Corrosion performance of blended concretes exposed to different aggressive environments

Construction and Building Materials 121 (2016) 704–716

Contents lists available at ScienceDirect

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

Corrosion performance of blended concretes exposed to different aggressive environments Ana María Aguirre-Guerrero a,⇑, Ruby Mejía-de-Gutiérrez a, Maria João Ribeiro Montês-Correia b a b

Composites Materials Group (CENM), School of Materials Engineering, Calle 13 # 100-00, Edif. 349, 2 piso, Universidad del Valle, Cali, Colombia Laboratório Nacional de Engenharia Civil, Avenida do Brasil 101, 1700-066 Lisboa, Portugal

h i g h l i g h t s
Impress voltage and wetting/drying cycles were used as chloride induced corrosion.
The MK, SF and OPC concretes subjected to accelerated corrosion were examined. 


In the simultaneous exposition to CO2 and Cl , SF-concrete greatly reduces corrosion.
Under carbonation MK and SF concretes shown higher corrosion rates than OPC.
Use of metakaolin and silica fume enhances chloride resistance of concrete.

a r t i c l e

i n f o

Article history: Received 2 February 2016 Received in revised form 23 May 2016 Accepted 14 June 2016

Keywords: Blended concrete Carbonation Chlorides Corrosion Metakaolin Silica fume Electrochemical impedance Polarisation resistance

a b s t r a c t The research study presented herein evaluates the corrosion behaviour of the reinforcing steel in blended concretes using two pozzolanic additives—metakaolin (MK) and silica fume (SF)—at 10% replacement of cement weight. They are exposed to CO2 and chlorides. The corrosion process was followed by monitoring of open-circuit potential (OCP), polarisation resistance (Rp) and electrochemical impedance spectroscopy (EIS). Electrochemical measurements show that the addition of MK and SF enhances corrosion resistance exposed to chlorides, however under accelerated carbonation these concretes show higher corrosion rates. In the simultaneous exposition to carbon dioxide and chlorides, SF-concrete shows a decrease of corrosion rate. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Concrete has become the most utilised material in the construction sector worldwide because of its versatility and low cost [1,2]. Portland cement concrete is a ceramic material that supports compressive stresses. However, it is susceptible to fracture under other types of mechanical loads, such as flexion, traction, torsion, and shear. As a result, reinforced concrete, which is a composite material that is composed of concrete and structural steel, has been developed. This material has been extensively utilised in the construction of bridges, buildings, and tunnels. One of the most

⇑ Corresponding author. E-mail addresses: [email protected] (A.M. Aguirre-Guerrero), [email protected] (R. Mejía-de-Gutiérrez), [email protected] (M.J.R. Montês-Correia). http://dx.doi.org/10.1016/j.conbuildmat.2016.06.038 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.

important characteristics of concrete, in addition to its mechanical properties, is its durability, which can be associated with the service life of a structure that has been exposed to aggressive environmental conditions [3,4]. The primary problem related to the durability of reinforced concrete is corrosion of the reinforcing steel, which causes the loss of mechanical and structural properties as corrosion progresses [5]. Corrosion of the reinforcing steel is primarily caused by exposure to aggressive environments that contain chloride ions and/or carbonation [4]. The presence of chlorides in the interior of concrete can originate from two main sources: the concrete mixture (contaminated aggregates, seawater or polluted water, and additives with a high chloride content) and the external environment. Once the chloride ions penetrate concrete, they spread as bound chlorides and free chlorides. The former correspond to chlorides that are chemically bonded via reaction with tricalcium aluminate (C3A) in cement, which

705

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

subsequently form calcium chloroaluminate—a compound known as ‘‘Friedel’s salt” that is not expansive. For this reason, it is advisable to utilise cements with high contents of C3A for marine environments. Conversely, free chlorides diffuse through the cementitious matrix until they reach steel; this diffusion causes localised dissolution of the passive layer and attacks at specific points, which can significantly reduce the structural properties of steel [4]. The process of carbonation is due to the entry of CO2 from the atmosphere to the interior of the concrete; urban environments and environmental pollution are necessary sources of this phenomenon. Concrete is an alkaline material with a pH between approximately 12.6 and 13.6, which provides natural protection against corrosion of reinforcing steel. However, carbonation decreases the pH of concrete to approximately nine; consequently, the passive layer is destabilised, which causes corrosion of reinforcing steel [4–7]. To achieve carbonation, carbon dioxide must contact the water and alkaline components in the concrete pores. The rate of CO2 penetration is dependent on environmental factors, such as relative humidity between 50% and 70%, the temperature, and CO2 concentration. The last factor can attain a maximum value of approximately 0.1% in urban environments [8–10]. Other factors that contribute to the propagation of CO2 are inadequate curing, poor compaction, and high water-to-cement (W/C) ratios, which generate highly permeable concrete. The incorporation of pozzolan can prolong the service life of a structure and contribute to the mechanical properties of a structure by decreasing the permeability of a material, which reduces the entry of aggressive agents from the environment. Despite the benefits of incorporating pozzolan in concrete, studies have indicated controversial findings [11–14]. The commercial pozzolans utilised in the production of high-performance concrete are metakaolin (MK) and silica fume (SF). In recent years new pozzolans have been studied as a cement replacement such as fluid cracking catalyst spent (SFCC), which shows an evidence of a pozzolanic activity and good corrosion performance [15–17]. MK is a material that is obtained from the calcination of kaolinitic clay at temperatures between 500 and 800 °C [1,4,18]. MK is an aluminosilicate-type pozzolan, whose particle size is slightly finer than the particle size of cement [1]; its primary characteristic is its high reactivity with the calcium hydroxide in cement and its ability to accelerate the hydration of cement [18]. Based on its characteristics, this additive increases the mechanical properties (compressive and flexural strengths) and the durability (lower permeability and greater resistance to chemical attack [4,18–20]) of concrete. The corrosion resistance of reinforcing steel increases with the addition of MK on the order of approximately 10–15% [4] by reducing the diffusion of chloride ions in the cementations matrix, which prolongs the service lives of structures. Güneyisi et al. [21] note that the rate of corrosion decreases by 50% for concretes that were blended with MK and exposed to chlorides. Kelesßtemur and Demirel [18] conclude that the replacement of cement with 15% MK increases the compressive and tensile strengths and improves the corrosion resistance of reinforcing steel. Conversely, controversial studies have investigated the performance of concrete in the presence of CO2. Mejía de Gutiérrez et al. [22] determined that the depth of carbonation for specimens blended with MK after 28 days of curing is slightly greater than the depth of carbonation for specimens of unblended concrete. However, the carbonation rate decreases after increasing the curing time. Kim et al. [23] assert that the carbonation depth increases in blended concretes. SF is a by-product of the silicon and ferrosilicon industry. The reduction of high-purity quartz at temperatures greater than 2000 °C produces silica vapours, which oxidise and condense to form small particles of amorphous silica [1,24]. SF has a high

