Accepted Manuscript

Electrochemical corrosion monitoring of steel fiber embedded in cement based

composites

Ahsanollah Beglarigale, Halit Yazici

PII: S0958-9465(16)30105-6
DOI: 10.1016/j.cemconcomp.2017.08.004
Reference: CECO 2882

To appear in: Cement and Concrete Composites

Received Date: 28 April 2016

Revised Date: 24 February 2017
Accepted Date: 7 August 2017

Please cite this article as: A. Beglarigale, H. Yazici, Electrochemical corrosion monitoring of steel fiber
embedded in cement based composites, Cement and Concrete Composites (2017), doi: 10.1016/
j.cemconcomp.2017.08.004.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Electrochemical corrosion monitoring of steel fiber embedded in cement based composites

Ahsanollah BEGLARIGALE* and Halit YAZICI

Department of Civil Engineering, Eng. Faculty, Dokuz Eylül University, Buca-İzmir, Turkey

Abstract

PT
Steel fibers are commonly used in cement based materials for many applications such as floors,

structural elements, repairing works, etc. The chloride-induced corrosion via ingress of seawater

RI
may become a risk for performance of the steel fiber reinforced cement based composites. Despite

of few studies that have dealt with the corrosion behavior of steel fibers embedded in cement based

C
composites, there are lack of information about the chloride-induced corrosion of steel fibers

US
embedded with sufficient cover in non-cracked matrixes, reactive powder concrete, and especially
AN
polymer-modified cement based mortars. The open-circuit potential and corrosion current density of

single steel fibers embedded in various cement based matrices were monitored after 200, 400, 600,
M

1200 wetting-drying cycles in 3.5% NaCl solution. The corrosion and microstructure analyses

revealed that the steel fibers can be protected by a well-designed mixture, non-cracked matrix and
D

sufficient cover. However, it could be corroded in long terms depending on the type and the dosage
TE

of the polymer latex used. In addition, the residual stresses in the deformed regions of the hooked-
EP

end steel fibers is critical in terms of the protection against to chloride-induced corrosion.

Keywords: Steel fiber, chloride-induced corrosion, wetting-drying cycles, cement based composites
C
AC

1. Introduction

Unreinforced cement based materials are known as brittle construction products with low tensile

strength. It is very well known that this drawback of cementitious materials can be mitigated by

incorporating different types of discontinuous fibers. Steel fibers are most commonly used in

cement based materials for applications such as pavements, slabs and also many structural elements

*
Corresponding author. Tel.: +90 232 3017044; Fax: +90 232 3017253.
E-mail address: ahsan.beglari@gmail.com.
1
 ACCEPTED MANUSCRIPT
[1].

The corrosion via ingress of seawater, which includes chloride ions, may become a risk for

performance of steel fiber reinforced cement based composites (SFRCC). A few scientific

researches revealed that the corrosion of steel fiber is much less harmful as compared with

traditional steel rebar reinforcement [2-4]. However, the fiber corrosion can affect negatively the

PT
crack bridging capacity of fiber with detrimental consequences in terms of the performance of

SFRCC structures [5]. If the fiber-matrix bond was affected with detrimental consequences by the

RI
corrosion of steel fiber, the properties of cementitious composites would be affected negatively

either. Banthia and Foy [6] reported that curing of SFRCC in high temperature seawater promotes

C
an early fiber corrosion that leads to substantial reductions in pull-out resistances. On the other

US
hand, Frazão et. al. [6] reported that the fiber-matrix bond strength of cracked self-compacting
AN
concrete, which had been subjected to 10-day 3.5% NaCl solution, increased due to the corrosion of

steel fiber. They declared that the corrosion products have increased the fiber surface roughness
M

which improved fiber-matrix bond characteristics [7].

The chloride corrosion of steel fibers, which is the most important concern in the case of the marine
D

structures, can be observed in three aspects. a) Surface corrosion: The fiber corrosion at surface of
TE

SFRCC was described by Balouch et. al [2] as appearance of rust spots at the surfaces of exposed
EP

SFRCC. b) Corrosion from the cracks: Granju and Balouch [1] stated the corrosion of steel fiber

from cracks may causes bursting of matrix due to swelling of corrosion products and also reduction
C

of fiber cross section which can affects the crack-bridging capacity of the fibers with detrimental
AC

consequences. However, the experimental research conducted by them revealed small sensitivity of

cracked SFRCC to corrosion after one year exposing to marine-like environment (salted fog) [1]. c)

Corrosion of steel fiber through the ingress of chloride ions into matrix: This issue, which is in the

scope of this research, is related directly with the properties of cementitious matrix and the location

of fiber (concrete cover). Depending on the characteristics of matrix, it seems that the corrosion of

steel fiber via ingress of chloride ions into matrix may be less harmful as compared to the cracked

2
 ACCEPTED MANUSCRIPT
matrix one. However, it can be dangerous in long terms due to the invisible and unrepairable

deterioration process of that corrosion mechanism. For example, Banthia and Foy [6] declared that

exposing concrete to seawater in early ages can lead to an easy ingress of corrosive agents to matrix

which has not achieved proper impermeability. Yalçınkaya and Yazıcı [8] reported major chloride-

induced corrosion of steel microfibers in the chloride-penetrated area of self-compacting repair

PT
mortar exposed to freeze-thaw cycles in sodium chloride solution. In the scope of this study

different types of polymer-modified cement based mortars, which can be used in many repair works

RI
such as restoration of marine structures, were investigated. Li et. al. [9] reported that flexural

strength of steel fiber reinforced concrete can be improved by incorporating of the styrene butadiene

C
rubber emulsions. In addition, the adhesiveness of polymer-modified cement based materials to

US
other materials is higher than the conventional mortar due to the formation of a film in the interface
AN
with other materials [10].
M

There are many different electrochemnical methods used for characterization of reinforcement

corrosion. Andrade and Alonso (1996) [11] categorized the methods as follows; a) Measurement
D

methods: Polarization curves [7, 12, 13], polarization resistance [13] and electrochemical
TE

impedance spectroscopy [13-15] b) Non-stationary methods: Galvanostatic [16] pulse and

EP

electrochemical noise [17]. Mott-Schottky analysis [18-20], cement-based piezoelectric composite

sensors and the acoustic emission can also be used for this purpose [21]. In addition, a novel test
C

method was proposed by Hong et al. [22] to monitor reinforcement corrosion using Ground-
AC

Penetrating Radar. Moreover, techniques such as corrosion potential, resistivity measurements and

galvanic current are categorized as the complementary test methods [11]. In this study, the open-

circuit potential measurement and polarization curves were used for corrosion monitoring of steel

fibers. Frazão et. al. [7] also used these techniques for characterization of fiber corrosion.

