Construction and Building Materials 101 (2015) 730–740

Contents lists available at ScienceDirect

Construction and Building Materials

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

Effect of fiber treatments on the sisal fiber properties and fiber–matrix

bond in cement based systems
Saulo Rocha Ferreira a, Flávio de Andrade Silva b,⇑, Paulo Roberto Lopes Lima c, Romildo Dias Toledo Filho a
a
Civil Engineering Department, 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

Significant improvements in the fiber–matrix interface were verified by the used treatments.

The polymer coating resulted in a better frictional mechanism and in a slip-hardening behavior.

The alkali treatment removed the amorphous constituents of the fiber and increased its crystallinity.

The hornification increased the elastic and frictional fiber–matrix bond.

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

Article history: This experimental research investigation aims to explain the influence of several treatments on the sisal
Received 23 March 2015 fiber properties and bonding strength with a cement based matrix free of calcium hydroxide. Four
Received in revised form 20 September different treatments were studied: hornification, alkali treatment with calcium hydroxide, polymer
2015
impregnation with styrene butadiene and a combination of hornification and polymer impregnation.
Accepted 18 October 2015
Modifications in the sisal fiber structure were investigated by X-ray diffraction (XRD), thermogravimetric
analysis (TGA), Fourier transform infra-red spectroscopy (FTIR) and scanning electron microscopy (SEM).
Water absorption and direct tensile tests were performed on the fibers to determine the influence of the
Keywords:
Natural fiber
treatments on their physical and mechanical properties. In order to study the sisal fiber–matrix bond,
Fiber treatments pullout tests were performed in fiber embedment lengths of 5, 10, 25 and 50 mm. All applied treatments
Pullout resulted in a reduction of the water absorption capacity and increase in tensile strength and stiffness.
Interface Significant improvements in the fiber–matrix interface were verified through the pullout test for the
several used treatments. The hornification treatment increased the elastic and frictional bond, whereas
the polymer and hybrid treatments resulted in a fiber slip-hardening behavior.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction have a low cost and present the great advantage of being
renewable.
In the line of development of materials reinforced with Sisal has been one of the most used fibers in the production of
vegetable fibers, many studies have been carried out on cementi- cement-based composites, especially due to its excellent tensile
tious and polymeric matrix composites and promising results have strength, around 400 MPa [6,7]. However, it has a weak chemical
been achieved, showing an improvement in durability, strength bond to cement based matrices, which according to the literature
and ductility [1–5]. There are many beneficial aspects in using varies between 0.32 and 0.72 MPa [8,9]. Due to chemical incom-
vegetable fibers including sustainability, and low energy consump- patibility, the chemical bonding between fiber and matrix is
tion. Vegetable fiber reinforced composites have the potentiality to mainly due to Van der Waals force and –OH interactions from
become the sustainable option for the construction industry. Due cellulose and calcium hydroxide. In addition to a low chemical
to its wide availability, especially in tropical countries, these fibers bond, the high water absorption capacity of sisal fibers causes
volume expansion when fibers are added to the fresh cementitious
⇑ Corresponding author. matrix and contraction when the matrix dries, resulting in a partial
E-mail address: [email protected] (F.A. Silva).
loss of physical contact with the matrix.

http://dx.doi.org/10.1016/j.conbuildmat.2015.10.120
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740 731

Different procedures to reduce the water absorption capacity of information on the sisal fibers microstructure can be found in the
natural fibers and to improve the fiber–matrix bond have been author’s previous works [2,3,7].
pursued by applying chemical and physical treatments, both in A matrix with a mix design of 1:0.5:0.4 (binder:sand:water/
the matrix and in the fiber [5,9–14]. The partial replacement of binder ratio) by weight was used. The binder was composed by a
cement by microsilica resulted in an increase of the pullout resis- combination of 30% of Portland cement CP-32 F II, 30% of
tance by about 24% [9]. This increase is related to the fineness of metakaolin and 40% of fly ash. This ratio of metakaolin and fly
the microsilica which reduced the porosity of the fiber matrix ash guarantee the durability of the fiber once a matrix free of
interface, improving its bond. The use of alkaline solutions calcium hydroxide is obtained, as shown in previous works
[10,15,16] removes most of the surface non-cellulosic substances [1,22]. The fly ash also ensures greater workability to the matrix
and increase the roughness of the fiber surface increasing the that, in the context of high-performance composites, is a desirable
fiber–matrix bond. Simple treatments such as soaking the fibers property as it ensures a better homogenization of the vegetable
in distilled water followed by a drying process also resulted in fibers [23]. The used fly ash has an amount of 74% of CaO, 13.6%
improvements in the fiber–matrix bond [17]. of SiO2, and 4% of SO3 and Al2O3. The used sand was ground to
A reduction in the fiber hydrophilicity can be achieved by obtain a maximum diameter of 840 lm. The granulometric compo-
rewetting and drying cycles to promote hornification which is sition was performed according to the Brazilian standard [24]. The
the stiffening of the polymeric structure present in lignocellulosic used superplasticizer was the Glenium 51 (type PA) with solids
materials [11]. This treatment promoted a reduction in volumetric content of 31%. A viscosity modifier Rheomac UW 410, (manufac-
changes of pulps and fibers of natural origin strengthening their tured by BASF), at a dosage of 0.8 kg/m3 was also used in order
interfacial bond. Polymer coating treatments reduce the water to avoid segregation and bleeding during molding. The matrix used
absorption and are effective in the sealing process of natural fibers in this study showed a flow table spread value of 450 mm accord-
[18]. Styrene butadiene polymers are used to improve the mechan- ing to the Brazilian standard NBR 13276/05 [25] and compressive
ical and physical properties in cement pastes as a result of the strength at 28 days of 31 MPa, according to NBR 7215/96 [26].
chemical reactions with calcium hydroxide and silicates [19,20]. The mixture were produced in a room with controlled temper-
Therefore, the simultaneous use of the two treatments may result ature (21 ± 1 °C) using a mixer with capacity of 5 L. The mixing
in a better fiber–matrix interface based in the cross-linking formed procedure is described as follow. All dry components were homog-
between the polymer coating and the cement matrix. enized in the mixer. The water and superplasticizer were added
However, there still exists a lack of information on the use of and mixed for 2 min at a speed of 125 RPM. After that, the process
such treatments to improve the interface of natural vegetable was stopped during 30 s to remove the material retained in the
fibers with a Portland cement based matrix. Therefore, the mixer. Then, the mixing procedure continued for 2 min at 220
objective of the present research is to investigate the effect of RPM and, finally for a further 5 min at 450 RPM.
different treatments on the sisal fiber physical and mechanical To mold the pullout specimens a special mold was developed
properties and fiber–matrix interface behavior with a cement (see Fig. 1(a)). After filling the mold with the matrix, the top cap
based matrix. 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 (Fig. 1(b)). Embedment lengths of 5,
2. Experimental program 10, 25 and 50 mm were studied. After 24 h, the specimens were
demolded and placed in a fog room (HR% P 95%) to moist curing
2.1. Materials and processing for 7 days for the pullout test.

