Corrosion of zinc in simulated soil at various degrees of saturation S_r was investigated by performing surface observations, corrosion depth analyses, electrochemical impedance spectroscopy, and potentiodynamic polarization tests. The aim was to evaluate the effects of soil moisture on the corrosion morphology and corrosion rate of zinc. At a high soil moisture content (S_r > 80%), the mean zinc corrosion rate was low and varied little as S_r varied. However, at a low moisture content (S_r < 80%), the mean corrosion rate increased markedly as the moisture content decreased and reached a maximum at S_r ∼ 50%. The zinc corrosion morphology became more heterogeneous, indicating that the corrosion depth increased, as S_r decreased. The effects of soil moisture on the zinc corrosion morphology and rate were assessed from changes in oxygen reduction (cathodic reaction) and zinc dissolution (anodic reaction) in simulated soil.

This Paper was Originally Published in Zairyo-to-Kankyo 73(1) (2024) 12–20. The abstract and captions of figures are slightly changed.

Fig. 3 Relationships between the degree of saturation S_r and (a) the corroded surface area S₁₀ and (b) the mean corrosion rate i_ave.

Metal structures including pipes, piles, transmission grids, and solar panel mounts that are fully or partly buried in soil need to be very resistant to corrosion to achieve the desired long operating lifespans. Steel is often used for such structures because it is strong, workable, and cost-effective. Galvanized or painted steel is more resistant than bare steel to corrosion. Galvanized steel is particularly widely used because it is cost-effective and corrosion resistant. Galvanized steel is more resistant than non-galvanized steel to corrosion because zinc (Zn) sacrificially protects the underlying steel and the Zn corrosion products form a barrier layer [1–4].

Much research into corrosion of Zn and galvanized steel in soil has been performed [5–13]. Metal corrosion in soil is very complex because numerous environmental factors are involved in complex combinations [14, 15]. Important soil characteristics are the chemical properties (e.g., pH and salts present), the pore structure (a physical property, which includes various characteristics such as void sizes, air content, and moisture content), and the types and concentrations of dissolved ions in the water contained in the soil. Denison et al. performed corrosion tests to investigate the corrosion resistance of galvanized steel sheets in various soil environments and found that galvanized steel was more resistant than bare steel to corrosion in alkaline soil [1]. Rossi et al. studied the effects of sulfate and chloride ions on corrosion of galvanized steel in soil and found that sulfate ions caused uniform corrosion but chloride ions caused localized corrosion. Corrosion was accelerated through synergistic effects of sulfate and chloride ions [16].

The moisture content of soil has a much stronger effect than other environmental factors on corrosion of metals in the soil [17, 18]. The soil moisture content controls the balance between the aqueous and air phases in the soil voids, and these control the ion conductivity and the rates of cathodic reactions, respectively, meaning that the moisture content is the dominant factor affecting corrosion of metals in the soil [19–23]. Hirata et al. studied the effect of the moisture content on corrosion of carbon steel in soil and found the highest corrosion rate at a moisture content of 90% [24, 25]. Azoor et al. found that cast iron had the highest corrosion rate at a moisture content that corresponded to the inflection point in the soil water retention curve and found that this value varied as the soil particle size varied [22]. However, few studies of the effect of the soil moisture content on corrosion of galvanized steel have been performed. Pereira et al. found that galvanized steel in soil collected in Brazil had a low corrosion rate at a low soil moisture content and a higher corrosion rate at a high soil moisture content [26]. However, only two moisture contents were used and the soil contained sulfate-reducing bacteria that accelerate corrosion, so it is not clear that the effect of the moisture content on the galvanized steel corrosion rate has been adequately assessed.

The aim of this study was to investigate the effect of the soil moisture content on the corrosion morphology and corrosion rate of Zn (the main component of the surface plating on galvanized steel) in soil. Corrosion and potentiodynamic polarization tests were performed using constant soil particle characteristics and chemical components but different soil moisture contents.

Zn plates (99.2% pure, 4 mm thick; The Nilaco Corporation, Tokyo, Japan) were used in the corrosion tests. The plates were cut into small pieces (10 mm long, 5 mm wide, 4 mm thick) using an electrical discharge machine. The Zn pieces were then embedded in parallel in epoxy resin with a 2 mm gap between each piece for the soil corrosion tests, as shown in Fig. 1(a). Additionally, no reference electrode (R.E.) was used to prevent contamination of the test solution [24, 25, 27]. Before a corrosion test was performed, the sample surfaces were wet polished with waterproof SiC abrasive paper (#2000) and then ultrasonically cleaned in ethanol for 5 min.

