Study on passivation mechanism of HRB500E rebar in highly alkaline concrete pore solution

Abstract

This study comprehensively investigates the passivation mechanism of HRB500E rebar in highly alkaline concrete pore solution by examining phase transformation thermodynamics, microstructural characteristics, and electronic properties, combined with passivation experiments. The results demonstrated that the microcell models exhibited three distinct passivation behaviors in the initial stage of passivation. Firstly, the certain inclusions/second-phase precipitates with smallest work functions underwent the preferential dissolution, exhibiting the following pitting susceptibility tendency: TiVN₂ > Ca₈MgAl₆Si₅O₂₈ > NbVCN > MnS > TiN. Secondly, SiO₂, CaAl₂SiO₆ and CaMgSiO₄ with higher work function served as cathodes, while the surrounding matrix functioned as anodes and underwent the rapid dissolution. Subsequently, the surface of rebar formed the polygonal passivation products, and gradually overlapping and covering the inclusions. Finally, the lamellar cementite, exhibiting the smallest work function, underwent the anodic dissolution. These findings indicated that the full passivation on the surface of rebar was developed through matrix dissolution induced by inclusions/second-phase precipitates, ultimately resulting in the nano-particle products aggregated into a continuous passivation film.

Introduction

HRB500E Rebars are protected by both physical and chemical actions in concrete structures. On the one hand, the presence of a concrete protective layer effectively shields the rebar from direct attack by harmful substances. On the other hand, the alkaline substances such as dicalcium aluminate (2CaO·Al₂O₃) and tricalcium silicate (3CaO·SiO₂) in the cement components provide the abundant Ca(OH)₂ in the hydration process. Additionally, the cement component often contains some soluble strong alkalis (Na₂O and K₂O), raising the pH of the concrete pore solution to above 13¹. Some O₂ permeates into the concrete and dissolves in the pore solution. In a highly alkaline concrete environment, a nanometer-scale (< 10 nm) dense layer of iron oxide (also known as a passivation film) forms on the surface of rebar, maintaining a very low corrosion rate, thereby playing the role of protecting the rebar matrix². However, the passivation film can be dissolved and damaged by concrete carbonation during service process, and the attack of chloride and sulfate directly destroys the passivation film, resulting in the occurrence of pitting corrosion.

In the highly alkaline environment of concrete pore solution, the passivation film spontaneously formed on the surface of steel reinforcement, and serving as a natural corrosion barrier, and exhibiting the excellent corrosion resistance and rapid self-healing properties^{3,4,5,6,7,8,9,10}. Since the passivation film is very thin and the microstructure is difficult to be observed, many scholars have been studied the passivation film using many advanced characterization techniques. Loh et al. used in-situ electron microscopy to investigate the nucleation and growth process of nanocrystals in aqueous solutions, the results indicated that the metastable phases decomposed into solute-rich and solute-poor liquid phases, followed by the nucleation of amorphous metal nanoclusters within the solute-rich phase, eventually these amorphous clusters formed the crystals¹¹. Xu et al. explored the fine structure of the passivation film and demonstrated the crystal growth modes, which will help in assessing corrosion initiation from the perspective of microscopic damage to the passivation film¹². Some researchers utilized the Mott-Schottky curves to investigate the semiconductor properties of the passivation film, evaluating its conductivity and corrosion resistance based on carrier density^13,14,15,16. Numerous studies reported that the addition of Cr and Ni played a crucial role for corrosion resistance of steel, and the results illustrated that the formation rate of passivation film in solutions with lower alkalinity gradually slowed down, thereby leading to the enrichment of Cr and Ni within the passivation film, which enhanced its corrosion resistance^{4,5,7,17,18,19,20}. The growth and formation process of the passivation film on the surface of steel can be explained by a dissolution-precipitation reaction mechanism, indicating that the metals underwent the rapid anodic dissolution, and releasing metal cations into the electrolyte, resulting in the precipitation of oxide/hydroxide formed a passivation film²⁰.

Currently, most research focuses on exploring the macroscopic corrosion resistance of steel and the passivation and breakdown behavior^19,21,22. However, there is no accurate understanding about the origin of rebar passivation and the evolution of crystal passivation film with time on the atomic scale^7,23. Because the mechanisms underlying the formation of the passivation films on the surface of rebar in concrete are not yet completely understood, there has been widely accepted theoretical framework describing the growth and failure of the passivation film. This lack of a systematic understanding leads to a reliance on empirical judgment of the critical conditions for corrosion initiation¹². Regarding pitting corrosion process in steel, it is widely acknowledged that MnS is prone to induce the initiation of pitting corrosion^24,25,26,27. The corrosion problem between matrix, inclusions, and second phase precipitates in steels has been a more controversial issue^{20,28,29,30,31,32,33,34,35}. The corrosion sequence between them has not been a comprehensive and accurate understanding, and their influence on the passivation process requires the further investigation. It is particularly noteworthy that microalloying, as a crucial method to enhance the strength of rebar, holds the significant application potential for both current and future scenarios. However, research on the corrosion issues of high-strength rebar is relatively limited at present and requires the urgent attention. Existing studies mostly focus on the corrosion models of the conventional non-metallic inclusion, lacking systematic research on the influence of microalloying elements on the passivation process. It is important to note that different inclusions may lead to significant variations in the passivation and corrosion process.

In this study, the phase transformation thermodynamic properties of HRB500E rebar were calculated using JMatPro software. The microstructures and inclusion types were characterized by scanning electron microscopy (SEM) and energy dispersive spectrometry (EDS). The passivation process in a highly alkaline simulated concrete pore (SCP) solution was observed by scanning electron microscopy. The macroscopic passivation process was tested by electrochemical test, and the conductivity and work function of the contained phases were calculated by first-principles. By focusing on the transformation thermodynamics, microstructures and electronic properties of each phase in the rebar, combined with the experimental verification of passivation process, this study systematically investigated the passivation mechanism of HRB500E rebar in SCP solution. Therefore, this research will provide a theoretical basis for the design and prediction of the service life of reinforced concrete, offering reliable theoretical guidance for the development of corrosion-resistant rebar and corrosion protection engineering practices.

