Influence of Zinc Content and Chloride Concentration on the Corrosion Protection Performance of Zinc-Rich Epoxy Coatings Containing Carbon Nanotubes on Carbon Steel in Simulated Concrete Pore Environments

The lifetime of reinforced concrete structures can be compromised by the initiation and propagation of corrosion processes in the reinforcing steel. It is well known that carbonation and the presence of chloride ions are the most common factors that initiate corrosion degradation processes.^1-2  Corrosion starts with the penetration and diffusion of these aggressive species through the concrete structure until they reach the reinforcing steel. At this point, an electrochemical cell is formed; anodic and cathodic sites are created at the reinforcing steel surface and they electrically interconnect through the steel itself. Concrete pores serve as reservoirs for the electrolytic medium, inside these pores; dissolved aggressive ions can accumulate, allowing for direct contact with the steel surface. The formation of iron corrosion products at the steel surface creates tensile stresses inside the concrete structure that eventually result in structural damage, such as cracking and spalling of the concrete surface.³ 

Zinc-rich epoxy primers (ZRPs) have been extensively used to protect carbon steel substrates from corrosion degradation.^4-8  They are commonly used as primers in combination with organic topcoats to provide longer corrosion protection and weathering resistance.⁹  The mechanism of corrosion protection of ZRPs is significantly different from the barrier protection mechanism provided by traditional organic coatings, in which a physical barrier isolating the metal substrate from the environment is the only form of corrosion protection. In contrast, ZRPs provide an active corrosion protection (galvanic protection) to the carbon steel substrate as a result of the preferential dissolution of zinc in the presence of oxygen and moisture. In addition, ZRPs can also provide barrier protection, as a result of the presence of the epoxy binder and the inhibition of the oxygen reduction reaction (ORR) at the metallic substrate. This inhibition process can be observed as a result of two effects:⁹  (1) electrochemical reactions between oxygen and the zinc particles that were not involved in the cathodic protection process that reduce the amount of oxygen reaching the metallic substrate, and (2) presence of zinc corrosion products at the carbon steel surface as a result of the galvanic protection process that provide an additional barrier layer to diffusion of oxygen.

The mechanism of corrosion protection of ZRPs has been defined in terms of three stages.⁹  The first stage corresponds to an intact condition in which the epoxy binder isolates the carbon steel substrate from the environment and represents a physical barrier for the diffusion of aggressive species.^10-11  Epoxy coatings have been commonly used as the polymeric matrices to host the zinc particles because of their excellent chemical and corrosion resistance in different aggressive environments, as well as their high cross-link density, outstanding adhesion to metallic substrates, and high compatibility with a variety of topcoats.^9-10,12-13  A second stage is developed once the electrolyte diffuses through the ZRP and the zinc particles are able to provide galvanic protection to the metallic substrate, where oxidation of the zinc particles occurs at the zinc/electrolyte interface and the reduction reaction takes place at the steel substrate. Finally, after the cathodic protection is no longer effective, the third stage corresponds to a long-term barrier protection period that is provided as a result of:⁹  (1) the presence of zinc corrosion products on the carbon steel surface that reduces the amount of ionic species reaching the metallic substrate, and (2) the formation of zinc corrosion products inside the coating system and at the coating surface that block ionic pathways for the diffusion of aggressive species across the material. During the second stage of this mechanism of corrosion protection, ZRPs must satisfy three conditions in order to provide an effective galvanic protection to the steel substrate:⁹  (1) the zinc particles must be electrically connected with each other in order to provide a conductive network for the electronic transfer between the zinc particles and the carbon steel substrate, (2) the zinc particles must also be in electrical contact with the carbon steel substrate, and (3) the ZRP must be wetted by an electrolyte to establish the ionic conduction between the substrate and the zinc particles to induce electrochemical reactions between the zinc particles and oxygen, leading to the formation of solid corrosion products that protect the steel.¹⁰  The last condition is relatively easy to satisfy once the coated steel substrate is exposed to an aggressive environment such as seawater. On the other hand, the first two conditions are more difficult to accomplish because they are related to the design, preparation, and application of the paint to the metallic substrate. These conditions imply that a large load of zinc particles must be added to the epoxy binder in order to guarantee the continuous electrical pathways between the zinc particles and the metal substrate for the cathodic protection process to take place. Several studies have reported that the zinc content must be higher than 90 wt% Zn to provide sacrificial protection to bare steel.^8,14-17  At this concentration condition, abrupt changes in the physicochemical properties of the epoxy binder will be observed; for example, blistering and gloss will decrease, whereas permeability and rusting will increase.¹⁸  In addition, high zinc content will reduce the adhesion to the metallic substrate resulting from the lack of epoxy binder and it will increase the viscosity of the coating, leading to difficulties in spraying and in the agglomeration of zinc particles during storage.¹⁹  Therefore, conductive particles have been proposed to improve the galvanic effect of the zinc particles and to reduce the required zinc content that can compromise the physicomechanical properties of the epoxy binder, such as adhesion, flexibility, and impact resistance.²⁰ 

Carbon-based pigments represent an alternative for enhancing the electrical conductivity between the zinc particles and the metallic substrate. Among the different carbon-based materials, carbon nanotubes (CNTs) are suitable for incorporation into zinc-rich epoxy primers, as they possess a high aspect ratio and excellent intrinsic electrical conductivity, which improves the electronic conduction of the epoxy matrix with a significantly small amount of CNTs.²¹  Owing to the high aspect ratio and high specific surface area of CNTs, they exhibit a lower percolation threshold compared to other conductive particles such as carbon black, carbon fibers, or metallic pigments.²²  Additionally, CNTs enhance the mechanical properties of the composite coating as a result of the strong interfacial interaction with the epoxy polymer that allows CNTs to be used as reinforcing material in composite coatings.^23-24  The use of carbon nanotubes and ZRPs would create an optimal symbiosis to reduce the corrosion of structures where these materials can be applied.

The purpose of this study is to investigate the corrosion protection performance of zinc-rich epoxy primers containing multiwall carbon nanotubes (CNT-ZRPs) during exposure for 150 d to simulated concrete pore (SCP) solutions with different chloride-to-hydroxide ([Cl⁻]/[OH⁻]) ratios using open-circuit potential (OCP) and electrochemical impedance spectroscopy (EIS). Coating systems with three different zinc contents (60 wt% Zn, 70 wt% Zn, and 80 wt% Zn) and a fixed composition of CNTs (<1 wt%) were studied. The physical and chemical transformations after 150 d of immersion in the different electrolyte solutions were also monitored by salt spray tests, morphology studies using scanning electron microscopy (SEM) equipped with an energy dispersive x-ray spectroscopy (EDS) analysis system, and x-ray diffraction (XRD).

The corrosion protection performance of zinc-rich epoxy primers containing multiwall carbon nanotubes with different zinc content was investigated; all of the coating samples were provided by Tesla Nanocoatings. The coating formulation has been reported elsewhere,⁶  which corresponded to a solvent-based, two component epoxy-polyamide primer that contained spherical zinc dust as the primary pigment and multiwall carbon nanotubes as additives. The different coating systems are listed in Table 1. AISI 1008 carbon steel (UNS G10080⁽¹⁾) panels (152 mm × 76 mm × 1 mm) were sandblasted by air spraying, and the coating systems were applied and allowed to cure for one week at 20°C, resulting in a dry film thickness of approximately 4 mils (101.6 μm).

EIS and OCP measurements were conducted using a Bio-Logic SP-200 Research Grade Potentiostat/Galvanostat/FRA^†. A glass electrochemical cell, exposing an area of 4.67 cm², was attached on the surface of each coating system by an O-ring and a metal clamp. The electrochemical cell was filled with a simulated concrete pore solution containing different chloride concentrations; Table 2 shows the chemical composition of the different electrolyte solutions. The coating systems were exposed to the electrolyte solutions for up to 150 d at 20°C. During this period, electrochemical measurements were regularly performed using a conventional three-electrode cell consisting of the coated steel panel as the working electrode, a saturated calomel electrode (SCE) as the reference electrode, and a Pt/Nb mesh electrode of 2 cm diameter as the counter electrode. The electrochemical cells were placed in a Faraday cage during the electrochemical measurements. First, OCP measurements were performed for 10 min (1 h for the first measurement), and then EIS measurements were conducted at OCP in a frequency range from 100 kHz to 10 mHz and with a sinusoidal amplitude of 10 mV. All electrochemical measurements were performed twice to verify reproducibility. Finally, the EIS results were fitted with EC-lab^† software.

