peaks are relative to lignin, correlated drying cycles.
558 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561
The prominent peak at 1742 cm1 in the spectrum is attributed
to the acetyl groups and hemicelluloses uronic ester or the ester
bond of the carboxylic groups of ferulic and p-coumaric acids of
lignin and/or hemicellulose. In the same spectrum the peak at
1644 cm1 is associated with the double bond C@C vibration of
the aromatic molecule lignin, and the wavelength at 1244 cm1
corresponds to the axial asymmetry of link @COC in ether, ester
and phenol groups present. The peak at 1044 cm1 is assigned to
the stretching of the CAO and vibration of the CAH cellulose group.
Treated fibers present a gradual increase in the intensity of the
peak at 3400 cm1. It is an evidence of the hydrogen bridges
hypothesis, also well correlated to NMR data, except for the sisal
NMR results. According to Zugernmaier [43], the hydroxyl groups
(AOH) in the cellulose, hemicelluloses and lignin build a large
amount of hydrogen bonds between the macromolecules in the
plant fiber cell wall. Subjecting the plant fibers to humidity causes
those bonds to break. The hydroxyl groups then form new hydro-
gen bonds with the water molecules, which induce swelling. It is
needed a high energy to brake those bonds, therefore the cycles
applied to the fiber increase the number of hydrogen bonds, struc-
tural changes in the fibers wall and an increase in the number of
AOH bonds, as can be seen in the FTIR analysis.
It is also observed a modification of the COOH-groups
(1750 cm1). The increase of this particular group results in an
increase of the hydrophobicity of lignin. Carboxyl groups are able
to connect to other functional via hydrogen bonds which may
result in an increased lignin macromolecule netting [44]. This
mechanism promotes a better interaction between lignin and cel-
lulose and hemicellulose, resulting in a stronger material and in a
higher crystallinity, stiffness and strain capacity.
The energy of hydrogen bonds (HB) for the natural and treated
fibers are presented in Table 2. Regarding to the natural fibers, jute
and curauá present, respectively, lower energy of hydrogen bonds
values at 3561 cm1. The same was observed by Poletto et al. [45].
According to the author this could be associated with a higher
quantity of absorbed water in the structure, since the band at
3567 cm1 is also assigned to the weakly absorbed water. The
author says also that fibers with a higher energy of HB usually pre-
sent higher quantities of lignin, contributing to form several
intramolecular hydrogen bonds. After 5 cycles jute fibers presented
a higher increase of h-bond to lignin and cellulose groups followed
by a reduction to the same value of natural fibers after 10 cycles.
Curauá and sisal presented the same behavior. After 5 cycles both
fibers presented a decrease of h-bond from lignin followed by an
increase after 10 cycles. Regarding to the cellulose h-bond energy
both fibers presented the same increasing rate after 5 and 10
cycles. In the case of jute and curauá fibers, even the T12H values
indicate a reduction in the quantity of h-bonds.
3.6. Fiber tensile strength and fiber-matrix bond
Typical stress-strain curves of untreated and treated fibers are
shown in Fig. 8. Results of individual treatments values and aver-
ages are presented in Table 3.
The mechanical and physical properties of the natural fibers are
influenced by many variables. One of the most important variable
is the fiber chemical composition (i.e. amounts of cellulose, hemi-
cellulose and lignin) [45,46]. As an example, higher tensile
strengths are obtained for fibers that contain more crystalline cel-
lulose [37].
For the curauá fibers, the hornification treatment resulted in an
increase of tensile strength and stiffness when treated after 5 and
10 cycles. Exceptionally, the curauá fiber after 5 cycles presented a
decrease in stiffness. Cycles of wetting and drying performed on
jute fibers resulted in an increase in the average strength and a
moderate reduction in the Young’s modulus after the cycles. The
tensile strength of the sisal fiber increased after the used treat-
