You are on page 1of 12

Electrochimica Acta 261 (2018) 66e77

Contents lists available at ScienceDirect

Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta

A comparative study of mild steel passivation embedded in


Belite-Ye'elimite-Ferrite and Porland cement mortars
Guilherme Yuuki Koga a, b, *, Blandine Albert a, Virginie Roche b,
Ricardo Pereira Nogueira b, c
a
LafargeHolcim Research Center, 95 Rue Du Montmurier, 38291 Saint Quentin Fallavier, France
b
LEPMI, UMR5279 CNRS, Grenoble INP, Universit e de Savoie, Universit
e Joseph Fourier, 1130, Rue de la Piscine, BP 75, 38402 Saint Martin d'H
eres, France
c
Gas Research Center, Khalifa University of Science and Technology, P.O. Box 2533, Abu Dhabi, United Arab Emirates

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

Article history: An extensive experimental program was designed to investigate the passivation of reinforced low
Received 9 November 2017 carbon-release Belite-Ye'elimite-Ferrite (BYF) cement mortars compared to the passivation of reinforced
Received in revised form conventional Portland equivalents (OPC). The influence of formulation parameters such as the water-to-
12 December 2017
cement ratio and the presence of fly ash on the steel rebar passivation was evaluated by conventional
Accepted 18 December 2017
Available online 19 December 2017
corrosion measurements such as corrosion potential readings and linear polarization resistance. Visual
observations were carried out as a complement. A negligible corrosion rate was reached after 28 days
with BYF cement against 14 days with OPC equivalent. This difference is attributed to the pH of the
Keywords:
Passivation
medium in the BYF mortars, which is lower than in the OPC mortars at a very early age, before becoming
Mild steel strongly basic. The exploitation of the Tafel slopes showed that the corrosion current can be evaluated by
Reinforced concrete the Stern and Geary equation with an identical B coefficient for the samples of BYF and OPC basis. Despite
Sulfoaluminate cement some differences compared to the passivation in Portland mortars, the present study clearly points to the
Portland cement BYF mortars intrinsically passivating nature regarding mild steel reinforcing bars.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction properties of fresh (e.g. workability) and hardened concretes (e.g.


mechanical and barrier resistances) [7,8]. However, if BYF is to
Belite-Ye'elimite-Ferrite cements, referred to as BYF, are a replace OPC binders in a wide range of concrete applications, it
promising lower carbon-release alternative to ordinary Portland must likewise be able to passivate ordinary black steel re-
cements (OPC) [1e4]. The base component is a clinker composed of inforcements. Once embedded in OPC-based concrete, the steel is
ye'elimite (Ca4Al6O12(SO4)), belite (Ca2SiO4), and calcium alumi- spontaneously passivated. The transposition to BYF-based con-
noferrite solid solution (Ca2(Al,Fe)2O5), stated here in decreasing cretes is questionable.
order of content [2]. The clinker production of BYF requires less The available literature on the investigation of the corrosion
limestone in the raw materials, a lower clinkering temperature, and behaviour of steel embedded in sulfoaluminate cement matrices (to
less energy to be ground compared to Portland clinker which BYF family belongs) are scarce and not conclusive [9e11].
manufacturing [4,5]. Industrial feasibility was checked and in-situ Based on visual inspection of the steel reinforcement embedded in
measurements confirmed overall CO2 savings between 25% and calcium sulfoaluminate (CSA) mortars after 14 years of service,
30% [6]. The partial substitution of BYF cement by supplementary Glasser and Zhang [9] concluded that steel remains passivated
cementitious materials (SCMs) brings an additional reduction of within the matrix. However, Kalogridis et al. [10] reported that the
the amount of CO2 through the clinker substitution and by the lower pH values and the higher average pore diameters of BYF
valorization of waste by-products such as fly ash. Moreover, the matrix, compared to Portland one, expose the steel rebar to a
addition of fly ash to cement may improve the engineering higher corrosion risk, as suggested by corrosion potential readings
and weight loss measurements. Janotka et al. [11] have performed
potentiodynamic polarization measurements using steel bars as
~o Carlos, Department of Mate-
working electrodes immersed in sulfoaluminate-belite mortar ex-
* Corresponding author. Federal University of Sa
rials Engineering, Rod. Washington Luis, km 235, S~
ao Carlos, CEP 13565-905, Brazil. tracts (pH~10.5). In line with Kalogridis et al. [10] results, it was
E-mail address: guilherme.koga@dema.ufscar.br (G.Y. Koga). observed that the steel did not exhibit passivation plateau upon

https://doi.org/10.1016/j.electacta.2017.12.128
0013-4686/© 2017 Elsevier Ltd. All rights reserved.
G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77 67

