Effect of Mixed Salts on Atmospheric Corrosion of 304 Stainless Steel

• Users may freely distribute the URL that is used to identify this publication. • Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. • User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?) • Users may not further distribute the material nor use it for the purposes of commercial gain.

In marine environments, airborne sea salt particles (salt aerosols) can be deposited on stainless steel surface by coastal winds. 1 Salt will absorb moisture from the environment when its deliquescence relative humidity (DRH) is below the relative humidity (RH) of the surrounding environment until the vapor pressure of the solution reaches that of the environment. If the solution is aggressive enough, particularly in the presence of chloride ions, localized corrosion can start. The degree of deliquescence varies with salt composition since salts have specific DRH values. For example, the DRH of the two major constituents of sea water, NaCl and MgCl 2, are 75% and 33%, respectively. 2 The composition of deposited aerosols varies with the studied area. 1,3,4 However, most work on chloride salts to date has focused either on single salts such as MgCl 2 5-8 and NaCl, [9][10][11] or artificial sea water. 5,12 For 304 and 403 stainless steel, pitting was more likely to take place under artificial sea water than MgCl 2 droplets. 5,12 There is a need to study the effect of mixed salts with different compositions on stainless steel. This work focuses on a mixture of NaCl and MgCl 2 as these are two major components of artificial sea water. 13 It is difficult to observe pit propagation with time since the growth takes place under droplets that develop rust layers. However, in situ synchrotron X-ray microtomography makes it possible to visualize the growth of pits in a non-destructive way. This method has been used in recent years to study in situ corrosion processes. [14][15][16][17][18][19][20][21] In the present work, the effect of mixed salts is investigated at an RH above the DRH of MgCl 2 but below that of NaCl. For ex situ lab-based tests, characterization was carried out at the end of the test whereas for synchrotron X-ray microtomography tests, in situ characterization was performed at several points during the exposure period.

Experimental
Material.-For lab-based tests, 304 stainless steel foils (annealed, 100 μm in thickness, supplied by Goodfellow) were cut into 3 cm × 6 cm. For synchrotron microtomography tests, 304 stainless steel pin samples (2 mm in diameter) were cut from a rod (drawn, 2 mm in diameter, supplied by Goodfellow Droplet deposition.-The foil and pin samples were ground with SiC papers to P800 and washed with de-ionized water (Millipore > 15 M .cm) and methanol. 0.5 M MgCl 2 and 0.25 M MgCl 2 + 0.5 M NaCl (1MgCl 2 :2NaCl) were used for droplet deposition. A droplet, containing 1.8 μL of solution, was deposited onto the foil/pin surface. In lab-based tests, the droplet diameter was 2.2-2.8mm, giving a chloride deposition density (CDD) of 900 to 1600 μg/cm 2 . In tomography tests, the droplet diameter was 2 mm, giving a CDD of 2000 μg/cm 2 .
Relative humidity and temperature.-In lab-based tests, after deposition of droplets, foils were put into a transparent desiccator. The 45% exposure RH was controlled by 4.25 M MgCl 2 5 or saturated K 2 CO 3 . 2 The 21 • C test temperature was controlled by an atmospheric chamber. OMEGA OM-EL-USB-2-LCD data loggers were used to monitor the RH and temperature during the corrosion test. In tomography tests, filter paper soaked with 4.25 M was fitted into a silicone tube, which were slipped over the pin. Tomography tests were carried out at 21 ± 1 • C.

Ex situ lab-based corrosion tests.-For lab-based experiments,
the corrosion behavior of 304 foils under MgCl 2 after 48 h exposure time and under mixed salts MgCl 2 + NaCl after 24 h, 48 h and 72 h exposure time was investigated. After exposure tests, a Leica DFC 420 light optical microscope (OM) was used to characterize the pit size. The width of a pit refers to the diameter of a circle which has the same area as the pit area. The depth of a pit is defined as the greatest depth found among different parts of a pit. Pit depths were measured by OM by using the depth of focus.
In situ synchrotron X-ray microtomography tests.-The design of the cell for in situ X-ray microtomography studies has been detailed previously. 22 Samples were scanned regularly to examine the timedependent pit growth. The tomography experiments using 70 keV X-rays were performed on the Beamline I12 23 at Diamond Light Source. The pin was rotated and radiographs were collected at intervals 0.1 • per step through 180 • . During scanning, two exposure times were used: 1.0 s exposure time was used with a pixel size of 1.8 μm, and 2.2 s exposure time was used with a pixel size of 0.9 μm. The total scanning time per sample was ∼40 minutes for the lower magnification scans and ∼80 minutes for the high magnification scans. Filtered back-projection was used during reconstruction. Fiji 24 was used for 2D visualization and characterization of the pit, and Avizo software was used for 3D characterization of the pit. 25 In tomography, the top surface area of the pit is defined as the pit area. The width of a pit refers to the diameter of a circle which has the same area as the pit area. Pit depth was measured from the pit mouth to the pit bottom. Figure 1 shows a typical droplet after deposition ( Figure 1a) and exposure at 45 ± 2% RH and 21 • C for 48 h ( Figure 1b) in a lab-based experiment. The droplet after deposition was not perfectly circular and showed elongation along the grinding direction. After exposure, a single pit and rust can be observed under the droplet (Figure 1b). The surface of the droplet appears to have a wrinkled skin. The nature of the skin is not known as it is difficult to isolate. Following removal of the droplet and corrosion products, the pit was further examined in SEM and shown in Figure 1c. 20 out of 23 droplets showed pitting corrosion after exposure. Only one pit was found under each droplet. Figure 1c shows a pit with "spiral" growth, consistent with growth under cathodicallylimited conditions. 6,8 When the pit width is calculated, both the shallow and the deep part are taken into account.

