Corrosion Performance of Nano-ZrO2 Modified Coatings in Hot Mixed Acid Solutions

A nano-ZrO2 modified coating system was prepared by incorporation of nano-ZrO2 concentrates into phenolic-epoxy resin. The corrosion performance of the coatings was evaluated in hot mixed acid solution, using electrochemical methods combined with surface characterization, and the effects of nano-ZrO2 content were specially focused on. The results showed that 1% and 3% nano-ZrO2 addition enhanced the corrosion resistance of the coatings, while 5% nano-ZrO2 addition declined it. The coating with 3% nano-ZrO2 presented the minimum amount of species diffusion, the lowest average roughness (5.94 nm), and the highest C/O ratio (4.55) and coating resistance, and it demonstrated the best corrosion performance among the coating specimens.


Introduction
The flue gas from oxy-fuels has higher levels of acid gases, mainly including SOx and NOx with smaller amounts of HF and HCl [1]. The hot flue gases may be compressed and cooled to form mixed acid environments, which are mainly composed of highly concentrated H 2 SO 4 , HNO 3 , HCl and HF. With pH values that are usually in the range of 0-2, the mixed acids environment promotes an extremely high rate of steel corrosion, threatening the safety and operation of equipment.
Organic coatings have been used to control the corrosion of steels in the highly acidic environment. Phenolic-epoxy resin is the most important and industrialized epoxy polymer and possesses excellent acid and alkali resistance [2,3]. It acts as a physical barrier layer between the metal substrate and solution to inhibit the permeation of water and aggressive ions [4]. However, aggressive ions can easily attack coatings through the pores, especially strong acid ions. Due to the presence of porosity [5] and inherent brittleness [6], many coatings used in these conditions are too intolerant for strong acid at high temperature, which leads to premature failure [7].
In recent years, nanoparticles in the coating systems have shown outstanding properties, i.e., excellent barrier resistance, flame retardancy, and wear resistance, etc. [8][9][10][11][12], which have attracted more and more attention. The barrier properties of the organic coatings can be improved by the inclusion of proper fillers, and nano-sized fillers have superior barrier properties than conventional fillers, even at low concentrations, due to their higher surface area [13]. Among these nanoparticles, ZrO 2 nanoparticles are one of the most promising types of pigment [14] because of its advantages, such as remarkable chemical stability, wear resistance, strength and fracture toughness, and chemical

Particle Size Analysis of Nano-ZrO 2 Concentrate
Particle sizes were measured using the dynamic light scattering (DLS) technique with the Malvern Zetasizer (Nano ZS). Before the experiment, a small amount of nano-ZrO 2 concentrate was dropped into solvent PMA and ultrasonic shaken for 20 min to produce a concentration of 0.5 × 10 −6 . The measurement of particle sizes was performed at room temperature with three repetitions.

Surface Properties
The surface and cross section panels were cut out of the coated plates before and after the static immersion test. In the case of cross-section, it was embedded in bakelite using an XQ-2B mounting press in order to grind with SiC paper up to 2000 grit, and then it was polished. The surface and cross section morphologies of samples were analyzed by scanning electron microscopy (SEM, XL30-FEG-ESEM, Philips, Amsterdam, Holland) at 20 kV equipped with an energy dispersive spectrometer (EDS) to evaluate element distribution, and a gold film was sprayed atop the surface of samples to make it electrically conductive. The surface morphologies of coatings were characterized by atomic force microscopy (PicoSPM II AFM), by Molecular Imaging Corp. (Bruker Corporation, Saarbrucken, Germany), with a scanning area of 5 µm × 5 µm.
X-ray photoelectron spectroscopy (XPS, Thermo, Waltham, MA, USA) analysis was performed with an ESCALAB250 spectrometer to study the carbon-to-oxygen (C/O) variation in polymers before and after acid immersion, using AlKα excitation radiation (hv = 1486.6 eV). The XPS plot size was 500 µm, and the constant pass energy was 50.0 eV. The XPS depth profile was carried out using Ar + ions. The sputtering rate and area were 0.2 nm/s and 2 × 2 mm, respectively. At the same time, the atomic percentages of elements were calculated automatically with a testing instrument. The data was fitted with XPSPEAK4.1. In addition, coatings before immersion were almost the same, and only the Z-0 coating needed to be tested.

Electrochemical Studies
The electrochemical properties of the coatings were investigated using an electrochemical workstation (PAR273A, Princeton Instruments, Princeton, NJ, USA). Measurements were carried out in a three-electrode cell with mixed acid solution (including 3 wt % H 2 SO 4 , 1 wt % HCl, 0.5 wt % HNO 3 , and 0.2 wt % HF) as an electrolyte. The three-electrode cell included a saturated calomel reference electrode (SCE) filled with saturated KCl solution which served as a reference electrode (RE), a platinum auxiliary electrode with an exposure surface of 13 mm × 13 mm as a counter electrode (CE), and the sample with an exposure surface of 12.56 cm 2 as working electrode (WC). Prior to the electrochemical measurements, the panels were kept in the solution for 30 min in order to stabilize the free corrosion potential.