This research is a part of a wider project which has dealt with the pull-out behavior of steel fiber

embedded in flowable reactive powder concrete and normal mortar [23] and also the effect of

3
 ACCEPTED MANUSCRIPT
chloride-induced corrosion on fiber-matrix bond characteristics. This study aims to analyze the

chloride-induced corrosion of steel fiber embedded with sufficient concrete cover in non-cracked

cement based matrices.

1.1 Research significance

Despite of a few studies related to the corrosion behavior of steel fibers embedded in cement based

PT
composites [1, 2, 6, and 7], these scientific research works have dealt with the cementitious

composites with cracked matrices and insufficient concrete cover or insufficient curing. In addition,

RI
there are lack of information about the chloride-induced corrosion of steel fiber embedded in

reactive powder concrete and polymer-modified cement based mortars or the SFRCC protected by

C
these cementitious materials. This study aims to fill above-mentioned gaps in the literature.

US
AN
2. Experimental

1.1 Materials
M

CEM I 42,5R ordinary portland cement with 371 m2/kg Blaine fineness was used. The silica fume

used in the flowable reactive powder concrete (R) mix design had 92.25% SiO2 content and 20,000
D

m2/kg nitrogen absorption fineness. Crushed limestone with maximum size of 5 mm was used in the
TE

normal mortar (N) mixtures, while four different sizes (1-3, 0.5-1, 0-0.4, and 0-0.075 mm) of quartz
EP

aggregate were used in the R mixtures. Polycarboxylic ether based superplasticizer was used to

reach the target workability. A single type of steel fiber with 1.05 mm diameter, 50 mm length, and
C

48 aspect ratio was used in the corrosion tests.

AC

Two types of the styrene-butadiene rubber (SBR1 and SBR2) and also two types of the acrylic

dispersion polymer (ADP1 and ADP2) which are commonly used in the construction industry were

used in this study. The chemical composition of waterproofing (WP) admixture used in this study is

an aqueous colloidal blend of inorganic silicate and fatty acids. Additionally, an alkanolamine

(nitrogen containing organic and inorganic substances) based corrosion inhibitor (CIN) were used in

this research.

4
 ACCEPTED MANUSCRIPT
1.2 Mixtures

N mixture with 0.5 water to binder ratio (W/B) and R mixture with 0.2 W/B ratio were selected as

the cementitious matrices. Maximum dosage recommended by the producers for SBR and ADP

admixtures are 5% and 1.5% by weight of cement content, respectively. The SBR admixtures were

added to the N and R mixtures by 5% and 15% weight of cement content. These ratios were 1.5%

PT
and 5% for ADP modified mixtures.

The producer recommended dosage of CIN was 12 kg/m3, while it was 3% of cement content (by

RI
weight) for the WP admixture. The producers recommended dosages were selected for both

admixtures. It must be noted that the water content of admixtures was deducted from the N and R

C
mixtures water content.

US
Cement hydration and polymer film formation processes are the main mechanisms responsible for
AN
binding phase of the polymer-modified mixtures. It has been declared by many researchers that the

mixed cure (water and air) is more effective for polymer-modified mixtures [24-26]. It can be
M

explained by the easier polymer film formation and the higher degree of cement hydration under

this curing condition [24]. For this purpose, the specimens containing SBR, ADP, and WP were
D

cured for 7 days in water and 22 additional days in air (Mix curing). Therefore, two groups of
TE

control specimens were prepared; a) Control specimens which were cured for 7 days in water and
EP

22 additional days in air (mix curing). b) Control specimens which were cured for 28 days in

20±2°C water (standard curing). This group of control specimens were prepared in order to
C

compare with the mixtures containing CIN admixture.

AC

Mixtures names were abbreviated throughout the paper according to their matrix type, curing

condition etc. (Table 2). The mix designs are presented in Table 3.

1.3 Specimens

All test batches were mixed by using a Hobart type mixer. Initially, binder and aggregate were

mixed in a dry state and then half of the mixing water was gradually added while mixing continued.

It should be noted that chemical admixtures were mixed with water. The mini-slump flow value of

5
 ACCEPTED MANUSCRIPT
N mixture was 150±5 mm, while the R mixture was an almost self-leveling mixture with a free flow

value of 220±5 mm (Fig. 1). The mixtures prepared for the corrosion test were poured into

50×50×50 mm cubic molds in two layers. After placing final layer, single steel fiber was centrally

embedded into the fresh mixture by an apparatus that allowed the fiber becomes perpendicular to

the surface of the specimen and adjusts the desired embedment length (25 mm) into the matrix and

PT
then, vibration was applied. In addition, the mechanical properties and water absorption tests were

determined on three 40×40×160 mm prismatic specimens.

RI
After casting, specimens were kept for 24h in 20 ±2°C and, after that, were demolded and then,

C
were cured in water. For corrosion tests, an electric wire was mounted on the free end of the fibers

which have been centrally embedded in the 50×50×50 mm cubic specimens. The upper surface of

US
specimens and free end of the fibers were carefully sealed by four layers of epoxy coating. The
AN
schematic figure of corrosion test specimen and its details are presented in Fig. 2. In addition, the

epoxy-coated free end of the fibers was also sealed with silicone coating for extra protection against
M

any possible crack formation in the brittle epoxy-coating (Fig 3a). The corrosion specimens were

subjected to wetting-drying cycles in 3.5% NaCl solution. The wetting-drying machine, which can
D

be seen from Fig. 3b, was adjusted to wet the specimens for 120 min and then, to dry them for 180
TE

min (20°). After 200, 400, 600, 1200 wetting-drying cycles, the corrosion process of specimens was
EP

monitored by using polarization technique. It should be noted that 8 specimens were used for each

mixture (192 specimens). Four specimens for each mixture were analyzed by the corrosion tests
C

after 200, 400, 600 cycles, and then were subjected to pull-out test. The pull-out test results will be
AC

reported in a detailed paper in future. The other four specimens (96 specimens) were kept in the

wetting-drying machine for 1200 cycles (Fig. 2c). These specimens were subjected to the corrosion

tests after 1200 cycles, and then were split to check the surface of the embedded steel fibers and the

fiber-matrix interfaces.

