Mechanics of Materials
In-situ Observation of Cosmetic Corrosion Behavior under Atmospheric Environment
2023 Volume 64 Issue 11 Pages 2596-2605
Details
2023 Volume 64 Issue 11 Pages 2596-2605
It is well known that the types of automotive corrosion can be divided into perforation and cosmetic corrosion. Although many studies concerning the mechanism of perforation corrosion have been conducted so far, very few have investigated the mechanism of cosmetic corrosion. In our previous work, the authors found that the initial corrosion behavior in cosmetic corrosion consists of three steps by conducting in-situ observation during a cyclic corrosion test. In the present work, in-situ observation of painted samples with a scratch was performed during an exposure test in Okinawa to examine the cosmetic corrosion behavior under an actual environment. It was found that corrosion started at the moment of salt deposition around the scribed part from air containing salts, and the initial corrosion behavior in the exposure test consisted of three steps, which were the same as in the cyclic corrosion test. In the 1st step, black rust of Fe3O4 formed at the scribed part, and in the 2nd step, under-film corrosion progressed from the scribed part where the black rust had formed. In the 3rd step, the tip of the under-film corrosion displayed swelling behavior. The behavior of each step was also discussed by combining the in-situ observation results with an analysis of environmental factors such as relative humidity and an EPMA analysis.
This Paper was Originally Published in Japanese in Zairyo-to-Kankyo 70 (2021) 18–27. The captions of Figs. 3, 4, 5, 6, 7, 12 and 13 are slightly modified.
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 is initiated from paint defects caused by chipping or from the material edge.1) Various technologies and products such as corrosion-resistant steel sheets, wax, sealers and painting have been developed for both types of automotive corrosion. However, from the viewpoints of resource saving and cost reduction, it is important to optimize corrosion resistance depending on the environments in which automobiles are to be used so as to avoid excessive corrosion resistance. In order to propose the optimum corrosion resistance, it is necessary to understand the environmental factors which affect perforation corrosion and cosmetic corrosion.
The mechanism of perforation corrosion has been studied extensively based on an analysis of actual vehicles,2) and the corrosion behavior of corrosion-resistant steel sheets under various environments has also been investigated.3) On the other hand, in the case of 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 also been studied in many reports.6,7) However, the specimens observed in these previous studies were usually examined after corrosion tests or exposure tests consisting of a few dozen cycles including both wet and dry conditions. Therefore, the results obtained by these conventional methods are not able to show the difference between the corrosion behaviors under wet and dry conditions because the specimens are taken out and analyzed after exposure to both conditions. This suggests that in-situ observation of the corrosion behavior would be a more suitable approach for clarifying the corrosion behavior in each wet or dry condition. In our previous work, 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 painted cold-rolled steel sheets with paint defects. We also discussed the mechanism of under-film corrosion considering both the results of in-situ observation of under-film corrosion and analysis of the specimens.8–11) The conclusions obtained in the investigation of scribed specimens exposed to a salt dip process, dry process and wet process were as follows:11)
These results were obtained in a cyclic corrosion test. Generally, the cyclic corrosion test is performed under controlled conditions such as constant temperature, constant humidity and a constant amount of deposited salt by dipping in salt water. However, the amount of deposited salt on automotive parts where cosmetic corrosion occurs, such as the front edge of the hood and the lower edge of doors, is not always constant in actual environment because the deposited salt is considered to be caused by airborne salt. In addition, in actual environments, the temperature and humidity also change from moment to moment by repeating day and night. Therefore, the under-film corrosion behavior in an actual environment may be different from that in a cyclic corrosion test.
In this study, in-situ observation of the under-film corrosion in an exposure test under actual environments was carried out using the in-situ observation technique used in our previous study.11) The temperature, relative humidity and amount of deposited salt were also measured during the exposure test. The relationship between these environmental factors and under-film corrosion behavior was discussed and compared with that of the under-film corrosion behavior in the cyclic corrosion test.11)
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 70 mm × 150 mm (thickness: 0.8 mm). The chemical composition is shown in Table 1. The specimens were alkaline degreased, phosphated and then coated by cathodic electrodeposition to a thickness of 15 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 the 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 exposure test was carried out at Uruma City, Okinawa Prefecture under a condition sheltered from rainfall for approximately 3 months beginning at 10:00 on July 7, 2017, as shown Fig. 2. The test specimens were set to face south.
