The role of inhibitors on the repassivation of pitting corrosion of carbon steel in synthetic carbonated concrete pore solution

Ze Hua Dong,Wei Shi,Guo An Zhang,Xing Peng Guo

DOI: https://doi.org/10.1016/j.electacta.2011.04.120

IF: 6.6

2011-01-01

Electrochimica Acta

Abstract:The semiconductive properties and pitting susceptibility of passive films formed on carbon steel in synthetic carbonated concrete pore solution containing Cl − ions were studied by Mott–Schottky (M–S) plots and electrochemical noise (EN). The results show that both nitrite and tetraethylenepentamine (TEPA) can reduce the donor density in passive film, leading to a more positive film breakdown potential. Moreover, nitrite can rehabilitate metastable pits and passivate the broken passive film more quickly than TEPA. Polarization curves also show that nitrite can increase the pitting potential more than TEPA at same concentration. Atomic force microscopy (AFM) imaging shows that Cl − ions could induce large metastable pits on passive film, but the size of pits decreases and the density of pits increases with addition of nitrite. However, AFM imaging fails to detect the metastable pits under TEPA adsorbed film in spite of a smooth-like image on the steel surface is observed. The force curves of AFM indicate a stiffness (elasticity) order of films: nitrite oxidation film > original passive film > enamine adsorbed film. Keywords Carbon steel Pitting corrosion Inhibition Mott–Schottky plot AFM 1 Introduction Corrosion of rebar in concrete is generally recognized as one of the major causes to deteriorate reinforced concrete structures, which is primarily induced by the carbonation of CO 2 from atmosphere and chlorides from de-icing salts or seawater [1] . Passive film formed on rebar in high alkaline concrete (pH > 12.6) can be destroyed in the presence of chlorides, especially when the pH in concrete pore solution (CPS) decreases to 9–11 due to the solution carbonation [2] . Localized corrosion starts at the active site of rebar when the chloride concentration at the rebar/concrete interface exceeds a critical threshold [3–6] , or when the ratio of [Cl − ]/[OH − ] in concrete is higher than 0.66–1.4 [7] . For rehabilitation of the concrete structures suffering severe corrosion, corrosion inhibitor is regarded as one of the simplest and most cost-effective techniques [8,9] . As one of most commonly used inhibitive admixtures, nitrite can react with ferrous ions generating a Fe 2 O 3 protective layer and reducing the corrosion rate of rebar [10–12] . Passive film on steel is a highly defective oxide film similar to n-type semiconductor (dopant level of 10 20 –10 21 cm −3 by Mott–Schottky analysis) [13–15] , and the electronic properties of passive film have strong ties with their stability in the presence of Cl − ions. Freire et al. [16] recently found that the ratio of Fe 2+ /Fe 3+ in passive film decreased with the increase of applied potential according to the results of XPS. Cheng [17] investigated the passive film formed on iron in borate buffer solution by Mott–Schottky (M–S) analysis and concluded that the donor density decreased exponentially with the increase of film forming potential (pre-passivated through potentiostatic polarization for 2 h). Although many studies concerning the film properties have been reported, the interaction between the passive film and inhibitors are rarely reported. Electrochemical noise (EN) is the spontaneous fluctuations of current and potential simultaneously without external polarization, which can provide principal information about the initiation, propagation and repassivation processes of pitting corrosion [18,19] . EN has also been used to estimate the rehabilitation role of inhibitors on the defects in passive film based on the statistic analysis of nucleation rate, lifespan and charge of metastable pits [20] . In spite of the advantages of the EN method, its interpretation is often difficult because of multiple factors influencing the EN signal such as cathodic process [21] , aliasing noises [22] and DC drift [23] , which make EN method suitable for qualitative other than quantitative analysis in most cases. Visual observation had been extensively introduced to image the nucleation and