Materials Chemistry
In-Situ Observation of Initial Stage Cosmetic Corrosion Behavior under Wet and Dry Conditions
2022 Volume 63 Issue 4 Pages 612-621
Details
2022 Volume 63 Issue 4 Pages 612-621
It is well known that the types of automotive corrosion can be divided into perforation corrosion and cosmetic corrosion. Although the mechanism of perforation corrosion has been studied extensively, the mechanism of cosmetic corrosion has not yet been clarified. The authors investigated the cosmetic corrosion behavior of cold-rolled steel sheets without galvanizing with an in-situ observation device. After coating the samples by cathodic electrodeposition (ED), the sample surface was scribed with a cutter. Corrosion resistance was evaluated under cyclic corrosion test. It was found that the progress of under-film corrosion consisted of 3 steps. The 1st step occurs in the initial stage of the dry process in the 1st cycle. In this step, red rust gradually changed to black rust. The 2nd step occurs in the latter stage of the dry process, and under-film corrosion progressed from the scribed part. The 3rd step occurs in the wet process, and in this step, the tip of the under-film corrosion displayed swelling behavior. In the 2nd cycle, the 2nd step and 3rd step of the 1st cycle were repeated. Under-film corrosion progressed at almost the same rate as in the 1st cycle.
This Paper was Originally Published in Japanese in Zairyo-to-Kankyo 69 (2020) 212–220.
Fig. 2 Appearance of samples after (a) 0 h, (b) 0.5 h, (c) 3 h, (d) 4.5 h, (e) 6 h, (f) 12 h, (g) 18 h, and (h) 24 h of cyclic corrosion test.11)
It is well known that the types of automotive corrosion can be divided into “perforation corrosion” which occurs in lapped panels and “cosmetic corrosion” which starts from paint defects caused by chipping and the material edge.1) Various countermeasures have been taken against both types of automotive corrosion, such as the use of anti-corrosion steel sheets, wax, sealers, and painting technology. However, from the viewpoint of resource saving and cost minimization, it is important to optimize corrosion resistance in accordance with the use environments of the automobile so as to avoid excessive corrosion resistance. In order to propose the optimum countermeasures, a correct understanding of how environmental corrosion factors affect “perforation corrosion” and “cosmetic corrosion” is necessary.
The mechanism of “perforation corrosion” has been studied extensively based on an analysis of actual vehicles,2) and the effects of anti-corrosion steel sheets under various corrosion environments have been clarified.3) Regarding “cosmetic corrosion”, in other words, “under-film corrosion”, it is generally known that the anodic and cathodic sites are localized and show the morphology of macro corrosion.4,5) This corrosion behavior is quite different from the corrosion morphology of non-painted steel sheets because the presence of the paint film assists the localization of the anodic and cathodic sites. “Cosmetic corrosion” has been studied in many reports. For example, Shastry6) et al. conducted an accelerated corrosion test using a painted cold-rolled steel sheet and galvanized steel sheet having a scribed part under wet and dry conditions for 20 cycles, and then attempted to identify the anodic and cathodic sites by analyzing the peeled position of the paint film and the concentrated position of chloride by cross-sectional EPMA. Hayashi7) et al. carried out salt spray exposure tests with a duration of 5 months to 2 years and accelerated corrosion tests under wet and dry conditions consisting 28 to 84 cycles, and then tried to identify the anodic and cathodic sites of the galvanized steel sheet by analyzing the concentrated position of chloride by cross-sectional EPMA in the same way as Shastry et al.
However, these previous studies observed specimens after corrosion tests consisting of more than a few dozen cycles under wet and dry conditions, and did not discuss how under-film corrosion occurs in the initial stage under wet and dry conditions or the respective morphologies of under-film corrosion in the dry and wet processes. It would be difficult to explain the corrosion behavior under wet and dry conditions by one corrosion model because the corrosion environment changes moment by moment in actual environments and or accelerated corrosion tests. In particular, it is uncertain whether the corrosion behavior under wet and dry conditions can be discussed based only on the conventional method, in which the specimens are taken out and analyzed in each cycle. In order to clarify the corrosion behavior in each wet and dry condition, in-situ observation of the corrosion behavior would be an effective approach.