Materials and methods

Materials and solution

The experimental materials were mainly HRB500E rebars produced by a steel plant, with a diameter of 22 mm. The partial composition of HRB500E rebars was provided by the steel plant, and the oxygen and nitrogen contents were analyzed using an oxygen-nitrogen analyzer (HORIBA EMGA-830, Japan), as shown in Table 1. During the refining process, the calcium-magnesium treatment was employed to regulate the non-metallic inclusions in the rebar. The yield strength and tensile strength of HRB500E rebars was 545 MPa and 685 MPa, and the elongation was 22%. According to the China’s new national standard (GB/T 1499.2–2018), the carbon content should be less than 0.25 wt% to maintain the welding performance of the rebar. Additionally, a moderate amount of carbon was beneficial for controlling the microstructure of the rebar and obtaining a high-volume fraction of precipitates. The addition of Si and Mn in HRB500E rebars primarily played the role of solid solution strengthening, the microalloying elements (such as Cr, Nb, V, Ti, etc.) mainly played the role of precipitation strengthening and grain refinement.

Table 1 Chemical composition of HRB500E rebar (wt%).

The calculation formula of carbon equivalent (C_eq) was given in Eq. (1).

$$\:{C}_{eq}=C+Mn/6+(Cr+V+Mo)/5+(Cu+Ni)/5$$

(1)

To simulate the high alkaline concrete environment, SCP solution was prepared with a composition of saturated Ca(OH)₂ + 0.1 NaOH + 0.2 KOH (in mol·L^− 1), with pH of 13.2 and Oxidation-Reduction Potential (ORP) of -392 mV, this indicated that the SCP solution displayed a strong reducing environment, and the SCP solution maintained the low potential, so as to ensure that the SCP solution had the real environmental characteristics of concrete. To minimize the effect of carbonation from CO₂ in the air on the SCP solution, a small amount of Ca(OH)₂ powder was added to SCP solution. All reagents used in the experiment were analytical grade reagents from China Sinopharm Chemical Reagent Co., Ltd. Deionized water (resistivity > 18.2 MΩ·cm, TOC < 10 p.p.b.) used for rinsing and preparing electrolyte solutions was prepared by a laboratory ultrapure water system (HQ-CCS, Shanghai Huaqiao Environmental Protection Equipment Co., Ltd. China).

Sample preparations

To avoid the influence of surface oxide layers on the experiment process, the HRB500E rebars were cut using an electrical discharge numerical control wire cutting machine (DK7732, Nanjing Jinyuxiang Technology Co., Ltd., China) to obtain 10 mm×10 mm×10 mm specimens from near the core of the rebar. All the specimens were sequentially ground with SiC paper to 2000 mesh, followed by polishing with 1.5 and 0.5 μm diamond polishing compounds on the PI-1 A polishing machine, and then quickly blown dry with compressed air. Metallographic samples were etched using a 4% nitric acid alcohol solution. For electrochemical testing, 100 mm long nickel strips were spot-welded as electrode leads, encased in 5 mm diameter insulated polytetrafluoroethylene tubing, and then the sample connecting the nickel sheet was sealed with epoxy resin, reserving 1 cm2 of exposed working surface. After immersed for different times in SCP solution, the surface of sample rinsed with deionized water, and dried with compressed air.

Research methods

The prepared samples were immersed in SCP solution for 0 ~ 10 days (d), and taken out according to the time gradient for testing. The morphologies and element distributions of the immersed samples were characterized by scanning electron microscope (TESCAN MIRA LMS, Czech Republic), and the working acceleration voltage was 15 kV. Electrochemical measurements were conducted using an electrochemical workstation (CHI660D 412081, Shanghai Chenhua Instrument Co., Ltd., China), with a classic three-electrode system. The test sample served as the working electrode (WE), a platinum electrode served as the counter electrode (CE), and a saturated calomel electrode (SCE) with a salt bridge of the measured solution served as the reference electrode (RE). The testing time of open circuit potential (OCP) was no less than 1800 s, and all electrochemical tests in this study were conducted in a water bath environment at 25 °C, with testing initiated after the OCP potential stabilized. In the SCP solution, the cyclic voltammetry scans were performed in the potential range of -1.25 V to 0.63 V, scan rate was 0.025 V/s, starting from the open circuit potential and completing 5 cycles. Electrochemical impedance spectroscopy (EIS) was tested using an amplitude of ± 5 mV vs. OCP and a frequency range of 100 kHz to 10 mHz. The scanning potential range of dynamic potential polarization curve was − 0.5 V to 1.2 V vs. OCP, with a scanning rate of 0.5 mV/s. The test results were fitted by the electrochemical workstation software (CHI660D, Version: 12.26). All electrochemical tests have been repeated at least three times, and three sets of parallel electrochemical tests were performed on all electrochemical samples under the condition of each immersion test to ensure the accuracy of the electrochemical test results.

Phase equilibrium calculations were performed using the General Steel module of JMatPro software (Version: 7.0). Based on Density Functional Theory (DFT), First-principles calculations were conducted using the CASTEP module of MedeA software (Version: 3.6.3), employing the Perdew-Burke-Ernzerhof (PBE) scheme as an approximation for electronic exchange and correlation within the generalized gradient approximation (GGA), with an energy cutoff of 600 eV. The interaction between ion cores was described by projection augmented wave (PAW). The convergence accuracy of self-consistent field (SCF) calculation was less than 10− 5 eV. For the optimization of the crystal structure and parameters of each phase, the maximum stress on each atom was within 0.02 eV Å−1. For the calculation of the work function, the unit cell structure of each phase cut the different surfaces, and the 17 Å vacuum layer on the surface was separated by 1 Å atomic layer between each layer, mainly avoiding the interaction of valence electrons between the atomic layers. To compare the properties of rebar matrix and inclusions, α-Fe with body-centered cubic structure (Bcc) at room temperature was used as a benchmark for this study.

Results and analysis

Phase transformation thermodynamics

Based on the composition of HRB500E rebar, the calculated phase transformation thermodynamics were given in Fig. 1.

