Journal of Materials Science & Technology  2019 , 35 (9): 1886-1893 https://doi.org/10.1016/j.jmst.2019.04.022

Orginal Article

Effect of corrosive media on galvanic corrosion of complicated tri-metallic couples of 2024 Al alloy/Q235 mild steel/304 stainless steel

Linjun Shiab, Xiuying Yangc, Yingwei Songa*, Dan Liuab, Kaihui Donga, Dayong Shana, En-Hou Hana

a Key Laboratory of Nuclear Materials and Safety Assessment, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China
b School of Materials Science and Engineering, University of Science and Technology of China, Hefei 230026, China
c Shenyang Institute of Technology, Fushun 113122, China

Corresponding authors:   ∗Corresponding author.E-mail address: ywsong@imr.ac.cn (Y. Song).

Received: 2019-01-27

Revised:  2019-03-23

Accepted:  2019-04-2

Online:  2019-09-20

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

More

Abstract

Galvanic corrosion of tri-metallic couples is more complicated than that of bi-metallic couples. In this study, the effect of the pH of corrosive media on the galvanic corrosion of 2024 Al alloy/Q235 mild steel/304 stainless steel tri-metallic couples was investigated using potentiodynamic polarization, scanning electron microscopy, scanning vibrating electrode technique and a multi-channel galvanic corrosion meter. The results show that 2024 always acts as the only anode in 3.5 wt% NaCl at pH 5.56, 9.72 and 12.0, while both Q235 and 2024 act as anodes at pH 2.39 in the initial stage and then the role of Q235 changes at longer coupling time, which can be attributed to the effect of pH on the surface film of 2024. It is also found that the galvanic current density of a tri-metallic couple is the superposition of two bi-metallic couples when cathodic reactions are controlled by the diffusion of oxygen, otherwise it is smaller than that of the sum of two bi-metallic couples. The localized corrosion instead of uniform corrosion of anodic metal is accelerated by galvanic corrosion.

Keywords: Galvanic corrosion ; Tri-metallic couples ; Corrosive media ; Scanning vibrating electrode technique ; Zero resistance ammeter

0

PDF (2918KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Linjun Shi, Xiuying Yang, Yingwei Song, Dan Liu, Kaihui Dong, Dayong Shan, En-Hou Han. Effect of corrosive media on galvanic corrosion of complicated tri-metallic couples of 2024 Al alloy/Q235 mild steel/304 stainless steel[J]. Journal of Materials Science & Technology, 2019, 35(9): 1886-1893 https://doi.org/10.1016/j.jmst.2019.04.022

1. Introduction

Galvanic corrosion is an accelerated corrosion of a metal when it is connected to other metals with higher potential [1]. Metal connectors, such as riveting components and weld assemblies, are inevitable in industries. Thus, galvanic corrosion is very common in aviation, petroleum, building and shipping industries, resulting in heavy losses in economy and safety [2].

Since 1763 when a ship failed owing to the galvanic corrosion of joint iron nails and copper cladding, considerable attention has been paid to bi-metallic galvanic corrosion. Researchers found that galvanic corrosion is affected by numerous factors, such as the potential difference and polarization abilities summarized as material properties [[3], [4], [5], [6], [7]]; area ratios of anode/cathode, insulation distance, and spatial positions summarized as geometric factors [4,[8], [9], [10], [11]]; and temperature, conductivity, oxygen contents and pH of electrolytes summarized as environmental factors [8,[12], [13], [14], [15], [16], [17]]. A formula of bi-metallic galvanic current density is derived on the basis of the electrochemical theories as follows [18].

