Cement and Concrete Composites 102 (2019) 134–144

Contents lists available at ScienceDirect

Cement and Concrete Composites

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

Impact of content and length of curauá fibers on mechanical behavior of T

extruded cementitious composites: Analysis of variance
R.S. Teixeiraa,∗, S.F. Santosb, A.L. Christoforoc, J. Payád, H. Savastano Jr.e, F.A. Rocco Lahra
a
University of São Paulo (USP), Department of Structural Engineering, Av. Trabalhador São Carlense, 400, 13566-590, São Carlos/SP, Brazil
b
São Paulo State University (UNESP), Department of Materials and Technology, Av. Ariberto Pereira da Cunha, 333, 12516-410, Guaratingueta/SP, Brazil
c
Federal University of São Carlos (UFSCar), Department of Civil Engineering, Rodovia Washington Luís, Km 235, s/n, 13565-905, São Carlos/SP, Brazil
d
Universitat Politècnica de València (UPV), Institute of Science and Technology of Concrete, Camino de Vera s/n, Edificio 4N, E-46022, Valencia, Spain
e
University of São Paulo (USP), Department of Biosystems Engineering, Av. Duque de Caxias Norte, 225, 13635-900, Pirassununga/SP, Brazil

ARTICLE INFO ABSTRACT

Keywords: The use of composite materials in construction has grown considerably in recent years, such as cementitious
Mechanical of fracture matrices and concrete reinforced with fibers. The vegetable fibers have become an alternative due to its
Extrusion process abundance, low cost and low energy consumption for its production, and appropriate properties mechanical.
Lignocellulosic fiber Curauá fiber is a plant native from Amazonas harvested manually in commercial farming and it is used in the
Amazonian fiber
manufacture of ropes and baskets or as reinforcement in composite with organic matrix of components for cars,
buses and trucks. On the other hand, the extrusion process can produce composites with high-density matrix
with fibers, low permeability and good interface between fiber and matrix. This process is also compatible with
the use of vegetable fibers as raw materials in the production of cost-effective construction elements such as
ceiling panels and drywalls. The objective of this research was use the analysis of variance (ANOVA) for eval-
uating with rigor mathematical the influence of content and length of curauá fibers on the mechanical behavior
of the extruded cementitious composites. Composites without fibers and reinforced with 1% and 2% by mass of
fibers as well as 6 mm and 10 mm of length these curauá fibers were evaluated. The composites with fibers of
10 mm have showed better mechanical results. Besides, the composites with fibers curauá after 200 accelerated
aging cycles were better than one non-aging.

1. Introduction The curauá fiber is a plant native from to Amazon and a hydrophilus
species which belongs to pineapple/bromeliad family. It is a lig-
Fiber-cement products have been widely used globally due to their nocellulosic fiber that has mechanical properties comparable to syn-
versatility as corrugated and flat roofing materials, cladding panels, and thetic polymeric fiber [7–10].
water containers presented in large number of building and agriculture As part of the global strategy to produce regional and environ-
applications [1,2]. mental-friendly materials, the extrusion technology has been success-
In order to improve the sustainability of construction materials, part fully used as an economical and efficient and processing method for
of the global strategy is to use regional, recyclable and renewable manufacturing sustainable fiber-reinforced cement based composites
materials from agroindustrial resource and environmentally appro- [11,12]. Other advantage of the extrusion process is its capacity of
priate technologies for civil construction. In recent years, new tech- producing not only flat shapes, but also structural and complex shapes.
nologies using lignocellulosic fibers have arisen enabling the use of Besides, this process allows the use of a variety of materials that have
composites with less environmental impact, low cost and low power been successfully incorporated such as lignocellulosic fiber, including,
consumption, allowing partial replacement of synthetic fibers, such as sugar cane fiber [4] and sisal fiber [13].
polypropylene or polyvinyl alcohol [1,3–5]. However, a major drawback in the use of lignocellulosic fibres in
The incorporation of lignocellulosic fibers which are mostly co- cement matrix is the durability of fibres in the alkaline cement matrix
products of agriculture and agro-industries, allows a valorization of (pH around 12). Partial alkaline hydrolysis leads to a reduction of the
these residues and a limitation of environmental damages [1,6]. macromolecular chains initially in lignin and followed by cellulose with

∗
Corresponding author.
E-mail addresses: [email protected], [email protected] (R.S. Teixeira).

https://doi.org/10.1016/j.cemconcomp.2019.04.022
Received 15 August 2018; Received in revised form 14 January 2019; Accepted 22 April 2019
Available online 23 April 2019
0958-9465/ © 2019 Elsevier Ltd. All rights reserved.
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

