Study of depassivation of carbon steel in simulated concrete pore solution using different equivalent circuits

Construction and Building Materials 157 (2017) 357–362

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Study of depassivation of carbon steel in simulated concrete pore solution using different equivalent circuits Guojian Liu, Yunsheng Zhang ⇑, Meng Wu, Ran Huang School of Materials Science and Engineering, Southeast University, Nanjing 21189, China

h i g h l i g h t s
Depassivation process of reinforcing steels subjected to chloride is investigated.
Chloride threshold value has been obtained by electrochemical monitoring methods.
Basic principles to determine the best fitted equivalent circuits were established.

a r t i c l e

i n f o

Article history: Received 2 July 2017 Received in revised form 6 September 2017 Accepted 18 September 2017 Available online 23 September 2017 Keywords: Corrosion Depassivation Electrochemical methods Equivalent circuits

a b s t r a c t Depassivation and corrosion of reinforcing steels submitted to chloride in simulated concrete pore solution was studied with electrochemical methods. Measurements included half-cell potential (HCP), linear polarization resistance (LPR) and electrochemical impedance spectroscopy (EIS). Three different equivalent circuits, i.e., R(QR), R(QR)(QR) and R(Q(R(QR))), were proposed to fit and elucidate the EIS data, respectively. Results showed that large-scale depassivation and corrosion of steel occurred when chloride content exceeded 0.05 mol/L. All the three equivalent circuits revealed a good fitting degree of Nyquist plots at early stage, while only circuit R(Q(R(QR))) still retained a most appropriate fitting at severe corrosion stage, of which the chi-square ranged from 3.1E4 to 8.5E4 throughout the whole exposure test. Verification could be obtained that circuit R(Q(R(QR))) with two time constants was effective to fit and explain the depassivation process in regard to steel corrosion induced by chloride in simulated concrete pore solution. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Reinforcing steels in cementitious materials normally maintains immune to corrosion due to the thin passive film attributed to the high degree of alkalinity of hydrated cement and concrete. The passive film is considered to be around 20 nm thick and mainly consists of an inner layer rich in FeII oxides/oxyhydroxides and an outer layer containing FeIII oxides/oxyhydroxides [1,2]. Depassivation or breakdown of passive film occurs when carbonation and/or chloride ingress take place to a certain extent. In scenario of chloride, it is believed that once the chloride content at the steel surface reaches chloride threshold value (CTV), with moisture and oxygen in presence, rapid localized corrosion of steel is initiated [3–6]. An accurate determination and a good understanding of the depassivation process of reinforcing steel together with the

⇑ Corresponding author. E-mail addresses: [email protected] (G. Liu), [email protected] (Y. Zhang). https://doi.org/10.1016/j.conbuildmat.2017.09.104 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.

CTV is of great importance for better clarifying corrosion behavior, assessing durability issues and predicting residual service life of reinforced structures. Researchers performed numerous investigations regarding CTV [7–11], but no consensus has been reached either on the experimental configuration or on the expression of CTV itself. For example, CTV in terms of total chloride or acidsoluble chloride in concrete ranged from 0.04% to 8.34% (bw%) and from 0.07% to 1.16% (bw%) expressed by free chloride or water-soluble chloride [12]. It should be noted that the large scatter of reported values of CTV is due to many influencing factors such as steel type, experiment setup, concrete mix, moisture availability and so forth [13–15]. Therefore, many researchers have conducted tests in a saturated Ca(OH)2 solution (pH  12.5) [16–18] as a simulated concrete pore solution. Apart from the reduction in experiment time, a good reproducibility and representativeness could also be ensured and thereby provide significant results for cementitious materials.

358

G. Liu et al. / Construction and Building Materials 157 (2017) 357–362

Equivalent circuits have been widely adopted as an effective approach to simulate and interpret electrochemical impedance results. The simple equivalent circuit of R(QR) is a modified Randles circuit [19] where a pure capacitance is replaced by a constant phase element. While Yuan [20] utilized Rs(RctQdl)(RfQf) to simulate corrosion process of 304 stainless steel and attribute the resistance at high frequency to charge transfer resistance, researchers [21–23] have assumed that charge transfer resistance is associated with the time constant at low frequency. The equivalent circuit of R (Q(R(QR))) was used to investigate the corrosion behavior of duplex stainless steel [24] and corrosion of coated steels [25]. However, a detailed comparison between different equivalent circuits needs to be discussed with regard to fitting degree and physical meaning of each circuit element. In the present work, the depassivation process of carbon steel was investigated and the electrochemical impedance results were analyzed using three different equivalent circuits. The fitting degree between measured data and the proposed simulation circuit was evaluated by Chi-square value method. Additionally, the physical meaning of each component obtained by different equivalents was tried to be explained.