evolution of pitting corrosion. For example, Wang et al. [24] found by SEM that the metastable pits of pure iron overlapped and appeared as cluster. However, SEM morphology cannot provide in situ images of the dynamic initiation and propagation of pits in solutions. On the other hand, atomic force microscopy (AFM) is capable of imaging surface topography in real time. Pidaparti and Patel [25] applied AFM to observe the growth characteristic of a single pit on aluminium. Martin et al. [26] studied the local corrosion of a 304L stainless steel in a chloride-containing solution, and found that most pits initiated at strain hardened areas. Cappella et al. [27] revealed that the topography of the stearic film in water was rougher than that in air due to the molecule rearrangement in water [28] . Many researchers have utilized the microscopic imaging function of AFM although many physical properties, such as the stability and adhesion strength of adsorbed film, can be studied by the force–distance ( F – D ) curves of AFM. In the present work, we focused on the interaction between inhibitors and the passive film to understand the structure of passive film and the rehabilitation mechanisms of organic and inorganic inhibitors on localized corrosion by means of M–S analysis, EN and AFM. 2 Experimental 2.1 Solutions and materials A synthetic carbonated concrete pore solution (CPS) was prepared by saturated Ca(OH) 2 with addition of NaHCO 3 to simulate a low degree of carbonated concrete with a resulting pH value of 12.0. All solutions were prepared from analytical grade reagents and deionised water. 0.01, 0.03 and 0.05 mol/L of NaNO 2 or tetraethylenepentamine (NH 2 (CH 2 CH 2 NH) 3 CH 2 CH 2 NH 2 , TEPA) were added as inhibitors. Two cylindrical specimens with the size of ϕ 8 mm × 5 mm were machined from a rebar of Q345 carbon steel with chemical composition (wt%): C 0.26, Mn 0.56, P 0.009, Si 0.063 and S 0.031, and Fe balance. These specimens, which were embedded in Teflon wedge-shaped rings leaving 0.5 cm 2 area exposed to solution, were used as working electrodes (WEs). The surfaces of WEs were ground with 400, 800, 1200 grit emery paper in turns, then cleaned in deionised water and degreased in ethanol. All 10 mm × 10 mm Q345 steel sheet specimens were firstly ground by the same procedures as mentioned above, and then cleaned in deionised water, mechanically polished with 0.5 μm and 0.05 μm alumina pastes for 1 h at 800 rpm in turns. The specimens were cleaned ultrasonically in deionised water, degreased in methanol, dried in cold air, and then mounted in a thin electrochemical cell filled with test solutions for in situ AFM observations. After AFM analysis, the specimens were taken out and cleaned with ethanol and deionised water, dried in N 2 for further investigation by scanning electron microscopy (Sirion200 SEM, FEI, The Netherlands). 2.2 Electrochemical measurements For electrochemical measurements, the edge of the specimen contacting to Teflon ring was coated with lacquer to avoid possible crevice corrosion. WEs were pre-passivated in CPS through potentiostatic polarization for 2 h before the EN and M–S measurements by a CS350 electrochemical workstation (Corrtest Instrument, China). Conventional three-electrode system was employed with a Pt plate as counter electrode and a saturated calomel electrode (SCE) as reference electrode. Potential and current noises were measured simultaneously by a voltage follower and an auto zero-resistance ammeter with a sampling frequency of 5 Hz. Two low-pass filters with cut-off frequency of 20 Hz were employed to remove aliasing noises. Potentiodynamic curves were obtained by polarization in the anodic direction at scan rate of 10 mV/min. M–S plots were measured to investigate the electronic properties of passive film under polarization potential from −0.8 V to 0.65 V (vs. SCE) at a 50 mV/s scanning rate. The scanning potential was superimposed with a 10 mV, 1 kHz AC signal. The space charge capacitance ( C SC ) in passive film as a function of applied potential ( E m ) was calculated from C SC = − 1/( ωZ ' ' ), where Z ' ' is the imaginary part of the AC impedance, ω is the angular frequency. Electric properties of semiconductors can be obtained from the M–S curves on the