Construction and Building Materials 116 (2016) 63–71

Contents lists available at ScienceDirect

Construction and Building Materials

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

Mechanical properties of alkali activated blast furnace slag pastes

reinforced with carbon fibers
J.L. Vilaplana a, F.J. Baeza a, O. Galao a, E.G. Alcocel b, E. Zornoza a, P. Garcés a,⇑
a
Department of Civil Engineering, University of Alicante, Alicante, Spain
b
Department of Architectural Constructions, University of Alicante, Alicante, Spain

h i g h l i g h t s

 The effect of CF aspect ratio and dosage in alkali-activated slag pastes is discussed.
 Blast furnace slag activated with water glass and CF reinforced to reduce shrinkage.
 CF could preserve AAS pastes from breaking due to high drying shrinkage.
 Compressive strength increases up to 19% were obtained with 3 mm long CF.
 Bending strength up to 18.6 MPa was achieved with 1% (by slag mass) 6 mm long CF.

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

Article history: Alkaline activation has become a useful procedure to obtain high strength cement composites without
Received 1 March 2016 Portland cement. Different industrial byproducts can be activated using this technique, e.g. blast furnace
Received in revised form 18 April 2016 slag, fly ash or metakaolin. Thus a new generation of more sustainable composites is being developed.
Accepted 18 April 2016
However, the main disadvantage of these composites is their high drying shrinkage strain. Among the
Available online 3 May 2016
different approaches to address this problem, shrinkage reducing admixtures are the most used, but
the addition of fibers can be useful for this purpose. In this work, alkali activated slag (AAS) pastes have
Keywords:
been reinforced with carbon fibers (CF). The effect of the concentration of alkaline activator (Na2O% and
Alkali-activated blast furnace slag
Carbon fiber
silica modulus) and the fiber aspect ratio (using different length fibers with the same diameter) on the
Multifunctional composites mechanical properties has been assessed. Mechanical characterization comprised bending and compres-
Mechanical properties sive strength tests, ultrasonic pulse velocity and density measurements, and drying shrinkage tests, in
Drying shrinkage which CF were capable of improving the mechanical strengths of AAS pastes while controlling the
specimens strain due to shrinkage. CF additions increased bending strength up to five times, and
increases up to 20% in compressive strength were observed. Furthermore, CF can be a convenient addition
to control the drying shrinkage of AAS composites. Even in adverse conditions, such us 50% RH, where
unreinforced specimens actually broke, CF could guarantee the stability of the specimen.
Ó 2016 Published by Elsevier Ltd.

1. Introduction
The application of industrial byproducts, some of which show
hydraulic properties, has grown as an alternative to Portland
cement in civil engineering applications [1–6]. Alkaline activation
has become a useful procedure to obtain high strength cement
composites without Portland cement [4–6]. Thus, besides of reus-
ing a waste product, the energy and environmental costs, related
to clinker production, can be reduced. And therefore, a new gener-
ation of more sustainable composites is being developed [5,7].
⇑ Corresponding author.
E-mail address: pedro.garces@ua.es (P. Garcés).
The use of water glass (sodium silicate base activators Na2SiO3-

xH2O + NaOH) as an alkaline activator for blast furnace slag (BFS)
has been recommended for the mechanical and durability proper-
ties of their composites, usually know as alkali activated slag (AAS)
[1,5]. The main advantage of AAS is their resistance against aggres-
sive agents (e.g. sulphates or chlorides), due to their low porosity
[3]. These composites also show high compressive strength, low
hydration heat, high carbonation or freeze-thaw resistances
[2–6]. But among their disadvantages, their important drying
shrinkage cannot be ignored [8–11].
The addition of short fibers to brittle composites represents a
commonly used technique to improve their mechanical properties,
especially those related to their tensile and shrinkage behaviors