$lni_{a1}=\frac{E_{corr2}-E_{corr1}}{\beta_{a1}+\beta_{c2}}+\frac{\beta_{a1}}{\beta_{a1}+\beta_{c2}}lnI_{corr1}+\frac{\beta_{c2}}{\beta_{a1}+\beta_{c2}}lnI_{corr2}+\frac{\beta_{c2}}{\beta_{a1}+\beta_{c2}}ln\frac{A2}{A1}$ (1)

where ia1 is anodic current density; Ecorr1 and Ecorr2 are corrosion potential of anodes and cathodes, respectively; βa1 and βc2 are Tafel slopes of anodes and cathodes, respectively; Icorr1 and Icorr2 are corrosion currents of anodes and cathodes, respectively; A1 and A2 are areas of anodes and cathodes, respectively.

Based on these studies and electrochemical theories, researchers have suggested two types of methods to suppress galvanic corrosion. One is insulating the direct electrical contact between corrosive media and metals using coatings or insulating washers [[19], [20], [21], [22]]. The another is designing reasonable structure to decrease or even avoid the risk of galvanic corrosion [23]. In a word, bi-metallic galvanic corrosion can be suppressed effectively based on the well understanding of its galvanic corrosion mechanisms.

However, although many complicated metallic couples inevitably exist in industries, such as riveting components, pipe-couples and weld assemblies, the investigations about the galvanic corrosion of complicated metallic couples are far from sufficient. Fortunately, some researchers had realized the necessity of investigating the galvanic corrosion of complicated metallic couples. El-Moneim et al. [24] studied the micro-galvanic interaction between three intermetallic phases in NdFeB-based permanent magnets. They concluded that the Nd-rich phase represents the typical anode, whereas the B-rich phase and the ferromagnetic phase act as cathodes. It indicates that the roles of metals are determined by their own material properties. However, Akid et al. [25] studied the electrochemical corrosion behavior of steel/Al/Al alloy tri-metallic couples in artificial saltwater, and found that the roles of Al and Al alloy are determined by their distances to the typical cathode (steel) and the conductivity of the solutions in tri-metallic couples. It indicates the role of metal cannot be ascertained only by the potential, and the insulation distance and the conductivity of solutions are key factors in tri-metallic galvanic corrosion. Besides, tri-metallic couples are possible to be serviced in different corrosive environments such as acidic, neutral and alkaline. On the one hand, the coupled metals exhibit different electrochemical feature under various pH value conditions. On the other hand, the galvanic corrosion status can change with various pH values. Thus, pH of solutions plays a key factor in galvanic corrosion [14,26], but few work has concentrated on the effect of pH on tri-metallic galvanic corrosion.

In this study, the effect of the pH of solutions on tri-metallic galvanic corrosion was investigated from a macroscale and a microscale, aiming to provide helpful theoretical information for diminishing the galvanic corrosion of complicated metallic couples.

2. Experimental

2.1. Materials and solutions

The experimental materials used for this investigation were Q235 mild steel, 2024 aluminum alloy and 304 stainless steel, which are widely used in industries. Their chemical composition is shown in Table 1, Table 2. The metals with working area of 1 cm2 were molded in epoxy, polished to 2000# grit SiC paper, washed with water and ethanol, and then dried in cold air.

Table 1   Chemical compositions of Q235 and 304 (wt%).

CSiMnPSCrNiCuFe
Q2350.210.060.230.0130.0080.030.010.01Bal.
3040.04210.46051.12350.03180.001518.1868.186Bal.

New window

Table 2   Chemical compositions of 2024 (wt%).

SiFeCuMnMgCrZnTiAl
0.4990.514.00.81.60.950.250.15Bal.

New window

All electrochemical measurements were carried out in 3.5 wt% NaCl which simulates a marine environment. The solutions were prepared with analytical NaCl reagent and deionized water. Four pH values (pH 2.39, pH 5.56, pH 9.72 and pH 12.0) were chosen to represent acidic, neutral, alkalescent and alkaline solutions, respectively. The pH of NaCl solutions were adjusted using analytical HCl and NaOH reagents. The conductivity of 3.5 wt% NaCl was measured to be from 50000 to 53400 μS·cm at 25℃ using a platinic conductivity cell.

