Electrochemical Study of Clean and Pre-Corroded Reinforcements Embedded in Mortar Samples with Variable Amounts of Chloride Ions

The present study investigates the possibility of re-surfacing previously corroded reinforcements and the suitability of the two electrochemical techniques that are widely used to determine the state of corrosion of steel (the corrosion potential Ecorr and the corrosion rate icorr). In order to test this, 32 pre-corroded B500SD reinforcing steel bars have been used for one year, where half of the bars have been cleaned to eliminate corrosion products. The other half have been maintained with the generated corrosion products. Subsequently, the bars have been embedded in cement mortar samples with variable amounts of chloride ion, and Ecorr and icorr have been measured for 250 days. The results showed that it is not possible to rework the reinforcement without removing corrosion products and that it is not possible to predict the passive or active state of steel by measuring Ecorr only.


Introduction
Reinforcement corrosion is accepted as the main cause of the reduction in the service life of reinforced concrete structures (RCSs) [1][2][3]. The enormous economic impact of this problem, due to the direct and indirect costs involved, has led to a vast development of new technologies and materials with the purpose of increasing the durability of RCSs [4][5][6][7]. In the USA, direct costs due to corrosion of RCSs infrastructure are estimated at 0.25% of GNP, which corresponds to USD 16.6 billion per year [8].
The steel reinforcements embedded in the RCS are in a passive state that are protected against corrosion. This protection is due to the existence of a passive layer formed at the steel/concrete interface, which is self-healing and very thin around 10 nm [9]. Its formation and stability are guaranteed by the high alkalinity of the concrete, usually between the range of pH 13-14, and by the existence of an appropriate electrochemical potential [10].
The loss of the passivity of the RCS reinforcement is due, in most cases, to the following factors: the presence of despassivating ions, essentially chlorides, in sufficient quantity to locally break up the passivating layers [11]; or the decrease in the pH of the concrete, due to the effect of the CO 2 present in the atmosphere [12][13][14]. In addition to the triggering factor that induces corrosion, the environment in which the structure is located determines the variables that most significantly influence its behavior. In the case of marine environments, the chloride diffusion coefficient, the concentration of chlorides on the surface and the thickness of the rebar coating are the most determining parameters to be able to evaluate its behavior, while in the case of underground tunnels, although corrosion is induced by chlorides, the formation of NaCl crystals is a parameter to take into account [15,16].
For decades, numerous investigations have been carried out to explain the role of the factors that trigger corrosion of rebars, especially chloride ions [17,18], whose mechanism can be seen in the Figure 1. is induced by chlorides, the formation of NaCl crystals is a parameter to take into account [15,16] For decades, numerous investigations have been carried out to explain the role of the factors that trigger corrosion of rebars, especially chloride ions [17,18], whose mechanism can be seen in the Figure 1. The critical threshold of chloride ions (Ccrit) needed to despassivate the steel is one of the critical parameters in the prediction and evaluation of corrosion [17]. The knowledge The critical threshold of chloride ions (C crit ) needed to despassivate the steel is one of the critical parameters in the prediction and evaluation of corrosion [17]. The knowledge of C crit is fundamental when establishing the requirements to achieve structures with sufficient durability and to evaluate the service life or residual life of existing structures [18]. The onset of corrosion due to the effect of chlorides in the concrete reinforcements has been related to an increase in the ratio [Cl − ]/[OH − ] above a certain value, which was initially set by Haussman at 0.6 [19][20][21][22][23][24]. However, it was later found that it can be affected by other parameters such as the amount of air retained in the holes of the steel-concrete interface or the presence of defects [23,[25][26][27][28][29][30]. Numerous studies depict that there is a significant dispersion in C crit values. In most cases, the limits are set in relation to the weight of the cement: For example, for concrete with a water/cement ratio of 0.5 mainly located in marine environments in the intertidal zone in Europe with a limit value of 0.5% suggested [31]. In Spain, the mandatory regulations establish a durability strategy based on the service life of the structure, but always with a limit of Cl − of 0.4% [32]. In the United States, the thresholds are set according to the exposure class of the structure, ranging from 0.15% to 0.30%, and may reach 1% for structures located in dry environments [33]. At present, there are numerous investigations that model the corrosion behavior of the reinforcements embedded in the reinforced concrete and the influence of the concrete quality, the concentration of chlorides and the coating and the cracking, on the rate of corrosion, under different environmental conditions [34][35][36][37]. In addition, these investigations, validated on the corresponding experimental works, have allowed to simulate the influence that the size of the pits have on the geometry of the fissures and how the accumulation of oxide in the pits influence the adhesion between concrete and steel [38].
The realization of potential maps according to ASTM C876-09 [39] is the electrochemical technique most commonly used to diagnose the risk of corrosion in RCSs [2,40]. However, the results of the evaluation of RCSs with corroded reinforcements may indicate different degrees of corrosion or probability of corrosion depending on the technique used for corrosion assessment [41,42]. It is generally accepted that the measurements of corrosion potential (E corr ) should be completed with other procedures [40,43]. The measurement of corrosion rate is a quantitative technique that is known for decades [44]. However there have been found large variations in the values of corrosion rate for narrow ranges of values of corrosion potential [45].
Taking into account the previous premises, the aims of the work are to verify the possibility of re-surfacing previously corroded reinforcements and to verify the suitability of two electrochemical techniques widely used to determine the corrosion status of reinforcement by correlating the E corr and the i corr measurements.