content of silica of approximately 85–95%, a fine particle size of approximately 0.1 to 0.5 lm, and a large specific surface area [1,4,24]. Because of its high reactivity, SF accelerates the hydration processes in cement; because of its small particle size, it refines the pores in the cementitious matrix and considerably decreases the porosity of the material [23]. The advantages of concrete that has been blended with SF are an increase in the mechanical properties of concrete and an improvement in the durability properties of concrete. To attain high compressive strengths and low chloridediffusion coefficients, 5–10% loading in place of cement is suggested [26,27]. Chao and Lin [28] have observed that SF-blended concrete reduces the permeability due to a decrease in capillary pores, increase the resistance to chloride and reduces the probability of corrosion of a reinforcing concrete. Regarding carbonation, Kulakowski et al. [29] assert that concrete with W/C ratios of 0.45–0.50 are more resistant to carbonation. However, some authors have noted controversial results [25]. The objective of this investigation was to compare the performance of concrete that is blended with 10% MK and SF with respect to the mass of cement with corrosion of the reinforcing steel. The evolution of the corrosion process was evaluated via nondestructive techniques, such as open-circuit potential (OCP), electrochemical impedance spectroscopy (EIS), and linear polarisation resistance (LPR), on specimens that were exposed to different aggressive environments, such as carbonation, chlorides, and a mixed environment (carbonation and chlorides). In all cases, a reference concrete that contains 100% Portland cement was employed. 2. Experimental procedure 2.1. Materials and specimen preparation Raw materials were selected for the production of the materials that are available on the domestic market, such as general-use Portland cement, reinforcing steel, and aggregates. Additives for reducing permeability, such as the commercial MK MetamaxÒ, supplied by BASF and the commercial SK from SikaÒ, were employed. Table 1 shows the chemical compositions of the utilised materials. A W/C ratio of 0.55 and a 10% replacement of cement with each of the additives were utilised in the production of the concrete, considering the results of previous studies [4,30,31]. The concrete was fabricated in a mixer with a capacity of 210 L. Table 2 shows the composition of the concrete mixtures. For the concrete mixture with SF, the superplasticiser additive SikaÒ ViscocreteÒ 20HE was employed. After mixing, the fresh concrete was poured into metal moulds with dimensions of 76.2  152.4 mm and compacted with a smooth metal rod. The specimens were unmoulded after 24 h and left to cure in water for 28 days. Each specimen had a steel rod with a length of 150 mm and a diameter of 6.4 mm embedded in its centre. These steel rods were cleaned with acetone to eliminate fats on the surface. In addition, a 60 mm-long exposure zone was defined, which corresponds to the area that was embedded within the concrete. The remainder of the rod was coated in corrosion-resistant epoxy paint.

2.2. Electrochemical measurements To monitor the evolution of corrosion in reinforced concrete, different non-destructive techniques have been utilised, such as OCP, LPR, and EIS. Table 1 Chemical composition of Portland cement, metakaolin and silica fume.

a

Chemical constituent (%)

Cement

Metakaolin

Silica fume

SiO2 Fe2O3 Al2O3 CaO MgO SO3 TiO2 K2O Na2O Cl LOIa

19.13 4.32 4.42 57.70 1.6 2.32 – – – – 9.78

52.36 0.37 44.25 – – – 1.76 – – – 0.86

85.3 0.5 0.3 0.5 3.6 0.5 – 2.4 2.1 0.8 3.78

LOI: Loss on ignition at 1000 °C.

706

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716 Steel

Table 2 Concrete mix composition. Material (kg/m3)

Cement Metakaolin Silica fume Water Coarse aggregate Sand Superplasticiser

Concrete OPC

MK (10%)

SF (10%)

389 – – 214 900 900 –

350 38.9 – 214 900 900 –

350 – 38.9 214 900 900 0.19

Power Supply DC

+

Concrete

-

Stainless Steel NaCl 3.5%

In construction, OCP is one of the most utilised procedures for the routine inspection of reinforced structures; it is governed by the American Society for Testing and Materials (ASTM) C876 standard [3,32]. In this study, OCP measurements were performed with a multimeter (UNIT-T MODELO UT70A) using an Ag/AgCl reference electrode. For the LPR and EIS measurements, an electrochemical cell with three electrodes was utilised: a working electrode (WE, steel embedded in concrete), a reference electrode positioned near the WE, and a stainless-steel counter-electrode located around the concrete. The Autolab Instrument PGSTAT128N Potenciostat/Galvanostat was employed with an FRA2 module. The LRP test was performed according to the ASTM G59 standard by applying overpotentials of 30 to +30 mV. The EIS measurements were performed in a frequency range of 105– 102 Hz, used for different researchers [33,34]. The sinusoidal voltage perturbation was 10 mV. The EIS data were analysed with Zview 2.1 software based on equivalent circuit models. These measurements were conducted with respect to time.

Fig. 1. Accelerated technique with impressed voltage.

2.2.1. Exposure to aggressive environments The unblended specimens—ordinary Portland cement (OPC)—and the blended specimens (MK and HS) were subjected to four different exposure environments: water (Ref), carbonation (CO2), chlorides (Cl), and a mixed environment (CO2/Cl).

2.2.2. Reference: water This environment is the reference. During the experiment, the samples were continually immersed in potable water. At the test times, the samples were removed to perform the electrochemical measurements; they were placed in the same environment after the tests.

2.2.3. Accelerated carbonation After cured for 28 days, the specimens were exposed to a carbonation chamber in controlled conditions: a CO2 concentration of 1%, a relative humidity of 65%, and an ambient temperature of 25 °C. The measurements were performed with respect to time. For the electrochemical measurements, the specimens were removed from the climatic chamber, in which a solution of potable water was utilised. After the tests, the specimens were placed in the climatic chamber.

Fig. 2. Intensity curve of current vs time.

3. Results and discussion 2.2.4. Chlorides. Accelerated, impressed-voltage technique – wetting/drying cycles At 28 days of curing, the specimens were partially immersed in a 3.5% NaCl solution. According to the NT BUILD 356 standard, a constant voltage of 5 V was applied with an external source between the anode (embedded steel) and the cathode (stainless steel plate) (Fig. 1). The time required for the chlorides to enter the material was determined from the current-time intensity curve throughout the immersion process, as shown in Fig. 2, which indicates the time at which the first crack appears for each type of concrete. In the accelerated chloride exposure test, 20% of this duration (time of initial crack appearance) was selected, and voltage was applied (IV). Subsequently, the process of wet/dry (W/D) cycling was continued. This process consisted of exposing the specimens to ambient temperatures for 15 days and immersing them for 15 days in a 3.5% aqueous NaCl solution, until 12 cycles were completed. To perform the set-up, monitoring, and data acquisition, an Agilent U2741A digital modular multimeter with a U2781A chassis was employed with the following parameters: DC voltage range – 1 lV to 300 V; AC – 1 lV rms to 250 V rms; DC current range – 1 lA to 2 A; AC – 1 lA rms to 2 A rms; and voltage source – Agilent E3645A; range from 0 to 35 V/2.2 A.

2.2.5. Mixed environment: carbonation and chlorides The specimens were exposed according to the procedure for accelerated carbonation over four months, which we refer to as CARB. Subsequently, they were exposed to wetting/drying cycles, in which they were exposed to ambient temperatures for 15 days and immersion for 15 days in an aqueous solution of 3.5% NaCl until 10 cycles were completed.