2.2. Potentiodynamic polarization

The potentiodynamic polarization curves of 2024, Q235 and 304 were measured in 3.5 wt% NaCl solutions (at four pH) using PARSTAT4000 electrochemistry test system (Princeton Applied Research, USA). A three-electrode electrochemical cell was used with a saturated calomel electrode (SCE) as the reference electrode, a platinum plate as the counter electrode, and metals with an exposed area of 1 cm2 as the working electrode. After an initial delay of 600 s, the potentiodynamic polarization curves of 2024, Q235 and 304 were measured. The scan rate was 0.5 mV/s.

2.3. Galvanic corrosion tests

The samples for tri-metallic galvanic corrosion tests were arranged in a one-dimensional array as shown in Fig. 1(a). Three metals with the same working areas of 1 cm2 were molded in epoxy in an insulation distance of 3 mm. The galvanic potential and currents of each electrode in 3.5 wt% NaCl at different pH were instantaneously and continuously recorded for 20 h using CST508. It is worth mentioning that the measurement method of tri-metallic galvanic corrosion is different from that of bi-metallic galvanic corrosion. The schematic for measuring the galvanic corrosion of tri-metallic couples is shown in Fig. 1(b). Three channels of CST508 were used for these tests, and each channel includes a work electrode connector1 (WEC1), a work electrode connector2 (WEC2) and a reference electrode connector (REC). 2024, Q235 and 304 were connected to three WEC2s, respectively, three WEC1s were connected together, and three RECs were connected to the same SCE. Zero Resistance Ammeters and Potential Followers in CST508 are used to record the current and potential, respectively.

Fig. 1.   Schematics of (a) sample and (b) connection method of tri-metallic galvanic corrosion measurements.

2.4. Surface morphologies

After galvanic corrosion tests, carbon was sputtered on the surface of the samples, and then the surface morphologies were observed using a scanning electron microscope (SEM, Phillips XL30FEG) with the acceleration voltage of 14 kV [27]. Corrosion products were not removed before the SEM observation.

2.5. SVET measurements

Scanning vibrating electrode technique (SVET) instrument (Applicable Electronics Inc., USA), which was controlled by the ASET 2.0 software (Science Wares Inc., USA), was used to measure the potential difference between the two extremes of vibration of the insulated Pt/Ir probe (Microprobes Inc., USA) due to ionic current in the electrolyte [28]. The insulated Pt/Ir probe, with a platinum black deposited on a spherical tip of 15 μm diameter, was used as the vibration electrode with a vibration amplitude of approximately 30 μm. The SVET scan height was fixed at 100 μm [29]. 2024, Q235 and 304 were cut to the size of 1.5 mm × 1.5 mm, electrically connected together, molded in epoxy resin and then polished to 2000# grit SiC paper for the SVET measurements. The electrolytes used for the investigations were 3.5 wt% NaCl at pH 2.39, 5.56, 9.72 and 12.0. Each test was performed at least three times to ensure good reproducibility.

3. Results

3.1. Zero resistance ammeter (ZRA) and potential measurements of tri-metallic couples in NaCl solution at different pH

The galvanic current density and potential of tri-metallic couples immersed in 3.5 wt% NaCl at pH 2.39, 5.56, 9.72 and 12.0 are shown in Fig. 2. The galvanic potential stabilizes at approximately -600 mV vs. SCE at pH 2.39 (Fig. 2(a)). 304 acts as a cathode, and its cathodic current density decrease from -106 to -63 μA cm-2. Both Q235 and 2024 act as anodes, and the current density of Q235 (approximately 56 μA cm-2) is larger than that of 2024 (approximately 32 μA cm-2) at the initial stage. With prolonging immersion time, the current density of Q235 gradually decreases from 56 μA cm-2 to nearly zero. Whereas the current density of 2024 increases from 32 to 60 μA cm-2, exceeds that of Q235 at 11000s, and then reaches a steady value of 60 μA cm-2 at about 20,000 s. In a word, the effect of Q235 on tri-metallic galvanic corrosion diminishes through time, whereas 2024 gradually turns to be a stronger anode and eventually becomes the unique anode.