ment. After 5 cycles, the sisal fiber showed an increase in the ten-
sile strength and stiffness. When treated with 10 cycles, the sisal
tensile strength increased by 5% and was also followed by a lower
data scattering. It was observed a reduction in the stiffness of 16%
and a higher strain capacity.
The hornification process modifies the polymeric structure of
the fiber-cells resulting in changes on tensile strength and strain.
According to Zugernmaier et al. [43], the hydroxyl groups (AOH)
in the cellulose, hemicelluloses and lignin build a large amount
of hydrogen bonds between the macromolecules in the plant fiber
cell wall, subjecting the plant fibers to humidity which causes
those bonds to break. The hydroxyl groups then form new hydro-
gen bonds with the water molecules when at wetting condition,
which induces swelling. It is needed a high energy to brake those
bonds, therefore the cycles applied to the fiber increase the num-
ber of hydrogen bonds, structural changes in the fibers wall and
an increase in the number of AOH bonds. Carboxyl groups are able
to connect to other functional groups by hydrogen bonds which
may result in increased lignin macromolecule netting [37]. That
mechanism promotes a better interaction between lignin and cel-
lulose and hemicellulose, resulting in a stronger fiber, with a
higher crystallinity, stiffness and strain capacity. In fact the appli-
cation of 10 cycles of wetting and drying results in a tendency on
that direction even though not all results showed an improvement
in the fiber mechanical properties. The reason is probably due to
the effects of fatigue stresses applied on fibers during the wet
and dry process. These fatigue stresses can result in a collapse of
the lumen structure or delamination between the lumen and the
middle lamella. According to Santos et al. [32], the excessive num-
ber of cycles may promote a decrease on the sisal fiber tensile
strength. In their work Santos and his co-works found that for
treatments with 20 cycles, the hornification promoted a higher
reduction on tensile strength of sisal fibers (336–234 MPa).
Typical load-slip curves of untreated and treated fibers are
shown in Fig. 9. Results of individual treatments values and aver-
ages are presented in Table 4. The hornified sisal after 5 cycles
showed a higher bond when compared to the untreated fiber.
The same behavior was observed for the curauá fiber. The jute
fibers showed a reduction in the bonding after 5 cycles.
After 10 cycles the sisal fiber presented a higher adhesional and
frictional phases, observed in the stiffness and higher pullout load.
For the other two fibers, the application of 10 cycles seems to be
excessive, leading to a decrease in bond.
The increase in the sisal fiber-matrix bond process can be traced
back to a packing of the fiber cells due to the formation of
hydrogen-bond, reducing the water absorption and dimensional
variation. Another reason for the better bonds is related to the
increase of the superficial roughness after 10 cycles of wetting
and drying. A direct consequence is the reduction of the lumen
size, which consequently reduces the capacity of water absorption
and dimensional change.
The same reduction in lumen size and packing of the fiber cells
was observed on curauá fibers. Even with an increase on the tensile
strength, the pullout behavior show a decrease on the fiber-matrix
bond. It is possible to correlate this result to a surface inactivation
process. Lignin, for example, can block the formation of hydrogen
bonds and prevent inter-fiber bonding. After the pullout test, the
jute and curauá fibers were analyzed by SEM observation. It was
observed a fiber cell delamination, justifying the loss of stiffness
as observed in Fig. 9.
Observing Fig. 3, one can conclude that fibers present a different
sensibility on water and to the numbers of cycles. Damage was
observed on jute fibers, initiated on cycle number 5. Several
changes in the curauá fiber-cells was observed, resulting on the
closure of the lumen and a reduction at the fiber cross section since
 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561 559