http://dx.doi.org/10.1016/j.conbuildmat.2016.04.066
0950-0618/Ó 2016 Published by Elsevier Ltd.
64 J.L. Vilaplana et al. / Construction and Building Materials 116 (2016) 63–71
[12–14]. There are multiple references reporting this mechanical
gains in Portland cement composites [12–22]. Even in some cases
the sustainability issue has been addressed by means of natural
[18] or polymeric recycled fibers [19]. Besides, some reports can
be found where different fiber types are used in AAS composites
with different levels of success reducing the shrinkage of the mate-
rial [23–31]. Fibers can be made on different materials such as:
steel [23–26], PE [27], PVA [28], PP [29], AR glass fibers [30] or even
carbon [31]. However there is a lack of information regarding the
use of short carbon fibers for this purpose. Besides ref [31], there
are few references using carbon fibers to reinforce geopolymers
[7,32,33], but they don’t use BFS as binder, e.g. in ref [7] alkali acti-
vated metakaolin was used, in ref [32] fly ash was activated, while
in [33] metakaolin was reinforced with continuous carbon fibers.
The current study uses CF with a twofold aim, first improve the
mechanical performance of AAS pastes, and second the use of these
CF reinforced composites as strain sensors (as will be introduced
below). Hence, the CF dosages and lengths have been chosen
according to previous research on Portland cement composites
for their strain sensing capacities [34], and differ greatly from
those used in ref [31].
Carbon fibers offer another extra advantage, as they can be used
to functionalize cement composites due to electrical properties
improvements [35]. These functional properties include strain or
damage sensing [34,36,37], heating control [38,39], EMI shielding
[40] or anode for electrochemical chloride extraction [41] among
others. And some of them have been successfully applied in civil
engineering structures, always using cement composites [36,42].
To achieve these functionalities, a certain level of electrical con-
ductivity is necessary. Therefore, as concrete is a bad electrical con-
ductor, conductive admixtures are needed. Several researchers
have focused on these conductive admixtures (e.g., steel fibers, or
carbon materials: carbon fibers, graphite powder, carbon nanofi-
bers or nanotubes) in order to achieve a better electrical behavior
without compromising the composite’s mechanical properties
[34,35]. Thus, for each type of admixture, the relationship between
their dosage and the composite’s conductivity has to be deter-
mined. Researchers have focused on the minimum amount of
admixture which guarantees low resistivity [34,35,43,44].
Strain and damage sensing, i.e. the relationship between the
material’s resistivity and its stress or strain state, have been inves-
tigated in several research papers using fiber-reinforced cementi-
tious materials based on Portland cement [34,35]. Nonetheless,
these applications of carbon fiber-reinforced cement composites
based on alkaline activation have been barely studied [32,45]. An
interesting preliminary study, regarding the self-sensing capacity
of AAS composites reinforced with CF, was reported with remark-
able results compared to their cement composites counterparts
[45].
Therefore, the main objective of the present paper is to evaluate
the influence of carbon fiber addition on the mechanical properties
of AAS composites, which will be used afterwards as strain sensors
(despite this application is not discussed in this work). In this man-
ner, a compromise between structural and functional applications
can be achieved, and the optimal dosage can be selected to com-
bine good strain sensing capacity and mechanical performance.
Thus, the multifunctionality of carbon fiber reinforced cement
composites would be improved by the sustainability and mechan-
ical properties enhancement of alkaline cements.
2. Experimental procedure
2.1. Materials and specimen preparation
Carbon fiber alkali activated slag (CF-AAS) pastes were made using blast furnace
slag (BFS) with an alkaline activator compound of commercial water glass (an aque-
ous solution with mass ratios of 27% SiO2 and 8% Na2O) and a NaOH solution (with
different concentration for each particular dosage). The oxide composition of the
BFS obtained by X-ray fluorescence (XRF) is included in Table 1. A test to assess
the specific surface area of the BFS was made according to the Blaine method spec-
ified in UNE-EN 196-6:2010 [46]. The average specific surface area was
525.7 ± 19.1 m2/kg.
A preliminary study of pastes without CF was performed to select the best AAS
dosages to include the fibers. In this first stage the silica modulus of the activator
(Ms = SiO2/Na2O mass ratio) was 1.0, 1.2 or 1.4, the Na2O dosage changed between
3, 4 and 5% (with respect to BFS mass), and the activator/BFS ratio was fixed at
0.56. Afterwards, two AAS dosages were selected (4% Na2O with Ms = 1.4, and 5%
Na2O with Ms = 1.0) and different unsized PAN-based CF were included in the mix
(see main properties in Table 2). Three CF lengths were used, 3, 6 and 12 mm (i.e.
⅛, ¼ and ½ inch), and three different mass ratios of each length were prepared,
0.2, 0.5 and 1.0% (by BFS mass).
In order to improve the CF dispersion in the mix an ultrasounds treatment was
applied [31]. The CF were stirred in part of the mixing water and an ultrasonic
device, model Hielscher UP200S was used at full power for 10 min. Afterwards,
all the remaining components of the alkali solution were mixed and stirred by hand
together with the treated CF. Then all materials, i.e. alkali solution (with or without
CF) and BFS, were put in an automatic mixer for five minutes. The fresh mix was
then poured into prismatic steel molds, 4  4  16 cm for the mechanical tests
and 2.5  2.5  25 cm for the shrinkage tests. A mechanical treatment to remove
any air bubbles was made using a vibration table, and the molds were kept in a con-
trolled environment room (20 °C and >99% RH) for 24 h. Afterwards, specimens
were demolded and conserved in the same conditions until tested.
2.2. Testing
First, a study of the paste workability has been carried out according to the flow
table test included in UNE83811:1992EX [47]. Afterwards, mechanical properties
tests – bending and compressive strengths (Rf and Rc), ultrasonic pulse velocity
(UPV) and density – were run at curing ages of 7, 28 and 60 days. All tests were
made according to European standards: Bending strength and compressive strength
UNE-EN 196-1:2005 [48], UPV UNE-EN 12504-4:2006 [49] and density UNE-EN-
1015-10:2000/A1:2007 [50]. Shrinkage tests according to UNE-80112:1989 EX
[51] were also made, in which case the specimens were kept, after demolding, into
a controlled environment of 20 °C and 50% RH, and the strain variations were mea-
sured daily until stabilization.
3. Results and discussion
3.1. Mechanical properties of unreinforced AAS
First, prior to any CF addition, the influence on mechanical
properties of the sodium oxide concentration in the activator and
its silica modulus were studied and compared with other
researches. Table 3 includes the results of this first experimental
campaign, where AAS pastes with no CF were tested. The average
value and standard deviation (SD) for three or six measures,
depending on each test set up, are included for three different
Na2O dosages (% by BFS mass) and three Ms for two different curing
ages. There are no results for the dosage with 3% Na2O and silica
modulus equal to 1.0 because the paste did not totally set and
hardened enough to be tested. On the other hand, the dosage with
Table 1
Chemical composition of the BFS obtained by XRF.
Chemical composition Mass%
CaO 45.20
SiO2 32.32
Al2O3 10.62
MgO 6.35
SO3 2.03
Fe2O3 0.86
TiO2 0.54
K2O 0.44
ZrO2 0.42
Na2O 0.32
MnO 0.31
BaO 0.15
P2O5 0.13
Cl 0.13
Other oxides 0.19
 J.L. Vilaplana et al. / Construction and Building Materials 116 (2016) 63–71 65

