Construction and Building Materials 143 (2017) 289297

Contents lists available at ScienceDirect

Construction and Building Materials

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

Corrosion behavior of steel bars immersed in simulated pore solutions of

alkali-activated slag mortar
Wenrui Wang a,c, Huisu Chen a,, Xiangyu Li b, Zhigang Zhu a
a
Jiangsu Key Laboratory of Construction Materials, School of Materials Science and Engineering, Southeast University, Nanjing 211189, China
b
College of Architecture and Civil Engineering, Taiyuan University of Technology, Taiyuan 030024, China
c
Institute of Advanced Engineering Structures and Materials, Zhejiang University, Hangzhou 310058, China

h i g h l i g h t s

 Simulated pore solutions of AAS mortar were first applied in this work.
 Both aluminate and silicate in AAS pore solution had corrosion inhibition effect.
 AAS had stronger capacity to passivate and protect steel bar than PC.

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

Article history: To study the corrosion behavior of steel bars in alkali-activated slag (AAS) system, simulated pore solu-
Received 15 November 2016 tions of AAS mortar were first applied in this work. The effects of the three major species in AAS pore
Received in revised form 10 March 2017 solutions, namely sulfur-containing species, aluminate, and silicate, were investigated. Electrochemical
Accepted 14 March 2017
impedance spectroscopy and potentiodynamic polarization results showed that the simulated pore solu-
tion of AAS mortar had a stronger capacity to passivate and protect steel bars than the simulated pore
solution of Portland cement (PC) mortar, mainly due to the higher concentration of silicate in the former,
Keywords:
as confirmed by surface analysis of SEM-EDXS.
Steel bar
Corrosion
2017 Elsevier Ltd. All rights reserved.
Simulated pore solution
Alkali-activated slag
Sulfur-containing species
Silicate

1. Introduction of long-term performance monitoring compared to that available

To meet the construction industrys demands for cement and to
ensure sustainable development of the environment, a greener
and energy-efficient cement, alkali-activated materials (AAMs),
has been proposed [1]. Its CO2 emission rate is approximately 6
times less than that of ordinary Portland cement (OPC) [2,3].
Because of the application of industrial by-products such as slag
[4], fly ash [5], and clay minerals [6], AAMs have the superior attri-
butes of a rich raw materials source, low price, and energy conser-
vation. They also have higher chemical resistance, higher
resistance to freeze-thaw cycles, and higher resistance to high tem-
perature and fire [7]. These excellent properties make AAMs a com-
petitive alternative for OPC [7,8]. At present, however, our
understanding of the durability of AAMs, especially the corrosion
behavior of steel in AAM concrete, is insufficient, due to the lack
Corresponding author.
E-mail address: chenhs@seu.edu.cn (H. Chen).
from the several hundred years applications of OPC [8]. This lack
is particularly true of AAMs based on blast furnace slag (BFS),
denoted as alkali-activated slag (AAS), containing sulfide, that gen-
erates a reduction environment within the binder and causes elec-
trochemical complexities that are not yet well understood [8].
Some researchers have studied the corrosion behavior of steel bars
embedded in AAS concretes under chloride ingress or carbonation
environment. The corrosion of reinforcing bars embedded in AAS
concrete subjected to chloride attack was studied by Chaparro
et al. [9]. Samples were immersed in 3.5% NaCl solution by weight
of the slag. It was found that AAS concrete presented passive cor-
rosion behavior in the first 3 months, after which it presented
decreased corrosion resistance, lower than that of OPC concrete.
Holloway et al. [10] applied detailed electrochemical characteriza-
tion to the analysis of mild steel in AAS mortars containing NaCl
admixtures, and found that the highest chloride levels produced
the lowest corrosion rates, a trend that could not be clearly
explained from a fundamental chemical perspective. Aperador