A different set of coated samples were placed in a salt spray chamber following the ASTM B117 “Standard Practice for Operating Salt Spray (Fog) Apparatus.” Intact and scribed samples were considered for this study. Before panels were introduced to the spray chamber, the backs and edges were covered with adhesive tape to isolate the exposed area. The samples were exposed for 30 d and photographs were taken after 3 d, 7 d, and 30 d of exposure to visually inspect the degradation of the coating systems. In order to guarantee the reproducibility of the results, three samples of each formulation were tested.

The morphology of the coating systems after being exposed to the electrolyte solutions for 150 d was inspected with a high-resolution SEM (Tescan LYRA-3 Model GMH Focused Ion Beam Microscope^†) at an accelerating voltage of 10.0 kV and under secondary electron mode. In addition, line scan analyses were performed using EDS (Standard EDS Microanalysis System with X- MaxN 50^†).

XRD was performed on a Bruker powder diffractometer utilizing a Cu-Kα source (λ = 1.54 Å = 0.154 nm) generated at 40 kV and 25 mA current. The coating samples before exposure and after 150 d of exposure to the simulated concrete pore solutions containing chloride ions were examined using XRD. All samples were scanned continuously from 10° to 90° with a step size of 0.03° per second. The evaluation of the data was performed using the EVA^† software, containing the JCPDS (ICDD) database files.

The cathodic protection behavior of zinc-rich epoxy primers when coupled with carbon steel substrates is commonly studied by monitoring the OCP of the system.^10,13  Metallic zinc shows highly negative OCP values (more than −1.4 V_SCE) during initial exposure to concrete environments,^25-28  passive reinforcing steel in concrete has a potential of −0.028 V_SCE,²⁹  and reinforcing steel actively corroded in concrete has a potential between −0.278 (intermediate [~50%] corrosion risk) and more than −0.428 V_SCE (high [>90%] risk of corrosion).²⁹  In order to control the effectiveness of the cathodic protection system, a criterion based on a maximum and a minimum potential has been established.²⁹  The maximum potential corresponds to the hydrogen evolution limit, and it has been set at −1.1 V_Ag/AgCl (−1.143 V_SCE).²⁹  The minimum potential has been set at −0.72 V_Ag/AgCl (−0.763 V_SCE) and corresponds to the cathodic protection limit.²⁹  Above this limit, zinc is no longer active to provide effective cathodic protection to the carbon steel substrate.¹² 

At pH 12.5 or higher and very negative potentials (more than −1.4 V_SCE), the dissolution of zinc that provides galvanic cathodic protection to the carbon steel substrate occurs vigorously in the presence of the hydrogen evolution reaction (HER):³⁰ 

The hydrogen bubbles produced from Reaction (1) are undesirable because they can promote early blistering underneath the coating, leading to adhesion loss between the carbon steel substrate and the coating system,³¹  and they can also induce hydrogen embrittlement when monoatomic hydrogen is adsorbed on the carbon steel surface and then diffuses into the steel.^12,25,32  The extension and activity of the HER will be limited by the alkaline environment and also by the progressive loss of electrical contact between the zinc particles and the metal substrate as they are providing galvanic protection to the carbon steel substrate.^25,33  Once the potential of the coating system becomes more positive than the cathodic protection limit (−1.143 V_SCE), the HER will be interrupted and the anodic dissolution of the zinc particles will proceed in the presence of the ORR:²⁷ 

The zinc dissolution process which takes place in the presence of the ORR will be also limited by the loss of the electrical contact between the zinc particles and the metallic substrate. This limit corresponds to the cathodic protection limit (−0.763 V_SCE); above this potential value, the zinc particles will be unable to provide further galvanic protection to the metallic substrate as a result of formation of nonconductive zinc corrosion products that isolate the zinc particles from each other and from the metal substrate.⁹ 

Figure 1 shows the OCP values for CNT-60ZRP immersed for 150 d in simulated concrete pore solutions with different chloride-to-hydroxide ratios. From Figure 1, it can be seen that OCP values for this coating system were above the cathodic protection limit for the entire immersion period. These results suggest that CNT-60ZRP was unable to provide cathodic protection to the carbon steel substrate, even when CNTs were added to the zinc-rich primer. This behavior can be explained as a result of the low pigment volume concentration of this material, which is around 25 vol% Zn, making the zinc particles completely embedded in the epoxy primer and highly separate from each other. As cathodic protection is effective when the zinc particles are in direct contact (or connected through CNTs) with other zinc particles and with the carbon steel substrate, it is expected that at very low zinc content, there are insufficient electronic pathways to drain cathodic protection current to the carbon steel substrate. However, it is important to notice that, even though cathodic protection was not provided to the metallic substrate, the OCP values for CNT-60ZRP were shifted to more positive values during the first 40 d of immersion, after which the potential was nearly constant during the remaining immersion time. This trend suggests that the exposed surface area of the metallic substrate to the electrolyte solution was negligible as a result of the highly cross-linked character of the epoxy polymer that provided good barrier protection to the carbon steel substrate.³⁴  Regarding the chloride-to-hydroxide ratio, the OCP values were similar during the entire exposure, even for the coating sample exposed to the highest chloride-to-hydroxide ratio that corresponds to a sodium chloride concentration higher than 5 wt%, which may indicate that the electrolyte did not reach the carbon steel surface as a result of the high barrier protection provided by the epoxy matrix.

Shreepathi, et al.,³⁵  studied the influence of Zn content on the corrosion protection of ZRPs immersed in 3.5% NaCl. They found that coating systems with 60 wt% and 70 wt% did not provide an efficient galvanic protection to the steel substrate. They also found poor barrier protection properties for these composite materials, concluding that ZRPs with zinc content between these values present a risk if used as protective coatings. These findings are in agreement with the electrochemical behavior observed for a ZRP with 70 wt% Zn and without the presence of CNTs (CNT-free 70ZRP). Figure 2 shows the OCP evolution for CNT-free 70ZRP immersed in the different electrolyte solutions. As it can be seen, OCP values were above the cathodic protection limit during the entire immersion period, meaning that this coating system cannot provide galvanic protection to the metallic substrate.

In contrast, the OCP evolution for CNT-70ZRP showed a different trend during the first 3 d to 7 d of immersion. From Figure 3, it can be observed that the OCP values shifted toward the cathodic protection region during the beginning of the immersion time. The observed behavior was likely a result of the activation of the zinc particles by the electrolyte solution, in which native zinc oxide surrounding the zinc particles reacted with chloride ions, allowing the zinc particles to provide cathodic protection to the carbon steel substrate.⁴  This reaction increased the zinc-to-iron area ratio, causing the potential to shift cathodically. After 2 d immersion, the OCP values shifted to anodic potentials as a result of the consumption of the zinc particles and formation of zinc corrosion products that decreased the zinc active area that was available to provide cathodic protection. From these results, it was concluded that CNTs improved the electrical contact between the zinc particles and the carbon steel substrate, allowing the zinc particles to afford galvanic protection at least during the early days of exposure. Following 3 d to 7 d immersion, the OCP values were no longer in the cathodic protection region, and they shifted to more positive potential values close to −0.03 V_SCE. After 20 d immersion, these OCP values stabilized in this passive region and remained constant for the rest of the immersion time. This trend suggests that either zinc corrosion products were formed, thereby protecting the carbon steel surface, or that hydroxyl ions coming from the simulated concrete pore solution reached a concentration high enough to passivate the carbon steel substrate.³⁶  It has been reported that under the pH conditions of simulated concrete pore solutions (usually between 12.5 and 13), the passive layer formed on carbon steel consists of an inner layer of magnetite (Fe₃O₄) and an outer layer of ferric oxides, such as maghemite (γ-Fe₂O₃) and hematite (α-Fe₂O₃).^37-38 