basis of the well-known M–S Eq. [29] : (1) 1 C SC 2 = 2 e N D ε 0 ε r ( E m − E fb − k T e )     For   n − type   semiconductor (2) 1 C S C 2 = − 2 e N A ε 0 ε r ( E m − E fb + k T e )     For   p − type   semiconductor where ɛ is the dielectric constant of the passive films, ɛ 0 is the permittivity of free space charge (8.854 × 10 −14 F/cm), e is the electron charge (1.602 × 10 −19 C), N D and N A are the donor and acceptor densities, respectively. E fb is the flat band potential, k is the Boltzmann constant (1.38 × 10 −23 J/K), and T is the absolute temperature. kT / e can be negligible as it is only 25 mV at room temperature. N D and N A can be determined from the slope of 1 / C SC 2 vs. E m assuming the dielectric constant of the passive film on the carbon steel as 15.6 [30] . E fb can be obtained from the linear extrapolation to 1 / C S C 2 = 0 . The validity of M–S analysis is based on the assumption that the capacitance of the space charge layer is much smaller than the double layer (Helmholtz) capacitance. All the electrochemical measurements were carried out at ambient temperature (∼22 °C), and all the potential values reported in this text, unless clearly indicated, are referring to SCE. 2.3 In situ AFM measurement AFM (SPA400, Seiko) was used to image the topography of metastable pits on a polished carbon steel specimen. For pitting nucleation, the specimen was pre-immersed in CPS containing 0.01 mol/L Cl − ions. All images were taken in contacting mode at a scan rate of 1 Hz. To avoid the erosion of Cl − ions on AFM tips, a NSC18/Cr-AuBS tip with force constant of 3.5 N/m was employed. The force–distance ( F – D ) curves were recorded from the vertical displacement ( Z ) of the specimen (1240 nm to −1240 nm) by measuring the cantilever deflection [28] . Z is obtained by approaching the specimen to the tip along the vertical axis (Z axis), which consists of three parts. (3) Z = D + Δ s c + Δ s s where D is the tip-to-specimen distance, Δ s c , Δ s s are the deformation of the cantilever and the specimen, respectively. When the tip contacts with the surface of specimen, Eq. (3) can be simplified as Z = Δs c + Δs s , and the force can be obtained by Eq. (4) developed by Cappella et al. [27] . (4) F = [ k c k s / ( k s + k c ) ] Z = k eff Z where k c and k s denote the elastic constants of the cantilever and the specimen, respectively. The effective elastic constant k eff is a parallel combination value of k c and k s . Assuming the k c is constant, the k eff is exclusively related to k s and it can be calculated from the slopes of contact line in the approaching process [28] . 3 Results 3.1 EN measurements Fig. 1 shows the potential and current fluctuations of Q345 steel after immersed in CPS containing 0.1 mol/L Cl − ions with addition of different inhibitors for 10 h. It can be seen that the strong potential transients have characteristics of fast drop and slow recovery (50–120 s lifespan), while the current transients are characterised by a simultaneous fast rise and then slow recovery with the same lifespan. Which are the characteristics of the nucleation/death of metastable pits under the attack of Cl − ions on passive film. However, with the addition of 0.03 mol/L TEPA or NO 2 − ions, the lifespan and amplitude of current transients decrease comparing to those without inhibitors, indicating that the inhibitors accelerate the rehabilitation of metastable pits [31] . In addition, Fig. 1 shows a negative shift for the electrode in CPS containing TEPA but a positive shift for the electrode in CPS containing NO 2 − ions. This difference could be ascribed to the fact that TEPA and NO 2 − ions mainly impede the cathodic and anodic process, respectively. Although the average amplitude of current transients decreases after inhibitors were added, interestingly, the average amplitude of potential transients is seen to increase in the presence of NO 2 − in Fig. 1 c comparing to that in Fig. 1 a. According to our previous work [20] , the amplitude of potential transients E amp = q pit / C dl , where q pit is the dissolution charge in a metastable pit, C dl the double layer capacitance of passive film. E amp is therefore inversely proportional to the value of C dl . Generally, the enhanced passivity of