http://dx.doi.org/10.1016/j.conbuildmat.2017.03.132
0950-0618/ 2017 Elsevier Ltd. All rights reserved.
290 W. Wang et al. / Construction and Building Materials 143 (2017) 289297
et al. [2] found that the carbonation rate was higher in AAS con-
crete than in the respective OPC concrete under both accelerated
carbonation (3% CO2, 65% RH, and 25 C) and a laboratory environ-
ment (0.03% CO2, 65% RH, and 25 C). In the authors view, the high
carbonation rate of AAS might be attributed to the low calcium
content in AAS concrete that leads to a low capacity to buffer the
pH of the pore solution in AAS and also leads to rare deposits of
CaCO3. Moreover, reduced carbonation resistance in AAS binders
can also be caused by the chemistry of the pore solution and the
pore structure of the binder matrix. The absence of calcium
hydroxide in AAS binders is also relevant. Though many studies
have been conducted, they cannot straightforwardly explain the
corrosion chemistry of steel embedded in AAS binders. Thus there
is a strong need for further detailed scientific and analytical work
to determine the mechanisms in order to better understand and
control steel corrosion within AAS [8].
Simulated pore solutions of AAS were applied in this work to
study the corrosion behavior of steel bar in AAS binders, because
the chemistry of pore solution around steel is closely related to
the corrosion behavior of rebar in concretes and directly influences
the passivation and the permanence of the passive state in the steel
surface. Few researchers have studied the corrosion of steel
immersed in simulated pore solutions of AAS binders. Chemical
analysis of pore solutions in AAS paste by Puertas et al. [11] and
Gruskovnjak et al. [12] showed that pore solutions of AAS had
higher concentrations of Na, Al, Si, and S elements than those in
OPC, with lower concentrations of K and Ca. It is interesting to find
that several sulfur-containing species, such as S2 , S2O23 , SO23 , and
SO24 , existed in the AAS system, whereas only sulfate was available
in the OPC system [12]. These differences in the chemistry of the
pore solutions may lead to considerable differences between the
corrosion behaviors of steel bars in these two cements.
The structure of this study is as follows. First, the chemistry of
real pore solution extracted from AAS mortar was studied. Simu-
lated pore solution was then prepared in accordance with the
chemistry of the extracted pore solution. Through immersing steel
bars in simulated pore solutions, various experimental methods
were applied to study the effect of the main anions (S2 , S2O23 ,
SO23 , SO24 , AlO2 , and SiO23 ) on the corrosion behavior of steel bars
to reveal the corrosion mechanism of steel in AAS system. Simu-
lated pore solution of Portland cement (PC) mortar was also pre-
pared for comparative purposes. Two nondestructive monitoring
techniques, electrochemical impedance spectroscopy (EIS) and lin-
ear polarization resistance (LPR), were utilized to continuously
monitor the corrosion processes of steel bars. To clarify the corro-
sion behavior, cyclic voltammetry (CV) and potentiodynamic
polarization (PDP) measurements were conducted. Scanning elec-
tron microscopy (SEM) and energy dispersive X-ray spectroscopy
(EDXS) were employed to observe the morphology of steel surfaces
and to analyze corrosion products after the immersion experi-
ments, in accordance with the EIS measurements.
2. Experimental
2.1. Pore solution extraction and analysis
Pore solution is usually squeezed from cylindrical samples [13]
or fragments [14] of pastes or mortars using a high-pressure
device. Concretes are not used, lest coarse aggregates damage the
device. In this study, squeezing was performed on AAS mortars
and also on PC mortars to extract pore solutions. The ground gran-
ulated blast furnace slag (GGBFS) used was S95 slag, conforming to
GB/T 18046-2008. P
II 52.5 cement conforming to GB175-2007 was
used to prepare the PC mortar. Chemical compositions, surface
areas and specific gravity are summarized in Table 1. The basicity
(CaO + MgO)/(SiO2 + Al2O3) and quality (CaO + MgO + Al2O3)/
(SiO2 + TiO2 + MnO) indexes of GGBFS were 0.94 and 1.73, respec-
tively. The waterglass used as activating solution was a mixture of
commercially available sodium silicate mix (34.8% SiO2, 11.6%
Na2O, and 63.6% water) with flaky analytical reagent NaOH to
obtain a SiO2/Na2O ratio of 1 and a 5% concentration of Na2O by
weight of slag. The specific gravity and bulk density of river sand
were 2630 kg
m 3 and 1480 kg
m 3 with a fineness modulus of
2.5. The sand/binder weight ratio was 2 for both AAS and PC mor-
tars. Based on previous study, a water/cement ratio of 0.3 was used
for AAS mortars to obtain pore solutions. PC mortars with water/
cement ratio of 0.3, 0.4, and 0.5 were prepared, in order to find a
suitable water content for the PC mortar to extract sufficient pore
solutions for testing. Mortar samples were cast into plastic molds
with the dimension U 50 mm  60 mm at room temperature. After
casting, the surfaces of the samples were sealed by plastic film and
the samples were then cured in a curing chamber at the tempera-
ture of 20 2 C and RH >95%. Taking into account that the
strength of AAS and PC mortars at age of 7 d can be almost 80%
of that at 28 d, and the chemistry of pore solutions in AAS is similar
from 7 d to 180 d [12], the pore solution of mortars was extracted
at 7 d because continuing hydration of the specimens could lead to
insufficient pore solution for squeezing out.
The pore solution extraction device in this experiment was sim-
ilar to that used by Barneyback et al. [15]. Before each extraction
test, all surfaces of the device were cleaned by ethanol at least
twice until no residue remained [15]. An approximate 6 mm-
thick Teflon seal was inserted between the top of the specimen
and the bottom surface of the piston, to reduce the damage to
the piston [14,15] and reduce the friction between mortar samples
and piston by facilitating transfer of the applied pressure to
samples.
The loading rate for extraction was about 12.8 MPa/s until the
maximum 550 MPa stress was achieved [14,15]. To eliminate
residual particles [14], the collected solution was filtered as soon
as possible using filter paper with pore size 0.45 lm, then the pH
of the pore solution was determined using a precision pH meter
(PHS-3E, Shanghai INESA Scientific Instrument CO., LTD, China).
All the liquid samples were sealed to minimize any carbonation
or oxidation [16]. The concentration of elements (K, Na, Ca, Al, Si,
and S) was measured by ICP-OES (700 Series, Agilent Technologies,
USA), and concentrations of S2 , SO23 , S2O23 , and SO24 were tested
by ion chromatography (IC) (ICS-3000, Thermo Fisher Scientific,
USA), as listed in Table 2. The tested concentration of ions in AAS
and PC mortars was consistent with the literature [12]. It should
be noted that only PC mortars with 0.4 and 0.5 water/cement ratio
could provide adequate pore solutions for chemical analysis.
2.2. Specimen preparation
The difficulty of studying the corrosion behavior of steel in pore
solutions stems from the complexity of effectively replicating the
chemical composition of the pore solution. For a PC system, satu-
rated Ca(OH)2 solution is usually applied to represent the chemical
environment of the concretes pore solution [17]. A certain amount
of NaOH and KOH and even a trace of CaSO4 may also be added
[18,19], to better represent the pore solution chemistry of PC. On
the other hand, simulated pore solution of the AAS system has
not previously been used. As already stated, the chemical compo-
sition of the pore solution of AAS is more complex than that of
PC. Therefore, we cannot use only one or a few species in pore solu-
tion to represent the real environment of AAS binders pore solu-
tion, as was done for the PC system. Thus when preparing
simulated pore solutions of AAS and PC mortars we considered
all the ions that exist in real pore solutions to better represent
the real environment in the binders pore solutions. This method
 W. Wang et al. / Construction and Building Materials 143 (2017) 289297 291