Exposure environment of test specimens.
The monthly amounts of airborne salt during the exposure test were measured by the dry gauze method. Humidity and temperature sensors (Syrinx Corp., STHM11S2) and an ACM sensor (Syrinx Corp., Fe–Ag coupling type) were also set up in the same environment. The measurement interval of each sensor was 10 minutes.
2.3 In-situ observation during atmospheric exposure testIn-situ observation under the condition sheltered from rainfall was carried out by using the same device13) applied in our previous study.11) The test specimen was set with an angle of 90° to the ground surface and the longitudinal direction of the specimen vertical. 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 minutes. The images were then combined to create a moving picture.
The under-film corrosion in one location was selected from the video images using a test specimen subjected to the exposure test, and scribe creep was measured at prescribed intervals. Cross-sections of the specimens were taken out after the exposure test observed and analyzed with a scanning electron microscope (SEM, Hitachi High-Tech Corp., S-3700N, accelerating voltage: 15 kV) and electron probe micro analyzer (EPMA, JEOL JXA8230, accelerating voltage: 15 kV, current: 5.007 × 10−8, dwell time: 40 ms).
Figure 3 shows the in-situ observation results of the surface of a specimen which was exposed under the eaves of the test facility to prevent contact with rainfall from July 7 to September 30, 2017, in Okinawa. For approximately two months, from the start of the exposure test at 10:00 on July 7 until September 13, initiation of corrosion was not observed within the in-situ observation area, which included the scribed part and the paint film (Fig. 3(a), (b)). Initiation of the corrosion was observed at the scribed part on September 14 (Fig. 3(c)), but the corrosion occurred in only one area in the scribed part (white circle in Fig. 3(c)). Afterwards, corrosion progressed partially in the scribed part, but did not occur some areas (Fig. 3(d)). The rust generated in this corroded area was a mixture of red rust (orange rust) and black rust. According to the obtained in-situ observation movie, the steel corroded at several positions in the scribed part from September 14 to September 15 in addition to the initial corroded area shown in Fig. 3(c), and the corrosion expanded gradually. On September 18, under-film corrosion occurred and progressed from some positions in the scribed part (Fig. 3(e), (f)). On September 19, only the tip of some under-film corrosion sites swelled in a circular shape (white circle in Fig. 3(g)). Then, until September 30, the under-film corrosion progressed again from the tip of the circular swelling as shown in Fig. 3(h). In addition, under-film corrosion was newly initiated from several positions in the scribed part, and both the progress of under-film corrosion and the circular swelling of the tip of the under-film corrosion were observed again.
Appearance of samples after (a) 10:00 July 7, (b) 0:00 September 13, (c) 0:00 September 14, (d) 12:00 September 14, (e) 0:30 September 18, (f) 20:30 September 18, (g) 6:30 September 19 and (h) 0:00 September 30, 2017 after exposure test in Okinawa.
The authors carried out in-situ observation of the initial stage of under-film corrosion and clarified the corrosion behavior under the environment of one cycle consisting of salt dip, dry and wet conditions.11) In this section, the under-film corrosion behavior in the exposure test in Okinawa is compared with that in the cyclic corrosion test reported previously. Figure 4 shows the in-situ observation results of the Okinawa exposure test (Fig. 4(a)–(c)14)) and the cyclic corrosion test (Fig. 4(d)–(f)11)) before and after the initiation of under-film corrosion. In our previous study, the under-film corrosion occurred and progressed after the black rust had covered the whole scribed part (Fig. 4(d)–(f)). Similarly, in the Okinawa exposure test, the under-film corrosion occurred near the black rust area in the scribed part (Fig. 4(a)–(c)). On the other hand, unlike the results of the cyclic corrosion test, the black rust did not cover the whole scribed part, and red rust and uncorroded areas remained. Initiation of under-film corrosion was not observed at positions near the red rust and uncorroded areas.