electrode by NO 2 − ions would decrease the C dl , leading to an increase of E amp , as shown in Fig. 1 c. This suggests that the intensity of current noises other than potential noises may be a suitable indication of the occurrence of pitting corrosion. From Fig. 1 a–c, the initiation rate (the number of current peaks per hour) is seen to increase with addition of TEPA or nitrite. However, this increase is not due to nitrite or TEPA accelerating the occurrence of pits, but due to the loss of cathodic protection effect of growing pits. In fact, once a metastable pit was re-passivated quickly, there would be no cathodic current from the growing pit to protect the rest surface of the electrode. Therefore, a new metastable pit initiates by the attack of Cl − ions. It is assumed that quick repassivation of metastable pits may result in the increase of nucleation rate in such case. 3.2 Mott–Schottky analysis Considering the applied potential range of M–S measurement should below the pitting potential ( E b ), a polarization curve is embedded in Fig. 2 to confirm the passive region of carbon steel in CPS. The decrease in the capacitance of passive film with the applied potential suggests that a depletion layer of donors is formed on the passive film/electrolyte interface with increasing potential. M–S plots for WEs, which were pre-passivated at various film formation potentials ( E ffp ), are shown in Fig. 3 . The positive slopes at applied potentials below the applied potential (0.45 V) indicate that the passive films are n-type semiconductors. In addition, when E ffp is above −0.8 V, two different M–S slopes in the applied potential range from −0.8 V to 0.45 V can be explained as the existence of donor states of different energy levels. The second slope may reveal the deep donor level of electrode being oxidized at higher anodic polarizations [32] . More positive slope in the passive film formed at higher E ffp suggests that the donor densities ( N d ) decrease with increasing film formation potential, which could be ascribed to the further oxidation of Fe 3 O 4 into γ-Fe 2 O 3 in the passive film at relatively high potential [13] . The slope becomes negative when the applied potential is higher than 0.45 V ( Fig. 3 ), indicating an initial breakdown of passive film due to the preponderance of iron ions vacancies over oxygen vacancies in the passive film [33] . It is generally accepted that the less the donor densities, the less susceptibility of the passive film to pitting corrosion [31] . According to the point defect model developed by Macdonald [34] , the n-type donors in the passive film are mainly oxygen vacancies at the metal/film interface, and the rupture of passive film can be ascribed to the anion-catalyzed cation (Fe 3+ ) vacancies concentrating on the metal/film interface which is responsible for the pitting events [35] . After the WEs were pre-passivated at various E ffp for 2 h in blank, 0.01 mol/L TEPA or 0.01 mol/L NaNO 2 containing CPS, M–S curves were measured and the N d versus E ffp plots are shown in Fig. 4 . It can be seen that N d decreases with increasing E ffp at different rates for all the three solutions. The higher values of N d at negative potential region for blank CPS indicate that the passive film formed in absence of inhibitors is less protective due to the fact that the inhibitors can prevent Cl − ions from undermining the passive film. It should be noted that the N d at the film formation potential of −0.8 V in the presence of NO 2 − ions is much lower than those in other two solutions, which is possibly ascribed to that the NO 2 − ions can repassivate the activated steel base (resulted from the dissolution of the passive film at −0.8 V) more quickly than TEPA and blank CPS. In addition, with the positive shift of E ffp , the N d values for all three solutions decrease with the growing passive films, and the difference of N d values also decrease with E ffp , as shown in Fig. 4 , which suggests that the higher film formation potential contributes to the passivity of steel much more than the inhibitors. In order to investigate the effect of Cl − ions on the donor density of the passive film, three WEs were pre-passivated at 0.2 V in blank CPS, 0.01 mol/L TEPA or 0.01 mol/L