anodic polarization. The available data are insufficient to an un- better understanding of the differences and similarities between
ambiguous conclusion to be drawn upon this issue and an extensive the passivation of the steel reinforcement embedded in BYF and
electrochemical survey is lacking to ascertain whether mild steel OPC mortars. Also, the suitability of the current recommendations
spontaneously passivates or not when embedded in BYF matrices in and standards to assess the corrosion risk and corrosion rate of the
normal service conditions. steel embedded in BYF mortars was discussed.
For a new family of binder arises also the question of the method A last important introductory remark is that, while in the pre-
of evaluation of the steel corrosion. The suitability of the standards sent study we focused on the inherent passivating properties of BYF
(e.g. ASTM C876 [12]) and recommendations (e.g. RILEM TC 154- matrix compared to OPC one, it is, of course, necessary to extend
EMC [13]) to the study the corrosion behaviour of reinforced BYF this study to include the presence of aggressive agents such as
structures should be experimentally verified since most of them chloride ions and their impact on the BYF matrix performance in
have been established for reinforced Portland structures. For terms of reinforcement corrosion. This is an issue of major concern
instance, the American Standard ASTM C 876-09 provides a crite- and will be a matter of another specific paper.
rion to correlate the risk of corrosion and the potential reading
values. However, this criterion was empirically derived from 2. Experimental
chloride-induced corrosion of mild steel embedded in bridges
decks in the USA for specific conditions of Portland concrete type, 2.1. Materials
layer, and exposition [14e16]. The RILEM TC 154-EMC [13], in turn,
provides recommendations to evaluate the corrosion rate of steel 2.1.1. Steel reinforcement
reinforcement in concrete by means of the linear polarization Thermo-mechanically-treated (TMT) steel ribbed rebars
resistance method (LPR). In this case, an issue of major concern is measuring 14 mm in nominal diameter and 130 mm long were
that the conversion of polarization resistance values, R, into used as mortar reinforcement, Fig. 1. The chemical composition of
corrosion current densities, icorr, requires the previous knowledge the mild steel used in this study is detailed in Table 1. The values are
of the coefficient B of the Stern-Geary relationship (icorr ¼ B/R), the average of four analyses, three at the edge and one at the centre
which is related to the Tafel slopes ba and bc [17]. For Portland- of cross-section samples, which gave the same chemical composi-
based reinforced structures, B values of 26 and 52 mV for steel in tion considering the precision of the equipment.
the active and passive state, respectively, have been recommended Fig. 2a and b show the microstructure of the steel reinforcement,
[13], but no reliable information about the B values for reinforced which is composed of a ductile and tough ferrite (brighter phase)
BYF structures is available. and pearlite (darker compound) matrix with a high resistant sur-
In this paper, the intrinsic ability of the BYF mortar to effectively face rim of tempered martensite. This resulting composite micro-
induce mild steel reinforcement passivation was investigated. The structure is characteristic of the thermomechanical processing in
study was conducted at different hydration conditions (water-to- which hot steel bars coming out of last rolling mill stands are
cement ratio, w/c, of 0.40, 0.50 and 0.67) and in the presence or not quenched through a series of water jets. During such slow cooling,
of fly ash (FA), in normal curing and service conditions, that is, in the heat released from the core tempers the hardened surface while
the absence of chloride or carbonation using OPC equivalents as the core itself turns into ferrite-pearlite aggregates, providing an
reference. Different electrochemical techniques such as corrosion optimum combination of high strength, ductility, and bendability.
potential measurements, LPR, and potentiodynamic polarization
curves were used to evaluate the corrosion risk and the corrosion 2.1.2. Cements
rate. Coupling different electrochemical techniques provided a An ordinary Portland cement (OPC), designation CEM I 52.5 CE

Fig. 1. Steel ribbed rebar used to reinforce mortar samples. Ruler in cm.
68 G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77

Table 1
Chemical composition of the steel rebar (working electrode) used to reinforce mortars.

Ca Sa Cu Ni Si Cr Mo Mn P V Ti Al Fe
0.21 0.03 0.43 0.21 0.2 0.068 0.041 0.72 0.011 <0.015 <0.010 <0.006 Bal.

Elementary composition determined by spark optical emission spectrometry.


a
Chemical determination by high-frequency furnace and infrared detection.

Fig. 2. Cross-sectional optical micrographs (500x, after Nital 2% etching) taken at a) the core of the sample and b) at the edge. A similar microstructure with equiaxial grains is
observed in longitudinal-section micrographs.

CP2 NF according to the standard NF EN 197-1 [18], and a recently in the mould and Fig. 3b shows the surface preparation of the rebar
developed BYF cement were used. The composition of the BYF and surface and the reinforced mortar specimen.
OPC cements and the calculated amount of the clinker phases in the Six different mortar mixes detailed in Table 3 were used. Each
cements are compiled in Table 2. binder was mixed with water and aggregate according to EN 196-1
[19] and cast in cylindrical moulds containing reinforcements
2.1.3. Preparation of reinforced mortars aligned with the axis of the cylinder. The moulds had a hole in its
Steel rebars were slightly brushed to remove the heteroge- bottom to facilitate the placement of the steel rebar before mortar
neously distributed spots of rust. Mill scale was however kept to pouring, as shown in Fig. 3a. The moulds were filled under vibra-
better mimic in-service conditions. After being cleaned in anhy- tion. Holed lids were used to cover the top of the moulds and to
drous ethanol, the ends of the rebar were tightly wrapped with keep them hermetically sealed with the steel well aligned. Just after
electrical tape before being insulated with epoxy immediately production and along the tests, the specimens were kept in a 20  C,
followed by the application of a heat-shrink tube over it. This 95% R.H. controlled chamber to ensure cement hydration while
procedure aimed to avoid boundary effects, such as crevice corro- keeping gaseous dioxygen available at the steel surface.
sion (differential aeration). The final non-insulated area of steel
rebars was about 31 cm2. The rebars were stored in a desiccator 2.2. Electrochemical measurements in reinforced mortars
until the production of the reinforced cylindrical mortars. Fig. 3a
illustrates the disposition of the reinforced mortar specimen curing Fig. 4 shows the corrosion cell used to carry out electrochemical

Table 2
Characteristics of the binders of the study (Rietveld quantitative phase analysis, readily soluble alkalis, and Blaine specific surface according to NF EN 196-6).

Rietveld Quantitative Phase Analysis


g/100 g

Phases OPC Phases BYF

Alite, Ca3SiO5 60.2 Ye'elimite, Ca4Al6O12(SO4) 31.1


b-Belite, Ca2SiO4 17.9 a0 -Belite, Ca2SiO4 29.0
Ferrite, Ca2(Al,Fe)2O5 10.5 b-Belite, Ca2SiO4 8.0
Aluminate, Ca3Al2O6 5.1 Ferrite, Ca2(Al,Fe)2O5 17.7
Gypsum, CaSO4.2(H2O) 1.8 Anhydrite, CaSO4 6.5
Hemihydrate CaSO4.0.5(H2O) 1.8 Gypsum, CaSO4.2(H2O) 1.5
Calcite, CaCO3 1.5 Periclase, MgO 0.9
Periclase, MgO 0.1 Other phases (inert) 5.9
Free lime, CaO 1.1

Readily Soluble Alkali

K2O 0.77 K2O 0.36


Na2O 0.07 Na2O 0.02

Blaine specific surface (cm2/g) 3680 Blaine specific surface (cm2/g) 4430
G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77 69

Fig. 3. a) Schematic illustration of the process to produce reinforced mortar specimens for electrochemical tests; b) Different steps of preparation of the electrochemical samples
since the surface normalization until the final specimen to be analysed on the right side. WE: working electrode.