The sisal fibers used in the present study were obtained from 2.2. Sisal fiber treatments
the sisal plant cultivated in farms located in the Bahia state, Brazil.
They were extracted from the sisal plant leaves in the form of long The methodology used to apply the several treatments used in
fiber bundles. The fiber extraction from the leaf was done by semi- the present paper as described as follow:
automatic scrapers. These fibers characterized mechanically by
Silva et al. [6,7,21] presents a fiber microstructure formed by (i) Hornification: The sisal fibers were placed in a container with
numerous individual fibers (fiber-cells) which varies from 6 to water (T = 22 °C) during three hours to reach its maximum
30 lm in diameter. The individual fiber-cells are linked together water absorption capacity. The drying process was carried
by means of the middle lamella. The chemical composition of out in a furnace at a temperature of 80 °C. The furnace used
the sisal fiber includes approximately 60.5% cellulose, 25.7% is equipped with an electronic temperature control and
hemicellulose, 12.1% lignin, 1% pectin and 1.6% ash [18]. More connected to a scale, with a precision of 0.01 g to record

Upper acrylic plate

a b c ∅ = 3/4"
5 mm

25 mm Mold

40 mm

Lower acrylic plate

Fig. 1. Casting procedure of specimens for the pullout test: (a) positioning the fiber in the mold, (b) filling the mold with the matrix and (c) detail of the mold.
732 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740

the moisture loss. The furnace was programmed to reach 2.3.4. Water absorption capacity
80 °C at a heating rate of 1 °C/min and to maintain this For the water absorption test a bundle of fibers weighing
temperature for 16 h. After 16 h of drying, the furnace was 200 g in the dry state was used. The specimens were dried under
cooled down to the temperature of 22 °C (at a cooling rate laboratory condition (23 °C and 56% RH until reach constant
of 0.23 °C/min) in order to avoid possible thermal shock to mass). The fiber bundle was then immersed in distilled water
the fibers. This procedure was repeated ten times. According to reach its absorption capacity. At the immersion time of 10,
to Ferreira et al. [14] this treatment is effective to promote a 30, 60, 120, 180, 240, 300, 360, 420, 480 and 1440 min the fibers
modification in fiber microstructure, resulting in a dimen- were removed, patted dry with lint free cloth, and weighed. For
sional stability. More details can be obtained elsewhere [14]. the drying process a furnace at a constant temperature of 80 °C
(ii) Alkali treatment: The sisal fibers were immersed in 0.73% was used. The used system is equipped with an electronic tem-
weight of Ca(OH)2 solution at laboratory controlled temper- perature control and connected to a scale, with precision of
ature (22 °C) for 50 min. According to Saha et al. [15], 0.01 g. The furnace, model FGT11000 (Fornos INTI, Brazil) was
immersion periods ranging from 30 to 60 min, in low alkali programmed to reach 80 °C at a heating rate of 1 °C/min and
concentrations (0.5–1%) does not degrade the natural fiber. to maintain this temperature for 16 h. The used furnace has no
The alkali treated fibers were not washed to keep the excess humidity control system. After that time the furnace was cooled
alkali on the fiber surface. The fibers were then air dried, in a to the room temperature of 23 °C ± 1 °C in order to avoid possi-
forced air flow chamber at 40 °C for 24 h. ble thermal shock to the fibers. The dry weight used was
(iii) Polymer impregnation: The polymer treatment used in the obtained by the last data recorded by the weight scale after each
present study consisted in an impregnation by immersing treatment. Water absorption is expressed as increase in weight
the fibers in an aqueous solution with a styrene butadiene percent:
polymer. The used polymer was liquid with a solid content
of 48.5 ± 1% of weight. The density, viscosity and pH were
saturated weight  dry weight
0.998 g/cm3, 400–850 mPas and 6.5 ± 1.5, respectively. The Water absorption ð%Þ ¼  100
dry weight
sisal fibers were placed in a container with the polymer
solution for 50 min. The fibers were then dried in a forced ð2Þ
air flow chamber during 24 h at 40 ± 1 °C and wind velocity
of 0.5 m/s. 2.3.5. Direct tensile tests
(iv) Hybrid treatment: The applied treatment consisted of a The tensile tests were performed in an electromechanical test-
combination of hornification and polymer impregnation. ing machine Shimadzu AG-X with a load cell of 1 kN. The tests
After 10 cycles of wetting and drying the fibers were were performed in 15 sisal fibers using a displacement rate of
immersed in the polymer solution, using the same proce- 0.1 mm/min. The fibers with a gage length of 50 mm were glued
dure described before. to a paper template for better alignment in the machine and for
a better griping with the upper and lower jaws in accordance with
2.3. Characterization testing ASTM C1557 [28]. To calculate the tensile strength of the fibers,
their area were measured by image analysis from micrographs
2.3.1. Thermogravimetric analysis (TGA) from a scanning electron microscope.
TGA was performed on natural and treated fibers. The tests
were performed in a SDT Q600 simultaneous TGA/DTA/DSC from
TA Instruments. Scissor cut samples weighing 10 mg were sub- 2.3.6. Single fiber pullout tests
jected to a heating rate of 10 °C/min until reaching 1000 °C in a The pullout tests were performed in an electromechanical test-
open platinum pan using 100 ml/min of nitrogen as the purge gas. 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
2.3.2. X-ray diffraction method (XRD) samples were fixed in the machine grips through a system with
The XRD measurements were performed on a Bruker D8 Focus hinged-fixed boundary conditions as shown in Fig. 2. Ten tests
(CuKa radiation = 1.54056 Å). The diffracted intensity of radiation were performed for each embedment length (5, 10, 25 and
was measured in a 2h range between 10° and 40° with a step size 50 mm) to understand the influence of the embedment length on
of 0.05° at 3°/min. For sample preparation the fibers were scissor the pullout behavior. For evaluating the efficiency of the several
cut to particle sizes of less than 0.5 mm (digitally measured from studied treatments, it was chosen the fiber embedment length of
optical microscopy by imageJ software). The test was performed 25 mm.
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 2.3.7. Scanning electron microscopy (SEM)
and the amorphous fraction, according to the method developed by The sisal fiber’s microstructure was investigated using an
Segal [27]: SEM Hitachi TM3000. The microscope was operated under an
accelerating voltage of 15 kV. A working distance of 3 mm was
Icr  Iam
Crl ð%Þ ¼  100 ð1Þ applied and tilt was set to 0°. A pre-coating with a thin layer
Icr of approximately 20 nm of gold was done to make the fiber con-
where, Icr is the intensity (peak height) of the crystalline peak at the ductive and suitable for analysis. In order to measure the fiber’s
maximum value (2h between 22° and 23°) and Iam is the intensity cross-sectional area, for each single fiber used in the pullout and
at the minimum value (2h between 18° and 19°). tensile test, an adjacent piece of the fiber (immediately next to
the one tested) was kept for future measurement and morphol-
2.3.3. Fourier transform infrared (FTIR) spectroscopy ogy characterization using the SEM. This procedure has been
FT-IR analyses were performed in a Nicolet iN10 Spectrometer established in the author’s previous work [7]. Fiber–matrix
(Thermo Scientific, U.S.A), using a MCT-B detector and a diffuse interface zone was also investigated. The obtained images were
reflectance probe, with resolution of 4 cm1 in the range of post-processed using ImageJ, a Java-based image processing
4000–675 cm1. Stored results were averages of 24 scans. program.
 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740 733

