Journal of Materials Science & Technology  2019 , 35 (7): 1228-1239 https://doi.org/10.1016/j.jmst.2019.01.008

Orginal Article

Effect of tin addition on corrosion behavior of a low-alloy steel in simulated costal-industrial atmosphere

Bo Liuab, Xin Mub*, Ying Yangc, Long Haob*, Xueyong Dinga, Junhua Dongb*, Zhe Zhangc, Huaxing Houc, Wei Keb

aSchool of Metallurgy, Northeastern University, Shenyang 110819, China
bEnvironmental Corrosion Research Centre of Materials, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China
cAnsteel Co., Ltd., Anshan 114021, China

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

Received: 2018-10-23

Revised:  2018-12-19

Accepted:  2018-12-29

Online:  2019-07-20

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

More

Abstract

The effect of tin addition on the atmospheric corrosion behavior of a low-alloy steel in simulated coastal-industrial atmosphere has been investigated by indoor wet/dry cyclic corrosion test (CCT). The results indicate that tin addition can obviously make the steel substrate more resistant to atmospheric corrosion by suppressing the cathodic H+ reduction reaction, and but tin addition is not of obvious beneficial effect when the steel is covered with a thicker rust layer during long-term corrosion process. The reason lies in the fact that the presence of un-reduced H+ can lower the electrolyte pH value and lead to a loose and porous rust layer on tin-containing steel sample than that on tin-free steel sample. In addition, the 120 CCT cycles corrosion process of the two steels can be divided into three stages. Both the tin-free and tin-containing steels show an increasing corrosion rate during the initial corrosion stage and then exhibit a decreasing corrosion rate during the second and third corrosion stages. Moreover, tin addition makes the tin-containing steel rust layer have a higher amount of α-FeOOH and lower amount of γ-FeOOH and Fe3O4 than the tin-free steel rust layer.

Keywords: Weathering steel ; Atmospheric corrosion ; Coastal-industrial atmosphere ; Tin addition ; Rusting evolution

0

PDF (5446KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Bo Liu, Xin Mu, Ying Yang, Long Hao, Xueyong Ding, Junhua Dong, Zhe Zhang, Huaxing Hou, Wei Ke. Effect of tin addition on corrosion behavior of a low-alloy steel in simulated costal-industrial atmosphere[J]. Journal of Materials Science & Technology, 2019, 35(7): 1228-1239 https://doi.org/10.1016/j.jmst.2019.01.008

1. Introduction

Weathering steel (WS), also known as atmospheric corrosion resistant low-alloy steel that contains one or several elements (Cu, P, Cr, Ni, etc.), was firstly developed by US Steel Company in 1930’s and named it Corten steel. The development of WS from the original Corten series to today’s high performance steel (HPS) series includes not only the improvement of atmospheric corrosion resistance but also the enhancement of mechanical properties, making WS used in a more widely application fields. Field exposure and application experiences [[1], [2], [3], [4], [5], [6], [7]] show that WS usually has a good corrosion resistance under single industrial or coastal atmosphere condition, where the corrosion level does not exceed the C3 category clarified by ISO 9223 However, the corrosion resistance of WS is sometimes not predominant under the compound effect of Cl- and SO2 in coastal-industrial atmosphere [4]. Moreover, the increased exploitation potential and the limited reserves of Cr and Ni elements have led to a worldwide need in the development of cost-effective WS with lower content of Cr and Ni or with other substitute elements, whilst retaining adequate corrosion resistance to atmosphere with single or co-existence of Cl- and SO2.

Tin has been used for a long time in steel industry, such as in the production of tinplate. Although tin, mainly in the formation of mesophases and solid solutions [8,9], is generally regarded as a harmful element to steel, it has been reported to be beneficial for corrosion resistance improvement of steel. For example, tin addition to stainless steel and oil tanker cargo tanks (COT) steel can increase their corrosion resistance in chloride-containing acidic solutions [10,11]. It has been reported [12] that a uniform distribution of tin in steel substrate can reduce the Fermi energy of steel, weaken the electrochemical activity and improve the corrosion resistance of steel in sea-water environment. Kamimura et al. [13] studied the atmospheric corrosion resistance mechanism of tin-containing low-alloy steels in high Cl- environment by the accelerating corrosion test method SAE J2334, and suggested that tin ions slow down the corrosion rate of steel by inhibiting the formation of adsorptive intermediates (FeOHad and FeClad) [[14], [15], [16], [17]]. Nam et al. [18] considered that SnO2, by oxidation of tin, formed a protective film on the steel surface, contributing to a slowing corrosion rate. In summary, the steel with tin addition has excellent corrosion resistance to chloride ion and acid solutions. As to the tin addition effect to steel resistance in atmospheric environment, it may be totally different as the steel is always covered with a thicker rust layer. Townsend et al. [19] confirmed the corrosion resistance improvement of WS by tin addition in industrial atmospheric environment. In marine atmospheric environment, a new WS with tin addition has been developed and but without the corrosion evolution behavior and mechanism having been obtained [20,21].

