Journal of Materials Science & Technology  2019 , 35 (10): 2345-2356 https://doi.org/10.1016/j.jmst.2019.05.039

Orginal Article

Corrosion kinetics and patina evolution of galvanized steel in a simulated coastal-industrial atmosphere

Chuang Qiaoab, Lianfeng Shena*, Long Haob*, Xin Mub, Junhua Dongb, Wei Keb, Jing Liuc, Bo Liubd

a College of Forestry, Henan Agricultural University, Zhengzhou, 450002, China
b Environmental Corrosion Centre of Materials, Institute of Metal Research, Chinese Academy of Sciences, Shenyang, 110016, China
c Electric Power Test and Research Institute, Guangzhou Power Supply Co. Ltd. (GZPS), Guangzhou, 510410, China
d School of Metallurgy, Northeastern University, Shenyang, 110819, China

Corresponding authors:   *Corresponding authors.E-mail addresses: shenlianfeng126@126.com (L. Shen), chinahaolong@126.com(L. Hao), jhdong@imr.ac.cn (J. Dong).*Corresponding authors.E-mail addresses: shenlianfeng126@126.com (L. Shen), chinahaolong@126.com(L. Hao), jhdong@imr.ac.cn (J. Dong).*Corresponding authors.E-mail addresses: shenlianfeng126@126.com (L. Shen), chinahaolong@126.com(L. Hao), jhdong@imr.ac.cn (J. Dong).

Received: 2019-02-22

Revised:  2019-04-29

Accepted:  2019-05-5

Online:  2019-10-05

Copyright:  2019 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

The corrosion kinetics and patina (corrosion products) layer evolution of galvanized steel submitted to wet/dry cyclic corrosion test in a simulated coastal-industrial atmosphere was investigated. The results show that zinc coating has a greater corrosion rate during the initial period and a lower corrosion rate during the subsequent period, and the patina composition and structure can greatly affect the corrosion kinetics evolution of zinc coating. Moreover, Zn5(OH)6(CO3)2 and Zn4(OH)6SO4 are identified as the main stable composition and exhibit an increasing relative amount; while Zn12(OH)15Cl3(SO4)3 cannot stably exist and diminish in the patina layer as the corrosion develops.

Keywords: Zinc ; Galvanized steel ; EIS ; Atmospheric corrosion ; Patina layer

0

PDF (5257KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Chuang Qiao, Lianfeng Shen, Long Hao, Xin Mu, Junhua Dong, Wei Ke, Jing Liu, Bo Liu. Corrosion kinetics and patina evolution of galvanized steel in a simulated coastal-industrial atmosphere[J]. Journal of Materials Science & Technology, 2019, 35(10): 2345-2356 https://doi.org/10.1016/j.jmst.2019.05.039

1. Introduction

Galvanized steel is widely used in main infrastructure constructions, such as power transmission towers, outdoor telecommunications towers and so on [1,2]. The enhanced corrosion resistance of galvanized steel in the atmospheric environment is brought by the surface deposited zinc coating, which is usually produced by the hot-dip process. The protective effect of zinc coating in resisting atmospheric corrosion of steel mainly lies in two aspects. Zinc coating has a more negative corrosion potential than steel substrate, and during the corrosion process, it preferentially corrodes and behaves as the anode, contributing to the retarded corrosion of steel substrate. In addition, the zinc coating corrosion product is usually of a complex structure [3,4] with a much lower conductivity, and it can greatly lower the electron transportation during electrochemical corrosion reactions and/or physically prohibit the migration of corrosive ions to the underlying substrate. Therefore, the combined chemical and physical characteristics of zinc coating and its corrosion product have always made galvanized steel as one of the most commonly employed structural material.

In fact, the life-span of galvanized steel is closely related to its corrosion evolution behavior in a given atmospheric environment [[5], [6], [7]]. Firstly, daily cyclic change in ambient humidity causes the galvanized steel surface to be in a state of wetting and drying alternation [8,9]. Electrochemical reactions including zinc coating dissolution and reduction of depolarizers (O2 and/or H+ in an acid environment) will control the corrosion process of galvanized steel under a thin liquid film and give birth to corrosion products [10]. Those corrosion products will accumulate at the zinc coating surface to form the patina layer as the corrosion proceeds. Then, characteristics of the patina layer will play a decisive role in the subsequent corrosion period [3,4,[11], [12], [13], [14], [15], [16]], i.e., both the structure and composition of the patina layer can affect the migration of ions from the electrolyte to zinc coating [3,4,[11], [12], [13], [14]] or even the underlying steel substrate [11,15,16]. In atmospheric environment, Zn5(OH)6(CO3)2 has been found to be protective [1,12] and regarded as the most common corrosion product in all environments due to the effect of carbon dioxide, and besides a large number of amorphous phases, such as ZnO and Zn(OH)2 also existed [1,17]. Zn4(OH)6SO4 has been characterized as the corrosion product in the industrial environment after years of exposure [18]. Zn4Cl2(OH)4SO4 is found in some industrial zones [19], where the deposition of Cl- is significant but not dominant. NaZn4Cl(OH)6SO4·6H2O [20] has been identified as the final corrosion product in the marine atmospheric environment with Zn5(OH)8Cl2·H2O generating in the intermediate steps [18]. In addition, Zn4(OH)6SO4 and Zn5(OH)6(CO3)2 were suggested to have the ability in repelling Cl- [12], and NaZn4Cl(OH)6SO4·6H2O was proved to have corrosion inhibiting effect in marine environment [21]. Furthermore, the zinc coating patina layer with an intact and compact structure was proposed to have anti-corrosion effect in blocking the penetration tunnel of corrosive electrolyte [3,4]. Therefore, available research of zinc atmospheric corrosion has mainly focused on the initial corrosion behavior [7,13,[22], [23], [24]], patina composition [10,[17], [18], [19], [20],25], and the empirical correlation between corrosion rate and environmental parameters [5,26].