Subsequently, the under-film corrosion behavior, including the circular swelling of the tip of under-film corrosion, was compared. Figure 5 shows the in-situ observation results of the Okinawa exposure test (Fig. 5(a), (b)14)) and the cyclic corrosion test (Fig. 5(c), (d)11)) before and after swelling of the tip of the under-film corrosion. Here, Fig. 5(a) and (c) show photographs before swelling of the tip of the under-film corrosion, and Fig. 5(b) and (d) show photographs after swelling of the tip of the under-film corrosion. In our previous study, the tip of the under-film corrosion swelled in a circular shape after under-film corrosion progressed (Fig. 5(c) and (d)). Similarly, in the Okinawa exposure test results shown in Fig. 5(a) and (b), the circular swelling grew gradually at the tip of the under-film corrosion with the filiform corrosion morphology, and this behavior was consistent with the results of cyclic corrosion test. However, the size of the circular swelling observed in the Okinawa exposure test was smaller than that in the cyclic corrosion test.
From the comparison of the under-film corrosion behavior in the exposure test with that in the cyclic corrosion test, the features of the corrosion behavior in the actual environment, such as the formation of a localized black rust area in the scribed part and the smaller size of the circular swelling, were different from those in cyclic corrosion test. However, the following three corrosion steps which occurred under the actual environment were the same as those observed in the cyclic corrosion test: (1) Change to black rust at the scribed part, (2) Initiation and progress of under-film corrosion from near the black rust area in the scribed part, (3) Circular swelling of the tip of under-film corrosion.
3.3 Comparison of appearance of specimens after exposure test with results of in-situ observationFigure 6 shows the appearance (Fig. 6(a)) and microscopic images (Fig. 6(b)–(e)) of the specimens after the Okinawa exposure test for three months. (In the microscopic images, Fig. 6(b): image of area in black rectangle indicated by symbol A in Fig. 6(a), Fig. 6(c): area indicated by symbol B in Fig. 6(a), Fig. 6(d): area indicated by symbol C in Fig. 6(b), Fig. 6(e): area indicated by symbol D in Fig. 6(c).) Figure 6(a) shows that under-film corrosion occurred and then progressed from several positions in the scribed part. Macroscopically, these under-film corrosion sites appear to have a mixture of two different morphologies: (1) under-film corrosion with the filiform corrosion morphology (hereinafter to referred as filiform-type corrosion) and (2) under-film corrosion with the circular swelling morphology (hereinafter referred to as swelling-type corrosion).
Appearance of samples exposed in Okinawa.
Detailed observation of (1) filiform-type corrosion (Fig. 6(d)) revealed that small circular swelling existed at the tip of the under-film corrosion and between the scribed part and the tip of the under-film corrosion (white circles with arrow in Fig. 6(d)). According to the in-situ observation results shown in Fig. 3(e)–(g), the tip of the under-film corrosion swelled in a circular shape after the progress of the under-film corrosion with the filiform corrosion morphology. In addition, once the under-film corrosion had progressed, the under-film corrosion progressed again from the tip of the circular swelling. It should be noted that the size and shape of the circular swelling shown in Fig. 6 were different from the results of the in-situ observation because the images in Fig. 6 show the appearance of the specimens after the full duration of the exposure test, and in this case, the amount of moisture and the kinds of corrosion products under the paint film would be different from those of the specimens during the exposure test. However, the existence of small circular swelling between the scribed part and the tip of the under-film corrosion is considered to be evidence of the repetition of both the progress of the under-film corrosion with the filiform corrosion morphology and the growth of circular swelling at the tip of the under-film corrosion obtained by in-situ observation.
On the other hand, when (2) swelling-type corrosion was observed in detail (Fig. 6(e)), small under-film corrosion was observed at the top end of the circular swelling (indicated by the arrow in Fig. 6(e)). According to the in-situ observation results, the under-film corrosion progressed again from the tip of the circular swelling. The circular swelling and the small under-film corrosion at the tip after the exposure test are considered to be the results of the repetition of both the progress of the under-film corrosion with the filiform corrosion morphology and the growth of circular swelling at the tip of the under-film corrosion, as in the case of (1) filiform-type corrosion described above.
Thus, the macroscopically observed under-film corrosion with the two different morphologies of (1) filiform-type corrosion and (2) swelling-type corrosion are considered to have been formed by the repetition of filiform corrosion and circular swelling, although their sizes and the number of repetitions were different.
As shown in Fig. 3, the initiation of corrosion was not observed within the in-situ observation area including the scribed part or on the paint film for a period of approximately two months from the start of the exposure test at 10:00 on July 7 until September 13 (Fig. 3(a), (b)). The initiation of corrosion was first observed at the scribed part on September 14 (Fig. 3(c)). Here, the influence of airborne salt is discussed in order to identify the factor responsible for the initiation of corrosion at the scribed part.