The electrochemical impedance spectroscopy (EIS) measurements were performed at open circuit potential in an applied frequency range from 100 to 10 mHz, and a sinusoidal perturbation signal with 20 mV amplitude was used. The obtained data were interpreted on the basis of equivalent electrical analogs using ZsimpWin to obtain the fitting parameters.

Characterization of Nano-ZrO 2 Concentrates
The size distribution, TEM micrographs, and chemical element mapping of nano-ZrO 2 concentrates are shown in Figure 1. The content of particles less than 100 nm in diameter reached 63.09%, with an average diameter of 74.5 nm (Figure 1a). The TEM observations showed that nano-ZrO 2 particles had a good dispersion state with dimensions of 10-50 nm, and the nano-ZrO 2 powder appeared to be gray and flat (Figure 1b). The EDS analysis and chemical element mapping showed the presence of elements like oxygen (68.07 at %), and Zirconium (31.93 at %) (Figure 1c), which provides direct evidence for the existence of nano-ZrO 2 .

Acid Immersion Test and Corrosion Morphologies
Typical SEM micrographs for different nano-ZrO2-modified coatings before immersion are displayed in Figure 2, along with the EDS analysis in coating Z-5. Pinholes were observed on the surface of the Z-0, Z-1, and Z-5 coatings (Figure 2a,b,d), except for the Z-3 coating (Figure 2c). The aggregation of nano-ZrO2 appeared on the surface of the Z-5 coating, which implies that the amount of aggregates increased with the nanoparticle content [17]. This was confirmed by the EDS analysis (Figure 2e), which showed the presence of elements like carbon (62.80 at %), zirconium (24.48 at %) and oxygen (12.73 at %).
SEM images of different nano-ZrO2-modified coatings after 192 h of acid immersion are shown in Figure 3. Various quantities and volumes of pinholes and cracks emerged on the surface of the coatings, which verifies that various levels of degradation occur. Pinholes uniformly distributed on the surface of the Z-0 and Z-1 coatings (Figure 3a,b). Furthermore, pinholes propagated to form cracks on the Z-0 coating. Since 1% nano-ZrO2 helps to inhibit the initiation and propagation of cracks, no obvious cracks appeared on the Z-1 coating, which demonstrates that the surface of coating Z-1 has better barrier properties than the Z-0 coating. It is necessary to emphasize that the degradation of the Z-3 coating mainly occurred on particular areas of existing fillers, whilst small pinholes formed, and their quantity and volume were much smaller than other coatings at high magnification ( Figure 3c). Three percent nano-ZrO2 dispersed uniformly in the Z-3 coating and generated a network nanostructure, which may be responsible for the improvement in corrosion resistance of the Z-3 coating. By contrast, the deterioration of Z-5 coating was most apparent. Non-uniform corrosion and the largest quantity and volume of pinholes were shown (Figure 3d), arising from the aggregation of excessive nanoparticles. Therefore, the surface of the Z-3 coating has the best barrier properties.

Acid Immersion Test and Corrosion Morphologies
Typical SEM micrographs for different nano-ZrO 2 -modified coatings before immersion are displayed in Figure 2, along with the EDS analysis in coating Z-5. Pinholes were observed on the surface of the Z-0, Z-1, and Z-5 coatings (Figure 2a,b,d), except for the Z-3 coating (Figure 2c). The aggregation of nano-ZrO 2 appeared on the surface of the Z-5 coating, which implies that the amount of aggregates increased with the nanoparticle content [17]. This was confirmed by the EDS analysis (Figure 2e), which showed the presence of elements like carbon (62.80 at %), zirconium (24.48 at %) and oxygen (12.73 at %).