Fig. 2.   Galvanic current density and potential of Q235/2024/304 in 3.5 wt% NaCl at different pH: (a) pH 2.39; (b) pH 5.56; (c) 9.72; (d) pH 12.0.

The galvanic corrosion of tri-metallic couples in 3.5 wt% NaCl at pH 5.56 and 9.72 (Fig. 2(b) and (c)) present a similar trend. The galvanic potential is approximately -600 mV vs. SCE. 2024 acts as an anode with current density of approximately 40 μA cm-2. Both Q235 and 304 act as cathodes with current density of approximately -20 μA cm-2.

As shown in Fig. 2(d), the galvanic potential at pH 12.0 gradually increases from -1.16 V vs. SCE to -1.02 V vs. SCE. 2024 acts as an anode with anodic current density decreasing from 480 to 56 μA cm-2. Q235 and 304 act as cathodes, and the cathodic effect of Q235 is stronger than 304 in the initial time. With prolonging immersion time, the cathodic current density of Q235 and 304 decreases from -280 and -110 μA cm-2 to the same value of -28 μA cm-2, respectively. The increase of galvanic potential and the decrease of anodic and cathodic current density indicate that the galvanic corrosion of tri-metallic couples is retarded with prolonging immersion time. It may result from the accumulation of corrosion products.

3.2. Comparison of corrosion status of Al in NaCl solution at different pH

The ZRA results show that 2024 always acts as the unique anode in four NaCl solutions at different pH except for the anodic role of Q235 in the acid solution during the initial stage. Therefore, the corrosion status of 2024 can represent the effect of pH on the galvanic corrosion of tri-metallic couples. The anodic current density of 2024 at four pH is compared in Fig. 3. 2024 presents the largest current density at pH 12.0, but the current density of 2024 dramatically decreases after long time immersion and even is smaller than the current density at pH 2.39 after 70,000 s. The current density of 2024 at pH 5.56 and 9.72 is the smallest.

Fig. 3.   Galvanic current density of 2024 coupled to Q235 and 304 in 3.5 wt% NaCl at different pH.

Furthermore, the corrosion morphologies of 2024 after tri-metallic galvanic corrosion measurements at four pH were observed using SEM as shown in Fig. 4. The overall surface of 2024 is visible with abundant corrosion pits in 3.5 wt% NaCl solution at pH 2.39. These pits are open without corrosion products piling up (Fig. 4(a)). 2024 suffers a slight attack at pH 5.56. Its surface is comparatively smooth, and some corrosion products can be seen on the corrosion pits (Fig. 4(b)). 2024 suffers slight corrosion at pH 9.72 as well. Corrosion products can be observed easily, but fewer pits can be seen (Fig. 4(c)) compared with Fig. 4(a) and (b). At pH 12.0, the overall surface of 2024 is covered by thick corrosion products with network cracks. Some grooves appear owing to the falling off of the corrosion products (Fig. 4(d)). Even so, these corrosion products still show a limited protection on 2024, which lead to the decrease of the anodic and cathodic current density with increasing immersion time. In a word, 2024 is corroded more heavily at pH 2.39 and pH 12.0, which is consistent with the electrochemical data shown in Fig. 3.

Fig. 4.   SEM surface morphologies of 2024 after tri-metallic galvanic corrosion in 3.5 wt% NaCl for 20 h at pH values of (a) 2.39, (b) 5.56, (c) 9.72, (d) 12.0.

3.3. Potentiodynamic polarization measurements

To clarify the reason for different galvanic corrosion behavior of 2024/Q235/304 tri-metallic couples in NaCl solution at different pH, the potentiodynamic polarization curves of 2024, Q235 and 304 were measured as shown in Fig. 5.