3.1. OCP measurements OCP is a technique that can be applied in the field; however, it only provides qualitative data, given that the speed or intensity of corrosion cannot be determined based on the results. Fig. 4a–d reveals the OCP evolution for MK- and SF-blended concretes and unblended OPC concrete that were exposed to different environmental conditions. As shown in Fig. 4a, which corresponds to the samples that were immersed in potable water, the three types of concrete are in an initial passive state; during the exposure time—between 100 and 360 days—the specimens are in an uncertain region of probability of corrosion. As shown in Fig. 4b, the specimens that were exposed to carbonation are observed in the region in which the corrosion probability is less than 10% until approximately 250 days of exposure, after which the corrosion potentials are located in an uncertain zone. The carbonation process produces calcium carbonate and water. The calcium carbonate has a low solubility and precipitates with the pores of the concrete, which reduces the porosity. This precipitation increases the strength of the concrete. When CO2 continues to diffuse through

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

707

Fig. 3. Concrete specimens after the carbonation susceptibility test (a), (b) and (c) OPC, MK and SF concrete, respectively, exposed for 45 days; (d), (e) and (f) OPC, MK and SF concrete, respectively, exposed for 165 days (g), (h) and (i) OPC, MK and SF concrete, respectively, exposed for 240 days.

the cementitious matrix, the pH continues to decrease, and after reaching the steel, the passive layer is destabilised, which initiates the propagation stage according to the model of Tutti [35]. To observe the processes, a carbonation susceptibility test, in which samples without steel were exposed to the same conditions and the carbonation depth at different times was evaluated, was performed (Fig. 3). Based on the measurement of carbonation depth, the carbonation coefficient Kc was obtained for accelerated conditions. The values of Kc for OPC, MK, and SF were 31.3 mm/year1/2, 35.6 mm/year1/2, and 37.7 mm/year1/2, respectively. Based on this value, the times at which the carbonation reaches the reinforcing steel were calculated, and values for OPC, MK, and SF of 441 days, 342 days, and 340 days, respectively, were obtained. These data coincide with the observations in Fig. 4b: at approximately 350 days, the corrosion process has begun. After applying the accelerated impressed-voltage (IV) technique, as shown in Fig. 4c, the three specimens are observed in the range in which the corrosion probability is 90%, which indicates that corrosion is active in this stage. From cycle 1 to 6, however, no major changes are observed among the specimens. Beginning from cycle 7, the MK and SF blends have greater positive potential values compared with OPC. Some fluctuations in the SF concrete are observed, which can be attributed to possible problems with the test. In Fig. 4d, the concrete that was exposed to CARB are in a passive state at the beginning of exposure. Immediately after the first cycle (chlorides), the values decrease to the uncertain region for the blended concretes. Beginning with cycle 2, all specimens are observed in the region in which the corrosion probability is greater than 90%, which indicates the acceleration of corrosion after exposure to chlorides.

3.2. Linear polarisation resistance and EIS LPR is a technique that has been applied since the 1970s; it provides information about the corrosion rate provides quantitative data [3,32,36]. The calculation of the corrosion current was performed by applying the Stern-Geary formula

Icorr ¼

B Rp

ð1Þ

where B is a constant that is dependent on the Tafel slopes, whose estimated value is 26 mV or 52 mV depending on the active or passive state of the steel. In this study, the utilised value ‘‘B” was 0.026 V, which simulates an active corrosion condition [37,38]. The Rp technique is not sufficient for discriminating the resistive contributions of the electrolyte solution [39]. For this reason, the Rp technique was performed in this study in combination with the EIS measurements to discriminate the resistance of the solution and obtain more accurate results for the corrosive evolution of the steel that is embedded in the concrete. EIS is a technique that works on the frequency domain, which provides abundant information about the system, such as the electrical resistivity, the dielectric properties of the concrete, the rate of corrosion, the properties of the steel/concrete interface, and kinetic information [32,36,40–42]. In this section, some elements of the equivalent circuit were determined, which correspond to the high-frequency domain in the EIS measurements. In this study, the results of the solution resistance (Rs) of the equivalent circuit are utilised and compensated with the polarisation resistance obtained with the

708

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

Fig. 4. Evolution of the OCP. (a) Ecorr for the corrosion evolution in the reference environment, i.e., potable water, (b) Ecorr for the corrosion evolution in the carbonation environment, i.e., CO2, (c) Ecorr for the corrosion evolution in the chloride environment, (d) Ecorr for the corrosion evolution in the mixed environment, i.e., CO2/Cl.

Fig. 5. Equivalent circuit schematic for the EIS analysis.

Rp technique. Fig. 5 represents the equivalent circuit schematic in this study. When Rs corresponds to the high-frequency domain, it is related to the bulk resistivity of the concrete (resistance of the electrolyte); Ri corresponds to the intermediate frequency domain (1kHz to 40 Hz) and represents the resistance of the passive layer or the layer of oxides that are formed by deterioration of the steel due to corrosion; and CPEi is a constant-phase element that represents the passive film or steel/concrete interphase. CPEdl represents the non-ideal behaviour of the double-layer capacitance of the charge transfer at low frequencies. Rct represents the charge transfer resistance or Rp [43]. The compensated resistance is calculated by subtracting the Rs in the EIS analysis from the polarisation resistance obtained by utilising the LRP, which yields an approximate estimation of the behaviour of the corrosion process. Fig. 6 shows the evolution of the corrosion rate icorr, which was calculated with the compensated resistance and the Stern-Geary

formula. In Fig. 6a, the behaviour over time is shown for the concrete specimens that were exposed to the reference environment. Throughout the time evolution, the OPC and SF specimens remained in the low corrosion region. Note that the SF specimen remained in the range of negligible corrosion. The specimens are located in the range of low corrosion in this environment as potable water contains a small amount of chlorides, which can have a small effect on the corrosion evolution with respect to time. The MK-10% specimens have a slightly better performance than the other specimens. Fig. 6b corresponds to the evolution of the concrete that was exposed to carbonation. After 365 days of exposure, the corrosion rate increases in the high corrosion range, which coincides with the time obtained by the carbonation susceptibility test. The blended concrete exhibits behaviour that is similar to the OPC. The blended concretes are more susceptible to carbonation, as indicated in the literature. Notwithstanding the composition, the characteristics of the concrete and their reactions determine the carbonation depth. San Nicolas et al. [44] assert that open porosity is necessary to promote carbonation in concrete, which permits diffusion of CO2 and the chemical reaction between calcium hydroxide and carbon dioxide. A concrete that possesses greater contents of portlandite is less susceptible to carbonation, whereas greater carbonation depths are obtained for concretes that are blended with MK, in which portlandite is partially consumed in the pozzolanic reactions. Mejía de Gutiérrez et al. [22] note that MK-blended concretes at curing ages greater than 90 days, when the pozzolanic reaction has been completed, display a better

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

709

Fig. 6. Evolution of the corrosion rate icorr. (a) icorr for the corrosion evolution in the reference environment, i.e., potable water, (b) icorr for the corrosion evolution in the carbonation environment, i.e., CO2, (c) icorr for the corrosion evolution in the chloride environment, (d) icorr for the corrosion evolution in the mixed environment, i.e., CO2/Cl.