NaNO 2 containing CPS for 2 h, respectively. Then 0.01 mol/L NaCl was carefully added to these solutions prior to M–S measurements. The donor densities as function of time are shown in Fig. 5 . Evidently, the N d increases with time in CPS containing Cl − ions, indicating that Cl − ions can increase the defect (mainly oxygen vacancies) density in the passive film possibly by preventing the influx of oxygen anions, and result in a local breakdown of passive film (formation of metastable pits) by Cl − ions entering into the surface oxygen vacancies [36] . The increase of N d in nitrite containing CPS shows that the defect density of passive film also increases even in the presence of nitrite, which is probably attributed to the gradual consumption of nitrite via a reduction reaction. However, the value of N d becomes stable after WE being immersed in CPS containing TEPA for 12 h. It is assumed that TEPA adsorbed film may prevent the aggression of Cl − ions to the passive film for a longer time than nitrite due to no consumption of TEPA molecules with time. 3.3 In situ AFM measurements For better identification of the nucleation of pitting corrosion, in situ AFM imaging of carbon steel surface in CPS was conducted. Fig. 6 shows that the surface of carbon steel specimen seems to be rough (at nanometre scale) and is full of tiny spikes along the mechanical scratches. The roughness is found to be ∼10 nm, which could be the limiting value corresponding to the abrasive grain in the alumina paste. Fig. 6 b shows the surface view of the specimen after being immersed in CPS containing 0.01 mol/L Cl − ions for 5 h, from which several micro hemispheric pits with only 2–3 μm in diameter can be observed. The small metastable pits, which are generally considered to be initiated by Cl − ions and then passivated by OH − ions, lead to a cluster of small pits on the specimen, which is consistent with the SEM results in Wang's work [24] . Fig. 7 a shows the AFM images of carbon steel specimen immersed in CPS containing 0.01 mol/L Cl − ions and 0.01 mol/L TEPA for 5 h. It can be seen that the AFM morphology was obviously different from the image in Fig. 6 . A smooth-like image was observed on the surface of specimen due to the adsorption of TEPA. No apparent pits can be observed from the AFM image. However, the frequent noise transients in Fig. 1 b suggest that TEPA was unable to inhibit effectively the pitting nucleation. It is assumed that AFM cannot detect pitting corrosion under TEPA adsorbed film. On the other hand, after NO 2 − ions were added to the CPS containing 0.01 mol/L Cl − ions, the AFM image of Fig. 7 b shows that the pits become smaller (∼1 μm in diameter) comparing to those in Fig. 6 b, which is consistent with the fact that the decreasing lifespan of current transients reduces the pit size due to the rehabilitation of NO 2 − ions on the growing metastable pits, as shown in Fig. 1 c. However, the density of pits is higher comparing to that in Fig. 6 b, which is in accord with the increasing number of transients in EN curves. As mentioned above, this increase could be ascribed to the interrupt of cathodic current from the growing pits by the repassivation process. The depth profiles of metastable pits on the steel specimen are shown in Fig. 8 based on the AFM images. The dash lines in Fig. 6 and Fig. 7 indicate the scanning tracks. Comparing to the deepest pits (−100 nm) appearing on the specimen under CPS containing Cl − ions, only shallow pits (−30 nm) can be observed after nitrite was added to CPS. On the other hand, small frequent fluctuations (<10 nm) in blank CPS show no existence of apparent pits. After TEPA was added, the fluctuations become very smooth, indicating that adsorbed TEPA molecules possibly shield the original scratches on the surface of specimen [27] . After AFM analysis, the specimen was withdrawn from the 0.01 mol/L nitrite containing CPS for a further SEM observation. The morphology of the metastable pit is shown in Fig. 9 . It is seen that a pit of ∼2 μm in diameter is covered by a broken rust layer, which is slightly larger than the estimated value by AFM. Fig. 9 b shows another uncovered metastable pit (no cap) with a diameter of 5–6 μm and a smooth bottom, which may be the real