During the exposure test, the amount of the airborne salt was 0.20 mdd from July 1 to July 31, 0.76 mdd from August 1 to August 31, and 2.24 mdd from September 1 to September 30, indicating that the amount of airborne salt in September was very large. On the other hand, although the initiation of corrosion in the scribed part was observed on September 14, it is not clear when the airborne salt was actually deposited on the specimen because the amount of the airborne salt was measured by the dry gauze method. Motoda et al.15) clarified the relationship between the sensor output (I) of Fe–Ag type ACM sensors and the relative humidity (RH) at each amount of deposited salt under the conditions of constant temperature and humidity, and that relationship was used to measure the amount of deposited salt during each duration in various exposure tests.15–17) The amount of salt deposited on the ACM sensor surface during the exposure test was calculated every 12 hours by using this calibration curve. As shown in Fig. 7, a large amount of salt was deposited on the ACM sensor surface on September 14, 2017, when the initiation of corrosion was observed in the scribed part. According to the climate records of the Japan Meteorological Agency, typhoon No. 18 formed on September 9 and was approaching Okinawa Island from September 13 to September 16 (Fig. 818)). This suggests that the deposition of a large amount of salt on the surface of the ACM sensor was caused by the typhoon. Although salt was also deposited on the surface of the ACM sensor before September 14 (for example, from August 21 to August 22), initiation of corrosion was not observed in the in-situ observation area. Here, the width of the scribed part of the painted specimen is about 100 µm, and the salt deposited within this narrow part would affect the corrosion of the scribed part. The salt could reach the scribed part for the first time in the exposure test on September 14, because the amount of the airborne salt on September 14 was larger than before that date. It would be enough to reach 100 µm of the scribed part, although the salt was deposited on the ACM sensor surface before September 14. As a result, corrosion was initiated at the scribed part. Furthermore, according to the in-situ observation results, in addition to the initial corroded area shown in Fig. 3(c), the number of corroded areas in the scribed part increased from September 14 to September 15, and their areas gradually expanded. This suggests that the salt was deposited on several positions in the scribed part on September 14, when a large amount of the airborne salt arrived at the test site. On the other hand, almost no salt was deposited on the ACM sensor after September 14, implying that little salt reached the test site. Based on these results, it is assumed that the corrosion behavior, including the progress of under-film corrosion after the initiation of corrosion in the scribed part, was caused by the salt that was deposited on the scribed part on September 14.
Time variation of stationary amount of sea salt deposited on ACM sensor in exposure test in Okinawa from July 7 to September 30, 2017.
Route of Typhoon No. 18 in 2017.18) Based on “TROPICAL CYCLONE TRACKS, International number ID: 1718” (Japan Meteorological Agency).
As shown in Fig. 4, the under-film corrosion occurred near the black rust area in the scribed part after the black rust had formed in the scribed part. This behavior is common to the Okinawa exposure test and the cyclic corrosion test. Here, the initiation of under-film corrosion from the scribed part is discussed in detail, comparing the results of the exposure test with those of the cyclic corrosion test.11)
Firstly, the behavior of the black rust that formed in the scribed part is discussed. Under-film corrosion occurred near the black rust area in the scribed part in the Okinawa exposure test (Fig. 4(a), (b)). This behavior was consistent with the results of the cyclic corrosion test, suggesting that iron oxyhydroxide reduced to magnetite based on the Evans model.19) On the other hand, red rust (orange rust) and uncorroded areas remained in the scribed part except at positions near those where under-film corrosion was initiated. According to the in-situ observation results during the cyclic corrosion test, the black rust covered the whole scribed part before the initiation of under-film corrosion. The behavior of the change to black rust in the scribed part was not completely consistent between the exposure test and the cyclic corrosion test. This difference is thought to be caused by the difference in the salt supply process. In the Okinawa exposure test, the salt could not be deposited uniformly on the whole scribed part of the specimen because airborne salt was transported to the scribed part little by little, as described above in section 4.1, whereas in the cyclic corrosion test, the salt was supplied almost uniformly over the whole scribed part because the specimen was dipped in the aqueous salt solution before the dry and wet processes. Therefore, in early stage of the Okinawa exposure test, uniform black rust like that in the cyclic corrosion test would not occur, because the corrosion reaction occurred only at local positions of the scribed part where salt was deposited. Although the salt deposition area was different, the fact that under-film corrosion was initiated near the black rust formation area in the scribed part in the Okinawa exposure test is similar to that in the cyclic corrosion test. Thus, it can be concluded that the initiation of under-film corrosion in the Okinawa exposure test and the cyclic corrosion test occurs by a similar process.