Fig. 1

Schematic of the experimental systems used in the (a) immersion tests and (b) polarization tests. (online color)

The artificial soil used in the tests was silica sand (Marutou, Gifu, Japan) with a mean particle size of 100 µm. A 3% NaCl solution was mixed with the artificial soil at a selected weight ratio to give simulated soil with the desired moisture content. The soil moisture content S_r was defined as the ratio between the volume of NaCl solution added and the total volume of voids in the soil. The volume of solution required to give S_r = 100% was determined by packing soil into a stainless steel pipe with filter paper at the bottom and allowing water to be absorbed through capillary action (Hilgard’s method). A 10 g aliquot of artificial soil required 2.94 mL of solution to reach S_r = 100%. The amounts of solution added to the soil for use in the corrosion tests were selected using this information to give S_r values between 30% and 100%.

A schematic of the cell used for the tests is shown in Fig. 1(a). A sample was placed at the bottom of an acrylic pipe with an inner diameter of 30 mm, then artificial soil with the selected S_r was placed over the sample so that the sample was buried under 10 mm of soil. The samples were kept at room temperature (25°C) for 14 d, with the top of each pipe covered with laboratory film (Parafilm M; Amcor, Zürich, Switzerland) to prevent water evaporating during the test. The Zn corrosion rate was determined by electrochemical impedance spectroscopy (EIS) each day. The test electrodes (Fig. 1(b)) were two Zn plates and a silver (Ag) plate (99.98% pure, 10 mm long, 3 mm wide; The Nilaco Corporation) 2 mm away from the Zn plates embedded in the epoxy resin. These plates were wet polished with waterproof SiC abrasive paper (#2000) and then ultrasonically cleaned in ethanol for 5 min. The Zn surfaces were then protected using NITOFLON fluororesin adhesive tape (Nitto Denko Corporation, Osaka, Japan), and the Ag plate was anodically oxidized in 1 kmol/m³ HCl at 0.4 mA/cm² for 30 min to cause silver chloride (AgCl) to form on the surfaces to allow the plate to act as the R.E. (in Fig. 1). EIS measurements were performed using the three-electrode cell, with one Zn plate acting as the working electrode (W.E. in Fig. 1) and the other acting as the counter electrode (C.E. in Fig. 1). An AC voltage with an amplitude of 10 mV_0–p was applied using a potentiostat/galvanostat with a frequency response analyzer (VSP; BioLogic, Seyssinet-Pariset, France) over the frequency range 100 kHz to 10 mHz.

After a corrosion test had been performed, the sample was removed from the cell and ultrasonically cleaned in ethanol for 5 min and then in 10% w/w NH₄Cl solution at 70°C for 5 min to remove soil particles and corrosion products from the electrode surfaces. The morphology of the Zn corrosion was then visualized using a VR-6000 three-dimensional optical profilometer (Keyence Corporation, Osaka, Japan) using a pitch of 11.8 µm.

Potentiodynamic polarization tests were performed using a three-electrode cell to investigate corrosion of Zn in artificial soil. The same samples used for the EIS measurements (Fig. 1(b)) were used in these tests. Similar to the corrosion tests described in Section 2.1, a cell was prepared containing artificial soil with a specified S_r and a platinum (Pt) plate was placed above the soil and used as the C.E.

Zn (the same as described in Section 2.1) or Pt (99.98% pure, 0.5 mm thick; The Nilaco Corporation) were used as the W.E. in the polarization tests. A test involved measuring the corrosion potential E_corr of one of the two Zn samples for 30 min before performing anodic polarization and then measuring the E_corr of the other sample for 15 min before performing cathodic polarization. The polarizations were performed using a VSP potentiostat/galvanostat (BioLogic) using a scan rate of 0.5 mV/s. Polarization tests were performed on samples buried in artificial soil with different S_r values immediately after the samples had been buried and 14 d later.

The surface corrosion morphologies, depth profiles, and binary images acquired after the corrosion tests using soil with various S_r values are shown in Fig. 2(a), 2(b), and 2(c), respectively. Each binary image was for an area in which corrosion had progressed ≥10 µm from the reference plane (the epoxy resin surface), determined from Fig. 2(b). As shown in Fig. 2, the Zn corrosion morphologies for high S_r values (80%–100%) and low S_r values (30%–70%) were markedly different. Little Zn corrosion occurred at high S_r values, and relatively shallow (maximum depth ∼70 µm) pit-like corroded areas were found. The number of corroded areas increased slightly as S_r decreased. However, the corroded area was much larger at low S_r values, and the corroded depths were up to five times larger than the corroded depths at high S_r values. The largest corroded area was found at S_r = 50%, and the corroded area was smaller at S_r = 30%.