A particular behavior was observed for the coating sample exposed to the simulated concrete pore solution with a chloride-to-hydroxide ratio of 10, in which the OCP shifted again to cathodic potential values after 70 d of immersion. This behavior can be associated with the activation of the carbon steel surface by the excessive amount of chloride ions reaching the metallic substrate. Additionally, a high amount of blisters were observed after the sample was removed from the electrolyte, indicating that iron corrosion products were formed underneath the coating inducing blister formation and delamination of the coating from the metallic substrate. As the corrosion process of reinforcing steel in concrete can be described as a competition between the stabilization and restoration provided by the hydroxyl ions and the depassivation of the carbon steel substrate by chloride ions,³⁹  one can conclude that in the CNT-70ZRP sample exposed to a chloride-free solution and in the sample exposed to a solution with a chloride-to-hydroxide ratio of 0.1, the influence of hydroxyl ions prevailed over the effect of chloride ions. In contrast, in the CNT-70ZRP sample exposed to the simulated concrete pore solution with a chloride-to-hydroxide ratio of 10, the degradation of the carbon steel surface by chloride ions predominated over the protection provided by the hydroxyl ions.

Figure 4 shows the OCP values for CNT-80ZRP immersed in the different electrolyte solutions for 150 d. From Figure 4, it can be observed that the OCP values decreased during the first 7 d of immersion in the chloride-free solution and during the first 3 d in the electrolyte solutions with chloride-to-hydroxide ratios of 0.1 and 10. Similar to CNT-70ZRP, this behavior is associated with the activation of the zinc particles by the electrolyte. As can be seen from Figure 4, the activation of the zinc particles was extended to longer immersion time for the CNT-80ZRP sample immersed in the blank solution as a result of the absence of chloride ions that can react with the zinc oxide covering the zinc particles that allow the rapid activation of zinc to provide cathodic protection to the carbon steel substrate. Based on the OCP values between 3 d and 7 d of immersion, the cathodic protection process occurred in the presence of the HER during this immersion time. However, after 7 d of immersion, the OCP values shifted to values that were within the cathodic protection region, suggesting the end of the HER. The carbon steel substrate was cathodically protected for 10 d when the coating system was immersed in the blank solution and in the electrolyte with low chloride concentration, and for 20 d when the coating system was exposed to the solution with the highest chloride concentration. This behavior shows that CNT-80ZRP samples provided galvanic cathodic protection for longer period of immersion compared to the CNT-70ZRP samples, as a result of the higher zinc content that increases the number of electronic pathways by either direct contact between zinc particles, or by interconnection of the zinc particles with CNTs. After 10 d of immersion, the OCP values for CNT-80ZRP samples exposed to the blank solution and the solution with low chloride concentration shifted to more anodic potentials and, thereafter, they remained almost constant at approximately −0.25 V_SCE. This trend can be associated with the formation of zinc corrosion products that prevented corrosion attack to the metallic substrate. In contrast, the OCP values for CNT-80ZRP exposed to the highest chloride concentration shifted slightly above the cathodic protection limit between 30 d and 70 d of immersion, indicating that zinc particles were losing the electrical contact between each other and with the carbon steel substrate. Nevertheless, after 70 d of immersion, the OCP values shifted again into the cathodic protection region, suggesting that the zinc particles were re-activated by the electrolyte medium. It is possible that the high chloride concentration could destroy the zinc corrosion products formed during the initial cathodic protection process, and therefore the protective layer of zinc corrosion products either cannot form or cannot remain stable. Additionally, chloride ions can react with the zinc particles even when they do not provide cathodic protection to the metallic substrate. This is a self-dissolution process of the zinc particles, resulting from the high concentration of chloride ions, in which zinc dissolution and the ORR occur on the surface of the zinc particles.

Figure 5 shows the EIS spectra for CNT-60ZRP immersed in the simulated concrete pore solutions for 150 d. As can be seen in Figure 5, similar impedance spectra were observed for the coating samples immersed in the different electrolyte solutions. The Nyquist representation showed only one time constant, and the Bode diagram exhibited high impedance magnitudes at 0.01 Hz, close to 10¹⁰ Ω·cm²; this behavior suggests that good barrier corrosion protection was provided to the metallic substrate and very limited charge transfer reactions were taking place at the coating/metal interface.⁴⁰  The barrier protection is a consequence of a relatively inert and dielectric material caused by the highly cross-linked epoxy matrix, where zinc particles are completely embedded in the epoxy primer and highly isolated from each other and from the metallic substrate.^40-41  In addition, it has been reported that CNTs can play an important role in providing corrosion protection to carbon steel surfaces as they are able to fill in micropores and flaws in the epoxy matrix.²³  Other studies reported that CNTs are also able to fill in microholes in the metallic surface, reducing the number of active sites for metal dissolution.⁴²  It is important to note that regardless of the high corrosion resistance shown by this coating in the different electrolytes, for the sample immersed in the highest chloride concentration, the impedance magnitude at 0.01 Hz decreased faster than in lower chloride concentrations; this behavior has been associated with the degradation of the epoxy matrix by chloride ions.^10,43  Some authors have reported that electrolytes with high chloride concentrations catalyze hydrolysis reactions that induce coating degradation and formation of microcavities in the coating that can work as pathways for the diffusion of electrolyte throughout the material.^43-44 

Figure 6 shows the Nyquist and Bode representations for CNT-70ZRP immersed in the simulated concrete pore solutions at different [Cl⁻]/[OH⁻] ratios. The impedance spectra at the first day of immersion showed high corrosion resistance of CNT-70ZRP in all of the electrolyte solutions, with impedance magnitudes at 0.01 Hz close to 10¹⁰ Ω·cm², similar to the values for CNT-60ZRP. This behavior was observed because of negligible diffusion of electrolyte through the coating system, which was described by the almost intact condition of the epoxy material. Unlike the EIS evolution of CNT-60ZRP, the impedance magnitude at 0.01 Hz of CNT-70ZRP significantly decreased after the first day of immersion in the different electrolyte solutions, owing to diffusion of electrolyte throughout the coating system, which is explained by the lower content of epoxy binder to encapsulate the zinc particles, resulting in a more porous material compared to CNT-60ZRP.⁴⁵  The Nyquist representation showed a depressed capacitive loop that can be associated with the dielectric properties of the zinc-rich epoxy material; because cathodic protection was active during the first 3 d to 7 d of immersion, several previous works have described that, in fact, this capacitive loop could correspond to a mixed impedance response combining the barrier properties of the epoxy matrix, contact impedances between the zinc particles, and electrochemical reactions between the zinc particles and the electrolyte.^4,8  CNTs could also play a role during this period by interconnecting the zinc particles between each other and with the carbon steel substrate, thereby increasing the number of electronic pathways that were available to provide cathodic protection to the metal substrate.⁴⁶ 

The impedance magnitude at 0.01 Hz and the capacitive loop in the Nyquist representation continuously decreased as the immersion time progressed. After 20 d of immersion, another time constant can be recognized at low frequencies (0.1 Hz). Based on the OCP values for this coating formulation, cathodic protection was no longer active after 20 d of immersion. However, zinc particles that did not contribute to the galvanic cathodic protection process could still interact with the electrolyte solution, in which dissolution of the zinc particles and the ORR both happened at the zinc particle/electrolyte interface.⁶  This dissolution process of the zinc particles can explain the appearance of the second time constant at low frequencies. A second interpretation of the presence of this time constant can be associated with electrochemical reactions between the carbon steel substrate and the electrolyte solutions. In this interpretation, the high concentration of hydroxyl ions present in the electrolyte solutions reacted with the carbon steel substrate to form a passive oxide layer on the metal surface, similar to the one that is formed in reinforcing steel embedded in concrete. This observation explains the highly anodic potential observed in Figure 3 for CNT-70ZRP immersed in the different electrolytes after 20 d, which is similar to the reported OCP values for passive reinforcing steel in concrete.²⁹ 