3. Results and discussion 3.1. HCP and corrosion current density

Steel specimens with a diameter of 16 mm and length of 10 mm were cut from carbon steel rod. Chemical composition of steel samples (by weight) was 0.20% C, 0.55% Si, 1.42% Mn, 0.028% S, 0.026% P and the balance Fe. One cross-section surface was polished to grade 1000 as exposure surface, degreased in acetone and then washed in distilled water. The remaining surface was sealed by epoxy resin. A wire was soldered to the other cross-section surface for electrochemical testing. The saturated Ca(OH)2 solution was prepared using distilled water with some insoluble Ca(OH)2. NaCl was added to saturated Ca(OH)2 solution stepwise, 0.01 mol/L each day. All chemical reagents applied were analytical reagent grade.

A series of identical steel samples with adequate prepassivation are submitted to chloride added stepwise by 0.01 mol/L each day. The evolutions of half-cell potential (Ecorr) are presented in Fig. 1. It can be seen that the average Ecorr of specimens after ten days of pre-passivation reads around 230 mV vs. SCE. With increasing concentrations of chloride, the Ecorr decreases gradually and shifts to more negative values, which implies a higher corrosion risk. Then a noteworthy drop of Ecorr occurs when the chloride content reaches 0.05 mol/L. This dramatic drop of open circuit potential is supposed to be indicative of relatively large scale breakdown of passive film and initiation of active corrosion. Samples subjected to chloride above 0.05 mol/L should be the most prone to corrosion, with respect to –those with lower chloride concentrations. It should be noted that the scatter of Ecorr values among the samples is a consequence of the stochastic nature of passive film. While corrosion thermodynamics is reflected by Ecorr, reaction kinetics of passive film could be further understood with Icorr changes depicted in Fig. 2. Icorr keeps between 0.25 lA/cm2 and 0.40 lA/cm2 at low chloride concentrations. However, once chloride exceeds 0.05 mol/L, Icorr increases to values greater than 0.97 lA/cm2. The calculated Icorr is an average corrosion rate, while the specific corrosion current density in corrosion pit could be several times higher [8]. From abovementioned evolution of Ecorr and Icorr, the CTV could be assumed to be in the vicinity of 0.05 mol/L in simulated concrete pore solution. Another commonly used expression of chloride threshold value is Cl/OH molar ratio, which means free chloride versus hydroxyl ion in the concrete pore solution. Cl/OH could reflect corrosion risk induced by either chloride or carbonation effect in reinforced concrete structures. Based on the simulated concrete pore solution of saturated Ca(OH)2 in this study, a sudden drop of Ecorr and a noteworthy increase in Icorr occurred when Cl/OH ratio reached in between 1.12 and 1.34, of which the chloride contents were 0.05 and 0.06 mol/L, respectively.

2.2. Electrochemical techniques

3.2. EIS analysis

All steel samples were immersed in saturated Ca(OH)2 solution for ten days to obtain stable pre-passivation before adding NaCl. After a sufficiently steady condition was achieved, electrochemical tests were performed using PARSTAT4000 immediately before addition of NaCl each day. A classic three-electrode system was adopted with steel sample being working electrode, a saturated calomel electrode (SCE) and a platinum electrode as reference and counter electrode, respectively. The Icorr as an index of corrosion current density was usually calculated by linear polarization based on the Stern–Geary method [26] as in Eq. (1)

3.2.1. Nyquist plot and data fitting The typical EIS plot in Nyquist format is shown in Fig. 3. Topology evolution of Nyquist plots also reveals substantial information about the depassivation process of steels [6,11]. Theoretically, the

2.1. Specimen preparation and exposure conditions

Icorr ¼

ba bc 1 B ¼ 2:303ðba þ bc Þ Rp Rp

ð1Þ

where Icorr is the corrosion current density, Rp the polarization resistance, Ba constant related to ba and bc, the anodic and cathodic slopes of Tafel curve, respectively. Though B is usually assumed to vary from 26 mV to 52 mV [27], a value of 26 mV was used in present study. The Rp was obtained directly by built-in software and the exposed area of nearly 2.0 cm2 was used as an input parameter. EIS test was performed with a perturbation amplitude 10 mV and a frequency range of 105–102 Hz. HCP was determined after stable open circuit potential was achieved. In LPR tests, reinforcing steels were polarized to ±10 mV at Ecorr with a scan rate of 10 mV/min.