Fig. 1

Thermodynamic equilibrium diagram. (a) Phase equilibrium, (b) Ferrite, (c) Cementite, (d) M₂O₃, (e) MN and (f) M (C, N). Note: M is alloy element (such as Cr, Nb, V, Ti, etc.).

It can be seen from Fig. 1(a) that after completing the transformation of ferrite, the main phases in the HRB500E rebar were ferrite and cementite (Fe₃C), with Fe₃C existing as an intermediate phase. Small amounts of MnS, M₂O₃, MN, M (C, N), and SiO₂ (M represented the alloying elements) were present. Mullite underwent a phase transformation after cooling to 750 °C, but at high cooling rates, it might exist as an intermediate phase in the rebar. Cementite may be transformed into M₇C₃ carbide and solid solution alloying elements after prolonged cooling or aging treatment. Before completing the transformation of ferrite (830.24 °C), there was a significant rate of nitride precipitation, and the fastest precipitation rate of carbonitrides occurred between 750 ~ 920 °C.

It can be seen from Fig. 1(b) that the highest solid solutions (wt%) of alloying elements in ferrite were 1.155 Mn, 0.780 Si, 0.061 V and 0.065 Cr. Figure 1(c) indicates that in the cementite, near 460 °C, the solid solubility of Mn, Cr, and V was 27.96, 2.46, and 0.015, respectively, and Mn and Cr exhibited the highest solid solubility at this temperature, and the highest solid solubility of V was 0.33 at 712.29 °C. Figure 1(d) shows that the presence of titanium oxide at high temperatures served as a metastable phase, which rapidly transformed into other phases as the temperature decreases, and the main oxide in the rebar at low temperatures was Al₂O₃. Figure 1(e) reveals that the nitrides precipitated between 800 ~ 1200 °C were (Nb, V, Ti) N, with the highest precipitation rate. Below 800 °C, the primary nitride was (Ti, V) N, and the atomic ratio of Ti to V was approximately 4:1 when the temperature was lower than 700 °C. Figure 1(f) demonstrates that the element content curve of V and Nb intersected around 922 °C, where the precipitation of carbonitrides was Nb_0.5V_0.5C_0.8N_0.2, and the carbonitrides precipitated above 922 °C were Nb-rich, and vice versa for V-rich.

Microstructure and typical inclusions

The microstructures of HRB500E rebar were depicted in Fig. 2.

Fig. 2

Microstructure of HRB500E rebar. (a) Microstructure, (b) Pearlite morphology. Note: PF → Polygonal ferrite, LP → Lamellar pearlite, d_PF → Ferrite grain size, d_LP → Pearlite grain size, S₀ → Cementite lamellae spacing.

It can be seen from Fig. 2(a) that the microstructure of HRB500E rebar was primarily composed of ferrite (black) and pearlite (gray-white), and the ferrite was polygonal, the pearlite was lamellar. The distributions of microstructure were uniform and the grains were fine. Figures 2(a) and (b) shows that the average grain sizes of ferrite and pearlite were 6.845 and 10.277 μm, and the spacing of cementite lamellae was about 0.221 μm, and the proportions of ferrite and pearlite were 67.49% and 32.51%, and the measurement results were statistically determined by Nano Measurer (Laboratory of Surface Chemistry and Catalysis, Department of Chemistry, Fudan University, China, Jie. Xu, Version 1.2).

The typical inclusion morphologies and elemental compositions in the HRB500E rebar were given in Fig. 3. Figure 3(a) shows that there were two elliptical MnS inclusions on the surface of HRB500E rebar. Figure 3(b) shows that there was a polygonal calcium silicoaluminate (0.23Al₂O₃·0.39SiO₂·0.38CaO) on the surface of HRB500E rebar. Figure 3(c) shows that there was a Al₂O₃ with irregular shape on the surface of HRB500E rebar. Figure 3(d) shows that there was a circular inclusion composed of silicoaluminates (0.18Al₂O₃·0.32SiO₂·0.47CaO·0.03MgO) with some attached MnS on the surface of HRB500E rebar. Figure 3(e) shows that there was a slender inclusion consisted of central MnS and silicoaluminates (0.32Al₂O₃·0.13SiO₂·0.42CaO·0.13MgO) at both ends on the surface of HRB500E rebar. Figure 3(f) shows that there was a square inclusion with (Nb, V, Ti) N and MnS wrapped around Al₂O₃ on the surface of HRB500E rebar. Figures 3(b) and (d)~(e) shows that the inclusions were modified during the HRB500E rebar refining process with calcium-magnesium treatment, and the types of inclusion in the HRB500E rebar were mainly micron-sized inclusions and some nano-sized inclusions.

Fig. 3

Morphology and elemental composition of typical inclusion. (a) MnS, (b) calcium silicoaluminate, (c) Al₂O₃, (d) and (e) silicoaluminate and MnS composite inclusions, (f) Composite inclusions of (Nb, V, Ti) N and MnS wrapped around Al₂O₃.

Current density response of passivation process

The cyclic voltammetry (CV) curves of HRB500E rebar in the SCP solution with pH13.2 were given in Fig. 4.

Fig. 4

Cyclic voltammetry curve of HRB500E rebar in the SCP solution with pH13.2. (a) 1 Cycles. (b) 5 Cycles.

It can be seen from Figs. 4(a) and (b) that the peak shapes of the CV curve were basically stable. However, as the number of cycles increased, the intensity of all peaks increased, and the intensity gap between the oxidation peak and the reduction peak obtained by scanning in the anode and cathode directions displayed larger, and the intensity of the reduction peak obtained by scanning in the cathode direction was significantly stronger than that of the oxidation peak. This may be due to the incomplete reduction of the reduction peak during the passivation process of the rebar surface in the cathode each scanning direction, which led to the increase of the resistance of the passivation film layer, thereby increasing the strength of the reduction peak. The above results indicated that the passivation and passivation breaking of HRB500E rebar in SCP solution was reproducible but not fully reversible. There was one oxidation peak (A₁) and one reduction peak (C₁) in the CV curve. At peaks A₁ and C₁, the current density increased significantly with the number of cycles. This indicated that the reaction process of the reaction peak was gradually strengthened, which demonstrated a strong capability of HRB500E rebar to self-repassivate.