4.3 Progress of under-film corrosion and circular swelling of tip of under-film corrosion under atmospheric corrosion environment 4.3.1 Change of morphology of under-film corrosion and influence of relative humidity on progress rate of under-film corrosionAs shown in Fig. 5, circular swelling of the tip of under-film corrosion was observed at the position where the under-film corrosion had progressed in both the Okinawa exposure test and the cyclic corrosion test. This section discusses the progress of under-film corrosion in the scribed part and the circular swelling of the tip of the under-film corrosion in the Okinawa exposure test in contrast to that in the cyclic corrosion test,11) focusing on the change of the relative humidity.
The change of relative humidity during the Okinawa exposure test is shown in Fig. 9. The moments of both the initiation of under-film corrosion and the occurrence of circular swelling obtained from in-situ observation are also shown in Fig. 9. It was found that the circular swelling of the tip of under-film corrosion occurred when the relative humidity exceeded 90% after under-film corrosion had initiated and progressed. In our previous report concerning under the cyclic corrosion test, the under-film corrosion with the filiform corrosion morphology progressed at 60°C and RH 50%, and then the tip of the under-film corrosion swelled in a circular shape when these conditions shifted to the 50°C and RH 100%.11) This behavior suggests that the swelling occurred due to the osmotic pressure because the soluble salt concentrated at the tip of the under-film corrosion and was diluted by the shift to 50°C and RH 100%.11) Unlike the humidity in the cyclic corrosion test, the humidity in the Okinawa exposure test was not constant. However, the under-film corrosion with the filiform corrosion morphology could also progress under a thin water film at lower humidity, and then the circular swelling of the tip of the under-film corrosion would occur under a thick water film at a high humidity above 90%.
Time variation of relative humidity at Okinawa from September 18 to September 20, 2017.
Next, the progress rate of under-film corrosion depending on the environment during the Okinawa exposure test is discussed. The relative humidity and scribe creep were measured during the exposure test, and the results are shown in Fig. 10 as a function of test time. Here, scribe creep was measured from the in-situ observation at intervals of 2.5 hours with an accuracy of 0.01 mm. Scribe creep increased during periods when the relative humidity was low in daytime or when the relative humidity increased from daytime to nighttime (areas highlighted in gray in Fig. 10). The increase of scribe creep was particularly high when the relative humidity kept a low value for a long time (duration designated by the two-way arrow in Fig. 10). This suggests that the progress rate of scribe creep is determined by the rate of under-film corrosion with the filiform corrosion morphology under a thin water film. This result agrees with that in our previous report, which demonstrated that the progress rate of under-film corrosion with the filiform corrosion morphology substantially determines the progress rate of scribe creep within 1 cycle.11) The reason why scribe creep hardly increased during periods when the relative humidity was high in nighttime or when the relative humidity decreased from nighttime to daytime is that the Cl which had concentrated at the tip was diluted by the osmotic pressure under a high relative humidity environment, and time was necessary for Cl to concentrate in the tip again under a low relative humidity environment.
Time variation of relative humidity and scribe creep at Okinawa from September 18 to September 30, 2017.
The periods when scribe creep increased were selected (areas highlighted in gray in Fig. 10), and the results of the increase of scribe creep are shown in Fig. 11 as a function of the average relative humidity. The ranges of the relative humidity in each period are also shown in Fig. 11 as error bars. The average relative humidity when scribe creep increased was less than RH 90%. This result is basically consistent with the work of Funke,12) which reported that filiform corrosion progressed at less than RH 85%. In addition, the average relative humidity during the period shown in the upper left side of Fig. 11, where the progress rate of scribe creep was the largest, was RH 51% and was the lowest compared to other periods. That is to say, from these results, it can be concluded that the rate of under-film corrosion with the filiform corrosion morphology also determines the progress rate of scribe creep under an atmospheric corrosion environment.