-150 -200 -250

Ecorr (mV)

2. Experimental program

-300 -350 -400 -450 -500 0.00

0.02

0.04

0.06

0.08

Chloride content (mol/L) Fig. 1. Evolution of Ecorr with increasing chloride concentrations.

0.10

359

G. Liu et al. / Construction and Building Materials 157 (2017) 357–362

8

Equivalent circuit A

Icorr (µA/cm2)

6

Equivalent circuit B

4

2

Equivalent circuit C

0 0.00

0.02

0.04 0.06 0.08 Chloride content (mol/L)

0.10

Fig. 2. Evolution of Ecorr with increasing chloride concentrations.

capacitive loop intersects the X-axis of real part at one point in low-frequency part which corresponds to the charge transfer property of the passive film. However, a series of incomplete depressed semicircles have been observed in the present study, partially due to the equipment limit. The radius of the semicircle in the lowfrequency domain gradually reduces as the concentration of chloride increases. The trend from ‘‘linear” to ‘‘semicircular” yielded decreasing values of potential intersection points, leading to reduced charge transfer resistance of passive film. Three equivalent circuits shown in Fig. 4 were proposed to simulate the measured EIS data for further understanding of depassivation process. Equivalent circuit A, referred to herein as EC A, contains one time constant, while both EC B and EC C contain two time constants. Rs is the solution resistance and the constant phase element (CPE) in EC A represents the double-layer capacitance of the steel/electrolyte interface. In EC B and EC C, CPE1 corresponds to capacitive behavior of the passive film and R1 deals with the ionic resistance of the paths through the passive film. CPE2 is the capacitive behavior of double layer capacitance in the localized corrosion area while Rct stands for the charge transfer resistance of the corrosion process. CPEs are widely used in data fitting to allow for depressed semicircles attributable to the heterogeneity of electrode interfaces. The impedance ZCPE is defined as in Eq. (2).

Z CPE ¼ Y 1 0 ðjxÞ

n

ð2Þ

Fig. 4. Equivalent circuits of reinforcing steel in simulated pore solutions.

where Y0 is the capacitance value of CPE, x is the angular frequency, j is the imaginary number and n is the CPE exponent. Y0 is proportional to the double layer capacitance of pure capacitive electrodes and n represents the deviated degree of the capacitance from the ideal condition, which ranges from 0 to 1. The CPE becomes a capacitor when n = 1, to a Warburg element representing semi-infinite length diffusion phenomena when n = 0.5 and to a resistor when n = 0. Figs. 5–7 include measured data as well as fitting curve yielded using the equivalent circuits shown in Fig. 4, respectively. The scattered symbols represent measured data and the solid lines are fitted curve obtained by applying equivalent circuits. In all cases, all three equivalent circuits, namely R(QR), R(QR)(QR) and R(Q(R(QR))), show a good agreement between experimental and theoretical curves in early stage of corrosion before chloride content increases up to 0.05 mol/L. Once large-scale localized depassivation happens, impedance plot depicts a sharp contraction. In Fig. 5b, a notable discrepancy has been observed between experimental and calculated values of the impedance when chloride exceeds 0.06 mol/L. While in Fig. 6b, the fitted curves obtained by EC B with two time constants connected in series reveal relatively better agreement than EC A with exception when chloride content reaches 0.08 mol/L. However, the measured data and simulated curve have shown a high fitting degree in Fig. 7a and b throughout the whole exposure experiment even when the chloride content is as high as 0.08 and 0.09 mol/L. From visual examination of fit plots yielded by different equivalent circuits proposed, the complex circuits of EC B and EC C can well simulate the measured data. However, more accurate

12.0k

a

250.0k

b

0 0.03M 0.05M 0.06M 0.07M 0.08M 0.09M

150.0k 0 0.03M 0.05M 0.06M 0.07M 0.08M 0.09M

100.0k 50.0k 0.0 0.0

40.0k

80.0k

120.0k 2

Zr (ohm·cm )

160.0k

-Zi (ohm·cm2)

-Zi (ohm·cm2)

200.0k 8.0k

4.0k

0.0 0.0

4.0k

8.0k

12.0k

2

Zr (ohm·cm )

Fig. 3. Typical measured Nyquist plots of specimen with additions of chloride: (a) full view of Nyquist plot; (b) enlarged view of high frequency response.