Table 3
Mortar mixes used in the production of different reinforced mortars.

Designation Cement (g) Fly ash (g) w/b ratio Sand (g)a Superplasticizerb (%, dry/cement)

OPC (0.50) 450.0 e 0.50 1350 0.00


OPC þ FA 337.5 112.5 0.50c 1350 0.05
BYF (0.40) 511.4 e 0.40 1350 0.30
BYF (0.50) 450.0 e 0.50 1350 0.10
BYF (0.67) 375.0 e 0.67 1350 0.00
BYF þ FA 337.5 112.5 0.50c 1350 0.15
a
Standard sand according to NF EN 196-6.
b
CHRYSO® Fluid Premia 196.
c
w/b ¼ water-to-binder ratio ¼ ratio of mixing water per binder content, with binder ¼ sum of cement and fly ash.

Fig. 4. View of the test set-up used to carry-out electrochemical measurements.


70 G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77

measurements. The steel reinforcement was the working electrode using different new samples each time. Corrosion current den-
(WE, 31 cm2 of surface) and the electrolyte the interstitial pore sity values (icorr) were determined by extrapolating the anodic
solution which composition was in equilibrium with the unhy- and cathodic Tafel's regions on the polarization curves around
drated and hydrated solid and gel phases. The counter-electrode Ecorr with the intersection being considered as icorr.
(CE) was placed so as to wrap the mortar surface which was sur-  Linear Polarization Resistance (LPR): LPR measurements were
rounded by a moist sponge band to ensure proper electrical con- carried out with a 10 mV/min scan with ±10 mV amplitude scan
tinuity. CE was prepared from a co-axial carbon sheet, providing a around the open circuit potential. The apparent polarization
uniformly distributed electric field. A saturated calomel electrode resistance, Rp(app), was straightforwardly determined from the
(SCE) was placed perpendicularly to the rebar axis in a hole of DE/Di slope in the linear region of the anodic domain. As dis-
15 mm diameter in CE, being the tip of the SCE reference electrode cussed later in the Results and Discussion section, this pro-
in contact with the same pre-wetted sponge that assured electrical cedure differs from the usual one (DE/Di slope determined
continuity with the counter-electrode. around the corrosion potential) due to the high dielectric char-
Electrochemical measurements were regularly performed up to acteristic of the mortar layer. The real polarization resistance,
365 days. At each predetermined age, the specimens were removed Rp(real), was obtained from Rp(app) after correction for the elec-
from the humid chamber and connected to the potentiostat. The trolyte resistance, Re (the procedure for the estimation of Re is
samples dedicated to the non-destructive techniques were imme- described in Section 3.1.3). The repeatability of the LPR curves
diately placed back in the humid chamber after measurement to was verified by analysing four different samples.
allow the hydration process to continue in well-controlled condi-  Visual examination: Visual inspection at 365 days of curing was
tions for subsequent measurements. The specimens used in used to confirm the electrochemical indications. The specimens
destructive electrochemical methods were discarded after each were split along the embedded rebar to observe the steel j
test. A BioLogic potentiostat SP-300 with impedance channels was mortar interface.
used in the present work.

 Corrosion potential measurements: Corrosion potential readings 3. Results and discussion


were recorded versus curing time. The data acquisition and
interpretation were based on the ASTM C876-09 [12] and RILEM 3.1. Effect of different w/c in BYF (0.40, 0.50 and 0.67) compared to
TC 154-EMC [20]. Four different samples were monitored to OPC (w/c ¼ 0.50)
assure the repeatability.
 Potentiodynamic polarization curves: Corrosion overall behaviour The large number of parameters and the long-term (one year)
of reinforced mortars was evaluated by means of large poten- analysis would yield a huge number of experiments and related
tiodynamic polarization. Anodic and cathodic curves were ob- data to be treated. We hence proceeded to the extensive experi-
tained separately using different samples because of the mental survey of BYF w/c 0.50 and 0.67 as well as OPC w/c 0.50.
irreversible changes after such wide polarizations. The polari- Some complementary less-extensive experiments for the BYF w/c
zation of each anodic or cathodic branch started with a small 0.40 have been also performed to extend the information about the
polarization in the opposite direction (jDEj ¼ 30 mV) before a impact of the hydration conditions on the passivation of reinforced
10 mV/min scan was applied in the desired direction up to BYF mortars.
jDEj ¼ 1000 mV, both with respect to the open circuit potential,
Eocp. The obtained curves were corrected for the iR drop after the
measurements with the help of the electrochemical impedance 3.1.1. Corrosion potential monitoring
spectroscopy. Reproducibility of data was ensured by repeating Corrosion potential values, Ecorr, measured over 365 days are
the tests three times for each polarization, cathodic and anodic, shown in Fig. 5. First measurements were conducted just after
demoulding of the samples (24 h after specimen fabrication). The
lines at Ecorr ¼ 126 and 276 mV are intended as eye guides for
the empirical criteria for probability of corrosion suggested by the
ASTM C876 [12]. Their applicability to alternative binders such as
BYF has to be defined.
From Fig. 5, one can see that the initial values for the BYF series
are lower than for the OPC one; the lowest corresponding to the
BYF made with the highest water-to-cement ratio (w/c ¼ 0.67).
Between 1 and 60 days, the potential values increase remarkably
for the four series, probably indicating the formation of a passiv-
ating layer and, therefore, a decrease of the corrosion risk. The
values remain the highest for Portland cement. The threshold
of 126 mV is nevertheless exceeded for all series after 60 days.
Between 60 and 180 days, the increase of the potential values is
more gradual. Potential values then stabilize for the OPC and the
BYF 0.40 mortars but continue to increase for the BYF series 0.50
and 0.67, becoming higher than the OPC and BYF 0.40 after 240
days. The corrosion potential for all the series exceeded the
threshold value of 126 mV after 14 (OPC) and 60 (BYF) days of
fabrication. According to the ASTM C876 criteria, this should
correspond to a lower probability of corrosion. The suitability of the
Fig. 5. Corrosion potential values of the steel rebar embedded in different mortars. The
horizontal dashed lines represent the potential ranges and the corresponding corro-
criteria must nevertheless be confirmed by other measurements
sion risks suggested by the ASTM C876 [12]. Points presented as mean values and which allow the estimation of the corrosion rate for the BYF-based
standard deviations from four different samples. specimen.
G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77 71