Therefore, in this study, the authors developed an in-situ observation technique, and carried out in-situ observation of the initial stage of under-film corrosion under wet and dry conditions using cold-rolled steel sheets without galvanizing with paint defects, and discussed the mechanism of the under-film corrosion considering both the results of in-situ observation of the process of under-film corrosion and the results of an analysis of specimens taken out in each wet and dry condition.
Cold-rolled steel sheets with tensile strength of 270 MPa grade without galvanizing were used as the test specimens. The dimensions of the specimens were 150 mm × 70 mm (thickness: 0.8 mm). The specimens were alkaline degreased, phosphated, and then coated by cathodic electrodeposition to a thickness of 20 µm, which is the same as the coating system used in automotive bodies. The central part of the specimens was scribed with a straight 60 mm line to the substrate steel, and an area with a width of 10 mm at the upper, lower, right, and left edges on the scribed side and all of the reverse side of specimens were sealed with a corrosion-protection film. Figure 1 shows the shape and dimensions of the test specimens.
Shape and dimensions of test specimens.
The cyclic corrosion test according to SAE J23348) [(1) wet process (50°C, RH 100%, 6 h), (2) Salt dip process (0.5 mass% NaCl + 0.1 mass% CaCl2 + 0.075 mass% NaHCO3, 15 min), (3) Dry process (60°C, RH 50%, 17 h 45 min)] was employed as the corrosion test. It has been reported that this corrosion test has a good correlation with corrosion in actual environments based on the results of a comparison with an on-vehicle test in North America and Canada.9) Since the presence of salt greatly affects corrosion, the following process was defined as one cycle in order to investigate the under-film corrosion behavior under wet and dry conditions after salt deposition: Salt dip process → Dry process → Wet process. The corrosion test was performed in one or two of these cycles.
2.3 In-situ observation during cyclic corrosion testIn order to enable in-situ observation during a cyclic corrosion test, it is necessary to protect a camera from dew condensation because the inside of the equipment is exposed to a high temperature and humidity environment. Therefore, an acrylic box was prepared to supply cool, dry air in the box to protect the camera from dew condensation. In addition, water droplets and salt deposited on the outside of the box were physically removed.10) This device made it possible to perform in-situ observation under wet and dry environments during the cyclic corrosion test. In the cyclic corrosion test, the specimen was set in the chamber with an angle of 90° to the ground surface so that the height of the specimen was 150 mm. A region with a size of 2 mm × 1.5 mm including the scribed part was observed by the in-situ observation device, and images were photographed and recorded in a personal computer at intervals of 10 min. The images were then combined to create a moving picture.
Generally, cosmetic corrosion of automotive steel sheets is evaluated by the “one-sided scribe creep of the paint film”, which is defined as the distance from the scribed part to the tip of under-film corrosion (hereinafter referred to as scribe creep). The under-film corrosion in one location was selected from the video images using a test specimen subjected to two cycles of the corrosion test, and scribe creep was measured at prescribed intervals. Specimens were taken out during the cycle at prescribed intervals, and the surface of the specimens including the scribed part was analysed after drying with a blower without washing. The analysis of the specimens included X-ray diffraction (XRD, RIGAKU RINT-RAPIDII-R, spot size: 0.3 mmϕ, tube: Co-Kα, (voltage: 45 kV, current: 160 mA, incident angle: 25°, 15°, 5°)) and cross-sectional analysis with a scanning electron microscope (SEM, JEOL JSA7001F, accelerating voltage: 15 kV), electron probe micro analyzer (EPMA; JEOL JXA8200, accelerating voltage: 15 kV, (current: 5.007 × 10−8, dwell time: 20 ms)) and Raman spectroscope (Thermo Fisher Scientific ALMEGAXR, laser wavelength: 532 nm, (exposure time: 30 s, number of exposures: 20)).