performance in carbonation. Kulakowski et al. [29] conclude that SF-blended concretes with W/C values greater than 0.5 are susceptible to carbonation. In Fig. 6c, which corresponds to the corrosion evolution in concrete that was exposed to chlorides, the icorr values are shown after application of the IV technique (VI). In these cases, a significant increase in the current with respect to the initial state is observed. Beginning with cycle 7, the benefits of pozzolanic additives are distinct. However, all specimens are observed at the high-corrosion level. At the end of cycle 12, icorr for MK and SF are approximately 85% less and 32% less, respectively, than the icorr reported for OPC. The pozzolanic additives MK and SF exhibit better performance in corrosion by chlorides, which is consistent with the results from other studies [18,20,22,25,28,45]. MKblended concretes display better performance in the corrosion of the reinforcing steel that is induced by chlorides. These results are attributed to the formation of Friedel’s salt, which restricts chloride diffusion in the interior of the concrete [46]. Fig. 6d corresponds to the corrosion evolution process in the mixed environment of CO2/Cl. This figure shows that the corrosion intensity values increase after CO2 exposure for four months (CARB) and subsequent exposure to wetting/drying cycles in the presence of chlorides. In general, an increase in the corrosion current is observed in all cases. At the end of cycle 10, MK shows an increase in the corrosion current that is approximately 5 times the corrosion current of OPC. This finding is attributed to the greater susceptibility to carbonation of MK-blended specimens (Fig. 3) and, therefore, the solubility of some carbonated species, which increase the porosity of the concrete [47], constituting a pathway for the entry of free chloride

ions towards the steel rod and initiate corrosion. As previously mentioned, the MK-blended concretes promote the formation of Friedel’s salt. However, the stability of this compound is dependent on the alkalinity of the concrete. As a result, the solubility of Friedel’s salt increases with the degree of carbonation and [48–50] a decrease in the pH of the concrete. These results demonstrate the competition between the effect of carbonation, the pozzolanic reaction, and the diffusion of chloride ions within the concrete. For concretes blended with fly ash that were exposed to carbonation and chlorides, Kuosa et al. [51] and Montemor et al. [47] observed that the corrosion rate notably increases with respect to the reference concrete, which is consistent with the concretes tested in this study. In this study, a better performance in the mixed environment is observed for concretes that are blended with SF. This finding can be attributed to a greater refinement of pores and to potential blockage of pores at this stage of carbonation by the precipitation of calcium carbonate in concrete pores [47]. An EIS analysis was performed at the final test age for the different environments. The Nyquist impedance spectra are shown in Fig. 7a–d. Fig. 7a shows the MK- and SF-blended concrete and unblended OPC that were exposed to the reference environment. In the majority of the spectra, with the exception of carbonation, these results are verified at high frequencies. The impedance for OPC is less than the impedance for the remaining concrete, and the impedance for MK is greater than the impedance for the remaining concrete. Pozzolans addition increases the value of the impedance at high frequencies for the specimens that were exposed to this environment. Fig. 7b shows the specimens that

710

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

Fig. 7. Nyquist spectra (a) after 450 days of exposure to potable water; (b) after 650 days of exposure to carbonation, CO2; (c) after 12 cycles of W/D exposure to chlorides; (d) after ten cycles of W/D exposure to the mixed environment of CO2/Cl.

were exposed to CO2 at the end of the exposure when active corrosion is observed. The difference between the conditions is less significant. In particular, SF exhibits a decrease in the value of the impedance at high frequencies with respect to the remaining specimens. Fig. 7c shows the specimens that were exposed to chlorides. The lowest value of impedance is obtained for OPC. Low values of impedance suggest active corrosion. The impedance values for blended concretes are greater than the impedance values for the remaining specimens. Triana et al. [52] assert that MK-blended concretes increase the impedance values over the entire frequency range, which is consistent with the results in this study. Fig. 7d shows the specimens that were exposed to the mixed environment. The blended specimens show an increase in the impedance values over the entire frequency range, as observed for the specimens that were exposed to chlorides. In this environment, the lowest impedance values are observed. Fig. 7b–d show the impedance spectra and values, which suggest active corrosion for all associated environments, especially at low frequencies. The reference environment shows the highest values of impedance with respect to other exposure environments. These results are confirmed in Fig. 8a–d, in which the Bode diagrams are shown for the final time of each exposure. The impedance modulus is less for OPC and greater for MK over the entire frequency range, with the exception of the specimens that were exposed to CO2 (Fig. 8b). The spectra suggest active corrosion in the steel in all environments and a greater resistance to corrosion, which is not observed in the case of the specimens that are exposed to the reference environment. A comparison of Fig. 8a–d shows that the reference environment is the least aggressive. The greatest impedance modulus is

observed for specimens in the reference environment; it decreases after exposure to carbonation and exposure to chlorides. The specimens in the mixed environment exhibit a lower impedance, which reinforces the observations for the Nyquist spectra. In the intermediate frequency range, a different behaviour is observed for MK, with the exception of the mixed CO2/Cl environment (Fig. 8d). This result may be attributed to the characteristics of the steel-concrete interface. Table 3 shows the fit with the equivalent circuit shown in Fig. 5 with a maximum percentage error of 20%. In the high-frequency range, the Rs corresponds to the electrolyte resistance in the concrete pores. Ri and Qi are attributed to the resistance of the passive layer and the constant phase element (CPE), respectively. Rct and Qdl correspond to the charge transfer resistance and the doublelayer characteristics, which are associated with the lowfrequency arc in the Nyquist plot [43]. Non-ideal behaviour is represented by the CPE, which has been extensively utilised for the study of the corrosion of reinforcing steel that is embedded in concrete [33,40,53–56]. The CPE impedance is given by

Z CPE ¼ 1=Q ðjxÞ

n

ð2Þ

where x is the angular frequency and Q and n are constants. The exponent n provides information about the degree of non-ideality. An ideal capacitor corresponds to n = 1, whereas n = 0.5 indicates that the CPE behaves as a Warburg diffusion component. If n = 0, its behaviour is indicative of a resistor. Intermediate values of n (0 < n < 1) are related to the non-homogeneity of the surface [40,55]. The results for each of the exposure environments at the initial and final times are shown. The initial state corresponds to time 0,

711

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

Fig. 8. Bode diagram (a) after 450 days of exposure to potable water, (b) after 650 days of exposure to carbonation CO2, (c) after 12 cycles of W/D exposure to chlorides, (d) after ten cycles of W/D exposure to the mixed environment of CO2/Cl.

Table 3 Best fitting results for the impedance spectra. Sample

Exposure environment

Exposure time

Rs (kX.cm2)

Qi (lF.cm2.sn)

ni

Ri (kX.cm2)

Qdl (lF.cm2.sn)

ndl

Rct (kX.cm2)

OPC

Reference

Initial Final Initial Final Initial Final Initial Final

24.6 6.9 17.0 8.1 28.6 1.3 26.7 0.9

16.7 87 15.3 78 11.1 484.2 16.7 1271.1

0.74 0.77 0.80 0.41 0.70 0.57 0.74 0.47

7.9 6.9 4.6 4.0 8.6 0.6 7.9 0.2

27.7 52 35.0 737.0 28.6 8110.0 27.7 7169.1

0.74 0.79 0.78 0.88 0.73 0.54 0.74 0.73

373.7 271 405.4 7.0 306.9 0.76 373.7 0.2

Initial Final Initial Final Initial Final Initial Final

24.0 29.3 21.3 9.2 30.5 5.4 25.9 2.2

17.3 4.0 16.1 87 16.1 296.5 17.3 459.3

0.73 0.95 0.79 0.55 0.72 0.38 0.73 0.40

5.9 3.2 4.7 1.8 9.6 1.9 5.9 0.6

35.2 79.0 32.8 367.0 35.8 872.9 35.2 10,529

0.72 0.55 0.78 0.78 0.73 0.60 0.72 0.68

388.3 223.0 503.7 5.2 556.1 3.41 388.3 0.97

Initial Final Initial Final Initial Final Initial Final

19.0 12.1 13.9 5.1 25.3 2.6 14.7 1.2

37.6 33 50.7 104.0 27.6 519.5 37.6 764.0

0.86 0.73 0.83 0.48 0.88 0.46 0.86 0.42

3.4 3.6 3.9 0.4 4.2 0.7 3.4 0.8

59.3 69.0 55.6 618.0 64.9 7334.7 59.3 1977.6

0.84 0.79 0.85 0.77 0.81 0.79 0.84 0.67

292.9 117 289.5 6.3 166.6 2.9 292.9 6.6

CO2 Cl CO2/Cl MK (10%)