In practice, the complexity of zinc corrosion process and the great diversity of corrosion product composition make that the corrosion evolution predicting research of galvanized steel is of great challenges. Furthermore, the diversified atmospheric environment of galvanized steel exposed to has also made the corrosion data be different from the historical one [25,26], especially under the combined effect of SO2, Cl-, high temperature and relative humidity (RH). Although field exposure tests have been extensively carried out worldwide [[25], [26], [27]] to assess the given environment corrosivity and materials resistance, the unfriendly, complicated, and uncontrolled environmental parameters have contributed to various opinions on corrosion characteristics of galvanized steel. Wet/dry cyclic corrosion test (CCT), under controlled environmental conditions, can simulate the atmospheric corrosion of galvanized steel and have been increasingly used in the atmospheric corrosion studies [28,29].

In the present investigation, the corrosion kinetics and patina characteristics evolution of galvanized steel exposed to a simulated coastal-industrial atmospheric environment has been carried out using CCT method, paying special focus on the relevance of corrosion kinetics evolution to the patina layer characteristics changing, such as surface and cross-sectional morphology, composition, corrosive element distribution, and electrochemical properties. Such kind of evolution information is vital for a comprehensive understanding of galvanized steel corrosion and for the accurate predicting how it will degrade in a given environment.

2. Experimental

2.1. Wet/dry cyclic corrosion test

The sample material was hot-dip galvanized steel with Q345 carbon steel as the substrate, and the zinc coating was about 100 μm in thickness including a Zn-Fe alloy layer at about 85-90 μm. The specimens were sectioned into rectangular coupons and then ultrasonically cleaned in alcohol. After being cleaned, samples for CCT tests were masked by 3M adhesive tape with an exposure area of 30 mm × 20 mm, and samples for electrochemical tests were encapsulated by silicone rubber with an exposure area of 10 mm × 10 mm. Before applied to CCT test, all samples were stored in a desiccator for 24 h and then recorded the initial weight of each sample. The CCT test is essentially a repeated process of wetting the sample surface by a prepared corrosion electrolyte and drying the sample in an atmosphere-simulating chamber under controlled temperature and RH. Each run of wetting and drying process is denoted as one CCT cycle and then continues to predefined cycle numbers. During the CCT test, the galvanized steel surface was wetted by the corrosion electrolyte of 40 μL cm-2 and dried at 25 °C and 80% RH for 12 h. The corrosion electrolyte is a mixture solution of 16.48 mg L-1 NaCl + 95.95 mg L-1 Na2SO3 + 546.69 mg L-1 NaHCO3 with pH 5.77 ± 0.02, based on the rainfall analysis in Guangzhou city, which locates at the south-eastern coastal line of China. The detailed experimental procedure for the present CCT test is available at Ref. [30].

At the end of 10, 20, 40, 60, 90, and 120 CCT, respectively, three paralleled samples were taken for thickness loss measurement. According to ISO 8407:2014, the patina layer was removed by ultrasonically cleaning in saturated CH3COONH4 solution at 80 °C for 10 min. The used balance for weight loss measurement was Sartorious BS 224 S analytical balance with a precision of 0.1 mg. The thinning value and the average corrosion rate of surface zinc coating were calculated by formulas Eqs. (1) and (2), respectively:

X =$\frac{ΔW}{ρS }$×10000(1)

V =$\frac{ΔW}{ρSN }$×10000 (2)

where X is the thinning value of zinc coating (μm); ΔW is the weight loss (mg), ρ is the density of the zinc coating (7140 mg cm-3); S is the uncovered surface area (cm2); V is the average corrosion rate (μm CCT-1) and N is the measured CCT cycle.

2.2. Patina layer analysis

The surface and cross-sectional morphologies of corroded galvanized steel were characterized by scanning electron microscopy (SEM, Inspect™ F, produced by FEI Company), equipped with the energy dispersion spectrum (EDS) analysis. For cross-sectional morphology characterization, the corroded galvanized steel sample was encapsulated in epoxy resin in priority, cut into a small sample of 20 mm × 10 mm × 5 mm, and then ground by 120# grade emery paper to a much smaller size of 12 mm × 10 mm × 5 mm in avoiding the damage of patina layer from the cutting operation. After that, the sample was ground to 2000# grade emery paper step by step and then polished by 1.5 μm particle size diamond paste in oil. The corroded sample for cross-sectional analysis was also observed using electron probe microanalyzer (EPMA, JXA-8530 F) for detecting the distribution of Cl and S elements at the test condition of U =15 kV, I =100 nA.

The patina composition was determined by multiple methods including synchrotron radiation X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), and Raman spectra. Before patina composition detection, the patina layer was carefully scraped from the corroded zinc coating samples and then ground into fine powders. The XRD patterns were obtained at beamline BL14B1 of the Shanghai Synchrotron Radiation Facility (SSRF) at a wavelength of 0.689 Å. The detailed information about beamline BL14B1 can be found in Ref. [31]. Compared with the conventional X-ray, the X-ray beam from the electron storage ring is of several orders of magnitude higher in intensity and brightness. The advantages in high intensity, brightness, collimation, and stability of X-ray beam promote the diffraction data collection with increased resolution and facilitate the analysis in phases of minor amount [32]. Transmission geometry with glass capillary sample holder of 0.5 mm in diameter [33] was used, and the powdered patina amount for each test was kept the same. In this condition, the peak intensity of a certain phase can semi-quantitatively reflect its relative content [30].

For FTIR analysis, the powdered patina was mixed with pure anhydrous KBr and pressed to pellets using a hydraulic press. Magna-IR 560 infrared spectrophotometer was used to measure the spectra of the powdered patina in the ranges from 400 cm-1 to 2000 cm-1 with an accuracy of 4 cm-1. Raman spectra were also used to analyze the as-corroded zinc coating patina layer on galvanized steel after 10 CCT. At least five paralleled test points were selected for the spectra characterization. The Raman spectra data were collected in a shift range from 100 cm-1 to 1200 cm-1 using the 632.8 nm line of a Melles Griot 35 mW HeNe laser with 60 s collecting time for each scan.