size of metastable pits covered by rust layer. It is also assumed that the death of metastable pits is owing to the breakdown or removal of rust layer, which cause the concentration of aggressive anolyte inside metastable pits to fall below a level where repassivation can take place [37] . At a scan rate of 800 nm/s, the force–distance curves between the AFM tip and the steel specimen in CPS containing 0.01 mol/L Cl − ions with and without different additives are shown in Fig. 10 . The statistics of effective elastic constant k eff based on 25 F – D curves are shown in Fig. 11 . It shows that the force at the same distance is the largest in the solution containing NO 2 − ions, which is possibly resulted from the formation of dense passive film by the oxidization of nitrite. The force of TEPA adsorbed film is the lowest because of the flexibility of organic molecules chain. It shows that the order of average k eff of films is: oxidation film of nitrite > original passive film > adsorbed film of TEPA, which indicates that oxidation film is generally the most rigid one. 4 Discussion 4.1 Rehabilitation of nitrite on metastable pits The localised corrosion diagram of steel in carbonate and chloride based electrolytes in the presence of nitrite ions have been presented by Reffass et al. [12] . As an anodic inhibitor, nitrite can react with dissolved Fe 2+ ions though Eq. (5) [38] . (5) 2 F e 2 + + 2 O H − + 2 N O 2 − → 2 NO + γ − F e 2 O 3 + H 2 O (6) 2 N O 2 − + 4 e − ⇒ N 2 O + 3 O 2 − Fe (II) produced from the dissolution of passive film can be oxidized to Fe(III) by nitrite, resulting in rehabilitation of the passive film and a decrease of the donor density of the passive film comparing to the blank CPS, as shown in Fig. 4 . Moreover, according to reaction (6) [12] , the reduction process of NO 2 − ions can release O 2− ions. These O 2− ions other than Cl − ions may enter into passive film and fill the oxygen vacancies due to the large radius of Cl − ions [39] , leading to a decrease of donor density in passive film [13,16] . With increasing applied potential, more Fe (II) ions interact with NO 2 − ions and are oxidized to γ-Fe 2 O 3 and facilitate the rehabilitation of metastable pits. (7) F e 2 + + HC O 3 − ⇔ FeCO 3 + H + However, even the bulk electrolyte is buffered by the carbonate species in CPS (pH 12.0). The acidification of the medium inside pits via reaction (7) was confirmed indirectly by the micro-Raman analyses [12] , which leads to a local acidic condition in the pit and a consequent pitting event, the growth current density of pits was controlled by concentration diffusion [40] . Once the Cl − ions accommodating in pit cavity cannot satisfy the acidulation of electrolyte, the metastable pit will stop growing, and then the rehabilitation process of metastable pits dominates, as shown in Fig. 1 . When nitrite was added to the CPS solution, they migrate into the pit cavity, reacting with the dissolved Fe 2+ ions and promoting the repassivation of metastable pits, as indicated by the fast decreasing current transients (sharp spikes) in Fig. 1 c. In addition, as the nitrite was consumed gradually according to reaction (5) , the residual nitrite concentration would decrease, consequently leading to a decrease of O 2− ions concentration based on reaction (6) . Thus, the oxygen vacancies in passive film would increase with immersion time, as shown in Fig. 5 , which is also consistent with the degradation of long-term performance of nitrite in concrete by de Rincon et al. [41] . According to the Faraday law, the radius of semi-spherical metastable pit r pit can be calculated through the following Eq. (8) , and the charge q during metastable pit growth can be obtained by integration of a current transient [42] . (8) r pit = 3 q M 2 π z F ρ 3 where z is the valence ( z = 2 for ferrous ions), ρ is the density ( ρ = 7.8 g/cm 3 for carbon steel), F is the Faraday constant, M the atomic weight of Fe. In the presence of nitrite, the calculated average of r pit is ∼2.6 μm according to the average charge (∼1 μC) of current transients in Fig. 1 c, which is consistent with the value (2.5 μm) from SEM in Fig. 9 b. The profile of pit in Fig. 8 also shows that the average depth of pits in the presence of nitrite is lower than that in the absence of nitrite, indicating that the metastable pits initiated by Cl − ions can be re-passivated immediately by NO 2 − ions. This small-size (2.5 μm) metastable pit suggests that a metastable pit on carbon steel in CPS containing nitrite can hardly grow into a stable pit. However, a large pit radius (∼13.5 μm) is calculated from the charge (140 μC) of the largest current peak in Fig. 1 a, suggesting that carbon steel in CPS without nitrite is readily to develop into stable pitting corrosion. 