Reference CO2 Cl CO2/Cl

SF (10%)

Reference CO2 Cl CO2/Cl

712

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

when the specimens have not been exposed to an aggressive environment. The final state corresponds to the final time of exposure: 450 days for the reference environment, 650 days for carbonation, after 12 cycles for chlorides, and after ten cycles for the mixed environment. 3.2.1. Carbonation Rs decreases for the OPC specimen by 52%, by 57% for MK, and by 63% for SF with respect to the initial value. This finding suggests an advanced stage of carbonation, in which the electrolyte resistance decreases due to the total diffusion of CO2 in the cementitious matrix and triggers the corrosion of steel, in which the steel-concrete interface experiences dramatic changes [57]. The comparison of the final MK values indicates an increase of 12% with respect to OPC, whereas SF exhibits a reduction of 37% with respect to OPC. The Qi values increase after CO2 exposure for all specimens with respect to the initial value. OPC shows the lowest value, which suggest the diffusion is more controlled than the blended concretes. The values of npf decrease for all specimens, which may be associated with a possible alteration in the steel/ concrete interface due to corrosion or to possible diffusion processes in the oxide layer, given that the values indicate the behaviour of a Warburg element. Ri decreases for all the specimens with respect to the initial value. For blended specimens, however, the effect is more pronounced, which suggests a greater degradation of the resistive characteristics of the interface in these specimens. The Rct values decrease for all specimens with respect to the initial value. Rct is inversely proportional to the corrosion current (icorr), which suggests an increase in the corrosion rate. The value of Rct for OPC is greater than the values of Rct for MK and SF by 25% and 10%, respectively. The blended specimens are slightly susceptible to carbonation with respect to OPC. The Qdl values for all types of concrete after CO2 exposure increase, which indicates the facilitation of charge transport in all cases. The highest Rs value corresponds to MK, and the lowest Rct value in the final stage suggests a greater accumulation of corrosion and reaction products, which block conductive pathways at the steel/concrete interface, as shown by the lowest value of Qdl relative to the transport capacity. In accordance with the literature, these products can increase the electrical resistance of the solution [15]. 3.2.2. Chlorides Rs is reduced for all specimens with respect to the initial value. Comparing the values after exposure to Cl, concrete with MK presents an increase of 86% with respect to the OPC, and SF presents an increase of 50% with respect to the OPC. Blended concretes have a greater Rs after exposure to chlorides, and the strength of MK increases because of the formation of Friedel’s salt, which precipitates in the pores of the concrete [58] and increases its resistance with respect to OPC. The Qi values increase after exposure to chlorides for all specimens with respect to the initial value. MK has the lowest value, which suggests the diffusion phenomena are more controlled compared with OPC. The values of n decrease for all specimens, whereas the values of n for OPC and SF exhibit the behaviour of a Warburg element. For MK, n = 0.38, which suggests a resistive behaviour at the steel/concrete interface. Ri decreases for all specimens compared with the initial state. MK yields the largest resistance value with respect to OPC, which may be associated with a resistive interface for these specimens. Rct values that are typical of active corrosion decrease in all specimens with respect to the initial value. MK yields the highest Rct value, with an increase of 77% over OPC. SF also exhibits improved performance with an increase in Rct of 74% compared with OPC. These findings indicate that the addition of MK and SF have a beneficial effect

during chloride exposure. The same trend is observed with the LRP technique. At the end of cycle 12, all specimens are in a state of active corrosion. However, action pozzolanic additives are observed. Sobhani and Najimi [56] presented concrete with added silica fume and conclude that blended concretes increase Rs and Rct compared with concrete without additives, which is consistent with this study. According to the results, the additives increase all resistive elements that are considered in the analysis. MK has the highest values for the resistive elements and the lowest values for CPEs relative to the transport properties at the steel/concrete interface and the double layer, which justifies the enhanced performance in the chloride environment. The OPC specimen shows the highest value of Qdl; this finding suggests a greater transfer capacity, which facilitates corrosion propagation, where indicates the passive film have been destroyed. The value of the exponent n for OPC is 0.54, which is close to a possible diffusion control. The Rct value for OPC is one of the lowest about 0.76 kX.cm2 which suggests the specimens were under an aggressive contamination of chlorides, due to use of combination of two accelerated techniques. Wei et al. [34] presented corrosion evolution in OPC concrete under dry/wet cyclic conditions. At the end of cycle 14 observed an Rct value about 0.11 kX.cm2 and Qdl value 117.8  104 X1.cm2.sn, which are lower Rct values and Qdl higher values results compared with this study. Pech-Canul and Castro [54] reported an Rct value about 0.87 kX.cm2 for an OPC concrete with a 0.7 w/c ratio exposed to a natural marine atmosphere. Sobhani and Najimi [56] reported 0.592 kX.cm2 Rct value for an OPC concrete with a 0.5 w/c ratio exposed to dry/ wet cycles at the ages of 90 days. 3.2.3. Mixed environment Rs is reduced for all specimens with respect to the initial value. The comparison of the values after exposure to CO2/Cl values indicates that the concrete with MK has the highest value of Rs. Qi values increase after exposure for all specimens with respect to the initial value. The highest value, which was observed in OPC, indicates a greater transport capacity at the steel/concrete interface, which facilitates the process of corrosion. The lowest value was obtained for MK. The values of n decrease for all specimens. Ri decreases for all specimens with respect to the initial state. SF has the largest resistance value with respect to OPC, which may be associated with a resistive interface for the type of specimen that is exposed to this environment. Rct values decrease for all specimens with respect to the initial value, which suggests an increase in the corrosion rate. SF has a higher value of Rct and an increase of 97% compared with OPC. MK also performs better with an increase in an Rct of 75% with respect to OPC. The Qdl values for all specimens after exposure to the mixed atmosphere markedly increase, which indicates that corrosion is likely in each case compared with other environments. The exposure condition made in this research was very aggressive due to use of combination of two accelerated techniques. The highest value for Qdl is obtained for the MK specimen, which suggests a greater charge transfer and contribution to corrosion propagation. In this environment, MK has high values of Rs, a lower value of Qi, and a higher value of Qdl. This finding suggests competition among the carbonation effect, the pozzolanic reaction, and the diffusion of chloride ions in concrete. The EIS results show that the corrosion rate after exposure is significant for OPC specimens. The Rct value for OPC is the lowest about 0.2 kX.cm2, which indicates the specimens were under an aggressive environment, which suggest a severe corrosion. Wei et al. [4] reported Rct values below the results presented in this study; the specimens were just exposed to chlorides. On the other hand, Montemor et al. [47] presented a different type of exposure for a combination of carbonation and chlorides. The samples were exposed in a chamber 5% CO2, room temperature and