Table 1
Chemical compositions, surface area and specific gravity of GGBFS and P
II 52.5 cement.

Chemical compositions (wt.%) LOI IR Surface area Specific gravity

(m2
kg 1) (kg
m 3)
Na2O K2O CaO SO3 SiO2 Al2O3 Fe2O3 MgO TiO2
GGBFS 0 0.23 35.60 2.61 34.73 15.94 0.23 8.80 1.72 1.40 1.29 580 2910
P
II 52.5 cement 7.80 1.01 50.26 3.53 18.71 6.30 2.52 3.20 0.21 3.01 0.79 324 3100

LOI: Loss on ignition.

IR: Insoluble residue.

Table 2
Chemical analysis of the solutions extracted from AAS and PC mortars (mmol/L).

K+ Na+ Ca2+ AlO2 SiO23 S2 SO23 S2O23 SO24 Stotal pH

AAS mortar 96.556 1552.826 0.0236 33.456 13.593 37.600 22.650 71.738 58.947 647.406 13.58
PC mortar W/B = 0.4 441.347 97.637 0.041 2.339 1.953 0.000 0.000 13.58
PC mortar W/B = 0.5 333.858 95.670 0.020 1.716 1.491 0.000 0.000 13.55

was also used in [20], and the results showed that the corrosion Table 4
behaviors of steel bars in simulated pore solutions and real con- Chemical composition of simulated pore solutions for PC.