As we all know that, for the corrosion of steel, an effective cathodic depolarizer is needed. Reduction of dissolved O2 is the most likely reaction to trigger the corrosion process of steel under atmospheric condition. Besides, the reduction of H+ can also be the effective cathodic reaction of the corrosion process of steel in the atmosphere with an acid pH value. As tin has a high over-potential for H+ evolution reaction, tin-containing steel may have a suppressed cathodic current density during the corrosion process in spite of the presence of O2 in the electrolyte. However, the un-reduced H+ can lower the electrolyte pH value and affect the formation and precipitation of rust layer (corrosion product) on tin-containing steel under conditions of high acidity, making the rust layer be of different characteristics, and at this time the effect of tin addition to the atmospheric corrosion of steel may be detrimental. However, to the best of our knowledge, such kind of deduction has not been verified by any atmospheric corrosion test of low-alloy steel. Therefore, in spite of the available data on the investigation of tin content in steels, there are few reports about the effect of tin addition on the corrosion and evolution behavior of low-alloy steel in atmospheric environment, especially under the compound effect of Cl- and SO2 in coastal-industrial atmosphere. In view of this, the present investigation employs the indoor wet/dry cyclic corrosion acceleration test (CCT) [22] to study the atmospheric corrosion behavior of tin-containing low-alloy steel in a simulated coastal-industrial atmosphere, paying particular attention to the corrosion evolution kinetics, rust layer structure, rust phase composition, and electrochemical characteristics. Besides, the corrosion mechanism of low-alloy steel under the effect of tin addition has been discussed. These expected results may have reference value in determining whether tin-containing low-alloy steel is suitable for application in coastal-industrial atmosphere or not, and provide theoretical understanding for the development of more resistant low-alloy steels.

2. Experimental

2.1. Materials

Table 1shows the chemical composition of the two low-alloy steels used for the present study, and the main difference for the two low-alloy steels is the addition of tin element. Steel No.1 is tin free and steel No.2 has a tin content of 0.074% wt. in mass.

Table 1   Chemical composition of tin-free and tin-containing low-alloy steels (wt.%).

SampleCSiMnPSNbTiAlCuSn
No. 10.0780.1711.160.0060.0040.0070.0180.0350.007--
No. 20.0740.1501.210.0080.0020.0110.0140.0280.0070.074

New window

The Zeiss light optical microscope (Axio Cam MRc 5, OM) was employed to observe the microstructure morphology of the two steel substrates. The samples used for microstructure observation were gradually ground to 2000# emery paper, mechanically polished with diamond paste of 2.5 μm particle size, and then etched in absolute ethanol solution with 4 vol.% nitric acid. Fig. 1 shows the metallographic photos of the two low-alloy steels. It can be seen from Fig. 1(a) and (b) that each of the steels microstructure is composed of ferrite (bright area) and pearlite (dark area) phases, with the basically same area fractions of the two phases. Therefore, for the following investigation, the effect of micro-structural difference, brought by 0.074 wt.% tin addition, on the corrosion behavior of steel can be neglected.

Fig. 1.   Microstructure of (a) tin-free low-alloy steel (No. 1) and (b) tin-containing low-alloy steel (No. 2).

2.2. Wet/dry cyclic corrosion test

In this investigation, wet/dry cyclic corrosion test (CCT) [22] was employed to simulate the atmospheric corrosion process of low-alloy steels, and to obtain the corrosion evolution kinetics and prepare the samples for rust layer characterization. The steel specimens were sectioned into 30 mm × 30 mm × 5 mm coupons for the CCT and 10 mm × 10 mm × 5 mm coupons for the electrochemical measurements. The coupons were encapsulated in epoxy resin and then ground with emery paper to 800#. After being cleaned with ethanol, the samples were stored in a desiccator for 24 h, and then subjected to the CCT [23]. It consists of (1) weighing the initial sample; (2) wetting the sample surface with 40 μL/cm2 of 0.05 mol/L NaCl + 0.01 mol/L Na2SO3 solution with pH = 4.0 ± 0.1 (simulating a coastal-industrial atmosphere); (3) drying the sample in a CSH-210 chamber maintained at 25 °C and 60% RH for 12 h; (4) re-weighing the sample after drying; (5) rinsing the sample with distilled water to prevent the accumulation of progressive salt followed by drying the sample before application of fresh corrosion electrolyte; (6) repeating the above steps from (2) to (5) for different CCT cycles, and during each CCT cycle for step (5) salt rinsing, careful operation should be taken to avoid the rust lost. In the following, N is the number of CCT cycles, and one cycle equals to 12 h.

2.3. Electrochemical measurements

The potentiodynamic polarization curves and electrochemical impedance spectroscopy (EIS) measurements were carried out using a classical three electrodes cell with platinum as the counter electrode, saturated calomel electrode (SCE) as the reference electrode, and steel samples with an exposed area of 1 cm2 after suffered different CCT cycles as the working electrode. The potentiodynamic polarization curves were obtained at a constant scan rate of 10 mV/min from the negative potential side to positive side in potential range from -300 mV to 500 mV vs. open circuit potential (OCP). The frequency range for EIS measurement was from 100 kHz to 10 mHz with a 10 mV amplitude signal at open circuit potential. All the measurements were carried out with a Gamry Reference 600 electrochemical workstation at room temperature (25 ± 1 °C) in the atmosphere-simulating electrolyte.

2.4. Corrosion product characterization and analysis

Macroscopic corrosion morphology of the rusted low-alloy steel surface was observed by high-resolution digital camera. The FEI scanning electron microscopy (Inspect F, SEM) equipped with an energy dispersive X-ray analysis (EDS) was employed to observe the cross-sectional morphologies of the rusted steels.

The scraped rust layer was ground to fine powders in a mortar with a pestle and then characterized by XRD technique to determine the rust phases. The XRD measurements were carried out using a Shimadzu XD-5 A diffractometer, with a CuKα target under 50 kV, 250 mA and 2θ = 5°-40° of range at a scanning speed of 2°/min.

The electron probe micro-analyzer (Shimadzu1610, EPMA) with accelerating voltage of 15 KV was used to detect the tin element distribution in the cross-sectional profile of the rusted steel sample after 120 CCT.

The X-ray photoelectron spectroscopy (Thermo VG ESCALAB 250, XPS) employing AlKα was used to analyze the chemical state of tin element in the rust layer on the rusted steel sample after 120 CCT. The XPS spectra were recorded in binding energy values under pass energy of 50 eV and power of 150 W.