consequent reduction in the degree of polymerization of both phases Mastersizer S long bed, version 2.19). The particle size distributions of
[14–16]. Furthermore, gradual filling of the inner cores of the fibres the raw materials are depicted in Fig. 2. Fig. 2a and b shows the discrete
with cement hydration production leads to embrittlement of fibres. particle size distributions and the cumulative percentage finer than
Both degradation mechanisms have an attendant effect on significant (CPFT) of the cement and limestone filler, respectively. Cement and
reinforcing properties of the fibres such as toughening mechanisms as limestone particles showed 50% of its mass less than 11.89 and
well as other long term mechanical properties of fibre cement compo- 12.38 μm, respectively. Both raw materials exhibit similar particle
sites [1,17]. distributions.
The degradation of the composite can be studied by accelerated The quantitative chemical analysis was performed of on OPC and
tests, whose advantage is to provide results in a smaller time interval limestone using PANalytical Axios Advanced X-ray fluorescence
[18,19]. The durability test (accelerating aging) of the cementitious equipment. The oxides are listed in Table 2. The specific surface area
composites shows their performance when subjected to wet/dry cycles (determined using the BET method) and specific density of raw mate-
and therefore may be recommended for both internal and external rials were measured. The OPC presented value of specific density of
building applications [16,20]. 3.10 g/cm3 and specific surface area of 1.10 m2g-1 whilst limestone had
In this study, Analysis of Variance (ANOVA) was used to statistically values of 2.80 g/cm3 and 1.14 m2g-1, respectively. The similar values of
evaluate the mechanical behavior of extruded fiber-cement composites the specific surface area may be important to avoid competition of
reinforced with curauá fiber with different content and lengths, before water between the raw materials in the system.
and after accelerated aging.
2.2. Formulation and preparation of composites cementitious
2. Experimental
The composite is composed of CPeV-ARI cement and limestone
2.1. Raw materials filler. The water-soluble polymers, High Range Water Reducer (HRWR)
provided by Aditex and polyether carboxylic provided by Grace, were
Curauá fibers (CF) (Ananas erectifolius) used in this study were ob- used as lubricant, representing 1% of cement mass.
tained from the Pematec Triangel Industry in Pará/PA, Brazil. Hydroxypropylmethyl cellulose (HPMC) with an average molecular
Mechanical and physical properties and chemical composition of the weight of 86,000 and a viscosity of 5.39 cps (at a concentration of 2% in
fiber are listed in Table 1. water at 20 °C) and carboxylate polyether (surfactant) commercially
The results of the chemical composition of curauá fibers was similar called ADVA 170, were used as rheological modifiers to promote
to values reported in literature. For instance, mass % of cellulose was pseudo plastic behavior of the composite. The mix design used in this
within 71%–73% whilst lignin content fell within the range of work is presented in Table 3.
7.5%–13% [24,25]. It is important to consider that there is a natural The sequence of mixing was powder/water [28]. The cement,
variation in the chemical composition of vegetable fibers as a result of limestone, curauá fiber (CF) and HRWR (by dry mass) were mixed and
quantification methods and characteristics of the fibers of a specific homogenized at low speed (distributive mixture), in a mechanical
region and harvested at different times throughout the season. Eirich intensive mixer (capacity of 10 L) for 5 min. After this stage,
Fig. 1a shows the surface roughness of the curauá fiber and Fig. 1b water and carboxylate polyether were added fractionally for 2 min. All
shows its cross section, approximately elliptical shape. These char- raw materials were mixed at high speed for another 5 min to achieve a
acteristics help to anchor better in cementitious matrix. Also, in Fig. 1b, high shear mixing to ensure break down of agglomerates generated in
it is possible to observe unit cells and lumens (cavities) within the fiber wet mixing stage. Before composites production, the composite was re-
cross section. homogenized in the extruder itself, feeding it and taking two times the
Unbleached unrefined eucalyptus (Eucalyptus grandis) Kraft pulp mass.
was provided by Fibria S/A, Brazil. The cellulose pulp was collected Composites with 15 mm thickness were extruded, as shown in
directly from the mill, prior to drying and pressing. It was extensively Fig. 3a and b. An extrusion helical screw equipment (Auger type),
washed with water and centrifuged to remove any residual chemicals Verdes, model 051, was used. The equipment comprised of a motor
from the pulping processes. speed regulator that was maintained at 4 mm/s during extrusion. Pads
Ordinary Portland cement (OPC) type CPeV-ARI, corresponding to with 200 mm × 50 mm x 15 mm were cured at water vapor saturated
ASTM C 150 [27], Type I was selected because of its finer particle size environment (in sealed plastic bags) at 25 ± 2 °C for two days. Sub-
and higher reactivity. Additionally, this type of cement contains higher sequently, the specimens were maintained in a water vapor saturated
levels of tricalcium silicate (C3S) and dicalcium silicate (C2S) for the environment (in sealed plastic bags) and placed in a chamber at 45 °C
formation of CeSeH. The cement and limestone filler particles dis- for five days (thermal curing) making a total of 7 days of curing [13].
tributions were evaluated by a laser particle size analyzer (Malvern
2.3. Mechanical characterization of the composites
Table 1
Mechanical, chemical and physical properties of the curauá fiber. Mechanical characterization tests were adopted according to Santos
et al., [13]. The fiber–cement composites were tested using a servo-
Curauá fiber (CF)
hydraulic mechanical testing machine MTS (810 series) controlled by
Ultimate tensile strength (MPa) 550 the TestStar IIs system. Prismatic specimens were prepared using a
Young's modulus (GPa) 64 diamond cut-off wheel before grinding and final polishing of the spe-
α-Cellulose a (% by mass) 68 cimen sides. The specimens had nominal dimensions of
Lignin b (% by mass) 14
Hemicellulose c (% by mass) 10
80 mm × 20 mm x 13 mm for all of the mechanical tests of the fiber-
Average length (mm) 6.00 ± 0.88 and 10.00 ± 0.46 cement composite. The modulus of rupture (MOR) was calculated by
Cross section (mmb) 0.1136 equation (1) and determined using a three point bending configuration
Thickness (μm) ∼75 with a span of 64 mm and cross-head speed of 5 mm min−1.
Density (g/cmc) 1.42
Aspect Ratio ∼80 and ∼133 3*Pmax *LV
MOR =
a
2*b*h2 (1)
TAPPI T 204 CM-97 [21].
b
Zimmermann et al., [22]. where Pmax is maximum load, b is width, h is height and Lv is span
c
TAPPI T 222 OM-02 [23]. between supports.

135
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

Fig. 1. Morphological characteristics of CF observed by scanning electron microscopy. (a) Lateral surface and (b) cross-section detaching the irregular lumens [26].

Fig. 2. (a) Particle size distribution of ordinary Portland cement and (b) limestone filler. The legend indicates the maximum equivalent diameter for each accu-
mulated percentage of particles.

Table 2 Table 3
Chemical analysis by means of X-ray fluorescence of the particulate raw ma- Formulations used in the production of cementitious composite.
terials (% by mass).
Raw material Content [% by mass]/length
Oxides compositions Ordinal Portland cement (OPC) CPeV-ARI Limestone
Ref 1%/6 mm 2%/6 mm 1%/10 mm 2%/10 mm
SiO2 (%) 14.70 9.40
CaO (%) 67.20 39.10 Portland cement 69.95 68.87 67.79 68.87 67.79
Al2O3 (%) 4.07 2.16 [CPeV-ARI]a
Fe2O3 (%) 3.50 1.25 Limestone fillerb 27.08 26.66 26.54 26.66 26.54
MgO (%) 3.13 8.90 Eucalyptus cellulosic 2.98 2.93 2.89 2.93 2.89
P2O5 (%) – 0.16 pulp
SO3 (%) 5.23 – Curauá fiber (CF)c – 1.00 2.00 1.00 2.00
K2O (%) 0.75 0.41 Water/cement ratio 0.33 0.34 0.36 0.34 0.36
MnO (%) – < 0.10
a
TiO2 (%) – 0.15 ASTM C-150 [27].
b
Provided by Infibra Ltda.
c
Volume fraction of fibers in study: 1.53% and 3.08%, respectively.
The fracture toughness (KIC), which is the critical stress intensity
factor value for crack growth in the material during mode-I failure was KIC =
Pmax
y( )
also used to characterize the cement based composites reinforced with bw1/2 (2)
curauá fiber. The SENB-type (single-edge notch bend) specimens were
where y(α) is the geometric factor and accounts for both shape of crack
prepared to establish the critical defect size and catastrophic fracture
and loading geometry, the tensile stress at fracture, and a0 the crack
[1]. The test configuration was the three-point bending. Prismatic
size. The ratio α = a0/w of the initial notch length to specimen height
specimens were prepared with a centered plan notch with a depth equal
was 0.1 (or 10% as mentioned before).
to 10% of the specimen height and notch tip profile in the shape of a
‘‘V’’ with an angle of approximately 30° using a diamond disk of 0.5 mm S 3 1/2

thickness to simulate a sharp crack. A cross-head speed of 15 mm min−1 y( ) = x 1.99 1.33 3.49 0.68
w 2(1 )3/2
was used. The values of the maximum load, Pmax, from the load–dis-
placement curves were applied in the calculation of the value of KIC 2 (1
+ 1.35
using the following equation, according to Santos et al. [13],: (1 + )2 (3)

where S is the span of 64 mm, and α is the relative length of the notch,

136
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

Fig. 3. (a) Front view of the vacuum extruder machine and (b) side view illustrating the composite exiting of the die [29].