3.1.2. Potentiodynamic polarization and ageing of the passive film, with values after 365 days ranging
Corrosion current density values have been straightforwardly between 415 and 680 mV/dec.
evaluated from the potentiodynamic polarization curves with the Potentiodynamic polarization results confirm that steel rein-
Tafel method. Fig. 6 shows the cathodic and anodic polarization forcement embedded in the BYF mortars exhibit negligible corro-
curves at 28 days for two reinforced OPC and two reinforced BYF sion rates as the steel rebar embedded in OPC mortars 14 days after
mortars (in each case, one specimen for the cathodic polarization fabrication up to 365 days. The corrosion process seems not to be
and one for the anodic one, all specimens w/c ¼ 0.50). Polarization limited by lack of oxygen or moisture content but controlled by the
curves for BYF (w/c 0.40 and 0.67) and OPC at different curing times presence of a passivating layer whose protective character in-
are qualitatively quite similar to those in Fig. 6 and are not shown creases with time.
here. The experimental values of the Tafel slopes allow us to calculate
Table 4 summarizes the ensemble of results, i.e., Tafel slopes, ba the coefficient B to be used in the Stern-Geary relationship together
and bc, and icorr. As previously mentioned, six samples were used with LPR measurements (a quicker and non-destructive alternative
for the measurements, three for the cathodic polarization and three to the Tafel extrapolations). The values obtained with the BYF vary
for the anodic one. The icorr values presented are therefore as the between 40 and 72 mV and are close to those obtained with OPC
average (and their standard deviations) of six curves. Anodic and which vary between 54 and 83 mV. In this sense, it is worth
cathodic Tafel slopes can be compared to the published data for noticing that the recommended B values for reinforced Portland
Portland cement [21e25], in the absence of values for steel concretes seem to be suitable to estimate the icorr values of rein-
embedded in BYF matrices. forced BYF mortars through the LPR technique.
From Table 4, it can be seen that all reinforced samples present
icorr values inferior to 0.1 mA/cm2, from 14 to 365 days, regardless
3.1.3. Linear polarization resistance (LPR)
of the cement type. The cathodic Tafel slopes show an increasing
LPR measurements were performed with the same samples
trend with time, indicating a slight decrease in the cathodic ki-
used for potential readings. Fig. 7 displays the polarization re-
netics. After 365 days, the values range between 156 and 270 mV/
sponses of both OPC and BYF series (w/c ¼ 0.50) at 28 days.
dec. These results are consistent with the reported values for oxy-
As expected for reinforced steel in concrete, the LPR curves in
gen reduction in alkaline solutions and for steel embedded in
Fig. 7a and b are not linear. Indeed, the shape of the curves for both
Portland mortars exposed for up to 12 months to 93% R.H [25,26].
OPC and BYF series at 28 days is typical of steel passivated in
Anodic Tafel slopes also increase with time, suggesting formation
mortar, when the reference electrode is placed on its surface [27]. In

Fig. 6. Semilogarithmic polarization curves of reinforced mortars after 28 days of curing and the respective Tafel extrapolation for a) OPC cathodic branch, b) OPC anodic branch, c)
BYF cathodic branch, and d) BYF anodic branch. Anodic and cathodic curves are obtained separately using different reinforced mortar samples. All samples produced with w/
c ¼ 0.50.
72 G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77

Table 4
Values of the electrochemical parameters from Tafel extrapolation of the cathodic and anodic potentiodynamic polarization curves.

Days Sample ba (mV/dec.) bc (mV/dec.) Bc (mV) icorr (mA/cm2)

14 OPC (w/c ¼ 0.50) 364 ± 3 190 ± 5 54 ± 1 0.09 ± 0.01


BYF(w/c ¼ 0.50)a 338 ± 8 156 ± 37 46 ± 8 0.04 ± 0.01
BYF (w/c ¼ 0.67) 360 ± 30 120 ± 10 40 ± 1 0.07 ± 0.01
28 OPC (w/c ¼ 0.50) 410 ± 20 190 ± 10 57 ± 1 0.077 ± 0.003
BYF(w/c ¼ 0.50) 373 ± 50 157 ± 37 48 ± 10 0.05 ± 0.01
BYF (w/c ¼ 0.67) 404 ± 40 150 ± 1 46 ± 2 0.079 ± 0.005
56 OPC (w/c ¼ 0.50) 420 ± 20 220 ± 4 62 ± 1 0.09 ± 0.01
BYF(w/c ¼ 0.50) Not meas. Not meas. Not meas. Not meas.
BYF (w/c ¼ 0.67) 403 ± 20 185 ± 3 55 ± 1 0.10 ± 0.02
84 OPC (w/c ¼ 0.50) 480 ± 10 241 ± 1 69 ± 1 0.06 ± 0.02
BYF(w/c ¼ 0.50) 403 ± 1 165 ± 40 50 ± 9 0.02 ± 0.01
BYF (w/c ¼ 0.67) 424 ± 36 190 ± 30 57 ± 8 0.09 ± 0.01
181 OPC w/c ¼ 0.50)b 470 ± 10 250 ± 10 71 ± 1 0.069 ± 0.003
BYF(w/c ¼ 0.50) 377 ± 8 194 ± 18 56 ± 4 0.056 ± 0.003
BYF (w/c ¼ 0.67) 473 ± 4 200 ± 10 60 ± 1 0.06 ± 0.01
365 OPC (w/c ¼ 0.50) 680 ± 13 270 ± 4 83 ± 12 0.09 ± 0.02
BYF(w/c ¼ 0.50) 415 ± 14 156 ± 10 49 ± 2 0.02 ± 0.01
BYF (w/c ¼ 0.67) 538 ± 16 241 ± 3 72 ± 1 0.066 ± 0.004
a
Measured at 15 days.
b
Measured at 177 days. Not meas. ¼ Not measured.
c
B ¼ (ba.bc/2.303(baþ bc)).