The first oxidation peak A₁ appeared around − 0.45 V, this was attributed to the formation of Fe²⁺ (Fe(OH)₂), the transformation of Fe²⁺ (Fe(OH)₂) to Fe³⁺ (Fe₃O₄ and γ-Fe₂O₃) and the formation of FeOOH (γ-FeOOH)^{20,36,37,38,39}, and the reaction process corresponding to the A₁ peak was as follows.

$$\:{\text{Fe }} + {\text{ 2OH}}^{ - } \to {\text{Fe}}\left( {{\text{OH}}} \right)_{{\text{2}}} + {\text{ 2e}}^{ - }$$

(2)

$${\text{3Fe}}\left( {{\text{OH}}} \right)_{{\text{2}}} + {\text{ 2OH}}^{ - } \to {\text{ Fe}}_{{\text{3}}} {\text{O}}_{{\text{4}}} + {\text{ 4H}}_{{\text{2}}} {\text{O }} + {\text{ 2e}}^{ - }$$

(3)

$${\text{Fe}}_{{\text{3}}} {\text{O}}_{{\text{4}}} + {\text{ OH}}^{ - } + {\text{ H}}_{{\text{2}}} {\text{O }} \to {\text{ 3}}\gamma - {\text{FeOOH }} + {\text{ e}}^{ - }$$

(4)

$${\text{2Fe}}_{{\text{3}}} {\text{O}}_{{\text{4}}} + {\text{ 2OH}}^{ - } + {\text{ 2H}}_{{\text{2}}} {\text{O }} \to {\text{ 3}}\gamma - {\text{Fe}}_{{\text{2}}} {\text{O}}_{{\text{3}}} + {\text{ 3H}}_{{\text{2}}} {\text{O }} + {\text{ 2e}}^{ - }$$

(5)

With the accumulation of Fe₂O₃, Fe₃O₄ and FeOOH, the passivation film was gradually formed, resulting to a continuous decrease in anodic current density. When further scanning in the anodic direction, a current density plateau (passivation region) with very low and stable current density appeared on the cyclic voltammetry curve. In the passivation region, i_flat was an important indicator of the corrosion resistance of the passivation film, reflecting the resistance of the passivation film to corrosion reactions. Generally, a smaller i_flat value indicated a greater resistance of the passivation film to corrosion reactions, indicating a stronger protective effect of the passivation film^36,40,41. The i_flat value was 0.6346 × 10^− 5 A·cm^− 2, indicating that the formed passivation film on the HRB500E rebar surface exhibited the excellent corrosion resistance. When the scanning potential reached the transpassive region (greater than E_transpassive), the anodic current density sharply increased, and showing a high peak, indicating the breakdown of the passivation film on the surface of HRB500E rebar.

During scanning the cathodic direction, the reduction peak C₁ corresponded to the oxidation peak A₁, and the reduction reaction from Fe³⁺ oxides to Fe²⁺ oxides occurred, as shown in Eq. (6)^20,38. Additionally, the reduction reactions in the cathodic scanning process (see Eqs. (7) and (8)) may also occurred at peak C₁^36,37.

$${\text{4FeOOH }} + {\text{ 2H}}_{{\text{2}}} {\text{O }} \to {\text{ 4Fe}}\left( {{\text{OH}}} \right)_{{\text{2}}} + {\text{ O}}_{{\text{2}}}$$

(6)

$$~{\text{Fe}}_{{\text{3}}} {\text{O}}_{{\text{4}}} + {\text{ 4H}}_{{\text{2}}} {\text{O }} + {\text{ 2e}}^{ - } \to {\text{ 3Fe}}\left( {{\text{OH}}} \right)_{{\text{2}}} + {\text{ 2OH}}^{ - }$$

(7)

$${\text{Fe}}_{{\text{3}}} {\text{O}}_{{\text{4}}} + {\text{ H}}_{{\text{2}}} {\text{O }} + {\text{ 2e}}^{ - } \to {\text{3FeO }} + {\text{ 2OH}}^{ - }$$

(8)

Effect of inclusions on passivation process

The dissolution processes of MnS and its composite inclusions in the SCP solution with pH13.2 were given in Fig. 5.

Fig. 5

Morphology and elemental composition of MnS and its composite inclusions after immersing for different time. (a) and (b) 1 h. (c) 3 h. (d) 12 h.

It can be seen from Fig. 5(a) that after immersing for 1 h, MnS was preferentially dissolved, and a small portion of undissolved MnS was still presented in the corrosion pits. There were no significant signs of dissolution on the rebar matrix, and the corrosion pits maintained the intact morphology. Figures 5(b)~(d) shows that the morphologies and elemental compositions of MnS composite inclusions after immersing different times, these composite inclusions were elongated MnS with Nb compounds attached at both ends, calcium silicoaluminate partially enveloping MnS, and elongated calcium silicoaluminate partially enveloping two MnS, respectively. The MnS portion of this composite inclusions was preferentially dissolved and the corrosion pits were formed, but no obvious traces of dissolution were found on the rebar matrix. The dissolution rate of MnS in composite inclusions existed the difference, this may be related to the size of MnS and the type of composite inclusions, which will be discussed in the first-principles calculations section. Notably, in Fig. 5(d), the polygonal iron oxides were found on the middle part of the calcium silicoaluminate after immersing for 12 h. This was due to the fact that most of MnS exposed at the ends have been dissolved and a corrosion microcell was formed between the middle portion of calcium silicoaluminate and the rebar matrix^41,42. When the anodic dissolution occurred on the rebar matrix, the producing iron oxides were gradually covered on the surface of rebar in a highly alkaline environment.

Figure 6 displays the process that a kind of inclusions were gradually covered by polygonal passivation layers.

Fig. 6

Formation process of polygonal passivation layers. (a) 1 h. (b) 3 h. (c) 12 h.