Figure 2 shows the in-situ observation results of a specimen surface for 1 cycle of the cyclic corrosion test.11,12) In the initial stage of the dry process at 60°C and RH 50% after the salt dip process, red rust (orange rust) was generated at the whole scribed part where the substrate was exposed (Fig. 2(a), (b)). Afterwards, this red rust gradually changed to black rust at the scribed part (Fig. 2(c) to (e)). According to the video record of the in-situ observation, the black rust was generated at multiple places in the scribed part and gradually expanded to the whole area. Although this continued until the black rust covered the whole scribed part, under-film corrosion was not observed. During this phenomenon, the whole scribed part appeared to be covered with a thin water film. After this, under-film corrosion occurred and progressed from the scribed part (Fig. 2(f), (g)). The morphology of this under-film corrosion can be recognized as filiform corrosion, and the under-film corrosion progressed in the direction away from the scribed part while keeping the same width. This continued until the end of the dry process in the 60°C and RH 50% environment. The corrosion behavior changed simultaneously with the shift to the wet process in the 50°C and RH 100% environment. The progress of the under-film corrosion in the direction away from the scribed part stopped, and only the tip of under-film corrosion swelled in a circular shape (Fig. 2(h)). Although the circular swelling continued to grow, the progress of the under-film corrosion in the direction away from the scribed part was relatively small, unlike the behavior of the under-film corrosion observed in the dry process of 60°C and RH 50% environment which was recognized as filiform corrosion. Unlike the 60°C and RH 50% environment, many small water droplets were observed on the paint film and scribed part during this period due to the high humidity environment of RH 100%.
Appearance of samples after (a) 0 h, (b) 0.5 h, (c) 3 h, (d) 4.5 h, (e) 6 h, (f) 12 h, (g) 18 h, and (h) 24 h of cyclic corrosion test.11)
From these in-situ observation results, three different corrosion steps were observed under the environment of 1 cycle consisting of salt dip, dry (60°C and RH 50%), and wet (50°C and RH 100%) processes: (1) Change to black rust at the scribed part, (2) Initiation and progress of under-film corrosion from the scribed part, and (3) Circular swelling of the tip of under-film corrosion.
3.2 In-situ observation of under-film corrosion for 2 cycles in cyclic corrosion testFigure 3 shows the in-situ observation results of the specimen surface for 2 cycles in the cyclic corrosion test using a specimen different from that in Fig. 2. In the 1st cycle, as in the results of Fig. 2, three different corrosion steps were reproduced in one cycle, which consisted of salt dip, dry (60°C and RH 50%), and wet (50°C and RH 100%) processes: (1) Change to black rust at the scribed part (Fig. 3(a), (b)), (2) Initiation and progress of under-film corrosion from the scribed part (Fig. 3(c), (d)), and (3) Circular swelling of tip of under-film corrosion (Fig. 3(e)).
Appearance of samples after (a) 15 min, (b) 6 h, (c) 9 h, (d) 18 h, and (e) 24 h in 1st cycle and (f) 15 min, (g) 6 h, (h) 9 h, (i) 18 h, and (j) 24 h in 2nd cycle of cyclic corrosion test.
In the 2nd cycle, there were no significant changes in appearance from immediately after the salt dip until 6 h at 60°C and RH 50% (Fig. 3(f), (g)). After this, under-film corrosion with the morphology of filiform corrosion progressed again from the tip where progress was stopped once by the circular swelling in the 1st cycle (Fig. 3(h), (i)). This continued until the end of dry process under the 60°C and RH 50% environment, as in the 1st cycle. Under the next 50°C and RH 100% environment, only the tip of the under-film corrosion swelled in a circular shape, as in the 1st cycle (Fig. 3(j)), and progress of under-film corrosion with the morphology of filiform corrosion was not observed.