(a) (b)
Load cell

Hinged-fixed

(c)

Fig. 2. The pullout testing system: (a) general view of the machine, (b) the pullout test setup and (c) detailed sketch of the pullout setup.

3. Results and discussion 3.2. X-ray diffraction analysis

3.1. Thermogravimetric analysis Fig. 4 presents the diffractogram for natural and treated sisal
fibers. The graph shows peaks representing the crystal planes
Fig. 3 shows the thermogravimetric analysis (TGA) performed in characteristic of lignocellulosic materials: 15°, 23° and 34°. Lignin
the natural and treated sisal fibers. It is possible to see three inter- and hemicellulose are amorphous macromolecules and cellulose
vals of decomposition, where the first stage occurs between 0 and molecules (a crystalline macromolecule) are randomly distributed,
100 °C, due to the water evaporation. The second peak occurs with a tendency to form intra- and intermolecular hydrogen
around 250 and 350 °C. It is related with the simultaneous decom- bonds. Through a peak fit performed using the ICDD database
position of cellulose, hemicellulose (denoted by a shoulder in the (PDF 4 – organics) it was found that the cellulose of sisal (2h  23°)
derivative thermogram at 300 °C) and pectin [29]. The third peak is of the type I beta with a PDF number of 00-056-1718. X-ray
occurs in intervals from 350 to 400 °C and it is characteristic of Diffraction reveals changes in the crystalline polymorphs after
cellulose degradation. To the fibers treated with polymer there is the treatments (see Table 1). As shown in Fig. 4, a clear reduction
another peak (350–500 °C) ascribed to the degradation of the in the crystallinity degree occurred after the polymer treatment.
polymer coating. Conversely, the effect seems to be more accentuated for sisal fiber
As showed per Claramunt et al. [11], the hornified fibers show hybrid treated fibers. This decrease may be a polymer reaction that
results that illustrate an initial thermal stability behavior after
treatment. This increase in the temperature of thermal decomposi-
tion after treatment is related to the partial removal of hemicellu-
lose and lignin from the fibers and to the higher crystallinity of the
cellulose. Thermal analysis of the polymer shows a simultaneous
decomposition of butadiene and styrene, beginning at around
310 °C and ending at 470 °C, where we can observe the residual
mass of 3.0%.

Fig. 3. Thermogravimetry analysis of the sisal fibers before and after the different
treatments. Fig. 4. X-ray diffractograms of the sisal fibers before and after treatments.
734 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740

Table 1
Crystallinity index of natural and treated sisal fibers.

Treatment Iam (2h = 18.5°) Icr (2h = 22.4°) Crystallinity index (%)
Natural 27 135 80
Hornified 19 155 87.74
Alkali treatment 15 126 88.10
Polymer Coated 13 86 84.88
Hybrid 14 72 80.56

Iam = intensity of amorphous/Icr = Intensity of crystalline fraction.

coat the cellulose chains on the surface and reduce the capacity of
refraction.
Interestingly, in the hornified and alkali treatments, an increase
in the intensity of the peak (600) around 2h = 23° is observed,
indicating an increase in the cellulose crystallites. According to
Ref. [30], it was observed that when sisal fiber is subjected to an
alkali solution, the cellulose relative peaks intensity and crystalline
fraction increased due to the lignin being partially removed in the
alkaline solution, i.e., the sisal fiber lost part of its amorphous
component.