which, in turn, is the ratio of the original length of the notch, a0, and the 2.4. Physical characterization tests
height of the specimen, w.
The fracture energy (FE) test was performed with the SENB type Water absorption (WA) apparent porosity (AP) and bulk density
specimen and three-point bending configuration, but the centered plan (BD) values were obtained from an average of five specimens for each
notch was of 30% of the specimen height. The span was of 64 mm. A formulation, following procedures specified by ASTM C 948 [32],
cross-head speed of 10 μm min−1 was used to guarantee stable growth Standards.
of the crack and to measure the energy required for extending this crack
over a unit area [30]. 2.5. Scanning electron microscopy (SEM)
The work performed by the machine to completely propagate the
crack along the specimen divided by two times the projected area of the Scanning electron microscopy (SEM) was used with secondary
fracture surface (cross section of the specimen, A) was used to de- electron (SE) detector, operated at 5.0 kV accelerating voltage, for ob-
termine the fracture energy, γWoF. The integration of the for- servation of the morphologies on the fractured surface of composites
ce–displacement curve was performed up to the point where the force generated in the mechanical tests. A back-scattered electron (BSE) de-
decreased to 5% of its maximum value reached during the test, ac- tector operated at around 15.0 kV and 20.0 kV was applied for viewing
cording to equation (4): cut and polished surfaces. The BSE image was used to study the fi-
ber–matrix transition zone. Energy dispersive X-ray spectroscopy (EDS)
1
= Pd analyses were also conducted. These were performed on the same flat
WoF
2A (4)
surface specimens in an effort to obtain semi-quantitative composi-
tional information. The preparation of specimens for BSE and EDS was
Additionally, a mechanical parameter was obtained the ‘‘relative
work of crack-propagation’’ [31]. This mechanical parameter is ob- accomplished with vacuum (80 kPa gauge) impregnation using cya-
noacrylate ester resin. BSE EDS samples were semi-automatic grinded
tained by dividing the work of crack propagation by the initial work
with silicon carbide abrasive paper with sequential grit sizes of 320,
(elastic energy stored). The initial work is that performed from zero up
600, 1200 and 2000 for 2 min each, using alcohol as lubricant. A final
to the point of the maximum load (Fig. 4). Although up to this point,
preparation was carried out using in turn 8–4, 4–2 and 1–0 μm diamond
some crack propagation could already occur, it is easy to determine.
polishing compound during 4, 2 and 1 min each size respectively.
This ratio considers all of the work performed for the effective crack
Fractured and polished samples were gold coated in a Bal-Tec Med020
propagation related to the elastic energy stored in the system. There-
coating system before being analyzed in a Hitachi TM 3000 microscope.
fore, a higher value of this relative work indicates that the material is
more resistant to propagation of a crack [31].
2.6. Accelerated aging testing

The accelerated aging testing involved a comparative analysis of

physical and mechanical performance of the composites, before and
after 200 soak/dry cycles. Specimens were successively immersed in
water at 20 ± 5 °C during 170 min, followed by the interval of 10 min,
and then exposed to temperature of 70 ± 5 °C for 170 min in a venti-
lated oven and with the final interval of 10 min. This procedure was
based on recommendations of the EN 494 [33], Standards. Each soak/
dry set represents one cycle and was performed for 200 cycles (200C)
[4].

2.7. Statistical analysis

The physical and mechanical properties evaluated were: water ab-

sorption (WA), apparent porosity (AP), bulk density (BD), modulus of
rupture (MOR), fracture energy (FE), fracture toughness (KIC) and re-
lative work of crack propagation (RWP). The factors and levels in-
vestigated consisted of curauá fiber fractions (% F) [1, 2%], fiber length
(FL) [6, 10 mm] and curing type (Cr) [7d (0), 200C (1)]. The product of
the three factor levels along with the reference conditions [Ref] (0%
fiber and 7 days cure [Tr1], 0% fiber resulted in an experimental design
Fig. 4. Schematic illustration of a typical load–displacement curve divided into consisting of 9 treatments, as explained in Table 4. It should be noted
two regions: initial work and work of crack propagation [13]. that the treatment Tr1 (reference) was used in the manufacture of

137
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

Table 4
Experimental treatments.
Formulation Treatment (Tr) %F FL Cr

Ref Tr1 0 0 7 days (0)

1%/6 mm Tr2 1% 6 mm 7 days (0)
1%/10 mm Tr3 1% 10 mm 7 days (0)
2%/6 mm Tr4 2% 6 mm 7 days (0)
2%/10 mm Tr5 2% 10 mm 7 days (0)
1%/6 mm Tr6 1% 6 mm 200C (1)
1%/10 mm Tr7 1% 10 mm 200C (1)
2%/6 mm Tr8 2% 6 mm 200C (1)
2%/10 mm Tr9 2% 10 mm 200C (1)

materials for the determination of physical and mechanical properties.

The methodology Design of Experiments (DoE), using Minitab®
software version 14, was used to establish the relationship between the
properties (physical and mechanical) and the factors evaluated, to un-
derstand the effects and to identify the factors and interactions con-
sidered significant and to identify the treatments that resulted in the
extreme values of the properties estimated by the models. The analysis
of variance (ANOVA) of regression models (Equation (5)) was evaluated
at the 5% level of significance (α), considering the non-significance (P-
value < 0.05) of the models and coefficients as null hypothesis (H0)
and significance as an alternative hypothesis (H1). For the validation of
the regression models, the normality of the generated residues was
tested with the aid of the Anderson-Darling normality test, also at the
5% level of significance, and for the hypotheses formulated, P-value
greater than or equal to 0.05 implies in the normality of the waste
distribution, validating the ANOVA model.