Relationship between relative humidity and increase of scribe creep.
As shown in Fig. 6, the specimens after the end of the exposure test have two macroscopically different morphologies of under-film corrosion, which shows the appearances of (1) under-film corrosion with the filiform corrosion morphology (filiform-type corrosion) and (2) under-film corrosion with the circular swelling morphology (swelling-type corrosion). On the other hand, from a comparison of the under-film corrosion behavior in the Okinawa exposure test with that in the cyclic corrosion test, the behavior of both is considered to be consistent in terms of repeating the processes of filiform corrosion and circular swelling. In order to confirm the agreement of these corrosion behaviors, a cross-sectional analysis was carried out for both the corrosion which was (1) filiform-type corrosion (part indicated by the arrow in Fig. 6(d)) and (2) swelling-type corrosion (part indicated by the arrow in Fig. 6(e)). Here, all the deposited salt is assumed to remain on the specimen because the exposure test was carried out under a roof to protect the specimens from rainfall. Since local differences in the amount of deposited salt would affect the morphology of the under-film corrosion, the relationship between the Cl distribution and the macroscopic morphology of the under-film corrosion was investigated.
Figure 12 shows cross-sectional SEM images and the results of Fe, O and Cl mapping by EPMA at the position which had the morphology of (1) filiform-type corrosion. Cl existed under the paint film and was concentrated up to the tip of the under-film corrosion. Our previous report11) confirmed that Cl moves to the tip of under-film corrosion, suggesting the migration of the anodic site. This result is similar to Funke’s filiform corrosion model.12) Therefore, it is suggested that the Cl which existed to the tip of the under-film corrosion in the Okinawa exposure test moved to the tip of the corrosion under the paint film.
(a) Cross-sectional SEM image and (b) cross-sectional EPMA mapping of Fe, (c) O and (d) Cl after exposure test in Okinawa, observing under-film corrosion with the filiform corrosion morphology.
Among the positions other than the tip, there was a position with a slightly higher Cl intensity than other positions, and the paint film was slightly swollen at that position (white circle in Fig. 12). This position corresponds to the circles on the side of the scribed part in Fig. 6(d). According to our previous report,11) it is clear that this circular swelling occurs due to the osmotic pressure caused by the shift to the wet process after the under-film corrosion with the filiform corrosion morphology. In addition, the progress of the under-film corrosion with the filiform corrosion morphology and the growth of circular swelling at the tip of the under-film corrosion are repeated under the dry and wet conditions. Based on that discussion, the position of the white circles in Fig. 12 is thought to correspond to the position where the circular swelling due to the osmotic pressure occurred after the progress of under-film corrosion with the filiform corrosion morphology. Furthermore, the progress of the under-film corrosion with the filiform corrosion morphology and the growth of circular swelling at the tip of the under-film corrosion are repeated again from the tip of the swelling. In addition to the appearance of the corrosion shown in Fig. 6(d), the results of the Cl distribution also showed that the progress of under-film corrosion with the filiform corrosion morphology and the growth of circular swelling at the tip of the under-film corrosion occur repeatedly.
Figure 13 shows the cross-sectional SEM images and the results of Fe, O and Cl mapping by EPMA at the position which had the morphology of (2) swelling-type corrosion. Cl also existed under the paint film and was concentrated up to the tip of the under-film corrosion as well as the results for the position with the morphology of (1) filiform-type corrosion shown in Fig. 12. A higher Cl intensity was detected at the position with the circular swelling appearance (white circle in Fig. 13) than at the position where the circular swelling in (1) filiform-type corrosion was observed, as shown in Fig. 12. The amount of locally deposited salt at the scribed part in the early stage would be relatively higher at this position than at the position of (1) filiform-type corrosion. As a result, a relatively higher amount of Cl was concentrated at the tip of the under-film corrosion during the progress of under-film corrosion. During the subsequent growth of the circular swelling by osmotic pressure, the swelling presumably grew larger because the highly concentrated soluble salts were diluted, resulting in a large circular swelling, which is different from (1) filiform-type corrosion. In addition, the Cl was further concentrated at the leading tip. It is thought that this occurred because the under-film corrosion with the filiform corrosion morphology progressed again from the tip of the swelling.
(a) Cross-sectional SEM image and (b) cross-sectional EPMA mapping of Fe, (c) O and (d) Cl after exposure test in Okinawa, observing under-film corrosion with the circular swelling morphology.