After 100 d of immersion, a diffusion-like behavior was observed at low frequencies (0.5 Hz to 0.01 Hz). This behavior has been associated with either diffusion of reacting chemical species through zinc corrosion products³³  or an oxide passive layer on the carbon steel/coating interface. The presence of corrosion products or an oxide layer at the metallic substrate provides an extra barrier protection to corrosion degradation processes. This behavior is in agreement with the almost constant impedance magnitude at 0.01 Hz and the highly anodic potential of CNT-70ZRP after 100 d of immersion in the blank solution and in the electrolyte with a chloride-to-hydroxide ratio of 0.1. Therefore, the influence of hydroxyl ions in forming an oxide layer at the carbon steel surface was superior compared to the influence of chloride ions in dissolving the metallic substrate. For the specific coating system immersed in the simulated concrete pore solution with a chloride-to-hydroxide ratio of 10, the EIS signal showed a severe drop in its magnitude at 0.01 Hz as a result of the high chloride concentration that can compromise the presence of corrosion products at the coating/metal interface or induce the passive layer breakdown. This behavior was in agreement with the OCP measurements reported in Figure 3, in which, after 70 d of immersion, the OCP shifted to more negative potentials that approached those of the OCP values reported for reinforcing steel actively corroded in concrete. This impedance signal can be also attributed to a negative effect of the CNTs when they are exposed to electrolytes with high ionic conductivities (i.e., high concentration of ionic species). In these environments, a high rate of cathodic reaction is prompted to occur, which is completed by the anodic reaction of either one or both metals (i.e., carbon steel surface and the zinc particles). Thus, the CNTs become electrodes where extensive iron and zinc dissolution can occur, leading to a shortened lifetime of the coating system.⁴⁶ 

The impedance spectra for CNT-80ZRP immersed in the different electrolyte solutions for 150 d is shown in Figure 7. From the Bode diagrams, it can be seen that diffusion of the electrolyte occurred right after immersion as a result of the high porosity of the coating system. The rapid diffusion of electrolyte suggests that the pigment volume concentration (PVC) for this specific zinc-rich epoxy primer was higher than the critical PVC for this material, which is usually in the range of 70 wt% to 80 wt% Zn,⁴⁷  leading to a higher number of ionic conductive pathways for electrolyte penetration.⁴⁸  Following the first day of immersion, there was a noticeable decrease in the impedance magnitude at 0.01 Hz to impedance values between 10⁴ Ω·cm² to 10³ Ω·cm². This behavior was attributed to the dissolution of the zinc oxide surrounding the zinc particles; this dissolution re-activated the particles and increased the zinc active area to provide galvanic cathodic protection to the metallic substrate. The complex representation showed one capacitive loop similar to the one for CNT-70ZRP that was associated with the mixed impedance response combining barrier properties of the epoxy matrix, contact impedances between the zinc particles, and Faradaic reactions between the zinc particles and the electrolyte solutions. Following the activation period, the Nyquist representation showed a significant change in the impedance spectra of CNT-80ZRP exposed to the different electrolyte solutions. Two capacitive loops were shown from high to medium frequencies (0.8 Hz to 1,300 Hz) and one inductive loop was observed at the lowest frequencies (0.02 Hz). Based on the OCP values reported in Figure 4, this period corresponded to the cathodic protection process in the presence of the HER. It was suggested that the capacitive loop observed at the highest frequency (400 Hz to 1,300 Hz) represented the dielectric properties of the zinc-rich epoxy material, the capacitive loop at intermediate frequencies (0.8 Hz to 1.2 Hz) described the charge transfer process resulting from the dissolution of zinc occurring at the surface of the zinc particles and the HER occurring at the carbon steel substrate, and the inductive loop observed in the lowest frequency range (0.02 Hz to 0.01 Hz) was attributed to adsorption of monoatomic hydrogen on the metallic substrate.^6,49 

Following the galvanic cathodic protection process combined with the HER, the impedance magnitude at 0.01 Hz increased as a result of the formation of zinc corrosion products that provided additional protection to the carbon steel substrate. Three time constants were defined in the Nyquist representation. The time constant at the lowest frequencies (0.07 Hz) was associated with the galvanic cathodic protection process taking place in the presence of the ORR, the time constant at the intermediate frequencies (600 Hz to 1,900 Hz) was related to the dielectric properties of the zinc-rich epoxy coating, and the very small time constant at high frequency (20 kHz) was associated with the formation of a layer of zinc corrosion products on top of the coating system. The cathodic protection process was effective for 10 d when the coating system was immersed in the blank solution and in the simulated concrete pore solution with a chloride-to-hydroxide ratio of 0.1, while this cathodic protection was extended to 20 d when the coating system was exposed to the electrolyte solution with the highest chloride-to-hydroxide ratio. This behavior can be explained by the high galvanic activity of the zinc particles in the presence of chloride ions.

Finally, after cathodic protection was no longer active, a significant difference in the impedance magnitude at 0.01 Hz in the Nyquist representation was observed for the coating system immersed in the different electrolyte solutions. For the coating systems immersed in the blank solution and in the simulated concrete pore solution with a chloride-to-hydroxide ratio of 0.1, the impedance magnitudes at 0.01 Hz increased to values between 10⁵ Ω·cm² to 10⁶ Ω·cm² owing to the formation of a layer of zinc corrosion products on the coating/electrolyte interface as a result of the effective galvanic cathodic protection provided by the zinc particles. Additionally, these impedance magnitudes remained almost constant as the exposure time progressed, suggesting that the protective layer of zinc corrosion products on the coating surface was stable for long-term exposure. From the Nyquist representation and phase angle diagram, three time constants were observed: the time constant at the high frequencies represented the layer of zinc corrosion products formed on the coating surface, the time constant at the intermediate frequencies was describing the electrolyte resistance through the zinc-rich epoxy primer, and the time constant at low frequency (0.02 Hz) was related to further dissolution of the zinc particles in which both zinc dissolution and the ORR took place on the surface of the zinc particles, leading to the formation of zinc corrosion products around their surface. For the condition in which the coating system was immersed in the electrolyte with the highest chloride-to-hydroxide ratio, the impedance magnitude at 0.01 Hz only increased to values close to 10⁴ Ω·cm², and after 65 d of immersion it started to decrease again until reaching impedance values as low as the ones obtained in the activation period of the zinc particles. This behavior was in agreement with the OCP values for this coating system, in which the potential remained close to the cathodic protection limit, suggesting that zinc particles were reacting during the entire immersion time; the high concentration of chloride ions can cause the dissolution of the protective layer comprising zinc corrosion products formed during the cathodic protection process.⁵⁰  Chloride ions could also react with the remaining zinc particles that were not involved in the cathodic protection process leading to significantly lower impedance values at 0.01 Hz compared to the corresponding values obtained in electrolyte solutions with low chloride concentration.