Scribe creep was measured from the observation image obtained in Fig. 3, and the results are shown in Fig. 413) as a function of test time. In the 1st cycle, scribe creep was not observed up to 6 h (Fig. 4(a)). In this period, under-film corrosion does not occur because black rust is generated in the scribed part. Subsequently, scribe creep increased at an almost constant rate beginning from 6 h (Fig. 4(b)). Under-film corrosion progresses with the morphology of filiform corrosion in this period. The progress rate during this period was 0.03 to 0.04 mm/h, and this phenomenon continued until the end of the dry process under the 60°C and RH 50% environment. Scribe creep increased slightly in the initial stage of the 50°C and RH 100% environment (Fig. 4(c)), but did not increase thereafter (Fig. 4(d)). During this period, only the tip of the under-film corrosion swelled in a circular shape, although scribe creep increased slightly in the initial stage. This suggests that the increase in scribe creep during this period is observed as a result of the growth of circular swelling. In addition, the increase in scribe creep was completed in the early stage of the wet process in the 50°C and RH 100% environment. This means that circular swelling occurs in the early stage of the wet process under the 50°C and RH 100% environment, and the size of that circular swelling is maintained afterwards.
Example of change of scribe creep with test time.13)
In the 2nd cycle, as in the 1st cycle, scribe creep was not observed up to 6 h after salt dip (Fig. 4(e)). Subsequently, again as in the 1st cycle, scribe creep increased at an almost constant rate beginning from 6 h (Fig. 4(f)). The progress rate can be calculated as 0.03 to 0.04 mm/h, and the behavior was substantially the same as in the 1st cycle. In the 50°C and RH 100% environment, as in the 1st cycle, scribe creep increased slightly in the initial stage due to the growth of circular swelling (Fig. 4(g)) and did not increase subsequently (Fig. 4(h)).
These results clarified the fact that the progress of under-film corrosion with the morphology of filiform corrosion and the growth of circular swelling at the tip of the under-film corrosion were repeated in the 1st cycle and the 2nd cycle. The progress rate of scribe creep in the 1st cycle was also repeated in the 2nd cycle corresponding to the change of dry and wet conditions, as shown by the fact that the scribe creep depended on the test time obtained from in-situ observation.
Moreover, in order to compare the corrosion behavior of the local area observed in-situ with the corrosion behavior of the whole scribed part with a 60 mm length, specimens were taken in each of the processes of salt dip, dry, and wet, and their appearances were observed. As results, in the specimen which had passed through 6 h of the dry process, the change to black rust covering the whole scribed part occurred, and under-film corrosion with the morphology of filiform corrosion was observed in the whole scribed part. In the specimen which had finished the wet process, circular swelling was found at the tip of the under-film corrosion with the morphology of filiform corrosion immediately after the specimen was taken. The agreement between these results for observation of the whole part and the results of the in-situ observation of different specimens shown in “In-situ Observation for 1 Cycle” in section 3.1 and “In-situ Observation for 2 Cycles” in section 3.2 suggests that the series of behaviors from the change to black rust at the scribed part to the progress of under-film corrosion with the morphology of filiform corrosion obtained in the area of in-situ observation are not specific local phenomena. However, the value of the progress rate of under-film corrosion obtained in Fig. 2 was different from that obtained in Fig. 3. This difference may have occurred because the data were affected by local conditions, such as a difference in the salt concentration under the paint film.
As mentioned above, based on the in-situ observation results in Fig. 2 and Fig. 3, three different corrosion steps were observed under the wet and dry conditions: (1) Change to black rust at the scribed part, (2) Initiation and progress of under-film corrosion from the scribed part, and (3) Circular swelling of the tip of under-film corrosion. Here, (1) Change to black rust at the scribed part is discussed.
Three specimens which were different from those used for the in-situ observation were prepared, and a cyclic corrosion test was carried out with in-situ observation. These specimens were taken out at prescribed times and analyzed from the view point of the rust composition and elemental distributions in the rust.