Fig. 5. Infrared spectroscopy of sisal fiber before and after the different treatments.
3.3. Fourier transform infra-red spectroscopy analysis

The FTIR spectra of sisal fibers are show in Fig. 5. The spectra of bond of the carboxylic groups of ferulic and p-coumaric acids of
the natural and treated fibers show the vibration modes in agree- lignin and/or hemicellulose. In the same spectrum the peak at
ment with the literature [30] and [31], characteristics of lignocellu- 1644 cm1 is associated with the double bond C@C vibration of
losic fibers. Sisal natural fibers have cellulose, hemicellulose and the aromatic molecule lignin, and the wavelength at 1244 cm1
lignin as the most important components. These three important corresponds to the axial asymmetry of link = COC in ether, ester
elements (cellulose, hemicellulose and lignin) have peaks in the and phenol groups present. The peak at 1044 cm1 is assigned to
largest region between 800 and 2000 cm1. Mostly of the peaks the stretching of the CAO and vibration of the CAH cellulose group.
are relative to lignin, correlated with Methoxyl groups (AOCH3), Regarding to the hornified fiber, an increase in the intensity of
CAOAC and C@C aromatic ring groups [32,33]. the peak at 3400 cm1 is observed. It is an evidence of the hydro-
The prominent peak at 1742 cm1 in the spectrum is attributed gen bridges hypothesis. According to [34], the hydroxyl groups
to the acetyl groups and hemicelluloses uronic ester or the ester (AOH) in the cellulose, hemicelluloses and lignin build a large

Fig. 6. Schematic illustration of the swelling process in a cell wall (based on Ref. [37]).
 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740 735

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, as
represented by the sketch in Fig. 6. It is needed a high energy to
brake those bonds, therefore the cycles applied to the fiber increase
the number of hydrogen bonds, structural changes in the fibers
wall and an increase in the number of AOH bonds, as can be seen
in the FTIR analysis. It is observed also 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
[35]. This mechanism promotes a better interaction between lignin
and cellulose and hemicellulose, resulting in a stronger material
and in a higher crystallinity, stiffness and strain capacity. Studies
have also suggested that covalent linkages between lignin and
hemicelluloses exist in native wood [36]. The lignin is covalently
bound to the hemicellulose which, in turn, is bound to cellulose
through extensive hydrogen bonding.
Alkali hydrolysis is able to produce reactive degradation. The
dissolution of the lignin phase is clearly observed due the absence
of their characteristic peak at 1727 cm1 and 1245 cm1. The Fig. 7. Mass gain of sisal fibers with different treatments at different times of
immersion.
hydroxyl groups, provenient from cellulose (1630 cm1) was
removed after the Ca(OH)2 treatment due the partial dissolution
of hemicellulose and lignin. After the alkali treatment, the bands alkali, polymer and hybrid treated fibers, respectively. The reduc-
at 1730 and 1245 cm1, attributed to the stretching vibrations of tion in the water absorption capacity can be correlated with the
C@O and CAO groups, respectively, disappeared; these results small loss of hemicellulose and lignin, presented by TG/DTG
were presented also by Ref. [37]. This phenomenon is an indication analysis as well as by changes in the chemical bonds that makes
of the removal of carbonyl and hydroxyl groups. The peak at the fibers more hydrophilic.
1249 cm1 corresponds to C@O stretching in the acetyl groups in The reduction in the water absorption capacity for the hornified
lignin [38]. The characteristic peak at 1430 cm1 is for ACH2 fibers can be explained as follows: according to [23] after several
bending of cellulose which remained unaffected after alkali cycles of wetting and drying, there is an increase packing of the
treatment [39]. The unaffected cellulose component of fiber may fiber-cells and a reduction of the lumen after submission to the
increase the fiber tensile strength [40]. All the groups peaks from treatment, which consequently reduces the capacity of water
lignin and hemicellulose were broaded or dissolved, resulted by absorption of the fiber. In the case of the alkali treatment, the
the removal of those components. mechanism is related to the removal of lignin and hemicelluloses,
According to Ref. [41] the major modification done by this which with higher contents absorb more moisture [45,46].
treatment is the disruption of hydrogen bonding in the network The use of polymer promotes a coating layer that decreases sig-
structure, thereby increasing surface roughness. It is possible to nificantly the water absorption capacity. Once combined with the
observe a broad of the hydroxyl band due to changes of the hornification process, it becomes an effective way to promote a
inter- and intra-molecular hydrogen bonding in polysaccharides high hydrophobicity behavior. It is thought that the used polymer
[42,43]. penetrates the fibers, specifically between the fiber cells and filling
As shown in Fig. 5, the absorption band at 3559 cm1 is the lumens, mitigating the water absorption.
assigned to be due to OAH stretching. The mode 2921 cm1 is
assigned to be due to CAH stretching vibration. The absorption
band at 1739 cm1 is absent from the natural sisal fiber and 3.5. Fiber’s tensile behavior
appears only in the polymer treated sisal fiber. The mode is
assigned to be due to the C@O stretching of ester linkage [44]. Typical stress–strain curves of untreated and treated sisal fibers
The appearance of this band is an indicative that an interaction are shown in Fig. 8. Results of individual treatments values and
between the polymer and the fibers are taking place. averages are presented in Table 2. The tensile strength of sisal fiber
increased after all treatments. For alkali treatments the tensile
3.4. Water absorption capacity strength of sisal fiber was improved by 32%. The same result was
found by Ref. [15]. According to Ref. [16], the reason of the incre-
Fig. 7 shows the results of the water absorption capacity of sisal ment in the tensile strength and elongation is the removal of
fibers for the different treatments used in the present work as a non-cellulosic materials and impurities, which remain dispensed
function of time. It is possible to observe a major increase of the in interfibrillar region of the fibers. After a low concentration alkali
sisal fiber water absorption after one hour. Under its natural dry treatment, the fibrils become more capable for rearranging
state the sisal fiber can absorb about 200% of its own weight. The themselves along the directions of tensile deformation since the
hemicelluloses are responsible for absorbing water, but crystalline interfibrillar region occurs to be less dense and less rigid due to
cellulose and lignin also play an important role in this process. the removal of hemicelluloses.
Moisture swells the cell wall of lignocellulosic fiber until its satu- The hornified treatment resulted in an increase of tensile
ration with water. Then the water occupies the free spaces strength and strain at failure by about 5% and 39%, respectively.
between the fibrils. All the applied treatments resulted in a Cycles of wetting and drying change the microstructure of
decrease in the water absorption capacity. It can be seen that the vegetable fibers, which modifies the polymeric structure of the
reduction rate was about 15%, 17.5%, 25% and 50% for the hornified, fiber-cells resulting in higher tensile strength and strain.
736 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740