Equivalent electrical circuits were used to reproduce the EIS signal of the different coating systems. Figure 8 shows the equivalent circuit used for CNT-60ZRP; in this circuit R_s corresponds to the resistance of the electrolyte, C_c represents the coating capacitance, and R_c describes the corrosion resistance of the zinc-rich epoxy material. This equivalent circuit was used for CNT-60ZRP during the entire immersion period because high impedance magnitudes at 0.01 Hz were observed and only one time constant was identified in the entire frequency range, suggesting a good barrier protection provided by this coating system. A constant phase element (CPE) was used instead of capacitance in order to consider deviation of the coating systems from ideal capacitive behavior.³³  The effective capacitance (C_eff)  associated with the CPE was calculated based on the expression reported by Orazem, et al.⁵¹ 

Figure 9 shows the coating resistance (R_c) and the effective capacitance of the coating (C_c,eff) values for CNT-60ZRP immersed for 150 d in the simulated concrete pore solutions with different chloride-to-hydroxide ratios. The R_c and C_c,eff are related to the corrosion-induced deterioration of the coating system that resulted from penetration of water and ionic species from the electrolyte medium.⁵²  From Figure 9(a), it can be seen that the coating resistance remained almost constant during the entire immersion time, with values between 10⁹ Ω·cm² and 10¹⁰ Ω·cm², suggesting that good barrier protection was provided to the metallic substrate. The coating resistance for the CNT-60ZRP sample immersed in the electrolyte solution with the highest chloride-to-hydroxide ratio decreased after 90 d of exposure, resulting in values slightly lower than 10⁹ Ω·cm². This behavior was associated with the degradation of the epoxy matrix during long-term immersion as a result of an excessive concentration of chloride ions. Figure 9(b) shows the effective coating capacitance values for the coating system immersed in the different electrolyte solutions, exhibiting values in the order of 10⁻¹⁰ F/cm². The coating capacitance slightly increased during the first 5 d after immersion, and then became almost constant during the remaining immersion period. This behavior indicates slow diffusion of electrolyte through the coating system resulting from high cross-linking of the epoxy material, and the potential influence of CNTs in filling micropores and flaws within the coating. It is important to notice that during the first 5 d of immersion, when the capacitance was increasing as a result of the permeation of electrolyte across the epoxy material, the coating system immersed in the blank solution showed the highest capacitance values as compared with the electrolyte solutions with chloride ions. This behavior is expected, as diffusion of the electrolyte is hindered when high concentrations of ionic species are present in the electrolyte medium.^43,53  Furthermore, the ionic radius of chloride (167 pm) is larger than the ionic radius of hydroxyl ions (133 pm) and the cations present in the simulated concrete pore solution (Ca²⁺: 99 pm, K⁺: 102 pm, and Na⁺: 138 pm); therefore, the diffusion rate of the electrolyte decreases as the concentration of chloride ions increase and increases when hydroxide ions and cations are the main chemical species in the electrolyte solution.

Figure 10 shows equivalent electrical circuits describing the impedance spectra of CNT-70ZRP. The equivalent circuit shown in Figure 10(a) was used to describe the almost intact condition during the first day of immersion. This equivalent circuit is similar to the one used for CNT-60ZRP. Figure 10(b) shows the equivalent circuit that describes the diffusion of electrolyte through the epoxy material and the activation of the zinc particles by the electrolyte solution. R_c and C_c combine the barrier properties of the epoxy matrix and the influence of contact impedances between the zinc particles, and R_ct and C_dl are associated with Faradaic reactions between the zinc particles and the electrolyte. The equivalent circuit shown in Figure 10(c) describes the stage in which a second time constant at low frequency was observed in the impedance spectra. R_ct,2 and C_dl,2 correspond to the charge transfer resistance and the double layer capacitance associated with electrochemical processes, such as dissolution of the zinc particles that did not contribute to the cathodic protection process or interaction between the electrolyte and the carbon steel substrate. Finally, the equivalent circuit shown in Figure 10(d) describes the last stage of the mechanism of corrosion protection of this coating system, in which the impedance magnitude at 0.01 Hz remained constant and a diffusion-like behavior was identified in the Nyquist representation. This behavior was associated with the formation of zinc corrosion products that provided additional barrier protection to the carbon steel surface and the formation of a passive layer at the carbon steel surface resulting from the high pH of the electrolyte solution. A Warburg element (W_s) was included in Figure 10(d), describing the mass transport process occurring at the carbon steel/coating interface.

Figure 11 shows the evolution of the equivalent circuit elements describing the impedance spectra for CNT-70ZRP. Figure 11(a) shows the coating resistance values combining the influence of contact impedances between the zinc particles; R_c rapidly decreased during the first few days of immersion because of the diffusion of electrolyte through the coating system, causing the activation of zinc particles and decreasing the contact impedances between them. After 20 d of immersion, when cathodic protection was no longer active and a second time constant was recognized in the Nyquist representation, R_c continuously decreased as a result of the progressive deterioration of the epoxy material by the electrolyte solutions. Finally, after 100 d of immersion, R_c remained almost constant for the CNT-70ZRP samples immersed in the blank solution and in the solution containing a chloride-to-hydroxide ratio of 0.1. This behavior can be associated with two processes: (1) formation of zinc corrosion products that improved the barrier properties of the epoxy material, or (2) formation of a passive layer on the carbon steel surface, resulting from the high concentration of hydroxyl ions in these electrolyte solutions that provided another barrier layer to prevent active dissolution of the metal substrate. In contrast, the coating resistance for the CNT-70ZRP sample immersed in the electrolyte solution with the highest chloride concentration progressively decreased during this period as a result of blister formation underneath the coating, which caused detachment of the epoxy coating from the metallic substrate. Figure 11(b) shows the capacitance values associated with the barrier properties of CNT-70ZRP. C_c,eff increased during the first few days of immersion as a result of diffusion of electrolyte through the coating system that increased the dielectric constant of the CNT-70ZRP.⁵⁴  With progression of the immersion time, the capacitance values of the coating system remained almost constant because of the saturation of the epoxy coating with the electrolyte. The coating system immersed in the highest chloride concentration showed a subsequent increase in the coating capacitance after 100 d of immersion. This trend has been associated with adhesion loss of the epoxy material from the metallic substrate resulting from blister formation.⁴¹ 

Figure 11(c) shows the evolution of the resistance describing charge transfer processes (R_ct) at either the zinc/electrolyte interface or at the metallic substrate. R_ct decreased during the activation of the zinc particles as a result of electrochemical reactions between the zinc particles and the electrolyte, which increased the zinc active area that was available to provide cathodic protection to the metal substrate. After 20 d of exposure, when a second time constant was recognized at low frequencies, R_ct also decreased as immersion time was progressing. It has already been mentioned that two mechanisms might have occurred: (1) dissolution of the zinc particles at the zinc/electrolyte interface, or (2) charge transfer processes between the carbon steel substrate and the electrolyte. During the last period of immersion, R_ct remained almost constant for the CNT-70ZRP samples immersed in the blank solution and in the solution with low concentration of chloride ions. This behavior can be explained as a result of formation of corrosion products on the surface of the zinc particles that avoided further dissolution processes or can be associated with the stability of the passive layer at the carbon steel substrate, preventing its active dissolution. For the CNT-70ZRP sample immersed in the electrolyte with the highest chloride concentration, R_ct was continuously decreasing even during the last period of immersion time. Two different processes were also proposed to describe this behavior: (1) further dissolution of zinc particles taking place at the zinc/electrolyte interface or (2) breakdown of the passive layer by the high concentration of chloride ions. This behavior suggests that the concentration of chloride ions in this electrolyte solution was higher than the chloride threshold value required for breakdown of the passive layer, leading to active dissolution of the carbon steel substrate. Finally, Figure 11(d) shows the evolution of the effective double layer capacitance  (C_dl,eff). For the CNT-70ZRP samples immersed in the blank solution and the electrolyte solution with low chloride concentration, C_dl,eff slightly increased during 80 d as a result of electrochemical reactions between the zinc particles and electrolyte that increased the active area of zinc to enable further dissolution of the zinc particles. After 80 d, C_dl,eff decreased as a result of the formation of zinc corrosion products on the surface of the zinc particles that diminishes the active area of the zinc/electrolyte interface.³³  C_dl,eff could also have decreased as a result of the formation of a passive layer at the metal substrate, which would have reduced the number of active zones available for the dissolution of the metal substrate. For the CNT-70ZRP sample immersed in the highest chloride concentration, C_dl,eff was mostly increasing during the entire immersion time, resulting from the continuous dissolution of the zinc particles and the breakdown of the passive layer creating active zones for iron dissolution.