The morphology and elemental composition of calcium silicoaluminate inclusions in Fig. 6(a) shows that after immersing for 1 h, the calcium silicoaluminate and the surrounding area was covered by granular calcium-rich precipitates. Simultaneously, around the calcium silicoaluminate, the polygonal passivation layers were covered on the rebar matrix and the calcium-rich precipitates. This was because that the calcium silicoaluminate and the rebar matrix formed the corrosion microcells, where the rebar matrix rapidly underwent the anodic dissolution, releasing a large amount of iron ions that combined with hydroxide ions, ultimately forming a passivation layer on the periphery of corrosion microcells. Meanwhile, the calcium ions in the solution aggregated at the inclusions and reacted to form the calcium precipitates. After immersing for 3 h (see Fig. 6(b)), the inclusions and calcium-rich precipitates were gradually covered by polygonal passivation layers, and the elemental compositions of inclusion were not clearly detected by EDS, indicating that the thickness of the passivation layer reached the micron level and was not penetrated by high-energy electrons. After immersing for 12 h (see Fig. 6(c)), the multiple layers of the polygonal passivation layers were formed, and some granular or floral calcium-rich precipitates still existed on the surface of rebar. According to the EDS in Fig. 6(c), the polygonal passivation layers were primarily consisted of iron oxides and calcium carbonate. Since the EDS data did not focus on the hydrogen elements, the iron hydroxides might also be present.

Formation process of passivation films

After passivation in the SCP solution, the passivation films on the surface of rebar were given in Fig. 7. The surface morphology in Figs. 7(a) and (b) displays that after passivation for 2 d, the polygonal passivation layers were gradually formed on the surface of rebar, and there were patches of microgaps in some areas. Combined with the microstructure of the rebar matrix in Fig. 2, this closely resembled the microstructure of pearlite. After passivation for 3 d (see Fig. 7(c)), the black areas represented the iron oxides, and the gray areas depicted the exposed rebar matrix. Figure 7(d) illustrates that the granular iron oxides were continuously covered on the surface of rebar, and gradually forming a localized passivation film. After passivation for 7 d (see Figs. 7(e) and (f)), the surface of rebar matrix was mainly consisted of two morphologies. Firstly, the microcrack region, namely the pearlite area and the ferrite area, exhibited the relative smooth. Due to the thinness of the passivation film, only a few nanometers thick, the observation of the passivation film on the surface of rebar matrix under a scanning electron microscope resembled a semi-transparent film. Secondly, due to the preferential dissolution of inclusions such as MnS on the surface of rebar, some corrosion pits formed on the surface of rebar, and the passivation film covering the surface of the corrosion pits increased with the passivation time (see Fig. 7(f)). Statistical analysis of microgaps revealed that the average gap spacing was 0.158 μm and the average gap width was 0.037 μm. These parameters were consistent with the lamellar cementite in the rebar. This confirmed that after passivation in SCP solution, the dissolution of cementite leaved the large areas of microgaps.

Fig. 7

Surface morphology of the passivation film of the rebar matrix. (a) and (b) 2 d, (c) and (d) 3 d, (e) and (f) 7 d.

Macroscopic electrochemical of passivation process

Electrochemical impedance spectroscopy (EIS) can analyze the controlling factors and mechanisms of the electrochemical reaction process. Capacitance arc radius and impedance modulus (|Z|) can reflect the resistance of electrochemical reaction, and the larger the capacitance arc radius and impedance modulus, the better the corrosion resistance⁴³. When the surface of the rebar was smooth, the absolute value of the maximum phase angle approached about 90°. A smaller absolute value of the maximum phase angle indicated a rougher surface film and poorer protective properties⁴⁴.

EIS results and the equivalent circuits were given in Fig. 8. It can be seen from the Nyquist plot in Fig. 8(a) that as the immersion time increased, the radius of capacitive arc first increased sharply and then reached the relative stability. It can be seen from the Bode plot in Fig. 8(b) that the |Z| value showed an increasing trend with the increases of immersion time. It can be seen from the Phase plot in Fig. 8(c) that the maximum phase angle gradually increased with the increases of immersion time, and the maximum phase angle stabilized around 78.5°. Overall, the EIS results indicated that the passivation process of HRB500E rebar tended to be stable and the surface roughness of rebar changed little with the passivation process, which confirmed that the formation process of the passivation film in Fig. 7.

Fig. 8

EIS results of passivation process. (a) Nyquist plot. (b) Bode plot. (c) Phase plot. (d) Equivalent circuit.

It can be seen from Figs. 8(a)~(c) that the capacitive arc of the surface passivation process of HRB500E rebar showed a large semi-circular arc, the impedance film value was large, and the maximum phase angle showed a certain width, this indicated that the passivation film formed on the surface of HRB500E rebar exhibited the double-layer capacitance characteristics, the film resistance and charge transfer resistance was large, thus the equivalent circuit in Fig. 8(d) was selected to fit the EIS results. Based on the passivation process of continuous local reaction, the local corrosion equivalent circuit in Fig. 8(d) was used to fit the EIS results in Figs. 8(a)~(c), and the fitting results were presented in Table 2, where R₁ represents the resistance of SCP solution; CPE₁ represents a constant phase angular capacitance element of the passivation film; R₂ represents the resistance of the passivation film (the electron transfer that hinders the electrode reaction); R₃ represents the ion charge transfer resistance of local corrosion reaction, which reflects the difficulty of electron transfer in anodic dissolution; CPE₂ represents a constant phase angular capacitance element of the electric double layer at the rebar-solution interface. Considering the unevenness of the surface film, its capacitance was not ideal, so the capacitance value of CPE can be characterized by basic admittance (Z₀) and dispersion coefficient (n). The closer n was to 1, the closer its system was to ideal capacitance. Therefore, CPE was defined as follows^43,44.

$$\:{Z}_{CPE}=\frac{1}{{Z}_{0}{\left(jw\right)}^{n}}$$

(9)