2.3. Electrochemical measurements

Electrochemical measurements were carried out employing the classic three electrodes system by Gamry Reference 600 electrochemical workstation. Platinum with 4 cm2 area and saturated calomel electrode (SCE) were selected as the counter electrode and the reference electrode, respectively. The un-corroded/corroded galvanized steel samples with an exposed area of 1 cm2 were taken as working electrodes. The polarisation data were recorded in a range of -0.2 V to +0.2 V vs. open circuit potential with a constant scan rate of 10 mV min-1, and then linear fitting method was used to obtain the electrochemical parameters of βa, βc, icorr, and Ecorr (Tafel slope of the anodic reaction and the cathodic reaction, corrosion current density, and corrosion potential) of each sample at the weak polarization region (±20 mV vs. Ecorr). The frequency range for EIS was from 100 kHz to 10 mHz with a 5 mV amplitude signal at open circuit potential, and the measured EIS data were fitted using the commercial software “Zsimpwin”. Before measurement, each sample was immersed in test electrolyte for 1200s to reach a steady state, and all the measurements were carried out at room temperature (25 °C ± 1 °C) in the atmosphere-simulating electrolyte. In addition, no observation of corrosion products falling off and no silicone rubber detachment were seen during the whole immersed process for each electrochemical test.

3. Results and discussion

3.1. Corrosion kinetics

The corrosion kinetics evolution including thickening data and average corrosion rate of galvanized steel exposed to the simulated coastal-industrial atmospheric environment as a function of CCT cycle is shown in Fig. 1 [34], with Fig. 1(a) showing the linear coordinates plot and Fig. 1(b) showing the log-log coordinates plot. Obviously, Fig. 1(a) shows that the zinc coating gradually thickens as the corrosion develops, and the calculated average corrosion rate by Eq. (2) gradually decreases as the corrosion develops. Fig. 1(b) indicates that the whole corrosion process can be divided into two periods with the fact that the slope of the first period from 0 to 60 CCT is higher than that of the second period from 60 to 120 CCT. Usually, the atmospheric corrosion kinetics of metals can be fitted with the power function law as shown in Eq. (3) [35].

ΔW=ANn (3)

where A and n are the constants. A is considered as a measure of the initial corrosion resistance of the sample, and n reflects the feature of the corrosion kinetics. Meanwhile, an expression of the average corrosion rate in logarithmic coordinates could be obtained by dividing the two sides of Eq. (3) by the test period N, as shown in Eq. (4) [35].

log$\frac{ΔW}{ N }$ =logA+(n-1)logN (4)

Fig. 1.   Thickness loss and average corrosion rate of galvanized steel in simulated coastal-industrial atmosphere as a function of CCT cycle: (a) linear coordinates plot; (b) log-log coordinates plot [34].

Clearly, Eq. (4) shows that the average corrosion rate has a linear changing relationship with N in log-log coordinates. Furthermore, n < 1 indicates a deceleration in corrosion rate; n > 1 indicates an acceleration in corrosion rate; n = 1 indicates that the corrosion rate keeps as a constant [33,36], and a smaller n value indicates a lower corrosion tendency during the period. The linear fitting expression of the calculated average corrosion rate for the first and second period during the whole corrosion process is also shown in Fig. 1(b) as follows:

logV = -0.86 - 0.44 log N (0≤N<60) (5)

logV =-0.21 - 0.80 log N (60≤N≤120) (6)

Obviously, the n-1 = -0.80 of the second period is much smaller than the n-1 = -0.44 of the first period, indicating that the corrosion tendency of the zinc coating during the second period is lower than it is during the first period. The two n values are both smaller than 1, indicating a deceleration in the corrosion rate of zinc coating during the whole corrosion process. In addition, an expression of instantaneous corrosion rate Vi (μm CCT-1) can be obtained by the differential of Eq. (3), as shown in Eq. (7):

Vi = $\frac{dΔW}{ dN }$ = AnN n-1 (7)

Based on the fitting results of n and A, Fig. 2 shows the evolution behaviour in the calculated instantaneous corrosion rate of galvanized steel exposed to the simulated coastal-industrial atmospheric environment. Clearly, the zinc coating corrodes in a decreasing corrosion rate during the whole corrosion process, and the corrosion rate during the first corrosion period is higher than that during the second corrosion period after 60 CCT. Such kind of evolution behaviour is closely related to the characteristics of the surface formed patina layer [3,4,[11], [12], [13],15,16].

Fig. 2.   Evolution of instantaneous corrosion rate of galvanized steel in simulated coastal-industrial atmosphere as a function of CCT cycle.

3.2. Cross-sectional and surface morphologies of the patina layer

The evolution of cross-sectional morphologies of the patina layer on galvanized steel is shown in Fig. 3. Clearly, the patina layer formed at 10 CCT in Fig. 3(a) is loose, thin, and discontinuous with very thin or almost no patina covering at some locations. As the corrosion prolongs to 20 CCT, although the patina layer in Fig. 3(b) is not compact and has obvious cracks, it has thickened in thickness and covered the whole coating surface. Fig. 3(c) indicates that the patina layer at 40 CCT is more compact and of fewer cracks or pores than the patina layer at 20 CCT in Fig. 3(b). For the patina layer at 60 CCT in Fig. 3(d), it is compact and smooth to underlying zinc coating with 10 μm thickness. When the corrosion prolongs to 90 CCT in Fig. 3(e), the patina layer is in the same thickness and almost has no difference with that of 60 CCT, indicating that the patina layer has entered into a stable state. As to the patina layer at 120 CCT in Fig. 3(f), it also has almost the same thickness as that of 60 CCT and 90 CCT. However, apparent pitting corrosion has been observed under the patina layer as indicated in Fig. 3(f).

Fig. 3.   Evolution in the cross-sectional morphology of the patina layer on galvanized steel as a function of CCT cycle: (a) 10 CCT, (b) 20 CCT, (c) 40 CCT, (d) 60 CCT, (e) 90 CCT, (f) 120 CCT.