cretes were similar, a finding that verified the reliability of using KOH = COH -CNaOH-2  CCa(OH)2 = 288.715 mM/L
simulated pore solutions to study the corrosion behavior of steel NaOH = CNa+-2  CNa2SiO3-CNaAO2 = 91.392 mM/L
bars. Na2SiO3 = CSiO32 = 1.953 mM/L
NaAlO2 = CAlO2 = 2.339 mM/L
Simulated pore solutions were prepared according to the chem- Ca(OH)2 = CCa2+ = 0.041 mM/L
ical analysis of extracted pore solutions. Here, the analysis results
of pore solution from PC mortar with W/B = 0.4 were adopted
because this W/B was close to the W/B of 0.3 for the AAS mortar,
behavior was studied by immersing steel bars in solutions, denoted
so comparisons could be drawn between AAS and PC in simulated
as S, with only sulfur-containing species (S2 , SO23 , S2O23 , and
pore solution experiments. The ion concentrations of simulated
SO24 ). Second, the study of steel corrosion behavior was conducted
pore solutions for AAS and PC mortars are listed in Tables 3 and
in two types of solution, one containing S-species and aluminate,
4, calculated in accordance with the concentration of ions as given
denoted S + Al, and the other containing S-species, aluminate,
in Table 2. The concentration of ions in the simulated pore solu-
and silicate, denoted S + Al + Si. Solutions were prepared in accor-
tions was consistent with that in the extracted pore solutions,
dance with Table 3. It should be noted that the S + Al + Si solution
except for Na+ in the AAS simulated pore solution and K+ in the
was used as the simulated pore solution of AAS because this solu-
PC simulated pore solution. The concentration of Na+ in the AAS
tion contained all the main ions identified in real AAS pore solu-
simulated pore solution was lower than that in real pore solution.
tion. Third, for comparison, the simulated pore solutions of PC
In the real pore solution of AAS, the total charge of cations was
mortar, denoted PC, were prepared in accordance with Table 4.
greater than that of anions (including OH calculated from the
All solutions were prepared with analytical reagents and deionized
pH value). This was due to the possible existence of other sulfur-
water with electrical resistivity of 18 MX.
containing species that were complex and difficult to determine
The steel bars used in this study were common HRB 400 (hot
[12]. Therefore, considering the charge balance of the simulated
rolled ribbed steel bars) with diameter 16 mm and length
pore solution, some cations were excluded. Because of the high
10 mm. Their chemical composition was tested by an optical emis-
content of Na+, this exclusion was applied to Na+. In the PC simu-
sion spectrometer (MAXx LMF 15, SPECTRO Analytical Instruments
lated pore solution, a similar application was conducted with
GmbH, Germany), and is listed in Table 5. Three replicates were
respect to K+ because of the similar charge imbalance in the real
tested in each solution. Tests were conducted on a cross-section
pore solution of PC mortar. This imbalance might also be due to
of steel bar with the surface area of 1 cm2, that was ground by
anions that could not be detected.
180#, 400#, 600#, 1200# SiC sandpapers for 23 min at each stage
To better understand the effects of various anions on the corro-
at the speed of 300 rpm, and then polished by 3 lm monocrys-
sion behavior of a steel bar embedded in AAS, the experimental
talline diamond suspension for 510 min at the speed of 150 rpm
scheme was designed as follows. First, the influence of the most
on the grinding and polishing machine (Phoenix 4000, Buehler
important component, sulfur-containing species, on the corrosion
Ltd., USA). After finishing, the steel bars were ultrasonically
behavior of the steel bar was investigated. The steel corrosion
cleaned with alcohol for 5 min [21] and then the lateral surface
was wrapped in insulating tape to prevent corrosion and other
unexpected effects.
Table 3
Chemical composition of simulated pore solutions for AAS. 2.3. Test methods
KOH = CK+ = 96.556 mM/L
Na2S = CS2 = 37.600 mM/L To simulate actual conditions for reinforced concrete structures,
Na2SO3 = CSO32 = 22.650 mM/L the steel bar samples were kept in solution for 2 weeks [19,22,23]
Na2S2O3 = CS2O32 = 71.738 mM/L
Na2SO4 = CSO42 = 58.947 mM/L
before the addition of chlorides, to passivate the steel bars and
Na2SiO3 = CSiO32 = 13.593 mM/L make the steel surface keep stable. Here, NaCl was gradually added
NaAlO2 = CAlO2 = 33.456 mM/L to simulate a severe marine environment and also to accelerate the
NaOH = COH -CKOH-2  CCa(OH)2 = 283.586 mM/L corrosion of the steel bars. With reference to Ghods et al. [19], after
Ca(OH)2 = CCa2+ = 0.024 mM/L
the passivation period, an additional 0.01 mol/L NaCl was added
292 W. Wang et al. / Construction and Building Materials 143 (2017) 289297
Table 5
Chemical composition (wt.%) of HRB400 steel bar.
Fe C Si Mn
98.8000 0.2430 0.2920 0.5590
every 7 d. 0.02 mol/L or more NaCl can be added at an older age to
accelerate the corrosion of steel bars.
All electrochemical measurements were carried out at room
temperature (25 5 C) in aerated and stagnant solutions with a
three-electrode system by Electrochemical Workstation (PARSTAT
4000, Princeton Applied Research, USA). The steel bar electrode,
saturated calomel electrode, and Pt electrode were used as the
working electrode, reference electrode, and counter electrode,
respectively. The corrosion potential (Ecorr) and polarization resis-
tance (RP) of steel bars were measured periodically to characterize
the corrosion tendency of steel in different solutions by the LPR
technique. The reported results are the average of three identical
bars. In LPR testing, the potential range was 10 mV vs. open cir-
cuit potential [24]. The sweep rate was 0.6 V/h [25].
Values of Ecorr and RP corresponding to the corrosion condition
are listed in Tables 6 and 7, respectively. Ecorr is an indication of