Table 2 Fig. 1 shows the strength results versus the silica modulus of
Properties of carbon fiber PANEX 35. the activator for the different sodium oxide dosages and curing
Property Value Unit ages. A clear trend in compressive strengths can be seen in
Diameter 7.2 lm Fig. 1(b), with higher strength for higher silica modulus and
Length 3–6–12 mm Na2O content. In this regard, the maximum compressive strength
Carbon content 95 % enhancement was up to a 116% increase (from 41.23 to
Tensile strength 3800 MPa 89.11 MPa at 28 days) or even 197% at 7 days, in both cases if
Elastic modulus 242 GPa
Resistivity 1.52  10 3
X cm
silica modulus was increased from 1.0 to 1.4 together with an
Density 1.81 g/cm3 increase in Na2O from 3% to 5% (by BFS mass). Similar conclu-
sions were drawn in [52], in which the activator concentration
was the second most significant factor, and higher strengths
5% Na2O and 1.4 silica modulus showed a setting time too short, were achieved for increasing dosages, finally Na2O content was
and consequently couldn’t be mixed properly and only one sample recommended between 3 and 5% of BFS weight. There is also a
was obtained. This issues were also indirectly observed in the slight increase in strength between 7 and 28 days of curing in
workability of the fresh mix, which was also measured by means 100% RH environment, in this case this improvement was
of flow table according to UNE 83811:1992 EX [47]. In this case a between 55% and 13% for increasing activator concentrations,
systematic decrease in the mix slump was measured when either i.e. pastes with higher concentrations of activator cured faster.
the sodium oxide dosage or silica modulus increased. Furthermore, These trends do not appear for the bending strength, where
all dosages with 3 or 4% Na2O were too fluid, with slump values the better results were measured with 4% Na2O and 1.2 silica
higher than 260 mm, maximum size of the testing table. While, if modulus paste. If the proper dosage is chosen, the mechanical
the sodium oxide amount was increased to 5% the slump decreased strength of these pastes – between 40 and 90 MPa at 28 days
slightly under 160 mm. – can be higher than similar pastes in Portland cement. This

Table 3
Mechanical properties of AAS pastes: average value ± standard deviation (SD) for three (Rf and UPV) or six (Rc and density) specimens.

Na2O Ms Age Rf Rc UPV Density

% days MPa ± SD MPa ± SD km/s ± SD g/cm3 ± SD
3 1 7 – – – –
28 – – – –
3 1.2 7 1.02 ± 1.76 26.52 ± 8.13 2.873 ± 0.059 1.347 ± 0.025
28 4.06 ± 0.19 41.23 ± 2.02 3.075 ± 0.015 1.344 ± 0.012
3 1.4 7 0.98 ± 1.70 33.78 ± 0.80 2.883 ± 0.032 1.313 ± 0.008
28 6.42 ± 1.56 44.67 ± 2.09 3.089 ± 0.036 1.338 ± 0.016
4 1 7 3.09 ± 0.19 40.14 ± 0.46 3.048 ± 0.021 1.394 ± 0.043
28 4.88 ± 0.37 46.85 ± 1.43 3.160 ± 0.026 1.431 ± 0.069
4 1.2 7 2.91 ± 0.15 48.15 ± 1.59 3.105 ± 0.013 1.402 ± 0.018
28 6.65 ± 0.57 56.71 ± 1.29 3.224 ± 0.016 1.417 ± 0.042
4 1.4 7 3.36 ± 0.35 53.46 ± 1.56 3.150 ± 0.021 1.413 ± 0.007
28 5.72 ± 0.41 61.27 ± 3.99 3.283 ± 0.017 1.415 ± 0.054
5 1 7 2.73 ± 0.60 52.39 ± 2.37 3.292 ± 0.012 1.469 ± 0.006
28 3.56 ± 0.79 65.74 ± 2.93 3.421 ± 0.011 1.488 ± 0.017
5 1.2 7 4.24 ± 0.92 66.25 ± 4.26 3.347 ± 0.012 1.502 ± 0.029
28 5.30 ± 0.55 79.44 ± 1.19 3.471 ± 0.015 1.505 ± 0.018
5 1.4 7 4.43⁄ 78.69 ± 5.05 3.390⁄ 1.574 ± 0.031
28 4.59⁄ 89.11 ± 0.50 3.516⁄ 1.522 ± 0.006
⁄
Only one sample could be tested for bending strength assessment.