360

G. Liu et al. / Construction and Building Materials 157 (2017) 357–362

250.0k

12.0k

a

0 0.03M 0.05M 0.06M 0.07M 0.08M 0.09M line: fitted curve

b

150.0k 0 0.03M 0.05M 0.06M 0.07M 0.08M 0.09M line: fitted curve

100.0k 50.0k 0.0 0.0

40.0k

80.0k

120.0k

-Zi (ohm·cm2)

-Zi (ohm·cm2)

200.0k 8.0k

4.0k

0.0

160.0k

0.0

2

4.0k

8.0k

12.0k

2

Zr (ohm·cm )

Zr (ohm·cm )

Fig. 5. Measured and fitted EIS spectra obtained by equivalent circuit A: (a) full view of Nyquist plot; (b) enlarged view of high frequency response.

250.0k

12.0k

a

0 0.03M 0.05M 0.06M 0.07M 0.08M 0.09M line: fitted curve

b

150.0k 0 0.03M 0.05M 0.06M 0.07M 0.08M 0.09M line: fitted curve

100.0k 50.0k 0.0 0.0

40.0k

80.0k

120.0k

-Zi (ohm·cm2)

-Zi (ohm·cm2)

200.0k 8.0k

4.0k

0.0 0.0

160.0k

4.0k

Zr (ohm·cm2)

8.0k

12.0k

Zr (ohm·cm2)

Fig. 6. Measured and fitted EIS spectra obtained by equivalent circuit B: (a) full view of Nyquist plot; (b) enlarged view of high frequency response.

250.0k

12.0k

a

b

0 0.03M 0.05M 0.06M 0.07M 0.08M 0.09M line: fitted curve

150.0k 0 0.03M 0.05M 0.06M 0.07M 0.08M 0.09M line: fitted curve

100.0k 50.0k 0.0 0.0

40.0k

80.0k

120.0k

-Zi (ohm·cm2)

-Zi (ohm·cm2)

200.0k 8.0k

4.0k

0.0 0.0

160.0k

Zr (ohm·cm2)

4.0k

8.0k

12.0k

Zr (ohm·cm2)

Fig. 7. Measured and fitted EIS spectra obtained by equivalent circuit C: (a) full view of Nyquist plot; (b) enlarged view of high frequency response.

evaluation method needs further discussion in regard to determining appropriate equivalent circuits. 3.2.2. Fitting degree verification The fitting degree between measured data and simulated results obtained by proposed equivalent circuits is quantitatively evaluated by chi-square (v2) method [28,29] as shown in Eqs. (3) and (4):

v2 ¼

N X ½Z re;i  Z re ðxi Þ2 þ ½Z im;i  Z im ðxi Þ2 i¼1

jZðxÞi j2

ð3Þ

Table 1 Chi-square value of different equivalent circuits at varying chloride contents. Chloride content mol/L

Chi-square (v2) value EC A

EC B

EC C

0 0.03 0.05 0.06 0.07 0.08 0.09

1.65E03 1.26E03 1.28E03 2.43E03 6.96E03 6.85E03 1.42E02

1.56E04 1.33E03 8.77E04 7.14E04 8.17E04 7.71E03 1.04E03

9.96E04 8.56E04 9.80E04 7.10E04 7.15E04 3.19E04 6.97E04

361

G. Liu et al. / Construction and Building Materials 157 (2017) 357–362

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi jZðxÞi j ¼ Z 2re;i þ Z 2im;i

ð4Þ

where Zre, i and Zim, i are measured data, Zre (xi) and Zim (xi) are fitted values, |Z (x)i| is the modulus of impedance used as a weighing factor herein, which indicates the value of systematic resistance to alternating current signal. It is obvious that v2 indicates the variance of data and lower v2 are attributed to a better quality of the fitting.