3. Results and discussion

3.1. Corrosion kinetics

Fig. 2gives the corrosion kinetics evolution of tin-free and tin-containing steels in the simulated coastal-industrial atmosphere as a function of CCT cycle. Fig. 2(a) shows that the corrosion weight gain of the two steels increases gradually with the CCT cycle increasing. Although the corrosion weight gain of tin-containing steel is lower than that of tin-free steel, the value difference is very small, indicating that the beneficial effect of tin addition on corrosion resistance improvement of the steel in simulated coastal-industrial atmosphere is not obvious. Usually, the atmospheric corrosion kinetics of steel can be fitted with the power function law as shown as follows:

ΔW=ANn (1)

Fig. 2.   Corrosion kinetics evolution of two low-alloy steels in simulated coastal-industrial atmosphere as a function of CCT cycle: (a) linear plot of weight gain; (b) average corrosion rate and linear fitting results in log-log coordinates.

where ΔW is the weight gain (mg/cm2), N is the number of CCT cycle, and A and n are constants. A is considered a measure of the initial corrosion resistance of the sample, and n reflects characteristic of the corrosion kinetics. When the value of n>1, that means a corrosion acceleration process; while when the value of n<1, that means a corrosion deceleration process [24]. A smaller n value indicates a lower atmospheric corrosion tendency. Besides, an expression of the average corrosion rate in logarithmic coordinates could be obtained through dividing the two sides of Eq. (1) by the test period N:

$log\frac{△W}{N}=logA+(n-1)logN$ (2)

Clearly, Eq. (2) shows that the average corrosion rate has a linear changing relationship with CCT cycle N in log-log coordinates. The calculated average corrosion rate as a function of CCT cycle in log-log coordinates is shown in Fig. 2(b), and the obtained linear fitting equations are listed in Table 2. Based on the slope variation of fitting lines (whether the value of n is higher than 1 or not), it can be found that the corrosion process of the two low-alloy steels is mainly divided into three stages, with an enhanced corrosion rate in the first corrosion stage and then a decreased corrosion rate with different descending tendency in the second stage and the third stage, respectively.

Table 2   Linear fitting results of corrosion kinetics data in Fig. 2(b).

SampleStage IStage IIStage III
Tin-free steellogΔW/N = -3.32 + 0.29logN
R2 = 0.98 (N≤ 4)
logΔW/N = -2.97-0.24logN
R2 = 0.96 (4<N≤ 20)
logΔW/N = -2.70-0.43logN
R2 = 0.99 (20<N≤120)
Tin-containing
steel
logΔW/N = -3.42 + 0.15logN
R2 = 0.94 (N≤ 10)
logΔW/N = -3.07-0.20logN
R2 = 0.92 (10<N≤ 20)
logΔW/N = -2.72-0.45logN
R2 = 0.99 (20<N≤120)

New window

The instantaneous corrosion rate Vi (mg/cm2) relative to the CCT can be obtained by the differential of Eq. (1) [25,26]:

$V_{i}=\frac{d△W}{dN}=AnN^{n-1}$ (3)

Based on the fitting results in Table 2, Fig. 3 shows the calculated instantaneous corrosion rate results for the two steels as a function of CCT cycle. Clearly, with the increase of CCT cycle, the whole corrosion process of the two steels consists of three stages. In the first stage, the corrosion rate rapidly increases; in the second stage, the corrosion rate rapidly decreases; in the third stage, the corrosion rate continues to decrease at a much lower value. From the corrosion kinetics analysis of the tin-free and tin-containing steels in Fig. 2, Fig. 3, it can be found that for the tin-free steel, the first corrosion stage is from 0 to 4 CCT, the second corrosion stage lasts from 4 to 20 CCT, and the third corrosion stage ranges from 20 to 120 CCT. However, for the tin-containing steel, the first corrosion stage is from 0 to 10 CCT, the second corrosion stage lasts from 10 to 20 CCT, and the third corrosion stage ranges from 20 to 120 CCT. Therefore, it can be concluded that tin addition has an obvious effect to the steel initial corrosion behavior. Fig. 3 shows that the tin-free steel corrodes at a higher corrosion rate than the tin-containing steel, and it has a shorter first corrosion stage duration of 4 CCT than that of 10 CCT for the tin-containing steel. Therefore, the tin-free steel has an earlier turning point from the first stage to the second stage than the tin-containing steel, resulting in the fact that the tin-containing steel corrodes faster than the tin-free steel over the same CCT duration from 5 to 10 CCT as shown in Fig. 3. In the second and third stages, the corrosion rate of tin-containing steel is lower than that of the tin-free steel. Therefore, it can be concluded from the corrosion kinetics results that tin addition contributes to a little enhancement in atmospheric corrosion resistance of low-alloy steel by affecting the steel initial corrosion process.

Fig. 3.   Calculated instantaneous corrosion rate evolution of two low-alloy steels in the simulated coastal-industrial atmosphere as a function of CCT cycle.

3.2. Corrosion morphology

The effect of tin addition on the surface macroscopic corrosion morphology of the two rusted steels in simulated coastal-industrial atmosphere as a function of CCT cycle is shown in Fig. 4. Clearly, the evolution in surface corrosion morphology of the rusted tin-free steel from Fig. 4(a)-(d) is similar to that of the tin-containing steel from Fig. 4(a')-(d'). For the two rusted steel samples at 10 and 20 CCT, the rust layer surface is relatively intact without obvious cracks. However, for the two rusted steel samples at 60 and 120 CCT, the rust layer surface is cracked and stratified. In addition, it can be found that, the color of rust layer for the two steels changes from light yellow to dark brown as the corrosion process proceeds. As to the difference of the two steels, it can be seen, especially for the 60 and 120 CCT samples, that the tin-containing steel has a darker inner rust layer and but the tin-free steel has a yellow inner rust layer. Usually, color difference is directly related to the variation in phase composition or content of the rust layer [27,28]. Therefore, tin addition to low-alloy steel may contribute to a rust layer of different phase composition or content that has a higher inhibitory effect on the atmospheric corrosion process.