Where, Z₀ is the basic admittance, n is the dispersion coefficient, j is the imaginary unit, w is the angular frequency. In order to further analyze the capacitance behavior of the surface passivation process of HRB500E rebar, the effective capacitance (C₁) of the passivation film resistance and the effective capacitance (C₂) of the charge transfer resistance were analyzed by Brug formula and the Hsu formula and the Mansfeld formula. The calculation formulas of effective capacitances C₁ and C₂ were given as follows^43,44.

$$\:{C}_{1}={CPE}_{1}^{\frac{1}{{n}_{1}}}{R}_{2}^{\frac{1-{n}_{1}}{{n}_{1}}}$$

(10)

$$\:{C}_{2}={CPE}_{2}^{\frac{1}{{n}_{2}}}{(\frac{1}{{R}_{1}}+\frac{1}{{R}_{3}})}^{\frac{{n}_{2}-1}{{n}_{2}}}$$

(11)

It can be seen from Table 2 that the R₂ and R₃ values generally increased with passivation time, R₂ value remained above 10⁴ Ω·cm², and R₃ value remained above 10⁵ Ω·cm². At different passivation time, the R₃ value was much larger than the R₂ value, indicating that the resistance of the passivation film on the surface of rebar was very large, which inhibited the charge transport of the film layer, thereby the passivation film formed on the surface of rebar effectively impeded the progress of electrochemical corrosion reactions. However, the basic admittance of CPE₂ and CPE₁ and the effective capacitance values C₁ and C₂ were below the order of 10^− 5 Ω⁻¹·cm^− 2·sⁿ, indicating that the different areas on the surface of the rebar reached a stable state, and forming relatively the complete passivation films on the rebar surface, thereby providing a good protection for the rebar.

Table 2 EIS fitting results of passivation process.

The dynamic potential polarization curve curves and the corresponding fitting parameters were given in Fig. 9.

Fig. 9

Dynamic potential polarization curve of passivation process. (a) Polarization curve. (b) Fitting parameters.

It can be seen from Fig. 9(a) that with increases of immersion time, the corrosion potential moved down (more negative), but there was a significant increase at 10 d. Above the self-corrosion potential (E_corr), a passive region was present, indicating that the passivation process occurred at all times. After passivation for 1 d, there were two segments in the cathodic polarization curve, this may be related to the localized reactivation. It can be seen from Fig. 9(b) that the E_corr value, self-corrosion current (i_corr), polarization resistance (R_p) and corrosion rate all decreased sharply before increasing with increases of immersion time. The corrosion rate was the highest at 6 d, and reaching 0.421 mm·year^− 1. Notably, after 10 d of passivation, the corrosion rate was low and the corrosion resistance was excellent, thus indicating that the surface of rebar formed a complete passivation film.

First-principles calculations

Based on experimental results, the crystal cell structures of the rebar matrix, solid solution elements and possible typical inclusions/secondary phase precipitates were established, and the construction processes were described as follows: (1) The cell parameters (a, b, c, α, β, γ) and space group codes of each phase were found through the MedeA software database; (2) The crystal cell was defined by the Structure Builder module, and the crystal cell parameters and space group codes of each phase were input at the corresponding lattice constant and space group position, and then adding the atomic position, thus establishing the crystal cell structure of each phase; (3) The established phase unit cell was optimized by CASTEP module, and the geometric optimization was selected for the calculation task, and the optimization functions were GGA and PBE, and then running the CASTEP calculation. Finally, the optimized unit cell structures and parameters of each phase were presented in Fig. 10; Table 3, respectively.

Fig. 10

Crystal cell structures of the rebar matrix, solid solution elements, and possible typical inclusions/secondary phase precipitates. (a) α-Fe. (b) Fe₃C. (c) MnS. (d) SiO₂. (e) Al₂O₃. (f) Mullite. (g) MgAl₂O₄. (h) CaMgSiO₄. (i) CaAl₂SiO₆. (j) Ca₈MgAl₆Si₅O₂₈. (k) TiVN₂. (l) NbVCN. (n) TiN. (m) V. (o) Mn. (p) Si. (q) Cr.

Table 3 Optimized crystal structure parameters.

The galvanic couple theory was commonly used to explain the mechanism of localized corrosion induced by inclusions, and two essential conditions was conductivity and potential difference. Traditionally, the mechanism of galvanic pitting corrosion of inclusions in steel often depended on the theory of potential difference^33,45,46. According to the galvanic couple theory, insulating inclusions did not form an electronegative coupling with the surrounding steel matrix. Therefore, the conductivity of inclusions was a critical property affecting pitting corrosion. The band structure can reflect the conductivity of inclusions, with a band gap ≥ 5.0 eV, indicating the non-conductive insulators^30,31. The calculation process of the band structure was described as follows: (1) Based on the optimized cell structure and parameters of each phase in Fig. 10; Table 3, the band structure was calculated by the CASTEP module in MedeA software, and the calculation functions were GGA and PBE; (2) The band gap of the band structure of each phase was read by the analysis tool of the CASTEP module, and the results were given in Table 4.

Table 4 Band gaps of band structure.

It can be seen from Table 4 that the band gaps of Al₂O₃, mullite, and magnesium aluminate spinel (MgAl₂O₄) were all ≥ 5.0 eV, indicating that these inclusions were non-conductive insulators. Fe₃C, MnS, TiN, NbVN₂, NbVCN, and other solid solution elements exhibited good conductivity, thus providing the electronic conduction pathways. The band gaps of CaAl₂SiO₆, CaMgSiO₄, and Ca₈MgAl₆Si₅O₂₈ modified by calcium-magnesium treatment ranged from 3.29 to 4.64 eV, thus showing the semiconductor characteristics.

The electron work function was the minimum energy required for an electron to escape immediately from the surface of a material, also known as the minimum work required for an electron to escape from the Fermi level to the vacuum level⁴⁷. The relevant literature reported that the electronic structure and surface characteristics of inclusions in steel were calculated by density functional theory (DFT), and the Fermi level and vacuum level of inclusions were determined^31,34,35. Therefore, the work function (W_F) of inclusions was calculated by Eq. (12).

$$\:{W}_{F}={E}_{vac}-{E}_{Fermi}$$

(12)

Where, E_vac represents the vacuum energy level and E_Fermi represents the Fermi energy level.