Fig. 8. Typical tensile behavior for treated and untreated fibers: (a) Curauá, (b) Jute and (c) Sisal.

Table 3
Average results and standard deviation of tensile tests for treated (5 and 10 cycles) and untreated fibers.

Fibers Cycles Max load (N) Tensile strength (MPa) Strain capacity (mm/mm) Young’s modulus (GPa)
Sisal 0 10.28 (3.43) 447.20 (23.90) 0.03 (0.01) 20.16 (3.52)
5 12.64 (2.14) 608 (23.10) 0.02 (0.01) 22.17 (2.65)
10 15.66 (2.61) 470.25 (16.56) 0.05 (0.01) 17.60 (1.22)
Curauá 0 2.54 (0.46) 632.14 (138) 0.021 (0.01) 38.1 (18.01)
5 5.15 (1.21) 689.30 (180.56) 0.028 (0.005) 52.79 (20.62)
10 3.62 (0.51) 815.12 (106) 0.019 (0.01) 81.4 (15.22)
0 1.05 (0.37) 249.23 (89.02) 0.006 (0.002) 43.9 (12.31)
Jute 5 2.26 (1.40) 488.82 (271.31) 0.008 (0.001) 40.32 (14.83)
10 0.98 (0.35) 543.13 (43.24) 0.01 (0.001) 38.41 (9.21)

Fig. 9. Typical pullout curves for treated and untreated fibers: (a) Curauá, (b) Jute and (c) Sisal.

the first wet and drying process. In the case of sisal, signals of fail-
ure due to swelling fatigue is observed only after 20 cycles, as
reported by Santos et al. [32].
No chemical bond exist between cellulose and hemicellulose,
but mutual adhesion is provided by hydrogen bonds and van der
Waals forces [47]. In this case the treatment induces a larger occur-
rence of intermolecular hydrogen bonding. The cellulose molecules
have a strong tendency to form intermolecular and intramolecular
hydrogen bonds. Cellulose and hemicellulose are non-covalently
bonded (i.e. hydrogen bounded) to each other [48,49], while lignin
and hemicellulose are covalently bonded [50].
Recently, however, it has been suggested that irreversible
hydrogen bonding occurs in wood during drying, similar to that
occurred in pulp fibers, though to a lesser extent [51]. The presence
of the ligno-hemicellulosic matrix in wood may restrict the forma-
tion of hydrogen bonds within the fibers [52]. This suggest that the
chemical composition rule the influence of cycles of wetting and
drying in the hornification efficiency.
560 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561

Table 4
Average results and standard deviation of pullouttest for treated (5 and 10 cycles) and untreated fibers.

Fibers Cycles Pad (N) Dad (mm) rad (MPa) Pfr (N) Dfr (mm) rfr (MPa) Stiffness (N/mm)
Sisal 0 3.73 (1.02) 0.66 (0.19) 0.30 (0.08) 2.56 (0.72) 0.92 (0.25) 0.18 (0.04) 9.31 (4.39)
5 4.89 (0.79) 0.73 (0.23) 0.46 (0.09) 2.64 (0.32) 1.03 (0.16) 0.20 (0.03) 12.19 (3.86)
10 4.53 (1.07) 0.23 (0.11) 0.42 (0.08) 3.88 (0.77) 0.41 (0.16) 0.27 (0.03) 20.60 (6.37)
Curauá 0 3.65 (1.21) 0.35 (0.19) 0.48 (0.22) 1.86 (0.95) 0.53 (0.21) 0.34 (0.19) 30.26 (0.36)
5 4.32 (1.32) 0.45 (0.18) 0.56 (0.13) 1.91 (0.83) 1.09 (0.16) 0.35 (0.11) 26.17 (0.44)
10 2.56 (1.07) 0.38 (0.10) 0.32 (0.18) 1.36 (1.44) 0.84 (0.22) 0.21 (0.09) 5.46 (0.12)
* * *
Jute 0 2.35 (0.79) 0.63 (0.32) 0.86 (0.23) 23.76 (8.99)
* * *
5 1.94 (0.58) 0.73 (0.31) 0.61 (0.17) 10.56 (6.23)
* * *
10 1.73 (0.67) 0.78 (0.42) 0.42 (0.12) 4.36 (0.23)

Pa – Adhesional force; Da – Displacement relative to the adhesional force; rad – Adhesional stress; Pfr – Frictional force; Dfr – Displacement relative to the frictional force; rfr –
Frictional stress.
*
Fiber fracture at pullout test.

4. Conclusions
The present work investigated the effect of hornification treat-
ment on natural fibers in their mechanical and fiber-matrix inter-
face properties. The following conclusions can be drawn from the
present research:

The natural fibers tensile strength increased with the applica-
tion of the hornification treatments. This can be explained by
the change in the cellulose crystallinity and the possible bonds
created between different polymer chains in the microfibrills as
observed in the DSC curves.