Fig. 1. (a) Bending and (b) compressive strength for AAS pastes versus silica modulus of the activator for different Na2O dosages and curing ages.
66 J.L. Vilaplana et al. / Construction and Building Materials 116 (2016) 63–71

enhanced mechanical performance is one of the aspects that
have drawn the attention of the researchers to these composites
despite their initial disadvantages [5,6].
UPV and density were also measured in the same samples, as
included in Fig. 2. In this case, compressive strengths are com-
pared with the UPV values – Fig. 2(a) – and densities – Fig. 2
(b) – for all the dosages tested. Almost a linear trend can be
adjusted for the UPV, with a systematic increase for higher
strength composites. A similar pattern can be seen between den-
sity and strength.
3.2. Mechanical properties of CF-AAS
After the first set of tests in unreinforced AAS pastes, two differ-
ent dosages were selected to study the effect of CF addition on
their mechanical properties. 4% Na2O – 1.4 silica modulus and a
5% Na2O – 1.0 silica modulus dosages were chosen because of their
similar mechanical performance, although they showed different
shrinkage behavior, as discussed below. Three different CF addi-
tions (0.2, 0.5 and 1.0% by BFS mass) and three different fiber
lengths (3, 6 and 12.5 mm) were combined and tested.

Fig. 2. Compressive strength for AAS pastes versus UPV and density of the composites for different Na2O dosages and curing ages.

Fig. 3. Bending strength for AAS pastes reinforced with different length CF for different alkali solution activator dosages (%Na2O – silica modulus) and curing ages: 7, 28 and
60 days.
 J.L. Vilaplana et al. / Construction and Building Materials 116 (2016) 63–71
Fig. 3 includes the bending strength results for each AAS type
with respect to the CF addition, and for each CF length and curing
age. First, as a general rule higher additions of CF led to higher
bending strengths. Flexural strength improvements were between
5% and 552% at 7 days, and between 22% and 410% at 28 days
(strength losses can be related to fiber dispersion issues as will
be discussed later). In addition, samples with the same CF addition
showed almost identical behavior. However differences between
both AAS dosages were only seen for 1% CF dosage when longer
fibers were used, 6 or 12 mm. In these cases some dispersions
issues may have appeared during the mixing process, as reported
in prior works in Portland cement composites [34]. When longer
fibers are used at higher amount, they may form clusters during
the paste fabrication, and consequently drop the mechanical
strength of the samples, mainly their bending strengths and
slightly their compressive strength. Nevertheless, the higher dis-
persion of the measurements for higher fiber additions is also a
sign of that mixing problems. The reinforcement effect of fibers
can be seen even at early ages reaching bending strength higher
than 16 MPa. In this regard short fibers are more efficient than
longer ones and result in higher strengths. On the other hand, a
slight strength loss was measured at 60 days, which may be related
to crack development due to shrinkage.
The average values and standard deviation for the compressive
strength tests are summarized in Fig. 4. In this case, between both
AAS proportions, the one with lower Na2O (4%) and higher silica
Fig. 4. Compressive strength for AAS pastes reinforced with different length CF for different alkali solution activator dosages (%Na2O – silica modulus) and curing ages: 7, 28
and 60 days.
67
modulus (1.4) was better in almost every case. The compressive
strength enhancement related to CF addition was between 18%
(for long CF, 12 mm) and 19% (for 3 mm CF). On the other hand,
the behavior of AAS pastes differs from prior experiences in Port-
land cement composites [34], where strength increases between
2 and 23% were detected for CF reinforced pastes, whichever their
dosage or fiber aspect ratio were. Actually, composites with at least
0.5% CF addition (by mass of cement) increased at least 13% the
strength of the unreinforced specimens [34]. In this case, AAS
pastes with CF not always improve the compressive strength of
their unreinforced counterparts. The best results were obtained
for the shortest fibers (3 mm), which showed a strength increase
of almost 19.3% for a 0.5% CF dosage, similar to the cited cement
composites [34]. The composites with 6 mm CF preserved the
strength of the unreinforced AAS, and one dosage showed a 10%
strength loss. However, the longest fibers, 12 mm, weren’t capable
of improving the compressive strength at all. In fact, strength
losses of almost 18% were registered for lower CF content.
This ineffectiveness of longer fibers to actually improve the
compressive strength can be explained by two different phenom-
ena. First, some issues may appear during the mixing process if
fibers are too long or the addition is too high. Thus fiber clusters
may appear as the one shown in Fig. 5(a), similar to the ones
reported in [34]. Second, apparently fibers show a preferential ori-
entation in the hardened paste, i.e. during the mixing-pouring-
compaction process fibers are not randomly oriented but follow
68 J.L. Vilaplana et al. / Construction and Building Materials 116 (2016) 63–71

the longitudinal axis of the specimen, Fig. 5(b). This effect can also
be observed in the SEM image included in Fig. 5(c). Therefore, both
situations combined can lead to the mechanical properties loss
aforementioned.
Fig. 6 includes the UPV results for each AAS type versus their
CF dosage. The fiber addition has very little effect on this param-
eter, which increases slightly for older samples. Besides, as a
general rule the dosages with 5% Na2O showed higher UPV, even
though the trend observed for the compressive strength was the
contrary.
3.3. Effect of CF on shrinkage of AAS pastes
It was already stated in the introduction that, as several authors
report, the main problem of AAS composites is their high drying
shrinkage [8–11]. In fact, in this study some specimens were
untestable because they were severely cracked due to shrinkage
before the age of testing. The standard UNE-80112:1989 EX [51]
establishes the shrinkage test conditions. In this case, samples
should be conserved at 20 °C and 50% RH between length measure-
ments. However, the samples without CF could not be tested in
these conditions, because they broke at early ages. Thus they wer-

Fig. 5. (a) Fiber cluster due to a difficult dispersion of long fibers in the mix. (b) Preferential orientation of fibers along the longitudinal axis of specimens. (c) SEM image of an
AAS paste reinforced with CF.