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

60–70% relative humidity and sprayed weekly with 15% NaCl aqueous solution; they reported an Rct value of 10 kX.cm2 at the age of 150 days approximately, which are higher than the results presented nevertheless they have a different age and exposure condition. The better performance of SF is justified by the greater values of the resistive components and the lowest value of the CPE relative to the charge transfer process in the double layer. 3.2.4. Reference environment The Rs in the reference environment was reduced by 72% for the OPC specimen, by 36% for SF, and an increase of 18% with respect to the initial for MK. The comparison of the final values of the specimens indicates that MK and SF showed an increase of 76% and 43%, respectively, compared with OPC, which suggests a greater resistance of blended concrete that was exposed to water. This resistance is dependent on the pore structure of the concrete and the solution [15]. The Qdl values for all specimens increase with respect to the initial value. The Rct values for OPC are higher than the Rct values for the blends but are not as low as observed with other exposure environments. The charge transfer capacity of the double layer in the blended concretes is significant, which should contribute to the development of corrosion. The difference between MK and SF may be attributed to the resistive characteristics of the concretes and the lower transport capacity at the steel/concrete interface. 3.2.5. Comparison of the exposure environments With few exceptions, the concrete strength decreases from the initial state to the final state. Minor differences are recorded for the reference environment, which denotes a better condition. Carbonated specimens also have relatively high values of Rs, which denote the consequences of the process of carbonation. Lower values of Rs in the chloride and mixed environments are expected because of contact with the solution of sodium chloride

713

and the state of degradation of the specimens caused by the exposure conditions. The resistance of the steel/concrete interface decreases for the final state, which shows higher values for the specimens in the reference environment. As in the case of Rs, specimens in the carbonation environment have the second-highest relative values after the specimens in the reference conditions. The only exception is the specimen with the addition of SF. The difference in the recorded range of values denotes a change in the steel/concrete interface for a non-ideal capacitive process due to diffusion control, with the exception of the specimens in the reference environment, which have the highest values, and the specimens with values similar to the initial values. Qi is greater in the final state, and the difference is most pronounced for specimens in the chloride and mixed environments. The addition of MK in the reference environment has a pronounced effect on the characteristics of the steel/concrete interface. In the double layer, systemic changes are recorded in the value of ndl; however, Qdl increases in the final state, especially in mixed and chloride environments. Specimens with additives, MK in the chloride environment and SF in the mixed environment have the lowest values. Carbonated specimens show intermediate values, and specimens in the reference condition have the lowest values. The charge transfer capacity and influence of the corrosion rate can show that the corrosion process is less favoured in the reference environment and carbonated specimens. For the remaining environments, improved behaviour is observed for specimens with MK in the chloride environment and specimens with SF in the mixed environment. Fig. 9a–c show icorr for the initial and final values of each exposure environment. Icorr was calculated from Eq. (2) using Rct. For the OPC and MK concrete, icorr is significantly higher for exposure to the mixed environment, which is the most aggressive environment. For SF concretes, the most aggressive environment is chloride exposure; its superior performance exceeds the performance

Fig. 9. Comparison of environments with respect to icorr calculated using EIS (a) OPC specimens exposed to different environments, (b) MK specimens exposed to different environments, (c) SF specimens exposed to different environments.

714

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

Fig. 10. Images of steel rods exposed to different environments (a), (b) and (c) steel embedded in OPC, MK and SF concrete, respectively, after exposure to CO2; (d), (e) and (f) steel embedded in OPC, MK and SF concrete, respectively, after exposure to chlorides; (g), (h) and (i) steel embedded in OPC, MK and SF concrete after exposure to the mixed environment.

of other concretes in the mixed environment. OPC exhibits a slightly better performance than the blended concretes in carbonation. The less aggressive environment for all specimens is the reference. For the specimens exposed to CO2, the Icorr results reveal that MK-blended concrete and SF-blended concrete show an increase in the corrosion rate of 24% and 8.5%, respectively, compared with OPC. In chlorides, concrete that contains MK has a decrease of 78% in the corrosion rate with respect to OPC. For the combined effect of exposure, the best performance of SF-blended concrete shows a decrease of 97% in the corrosion rate with respect to OPC. In Fig. 10, a visual inspection of the reinforcements after exposure to the harshest environments is presented. Signs of corrosion in the reinforcement is confirmed for all environments. For the mixed environment, a substantial deterioration of the steel with respect to OPC and MK specimens is observed. Steel that was exposed to carbonation shows uniform corrosion, whereas pitting is observed in the steel that is exposed to chlorides and the mixed environment. This finding confirms the results from different techniques for monitoring the corrosion of reinforcing steel that is embedded in reinforced concrete. 4. Conclusions Accelerated corrosion tests were performed on concrete with and without pozzolanic additives in different aggressive environments. Findings apply to materials, region and conditions tested. Extrapolation to other circumstances must be careful. The results of OCP, LPR, and EIS measurements show that unblended concrete that were exposed to carbonation showed a slightly better

performance in corrosion of the reinforcing steel than the performance of MK- and SF-blended concrete. In harsh environments, such as environments with chlorides, the pozzolanic additives showed a better corrosion performance and the MK-blended concrete exhibited the best performance. Moreover, the SF-blended concrete showed the best corrosion performance in the combined environment with carbonation and chlorides. EIS proved to be a technique that provides more information about the corrosive process in reinforced concrete, where the blended concretes limit transport processes in the double layer and at the steel/concrete interface in harsh environments, such as environments with chlorides. In certain cases, they have better resistive characteristics, which provides greater charge-transfer resistance than chargetransfer resistance of the unblended concrete. For the combined environment, competition between the effect of carbonation and the pozzolanic reactions during exposure was evident. Particularly for concrete with SF, the effect produces a limitation of the chargetransfer capacity in the double layer with an increase of 97% of the charge-transfer resistance with respect to OPC in the final exposure cycle. The Rct values for OPC and MK at the final exposure were the lowest indicating a severe corrosion. After exposure to each of the environments, the deterioration of all concrete due to corrosion was observed.

Acknowledgements The authors express their gratitude to the Civil Engineering Laboratory (LNEC, Lisbon, Portugal) for their assistance with the discussion of the results, the Universidad del Valle (Cali, Colombia),