In fact, the compactness, thickness, and adherence of zinc coating patina layer are closely related to the corrosion kinetics of galvanized steel [7,14,18]. It is believed that an adherent and compact patina layer can generally contribute to a higher corrosion resistance of the corrosion galvanized steel because of the physical restriction ability to corrosive electrolyte provided by the patina layer [3,4]. During the first corrosion period at 10 CCT, the corrosion product cannot completely cover the zinc coating surface, and the patina layer is loose, thin, porous, and thus of less protective ability. The penetration of aggressive ions, Cl-, SO42-, and dissolved oxygen cannot be effectively hindered, resulting in a greater corrosion rate as shown in Fig. 2. As the corrosion prolongs to 60 CCT, the patina layer gradually grows in thickness and becomes compact and adherent to the underlying zinc coating, contributing to the gradually enhanced corrosion resistance [4], and thus a decreasing corrosion rate. During the second corrosion period from 60 CCT to 120 CCT, the surface patina layer has entered into a stable state with almost the same thickness, and such a stable and corrosion resistant patina layer can protect the galvanized steel from severe corrosion attack. So, the corrosion rate during this period maintains at a lower value and the thickness loss kinetics from 60 CCT to 120 CCT in Fig. 1 increases at a lower rate. As to the observed corrosion pits for the 120 CCT sample in Fig. 3(f), it may lie in the existence of tiny cracks in the patina layer, facilitating the penetration of Cl-, SO42- and dissolved oxygen to the patina/zinc coating interface. Therefore, the evolution of cross-sectional morphologies of the patina layer on galvanized steel in Fig. 3 is in accordance with its corrosion kinetics in Fig. 1, Fig. 2.

Fig. 4 shows the surface morphology characterization of the corrosion product on zinc coating observed by SEM and EDS during the corrosion process of galvanized steel. Fig. 4(a) illustrates the surface morphology of corrosion product at 10 CCT, and three kinds of morphologies with different shapes of nodular-like, lamellar-like and filamentary-like, respectively have been observed as indicated by a, b, and c. In addition, the red dotted ring indicates the residual zinc coating with almost no patina covering. Fig. 4(b) gives the surface morphology of corrosion product at 60 CCT. Obviously, the whole zinc coating surface has been covered with the corrosion product, and the amount of both nodular-like and lamellar-like corrosion product increases, while the amount of filamentary-like corrosion product greatly decreases. When the corrosion prolongs to 120 CCT, the whole zinc coating surface has been covered with a thick and compact patina layer as indicated in Fig. 4(c). Usually, the particular shape and morphology of certain corrosion product on zinc coating in an atmospheric environment has a close relationship with its composition and compound type [15,22,24]. EDS has been employed to identify the compositional elements of the three kinds of corrosion products. Fig. 4(a') corresponds to the result of the nodular-like product, consisted of Zn, C, and O elements; Fig. 4(b') corresponds to the result of the lamellar-like product, consisted of Zn, S, and O elements; Fig. 4(c') corresponds to the result of the filamentary-like product, consisted of Zn, S, Cl, and O elements. To further clarify the patina composition and its evolution under the combined effect of Cl- and SO2, synchrotron radiation XRD, Raman spectra, and FITR have been used to characterize the powdered corrosion product.

Fig. 4.   Surface morphology observation and EDS characterization of the patina layer on galvanized steel as a function of CCT cycle: (a) 10 CCT, (b) 60 CCT, (c) 120 CCT; (a′), (b′) and (c′) corresponding to EDS spectra of the patina labelled as a, b, and c, respectively in Fig. 4(a).

3.3. Patina composition

Fig. 5 shows the synchrotron radiation XRD patterns of the powdered patina on galvanized steel exposed to the simulated coastal-industrial atmospheric environment as a function of CCT cycle. Clearly, phases of Zn5(OH)6(CO3)2 (JCPDS No. 194158), Zn4(OH)6SO4 (JCPDS Nos. 440673, 440674, 440675), and Zn12(OH)15Cl3(SO4)3 (JCPDS No. 420152) have been detected. As to Zn12(OH)15Cl3(SO4)3, its intensity at 10 CCT is strong and gradually decreases as CCT cycle increases and after 60 CCT the intensity cannot be detected, indicating the formation of a large amount of Zn12(OH)15Cl3(SO4)3 at the initial corrosion period and its diminishing during the second corrosion period. As to Zn5(OH)6(CO3)2 and Zn4(OH)6SO4, each of the intensity increases a little as the CCT cycle increases to 60 CCT and then almost keeps at a steady level during the subsequent CCT process, although the intensity of each phase does not dramatically evolve during the whole corrosion process compared with that of Zn12(OH)15Cl3(SO4)3. The non-obvious evolution of amounts of Zn5(OH)6(CO3)2 and Zn4(OH)6SO4 may also indicate that the zinc coating patina layer has entered into a stable state during the second period. Zn12(OH)15Cl3(SO4)3 has been identified by Odnevall and Leygraf [38], and is considered as corrosion product composition exclusively existed in urban and industrial environments as result of human activity. Its identification in this investigation relates to the atmosphere-simulating electrolyte with dominant SO42- content.

Fig. 5.   Synchrotron radiation XRD patterns of the powdered corrosion product on galvanized steel in simulated coastal-industrial atmosphere as a function of CCT cycle.

Based on the surface morphology characterization and EDS analysis, as well as the synchrotron radiation XRD results, it is believed that in Fig. 4 the nodular-like product is Zn5(OH)6(CO3)2, the lamellar-like product is Zn4(OH)6SO4 [37], and the filamentary-like product is considered as Zn12(OH)15Cl3(SO4)3. Besides, both the amount of Zn5(OH)6(CO3)2 and Zn4(OH)6SO4 increases and the amount of Zn12(OH)15Cl3(SO4)3 decreases as the corrosion develops.