With the hornification treatment it was possible to reach a
higher fiber-matrix bond, improving also the frictional mecha-
nism, once the number of cycles does not promote damage to
the fiber structure, as no damage was observed in the case of
the curauá and jute fibers. It was concluded that 5 cycles pro-
motes a better performance to curauá and sisal fibers. Only
the sisal fibers show improvement on its bond mechanisms
after 10 cycles.

The hornification treatment is presented as a simple solution to
improve the mechanical behavior and bond to cement based
systems. The cycles of wetting and drying can improve the
mechanical properties and fiber-matrix bond strength for dif-
ferent types of natural fibers.
References
[1] S.R. Ferreira, P.R.L. Lima, F.A. Silva, R.D. Toledo Filho, Effect of sisal fiber
hornification on the adhesion with Portland cement matrices, Revista Materia
17 (2012) 1024–1034.
[2] S.R. Ferreira, F.A. Silva, P.R.L. Lima, R.D. Toledo Filho, Effect of fiber treatments
on the sisal fiber properties and fiber–matrix bond in cement based systems,
Constr. Build. Mater. 101 (2015) 730–740.
[3] M.T. Ton-That, J. Denault, Development of Composites Based on Natural Fibers,
National Research Counsil, Industrial Materials Institute, The Institute of
Textile Science, Ottawa, Canada, 2007.
[4] S.P. Kundu, S. Chakraborty, A. Roy, B. Adhikari, S.B. Majumder, Chemically
modified jute fiber reinforced non-pressure (NP) concrete pipes with improved
mechanical properties, Constr. Build. Mater. 37 (2012) 841–850.
[5] F.F.M. Lopes, G.T. Araujo, J.W.B. Nascimento, T.S. Gadelha, V.R. Silva, Study of
the effects of acetylation treatments on sisal fiber, Revista Brasileira de
Engenharia Agrícola e Ambiental 14 (2010) 783–788 [in Portuguese].
[6] S. Altundal, A. Krasnikovs, O. Lininsh, I. Telnova, Transport and Engineering
Production Technologies – Mechanical Properties of Flax Fiber Based Polymer
(With Nano Additives) Matrix Composites, Riga Technical University, Building
Research Center, 2013.
[7] J.M. Ferraz, C.H. Del Menezzi, D.E. Texeira, S.A. Martins, Effects of treatment of
coir fiber and cement/fiber ration on properties of cement-bonded composites,
Bioresources 29 (2006) 495–501.
[8] B.N. Barra, S.F. Santos, P.V.A. Bergo, C. Alves Jr., K. Ghavami, H. Savastano Jr.,
Residual sisal fibers treated by methane cold plasma discharge for potential
application in cement based material, Ind. Crops Prod. 77 (2015) 691–702.
[9] H. Abou-Yousef, M. El-Sakhawy, S. Kamel, Multistage pulping by using
alkali/cairo’s acid treatment, Ind. Crops Prod. 21 (3) (2004) 337–341.
[10] G. Garrote, M.E. Eugenio, M.J. Díaz, J. Ariza, F. López, Hydrothermal and pulp
processing of eucalyptus, Bioresour. Technol. 88 (2003) 61–68.
[11] A. Ouensanga, C. Picard, Thermal degradation of sugar cane bagasse,
Thermochim. Acta 125 (1988) 89–97.
[12] K. Bilba, M.-A. Arsene, A. Ouensanga, Sugar cane bagasse fibre reinforced cement
composites. Part I. Influence of the botanical components of bagasse on the
setting of bagasse/cement composite, Cem. Concr. Compos. 25 (2003) 91–96.
[13] J. Claramunt, M. Ardanuy, J.A. Garcia-Hortal, Effect of drying and rewetting
cycles on the structure and physicochemical characteristics of softwood fibres
for reinforcement of cementitious composites, Carbohydr. Polym. 79 (2010)
200–205.