Table 2
Average results of fifteen tensile test performed on natural and treated sisal fibers.

L Treatment Max. Tensile Strain Young’s

(mm) load strength capacity modulus
(N) (MPa) (mm/mm) (GPa)
50 Natural 10.28 453.23 0.03 (0.01) 20.16 (3.52)
(3.43) (66.49)
Hornified 9.82 474.86 0.05 (0.01) 18.05 (2.75)
(2.22) (26.46)
Alkali 15.36 708.70 0.03 (0.01) 22.66 (3.97)
treatment (2.32) (103.29)
Polymer 15.66 722.07 0.03 (0.01) 23.98 (4.58)
Coated (2.61) (98.27)
Hybrid 11.38 727.46 0.05 (0.02) 25.04 (5.02)
(3.11) (206.32)

The values under parentheses are standard deviation.

the fiber structure rigid and dense. Probably that makes the struc-
ture more bundled, not permitting accommodation of fiber cells
during the initiation of tensile test, thus justifying the increase in
fiber stiffness.
The hybrid treatment showed almost the same increase when
Fig. 8. Influence of fiber treatments on the sisal fiber tensile behavior.
compared to the polymer treatment, around 60%. It is possible to
assume that there is no significant difference on tensile strength,
strain capacity and stiffness for both treatments.
The treatment with polymer resulted in an increase of 50% of
the tensile strength. This increase was also observed in stiffness,
which increased by 15%. During the polymer treatment, the fiber 3.6. Pullout tests
is immersed in-aqueous solution of styrene butadiene. During this
period, the fiber absorbs the solution, allowing the penetration of A detailed analysis of pullout load vs. fiber embedded length
the polymer between the fiber cells and the internal walls of the curves has been presented in Figs. 9 and 10 (see results in
lumen. After drying, the polymer becomes more rigid, making Tables 3 and 4). A typical pullout curve has four distinct regions,

L = 5mm L = 10 mm

L = 25 mm L = 50 mm

Fig. 9. Influence of different embedment length on the untreated sisal fiber pullout behavior. Ten samples have been tested for each condition.
 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740 737

Fig. 10. Idealized pullout behavior of a sisal fiber from a Portland cement matrix (a); typical curves for embedment lengths of 10, 25 and 50 mm and (b) sketch showing the
debonding process (L = embedment length; Pad = maximum adhesional load; Pfr = frictional load).

as can be seen in Fig. 10: The region I corresponds to the elastic- the post-peak region is governed by the frictional shear strength
linear range with a rapid rate of ascent of the load. That first of the interface and continues till the fiber is completely
region is almost similar for all studied embedment lengths. As debonded from the interface. The shear strength at this region
the load increases, beyond the linear region, a degree of nonlin- is defined as frictional resistant strength (rfr).
earity is observed and this range is designated as region II which The variation of embedment lengths can result in different pull-
defines the initial point of fiber debonding. The peak response is out mechanisms. Fig. 9 show the pullout curves of untreated sisal
reached at region III under partial debonding conditions, where fibers in a cement based matrix with an embedment length of 5,
the pullout force reaches a maximum value (Pad). The slip (s) of 10, 25 and 50 mm. For the fibers with 5 and 10 mm of embedment
the fiber at this point is defined as the slip at the peak and can length an ideal friction behavior is observed. For the 25 mm fiber
be considered as the critical debonded length. For a given matrix embedment length (Fig. 9b) the curve shows a slip softening pro-
and interface condition the peak pullout force depends on the cess. After reaching the maximum adhesional load, the fiber slips
embedded length, diameter of the fiber and the curing period. from the matrix with a constant load. When the embedment length
The nominal shear strength computed at Pad is defined as the is higher than 50 mm the sisal fiber fracture after the elastic linear
adhesional strength (rad). In region IV the load drops to a fixed range (zone I).
value and remains constant thereafter. The pullout behavior in Fig. 11 shows a comparison of the pullout test results for sisal
fibers (embedment length of 25 mm) with the different surface
treatments investigated in the present work. It can be observed
Table 3
that all treatments resulted in improvements of the bond between
Average results of ten pullout tests performed on natural sisal fiber with different
embedment length.
the sisal fiber and the Portland cement matrix. Those improve-
ments are related to increases in stiffness, adhesional and frictional
L Pad (N) Dad rad Pfr (N) Dfr rfr Stiffness
bonds.
(mm) (mm) (MPa) (mm) (MPa) (N/mm)
The hornified sisal showed a higher adhesion bond by the stiff-
5 1.62 0.26 0.37 1.12 0.63 0.41 8.69 (6.78) ening of the polymeric structure of the fiber-cells. That process
(0.49) (0.14) (0.28) (0.36) (0.26) (0.25)
10 2.87 0.43 0.26 2.29 0.86 0.28 60.00
result in a packing of the fibercells due the formation of
(0.48) (0.23) (0.03) (0.71) (0.14) (0.05) (3.46) hydrogen-bonds. A direct consequence is the reduction of the
25 3.73 0.66 0.30 2.56 0.92 0.18 9.31 (4.39) lumen, which consequently reduces the capacity of water absorp-
(1.02) (0.19) (0.08) (0.72) (0.25) (0.04) tion and dimensional change (refer to Fig. 6).
50 6.35 0.97 0.32 – – – 33.83
The alkali treatment resulted in slightly higher crystallinity
(1.15) (0.46) (0.09) (16.04)
indexes of the fibers (see Table 1). It can improve the pullout
The values under parentheses are standard deviation. behavior but the main mechanism was the chemical interaction