0.016 R(QR) R(QR)(QR) R(Q(R(QR)))

Chi-square

0.012

0.008

0.004

0.000 0.00

0.02

0.04

0.06

0.08

0.10

Chloride content (mol/L) Fig. 8. Evolution of Chi-square values at varying chloride contents.

The calculated Chi-square values of different equivalent circuits at varying chloride contents are listed in Table 1. It can be seen that chi-square value of EC A has fallen into the range of 103 and reaches 0.0142 at 0.09 mol/L of chloride content. While fitting degree of EC B fluctuates between 7.71  103 and 1.56  104, the values of v2 yielded by EC C stabilizes within order of magnitudes of 104. Evolution of Chi-square values at varying chloride contents is shown in Fig. 8, which depicts the best fitting agreement of EC C more intuitively. 3.2.3. Simulated equivalent circuit parameters Visual observation to determine the consistency of experimental results and fitted plots has been carried out to evaluate feasibility of proposed equivalent circuits as well as quantitative verification by Chi-square method. However, another important selection criterion for potential equivalent circuit is the physical meaning of simulated circuit element parameters. Bearing this in mind, the calculated values of each circuit element obtained by EC A, EC B and EC C are listed in Tables 2–4, respectively. In Table 2, the value of Rs undergoes a steady decreasing evolution with increasing additions of chloride ions. Other circuit element parameters including Y0, n and especially Rct depict a good changing trend. However, this is not convincing in predicting actual electrochemical reaction due to the high discrepancy between experimental data and fitting results. While in Table 3, it is difficult to explain that the value of Rs increased from 0.1313 X cm2 to 158.2 X cm2 and then decreased from 134.1 X cm2 to 0.10 X cm2. Furthermore, similar irregular fluctuation occurred in the values of Y0 and n of

Table 2 Fitting results at different chloride contents by EC A. Cl (mol/L)

Rs (Ohmcm2)

Y0 (Scm 0.00 0.03 0.05 0.06 0.07 0.08 0.09

Rct (ohmcm2)

CPE

211.6 175.1 158.1 142.6 132.1 121.1 111.3

2

n

sec )

n

3.67E05 3.66 E05 3.71 E05 6.15 E05 7.40 E05 19.24 E05 12.54 E05

0.9296 0.9299 0.9305 0.8670 0.8325 0.7949 0.7979

724400 527800 391700 10740 7887 9924 10460

Table 3 Fitting results at different chloride contents by EC B. Cl (mol/L)

Rs (ohmcm2)

R1 (ohmcm2)

CPE1 2

sec )

Y0 (Scm 0.00 0.03 0.05 0.06 0.07 0.08 0.09

0.0103 0.1313 158.2 143.5 134.1 0.10 0.0264

n

4.76E10 3.38E10 1.26E04 3.26E04 6.64E05 1.92E04 1.25E04

2

n 1 1 1 0.8024 0.9066 0.795 0.798

Rct (ohmcm2)

CPE2 sec )

Y0 (Scm 211.6 175.1 48430 5597 4256 9921 12460

3.67E05 3.66E05 5.11E05 6.60E05 5.01E04 3.68E10 7.91E10

R1 (ohmcm2)

CPE2

n

n 0.9256 0.93 0.9153 0.9203 0.7893 1 1

724300 527700 425700 6088 4962 121 111.3

Table 4 Fitting results at different chloride contents by EC C. Cl﹣(mol/L)

Rs (ohmcm2)

CPE1 2

Y0 (Scm 0.00 0.03 0.05 0.06 0.07 0.08 0.09

212.3 175.6 158.5 143.9 134.6 123.3 114.1

3.56E05 3.57E05 3.63E05 5.43E05 5.61E05 1.51E04 8.87E05

n

sec )

n 0.9355 0.9376 0.9373 0.8971 0.8974 0.8636 0.8815

Rct (ohmcm2) 2

Y0 (Scm 528600 437500 341300 8555 5121 5313 5069

4.92E05 9.42E05 1.56E04 6.16E04 5.17E04 5.40E04 4.09E04

n

sec )