Fig. 2

Corrosion morphologies, depth profiles, and thresholded binary images of Zn at 14 d after Zn plates were buried in soil at different degrees of saturation S_r. (online color)

The effect of S_r on the corrosion morphology was quantified using Fig. 2(b) and 2(c), and the results are shown in Fig. 3. The area ratio S₁₀, defined as the ratio between the area corroded to ≥10 µm deep and the geometric surface area S_geo of Zn, is plotted against S_r in Fig. 3(a). Extremely low S₁₀ values were found at high S_r values (80%–100%), and S₁₀ increased as S_r decreased below 70%, reaching a maximum at S_r = 50%. S₁₀ decreased as S_r decreased further. The volume of the corroded area, calculated from the corrosion depth distribution determined from Fig. 2(b), was divided by S_geo to give the mean corrosion rate (i_ave) over 14 d, assuming that dissolution of Zn to Zn²⁺ occurred at a current efficiency of 100%. i_ave is plotted against S_r in Fig. 3(b), and it can be seen that the relationship between i_ave and S_r was very similar to the relationship between S₁₀ and S_r. i_ave was very low (∼2 µA/cm²) and was relatively independent of S_r values of 80%–100% but increased markedly as the S_r value decreased below 70% and reached a maximum of almost 60 µA/cm² at S_r = 50%.

Fig. 3

Relationships between the degree of saturation S_r and (a) the corroded surface area S₁₀ and (b) the mean corrosion rate i_ave.

The Zn corrosion morphology was investigated further by creating a depth distribution of the corroded area defined in Fig. 2(c). The depth distribution is shown in Fig. 4. The depth distribution strongly depended on S_r. Most of the corroded area was shallow (10–20 µm), and the maximum depth was a little more than 70 µm at S_r > 80%. In contrast, the proportion of corroded areas >100 µm deep increased significantly as S_r decreased below 70%. The maximum corrosion depth d_max was determined from the depth distribution (Fig. 4), and d_max is plotted against S_r in Fig. 5. The relationship between S_r and d_max was very similar to the relationship between S₁₀ and i_ave shown in Fig. 3(a) and 3(b). d_max was <70 µm at S_r > 80% and markedly higher at S_r < 70%, reaching a maximum of ∼320 µm at S_r = 50%. These results clearly indicated that the Zn corrosion morphology and corrosion rate strongly varied as S_r varied.

Fig. 4

Histograms of the corrosion depth distributions determined from the corrosion depth images shown in Fig. 2(b).

Fig. 5

Relationship between the maximum corrosion depth d_max and the degree of saturation S_r.

The anodic and cathodic polarization curves acquired immediately after the samples were buried in soil at various S_r values are shown in Fig. 6. Both anodic and cathodic polarization were clearly strongly affected by S_r. The dissolved oxygen reduction reaction (ORR) and the Zn deposition reaction [28], which are involved in cathodic polarization, occurred near E_corr. The ORR, which is considered to be the dominant cathodic reaction involved in Zn corrosion in soil, had a diffusion-limiting current of ∼10 µA/cm² between −1.0 and −1.05 V vs. the Ag/AgCl electrode at S_r 80%–100%. However, markedly more ORR occurred as S_r decreased below 70%, and the maximum diffusion-limiting current was ∼3 mA/cm² at S_r = 50%. The cathodic polarization curves for Pt at the different S_r values are shown in Fig. 7. The cathodic polarization curves were measured using the same method as described above to investigate the certainty with which the ORR current was dependent on S_r. A clear diffusion-limiting current for ORR was found between −0.2 and −0.8 V vs. the Ag/AgCl electrode at S_r > 80%, and the magnitude was ∼10 µA/cm². In contrast, the diffusion-limiting current for the ORR increased as S_r decreased below 70% and reached a maximum of ∼2 mA/cm² at S_r = 50%. The relationships between the ORR diffusion-limiting current and S_r were similar for Pt and Zn. This indicates that the ORR diffusion-limiting current on the Zn surface was dependent on S_r and that the ORR rate increased markedly as S_r decreased below 70%.