the relative probability of corrosion activity that could be empiri-
cally obtained through measurement of the potential difference
between the reference electrode and the steel bar [26]. That is to
say, the results of Ecorr are qualitative and cannot provide an accu-
rate corrosion rate of the steel bar. Ecorr data can be used to provide
a rapid preliminary judgment of the extent of corrosion and to con-
firm surveys obtained by other techniques [26]. Factors such as
chloride and moisture content, cell placement, electrical continu-
ity, and temperature are known to have a significant influence on
Ecorr readings [26]. Cathodic protection systems, stray currents,
previous repair patches, etc., can also cause interference. Confirma-
tion by other nondestructive corrosion rate determination meth-
ods (such as RP) was therefore necessary in evaluating the Ecorr
results [26].
EIS measurement was conducted with a scan frequency from
10 mHz to 100 kHz, an amplitude of 10 mV RMS [29,30]. CV plots
can provide information about the possible redox reaction on steel
surface. In this study, CV data was obtained after 10 cycles
between the voltages of 1.4 V and 0.6 V (vs. reference electrode)
with a 50 mV/s scanning rate at 1 h. The PDP test was started from
0.25 V (vs. open circuit potential) to 1.2 V (vs. reference elec-
trode) at a scan rate of 0.5 mV/s [31] to study the passive and pit-
ting behavior of steel bars in different solutions.
The topographies of the surface of steel bar specimens and the
chemical composition at specific location were analyzed by SEM
(Quanta 3D FEG, FEI, USA) at an acceleration voltage of 20 kV.
3. Results and discussion
3.1. Linear polarization resistance measurements
The development of Ecorr and RP for steel bars in different solu-
tions is shown in Fig. 1. From Fig. 1(a) we can see that the presence
of AlO2 led to a decrease in Ecorr of the steel bar, whereas the pres-
Table 6
Ecorr values for different corrosion conditions (calomel electrode) [27].
Half cell potential Corrosion condition
> 126 mV Low (10% risk of corrosion)
126 mV to 276 mV Intermediate corrosion risk
< 276 mV High (<90% risk of corrosion)
< 426 mV Severe corrosion
P S Cu Ni
0.0289 0.0178 0.0061 0.0069
Table 7
RP values for different corrosion conditions
[28].
RP (kX
cm2) Rate of corrosion
0.25 < RP < 2.5 Very high
2.5 < RP < 25 High
25 < RP < 250 Low/moderate
250 < RP Passive
ence of SiO23 showed an opposite trend but was still lower than
that in PC. This finding might result from the presence of S2 , as
mentioned by Glasser [32], which brought about a lower Ecorr in
the steel bars in AAS binders than that in PC binders. Although
the RP of steel bars in different solutions showed a trend similar
to that with Ecorr, for S + Al with 14 d passivation and the addition
of NaCl (0.04 mol/L), RP became larger than that in S. Based on
above analysis, it can be concluded that AlO2 and SiO23 could effi-
ciently protect steel bars from corrosion caused by chloride attack.
It is well recognized that the presence of chloride in reinforced
concrete can lead to corrosion of the steel bar by destruction of
the passive film on the steel surface [33]. Here, a parameter is very
important for the design and monitoring of the durability of rein-
forced concrete structures. It is the critical chloride concentration/
content (Ccrit), indicating that at this concentration the depassiva-
tion of the steel surface has developed while corrosion is just
beginning. From Fig. 1(b) we can see that the Ccrit in AAS is about
0.33 mol/L (at 52 d), while it is 0.53 mol/L (at 58 d) or even higher
in PC. These LPR results reflected that the steel bars were more vul-
nerable to corrosion in the simulated pore solution of AAS than in
the simulated pore solution of PC.
It is worth noting that the steel bars in both S and S + Al suffered
from severe corrosion as shown in Fig. 1(a), whereas a low or mod-
erate degree of corrosion was found before 35 d as shown in Fig. 1
(b). In S + Al + Si solution, a similar tendency was also found. This
phenomenon may result from the decrease of potential by S2
[33]. Additionally, the redox actions of S-containing species may
consume O2, which would inhibit the cathodic reactions [10]. This
result indicates that when S-containing species (mainly referring
to S2 ) exist, Ecorr can just be used to judge the relative corrosion
tendency of the steel bar and RP can be applied to estimate the cor-
rosion degree more accurately.
3.2. Electrochemical impedance spectroscopy
The EIS behavior of the steel bars shown in Fig. 2 was basically
similar to the LPR results in Fig. 1: the extent of corrosion of steel
bars in S and S + Al was similar, while it was similar in S + Al + Si
and PC. As shown in the Nyquist plots of Fig. 2(a), after passivation
for 14 d, the radius of capacitance arc of steel bars in different solu-
tions increased in the following order: S + Al < S < S + Al + Si < PC.
Steel bars in S + Al and S solutions showed similar impedance val-
ues, but the impedance value in the S + Al + Si solution was much
higher than that in S + Al and S. These results were in line with
their Bode modulus diagrams as shown in Fig. 2(a), indicating that
the presence of silicate led to a significant increase in the capacity
of the solution to passivate the steel bar. In the high frequency
region (>10 Hz [34]) of Bode diagrams in Fig. 2(a), the impedance
modulus and phase angle value of steel bars in S + Al + Si were
 W. Wang et al. / Construction and Building Materials 143 (2017) 289297
(a) Corrosion potential
Fig. 1. Results of linear polarization resistance tests in different solutions.
higher than those in PC, implying that the passivation film on the
surface of steel bars in S + Al + Si had a high impedance value
and was more intact. In the low frequency region (<10 Hz [34]),