Fig. 4.   Macroscopic corrosion morphologies of rusted tin-free steel at (a) 10 CCT, (b) 20 CCT, (c) 60 CCT, (d) 120 CCT and rusted tin-containing steel at (a') 10 CCT, (b') 20 CCT, (c') 60 CCT, (d') 120 CCT as a function of CCT cycle.

Fig. 5 shows the evolution in cross-sectional morphologies of the rusted two steels in the simulated costal-industrial atmosphere as a function of CCT cycle. In general, for the two steels, the rust layer is thin in thickness at 10 CCT and it grows gradually in thickness as the corrosion process proceeds to 120 CCT. Besides, it can be found that the rust layer is not dense and always has some cracks for both of the two steels. As to the effect of tin addition on the cross-sectional morphology of the rusted steel, the tin-containing steel at 10 CCT in Fig. 5(a') seems has a thinner rust layer, and at 20 CCT the rust layer in Fig. 5(b') has obvious larger cracks than the tin-free steel. Although the tin-containing steel rust layer at 60 CCT in Fig. 5(c') is relative adherent to steel substrate, it is of much more pores. Moreover, the tin-containing steel rust layer at 120 CCT in Fig. 5(d') has pores in much larger size than the tin-free steel rust layer in Fig. 5(d). Therefore, from the evolution in cross-sectional morphologies of the two steels as a function of CCT cycle, it can be found that the tin-containing steel may have a rust layer of much more pores than the tin-free steel.

Fig. 5.   Cross sectional morphologies of rusted tin-free steel at (a) 10 CCT, (b) 20 CCT, (c) 60 CCT, (d) 120 CCT and rusted tin-containing steel at (a') 10 CCT, (b') 20 CCT, (c') 60 CCT, (d') 120 CCT as a function of CCT cycle.

3.3. Composition analysis by XRD

It is well known that the rust on low-alloy steels under atmospheric corrosion is composed of a fewer amounts of α-FeOOH, γ-FeOOH, Fe3O4 and/or β-FeOOH, and more amounts of X-ray amorphous substance, which does not give a well-defined Bragg peak [23]. Among these corrosion products, α-FeOOH is stable both electrochemically and thermodynamically [29]. Therefore it is an important part of the stable rust layer of weathering steel, which directly determines whether the rust layer is dense and protective. β-FeOOH is usually generated in a solution containing Cl-, and once it is formed, it will squeeze the surrounding rust layer, resulting in cracks or even rust stripping, reducing the protective layer of rust on the substrat [30,31]. Moreover, these phases show a wide distribution in particle sizes. Therefore, it is difficult to quantitatively determine the amount of the phases in the rust. In this investigation, we use the same amount of powdered rust for XRD analysis and compare the intensity variation for different samples to semi-quantitatively reflect the relative amount of the phases [32], and thus analysis can only give the evolution in rust composition without exact content [24,25].

Fig. 6 shows the XRD results of the powdered rust formed on the tin-free and the tin-containing steels after 10, 20, 60 and 120 CCT. The α-FeOOH, γ-FeOOH and Fe3O4 peaks were detected for all the measured samples, and tin or tin oxides were not detected. It can be found from Fig. 6 that the intensity of α-FeOOH increases and Fe3O4 decreases with increasing the CCT cycle, and but the intensity of γ-FeOOH does not change dramatically. Thus, for the rust sample at 10 CCT and 20 CCT, it has a relatively higher amount of Fe3O4 and a lower amount of α-FeOOH. As the CCT cycle increases to 60 and 120 CCT, the rust has a relatively higher amount of α-FeOOH but lower amount of Fe3O4. Therefore, the evolution of the XRD results demonstrates that the increased relative amount of α-FeOOH and decreased amount of Fe3O4 is related to the decreased corrosion rate of the two steels as the CCT cycle increases. As to the effect of tin addition, it can be found that, for the same CCT cycle, the tin-containing steel has a lower amount of Fe3O4 and γ-FeOOH and but a higher amount of α-FeOOH than the tin-free steel, indicating that tin addition contributes to a higher amount of α-FeOOH and a lower amount of Fe3O4 and γ-FeOOH in the rust layer on steel in simulated coastal-industrial atmosphere.

Fig. 6.   XRD patterns of powdered rust on two low-alloy steels as a function of CCT cycle: (a) 10 CCT; (b) 20 CCT; (c) 60 CCT; (d) 120 CCT.

3.4. Potentiodynamic polarization curves

Fig. 7 shows the evolution and comparison in potentiodynamic polarization curves of the tin-free steel and tin-containing steel as a function of CCT cycle in the atmosphere-simulating electrolyte. Generally, for the tin-free steel samples and the tin-containing steel samples from Fig. 7(a)-(f), they show almost the similar evolution behavior in open corrosion potential (OCP), cathodic current density, and anodic current density. The anodic steel dissolution and the cathodic reduction of H+ and dissolved O2 dominate the corrosion process of the naked steel i.e. 0 CCT sample [33]. For the steel sample at 4 CCT, a higher cathodic current density caused by the rust reduction became the predominant cathodic reduction process with the disappearance of dissolved O2 reduction, and the H+ reduction also exists due to the measured much lower OCP than the equilibrium potential of hydrogen evolution reaction of about -0.55 V vs. SCE at pH 4.0. Fig. 7(a)-(f) also indicate that the OCP of the steel sample gradually moves to the positive direction to higher than -0.55 V, indicating that the hydrogen evolution reaction gradually weakens. In addition, as the CCT cycle proceeds to 8, 20, 60 and 120 CCT, a gradually increased and then stabilized cathodic current density has been observed due to the gradually stabilized rust layer with an increasing content of α-FeOOH. Moreover, a generally decreasing anodic current density has also been observed as the CCT cycle increases, and the surface formed rust layer with physical barrier effect to the penetration of Cl-, SO42- and H+ may be responsible for this. Furthermore, it can also be found that the anodic current density of the two steel samples does not always show a suppressed value as the CCT cycle increases, and the rust layer evolution behavior that does not always have a growing compactness in Fig. 5 may be responsible for this.