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716

the Centre of Excellence in New Materials (CENM), and Colciencias for their assistance with the development of this study. References [1] X. Shi, N. Xie, K. Fortune, J. Gong, Durability of steel reinforced concrete in chloride environments: an overview, Constr. Build. Mater. 30 (2012) 125–138, http://dx.doi.org/10.1016/j.conbuildmat.2011.12.038. [2] M. Glavind, Sustainability of cement, concrete and cement replacement materials in construction, in: J. Khatib (Ed.), Sustainability of Construction Materials, Woodhead Publishing, Cambridge, UK, 2009. [3] L. Bertolini, B. Elsener, P. Pedeferri, P. Polder, Corrosion of Steel in Concrete: Prevention, Diagnosis, Repair, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2004. [4] A.M. Aguirre, R.M.d. Gutiérrez, Durability of reinforced concrete exposed to aggressive conditions, Mater. Constr. 63 (309) (2013) 7–38, http://dx.doi.org/ 10.3989/mc.2013.00313. [5] M.A. Baccay, N. Otsuki, T. Nishida, S. Maruyama, Influence of cement type and temperature on the rate of corrosion of steel in concrete exposed to carbonation, Corrosion 62 (9) (2006) 811–821, http://dx.doi.org/10.5006/ 1.3278306. [6] S.-H. Han, W.-S. Park, E.-I. Yang, Evaluation of concrete durability due to carbonation in harbor concrete structures, Constr. Build. Mater. 48 (2013) 1045–1049, http://dx.doi.org/10.1016/j.conbuildmat.2013.07.057. [7] M.G. Alexander, J.R. Mackechnie, W. Yam, Carbonation of concrete bridge structures in three South African localities, Cement Concr. Compos. 29 (10) (2007) 750–759, http://dx.doi.org/10.1016/j.cemconcomp.2007.06.005. [8] M. El-Reedy, Steel-Reinforced Concrete Structures – Assessment and Repair of Corrosion, Taylor & Francis Group, LLC, 2008. [9] García T. Carmona, Modelos de Previsão da Despassivação das Armaduras em Estruturas de Concreto Sujeitas à Carbonatação, Universidade de São Paulo, São Paulo, 2005. [10] A. Nasser, A. Clément, S. Laurens, A. Castel, Influence of steel–concrete interface condition on galvanic corrosion currents in carbonated concrete, Corros. Sci. 52 (9) (2010) 2878–2890, http://dx.doi.org/10.1016/j.corsci.2010.04.037. [11] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials, Cem. Concr. Res. 41 (12) (2011) 1244–1256, http://dx.doi.org/ 10.1016/j.cemconres.2010.12.001. [12] F. Massazza, Pozzolanic cements, Cement Concr. Compos. 15 (4) (1993) 185– 214, http://dx.doi.org/10.1016/0958-9465(93)90023-3. [13] M. Cyr, 8 – Influence of supplementary cementitious materials (SCMs) on concrete durability, in: F. Pacheco-Torgal, S. Jalali, J. Labrincha, V.M. John (Eds.), Eco-Efficient Concrete, Woodhead Publishing, 2013, pp. 153–197. [14] M.C.G. Juenger, R. Siddique, Recent advances in understanding the role of supplementary cementitious materials in concrete, Cem. Concr. Res. 78 (Part A) (2015) 71–80, http://dx.doi.org/10.1016/j.cemconres.2015.03.018. [15] Y. Morozov, A.S. Castela, A.P.S. Dias, M.F. Montemor, Chloride-induced corrosion behaviour of reinforcing steel in spent fluid cracking catalyst modified mortars, Cem. Concr. Res. 47 (2013) 1–7, http://dx.doi.org/10.1016/ j.cemconres.2013.01.011. [16] S. Izquierdo, J. Diaz, R. Mejía de Gutiérrez, J. Torres-Agredo, Blended cement containing fluid catalytic cracking catalyst residue (FCC): hydration and paste microstructure, Rev. Ing. Constr. 28 (2) (2013) 141–154. [17] N. Torres-Castellanos, S. Izquierdo, J. Torres-Agredo, R. Mejía de Gutiérrez, Resistance of blended concrete containing an industrial petrochemical residue to chloride ion penetration and carbonation, Ing. Invest. 34 (1) (2014) 11–16. [18] O. Kelesßtemur, B. Demirel, Effect of metakaolin on the corrosion resistance of structural lightweight concrete, Constr. Build. Mater. 81 (2015) 172–178, http://dx.doi.org/10.1016/j.conbuildmat.2015.02.049. [19] C.S. Poon, S.C. Kou, L. Lam, Compressive strength, chloride diffusivity and pore structure of high performance metakaolin and silica fume concrete, Constr. Build. Mater. 20 (10) (2006) 858–865, http://dx.doi.org/10.1016/ j.conbuildmat.2005.07.001. [20] A.A. Ramezanianpour, Jovein H. Bahrami, Influence of metakaolin as supplementary cementing material on strength and durability of concretes, Constr. Build. Mater. 30 (2012) 470–479, http://dx.doi.org/10.1016/ j.conbuildmat.2011.12.050. [21] E. Güneyisi, M. Gesog˘lu, F. Karabog˘a, K. Mermerdasß, Corrosion behaviour of reinforcing steel embedded in chloride contaminated concretes with and without metakaolin, Compos. B Eng. 45 (1) (2013) 1288–1295, http://dx.doi. org/10.1016/j.compositesb.2012.09.085. [22] R. Mejía de Gutiérrez, C. Rodríguez, E. Rodríguez, J. Torres, S. Delvasto, Concreto adicionado con metacaolín: Comportamiento a carbonatación y cloruros, Rev. Facultad de Ing. Univ. Antioquia 48 (2009) 55–64. [23] H.-S. Kim, S.-H. Lee, H.-Y. Moon, Strength properties and durability aspects of high strength concrete using Korean metakaolin, Constr. Build. Mater. 21 (6) (2007) 1229–1237, http://dx.doi.org/10.1016/j.conbuildmat.2006.05.007. [24] R. Siddique, Utilization of silica fume in concrete: review of hardened properties, Resour. Conserv. Recycl. 55 (11) (2011) 923–932, http://dx.doi. org/10.1016/j.resconrec.2011.06.012. [25] M.I. Khan, R. Siddique, Utilization of silica fume in concrete: review of durability properties, Resour. Conserv. Recycl. 57 (2011) 30–35, http://dx.doi. org/10.1016/j.resconrec.2011.09.016.