SEM images of different nano-ZrO 2 -modified coatings after 192 h of acid immersion are shown in Figure 3. Various quantities and volumes of pinholes and cracks emerged on the surface of the coatings, which verifies that various levels of degradation occur. Pinholes uniformly distributed on the surface of the Z-0 and Z-1 coatings (Figure 3a,b). Furthermore, pinholes propagated to form cracks on the Z-0 coating. Since 1% nano-ZrO 2 helps to inhibit the initiation and propagation of cracks, no obvious cracks appeared on the Z-1 coating, which demonstrates that the surface of coating Z-1 has better barrier properties than the Z-0 coating. It is necessary to emphasize that the degradation of the Z-3 coating mainly occurred on particular areas of existing fillers, whilst small pinholes formed, and their quantity and volume were much smaller than other coatings at high magnification ( Figure 3c). Three percent nano-ZrO 2 dispersed uniformly in the Z-3 coating and generated a network nanostructure, which may be responsible for the improvement in corrosion resistance of the Z-3 coating. By contrast, the deterioration of Z-5 coating was most apparent. Non-uniform corrosion and the largest quantity and volume of pinholes were shown (Figure 3d), arising from the aggregation of excessive nanoparticles. Therefore, the surface of the Z-3 coating has the best barrier properties.       Figure 4 exhibits the cross section morphology and elemental mapping for different nano-ZrO 2 -modified coatings after 192 h of acid immersion. The cross section morphology showed that obvious cracks appeared at the interface between the substrate and coatings for the Z-0, Z-1 and Z-5 coatings, especially for the Z-5 coating, while the Z-3 coating remained unchanged. Figure 4 exhibits the cross section morphology and elemental mapping for different nano-ZrO2modified coatings after 192 h of acid immersion. The cross section morphology showed that obvious cracks appeared at the interface between the substrate and coatings for the Z-0, Z-1 and Z-5 coatings, especially for the Z-5 coating, while the Z-3 coating remained unchanged. The elemental mapping revealed that C and O are the major constituents of coatings, and minor amounts of Cl and Fe are also found, which may be introduced from substrates and electrolytes. The big white patches in the oxygen map are glass flakes, and consist of SiO2, CaO, and Na2CO3, corresponding to the low carbon content in carbon map. After 192 h of immersion, corrosive medium went through the coating and reached the metal/coating interface, resulting in the corrosion of substrates. Many iron oxides and hydroxides were generated above the metal substrate, whilst iron ion diffused into the coating via diffusion paths. The concentration of O tended to increase slightly at the metal/coating interface during the immersion period, with the following order: Z-5 coating > Z-0 coating > Z-1 coating > Z-3 coating, corresponding to the distribution sequence of Fe in the coatings. Based on the elemental distributions of O and Fe, different contents of rust formed at the metal-coating interface for these coatings, which revealed that 3% nano-ZrO2 effectively prevents the permeation of corrosive medium and reduces the generation of rust, whereas 5% nano-ZrO2 promotes the degradation of coatings.
The same diffusion property was observed for Cl and C. The behavior of Cl − may be attributed to its penetration through pinholes [18], which implies that the Z-3 coating possesses remarkable Cl − penetration resistance. Due to the phenomenon of C penetration, degradation of the coatings occurred in the mixed acid, which provided diffusion paths for carbonaceous species dissociated from coatings, and caused C to accumulate at the metal/coating interface. The distributions of Cl and The elemental mapping revealed that C and O are the major constituents of coatings, and minor amounts of Cl and Fe are also found, which may be introduced from substrates and electrolytes. The big white patches in the oxygen map are glass flakes, and consist of SiO 2 , CaO, and Na 2 CO 3 , corresponding to the low carbon content in carbon map. After 192 h of immersion, corrosive medium went through the coating and reached the metal/coating interface, resulting in the corrosion of substrates. Many iron oxides and hydroxides were generated above the metal substrate, whilst iron ion diffused into the coating via diffusion paths. The concentration of O tended to increase slightly at the metal/coating interface during the immersion period, with the following order: Z-5 coating > Z-0 coating > Z-1 coating > Z-3 coating, corresponding to the distribution sequence of Fe in the coatings. Based on the elemental distributions of O and Fe, different contents of rust formed at the metal-coating interface for these coatings, which revealed that 3% nano-ZrO 2 effectively prevents the permeation of corrosive medium and reduces the generation of rust, whereas 5% nano-ZrO 2 promotes the degradation of coatings.
The same diffusion property was observed for Cl and C. The behavior of Cl − may be attributed to its penetration through pinholes [18], which implies that the Z-3 coating possesses remarkable Cl − penetration resistance. Due to the phenomenon of C penetration, degradation of the coatings occurred in the mixed acid, which provided diffusion paths for carbonaceous species dissociated from coatings, and caused C to accumulate at the metal/coating interface. The distributions of Cl and C also demonstrated that the coatings had suffered various levels of degradation, in the order of 5%, 0, 1%, and 3% nano-ZrO 2 contents.
Internal voids inevitably exist in the coatings, owing to the volatilization of dissolvent and other reasons which provide diffusion paths for small corrosive molecules, such as H + , Cl − , and O 2 . The diffusion of these corrosive molecules promotes the failure of coatings. The presence of H + evidently accelerates the self-degradation rate of coatings [19], and Cl − reaches the interior of coating through these micro-voids, because of its strong penetration ability. Meanwhile, complete coatings cannot block the attack of acid medium, which causes the nucleation of new micro-voids. Micro-voids grow up with the immersion time, and larger corrosive molecules go through and accumulate at the metal-coating interface, such as carbonaceous species. The nano-ZrO 2 addition changes the size and amount of internal voids, which has an impact on the corrosion resistance of coatings. The specific effect of nano-ZrO 2 content on the variation of voids for coatings is discussed in the following EIS sections.