[14] J. Claramunt, M. Ardanuy, J.A. García-Hortal, R.D. Tolêdo Filho, The
hornification of vegetable fibers to improve the durability of cement mortar
composites, Cem. Concr. Compos. 33 (2011) 586–595.
[15] M.E.A. Fidelis, T.V.C. Pereira, O.F.M. Gomes, F.A. Silva, R.D. ToledoFilho, The
effect of fiber morphology on the tensile strength of natural fibers, J. Mater.
Res. Technol. 2 (2) (2013) 149–157.
[16] R.D. Toledo Filho, F.A. Silva, J.A. Melho Filho, E.M.R. Fairbairn, Durability of
compression molded sisal fiber reinforced mortar laminates, Constr. Build.
Mater. 23 (2009) 2409–2420.
[17] K.G. Satyanarayana, J.L. Guimarães, F. Wypych, Studies on lignocellulosic fibers
of Brazil. Part I: source, production, morphology, properties and applications,
Compos. A Appl. Sci. Manuf. 38 (2007) 1694–1709.
[18] P.S. Mukherjee, K.G. Satyanarayana, Structure and properties of some
vegetable fibers, J. Mater. Sci. 19 (1984) 3925–3934.
[19] F.A. Silva, N. Chawla, R.D. Toledo Filho, An experimental investigation of the
fatigue behavior of sisal fibers, Mater. Sci. Eng. A 516 (2009) 90–95.
[20] F.A. Silva, N. Chawla, R.D. Toledo Filho, Tensile behavior of high performance
natural (sisal) fibers, Compos. Sci. Technol. 68 (2008) 3438–3443.
[21] S.R. Ferreira, Effect of surface treatments on the structure, durability and bond
behavior of vegetable fibers for cementitious composites (Ph.D. thesis),
Department of Civil Engineering, COPPE/UFRJ, 2016.
[22] NBR 11578, Portland Composite Cement. Brazilian Society of Technical
Standards (ABNT), 1997 [in Portuguese].
[23] ASTM C1437-15, Standard Test Method for Flow of Hydraulic Cement Mortar,
ASTM International, West Conshohocken, PA, 2015.
[24] ASTM C39/C39M-15a, Standard Test Method for Compressive Strength of
Cylindrical Concrete Specimens, ASTM International, West Conshohocken, PA,
2015.
[25] C.M.O. Müller, J.B. Laurindo, F. Yamashita, Effect of cellulose fibers on the
crystallinity and mechanical properties of starch-based films at different
relative humidity values, Carbohydr. Polym. 77 (2) (2009) 293–299.
[26] H. Struszczyk, Modification of lignins III. Reaction of lignosulfonates with
chlorophosphazenes, J. Macromol. Sci. 23 (1986) 973–992.
[27] ASTM standard C1557, Standard Test Method for Tensile Strength and Young’s
Modulus of Fibers, ASTM International, West Conshohocken, PA, 2013.
[28] D.H. Page, P.A. Tydeman, Transverse swelling and shrinkage of softwood
tracheids, Nature 199 (4892) (1963) 471–472.
[29] C. Ackermann, L. Göttsching, H. Pakarinen, Recycled fiber and deinking, in: L.
Göttsching, H. Pakarinen (Eds.), Papermaking Science and Technology.
Recycled Fiber and Deinking (Book 7, pp. 359–438), Finland, Finnish Paper
Engineers’ Association, 2016.
[30] R.C. Howard, The effect of recycling on paper technology, Pap. Technol. 32 (4)
(1991) 20–25. ISSN 0031-1189.
[31] K. Murata, M. Masuda, Observation of microscopic swelling behavior of the cell
wall, J. Wood. Sci. 47 (2001) 507–509.
[32] R.J. Santos, P.R.L. Lima, Effect of treatment of sisal fiber on morphology,
mechanical properties and fiber-cement bond strength, Key Eng. Mater. 634
(2015) 410–420.
[33] P.R.L. Lima, R.J. Santos, S.R. Ferreira, R.D. Toledo Filho, Characterization and
treatment of sisal fiber residues for cement-based composite application, Eng.
Agríc. 34 (5) (2014) 812–825.
[34] A. Arbelaiz, B. Fernández, J.A. Ramos, I. Mondragon, Thermal and