Equivalent circuits describing the mechanism of corrosion protection of CNT-80ZRP at long-term immersion time are shown in Figure 12. The equivalent circuit associated with an intact condition was not considered because diffusion of electrolyte through the epoxy material occurred in the first few minutes of immersion. The equivalent circuit describing the activation of the zinc particles is shown in Figure 12(a); it has the same phenomenological interpretation as the one described in Figure 10(b) for CNT-70ZRP. Figure 12(b) shows the equivalent circuit representing the galvanic cathodic protection process provided by the zinc particles to the metal substrate in the presence of the HER. R_ct,2 and C_dl,2 describe the zinc dissolution process in the presence of the HER, and L and R_L correspond to the inductance and the resistance resulting from adsorption of monoatomic hydrogen on the carbon steel substrate. Figure 12(c) shows the equivalent circuit associated with the cathodic protection process in the presence of the ORR. R_ct,3 and C_dl,3 represent the cathodic protection process in which dissolution of zinc took place on the surface of the zinc particles while the ORR occurred at the carbon steel substrate. R_ox and C_ox were related to the formation of a layer of zinc corrosion products on the coating surface. Finally, the equivalent circuit in Figure 12(d) describes the last period of immersion in which a layer of zinc corrosion products was covering the majority of the coating exposed area. R_ct,4 and C_dl,4 represent the dissolution of the zinc particles that did not contribute to the cathodic protection process; this charge transfer process occurred at the interface between the surface of the zinc particles and the surrounding electrolyte. This process can be controlled by mass transfer through the zinc corrosion products formed on the surface of the particles; hence, a Warburg impedance was added to account for radial diffusion of ionic species through these corrosion products.⁴ 

Figure 13 shows the evolution of the different equivalent circuit elements describing the corrosion protection mechanism of CNT-80ZRP. Coating resistance values are shown in Figure 13(a); it is observed that R_c rapidly decreased during the activation period and the cathodic protection process in the presence of the HER as a result of the high diffusivity of electrolyte through the large number of pores inside the epoxy coating. These pores, created by the large PVC of the zinc particles, provided effective conductive pathways for ionic transport across the epoxy material. After the HER ceased, R_c significantly increased because of the formation of solid corrosion products that sealed the pores of the coating system. After 20 d of exposure, when zinc particles were unable to provide further cathodic protection, a different evolution of R_c was observed for the coating samples immersed in the electrolyte solutions. For the CNT-80ZRP samples immersed in the blank solution and in the electrolyte solution with low chloride concentration, R_c progressively increased during the remaining exposure time as a result of the formation of a protective layer of zinc corrosion products on the coating surface that prevented additional diffusion of the electrolyte. In contrast, for CNT-80ZRP immersed in electrolyte solution with the highest chloride concentration, R_c decreased continuously as immersion time progressed. This behavior is associated with degradation of the epoxy matrix by the high chloride concentration. Figure 13(b) shows the evolution of the effective capacitance of the coating for the CNT-80ZRP samples immersed in the three different electrolyte solutions. The capacitance values for the coating systems immersed in the blank solution and in the solution with low chloride concentration increased during the first few days, then these values decreased for the following days, and finally they were nearly constant during the remaining immersion time. This behavior was likely a result of the diffusion of electrolyte at the beginning of immersion time, the blocking of coating pores by zinc corrosion products, and the saturation of the coating system with the electrolyte solutions. The coating capacitance of the CNT-80ZRP sample exposed to the highest chloride concentration also increased during the first few days of immersion as a result of electrolyte penetration, and then it reached a relatively constant value because of the saturation of the epoxy material with the electrolyte. These coating capacitance values were significantly higher than the corresponding capacitance values for coating samples immersed in the blank solution and in the solution with low chloride concentration. This behavior confirms the degradation of the epoxy material by the high chloride concentration at long immersion time.

Figure 13(c) shows the R_ct values for CNT-80ZRP immersed in the different electrolyte solutions. Once again, different behavior was observed between the coating samples immersed in the three electrolyte solutions. For the CNT-80ZRP samples immersed in the blank solution and in the electrolyte solution with low chloride concentration, R_ct decreased during the activation period as a result of the removal of native zinc oxide covering the zinc particles, increasing the zinc active area and, therefore, decreasing the charge transfer resistance associated with zinc dissolution. In addition, R_ct also decreased during the sacrificial protection process in the presence of the HER as a result of the continuous dissolution of the zinc particles. During the cathodic protection process in the presence of the ORR, R_ct started to increase as a result of a reduction in the zinc active area. Later, when cathodic protection was overcome, R_ct gradually increased with increasing immersion time resulting from the loss of electric contact between the zinc particles and the carbon steel substrate.⁴  Finally, during the last period of immersion, R_ct remained constant as a result of the formation of a layer of zinc corrosion products on the coating surface that prevented additional diffusion of electrolyte and, therefore, further dissolution of the zinc particles. For the CNT-80ZRP sample immersed in the electrolyte solution with the highest chloride concentration, a similar behavior was observed in the first 20 d of exposure, as was observed in the coating samples immersed in the blank solution and for the simulated concrete pore solution with low chloride concentration. After 20 d of immersion, corresponding to the end of the galvanic cathodic protection, R_ct progressively decreased during the remaining immersion time. This behavior was associated with the further dissolution of the zinc particles that did not contribute to the cathodic protection process. Abreu, et al.,⁴  and Kannan, et al.,⁵⁵  have called this process a self-corrosion process, in which zinc was consumed by charge transfer processes in the presence of the ORR occurring on the surface of the zinc particles.

Finally, Figure 13(d) shows the evolution of the resistance associated with the protective layer of zinc corrosion products formed on the surface of the coating samples  (R_ox); the formation of this layer started during the cathodic protection process in the presence of the ORR by the combination of charged species. Hydroxyl ions formed on the carbon steel substrate and Zn²⁺ ions produced in the coating matrix diffused across the coating in order to balance the ionic charge; they combined to form insoluble zinc corrosion products on the coating surface.⁴  For these reasons, R_ox increased during the cathodic protection period as a result of the development of the protective layer, and then it became almost constant with progressing immersion time, suggesting high stability of the protective layer. For the specific situation of the CNT-80ZRP sample immersed in the electrolyte solution with the highest chloride concentration, R_ox slightly decreased with time, which can be associated with the re-dissolution of the zinc corrosion products by the large concentration of chloride ions.

The impedance magnitude at 0.01 Hz (|Z|_0.01 Hz) was also used in order to compare the corrosion protection performance of the different coating systems. This parameter has been reported for several research papers to describe the overall corrosion resistance of a coating system including electrolyte resistance, coating resistance, resistance of the coating/metal interface, and charge transfer resistance of the metallic substrate.^40,56  Figure 14 shows the evolution of |Z|_0.01 Hz for the different CNT-ZRPs immersed in the electrolyte solutions, where three mechanisms of corrosion protection can be recognized. The coating with 60 wt% Zn exhibited large |Z|_0.01 Hz values close to 10¹⁰ Ω·cm² throughout the entire immersion period, attributed to the highly cross-linked nature of the epoxy binder and the presence of CNTs that may act as pore fillers, providing a good barrier layer that isolates the steel substrate from the electrolyte. For the coating with 70 wt% Zn, it can be seen that |Z|_0.01 Hz decreased during 100 d of immersion (from 10¹⁰ Ω·cm² to 10⁷ Ω·cm²) as a result of water uptake and the presence of different charge transfer processes (galvanic protection, electrochemical reactions between the substrate and the electrolyte, and self-corrosion process of the zinc particles), and then it remained nearly constant (|Z|_0.01 Hz ≈ 10⁷ Ω·cm²) as a result of the passivation of the steel substrate and the formation of zinc corrosion products. The mechanism of corrosion protection for this coating can be summarized as a short-term cathodic protection (between 3 d and 7 d), followed by an intermediate barrier protection. Figure 14 also shows the evolution of |Z|_0.01 Hz for the coating with 70 wt% Zn and without the presence of CNTs. It is clearly seen that |Z|_0.01 Hz values for this coating system immersed in the different electrolytes were significantly lower compared to the corresponding values for CNT-70ZRP. Based on these results, it was concluded that CNTs provided a positive contribution to the corrosion protection performance of the coating with 70 wt% Zn; they improved the electron transfer between the zinc particles and the carbon steel substrate, as was shown from the OCP measurements, allowing the zinc particles to provide galvanic protection during the early stage of immersion. Furthermore, CNTs enhanced the total resistance of the coating, as can be seen from the evolution of |Z|_0.01 Hz, suggesting that CNTs can act as pore fillers hindering the diffusion of electrolyte throughout the material. The presence of CNTs in the coating with 70 wt% Zn provided a good balance between a moderated galvanic protection and a long-term barrier protection, with the additional advantage that unreacted zinc can be used to protect the steel surface from future mechanical damage of the coating or chemical degradation of the corrosion products that may occur from the continuous exposure to the electrolyte.^9-10  Finally, the coating with 80 wt% Zn showed an efficient galvanic protection, in which |Z|_0.01 Hz decreased at the beginning of the immersion period (|Z|_0.01 Hz ≈ 10³ Ω·cm²) because of the active dissolution of the zinc particles, then gradually increased over time, reaching an asymptotic value close to 10⁵ Ω·cm² for the coating samples exposed to the blank solution and the solution with low chloride concentration, as a result of the formation of insoluble zinc corrosion products that blocked pores in the material enhancing the barrier properties of the coating.⁵⁴  The barrier protection provided by these solid corrosion products was compromised in the coating sample exposed to the highest chloride concentration. |Z|_0.01 Hz decreased to very low values (below 10³ Ω·cm²) during the last period of immersion as a result of the dissolution of the zinc corrosion products previously formed during the cathodic protection process by the excessive concentration of chloride ions.