Fig. 6 shows the FTIR spectra of the powdered patina on galvanized steel exposed to the simulated coastal-industrial atmospheric environment as a function of CCT cycle. All of the test samples show the similar absorption spectrum. Clearly, the most prominent bands in the spectra are due to CO32- and SO42- from Zn5(OH)6(CO3)2 and Zn4(OH)6SO4, respectively. The bands at 833 cm-1 and 707 cm-1 belong to the out-of-plane and in-plane vibrations of CO32- in Zn5(OH)6(CO3)2 [25], while the bands at 1513 cm-1 and 1388 cm-1 correspond to the asymmetric stretching vibrations of CO32- in Zn5(OH)6(CO3)2 [39]. The absorption bands at 1120 cm-1, 1047 cm-1 and 960 cm-1 are identified as the vibration of SO42- [40] and the band at 476 cm-1 may associate with Zn(OH)2 as the Zn-O bond corresponds in 350-600 cm-1 [7].

Fig. 6.   FTIR characterization of the powdered corrosion product on galvanized steel in simulated coastal-industrial atmosphere as a function of CCT cycle.

Raman spectrum has also been used to further characterize the as-corroded zinc coating of galvanized steel submitted to the simulated coastal-industrial atmospheric environment after 10 CCT, and the result is shown in Fig. 7. The phase identification of corrosion products is obtained by comparing Raman shift (cm-1) with the reference data published in literature as provided in Table 1. Typically, SO42- has four kinds of vibration modes: v1 and v2 represent the symmetric stretching and deformation vibrations, and v3 and v4 represent the asymmetric stretching and deformation vibrations, respectively [41]. Besides, the vibration frequency of SO42- is also influenced by the reduced symmetry when it is bounded into a structure. The fundamental vibration frequencies for SO42- of reduced symmetry have been reported to occur in a range of values, viz., v1, 971-993 cm-1; v2, 445-490 cm-1; v3, 1070-1190 cm-1; v4, 613-648 cm-1 [42,43]. In the Raman spectrum in Fig. 6, the peaks at 451 cm-1, 472 cm-1, 620 cm-1, 1150 cm-1 can be attributed to v2, v2, v4, v3 vibration of SO42-, respectively. By comparison with vibration data from the literature, the peaks at 472 cm-1 can be assigned to ZnSO4 [44], while peaks at 451 cm-1, 601 cm-1, 620 cm-1 and 1150 cm-1 can be attributed to Zn4(OH)6SO4 [44,45]. The intense sharp peaks at 736 cm-1, 980 cm-1, and 1078 cm-1 give the presence of Zn5(CO3)2(OH)6 [44,46]. The peaks at 736 cm-1, 1078 cm-1 can be boiled down to v4, v1 vibration of CO32- in Zn5(OH)6(CO3)2 [44,46], and the peak at 980 cm-1 is assigned to the δ OH deformation mode [46]. The medium peaks at 574 cm-1 and 380 cm-1 can be assigned to ZnO [47], and the intensity peaks at 150 cm-1 and 210 cm-1 confirm the presence of Zn(OH)2 [48]. Moreover, the phase of Zn5(OH)8Cl2·H2O is found with a prominent peak at 910 cm-1 [44,45]. Therefore, for the as-corroded zinc coating after 10 CCT, phases of ZnO, Zn(OH)2, ZnSO4, Zn5(OH)6(CO3)2, Zn4(OH)6SO4, and Zn5(OH)8Cl2·H2O have been detected in the zinc coating corrosion product by Raman spectrum.

Fig. 7.   Raman spectra characterization of the powdered corrosion product on galvanized steel in simulated coastal-industrial atmosphere at 10 CCT cycle.

Table 1   Raman shift (cm-1) data for zinc corrosion products [[45], [46], [47], [48], [49]].

ZnOZn(OH)2ZnSO4Zn5(OH)6(CO3)2Zn4(OH)6SO4Zn5(OH)8Cl2·H2O
[48][49][45][45,47][45,46][45,46]
101395
150
210218227210
250244231255
278272267
370346
380388395
407380385403
421
437445451467
472
574507543
583610601
626636620
720707730
760736
985980965910
103010221005
1050107110621065
1080108410781130
1150

New window

Usually, Zn5(OH)8Cl2·H2O has a lower thermodynamic stability compared with Zn4(OH)6SO4 [5,18]. The structural resemblance between the phases also permits a gradual transformation from the chloride sequence to sulphate sequence [17]. In the present simulated atmosphere with a higher SO2 content and a lower NaCl content, Zn5(OH)8Cl2·H2O may only exist as an intermediate product with poor crystallization as only detected by the Raman spectrum. As to the crystallization of ZnO, Zn(OH)2, and ZnSO4, each of them is also not well defined. The detection of Zn5(OH)6(CO3)2 and Zn4(OH)6SO4 by synchrotron radiation XRD confirms the good crystallization of the two phases. In addition, Zn12(OH)15Cl3(SO4)3 was only detected by synchrotron radiation XRD at the initial corrosion period, and it was absent during the second corrosion period, indicating its instability under long-term atmospheric corrosion condition, and its existence in zinc coating patina layer is closely related to the corrosion rate evolution of galvanized steel. Therefore, the combination of synchrotron radiation XRD, Raman spectra, and FTIR gives a full identification of the zinc coating atmospheric corrosion product.