4.2 Rehabilitation of TEPA on metastable pits The passive film of carbon steel is generally characterized as nanocrystalline composed of the outer primary γ-Fe 2 O 3 layer and the inner FeO/Fe 3 O 4 layer [43] . Oxygen vacancies are produced at the interface of metal and passive film because that the flux of Fe(II) cations is higher than the diffusivity of O 2− anions. A simple passive film model is schematically shown in Fig. 12 according to the PDM [36] , where anion vacancies and metal base constructs a space charge depletion layer (n-type semiconductor). Meanwhile, TEPA molecules adsorb on the passive film surface via the electronegative lone-pairs electrons of nitrogen atoms of amine groups, and form a close double layer (Helmholtz). Since the space charge region developed in the oxide film and the Helmholtz layer can be considered as two capacitors in series, the measured capacitance ( C m ) was determined by Eq. (9) to assess the semiconductive properties of the oxide films [44] . (9) 1 C m = 1 C SC + 1 C dl where C sc and C dl are the space charge and the Helmholtz capacitances of TEPA adsorbed layer, respectively. Generally, the space charge capacitance C sc in passive film is very small comparing to the double layer capacitance ( C dl ) because of the relatively thick space charge layer (nm range or larger) [45] . The contribution of C dl can be therefore neglected and C m is taken as C sc . However, when the passive film was covered by TEPA molecules, the C dl could be reduced much more comparing to that by water molecules due to the thick adsorbed layer and low permittivity of enamine molecules (from 80 μF/cm 2 to 20 μF/cm 2 ), resulting in an minor decrease of calculated C m and an additional decrease of the calculated donor density N d according to Eq. (1) . Fig. 4 shows that the N d values in the presence of TEPA are even lower than those in the presence of NO 2 − ions when E ffp is higher than −0.40 V. Meanwhile, Fig. 5 also shows lower N d values in CPS containing TEPA. It seems that TEPA can protect the passive film more effectively than nitrite. Nevertheless, the polarization curves in Fig. 13 indicate that the pitting potential in the presence of NO 2 − ions is more positive than that in the presence of TEPA at the same concentration, indicating that nitrite should be more effective to inhibit pitting corrosion than TEPA. Evidently, there is a big difference between the pitting susceptibility estimated by polarization curves and by M–S plots, which suggests that the N d calculated from the slope of C m −2 vs. applied potential may be misunderstood due to the non-negligible decrease of C dl by the adsorption of TEPA molecules on the passive film. TEPA adsorbed film can inhibit the regeneration of oxygen vacancies in passive film, resulting in an increase of pitting potential and rehabilitation of metastable pits, although this process is not as quick as nitrite ions. However, the image in Fig. 7 a looks smooth and no apparent pits are observed. In fact, this is just a pseudo image of TEPA film. It may be owing to that the AFM tip-followed viscous flow wake of TEPA solution obscures the details and/or that TEPA molecules adhering to the tip reduce the resolution of AFM imaging. In addition, TEPA adsorbed film is more flexible and mobile than nitrite oxidation film according to the k eff values in Fig. 10 , which suggests that TEPA is unable to impede the ingression of Cl − ions as effectively as nitrite. As a result, metastable pits would nucleate under TEPA film, as indicated by the current transients in Fig. 1 b. 5 Conclusions Cl − ions can promote the growth of metastable pits on carbon steel in synthetic carbonated concrete pore solution. On the other hand, nitrite and