Salt spray exposure was performed to provide a visual assessment of the corrosion protection performance of the coating systems as a function of exposure time.⁵⁷  Figure 15 shows photographs of the different coating panels after exposure to salt spray fog chamber following ASTM B117 standard. Two conditions were examined in the test chamber: intact surface and scribed surface. From Figure 15, it can be seen that coating systems with an intact surface did not show significant damage after exposure for 30 d to the salt spray fog chamber. However, further visual observations showed some differences in the damage condition of the coating samples. The CNT-60ZRP panel showed excellent barrier corrosion protection and no corrosion products were observed after exposure. The CNT-70ZRP panel showed little amount of iron corrosion products at the edges of the panel, resulting from inappropriate masking during the sample preparation. Moreover, after 30 d of exposure, the CNT-70ZRP panel showed white corrosion products that resulted from the corrosion process between the zinc particles and the aggressive environment. Finally, the CNT-80ZRP panel developed high amounts of zinc corrosion products that covered the entire exposed area of the panel.

The deterioration process of the coatings was also investigated in the scribed surfaces. Figure 15 showed severe corrosion attack on the CNT-60ZRP panel. Large amounts of iron corrosion products and a lower amount of zinc corrosion products were formed at the scratched area. This visual observation confirmed that CNT-60ZRP provided excellent barrier protection to the carbon steel substrate but was unable to provide sacrificial protection to the metal substrate. The CNT-70ZRP panel developed a lower amount of iron corrosion products compared to the CNT-60ZRP panel. In addition, the visible amount of zinc corrosion products on the CNT-70ZRP panel was slightly higher than in the CNT-60ZRP panel. Based on the surface appearance, it can be suggested that the CNT-70ZRP panel was able to provide galvanic cathodic protection to the metallic substrate, at least during the first few days (or hours) of exposure. Finally, the CNT-80ZRP panel showed clear evidence that effective sacrificial protection was provided to the carbon steel substrate at least during the first 7 d of immersion. As a final remark, it should be noted that even though iron and zinc corrosion products were formed at the scribes of all of the coating panels, no signs of blister formation, color change, or delamination were identified in any of the exposed panels.

SEM micrographs of the surface of CNT-60ZRP before and after immersion for 150 d in the different electrolyte solutions are shown in Figure 16. The SEM morphology for the intact condition (Figure 16[a]) shows that the zinc particles were highly embedded in the epoxy matrix and were separated by large distances between each other. After the immersion period, it can be seen that the surface morphology for the coating samples exposed to the chloride-free solution (Figure 16[b]) and to the solution with the low chloride concentration (Figure 16[c]) did not significantly change compared to the intact surface. Additionally, zinc corrosion products were not detected on the coating surface. These SEM images confirmed the good barrier protection of CNT-60ZRP, which was a result of the dense epoxy matrix and the inability of this coating system to provide galvanic cathodic protection, as the zinc particles were highly separated from each other and completely isolated by the epoxy resin. Similar to these coating samples, no corrosion products were found on the coating surface exposed to the electrolyte solution with the highest chloride concentration, as shown in Figure 16(d); however, a large amount of pores were identified as a result of the degradation of the epoxy polymer by the high concentration of chloride ions.

Figure 17 shows the SEM micrographs of CNT-70ZRP before and after immersion for 150 d in the electrolyte solutions with low and high chloride concentration. Figure 17(a) shows the intact condition of the coating where the higher zinc content of CNT-70ZRP compared to CNT-60ZRP can be seen. After immersion for 150 d in the electrolyte solution with low chloride concentration, zinc corrosion products were found on the coating surface as shown in Figure 17(b). It has been reported that these corrosion products formed on top of the coating are related to the sacrificial protection process afforded by the zinc particles, where Zn²⁺ ions produced on the surface of the zinc particles and OH⁻ ions formed on the carbon steel surface diffuse in the electrolyte to neutralize the charge within the epoxy binder, and then they precipitate, forming insoluble zinc corrosion products on the coating surface.^4,6  The EDS analysis in Figure 17(d) shows that the zinc corrosion products were mainly composed of zinc and oxygen, suggesting the formation of ZnO and Zn(OH)₂. In contrast, zinc corrosion products were not found on the coating surface exposed to the highest chloride concentration, as shown in Figure 17(c), even though sacrificial protection was identified from the OCP measurements during the first 7 d of immersion, proving that zinc corrosion products were not stable in the highly concentrated chloride solution. Additionally, the large amount of chloride ions caused blister formation at the metal/coating interface and increased in the coating porosity similar to the corresponding coating with 60 wt% Zn shown in Figure 16(d).

Figure 18 shows the surface morphologies of CNT-80ZRP before and after immersion in the different simulated concrete pore environments. As shown in Figure 18(a), the intact coating exhibited a significantly higher amount of zinc particles compared to the coatings with 60 wt% Zn and 70 wt% Zn that were less isolated by the epoxy binder. For CNT-80ZRP samples immersed in the blank solution (Figure 18[b]) and in the electrolyte with low chloride concentration (Figure 18[c]), a large amount of zinc corrosion products were covering the majority of the coating surface. These observations confirmed that the zinc particles provided efficient galvanic cathodic protection to the carbon steel substrate. The EDS analysis for the sample exposed to the electrolyte solution with low chloride concentration (Figure 18[d]) showed zinc and oxygen as the main elements of the zinc corrosion products, indicating the presence of ZnO/Zn(OH)₂. On the other hand, for the coating sample immersed in the solution with high chloride concentration, a lower amount of these corrosion products were found on the coating surface, as shown in Figure 18(e), verifying that re-dissolution or degradation of the zinc corrosion products occurred as a result of the excessive concentration of chloride ions that were in direct contact with the coating surface. The EDS analysis of the sample immersed in the highest chloride concentration (Figure 18[f]) showed that the primary elements in the corrosion products were Zn, O, and Cl, suggesting the formation of zinc oxides/hydroxides and chloride-based complexes.

The electrochemical behavior of the CNT-80ZRP sample immersed in the simulated concrete pore solution with a chloride-to-hydroxide ratio of 10 suggested that further dissolution of the zinc particles that did not contribute to the cathodic protection process occurred as a result of the high chloride concentration diffusing through the epoxy material. This zinc dissolution process, combined with the ORR, occurred on the surface of the zinc particles. To confirm this mechanism, SEM cross-sectional analyses were performed on this sample. Figures 19(a) and (b) show the cross section of this coating sample, in which the presence of zinc corrosion products on the surface of the zinc particles was clearly observed. A line scan analysis was performed on one of the zinc particles that was covered with zinc corrosion products. Figure 19(c) shows the EDS spectra for zinc and oxygen along the line scan. From Figure 19(c), it can be seen that the dark layer covering the zinc particle contained a lower amount of zinc and higher amount of oxygen in comparison to the composition at the center of the particle, which means that this layer indeed corresponds to zinc corrosion products that were formed on the surface of the zinc particles as a result of a self-corrosion process of the zinc particles (i.e., zinc dissolution and the ORR occurred on the surface of the zinc particles) at the zinc/electrolyte interface.