Fig. 6. Ultrasounds Pulse Velocity for AAS pastes reinforced with different length CF (3, 6 and 12 mm) for different alkali solution activator dosages (%Na2O – silica modulus)
and curing ages: (a) 7 days; (b) 28 days; and (c) 60 days.
 J.L. Vilaplana et al. / Construction and Building Materials 116 (2016) 63–71 69

en’t measured at the expected time. In order to compare, unrein- during the first weeks of autogenous shrinkage process for each
forced AAS pastes were kept in 20 °C and 100% RH for shrinkage fiber addition and compares them with their unreinforced counter-
tests, and their results are included in Fig. 7. Only six different parts. All of them were kept the whole time between measure-
dosages could be tested, pastes with 3% Na2O and alkali moduli ments in controlled environment (20 °C and 100% RH). Each
lower than 1.4 actually broke in the first 24 h or did not harden graph corresponds to a different fiber length, and different behav-
enough to be demolded. While composites with 5% Na2O and silica iors can be observed in each one. For example, all samples with
modulus of 1.4 showed a setting time too short to even pour the 3 mm CF showed a better behavior than the unreinforced pastes.
mix into the molds properly. Fig. 7 includes the average curve (of The maximum strain of the control AAS was between 3 and
two specimens) for each dosage. As a general rule, a higher autoge- 5 mm/m, while the specimens with 3 mm fibers showed maximum
nous shrinkage can be related to higher values of both, silica mod- strains lower than 2.5 mm/m. The 6 mm CF are also effective for
ulus and Na2O% , specially the first one. Because, for example, for a shrinkage control, but in this case there are significant differences
1.2 silica modulus there was almost no difference between 4 and 5 between dosages. It is noteworthy that not always a greater fiber
Na2O% . These results are very similar to those reported by Krizan addition means a lower shrinkage strain, as can be seen in [31].
[11] where 4% and 1.2 AAS showed approximately 3 mm/m Sometimes 0.2% CF addition (by BFS mass) was enough to obtain
shrinkage strain. good results, with maximum strains between 1 and 1.5 mm/m at
Fig. 8 includes the average shrinkage curves for each dosage of
AAS paste reinforced with different amounts of three different
length CF. In these specimens it was possible to maintain the
50% RH condition referred in UNE 80112:1989 EX [51]. Therefore,
the final values of this figure were higher than the ones presented
in Fig. 7 for unreinforced samples in water saturated environment.
In this second study two different Na2O% – silica modulus combi-
nations were selected (4%-1.4 and 5%-1.0), both of which showed
very different shrinkage behaviors without CF addition (Fig. 7).
Nonetheless, if the CF addition was high enough, these differences
between AAS dosages disappeared. Thus, the effect of fiber rein-
forcement on the shrinkage deformation can be observed compar-
ing the series with the same color in each graph. In Fig. 8(a), 0.2% of
3 mm CF were not capable of controlling the shrinkage of the spec-
imens, and the effect of AAS proportion prevailed, thus different
strains were registered for each AAS dosage (i.e. 12 mm/m strain
or almost 16 mm/m for the 1.0 or 1.4 Ms composites, respectively).
However, if the 3 mm CF addition was augmented to 0.5% (by BFS
mass), the effect of CF prevailed, and the final strain at 28 days was
almost identical for both AAS dosages. Nonetheless, the strain rate
was much higher for the highest silica modulus paste. Finally, there
weren’t almost any differences if 1% CF was added to the mixes. On
the other hand, if the fiber length was increased the behavior of the
composite was ruled by the CF%, regardless of the AAS paste
dosage. Furthermore, the longer the fibers were, the lower effect
the fiber dosage had in the shrinkage behavior, i.e. the 12 mm CF
pastes showed very similar strain curves for all three fiber
amounts.
In addition, a new set of tests with samples conserved in satu-
rated environment (20 °C and 100% RH) was performed, in order to
make a proper assessment of the effectiveness of CF as shrinkage
control method in AAS. Thus a comparison with the results of the
unreinforced pastes can be made. Fig. 9 includes the strain curves