The XRD diffraction (incident angle: 25°) patterns of the 1st cycle are shown in Fig. 5(a) before the corrosion test, (b) after 3 h, and (c) after 6 h at 60°C and RH 50%. Most of the peaks such as TiO2 and Al2Si2O5(OH)4, originated from the paint,14) but peaks of iron oxyhydroxide (β-FeOOH (2θ = 39.698°), γ-FeOOH (2θ = 55.205°)) and magnetite (Fe3O4 (2θ = 41.375°)) were also detected as components of the rust. The ratio of the net peak intensity of Fe3O4 (2θ = 41.375°) to that of γ-FeOOH (2θ = 55.205°) is shown in Fig. 6 as a function of the test time, because it is known that γ-FeOOH is red rust and Fe3O4 is black rust.15) Here, the background was defined as a straight line connecting the intensities of 2θ = 38° and 2θ = 58°, and the net peak intensity ratio was calculated from the background and gross intensity of Fe3O4 (2θ = 41.375°) and γ-FeOOH (2θ = 55.205°). The same measurement was carried out 2 times considering repeatability. The results showed that the ratio of Fe3O4 to γ-FeOOH increased with increasing test time. This result suggests that the black rust observed in-situ consists of Fe3O4, because the increase in the ratio of Fe3O4 to γ-FeOOH was similar to the expansion of the black rust observed in in-situ as the test time increased as shown in Fig. 2(b) to (e).
X-ray diffraction patterns after (a) 0 h, (b) 3 h, and (c) 6 h of cyclic corrosion test.
Change of Fe3O4/γ-FeOOH peak intensity ratio with test time.
The diffraction patterns obtained by XRD with the different incident angles of (a) 25°, (b) 15°, and (c) 5° using the specimen after 6 h are shown in Fig. 7, and the peak intensity ratio of Fe3O4 (2θ = 41.375°) to γ-FeOOH (2θ = 55.205°) obtained by the same procedure as in Fig. 6 is shown in Fig. 8 as a function of the incident angles. It was found that the ratio of Fe3O4 increased as the incident angle decreased. This means that the ratio of magnetite (Fe3O4) on the surface side of the rust is larger than that at the substrate side of the rust.
X-ray diffraction patterns of incident angle of (a) 25°, (b) 15°, and (c) 5° after 6 h in cyclic corrosion test.
Relationship between incident angle and Fe3O4/γ-FeOOH peak intensity ratio.
The cross-sectional SEM images and the results of Fe, O, Na, Cl mapping by EPMA using the specimen after 6 h are shown in Fig. 9. Rust was formed on the scribed part, and Na and Cl were segregated to the surface and substrate sides of the rust, respectively. From this result, it is assumed that the cathodic reaction occurs on the surface side of the rust and the anodic reaction occurs on the substrate side of the rust. Figure 10 shows the results of Raman spectroscopy of the Na-enriched rust layer on the surface side and the Cl-enriched rust layer on the substrate side, together with the spectra of rust standard materials. The spectrum of the Na-enriched rust layer on the surface side shown in Fig. 10(a) was in relatively good agreement with spectrum (665 cm−1) of the magnetite (Fe3O4) standard material shown in Fig. 10(c). However, the spectrum of the Cl-enriched rust layer on the substrate side showed a mixed spectrum of the γ-FeOOH standard material spectrum (253 cm−1) and the β-FeOOH standard material spectra (720, 383, 309 cm−1), as shown in Fig. 10(b).
(a) Cross-sectional SEM image and cross-sectional EPMA mapping results for (b) Fe, (c) O, (d) Cl, and (e) Na after 6 h of cyclic corrosion test.
Raman spectra of (a) surface layer rust and (b) substrate side rust after 6 h of cyclic corrosion test and standard samples of (c) Fe3O4, (d) γ-FeOOH, and (e) β-FeOOH.
From these results, it can be said that the change to black rust at the scribed part is a stage in which magnetite (Fe3O4) is generated on the surface side of the rust layer.
From the in-situ observation results, red rust (orange rust) was generated at the scribed part where the substrate was exposed immediately after the dry process following salt dip, and the whole scribed part was covered with red rust (Fig. 2(a), (b)). This is thought to be due to the formation of iron oxyhydroxide by the initial dissolution of Fe. Afterwards, the whole scribed part changed from red rust to black rust (Fig. 2(c)–(e)). Evans has proposed that atmospheric corrosion of carbon steel progresses by the redox reaction of iron rust.16) The redox reaction includes a cathodic reaction in which iron oxyhydroxide (FeOOH) changes to magnetite (Fe3O4) under wet conditions.