1.4 Evaluation of the hardened properties of mixtures

6
 ACCEPTED MANUSCRIPT
The water absorption were obtained through the following procedure: Specimens were first kept at

105°C until obtain a constant weight (W1). After cooling, they were immersed in water for 48 h to

achieve the saturated surface dry condition, and then weighed again (W2). The water absorption of

mixtures (WA) was calculated according to Equations 1.

−
( )= × 100 [ . 1]

PT
The dry bulk density (DBD) and porosity (P) values were determined through the Equations 2 and

Equations 3, respectively.

RI
( )= [ . 2]

C
US
( )= 1− × 100 [ . 3]

Where V is the volume of the specimens and γ is the saturated surface drying density of the
AN
specimens.

The compressive strength tests were applied on the two pieces (40×40 mm2 loading area) that were
M

left from the three point flexural tests. Therefore, the flexural and compressive strengths test results
D

are the average of three and six specimens, respectively.

TE

1.5 Characterization of steel fiber corrosion

The open-circuit potential (EOCP) and corrosion current density (icorr) values of the specimens were
EP

obtained using a computer-controlled potentiostats though a three-electrode system (Fig. 4). The

EOCP measurements were carried out for at least 50 min to obtain a steady state condition.
C
AC

Immediately after EOCP measurement the potentiodynamic polarization curves were obtained in the

applied potential range from ±200 mV with scan rate of 0.2 mV/s. The ±200 mV, which were

applied to obtain appropriate anodic and cathodic polarization curves, can be considered as a high

potential range for a steel fiber. However, any damage was not observed on the fiber. As discussed

earlier, there was 8 specimens for each mixture. Four specimens were analyzed by the corrosion

tests after 200, 400, 600 cycles, and then were subjected to pull-out test. In spite of the fact that

these specimens (96 specimens) were subjected to the test after 200, 400, and 600 cycles, there was
7
 ACCEPTED MANUSCRIPT
not any damage on the fibers. The other four specimens were kept in the wetting-drying machine

for 1200 cycles (Fig. 2c). These 96 specimens were subjected to the corrosion tests only after 1200

cycles, and then were split for observation of the embedded steel fibers and the fiber-matrix

interfaces. The fibers were carefully monitored by a digital optical microscope. Except of the

chloride-induced corrosion on the fiber of some of the N mixtures, there was not any corroded

PT
resign or pit on the fiber of other N mixtures. Besides, no matter which mixtures, there were not any

measured or observed damage in the R specimens after 1200 cycles.

RI
The corrosion current density has been analyzed through the Tafel analysis. The current density can

be expressed as a function of the overpotential, η , where η=Eapplied - Eopen circuit as follows [27];

C
η = ß log
!
US [ . 4]
AN
This expression is known as the Tafel equation, where ß is the Tafel slope, i the applied current
M

density,and i0 the exchange current density [27]. Firstly, the Icorr was obtained by extrapolating the
D

linear regions of polarization curves (plot of potential versus logarithmic current) back to their
TE

intersection. The current density icorr (µA/cm2) were calculated by dividing the Icorr value to the

surface area of embedded portion of steel fiber (0.832904625 cm2).

EP

The working electrode is the steel fibers embedded in cementitious matrices. Saturated calomel
C

electrode (SCE) and graphite was used as the reference electrode and counter electrode,
AC

respectively. It must be noted that the electrolytic environment of three-electrode system was the

3.5% NaCl solution.

1.6 Microstructure analysis

The optical microscope images were taken by a prototype digital optical microscope. The

microstructure were studied by using JEOL JSM 6060 electron microscope. The samples for

Scanning electron microscopy (SEM) analysis were prepared by taking small pieces from the fiber

8
 ACCEPTED MANUSCRIPT
and fiber-matrix interface zone. Original microstructure and morphology of the products were

observed on fractured surfaces using secondary electron imaging. Fractured small samples were

mounted on the SEM stubs using carbon paint, and then, the samples were coated with gold. Energy

dispersive spectrometry (EDS) analyses given here are the average of three analyses result. Limited

micrographs are presented here to give only an opinion about microstructure of the specimens.

PT
3. Results and discussion

3.1 Fresh and hardened properties of mixtures

RI
As mentioned earlier the mini-slump flow value of N mixtures was 150±5 mm, while R was an

almost self-leveling mixture with a free flow value of 220±5 mm. Generally, polymer based

C
admixtures improve the workability of the cement based composites as a result of air entraining and

US
friction reducing effects of them [28]. For example, Fig. 4 shows the SEM images of the A2-N and
AN
A2-R specimens which have considerable amount of entrained air voids. As can be seen from Fig. 5

the polymer latex emulsions reduced the superplasticizer demand by 16% – 34% depending on the
M

polymer/cement dosage, latex type, and matrix type (N or R). It must be noted that both of the CIN

and WP admixtures did not affect the superplasticizer demand. The superplasticizer dosage
D

reduction effect of ADP was more than SBR one. This behavior can be explained by the higher air
TE

entraining effect of acrylic based polymers than SBR. It can be seen from Fig. 5 that the
EP

superplasticizer dosage reduction effect of polymer latexes was much more obvious in N mixtures

as compared to the R ones. The polymer-modified mixtures were also very sticky. This behavior is
C

one of the positive effects of polymers in the case of using them in repair works.
AC

Dry bulk density values of all mixtures are presented in Fig. 6. As can be seen from the figure the

dry bulk density of N mixture was decreased by using polymers and WP admixtures depending on

their type and dosage. It can be explained by the air entraining effect of these admixtures and also

low specific gravity of the solid content of the admixtures. Meanwhile, the CIN admixture did not

affect the dry bulk density value of N mixture significantly. The decrements in the bulk density

values were also valid in the case of R mixtures. As can be seen from SEM images presented in Fig.

9
 ACCEPTED MANUSCRIPT
7, the R mixture containing 5% ADP2 (R2-R) has many air voids as compared to the control R

mixture. In addition, CIN admixture increased the bulk density of R mixture.

The porosity (P) and water absorption (WA) values of mixtures are presented in Figs. 8 and 9. It

can be seen from the figures that the CIN did not affect the P and WA values of N mixtures, while it

decreased these values of R by more than 50%. The P and WA of the N mixture containing

PT
polymers were higher than the control mixture. However, the P and WA values of B1-R, L2-R, R2-

R mixtures were lower than the R control mixture. Although the WP admixture decreased the P and

RI
WA values of R mixture, it affects negatively these properties of the N mixture. As can be seen

from Fig. 10 there is a strong correlation between the porosity and water absorption values of all N

C
and R mixtures.