Fig. 7.   Polarization curves of tin-free steel and tin-containing steel as a function of CCT cycle: (a) 0 CCT; (b) 4 CCT; (c) 8 CCT; (d) 20 CCT; (e) 60 CCT; (f) 120 CCT.

The effect of tin addition on the evolution behavior of potentiodynamic polarization curves of the two steels as a function of CCT cycle is also shown in Fig. 7. For the naked steel substrate samples in Fig. 7(a), a much lower cathodic current density was observed for the tin-containing steel substrate, indicating that tin-addition has obviously inhibited the cathodic reduction process and contributed to a lower corrosion current density of the steel. For the rusted steel samples at 4 CCT in Fig. 7(b), the lower cathodic current density of the tin-containing steel than that of the tin-free steel also lies in the inhibition effect of H+ reduction by tin addition, and but the tin-containing steel seemingly has a higher anodic current density than the tin-free steel. As the CCT cycle proceeds to higher number, it can be found that tin addition always makes a lower cathodic current density of tin-containing steel than that of the tin-free steel. However, the anodic current density of tin-containing steel does not always show a lower value than that of tin-free steel, indicating the complicated effect of tin addition to the anodic dissolution process of low-alloy steel in coastal-industrial atmosphere. The lasting inhibition of H+ reduction at cathode and the higher amount of α-FeOOH by tin addition have always made a lower cathodic current density of the tin-containing steel than that of the tin-free steel. Therefore, for the samples at 8 CCT in Fig. 7(c) and 20 CCT in Fig. 7(d), they have a similar evolution behavior in cathodic current density with the 4 CCT sample in Fig. 7(b), i.e. the tin-containing steel has a lower cathodic current density than the tin-free steel. For the tin-containing steel, the presence of a large amount of un-reduced H+ can lower the electrolyte pH value and affect the formation and precipitation of rust layer on tin-containing steel under conditions of high acidity, and leads to the formation of a porous and discontinuous rust layer on tin-containing steel. For the tin-free steel, it has a higher cathodic current density, and thus the lower amount of remaining H+ can increase the electrolyte pH value and contribute to the formation of a compact and continuous rust layer on tin-free steel under conditions of low acidity. Therefore, a higher anodic current density of tin-containing steel sample than that of tin-free steel sample has been observed at 8 and 20 CCT. In addition, the compact inner rust layer on tin-containing steel sample in Fig. 5(c') and the cracked rust layer on tin-free steel sample in Fig. 5(c) may be responsible for the observed lower anodic current density of tin-containing steel sample than that of the tin-free steel sample at 60 CCT in Fig. 7(e). It can also be concluded that the tin-containing steel has a porous rust layer of more cracks and cannot effectively hinder the penetration of Cl-, SO42- and H+ to the steel substrate, leading to the higher anodic current density of the tin-containing steel sample at 120 CCT as shown in Fig. 7(f).

3.5. Electrochemical impedance spectroscopy

Fig. 8 shows the evolution in Bode plots, including the modulus plot and phase angle plot, of EIS results for the tin-free steel (Fig. 8(a) and (a')) and the tin-containing steel (Fig. 8(b) and (b')) as a function of CCT cycle measured in the atmosphere-simulating electrolyte. According to the impedance modulus for 0 CCT tin-free steel sample in Fig. 8(a), it can be seen that the |Z| value at high frequency impedance (|Z|HF) corresponding to the solution resistance (Rs) is about 40 Ω cm2, and the |Z| value at low frequency impedance (|Z|LF) corresponding to the sum of polarization resistance (Rp) and Rs [34] is about 130 Ω cm2. Fig. 8(a') shows that the phase angle plot for 0 CCT tin-free steel sample shows a single symmetric peak at 3.76 Hz with a maximum value of 26.71°, indicating that the corrosion process contains one time constant. For the 4 CCT tin-free steel sample, |Z|HF increases to about 50 Ω cm2, indicating the presence of rust layer resistance Rr at high frequency impedance as the same test electrolyte with almost unchanged Rs. However, |Z|LF decreases to about 80 Ω cm2, indicating the lowered Rp and increased corrosion rate of the steel caused by the depolarizer reduction of γ-FeOOH during electrochemical corrosion process. Fig. 8(a') shows the obvious influence in phase angle plot for the 4 CCT sample, caused by the presence of rust layer covering with the phase angle peak disappearance at 3.76 Hz compared with the 0 CCT sample. Moreover, as the corrosion process proceeds to higher CCT cycle, Fig. 8(a) shows that both |Z|HF and |Z|LF increase to higher impedance value, indicating the increased Rr and Rp of the rusted steel sample with higher CCT cycle.

Fig. 8.   (a, b) Modulus and (a', b') phase angle diagrams for (a, a') tin-free and (b, b') tin-containing low-alloy steels in corrosion electrolyte as a function of CCT cycle.