Table 4
Average results of ten pullout tests performed on natural and treated sisal fibers.

L (mm) Treatment Pad (N) Dad (mm) sad (MPa) Pfr (N) Dfr (mm) sfr (MPa) Stiffness (N/mm)
25 Natural 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)
Hornified 4.53 (1.07) 0.26 (0.12) 0.42 (0.08) 3.88 (0.77) 0.37 (0.13) 0.27 (0.03) 20.60 (6.37)
Alkali treatment 4.84 (0.42) 0.45 (0.15) 0.39 (0.03) 4.31 (0.76) 1.08 (0.32) 0.32 (0,06) 27.34 (7.37)
Polymer Coated 5.77 (1.34) 0.75 (0.37) 0.49 (0.12) 4.16 (0.64) 1.13 (0.21) 0.34 (0.04) 31.53 (5.24)
Hybrid 10.57 (1.72) 0.85 (0.32) 0.86 (0.14) 4.37 (0.65) 1.18 (0.12) 0.31 (0.13) 34.82 (4.21)

The values under parentheses are standard deviation.

Pad: adhesional load; Dad: displacement correspondent to the adhesional load; sad: adhesional shear stress; Pfr: frictional load; Dfr: displacement correspondent to the
frictional load and sfr: frictional shear stress.
738 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740

Fig. 11. Influence of fiber surface treatments on the sisal fiber–matrix pullout behavior: (a) pull-out slip curves (L = 25 mm) and (b) detailed micrograph of fibrillation and
peeling off surface due a high mechanical anchoring mechanism in a polymer treated sisal fiber.

between the remaining calcium hydroxide and also the calcium between the fiber and the matrix with the consequence of an
carbonate attached on the surface of sisal fibers. On the other hand, ascending post-peak in the pull-out curve.
the reaction of a low concentrated alkaline solution can clear and The treatment with polymers after the hornification of the
increase the roughness of sisal fibers. The result after the treatment fibers (hybrid treatment), lead to a greater adhesion bond which
is a higher frictional pullout, leaded by the abrasion and grinding resulted in a fiber fracture for the studied embedment length. This
resulting in defibrillation (see Fig. 12c). result is due to the dimensional stability promoted by hornification
The treatment with polymer shows an improvement in pullout and the chemical interaction between the polymer coating and the
behavior. It is possible to assume that the polymer acts as a bridge Portland cement matrix.
between fiber and matrix, strengthening the interfacial bonding. It is postulated that the polymer coating acts as a bridge
Indeed in the SEM micrograph, it was observed a formation of a between fiber and cement matrices to strengthen the interfacial
SBR polymer film on the fiber surface. Polymer anchor points were bonding between them. In the SEM micrograph of the polymer
also identified, possible being an agent that leads to the improve- modified sisal fiber (Figs. 11b, 12b and 13a), a formation of a
ment of the adhesion, resulting in a slip-hardening on pullout polymer film between the fiber and cement matrix can be easily
(see Figs. 11b and 13a). identified. The strong interfacial bonding between the reinforcing
The highest improvements were found in particular for the agent and cement matrix leads to the improvement in the pullout
treatment with polymer and hybrid treatments. The addition of load. The formation of such interfacial bond can be understood by
styrene butadiene copolymer in the fibers promoted a slip- following the plausible reaction sequence presented schematically
hardening behavior during the pullout test. This type of in Fig. 13b as described by Ref. [16]. The main constituent of a sisal
behavior can be explained from microscopical observations fiber is cellulose which contains a large number of intra and
(see Figs. 11b and 12b). The pullout of the fiber from the matrix inter-molecular bonded hydroxyl groups. During polymer modifi-
results in damage to the fiber showing some fibrillation and peel- cation, the cellulose eventually reacts with the carboxylic acid
ing off the surface leading to a mechanical anchoring mechanism. group of the polymer latex (carboxylated styrene butadiene
This behavior was described elsewhere [47], which says that fibers copolymer) to form a linkage. Another part of this co-polymer
with low modulus and swelling issues (e.g. nylon) presents a fiber forms bond with calcium ion of the hydrated cement. Thus the
damage at pullout test, showing fibrillation and peeling on surface, polymer forms a bridge between the fiber surface and cement
being a result of a good mechanical and chemical anchoring matrix.

Natural sisal Hybrid Alkali treatment Hornified sisal

(a) (b) (c) (d)

200 µm

Polymer coating

Fig. 12. Micrograph showing the sisal fiber surface after pullout test for the different used treatments.
 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740 739

(a) (b)

Cement
Matrix

200µm
Matrix Fiber Polymer coating Matrix Fiber SBR
(Cellulose) Polymer

Fig. 13. (a) Micrograph showing the fiber matrix interface of polymer modified sisal fiber and (b) a schematic model used to explain the fiber–polymer–concrete interfacial
bonding.