The work function mainly reflected the inherent corrosion tendency of the inclusions in steel. Generally, the lower the surface energy of a Miller index surface, the more likely it is to be an exposed surface, making it more susceptible to participating in electrochemical corrosion processes^31,35. Some studies suggested that the surface energy of the α-Fe(100) was the lowest^33,35,48. However, other research indicated that the surface energies of low-index α-Fe surfaces were compared as follows: (111) >(100) >(110), and the easily exposed surfaces were (110) and (100)^34,49,50,51. In this study, the moderately exposed α-Fe(100) surface with intermediate surface energy was used as a reference state to compare the work functions between the steel matrix and other phases, and the calculated work functions of all phases in the rebar were given in Fig. 11.

Fig. 11

Work functions of each phase at three different low-indexed surfaces. (a) Fe matrix and inclusions. (b) Elementary substances.

It can be seen from Fig. 11 that the inclusions with band gaps ≥ 5.0 eV in Table 4, except for the (100) and (110) surfaces of MgAl₂O₄, the work functions of the others were all greater than 5.19, making them less likely to participate in galvanic corrosion and less prone to self-dissolution. The experimental value for the work function of α-Fe was approximately 4.8^51,52. It was observed that the work function of the α-Fe(111) surface was much lower than the experimental value. While this discrepancy has been reported in various literature^{30,31,35,50,53}, the impact on this study was minimal as the most likely exposed surfaces of α-Fe were (110) and (100) based on the surface energy of α-Fe. The average work function order of each phase was as follows: SiO₂ > CaAl₂SiO₆ > CaMgSiO₄ > Fe₃C > α-Fe > MnS > NbVCN > Ca₈MgAl₆Si₅O₂₈ > TiVN₂ > TiN. The average work functions of the solid solution elements in the ferrite and cementite phases were compared as follows: Si > α-Fe > Cr > Mn > V, which accorded with the order of metal activity.

The work functions of Fe₃C and MnS were around α-Fe(100), and there were overlaps which need further discussion. In the microcell system within the pearlite, the surface energies of Fe₃C were ranked as (100) >(111) >(110)⁵⁴. Comparing the two easily exposed low-index crystal planes, the average work function of α-Fe was 0.24 eV higher than that of Fe₃C. The surface energy of MnS was reported as (111) >(110) >(100)⁴⁹. Comparing the average work functions of the two exposed low-index crystal planes, MnS was 0.29 eV lower than α-Fe.

The work functions of NbVCN, Ca₈MgAl₆Si₅O₂₈, TiVN₂, and TiN were significantly lower than α-Fe. For the TiN inclusions, the surface energy order was (111) >(110) >(100), and TiN(100) and (110) were easily exposed surfaces^51,53. Therefore, the work function of TiN was much lower than that of α-Fe. Moreover, in the process of ferrite nucleation induced by TiN, Fe atoms were more likely to gather on the surface of TiN(100), and the α-Fe(100)/TiN(100) interface was the most stable⁵¹. However, due to the thermodynamic instability of TiN, it was easy to form a layer of TiO₂ film on the surface of TiN, which displayed the cathode characteristics⁵³. The work functions of SiO₂, Mullite, CaAl₂SiO₆ and CaMgSiO₄ were much larger than that of α-Fe, and they could be used as cathodes in microcells, which greatly accelerated the anodic dissolution rate of rebar matrix. This phenomenon was also confirmed by the experimental results in Fig. 6.

In summary, the tendency of anodic dissolution of each phase was TiVN₂ > Ca₈MgAl₆Si₅O₂₈ > NbVCN > MnS > Fe₃C > TiN. For the phase that participated in galvanic corrosion served as cathode, the tendency to induce matrix dissolution was SiO₂ > CaAl₂SiO₆ > CaMgSiO₄. Al₂O₃, Mullite and MgAl₂O₄ were non-conductive insulators and did not participate in galvanic corrosion.

Discussions

The microstructure of HRB500E rebar was primarily composed of ferrite and pearlite, and pearlite was a mechanical product of α-Fe and cementite. Some elements of Mn, Si, V and Cr were solid-solution in ferrite, and a large amount of Mn and a small amount of Cr and V were solid-solution in cementite. According to the order of work function and metal activity, Mn, Cr, and V were more reactive than α-Fe. During the passivation process, these solid solution metal elements can act as anodes and induce the initiation of dissolution, especially in the cementite region with a significant amount of Mn. In contrast, Si has a high work function and can accumulate in the passivation film during the passivation process. This may make the rebar surface uneven, resulting in a rougher passivation film. The typical inclusions/secondary phase precipitates in the HRB500E rebar mainly included MnS, Al₂O₃, SiO₂, Mullite, MgAl₂O₄, CaMgSiO₄, CaAl₂SiO₆, Ca₈MgAl₆Si₅O₂₈, and (Nb, V, Ti) N.

Based on the experimental results of the passivation process from Figs. 5, 6 and 7, and the first-principles calculation results in Table 4; Fig. 11, three main microcell models were identified during the initial stages of passivation. First, the typical inclusions with lower work functions such as TiVN₂, Ca₈MgAl₆Si₅O₂₈, NbVCN, and MnS underwent the anodic dissolution, and generating the corrosion pit in the inclusion area, thereby inhibiting the dissolution of the matrix. Second, the inclusions with higher work functions such as SiO₂, CaAl₂SiO₆, and CaMgSiO₄ led to the rapid dissolution of the surrounding rebar matrix, a polygonal passivation layer began to be formed at the periphery of the inclusion, and the inclusion was covered by passivation layer. Third, the cementite lamellae and the surrounding ferrite formed a microcell and the cementite underwent the anodic dissolution, resulting in the granular passivation products attached on the surface of rebar. Additionally, after these reactions, the exposed rebar matrix underwent the comprehensive passivation, releasing a large amount of iron ions into the solution, forming a passivation film with continuous nanoscale granular products in the highly alkaline environment. The macroscopic electrochemical results verified this process, indicating the presence of stable and reactivation stages during the passivation process of the rebar matrix. Based on the research findings, the passivation process mechanisms were illustrated in Fig. 12.