AFM images of different nano-ZrO 2 -modified coatings after 192 h immersion are shown in Figure 5, as well as the average roughness (R a ) and root mean square (RMS), as deduced from the AFM analysis, are presented in Table 2. Small hill-shaped degradation products were observed on the surface of the Z-0coating, and the main component was carbohydrates. Some flaky degradation products completely separated from the binder due to embrittlement and poor adhesion [20], leading to the highest R a (24.58 nm) and RMS (73.60 nm) values for the Z-0 coating. With the addition of 1% nano-ZrO 2 , the surface topography for the Z-1 coating changed and refined. Only smaller, regular micro-bulges like hills appeared on the surface, which resulted in a lower R a value of 12.79 nm and RMS value of 27.44 nm. With an increase in nano-ZrO 2 content, the smallest micro-bulges were observed on the surface of the Z-3 coating, and the R a and RMS values reduced to 5.94 nm and 16.10 nm, respectively, in spite of existing amorphous bulges and micro-holes. However, when the nano-ZrO 2 content reached 5%, the corrosion resistance of the Z-5 coating declined significantly. Patches of micro-bulges became larger, and flaky degradation products were faintly visible, which led to an increase in the R a (20.42 nm) and RMS values (52.03 nm). In addition, the R a value of the Z-5 coating was lower than that of the Z-0 coating, which resulted from the severe inhomogeneous corrosion of the Z-5 coating and a limited scanning area for AFM. In short, the addition of 3% nano-ZrO 2 minimized the average roughness and root mean square values, which effectively improved the corrosion resistance of the coating. Due to high surface sensitivity and chemical specificity, XPS was used to better understand the exposure-induced bond failure and component variation in the organic coatings during the corrosion test [7]. Figure 6 shows the XPS spectra (C1s and O1s) and atomic ratio (C/O) of the Z-0, Z-1, Z-3, and Z-5 coatings before and after acid immersion, and the corresponding atomic compositions are given in Table 3. The C1s and O1s XPS peaks of four coatings were at about 281.4 eV and 528.6 eV, respectively. The XPS intensity data were converted into units of at %, and the images provide detailed elemental components. The C1s and O1s spectra were peak-fitted to deduce the concentrations of carbon and oxygen.  Table 3. Atomic compositions (at %) derived from XPS survey spectra for the Z-0, Z-1, Z-3 and Z-5 coatings before and after acid immersion.

Samples
Element At % Before immersion Z-0, Z-1, Z-3 and Z-   Due to strong oxidation and ion permeation at high temperatures, mixed acid initiates the scission of molecular chain and degradation of the polymer network [7]. The general trend is a decrease in C content, which is related to the migration of carbonaceous species from the surface. The amount of O on the surface generally increases, due to the superficial oxidation and water adsorption upon storage [21]. The contents of C, O, and N elements of the Z-0 coating before acid immersion were 79.86%, 17.01%, and 1.80%, respectively, while after acid immersion, these changed to 77.82%, 20.47%, and 1.71%. This is a consequence of the migration of carbonaceous species and superficial oxidation. In addition, the C/O ratio of the Z-0 coating decreased from 4.69 (before acid immersion) to 3.89 (after acid immersion), suggesting that the stability and component of the Z-0 coating was strongly damaged, due to the chemical decomposition.
With the addition of 1% nano-ZrO 2 , the C and O contents after acid immersion for the Z-1 coating increase to 80.37 at % and 18.46 at%, respectively. The C/O value decreases from 4.69 (before acid immersion) to 4.30 (after acid immersion), which is obviously larger than that of the Z-0 coating, demonstrating that 1% nano-ZrO 2 can inhibit the migration of carbonaceous species and superficial oxidation to a certain degree. With an increase in the nano-ZrO 2 content, the concentrations of C and O after acid immersion for the Z-3 coating increased to 80.77 at %, and 17.31 at %, respectively. The reduction degree of the C/O value of the Z-3 coating was the smallest, decreasing from 4.69 (before acid immersion) to 4.55 (after acid immersion). The high C/O ratio was directly connected with good anti-oxidation and corrosion resistance, implying that 3% nano-ZrO 2 provided the physical barrier for corrosive media to permeate, and suppressed the migration of carbonaceous species and superficial oxidation. When the nano-ZrO 2 content reached 5%, the C content after acid immersion of the Z-5 coating reduced significantly, while the O content increased greatly. Correspondingly, the C/O value after acid immersion (3.88) decreased to the minimum, and it was even lower than that of the Z-0 coating. Excessive nano-ZrO 2 led to a decline of the anticorrosion property of coatings, and promoted the migration of carbonaceous species and superficial oxidation. In conclusion, XPS analyses showed that Z-3 coating had the best anticorrosion property. Due to strong oxidation and ion permeation at high temperatures, mixed acid initiates the scission of molecular chain and degradation of the polymer network [7]. The general trend is a decrease in C content, which is related to the migration of carbonaceous species from the surface. The amount of O on the surface generally increases, due to the superficial oxidation and water adsorption upon storage [21]. The contents of C, O, and N elements of the Z-0 coating before acid immersion were 79.86%, 17.01%, and 1.80%, respectively, while after acid immersion, these changed to 77.82%, 20.47%, and 1.71%. This is a consequence of the migration of carbonaceous species and superficial oxidation. In addition, the C/O ratio of the Z-0 coating decreased from 4.69 (before acid immersion) to 3.89 (after acid immersion), suggesting that the stability and component of the Z-0 coating was strongly damaged, due to the chemical decomposition.