4. Conclusions [2] F.A. Silva, B. Mobasher, R.D. Toledo Filho, Cracking mechanisms in durable
sisalreinforced cement composites, Cement Concr. Compos. 31 (2009)
721–730.
The work in hand investigated the effect of surface treatments [3] F.A. Silva, R.D. Toledo Filho, F.A. Melo Filho, E.M.R. Fairbairn, Physical and
on sisal fibers chemical, mechanical and fiber–matrix interface mechanical properties of durable sisal fiber cement composites, Constr. Build.
properties. The following conclusions can be drawn from the pre- Mater. 24 (2010) 777–785.
[4] A.N. Netravali, S. Chabba, Composites get greener, Mater. Today 6 (2003)
sent research:
22–29.
[5] P.A. Sreekumar, S.P. Thomas, J.M. Saiter, K. Joseph, G. Unnikrishnan, T. Sabu,

After the surface treatments the sisal fibers showed some Effect of fiber surface modification on the mechanical and water absorption
reduction on their chemical compounds, such as hemicelluloses characteristics of sisal/polyester composites fabricated by resin transfer
molding, Compos. A Appl. Sci. Manuf. 40 (2009) 1777–1784.
and lignin. This reduction was probably not enough to affect the [6] F.A. Silva, N. Chawla, R.D. Toledo Filho, Mechanical behavior of natural sisal
mechanical behavior but enough to increase the crystallinity of fibers, J. Biobased Mater. Bioenergy 4 (2010) 106–113.
cellulose, changing its chemical bonds, especially the acetyl [7] F.A. Silva, N. Chawla, R.D. Toledo Filho, Tensile behavior of high performance
natural (sisal) fibers, Compos. Sci. Technol. 68 (2008) 3438–3443.
groups, and promoting a reduction on the fiber water absorp- [8] F.A. Silva, B. Mobasher, C. Soranakom, R.D. Toledo Filho, Effect of fiber
tion capacity. shape and morphology on interfacial bond and cracking behaviors of sisal

The fiber tensile strength increased with the application of all fiber cement based composites, Cement Concr. Compos. 33 (2011)
814–823.
treatments. This can be explained by the change in the cellulose [9] R.D. Toledo Filho, Materiais compósitos reforçados com fibras naturais:
crystallinity, reduction of amorphous material and the stiffen- caracterização experimental (Doctoral thesis), Civil Engineering Department,
ing provided due the polymer interaction with the sisal fiber PUC-Rio, Rio de Janeiro, Brasil, 1997 (in portuguese).
[10] C.A. Angrizani, C.A.B. Vieira, A.J. Zattera, E. Freire, R.M.C. Santana, S.C. Amico,
and to the hornification phenomena.
Influência do comprimento da fibra de sisal e do seu tratamento químico

As compared to the natural sisal fiber, significant improvement nas propriedades de compósitos com poliéster, 17° CBECIMat, Congresso
in the fiber–matrix interface was verified through the pullout Brasileiro de Engenharia e Ciência dos Materiais, Foz do Iguaçu, PR, Brasil,
tests. With the treatments it was possible to reach a higher fiber 2006.
[11] J. Claramunt, M. Ardanuy, J.A. Garcia-Hortal, Effect of drying and rewetting
matrix bond, showing not only a better frictional mechanism cycles on the structure and physicochemical characteristics of softwood fibres
but also a slip-hardening behavior, in the case of the polymer for reinforcement of cementitious composites, Carbohydr. Polym. 79 (2010)
treatment. The alkali treatment modified the fiber composition 200–205.
[12] F.F.M. Lopes, G.T. Araujo, J.W.B. Nascimento, T.S. Gadelha, V.R. Silva, Estudo dos
by removing the amorphous constituents of the fiber and
efeitos da acetilação em fibras de sisal, Rev. Bras. Eng. Agríc. Ambient. 14
thereby increasing its crystallinity. The polymer coating acted (2010) 783–788.
as a bridge between fiber and cement matrices to strengthen [13] J.M. Ferraz, C.H.S. Menezzi, D.E. Texeira, S.A. Martins, Effects of treatment of
the interfacial bonding between them. coir fiber and cement/fiber ration on properties of cement-bonded composites,
BioResources 3 (2011) 3481–3492.
[14] 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, Rev. Mater. 17
(2012) 1024–1034.
Acknowledgements
[15] P. Saha, S. Manna, S.R. Chowdhury, R. Sen, D. Roy, B. Adhikari, Enhancement of
tensile strength of lignocellulosic jute fibres by alkali-steam treatment,
The authors thank the National Council for Scientific and Tech- Bioresour. Technol. 101 (2010) 3182–3187.
nological Development (CNPq, Brazilian National Science Agency) [16] S.P. Kundu, S. Chakraborty, A. Roy, B. Adhikari, S.B. Majumber, Chemically
for their support. modified jute fibre reinforced non-pressure (NP) concrete pipes with
improved mechanical properties, Constr. Build. Mater. 37 (2012) 841–850.
[17] Y. Li, C. Hu, Y. YU, Interfacial studies of sisal fiber reinforced high density
References polyethylene (HDPE) composites, Compos. A Appl. Sci. Manuf. 4 (2008)
570–578.
[1] J.A. Melo Filho, F.A. Silva, R.D. Toledo Filho, Degradation kinetics and aging [18] M.E.A. Fidelis, Development and Mechanical Characterization of Jute Textile
mechanisms on sisal fiber cement composite systems, Cement Concr. Compos. Reinforced Concrete (Doctoral Thesis), Civil Engineering Department,
40 (2013) 30–39. Universidade Federal do Rio de Janeiro (COPPE/UFRJ), 2012 (in portuguese).
740 S.R. Ferreira et al. / Construction and Building Materials 101 (2015) 730–740