US
The compressive and flexural strengths of N and R mixtures are presented in Figs. 11 and 12,
AN
respectively. It can be seen from figures that the compressive and flexural strengths was not

affected significantly by using CNI (F-N and F-R mixtures) as compared to the control mixtures
M

(CN-N and CN-R).

The WP admixture decreased the compressive and flexural strength of N due to the excessive
D

entrained air voids that can be seen from the SEM images presented in Fig. 13a. In the case of R
TE

mixture, there was not any remarkable difference between the mechanical strengths of CM-R and
EP

SI-R (Fig. 12). As can be seen from Fig. 13b there is not considerable amount of the entrained air

voids in the SI-R matrix as compared to the A2-R and R2-R ones (Figs. 4b and 7b), respectively.
C

The compressive strength of the N mixtures containing SBR admixtures were decreased as
AC

compared to the control mixture due to the air entraining effect of SBR admixtures, while there is

an increment in flexural strength of the N mixtures containing SBR1 (A1-N and A2-N). This

behavior is in accordance with the related literature. The compressive strength of polymer-modified

cement based composites can be decreased as a result of the air entraining effect of polymer based

admixtures [25, 29, and 30], while they can have a positive effect on the flexural strength of

concrete and mortar due to the better bond strength between aggregate and matrix as a result of

10
 ACCEPTED MANUSCRIPT
polymer film formation (28 29, and 31). However, there are significant decreases in flexural

strength of R as much as the compressive strength. Furthermore, there was a massive Ca(OH)2

crystals formation in the N mixtures containing SBR admixtures, especially in air voids (Fig. 14a

and 14b). This phenomenon did not occur in the R mixtures (Fig. 14c).

The ADP admixtures decreased the compressive and flexural strengths of N mixture by 3-11% and

PT
12-17%, respectively. It can be noted that the 1.5% ADP1 had the minimum negative effect on the

mechanical strength of N mixture. As can be seen from Fig. 15, the N mixture containing %5 ADP

RI
admixture has many entrained air voids which include many Ca(OH)2 crystals. No matter the

dosage and type of ADP admixture used, the flexural strength of the R mixture decreased by 30 ±2

C
%. However, incorporating 1.5 % acrylic based polymers (L1-R and R1-R) increased the

US
compressive strength of R mixture by approximately 5%.
AN
Corrosion monitoring of steel fibers

The average open-circuit potential (EOCP) and corrosion current density (icorr) values are presented
M

in Table 3. The EOCP, which was measured relative to the SCE as the reference electrode, is a

parameter to predict the corrosion potential of reinforcing steel in concrete [32]. ASTMC876 [32]
D

recommends that there is a greater than 90 % probability of corrosion in the case of potential less
TE

than −276 mV. A corrosion potential more positive than −120 mV shows that there is a greater than
EP

90 % probability that no corrosion is occurring. Also, an EOCP value between −120 mV to −276 mV

represents uncertain corrosion activity [32]. Fig. 16 presents a few of the EOCP measurements
C

executed after 1200 cycles. As can be seen from Fig. 16 a steady state condition for EOCP was
AC

obtained even earlier than 50 min. In addition, Fig. 17 shows some of the potentiodynamic

polarization curves that were used to calculated icorr values. For this purpose, Icorr values (µA),

which obtained from Tafel analysis, were divided by the surface area of embedded steel fiber.

According to the EOCP values, there is a greater than 90 % probability that the steel fibers embedded

in some of mixtures were not corroded at 200, 400 and even 600 wetting-drying cycles (Table 3).

However, many of the EOCP values are in the uncertain corrosion activity zone. It should be noted

11
 ACCEPTED MANUSCRIPT
that the EOCP values of steel fibers, no matter embedded in which mixture, did not exceed the −276

mV value up to 600 wetting-drying cycles. In spite of some of the EOCP values that are in the

uncertain corrosion activity zone, it can be seen from Fig. 17 that the icorr values of them were lower

than the 0.1 µA/cm2 which indicates negligible corrosion level [33]. It can be concluded from test

results of the specimens subjected to 600 wetting-drying cycles in 3.5% NaCl solution that the

PT
parameters such as un-cracked matrix used in this study and the location of the fiber in that matrix

are more predominant than the properties of the cementitious matrix. Balouch et. al. [2] reported

RI
that decreasing W/C ratio from 0.78 to 0.5 leads to decrease the minimum necessary cover to

prevent surface corrosion from 1 mm to 0.1-0.2 mm. However, they declared that further decrease

C
of W/C does not reduced the minimum necessary cover. On the other hand, it can be seen from the

US
study conducted by Frazão et. al. [7] that the steel fiber with a pre-cracked concrete cover (1 cm)
AN
can be remarkably corroded even at 10-day 3.5% NaCl solution without any wetting-drying cycles.

As discussed earlier the specimens (4 specimens for each mixture) that were analyzed by the
M

corrosion tests after 200, 400, 600 cycles, were subjected to pull-out test. Limited results of the pull-

out test were reported in a conference paper [34]. After 600 cycles the fibers, which had been pulled
D

out from matrixes, were monitored by digital optical microscope. There were not any corrosion
TE

products found on the surface of the embedded steel fibers and the fiber-matrix interfaces. These
EP

findings are in accordance with the corrosion test results. In addition, the pull-out behavior of the

specimens subjected to wetting-drying cycles and the specimens that were stored in water up to the
C

age of 600-cycle were similar to each other [34]. The pull-out test results will be reported in a
AC

detailed paper in future.

As can be seen from Fig. 18 and Table 3 the EOCP and icorr values of some of the N mixtures

increased at 1200 cycles. The average icorr value of steel fiber embedded in the L2-N mixture is

2.825 µA/cm2 which shows a high corrosion level. Corrosion current density higher than 1 µA/cm2

indicates high corrosion level [33]. Moreover, the EOCP value exceeded the -270 mV value which

had already revealed that there is a greater than 90 % probability that corrosion is occurring in the

12
 ACCEPTED MANUSCRIPT
steel fiber embedded in L2-N mixture. It must be mentioned that the porosity and water absorption

values of the L2-N mixture is, respectively, 49% and 70% higher than the control mixture (CM-N)

ones. Fig 19. shows the digital microscope images of the steel fiber embedded in the L2-N mixture

after 1200 wetting-drying cycles. As can be seen from figures the corrosion occurred at the hooked

regions of the fibers most likely due to the residual stresses at these locations. This phenomenon

PT
was also reported by the Banthia and Foy [6]. They declared the residual stresses at hooked regions

cause generating different electrochemical potential than the rest of the steel fiber which leads to an

RI
anodic formation. In addition, pitting corrosion can be observed at this region (Fig.19). As can be

seen from Fig. 20 the corrosion products exist throughout of the fiber-matrix interface of the hooked

C
regions. Massive corrosion products with different morphology can be observed at the hooked

US
region of the steel fiber embedded in the L2-N mixture (Fig. 21). However, the corrosion products
AN
were decreased with distance from the hook (Fig. 22).