crystallization studies of short flax fibre reinforced polypropylene matrix
composites: effect of treatments, Thermochim. Acta 440 (2006) 111–121.
 S.R. Ferreira et al. / Construction and Building Materials 139 (2017) 551–561
[35] H.-S. Kim, S. Kim, H.-J. Kim, H.-S. Yang, Thermal properties of bio-flour-filled
polyolefin composites with different compatibilizing agent type and content,
Thermochim. Acta 451 (2006) 181–188.
[36] H. Yang, R. Yan, H. Chen, C. Zheng, D.H. Lee, D.T. Liang, In-depth investigation
of biomass pyrolysis based on three major components: hemicellulose,
cellulose and lignin, Energy Fuels 20 (2006) 388–393.
[37] E. Gümüskaya, M. Usta, H. Kirei, The effects of various pulping conditions on
crystalline structure of cellulose in cotton linters, Polym. Degrad. Stab. 81
(2003) 559–564.
[38] M. Poletto, Heitor L. Ornaghi Júnior, Ademir J. Zattera, Native cellulose:
structure, characterization and thermal properties, Materials 7 (2014) 6105–
6119.
[39] A.C.H. Barreto, D.S. Rosa, P.B.A. Fechine, S.E. Mazzetto, Properties of sisal fibers
treated by alkali solution and their application into cardanol-based
biocomposites, Compos. A 42 (2011) 492–500.
[40] P.J. Herrera-Franco, A. Valadez-González, A study of the mechanical
properties of short natural-fiber reinforced composites, Compos. B 8 (2005)
597–608.
[41] H. Yang, R. Yan, H. Chen, D.H. Lee, C. Zheng, Characteristics of hemicellulose,
cellulose and lignin pyrolysis, Fuel 86 (12–13) (2007) 1781–1788.
[42] K.V. Sarkanen, Lignins: Occurrence, Formation, Structure and Reactions,
University of Washington, Seattle, Washington, 1987.
[43] P. Zugernmaier, Crystalline cellulose and derivatives: characterization and
structures, Springer Series in Wood Science, Springer, 2008.
561
[44] O. Derkacheva, Studying the structure of native, residual and dissolved lignins
by FTIR spectroscopy method, Proceedings of the First International Workshop
on Physicochemical Properties of Lignins (2005) 255–257.
[45] M. Poletto, A.J. Zattera, R.M.C. Santana, Structural differences between wood
species: evidence from chemical composition, FTIR spectroscopy, and
thermogravimetric analysis, J. Appl. Polym. Sci. 126 (2012) E336–E343.
[46] H.L. Ornaghi, M. Poletto, A.J. Zattera, S.C. Amico, Correlation of the thermal
stability and the decomposition kinetics of six different vegetal fibers,
Cellulose 21 (2014) 177–188.
[47] S. Dumitriu, Polysaccharides: Structural Diversity and Functional Versatility,
second ed., 2005. p. 999.
[48] D.H. Page, A note on the cell-wall structure of softwood tracheids, Wood Fiber
7 (1976) 246–248.
[49] T. Ishii, K. Shimizu, Chemistry of cell wall polysaccharides, in: D. N.-S. Hon, N.
Shiraishi (Eds.), Wood and Cellulosic Chemistry, Marcel Dekker Inc., New York,
2001, pp. 175–212.
[50] K.G. Johnsson, R.P. Overend, Lignin-carbohydrate complexes from Populus
deltides. I. Purification and characterization, Holzforschung 45 (1991) 469–
475.
[51] M. Suchy, E. Kontturi, T. Vuorinen, Impact of drying on wood ultrastructure:
similarities in cell wall alteration between native wood and isolated wood-
based fibers, Biomacromolecules 11 (2010) 2161–2168.
[52] I. Duchesne, E. Hult, U. Molin, G. Daniel, T. Iversen, H. Lennholm, The influence
of hemicellulose on fibril aggregation of kraft pulp fibres, Cellulose 8 (2001)
103–111.