For the tin-containing steel sample, Fig. 8(b) shows that the 0 CCT tin-containing steel sample has almost the same |Z|HF of about 40 Ω cm2 and predominately higher |Z|LF of about 270 Ω cm2, compared with the 0 CCT tin-free steel sample, indicating the almost same Rs of the two electrolytes and the much higher corrosion resistance of tin-containing steel substrate than that of the tin-free steel substrate as Rp usually reflects the corrosion resistance characteristics. Fig. 8(b') shows that the phase angle plot for 0 CCT tin-containing steel sample shows a single symmetric peak at 1.915 Hz with a maximum value of 34.27°, indicating the corrosion process also contains one time constant. Moreover, the fact that the phase angle peak shifts to the lower frequency region and shows a higher value than the 0 CCT tin-free steel sample also indicates the enhanced corrosion resistance by tin addition to the steel substrate in the coastal-industrial atmosphere-simulating electrolyte with pH 4.0. For the 4 CCT tin-containing steel sample, |Z|HF increases to about 50 Ω cm2, indicating the presence of rust layer resistance Rr at high frequency impedance; nevertheless, |Z|LF decreases to about 90 Ω cm2, indicating the increased corrosion rate caused by the γ-FeOOH reduction during corrosion process. As the corrosion process proceeds to higher CCT cycle, Fig. 8(b) shows that both |Z|HF and |Z|LF increase to higher impedance value, indicating the increased Rr and Rp of the rusted tin-containing steel sample with higher CCT cycle. Fig. 8(b') shows that the tin-containing steel sample exhibits the similar evolution behavior in phase angle plot with the tin-free steel sample.

To further understand the electrochemical characteristics evolution of the rusted tin-free and tin-containing steels, the equivalent electrical circuits listed in Fig. 9 have been employed to fit the EIS data as a function of CCT cycle. In the circuit, R represents the Faraday resistance of different substances that affect the corrosion process. Q represents the constant phase element (CPE) caused by diffusion effect and its impedance (Z) is defined by the following function [35]:

ZQ = Y0-1()-n (4)

Fig. 9.   Equivalent electrical circuits for EIS data of two low-alloy steels at different CCT cycle: (a) un-corroded 0 CCT sample; (b) corroded samples with different CCT cycle.

where Y0 is the magnitude of CPE, reflecting the property of surface and electro-active species, with the dimension of S sn cm-2; is the variables of sinusoidal perturbation with ω = 2πf and f is the voltage signal frequency; n is the exponent of Q. When n equals to 0.5, Q could be interpreted as a semi-infinite Warburg element (W), while when n equals to 1, Q could be regarded as an ideal capacitance (C). Therefore, the specific elements used in Fig. 9 are as follows: Rs is the electrolyte resistance; RO and QO are the Faraday resistance and the CPE of dissolved oxygen reduction, respectively; RH and QH are the Faraday resistance and the CPE of hydrogen irons reduction, respectively; Rrust and Qrust are the Faraday resistance and the CPE of rust layer, respectively; Rct is the charge transfer resistance and Qdl is the CPE of the double layer at electrolyte/steel substrate interface. Fig. 9(a) exhibits the equivalent circuit used for fitting the EIS data of un-corroded tin-free and tin-containing steel substrates, 0 CCT. As the steel substrate at this time has no rust covering, the cathodic corrosion process only includes the dissolved oxygen reduction and H+ reduction, which can be denoted by component of (QO(ROWO)) and (QHRH), respectively, and the component of (QdlRct) reflects the anodic reaction of steel substrate dissolution process. However, for the corroded tin-free and tin-containing steel samples with rust layer covering, the oxidation of ferrous ion to ferric ion by dissolved oxygen is always of a much greater rate than the oxygen reduction rate during cathodic process. Therefore, for the EIS data fitting of corroded tin-free and tin-containing steel samples after different CCT cycle, the oxygen reduction at cathodic corrosion process is no longer considered and but the rust layer reduction is taken consideration and denoted by the component of (QrustRrust) as shown in Fig. 9(b), and the component of (QH(RHWH) is used to reflect the cathodic H+ reduction process and the diffusion feature under the effect of rust layer covering, with the component of (QdlRct) also reflecting the anodic reaction of steel substrate dissolution process.