Fig. 12

Passivation mechanism. (a) and (b) Inclusions (SiO₂, CaMgSiO₄, CaAl₂SiO₆) microcell model. (c) Cementite microcell model. (d) Comprehensive passivation model.

Figures 12(a) and (b) depicted a formation process of microcells. In the microcell, the typical inclusions with high work functions such as SiO₂, CaMgSiO₄, and CaAl₂SiO₆ acted as the cathode phase, while the rebar matrix played the role of the anode phase, and the two constituted the micro-corrosion, resulting in the anodic dissolution of the rebar matrix, which further developed into a microcell. With the transfer of electrons within the matrix, there was ion transfer in the solution, and forming a microcell circuit. In the cathode region, due to the oxygen reduction reaction (O₂ + 2H₂O + 4e⁻→4OH⁻) and the hydrogen evolution reaction (2 H⁺+2e⁻→H₂), the OH⁻ ion generated by the cathode reaction will rapidly increase the local negative charge density. In order to maintain the electrical neutrality of the solution, Ca²⁺ and H⁺ in the SCP solution will migrate to the cathode region and accumulate under the action of the electric field. At this time, H⁺ will continue to consume and generate H₂, resulting in an increase in the relative concentration of Ca²⁺, thus Ca²⁺ and dissolved CO₂ formed the insoluble CaCO₃ (Ca²⁺+CO_2(diss)→CaCO₃↓). Due to the floating of H₂ during this process, the formation of CaCO₃ as a large solid precipitate was hindered, resulting in small spherical deposits distributed around the inclusions. In the anodic region, due to the electron transfer within the rebar matrix, the rebar matrix lost the electrons and underwent an oxidation reaction to form Fe²⁺ (Fe→Fe²⁺+2e⁻), and OH⁻ and Fe²⁺ in the SCP solution formed the unstable Fe(OH)₂ (Fe²⁺+2OH⁻→Fe(OH)₂). Under the action of dissolved oxygen, Fe(OH)₂ was further oxidized to the stable Fe(OH)₃ (4Fe(OH)₂+O₂+2H₂O→4Fe(OH)₃), resulting in the deposition of these products around the anode region. It was noteworthy that the formation of these polygonal products was related to their crystallographic structure. Figure 12(c) illustrated a schematic of the microcell model formed by the lamellar cementite and the ferrite within the pearlite. After the reaction of the microcell model composed of inclusions, a large number of microcells were formed by the lamellar cementite and the ferrite. The dissolution of the cementite released some iron ions, and combining with OH⁻ to form precipitates of Fe₃O₄, FeOOH, and γ-Fe₂O₃, thus presenting the nanoscale granular passivation products. Figure 12(d) displayed the mechanism of comprehensive passivation, in which iron ions within the rebar matrix were continuously released to generate the granular products, which were finally connected together to form a passivation film.

Currently, there were several contentious issues in the research. Firstly, the prevailing view suggested that MnS was the source of pitting corrosion in the steel matrix, but TiVN₂, Ca₈MgAl₆Si₅O₂₈, NbVCN, and others were more susceptible to pitting corrosion than MnS. This was due to the stable interface structure formed by the orientation relationship of α-Fe(110)/MnS(110)⁴⁹, and the fact that MnS in the steel often served as micrometer-sized inclusions or composite inclusions. Secondly, regarding the cementite, many studies^41,55,56,57 suggested that the potential of cementite was higher than that of ferrite, making it more corrosion-resistant, which contradicted the findings of this study. This discrepancy may be due to the different forms of cementite as intermediate phase, as well as the influence of different solid-solution elements on its overall potential. There were two limitations in this study. Firstly, the crystallographic structures of some inclusions were extremely complex, and there was a lack of research on their crystallographic structures. As a result, the computational results from different models may deviate from the actual inclusions in the steel. Secondly, there was insufficient evidence to determine whether the comprehensive passivation was a result of chemical dissolution or electrochemical action, and there was a lack of evidence to prove whether the generated granular passivation products had an impact on electrochemical action.

Conclusions

In this study, thermodynamic calculation, scanning electron microscope, electrochemical test and first-principles calculation were used to study the passivation process of HRB500E rebar in high alkaline concrete pore solution. The main conclusions were summarized as follows.

(1) At the early stage of passivation, there were mainly three kinds of passivation behaviors of microcell models. Firstly, the inclusions/second phase precipitates such as TiN, MnS, NbVCN, Ca₈MgAl₆Si₅O₂₈ and TiVN₂ in the rebar were preferentially dissolved. Secondly, SiO₂, CaAl₂SiO₆ and CaMgSiO₄ as cathodes participated in galvanic corrosion, and the surrounding ferrite was dissolved, and the polygonal passivation products were generated to cover the inclusions. Finally, the lamellar cementite as anode, exhibiting the smallest work function, was dissolved, and some particle passivation products were generated on the surface of rebar. After the formation of microcells induced by inclusions/second phases, the polygonal products containing iron compounds were formed on the surface of rebar. With the increases of passivation time, this product gradually aggregated to form a nano-particle passivation products, thereby a layer of passivation film was covered on the surface of rebar.

(2) The pitting corrosion tendency of all phases in the rebar was as follows: TiVN₂ > Ca₈MgAl₆Si₅O₂₈ > NbVCN > MnS > Fe₃C > TiN. The tendency of anodic dissolution of α-Fe induced by inclusions was SiO₂ > CaAl₂SiO₆ > CaMgSiO₄. The conductivity and work function were two necessary conditions for forming a microcell. Al₂O₃, Mullite and MgAl₂O₄ did not conductivity and did not participate in galvanic corrosion. Calcium magnesium treatment, especially calcium treatment, was not conducive to reduce the corrosion resistance of rebar. At the same time, the solid elements in the rebar had an important influence on the passivation and corrosion resistance of HRB500E rebar.