Fig. 7. Experimental polarization curves around Eocp for a) reinforced OPC mortar, and b) reinforced BYF mortar. Measurements carried out at 28 days of curing using jDEj ¼ 20 mV
and scan rate of 10 mV/min.

these conditions, it senses the capacitive response of the subjacent As mentioned before, B values used for OPC mortars seem to be
mortar layer which is dominated by the low-frequency transients also suitable for BYF cements. The icorr values have thus been
produced by the sudden application of the polarization at the calculated from LPR measurements, taking B ¼ 26 mV and
beginning of the scan. These transients have been associated with assuming a maximum error factor of two [13], and they were
the capacitive response of the mortar [27,28]. As the potential ramp compared with those obtained directly by Tafel extrapolation
gets closer to the open circuit potential (Eocp), the current density (Table 5). Good agreements between the values from the two
presents a linear behaviour, which means that the potential scan methods are observed. LPR measurements are more interesting to
rate allows the current response to approximate that of steady- obtain instantaneous icorr values as the method is non-destructive
state conditions. For potentials beyond Eocp, the measured slope and quick compared to the Tafel method.
of current density vs. potential, Rp, has been considered to as the Fig. 9 shows the evolution of the corrosion current density (icorr)
sum of Rp,true and Re [28], being Re taken as the real axis interception over time. The values can be compared to the RILEM recommen-
of the high-frequency loop d assimilated to the dielectric response dations [13]. Values of icorr below 0.1 mm/cm2 correspond to negli-
of the mortar layer d of the Nyquist electrochemical impedance gible corrosion rates while values between 0.1 and 0.5 mA/cm2
diagram [29], as illustrated in Fig. 8. correspond to low corrosion level.
The corrosion current density, icorr, is then calculated over 365 The icorr values decrease sharply during the first 7 days, being
days according to the Stern-Geary equation [30]: the highest for the BYF (w/c ¼ 0.67) while the values are similar for
the others BYF and OPC series. Although further research is needed
RpðrealÞ ¼ Rp  Re (1) to understand this difference, a reasonable explanation is that
higher w/c ratios possibly induces bleeding water and plastic set-
!
B h i tlement which generates local low pH values at the steel j mortar
icorr ¼ mA=cm2 (2) interface [31,32].
RpðrealÞ
G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77 73

Fig. 8. Determination of Re from the Nysquist diagram for a) reinforced OPC mortar, and b) reinforced BYF mortar.

Table 5
Comparison between the corrosion current densities, icorr, calculated from LPR and obtained by Tafel extrapolation.

Samples Method icorr, 102 (mA/cm2)

14 days 28 days 84 days ~181 days 365 days

OPC (w/c ¼ 0.50) LPR 8.1 ± 0.4 6.8 ± 0.4 7.2 ± 0.9 6.1 ± 0.9 4.9 ± 0.5
Tafel 9±1 7.7 ± 0.3 6±2 6.9 ± 0.3 9±2
BYF (w/c ¼ 0.50) LPR 10 ± 2 8±1 6±1 6±1 6±1
Tafel 5±1 4±1 3±1 5.6 ± 0.3 3±1
BYF (w/c ¼ 0.67) LPR 13 ± 03 10 ± 2 8±2 6±1 5±1
Tafel 7±1 7.9 ± 0.5 9±1 6±1 6.6 ± 0.4

corrosion risk levels indicated by the corrosion potential mea-


surements, Fig. 5, performed on the same samples at the same
curing times. The corrosion level is then negligible after one or
several weeks, depending on the cement type and water-to-cement
ratios.

3.1.4. Impact of the addition of fly ash


The addition of fly ash (FA) had no significative impact on the
corrosion-related parameters, Ecorr and icorr, for the OPC samples
(Fig. 10a and c, respectively). OPC samples with FA showed steady
potential values slightly higher than in the absence of FA and clearly
and permanently in the low-risk range, with corrosion current
values almost identical to the ones in the absence of FA all over the
365 days of testing.
It is worth noticing however that, for the BYF matrices, even if
no clear corrosion tendency has been identified, the effect of the
presence of FA was not negligible as for the OPC. As seen in Fig. 10b,
the increase of Ecorr with time was much less pronounced than in
the absence of FA and the final values were only slightly higher than
Fig. 9. Indication of the variation of icorr with time for reinforced BYF and OPC cement the threshold of the intermediate/unknown corrosion risk. In a
mortars. The values are calculated from LPR measurements and Stern-Geary rela- similar tendency, icorr values in the presence of FA (Fig. 10d),
tionship, taken B ¼ 26 mV. Points are presented as mean values and standard de-
viations from four different samples.
although still low, have shown a slight but monotonic increase with
time reaching the lower limit of the uncertainty domain. No
significative corrosion attack has nevertheless been identified after
Corrosion current densities for the four series continue to visual observation as shown in the next section.
decrease between 7 and 28 days, but more gradually and then Another interesting effect of the FA additions is on the electro-
stabilized more or less between 28 and 365 days. Despite the initial chemical response of the cementitious bulk (i.e. of the mortar
differences among the samples, the icorr finally reaches similar resistance, Re). For OPC samples with no FA, the Re increases by a
values after 120 days, indicating negligible corrosion rates for all factor of 5 between 1 and 365 days, whereas for OPC samples with
investigated specimens. FA addition, the factor is about 37 (Fig. 11a). Similar trends have
Before reaching a negligible corrosion level, the LPR results been reported [33,34] and related to the pozzolanic reaction which
show that the initial icorr values are superior to 0.1 mA/cm2, corre- decreases the porosity and the connectivity of pores. Also, fly ash
sponding to non-negligible corrosion rates. This is in line with the may be used as fine granulates and upon hydration it has the
74 G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77

Fig. 10. Effect of addition of fly ash in the corrosion potential (A, B) and corrosion current density (C, D) of reinforced OPC and BYF mortars.