In the process in which the whole scribed part changes from red rust to black rust, iron oxyhydroxide act as an oxidant causing anodic dissolution of Fe, and the iron oxyhydroxide is reduced to magnetite as described in the Evans model, the anodic dissolution of Fe occurs on the substrate side of the rust because Cl was concentrated on the substrate side after the change to black rust, as shown in Fig. 9. On the other hand, because Na was concentrated and magnetite (Fe3O4) was detected on the surface side of the rust, the cathodic reaction by which iron oxyhydroxide changes to magnetite could occur on the surface side, in addition to the reduction reaction of oxygen based on the Evans model, shown by the following eqs. (1) to (3):
Anodic reaction:
| \begin{equation} \text{Fe}\to \text{Fe$^{2+}$} + \text{2e$^{-}$} \end{equation} | (1) |
Cathodic reaction:
| \begin{equation} \text{O$_{2}$} + \text{4H$_{2}$O} + \text{4e$^{-}$}\to \text{4OH$^{-}$} \end{equation} | (2) |
| \begin{equation} \text{8FeOOH} + \text{Fe$^{2+}$} + \text{2e$^{-}$}\to \text{3Fe$_{3}$O$_{4}$} + \text{4H$_{2}$O} \end{equation} | (3) |
Here, the cathodic reaction by which iron oxyhydroxide is changed to magnetite requires the supply of Fe2+, and Fe2+ would be supplied by the diffusion of Fe2+ generated by the dissolution reaction of Fe on the substrate side to the surface side. This process occurred during the dry process of the 60°C and RH 50% environment, which may be related to the existence of CaCl2 in the test solution. CaCl2 absorbs water in the relative humidity of 31 to 33% at room temperature (20 to 25°C).17) From the in-situ observation results, it appeared that the whole scribed part was covered with a thin water film. This means that a very thin water film was formed due to the existence of CaCl2 even under the 60°C and RH 50% environment, and as the local reaction progressed gradually, the black rust expanded over the whole scribed part.
In the Evans model, it is reported that, under a dry condition, the reduced magnetite in the outer layer is further oxidized to iron oxyhydroxide by oxygen in the gas phase.16) However, according to the in-situ observation results, the black rust [reduced rust] remained without any changes after the scribed part once changed to black rust, and re-oxidation to iron oxyhydroxide was not observed in the subsequent dry process. This behavior can be explained by one of the following two theories. The first possibility is that the oxidation reaction by oxygen in the gas phase or water film as described in the Evans model does not occur under the 60°C and RH 50% environment. The second possibility is that the oxidation reaction of magnetite does not occur even in contact with oxygen in the gas phase or water film. This second possibility is discussed by Suzuki et al.18) as follows. Suzuki et al. analyzed the rust layer of pure iron formed in an exposure test by a cathodic constant current method and concluded that magnetite is not easily oxidized by air when it has been reduced to crystalline magnetite.18) As mentioned above, crystalline magnetite was detected by the XRD analysis in the present study. Therefore, the magnetite was not oxidized by air or the water film in this case because the magnetite consisted of crystalline magnetite.
4.2 Initiation and progress of under-film corrosion from scribed part and circular swelling of tip of under-film corrosionAmong the three steps, this section discussed (2) Initiation and progress of under-film corrosion from the scribed part, and (3) Circular swelling of the tip of under-film corrosion.