Experimental Program
In this study, two issues related to corrosion of RCS reinforcements have been investigated: (i) The possibility of re-surfacing previously corroded reinforcements and (ii) the verification of the suitability of two electrochemical techniques widely used to determine the corrosion status of reinforcements. In order to test this, 32 B500SD reinforcing steel bars of 6 mm diameter and 120 mm length have been used. The chemical composition of the steel is shown in Table 1. The analyses have been carried out with an Optical Emission Spectrometer by Arc/Spark model SPECTROMAX. The rebars have been embedded in cement mortar samples of 80 × 55 × 20 mm 3 with a cement/sand/water dosage of 1/3/0.5. CEM I 42.5R Portland cement according to RC-16 standard [46], siliceous sand with a maximum size of 0.4 mm and drinking water supplied by Canal de Isabel II in Madrid were used. The physical characteristics and chemical composition of the cement and the sand used can be seen in Table 2. The steel rebars were embedded in the specimens, leaving 5 cm outside, in which the risk of differential aeration at the triple atmosphere/mortar/steel interface was eliminated by means of adhesive tape. The specimens were demolded after 24 h and cured in a wet chamber for 28 days at a temperature of 20 ± 2 • C and a humidity above 90%. Subsequently, they were subjected to a constant anodic polarization of 20 µA/cm 2 during one year. In the photograph (Figure 2a), one of the specimens can be seen after the passage of the electric current. The samples were mixed with an addition of 2% of Cl − in relation to the weight of the cement to ensure that the anodic polarization would lead to corrosion of the reinforcements and would not cause the electrolysis of the water in the pore network.  The steel rebars were embedded in the specimens, leaving 5 cm outside, in which the risk of differential aeration at the triple atmosphere/mortar/steel interface was eliminated by means of adhesive tape. The specimens were demolded after 24 h and cured in a wet chamber for 28 days at a temperature of 20 ± 2 °C and a humidity above 90%. Subsequently, they were subjected to a constant anodic polarization of 20 µA/cm 2 during one year. In the photograph (Figure 2a), one of the specimens can be seen after the passage of the electric current. The samples were mixed with an addition of 2% of Cl − in relation to the weight of the cement to ensure that the anodic polarization would lead to corrosion of the reinforcements and would not cause the electrolysis of the water in the pore network. Once the pre-corrosion of the rebars was complete, the mortar samples were then broken and the rebars removed. Half of the rebars (16 pieces) were chosen at random, and the corrosion products were completely removed by dissolving the iron oxides and hydroxides in 50% HCl inhibited with 4g/l urotropin (hexamethylenetetramine). With the clean rebars (CLN), 8 specimens similar to those of the first phase were performed, but with variable amounts of chloride ion of 0.0%; 0.2%; 0.4%; 0.6%; 0.8%; 1.0%; 1.5%; and 2.0% in relation to the weight of cement and with the incorporation of a third rebar located in the center to facilitate the electrochemical measurements (Figure 2b). The same process has been carried out with the rebars that hold the corrosion products. Then the 16 specimens were kept in a wet chamber for 250 days at a temperature of 20 ± 2 °C and a humidity Once the pre-corrosion of the rebars was complete, the mortar samples were then broken and the rebars removed. Half of the rebars (16 pieces) were chosen at random, and the corrosion products were completely removed by dissolving the iron oxides and hydroxides in 50% HCl inhibited with 4g/l urotropin (hexamethylenetetramine). With the clean rebars (CLN), 8 specimens similar to those of the first phase were performed, but with variable amounts of chloride ion of 0.0%; 0.2%; 0.4%; 0.6%; 0.8%; 1.0%; 1.5%; and 2.0% in relation to the weight of cement and with the incorporation of a third rebar located in the center to facilitate the electrochemical measurements (Figure 2b). The same process has been carried out with the rebars that hold the corrosion products. Then the 16 specimens were kept in a wet chamber for 250 days at a temperature of 20 ± 2 • C and a humidity above 90%. During this time the corrosion potential (E corr ), and corrosion rate were