Fig. 11. Impact of the addition of FA on the mortar resistance values of a) reinforced OPC specimens, and b) reinforced BYF specimens.

capability of partially obstructing voids and pores. Even before in the hydration process of CSA cement [37].
reacting, fly ash particles that are composed of cenospheres (hollow In any case, because of the non-negligible resistance of the
spheres) can modify the dielectric behaviour of a material in which mortar compared to the measured polarization resistance from LPR
they are incorporated [35]. in presence of fly ash, the correct IR compensation becomes
Considering the samples based on BYF cements, Fig. 11b, for mandatory to estimate the icorr even when the steel rebar is highly
samples with no FA addition, the resistance increases by a factor of passivated.
about 10 from 1 day to 365 days of curing whereas, for BYF samples
with FA, the factor is about 70. It is hard to find an explanation in 3.2. Visual examination
the literature as the data on the role of fly ash blended to BYF or CSA
cements are still limited. A study pointed out that FA may accelerate Different electrochemical measurements point out that the
the formation of ettringite, Ca6Al2(SO4)3(OH)12.26H2O, yielding to a corrosion of the steel embedded in the BYF and OPC mortars is
denser microstructure which could explain, in part, the initial in- negligible. Additional reinforced mortar samples were made and
crease of Re observed [36]. At longer-term, fly ash can be expected split after 365 days of curing to observe the presence of any
to remain inert particles due to the absence of portlandite, Ca(OH)2, corrosion product.
known to initiate pozzolanic reactions. But this assumption is For OPC specimens with and without FA, no visible corrosion
contradicted by a recent study showing the contribution of fly ash product is observed, as seen in Fig. 12a and b. Similarly, no corrosion
G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77 75

products are observed after splitting reinforced BYF with w/c of a decrease of corrosion risk) and the decrease of icorr (cf Fig. 9,
0.40, 0.50, and containing FA (Fig. 12c, d, and f, respectively). For decrease of the corrosion rates) suggest the formation of an effec-
these samples, the active corrosion during the first days of hydra- tive passivating layer on the reinforcing steel surface embedded in
tion is not enough to promote visible “rust” after 365 days of the BYF matrices used in the present study. The presence of this passive
survey. These observations confirm the fact that the corrosion rate layer is also consistent with the results from polarization curves. In
of the steel rebar embedded in OPC and BYF cement mortars drops addition, the usual evaluation criteria (ASTM C876 and RILEM) for
rapidly ensuring steel protection since early ages. the Portland cement matrices appear to be applicable for the BYF
Some signs of corrosion are found on the steel rebar surface cement matrices of our study, as shown in Fig. 13 by the good
embedded in BYF (w/c ¼ 0.67) mortar for 365 days, Fig. 12e. Pre- correlation between the Ecorr and icorr,LPR. Indeed, as it clearly ap-
sumably, the corrosion spots were certainly formed at early age pears from the figure, the points for all different samples merge in a
while the pH of the pore solution is below 12. It is interesting to single cloud indicating that the higher the corrosion potential, the
note that the corrosion products are exclusively concentrated at the lower the corrosion current density with values in overall good
lower half of ribs of the rebars. This observation points out that agreement with the ones stated in the usual procedures.
steel j mortar interface is one of the most dominant influencing This means that the survey and study of the corrosion behaviour
parameters together with pH [38]. The ribs regions are more prone of steel embedded in reinforced BYF matrices are possible with the
to present higher density of metallurgical defects and residual current standards (e.g. ASTM C876 [12]) and recommendations (e.g.
stress from the thermomechanical processing and heat treatment. RILEM TC 154-EMC [13]) in the same conditions and with the same
Additionally, porosity and bleeding may be more pronounced at assessment criteria. Furthermore, the B value required to calculate
lower rib region due to the casting direction [31]. the corrosion current density from LPR measurements is quite the
same for the BYF-based samples than for the OPC ones.
4. Final remarks
4.2. Influence of the pore solution chemistry on steel passivation
4.1. Suitability of the current practices to assess steel corrosion in
BYF mortars The time required for steel passivation is an indicator of the
protectiveness ability of the interstitial pore solution. The pore
Both the increase of Ecorr during hydration (cf Fig. 5, indication of solution chemistry affects the initial rate of the oxide film growth so

Fig. 12. Steel j mortar interfaces, after 365 days, of a) OPC (w/c ¼ 0.50), b) OPC þ FA, c) BYF (w/c ¼ 0.40), d) BYF (w/c ¼ 0.50), e) BYF (w/c ¼ 0.67), and f) BYF þ FA.
76 G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77

Fig. 14. pH evolution over one day (1440 min.) of pore solutions extracted from BYF
Fig. 13. Experimental correlation between Ecorr and icorr obtained by LPR, taking B (w/c of 0.50 and 0.67) and OPC (w/c ¼ 0.50) cement pastes.
equal to 26 mV. Points of all sample series plotted regardless curing time.