Fig. 7. Autogenous shrinkage of AAS pastes without CF in a 20 °C and 100% RH Fig. 8. Drying shrinkage (at 20 °C and 50% RH) for AAS pastes reinforced with
environment for different alkali solution activator dosages (%Na2O – silica different length CF: (a) 3 mm, (b) 6 mm, and (c) 12 mm. for different alkali solution
modulus). activator dosages (Na2O% – silica modulus): 4% – 1.4 and 5% – 1.0.
70 J.L. Vilaplana et al. / Construction and Building Materials 116 (2016) 63–71
Fig. 9. Effect of different length CF – (a) 3 mm, (b) 6 mm and (c) 12 mm – on the autogenous shrinkage (at 20 °C and 100% RH) of AAS pastes with different alkali solution
activator dosages (Na2O% – silica modulus): 4% – 1.4 and 5% – 1.0, and CF% (by BFS mass).
28 days. Even more, in some cases, when 12 mm CF was used, the
addition of fibers made the shrinkage worse. Probably, this effect is
related to the paste mixing issues explained previously, i.e. fiber
clusters as in Fig. 5(a).
As a summary, a CF addition can highly reduce the autogenous
and drying shrinkage strains of AAS pastes, and guarantee their
stability even in conditions of 50% RH. As the CF dosage or length
are increased, the final strains are controlled and diminished. From
almost 16 mm/m for 0.2% 3 mm CF, an increase in fiber addition up
to 1% (by BFS mass) can limit the strain to values lower than 6 mm/
m. Nonetheless, a more extensive study should be made to opti-
mize the dosage of the composites, as a compromise between the
amount of reinforcement and a proper mixing process.
4. Conclusions
The aim of the present work is to study the influence of carbon
fiber addition on the mechanical properties of alkali activated blast
furnace slag pastes. Specimens of different AAS types, with varia-
tions in the alkali solution, were prepared and tested to determine
the experimental dosages that were used afterwards as control val-
ues for the CF reinforced pastes. After discussing the results, the
following conclusions can be drawn:
- With unreinforced pastes, an increase on either silica modulus
or Na2O content led to higher compressive strength values, as
expected from references. This tendency wasn’t observed in
the case of bending strength, whose highest value was obtained
with an intermedium dosage, i.e. 4% Na2O and 1.2 silica modu-
lus. Furthermore, a linear correlation between compressive
strength and UPV or density was observed.
- The bending strength was systematically increased with CF
dosage (up to 410% at 28 days and 522% at 7 days with respect
unreinforced pastes). The higher the fiber dosages were, the
higher bending strengths achieved, up to values close to
18 MPa.
- Compressive strengths were only improved significantly if 3 mm
CF were used. In this case, increases up to 19% were measured just
like previously reported in Portland cement composites. How-
ever, if longer fibers were added the strength was maintained
or even decreased (up to 18%), probably due to mixing problems
such as fiber cluster formation or preferential orientation.
- Carbon fibers can be a convenient addition to control the drying
shrinkage of AAS composites. 0.2% addition of 3 mm or 6 mm
long fibers can be enough to reduce the maximum strain below
1.5 mm/m, in a 100% RH favorable environment. Even more, in
harder conditions, such us 50% RH, where unreinforced speci-
mens actually broke, CF could guarantee the stability of the
specimen regardless of the damaged they had suffered.
Once the effect of carbon fibers on the mechanical properties of
alkali activated BFS composites has been determined, the conclu-
sions of this work will be used to design conductive composites,
which combine the enhanced mechanical properties of AAS with
the electrical conductivity of carbon fibers. And therefore these
new composites could be used as multifunctional materials, such
as strain sensors. Besides, the combined use of carbon fibers and
shrinkage reducing admixtures should be addressed to optimize
shrinkage control of these composites, and assessing the influence
they could have on their strain sensing properties.
Acknowledgements
The authors want to acknowledge Generalitat Valenciana
(Spain) for the economic support, grant PROMETEO/2013/035.
References
[1] D. Ravikumar, N. Neithalath, Effects of activator characteristics on the reaction
product formation in slag binders activated using alkali silicate powder and
NaOH, Cem. Concr. Compos. 34 (2012) 809–818.
[2] C. Bilim, C. Duran, Alkali activation of mortars containing different
replacement levels of ground granulated blast furnace slag, Constr. Build.