XRD analysis was performed to identify the crystalline phases of the zinc corrosion products formed after immersion in the electrolyte solutions. Figure 20 shows the XRD spectra for CNT-60ZRP before and after exposure to the simulated concrete pore solution with the low chloride concentration. The XRD diffractograms obtained on CNT-60ZRP before and after exposure were almost identical, showing that the main crystalline phase corresponded to metallic zinc. These XRD results make evident the good barrier protection of this coating system, which is in agreement with the electrochemical findings and the morphology analysis on the coating surface. The XRD spectra for CNT-70ZRP after the exposure to the different electrolyte solutions (not shown here) were similar to the corresponding XRD spectra for CNT-80ZRP. Figure 21 shows the XRD spectra obtained on CNT-80ZRP after 150 d of exposure to the simulated concrete pore solutions with low and high chloride concentration. In addition, Figure 21 also shows the XRD spectrum of the coating system before immersion to elucidate the chemical transformations of the zinc particles after the immersion in the electrolyte solutions. Similar to the coating with 60 wt% Zn, the XRD spectrum for CNT-80ZRP before immersion showed the presence of metallic zinc as the primary crystalline phase. After the immersion in the simulated concrete pore solution with low chloride concentration, the XRD pattern showed the presence of ZnO and Zn(OH)₂, which is in agreement with the EDS analysis shown in Figure 18(d). Additionally, a significant amount of metallic Zn remained after the immersion period, meaning that the zinc particles were not totally consumed during the galvanic protection process or the self-corrosion process. It has been reported that these unoxidized particles represent a reservoir to protect the steel substrate in case mechanical damage of the coating occurs.⁴⁰  The XRD spectrum for the coating sample exposed to the simulated concrete pore solution with the high chloride concentration showed the presence of metallic Zn, ZnO, Zn(OH)₂, and zinc hydroxide chloride, also called simonkolleite (Zn₅(OH)₈Cl₂·H₂O). From Figure 21, the intensities of Zn(OH)₂ at 20.18° was found to decrease dramatically from approximately 1,143 counts to 121 counts as the chloride concentration increased. These XRD patterns confirm that the high amount of chloride ions hindered the formation of solid corrosion products that can protect the carbon steel substrate. In addition, it is also noteworthy that the peak intensity of simonkolleite is very low despite the high chloride concentration in the simulated concrete pore solution.

The crystalline phases detected for the coating system with 80 wt% Zn were also reported in previous studies.^50,58-59  Mouanga, et al.,⁵⁹  reported that the overall reaction for the anodic dissolution of Zn immersed in a simulated concrete pore solution is given by:

This reaction proceeds by the initial formation of zinc hydroxide and its further reaction with hydroxyl ions to produce zincate ions:^60-61 

The zinc hydroxide detected by XRD on CNT-80ZRP after exposure to the electrolyte solutions was formed following Reaction (4). Zinc oxide was also detected by XRD after 150 d of immersion in the simulated concrete pore solutions with low and high chloride concentration. It has been reported that the formation of zinc oxide occurs by further transformation of the zincate ions as follows:⁶⁴

The XRD pattern obtained for CNT-80ZRP after 150 d of immersion in the simulated concrete pore solution with high chloride concentration showed the presence of simonkolleite. Mouanga, et al.,⁵⁹  reported that the formation of zinc hydroxide chloride occurs according to the following reaction:

Reaction (7) can explain the decrease in the intensities for the XRD peaks that corresponds to zinc hydroxide. From this reaction, it can be seen that the chloride ions react with water (or moisture) and the zinc hydroxide previously formed according to Reaction (4) to produce zinc hydroxide chloride. Several authors have reported that simonkolleite is not thermodynamically stable in alkaline environments as a result of its high solubility.^9,62  In high alkaline environments, simonkolleite is dissolved, releasing chloride ions that can promote further attack to the zinc particles or to the metallic substrate.⁶²  This observation explains the very low intensity of simonkolleite found in the XRD spectrum for CNT-80ZRP after 150 d of exposure in the simulated concrete pore solution with high chloride concentration, even though the concrete pore solution contained an enormous amount of chloride ions.

The corrosion protection performance of CNT-ZRPs with different zinc contents was investigated at long-term exposure in simulated concrete pore solutions with various chloride-to-hydroxide ratios. Electrochemical measurements, accelerated tests, morphology analyses, and chemical characterization were performed to investigate the mechanisms of corrosion protection of these coatings. The following conclusions were found:

• Based on the electrochemical measurements, three different corrosion protection mechanisms were identified. An excellent barrier corrosion protection was provided by CNT-60ZRP as a result of the low porosity of the epoxy binder and the likely positive effect of the CNTs who might fill out pores and flaws within the organic coating. A mixed corrosion protection mechanism was identified for CNT-70ZRP, where short-term cathodic protection, barrier protection, and Faradaic processes of the zinc particles were identified. Finally, for CNT-80ZRP, long-term galvanic cathodic protection occurred, leading to the formation of a protective layer of zinc corrosion products on top of the coating surface. It was presumed that CNTs improved the electrical contact between the zinc particles and the carbon steel substrate.

• The simulated concrete pore solution with a chloride-to-hydroxide ratio of 10 significantly changed the corrosion protection performance of the coating systems. The CNT-60ZRP sample showed a lower coating resistance as a result of the degradation of the epoxy matrix by the chloride ions. For CNT-70ZRP, blister formation occurred as a result of the breakdown of the passive layer at the carbon steel substrate and the active dissolution of the metal surface. For CNT-80ZRP, the solid corrosion products formed at the coating surface were dissolved by the high chloride concentration. In addition, zinc particles that did not contribute to the cathodic protection process were further dissolved inside the epoxy matrix.

• The morphology analyses correlated well with the electrochemical measurements. The surfaces of the CNT-60ZRP samples exposed to the blank solution and to the solution with low chloride concentration did not show significant changes compared to the coating surface before exposure, demonstrating the good barrier properties of this coating. The surface of the CNT-60ZRP sample exposed to the electrolyte with high chloride concentration showed higher porosity as a result of polymer degradation by the chloride ions. Zinc corrosion products were found on the surfaces of CNT-70ZRP and CNT-80ZRP samples exposed to the blank solution and to the simulated concrete pore solution with a chloride-to-hydroxide ratio of 0.1, as a result of the sacrificial protection process. The amount of zinc corrosion products on the coating surface was reduced for the samples exposed to the simulated concrete pore solution with the high chloride concentration. Finally, for the CNT-80ZRP sample exposed to the high chloride concentration, the presence of a self-corrosion process of the zinc particles that did not contribute to the sacrificial protection process was confirmed.

• The XRD spectrum obtained for CNT-80ZRP after 150 d of immersion in the simulated concrete pore solution with a chloride-to-hydroxide ratio of 0.1 showed the presence of remaining metallic zinc, zinc oxide, and zinc hydroxide, as a result of the galvanic protection process provided by the zinc particles. The XRD spectrum for the CNT-80ZRP sample exposed to the simulated concrete pore solution with a chloride-to-hydroxide ratio of 10 showed a decrease in the intensities of the Zn(OH)₂ peaks and the appearance of XRD peaks that corresponds to simonkolleite.

The authors would like to acknowledge the support of the CERL-U.S. Department of Defense Office of Corrosion Policy and Oversight and Tesla Nanocoatings Inc. for their special contribution in the coating preparation. Use of the TAMU Materials Characterization Facility is also acknowledged.