Localized corrosion under MgCl 2 .-
Localized corrosion under mixed salt: MgCl 2 + NaCl.-Under mixed salt droplets containing MgCl 2 + NaCl, crystals form as the RH is reduced to 45% RH as shown in Figure 2a. The process where water evaporates from the salt solution is called efflorescence. The crystals are likely to be NaCl particles, given that the RH is in the range of the reported efflorescence relative humidity (ERH) of NaCl, which is 41% to 51% RH. [26][27][28][29][30][31] MgCl 2 is expected to remain in solution since the DRH of MgCl 2 (33%) is lower than the exposure RH. Therefore, solution layers can also be observed. Figure 2b shows the vertical section of an X-ray tomogram of a pin sample, deposited with a 1MgCl 2 :2NaCl droplet, after exposure at 45 ± 2% RH and 21 ± 1 • C for 17 h. It clearly shows the presence of crystals on top of the pin sample.
In lab-based experiments, mixed salt droplets were deposited onto 304 foils and exposed at 45 ± 2% RH and 21 • C for 24 h, 48 h and 72 h. 18 droplets were studied for 24 h or 72 h exposure time. 41 droplets were studied for 48 h exposure time. After exposure, all droplets showed pitting between or under crystals, or crevice corrosion under crystals. Figure 3 shows three typical forms of corrosion attack found under droplets after exposure. Figures 3a-3c show micrographs of selected regions of the three droplets after exposure for 48 h. Figures 3d-3f show the foil surfaces of the same droplets following washing with deionized water at the end of the test. Figures 3g-3i are SEM images that show the corrosion morphology of the same samples, illustrating the three typical types of corrosion behavior: a dish-shaped pit between crystals, shallow crevice-like attack under a crystal, and crevice-like attack containing a pit under a crystal. Figures 3a, 3d and 3g show a pit found between crystals. It is dish-shaped and is ∼100 μm in width and ∼50 μm in depth. The morphology of the pit is similar to the deep region of the pit in Figure 1. However, there is no indication of the large shallow region that is generally observed to develop initially during pit growth in MgCl 2 . 8 The reason for this is not clear.  Figures 3b, 3e and 3h show a crevice under a crystal: it is wide and shallow (∼200 μm in width and ∼20 μm in depth). The depths of four crevice-like attack sites after exposure for 48 h are from 5 to 30 μm. Most of the crevice-like attack was irregular and showed crystallographic etching (Figure 3h). This type of corrosion attack was only found under crystals, which were regarded as crevice-formers. Crevice-like corrosion on 304L under salt crystals was also reported by Cook 32 in his study of MgCl 2 and MgSO 4 mixtures. Figures 3c, 3f and 3i illustrate crevice-like attack containing a dish-shaped pit (∼200 μm in width and ∼35 μm in depth). The depths of three different sites of crevice-like attack containing dishshaped pits (after exposure for 48 h) varied from 10 to 35 μm. It was not clear whether the crevice-like attack or the dish-shaped pit formed first. The morphology of these pits is somewhat similar to the spiral pits observed under pure MgCl 2 . However, the crevice-like attack containing a dish-shaped pit was only found under crystals.
Pin samples under mixed salt droplets were imaged with X-ray microtomography, as shown in Figure 4. The left and right images are the horizontal and vertical sections of the tomogram, respectively. Figure 4a shows a dish-shaped pit between crystals. Figure 4b illustrates crevice-like attack under a crystal. Compared with the dishshaped pit shown in Figure 4a, the crevice-like attack found here is wider and shallower. Figure 4c shows crevice-like attack containing a pit under a crystal. Pitting corrosion under droplet electrolyte layers and crevice corrosion under crystals in synchrotron X-ray microtomography is consistent with that found in following lab-based experiments. However, it should be noted that multiple pits (more than four) were found on each of the pin samples in tomography experiments while the observation of single pits was most frequent in lab-based tests. The observation of multiple pits in tomography experiments might be due to the microstructure difference between the pin and foil samples. This observation might also indicate the presence of beam-induced pit initiation. 15,33,34 The interaction between an intense X-ray beam and water might produce radiolytic products, and affect the electrochemical processes. 34 Figure 5 gives a summary of the corrosion behavior on 304 foils under mixed salt droplets after exposure for 24, 48 or 72 h in labbased experiments. One single dish-shaped pit was most frequently observed under the experimental conditions studied here. All droplets contained at least one corrosion site (pit or crevice).
Since one single pit under a droplet after exposure was the most common observation for both MgCl 2 and 1MgCl 2 :2NaCl droplets, the widths and depths of these pits were studied further. Figure 6 is a summary of the widths and depths of 10 single pits, found under MgCl 2 droplets or 1MgCl 2 :2NaCl mixed salt droplets (under electrolyte layers between crystals). The overall width of pits observed under MgCl 2 are commonly greater than the dish-shaped pits under 1MgCl 2 :2NaCl because of the wide shallow region (Figure 1c) that forms initially in these droplets. 8 However, there is no obvious trend for pit depths.
Pits in mixed salt droplets are under electrolyte layers between crystals and therefore diffusion of ions away from the pit is expected to be decreased due to the diffusion barrier provided by the neighboring NaCl crystals. Therefore, an aggressive environment for pit initiation and propagation is more easily developed and maintained. The observation ( Figure 6) that the widths of the dish-shaped pits  under mixed salt droplets are generally smaller than those of the spiral pits under pure MgCl 2 droplets might be due to the smaller surface area available for the cathodic reaction (reduction of oxygen), 35 and increased solution resistance, again provided by the crystals, which will lead to a greater IR drop between the cathodic areas and the pit. Furthermore, while the initial volume and CDD of the droplets are the same for the mixed salt and pure MgCl 2 droplets, precipitation of NaCl crystals will remove most of the sodium ions and nearly half of the chloride ions from the electrolyte. This will reduce the volume of the electrolyte layer by nearly half compared with that for pure MgCl 2 solutions since the final exposure RH, which controls the concentration of the electrolyte layers, was the same. The decreased electrolyte volume/thinner electrolyte layer also leads to a greater IR drop. Previous studies have also reported that the pit width is IR-controlled 8,15,36 which may provide an explanation for the observation that the widths of the pits under mixed MgCl 2 + NaCl were smaller than those under pure MgCl 2 .
Pit growth with time.-Samples were examined at different exposure times: 3 h, 17 h, 34 h and 69 h or 79 h in tomography experiments. Figure 7a shows the growth of the dish-shaped pit shown in Figure 4a. There is no obvious sign of pitting after 3 h. After 17 h, a very small hole surrounded by a very shallow pit area can be observed from the vertical section. After 34 h, a dish-shaped pit was found. The pit grew between 34 h and 79 h. Figure 7b shows the pit size after exposure for 34 h and 79 h. Although there is little increase in pit width and pit depth, there is a substantial pit volume increase from ∼16000 μm 3 to 26000 μm 3 . Figure 8 shows the growth of crevice-like attack under the crystal shown in Figure 4b. After exposure for 3 h, there was no obvious crevice corrosion, while crystals formed. After exposure for 17 h, crevice-like attack was found. Both the width and depth of the crevice-like attack increased during the subsequent exposure. The pit volume changed greatly from 7300 μm 3 (after exposure for 17 h) to 38000 μm 3 (after exposure for 69 h). It is seen that growth, once initiated, continued throughout the experiment. According to the data in Figure 8, the square of pit depth is nearly linear with exposure time. This dependence is consistent with the idea that pit depth is under diffusion control, 36 but could also be associated with decreasing cathodic current. However, the data are too limited to draw any broader conclusion.