Y = β0+β1∙%F+β2∙FL+β3∙Cr+β4∙%F∙FL+β5∙%F∙Cr+β6∙FL∙Cr+β7∙%
F FL∙Cr +ε (5) Fig. 5. Average values and standard deviation (MOR) modulus of rupture (a)
From Equation (5), Y denotes the estimated physical and mechan- and (KIC) fracture toughness (b).
ical properties, βi are the coefficients obtained from the least squares
method and ε consists of the random error. The coefficient of de- considered non-significant by ANOVA of the regression models were
termination (R2) was used to measure the quality of the adjustments underlined.
obtained, and it should be noted that the reference conditions are not
MOR = 1.39775 + 9.30425∙%F+1.55238∙FL+20.7527∙Cr–1.05938∙%
incorporated in the regression models.
F∙FL–14.2959∙%F∙Cr 1.23217∙FL∙Cr+1.14854∙%F∙FL∙Cr (R2 = 79.72%);
In planning involving the 9 treatments after understood the effects
MOR = [12.19; 23.18] (6)
of factors and interactions between them on each property investigated.
It can be noted that 6 or more determinations were obtained by treat- KIC = 0.878467–0.0865817∙%F–0.0040055∙FL+0.438194∙Cr
ment and property investigated, totaling 382 determinations. +0.0117967∙%F∙FL 0.332244∙%F∙Cr–0.0522791∙FL∙Cr+0.0312871∙%F
FL∙Cr (R2 = 81.65%); KIC = [0.64; 1.08] (7)
3. Results and discussion ECC reinforced with 2% of curauá fiber (CF) at 7d showed a sig-
nificant increase of MOR values in relation to the ECC with 1% of CF
3.1. Mechanical properties and reference (Ref), according the ANOVA test shown in Equation (6)
and Fig. 5a. However, the results of fiber length did not differ sig-
Fig. 5 shows the mechanical properties (modulus of rupture (MOR), nificantly between the formulations. The MOR is the tensile strength in
fracture toughness (KIC), fracture energy (FE), and relative work of bending as well as it is influenced by interaction and distribution of
crack propagation (RWP)) of extruded cementitious composite (ECC) stresses between fiber-matrix and matrix porosity. The inclusion of fi-
cured at 7 days (7d), after accelerated aging (200C), the confidence bers increases the toughness and the reinforcement of composites, but
intervals of the mean (at the 95% confidence level) and the range of also increases the porosity because the dispersion deficiency of the fi-
variation of the coefficient of variation (CV). Table 5 lists the respective bers in cementitious matrix and, consequently, generating lack of stress
average values and standard deviations of the mechanical properties transfer between the fibers and matrix. After 200C, (Fig. 5a), the MOR
and physical characteristics. The P-values of the Anderson-Darling values of the ECC increased due to the combined effect better adhesion
normality test of the ANOVA residues for the four mechanical proper- (increase in chemical bonding) of CF in the cementitious matrix, the
ties (MOR, KIC, FE and RWP, respectively) evaluated ranged from 0.265 continued hydration process in the fiber-matrix interface and petrifi-
to 0.781, validating the ANOVA model, and by the results of the de- cation or mineralization these fibers [34].
termination coefficients [79.72%; 99.02%], the satisfactory estimation According to Melo Filho et al. [17], mineralization occurs under
of the properties provided by the models, is verified; being all con- conditions in which the cement hydration products migrate to the more
sidered significant by ANOVA. Equations (6) and (7) expresses the re- porous regions within the fibers (surface pores and lumens).
gression model obtained in the estimation of the MOR and KIC values, From the interaction between the three statistical factors (fiber
respectively, and also the coefficients of determination (R2) and the fractions (% F) [1, 2%], fiber length (FL) [6, 10 mm] and curing type
numerical intervals of this property are presented. In Equations (6) and (Cr) [7d, 200C]), the highest value occurs from the combination with
(7), the outliers were excluded from the set of results, and the terms

138
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

Table 5
Average values and standard deviations of modulus rupture (MOR), fracture toughness (KIC), fracture energy (FE), relative work of propagation (RWP), water
absorption (WA), bulk density (BD) and apparent porosity (AP) of the extruded cementitious composite (ECC) reinforced with curauá fiber (CF), 6 mm and 10 mm of
length in the conditions at 7 days (7d) of curing and after 200 accelerated aging cycles (200C).
Formulations Condition MOR (MPa) KIC (MPa.m1/2) FE (J/m2) RWP WA (%) BD (g/cm3) AP (%)

Reference (Ref) 7d 18.75 ± 0.75 1.05 ± 0.04 187.61 ± 16.07 9.77 ± 4.22 16.14 ± 0.28 1.702 ± 0.013 27.47 ± 0.34

1%/6 mm 7d 13.82 ± 0.57 0.83 ± 0.02 108.20 ± 13.58 3.21 ± 0.67 16.49 ± 0.20 1.708 ± 0.007 28.16 ± 0.29
1%/10 mm 15.85 ± 1.66 0.88 ± 0.03 309.98 ± 15.98 14.65 ± 5.59 17.22 ± 0.15 1.729 ± 0.055 29.78 ± 0.84
2%/6 mm 16.01 ± 1.92 0.78 ± 0.08 102.05 ± 9.05 3.66 ± 0.80 18.48 ± 0.72 1.726 ± 0.016 31.90 ± 0.96
2%/10 mm 15.21 ± 2.26 0.92 ± 0.10 245.09 ± 89.49 14.04 ± 3.13 17.04 ± 0.24 1.720 ± 0.031 29.31 ± 0.59

1%/6 mm 200C 19.52 ± 1.63 0.82 ± 0.03 45.67 ± 4.28 1.41 ± 0.42 16.73 ± 0.14 1.740 ± 0.009 29.10 ± 0.23
1%/10 mm 20.45 ± 2.17 0.77 ± 0.03 56.75 ± 5.83 1.43 ± 0.63 17.41 ± 0.10 1.754 ± 0.044 30.54 ± 0.72
2%/6 mm 15.16 ± 1.27 0.70 ± 0.06 20.18 ± 2.51 0.61 ± 0.25 17.77 ± 0.27 1.725 ± 0.008 30.66 ± 0.40
2%/10 mm 18.08 ± 2.77 0.78 ± 0.05 41.75 ± 3.89 1.67 ± 1.17 17.59 ± 0.19 1.723 ± 0.027 30.31 ± 0.39

Fig. 6. SEM of the composites at 7d and after 200C, respectively. (a) and (b) reference composite and (c) and (d) composite reinforced with CF.

2% of fibers, 10 mm in length and after 200C. produced cement composites, reinforced with 3% eucalyptus pulp and
MOR values obtained in this study were 16 MPa at 7d and 20 MPa 2% sisal fibers, produced by the extrusion method and subjected to
after 200 cycles for the formulation 1%/10 mm. These MOR values accelerated carbonation curing in the supercritical condition. The au-
were higher than those found in previous studies with ECC reinforced thors also analyzed the composite before and after 200 cycles of im-
with sisal and sugarcane fibers, with respective content of 1% by mass mersion and drying. The average results obtained was 0.9 MPa m1/2 and
and 1.5% by mass and the distribution of lengths between 15 mm to 0.85 MPa m1/2, for the unaged and aged composites, respectively.
18 mm and 10 mm–15 mm that showed, respectively, MOR values of Fig. 6a–d shows the micrographs of ECC with polished surface,
8 MPa–11 MPa for 28 days (28d) and 4 MPa–15 MPa after 200 cycles matrix reinforced with CF at 7d and 200C respectively, revealing a
[4,35]. homogenous microstructure of the cement matrix at 7d and more
The individual statistical interactions as the higher content (%F) densified after 200C. The composite reinforced with CF (Fig. 6c and d),
and fiber length (FL) decrease KIC values, i.e., initial crack growth re- reveals a high porosity in the fiber-matrix interface, i.e., a poor adhe-
sistance in cement matrix, according to ANOVA, as calculated by sion which does not lead to satisfactory values of MOR and KIC in re-
equation (7) and whose values are shown in Fig. 5b. The KIC value of lation to the Ref. In the mixing process, the ECC with CF consumed
the reference composites (1.05 MPa m1/2) suggests that these results more water, which promotes the formation of porosity and interferes
may be related to distribution of defects in the matrix produced by strongly in the w/c ratio (water/cement) and the packaging of the raw
several factors such as the difficulty in packaging fiber in the matrix materials in the matrix.
with particles and negative interference in cement hydration process Fig. 6d indicates that the vegetable fiber suffers a dimensional
caused by absorption of water by CF. However, the statistical combi- variation because it tends to lose water to the system of the cementi-
nation of 2% by mass of fiber with 10 mm length after 200C increases tious matrix which in turn is submitted to a rehydration process. This
the KIC values. It indicates that the matrix was improved mainly after dimensional variation process of the fiber promotes detachment of it in
200 cycles of immersion and drying. the cementitious matrix, consequently, it affects the mechanical beha-
The results of fracture toughness of the composites reinforced with vior of the cement composite [19].
CF are similar to the results obtained by Santos et al., [13]. The authors The average values of fracture energy values (FE) and relative work