n 0.986 1 0.9999 0.716 0.6639 0.5217 0.5078

576800 249700 131000 3326 4521 14390 33840

362

G. Liu et al. / Construction and Building Materials 157 (2017) 357–362

CPE2. Fitted results yielded by applying EC C is shown in Table 4. Declining Rs values are attributable to decreasing solution resistance due to added chloride. Under chloride attack, passive film gradually breaks down, resulting in higher value of Y0 and lower value of n in CPE2. The Rct value shows a noteworthy decrease by two orders in magnitude, reflecting initial depassivation of steel reinforcements. From abovementioned discussion, physical meaning of fitted results obtained by different equivalent circuits should be borne in mind even though Chi-square is relatively low. 4. Conclusion Electrochemical methods were utilized to monitor depassivation process of carbon steel submitted to chloride attack. The impedance results were simulated using three different equivalent circuits. Furthermore, fitting degree between measured data and calculated data by proposed equivalent circuit was evaluated by Chi-square value method. According to the results, the following concluding remarks could be drawn: (1) Large-scale localized depassivation of carbon steel occurs in simulated concrete pore solution when the chloride concentration reaches around 0.05 mol/L, accompanied by a dramatic drop of half-cell potential and an increase in corrosion rate. Thus the CTV could be determined as 0.05 mol/L for steel depassivation in this case. (2) The proposed equivalent circuit of R(Q(R(QR))) is effective to simulate and interpret EIS results of steel corrosion. Furthermore, it is recommended that the time constant within low frequency domain is more appropriate to be associated with charge transfer resistance of passive film. (3) Principles to determine applicable equivalent circuit for EIS data have been obtained. The first step includes visual observation and quantitative comparison between experimental data and simulated results. Chi-square method allows for evaluation of abovementioned curves and a good fitting degree is reached when the value of v2 falls within the range of 104. Nevertheless, physical meaning of each circuit element should be assured prior to choosing best fitted equivalent circuit. Pointless or abnormal data along with experiment time elapse indicate inadaptability of proposed circuit.

Acknowledgments The authors gratefully acknowledge financial support from 973 Program (2015CB655102) and Natural Science Foundation of China (51378116 and 51678143). References [1] P. Ghods, O. Burkan Isgor, F. Bensebaa, D. Kingston, Angle-resolved XPS study of carbon steel passivity and chloride-induced depassivation in simulated concrete pore solution, Corros. Sci. 58 (2012) 159–167. [2] S. Joiret, M. Keddam, X.R. Nóvoa, M.C. Pérez, C. Rangel, H. Takenouti, Use of EIS, ring-disk electrode, EQCM and Raman spectroscopy to study the film of oxides formed on iron in 1 M NaOH, Cem. Concr. Compos. 24 (2002) 7–15. [3] C. Arya, F.K. Ofori-Darko, Influence of crack frequency on reinforcement corrosion in concrete, Cem. Concr. Res. 26 (1996) 345–353.