Localized corrosion behavior in realistic atmospheric
conditions.-The current study shows that different morphologiesdish-shaped pits and crevice-like attack -can both be observed in salt droplets in atmospheric conditions where mixed salts are present. This is a particular concern for surfaces that are under stress where there is a risk of environmentally-assisted cracking. The morphology of the pit/crevice-like attack might affect the susceptibility of the pit developing into a crack since the morphology might affect the stress/strain distribution. 37,38 This indicates the importance of studying the effect of composition variation as a function of relative humidity for conditions where mixed salts are found.

Conclusions
r Atmospheric corrosion of stainless steel has been investigated under salt droplets containing a mixture of NaCl and MgCl 2 at 21 ± 1 • C and 45 ± 2% RH. This value is above the deliquescence relative humidity of MgCl 2 and in a region where the formation of solid NaCl crystals is expected.
r Ex situ lab-based tests showed three types of corrosion attack: dish-shaped pits (usually) between NaCl crystals, crevice corrosion under NaCl crystals, and pits within crevices under NaCl crystals. The widths of the pits found between crystals were smaller than the pits observed in droplets containing only MgCl 2 with the same CDD. This can be attributed to an increased IR drop (due to the presence of crystals and a lower electrolyte volume) and a lower area for oxygen reduction for the mixed salt droplets.
r In situ X-ray microtomography can be used to monitor the pit growth on metal pins. The morphologies of pits in the tomography observations are consistent with lab-based measurements.