periodically recorded using an AUTOLAB/PGSTAT302N potentiostat in which the outer bars were used as working electrodes and the central bar as an auxiliary electrode  Figure 3). The reference electrode used was Cu/CuSO 4 . The corrosion rate was measured by the corrosion current density (i corr ) obtained by measuring the polarization resistance (R p ) by the Stern and Geary equation [47]: where the value of 26 mV for the constant B was chosen [48]. The value of R p was obtained by applying polarization of 10 mV and measuring the current response after 15 s. The reference values for predicting the corrosion state as a function of E corr and i corr are shown in Table 3 [39,49].
odically recorded using an AUTOLAB/PGSTAT302N potentiostat in which the outer bars were used as working electrodes and the central bar as an auxiliary electrode ( Figure 3). The reference electrode used was Cu/CuSO4. The corrosion rate was measured by the corrosion current density (icorr) obtained by measuring the polarization resistance (Rp) by the Stern and Geary equation [47]: where the value of 26 mV for the constant B was chosen [48]. The value of Rp was obtained by applying polarization of 10 mV and measuring the current response after 15 s. The reference values for predicting the corrosion state as a function of Ecorr and icorr are shown in Table 3 [39,49]. With respect to the thresholds of the Ecorr value there is a very broad consensus throughout the scientific and technical community, most likely due to the enormous diffusion at the international level of the ASTM C 876-09 standard. Regarding icorr thresholds, it is widely accepted that icorr < 0.1 µA/cm 2 corresponds to steel in a passive state, and above this value steel corrodes and icorr > 1 µA/cm 2 is very dangerous [42,44,[50][51][52][53].   Figure 4 shows the evolution in time of the icorr of all the analyzed rebars. Each of the data have been obtained as the arithmetic mean of the two rebars in the same state and embedded in the same specimen. The abbreviation CLN indicates that the rebar has been  With respect to the thresholds of the E corr value there is a very broad consensus throughout the scientific and technical community, most likely due to the enormous diffusion at the international level of the ASTM C 876-09 standard. Regarding i corr thresholds, it is widely accepted that i corr < 0.1 µA/cm 2 corresponds to steel in a passive state, and above this value steel corrodes and i corr > 1 µA/cm 2 is very dangerous [42,44,[50][51][52][53]. Figure 4 shows the evolution in time of the i corr of all the analyzed rebars. Each of the data have been obtained as the arithmetic mean of the two rebars in the same state and embedded in the same specimen. The abbreviation CLN indicates that the rebar has been cleaned while the abbreviation COR indicates that the rebar keeps the corrosion products. The final numbers indicate the amount of Cl − , expressed as a percentage by weight of cement, present in the specimens. It can be seen that, in most cases, the starting point is an i corr that decreases in a very similar way during the first 35 days of exposure. After this period of exposure, the i corr remains at very stable values most probably because of the humidity of the environment remaining constant. Figure 5 shows the values corresponding to the i corr of all the rebars after 250 days. cleaned while the abbreviation COR indicates that the rebar keeps the corrosion products. The final numbers indicate the amount of Cl − , expressed as a percentage by weight of cement, present in the specimens. It can be seen that, in most cases, the starting point is an icorr that decreases in a very similar way during the first 35 days of exposure. After this period of exposure, the icorr remains at very stable values most probably because of the humidity of the environment remaining constant. Figure 5 shows the values corresponding to the icorr of all the rebars after 250 days.  [32,33,54]. The embedded rebars in specimens with 0.6% Cl − are maintained with low corrosion levels. Rebars in specimens with 0.8% Cl − are maintained with icorr between 1 and 2 µA/cm 2 . With higher amounts of Cl − , the rebars maintain icorr at approximately 10 µA/cm 2 . Similar results were obtained in previous investigations, observing that a greater amount of chloride ion present in the mortar, generates a higher rate of corrosion in the rebars [49,55,56].