[9e11].
that differences in the electrochemical responses can be ascribed to
the differences of the electrolyte upon hydration [39,40].
The interstitial solution during the hydration of the Portland 5. Conclusions
cement rapidly becomes highly alkaline, reaching pH values above
13 after the first minutes. This strongly alkaline medium is known  The results pointed out that mild steel effectively passivates in
to promote the formation of a passivation layer which assures BYF cement mortars. The time required for the corrosion current
negligible corrosion rates. The corrosion potential values measured density to drop to values typically considered to be passive
correspond well to a low corrosion risk and the current density to (<0.1 mA/cm2) was approximately 28 days regardless the w/c
negligible corrosion rates. ratio and fly ash additions. Corrosion potential took even longer
The results of the first week after the fabrication of the samples time to reach the values related to low probability of corrosion
indicate, however, the possibility of a slight corrosion to take place (Ecorr > 126 mV/SCE), especially for BYF þ FA.
due to the time needed to form a sufficiently protective layer. The  Cathodic Tafel slopes were ranged between 156 and 270 mV/dec
reinforced OPC mortars require 2 days to their icorr values to reach which are consistent with the reported Tafel slopes for oxygen
the transition between passive and active corrosion (0.1 mA/ reduction. Anodic Tafel slopes, after 365 days, were ranged be-
cm2  icorr  0.2 mA/cm2) and 7 days to reach values below 0.1 mA/ tween 415 and 680 mV/dec. It was found that ba increases with
cm2. Similarly, Poursaee and Hansson [41] have shown that the time (formation and ageing of the passive film). Calculation of
time required for the corrosion rate of steel in Portland mortar to the B values from experimental Tafel slopes showed that the
drop to a rate typically considered to be passive is approximately 7 recommended values (26 mV for active and 52 mV for passive
days and that it continues to decrease slowly for a long period. states) are a reasonable choice for reinforced BYF as for OPC
For the BYF cement, the pH of the interstitial solution is less mortars.
alkaline during the first hours (around 10.5) and then it increases  Visual inspection indicated that the corrosion rates of steel
sharply. Fig. 14 shows the pH evolution over 1440 min (1 day) for embedded in OPC and BYF drop quickly to negligible values and
BYF and OPC pore extracts. no corrosion products are visible after 365 days, except for BYF
The initial lower pH values could be an explanation for the initial (w/c ¼ 0.67). In this case, the local low pH and the bleeding
lower corrosion potential values and higher current density surrounding the ribs of the steel rebar embedded in BYF (w/
compared to the OPC samples. The electrochemical responses c ¼ 0.67) mortars generated non-negligible initial corrosion
(corrosion potential measurements and LPR curves) confirmed that rates for a sufficient period to form visible corrosion products.
the initial low alkalinity of BYF matrices may expose the steel Steel was however fully passivated after 28 days and the
reinforcement to a higher corrosion risk and higher corrosion rate corrosion products observed after 365 days are attributed to the
compared to OPC analogous. However, the subsequent increase of active initial period.
pH after the first hours assures the protection of the steel  Good agreement between the electrochemical techniques and
embedded in BYF mortars thanks to the formation of an effective visual inspections indicated that the ASTM C876 and RILEM
passive layer as for the reinforced Portland mortars. recommendations can be used to assess the corrosion of steel in
Thus, the scenario seen for the BYF matrices of this study is as reinforced BYF cement mortars. Some controversial aspects of
follows: a non-passivating pore solution at very early age (initial steel protection in sulfoaluminate matrices were discussed.
few hours) turning into a protective electrolyte within the hydra-
tion progress. It is also worth mentioning that the different possible Acknowledgements
pH values depending on the composition of the BYF cements and
the progress of the hydration reactions may be an explanation of The authors gratefully acknowledge the financial support of the
the controversial conclusions found for different studies in the Brazilian institution CNPq (grant number 229886/2013-2), of the
literature about steel corrosion in CSA or BYF cement matrices French institution ANRT (grant number 2013/1440), and of the
G.Y. Koga et al. / Electrochimica Acta 261 (2018) 66e77 77