[19] Y. Ohama, Handbook of polymer-modified concrete and mortars., Properties [34] P. Zugernmaier, Crystalline Cellulose and Derivatives: Characterization and
and Process Technology, first ed., Noyes Publications, New Jersey, 1995. Structures. Springer Series in Wood Science, Springer, 2008.
[20] C.E.M. Gomes, O.P. Ferreira, Influência das adições do copolímero VA/VEOVA e [35] O. Derkacheva, Studying the structure of native, residual and dissolved lignins
fibras sintéticas nas propriedades da pasta de cimento Portland, Rev. Iberoam. by FTIR spectroscopy method, Proc. First Int. Workshop Physicochem. Prop.
Polím. 7 (3) (2006) 162–173. Lignins (2005) 255–257.
[21] F.A. Silva, N. Chawla, R.D. Toledo Filho, An experimental investigation on the [36] A. Bjorkman, Studies o finely divided wood part 3. Extraction of lignin–
fatigue behavior of sisal fibers, Mater. Sci. Eng., A 516 (2009) 90–95. carbohydrate complexes with neural solvents, Svensk Papperstidning 60 (7)
[22] Filho Toledo, F.A. Silva, J.A. Melho Filho, E.M.R. Fairbairn, Durability of (1957) 243–251.
compression molded sisal fiber reinforced mortar laminates, Constr. Build. [37] J.T. Kim, A.N. Netravali, Mercerization of sisal fibers: effect of tension on
Mater. 23 (2009) 2409–2420. mechanical properties of sisal fiber and fiber-reinforced composites, Compos.
[23] S.R. Ferreira, Influência da hornificação na aderência fibra-matriz e no A Appl. Sci. Manuf. 41 (2010) 1245–1252.
comportamento mecânico de compósitos cimentícios reforçados com fibras [38] E. Sinha, S.K. Rout, Influence of fibre-surface treatment on structural, thermal
curtas de sisal (Master dissertation), Civil and environmental Engineering and mechanical properties of jute, J. Mater. Sci. 43 (2008) 2590–2601.
Department, Universidade Estadual de Feira de Santana, Feira de Santana, [39] Y. Leonard, Mwaikambo, P. Martin, Ansell, Chemical modification of hemp,
Brasil, 2012 (in portuguese). sisal, jute and kapok fibers by alkalization, J. Appl. Polym. Sci. 84 (12) (2002)
[24] NBR 7217: agregados – determinação da composição granulometrica, ABNT, 2223–2233.
Rio de Janeiro, 2003. p. 6. [40] S. Alix, E. Philippe, A. Bessadok, L. Lebrun, C. Morvan, S. Marais, Effect of
[25] NBR 13276: argamassa para assentamento e revestimento de paredes e tetos: chemical treatments on water sorption and mechanical properties of flax
preparo da mistura e determinação do índice de consistência, ABNT, Rio de fibres, Bioresour. Technol. 100 (2009) 4742–4749.
Janeiro, 2005. p. 3. [41] Por Khalid Rehman Hakeem, Mohammad Jawaid, Umer Rashid, in: Biomass
[26] NBR 7215: cimento Portland: determinação da resistência à compressão, and Bioenergy: Processing and Properties, springer, 2014.
ABNT, 1996. p. 8. [42] M.A. Esmeraldo, A.C.H. Barreto, J.E.B. Freitas, P.B.A. Fechine, A.S.B. Sombra, E.
[27] L. Segal, J.J. Creely, A.E. Martin, C.M. Conrad, An empirical method for Corradini, et al., Dwarf-green coconut fibers: a versatile natural renewable raw
estimating the degree of crystallinity of native cellulose using the X-ray bioresource. Treatment morphology and physicochemical properties, Bioresources
diffractometer, Text. Res. J. 29 (1959) 786–794. 5 (4) (2010) 2478–2501.
[28] ASTM standard C1557, Standard Test Method for Tensile Strength and Young’s [43] O. Erikson, B.O. Lindgren, About the linkage between lignin and hemicelluloses
Modulus of Fibers, ASTM International, West Conshohocken, PA, 2013. in wood, Svensk Papperstidning 80 (2) (1977) 59–63.
[29] A. Arbelaiz, B. Fernández, J.A. Ramos, I. Mondragon, Thermal and [44] A. Roy, S. Chakraborty, S.P. Kundu, R.K. Basak, S.B. Majumder, B. Adhikari,
crystallization studies of short flax fibre reinforced polypropylene matrix Improvement in mechanical properties of jute fibres through mild alkali
composites: effect of treatments, Thermochim. Acta 440 (2006) 111–121. treatment as demonstrated by utilisation of the Weibull distribution model,
[30] A.C.H. Barreto, D.S. Rosa, P.B.A. Fechine, S.E. Mazzetto, Properties of sisal fibers Bioresour. Technol. 107 (2012) 222–228.
treated by alkali solution and their application into cardanol-based [45] P. Methacanon, U. Weerawatsophon, M. Smransin, C. Prahsarn, D.T. Bergado,
biocomposites, Composites Part A (2011). Properties and potential application of the selected natural fibers as limited
[31] P.J. Herrera-Franco, A. Valadez-González, A study of the mechanical properties life geotextiles, Carbohydr. Polym. 82 (2010) 1090–1096.
of short natural-fiber reinforced composites, Composites Part B 8 (2005) [46] M.D.H. Beg, K.L. Pickering, Accelerated weathering of unbleached and bleached
597–608. kraft wood fibre reinforced polypropylene composites, Polym. Degrad. Stabil.
[32] H. Yang, R. Yan, H. Chen, D.H. Lee, C. Zheng, Characteristics of hemicellulose, 93 (2008) 1939–1946.
cellulose and lignin pyrolysis, Fuel 86 (12–13) (2007) 1781–1788. [47] A. Bentur, S. Mindess, Fibre Reinforced Cementitious Composites, 2nd ed.,
[33] K.V. Sarkanen, Lignins: Occurrence, Formation, Structure and Reactions, Taylor & Francis, 2007.
University of Washington, Seattle, Washington, 1987.