however, both impedance modulus and phase angle value were
lower than in the PC solution in the low frequency regions
(<10 Hz and <100 mHz, respectively), suggesting that redox reac-
tions on steel bar surface could occur more easily in AAS system.
After 14 d passivation and the introduction of NaCl (0.04 mol/L),
as shown in Fig. 2(b), the impedance values of the steel bars in S
+ Al + Si and PC solutions were similar to those of the
passivation-only solutions. However, for the solution with NaCl,
the radius of capacitance arc for the steel bars in S was lower than
that in S + Al solution and their impedance modulus and phase
angle value showed similar phenomena in the low frequency
region, indicating that the introduction of aluminate could effec-
tively inhibit chloride attack, a finding that was consistent with
the RP results in Fig. 1(b). On the other hand, the radius of capaci-
tance arc and impedance modulus of the steel bars in S + Al + Si
were always higher than those in PC solution when the concentra-
tion of NaCl was in the range from 0.53 mol/L (Fig. 2(c)) to
0.83 mol/L (Fig. 2(d)), implying that the AAS system had a higher
capacity for inhibiting corrosion caused by chloride than the PC
system with the increase of chloride concentration, although the
LPR results reflected that the steel bars in simulated pore solution
of AAS were more vulnerable to corrosion than those in the simu-
lated pore solution of PC.
SEM observations were carried out to interpret the corrosion
inhibition effects of aluminate and silicate in AAS pore solutions.
It can be seen from Fig. 3(a) that the presence of aluminate in S
+ Al brought about the formation of a discrete film on the steel sur-
face. This film was the reason why the steel bars in S + Al solution
exhibited better performance in inhibiting corrosion caused by
chloride than those in solution S, as shown in Fig. 2(b). According
to the EDXS results shown in Fig. 3(d), apart from the element Fe,
elements Na, O, and Al were also included in this film. The large
amount of Fe identified in the EDXS might result from the steel
bar itself, in view of the fact that the film was thin and discontin-
uous. Therefore, the substances existing in the film might be NaOH
and Al(OH)3 or aluminates. With S + Al + Si, a thick but easily
cracked film formed on the surface of the steel bars, with some
colloid-like substances attached to it. Apart from the element Fe,
this film contained much Na, Al, Si, and O, indicating that the main
components of the film might be silicates and aluminates (or alu-
minum hydroxide). The colloid-like substances might be silicates,
aluminates (or aluminum hydroxide), or some salts containing sul-
fur. Only a very thin macula formed on the surface of the steel bars
in the simulated solution of PC. By EDXS analysis (Fig. 3(g)), its
main components were elements Fe, Na, and O, implying the pres-
ence of NaOH or oxidation products of iron.
293
(b) Polarization resistance
Based on the EIS and SEM results, it can be concluded that the
protective films of aluminates (or aluminum hydroxide) and sili-
cates that formed on the steel surface endowed the solutions con-
taining AlO2 and SiO23 with good corrosion resistance capacity.
And it was the high concentration of silicate that endowed the sim-
ulated pore solution of AAS with greater steel-protection capacity
than that of the simulated pore solution of PC. Furthermore, the
results of EIS indicated that the AAS system would show better
steel-protection capacity than the PC system under seawater envi-
ronment (average salinity 3.5% [35], corresponding to 0.6 mol/L
Cl ). This result would provide a valuable reference for wide appli-
cations of AAS and improvement of the durability of structures
under severe marine environments.
3.3. Potentiodynamic polarization results
Potentiodynamic polarization measurement can provide further
information, including pitting potential (Epit) and the kinetics of
anode and cathode on steel surface to clarify the corrosion behav-
ior exhibited by EIS results. The increase of passive current density
(iP) may reflect greater difficulty in obtaining passivation in solu-
tions [36]. A higher Epit indicates that the passivity of steel is stable.
As shown in Fig. 4(a), at 1 h, the polarization behaviors of steel
bars in S and S + Al were similar: low Epit and high iP., whereas the
steel bars in S + Al + Si and PC showed the opposite trend: high Epit
and low iP. Among various solutions, S + Al + Si had the highest Epit.
These results suggested that the capacity of the solution to passi-
vate the steel bar and inhibit the attack of chloride was improved
significantly by the introduction of silicate, a conclusion that was
consistent with the LPR and EIS results. When the PDP results of
14 d passivation are compared with those of 1 h passivation, it
can be seen that steel bars in the simulated pore solution of AAS
showed better chloride inhibiting performance than that in the
simulated pore solution of PC, a phenomenon that became more
remarkable after the introduction of NaCl (0.83 mol/L). In this pro-
cess, the presence of silicate played a key role in simulated pore
solution of AAS.
3.4. Cyclic voltammetry
According to the results of the voltammetric curve shown in
Fig. 5(a), steel bars in S and S + Al showed similar redox reaction
processes, but the anodic current density of the steel bars in S
+ Al + Si was significantly lower than that of the steel bars in S
and S + Al. This finding might result from the fact that the silicates
film on the steel surface (as shown in Fig. 3(b)) inhibited contact
between the steel bars and the surrounding solution [37,38] so
reactions were less likely to occur. Additionally, for S + Al + Si, the
positions of redox peaks in the cyclic voltammetry curve were dif-
294 W. Wang et al. / Construction and Building Materials 143 (2017) 289297