The above results imply that the amount of locally deposited salt under the paint film is dependent on the position where under-film corrosion progresses because salt is deposited locally on the specimen surface by airborne salt in the actual environment. Therefore, the size of circular swelling due to osmotic pressure associated with dilution depends on the amount of deposited salt. It is not clear why the progress of under-film corrosion after the large growth of the circular swelling of (2) swelling-type corrosion in Fig. 6(e) was delayed compared with the progress of the under-film corrosion of (1) filiform-type corrosion in Fig. 6(d). However, to summarize these results, it can be said that both types of under-film corrosion progress by a similar process, although the macroscopic morphology is different, and it is assumed that the difference in the macroscopic morphology may be caused by the difference in the amount of locally deposited salt in the scribed part in the early stage.
4.4 Mechanism of cosmetic corrosion under atmospheric corrosion environmentFigure 14 shows schematic images of the mechanism of under-film corrosion of a cold-rolled steel sheet without galvanizing under an atmospheric corrosion environment. As in the previously-reported cyclic corrosion test,11) the cosmetic corrosion behavior under the actual environment consisted of the following three steps although some specific behaviors were observed in the actual environment, as noted in the following.
Model of initial stage cosmetic corrosion under actual environment. (a) Deposition of salt at scribed part. (b) Initiation of corrosion and formation of black rust at scribed part. (c) Expansion of black rust area. (d) Initiation of under-film corrosion from scribed part. (e) Progress of under-film corrosion from scribed part. (f) Growth of tip of under-film corrosion.
(1) Change to black rust at the scribed part (Fig. 14(a)–(c))
“Change to black rust” occurs as a result of the reduction reaction of iron oxyhydroxide to magnetite in the scribed part before the beginning of under-film corrosion. Unlike the cyclic corrosion test, in the atmospheric exposure test, the change to black rust occurs locally in the scribed part because salt is deposited locally in the early stage of the exposure test of the scribed part under an actual environment.
(2) Initiation and progress of under-film corrosion from the scribed part (Fig. 14(d), (e))
When the cathodic reaction due to the formation of magnetite in the scribed part is completed, the cathodic site moves to the interface of the paint film and the substrate, and then under-film corrosion is initiated. With the movement of the cathodic reaction, the anodic site also moves from the scribed part to under the paint film, and 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 at a humidity of less than RH 90%. The progress of under-film corrosion with the filiform corrosion morphology determines the progress rate of scribe creep.
(3) Circular swelling of the tip of under-film corrosion (Fig. 14(f))
The phenomenon of circular swelling is caused by the shift to the wet condition under a thick water film at a humidity above RH 90%. The soluble salt which 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.
Following this sequence of corrosion processes, (2) Progress of under-film corrosion and (3) Circular swelling of the tip of under-film corrosion are repeated.
On the other hand, two macroscopically different morphologies of under-film corrosion are observed, which show the appearances of (1) under-film corrosion with the filiform corrosion morphology (filiform-type corrosion) and (2) under-film corrosion with the circular swelling morphology (swelling-type corrosion), as described in 4.3.2, although under-film corrosion of both progresses by a similar process. While the reason for the difference in the macroscopic morphology is not clear, it is thought to be caused by the difference in the amount of local salt deposition on the scribed part in the early stage.
It should be noted that this paper has discussed under-film corrosion behavior under a roof to protect the specimens from rainfall. In studying the behavior of under-film corrosion under an environment without a roof, it is necessary to consider the effect of rainfall washing away the deposited salt. Therefore, the moment of corrosion initiation and the progress rate of under-film corrosion would be different from that under the roof. In addition, the frequency of swelling by osmotic pressure under a thick water film would also be different because of the moisture supplied by rain, so the macroscopic morphology of the under-film corrosion might be also different. However, as described in this paper, the basic mechanism of under-film corrosion is similar because under-film corrosion progresses under an environment of repeated dry and wet processes with deposited salt.
In order to clarify the mechanism of under-film corrosion under an atmospheric corrosion environment, in-situ observation during an exposure test in Okinawa was carried out using cold-rolled steel sheets without galvanizing as the test specimens. The behavior of under-film corrosion under the atmospheric corrosion environment was compared with that in the cyclic corrosion test. The following conclusions were obtained.