3.4. EPMA characterization of the patina layer

The above discussion indicates that corrosion evolution kinetics of galvanized steel is closely related to the characteristics of the surface formed patina layer. The higher corrosion rate reflects the formation of a loose and un-compact patina layer in a growing status, and a lower corrosion rate reflects the formation of a continuous and compact patina layer in a stable status. To further illustrate the participation of SO42- and Cl- in the formation of corrosion product and the effect of patina layer in resisting the ingression of SO42- and Cl-, EPMA has been employed to detect the distribution of S and Cl elements in the patina layer as a function of CCT cycle, and the results are shown in Fig. 8. Clearly, for the samples at 10 CCT and 20 CCT during the first corrosion period with a higher corrosion rate, S and Cl elements distributed all over the zinc coating patina layer and accumulated at the certain surface and bottom sites, indicating the formation of S and/or Cl--containing corrosion product covered the whole coating surface. As the patina layer during the first corrosion period is thin, loose, and porous, and the barrier effect is poor, SO42- and Cl- can easily get through, leading to the accumulation of S and Cl elements at the bottom of certain sites [25,26]. As the corrosion prolongs to 60 CCT at the beginning of the second corrosion period, the patina layer has grown in thickness and become continuous and compact, and the corrosion of galvanized steel has reached a stable state. EPMA mapping of the patina layer shows the presence of only S element, at the crack bottom with penetration tunnel for SO42-, and the almost absence of Cl element in the patina layer. Therefore, EPMA mapping shows Cl and S presence in the patina layer at 10 CCT and 20 CCT, but Cl absence and S presence in the patina layer at 60 CCT, indicating the gradual vanishing of Cl--containing corrosion product and the stable existence of SO42--containing corrosion product. Based on the synchrotron radiation XRD and Raman spectra, EPMA results also confirm the gradual vanishing of Zn12(OH)15Cl3(SO4)3 and the stable existence of Zn4(OH)6SO4 in the patina layer during the whole corrosion process. When the corrosion prolongs to 120 CCT, corrosion pits at the patina layer bottom and cracks in the patina layer, caused by the inner stress triggered by long-term wet/dry cyclic corrosion process and transformation of phases from less stable one to more stable one, were clearly observed, facilitating the penetration of Cl- and SO42- through the patina layer. Correspondingly, EPMA mapping of the patina layer at 120 CCT shows the presence of accumulated S and Cl elements, which locate at the bottom of the patina layer with the existence of tinny cracks.

Fig. 8.   EPMA mapping of S and Cl elements in the patina layer on galvanized steel in simulated coastal-industrial atmosphere as a function of CCT cycle.

Therefore, EPMA mapping indicates that a continuous and compact patina layer of minimum defects such as cracks or voids is necessary for protecting the galvanized steel from further corrosion [4,13]. Besides, the presence of S element throughout the whole corrosion process and the gradual vanishing and again the presence of Cl element gives the conclusion that both S and Cl elements participate in the formation of corrosion products. The SO42--containing corrosion product Zn4(OH)6SO4 can stably exist in the long-term corrosion of zinc coating. However, the Cl--containing corrosion product Zn12(OH)15Cl3(SO4)3 is only detected during the initial corrosion process and then cannot be detected during the long-term corrosion process.

3.5. Electrochemical properties of the patina layer

Fig. 9 gives the evolution of potentiodynamic polarization curves of galvanized steel after suffering different CCT cycles measured in the atmosphere-simulating electrolyte at pH of 5.77 ± 0.02, and the un-corroded galvanized steel is used for comparison. The obtained electrochemical parameters for each sample are shown in Table 2. In general, Fig. 9 shows that both cathodic current density and anodic current density decrease as the CCT cycle increases, with the fact that the cathodic process was dominated by the reduction of H+ and/or dissolved oxygen in the electrolyte and the anodic process was dominated by the dissolution of the galvanized steel. Moreover, the obtained icorr shows a decreasing tendency with the increase of CCT cycle as indicated in Table 2. In addition, the polarization curves comparison between corroded and un-corroded galvanized steel sample indicates that the formation of patina layer on galvanized steel greatly affects the cathodic reduction rate and anodic dissolution rate. As to the un-corroded galvanized steel, it can also be found that the limiting diffusion feature of oxygen reduction is not obvious with the existence of H+ reduction in the cathodic branch curve, and the formation and existence of a very thin patina layer on galvanized steel surface due to the high corrosion tendency of zinc under natural atmospheric environment may be responsible for this. But, the gradually decreased icorr and increased βa and βcof the sample with higher CCT cycle relates to the gradually increased compactness and lower conductivity of the zinc coating patina layer as the difficulty of depolarizer reduction process enhances. As to the anodic dissolution process, the 20 CCT sample shows no obviously lower anodic current density and higher βa than the un-corroded 0 CCT sample, indicating the almost non-protective of the formed patina layer at 20 CCT. When the corrosion prolongs to 40 CCT, the anodic current density decreases and βa increases compared with that of the 20 CCT sample. Also, a further decrease of anodic current density and βa increase was observed for the 60 CCT sample. The gradual decrease in icorr and increase in βa of samples from 0 CCT to 60 CCT during the first corrosion period reflects the gradually enhanced protective ability of the patina layer. And a further decrease in icorr and increase in βa for the 90 CCT sample compared to the 60 CCT sample also indicates the further decreased corrosion rate and improved corrosion resistance of the 90 CCT sample. However, the 120 CCT sample has a little higher icorr than the 90 CCT sample, indicating the dissolution of new zinc coating due to the penetration of free SO42- and Cl- through the cracks in the patina layer.

Fig. 9.   Potentiodynamic polarization curves of the un-corroded and corroded galvanized steel in the coastal-industrial atmosphere-simulating electrolyte as a function of CCT cycle.

Table 2   The evolution of fitted parameters from potentiodynamic polarization curves of un-corroded and corroded galvanized steel as the function of CCT cycle.

CCT cycle (N)βa (V·dec-1)βc (V·dec-1)Ecorr (V vs. SCE)icorr (A·cm-2)
0 CCT0.05440.5467-1.0052.85 × 10-6
20 CCT0.06950.0570-1.0421.36 × 10-6
40 CCT0.08600.1575-1.0551.31 × 10-6
60 CCT0.10350.1642-1.0648.93 × 10-7
90 CCT0.20590.1740-1.1234.00 × 10-7
120 CCT0.13360.1535-1.1194.10 × 10-7

New window

The evolution of potentiodynamic polarization behaviour is closely related to the formation and growth of the surface patina layer [3,4,[11], [12], [13], [14], [15], [16]]. During the initial period, the patina layer at 10 CCT and 20 CCT are thin, loose and not compact, with almost no ability in resisting the penetration of SO42- and Cl- to corrode the underlying zinc coating. As the corrosion prolongs to 40 CCT and 60 CCT, the patina layer grows in thickness and becomes compact and dense with improved resistance. Therefore, the icorr decreases gradually as the corrosion prolongs from 10 CCT to 60 CCT during the initial period as shown in Table 2 and Fig. 9. During the second corrosion period from 60 CCT to 90 CCT, the patina layer has entered a stable stage with improved corrosion resistance and the icorr further decreases for the 90 CCT sample. As to the sample at 120 CCT, apparent pitting corrosion has been observed under the patina layer as indicated in Fig. 3(e), leading to a higher icorr as shown in Table 2 and Fig. 9.