Mater. 28 (2012) 708–712.
[3] F. Puertas, R. de Gutiérrez, A. Fernández-Jiménez, S. Delvasto, J. Maldonado,
Alkaline cement mortars. Chemical resistance to sulfate and seawater attack,
Mater. Constr. 52 (2002) 55–71.
[4] E. Rodríguez, S. Bernal, R.M. de Gutiérrez, F. Puertas, Alternative concrete based
on alkali-activated slag, Mater. Constr. 58 (2008) 53–67.
[5] A.M. Rashad, A comprehensive overview about the influence of different
additives on the properties of alkali-activated slag – a guide for civil engineer,
Constr. Build. Mater. 47 (2013) 29–55, http://dx.doi.org/10.1016/
j.conbuildmat.2013.04.011.
[6] A. Fernández-Jiménez, A. Palomo, Properties and uses of alkali cements, Rev.
Ing. Constr. 24 (2009) 213–232.
[7] A. Natali, S. Manzi, M.C. Bignozzi, Novel fiber-reinforced composite materials
based on sustainable geopolymer matrix, Proc. Eng. 21 (2011) 1124–1131,
http://dx.doi.org/10.1016/j.proeng.2011.11.2120.
[8] A.A. Melo Neto, M.A. Cincotto, W. Repette, Drying and autogenous shrinkage of
pastes and mortars with activated slag cement, Cem. Concr. Res. 38 (4) (2008)
565–574, http://dx.doi.org/10.1016/j.cemconres.2007.11.002.
[9] M. Palacios, F. Puertas, Effect of shrinkage-reducing admixtures on the
properties of alkali-activated slag mortars and pastes, Cem. Concr. Res. 37
(2007) 691–702, http://dx.doi.org/10.1016/j.cemconres.2006.11.021.
[10] M. Palacios, F. Puertas, Effect of superplasticizer and shrinkage-reducing
admixtures on alkali-activated slag pastes and mortars, Cem. Concr. Res. 35
(2005) 1358–1367, http://dx.doi.org/10.1016/j.cemconres.2004.10.014.
 J.L. Vilaplana et al. / Construction and Building Materials 116 (2016) 63–71
[11] D. Krizan, B. Zivanovic, Effects of dosage and modulus of water glass on early
hydration of alkali-slag cements, Cem. Concr. Res. 32 (8) (2002) 1181–1188,
http://dx.doi.org/10.1016/S0008-8846(01)00717-7.
[12] F.B. Varona, F.J. Baeza, S. Ivorra, D. Bru, Experimental analysis of the loss of
bond between rebars and concrete exposed to high temperatures, Dyna
(Spain) 90 (1) (2015) 78–86, http://dx.doi.org/10.6036/7184.
[13] P.W. Chen, D.D.L. Chung, Low-drying-shrinkage concrete containing carbon
fibers, Compos. Part B Eng. 27 (3–4) (1996) 269–274.
[14] T. Aly, J.G. Sanjayan, F. Collins, Effect of polypropylene on shrinkage and
cracking of concretes, Mater. Struct. 41 (10) (2008) 1741–1753.
[15] V.T. Giner, F.J. Baeza, S. Ivorra, E. Zornoza, O. Galao, Effect of steel and carbon
fiber additions on the dynamic properties of concrete containing silica fume,
Mater. Des. 34 (2012) 332–339, http://dx.doi.org/10.1016/
j.matdes.2011.07.068.
[16] A.M. Brandt, Fibre reinforced cement-based (FRC) composites after over 40
years of development in building and civil engineering, Compos. Struct. 86
(2008) 3–9.
[17] N. Banthia, J. Sheng, Fracture toughness of micro-fiber reinforced cement
composites, Cem. Concr. Compos. 18 (4) (1996) 251–269.
[18] R.D. Toledo, K. Ghavami, M.A. Sanjuan, G.L. England, Free, restrained and
drying shrinkage of cement mortar composites reinforced with vegetable
fibres, Cem. Concr. Compos. 27 (5) (2005) 537–546.
[19] D. Foti, F. Paparella, Impact behavior of structural elements in concrete
reinforced with PET fibers, Mech. Res. Commun. 57 (2014) 57–66, http://dx.
doi.org/10.1016/j.mechrescom.2014.02.007.
[20] T.J. Kim, C.K. Park, Flexural and tensile strength developments of various shape
carbon fiber-reinforced lightweight cementitious composites, Cem. Concr. Res.
28 (7) (1998) 955–960.
[21] S. Yazici, G. Inan, V. Tabak, Effect of aspect ratio and volume fraction of steel
fiber on the mechanical properties of SFRC, Constr. Build. Mater. 21 (2007)
1250–1253.
[22] A. Katz, A. Bentur, Effect of matrix composition on the aging of CFRC, Cem.
Concr. Compos. 17 (1995) 87–97.
[23] S. Bernal, R. de Gutiérrez, E. Rodríguez, S. Delvasto, F. Puertas, Mechanical
behaviour of steel fibre-reinforced alkali activated slag concrete, Mater.
Constr. 59 (2009) 53–62.
[24] S.W. Kim, S.K. Jang, D.H. Kang, K.L. Ahn, H.D. Yun, Mechanical properties and
eco-efficiency of steel fiber reinforced alkali-activated slag concrete, Materials
8 (2015) 7309–7321, http://dx.doi.org/10.3390/ma8115383.
[25] S. Bernal, R. De Gutierrez, S. Delvasto, E. Rodriguez, Performance of an alkali-
activated slag concrete reinforced with steel fibers, Constr. Build. Mater. 24
(2010) 208–214, http://dx.doi.org/10.1016/j.conbuildmat.2007.10.027.
[26] S. Aydin, B. Baradan, The effect of fiber properties on high performance alkali-
activated slag/silica fume mortars, Compos Part B 45 (2013) 63–69, http://dx.