Results and Discussion
All the rebars that have kept the corrosion products (COR) exhibit very high icorr values (10 µA/cm 2 ) regardless of the amount of Cl − present in the mortar. Moreover, the icorr of the bars does not depend on the amount of chloride ion, which proves the impossibility of reworking reinforcements with thick corrosion products when these are not eliminated. This is because the icorr in the reinforcements that corrode in an active state are sufficient to maintain an acid pH in the steel/corrosion products interface within such an alkaline medium as concrete, so that, once corrosion has been triggered, Cl − are not necessary to maintain it [49,55,57]. Similar results were obtained by Miranda et al., when they observed that the higher the degree of pre-corrosion of the steel, the higher its corrosion rate, even in chloride-free environments [56,57]. If we compare the evolution in time of the graphs in Figures 4 and 6, we can see that the Ecorr suffers a greater variation than the icorr, reaching, in some cases (corroded rebar introduced in a specimen with 0.4% of Cl − ), values, during the whole studied period, typical of the passive, uncertain and active state, since the measurement of the potential for corrosion varies enormously depending on various factors, such as temperature and humidity [58]. It can also be seen that at 250 days, the Ecorr value depends much more on the amount of Cl − in the specimen than on the passive or active state of the rebar, so that the Ecorr of rebars in specimens containing the same amount of Cl − are very similar, regardless of whether the rebar has been embedded in the clean specimen or with the corrosion products (Figures 6 and 7).  [32,33,54]. The embedded rebars in specimens with 0.6% Cl − are maintained with low corrosion levels. Rebars in specimens with 0.8% Cl − are maintained with i corr between 1 and 2 µA/cm 2 . With higher amounts of Cl − , the rebars maintain i corr at approximately 10 µA/cm 2 . Similar results were obtained in previous investigations, observing that a greater amount of chloride ion present in the mortar, generates a higher rate of corrosion in the rebars [49,55,56].
All the rebars that have kept the corrosion products (COR) exhibit very high i corr values (10 µA/cm 2 ) regardless of the amount of Cl − present in the mortar. Moreover, the i corr of the bars does not depend on the amount of chloride ion, which proves the impossibility of reworking reinforcements with thick corrosion products when these are not eliminated. This is because the i corr in the reinforcements that corrode in an active state are sufficient to maintain an acid pH in the steel/corrosion products interface within such an alkaline medium as concrete, so that, once corrosion has been triggered, Cl − are not necessary to maintain it [49,55,57]. Similar results were obtained by Miranda et al., when they observed that the higher the degree of pre-corrosion of the steel, the higher its corrosion rate, even in chloride-free environments [56,57]. The evolution of the E corr in the same period of time can be seen in Figure 6. Figure 7 depicts in more detail the E corr values reached by all the bars after 250 days.
the Ecorr suffers a greater variation than the icorr, reaching, in some cases (corroded rebar introduced in a specimen with 0.4% of Cl − ), values, during the whole studied period, typical of the passive, uncertain and active state, since the measurement of the potential for corrosion varies enormously depending on various factors, such as temperature and humidity [58]. It can also be seen that at 250 days, the Ecorr value depends much more on the amount of Cl − in the specimen than on the passive or active state of the rebar, so that the Ecorr of rebars in specimens containing the same amount of Cl − are very similar, regardless of whether the rebar has been embedded in the clean specimen or with the corrosion products (Figures 6 and 7).  If the values of Ecorr and icorr are represented in a system of axes together with the thresholds that delimit the passive and active states, it is possible to check the validity of the Ecorr measurements as a predictor of the corrosion state of a rebar (Table 3). Choosing as ordinate axis the value of Ecorr and as x-axis the value of icorr and marking by vertical and horizontal lines the thresholds of these values, the space is divided into 6 quadrants ( Figure 8). In quadrants A and F, corrosion states coincide that predict the measurements of the Ecorr and icorr: Quadrant A corresponds to measurements that indicate the passive state If we compare the evolution in time of the graphs in Figures 4 and 6, we can see that the E corr suffers a greater variation than the i corr , reaching, in some cases (corroded rebar introduced in a specimen with 0.4% of Cl − ), values, during the whole studied period, typical of the passive, uncertain and active state, since the measurement of the potential for corrosion varies enormously depending on various factors, such as temperature and humidity [58]. It can also be seen that at 250 days, the E corr value depends much more on the amount of Cl − in the specimen than on the passive or active state of the rebar, so that the E corr of rebars in specimens containing the same amount of Cl − are very similar, regardless of whether the rebar has been embedded in the clean specimen or with the corrosion products (Figures 6 and 7).
If the values of E corr and i corr are represented in a system of axes together with the thresholds that delimit the passive and active states, it is possible to check the validity of the E corr measurements as a predictor of the corrosion state of a rebar (Table 3). Choosing as ordinate axis the value of E corr and as x-axis the value of i corr and marking by vertical and horizontal lines the thresholds of these values, the space is divided into 6 quadrants (Figure 8).
the Ecorr measurements as a predictor of the corrosion state of a reb as ordinate axis the value of Ecorr and as x-axis the value of icorr an and horizontal lines the thresholds of these values, the space is div (Figure 8). In quadrants A and F, corrosion states coincide that predict th Ecorr and icorr: Quadrant A corresponds to measurements that ind while the measurements of quadrant F indicate the active state. In and E define corrosion states with total discrepancy between the pr icorr. Quadrants C and D define corrosion states where the Ecorr pred Figure 9 depicts the Ecorr and icorr values for each of the spe measurements. The graphs on the left column correspond to the cle the right column corresponds to the values of the rebars with corr In the two graphs located in the same row, the amount of Cl − pre coincides. Figure 10 shows the values of all the rebars together. In quadrants A and F, corrosion states coincide that predict the measurements of the E corr and i corr : Quadrant A corresponds to measurements that indicate the passive state while the measurements of quadrant F indicate the active state. In contrast, quadrants B and E define corrosion states with total discrepancy between the prediction of the E corr and i corr . Quadrants C and D define corrosion states where the E corr predicts an uncertain state. Figure 9 depicts the E corr and i corr values for each of the specimens obtained in all measurements. The graphs on the left column correspond to the clean rebars (CLN) while the right column corresponds to the values of the rebars with corrosion products (COR). In the two graphs located in the same row, the amount of Cl − present in the test pieces coincides. Figure 10 shows the values of all the rebars together.
The analysis of all the graphs in Figure 9, shows that only the predictions of the E corr and i corr for rebars with an ongoing corrosion process and containing at least 0.8% Cl − coincide. In these cases, all the data points of the graphs are located in quadrant F indicated in Figure 8, regardless of whether the rebars are clean or with corrosion products. In addition, the E corr value becomes more electronegative as the amount of Cl − present in the specimen increases and regardless of the state of the rebar. This trend is more pronounced for very low amounts of Cl − and does not depend on the corrosion state of the reinforcement.
The analysis in Figure 10 allows verification of (i) the rebars in passive state (i corr < 0.1 µA/cm 2 ) show a large number of E corr values corresponding to high probabilities of corrosion (quadrant A) or uncertain states (quadrant C), and (ii) the rebars in active state (i corr > 0.1 µA/cm 2 ) that show a large number of E corr values corresponding to low probabilities of corrosion (quadrant F) or uncertain states (quadrant D). These results show the impossibility of predicting the passive or active state of the steel only by means of E corr measurements [58].