Fig. 10 shows the Bode plots, including impedance modulus vs. frequency plot in Fig. 10(a) and phase angle vs. frequency plot in Fig. 10(b), for the electrochemical impedance spectra (EIS) results of the un-corroded and corroded galvanized steel as a function of CCT cycle obtained in the atmosphere-simulating electrolyte. Fig. 10(a) indicates that the impedance modulus value at low frequency region gradually increases during the first corrosion period from 0 CCT to 60 CCT, and rapidly increases afterward, indicating a corrosion resistance improvement of the patina layer and the further decreased of corrosion rate of galvanized steel in the second corrosion period as shown in Fig. 2. Fig. 10(b) indicates that the phase angle plot for the un-corroded galvanized steel sample shows a strong capacitive peak locating at about 40 Hz with a maximum angle of 59° and an inductive feature at the low frequency region. When the corrosion prolongs to 10 CCT, the singular phase angle peak for the un-corroded sample has separated into two peaks, with the first one locating at about 8 Hz with a maximum angle of 39° and the second one locating at about 140 Hz with a maximum angle of 32°. As to the 20 CCT sample, the first peak has shifted to the location at about 2 Hz with a maximum angle of 34°, and the second peak has shifted to about 52 Hz with a maximum angle of 39°. Besides, the inductive feature at the low frequency region has also been observed for the 10 CCT and 20 CCT samples. In fact, the inductive feature can be found for all the measured samples from the Nyquist plot. For the 40 CCT sample, however, three phase angle peaks have been observed, locating at 0.1 Hz, 71 Hz, and 5203 Hz, respectively, and each of the maximum phase angle values is smaller than 34°, the lowest phase angle value of the 20 CCT sample. In addition, the three phase angle peaks are located at high frequency region, low frequency region and middle frequency region, respectively arranged by the maximum phase angle value from high to low. The 60 CCT sample also shows three phase angle peaks located at low frequency region, middle frequency region and high frequency region, respectively. However, the 60 CCT sample presents two more intensive peaks at 0.19 Hz and 26,860 Hz with phase angles of 25° and 34°, respectively, and a smaller peak at 192 Hz with 21° phase angle compared with the 40 CCT sample. As to samples of 90 CCT and 120 CCT, only two phase angle peaks could be found in the phase angle plot. For the 90 CCT sample, a smaller phase angle peak locating at 0.19 Hz with a maximum angle of 19° and a strong phase angle peak locating at about 3722 Hz with a maximum angle of 58° can be found. The maximum phase angle value of the two peaks at 120 CCT was decreased to 45° and 22°, locating at about 1938 Hz and 0.19 Hz, respectively.

Fig. 10.   Bode plots of EIS results for the un-corroded and corroded galvanized steel as a function of CCT cycle: (a) impedance module plot; (b) phase angle plot.

In fact, the surface state including the patina layer structures and characteristics of the working electrode is responsible for the phase angle evolution, which further associates with the time constants for the electrochemical reaction process [49]. Based on the patina layer characterizations in this investigation, it can be found that the corrosion evolution behaviour of galvanized steel is closely related to the surface formed zinc coating patina layer, of which the composition has a great relation to the enhanced corrosion resistance. Besides, the corrosion resistance enhancement by precipitated sulphate-based corrosion product of zinc has also been reported by Svensson and Johansson [50]. Moreover, the synchrotron radiation XRD indicates that both Zn5(OH)6(CO3)2 and Zn4(OH)6SO4 show an increasing amount during the initial corrosion period and maintain at a steady level during the second corrosion process as a function of CCT cycle. Accordingly, in this investigation, an equivalent circuit is shown in Fig. 11 was proposed to fit the EIS data of the corroded galvanized steel samples. In this circuit, Rs the electrolyte solution, RZHS and QZHS the Faraday resistance and capacitance caused by the coverage and amount of Zn4(OH)6SO4 in the patina layer, RHZ and QHZ the Faraday resistance and capacitance caused by the coverage and amount of Zn5(OH)6(CO3)2 in the patina layer, Rct the charge transfer resistance and Qdl the double layer capacitance at patina/zinc coating interface, L the inductor. The fitted parameters RZHS and QZHS, RHZ and QHZ, Rct and Qdl for the corroded galvanized steel samples are shown in Fig. 12. The standard deviations x2 are in the order of 10-4, and the relative error for each parameter is less than 10%, suggesting that the used equivalent circuit is suitable and the fitting parameters are reasonable.

Fig. 11.   Equivalent circuit used to fit the EIS data of corroded galvanized steel samples with different CCT cycle.

Fig. 12.   Evolution in the fitting parameters of (a) RZHS and QZHS, (b) RHZ and QHZ, (c) Rdl and Qct obtained from EIS data of corroded galvanized steel samples as a function of CCT cycle.

Clearly, Fig. 12(a)-(c) shows that RZHS, RHZ, and Rct have an inverse proportional relationship with QZHS, QHZ, and Rct, respectively. In addition, the increased RZHS in Fig. 12(a) and RHZ in Fig. 12(b) indicate an increased relative amount and coverage area of Zn4(OH)6SO4 and Zn5(OH)6(CO3)2 in the patina layer as a function of CCT cycle. Fig. 12(c) shows that Rct also increases as a function of CCT cycle, indicating the improved corrosion resistance of the zinc coating patina layer on galvanized steel. However, the observation of a decreased value of RZHS, RHZ, and Rct for the sample at 120 CCT is due to the change in structure of the patina layer. For the 10 CCT sample, the surface patina layer on galvanized steel is thin, loose with a lower relative amount of Zn4(OH)6SO4 and Zn5(OH)6(CO3)2, and the penetration resistance and protection effect provided by the patina layer is weak. As the corrosion develops, the patina layer thickens and becomes compact with a higher relative amount of Zn4(OH)6SO4 and Zn5(OH)6(CO3)2, and the physical barrier effect to electrolyte enhances, contributing to the increased value of RZHS, RHZ, and Rct, respectively with the CCT cycle increasing. Nevertheless, a long-term corrosion process and repeated wet/dry cyclic process may cause tiny inner stress cracks in the patina layer [1], facilitating the penetration of Cl- and SO42- and leading to the formation of corrosion pits as shown in Fig. 3(f) [25,26]. Therefore, a little-decreased value of RZHS, RHZ, and Rct has been observed for the sample at 120 CCT.