(a) Passivation for 14 d

(b) Addition of 0.04 mol/L NaCl after passivation for 14 d

(c) Addition of 0.53 mol/L NaCl after passivation for 14 d

(d) Addition of 0.83 mol/L NaCl after passivation for 14 d

Fig. 2. Impedance spectra of steel bars in different solutions at different stages.

ferent from those in PC, indicating that different kinds of redox At the 10th cycle, the current density of oxidation peak at
reactions occurred in these two solutions. 0.9 V of steel bars in S + Al + Si was higher than that at the first
 W. Wang et al. / Construction and Building Materials 143 (2017) 289297 295

(a) S+Al solution (b) S+Al+Si solution (c) PC solution

(d) EDXS 1 (e) EDXS 2 (f) EDXS 3

(g) EDXS 4
Fig. 3. Surface morphology of steel bars immersed in different solutions (a), (b), and (c) with 0.83 mol/L NaCl after passivation for 14 d, and corresponding EDXS results (d),
(e), (f), and (g).

(a) Passivation for 1 h (b) Passivation for 14 d (c) Addition of 0.83 mol/L
NaCl after passivation for 14 d
Fig. 4. Potentiodynamic polarization curves of steel bars in different solutions at different stages.
296 W. Wang et al. / Construction and Building Materials 143 (2017) 289297
(a) 1st
Fig. 5. Voltammograms for steel bars in different solutions at different cycles.
cycle. This peak, the highest one, corresponded to the oxidation of
Fe to Fe (II) [39]. The highest oxidation peak appearing in the PC
solution was at 0.64 V, corresponding to the oxidation of Fe (II)
to Fe (III) [39]. From Fig. 5, therefore, it can be concluded that for
the AAS system the oxidation of Fe to Fe (II) was the main reaction
occurring on the surface of the steel bars whereas oxidation of Fe
(II) to Fe (III) was the main reaction occurring on the surface of
the steel bars in the PC system. In other words, in the AAS system
Fe existed mainly in the form of Fe (II), but Fe (III) was the most
stable state in the PC system, a phenomenon that might be caused
by the presence of S2 in slag that kept Fe in reduction state [40].
Additionally, when the potential was higher than 0.3 V, the ano-
dic current density of the steel bars in S and S + Al increased with
the increase of potential, but the opposite trend was found in S
+ Al + Si, indicating that the formation of oxide film on the surface
of steel bars in S + Al + Si could passivate the steel bar.
4. Conclusions
Simulated pore solutions of AAS and PC mortars were used to
study the corrosion behavior of steel bars in AAS and PC binders.
NaCl was gradually added to simulate a severe marine environ-
ment and also to accelerate the corrosion of the steel bars. Electro-
chemical measurements of LPR, EIS, PDP, and CV showed good
consistency. SEM observation provided further proof for EIS
results. From this study, the following conclusions can be drawn:
1. RP and EIS results showed that aluminate in pore solution of
AAS mortar had an inhibition effect on the corrosion of steel
bars after the addition of NaCl. SEM results showed that a dis-
crete film was formed on the steel surface.
2. All electrochemical tests indicated that silicate, another major
ion in the AAS pore solution, provided a good protective effect
on steel bars in both passivation and chloride attack stages.
During chloride attack, the formation of a protective film con-
taining aluminates and silicates on steel in S + Al + Si solution
was the main reason for the inhibition effect.
3. EIS results revealed that the simulated pore solution of AAS
mortar had a greater capacity to passivate steel bars than the
simulated pore solution of PC mortar, and had better protective
effect on steel bars from chloride ingress, especially under
higher chloride concentration (consistent with the PDP results),
although LPR results showed that the Ccrit in AAS was lower
than that in PC, mainly due to the existence of a higher concen-
tration of silicate in the AAS pore solution.
(b) 10th
Acknowledgements
The authors gratefully acknowledge the financial support from
the National Nature Science Foundation of China via Grant No.
51461135001, the Ministry of Science and Technology of China
973 Project via Grant No. 2015CB655102, and the Fundamental
Research Funds for the Central Universities via Grant No.
2242016K41054.
References
[1] F. Puertas, A. Gil-Maroto, M. Palacios, T. Amat, Alkali-activated slag mortars
reinforced with ar glassfibre. Performance and properties, Mater. Constr. 56
(283) (2006) 7990.
[2] W. Aperador, R. Meja de Gutirrez, D.M. Bastidas, Steel corrosion behaviour in
carbonated alkali-activated slag concrete, Corros. Sci. 51 (9) (2009) 2027
2033.
[3] F.P. Torgal, S. Jalali, Eco-efficient Construction and Building Materials, Springer,
2011, pp. 75115.
[4] S.A. Bernal, R. Meja de Gutirrez, J.L. Provis, Engineering and durability
properties of concretes based on alkali-activated granulated blast furnace slag/
metakaolin blends, Constr. Build. Mater. 33 (2012) 99108.
[5] C. Monticelli, M. Criado, S. Fajardo, J.M. Bastidas, M. Abbottoni, A. Balbo,
Corrosion behaviour of a low Ni austenitic stainless steel in carbonated
chloride-polluted alkali-activated fly ash mortar, Cem. Concr. Res. 55 (2014)
4958.
[6] L. Reig, M.M. Tashima, M.V. Borrachero, J. Monz, C.R. Cheeseman, J. Pay,
Properties and microstructure of alkali-activated red clay brick waste, Constr.
Build. Mater. 43 (2013) 98106.
[7] F. Pacheco-Torgal, Z. Abdollahnejad, A.F. Cames, M. Jamshidi, Y. Ding,
Durability of alkali-activated binders: a clear advantage over Portland
cement or an unproven issue?, Constr Build. Mater. 30 (2012) 400405.
[8] J.L. Provis, J.S.J. van Deventer, Alkali-Activated Materials: State-of-the-Art
Report, RILEM TC 224-AAM, Springer/RILEM, Dordrecht, 2014, pp. 242249.
[9] W.A. Chaparro, J.H.B. Ruiz, R.J.T. Gmez, Corrosion of reinforcing bars