With the addition of 1% nano-ZrO2, the C and O contents after acid immersion for the Z-1 coating increase to 80.37 at % and 18.46 at%, respectively. The C/O value decreases from 4.69 (before acid immersion) to 4.30 (after acid immersion), which is obviously larger than that of the Z-0 coating, demonstrating that 1% nano-ZrO2 can inhibit the migration of carbonaceous species and superficial oxidation to a certain degree. With an increase in the nano-ZrO2 content, the concentrations of C and O after acid immersion for the Z-3 coating increased to 80.77 at%, and 17.31 at %, respectively. The reduction degree of the C/O value of the Z-3 coating was the smallest, decreasing from 4.69 (before acid immersion) to 4.55 (after acid immersion). The high C/O ratio was directly connected with good anti-oxidation and corrosion resistance, implying that 3% nano-ZrO2 provided the physical barrier for corrosive media to permeate, and suppressed the migration of carbonaceous species and superficial oxidation. When the nano-ZrO2 content reached 5%, the C content after acid immersion of the Z-5 coating reduced significantly, while the O content increased greatly. Correspondingly, the C/O value after acid immersion (3.88) decreased to the minimum, and it was even lower than that of the Z-0 coating. Excessive nano-ZrO2 led to a decline of the anticorrosion property of coatings, and promoted the migration of carbonaceous species and superficial oxidation. In conclusion, XPS analyses showed that Z-3 coating had the best anticorrosion property.  Figure 7 shows the OCP variation in the nano-ZrO2-modified coatings with the immersion time. The initial OCP of the bare substrate was -485 mV vs. SCE, and this shifted towards a more cathodic  Figure 7 shows the OCP variation in the nano-ZrO 2 -modified coatings with the immersion time. The initial OCP of the bare substrate was -485 mV vs. SCE, and this shifted towards a more cathodic region with time. After 24 h of immersion, it became a steady-state value (-601 mV vs. SCE). In the beginning, the OCP values of the Z-1 coating, the Z-0 coating, the Z-5 coating, and the Z-3 coating were more positive than that of the bare substrate, and increased by 133 mV, 149 mV, 157 mV, and 184 mV, respectively, which clearly indicates the high corrosion resistance provided by these coatings [22]. With a prolonged immersion time, the OCPs of all the coatings presented a similar tendency towards lower values, arising from the diffusion of electrolytes and corrosive ions through coating defects and pinholes [23]. After 120 h of immersion, the potential difference between coatings and bare substrate decreased to 54 mV, 81 mV, 85 mV, and 101 mV for the Z-5 coating, Z-0 coating, Z-1 coating, and Z-3 coating, respectively, indicating that the protection performance of coatings deteriorates continuously with immersion time [24]. After 192 h of immersion, the OCPs of all the coatings reduced towards a negative potential to be lower than bare substrate, except for the Z-3 coating. This is probably due to the heterogeneous reactions arising from the formation of large diffusion paths within coatings [25,26] or the occurrence of some surface phenomena, such as the diffusion of chloride ions through the coating [27]. Consequently, the Z-3 coating exhibited the highest OCP value during the immersion, which demonstrates that the addition of 3% nano-ZrO 2 effectively improves the anticorrosion property of coatings. In the beginning, the OCP values of the Z-1 coating, the Z-0 coating, the Z-5 coating, and the Z-3 coating were more positive than that of the bare substrate, and increased by 133 mV, 149 mV, 157 mV, and 184 mV, respectively, which clearly indicates the high corrosion resistance provided by these coatings [22]. With a prolonged immersion time, the OCPs of all the coatings presented a similar tendency towards lower values, arising from the diffusion of electrolytes and corrosive ions through coating defects and pinholes [23]. After 120 h of immersion, the potential difference between coatings and bare substrate decreased to 54 mV, 81 mV, 85 