It can be seen from Fig. 18 and Table 3 that the EOCP values of steel fibers embedded in the R2-N
M

and SI-N mixtures exceeded -270 mV after 1200 wetting-drying cycles similar to the L2-N mixture.

However, the average icorr values of steel fiber embedded in the R2-N and SI-N mixtures are,
D

respectively, 0.584 µA/cm2 and 0.511 µA/cm2 which show a moderate corrosion level. Corrosion
TE

current density between 0.5 to 1 µA/cm2 indicates moderate corrosion level [33]. For example, as
EP

can be seen from Fig. 23 local corrosion occurred at the hooked region of the fiber embedded in the

R2-N mixture. As mentioned earlier, this type of corrosion, which is more obvious in Fig. 23, can
C

occur as a result of the different electrochemical potential at the hooked region due to the residual
AC

stresses. In addition, Fig. 23 shows cracks on the surface of hooked region of steel fiber surface that

may be occurred by the expansive products of corrosion. These kinds of cracks and corrosion types

on the hooked regions of steel fibers can affect adversely the pull-out behavior of the steel fibers. In

other words, fiber rupture can occur from these regions during the pulling out of the steel fibers

from the cementitious composite. It seems that, besides of the factors such as the properties of

cementitious mixture or the thickness of concrete cover, the chloride-induced corrosion resistance

13
 ACCEPTED MANUSCRIPT
of the hooked-end steel fibers is also depend on the residual stresses induced in them at the

manufacturing stage.

According to the analysis conducted at 1200 wetting-drying cycles, the average icorr values of steel

fibers embedded in the B1-N and R1-N mixtures are in the range of 0.1 to 0.5 µA/cm2 which

indicate a low corrosion level. However, The EOCP values of them are in the uncertain corrosion

PT
activity zone. On the other hand, the average icorr values of steel fibers embedded in the CS-N, CM-

N, A1-N, A2-N, B2-N, L1-N, and F-N mixtures are lower than the 0.1 µA/cm2 value (negligible

RI
corrosion level). It must be noted that the average EOCP value of steel fibers embedded in the above-

mentioned mixtures are also in the uncertain corrosion activity zone. The corrosion test results were

C
in accordance with the digital optical microscope analyses that were conducted on the surface of the

US
steel fibers and the fiber-matrix interfaces.
AN
As can be seen from the corrosion test results at 1200 wetting-drying cycles, both of the EOCP and

icorr values of steel fibers embedded in any of the R mixtures, no matter which polymer or admixture
M

have been used in them, did not exceed the critical values. Whereas, as mentioned earlier, some of

admixture used in this study had negative effects on the properties of R. The EOCP and icorr values of
D

steel fibers embedded in the R mixtures are lower than, respectively, -270 mV and 0.1 µA/cm2
TE

values. The superior resistance of R-series specimens against chloride corrosion can be explained
EP

by their very low permeability and porosity (0.5-1.8%) owing to the factors such as the low

water/binder ratio and high binder (cement and SF) content of R mixtures. In addition, it is well
C

known that silica fume improves the chloride penetration resistance of concrete [35-37]. Besides,
AC

the filler effect of quartz powder and silica fume should not be ignored. In accordance with the

corrosion test results, digital microscope, SEM, and EDS analyses revealed that there were not any

corrosion products in the fiber matrix interface of R mixtures and also on the surface of the steel

fibers. For example, the SEM and EDS analysis performed on the steel fibers embedded in the B2-R

mixture are presented in Fig. 24 to give a picture of microstructure of non-corroded fibers that have

been protected by the R matrixes during the 1200 wetting-drying cycles. As can be seen from the

14
 ACCEPTED MANUSCRIPT
SEM images the steel fiber has a rough microstructure (Fig. 24). Also, the EDS analysis shows the

adhered cement hydrates on the rough regions of the steel fibers. In addition, Fig. 25 shows the

fiber-matrix interface of the B2-R mixture without any corrosion products in it. It should be noted

that further cycles are necessary to determine the effect of corrosion inhibitor admixture.

PT
4 Conclusions

Based on the test results of this study, the following conclusions can be drawn:

RI
• Steel fibers, which had been embedded with sufficient cover in the non-cracked normal

mortar specimens, could be protected against chloride-induced corrosion by a well-designed

C
mixture with and without polymer latex admixtures. However, the electrochemical and

US
microstructural analyses conducted at long-term (1200 wetting-drying cycles) revealed fiber
AN
corrosion in some of mixtures containing various admixtures. It is obvious that corrosion

resistance of these mixtures depends on the type and dosage of the admixtures.
M

• The steel fibers embedded in the flowable reactive powder concrete, no matter which
D

admixture used in them, were not corroded during the 1200 wetting-drying cycles in
TE

artificial seawater. It can be explained mainly by the low porosity of these mixtures that
EP

leads to a superior protection against chloride-induced corrosion of the embedded steel

fibers. However, further investigation is needed for fibers with a thinner concrete cover.
C
AC

• Any residual stresses in the deformed regions of hooked-end steel fibers is critical as much

as the factors such as the properties of cementitious mixture or the thickness of concrete

cover. The chloride-induced corrosion risk of the hooked-end steel fibers can be increased

due to the residual stresses generated at the manufacturing stage of steel fibers. This type of

corrosion on the hooked regions of steel fibers can lead to fiber rupture from these regions

due to the mechanical loadings.

15
 ACCEPTED MANUSCRIPT
• The electrochemical test results were in accordance with the optical microscopy, SEM and

EDS analyses that were conducted on the surface of embedded steel fibers and the fiber-

matrix interfaces.

PT
RI
Acknowledgement

C
This study is a part of the TÜBİTAK MAG project No: 112M598. The author acknowledges to

US
TÜBİTAK for their support and financial aids. Special thanks are also given to Assist. Prof. Dr.

Hüseyin YİĞİTER for his support throughout of this study. The authors also are grateful to Mr.
AN
Mehmet YERLIKAYA from Bekaert (Turkey), Mr. Okan DUYAR from BASF (Turkey), Mr. Ümit

ÜRÜN from Pomza Export, and Modern Beton for their material supports.
M
D

References
TE

[1] J.L. Granju, S.U. Balouch, Corrosion of steel fibre reinforced concrete from the cracks, Cem.