139
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

The outliers were excluded from the set of results, and the terms
considered non-significant by ANOVA of the regression models were
underlined.
FE is the energy per unit area needed to completely fracture the
composite in quasi stable crack propagation process in order to record
the contribution of all toughness mechanisms mainly promoted by the
fibers. The RWP is the ratio of the plastic working and the elastic work
i. e. indicates the degree of pseudoplastic deformation, that is, it shows
that capability of the cement composite to absorb energy.
The ECC reinforced with 2% of fibers presented a significant sta-
tistical FE value in relation to ECC with 1% at 7d and after accelerated
aging (200C), based on Equation (8) and Fig. 7a. According to Ro-
drigues and Montardo [36] and Bentur and Mindess [37], the fiber
content provides greater post-cracking energy and smaller size of the
cracks, since the fibers help to absorb the elastic energy necessary to
propagate cracks, which occur between fiber fractions. Regarding the
fiber length, the average value of FE of the composite with fiber of
10 mm is higher than one with fibers of 6 mm, according to the ANOVA,
as calculated by Equation (9) and demonstrated in Fig. 7b, before ac-
celerated aging. However, FE values of composites with 10 mm fibers
presented major standard deviations than ones with fibers of 6 mm
length. Fiber with length of 6 mm has a higher number of filaments per
mass, but fiber of 10 mm presented more efficiency in relation to
pullout mechanism that increased average values of the FE and RWP.
Fiber with 10 mm has a better degree of adhesion in the cementitious
matrix between the different lengths of fibers, i.e. there is a greater
probability of shorter fibers be pulled without an effective frictional
energy. For a surface shear stress applied to the fiber, this will be more
efficient if its length is capable of allowing the shear stress permits the
development of a tensile stress equal to its tensile strength [37,38].
Fig. 7. Average values and standard deviation (FE) Fracture energy (a) and From the individual statistical factors, increases in fiber content and
(RWP) relative work of crack propagation (b) in the composite extruded. length promote increases in FE and RWP values and the best type of
cure is 7d, however, the interaction between the three statistical factors
of crack propagation (RWP) of the composite at 7d and after 200C are (fiber fractions (% F) [1, 2%], fiber length (FL) [6, 10 mm] and curing
shown in Fig. 7a and b, respectively. Equations (8) and (9) expresses the type (Cr) [7d, 200C]), the highest value for FE and RWP occurs from the
regression analysis obtained in the estimation of the FE and RWP va- combination with 2% of fibers, 10 mm in length and after 200C.
lues, and the coefficients of determination (R2) and the numerical in- ECC reinforced with 1%/10 mm length fibres presented the highest
tervals of these properties are also presented, respectively. A regression average FE value of 310 J/m2 and 57 J/m2, at 7d and after 200C, re-
analysis generates an equation to attempts to explain the statistical spectively. The value of FE at 7d was higher than values determined by
relationship between one or more predictors and the response variable. Santos et al. (2015), which produced extruded composite reinforced
with 2% by mass of eucalyptus cellulosic pulp and 3% by mass of sisal
FE = −314.084 + 113.851∙%F+70.6381∙FL+384.347∙Cr–19.6163∙% fibers with a length distribution between 1 mm and 14 mm, 7 days cure
F∙FL–155.064∙%F∙Cr–70.4887∙FL∙Cr+22.2370∙%F FL∙Cr (R2 = 99.02%); and exposed to supercritical carbonation. They reported an average
FE = [16.68; 391.23] (8) value of FE approximately 230 J/m2. Correia et al. [39], worked with
RWP = −21.7027 + 4.75874∙%F+4.01494∙FL+25.4552∙Cr– extruded composites reinforced with hybrid cellulosic fibers (8% cel-
0.686908∙%F∙FL–7.11968∙%F∙Cr–4.27127∙FL∙Cr+0.946910∙%F FL∙Cr lulosic fibers + 1% bamboo nanofibers) that presented FE value around
(R2 = 98.21%); RWP = [0.33; 19.23] (9) 430 J/m2 after 28 days curing, but decreased to 271 J/m2 after ac-
celerated aging of 200C. The reduction of the FE value after accelerated

Fig. 8. SEM micrographs of composite extruded at 7d (a) the arrow shows deboning fiber from cement matrix after 200 cycles (b) breaked fiber.

140
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

ANOVA model. Although the coefficient of determination obtained for

the estimation of apparent porosity (AP) was 33.22%, all regression
models were considered significant by ANOVA. Equations (10)–(12)
expresses the regression analysis obtained in the estimation of the WA,
BD and AP values, and also the coefficients of determination (R2) and
the numerical intervals of this property are presented, respectively.

WA = 10.0648 + 5.32864∙%F+0.743583∙FL+3.14206∙Cr–0.560438∙
%F∙FL–2.83893∙%F∙Cr–0.341080∙FL∙Cr+0.329509∙%F FL∙Cr (R2 =
92.09%); WA = [16.24; 18.47] (10)

BD = 17.5215 + 8.65426∙%F+1.21420∙FL+5.76981∙Cr–0.883184∙%
F∙FL–4.71053∙%F∙Cr–0.506878∙FL∙Cr+0.486204∙%F FL∙Cr (R2 =
84.44%); BD = [27.86; 31.98] (11)

AP = 1.61519 + 0.0700806∙%F+0.0122749∙FL+0.160396∙Cr–
0.00855097∙%F∙FL 0.0935377∙%F∙Cr–0.0158791∙FL∙Cr+0.0100770∙%F
FL∙Cr (R2 = 33.22%); AP = [1.67; 1.82] (12)