mV, and 101 mV for the Z-5 coating, Z-0 coating, Z-1 coating, and Z-3 coating, respectively, indicating that the protection performance of coatings deteriorates continuously with immersion time [24]. After 192 h of immersion, the OCPs of all the coatings reduced towards a negative potential to be lower than bare substrate, except for the Z-3 coating. This is probably due to the heterogeneous reactions arising from the formation of large diffusion paths within coatings [25,26] or the occurrence of some surface phenomena, such as the diffusion of chloride ions through the coating [27]. Consequently, the Z-3 coating exhibited the highest OCP value during the immersion, which demonstrates that the addition of 3% nano-ZrO2 effectively improves the anticorrosion property of coatings. The typical EIS diagrams measured for various nano-ZrO2-modified coatings after 30 min, 24 h, 120 h, and 192 h immersion are presented in Figure 8. The equivalent electrical circuits shown in Figure 9 were used to fit the measured data, and the fitted lines are presented along with the measured data points. The equivalent electrical circuits are addressed in detail in the next section. The typical EIS diagrams measured for various nano-ZrO 2 -modified coatings after 30 min, 24 h, 120 h, and 192 h immersion are presented in Figure 8. The equivalent electrical circuits shown in Figure 9 were used to fit the measured data, and the fitted lines are presented along with the measured data points. The equivalent electrical circuits are addressed in detail in the next section.      The EIS diagrams for all of the coatings showed two time constants. The high frequency time constant shows the barrier properties of the coating, and the low frequency time constant corresponds to the polarization resistance of steel surface beneath the coating layer [23,[27][28][29][30]. A shift of phase angle to higher frequency region demonstrates that an increased area of coated substrate is exposed to the corrosive environment [31]. The Z-3 coating showed a high frequency phase angle in the lower frequency region, indicating its superior coating barrier property. The low frequency impedance modulus, |Z| LF (e.g., |Z| 0.01Hz ) is commonly used to roughly estimate the coating resistance [32]. The |Z| LF of the Z-0 coating presented a clear tendency to decrease with the immersion time. After 24 h of immersion, |Z| LF decreased below 10 6 Ω·cm 2 , and a horizontal line section appeared at the middle frequency, which is characteristic of the delamination of a coating [2,33]. After 192 h of immersion, the |Z| LF was even lower than 10 4 Ω·cm 2 . The |Z| LF of the Z-1 coating increased firstly and then decreased with the immersion time. Though it was lower in the beginning, the |Z| LF was much higher than that of the Z-0 coating after 192 h of immersion, which testifies that 1% nano-ZrO 2 has a limited improvement on the anticorrosion property. The |Z| LF of the Z-3 and Z-5 coatings kept decreasing with the immersion time. The |Z| LF of the Z-3 coating was about 10 8 Ω·cm 2 in the beginning, and it remained the largest among all the coatings, even after 192 h of immersion, which demonstrates that 3% nano-ZrO 2 efficiently improves the coating resistance. The |Z| LF of coating Z-5 decreased rapidly from 10 7 Ω·cm 2 , to a value lower than that of coating Z-0 after 192 h of immersion, which means that an excessive concentration of nano-ZrO 2 has a negative effect on the anticorrosion property of a coating. Therefore, when the nano-ZrO 2 content reached 3%, the anticorrosion property of coatings performed best, and then decreased with an increase in the nano-ZrO 2 content.

Electrochemically-Evaluated Corrosion Response
The EIS data were fitted by the equivalent circuit R s (Q coat (R coat (Q dl R ct ))) shown in Figure 9 using ZsimpWin [34]. In these circuits, R s is the solution resistance, Q coat is the constant phase element (CPE) of a coating, R coat is the resistance of the electrolyte in the coating pores, Q dl is the CPE of the electrical double layer, and R ct is the charge transfer resistance.