Concr. Res. 35 (2005) 572– 577.

EP

[2] S.U. Balouch, J.P. Forth, J.-L. Granju, Surface corrosion of steel fibre reinforced concrete, Cem.

Concr. Res. 40 (2010) 410–414.

C

[3] G.T. Halvorsen, C.E. Kesler, A.R. Robinson, J.A. Stout, Durability and physical properties of
AC

steel fiber reinforced concrete, Report n8 DOTTST 76T-31, US. Department of transportation,

Federal railway administration, Washington, DC, 1976.

[4] D.C. Morse, G.R. Williamson, Corrosion behaviour of steel fibrous concrete, Technical report,

M-217, US Army Corps of Engineers, Construction Engineering Research Laboratory, Chaimpaign,

IL, 1977.

16
 ACCEPTED MANUSCRIPT
[5] C. Frazão, A. Camões, J. Barros, D. Gonçalves, Durability of steel fiber reinforced self-

compacting concrete, Cons. Build. Mat. 80 (2015) 155–166.

[6] N. Banthia, C. Foy, Marine Curing of Steel Fiber Composites, J. Mater. Civ. Eng. 86 (1989) 86-

96.

[7] C. Frazão, J. Barros, A. Camões, A.C. Alves, L. Rocha, Corrosion effects on pullout behavior of

PT
hooked steel fibers in self-compacting concrete, Cem. Concr. Res. 79 (2016) 112–122

[8] Ç. Yalçınkaya, H. Yazıcı, Effect of early-age freeze-thaw exposure on the mechanical

RI
performance of self-compacting repair mortars, Sci. Eng. Compos. Mater. 23 (2016), 335–344.

[9] G. Li, X. Zhao, C. Rong, Z. Wang, Properties of polymer modified steel fiber-reinforced cement

C
concretes, Cons. Build. Mat. 24 (2010) 1201–1206.

US
[10] E. Sakai, J. Sugita, Composite mechanism of polymer modified cement, Cem. Concr. Res. 25
AN
(1995) 127–135.

[11] C. Andrade, C. Alonso, Corrosion rate monitoring in the laboratory and on-site, Cons. Build.
M

Mat. 10 (1996) 315-328.

[12] Z. Wang, Q. Zeng, L. Wang, Y. Yao, K. Li, Corrosion of rebar in concrete under cyclic
D

freeze–thaw and Chloride salt action, Cons. Build. Mat. 53 (2014), 40–47.
TE

[13] M. P. López , J. M. Ortega, I. Sánchez, M. A. Climent, Steel corrosion rate measurements in

EP

FA concrete using three electrochemical techniques, in: C. Andrade, J. Gulikers, R. Polder,

Durability of Reinforced Concrete from Composition to Protection, Springer International

C

Publishing, Switzerland, 2014, pp. 45-56.

AC

[14] B. Dong, Q. Qiu, Z. Gu, J. Xiang, C. Huang, Y. Fang, F. Xing, W. Liu, Characterization of

carbonation behavior of fly ash blended cement materials by the electrochemical impedance

spectroscopy method, Cem. Concr. Comp. 65 (2016), 118–127.

[15] D.V. Ribeiro, J.C.C. Abrantes, Application of electrochemical impedance spectroscopy (EIS)

to monitor the corrosion of reinforced concrete: A new approach, Cons. Build. Mat. 111 (2016) 98–

104.

17
 ACCEPTED MANUSCRIPT
[16] S. Sathiyanarayanan, P. Natarajan, K. Saravanan, S. Srinivasan, G. Venkatachari, Corrosion

monitoring of steel in concrete by galvanostatic pulse technique, Cem. Concr. Comp. 28 (2006),

630–637.

[17] A. Legat, , M. Leban, Ž. Bajt, Corrosion processes of steel in concrete characterized by means

of electrochemical noise, Electrochimica Acta, 49 (2004), 2741–2751.

PT
[18] N. Etteyeba, X.R. Nóvoa, Inhibition effect of some trees cultivated in arid regions against the

corrosion of steel reinforcement in alkaline chloride solution, Corro. Sci. 112 (2016), 471–482.

RI
[19] A. Poursaee, Temperature dependence of the formation of the passivation layer on carbon steel

in high alkaline environment of concrete pore solution, Electrochemistry Communications, 73

C
(2016), 24–28.

US
[20] J. Williamson, O. B. Isgor, Investigation of Mott–Schottky analysis test parameters to study the
AN
semiconductive properties of passive films of carbon steel in highly alkaline environments,

Advances in Civil Engineering Materials, 5 (2016), 80-106.

M

[21] Y. Lu, J. Zhang, Z. Li, B. Dong, Corrosion monitoring of reinforced concrete beam using

embedded cement-based piezoelectric sensor, Mag. Concr. Res. 65 (2013), 1265-1276.

D

[22] S. Hong, H. Wiggenhauser, R. Helmerich, B. Donga, P. Donga, F. Xing, Long-term

TE

monitoring of reinforcement corrosion in concrete using ground penetrating radar, Corro. Sci. 114
EP

(2017) 123–132.

[23] A. Beglarigale, H. Yazıcı, Pull-out behavior of steel fiber embedded in flowable RPC and
C

ordinary mortar, Constr. Build. Mater. 75 (2015) 255-265.

AC

[24] R. Wang, P.M. Wang, X.G. Li, Physical and mechanical properties of styrene–butadiene

rubber emulsion modified cement mortars, Cem. Concr. Res. 35 (2005) 900-906.

[25] A. Çolak, Properties of plain and latex modified Portland cement pastes and concretes with and

without superplasticizer, Cem. Concr. Res. 35 (2005) 1510-1521.

[26] J.H. Kim, R.E. Robertson, Prevention of air void formation in polymer-modified cement

mortar by pre-wetting, Cem. Concr. Res. 27 (1997) 171-176.

18
 ACCEPTED MANUSCRIPT
[27] D. Enos, The Potentiodynamic Polarization Scan, Technical Report 33, Center of

Electrochemical Science & Engineering, Department of Materials Science & Engineering,

University of Virginia, Charlottesville, VA, 1997.

[28] Y. Ohama, Recent progress in concrete-polymer composites. Adv. Cem. Based Mater. 5 (1997)

31-40.

PT
[29] S. Pascal, A. Alliche, P. Pilvin, Mechanical behaviour of polymer modified mortars, Mater.