doi.org/10.1016/j.compositesb.2012.09.080.
[27] J.I. Choi, K.I. Song, J.K. Song, B.Y. Lee, Composite properties of high-strength
polyethylene fiber-reinforced cement and cementless composites, Compos.
Struct. 138 (2016) 116–211, http://dx.doi.org/10.1016/
j.compstruct.2015.11.046.
[28] S.J. Choi, J.I. Choi, J.K. Song, B.Y. Lee, Rheological and mechanical properties of
fiber-reinforced alkali-activated composite, Constr. Build. Mater. 96 (2015)
112–118, http://dx.doi.org/10.1016/j.conbuildmat.2015.07.182.
[29] F. Puertas, T. Amat, A. Fernández-Jiménez, T. Vázquez, Mechanical and durable
behaviour of alkaline cement mortars reinforced with polypropylene fibres,
Cem. Concr. Res. 33 (2003) 2031–2036.
[30] F. Puertas, A. Gil-Maroto, M. Palacios, T. Amat, Alkali-activated slag mortars
reinforced with ar glassfibre. Performance and properties, Mater. Constr. 56
(2006) 79–90.
71
[31] J.S. Alcaide, E.G. Alcocel, F. Puertas, R. Lapuente, P. Garcés, Carbon fibre-
reinforced, alkali-activated slag mortars, Mater. Constr. 57 (2007) 33–48.
[32] S. Vaidya, E.N. Allouche, Strain sensing of carbon fiber reinforced geopolymer
concrete, Mater. Struct. 44 (2011) 1467–1475, http://dx.doi.org/10.1617/
s11527-011-9711-3.
[33] P. He, D. Jia, T. Lin, M. Wang, Y. Zhou, Effects of high-temperature heat
treatment on the mechanical properties of unidirectional carbon fiber
reinforced geopolymer composites, Ceram. Int. 36 (2010) 1447–1453, http://
dx.doi.org/10.1016/j.ceramint.2010.02.012.
[34] F.J. Baeza, O. Galao, E. Zornoza, P. Garcés, Effect of aspect ratio on strain sensing
capacity of carbon fiber reinforced cement composites, Mater. Des. 51 (2013)
1085–1094.
[35] F.J. Baeza, D.D.L. Chung, E. Zornoza, L.G. Andión, P. Garcés, Triple percolation in
concrete reinforced with carbon fiber, ACI Mater. J. 107 (2010) 396–402.
[36] F.J. Baeza, O. Galao, E. Zornoza, P. Garcés, Multifunctional cement composites
strain and damage sensors applied on reinforced concrete (RC) and structural
elements, Materials 6 (2013) 841–855.
[37] F.J. Baeza, E. Zornoza, L.G. Andión, S. Ivorra, P. Garcés, Variables affecting strain
sensing function in cementitious composites with carbon fibers, Comput.
Concr. 8 (2011) 229–241.
[38] J. Gomis, O. Galao, V. Gomis, E. Zornoza, P. Garcés, Self-heating and deicing
conductive cement. Experimental study and modeling, Constr. Build. Mater. 75
(2015) 442–449, http://dx.doi.org/10.1016/j.conbuildmat.2014.11.042.
[39] O. Galao, L. Bañón, F.J. Baeza, J. Carmona, P. Garcés, Highly conductive carbon
fiber reinforced concrete for icing prevention and curing, Materials 9 (4)
(2016) 281, http://dx.doi.org/10.3390/ma9040281.
[40] E. Zornoza, G. Catalá, F. Jiménez, L.G. Andión, P. Garcés, Electromagnetic
interference shielding with Portland cement paste containing carbon materials
and processed fly ash, Mater. Constr. 60 (2010) 21–32.
[41] A. Pérez, M.A. Climent, P. Garcés, Electrochemical extraction of chlorides from
reinforced concrete using a conductive cement paste as an anode, Corros. Sci.
52 (2010) 1576–1581.
[42] C.Y. Tuan, S. Yehia, Evaluation of electrically conductive concrete containing
carbon products for deicing, ACI Mater. J. 101 (4) (2004) 287–293.
[43] B. Chen, K. Wu, W. Yao, Conductivity of carbon fiber reinforced cement based
composites, Cem. Concr. Compos. 26 (2004) 291–297.
[44] M. Chiarello, R. Zinno, Electrical conductivity of self-monitoring CFRC, Cem.
Concr. Compos. 27 (2005) 463–469.
[45] J.L. Vilaplana, F.J. Baeza, O. Galao, E. Zornoza, P. Garcés, Self-sensing properties
of alkali activated blast furnace slag (BFS) composites reinforced with carbon
fibers, Materials 6 (2013) 4776–4786, http://dx.doi.org/10.3390/ma6104776.
[46] UNE-EN 196-6-2010, Methods of Testing Cement – Part 6: Determination of
Fineness, AENOR.
[47] UNE 83811:1992 EX, Mortars. Test Methods. Fresh Mortars. Determination of
Consistence. Flow Table (Reference Method), AENOR.
[48] UNE-EN 196-1:2005, Methods of Testing Cement – Part 1: Determination of
Strength, AENOR.
[49] UNE-EN 12504-4:2006, Testing Concrete: Part 4: Determination of Ultrasonic
Pulse Velocity, AENOR.
[50] UNE-EN-1015-10:2000/A1:2007, Methods of Test for Mortar for Masonry –
Part 10: Determination of Dry Bulk Density of Hardened Mortar, AENOR.
[51] UNE-80112:1989 EX, Test Methods of Cements. Physical Analysis.
Determination of Contraction in Air and Expansion in Water, AENOR.
[52] A. Fernández-Jiménez, J.G. Palomo, F. Puertas, Alkali-activated slag mortars:
mechanical strength behavior, Cem. Concr. Res. 29 (1999) 1313–1321, http://
dx.doi.org/10.1016/S0008-8846(99)00154-4.