3.6. Corrosion process of galvanized steel

The atmospheric corrosion process of galvanized steel can be interpreted as an electrochemical corrosion process of zinc coating under cyclic wetting/drying of a thin electrolyte film, and the main electrode reactions can be described as follows:

Zn→Zn2++2e (Anodic reaction)(8)

2H2O+O2+4e→4OH- (Cathodic reaction) (9)

2H++2e→H2 (Cathodi creaction) (10)

Under the effect of weak acid electrolyte, Zn2+ can chemically or electrochemically react with OH- to form the initial phases of Zn(OH)2 or ZnO as the main corrosion product composition [1,17]. With the presence of CO32-, SO42-, and Cl- in the electrolyte and its cyclic wet/dry process, ZnSO4, Zn4(OH)6SO4, Zn5(OH)6(CO3)2, and Zn5(OH)8Cl2·H2O can be formed [10,17], respectively by the following reactions:

Zn2+ + $SO_{4}^{2- }$ + xHdO → ZnSO4·xH2O (11)

Zn(OH)2 + 3Zn2+ + 4OH- + $SO_{4}^{2- }$ → Zn4(OH)6SO4 (12)

Zn(OH)2 + 4Zn2+ + 4OH- + 2$SO_{3}^{2- }$ → Zn5(OH)6(CO3)2 (13)

Zn(OH)2 + 4Zn2+ + 6OH- + 2Cl- → Zn5(OH)8Cl2 (14)

Besides, ZnO can react with the newly dissolved Zn2+ to form Zn12(OH)15Cl3(SO4)3 under the combined effect of SO42- and Cl- in the electrolyte [7]:

9Zn2+ + 15ZnO + 6$SO_{4}^{2- }$ + 15H2O + 6Cl- → 2Zn12(OH)15Cl3(SO4)3 (15)

Although ZnO, Zn(OH)2, and ZnSO4 are all of less solubility and can deposit on the zinc coating surface of galvanized steel, the patina layer is porous, thin and discontinuous during the initial corrosion period (Fig. 3(a)), and of less barrier ability in resisting the penetration of SO42- and Cl- to the underlying zinc coating as indicated in the 10 CCT sample (Fig. 8). Therefore, the galvanized steel exhibits a greater corrosion rate during the initial corrosion period (Fig. 1, Fig. 2), and the patina layer at this time has lower RZHS, RHZ, and Rct (Fig. 12). In addition, ZnO, Zn(OH)2, ZnSO4, and the poor crystallized Cl--containing phases are not stable during a long-term wet/dry cyclic corrosion process compared with Zn4(OH)6SO4 and Zn5(OH)6(CO3)2:

Zn12(OH)15Cl3(SO4)3 + 3H2O → 3Zn4(OH)6SO4 + 3Cl- + 3H+ (16)

Therefore, the synchrotron radiation XRD results (Fig. 5) shows the diminishing of Zn12(OH)15Cl3(SO4)3, and an increase of Zn4(OH)6SO4 and Zn5(OH)6(CO3)2 in the patina layer. As the corrosion develops, the patina layer thickens and becomes compact (Fig. 3(d)) with an increased relative amount of Zn4(OH)6SO4 and Zn5(OH)6(CO3)2 (Fig. 5), and thus the corrosion resistance to SO42- and Cl- attack in the electrolyte enhances (Fig. 8). At this time, the galvanized steel has entered into the second corrosion period with a stable patina layer and a lower corrosion rate (Fig. 2), the parameters of RZHS, RHZ, and Rct, relating to the coverage and amount of Zn4(OH)6SO4 and Zn5(OH)6(CO3)2 in the patina layer and the charge transfer resistance, respectively increase gradually as the CCT cycle increases as shown in Fig. 12. However, as the corrosion prolongs to 120 CCT, the long-term wet/dry cyclic corrosion process and the transformation of phases from less stable one to more stable one lead to the formation of tiny inner stress cracks in the patina layer. The tiny cracks act as the penetration tunnels for Cl- and SO42- to corrode new zinc coating, leading to the formation of corrosion pits [25,26] (Fig. 3(f)), with the observation of a large amount of S and Cl elements (Fig. 8), and a little decreased value of RZHS, RHZ, and Rct (Fig. 12) and icorr (Table 2).

4. Conclusions

The corrosion kinetics and patina characteristics evolution of galvanized steel exposed to a simulated coastal-industrial atmospheric environment have been investigated, and it is found that the patina composition and structure can greatly affect the corrosion behaviour of galvanized steel.

(1)The corrosion process of galvanized steel can be divided into two periods with a greater corrosion rate during the first period and a lower corrosion rate during the second period.

(2) The increased relative amount and coverage of Zn4(OH)6SO4 and Zn5(OH)6(CO3)2 in zinc coating patina layer were responsible for the improved corrosion resistance of galvanized steel.

(3) Zn5(OH)6(CO3)2 and Zn4(OH)6SO4 were identified as the main stable composition in the patina layer throughout the whole corrosion process, and Zn12(OH)15Cl3(SO4)3 was only generated in initial corrosion period and diminishes as the corrosion develops.

Acknowledgements

This work was supported by the National Natural Science Fundation of China (Nos. 51501204 and 51671200), the Research Program of Corrosion Distribution and Anti-corrosion Measures of Power Transmission in Complex Atmospheric Environment of Large Coastal Cities (No. GZM2014-2-0004), and the Science and Technology Department of Henan Province (No. 172102310726).


/