embedded in alkali-activated slag concrete subjected to chloride attack,
Mater. Res. 15 (1) (2012) 5762.
[10] M. Holloway, J.M. Sykes, Studies of the corrosion of mild steel in alkali-
activated slag cement mortars with sodium chloride admixtures by a
galvanostatic pulse method, Corros. Sci. 47 (12) (2005) 30973110.
[11] F. Puertas, A. Fernndez-Jimnez, M.T. Blanco-Varela, Pore solution in alkali-
activated slag cement pastes. Relation to the composition and structure of
calcium silicate hydrate, Cem. Concr. Res. 34 (1) (2004) 139148.
[12] A. Gruskovnjak, B. Lothenbach, L. Holzer, R. Figi, F. Winnefeld, Hydration of
alkali-activated slag: comparison with ordinary Portland cement, Adv. Cem.
Res. 18 (3) (2006) 119128.
[13] C. Arya, J.B. Newman, An assessment of four methods of determining the free
chloride content of concrete, Mater. Struct. 23 (1990) 319330.
[14] M. Cyr, P. Rivard, F. Labrecque, A. Daidi, High-pressure device for fluid
extraction from porous materials: application to cement-based materials, J.
Am. Ceram. Soc. 91 (8) (2008) 26532658.
[15] R.S. Barneyback, S. Diamond, Expression and analysis of pore fluids from
hardened cement pastes and mortars, Cem. Concr. Res. 11 (2) (1981) 279285.
[16] S. Song, H.M. Jennings, Pore solution chemistry of alkali-activated ground
granulated blast-furnace slag, Cem. Concr. Res. 29 (2) (1999) 159170.
[17] A. Alhozaimy, R.R. Hussain, A. Al-Negheimish, R. Al-Zaid, D.D.N. Singh, Effect of
simulated concrete pore solution chemistry, chloride ions, and temperature on
passive layer formed on steel reinforcement, ACI Mater. J. 111 (4) (2014) 411
421.
 W. Wang et al. / Construction and Building Materials 143 (2017) 289297
[18] R. Liu, L. Jiang, J. Xu, C. Xiong, Z. Song, Influence of carbonation on chloride-
induced reinforcement corrosion in simulated concrete pore solutions, Constr.
Build. Mater. 56 (2014) 1620.
[19] P. Ghods, O.B. Isgor, G.A. McRae, G.P. Gu, Electrochemical investigation of
chloride-induced depassivation of black steel rebar under simulated service
conditions, Corros. Sci. 52 (5) (2010) 16491659.
[20] N.S. Allan, The Influence of Binder Type and Cracking on Reinforcing Steel
Corrosion in Concrete (Ph.D. thesis), Department of Civil Engineering,
University of Cape Town, 2004.
[21] C.Q. Ye, R.G. Hu, S.G. Dong, X.J. Zhang, R.Q. Hou, R.G. Du, C.J. Lin, J.S. Pan, EIS
analysis on chloride-induced corrosion behavior of reinforcement steel in
simulated carbonated concrete pore solutions, J. Electroanal. Chem. 688 (2013)
275281.
[22] A. Poursaee, C.M. Hansson, Reinforcing steel passivation in mortar and pore
solution, Cem. Concr. Res. 37 (7) (2007) 11271133.
[23] P. Ghods, O.B. Isgor, G. McRae, T. Miller, The effect of concrete pore solution
composition on the quality of passive oxide films on black steel reinforcement,
Cem. Concr. Compos. 31 (1) (2009) 211.
[24] ASTM G102-89, Standard Practice for Calculation of Corrosion Rates and
Related Information from Electrochemical Measurements, American Society
for Testing and Materials, West Conshohocken, 2015. 7 pages.
[25] ASTM G59-97, Standard Test Method for Conducting Potentiodynamic
Polarization Resistance Measurements, American Society for Testing and
Materials, West Conshohocken, 2014. 4 pages.
[26] V.S. Ramachandran, J.J. Beaudoin, Handbook of Analytical Techniques in
Concrete Science and Technology: Principles, Techniques and Applications,
Elsevier, 2000, pp. 468472.
[27] K.K. Sideris, A.E. Savva, Durability of mixtures containing calcium nitrite based
corrosion inhibitor, Cem. Concr. Compos. 27 (2005) 277287.
[28] P.K. Mehta, P.J.M. Monteiro, Concrete: Microstructure, Properties,
and Materials, third ed., McGraw-Hill, The United States of America, 2006, p.
178.
297
[29] ASTM G106-89, Standard Practice for Verification of Algorithm and
Equipment for Electrochemical Impedance Measurements, American Society
for Testing and Materials, West Conshohocken, 2015. 11 pages.
[30] P. Garcs, M.C. Andrade, A. Saez, M.C. Alonso, Corrosion of reinforcing steel in
neutral and acid solutions simulating the electrolytic environments in the
micropores of concrete in the propagation period, Corros. Sci. 47 (2) (2005)
289306.
[31] ASTM F2129-15, Standard Test Method for Conducting Cyclic
Potentiodynamic Polarization Measurements to Determine the Corrosion
Susceptibility of Small Implant Devices, American Society for Testing and
Materials, West Conshohocken, 2015. 9 pages.
[32] F.P. Glasser, Mineralogical aspects of cement in radioactive waste disposal,
Mineral. Mag. 65 (5) (2001) 621633.
[33] M.G. Alexander, H.-D. Beushausem, F. Dehn, P. Moyo, Concrete Repair,
Rehabilitation and Retrofitting II, Taylor & Francis Group, London, 2009, pp.
311317.
[34] F. Wenger, J. Galland, Study of corrosion of steel in concrete by electrochemical
impedance measurements, Mater. Sci. Forum 44 & 45 (1989) 375380.
[35] Seawater. https://en.wikipedia.org/wiki/Seawater.
[36] Q. Xie, W. Chen, Corrosion behavior of 16Mn low alloy steel in sulfide-
containing Bayer solutions, Corros. Sci. 86 (2014) 252260.
[37] M. Salasi, T. Shahrabi, E. Roayaei, M. Aliofkhazraei, The electrochemical
behaviour of environment-friendly inhibitors of silicate and phosphonate in
corrosion control of carbon steel in soft water media, Mater. Chem. Phys. 104
(1) (2007) 183190.
[38] W. Stericker, Protection of small water systems from corrosion, Ind. Eng.
Chem. 37 (8) (1945) 716720.
[39] F.R. Foulkes, P. McGrath, A rapid cyclic voltammetric method for studying
cement factors affecting the corrosion of reinforced concrete, Cem. Concr. Res.
29 (6) (1999) 873883.
[40] D.E. Macphee, H.T. Cao, Theoretical description of impact of blast furnace slag
(BFS) on steel passivation in concrete, Mag. Concr. Res. 45 (162) (1993) 6369.