Fig. 5.   Potentiodynamic polarization curves of 2024, Q235 and 304 in 3.5 wt% NaCl at different pH: (a) pH 2.39; (b) pH 5.56; (c) pH 9.72; (d) pH 12.0.

At pH 2.39 (Fig. 5(a)), 2024, which shows the lowest potential of -875 mV vs. SCE, exhibits a short passivation platform at its anodic side. The current density of 2024 rapidly increases once the film is broken down (approximately -600 mV vs. SCE). Q235, with an intermediate potential of -609 mV vs. SCE, shows the largest corrosion current density among three metals. The anodic side of Q235 in acid solution is controlled by the metallic dissolution reaction. 304, with the highest potential of -344 mV vs. SCE, shows a short passivation platform at its anodic side. The galvanic potential intersects with the cathodic side of 304, the anodic side of Q235 and the pitting side of 2024. 304 acts as a cathode, while 2024 and Q235 act as anodes under the galvanic potential. Interesting enough, the cathodic reduction reactions of three metals are all controlled by the diffusion of oxygen in acid solution as that in neutral solution. However, the cathodic current density in acid solution is larger than that in neutral solution. During the cathodic polarization period, the emission of a few bubbles can be observed on the surface of metals. The reason might be that part of H+ is reduced along with the reduction of oxygen, and then the diffusion of H2 stirs the stagnant layer of oxygen, which leads to the decrease of stagnant layer thickness. Eventually, the decrease of stagnant layer thickness leads to the increase of the limit diffusion current density of oxygen IL [18].

At pH 5.56 (Fig. 5(b)), 2024 shows the lowest corrosion potential of -690 mV vs. SCE, and its anodic side is controlled by metallic dissolution reaction. Q235 and 304 present higher corrosion potential of -519 mV and -249 mV vs. SCE, respectively, and their cathodic sides are controlled by the oxygen reduction reaction. The galvanic potential intersects with the anodic side of 2024, the cathodic sides of both Q235 and 304. Therefore, it indicates that 2024 acts as an anode, whereas Q235 and 304 act as cathodes under the galvanic potential of -595 mV vs. SCE.

At pH 9.72 (Fig. 5(c)), 2024 shows the most negative potential of -655 mV vs. SCE. Its anodic side is controlled by the metallic dissolution reaction, and then pitting corrosion potential appears at around -600 mV vs. SCE. Q235, with an intermediate potential of -523 mV vs. SCE, is controlled by the active dissolution reaction at its anodic side. 304, with the most positive potential of -247 mV vs. SCE, exhibits a passivation platform at its anodic side. The cathodic reactions of three metals are controlled by the diffusion of oxygen. The galvanic potential of -600 mV vs. SCE intersects with the anodic side of 2024, the cathodic sides of both Q235 and 304. 304 and Q235 act as cathodes, whereas 2024 acts as an anode under the galvanic potential.

At pH 12.0 (Fig. 5(d)), both Q235 and 304 exhibit the corrosion potential of -543 mV vs. SCE and passivation platforms, but the passivation platform of 304 is wider than that of Q235. Besides, the cathodic sides of Q235 and 304 are controlled by the diffusion of oxygen at potential from -664 to -950 mV vs. SCE and then followed by the reduction reaction of H+ at the more negative potential regions. On the contrary, 2024 exhibits a lower potential of -1.355 V vs. SCE, and the galvanic potential intersects with the anodic side of 2024, the cathodic sides of both Q235 and 304., that is, 2024 acts as an anode, whereas 304 and Q235 act as cathodes under the galvanic corrosion. Furthermore, the cathodic current density of Q235 is larger than that of 304 under the galvanic potential, which is accordant with the ZRA result during the initial immersion time (Fig. 6(c)).