In Equations (10)–(12) the outliers were excluded from the set of

results, and the terms considered non-significant by ANOVA of the re-
gression analysis were underlined.
ECC reinforced with 2% of curauá fiber (CF) presented higher values
of WA in relation to the formulation with 1% by mass showed in
Equation (10) and Fig. 9a. This may be attributed to the greater number
of fibers per unit volume resulting in inefficient packaging of the ma-
trix, as well as varying aspect ratios of the fibers, CF (80 and 133, re-
spectively for lengths of 6 mm and 10 mm) and, consequently more
defects appears in the interface fiber and matrix [40,41]. The results of
fiber length and type of cure did not differ significantly between the
formulations according to results from ANOVA. Analysis of the inter-
action between the three statistical factors; fiber fractions (% F) fiber
length (FL) and curing type (Cr), reveals that the highest value occurs
from the combination with 2% of fibers, 6 mm or 10 mm in length at 7d
or after 200C.
Composites with 2% by mass of fibers after 200C presented a sig-
nificant increase of the BD values as shown in Equation (11) and
Fig. 9b, which can be attributed to the filling of the matrix pores by the
continued hydration process during accelerated aging cycles in forma-
tion of calcium hydroxide (Ca(OH)2), calcium silicate hydrate (CSH)
and calcium carbonate (CaCO3) [13]. These phenomena has been re-
ported extensively in studies focused on accelerated aging of cementi-
tious composites; Soto et al., [35]; Dias et al. [42], and Teixeira [4]. The
only individual factor that did not statistically affect the bulk density
(BD) values was the fiber length. Thus, the interaction between the
three statistical factors (fiber fractions (% F) [1, 2%], fiber length (FL)
[6, 10 mm] and curing type (Cr) [7d, 200C]), the highest value occurs
from the combination with 2% of fibers, 6 mm or 10 mm in length and
after 200C.
According to ANOVA analysis, considering the individual statistical
factors and the interaction, only accelerated aging (200C) promoted
Fig. 9. Results of the physical properties: (WA) water absorption (a); (BD) bulk significant differences in the values of the apparent porosity (AP) of the
density (b) and (AP) Apparent porosity (c). composites shown in Equation (12) and Fig. 9c. It is believed that the
immersion and drying cycles caused higher incidence of pores as result
aging indicates that degradation of the fibers in the ECC caused de- of microcracks caused by the aging cycles [39]. At the initial period of
bonding and breakage of fibers as illustrated in Fig. 8a and b. the immersion cycles, lignocellulosic fiber could absorb the water from
the cementitious microstructure since the surrounding water con-
centration is greater than within the fibers. At a later stage of drying
3.2. Physical characterization cycles, the moisture content within the fibers dry out and thus, fibers
shrink to a smaller size. The shrinkage of the fiber generates micro-
Fig. 9 presents graphs with average values of the physical para- cracks between the fiber and cement matrix, creating voids that explain
meters: water absorption (WA), bulk density (BD) and apparent por- the low results of FE and RWP in the transition zone [29,39]. From the
osity (AP) of the composites at 7 days and after 200C for different fibre interaction between the three statistical factors (fiber fractions (% F) [1,
fractions and lengths of curauá fiber (CF). The confidence intervals of 2%], fiber length (FL) [6, 10 mm] and curing type (Cr) [7d, 200C]), the
the mean (at the 95% confidence level) and the range of variation of the highest value occurs from the combination with 1% or 2% of fibers,
coefficient of variation (CV). The P-values of the Anderson-Darling 6 mm or 10 mm in length at 7d.
normality test of the ANOVA residues for the three physical properties
(WA, BD and AP) evaluated ranged from 0.109 to 0.918, validating the

141
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

Fig. 10. (a) Image Scanning electron microscopy (SEM-BSE) polished surfaces of composites reinforced with CF and points of EDS analyzes that are marked on the
images (1–4): (1) limestone; (2) anhydrous grain; (3) cementitious matrix and (4) cellulosic pulp.

3.3. Scanning electron micrograph SEM atomic number of the predominant chemical elements) correspond to
the longitudinal sections of the fibers.
Fig. 10 shows the micrographs (SEM) with EDS in the composites Point 1 in Fig. 10 shows a grain with a high calcium index from the
reinforced with CF detaching the points of chemical elements in ce- formulation as a result of 27% of limestone in the mix. Point 2 indicates
mentitious structures. The dark areas in the image (associated with low an anhydrous grain that was not completely hydrate comprising

142
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

Fig. 11. Micrograph of scanning electron microscopy (SEM-BSE) polished surfaces of composites reinforced with CF: (a) CF (arrow indicates adhesion fiber-cement)
and (b) CF after 200C with points of EDS analysis that is signaled the image (point 1 and 2).

predominantly of Si and Ca elements. result of absorption of cement hydration products) which consequently
Point 3 clearly shows a grain comprising cement constituents. In caused the fiber to decrease its properties. The fiber content directly
point 4, the cellulose pulp with a high calcium content and silicon are influenced the mechanical performance and fibers with greater lengths
shown, indicating that the fiber absorbed cement hydration water, showed better mechanical results for modulus of rupture (MOR) and
which was similar to that presented by Tonoli et al. [43], and Teixeira fracture energy (FE) according to ANOVA. Cementitious composites
et al. [29], who studied cementitious composites reinforced with eu- reinforced with curauá fibers showed superior mechanical performance
calyptus cellulosic pulp. compared to available literature. The modulus of rupture (MOR),
In Fig. 11a, it is observed a resulted of the dimensional variation of fracture energy (FE) and relative working crack propagation (RWP)
fiber according to its humidity content. Resulting pores lead to higher results of the composites reinforced with curauá fibers after 200 ac-
water absorption, greater porosity and low resistance. celerated aging cycles were better in relation of the composites at 7
Savastano and Agopyan [44], explain that the best performance is days, because of the cement hydration, which filled the pores, densified
achieved by better adhesion of the fiber-matrix. The improved ad- its structure, which improved the transition zone fiber matrix. On the
herence is achieved by reducing the porosity and the lowest con- other hand, the aging promoted mineralization of the fiber, which re-
centration of Portlandite (calcium hydroxide crystals) approximately duced the mechanical performance of composites with curauá com-
the fiber. pared to values reported in literature. Thus, the best results were ob-
In Fig. 11b, after 200 cycles, there is better adhesion of the fiber in tained for composites reinforced with 2% of curauá fiber with 10 mm of
the matrix. Point 1 shows the absence of cement hydration products in length after 200C.
the central part of the fiber as indicated by the EDS, the large presence In the physical results, the composite with 2% of fiber showed an
of C. However, at point 2, the border of the fiber presents the main increase in water absorption and bulk density due to the greater
chemical elements of cement, silicon and calcium [19,45]. EDS also number of fibers per unit volume as well as the filling of the matrix
reveals the presence of other elements such as, Al, S, and Mg. This pores by the continued hydration process during accelerated aging
phenomenon can be associated with mineralization fibers as indicated cycles (200C) in formation of calcium hydroxide (Ca(OH)2), calcium
by Bentur and Akers [46]. On immersion in water, free ions formed by silicate hydrate (CSH) and calcium carbonate (CaCO3) respectively.
the dissolution of the cementitious phases of Portland cement pene- Considering the individual statistical factors and the interaction, only
trated into the lumen of the fibers, leading to the formation of et- the accelerated aging (200C) promoted significant differences in the
tringite/monosulfates and calcium hydroxide Ca(OH) [29,45]. Batic values of the apparent porosity (AP). It is believed that the immersion
et al. [47], showed that the re-precipitation of ettringite in microcracks and drying cycles caused higher incidence of pores as result of micro-
and pores of the cementitious composite may occur under normal cracks caused by the aging cycles. According to the ANOVA, the only
conditions (ambient temperature) curing. Such training has previously individual factor statistical that did not affect the physical values was
been suggested as one of the degradation mechanisms of the fibers the fiber length. Thus, the interaction between the three statistical
within the concrete matrix [14,15]. factors, the highest value occurs from the combination with 2% of
curauá fibers, 6 mm or 10 mm in length after 200C.
Scanning electron microscopy of the fracture surface of curauá fi-
4. Conclusions
bers showed chemical elements from the cement inside the fiber. Also
showed detachment of the fibers from the cement matrix indicating low
Curauá fiber suffered mineralization process during curing (as a