LafargeHolcim group. [22] A.A. Gürten, K. Kayakirilmaz, M. Erbil, The effect of thiosemicarbazide on
corrosion resistance of steel reinforcement in concrete, Construct. Build.
Mater. 21 (2007) 669e676, https://doi.org/10.1016/j.conbuildmat.2005.
References 12.010.
voa, The electrochemical behaviour of steel
[23] J.M. Deus, B. Díaz, L. Freire, X.R. No
[1] E. Gartner, H. Hirao, A review of alternative approaches to the reduction of rebars in concrete: an Electrochemical Impedance Spectroscopy study of the
CO2 emissions associated with the manufacture of the binder phase in con- effect of temperature, Electrochim. Acta 131 (2014) 106e115, https://doi.org/
crete, Cem. Concr. Res. 78 (2015) 126e142, https://doi.org/10.1016/ 10.1016/j.electacta.2013.12.012.
j.cemconres.2015.04.012. [24] M. Pour-Ghaz, O.B. Isgor, P. Ghods, The effect of temperature on the corrosion
[2] E.M. Gartner, D.E. MacPhee, A physico-chemical basis for novel cementitious of steel in concrete. Part 1: simulated polarization resistance tests and model
binders, Cem. Concr. Res. 41 (2011) 736e749, https://doi.org/10.1016/ development, Corrosion Sci. 51 (2009) 415e425, https://doi.org/10.1016/
j.cemconres.2011.03.006. j.corsci.2008.10.034.
[3] E. Gartner, Industrially interesting approaches to “low-CO2” cements, Cem. [25] B. Elsener, Corrosion rate of steel in concrete - measurements beyond the
Concr. Res. 34 (2004) 1489e1498, https://doi.org/10.1016/j.cemconres. Tafel law, Corrosion Sci. 47 (2005) 3019e3033, https://doi.org/10.1016/
2004.01.021. j.corsci.2005.06.021.
[4] M. Schneider, M. Romer, M. Tschudin, H. Bolio, Sustainable cement production [26] P. Garces, M.C. Andrade, A. Saez, M.C. Alonso, Corrosion of reinforcing steel in
- present and future, Cem. Concr. Res. 41 (2011) 642e650, https://doi.org/ neutral and acid solutions simulating the electrolytic environments in the
10.1016/j.cemconres.2011.03.019. micropores of concrete in the propagation period, Corrosion Sci. 47 (2005)
[5] C. Shi, A.F. Jimenez, A. Palomo, New cements for the 21st century: the pursuit 289e306, https://doi.org/10.1016/j.corsci.2004.06.004.
of an alternative to Portland cement, Cem. Concr. Res. 41 (2011) 750e763, [27] K.V. Subramaniam, M. Bi, Investigation of the local response of the steel-
https://doi.org/10.1016/j.cemconres.2011.03.016. concrete interface for corrosion measurement, Corrosion Sci. 51 (2009)
[6] E. Gartner, C. Comparet, B. Albert, Project aether, Int. Cem. Rev. (2014) 1976e1984, https://doi.org/10.1016/j.corsci.2009.05.025.
106e107. [28] J.A. Gonza lez, A. Molina, M.L. Escudero, C. Andrade, Errors in the electro-
[7] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious mate- chemical evaluation of very small corrosion rates - I. polarization resistance
rials, , Cem. Concr. Res. 41 (2011) 1244e1256, https://doi.org/10.1016/ method applied to corrosion of steel in concrete, Corrosion Sci. 25 (1985)
j.cemconres.2010.12.001. 917e930, https://doi.org/10.1016/0010-938X(85)90021-6.
[8] K.L. Scrivener, A. Nonat, Hydration of cementitious materials, present and [29] S.J. Ford, J.D. Shane, T.O. Mason, Assignment of features in impedance spectra
future, Cem. Concr. Res. 41 (2011) 651e665, https://doi.org/10.1016/ of the cement-paste/steel system, Cem. Concr. Res. 28 (1998) 1737e1751,
j.cemconres.2011.03.026. https://doi.org/10.1016/S0008-8846(98)00156-2.
[9] F.P. Glasser, L. Zhang, High-performance cement matrices based on calcium [30] M. Stern, A.L. Geary, Electrochemical polarization I. A theoretical analysis of
sulfoaluminate-belite compositions, Cem. Concr. Res. 31 (2001) 1881e1886, the shape of polarization curves, J. Electrochem. Soc. 104 (1957) 56e63,
https://doi.org/10.1016/S0008-8846(01)00649-4. https://doi.org/10.1149/1.2428653.
[10] D. Kalogridis, G.C. Kostogloudis, C. Ftikos, C. Malami, Quantitative study of the [31] U.M. Angst, B. Elsener, C.K. Larsen, Ø. Vennesland, Chloride induced rein-
influence of non-expansive sulfoaluminate cement on the corrosion of steel forcement corrosion: electrochemical monitoring of initiation stage and
reinforcement, Cem. Concr. Res. 30 (2000) 1731e1740, https://doi.org/ chloride threshold values, Corrosion Sci. 53 (2011) 1451e1464, https://
10.1016/S0008-8846(00)00277-5. doi.org/10.1016/j.corsci.2011.01.025.
[11] I. Janotka, L. Kraj ci, A. Ray, S.C. Mojumdar, The hydration phase and pore [32] D. Thomas, Concrete Durability, first ed., CRC Press, New York, 2014.
structure formation in the blends of sulfoaluminate-belite cement with [33] M.F. Montemor, A.M.P. Simo ~es, M.M. Salta, Effect of fly ash on concrete
Portland cement, Cem. Concr. Res. 33 (2003) 489e497, https://doi.org/ reinforcement corrosion studied by EIS, Cement Concr. Compos. 22 (2000)
10.1016/S0008-8846(02)00994-8. 175e185, https://doi.org/10.1016/S0958-9465(00)00003-2.
[12] ASTM C876-09, Standard Test Method for Half-cell Potentials of Uncoated [34] M.F. Montemor, A.M.P. Simo ~es, M.M. Salta, M.G.S. Ferreira, The assessment of
Reinforcing Steel in Concrete, ASTM International, West Conshohocken, 2009, the electrochemical behaviour of fly ash-containing concrete by impedance
https://doi.org/10.1520/C0876-09.2. spectroscopy, Corrosion Sci. 35 (1993) 1571e1578, https://doi.org/10.1016/
[13] C. Andrade, C. Alonso, R. Polder, R. Cigna, O. Vennesland, M. Salta, 0010-938X(93)90386-U.
A. Raharinaivo, B. Elsener, Test methods for on-site corrosion rate measure- [35] W. Shoude, L. Lingchao, C. Xin, Effect of ultra-fine fly ash on the dielectric
ment of steel reinforcement in concrete by means of the polarization resis- behavior of CFSC under stress, Adv. Mater. Lett. 2 (2011) 12e16, https://
tance method, Mater. Struct. 37 (2004) 623e643. doi.org/10.5185/amlett.2010.9163.
[14] R.F. Stratfull, The corrosion of steel in a reinforced concrete bridge, Corrosion [36] S. Ioannou, L. Reig, K. Paine, K. Quillin, Properties of a ternary calcium
13 (1957) 173e179. sulfoaluminate-calcium sulfate-fly ash cement, Cem. Concr. Res. 56 (2014)
[15] R.F. Stratfull, W.J. Jurkovich, D.L. Spellman, Corrosion Testing of Bridge Decks, 75e83, https://doi.org/10.1016/j.cemconres.2013.09.015.
1975. California. [37] L.H.J. Martin, F. Winnefeld, E. Tschopp, C.J. Müller, B. Lothenbach, Influence of
[16] D.L. Spellman, R.F. Stratfull, Concrete variables and corrosion testing, in: fly ash on the hydration of calcium sulfoaluminate cement, Cem. Concr. Res.
Highw. Res. Rec, 1973, pp. 27e45. 95 (2017) 152e163, https://doi.org/10.1016/j.cemconres.2017.02.030.
[17] M. Stern, R.M. Roth, Anodic behavior of iron in acid solutions, J. Electrochem. [38] U. Angst, B. Elsener, C.K. Larsen, Ø. Vennesland, Critical chloride content in
Soc. 104 (1957) 390e392, https://doi.org/10.1149/1.2428588. reinforced concrete - a review, Cem. Concr. Res. 39 (2009) 1122e1138,
[18] Cement d Part 1: Composition, Specifications and Conformity Criteria for https://doi.org/10.1016/j.cemconres.2009.08.006.
Common Cements, NF EN 197e1, 2012. [39] S.M. Abd El Haleem, E.E. Abd El Aal, S. Abd El Wanees, A. Diab, Environmental
[19] Methods of Testing Cementd Part 1: Determination of Strength, NF EN factors affecting the corrosion behaviour of reinforcing steel: I. The early stage
196e1, 2006. of passive film formation in Ca(OH)2 solutions, Corrosion Sci. 52 (2010)
[20] B. Elsener, C. Andrade, J. Gulikers, R. Polder, M. Raupach, Half-cell potential 3875e3882, https://doi.org/10.1016/j.corsci.2010.07.035.
measurements e potential mapping on reinforced concrete structures, Mater. [40] A. Poursaee, Corrosion of steel bars in saturated Ca(OH)2 and concrete pore
Struct. 36 (2003) 1e11. solution, Concr. Res. Lett. 1 (2010) 90e97.
[21] C. Cao, M.M.S. Cheung, B.Y.B. Chan, Modelling of interaction between [41] A. Poursaee, C.M. Hansson, Reinforcing steel passivation in mortar and pore
corrosion-induced concrete cover crack and steel corrosion rate, Corrosion Sci. solution, Cem. Concr. Res. 37 (2007) 1127e1133, https://doi.org/10.1016/
69 (2013) 97e109, https://doi.org/10.1016/j.corsci.2012.11.028. j.cemconres.2007.04.005.