Fig. 6.   SVET of Q235/2024/304 tri-metallic couples in 3.5 wt% NaCl at pH 2.39, 5.56, 9.72 and 12.0 at different immersion time: (a) schematic illustration of coupled samples for SVET measurement; current density distribution after immersion for 30 min (b, d, f, h) and 20 h (c, e, g, i).

3.4. SVET measurements

The SVET results of Q235/2024/304 tri-metallic couples immersed in 3.5 wt% NaCl at different pH are shown in Fig. 6. The spatial distribution of 2024, Q235 and 304 are shown in Fig. 6(a). Since adjusting the probe to suitable sites of the SVET measurements spent approximately 30 min, the current density mappings were obtained at 30 min and 20 h during the immersion time.

At pH 2.39, both of 2024 and Q235 act as anodes, while 304 acts as a cathode in the initial immersion time (Fig. 6(b)). With prolonging immersion time, the anodic current density of Q235 decrease and Q235 eventually turns to be a weak cathode after immersed for 20 h (Fig. 6(c)). At pH 5.56, it can be observed that 2024 acts as an anode, while Q235 and 304 act as cathodes during the whole immersion time. The anodic and cathodic currents decrease with prolonging immersion time (Fig. 6(d) and (e)). At pH 9.72, 2024 acts as an anode while Q235 and 304 act as cathodes. The anodic and cathodic currents in the initial time (Fig. 6(f)) are stronger than that at 24 h (Fig. 6(g)). At pH 12.0, 2024 acts as an anode while Q235 and 304 act as cathodes. The anodic and cathodic currents decrease with prolonging immersion time (Fig. 6(h) and (i)). Otherwise, the anodic current density at pH 2.39 and 12.0 are obviously larger than that at pH 5.56 and 9.72 indicating the severer corrosion, which confirms the conclusion deduced from the data of ZRA measurements (Fig. 2).

The non-uniform distribution of anodic current density with higher local current density exhibits the feature of localized corrosion of anodic metals. Severe localized corrosion occurs on the surface of 2024 at pH 2.39, 5.56 and 9.72 except for pH 12.0, which can be observed from SEM as well (Fig. 4(a)-(c)). Moreover, the non-uniform distribution of anodic currents proves that only the local pitting corrosion is accelerated instead of the uniform corrosion of whole surface of anodic metal. On the country, the cathodic current density distributes uniformly, which indicates the oxygen are reduced uniformly on the surfaces of cathodic metals.

4. Discussion

4.1. Galvanic corrosion of the tri-metallic couples

According to the mixed potential theory, 304 always acts as a cathode, whereas 2024 always acts as an anode in 3.5 wt% NaCl at pH 5.56 (Fig. 2(b)). The potential difference between 304 and Q235 is larger than that between Q235 and 2024. If the potential difference (thermodynamics) is considered only, Q235 is more likely to be an anode. However, according to the electroneutrality principle (dynamics), the galvanic potential can be estimated to be approximately -600 mV from the potentiodynamic polarization curves (Fig. 5(b)), and Q235 acts as a cathode under this galvanic potential. This is consistent with the electrochemical data (Fig. 2(b)), SEM (Fig. 4(b)), and SVET result (Fig. 6(d) and (e)). But Q235 acts as an anode at the initial immersion time in the NaCl solution at pH 2.39 (Fig. 5(a)). In summary, the role of the metal with an intermediate potential in tri-metallic couples cannot be confirmed simply through their corrosion potential.

It is note that, although the potential of 304 is more positive than that of Q235, the cathodic current density of 304 and Q235 are almost the same (Fig. 2(b)). The reason is that the cathodic reduction reactions on both 304 and Q235 are controlled by the diffusion of oxygen under the galvanic potential (Fig. 5(b)). Therefore, the cathodic current density of both 304 and Q235 equal to the limit diffusion current density of oxygen (iL). More than that, it seems that the tri-metallic couple Q235/2024/304 is the superposition of bi-metallic couples 2024/304 and 2024/Q235 (Fig. 2(b)) and the anodic current Ia is proportional to the cathodic working areas Ac.