143
R.S. Teixeira, et al. Cement and Concrete Composites 102 (2019) 134–144

mechanical performance. J. Maldonado, F.A.R. Lahr, S. Delvasto, Different ageing conditions on cementitious
These results encourage us to use composites reinforced with curauá roofing tiles reinforced with alternative vegetable and synthetic fibres, Mater.
Struct. 47 (2014) 433–446, https://doi.org/10.1617/s11527-013-0070-0.
fiber in constructions in indoor environments as ceiling and partitions. [20] L.K. Aggarwal, Bagasse-reinforced cement composites, Cement Concr. Compos. 17
New tests should be used to optimize results. (1995) 107–112.
[21] TAPPI T 204 Cm-97, Solvent Extractives of Wood and Pulp, TAPPI TEST METHODS.
Tech. Assoc. Pulp an Pap. Ind, Tappi Press, Atlanta, GA, 1997, pp. 1–4.
Acknowledgements [22] T. Zimmermann, N. Bordeanu, E. Strub, Properties of nanofibrillated cellulose from
different raw materials and its reinforcement potential, Carbohydr. Polym. 79
(2010) 1086–1093, https://doi.org/10.1016/j.carbpol.2009.10.045.
The authors acknowledge the financial support provided by [23] TAPPI T 222 Om-02, Acid-Insoluble Lignin in Wood and Pulp, TAPPI TEST
Brazilian Agencies: Fundação de Amparo à Pesquisa do Estado de São METHODS. Tech. Assoc. Pulp an Pap. Ind, Tappi Press, Atlanta, GA, 2002, pp. 1–5.
[24] O. Faruk, A.K. Bledzki, H.-P. Fink, M. Sain, Progress report on natural fiber re-
Paulo (FAPESP, Grant nº 2013/03823-8 and 2012/51467-3);
inforced composites, Macromol. Mater. Eng. 299 (2014) 9–26, https://doi.org/10.
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, 1002/mame.201300008.
Grant nº 3886/2014); and Conselho Nacional de Desenvolvimento [25] J.C. Caraschi, A.L. Leão, Characterization of curaua fiber, Mol. Cryst. Liq. Cryst. Sci.
Technol. Mol. Cryst. Liq. Cryst. 353 (2000) 149–152, https://doi.org/10.1080/
Científico e Tecnológico (CNPq, Grant nº 406429/2015 and 312151/ 10587250008025655.
2016-0). The authors thank the Brazilian companies Fibria S.A., Infibra [26] S.F. Santos, R.S. Teixeira, H. Savastano Jr., Interfacial transition zone between
S.A. and Imbralit Ltda. for technical support to the development of this lignocellulosic fiber and matrix in cement-based composites, Sustain. Nonconv.
Constr. Mater. Using Inorg. Bond. Fiber Compos. vol. 6, Woodhead Publishing,
work. 2017, pp. 27–68.
[27] ASTM C 150, Am. Soc. Test. Mater. – ASTM C150/C150M-11, Stand. Specif. Portl.
Cem, 2011.
References [28] F.A. Cardoso, R.G. Pileggi, V.M. John, Caracterização reológica de argamassas pelo
método de squeeze flow, VI Simpósio Bras. Tecnol. Argamassa, vol. 2005,
[1] S.F. Santos, J.D.A. Rodrigues, G.H.D. Tonoli, A.E.F.D.S. Almeida, H. Savastano Jr., Florianópolis, 2005, pp. 121–143 121–143. (in portuguese).
Effect of colloidal silica on the mechanical properties of fiber–cement reinforced [29] R.S. Teixeira, G.H.D. Tonoli, S.F. Santos, E. Rayón, V. Amigó, H. Savastano Jr.,
with cellulosic fibers, J. Mater. Sci. 49 (2014) 7497–7506, https://doi.org/10. F.A. Rocco Lahr, Nanoindentation study of the interfacial zone between cellulose
1007/s10853-014-8455-1. fiber and cement matrix in extruded composites, Cement Concr. Compos. 85 (2018)
[2] S. Ikai, J.R. Reichert, A.V. Rodrigues, V.A. Zampieri, Asbestos-free technology with 1–8, https://doi.org/10.1016/j.cemconcomp.2017.09.018.
new high toughness polypropylene (PP) fibers in air-cured Hatschek process, [30] J. Nakayama, Direct measurement of fracture energies of brittle heterogeneous
Constr. Build. Mater. 24 (2010) 171–180, https://doi.org/10.1016/j.conbuildmat. materials, J. Am. Ceram. Soc. 48 (1965) 583–587, https://doi.org/10.1111/j.1151-
2009.06.019. 2916.1965.tb14677.x.
[3] F. Pacheco-Torgal, S. Jalali, Cementitious building materials reinforced with ve- [31] S. Ribeiro, J.A. Rodrigues, The influence of microstructure on the maximum load
getable fibres: a review, Constr. Build. Mater. 25 (2011) 575–581, https://doi.org/ and fracture energy of refractory castables, Ceram. Int. 36 (2010) 263–274, https://
10.1016/j.conbuildmat.2010.07.024. doi.org/10.1016/j.ceramint.2009.07.033.
[4] R.S. Teixeira, G.H.D. Tonoli, S.F. Santos, J. Fiorelli, H. Savastano Jr., F.A.R. Lahr, [32] ASTM C 948-81, Am. Soc. Test. Mater. ASTM C 948-82 Test Method Dry Wet Bulk
Extruded cement based composites reinforced with sugar cane bagasse fibres, Key Density, Water Absorption, Apparent Porosity Thin Sect, vol. 1982, Glas. Reinf.
Eng. Mater. 517 (2012) 450–457, https://doi.org/10.4028/www.scientific.net/ Concr., West Conshohocken, PA, USA, 1982.
KEM.517.450. [33] EN 494, Eur. Comm. Stand. EN 494 Fibre-Cement Profiled Sheets Fitt. Roof. – Prod.
[5] T. Tan, S.F. Santos, H. Savastano Jr., W.O. Soboyejo, Fracture and resistance-curve Specif. Test Methods. London, UK, vol. 1994, BSI – Br. Stand. Institution., 1994.
behavior in hybrid natural fiber and polypropylene fiber reinforced composites, J. [34] A. Bentur, S.A.S. Akers, The microstructure and ageing of cellulose fibre reinforced
Mater. Sci. 47 (2012) 2864–2874, https://doi.org/10.1007/s10853-011-6116-1. autoclaved cement composites, Int. J. Cem. Compos. Lightweight Concr. 11 (1989)
[6] C.S. Fonseca, T.F. Silva, M.F. Silva, I.R.C. Oliviera, R.F. Mendes, P.R.G. Hein, 111–115, https://doi.org/10.1016/0262-5075(89)90121-8.
L.M. Mendes, G.H.D. Tonoli, Micro/Nanofibrilas celulósicas de eucalyptus em fi- [35] Y.J.M. Soto, G.H.D. Tonoli, R.S. Teixeira, S.F. Santos, H. Savastano Jr., Prospective
brocimentos extrudados, Cerne 22 (2016) 59–68, https://doi.org/10.1590/ study on vegetable wastes as reinforcement in extruded fibre-cement, Int. Conf.