Fig. 6

Potentiodynamic polarization curves for Zn in soil at different degrees of saturation S_r: comparison of polarization behavior measured immediately after Zn plates were buried in soil at different S_r values. (online color)

Fig. 7

Potentiodynamic cathodic polarization curves for Pt in soil at different degrees of saturation S_r: comparison of polarization behavior measured immediately after Pt was buried in soil at different S_r values. (online color)

The anodic reaction of Zn at S_r > 80% caused typical polarization behavior found in a NaCl solution, with little dependence on S_r (Fig. 6). However, the anodic polarization curves shifted and the overall anodic current markedly decreased as S_r decreased below 70%. The corroded area depended on S_r (Figs. 2 and 3), but it was unclear whether the measured anodic current caused by Zn dissolution was affected by changing S_r. The Zn corrosion morphology after anodic polarization was therefore assessed from the corrosion depth profile.

When Zn in soil is anodically polarized, the areas on the Zn in which dissolution reactions occur (i.e., the areas in contact with the solution) become evident because anodic polarization enhances Zn dissolution [18]. A binary image of a dissolution area when anodic polarization was applied from E_corr to the potential when a current of 5 mA passed through the sample is shown in Fig. 8(a). The binarization threshold was the epoxy resin surface in the depth distribution, and the dissolved area is shown in black in the figure. At S_r > 80%, dissolution occurred over the whole surface. However, dissolution areas were heterogeneously distributed on the Zn surface at S_r < 70%. The ratio between the contact area and S_geo for Zn (called S_anode) was calculated and is plotted against S_r in Fig. 8(b). S_anode was ∼80% when S_r was 90%–100%, and S_anode decreased linearly as S_r decreased. These results indicated that the contact (wetted) area and corroded depth distribution were markedly different at S_r ∼ 80%. The anodic polarization behavior shown in Fig. 6 therefore reflected a marked change in the contact area between the Zn surface and NaCl solution, i.e., the dissolution area decreased. The ratio between S_anode at S_r = 30% and S_r = 100% was about 0.25, and the anodic current at S_r = 30% was little more than a tenth of the current at S_r = 100%. Therefore, as the dissolution area decreased, the water film on the electrode surface also became thinner and anodic Zn dissolution might have been suppressed [29, 30].

Fig. 8

(a) Thresholded binary images of the corroded areas of Zn after anodic polarization in soil at different degrees of saturation S_r, and (b) changes in the corroded surface area on Zn determined from the binarized images plotted against S_r.

The anodic and cathodic polarization curves for Zn in soil at S_r = 90%, S_r = 70%, and S_r = 50% measured immediately after the samples were buried and 14 d later are shown in Fig. 9(a)–9(c). The cathodic polarization behaviors of Zn after 14 d were markedly different at the different S_r values. At S_r = 90%, the ORR current was lower after 14 d than immediately after burial. Most of the voids in the soil were filled with solution at S_r = 90%. Convection is less likely to occur in a solution in voids than in a bulk solution, so the oxygen-diffusion layer was thicker toward the top of the soil because O₂ was consumed during the corrosion reaction [24, 25]. More O₂ would have been present in the voids immediately after burial at S_r = 70% than at S_r = 90%, so the ORR diffusion-limiting current would have been higher because more O₂ would have been supplied from voids near the electrodes at S_r = 70% than at S_r = 90%. O₂ in voids immediately after burial would have been consumed in 14 d regardless of the S_r. The ORR diffusion-limiting currents at 14 d at S_r = 70% and S_r = 90% were almost the same, suggesting that the oxygen-diffusion layer became thicker from the Zn surface toward the top of the soil and was similar at S_r = 70% and S_r = 90%. In other words, there was no continuous air phase (O₂ supply path) from the electrode/solution interface to the top of the soil even at S_r = 70%, and the cathodic reaction was markedly suppressed at 14 d. Decreasing S_r to 50% caused there to be more air in the voids, meaning that an O₂ supply path from the electrode/solution interface to the top of the soil was present, and there was almost no difference between the cathodic polarization behaviors immediately after burial and 14 d later (Fig. 9(c)).

Fig. 9

Potentiodynamic polarization curves for Zn in soil at different degrees of saturation S_r: (a)–(c) polarization measured immediately after Zn plates were buried in soil at different S_r values and 14 d later. (online color)

The anodic reaction of Zn was suppressed at 14 d compared with immediately after burial at all S_r values, probably because of corrosion products accumulating on the Zn surface.