The CPE was substituted for pure capacitance, due to surface heterogeneities, deviation from capacitive behavior, and dispersion effects [35]. The impedance of CPE can be written as: where Q is the CPE constant, j= √ −1, ω is the angular frequency (rad/s), and α is a CPE exponent associated with the surface heterogeneity or roughness [36]. When α < 1, the CPE parameter (Q) cannot represent the capacitance, and the effective capacitance (Ceff ) associated with the CPE can therefore be expressed as [37]: Table 4 presents the electrical element parameters obtained from fitting the measured EIS data in Figure 8. The change in electrical element parameters reflects the change in the electrochemical properties of the coated system [38]. It is generally accepted that the increase in Ceff coat with time is related to the water uptake [39] and the dielectric constant [40] of the coating; R coat is attributed to the electrical resistance to ionic transfer through the coating pores, which reflects the porosity of coatings and the anti-penetrating ability to the electrolyte solution [31,41]; R ct is used to specify the delamination of the top coat and the onset of substrate corrosion [42], which is inversely proportional to the surface area of the sample [43]. In general, a good coating system is characterized by high resistances (R coat and R ct ), lower capacitances (Ceff coat ) [44]. The plots in Figure 10a-c clearly show the variation trends of Ceff coat , R coat and R ct for the coatings versus the immersion time. The Ceff coat of all of the coatings was very low (less than 10 −8 ), indicating the insulating nature of the coatings [28]. At the beginning of the test, all of the coatings showed high resistances (R coat and R ct ) and low Ceff coat , except for the Z-1 coating. The Ceff coat value of the coatings increased in the following order: Z-3 coating, Z-5 coating, Z-0 coating, and Z-1 coating. This shows that the water resistance of coatings does not increase with the nano-ZrO 2 content. For the 1% nano-ZrO 2 content, the nano-ZrO 2 particle tended to sink into the coating due to its high density during the curing process, and larger micro-channels emerged on the top of the coating in the initial stage, which led to the Z-1 coating having the lowest resistances (R coat and R ct ) and highest Ceff coat . When the nano-ZrO 2 content increased to 3%, nano-ZrO 2 particles were able to uniformly distribute in the coating and the hole sealing effect had an absolute advantage over agglomeration, which effectively improved the resistances (R coat and R ct ) and led to a decline in Ceff coat in the Z-3 coating. As the nano-ZrO 2 content reached 5%, the dominant status of the hole-sealing effect in the nano-particles was challenged by aggregation, which led to a decline in resistance (R coat and R ct ) and an increase in Ceff coat of the Z-5 coating to some degree. The plots in Figure 10a-c clearly show the variation trends of Ceffcoat, Rcoat and Rct for the coatings versus the immersion time. The Ceffcoat of all of the coatings was very low (less than 10 −8 ), indicating the insulating nature of the coatings [28]. At the beginning of the test, all of the coatings showed high resistances (Rcoat and Rct) and low Ceffcoat, except for the Z-1 coating. The Ceffcoat value of the coatings increased in the following order: Z-3 coating, Z-5 coating, Z-0 coating, and Z-1 coating. This shows that the water resistance of coatings does not increase with the nano-ZrO2 content. For the 1% nano-ZrO2 content, the nano-ZrO2 particle tended to sink into the coating due to its high density during the curing process, and larger micro-channels emerged on the top of the coating in the initial stage, which led to the Z-1 coating having the lowest resistances (Rcoat and Rct) and highest Ceffcoat. When the nano-ZrO2 content increased to 3%, nano-ZrO2 particles were able to uniformly distribute in the coating and the hole sealing effect had an absolute advantage over agglomeration, which effectively improved the resistances (Rcoat and Rct) and led to a decline in Ceffcoat in the Z-3 coating. As the nano-ZrO2 content reached 5%, the dominant status of the hole-sealing effect in the nano-particles was challenged by aggregation, which led to a decline in resistance (Rcoat and Rct) and an increase in Ceffcoat of the Z-5 coating to some degree.  Table 4).
With a prolonged immersion time, there was a prominent decrease in the resistances (Rcoat and Rct) and an increase in Ceffcoat for most coatings, due to the development of pathways. In particular, the Ceffcoat of the Z-1 coating decreased firstly and then increased with an increased immersion time, while the resistances (Rcoat and Rct) increased and then decreased. This implies that 1% nano-ZrO2 takes effect when the corrosive medium permeates into the lower part of the Z-1 coating. Additionally, the hole-sealing effect of insoluble corrosion products [45] should not be neglected, as confirmed in the shift of the Rcoat value.
After 192 h of immersion, the significant decrease in the resistance and increase in Ceffcoat indicated that the corrosion process had occurred. With an increase in the nano-ZrO2 content, the corrosion process was inhibited firstly and then promoted, and the inhibition effect of 3% nano-ZrO2 was the most significant. The Ceffcoat, Rcoat and Rct of the Z-3 coating were 0.25, 31.90 and 9.86 times that of coating Z-0, respectively, which indicates a markedly enhanced water resistance, lower porosity and larger charge transfer resistance.  Table 4).
With a prolonged immersion time, there was a prominent decrease in the resistances (R coat and R ct ) and an increase in Ceff coat for most coatings, due to the development of pathways. In particular, the Ceff coat of the Z-1 coating decreased firstly and then increased with an increased immersion time, while the resistances (R coat and R ct ) increased and then decreased. This implies that 1% nano-ZrO 2 takes effect when the corrosive medium permeates into the lower part of the Z-1 coating. Additionally, the hole-sealing effect of insoluble corrosion products [45] should not be neglected, as confirmed in the shift of the R coat value.
After 192 h of immersion, the significant decrease in the resistance and increase in Ceff coat indicated that the corrosion process had occurred. With an increase in the nano-ZrO 2 content, the corrosion process was inhibited firstly and then promoted, and the inhibition effect of 3% nano-ZrO 2 was the most significant. The Ceff coat , R coat and R ct of the Z-3 coating were 0.25, 31.90 and 9.86 times that of coating Z-0, respectively, which indicates a markedly enhanced water resistance, lower porosity and larger charge transfer resistance.