Sci. Eng. A. 380 (2014) 1-8.

RI
[30] M. Pei, W. Kim, W. Hyung, A.J. Ango, Y. Soh, Effects of emulsifiers on properties of

poly(styrene– butyl acrylate) latex-modified mortars, Cem. Concr. Res 32 (2012) 837-841.

C
[31] Y. Ohama, Polymer-based admixtures. Cem. Concr. Compos. 20 (1998) 189-212.

US
[32] ASTM C876-09, Standard Test Method for Corrosion Potentials of Uncoated Reinforcing
AN
Steel in Concrete, ASTM International, 2009.

[33] C. Andrade, C. Alonso, Test methods for on-site corrosion rate measurement of steel
M

reinforcement in concrete by means of the polarization resistance method, Mater. Struct. 37(2004)

623-64.
D

[34] A. Beglarigale, H. Yazıcı, Chloride-induced corrosion and bond performance of single steel
TE

fiber embedded in cement based composites, 12th International congress on advances in civil
EP

engineering ACE 2016, İstanbul, Turkey, September 2016.

[35] A.A. Ramezanianpour, Effect of curing on the compressive strength, resistance to chloride-ion
C

penetration and porosity of concretes incorporating slag, fly ash or silica fume, Cem. Concr. Comp.
AC

17 (1995), 125-133.

[36] C.S. Poon, S.C. Kou, L. Lam, Compressive strength, chloride diffusivity and pore structure of

high performance metakaolin and silica fume concrete, Cons. Build. Mat. 20 (2006), 858–865.

[37] H. Yazıcı The effect of silica fume and high-volume Class C fly ash on mechanical properties,

chloride penetration and freeze–thaw resistance of self-compacting concrete, Cons. Build. Mat. 22

(2008), 456–462.

19
 ACCEPTED MANUSCRIPT

PT
RI
Table 1. Mixtures abbreviations

Abb. Matrix Chemical Admixture Curing Abb. Matrix Chemical Admixture Curing

C
CS-N N Control mixture Standard CS-R R Control mixture Standard
CM-N N Control mixture Mix CM-R R Control mixture Mix

US
A1-N N SBR1–5%* Mix A1-R R SBR1–5%* Mix
A2-N N SBR1-15%* Mix A2-R R SBR1-15%* Mix
B1-N N SBR2-5%* Mix B1-R R SBR2-5%* Mix
B2-N N SBR2-15%* Mix B2-R R SBR2-15%* Mix
AN
L1-N N ADP1-1.5%* Mix L1-R R ADP1-1.5%* Mix
L2-N N ADP1-5%* Mix L2-R R ADP1-5%* Mix
R1-N N ADP2-1.5%* Mix R1-R R ADP2-1.5%* Mix
R2-N N ADP2-5%* Mix R2-R R ADP2-5%* Mix
CNI-12 kg/m3 CNI-12 kg/m3
M

F-N N Standard F-R R Standard

SI-N N WP-3%* Mix SI-R R WP-3%* Mix
*Admixture/cement ratio
D

Table 2. Mix design of mixtures.

TE

Aggregate (kg/m3) SP/C

Mixtures

Cement
(kg/m3)

(kg/m3)

Lime Stone Quartz (mm) (%)**

Water

3
Silica
Fume

Chemical
0-5 mm
EP

1-3 0.5-1 0-0.4 0-0.075

Admixture

CS-N 250 0 500 - 1466 - - - - 1.2

CM-N 250 0 500 - 1466 - - - - 1.2
C

A1-N 237 SBR1–5%* 500 - 1466 - - - - 1

A2-N 213 SBR1-15%* 500 - 1466 - - - - 0.9
AC

B1-N 237 SBR2-5%* 500 - 1466 - - - - 1

B2-N 213 SBR2-15%* 500 - 1466 - - - - 0.9
L1-N 246 ADP1-1.5%* 500 - 1466 - - - - 1
L2-N 238 ADP1-5%* 500 - 1466 - - - - 0.8
R1-N 246 ADP2-1.5%* 500 - 1466 - - - - 1
R2-N 238 ADP2-5%* 500 - 1466 - - - - 0.8
F-N 242 CNI-12 kg/m3 500 - 1466 - - - - 1.2
SI-N 243 WP-3%* 500 - 1466 - - - - 1.2
CS-R 184 0 824 107 - 603 347 174 87 2.2
CM-R 184 0 824 107 - 603 347 174 87 2.2
A1-R 160 SBR1–5%* 824 107 - 603 347 174 87 2.1
A2-R 114 SBR1-15%* 824 107 - 603 347 174 87 1.8
B1-R 160 SBR2-5%* 824 107 - 603 347 174 87 2.1
B2-R 114 SBR2-15%* 824 107 - 603 347 174 87 1.8
L1-R 177 ADP1-1.5%* 824 107 - 603 347 174 87 2.1

20
 L2-R 161 ADP1-5%*
ACCEPTED
824 107
MANUSCRIPT
- 603 347 174 87 1.7
R1-R 177 ADP2-1.5%* 824 107 - 603 347 174 87 2.1
R2-R 161 ADP2-5%* 824 107 - 603 347 174 87 1.7
F-R 178 CNI-12 kg/m3 824 107 - 603 347 174 87 2.2
SI-R 170 WP-3%* 824 107 - 603 347 174 87 2.2
* Admixture/cement ratio
** Superplasticizer/cement

PT
RI
Table 3. The average open-circuit potential and corrosion current density values.

C
200 cycles 400 cycles 600 cycles 1200 cycles
EOCP icorr EOCP icorr EOCP icorr EOCP icorr

US
mV(SCE) µA/cm² mV(SCE) µA/cm² mV(SCE) µA/cm² mV(SCE) µA/cm²
CS-N -119 0.0011 -138 0.013 -203 0.021 -204 0.091
CM-N -128 0.012 -143 0.017 -190 0.028 -210 0.021
AN
A1-N -136 0.019 -135 0.026 -196 0.044 -220 0.073
A2-N -110 0.004 -134 0.007 -181 0.04 -212 0.028
B1-N -138 0.023 -140 0.029 -185 0.032 -188 0.173
M