The fitting results of the EIS plots for tin-free and tin-containing steel samples by software of Zsimpwin are shown in Table 3, with each of the data having a fitting error less than 10-4, and Fig. 8 also shows that the equivalent circuits are relatively well fitted with the measured EIS data. Clearly, the 0 CCT tin-containing steel substrate has a higher RH than the tin-free steel substrate, indicating the greater difficulty in the occurrence of H+ reduction on tin-containing steel substrate surface and thus the greater beneficial effect of tin addition in resisting H+-induced electrochemical corrosion of steel substrate in acid corrosion electrolyte. The greater Rct and RO together with the greater Rs (lower electrical conductivity caused by the lower content of dissolved Fe2+ and the lower OH- concentration generated by H+ and oxygen reduction) of tin-containing steel substrate system also indicates the lowered corrosion current density of tin-containing steel substrate brought by tin addition as illustrated in Fig. 7(a). In addition, for the 4 CCT tin-free and tin-containing steel sample systems compared with the 0 CCT system, the obviously lowered Rs value also lies in the higher electrical conductivity caused by the higher content of dissolved Fe2+ and the higher OH- concentration generated by H+ and oxygen reduction in the electrolyte. Besides, the obviously lowered Rct value lies in the steel sample surface formed rust layer, which contains a large amount of reducible composition, such as γ-FeOOH that can act as depolarizer and accelerate the corrosion rate during the corrosion process. Moreover, the obviously lowered Rct value also indicates the great effect of rust layer on atmospheric corrosion process of steels. As the corrosion process proceeds to higher CCT cycle, Table 3 shows that Rrust and Rct exhibit an increasing tendency, indicating the gradually improved rust layer resistance and decreased corrosion rate for both tin-free and tin-containing steels. RH shows an increasing tendency as the CCT cycle increases to 20 CCT and then maintains at a steady level as the CCT cycle further increases to 120 CCT. As to the effect of tin addition, it can be found that tin-containing steel sample always has a greater RH than tin-free steel sample, and but the two RH values for the un-corroded steel samples at 0 CCT have the greatest difference, which then becomes smaller for the corroded samples, indicating that tin addition in resisting corrosion is predominant for the steel substrate corrosion and such beneficial effect lowers but still exists for the rust covered steel. It is also noted that the tin-containing steel sample in general has a higher Rct value than the tin-free steel sample, however, it has a lower Rrust value and the difference enlarges as the CCT cycle increases. As to the reasons, we think that although the tin-containing steel can greatly inhibit the reduction of H+ at the electrolyte/electrode interface, the un-reduced H+ are also existed in the electrolyte, and the effect of H+ to rust layer itself cannot be neglected. The presence of a large amount of un-reduced H+ can lower the electrolyte pH value and affect the formation and precipitation of rust layer on tin-containing steel under conditions of high acidity, and lead to the formation of a porous and discontinuous rust layer. Therefore, the rust layer on tin-containing steel surface is loose and of higher porosity than that on the tin-free steel surface, and thus leads to the lowered Rrust on tin-containing steel surface. However, for the tin-free steel, it has a higher hydrogen evolution rate and cathodic current density, and thus the lower amount of remaining H+ can increase the electrolyte pH value and contribute to the formation of a compact and resistant rust layer on tin-free steel under conditions of low acidity, and thus a higher Rrust has been observed on the tin-containing steel surface.

Table 3   Fitting results of EIS data of rusted tin-free steel and tin-containing steel samples as a function of CCT cycle.

CCT cycleRsQO (mS sn cm-2)ROYw-OQH (mS sn cm-2)RHYw-HQrust (mS sn cm-2)RrustQdl (mS sn cm-2)Rct
(Ω cm2)YOnO(Ω cm2)(mS s0.5 cm-2)YHnH(Ω cm2)(mS sn cm-2)Yrustnrust(Ω cm2)Ydlndl(Ω cm2)
Tin-free steel
036.606.34800.66885.1570.85380.869500.955914.25----1.02800.960072.41
420.01----0.115200.605124.850.00262.3410.191929.970.11760.17759.450
819.21----0.000050.909833.68100.7072.130.456041.1726.5200.280129.57
2021.34----0.061550.333442.580.062632.190.5803139.217.5500.3863160.6
6021.59----12.64000.441943.332.02500.94370.4856223.90.00520.4800212.5
12021.01----5.263000.398043.890.23942.5820.6626413.30.00410.5028267.7
Tin-containing steel
039.459.78400.104811.640.38760.530800.919950.09----0.90400.9800210.4
421.05----0.153800.594728.580.01871.4180.657625.080.00900.181513.97
820.62----0.000080.783437.530.0898108.80.518447.4235.2200.280223.75
2022.43----0.026320.388848.200.069315.120.3222133.532.4300.1877156.7
6022.09----12.18000.799253.300.006320.020.5729190.71.32900.1094232.7
12021.84----12.85000.354853.500.02341.7620.5205188.10.21880.2195230.3

New window

3.6. Discussion on the effect of tin addition

Fig. 10 shows the distribution of tin element in the rust layer on the tin-containing steel after 120 CCT cycle characterized by EPMA. In Fig. 10(a), the light color represents the steel substrate, and the dark color represents the rust layer. Fig. 10(b) shows that the tin element is uniformly distributed in the steel substrate and shows slight accumulation in the inner rust layer, while its distribution in the outer rust layer is relatively scare and loose. Accordingly, it can be found that, during a long term of 120 CCT cycles corrosion, slightly accumulation of tin element was found in the inner rust layer, indicating that the formed tin-containing oxides is mainly distributed in the inner rust layer on tin-containing steel.

Fig. 10.   EPMA results of tin element distribution in rust layer on tin-containing steel after 120 CCT cycle.

Fig. 11 shows the XPS spectrum analysis of the chemical states of tin element in the rust layer, and the sub-peak fitting analysis indicates that SnO2, SnO and elemental tin are probably the main forms of tin element in the rust layer [[36], [37], [38]] by means of the following corrosion reactions:

Sn + 4H2O → Sn(OH)4 + 4e- + 4H+ (5)

Fig. 11.   XPS spectra and sub-peak fitting analysis of Sn element in rust layer on tin-containing low-alloy steel after 120 CCT cycles.

Sn may also be oxidized according to the following reactions:

Sn + 2H2O → Sn(OH)2 + 2e- + 2H+ (6)

Sn(OH)2 + 2H2O → Sn(OH)4 + 2e- + 2H+ (7)

The following dehydration reaction can also take place:

Sn(OH)2 → SnO + H2O (8)

Sn(OH)4 → SnO2 + 2H2O (9)

Therefore, oxides of SnO and SnO2 may be formed during the wet/dry cyclic corrosion process of the steel [12]. Usually, the compounds of SnO and SnO2 are regarded as immune in aqueous, and SnO2 can remain stable in dilute acid solution [39]. Therefore, it may be inferred from the XPS analysis that the formation of tin oxides not only can reduce the number of active points on steel substrate where corrosion can be initially induced, but also greatly inhibits H+ reduction at tin-containing steel surface, contributing the lower cathodic current density of tin-containing steel than that of tin-free steel.