To further understand the anticorrosion mechanism of the nano-ZrO 2 -modified coatings in mixed acid solution, a schematic interpretation is depicted in Figure 11. As shown in Figure 11a, the Z-0 coating has abundant voids and defects, which provides preferential diffusion paths for corrosive species of Cl − , H + , and O 2 , etc. When a corrosive medium reaches the metal-coating interface, the following corrosion process for the substrate is proposed: To further understand the anticorrosion mechanism of the nano-ZrO2-modified coatings in mixed acid solution, a schematic interpretation is depicted in Figure 11. As shown in Figure 11a, the Z-0 coating has abundant voids and defects, which provides preferential diffusion paths for corrosive species of Cl − , H + , and O2, etc. When a corrosive medium reaches the metal-coating interface, the following corrosion process for the substrate is proposed: Anodic reaction: Cathodic reaction: The nano-ZrO2 addition did not change the reaction mechanism of the substrates. With an increase in the nano-ZrO2 content, the micro-pore channels gradually sealed, while excessive nano- To further understand the anticorrosion mechanism of the nano-ZrO2-modified coatings in mixed acid solution, a schematic interpretation is depicted in Figure 11. As shown in Figure 11a, the Z-0 coating has abundant voids and defects, which provides preferential diffusion paths for corrosive species of Cl − , H + , and O2, etc. When a corrosive medium reaches the metal-coating interface, the following corrosion process for the substrate is proposed: Anodic reaction: Cathodic reaction: The nano-ZrO2 addition did not change the reaction mechanism of the substrates. With an increase in the nano-ZrO2 content, the micro-pore channels gradually sealed, while excessive nano- To further understand the anticorrosion mechanism of the nano-ZrO2-modified coatings in mixed acid solution, a schematic interpretation is depicted in Figure 11. As shown in Figure 11a, the Z-0 coating has abundant voids and defects, which provides preferential diffusion paths for corrosive species of Cl − , H + , and O2, etc. When a corrosive medium reaches the metal-coating interface, the following corrosion process for the substrate is proposed: Anodic reaction: Cathodic reaction: The nano-ZrO2 addition did not change the reaction mechanism of the substrates. With an increase in the nano-ZrO2 content, the micro-pore channels gradually sealed, while excessive nano-: negative charge.
The nano-ZrO 2 addition did not change the reaction mechanism of the substrates. With an increase in the nano-ZrO 2 content, the micro-pore channels gradually sealed, while excessive nano-ZrO 2 caused agglomeration and produced larger micro-pore channels when corrosion occurred. The competition relationship between agglomeration and the hole-sealing effect of nano-particles was obviously visible. The addition of 1% nano-ZrO 2 partially sealed micro-pore channels of the Z-1 coating (Figure 11b), and generated an incomplete nanostructure network, which was responsible for the limited improvement of corrosion resistance. In contrast, the addition of 3% nano-ZrO 2 perfectly balanced the relationship between agglomeration and the sealing of micro-pore channels of nano-particles, and formed a relatively complete nanostructure network, which effectively improved the corrosion resistance of the Z-3 coating (Figure 11c). When the nano-ZrO 2 content reached 5%, the corrosion resistance of the Z-5 coating dramatically reduced. Nano-particles still sealed micro-pore channels, while lots of agglomerations occurred for the Z-5 coating (Figure 11d). Thus, the Z-3 coating showed the best anticorrosion property.

Conclusions
The corrosion protection characterizations for nano-ZrO 2 -modified coatings were examined in mixed acid solution. Corrosion resistance was enhance for the coatings with 1% and 3% nano-ZrO 2 contents, while it declined for the coating with 5% nano-ZrO 2 content. The 3% nano-ZrO 2 particle modified coating showed the best corrosion protection performance, as evidenced by EIS results. Visual assessments through SEM, AFM and elemental mapping observations were in accordance with the electrochemical results.
One percent nano-ZrO 2 , with an incomplete nanostructure network in the coating, showed a limited improvement to the anticorrosion property, and 5% nano-ZrO 2 led to excessive aggregation and declined the adhesion of reinforced nanoparticles, which damaged the corrosion resistance. Three percent nano-ZrO 2 possesses remarkable dispersion properties and a relatively complete nanostructure network in the coating, and was shown to balance the relationship between agglomeration and sealing of the micro-pore channels of nano-particles perfectly, which resulted in a minimum amount of diffusion of Cl, C, O and Fe, the lowest average roughness (5.94 nm), the highest C/O radio (4.55), and the best electrochemical properties (highest resistances and lowest Ceff coat ).
Author Contributions: Z.W. and W.X. designed the research. W.X. and S.W. performed the mathematical calculations. W.X., Z.W., S.W. and Q.L. discussed and analyzed the data. W.X. wrote the manuscript. Z.W. and E.-H.H. revised the paper.