[4] G. Blancoa, A. Bautista, H. Takenouti, EIS study of passivation of austenitic and duplex stainless steels reinforcements in simulated pore solutions, Cem. Concr. Compos. 28 (2006) 212–219. [5] 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 (2010) 1649–1659. [6] G. Liu, Y. Zhang, Z. Ni, R. Huang, Corrosion behavior of steel submitted to chloride and sulphate ions in simulated concrete pore solution, Constr. Build. Mater. 115 (2016) 1–5. [7] D. Trejo, R.G. Pillai, Accelerated chloride threshold testing: Part I. ASTM A 615 and A 706 reinforcement, ACI Mater. J. 100 (2003) 519–527. [8] C. Alonsoa, C. Andradea, M. Castellotea, P. Castrob, Chloride threshold values to depassivate reinforcing bars embedded in a standardized OPC mortar, Cem. Concr. Res. 30 (2000) 1047–1055. [9] D. Trejo, P.J. Monteiro, Corrosion performance of conventional (ASTM A615) and low-alloy (ASTM A706) reinforcing bars embedded in concrete and exposed to chloride environments, Cem. Concr. Res. 35 (2005) 562–571. [10] W. Morris, A. Vico, M. Vazquez, S.R. de Sanchez, Corrosion of reinforcing steel evaluated by means of concrete resistivity measurements, Corros. Sci. 44 (2002) 81–99. [11] J. Shi, W. Sun, J. Jiang, Y. Zhang, Influence of chloride concentration and prepassivation on the pitting corrosion resistance of low-alloy reinforcing steel in simulated concrete pore solution, Constr. Build. Mater. 111 (2016) 805–813. [12] U. Angst, B. Elsener, C.K. Larsen, Ø. Vennesland, Critical chloride content in reinforced concrete — a review, Cem. Concr. Res. 39 (2009) 1122–1138. [13] K.Y. Ann, H.W. Song, Chloride threshold level for corrosion of steel in concrete, Corros. Sci. 49 (2007) 4113–4133. [14] D. Izquierdo, C. Alonso, C. Andrade, M. Castellote, Potentiostatic determination of chloride threshold values for rebar depassivation: experimental and statistical study, Electrochim. Acta 49 (2004) 2731–2739. [15] Z. Jin, X. Zhao, T. Zhao, L. Yang, Corrosion and 3D crack-propagation behaviors in reinforced concrete subjected to bending load in simulated marine environment, Int. J. Electrochem. Sci. 11 (2016) 8779–8796. [16] S.M.A. Haleem, E.E. Abd El Aal, S. Abd El Wanees, A. Diab, Environmental factors affecting the corrosion behavior of reinforcing steel: I. The early stage of passive film formation in Ca(OH)2 solutions, Corros. Sci. 52 (2010) 3875– 3882. [17] L. Jiang, G. Huang, J. Xu, et al., Influence of chloride salt type on threshold level of reinforcement corrosion in simulated concrete pore solutions, Constr. Build. Mater. 30 (2012) 516–521. [18] L.V. Zˇulj, M. Serdar, S. Martinez, Effect of tartrate on the electrochemical behaviour and semiconductive properties of passive film on steel in saturated calcium hydroxide, Mater. Corros. 66 (2015) 1344–1353. [19] J.A. Gonzalez, A. Molina, M.L. Escudero, C. Andrade, Errors in the electrochemical evaluation of very small corrosion rates – I. Polarization resistance method applied to corrosion of steel in concrete, Corros. Sci. 25 (1985) 917–930. [20] S. Yuan, B. Liang, Y. Zhao, S.O. Pehkonen, Surface chemistry and corrosion behaviour of 304 stainless steel in simulated seawater containing inorganic sulphide and sulphate-reducing bacteria, Corros. Sci. 74 (2013) 353–366. [21] O. Poupard, A. Aït-Mokhtar, P. Dumargue, Corrosion by chlorides in reinforced concrete: determination of chloride concentration threshold by impedance spectroscopy, Cem. Concr. Res. 34 (2004) 991–1000. [22] C. Monticelli, M.E. Natali, A. Balbo, C. Chiavari, F. Zanotto, S. Manzi, M.C. Bignozzi, A study on the corrosion of reinforcing bars in alkali-activated fly ash mortars under wet and dry exposures to chloride solutions, Cem. Concr. Res. 87 (2016) 53–63. [23] L. Yohai, M.B. Valcarce1, M. Vázquez1, Testing phosphate ions as corrosion inhibitors for construction steel in mortars, Electrochim. Acta 202 (2016) 316– 324. [24] S. Jin, C.-M. Liu, X. Lin, M.-J. Li, H.-B. Cao, H.-Z. Wang, Effect of inorganic anions on the corrosion behavior of UNS S32750 duplex stainless steel in chloride solution, Mater. Corros. 66 (2015) 1077–1083. [25] C. Liu, Q. Bi, A. Leyland, A. Matthews, An electrochemical impedance spectroscopy study of the corrosion behaviour of PVD coated steels in 0.5 N NaCl aqueous solution: Part II. EIS interpretation of corrosion behavior, Corros. Sci. 45 (2003) 1257–1273. [26] M. Stern, A.L. Geary, Electrochemical polarization I A theoretical analysis of the shape of polarization curves, J. Electrochem. Soc. 104 (1957) 56–63. [27] C. Andrade, J.A. Gonzalez, Quantitative measurements of corrosion rate of reinforcing steels embedded in concrete using polarisation resistance measurements, Mater. Corros. 29 (1978) 515–519. [28] S. Surviliene, A. Lisowska-Oleksiak, A. Cˇešuniene, Effect of ZrO2 on corrosion behaviour of chromium coatings, Corros. Sci. 50 (2008) 338–344. [29] X. Zhang, S.O. Pehkonen, N. Kocherginsky, G.A. Ellis, Copper corrosion in mildly alkaline water with the disinfectant monochloramine, Corros. Sci. 44 (2002) 2507–2528.