01047760201622012084 (in portuguese). Non-conventional Mater. Technol. Ecol. Mater. Technol. Sustain. Constr. vol. 2007,
[7] C.R. Araújo, L.C. Perlaza, C.G. Mothé, Thermal properties of commercial and castor Anais do Brasil-NOCMAT. (Abmtenc), Maceió, 2007.
oil polyurethane composites with curauá fiber, Fourth Int. Symp. Nat. Polym. [36] P.P.F. Rodrigues, J.P.A. Montardo, Influência da adição de fibras de polipropileno
Compos, vol. 2002, São Paulo, 2002. nas propriedades dos concretos para pisos e pavimentos, 44o Congr. Bras. Do
[8] F. Tomczak, K.G. Satyanarayana, T.H.D. Sydenstricker, Studies on lignocellulosic Concreto, vol. 2002, Instituto Brasileiro do Concreto – IBRACON, Belo Horizonte,
fibers of Brazil: Part III - morphology and properties of Brazilian curauá fibers, 2002(in portuguese).
Compos. Part A Appl. Sci. Manuf. 38 (2007) 2227–2236, https://doi.org/10.1016/j. [37] A. Bentur, S. Mindess, Fibre Reinforced Cementitious Composites, vol. 449, Elsevier
compositesa.2007.06.005. Applied Science, England, 1990.
[9] M.A.S. Spinacé, C.S. Lambert, K.K.G. Fermoselli, M.A. De Paoli, Characterization of [38] H.F.W. Taylor, Cement Chemistry, vol. 475p, Academic press limited, London,
lignocellulosic curaua fibres, Carbohydr. Polym. 77 (2009) 47–53, https://doi.org/ 1990.
10.1016/j.carbpol.2008.12.005. [39] V. da C. Correia, S.F. Santos, R.S. Teixeira, H. Savastano Jr., Nanofibrillated cel-
[10] D.G. Soltan, P. das Neves, A. Olvera, H. Savastano Jr., V.C. Li, Introducing a curauá lulose and cellulosic pulp for reinforcement of the extruded cement based materials,
fiber reinforced cement-based composite with strain-hardening behavior, Ind. Crops Constr. Build. Mater. 160 (2018) 376–384, https://doi.org/10.1016/j.conbuildmat.
Prod. 103 (2017) 1–12, https://doi.org/10.1016/j.indcrop.2017.03.016. 2017.11.066.
[11] K.G. Kuder, S.P. Shah, Processing of high-performance fiber-reinforced cement- [40] E.M. Bezerra, A.P. Joaquim, H. Savastano Jr., V.M. John, V. Agopyan, The effect of
based composites, Constr. Build. Mater. 24 (2010) 181–186, https://doi.org/10. different mineral additions and synthetic fiber contents on properties of cement
1016/j.conbuildmat.2007.06.018. based composites, Cement Concr. Compos. 28 (2006) 555–563, https://doi.org/10.
[12] C. Aldea, S. Marikunte, S.P. Shah, Extruded fiber reinforced cement pressure pipe, 1016/j.cemconcomp.2006.02.001.
Adv. Cem. Base Mater. 8 (1998) 47–55, https://doi.org/10.1016/S1065-7355(98) [41] G.H.D. Tonoli, H. Savastano Jr., S.F. Santos, C.M.R. Dias, V.M. John, F.A.R. Lahr,
00006-6. Hybrid reinforcement of sisal and polypropylene fibers in cement-based composites,
[13] S.F. Santos, R. Schmidt, A.E.F.S. Almeida, G.H.D. Tonoli, H. Savastano Jr., J. Mater. Civ. Eng. 23 (2011) 177–187, https://doi.org/10.1061/(ASCE)MT.1943-
Supercritical carbonation treatment on extruded fibre–cement reinforced with ve- 5533.0000152.
getable fibres, Cement Concr. Compos. 56 (2015) 84–94, https://doi.org/10.1016/ [42] C.M.R. Dias, H. Savastano Jr., V.M. John, Exploring the potential of functionally
j.cemconcomp.2014.11.007. graded materials concept for the development of fiber cement, Constr. Build. Mater.
[14] B.J. Mohr, H. Nanko, K.E. Kurtis, Durability of kraft pulp fiber – cement composites 24 (2010) 140–146, https://doi.org/10.1016/j.conbuildmat.2008.01.017.
to wet/dry cycling, Cement Concr. Compos. 27 (2005) 435–448, https://doi.org/ [43] G.H.D. Tonoli, U.P. Rodrigues Filho, H. Savastano Jr., J. Bras, M.N. Belgacem,
10.1016/j.cemconcomp.2004.07.006. F.A. Rocco Lahr, Cellulose modified fibres in cement based composites, Compos.
[15] B.J. Mohr, J.J. Biernacki, K.E. Kurtis, Microstructural and chemical effects of wet/ Part A Appl. Sci. Manuf. 40 (2009) 2046–2053, https://doi.org/10.1016/j.
dry cycling on pulp fiber-cement composites, Cement Concr. Res. 36 (2006) compositesa.2009.09.016.
1240–1251, https://doi.org/10.1016/j.cemconres.2006.03.020. [44] H. Savastano Jr., V. Agopyan, Transition zone studies of vegetable fibre-cement
[16] G.H.D. Tonoli, M.N. Belgacem, G. Siqueira, J. Bras, H. Savastano Jr., F.A. Rocco paste composites, Cement Concr. Compos. 21 (1999) 49–57, https://doi.org/10.
Lahr, Processing and dimensional changes of cement based composites reinforced 1016/S0958-9465(98)00038-9.
with surface-treated cellulose fibres, Cement Concr. Compos. 37 (2013) 68–75, [45] R. MacVicar, L.M. Matuana, J.J. Balatinecz, Aging mechanisms in cellulose fiber
https://doi.org/10.1016/j.cemconcomp.2012.12.004. reinforced cement composites, Cement Concr. Compos. 21 (1999) 189–196,
[17] J. de A. Melo Filho, F. de A. Silva, R.D. Toledo Filho, Degradation kinetics and aging https://doi.org/10.1016/S0958-9465(98)00050-X.
mechanisms on sisal fiber cement composite systems, Cement Concr. Compos. 40 [46] A. Bentur, S.A.S. Akers, The microstructure and ageing of cellulose fibre reinforced
(2013) 30–39, https://doi.org/10.1016/j.cemconcomp.2013.04.003. cement composites cured in a normal environment, Int. J. Cem. Compos.
[18] A.E.F.S. Almeida, G.H.D. Tonoli, S.F. Santos, H. Savastano Jr., Improved durability Lightweight Concr. 11 (1989) 99–109, https://doi.org/10.1016/0262-5075(89)
of vegetable fiber reinforced cement composite subject to accelerated carbonation 90120-6.
at early age, Cement Concr. Compos. 42 (2013) 49–58, https://doi.org/10.1016/j. [47] O.R. Batic, C.A. Milanesi, P.J. Maiza, S.A. Marfil, Secondary ettringite formation in
cemconcomp.2013.05.001. concrete subjected to different curing conditions, Cement Concr. Res. 30 (2000)
[19] R.S. Teixeira, G.H.D. Tonoli, S.F. Santos, H. Savastano Jr., T.P. Protásio, E.F. Toro, 1407–1412, https://doi.org/10.1016/S0008-8846(00)00343-4.

144