B2-N -105 0.009 -124 0.01 -169 0.021 -190 0.063

L1-N -122 0.011 -126 0.016 -177 0.031 -244 0.055
L2-N -138 0.018 -135 0.042 -181 0.069 -363 2.825
D

R1-N -109 0.023 -144 0.031 -190 0.035 -220 0.280

R2-N -122 0.041 -131 0.055 -172 0.066 -452 0.584
TE

F-N -106 0.015 -126 0.019 -186 0.022 -203 0.019

SI-N -155 0.042 -160 0.054 -206 0.079 -325 0.511
CS-R -83 0.011 -120 0.014 -165 0.018 -201 0.026
EP

CM-R -99 0.015 -110 0.016 -151 0.016 -195 0.010

A1-R -105 0.021 -131 0.021 -174 0.026 -185 0.025
A2-R -124 0.025 -130 0.029 -180 0.034 -220 0.061
-141 0.024 -133 0.028 -191 0.031 -250 0.043
C

B1-R
B2-R -129 0.007 -130 0.012 -181 0.023 -198 0.048
AC

L1-R -102 0.0015 -124 0.0018 -175 0.011 -198 0.006

L2-R -113 0.021 -124 0.022 -174 0.025 -176 0.024
R1-R -96 0.007 -113 0.009 -169 0.011 -175 0.042
R2-R -134 0.018 -147 0.023 -192 0.024 -207 0.047
F-R -110 0.022 -123 0.019 -180 0.031 -191 0.04
SI-R -128 0.023 -137 0.029 -192 0.036 -193 0.034

21
 ACCEPTED MANUSCRIPT

PT
C RI
Figure 1. Details of corrosion test specimens.

US
AN
a b
M
D
TE
C EP

c
AC

Figure 2. a) a corrosion test specimen b) the wetting-drying machine c) corrosion test specimen
after 1200 wetting-drying cycles.

22
 ACCEPTED MANUSCRIPT

PT
Figure 3. Electrochemical corrosion testing setup.

C RI
US
AN
M

≅ 220 mm
D
TE
CEP

Figure 4. Free flow test of control R mixture (CS-R.)

AC

a b

23
 ACCEPTED MANUSCRIPT

PT
RI
Figure 5. a) SEM images of the a) %15 SBR1 containing N mixture (A2-N) b) %15 SBR1

containing R mixture (A2-R).

C
US
AN
M
D
TE
C EP
AC

Figure 6. Decrease in superplasticizer demand.

24
 ACCEPTED MANUSCRIPT

PT
C RI
US
AN
M
D

Figure 7. Dry bulk density values.

TE
EP

a b
C
AC

Figure 8. SEM images of the a) control R mixture (CM-R) b) %5 ADP2 containing R mixture (R2-

R).
25
 ACCEPTED MANUSCRIPT

PT
C RI
US
AN
M
D
TE

Figure 9. Porosity values of the hardened mixtures.

EP
C
AC

26
 ACCEPTED MANUSCRIPT

PT
C RI
US
AN
M
D

Figure 10. Water absorption values of the hardened mixtures.

TE
C EP
AC

Figure 11. Correlation between porosity and water absorption values.

27
 ACCEPTED MANUSCRIPT

PT
C RI
US
AN
M
D

Figure 12. Compressive and flexural strength of the hardened N mixtures.

TE
C EP
AC

28
 ACCEPTED MANUSCRIPT

PT
C RI
US
AN
M
D

Figure 13. Compressive and flexural strength of the hardened R mixtures.

TE

a a
C EP
AC

b b

Figure 14. SEM images of the a) WP containing N mixture (SI-N) b) WP containing R mixture (SI-
R).
29
 ACCEPTED MANUSCRIPT

a a

PT
b b

C RI
b US b
AN
M
D
TE

c c
C EP
AC

Figure 15. SEM images of the a) %15 SBR1 containing N mixture (A2-N) b) %15 SBR2 containing
N mixture (B2-N) c) %15 SBR1 containing R mixture (A2-R)

30
 ACCEPTED MANUSCRIPT

a a

PT
b b

C RI
US
Figure 16. SEM images of the a) %5 ADP2 containing N mixture (R2-N) b) %5 ADP1 containing
N mixture (L2-N).
AN
M
D
TE
C EP
AC

Figure 17. Open-circuit potential measurements.

31
 ACCEPTED MANUSCRIPT

PT
C RI
US
Figure 18. Polarization curves.
AN
M
D
TE
EP
C
AC

32
 ACCEPTED MANUSCRIPT

PT
C RI
US
AN
M
D
TE
C EP

Figure 19. The open-circuit potential and corrosion current density values.
AC

33
 ACCEPTED MANUSCRIPT

PT
1
1.0 mm 0.2 mm

RI
2

C
US
2 AN
M

1.0 mm 0.2 mm

Figure 20. Digital microscope images of the steel fiber embedded in the %5 ADP1 containing N
D

mixture (L2-N) after 1200 wetting-drying cycles.

TE
C EP
AC

34
 ACCEPTED MANUSCRIPT
0.5 mm
1

PT
0.5 mm

RI
2

C
2
US
AN
M

Figure 21. Digital microscope and SEM images of the fiber-matrix interface of the %5 ADP1
containing N mixture (L2-N) after 1200 wetting-drying cycles.
D
TE
C EP
AC

35
 ACCEPTED MANUSCRIPT
1
2
1

PT
RI
2 3

C
3
US
AN
M

Figure 22. SEM images of the first hooked region of the steel fiber embedded in the %5 ADP1
containing N mixture (L2-N) after 1200 wetting-drying cycles.
D
TE
C EP
AC

36
 ACCEPTED MANUSCRIPT
3 2
Non-corroded Corroded

PT
RI
1
3

C
Corroded
1

US
4
2
AN
M

4 4
D

Non-corroded
TE

Non-corroded
C EP

Figure 23. SEM images of the steel fiber embedded in the %5 ADP1 containing N mixture (L2-N)
AC

after 1200 wetting-drying cycles.

37
 ACCEPTED MANUSCRIPT

1 2

PT
1.0 mm 0.2 mm

RI
2 2

C
US
AN
M

1.0 mm 0.2 mm
3
D
TE
EP

3
C

Figure 24.Digital microscope and SEM images of the fiber-matrix interface of the %5 ADP2
AC

containing N mixture (R2-N) mixture after 1200 wetting-drying cycles.

38
 ACCEPTED MANUSCRIPT
1 2

1 3

PT
2 3

C RI
US
AN
M
D
TE

Figure 25. SEM images and EDS analyses of the steel fibers embedded in the %15 SBR2
containing R mixture (B2-R) mixture after 1200 wetting-drying cycles.
C EP

1
AC

Figure 26. SEM images of the fiber-matrix interface of the %15 SBR2 containing R mixture (B2-R)
mixture after 1200 wetting-drying cycles.

39