715

[26] W.-m. Zhang, H.-j. Ba, S.-j. Chen, Effect of fly ash and repeated loading on diffusion coefficient in chloride migration test, Constr. Build. Mater. 25 (5) (2011) 2269–2274, http://dx.doi.org/10.1016/j.conbuildmat.2010.11.016. [27] M. Shekarchi, A. Rafiee, H. Layssi, Long-term chloride diffusion in silica fume concrete in harsh marine climates, Cement Concr. Compos. 31 (10) (2009) 769–775, http://dx.doi.org/10.1016/j.cemconcomp.2009.08.005. [28] S. Chao, W. Lin, Effects of silica fume and steel fiber on chloride ion penetration and corrosion behaviour of cement-based composites, J. Wuhan Univ. Technol. 28 (2) (2013) 279–284, http://dx.doi.org/10.1007/s11595-013-0679-4. [29] M.P. Kulakowski, F.M. Pereira, D.C.C.D. Molin, Carbonation-induced reinforcement corrosion in silica fume concrete, Constr. Build. Mater. 23 (3) (2009) 1189–1195, http://dx.doi.org/10.1016/j.conbuildmat.2008.08.005. [30] M. Valipour, F. Pargar, M. Shekarchi, S. Khani, Comparing a natural pozzolan, zeolite, to metakaolin and silica fume in terms of their effect on the durability characteristics of concrete: a laboratory study, Constr. Build. Mater. 41 (2013) 879–888, http://dx.doi.org/10.1016/j.conbuildmat.2012.11.054. [31] R. Siddique, J. Klaus, Influence of metakaolin on the properties of mortar and concrete: a review, Appl. Clay Sci. 43 (3–4) (2009) 392–400, http://dx.doi.org/ 10.1016/j.clay.2008.11.007. [32] H.-W. Song, V. Saraswathy, Corrosion monitoring of reinforced concrete structures – a review, Int. J. Electrochem. Sci. 2 (2007) 1–28. [33] R. Vedalakshmi, N. Palaniswamy, Analysis of the electrochemical phenomenon at the rebar–concrete interface using the electrochemical impedance spectroscopic technique, Mag. Concr. Res. 62 (3) (2010) 177–189, http://dx. doi.org/10.1680/macr.2010.62.3.177. [34] J. Wei, X.X. Fu, J.H. Dong, W. Ke, Corrosion evolution of reinforcing steel in concrete under dry/wet cyclic conditions contaminated with chloride, J. Mater. Sci. Technol. 28 (10) (2012) 905–912, http://dx.doi.org/10.1016/S1005-0302 (12)60149-2. [35] K. Tutti, Corrosion of Steel in Concrete, Swedish Cement and Concrete Research Institute, 1982, p. 4. [36] M.F. Montemor, A.M.P. Simões, M.G.S. Ferreira, Chloride-induced corrosion on reinforcing steel: from the fundamentals to the monitoring techniques, Cement Concr. Compos. 25 (4–5) (2003) 491–502, http://dx.doi.org/10.1016/ S0958-9465(02)00089-6. [37] C. Andrade, J.A. González, Quantitative measurements of corrosion measurements, Werkst. Korros. 29 (1978) 515–519. [38] C. Andrade, C. Alonso, Corrosion rate monitoring in the laboratory and on-site, Constr. Build. Mater. 10 (5) (1996) 315–328, http://dx.doi.org/10.1016/09500618(95)00044-5. [39] F. Rodríguez Gómez, Técnicas electroquímicas de corriente directa para la medición de la velocidad de corrosión. Resistencia a la polarización, in: J. Genescá (Ed.), Tecnicas electroquímicas para el control y estudio de la corrosión, UNAM, México D.F., México, 2003. [40] S.J. Ford, J.D. Shane, T.O. Mason, Assignment of features in impedance spectra of the cement-paste/steel system, Cem. Concr. Res. 28 (12) (1998) 1737–1751, http://dx.doi.org/10.1016/S0008-8846(98)00156-2. [41] P. Lay, P.F. Lawrence, N.J.M. Wilkins, D.E. Williams, An a.c. impedance study of steel in concrete, J. Appl. Electrochem. 15 (5) (1985) 755–766, http://dx.doi. org/10.1007/BF00620572. [42] M. Ismail, M. Ohtsu, Corrosion rate of ordinary and high-performance concrete subjected to chloride attack by AC impedance spectroscopy, Constr. Build. Mater. 20 (7) (2006) 458–469, http://dx.doi.org/10.1016/ j.conbuildmat.2005.01.062. [43] D.A. Koleva, J.H.W. de Wit, K. van Breugel, L.P. Veleva, E. van Westing, O. Copuroglu, et al., Correlation of microstructure, electrical properties and electrochemical phenomena in reinforced mortar. Breakdown to multi-phase interface structures. Part II: pore network, electrical properties and electrochemical response, Mater. Charact. 59 (6) (2008) 801–815, http://dx. doi.org/10.1016/j.matchar.2007.06.016. [44] R. San Nicolas, M. Cyr, G. Escadeillas, Performance-based approach to durability of concrete containing flash-calcined metakaolin as cement replacement, Constr. Build. Mater. 55 (2014) 313–322, http://dx.doi.org/ 10.1016/j.conbuildmat.2014.01.063. [45] G. Batis, P. Pantazopoulou, S. Tsivilis, E. Badogiannis, The effect of metakaolin on the corrosion behaviour of cement mortars, Cement Concr. Compos. 27 (1) (2005) 125–130, http://dx.doi.org/10.1016/j.cemconcomp.2004.02.041. [46] R. Talero, Synergic effect of Friedel’s salt from pozzolan and from OPC coprecipitating in a chloride solution, Constr. Build. Mater. 33 (2012) 164–180, http://dx.doi.org/10.1016/j.conbuildmat.2011.12.040. [47] M.F. Montemor, M.P. Cunha, M.G. Ferreira, A.M. Simões, Corrosion behaviour of rebars in fly ash mortar exposed to carbon dioxide and chlorides, Cement Concr. Compos. 24 (1) (2002) 45–53, http://dx.doi.org/10.1016/S0958-9465 (01)00025-7. [48] A.K. Suryavanshi, R. Narayan Swamy, Stability of Friedel’s salt in carbonated concrete structural elements, Cem. Concr. Res. 26 (5) (1996) 729–741, http:// dx.doi.org/10.1016/S0008-8846(96)85010-1. [49] M. Frías, S. Goñi, Accelerated carbonation effect on behaviour of ternary Portland cements, Compos. B Eng. 48 (2013) 122–128, http://dx.doi.org/ 10.1016/j.compositesb.2012.12.008. [50] M. Saillio, V. Baroghel-Bouny, F. Barberon, Chloride binding in sound and carbonated cementitious materials with various types of binder, Constr. Build. Mater. 68 (2014) 82–91, http://dx.doi.org/10.1016/j.conbuildmat.2014.05.049. [51] H. Kuosa, R.M. Ferreira, E. Holt, M. Leivo, E. Vesikari, Effect of coupled deterioration by freeze–thaw, carbonation and chlorides on concrete service

716

[52]

[53]

[54]

[55]

A.M. Aguirre-Guerrero et al. / Construction and Building Materials 121 (2016) 704–716 life, Cement Concr. Compos. 47 (2014) 32–40, http://dx.doi.org/10.1016/j. cemconcomp.2013.10.008. V. Triana, J. Lizarazo-Marriaga, J. Olaya, Steel corrosion assessment by electrochemical impedance on metakaolin blended mortars, Mater. Res. 16 (6) (2013) 1457–1464, http://dx.doi.org/10.1590/S1516-14392013005000178. M.F. Montemor, A.M.P. Simões, M.M. Salta, Effect of fly ash on concrete reinforcement corrosion studied by EIS, Cement Concr. Compos. 22 (3) (2000) 175–185, http://dx.doi.org/10.1016/S0958-9465(00)00003-2. M.A. Pech-Canul, P. Castro, Corrosion measurements of steel reinforcement in concrete exposed to a tropical marine atmosphere, Cem. Concr. Res. 32 (3) (2002) 491–498, http://dx.doi.org/10.1016/S0008-8846(01)00713-X. V. Feliu, J.A. González, C. Andrade, S. Feliu, Equivalent circuit for modelling the steel-concrete interface. I. Experimental evidence and theoretical predictions,

Corros. Sci. 40 (6) (1998) 975–993, http://dx.doi.org/10.1016/S0010-938X(98) 00036-5. [56] J. Sobhani, M. Najimi, Electrochemical impedance behaviour and transport properties of silica fume contained concrete, Constr. Build. Mater. 47 (2013) 910–918, http://dx.doi.org/10.1016/j.conbuildmat.2013.05.010. [57] C. Alonso, C. Andrade, J.A. González, Relation between resistivity and corrosion rate of reinforcements in carbonated mortar made with several cement types, Cem. Concr. Res. 18 (5) (1988) 687–698, http://dx.doi.org/10.1016/0008-8846 (88)90091-9. [58] J. Torres, R. Mejía-de-Gutiérrez, F. Puertas, Effect of kaolin treatment temperature on mortar chloride permeability, Mater. Constr. 57 (285) (2007) 61–69, http://dx.doi.org/10.3989/mc.2007.v57.i285.39.