Is that law universal? To explain the relationship between the tri-metallic couple M1/M2/M3 and the bi-metallic couples M1/M2 and M1/M3 (or M1/M3 and M2/M3), three schematics were drawn as shown in Fig. 7. The working areas of M1, M2 and M3 are 1 cm2. If the reduction reactions on cathodic metals are controlled by the diffusion of oxygen (Fig. 7(a)), the galvanic current density of M1/M2 couple is iM1/M2, which equals to the anodic current density of M1 and the absolute value of the cathode current density of M2, iM1/M2=iaM1=|icM2|; the galvanic current density of M1/M3 couple is iM1/M3, which equals to the anodic current density of M1 and the absolute value of the cathode current density of M3, iM1/M3=iaM1=|icM3|=|icM2|=iM1/M2=iL. There exists only one galvanic potential (EM1/M2/M3) which conforms to the electroneutrality principle when M1, M2 and M3 are coupled together. Under this galvanic potential EM1/M2/M3, the anodic current density iaM1’ equals to the superposition of the absolute value of the cathode current density of M2 and M3, |icM2|’ and |icM3|’, respectively, that is, iaM1’=|icM2|’+|icM3|’. |icM2|, |icM3|, |icM2|’ and |icM3|’ equal to the limit diffusion current density of oxygen iL. Therefore, it can be concluded that if the reduction reactions on cathodic metals are controlled by the diffusion of oxygen, the galvanic current density of tri-metallic couple (M1/M2/M3) is the superposition of bi-metallic couples (M1/M2 and M1/M3), and the anodic current density ia in a bi-metallic or complicated metallic galvanic corrosion can be calculated as the linear relationship of ia=Ac×iL, no matter what cathodic metals are. However, the linear relation is invalid if reduction reactions on cathodic metals are not controlled by the diffusion of oxygen (Fig. 5(d)).

Fig. 7.   Schematics of polarization curves of tri-metallic couples: (a) cathodic reactions controlled by diffusion of oxygen; (b) liner system with two anodes and a cathode; (c) liner system with two cathodes and an anode.

On the contrary, assuming that M1 and M2 act as anodes while M3 acts as a cathode (Fig. 7b), the current density of M1 and M3 in the bi-metallic couple M1/M3 is iaM1 and icM3 under the galvanic potential EM1/M3 and iaM1=|icM3|, and the current density of M2 and M3 in the bi-metallic couple M2/M3 are iaM2 and icM3 under the galvanic potential EM2/M3 and iaM2=|icM3|. When M1, M2 and M3 are coupled together, only one galvanic potential EM1/M2/M3 conforms to the electroneutrality principle. Under this galvanic potential EM1/M2/M3, the absolute value of the cathodic current density of M3 |icM3’| equals to the superposition of the anodic current density of M1 and M2, iaM1’ and iaM2’, respectively, that is, |icM3’|=iaM1’+iaM2’< iaM1+iaM2. Assuming that M1 acts as an anode and M2 and M3 act as cathodes (Fig. 7(c)), the conclusion of iaM1’=|icM2’+icM3’|<|icM2+icM3| can be gotten as well. Therefore, the total galvanic current density of a tri-metallic couple is smaller than that of the sum of both bi-metallic couples when their cathodic reactions are not controlled by the diffusion of oxygen.

4.2. Effect of corrosive media on 2024/Q235/304 tri-metallic galvanic corrosion

The 2024/Q235/304 tri-metallic couples show different galvanic corrosion in different corrosive media.