The relationship between E_corr and S_r is shown in Fig. 10. E_corr was ∼−1 V vs. the Ag/AgCl electrode, which is typical for Zn in a NaCl solution, and the values immediately after burial and 14 d later were the same at S_r > 80%. However, E_corr increased as S_r decreased below 70%, and E_corr was higher at 14 d than immediately after burial. The increases in E_corr immediately after burial as S_r decreased were attributed to more cathodic reactions occurring. In contrast, the stronger increase in E_corr at 14 d than immediately after burial at a low S_r was probably caused by suppression of both anodic and cathodic reactions and particularly strong suppression of the anodic reaction, presumably due to the effects of corrosion products.

Fig. 10

Relationships between the degree of saturation S_r and the Zn corrosion potential immediately after Zn plates were buried (○) and 14 d later (△).

The effects of S_r on Zn corrosion morphology were investigated by performing corrosion tests with Zn plates in artificial soil. The results indicated that the corrosion morphologies at S_r > 80% and S_r < 70% were markedly different. Little Zn corrosion occurred at S_r > 80% but relatively severe localized corrosion occurred at S_r < 70%. Here, we discuss the effect of S_r on corrosion morphology.

The cathodic polarization curves (Figs. 6 and 7) indicated that the ORR diffusion-limiting current sharply increased at S_r ∼ 70%, indicating that more O₂ was supplied to the Zn surface at S_r < 70% than at S_r > 70%. This was probably because the less NaCl solution was present the more air (containing O₂) would fill the voids and this made a marked difference at S_r ∼ 70% because of the physicochemical properties and particle size (and therefore packing structure) of the soil used in this study. In saturated soil (S_r = 100%), dissolved O₂ in the soil solution needs to be transported to the Zn surface through diffusion for corrosion to occur because the soil voids are filled with NaCl solution. However, air in the voids will create a path for fast transport of O₂ to the Zn surface, which will increase the ORR diffusion-limiting current. A decrease in S_r will increase the proportion of air in the voids and therefore increase the rate at which O₂ diffuses to the Zn surface.

In contrast, S_anode decreased as S_r decreased (Fig. 8(b)), suggesting that the Zn surface area wetted with NaCl solution decreased. Even though the rate at which O₂ diffused to the Zn surface was expected to increase, the ORR diffusion-limiting current decreased as S_r decreased below 50%. These results indicated that decreasing the volume of NaCl solution in the soil voids caused the amount of continuous liquid phase available for ionic conduction to decrease and the wetted area of the electrode/solution interface to decrease.

Azoor et al. studied the effects of the soil texture and S_r on the corrosion rate of cast iron in soil. They measured corrosion currents and potentials in tests using sand, clay, and silt with different particle sizes at various S_r values [21, 22]. The results indicated that the corrosion rate reached a maximum at a particular S_r for each soil type and the S_r value at which the corrosion rate reached a maximum increased as the soil particle size decreased. They explained this in relation to the ratio between the air and liquid phases and assessed the relationship between the changes in S_r at which the corrosion rate reached a maximum and soil texture and particle structure. They found that larger particles caused larger voids, which decreased water retention and decreased the S_r at which the corrosion rate reached a maximum. The corrosion rate of cast iron in sand with particles of similar sizes to the artificial soil we used reached a maximum at S_r ∼ 50%. This suggested that the relationship between Zn corrosion morphology and S_r could be explained through a similar mechanism.

The effect of S_r on the Zn corrosion rate was evaluated using the corrosion depth data from the corrosion tests. The results indicated that the corrosion rate depended on S_r and reached a maximum at S_r = 50%, which is discussed below.

Corrosion of Zn immersed in a neutral aqueous solution is generally considered to proceed through coupling of anodic dissolution of Zn (eq. (1)) and the ORR (eq. (2)), with the ORR being diffusion-limiting and determining the corrosion rate. Corrosion in soil can be considered similar to corrosion in an aqueous solution. The ORR diffusion-limiting current was almost unaffected by S_r at S_r > 80% (Figs. 6 and 7), suggesting that the soil was almost saturated, meaning the solution filled most of the voids in the soil. O₂ would have diffused through the solution to be supplied to the Zn surface. Zn dissolution occurred over the entire Zn surface through anodic polarization, and the corroded area was rather uniform at S_r > 80% (Fig. 8(a)), suggesting that the whole surface was in contact with the solution. The Zn corrosion rate would have been determined by the O₂ diffusion rate in the solution in the voids at S_r > 80%.