In order to understand the very initial behavior of under-film corrosion, Fig. 11 shows cross-sectional SEM images and the results of Fe, O, Na, Cl mapping by EPMA at the position where under-film corrosion occurred slightly after 6 h. The specimen analysed here was the same as that used in Fig. 9. The analysis position is located away from the in-situ observation position. Similar to the Cl distribution observed before the progress of under-film corrosion in Fig. 9, Cl existed in the scribed part. However, Na was distributed not only at the scribed part (A in Fig. 11(e)) but also just below the under-film corrosion (B in Fig. 11(e)). Considering this result, it is assumed that the tip becomes a cathodic site in the initial stage of under-film corrosion. As shown in Fig. 2(b) to (e), under-film corrosion occurs after the change to black rust at the scribed part has been completed. Therefore, when the cathodic reaction due to the formation of magnetite in the scribed part has been completed, the cathodic site moves to the paint film/substrate interface and under-film corrosion starts. As described above, after the cathodic reaction in the scribed part has been completed, a very thin water film continues to exist even in the dry process of the 60°C and RH 50% environment. The existence of this water film prevents the re-oxidation of magnetite by preventing contact with the oxygen in the gas phase or by reduction to crystalline magnetite, suggesting that the reduction reaction cannot be supplemented on the surface side of the rust. That is to say, in the presence of a water film, dissolution of Fe continues to occur, but the cathodic reaction at the scribed part has been completed, so the cathodic site moves to the paint film/substrate interface. Although it is not clear why under-film corrosion starts locally, it is estimated that the cathodic reaction involving iron dissolution occurred at the paint film/substrate interface due to factors such as alkalinization of the paint film/substrate interface by the oxygen reduction reaction, dissolution of the phosphating film, and a slight decrease in the local adhesion of the paint film.
(a) Cross-sectional SEM image and cross-sectional EPMA mapping of (b) Fe, (c) O, (d) Cl, and (e) Na in initial stage after under-film corrosion.
Figure 12 shows the cross-sectional SEM images and results of Cl mapping by EPMA using the specimen taken out in the stage of progress of under-film corrosion. Before progress of under-film corrosion, Cl was concentrated in the rust on the substrate side of the scribed part (Fig. 12(a)), but after progress of under-film corrosion, Cl was concentrated at the tip of under-film corrosion and the Cl in the scribed part disappeared (Fig. 12(b), (c)). This change occurs under the 60°C and RH 50% environment, and salt is not supplied externally. This means that the Cl moves from the scribed part to the tip of under-film corrosion, suggesting the migration of the anodic site. As shown in Fig. 11, it is estimated that the furthest tip of under-film corrosion is cathodic site, and the anodic reaction follows behind it. This result is similar to Funke’s filiform corrosion model.19)
Cross-sectional SEM image and Cl mapping after (a) 6 h, (b) 12 h, and (c) 18 h of cyclic corrosion test.
Since the circular swelling under the 50°C and RH 100% environment occurred only at the tip of the under-film corrosion, it was estimated that the Cl which concentrated in the stage of progress of under-film corrosion affected the growth of circular swelling. Funke reported that a paint film swells by osmotic pressure when a residual solvent exists under the film due to failure of the painting process.19) Although different from a residual solvent, soluble NaCl and CaCl2 seem to be concentrated at the tip, and osmotic pressure may occur when the environment shifts to the wet process with conditions of 50°C and RH 100%. As shown in Fig. 4, scribe creep tended to increase at an almost constant rate under the 60°C and RH 50% environment, after which only the tip of the under-film corrosion swelled in the circular shape and scribe creep increased slightly in the initial stage of the 50°C and RH 100% environment. It may be noted that this swelling occurs in the initial stage of the 50°C and RH 100% environment, but after that stage, its size remains substantially the same. This suggests that the main factor controlling the swelling of the tip is not the electrochemical factor but the physical factor. That is to say, it is considered that the soluble salt concentrated at the tip in the stage when under-film corrosion progresses under the 60°C and RH 50% environment is diluted by the shift to the 50°C and RH 100% environment. The tip initially swells due to the osmotic pressure associated with the dilution, but does not grow after this dilution is completed.
In the 2nd cycle, as shown in Fig. 3 and Fig. 4, progress of under-film corrosion and circular swelling were repeated, and which can be explained by the behavior observed in the 1st cycle. On the other hand, an increase in scribe creep was not observed in the initial stage of the 60°C and RH 50% environment immediately after salt dip (Fig. 4(e)). From the results of the in-situ observation, this corrosion behavior is considered to be different from that of the scribed part observed in the 1st cycle because the scribed part is already black rust in the 2nd cycle (Fig. 3(f), (g)). It is assumed that the concentrated salt at the tip of the under-film corrosion in the 1st cycle was diluted during the salt dip process at the beginning of the 2nd cycle, and as a result, reconcentration of the salt in the tip was delayed.