Conclusions
Previously corroded reinforcing steel bars have been soaked in mortar samples with varying contents of Cl-in two different states: previously pickled and with the corrosion products. They have been kept for 250 days in a humid chamber, and E corr and i corr have been measured and obtained with the following conclusions being drawn:

•
All the rebars that have kept the corrosion products showed a very high i corr and a similar value, approximately 10 µA/cm 2 , regardless of the number of chlorides present in the specimen. • Clean rebars (CLN) embedded in specimens with 0.6 Cl − remain with uncertain i corr (values between 0.1 and 1.0 µA/cm 2 ).

•
The values of the E corr measurement depended more on the amount of Cl − present in the specimen than on the passive or active state of the rebars.

•
Only the predictions of the E corr and i corr coincided in bars embedded in specimens with at least a 0.8% Cl − by weight cement ratio, regardless of whether the rebar is clean or maintains the corrosion products.

•
The low correlation of the results obtained in the E corr and i corr in different situations, makes it impossible to predict the passive or active state of the steel solely based on E corr measurements.

•
To repair a concrete structure corroded by the effect of chloride ions, the concrete that surrounds the rebars must be removed so that all the corrosion products generated on the surface of the rebars can be eliminated. If the complete removal of the corrosion products is not achieved, even if a repair mortar is placed on it, the rebars will remain in active state.