The initial cathodic polarization curves measured immediately after the Zn plates were buried in soil indicated that the ORR diffusion-limiting current was ∼10 µA/cm². This suggested that Zn corrosion started at this rate. The inverse of the impedance at 10 mHz (Z⁻¹_10mHz) and the impedance at 100 kHz (Z_100kHz) are plotted against time for the tests at various S_r values in Fig. 11. The impedances were derived from the EIS measurements for the samples used in the polarization tests (Fig. 1(b)). At S_r = 90% and S_r = 100%, Z_100kHz was small and constant throughout the test period, indicating that the ion conductivity and contact interface between the Zn surface and NaCl solution did not change. Z⁻¹_10mHz, which is considered to indicate the corrosion rate, was extremely small (in the 10⁻⁴ Ω⁻¹cm⁻² range) and remained constant for 14 d, suggesting that the corrosion rate hardly changed during the test. This was consistent with the results of the potentiodynamic polarization tests (Fig. 9(a)). However, Z⁻¹_10mHz was higher at lower S_r values than at higher S_r values, particularly in the early stages of the tests. This corresponded to the i_ave (Fig. 3(b)) increasing as S_r decreased at lower S_r values.

Fig. 11

Transient electrochemical impedances for Zn buried in soil at different degrees of saturation S_r measured at (a) 100 kHz and (b) 10 mHz. (online color)

Z_100kHz increased as S_r decreased, particularly at S_r < 70%. This suggested that the proportion of air phase in the soil was higher at S_r < 70% than at S_r > 70%. At S_r < 70%, Z_100kHz gradually increased up to 5 d and then remained constant. In contrast, Z⁻¹_10mHz gradually decreased for 5 d and then remained constant. The increase in Z_100kHz indicated that the solution (soil) resistance increased or the contributions of the resistances of the corrosion products increased. The initial changes in Z⁻¹_10mHz at low S_r values indicated that i_ave was high, at least immediately after the Zn plate had been buried. Z_100kHz remained constant at a high S_r, indicating that almost no changes occurred in the soil for 14 d. This indicated that Zn corrosion in the initial stages of the tests probably led to the formation of corrosion products and that the resistances of the corrosion products were probably included in Z_100kHz at low S_r values. We therefore concluded that Z⁻¹_10mHz became low and constant after 5 d because of the formation of corrosion products. As shown in Fig. 9(c), the anodic reaction was suppressed at 14 d. The decrease in corrosion rate at between S_r = 30% and S_r = 50% was therefore not caused by suppression of the cathodic reaction by growth of the oxygen-diffusion layer but by suppression of the anodic reaction by the corrosion products.

The effects of S_r on Zn corrosion in soil determined from the results are shown schematically in Fig. 12. At S_r > 80%, the Zn surface would have been almost entirely in contact with the liquid phase and O₂ would have needed to diffuse through this phase to corrode the Zn. The low diffusion rate would have caused the corrosion rate to be low, so little corrosion occurred. Conversely, as the proportion of air in the soil increased and the wetted area between the Zn surface and NaCl solution decreased at S_r < 70%, the anodic dissolution site area decreased and localized dissolution occurred. Despite increased O₂ diffusion, the decreased wetted area and liquid volume on the Zn surface inhibited diffusion of Zn²⁺, so the corrosion rate decreased as S_r decreased further. The corrosion depth and localized corrosion rate reached maxima at S_r ∼ 50%.

Fig. 12

Schematic of the effects of the degree of saturation on the Zn corrosion morphology and corrosion rate. (online color)

If the S_r value at which the maximum corrosion rate was controlled by the O₂ diffusion rate in the soil, different metals such as Fe and Zn would have maximum corrosion rates at the same S_r. Hirata et al. performed similar experiments and made EIS measurements and surface morphology observations to investigate the effect of S_r on corrosion of carbon steel (SM490A) [24, 25]. They found that uniform corrosion occurred on carbon steel at S_r = 100% (saturated soil) and that non-uniform corrosion occurred in unsaturated soil (S_r < 100%). The maximum corrosion rate was found at S_r = 90%. The soil properties were similar and S_r was controlled, so the difference between the S_r values at which the corrosion rates reached maximum in that study and our study was probably caused by the different characteristics of the corrosion products.