From the above, under-film corrosion under the wet and dry conditions is caused by the repetition of two different steps of under-film corrosion with the morphology of filiform corrosion and corrosion with swelling by osmotic pressure. The results of in-situ observation revealed the existence of many small water droplets on the paint film and scribed part under the 50°C and RH 100% environment, which is different from the condition under the 60°C and RH 50% environment. This suggests that different modes of under-film corrosion occur depending on the difference in the thickness of the formed water film. That it, under-film corrosion with the morphology of filiform corrosion occurs under a thin water film, whereas the soluble salt concentrated in the tip is diluted under a thick water film. In other words, osmotic pressure due to the dilution of the soluble salt, and as a result, water flows under the paint film, causing circular swelling, and the size of the circular swelling seems to be determined by the concentration of the salt in the stage of progress of under-film corrosion with the morphology of filiform corrosion.
In addition, this study clarified for the first time the fact that the progress rate of scribe creep is determined by the rate of under-film corrosion with the morphology of filiform corrosion, as shown in Fig. 4. In general, the cosmetic corrosion resistance of steel sheets is evaluated by the width of scribe creep in each cycle in “one-sided scribe creep”, which is an index for evaluating the cosmetic corrosion resistance of automotive steel sheets. However, we believe that it is important to understand the under-film corrosion behavior in one cycle consisting of dry and wet conditions in order to evaluate materials accurately and to discuss the mechanism of corrosion under these different conditions.
4.3 Mechanism of initial stage cosmetic corrosion under wet and dry conditionsFigure 13 shows schematic images of the mechanism of initial under-film corrosion of a cold-rolled steel sheet without galvanizing under wet and dry conditions, as summarized below.
Model of initial stage cosmetic corrosion under cyclic wet and dry environment. (a) Change of red rust to black rust at scribe creep. (b) Progress of under-film corrosion from scribe creep. (c) Growth of tip of under-film corrosion.
(1) Change to black rust at the scribed part (Fig. 13(a))
Iron oxyhydroxide is generated in the initial stage of the dry process following the salt dip process. Iron oxyhydroxide acts as an oxidant, causing anodic dissolution of Fe, and the iron oxyhydroxide of the oxidant is reduced to magnetite. Under-film corrosion does not occur during this period.
(2) Initiation and progress of under-film corrosion from the scribed part (Fig. 13(b))
When the cathodic reaction due to the formation of magnetite in the scribed part is completed, the cathodic site moves to the paint film/substrate interface and under-film corrosion starts. With the movement of the cathodic reaction, the anodic site also moves from the scribed part to under the paint film, and the under-film corrosion progresses. This under-film corrosion progresses based on Funke’s filiform corrosion model. It should be noted that this sequence of processes occurs under a thin water film.
(3) Circular swelling of the tip of under-film corrosion (Fig. 13(c))
The phenomenon of circular swelling is caused by the shift to the wet process under a thick water film. The soluble salt concentrated at the tip in the stage when under-film corrosion progresses is diluted, and swelling occurs due to the osmotic pressure associated with this dilution.
From the 2nd cycle, (2) Progress of under-film corrosion and (3) Circular swelling of the tip of under-film corrosion are repeated. The rate of under-film corrosion with the morphology of filiform corrosion determines the progress rate of scribe creep. The sequence of processes from the change to black rust to the progress of under-film corrosion occurs under a thin water film because these processes occur in the dry process with the 60°C and RH 50% environment. Because it has been reported that the morphology of under-film corrosion of cold-rolled steel sheet without galvanizing observed in automotive outer panels in actual environments is filiform,20) it is thought that under-film corrosion would progress in the same manner as described above in the actual environment.
In order to clarify the mechanism of under-film corrosion, an in-situ observation technique was developed. The following conclusions were obtained through in-situ observation of the process of under-film corrosion under wet and dry conditions.