TEPA can accelerate the rehabilitation process of metastable pits, resulting in an increase of pitting corrosion resistance of carbon steel. Moreover, nitrite can rehabilitate the metastable pits more quickly than TEPA, leading to a shorter lifespan of metastable pits. Nitrite and TEPA can reduce the donor density in passive film, enhancing the stability of passive film in the presence of Cl − ions. However, the donor density calculated from M–S plots could be underestimated due to the decrease of Helmholtz capacitance from the adsorption of TEPA molecules on the passive film. AFM imaging can detect the in situ growing metastable pits on the surface of carbon steel in the presence of Cl − ions and NO 2 − ions. However, it fails to detect the metastable pits under TEPA adsorbed film. The force–distance curves between AFM tip and specimen show the stiffness (elasticity) order of films: nitrite oxidation film > original passive film > enamine adsorbed film. Acknowledgements The authors want to thank the National Natural Science Foundation of China for the financial supports to this work (grant no. 50971064 ). References [1] V.T. Ngala C.L. Page M.M. Page Corros. Sci. 44 2002 2073 [2] M.B. Valcarce M. Vazquez Mater. Chem. Phys. 115 2009 313 [3] E.E.A. El Aal S.A. El Wanees A. Diab S.M.A. El Haleem Corros. Sci. 51 2009 1611 [4] K.Y. Ann H.W. Song Corros. Sci. 49 2007 4113 [5] J.A. González E. Otero S. Feliu A. Bautista E. Ramirez P. Rodriguez W. Lopez Mag. Concr. Res. 50 1998 189 [6] G.K. Glass N.R. Buenfeld Corros. Sci. 39 1997 1001 [7] D.A. Hausmann Mater. Perform. 37 1998 64 [8] J.G. Cabrera A.S. Al-Hasan Constr. Build. Mater. 11 1997 283 [9] M. Ormellese L. Lazzari S. Goidanich G. Fumagalli A. Brenna Corros. Sci. 51 2009 2959 [10] N.S. Berke M.C. Hicks Cem. Concr. Compos. 26 2004 191 [11] M.A. Deyab S.S.A. El-Rehim Electrochim. Acta 53 2007 1754 [12] M. Reffass R. Sabot M. Jeannin C. Berziou P. Refait Electrochim. Acta 52 2007 7599 [13] S.J. Ahn H.S. Kwon Electrochim. Acta 49 2004 3347 [14] W.S. Li J.L. Luo Corros. Sci. 44 2002 1695 [15] M. Metikos-Hukovic Z. Grubac J. Phys. Chem. B 102 1998 7406 [16] L. Freire M.J. Carmezim M.G.S. Ferreira M.F. Montemor Electrochim. Acta 55 2010 6174 [17] Y.F. Cheng C. Yang J.L. Luo Thin Solid Films 416 2002 169 [18] R.A. Cottis Electrochim. Acta 52 2007 7585 [19] A. Legat M. Leban Z. Bajt Electrochim. Acta 49 2004 2741 [20] Z.H. Dong W. Shi X.P. Guo Corros. Sci. 53 2011 1322 [21] H.S. Klapper J. Goellner A. Heyn Corros. Sci. 52 2010 1362 [22] I.N. Bastos F. Huet R.P. Nogueira P. Rousseau J. Electrochem. Soc. 147 2000 671 [23] U. Bertocci F. Huet R.P. Nogueira P. Rousseau Corrosion 58 2002 337 [24] H. Wang J. Xie K.P. Yan M. Duan Y. Zuo Corros. Sci. 51 2009 181 [25] R.M. Pidaparti R.K. Patel Corros. Sci. 52 2010 3150 [26] F.A. Martin C. Bataillon J. Cousty Corros. Sci. 50 2008 84 [27] B. Cappella P. Baschieri M. Ruffa C. Ascoli A. Relini R. Rolandi Langmuir 15 1999 2152 [28] J.E. Qu X.P. Guo Z.Y. Chen Mater. Chem. Phys. 93 2005 388 [29] N.E. Hakiki M.F. Montemor M.G.S. Ferreira M.D. Belo Corros. Sci. 42 2000 687 [30] A.M.P. Simões M.G.S. Ferreira B. Rondot M. da Cunha Belo J. Electrochem. Soc. 137 1990 82 [31] Y.F. Cheng J.L. Luo Appl. Surf. Sci. 167 2000 113 [32] Y.F. Cheng J.L. Luo Electrochim. Acta 44 1999 2947 [33] M.J. Carmezim A.M. Simoes M.F. Montemor M.D. Cunha Belo Corros. Sci. 47 2005 581 [34] D.D. Macdonald J. Electrochem. Soc. 139 1992 3434 [35] Y.F. Cheng J.L. Luo Electrochim. Acta 44 1999 4795 [36] S. Ahn H. Kwon J. Electroanal. Chem. 579 2005 311 [37] H.S. Isaacs Corros. Sci. 29 1989 313 [38] T.A. Soylev M.G. Richardson Constr. Build. Mater. 22 2008 609 [39] S. Fajardo D.M. Bastidas M.P. Ryan M. Criado D.S. McPhail J.M. Bastidas Appl. Surf. Sci. 256 2010 6139 [40] Y.F. Cheng J.L. Luo M. Wilmott Electrochim. Acta 45 2000 1763 [41] O.T. de Rincon O. Perez E. Paredes Y. Caldera C. Urdaneta I. Sandoval Cem. Concr. Compos. 24 2002 79 [42] Z.H. Dong X.P. Guo J.X. Zheng L.M. Xu J. Appl. Electrochem. 32 2002 395 [43] M.F. Toney A.J. Davenport L.J. Oblonsky M.P. Ryan C.M. Vitus Phys. Rev. Lett. 79 1997 4282 [44] N.B. Hakiki S. Boudin B. Rondot M. Da Cunha Belo Corros. Sci. 37 1995 1809 [45] J. Wielant V. Goossens R. Hausbrand H. Terryn Electrochim. Acta 52 2007 7617