Besides, the wetted area at the electrode/soil interface became more unevenly distributed as S_r decreased (Fig. 12(b)), and there were probably areas with electrolyte films of different thicknesses. O₂ diffused readily in thin electrolyte film areas, so these areas became preferential cathodic reaction sites. Thick electrolyte areas became anodic reaction sites. Cathodic reaction sites on carbon steel tend to passivate because of alkalinization associated with the ORR, making it unlikely that such sites change quickly into anodic reaction sites. However, even cathodic reaction sites on a Zn surface that become alkaline can still become anodic reaction sites because Zn is amphoteric. This suggests that the corrosion property of a metal may affect the S_r dependence of corrosion.

The effect of S_r on Zn corrosion in soil was investigated, and the following conclusions were drawn.

• 1)   I.A.  Denison and  M.  Romanoff: Corrosion of galvanized steel in soils, J. Res. Natl. Bur. Stand. 49 (1952) 299–314. doi:10.6028/jres.049.031
• 2)  M. Romanoff: Underground Corrosion, (National Bureau of Standards, United States, 1957) p. 107.
• 3)   I.  Sekine,  T.  Kobayashi and  H.  Suda: Corrosion Protective Property of Various Covering Materials of Steels in Soil, Boshoku Gijutsu 36 (1987) 3–9. doi:10.3323/jcorr1974.36.1_3
• 4)  X.G. Zhang: Corrosion and Electrochemistry of Zinc, (Plenum Press, United States, 1996) pp. 305–312.
• 5)   A.  Farooq,  U.  Masood Chaudry,  A.  Saleem,  K.  Mairaj Deen,  K.  Hamad and  R.  Ahmad: Sacrificial Dissolution of Zinc Electroplated and Cold Galvanized Coated Steel in Saline and Soil Environments: A Comparison, Materials (Basel) 14 (2021) 744. doi:10.3390/ma14040744
• 6)   L.-D.  Arriba-Rodríguez,  F.  Ortega-Fernández,  J.M.  Villanueva-Balsera,  V.  Rodríguez-Montequín and  A.  Bueno: Corrosion Predictive Model in Hot-Dip Galvanized Steel Buried in Soil, Complexity 2021 (2021) 9275779. doi:10.1155/2021/9275779
• 7)   V.  Padila,  P.  Ghods and  A.  Alfantazi: Parametric Studies and Application of a Practical Model for Corrosion of Galvanized Steel in Soil, Corrosion 70 (2014) 1189–1202. doi:10.5006/1284
• 8)   D.  Nakhaie,  A.  Kosari,  J.M.C.  Mol and  E.  Asselin: Corrosion resistance of hot-dip galvanized steel in simulated soil solution: A factorial design and pit chemistry study, Corros. Sci. 164 (2020) 108310. doi:10.1016/j.corsci.2019.108310
• 9)   I.M.F.  Lopes,  C.R.O.  Loureiro and  R.M.R.  Junqueira: Corrosion monitoring of galvanized steel in soil extract solutions by electrochemical impedance spectroscopy, Materialwiss. Werkstofftech. 45 (2014) 619–627. doi:10.1002/mawe.201400243
• 10)   N.  Palsson,  P.  Khamsuk,  S.  Sorachot,  K.  Wongpinkaew,  W.  Thaiwirot,  K.  Phaebua and  P.  Klomjit: Corrosion behavior of zinc and copper coated structural steels in soil environments, Materialwiss. Werkstofftech. 53 (2022) 68–79. doi:10.1002/mawe.202100145
• 11)   U.  Nürnberger: Long-time behavior of non-galvanized and galvanized steels for geotechnical stabilization applications, Mater. Corros. 63 (2012) 1173–1180. doi:10.1002/maco.201206680
• 12)   C.  Soriano and  A.  Alfantazi: Corrosion behavior of galvanized steel due to typical soil organics, Constr. Build. Mater. 102 (2016) 904–912. doi:10.1016/j.conbuildmat.2015.11.009
• 13)   S.  Karthick,  S.  Muralidharan and  V.  Saraswathy: Corrosion performance of mild steel and galvanized iron in clay soil environment, Arab. J. Chem. 13 (2020) 3301–3318. doi:10.1016/j.arabjc.2018.11.005
• 14)   I.S.  Cole and  D.  Marney: The science of pipe corrosion: A review of the literature on the corrosion of ferrous metals in soils, Corros. Sci. 56 (2012) 5–16. doi:10.1016/j.corsci.2011.12.001