As to the effect of tin addition on the atmospheric corrosion behavior of low-alloy steel in coastal-industrial atmosphere, the corrosion kinetics obtained by weight gain evolution in Fig. 2, Fig. 3 shows the lower corrosion weight gain of tin-containing steel than that of tin-free steel, indicating the seemingly beneficial effect of tin addition on the corrosion improvement of low-alloy steel in coastal-industrial atmosphere. However, detailed analysis including electrochemical methods and physical characterization of the formed rust layer shows that tin addition only has obvious beneficial effect to the steel substrate without rust layer covering, and it has almost no beneficial effect to the steel with a rust layer covering from the aspect of long-term atmospheric corrosion.

For the atmospheric corrosion of low-alloy steel in the present simulated coastal-industrial atmosphere with pH = 4.0, the anodic dissolution of steel, and the cathodic reduction of H+ and dissolved O2 dominate the corrosion process [33]. Tin addition to steel substrate greatly suppresses the cathodic reduction of H+ in the corrosion electrolyte, and thus the tin-containing steel shows an obviously lower corrosion rate as shown at 0 CCT in Fig. 3, a lower cathodic corrosion current density in Fig. 7(a), and a higher Rct in Table 3 than the tin-free steel substrate. As the corrosion process proceeds to higher CCT cycles, a rust layer gradually forms and thickens and the tin addition effect still exists, contributing to the always lower cathodic current density of the tin-containing steel sample than the tin-free steel sample with rust layer covering. However, the presence of un-reduced H+ can lower the electrolyte pH value and affect the formation and precipitation of rust layer on tin-containing steel under conditions of high acidity, leading to a loose and porous rust layer on tin-containing steel surface as indicated by the lower Rrust of tin-containing steel sample than that on tin-free steel sample in Table 3. In such condition, micro-galvanic corrosion between tin oxides (SnO2 and SnO) and steel substrate may take place at the electrolyte/tin-containing steel substrate interface under the effect of available Cl-, SO42- and H+, leading to the higher anodic current density of tin-containing steel sample than that of the tin-free steel sample as shown in Fig. 7(b), (c) and (f). Only when a compact inner rust layer forms, it can greatly hinder the penetration of H+, SO42- and Cl- through the rust layer to underlying tin-containing steel substrate and contribute to a lower anodic current density of tin-containing steel sample than that of tin-free steel sample at 60 CCT in Fig. 7(e).

In addition, the dissolved ferrous ions are hydrolyzed into a deposit of Fe(OH)2, and during the drying of the aqueous film, Fe(OH)2 is quickly oxidized by dissolved oxygen into electrochemically active Fe(III) oxides [29]. These oxides transform into a large amount of amorphous substance and a small amount of α-FeOOH, γ-FeOOH and Fe3O4 as the corrosion process proceeds. The formation of uniformly distributed elemental tin and/or oxides in the rust layer not only can reduce the number of active sites on steel substrate where corrosion can be induced, but also can promote the formation of α-FeOOH and inhibit the formation of γ-FeOOH and Fe3O4 in the rust layer. As the corrosion process proceeds, the amount of α-FeOOH gradually increases and the amount of Fe3O4 gradually decreases as shown in Fig. 6. Moreover, under the effect of tin addition, the tin-containing low-alloy steel will have a higher amount of α-FeOOH and a lower amount of γ-FeOOH and Fe3O4 than the tin-free low-alloy steel as shown in Fig. 6. As the CCT cycle proceeds to the third corrosion stage, the rust layer on tin-containing low-alloy steel has a higher amount of α-FeOOH, and a lower amount of γ-FeOOH and Fe3O4 than the rust layer on the tin-free low-alloy steel.

To sum up, tin addition can obviously suppress the cathodic H+ reduction reaction and contribute to a lower corrosion rate of tin-containing steel during the first several CCT cycles. For the tin-free steel, it has a higher hydrogen evolution rate and cathodic current density, and thus the lower amount of remaining H+ can increase the electrolyte pH value and contribute to the formation of a compact and continuous rust layer on tin-free steel under conditions of low acidity. However, for the tin-containing steel, the un-reduced H+ can lower the electrolyte pH value and affect the formation and precipitation of rust layer on tin-containing steel under conditions of high acidity, and make the rust layer to be loose and porous. Therefore, corrosive ions of SO42- and Cl- can easily penetrate the rust layer to the steel substrate surface to corrode more steel. It can be concluded that tin addition to low-alloy steel can make the steel substrate more resistant to atmospheric corrosion, and but tin addition is not of obvious beneficial effect when the steel is covered with a thicker rust layer during long-term corrosion process.

4. Conclusions

The effect of tin addition to the corrosion behavior of low-alloy steel submitted to the wet/dry cyclic corrosion tests in a simulated coastal-industrial atmosphere has been investigated, and the following conclusions can be obtained:

(1)The 120 CCT cycles corrosion process of the two steels can be divided into three stages. Both the tin-free and tin-containing steels show an increasing corrosion rate during the initial corrosion stage and then exhibit a decreasing corrosion rate during the second and third corrosion stages as the corrosion process proceeds.

(2)Tin addition can obviously suppress the cathodic H+ reduction reaction and make the steel substrate more resistant to atmospheric corrosion, and but tin addition is not of obvious beneficial effect when the steel is covered with a thicker rust layer during long-term corrosion process.

(3)The presence of un-reduced H+ can lower the electrolyte pH value and affect the formation and precipitation of rust layer on tin-containing steel under conditions of high acidity, leading to a loose and porous rust layer on tin-containing steel sample than that on tin-free steel sample.

Acknowledgements

This work was supported financially by the National Natural Science Fundation of China (Nos. 51501204, 51501201 and 51671200) and the National Key Research and Development Program of China (No. 2017YFB0702302).

The authors have declared that no competing interests exist.


/