Q235 always acts as a cathode in neutral, alkalescent and alkaline solution in macroscale (Fig. 2(b)-(d)), but it act as an anode in acidic solution (Fig. 2(a)) during the initial immersion time. With prolonging immersion time, the current density of Q235 gradually decrease to zero, while the current density of 2024 increases to be a stronger anode. Potentiodynamic polarization results (Fig. 5(a)) show that the galvanic potential intersects with the cathodic side of 304, the anodic activation dissolution side of Q235 and the pitting side of 2024, and the anodic current density icorr of Q235 is larger than that of 2024. It is due to the existence of oxidation film on 2024 during the initial immersion time, which leads to the surface activity of Q235 is higher than 2024. Thus, the anodic effect of Q235 is stronger than 2024. With prolonged immersion time in an acidic solution, the oxidation film of 2024 is gradually dissolved. The accelerated dissolution of matrix and the cathodic effect of precipitated phase of 2024 lead to the increase of the anodic currents of 2024. Consequently, the anodic effect of Q235 is diminished.

Ordinarily, Al as an amphoteric metal should be corroded more seriously in alkaline and acid solution [30,31], which should increase the galvanic currents. However, the galvanic corrosion current density of 2024, Q235 and 304 in neutral solution (Fig. 2(b)) are almost the same as that in alkalescent solution (Fig. 2(c)). It may result from the interaction between the pH and the Cl-. Firstly, the difference between the alkalescent solution at pH 9.72 and the neutral solution at pH 5.56 is slight. Secondly, pitting corrosion occurs on the surface of 2024 (Fig. 4(b), (c) and Fig. 7), which is due to the interaction of Cl- with an oxide film as well the micro-galvanic corrosion produced by the second phases [32]. Therefore, the effect of pH is weakened in comparison with the role of Cl in the pitting corrosion of 2024.

The chemical dissolution reaction of the aluminum hydroxide in alkaline solution is generally given by:

Al(OH)3+OH-(aq)=Al(OH)-4(aq) (2)

However, the concentration of Al(OH)-4(aq) will be increased at the interface by the applied electric field, thereby reducing the rate of chemical film thinning by the OH- attack, and the applied electric field can accelerate the growth of the Al(OH)3 film [33]. The anodic polarization carve of 2024 in alkaline solution (Fig. 5(d)) shows a passivation region, which indicates a film formed by the applied electric field as shown in Fig. 4(d). During the galvanic corrosion period, 2024 is in the state of anodic polarization and the film thickens continuously, resulting in the decrease of the anodic current density of 2024 and the increase of the galvanic potential (Fig. 2(d)). The increase of galvanic potential transform the cathodic reactions of Q235 and 304 from the hydrogen reduction reaction to the oxygen reduction reaction (Fig. 5d), and this is the reason why the cathodic current density of Q235 and 304 increase to the same value gradually (Fig. 2(d)). The galvanic corrosion in alkaline solution depends on the surface state of 2024.

5. Conclusions

(1)The galvanic corrosion of 2024/Q235/304 tri-metallic couple in the alkaline solution is severest, but its galvanic current density decreases with the prolonging immersion time because of the accumulation of corrosion products. The galvanic corrosion in acidic solution is followed, and in alkalescent and neutral solutions are slightest.

(2)2024 always acts as the unique anode in four NaCl solutions at different pH except for the anodic role of Q235 at the initial stage of galvanic corrosion in the acid solution, and the galvanic corrosion accelerates localized corrosion rather than uniform corrosion of 2024.

(3)When the cathodic reactions are controlled by the diffusion of oxygen under the galvanic potential, the galvanic current density of a tri-metallic couple (M1/M2/M3) is a superposition of the two bi-metallic couples (M1/M2 and M1/M3), and the anodic currents of the tri-metallic couple are proportional to the cathodic working areas no matter what types of cathodic metals are. When their cathodic reactions are not controlled by the diffusion of oxygen, the total galvanic current density of a tri-metallic couple is smaller than the sum of that of both bi-metallic couples.

Acknowledgements

This work was supported financially by the National Key Research and Development Program of China (No. 2017YFB0702100) and the Natural Science Foundation of Liaoning Province (No. 20170540666).

The authors have declared that no competing interests exist.


/