• 15)   S.  Tsunoda and  T.  Akiba: Some Problems for Evaluating Soil Aggresivity, Boshoku Gijutsu 36 (1987) 168–177. doi:10.3323/jcorr1974.36.3_168
• 16)   S.  Rossi,  M.  Pinamonti and  M.  Calovi: Influence of soil chemical characteristics on corrosion behaviour of galvanized steel, Case Stud. Constr. Mater. 17 (2022) e01257. doi:10.1016/j.cscm.2022.e01257
• 17)   S.K.  Gupta and  B.K.  Gupta: The critical soil moisture content in the underground corrosion of mild steel, Corros. Sci. 19 (1979) 171–178. doi:10.1016/0010-938X(79)90015-5
• 18)   J.  Jiang,  J.  Wang,  W.W.  Wang and  W.  Zhang: Modeling influence of gas/liquid/solid three-phase boundary zone on cathodic process of soil corrosion, Electrochim. Acta 54 (2009) 3623–3629. doi:10.1016/j.electacta.2009.01.033
• 19)   M.  Aachib,  M.  Mbonimpa and  M.  Aubertin: Measurement and Prediction of the Oxygen Diffusion Coefficient in Unsaturated Media, with Applications to Soil Covers, Water Air Soil Pollut. 156 (2004) 163–193. doi:10.1023/B:WATE.0000036803.84061.e5
• 20)   R.  Akkouche,  C.  Rémazeilles,  M.  Jeannin,  M.  Barbalat,  R.  Sabot and  P.  Refait: Influence of soil moisture on the corrosion processes of carbon steel in artificial soil: Active area and differential aeration cells, Electrochim. Acta 213 (2016) 698–708. doi:10.1016/j.electacta.2016.07.163
• 21)   R.M.  Azoor,  R.N.  Deo,  N.  Birbilis and  J.K.  Kodikara: Coupled Electro-Chemical-Soil Model to Evaluate the Influence of Soil Aeration on Underground Metal Pipe Corrosion, Corrosion 74 (2018) 1177–1191. doi:10.5006/2860
• 22)   R.M.  Azoor,  R.N.  Deo,  N.  Birbilis and  J.  Kodikara: On the optimum soil moisture for underground corrosion in different soil types, Corros. Sci. 159 (2019) 108116. doi:10.1016/j.corsci.2019.108116
• 23)   R.  Akkouche,  C.  Rémazeilles,  M.  Barbalat,  R.  Sabot,  M.  Jeannin and  P.  Refait: Electrochemical Monitoring of Steel/Soil Interfaces during Wet/Dry Cycles, J. Electrochem. Soc. 164 (2017) C626–C634. doi:10.1149/2.0531712jes
• 24)   R.  Hirata,  A.  Ooi,  E.  Tada and  A.  Nishikata: Influence of the degree of saturation on carbon steel corrosion in soil, Corros. Sci. 189 (2021) 109568. doi:10.1016/j.corsci.2021.109568
• 25)   R.  Hirata,  W.  Yonemoto,  A.  Ooi,  E.  Tada and  A.  Nishikata: Influence of Soil Particle Size, Covering Thickness, and pH on Soil Corrosion of Carbon Steel, ISIJ Int. 60 (2020) 2533–2540. doi:10.2355/isijinternational.ISIJINT-2020-261
• 26)   R.F.  da Costa Pereira,  E.S.D.  de Oliveira,  M.A.G.  de Andrade Lima and  S.L.D.C.  Brasil: Corrosion of Galvanized Steel Under Different Soil Moisture Contents, Mat. Res. 18 (2015) 563–568. doi:10.1590/1516-1439.341714
• 27)   A.  Nishikata: IV. Electrochemical Measurements in Various Environments—Soil Environments—, Zairyo-to-Kankyo 67 (2018) 362–367. doi:10.3323/jcorr.67.362
• 28)  In the potentiodynamic polarization tests, one of the two adjacent Zn electrodes shown in Fig. 1(b) was first subjected to anodic polarization and the other was subsequently subjected to cathodic polarization. The reduction (deposition) current for Zn²⁺, which would have dissolved in the soil moisture during anodic polarization, was therefore determined during cathodic polarization.
• 29)   A.  Nishikata,  Y.  Ichihara and  T.  Tsuru: An application of electrochemical impedance spectroscopy to atmospheric corrosion study, Corros. Sci. 37 (1995) 897–911. doi:10.1016/0010-938X(95)00002-2
• 30)   A.  Nishikata,  Y.  Ichihara,  Y.  Hayashi and  T.  Tsuru: Influence of Electrolyte Layer Thickness and pH on the Initial Stage of the Atmospheric Corrosion of Iron, J. Electrochem. Soc. 144 (1997) 1244–1252. doi:10.1149/1.1837578